144 



THE AMERICAN NATURALIST 



[Vol. LIII 



derived from the Ctrlenterata-Porifera group— i. e., from 

 a ccelenterate (cnidarian ?) type of animal through 

 Ctenophora-like (?) forms; or more directly from a 

 colonial protozoan type through forms comparable to the 

 "Mesozoa" and their relatives, such as Dicyema, etc.; 

 although there is a strong probability that the lower 

 worm-like forms arose from ancestors occupying a posi- 

 tion intermediate between these types of animals. From 

 the standpoint of evolution, the Kotifera and Platyhel- 

 minthes (also such worms as DinopMlus, etc.) are among 

 the most important of the lower worm-like forms, since 

 they have departed as little as any from the condition 

 characteristic of the ancestors of the "Annelida" and 

 many other invertebrates; and even the line of descent 

 of the vertebrates themselves may ultimately lead back 

 to forms not unlike the members of tliis group. A higher 

 stage of development is represented by the ''Annelida" 

 (including the Sternaspididae, Gephyrea, etc.), which are 

 a group of the greatest phylogenetic importance due to 

 the fact that their line of development is approached by, 

 or is paralleled by, those of many other invertebi;ate 

 groups, and to the fact that they have retained a great 

 number of characteristics apparently typical of the an- 

 cestors of the Arthropoda. Their forebears probably 

 occupied a position intermediate between the Rotifera 

 and the Platyhelminthes, and indeed, some investigators 

 have even placed the ''archiannelid" DinopMlus among 

 the planarian Platyhelminthes, although its closest affini- 

 ties seem to be with the annelidan worms Profodrilus and 

 Polygordius. 



From their nimelid-like forebears, there have branched 

 off two important liiif- of descent, wliich have ap- 

 proached very do-c to the arthropodan type, and which 

 have even luvii cln-c.l among the Arthropoda by some 

 investigatoi->. One (»!' tlic^e lines of development is rep- 



artliropod-liko in many jiarticulnrs. altliough they have 

 retained many annelidan characters; while the other line 



