APPEXDIX. 45 



\ 



entertained by M. Dunal, in common with many botanists, to 

 diminish as much as possible the nmnber of natural orders, a 

 veiy proper and meritorious caution, but when carried to excess, 

 as in this instance_, is productive of mischief; for by \uuting 

 several families into one, which are composed of very opposite 

 and dissimilar characters, we destroy the very object we attempt 

 to establish, viz. to mark the limits of distinction between differ- 

 ent groups of plants. The selection of a few decided and con- 

 stant characters, that can serve to distinguish each order, tribe or 

 section, must infallibly tend to the greatest simplicity of arrange- 

 ment ; and if in accomplishing this purpose, we should thus be 

 led to increase the number of famihes, in order to ensure the 

 means of certain discrimination, it is indubitably better to do so, 

 rather than, by pursuing the opposite extreme, to render all fixed 

 landmarks useless. It was upon this conviction that I proposed 

 {huj. op. i. 163) to reduce the Scrophulariace(e within more cer- 

 tain limits than Mr. Bentham had employed in his admirable 

 monograph of the order in the 10th volume of the ' Prodromus ' 

 of M. DeCandolle, and also to confine the Solanacece within 

 strictly definable bounds. The difficulty of establishing an ob- 

 vious line of demarcation between these two great families, w^as 

 there discussed at some length, when I showed how unsuccessful 

 had been the attempts of botanists to remedy so manifest a de- 

 fect in the system. Mr, Bentham, it is true, adopted with this 

 view, the plan of associating the few aberrant cases then known, 

 in a distinct tribe, his Salpiglossideie : the heterogeneous features 

 of that tribe have been fully demonstrated, proving that the at- 

 tempted remedy has been wholly inefficacious : among the many 

 instances that could be cited, it is only necessary to point out, 

 how impossible it is to retain Salpiglossisy Anthocercis, Sckwenkia 

 and others in Scrophulariacete^ while Petunia, Nierembergia, and 

 numerous others are placed in Solanacece. At the time of Mr. 

 Bentham^s monograph the exceptional genera were few, but 

 since that period they have become so multiplied as to equal in 

 number those belonging to true Solanacece. 



As a desirable test towards the attainment of this great desi- 

 deratum, I suggested the constant character of the aestivation of 

 the corolla, which, combined with other well-selected features, 

 will be found to reduce these two extensive orders within de- 

 finable bounds : for this purpose, it is only requisite to detach 

 from each their several aberrant cases, and comprise these in an 

 intermediate family, where they are easily separable into tribes, 

 distinguishable by marked peculiarities. M. Dunal duca not 

 seem to have been aware of this suggestion, or at least, no such 

 expedient appears to have entered into his contemplation ; and his 

 ordinal diagnosis of the Solanac€(B, aggravated still further by 



