* BOTANY OF THE LACCADIVES. 61 
figure from a tree in Amboina, there are before the writer Specimens of un- 
doubtedly this species from Java (Zollinger n, 2861, which, however, Zollinger 
hinself has identified with H. sonora) and specimens recently collected by the 
officers of the “ Eyeria” in Christmas Island, where H. ovigera occurs 
(Hemsl., Journ. Linn, Soc., xxv, 357) on the summit, elevation about 1,290 feet, 
a rather remarkable fact, since, according to Rumf, it occurs, like the other 
Hernandias, “ semper in arenoso solo circa litora.”* 
As has been remarked, the belief of Linnaeus, and of Roxburgh that Rumf’s 
Arbor regis is a Hernandia is probably partly justified, for it is possible, from his 
account of the habitat of his tree—“ occurrit tam in litore inter leves ac humiles 
“stlvas ”’ (quite the situation affected by Hernandia peltata) “ quam in montibus 
“a altioribus silvis”’ (where to find H. peltata would be somewhat surprising) — 
that Rumf has included two trees in his description. His figure, moreover, bears 
out this, for some of the leaves are without, while others exhibit, a pair of glands 
where the petiole joins the leaf. The figure as a whole, however, suggests at once, 
as Lamarck (Hneye, Meth., iii, 128) a century ago pointed out, a Euphorbiaceous 
plant, while Rumf’s description of the fruit is altogether suitable to that of a 
Species of this order. The first authors to recognise Rumf’s Arbor regis, 
however, were Teysmann and Binnendyk, who described it as Capellenia moluccana 
(Nat. Tyds, Ned, Ind., xxix, 239), founding a new genus to accommodate it ; as, 
however, Capellenia does not differ generically from Endospermum, the tree has 
been re-described by Beccari as Endospermum moluccanum (Malesia, ii, 38) in his 
treatise Peante Ospitatrici, where another species from New Guinea (Endospermum 
Sormicarum Bece., Malesia, ii, 44, t. 2) is described, which shares with Rumf’s 
tree the character of sheltering a species of ant in its hollowed stems and branches, 
Teysmann and Binnendyk described their species from trees grown in the Botanic 
Garden at Buitenzorg; Beccari does not mention the habitat of the N ew 
Guinea species; in the Calcutta Herbarium there are, however, examples 
of another closely related species, with the same hollow branches, collected. in 
Sumatra by H. O. Forbes, which were obtained on the volcano of Kaba at 3,500 
feet elevation, This fact, therefore, does not oppose, if it does not corroborate, the 
surmise that Rumf under Arbor regis has included. two trees, one found only on, 
SER es ee I LY NCE) ere 1 ened 
* Two parallel instances known to the writer of littoral species ascending to a con- 
siderable height are met with in N arcondam, where Morinda bracteata ascends to 2,300 
feet and in Barren Island, where Zerminalia Catappa ascends 1,100 feet, The expla- 
nation of all three cases is doubtless the same ; these “ littoral” species being amongst the 
first to appear on the respective islands were able to spread unchecked from the shore to 
the summit of their peaks, and the invasion of inland species has not subsequently 
been sufficiently great to compel them to retire completely from the unusual localities 
they had at first invaded, 
