Publicly accessible
URL: http://www.nzetc.org/collections.html
copyright 2006, by Victoria University of Wellington
All unambiguous end-of-line hyphens have been removed and the trailing part of a word has been joined to the preceding line, except in the case of those words that break over a page. Every effort has been made to preserve the Māori macron using unicode.
Some keywords in the header are a local Electronic Text Center scheme to aid in establishing analytical groupings.
Botany Department, Victoria University of Wellington
No. 2 R.D., Ashburton
With about 40 species the genus Aciphylla, popularly known as Spaniard or Speargrass, is one of the larger genera of New Zealand flowering plants. It is also one of the most distinctive as some of the species are so spinescent that it is difficult at first sight, even for botanists, to believe that they belong to the Umbelliferae (carrot family).
Much remains to be done in defining and understanding the relationships of the species, but it was thought that it would be useful to report the current state of knowledge as a matter of interest and also in the hope that it might stimulate observations and even perhaps collection of material by others.
In what follows, except where otherwise indicated, the observations are those of the first author. The Chatham Island species will be considered separately after those of the main islands.
On the main islands these are spinescent, mostly short-stipuled, and have narrow elongate inflorescences from a metre to 3 or more metres tall.
It was noted that some species exude a milky juice when cut, others a clear juice. Correlated features are short (4-5 mm), ribbed fruits and long (8-10 mm), winged fruits respectively. The milky juice group comprises A. aurea (figs. 1, 2), A. ferox, and A. horrida; the clear juice group A. squarrosa, A. squarrosa var. flaccida, A. glaucescens, A. subflabellata (fig. 3), A. scott-thomsonii and A. colensoi.
In the former group some populations of A. aurea are distinctive, probably in the genus as a whole, in having the stomata of the leaflets sunken in grooves. The latter group can be subdivided further into species with leaves three or more times pinnate, with the petiolules flattened in a plane at right angles to the lamina. (A. squarrosa, A. squarrosa var. flaccida, A. glaucescens, A. subflabellata); and species with leaves one to two pinnate (lower leaflets only), with the petiolules flattened in the plane of the lamina (A. scott-thomsonii, A. colensoi).
Aciphylla aurea. Mt. St. Bathans
Aciphylla squarrosa occurs along the coasts and on adjacent hills on both sides of Cook Strait. The species has also been recorded on the Volcanic Plateau and the axial ranges, but this appears to be a
A. squarrosa var. flaccida by Kirk and perhaps the variety should be raised to specific status. Records of A. glaucescens in the North Island are probably of this variety, but A. glaucescens mostly has leaf segments much wider than those of typical A. squarrosa, not narrower. A. colensoi occurs at a number of localities along the axial ranges.
Aciphylla ferox with its wide leaf segments and ascending to appressed bracts is common on the wet mountains of north-west Nelson. Records from Marlborough are very doubtful.
A. aurea, with narrower leaf segments and spreading bracts, is quite common on the drier Marlborough mountains. With predominantly short stipules this is a different form from the populations further south. It comes near to sea level at Woodside Gorge. The plants here have rather drawn-out flexible leaves, but this may be due to the shady habitat. Small plants similar to A. aurea have been collected at a few localities further to the west — Mt. Cann (see last section this article) and in marble crevices on Mt. Owen and Hoary Head. This may be a distinct variety.
A. glaucescens is abundant in places in north-west Nelson, but also occurs less abundantly in Marlborough in wet sites.
A. subflabellata, common in the eastern South Island, apparently reaches its northern limit in the upper Wairau Valley.
A. colensoi reaches its southern limit in this region.
The typical form of Aciphylla aurea with long stipules and A. horrida are the large species with milky juice in this region. A. aurea can be common on the drier eastern side of the mountains. A. horrida favours wetter habitats in the west from Arthur Pass southwards. In its typical form it is clearly distinct from A. aurea, being larger with much broader leaf segments in which the stomata are not sunken in grooves. The leaves too are quite a dark green compared with the orangey green of A. aurea. Oliver states that this species often has only two pairs of leaflets per leaf. In the Eyre Mountains we found three pairs of leaflets to be equally common often with secondary pinnules on the lowermost pair. Stipules are more variable than those of typical A. aurea. Between or even within plants stipules may vary from long to short or even be absent.
At some locations of intermediate rainfall we seem to have encountered a problem in distinguishing the species. On Mt. Pisa
A. aurea, but on closer examination began to have doubts. The plants were similar to A. aurea in size and narrowness of the leaf segments but were a darker green, had variable stipules and did not have the stomata sunken in grooves as did populations of A. aurea further east. This problem will need further investigation.
Aciphylla aurea. Seed head. Porter's Pass.
Aciphylla subflabellata. Seed head with appressed bract segments forming a kind of cage. Porter's Pass.
Aciphylla scott-thomsonii, for which an inflorescence 4m high has been recorded, represents the once-pinnate subgroup of those species with clear juice. This species occurs east of the divide, and is abundant in some wetter western localities, but is restricted to locally wet sites further east.
J.R.L. recently repotted more than 100 two-year seedlings of A. glaucescens and A. subflabellata belong to the several times pinnate subgroup. They are both eastern species, but the latter is like A. aurea in its tolerance of dry sites at low and middle elevations. The relatively short leaves with very compound stipules almost equalling the rest of the lamina are distinctive. An unusual feature is that the centre of the leaf rosette often lies several centimetres below ground level, presumably as a result of root contraction. This may be an adaptation related to dryness of habitat.A. subflabellata and the crown of each was well below the surface even though the mixture was always moist. Also, most of the seedlings had sent out a long ‘questing’ root from 60 cm to over a metre long. This is by far the most vigorous root development encountered so far.
A. glaucescens, as previously noted, requires moister conditions. In northern Southland we encountered plants near a rock outcrop, which had the tall inflorescences and spreading bract segments of A. glaucescens, but the leaf segments were only as wide as those of A. squarrosa.
It may be appropriate to consider the problem of Aciphylla inermis at this point. Oliver (1956) described the species from one herbarium specimen, but it has since been found (singly or in pairs) at three other localities in shrub vegetation near Mt. Cook. The leaves are soft, grey-green and two-pinnate with the petiolules laterally flattened. This suggests an affinity with the A. squarrosa group of the large species, although the inflorescences are only about half a metre tall. However, A. inermis has milky juice and the male inflorescences are broad and these features plus the soft texture raise the possibility, admittedly unlikely, of a relationship with A. dieffenbachii of the Chatham Islands. Unfortunately no female plants have been seen and until some are discovered the status and relationships of the species will be uncertain.
It has been suggested that Hybrids in many genera show reduced pollen fertility as judged by the percentage of shrunken or empty pollen grains. The pollen of a few Aciphyllas, including A. inermis is a hybrid between Aciphylla aurea and Anisotome haastii (Wilson, 1976). If it is a hybrid we find it difficult to envisage the rigid leaved, spinescent A. aurea as one parent.A. inermis, has been examined and the percentage sterilities were — A. lyallii, 1%; A. crenulata, 4%; A. inermis (Type), 4%; A. latibracteata (Type. Suspected hybrid), 31%; A. scott-thomsonii x A. poppelwellii (Old Man Range), 78%.
A few small plants of a soft, grey-leaved, as yet undetermined Aciphylla, reminiscent of A. inermis, have been collected on peaks on Banks Peninsula.
These are very diverse, but have inflorescences that are mostly less than half a metre high. They can be arranged into five groups, two of which appear to be related to the two groups of the large species.
The clear juice and relatively long, winged mericarps indicate a relationship with the comparable group of the large species. Inflorescences are narrow and elongate. None of the group is found in the North Island and in the South Island the species are mainly along and west of the main divide.
A. indurata is intermediate in size, having inflorescences up to 1 m tall and with its bright yellow midribs it is rather like a small version of A. colensoi. J. W. D. has collected this species on the Lyell Range, the type locality, and it is also known from other localities in south Nelson and northern Westland.
Three other species of the group also occur on the Lyell Range — A. hookeri, A. townsonii and A. trifoliolata. The first has distinctive short and broad coriaceous primary and secondary leaflets; the second very soft leaves with longer, narrow, almost threadlike leaflets; the third, two or three pairs of long, relatively broad, rigid leaflets.
A. hookeri occurs elsewhere in south and north-west Nelson and north Westland and A. townsonii has a similar range, although probably not extending to north-west Nelson. A. trifoliolata was not seen by J. W. D., but may occur on rocky outcrops at higher elevations. The Lyell Range is the type and at present only locality for A. trifoliolata.
A species of this group has recently been discovered in the south of Stewart Island and it is interesting that it bears some resemblance to the type specimens of A. trifoliolata.
A. crenulata ranges from Arthur Pass to Fiordland. It has simple to once-pinnate, often long and rather flexible leaves. Some segments of the inflorescence bracts extend at right angles to the stem.
A closely related species is A. anomala of south and north-west Nelson. It has less coriaceous leaves and the bract segments are not spreading.
A. takahea of Fiordland may also be related to A. crenulata, but we have little field experience of this species. Its leaves are more compound than those of A. crenulata.
This group is restricted to the southern third of the South Island in alpine habitats on and east of the divide. Milky juice and short, ribbed mericarps suggest a link with the second group of the large species. Stipules are long, which also accords with some members of the latter group. The leaves are once-pinnate and the inflorescences are narrow.
Aciphylla kirkii has rigid broad leafllets and is found at high elevations in the western part of the group's range. We have observed it at about 1800 m on rocky outcrops on the Remarkables, where the leaves had a distinct greyish bloom. A. kirkii has also been recorded at lower elevations on Coronet Peak and Ben Lomond, but the plants here grow among snow tussocks and have much longer leaves without any bloom, so their identity is uncertain.
Plants of this group on mountains further east are smaller than A. kirkii and were referred to A. hectori by Oliver. However, our observations suggest that there are two entities involved. On the Old Man Range there is one form at about 1000 m in moist grassy situations. This has ascending leaves with narrow leaflets marked with yellow cross veins. At about 1600 m in alpine tundra is an apparently different form with spreading leaves and wider leaflets lacking cross veins.
In cultivation in J.R.L.'s garden the differences between the two forms if anything increase. The larger form agrees fairly well with typical A. hectori, although plants of the latter from the type locality (Hector's Col, Mt. Aspiring) have acute not rounded leaflets. The other form matches the type of A. poppelwellii, which was merged with A. hectori by Oliver.
Ian Spence of Wendonside, nothern Southland, has observed the two forms on the Garvie Mountains and has also found that their differences increase in cultivation. It has yet to be decided whether the two forms should be treated as varieties or species.
A. verticillata is the last species of this group. It was described by Oliver and is only known from the type specimen from Mt. Kyeburn and a garden grown plant from ‘somewhere in Otago’. (Other records for the species in Otago and Westland are based on mixed sheets of uncertain origin and the leaves identified as A. verticillata match those of A. hectori.) We have visited Mt. Kyeburn and found the A. poppelwellii form, but nothing to match the larger A. verticillata. The answer to this problem may have been found in Ian Spence's garden where a plant of A. poppelwellii from the Garvies growing in a shaded situation had developed unusually long leaves similar to those of A. verticillata and also an inflorescence with the verticillate arrangement of the lower bracts on which the name of the latter was based. As one of the original specimens of A. verticillata was also garden grown in Southland this species may be based on an unusually large specimen of A. poppelwellii.
A frequent feature of this group is variability in numbers of leaflets and stipules within individual plants.
In the preceding groups of large and small species the mericarps usually have fewer than five ribs or wings and have many oil-tubes. In the following groups the mericarps usually have five ribs and few oil-tubes.
Here stipules are mostly short and the juice is clear. Some species have broad and others narrow inflorescences.
A. dissecta is the only species of the group in the North Island, being common in the southern Tararuas. It has broad inflorescences, soft, 3-4 pinnate leaves and compound stipules. The one record for the South Island (Boulder Lake) is probably incorrect.
In the northern South Island A. polita has been recorded from the Richmond Range in Marlborough and from the mountains of north-west Nelson. These plants also have broad inflorescences but are smaller than A. dissecta, with 2-3 pinnate leaves and fewer pairs of primary leaflets.
The type locality is in the Richmond Range and the populations here differ from those of north-west Nelson in a number of features. The former have fine, hair-like leaf segments and compound stipules, the latter coarser segments and mostly simple stipules. Within northwest Nelson there is some variation between localities in the width of leaf segments and in the north near Boulder Lake there is a form with particularly coarse segments. Perhaps the north-west Nelson populations should constitute a new species.
The record of A. polita in the Tararuas is probably erroneous.
A. monroi is common on the drier eastern mountains from Marlborough to mid-Canterbury. Records from north-west Nelson and Westland are doubtful. Leaves are once-pinnate, greyish green and the inflorescences are broad.
A. similis (fig. 4) is similar to A. monroi with once-pinnate leaves and broad inflorescences, but it is a larger plant and the leaflets are more numerous, yellow-green, broader and set at quite a wide angle to give the leaves a palm-like appearance. It is found on the main divide, but its range is probably much less than that given by Oliver. The type locality is Arthur Pass and it is found on other mountains in north Canterbury and adjacent Westland.
Oliver's treatment of certain species of this group south of Arthur Pass — A. lyallii, A. similis, A. gracilis, A. flexuosa — is confusing. These all have once-pinnate leaves, but, except for A. similis, narrow inflorescences.
The type locality for A. lyallii is in Fiordland and Oliver regarded the species as extending from there to the mid-Canterbury mountains with a doubtful disjunct record from the St. Arnaud Range in Nelson. It seems to us that the Fiordland populations are different from those further north. Typical A. lyallii has short female flower heads with broad bract sheaths which largely conceal the fruits. The leaves have evident joints at the top of the sheath and the petioles are concave above with acute, often cartilaginous, margins. The plants in populations further north have longer female heads with narrower bract sheaths and visible fruits and in the leaves the sheath joint is obscure and the petiole is flat to convex above with rounded margins.
