130 VARIATION AND DIFFERENTIATION IN CERATOPHYLLUM. 



tion of the comparative variability of a series of metamerically repeated 

 and slightly differentiated organs or characters in an animal form has 

 ever been made. Hence it is hopeless at present to look to this source 

 for confirmatory evidence for the law. On the botanical side the situation 

 is somewhat better. Thus, for example, in a recent study of "Stages in 

 the Development of Sium cicutsefolium' ' Shull (:05) has investigated the 

 relative variability of the successive "nepionic" leaves in this form. 

 In his fig. 1 (p. 9) he gives the frequency polygons for variations in the 

 first eight leaves. These polygons very obviously substantiate his con- 

 clusion that "there is a progressive lessening of the variability from the 

 first leaf onward. " In this case we clearly have a direct confirmation 

 of the law of diminishing variability. Similarly, in the papers of Cush- 

 man (:02, :03, and :04) there are several statements which appear to 

 indicate that in many of the large number of plants studied by this 

 worker the variation diminished in successive nepionic leaves. In a less 

 direct way the operation of this law is to be seen, I believe, in a wide 

 variety of morphological phenomena, including particularly what we have 

 called above "individual" growth. Thus we have one example of it 

 in the well-known fact that embryonic characters are much more vari- 

 able than adult characters. That this is true has been directly proven 

 by an elaborate system of measurements on duck embryos made some 

 years ago by Fischel ('96). 



Minot found the same thing in his guinea-pig measurements. On 

 this point he says (loc. cit, p. 140): 



This diminution of variability with age is demonstrable in the growth of other 

 mammals, hence it probably occurs in all. We ;are led by this to put the question 

 whether all variability of higher animals does not lessen with the age of the individual. 

 In view of the extreme variations of structures which occur in all vertebrate embryos, 

 and which, as all embryologists know familiarly, are far greater and more frequent 

 than the variations of the adult, we are justified in asserting that there is a diminution 

 of variability with age. 



More evidence in the same direction might be cited, but it is unneces- 

 sary. If what has frequently been asserted should be proven by exact 

 measurements, namely, that phylogenetically "young" organs are 

 more variable than "old" organs, this again would fall in line with 

 the general law of diminishing variability. It has recently been 

 shown by Shull (:05a) that mutant forms of (Enothera are more vari- 

 able than the parent form. One might hazard the suggestion that 

 this fact is another expression of the operation of this same law. The 

 matter could easily be tested by a biometrical investigation of mutants 

 and their descendants extending through a considerable period of 

 time. 



