Studiea on the Development of Larvai Nephridia. 501 



of tlie rectum [tr.c). This is from the first unpaired, and as it 

 rapidly grows it surroimds the rectum on either side, so that in 

 advanced stages these lateral portions meet on the median ventral 

 line, forming the ventral mesentery. Thus tlie trunk coelom soon 

 comes to consist of a large unpaired dorsal portion, and two lateral 

 parts surrounding the gut (figs. 33, 39, 41). 



In a series of coronai sections, represented in pl. 31 figs. 19 to 

 24, taken through a young A. hrauchiata^ the shape of this cavity 

 can be clearly followed. 



The first sectiou passes through a ventral plane, so that the 

 ventral wall of the gut just appears cut in the section. The ventro- 

 lateral portions of the coelom (tr.c.v) appear as two narrow slit like 

 sacs one on either side of the gut. As the sections gradually pro- 

 ceed dorsalwards these lateral portions get larger tili they finally 

 joiu in the dorsal cavity [tr.c.d]. While the coelom is surrounding 

 the gut it is at the same time extending on to the dorsal portion 

 of the stomach, so that it soon comes to fili up all the remaining 

 space in the posterior end of the larva, There -is ne ver any trace 

 of a dorsal mesentery in the dorsal portion of the coelom (figs. 14, 35, 

 40, 42, 45). CowLES (6) in a late stage of A. arddtecta^ has figured 

 (fig. 5) what may he a dorsal mesentery in this cavity, but it is of 

 a very doubtful nature, and has, I think, no hearing on the double 

 origin of the cavity. 



In larvae of the same age and size, the shape of the coelom 

 varies very much. In some it is well developed, while in others of 

 the same relative size it is still small. In some it is even imper- 

 fect. In these last it is the dorsal wall that seems to be wanting, 

 so that in these larvae the coelom seems to consist of two lateral 

 portions surrounding the gut in communication with a dorsal cavity 

 which is in free connection with the original blastocoelic space. I 

 believe that this condition is quite abnormal if it really exists during 

 i! life. It is certain that in the majority of larvae it is possible to see 

 Ij both in sections, and still more readily in suitably cleared whole 

 preparations , that the dorsal wall of the coelom is quite intact 

 throughout. If this condition does exist during life, and De Selys 

 LoNGCHAMPS Claims it does, I think no importance is to be attached 

 to it. De Selys Longchamps has never been able to find this 

 cavity perfect in the early stages, and in the dorsal region and 

 also in an anterior direction, it is always up to a later date in free 

 communication with the blastocoel. Certainly, in most of the larvae 



