314 
PACIFIC SCIENCE, Vol XIII, October, 1959 
American mainland. This distributional pat- 
tern seems to be related to the bottom to- 
pography within the south temperate zone 
of the Pacific: there is a very deep trench off 
the Chilean coast, but from the vicinity of 
islas Juan Fernandez and San Felix westward 
numerous submarine ridges and peaks rise to 
various heights below and above sea level. It 
may be predicted that the species will be 
found not only on the guyot nearly 600 miles 
northwest of Juan Fernandez, but also on 
other rises between the offshore islands of 
Chile and New Zealand. It is not known 
whether seamounts intervene between Juan 
Fernandez and San Felix, and it is now 
doubted that the long charted intervening 
ridge exists. 
This young specimen differs in a number 
of respects from the characters indicated in 
the original description and figure of Trigia 
picta, but these differences seem attributable 
to age and individual variation. The dorsal 
spines number VIII instead of VII, but the 
eighth is short and very slender (and is per- 
haps concealed in the type); the number of 
dorsal soft rays (12) agrees. The anal rays are 
11, not 12, but some variation is expected. 
(The principal caudal rays number 6 T- 5 = 
11, and each pectoral has 15 rays, in the 
formula i, 9, ii + 3.) The nuchal spine is 
much longer, reaching to below the base of 
the third dorsal spine (probably as a function 
of youth), and is about as long as the humeral 
spine, (The pectoral fin, as in the type, reaches 
to about the seventh anal ray.) The preopercu- 
lar spine (at the end of the cheek ridges) is 
double, rather than single, but the lower point 
is definitely the smaller, and it probably be- 
comes obsolete in larger fish. The body is 
similarly marked with light-ringed blackish 
spots, but these spots (as expected) are fewer. 
On the head the spots are discernible, though 
faint, as though just developing. The fins, 
unlike those on the type, are unspotted: in 
the spinous dorsal the membranes are sooty 
outward; the second dorsal has a submarginal 
dark band; the pectoral is very dark, except 
for a light border; the other fins are almost 
wholly clear. In coloration the young speci- 
men corresponds better with the figure given 
by Mann (1954: 309) for the ”pez mariposa 
de Juan Fernandez,” which he designated 
Chelidonichthys pictus. 
Trigia pkta has been referred, properly I 
think, to the genus Pterygotrigla Waite (1899: 
108), which was based on the Australian spe- 
cies Trigia polyommata Richardson, Although 
quite different in coloration, P. pkta and P. 
polyommata seem very similar in structure. 
The genus Pterygotrigla has been attributed 
to Japan (Matsubara and Hiyama, 1932: 8-14, 
figs. 2-5), but the Japanese species referred to 
the genus differ trenchantly from P. pkta and 
P. polyommata in lacking vomerine teeth. 
They are also said to have a "distinct anal 
spine” (a hardened soft-ray?), which does not 
seem to be true of the Southern Hemisphere 
species. Though the two groups agree in 
many characters it seems advisable to put 
them in separate genera. The name Otohime 
Jordan and Starks (1907) is available for the 
common Japanese species hemisticta (the 
type species), and probably for Pterygotrigla 
ryukyuensis Matsubara and Hiyama. 
Otohime differs further from Pterygotrigla in 
the small size of the dorsal bucklers. The 
distinction in the first anal ray may not be 
valid, for a young specimen of Otohime hemi- 
stkta has this ray paired, articulated, and 
flexible. 
It seems clear from the original account 
that Trigia guttata, which was described with- 
out comparisons by Philippi (1896: 375- 
376) from Juan Fernandez, is a synonym of 
Pterygotrigla pkta (Gunther) — as was assumed 
by McCulloch (1929: 393). But the gurnard 
from Islas Juan Fernandez that was figured 
and discussed by Mann (1954: 309) as Trigia 
punctata Philippi obviously belongs in a very 
different species. From the figure it seems 
clearly referable to Chelidonichthys, as defined 
by Matsubara and Hiyama (1932: 4). It even 
seems probable that the specimen may be re- 
ferred to C, kumu (Lesson and Garnot), an 
