DECAPODA. 133 



the entoderm-cells do not here contain these masses of food-yolk, Liit occupy 

 the intervening spaces. The same is perhaps the case in Atycphyra (Ischikawa, 

 No. 51). In Crangon and Alpheus, on the contrar}', nothing in any May 

 lesembling the secondary cleavage of the yolk is apparently to be found 

 (KiNGSLEY, N"o. 53, and Hekkick, No. 49). 



In all these cases, the formation of the germ-layers begins with the develop- 

 ment of a gastrula. Some recent researches by Lebedinsky (No. 57) as to the 

 formation of the germ-layers of a Brachyuran, Eriphia spinifrons. seem to us to 

 require further explanation. Here also there is a gastrula-invagination, from 

 the inner portion of ■which the entoderm-cells that immigrate into the yolk 

 proceed, while, from the lateral walls of the vesicle, the mesoderm proliferates. 

 Nevertheless, even before the develojmient of this invagination, a separation 

 into three layers lying one above the other, and corresponding to the three 

 germ-layers, could already be made out in the multilaminar germ-disc. Further, 

 the ectoderm of the germ-band is said to give oft' mesodermal elements along its 

 ■whole inner surface by the division of its cells. The entoderm-cells scattered in 

 the food-yolk linally rise to its suiface and there become arranged to form the 

 wall of the mid-gut. At the same time, however, the food-yolk breaks up into 

 columnar masses corresponding to the separate cell-areas, so that here also, at a 

 late stage, secondary yolk-pyramids are developed. 



The mesoderm arises in Astacus (Bobretzky, Xo. 41, and 

 Kbichenbach, Nos. 64 and 65) and in Eupaguriis (Mayer, No. 59) 

 from a definite point at the anterior margin of the blastopore 

 (Figs. 60, BM, and 62, mes). Other authors have less definitely 

 assigned its origin to the neighbourhood of the blastopore (Kingslev, 

 No. 53, for Cranr/071, Haeckel, No. 47, for Penaeus), or to the 

 lateral walls of the archenteron (Bobretzky for Palaemon, No. 41, 

 and Lebedinsky for Eriphia, No. 57). The first rudiment of the 

 mesoderm is always multicellular. Its cells, which rapidly increase 

 by division, spread out apparently in an irregular manner between 

 the ectoderm of the germ-band and the food-yolk (Fig. 63, m). 

 Only in a few stages is the distribution of the mesodermal elements 

 into paired mesoderm-bands perceptible. The indications of its 

 division into segments is equally slight. In Astacus, a grouping 

 of the mesodermal elements into paired segnientally -arranged 

 coelomic vesicles was found by Reichenbach only in the abdomen 

 at quite a late stage, when the abdominal limbs had already begun to 

 form. In the anterior region of the body, coelomic vesicles evidently 

 do not develop, the body-cavity at this part having the character of 

 a pseudocoele. 



Besides the usual mesodermal cells, Reichenbach found, in the mesodermal 

 rudiment, smaller cells wliose protoplasm exhibited a peculiar foam-like structure 

 (Fig. 62, rii, and Fig. 82 A, sm, p. 161), and which contained several very small 

 deeply staining nuclei. These, which were defined by him as elements of the 

 secondary mesoderm, are said to originate by a process of endogenous cell- 



