‘Let anyone pass slowly over the surface of the earth, especially in a north-south direction, stopping from time to time to give himself leisure to observe; he will invariably see (each) species varying little by little and more and more as he is further from his starting point.’

Lamarck. 1815.

It Has Long Been Recognised that the morphology of animals is affected by environment. Size particularly varies in many species. Mayr (1963) considers “Size is perhaps the character most universally subject to geographic variation. It varies in virtually every species of animal that has an extensive geographic range.’ In most cases geographic variation of a species can be correlated with changing climate and usually with temperature. If a close correlation between size and temperature can be demonstrated at the present, it is reasonable to suppose that the same relationship existed in the past. Therefore the size of fossil animals may often give an indication of the temperature at which they lived.

Species vary in size with temperatures in two basically different ways. Regional differences may be purely the response of individuals to local conditions or size may be controlled genetically, in which case the species rather than the individual has responded by “natural selection’ to its temperature environment. The result in either case can be an increased or a decrease in size with increased temperature. However, in the majority of groups studied (Ray, 1960) size is inversely proportioned to temperature, a relationship referred to as Bergmann's Rule (1847).

The size of individuals is affected by temperature because the rates of chemical processes and hence metabolism and growth are also affected by temperature. Laboratory experiments have shown that such diverse forms as fowl and mice (Harrison 1959) and foraminifers (protozoans) (Bradshaw 1957, 1961) raised at high temperatures are smaller than their brothers raised at lower temperatures. Although growth may be slower in the cooler conditions it is normally continued for a longer period before the animal is capable of reproduction (Bradshaw loc. cit.). In the natural environment any species reacting in this way, which had an extensive, temperature varying range, would show geographic variation according to Bergmann's Rule. In warm waters the New Zealand benthonic foraminifer Nonionellina flemingi (Vella) reproduces at a size and form which would be regarded as juvenile page 63 in cooler regions; a clear inverse size/temperature relationship is apparent for the species (Lewis and Jenkins, in press). If young are transferred to another region they will grow in a form appropriate to their new rather than ancestral environment; for instance Danish sea trout raised at the appropriate temperatures produce isomorphs of Mediterranean or north Scandinavian forms (Taning, 1952). It is probable that many cold blooded animals react in this way.

Other cold blooded animals are larger in warmer areas, which is the opposite to Bergmann's Rule. These are typically terrestrial or shallow water species with an annual breeding cycle, for instance many molluscs. In a warmer climate growth is more rapid so the animal has reached a larger size at the start of the breeding season.

Geographic variation of a species can be fixed genetically by selection of suitable forms in particular areas. This means that young transferred to another area would grow like their parents and not like the native race. However they would not be well adapted to the new environment. Most warm blooded animals have a genetically controlled inverse size/temperature relationship (i.e. according to Bergmann's Rule). Good examples are the Puffin, a European seabird, (de Beer 1964) and the European Wren (Armstrong 1955). The largest Wrens from Iceland are about twice as heavy as the largest specimens from southern England. The reason is obvious when it is realised that body temperature can be regulated by the ratio of heat losing surface to heat producing volume. Surface area increases roughly as the square of the length; volume is related to the cube of the length. Therefore the largest animals have the smallest surface area per unit volume, and use a smaller proportion of their energy maintaining body temperature in cooler regions. When the temperature is near the lowest endurable by the species the largest animals (with the shortest extremities — Allen's (1877) Rule) have the best chance of survival. The local gene pool is modified in favour of larger size and so long as the group is not isolated a gradation (or cline) is produced to small size in warmer regions, where the loss of heat is an advantage.

Once the relationship of size to temperature has been formulated for a particular living species, or group of closely related species, the same relationship can be assumed in fossil populations. It is not surprising that mammals, which are (of course) warm blooded, were larger during the Pleistocene glaciations than animals of the same species living in the same area today. Porcupines, tapirs, and orang-utans from Sumatran cave deposits are larger than the present native races although smaller than fossil forms from China. Fragments from tigers and rhinoceros were also usually large and compared more closely with present Bengalese rather than Sumatran page 64

Fig. 1: Upper Tertiary and Quaternary temperature fluctuations at sea surface in the region of Wellington, based on the size of Nonionellina flemingi (Vella).

races. (Hooijer 1949). It is clear that the cave deposits were formed during a cool phase when climatic conditions in Sumatra were perhaps equivalent to those found at present in Bengal.

A more statistical approach by Lewis and Jenkins (in press) relates the mean length of Upper Miocene to Recent populations of the foraminifer (shelled protozoan) Nonionellina flemingi (Vella) to temperature. The relationship in the Recent is apparently non genetic and mean length of tests is an inverse exponential function of the temperature at which the animals lived. The species has been abundant in muddy shelf sediments around New Zealand since the Upper Miocene. Assuming an identical length / temperature relationship since then it is possible, from the changes of mean length with time in the same area, to elucidate the Plio-Pleistocene climate changes. Some two dozen samples, rich in N. flemingi from the Upper Cainozoic of the Wairarapa and Wanganui Basins were compared. All of the populations have a mean size greater than Recent samples from the same area. It is therefore suggested that all of these populations lived in water cooler than that which exists in the same latitude today. In samples from the Waipipian, Uppermost Waitoraran, Upper Okehuan and Lower Castlecliffian (Fig. 1) the mean lengths were greater than that of the present population on Campbell Island. These measurements are not comparable with Recent data and it can only be assumed that conditions during these times were cooler than they are at Campbell Island today. A cool Waipipian is in agreement with the planktonic foraminiferal evidence of Jenkins (1967). Both are at variance with the Molluscan evidence of Fleming (1953). All workers agree page 65 on cool conditions at some stage in the Hautawan although present evidence suggests that the major cooling was immediately prior to this stage.

