of the relationship, by its strong garlic odour, and its power of making 
the eyes water. Nectaroscordum siculum, Lindl., has only been discovered 
as yet in this one station among the Esterels, where the dense brush- 
wood and the distance from any roads make the search for specimens 
one of great difficulty, and also among the mountains of Sicily, and on 
Monte Gennargentu in Sardinia. I have, however, seen a plant in 
M. Alphonse de Candolle’s herbarium bearing the following label :—‘“ No. 
2870, Allium, in Mte. Nalkou (Assyrie), M. Aucher Eloy, 1836,” which 
is attributed to this species, and which, as far as one may judge from 
specimens in an imperfect condition, appears to be indeed Nectaroscordum 
siculum. If this plant from Kourdistan* be really our species, we may 
perhaps have here an indication that this plant isa relic of the times when 
the east was more completely one in its vegetation with the western 
portion of the Mediterranean than it now is, and when the Cedars of 
Atlas, now distant 1400 miles from their relations, spread in more or less 
continuous forests to join the Cedars of Lebanon and the Taurus range, 
while these again were united with the Deodar Cedars of Affghanistan 
and the Himalaya, bridging over another space, now blank, of 1400 
miles!t Besides the partially inferior ovary, the dissimilar divisions of 
the perianth and the sheathing leaf or bract from which the flower scape 
of this plant emerges, another interesting feature in its structure, and one 
which appears to have hitherto escaped observation, is the difference which 
exists between the central and outer flowers of well-furnished umbels, in 
respect of the number of their parts. I find that in all stout and well grown 
plants of Mectaroscordum, the two or three central flowers have 8 divisions 
of the perianth, 8 stamens, and an 8-celled ovary (becoming 4-celled 
in the ripe capsule by abortion), while the outer flowers have their parts 
in multiples of 3, there being 6 divisions of the perianth, 6 stamens, and 
a 3-celled ovary. This difference between the innermost and the outer 
flowers might be supposed to be due to an excess of nourishment which 
is supplied to those flowers, which most nearly spring from the summit 
of the central axis of the plant; but this alone will not serve to explain 
the assumption of a fixed number of additional parts. For, if this were 
the case, one might expect to find the numbers of the parts proportioned 
to the amount of nutriment received, so that in large vigorous umbels 
the maximum would be reached by the central flowers, while the adjacent 
flowers would have their parts in sevens, eights, or sixes; but this 
apparently never takes place, the flowers being either trimerous or 
tetramerous in their plan. Among Monocotyledons the trimerous arrange- 
ment is almost universal, but among the few exceptions we find Smilacina 
bifolia, Roem. et Schult., and Paris quadrifolia, L., which have the parts 
of their flower either on a dimerous or trimerous plan. I do not know of 
any Monocotyledonous plant which has throughout its floral organs the 
* Mte. Nalkou is near and north of Kermandshah on the western borders of Persia. 
+ For an account of these Cedars, their affinities and distribution, see Dr. Hooker 
in Natural History Review, Il. p. 11. (1862). 
