<?xml version="1.0" encoding="UTF-8"?><TEI xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.tei-c.org/ns/1.0 nzetc-p5.xsd" xml:id="BucMani" xml:lang="en">
  <teiHeader>
    <fileDesc xml:id="fileDesc-0001">
      <titleStmt>
        <title type="marc245"><name key="name-101721" type="work">Ethnology of Manihiki and Rakahanga</name></title>
        <title type="sort">Ethnology of Manihiki and Rakahanga</title>
        <title type="gmd">[electronic resource]</title>
        <author><name key="name-202886" type="person">Te Rangi Hiroa (Sir Peter Henry Buck)</name></author>
        <respStmt xml:id="respStmt-0001">
          <resp>Creation of machine-readable version</resp>
          <name key="name-121582" type="organisation">TechBooks, Inc.</name>
        </respStmt>
        <respStmt xml:id="respStmt-0002">
          <resp>Creation of digital images</resp>
          <name key="name-141366" type="person">Samantha Callaghan</name>
        </respStmt>
        <respStmt xml:id="respStmt-0003">
          <resp>Conversion to TEI.2-conformant markup</resp>
          <name key="name-121582" type="organisation">TechBooks, Inc.</name>
        </respStmt>
      </titleStmt>
      <extent>ca. 809 kilobytes</extent>
      <publicationStmt>
        <publisher><name key="name-121602" type="person">New Zealand Electronic Text Centre</name></publisher>
        <pubPlace><name key="name-008844" type="place">Wellington</name>, New Zealand</pubPlace>
        <idno type="etc">Modern English, BucMani</idno>
        <availability status="unknown">
          <p>Publicly accessible</p>
          <p n="public">URL: http://www.nzetc.org/collections.html</p>
          <p>copyright 2006, by Victoria University of Wellington</p>
        </availability>
        <date when="2006">2006</date>
      <idno type="vuw-bbid">1008234</idno></publicationStmt>
      <seriesStmt xml:id="seriesStmt-0001">
        <title><name key="name-101719" type="work">Bernice P. Bishop Museum Bulletin</name></title>
        <idno type="vol">99</idno>
      </seriesStmt>
      <notesStmt>
        <note xml:id="note-0001">Line breaks have only been retained for non-prose elements.</note>
      </notesStmt>
      <sourceDesc xml:id="sourceDesc-0001">
        <biblFull>
          <titleStmt>
            <title><name key="name-101721" type="work">Ethnology of Manihiki and Rakahanga</name></title>
            <author><name key="name-202886" type="person">Te Rangi Hiroa (Sir Peter Henry Buck)</name></author>
          </titleStmt>
          <publicationStmt>
            <pubPlace>Honolulu, Hawaii</pubPlace>
            <publisher><name key="name-101718" type="organisation">Bernice P. Bishop Museum</name></publisher>
            <date when="1932">1932</date>
            <idno type="callno">Source copy consulted: Victoria University of Wellington, Q11 B5 B 99</idno>
          </publicationStmt>
        </biblFull>
      </sourceDesc>
    </fileDesc>
    <encodingDesc>
      <editorialDecl>
        <p>All unambiguous end-of-line hyphens have been removed, and
          the trailing part of a word has been joined to the preceding
          line.</p>
        <p xml:id="ETC">Some keywords in the header are a local Electronic
          Text Centre scheme to aid in establishing analytical
          groupings.</p>
      </editorialDecl>
      <classDecl>
        <taxonomy xml:id="nzetc-subjects">
          <bibl>
            <title>NZETC Subject Headings</title>
          </bibl>
        </taxonomy>
      </classDecl>
    </encodingDesc>
    <profileDesc xml:id="profileDesc-0001">
      <langUsage>
        <language ident="en">English</language>
      </langUsage>
      <textClass>
        <keywords scheme="http://www.nzetc.org/nzetc-subjects">
          <list>
            <item><rs key="subject-000010" type="subject">Historical Māori and Pacific Islands</rs></item>
          </list>
        </keywords>
      </textClass>
    </profileDesc>
    <revisionDesc>
      <change n="quickProof"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Text-proofing of a sample of the text</change>
      <change n="teiMarkup"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Conversion to TEI.2-conformat markup</change>
      <change n="scriptedMarkup"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Adding scripted markup</change>
      <change n="encodingDesc"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Addition of encodingDesc</change>
      <change n="addBibls"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Addition of bibls</change>
      <change n="assembleImages"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Assembled all images</change>
      <change n="derivativeCreation"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Creation of derivative images</change>
      <change n="teiValidation"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Validation of TEI</change>
      <change n="nameValidation"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Validation of names</change>
      <change n="utf8Conversion"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Conversion to Unicode (utf-8)</change>
      <change n="makeProduction"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Promotion to production</change>
      <change n="drmAddition"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Addition of text to access control</change>
      <change n="harvestTopicMap"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Harvest into Topic Map</change>
      <change n="browserCheck"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Checking of text using browser</change>
      <change n="corpusAddition"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Addition of text to corpus</change>
      <change n="catalogueAddition"><date when="2007-08-07T21:17:48">21:17:48, Tuesday 7 August 2007</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Addition of text to Library Catalogue<!-- BBID=1008234 --></change>
      <change n="live"><date when="2008-09-23T14:47:07">14:47:07, Tuesday 23 September 2008</date><label>editorial</label><name type="organisation" key="name-121602">NZETC</name>Make text available on NZETC website</change>
    <change n="epubPreparation"><date when="2009-08-04T14:07:12">14:07:12, Tuesday 4 August 2009</date><name type="organisation" key="name-121602">NZETC</name>Preparation of EPUB (and other formats such as DaisyBook)</change></revisionDesc>
  </teiHeader>
  <text xml:id="t1">
    <front xml:id="t1-front">
      <div xml:id="t1-front-d1" type="covers">
        <p>
          <figure xml:id="BucManiFCo">
            <graphic url="BucManiFCo.jpg" mimeType="image/jpeg" xml:id="BucManiFCo-g"/>
            <figDesc>Front Cover</figDesc>
          </figure>
        </p>
        <p>
          <figure xml:id="BucManiSpi">
            <graphic url="BucManiSpi.jpg" mimeType="image/jpeg" xml:id="BucManiSpi-g"/>
            <figDesc>Spine</figDesc>
          </figure>
        </p>
        <p>
          <figure xml:id="BucManiBCo">
            <graphic url="BucManiBCo.jpg" mimeType="image/jpeg" xml:id="BucManiBCo-g"/>
            <figDesc>Back Cover</figDesc>
          </figure>
        </p>
        <p>
          <figure xml:id="BucManiTit">
            <graphic url="BucManiTit.jpg" mimeType="image/jpeg" xml:id="BucManiTit-g"/>
            <figDesc>Title Page</figDesc>
          </figure>
        </p>
      </div>
      <pb xml:id="n1"/>
      <pb xml:id="n2"/>
      <pb xml:id="n3"/>
      <titlePage xml:id="t1-front-d1-d1">
        <docTitle>
          <titlePart type="main">ETHNOLOGY OF MANIHIKI<lb/>
AND RAKAHANGA</titlePart>
        </docTitle>
        <byline>BY<lb/>
<docAuthor>TE RANGI HIROA<lb/>
(PETER H. BUCK)</docAuthor></byline>
        <docImprint><hi rend="sc">Bernice P. Bishop Museum<lb/>
Bulletin</hi> 99<lb/>
<pubPlace><hi rend="lsc">Honolulu, Hawaii</hi></pubPlace><lb/>
<publisher><hi rend="sc">Published by the Museum</hi></publisher><lb/>
<date when="1932">1932</date></docImprint>
      </titlePage>
      <pb xml:id="n4"/>
      <pb xml:id="n5" n="[i]"/>
      <div xml:id="t1-front-d2" type="contents">
        <head><hi rend="c">Contents</hi></head>

          <table rows="135" cols="2">
            <row>
              <cell/>
              <cell rend="right"><hi rend="sc">Page</hi>
              </cell>
            </row>
            <row>
              <cell>Introduction</cell>
              <cell rend="right"><ref target="#n11">3</ref></cell>
            </row>
            <row>
              <cell> Acknowledgments</cell>
              <cell rend="right"><ref target="#n11">3</ref></cell>
            </row>
            <row>
              <cell> Culture</cell>
              <cell rend="right"><ref target="#n12">4</ref></cell>
            </row>
            <row>
              <cell> Geography</cell>
              <cell rend="right"><ref target="#n12">4</ref></cell>
            </row>
            <row>
              <cell> Contact with Western culture</cell>
              <cell rend="right"><ref target="#n13">5</ref></cell>
            </row>
            <row>
              <cell> The people</cell>
              <cell rend="right"><ref target="#n19">11</ref></cell>
            </row>
            <row>
              <cell> Language</cell>
              <cell rend="right"><ref target="#n19">11</ref></cell>
            </row>
            <row>
              <cell>Traditional history</cell>
              <cell rend="right"><ref target="#n22">14</ref></cell>
            </row>
            <row>
              <cell> Discovery</cell>
              <cell rend="right"><ref target="#n22">14</ref></cell>
            </row>
            <row>
              <cell> Genealogical records</cell>
              <cell rend="right"><ref target="#n28">20</ref></cell>
            </row>
            <row>
              <cell> Marriages of Toa</cell>
              <cell rend="right"><ref target="#n31">23</ref></cell>
            </row>
            <row>
              <cell>Social Organization</cell>
              <cell rend="right"><ref target="#n36">28</ref></cell>
            </row>
            <row>
              <cell> Family pedigrees</cell>
              <cell rend="right"><ref target="#n36">28</ref></cell>
            </row>
            <row>
              <cell> Relationship terms</cell>
              <cell rend="right"><ref target="#n37">29</ref></cell>
            </row>
            <row>
              <cell> The biological family</cell>
              <cell rend="right"><ref target="#n44">36</ref></cell>
            </row>
            <row>
              <cell>  The household</cell>
              <cell rend="right"><ref target="#n44">36</ref></cell>
            </row>
            <row>
              <cell>  Birth</cell>
              <cell rend="right"><ref target="#n45">37</ref></cell>
            </row>
            <row>
              <cell>  Childhood</cell>
              <cell rend="right"><ref target="#n47">39</ref></cell>
            </row>
            <row>
              <cell>  Puberty</cell>
              <cell rend="right"><ref target="#n48">40</ref></cell>
            </row>
            <row>
              <cell>  Marriage</cell>
              <cell rend="right"><ref target="#n50">42</ref></cell>
            </row>
            <row>
              <cell> Titular system</cell>
              <cell rend="right"><ref target="#n51">43</ref></cell>
            </row>
            <row>
              <cell>  History of the single and dual ariki titles</cell>
              <cell rend="right"><ref target="#n51">43</ref></cell>
            </row>
            <row>
              <cell>  The Whainga-aitu ariki</cell>
              <cell rend="right"><ref target="#n56">48</ref></cell>
            </row>
            <row>
              <cell>  The Whakaheo ariki</cell>
              <cell rend="right"><ref target="#n60">52</ref></cell>
            </row>
            <row>
              <cell>  Other titles</cell>
              <cell rend="right"><ref target="#n62">54</ref></cell>
            </row>
            <row>
              <cell> Tribal history</cell>
              <cell rend="right"><ref target="#n65">57</ref></cell>
            </row>
            <row>
              <cell>  First divisions</cell>
              <cell rend="right"><ref target="#n65">57</ref></cell>
            </row>
            <row>
              <cell>  The tribe</cell>
              <cell rend="right"><ref target="#n67">59</ref></cell>
            </row>
            <row>
              <cell>  Subtribes</cell>
              <cell rend="right"><ref target="#n69">61</ref></cell>
            </row>
            <row>
              <cell> Population and land</cell>
              <cell rend="right"><ref target="#n73">65</ref></cell>
            </row>
            <row>
              <cell>Material culture</cell>
              <cell rend="right"><ref target="#n78">70</ref></cell>
            </row>
            <row>
              <cell>  Houses and furniture</cell>
              <cell rend="right"><ref target="#n78">70</ref></cell>
            </row>
            <row>
              <cell>  Introduction</cell>
              <cell rend="right"><ref target="#n78">70</ref></cell>
            </row>
            <row>
              <cell>  Original house type</cell>
              <cell rend="right"><ref target="#n78">70</ref></cell>
            </row>
            <row>
              <cell>   Framework</cell>
              <cell rend="right"><ref target="#n78">70</ref></cell>
            </row>
            <row>
              <cell>   Thatch</cell>
              <cell rend="right"><ref target="#n84">76</ref></cell>
            </row>
            <row>
              <cell>   Interior</cell>
              <cell rend="right"><ref target="#n86">78</ref></cell>
            </row>
            <row>
              <cell>   Wall and door screens</cell>
              <cell rend="right"><ref target="#n88">80</ref></cell>
            </row>
            <row>
              <cell>  Whare taka</cell>
              <cell rend="right"><ref target="#n88">80</ref></cell>
            </row>
            <row>
              <cell>  Whare tuku whakararo</cell>
              <cell rend="right"><ref target="#n89">81</ref></cell>
            </row>
            <row>
              <cell>   Rarotongan types</cell>
              <cell rend="right"><ref target="#n89">81</ref></cell>
            </row>
            <row>
              <cell>  House furniture</cell>
              <cell rend="right"><ref target="#n90">82</ref></cell>
            </row>
            <row>
              <cell> Food</cell>
              <cell rend="right"><ref target="#n89">81</ref></cell>
            </row>
            <row>
              <cell>  Introduction</cell>
              <cell rend="right"><ref target="#n91">83</ref></cell>
            </row>
            <row>
              <cell>  Fire</cell>
              <cell rend="right"><ref target="#n92">84</ref></cell>
            </row>
            <row>
              <cell>  Fire myth</cell>
              <cell rend="right"><ref target="#n93">85</ref></cell>
            </row>
            <row>
              <cell>  Cooking utensils and food accessories</cell>
              <cell rend="right"><ref target="#n94">86</ref></cell>
            </row>
            <row>
              <cell>  Flesh foods</cell>
              <cell rend="right"><ref target="#n98">90</ref></cell>
            </row>
            <pb xml:id="n6" n="[ii]"/>
            <row>
              <cell>  Vegetable foods</cell>
              <cell rend="right"><ref target="#n100">92</ref></cell>
            </row>
            <row>
              <cell>   Variety</cell>
              <cell rend="right"><ref target="#n100">92</ref></cell>
            </row>
            <row>
              <cell>   The coconut</cell>
              <cell rend="right"><ref target="#n100">92</ref></cell>
            </row>
            <row>
              <cell>   Puraka</cell>
              <cell rend="right"><ref target="#n104">96</ref></cell>
            </row>
            <row>
              <cell>   Hala (Pandanus)</cell>
              <cell rend="right"><ref target="#n105">97</ref></cell>
            </row>
            <row>
              <cell>  Food preparations</cell>
              <cell rend="right"><ref target="#n105">97</ref></cell>
            </row>
            <row>
              <cell>  Fats and oil</cell>
              <cell rend="right"><ref target="#n108">100</ref></cell>
            </row>
            <row>
              <cell>  Feeding of children</cell>
              <cell rend="right"><ref target="#n109">101</ref></cell>
            </row>
            <row>
              <cell> Plaiting</cell>
              <cell rend="right"><ref target="#n110">102</ref></cell>
            </row>
            <row>
              <cell>  Introduction</cell>
              <cell rend="right"><ref target="#n110">102</ref></cell>
            </row>
            <row>
              <cell>  Plaiting with coconut leaves</cell>
              <cell rend="right"><ref target="#n110">102</ref></cell>
            </row>
            <row>
              <cell>   Plaited articles</cell>
              <cell rend="right"><ref target="#n110">102</ref></cell>
            </row>
            <row>
              <cell>   The leaf</cell>
              <cell rend="right"><ref target="#n110">102</ref></cell>
            </row>
            <row>
              <cell>   Technical terms and methods</cell>
              <cell rend="right"><ref target="#n111">103</ref></cell>
            </row>
            <row>
              <cell>   Commencement edge</cell>
              <cell rend="right"><ref target="#n112">104</ref></cell>
            </row>
            <row>
              <cell>   Finishing edge</cell>
              <cell rend="right"><ref target="#n116">108</ref></cell>
            </row>
            <row>
              <cell>   Sheets, screens, and mats</cell>
              <cell rend="right"><ref target="#n118">110</ref></cell>
            </row>
            <row>
              <cell>   Cooking receptacles</cell>
              <cell rend="right"><ref target="#n122">114</ref></cell>
            </row>
            <row>
              <cell>   Baskets</cell>
              <cell rend="right"><ref target="#n127">119</ref></cell>
            </row>
            <row>
              <cell>   Eye shades</cell>
              <cell rend="right"><ref target="#n132">124</ref></cell>
            </row>
            <row>
              <cell>  Plaiting with lauhala (Pandanus leaves)</cell>
              <cell rend="right"><ref target="#n133">125</ref></cell>
            </row>
            <row>
              <cell>   Material</cell>
              <cell rend="right"><ref target="#n133">125</ref></cell>
            </row>
            <row>
              <cell>   Mats</cell>
              <cell rend="right"><ref target="#n134">126</ref></cell>
            </row>
            <row>
              <cell>   Baskets</cell>
              <cell rend="right"><ref target="#n135">127</ref></cell>
            </row>
            <row>
              <cell>  Fans (coconut leaf)</cell>
              <cell rend="right"><ref target="#n139">131</ref></cell>
            </row>
            <row>
              <cell> Clothing and adornment</cell>
              <cell rend="right"><ref target="#n142">134</ref></cell>
            </row>
            <row>
              <cell>  Introduction</cell>
              <cell rend="right"><ref target="#n142">134</ref></cell>
            </row>
            <row>
              <cell>  Material</cell>
              <cell rend="right"><ref target="#n142">134</ref></cell>
            </row>
            <row>
              <cell>  Women's garments</cell>
              <cell rend="right"><ref target="#n142">134</ref></cell>
            </row>
            <row>
              <cell>  Men's garments</cell>
              <cell rend="right"><ref target="#n144">136</ref></cell>
            </row>
            <row>
              <cell>  Accessories</cell>
              <cell rend="right"><ref target="#n144">136</ref></cell>
            </row>
            <row>
              <cell>  Personal decoration</cell>
              <cell rend="right"><ref target="#n144">136</ref></cell>
            </row>
            <row>
              <cell> Stonework</cell>
              <cell rend="right"><ref target="#n146">138</ref></cell>
            </row>
            <row>
              <cell>  Material</cell>
              <cell rend="right"><ref target="#n146">138</ref></cell>
            </row>
            <row>
              <cell>  Miscellaneous</cell>
              <cell rend="right"><ref target="#n146">138</ref></cell>
            </row>
            <row>
              <cell>  Adzes</cell>
              <cell rend="right"><ref target="#n147">139</ref></cell>
            </row>
            <row>
              <cell>   Stone adzes</cell>
              <cell rend="right"><ref target="#n147">139</ref></cell>
            </row>
            <row>
              <cell>   Shell adzes, haft</cell>
              <cell rend="right"><ref target="#n150">142</ref></cell>
            </row>
            <row>
              <cell> Canoes</cell>
              <cell rend="right"><ref target="#n154">146</ref></cell>
            </row>
            <row>
              <cell>  Material</cell>
              <cell rend="right"><ref target="#n154">146</ref></cell>
            </row>
            <row>
              <cell>  Terms and technique</cell>
              <cell rend="right"><ref target="#n155">147</ref></cell>
            </row>
            <row>
              <cell>  Types</cell>
              <cell rend="right"><ref target="#n156">148</ref></cell>
            </row>
            <row>
              <cell>   Introduction</cell>
              <cell rend="right"><ref target="#n156">148</ref></cell>
            </row>
            <row>
              <cell>   One-piece dugout canoe</cell>
              <cell rend="right"><ref target="#n156">148</ref></cell>
            </row>
            <row>
              <cell>   Five-piece canoe</cell>
              <cell rend="right"><ref target="#n157">149</ref></cell>
            </row>
            <row>
              <cell>   Double sailing canoe</cell>
              <cell rend="right"><ref target="#n157">149</ref></cell>
            </row>
            <row>
              <cell>   Modern plank canoes</cell>
              <cell rend="right"><ref target="#n164">156</ref></cell>
            </row>
            <row>
              <cell>  Paddles</cell>
              <cell rend="right"><ref target="#n165">157</ref></cell>
            </row>
            <row>
              <cell>  Bailers</cell>
              <cell rend="right"><ref target="#n166">158</ref></cell>
            </row>
            <row>
              <cell> Fishing</cell>
              <cell rend="right"><ref target="#n166">158</ref></cell>
            </row>
            <row>
              <cell>  Introduction</cell>
              <cell rend="right"><ref target="#n166">158</ref></cell>
            </row>
            <row>
              <cell>  Walled traps</cell>
              <cell rend="right"><ref target="#n167">159</ref></cell>
            </row>
            <row>
              <cell>  Nets</cell>
              <cell rend="right"><ref target="#n168">160</ref></cell>
            </row>
            <pb xml:id="n7" n="[iii]"/>
            <row>
              <cell>  Angling</cell>
              <cell rend="right"><ref target="#n171">163</ref></cell>
            </row>
            <row>
              <cell>   Hooks and lines</cell>
              <cell rend="right"><ref target="#n171">163</ref></cell>
            </row>
            <row>
              <cell>   Ruvettus hook</cell>
              <cell rend="right"><ref target="#n172">164</ref></cell>
            </row>
            <row>
              <cell>   One-piece hooks</cell>
              <cell rend="right"><ref target="#n178">170</ref></cell>
            </row>
            <row>
              <cell>   Bonito hooks</cell>
              <cell rend="right"><ref target="#n180">172</ref></cell>
            </row>
            <row>
              <cell>   Bonito hooks, comparative study</cell>
              <cell rend="right"><ref target="#n195">187</ref></cell>
            </row>
            <row>
              <cell> War and weapons</cell>
              <cell rend="right"><ref target="#n202">194</ref></cell>
            </row>
            <row>
              <cell>  Warfare</cell>
              <cell rend="right"><ref target="#n202">194</ref></cell>
            </row>
            <row>
              <cell>  Spears</cell>
              <cell rend="right"><ref target="#n203">195</ref></cell>
            </row>
            <row>
              <cell>  Clubs</cell>
              <cell rend="right"><ref target="#n203">195</ref></cell>
            </row>
            <row>
              <cell> Recreations</cell>
              <cell rend="right"><ref target="#n204">196</ref></cell>
            </row>
            <row>
              <cell>  Play</cell>
              <cell rend="right"><ref target="#n204">196</ref></cell>
            </row>
            <row>
              <cell>   Dances</cell>
              <cell rend="right"><ref target="#n205">197</ref></cell>
            </row>
            <row>
              <cell>   Drama</cell>
              <cell rend="right"><ref target="#n206">198</ref></cell>
            </row>
            <row>
              <cell>   Musical instruments</cell>
              <cell rend="right"><ref target="#n211">203</ref></cell>
            </row>
            <row>
              <cell>Religion</cell>
              <cell rend="right"><ref target="#n213">205</ref></cell>
            </row>
            <row>
              <cell> The gods</cell>
              <cell rend="right"><ref target="#n213">205</ref></cell>
            </row>
            <row>
              <cell> Maraes</cell>
              <cell rend="right"><ref target="#n216">208</ref></cell>
            </row>
            <row>
              <cell> Priesthood</cell>
              <cell rend="right"><ref target="#n217">209</ref></cell>
            </row>
            <row>
              <cell> Ritual</cell>
              <cell rend="right"><ref target="#n218">210</ref></cell>
            </row>
            <row>
              <cell> Minor gods</cell>
              <cell rend="right"><ref target="#n218">210</ref></cell>
            </row>
            <row>
              <cell> Prohibited foods</cell>
              <cell rend="right"><ref target="#n222">214</ref></cell>
            </row>
            <row>
              <cell> Religious objects</cell>
              <cell rend="right"><ref target="#n223">215</ref></cell>
            </row>
            <row>
              <cell>Sickness and death</cell>
              <cell rend="right"><ref target="#n224">216</ref></cell>
            </row>
            <row>
              <cell> Medicine</cell>
              <cell rend="right"><ref target="#n224">216</ref></cell>
            </row>
            <row>
              <cell> Death</cell>
              <cell rend="right"><ref target="#n225">217</ref></cell>
            </row>
            <row>
              <cell>The calendar</cell>
              <cell rend="right"><ref target="#n226">218</ref></cell>
            </row>
            <row>
              <cell> Nights of the moon</cell>
              <cell rend="right"><ref target="#n226">218</ref></cell>
            </row>
            <row>
              <cell> Annual cycle</cell>
              <cell rend="right"><ref target="#n230">222</ref></cell>
            </row>
            <row>
              <cell> Rakahangan intercalation</cell>
              <cell rend="right"><ref target="#n235">227</ref></cell>
            </row>
            <row>
              <cell> Stars</cell>
              <cell rend="right"><ref target="#n239">231</ref></cell>
            </row>
            <row>
              <cell>Literature cited</cell>
              <cell rend="right"><ref target="#n240">232</ref></cell>
            </row>
            <row>
              <cell>Plate legends</cell>
              <cell rend="right"><ref target="#n241">233</ref></cell>
            </row>
          </table>
      </div>
      <div xml:id="t1-front-d3" type="section">
        <head><hi rend="c">Illustrations</hi></head>

          <table rows="122" cols="3">
            <row>
              <cell/>
              <cell/>
              <cell rend="right"><hi rend="sc">Page</hi></cell>
            </row>
            <row>
              <cell rend="right"><hi rend="sc">Figure</hi> 1.</cell>
              <cell>Map of part of the central Pacific Ocean</cell>
              <cell rend="right"><ref target="#BucMani005a">5</ref></cell>
            </row>
            <row>
              <cell rend="right">2.</cell>
              <cell>Map of Manihiki</cell>
              <cell rend="right"><ref target="#BucMani006a">6</ref></cell>
            </row>
            <row>
              <cell rend="right">3.</cell>
              <cell>Map of Rakahanga</cell>
              <cell rend="right"><ref target="#BucMani007a">7</ref></cell>
            </row>
            <row>
              <cell rend="right">4.</cell>
              <cell>House framework</cell>
              <cell rend="right"><ref target="#BucMani071a">71</ref></cell>
            </row>
            <row>
              <cell rend="right">5.</cell>
              <cell>House framework, lashing</cell>
              <cell rend="right"><ref target="#BucMani072a">72</ref></cell>
            </row>
            <row>
              <cell rend="right">6.</cell>
              <cell>House framework, supporting post</cell>
              <cell rend="right"><ref target="#BucMani073a">73</ref></cell>
            </row>
            <row>
              <cell rend="right">7.</cell>
              <cell>Roof framework, rafters and lashings</cell>
              <cell rend="right"><ref target="#BucMani073b">73</ref></cell>
            </row>
            <row>
              <cell rend="right">8.</cell>
              <cell>Roof framework</cell>
              <cell rend="right"><ref target="#BucMani074a">74</ref></cell>
            </row>
            <row>
              <cell rend="right">9.</cell>
              <cell>End framework of house</cell>
              <cell rend="right"><ref target="#BucMani075a">75</ref></cell>
            </row>
            <row>
              <cell rend="right">10.</cell>
              <cell>Roof sheet needle</cell>
              <cell rend="right"><ref target="#BucMani076a">76</ref></cell>
            </row>
            <row>
              <cell rend="right">11.</cell>
              <cell>Thatching needle</cell>
              <cell rend="right"><ref target="#BucMani077a">77</ref></cell>
            </row>
            <row>
              <cell rend="right">12.</cell>
              <cell>Lifting pole for thatch</cell>
              <cell rend="right"><ref target="#BucMani077b">77</ref></cell>
            </row>
            <row>
              <cell rend="right">13.</cell>
              <cell>House platform, or upper story</cell>
              <cell rend="right"><ref target="#BucMani078a">78</ref></cell>
            </row>
            <row>
              <cell rend="right">14.</cell>
              <cell>Interior arrangement of house</cell>
              <cell rend="right"><ref target="#BucMani079a">79</ref></cell>
            </row>
            <row>
              <cell rend="right">15.</cell>
              <cell>Portable house (whare taka)</cell>
              <cell rend="right"><ref target="#BucMani080a">80</ref></cell>
            </row>
            <row>
              <cell rend="right">16.</cell>
              <cell>House without supporting posts (whare tuku whakararo)</cell>
              <cell rend="right"><ref target="#BucMani081a">81</ref></cell>
            </row>
            <row>
              <cell rend="right">17.</cell>
              <cell>Rarotongan type of house, surrounding veranda</cell>
              <cell rend="right"><ref target="#BucMani082a">82</ref></cell>
            </row>
            <pb xml:id="n8" n="[iv]"/>
            <row>
              <cell rend="right">18.</cell>
              <cell>Wooden seat</cell>
              <cell rend="right"><ref target="#BucMani082a">82</ref></cell>
            </row>
            <row>
              <cell rend="right">19.</cell>
              <cell>Wooden bowls</cell>
              <cell rend="right"><ref target="#BucMani087a">87</ref></cell>
            </row>
            <row>
              <cell rend="right">20.</cell>
              <cell>Pearl-shell hand graters</cell>
              <cell rend="right"><ref target="#BucMani089a">89</ref></cell>
            </row>
            <row>
              <cell rend="right">21.</cell>
              <cell>Wooden pounder</cell>
              <cell rend="right"><ref target="#BucMani089b">89</ref></cell>
            </row>
            <row>
              <cell rend="right">22.</cell>
              <cell>Whole midrib commencement, plaiting with coconut</cell>
              <cell rend="right"><ref target="#BucMani105a">105</ref></cell>
            </row>
            <row>
              <cell rend="right">23.</cell>
              <cell>Single-strip midrib commencement</cell>
              <cell rend="right"><ref target="#BucMani106a">106</ref></cell>
            </row>
            <row>
              <cell rend="right">24.</cell>
              <cell>Simple two-strip commencement</cell>
              <cell rend="right"><ref target="#BucMani107a">107</ref></cell>
            </row>
            <row>
              <cell rend="right">25.</cell>
              <cell>Twisted two-strip commencement</cell>
              <cell rend="right"><ref target="#BucMani107b">107</ref></cell>
            </row>
            <row>
              <cell rend="right">26.</cell>
              <cell>Simple two-pair commencement</cell>
              <cell rend="right"><ref target="#BucMani108a">108</ref></cell>
            </row>
            <row>
              <cell rend="right">27.</cell>
              <cell>Spaced three-ply braid finish of mats, sheets, and platters</cell>
              <cell rend="right"><ref target="#BucMani109a">109</ref></cell>
            </row>
            <row>
              <cell rend="right">28.</cell>
              <cell>Close three-ply braid finish</cell>
              <cell rend="right"><ref target="#BucMani110a">110</ref></cell>
            </row>
            <row>
              <cell rend="right">29.</cell>
              <cell>Commencement of screen</cell>
              <cell rend="right"><ref target="#BucMani111a">111</ref></cell>
            </row>
            <row>
              <cell rend="right">30.</cell>
              <cell>Keel commencement of tapakau mat</cell>
              <cell rend="right"><ref target="#BucMani112a">112</ref></cell>
            </row>
            <row>
              <cell rend="right">31.</cell>
              <cell>Twill plaiting of tapakau mat</cell>
              <cell rend="right"><ref target="#BucMani113a">113</ref></cell>
            </row>
            <row>
              <cell rend="right">32.</cell>
              <cell>Fish platter, first course of two-course braid finish</cell>
              <cell rend="right"><ref target="#BucMani116a">116</ref></cell>
            </row>
            <row>
              <cell rend="right">33.</cell>
              <cell>Pointed pite receptacle</cell>
              <cell rend="right"><ref target="#BucMani117a">117</ref></cell>
            </row>
            <row>
              <cell rend="right">34.</cell>
              <cell>Four-cornered receptacle, technique of corners</cell>
              <cell rend="right"><ref target="#BucMani118a">118</ref></cell>
            </row>
            <row>
              <cell rend="right">35.</cell>
              <cell>Whatu technique, round basket, twisted two-strip commencement</cell>
              <cell rend="right"><ref target="#BucMani119a">119</ref></cell>
            </row>
            <row>
              <cell rend="right">36.</cell>
              <cell>Round basket, bottom closure</cell>
              <cell rend="right"><ref target="#BucMani120a">120</ref></cell>
            </row>
            <row>
              <cell rend="right">37.</cell>
              <cell>Braid commencement of rim of kete ngahengahe (clothes basket)</cell>
              <cell rend="right"><ref target="#BucMani121a">121</ref></cell>
            </row>
            <row>
              <cell rend="right">38.</cell>
              <cell>Twilled plaiting of kete ngahengahe</cell>
              <cell rend="right"><ref target="#BucMani122a">122</ref></cell>
            </row>
            <row>
              <cell rend="right">39.</cell>
              <cell>Rim wrapping of kete ngahengahe</cell>
              <cell rend="right"><ref target="#BucMani123a">123</ref></cell>
            </row>
            <row>
              <cell rend="right">40.</cell>
              <cell>Braid technique of full-leaf braided basket</cell>
              <cell rend="right"><ref target="#BucMani124a">124</ref></cell>
            </row>
            <row>
              <cell rend="right">41.</cell>
              <cell>Simple eye shade</cell>
              <cell rend="right"><ref target="#BucMani125a">125</ref></cell>
            </row>
            <row>
              <cell rend="right">42.</cell>
              <cell>Two-cornered lauhala satchel, bottom closure</cell>
              <cell rend="right"><ref target="#BucMani127a">127</ref></cell>
            </row>
            <row>
              <cell rend="right">43.</cell>
              <cell>Four-cornered satchel, bottom and corner technique</cell>
              <cell rend="right"><ref target="#BucMani129a">129</ref></cell>
            </row>
            <row>
              <cell rend="right">44.</cell>
              <cell>Lauhala satchels, color designs</cell>
              <cell rend="right"><ref target="#BucMani130a">130</ref></cell>
            </row>
            <row>
              <cell rend="right">45.</cell>
              <cell>Single-section coconut leaf fan</cell>
              <cell rend="right"><ref target="#BucMani132a">132</ref></cell>
            </row>
            <row>
              <cell rend="right">46.</cell>
              <cell>Two-section coconut leaf fan</cell>
              <cell rend="right"><ref target="#BucMani133a">133</ref></cell>
            </row>
            <row>
              <cell rend="right">47.</cell>
              <cell>Kilt and fringe technique</cell>
              <cell rend="right"><ref target="#BucMani135a">135</ref></cell>
            </row>
            <row>
              <cell rend="right">48.</cell>
              <cell>Long perineal band</cell>
              <cell rend="right"><ref target="#BucMani137a">137</ref></cell>
            </row>
            <row>
              <cell rend="right">49.</cell>
              <cell>Ornament of pearl shell</cell>
              <cell rend="right"><ref target="#BucMani138a">138</ref></cell>
            </row>
            <row>
              <cell rend="right">50.</cell>
              <cell>Triangular tanged stone adz</cell>
              <cell rend="right"><ref target="#BucMani139a">139</ref></cell>
            </row>
            <row>
              <cell rend="right">51.</cell>
              <cell>Triangular adz, without tang</cell>
              <cell rend="right"><ref target="#BucMani140a">140</ref></cell>
            </row>
            <row>
              <cell rend="right">52.</cell>
              <cell>Quadrangular adz, with tang</cell>
              <cell rend="right"><ref target="#BucMani141a">141</ref></cell>
            </row>
            <row>
              <cell rend="right">53.</cell>
              <cell>Quadrangular tanged stone adz with bilateral lugs, Nassau Islands</cell>
              <cell rend="right"><ref target="#BucMani142a">142</ref></cell>
            </row>
            <row>
              <cell rend="right">54.</cell>
              <cell>Adzes of Tridacna shell</cell>
              <cell rend="right"><ref target="#BucMani142b">142</ref></cell>
            </row>
            <row>
              <cell rend="right">55.</cell>
              <cell>Adz haft</cell>
              <cell rend="right"><ref target="#BucMani143a">143</ref></cell>
            </row>
            <row>
              <cell rend="right">56.</cell>
              <cell>Lashing of simple haft</cell>
              <cell rend="right"><ref target="#BucMani144a">144</ref></cell>
            </row>
            <row>
              <cell rend="right">57.</cell>
              <cell>Haft with socket, and lashing technique</cell>
              <cell rend="right"><ref target="#BucMani145a">145</ref></cell>
            </row>
            <row>
              <cell rend="right">58.</cell>
              <cell>Dugout canoe hull</cell>
              <cell rend="right"><ref target="#BucMani148a">148</ref></cell>
            </row>
            <row>
              <cell rend="right">59.</cell>
              <cell>Double canoe, hull</cell>
              <cell rend="right"><ref target="#BucMani149a">149</ref></cell>
            </row>
            <row>
              <cell rend="right">60.</cell>
              <cell>Double canoe, stern piece</cell>
              <cell rend="right"><ref target="#BucMani150a">150</ref></cell>
            </row>
            <row>
              <cell rend="right">61.</cell>
              <cell>Double canoe, bow piece</cell>
              <cell rend="right"><ref target="#BucMani150b">150</ref></cell>
            </row>
            <row>
              <cell rend="right">62.</cell>
              <cell>Double canoe, gunwale plank</cell>
              <cell rend="right"><ref target="#BucMani150c">150</ref></cell>
            </row>
            <row>
              <cell rend="right">63.</cell>
              <cell>Double canoe, hull elements assembled</cell>
              <cell rend="right"><ref target="#BucMani151a">151</ref></cell>
            </row>
            <row>
              <cell rend="right">64.</cell>
              <cell>Double canoe, gunwale and hull lashing</cell>
              <cell rend="right"><ref target="#BucMani151b">151</ref></cell>
            </row>
            <row>
              <cell rend="right">65.</cell>
              <cell>Double canoe, bulwark</cell>
              <cell rend="right"><ref target="#BucMani152a">152</ref></cell>
            </row>
            <row>
              <cell rend="right">66.</cell>
              <cell>Double canoe, cross brace and lashings</cell>
              <cell rend="right"><ref target="#BucMani153a">153</ref></cell>
            </row>
            <row>
              <cell rend="right">67.</cell>
              <cell>Double canoe, mast</cell>
              <cell rend="right"><ref target="#BucMani154a">154</ref></cell>
            </row>
            <row>
              <cell rend="right">68.</cell>
              <cell>Double canoe, setting of mast and sail</cell>
              <cell rend="right"><ref target="#BucMani155a">155</ref></cell>
            </row>
            <row>
              <cell rend="right">69.</cell>
              <cell>Modern plank canoe, indirect connections between boom and float</cell>
              <cell rend="right"><ref target="#BucMani156a">156</ref></cell>
            </row>
            <row>
              <cell rend="right">70.</cell>
              <cell>Steering paddle used with double canoe</cell>
              <cell rend="right"><ref target="#BucMani157a">157</ref></cell>
            </row>
            <row>
              <cell rend="right">71.</cell>
              <cell>Walled fish trap</cell>
              <cell rend="right"><ref target="#BucMani159a">159</ref></cell>
            </row>
            <pb xml:id="n9" n="[v]"/>
            <row>
              <cell rend="right">72.</cell>
              <cell>Wooden netting needle</cell>
              <cell rend="right"><ref target="#BucMani160a">160</ref></cell>
            </row>
            <row>
              <cell rend="right">73.</cell>
              <cell>Netting technique, first row</cell>
              <cell rend="right"><ref target="#BucMani161a">161</ref></cell>
            </row>
            <row>
              <cell rend="right">74.</cell>
              <cell>Netting technique, second row</cell>
              <cell rend="right"><ref target="#BucMani162a">162</ref></cell>
            </row>
            <row>
              <cell rend="right">75.</cell>
              <cell>Netting technique, extra meshes</cell>
              <cell rend="right"><ref target="#BucMani162b">162</ref></cell>
            </row>
            <row>
              <cell rend="right">76.</cell>
              <cell>Flying fish net</cell>
              <cell rend="right"><ref target="#BucMani163a">163</ref></cell>
            </row>
            <row>
              <cell rend="right">77.</cell>
              <cell>Ruvettus hook and point lashing</cell>
              <cell rend="right"><ref target="#BucMani164a">164</ref></cell>
            </row>
            <row>
              <cell rend="right">78.</cell>
              <cell>Ruvettus hook, snood wrapping, Manihiki</cell>
              <cell rend="right"><ref target="#BucMani165a">165</ref></cell>
            </row>
            <row>
              <cell rend="right">79.</cell>
              <cell>Ruvettus hook, snood lashing</cell>
              <cell rend="right"><ref target="#BucMani166a">166</ref></cell>
            </row>
            <row>
              <cell rend="right">80.</cell>
              <cell>Ruvettus hook with pearl-shell point</cell>
              <cell rend="right"><ref target="#BucMani167a">167</ref></cell>
            </row>
            <row>
              <cell rend="right">81.</cell>
              <cell>Ruvettus hook, snood wrapping, Rakahanga</cell>
              <cell rend="right"><ref target="#BucMani168a">168</ref></cell>
            </row>
            <row>
              <cell rend="right">82.</cell>
              <cell>Ruvettus hook, snood lashing</cell>
              <cell rend="right"><ref target="#BucMani169a">169</ref></cell>
            </row>
            <row>
              <cell rend="right">83.</cell>
              <cell>One-piece pearl shell hooks</cell>
              <cell rend="right"><ref target="#BucMani170a">170</ref></cell>
            </row>
            <row>
              <cell rend="right">84.</cell>
              <cell>One-piece hook, snood lashing</cell>
              <cell rend="right"><ref target="#BucMani171a">171</ref></cell>
            </row>
            <row>
              <cell rend="right">85.</cell>
              <cell>One-piece hook, snood lashing</cell>
              <cell rend="right"><ref target="#BucMani172a">172</ref></cell>
            </row>
            <row>
              <cell rend="right">86.</cell>
              <cell>Bonito hook shanks, front</cell>
              <cell rend="right"><ref target="#BucMani174a">174</ref></cell>
            </row>
            <row>
              <cell rend="right">87.</cell>
              <cell>Bonito hook shanks, side</cell>
              <cell rend="right"><ref target="#BucMani175a">175</ref></cell>
            </row>
            <row>
              <cell rend="right">88.</cell>
              <cell>Bonito hook points</cell>
              <cell rend="right"><ref target="#BucMani176a">176</ref></cell>
            </row>
            <row>
              <cell rend="right">89.</cell>
              <cell>Bonito hook, attachments of snood to point</cell>
              <cell rend="right"><ref target="#BucMani177a">177</ref></cell>
            </row>
            <row>
              <cell rend="right">90.</cell>
              <cell>Bonito hook, single lashing through proximal hole</cell>
              <cell rend="right"><ref target="#BucMani178a">178</ref></cell>
            </row>
            <row>
              <cell rend="right">91.</cell>
              <cell>Bonito hook, distal single lashing and hackle fixation</cell>
              <cell rend="right"><ref target="#BucMani179a">179</ref></cell>
            </row>
            <row>
              <cell rend="right">92.</cell>
              <cell>Bonito hook, continuous lashing, proximal hole commencement</cell>
              <cell rend="right"><ref target="#BucMani180a">180</ref></cell>
            </row>
            <row>
              <cell rend="right">93.</cell>
              <cell>Bonito hook, continuous lashing, distal hole commencement</cell>
              <cell rend="right"><ref target="#BucMani181a">181</ref></cell>
            </row>
            <row>
              <cell rend="right">94.</cell>
              <cell>Bonito hook, variations in point lashing</cell>
              <cell rend="right"><ref target="#BucMani182a">182</ref></cell>
            </row>
            <row>
              <cell rend="right">95.</cell>
              <cell>Bonito hook, modern attachment of hackle</cell>
              <cell rend="right"><ref target="#BucMani182b">182</ref></cell>
            </row>
            <row>
              <cell rend="right">96.</cell>
              <cell>Bonito hook, simple snood-head lashing</cell>
              <cell rend="right"><ref target="#BucMani183a">183</ref></cell>
            </row>
            <row>
              <cell rend="right">97.</cell>
              <cell>Bonito hook, complicated snood-head lashing</cell>
              <cell rend="right"><ref target="#BucMani183b">183</ref></cell>
            </row>
            <row>
              <cell rend="right">98.</cell>
              <cell>Bonito hook, neatest snood-head lashing</cell>
              <cell rend="right"><ref target="#BucMani184a">184</ref></cell>
            </row>
            <row>
              <cell rend="right">99.</cell>
              <cell>Bonito hook, complete lashing</cell>
              <cell rend="right"><ref target="#BucMani184b">184</ref></cell>
            </row>
            <row>
              <cell rend="right">100.</cell>
              <cell>Bonito hook, filler sticks</cell>
              <cell rend="right"><ref target="#BucMani184c">184</ref></cell>
            </row>
            <row>
              <cell rend="right">101.</cell>
              <cell>Filler sticks, Pukapukan and Ellice Islands bonito hooks</cell>
              <cell rend="right"><ref target="#BucMani185a">185</ref></cell>
            </row>
            <row>
              <cell rend="right">102.</cell>
              <cell>Melanesian and Micronesian bonito hooks</cell>
              <cell rend="right"><ref target="#BucMani188a">188</ref></cell>
            </row>
            <row>
              <cell rend="right">103.</cell>
              <cell>Eastern Polynesian bonito hooks</cell>
              <cell rend="right"><ref target="#BucMani190a">190</ref></cell>
            </row>
            <row>
              <cell rend="right">104.</cell>
              <cell>Eastern Polynesian bonito hook points</cell>
              <cell rend="right"><ref target="#BucMani191a">191</ref></cell>
            </row>
            <row>
              <cell rend="right">105.</cell>
              <cell>Western Polynesian bonito hook points</cell>
              <cell rend="right"><ref target="#BucMani192a">192</ref></cell>
            </row>
            <row>
              <cell rend="right">106.</cell>
              <cell>Fagatau bonito hooks</cell>
              <cell rend="right"><ref target="#BucMani193a">193</ref></cell>
            </row>
            <row>
              <cell rend="right">107.</cell>
              <cell>Western Polynesian bonito hook points</cell>
              <cell rend="right"><ref target="#BucMani193b">193</ref></cell>
            </row>
            <row>
              <cell rend="right">108.</cell>
              <cell>Weapons</cell>
              <cell rend="right"><ref target="#BucMani195a">195</ref></cell>
            </row>
            <row>
              <cell rend="right">109.</cell>
              <cell>Short clubs</cell>
              <cell rend="right"><ref target="#BucMani106a">106</ref></cell>
            </row>
            <row>
              <cell/>
              <cell/>
              <cell rend="right"><hi rend="sc">Following Page</hi></cell>
            </row>
            <row>
              <cell rend="right"><hi rend="sc">Plate</hi> 1.</cell>
              <cell>Houses and implements</cell>
              <cell rend="right"><ref target="#BucManiP001a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">2.</cell>
              <cell>Coconut leaf mats</cell>
              <cell rend="right"><ref target="#BucManiP002a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">3.</cell>
              <cell>Coconut leaf mat and cooking receptacles</cell>
              <cell rend="right"><ref target="#BucManiP003a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">4.</cell>
              <cell>Coconut leaf platters, receptacles, eye shade, and baskets</cell>
              <cell rend="right"><ref target="#BucManiP004a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">5.</cell>
              <cell>Satchels and fans</cell>
              <cell rend="right"><ref target="#BucManiP005a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">6.</cell>
              <cell>Poncho, kilt, and mat</cell>
              <cell rend="right"><ref target="#BucManiP006a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">7.</cell>
              <cell>Canoes and bailer</cell>
              <cell rend="right"><ref target="#BucManiP007a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">8.</cell>
              <cell>Fish nets and netting needle</cell>
              <cell rend="right"><ref target="#BucManiP008a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">9.</cell>
              <cell>Paddles and hooks</cell>
              <cell rend="right"><ref target="#BucManiP009a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">10.</cell>
              <cell>Weapons and adzes</cell>
              <cell rend="right"><ref target="#BucManiP010a">238</ref></cell>
            </row>
            <row>
              <cell rend="right">11.</cell>
              <cell>Wooden gongs and historical play</cell>
              <cell rend="right"><ref target="#BucManiP011a">238</ref></cell>
            </row>
          </table>
      </div>
      <pb xml:id="n10" n="[vi]"/>
      <div xml:id="t1-front-d4" type="section">
        <head><hi rend="c">Tables</hi></head>

          <table rows="23" cols="3">
            <row>
              <cell rend="right"><hi rend="sc">Table</hi> 1.</cell>
              <cell>Descent from Toa</cell>
              <cell rend="right"><ref target="#n30">22</ref></cell>
            </row>
            <row>
              <cell rend="right">2.</cell>
              <cell>Issue of Toa</cell>
              <cell rend="right"><ref target="#n34">26</ref></cell>
            </row>
            <row>
              <cell rend="right">3.</cell>
              <cell>Second families</cell>
              <cell rend="right"><ref target="#n34">26</ref></cell>
            </row>
            <row>
              <cell rend="right">4.</cell>
              <cell>Relationship terms</cell>
              <cell rend="right"><ref target="#n38">30</ref></cell>
            </row>
            <row>
              <cell rend="right">5.</cell>
              <cell>Single ariki title holders</cell>
              <cell rend="right"><ref target="#n52">44</ref></cell>
            </row>
            <row>
              <cell rend="right">6.</cell>
              <cell>The Whainga-aitu title holders</cell>
              <cell rend="right"><ref target="#n57">49</ref></cell>
            </row>
            <row>
              <cell rend="right">7.</cell>
              <cell>Manner in which succession passes over female issue</cell>
              <cell rend="right"><ref target="#n60">52</ref></cell>
            </row>
            <row>
              <cell rend="right">8.</cell>
              <cell>The Whakaheo title holders</cell>
              <cell rend="right"><ref target="#n61">53</ref></cell>
            </row>
            <row>
              <cell rend="right">9.</cell>
              <cell>Succession to Takai-whakaheo</cell>
              <cell rend="right"><ref target="#n61">53</ref></cell>
            </row>
            <row>
              <cell rend="right">10.</cell>
              <cell>Whainga-aitu tribes and subtribes</cell>
              <cell rend="right"><ref target="#n69">61</ref></cell>
            </row>
            <row>
              <cell rend="right">11.</cell>
              <cell>Origin of subtribes</cell>
              <cell rend="right"><ref target="#n69">61</ref></cell>
            </row>
            <row>
              <cell rend="right">12.</cell>
              <cell>Te-whare-nui subtribe</cell>
              <cell rend="right"><ref target="#n70">62</ref></cell>
            </row>
            <row>
              <cell rend="right">13.</cell>
              <cell>Whatiakau subtribe</cell>
              <cell rend="right"><ref target="#n71">63</ref></cell>
            </row>
            <row>
              <cell rend="right">14.</cell>
              <cell>Whakaheo tribes and subtribes</cell>
              <cell rend="right"><ref target="#n73">65</ref></cell>
            </row>
            <row>
              <cell rend="right">15.</cell>
              <cell>Stages in growth of the coconut</cell>
              <cell rend="right"><ref target="#n103">95</ref></cell>
            </row>
            <row>
              <cell rend="right">16.</cell>
              <cell>Rakahangan nights of the moon</cell>
              <cell rend="right"><ref target="#n226">218</ref></cell>
            </row>
            <row>
              <cell rend="right">17.</cell>
              <cell>Rakahangan nights of the moon</cell>
              <cell rend="right"><ref target="#n229">221</ref></cell>
            </row>
            <row>
              <cell rend="right">18.</cell>
              <cell>Rakahangan annual cycle, three-period classification</cell>
              <cell rend="right"><ref target="#n231">223</ref></cell>
            </row>
            <row>
              <cell rend="right">19.</cell>
              <cell>Rakahangan annual cycle, two-period classification</cell>
              <cell rend="right"><ref target="#n232">224</ref></cell>
            </row>
            <row>
              <cell rend="right">20.</cell>
              <cell>Twelve-month cycle, comparison and stars</cell>
              <cell rend="right"><ref target="#n233">225</ref></cell>
            </row>
            <row>
              <cell rend="right">21.</cell>
              <cell>Observations regarding the Pleiades</cell>
              <cell rend="right"><ref target="#n234">226</ref></cell>
            </row>
            <row>
              <cell rend="right">22.</cell>
              <cell>Rakahangan annual cycle of thirteen months</cell>
              <cell rend="right"><ref target="#n236">228</ref></cell>
            </row>
            <row>
              <cell rend="right">23.</cell>
              <cell>Range of lunar months</cell>
              <cell rend="right"><ref target="#n237">229</ref></cell>
            </row>
          </table>
      </div>
    </front>
    <body xml:id="t1-body">
      <pb xml:id="n11"/>
      <head>Ethnology of Manihiki and Rakahanga</head>
      <byline>By<lb/>
<hi rend="sc"><name type="person" key="name-202886">Te Rangi Hiroa</name> (<name type="person" key="name-202886">P. H. Buck</name>)</hi></byline>
      <div xml:id="t1-body-d1-d1" type="section">
        <head>
          <hi rend="c">Introduction</hi>
        </head>
        <div xml:id="t1-body-d1-d1-d1" type="section">
          <head>
            <hi rend="sc">Acknowledgments</hi>
          </head>
          <p>This study of the culture of Manihiki and Rakahanga represents part of the material gathered on the Cook Islands Expedition of Bernice P. Bishop Museum in 1929.</p>
          <p>I landed in Manihiki on May 31 on the schooner <hi rend="i">Tiare Taporo</hi> under Captain Viggo. While I was in Rakahanga, Judge <name type="person" key="name-207309">Hugh Ayson</name>, Resident Commissioner for the Cook Islands, convened the Native Land Court to inquire into family pedigrees for the purpose of forming bases for land claims. Through his courtesy and the assistance of <name type="person" key="name-121166">Stephen Savage</name>, Registrar of the Court, I was able to acquire a complete set of island pedigrees. To <name key="name-101696" type="person">Henry Williams, Jr.</name>, I am under obligation not only for maps of Manihiki and Rakahanga, but also for assistance in recording anthropometrical measurements. To Tupou-rahi, Sergeant of Police at Rakahanga, and his family thanks are due for accommodation, hospitality, and much information. The hospitality of the kindly people of Rakahanga manifested itself in feasts and presents of artifacts. The people were eager to impart what they knew, but owing to the exigencies of inter-island transport three weeks were all that could be devoted to the atoll. Only two nights were spent in Manihiki before the schooner moved on to Tongareva. In Manihiki the people of the villages, Tauhunu and Tukou, were also most hospitable and would not let their visitors go without weighing them down with food and presents. The gratitude of Bernice P. Bishop Museum is due to the people of the two atolls whose gifts have materially enriched the Polynesian collection. I have also to thank Mr. Murray of Rakahanga for the replica of a club and <name key="name-209643" type="person">Henry Williams, Sr.</name>, and his family for hospitality and assistance.</p>
          <p>The time spent in the two atolls was all too short to do justice even to the abridged field information available. This study can only hope to record some of the main points in the culture of the people. Details as to spread of the coconut and the ownership of land await the further investigator of land claims, when more information is available.</p>
          <p>Meanwhile, the Huru-awatea waves its fronds in the shade of Arai-awa— “Tera pa te Huru-awatea te tahirihiri mai ra i te maru o Arai-awa.”</p>
        </div>
        <pb xml:id="n12" n="4"/>
        <div xml:id="t1-body-d1-d1-d2" type="section">
          <head>
            <hi rend="sc">Culture</hi>
          </head>
          <p>Manihiki and Rakahanga, atolls 25 miles apart, are occupied by a people sprung from one family of settlers. The whole population once lived on one atoll at a time and moved back and forth from one atoll to the other when forced to do so by depletions of the coconut and <hi rend="i">puraka</hi> (species of taro) supplies. Manihiki and Rakahanga are so low that one atoll cannot be seen from the other, but from a point halfway between them both are visible. The people, in going from one atoll to the other, used the Magellan clouds (<hi rend="i">Na Mahu</hi>) as guides. The voyages were made in fleets of double sailing canoes, and the <hi rend="i">Whakaheo ariki</hi> (p. <ref target="#n60">52</ref>), because he was believed to control the weather conditions that insured a successful voyage, took command. Now and again the voyagers were unexpectedly overtaken by storms, but such disasters did not deter the people from making their inter-atoll voyages, for they were impelled by an important need, the urge for food. The occasional loss of life was regarded merely as the natural toll of the sea which the ancestors had paid from time immemorial. It remained for outside influence, in 1852, to use the loss of life as a means of dividing the population into permanent settlements on each atoll. Thus, though Manihiki and Rakahanga are two atolls, the culture is one. References in this text to the culture of either atoll may be taken to apply to the culture of both areas.</p>
        </div>
        <div xml:id="t1-body-d1-d1-d3" type="section">
          <head>
            <hi rend="sc">Geography</hi>
          </head>
          <p>Manihiki and Rakahanga are atolls now politically included in Cook Islands, but, together with Tongareva (Penrhyn), Pukupuka, Nassau, Suvarov (Suwarrow), and Palmerston, they are not geographically part of Cook Islands. References in the text to Cook Islands apply to the geographical division unless it is otherwise stated. Manihiki lies 650 miles north of Rarotonga. Rakahanga is 25 miles north northwest of Manihiki. The two atolls are south of latitude 10° S. and west of longitude 160° W. (See <ref target="#BucMani005a">fig. 1</ref>.)</p>
          <p>Manihiki is the larger of the two atolls and contains about 1,250 acres of land. It has a fine lagoon abounding in pearl shell and <hi rend="i">Tridacna</hi>, but there are no large natural passages through the reef. The boat passages opposite the two villages are short and boats or canoes must be run up onto the reef, from which they are dragged to the deeper water on the inner side of the reef flat. Shallow passages through which the tide reaches the central lagoon separate the islands. (See <ref target="#BucMani006a">fig. 2</ref>.)</p>
          <p>Many of the islands shown in <ref target="#BucMani006a">figure 2</ref> are named in pairs with the qualifying terms <hi rend="i">rahi</hi> (large) and <hi rend="i">iti</hi> (small), as in Hohahake rahi and Hohahake iti. Some receive the name <hi rend="i">motu</hi> (island) with a qualifying term, as in Motu-roa (Long Island), Motu Fara (<hi rend="i">Pandanus</hi> Island), and Motu-o-poia (Poia's Island). The large island, Porea, has a fishpond in its interior.</p>
          <pb xml:id="n13" n="5"/>
          <p>Brigham (3, p. 106)<note xml:id="fn1_5" n="1"><p>Numbers in parentheses refer to Literature Cited, page 232.</p></note> has given the atoll the incorrect name, Monahiki, and in writing of model double canoes inlaid with pearl shell he has evidently confused the correct name, Manihiki, with Manihi in the Tuamotus.</p>
          <p>Rakahanga, with an area of 1,000 acres, is smaller than Manihiki. It is written “Rakaanga” on the maps, from a failure to appreciate the presence of the <hi rend="i">h</hi> sound. Gill, who did not visit the atoll, makes the curious statement that there is no lagoon (11, p. 12). Brigham (3, p. 35) has evidently copied the erroneous statement. Rakahanga is an atoll, and the inclosed lagoon is a characteristic feature. (See <ref target="#BucMani007a">fig. 3</ref>.)</p>
          <p>
            <figure xml:id="BucMani005a">
              <graphic url="BucMani005a.jpg" mimeType="image/jpeg" xml:id="BucMani005a-g"/>
              <head><hi rend="sc">Figure</hi> 1. Map of part of the central Pacific Ocean showing the position of Manihiki and Rakahanga.</head>
            </figure>
          </p>
          <p>The small island, Te Kainga, in the southwest, was the original home of the people, but the site of the single village was changed to Rakahanga across the inter-island channel to the south. Opposite Te Kainga is the shallow boat passage. The reef has no deep passages. The presence of the place names Tongareva and Tua-i-Omoka on the south coast suggests some connection with the Tongarevan atoll, in which Omoka is the principal village. The historical connection is further supported by the name Tua i te Ara o Mahuta (the back of the path of Mahuta) on the northeast. Mahuta, the Tongarevan ancestor, is stated by both Tongarevan and Rakahangan traditions to have lived in Rakahanga prior to his voyage to Tongareva.</p>
        </div>
        <div xml:id="t1-body-d1-d1-d4" type="section">
          <head>
            <hi rend="sc">Contact With Western Culture</hi>
          </head>
          <p>Bellinghausen, the Russian explorer, visited Rakahanga in 1820 and named it Grand Duke Alexander. In 1822 Captain Patrickson in the <hi rend="i">Good</hi>
<pb xml:id="n14" n="6"/>
<figure xml:id="BucMani006a"><graphic url="BucMani006a.jpg" mimeType="image/jpeg" xml:id="BucMani006a-g"/><head><hi rend="c">Manihiki Islets</hi><lb/>
1. Rangahoe<lb/>
2. Name unknown (not recorded)<lb/>
3. Hohahake-rahi<lb/>
4. Hohahake-iti<lb/>
5. Te Motu-o-Poia<lb/>
6. Tima<lb/>
7. Tapa<lb/>
8. Paeke-rahi<lb/>
9. Paeke-iti<lb/>
10. Hiropotiki<lb/>
11. Tikapai<lb/>
12. Name Unknown<lb/>
13. Moturoa<lb/>
14. Mairere-tou<lb/>
15. Motu-fara [Motu-whara]<lb/>
16. Te Motu-o-Pae<lb/>
17. Fakifaki [Whakiwhaki]<lb/>
18. Taingaru-iti<lb/>
19. Taingaru-rahi<lb/>
20. Topuaikaha<lb/>
21. Name Unknown<lb/>
22. Raukotaha<lb/>
23. Rifa-iti [Riwha-iti]<lb/>
24. Toruerue-iti<lb/>
25. Toruerue-rahi<lb/>
26. Punganui-rahi<lb/>
27. Punganui-iti<lb/>
28. Name unknown<lb/>
29. Nafarakura [Nawharakura]<lb/>
30. Tavahavaha<lb/>
31. Tarakite-iti<lb/>
32. Kopu-Ngaha<lb/>
33. Tarakite-iti<lb/>
34. Hakari-manu<lb/>
35. Motu-tou<lb/>
36. Toputangaroa<lb/>
37. Aratini<lb/>
38. Nuku-Hiro<lb/>
39. Iotia<lb/>
40. Moina<lb/>
<hi rend="sc">Figure</hi> 2. Map of Manihiki showing villages Tauhunu and Tukou. (Based on a sketch by <name type="person" key="name-101696">Henry Williams</name>, Government Surveyor, Cook Islands.) The spelling “Tukao” on the map should be “Tukou.”</head></figure>
<pb xml:id="n15" n="7"/>
<figure xml:id="BucMani007a"><graphic url="BucMani007a.jpg" mimeType="image/jpeg" xml:id="BucMani007a-g"/><head><hi rend="sc">Figure</hi> 3. Map of Rakahanga, much of the land area shown without the shallow channels which divide the small islands. (Based on an uncompleted survey by <name type="person" key="name-101696">Henry Williams</name>, Government Surveyor, Cook Islands.)</head></figure>
<pb xml:id="n16" n="8"/>
<hi rend="i">Hope</hi> saw both atolls. He gave Rakahanga the name Reirson and called Manihiki Humphrey.</p>
          <p>According to Gill (10, vol. 2, p. 266), the first person to tell the islanders of the “white man's God” was a Tahitian lad who ran away from a whaling ship that called at Manihiki to get a supply of coconuts. Some of the young men of Manihiki set out in a canoe expedition in the hope of getting to some of the islands of which they were told. “In this and other enterprises of kindred character,” writes Gill, “many of them lost their lives, and on one occasion some were taken to the distant islands of Samoa.”</p>
          <p>The first actual contact with Western culture was brought about by an accident which befell a party sailing from Manihiki to Rakahanga in 1849. A canoe was blown out of its course and was picked up by a whaling ship 80 miles from land. The whaler picked up the crew of five men and four women. Gill (10, vol. 2, p. 268) states that the captain intended to land the Manihikians at Aitutaki or Rarotonga, but that, not being able to make either island, he landed them at Manuae, where an American salesman was living, collecting coconuts, and feeding swine for the Tahitian market. The missionary ship <hi rend="i">John Williams</hi> eventually called at Manuae and took the Manihikians to Aitutaki, which had come under Christian influence in 1821. After a fortnight on Aitutaki they were returned to Manihiki on the mission ship, accompanied by two native teachers named Aporo and Tahiri. The people allowed the teachers to land, and the London Missionary Society thus established itself on Manihiki.</p>
          <p>The teachings of the missionaries were accepted; in less than twelve months after the missionaries landed, most of the material representations of the native gods were destroyed. Churches and schools were established, and the foundations of Western culture were laid by the native teachers. Gill, on a voyage to Sydney in the mission ship, visited Manihiki in 1852. He remarks (10, vol. 2, p. 276) that in the space of three years all the inhabitants, with the exception of one hundred persons, were under Christian instruction. After the establishment of the teachers, twenty persons out of two hundred who were overtaken by a storm were drowned in a voyage to Rakahanga. As a result of missionary representations in 1852, the Christians in Sydney purchased a boat for the teachers, and the Aitutakians purchased another. With the introduction of European boats commenced the passing away of the old-time double sailing canoes in which the voyages between the two atolls had been made. Gill and the native teachers also used their influence in persuading the people to abandon the voyages between the two atolls. To prevent the necessity for the voyages, the population was divided, and each atoll was occupied permanently. The religious influence of the introduced culture resulted in the destruction of the maraes and of
<pb xml:id="n17" n="9"/>
the native gods and in the loss of the priestly functions of the two <hi rend="i">ariki</hi> known as the <hi rend="i">Whainga-aitu</hi> and the <hi rend="i">Whakaheo</hi> (p. <ref target="#n56">48</ref>). Houses and clothing were affected by the introduction of new modes and materials. The abandonment of the inter-atoll voyages with permanent occupation of the two atolls by the divided people led to further changes in social organization in which the power of the high chiefs lessened as the influence of the missionaries increased. One of the <hi rend="i">Whakaheo</hi> even offered the succession of the title to the missionary teacher, Tahiri. Tahiri refused the title but later used his influence to have succession conferred on the female line, as his selected candidate was a deacon in the church (p. <ref target="#n62">54</ref>).</p>
          <p>The missionary teachers, following the procedure adopted by the London Missionary Society in Aitutaki, Rarotonga, and Mangaia, formulated a number of moral laws, the infringing of many of which we would regard as “sins” to be dealt with by moral teaching or church discipline. The chiefs, however, who had become deacons of the church, could not let such offences go without inflicting material punishment. A list of fines to be paid in money and trade was instituted. Some elements of the Western culture of the early white missionaries were evidently adopted, for accused persons were put in the stocks to await trial, and women found guilty of sex offences were drummed through the village after they had been fined. If an offender was caught on a Saturday evening, he or she remained in the stocks all day Sunday, for the Sabbath could not be desecrated by the holding of a court on that day. The administration of the laws was in the hands of people termed Turimen who were elected annually by the heads of households. A Turiman held the double office of police and judge. In Manihiki, Turimen were divided into four vigilance committees which took alternate weeks of duty. An offender was reported by a Turiman to his committee, and the committee promptly exacted the fine laid down for the offence. There was no formal trial and the accused had no appeal. Half of the fines went to the <hi rend="i">ariki</hi>, and the other half was divided among the Turimen. The system led to espionage of sexual behavior in order that the amount of fines might be increased as much as possible, and a mean spirit foreign to Polynesian psychology was engendered. However, in the curious adjustment that took place was the retention of a Polynesian trait, inasmuch as the relatives of the offender shared his burden by assisting him to pay the fine in order to save the family name.</p>
          <p>Moss (20, pp. 117, 118) sums up the situation as follows:</p>
          <p>The laws are objectionable, and their modes of enforcing them, putting men and women in the public stocks or drumming them through the public street, are bad enough; but the methods of prevention and discovery which these Turimen adopt, are worse. If a Turiman suspects a man of having taken liquor, he will stop him at any time and order him to “blow” so that he may discover if his breath has lost its normal sweetness.
