MEAN TEMPEEATUEB. 103 



adapt themselves to it without any change of structure being 

 the inevitable result, although such changes are clearly recognis- 

 able in function ; and, in the third place, that the same degree 

 or variation of temperature affects different organisms in 

 different ways. We might perhaps be disposed to assume, on 

 the ground of theoretical conjecture, that all the animals living 

 together in the same climate must be affected in the same ' 

 manner by the normal variations of its temperature ; but such 

 an assumption would be, as everyone knows, altogether false. 

 On the contrary, well-known facts tend to show that there are 

 enormous differences in this respect, and the same facts teach us 

 at the same time, that the well-being of animals living in associa- 

 tion depends far more essentially On the variations and extremes 

 of temperature than on the absolute degree of heat to which 

 they may be simultaneously exposed at any given time. 



The results thuS laid down, somewhat dogmatically perhaps — 

 but the reasons on which they are founded will be given pre^ 

 sently — -justify us (only hypothetically, it is true, for the pre- 

 sent) in denying the value frequently attributed, even quite 

 lately, to the curves of tempei'ature as constructed by meteoro- 

 logists. Annual isothermal (isochimenal or isctheral) curves 

 are constructed by estimatin'g the mean temperature for the 

 days first, then for the weeks, months, seasons, or for the whole 

 year ; but these curves calculated from mean temperatures are 

 in truth, if of any, only of very small importance to the matter 

 in hand. Thus, for instance, it is certain that a given degree 

 of temperature, conceived of as absolute in its effects, will have 

 a favourable effect on one animal, while on another it is less 

 favourable or even injurious. Now the mean temperature of 

 a day, as calcidated by the meteorologist and assumed as the 

 basis of all his curves, can afford no standard by which to 

 measure the iniluence of the heat dm-ing that day, since it is 

 not the same as any of the different temperatures observed 

 during the day, and that mean of temperature may be the re- 

 sult of very dissimilar extremes. A pond-snail is developed, 

 lives, and feeds best in a mean temperature of about 20° centi- 

 grade ; but this, as a daily mean of heat, might be the mean of 

 two extremes lying far apart from each other. This water-snail 



