20 ON THE PROTEACE.E OF JUSS1EU.
cases, but fails in Persoonia, the species of which differ in having one or two seeds : it would seem however, in this case, that an irregularity in a point of such importance could not take place unaccompanied with other anomalies in the same organ, and accordingly such are found to exist in this genus, and will be mentioned when treating of the fruit.
Besides number, the insertion of the ovula is also to be attended to ; for though this may generally be presumed from the situation of the radicula in the ripe seed, yet to this criterion there are several exceptions, even in the present order : thus, while the radicula constantly points downward in the whole of the order, the insertion of the ovulum is in many cases at the top or side of the cell of the ovarium. My observations on this subject are as yet incomplete ; but, from those that I have made, I am in- clined to think such differences will be connected with genera, or rather perhaps with particular kinds of fruit. Thus I conjecture, in Leucospernvnn, Mimetes, Nivenia, and Spatalla, the insertion to be uniformly lateral.
The style, though not subject to much variety in this family, will be found in a few cases to furnish generic cha- racters. Thus in Protect, strictly so called, the persistent subulate style forms an important part of its character: and the persistency of the whole of the style in the greater number of species of Grevittea will probably be used by 33] future botanists in distinguishing them from that re- markable section of the genus, which I have at present united with them and called Cyclojpterce. Its length also, when compared with that of the calyx, seems in some cases to be of importance, as in distinguishing AdenantJios from Spatalla ; but in general this circumstance can hardly be had recourse to except in specific characters.
The form of the stigma is in many cases of considerable importance in characterising genera, a fact which could not escape the penetration of Dr. Smith when establishing his new genera of this order : thus its conical papilla in his Concilium (the Hakea of Schrader) will in many, though certainly not in all cases, distinguish it from Grevillea ; but
�� �
