604 Wm. E. KelUcott 



may be true for such characters as the weight of the spleen or 

 liver it would be less true for such as the weight of the heart or 

 brain or length of the alimentary canal. And farther, the values 

 of many external characters are easily subject to modification by 

 the general metabolic condition of the organism, particularly 

 during the early period of growth (/. g., Vernon ('95)) and that 

 many of these modifications may be overcome by compensatory 

 growth should conditions of life change. The second sugges- 

 tion carries more force. It amounts practically to saying that at 

 present we have no means of measuring the actual value of these 

 internal characters which approaches accuracy. A measurement 

 of the mass of the liver, for example, gives no exact information as 

 to its functional worth. There is no legitimate reason for suppos- 

 ing that there is an exact ratio between the size of a viscus and its 

 functional value, indeed the ratio between mass and efficiency 

 may be inverse and the better the tissue is functioning the smaller 

 need its mass be to carry on its work in the life of the organism. 



In the statistical study of the characteristics of animals it 

 should be borne in mind constantly that it is the functional value 

 of a character to its possessor which is the bearing point of natural 

 or any other form of selection. It makes no difference to a toad 

 how long his legs may be but only how far or how fast he can 

 jump. If the length of the legs or their segments is an exact indi- 

 cation of their ability to function, then only are we justified in using 

 their lengths as data of actual evolutionary significance. Simi- 

 larly, to the individual toad the weight alone of the liver is a matter 

 of no consequence, only its ability to secrete and metabolize in 

 both qualitative and quantitative relations, and unless we can 

 demonstrate a close relation between bulk and efficiency, again the 

 data themselves do not afford material for study of the evolution- 

 ary significance of its variability in weight. 



That there is a very close relation between the dimensions of 

 external characters and their functional value is probably true but 

 the fact of their low degree of variability is not sufficient alone to 

 prove it. The segments of the legs, for example, form a system of 

 levers whose action depends largely upon their relative lengths and 

 proportions. Their value lies almost wholly in their purely mechan- 



