The interactive effects of temperature, food level and maternal phenotype on offspring size in Daphnia magna.

Invertebrate offspring are usually larger in colder environments. To test for possible effects of covariates (e.g. maternal phenotype and feeding conditions) on this pattern, we performed a laboratory experiment to look at the effect of temperature on newborn weight in the planktonic crustacean Daphnia magna. Three tempèratures (12°C, 16°C and 22°C) and two food levels (10,000 cells ml and 150,000 cells ml) were used, and offspring were examined from the first five clutches of mothers that had been maintained under the constant experimental conditions for three generations. Preliminary analysis suggested that newborn weight was significantly affected by temperature although patterns in the data were not clear cut. In addition, the covariates mother weight and clutch size were positively and negatively correlated with newborn weight, respectively; and later clutches tended to contain heavier offspring. Therefore, in an effort to control for the effects of the covariates, repeated-measures analysis of covariance was performed using ratio values of newborn weight/mother weight (relative newborn weight) as the dependent variable, clutch size as the covariate and clutch number as the repeated measures term. Now, temperature as a main effect in an ANCOVA model did not significantly influence relative newborn weight. The repeatedmeasure term clutch number also became nonsignificant, indicating that when differences in mother weight due to age were accounted for there were no overall differences in relative newborn weight between clutches from a particular mother. Temperature effects on relative newborn weight were only significant as part of interaction terms with food concentration and with clutch number. Thus there were different weight responses to temperature within food levels, and between clutch numbers within food levels. Under the low-food conditions newborn were heaviest at 16°C, lightest at 12°C and intermediate at 22°C. Conversely, under the high-food condition newborn were lightest at 16°C, heaviest at 12°C and again intermediate at 22°C. However, newborn tended to be heavier under the low food condition, and food concentration was highly significant as a main effect. Mother growth rate showed no significant relationship with newborn weight. It is concluded that direct temperature effects on relative newborn weight are marginal and nonsignificant. Temperature effects through interactions with food concentration and clutch number are important determinants of newborn weight, but relatively speaking account for only a small proportion of observed variance in newborn weight (25%), compared with the direct effect of food concentration (67%).