18S CTENOPHOR.E. Part II. 



not .sufficient to deserilje the mere outlines or the figure of its representatives, but 

 it is indispensable to point out the structural elements Avhich determine the par- 

 ticular form, characteristic of the family under consideration. I may add, from 

 the attemi)ts 1 have made to characterize the natural families of different classes 

 of animals, that this is one of the most difficidt tasks a zoologist can vmdertake ; 

 but I am at the same time satisfied, from the results at which I have already 

 arrived, that it is one of the most profitable sources of new and intei'csting 

 discoveries. 



It may perhaps be oljjected, that the limitation of families depends, in a great 

 measure, upon preconceived views; and that naturalists disagree entirely in their 

 estimate of the natural range of most of them. This is undoubtedly the case 

 at jn'esent ; but let those who think that they may divide animals as they please, 

 conscientiously try to distinguish the different categories of characters of those 

 classes of animals with which they are particularly familiar, and they will soon 

 perceive how much this method will improve their studies, and how easily order 

 will replace the chaotic accumulation of characters which are generally given 

 as diagnoses of the groups they consider, whether they ap^^ly themselves to the 

 description of species or genera or families or orders or classes or branches. And 

 as to families in particidar, they Avill soon find out how fully the term form, 

 understood as the pattern of a definite figure, expresses the general character of 

 families ; and they will also be made to feel how difficult it is correctly to investi- 

 gate the essential elements of these family forms, and what extensive anatomical 

 investigations are requrred before a single one can be satisfactorily described. 

 An acquaintance with the changes which animals undergo during their embryonic 

 growth is particularly useful in this kind of investigations. In Ctenophora?, the 

 family characters rest chiefly in the various combinations of the difierent sys- 

 tems of motory cells which make up the bidk of the spherosome. 



Another objection may perhaps be raised upon the ground, that if every modi- 

 fication in the form of animals is to be considered as the basis of a distinct family, 

 the number of families will be increased beyond measure. Supposing, for a moment, 

 it were so ; if the investigation of the structural elements determining the form 

 should reveal to us, in the course of time, an unexpected number of structural 

 patterns unknown at present, this would be a decided gain for science, and not 

 an objection to our method. But I may say that notliing of the kind is to be 

 apprehended, if I may judge from those classes to which I have thus far jDaid the 

 most special attention. The analytical method I propose has in most cases only 

 helped nie to define with greater precision, families already pointed out or jiartially 

 characterized, and in a few instances only rendered necessary the further subdivision 

 of certain families and the reunion of others. In this connection, it is of the 



