GALIUM. 
389 
note “near G. eleyans of Wallicli”) supposed to have come 
from thence. This Can. pi. seems indeed at first sight very 
different in its grey or hoary aspect from the Mad. pi. above 
described ; but I have failed entirely to discover any permanent 
or constant distinction stronger than varietal between the 
two. It will therefore stand as G. ellipticum Willd., 
(3. villosum ; villous-hoary or canescent ; st. grey-lanuginous 
or densely villous ; 1. all over villous especially beneath, dis- 
tinctly petiolate, mostly acute; branches of panicle, ped., pedic., 
ov. and cor. outside hoary villous. — G. Neesianum Req., DC. 
iv. 600. u G. hirsutum Nees et Buch in Hort. Ber. p. 113. t. 22 
(non Ruiz et Pav.).” G. rotundifolium var. a villosum WB. ii. 
185 (not Linn.). G. ellipticum E. Bourgeau ! PI. Can. ex Itin. 
2 cl °, Teneriffa (Coss.) in H. K. et Herb. Mus. Paris. — Hab. in 
Canariis in sylvis latebrosis salebrosis. Necnon fortasse hue 
quoque spectat G. rotundifolium Webb Spicil. Gorg. 133 ; 
Selim. FI. Yerd. 209, planta microphylla valde incano-villosa 
ad alt. 5000-6000 ped. in dumetis Montis Gordo S u Nicolai 
Insularum Yiridensium obvia. Sed ob defectum infiorescentise 
fructusque in exemplaribus adhuc visis suspensus hsereo. 
The original G. rotundifolium Linn. Sp. PI. ed. 1. p. 108 is 
assuredly, both by his diagnosis and quotation of Morison t. 21. 
f. 4, simply and purely G. rotundifolium of DeCandolle, Koch, 
Grenier et Godron &c. ; mainly distinguished by its annual 
herbaceous erect simple st. not above 8 or 10 in. high, and few- 
fid. less developed cymes, and not found either in the Canaries 
or Mad. 
G. rotundifolium Linn. Sp. PI. ed. 2. pp. 156, 157, is a com- 
pound : formed primarily (p. 156) of a pi. not before contem- 
plated by him in ed. 1, for which he quotes Moris, t. 21. f. 5 
and Boccone, adding in MS. propria manu in his own copy of 
Sp. PI. ed. 2, to his imperfect printed reference to the last 
named author, u ic. 10. t. 11. f. 5,” — which is clearly an error 
for t. 6. f. 1, caused by his having mistaken the number 11 of 
the page close above the plate at the right hand corner for 
that of the plate itself *- which is in small inconspicuous 
print at its left hand upper corner, and then having counted 
the separate figures of the plate backwards from right to left. 
* Lamarck (Diet. i. 299), quoting Boccone, has made precisely the 
same mistake as to the number of this plate. 
