14 CRETACEOUS GASTROPODA 
respects preferable to that of PROBOSCIDIFERA, ToXIFERA, RosTRIFERA, &c., because, if 
we depend on this single character of dentition, groups otherwise very closely allied are 
totally separated from each other.* To the palzeontologist also, who has principally to 
deal with the shells only, and but very rarely with the animals, the first mentioned 
division is—no doubt only for the present—of far greater use, while he could scarcely 
take the desirable advantage of the other system in classifying his fossil remains. 
Kefersteint distinguishes in the two divisions of the SipHonosromaTa and 
Ho osromMatTa a number of smaller groups, depending on the variations in the 
dentition of the radula, as the indefatigable labours of Drs. Gray, Troschel, Lowen, 
Morch and others have made them known. It is certain, that these differences in 
the plaits or teeth of the radula, and their connection with the food and general 
living of the animal form a very important character in the classification: it would 
be quite an exception in this, as compared with the other classes, if they did not. 
Still so much remains to be done in this direction, as we have already noticed, that 
Keferstein’s sub-divisions into TanrociossA, Racutetossa, &e., offer disadvantages 
similar to those presented by Adams’ higher divisions. They widely separate 
families, which are closely allied in every respect save the dentition : so widely 
indeed, that even a partially natural arrangement from lower to more highly 
organized forms is unobtainable. We shall not enter further upon the discussion 
of these divisions depending on the dentition, but shall simply quote the families 
in what we believe to be their respective relations to each other, for which we shall 
endeavour to give more detailed proofs. We cannot pretend to call this arrange- 
ment more natural than many others. Our only reason for adopting it is, that m 
our present case of describing a local cretaceous fauna it appears to be more suitable 
for our purposes than that adopted by Keferstein (loc. cit.). As regards families and 
sub-families, which will be referred to, we must state in each individual case, in 
what sense these divisions are to be understood. 
Tribe, Siphonostomata. 
The CrenoBRANCHTA included in this tribe have usually a small head with either 
a proboscis or a long rostrum, and with a respiratory sipho of different lengths, 
sometimes represented by a siphonal fold only. The shells are convolute, involute, 
or coiled up into a conical spire. Depending on the development of the respira- 
tory sipho, the aperture is either provided on the anterior termination with a canal, 
or is simply a little produced and notched. The operculum is sometimes wanting ; 
but usually it is present, of horny or often of lamellose structure. Nearly all the 
species inhabit the sea and they are mostly carnivorous, feeding upon other mollusca 
or similar animals. 
Geologically speaking, the SrpHonosromATA are younger than the HoLostomata. 
Of the former, several families scarcely appear in deposits of date earlier than ter- 
* This point has been most recently discussed again by Mr. Crosse (Journal de Conchiliologie, 3me. Ser. 
Vol. VL, p. 216, 1866), contesting the value of the dentition as the only or even principal base of classification 
in Mollusea. 
+ Bronn, Klassen und Ordnungen des Thierreiches, Vol. III, p. 1030, &e. 
