PE:dIGR^e: BRi:E;DING. 2^^ 



From lot 2b, continued in the sixth generation, i got of B 2h' , L. pallida, 

 12 male and 14 female ; B 2h" and L. decemlineata, 41 male and 43 female ; 

 and from B 2a, after hibernating, I reared a sixth generation of L. pallida, 

 82 male and 61 female, which were all killed by a fungus disease during the 

 pupal period of the following generation. 



These experiments with L. pallida cover a wide range of phenomena and 

 bring forth many points of interest. Some, especially the data bearing upon 

 heredity, we shall touch upon only slightly in this series of papers, and 

 only as far as it bears directly upon the main point of investigation — the 

 origin in evolution of new characters and species. If, therefore, I seem to 

 pass lightly over evident facts and interesting experiments instructive from 

 the standpoint of heredity, it is not because I am unmindful of them. I shall 

 present the evidence gathered concerning heredity in following papers. 



I can give a clearer and more diagrammatic idea of the nature of L. pallida 

 and the results obtained in these experiments by the aid of text-figure 19. 



On the basis of these cultures of L. pallida it is shown that it behaves in 

 crossing exactly as we should expect independent and isolated species or ele- 

 mentary species to do. Without selection it remains true to type, differing 

 thus in fundamental attributes from races and many of the so-called discon- 

 tinuous variations. Moreover, pallida behaves like an independent species 

 when reared with decemlineata in a general culture where free hybridiza- 

 tion is possible, and we see that it not only is not obliterated, but eventually 

 becomes even more numerous than decemlineata. I see no other possible 

 conclusion than that pallida is a true elementary species, developing suddenly 

 from decemlineata, and behaving from the start as an independent specific 

 unit. 



The question may be raised that if pallida is so well able to stand alone, 

 so well able to increase even when there is opportunity to cross freely with 

 the parent species, why do we not find it becoming estabHshed in nature? 

 The answer to this question would be difficult to frame did we not know well 

 the natural history of the material. I have shown that, taking all the nine 

 variations of this class, we find only one of them in 6,000 cases, excepting 

 under rare circumstances, so that we may fairly expect to find a particular 

 variation only once in 54,000 cases. Pallida, however, is the most common 

 one, occurring once in every 5,000 cases. Suppose one to arise in nature. 

 I have shown that there is close selective mating in decemlineata, and that 

 in confinement the chances are 7 to i against pallida mating v/ith decem- 

 lineata. When we add to these conditions the great mortality during hiber- 

 nation which is especially fatal to extremes of variability, the probability of 

 a single variation of this class being able to propagate itself is so remote as 

 to become a real impossibility. It is only under rarely realized conditions, 

 when, as at Cabin John Bridge, variations are for some reason especially 

 numerous, that they would perhaps be able to get the necessary foothold to 



