LEADING GROUPS OF EXISTING SYSTEMS 157 



phorae, and the Ctenoids among Acalephs; the Crinoids, Asterioids, 

 Echinoids, and Holothurke among Echinoderms; the Bryozoa, Bra- 

 chiopods, Tunicata, Lamellibranchiata among Acephala; the Bran- 

 chifera and Puhnonata among Gasteropods; the Ophidians, the Sauri- 

 ans, and the Chelonians among Reptiles; the Ichthyoids and the 

 Anoura among Amphibians, etc. 



Having shown in the preceding paragraph that classes rank next 

 to branches, it would be proper I should show here that orders are 

 natural groups which stand above families in their respective classes; 

 but for obvious reasons I have deferred this discussion to the follow- 

 ing paragraph, which relates to families, as it will be easier for me 

 to show what is the respective relation of these two kinds of groups 

 after their special character has been duly considered. 



From the preceding remarks respecting orders it might be inferred 

 that I deny all gradation among all other gioups, or that I assume 

 that orders constitute necessarily one simple series in each class. Far 

 from asserting any such thing, I hold on the contrary, that neither is 

 necessarily the case. But to explain fully my views upon this point 

 I must introduce here some other considerations. It will be obvious, 

 from what has already been said (and the further illustration of this 

 subject will only go to show to what extent this is true) that there 

 exists an unquestionable hierarchy between the different kinds of 

 groups admitted in our systems, based upon the different kinds of 

 relationship observed among animals; that branches are the most 

 comprehensive divisions, including each several classes; that orders 

 are subdivisions of the classes, families subdivisions of orders, genera 

 subdivisions of families, and species subdivisions of the genera; but 

 not in the sense that each type should necessarily include the same 

 number of classes, nor even necessarily several classes, as this must 

 depend upon the manner in which the type is carried out. A class, 

 again, might contain no orders,^^ if its representatives presented no 

 different degrees characterized by the greater or less complication of 

 their structure; or it may contain many or few, as these gradations are 

 more or less numerous and well marked; but as the representatives 

 of any and every class have of necessity a definite form, each class 

 must contain at least one family or many families, indeed, as many 

 as there are systems of forms under which its representatives may be 



1^ See Chap. I, Sect. i. 



