35 



Table VI. 



m 



The Platano-Proteal, or Labiatifloral Division. 



That the extensive range of families here regarded as forming one Division are a strictly natural assemblage will probably be admitted by whatever name it may be called ; and that they take 

 their origin from the Conifers through Platanacece and Proteacem, can scarcely be doubted, if it is admitted that there are no other families, either Gymnospermous or Cryptogamous, of which they 

 could be considered as branches. The approach between the Conifers and Proteacece is not very near except in habit, but Platanus agrees with Pinus in its male flowers, the apex of the anther 

 being broad and flattened, unlike those of any other families, and in the ovule having an inferior foramen and the seed containing albumen.* 



If any part of this Division could be separated, it would be the Geranial Subdivision, which is a very natural assemblage of families, the polypetalous forms being I believe, distinctly con- 

 nected with Nolanacea through Kitaibelia among Malvacem, which with several rows of carpels on the same plane has but one style, and Limnanthes may also be an approach to the same kind of 

 structure. But it seems not advisable to separate it, because Geraniacea and their allies scarcely differ from Rutacece and their allies. 



Forestiera agrees with the Oleacece, an affinity suggested by Mr. Bentham, in having suspended ovules with the raphe obviously lateral, occasionally inclining to dorsal, in which it differs from 

 other families with which it has been compared. 



Phytocrene^e. 



This family is very close to Aquifoliacece, agreeing with them in the number of ovules and the position of the raphe, and the porous stem of Phytocrene like that of Vitis, produces a large 

 quantity of watery fluid ; and these with other characters seem to indicate an approach to Anacardiacece and their allies. 



Salvadorace^e. 



The families with which Salvador a has been associated have compound ovaries, with the carpels united by their margins, but its ovary consists of a single carpel (there being in the unopened 

 flower a furrow on one side, relatively with which the raphe is lateral which shows that it is not accidental), which alone would make it very improbable that its affinities had been correctly 

 understood. I have no doubt but that M. Planchon is right in regarding it as allied to Oleacece, and he informs me that he believes it to be near Loganiacece, and this agrees with a suggestion o 



* I regard the scale of the female cone of Pinus as two bracts analogous to those of Welwitschia. 



