18 ILLINOIS BIOLOGICAL MONOGRAPHS [18 



workers. Many problems still remain unsolved, — for the recapitulation 

 is neither direct nor easy to interpret and is often covered by such a 

 multitude of complications, reversals, and omissions that we are baffled 

 in an attempt at their solution. For this reason, before an argument 

 can be based on ontogeny, it must be shown that the condition discussed 

 is necessarily a recapitulation. 



When the larva of a moth hatches from the egg it is somewhat 

 different, in most cases, from the mature caterpillar. The ' ' woolly bear ' ' 

 is as naked as a cutworm and the butterfly larva could be mistaken for 

 a tortricid. Sometimes indications of this peculiar condition remain 

 after the first molt, but usually this glimpse of the past is as evanescent 

 as it is surprising. Let this minute "worm" reach the second instar 

 and nothing remains to show that the arctian was not always "hairy", 

 or that the ancestors of the saturnian did not possess prominent scoli. 



The natural supposition that this first stage is a recapitulation of 

 the past has seldom been doubted. Several arguments, or, rather, sug- 

 gestions, may however be advanced against it, and these must be dis- 

 proven before this stage is admitted as evidence. 



In the first place it may be urged that this stage represents an 

 adaptive condition. While it is true that the thick setae of an arctian or 

 the spiny processes of a nymphalid might prove an obstacle to hatching, 

 this would merely show that the former condition had been retained 

 in the first instar after the appearance of the new armature in later 

 development, rather than that a new elementary stage had been acquired. 

 At the same time such an interpretation can not give us a clue to the 

 reason for the lack of one of the setae (mu) (cf. Figs. 29 and 31) on the 

 abdomen of all the Frenatae in the first stage, and its presence after the 

 first molt. Nor is it clear how the absence of theta from slightly 

 different positions on all the segments of Hepialus (cf. Figs. 2, 3, 4 

 with Fig. 6) could assist in emerging from the egg. Secondary adap- 

 tation might explain a transformation in the entire style of armatm-e 

 but only recapitulation can suggest a reason for changes in the presence 

 and position of a single seta. 



Again, the differences between the elementary stages of different 

 forms may be pointed out and the conclusion reached that they can 

 not therefore represent an ancestral condition. While the first stages 

 are not, it is true, identical throughout the order, they do not vary 

 one-tenth as much as the mature larvae. These first instars diverge 

 slightly in several directions from the ancestral type, while the mature 

 larvae diverge rapidly and extensively from the type. The former are 

 so nearly uniform that their evidence in regard to the past is invaluable. 



