396 DR. carpenter's researches : 



That tliey are not mere channels or excavations in the shell- 

 substance is proved by the fact that they may be seen in the 

 decalcified membrane. I have frequently seen in them indi- 

 cations of a cellular origin, as if they had been formed by 

 the coalescence of a number of cells arranged in a linear 

 direction ; and I find that Mr. Bowerbank has come to the 

 same conclusion. 



" The tubular structure is usually found only in the 

 ordinary membranous shell-substance ; in fact, I have seldom 

 observed it in the nacre, except where the tubes penetrate 

 this, to be distributed in a layer external to it, as is the 

 case, for example, in Anomia and Trigonia. I have nowhere 

 found it coexisting in the same shell with any great amount 

 of prismatic cellular substance ; consequently it is for the 

 most j)art absent in the Margaritaceae and Nayadeae, and 

 but very slightly manifested in the true Ostraceae. In most 

 of the families of bivalves, however, in which the lobes of 

 the mantle are united, some traces of it may be detected ; 

 though these are often very scanty. There is less regularity 

 in regard to this character, than in respect to most others 

 furnished by the microscopic examination of the shell. Thus 

 I have found a little collection of tubes in one spot of the 

 nacre of an Avicula, in no other part of which did I meet 

 with any ; and I have frequently found one species of a 

 genus extremely tubular, whilst another, closely allied to it, 

 was almost or entirely destitute of tubes." 



The third kind, or cancellated structure, resembles the 

 cancellated texture of bone, and is characteristic of a very 

 peculiar group of shells, named the Rudistes.* Dr. Car- 

 penter compares it to the prismatic cellular structure on a 

 large scale, with this important difference, however, that the 

 prismatic cells are not solid but hollow. " In what manner 

 these minute chambers were occupied during the life of the 

 animal, it is impossible now to say ; as there is no existing 

 group, to which the Rudistes seem to bear any close resem- 

 blance. The shape of each is usually that of a very short 

 hexagonal prism, terminated at each end by a flat partition : 

 consequently a section in one direction will exhibit the 

 walls of the chambers disposed in a hexagonal network ; 

 whilst, when the section passes in the opposite direction, 

 the transverse partitions come into view. The cancellated 

 structure is externally and internally covered with a shelly 

 plate, in which no perforations whatever can be seen. It 



* This structure was first described by Mr. J. E. Gray in tlie Magazine 

 of Zoology and Botany, ii. 228—32. 1838. 