<pb xml:id="n18" n="10"/>
The decision then come to is conclusive, adopted as a judgment by his fellow Turimen and the culprit fined accordingly. A “curfew” drum is beat at eight o'clock, and after that hour if anyone is seen abroad the Turimen are down upon him with a heavy fine next day. Their lovely moonlight nights bring no enjoyment to these people.</p>
          <p><name type="person" key="name-209643">Henry Williams</name>, Sr., informed me that three Peruvian slavers visited the atolls, but as one ship came within an ace of being wrecked by drifting onto the reef, the slavers withdrew without effecting depredations as they did in Tongareva.</p>
          <p>The notorious Captain Bully Hayes in the brig <hi rend="i">Rona</hi> foundered at sea near Manihiki, according to Moss (20, p. 86). Captain Hayes was kindly treated by the Manihikians, who helped him to build another small craft. The vessel completed, he started with a party of Manihikians for a marriage feast at Rakahanga. He purposely missed Rakahanga, however, and made Samoa. He induced the Manihikians to work on one of the plantations and charged their employer a good round sum for bringing laborers to Samoa.</p>
          <p>In marked contrast to scoundrels of the Hayes type were men of the stamp of Henry Greig and <name type="person" key="name-012338">George Ellis</name>. Greig employed Manihikian labor on Fanning Island. He married a Manihikian woman of high rank, and his descendants on Manihiki are respected people. <name type="person" key="name-012338">George Ellis</name> was a trader on Manihiki and taught the people much in the way of useful crafts. His two sons, Ben and Dan, are leaders in the atoll.</p>
          <p>A certain amount of friction existed at one time between the supporters of the native missionary and a faction opposed to him. The French at Tahiti were invited to annex the atolls, but when the French warship appeared for this purpose the native missionary hauled up the British flag and dared anyone to lower it. The crisis passed, and French annexation did not materialize.</p>
          <p>In the copra and pearl shell trade which developed, trading firms established touch with Manihiki and Rakahanga from Rarotonga. Thus for both religious and commercial purposes the atolls were connected with Rarotonga. As the result of a petition in 1900 by the <hi rend="i">ariki</hi> of the Cook Islands, Manihiki and Rakahanga were included in the boundaries of New Zealand by an Imperial Order in Council in 1901. The atolls are governed by New Zealand through the Resident Commissioner at Rarotonga. <name type="person" key="name-209643">Henry Williams</name>, Sr., who is of part-Manihikian blood, is Government Agent for the two atolls. He is assisted in each atoll by a council of eight nominated persons. The villages in the two atolls are models of orderly arrangement and cleanliness and are not surpassed in any of the Cook Islands.</p>
        </div>
        <pb xml:id="n19" n="11"/>
        <div xml:id="t1-body-d1-d1-d5" type="section">
          <head>
            <hi rend="sc">The People</hi>
          </head>
          <p>The population given in the last five government censuses is as follows:</p>
          <p>
            <table rows="4" cols="9">
              <row>
                <cell/>
                <cell/>
                <cell/>
                <cell/>
                <cell/>
                <cell/>
                <cell role="label" rend="center">1926</cell>
                <cell/>
                <cell/>
              </row>
              <row>
                <cell/>
                <cell role="label" rend="center">1906</cell>
                <cell role="label" rend="center">1911</cell>
                <cell role="label" rend="center">1916</cell>
                <cell role="label" rend="center">1921</cell>
                <cell role="label" rend="center">Males</cell>
                <cell role="label" rend="center">Females</cell>
                <cell role="label" rend="center">Whites</cell>
                <cell role="label" rend="center">Total</cell>
              </row>
              <row>
                <cell>Manihiki</cell>
                <cell rend="right">521</cell>
                <cell rend="right">444</cell>
                <cell rend="right">493</cell>
                <cell rend="right">432</cell>
                <cell rend="right">214</cell>
                <cell rend="right">199</cell>
                <cell rend="right">3</cell>
                <cell rend="right">416</cell>
              </row>
              <row>
                <cell>Rakahanga</cell>
                <cell rend="right">352</cell>
                <cell rend="right">315</cell>
                <cell rend="right">295</cell>
                <cell rend="right">310</cell>
                <cell rend="right">172</cell>
                <cell rend="right">153</cell>
                <cell rend="right">2</cell>
                <cell rend="right">327</cell>
              </row>
            </table>
          </p>
          <p>Although during the 20 years from 1906 to 1926 the total population of the administrative territory of the Cook Islands has shown an increase, the population of Manihiki has decreased by 115 and that of Rakahanga by 25. Part of the decrease has been due to emigration to Rarotonga, where the better opportunities of obtaining employment have attracted settlers. This emigration has resulted in a Manihikian colony in Rarotonga.</p>
        </div>
        <div xml:id="t1-body-d1-d1-d6" type="section">
          <head>
            <hi rend="sc">The Language</hi>
          </head>
          <p>The language is a pleasing dialect and has closer affinities with Maori than with the dialects of Tongareva, Tahiti, and the Cook Islands. The dialect differs from Tongarevan in using <hi rend="i">h</hi> instead of <hi rend="i">s</hi> and <hi rend="i">wh</hi> instead of <hi rend="i">h</hi>, from Tahitian in retaining <hi rend="i">k</hi> and <hi rend="i">ng</hi> and using <hi rend="i">wh</hi> instead of <hi rend="i">f</hi>, and from Cook Islands dialects in the presence of <hi rend="i">wh</hi> and a more definitely sounded <hi rend="i">h.</hi> All these differences are shared by the Maori dialect. Also, a number of words that are not shared by the Rarotongan and Tahitian dialects are common to the Maori and the Manihiki-Rakahangan dialects.</p>
          <p>The native pastors, educated by the London Missionary Society in Rarotonga, have introduced the alphabet adopted for Rarotonga. This alphabet is without the letter <hi rend="i">h</hi>, and <hi rend="i">v</hi> is used instead of <hi rend="i">w.</hi> Thus both the <hi rend="i">h</hi> and <hi rend="i">wh</hi> sounds which are present in the dialect have no letters to represent them. No organized effort has been made by the church or the state to remedy the deficiencies. As in others of the Cook Islands, the local dialect is being assimilated rapidly by the Rarotongan dialect. The Bible, which is printed in Rarotongan, exercises a great influence in standardizing Rarotongan as the accepted dialect.</p>
          <p>The alphabet in use contains the vowels <hi rend="i">a, e, i, o</hi>, and <hi rend="i">u</hi>, and the consonants <hi rend="i">k, m, n, ng, p, r, t</hi>, and <hi rend="i">v.</hi></p>
          <p>The consonants not represented are <hi rend="i">h</hi> and <hi rend="i">wh</hi>, and the <hi rend="i">v</hi> should be <hi rend="i">w.</hi> <name type="person" key="name-121166">Stephen Savage</name>, official interpreter to the Cook Islands Administration, holds that <hi rend="i">w</hi> should have been adopted for the Rarotongan dialect instead of <hi rend="i">v.</hi> This applies with even more force to Manihikian. At the same time, there may be some such words as <hi rend="i">vero</hi> (stern piece of a canoe) that are pronounced with a <hi rend="i">v</hi> sound. The older people pronounce the words with a <hi rend="i">w</hi> and write them with a <hi rend="i">v.</hi> With the modern method of teaching by <choice><orig>alpha-
<pb xml:id="n20" n="12"/>
betical</orig><reg>alphabetical</reg></choice> sounds, the tendency is for the younger people to adopt the <hi rend="i">v</hi> sound as taught to them. Europeans have recorded the <hi rend="i">h</hi> sound in Manihiki on official maps and have omitted the equally obvious <hi rend="i">h</hi> sound in Rakahanga by printing it “Rakaanga.” The people pronounce the name of their atoll “Manihiki” and write it “Maniiki” because the schools do not include the <hi rend="i">h</hi> when teaching the alphabet. An extra emission of the breath gives <hi rend="i">h</hi> the sound <hi rend="i">hi</hi> before the regular vowel, and it has become usage to say “hi,” as in “Hiuku,” for Huku, a word variously written as “Iku,” “Hiku,” and “Huku.” Smith, in editing Gill's account of the origin of Manihiki (13, p. 140), states that the name should be spelled “Hiku”; but though this represents the name as it would appear in other dialects, the local pronunciation is “Hiuku.” The people, not having been provided with the letter <hi rend="i">h</hi> by the teachers of the Rarotongan alphabet, usually spelled it “Uku,” or even “Iuku.” It is, perhaps, more convenient to spell it “Huku,” but the correct pronunciation must be borne in mind (see page <ref target="#n22">14</ref>). This usage resembles that of the Rarawa tribe of the Maoris of northern Auckland, who have a tendency to use <hi rend="i">he</hi>, as in “Heone” for Hone, a pronunciation used by older people but not followed by the younger generation of Maoris.</p>
          <p>The <hi rend="i">wh</hi> sound has been recorded for New Zealand and the Chatham Islands by the double letter <hi rend="i">wh.</hi> Of this consonant Williams says (31, p. 568):</p>
          <p><hi rend="i">Wh</hi> represents the voiceless consonant corresponding with <hi rend="i">w</hi>, and is produced by emitting the breath sharply between the lips. It is a mistake to assimilate the sound to that of <hi rend="i">f</hi> in English, though it has become fashionable in recent years with some of the younger Maoris. In some words <hi rend="i">wh</hi> and <hi rend="i">h</hi> are interchangeable, as <hi rend="i">kohatu, kowhatu; mahiti, mawhiti.</hi> In a few words there is confusion between <hi rend="i">wh</hi> and <hi rend="i">w</hi>, but this may be due to the fact that in early works printed in Maori no distinction was made between the two, both being printed as <hi rend="i">w.</hi> <hi rend="i">Wh</hi> is never found in Maori followed by o or <hi rend="i">u.</hi></p>
          <p>It was evident to both natives and Europeans that an extra sound not provided for by the Rarotongan alphabet was present. The mistake of assimilating the sound to that of <hi rend="i">f</hi> in English was committed by Europeans, and the few natives who write have followed suit. Thus we have the word for hala (<hi rend="i">Pandanus</hi>) written as “fara” and it appears in <ref target="#BucMani006a">figure 2</ref> as Motu Fara (<hi rend="i">Pandanus</hi> Island). A certain amount of influence may have come from Tahiti, where the sound exists as an actual <hi rend="i">f</hi>, and the word is pronounced “fara.” Knowing this, I was prepared to accept the sound as <hi rend="i">f</hi> until I heard the words pronounced in the atolls. While I was recording pedigrees in the Land Court at Rakahanga it became evident to me that the sound was not the English or Tahitian <hi rend="i">f</hi> but resembled the Maori <hi rend="i">wh.</hi> However, lest my own Maori background might have influenced me, I asked <name type="person" key="name-121166">Stephen Savage</name> and <name type="person" key="name-101696">Henry Williams</name>, Jr., who is of Manihikian extraction,
<pb xml:id="n21" n="13"/>
to check up on the words containing the sound. They agreed that the sound was <hi rend="i">wh</hi> and not <hi rend="i">f.</hi></p>
          <p>The remarks of Williams (31) about the confusion between the Maori <hi rend="i">wh</hi> and <hi rend="i">w</hi> apply with still more force to the Manihikian <hi rend="i">wh</hi> and <hi rend="i">v.</hi> The <hi rend="i">v</hi> is wrong in the first place, but it was the only letter that could be used to represent both the <hi rend="i">w</hi> and <hi rend="i">wh</hi> sounds, until a few people began to use <hi rend="i">f.</hi></p>
          <p>The interchange in some Maori words between <hi rend="i">wh</hi> and <hi rend="i">h</hi>, noted by Williams, applies also to inter-dialectical variations. Although the distinction between <hi rend="i">wh</hi> and <hi rend="i">f</hi> is marked as regards sound and the position of the lips and teeth, the fashionable interchange in recent years by the younger Maoris has evidently followed a general Polynesian tendency. K. P. Emory of Bernice P. Bishop Museum informs me that the Tahitian <hi rend="i">f</hi> sound prevails over most of the Tuamotuan archipelago, but that at Reao in the east the <hi rend="i">wh</hi> sound is used. This was checked by Mr. Emory and F. J. Stimson, both of whom were accustomed to the Tahitian <hi rend="i">f.</hi> Interchanges have thus occurred between <hi rend="i">h</hi>, <hi rend="i">wh</hi>, and <hi rend="i">f.</hi> A good test word is the widespread Polynesian name for house, which consists of one of the three interchangeable consonants followed by <hi rend="i">are</hi> or <hi rend="i">ale</hi>, according to the dialectical selection between <hi rend="i">r</hi> and <hi rend="i">l.</hi> The main dialects interchange as follows:</p>
          <p>
            <table rows="6" cols="3">
              <row>
                <cell role="label" rend="center">Hare</cell>
                <cell role="label" rend="center">Whare</cell>
                <cell role="label" rend="center">Fare</cell>
              </row>
              <row>
                <cell>Cook Islands</cell>
                <cell>New Zealand</cell>
                <cell>Society Islands</cell>
              </row>
              <row>
                <cell>Tongareva</cell>
                <cell>Chatham Islands</cell>
                <cell>Tuamotus</cell>
              </row>
              <row>
                <cell>Hawaii (hale)</cell>
                <cell>Manihiki</cell>
                <cell>Marquesas (fa'e)</cell>
              </row>
              <row>
                <cell/>
                <cell>Reao (Tuamotus)</cell>
                <cell>Samoa (fale)</cell>
              </row>
              <row>
                <cell/>
                <cell/>
                <cell>Tonga (fale)</cell>
              </row>
            </table>
          </p>
          <p>Although <hi rend="i">wh</hi> occupies the intermediate position between <hi rend="i">h</hi> and <hi rend="i">f</hi>, a direct interchange between <hi rend="i">h</hi> and <hi rend="i">f</hi> is seen in the eastern Polynesian word <hi rend="i">aroha</hi> and the western word <hi rend="i">alofa.</hi> It is tempting to think of <hi rend="i">h</hi> as the simplest, oldest form, retained in the northern remote area, Hawaii, and surviving in Tongareva and Cook Islands; of the <hi rend="i">wh</hi> as an old form retained in the southern remote area of New Zealand and Chatham Islands, the remote eastern area of Reao, and surviving in Manihiki; and of the <hi rend="i">f</hi> as coming in as a later intrusive element from the west, establishing itself in the Society Islands, when it spread through the Tuamotus and Marquesas, displacing the <hi rend="i">wh</hi> but failing to extinguish it in the far east at Reao.</p>
          <p>Some inconsistency in the spelling of native words will be observed in this study. In quotations from manuscripts or printed works the original spelling has been kept, but in my own observations the <hi rend="i">h</hi> and <hi rend="i">wh</hi> have been used in words in which they were sounded. As in Tongareva, a study of the dialect and local vocabulary awaits the linguist.</p>
        </div>
      </div>
      <pb xml:id="n22" n="14"/>
      <div xml:id="t1-body-d1-d2" type="section">
        <head>
          <hi rend="c">Traditional History</hi>
        </head>
        <div xml:id="t1-body-d1-d2-d1" type="section">
          <head>
            <hi rend="c">Discovery</hi>
          </head>
          <p>The traditional story of the discovery of Rakahanga and Manihiki is a blend of historical narrative and myth. The human discoverer, Huku (see p. <ref target="#n27">19</ref>), is stated to have sailed from Rarotonga on a fishing expedition. When he came to a part of the ocean referred to as “te tukuanga i Whaka-hotu” he noticed an upgrowth of rock or land (<hi rend="i">tapua whenua</hi>) projecting from the sea bottom but not rising above the water, an image evidently culled from the experience of an atoll-dwelling people to whom coral upgrowths on an encircling reef were familiar. Huku, on seeing the coral upgrowth, recited these words (<hi rend="i">pehe</hi>), which are always recited by native historians: “Titiro iho Huku, tapua e—” (Huku gazed down, [and saw] an upgrowth.) From this incident Huku named his canoe <hi rend="i">Tapua.</hi> Gill's informant (13, p. 140) gave the canoe name as <hi rend="i">Tapuaua</hi>, which is really <hi rend="i">Tapua-hua</hi> and means “an upgrowth only.” Huku returned to his home with the idea that the upgrowth would eventually reach the surface and become land.</p>
          <p>Pure myth is introduced by interpolating the island-fishing exploits of the well-known Polynesian hero, Maui. The myth states that Tangaroa-tuhi-mata, with his wife, Hina-mata-porari, dwelt in Hawaiki-ki-raro, which was under the earth's surface. Their son, Tongoi-whare, dwelt with his wife, Makuwai-whare, in Hawaiki-ki-runga, which was on the earth's surface. Tongoi-whare, had three sons, Maui-mua, Maui-roto, and Maui-muri, and a daughter named Hina-mai-raro-te-takere. The three brothers planned a fishing expedition and caught some flying fish (<hi rend="i">maroro</hi>) for bait. Maui-muri, unknown to his brothers, went to Hina-i-te-papa who dwelt at the bottom of the sea. He told her of the proposed expedition and asked her to put certain fish on the hooks according to the manner in which they were baited. On the first hook let down, baited with flying fish, she was to put a shark (<hi rend="i">mango</hi>), and on the second hook, also baited with flying fish, she was to put an <hi rend="i">urua.</hi> His own hook was to be baited with a small branch of the <hi rend="i">puka</hi> tree (<hi rend="i">tauru raupuka</hi>), dried coconut husk (<hi rend="i">puakoua</hi>), coconut flower stalk (<hi rend="i">puroro</hi>), and a dry, immature young coconut (<hi rend="i">aoa</hi>). When she saw it, she was to hook it into the rock bottom of the sea.</p>
          <p>The next day the brothers set out in a canoe named <hi rend="i">Pipi-ma-hakohako.</hi> Tupou-rahi stated that the name of the canoe was <hi rend="i">Whakahotu</hi> and quoted the following chant as proof:</p>
          <lg type="verse">
            <l>Wharekura-ariki i noho ia Hina-i-te-papa,</l>
            <l>Nohona te waka o nga atua;</l>
            <l>Whakahotu, nohona te tukunga.</l>
            <l>To rire to, e tapu.</l>
          </lg>
          <lg type="verse">
            <l>Wharekura the ariki married Hina-i-te-papa,</l>
            <l>His was the canoe of the gods;</l>
            <l>Whakahotu, his was the fishing ground.</l>
          </lg>
          <pb xml:id="n23" n="15"/>
          <p>Tupou-rahi has evidently associated the name Whakahotu with the canoe, whereas it is really associated with the fishing ground. The historic spot in the ocean visited by Huku was termed “Te tukunga i Whakahotu.” The construction of the chant shows that the <hi rend="i">tukunga</hi> (fishing ground) was owned by Whakahotu. Although Whakahotu does not appear elsewhere in Rakahangan tradition, in neighboring Tongareva (29) the name appears as Hakahotu. Hakahotu was the wife of Atea, and the two were the primary parents whose progeny were the gods Tangaroa, Tane, Rongo, and others. It may be that Rakahanga had some legend of Whakahotu similar to the Tongarevan story, and that the details have been lost.</p>
          <p>Arrived at the fishing ground, Maui-mua (Maui-the-first-born) let down his hook baited with flying fish. When a fish took the hook, he began to chant a question (<hi rend="i">tautopa</hi>):</p>
          <lg type="verse">
            <l>Maui-roto, Maui-muri, e uia mai</l>
            <l>Te ingoa i taku ika, mei aha?</l>
          </lg>
          <lg type="verse">
            <l>Maui the middle, Maui the last,</l>
            <l>ask The name of my fish, what is it?</l>
          </lg>
          <p>Maui-roto remained silent but Maui-muri, relying on his compact with Hina-i-te-papa, replied:</p>
          <lg type="verse">
            <l>E haha mango tau ika tutae, hutia!</l>
            <l>A large shark is your filthy fish, haul it up!</l>
          </lg>
          <p>A shark was duly hauled up. Tupou-rahi said that the fish was an albacore (<hi rend="i">kakahi</hi>). Maui-roto let down his hook baited with flying fish, and when he hooked a fish he also chanted a question. Maui-muri named an <hi rend="i">urua</hi>, which was duly hauled in. Maui-muri then baited his hook with the leaves and husk according to plan. Hina-i-te-papa, on seeing the hook so baited, stuck it into the <hi rend="i">papa</hi> or rock at the bottom of the sea. From what follows, it is to be inferred that the rock was the upgrowth seen by Huku. When Maui-muri felt the hook take hold, he called (<hi rend="i">uru</hi>) a question to his elder brothers, who named a shark and an <hi rend="i">urua.</hi> Maui-muri thereupon hauled on his line. As the fish rose, the sea began to boil and foam (<hi rend="i">kua wheta te moana</hi>), and land appeared (<hi rend="i">kua haha te whenua</hi>). The rising land lifted the canoe on its steep edge. Maui-mua and Maui-roto were in the bow and Maui-muri in the stern. The canoe broke in the middle. The bow part with the two elder brothers fell into the seething water, and they were swept away. The stern part remained on the land, and Maui-muri stepped ashore. He then recited a chant (<hi rend="i">amu</hi>):</p>
          <lg type="verse">
            <l>Tataka. tataka e, tataka ki muri,</l>
            <l>Tataka, tataka e, tataka ki muri.</l>
            <l>Tokomiti, tokomiti,</l>
            <l>Tokowheta, tokowheta.</l>
            <l>Haha, haha te whenua,</l>
            <l>Tutu Maui.</l>
          </lg>
          <lg type="verse">
            <l>Fall off, fall off, fall off to the stern,</l>
            <l>Fall off, fall off, fall off to the stern.</l>
            <l>The sea recedes,</l>
            <l>The sea seethes.</l>
            <l>It appears, the land appears,</l>
            <l>And Maui stands upon it.</l>
          </lg>
          <pb xml:id="n24" n="16"/>
          <p>In the chant recorded by Gill (13, p. 141) the words “Motu Manihiki, motu Rakahanga” (Manihiki is severed, Rakahanga is severed) are added, but they are not severed from each other until the later reappearance of Huku. Maui-muri, having lost his brothers and his canoe, explored the new land. A curious reference is made to his seeing a house of sand (<hi rend="i">whare one</hi>). Gill (13, p. 148) states that the house contained eighty spirits. Tupou-rahi stated that Maui saw an opening in the sand and dug down. The hole contained ugly fish from the bottom of the sea (<hi rend="i">ko te au ika kikino anake i raro i te moana</hi>). Such details were evidently intended to embellish the tale and perhaps to emphasize the fact that there were no human occupants on the land discovered. Maui-mua remained on the land.</p>
          <p>The story goes back to Huku, who, in his sleep (<hi rend="i">turamoe</hi>), had a dream (<hi rend="i">rikamoe</hi>) that the upgrowth he had seen had reached the surface. His dream was expressed in the chant:</p>
          <lg type="verse">
            <l>Whakarika mai ana, turamoe ana,</l>
            <l>Turamoe ia Rakahanga, ia Rakahanga.</l>
            <l>Kua haha, kua haha,</l>
            <l>Kua roharoha.</l>
          </lg>
          <lg type="verse">
            <l>It came as a dream while sleeping,</l>
            <l>A dream about Rakahanga, about Rakahanga,</l>
            <l>Which has emerged and risen,</l>
            <l>And lies spread out.</l>
          </lg>
          <p>Tupou-rahi states that Huku did not think out the name Rakahanga, but that the name came to him in a dream.</p>
          <p>Huku sailed back to the site of the upgrowth and saw that the land had risen above the surface. He landed at Waiawa and commenced to explore. He met Maui-muri and immediately attacked him to expel him from the land which he considered to be his. In the struggle a portion of the land broke off. It floated away (<hi rend="i">rewa atu</hi>) and became Manihiki. Maui fled successively to the places on the atoll named Tumu-whenua, Kaeru, Tumukau, and Paaki. As Huku pursued Maui to Paaki, the rain fell (<hi rend="i">kuru te ua</hi>), the lightning flashed (<hi rend="i">rapa te uira</hi>), the thunder rolled (<hi rend="i">tuki te whatitiri</hi>), and Maui fled away to the heavens (<hi rend="i">kua rere Maui ki te rangi</hi>).</p>
          <p>After his victory, Huku, returning along the coast, saw a drift coconut (<hi rend="i">ponga</hi>) cast ashore. After his return to Nukuangaanga he planted the nut at a spot which he named Te-maru-o-araiawa. He named the nut, or the plant which was to grow from it, Te-huru-awatea. In Gill's account (13, p. 148) no mention is made of a drift coconut. Gill states that Huku returned to Rarotonga because the land was desert (<hi rend="i">ha</hi>) and no coconuts had yet been planted. The coconut named Te-huru-awatea he includes with seven others that were subsequently brought. The story of the drift coconut as given by Tupou-rahi fits in better with the subsequent narrative.</p>
          <p>According to Tupou-rahi's version, Huku, after planting Te-huru-awatea, returned to Rarotonga. He then made a third voyage in a canoe named
<pb xml:id="n25" n="17"/>
Hotu-rangaranga, taking a supply of planting nuts with him. He also took two paddlers (<hi rend="i">hoe waka</hi>) named Ruia and Papera. Seven nuts named Tiro-hanga, Turuki-wairaro, Papuka, Kai-akuaku, Tumata-whare, and Nuku-angi-angi were planted at Te-maru-o-araiawa. Gill (13, p. 148) gives names which differ slightly, but includes Te-huru-awatea, the name given by Tupou-rahi to the drift nut of a previous voyage. Te-maru-o-araiawa, where the nuts were planted, is situated on the island of Te Kainga in the atoll of Raka-hanga. The two paddlers both died and were buried on Te Kainga. Tupou-rahi quoted this <hi rend="i">pehe</hi> as proof of his tale:</p>
          <lg type="verse">
            <l>Tanu Ruia tei te turuki,</l>
            <l>Tanu Papera tei te paapuka.</l>
            <l>Haroi ha.</l>
          </lg>
          <lg type="verse">
            <l>Ruia was buried at the <hi rend="i">turuki</hi>,</l>
            <l>Papera was buried at the <hi rend="i">paapuka.</hi></l>
          </lg>
          <p>Some of the Rakahangans were inclined to think that the inclusion of the paddlers in the tale was an elaboration, as <hi rend="i">ruia</hi> and <hi rend="i">papera</hi> are names given to two species of shark.</p>
          <p>Huku returned to Rarotonga. Sitting one evening in front of his house, he began to wonder whether the first nut he had planted on Rakahanga had grown into a stately palm. His thoughts found expression in the following chant (<hi rend="i">haka</hi>):</p>
          <lg type="verse">
            <l>Tera paa Te-huru-awatea te tahirihiri mai ra</l>
            <l>i Te-maru-o-araiawa.</l>
            <l>Perchance Te-huru-awatea is waving its fronds</l>
            <l>over there at Te-maru-o-araiawa.</l>
          </lg>
          <p>The literal translation of the <hi rend="i">haka</hi> conveys merely the thought, but to the descendants of Huku there is music and sentiment in the native words. They take pride in quoting them over and over again. During my stay in Rakahanga the words were called out to me from houses along the village street, for they formed a bond of our mutual interest in the past. When I called them out to my history teachers, their gratification was apparent.</p>
          <p>Huku had kept his discovery of a new land secret. A man named Wheatu heard Huku and conjectured that the chant referred to some other land. He embarked in a canoe named Paparinga-tahi, eventually came to Manihihi, and landed at Tarakite. From there he set out and came to Rakahanga, where he saw the palm named Te-huru-awatea waving in the breeze at Te-maru-o-araiawa. He landed at Omoka and dragged his canoe ashore, naming the place Te-amonga-waka (the carrying of the canoe). He went on to a place on the reef called Awanui. There he commenced to cut a channel through the reef to connect the lagoon with the sea and so to provide a passage for canoes. As he worked he sang the following chant, which Gill (13, p. 142) has recorded in full:</p>
          <pb xml:id="n26" n="18"/>
          <lg type="verse">
            <l>Ana mai, ana mai, kurua!</l>
            <l>Ana mai ko Wheatu, kurua!</l>
            <l>E ano ki Rakahanga,</l>
            <l>Kurua iho, kurua e koe,</l>
            <l>Kurua te papa i Awanui,</l>
            <l>Kurua, kurua, kurukurua!</l>
          </lg>
          <lg type="verse">
            <l>E noho i Tarakite, kurua!</l>
            <l>Takahia e koe, kurua!</l>
            <l>Te matangi ko te tonga,</l>
            <l>Kurua, kurua, kurukurua!</l>
          </lg>
          <lg type="verse">
            <l>Ki waenga moana</l>
            <l>Kurua ko te mata o Wheatu,</l>
            <l>Kurua te uru o Rakahanga,</l>
            <l>Kurua te awa i Omoka,</l>
            <l>Kurua, kurua, kurukurua!</l>
          </lg>
          <lg type="verse">
            <l>Come along, come along, batter away!</l>
            <l>Come along, it is Wheatu, batter away!</l>
            <l>Go to Rakahanga,</l>
            <l>Batter it down, batter O thou,</l>
            <l>Batter the rock at Awanui.</l>
            <l>Batter, Oh, batter, Oh, batter away!</l>
          </lg>
          <lg type="verse">
            <l>Stay at Tarakite, batter away!</l>
            <l>Stamp on it, thou, and batter away!</l>
            <l>The wind is the south wind,</l>
            <l>Batter, Oh, batter, Oh, batter away!</l>
          </lg>
          <lg type="verse">
            <l>In the midst of the ocean,</l>
            <l>Batter the face of Wheatu,</l>
            <l>Batter the head of Rakahanga,</l>
            <l>Break out the channel at Omoka,</l>
            <l>Batter, Oh, batter, Oh, batter away!</l>
          </lg>
          <p>The words <hi rend="i">kurua, kurua, kurukurua</hi> of the refrain mean literally “to batter,” for Wheatu used a piece of coral boulder to hammer away at the reef and break pieces off in his attempt to cut out a channel. In view of the fact that the Cook Islands Administration has considered the advisability of cutting channels through the reef, the chant is interesting, showing that the problem exercised the minds of the early Polynesian settlers. The early voyager, Te Herui, is credited by tradition with having cut the passage called Te-rua-i-kakau through the reef at Aitutaki, but he had the advantage over Wheatu of having an adz. Another ancestor, Ruatapu, commenced a similar engineering task on the reef at Atiu, but as the food supplies of his employer fell short, he abandoned the work.</p>
          <p>Meanwhile, Huku had a premonition that something was happening on the land he regarded as his. According to Tupou-rahi, Huku set sail in a canoe named Te Rawhiti and was accompanied by his sister, Tapairu, and her husband, Toa. According to Gill (13, p. 149), Huku came without them. In Rakahanga, Huku found Wheatu still battering away at the reef at Awanui. To Huku's fierce inquiry as to what he was doing on his land, Wheatu diplomatically replied, “I am battering this rock to make a canoe channel for you.” (<hi rend="i">E tuki ana au i te papa nei tei ara waka nohou.</hi>) According to one version, Wheatu was allowed to remain on condition that he did not go inland, as Huku was afraid he might pull up the coconuts that had been planted. The version adopted by the people in their historical dramas (see p. <ref target="#n206">198</ref>) is that Wheatu was driven off the island. A short, wide indentation on the inner side of the reef at Awanui is pointed out as the scene of Wheatu's labor, and the opinion is held that he would have completed the channel had he not been driven away. The name of Wheatu does not appear in the local genealogies.</p>
          <pb xml:id="n27" n="19"/>
          <p>Huku subsequently departed for Rarotonga, leaving the land in charge of his sister and her husband. Gill (13, p. 149) states that Huku sent his sister and her husband to Rakahanga after his return to Rarotonga, and that the canoe in which they came was named Reiapata. Huku here disappears from history, and the peopling of the atoll commences with Toa and his wife.</p>
          <p>Huku is generally believed by the people of Rakahanga and Manihiki to have come from Rarotonga. Tupou-rahi stated that he belonged to the land division of Tukuvaine, that Ikurangi was his chiefly mountain (<hi rend="i">maunga ariki</hi>), and that his god, Mokoroa-i-taupo, was located at Maungatea. Tukuvaine, Ikurangi, and Maungatea are all well-known localities close to the main village, Avarua, in Rarotonga. So much communication, however, has existed between the atolls and Rarotonga since the advent of Christianity that the possibility of elaboration between the two peoples cannot be disregarded. Mr. Savage told me that Huku appears as “Iku” in the Rarotongan genealogies, but careful checking up is required before the two can be accepted as identical. The number of voyages Huku is reputed to have made between Rarotonga and Rakahanga is probably exaggerated.</p>
          <p>The interpolation of the Maui myth is interesting in view of the dialectical affinity between the atolls and New Zealand. Both areas have the story of Maui fishing up the land with a hook and line. The New Zealand story differs in that it presents five Maui brothers instead of three. In both areas the hero of the adventure is the youngest son, Maui-potiki in one, and Maui-muri in the other. Both <hi rend="i">potiki</hi> and <hi rend="i">muri</hi> refer to the last-born of a family. In the New Zealand story the hook entered a house at the bottom of the sea and caught on the back of the house, which suggests the Rakahangan incident in which Hina-i-te-papa hooked the line into the <hi rend="i">papa</hi> or rock where she dwelt at the bottom of the sea. In outline the two stories show an affinity, but they naturally differ in minor details.</p>
          <p>Toa and Tapairu settled as permanent colonists on the island of Te Kainga in the atoll of Rakahanga. Tradition states that Tapairu was the sister of Huku and that it was through this blood kinship that Huku gave her the land. Pedigrees of the ancestry of both Tapairu and Toa are lacking. Some informants stated that Huku and Tapairu were the children of Hiro of Hawaiki, but Gill (13, p. 143) records that Tapairu was the daughter of a son of Hiro who came from Hawaiki, presumably, to Rarotonga. An interesting complication is due to the definite mention of a son of Hiro in Rarotongan history. Hiro, the celebrated explorer, was a contemporary of Tangiia, the great Rarotongan ancestor. Tangiia adopted Tai-te-ariki, a son of Hiro, and made him high chief over a section of his people. The Pa-ariki title of the Takitumu tribes of east and south <choice><orig>Raro-
<pb xml:id="n28" n="20"/>
tonga</orig><reg>Rarotonga</reg></choice> descends from Tai-te-ariki. If Huku and Tapairu belonged to Rarotonga, they should have been associated with the Takitumu people. Tupourahi, however, places them in Tukuvaine in north Rarotonga, which is Makean territory, derived historically from a totally different family. It is evident, however, that the Rakahangan historians knew of Hiro, but that knowledge could have been derived from other areas besides Rarotonga. They may have followed the strong tendency that existed among genealogists to connect their ancestors with celebrated historical characters, thus adding luster to their pedigrees.</p>
          <p>According to Gill (10, vol. 2, pp. 281, 282), Toa was a great warrior but was defeated in Rarotonga. Gill gives this defeat as the reason why Toa and his wife came to settle in the land discovered by Iku (Huku). He says further, “The Rarotongan chiefs confirm this testimony by relating the departure of ‘Toa’ from this island many generations ago. He was a great warrior of the Ngati-Tinomana tribe.” The tribe referred to inhabits the western part of Rarotonga. Gill also mentions the habit of saying something after sneezing, as though addressing a spirit. On Manihiki it was usual to say, “Alas! alas! Go to Rarotonga.” The evidence cited would appear to confirm Rarotonga as the original habitat of Toa and Tapairu.</p>
        </div>
        <div xml:id="t1-body-d1-d2-d2" type="section">
          <head>
            <hi rend="c">Genealogical Records</hi>
          </head>
          <p>Because traditional history in Rakahanga attributes the settlement of the atolls to one biological family, the head of that family a defeated warrior accompanied by no priests, the genealogical records of the two atolls thus peopled are poor in extent and detail. As in Tongareva (29, p. 16), the genealogical records cover three periods, settlement, exploratory or migratory, and mythical. The settlement period commenced with the historical settlement of the land and extends to the present time. The exploratory phase covered the period from the departure from an ancient homeland to the beginning of the present occupation. The mythical period dealt with the creation of the gods and the origin of human life. Even in the mythical period the order of creation and natural phenomena were personified and arranged in genealogical order. In the preservation of genealogies, an outstanding element of all Polynesian culture, native scholars have provided a chronological skeleton for traditional history. Whereas the Western historian refers to a calendar date of solar years, the Polynesian refers, then, to the number of generations of human beings (<hi rend="i">uki tangata</hi>) in dating past events. Because the genealogies covering the three periods are elaborate, their preservation and teaching were delegated to persons trained to feats of memory. These people, in Polynesia, formed the educated priesthood. It can <choice><orig>there-
<pb xml:id="n29" n="21"/>
fore</orig><reg>therefore</reg></choice> be understood that unless priests accompanied the first settlers to an island the historical records of that island are likely to be abridged and unsatisfactory. The noted explorers of high rank, accompanied by chiefs of status and skilled priests and astronomers, did not seek out small atolls. They preferred the high islands with ample area for distribution and fertility for the production of food supplies. The character of the explorers is reflected in the traditional history, genealogies, and social structure subsequently revealed by the field research of foreign students.</p>
          <p>In Rakahanga the only voyages of the long migratory period recorded are those of Huku and Wheatu from Rarotonga. Beyond the statement that Huku and Tapairu were either the children or the grandchildren of Hiro, no pedigrees are recorded to extend over the migratory period from the mythical past. The widespread myth of Maui was evidently one of the few things remembered, and it was interpolated to link with the recent period of Huku. No definite mythical period is recorded, nor are primary parents, represented in some areas by Atea and Papa, remembered. Whakahotu appears in the name of the spot in the ocean whence Rakahanga emerged, but though there is an indication that Whakahotu was personified, she does not appear as a primary nature mother as she does in Tongarevan myth. The god Tangaroa appears as the father of Maui as he does in Rarotongan myth. No reference is made, however, to his brothers, Tane, Rongo, and others who occupy such a prominent position in the Polynesian pantheon. The gods worshiped were local gods not elsewhere known. Rakahanga has preserved even less of Polynesian myth than Tongareva (29, p. 85). The first settler, Toa, must be held responsible for what knowledge of outside history and institutions was brought into the country. Rakahangan chiefs and heads of families subsequently learned the family pedigrees covering the settlement period, for inheritance and succession were based on these, but much has been forgotten since the advent of Christianity.</p>
          <p>The family pedigrees covering the settlement period were recorded for this study from evidence given by witnesses before the Native Land Court held in Rakahanga in June, 1929. All the witnesses were acquainted with the main facts concerning Toa's marriages, but only a few could trace a connected line from Toa to themselves. Most could trace pedigrees to ancestors through 6 to 12 generations but could not connect them with the main lines. The families of Toa's daughters were confused, and some ancestors were given different names and different parents.</p>
          <p>Out of this confusion a lineage given by the pastor, Kairenga, is selected, as it gives his lineal descent from Toa (Table 1). Another connected line was given by Haumata-tua, but it includes a list of persons who held one of the <hi rend="i">ariki</hi> titles. In pedigrees which include lists of <hi rend="i">ariki</hi>, younger brothers
<pb xml:id="n30" n="22"/>
who may have held the title are apt to be recited in a direct lineal descent, and the number of generations is thereby incorrectly increased. Unless the collaterals can be checked, it is safer to use a junior line to judge length of occupation.</p>
          <p>
            <figure xml:id="BucMani022a">
              <graphic url="BucMani022a.jpg" mimeType="image/jpeg" xml:id="BucMani022a-g"/>
              <head>Table 1. Descent from Toa</head>
            </figure>
          </p>
          <p>If the allowance of the Polynesian Society of 25 years to a generation is accepted, the 22 generations from Toa to Kairenga's adult nephew, Mata, would make a settlement period of 550 years. Toa would have landed in Rakahanga in about the middle of the fourteenth century. This date agrees with the date of the last migration to New Zealand and with the approximate date of settlement by Mahuta in Tongareva. Rarotonga was settled by Tangiia about a century earlier.</p>
          <p>The first part of Table 1 ends with Kanohi in the 10th generation. Kanohi married Tautape, the last <hi rend="i">ariki</hi> to rule singly over the atolls. The period of the dual <hi rend="i">ariki</hi> commences with the 11th generation and consists of
<pb xml:id="n31" n="23"/>
12 generations. The column on the right of the table gives the descent in the 11th generation from Te-patiti, a younger brother of Tuteru-matua. This line is three generations shorter than the line given in the left column. Two other lines which connected with the Kanohi-Tautape generation (10) are also about nine generations long, so the second period may be only a little more than 200 years, instead of 300 years as the Kairenga line would indicate. A discrepancy as to the length of the first period from Toa to Tautape on another line is discussed on page <ref target="#n54">46</ref>.</p>
          <p>A pedigree skeleton has now been provided to which collaterals may be linked as the story develops.</p>
        </div>
        <div xml:id="t1-body-d1-d2-d3" type="section">
          <head>
            <hi rend="c">Marriages of Toa</hi>
          </head>
          <p>The human population is referred to in Rakahanga as the <hi rend="i">kura tangata</hi>, the line of descent as the <hi rend="i">katiri tangata</hi>, and the biological family as the <hi rend="i">puna.</hi> Toa and Tapairu and their four children are referred to historically as the <hi rend="i">puna mua</hi> (the first family). In New Zealand <hi rend="i">puna</hi> means a wife and <hi rend="i">puna rua</hi>, a second wife. In Hawaii <hi rend="i">punalua</hi> means the two wives of one husband or the two husbands of one wife. The two husbands of one wife may be regarded as a late development with a limited distribution. To the Polynesian in general the dominant reason for marriage was the procreation of children to perpetuate the line of descent. Though to the native mind the term <hi rend="i">puna</hi> may apply particularly to the offspring, the parents are not entirely disassociated. The very mention of the <hi rend="i">puna</hi> recalls the parents, and the <hi rend="i">puna</hi> cannot be described without mentioning the parents—“Te puna mua a Toa raua ko Tapairu, e wha tamahine: ko Kae, ko Poe, ko Naunau, e Nanamu.” (The first family of Toa and Tapairu consisted of four daughters: Kae, Poe, Naunau, and Nanamu.) The children of this marriage were all girls, and their names are given in order of birth. Some pedigrees give a different order, making Poe first and Naunau last. Gill (10, vol. 2, p. 281) gives their names as Vai, Navenave, Pae, and Nannau, but as every consonant must be followed by a vowel, Nannau at least has been spelled incorrectly. Throughout this early period considerable confusion exists in the pedigrees given by different families.</p>
          <p>It is obvious that in the settlement of an uninhabited island by one biological family the perpetuation of the human stock must be continued through incestuous marriages. The members of Toa's first family were all females and he desired male issue (<hi rend="i">kua hinangaro ki te kapi tane</hi>). Had some of the first family been males, it is probable that brother and sister marriages would have been consummated to meet the problem. No sons having been born to Tapairu, Toa married his eldest daughter to obtain male issue (<hi rend="i">kapi tane</hi>) whereby the line of descent (<hi rend="i">katiri tangata</hi>) might be continued.</p>
          <pb xml:id="n32" n="24"/>
          <p>With a few exceptions, as in Tongareva, close marriages were favored by Polynesians. In the highly sophisticated cultures of Hawaii and Rarotonga brother and sister marriages took place for the purpose of perpetuating chiefly lines of high rank. The theory that peoples of lower cultures have an instinctive horror of incest is not substantiated in Polynesia; where prohibitions exist they may be attributed to a cultural development. Father-daughter marriages probably would not have occurred in Rakahanga, however, except that there was apparently no other solution to the difficulty. The eldest daughter, Kae, bore a female child, and the problem remained unsettled. The child was named Tupunoa, which carries the idea of “growth to no purpose” and bears witness to the disappointment that must have been engendered by the birth of a girl child. Any psychological inhibition against a father-daughter marriage having been once broken down, Toa married his second daughter in search of male issue. Again the child was a daughter. He married his third daughter, Naunau, and, according to some authorities, the issue was again a daughter. Other genealogists state that male issue was born. Toa then married his youngest daughter, Nanamu, in his determination to obtain male issue. Later, he even married his daughter-granddaughter, Tupunoa. The necessity of providing male issue to perpetuate the human species was the dominant consideration.</p>
          <p>The people of Manihiki and Rakahanga do not express any repugnance at the action of their ancestors, for they say that had such marriages not been made there would have been no <hi rend="i">kura tangata.</hi> By both Nanamu and Tupunoa, Toa had male issue. The problem of the perpetuation of the stock having been settled, father and daughter marriages ceased and were not repeated from Toa's period to the present day.</p>
          <p>Considerable contradiction exists regarding Toa's families by his daughters. Gill, who visited the atolls in 1852, recorded the following genealogy (10, vol. 2, p. 281):</p>
          <p>
            <figure xml:id="BucMani024a">
              <graphic url="BucMani024a.jpg" mimeType="image/jpeg" xml:id="BucMani024a-g"/>
            </figure>
          </p>
          <pb xml:id="n33" n="25"/>
          <p>Ikutau (Hukutahu) married his first cousin and had five children, whose sexes are not given. Attention has already been drawn to the incorrect spelling in this list. The Ikutau family given cannot be traced by the present generation.</p>
          <p>From a manuscript written in the native language by Tairi, one of the two first Rarotongan missionaries who went to Rakahanga in 1849, Gill records the following genealogy (13, pp. 143, 144):</p>
          <p>
            <figure xml:id="BucMani025a">
              <graphic url="BucMani025a.jpg" mimeType="image/jpeg" xml:id="BucMani025a-g"/>
            </figure>
          </p>
          <p>The parentage of Hoturangaranga and Tutonga is not given. They may have been among the retainers of Toa, if he had any retainers. After male issue had been produced by Toa, any restriction that may have existed against marrying with retainers' stock may have been relaxed.</p>
          <p>Kairenga gave his own pedigree from Matangaro. He stated that Matangaro married his aunt, Poe, but Gill's pedigree showing that Matangaro married Poe's daughter is more likely to be correct. Kairenga gave Matangaro's family as consisting of two daughters, Paevaka and Horoeka, and a son, Rua-ariki.</p>
          <p>Haumata-tua gave the pedigree in Table 2 before the Native Land Court.
