566 [December 



gate, slender tip of the cell, where, the air having free access to it. 

 it dries rapidly, so as to form a subterminal diaphragm across the 

 mouth of the cell, as in C. s. nliqua n. sp. and C. s. cornu n. sp. 

 When its expansive force is lost before the walls of the cell are reached, 

 or when the walls of the cell are not adapted to adhere to a glutinous 

 substance, or when, from the free admission of air, the glutinous matter 

 dries too rapidly to have time to adhere, then the cocoon i-emains sepa- 

 rate and distinct from the walls of the cell, as in C. s. triticoides n. sp,, 

 or adheres to it only here and there, as in C. s. strohiloidea, etc. The 

 fact just now referred to of there being a double diaphragm formed by the 

 thin pellicle of the cocoon at both ends of the cell in two specimens of 

 the gall ^S*. siliqua found on S. cordata, seemed at first sight opposed to 

 the above hypothesis ; but we may get over the difficulty by supposing 

 some abnormal aftection of the larva, so that its gas began to be dis- 

 charged before it had done secreting its glutinous matter, and that it thus 

 formed two cocoons one after the other, and one inside the other. In 

 any case, uo matter how the cocoon was formed, there must have been 

 here two separate cocoons formed one after the other, and one within 

 the other ; and the fact of the exterior one of the two not having ex- 

 tended to the base of the cell, as it invariably did in scores of other 

 specimens examined by me, proves that when it was formed there must 

 have been a scant supply of material. On the whole, it is impossible 

 to look at the thin, filmy cocoons of 0. .s. strobiloides and its allies, 

 which are not thicker here and thinner there, but of one uniform, ho- 

 mogeneous thinness, without being impressed by the idea that they are 

 mere bubbles, blown by some wonderful and hitherto undreamt of pro- 

 cess within the lanceolate cell in which the animal resides. A larva 

 might spin such a homogeneous cocoon with its mouth, as many Hy- 

 menopterous cocoons of nearly as great tenuity and equally homogeneous 

 are spun, e. g. that o^ Pelo^^seus luiiafus Fabr. 5 but it is, I think, proved 

 that the cocoon of the Gall-gnats is exuded and not spun. It must, there- 

 fore, be either blown like a bubble or be daubed on the walls of the cell 

 by the body of the insect. But no mere smearing and daubing process 

 could spread that mortar in such a regular manner, as to be precisely of 

 the same tenuity, where it forms a diaphragm across the upper end of 

 the lanceolate cell, as in 0. s. strobduides, &c., that it maintains every- 

 where else. Consequently it must be blown like a bubble. 