Aciphylla similis. Male, left; female, right. Temple Basin.
If it is decided that the two forms are specifically distinct then A. montana Armstrong, based on plants from the Rangitata Mountains, should be reinstated for the extra Fiordland plants. In what follows the name A. montana is used for convenience of reference.
Most records of A. similis from the Southern Alps south of Arthur Pass, often based on sterile material, are probably A. montana, e.g. at Mt. Cook. Apart from the marked difference in their inflorescences, adult, well developed leaves of the two species can also be distinguished. The former has 6-10 pairs of leaflets and the latter 3-4.
Some records of As well as being broader, broad inflorescences also differ from narrow in usually having fewer compound umbels and in the terminal umbel being the largest.A. similis from mountains of West Otago and northern Southland east of the main divide are probably of an undescribed species. We have collected this on the Remarkables and there are also herbarium specimens from the Hector Range and the Garvie Mountains. This species has 3-4 pairs of leaflets often widely spreading, a broad male inflorescence and a female inflorescence basically of the broad type A. similis in having fewer pairs of leaflets and a condensed female inflorescence. Plants from the Eyre Mountains with rather broader leaflets may be of a related form.
We have not studied A. lyallii in Fiordland yet, but as Oliver noted there seem to be several forms there.
Oliver based A. gracilis on two herbarium specimens, the type being from the Kirkliston Range. J. R. L. has examined populations at the type locality and we have both collected similar plants at the Hakataramea Pass and Mt. St. Bathans. This form seems to us close to A. montana, although the leaves do frequently have only two pairs of leaflets so it may warrant varietal status.
A. flexuosa was based on a single herbarium specimen from Mt. Alta. It agrees well with A. montana except that the stipules are unusually long. However sometimes leaves of A. montana have stipules longer than usual and, in addition, there is another herbarium specimen from the type locality collected by the same collector on the same day as the type, in which the stipules are quite short.
The problem will have to be investigated in the field before the status of A. flexuosa can be decided.
We have not yet investigated A. divisa and A. multisecta in the field. A possible new species in this group has recently been discovered on Mt. Alexander in Westland. It is very small, with soft 2-3 pinnate leaves and broad inflorescences. More material is required.
In this group A. pinnatifida occurs on mountains from Central Otago to Fiordland; A. traillii on Mt. Anglem in northern Stewart Island and A. traillii var. cartilaginea on mountains in southern Stewart Island. They have clear juice, but long stipules and narrow inflorescences in which broad bract sheaths largely obscure the fruits. The laminae are once-pinnate or simple to once-pinnate and are unusual in that the leaflets increase very little in length downwards and may be sub-opposite to alternate. In A. pinnatifida the stipules are also pinnate and very similar to the lamina. In A. traillii the lamina and stipules are often simple, although a few pairs of leaflets are often present in well-grown plants.
A. pinnatifida usually grows in very wet places and is strongly rhizomatous. A. traillii is not rhizomatous.
The two varieties of A. traillii require more detailed comparison. The var. cartilaginea appears to be a stouter plant with shorter flower heads, wider leaf segments and with more frequent development of leaflets.
The three species of this group are restricted to the mountains west of Lake Wakatipu and in Fiordland. The stipules are long and simple (sometimes with an accessory pinnule in A. congesta) and the lamina once-pinnate with none of the variation noted for Group B. The juice is milky and inflorescences broad.
A. crosby smithii has coriaceous yellow-orange leaves and forms large cushions. A. congesta and A. spedenii have softer grey-green leaves and often also form cushions. A. spedenii is unusual in having the petiole and intervals between the leaflets very reduced.
The species in this group also form cushions, which are very dense. The leaves are much smaller than those of Group D, with simple laminae and long stipules, except for A. simplex where stipules are absent. The juice is milky and inflorescences broad and the compound umbels are very compact due to the short peduncles of the simple umbels. Mericarps have not been seen for A. leighii but in the other two species — A. dobsonii and A. simplex — they are unusual in that there are usually no oil tubes in the intervals between the ribs.
All the species grow in rocky places and A. dobsonii (fig. 5) and A. leighii often in screes.
A. dobsonii has the widest distribution, being found at a number of localities near and east of the divide from near Lake Tekapo to North Otago.
A. simplex occurs on eastern mountains in north and central Otago.
A. leighii is only known from the Darran Range in Fiordland and female flowers and fruits have not yet been collected. Allan (1961) suggested that it might be just a reduced form of A. dobsonii, but new material shows it to be quite distinct. The leaves are smaller and paler and the leaf segments are rounded to notched at the tip rather than acute. The species needs fuller investigation.
The two species here are A. dieffenbachii and A. traversii. They have inflorescences up to 1 m high so can be regarded as large.
A. dieffenbachii has soft, several times compound leaves with stipules reduced to rounded lobes. Its fruits are distinctive in having mericarps with three and two wings and in being very broad.
Aciphylla dobsonii. Cushion about 60 cm diameter. Mt. St. Bathans.
The species has sometimes been treated as a distinct genus, Coxella.
A. traversii, with its coriaceous, once-pinnate leaves with short but evident stipules and narrow 3-4 winged mericarps, has been regarded as very distinct from A. dieffenbachii. However closer investigation shows that the two species share a number of features, one of which is peculiar to them. They both have milky juice, broad inflorescences, winged mericarps with fewer than five wings and differ from other species of Aciphylla in having the mericarp veins out towards the tips of the wings instead of at their bases.
Some of the species have been described as having serrulate or crenulate leaflets. The latter are not truly toothed, but the appearance is due to the presence of small, more or less hemispherical multicellular protuberances on the strips of lignified tissue near the surface of the leaves and inflorescence axes. These tubercles are most abundant at the margins of leaflets but can also occur on the midribs and sometimes on all the lignified strips.
Not all of the groups of species have tubercles and of those that do some have coarse tubercles (fig. 6), visible to the naked eye, and others fine tubercles (fig. 7), visible only with magnification.
The groups of large species all have coarse tubercules as do groups A and B of the small species. Groups C and E have fine tubercles and the remaining groups no tubercles (fig. 8).
Aciphylla takahea. Scanning electron micrograph of ventral leaflet surface. Margin to right. X 50.
Some species have soft tissue at the front of the base of the petiole extending a little to the back at each side. As a leaf ages this pulvinate tissue enlarges and the lamina curves away from the younger leaves. In some cases the pulvini are very pronounced, particularly in dried specimens, where they become brown and shrunken. In other species they are less clearly defined or absent.
Most of the large species appear to have pulvini, although they are often not very clearly defined morphologically. Most species of Group C of the smaller species have pulvini, sometimes very clearly defined, but in the remaining groups they seem to be absent.
