54 ARACHNIDA. 



by a widening (due to growth) of the median ventral groove. This 

 latter now becomes converted into a swelling, owing to the yolk 

 pushing out the thin layer of cells, which unites the germ-bands 

 ventrally, and forming a protrusion, the ventral yolk-sac (Barrois, 

 Balfour), which soon completely separates the two halves of the 

 germ-band from one another, causing the limb-rudiments, which 

 formerly overlapped ventrally, to become widely separated (Figs. 29 

 and 31 A) and to move apart. 



The middle portions of the two halves of the germ-band now 

 gradually move dorsally, a change brought about more by the 

 shifting of the yolk than by any growth of their own. And at 

 the same time the two extremities of this band are pushed still 

 further down until they reach the ventral surface. 



As a direct consequence of these changes, the previous dorsal flexure 

 of the germ-band gradually undergoes a complete reversal of its curva- 

 ture, and the germ-hand finally assumes a well-marked ventral flexure. 

 These changes constitute the "reversion" of the embryo. 



During these processes the growth of the dorsal surface proceeds 

 only in a line with the long axis of the body, while it becomes 

 narrower by the displacement of the two halves of the germ-band 

 (cf. Figs. 27 and 30). The opposite condition obtains for the ventral 

 surface, which we find shortening and widening, and finally converted 

 into the temporary yolk-sac, which completely separates the two 

 halves of the germ-band. 



The completion of the reversal of flexure is reached when the abdo- 

 men is bent forwards ventrally (Fig. 28 C), i.e., in a direction opposite 

 to that which it formerly assumed. In Agalena, to which this descrip- 

 tion specially applies, a great reduction of the posterior abdominal 

 segments is connected with this process (Fig. 31). These segments, 

 when the change begins, rise somewhat from the yolk, so that the 

 tail of the embryo resembles a loose flap lying on the yolk (Locy). 

 The flexure-reversal of the posterior region of the body may thus be 

 facilitated, and such a loosening helps us to understand the condition 

 of Epeira. In this form the post-abdomen only degenerates at a 

 later stage. Barrois describes a stage in the development of this 

 Araneid in which the ventrally-curved embryo lying on the yolk-sac 

 has, besides the four limb-bearing abdominal segments, at least six 

 (perhaps even eight) others. The early development of the dorsal 

 side of the abdomen, which has here taken place, is more easily 

 understood on account of its loose attachment to the yolk. The 

 change in the position of the yolk from the dorsal to the ventral side 



