204 COI.ORATION IN LE^PTINOTARSA. 



to experiment; and likewise lots were taken from the control and placed in 

 experiment, and subsequently returned to control. In this way a complete 

 check was kept on the experiments. In text-figure 17 are represented the 

 generations experimented upon and the proceedings followed with each. 



Experiment 2"] was divided into two parts — the experiment proper {2ya) 

 and the control (2yb). In 2ya the beetles were subjected to the conditions of 

 experiment during ten lineal generations, with results shown in text-figure 

 18. A maximum deviation in coloration was produced at once toward a me- 

 lanic state (class 15), from which there was no deviation either above or 

 below in the succeeding generation. In the third generation of 27a the prog- 

 eny were divided into two lots of equal size, one of which was kept in the con- 

 ditions of experimentation, and the other returned to normal conditions. This 

 second lot, known as 2ya^, after being bred during four generations in nor- 

 mal surroundings, was further separated into two portions, one of which was 

 still kept in normal conditions as 2'/a^^, while the other was returned to the 

 conditions of experimentation as 2ya^^. When the beetles in 2ya^ were 

 returned to normal surroundings they at once resumed their natural characters 

 (see text-fig. 1 8), and did not deviate therefrom, during the four generations of 

 2ya and the three of 2ya^^, or seven in all. However, the effect upon 2ya^^ 

 of being returned to the conditions of experiment was an immediate return to 

 the maximum melanic tendency before observed. From the sixth generation 

 in 2ya another lot of beetles, 2ya^, were taken and reared in normal conditions, 

 with the result that they also immediately reverted to the parental condition, 

 and the same was true of 2^0^ in the ninth generation. In experiment 2yb 

 there appears a slight oscillating variability (text-fig. 18), which, however, is of 

 no consequence. In the second generation 2'yb was likewise separated into two 

 lots of equal size, one of which, 2'/b, was retained as a control, while the other, 

 2yb'^, was placed in the conditions of experimentation for four generations, 

 and later in the seventh generation returned to control with 2yb. The effect 

 upon 2yb'^ was an immediate production of the maximum melanic condition 

 (class 15), which was retained throughout the four generations of experi- 

 mentation, and lost only when 27^^ was returned to control. 



In experiment 2y there was no artificial selection, all imagines being allowed 

 freedom to mate and breed as in nature ; hence the only selective influences 

 present were those exercised in the mating of the beetles and by the condi- 

 tions of the experiment, which eliminated a small percentage. 



From the data of this experiment the following conclusions are drawn : 



(i) A deviation in an environmental complex at once causes the polygon 

 of somatic variation and the modal class to shift as far from the normal as it 

 can go under the given condition, and keeps them there until there is a return 

 to the normal environmental complex, when the somatic variations also at 

 once return to their normal state. 



