104 AN INTRODUCTION 



rolla; and Helleborus and Nigella also, by their 

 tubulose nectaria ; in Pancratium the stamina are 

 inserted in the nectarium, which distinguishes it 

 from Narcissus ; in Hyoscyamus there is a covering 

 to the capsules, by which it is known from Physalis ; 

 the Reseda has always a lateral nectarium, but va- 

 nes in its corolla and pistillum ; the Campanula has 

 a quinquevalved nectarium, but is inconstant in the 

 corolla and capsule ; and lastly, the Iris has a stigma 

 of singular construction, but varies in the beard of 

 its corolla. 



There is, however, no one part of fructification 

 that can be relied on as a constant characteristic 

 mark for all genera ; it being found, that the part 

 which is -constant in some genera will be inconstant 

 in others : Thus in Carica the flowers of the male 

 plant are monopetalous, and those of the female 

 pentapetalous ; in Myrica some species have naked 

 seeds, others berries; in Fraxinus some have a 

 naked flower, and others a corolla; in Geranium 

 some have regular corollae, and others irregular ; in 

 Linum some are pentapetalous, others tetrapetalous ; 

 inAconitum some are tricapsular, and others quinque- 

 capsular ; and in Trifolium some are monopetalous, 

 others polypetalous, some monospermous, and others 

 polyspermous. 



This inconstancy of particular parts in many ge- 

 nera has been another source of error amongst the 

 earlier botanists, who have parted many plants from 

 their congeners on this account : of these mistakes 

 vie shall give an ample list in chap. 32. 



When the characteristic mark of any genus is 

 wanting in any particular species, we should proceed 

 with caution, least we confound genera that should 

 be distinguished : for want of this caution, the Erica 

 and Andromeda had been joined, but were parted 

 afterwards on account of the two horns in the an- 

 therae of the Erica; the Adonis had been joined to 

 the Ranunculus, but was parted from it again on ob 



3 



