107 



tlie family should be broken up and its genera should be referred respectively 

 to the Apocvnacese, Gentianacese, Solanacese Rubiacese, and Asclepiadacese. 

 The limits of the families just referred to and their members to some extent 

 vary, according as we hold this view or that. 



Now, take examples of another kind. The Labiatge i;) and Verbenaceae 2) 

 are, according to one way of looking at them, to be regarded as two dis- 

 tinct families ; but in another way - - which is that observed in the 

 Borraginacete s) , which is established regardless of the position of the styles 

 (whether the latter are terminal or gynobasic) the former two families (i.e. 

 Labiatse and Verbenaceae) should be united. Should the former view be held 

 in the case of the Borraginacese, the latter family should be divided into two. 

 But, such a division, in fact, could never be cansidered natural. The same is 

 true of the separation of the Labiatse from the Verbenacese. The two families 

 are only artificially or superficially separated, but in reality, they are closely 

 inter-related like the meshes of a net. This shows plainly that it is impossible 

 to classify plants according to one sole view so naturally that the classification 

 should denote their natural relations ; that the latter relations are not to be 

 understood in a static sense, but are only conceivable in dynamic senses ; and 

 finally, that natural groups are only thinkable in a dynamic sense. 



Now let me give other examples showing how the natural arrangement or 

 natural system of such natural dynamic groups should also be a dynamic one. 



Julianiaceae^ (Julianiales) : This group is closely related to the Anacardiacese 

 and Juglandacea3 (the latter two families are widely separated from each 

 other, so far as the existing system is concerned) in its resiniferous character, 

 in its unisexual flowers with reduced envelopes, and in its solitary exalbuminate 

 seeds. Other points of resemblance or similarity between the Julianiacese and the 

 Juglandaceaa are the dissimilar male and female flowers, the broad stigmatic lobes 

 of the styles and single coated ovules. In anatomical characters, the Juliauiacese 

 and Anacardiacese are very much alike, and the singular funicular development of 

 the ovules, the seeds and embryo, are very much the same in both families. 



1) Nat. Pfl.-fam. IV. 3, a, p. 205. 2) Nat. Pfl.-fam. IV.- 3, a, 143. 



3) Nat. Pfl.-fam. IV. 3, a, p. 80. 4) HEMSLEY, W. B . On the Julianiacese : A New 



Natural Order of Plants, in Phil. Trans. Koy. Soc. Loncl. Series B, CXCIX. pp. 169-197, Plates 18-24. 



