INTESTINAL RESPIRATION IN ANNELIDS. 787 



determinations of oxygen and carbon dioxide, and so do not directly touch the present 

 subject. 



(2) Refipiration by the introduction of water at the anus into a ventral groove of 

 the intestine, or into a ventral groove and accessory intestine. 



The chief source of our knowledge of this phenomenon is Eisig's work on the 

 Capitellids (17). 



In Notomastus, a worm about four or five inches long, a ventral ciliated groove 

 begins at the anus, and extends forwards for a distance of two centimeters ; here an 

 accessory intestine (Nebendarm), which may be regarded as the forward continuation 

 of the groove, separates from the gut, which it accompanies on its ventral side as far 

 as the posterior end of the oesophagus, where it again joins the alimentary tube. 

 This accessory channel lies close to, or at some distance from, the intestine ; in the 

 latter case it is connected with the intestine by a sheet of peritoneum. It has the 

 same structure as the intestine ; it contains no food, but may be filled with structureless 

 matter cast off by the lining cells ; its epithelium is furnished with minute cilia, which 

 are recognisable only with difficulty. The groove in the posterior part of the intestine 

 is provided with long cilia, which work in an ascending direction — an arrangement 

 probably serving to introduce into the accessory gut water which enters the intestine 

 at the anus. The animal possesses gills ; intestinal respiration is therefore not of such 

 importance here as in Capitella, which has none. 



Dasyhranchus is similar ; so also is Mastohranchus , in which the accessory gut 

 keeps close to the intestine. In Heteromastus the connection is still closer ; here the 

 cavity of the accessory gut often bursts through into the intestine, and sections may 

 thus show the " Nebendarm " merely as a groove of the intestinal wall. 



In Capitella the intestinal groove is well marked, and extends through one-third of 

 the body length ; its epithelium is high and markedly ciliated. The accessory gut lies 

 close to the intestine, and here also often appears, through the solution of the separating 

 layer, as a ventral groove of the intestine. It is continued forwards by an oesophageal 

 groove, lined by a low epithelium, which is ciliated either only very faintly or not at 

 all. In this animal the hinder portion of the intestine frequently contracts rhythmically, 

 or in a " pumping " fashion. The ciliary current can often, especially in young animals, 

 be followed upwards as far as the region of the "Nebendarm." Capitella may some- 

 times be seen to wave the hinder end of its body about in the water, like certain aquatic 

 Oligochseta ; it has no gills, and so depends entirely on cutaneous and intestinal 

 respiration. 



Generally in this family the accessory gut has a diameter about one-fifth that of 

 the intestine, with which it corresponds in structure. It never contains food ; but it 

 may, like the intestine, be filled with a spongy mass which has been cast off by the 

 epithelial cells. 



The question of the comparative anatomy and homologies of the accessory gut, dis- 

 cussed at some length by Eisig, lies rather apart from our present subject. We may briefly 



