48 



of difference. The inquirer is thus further relieved from the 

 detail of specific differences by the division of each genus into 

 sections. The still subordinate but constant points of difference 

 last named will be characteristic of species. 



I have already alluded to the important and valuable test of 

 the soundness of these principles of classification afforded, un- 

 expectedly and after the work was completed, by the stratigra- 

 phical harmony exhibited by the table of classification. It will 

 be sufficiently obvious that the ocean of different ages would 

 have such modifications as would not be adapted equally to all 

 varieties. We accordingly find among the Ventriculidse, as in 

 other divisions of palaeontology, a few species enduring through 

 many changes; others dying out; while with every fresh sera 

 fresh forms display themselves. 



It will be understood from this, that mere size does not enter 

 as an element into the determination of genus or species. Of 

 many species I have specimens from an inch to eight or nine 

 inches in diameter. It is not necessary to enter very fully, there- 

 fore, into the question of growth. That question, always a difficult 

 one in paleontology, is difficult even in recent forms of the families 

 allied to the Ventriculidse. It would be vain to hope to throw 

 much light upon it by fossil forms. Where constant differences 

 are found under all varieties of size, we are bound to consider 

 them as distinct species. I shall touch briefly on the question 

 of growth in introducing each separate genus. 



It will be also understood that the mere external (outward or 

 inward) general form of the fossil does not enter as an element 

 into the determination of genus or species. I have shown how 

 deceptive that criterion must ever be. In the present instance 

 the same general external form conceals essential differences in 

 the mode and degree of folding of the membrane. 



It will occur to the reader that to follow the fold of a mem- 

 brane, the trace of which is preserved only in a hard and solid 

 matrix, must be a work of great difficulty ; and especially when 

 that matrix is either so friable as the chalk, or so impracticable as 

 the flint. The actual amount of the difficulty * cannot however be 

 fully appreciated without actual experiment. The presence of that 

 very oxide of iron, without which the forms could not be, in general, 



* In order that the actual nature, importance, and results of the present 

 investigation should he properly understood, it is necessary to remind the 

 reader that from the time of Dr. Mantell's first work to his latest, and 

 either by him or the other latest writers (see Portlock's 'Report, &c.'p. 342), 

 it has never been suggested or suspected that any membrane whatever 

 existed in any of the Ventriculidse. They all describe them as composed 

 of anastomosing " cylindrical fibres," (see ante, p. 4,) between which, on 

 the inside, papillae or tubuli arise. I have demonstrated that the basis of 

 the Ventriculidse is a simple unperforated membrane ; that, therefore, the 



