MR. J. MIERS ON THE HIPPOCRATEACEA OF SOUTH AMERICA. 327 
seeds are borne upon a large membranaceous wing, or they are nuciform upon a coria- 
ceous stipitate support; while in the drupiform fruits the exalbuminous exalate seeds 
are often extremely large, with two, sometimes three, thick cotyledons, with a minute 
radicle,—conditions quite at variance with the structure of the seeds in Celastracee. 
8. The seeds in all the genera of the true Celastracee, which have mostly capsular 
fruits, are partially or wholly covered by a conspicuous lax arillus. From this category 
I exclude all the genera of the Eleodendree, which are marked by exclusive characters. 
No such arillus occurs in Hippocrateacee—although, in some of its drupaceous genera 
where they have been examined in a fresh state, the seeds are covered by an adherent 
entire fleshy coating, an arilline, to which 1 have before alluded (supra, p. 324), which 
is of a nature and origin quite different from that of a true arillus. 
9. Another circumstance, though of less importance, and to which I have previously 
alluded, still adds to the sum of differences between the two families—the ordinary per- 
sistence of the petals and stamens upon the base of the ripe fruit, which I believe never 
occurs in Celastraceæ. This persistence may perhaps be due to the presence of the 
vessels mentioned in the next paragraph. 
10. There is yet another difference of structure in the two families which deserves 
especial notice. In the Hippocrateacee there is a general (and, I suspect, universal) 
presence of numerous spiral vessels in the leaves, the ramifications of the inflorescence, 
and in the parts of the flower, which in many cases are equally abundant in the pericarp 
of the fruit, in the integument of the seeds, and sometimes even in the cotyledons of the 
embryo; but I am not aware that they are ever seen in like manner in the Celastracee. 
If, for instance, we break off the pedicel of a flower of Hippocratea, or a sepal, or petal, 
and pull it away carefully, these will be found suspended by many fine capillary threads, 
which have been drawn out; orif we make a longitudinal section through a flower, in 
many cases similar threads are drawn out, adhering to the knife. "This peculiarity alone 
is sufficient to indicate a considerable organic difference in structure between the Hip- 
pocrateaceæ and Celastracee. 
authors, makes the slightest mention of albumen one way or the other. Subsequent authors have blindly followed 
Endlicher’s character. Harvey, in his * Genera,’ p. 59, is silent on the subject. Ecklon & Zeyher (Enum. 980), in 
describing the frnit, do not allude to the absence of albumen. Sonder € Harvey (Fl. Cap. i. 464) also describe 
Hartogia capensis in full, preceded by a generic character translated from that of Endlicher, which is quoted as 
their authority. They do not appear to have had any evidence of their own concerning the point in question ; for in 
their copious description they make no mention of albumen. In counterpoise to this negative evidence, we have the 
more positive testimony of Thunberg. Hartogia was first proposed in 1781 by the younger Linnæus (Syst. Suppl. 
128). Hartogia capensis was described and figured in 1784 by Thunberg (Nov. Gen. v. 86), and repeated verbatim 
in Diss. v. 35; but no mention is there made of albumen in the seeds. But in his Flor. Cap, 28 (1794) he gives a 
drawing, with analytical figures, of Schrebera schinoides, Linn. Syst. 265, where, in plate ii. fig. h, he gives a 
transverse section of a seed, in which the embryo is shown distinctly enveloped in copious albumen. This was 
drawn at the time when he thought Schrebera was generically distinct from Hartogia, the seed of which he does 
not appear to have examined. This fact appears to me decisive on the question. 
In regard to Kokoona, I have evidence to offer, proving it to be a true Hippocrateaceous genus. Consequently this 
particular rule of distinction between the two families, as above defined, appears to be universal, as far as cur present 
knowledge extends. 
2x2 
