246 HYMENOPTEKA. 



of the genuine Orthoptera than of the otherwise nearer 

 alhed Lepidoptera and Diptera, 



Thus while the habits, general aspect, and compH- 

 cated nature of their external skeleton leads one to 

 consider the Hymenoptera as a highly specialized type, 

 it is evidently one that exemplifies specialization by 

 addition and not by reduction. So far as other struc- 

 tural and larval characteristics are concerned, they are 

 not therefore entitled, according to the standard here 

 adopted, to be considered the most highly specialized 

 of all insects. 



Although entomologists as a rule do not seem dis- 

 posed to consider that the three rings of the thorax 

 and the caterpillar-like larvae of the saw-flies indicate 

 affinities with the Lepidoptera, nevertheless this family 

 is evidently the most generalized of its order, and the 

 thorax, sessile abdomen, and mode of development 

 not only separate it widely from the great body of the 

 Hymenoptera, but remind one strongly of the Lepi- 

 doptera. Thus, although it would be an error, as 

 pointed out by Brauer and others, to consider the 

 saw-flies as transitions to the Lepidoptera, this family 

 is obviously more closely alhed to Lepidoptera than 

 any other of its order. One cannot avoid also giving 

 some weight ta the larval form, especially when it 

 occurs in association with such a generalized adult 

 form, and this leads one to suspect that the Hymen- 

 optera and Lepidoptera may have had a common an-' 

 cestor in spite of the anatomical differences which 

 now distinguish them. The absence of the caterpillar- 

 like stage of the Tenthredinidae may have been due 

 to its obliteration by the law of acceleration acting 



