580 Wm. E. K ell icon 



10 Weight of gastrocnemii. The ankle-joint was completely flexed and the tendo Achilles sectioned 

 in a radius of the joint. The attachment of the muscle to the femur was then cut along the bone and 

 the muscle removed. Both muscles were weighed on a balance sensitive to one milligram. They were 

 handled only with forceps, by the tendo Achilles and were not allowed to dry. 



11 Weight of liver. The abdominal cavity was then opened and the liver removed by cutting 

 through its attachments and blood-vessels closely along its surface. It was then gently rolled in a towel 

 until it ceased to stain and weighed as above. 



12 Weight of ovaries. These were removed in same manner as liver except that being practically 

 bloodless, they were not rolled. 



13 Length of alimentary canal. The trunk was then completely divided, the mesentery cut through 

 and the alimentary canal straightened out to its full extent by pulling gently. A condition is soon 

 reached where no farther stretching occurs and the length was measured when this point was reached. 

 This character may be considered as little subject to error in determination as any visceral character, 

 since a number of tests showed that this method of measurement gave a very reliable datum, much more 

 reliable than was expected. 



A resume of the data will be given first as briefly as possible, 

 discussion of their significance being deferred until they have been 

 completely presented. 



3 Numerical Ratio Betzveen Sexes 



Preliminary examination of the data shows at once that the 

 sexes must be treated separately. Of the total number measured 

 173 were males and 252 females but these numbers are not quite 

 indicative of the actual ratio, as sixteen additional individuals 

 were collected and used for purposes that might have affected the 

 values of some of the internal characters and w^hich therefore were 

 not measured. Of the total 441 collected 175 were males and 266 

 females, giving a ratio of 658 males to 1000 females in the entire 

 colony. 



^ Variability 



The means and standard deviations of all the characters in the 

 male and female series, calculated according to the usual methods^ 

 are given in Table I and the coefficients of variability in Table II. 



^Formulae and methods from Davenport ('04). 



I (F. /) 

 (V/ = 



[■•(yr 



