INSECTA. 245 



lescence longitudinally, and by approximating transversely, and ulti- 

 mately uniting at the middle line, first form an eight-chambered, and 

 afterwards a spherical, gland, (^fig^ 105. s.) The ovaria, retaining 

 their primitive separate state, increase in length, and assume the spiral 

 disposition in the pupa state. 



I have already* described the chief changes which the nervous 

 system undergoes during the transformation of the larva into the 

 imago, and need only now observe that the general principle of those 

 changes is like that which governs the modifications of the muscular 

 system, viz. a localisation of special masses at particular parts for 

 special purposes, the result of which is the dejDarture from a common 

 to a particular type of arrangement. 



One of the most obvious and remarkable phenomena in the larval 

 life of an insect is the successive sheddings of the skin. The number 

 and frequency of the ecdyses varies in different species, and relates to 

 two circumstances, viz. the rapidity of the growth of the body, and 

 the susceptibility or otherwise of the skin to be distended or to grow 

 with the increase of the body. 



The soft-skinned maggots of many flies, which acquire a vast 

 increase of size during their brief larval state, never moult until they 

 change into pupae, when the exuvium forms the pupa-case. In like 

 manner the soft-skinned apodal larvae of the Hymenoptera do not 

 moult until they have acquired their full size. The caterpillars of 

 the Lepidoptera moult at least three times, and some more frequently ; 

 the Bombyx villica, for example, from five to eight times, and the 

 tiger-moth (Arctia caja) ten times. 



With regard to the nature of the mutations and ecdyses which 

 culminate in the perfect insect, I should hardly have felt justified, 

 after what has been already detailed respecting the development of 

 the larva in the egg, in referring to the hypothesis of Swammerdam, 

 — that the imago was actually included in the larva, and that all new 

 skins pre-existed beneath the old one, — if such opinion had not been 

 adopted to explain the metamorphoses of insects in the admirable 

 work of our countrymen, Kirby and Spence.* The accurate observ- 

 ations of Herold on the changes and development of the organs 

 during the pupa state show these to be, like the original processes of 

 the development of the larva itself, the results of a transmutation, 

 increase, and coalescence of primitive elements of the different tissues. 



The few instances of the reproduction of mutilated parts in in- 

 sects have been observed to take place only at the period of the moult, 

 and are never manifested by the imago. A young Blatta^ in which 



* p. 209. t Introduction to Entomology, vol. ii,, p. 61. 



R 3 



