28 Records of the Indian Museum. [Vol. Ill, 



Nesokia and Gunomys all show that the range of variation in 

 length is about 40 % of the mean length ; the same is probably 

 true of all small mammals. This is not usually admitted by sys- 

 tematists at the present day, although in 187 1 J. A. Allen measured 

 the length of twent^^-eight adult squirrels and found that this com- 

 paratively small number exhibited a range of over 20 % of the 

 mean [12], 



In the case of Nesokia, however, there is considerable evidence 

 that a pigmy race has arisen or is arising in a certain part of the 

 Amritsar district. Some families of Nesokia were sent from a field 

 near the village of Atari; these included four adults. In length 

 these four measure 146, 147^ 151, 1^5 mm., that is to sa^^ they are 

 all well below the mean. The similarity of the skulls of these four 

 is remarkable. 



The pigmy which measures 134 mm. is, however, from another 

 part of the district. It was speciall}' sent by Davys as it was 

 pregnant, bearing two embryos. The third molars of this specimen 

 are well worn down. Its maturity cannot be questioned. 



Plate ii shows a series of seven Nesokice selected from the 

 twent}^- three to show gradation in length. Beneath them are 

 shown the skulls of the same seven specimens. These skulls ex- 

 hibit not only a difference in size, but a difference in form, for in the 

 two marked S. B. B. i and S. B. B. iS the ridges which' mark the 

 upper limit of the temporal muscles are visible but not prominent, 

 and are set widely apart. In the skull P. C. M. 2 these ridges are 

 more prominent and much closer together ; this is the skull of the 

 largest Nesokia we have received. The rats designated S. B. B. are 

 the small ones from Atari already mentioned. Other skulls are 

 intermediate in form as well as size between the two extremes. 

 The measurements given below show how the small skulls resemble 



ERRATUM. 



In the table on page 29 the designation " S. B. B. 3 " refers 

 to the fifth series of measurements only. 



tor tne aboreviation of the palatine foranien which has obviously 

 been caused by the closing of the posterior half of that space. A 