<pb xml:id="n34" n="26"/>
<figure xml:id="BucMani026a"><graphic url="BucMani026a.jpg" mimeType="image/jpeg" xml:id="BucMani026a-g"/><head>Table 2. Issue of Toa</head></figure>
If this pedigree is compared with that given by Gill, the confusion is apparent. In the two tables, though the names of the mothers are reversed, the brothers Matangaro and Hukutahu are given as the sons of the youngest daughter.</p>
          <p>Of the second families, those who became most important were the male children of the youngest daughter. Of these, Te-pori-o-kaivai evidently died young and without issue. The outstanding ancestors are the second and third sons, Matangaro and Hukutahu. On the evidence adduced, Table 3 probably represents more nearly what took place with the main characters in the second group of four families.</p>
          <p>
            <figure xml:id="BucMani026b">
              <graphic url="BucMani026b.jpg" mimeType="image/jpeg" xml:id="BucMani026b-g"/>
              <head>Table 3. Second Families</head>
            </figure>
          </p>
          <p>In the second generation, except for the marriage of Toa, incest was avoided, as first cousins were available for marriages. It is likely that other first cousin marriages took place, but the records are as confused. It may
<pb xml:id="n35" n="27"/>
possibly be, as Kairenga maintained, that aunt and nephew marriages took place and that the confusion in the pedigrees is but a reflex of the confusion that actually existed at that period. Matters subsequently righted themselves and it is now immaterial which daughter of Toa produced which family. Certain it is that Matangaro and Hukutahu are the outstanding ancestors from whom descent is traced. Their brothers and cousins, whose alliances are not clear, nevertheless had families whose offspring formed the mass of the community which clustered around the more dominant, or chiefly, lines.</p>
        </div>
      </div>
      <pb xml:id="n36" n="28"/>
      <div xml:id="t1-body-d1-d3" type="section">
        <head>
          <hi rend="c">Social Organization</hi>
        </head>
        <div xml:id="t1-body-d1-d3-d1" type="section">
          <head>
            <hi rend="c">Family Pedigrees</hi>
          </head>
          <p>Family pedigrees record the elements and mechanism of social structure, lineal descent, marriages, sex, seniority, and collaterals in each generation. The sex and order of birth in each generation were always recited, but collaterals were left for their own descendants to trace. The skilled genealogist, however, besides knowing his own lineal descent, memorized the descent of the leading families in the community and could, by acquaintance with the collaterals of each generation, show how all the families were linked together by common ancestors of more recent date and how they all descended from the original biological family. He could demonstrate how all the inhabitants of the atoll were linked together by a common blood tie.</p>
          <p>Owing to the partial dislocation of native culture that followed the advent of Christianity in Manihiki and Rakahanga, however, many of the distant marriages in the pedigrees have been forgotten. One native, Kairenga, for example, was unable to give the marriages in his pedigree from the 4th to the 9th generation. He was doubtful about the 11th and 12th generations, but from the 13th generation onward he recited them as a matter of course. Some children who died young, or adults who had no issue, have been dropped out of the records in the course of time. It is the living descendants who perpetuate lineal descent.</p>
          <p>The value of even the last part of Kairenga's pedigree in establishing his social position is illustrated in the following summary of the sexes and order of births in each generation. Kairenga's own ancestors are represented by capital letters under “order of birth,” and Kairenga is represented by the capital in the 21st generation.</p>
          <p>
            <table rows="12" cols="4" rend="complex">
              <row>
                <cell role="label" rend="center">Generation</cell>
                <cell role="label" rend="center">Males</cell>
                <cell role="label" rend="center">Females</cell>
                <cell role="label" rend="center">Order of birth</cell>
              </row>
              <row>
                <cell rend="center">12</cell>
                <cell rend="center">2</cell>
                <cell rend="center">1</cell>
                <cell>M, f, m</cell>
              </row>
              <row>
                <cell rend="center">13</cell>
                <cell rend="center">1</cell>
                <cell rend="center">1</cell>
                <cell>M, f</cell>
              </row>
              <row>
                <cell rend="center">14</cell>
                <cell rend="center">1</cell>
                <cell rend="center">2</cell>
                <cell>M, f, f,</cell>
              </row>
              <row>
                <cell rend="center">15</cell>
                <cell rend="center">3</cell>
                <cell rend="center">3</cell>
                <cell>M, f, m, f, f, m</cell>
              </row>
              <row>
                <cell rend="center">16</cell>
                <cell rend="center">3</cell>
                <cell rend="center">1</cell>
                <cell>M, m, f, m</cell>
              </row>
              <row>
                <cell rend="center">17</cell>
                <cell rend="center">2</cell>
                <cell rend="center">2</cell>
                <cell>m, f, M, f</cell>
              </row>
              <row>
                <cell rend="center">18</cell>
                <cell rend="center">6</cell>
                <cell rend="center">1</cell>
                <cell>M, m, m, m, f, m, m</cell>
              </row>
              <row>
                <cell rend="center">19</cell>
                <cell rend="center">2</cell>
                <cell rend="center">1</cell>
                <cell>M, m, f</cell>
              </row>
              <row>
                <cell rend="center">20</cell>
                <cell rend="center">2</cell>
                <cell rend="center">1</cell>
                <cell>F, m, m</cell>
              </row>
              <row>
                <cell rend="center">21</cell>
                <cell rend="center">2</cell>
                <cell rend="center">1</cell>
                <cell>M, m, f</cell>
              </row>
              <row>
                <cell rend="center">10</cell>
                <cell rend="center">24</cell>
                <cell rend="center">14</cell>
                <cell/>
              </row>
            </table>
          </p>
          <p>Kairenga was senior to all descendants of the younger brothers and sisters of the five generations from the 12th to the 16th. In the 17th <choice><orig>genera-
<pb xml:id="n37" n="29"/>
tion</orig><reg>generation</reg></choice> his male ancestor was born third in a family of four, but he records that the elder brother and sister of that generation had no issue, so he retained his seniority in the 17th generation. In the 18th and 19th generations his ancestors were again first-born males. In the 20th generation the male line was broken, but Kairenga's mother was the first-born and senior in birth to her two brothers. If a patrilineal title had descended in the line succession would have gone to the oldest brother of Kairenga's mother and to his male issue. There was no title, however, and Kairenga's mother, from her position, had an important share of family land, which was inherited by her son. Kairenga himself was the senior member of a large family community. This would only have been possible under a bilateral system of tracing descent.</p>
          <p>The family, because the parts played by husband and wife in reproduction were fully recognized and the child regarded as flesh and blood of both parents, was bilateral and traced its descent through both parents. Other things being equal, more importance was attached to patrilineal descent. If, however, the mother came of a more important lineage than the father and more land and property were inherited from her, matrilineal descent assumed a correspondingly greater importance. A mother of high rank might form a break in a male line, but even so it was more important to be joined to an illustrious line by a female link than to be connected by a male link to an insignificant line. Patrilineal descent was, however, all-important in succession to rank and title. (See pp. <ref target="#n60">52</ref>, <ref target="#n62">54</ref>.) A female break in the <hi rend="i">ariki</hi> line was usually fatal to succession.</p>
        </div>
        <div xml:id="t1-body-d1-d3-d2" type="section">
          <head>
            <hi rend="c">Relationship Terms</hi>
          </head>
          <p>The enumeration of the names of individuals in pedigrees is sufficient record for the dead. For the living, who are in constant communication with each other, terms to indicate degrees of relationship are derived from the pedigrees.</p>
          <p>As about five generations are all that are normally contemporary, it is for five generations that relationship terms have been provided. The terms denote the relationship of any member of the blood community to a person speaking or a person spoken to, and they give a key to social structure. An adult male of middle age may, in his lineal descent, have both parents and grandparents alive who have lineal pedigrees one and two generations respectively shorter than his. If he has children and grandchildren, their lines of descent are one and two generations respectively longer than his. The names of ancestors are checked off on the fingers as relationship is being determined. The person speaking or spoken to is placed in the middle stratum, regarded as o. His parents are on the — 1 and his grandparents
<pb xml:id="n38" n="30"/>
on the − 2 stratum before him. His children are on the + 1 and his grand-children on the + 2 stratum after him. With him on the o stratum are his brothers and sisters. (See Table 4.) The −3 and + 3 strata are included in Table 4 to show terms used for extensions beyond five generations. Wissler (33, p. 160), in showing the same idea of genealogical strata in relationship, while treating the person speaking (self) as o, places the older brothers and sisters on the − 1 stratum above, and the younger brothers and sisters on the + 1 stratum below. This makes the parents − 2, the grandparents −3, the children + 2, and the grandchildren + 3. This numbering does not express the Polynesian system of counting, for brothers and sisters are collaterals and are not on the vertical line of descent. They must occupy the same horizontal on the o stratum but will fall into vertical collateral columns on the left or right of the speaker, according as they are older or younger than he. Both sex and seniority are given expression in relationship terms.</p>
          <p>
            <table rows="8" cols="6">
              <head>Table 4. Relationship Terms</head>
              <row>
                <cell role="label" rend="center">Generation stratum</cell>
                <cell role="label" rend="center">Senior collateral</cell>
                <cell role="label" rend="center">Lineal</cell>
                <cell role="label" rend="center">Junior collateral</cell>
                <cell role="label" rend="center">Marriage</cell>
                <cell role="label" rend="center">Adoption</cell>
              </row>
              <row>
                <cell rend="center">− 3</cell>
                <cell>Great-grand-uncles and aunts <hi rend="i">tupuna</hi></cell>
                <cell>Great-grandparents, <hi rend="i">tupuna</hi></cell>
                <cell>Great-granduncles and aunts, <hi rend="i">tupuna</hi></cell>
                <cell/>
                <cell/>
              </row>
              <row>
                <cell rend="center">− 2</cell>
                <cell>Granduncles and aunts <hi rend="i">tupuna</hi></cell>
                <cell>Grandparents, <hi rend="i">tupuna</hi></cell>
                <cell>Granduncles and aunts <hi rend="i">tupuna</hi></cell>
                <cell/>
                <cell/>
              </row>
              <row>
                <cell rend="center">− 1</cell>
                <cell>Uncles, <hi rend="i">metua-tane;</hi> aunts, <hi rend="i">metua-wahine</hi></cell>
                <cell>Father, <hi rend="i">metua-tane;</hi> mother, <hi rend="i">metua-wahine</hi></cell>
                <cell>Uncles, <hi rend="i">metua-tane;</hi> aunts, <hi rend="i">metua-wahine</hi></cell>
                <cell>Father-in-law and mother-in-law, <hi rend="i">metua-huangai</hi></cell>
                <cell>Adoptive parents, <hi rend="i">metua-whangai</hi></cell>
              </row>
              <row>
                <cell rend="center">0</cell>
                <cell>First and second cousins, <hi rend="i">tuakana</hi></cell>
                <cell>Self older, younger, <hi rend="i">tuakana teina</hi></cell>
                <cell>First and second cousins, <hi rend="i">teina</hi></cell>
                <cell/>
                <cell/>
              </row>
              <row>
                <cell rend="center">+ 1</cell>
                <cell>Nephews and nieces, <hi rend="i">iramutu</hi></cell>
                <cell>Son, <hi rend="i">tama;</hi> daughter, <hi rend="i">tamahine</hi></cell>
                <cell>Nephews and nieces, <hi rend="i">iramutu</hi></cell>
                <cell>Son-in-law and daughter-in-law, <hi rend="i">hunonga</hi></cell>
                <cell>Adopted children, <hi rend="i">tamaiti-whangai</hi></cell>
              </row>
              <row>
                <cell rend="center">+ 2</cell>
                <cell>Grandnephews and nieces, <hi rend="i">mokopuna</hi></cell>
                <cell>Grandchildren <hi rend="i">mokopuna</hi></cell>
                <cell>Grandnephews and nieces, <hi rend="i">mokopuna</hi></cell>
                <cell/>
                <cell/>
              </row>
              <row>
                <cell rend="center">+ 3</cell>
                <cell>Great-grand-nephews and nieces, <hi rend="i">mokopuna</hi></cell>
                <cell>Great-grand-children, <hi rend="i">mokopuna</hi></cell>
                <cell>Great-grand-nephews and nieces, <hi rend="i">mokopuna</hi></cell>
                <cell/>
                <cell/>
              </row>
            </table>
          </p>
          <p>When two individuals wish to decide their collateral relationship, they count generations from a common ancestor. The two counts are compared and one subtracted, and as the difference is usually between −2 and +2,
<pb xml:id="n39" n="31"/>
the collateral relationship terms can be determined. Most greater divergences that took place during the long period from remote ancestors have been closed by marriages in the pedigree line. It is from the nearest common ancestors that collateral relationship is decided; wider divergences from remote ancestors are matters of only academic interest. Thus the simple mathematical system in use could establish the relationship term between any two individuals so long as they could trace lineal descent from a common ancestor.</p>
          <p>The following correlative terms denote sex: <hi rend="i">tane</hi> (male or husband), <hi rend="i">wahine</hi> (adult female or wife); <hi rend="i">tuangane</hi> (brother of sister), <hi rend="i">tuahine</hi> (sister of brother); <hi rend="i">ure</hi> (son), and <hi rend="i">hika</hi> (daughter). <hi rend="i">Tane</hi> and <hi rend="i">wahine</hi> may be used as qualifying terms to distinguish sex in relationship terms of common gender.</p>
          <p>The position on the horizontal levels of the following terms, which distinguish lineal and collateral descent in generations and terms of relationship by marriage and adoption, depends on their vertical count from the nearest common ancestor:</p>
          <p><hi rend="sc">Tupuna</hi>. Includes the grandparents in lineal descent and their brothers and sisters. Also applies to all collaterals who count two generations less from a common ancestor or to all who fall in the −2 stratum. Sex is indicated by adding qualifying terms as in <hi rend="i">tupunatane</hi> (grandfather) and <hi rend="i">tupunawahine</hi> (grandmother). <hi rend="i">Tupuna</hi> also applies to all generations beyond that of grandparents and is used as a general term to denote ancestors. Great-grandparents, however, may be designated as <hi rend="i">tupuna-tuarua</hi> (second grandparents) and great-great-grandparents as <hi rend="i">tupuna-tuateru</hi> (third grandparents).</p>
          <p><hi rend="sc">Metua</hi> (a dialectical form of <hi rend="i">matua</hi>). Applies to parents, uncles, aunts, second cousins, and other collaterals on the −1 stratum. In some regions <hi rend="i">matua</hi> applies particularly to the male sex, and the female sex may be indicated by a different word, such as <hi rend="i">whaea</hi> or <hi rend="i">whaene</hi> in New Zealand. In Rakahanga, however, <hi rend="i">metua</hi> is common gender and sex is indicated by adding <hi rend="i">tane</hi> or <hi rend="i">wahine.</hi></p>
          <p><hi rend="sc">Metua</hi>-<hi rend="sc">Huangai</hi>. Applies to fathers-in-law and mothers-in-law. Here <hi rend="i">metua</hi> indicates the −1 stratum, for marriage puts the husband and wife on the same stratum as regards relationship with the parents of both contracting parties. The husband and wife still retain the relationship to each other indicated by their respective generation strata. The marriage relationship is indicated by <hi rend="i">huangai</hi>, which corresponds to the English “in-law”.</p>
          <p><hi rend="sc">Metua-Whangai</hi>. An adoptive parent; <hi rend="i">metua</hi> shows the generation stratum and <hi rend="i">whangai</hi> (to feed) refers to the outstanding feature in adoption. Polynesians in speaking English often refer to their “feeding father” or “feeding mother,” which conveys the meaning of <hi rend="i">metua-whangai</hi> better from a Polynesian angle than “adoptive parent”.</p>
          <p><hi rend="sc">Tama</hi>. Son.</p>
          <p><hi rend="sc">Tamahine</hi>. Daughter, <hi rend="i">tama</hi> with the female suffix <hi rend="i">hine.</hi> Also used to denote female first cousins once removed, female second cousins once removed, and other more distant collaterals who are on the + 1 stratum.</p>
          <pb xml:id="n40" n="32"/>
          <p><hi rend="sc">Iramutu</hi>. A special term to denote nephews and nieces, fairly widespread. Applies to the sons and daughters of brothers and sisters, whether they are older or younger than the speaker.</p>
          <p><hi rend="sc">Hunonga</hi>. Son-in-law or daughter-in-law, expresses both the generation level and the marriage relationship.</p>
          <p><hi rend="sc">Tamaiti-Whangai</hi>. Literally, “feeding child”. Applies to an adopted son or daughter.</p>
          <p><hi rend="sc">Tamaiti</hi>. Used in <hi rend="i">tamaiti-whangai</hi> to denote the generation level. Also denotes collaterals beyond the degree of consanguinity of nephews and nieces. May be used as the male correlative to <hi rend="i">tamahine</hi>, and thus includes first and second cousins once removed and other collaterals on the + 1 level.</p>
          <p><hi rend="sc">Mokopuna</hi>. Denotes grandchildren in lineal descent and extends laterally to include grandnephews, grandnieces, first and second cousins twice removed, and all collaterals on the + 2 level. The term also applies to great-grandchildren and relatives on the + 3 level. It may extend further to great-great-grandchildren and relatives on the + 4 level. The general term may be qualified by a number to denote the degree or removal as follows: + 2. Grandchildren−<hi rend="i">mokopuna.</hi> + 3. Great-grandchildren−<hi rend="i">mokopuna-tuarua</hi> (second). + 4. Great-great-grandchildren− <hi rend="i">mokopuna-tuateru</hi> (third).</p>
          <p>The method of deciding relationship terms which denote seniority within a family works horizontally instead of vertically. Though only two terms are used, they have a far-reaching application. Their first application is shown on the o stratum in Table 4. The seniority terms are:</p>
          <p><hi rend="sc">Tuakana</hi>. Elder brother of a male; elder sister of a female; a relative of the same sex, who is an equal number of generations removed from a common ancestor but who is descended from an older brother or sister in the common ancestor's biological family.</p>
          <p><hi rend="sc">Teina</hi>. Younger brother of a male; younger sister of a female; a relative of the same sex who is an equal number of generations removed from a common ancestor, but who is descended from a younger brother or sister of the common ancestor's own children.</p>
          <p>Seniority terms are correlative as regards age within the same sex. They are not applied in ordinary speech to brother and sister, though they may be referred to theoretically to denote priority in date of birth of individuals of opposite sexes. Seniority was all-important in the old social structure in deciding rank and title and relative degrees of influence in family and community gatherings. Certain terms were applied to the first-born son of a family, but such terms were restricted to one person. Dates of birth could not be recorded, but as children were born the terms used automatically placed them in their order of seniority. Thus when the eldest son died seniority within his generation naturally passed to the next brother, who was <hi rend="i">tuakana</hi> to all the others.</p>
          <p>The seniority inherited through order of birth was transmitted through the individual members of the family to their descendants. Seniority as
<pb xml:id="n41" n="33"/>
expressed by the <hi rend="i">tuakana</hi> and <hi rend="i">teina</hi> terms was horizontal in one biological family but was traced vertically in collaterals back to the family of the common ancestor. Thus in Table 4 the terms have been arranged in senior and junior collateral columns on either side of the lineal column. The senior collaterals are <hi rend="i">tuakana</hi> to those in both the other columns and the junior collaterals are <hi rend="i">teina</hi> to the other two. The descendants of the senior collaterals will always be on the left of the lineal column in the senior position, and similarly the descendants of the junior collaterals will always be on the right in the junior position. As they pass down through the generation levels they receive the relationship term of the particular level, but their senior or junior relationship will always be recognized. Order of birth in each biological family places individuals and their families in senior or junior positions with regard to other individuals and families.</p>
          <p>On the o level the truly collateral terms, <hi rend="i">tuakana</hi> and <hi rend="i">teina</hi>, are applied to brothers and sisters. The nearest of kin on this level are first cousins who are 2 generations removed from the common grandparents in both the lineal and collateral columns. Their relative position is merely an extension of the principle governing the individual members of one biological family. The children of the uncles and aunts who are older than the lineal father are classed as <hi rend="i">tuakana</hi>, and the children of the younger uncles and aunts are <hi rend="i">teina.</hi> The next nearest of kin are second cousins who are 3 generations removed from the common great-grandparents on both the lineal and collateral lines. They are divided into <hi rend="i">tuakana</hi> and <hi rend="i">teina</hi> according to whether their grandparents were senior or junior to the grandparents in the lineal line. In a similar way the classification extends to the third, fourth, fifth, and sixth cousins, and so on, as long as two lineal lines have the same number of generations. Theoretically, it could extend back to the first family of occupation, in which circumstance the seniority would be derived from the children of that family. What happens in practice, however, is that more recent marriages lead to a readjustment of relationship strata from a more convenient point of calculation. When members of a junior line marry into a senior line, it is natural for their descendants to trace descent through the senior line if they gain extra advantages by doing so. A person was often <hi rend="i">tuakana</hi> to a relative on one line of descent and <hi rend="i">teina</hi> to him on another. In some Polynesian areas, for example, New Zealand, a special term was used to designate such a double relationship, but no specific term seems to have been used in Rakahanga. Though the <hi rend="i">ariki</hi> title went by seniority, an <hi rend="i">ariki</hi> might have several <hi rend="i">tuakana</hi> in the community, owing to the marriages of his ancestors. His father might have married a younger sister, and thus all the children of his mother's older brothers and sisters would be <hi rend="i">tuakana</hi> to
<pb xml:id="n42" n="34"/>
him. As the title came from his father's side, this seniority would not affect him except in courtesy toward them. Similarly, the children of his father's elder sister would be <hi rend="i">tuakana</hi> to him, but again seniority through a female line would not affect him as regards his office. Thus the <hi rend="i">tuakana-teina</hi> relationship permeated the whole of society, and the respect and deference paid to seniority was indicated by its wide application not only to individuals on the same generation level but to families and groups. Its power, however, varied with the degree of consanguinity and its derivation through a male or female line.</p>
          <p>Two terms which did not fit into the collateral system based on seniority were used to denote brother and sister, <hi rend="i">tuangane</hi> (brother of a sister or the male relaitve of a female, both an equal number of generations removed from a common ancestor), and <hi rend="i">tuahine</hi> (sister of a brother or the female relative of a male, both an equal number of generations removed from a common ancestor). These have not been included in the table of relationship terms. Personally, I feel that there is an implied sex superiority in the term <hi rend="i">tuangane</hi> and an implied inferiority in <hi rend="i">tuahine.</hi> These terms fill in the gap left by the restrictions applied to <hi rend="i">tuakana</hi> and <hi rend="i">teina</hi>, which are fundamentally correlatives to denote superiority of birth. The restriction of <hi rend="i">tuakana</hi> and <hi rend="i">teina</hi> to members of the same sex is to prevent their use between opposite sexes. A logical reason for this usage would be to prevent some danger. The danger, as I see it, was to the male succession to rank and title through seniority of birth. No danger could arise from the use of terms among sisters to denote their relative positions in the female sphere of activity. If, however, a sister were termed <hi rend="i">tuakana</hi> to a younger brother, her seniority to him in the family would be admitted, regardless of sex. A first-born female would be <hi rend="i">tuakana</hi> to the rest of the family, and her claims would be hard to combat. It seems plausible, therefore, that those who guided the evolution of social structure provided against such a contingency.</p>
          <p>In most Polynesian dialects proper names as applied to the different sexes have been in use so long that masculine and feminine names have become differentiated by usage. Masculine names have, however, been given to females, and feminine names to males. In Rakahanga the naming of either sex by the same name is so prevalent that there are very few names which are not common to both sexes. Of the few which are not common, a male name is Atua, as in Te-atua-a-tupou, Te-atua-a-maheanga, and others. Though <hi rend="i">tane</hi> is used to distinguish the male sex in relationship terms, in proper names the male suffixes are <hi rend="i">matua</hi> and <hi rend="i">tangata.</hi> The term <hi rend="i">matua</hi> is a common term applying to the −1 generation, and when used as a relationship term it has to be qualified by <hi rend="i">tane</hi> or <hi rend="i">wahine</hi> to distinguish sex. In proper names, however, such as Temu-matua, it invariably indicates the male
<pb xml:id="n43" n="35"/>
sex. Similarly, <hi rend="i">tangata</hi> refers primarily to human beings as distinguished from other animals or the gods, but in proper names it is restricted to the male sex, as in Munokoa-tangata. In feminine names the general sex suffix is <hi rend="i">wahine</hi>, as in Temu-wahine, but the senior daughters of chiefly families have the special suffix <hi rend="i">tapairu</hi>, as in Haumata-tapairu and Takai-tapairu.</p>
          <p>The word <hi rend="i">ure</hi> (boy) and <hi rend="i">hika</hi> (girl), formerly used to indicate the sex of children in pedigree recitations or conversation, indicate the male and female organs of generation. There was no diffidence in using them to denote sex until prurient ideas of Western culture had affected the natives. The people are now ashamed to use these terms.</p>
          <p>It is reasonable to think that the terms <hi rend="i">tupuna, metua, tama, tamahitie</hi>, and <hi rend="i">mokopuna</hi> were originally used descriptively in lineal descent and that the collateral terms <hi rend="i">tuakana, teina, tuangane</hi>, and <hi rend="i">tuahine</hi> were confined to members of the same biological family. The collateral extension of the parent term <hi rend="i">metua</hi> to include uncles and aunts is natural, when the conditions under which the families developed are considered. The people lived in composite households or in households close together. The uncles and aunts extended to nephews and nieces the care and affection they gave to their own children. Uncle and nephew had mutual obligations toward each other, and the extension of the parent term toward the uncle stressed the closeness of relationship. As the child grew up he knew which of his <hi rend="i">matua</hi> was his actual father and he learned that other <hi rend="i">matua</hi> were brothers of his parents. The learning of degrees of consanguinity was part of the cultural education, knowledge that opened out gradually before the child with the learning of his pedigree. It evidently did not occur to the ancestors that there was a need for a descriptive term to distinguish between uncle and father to avoid confusion, for there was no confusion in their minds. Further extension of relationship terms to group all relatives on certain levels was a continuation of the principle of binding the people together through a common blood tie. The relationship terms were used to express blood obligations inherited through mutual ancestors. Social structure was based upon cooperation and support of institutions by as many of the blood-kinsmen as possible.</p>
          <p>The derivation of relationship terms from pedigrees which stratify relatives in generation levels, called by ethnological writers a “classificatory” system in contrast with the European “descriptive” system, is typical of Polynesia. It is true that the classificatory system lacks the exactness of the descriptive system. The classificatory system contains a few terms which apply to many relatives, whereas the descriptive system contains more terms and brings out remoteness of relationships more carefully than does the classificatory system. The contention of Kroeber (17) that relationship
<pb xml:id="n44" n="36"/>
terms are influenced by psychology is upheld in explanations of the usefulness to the Polynesians of their system. It was to the desire to maintain touch between collaterals no matter how far distant that the Polynesian relationship terms gave expression, and there was no need for exactness inherent in the terms themselves.</p>
        </div>
        <div xml:id="t1-body-d1-d3-d3" type="section">
          <head>
            <hi rend="c">The Biological Family</hi>
          </head>
          <div xml:id="t1-body-d1-d3-d3-d1" type="section">
            <head>
              <hi rend="sc">The Household</hi>
            </head>
            <p>It is probably largely because of the nature of the Polynesian household that foreigners have been led to regard the Polynesian as overlooking the importance of the biological family as a unit of social structure. Western culture is individualistic, and from the Polynesian point of view, selfish. The European tendency is not to share dwellings, food, property, or children. Relatives are not encouraged to stay indefinitely with the European family, a procedure almost incomprehensible to the Polynesian. Europeans, except in families holding ancient titles or inherited social position, do not keep family pedigrees. Relationship terms are restricted to a few degrees, and collateral terms have no vital meaning. The idea of kinship obligation and coöperation has weakened. The unit of money has displaced community coöperation in supplying man power for tasks beyond the scope of the individual. Villages and towns are accumulations of individual family units, independent of each other and having no common blood tie to unite the whole community. When members of one family do establish their households in the same village or town, they become separate independent units. Although the separate households recognize close relationship and may coöperate more closely with one another than with unrelated households, they develop different occupations and interests, and one household cannot dictate the policy of another by right of kinship. In the course of time, when relationship becomes remote, the diluted blood tie assumes a merely historical interest. Migrations are frequent, and separation further weakens any influence that blood kinship might have exerted. Thus collateral relationship ceases to exercise an active influence in social structure.</p>
            <p>The Polynesian household offers a marked contrast. Although the nucleus of the household is the <hi rend="i">puna</hi>, or single biological family, it is augmented in various ways. A son may elect to live with his parents after reaching adulthood. He marries and brings up his children in the same household, which thus comprises three generations. As the couple of the first generation grow old they relinquish the active management of the household to the couple of the second generation. The younger couple does the harder manual work in providing food and other necessities. The old people
<pb xml:id="n45" n="37"/>
do the light tasks and take care of the grandchildren. Parental authority becomes divided, and the grandparents exercise an active share in the upbringing of the children of the third generation. Two brothers of the second generation or a sister with her husband may also remain with the parents. If the family is wealthy in land, some poor collateral may live in the household and assist in cultivating the land and procuring food supplies. Some of the children of the third generation may be adopted by other relatives and thus pass away from their own family. On the other hand, the family may adopt children of other relatives and bring them up in their own household. The household, through the recognition of the blood tie and the prevalence of adoption, is thus rarely confined to one biological family. The children come under the influence and instruction of grandparents, aunts, and uncles, so that the clean-cut parental control characteristic of Western culture is not sharply defined, though it is recognized. The composite household, comprising members of three and even four generations, is in a fluid state, affected by the ebb and flow of blood kin. Seniority of birth indicates the family head. The source of the complex household is the first biological family in occupation. The bilateral biological family is the unit on which family pedigrees are built up, and the position and relationship of each individual member is clearly defined by the family pedigree.</p>
            <p>The household residence consists of separate buildings devoted to cooking and sleeping. The cooking house is shared by all. The food supplies are pooled, cooked in one earthen oven, and shared among the members of the household. The sleeping house accommodates all.</p>
          </div>
          <div xml:id="t1-body-d1-d3-d3-d2" type="section">
            <head>
              <hi rend="sc">Birth</hi>
            </head>
            <p>The great desire of the single biological family (<hi rend="i">puna</hi>) was to have male issue (<hi rend="i">kapi tane</hi>). The intensity of this desire led to incestuous marriages in the beginning of the history of the atolls. The general Polynesian attitude toward children was that females, through marriage, were likely to be lost to the local group or subtribe, whereas males, owing to the prevalence of patrilocal residence, strengthened their own community. In prominent families the desire for male issue was increased by the law of male succession to rank and title. A young married couple occupied by love and appreciation of each other's physical perfections probably dwelt less on the idea of male issue, though the desire must have existed. To the parents of the couple, however, the desire for male issue was dominant. The romantic side of the marriage did not affect them. The chiefly parents of the husband looked forward to a grandson to carry on the line, to succeed to the title, and to inherit the estate. A granddaughter would merely be a wife for some other family. The feeling was naturally shared by the husband's family and
<pb xml:id="n46" n="38"/>
tribe. The desire for male issue, curious though it may seem, was also shared by the parents of the wife and her tribe. They were actuated by the wish to see one of their blood occupying a high position, even though it be in another tribe. Although the sex of the wife may have been regarded as a mistake at the time of birth by her people, once she was married to a high chief she assumed importance in their eyes. She was the potential mother of a chief who would be of their blood. Considerable friction occurred from time to time between the families of two wives married to one chief, owing to the interest taken in the first son born of each marriage, and led to the establishment of a land distributor (<hi rend="i">tuha whenua</hi>) in the 6th generation and of the dual <hi rend="i">ariki</hi> in the 11th generation.</p>
            <p>When the news reached the parents of the newly married couple that the wife was pregnant, general satisfaction was expressed. Both mothers-in-law commenced to plait garments of split hala leaves (<hi rend="i">papa</hi>) for both the prospective mother and child. As the time of confinement arrived, the parents-in-law prepared to give a feast in celebration. If the child born was a male there was great rejoicing. The news was called (<hi rend="i">ka karangatia</hi>) throughout the village. A feast was the material expression of the general rejoicing. If the child was a girl, however, the disappointing news filtered out, but there was no public announcement. The food prepared was eaten, but there was no feast. It was stated by my informant that after the birth of a daughter the husband sometimes left his wife in disgust.</p>
            <p>The accouchement was carried out in the squatting position. The patient sat on the floor reclining against a relative of experience. When labor pains came on the patient was pushed up into the squatting position, and she supported herself also by holding to a rope hanging down from a rafter. The support assisted her in bearing down. Older female relatives assisted her by massaging the back and lower limbs between pains. The child was received on a plaited mat held by an assistant.</p>
            <p>The placenta was taken away and placed in a hole dug for it at some appropriate place. A little earth was filled in, and a coconut was planted above the placenta. Mr. Savage informs me that the coconut is always referred to as the <hi rend="i">weri</hi> of the afterbirth of the child. As the plant grew it was observed from time to time. Its manner of growth was supposed to indicate the nature of the growth of the child. If growth was vigorous the child would be healthy and strong, but if it was poor the child would be correspondingly weak and ailing.</p>
            <p>When the umbilical cord was tied and cut, the short end which remained attached to the child was termed <hi rend="i">pito</hi> and when it dropped off the umbilicus was also termed <hi rend="i">pito.</hi> The short length of cord that dropped off in natural course of time was differently disposed of for the two sexes. The male
<pb xml:id="n47" n="39"/>
<hi rend="i">pito</hi> was taken to the ocean side of the island and cast into the sea. This action, Mr. Savage informs me, was referred to as “titiri ki te moana roa” (casting into the long ocean). It was a form of sympathetic magic to insure that the child would develop into an expert fisherman or find success if he made voyages on the long ocean. The female <hi rend="i">pito</hi> was taken to the lagoon side of the island and cast into the lagoon waters. This was termed “titiri ki te tai roto” (casting into the inland waters). Man's sphere was in the ocean waters outside the bounding reef, but woman's sphere was in the waters of the lagoon. The treatment applied to the female <hi rend="i">pito</hi> insured that when she grew up the girl would become an expert diver in procuring <hi rend="i">pahua</hi> (<hi rend="i">Tridacna</hi>) and would be able to remain under water for a long time.</p>
            <p>In addition to the placental tree (<hi rend="i">weri</hi>), one or more coconut trees were planted by the parents or an elder of the family on family land to commemorate the birth of the child. Such trees were regarded as the special property of particular children, to whose exclusive use they were reserved. The custom and the trees, which were to provide a beverage for the children in years to come, were termed <hi rend="i">wai</hi> (water, coconut fluid).</p>
          </div>
          <div xml:id="t1-body-d1-d3-d3-d3" type="section">
            <head>
              <hi rend="sc">Childhood</hi>
            </head>
            <p>Children, as they grew up, were instructed in the observances of life as they came under the influence of the elder members in the composite household. Grandparents, who had more time than the parents, told them tales and myths and probably gave more instruction than the parents. The children played with tops and darts on land. Within the calm waters of the lagoon they early learned to swim and to handle canoes, and so laid the foundation to future success as fishermen and seamen.</p>
            <p>Mr. Savage writes me that girls of high rank were specially cared for from childhood. A female child was kept in a private house and attended by an elder woman of the family. During the period of childhood she was carefully fed, and her limbs were massaged with the coconut oil (<hi rend="i">romonga</hi>) prepared from mature nuts. She received careful attention in early childhood as regards the processes of excretion. After micturition or defaecation the parts were washed and gently patted during the washing. Such action is the natural care of children during babyhood, but its continuance with a selected attendant went with higher rank. The careful feeding and massaging was influenced not only by affection, but in order that the girl should become well-favored for some future marriage of note.</p>
            <p>The care lavished on a female child of rank was increased as she reached puberty. At about the age of the first menses she was kept guarded in a house by the mother and female attendants. This seclusion was for two purposes, to keep her from exposure to the hot sun to render her skin fair, and
<pb xml:id="n48" n="40"/>
to prevent her from obtaining any premature sex experiences. When she was allowed outside for physiological reasons she was accompanied by female attendants to guard her from any love affairs that the desires of her age might dispose her to entertain. Walks for the purpose of exercise were taken in the evening, to protect her skin and also to prevent others from seeing her too closely. Always, she was guarded by female attendants.</p>
            <p>A third reason for the seclusion and night walks was to keep the form and beauty of the girl a secret from other families until she should make her public début. During the period of seclusion the best foods obtainable were contributed by the family and subtribe to make her well-nourished, for plumpness was one of the standards of beauty acquired and required by the upper classes. The greater the impression created at the official début of a girl of rank, the greater the credit and satisfaction to her family and subtribe. That a girl of high rank was being kept in seclusion was known throughout the village. All in the community, particularly the families of rank who had marriageable sons, were interested. Thus the whole population looked forward to the time when the girl would make her public appearance.</p>
          </div>
          <div xml:id="t1-body-d1-d3-d3-d4" type="section">
            <head>
              <hi rend="sc">Puberty</hi>
            </head>
            <p>The public appearance of the girl took place at puberty. She had had her first menses and was considered fit for marriage. The ceremony of showing her to the people was termed <hi rend="i">whakahinga</hi>, according to Mr. Savage, and was associated with the father of the girl. Thus, if Mika was the father of the secluded girl, the approaching ceremony was promulgated through the village as “te whakahinga a Mika” (Mika's <hi rend="i">whakahinga</hi>). The event took place at a religious inclosure or marae at which the people gathered, after the announcement of time and place. The maid was dressed in a particular form of <hi rend="i">tipora</hi> garment, consisting of a sennit belt (<hi rend="i">tu kaha</hi>) supporting a small rectangular apron of plaited material (<hi rend="i">tautape</hi>) in front and a similar <hi rend="i">tautape</hi> behind. She was escorted to the gathering place by her family and on the marae was subjected to the scrutiny of the people, who appraised her physical beauty. It was stated that girls treated with seclusion and good feeding developed beautiful figures with skins of velvety softness. The brother of the girl's mother had the right to remove the maid's garment. He untied the <hi rend="i">tukaha</hi> belt and removed it, with the two <hi rend="i">tautape</hi> aprons. The girl thus stood entirely naked, but she folded her hands in front of her for concealment. The adults inspected her figure and shape. Viewed from the back, if her lower limbs were close together (<hi rend="i">piri</hi>) it was held that she had been virtuous. Her general demeanor was also observed. From the native point of view, no girl had any need to feel shame at the exposure of her naked form which custom allowed on such an occasion. If well formed,
<pb xml:id="n49" n="41"/>
she had every reason to be proud, for the ceremony was a public honor from which she would derive the appreciation and admiration of the people. She was getting publicity that would insure an advantageous marriage. Her only cause for fear was that if she had managed to elude her guards and had given away her virginity, the fact might in some way become known to the public The shrewd onlookers studied her with the knowledge that guilt is sometimes manifested in psychological behavior.</p>
            <p>During the ceremony on the marae the maid, from association with a religious structure, was tapu. During the period of seclusion she was also tapu, in the sense of being restricted from intercourse with outside people. After the inspection on the marae had been concluded, the maid was conducted through a complementary ceremony. Presumably her garments were restored to her, and she was taken to a place called Taipari. This excursion found expression in the words, “E ue haere ki to koutou whenua ko Taipari.” (Go to your land at Taipari.) There she was bathed, and a ceremony in which incantations and leaves were used was performed over her. No details were available, but I was informed that the ceremony made the girl <hi rend="i">noa</hi> (common), to enable her once more to mix with the public. I understood from my informants that the second ceremony was termed <hi rend="i">whakapu</hi>, but Mr. Savage seems to indicate that the word <hi rend="i">whakahinga</hi> used for the first part is an alternate term with <hi rend="i">whakapu.</hi> After the ceremony, though the girl moved about with fewer restrictions, she was still watched, as all the trouble taken was for the purpose not only of advertising the family but of making an advantageous marriage. A certain amount of restriction was enforced until the marriage was arranged and completed. The family and the social group had contributed food in order that she should be kept fair and virtuous for marriage. They had a share in her, and a worthy alliance was necessary to justify the interest and support they had given.</p>
            <p>The <hi rend="i">whakahinga</hi> and <hi rend="i">whakapu</hi> ceremonies were observed in the few families of high rank, particularly for the high-born <hi rend="i">tapairu</hi> of the <hi rend="i">ariki</hi> families. Among the general mass of the people such restrictions were not imposed. Both sexes had early love affairs with sex experience. Sex restrictions were not actively imposed as a rule until after marriage.</p>
            <p>I was informed that young men were also treated at puberty to seclusion and the subsequent public appearance on the marae. It was after or during the visit to Taipari that a father took his son to all the islands on which he owned property. He pointed out his shares of land and indicated the landmarks. The visit formed an instructional tour during which the boy committed the information to memory for future guidance. After this, the boy was qualified for marriage.</p>
          </div>
          <pb xml:id="n50" n="42"/>
          <div xml:id="t1-body-d1-d3-d3-d5" type="section">
            <head>
              <hi rend="sc">Marriage</hi>
            </head>
            <p>Among junior branches of families which had come to form the mass of the people, it is probable that matings took place without much ceremony. Young people who wished to prolong love affairs and to continue to live together probably did so without much ado. The cohabitation was soon noticed by their respective parents and perhaps ratified at a small feast by the two families. The couple might also obtain the consent of their parents beforehand, and the mutual consent of both families would ratify the marriage.</p>
            <p>For children of senior and chiefly families, marriage arrangements were much more formal. Males in the line of succession to the titles of <hi rend="i">ariki, tukuwhare</hi>, and to leading positions in subtribes, and girls of chiefly family who had been subjected to the <hi rend="i">whakapu</hi> ceremony, were public characters in whom the greatest interest was taken by the families and social groups to which they belonged. Any loose alliance conducted in a common way was not conducive to the maintenance of family and group prestige. Marriages of those of high rank were thus of public concern.</p>
            <p>Marriage alliances were discussed at meetings called for the purpose. When the parents or elders of the family came to a preliminary decision, they discussed the matter with the leaders of the families within the group or groups concerned. The decision was made at a public meeting, and carried the support of the group. Most decisions were quickly arrived at by those most intimately concerned, but the longer course of submitting matters to the group was necessary in order to give them a share in discussion and, by their public ratification, to obtain their hearty support in providing the food and property necessary to the marriage ceremony. Objections raised by parents to what they considered unsuitable marriages have, on occasion, led to romantic marriages in which the parental wrath was braved.</p>
            <p>The marriage having been consented to by all concerned, the members of the two family groups set to work to plait garments and mats for a marriage dowry. The dowry is termed <hi rend="i">takahinga</hi> (<hi rend="i">takahi</hi>, to stand on), as the mats form the material on which either party will stand in the house of the other. The women prepare the garments and mats, and the men collect and prepare food to accompany the <hi rend="i">takahinga.</hi> The bride's tribe (<hi rend="i">matakeinanga</hi>) escorts the girl in procession to the house of the groom's father, carrying dowry and food, to the accompaniment of singing and dancing. The <hi rend="i">takahinga</hi> and the food is formally presented to the bridegroom's family. The bride's people then retire to their own part of the village to await the reciprocal part of the ceremony.</p>
            <p>The bridegroom's tribe now escorts the bridegroom to the house of the bride's father. They, in turn, take a <hi rend="i">takahinga</hi> of garments and mats and
<pb xml:id="n51" n="43"/>
also food, which is presented to the bride's family with the appropriate speeches. In this manner both tribes share equally in the expenses, and each tribe contributes the food for the feasting of the other. The dowries are also distributed in the two groups, and though some may not receive as much as they gave, all get an ample share of the food to which they have contributed. The principle governing such ceremonies is one of public contribution followed by public redistribution. From the public speaking, singing, dancing, and feasting, the groups derive much pleasure. The exchange of gifts and accompanying speeches and reciprocal feasts give the alliance public recognition and constitute the main features of the marriage complex among the upper classes.</p>
            <p>Later, the husband is escorted back to his father's house, where the parents of both parties weep loudly and even cut themselves as a sign of grief that their direct authority over their children has ceased. This public demonstration of removal of parental control indicates that the married couple can now set up a household.</p>
          </div>
        </div>
        <div xml:id="t1-body-d1-d3-d4" type="section">
          <head>
            <hi rend="c">Titular System</hi>
          </head>
          <div xml:id="t1-body-d1-d3-d4-d1" type="section">
            <head>
              <hi rend="sc">History Of The Single And Dual Ariki Titles</hi>
            </head>
            <p>Primogeniture, or seniority in the male line, governs succession to rank and title in the Cook Islands and in the Society Islands. In the Cook Islands <hi rend="i">matahiapo</hi> is a general term for “eldest son,” but in certain family groups the <hi rend="i">matahiapo</hi> is a title held in the senior family. The <hi rend="i">ariki</hi> (chief) is simply the senior <hi rend="i">matahiapo</hi> of a number of family groups which have branched out from the original family. The more numerous the family groups, the greater the seniority and power of the <hi rend="i">ariki.</hi> The leaders of the expeditions which settled the islands must have had rank originally to enjoy the position of leadership. No matter, however, what their rank may have been in their original homeland, they have assumed the <hi rend="i">ariki</hi> title in the new lands on which they settled. The early use of the <hi rend="i">ariki</hi> title in Rakahanga shows that the institution was known to Toa and Tapairu and was introduced by them from their island of origin. In the later development of the dual <hi rend="i">ariki</hi> titles and the tribal titles, the names used indicate a local development and that the people were not guided in the formation of nomenclature by any memory of past tradition. The Rakahangans instituted new offices based on Polynesian principles but had to coin new local terms.</p>
            <p>The term <hi rend="i">ariki</hi> was regarded as a <hi rend="i">taohanga</hi> (title). Toa was a warrior, as his name implies, and as Gill (10, vol. 2, p. 281) records. Toa himself may have had no title, but his wife, from her name, Tapairu, was evidently of high rank. The establishment of the <hi rend="i">ariki</hi> title dates from Toa's second family. The title-holders are shown in capitals in Table 5.</p>
            <pb xml:id="n52" n="44"/>
            <p>
              <figure xml:id="BucMani044a">
                <graphic url="BucMani044a.jpg" mimeType="image/jpeg" xml:id="BucMani044a-g"/>
                <head>Table 5. Single Ariki Title Holders</head>
              </figure>
            </p>
            <p>Haumata-tua in Table 2 gives a different family of male children to Nanamu and derives Te-pori-o-kaivai and his brothers from Naunau, who is given as the youngest daughter. If correct, they were passed over for some reason and the title went to the sons of the youngest daughter. Of these, Te-pori-o-kaivai was the first-born, but he disappeared from the story and left no issue. Kairenga held that the title went first to the second son, Matangaro, and that his descendants were <hi rend="i">hui ariki</hi> (of the assembly of <hi rend="i">ariki</hi>) until the period of Kanohi, a female in the 10th generation. The title then passed to the Hukutahu line upon Kanohi's marriage to Tautape, a high chief of that line. The majority, however, state that the title went primarily to Huku-tahu, a younger brother of Matangaro. It is not clear why it should have passed to a junior. It was held that his grandmother, Tapairu, named him for the position. Tapairu's influence probably came from her supposed descent from Hiro and from the fact that she was a sister of the discoverer of the land. She was probably of higher rank than her husband, and the title may have been instituted from her side of the family. The phrase used to confirm Huku-tahu's appointment runs, “Te
<pb xml:id="n53" n="45"/>
kapi o te hui ariki kei a Huku-tahu, kei a ia hoki te pohatu.” (The male issue of the <hi rend="i">ariki</hi> was with Huku-tahu, as he also had the symbol of office.)</p>
            <p>The term <hi rend="i">pohatu</hi> (stone) was used to designate the symbol of office. Nothing is known of a stone symbol of office, but the term was at least used metaphorically. The <hi rend="i">ariki</hi> had priestly as well as chiefly duties.</p>
            <p>In the 4th generation the title was held by Rua. Kairenga held that this was Rua-ariki, the son of Matangaro. The majority held that the second <hi rend="i">ariki</hi> was Rua-a-makoha, the elder son of Huku-tahu. He went on an expedition to Hawaiki and never returned. Some of the witnesses at the Land Court held that Hawaiki was New Zealand, but others referred to Rua as having gone to Tokelau. Rua left word that he would send back some sign if he arrived safely at Hawaiki. Some time after Rua's departure, a shoal of fish named <hi rend="i">marau-awa</hi> appeared, and the people, accepting them as the promised sign, refer to them as the excrement of Rua (Tutae-o-Rua).</p>
            <p>Rua left no issue, so the title passed to the son of his younger brother, Kakahi. The <hi rend="i">ariki</hi> of the 5th generation had two names, Huku-tahu-rourou-a-whara and Tapu-mahanga. He had two wives. By the first wife he had a son named Kaitapu and by the second wife a son named Huku-potiki. The title passed to Kaitapu, who had the <hi rend="i">pohatu</hi>, but the powerful family of the second wife brought such influence to bear that Huku-potiki was given the office of attending to the distribution of land among the families. The office was termed <hi rend="i">tuha-whenua</hi> (land distributor), and Huku-potiki received special grants of land to go with his office, Paerangi in Rakahanga and Haroi in Manihiki. Huku-potiki was also entitled to the <hi rend="i">mata kairau</hi>, a contribution of food from the people in recognition of his rank. Thus the first division of authority is said to come from the two <hi rend="i">puna</hi> (families) of Huku-tahu-rourou-a-whara. This was the period in which the Hukutahu group was dividing into two subgroups which formed the bases of two tribes.</p>
