TUE CEYLON PEARL-OYSTER. 309 



The outermost layer of the periostracum (which in Margariti- 

 J'era is a negUgible quantity) and the hyposti-acum probably 

 arise by direct transformation of the outermost portions of 

 specialised epidermal cells, and on this account it may prove 

 necessary to draw a sharper morphological distinction between 

 them and the rest of the shell than has hitherto been done *. 



The prismatic layer and the nacre, together with the inner 

 layers of the periostracum, more probably arise as a seci'etion 

 Avliich first hardens into a membnine ht situ, and then forms 

 the delicate skin which Huxley observed between the mantle 

 and the shell in the freshwater mussels. 



It would appear that the lime-salts and albumirious fluid 

 which hardens to form the conchyolin are independeut of each 

 other, and may be secreted in varying proportions. Where 

 these two constituents are seci-eted under circumstances wdiich 

 inhibit the control of the shell -secreting epidermis, or where 

 the secretion takes place so copiously and rapidly that the 

 epidermis is unable to regulate the deposition (as in the patho- 

 logical cases described below), lime-salts are precipitated in a 

 columnar form, much as in Hnrting's bodies, and, concurrently 

 with this, the albuminous fluid is transformed into an insolul)le 

 substance resembling conchyolin. The process of shell- secretion 

 at the rapidly growing edge of the shell resulting in the 

 formation of the prismatic layer — which in Margaritifera 

 v^dgaris measures as much as 1 mm. or more in thickness — is 

 probably in some degree analogous to the pi'ocess of secretion of 

 repair-substance, the epithelium exercising comparatively little 

 control over the arrangement of the elements. 



But in the case of the nacre it is different. Here the epithe- 

 lium seems to exert a definite and very strict selective influence 

 resulting in the finely stratified and chambered structure which 

 can, I think, best be interpreted as arising fx^om rhythmically 

 intermittent secretory action on the part of the controlling 

 epidermis. Any disturbance of the normal rhythm of this 

 secretion, e. g. the stimulation of an intrusive particle between 

 shell and epidermis, i-esults in the formation of the iriegular 

 substances described below, such as granular repair-nacre, the 

 several varieties of columnar repair-substance, or the amorphous 

 non-calcified substance. 



It would thus seem as though the structiu-e of the shell- 

 substance, and its variations, normal and pathological, could 

 be expressed in terms of the proportions of lime-salts and organic 



* The difference between the outermost layer of tlie periostracum and the hj'po- 

 stracum on the one hand, and tlic remainder of the shell on the other, the former parts 

 arising by dii-ect cell-transformation or cuticnlarisation of cell-protoplasm, the latter 

 as a secretion poured out by the cells, svigojests a line of inquiry that mif^ht yield 

 interesting results. Can these two constituents of the shell be separated moi-pho- 

 logically and phylogenetically, and, if so, can the former be regarded as in any sense 

 homologous with the cuticular exoskeleton of an ancestor common to MoUu^ca and 

 Arthropoda, the latter being a subsequent addition peculiar to the Mollusca. 

 associated with their more sedentary modes of life, which has )iow, for all practical 

 purposes, replaced the more strictly cuticular element as an exoskeleton ? 