Probable hybrids involving a number of species pairs have been observed in the field and some of these have been described as species. In the usual situation there are more or less uniform populations of two species and an occasional plant of intermediate appearance. Others can probably be added to the following list. None of the suggested hybrids has been investigated experimentally.
Aciphylla montana. Scanning electron micrograph of dorsal leaflet surface. Margin to right. X 50.
Aciphylla pinnatifida. Scanning electron micrograph of dorsal leaflet surface. Margin to right. X 50.
A. aurea x A. crenulata? Mt. Cook. (H. D. Wilson)
A. aurea x A. montana? Mt. Cook. (H. D. Wilson)
A. aurea x A. scott-thomsonii Hakataramea Pass. (
A. aurea x A. subflabellata Rangitata. (J. R. Le Comte)
A. aurea x Anistotome haastii? Mt. Cook, = A. inermis? (H. D. Wilson)
A. aurea x A. poppelwellii Dansey Pass. (
A. colensoi x A. dissecta Mt. Hector; Mt. Holdsworth, = A. intermedia (
A. colensoi x A. polita N.W. Nelson form. Glenroy Area. (J. R. Le Comte)
A. crenulata x Anisotome haastii Mt. Cook. (H. D. Wilson)
A. dissecta x Anisotome aromatica Mt. Hector. (
A. divisa x A. montana Mt. Cook. (H. D. Wilson)
A. divisa x Anisotome haastii Mt. Cook. (H. D. Wilson)
A. ferox x A. anomala Mt. Peel, N.W. Nelson. (
A. hookeri x A. townsonii Lyell Range. (
In the last case the probable hybrids were growing in large patches of granite boulders and were so varied that they must include back crosses.
A. hookeri x A. indurata Mt. Bovis. (
A. horrida x A. similis Temple Basin, = A. latibracteata. (
A. scott-thomsonii x A. poppelwellii Old Man Range. (
Some backcrossing seems to have been involved here too. Some plants we regard as hybrids have been identified as A. kirkii from this locality.
A. squarrosa var. flaccida x A. dissecta Mt. Holdsworth (
A. squarrosa var. flaccida x A. colensoi Mt. Holdsworth (
The higher alpine species are noted for their lack of flowers, sometimes for years. This would seem to pose a problem for identification but in fact years of field-work suggest that most of the species can be distinguished easily, even by people with no botanical training, by leaf shape and arrangement alone. I am working on a vegetative key, using the inflorescence only as a ‘back-up’ when flowers are present.
The writer has 31 species growing successfully in the garden, as well as several variants, hybrids, and at least 2 unnamed species. With the exception of A. dobsonii and A. simplex all have proved relatively easy to establish in cultivation. Collected rosettes are
A. dobsonii and A. simplex have not been so easy. (A. leighii has not been tried at all.) They usually develop only fine white roots that will not sustain the plant in open ground. However material of A. dobsonii collected from above Lake Tekapo developed a much stronger root system and it is felt that these plants will establish in the garden. This adaptability is probably due to the fact that material collected was from small, young plants only. Because these species usually fail to show strong root development it is obvious that propagation from seed, though slow, offers the best chance of success. Seed collected in the summer of 1976 has germinated and growth will be watched with interest. I have grown many species from seed, but urge patience because often germination does not take place until two winters have elapsed.
Several Aciphylla species have flowered in the garden, and in cases where both sexes were flowering together seed has been produced.
Growing all these species in one uniform garden area is of real scientific value and may yet resolve some of the existing anomalies. For example, if habitat is the cause of observed regional variation this should even out after plants have been growing together for some years in the garden. If genetically divergent forms are involved, however, clearly one would expect them to remain distinct even under the same growing conditions.
I have noted that many high alpine species grow vigorously in a very narrow altitudinal band and that plants growing above or below this are much reduced and unlikely to flower. The leaves of such plants are often not at all characteristic of the species. In the group that has once-pinnate leaves and short, simple stipules, i.e. A. monroi, similis, montana, gracilis, the ‘out of bounds’ plants closely resemble A. monroi, so much so that they have sometimes been incorrectly identified as such. When grown in the garden, however, the plants became ‘normal’ in one season.
An interesting point to ponder is why a species will be stunted when growing 20 metres out of its zone, but quite normal when growing in the garden, about 1,300 metres below its zone!
I have collected A. similis from several areas north of the type locality (Arthur's Pass) and in general terms it can be said that the further north the station, the more reduced the plant. Specimens collected on the Mt. Cann Range (north of Lewis Pass) were so reduced that it was at first thought that the plant was A. monroi until closer examination proved it to be A. similis. After two seasons in the garden these plants now show that they are definitely A. similis, having grown to similar size and habit as plants of that species from other localities. When the Mt. Cann plants were collected we thought
A. monroi and A. similis; not hybrids, but part of a cline. The situation has been clarified in the garden.
A year earlier, I collected plants approximately 4 miles north of Mt. Cann in an area just south of the southern boundary of the Nelson Lakes National Park. All these specimens matched A. monroi. Garden cultivation has shown them to be the north-west Nelson form of A. polita (bipinnate), not A. monroi. Presumably they were reduced because of growing out of their usual habitat.
It is known, then, that A. similis extends northwards to a point only some 4 miles from the most southern habitat of the north-west Nelson form of A. polita. The intervening mountains would be interesting places to botanise and Aciphylla specimens from there should be grown on in the garden.
Also on the Mt. Cann Range, amongst hundreds of acres of A. colensoi, I found a small colony of a form of the northern A. aurea. No more than 30 cm high, and without the typical stipules of A. aurea, these plants were growing at over 1,600 metres and may deserve varietal status. Specimens are now well established in the garden and future growth will be watched with interest.
The most common problem in growing Aciphylla seems to be collar rot but this condition is not very prevalent in the author's well-drained mid-Canterbury garden. The most troublesome ‘beasties’ are mealy bugs that live and multiply alarmingly between the leaf sheaths, slowly killing the plant. Their presence is evident by the white ‘meal’ on the lower part of the leaves and a systemic spray should be used, thoroughly drenching the plant.
In the revision of the genus, the emphasis is on field-work and both authors have attempted to visit as many mountains as possible. Middle-age and available time place certain restrictions upon such plans.
We have had wonderful assistance from a West Coast helicopter firm which has enabled J.R.L. to visit many otherwise inaccessible areas, including many isolated peaks above bushline.
Every summer swarms of vigorous trampers and climbers are loose in the mountains; if even a little of this energy could be harnessed for botanical purposes, what a full and complete revision could be made. TRAMPERS please note: no offer of assistance will be refused.
We are grateful to Dr.
From studies on several populations of this New Zealand Mygalomorph spider the following characteristics emerged: (i) age structures showed marked seasonal variation, (ii) sampling for density indicated the spider is most successful in open habitats, (iii) migratory activity is a feature of many members of these populations, and dispersal rates of up to 100 metres/year are possible, (iv) over the summer months, a substantial change in numbers of mature spiders was noted in many populations. This was regarded as being due to the activities of Pompilid wasps.
