102 
dense croceo-velutinis, lana (% unc. long.) crocea. — 
G. religiosum, Roxb. FI. Ind. vol. iii, p. 185, non Linn. 
G. parvifolium, Nutt. Herb. — Viti Levu, on the Raki- 
raki coast (Smythe!), Kadavu (Pritchard! Seemann! 
n. 28). Also collected in Oahu, Atoi, Hawai (Diell! 
Nuttall!), Maui, Sandwich Islands (D. Nelson! Men- 
zies!). 
This is the plant which A. Gray (Bot. Wilkes, p. 
179) calls G. religiosum, but it is not that of Cavanil- 
les, which is more glabrous, has a deeper-cleft calyx, 
white flowers, and the seeds quite glabrous, after the 
removal of the wool. 3 The Sandwich Islands plant is 
apparently identical with that described by Roxburgh 
1. c. under the name G. religiosum, which, he says, 
has "seeds free, clothed with firmly-adhering, short, 
tawny down, and long wool of the same colour." 
There is a specimen of "Yellow Cotton” from Joy- 
negau (Trove!) at the British Museum, which has 
very small leaves, — the smallest I have seen in this 
genus, — agreeing as far as it goes with the above 
species; and there is a starved specimen of G. tomen- 
tosum from Hawai (Diell!) which has the leaves 
almost as small, and which Nuttall had provisionally 
named G. parvifolium. But generally the leaves and 
flowers of G. tomentosum are those of the size usual 
in this genus. 4 
A considerable number of the features men- 
tioned in the quoted description are definitely 
not diagnostic for they are to be found in many 
species of Gossypium. More important, many of 
these described features are to be found in both 
the Hawaiian endemic and the Fijian cotton 
which Seemann, Pritchard, and Smythe all col- 
lected 100 years ago. Certain of the features 
described in the original publication, however, 
clearly exclude either the Hawaiian endemic or 
the Fijian introduction and these features will 
be discussed briefly below. 
The stipules of the Hawaiian species are de- 
scribed as "minute” by Watt ( 1907 ) and "subu- 
late” by Degener (1933). The stipules of the 
variant of G. hirsutum to which Seemann’s No. 
28 belongs are described by Watt as "broad, 
oblique, ovate lanceolate . . . subcordate.” Ob- 
viously Seemann’s description of the stipules of 
G. tomentosum in his original description were 
not from the Hawaiian plant to which the name 
has been applied but from the Fijian cotton. 
3 This footnote quoted Solander’s manuscript de- 
scription of the cotton found in the Society Islands, 
which is apparently what has been called G. taitense 
Pari. — R.L.W. 
4 This footnote provided the original description of 
Gossypium drynarioides Seem. = Kokia drynarioides 
(Seem.) Lewton. — R.L.W. 
PACIFIC SCIENCE, Vol. XVIII, January 1964 
The involucral bracts of the Hawaiian species 
possess small triangular teeth, each at most two 
or three times longer than their width and not 
deeply laciniate as they are in the cotton from 
Fiji collected by Seemann. Again, it is obvious 
from Seemann’s account that this described fea- 
ture of G. tomentosum was taken from the in- 
troduced Fijian plant. 
Seemann’s key to the species of Gossypium 
found in Fiji states that the seeds of his G. 
tomentosum retain tawny "moss” after the re- 
moval of the wool. All authors are agreed that 
the seed-hairs of the Hawaiian species are not 
separable into "fuzz” and "floss.” Seemann’s 
description states that the trichomes of G. to- 
mentosum are about 5 A inch long (= c. 1.9 cm), 
while those of the Hawaiian endemic are stated 
seldom to exceed 1 cm in length. 
In contrast to all of the above described 
features, which could never have been observed 
on the Hawaiian plant, the subtruncate calyx 
could scarcely have been observed except on 
the Hawaiian specimens, inasmuch as the Fijian 
plants would be expected to possess the pro- 
nounced calycine teeth characteristic of most of 
the hirsutum cottons. 
Watt (1907:69-71) first pointed out that 
Seemann was in error in associating specimens 
from Fiji with collections made by Nuttall in 
Hawaii. Watt then in effect proposed to typify 
the name G . tomentosum by the Hawaiian ele- 
ment included in the original description and in 
this interpretation he has been followed by all 
subsequent authors. However, it is readily ap- 
parent that the greater portion of Seemann’s 
original diagnosis can apply only to the Fijian 
cotton and clearly excludes the Hawaiian en- 
demic. The typification of such taxa which, when 
originally published, contained two or more 
elements has often proven itself to be a most 
difficult problem. In an effort to provide broad 
principles for the solution of such problems a 
guide for the determination of types has been 
added to the International Code (Lanjouw, 1961: 
65) and includes the following instruction as 
to proper procedure: 
d. In choosing a lectotype, any indication of intent 
by the author of a name should be given preference 
unless such indication is contrary to the protologue. . . . 
e. In cases when two or more elements were in- 
cluded in or cited with the original description, the 
