ECTOTHERMS FOLLOW THE CONVERSE TO BERGMANN'S RULE.

In a recent paper, Van Voorhies (1996) suggested that Bergmann size clines in ectotherms might result from developmental processes that cause cells to grow larger at lower temperatures. Van Voorhies (1996) found that when Caenorhabditis elegans was grown at cool temperatures cell sizes were larger than control groups reared in a warmer environment, and that this observation may be generally true for other ectotherms (e.g., nematode egg size and fish red blood cells). This is an interesting observation that may play an important role for intrapopulation phenotypic plasticity in body size. However, I would suggest that temperature induced variation in cell size is not a likely mechanism underlying observed patterns of body size variation along latitudinal and altitudinal clines that have been reported for many invertebrates. There are three lines of evidence to support my proposition. First, in many studies of geographic variation in body size in insects that I have reviewed, body size generally decreases with decreasing mean temperature. For example, in an early review of North American field crickets (genus Gryllus), Lutz (1908) found that populations from high latitudes had much smaller tegmina and femora than their southern counterparts. In studies of several Japanese cricket species (e.g., Teleogryllus emma and T. yezoemma), Masaki and others (see Masaki 1978 for a review) found that body size also decreased at higher latitudes. Crickets from northern sites were generally smaller that their southern conspecifics. Similarly, in the North American striped ground cricket, Allonemobius socius, body size shows a "saw-tooth" pattern, with size generally decreasing with decreasing summer season length within a life-cycle type (i.e., Univoltine or bivoltine; Mousseau and Roff 1989). Saw-tooth body size clines are expected for any Organism capable of adjusting the number of generations per growing season (Roff 1980), with a dip in body size corresponding to the transition from bivoltine to Univoltine life cycles. Curiously, one could erroneously conclude that Bergmann's Rule applied if one only examined body size variation without regard to shifts in voltinism (i.e., in the transition zone, body size shifts from small, bivoltine individuals to large, Univoltine individuals; see Fig. 1). Similar patterns have been observed for other North American crickets by Alexander and Bigelow (1960). In the lesser migratory grasshopper, Melanoplus sanguinipes, body size also decreases with decreasing mean annual temperatures (Dean 1982; Scott and Dingle 1990; Orr 1996). Scott and Dingle (1990) and Orr (1996) found that grasshoppers from high altitudes in the Sierra Nevada's of California were considerably smaller (and had much shorter development times) than populations from valley sites (Fig. 2), while Dean (1982) found a similar pattern of decreasing size with increasing latitude. Thus, there is strong empirical evidence that refutes the Bergmann's Rule paradigm, and in fact points to a converse of Bergmann's Rule for ectotherms, as has been suggested by Masaki and others (Masaki 1978; Roff 1980, 1986; Scott and Dingle 1990). There is a second line of evidence that refutes the applicability of Bergmann's Rule to ectotherms, that also points to the developmental mechanism underlying geographic variation in body size. When reared in a common garden in the lab, body-size variation among populations tends to follow that observed in the field, suggesting that much of the variation observed in the wild reflects genetic differentiation rather than environmentally induced phenotypic plasticity. In a study of 10 populations of Allonemobius socius, lab-reared (after two generations of common garden conditions to remove maternal effects) and field-collected crickets showed essentially the same pattern of body size variation (Mousseau and Roff 1989). Similar findings have been reported for grasshoppers (Dingle et al. 1990; Orr 1996) and several Japanese cricket species (Masaki 1967, 1983). These observations add further support to the notion that ectotherms tend to follow the converse of Bergmann's Rule, and indicate a high degree of genetic determinism underlying this pattern. Such patterns of body-size variation are an expected and predictable consequence of the interaction between season length at a given latitude (or altitude) and the physiological time available for development in an ectothermic Organism (Masaki 1978; and see Roff 1980, 1986 for detailed models). At high latitudes (or altitudes), the growing season is short, and for a Univoltine Organism development is constrained to a single season or less. At lower latitudes (or altitudes), the growing season is relatively longer and an individual's development time can be extended. Given a positive relationship between development time and body size that is frequently observed in ectotherms (e.g., Peters 1983), body size tends to directly parallel development time, and likely leads to the converse of Bergmann's Rule that has been reported for insects. Although in crickets and grasshoppers, body size appears to have a large genetic basis (with northern and high altitude populations being genetically smaller), this does not need to be so: the converse to Bergmann's Rule would follow for any Organism whose life cycle is linked to season length, given a relationship between development time and body size, even if variation in development time reflects environmentally modulated plasticity rather than genetic determinism. A third line of evidence refuting the importance of temperature induced cell size effects comes from studies of rear-