42 APPENDIX. 



it is known^ that in his anxious endeavours to bring it to a close^ 

 he has laboured beyond his physical powers, persevering in this 

 task under the pressure of long and continued illness. Still it 

 is to be regretted, that circumstances should have operated to 

 keep its distinguished author in ignoi-ance of the facts and rea- 

 sonings that have been published within the last few years. If, 

 therefore, I now proceed to point out several inconsistencies in 

 the distribution exhibited in this volume, I can truly affirm, that 

 it is not from any desire to criticize the labours of M. Dunal, but 

 to justify what is due to the advancement of science, and to sup- 

 port the inferences drawn and the facts collected by me towards 

 the history of this family. 



The gi-eat object of all scientific arrangement is to group 

 together individuals possessing some common conspicuous fea- 

 tures, by which they may be i^eadily distinguished : these may 

 again be subdivided by other partial characters into sections and 

 subsections, but all such characters should be clearly definable. 



The ordinal character of the Solanace(By as given by M. Dunal, 

 like that of his predecessors, falls very far short of this deside- 

 ratum, and the Conspectus of the classification is deficient of 

 those tangible features that should serve the purpose of discri- 

 mination. Its limits are by far too general, embracing within 

 its scope individuals belonging to other orders. In the cha- 

 racter of the corolla, for instance, the more important features 

 are neglected or merely hinted at, while others more especially 

 selected are frequent among Scrophulariacecs and other orders : 

 hence they are of little value for distinguishing the precise family 

 to which species belong. In the characters given of the struc- 

 ture of the seed and embryo, many peculiarities are altogether 

 omitted, while others are inaccurately described : these will be 

 more particularly noticed presently. In his Conspectus, M. Dunal 

 divides the family into two tribes, the Nolane(S and the Solanei^, 

 the latter being separated into nine subtribes, which are marked 



very 



Of 



these subtribes the most numerous in genera is the SolanecB, 



JTular 



indication of little utility, as it exists equally in other tribes ; 

 2. a bilocular berry; 3. a semicircular or spiral embryo, a feature 

 also retained by other subtribes, and even here too its applica- 

 bility as a test is rendered nugatory by the knowledge that 



Marckea have a nearly straight embry 



Nolanacece 



in the structure of the pistil and the fruit. The Sa^ophulariacece, 

 AtropacecB and SolanacecB, placed as I have suggested under more 

 strictly defined and simple limits, form, together with some "other 