            <p>From the period of Kaitapu in the 6th generation there follows a list of four <hi rend="i">ariki</hi>, Nikai-patu, Touiho, Te Renga, and Tautape. According to Gill (13, p. 144) the name of Ruare-tapu comes before Tautape. These names are placed in brackets in the lineal line because it is not certain whether or not they are father and son. As the column on the right gives Niho-tapu as Tautape's father, the <hi rend="i">ariki</hi> who immediately preceded Tautape, whether it was Te Renga or Ruare-tapu, could not have been his father. This raises the problem of how Tautape became <hi rend="i">ariki</hi> when his male line of descent is evidently derived from the junior line of Huku-potiki in the 6th generation. It may have been that Tautape's mother belonged to the senior line in which male issue failed after Ruare-tapu. It will be noted that in the middle column of Table 5, if the doubtful Ruare-tapu is left out, Tautape coincides with the 10th generation, which is the same as that of his two wives,
<pb xml:id="n54" n="46"/>
Heitutae and Kanohi, who are descended from Matangaro. His lineal descent, however, through Huku-potiki, is 4 generations longer. This makes a considerable difference in time if calculated in generations. However, it was stated that Niho-tapu, father of Tautape, and Poupou-whenua, father of the girls who married Tautape, were contemporaries. Thus the difference of 4 generations may be due to earlier marriages on the longer line or lapses of memory on the part of genealogists. For the purposes of following out the titles, Tautape will be regarded as living in the 10th generation.</p>
            <p>Tautape is also known as Rahui-ariki, and some genealogists confuse him with Temu. The middle and right columns in Table 5 show the descent from the two families of Huku-tahu-rourou-a-whara, among whom the power was divided. The senior line of Matangaro, in the left column, joins the Huku-tahu-rourou-a-whara line through the two daughters by different wives of Poupou-whenua. Both daughters married Tautape. As a result of this double marriage a further rearrangement was made in social structure. Tautape was the last of the <hi rend="i">ariki</hi> to rule singly over the people. If Ruare-tapu is counted, there were nine <hi rend="i">ariki</hi> who held the single <hi rend="i">pohatu</hi> from Haku-tahu to Tautape. The number of generations, if taken on the shorter side, indicates that the period of the single <hi rend="i">ariki</hi> occupied from 250 to 300 years, roughly.</p>
            <p>By the time of Tautape the population had increased (<hi rend="i">kua tupu te kura tangata</hi>) and aggregations of one blood (<hi rend="i">kura toto</hi>) had begun to develop into separate family groups. The individual households within the same group had, in turn, built their houses around the group nucleus within the common village. When a group moved out to secure more room for expansion, all the individual members linked together by a more recent blood tie (<hi rend="i">kura toto</hi>) moved and built their homes in proximity to each other. This expansion with the establishment of extra households is referred to in the phrase, “Kua tere te tangata me tona nani.” (People moved with their households.) Here <hi rend="i">nani</hi> is the equivalent of <hi rend="i">kainga</hi> (household) in other dialects. All the elements were present for a quarrel between the descendants of Matangaro and those of Huku-tahu, unless some arrangement was arrived at whereby the increased descendants of Matangaro could be pacified and given an active share in the government.</p>
            <p>The head of the Matangaro stock at this period appears to have been Poupou-whenua, who is shown in the ninth generation in the Matangaro line in Table 5. Poupou-whenua had two wives, by each of whom he had a family of three children. A daughter from each family was united in marriage to the <hi rend="i">ariki</hi> Tautape, and in this way the two divisions descended from Matangaro and Huku-tahu came together. What subsequently transpired is as follows:</p>
            <pb xml:id="n55" n="47"/>
            <p>
              <figure xml:id="BucMani047a">
                <graphic url="BucMani047a.jpg" mimeType="image/jpeg" xml:id="BucMani047a-g"/>
              </figure>
            </p>
            <p>The descent of the two wives of Poupou-whenua could not be traced, but it seems probable from what transpired that the mother of Heitutae came of Matangaro stock and the mother of Kanohi of Huku-tahu stock. The reason for this statement is that in quarrels over succession to a title it is usually the families of the respective mothers of two heirs who create the trouble which ensues. Heitutae had a first-born son, Temu-matua, and Kanohi had a first-born son, Tianewa-matua. Under normal circumstances the son of the first wife should have succeeded to the title. On the death of Tautape, however, complications came up as to the succession. Though both Temu-matua and Tianewa-matua were descended from Matangaro through the maternal grandfather, Poupou-whenua, a single title would have left the power with the Huku-tahu stock through their father, Tautape. It is evident that the Matangaro stock wanted direct representation through Temu-matua, which makes me think that his mother, Heitutae, was of Matangaro stock on her mother's side as well as on her father's. The other claimant to succession, Tianewa-matua, was associated with the Huku-tahu stock which had become divided into the two groupings known as Nu-matua and Tia-ngaro-tonga. If Kanohi, mother of Tianewa-matua, was of Huku-tahu descent on her mother's side, all the elements were present for the factional division of the Matangaro and Huku-tahu people, each faction demanding the succession of its respective close blood kinsman. It must have required strong expression of divided opinion to bring about the change in social organization which occurred. The opposing factions were pacified by a compromise, for the native historians state that in the period of Temu-matua and Tianewa-matua the authority (<hi rend="i">pohatu</hi>) was divided (<hi rend="i">I to raua tuatau i ngaha te pohatu</hi>). The compromise was the creation of a dual <hi rend="i">ariki</hi>ship. Temu-matua was made an <hi rend="i">ariki</hi> and was the first to hold the <hi rend="i">Whainga-aitu</hi> title. Tianewa-matua was made an <hi rend="i">ariki</hi> and was the first to hold the <hi rend="i">Whakaheo</hi> title.</p>
            <p>The people also divided into four tribes, and two tribes supported each title. The old tribes, Numatua and Tia-ngaro-tonga, upheld the <hi rend="i">Whakaheo</hi> title. Two newly-created tribes, Heahiro and Mokopuwai, banded together under the leadership of the <hi rend="i">Whainga-aitu.</hi> The native historians state, “Ko Heitutae e Poupou-whenua, ko te tumu ia o Mokopuwai.” (Heitutae and
<pb xml:id="n56" n="48"/>
Poupou-whenua, they were the foundation of the Mokopuwai tribe.) The younger brothers of Temu-matua became heads of subtribes in the Heahiro tribe. Thus the descendants of Matangaro supported their own direct representative in the person of the <hi rend="i">Whainga-aitu</hi>, and the descendants of Hukutahu supported the continuation of the old single title under the new <hi rend="i">Whakaheo.</hi> The lands in both Rakahanga and Manihiki became divided among the tribes. The dual <hi rend="i">ariki</hi> ruled over their respective tribes, with evidently no serious friction, as there are no traditional records of local wars. The <hi rend="i">ariki</hi> were supported by their <hi rend="i">hui rangatira</hi> (assembly of chiefs), which included the special officers dealing with land and food, and the heads of subtribes. There was some differentiation in the powers of the two <hi rend="i">ariki.</hi> (See p. <ref target="#n218">210</ref>.) The dual <hi rend="i">ariki</hi>ship existed down to the advent of Christianity, when the offices gradually fell into abeyance owing to changed conditions affecting the social structure of the people.</p>
            <p>Lists of the successive holders of the two titles were obtained. Of the <hi rend="i">Whakaheo</hi>, eleven title-holders are listed from the time of Tianewa-matua to the last holder, Iese, who was alive in 1898. Of the <hi rend="i">Whainga-aitu</hi>, fourteen held office from the inauguration of the title to Tupou-aporo, the last holder. It is not clear whether the earlier names in the list are direct successions of fathers and sons, but the pedigrees of later members of both titles will be referred to to throw light on the question of succession. It may be said, however, that the period of the dual titles extended over more than 200 years.</p>
          </div>
          <div xml:id="t1-body-d1-d3-d4-d2" type="section">
            <head>
              <hi rend="sc">The Whainga-Aitu Ariki</hi>
            </head>
            <p>The <hi rend="i">Whainga-aitu</hi> title was a full <hi rend="i">ariki</hi> title, but the special term must have been developed locally, for it does not appear, so far as I know, in any other part of Polynesia. The term is said to be derived from the words <hi rend="i">whai</hi> (to follow), <hi rend="i">nga</hi> (the), and <hi rend="i">aitu</hi> (gods). The title thus stresses the priestly functions of the office, for the <hi rend="i">Whainga-aitu</hi> was the principal medium between his people and the gods. The word <hi rend="i">whai</hi> has been translated “to follow,” but it seems to carry the additional meaning conveyed by the Tongarevan word “hai” which means to recite in the religious inclosures ritual whereby the gods are placated and success is assured. Thus the phrase, “kua whai i to ratou atua,” does not mean so much “to follow,” as that “they had conducted the regular ritual to their god.” Of the powers of the <hi rend="i">Whainga-aitu</hi> it was stated, “Te Whainga-aitu nona te papa.” (The <hi rend="i">Whainga-aitu</hi>, his was the lower stratum.) The <hi rend="i">papa</hi> is here translated as the lower stratum because it includes the sea as well as the land. The <hi rend="i">Whainga-aitu</hi> had powers over things terrestrial, as opposed to things celestial. He was supposed to concern himself with the welfare of the land and the
<pb xml:id="n57" n="49"/>
food growing thereon, as well as with promoting the food-productivity of the sea. He ruled over the two tribes, Heahiro and Mokopuwai, but did not interfere in the sharing out of land, which was left to the tribal heads, or <hi rend="i">whakamaru.</hi> The <hi rend="i">Whainga-aitu</hi> was the spiritual head and could command the assistance of the supernormal powers that commanded the productivity of land and sea. The <hi rend="i">whakamaru</hi> were the executives who superintended economic details.</p>
            <p>The list of <hi rend="i">Whainga-aitu</hi> from Temu-matua is given in Table 6.</p>
            <p>
              <figure xml:id="BucMani049a">
                <graphic url="BucMani049a.jpg" mimeType="image/jpeg" xml:id="BucMani049a-g"/>
                <head>Table 6. The Whainga-aitu Title Holders</head>
              </figure>
            </p>
            <p>The list shows 15 title-holders, but Mr. Savage, in a table compiled for me, omits Munokoa, Utua, and Putuhonu and makes the next holder of the title Maraerau, a son of the first <hi rend="i">ariki</hi>, Temu-matua, by a wife named Teraro-puka. In evidence given before the Court, Haumata-tua gave Munokoa and Utua as sons of Temu-matua and title-holders. The relationship of Putuhonu was not given. From Putuhonu to Whaitika the names follow
<pb xml:id="n58" n="50"/>
as a list, but from Whaitika the pedigrees are given. The pedigrees illustrate the rules governing succession to the title, which eliminated female succession and considered male seniority.</p>
            <p>The 9th holder, Whaitika, had three children. The first-born was a female, so succession went to the senior male child, Whairoa-enemea. The 10th holder, Whairoa-enemea, had four children, but as the first two were females, the title went to the third child, Whaireka. Whaireka, the 11th holder, had two families, but as his younger brother, Whaipoto, appears on the list as the 12th title-holder, Whaireka must have died before his children were old enough to hold office. Whaipoto would thus have succeeded as a regent until the heir of the senior line was old enough to assume office. The title went to the family of Whaireki's first wife, of which the first two were females. The third child, a son named Tupou-ma-te-tika, succeeded as the 13th holder. He had three wives. The first wife had one daughter, and the second wife had five daughters. Neither had a son. The succession therefore passed to the family of the third wife, Makirau, who had a son named Tupou-aporo. Before Tupou-aporo was old enough to assume office, his father died. The privilege of maintaining the position was thereupon assumed by Whaireka's second family, and Ieremia acted as regent for ten years. Tupou-aporo and Ieremia really acted together, and at the end of ten years Tupou-aporo, having reached man's estate, took over full control of the office. Ieremia objected, but as he was not supported by the families concerned, he left the island.</p>
            <p>A story in connection with the first <hi rend="i">Whainga-aitu</hi>, Temu-matua, illustrates the connection between the temporal and religious sides of the office. Temu-matua was a weakly child, so his maternal uncle, Rikiriki, was sent for. Rikiriki was the male representative of the Matangaro stock, which again shows how the Matangaro division had concentrated attention on Temumatua as their particular representative in the <hi rend="i">ariki</hi> families of his father, Tautape. Rikiriki took the child before the god Hikahara. The child recovered health. Some time afterwards, Rikiriki built a voyaging canoe (<hi rend="i">kua tuki pahi</hi>) with the object of visiting foreign lands (<hi rend="i">heaheake</hi>). He selected his party (<hi rend="i">tere</hi>), which included his nephew, Temu-matua, and went through the appropriate ritual before the god Hikahara to insure success (<hi rend="i">kua hakairo ia ratou i mua i taua atua</hi>). On the date of departure from Tauhunu in Manihiki they sang a song (<hi rend="i">pehe</hi>) in the channel or lagoon at Awanui. The tribe then realized that Rikiriki was taking the boy away with him on his travels, so the people begged him to allow the boy ashore that they might press noses (<hi rend="i">hohongi</hi>) with him in farewell.</p>
            <p>This story has a significant bearing on the creation of the <hi rend="i">Whainga
<pb xml:id="n59" n="51"/>
aitu</hi> title. The Polynesian historians have a habit, at times, of telling a straight narrative of a historical incident instead of discussing the details of the origin of an institution. Rikiriki was the eldest brother of Hei-tutae and the eldest son of Poupou-whenua's first wife. (See p. <ref target="#n55">47</ref>.) He was therefore head of the strong Matangaro group, and he cured his nephew Temu-matua. That he had prepared to go on a voyage to other lands and take his nephew with him shows that the dual <hi rend="i">ariki</hi>ship had not been established at that time. It is probable that the ambition Rikiriki may have entertained for his nephew had not received sufficient whole-hearted support to result in tribal action. The intended voyage, therefore, may have been due to spleen. The singing of the song in the channel drew full attention to the voyaging canoe. When it was perceived that Temu-matua was on board, the two Matangaro tribes realized that they were about to lose him. They were galvanized into action then by the imminence of a disaster. They begged that the boy be allowed ashore that they might press noses with him in farewell. The historian states that, on getting him ashore, “<hi rend="i">Kua tohi te matakeinanga.</hi>” The <hi rend="i">matakeinanga</hi> is the large group or tribe, and the tribe evidently went through a ceremony termed <hi rend="i">tohi</hi> in order to detain Temu-matua. Rikiriki, seeing what was happening, called out, “Ka tohi kotou, ono reka iho. Kua whakairo au i taua tamaiti ki te atua.” (You are doing the <hi rend="i">tohi</hi>, he may remain. I have, however, already dedicated that boy to the god.) I did not get the full meaning of <hi rend="i">tohi</hi> in Rakahanga. In New Zealand <hi rend="i">tohi</hi> refers to ceremony performed over a new-born infant or over adults on certain occasions to make them successful. The <hi rend="i">tohi</hi> over Temu-matua was undoubtedly a ceremony proclaiming him high chief over the Matangaro <hi rend="i">matakeinanga</hi>, and so making it impossible for him to leave. Rikiriki had probably achieved what he desired, to establish Temu-matua, and he volunteered the further information that the child was already in close connection with the god and so fitted for his position.</p>
            <p>At this time the two Matangaro tribes, Heahiro and Mokopuwai, were affected by a serious sickness (<hi rend="i">uiha</hi>). No one had been able to relieve (<hi rend="i">tunoko</hi>) them. The tribes, as shown by the story, were living on Tauhunu. Temu-matua then devised a plan for alleviating their distress. He went to the island of Te Puka and sought out a coconut that grew singly on one flower stalk. This nut (<hi rend="i">tautahi</hi>) he took, with some <hi rend="i">puraka</hi>, to Tukou as an offering to the god Hikahara. The disease thereupon cleared up (<hi rend="i">moki</hi>). This incident established the custom of traveling from Tauhunu to Tukou via Te Puka. The marked success of Temu-matua proved that he had power with the god and confirmed his authority. He became <hi rend="i">kana</hi> or <hi rend="i">tutara</hi> to the gods, or in other words, he became priest as well as <hi rend="i">ariki.</hi></p>
            <pb xml:id="n60" n="52"/>
            <p>Table 7 gives the pedigree of the three families of Tupou-ma-te-tika in full to show the manner in which succession passes over female issue.</p>
            <p>
              <figure xml:id="BucMani052a">
                <graphic url="BucMani052a.jpg" mimeType="image/jpeg" xml:id="BucMani052a-g"/>
                <head>Table 7. Manner in which Succession Passes over Female Issue</head>
              </figure>
            </p>
          </div>
          <div xml:id="t1-body-d1-d3-d4-d3" type="section">
            <head>
              <hi rend="sc">The Whakaheo Ariki</hi>
            </head>
            <p>The term <hi rend="i">Whakaheo</hi> is said to be derived from <hi rend="i">waka</hi> (canoe) and <hi rend="i">heo</hi> (to surround). In the voyages made by the whole population between the two atolls, the <hi rend="i">ariki</hi> holding the title surrounded the canoes with his priestly and supernormal powers and thus insured safe transport. The <hi rend="i">ariki</hi> was therefore termed originally the “waka-heo.” In the course of time <hi rend="i">waka</hi> (canoe) became changed to <hi rend="i">whaka</hi> (the causative prefix), but the title of <hi rend="i">whakaheo</hi> has not changed materially in meaning. It now means “to cause to surround with priestly influence,” and its original application to the voyaging canoes is understood. Of the powers of the position it is said, “Te Whaka-heo, nona te tira.” (The Whaka-heo, his is the <hi rend="i">tira.</hi>) The <hi rend="i">tira</hi> is in contradistinction to the <hi rend="i">papa</hi> of the <hi rend="i">Whainga-aitu.</hi> Just as the <hi rend="i">papa</hi> refers to things terrestrial, so the <hi rend="i">tira</hi> refers to things celestial. The <hi rend="i">Whakaheo</hi> had power over the phenomena of nature. He could demonstrate his power by causing the lightning to flash, the thunder to sound, and the rain to fall. He thus controlled the winds and storms, and it was through this power that he was able to surround the voyaging canoes with his priestly protection and insure a safe passage between the atolls.</p>
            <p>The <hi rend="i">Whakaheo</hi> trace their descent through Tianewa-matua, the first of the dual <hi rend="i">ariki</hi> who represented the Huku-tahu division of the people. Under his <hi rend="i">ariki</hi>ship were the two tribes, Nu-matua and Tia-ngaro-tonga. Patrilineal descent and seniority decided succession to the title, and women could not succeed to it. The list of title-holders is given in Table 8.</p>
            <pb xml:id="n61" n="53"/>
            <p>
              <figure xml:id="BucMani053a">
                <graphic url="BucMani053a.jpg" mimeType="image/jpeg" xml:id="BucMani053a-g"/>
                <head>Table 8. The Whakaheo Title Holders</head>
              </figure>
            </p>
            <p>Table 8 is only a list, for the exact pedigrees from Tapaha-matua (2) to Takai-whakaheo (9) could not be obtained. It was assumed that sons followed fathers throughout, but this is not certain. A detailed pedigree from Takai-whakaheo which throws further light on the complications that sometimes arise with regard to succession was obtained. (See Table 9.) Christianity entered the atolls in 1849, and in this year Takai-whakeheo (9) was holding the title. lete was alive in 1898.</p>
            <p>
              <figure xml:id="BucMani053b">
                <graphic url="BucMani053b.jpg" mimeType="image/jpeg" xml:id="BucMani053b-g"/>
                <head>Table 9. Succession to Takai-whakaheo</head>
              </figure>
            </p>
            <p>Takai-whakaheo (9) had a family of eight, the first child a daughter. Three sons followed. The senior son, Whakaheo-tama, went to Malden Island. During his absence his father died, and the next son in seniority,
<pb xml:id="n62" n="54"/>
Tuteru-utua (10), who was on the spot, was raised to the title. When Whakaheo-tama (11) returned from Malden Island seniority asserted itself, and he was made <hi rend="i">ariki</hi> in place of his younger brother. Tuteru-utua, having had to relinquish office, left the country and went to Tonga. Later, he returned and settled down. Both Tuteru-utua and the younger brother, Tarau, died before Whakaheo-tama, the title-holder. Whakaheo-tama had no children of his own, but he adopted Takai-taupe, the eldest son of his younger brother, Tuteru-utua. Whakaheo-tama is stated to have wished to confer his title on Tairi-orometua, the native missionary who came from Raro-tonga to spread the gospel in 1849. Tairi rightly refused the honor. On the death of Whakaheo-tama, however, Tairi used the strong missionary influence that existed, and the title passed to Iete (12), the son of Whakaheo-tama's elder sister. Had either Tuteru-utua or his younger brother, Tarau, survived, the title would undoubtedly have gone to them. Under the circumstances the rightful heir according to the laws of succession was Takai-taupe, the son of Tuteru-utua, who had been adopted by Whakaheo-tama. Probably Iete was an elder of the church, but certain it is that missionary influence interfered, and succession took place through an elder sister's line. The title ended with lete. Had the title continued, it would have been interesting to learn whether it would have continued down the Iete line or returned to the Tuteru-utua line.</p>
          </div>
          <div xml:id="t1-body-d1-d3-d4-d4" type="section">
            <head>
              <hi rend="sc">Other Titles</hi>
            </head>
            <p>It was held that each <hi rend="i">ariki</hi> had his <hi rend="i">hui rangatira</hi> (assembly of chiefs). The term <hi rend="i">hui rangatira</hi> is a Rarotongan one which includes the heads of families who are closely related to the ruling <hi rend="i">ariki.</hi> It is probable that the term was adopted from Rarotonga but that the principle involved was in force. Thus, the younger brothers and paternal uncles of the <hi rend="i">ariki</hi> were of high rank and would be closely associated with the official head of the family who held the title. Associated with the <hi rend="i">ariki</hi> were certain chiefs who had to deal with the economic details within the tribes. We have seen that the special office of <hi rend="i">tuha whenua</hi> was given to Huku-potiki in the 6th generation. When the spread of population resulted in family groups or tribes, the heads of the tribes functioned as <hi rend="i">tuha whenua</hi>, but the special title, <hi rend="i">whakamaru</hi>, was evolved. The <hi rend="i">tuha-whenua</hi> title that was instituted in the 6th generation became merged in one of the <hi rend="i">whakamaru</hi> titles that were evolved about the 11th generation.</p>
            <p><hi rend="i">Whakamaru</hi> was the local <hi rend="i">taohanga</hi> (title) given to the heads of tribes. It corresponds to the Rarotongan title, <hi rend="i">mata(h)iapo</hi>, which was not known in the atolls until after the introduction of Christianity. <hi rend="i">Whakamaru</hi> (to give shelter or shade) is thus an expressive term, as the head of the tribe
<pb xml:id="n63" n="55"/>
ought to shelter his people. According to some informants, there were two <hi rend="i">whakamaru</hi> under each of the dual ariki, thus making one representative for each of the four tribes. Others seemed to think that the heads of subtribes were also <hi rend="i">whakamaru.</hi> Some of the subtribes must have been fairly small in number, and it is hardly likely that the term would have been applied to the heads of many subtribes. It was stated that on the death of a <hi rend="i">whakamaru</hi>, his relatives (<hi rend="i">huanga</hi>) elected (<hi rend="i">mono</hi>) his successor. Others held that the <hi rend="i">whakamaru</hi> was elected by the subtribes, as he had power over their lands. It is unfortunate that no tribe was able to give a list of its <hi rend="i">whakamaru.</hi> No check data are available for pedigrees to indicate on what principles the <hi rend="i">whakamaru</hi> was really appointed. It is to be presumed, however, that the <hi rend="i">whakamaru</hi> was appointed by succession in the male line from the leading family in each group. Relatives and heads of subtribes probably met to discuss the pedigree and ratify the election of the person who was entitled by birth to succeed.</p>
            <p>The duties of the <hi rend="i">whakamaru</hi> were to act as public custodian over tribal lands, to settle disputes, and to prevent outside interference from another tribe. He had to do with directing the planting of food crops and the protection of the coconut plantations and <hi rend="i">puraka</hi> swamps from theft. He had power also over the redistribution of tribal lands which had to be adjusted to the ebb and flow of population. His decision was final, and not even his <hi rend="i">ariki</hi> could interfere with him in matters that concerned the interior economy of the tribe. His status was as high or even higher than that of the <hi rend="i">ariki</hi> in local matters. When it came to questions which concerned the intervention of the tribal gods, however, the <hi rend="i">ariki</hi> was superior, owing to his special priestly functions. From a modern point of view, the <hi rend="i">whakamaru</hi> in his own tribal district was judge of the native land court, Crown ranger, and director of agriculture. He also had a priestly function. When the whole population crossed from one atoll to the other the <hi rend="i">whakamaru</hi> from the tribes went first to the marae to conduct the appropriate ritual to the god (<hi rend="i">ka whai i to ratou atua</hi>). It was also stated that certain lesser tribal gods were in the keeping of the <hi rend="i">whakamaru.</hi></p>
            <p>The <hi rend="i">moa</hi> was a speaker or messenger between the <hi rend="i">whakamaru</hi> and the ariki. He was a <hi rend="i">whakamaru</hi> appointed by <hi rend="i">whakamaru.</hi> In discussions among the <hi rend="i">whakamaru</hi> he assumed seniority, and his decision was final. The title-holder seemed to be associated particularly with the <hi rend="i">Whainga-aitu.</hi></p>
            <p>The <hi rend="i">tira</hi> was an honorific title applied to the <hi rend="i">Whakaheo</hi> on account of his supernormal powers in dealing with natural phenomena. He was the astronomer who studied the heavens with regard to star signs of the seasons and the advent of winds.</p>
            <pb xml:id="n64" n="56"/>
            <p>The <hi rend="i">papa</hi> was a special title held by one Tuteru-te-tahua. He dealt with economic matters with which the ariki, by virtue of his priestly position, was not allowed to concern himself. He thus controlled land and food supplies. He had power with the <hi rend="i">Whainga-aitu</hi> division through his father and power with the <hi rend="i">Whakaheo</hi> division through his mother. He could thus conduct the religious ritual at the Poutu marae at Tauhunu and at the Akaroa marae at Tukou. He probably was a super <hi rend="i">whakamaru</hi> who by birth and ability carried influence with the <hi rend="i">whakamaru</hi> of the four tribes.</p>
            <p>It has been pointed out that a woman could not succeed to an <hi rend="i">ariki</hi> title even when she was the first-born child of an <hi rend="i">ariki</hi> by his first wife. This rule held in the Cook Islands, but was broken after the advent of Christianity. In Manihiki and Rakahanga it was observed until the extinction of the titles. The first-born females of the <hi rend="i">ariki</hi> lines, however, had a high status. They received personal names which became established as set names for those born in that line. The eldest daughters of the <hi rend="i">Whainga-aitu</hi> were named Haumata, and the eldest daughters of the <hi rend="i">Whakaheo</hi> were named Takai. Toa's wife was named Tapairu-taki-hetu. The first part of the name, Tapairu, means, in the Cook Islands, a first-born female. This meaning of the word was evidently carried on in the memory of Tapairu's descendants, for the special title, <hi rend="i">whaka-tapairu</hi>, was given to the first-born daughters of the <hi rend="i">ariki.</hi> By receiving special titles of dignity, the first-born daughters were effectively eliminated from any possibility of succeeding to the male <hi rend="i">ariki</hi> title, “Kare te wahine i te taohanga ariki, ka noho ratou i te taohanga whaka-tapairu.” (Women could not occupy the <hi rend="i">ariki</hi> title, they remained with the title of <hi rend="i">whakatapairu.</hi>) The first-born daughter of the <hi rend="i">Whainga-aitu</hi> was thus named Haumata-whakatapairu and that of the <hi rend="i">Whakaheo</hi>, Takai-whaka-tapairu. The prefix <hi rend="i">whaka</hi> is causative and carries the idea that the bearers of the titles had been made <hi rend="i">tapairu</hi> or first-born.</p>
            <p>The <hi rend="i">whakatapairu</hi> had certain privileges and performed certain functions connected with the title. Thus, after the <hi rend="i">whakamaru</hi> chiefs had visited the marae on landing at one of the atolls, they returned to the canoes. The <hi rend="i">whakatapairu</hi> then visited the marae in turn to perform their duties, and only after that could the people leave the beach and explore the land. The details of what they did on the marae are not clear.</p>
            <p>The <hi rend="i">Takai-whakatapairu</hi> had special powers to calm the sea. If the sea was rough during a bad season when food was low, the <hi rend="i">whakatapairu</hi> went out to the reef and beat upon the waters of the lagoon and the reef channel with a coconut leaf. At the same time she recited a chant (<hi rend="i">pehe</hi>). Her action caused the seas to subside and enabled the men to go fishing. Her power (<hi rend="i">mana</hi>) was derived from her position as direct representative of the <hi rend="i">Whakaheo</hi> who possessed the <hi rend="i">tira</hi>, or power over natural phenomena.</p>
          </div>
        </div>
        <pb xml:id="n65" n="57"/>
        <div xml:id="t1-body-d1-d3-d5" type="section">
          <head>
            <hi rend="c">Tribal History</hi>
          </head>
          <div xml:id="t1-body-d1-d3-d5-d1" type="section">
            <head>
              <hi rend="sc">First Divisions</hi>
            </head>
            <p>Rakahanga affords the unique example of the development of an insular population from one biological family. Its isolation and sea-girt bounds presumably prevented the intrusion of foreign elements. The blood tie was apparently realized to such a degree that no internecine wars broke out. The people developed for about nine or ten generations in one village on one island in peace and harmony.</p>
            <p>The absence of descriptive words in the relationship terminology must not be regarded as an indication that degrees of remoteness in kinship were not recognized. When crowding of population took place, some households moved away to obtain more space. The households that were close of kin accompanied them to the new center instead of remaining in proximity with people remote in blood. Thus the very feeling of close consanguinity that led family groups to establish themselves close together, by removal from the vicinity of other families gave expression also to the recognition of remote degrees of blood relationship. Upon this principle family groups have been segregated, have developed into tribes, and have later automatically divided into subtribes.</p>
            <p>The budding off of the families (<hi rend="i">puna</hi>) took place on the one small island, Te Kainga. Kainga means “home,” or the place where people dwell. In the course of time the original household developed into a village, but the village was automatically divided into separate groupings that claimed common blood kinship from more recent ancestors. The island of Te Kainga became established as the island of occupation, whereas all the other islands were reserved for the production of food. The regular passages back and forth to Manihiki were later established, though food considerations and the establishment of two villages in Manihiki were the result of the still later development of tribes and a dual <hi rend="i">arikiship.</hi></p>
            <p>Tradition shows that the families in the first few generations arranged themselves into two groupings according to their descent from the two brothers, Matangaro and Hukutahu. This dual grouping was rendered the more conscious because the <hi rend="i">ariki</hi> title descended in the Hukutahu family. The Matangaro families built their habitations on the sea side of the island, and the Hukutahu families took the inland lagoon side. In each generation families of close consanguinity naturally built their dwellings in proximity to each other. Occupation of particular localities led to the establishment of the rights of ownership to the particular portion of land on which the dwellings were built, and the theoretical division into two groups according to pedigree was thus carried into practice in the grouping of habitations. The
<pb xml:id="n66" n="58"/>
factors which led to the crossing of the group boundaries were marriage and adoption. Wives went to live in their husbands' habitations in their group localities, and conversely, younger brothers sometimes went to live with their wives' relatives. A brother or a father might adopt a sister's or a daughter's child, who otherwise would have remained with the group of the child's blood father.</p>
            <p>The automatic grouping took place in each of the two divisions. In each generation, while new families were created with their respective degrees of closeness, the existing degrees of remoteness were rendered another step more remote. Thus, if the Hukutahu families had been examined in the 5th generation, it would have been found that the families at one end of the territory allotted to them were of closer kin to each other than they were to the families at the other end. Somewhere between them there was a potential line of cleavage which became intensified with each succeeding generation. In the 5th generation the <hi rend="i">ariki</hi> Hukutahu-rourou-a-whara married two wives. The tendency to split is exemplified by the division of power between the eldest sons of the two wives. Kaitapu, the son of the first wife, retained the <hi rend="i">ariki</hi> title. All the families connected by close kinship to his mother formed a closer group around Kaitapu. Huku-potiki, the son of the second wife, became the land distributor, and all the families closely related to his mother formed another group around Huku-potiki. This secondary grouping within the Hukutahu group was simple, because evidently both mothers were from within the Hukutahu group. The automatic grouping or arrangement within a recognized larger group is unconscious until growth of population or the ambition of families makes the secondary grouping so conscious that it leads to a definite cleavage of the main group. Thus the establishment of two titles or offices led to the commencement of conscious grouping within the Hukutahu group. By the time the 10th generation was reached, the cleavage had become so definite that the original Hukutahu group had split into two groups which received the specific names, Nu-matua and Tia-ngaro-tonga. The formation of two definite groups with distinctive names led to the abandonment of the Hukutahu classification, but it was known that these two organized groups had descended from Hukutahu. The two groups continued to dwell on the lagoon side of the village that had grown up, and, owing to the method of budding off, each group occupied its own definite area.</p>
            <p>A similar process had taken place in the Matangaro group, but the Matangaro people had no special titles or offices which might serve to accentuate a conscious cleavage in the earlier stages of growth. In the 11th generation, however, the tendency to segregate was brought to a head through the two families of the <hi rend="i">ariki</hi> Tautape. The native historians state that the
<pb xml:id="n67" n="59"/>
old group of Nu-matua and Tia-ngaro-tonga ranged to the support of one of the sons of Tautape named Tianewa-matua. This was a carrying on of the support that they had always rendered to the <hi rend="i">ariki</hi> line, which came from within themselves. The Matangaro group, however, became attached to Temu-matua, the son of Tautape's other wife. Spurred to action by the fear of losing Temu-matua through the proposed voyage with his maternal uncle, Rikiriki, they kept him by making him head over their group. As the historians say, the Matangaro people formed the new groups (<hi rend="i">mata-keinanga</hi>) of Heahiro and Mokopuwai to support Temu-matua. This created a dual <hi rend="i">arikiship</hi> with two groups supporting each <hi rend="i">ariki.</hi> From the statement regarding new tribes, it would appear that the two Hukutahu groups had been named some time before but that the circumstances surrounding the creation of the second <hi rend="i">ariki</hi> led to the definite naming of the two groups into which the Matangaro people divided.</p>
            <p>The scattered habitations had by this time fused into a large village on Te Kainga, in which the population lived close together (<hi rend="i">piri te tangata</hi>). The division between the four groups was maintained in the arrangement within the village. A boundary stone (<hi rend="i">tuakoi</hi>; Rarotongan, <hi rend="i">kena</hi>) had been erected in the middle of the village to mark the boundary between the two original groups. It still remains in its original position and consists of a wide coral slab standing about 4 feet above ground. The two groups on either side of the boundary had automatically erected their dwellings in group clusters. The native phrase “ka tere te tangata e tona nani” (the individual went with his household) indicates the process already described. In the more complicated arrangement of a closely settled village, the leaders of the groups set up boundaries between the four groups and later between the subgroups which subsequently developed.</p>
            <p>In the 11th generation, therefore, the population had organized into four named groups (<hi rend="i">matakeinanga</hi>) which for the purposes of this study may be regarded as tribes.</p>
          </div>
          <div xml:id="t1-body-d1-d3-d5-d2" type="section">
            <head>
              <hi rend="sc">The Tribe</hi>
            </head>
            <p>The term “tribe” is usually applied to a fairly large number of people who occupy a territory defined by boundaries, speak a common language or dialect, are governed by one head, and share a common culture. This definition would cover all the inhabitants of Rakahanga and Manihiki. The four <hi rend="i">matakeinanga</hi> speak a common dialect of the Polynesian language. They originally had one common <hi rend="i">ariki</hi>, but the creation of a dual <hi rend="i">arikiship</hi> broke this unity. The general culture is identical, but there are differences with regard to leadership and the worship of gods. Furthermore, the land in the two atolls was definitely divided among the four <hi rend="i">matakeinanga.</hi></p>
            <pb xml:id="n68" n="60"/>
            <p>The term <hi rend="i">matakeinanga</hi>, in Mangaia, means a group of kinsmen, but evidently it was not in use as a tribal term. In Rakahanga the basic meaning of the term was similar to that in Mangaia, but when the four groups of kinsmen were distinguished by individual names these named groups were definitely referred to as four <hi rend="i">matakeinanga</hi> (<hi rend="i">e ha matakeinanga</hi>).</p>
            <p>Each <hi rend="i">matakeinanga</hi> elected its head independently of the others. The special term <hi rend="i">whakamaru</hi> was coined to distinguish him. His powers have been defined. For practical purposes, the <hi rend="i">matakeinanga</hi> was a small tribe, independent of the others as regards local government, but uniting with another <hi rend="i">matakeinanga</hi> under the two priestly <hi rend="i">ariki</hi> for religious purposes. All four <hi rend="i">matakeinanga</hi> federated for voyages back and forth between Mani-hiki and Rakahanga. The <hi rend="i">matakeinanga</hi> might have been regarded as sub-tribes were it not that each of them split into named subdivisions to which the term subtribe is better applied.</p>
            <p>The naming of the tribes creates a problem. In Tongareva, owing to the spread of secondary centers of habitation, the groups which grew up were designated by the territorial name of the island they occupied. In Rakahanga, as the whole population lived on one island and not in territories, the territorial designation of groups was not applicable. Under such circumstances, a group designation conveying descent from a common ancestor might be expected. In New Zealand and other areas, tribes are commonly designated by the application of a plural prefix to the name of an eponymous ancestor. Latent in the four Rakahangan tribal names, Numatua, Tia-ngarotonga, Heahiro, and Mokopuwai, is a possibility of derivation from ancestral names, but the people themselves were unable to settle the difficulty by locating the eponymous ancestors, if such they were, on their family pedigrees. Nu-matua looks like a personal name but does not appear in the pedigrees. Ngarotonga appears in the 5th generation (p. <ref target="#n34">26</ref>) as a grandson of Matangaro, but the Tia-ngarotonga is a Hukutahu tribe. Haumata-tua stated that Heitutae and Poupou-whenua were the stock through which the Mokopu-ngarotonga descended, and that Poupou-whenua, the father of Heitutae, was of Matangaro stock. This will bring in the name Ngarotonga as an eponymous ancestor, but when I pointed out that the Tia-ngarotonga tribe was not of Matangaro descent, the reply was given that Mokopu-ngarotonga did not refer to the Tia-ngarotonga but to the Mokopuwai, which is of Matangaro stock. No explanation was offered as to the Heahiro tribe. It is probable that the tribal names are derived from ancestors who lived between the 5th and 9th generations, but the imperfect transmission of the pedigrees prevents illustration of this.</p>
          </div>
          <pb xml:id="n69" n="61"/>
          <div xml:id="t1-body-d1-d3-d5-d3" type="section">
            <head>
              <hi rend="sc">Subtribes</hi>
            </head>
            <p>After the 11th generation, when the four tribes were established in name, a further subdivision took place within each of the tribes. These smaller groups were termed <hi rend="i">tukuwhare</hi>, which is again a local term coined to meet the local development. In the term <hi rend="i">tukuwhare</hi>, whare means a house, and the idea is conveyed of the kinsmen of the one inatakeinanga being grouped together in separate houses. The two <hi rend="i">Whainga-aitu</hi> tribes were divided into seven and four subtribes respectively (Table 10).</p>
            <p>
              <table rows="12" cols="2">
                <head>Table 10. The Whainga-aitu Tribes and Subtribes</head>
                <row>
                  <cell>Matakeinanga</cell>
                  <cell>Tukuwhare</cell>
                </row>
                <row>
                  <cell>Heahiro</cell>
                  <cell>1. Te-whare-ariki</cell>
                </row>
                <row>
                  <cell/>
                  <cell>2. Kai-wai-pa-honu</cell>
                </row>
                <row>
                  <cell/>
                  <cell>3. Po-te-noa</cell>
                </row>
                <row>
                  <cell/>
                  <cell>4. Mau-kino</cell>
                </row>
                <row>
                  <cell/>
                  <cell>5. Whare-nui</cell>
                </row>
                <row>
                  <cell/>
                  <cell>6. Whati-akau</cell>
                </row>
                <row>
                  <cell/>
                  <cell>7. Te-ure-roto</cell>
                </row>
                <row>
                  <cell>Mokopuwai</cell>
                  <cell>1. Hihahuke</cell>
                </row>
                <row>
                  <cell/>
                  <cell>2. Nga-whare-ririki</cell>
                </row>
                <row>
                  <cell/>
                  <cell>3. Taupo</cell>
                </row>
                <row>
                  <cell/>
                  <cell>4. Tutonga</cell>
                </row>
              </table>
            </p>
            <p>The origin of five of the seven subtribes of the Heahiro were demonstrated by Haumatatua (Table 11).</p>
            <p>
              <figure xml:id="BucMani061a">
                <graphic url="BucMani061a.jpg" mimeType="image/jpeg" xml:id="BucMani061a-g"/>
                <head>Table 11. Origin of Subtribes</head>
              </figure>
            </p>
            <p>Temu-matua and Tianewa are the first two dual <hi rend="i">ariki.</hi> The Heahiro and Mokopuwai tribes clustered around the family of Heitutae. By her, Tautape had five sons. The eldest son, Temu-matua, was raised by the two tribes to the position of <hi rend="i">ariki</hi>, and he became the first <hi rend="i">Whainga-aitu.</hi> Haumata-tua stated that the other sons were made heads of <hi rend="i">tukuwhare.</hi> Thus it is evident that minor subdivisions had already been following a natural process, but the five sons were made heads of five subdivisions in the Heahiro tribe. Succession to leading rank in those subtribes would subsequently be traced through the brothers. The group associated with
<pb xml:id="n70" n="62"/>
Temu-matua was named Te-whare-ariki (the house of the <hi rend="i">ariki</hi>) because the <hi rend="i">ariki</hi> title would descend in it from Temu-matua. The group associated with the second son, Pa-honu, was named after him and became Kai-wai-pa-honu, in which <hi rend="i">kai-wai</hi>, for some reason not explained, was prefixed to the ancestral name. Similarly, the Po-te-noa took its name from the third son by the prefix <hi rend="i">Po</hi> before the personal name of Te-noa. The Mau-kino subtribe took the name of the fourth son without any prefix. The fifth subtribe took the fifth son, Te-patiti, as their head but adopted the term of Te-whare-nui (the big house) for a reason not explained. Unfortunately, no clear pedigrees were furnished from these original heads of subtribes except the one from Te-patiti, shown in Table 12.</p>
            <p>
              <figure xml:id="BucMani062a">
                <graphic url="BucMani062a.jpg" mimeType="image/jpeg" xml:id="BucMani062a-g"/>
                <head>Table 12. Te-whare-nui Subtribe</head>
              </figure>
            </p>
            <p>The line is short. Pukerua in the 17th generation is still alive. As Te Patiti in the 11th generation was a fifth brother, the line may be expected to be shorter than those from senior sources. The line shows three females in it, and I do not know the status of the last issue in the subtribe.</p>
            <p>The sixth subtribe of Whatiakau, according to Tupou-rahi, comes from an important ancestor named Whatiakau, from whom Tupou-rahi gave his descent (Table 13).</p>
            <pb xml:id="n71" n="63"/>
            <p>
              <figure xml:id="BucMani063a">
                <graphic url="BucMani063a.jpg" mimeType="image/jpeg" xml:id="BucMani063a-g"/>
                <head>Table 13. Whatiakau Subtribe</head>
              </figure>
            </p>
            <p>Tupou-rahi could not link up the beginning of his pedigree with the main lines of descent from Toa. However, in the 14th generation Pupuke-papake married Te-atua-a-Tupou, who traced back to Tautape, the last of the single <hi rend="i">ariki.</hi> This places Te-papa-i-wairaro in the 7th generation. According to Tupou-rahi, Te-papa-i-wairaro was a very important man. He had four sons to whom he distributed land and authority. To the eldest son he gave the authority over his lands (<hi rend="i">tuku te whenua kia Whatiakau</hi>), to the second son he delegated authority over his group of people (<hi rend="i">whaka-tere te matakeinanga kia Tangihoro</hi>), to the third son he gave the care of the family gods (<hi rend="i">te whare urunga kia Ura</hi>), and to the fourth son he gave the position of herald (<hi rend="i">te horohoro kia Ngaropuruhi</hi>). If this is correct, Te-papa-i-wairaro must have belonged to the Matangaro group, for the Hukutahu group had already divided the authority over the people and the land between Kaitapu and Huku-potiki in the 6th generation. That the Whatiakau subtribe belongs to one of the <hi rend="i">Whainga-aitu</hi> tribes is a further substantiation. It will be noted that at this period there were evidently group gods which are referred to as the “whare-urunga.” Evidently Ura
<pb xml:id="n72" n="64"/>
performed the duties subsequently delegated to the <hi rend="i">whakamaru</hi> who had charge of the tribal gods and kept them in a house. No details were obtained of the functions of the herald (<hi rend="i">horohoro</hi>) beyond that he was the official messenger between the chiefs and the people, calling people together or promulgating decisions arrived at concerning group policy and action. Tangihoro was stated to have made voyages to foreign lands and to have come back among the people. Such voyages may have been made to the neighboring atoll of Manihiki and thus paved the way for the planting of that atoll and the subsequent regular visits to alternate the food supplies. Tupou-rahi in the 17th generation is an old man, and two generations may be added to the line to bring it to about 1900. A line of 19 generations makes it coincide in length with the Wharenui line in Table 12. In Tupou-rahi's descent, his male line from Te-atua-a-Tupou (14th generation) comes from the Whare-ariki subtribe through Temu-matua (11th generation), but he claims Whatiakau descent through the female line of Pupuke-papake (14th generation). He attaches great importance to her, as she was the first-born of her family. The leading line in the subtribe should come down through one of the elder brothers of Te-atua-maheanga in the 12th generation. It will be noted that the name of the mother of Temu-matua (11th generation) is given as Kanohi, whereas in Table 6 it is given as Hei-tutae. This is another instance of the confusion that exists.</p>
            <p>The seventh subtribe, Te-ure-roto, completes the Heahiro tribe. I am unable to give the circumstances which led to the adoption of the name.</p>
            <p>Of the four Mokopuwai subtribes, Tupou-rahi gave a descent to himself which placed Tutonga in the 11th generation. Tutonga was thus an ancestor who lived at the period when the subtribes were formed, and his name was adopted for one of them. Of the others, Nga-whare-ririki means “the small houses” and was probably so named from the linking of a number of small families to form a subtribe. The remaining two, Hihahuke and Taupo, sound like proper names but do not occur in the pedigrees submitted to the Land Court.</p>
            <pb xml:id="n73" n="65"/>
            <p>The two <hi rend="i">Whakaheo</hi> tribes have seven subtribes (Table 14).</p>
            <p>
              <table rows="15" cols="2">
                <head>Table 14. The Whakaheo Tribes and Subtribes</head>
                <row>
                  <cell role="label" rend="center">
                    <hi rend="lsc">Matakeinanga</hi>
                  </cell>
                  <cell role="label" rend="center">
                    <hi rend="lsc">Tukuwhare</hi>
                  </cell>
                </row>
                <row>
                  <cell>Nu-matua</cell>
                  <cell>1. Te-pu-tauhunu</cell>
                </row>
                <row>
                  <cell/>
                  <cell>2. Purenga</cell>
                </row>
                <row>
                  <cell/>
                  <cell>3. Kaupapa</cell>
                </row>
                <row>
                  <cell/>
                  <cell>4. Hitiki</cell>
                </row>
                <row>
                  <cell/>
                  <cell>5. Popo-iti</cell>
                </row>
                <row>
                  <cell/>
                  <cell>6. Nga-hoe-e-wha</cell>
                </row>
                <row>
                  <cell/>
                  <cell>7. Whati-kaua</cell>
                </row>
                <row>
                  <cell>Tia-ngaro-tonga</cell>
                  <cell>1. Wai-a-Matua</cell>
                </row>
                <row>
                  <cell/>
                  <cell>2. Ngaro-Tapaha</cell>
                </row>
                <row>
                  <cell/>
                  <cell>3. Nga-whare-ririki</cell>
                </row>
                <row>
                  <cell/>
                  <cell>4. Tuteru-matua</cell>
                </row>
                <row>
                  <cell/>
                  <cell>5. Tianewa-matua</cell>
                </row>
                <row>
                  <cell/>
                  <cell>6. Tihauma</cell>
                </row>
                <row>
                  <cell/>
                  <cell>7. Hua-tane</cell>
                </row>
              </table>
            </p>
            <p>No details of definite pedigrees of these subtribes were obtained. It will be noted that some, such as Tuteru-matua, Tianewa-matua, and Tihauma, are definitely the names of ancestors. The Tia-ngaro-tonga tribe resembles the Mokopuwai tribe in having a subtribe called Nga-whare-ririki (the small houses). The Nu-matua tribe has a subtribe called Te-pu-tauhunu, in which <hi rend="i">tauhunu</hi> is a plant and <hi rend="i">pu</hi>, a bush. Another subtribe name, Nga-hoe-e-wha, means “the four paddles.” It is thus evident that some subtribes were named after ancestors and others from incidents or things that are not clear to the present generation.</p>
            <p>Our stay on Rakahanga was too short for us to gather all the fragments of information that might throw more light on the details of subtribal evolution. It was clear to the Native Land Court that the witnesses were hazy about linking up their pedigrees with leading lines that would connect with descent from Toa. Most started with ancestors from 6 to 8 generations back who formed independent units and were left floating in the air. Certain families supplied the leaders for the subtribes, but no connected list of tribal or subtribal leaders could be conjured up out of the mass of family pedigrees adduced before the Court. It is probable that when the Court goes into the question of land ownership on the various islands, subtribes will be associated with definite islands, and the present family pedigrees may then be more clearly arranged to show the structure of subtribes.</p>
          </div>
        </div>
        <div xml:id="t1-body-d1-d3-d6" type="section">
          <head>Population And Land</head>
          <p>The growing needs of the increasing population led to the planting of Manihiki with coconuts and <hi rend="i">puraka.</hi> When food supplies ran short on
<pb xml:id="n74" n="66"/>
Rakahanga, the whole population voyaged across to Manihiki to partake of the food which had been planted there. It is evident that the migrations to Manihiki took place regularly after the population had increased sufficiently to necessitate such a step. On Rakahanga, sentimental considerations probably induced the four tribes to remain in the one village of Te Kainga. All had an equal right to it, and the law making the other islands tapu for food had been enacted before the development of the four tribes. Manihiki, however, had no sentiment attached to it, so that when the four tribes went there they founded two separate villages on different islands. Each village was occupied by an <hi rend="i">ariki</hi> and his two tribes.</p>
          <p>The islands were distributed among the tribes and subtribes and they planted them with coconuts. Certain parts were also excavated for the growth of the <hi rend="i">puraka</hi> (species of taro) in the brackish subsoil water. When the mature tubers were dug up, the tops were replanted. The tribes continued in occupation until the <hi rend="i">puraka</hi> and coconut supplies were depleted. They then returned to the other atoll, on which the food supplies had matured during their absence.</p>
          <p>The distribution of land had come about by a gradual process of evolution with the growth of population. When Toa, the first settler, established his family at Te Kainga, the problem of land distribution did not obtrude itself. In the 6th generation, however, the need of defining spheres of influence had evidently become necessary, and the office of land distributor (<hi rend="i">tuha whenua</hi>) was given to Huku-potiki, whose elder half-brother, Kaitapu, retained the rank of <hi rend="i">ariki.</hi> Power over the land thus became early disassociated from the office of <hi rend="i">ariki.</hi> It must be assumed that successive families under their respective heads had planted in the near-by islands that they considered suitable. There was room for all and no cause for trouble, as there were no previous settlers. The creation of a specific land officer led to organized disposal of the available lands. The land distributor, taking into account the planting that had already been done, associated certain families with particular localities. Probably the heads of families threshed out their arguments in public meetings. The official land distributor had the final word, and what was decided upon by him became established. As the families developed into larger groups, which subsequently became tribes, the tribal boundaries simply included the lands of the families comprising the tribes. With the increasing complexity of social structure to meet the growth of population, the power over tribal lands passed from one individual into the hands of the tribal heads, or <hi rend="i">whakamaru.</hi></p>
          <p>The functions of the <hi rend="i">whakamaru</hi> with regard to land have already been defined (p. <ref target="#n63">55</ref>). He acted as public custodian over tribal lands, settled disputes within the tribe, and prevented outside interference from other tribes.
