Effects of temperature on rates in poikilotherm nervous systems: Evidence from the calling songs of meadow katydids (Orthoptera: Tettigoniidae:Orchelimum) and reanalysis of published data

Abstract

1. Male calling songs of 10 species ofOrchelimum were tape-recorded 9 to 52 times over ranges of ambient temperatures as great as 20 °C. 2. For each species the relation between wingstroke rate and temperature was approximately linear. Calculated regression lines had coefficients of determination (r2) between 0.76 and 0.99 (aver.r2=0.93). The slopes of the regression lines varied from 0.9 to 5.1, but when extrapolated downward toŷ=0, the lines tended to converge at 11 °C (x±SD=10.9±1.9). 3. These results are similar to those published for 27 species of crickets and known for 20 additional species of katydids. 4. In previous studies, by others, the effects of temperature on rates within poikilotherm nervous systems have generally been reported asQ10 values. Such values are appropriate if rates approximate an exponential function of temperature. However, reanalysis of 15 published data sets that were adequate to justify a comparison showed that 13 fit a linear model more closely than an exponential one. Furthermore, if one half a data set was used to predict the other half, the linear model was superior to the exponential model 26 of 30 times. The temperature ranges of the 15 data sets were as small as 12 and great as 26 Celsius degrees. 5. For those 13 data sets with anr2>0.90, the regression lines tended to converge like those of katydids and crickets. For seven species of electric fish the values ofx whenŷ (discharge rate)=0 clustered about 4 °C (x±SD=4.0±1.8 C). 6. Empirically the linear model is superior to the exponential one in (1) goodness of fit, (2) prediction of effects beyond the temperature range studied, and (3) prediction of data sets from a single or no datum. 7. The origin and slope of the linear regression are generally more useful thanQ10 values as short-hand expressions of temperature effects on rates in poikilotherm nervous systems. 8. Rates in poikilotherm nervous systems are more complex in their determinants than are rates ofin vitro chemical reactions. It is therefore not surprising that the former rates fail to show the simple exponential relations to temperature that the latter show. On the other hand, that poikilotherm nervous systems should generally have approximately linear relations to temperature and that the extrapolations of linear regression lines should tend to converge at zero rate are unexpected and reveal a need for more study of underlying causes.