54 BULLETIN : MUSEUM OF COMPARATIVE ZOuLOGY. 



The Malpighian body is squeezed against the peritoneum, and the outer 

 funnel opens slightly posterior and generally mesad to it.^ 



Nephrostomes of Secondary Units. 



Soon after the primary unit has become functional, the secondary 

 begins to go through the same process of development. Its behavior 

 differs from that of the primary in two respects, however : its lateral 

 arm empties into the collecting trunk of the primary unit instead of into 

 the Wolffian duct ; its outer tubule is not formed until the rest of its 

 development is nearly completed. The development of this outer tubule 

 is very difficult to decipher from the fact that, at the period w'hen it 

 begins, the Malpighian body has become imbedded in a mass of primary- 

 tubule coils. I think, however, that I am not mistaken in saying that 

 a hollow evagination, tipped with a solid, conical, deeply staining point, 

 grows out from the neck of the Maljjighian body, and, pushing its way 

 through the mass of tubules, opens on the peritoneum mesad to the 

 primary Malpighian body (Fig. 24, nph'stm. 2). This process begins in 

 larvae about 45 mm. in length. 



The later, more dorsal sets of units develop precisely as do the 

 secondary. Each opens into the collecting trunk of the next ventral, 

 so that there is formed a long compound collecting trunk reaching from 

 the most dorsal set down to the duct. 



Whether or not any or all of the units dorsal to the secondary one 

 produce outer tubules, I do not know. In my oldest specimen^ (from 

 ■which Figure 24 was drawn), which had lost its gills, having left the water 

 and assumed the adult markings, there are occasionally three outer funnels 

 cut in one cross-section. In these cases the two younger nephrostomes 

 are closely approximated to each other. This would seem to suggest that 

 some of the tertiary units had developed outer tubules. The only way to 

 settle this point with absolute certainty w^ould be to follow each unit from 

 its outer funnel to the common collecting trunk. This is practically im- 

 possible. A less certain but fairly satisfactory way would be to plot the 

 outer funnels and all those glomeruli which, from their position, seem to 

 belong to primary and secondary units. By comparing the number of 

 outer funnels thus obtained with the sum of primary and secondary 

 units, an inference could be drawn as to whether these two sets alone 



^ The section figured passes tangentially through the posterior wall of Bowman's 

 capsule {(jlm. 1), which touches the peritoneum in the next anterior section. 



^ I unfortunately have no record of tlie length of the specimen, but should 

 estimate it at about 50 mm. 



