177 



production of the mean of the whole. Under these condi- 

 tions heredity tends to uniformity of type if the numbers 

 are large. If, however, the parent stock be separated into 

 parts, which are prevented from interbreeding, or a part of 

 whose progeny becomes less fertile or less adapted to altered 

 conditions, then a new factor, namely, isolation, comes into 

 play, and an evolution or evolutions of type appear as a re- 

 sult. The diversities of these results will depend upon the 

 size of the separated parts. For instance, a parent stock 

 divided into two equal parts by isolation will evolve into 

 two different types because the mean of variability of a-half 

 is different to that of a whole, but the evolution will be slow, 

 because the differences would be slight between the two, 

 other things being equal. If, on tjhe other hand, the 

 separated part is comparatively small, the divergent evolution 

 would be proportionately marked and appear at an earlier 

 date. 



To fully appreciate the effects of isolation, considerable em- 

 phasis is laid on the species of isolation as distinguished from 

 the form of isolation, the former being much more important 

 as an evolutionary factor. In other words, is the isolation 

 discriminate or indiscriminate ; in the discriminate isolation 

 it is a breeding of like with like; in the indiscriminate it is 

 an interbreeding between all the members of a simply 

 separated part. If a breeder divides his nock of sheep into 

 a white part and a black part, it will be a case of discriminate 

 isolation; if simply into numerically two equal parts it will 

 be a case of indiscriminate isolation. Unfortunately, dif- 

 ferent names have been given to these two kinds of isolation, 

 although practically the same thing is understood. For in- 

 stance, discriminate isolation is called by Gulich segregate 

 breeding, by Weissman amixia, and by Romanes homo- 

 gamy ; whilst indiscriminate isolation is called by Gulich 

 separate breeding, by Weissmann panmixia, and by Romanes 

 apogamy. To clearly define what he implies by the use of 

 different terms applied to isolation, Romanes makes the fol- 

 lowing note : — Let isolation be considered a generic term ; 

 species of* isolation will be homogamy and apogamy or isola- 

 tion as discriminate and indiscriminate, and that natural 

 selection may be considered a sub-species of homogamy ; forms 

 of isolation are modes of isolation, such as the geographical, 

 the sexual, the instinctive, or any other of the numerous 

 means by which isolation of either species may be secured ; 

 cases of isolation are the instances in which any of the forms 

 of isolation may be at work ; thus, if a group of intergenerants 

 be segregated into five groups, a, b, c, d, e, then before tne 



M 



