308 PROBLEM OF THE ORIGIN OF SPECIES. 



parental characteristics. In Lcptinotarsa these variations, which came from 

 the various species of the genus, are orthogenetic. 



The question whether or not we shall consider these extreme variations as 

 fundamentally different from those nearer to the parental mean is crucial. 

 While I have no basis for judging the nature of these extreme variations 

 ("mutants") in plants, those which have appeared in my cultures are all 

 the results of responses to stimuli acting upon the germ plasm, as shown in 

 Chapter V, and, as far as I am able to discover, the)' differ from smaller varia- 

 tions only in the amount of the deviation from the mean of the parents. All 

 of my "mutants" are found to deviate from the parents most in one character 

 and to a far less degree in others. I am convinced, therefore, that the one is 

 the real permanently changed character, and that the less strongly marked 

 changes are due to correlation in variability. This point I can not at present 

 be positive of, because it needs sufficient material for adequate quantitative 

 treatment, and an attempt to make positive assertions on any other basis is not 

 warranted. It offers, however, a most plausible explanation of the fact that 

 mutants differ from the parent in several characters. 



There is no reason for regarding the "mutants" of Lcptinotarsa as other 

 than responses to stimuli, which, acting upon the germ plasm, produce the 

 variations found. I have demonstrated in Chapter V that all inheritable 

 variations behave alike, and in no case is there any evidence that there is a 

 fundamental difference between "mutants" and any other heritable variation. 

 The mutations in Lcptinotarsa are solely responses to stimuli, orthogenetic 

 (Chapter II), and are distributed in their distances from the parental mean in 

 full conformity with the law of the distribution of error. 



In text-figure 31 I have represented the distribution of the heritable varia- 

 tions about the mean of Lcptinotarsa dccemlincata. This figure is based upon 

 the inheritable variations of general color (albinism and melanism), which 

 carry in correlative variability other characters, and with this distribution is 

 given the occurrence of the "mutants" in their order and degree of removal 

 from the parental mean. To the left of the polygon are the variations in 

 increased coloration, melanism, with which are also correlated increased size, 

 etc., and to the right, albinism, with the correlated variation in decreased size 

 and punctation. The polygon is based upon one character, and, although 

 through correlation other characters also vary, it represents the general dis- 

 tribution of permanent variations about the mean of the species. It is shown 

 that in nature we find about one of these "mutants" in every 6,000 (5,447) 

 beetles examined. So when by chance in any lot of variates one of these 

 extreme variations is found, we ought not to jump to the conclusion that it is 

 not a part of the normal variability, because it is. To determine the distribu- 

 tion of the variations in a few hundred variates is perhaps useful, but it 

 represents only the distribution of the variations on the average, and not the 

 distribution of variation in the long run. 