<pb xml:id="n75" n="67"/>
Thus the tribes had their tribal lands on both atolls, and the problems of partitions and adjustments were settled within the tribe.</p>
          <p>Within the tribe itself, secondary groupings that resulted in subtribes also naturally led to the defining of subtribal lands within the tribal area. These subtribal problems were settled by the heads or leading families within the subtribe acting with the <hi rend="i">whakamaru.</hi> Within the subtribe, every individual had a right to a share of land, but the individual shares were grouped together in family holdings. The location of family shares was guided in later years by the inheritance of the areas planted by ancestors. In the beginning, however, the families were probably well-spaced in the lands that were being planted for the first time. In that period there were evidently no exact boundaries defined by equivalents of survey pegs or boundary stones. The boundaries between neighboring families were, to a certain extent, created by the spreading and meeting of the coconut trees that had been planted by successive generations. The self-planted trees that grew up naturally from the nuts that fell in an outward direction from marginal trees were claimed by the owners of the trees from which the nuts had fallen. The boundaries between neighboring family properties are thus often irregular. The evolution of irregular family boundaries is exemplified by a case that came up before Judge Ayson during our stay on Rakahanga. Two families desired a survey to define the boundary between their respective properties. As defined by the coconut trees claimed by each party, the boundary was marked by curves and projections of which the government surveyor, who naturally desired straight lines between angles, despaired. Each party claimed coconut trees that had encroached beyond the natural line. Neither would give way. A compromise was eventually made on the advice of the Court by the exchange, or even sale of trees, in order to get a satisfactory survey boundary. The fairness of the survey line with regard to the land was admitted by both, but the coconut trees were the bone of contention.</p>
          <p>The influence of the ownership of coconut trees in establishing a right to land is shown by the case cited above. The rights to the immovable property combined in the coconut and the land on which it grows is further exemplified by the <hi rend="i">pakewa</hi> custom. As defined by Mr. Savage in correspondence, it is a custom under which a person with no rights to land is given</p>
          <p>… the right to the use of a certain tree or trees, that he may pick the fruit thereof and use same as refreshment in his passing over the land. This custom gave only the right to use the tree but no right whatever to the land.</p>
          <p>The value of the <hi rend="i">pakewa</hi> custom will be fully appreciated by those who have experienced the heat of a tropical sun while walking over atoll islands. Out of the dual feelings of kindness of heart combined, doubtless, with the
<pb xml:id="n76" n="68"/>
idea of protecting one's trees from indiscriminate picking, the originators of the <hi rend="i">pakewa</hi> custom allocated a certain tree or trees to the use of the outsider. The tree was vested in the individual by name and could be used by him and his family. The use was enjoyed by his descendants, who traced their rights to the original agreement. The recognition of the equity of the custom has survived, but it has caused complications, owing to the increasing commercialism engendered by Western culture. Where self-grown trees have sprung up around the <hi rend="i">pakewa</hi> tree, these have been claimed by persons entitled to use the nuts. The need of the nuts for drinking while passing over the land may be rare, yet the mature nuts are gathered for the modern commercial purpose of making copra for sale to the traders. Some tree-owners accuse the landowners of having, in their turn, gathered the mature nuts for making copra. The situation is extreme when the owners of the <hi rend="i">pakewa</hi> trees deny that the trees were given by the <hi rend="i">pakewa</hi> custom and claim the land because their ownership of the trees is admitted.</p>
          <p>An interesting modification of the custom occurs. Visitors are sometimes asked to plant a coconut so that the tree may serve as a memorial of their visit. An old lady persuaded me to plant a nut in front of her house at Rakahanga and referred to it as my tree. Should it grow and bear nuts, I should on a subsequent visit probably be given drinking nuts from it to quench my thirst while there, but my courtesy rights would end at that.</p>
          <p>Property rights in land and coconut trees were definitely recognized. The loose attribution of commercial rights over property and land to Polynesian communities cannot be upheld. When within the subtribal areas, the land was distributed (<hi rend="i">tuha</hi>) among families. Each individual had a right to a share in the land, but the individuals were grouped in families of which the heads were the administrators. The members of each family lived together, worked together, and shared the produce of the land allocated to them. The things made by the individual were private property that he could dispose of as he deemed fit. Land and coconut trees, however, were family properties that descended by inheritance and were transmitted by inheritance. An individual had no right to dispose of land to outsiders without the consent of the family. He or the family had no right to dissipate the inheritance of posterity.</p>
          <p>Although the boundaries between tribal lands became more or less fixed, the boundaries between family holdings were not fixed immutably for all time. The shares allocated to families within the larger family groups comprising subtribes were subject to change in each generation, in accordance with the ebb and flow of population. Just as the small family group interested in a family holding was constantly changing from generation to
<pb xml:id="n77" n="69"/>
generation, not only in personnel but in numbers, so the shares within the family area required redistribution from time to time. It was part of the duty of the <hi rend="i">whakamaru</hi> to see that these redistributions and readjustments were conducted with equity.</p>
          <p>Owing to the recognition of bilateral descent, children could inherit a mother's share in her paternal estate, and families came eventually to have holdings of land in more than one district. How far this system of multiple holdings extended can only be fully decided when the Land Court investigates the land titles of the atolls. It is likely that the islands nearer to the villages were planted first, when the population was small. As the population increased, other islands were planted successively, and it is extremely probable that the descendants of the first planters have had a share in the planting of these islands. If so, it is extremely likely that the different tribes may have shares in a number of islands instead of having their shares consolidated in separate and adjacent islands, as they would had the families moved off and settled upon the lands they cultivated. This would further increase the number of holdings to which a family is entitled. The subject is one for further investigation.</p>
          <p>In the Cook Islands the dominant <hi rend="i">ariki</hi>, when they established themselves, took care to reserve definite areas of land for themselves. Such lands went with the <hi rend="i">ariki</hi> title and were inherited by the successor to the title. I am not sure whether or not specific lands went with the <hi rend="i">Whainga-aitu</hi> and <hi rend="i">Whakaheo</hi> titles. It was definitely stated, however, that when the office of land distributor was conferred upon Huku-potiki, he was given the special grants of Paerangi in Rakahanga and Haroi in Manihiki to go with his office. It is probable, therefore, that special grants of land were made to the holders of the two <hi rend="i">ariki</hi> titles.</p>
        </div>
      </div>
      <pb xml:id="n78" n="70"/>
      <div xml:id="t1-body-d1-d4" type="section">
        <head>
          <hi rend="c">Material Culture</hi>
        </head>
        <div xml:id="t1-body-d1-d4-d1" type="section">
          <head>
            <hi rend="c">Houses and Furniture</hi>
          </head>
          <div xml:id="t1-body-d1-d4-d1-d1" type="section">
            <head>
              <hi rend="sc">Introduction</hi>
            </head>
            <p>The houses now in use in Manihiki and Rakahanga follow the architectural principle of the rectangular house used in the Cook Islands. These houses are characterized by long upright posts erected at the middle of each end to support directly the main ridgepole. It was held that this type was introduced, together with houses built of lime, by the Rarotongan missionaries, Aporo and Tairi. The original type of house, which was made in both atolls until comparatively recently, has been gradually displaced, but fortunately a single example survives in Rakahanga, characterized by the absence of end posts supporting the ridgepole and of wall posts supporting the wall plates.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d1-d2" type="section">
            <head>
              <hi rend="sc">Original House Type</hi>
            </head>
            <div xml:id="t1-body-d1-d4-d1-d2-d1" type="section">
              <head>
                <hi rend="lsc">Framework</hi>
              </head>
              <p>The framework of the original type of house is composed of paired opposite side posts supporting two longitudinal beams, which in turn support a number of crossbeams. The crossbeams support the wall plates on their ends. The principal rafters rest on the wall plates, and their crossed upper ends support the ridgepole without the assistance of king-posts. The architectural principle followed resembles that of the Samoan long house, except that the Samoans use king-posts to support the ridgepole and cross the principal rafters above the ridgepole instead of below it (28, p. 20). The ground foundation upon which the Rakahangan house is erected is termed <hi rend="i">tango.</hi></p>
              <p>The framework technique, described from the surviving original type house in Rakahanga, is as follows:</p>
              <p>Skeletal framework (<ref target="#BucMani071a">figs. 4</ref>-<ref target="#BucMani073a">6</ref>). The longitudinal beams, crossbeams, and wall plates are lashed together on the ground before the supporting posts.are erected (<ref target="#BucMani071a">fig. 4</ref>). During the lashing each beam is supported by two short temporary posts slightly hollowed on the upper end to keep the round beams from rolling off. The temporary posts serve to raise the beams sufficiently high to allow the lashing patterns to be developed on the under surface. The two longitudinal beams and the crossbeams are of hala (<hi rend="i">Pandanus</hi>) trunks in the round. Nine crossbeams are used; of these, two are lashed close to the ends of the longitudinal beams and the remaining seven are evenly spaced to fill up the intervening space. Each crossbeam is securely lashed to each longitudinal beam with sennit braid. The somewhat thin wall plates are made of split hala wood. The two wall plates are laid over the ends of the crossbeams and lashed to each crossbeam with sennit braid. Three-ply braid (<hi rend="i">kaha</hi>) of a thicker size than that used on other parts of the framework is used for the beam and wall plate lashings (<ref target="#BucMani072a">fig. 5</ref>). The turns of the braid are arranged to form decorative patterns on
<pb xml:id="n79" n="71"/>
the under surface of the longitudinal and cross beams. The pattern is formed by alternating curves; the common lozenge pattern of other areas is not used. The term for the curved turns is <hi rend="i">ua</hi>, and the phrase used for tightening up such turns is <hi rend="i">uahia kia whakaketa.</hi></p>
              <p>Four supporting posts (<hi rend="i">pou</hi>) (<ref target="#BucMani073a">fig. 6</ref>) made of coconut trunks are spaced on the house site in the same manner as the temporary supporting posts, so as to coincide with the lashed frame of beams and wall plates, and arranged in two side pairs, the distance between the posts of one side being 19 feet and between opposite sides, 12.5 feet. The posts when fixed in the ground are 6 feet high. The upper ends are notched to take the longitudinal beams, which are thus kept in position without any subsequent lashing. In extra long houses, two posts instead of one are used at either end of the longitudinal beam, and the longitudinal beam (<ref target="#BucMani073a">fig. 6</ref>) is formed of coconut trunk instead of hala wood. When the posts are in position the whole beam frame (<ref target="#BucMani071a">fig. 4</ref>) is lifted and the longitudinal beams are fitted into the notches of the posts.</p>
              <p>
                <figure xml:id="BucMani071a">
                  <graphic url="BucMani071a.jpg" mimeType="image/jpeg" xml:id="BucMani071a-g"/>
                  <head><hi rend="sc">Figure</hi> 4. House framework: <hi rend="i">a</hi>, view from above; <hi rend="i">b</hi>, side view. 1, temporary posts; 2, longitudinal beams (<hi rend="i">hapai</hi>), 24 feet 3 inches long, 7 inches in diameter, laid parallel 12.5 feet apart; 3, crossbeams (<hi rend="i">vae</hi>), 17 feet 4 inches long, 5 inches in diameter, ends project 2 feet 5 inches on each side beyond longitudinal beams; 4, wall plates (<hi rend="i">kaupapa</hi>), 25 feet 2 inches long, 4 inches wide, and 1 inch thick; 5, lashing of crossbeams to longitudinal beams; 6, lashing of wall plate to crossbeams.</head>
                </figure>
              </p>
              <p>Roof framework (<ref target="#BucMani073a">figs. 7</ref>, <ref target="#BucMani074a">8</ref>). A scaffolding (<hi rend="i">turanga</hi>, standing place) is erected to facilitate the erection of the roof framework. This scaffolding takes the form of a roof framework with two end uprights supporting a beam in the manner of a ridgepole. Poles resembling rafters are placed in a slanting position from the ground to the ridgepole on either side, and lashed. Horizontal poles are lashed like purlins to
<pb xml:id="n80" n="72"/>
the supporting rafters to form steps. The height of the end uprights and the slant of the side poles are adjusted so that the carpenters standing on the scaffolding can reach any part of the house roof from within. The method is widespread and has been described in detail for Samoa (28, p. 24).</p>
              <p>
                <figure xml:id="BucMani072a">
                  <graphic url="BucMani072a.jpg" mimeType="image/jpeg" xml:id="BucMani072a-g"/>
                  <head><hi rend="sc">Figure</hi> 5. House framework, lashing of crossbeams (<hi rend="i">vae</hi>) to longitudinal beams (<hi rend="i">hapai</hi>) with curved pattern (<hi rend="i">ua</hi>). <hi rend="i">a</hi>, view from above: sennit braid (3) tied around <hi rend="i">vae</hi> (1) with running noose (4) and makes curved turn on under side of <hi rend="i">hapai</hi>, middle of curve crossing middle transverse line of <hi rend="i">hapai</hi> which corresponds to middle longitudinal line of <hi rend="i">vae</hi>; curved turn held in position by left hand. <hi rend="i">b</hi>, view from below, curved first turn (5) shown in position on under surface of <hi rend="i">hapai</hi> (2): braid (3) brought up on near side of <hi rend="i">hapai</hi> and on same side of <hi rend="i">vae</hi> (1) on which it commenced turn. <hi rend="i">c</hi>, view from above: braid (3) brought up over <hi rend="i">vae</hi> and makes diagonal turn (6) over <hi rend="i">vae</hi> to far right corner. <hi rend="i">d</hi>, view from below: braid (3) descends in far right corner and makes second curved turn (7) on under surface of <hi rend="i">hapai</hi>, crossing middle part of curve of first turn (5) and so fixing it in position; braid completes its turn by passing upward at near right corner. <hi rend="i">e</hi>, view from above: braid passes from near right corner diagonally over <hi rend="i">vae</hi> to far left corner making diagonal turn (8) which crosses over previous diagonal turn (6). <hi rend="i">f</hi>, view from below: diagonal turn over <hi rend="i">vae</hi> has brought braid back into far left corner from which it made first curved turn (5) in <hi rend="i">b</hi>; braid makes similar curved turn (9) but keeps closely to outer side of first turn (5) and in so doing crosses and fixes previous turn (7); from this position braid will cross back of <hi rend="i">vae</hi> and, appearing on under surface at far right corner, will make crossing curved turn to outer side of previous turn (7) from same side. <hi rend="i">g</hi>, view from below: repetition of first two curved turns successively applied on outer side of previous turns from same side results in pattern (10), which is continued until lashing sufficiently firm and pattern developed to taste of craftsman; lashings averaged 8 from one side and 7 from other, making 15 in all. <hi rend="i">h</hi>, side view showing crossings of turns on one side and half of main design (10), braid (3) in position to make horizontal fixation turns to fix lashing. <hi rend="i">i</hi>, side view: braid makes horizontal circumferential turn (11) around lashing turns where they pass between two wooden elements of frame and so tightens lashing; few turns made; braid passed through one or two loops to form half-hitches which fix end of braid. <hi rend="i">j</hi>, side view: alternate lashing formed by running turns from either side straight around <hi rend="i">hapai</hi> instead of using curved turns, crossing above <hi rend="i">vae</hi> being exactly similar to those in previous lashing; two parallel bands (3, 4) on under surface of <hi rend="i">hapai</hi> thus formed; circumferential horizontal turns (5) made as in <hi rend="i">i</hi>; diagonal crossing turns above <hi rend="i">vae</hi> not arranged in any definite lozenge pattern in either lashing, as they cannot be seen from below.</head>
                </figure>
              </p>
              <pb xml:id="n81" n="73"/>
              <p>
                <figure xml:id="BucMani073a">
                  <graphic url="BucMani073a.jpg" mimeType="image/jpeg" xml:id="BucMani073a-g"/>
                  <head><hi rend="sc">Figure</hi> 6. House framework: upper ends of supporting post (<hi rend="i">pou</hi>) notched to receive longitudinal beams (<hi rend="i">hapai</hi>). <hi rend="i">a</hi>, side view; <hi rend="i">b</hi>, end view; <hi rend="i">c</hi>, side view with chock to steady beam; <hi rend="i">d</hi>, side view of long house, each post reinforced by another placed close to it. 1, notched post, 6 feet above ground, 11.6 inches in diameter; 2, longitudinal beam; 3, triangular projection of notch; 4, curve of notch, fits beam; 5, chock, segment of hala trunk inserted to steady beam when notch and beam do not fit; 6, extra post used for long houses.</head>
                </figure>
              </p>
              <p>The principal rafters (<hi rend="i">oka</hi>) are natural hala poles in the round. The two end pairs of rafters and six intermediate pairs are evenly spaced. The top ends of each pair are cut to fit against each other, and lashed. An end pair of rafters is 17 feet 6 inches long and 5.3 inches in diameter at the lower thicker ends. Toward the lower ends, the rafters are notched to fit against the wall plate (<hi rend="i">kaupapa</hi>) and lashed, having a projection below the wall plate of 4 feet 3 inches. (See <ref target="#BucMani073a">fig. 7</ref>.) The pairs of end rafters are lashed in position, and a sennit line is stretched between their apices. The intermediate pairs are spaced, and after their upper ends are lashed they are raised to touch the stretched line. The parts touching the wall plate are then marked and notched on the inner side to fit against the wall plate, to which they are lashed. The intermediate pairs are not cut to exactly the same length as the end pairs, but their extra length simply projects beyond the wall plate. They are, however, graduated, for the pairs next the end pairs are 4 inches longer, the next two pairs from either end are 5 inches longer, and the two middle pairs are 8 inches longer. These projections thus give the line of the lower ends a slight convex curve from end to end. The main ridgepole (<hi rend="i">tauhuhu</hi>) is laid in the forks formed by the crossed upper ends of the principal rafters, and lashed to them. It is made of split hala wood 3 inches wide by 1 inch thick. The purlins (<hi rend="i">tarawa</hi>) consist of five on either side. The material and dimensions are the same as the main ridgepole. The purlins are laid horizontally over the principal
<figure xml:id="BucMani073b"><graphic url="BucMani073b.jpg" mimeType="image/jpeg" xml:id="BucMani073b-g"/><head><hi rend="sc">Figure</hi> 7. Roof framework: principal rafters (<hi rend="i">oka</hi>), shaping of upper ends and lashings. <hi rend="i">a</hi>, top ends of pair of rafters (1, 2) with about half of section cut away to form flat surfaces to fit against each other. <hi rend="i">b</hi>, pair with flat surfaces of top ends laid against each other when crossed; braid lashing fixed to one rafter with running noose; alternate horizontal (3) and vertical (4) turns made around opposite angles to lash pair together. <hi rend="i">c</hi>, side view, lower part of rafter (1) where notched to fit against wall plate (5) <hi rend="i">d</hi>, outer side of rafter (1) resting against wall plate (5), lashing of diagonal turns (3, 4) cross on outer surface of rafter and pass transversely around wall plate; after first crossing turns, subsequent turns made to outer side above and below; first turns to form lozenge pattern shown.</head></figure>
<pb xml:id="n82" n="74"/>
rafters, spaced so that three are above the level of the wall plate and two below it, and lashed to each principal rafter with sennit braid. The <hi rend="i">ua</hi> lashing consists of two curved turns from either side, or four altogether. The purlins are composed of two pieces overlapped and lashed together with transverse turns. Diagonal struts (<hi rend="i">toko</hi>) made of split hala 4 inches wide and 1 inch thick are used to steady the rafters on each side of the roof. On one side the strut is lashed above to the main ridgepole on the inner side of the second pair of rafters from one end. Thence the strut extends diagonally downward to meet the wall plate, to which it is lashed between the second and third rafters from the other end. It thus crosses four intermediate rafters, to each of which it is lashed. The strut on the opposite side of the roof is also lashed above to the ridgepole on the inner side of the second rafter from one end, and its lower end is lashed to the fourth rafter from the other end just above the wall plate, which it does not meet. This strut thus crosses and is lashed to three rafters. The lashings to ridgepole and rafters are by means of oblique turns passed around both elements in one direction only, and finished off with circumferential turns passing around the previous lashing turns and between the two wooden elements. There are 18 rafters (<hi rend="i">whakakaho</hi> or <hi rend="i">tokotoko</hi>), made of split hala wood 1 inch wide by 1 inch thick, on either side. One rafter is outside of the end walls on either side at either end. The rafters are then spaced about 1 foot 7 inches apart. The upper ends are crossed above the main ridgepole and lashed together at the crossing. The braid, after each lashing, is brought down and takes a complete turn around the main ridgepole. From there it passes down to the uppermost purlin and is used to lash the thatch rafter to the purlin with three or four curved turns in the <hi rend="i">ua</hi> pattern. By this means the thatch rafters are securely anchored down to the framework. The thatch rafters are all lashed to each purlin with the same <hi rend="i">ua</hi> lashing in three or four curved turns. For the thatch-rafter lashing, two-ply twisted sennit cord (<hi rend="i">whauhoto</hi>) is used instead of three-ply braid. The upper ridgepole (<hi rend="i">tauhuhu iti</hi> or <hi rend="i">takiri kaho</hi>), a long slender rod of split hala, is laid in the forks formed by the crossed thatch rafters, and lashed to them. The braid, after each lashing, is brought around the main ridgepole to anchor the upper ridgepole securely in position. The upper ridgepole is also called <hi rend="i">tokotoko</hi>, as it is of the same size as the thatch rafters. The eaves rod (<hi rend="i">turuturu iti</hi>), of the same material and size as the thatch rafters, is lashed to the lower ends of the thatch rafters on
<figure xml:id="BucMani074a"><graphic url="BucMani074a.jpg" mimeType="image/jpeg" xml:id="BucMani074a-g"/><head><hi rend="sc">Figure</hi> 8. Roof framework of house: <hi rend="i">a</hi>, side view, left end near principal rafter, four thatch rafters and end of fourth purlin cut away to show underlying parts of frame; <hi rend="i">b</hi>, end view. 1, supporting posts (<hi rend="i">pou</hi>); 2, longitudinal beams (<hi rend="i">hapai</hi>); 3, crossbeams (<hi rend="i">vae</hi>); 4, wall plate (<hi rend="i">kaupapa</hi>); 5, principal rafters (<hi rend="i">oka</hi>); 6, main ridgepole (<hi rend="i">tauhuhu</hi>); 7, purlins (<hi rend="i">tarawa</hi>); 8, diagonal strut (<hi rend="i">toko</hi>) attached to main ridgepole above and to wall plate below, crossing four rafters to which it is also attached; 9, thatch rafters (<hi rend="i">whakakaho</hi>); 10, upper ridgepole (<hi rend="i">tauhuhu iti</hi> or <hi rend="i">takiri kaho</hi>); 11, eaves rod (<hi rend="i">turuturu iti</hi>).</head></figure>
<pb xml:id="n83" n="75"/>
their outer side. The attaching of the eaves rods completes the roof framework (<hi rend="i">tua</hi>). The wooden elements composing the <hi rend="i">tua</hi> are shown in <ref target="#BucMani074a">figure 8</ref>.</p>
              <p>End framework (<ref target="#BucMani075a">fig. 9</ref>). The framework which closes in the ends (<hi rend="i">tara</hi>) of the house is composed of horizontal purlins and vertical rods which supply the thatch rafters, to which the thatching is attached. Wooden chocks (<hi rend="i">pono</hi>) are used to push out the lower end of the framework. Four horizontal purlins are stretched across the space between the end pair of rafters with the ends resting on the ends of the side purlins, to which they are lashed. The lowest purlin in the actual house is level with the end crossbeam, but in <ref target="#BucMani075a">figure 9, <hi rend="i">a</hi></ref>, it is a little too high. A mesial, rather thick rod is attached above to the crossing of the end rafters and lashed to the outer side of the four purlins. Its lower end projects a short distance below the lowest purlin. Two other vertical thin rods of thatch rafter material and size are spaced on the left and four others on the right. These lie outside of the four purlins and are lashed to them at the crossings. Above, they meet the end rafters, to which they are lashed. Below, they project slightly beyond the upper level of the lowest side purlins. Two additional purlins are next attached to the vertical rods on either side of the mesial rod, and their outer ends to the end rafters. The lowest purlins are on the upper level of the lowest side purlins. The cutting short of the mesial vertical rod leaves a space for the end entrance. The purlins are pliable, and chocks (<hi rend="i">pono</hi>) are placed between the end crossbeam and the fourth purlin to push the framework out a little from the perpendicular (<ref target="#BucMani075a">fig. 9, <hi rend="i">b</hi></ref>). The chocks, pieces of hala wood 4 inches by 2 inches, are three in number. The longest chock (<ref target="#BucMani075a">fig. 9, <hi rend="i">b</hi></ref>, 13) is 11 inches long and is placed immediately behind the mesial rod. Two others, 9 inches long, are placed also between the end crossbeam and fourth end purlin but directly behind the vertical rods on either side of the mesial rod. The fourth purlin gives a little extra length, and any adjustment of the lashing at the ends can be dealt with after the chocks are placed in position. Thus the lower end of the framework is not only pushed out slightly from the vertical, but by means of the graduated chocks the fourth end purlin has a slight outward, convex curve.</p>
              <p>Dimensions of completed framework: total length in middle line, 26 feet 2 inches; width between eaves rods, 23 feet 11 inches; height from floor to main ridgepole, 16 feet 5 inches; height from crossbeams to main ridgepole, 10 feet; height of eaves rods from ground, 3 feet.</p>
              <p>
                <figure xml:id="BucMani075a">
                  <graphic url="BucMani075a.jpg" mimeType="image/jpeg" xml:id="BucMani075a-g"/>
                  <head><hi rend="sc">Figure</hi> 9. End framework of house: <hi rend="i">a</hi>, end view from outside; <hi rend="i">b</hi>, section showing mesial rod pushed out from side with inserted chock. 1–4, horizontal purlins; 5, 6, additional purlins; 7, mesial rod; 8, vertical rods; 9, end pair of rafters; 10, side purlins; 11, end crossbeam; 12, lowest side purlins; 13, chock (<hi rend="i">pono</hi>); two vertical rods, one on either side of mesial rod, pushed out from side with shorter chocks.</head>
                </figure>
              </p>
            </div>
            <pb xml:id="n84" n="76"/>
            <div xml:id="t1-body-d1-d4-d1-d2-d2" type="section">
              <head>
                <hi rend="lsc">Thatch</hi>
              </head>
              <p>The roof and ends were thatched with sheets of lauhala (<hi rend="i">rau whara; Pandanus</hi> leaves) in which the leaves were bent over strips of aërial hala rootlets. The rootlets (<hi rend="i">kawhara</hi>, not <hi rend="i">kai whara</hi>, or rootlets with lower ends still in the air) which had reached the ground were used. The prepared strips (<hi rend="i">kaho</hi>, pronounced “kaheo”) measured 4 finger spans (<hi rend="i">anga honu</hi>, finger span), or about 3 feet.</p>
              <p>
                <figure xml:id="BucMani076a">
                  <graphic url="BucMani076a.jpg" mimeType="image/jpeg" xml:id="BucMani076a-g"/>
                  <head><hi rend="sc">Figure</hi> 10. Roof sheet needle (<hi rend="i">tuiau</hi>) : <hi rend="i">a</hi>, upper view, needle pointed at both ends, left end of type needle (B. P. Bishop Museum, C. 2811) slightly blunter than other; <hi rend="i">b</hi>, side view, upper surface slightly concave longitudinally; <hi rend="i">c</hi>, cross-sections with upper surface to right, upper surface slightly flatter than other, side edges rounded off. Total length, 10 inches; middle width, 0.9 inch; middle thickness, 0.4 inch; section about halfway between middle and points, 0.8 inch wide and 0.3 inch thick.</head>
                </figure>
              </p>
              <p>A roof sheet needle (<hi rend="i">tuiau</hi>) made of <hi rend="i">ngangie</hi> wood was used to sew the leaves together over the <hi rend="i">kaho</hi> strip. (See <ref target="#BucMani076a">fig. 10</ref>; <ref target="#BucManiP001a">pl. 1</ref>.) The needle was exactly similar in shape to that of Tongareva. The butt ends of the lauhala were doubled over the <hi rend="i">kaho</hi> strip in the same way as in Tongareva and the Cook Islands (27, p. 15). When the leaves were pierced transversely with the needle, dry coconut leaf midribs (<hi rend="i">tunikau</hi>) or thin strips of aerial hala rootlets were passed through the holes made and kept the leaves together. The sheet was made the full width of the <hi rend="i">kaho</hi> strip (3 feet) and when completed was also termed a kaho. The full supply of sheets was made and the thatching (<hi rend="i">ato</hi>) commenced. The technique is as follows:</p>
              <p>Thatching (<hi rend="i">ato</hi>) commences from below and works upward. The first sheet of lauhala is laid across the thatch rafters just above the eaves rod at one end of the <hi rend="i">tua</hi> frame. The part doubled over the hala rod is toward the top, and the doubled-over butt ends of the leaves are outside. The end of a piece of sennit braid is tied to the thatch rafter with a running noose. The braid is carried over the edge of the sheet to the right of the rafter. A hole is punctured through the sheet below the contained rod and to the left of the rafter. The braid is passed back through the hole from the outside and makes either a half-hitch or an overhand knot with its standing part and so fixes the stiff upper edge of the sheet to the thatch rafter. Subsequent sheets are added a little higher up the rafter than the one preceding, and a similar fastening is made with the continuous braid. The method is exactly similar to that used in the
<pb xml:id="n85" n="77"/>
Cook Islands (27, pp. 20–23). For piercing the holes through the sheets, the fingers are generally used, but it was said that a hooked thatching needle was sometimes used. (See <ref target="#BucManiP001a">pl. 1</ref>.) No local name could be remembered for the implement, and as the people boasted that their method of piercing holes with the fingers was much quicker than the Rarotongan method with the needle, it is probable that the thatching needle was introduced from that area but never became popular. A Rakahangan thatching needle made of <hi rend="i">ngangie</hi> wood is shown in <ref target="#BucMani077a">figure 11</ref>. It resembles the roof thatch needle in its width, which exceeds that of similar implements from the Cook Islands (27, p. 19).</p>
              <p>
                <figure xml:id="BucMani077a">
                  <graphic url="BucMani077a.jpg" mimeType="image/jpeg" xml:id="BucMani077a-g"/>
                  <head><hi rend="sc">Figure</hi> 11. Thatching needle: <hi rend="i">a</hi>, upper surface view; <hi rend="i">b</hi>, side view. 1, notch point, used for puncturing thatch sheet; 2, notch, hooks up braid and draws it back through sheet; 3, upper surface, concave longitudinally and fairly flat transversely; 4, unnotched point, may be used in sewing sheet together; 5, lower surface, convex transversely, side edges rounded off. Length, 11 inches; mesial width, 1.2 inches; mesial thickness, 0.4 inch; notch point, 1.2 inches long; base of notch point, 0.7 inch wide and 0.4 inch thick; other point at same distance (1.2 inches) from end, only 0.2 inch thick; notch, 0.3 inch deep.</head>
                </figure>
              </p>
              <p>The thatcher works a width of one sheet until he reaches the ridgepole, when another strip of thatching is commenced with the adjacent edges of the new sheets slightly overlapping those already fixed. The completed upward width of one sheet is known as a <hi rend="i">marewa</hi>, and the length of the house is referred to in the number of <hi rend="i">marewa</hi> it takes to thatch one side of the roof. The thatcher works from the inside of the house, and as he works upward he ascends on the scaffolding within, which is allowed to remain until the thatching is completed. An assistant carries the roof sheets and places them in position from the outside. The thatcher within usually spaces the upper edge of the sheet from the one below by the number of fingerbreadths decided upon. The closer the sheets are together, the better the house, but close thatching takes more material in sheets and sennit and involves a corresponding increase in labor.</p>
              <p>The assistant uses a lifting pole (<hi rend="i">tukutuku</hi>) made of <hi rend="i">whano</hi> wood (<ref target="#BucMani077a">fig. 12</ref>) to place the sheets in position when the thatching rises beyond his reach. The point of the long pole is stuck into the outer surface of the sheet in the middle line just below the site of the transverse hala rod at the upper edge. The shoulder prevents the pole from going in too far. The assistant lifts up the sheet with the point and swings it into position above the last sheet tied. The thatcher within adjusts it accurately and then lashes it.</p>
              <p>
                <figure xml:id="BucMani077b">
                  <graphic url="BucMani077b.jpg" mimeType="image/jpeg" xml:id="BucMani077b-g"/>
                  <head><hi rend="sc">Figure</hi> 12. Point of lifting pole (<hi rend="i">tukutuku</hi>) for thatch: <hi rend="i">a</hi>, front view; <hi rend="i">b</hi>, side view. Thinner end of convenient-sized pole, about 14 feet long, is cut at one end for about 2 inches into point with sharp curve proximally to form distinct shoulder (i).</head>
                </figure>
              </p>
              <pb xml:id="n86" n="78"/>
              <p>The roof ridging is attended to after both sides of the roof have been thatched. The lauhala sheets have been extended as high as the ridgepole. The first stage in the roof ridging consists of applying a double sheet of lauhala on either side with the upper edges meeting in an apex above the upper ridgepole. The two sheets of either side are then sewed together with two-ply twisted sennit cord. The two double sheets thus sewed together (<hi rend="i">uihau</hi>) extend from end to end of the ridge. The second stage consists of reversing two plaited coconut leaf mats (<hi rend="i">pora</hi>) (see p. <ref target="#n119">111</ref>) with the shiny upper surface of the leaflets on the outside. These are doubled over the <hi rend="i">uihau</hi> and kept in position by passing pointed wooden pins about 2 feet long through the mats from side to side, taking care that the pins pass through between the upper and main ridgepoles. The upper ridgepole prevents the pins from working upward and thus anchors the mats in position. The reason for reversing the two <hi rend="i">pora</hi> mats is that there may be one unsplit midrib edge on either side of the ridgepole to hold against the pins. The double mats are laid on with a slight overlap until the ridgepole is covered from end to end.</p>
              <p>The ends of the house are also thatched with lauhala sheets, the sheets being tied to the vertical elements of the framework in the same way as in the roof thatching. Some adjustment is required in cutting some of the sheets to fit in with the changing width of the area covered. The ends of the lowest sheets on both the sides and the ends are cut off a little below the eaves rod in a straight line to form the eaves (<hi rend="i">turuturu</hi>). At the ends the eaves are 3 feet 2 inches above the ground, and on the sides, 3 feet.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d1-d2-d3" type="section">
              <head>
                <hi rend="lsc">Interior</hi>
              </head>
              <p>A high platform (<hi rend="i">whata</hi>) was constructed in the old houses by laying longitudinal joists (<hi rend="i">tarawa</hi>), five on one side and four on the other to leave a space for an opening, over the crossbeams (<hi rend="i">vae</hi>), and then laying cross pieces or poles closely over the joists to form a floor (<hi rend="i">papa</hi>). (See <ref target="#BucMani078a">fig. 13</ref>.)</p>
              <p>
                <figure xml:id="BucMani078a">
                  <graphic url="BucMani078a.jpg" mimeType="image/jpeg" xml:id="BucMani078a-g"/>
                  <head><hi rend="sc">Figure</hi> 13. House platform (<hi rend="i">whata</hi>), or upper story: 1, longitudinal joists (<hi rend="i">tarewa</hi>); 2, crossbeams (<hi rend="i">vae</hi>); 3, cross pieces forming floor (<hi rend="i">papa</hi>); 4, opening (<hi rend="i">ngutu whata</hi>, door of platform); 5, ladder (ara); 6, longitudinal beams (<hi rend="i">hapai</hi>); 7, posts; 8, wall plates; 9, principal rafters.</head>
                </figure>
              </p>
              <pb xml:id="n87" n="79"/>
              <p>Through the left side of the rectangular opening (<hi rend="i">ngutu whata</hi>, door of the platform) made in the middle of the floor by the use of shorter cross pieces on either side, the upper chamber formed by the platform was reached by means of a ladder (<hi rend="i">ara</hi>) placed in a slanting position. The ladder was formed of a stout piece of hala trunk with short cross pieces lashed on as steps. The upper story was also used as a dormitory and as a storage space, especially for the mature coconuts kept to form the <hi rend="i">takataka</hi> stage. In special houses set apart for the tribal gods, the bodies of dead chiefs were laid out on the high platform.</p>
              <p>
                <figure xml:id="BucMani079a">
                  <graphic url="BucMani079a.jpg" mimeType="image/jpeg" xml:id="BucMani079a-g"/>
                  <head><hi rend="sc">Figure</hi> 14. Interior arrangement of house: <hi rend="i">a</hi>, ground plan; <hi rend="i">b</hi>, detail of fence (<hi rend="i">apaapa</hi>). 1, inner sleeping division; 2, fence formed of uprights 2 feet to 2 feet 3 inches high, some of which are driven into ground to act as posts, and two horizontal rods lashed to uprights; 3, supporting posts; 4, openings in fence; 5, doorways of house.</head>
                </figure>
              </p>
              <p>The house platform (<hi rend="i">paepae</hi>) now seen in dwelling houses of the Raro-tongan type, is an innovation which accompanied the houses introduced by the missionary, Aporo, from Rarotonga. The old type of house was built on the flat ground without any raised platform. Curbstones, such as mark some of the Tongarevan houses, were not used.</p>
              <p>The ground floor was usually divided by a low fence (<hi rend="i">apaapa</hi>) into an inner rectangular sleeping part and an outer part contiguous to the sides and used as a dining room. The site of the fence was guided by the position of the supporting posts of the roof. The fence had middle gaps opposite the house doorways in all four sides to give admittance to the inner sleeping part. (See <ref target="#BucMani079a">fig. 14</ref>.)</p>
            </div>
            <pb xml:id="n88" n="80"/>
            <div xml:id="t1-body-d1-d4-d1-d2-d4" type="section">
              <head>
                <hi rend="lsc">Wall And Door Screens</hi>
              </head>
              <p>Wall screens (<hi rend="i">pataro whani</hi>) of plaited coconut leaves were used to form side or end walls when required, there being no permanent walls as in the Rarotongan type of house. The eaves of the house were fairly low, so that the depth of one screen was enough to reach the ground. Above, the screens were tied to a convenient part of the framework and simply hung down. The wall so formed was termed <hi rend="i">haihai.</hi> For the technique of the screen, see page <ref target="#n119">111</ref>, and for their appearance when hung, see <ref target="#BucManiP001a">plate 1</ref>.</p>
              <p>Doorways (<hi rend="i">ngutupa.</hi>) The structure of the end framework to provide for end entrances has been shown in <ref target="#BucMani075a">figure 9</ref>. The middles of the sides were also used as entrances, but here one had to stoop below the eaves to enter. The doorways were closed with <hi rend="i">pataro whani</hi> mats identical in structure with the wall screens.</p>
              <p>
                <figure xml:id="BucMani080a">
                  <graphic url="BucMani080a.jpg" mimeType="image/jpeg" xml:id="BucMani080a-g"/>
                  <head><hi rend="sc">Figure</hi> 15. Portable house (<hi rend="i">whare taka</hi>) : 1, lower horizontal beam (<hi rend="i">hapai</hi>); 2, upright supporting post; 3, upper longitudinal beam (<hi rend="i">hapai</hi>); 4, 5, diagonal braces; 6, cross pieces; 7, longitudinal wall plates; 8, end pairs of principal rafters; 9, main ridgepole. (Native terminology as in <ref target="#BucMani074a">figure 8</ref>.)</head>
                </figure>
              </p>
            </div>
          </div>
          <div xml:id="t1-body-d1-d4-d1-d3" type="section">
            <head>
              <hi rend="sc">Whare Taka</hi>
            </head>
            <p>A small portable house (<hi rend="i">whare taka</hi>) was constructed without supporting posts fixed in the ground. (See <ref target="#BucMani080a">fig. 15</ref>.) Short supporting posts were lashed to either end of longitudinal beams (<hi rend="i">hapai</hi>) laid on the ground. Other longitudinal beams, also termed <hi rend="i">hapai</hi>, were lashed to the tops of the short posts. Two diagonal crossed braces used on each side to brace the two longitudinal beams and posts together were lashed at the corners and at the point of crossing. The two sides so formed were held upright, and cross pieces were laid across the upper longitudinal beams and lashed. The longitudinal wall plates were laid over the ends of the cross beams. The addition of principal rafters, purlins, thatch rafters, upper ridgepole, and eaves rod follows the method described on page <ref target="#n82">74</ref>.</p>
          </div>
          <pb xml:id="n89" n="81"/>
          <div xml:id="t1-body-d1-d4-d1-d4" type="section">
            <head>
              <hi rend="sc">Whare Tuku Whakararo</hi>
            </head>
            <p>A low house without supporting posts (<hi rend="i">whare tuku whakararo</hi>, house let down) was, as the name implies, let down to the ground by doing away with the supporting posts (<hi rend="i">pou</hi>), longitudinal beams (<hi rend="i">hapai</hi>), crossbeams (<hi rend="i">vae</hi>), and wall plate (<hi rend="i">kaupapa</hi>). Thus the principal rafters rested directly on the ground and formed a roof without walls. The house (<ref target="#BucMani081a">fig. 16</ref>) was built for use in the hurricane season. As it had no side walls, the wind could not get under the roof to blow it away. The framework technique follows that already described.</p>
            <p>
              <figure xml:id="BucMani081a">
                <graphic url="BucMani081a.jpg" mimeType="image/jpeg" xml:id="BucMani081a-g"/>
                <head><hi rend="sc">Figure 16</hi>. House without supporting posts (<hi rend="i">whare tuku whakararo</hi>), end view of framework: 1, principal rafters; 2, principal ridgepole; 3, purlins; 4, thatch rafters; 5, upper ridgepole.</head>
              </figure>
            </p>
          </div>
          <div xml:id="t1-body-d1-d4-d1-d5" type="section">
            <head>
              <hi rend="sc">Rarotongan Types Of Houses</hi>
            </head>
            <p>Two variations of the Rrarotongan type of house introduced by the missionaries are shown in <ref target="#BucManiP001a">plate 1</ref>. The ridgepole is supported directly by end posts. Wall posts along the sides support a wall plate, also directly. The principal rafters extend between the ridgepole and wall plate, but their upper ends are crossed above the main ridgepole instead of below it. The upper ridgepole then rests in the forks formed by the principal rafters. The details of the Rarotongan form of framework have been described (27, p. 4).</p>
            <p>One of the houses, used as an assembly place, has straight thatching at the ends and to almost the level of the side eaves. Below the thatching of the sides and ends the house is open. The other house is used for sleeping, and as a result of missionary influence, no doubt, the sides and ends are walled in. The walls receive the Rarotongan name, <hi rend="i">paruru</hi>, as they differ from the local <hi rend="i">haihai</hi> of movable screen sheets. The walls have also departed from the Rarotongan pattern in that the rods used to fill in the spaces between the wall posts are attached horizontally instead of vertically. The end thatching is also projected outward at its lower end to form a veranda. A raised platform of lime surrounds the house.</p>
            <pb xml:id="n90" n="82"/>
            <p>A third variation (<ref target="#BucMani082a">fig. 17</ref>), which is popular in small dwelling houses, consists of carrying the outward projection of the <hi rend="i">tara</hi> (p1. 1, <hi rend="i">B</hi>, 1) to both ends and on the two sides. The small house is thus surrounded by a wide veranda on all four sides, and when set on a raised platform, the covered area looks more extensive than the space actually contained within the walls. The principle, however, of providing plenty of veranda space is a sound one for the tropics.</p>
            <p>
              <figure xml:id="BucMani082a">
                <graphic url="BucMani082a.jpg" mimeType="image/jpeg" xml:id="BucMani082a-g"/>
                <head><hi rend="sc">Figure</hi> 17. Rarotongan type of house, with surrounding veranda: 1, end, with lower part projected to form veranda roof; 2, veranda roof; 3, house platform; 4, posts supporting outer edges of veranda roof.</head>
              </figure>
            </p>
          </div>
          <div xml:id="t1-body-d1-d4-d1-d6" type="section">
            <head>
              <hi rend="sc">House Furniture</hi>
            </head>
            <p>Seats (<hi rend="i">nohoanga</hi>). Though the people usually sat on mats on the ground, a wooden seat (<ref target="#BucMani082b">fig. 18</ref>) was made of <hi rend="i">tou</hi> timber. The seat was rectangular with the slightest curve from side to side and was supported by four legs all cut out of the solid timber with the seat. It was maintained that these were made in Rakahanga in olden times.</p>
            <p>The seats are much less curved than those of the Cook Islands (27, p. 43), and the lower ends of the legs are plain and without the heart-shaped feet characteristic of the islands to the south.</p>
            <p>
              <figure xml:id="BucMani082b">
                <graphic url="BucMani082b.jpg" mimeType="image/jpeg" xml:id="BucMani082b-g"/>
                <head><hi rend="sc">Figure</hi> 18. Wooden seat (<hi rend="i">nohoanga</hi>) : <hi rend="i">a</hi>, side view; <hi rend="i">b</hi>, end view. 1, rectangular seat, 24.25 inches long, 9.25 inches wide, and 1 inch thick, upper surface 7 inches from ground at either end and 6.5 inches at middle; 2, legs; 3, elliptical cross section of leg at junction with seat, 3.4 inches in transverse axis of seat and 1.9 inches in longitudinal axis of seat; 4, round cross-section of leg base, diameter 1.5 inches.</head>
              </figure>
            </p>
            <pb xml:id="n91" n="83"/>
            <p>A round wooden box (<hi rend="i">puiha</hi>) 10 inches in diameter at the bottom, 8 inches at the top, and 9 inches high, was seen. It had ten short legs 0.25 inch high and 0.5 inch by 0.4 inch in cross section. The upper edge had an inner raised rim which fitted against an outer raised rim on the edge of a wooden lid, which was provided with a perforated knob in the center of its upper surface. A perforated lug at the opposite ends of a diameter had been cut out of the solid on the outer surface of the box. A cord passed through the lugs and the lid knob to fasten the lid in position. As <hi rend="i">turuma</hi> was an alternative name given to the box, I pointed out that the term <hi rend="i">tuluma</hi> and the details of the box construction existed in the Tokelau Islands. At first it was stoutly maintained that the box was native to Rakahanga, but later an old man conversant with its history was found who stated that it was brought from Tokelau by Tuteru-utua, who acquired it during his wanderings after being deposed from the position of <hi rend="i">Whakaheo</hi>. The wooden box, therefore, does not belong to the local culture, but the box and the incident are described here to clarify the position and prevent its being accepted when no further historical check may be available.</p>
            <p>Brooms (<hi rend="i">ruruku tu nikau</hi>) were made of coconut leaflet midribs (<hi rend="i">tu nikau</hi>) which were bound together (<hi rend="i">ruruku</hi>). The midribs, as they were detached, were torn off with a jerk which brought away a thin strip of the main leaf midrib. Then the thin ends were plaited together in a three-ply braid until a sufficient quantity was thus provided. The braided end was then rolled to form the midribs into a bundle, which was bound below the braided ends. With such brooms the houses were swept clean. The same form of broom is common in the Cook Islands and other islands, and exists in Samoa.</p>
          </div>
        </div>
        <div xml:id="t1-body-d1-d4-d2" type="section">
          <head>
            <hi rend="c">Food</hi>
          </head>
          <div xml:id="t1-body-d1-d4-d2-d1" type="section">
            <head>
              <hi rend="sc">Introduction</hi>
            </head>
            <p>In comparison with high islands, the low volcanic islands or atolls are greatly restricted as to natural variety in food. Manihiki and Rakahanga were not on the ancient sea routes followed by the voyaging canoes that sailed between Society Islands and Cook Islands, and this may in part account for the failure of many of the foods carried by the early Polynesians to reach these two atolls. The pig, the dog, and the domestic fowl were not known. A small rat, no doubt introduced accidentally, was not eaten, according to the inhabitants. For flesh foods, the people depended on the sea and the lagoon. Certain sea birds with their eggs also augmented the larder. Of vegetable foods, the introduction of the coconut was attributed to Huku, who, besides planting a <hi rend="i">ni ponga</hi> which floated ashore, brought others from his land of birth. Tradition does not connect Huku with the <hi rend="i">puraka</hi> (species of taro)
<pb xml:id="n92" n="84"/>
which was grown on the atolls. It is probable that this plant was brought in by some of the local voyagers who were reputed to have visited other lands, such as Tokelau, and to have returned. If Huku did come from Rarotonga, his not bringing the <hi rend="i">puraka</hi> is in accordance with the fact that the <hi rend="i">puraka</hi> was not grown on that island. If he brought other species of taro they perished, for the <hi rend="i">puraka</hi> seems to be the only species that thrives in Manihiki and Rakahanga. The presence of the <hi rend="i">puraka</hi> resulted in the use of wooden pounders which are not found on Tongareva, owing to the absence of that food plant, and of coral slabs for grating. Of the native plants, the hala (<hi rend="i">Pandanus</hi>) and the <hi rend="i">Morinda citrifolia</hi> were the only ones that could be utilized. If the hala was introduced, there is no historical narrative supporting its diffusion as there is in Tongareva. With the main vegetable supplies consisting of the coconut and the <hi rend="i">puraka</hi>, the people made the best of their limited resources by serving them in as many different forms as possible to add variety to their diet.</p>
            <p>Variety of cooking utensils was also restricted by paucity of material. Much of the food was prepared in the kitchen, a separate establishment set at the back of the dwelling house to shelter the earth oven from the rain.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d2-d2" type="section">
            <head>
              <hi rend="sc">Fire</hi>
            </head>
            <p>Fire was generated (<hi rend="i">hika</hi>) by means of a typical Polynesian fire plough. A groove was rubbed on a lower piece of dry wood (<hi rend="i">kauneti</hi>) by means of a pointed rubbing stick (<hi rend="i">kaurima</hi>) of hard wood. It was customary for an assistant to steady the far end of the <hi rend="i">kauneti</hi> by pressing it firmly on the ground with the foot. This process of pressing down (<hi rend="i">tomi;</hi> Maori, <hi rend="i">taomi</hi>) the under piece was regarded as necessary, as shown by quotations from fire myths. By the friction of the <hi rend="i">kaurima</hi>, particles of wood collected at the far end of the groove, smouldered, and ignited (<hi rend="i">kua tu te ahi</hi>). The ignited particles were blown upon gently to light up the wood dust and were then emptied onto a piece of dry coconut husk (<hi rend="i">puru</hi>), which, by gentle waving to and fro, was also ignited. From the flaming <hi rend="i">puru</hi> the fire was lit.</p>
            <p>Firewood was a problem, as native forests were not extensive. Dry branches of trees were used, but the coconut, besides supplying food, also provided much of the firewood used. The dry flower sheath (<hi rend="i">taume</hi>) and the dry flower stalks (<hi rend="i">roro</hi>) were gathered for heating the oven. Empty coconut shells (<hi rend="i">ipu</hi>), thrown in heaps near the kitchen, were utilized for firewood as they are even to the present day.</p>
            <p>The earth oven (<hi rend="i">umu</hi>) was used, but suitable stones which exist in abundance in high islands were absent on the coral atolls. Recourse was had, as in Tongareva, to coral and <hi rend="i">Tridacna</hi> shells. The coral, after one heating, became friable, and fresh material had to be collected for each oven. The lack of suitable large leaves in the local flora led to the plaiting of special sheets from coconut leaves to provide covers for the food while it was cooking, in order to retain the heat. Thus the people of the atolls of Rakahanga and Tongareva used plaited oven covers like their kinsmen in New Zealand, whereas those of Cook Islands and Society Islands used covers of breadfruit, banana, and other large leaves which their flora provided in abundance.</p>
            <p>Tongs (<hi rend="i">pingohi</hi>) made by doubling pieces of coconut leaf midrib were used to lift heated stones from the oven in some of the culinary operations and in the making of coconut oil.</p>
          </div>
          <pb xml:id="n93" n="85"/>
          <div xml:id="t1-body-d1-d4-d2-d3" type="section">
            <head>
              <hi rend="sc">Fire Myth</hi>
            </head>
            <p>The widespread fire myth associated with the culture hero, Maui, is present in Rakahanga, but it differs somewhat from versions of other areas in that the fire was obtained from Tangaroa-tuhi-mata. According to a manuscript written by Tupou-rahi, Tangaroa-tuhi-mata was the paternal grandfather of Maui and dwelt in the underworld, Hawaiki-i-raro. The underground residence coincides with that of Mahuika in Maori myth and of Mafui'e in Samoan myth. The Rakahangan version has evidently mixed up the names of a widespread legend. Makuai-whare, unable to light her fire, sent her son Maui to his grandfather in the underworld to obtain a lighted piece of firewood (<hi rend="i">motumotu</hi>), instructing him not to go by the forbidden path (<hi rend="i">ara tapu</hi>) but by the common road (<hi rend="i">ara noa</hi>). Maui, true to the mischievous perversity with which he has been credited by all branches of the Polynesians, went by the forbidden path. Tangaroa, seeing a man approach by the path restricted to the gods, commenced to work magic. He raised his right hand, but Maui likewise raised his hand and continued to walk toward Tangaroa. Tangaroa raised his left hand and then both hands, but Maui raised both hands and kept moving. Tangaroa turned one side (<hi rend="i">koko</hi>), but Maui also turned, and approached sideways. Tangaroa turned his other side and Maui followed suit without slowing his progress. Tangaroa turned his back; Maui copied him and walked backwards toward Tangaroa. The various magical movements having failed to impede Maui's progress, Tangaroa desisted and asked Maui who he was and what he wanted. Maui informed his grandfather of their relationship and asked for a light to kindle his mother's fire. Tangaroa proceeded to generate fire while Maui held down (<hi rend="i">tomi</hi>) the lower fire stick. A piece of lighted husk was then given to Maui, who, when he was out of Tangaroa's sight, extinguished (<hi rend="i">pokia</hi>) the brand. Maui returned with the excuse that he had fallen down and accidentally put out the fire. Tangaroa told him to take the <hi rend="i">kauneti</hi> and generate the fire himself. Maui proceeded to do so, but as
<pb xml:id="n94" n="86"/>
Tangaroa did not press down the fire stick, it moved about (<hi rend="i">hinga-hinga</hi>). Maui complained, and Tangaroa said, “Call those birds to fly hither to steady it.” (<hi rend="i">Uru atu ki na maimua, kia rere mai kia tomia.</hi>) Maui perceived two large sea birds (<hi rend="i">kakave</hi>), which, when Maui called, flew down and steadied the fire plough by standing on the far end. Maui, having successfully generated the fire, rewarded his assistants by striking them on the head with the hand stick (<hi rend="i">kaurima</hi>). One flew north and the other south, and to this day the <hi rend="i">kakave</hi> birds bear on their heads the marks (<hi rend="i">whakairo</hi>) of Maui's fire stick. After an interlude, Maui returned to his mother with the lighted husk.</p>
            <p>A further mythical incident in the Maui exploits associates fire with the great exploit of snaring the sun. The Maoris, who were agriculturists, say that the reason for the snaring was to retard the speed of the too rapidly moving sun in order that man might have more time to cultivate food. The Rakahangan myth states that there was barely time in the course of the sun's journey across the sky to light one fire. Maui journeyed to the opening (<hi rend="i">rua</hi>) from which the sun emerged and twice snared him with ropes of sennit braid (<hi rend="i">kaha</hi>), but each time the sun broke away. The sun is referred to as “te tama nui a Hina ko te ra” (the great son of Hina was the sun). Hina was the caretaker of the opening through which the sun emerged. After his failures, Maui sought counsel with her. Her advice (<hi rend="i">wananga</hi>) was to make a snare <hi rend="i">(here)</hi> with strands of her hair (<hi rend="i">iho rauru</hi>). With the hair snare, Maui captured the sun and made the following terms before he let him go. The sun was to proceed slowly (<hi rend="i">haere maria</hi>) to enable mankind to light a morning fire (<hi rend="i">ahi popongi</hi>), a day fire (<hi rend="i">ahi awatea</hi>), and an evening fire (<hi rend="i">ahi ahiahi</hi>). The sun consented and has ever since carried out his bargain.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d2-d4" type="section">
            <head>
              <hi rend="sc">Cooking Utensils And Food Accessories</hi>
            </head>
            <p>Nature provided in the hard shell of the coconut a natural cooking receptacle. The stalk end of the nut was tapped all around with a stone and removed. After the flesh within had been grated with a hand grater, the fluid was poured back and the top of the shell was replaced as a cover. The shell was then placed in the <hi rend="i">umu</hi> with the other food, and the coconut preparation was cooked in its own natural container, which served as pot or casserole. Fish were sometimes cooked in coconut containers.</p>
            <p>A coconut shell cup (<hi rend="i">ipu</hi>) used as a drinking cup for water was made by cutting off the base end of the mature husked shell. Coconut fluid was drunk directly from the opened green nut.</p>
            <p>A water vessel (<hi rend="i">ohonu</hi>) was provided by the larger mature nut, the eye hole being punctured and the flesh rotted out with sea water.</p>
            <pb xml:id="n95" n="87"/>
            <p>Wooden bowls (<hi rend="i">kumete</hi>) were made of trunk sections of <hi rend="i">whano</hi> and <hi rend="i">tou</hi> timber. (See <ref target="#BucMani087a">fig. 19</ref>.) They are now fairly scarce, for trade dishes are easily procured. The bowls seen were elliptical, with projecting lugs for handles, flat bottoms, and no legs. They somewhat resemble the type in common use in Samoa. Other types are figured by Edge-Partington (6, series 1, p. 62, nos. 2–4). Bowl number 2 (<ref target="#BucMani087a">fig. 19, <hi rend="i">c</hi></ref>) Edge-Partington describes: “Oval bowl carved out of cinnamon-colored wood, the carved design is similar to that on the High Island and Mangaian adzes.” The bowl
<figure xml:id="BucMani087a"><graphic url="BucMani087a.jpg" mimeType="image/jpeg" xml:id="BucMani087a-g"/><head><hi rend="sc">Figure</hi> 19. Wooden bowls (<hi rend="i">kumete</hi>). <hi rend="i">a</hi>, large elliptical bowl with handles (1, 2) and flat rim (3) which widens out at ends, surface on same plane as upper surface of handles; left handle (1) grooved on upper surface to serve as spout for pouring off liquids. <hi rend="i">b</hi>, smaller elliptical bowl with end handles (1, 1) projecting outward and no groove; flat rim; convex outer surface (3); bottom (4) flat without legs. <hi rend="i">c</hi>, bowl apparently elliptical, handle (1) appears to project upward from rim instead of outward; four-sided legs (2); carving (3) of lozenges cut out of wood and narrow raised borders between contiguous lozenges on outer surface; diameter, 11.25 inches; height, 4.5 inches. <hi rend="i">d</hi>, oval bowl, handle (1) projects upward from wide end and narrow end (2) has groove cut below rim surface to form spout; no legs; length, 14.5 inches; width, 11 inches; height, 4.5 inches. <hi rend="i">e</hi>, flat elliptical bowl, handles (1, 2) project outward, left handle (1) formed of double projection; length, 20 inches; width, 10.25 inches; height, 3 inches. (Bowls <hi rend="i">c-e</hi> in British Museum, after Edge-Partington.)</head></figure>
<pb xml:id="n96" n="88"/>
appears to be elliptical instead of oval, and it has four-sided legs. Bowl number 3 (<ref target="#BucMani087a">fig. 19, <hi rend="i">d</hi></ref>) he describes: “Oval bowl with end scooped out for pouring.” The narrow end of the bowl is shaped after the beaker type of bowl seen in Cook Islands, with a groove cut below the rim surface to form a spout. Bowl number 4 (<ref target="#BucMani087a">fig. 19, <hi rend="i">e</hi></ref>) he describes: “Flat oval bowl ‘Cuvette à eau’, carved out of wood resembling satin wood and inlaid with discs of pearl shell.” Wooden bowls were used as containers in which <hi rend="i">puraka</hi> and <hi rend="i">uto</hi> preparations were pounded. They were also necessary for the cooking of some food preparations, especially those which were heated or cooked by dropping heated stones into a liquid. Round bowls for coconut cream (<hi rend="i">roro</hi>), characteristic of Tongareva, are absent.</p>
