20 CEYLON PEARL OYSTEE REPORT. 



the attachments of the ligament foreshadows the formation of the posterior cardinal 

 tooth and the ligamentar fossa. In later stages, with seven or eight growth laminae 

 outside the prodissoconch, the adult characters of the hinge are fully established. 

 It would appear, then, that, as compared with the size of the shell, the visceral 

 organs and the permanent hinge teeth are precociously developed in Jousseaumia, 

 and the suppression of the provinculum and consequent abbreviation of the several 

 stages in the evolution of the heterodont hinge may account for the ligament 

 remaining internal and therefore in its larval condition instead of being shifted to 

 an external position. 



CONCLUSION. 



I have given a full description of this interesting little Lamellibranch, because, as it 

 seems to me, it is incumbent on students of this class to give a full anatomical 

 account of the various forms that come under their notice. A detailed account of the 

 anatomy of various Lamellibranchs is needed before many questions of classification 

 can be finally settled. The researches of Pelseneer (10) have broken ground in this 

 direction, but subsequent authors have not followed his example by dealing with the 

 whole anatomy of the species they have investigated. The work of Ridewood (15), 

 dealing with a large number of species of all orders of the Lamellibranchia, is con- 

 fined to a detailed exposition of the gill-structure, and though it forms a valuable 

 contribution to our knowledge of this single feature of Lamellibranch anatomy, its 

 main result has been to show how little the characters of a single organ are to be 

 relied upon in framing the smaller subdivisions of a system of classification. There 

 are at the present time few malacologists who will question the importance of gill- 

 structure as a basis of the general classification of the Lamellibranchia and their 

 division into the orders Protobranchia, Filibranchia, Eulamellibranchia and Septi- 

 branchia meets with general acceptance, the more so because these orders correspond 

 very closely with those based upon a stiidy of the hinge characters. But when we 

 come to subdivide the orders into sub-orders and to arrange the latter in families, 

 and especially when we attempt to estimate the relationship and probable lines of 

 descent of the various families grouped together in the sub-orders, the structure of 

 the gills becomes of less value to us. Thus, to take a single instance, in the family 

 Donacidse we find plicate and non-plicate, homorhabdic and heterorhabdic gills with 

 almost every variety of interfilamentar and interlamellar connection. In the large 

 sub-order Submytilacea, we find the simple Astartiform gill at one end of the order 

 and the extremely specialised complex gills of the Unionidai at the other, and no 

 very definite series connecting the two. Pelseneer (10 and 12) characterises the 

 Submytilacea as Eulamellibranchs with smooth, i.e., non-plicate gills, but Ridewood 

 (15) has shown that the gills of Diplodonta oblonga and Monocondylcea are slightly 

 and those of Corbicula lydigina markedly plicate. On the other hand, smooth or 



