384 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



individuals found iu tbein belong to this variety. The color variations of this form have already 

 been given iu section i. 



The rostrum is sometimes wanting, as iu the individual from which Fig. 11, PI. xxn, was 

 drawn. This variation has been noticed in other species and is interesting, since the absence of 

 the rostrum is a constant character iu a closely related series of forms, which are placed by Dana 

 in a separate genus (Bctcvus). These variations indicate that the uniform presence or absence of a 

 rostrum is a specific and not a generic character, as has already been shown by Kiugsley (Hi)). 

 The structural poiuts of difference between the lomjicarpiis and the other form lie chiedy iu the 

 antenna? and lirst pair of walking legs. These may be seen by a comparison of Figs. 11, 13, IS, 

 PI. xxn, and Fig. .!, PI. xxiv, with Figs. 4, 8, PI. xxm, and Fig. 3, PI. xxiv. 



In the. first pair of antenna} the aural spine (Fig. 11, PI. xxn) is scarcely more than half the 

 length of the first segment of the stem. It is blunt and somewhat ovate in shape, as seen from 

 above. 



(2) In the other form (var. brevicarpus) the aural spine (Fig. 4, PI. xxm) has a different shape, 

 and is relatively nearly twice as long. In this case it extends beyond the first segment to two- 

 thirds the length of the second. The second or outer antenna of the lony ion-pus is armed with two 

 spines at its base (Fig. 11, PI. xxn) ; an inferior and outer basal spine, and a slightly longer one, 

 the sqiKimoiis xjiitu-, articulated to the joiut carrying the latter. 'There is no scale. The basal spine 

 is rather more than one-half the length of the autennal stalk. There may be present a small 

 tubercle on the upper surface of the segment bearing the basal spine, near the articulation. 



In variety hrcricin-jnis (Fig. 8, PI. xxm) the sqiiaiiious spine is stout ami reaches nearly to the. 

 end of the antenual stalk. There also springs from its inner and proximal margin an elongate 

 plate or scale, the inner free edge of which is fringed with plumose seta-; scale not quite as long 

 as spine. The inferior basal spine not one-half the length of the squamousspiue. There is a rounded 

 or pointed tubercle over basal spiuenear the joint. 



The small chela of the first pair of thoracic legs of the longicarpus (Fig. 2, PI. xxiv) is short 

 and broad. The finger ends in two or three horny teeth or prongs, which interlock those of the 

 opposing thumb. The dactyle bears on its outer surface a tuft of peculiar hairs. The latter are 

 fiuely serrate and have bent or hooked tips. The carpus is relatively very long, quite as long as 

 the palmar portion of the propodus. 



In the brer ion-pus the small chela is long and somewhat narrower (Fig. 3, PL xxiv). Tips 

 of fingers usually simple, but sometimes notched: the peculiar tuft of hairs is wanting. Carpus 

 relatively short ; about one-third the length of the palm. 



The large chela of the longicarpus may also differ noticeably from the brcricarpus type. (Com- 

 pare Fig. 8, PL xxiv with the figures on PL IV.) Fig. 8 represents a common form of this append- 

 age. The propodus is long, cylindrical, slightly twisted, very smooth, and polished; ends above 

 dactyle in a short spine and below iu a rudimentary thumb with claw-like tip. Dactyle overreaches 

 propodus, and its inner margin is not concave, or but slightly so. 



These two forms, differing in the particulars just mentioned, would doubtless be considered 

 as two distinct species if only these facts were known. A prolonged study, however, of a large 

 number of individuals, collected both in sponges and from porous rocks on a number of reefs, has 

 resulted in the discovery of a complete series of intermediate links. These connecting forms sug- 

 gest a nuiuber of important questions relating to the causes and significance of variation. 



By far the greater number of individuals of this species have the characteristics of the two 

 varieties just described, but abouttive per cent of the col lection made at different points near Nassau 

 present intermediate characters.* 



More fully stated, the noticeable points of variation are as follows: (1) The relative length of 

 the auteuiiular stalk and aural spine; (2) the lengths of the antenna! spines relative to each 

 other and to the peduncle of the antenna; the presence or absence of a squame or scale; (3) 

 the character of the dactyle and propodus of the small chela of the first pair of pereiopods ; (4) 

 the length of the carpus of the small cheliped ; (5) the general shape and character of the large 



* While tln> .species live, as ;i rule, in tlie interior of the green and brown sponges, a few undersized individuals 

 may be found, by careful .searching, among the loose blocks of porous coral which are scattered over the reef, and it^ 

 frequently happens that these individuals possess intermediate characters hetween the two varieties just described. 



