' t 



tl„- tme that in the oldest yel observed the fourth couple is shorter than 



the second, but this may justifiably be regarded .is due to the known stages being too young 



to show the definite length relations of the mesenteries. The fourth couple is formed later than 



th,- ; and must for a time therefore, be shorter than the second in all species, and it ma) 



ed that its shortness in the tonus under consideration is .1 condition whicfa will later 



,1. in 1 bermudensis we certainly cannot consider the arrangemenl seen in the 



oldest know larva .is necessarily representing the aduh relations, and this is equally true for 



Da lermis and D. elegans and perhaps also for Solasteractis macropoda in regard i" 



which there is yet no information as to the development <>( tin- acontia. It is true that in the 



oldest known examp Ovactis brasiliensis described l>y van Beneden the second couple is 



still longer than the fourth, hut in the monstrous form of the same species it is noticeable that 



one side of the sagittal plane the mesentery of the fourth couple is the longer, and this 



being the relative length of the couple in other species of the genus, it may be regarded as 



indicating what the adult relations will be. 



There are, then, in the protocnemes of quite an extensive series of forms two associated 

 iliarities which seem to be constant, and are therefore to be regarded as characteis oi 

 taxonomie importance. Hut before taking up this question further I wish to consider the arran- 

 nent of the deuteroencmes and certain other structural features ot the order. 



The Deuterocnemic Mesenteries. The development of our knowledge of the 

 arrangement of the deuterocnemic mesenteries has already been described and the important 

 dis of Faui 91) of the quartette arrangement has been noted. It has also been pointed 



out that Faurot in his later paper 1 [895) and \ \n 1!im i>i \ [898 begin the enumeration of the 

 (|iiariett<-s with the last protocnemes, a method which is inconsistent with the true relationship ot 

 the two sets of mesenteries. Correctin^ this error, in each of the species hitherto described in which 

 the quartettes have been sufticiently Studied each is composed of two sterile and two fertile 

 mesenteries, the sterile and the fertile alternating. In the different species considerable differen> 

 occur in the relative lengths of the mesenteries composing each quartette, and it seems probable 

 that these differences may have a classificatory value, although this cannot be definitely determined 

 until a greater number of individuals of each species has been studied. 



B( • illing attention to these differences a few words are necessary as to the 



of the quartette mesenteries. Von Heider's terms Filamentsepta and Genitalsepta 

 are not quite opposite for several reasons, the fact that in one of the species to be described 

 later the Filamentsepta as well as the Genitalsepta bear reproductive organs being the most 

 important one. and the same fact makes the use of the terms sterile and fertile also inappropriate. 

 In what follows I shall term the longer mesenteries (von Heider's Genitalsepta) macrocnemes 

 and the- shorter ones von Heider's Filamentsepta) brachyenemes. 



In C. membranaceus the first quartette, that is to say that nearest the protocnemic group, 



f first a brachyeneme, to which succeeds a macrocneme, then a d brachyeneme 



horter than the first and then a second macrocneme longer than the first one 



111 . [f the brachyenemes be represented by b and B lor the shorter and longer 



m< tively and the macrocnemes be similarly represented by m and .1/. the 



'4 



