280 OBSERVATIONS ON THE NATURAL FAMILY
In conformity with this view of its composition, and with the relation above attempted to be established, the deve- lopment of the pistillnm precedes that of the stamina in many species of the genus.
It is more difficult to determine whether this order of expansion and relative position of sexes in Euphorbia be in conformity with the general rule, or an exception to it. For its fasciculus of flowers may be considered as analogous either to the simple spike, and consequently having an inverted order of expansion, as in Allium descendens, and certain species of Grevillea and Anadenia : or it may be assimilated to the compound spike, as in several species of the genus the male flowers appear to be separated into ioo] fasciculi ; and according to this view the order of expan- sion is direct, the central female flower being the repre- sentative of the terminal partial spike.
There is even a third species of inflorescence with which the fasciculus of Euphorbia may be compared, namely, that consisting of one or more verticilli with a single flower in the centre. In this, which may be considered a modi- fication of the spike or umbel, the usual order of expansion seems to be from centre to circumference. Its simplest form occurs in an unpublished New Holland genus of the same natural family with Euphorbia, in which a single verticillus of male flowers surrounds the central female flower. Lambertia may be considered as another instance of the same mode, and as far as can be determined, in a case where the flowers are hermaphrodite and their ex- pansion nearly synchronous, following the same order. In all the known species of this genus the leaves are verticil- late, and uniformly in threes : in L. formosa and inermis the involucrum constantly contains seven flowers, while in L. uni flora it is reduced to one flower. The seven flowers of the two former species I consider as made up of tw^o verticilli, in number of fiWers corresponding with that of the leaves, and of a single central or terminal flower ; to which terminal flower L. uniflora appears to be reduced. From this order of reduction it may be assumed as more probable that species of Lambertia should be found with
�� �
