84 
CONTRIBUTIONS TO BOTANY. 
a suspicion that the seed so examined may have been a malfor-^ 
mation, but there appeared in this case no indication of any ab- 
normal deformity : how then can we account for the existence of 
an embryo so different in size and form from that figured by 
Dr. Wight ? Is it possible that this distinguished botanist, or 
more probably his draughtsman, can have mistaken the radicle 
for the entire embryo ? This will appear probable when we care- 
fully examine fig. 10 of his plate 954 above refeiTed to, which 
gives a transverse section of the seed, where exactly the same 
lunated space is shown across the middle, which I found to exist, 
lined with the two thin membranes above described, and which 
I conceive can be nothing hut cotyledons ; and again, if we com- 
pare this with fig. 11 of the same plate, which is a longitudinal 
section of the same, we perceive a line or long space descending 
from the supposed embryo at the summit to the base of the 
albumen, a fact which precisely accords with what I have de- 
scribed in the preceding analysis. Upon such grounds I am 
inclined to believe, that what I have detailed above is the real 
structure of the seed in Stemonurus. In anatropal suspended 
seeds with a superior radicle, it is usual to observe the raphe ter- 
minate at the opposite extremity to the point of suspension ; but 
in this case we find an exception to this general rule, which seems 
opposed to the established theory ; here the dii’ection of the 
raphe would seem to indicate a double retroversion of the ovule : 
so singular a fact may be of more frequent occurrence, but I 
confess that I have never met with, nor seen the record of, any 
such development. On the other hand, again, we have an ana- 
lysis given by Blume of the seed of Stemonurus secundijlorus in 
his ‘ Mus. Lugd. Bat.,’ in which the embryo is small in the sum- 
mit of the albumen, as represented by Dr. Wight, the radicle 
being terete and the cotyledons exceedingly small. 
The flowers of Stemonurus are sometimes 4-, often 5-merous, • 
but I am not aware whether this can be depended on as a good 
specific character ; all I can affirm is, that in those specimens I 
have seen, where 4-merous flowers prevail, I have occasionally 
met with some that are 5- or even 6-merous. Generally, the 
inflorescence is so short, as often to appear like a cluster of ax- 
illary fasciculated flowers ; in other species it consists of long 
branching panicles, in which the flowers are sometimes secundly 
disposed. The flowers are always glabrous, and each articulated 
upon its separate pedicel, which is often pubescent. I have seen 
but few of the species on record, and those mostly imperfect spe- 
cimens. In the following enumeration the characters are therefore 
given as described by their several authors ; they require doubt- 
lessly a more careful revision, for as they generally resemble 
each other so much in the appearance of the leaves, the shape of 
