10 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



evorj' sort, of combinatiou is possible, ami a vcrj' large variety is found, though the 

 tendency is for tlie characters to form more or less definitely correlated groupings, 

 and to crj'staUizc into certain tlofinito types. 



In other genera (raaiiily raacrophreate) all the characters are more definitely 

 correlated with each other, and then the nodes on the species curve will be found 

 to be ver\^ shar]) and almost or entkely distinct from each other, the various species 

 indicated e.\lubithig little or no tendency toward intergradation. 



This typo of variability is not cormected with the geographical origin of the 

 specimens except hi a very' general way, and tlierefore the several forms can not be 

 considered as subspecies as that term is commonly understood. It is practically 

 confined to the multibracliiate Oligophreata, and to specimens of oligophrcate 

 species from the East Indian region. These same species when extending their 

 range outside of this region gradually become more fixed and definite in their 

 characters, so that individuals from, for example, Madagascar or southern Japan 

 will all be fouml to be practical^ uniform in their various features, and to represent 

 the mean of the two extremes seen in a scries from the central East Indian region. 



The recent representatives of a few families appear to have suddenly deviated 

 from any ty])e which we might reconstruct as the ])hylogenetic stock whence they 

 had been derived by a process of "explosion" of their characters which have become 

 recombined in a curiously unbalanced manner, exactly as we see to be the case in 

 several fossil groups. A tendency to form an exi)losive or very aberrant offshoot 

 is more or less evident hi every group of animals, but it rarely affects more than a 

 small minority of the genera or of the species. 



An earnest effort has been made to avoid the common error of takmg into 

 account only ob\'ious dilTerential characters, thereby becoming blmded to the less 

 obvious, but often more reliable, systematic features, by carefully examming every 

 <letail of the animal and every pomt offered by its structure apart from all the 

 othei-s, though hi many cases, so far as reganls comi)arative descriptive work, 

 no use has subsequently been found for the data acquired. 



Great care has been used in the selection of new generic names, and especially 

 hi the selection of the types of new genera; the types arc, whenever possible, the 

 first species to have been described, and the commonest species; but in cases where 

 the original description is deficient or the identification doubtful I have taken one 

 of the later species, where circumstances permitted one considered as a synonym of 

 the first. Preference has always been given to species at hand to guard against the 

 possibility of nomenclatorial disturbance through misconception of species not 

 personally known to me. 



EMBRYOLOGY, DEVELOPMENT AND ANATOMY. 



The systematic study of the comatuhds is, no less than that of other grou])s, 

 based largely upon a knowledge of the development and of the external and uiternal 

 anatomy; the comatuUds, through uniformity of habit, are all built upon the same 

 general plan, and hence the knowledge of their development and anatomy must 

 be comparatively exhaustive m order that the systematic differentiation, at first 

 sight apparently very slight, may properly be appreciated, when it becomes obvious 



