Royal Society Publishing

Gigantism, temperature and metabolic rate in terrestrial poikilotherms

Anastassia M Makarieva, Victor G Gorshkov, Bai-Lian Li


The mechanisms dictating upper limits to animal body size are not well understood. We have analysed body length data for the largest representatives of 24 taxa of terrestrial poikilotherms from tropical, temperate and polar environments. We find that poikilothermic giants on land become two–three times shorter per each 10 degrees of decrease in ambient temperature. We quantify that this diminution of maximum body size accurately compensates the drop of metabolic rate dictated by lower temperature. This supports the idea that the upper limit to body size within each taxon can be set by a temperature-independent critical minimum value of mass-specific metabolic rate, a fall below which is not compatible with successful biological performance.


1. Introduction

The relationship between metabolic rate and body size has been predominantly studied in terms of body size as independent, and metabolic rate as dependent, variables and not vice versa, e.g. Peters (1983). However, several studies suggest that certain critical values of metabolic rate may set limits to animal body size. For example, in a study of unicells, poikilotherms and homeotherms, Robinson et al. (1983) noted the smallest representatives of each group have a nearly uniform mass-specific metabolic rate. This suggests that the smallest size within each group can be dictated by this uniform value of mass-specific metabolic rate. Geiser (1988) found that mammals hibernating at low body temperature reduce their metabolic rate down to approximately 0.1 W kg−1 irrespective of body size. Singer et al. (1993) suggested that the maximum body size in mammals is prescribed by this minimum value: as far as mass-specific metabolic rate decreases with growing body size, no further growth of body size is expected when this critical value is reached. Makarieva et al. (2003) proposed that the maximum amount of metabolically active biomass of plants attainable at a given ambient temperature and solar irradiance is similarly dictated by a minimum temperature-independent mass-specific metabolic rate qmin compatible with viability of living tissues.

In this paper we report evidence which further supports the idea that the upper limit to body size within each taxon can be set by a temperature-independent critical minimum value of mass-specific metabolic rate qmin, a fall below which is not compatible with successful biological performance. Mass-specific metabolic rate decreases with increasing body size but, in poikilotherms, grows with ambient temperature. Compensation of the size-related drop in mass-specific metabolic rate by higher temperature extends the permitted range of body sizes for which qqmin. Hence, the maximum body sizes attained by species inhabiting warmer environments should be larger than the maximum body sizes attained by species from the same taxon but living at lower temperatures. Here we test these predictions by analysing body lengths of the largest representatives of 24 poikilotherm taxa from the tropical, temperate and polar environments.

Investigation of maximum body sizes across diverse taxa and climatic zones requires extensive faunistic descriptions of the studied areas. We chose Great Britain and Colorado, USA as two well studied sites in the temperate zone and Wrangel Island (71°N, 179°W, Russia) as a representative polar territory (the whole island (22 257 km2) is occupied by state nature reserve with relatively well studied fauna). As far as species lists for tropical countries are relatively fragmentary, we identified world's largest species in each of 24 taxa and determined their geographic ranges, to find that all of them inhabit tropical areas. For each region we calculated mean daily temperature T of the six warmest months presumably corresponding to maximum activity of poikilotherms. For the tropics, T was averaged over typical locations of the investigated species. For Wrangel Island, T was calculated as the mean for June, July and August, the only three months with mean temperature above 0°C. Temperature data were taken from Landsberg (1969–1984).

2. Results and discussion

Our analysis revealed that by linear body size, table 1, the largest tropical species of terrestrial poikilotherms exceed the largest representatives of the same taxa from Colorado, Great Britain and Wrangel Island by 2.3, 2.9 and 6.1 times, respectively. Body size change parallels the progressive decrease in mean ambient temperature: T=26°C for the tropics, T=18°C for Colorado, T=14°C for Great Britain and T=2°C for Wrangel Island.

View this table:
Table 1

Maximum body size of terrestrial poikilotherms living at different ambient temperatures.

Within poikilothermic taxa mass-specific metabolic rate q grows with decreasing body size, but declines with decreasing temperature, Embedded Image, where M is body mass, T0 is reference temperature, Q10 is typically 2–2.5 and α is typically in the vicinity of 1/4 or 1/3 (Peters 1983). At a given temperature the largest species feature minimum mass-specific metabolic rate qmin. If qmin is universal for the taxon and independent of temperature, qmin=const, then for the largest species with body masses M1 and M2 living at temperatures T1 and T2 we haveEmbedded Image(2.1)

Assuming that body shape is conserved within a given taxon, LM1/3, we obtain the following ratio for maximum linear body sizes L1 and L2 found at temperatures T1 and T2:Embedded Image(2.2)where ΔTT1T2.

Theoretically predicted ratios RTH calculated using mean representative values Q10=2.3 and α=0.3 for the temperature differences ΔT between the studied areas agree well with the observed mean ratios ROB. For example, the largest terrestrial poikilotherms in Colorado are on average ROB=1.42±0.13 (s.e.) times longer than their counterparts in Great Britain (averaging is done over n=17 taxa studied in both Colorado and Great Britain, table 1). Temperature difference between Colorado and Great Britain is ΔT=4°C, which gives Embedded Image. Relative difference δ=((RTHROB)/RTH)×100% between the theoretical and observed value is δ=+2%. Corresponding figures for comparison between the other geographic regions studied are given in table 2. The discrepancy between theoretical and observed values ranges from −14% to +31% and is larger for comparisons involving fewer taxa. For the three comparisons involving more than 15 taxa, the discrepancy between the observed and theoretical values ranges from −14% to +2%, table 2.

