O. A. Merritt Hawkes 149 



valve in the erection of the tubercles. The tubercles are well supplied 

 by tracheae and nerves. The fat-body and muscles extend into the 

 basal, but not into the distal portion of the tubercles. The tubercles 

 are similar to those of the Tortoise-shell butterfly {Vanessa urticae), 

 described by Berlese(l). This being their anatomy the tubercles are 

 not arsimple character like colour, etc., but they have a certain anatomical 

 importance, and must be represented in the chromosome by a large 

 number of correlated factors. 



An attempt was made to mate these abnormal forms, but two were 

 never fit for mating at the same time and it was not until the following 

 generation {F-^ that resort was finally made to a mating between a 

 normal and an abnormal larva. The abnormal form used in the suc- 

 cessful mating (m. 79) was not one with perfectly reduced tubercles, but 

 had them represented only by stumps (papillae) except in the infra- 

 spiracular row, where the position of each tubercle was indicated by a 

 chitinous ring. In this family 86 adult larvae were reared (see 

 Table III). These larvae all had normal tubercles, the abnormal form 

 had disappeared. These were then inbred, ten families being reared, 

 and amongst them, larvae occurred with long, with short, and with no 

 tubercles — an obvious Mendelian segregation. The individuals with 

 some or all tubercles absent were separated from those that had tubercles, 

 but the long and short tubercled forms were not separated from one 

 another — partly because of lack of space and partly because I had no 

 satisfactory standard of size by which to effect a satisfactory separation. 

 The tubercles vary in distention and also in individuals. Unfortunately 

 the completely non-tubercled forms failed to mate with one another. 

 A large number of the long and short tubercled forms were eaten by 

 mice, and those that appeared as imagines did not mate well and laid 

 very few eggs. These results, as stated in the earlier section, may have 

 been due to inbreeding for five generations or simply to the unsatis- 

 factory heat of the breeding room. As a result of breeding in F<i four 

 small families were reared (98, 96, 100, 103). In two of the families, 100 

 and 103, the larvae all had short tubercles, but no tubercles were absent. 

 In another family (96) most of the larvae were normal, but a few had short 

 tubercles. In mating 98 one, of the parents (83*7) was fi-om the cage 

 which contained a mixture of larvae with both long and short tubercles, 

 and the other was from a larva (81*21) with reduced tubercles; this 

 larva was as follows : the two rows of dorsal tubercles were very short, 

 those on the seventh and eleventh segments being represented only by 

 minute protrusions, the other rows of tubercles were not quite as short 



