( 445 ) 



divergent development of the variety we will assume (1) tliat a number of varietal 

 specimens are completehj isolated from the rest of the species ; (2) that 80 per cent. 

 of the oifspring of these specimens belong to the varietal form. The specimens 

 now have to propagate under the further premiss (3) that the normal and the 

 varietal forms exist under exactly the same conditions of life, so that every other 

 transmuting factor besides mechanical isolation is excluded. By mechanical 

 isolation we understand a sei)aration of the animals or plants in question by a 

 mechanical barrier, so that an intercrossing with the original stock is prevented; 

 experimentally the case could be demonstrated by rearing wingless animals side by 

 side, but separated by an adequate fence. Under the above premisses ten femalrs of 

 a variety, each producing twenty females, kept in an enclosure rejiresenting the 

 isolated locality, would give birth to a hundred and sixty females of the varietal 

 (T') and forty of the normal form (iV). If the latter produce also each twenty 

 females, of which 80 per cent, might be taken as normal and 20 per cent, as 

 varietal, and if the locality is fit to provide food for a thousand females (and a 

 thousand males, which are not taken into account), the numbers of both forms would 

 in the succeeding broods be as follows : — 



The result is here again that after a small number of broods both the varietal 

 and normal forms will exist in equal numbers in the isolated district. The varietal 

 form can never become the sole inhabitant of that district unless the circumstances 

 of life are such that the normal form is less favoured b_v' them, i.e. unless there is 

 some transmuting factor active besides isolation. Isolation as such is not an 

 active factor which produces a character, but is a factor which merely preserves 

 a character produced by some other factor ; isolation has, therefore, no direct 

 effect. The reason that the effect of isolation has by many authors been so much 

 overestimated is so obvious that we scarcely need mention it : the differences exhibited 

 by geographically isolated forms of the same species are often attributed to the direct 

 effect of isolation, because isolation and morphological difference were seen to be 

 always associated, while no other transmuting factor seemed to be obvious to those 

 who were unacquainted with the experiments made in this direction. Apart from 

 experiments, there are many geographical races which, on closer e.xamination, show 

 at once that their characters cannot be the outcome of the isolation of some ancestral 

 specimens which accidentally exhibited the respective distinguishing character. 

 Wallace * was the first to draw attention to a peculiarity common to a great many 

 species of butterflies on the island of Celebes : these species or varieties have much 

 longer and more falcate forewings than the races from the other islands of the Indo- 

 Australian region. On the island of Sumba, which lies south of Flores, that 

 character is also found in some species. The Chinese races of butterflies and moths 

 have generally tlie black colour more extended than the respective Indian races. 

 The butterflies and moths of Sumatra and Borneo t are mostly much darker than the 

 races of the same species from Malacca and Java. The Queenslandian races are 

 often pale, those from the Kci Islands have the markings often restricted, and so on. 



* Proc. Linn. Soe. Land. XXV. p. 18. 

 t Hagen, /rw 18W. p. 17. 