P. antipodiana is a large, hairy mygalomorph spider of the family Dipluridae. It commonly grows to 30 mm in body length and is readily identifiable by its orange-brown cephalothorax, and a coating of long black hairs on its legs and abdomen. It spins a tubular or tunnel web which connects to a flat sheet web used for prey capture. The tunnel in which the spider shelters may be under a log or rock; in a crevice on a bank; in the hollow of a tree; and sometimes the spider may take up residence in or on buildings, a feature which often brings it to the notice of urban dwellers.
Published material on the ecology and behaviour of the spider includes two articles by this author (1973, 1975) and also the discussion of New Zealand mygalomorphs in the book by the Forsters (1973). The distribution, feeding patterns, postures, and behavioural repertoire have been described in these articles. A little is known of the spider's reproductive behaviour and of its movements about the
Information on the spider's life history has been presented in the chart below. The findings of the Forsters have been arranged side by side with information derived from the present study.
The information from the two sources is quite comparable, and the differences in the egg-sac construction time, for instance, could well be due to climatic differences between the Wellington region and the lower South Island. In most of the populations studied, very few mature males were found. The ratio of mature males to mature females was often as great as 1 : 100, and very commonly around the 1 : 50 mark. The population figures quoted in this articles will therefore refer mainly to female spiders.
The method of assessing age structures for these studies was to use body length as a crude indicator of age. It has been found, from records kept over the past six years, that P. antipodiana will grow to 12 mm long by the end of year one (see Fig. 1 for details). By the end of year two, the spiders may be upwards of 20 mm long, and the females may be found in possession of egg-sacs at this age. These statistics give three reasonably natural groups which were used in the age structure surveys:
P. antipodiana. Points plotted on the graph are mean, and the upper and lower ends of the range found in each age group that was measured. Numbers of spiders measured were as follows: 3 months, 110; 6 months, 80; 9 months, 68; 12 months, 45; 24 months, 31. The upper growth limits at 12 and 24 months are the ones referred to in the text for determining age structures of populations. Note how the growth curve flattens during the winter months of the first year. After sexual maturity is reached by the end of year two, the growth rate appears to slow down but not enough measurements were available to reliably plot the curve beyond this point.
There are several methods of assessing body length in spiders for age structure studies. They are: (1) by capture and direct measurement; but this usually results in damage to the web and so was used infrequently. (2) Inspection at night by torch light and visual estimation of body length. (3) Measurement of the tunnel that the spider lives in. This method proved to be a reliable estimate of body length. It involves the assumption that the width of the tunnel opening is equal to three-quarters of the body length of the spider living there.
One thing that soon became apparent when populations were counted to determine age structures was the highly variable nature of the age structure. The main factor involved was the time of the year when counts were made. For example, if a count was made in late summer to early autumn, then large numbers of juvenile spiders from the summer reproductive activities feature in the figures.
P. antipodiana, as at January. Small spiders (S) numbered 167, medium spiders (M) numbered 18, large spiders (L) numbered 18.
P. antipodiana as in Fig. 2. Age structure as at August. Small spiders numbered 46, medium spiders numbered 12, large spiders numbered 18.
If, however, a count was taken in winter or spring, then many of these juvenile spiders would have died and the resulting figures show a bias toward the mature spiders. In the Johnsonville broken rock bank population, shown in figures 2 and 3, the decrease in the number of small spiders present from January to August was 73 per cent. Winter mortality of young spiders is due to several factors. Starvation and disease must play a part, but predation by planarian worms accounts for a good number of the deaths. Predation by other spiders — sometimes their own species and sometimes from other species — certainly takes place. The brown grass spider Miturga has been observed feeding on young tunnel web spiders several times during the course of this survey. See Table 1 for detailed mortality figures.
Age structure can be modified substantially by the nature of the habitat. For example, a population in a brick wall at Northland, Wellington is always characterised by having more individuals in the large spider group than in any of the other groups. This is due in there being a lack of web sites on the wall; a row of drainage holes at the bottom, and a few cracks in the wall being the only sites. Because of this, there is strong competition for sites, and they are usually filled by large spiders, giving the unusual age structure as shown in Figure 4.
Causes of mortality in Porrhothele antipodiana. Figures were derived from observations on several populations, 1971-76.
Similar effects have been noticed in populations in more natural habitats, such as rock faces and hillsides with accumulations of loose rocks. In many such areas a normal age structure may not be found. It seems that the maturing smaller spiders, and the medium spiders, are forced to migrate in search of suitable sites. This hypothesis gains support from the observation that most of the spiders found moving at night do fall into the medium size groupings.
Note: While the small and medium groups are composed of single year groups, the large group is made up of third, fourth and possibly older age groups. For this reason, the large group does take up a disproportionate amount of space on the age structure diagrams. The problem could not be avoided because it was not possible to accurately determine the age of spiders older than two years.
Population density is highly variable in P. antipodiana. It appears to be most closely related to the number of sites available for web building, and also to the local food supply. The population on the clay/broken rock bank at Johnsonville (figs. 2 and 3) seems to have been limited by the number of suitable sites available for the mature spiders. These were always filled and the variation in spider density was due not to changes in the numbers of mature spiders, but to changes in the numbers of small spiders. It is doubtful if the numbers of mature spiders in this population would have changed unless there were physical changes to the bank to increase the number of crevices or holes for the large and medium spiders to enter. Small spiders can make do with a site that is quite unsuitable for a mature spider to live in, but as they grow they are forced to (a) move from the locality, or (b) try to take over a site from an incumbent spider. As the older spiders are far stronger, they usually retain possession of their sites.
P. antipodiana population on brick wall, Northland, Wellington, as at August. Total population: 40.
Similar factors appear to govern density on habitats such as hillsides covered by logs or stones; density will be limited by the same combination of suitable sites and food supply.
Where density has been measured, figures have ranged from 4.4 spiders per square metre (clay/rock bank) to one spider per 150 square metres (regenerating broadleaf bush). If density is considered for a small area such as around the web of a mature female after the young have dispersed, then densities of 20 spiders per square metre are not uncommon.
Below are further measurements of density, based on the number of medium and large spiders per 20 x 20 metre sample area, 10 samples having been taken where possible:
These figures reinforce the general impression held by this writer — that P. antipodiana is most successful in the open habitat rather than in bush areas.
P. antipodiana populations on Mt. Kaukau, Wellington. Clumpings of spiders (one dot: one mature spider) are where loose rocks were concentrated below the steeper slopes and in gullies like the one shown in Plate 2.
It has already been mentioned that the young spiders disperse up to 1.5 metres from the parental web. This is not the end of the migratory behaviour in P. antipodiana, for although the majority of spiders do not move very often, and indeed may occupy the same site for a year or more, some of the medium spiders, in particular, move several times from site to site until a suitable site is found. The distance of these site to site movements is commonly 1-4 metres, and movements may be made as often as every 1-3 weeks. During the course of a year, it is possible that highly mobile individuals could disperse for distances of 20-100 metres from their original web sites.
