50 THE ECHINODERMS OF TORRES STRAIT. 
young of a Certonardoa; at any rate, the species may be placed in that genus. The species 
frianti is a well-marked one but lemonnieri is, I believe, undoubtedly synonymous with 
mollis. Koehler’s reference to ‘‘N. indica (Perrier),”’ p. 171, apparently means Fromia 
indica, although on page 140 and on plate xvii it is considered a Fromia without a hint 
that it might be a Nardoa. The species ‘‘N. bellone de Loriol” (p. 164-165) is purely 
imaginary and almost certainly refers to N. mollis, which is figured by de Loriol on the 
same plate with Luidia bellone Liitken. In 1917, Fisher described a Nardoa tumulosa 
from the Sulu Archipelago, which is allied to friantt but seems to be quite distinct. 
Although no species of Nardoa has previously been recorded from the Torres Strait 
region, on the reef flat at Mer the genus is well represented, 4 species occurring, not in great 
abundance, but 3 at least by no means rarely. They are found under large rock fragments 
or, less commonly, partly or wholly exposed on the sand. The most common species 13 
pauciforis, but novecaledonia is little more rare. These two species are so much alike in 
size, color, and form that the difference in size of the abactinal plates of disk and ray-bases, 
which is remarkably constant, suggested a sex difference. But Dr. E. N. Harvey very 
kindly made examination of the gonads of a number of living specimens and found there 
was no correlation between sex and size of plates. While it is possible that the two forms 
represent a single dimorphic species, that can only be determined by breeding experiments; 
it is therefore necessary for the present to designate them by different specific names. No 
individuals were found which could be considered intermediate between the two forms, 
which were invariably recognizable at a glance. The fine species, mollis, is much less 
common than pauciforis; it reaches a larger size and is always of much darker color and 
more elegant appearance. The fourth species occurring at Mer is the least common and 
the handsomest. It appears to be undescribed and because of its strikingly beautiful colora- 
tion in life I am calling it rosea. It was interesting to see that while these 4 species of 
Nardoa occurred indiscriminately on the reef flat, they do not seem to hybridize or inter- 
grade in any way. There is no difficulty, either in living or preserved material, in assigning 
each specimen at once to its proper species. 
The 10 species of Nardoa which seem to me valid may be distinguished from each 
other as follows: 
Key to the Species of Nardoa. 
A. No abactinal or marginal plates noticeably tuberculiform.t 
B. Superomarginal plates not alternately large and small. 
C. Adambulacral armature in 3 or 4 series of 3 or, in the furrow series, 4 spinelets each. 
D. Rays long, R=5-9 r; actinolateral series extending far beyond middle of ray. 
E. Abactinal plates on distal part of ray not markedly and rather abruptly different 
from those on basal part. 
F. Many large abactinal plates (3 to 5 mm. in diameter) on disk and rays....... variolata 
FF. No large abactinal plates, few exceeding 2 mm. across. pauciforis 
EE. Abactinal plates on distal part of ray markedly, and somewhat ‘abruptly, ‘smaller 
and more crowded than basally. 
F. Many large abactinal plates 2.5 mm. in diameter or more; ground-color fawn 
OMlight DIOWcsenice mete rece reticent Mite eh ner ere ee Ee Cer novecaledonie 
FF. Few or no large abactinal plates; usually dorsal plates are under 2 mm. in 
diameter; ground-color deep rich brown, in life.....................000-005 mollis 
DD. Rays short, R=4 r+; actinolateral series not extending beyond middle of ray...... obtusa 
CC. Adambulacral armature in 2 series of 4 spinelets each.................00e cece eeseeee galathee 
BB. Superomarginal plates more or less regularly alternating, large and small................-. rosea 
AA. Some, and often many, abactinal plates (and marginals, in frianti) swollen and more or less con- 
spicuously tuberculiform. 
B. Tubercles scattered, small (2 mm. or less in diameter), none in superomarginal series....... tuberculata 
BB. Tubereles large, 3 to 5 mm. in diameter. 
C. Tubercles numerous, occurring near tip of ray and on marginal plates as well as elsewhere Jrianti 
CC. Tubercles few, confined to abactinal surface of basal two-thirds of ray...............-- tumulosa 
1 Wisher (1919) suggests that the type of obtusa is a young tuberculata with the tubercles “a little more prominent 
than usual.” Perrier’s description implies an absence of tubercles, and I have therefore included obtusa in this section 
of the genus, but Fisher has seen the type and his views of its real affinities are no doubt correct. 
