Oxygen Isotope Palaeotemperatures obtained from New Zealand belemnites are plotted on Fig. 1. There is a gap in the New Zealand belemnite succession apparently corresponding to Upper (perhaps also Middle) Tithonian. Neocomian and Aptian (Stevens, 1967). For the purpose of this study this gap can be partially closed by indicating palaeotemperatures obtained from Australian Dimitobelid belemnites. Inclusion of these results is thought to be a valid procedure as Dimitobelidae appeared in Australia in the Aptian and were shared by New Zealand and Australia in the Albian and Cenomanian. It is therefore likely that New Zealand and Australia would have also shared the same belemnites in the Aptian, had migration routes and sedimentation conditions in New Zealand been favourable at that time.

Each sample point on the graph (Fig. 1) represents an analysed cross-sectional cut of a belemnite guard. The resulting determination is therefore a mean of the temperature record preserved in the guard.

As may be seen from the graph, there are marked differences in values obtained from individual specimens from the same stratigraphic horizon. Analyses of samples taken serially across the growth rings of individual belemnite guards have also shown considerable isotopic variation between and within individual growth layers (Clayton and Stevens, 1968, and in preparation). This variability is interpreted as indicating that some isotopic exchange has taken place in most of the specimens. Because it is difficult to determine the degree of isotopic exchange a specimen has undergone, and to correct for this in the palaeotemperatures, only the minimum values are regarded as having significance, as these represent the least exchanged carbonates. In Fig. 1 the minimum values are outlined by the shaded curve and in the interpretation of results below, the inferred temperatures discussed are based entirely on minimum values. Averaging of palaeotemperatures, such as that carried out by Bowen (1961b, fig. 2, p. 82; see Fig. 1 of this paper) to obtain a temperature curve can only be misleading, as the page 4

FIG. 1: Values for Jurassic and Cretaceous temperatures based on oxygen isotope analyses of belemnites. Analyses of Australian Dimitobelindae, published by Dorman and Gill (1959, p. 91); Lowenstam and Epstein (1954, p. 222), and Bowen (1961 a), have been included to help fill a gap in the New Zealand belemnite sequence in the Lower Cretaceous.
The minimum values, representing the least exchanged carbonates, are shown by the shaded curve. The continuous curve shows the minimum values of West European belemnite palaeotemperatures determined by Lowenstam and Epstein (1954, fig. 10, p. 226). The broken curve is from Bowen (1961b, fig. 2, p. 82) and is the result of an averaging of the results then available.

averaged result will undoubtedly include many determinations affected by exchange to a varying, but unknown, extent.

The world-wide palaeotemperature results obtained to date, while showing the existence of temperature fluctuations in the Cretaceous, reveal many anomalies, and these have not been satisfactorily resolved (Lowenstam, 1964; Voigt, 1965; Naidin et al., 1966; Berlin’ et al., 1966). Nevertheless the results agree in indicating page 5 a period of general cooling towards the end of the Cretaceous, and the New Zealand results show a similar trend (Fig. 1). But this simple picture is apparently complicated, according to recent Russian work (Naidin et al., 1964, 1966; Teis et al., 1965) by a rise from temperatures of 11.0-13.0°C in the Lower Maastrichtian to temperatures between 13.5-15.5°C in the Upper Maastrichtian^*.

The present distribution of coastal seawater surface temperatures in the New Zealand region can be used as a basis for comparison with palaeotemperatures. As each minimum value on Fig. 1 represents a mean of the temperature record preserved during the life of an individual belemnite, the proper comparison is with present mean annual isotherms, not individual summer and winter isotherms. The approximate positions of the relevant mean annual isotherms have been calculated from the charts in Garner and Ridgway (1965).

In making these comparisons it must be recognised that the temperature patterns of Mesozoic seas were in all probability markedly different from those of the present day and that the temperatures obtained from belemnites need not necessarily reflect surface sea water temperatures but may be the temperature of deeper water.

