LEADING GROUPS OF EXISTING SYSTEMS 165 



zoa, Tunicata, Brachiopods, and Lamellibranchiata would be also 

 very generally considered to be natural orders. Among Echinoderms, 

 I suppose Crinoids, Asterioids, Echinoids, and Holothurioids would 

 be conceded also as such natural orders; among Acalephs the Beroids, 

 and perhaps also Discophoras and Hydroids; while among the Polypi, 

 the Halcyonoids constitute a very natural order when compared 

 with the Actinoids. 



Let us now consider these orders with reference to the character- 

 istic forms they include. The forms of the genuine Testudo, of 

 Trionyx, and of Chelonia are very different, one from the other, 

 and yet few orders are so well circumscribed as that of Chelonians. 

 The whole class of Fishes scarcely exhibits greater differences than 

 those observed in the forms of the common Sharks, the Sawfishes, 

 the common Skates, and the Torpedo, not to speak of the Cyclo- 

 stomes and Myxinoids, if these families were also considered as 

 members of the order of Placoids. Ganoids cannot be circumscribed 

 within narrower limits than those assigned to them by J. M tiller, 

 and yet this order, thus limited, contains forms as heterogeneous as 

 the Sturgeons, the Lepidosteus, the Polypterus, the Amia, and a 

 host of extinct genera and families, not to speak of those families I 

 had associated with them and which Prof. Miiller would have re- 

 moved, which, if included among Ganoids, would add still more 

 heteromorphous elements to this order. Among Decapods we need 

 only remember the Lobsters and Crabs to be convinced that it is not 

 similarity of form which holds them so closely together as a natural 

 order. How heterogeneous Bryozoa, Brachiopods, and Tunicata are 

 among themselves, as far as their form is concerned, everybody 

 knows who has paid the least attention to these animals. 



Unless, then, form be too vague an element to characterize any 

 kind of natural group in the animal kingdom, it must constitute a 

 prominent feature of families. I have already remarked that orders 

 and families are the groups upon which zoologists are least agreed 

 and to the study and characterizing of which they have paid least 

 attention. Does this not arise simply from the fact that, on the one 

 hand, the difference between ordinal and class characters has not 

 been understood and only assumed to be a difference of degTce; and, 

 on the other hand, that the importance of the form, as the prominent 

 character of families, has been entirely overlooked? For, though so 