In addition to these short site to site migrations, longer migrations also take place. In these, the individuals concerned may cover distances of 10-20 metres in one move. During the breeding season (late spring), the males may wander considerable distances. In some of the populations studied by this writer, males were often found up to 30 metres away from their known web sites.
When this migratory activity is considered, it is apparent that the clumpings of spiders — as in the Mt. Kaukau example, Fig. 5 — are not too far apart for gene flow between them to take place. P. antipodiana population units seem to be similar to the deme
P. antipodiana populations is not clear, but Forster (pers. comm.) has noted that some populations of this species show a very pronounced chevron pattern on the abdomen, while other populations possess almost uniformly dark abdomens.
Keeping track of the movements of these spiders sometimes takes up a considerable amount of time. It involves regular checking of all the known webs in the locality. Then, when a spider is found to have left its web, a search of the area must be made to find a newly established web. Checking on the size and colour of the inhabitant is usually sufficient to establish whether it is the individual from the empty web or not. When dealing with these relatively small groups, it is possible to become familiar with them as individuals. Where spiders happen to be very similar in body length and colour, they can be separated by a dab of white acrylic paint on the cephalothorax or the abdomen.
Mapping of the web sites in the locality being studied proved to be a definite advantage in keeping track of spider movements. The example in Fig. 6 was used to help keep track of the spiders in part of population 2 on the crib wall at Johnsonville.
In 1971, a broken rock/clay cliff face at Paremata measuring approximately 20 metres wide and 40 metres high was surveyed for numbers of P. antipodiana in excess of 20 mm body length. Counts were made at two-monthly intervals and the results were graphed (Fig. 7). The number of mature spiders was relatively stable over most of the year, except for the fall off in numbers over the summer. The peak number recorded was 60 spiders, in July-September. This decreased by 30 per cent, to 42 spiders, by December.
During late October-December, several large black wasps were seen searching the cliff face, running into P. antipodiana webs, and occasionally dragging paralysed spiders along. These Pompilid wasps (species: Salius monachus) are known hunters of the larger New Zealand spiders, so it was thought they might be the cause of the summer decrease in the spider numbers. Other factors investigated were: predators, but none were observed working in the area; dehydration due to summer drought, but the remaining spiders appeared to be in good condition so this factor was not regarded as being responsible for the population loss. Starvation/summer drought effects can be identified by the considerable shrinkage of the spider's abdomen that takes place under these conditions.
P. antipodiana. Low points are at January (J) and later in year at November-December.
With the Paremata study in mind, a longer term study was begun at Johnsonville on a more accessible population, located on a concrete crib wall. The wall was six cribs high, the spaces between cribs being 14 cm high by 60 cm wide. The assumption was that in the web sites it provided, this crib wall was not dissimilar to a broken cliff face. Methods of study used were: spiders assessed for body length on the basis of width of tunnel opening; the state of their web was taken as the main indicator of whether the spider was still resident in the site. If the spider does not continually add new silk to the web it soon deteriorates, and this is a reliable sign that the spider has either migrated or has been captured.
Wasps were captured, marked on the thorax with white acrylic paint and released. Particular attention was to be paid to spider numbers in September (prior to the onset of wasp activity) and late in February (after wasp activity ceases in this locality). Two wasp species were active in the area: Salius monachus (black) and Salius fugax (red).
Since the last survey, the crib wall has been planted in thick creepers and is no longer suitable for observation or for counting numbers of spiders present. Unfortunately, this study had to be terminated with the spider population consisting of a few very young individuals. It would have been interesting to plot the populations for another two or three years to see if the trends displayed were repeated; that is, a second rise in the spider population followed by a rise in the wasp numbers. As it is, the trends are reminiscent of classical prey-predator cycles.
The spider population at Johnsonville appears to have been a young one, undergoing expansion during 1971-74, by which stage it began to support an unduly large wasp population. No doubt the wasps were also taking spiders from adjacent populations, for it is doubtful that a peak population of 27 medium and large spiders could provide for the larvae of up to 30 wasps. The wasp activity was intense on the crib wall; sometimes as many as 10 wasps at a time were running over the bank, investigating webs and searching for spiders. Several spiders were observed being captured, and others were seen to escape the wasps by vacating their tunnels and leaving the wall on the run.
Why this spider population attracted so many wasps is not clear, but the effect of such activity was to reduce the spider population down to a few very young individuals. This is, in all probability, a rare situation.
This was another crib wall population, less than 100 metres distant from population 2. Occasional visits by red wasps were noted, but black wasps were not found active in this area. This population acted as a valuable control, for the influence of other factors such as summer drought/food shortage would have shown up in this population if they had been involved in the decline of population 2. However, as population 3 did not exhibit undue mortality over summer, then drought and food shortage were ruled out as contributors to the decline of population 2. The figures taken for population 3 in the years 1975-77 are as follows:
P. antipodiana and Pompilid wasps, with numbers of the two species of wasp being added together. Locality: Johnsonville, on crib wall. September figures only plotted. By January, 1976, numbers of medium/large spiders had reached zero.
P. antipodiana on bank at Northland, Wellington. Summer drop in population of medium/large spiders notable in January figure.
Porrhothele antipodiana, body length: 25 mm, seen here emerging from her tunnel or tube to gather in a slater moving across the sheet or expanded portion of the web.
P. antipodiana is very successful in — the looser rocks, particularly those associated with clumps of grass or low shrubs, are the most favoured ones. The slaters, millipedes and beetles which form the bulk of the spider's food are plentiful around the rocks. In addition, the drainage is good and this appears to be an important factor in determining whether P. antipodiana populations will attain high densities or not.
A fourth small population on a broken greywacke bank at Northland was observed for a year from July, 1976, to July, 1977. The familiar pattern of a summer population decrease was evident (Fig. 9). Again, in this area small numbers of Pompilid wasps were seen searching for spiders during the summer.
From these four population studies, it appears that populations of P. antipodiana can usually maintain stable numbers of mature indivudals — that is except when the Pompilid wasps are active during the summer months. Under these conditions the decrease in numbers is usually made up, partly from the pool of maturing younger spiders in the population, and probably in part from immigration of mature spiders from adjacent populations.
In part two of this study, detailed consideration will be given to the effects of wasp activity in modifying the P. antipodiana populations. Attention will also be given to the mechanisms by which the spider populations survive wasp activity — without suffering undue mortality rates.
‘Biogeography is like chess: he who really understands the game can tell in a matter of a few moves whether he stands before a genuine master or a pretentious patzer.’
