416 THE ANATOMY OF IXERVTEBRATED ANIMALS. 



to the oral cavity, takes place. The greater part of the meso- 

 blastic cells become the adductor muscle, which is at first sin- 

 gle and answers to the posterior adductor of the adult. There 

 seems to be no shell gland. The shell appears at first as a 

 membranous cuticula, continuous from side to side, and there- 

 fore undivided into two valves. Subsequently it becomes 

 calcified and bivalve. The byssus gland is developed as an 

 involution of the octoderm at the posterior end of the body; 

 and the ventral hemisphere, or that opposite the shell, be- 

 comes divided by a deep median fold into the two lobes of the 

 mantle on which the characteristic pencil-like papilla? appear. 

 In front of the rudimentary mouth are two ciliated depres- 

 sions of the ectoderm, which are possibly the rudiments of the 

 nervous ganglia. 



In Unio and Anodonta the young are hatched in the outer 

 gill pouches of the parent, from which they are so dissimilar 

 that they were at one time considered to be parasites ( Glochi- 

 dium). The valves of the shell are triangular, and have in- 

 curved and serrated apices, by the help of which the larvae, 

 after they leave the parent, attach themselves to fishes and 

 other floating bodies. In this position they undergo a sort of 

 metamorphosis, and eventually fall off and sink to the bottom 

 as minute fresh -water Mussels. 



On comparing the Lamellihranchiata with the Brachio- 

 poda, it is obvious that the two have, in common with one 

 another and with the Annelida, the ciliated or veligerous 

 larval form. If the shell gland is, as Mr. Lankester suggests, 

 the homologue of the peduncular gland of Loxosoma and of 

 the Brachiopod larva?, it follows that the peduncle of the 

 Brachiopod corresponds with the centre of the pallial surface 

 of the Lamellibranch, and that the so-called dorsal and ven- 

 tral lobes of the mantle in the Brachiopod correspond with 

 the anterior and posterior halves of the mantle in the Lamel- 

 libranch. The Brachiopod hinge will therefore be transverse 

 to the axis of the body, while the Lamellibranch hinge is 

 parallel with it. If this comparison be just, however, the 

 three segments of the Brachiopod larva cannot answer to the 

 segments of an Annelid larva, but the two posterior seg- 

 ments of the Brachiopod larva must represent an outgrowth 

 of the haemal side of the body; and this would correspond 

 very well with the arrangement of the intestine in the artic- 

 ulated Brachiopcda. 



In the simplest forms of the Lamellihranchiata, as Tri- 

 gonia, Nucula, and Pecten, the mantle-lobes are almost, or 



