506 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1920. 



thoracic pairs end with stiff, projecting cuffs which cover the claws 

 above and on the sides (B, C, D). Since the caterpillar uses only 

 the front feet for supporting herself against her work while spin- 

 ning, or perhaps for holding the fresh thread in place, one might 

 suppose that the cuffs on the other feet serve as guards to prevent 

 their claws from becoming entangled in the thread. But the cuffs 

 themselves are complicated, each being split into two lobes, and it 

 really looks as if they must offer only increased chances for possible 

 entanglement. Other caterpillars get along without claw guards, 

 and none are so clumsy, anyhow, as to get their feet snarled in their 

 knitting. 



When the Bucculatrix caterpillar has woven about 60 cross strands 

 in each direction, or 120 in all, including over 800 rib loops, the 

 cocoon has reached about two-thirds of its final length and has 

 crowded the worker mostly off the carpet. It is evident that if she 

 keeps on in this fashion she will eventually shut herself out of her 

 own house. But just at this point it appears that the same idea 

 occurs to the caterpillar, for she stops work and crawls into the 

 cocoon till her head touches the inner end. Then she turns " back to 

 back" and emerges in the opposite direction till she reaches the 

 other end of the carpet, which leaves about two-thirds of her body 

 under cover. In this reversed position she weaves the second end 

 of the cocoon in the same way that the first was made, except that 

 she now backs into the part already completed as her work advances 

 (pi. 3, E). Thus she is at least assured of being inside when the 

 thing is done. But now the question arises of how she will be able 

 to bring the opposing edges together when the space between them 

 becomes too narrow to emit her head and shoulders. When this 

 stage is reached, however, and just when the observer is most keen 

 to witness some clever trick, the caterpillar calmly withdraws her 

 head and deliberately bridges the space with a network of ordinary 

 figure-8 loops, continuing till the interval is closed by a plain sheet 

 of silk. A commonplace finish enough, and one that makes a very 

 weak ending to an otherwise highly entertaining show, yet the sim- 

 plest thing to be done in the face of an impossibility. The cater- 

 pillar observed for this description connected the two ends of her 

 cocoon very neatly and finished with the ends of the opposing ribs 

 in good alignment. An examination of other cocoons, however, 

 shows that the weaver is often somewhat bungling in this last act 

 of her work, for the lines of the ridges frequently do not match 

 at the bridge, and sometimes those of one end run into the grooves 

 of the other. Usually a slight sag is evident at this point. 



But the closing of the ribbed cocoon is not the end of the cater- 

 pillar's work. When the bridge is finished she keeps right on with 



