THE AMERICAN NATURALIST. [Vol. XXXVII. 



33-43 joints in the antennae. In the table of data (Table 9) 

 it is shown by underlining that two of the right antennae 

 counted ended abruptly (Numbers 8 and 50). On the left 33, 

 50, 57 and 86 had broken ends. The presence of a rounded tip 

 does not necessarily indicate perfectness but possibly merely 

 that regeneration took place at the last moult. Since the 

 so liable to injury not much stress can be laid on 



derived from them. The mode lies 

 an average of 41-23. There are equal 

 numbers in classes 39 and 

 JJ I I 1 1 4 1 on tne leit side and the 

 fill Ii: average is 40.77 . 



The number of joints in 



«r. 40. «. 44. antennae make a good 

 "quick" distinction be- 

 tween L.forficatus and L. multidentatus . The latter has fewer 

 joints (about as 21 to 43) but the individual joints are longer. 

 This can be seen on the left side of Figures 6 and 7. 



the polygons (Fig. 

 at 41 in the right \ 



Polygons o 



In the very young individual the hinder pairs of legs are not 

 vet budded out. The 13th, 14th, and 15th pairs of legs grow 

 in rapid succession, they may possibly all be indicated at the 

 same moult. This increase in the number of legs takes place 

 when the animal is less than 10 mm. long. 1 The 12th pair of 

 legs is the first pair to bear the pits on the coxae so that these 

 are the oldest pits of the series ontogenetically. The youngest 

 specimens that showed coxal pores at all had two pores, one on 

 each 1 2th coxa. These were round. They must of course 

 increase in number at the times of moulting until the adult con- 

 dition is reached. For the condition of the coxal pores in the 

 three hinder pairs of legs in both species under consideration, 

 see Figs. 9 and 10.- 



