ON MELIANTHUS TRIMENIANUS. 3900 
into the order Melianthea. This he separated definitely from Sapin- 
dacee by the albuminousseed, and narrow straight embryo : at the same 
ime he indicated with great sagacity the many salient points of 
affinity in every organ between the two orders. 
In the ‘‘Genera Plantarum” I regarded Melianthee asa tribe of 
Sapindacee, after a re-examination of the whole of the latter order, in 
which I found so many deviations from the normal character of the 
seed and disk attributed to it by Planchon and others, that it appeared 
to me to be more natural to unite’ Melianthus with it; except indee 
would have more than half-a-dozen genera. In this case there would 
result, 1. Sapindacee (from which some would separate Hippo- 
disk or in sinuosities of its marg seeds are exalbuminous ; 
whilst in Staphylee the stamens are inserted in the outer base of the 
disk, and the seeds are a inous as in DMelianthee. Uniting all, I 
e latte ‘ 
the position I would still assign to it. Of the above characters, that 
of a straight or curved embryo, though so prevalent in Sapindacee, 
isa very variable one in degree. In both Alvaradoa and Aitonia, 
t t 1, a capsule 4-lobed at the top and dehiscing by an 
internal suture, and connate intra-petiolar stipules; an lerisma, 
w not conspicuously gibbous, no ru nt of a fifth petal, 
capsule is 4-lobed though not acutely at the top ; in comosus and 
Trimenianus, and it opens ventrally as well as dorsally in D. minor 
and perhaps others; whilst the stipules, though not ate, are 
enus. 
atalia, Hochét., the third genus included by Planchon Pad 
: a foe tk 
ae t 
Fresen., as the type of a separate tribe, Bersamec, distinguished by 
flower not being resupinate, by the more regular calyx, five 
equidistant petals, as:many equal stamens, a very large oe os 
A 
