II. A NOTE ON THE VARIATION OF A LOCAL 



RACE OF EPIMYS RATI I i - RATTl S 



J ARAN (BONHOTE), i ROM PULAU JARAK, 



STRAITS Of M vL.U CA. 



By H. C. Robinson, C.M.Z.S. 



In an earlier number ol vol. i, pp. 70, 71 



119051. Mr. j. L. Bonhote described this rat on a single 

 specimen obtained by me ii 1904. 



The mth 1 1 ■ ,:. rded it as a rao t the Sumatran 



I rttink), but the acquisition ol lar.ye series of 



closely allied foi ud and from various 



groups of islands show that it is rather to be regarded as 



a form of the cosmopolitan E. rattus. 



In view of the fact that Pulau Jarak is very isolated and 

 is practically never landed on, 1: 31 mm h as it possesses no 

 beach and is steep to right up to the masses of granite 

 boulders that form the shore, it is probable that the local rat 

 population is hardly, it ever, contaminated by the introduction, 

 whether by man or by natural agencies, of fresh blood. The 

 race has therefore, in bility, had time to attain 



a position of more murium and I have 



therefore compiled the follow ins; tables based on a very consi- 

 derable series obtained during two 01 three days in April, 1915. 



The specimens wi by two natives and were 



measured by them, but 1 ( ought it advisable to 



submit their* figures to analysis as the personal error is 

 probably large nstant a race almost 



certainly masks the individual variation. It is, moreover, 

 difficult to tell from skins whether tne tails are really perfect, 

 while the fo it-measure, even for experienced European 

 collectors, is subject to a. personal or individual error, which is 

 relatively considerable. 



The measurements on the skulls have all been taken by 

 myself with fine pointed dividers on a metal scale, the tenths 

 of millimetres being estimated and in this connection it is well 

 to consider the errors inherent to the methods of measure- 

 ment, as ring on modern work 

 in mammalol gy, lo al rai ften founded on small 

 differences in measurements based on series which from a 

 biometrical point of view are frequently small. 



Masking errors may therefore be introduced from the 

 following cases : — 



(1) A skull which has been comparatively recently 

 cleaned, has been overboiled in the 



process of cleaning, will always give slightly 

 larger measun to opening of the 



sutures; in the case of badly overboiled skulls 

 this increase is permanent. 

 August, 1916. 2 



