GENERAL CONSIDERATIONS — FORMATION OF THE GERM-LAYERS. 309 



other, they envelop the embryo, as has already been mentioned in 

 the case of Hylotoma, lip to the time of hatching. This type occurs 

 in the Hymenoptera and the Lepidoptera. In the latter (Fig. 142 C, 

 p. 285), in which the germ-band is immersed, remains of food-yolk 

 are retained between the amnion and serosa, which, together with 

 the large cells of the serosa, serve as the first nourishment for the 

 young caterpillar (Ganin, No. 23). 



E. General Considerations. 



We must regard as the most primitive the first of these types 

 of development, in which, after rotation of the germ-band and the 

 development of a continuous amnion-serosa-sac, conditions are brought 

 about resembling those usually found in the other Arthropoda, the 

 serosa being invaginated and gradually degenerating. This is con- 

 firmed by the fact that this type is frequent in those orders of 

 Insects which are generally claimed as the more primitive. The 

 fourth type of development, on the other hand, in which the 

 embryonic envelopes lose their continuity with the embryo and 

 form two free membranes enclosing the latter, must be regarded as 

 the most specialised. The third type is intermediate between these 

 two. With regard to the way in which the serosa degenerates, it 

 approaches the first type, but in the separation of the amnion 

 resembles the fourth. The second seems to represent a type of 

 formation of the dorsal integument independently acquired among 

 the Coleoptera. 



In the first type, the development of the amnion-serosa-sac is 

 introduced by the rupture of the two fused embryonic envelopes. 

 This rupture in the ventral median line in the Libettulidae takes 

 place only in the cephalic region. In the second type only the 

 serosa is affected by the rent, while, in the fourth type, both 

 embryonic envelopes remain intact up to the time of hatching. 



6. The formation of the Germ-layers. 



The older accounts of the formation of the layers in the germ- 

 band of the Insecta were very incomplete. Butschli (No. 11) first 

 found that, in Apis, a loAver or inner layer is produced in the 

 germ-band by an infolding. Soon after, Kowalevsky's researches 

 (No. 48), carried out by means of sections, laid the foundation for 

 more accurate knowledge. Kowalevsky found that, in Hydropliilus, 

 a furrow appears running along the whole length of the rudiment 

 of the germ-band (Fig. 134 A, B, r, p. 270), which, sinking in, 

 yields the loicer or inner layer of the germ-band, i.e. t the common 



