292 МВ. ©. Т, DUNN: A REVISION 
to their base, or free, and in one case enclosed by the revolute margins of the 
flat filament. They apparently arise as papillæ at the base of the developing 
stamens. When fully developed, they form shortly stalked membranous cups 
with open spathe-like acute or truncate entire or serrulate hoods, or else, the 
spathe being nearly closed, they become oblique elub-shaped receptacles with 
longitudinal openings; in all cases, the insertion is slightly behind or outside 
of the filaments and the opening is always extrorse. When rudimentary, the 
nectaries form small hemispherical or club-shaped exerescences or stalked 
globular appendages at the base of the stamens. 
Twelve of the species now known inhabit a well-defined geographical area, 
including N.E. India, Burma, the Malay Peninsula and Archipelago, the 
Philippine Islands, Formosa, and tropical China. The remaining species is 
widely distributed in tropical Africa. 
The systematie position of the genus has been the subject of much 
discussion from the time of its discovery, and succeeding botanists have 
placed it in association with widely different groups according as they were 
more impressed by one or other of its characteristics. Blume located it near 
Lauraeez, in consequence of the valvular dehiscence of its stamens and its 
solitary pendulous ovule, and in spite of its inferior ovary, which gave it, he 
said, the same relation to that order as Vaccinium has to Ericaceæ (Ann. Sci. 
Nat. 2 ser. ii. (1834) 95). Не recognized also its relation to Gyrocurpus, 
and placed them together under the name of ППрегег». Endlicher (Gen. 
(1836) 2069) associated the same two genera under the name of Gyrocarpee 
and placed them near Daphnoidew. The close relation of Hernandia with 
llligera was not seen till thirty years later, but in order to understand the 
systematie history of the latter it will be necessary to trace also that of 
the former. Endlicher (l. с.) placed Zernandia in Santalaceæ. Nees (Syst. 
Laur. (1836) 703), and originally Lindley (Key to Nat. Syn. (1836) 202), 
followed Blume. Later, in his Vegetable Kingdom (1846, p. 718), Lindley 
referred Gyrocarpez to Combretacere, Hernandia to Thymelieaceze. Meissner 
(DC. Prodr. ху. i. (1864) 251) restored Gyrocarpeæ to Lauraeeze, and made 
Hernandia the type of a natural order Hernandiacew in the same affinity. 
Bentham and Hooker (Gen. Pl. i. (1865) 689), following Lindley, placed 
Gyrocarpez again in Combretaceze, adding to them Martius's Sparattanthelium, 
while Hernandia was referred to Lauraceæ. Meanwhile Baillon had pointed 
out (Adans. v. (1865) 186) that Hernandia might be regarded as a diclinous 
reduced type of ИНдета. Ш, indeed, a hermaphrodite flower be theoretically 
constructed from the male and female flowers of //ernandia, the peculiar 
floral characters of ligera are almost exactly reproduced. In 1885 
Solereder (in Bot. Centralbl. xiii. 163) showed that the Gyrocarpese differed 
trom the Combretacere and agreed with the Lauracez in important anatomical 
characters, namely, the absence of internal phloem and the presence of 
secreting cells. Four years later (in Engl. Bot. Jahrb. x. (1889) 511), he 
