316 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1924 



able depths off Celebes, one near Menado on the north coast, the 

 other in the large Tomini Baj- on the east coast. The lengths were 

 53 and 55 mm. Both specimens were preserved in formol and well 

 enough for sufficiently accurate determination of the numerical char- 

 acters. Reserving detailed description for a later occasion, I will 

 here refer to Figures 1 and 2, plate 7, and give the following data, 

 determined by Mr. A. V. Taning, M Sc. 



The specimen of 53 mm. is a fully developed larva, metamorphosis 

 not yet begun, larval dentition intact, 16-17 teeth in each half of 

 jaw behind the large grasping teeth; in other words, about the 

 same number as in A. vulgaris and rost7'ata. It has 106 (69+37) 

 myomeres, answering to 105 vertebrae, 285 dorsal, 218 anal, and 10 

 (5-|-5) caudal rays. 



The other specimen, 55 mm. (more correctly 55.5 mm.) has begun 

 metamorphosis and lost the larval teeth. Number of myomeres 

 about the same as in the first larva, about 106; like the Atlantic 

 species at the same stage, it has some black stellate chromatophores 

 at tip of tail. 



The numerical values of the larvae enable us to identify them, at 

 least the first specimen, with sufficient certainty as belonging to the 

 species AnguiUa niauritiana Bennett, a widely distributed tropical 

 species, common in the Dutch East Indies. These values agree com- 

 pletely with those we have found in AnguiUa mauHtiwna^ and pre- 

 clude other species, especially the two common in this area — AnguiUa 

 cmstralis Autt. and AnguiUa celebesensis Kaup. The advanced posi- 

 tion of the dorsal fin in our larva (fig. 1, plate 7) also shows that it 

 belongs to AnguiUa niawitiana. 



The identification of the larvae from Celebes reveals several inter- 

 esting points which I will briefly note here. 



1) We have learned that the larvae of Indo-Pacific fresh-water 

 eels entirely resemble those of the Atlantic species. This is a fur- 

 ther proof of the extremely close relationship between all AnguiUa 

 species, for AnguiUa mamntiana is probably one of the fonns most 

 widely removed from ours. 



2) The length of the AnguiUa mau7'itiana larva is about 514 cm. 

 This agrees well with the average length of five elvers of this species 

 from the Dutch East Indies (Zool. Mus. Amsterdam), about 5 cm. 



3) The identification of the larva of AnguiUa niauritiana is an 

 incontestable proof that the Indo-Pacific AnguiUa species, like the 

 Atlantic, breed in the sea and undergo a larval development with 

 metamorjohosis. Other facts which I have brought to light show 

 that the Indo-Pacific fresh-water eels, like those of the Atlantic, 

 require deep water to breed. I will deal with this point in a sub- 

 sequent paper. 



