562 [December 



seed" envelop is carefully opened, the included insect will he seen to 

 be still in the larva state. (Harris. l}>id ; Dipt. N. A. p. 185; Westw. 

 Intr. II. p. 529.) Now how is it possible for the "flax-seed" envelop to 

 be composed of the external integument, or '• skin," if you choose to call 

 it by that name, of the larva, when that larva exists in its normal con- 

 dition inside the ''flax-seed" envelop? To believe this, we must be- 

 lieve that the larva moults twice over to pass into the pupa, once to 

 form its pupal envelop, and once to pass into the pupa state, which is 

 contrary to all analogy. Of one thing I am, at all events, quite certain, 

 viz: that with the Grall-gnats of the Willow it is impossible that the 

 cocoon can be formed of the external integument of the larva; for, not 

 only is there an utter absence of the transverse sutures which we find 

 in all coarctate pupae, representing the sutures between the joints of 

 the larva, but in several species the cocoon is 2 — 4 times as long as the 

 body of the larva when that body is stretched out to its fullest extent. 

 Moreover in two specimens of the gall S. siliqua, (see below No. 8,) I 

 found two cocoons, one inside the other ; so that if the cocoon of this 

 species is always formed of the larval integument, the larva must, in 

 these two cases, have moulted twice over to form its two cocoons; which is 

 absurd. Osten Sacken observes that " the larva of C.pini inopis 0. S. fas- 

 tens itself to a pine leaf, and remains motionless until the resinous sub- 

 stance, which it exudes abundantly, begins to harden ; the larva then gra- 

 dually frees itself from the contact of the cocoon-like case thus formed ' 

 {Dipt. N. A. p. 185.) These observations are in complete harmony 

 with the theory of Winnertz, quoted above; but when Osten Sacken 

 adds that "it is very probable that this cocoon is nothing but the outer 

 larva-skin, saturated with resin," I think he has been inadvertently led 

 into error by the theories of Harris and Fitch. 



I am also very skeptical as to certain assertions of Harris and Fitch, 

 that the larva of Cecidomyia transforms gradually into the pupa state, 

 by a kind of budding process, without moulting the larval integument, 

 instead o? suddenly moulting into the pupa state, as in all other insects. 

 This theory seems to have been devised in order to harmonize with the 

 erroneous hypothesis already referred to, (viz : that the cocoon of the 

 Hessian fly is made out of the external integument of the larva,) and 

 so prevent the necessity of assuming that the larva moulted twice over 

 to pass into the pupa state. (See Harr. Lij. Ins. p. 577.) Thus, per- 