View this table:
Table 2

Differences in maximum linear body sizes versus differences in environmental temperatures in terrestrial poikilotherms.

The obtained theoretical ratios RTH between body lengths of the largest species of a given taxon living at different ambient temperatures can be independently tested by involving geographic regions different from those studied in the present paper. For example, the Antarctic Peninsula is situated at approximately the same latitude as Wrangel Island and features comparable temperatures. Hence, the prediction obtained for Wrangel Island that the largest polar species should be on average 7.7 times shorter in body length than the largest tropical species from the same taxon should apply to the Antarctic Peninsula as well. If, e.g. the largest tropical mite is about 16 mm in length, table 1, the largest mite on the Antarctic Peninsula should be able to reach about 2 mm in body length. This prediction is excellently confirmed by Alaskozetes antarcticus, the largest free-living mite on the Antarctic Peninsula (Block & Convey 1995).

The agreement between theory and data indicates that in the largest terrestrial poikilotherms from different climatic zones the expected decrease in metabolic rate caused by lower ambient temperature can be fully compensated by their smaller maximum body sizes. This suggests that the upper limit to body size within taxa can be set by a critical temperature-independent minimum of mass-specific metabolic rate qmin, which prohibits attaining larger size at lower ambient temperatures (Singer et al. 1993; Seebacher et al. 1999; Makarieva et al. 2003).

Possible uniformity of qmin across different taxa, i.e. whether the largest representatives of different taxa in different climatic zones feature similar mass-specific metabolic rates, warrants investigation. Singer et al. (1993) found that mammals, independent of body size, do not tolerate a decrease of mass-specific metabolic rate below approximately 0.1 W kg−1. It is interesting that the largest African centipede with measured metabolic rate (Cormocephalus morsitans, body mass 3.7 g, T=20°C) has a resting metabolic rate of approximately 0.3 W kg−1 (Klok et al. 2002), which coincides by the order of magnitude with the critical qmin value for mammals. For comparison, resting metabolic rate of males of the largest Antarctic tick Ixodes uriae (body mass 7 mg, T=5°C) is 0.22 W kg−1 (Lee & Baust 1982) while resting metabolic rate of one of the world's largest frogs, the African bullfrog Pyxicephalus adspersus (body mass 1 kg, T=20°C) is 0.14 W kg−1 (Loveridge & Withers 1981). These values for the largest representatives of taxa characterized by strikingly different body sizes and environmental temperatures are remarkably similar to each other and to the qmin value for mammals.

Oxygen concentration in the air is over an order of magnitude higher than that of water-dissolved oxygen. This means that per unit exerted drag force of the ventilatory muscles aquatic animals are unable to inhale as much oxygen as do air-breathing animals. Thus, even if at the microscopic scale the assimilation of oxygen by body cells of aquatic and terrestrial organisms is equally rapid (i.e. independent of the ambient oxygen concentration), metabolism of aquatic animals can be nevertheless limited by low oxygen concentrations in their environment due to the higher cost of delivering a unit oxygen mass into the organism via the body–environment interface. This can explain the difference in gigantism patterns between our study (larger poikilothermic giants at higher temperatures) and the study of benthic amphipods (Chapelle & Peck 1999), where largest body sizes were observed at lowest ambient temperatures associated with highest concentrations of dissolved oxygen. In terrestrial poikilotherms, the rate of oxygen uptake from the environment likely becomes a limiting factor during periods of maximum activity (e.g. flight), when metabolic rates in the regime of oxygen balance can exceed 500 W kg−1 and oxygen demand is very high (Harrison & Lighton 1998). (Note that the fact that such high rates have never been observed in oxygen-balanced aquatic animals unambiguously points to limitation of metabolic rate by low oxygen concentration in aquatic media.) Gigantism of extinct winged insects can be therefore related to atmospheric hyperoxia (Dudley 1998; Harrison & Lighton 1998). However, how elevated oxygen concentration translates into the observed gigantic insect sizes has not been theoretically quantified. Our study provides a quantitative tool for analysing higher ambient temperature as another factor possibly responsible for gigantism in extinct air-breathing poikilotherms. Under otherwise similar environmental conditions elevation of ambient temperature by 10°C could bring about a several-fold rise in maximum linear body size depending on the characteristic Q10 value for each taxon.


With sincere gratitude we acknowledge the invaluable help in identification of the largest animals from different taxa and geographic zones willingly provided to the authors by A.B. Babenko, A.S. Baker, A.D. Barber and the British Myriapod and Isopod Group, O.E. Berlov, P. Convey, S.A. Corbet, S.P. Hopkin, I.M. Kerzhner, V.I. Kuznetsov, J.A. Marshall, A. Moreno, G. Ramel, D.E. Walter and Z.-Q. Zhang. R. Poulin has generously shared with us his unpublished dataset on body sizes in ticks. Special thanks are due to T.I. Blumental and Joan and Ron Engel for facilitation of data access and C. Loehle for commenting the earlier version of this manuscript. This work was partially supported by US National Science Foundation (B.L.L.) and Russian Science Support Foundation (A.M.M.).


    • Received January 4, 2005.
    • Accepted June 22, 2005.


View Abstract