322 
(1938). Ciliated groove: An open crescent; a 
broad languet below ciliated groove at junc- 
tion with gill (Tokioka, 1938, fig. 5). Viscera: 
Gut forming elongate loop in which course 
of intestine is obscured by large number of 
testis lobules; in condition D of Ihle and 
Ihle-Landenberg (Tokioka, 1938); gut con- 
nected by long mantle projection with pos- 
terior gut projection; the mantle projection 
extending into median posterior tentacle, if 
present, as shown by Tokioka (1938); endo- 
style weakly sinuous; embryos two, lateral to 
joint of M III-IV on right side. 
All the solitary forms described by Met- 
calf (1918) are probably one species, as Sewell 
(1926) believed. Oka (1921), however, recog- 
nized two species, Salpa {Traustedtia) multiten- 
taculata and 5. (T.) henseni, the latter being 
Metcalf’s S. (T.) radiata, on the basis of the 
presence of eight tentacles in the former (ac- 
cording to my classification, 13, as he did not 
include in his count two paired and one un- 
paired pointed processes which are essentially 
similar to tentacles) and 20-23 in the latter. 
The conclusion that there is only one species 
is supported by the great variation in number 
of tentacles (varying from 7 to 25 in POFI 
specimens) and the probability that the radi- 
ate forms described by Apstein (1906^) and 
Dober (1912) simply were flattened dorso- 
ventrally, giving the impression of a radiate 
form; this probably also explains the hori- 
zontal position of the gut of these specimens, 
as Sewell (1926) remarked. Sewell stated that 
differences in details of musculature are often 
due to conditions of preservation, which per- 
haps is the cause of differences described by 
Metcalf (I9I8). Both Oka (1921) and Sewell 
(1926) considered Metcalf’s subspecies bicris- 
tata to be invalid. In POFI specimens the 
crested condition is one of intergradation, but 
crests are always present, apparently due to 
age differences. In one small (2 mm.) indi- 
vidual the crests are very small; they become 
more pronounced in larger individuals. 
The only basis for considering the various 
forms distinct, according to Oka (1921) and 
PACIFIC SCIENCE, Vol. VIII, July, 1954 
Sewell (1926), is the different number of ten- 
tacles. Sewell (1926) distinguished between 
"true tentacles” and projections that pene- 
trate the test but do not extend beyond the 
body margin. This, however, is probably only 
a condition of age and is not considered a 
useful distinction except for descriptive pur- 
poses. Oka (1921) remarked on the probable 
similarity of his "tentakeln” and "kurze . . . 
Anhange” which projected from each side- 
wall of the cloacal opening. He also referred 
to "verborgene Tentakel” (those projecting 
into the crests and the unpaired posterior 
one). He did not describe any other projec- 
tions. In the largest POFI specimen (Fig. 
2‘ba) all projections are of the true tentacle 
type, although many project only slightly be- 
yond the test margin. In Sewell’s (1926) and 
Thompson’s (1948) specimens, there were 
nine tentacles present, including those not 
projecting beyond the body margin; in Ap- 
stein’s (1894^) specimen there were eight, 
but, as Sewell (1926) remarked, there possibly 
were 13; in Metcalf’s (1918) specimen and 
in Oka’s (1921), 13 tentacles; and in Apstein’s 
(1906^) and Dober’s (1912) specimens, there 
were 20 and 23 tentacles present. Sewell (1926) 
thought there was a possibility that the differ- 
ences were due to either geographic variation 
or variation in size. POFI specimens seem to 
confirm the latter conclusion as the number 
of tentacles is less with the smaller individuals 
and greater with larger ones. The smallest 
(2 mm.) specimen examined bore no "true 
tentacles” but bore seven mantle projections: 
one unpaired posterior; one pair of postero- 
lateral projections (the "posterior tentacles” 
of Sewell and Oka and tentacle 6 of Metcalf) ; 
one pair of projections in the cloacal siphon 
(5 of Metcalf); one pair in the posterior part 
of the crests (4 of Metcalf) . 
Further, in a specimen of 7 millimeters 
length, the projections are still not "true ten- 
tacles.” This specimen is so poorly preserved, 
however, that I am unable to ascertain the 
number of projections present. 
In another specimen, 11 millimeters long, 