            <p>Hand graters (<hi rend="i">tuai</hi>) made of pearl shell are still in everyday use, due to the extensive use of unripe coconut as food. The soft coconut flesh is better dealt with by the hand grater than by a stand grater. The hand grater (<ref target="#BucMani089a">fig. 20</ref>) is shaped somewhat like a European shoe horn. The part of the shell toward the hinge forms the grip, and the part toward the free edge forms the cutting edge. The grip end may be pointed or rounded and in the long graters may include a portion of the shell hinge. The curved grating edge is sharpened by grinding down the back of the shell, and it may be plain or serrated. The women are very expert in the use of the hand grater. By altering the pressure they make the grating fine or coarse. In addition to grating coconut flesh, the implement is used to grate the edible soft husk and also the <hi rend="i">uto</hi>, absorbing organ of the coconut. (For types, see <ref target="#BucManiP001a">pl. 1</ref>; <ref target="#BucMani089a">fig. 20</ref>.)</p>
            <p>The stand grater (<hi rend="i">kautuai</hi>) was used to grate the mature coconut in preparing coconut cream and obtaining coconut oil. It evidently resembled the Tongarevan stand grater in which a piece of coral was lashed to a wooden limb. The Cook Islands type of stand grater with a serrated metal grating edge is now in common use.</p>
            <p>Another form of grater was provided by a block of sharp coral (<hi rend="i">punga taratara</hi>, on the rough surface of which raw <hi rend="i">puraka</hi> was rubbed to procure the grated form necessary in a particular food preparation and <hi rend="i">takataka</hi> oil.</p>
            <p>The rough skin from the tail of the ray fish (<hi rend="i">hiku whai</hi>) was also used as a grater. A wooden scraper was used to detach the flesh from the keys of the ripe hala (<hi rend="i">Pandanus</hi>) fruit, but no specimen was seen. The combined wringer and strainer (<hi rend="i">kainga</hi>) was made from the pounded husk of the <hi rend="i">ni mata</hi> stage of the coconut. The <hi rend="i">kainga</hi> was used in expressing the coconut cream from the grated mature nut. The sheets of the fibrous textile-like stipule (<hi rend="i">kaka</hi>) at the base of the growing coconut leaves was also used as a strainer, especially in the preparation of coconut oil.</p>
            <p>Pounders (<hi rend="i">reru</hi>) were rendered necessary by the presence of the <hi rend="i">puraka</hi>, which was mashed in certain food preparations. The <hi rend="i">uto</hi>, though fairly soft
<pb xml:id="n97" n="89"/>
when cooked, was also mashed in the <hi rend="i">oveke</hi> preparation. The pounder (<ref target="#BucManiP001a">pl. 1</ref>; <ref target="#BucMani089a">fig. 21</ref>) was made of wood in the general form of a potato masher with a narrow neck and a terminal knob. The food was pounded directly in a wooden bowl without the intervention of a special pounding table (<hi rend="i">papahia</hi>) such as was used in Cook Islands. Pounders of coral were not made.</p>
            <p>
              <figure xml:id="BucMani089a">
                <graphic url="BucMani089a.jpg" mimeType="image/jpeg" xml:id="BucMani089a-g"/>
                <head><hi rend="sc">Figure</hi> 20. Pearl-shell hand graters (<hi rend="i">tuai</hi>). <hi rend="i">a</hi>, front and side views of short grater (C.2788) with pointed grip, thick hinge part of shell not used: 1, pointed grip end; 2, cutting edge, serrated but worn through use; 3, longitudinally concave front surface of nacreous inner coating of shell; 4, convex back, outer shell surface ground to remove roughness; 5, posterior grinding of cutting edge; length of grater, 4.3 inches; width at lower end, 1.3 inches; greatest thickness, slightly more than 0.1 inch. <hi rend="i">b</hi>, front and side views of long grater: 1, grip, rounded; 2, edge, smooth; 3, marked front longitudinal concavity; 5, posterior grinding; length of grater, 5.7 inches; width at lower end 1.6 inches; width at upper end, 0.7 inch; greatest thickness, 0.2 inch. <hi rend="i">c</hi>, front and side views of long grater with part of thick hinge included: 1, thick upper end; 2, serrated edge; 3, front longitudinal convexity, not marked; 5, posterior grinding of cutting edge; 6, thick hinge included because of extra length of grater, ground down on front surface; 7, point at which hinge part is 0.7 inch wide; length of grater, 6.5 inches; width at lower end, 1.4 inches; thickness of hinge, 0.4 inch; thickness of blade, 0.2 inch.</head>
              </figure>
            </p>
            <p>
              <figure xml:id="BucMani089b">
                <graphic url="BucMani089b.jpg" mimeType="image/jpeg" xml:id="BucMani089b-g"/>
                <head><hi rend="sc">Figure</hi> 21. Pounder (<hi rend="i">reru</hi>) made of <hi rend="i">ngangie</hi> wood (C.3020), rounded, with fairly equal diameters: 1, greatest diameter at base, 2.8 inches; 2, neck, diameter, 1.2 inches; 3, knob, diameter, 2.3 inches; 4, base, flatly convex; 5, groove between knob and neck to take supporting cord by which implement may be hung up, probably modern, due to metal nails or hooks in kitchen; length of pounder, 13 inches.</head>
              </figure>
            </p>
            <pb xml:id="n98" n="90"/>
            <p>Special containers for <hi rend="i">puraka</hi> were plaited from coconut leaves to hold mashed and grated preparations for consignment to the cooking oven. These <hi rend="i">raurau</hi> and <hi rend="i">pite</hi> containers are described on pages <ref target="#n122">114</ref>–<ref target="#n126">118</ref>.</p>
            <p>The coconut husker (<hi rend="i">ko</hi>) was a pointed stick driven into the ground like the common Polynesian type of husker. The short husker held between the feet by the Tongarevans was not used. As the green husk at the base of the nut in the early stages of growth was used for food, it is probable that different methods of husking nuts prevailed as in Tongareva (29, p. 117), but details were not procured.</p>
            <p>A short piercer (<hi rend="i">ko poto</hi>) made of a sharpened piece of <hi rend="i">ngangie</hi> wood was used to pierce the <hi rend="i">ni mata</hi> nut for drinking. This was used in the plantations and obviated husking the nut. The nut was held close to the face; the piercer was driven through the <hi rend="i">mokomoko</hi> butt end of the nut. As the piercer was removed the mouth was quickly applied over the hole, for the liquid spurted out and was likely to be wasted. The method of drinking (<hi rend="i">mokomoko kiri</hi>) was not without danger. The movements had to be so quick that, with the face held close to the nut, the exact site for stabbing was judged and not seen. Sometimes judgment erred and the cheek was stabbed instead of the nut.</p>
            <p>Climbing bandages were used to assist in ascending coconut trees to pluck the nuts. Two forms are now in use:</p>
            <p>1. Common <hi rend="i">tari</hi> bandage. This bandage is obtained from material stripped from the leaf of young growing plants which throughout the plantations have sprung from self-grown nuts. A leaf from the outer side of the closed central leaves (<hi rend="i">rito</hi>) is torn off, and the skin from the upper surface (<hi rend="i">aro</hi>) of the midrib is peeled off from the smooth part below the first two leaflets (<hi rend="i">puwha</hi>). About six strips (<hi rend="i">tari</hi>) 3 feet long are beaten against a tree trunk, chewed, and rubbed between the hands to render them soft and pliable. The ends are tied in a reef knot to form a closed loop about 16 inches long. The band is slipped over the dorsum of each foot and gives the climber a purchase against the palm trunk as he straightens his knees and seeks a higher grip with his hands.</p>
            <p>2.<hi rend="i">Kaha-piki</hi> bandage (<hi rend="i">kaha</hi>, sennit braid; <hi rend="i">piki</hi>, to climb). The bandage is made of a length of sennit braid wound loosely into five loops about 16 inches long when pulled taut. The braid ends are knotted together and seized for a few turns around the five strands at the knot.</p>
            <p>The carrying pole (<hi rend="i">amo</hi>) was formed of a convenient length of coconut leaf midrib. No shaped wooden poles were seen in use.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d2-d5" type="section">
            <head>
              <hi rend="sc">Flesh Foods</hi>
            </head>
            <p>In the absence of cannibalism and many flesh foods, the proteid part of the diet was based on fish and shellfish augmented sometimes by wild fowl.</p>
            <p>Fish were plentiful within the lagoon and in the deep sea. They were caught by a variety of methods (see pp. <ref target="#n166">158</ref>–<ref target="#n202">194</ref>). Certain fish were
<pb xml:id="n99" n="91"/>
eaten raw as a change in diet and because the taste was appreciated. The general routine, however, was to cook them in the earth oven. When the sea and season were favorable, the people fared well. When, however, the sea was rough for some time in particular seasons, want of a flesh complement (<hi rend="i">ninaki</hi>) to eat with the vegetable food was felt. On such occasions the <hi rend="i">Takai-whakaheo</hi> chiefess was called upon to exercise her power in calming the sea to enable the hungry fishermen to supply the larder. Cooked fish left over from meals were dried and kept in baskets for future use.</p>
            <p>The waters of the outer lagoon were rich in crayfish, which were caught at night by torching and formed an important addition to the diet. They were cooked in the oven.</p>
            <p>Two species of land crab, the <hi rend="i">tupa</hi> and the <hi rend="i">koveu</hi>, are plentiful, but the <hi rend="i">koveu</hi> (coconut crab) was more sought after. Certain islands were closed until the <hi rend="i">koveu</hi> were plentiful, when the restrictions were removed. Motungangie was opened for our party. We went crabbing at night with torches of dry coconut leaves. The ground was damp from rain, and the crabs, bloated, purplish objects with powerful claws, were found out of their holes, resting on tree trunks. We procured a large number by picking the crabs from the tree trunks or poking them down with the torch or a stick when out of reach. The natives grasp them behind the claws and then tie them with strips of bark or <hi rend="i">tari</hi> so that their claws are pinioned. The crabs are cooked in the oven. The huge claws contain a delicate flesh, but the rich fatty material in the round bloated part is considered the best portion. It is very rich and oily, but palatable. On account of the oil, I should imagine that the <hi rend="i">koveu</hi> supplies an important addition to the atoll diet.</p>
            <p>The inner lagoons in both atolls were rich in pearl oyster (<hi rend="i">parau</hi>) and <hi rend="i">Tridacna</hi> (<hi rend="i">pahua</hi>). The <hi rend="i">Tridacna</hi> formed an important food supply and were eaten either raw or cooked in the oven. <hi rend="i">Tridacna</hi> were also strung on strips of <hi rend="i">tari</hi> after cooking and dried in the sun to form a reserve supply. When dried, <hi rend="i">Tridacna</hi> become as hard as leather and keep for a considerable time. After they are recooked in the oven, they become soft again. We found <hi rend="i">Tridacna</hi> a useful change while on Rakahanga. The women of our household were eager to get us a supply when we expressed the wish. The <hi rend="i">Tridacna</hi> were of medium size and were obtained from the sandy bottom of the inner lagoon in comparatively shallow water. Almost every household had strings of dried <hi rend="i">Tridacna</hi> hanging up in the cooking house. In eating raw shellfish, a certain amount of salty liquid which satisfies the craving for salt is ingested. Other smaller shellfish are found in smaller quantities and are collected by the women.</p>
            <p>Turtles were caught and cooked, but did not seem to play as important a part in the marae ceremony as they did in Tongareva.</p>
            <pb xml:id="n100" n="92"/>
            <p>All wild fowl that could be caught were grist to the mill. The man-of-war hawk (<hi rend="i">kotaha</hi>) and the brown booby (<hi rend="i">toroa</hi>) were caught at night on their rookeries. A strip of coconut midrib skin (<hi rend="i">tari</hi>) tied in a running noose was attached to a handle of <hi rend="i">ngangie</hi> wood. The stiff <hi rend="i">tari</hi> material kept the noose open. The birds roost on low <hi rend="i">tauhunu</hi> bushes and along the leaves of coconut trees. On the low plants, they were easily reached with the noose. The higher coconut trees were climbed by the fowler, and birds within reach were caught both by hand and by snare. The coconut leaf was bent down, which caused the birds farther out to climb up toward the butt, when they came within reach. The <hi rend="i">kotaha</hi> colored white on the neck under the bill are termed <hi rend="i">kotaha mari</hi> and those colored red, <hi rend="i">kotaha tarakura.</hi> Two smaller seabirds, the <hi rend="i">ngoio</hi> and <hi rend="i">rakie</hi>, were also caught.</p>
            <p>Other birds were caught with set snares or killed with stones (<hi rend="i">pehi ki te toka</hi>). The snares consisted of running loops of coconut husk fiber arranged in rows on sticks with the loops directed upward. Barriers were made on opposite sides of the set snares to force the birds to pass through the snares as they moved about on their quest for food. The birds said to be caught in the snares were the <hi rend="i">torea, kihi, parauanga, kururi, rahurahu</hi>, and <hi rend="i">moiho.</hi></p>
          </div>
          <div xml:id="t1-body-d1-d4-d2-d6" type="section">
            <head>
              <hi rend="sc">Vegetable Foods</hi>
            </head>
            <div xml:id="t1-body-d1-d4-d2-d6-d1" type="section">
              <head>
                <hi rend="lsc">Variety</hi>
              </head>
              <p>The vegetable food supplies consisted of the coconut, the <hi rend="i">puraka</hi>, the fruit and tips of the aërial rootlets of the hala (<hi rend="i">Pandanus</hi>), and the fruit of the <hi rend="i">nenu</hi> (<hi rend="i">Morinda citrifolia</hi>) which was used as a flavoring agent with one of the food preparations.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d2-d6-d2" type="section">
              <head>
                <hi rend="lsc">The Coconut</hi>
              </head>
              <p>The coconut was introduced by Huku and planted by successive generations, so that at the present time all the islands are thickly covered by coconut groves. Many of the trees have been self-planted and there appears to have been no particular system followed.</p>
              <p>The green nuts were plucked from the trees as required by the owners. The mature nuts that fell to the ground were gathered in heaps and removed as required. A heap of nuts (<hi rend="i">komua</hi>) belonged definitely to the owner of the trees, as expressed by the phrase, “Te komua ni nei na Tuteru.” (This heap of coconuts belongs to Tuteru.) A large heap is termed <hi rend="i">whetomo</hi>, but the term is sometimes loosely applied to a small heap.</p>
              <p>In removing nuts, a strip of husk was peeled down, leaving one end attached. By means of the strip, the nuts were tied together in pairs. The
<pb xml:id="n101" n="93"/>
connecting strip of husk was termed a <hi rend="i">whakahani</hi> and the term was used in counting the nuts in pairs as follows:</p>
              <p>
                <table rows="11" cols="3">
                  <row>
                    <cell>E tahi (1) whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">2</cell>
                  </row>
                  <row>
                    <cell>E rua (2) whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">4</cell>
                  </row>
                  <row>
                    <cell>E teru (3) whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">6</cell>
                  </row>
                  <row>
                    <cell>E ha (4) whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">8</cell>
                  </row>
                  <row>
                    <cell>Purupuru whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">10</cell>
                  </row>
                  <row>
                    <cell>Purupuru ma tahi whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">12</cell>
                  </row>
                  <row>
                    <cell>Purupuru ma rua whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">14</cell>
                  </row>
                  <row>
                    <cell>Purupuru ma teru whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">16</cell>
                  </row>
                  <row>
                    <cell>Purupuru ma ha whakahani</cell>
                    <cell>=</cell>
                    <cell rend="right">18</cell>
                  </row>
                  <row>
                    <cell>Takau</cell>
                    <cell>=</cell>
                    <cell rend="right">20</cell>
                  </row>
                  <row>
                    <cell>Rau</cell>
                    <cell>=</cell>
                    <cell rend="right">200</cell>
                  </row>
                </table>
              </p>
              <p>The coconut tree is termed <hi rend="i">ni</hi>, the leaves, <hi rend="i">nikau</hi>, and the textile-like stipule at the base of the leaf, <hi rend="i">kaka.</hi> The flower is first inclosed in a small sheath (<hi rend="i">poepoe</hi>), from which an inner sheath (<hi rend="i">taume</hi>) emerges. The <hi rend="i">taume</hi> grows to a much greater length than the <hi rend="i">poepoe.</hi> From the <hi rend="i">taume</hi> emerges the flower (<hi rend="i">roro</hi>), which consists of a main spike with a number of smaller spikes. The secondary spikes have a number of male flowers (<hi rend="i">pua ni</hi>), with one female flower (<hi rend="i">huariri</hi>) at the base of each secondary spike. When the female flower has been fertilized and commences to grow, the male flowers with the main and secondary spikes dry up, and, still under the term <hi rend="i">roro</hi>, are used as firewood when the nuts are removed. The long boat-shaped <hi rend="i">taume</hi> also dries and is used as firewood. The growing stipules are cut off for use as strainers for coconut cream, and dry stipules are used as kindling for the fire generated with the fire plough.</p>
              <p>The nut is contained within a thick envelope of husk (<hi rend="i">puru</hi>) which is covered by an outer skin (<hi rend="i">kiri</hi>). Nuts in the green stages (<hi rend="i">ni mata</hi>) are classified into two kinds, those in which the husk has a sweet taste (<hi rend="i">mangaro</hi>), and those which are not sweet (<hi rend="i">kawa</hi>, bitter). The husk of the green nuts is soft and contains a certain amount of moisture, whereas that of the mature nuts is hard and dry. The <hi rend="i">mangaro</hi> husk, owing to its sweetness, is used as food. It may be chewed raw like sugar cane to extract the sweet juice, or it may be cooked. When eaten raw, the husk is removed in segments with a husking stick and the outer skin (<hi rend="i">kiri</hi>) is peeled off. The term for chewing is <hi rend="i">ngaungau</hi>, but chewing <hi rend="i">mangaro</hi> husk receives the special term, <hi rend="i">kopani.</hi> To cook the husk, the nut is cooked whole without husking. The husk is then removed and chewed. The fluid contained within the nut may be drunk in the ordinary way.</p>
              <p>The husk at the base or stalk end of the green <hi rend="i">ni mata</hi> stage contains less fiber and more parenchymatous material (<hi rend="i">mokomoko</hi>; Tongarevan, <hi rend="i">karisi</hi>) which is utilized as food. The green nuts are husked in such a way as to leave the <hi rend="i">mokomoko</hi> end still attached to the shell. The outer end which carried the stalk is broken off, and the <hi rend="i">mokomoko</hi> material is scraped off in small pieces with a hand grater to be mixed with dishes prepared from the grated nut. The material has a slightly bitter taste and thus provides a flavoring agent, as in Tongarevan cookery. In the later stages of growth, the <hi rend="i">moko
<pb xml:id="n102" n="94"/>
moko</hi> is invaded by more fibrous material and becomes too hard and dry for food.</p>
              <p>The nut, as it develops, contains a fluid within the nut cavity which at first is too bitter to drink but which in the green nut stages forms a cool, refreshing beverage. The fluid also enters into the composition of dishes made from the grated flesh. In the mature stages of the nut the fluid becomes too bitter for use and is gradually absorbed by a part of the growing embryo and diminishes in quantity until it entirely disappears. The various stages of the nut are distinguished not only by external appearance, but also by shaking the nut to determine by the sound and feel the quantity of fluid present.</p>
              <p>The fluid appears in the earliest stages before the flesh. The flesh then appears at the base end of the shell as a thin, slimy layer (<hi rend="i">havarevare</hi>). In the <hi rend="i">ni mata</hi> stages of growth, the flesh completely covers the interior of the shell and gradually thickens and becomes firmer. In this condition, the flesh is grated and utilized in various dishes. When the <hi rend="i">hakari</hi> stages are reached, the flesh assumes its maximum thickness and hardness and is still used as food. From then on, two courses are open. If the embryo grows, it gradually absorbs the flesh until only a thin, hard layer is left. If it does not grow, the flesh remains thick, hard, and well preserved. It shrinks slightly from the shell, and the loosened flesh rattles within the shell when the nut is shaken. This preserved form provides a useful reserve food.</p>
              <p>The embryo develops in the flesh at the base of the nut. One part grows out externally through the patent eye of the shell base to form the root, stem, and leaves of the growing plant. Another part appears as a rounded protrusion (<hi rend="i">mata uto</hi>) at the base of the cavity. This is the growing absorbing organ which gradually fills the cavity with a soft, spongy mass (<hi rend="i">uto</hi>). The <hi rend="i">uto</hi> is edible raw or cooked and forms a welcome variation from the ordinary flesh of the nut. The <hi rend="i">uto</hi>, in expanding, absorbs all the remaining fluid, and unless used at the stage when it fills the cavity, it will gradually absorb the mature hard flesh with its contained fat crystals.</p>
              <p>The coconut provides food from the husk, flesh, fluid, and <hi rend="i">uto</hi> which are present in different stages of growth. The particular stages of growth were thus of the greatest practical importance. The individual was not forced to make the food preparations from a haphazard collection of nuts, but he selected his nuts for the particular dishes that he desired. Experience had taught him exactly what parts of the nut were still edible in the nuts that he saw growing on the trees. Accumulated knowledge led him to recognize the different stages from their practical value as food. From the female flower to the nut only fit for planting, thirteen names distinguish the stages of the coconut growth (Table 15).</p>
              <pb xml:id="n103" n="95"/>
              <p>
                <table rows="13" cols="2">
                  <head>Table 15. Stages in Growth of the Coconut</head>
                  <row>
                    <cell> 1. Huariri</cell>
                    <cell>Female flower.</cell>
                  </row>
                  <row>
                    <cell> 2. Mokomoko puapua</cell>
                    <cell>Small, no flesh.</cell>
                  </row>
                  <row>
                    <cell> 3. Havarevare</cell>
                    <cell>Flesh starting to grow; thin, slimy.</cell>
                  </row>
                  <row>
                    <cell> 4. Ni mata mua<lb/>
   Ni mata muri</cell>
                    <cell>Thin flesh covers interior of shell.</cell>
                  </row>
                  <row>
                    <cell> 5. Ni mata matua</cell>
                    <cell>Flesh thicker.</cell>
                  </row>
                  <row>
                    <cell> 6. Ni momoto</cell>
                    <cell>Shell darker, flesh mature.<lb/>
Slight rattle of fluid.</cell>
                  </row>
                  <row>
                    <cell> 7. Hakari kahatea</cell>
                    <cell>Flesh thicker, fluid rattles more.</cell>
                  </row>
                  <row>
                    <cell> 8. Hakari <hi rend="i">uri</hi></cell>
                    <cell>Flesh hard (copra).</cell>
                  </row>
                  <row>
                    <cell> 9. Takataka</cell>
                    <cell><hi rend="i">Hakari uri</hi> kept in house, fluid dries, flesh loosens from shell and can be rattled, hence <hi rend="i">takataka.</hi></cell>
                  </row>
                  <row>
                    <cell>10. Mata uto</cell>
                    <cell>Commences to sprout, absorbing organ shows at base inside, fluid not all dry.</cell>
                  </row>
                  <row>
                    <cell>11. Uto pine</cell>
                    <cell>Cavity filled with <hi rend="i">uto</hi> (<hi rend="i">pine</hi>, filled up), no rattle. <hi rend="i">Katinga</hi> flesh still present. <hi rend="i">Uto</hi> may be eaten and flesh made into copra.</cell>
                  </row>
                  <row>
                    <cell>12. Uto <hi rend="i">puni</hi></cell>
                    <cell><hi rend="i">Katinga</hi> thinner and harder, too hard to eat. More <hi rend="i">uto</hi> but too tough. Kept too long, only fit to grow.</cell>
                  </row>
                  <row>
                    <cell>13. Purapura uri</cell>
                    <cell>Uto <hi rend="i">puni</hi> when planted.</cell>
                  </row>
                </table>
              </p>
              <p>The flesh can be eaten uncooked at any stage from the thin <hi rend="i">havarevare</hi> to the hard <hi rend="i">takataka</hi> and the <hi rend="i">katinga</hi> of the <hi rend="i">uto pine.</hi> The fluid is drunk from the <hi rend="i">havarevare</hi> to the <hi rend="i">ni momoto</hi> stages. Out in the plantations where the nuts are opened for drinking, the flesh is generally eaten on the spot, or the nut is carried home for a later meal. In the earliest stages, when the flesh is thin, it is readily removed from the shell with the fingers. From the <hi rend="i">ni mata muri</hi> stage, it is customary to grate the flesh with a hand grater (<hi rend="i">tuai</hi>). The slices of flesh removed may be made fine or thick by altering the pressure on the grater. To grate finely is termed <hi rend="i">varuvaru</hi>, and to remove in coarser slices is <hi rend="i">tupere</hi>. The finer grating is used with the softfleshed <hi rend="i">ni mata</hi> stages and the coarser grating with the firmer <hi rend="i">ni momoto.</hi> Hence the following chant or <hi rend="i">pese:</hi></p>
              <lg type="verse">
                <l>Tupere au te ni momoto,</l>
                <l>Maroro au,</l>
                <l>Ki te kopu,</l>
                <l>Pakari te waewae,</l>
                <l>Matutu te kopapa.</l>
              </lg>
              <lg type="verse">
                <l>I grate the <hi rend="i">ni momoto</hi> coarsely,</l>
                <l>That I may be strong,</l>
                <l>That the stomach be filled,</l>
                <l>The legs strengthened,</l>
                <l>And the body fattened.</l>
              </lg>
              <p>Before a green nut is grated the fluid may be poured out into another shell, and the process of grating is then termed <hi rend="i">varuvaru maro</hi> (dry grating). In some preparations a little of the fluid is drunk to prevent spilling, and the grating process with the rest of the fluid in the shell is termed <hi rend="i">varuvaru tavai</hi> (wet grating).</p>
              <p>The grated flesh of green nuts with the fluid retained (<hi rend="i">whakaehu</hi>) forms the basis of a number of preparations. The grated flesh of green nuts
<pb xml:id="n104" n="96"/>
without the fluid is termed <hi rend="i">takarari.</hi> When mixed with other food such as hala or <hi rend="i">puraka</hi>, the term <hi rend="i">pana</hi> is used to denote the mixture, but the actual mixing process with two grated foods is <hi rend="i">kahiro.</hi> When required, the liquid contained in the grated nut is squeezed through a stipule strainer. The remaining dry flesh is termed <hi rend="i">ota.</hi></p>
              <p>The mature <hi rend="i">hakari</hi> flesh is eaten (<hi rend="i">ka ngaungau te katinga</hi>). The fluid is apt to be bitter and is never sought after as a drink, though it may be used with the flesh. The Tongarevans do not drink it.</p>
              <p>The <hi rend="i">takataka</hi> nut has three recognized stages: <hi rend="i">takataka maimaeha</hi>, in which the flesh is still completely soft; <hi rend="i">takataka whati</hi>, in which the flesh hardens but has not changed color; and <hi rend="i">takataka kura</hi>, in which the flesh darkens and turns somewhat reddish (<hi rend="i">kura</hi>) after being kept some time. The <hi rend="i">takataka kura</hi> is regarded as the best form for eating. The flesh is eaten without any special preparation. The thin inner part left after rubbing off the outer part for oil on sharp coral or ray skin is termed <hi rend="i">uhio</hi> and is eaten. The hard nut is eaten with dried fish, dried <hi rend="i">Tridacna</hi> (<hi rend="i">pahua maro</hi>), and <hi rend="i">puraka.</hi> A person with a large stock of <hi rend="i">takataka</hi> used it as a medium of exchange for fish.</p>
              <p>The <hi rend="i">uto</hi> stages of the nut are, of course, known by the presence of the growing plant projecting from the base of the unhusked nut. When the nut is unhusked it is termed <hi rend="i">uto kiri</hi> (<hi rend="i">kiri, skin</hi>), and when husked it is <hi rend="i">uto ko</hi> (<hi rend="i">ko</hi>, to husk). The growing part is jerked off (<hi rend="i">huhuti</hi>) from the outside. A part left attached to the <hi rend="i">uto</hi> is called the <hi rend="i">pito</hi> (navel) and is removed (<hi rend="i">huhuti</hi>) after the shell is opened. To open the husked shell is <hi rend="i">kohoa</hi> or <hi rend="i">whowhoa.</hi> The end of the <hi rend="i">uto</hi> toward the base of the shell is the <hi rend="i">mata</hi>, as it is near the <hi rend="i">mata</hi> hole in the shell. The other end is the <hi rend="i">take.</hi> To cut in pieces is <hi rend="i">tehi</hi>, as in <hi rend="i">tehia mai na uto</hi> (to cut up the <hi rend="i">uto</hi>). Several preparations are made with the <hi rend="i">uto.</hi></p>
              <p>The <hi rend="i">uto puni</hi> is used medicinally but is of no use as food. The Manihi-ki-Rakahangan people considered it too tough; but as they had the <hi rend="i">puraka</hi>, they did not have the same incentive to experiment with pitting the <hi rend="i">uto</hi> as the Tongarevans, who had to rely on the coconut alone.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d2-d6-d3" type="section">
              <head>
                <hi rend="lsc">Puraka</hi>
              </head>
              <p>The <hi rend="i">puraka</hi> species of taro was extensively grown on both Manihiki and Rakahanga. Large areas had been excavated, especially on the island of Rakahanga, so that the level which was well below the surface reached the brackish water beneath. The large plots must have entailed considerable community labor, for the areas so dealt with are quite extensive. The spoil has been thrown out so long ago that it has formed natural-looking mounds and ridges at the sides of the excavations. These mounds on Rakahanga
<pb xml:id="n105" n="97"/>
form the highest part of the island, and it is here that the people retire to avoid the high tidal waves that inundate the land during some of the severe hurricanes. When the mature <hi rend="i">puraka</hi> was dug up for food, the top of the leaves termed the seed or <hi rend="i">puraka purapura</hi> was replanted. The tubers were then pitted in a damp place in the sand near the cookhouse for use as required. The presence of the <hi rend="i">puraka</hi> gave the people a great advantage over the atoll of Tongareva, which the <hi rend="i">puraka</hi> never reached in pre-European times. A number of preparations were made from it.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d2-d6-d4" type="section">
              <head>
                <hi rend="lsc">Hala (Pandanus)</hi>
              </head>
              <p>Hala fruit is eaten raw when ripe, or it may be cooked on the stones of an oven without covering. Hala gratings were also dried and kept as reserve stock.</p>
            </div>
          </div>
          <div xml:id="t1-body-d1-d4-d2-d7" type="section">
            <head>
              <hi rend="sc">Food Preparations</hi>
            </head>
            <p>1. <hi rend="i">Whakaehu mata.</hi> The grated flesh of the <hi rend="i">ni mata muri</hi> coconut is eaten uncooked (<hi rend="i">mata</hi>) by itself after drinking the fluid of the nut or with the fluid mixed with it. As a flavoring agent, the grated husk (<hi rend="i">mokomoko</hi>) from the base of a green <hi rend="i">mangaro</hi> nut may be mixed with it.</p>
            <p>2. <hi rend="i">Wai-tahi.</hi> The grated material of the <hi rend="i">whakaehu mata</hi> may be cooked in an oven in its own shell by itself or with the husk flavoring. The preparation is then termed <hi rend="i">wai-tahi.</hi></p>
            <p>3. <hi rend="i">Whakaehu pana puraka.</hi> The <hi rend="i">puraka</hi> is cooked whole and the <hi rend="i">whakaehu</hi> preparation made from <hi rend="i">ni mata muri</hi> is eaten with it. It is really a combination and not a single preparation.</p>
            <p>4. <hi rend="i">Takarari.</hi> The flesh of the <hi rend="i">ni mata muri</hi> is grated dry and worked (<hi rend="i">oi</hi>) with the hands. The difference between it and <hi rend="i">whakaehu</hi> is simply that of fluid. Both are eaten uncooked.</p>
            <p>5. <hi rend="i">Verovero puraka.</hi> The <hi rend="i">ni mata muri</hi> is wet-grated and raw <hi rend="i">puraka</hi> is cut into thin slices (<hi rend="i">tipu rahirahi</hi>). A layer of grated coconut is placed in the bottom of a coconut shell and alternate layers of coconut and <hi rend="i">puraka</hi> made until the shell is filled. The mixture is then cooked.</p>
            <p>6. <hi rend="i">Romanga.</hi> The full name of this mixture is <hi rend="i">whakaehu pana ota romanga.</hi> As the name implies, the grated green coconut flesh (<hi rend="i">whakaehu</hi>) is mixed (<hi rend="i">pana</hi>) with the dry gratings (<hi rend="i">ota</hi>) left after expressing the oil from the grated mature nut (<hi rend="i">hakari</hi>). The grated mature nut is treated with heated stones to bring out the oil. The mixture prevents waste and supplies the <hi rend="i">whakaehu</hi> with a different flavoring agent.</p>
            <p>7. <hi rend="i">Whakaehu pana nenu.</hi> The ripe fruit of the <hi rend="i">Morinda citrifolia</hi> (<hi rend="i">nenu</hi>) is squeezed in a wooden bowl with the hands to express the juice. The fluid from a green nut is poured in over the fruit to assist the extraction of the juice, as the fruit is fairly dry. The flavored fluid is then put in the empty shell of a green nut with the shell cover over it and cooked. The cooked mixture will remain good for two days. This is used as a flavoring agent with the <hi rend="i">whakaehu</hi> preparation of uncooked grated green nut (<hi rend="i">ka kahiro ki te whakaehu</hi>).</p>
            <p>8. <hi rend="i">Tupere.</hi> The apical end of a <hi rend="i">ni mata matua</hi> is cut off, some of the fluid drunk, and the flesh wet-grated in coarse pieces. It receives its name from the coarse grating (<hi rend="i">tupere</hi>) as compared with the finer grating (<hi rend="i">varuvaru</hi>) of the <hi rend="i">whakaehu</hi> preparations which are made from <hi rend="i">ni mata muri</hi> or <hi rend="i">ni mata mua.</hi> The <hi rend="i">tupere</hi> preparation is largely used by nursing women to increase the milk flow, for which it is said that the <hi rend="i">ni mata muri</hi> of the <hi rend="i">whakaehu</hi> preparations is useless. <hi rend="i">Tupere</hi> is also used as a food complement (<hi rend="i">ninaki</hi>; Maori, <hi rend="i">kinaki</hi>) with fish.</p>
            <p>9. <hi rend="i">Pana mokomoko.</hi> The <hi rend="i">ni mata mua</hi> is finely wet-grated, and the gratings of the <hi rend="i">mokomoko</hi> husk of a green bitter nut
<pb xml:id="n106" n="98"/>
(<hi rend="i">ni kava</hi>) are mixed with it. The mixture turns reddish in color. The <hi rend="i">mokomoko</hi> by itself is bitter, but when mixed with the grated coconut flesh it turns sweet.</p>
            <p>10. <hi rend="i">Huripaka.</hi> The flesh of a <hi rend="i">ni mata matua</hi> is dry-grated with the husk of a green sweet coconut (<hi rend="i">ni mangaro</hi>). The two are mixed, placed in a specially made coconut leaflet container (<hi rend="i">rarau</hi>), and cooked in an oven. The cooked preparation is very dark in color.</p>
            <p>11. <hi rend="i">Puraka pana ota.</hi> Cooked <hi rend="i">puraka</hi> is pounded (<hi rend="i">tukituki</hi>) and mixed with the coarsely grated flesh of the <hi rend="i">ni momoto.</hi></p>
            <p>12. <hi rend="i">Wai ta.</hi> The grated flesh of a <hi rend="i">ni momoto</hi> with its fluid is put into a wooden bowl and mixed with grated husk of a green bitter nut. The mixture is then squeezed through a strainer to separate the flesh. The strained fluid is termed <hi rend="i">wai ta</hi> and is used medicinally by adults for pain in the back. The dry <hi rend="i">ota</hi> flesh is used for the <hi rend="i">puraka pana ota.</hi></p>
            <p>13. <hi rend="i">Roro.</hi> The <hi rend="i">roro</hi> is the creamy fluid from the mature nut grated on a standgrater and expressed through a stipule strainer. The cream may be used uncooked (<hi rend="i">roro mata</hi>), or a hot stone may be dropped into the bowl to heat it (<hi rend="i">roro hehengi</hi>). The <hi rend="i">roro mata</hi> may be used as a laxative. The preparation did not assume the social importance that it did in Tongareva, and no special bowls were made for it.</p>
            <p>14. <hi rend="i">Uto kai mata.</hi> The <hi rend="i">uto</hi> is cut up and eaten raw.</p>
            <p>15. <hi rend="i">Uto tupere.</hi> The opened shell is held in the hand and with a hand grater the <hi rend="i">uto</hi> is coarsely grated within the shell, as is <hi rend="i">katinga</hi> flesh. The two are mixed with the hand grater, which also acts as a spoon in conveying the mixture to the mouth. The person does not wait to grate the whole quantity but eats as he grates.</p>
            <p>16. <hi rend="i">Turu uto.</hi> The <hi rend="i">uto</hi> is cooked in the shell in the oven. The <hi rend="i">mata</hi> ends are cut off with a hand grater and eaten by the family. The other ends are arranged around cooked <hi rend="i">puraka</hi> in the middle of a bowl and conveyed to the eating place (<hi rend="i">whainga</hi>) where guests are assembled.</p>
            <p>17. <hi rend="i">Uto haehae.</hi> The uncooked <hi rend="i">uto</hi> is shredded (<hi rend="i">haehae</hi>) with the fingers, placed in a bowl, and the grated hard flesh of the nut is strained over the shredded <hi rend="i">uto.</hi> The fluid is very oily.</p>
            <p>18. <hi rend="i">Oveke.</hi> Cooked <hi rend="i">uto</hi> is pounded (<hi rend="i">tukituki</hi>) in a wooden bowl and the grated mature flesh of the nut is mixed with it.</p>
            <p>19. <hi rend="i">Pitei.</hi> A <hi rend="i">ni mata</hi> nut is dry-grated, put in a <hi rend="i">pite</hi> (coconut leaf container) and cooked in the oven. The <hi rend="i">uto</hi> is also cooked in the shell and afterwards pounded in a wooden bowl. The cooked coconut gratings (<hi rend="i">takarari</hi>) are mixed with the pounded <hi rend="i">uto</hi>, and the resulting mixture is <hi rend="i">pitei.</hi></p>
            <p>20. <hi rend="i">Pupu.</hi> Cooked <hi rend="i">puraka</hi> and <hi rend="i">uto</hi> are pounded and the <hi rend="i">takarari</hi> preparation from <hi rend="i">ni mata</hi> is prepared. The directions are as follows: “Oti, kua pana te puraka ki te uto, kua pana te takarari ki te puraka e te uto. E teru mea kua pana.” (When finished, the <hi rend="i">puraka</hi> is mixed with the <hi rend="i">uto</hi>, the grated coconut is mixed with the <hi rend="i">puraka</hi> and the <hi rend="i">uto.</hi> Three things are mixed.) “Raranga teraurau kapukapu. Kua hua ki roto, kua ruruku. Kia pu te umu, kua to. Huke ake, kua kai.” (Plait the <hi rend="i">kapukapu</hi> basket. Place within and fold up. When the oven is ready, cook. Open up the oven, eat.)</p>
            <p>21. <hi rend="i">Uto whakapapa.</hi> The cooked <hi rend="i">uto</hi> is placed on a piece of <hi rend="i">katinga</hi> (hard flesh of the nut) and the two are eaten together, just as one would eat bread and butter.</p>
            <p>22. <hi rend="i">Puraka mata.</hi> The <hi rend="i">puraka</hi> is eaten uncooked.</p>
            <p>23. <hi rend="i">Puraka to.</hi> The <hi rend="i">puraka</hi> is scraped to remove the skin and cooked in an oven. It is eaten with <hi rend="i">uto</hi>, hard coconut, and fish. It is also dipped into <hi rend="i">roro.</hi> The <hi rend="i">puraka</hi> is grated on a coral slab to provide a number of dishes. When grating, the slab is placed on a <hi rend="i">kurei</hi> mat made of coconut leaflets so as to catch the grated material.</p>
            <p>24. <hi rend="i">Papa puraka.</hi> The grated <hi rend="i">puraka</hi> is mixed with coconut cream (<hi rend="i">roro</hi>) and folded up in a <hi rend="i">puraka</hi> leaf. This is placed in the oven, with the end of the folded leaf below to prevent the parcel's unraveling. The cooked preparation is <hi rend="i">papa puraka.</hi></p>
            <pb xml:id="n107" n="99"/>
            <p>25. The raw grated <hi rend="i">puraka</hi> is kneaded (<hi rend="i">oi</hi>) with the hands in a wooden bowl, placed in a <hi rend="i">pite</hi> (container made of coconut leaflets), and cooked. Sometimes only the outer parts of the <hi rend="i">puraka</hi> are grated off, when the preparation is termed <hi rend="i">oro</hi>, and the middle inner part (<hi rend="i">tahi</hi>) is cooked separately in another <hi rend="i">pite</hi> container. (<hi rend="i">Pite ke to ti oro, pite ke to ti tahi</hi>). When cooked, both the <hi rend="i">oro</hi> and the <hi rend="i">tahi</hi> are put in a wooden bowl, coconut cream is poured over them, and the contents of the bowl are stirred up and mixed (<hi rend="i">karo</hi>) with a piece of coconut midrib or a wooden pounder. The heat brings out the oil in the cream and the preparation is exceedingly sweet. (<hi rend="i">Kua mea te hinu, kua none whakarere</hi>).</p>
            <p>26. <hi rend="i">Poke.</hi> The raw grated <hi rend="i">puraka</hi> is kneaded with the hands in a bowl and then placed on a plaited coconut platter termed a <hi rend="i">papa.</hi> The material is placed near the midrib edge and the far edge of the <hi rend="i">papa</hi> is folded over and tied. When cooked, the <hi rend="i">puraka</hi> is again kneaded and stirred up in a bowl, as the under part is somewhat harder from its position on the heated stones. The cooked <hi rend="i">puraka</hi> is made up into balls (<hi rend="i">popo</hi>) about the size of oranges. Heated coconut cream (<hi rend="i">roro hehengi</hi>) is then spread over the balls. The preparation is termed <hi rend="i">poke.</hi></p>
            <p>27. <hi rend="i">Mahu.</hi> The raw <hi rend="i">puraka</hi> is pounded. A receptacle is formed by placing a close <hi rend="i">kurei</hi> in a coconut leaflet basket (<hi rend="i">kete</hi>). The <hi rend="i">puraka</hi> is placed within and the <hi rend="i">kurei</hi> folded (<hi rend="i">whatuwhatu</hi>) over it. The basket is closed (<hi rend="i">tia</hi>) with two-ply twisted sennit (<hi rend="i">whauhoto</hi>) and laid aside (<hi rend="i">ka tuku ki waerenga</hi>) until the <hi rend="i">puraka</hi> becomes soft (<hi rend="i">whakape</hi>) and emits an odor (<hi rend="i">haunga piropiro</hi>). When required, some is taken out, kneaded in a bowl, placed in a <hi rend="i">pite</hi> (container) and cooked. When cooked, it is served with coconut cream. The food is fermented in the same manner as breadfruit but is kept in a closed basket instead of a pit.</p>
            <p>28. <hi rend="i">Whakarikoriko.</hi> The <hi rend="i">puraka</hi> is pounded raw, placed in a large coconut shell container (<hi rend="i">ipu</hi>) with coconut cream, and cooked.</p>
            <p>29. <hi rend="i">Verovero.</hi> The <hi rend="i">puraka</hi> is cut into very thin slices. Commencing with the <hi rend="i">puraka</hi>, alternate layers of <hi rend="i">puraka</hi> and grated coconut (<hi rend="i">whakaehu</hi>) are used to fill a coconut container. A topmost layer must be of coconut to prevent the <hi rend="i">puraka</hi> from becoming too hard. When cooked, the slices of <hi rend="i">puraka</hi> are removed with the layer of grated coconut above it and both are eaten together.</p>
            <p>30. <hi rend="i">Puraka whakapara. Puraka</hi> is cooked whole in the oven and while still hot is placed in baskets which are closed with ties (<hi rend="i">tia more</hi>). All is laid aside until the <hi rend="i">puraka</hi> is dark, when it is recooked after any mould that may have appeared has been washed off. When being eaten, the <hi rend="i">puraka</hi> is dipped in coconut cream. The preparation is sweet.</p>
            <p>31. <hi rend="i">Turoro. Puraka</hi> is cut very thin (<hi rend="i">kotikoti rahirahi</hi>) and laid in a coconut shell container. Coconut cream is expressed over the sliced <hi rend="i">puraka.</hi> The oven stones are covered with the husk (<hi rend="i">kainga</hi>) of the <hi rend="i">ni mata.</hi> The vessels are placed upon the husk and the oven closed. When cooked, the preparation forms <hi rend="i">turoro.</hi></p>
            <p>32. <hi rend="i">Popo puraka.</hi> Cooked <hi rend="i">puraka</hi> is pounded and mixed with heated coconut cream in which the oil is brought out.</p>
            <p>33. <hi rend="i">Turu puraka.</hi> Large-sized <hi rend="i">puraka</hi> called <hi rend="i">tahua</hi> are cooked, placed in a wooden bowl with the larger ends uppermost, and heated coconut cream is poured over them.</p>
            <p>34. <hi rend="i">Reru.</hi> The <hi rend="i">puraka</hi> is cooked whole and pounded after the skin is peeled off (<hi rend="i">teretere</hi>). Heated coconut cream is mixed with it and the mixture rounded off (<hi rend="i">popo</hi>) into dumplings. The separate dumplings are placed on pieces of <hi rend="i">puraka</hi> leaf which are brought together at the top and tied. They are then cooked and served.</p>
            <p>35. <hi rend="i">Puraka pana whakaehu.</hi> The pounded cooked <hi rend="i">puraka</hi> is mixed with grated green coconut (<hi rend="i">whakaehu</hi>).</p>
            <p>36. <hi rend="i">Penu whara.</hi> The keys of the hala (<hi rend="i">Pandamis</hi>) fruit are pounded off the central core, and the inner soft ends are grated on a special instrument into a container. The <hi rend="i">penu whara</hi> is sweet-smelling as well as palatable.</p>
          </div>
          <pb xml:id="n108" n="100"/>
          <div xml:id="t1-body-d1-d4-d2-d8" type="section">
            <head>
              <hi rend="sc">Fats And Oil</hi>
            </head>
            <p>The need for fat in the diet was evidenced by the fondness of the people for fatty foods. During the seasons when certain fish were fat, they were eagerly sought after. The oily part of the <hi rend="i">koveu</hi> (land crab) and the fat of the sea birds assisted in providing the necessary ingredients to the diet. An ever-present source of fat was the coconut, the mature flesh of which contains abundant fat crystals. The fat crystals were available not only in the mature flesh (<hi rend="i">katinga</hi>) but also in the coconut cream expressed from the grated flesh. In some food preparations the oily flavor was brought out by heating the coconut cream with hot stones. The <hi rend="i">uto</hi> (absorbing organ) also absorbs oil from the mature flesh and thus supplies it to the diet. Even in the preparation of coconut oil, the burned, grated material, after being strained and squeezed, is not wasted but is used in the food preparation known as <hi rend="i">romonga.</hi> As the coconut formed the staple food, it may be said that the atoll diet was fairly rich in oil.</p>
            <p>Coconut oil was used medicinally and also for rubbing over the body. Three forms of oil were prepared:</p>
            <p>1. <hi rend="i">Hinu takataka.</hi> As the name implies, the oil (<hi rend="i">hinu</hi>) is obtained from the mature nut (<hi rend="i">takataka</hi>), the flesh of which is very oily. It is obtained by chewing the flesh, and the chewed material is rubbed over the head and body. This is especially done by young people to render themselves attractive in the evening after the labors of the day have ended. Any extra material left over is put in the body belt (<hi rend="i">taoa</hi>) by young adults going out to keep love appointments (<hi rend="i">whakaturi</hi>).</p>
            <p>The <hi rend="i">takataka</hi> oil is also prepared by grating the flesh on a slab of sharp coral or on the skin from the tail of the ray fish (<hi rend="i">hiku whai</hi>). The quantity required is small and is used immediately after its extraction from the grated nut. To rub is <hi rend="i">ukui</hi>, but the special grating process with the <hi rend="i">takataka</hi> nut is termed <hi rend="i">puoro</hi> or simply <hi rend="i">oro.</hi> The oil is contained, for the most part, in the middle of the flesh. As the pieces of flesh detached from the shell are ground individually, it is customary to rub off the outer and middle layers while the thin (<hi rend="i">angiangi</hi>) inner layer of flesh is being eaten. The oil may also be directly expressed by gathering up the gratings in a stipule strainer and squeezing the liquid through into a vessel. The oil may be kept in a coconut cup or container. This oil is also used medicinally.</p>
            <p>2. <hi rend="i">Hinu romonga.</hi> The oil is procured from the hard flesh of the mature <hi rend="i">hakari kahatea</hi> and <hi rend="i">hakari uto</hi> nuts as well as from the growing nuts in the <hi rend="i">mata uto</hi> and <hi rend="i">uto pine</hi> stages. All these nuts are now used in making trade copra. The nuts are grated on a stand grater (<hi rend="i">kautuai</hi>) and the gratings fall on a <hi rend="i">pataro</hi> mat spread below. The gratings (<hi rend="i">ota</hi>) are exposed to the sun on the mat and turned now and again until the oil begins to show. This process is termed <hi rend="i">haumake.</hi> The mats are then drawn into the shade to prevent the grated material from becoming too hard. Stones heated in the oven are lifted with <hi rend="i">pingohi</hi> tongs, placed on the gratings, and the material at the sides is heaped up over the stones. The material is carefully watched and the stones are moved about to prevent the grated nut from becoming burned. When all is heated to the right appearance, it is transferred to a wooden bowl. The oil is expressed by squeezing (<hi rend="i">tatau</hi>) it through a stipule strainer into a coconut cup where it is allowed to settle. The grated material in which some oil still lingers is used as food in the <hi rend="i">ro</hi>
<pb xml:id="n109" n="101"/>
<hi rend="i">monga</hi> preparation. The heating process is also termed <hi rend="i">romonga</hi> and gives its name to the oil.</p>
            <p>3. <hi rend="i">Hinu pipiro.</hi> The oil is prepared by mixing the inner parts (<hi rend="i">tutae</hi>) of the <hi rend="i">tupa</hi> (land crab) with the grated nut prepared from the same stages of nuts as those used in the extraction of <hi rend="i">romonga</hi> oil. The mixture is wrapped in the leaves of the <hi rend="i">whano</hi>, <hi rend="i">nenu</hi>, or <hi rend="i">puraka</hi> and covered over in a basket, bowl, or coconut leaf. It is left for two or three days to become ripe (<hi rend="i">para</hi>), a state influenced perhaps by the decomposition of the <hi rend="i">tupa.</hi> The material is then heaped up on the two sides of the broad upper concave surface of the butt end of a coconut leaf. The oil runs down the middle clear space into a coconut shell container placed below. The shell full of oil is then stoppered for future use. It was stated that this oil was used for lamps after their introduction. Because of its unpleasant odor (<hi rend="i">pipiro</hi>), it was not likely to have been made for bodily use in ancient times. The dry gratings were eaten and the thick sediment in the oil was applied to ringworm (<hi rend="i">hune</hi>).</p>
          </div>
          <div xml:id="t1-body-d1-d4-d2-d9" type="section">
            <head>
              <hi rend="sc">Feeding Of Children</hi>
            </head>
            <p>Children received careful attention from birth, and food preparations to suit children's ages were made from the coconut.</p>
            <p>It was considered advisable to give a laxative to the baby after the cord had been cut, in order to get rid of what was termed the dark excreta. The laxative was obtained from the <hi rend="i">uto pine</hi> nut, which contains both the <hi rend="i">uto</hi> and the <hi rend="i">katinga.</hi> The <hi rend="i">uto</hi> and a little of the <hi rend="i">katinga</hi> were grated with a hand grater. The mixed grated material was squeezed with a stipule strainer to express the liquid into a coconut shell cup. A piece of the free end of an aërial hala rootlet (<hi rend="i">kai whara</hi>) was used as a teat and dipped into the fluid (<hi rend="i">whakaoma</hi>). The child was encouraged to suck (<hi rend="i">whakaroroma</hi>) the teat and thus absorb the laxative.</p>
            <p>The child was not fed at the breast for three days. The colostrum milk of the mother (<hi rend="i">mea renga</hi>, yellow material) was sucked out by a friend during this period until the normal milk was established. During this threeday period the child was fed on <hi rend="i">waita</hi>, which consisted of the fluid expressed from a grated <hi rend="i">ni momoto</hi> nut, mixed with water, and squeezed through a stipule strainer into a coconut shell cup. The hala rootlet teat was used.</p>
            <p>The child was fed at the breast entirely for four months, after which the breast milk was augmented by giving the child a little grated <hi rend="i">whakaehu</hi> made from the <hi rend="i">ni mata</hi> nut. This was continued until the child was nine months old. Variety was provided by mixing grated <hi rend="i">uto</hi> with the coconut <hi rend="i">whakaehu.</hi> The <hi rend="i">uto</hi> was carefully selected from those which had reached the right full size but had not become tough. Both the <hi rend="i">uto</hi> and the <hi rend="i">ni mata</hi> were finely grated and mixed together.</p>
            <p>After the child was nine months old, the best food for it was cooked <hi rend="i">uto</hi> (<hi rend="i">uto to</hi>).</p>
          </div>
        </div>
        <pb xml:id="n110" n="102"/>
        <div xml:id="t1-body-d1-d4-d3" type="section">
          <head>
            <hi rend="c">Plaiting</hi>
          </head>
          <div xml:id="t1-body-d1-d4-d3-d1" type="section">
            <head>
              <hi rend="sc">Introduction</hi>
            </head>
            <p>Plaiting is woman's craft. It provides the technique by which many articles required in domestic life are made. The men are also skillful with their fingers and make thatch sheets and baskets that may be immediately required, but the production of plaited articles is essentially woman's work. The scope of the craft was wide when no form of weaving was in use. Certain articles of clothing were originally provided by plaiting. The advent of Western culture has affected the form of clothing, but all other plaited articles are as much in use today as they were in pre-European times. Plaiting has thus always been a live craft. It has recently been developed further by the making of more elaborate mats and satchels in lauhala (<hi rend="i">Pandanus leaf</hi>) and new types of fans and hats in prepared coconut leaf. The women of Manihiki and Rakahanga have established a wide reputation as skilled plaiters and far surpass their neighbors of Tongareva. The materials used, coconut leaf or lauhala, influence technique.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d3-d2" type="section">
            <head>
              <hi rend="sc">Plaiting With Coconut Leaves</hi>
            </head>
            <div xml:id="t1-body-d1-d4-d3-d2-d1" type="section">
              <head>
                <hi rend="lsc">Plaited Articles</hi>
              </head>
              <p>The coconut provides a great number of plaited articles which are necessities in everyday life. For houses the coconut leaf furnishes an alternative thatching sheet and supplies wall screens. Within the house, coconut leaf mats form part of the necessary furnishings. A coconut leaf mat is also used to sit on out of doors and during certain ceremonials. The leaf supplies containers in which food is cooked, covers to retain the heat of the oven during cooking, and platters on which cooked food is served. A variety of baskets is made to contain foods and stored clothing. Last of all, the coconut leaf furnishes a shade to protect the eyes during operations conducted during the heat of the day, especially fishing for bonito.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d2" type="section">
              <head>
                <hi rend="lsc">The Leaf</hi>
              </head>
              <p>The leaf has received so much attention that the parts have been named for convenience:</p>
              <list type="simple">
                <item>
                  <p>nikau, full leaf</p>
                </item>
                <item>
                  <p>whani, main midrib</p>
                </item>
                <item>
                  <p>tua (back), upper shiny surface</p>
                </item>
                <item>
                  <p>aro (front), under surface</p>
                </item>
                <item>
                  <p>pihonga, under midrib surface</p>
                </item>
                <item>
                  <p>hikuhiku nikau, tip end of leaf</p>
                </item>
                <item>
                  <p>katuri, leaflet from tip end</p>
                </item>
                <item>
                  <p>rau papata longer</p>
                </item>
                <item>
                  <p>patapata leaflets</p>
                </item>
                <item>
                  <p>ruruku, bundle of leaflets</p>
                </item>
                <item>
                  <p>tuaniu, leaflet midrib</p>
                </item>
              </list>
              <pb xml:id="n111" n="103"/>
              <p>The naming of the two surfaces of the leaf has been reversed from that of Cook Islands. The skin of the midrib is used for rough lashings. The leaflets toward the butt end of the midrib become hard, brittle, and frayed, and most articles are therefore made from the leaflets toward the tip end of the leaf. Thus a woman requiring individual leaflets for a particular type of mat will say to an assistant, “Haehaengina mai na katuri.” (Tear off the leaflets from the tip end of the coconut leaf.)</p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d3" type="section">
              <head>
                <hi rend="lsc">Technical Terms And Methods</hi>
              </head>
              <p>Below the attachment of the leaflet midrib to the leaf midrib the two sides of the leaflet are attached close together in a vertical line to the side of the leaf midrib. The leaflet is therefore folded together at its commencement but soon opens out into its full width.</p>
              <p>To obtain a woven or plaited surface, two sets of elements have to be provided for interlacement. In plaiting, the two sets of elements are fixed at the same commencement edge; one set is turned diagonally toward the left and the other toward the right. Individual strips of material, or wefts, inclined toward the left are sinistral and those toward the right, dextral. The technique is as follows:</p>
              <p>The plaiter sits on the ground with the commencement edge lying transversely toward her and works from left to right in a section of convenient depth. When she attains the depth of a working section on the left, she also defines the left margin of the section by successively turning in the sinistral wefts as they reach the left margin. By turning in the sinistral wefts at right angles to their previous course, she not only defines the left margin of the working section, but she turns the sinistrals back into the plaiting where they now incline toward the right and thus function as dextrals. From the method of beginning, the working edge of the section is on the right of the plaited portion and is inclined obliquely upward toward the left, being formed by the last sinistral weft crossed with a number of dextrals. A triangular section is built up on the left until the worker has a convenient number of dextrals engaged, usually six to eight. With each subsequent movement, a dextral and a sinistral are added on the right of the working edge. From its position, the new dextral is added to the lower end of the oblique working edge, and for convenience the worker drops the top dextral in the last movement so as to keep to the same number of working dextrals. The series of working dextrals is manipulated by the left hand into alternating sets, the arrangement of which depends upon the plaiting stroke used. In check plaiting, the left hand picks up every alternate dextral and leaves the others down. The raising of one set of alternate wefts forms a shed between the two sets in which the new sinistral is laid by the right hand. In the next movement the plaiter drops the top dextral altogether, as she will pick up a new dextral at the lower end of the working edge. Commencing with the upper end of the working edge, she successively drops the raised dextrals over the sinistral placed in the shed and also picks up successively the recumbent set of dextrals from beyond the right edge of the sinistral which lies upon them. In this way, the added new sinistral passes over one and under one crossing element throughout its course in the shed, and the plaiting stroke is thus a check. In covering the last sinistral, the two sets of working dextrals have been reversed in position and a shed provided for the next sinistral, which will continue the check technique. When the wefts pass under
<pb xml:id="n112" n="104"/>
and over more than one crossing element, the stroke is termed a twill. A twilled-two is formed by raising the dextral wefts at the working edge in alternate pairs and leaving the alternate pairs down. When the new sinistral is laid in the shed, it thus passes over two and under two crossing elements. To continue the twill, the technique demands that in the next movement one element of each raised dextral pair be kept raised and a pair provided for it by picking up an adjacent recumbent weft. Two courses are thus open. If the lower member of each pair is retained, the recumbent weft below it must be picked up. This will result in the rows of twilled-twos being horizontal in direction. If the upper element of each raised pair is retained, the recumbent weft above it must be picked up, and the result will be that the twilled rows will run vertically. To form twilled-threes, twilled-fours, and higher combinations, the number indicated is held up and a similar number left recumbent. In each subsequent movement, one of each raised group is dropped and a recumbent picked up to replace it. If the recumbent weft is picked up from below, the twilled row will run horizontally, and if from above, it will run vertically.</p>
              <p>After establishing the technical stroke and the depth of the working section, the craftswoman works along the commencing edge from left to right by successive movements in which she drops the top working dextral, arranges the next shed, picks up a new dextral from below, and places a new sinistral in the shed. She is concerned with arranging the sheds and the plaiting stroke pattern develops automatically. On reaching the right margin, she successively turns in the dextrals as they reach the margin, and by bending them in to function as sinistrals she defines the right marginal edge. The defining of the right and left marginal edges is necessary with each working section of a mat. With baskets, the margins are not turned but are left free and oblique for subsequent treatment. Successive working sections are added until the required depth of the article is reached. The far or finishing edge is then completed, ordinarily by a three-ply braid.</p>
              <p>Plaiting vocabulary used by natives:</p>
              <list type="simple">
                <item>
                  <p>Puwhera (open), open leaflet</p>
                </item>
                <item>
                  <p>Piri (stuck together), closed leaflet</p>
                </item>
                <item>
                  <p>Kohiti, check plaiting</p>
                </item>
                <item>
                  <p>To, prefix before number to distinguish twill</p>
                </item>
                <item>
                  <p>Torua, toteru, twilled-two, twilled-three</p>
                </item>
                <item>
                  <p>Whakatutu (to make standing up) toteru, vertical twilled-three (without qualification <hi rend="i">toteru</hi> refers to horizontal twills)</p>
                </item>
                <item>
                  <p>Peperu, to turn side margins</p>
                </item>
                <item>
                  <p>Hiri, three-ply braid</p>
                </item>
                <item>
                  <p>Take, free braid tail or corner</p>
                </item>
                <item>
                  <p>Huru, additional separate leaflets added to plaiting</p>
                </item>
                <item>
                  <p>Potiki, to narrow plaiting by bringing two wefts together to function as one</p>
                </item>
                <item>
                  <p>Hatu, to commence plaiting two leaflet-bearing strips</p>
                </item>
                <item>
                  <p>Tutaki, to close in ends of basket</p>
                </item>
