Pearlfishes from Guam—SMITH 
35 
information on habits and characters for proper 
identification of species/' 
Much of the difficulty in identifying pearl- 
fishes lies in the fact that they undergo dramatic 
metamorphoses during their life history. Al- 
though the life histories of most pearlfishes are 
unknown, it is generally assumed that they are 
similar to that of Carapus acus (Brunnich), 
which has been studied in detail by Emery 
( 1880) and others. The floating eggs hatch into 
pelagic larvae characterized by a caudal filament 
(Parr, 1927:133) and a remarkable dorsal ap- 
pendage, the vexillum. At the end of the first 
( vexillifer ) stage, these appendages are lost and 
the larva, now called a tenuis, assumes the in- 
quiline habit. The tenuis is elongate with short 
rounded pectoral fins and has the anus at the 
rear of the body cavity. As the pearlfish trans- 
forms into its definitive form, here called the 
adult, the pectoral become longer, the anus 
moves forward, and there are marked changes in 
body proportions accompanied by a drastic re- 
duction in? total length. Thus, at some sizes, a 
given fish may be in the vexillifer or early 
tenuis stage, the late tenuis, or the adult stage. 
(See Arnold, 1956: fig. 1, and Fig. 2 of the 
present discussion.) Several specimens collected 
in Guam that seemed at first to represent a 
distinct species were found to be the late tenuis 
stages of Carapus homei. 
In October, I960, 1 found pearlfishes inhabit- 
ing the holothurian Stick opus chloronotus 
Brandt in certain areas of the fringing reefs of 
Guam. Periodic collections of this and other 
echinoderms until July, 1961, yielded 230 speci- 
mens from 567 possible hosts. Four species of 
carapids were represented as follows (numbers 
and sizes in parentheses ) : 
Encheliophis gracilis 
(Bleeker) 
Carapus homei 
(Richardson) 
Carapus mourlani (Petit) 
Carapus parvipinnis 
(Kaup) 
(15, 167-220 mm) 
(208, 68-197 mm) 
( 5, 74-94 mm) 
( 2, 63-238 mm) 
Schultz (1960:392) reported Encheliophis 
vermicularis Muller from Guam, but since I 
have not collected it, it will be omitted in the 
following discussion. 
NOMENCLATURE 
The form referred to here as Carapus mour- 
lani was considered by Arnold (1956:274) to 
be the same as C. homei. In structural features 
it seems to be nearly identical with homei, but 
it differs in having superficial melanophores on 
the head and body and it lives in the starfish 
Culcita novaguineae Muller and Troschel in- 
stead of in holothurians. If the differences are 
due to the effects of the different hosts these 
forms may be the same species, although there 
is still the possibility that they are reproductively 
isolated through active host selection. It is per- 
haps significant that only the mourlani form 
seems to occur in Hawaiian waters; however 
this could be due to the scarcity of suitable 
holothurian hosts. For the moment it seems pref- 
erable to follow Schultz ( 1960:393 ) in recog- 
nizing two species, although he used the name 
mourlani with some hesitation since he did not 
see the type which came from Madagascar. 
The nomenclature of other species is that of 
Arnold (1956). 
It is usually stated that carapids have no 
pelvic girdles, but alizarin staining reveals the 
presence of small rodlike bones between the 
lower ends of the cleithra. These resemble the 
pelvic bones of Dinematic hthys (Gosline, I960: 
fig. 3b) except that there are no pelvic fins. 
Alizarin staining also reveals a row of deeply 
imbedded plaques along the lateral line. These 
seem to be vestigial scales and are found in 
Encheliophis gracilis, Carapus mourlani, C. 
homei, and in a slightly different form in C. 
parvipinnis. 
The five species reported from Guam can be 
distinguished by means of the following key 
and the features summarized in Table 1. 
HOSTS 
Pearlfishes were found in four species of holo- 
thurians and in the pillow starfish, Culcita no- 
vaguineae. Occasional sampling of nine other 
species of holothurians revealed no inquilines 
but the numbers sampled (except Holothuria 
atra, see below) were small and low incidences 
could have been missed. The numbers of hosts 
examined and inquilines found are summarized 
in Table 2. 
