4 THE MEASUREMENT OF VARIATION. 



supposing it be said that out of a large number of indi- 

 viduals half varied in size by 5 per cent, from the 

 average of the whole, then there is afforded a numerical 

 expression of the degree of variation, which can readily 

 be compared with similar expressions concerning the 

 variability of other parts of the same organism, and 

 with those of quite distinct organisms. 



Let us first of all, therefore, examine one or two 

 simple series of measurements made on a group of indi- 

 viduals of a species, so as to get some idea of the actual 

 differences exhibited by the varying characters, or, as 

 they have been termed, the variants. Some of the most 

 striking are those obtained by J. A. Allen,* concerning 

 the variation in certain mammals and winter birds of 

 East Florida. Of a species of squirrel (Sciurus caroli- 

 nensis), for instance, 28 individuals were measured, 

 and these . neasurements are reproduced to scale in the 

 accompanying diagram. Here the animals are ar- 

 ranged in order according to the length of their body 

 in inches, and the corresponding values for the head, 

 tail, and forefoot are given on the same ordinates. By 

 means of this diagram, the magnitude of each and all 

 of the measurements made can be read off at a glance. 

 The body was on an average 9.15 inches long, but the 

 extreme values were 8.25 and 10.20 inches, or respect- 

 ively 9.8 per cent, and 11.5 per cent, less and greater 

 than the mean. The tail measurements were even 

 more variable than this, the extremes varying from 6.75 

 to 8.75 inches, or by respectively 14.3 per cent, and 11 

 per cent, from the mean. In the forefoot the range of 

 variation was less, and in the head smaller still; but 

 * Bulletin Museum Comp. Zo51,, Harvard, 1871. 



