GIBSON — CEPHALOCHOEDA : " AMPHIOXIDES." 233 



perigonadial coelom, is represented only by a slight downward bulge of the posterior end 

 of the myotome in which each gonad lies. 



In the great majority of specimens, including all those of A. valdivice, gonads could 

 not be seen at all. In sections of an A. valdivice 10"6 mm. long there was no trace of 

 them ; Goldschmidt, however, found them, in an early stage of development, in animals 

 only 7'5 and 8 mm. long. In A. pelaglcus their differentiation appears to commence 

 when the animal is about 9 mm. long ; they occur in all the animals alluded to on 

 p. 224, as probably rather older than the rest, and in a few of the more normal 

 specimens. In one of the former category, 9-1 mm. long, they could be seen in the 

 liinder corners of segments 17-42 (their mother-segments therefore being 18-43) ; in 

 one of the latter, 8-9 mm. long, they are found in segments 20-43. In both animals 

 traces of them could be seen in a few more anterior segments in the shape of cell- 

 tliickenings still lying in the antei-ior wall of their parent myotomes, as in Boveri's 

 ligs. 3-6. 



Finally, Goldschmidt found the gonads on the right side only. In this I can fully 

 corroborate him ; no traces of them were ever to be seen on the left. This character is 

 important as allying Amphioxides to Asymmetron and Heteropleuron in contradistinction 

 to Branchiostoma. 



(2) The Metamorphosed Animals. 



Of these there were unfortunately only two in the collection. They present many 

 points of agreement with A. pelaglcus, and I am inclined, in spite of certain difficulties, 

 to assign them to the same species. To the discussion of their nature we shall return 

 after describing them. 



External Characters. 



These were examined in the smaller of the two specimens shown in PI. 15. fig. 2. Its 

 measurements, when mounted, were 8X'64 mm., the post-anal length being 1-57 mm. ; 

 thus the ratio of length to breadth — 12-5 — though large, is not unparalleled in 

 A. pelaglcus {vide table on p. 218) ; that of total to post-anal length — 5-1 — is a lower 

 hgure than was ever found in the latter. 



Apart from differences due to metamoi-phosis, there is a considerable general 

 resemblance to A. pelaglcus. Thus the fin-ray boxes have a similar character and 

 arrangement, disappearing some little distance behind the anterior end of the nervous 

 system. The dorsal fin-membrane starts at the same level, and there are similarly 

 disposed fin-strands. The tail, however, is longer in proportion, and a good deal more 

 tapering and sharply pointed. The number of myotomes appears to be 70, of which 

 16 are post-anal, a larger number than was ever found in A. pelaglcus. 



The atrium is evidently completely closed and extends back a long way behind the 

 gill-region to the atriopore, situated opposite the 46th myotome. Just in front of the 

 latter the thickened rectal patch, so conspicuous in Amphioxides, can still be seen. 



