THE GILLS. 71 



In cases in which, as in Cyclas, the rudiment of the gill is leaf-like 

 and only breaks up later into consecutive lobes through the slits 

 which arise in it, we may assume that these lobes unite later, like 

 the papillae, to form the branchial leaf. 



If we compare the origin of the gills in Teredo and Cyclas on the one hand 

 and Mytilus, etc., on the other, we might at first feel inclined to regard the 

 method seen in the foi-mer as the more primitive, since the formation of the 

 leaf precedes that of the papillae. The gill originates as a leaf, and is only 

 later broken up by incisions into separate lobes which are arranged in the 

 same way as the papillae in other cases. This view, which is founded on the 

 ontogeny of a few forms such as Teredo and Cyclas, which in other respects 

 are undoubtedly specialised, cannot, however, in any way be reconciled with 

 the morphological conditions of the definitive gill in the different Lamelli- 

 branchs. A comparative study of these latter suggests rather that the origin 

 of the gills in the form of papillae; as in Mytilus, was the primitive condition. 



Unfortunately very little is as yet known as to the mode of formation of the 

 gills, but if we examine the apparently carefully investigated development of 

 these organs in Mytilus and Ostrea, we find that certain ontogenetic stages can 

 be most exactly matched in the shape of the gills of certain adult Lamelli- 

 branchs. Thus, in Dimya, according to Dall, the gill on each side consistsof one 

 row of branchial filaments (Fig. 30 B) and in Amusium Dalli (and as it appears 

 also in Area ectocomata) there are two such rows on each side (Fig. 30 C).* 

 The brauchial filaments are not connected, and thus represent the ontogenetic 

 stage at which there are one or two rows of papillae. The further develop- 

 ment of the gills may be imagined to have taken place by the free ends of the 

 branchial filaments becoming connected, in the manner illustrated in the 

 ontogeny of Mytilus (p. 68). In this way the row of branchial filaments gave 

 rise to the branchial leaf. This leaf doubled back on itself, when an increase 

 of surface was needed and growth in a straight direction was not possible on 

 account of the want of room in the shell (Fig. 29 B-E). The ascending (re- 

 flected) lamella of the branchial leaf thus arose ; the free edge of which may 

 finally fuse with the mantle, as is the case, for example, with the ascending 

 lamella of the outer branchial leaf in the I nionidae (Fig. 30 E). 



That form of gill which consists of single filaments, bent back upon them- 

 selves, thus indicating the two lamellae of the later branchial leaf (Fig. 30 />) 

 has repeatedly been held to be very primitive and has been thought to repre- 

 sent the stage succeeding that in which the gills consisted of two straight rows 

 of filaments (Fig. 30 C). Such gills are found in Trigonia (Pelseneer) and 

 Area noae which may be considered as very old forms. The gill-leaf consisting 

 of two lamellae was thought to have arisen from the union of these reflected 

 filaments. To us, the reflection of the single filaments and their regular. 

 almosl leaf-like arrangement, such as is seen in the gills of Pecten and Mytilus 



' We follow here the accounts given by Pelseneer, Dall and Mitsuklhi 

 of the morphological conditions of the Lamellibranch gills. It is impos- 

 sible to decide how far these may represent primitive conditions or may to 

 some degree be degeneration-phenomena, for it is evident that these latter d<> 

 occur and cause a reduction of the gill-leaves. 