              </list>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d4" type="section">
              <head>
                <hi rend="lsc">Commencement Edge</hi>
              </head>
              <p>When individual separate wefts are used, as with lauhala, a special technique is needed for joining the two sets of individual elements along a commencement edge. In the coconut leaf, however, nature has provided an established commencement edge in the leaf midrib on which the leaflet wefts are growing at regular intervals. The fixation problem was thus solved by nature. In the local development of the craft of plaiting to satisfy various
<pb xml:id="n113" n="105"/>
needs, the midrib, with rare exceptions, has been used as the fixed commencement edge, but the need for different types of plaited material has led to a diversity of treatment in utilizing the leaf midrib. The whole midrib may be retained and the leaflets of both sides used, or the midrib may be split and the leaflets of one side used. For better work, a thin strip sufficient to carry the leaflets may be carefully split off the main midrib and will thus form a neat commencement edge. Two or four of these midrib strips may be combined. The forms of commencement technique are as follows:</p>
              <p>
                <figure xml:id="BucMani105a">
                  <graphic url="BucMani105a.jpg" mimeType="image/jpeg" xml:id="BucMani105a-g"/>
                  <head><hi rend="sc">Figure</hi> 22. Whole midrib commencement. <hi rend="i">a</hi>, section of coconut leaf, butt end to left, natural upper shiny surface uppermost: in this position, leaflets of both sides naturally directed toward right; leaflets on far side (1–3) retain their natural direction as dextral wefts, and those on near side (1′–3′) will have to provide sinistral crossing wefts; leaflets shown with leaflet midribs (5) running down middle line; part near leaf midrib (4) narrow but opens out away from leaf midrib, <hi rend="i">b</hi>, in plaiting, leaflets kept closed so that leaflet midribs (5) are on right or side to which leaflets naturally directed; first near leaflet on left (1′) used as first sinistral, bent across leaf midrib (4) to pass under first dextral (1), and thus directed towards the left to function as sinistral; bent direct, without any half-turn, so leaflet midrib (5) of sinistral (1′) still remains on right edge of weft. <hi rend="i">c</hi>, next near leaflet (2′) makes direct bend across leaf midrib, passes under next dextral (2), and if a check stroke is maintained, passes above next dextral (1); third sinistral (3′) crosses over leaf midrib, passes under opposite dextral (3), over next dextral (2), and under next (1); in this manner, by direct bends, near leaflets cross over leaf midrib, under opposite dextral leaflets, and subsequent interlacing depends on stroke used.</head>
                </figure>
              </p>
              <p>1. Whole midrib commencement (<ref target="#BucMani105a">fig. 22</ref>). A section of leaf for the required length of the screen is cut off. The section is laid transversely before the plaiter with the tip end toward the right and the shiny surface uppermost. In this position the leaflets on the far side of the midrib form natural dextral wefts. The opposing set of sinistral wefts is formed by turning the near set of leaflets over the midrib and bending them to the left to engage with the natural dextrals in the manner shown in <ref target="#BucMani105a">figure 22</ref>. The leaflets are kept closed (<hi rend="i">piri</hi>).</p>
              <p>2. Single-strip midrib commencement (<ref target="#BucMani106a">fig. 23</ref>). The leaf section is cut to the required length and split down the middle. Only one side of the leaf is utilized, so the crossing wefts are provided by bending alternate leaflets in the opposite direction to their natural course. When only one article is required, the leaf strip with the leaflets directed naturally to the right is preferred as the leaflets form natural dextrals and plaiting can proceed from left to right in the orthodox manner. If the other strip with
<pb xml:id="n114" n="106"/>
the leaflets directed toward the left is used, the plaiting must proceed from right to left. As the bending of alternate leaflets in the opposite direction increases the inter-weft spaces, the open leaflet is used to diminish the spaces.</p>
              <p>3. Simple two-strip commencement (<ref target="#BucMani107a">fig. 24</ref>). The problem of supplying the second set of crossing elements has so far been met by twisting leaflets from the opposite side of a whole midrib or alternate leaflets from the same side of a single strip. A simple advance in technique is to split the leaf midrib and utilize both sides. When the strips are laid with the midrib strips together and the shiny surface of the leaflets up, the strip from the left side of the leaf provides a set of dextrals naturally directed, and the strip from the other side a set of naturally directed sinistrals. This obviates bending leaflets in the opposite direction to their natural inclination. The commencement of a food platter in check with the open leaflet is shown in <ref target="#BucMani107a">figure 24</ref>. In such commencements, it is customary for the sinistral-bearing strip to be placed above the other.</p>
              <p>
                <figure xml:id="BucMani106a">
                  <graphic url="BucMani106a.jpg" mimeType="image/jpeg" xml:id="BucMani106a-g"/>
                  <head><hi rend="sc">Figure</hi> 23. Single-strip midrib commencement. <hi rend="i">a</hi>, midrib strip (8) with leaflets directed naturally toward right, shiny surface up, is selected; leaving out first leaflet (1) on left, following leaflets dealt with in pairs, one on right being converted into sinistral by direct bend over its dextral pair on its left; thus, first sinistral is third leaflet (2′) from left, which is right member of first pair (2, 2′) and is bent to left over its dextral pair (2) on its left; purpose of leaving first leaflet (1) out of pairing is now obvious; as check stroke is used, first sinistral (2′) passes under it and is thus kept down in position, <hi rend="i">b</hi>, right hand members of subsequent pairs (2′–7) have been converted into sinistrals by passing over dextral members (2–7) on their left, and each additional sinistral as formed is interlaced with preceding dextrals in check stroke; in working edge on right one set of dextral alternates (2, 4, 6) has been raised, and other set (1, 3, 5, 7) has been left down; next sinistral (7′) has been bent to take its position in shed formed; as working edge of 6 dextrals is enough, top recumbent dextral (1) will be left out of next movement; top raised weft (2) dropped over sinistral (7′); recumbent weft (3) raised; 4, dropped: 5, raised; 6, dropped; 7, raised; movement is thus from above down in order of wefts; left hand will keep wefts 3, 5, and 7 raised; next pair (8, 8′) will then enter plaiting, one on right (8′) forming sinistral which will be placed in new shed; sinistrals (2′–7′) on left project beyond left marginal dextral (1). <hi rend="i">c</hi>, formation of left edge: lowest free sinistral (2′) bent in with half-turn over sinistral above it (3′) and functions as dextral; next sinistral (3′) also turned over sinistral above it (4′); similarly, next sinistral (4′) will be turned over one above (5′); thus left edge formed by series of half-turns which converts free sinistrals into functioning dextrals.</head>
                </figure>
              </p>
              <pb xml:id="n115" n="107"/>
              <p>4. The twisted two-strip commencement (takawiri, to twist). (See <ref target="#BucMani107b">fig. 25</ref>.) Two midrib strips from opposite sides are used, but before plaiting, each strip is prepared by twisting each leaflet forward under the strip attachment of the leaflet in front (tip end of leaf in direction in which leaflets are naturally inclined). The two strips so dealt with are placed together with the sinistral-bearing strip above as in the simple two-strip commencement. The leaflets cross naturally, and in the subsequent plaiting they are kept closed.</p>
              <p>
                <figure xml:id="BucMani107a">
                  <graphic url="BucMani107a.jpg" mimeType="image/jpeg" xml:id="BucMani107a-g"/>
                  <head><hi rend="sc">Figure</hi> 24. Simple two-strip commencement. <hi rend="i">a</hi>, midrib strip (9) from right side of leaf with leaflets naturally directed toward left placed above strip (10) from other side of leaf, with leaflets directed toward right; leaflets thus cross naturally to be interlaced with required stroke. <hi rend="i">b</hi>, left ends of midrib strips do not exactly coincide, upper strip a little to right in order that its first leaflet (1′) may pass under first leaftet (1) of under strip; leaflets of upper strip raised with right hand; left hand raises first leaflet (1) of lower strip, and first sinistral (1′) dropped under it; left hand drops first dextral (1), raises second dextral (2), and right hand drops next sinistral (2′); second dextral (2) dropped over sinistral (2′) and raises next dextral (3); process carried on to right in order thus begun; left hand picks up next dextral, right drops sinistral, left drops raised dextral over it and picks up next dextral; plaiting in check resembles technique of single split midrib commencement (<ref target="#BucMani106a">fig. 23, <hi rend="i">b</hi></ref>), and side edges formed by upward half-turns; thus on left, lowest sinistral (1′) turned in over sinistral (2′) above it; on right, lowest dextral (8) turned in below dextral (7) above to comply with check technique.</head>
                </figure>
              </p>
              <p>
                <figure xml:id="BucMani107b">
                  <graphic url="BucMani107b.jpg" mimeType="image/jpeg" xml:id="BucMani107b-g"/>
                  <head><hi rend="sc">Figure</hi> 25. Twisted two-strip commencement, <hi rend="i">a</hi>, in dextral-bearing midrib strip (9), twisting commences at butt end so that leaflet 1 passes forward under strip attachment of leaflet 2 in front of it, and thus 2 under 3, 3 under 4, 4 under 5, and 5 will pass under 6. <hi rend="i">b</hi>, in sinistral-bearing strip, commencing from butt end on right, leaflets will successively pass under leaflets in front as shown numerically, <hi rend="i">c</hi>, sinistral-bearing strip (1) placed above dextral-bearing strip (2) and their leaflets cross naturally.</head>
                </figure>
              </p>
              <pb xml:id="n116" n="108"/>
              <p>5. The simple two-pair commencement (<ref target="#BucMani108a">fig. 26</ref>). The number of wefts is reinforced by the use of a pair of midrib strips with natural dextrals and another pair with natural sinistrals. No twist is used, and each pair is so placed that the leaflets of one strip alternate with those of the other. As a double quantity of wefts is thus provided, the edges of the leaflets may be trimmed and made narrower. In a fish platter, the open leaflet was used and both free edges of the leaflets split off with the thumb nail.</p>
              <p>
                <figure xml:id="BucMani108a">
                  <graphic url="BucMani108a.jpg" mimeType="image/jpeg" xml:id="BucMani108a-g"/>
                  <head><hi rend="sc">Figure</hi> 26. Simple two-pair commencement, <hi rend="i">a</hi>, two sinistral-bearing strips placed together so that leaflets of lower one (2) alternate and fill intervals between leaflets of upper strip (1). <hi rend="i">b</hi>, two dextral-bearing strips similarly placed, lower leaflets (4) alternate with those of upper strip (3). <hi rend="i">c</hi>, sinistral-bearing strips (1, 2) placed above dextral-bearing strips (3, 4); naturally crossed leaflets plaited in close check.</head>
                </figure>
              </p>
              <p>6. The twisted two-pair commencement used in Samoa (28, p. 185) and Aitutaki (27, p. 183) is not used in Rakahanga, but completes the series of commencement methods. In the twisted two-strip commencement, the twist is used with each single strip. In the twisted two-pair commencement, each pair is twisted by placing the two strips from the same side with their midrib strips together but the leaflets alternating. Commencing, the back leaflet of one strip is twisted forward under the two leaflets in front of it, one from each strip. The next leaflet from the other strip is then twisted forward under the two leaflets in front of it, one from each strip. The twisting goes on automatically by taking the nearest leaflet from each strip alternately and twisting it forward under the next two leaflets, one from each strip.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d5" type="section">
              <head>
                <hi rend="lsc">The Finishing Edge</hi>
              </head>
              <p>In mats, sheets, and platters the plaiting ends when sufficient depth has been secured. The commencing edge has been defined by the midrib strip in the various forms of commencement. The side edges have been formed by half-turns or direct bends. The far edge now ends in the free dextrals and sinistrals which project beyond the last horizontal row of check plaiting. The problem of disposing of the weft ends was met by plaiting them into a three-ply braid:</p>
              <p>The three-ply braid finish (<ref target="#BucMani109a">figs. 27</ref>, <ref target="#BucMani110a">28</ref>). The plaiting is turned so that the incomplete far edge lies longitudinally to the craftswoman. In Rakahanga, the plaited article is so placed that this far edge is on the right of the plaited portion. In this position, the dextral wefts are directed obliquely toward the plaiter, and the sinistral wefts are directed away from her. It is convenient to allude to the dextrals as near wefts and the sinistrals as far wefts. The side of the braid toward the plaiting (left) will be termed the inner side, and that on the other side (right), the outer. Thus, if the free
<pb xml:id="n117" n="109"/>
finishing edge were turned to the left, the terms inner and outer would apply equally well. The plies will also be termed inner and outer, according to the side from which they enter the braid. In plaiting a three-ply braid, the plaiting moves toward the plaiter, and there are two plies alternately on either side of the braid. To advance a step, the back ply crosses over the one in front of it to take up the middle position. The free weft ends are disposed of by adding them successively to a ply which has been brought into the middle position. The problem of the near and far wefts is settled by adding the near wefts to the inner plies and the far wefts to the outer plies.</p>
              <p>
                <figure xml:id="BucMani109a">
                  <graphic url="BucMani109a.jpg" mimeType="image/jpeg" xml:id="BucMani109a-g"/>
                  <head><hi rend="sc">Figure</hi> 27. Spaced three-ply braid finish of mats, sheets, and platters, <hi rend="i">a</hi>, plaiting placed with finishing edge longitudinal and free weft ends to right; dextral wefts (1–4) near and sinistral wefts (1′–4′) far; braiding commences at far end and three plies will be formed by first far weft (1′) and first two near wefts (1, 2). <hi rend="i">b</hi>, first far weft (1′) bent over crossing weft (1) to form first ply; second ply formed by second near weft (2) which is lifted over first ply (1′); third ply formed by turning in first near weft (1) over second ply (2); the three plies now established; two plies on inner side with weft (1′) as back ply. <hi rend="i">c</hi>, inner back ply (1′) brought over ply (1) in front of it to middle position; this shifts the two plies to outer side so back outer ply (2) crosses over to middle position, again changing pair position; back inner ply (1) crosses over to middle position; this, however, is second turn of an inner ply and technique demands near weft added to every second turn of an inner ply. <hi rend="i">d</hi>, next near weft (3) therefore added to inner ply (1) by lifting it up from outer side of last weft (2′) that crosses it and laying it on inner ply (1) in middle position, <hi rend="i">e</hi>, back outer ply (1′) brought over last double ply (1, 3) to middle position but as it is second turn of an outer ply, next far weft (2′) must be added to it; far weft (2′) under braid is pulled down to level of outer ply (1′) on outer side and will be turned over to join it in middle position, <hi rend="i">f</hi>, technique established; near wefts (3, 4) have been added to alternate turns of inner plies and far wefts (2′, 3′) to alternate turns of outer plies, <hi rend="i">g</hi>, established technique has been followed throughout; near wefts enter inner side of braid with every second or alternate turn of inner plies and leave spaces (6) between them, shaded alternate outer plies (7) show where far wefts have entered; on reaching near end, plies continued on in free tail (8) and knotted.</head>
                </figure>
              </p>
              <pb xml:id="n118" n="110"/>
              <p>Two varieties of technique have been developed. In one, the wefts are added to every turn of the plies. In the other, the wefts are added to alternate turns. These variations are influenced by the closeness of the wefts in the plaited article. Thus in the oven cover meant to retain the steam in the cooking oven, the plaiting is close and the wefts are added to every turn of the plies. In most of the other articles, the wefts are added to alternate plies. It is convenient, therefore, to allude to the two variations as the close three-ply braid and the spaced three-ply braid. Sometimes, owing to variations in the width of the wefts, both techniques (<hi rend="i">whiri</hi>) may be used in the one braid finish. The more common spaced three-ply braid is shown in <ref target="#BucMani109a">figure 27</ref>. The close three-ply braid finish is shown in <ref target="#BucMani110a">figure 28</ref>.</p>
              <p>
                <figure xml:id="BucMani110a">
                  <graphic url="BucMani110a.jpg" mimeType="image/jpeg" xml:id="BucMani110a-g"/>
                  <head><hi rend="sc">Figure</hi> 28. Close three-ply braid finish, <hi rend="i">a</hi>, commencement technique and figuring exactly the same as in <ref target="#BucMani109a">figure 27</ref>; when ply 1′ is turned in from inner side, near weft (3) added to it; outer ply (2) has been turned to middle position and far weft (2′) added to it; in next turn, back inner ply (1) will be turned into middle position but instead of waiting for another turn, next near weft (4) will be added to it; similarly, when next back outer ply (1′, 3) turned into middle position, next far weft (3′) will be added to it. <hi rend="i">b</hi>, near wefts (2–6) enter with successive inner plies; far wefts (2′–5′) enter with successive outer plies, there being no spacing.</head>
                </figure>
              </p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d6" type="section">
              <head>
                <hi rend="lsc">Sheets, Screens And Mats</hi>
              </head>
              <p>From the coconut leaf, thatch sheets, screens, oven covers, and two kinds of mats are made:</p>
              <p>1. Thatch sheet. Thatch sheets for cooking houses and temporary houses out in the plantations are made of split half leaves. The leaves are cut off in 6-foot or 7-foot lengths. Each section is split down the midrib and each half leaf plaited separately with the split midrib commencement in check with the open leaflet. In the left half leaf alternate leaflets are bent to the left to provide the crossing sinistrals, and the plaiting proceeds normally from left to right. (See <ref target="#BucMani106a">fig. 23</ref>.) In the right half leaf the natural direction of the leaflets is sinistral, so the alternate leaflets are bent to the right to provide crossing dextrals. The plaiting proceeds from right to left, the right hand manipulating a working set of sinistrals at the working edge and the left hand placing a dextral in the shed formed. The side margins are turned in with the upward half-turn. (See <ref target="#BucMani106a">fig. 23, <hi rend="i">c</hi></ref>.) After plaiting a few inches in depth with one working section, the ends of the leaflets are left free. The narrow band of plaiting insures that the leaflets are crossed to form a better thatch. The technique is the same as in Cook Islands, Tahiti, and Samoa.</p>
              <pb xml:id="n119" n="111"/>
              <p>2. The sitting mat (<hi rend="i">pora</hi>) is made by continuing the thatch sheet technique to a greater depth and then finishing off the far edge with a spaced three-ply braid. (See <ref target="#BucMani109a">fig. 27</ref>.) It resembles, moreover, the Aitutaki thatch sheet in the principle of joining two sheets together after commencing each separately. The split midrib commencement is used with the open leaflet in check, and the side margins are formed with an upward half-turn. The two separate sheets, after being plaited for a depth of about 1 foot, are placed one above the other. The wefts of the two sheets which coincide in position are then treated as double wefts and the plaiting in check is continued to reach the full depth, when the braid finish is used. The process of combining the two sheets by using double wefts is termed <hi rend="i">kohiti</hi>, a term also applied to check plaiting. (See <ref target="#BucManiP002a">p1. 2, <hi rend="i">A</hi></ref>.)</p>
              <p>
                <figure xml:id="BucMani111a">
                  <graphic url="BucMani111a.jpg" mimeType="image/jpeg" xml:id="BucMani111a-g"/>
                  <head><hi rend="sc">Figure</hi> 29. Commencement of screen (<hi rend="i">pataro whani</hi>), a, full midrib (<hi rend="i">whani</hi>) (9) with far leaflets (1, 2) and near leaflets (1′, 2′); far leaflets form natural dextrals; first near leaflet (1′) on left crossed over midrib and under its opposite leaflet (1) to form first sinistral weft, <hi rend="i">b</hi>, second near leaflet (2′) passed over midrib and also under first dextral (1) as technique is to be twill; two extra free leaflets (1″, 2″) introduced as dextrals to fill wide space between two natural dextrals (1, 2); extra leaflet (3″) also added as extra sinistral; first sinistral (1′) bent in to function as dextral and commence left edge of screen, <hi rend="i">c</hi>, third near leaflet (3′) has been crossed over midrib to act as sinistral and second sinistral (2′) has been turned in to define further the left edge. <hi rend="i">d</hi>, another free sinistral (4″) added; extra sinistral (3″) turned in to continue definition of left edge, <hi rend="i">e</hi>, fourth near leaflet (4′) has been added as sinistral and sinistral (3′) turned in to define further the left edge, <hi rend="i">f</hi>, development of twilled technique and use of additional wefts.</head>
                </figure>
              </p>
              <p>3. Wall screen (<hi rend="i">pataro whani</hi>). (See <ref target="#BucMani109a">fig. 29</ref>.) The term <hi rend="i">pataro</hi> is identical with that used in Tongareva and distinguishes coconut leaf sheets or mats used as screens for the sides of houses, doorways, and also for sitting mats. The leaf midrib (<hi rend="i">whani</hi>) is not split, and its retention in the unsplit form gives the qualifying term, <hi rend="i">whani</hi>, to the screen. The same mat was also used as a seat, and was then termed a <hi rend="i">pataro noho.</hi> The whole midrib commencement, whereby the naturally directed leaflets of one side form the dextrals, is used. The leaflets from the other side are crossed over the midrib to engage the natural dextrals as sinistrals. The leaflets are kept closed (<hi rend="i">piri</hi>), and as
<pb xml:id="n120" n="112"/>
this method increases the spaces between the leaflets, extra leaflets torn off another section of leaf are introduced as both dextrals and sinistrals to close the spaces as the plaiting proceeds. A twilled-two stroke is used and a decorative effect is introduced by changing the direction of the rows of plaiting. Thus in the type mat in <ref target="#BucManiP002a">plate 2, <hi rend="i">B</hi></ref>, the twilled-twos at the midrib commencement are in horizontal rows, which are changed to vertical rows and then to a horizontal row of twilled-four sinistrals, followed by a row of twilled-three dextrals and finally finished off in check. The change in the twill and in direction gives variety to the surface appearance of the mat and is introduced for the aesthetic effect. The side edge on the left is formed by upward half-turns and on the right by direct bends. The far edge is finished off with the spaced three-ply braid technique. (See <ref target="#BucMani109a">fig. 27</ref>.)</p>
              <p>4. The oven cover (<hi rend="i">pataro umu</hi>) carries the screen name, <hi rend="i">pataro</hi>, qualified by the word <hi rend="i">umu</hi> (oven) to denote its use as a cover placed over the food in the oven to keep in the heat during cooking. (See <ref target="#BucManiP002a">p1. 2, <hi rend="i">C</hi></ref>.) The technique departs from that of the screen <hi rend="i">pataro</hi> and is exactly similar to that of the Tongarevan oven cover (<hi rend="i">toto umu</hi>). The cover is made of two leaflet-bearing strips in the twisted two-strip commencement (<ref target="#BucMani107b">fig. 25</ref>). The closed leaflets are used with both check and twill, both side edges are formed by direct bends, and the far edge is finished off with the close three-ply braid technique. (See <ref target="#BucMani110a">fig. 28</ref>.) In addition to being used to cover (<hi rend="i">tapoki or puroku</hi>) food while it is cooking, the closely plaited mat is used for other purposes. The <hi rend="i">puraka</hi> may be grated on it (<hi rend="i">oro te puraka</hi>) and also the mature coconut (<hi rend="i">waru te hakari</hi>), so that the mat receives the grated material. It is also used as a mat on which grated coconut is exposed to the sun (<hi rend="i">tauraki ki te ra</hi>) to make the oil run, and for drying fish and <hi rend="i">Tridacna</hi> (<hi rend="i">tauraki te ika, te pahua</hi>).</p>
              <p>
                <figure xml:id="BucMani112a">
                  <graphic url="BucMani112a.jpg" mimeType="image/jpeg" xml:id="BucMani112a-g"/>
                  <head><hi rend="sc">Figure</hi> 30. Keel commencement of <hi rend="i">tapakau</hi> mat, <hi rend="i">a</hi>, three strips of coconut or lauhala commence three-ply braid; when short length established, wefts added on either side after back ply is twisted over into middle position; thus when ply (1) twisted over into position from right, weft (4) from bundle on right is added to it, its midrib strip (4′) being placed proximally to lie on ply; back ply (2) then twisted over it from left, and weft (5) from left bundle added to it with its midrib strip lying on ply; all closed wefts added with their own midrib edges distal to worker, <hi rend="i">b</hi>, wefts added alternately on either side; thus weft 13 has been added from left to ply 1, weft 14 from right to ply 2, and ply 3 twisted into middle position to await addition of weft from left; in close addition of wefts, braid does not show on upper surface; fibrous midrib strips of wefts continue in braid for their entire length, and as earlier ones give out later additions continue plies, <hi rend="i">c</hi>, required length attained; braid plies continued on as free tail (6), finished off with overhand knot (7); in type mat, length of weft-bearing part of braid is 45 inches and both commencing and finishing free tails, 2.5 inches, <hi rend="i">d</hi>, under surface, braid stands out as contrasted with upper surface <hi rend="i">c</hi>.</head>
                </figure>
              </p>
              <pb xml:id="n121" n="113"/>
              <p>5. The large sitting mat (<hi rend="i">tapakau</hi>) shows a marked departure from the technique ordinarily associated with coconut leaf material. (See <ref target="#BucMani112a">fig. 30</ref>.) It is evident that whereas the midrib strips offer certain advantages in providing a natural commencement edge, their presence also imposes restrictions. The disadvantages of the wide spacing between the wefts has been met in making some articles by using the two-strip twisted commencement further reinforced by the addition of extra individual wefts to fill up interweft spaces. There are, however, further limitations as to the length of the midrib and the alternation in length and width of leaflets as they proceed toward the tip and the butt. Furthermore, the depth of the plaiting is limited to the length of the leaflets on one side of the midrib. To meet these disadvantages and provide a larger and neater mat, an artificial midrib is made, of any length required and with selected leaflets on either side set closely together, a distinct advance in technique. The new principle involved consists of plaiting a three-ply braid and adding selected leaflets on either side to each ply of the braid as it twists into the middle position. The appropriate <hi rend="i">katuri</hi> leaflets from toward the tip ends of the leaves are pulled off individually so that the butt ends of the leaflets carry short strips of the leaf midrib. In the closed leaflet wall screens and oven covers, the leaflet wefts average 0.8 inch in width, but in the <hi rend="i">tapakau</hi> mat the edges of the leaflets are split off with the thumb nail to form closed leaflet wefts from 0.4 inch to 0.5 inch wide. The leaflets from the opposite sides of the leaf are kept distinct, those from the right side being placed in a bundle (<hi rend="i">ruruku</hi>) on the left side of the craftswoman and those from the left side of the leaf on her right. When plaiting the braid commencement, the worker automatically adds the leaflets on her right to the right side of the braid and those on her left to the corresponding left side of the braid. The women are methodical in their work and, by attention to minor details beforehand, they not only quicken the work but avoid confusion and bad craftsmanship. The commencement is here termed, from its appearance, the keel commencement (<ref target="#BucMani112a">fig. 30</ref>), to distinguish it from other braid commencements. Plaiting is commenced by laying the braided keel transversely in front of the plaiter. She sits on the near set of wefts and plaits the far wefts from left to right in twill. (See <ref target="#BucMani113a">fig. 31</ref>.)</p>
              <p>
                <figure xml:id="BucMani113a">
                  <graphic url="BucMani113a.jpg" mimeType="image/jpeg" xml:id="BucMani113a-g"/>
                  <head><hi rend="sc">Figure</hi> 31. Twill plaiting of <hi rend="i">tapakau</hi> sitting mat. <hi rend="i">a</hi>, with finishing braid tail (10) to left and thick braid beneath, plaiter sits on near wefts (11) and plaits wefts on far side of braid, commencing on left; all wefts directed toward right from method of fixing them in keel braid; first weft (1) in position as natural dextral and next weft (2) turned under it toward left as sinistral; next two wefts in position as dextrals and next one (3) twisted under them to left to form sinistral and comply with twilled-two stroke used at commencement; sinistral (3) passes over first dextral (1) and comes to left edge; regular technique affected by necessity of turning left edge of mat; method of turning weft affected by nature of outer weft edge; thus midrib edge of closed leaflet weft being on right, open free edges are on left or outer side; in Rakahangan technique, such an outer edge is not considered suitable to direct bend as it
<pb xml:id="n122" n="114"/>
would open out and fray edges; upward half-turn adopted; in making this turn, weft naturally crosses over next weft; thus sinistral 3 having crossed over dextral 1, weft (2) below it turned in over it with upward half-turn to define left edge; similarly, weft 3 turned over weft 4 above and so successively as sinistrals reach left edge; from weft 3, every alternate weft turned to left to form sinistral and passes under two dextrals to make horizontal row of dextral twilled-twos (5); sinistrals passed under two dextrals, lifted, and passed over two dextrals to form horizontal row of sinistral twilled-twos (6); exception to rule is sinistral 4, which, being near edge, has been left over three dextrals instead of two; such irregularities take place near edge, for not only must turned weft defining left edge be turned over sinistral above it, but it must pass under next sinistral to form better edge; above row of sinistral twilled-twos (6) dextrals emerge to form horizontal row of twilled-twos (7), succeeded by horizontal row of sinistral twilled-threes (8), then dextral twilled-threes (9). <hi rend="i">b</hi>, plaiting proceeds to right edge defined first by last dextral (1) turned in under dextral weft (2) above it; technique of turning in wefts on right edge again influenced by nature of outer weft edge formed here by leaflet midrib; wefts therefore turned in with direct bend without any half-turn, as leaflet midrib forms good outer edge, direct bends having been commenced by turning bent-in weft under crossing dextral above it; this course followed throughout; thus weft 2 passes under weft 3 above it and successively along right edge; stroke technique established on left carried out along full course of working section, but when plaiting reaches right edge, departures from regular stroke fit in with formation of right edge; continuation of horizontal rows of twill from left, commencing braid tail (10), and unplaited near wefts (11) shown; depth of plaiting from braid keel to last horizontal row of dextral twilled-threes (9), 4.5 inches.</head>
                </figure>
              </p>
              <p>From the last horizontal row of twilled-threes in <ref target="#BucMani113a">figure 31</ref>, the stroke technique is changed to vertical rows of twilled-twos for a depth of 12 inches. The stroke then changes to 5 horizontal rows of twilled-twos followed by about 10 rows of check. The dextral wefts in the check portion are narrowed by folding them over slightly at one edge. The depth of plaiting to the last check row is 22.5 inches. The far edge of the plaiting is then finished off with the spaced three-ply braid as in the preceding mats. (See <ref target="#BucMani109a">fig. 27</ref>.) The free tail end of the braid is knotted and pushed back through the plaiting below the braid edge with the knot to the under surface. This completes one half of the mat. The unplaited side is then turned so that the worker sits on the plaited part. The second half is plaited in exactly the same manner as the first. (See <ref target="#BucManiP003a">p1. 3, <hi rend="i">A</hi></ref>.) The <hi rend="i">tapakau</hi> is the best mat I have seen in coconut leaflet material. The long leaflets of the Rakahangan palms render the depth possible. Though the plaiting depth is 22.5 inches, the diagonal course of a single leaflet weft is 32 inches and at least an extra 2 inches are included in the braid.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d7" type="section">
              <head>
                <hi rend="lsc">Cooking Receptacles</hi>
              </head>
              <p>A group of plaited articles known geneically as <hi rend="i">raurau</hi> divides into two classes, cooking receptacles and serving platters. With the exception of one used for cooking fish, the cooking receptacles have been created by the need for containers in which to cook the grated <hi rend="i">puraka</hi> preparations. In Tongareva, where until recently the <hi rend="i">puraka</hi> was not present, the technique of the <hi rend="i">raurau</hi> cooking receptacles was not known. The <hi rend="i">raurau</hi> food platter upon which the cooked food is served to guests follows the technique of the cooking <hi rend="i">raurau</hi>, but the finishing edge is made neater by finishing it off with a three-ply braid. Three forms of cooking <hi rend="i">raurau</hi> and three forms of platters are made:</p>
              <pb xml:id="n123" n="115"/>
              <p>1. The fish receptacle (<hi rend="i">raurau to ika</hi>) is the roughest form of <hi rend="i">raurau.</hi> (See <ref target="#BucManiP003a">p1. 3, <hi rend="i">C</hi></ref>.) Two midrib strips from opposite sides of the leaf, each carrying 10 or 11 leaflets, are plaited together in check, the sinistral-bearing strip above the other. A leaflet at either end of the sinistral strip is left free for tying leaflets. The receptacle is about 12 to 13 inches wide. The side edges are turned in, and the plaiting proceeds until it is about 12 inches deep. The free ends of the dextrals are drawn together and tied in an overhand knot close to the plaiting, and the sinistrals are dealt with similarly. The fish are placed in the receptacle, which is hollow on the back surface of the leaflets, and the leaflet ends are simply tucked in around the fish. The plaiting is folded over the fish, the free tying leaflets are passed around parallel with the midrib strip, and the ends tied with a reef knot. The receptacle then serves as a cooking dish which is placed in the earth oven for cooking.</p>
              <p>2. The large <hi rend="i">puraka</hi> receptacle (<hi rend="i">raurau papa</hi>) is also made of two opposite strips bearing 11 or 12 leaflets. (See <ref target="#BucManiP003a">p1. 3, <hi rend="i">B</hi></ref>.) The sinistral-bearing strip is placed above the other, with the shiny surface of the leaflets upward. All the wefts are plaited in check with open leaflets. After the plaiting has reached a depth of about 5 inches, a leaflet is dropped out at either side edge for tying. The plaiting continues for a full depth of about 22 inches, the side edges being defined by turning in the wefts with a half-turn. The far finishing edge is fixed by tying the crossing wefts together with a reef knot in pairs, a sinistral and a dextral being included in each knot. For use, the <hi rend="i">raurau</hi> is turned over with the dull surface upward, as this surface is slightly hollow. The grated food (<hi rend="i">puraka oro</hi>) is laid on the midrib strip half of the sheet, the other end is folded over, and the two side leaflets that were left out of the plaiting are brought over to tie the receptacle together. The food in the receptacle is then placed in the earth oven for cooking.</p>
              <p>3. The smaller cooking receptacle (<hi rend="i">raurau kapukapu</hi>) is made in a similar manner but with shorter midrib strips carrying about 7 leaflets. (See <ref target="#BucManiP003a">p1. 3, <hi rend="i">D.</hi></ref>) The plaiting in check with open leaflets proceeds for a depth of 14 inches. The two leaflets which cross in the middle of the finishing edge are knotted together with a reef knot, which braces the plaiting edge together and prevents it from unraveling. The food cooked in the <hi rend="i">raurau kapukapu</hi> consists also of grated <hi rend="i">puraka</hi> when not so much quantity is required, and also of the preparation (<hi rend="i">pupu</hi>) of grated <hi rend="i">puraka</hi> mixed with coconut. The dull surface is again the hollow part, so the receptacle is turned with that surface uppermost. The food is placed on the midrib half and the other end doubled over. The free weft ends, which have been left long and brought around from either side, are passed around the sides above the projecting ends of the midrib strip, and tied at the back with a reef knot. The receptacle is neat and effective.</p>
              <p>4. The ordinary platter is made of two opposite midrib strips, each carrying 8 leaflets. (See <ref target="#BucManiP004a">p1. 4, <hi rend="i">A</hi></ref>, I.) With the shiny surface of the leaflets up, the sinistral-bearing strip is laid above the other. The stroke is a check with open leaflets, and the side edges are turned in with half-turns. The platter is about 12 inches wide at the midrib strip commencement. The plaiting continues for a depth of about 13 inches. The free ends of both sets of leaflets are then plaited in a single course three-ply braid, commencing on the left with the shiny surface upward. As in screens, the dextral wefts go straight into the braid from the plaiting edge, but the sinistrals pass under the braid and enter the braid in the outer plies. On reaching the right edge, the wefts are continued as a free braid tail for about 4 inches and tied in an overhand knot. The tail is then doubled in on the dull surface side, pushed through under a couple of wefts, and the knotted end is brought back again. The platter is hollow on the dull surface, which is turned uppermost in use.</p>
              <pb xml:id="n124" n="116"/>
              <p>5. A platter for serving fish (<hi rend="i">raurau rava ika</hi>) is more elaborate in that two sinistral-bearing and two dextral-bearing strips are used to supply a double quantity of wefts. (See <ref target="#BucMani116a">fig. 32</ref>; <ref target="#BucManiP004a">p1. 4, <hi rend="i">A</hi></ref>, 2). Owing to the extra material, the edges of the leaflets are split off to form narrower and neater wefts. The two-pair simple commencement shown in <ref target="#BucMani108a">figure 26</ref> is used. The side edges are turned in with half-turns, and the check stroke with open leaflets is used. Each midrib strip carries 7 leaflets, thus making 14 wefts in each series. The platter is a little more than 12 inches wide at the strip commencement. The plaiting continues for a depth of 15 inches where the plaiting narrows down to 9 inches in width. The finishing edge is completed by a two-course three-ply braid, shown in <ref target="#BucMani116a">figure 32</ref>.</p>
              <p>
                <figure xml:id="BucMani116a">
                  <graphic url="BucMani116a.jpg" mimeType="image/jpeg" xml:id="BucMani116a-g"/>
                  <head><hi rend="sc">Figure</hi> 32.—Fish platter, first course of two-course braid finish, <hi rend="i">a</hi>, left corner of finishing edge; sinistral weft (1′) has been bent at right angles over next sinistral (2′) to form first ply of three-ply braid, <hi rend="i">b</hi>, sinistral (2′) doubled over crossing ply (1′) to form second ply, and first dextral weft (1) lifted over second ply to form middle ply and so establish commencement of three-ply braid, <hi rend="i">c</hi>, back ply is outer ply (1′), crossed over middle ply (1) to take middle position, <hi rend="i">d</hi>, back ply is now inner ply (2′), crossed over middle ply (1′) and next dextral weft (2) added to it; technique now established; back ply comes alternately from outer and inner sides to take middle position by crossing over last ply placed in middle position; every time inner ply placed in middle position, next dextral weft automatically added to it. <hi rend="i">e</hi>, continuation of braidiing shows successive addition of dextrals (1–4) to inner plies; outer plies receive no additions as sinistrals and except for first (1′, 2′) are left out for second course; dextral wefts inclined in course of braiding; sinistrals directed in opposite direction; braiding continued to right end of finishing edge and thus uses up all dextral elements; plaiting then turned over and braid bent around in the opposite direction on reversed surface of plaiting; free sinistrals now run in same direction as plies of braid; braid thus simply continued in second course which uses up remaining sinistral wefts; when all sinistrals braided, free tail continued for short distance and fixed with overhand knot, excess of material being cut off; finishing edge thus consists of distinct braid on each surface.</head>
                </figure>
              </p>
              <p>6. A small, neat platter (<hi rend="i">raurau mereki: mereki</hi>, modern term for plate) was made of two opposite midrib strips each carrying 6 leaflets. (See <ref target="#BucManiP004a">p1. 4, <hi rend="i">A</hi></ref>, 3). In the type specimen examined, the dextral strip was placed above the other. The check stroke was used with the open leaflets and the side edges were turned in with half-turns. The platter at the commencement edge is 8 inches wide. The plaiting is continued for a depth of 8 inches. The 6 sinistral wefts are then braided in three-ply from the middle line toward the left. After all the six wefts have been included in the braid, the ends are continued as a free braid tail for a length of about 11 inches. The 6 dextral wefts are similarly treated on the right. The two tails are then brought back along their respective sides to the midrib strip, where they are passed through from the dull, concave side of the platter to the inner side of the marginal wefts. The two ends are drawn taut along the outer side of the strip margin and tied together in the middle line with a reef knot. This type of platter is probably modern but illustrates a form of technique seen in the <hi rend="i">pite</hi> group of plaited articles.</p>
              <pb xml:id="n125" n="117"/>
              <p>The <hi rend="i">raurau</hi> are characterized by the natural concavity formed on the dull side of the leaflets when plaited together or by folding. They are really plaited as sheets of material. The <hi rend="i">pite</hi> group consists of two types which commence with the <hi rend="i">raurau</hi> technique but are definitely converted into shaped receptacles by the introduction of corners or points into the plaiting. In both, the two-strip simple commencement is used, and the check stroke with open leaflets. Both utilize the <hi rend="i">raurau mereki</hi> method of disposing of the leaflets at the finishing edge by braiding the dextrals and sinistrals separately
<figure xml:id="BucMani117a"><graphic url="BucMani117a.jpg" mimeType="image/jpeg" xml:id="BucMani117a-g"/><head><hi rend="sc">Figure</hi> 33. Pointed <hi rend="i">pite</hi> receptacle, <hi rend="i">a</hi>, sinistral-bearing strip (8) laid over dextral-bearing strip (8′) with shiny surface of leaflets up; ends do not coincide but project at either end so as to allow for diagonal interlacing of leaflets; left dextral (1′) lifted and left sinistral (1) dropped under it; check stroke continued until seven wefts from both strips interlaced in check to form triangular piece of plaiting with apex upward; projecting free leaflets have now to be doubled back and plaited again in check to form double layer of plaiting, <hi rend="i">b</hi>, plaiting is turned over; plaiting commences on left margin by doubling over wefts 1′, 3′, 5′, and 7′, which have passed under left marginal weft (7); alternate wefts (2′, 4′ and 6′) above marginal crossing weft (7), left down; doubling over of set of alternating wefts forms shed in which crossing weft (7) laid by doubling it back over right marginal weft (1′), thus forming point of <hi rend="i">pite</hi> at apex of plaiting; seven wefts on left have now been divided into two alternating sets which will continue check stroke; in next movement three projecting wefts (2′, 4′, 6′) will be doubled over weft 7 in shed and four recumbent wefts (1′, 3′, 5′, 7′) raised, thus forming shed into which next crossing weft (6) will be placed by doubling it back over its previous course, <hi rend="i">c</hi>, second layer of plaiting completed by manipulation of two alternating sets of wefts on left until all wefts which projected on right margin included; in figure, completed plaiting projects below base of midrib strip, but in actual plaiting, leaflets drawn close together and do not project so far; set of wefts (7′–1′) on right now plaited in three-ply braid; thus left weft (7′) twisted over next weft (6′); third weft (5′) passes over first (7′) and second (6′) twisted over it to middle position; three-plies established and other wefts successively included by being added to ply which comes in from left; when all included, braid passed through under first part of right marginal weft (1′) and continued on as free tail (9); wefts 1–7 on left dealt with similarly and free tail passed through under first part of left marginal weft (7); opening of <hi rend="i">pite</hi> thus defined by midrib strip on one side and three-ply braid on other, <hi rend="i">d</hi>, plaiting turned over and two braid tails (9) brought together in middle line where tied in reef knot (10); leaflet strips about 7 inches long; <hi rend="i">pite</hi> when opened out, about 5 inches in diameter across rim and 6 inches deep.</head></figure>
<pb xml:id="n126" n="118"/>
and fastening the free braid tails through at the back of the midrib commencement edge.</p>
              <p>The two forms of <hi rend="i">pite</hi> consist of a pointed conical article and a four-cornered one:</p>
              <p>1. The pointed <hi rend="i">pite (pite pahua)</hi> is used by women when gathering <hi rend="i">Tridacna (pahua)</hi> for food. (See <ref target="#BucMani117a">fig. 33</ref>; p1. 4, <hi rend="i">A</hi>, 5.) The receptacle is hung with the open mouth upward on a stick in the sand. After shellfish have been collected, the shells are opened and the flesh put in the <hi rend="i">pite</hi>. This saves carrying the heavy shells back to the cooking houses.</p>
              <p>2. The four-cornered <hi rend="i">pite (pite pupu)</hi> is used for cooking the preparation (<hi rend="i">pupu</hi>) of grated <hi rend="i">puraka</hi> and coconut. (See <ref target="#BucMani118a">fig. 34</ref>; p1. 4, <hi rend="i">A, 4</hi>). The receptacle has four corners (<hi rend="i">e ha take</hi>). The technique of the corner is simple and is similar to that used in satchels in New Zealand. In any plaiting surface, if the normal lean of a weft is turned in the opposite direction and the crossing weft is similarly dealt with, a corner will result.</p>
              <p>
                <figure xml:id="BucMani118a">
                  <graphic url="BucMani118a.jpg" mimeType="image/jpeg" xml:id="BucMani118a-g"/>
                  <head><hi rend="sc">Figure</hi> 34. Four-cornered receptacle (<hi rend="i">pite pupu</hi>), technique of corners, <hi rend="i">a</hi>, simple two-strip commencement with dextral-bearing strip (7) placed over sinistral strip (8); each strip carries six leaflets kept spread open and interlaced in check, <hi rend="i">b</hi>, if lowest weft (1′) on left bent in, ordinary edge will result; to form corner next weft (2′) turned in at right angles and plaited in check with crossing sinistrals; lowest weft (1′) bent in parallel with 2′ to act as dextral and cross sinistrals in check; figure shown spread out on the flat, but in actual plaiting both turned-in wefts drawn taut to form definite corner where weft 2′ crosses weft 3′; weft 3′ and those above it (4′–6′) also forced to change direction and come down closer toward midrib strip; on right, similar plan pursued to form corresponding corner; second lowest dextral (5) turned in to function as sinistral and its crossing with dextral 4 forms point of second corner; lowest dextral (6) turned in and drawn taut to deepen corner; with all wefts crossing in check, second pair of corners may be formed in exactly same manner as first; thus lowest free weft on left is 3′ and corner key weft above it is 4′; similarly on right, lowest free weft is 4 and key weft above it is 3, <hi rend="i">c</hi>, on left, key weft (4′) turned in as dextral and forms corner with crossing weft 5′; weft (3′) below it follows suit and deepens third corner; both wefts (4′, 3′) cross sinistrals in check; on right, key weft (3) turned to act as sinistral and form fourth corner with crossing weft (2); weft 4 below it follows suit; with all wefts drawn taut, four corners distinctly defined and margin formed by far edge of plaiting curved around to more nearly approach midrib edge; four-cornered hollow receptacle formed; six free weft ends on either side plaited as free braid tails and brought around behind midrib edge in exactly same technique as in <ref target="#BucMani117a">figure 33, <hi rend="i">c, d</hi></ref>.</head>
                </figure>
              </p>
            </div>
            <pb xml:id="n127" n="119"/>
            <div xml:id="t1-body-d1-d4-d3-d2-d8" type="section">
              <head>
                <hi rend="lsc">Baskets</hi>
              </head>
              <p>The term basket (<hi rend="i">kete</hi>) is here used to include deeper and more permanent receptacles than the <hi rend="i">raurau</hi> and <hi rend="i">pite</hi> articles just described, and it includes articles usually described as satchels. Four types are made from coconut leaflets and two from lauhala. Of the four coconut leaf types, a round basket and a satchel clothes basket are local. The third type (<hi rend="i">to-ngini</hi>) was introduced during the post-missionary period, and the rough braided basket was also probably introduced.</p>
              <p>1. The round basket (<hi rend="i">kete</hi>). The round basket is the common receptacle for food and other objects, the common baskets of other areas made with open leaflets in check not having been adopted. (See <ref target="#BucMani119a">figs. 35</ref>, <ref target="#BucMani120a">36</ref>; <ref target="#BucManiP004a">pl. 4, <hi rend="i">B</hi></ref>, 2–3.) The basket is made of two midrib strips from opposite sides of the leaf. The two strips are dealt with by the twisted two-strip commencement (<hi rend="i">takaviri</hi>) and a twilled-two plait is used after the <hi rend="i">whatu</hi> technique and join (<ref target="#BucMani119a">fig. 35</ref>) are established. The horizontal row of dextral checks which completed the <hi rend="i">whatu</hi> is followed by a horizontal row of sinistral twilled-twos, another row of dextral checks, and then alternating rows of sinistral and dextral twilled-twos of two each. This forms a plaited cylinder about 6 inches deep.</p>
              <p>
                <figure xml:id="BucMani119a">
                  <graphic url="BucMani119a.jpg" mimeType="image/jpeg" xml:id="BucMani119a-g"/>
                  <head><hi rend="sc">Figure</hi> 35. The <hi rend="i">whatu</hi> technique, round basket, twisted two-strip commencement: most twisted strips, about 50 inches long; sinistral-bearing strip (5) placed above dextral-bearing strip (6); <hi rend="i">whatu</hi> technique consists of joining strip together with check stroke; to establish this, right hand holds up number of sinistrals on left and left hand holds up some dextrals; second sinistral (2) dropped and first dextral (1′) dropped over it; right hand drops third sinistral (3) and left hand drops second dextral (2′) over it; from this commencement, each hand alternately drops next weft on right; horizontal row of dextral checks (4) produced; this continued to end of strip; two ends of double strip brought together and joined; if any interweft space too wide, extra single weft (<hi rend="i">huru</hi>) interpolated and its butt end rests between the two midrib strips; in preparing midrib strips, each weft twisted forward in front of <hi rend="i">huru</hi> and so successively fixed; end leaflet remains free, having no leaflet in front of it under which it may be twisted; similarly, first leaflet to be twisted forward has no leaflet passing under it; thus on sinistral strip, end leaflet (1) on left free; first twisted leaflet (11) on right has no leaflet passing under its loop between it and midrib strip (5); when the two ends brought together, free leaflet (1) passed up under loop of leaflet 11 in place indicated by arrow; in this way, twisting of strip completed by joining it into circle; in lower dextral strip (6), free leaflet (11′) on right, and free loop formed by first dextral (1′) on left; with ends together, free leaflet (11′) threaded up through loop of dextral 1′ in place indicated by arrow; this completes join and crossing leaflets plaited in check to complete <hi rend="i">whatu</hi> join.</head>
                </figure>
              </p>
              <pb xml:id="n128" n="120"/>
              <p>The bottom of the basket is formed by decreasing the diameter with each circle of weft strokes. Thus from the last circle of twilled-twos with single wefts, the next circle of twilled-twos is formed by drawing two adjacent wefts together to form a double weft with both dextrals and sinistrals. This rounds the bottom and narrows (<hi rend="i">potiki</hi>) the circle of plaiting. After this round, the narrowing is repeated (<hi rend="i">ka potiki whakohou</hi>) by bringing two double wefts together to form compound wefts containing four leaflets in each weft. The four-element wefts are now plaited in check for three rounds. The three rounds of check plaiting form a raised cuff. The cuff is then flattened longitudinally, taking the join (<hi rend="i">tutaki</hi>) on the rim as the far end. From the far end of the unclosed slit bottom, the plaiter commences a three-ply braid (<hi rend="i">whiri</hi>) with the wefts directed toward her on either side. When she gets to the near end and has incorporated all the wefts directed toward her in the braid, she continues the braid in a free, thick, braided tail (<hi rend="i">take</hi>) and fixes the end with an overhand knot. The free tail is about 11 inches long. The ends are then reversed, and the remaining wefts on either side, which are now directed toward the plaiter, are braided in a similar way. The bottom is thus closed by a two-course braid, each course being completed by its own free tail. (See <ref target="#BucMani120a">fig. 36</ref>.) Each tail is pushed through over two crossing wefts into the inside of the basket. The cuff is stretched and flattened out and the free tails pushed through to the outside at the ends of the bottom. The tails are turned upward and after a short course of 3.5 inches on the outside the knotted ends are pushed through to the inside. (See <ref target="#BucManiP004a">pl. 4, <hi rend="i">B</hi></ref>, 2–3.) Owing to the longitudinal course of the thick braid tails on the bottom, the basket described is more elliptical than circular at the bottom. In the type basket, the wefts were 0.8 inch wide, but in a smaller, neater basket the wefts were 0.4 inch wide. In the neater basket, the narrower wefts were rendered possible by introducing a larger number of the extra wefts (<hi rend="i">huru</hi>).</p>
              <p>
                <figure xml:id="BucMani120a">
                  <graphic url="BucMani120a.jpg" mimeType="image/jpeg" xml:id="BucMani120a-g"/>
                  <head><hi rend="sc">Figure</hi> 36. Round basket, bottom closure: rounded body (1) of basket with midrib strip rim (2) shown with cuff-like bottom (3) formed of three rows of check with four-element wefts; bottom closed by bringing sides together; first braid (4) formed by plaiting from right to left with wefts directed toward left; remaining wefts directed toward right, plaited from left to right to form second braid (5); both braid tails finished off with overhand knots (6).</head>
                </figure>
              </p>
              <p>2. Clothes basket (<hi rend="i">kete ngahengahe</hi>). The basket is made in the form of a satchel. (See figs. 37–39; <ref target="#BucManiP005a">pl. 5, <hi rend="i">A</hi></ref>, 1.) It shows a marked advance in technical detail. The basket is distinguished by the braided rim commencement, the use of changing twill, and the advanced type of braid finish at the bottom. In addition, the rim is wrapped with colored lauhala and the satchel decorated with wrapped strips of aërial hala roots, tufts of colored fiber, and pearl shell ornaments. The leaflets of young coconut leaves are split with the thumb nail (<hi rend="i">tete</hi>) and passed over the fire (<hi rend="i">parara</hi>) to soften them. They are then dried in the sun (<hi rend="i">haumaki ki te ra</hi>). The leaflets are split from the leaflet midrib, and any thick strips from the vicinity of the midrib are carefully removed so that the strips are of even thickness. A number of individual wefts are thus prepared, which are cut off square toward the butt ends and average about 0.2 inch in width. The commencement at the rim is a unique feature,
<pb xml:id="n129" n="121"/>
<figure xml:id="BucMani121a"><graphic url="BucMani121a.jpg" mimeType="image/jpeg" xml:id="BucMani121a-g"/><head><hi rend="sc">Figure</hi> 37. Braid commencement of rim of <hi rend="i">kete ngahengahe. a</hi>, three strips of leaflet (1–3) commence three-ply braid; back ply (1) on left crossed over into middle position. <hi rend="i">b</hi>, two wefts (4, 4′) added together to middle ply (1) which came in from left; cut-off butt ends rest on turn of ply and long length of wefts joins long length of ply; care taken in adding wefts to keep shiny surface of wefts uppermost; though leaflet midribs with thicker parts on either side split off, edge of weft toward midrib still shows different consistency to edge away from it; to obtain neat plaiting, midrib edges of each series of wefts must all face same direction; of two wefts added to braid, under one will subsequently function as dextral and upper as sinistral in all pairs added to braid; in adding pairs to braid, craftswoman selects wefts with midrib edge to right for lower element (4) and wefts with midrib edge to left for upper element (4′); detail indicates care taken to enhance effect, <hi rend="i">c</hi>, two wefts (4, 4′) being in position, back ply (3) on the right crossed over them to middle position; back ply on left (2) crossed over 3 to take middle position. <hi rend="i">d</hi>, all additional pairs of wefts added to ply which comes in from left as in <hi rend="i">b</hi>; two wefts (5, 5′) therefore added to middle ply from left (2). <hi rend="i">e</hi>, two plies being now on right, back ply (1) with its two added wefts (4, 4′) brought over to middle with half-twist to round braid; position of wefts now reversed, weft 4 being above 4′ and ply 1 above 4. <hi rend="i">f</hi>, back ply (3) brought over from left to middle position and two wefts (6, 6′) added to it. <hi rend="i">g</hi>, back ply (2) on right brought over to middle position bringing two wefts (5, 5′) with it; back ply (1) on left brought over to middle position, but two wefts (4, 4′) associated with it, now sufficiently fixed by their course in braid, left out as shown; fixation technique established; two more wefts will be added to middle from left (1), and back ply on right (3) will be brought over to middle position with its accompanying wefts (6, 6′); back ply (2) will then be brought over to middle position from left, leaving two wefts (5′, 5) fixed on left; two new wefts will be added to middle ply (2) to take place of those left out; technique continued until required length reached; thus each ply contains original ply element with two added wefts; in every movement from left, two fixed wefts left out, original ply element brought over to middle position, and two new wefts added to it. <hi rend="i">h</hi>, section of braid with paired weft added.</head></figure>
<pb xml:id="n130" n="122"/>
for in the preparation of individual separate wefts the ordinary basket commencement by utilizing a midrib strip carrying the leaflets has been abandoned. The individual wefts are therefore connected by means of a three-ply braid as shown in <ref target="#BucMani121a">figure 37</ref>. The three-ply braid is continued for twice the length of the basket. The material is then turned over with the braid toward the worker, and the shiny surface of the wefts, which were reversed by the half-twists in the braid, is again uppermost. This surface will form the outer side of the satchel, and the thick part of the braid will also be reversed to the inside of the rim. The paired wefts are now all directed toward the right. The upper elements of the pairs function as naturally directed dextrals, and the lower elements are turned to the left to function as sinistrals. Plaiting commences with a horizontal twilled-two. (See <ref target="#BucMani122a">fig. 38</ref>.) The body of the basket, after the rim
<figure xml:id="BucMani122a"><graphic url="BucMani122a.jpg" mimeType="image/jpeg" xml:id="BucMani122a-g"/><head><hi rend="sc">Figure</hi> 38. Twilled plaiting of <hi rend="i">kete</hi> ngahengahe. <hi rend="i">a</hi>, braided rim commencement turned over in position for plaiting with all wefts directed toward right; upper elements (4′) form dextrals and lower elements (4) will supply sinistrals. <hi rend="i">b</hi>, technical twilled-two stroke; commencing on left, number of upper laying elements raised with right hand; lower element on left (4) turned to left with left hand, and next lower element to right (5) also turned to left; left upper weft (4′) dropped over the required two, and left down; next lower element (6) turned to left and next upper element (5′) dropped over it from right hand and left down; similarly, upper weft (6′) dropped over lower weft (7) and so plaiting moves to right by successively turning lower weft to left and dropping upper weft over it; once commencement made, twilled-two proceeds automatically. <hi rend="i">c</hi>, when a sufficient depth has been secured, two pairs of dextrals raised with intervening pair left down; thus in last movement, two pairs (1′, 2′, 5′, 6′) were kept raised and intervening pair (3′, 4′) left down; sinistral weft (6) was then bent into shed prepared; in next movement in order to continue horizontal row of twill, upper or far numbers of raised pairs (1′, 5′) will be dropped and recumbent wefts below or on near side of raised wefts (3′, 7′) raised to complete raised pairs; the next sinistral (7) will be bent into shed and twilled-two technique thus continued; result as shown on surface is two horizontal rows of twilled-two dextrals and one intervening row of twilled-two sinistrals; horizontal twilled-twos continued for full length of braid; ends of braid then brought together and free wefts from each end plaited in same twill so as to close space (<hi rend="i">tutaki</hi>).</head></figure>
fixation by 3 horizontal rows of twilled-twos, was changed in the type basket to vertical rows of twill for about 2.4 inches in depth. The plait was then changed to 3 rows of horizontal twill for another 2.4 inches. This was succeeded by 3 horizontal rows of check, preparatory to closing the bottom of the satchel. The bottom was closed by a two-course braid finish (<hi rend="i">hiri teru</hi>) in which the end of the first course was turned over to form the commencement of the second. The second course was finished off with a knot close to the plaiting and not pushed through to the inside. During the second course, the wefts, after a few turns in the braid, were successively dropped out on either side and then cut off close to the braid turns. This made a neater finish than the ordinary two-course braid in which the wefts are kept in the braid for their full length. The technique has been described for the baskets of Aitutaki (27, p. 186).
