MOKPHOLUGV OF LAMELLlUKA.NClllATE MOLLU«KS. 427 



Jt is pos.siblt' that tlic oyster came, iu its degeueiation, tbiougb a form in many 

 respects similar to Mytiliis, from some active ancestor with a Jocoinotor foot. Judg- 

 ing from the most primitive of existing lamellibranchs — Nucula, )SoleHO)nya,etii. — the 

 immediate ancestors of the gn^u}) probably possessed a greatly developed foot. Osfrea, 

 therefore, gives e\idence of having departed much further from this ancestor in its 

 degeneration than has Mi/tih(s. This latter form, in turn, is less changed by its mode 

 of life, and, noting here and there a form on the way to the most archai*-, Venus, Area, 

 Xin-iila, still less and less so. 



And yet the foot is too variable an organ to be made an exclusive means of dassi- 

 lication or even the chief one. Anomia has lost the foot on account of its fixed habit, 

 and yet it possesses some structures which indicate a more i)rimitive position for the 

 tbrm than the one commonly supposed for it. Pelseneer goes so far as to place it 

 inunediately next the ijrimitive forms with ])late gills. 



But if there is one organ of lamellibranchs which is most subject to variation 

 by secondary modification, it is the gill. While it is an im])ortant organ, it seems as 

 if it were hardly possible to use it as ;». basis of classifi{;atioii for the whole grou]), as 

 Pelseneer has done. In a group where any and all organs are so subject to secondary 

 modification, there must be a careful comparison of many of them, instead of one oi' 

 two, perhaps; and even a complete knowledge of the comparative anatomy, which we 

 by no means possess, can not be safely used as a basis for classificatiim without the 

 aid of comparative embryology, which is still less known. 



The i)hylo(jeny of the gilh, — It is generally considered that the anatomy of tliose 

 lamellibranchs which possess jjlate gills (X;/e;f/«, ^Solenonif/a, vtr.) shows conclusively 

 that the group which they fornj nuist be the most i)rimitiv«^ one of all living lauielli 

 branchs. More especially since the appearance of .Mitsukuri's paper i»n these plate 

 gills (No. 13) has this been the general opinion. Tlieir anahnnical and histidogical 

 similarity to certain gasteropod gills is one of the strong i>oints for such a belief 



But there is a very great gai» between these plate gills and the strictly filamen- 

 tous type, which, it seems to me, can not be explained by any facts which we now 

 possess, either anatomical or embryological. 



In his GhaUenger re]>ort on the Mollusca, Pclseneei- attempts to show by a series 

 of diagrams the phylogenetic <levelo]»ment of the gill. Beginning with 2IaUetia, with 

 plates extending laterally, he derives Niicula from it, in which the outer end of the 

 plate is turned slightly downward. He now has to interpolate an hypothetical gill in 

 which the i)lates have developinl \'entrally for some distance, but which shows no sign 

 of an ascending i^ortion. The luwt stage which he takes in his lihylogenetic develop- 

 ment is represented by the gill of Area, in which there is a fully formed ascending 

 liud) of the filament and neither limb shows anything plate-like, but both are cylin- 

 drical. But right here is the gap referred to, and it would still remain very great 

 even if his hypothetical gill just preceding it were really known to exist. 



This scheme of historical development was molded ui)on what little knowledge 

 we have of th«^ ontogeny of the gill. This knowledge, except for a few fragmentary 

 observations, we owe to Lacaze-Duthiers (No. 7). In his study of the development of 

 the gill of Mytilus, he shows that the gills appear from a ridge running horizontally 

 along the side of the body, from which rods grow out and descend ventrallj-. These 

 are separate from one another and become the filaments. Afrer attaining to a certain 

 length their enlarged ends fuse together and form a solid membrane. The lower edge 