New Zealand palaeotemperatures obtained by the analysis of belemnite guards can be summed up as follows:

Inferred temperatures for the Upper Jurassic (Heterian-Puaroan in the New Zealand stage terminology), based on specimens from Kawhia Harbour, 38°5′S. Lat., correspond approximately to mean annual seawater temperatures off Greymouth, 42°27′ S. Lat. (14.7°C), so that seawater temperatures at that time may have been slightly cooler than those of today.

Using the Australian results for the Albian and Aptian, the Lower Cretaceous was apparently a cooler period, with a temperature minimum in the Aptian, but with a slight improvement in the Albian. Inferred temperatures for the Aptian, based on specimens from Lake Eyre, 28° S. Lat., correspond approximately to mean annual temperatures in the region between The Snares and Stewart Island, 47°30′ S. Lat. (12.2°C). Inferred temperatures for the Albian, again based on specimens from Lake Eyre, 28° S. Lat., correspond approximately to mean annual temperatures to the south of Cape Campbell, 41°50′ S. Lat. (15°C).

A range of similar inferred temperatures, 10.4-14.0°C, has been recorded from the Neocomian and Aptian of Russia (Teis et al., 1957; Berlin et al., 1966).

As may be seen from Fig. 1, a serious discrepancy exists between the Australian temperatures and the temperature graphs of page 6 Lowenstam and Epstein and of Bowen. Lowenstam (1964) has attempted to explain the Australian temperatures as being a result of water isolation during a climatic maximum, giving an artificially low temperature. But in the writer's opinion he has not satisfactorily disposed of the possibility that the high temperatures plotted by himself, Epstein and Bowen are due to varying degrees of isotopic exchange. Ludbrook (1966, p. 25) stated that the temperatures recorded by the Australian Aptian-Albian belemnites represent those existing at water depths in excess of 100 fathoms. This overlooks the possibility that the belemnites lived in higher levels in the sea than indicated by the sediments in which they are preserved. Also, the palaeoecology of belemnites suggest that they preferred shallow waters (Stevens, 1965). Moreover, Dorman and Gill (1959, p. 94) recorded well-defined seasonal fluctuations in the three Australian specimens used for growth-ring analyses; such fluctuations would not be expected at substantial depths below sea level.

The New Zealand inferred temperatures for the post-Albian Cretaceous (Urutawan-Haumurian) show a gradual increase until in the Turonian-Santonian (Arowhanan-Teratan) the lowest recorded values, based on specimens from the Wairarapa and Marlborough districts. 40°45′, 42°S Lat. respectively, correspond approximately to mean annual temperatures just south of Norfolk Island, ca. 30°S Lat. (22.0°C). In the Campanian-Maastrichtian the temperatures, based on specimens from E. Otago (Haumurian) and N. Canterbury Piripauan and Haumurian), 45°55′, 43°5′S respectively, gradually decline to reach in the Maastrichtian figures corresponding approximately to mean annual temperatures off Kaikoura, 42° 25′S Lat. (14.3′C).

Discussion

(Dr. I. Speden (N.Z. Geological Survey) presented the paper on behalf of Drs. Clayton and Stevens as both were overseas at the time of the Conference.)

Dr. P. Webb. I would like to emphasise the point that the Haumurian samples are restricted both geographically and stratigraphically and that the palaeotemperatures obtained may be of local, not regional, significance for the Haumurian.

Dr. I. Speden. Dr. Stevens is aware that the sampling is not fully representative.

Dr. C. A. Fleming. I believe Dr. Stevens’ careful sampling through individual belemnite guards is possibly the first such study anywhere in the world and his results suggest that previous results should be re-examined, especially where bulk samples have been used.

Dr. I. Speden. I would agree with Dr. Fleming and also say that Dr Stevens’ work on the various species of belemnites is also of page 7 considerable importance. The fact is that some species show seasonal variations and some don't. These latter results may possibly tie in with some results of work by a Russian team working with squids on the Great Banks. Some squids were found to migrate up and down through the water column from season to season, evening out any temperature fluctuations.