GIBSON — CEPHALOCHORDA : " AMPIIIOXIDES." 241 



hence it is only natural that, in the absence of intermediate stages, its proper assignment 

 should he far from obvious. One valuable criterion remains to us, however, viz., the 

 myotome formula. In view of the constancy of the number of myotomes in the larger 

 specimens of Amphlo.vides, and the fact that even the much smaller Branchiostoma larva 

 already possesses its adult number before metamorphosis, there can hardly be any doubt 

 that Ampldoxkles, as we know it, also possesses its full adult number. 



The following table (p. 242) shows the myotome formulae and distribution of the 

 various species of Heteropleuron, Asymmetroii, and Amphioxldes. 



Two correspondences are at least close enough to be worth considering, and are 

 iurtber supported by coincidences of distribution, viz., those between Astjmmetron 

 lucayanum and Amphioxides pelagicus, and between Heteropleuron parvum and 

 Amphioxides valdivice. We will deal first with the former. 



Cooper (1903) describing as pelagic larvae specimens of Amphioxides pelagicus taken 

 off the Maldives, has already speculated as to whether they could be the larvae of 

 Asymmetron lucayanum — by far the most abundant Cephalochordate of that Archipelago 

 — but decided the question, provisionally, in the negative. One of his reasons for doing 

 so is the absence of an olfactory pit in Asymmetron : since, however, the olfactory pit is 

 a remnant of the neuropore, its absence cannot be a primitive feature, and it is bound to 

 occur in the earlier stages of the life-history ; its retention up to the end of larval life, 

 and subsequent atrophy, would be in no way remarkable. 



Apart from the sufficiently close correspondences in the myotome formulae, a relation- 

 ship between Amphioxides pelagicus and Asymmetron is, as we have seen, indicated by 

 certain anatomical features. With A. lucayanum in particular there is a very remark- 

 able correspondence in the structure of the snout, as Andrews's fig. 10, in the paper in 

 which he first described the species (1893), will show. This figure might almost have 

 been drawn from a specimen of A. pelagicus, so close is the similarity of the long 

 tapering canal which takes the place of fin-ray boxes anteriorly. A further noteworthy 

 correspondence is in the median position of the anus and consequent right-handed 

 position of the ventral fin in its vicinity. 



But perhaps the most remarkable correspondence is that of distribution. A glance at 

 Andrews's fig. 5 is, I think, sufficiently convincing evidence that Amphioxides pelagiciis 

 occurs in association with Asymmetron lucayanum in the Bahamas, as well as in the 

 Maldives. The animal there depicted is clearly recognisable, and in the numerical 

 characters given (length, 6 mm. ; no. of gill-slits, 22 ; myotome formula, 51+13) it 

 agrees excellently with similar-sized specimens of mine {of. Table, p. 218). Andrews 

 assigns it without question to Asymmetron lucayanum ; he may, of course, have been 

 incorrect in so doing, since he did not follow the metamorphosis ; but the occurrence of 

 practically identical adults in conjunction with larvae, also apparently identical, in two 

 localities so widely separated as the Bahamas and the Maldives, is at least highly 

 remarkable, if adults and larvae do not belong to a single species. 



At this point it will be convenient to consider the metamorphosed animals which I 

 have described above. There are undoubtedly some difficulties in the way of regarding 

 them as metamorphosed A. pelagicus. No specimen of the latter was found in which 



