iQog.] R. E. lyLOYD : The Races of Indian Rats. 91 



writer does not think that Dr. Hossack could always select such a 

 series of nine, as he has figured, from any 100 chance-taken rats of 

 Calcutta. It is true that a similar series cannot be selected from 

 nearly 2,000 specimens of AIus rattiis from different parts of India. 

 However, a similar series could be selected from 58 rats from Simla. 

 And yet these nine are " the total bag for one day." The ex- 

 planation seems obvious : this ' ' bag ' ' was made in one house or 

 granary, and the contents of it represent a parallel case with the 

 Ayapata family group. 



lyCt us consider how similar these two cases are. Rats with 

 dark {i.e., any colour but pure white) under parts occur in vast 

 majority throughout northern India, but white-bellied rats are 

 common in certain places ; Calcutta is one of these. It may be safely 

 assumed that the white-bellied type was derived from the dark- 

 bellied type, and that the Ayapata race with their bicoloured tails 

 were derived from the rats which possessed unicoloured dark tails. 

 It seems to the writer that in the groups of the eight Ayapata 

 rats and of the nine Calcutta rats, — groups which cannot each 

 represent more than two or at the most three generations, — we 

 can see the change from one race to another actually in progress. 

 In both cases the change is associated with a variation of much • 

 wider range than that which is spoken of as normal variation. 



It is not claimed that this assumption is already far justified by 

 facts. Such communities might be looked upon as groups of hy- 

 brids between two established species. These two cases, however, 

 wonder fuU}^ resemble a case recently described from another part 

 of the animal kingdom, which cannot be explained in this way. 



In a certain part of the Arabian Sea a community of the crus- 

 tacean Squilla [16] was found. This community shows an extra- 

 ordinary variabilit}^ in one respect. Species of the genus Squilla 

 possess as a rule six-spined claws, the number of spines being very 

 constant for each species. The members of this particular community, 

 however, all possess a greater number of raptatorial spines, the 

 number varying from 10 to 19, among the seventeen specimens ob- 

 tained. The number of spines is so variable that there are eleven 

 ' ' types ' ' among the seventeen, some of them being unsymmetrical, 

 but the number of spines in every individual is well above the 

 number found in any other species of the genus. Except for these 

 spines the specimens resemble one another as closely as the members 

 of a species usually do. It seems that there are certain definite 

 resemblances between this crustacean community and the. two com- 

 munities of rats which have just been referred to. In all three cases 

 there is a local community differing from the general group of which 

 it is a part, in that it possesses a certain character (presumably a 

 newly acquired one)— albiventralism and caudal bicoloration in 

 the case of the rats, multispinity in the case of the crustacean ; but 

 in all three that character is in a most unstable condition, so that, 

 as regards it, each community contains several types. At the same 

 time there is little doubt that in each case the members of the com- 

 munity are of one stock and are not separated by more than two 



