LARVAE OF DECAPOD CRUSTACEA 393 



maxillipede 3 and legs 1-5, and stated that pleopod i is absent and that leg 5 is biramous. 

 His conclusion that the oldest individuals are mature and related to Sergestidae was 

 quite unjustified. 



Sund's proof (1915) that Eryoneicus is the larva of Polycheles disposes of Boas' 

 suggestion, and the real position of Amphion remains as uncertain as ever it was. 



I have suggested myself (1924, p. 105) that Korschelt and Heider were right in 

 referring Amphion to the Caridea. That the oldest known specimens are immature 

 there can be no doubt. The form of all the appendages is definitely larval, and even if 

 the presence of a gonad can be demonstrated, that does not necessarily imply maturity. 

 It is known that certain carid larvae of the high seas continue to grow and retain larval 

 characters until the sexes are recognizable by the form of pleopod 2, and it is probable 

 that the same is the case with Amphion. If Amphion is a carid, as its gills and mouth- 

 parts so strongly suggest, the only known genus to which it can possibly be attached is 

 Amphionides, Zimmer, of which a description is given below. 



Amphion 



The Discovery material contains about a hundred specimens of Amphion, mainly of 

 the later stages. The smallest specimen, in which the three pairs of maxillipedes only are 

 developed, and without uropods, represents the earliest stage described (Bate, 1888, 

 pi. 146, fig. i), but it appears from the condition of the eyes and antennae, and the 

 presence of rudiments of the uropods under the cuticle, to be actually stage II (Fig. iza). 

 The next stage, of which there are two specimens only, agrees with the normal stage III 

 of Caridea in having uropods with the endopod not fully developed. Among the older 

 specimens it is possible to distinguish sk stages, making nine in all, but these stages are 

 not well defined and it is not possible to fit every individual into a clear-cut group. For 

 instance, in two specimens in which the legs may be at the same stage of development, 

 the antennules may be much more developed in one than in the other, and the appearance 

 and size of the pleopods does not always correspond in specimens judged on other 

 grounds to be in the same stage. Assuming a growth factor of about 1-26 the range of 

 size between the smallest and largest specimens may be divided into eight stages 

 corresponding very closely with the average sizes observed, so that I feel confident that 

 a total of nine stages, or eight moults, is correct, even though there may be great 

 variation in the degree of development attained by individuals at each moult. 



The rudiment of leg 5 may be quite distinct in specimens smaller, and apparently 

 younger, than others in which it is not present at all, and it may still be absent in speci- 

 mens as large as 23 mm. There are only about fifteen specimens old enough for these 

 details to be made out, and of these five only have no trace of leg 5. In these five pleopod 

 I has no endopod, whereas it is distinctly biramous in all the others. It seems possible 

 that the differences in these two characters may be indicative of sex, those without 

 leg 5 being males. 



Both Bate and Suhm believed that they could recognize a testis, and Koeppel de- 

 scribed an ovary, but the testis of Bate and the ovary of Koeppel were probably really 



