278 



PROCEEDINGS OF THE ACADEMY OF 



[June, 



The Method of Measurement. 



The method adopted in the measurement is as follows: the right 

 fore wing of the butterfly, carefully removed from the body, is put 

 upon the glass stage of a dissecting microscope and covered with 

 a thin plate of mica; the mirror of the microscope is arranged to 

 so reflect light from its surface as to make the outline of the wing 

 clear; then the length from the base of the wing to the farthest 

 point on its outer margin, without taking the length of the marginal 

 hairs into account (this point is always a little behind the anterior 

 angle), is measured with compasses from over the mica to the nearest 

 half -millimeter. 



The Differences Among the Means. 



Table II shows how the variants of each lot are distributed through 

 the classes, while in Table III are exhibited the means of measure- 

 ments, the standard deviations, and the coefficients of variation 

 with their probable errors. One may notice in these tables, but 

 more clearly in Table IV, how some fairly considerable differences 

 occur among the mean values of several measurements. These 

 differences are always great enough to be considered significant, as 

 they exceed three times the corresponding probable errors except 

 in one case only (a) where two groups of males from the same locality 

 and captured in the same season two successive years are to be 

 compared with each other. It might be supposed that some 

 intrinsic causes lying in the bodies of the animals must have had 

 influence in determining the wing-length of the butterflies; but 

 several external factors, too, e.g., the temperature, the quantity of 

 food, etc., working on the developing organisms surely have had 

 some important share in it. 



