A TAXONOMIC STUDY ON PYKOSOMA — METCALF AND HOPKINS. 215 



cloacal sphincter, and the eighth, reduced in size, goes to the stolon 

 process of the bud. In other respects the distribution of the nerves 

 is somewhat similar in the two subgenera. 



The cloaca is diverse in size and shape (fig. 3, pi. 16; fig. 8, pi. 19). 

 Its aperture is broad and bears on its dorsal edge a flat tapering lan- 

 guet (cloacal tentacle). The cloacal sphincter is incomplete, being 

 broad and well developed on the dorsal edge, and lacking on the 

 ventral edge (fig. 8, pi. 19). 



Tie musculature is unique in some respects (figs. 2 and 3, pi. 16; 

 fig. 8, pi. 19). The pair of muscles found one on each side of the cloaca 

 in the Pyrosomata ambulata, are represented here by similar bands, 

 but lying over the peribranchial spaces, near their middle. A vari- 

 able number of branching and anastomosing muscle fibers surround 

 the oral sphincter. Between the mouth and the ganglion, and be- 

 tween the mouth and the endostyle, are muscle fibers which branch 

 on each side of the body. These constitute the lateral muscle system, 

 which is peculiar to this group of the Pyrosomas. 



Gonads. The testis forms between the stomach and the intestine 

 (fig. 9, pi. 20; fig. 4, pi. 17), and consists of numerous, short lobes radi- 

 ating toward the ventral side of the zooid. The egg lies posterior to 

 it. The large embryo is detached from the parent relatively early. 

 Zooids protandrous. 



PYROSOMA AGASSIZI (Ritter and Byxbee, 1905). 



Platea 16, 17, 18, and fig. 42, plate 33. 



This species was first described from material taken from the Pacific 

 Ocean. Subsequently it has been reported from the Atlantic and 

 Indian Oceans. Our specimens come from the Pacific. We have 

 seventeen in all, from- ten different stations, and these agree very 

 closely with the original description of the species. Some are pre- 

 served in alcohol, others in formalin. 



The colony is rather long and slender, and is always limp because 

 of the soft, gelatinous nature of the test. The test itself is very trans- 

 parent, the zooids being rather opaque. They are yellowish in the 

 colonies preserved in alcohol, whitish in the formalin specimens. 

 Usually there are present, at the open end of the colonial tube, four 

 rather large processes of the test which are of variable length, and are 

 quadrangular in cross section (fig. 6, pi. 18). One angle is toward the 

 colonial aperture, one is on the opposite side, and two are lateral. 

 The outer and the lateral of these ridges continue back onto the sur- 

 face of the colony, which bears generally twelve ridges near its open 

 end. Toward the closed end of the colony the arrangement of the 

 ridges is less regular, instead of twelve, ten or less being found. Some 

 of our colonies have the ridges less marked. The ridges at the closed 

 end of the colony and those at the open end are for the most part not 



