266 PRODUCTION 01^ RACES AND SPi:ClKS IN LKPTINOTARSA. 



resented in plate 28. From the parent generation two selected groups, one 

 melanic (B), the other albinic (A), were taken, and from these, two clearly 

 defined races without trace of intermediate condition were produced. During 

 each of eight consecutive generations slightly variable, light and dark races 

 were maintained. At the end of this time the material was divided and selec- 

 tion was stopped in one group and continued in the other, but the lots were 

 not allowed to interbreed. The removal of the selective factor at once 

 resulted in a regressive shifting of the mode of each unselected race and in 

 increased variability, and this change continued through the eleventh genera- 

 tion, when both unselected lots had moved back to the mode of the species. 



These experiments with color characters show very clearly that artificial 

 selection is with transmissible variations a powerful factor and can greatly 

 accentuate any character and maintain it in an extreme condition, but that 

 there are limits beyond which I was not able to modify the characters by this 

 agency. The experiments also show that artificial selection works rapidly, 

 and not, as has been so often assumed, with extreme slowness. True, in 

 experiment I practiced a most rigorous selection, but not more rigorous than 

 that which the natural selectionists believe exists in nature. The general 

 feature of these experiments I shall postpone discussing until after the data 

 from other selection experiments have been considered. 



The experimental production of general color variations and their preserva- 

 tion by selective breeding give many points of interest. In this I have con- 

 fined my attention almost entirely to extreme light and dark forms. To pro- 

 duce light forms I have used hot and dry conditions, and for dark forms, 

 warm and moist. The conditions productive of these variations have already 

 been discussed and recorded ( Chapter III). The experiments herein recorded 

 differ from those already given in that the entire life of the beetles was passed 

 in the conditions of experiment, and not the larval and pupal stages alone, as 

 in the experiments upon coloration. 



In plate 29 are brought together in diagrammatic form the data and general 

 history of cultures where both light and dark forms were produced and 

 further subjected to experiment. The black polygons represent the selected 

 groups of parents, the ruled polygons the offspring. The appearance of 

 "mutants" beyond the normal range of variability is indicated by the small 

 white polygons, pa = L. Pallida, with its range of variation ; de = defecto- 

 punctata; ni = L. melanicum. The series is a complex one, involving pro- 

 cesses other than artificial selection and introducing factors of interest which 

 are the key to further experimental study. At present we shall consider 

 only that portion directly concerned with selection or selective processes. 



In the first, or parent, generation I selected 6 copulating pairs of beetles 

 from the hibernating population, and kept them and their progeny in natural 

 conditions. From the 6 pairs were obtained in the second generation 1,320 



