226 Dr Daubeny on the Writings and 



DeeandoUe, in his Theorie Elementaire, and other writings, 

 has followed u^j this notion in his usual masterly manner ; but 

 aware that a complete map of this kind can only be con- 

 structed when the whole vegetable kingdom is fully investi- 

 gated, just as it would be impossible to make a general map 

 of the globe, till every part of it had been explored, he has 

 contented himself with the preliminary labour of constructing 

 local maps, as it Avere, of different countries, by pointing out 

 the relations of each of the families which he notices to the rest. 



In some cases, indeed, the characters of a family are so 

 well marked, and so distinct from those of most others, that 

 it is difficult to say to what the plants included in it are most 

 allied. They stand in the relation of islands in the midst of 

 a vast ocean, " penitus toto divisos orbe," and in any map of 

 the vegetable kingdom which may hereafter be constructed 

 would be represented as detached groups. 



Such is the case with the Cruciferje, a family which, al- 

 though it bears some analogies to the Papaveraceae in the num- 

 ber of its petals, and in the structure of its fruit, possesses 

 nevertheless so many peculiarities of structure, that it seems 

 to lie apart from all other natural groups, except, perhaps, 

 from the small one of the Capparideas. 



But in the greater number of instances, the characters of 

 the natural groups graduate into those of several others, like 

 the countries included within the same continent ; and in 

 these cases, Decandolle has presented us with a sort of pic- 

 torial view of the entire family, representing the sevei'al 

 genera of which it consists diverging from a common centre, 

 around the circumference of which are placed the families 

 with which these genera bear respectively the nearest con- 

 nection. 



Thus in the case of the Crassulacea?, the division into tribes 

 is drawn from the relation of the stamens, in point of num- 

 ber, to the petals, they being in some cases the same, in 

 others double of the latter ; and these two tribes are again 

 subdivided, according as the petals are united or distinct. 