<pb xml:id="n131" n="123"/>
The rim (<hi rend="i">kaungutu</hi>) was finished off and embellished by a strip of the thin <hi rend="i">papa</hi> material obtained by splitting off the upper surface of the lauhala. It was dyed red and attached as shown in <ref target="#BucMani123a">figure 39</ref>. Ornamentation was provided by the use of splints, fringes, and pearl shell. Ornamental splints (<hi rend="i">weri</hi>) were made of strips of aërial hala rootlets about 0.3 inch wide and 0.05 inch thick. These were wrapped spirally with a strip of dyed <hi rend="i">papa</hi> material. One strip passed along under the bottom, was bent up sharply at each bottom corner, and ascended at the ends to meet the rim where the ends were cut off level. A thin strip of material passed through the basket and around the splint fastened them at the rim and at each lower corner. A horizontal splint was then passed around the basket about halfway down. The ends overlapped and were tied together. They were also tied to the vertical limbs of the first splint at the ends. Two other vertical splints were passed from rim to rim under the bottom, dividing each side into three approximately equal parts. Short fringes of bast fiber, dyed red, were attached at the rim to the upper ends of the vertical splints and to the crossings with the horizontal middle splint. Pearl shell pendants (<hi rend="i">tau parau</hi>) triangular in shape but with the short base curved were attached to the fringes in sets of three by threads passing through two holes bored in each pendant. The pendants average 0.6 inch by 0.4 inch. The satchels are neat and attractive. The one figured (<ref target="#BucManiP005a">pl. 5, <hi rend="i">A</hi></ref>, 1) is small and was made as a sample by Tina, the only woman on Rakahanga who knew how to make them. The baskets were made much longer and were used to contain garments (<hi rend="i">ngahengahe</hi>), hence the name, <hi rend="i">kete ngahengahe.</hi> Traditional history states that a woman named Utua-vaine once avoided the regular migration from Te Kainga to Manihiki by slipping away from the canoes to get her <hi rend="i">kete ngahengahe.</hi> This was adduced as proof of the age of the basket technique. The wrapping with dyed lauhala and the colored fringes are probably of more recent origin.</p>
              <p>
                <figure xml:id="BucMani123a">
                  <graphic url="BucMani123a.jpg" mimeType="image/jpeg" xml:id="BucMani123a-g"/>
                  <head><hi rend="sc">Figure</hi> 39. Rim wrapping of <hi rend="i">kete ngahengahe</hi>: strip of <hi rend="i">papa</hi> material (1) 0.8 inch wide folded longitudinally over rim to conceal rim braid (2) on both sides; length of thin two-ply twisted sennit cord (3) knotted at one end, passed through hole made through rim just above two lower edges of <hi rend="i">papa</hi> strip, knot to inside; a series of half-hitches made through similar holes about 0.3 inch apart, crossing (4) of half-hitches being on outer side of rim; wrapping worked to left to complete circuit of rim.</head>
                </figure>
              </p>
              <p>4. Foreign basket (<hi rend="i">tongini</hi>). This type of basket has been described by Handy (15, pp. 21–35) for Tahiti, and for Cook Islands by Hiroa (27, p. 168), and the Rakahangan technique differs in no way from that described. In Rakahanga the basket is made with four interlocking pairs of midrib strips, each pair carrying 4 and 3 leaflets respectively. (See <ref target="#BucManiP004a">pl. 4, <hi rend="i">B</hi></ref>, 1.) The handles are plaited in flat bands, and the leaflets that pass around on the outside are plaited in three-ply braid and knotted together under the basket bottom. The baskets are large for such baskets. The basket is frankly acknowledged not to be native to Rakahanga. In Tahiti it is called <hi rend="i">oini</hi> and in Aitutake, <hi rend="i">ohini.</hi> The Rakahangan name, <hi rend="i">tongini</hi>, is evidently a local adaptation of the other name.</p>
              <p>5. Full-leaf braided basket (<ref target="#BucMani124a">fig. 40</ref>). This quickly made receptacle is made from a section of coconut leaf midrib with the leaflets on either side intact. It has been described by Handy (15, p. 62, fig. 27) for Tahiti. I could not determine whether it was introduced into Rakahanga from that area with the <hi rend="i">tongini</hi>, but the identity of technique favors the probability of diffusion in the post-missionary period.</p>
              <pb xml:id="n132" n="124"/>
              <p>
                <figure xml:id="BucMani124a">
                  <graphic url="BucMani124a.jpg" mimeType="image/jpeg" xml:id="BucMani124a-g"/>
                  <head><hi rend="sc">Figure</hi> 40. Braid technique of full-leaf braided basket. <hi rend="i">a</hi>, section in middle defined by counting off about five opposite pairs of leaflets and cutting nicks (6, 7) on under surface of midrib to outer side of end leaflets (1, 5); the two ends bent up at right angles to middle section; seven opposite pairs of leaflets counted on either vertical limb and ends cut off; horizontal middle section forms bottom of basket; commencing with leaflets on one side of midrib, three-ply braid started at bottom; middle leaflet (3) forms first ply, left leaflet (2) brought over it to form second ply, and leaflet on right (4) brought over to middle position to form the third ply; the back ply (3) brought over to the middle position from left ready to have weft from left added to it. <hi rend="i">b</hi>, from last position, leaflet 1 added to middle ply; back ply on right (2) crossed to middle position from right and next leaflet on right (5) added to it; this includes last leaflet from bottom section; from now on, with each turn of braid, leaflet added alternately from either side commencing with lowest leaflet (1′) on left and then lowest leaflet (1″) on right; when all side leaflets have been added to braid, ends continued on as free braid tail (8) and knotted. Basket turned over and leaflets on other side braided in similar manner; the two braid tails knotted together to form handle for basket.</head>
                </figure>
              </p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d2-d9" type="section">
              <head>
                <hi rend="lsc">Eye Shades</hi>
              </head>
              <p>Plaited shades for the eyes (<hi rend="i">taumata</hi>) are very quickly and roughly made, as they are not permanent articles but are discarded after use. Two forms were seen. The simplest form was made from one midrib strip carrying 7 leaflets. (See <ref target="#BucMani125a">fig. 41</ref>.) The other is made with the simple two-strip commencement, each strip carrying four leaflets. The check stroke with the open leaflets is used. The technique is similar to that of the first shade in commencement, but the plaited band is made longer. Thus in the simple eye shade only two half-turns are made on the short side of the shade (<ref target="#BucMani125a">fig. 41, <hi rend="i">b</hi></ref>), but in the second form six half-turns are made. The end edge (<ref target="#BucMani125a">fig. 41, <hi rend="i">c</hi></ref>) is formed in the same way, but instead of finishing with a three-ply braid, a narrow band is plaited by closing the leaflets and thus narrowing the wefts. The ends of the leaflets are knotted and form one tie, and the other tie is provided by a long continuation of the midrib strip. (See <ref target="#BucManiP004a">pl. 4, <hi rend="i">A</hi></ref>, 6.) These shades were used in fishing to protect the eyes from the glare of the sun, but plaited hats with brims are taking their place.</p>
              <pb xml:id="n133" n="125"/>
              <p>
                <figure xml:id="BucMani125a">
                  <graphic url="BucMani125a.jpg" mimeType="image/jpeg" xml:id="BucMani125a-g"/>
                  <head><hi rend="sc">Figure</hi> 41. Simple eye shade (<hi rend="i">taumata</hi>). <hi rend="i">a</hi>, midrib strip split off from right side of leaf and section bearing seven leaflets cut off; four leaflets (1–4) on right function as natural sinistrals; strip bent up on left and three leaflets (5–7) bent in as dextrals; the four sinistrals and three dextrals interlaced in check with open leaflet, <hi rend="i">b</hi>, plaiting proceeds by turning lowest sinistral (4) on left with upward half-turn to cross sinistral 3 above it in check; left edge formed in same way by successive half-turns; on right, edge formed by turning lowest dextral (5) with upward half-turn to cross dextral (6) above it and function as sinistral; next dextral (6) turned and plaiting continued until other six wefts have crossed sinistral (6) which is key weft for change in position, <hi rend="i">c</hi>, plaiting moved so that what was left edge is toward plaiter; key weft (6) is bent upward over crossing weft (5) and returns parallel to its previous course; other six wefts now sinistrals; edge formed on left by turning up lowest sinistral (5) with upward half-turn to cross sinistral 1 above it and function as dextral; other sinistrals (1–4) successively turned up with upward half-turn and function as dextrals; lowest dextral on right (6) takes upward half-turn and crosses the others as last sinistral, when it is left projecting at left edge; sinistral 7 takes half-turn across top sinistral (6); again, six dextral wefts; these plaited in three-ply braid, first ply being formed by middle wefts (2, 3); two lowest wefts (5, 1) form second ply which crosses first while remaining two upper wefts (4, 7) form third ply which crosses second ply; three plies plaited in free tail (8) and knotted; end of free weft (6) on right simply knotted to crossing weft (7); coconut leaflet (9) knotted around first leaflet (1) on right; leaflet 9 and braid tail (8) used as strings to tie at back of head to keep shade in position.</head>
                </figure>
              </p>
            </div>
          </div>
          <div xml:id="t1-body-d1-d4-d3-d3" type="section">
            <head>
              <hi rend="sc">Plaiting with Lauhala</hi>
            </head>
            <div xml:id="t1-body-d1-d4-d3-d3-d1" type="section">
              <head>
                <hi rend="lsc">Material</hi>
              </head>
              <p>The hala (<hi rend="i">Pandanus</hi>) grows on all the islands without cultivation. Though <hi rend="i">whara</hi> is the general name, that term is especially applied to the mature plant which has developed a long trunk. The young plant which has no length of trunk is distinguished as <hi rend="i">puwhara</hi>; and it is from this plant that mats and clothing are made, whereas the leaves of the <hi rend="i">whara</hi> serve for roof thatch. The preparation, according to my informants, is as follows:</p>
              <p>The <hi rend="i">puwhara</hi> is cut down and the leaves cut off. The leaves are individually passed over the fire (<hi rend="i">parara ki te ahi</hi>) and then exposed to the sun (<hi rend="i">haumtaki ki te ra:haumaki</hi>, expose to the sun; Cook Islands, <hi rend="i">tauraki</hi>) for two days. The spines on the
<pb xml:id="n134" n="126"/>
back of the leaf midrib are removed by splitting off with the teeth the strip carrying the spines. A pearl shell implement is used to smooth out the leaves. If the shell is large it is called <hi rend="i">parau</hi> but if small, <hi rend="i">kati.</hi> The butt end of the leaf is held with the left hand and the shell is run along the under side of the leaf. This process (<hi rend="i">vautua</hi>) spreads out the wrinkles caused by drying and smooths and softens the material. The leaves are again exposed to the sun for three or four days.</p>
              <p>Each leaf is then wound around the hand, commencing with the butt end. This opens out the leaf. A reverse winding is then made on the other hand, which leaves the butt end on the outside. The side edges (<hi rend="i">papita</hi>) of the leaf are split off (<hi rend="i">tete</hi>) with a <hi rend="i">kuku</hi> shell. Other leaves are added to form a large roll (<hi rend="i">pipiti</hi>). The process of rolling is termed <hi rend="i">tupe.</hi> The end of the added leaf is always placed under the end of the preceding leaf. The <hi rend="i">pipiti</hi> rolls are fastened with a strip of lauhala passed through the central hole and tied around one radius of the roll. These are stored for use.</p>
              <p>In preparing the material for plaiting, the roll is first beaten on the ground (<hi rend="i">ka tuki te pipiti</hi>). The leaves are again scraped (<hi rend="i">haro</hi>) along the back with a shell to soften the material (<hi rend="i">kia paruparu</hi>). The leaf is split off on either side of the midrib, which is discarded. The worker then deals with the half leaves. The butt end of the leaf is the <hi rend="i">pu</hi> and the tip end, the <hi rend="i">hiku.</hi> The half leaf is split (<hi rend="i">tete</hi>) into wefts of the desired width. Mats with wide wefts are termed <hi rend="i">kiri maraia</hi> and those with narrow wefts, <hi rend="i">moenga kuti.</hi></p>
              <p>In plaiting mats, double wefts (<hi rend="i">tuarua</hi>) are used. To form a double weft, the dull surfaces of the wefts are placed in opposition so that the shiny (<hi rend="i">tua</hi>) surface will be exposed on both sides of the double weft.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d3-d2" type="section">
              <head>
                <hi rend="lsc">Mats</hi>
              </head>
              <p>The mats made from lauhala are used for sleeping upon and are hence termed <hi rend="i">moenga</hi> (<hi rend="i">moe</hi>, to sleep). The <hi rend="i">tua</hi> surface of the lauhala of the warmer atolls seems to have a better appearance than the material used in the Cook Islands. The women of Manihiki and Rakahanga are regarded as good plaiters. The sleeping mats made of double wefts and embellished with colored designs in overlaid plaiting are much admired and sought after. The commencement technique is as follows:</p>
              <p>As applied with double wefts, the commencement technique (<hi rend="i">whatu</hi>) does not differ from that of the Samoan <hi rend="i">papa</hi> mats (28, pp. 214–216) which may be compared for technical details. (See <ref target="#BucManiP006a">pl. 6, <hi rend="i">C.</hi></ref>)</p>
              <p>The plaiting stroke throughout is the check. Colored designs are worked in after the completion of the mat with overlaid material formed of the upper layer (<hi rend="i">papa</hi>) of the lauhala separated in thin sheets after heating the leaf. The material is now dyed with trade dyes. It is split to the same width as the mat wefts, and geometrical designs are worked by filling in a series of weft squares. Dextral strips to cover the dextral squares are pushed under the crossing sinistrals. The sinistrals are then covered by sinistral strips which are pushed under the crossing dextrals. Large designs are worked and give a pleasing effect.</p>
              <p>The length of a mat is <hi rend="i">roa</hi> and the width, <hi rend="i">kakano.</hi> The edges are termed <hi rend="i">tamore.</hi></p>
            </div>
            <pb xml:id="n135" n="127"/>
            <div xml:id="t1-body-d1-d4-d3-d3-d3" type="section">
              <head>
                <hi rend="lsc">Baskets</hi>
              </head>
              <p>Baskets are now extensively made from lauhala, but little doubt can exist that this form of the craft has come in by diffusion from Cook Islands in the post-missionary period. However, the baskets warrant a brief description, for, though comparatively recent, they are nevertheless essentially Polynesian in technique and have been adopted and varied in the local culture. The lauhala baskets are of two types, the two-cornered and the four-cornered satchels:</p>
              <p>1. The two-cornered satchel (<ref target="#BucMani127a">fig. 42</ref>; <ref target="#BucManiP005a">pl. 5, <hi rend="i">A</hi></ref>, 2). This basket conforms in all details except ornamentation to the baskets made in Cook Islands. The wefts are single strips 0.15 inch to 0.2 inch wide. The check stroke is used throughout. The basket is plaited like a strip of mat for twice the required length. The ends are brought together and the triangular gap filled in by plaiting the free marginal wefts of each end. This forms a plaited cuff, or cylinder, which is plaited in sections to the required depth. The cuff is turned inside out and the sides brought together to compress the bottom opening into a longitudinal line.</p>
              <p>The bottom is closed (<ref target="#BucMani127a">fig. 42</ref>) along the longitudinal line in a simple manner by check plaiting in two courses, which is really the beginning technique of a braid which is not developed by twisting in the plies. The dextral and sinistral wefts on either side of the median line run in opposite directions. In the first course the set of wefts on either side that inclines toward the plaiter is used, and in the second course both sets are finally disposed of.</p>
              <p>
                <figure xml:id="BucMani127a">
                  <graphic url="BucMani127a.jpg" mimeType="image/jpeg" xml:id="BucMani127a-g"/>
                  <head><hi rend="sc">Figure</hi> 42. Two-cornered lauhala satchel, bottom closure, <hi rend="i">a</hi>, to avoid confusion, it is convenient to apply term “inward” to wefts on either side directed toward the plaiter and “outward” to those directed away from her; plaiting commences on far end of longitudinal line forming unclosed bottom; in making start, certain wefts (10) do not readily fit into scheme of check crossing and are left out to be disposed of later; selecting inward weft (1) on right, this is crossed diagonally over middle line in its natural course; on left, inward weft meets outward weft (1′) which is bent back over its crossing weft (4) and lies paired with inward weft (1) above it; next inward weft (2) taken from left and crossed over previous inward weft (1) in check stroke; meets outward weft (2′) on right, which, in turn, is bent back to form paired weft; right inward weft (3) crosses over in check and turns back outward weft (3′); similarly inward wefts (4, 5, 6) crossed in check from alternate sides and each pair with outward wefts (4′, 5′,
<pb xml:id="n136" n="128"/>
6′) on opposite sides which are turned back under them. <hi rend="i">b</hi>, right inward weft (5) raised; complementary outward weft (5′) doubled back over its crossing weft (7) and ready for inward weft (5) to be crossed over weft 4 in check technique; next inward weft (6) raised while its complementary outward weft (6′) bent back ready to receive inward weft (6) on its upper surface when next check stroke is made; thus inward wefts form check crossings of single elements to close longitudinal open slit of bottom, but when they cross, they turn back corresponding outward weft and form double wefts on far side of middle line; this continued for length of bottom; first course thus closes opening with crossed single wefts, but technique arranges all wefts into double wefts directed toward plaiter on both sides of middle line, <hi rend="i">c</hi>, bottom reversed so double wefts on either side of middle line of single checks directed away from plaiter. <hi rend="i">d</hi>, second course commences at far end by doubling first two double wefts (1, 2) down middle line, <hi rend="i">e</hi>, double wefts plaited in check alternately from either side, but as they consist of double elements, it is better to deal with each individually; double weft on left selected as first to cross; its upper element (3) doubled back across middle line thus exposing under element (3′) which was originally a doubled-back outward weft. <hi rend="i">f</hi>, doubled-back under element (3′) straightened out over its pair mate (3). <hi rend="i">g</hi>, double weft (4,4′) on right crossed over in two similar movements, <hi rend="i">h</hi>, technique established and double wefts (5, 6, 7, 8, 9) successively crossed from alternate sides; under elements of wefts pulled taut so as to make check technique close and firm; double weft (9) last to be brought over from left to cross preceding weft (8); next weft (10) from the right has been covered by weft 7, which has to be lifted to allow weft 10 to be turned back over weft 9 in middle line; so throughout plaiting. <hi rend="i">i</hi>, last two double wefts (1, 2) on near end of mesial line crossed and continued with loose elements left out of commencement (<hi rend="i">a</hi>, 10) shown as 3, continued as free braid tail (4), and finished off with overhand knot; in second course, check stroke made with double wefts over middle line; weft ends, if too long, cut to form short fringe on either side as shown.</head>
                </figure>
              </p>
              <p>The rim is finished off with a last working section to form serrations after the style of the pointed <hi rend="i">pite</hi> (<ref target="#BucMani117a">fig. 33</ref>). Below the serrations the wefts are carried on in a plaited band to fix the weft ends. The technique is exactly the same as in the lauhala baskets of Aitutaki (27, p. 197). The basket is turned again and the bottom closure with the finishing rim cuff is concealed within. A handle of plaited lauhala is sewn on by the ends to the middle of the inner sides of the rim. Various designs in overlaid plaiting are made on the sides with the thin <hi rend="i">papa</hi> material in color.</p>
              <p>2. The four-cornered satchel (<ref target="#BucMani129a">figs. 43</ref>, <ref target="#BucMani130a">44</ref>; <ref target="#BucManiP005a">pl. 5, <hi rend="i">A</hi></ref>, 5–6). This basket has come into great favor and large numbers are made to give away as gifts to friends and visitors. Single wefts are used with the check stroke. The technique differs in principle from the two-cornered basket. In the two-cornered satchel the corners are formed by the ends of the straight braid which closes one opening of a plaited cuff. In the four-cornered satchel the bottom is formed of plaited material first, and four corners are then turned, from which the wefts are directed upward at right angles to the bottom plane to form the sides and ends. (See <ref target="#BucMani129a">fig. 43</ref>.) The size of the bottom varies. A fair average for the smaller satchels is 8 inches by 2.5 inches. The shape also varies, from rectangular to square. The plaiting of the sides and ends continues by working sections until a depth of 5.5 inches or more is reached. The upper edge is formed by turning down the top dextral at the working edge to lie on the sinistral placed in the shed. This is done successively throughout the length of the rim. The lower ends of the turned-down dextrals are left projecting on the outer surface. These are afterwards cut off, and the free ends are tucked in under a crossing weft and are thus not seen from the outside. The rim forms a smooth, even edge from the downward half-turn of the dextrals. (For detailed technique see 27, p. 202.)</p>
              <p>Some of the satchels are finished off at the rim by a wrapped band laid along the outer side with another laid along the inner side. These two bands are stitched to the rim with a steel needle and cotton thread. Handles are formed of plaited bands from 0.3 inch to 0.6 inch plaited in the round in check and then flattened out. In
<pb xml:id="n137" n="129"/>
<figure xml:id="BucMani129a"><graphic url="BucMani129a.jpg" mimeType="image/jpeg" xml:id="BucMani129a-g"/><head><hi rend="sc">Figure</hi> 43. Four-cornered satchel, bottom and corner technique. Wefts plaited in check until area indicated by dotted lines is large enough to form bottom; with plaited bottom before her, plaiter selects suitable line of width on left; this line, indicated by left vertical dotted line, runs out at edges of plaiting above (far) and below (near); at these points corners formed by crossing two adjacent wefts at right angles to their previous course—in upper left corner, 4 and 5; to comply with check technique, weft 5 crossed over 4 and each plaited in check with set of wefts it encounters; lower left corner formed in similar way, 4′ over 5′; plaiting sections to outer sides of left and far dotted lines marking bounds of bottom are bent up at right angles, and gap at upper left corner between near series (5′–3) and far series (6–9) is closed and the two series will cross each other diagonally; nothing remains but to interlace them in check plaiting until corner clearly defined and part of far side and left end of satchel completed; near left corner treated similarly; to form corners on right, plaiting of bottom must be continued to right until length of bottom secured by coincidence of right vertical dotted line with far and near edges of plaiting; in figure, bottom will be slightly wider on right than left; this due to error in counting, illustrates what sometimes happens in actual plaiting; plaiter, realizing mistake, would uncross wefts 5 and 6 and remove weft 19, uncross wefts 5′ and 6′ at near corner and remove weft 17′; corners would be built up in same way on left.</head></figure>
some bands, colored elements are introduced which form a check pattern. The handle is single and long, the ends being sewed onto the inner side of the middle of the side rim with cotton thread.</p>
              <p>A still more modern note is struck in some of the wider bottomed satchels. The ends are bent inward by creasing the middle. The sides of the rim are brought together and kept closed with a dome fastener sewed to the inner side of the rim. The resemblance to the ladies' handbags of western culture is striking, and the idea was probably obtained from them.</p>
              <p>The wide range of the geometrical designs used included the chevron, triangle, and lozenge in various combinations, together with more complicated figures, as shown in <ref target="#BucMani130a">figure 44</ref>.</p>
              <p>3. Fine coconut leaf satchel, technique derived from the lauhala two-cornered satchel. The satchel figured in <ref target="#BucManiP005a">plate 5, <hi rend="i">A</hi></ref>, 3 is fairly modern. The material is obtained from the leaflets of the young unopened leaves in the middle of the leaf head. These are cut off from the midrib and boiled for a few minutes, when the upper shiny surface of the leaflets can be split off from the under surface in much the same manner as in procuring the <hi rend="i">papa</hi> material from lauhala to provide the colored wefts for overlaid plaiting. The thin coconut material becomes very white on drying, and narrow
<pb xml:id="n138" n="130"/>
<figure xml:id="BucMani130a"><graphic url="BucMani130a.jpg" mimeType="image/jpeg" xml:id="BucMani130a-g"/><head><hi rend="sc">Figure</hi> 44. Lauhala satchels, color designs. (Drawn by C. K. Hedemann.)</head></figure>
<pb xml:id="n139" n="131"/>
strips are folded longitudinally with the shiny surface showing on both sides. Much finer work can be done with the material, and in the satchel figured, 17 or 18 wefts to the inch were used.</p>
              <p>The satchel is made in exactly the same way as the two-cornered lauhala satchel as regards the serrated rim and bottom closure, but the plait is in twill, which is changed to various combinations to obtain structural decoration.</p>
              <p>The use of the specially prepared coconut leaflet material is a comparatively recent introduction and was stimulated by the fashion of making hats of the Panama type. The first person to use the coconut leaf was a woman from the Gilbert Islands who had married a Rakahangan. She, however, kept her process of preparing the material a secret. For a considerable time the Rakahangan craftswomen puzzled over the excellence of the material and endeavored to solve the secret. At last one struck upon the method of boiling the leaflets to enable them to be split. The discoverer of the method proudly told me that her method was superior to that of the Gilbert Islands woman who, as it was subsequently found when secrecy was useless, ran the leaflets quickly over a lighted fire. From then on, the people of the atolls used coconut leaflet instead of lauhala for their hats, which are beautifully made and provide some income.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d3-d3-d4" type="section">
              <head>
                <hi rend="lsc">Fans</hi>
              </head>
              <p>Fans (<hi rend="i">tahiri</hi>) are made from the ordinary coconut leaf dried to a brown color in the sun and from the beautifully white prepared leaflets now used in the manufacture of hats. The brown fans are made of sections of leaf midrib with the bilateral leaflets attached naturally. A part of the midrib serves as a handle. The closed leaflets are generally plaited in check, but sometimes the twill is used. Two varieties of brown fan were seen, one composed of a single section of leaf and another of two sections, and one type of white fan:</p>
              <p>1. Single-section fan (<ref target="#BucMani132a">fig. 45</ref>) made of a midrib section 11 inches long with nine pairs of leaflets. Technique shown in the figure.</p>
              <p>2. Two-section fan (<ref target="#BucMani133a">fig. 46</ref>) made of a piece of midrib 8.25 inches long carrying five pairs of leaflets and a shorter piece 7.5 inches long carrying four pairs of leaflets. Technique shown in the figure. A larger two-section fan, plaited partly in twill, is shown in <ref target="#BucManiP005a">plate 5, <hi rend="i">B</hi></ref>, 3.</p>
              <p>3. The white fan (<ref target="#BucManiP005a">pl. 5, <hi rend="i">B</hi></ref>, <hi rend="i">I-2</hi>) is neatly made of individual doubled wefts doubled around a thin wooden handle. The handle consists of a rod about 0.6 inch in diameter, which is left at its natural thickness for 4 inches and then cut down to a thin rod for about 9 inches to give support to the wefts. The wefts are added from the handle upward, each weft being doubled around the rod and plaited in check over the front in much the same manner as the two-section brown fan, the rod supplying the place of the leaf midrib. The wefts are then plaited in twilled-twos and twilled-threes and finished off in check toward the curved distal end. A fringe of dyed <hi rend="i">tou</hi> bast 2.75 inches long is caught in the plaiting and gives a characteristic finish. The fan is really a modern adaptation of the technique of the brown fan in <ref target="#BucManiP005a">plate 5, <hi rend="i">B</hi></ref>, 3, and a detailed description of technique is unnecessary.</p>
              <pb xml:id="n140" n="132"/>
              <p>
                <figure xml:id="BucMani132a">
                  <graphic url="BucMani132a.jpg" mimeType="image/jpeg" xml:id="BucMani132a-g"/>
                  <head><hi rend="sc">Figure</hi> 45. Single-section coconut leaf fan, leaflets short in figures to save space. <hi rend="i">a</hi>, butt end of midrib proximal with natural upper surface upward; leaflets kept closed with leaflet midrib forming far edge of wefts; leaflets on right twisted forward under butt end of leaflet in front, commencing from butt; thus nearest leaflet (1) twisted under leaflet 2 in front of it and 2 under 3, 3 under 4, in succession until leaflet 6 twisted under leaflet 7 in front of it; this ends twisting, for leaflet 7 acts as locking weft by passing over 8, under 9, to be drawn toward left to lie parallel with left end weft (9′). <hi rend="i">b</hi>, right leaflets plaited in check, nearest (1) being bent with upward half-turn to run parallel with midrib (10) and pass over and under intervening wefts (2–9) in check technique; next weft (2) similarly treated and so successively are wefts 3–6, upward half-turn of each weft places midrib edge of closed leaflets on right; when sixth weft (6) turned, remaining two wefts (8, 9) left as they are. <hi rend="i">c.</hi> left leaflets twisted in similar manner to those on right as far as sixth (6′); seventh (7′) crossed over to other side as locking weft but in fan examined crossed over both 8′ and 9′ instead of under 8′ and over 9′ due to slip in technique; wefts 1′ to 6′ bent successively with upward half-turn and continue check technique; last two wefts (8′, 9′) left as they are; left wefts (7′, 1′) carried across in check technique through right wefts to reach edge on right defined by weft 6; left wefts (2′–6′) cut short after crossing weft 9′, but in practice are full length; when plaited portions tightened up, right weft (7) will come over to left and lie parallel to 9′. <hi rend="i">d</hi>, wefts from right have crossed over to left and those from left, to right, check stroke being continued; dextral wefts (8, 9, 7′, 1′) on right edge have been doubled over weft 6 to define right far edge and ends passed down under crossing wefts which keep them in position, any excessive length being torn off; similarly on left, sinistrals 8′, 9′, 7, and 1 will be doubled over bounding weft (6′); when wefts above crossing weft (6′) as 8′ and 7, they are doubled over backward and when they are behind as 9′ and 1, they are doubled forward; bounding weft (6′) doubled backward on itself and passed under crossing wefts to fix it. <hi rend="i">e</hi>, in completed fan bounding weft (6) on right has been doubled over first weft (1′) and then passed under crossing wefts (7′, 9, 8) where its cut-off end projects on near side of 8; similarly on left, bounding weft (6′) has been bent backward around the first weft (1) and at back under crossing wefts (7,9′,8′); intervening weft ends dealt with in pairs of which sinistral weft under dextral; first pair on right (5, 2′ in <hi rend="i">d</hi>) supplies the technique; under sinistral 5 doubled over dextral 2′ and follows its first part by passing under crossing wefts 7′ and 8; the dextral weft (2′) doubled over second part of 5 and passes under crossing wefts; other three pairs similarly treated and fan completed with straight serrated edge.</head>
                </figure>
              </p>
              <pb xml:id="n141" n="133"/>
              <p>
                <figure xml:id="BucMani133a">
                  <graphic url="BucMani133a.jpg" mimeType="image/jpeg" xml:id="BucMani133a-g"/>
                  <head><hi rend="sc">Figure</hi> 46. Two-section coconut leaf fan. <hi rend="i">a</hi>, longer section (10) with five pairs of leaflets placed above shorter section (11) with four pairs of leaflets, so that leaflets of under section fall into intervals of those of upper section; lower section leaflets shaded in figure; thus nine leaflets on either side. <hi rend="i">b</hi>, leaflets alternately crossed to opposite sides, commencing with right side; right upper leaflet (1) crossed to left and left upper leaflet (1′) crossed over it to right; next lower leaflet (2) on right crossed over preceding crossed leaflet (1′) on its way to left and next left lower leaflet (2′) in its turn crosses over preceding crossed leaflet (2) on its way to right; each leaflet crosses alternately from either side and either section until all leaflets have crossed concluding with 9′ over 9. <hi rend="i">c</hi>, nearest right leaflet (1′) bent with upward half-turn, passes under next leaflet (2′), and runs parallel with midrib to cross all successive leaflets (3′–9′) in check stroke; next right leaflet (2′) similarly treated and so in succession to sixth leaflet (6′); last three leaflets (7′, 8′, 9′) left as they are; first leaflet (1′) inclined to left as locking weft and made to lie parallel with left last leaflet (9); other right leaflets (2′–6′) turned in plaiting also incline to left. <hi rend="i">d</hi>, first six leaflets (1–6) on left successively bent with backward half-turn and plaited in check in similar manner to those on right; first of series (1) acts as locking weft and is inclined to right where it reaches right far edge defined by weft 6′; others follow suit and last three wefts (7, 8, 9) on left unturned; all crossing wefts plaited in check and point of plaiting formed by crossing of bounding wefts (6′, 6). <hi rend="i">e</hi>, of two bounding wefts, under weft (6′) doubled over crossing weft (6) and brought back along its previous marginal course to pass under crossing wefts (4, 2, 9′, 7′) where its cut-off end projects on near side of last weft (7′); weft ends on right doubled over double marginal weft (6′), those above it (4, 2, 9′, 7′) being doubled back and passed under one or more crossing wefts to fix ends; alternate wefts (8′, 1, 3, 5) under marginal weft (6′) doubled forward over it and ends tucked under nearest crossing weft; left marginal weft (6) doubled over second crossing part of weft 6′ and interlaced in check corresponding to its former course where its end (6) projects on near side of last crossing weft (7); weft ends on left far edge treated in exactly same manner as on the right, wefts behind marginal weft (6) being doubled upward over it and those in front being doubled backward; fan thus completed with straight far edges which meet in median point.</head>
                </figure>
              </p>
            </div>
          </div>
        </div>
        <pb xml:id="n142" n="134"/>
        <div xml:id="t1-body-d1-d4-d4" type="section">
          <head>
            <hi rend="c">Clothing and Adornment</hi>
          </head>
          <div xml:id="t1-body-d1-d4-d4-d1" type="section">
            <head>
              <hi rend="sc">Introduction</hi>
            </head>
            <p>Clothing was further developed than in Tongareva. The Tongarevan perineal band of coconut stipule (<hi rend="i">kaka</hi>), which was the general dress of the men, was not used in Rakahanga. A more elaborate <hi rend="i">maro</hi> of fine plaiting was used instead. Kilts, capes, and poncho-like <hi rend="i">tiputa</hi> were also made. No form of weaving was known, and the craft by which the garments were made was plaiting. European textiles have completely displaced the old forms of clothing, but kilts and ponchos are sometimes plaited for festival dances and for presentation to visitors. Thus I was agreeably surprised on calling at Manihiki after our sojourn in Rakahanga to find that the people had of their own volition plaited kilts and ponchos for me as examples of their old-time garments. Though highly ornate, they gave some idea of the craft employed. Tupou-rahi also possessed a fairly old plaited <hi rend="i">maro.</hi></p>
          </div>
          <div xml:id="t1-body-d1-d4-d4-d2" type="section">
            <head>
              <hi rend="sc">Material</hi>
            </head>
            <p>Two plants which enter largely into Polynesian clothing, the paper mulberry (<hi rend="i">Broussonetia papyrifera</hi>) and the ti (<hi rend="i">Cordyline terminalis</hi>), were absent from the two atolls as from Tongareva. Their absence precluded the manufacture of bark cloth (<hi rend="i">tapa</hi>) and ti leaf kilts. For clothing material, the plants available were the coconut and hala (<hi rend="i">Pandanus</hi>). It is curious that in Tongareva the people selected the coconut palm and used the stipule for the men's <hi rend="i">maro</hi> and the leaflets for the women's kilts and capes. In Rakahanga, owing perhaps to the greater plaiting dexterity of the people, the material selected was lauhala (<hi rend="i">Pandanus</hi> leaf). The leaves of the young hala (<hi rend="i">puwhara</hi>) entered into all the articles made, and the coconut was disregarded except to provide fiber for a belt. The upper layer of the lauhala was also split off, and the thin material (<hi rend="i">papa</hi>) was used for decorative purposes. The bast of the <hi rend="i">tou</hi> (<hi rend="i">Cordia subcordata</hi>) was soaked in sea water and used to provide fringes for the garments. A reddish-brown dye was obtained from the <hi rend="i">nenu</hi> (<hi rend="i">Morinda citrifolia</hi>), but now imported trades dyes have taken its place.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d4-d3" type="section">
            <head>
              <hi rend="sc">Women's Garments</hi>
            </head>
            <p>Women's garments enumerated were the double apron and kilt (<hi rend="i">tipora</hi>), the cape (<hi rend="i">pikipiki</hi>), and the poncho (<hi rend="i">tiputa</hi>).</p>
            <p>1. The double apron form of <hi rend="i">tipora</hi> consisted of two rectangular pieces (<hi rend="i">tautape</hi>) of plaited lauhala. These were used in conjunction with a sennit belt (<hi rend="i">tukaha: tu</hi>, belt; <hi rend="i">kaha</hi>, sennit braid). Whether the <hi rend="i">tukaha</hi> was a particular form of belt like the many-stranded <hi rend="i">tu</hi> belts of New Zealand or merely a length of sennit braid was not made clear.
<pb xml:id="n143" n="135"/>
The two <hi rend="i">tautape</hi> were of different sizes, the longer one being slung over the belt in front to conceal the genitals, and the smaller one being hung over the back. The two <hi rend="i">tautape</hi>, with the sennit belt, constitute the <hi rend="i">tipora</hi>; but sometimes the term <hi rend="i">tautape</hi> is loosely applied to the whole combination.</p>
            <p>2. The kilt form of <hi rend="i">tipora</hi> (<ref target="#BucMani135a">fig. 47</ref>) consisted of a long <hi rend="i">tautape</hi> (<hi rend="i">tautape roa</hi>) piece of plaiting which completely encircled the waist and thus acted as a narrow kilt. Some confusion existed between different informants, one maintaining that the narrow kilt with a hanging fringe was termed a <hi rend="i">mahere</hi>, and another that the <hi rend="i">mahere</hi> was the short perineal band. The kilt in <ref target="#BucManiP006a">plate 6, <hi rend="i">B</hi></ref>, consists of a long, narrow band plaited in check with overlaid wefts of the thin <hi rend="i">papa</hi> material dyed red. The separate fringe of dyed <hi rend="i">tou</hi> bast, which is attached in a continuous piece to the upper and lower borders and both ends, is described in <ref target="#BucMani135a">figure 47</ref>. The two-cord attachment to fix bast elements in a short fringe is exactly similar to the technique used with longer strips to make kilts in the Cook Islands (27, p. 88) and Samoa (28, p. 254). The method of sewing the lauhala strips to the plaited band is modern and is due to the introduction of needles and cotton thread. It is probable that in the original garments the wide lauhala strips were split at one end into wefts, which were plaited to form the band to keep the hanging lauhala strips together.</p>
            <p>3. The cape (<hi rend="i">pikipiki</hi>) was also plaited with lauhala wefts, but the exact technique has been forgotten, as capes have long been out of use. Capes were worn over the shoulders and tied (<hi rend="i">ruruku</hi>) in front at the neck. They were stated to be women's garments, but were also worn ceremonially by men when going to the religious inclosures, or maraes.</p>
            <p>
              <figure xml:id="BucMani135a">
                <graphic url="BucMani135a.jpg" mimeType="image/jpeg" xml:id="BucMani135a-g"/>
                <head><hi rend="sc">Figure</hi> 47. Kilt and fringe technique. <hi rend="i">a</hi>, strip of <hi rend="i">tou</hi> bast (3) about 3 inches long, dyed red, doubled under two cords (1, 2) composed of strips of young coconut leaflets boiled and bleached white; doubling of strip of bast in middle forms far limb (4) and near limb (5). <hi rend="i">b</hi>, both bast limbs (4, 5) brought over two cords (1, 2) and passed down between them. <hi rend="i">c</hi>, strips of bast added successively to right and kept close together so that cords (1, 2) hidden by turns of the bast; in type kilt, fringe is 86 inches long. <hi rend="i">d</hi>, strips of <hi rend="i">papa</hi> lauhala (6) attached to inner side of lower edge of plaited band (7); strips range in width from 2.5 inches to 3 inches and are 28 inches long; strips doubled to half their length and the two ends sewn together with cotton thread to plaited band; two-ply twisted cords of white coconut leaflet used for decorative effect; four cords (8) bunched together are run along just above edge of plaited band on outer side; bast fringe (3) placed in position with its supporting cords (1) laid against edge of band (7) and just below decorative cords (8); cords and fringe attached to plaited band by single continuous thread of white coconut leaflet; stitch (9) passed through plaiting and around cords (8) from below upwards on outer side; after passing through plaiting above decorative cords, thread descends obliquely to right on back, passes under fringe cord (1) and makes stitch (10) over it to fix it to band; stitch around decorative cords (9) made and again thread passes to right on back to appear under fringe cord; in this way, by stitches 0.3 inch apart, both cords and fringe attached along all four edges of plaited band.</head>
              </figure>
            </p>
            <pb xml:id="n144" n="136"/>
            <p>4. The poncho (<hi rend="i">tiputa</hi>) was plaited in check with prepared lauhala. (See <ref target="#BucManiP006a">pl. 6, <hi rend="i">A</hi></ref>). A hole was left for the head so that the garment hung down in front and behind to the thighs or knees. The square-cut hole was decorated by a <hi rend="i">tou</hi> bast fringe made with the two-cord attachment as in the kilt fringe. The fringe was attached to the edge of the plaiting by a continuous colored bast thread which passed spirally around the fringe cord and through the edge of the plaiting. Another fringe was attached to the outer edges of the plaiting. This was composed of single strands of colored bast which were held down over the middle by two colored bast threads which passed along the edge in the plaiting strokes. The ends of the colored strands on the plaiting side were then turned outward.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d4-d4" type="section">
            <head>
              <hi rend="sc">Men's Garments</hi>
            </head>
            <p>The men's garments consisted of two forms of perineal band. The perineal band is known generally as a <hi rend="i">maro</hi>, but the shorter form received the specific name <hi rend="i">mahere</hi> and the longer one, <hi rend="i">taoa.</hi></p>
            <p>1. The short band (<hi rend="i">mahere</hi>) was plaited in lauhala. It was 1.5 arm spans (<hi rend="i">maro</hi>) in length and 1 finger span (<hi rend="i">ngahonu</hi>) in width. The band was passed between the legs and the front and back ends were held in position by a sennit belt (<hi rend="i">tukaha</hi>), the ends falling over it. (<hi rend="i">Ka huru mai i mua, ka huru mai i muri.</hi>) If the band was long enough, it was passed around the waist after passing between the legs.</p>
            <p>2. The long band (<hi rend="i">taoa</hi>) was from 4 to 7 yards long and about 2 finger spans wide. (See <ref target="#BucMani137a">fig. 48</ref>.) The specimen of <hi rend="i">taoa</hi> owned by Tupou-rahi was made of fine plaiting with 10 wefts to the inch. Overlaid plaiting in thin <hi rend="i">papa</hi> material stained with native <hi rend="i">nenu</hi> dye was used for decorative effect. Twill strokes were used to relieve the ordinary check plait, and a colored fringe of <hi rend="i">tou</hi> bast was attached at either end.</p>
            <p>The plaited <hi rend="i">taoa</hi> was a well-made garment and a marked advance over the coconut stipule <hi rend="i">maro</hi> of Tongareva. Being long, it was wound around the waist besides being passed back between the limbs. The long <hi rend="i">taoa</hi> with colored wefts and end fringes was used by people of higher status on festive or ceremonial occasions, whereas the short <hi rend="i">mahere</hi> was used as ordinary clothing.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d4-d5" type="section">
            <head>
              <hi rend="sc">Accessories</hi>
            </head>
            <p>Headdress. Pieces of net (<hi rend="i">kupenga</hi>) were worn around the head like turbans by men on particular occasions such as consulting the gods.</p>
            <p>Eye shades (<hi rend="i">taumata</hi>) have been described. (See p. <ref target="#n133">125</ref>.)</p>
            <p>Neck ornaments (<hi rend="i">takawe</hi>) made of coconut leaf and hala were worn around the neck by the <hi rend="i">whakamaru</hi> chiefs when about to consult the gods. Their use inspired dread among the people, from association with this function.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d4-d6" type="section">
            <head>
              <hi rend="sc">Personal Decoration</hi>
            </head>
            <p>It is unfortunate that no accounts of the first contacts of European voyagers with the atolls of Manihiki and Rakahanga seem to be available. The present population has had little handed down regarding the ornaments and personal decoration of early times.</p>
            <pb xml:id="n145" n="137"/>
            <p>Ribbons consisting of the thin white material peeled off from the <hi rend="i">tua</hi> surface of the closed young coconut leaves of the growing center (<hi rend="i">rito</hi>) were used. They were tied together in long thin streamers and stuck in the hair by women on festive occasions. The material (<hi rend="i">kamuka</hi>) corresponds in form and use to the <hi rend="i">revareva</hi> of Rarotonga.</p>
            <p>
              <figure xml:id="BucMani137a">
                <graphic url="BucMani137a.jpg" mimeType="image/jpeg" xml:id="BucMani137a-g"/>
                <head><hi rend="sc">Figure</hi> 48. Long perineal band (<hi rend="i">taoa</hi>): from left end (1) to middle (4), length is 11 feet 2.5 inches, total length is 22 feet 5 inches; width (<hi rend="i">patapata</hi>) at ends, 14.5 inches, and in middle, 10.5 inches; end from points 1 to 2, 33.5 inches long, maintains width of 14.5 inches; this part colored with overlaid wefts and at intervals of about 8 inches, transverse rows (5) of twilled-threes relieve check; length between points 2 and 3, 45.5 inches, and width gradually diminishes from 14.5 inches on left to 10.5 inches on right; narrow parts (6, 6) at either edge carry on colored overlaid plaiting, but middle part (7) is plain white check plaiting relieved by four longitudinal rows of twilled-twos; beyond ends of colored strip (6, 6), middle part of band (8) plain; right half of band repetition of left half; fringe (9) of <hi rend="i">tou</hi> bast dyed with <hi rend="i">nenu</hi>, 4 inches deep; extends across ends, continued for 3 inches along sides; narrow strip (<hi rend="i">peipei</hi>) (10) consists of stained <hi rend="i">papa</hi> material, doubled and split on free edges to within 0.15 inch of doubled edge; this laid on edge of plaiting with split parts directed outward; as each side weft of plaiting reaches edge margin, turned in over split portion of <hi rend="i">peipei</hi> and so fixes it in position as part of fringe; strips of <hi rend="i">tou</hi> bast also added (<hi rend="i">whakaumu</hi>) to split portions of <hi rend="i">peipei</hi> and fixed by plaiting edge; bast strips caught by middle, and end which projects in over plaiting turned out to join other limb in fringe; when <hi rend="i">peipei</hi> split portions and bast have been fixed by plaiting, unsplit portion which lies on plaiting folded outward and thus covers split parts and makes neater finish; in finishing off edges, beyond <hi rend="i">peipei</hi> (10), plain wefts turned back in usual finishing edge technique observed in band finish of two-cornered satchels; in wefts overlaid with colored <hi rend="i">papa</hi>, colored material turned back but plain weft beneath left out and subsequently cut off; in narrow colored strips (6, 6) check stroke used throughout; edges of plain part (8) turned in finishing technique on under surface making narrow finishing strip about 0.2 inch deep; wefts cut off.</head>
              </figure>
            </p>
            <p>Neck ornaments (<hi rend="i">takawe</hi>). A form of ornament plaited from coconut leaf or lauhala was worn around the neck (<hi rend="i">hei ki te kaki</hi>) by the <hi rend="i">whakamaru.</hi> The appearance of these ornaments was regarded with a certain amount of fear by the public, as it was known that the wearer was about to consult his god.</p>
            <p>Breast ornament (<ref target="#BucMani138a">fig. 49</ref>). Edge-Partington (6, series 1, p. 62, no. 5) figures a breast ornament and describes it as, “Breast ornament of pearl shell in shape of fish.” The ornament is 7 inches long, much larger than the shank of an ordinary bonito hook. Such ornaments were not mentioned
<pb xml:id="n146" n="138"/>
to me. The addition of two short strings of beads attached at one end to a hole in the shell resembles the post-European technique of Melanesia, to which area the ornament may belong.</p>
            <p>Tattooing was not indulged in until after European contact.</p>
            <p>
              <figure xml:id="BucMani138a">
                <graphic url="BucMani138a.jpg" mimeType="image/jpeg" xml:id="BucMani138a-g"/>
                <head><hi rend="sc">Figure</hi> 49. Ornament of pearl shell: 1, head cut through hinge of shell; 2, apparently two strings of beads, carried by hole pierced through projecting edge of hinge. (After Edge-Partington.)</head>
              </figure>
            </p>
          </div>
        </div>
        <div xml:id="t1-body-d1-d4-d5" type="section">
          <head>
            <hi rend="c">Stonework</hi>
          </head>
          <div xml:id="t1-body-d1-d4-d5-d1" type="section">
            <head>
              <hi rend="sc">Material</hi>
            </head>
            <p>Though a low-lying coral atoll does not supply suitable basaltic material, a good deal can be done with coral boulders and coral limestone slabs. Rakahanga and Manihiki, however, do not show the activities in available material that characterize Tongareva. The outstanding Tongarevan features, house sites and religious inclosures, are lacking, and stone piers for canoes were not observed.</p>
          </div>
          <div xml:id="t1-body-d1-d4-d5-d2" type="section">
            <head>
              <hi rend="sc">Miscellaneous</hi>
            </head>
            <p>House sites. The stone platform, according to my informants, was not used. The surviving native house on Rakahanga is characterized by the absence of any platform or stones marking the wall boundaries. This is in marked contrast to the platforms of coral boulders associated with the Rarotongan types of houses which have become popular. The platform was introduced with the type of house. Because the present modern villages occupy the sites of the old villages, any traces of the past that might have survived the changes of time have been obliterated. The custom of segregation within the present village sites from the earliest times prevented the establishment of houses on other islands that might have left traces of interest to the archaeologist. It seems true, however, that not only were no house platforms made formerly, but that even the marking of the wall boundaries with stone was non-existent.</p>
            <p>Maraes. The scanty material concerning maraes is presented on page <ref target="#n216">208</ref>. Here again the segregation into three villages prevented the survival of a marae on an outer island.</p>
            <p>Graves. (See p. <ref target="#n225">217</ref>.)</p>
            <p>Boundary stones (<hi rend="i">tuakoi</hi>) were used in the village of Te Kainga.</p>
          </div>
          <pb xml:id="n147" n="139"/>
          <div xml:id="t1-body-d1-d4-d5-d3" type="section">
            <head>
              <hi rend="sc">Adzes</hi>
            </head>
            <div xml:id="t1-body-d1-d4-d5-d3-d1" type="section">
              <head>
                <hi rend="lsc">Stone Adzes</hi>
              </head>
              <p>Owing to the lack of suitable material, stone adzes are not to be expected on an atoll isolated from frequent communication with high volcanic islands. For material, the atolls had recourse to the shell of the <hi rend="i">Tridacna.</hi> It came as a surprise, therefore, that three stone adzes were procured in Rakahanga. I picked up a tanged adz on the old village site of Te Kainga. My attention was drawn to it from its color, which was darker than the surrounding pieces of white coral. A few feet away, Judge Ayson picked up a piece
<figure xml:id="BucMani139a"><graphic url="BucMani139a.jpg" mimeType="image/jpeg" xml:id="BucMani139a-g"/><head><hi rend="sc">Figure</hi> 50. Triangular tanged adz (C. 2743). <hi rend="i">a</hi>, front: surface has marked longitudinal convexity; blade and butt surfaces continuous without any distinction except at side edges where shoulders formed by inward inclination of butt; blade slightly convex transversely, defined by well-marked side edges; blade width fairly even throughout; sides curve in slightly to cutting edge, more marked on one side; butt more convex transversely than blade; sides rounded off; adz narrows progressively toward convex poll. <hi rend="i">b</hi>, back: somewhat rounded median edge commences on back of butt, is fairly distinguishable in upper part of blade; in lower part of blade, median edge becomes less distinct and inclines slightly to one side of median line to meet bevel surface; median edge divides back into two postero-lateral surfaces which are convex transversely and meet front at well defined longitudinal edges; front side edges give adz affinity with triangular adz technique and make up for rounded posterior median edge; bevel surface evident, but lower part broken by concave depression which has marred cutting edge; owing to flattening out of median edge, chin curved in outline; upper part of butt slopes upward and forward to meet convex curve of poll. <hi rend="i">c</hi>, right side: front longitudinally convex; back concave; side edge continuous from blade to butt and over poll; butt slopes upward and forward to meet poll edge. <hi rend="i">d</hi>, cross-sections.</head></figure>
<pb xml:id="n148" n="140"/>
of similar stone, smaller and without a tang. The third adz had also been picked up on Te Kainga and was presented to me when the Rakahangan people made their parting gifts of farewell. Adz terminology is given by Hiroa (28, pp. 333–335).</p>
              <p>1. Tanged adz (<ref target="#BucMani139a">fig. 50</ref>; <ref target="#BucManiP010a">pl. 10, <hi rend="i">C</hi></ref>, 2). The adz is made of a gray stone and has been so weathered that the general surface is smooth. It gives the general impression of a triangular tanged adz with a posterior median edge. The posterior median edge is not pronounced owing to the comparative thinness of the blade and the rounding off of the lower part.</p>
              <p>2. Small triangular adz, without tang (<ref target="#BucMani140a">fig. 51</ref>; <ref target="#BucManiP010a">pl. 10, <hi rend="i">C</hi></ref>, 1). This fragment is regarded as an adz because it was found near the tanged adz and was of similar stone, though incrusted with a deposit of lime. The lower end was unfortunately chipped off some considerable time ago, for the white deposit completely incrusts it. The posterior median edge is well marked though rounded off by the incrustation. The lower end shows a trace of a triangular bevel surface.</p>
              <p>3. Large quadrangular adz, with tang (<ref target="#BucMani141a">fig. 52</ref>; <ref target="#BucManiP010a">pl. 10, <hi rend="i">C</hi></ref>, 3). Whatever the doubts that may be occasioned by the preceding artifacts, none can exist with regard to the third adz. It is made of gray stone which has been ground throughout. It is characterized by a well-formed tang defined below by a deep convex shoulder and ending above in bilateral lugs at the junction with the poll. The longitudinal edges separating the four surfaces of the blade are well ground, and the two posterior ones are markedly concave. Owing to the pronounced longitudinal concavity of the back, the chin is prominent. The bevel surface is large and with the blade increases in width towards the cutting edge.</p>
              <p>
                <figure xml:id="BucMani140a">
                  <graphic url="BucMani140a.jpg" mimeType="image/jpeg" xml:id="BucMani140a-g"/>
                  <head><hi rend="sc">Figure</hi> 51. Triangular adz, without tang (C. 2744). <hi rend="i">a</hi>, front: surface slightly convex longitudinally and transversely; lower end shows loss of cutting edge by large chip fractured obliquely downward from before back; sides marked by distinct edges; surface narrows from cutting edge to poll formed by convex curve somewhat lower on one side. <hi rend="i">b</hi>, back: posterior median edge extends from just below poll and meets what was evidently triangular bevel surface at its apex; of bevel surface only trace of upper part remains but is sufficient to indicate that it was triangular; median edge divides back with two transversely convex postero-lateral surfaces which meet front in typical acute angled edges of triangular adzes. <hi rend="i">c</hi>, right side: longitudinal front edge continuous between blade and butt and continuous over poll. <hi rend="i">d</hi>, cross-sections.</head>
                </figure>
              </p>
              <pb xml:id="n149" n="141"/>
              <p>
                <figure xml:id="BucMani141a">
                  <graphic url="BucMani141a.jpg" mimeType="image/jpeg" xml:id="BucMani141a-g"/>
                  <head><hi rend="sc">Figure</hi> 52. Quadrangular adz, with tang (C.2742). <hi rend="i">a</hi>, front: edge narrows slightly owing to rounding off of corners; adz slightly concave longitudinally and flat transversely except for slight rounding off of well defined side edges (3); long flaw (4) along one of side edges; shoulder (2) flat and forms convex curve toward tang (5) formed by grinding away front of butt and rounding off front longitudinal edges; tang inclines backward at angle with blade and narrows toward poll, where it ends in two well shaped bilateral lugs (6,6) which come to rounded point with hollow between them. <hi rend="i">b</hi>, back: tang and blade surfaces continuous, concave longitudinally and slightly convex transversely owing to slight rounding off of back longitudinal edges (7); back narrower than front, and parts of lateral surfaces (9) may be seen; bevel (10) forms large surface, quadrilateral in shape, bounded above by straight chin (8), below by wide cutting edge, and at sides by well marked edges formed with lateral surfaces; bevel convex longitudinally and flat transversely, except near cutting edge where slightly concave. <hi rend="i">c</hi>, right side view: shoulder (2) and lugs (6) prominent; back has marked longitudinal concavity; poll (11) forms surface bounded by straight edge at back and concave edge in front due to contour of lugs; cutting edge well ground but blunt and shows no flaws or chips.</head>
                </figure>
              </p>
              <p>The quadrangular adz is exactly similar to an adz (<ref target="#BucMani142a">fig. 53</ref>) discovered by the late Captain Allen on the island of Nassau, which lies to the west of Rakahanga. This adz was revealed by the slipping away of a bank, when it was seen sticking out from below the surface level. The adz was deposited by Captain Allen in the Australian Museum and afterwards given by him to the Auckland Museum, where it now is. It has been figured by Skinner (23, p. 92), who drew attention to the two lugs (poll-knobs) resembling those in Moriori adzes from Chatham Islands and in a less pronounced form from the South Island of New Zealand. The presence of lugs in the triangular adzes of the Austral Islands has also been remarked by Stokes (manuscript in Bernice P. Bishop Museum). The Nassau adz is a little larger than the Rakahangan adz, as the following figures, compared with measurements made by Skinner, show:</p>
              <pb xml:id="n150" n="142"/>
              <p>
                <table rows="6" cols="3">
                  <row>
                    <cell/>
                    <cell role="label" rend="center">
                      <hi rend="lsc">Nassau</hi>
                    </cell>
                    <cell role="label" rend="center">
                      <hi rend="lsc">Rakahanga</hi>
                    </cell>
                  </row>
                  <row>
                    <cell>Length</cell>
                    <cell rend="right">12 inches</cell>
                    <cell rend="right">10¼ inches</cell>
                  </row>
                  <row>
                    <cell>Cutting Edge</cell>
                    <cell rend="right">3⅜ inches</cell>
                    <cell rend="right">3 inches</cell>
                  </row>
                  <row>
                    <cell>Width at poll</cell>
                    <cell rend="right">1¾ inches</cell>
                    <cell rend="right">1¼ inches</cell>
                  </row>
                  <row>
                    <cell>Thickness at bevel</cell>
                    <cell rend="right">2 inches</cell>
                    <cell rend="right">2 inches</cell>
                  </row>
                  <row>
                    <cell>Weight</cell>
                    <cell rend="right">5¼ pounds</cell>
                    <cell rend="right">4 pounds (±)</cell>
                  </row>
                </table>
              </p>
              <p>
                <figure xml:id="BucMani142a">
                  <graphic url="BucMani142a.jpg" mimeType="image/jpeg" xml:id="BucMani142a-g"/>
                  <head><hi rend="sc">Figure</hi> 53. Quadrangular tanged stone adz with bilateral lugs, from Nassau Island: <hi rend="i">a</hi>, front; <hi rend="i">b</hi>, right side view. Convex shoulder (1) formed by cutting away front of butt to form tang with projecting bilateral lugs (2), marked longitudinal concavity of back (3), clearly defined heel (4) with corresponding thickness of blade and long slope of bevel (5) are also features of Rakahangan adz. (After Skinner.)</head>
                </figure>
              </p>
              <p>
                <figure xml:id="BucMani142b">
                  <graphic url="BucMani142b.jpg" mimeType="image/jpeg" xml:id="BucMani142b-g"/>
                  <head><hi rend="sc">Figure</hi> 54. Adzes of <hi rend="i">Tridacna</hi> shell. <hi rend="i">a</hi>, adz in B. P. Bishop Museum (C. 2767): shell about 8 mm. thick but thinner in places owing to natural grooving of outer surface; outer surface of shell forms front, ground down in places to remove natural roughness; lower cutting edge fairly straight but curved up slightly at one end; shell fairly straight; bevel to form cutting edge, on front. <hi rend="i">b</hi>, adz in B. P. Bishop Museum (C.2768): outer surface of shell forms front; shell curved, front distinctly convex longitudinally and transversely; lower cutting edge convex; bevel grinding on back; thickness, 5 mm.</head>
                </figure>
              </p>
              <pb xml:id="n151" n="143"/>
              <p>Skinner states that the Nassau adz is of light gray volcanic tuff, and it is evident that the Rakahangan adz is of the same material. Little doubt can exist that the two adzes came from the same workshop, being formed of the same material and by the same technique. It would be interesting to know where that workshop was, for both adzes have evidently been introduced, if not by the same people, at least by people who had mutual contact in some part of Polynesia.</p>
            </div>
            <div xml:id="t1-body-d1-d4-d5-d3-d2" type="section">
              <head>
                <hi rend="lsc">Shell Adzes, Haft</hi>
              </head>
              <p>Two <hi rend="i">Tridacna</hi> shell adzes (<ref target="#BucMani142b">fig. 54</ref>) were among the presents given to me on Manihiki. Though they were evidently made for the occasion, it is presumed that the form and hafting reflects the technique of the past. One of the adzes was hafted to a movable socket.</p>
              <p>
                <figure xml:id="BucMani143a">
                  <graphic url="BucMani143a.jpg" mimeType="image/jpeg" xml:id="BucMani143a-g"/>
                  <head><hi rend="sc">Figure</hi> 55. Adz haft, with simple mesial peg. <hi rend="i">a</hi>, side view: rou