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



On examining the measurements we see that the difference 

 between the largest and the smallest comberi is 7 mm., and that the 

 difference between the smallest listoni and the largest comheri is 

 also 7 mm. Why, then, is it necessary to seek further for 

 "intermediates"? In one case a difference of 7 mm. in length 

 between two individual rafs is not considered sufficient to separate 

 them as species , while in the case of another pair of rats a precisely 

 similar diff'erence is the sole reason given for separating them. 



Introductory statements must be made regarding the waj' in 

 which the subject has been dealt with in this paper. 



House Rats [Mus rattus). 



It will be found that the house rats from each place, if 

 received in large numbers, have been sorted into groups according 

 to their lengths, the numbers in each group being represented 

 graphicall}' by an upright line of proportionate length. This 

 method has been much used by biologists of recent years. It is 

 generally applied to large numbers of accurate measurements. 

 In the present instance it will be seen that the numbers that have 

 been so arranged are often not large, nor can the measurements 

 be very accurate. The measurements here recorded are of the 

 distance between the centre of the anus and the tip of the nose 

 in freshl}^ killed rats. If a careful observer measures and records 

 this length in ten freshly killed rats, and, without referring to the 

 first record, repeats the measurement two or more times, he will 

 find that his records do not agree, if the bodies of the rats are 

 disturbed in any way (as b}^ being lifted and replaced on the table) 

 between each set of measurements. The differences observed may 

 be as wide as five millimetres. The error is principally due 

 to the varying attitude of the head and neck. In taking the 

 measurement the head must be straightened, but not stretched 

 from the neck. In his efforts to perform this the recorder arrives 

 at a different result each time. Hence we are dealing with a dimen- 

 sion which cannot be accurately measured. This error contra- 

 venes in measuring the limp bodies of rats immediateh' after 

 their death. If a rat is measured two or three hours after death 

 when the neck is shortened and the spine curved by rigor mortis, it 

 may have lost as much as 10 % of its original length. In this case 

 the recorder, while measuring a large rat, is in doubt as to whether 

 10 mm. or 20 mm. should be added by stretching out the bod}'. 



However, as the measurements of diff'erent observers must be 

 marred by these accidents by an equal degree, it is clear that a 

 comparison of them must i3e of value. 



Although the sets of measurements were made b}* diff'erent 

 observers working quite independently, they show a striking re- 

 semblance to one another. They show that the adult house rats of 

 most districts vary in length from less than 145 mm. to more than 

 215 mm., but that those of 175 — 185 mm. are the most numerous. 

 As the same result is obtained from such widely separated districts 



