THE ANATOMY. RESPIRATORY ORGANS 83 



processes, and one on each side passing obliquely across the hood ; they then bend 

 upwards and become continuous with the dorsal vessel ; each branchial capillary 

 runs up the branchia to its tip and then bending upon itself runs down the opposite 

 side of the branchiae ; these and the other branches fuse together and form the 

 dorsal blood-vessel. 



In Dero obtusa the circulation in the branchiae is somewhat simpler; it was 

 described and figured by D'UDEKEM (1). The ventral vessel divides as in D. perrieri 

 but simply pursues a sinuous course, passing up and down each branchia ; the right 

 and left halves of the circle unite at the base of the branchial hood to form the 

 dorsal vessel. Dero digitata has more complicated vascular loops than even D, 

 perrieri. The arrangement of these has been worked out by STOLC (2). Each of 

 the four branchial processes has two capillary loops instead of only one ; and there 

 are two circular vessels, derived from the ventral vessel, which form complete circles 

 and do not communicate with the dorsal vessel except by the recurrent loops from 

 the branchiae. 



It seems clear from the habits of the worm and from the structure of these 

 circumanal processes that they must be regarded as branchiae ; the expanded hood 

 which is itself in some species furnished with elongated processes is no doubt also 

 branchial in function. The next Oligochaet, in order of discovery, which is 

 branchiate, is Alma nilotica. In the systematic part of this work I discuss 

 whether this form (recently re-described by LEVINSEN as Digitibranchus niloticus) be 

 really an Oligochaet. The conclusion arrived at is that it is an Oligochaet. The 

 branchiae, like those of Dero, are found at the posterior end of the body. It is very 

 desirable that this form should be re-investigated, for at present there are no details 

 concerning the structure of the supposed branchiae to hand. There is, however, 

 hardly room for doubt as to the branchial nature of the short cylindrical processes 

 which are found upon the last sixty or seventy segments of the body. They occur just 

 to the dorsal side of each dorsal pair of setae and there are four or five on each side, 

 sometimes simple and sometimes branched. Chaetobranchus semperi is a remarkable 

 branchiate Naid, apparently known to SEMPEE, and lately described by BOURNE 

 (1). I can confirm from personal observations BOURNE'S description of the branchial 

 organs, which, unlike those of Dero, are confined to the anterior part of the body. 

 The most anterior, after the first one or two pairs, are the longest ; and in those 

 anterior ones are imbedded the long capilliform setae of the worm ; further back the 

 setae are independent of the branchial processes. The inclusion of the setae within 

 them is not unsuggestive of a comparison with parapodia of the Polychaeta. These 

 branchia in Chaetobranchus are ciliated externally, they contain a capillary loop 



M 2 



