RARITY OF MUTANTS. 



273 



be crossed with the parent stock, because similar variations of the other sex 

 are usually wanting. Cultures of these become, then, largely experiments 

 in crossing these variations with the parent species, as they must and will be 

 in nature. 



The variations of this kind observed to arise from decemlineata and the 

 number of specimens are: of melaniciim, 31 ; tortuosa, 3; minuta, 2; immac- 

 ulothorax, 2; pallida, 63; mhrivittata, i; defectopunctata, i; alhida, i, and 

 ohscurata, 4. Some of these I have been able to propagate with success; 

 others, like defectopunctata and ohsciirata, I have been unable to cross suc- 

 cessfully with the parents. It is, however, certain that these variations are 

 a normal product found in nature produced in different ratios — a ratio seem- 

 ingly dependent upon fluctuations in the environment. The point is that 

 these variations are natural and not the result of domestication or confine- 

 ment, and it can not be charged that in this material their appearance is due 

 to cultivation, because the foregoing data are drawn directly from nature, 

 and the material for the following pedigree cultures came from the same 

 source. My material is pure, never having been subjected to the supposed 

 vicious influences of cultivation or domestication. 



It should also be noted here that the same kind of variations were found 

 to arise in the selection experiments. These were especially numerous in the 

 experiments where I attempted to produce extreme racial conditions through 

 changed surroundings, accompanied by selection, and they are also numerous 

 in the other selection experiments as the races diverge from the mean 

 and mode of the parent species; but they are not numerous in any case, 

 excepting where changed environmental complexes have been employed in 

 conjunction with selection. In many respects these results resemble De 

 Vries's experiments with CBnothera, excepting that here we have evidence 

 converging from many sources and pointing strongly to the conclusion that 

 the production of these rare variations is in some way connected with varia- 

 tions in the environmental complexes. This point we shall investigate in a 

 subsequent section of this chapter. 



One of the earliest observed variations of this class from decemlineata is 

 the form pallida, which was used in some of my earlier cultures. (See plate 

 16, fig. 7.) It is not a variety of decemlineata formed by the taking away of 

 characters, but it is a step forward in the line of evolution represented by 

 L. multitceniata, intermedia, and decemlineata — that is, pallida is the next 

 logical step in the orthogenetic evolution of the series of species. This spe- 

 cies has been found in the last ten years many times and in divers parts of the 

 United States, and is not limited to any one region. 



At various times I have been able to secure 2 or 3 and once 6 specimens of 

 pallida at the same time, and I have made experimental cultures of these with 

 interesting results. In August, 1901, I found 6 specimens of pallida at CHf- 

 ton, Ohio. Four of these were males and 2 females ; all were well nourished, 