All methods of estimating paleotemperature assume a particular process has been constant with time, all require a knowledge of animal ecology and hydrology which is rarely available. The normal paleontological technique for assessing paleo-temperatures is to compare the distribution of ancient forms with the distribution of comparable Recent material. When a whole fossil fauna is compared to a Recent one this is probably the most accurate method available. An element of doubt is obviously included when conclusions must be drawn from one or two “temperature sensitive’ species, particularly when the species are no longer extant. If a fossil species is still living, its present distribution may be restricted for reasons completely unrelated to temperature; for instance shallow water or terrestrial forms which became extinct in New Zealand during the Pleistocene can not recolonise the area from Australia or the islands to the north because an expanse of deep water is an effective barrier. An extinct form can be assumed to have a temperature tolerance within that of living representatives of the species, genus or family. However, tigers, rhinoceros and elephants, all thought of as tropical groups, had large “woolly’ forms adapted to cold Pleistocene conditions. Separate races or species must have been adapted to a different environment, possibly different temperature environment.

An alternative method of estimating palaeotemperatures using oxygen isotopes still involves many practical difficulties including secondary addition of material, recrystalization, processing of the sample (Wiseman, 1966). Salinity as well as temperature will affect the isotope ratio (Shackleton, 1967). Interpretation must involve a knowledge of the ecology of the beast whose shell is being used.

Geographical variation of a species is an inevitable result of variation of environment. Steady gradation (or clines) are usually due to climatic factors particularly temperature; irregular changes to local habitat, food or predators. Obviously local effects can be superimposed on the wider climatic effects. In the case of shallow water marine animals changes of salinity will often have a drastic local effect. Bradshaw (1957, 1961) has shown that extremes of salinity as well as food supply, can have the same effect on foraminifers as reduced temperatures. However abnormal salinities will produce distinctive total faunas. If the climatic effect on a single species is being studied the number of variables must be kept to a minimum and samples from brackish and other unusual environments must be ignored.

The effect of sorting by currents or during fossilization has a greater effect on statistical estimates of size, than it does on page 66 qualitative paleontological methods. With micro - organisms redeposition may also be a problem. These possibilities must be considered, particularly where the evidence from geographical variation differs from that of other methods.

The size of animals is often temperature controlled. Therefore a change in size of a fossil group should not be regarded as simply evolutionary but an explanation should be sought in terms of variation, particularly temperature variation, of the environment. If size is always noted extra information may be gleaned from the available fossil material.

Discussion

Dr D. G. Jenkins. I'd like to congratulate Mr Lewis on his work and note that he also has a cool Waipipian. I'm wondering if the macropaleontologists can confirm this.

Dr. J. Marwick. There is a difference between the Waitotaran and the Miocene but not much between the Waitotaran and the Taranaki Series. The most spectacular effect is the cooling at the end of the Waitotaran. The Waipipian seems if anything warmer than the Waitotaran.

Dr. D. G. Jenkins. It seems that we are out of step. It is possible that the Mollusca are not so sensitive to temperature changes.

Prof. P. Vella. I'm sure that is true Dr. Jenkins!

Mr. J. Grant-Mackie. Mr. Lewis mentioned that the animals in question lived in silty sediments and it is possible that if samples of fossil assemblages are not taken from similar sediments then some current action may have occurred causing preferential removal of the smaller ones.

Mr. K. B. Lewis. There is quite a lot of fine sediment in the ones I have collected but some of the samples came from the Geological Survey and they were already washed. They include some of the Waipipian samples and I don't know how much mud they had originally contained.

Dr. C. A. Fleming. There is often quite a lot of mud but there are some anomalies because the shell beds are largely of coarse elements and there is doubt as to whether they represent the true bottom conditions.

Dr. P. Suggate. I am interested in the question of salinity changes and their effects. As the cooling occurred from the Miocene onwards, at some stage ice would begin to accumulate at the poles and this would affect the salinity of the oceans. There would presumably be some threshold where the change in salinity would have some effect on the animal. This type of effect could also occur during the climatic fluctuations of the Pleistocene.

Mr. K. B. Lewis. Extremes of salinity do have a marked effect on the rate of growth just as temperature does. For some animals examined by other workers the salinity has to be markedly different to get much effect I don.t know in this case what salinity change would be significant.

Dr. D. F. Squires. Could you get some salinity effect by altering the position of the sub-tropical convergence?

Mr. K. B. Lewis. I think the salinity change across the convergence is fairly small — only about one part per thousand. I don't think this would be significant.

Dr. N. deB. Hornibrook. It is possible that during the Pleistocene there was much more rainfall than at present and so the river outflow would have been much greater and so affect the salinity of the inshore waters.

Mr. K. B. Lewis. It is possible I suppose. Drastically lowered salinity will produce the same effect as lowered temperature. Recent material from Manakau Harbour contained Nonionellina larger than those on the open shelf in the same neighbourhood. However salinity is unlikely to affect the size of Nonionellina without producing a distinctly brackish fauna. If large numbers of Ammonia are present the samples must be treated with caution.