REACTIONS OF THE HORNED LIZARD. IO5 



starvation, as it was difficult to induce the animal to take food 

 under the conditions of the experiment. The weight in this time 

 was reduced to 2 grams. 



The broad limits of the evaporation optimum make it rather 

 unlikely that this factor is the principal one in determining the 

 local distribution of the species. It would not be expected that 

 this would be the case in a reptile from such a habitat, where the 

 evaporating power of the air is subject to such great variations in 

 the course of twenty-four hours. Reptiles, also, are, in general, 

 especially adapted to withstand a considerable amount of drouth, 

 the water loss through the integument being very slight and that 

 through the feces and urine almost negligible, as compared to the 

 similar losses of other animals. 



2. Air Temperature Gradients. (Evaporation Varied by Differ- 

 ences in Air Temperature.} 



Experiments with an air temperature gradient were carried out 

 in the same apparatus as the previous experiments, the air passing 

 through the cage being heated or cooled by passing through coils 

 immersed in water of high or low temperature. The air came 

 directly from the storage tank and was unmodified except as 

 above noted and flowed at the same rate as in the humidity 

 experiments. The experiments (34, a, b, c) summarized in 

 Table II. together with the many others performed in the same 

 manner show the optimum to be between 33 and 38 when the 

 substratum is not cooled, i. e., when the temperature of the sub- 

 stratum varies approximately with the air. When the substra- 

 tum is cooled the optimum more closely approaches 30, which 

 corresponds, under the conditions of the experiment, to air with 

 an evaporation of i.o c.c. for the experimental period. Both 

 humidity and temperature gradients are essentially evaporation 

 gradients and optima are most conveniently expressed in terms 

 of evaporation (Shelford '136). 



