Fifteen methods of maintaining control populations have been studied over eight generations using the flour beetle, Tribolium castaneum. Populations were reproduced each generation from 50 males and 50 females. The methods were compared as to their ability to establish the level of the environment with respect to the base population and to separate environmental and genetic effects in two directionally selected lines.

It was demonstrated that a foundation stock of these beetles produced pupae which weighed about 10% more when grown in 70% relative humidity than when grown in 40% relative humidity. All lines were grown in one environment for two generations and then in the other environment for the next two generations. The mean pupa weights of two master controls were averaged to give an average master value which served as a standard against which the other controls were compared.

Inbred lines derived from a separate source of stock responded differently to the environments than did the foundation stock from which selected lines were drawn.

Evidence of genetic drift in mass mated lines was presented. No differences among the other methods of maintaining the original population were observed. It was suggested that if a smaller number of families had been used, additional differences might have been detected.

Response to selection in both directions over eight generations was symmetrical. Symmetry was also demonstrated for within environment regressions and it appears that greater progress was made in the dry environment than in the wet.

Initially the foundation stock weighed more in the wet environment than in the dry environment. After selection, the large line weighed more in the dry than in the wet. It was shown that the original population contained the basis for this interaction of directionally selected lines with environments since some ofthe initial famllies weighed more in dry than in the wet environment. It was suggested that the apparent differential response to selection, dependent upon the environment, may have been due to the larger variance in the dry environment. This variance was shown to be about twice as large as that in wet.

Due to the presence of the interaction, the base control populations were ineffective in separating genetic and environmental effects in the later generations of the selected lines. Repeated and relaxed methods of maintaining controls more closely indicated how environmental shifts affected the later generations of the selected lines.

It is concluded that a control line must be closely related to the selected line in origin and time in order for its reactions to environmental shifts to be similar in nature to the selected line. If these conditions are not met, undetected genotype by environment interactions may contribute to faulty comparisons. In addition, some method of reproducing the original population must be followed if interactions with respect to the origin are to be detected.

It is further concluded that it is unnecessary to place restrictions on the method of mating base populations for the purpose of maintaining control stocks, provided that the size of the breeding population is not less than 50 males and 50 females and that mass mating is not used.