Biogeography is the study of the distribution of organisms upon the surface of the globe. It can be divided into historical and geographical/ecological biogeography: ‘… if our time scales are broad, then we are concerned with the subject matter of historical biogeography; if narrow, then it is the subject of geographical ecology that is of most interest.’ (Cracraft, 1974: 215). Biogeography is often divided into phytogeography (plants) and zoogeography (animals) but as the problems both face are essentially the same the division is not recognised in this paper. (See Croizat, 1958; Fleming, 1962, and Banarescu, 1975, for further discussion of this.)
Since the inclusion of two chapters on ‘Geographical Distribution’ in ‘The Origin of the Species by means of natural selection’ (Darwin, 1859) animal and plant distributions have usually been interpreted within the context of the ‘centres of origin’ concept. Darwin believed that species originated in centres from which they dispersed by either passive or active dispersal mechanisms:
‘When we feel assured that all the individuals of the same species, and all the closely allied species of most genera, have within a not very remote period descended from one parent, and have migrated from some one birth-place; and when we better know the many means of migration, then, by the light which geology now throws, and will continue to throw on former changes of climate and of the level of the land, we shall surely be enabled to trace in an admirable manner the former migrations of, the inhabitants of the whole world.’
(‘Origin’, Chap. XV)
Darwin's biogeographic concepts were expanded and elaborated by a number of subsequent authors in books and papers which have had a significant impact on the development of biogeography. Some of the most important of these are ‘The Geographical Distribution of Animals’ (Wallace, 1876), ‘Island Life’ (Wallace, 1880), ‘Climate and Evolution’ (Matthew, 1915), ‘Antarctica as a faunal migration route’ (Simpson, 1940b), ‘Zoogeography: the geographical distribution of animals’ (Darlington, 1957), ‘Biogeography of the southern end of the World’ (Darlington, 1965), ‘Island Biology’ (Carlquist, 1974) and ‘Marine Zoogeography’ (Briggs, 1974). The fundamental assumptions underlying all Darwinian biogeography as noted by Croizat (1968a: 430) are. ‘(1) every taxon — the species first and foremost as the paragon of biological origin and related questions — comes into being in a center of origin of its own; (2) if “successful” it sooner or later outgrows it, and undertakes migration in order to occupy other regions of the earth, whether near or far; (3) the migration is accomplished thanks to appropriate means of dispersal. Since it is normal to find that plants and animals lacking said means are very widely distributed the conclusion is normally reached by orthodox — so called — phytogeographers and zoogeographers that means of dispersal are essentially mysterious.’
The centres of origin concept was challenged as early as 1901 (Croizat et al., 1974) but the concept was not critically analysed until the early 1940's in the work of Cain (1943, 1944). Cain demonstrated that none of the 13 criteria advanced for recognition of a centre of origin were of any real value in biogeographic studies.1 Cain's criticisms were to a large extent responsible for the rejection of the concept of dispersals from centres of origin by
Croizat's work finds its immediate historical roots not only in the work of Cain but also in Rosa's theory of hologenesis (i.e. primitive cosmopolitanism of life) and two works by Andersen (1908, 1912) on bats. Biogeographical implications of hologenesis are that:
‘Orthodox biogeography is monogenetic, it states that each species and each group of species arises from a small number of individuals, in a very small area (center of origin), from which the species (or group) spreads by active or passive migrations, over the whole area of
‘Hologenesis on the other hand, is extremely polygenetical and leads to a biogeographical theory which is opposed to the former one. Without denying migrations, hologenesis ought to make as a basic principle of biogeography not a distribution arising from different centers, but a primitive cosmopolitanism succeeded by a process of localisation taking place at the same time as the multiplication and differentiation of species.
‘Working from this basis we are able, without always resorting to often unlikely migrations and without having to invent continents to explain very satisfactorily how sometimes the same species and more often different species of the same genus or family can turn up in zones which are distant from one another and discontinuous.'
(Rosa, 1933: 120)
Croizat (1963: 93) considered that Andersen's work on bats was ‘a model of biogeographic method and analysis’. During a period when Gadow (1913) and Matthew (1915) were advocating extensive migrations/dispersal of animals Andersen (1912) was astute enough to comment: … that the spreading of bats from one locality to another must obviously have been greatly facilitated by their possession of wings, may in theory appear plausible enough, but when tested on the actual distribution of the species and subspecies it proves to be of much less importance than commonly supposed; it rests, in reality, on a confusion of two different things: the power of flight no doubt would enable a bat to spread over a much larger area than non-flying Mammalia, but, as a matter of fact, only in very few cases is there any reason to believe that it has caused it to do so' (p. lxxvii) and ‘… the present distribution of the Megochiroptera has not been influenced to any great, and as a rule not even to any appreciable extent by their power of flight …’ (p. lxxviii).
‘During the past 5 years, however, the emergence of the theory of plate tectonics has provided a better understanding of the history of the earth's crust and, hence, a more reasonable basis for explaining many problems of biogeography.’
(Raven and Axelrod, 1972: 1379)
‘I have never been able to understanding why certain authors erected continental “drift” or the like into paragons of biogeography. Whether the Bay of Bengal came about because the Deccan and the Indo-Chinese Peninsula, respectively, did drift apart, or whether land once in between its current shores crumbled, is not a question to interest the student of dispersal very deeply. The student takes stock of the
records of the geographic distribution of plants and animals, and objectively analyses them on a comparative basis, drawing the conclusions which the facts advise and common sense endorses.’ (Croizat, 1968b: 578)
One of the major problems in biogeography is finding an explanation for the faunal and floral resemblances between widely separated tropical and Southern Hemisphere temperate regions. There are two major ‘schools’ of thought on this problem: one, which traces the origin of its concepts of land connections of some sort during previous geological periods back to the work of Hooker in the 1850-60's (Hooker, 1860); and the other, which derives its view that present patterns of plant and animal distribution can be explained by progressive dispersal of life from an Holarctic center of origin along three independent routes into South America, Africa and the Indo-Australian region over a geography that is essentially modern from statements made by Darwin in chapter twelve of ‘The Origin of Species’.
Darwin believed that cold, glacial periods drove organisms from areas around the north and south poles into equatorial regions where they mingled together. When climatic conditions improved northern organisms were able to migrate southwards with the southern organisms; but southern organisms were unable to migrate into the Northern Hemisphere. This inability of southern life to migrate northwards he attributed to the great competitive, and hence ‘dominating’ powers of the northern organisms, such ‘powers’ having evolved in the northern plants and animals because they occupied the much larger land-mass areas of the Northern Hemisphere. Similarities between, and the presence of related organisms, in two or more of the Southern Hemisphere land areas was attributed to long-distance dispersal across water gaps.
