46 CALIFORNIA ACADEMY OF SCIENCES. 



ature of the antero-posterior axis itself. As compared 

 with the form of P. Listcri (Lister, '34, pi. xi, fig. 2), 

 it is as though the dorsally directed atrial siphon of this 

 latter species had been brought to an anterior direction 

 by bending dorsalward the entire posterior portion of the 

 body. When well distended, the siphons are distinct in 

 specimens removed from the test; and in such cases the 

 marginal lobes of them, though not large, may yet be 

 clearly seen with a slight magnification. In the individual 

 shown in fig. 3 there were six lobes on the branchial 

 siphon and five on the atrial. These are the inost usual 

 numbers, but they are not altogether constant. In several 

 instances I have found six on the atrial also. The size 

 of the lobes and the spacings between them vci'A.y vary 

 considerably. Thus it will be observed in fig. 3, that 

 the interval between two of the atrial lobes is considerably 

 wider and deeper than are those between the others. 



2. THE TEST AND THE ORIGIN OF ITS CELLS. 



In a majority of the sections which I have examined, the 

 test presents a uniformly hyaline matrix, in which are 

 scattered a few cells (figs. 29, 30, 31 and 32, ts., m. c.) 

 The structure is not, however, so simple as this in all 

 cases, for in several instances a vast number of exceed- 

 ingly fine granules have been found in the matrix. This 

 condition I first observed in sections stained on the slide 

 in Delafeld's htematoxylin ; and as the granules appeared 

 to be of the color of the stain, I was inclined to think that 

 they had been deposited from the stain. However, 

 further examination of sections stained by various other 

 methods shows that this, at least, is not their origin. In 

 some cases they are not stained at all, but are seen be- 

 cause of their being somewhat more refractive than the 

 testicular matrix. They are not cut fibers, since they 



