296 A COMMENTAllY ON THE SECOND BOOK 
cies, is as follows : " Ramuli novi floribus sessilibus sparsis 
tecti, apice foliolis novellis comosi, post florescentiam in 
i'amum foliosum basi fructiferum excrescentes," as may be 
seen in other three parts of the figure, even in Rumphius 
(t. 16). This structure did not escape Garcin, who in his 
last paragraph describes it excellently. 
M. Desrousseaux {Enc. Meth. iv. 17) for his Melaleuca 
Leucodendra ]d\x\s together the Arbor alba (H. A. ii. 1. 16) 
Jhliis mudtmermis^ the Cajwputi Jbliis trinerviis (Ibid. 1. 17, 
f. 1), and the specimens from Java, of which M. Lamarck 
has published an engraving (111. Gen. t. 641, f. 4), to which 
we shall again return. These are three distinct species ; 
and he quotes the M. viridiflora of G^ertner with doubt. 
Mr Robert Brown (Hort. Kew. iv. 410), although he 
quotes only the 16th plate, yet, in the specific character, 
he, in his Melaleuca Leucodendro7i, includes not only the 
kind with five large nerves, but also that with only three 
nerves in its leaves, which is that represented in Plate 17, 
%• 1- 
M. Lamarck, having procured from Sonnerat, a speci- 
men of the Arbor alba from Java, published a figure of it 
for the Melaleuca Leucodendra (111. Gen. t. 641, f. 4). 
It differs from the plant of Sir J. E. Smith, in having its 
younger parts thickly covered with erect hairs, I found 
this plant in the botanical garden at Calcutta, to which it 
had been sent by Dr John Fleming, as the plant really 
producing the cajeput oil ; and such may be the case, al- 
though, according to Rumphius, an oil is produced by the 
species with three nerves; while this has five principal 
nerves, besides several veins disposed longitudinally be- 
tween them. These veins, indeed, are not continued the 
whole length of the leaf, but form occasional anastomoses. 
Still, however, they give the leaf the appearance of having 
nine or ten longitudinal nerves, as represented by M. La- 