Wallace (1876) presented the view that organisms had dispersed within the Holarctic region from Eurasia (Palearctic) to North America (Nearctic) via the Bering Land Bridge. From here animals migrated into South America, while they also migrated from Eurasia into Africa and along the Indonesian chain of islands into the Australasian region. Wallace (1880) adhered generally to a permanence of continental areas but was prepared to accept ancient land connections with regard to New Zealand and the Celebes which he class as anomalous islands.2
Matthew (1915) developed these ideas even further. According to Matthew the main groups of vertebrates arose in centres of origin in northern parts of the Holarctic and dispersed southwards along the three independent routes mentioned above. He believed that (1) animals disperse from large areas, into small areas, of land; based on the assumption that animals tend to disperse from continents to islands and not vice-versa, and (2) Matthew considered that alternations of arid and cold climatic zones in Holarctica resulted in the
Darlington (1957, 1965) modified Matthew's area-climate view of animal distribution. Arguing from the observable facts that larger land areas have more animals and that areas of favourable climate have more animal diversity, he suggested that animals disperse from large to small areas and from favourable climatic to unfavourable climate areas. Matthew's idea of dominant, competitive forms evolving under adverse climatic conditions was rejected by Darlington who argued that such conditions would result in cold-adapted tolerant forms. Darlington believed that dominant, more competitive forms evolved in warm climatic regions. Being an aprioristic believer in continental permanence (he admitted the possibility of continental drift/rafting in 1965 but so long ago that it could not have affected animal distribution) Darlington put forth the view that vertebrates (1957) and some invertebrates (1952, 1965, 1971) arose in the Old World tropics and dispersed from this centre of origin along three independent routes.
Simpson (1940 a and b) argued that marsupials had entered both South America and Australia by northern routes of entry. He also (1952) held that any biotic resemblances between widely separated continental areas could be explained in terms of the probability of long-distance dispersal.
All the above views depend on the idea of the permanence of continents and oceans through time. The recent widespread acceptance of the concepts of plate tectonics and continental drift/rafting have resulted in numerous reinterpretations of animal and plant distribution.3 For instance certain groups of carabid beetles considered by Darlington (1965, 1970) to illustrate his views can have their present distribution satisfactorily explained in terms of plate tectonics (see e.g. Jeannel, 1961; Noonan, 1973). Keast (1973) in an analysis of contemporary biotas showed that they were not inconsistent with what we know of the breakup of the large, original land mass Pangea and the two smaller super continents. Laurasia and Gondwanaland. A number of authors (e.g. Brudin, 1966, 1972; Cracraft, 1973 a and b, 1974; Fooden, 1972; Jardine and MacKenzie, 1972; Raven and Axelrod, 1972; Schuster, 1976) have interpreted present-day plant and animal distributions in terms of continental drift. Some (e.g. Brundin, 1966; Cracraft, 1973 a and b) postulate a southern origin
Despite the controversy between the advocates of these two viewpoints in recent years the differences that exist are not as significant as they appear to be from the polemical publications of Brundin (1972) and Darlington (1970). As correctly noted by Cracraft (1975), both involve the concept that as organisms evolve they disperse over the globe. Both schools have as their basis the recognition of centres of origin, dispersal mechanisms and pathways of dispersal, and interpret the biogeographic history of taxa within the context of this framework. Those who support continental drift as the key to historical biogeography merely apply the centre of origin/dispersal framework to a mobilist concept, as opposed to the Matthew/Darlington stabilist concept of past geography (Nelson, 1975). However, Matthew/Darlington biogeography is not incompatible with continental drift concepts. McKenna (1973) attempts a reconciliation and Noonan (1973) notes that Gondwanaland satisfies the criterion for Darlington's large land mass with favourable climates.4 Horton (1974), however, has challenged Darlington's dominance views and noted a number of cases where he considers animals have dispersed from small into large land areas.
Darwin, Hooker and Wallace discussed the origins of the New Zealand biota in their work. Hooker (e.g. 1860) considered that an Antarctic continental extension was needed to explain the similarities between the plant life of New Zealand, South America and the Sub-Antarctic Islands. Darwin regarded New Zealand as an ‘oceanic’ island, populated by dispersal over the sea, but he had doubts: ‘It is only against the former union with the oceanic volcanic islands that I am vehement. What a perplexing case New Zealand does seem: is not the absence of leguminosae, etc. fully as much opposed to continental connection as to any other theory? … The presence of a frog in New Zealand seems to me a strongish fact for continental connection, for I assume that seawater would kill spawn, but shall try.’ (Darwin, 1903: 418)
Hutton (1872: 228) disputed the Darwinian classification of New Zealand as an ‘oceanic’ island and commented that ‘The New Zealand fauna may be correctly called the remnants of a continental fauna.’ He attributed the origin of the New Zealand fauna and flora to migration across landbridges into New Zealand. In 1872 Hutton proposed that in the lower Cretaceous period an Antarctic continent extended northwards into Polynesia, connecting
‘I now abandon the idea of an extensive Antarctic continent, because the soundings that have been lately taken in the Pacific Ocean have shown that such a supposition is highly improbable. At the same time, these soundings have made it clear how the connection really took place.’ (pp. 7-8.)
His mature view was that:
‘New Zealand, which formerly existed as the southern part of a continent extending through Australia to India, was isolated from Australia towards the close of the Jurassic period, but was attached to a South Pacific continent and received a stream of immigrants from the north. None arrived from the south because Fuegia was not then in existence. In the upper Cretaceous the land shrank to a size considerably smaller than at present. In the Eocene, elevating took place and New Zealand extended outwards in all directions but remained isolated from other lands. Plants and animals came in both from the north and from the south. In the Oligocene and Miocene periods New Zealand was, except for a short interval, a cluster of islands, but was upraised once more and obtained more immigrants from north and south during the Pliocene.’ (p. 273.)
Hutton conceived of biotic immigration into New Zealand as a series of ‘waves’ rather than as a continuous process and although Darwinian in his attribution of organisms to centres of origin outside New Zealand he was strongly opposed to long-distance transoceanic dispersal:
‘Our fauna and flora is indeed a standing protest against the views of those naturalists who would make the winds scatter abroad insects and seeds of plants over hundreds of miles, and who imagine land shells and lizards to float about on logs for days and weeks together without being killed.’ (p. 274.) Hutton's views on land connections as opposed to transoceanic dispersals to account for the New Zealand biota were supported by von Jhering (1892) and Hedley (1899).
Around the time that Hutton was invoking landbridges Wallace (1876) attributed the origin of New Zealand birds (‘still less does any other form of animal inhabiting New Zealand require a land connection with distant countries to account for its presence’, p. 431) and insects (‘The poverty of insect life in New Zealand must therefore, be a very ancient feature of the country, and it furnishes an additional argument against the theory of land connection with, or ever any near approach to either Australia, South Africa or South America’, p. 463) to trans-oceanic dispersal. Wallace later (1880) changed his views somewhat and accepted land connections between New Zealand and other land areas though at ‘a very remote epoch’.
Cockayne (1958) and Oliver (1925) both interpreted the origins
In a series of papers Fleming (1949-76) and to a lesser extent Gaskin (1970, 1975) systematised and concretised a centre of origin dispersal framework as the biogeography of New Zealand. Fleming's method has been to group New Zealand organisms into categories according to his understanding of