pedigree; cultures. 261 



Two questions which can be solved only by experiment are before us : ( i ) 

 Do new characters, races, and species arise by rapid development only, or do 

 they arise by both rapid development and by slow variation and isolation 

 through extinction? (2) What is the cause of these transformations of char- 

 acters and species? is it external or internal, or both? and can they be pro- 

 duced in experiment ? 



PEDIGREE BREEDING OF L. DECEMLINEATA. 



EXPERIMENTS IN RACE PRODUCTION AND MODIFICATION BY 

 ARTII^ICIAI, SEI.ECTI0N. 



From 1896 to 1904 I carried on continuous and rather extensive experi- 

 ments in selection. In these experiments the beetles were kept under con- 

 stant conditions, especially during ontogeny, and in each generation indi- 

 viduals which showed extreme variations were isolated and became the 

 parents of following generations. This is a comparatively simple process, 

 and one which can without doubt be kept up for a long series of generations. 

 It is these experiments which will now be considered. 



I began with the idea of showing experimentally that where there are 

 closely related species similar in most of their characters, but isolated by 

 gaps, evolution has been through slow progressive modal shifting and the 

 extinction of intermediate forms by selective processes. Thus, for example, 

 in the series multitcBniata, intermedia, and decemlineata, multitceniata is most 

 melanic in general coloration, and decemlineata least. It is possible, by envi- 

 ronmental influence and selection, to change species one into the other ? 



SEi<ECTiON Experiments with Coi<or Characters. 



My experiments with color characters dealt both with special areas of color 

 and with general color tendencies, such as melanism, albinism, xanthism, and 

 rufism. 



I attempted by selection to create a race of decemlineata in which the spots 

 a and a' upon the pronotum would always be fused posteriorly, as they are in 

 some species, and at the same time to increase their size. In this experiment 

 the most rigorous selection was practiced, and every opportunity given for 

 selection to produce the desired result, but to no purpose. In one part of this 

 experiment the selection was carried out in a most rigorous way for eleven 

 generations without any discoverable result (plate 25, A). In each genera- 

 tion the progeny of extreme melanic parents presented, as far as the character 

 in question was concerned, about the normal range of variation. The entire 

 negative result in this experiment may be interpreted as an example of the 

 total inefficiency of selection, or, what is more probable, as indicating that the 

 selected variations of the character were not capable of transmission, being 

 purely fluctuating somatic variations. In the chapter on coloration I have 



