228 Dr Daubeny 07i the Writings and 



see on the margins the genera Dasystemon, Bulliarda, and 

 Tilloea, which deviate somewhat from the standard. 



Amongst the Diplostemones gamopetalae, are placed on the 

 borders of the circle the genera Kalanchoe and Bryophyllum, 

 which by their pinnated leaves deviate from the rule ; amongst 

 the Diplostemones polypetalag we place the genus Penthorum 

 and Diamorpha half way beyond the circle, to shew that it 

 may be doubted whether they belong to the family at all. 



The genera that more immediately surround the centre, are 

 disposed in the order of their natural affinities, which it would 

 have l:)een impossible to have done if a linear arrangement 

 had been adopted. 



But with reference to the point more immediately under 

 consideration, it may be remarked, that this expedient of 

 Decandolle's enables him to shew to what other natural 

 families the Crassulaccae are most allied, and which of the 

 genera included in it constitute the connecting links. Thus 

 they approach to the Saxifrageas by means of the genera Dia- 

 morpha and Penthorum, which are not succulent, and have 

 the ovaries concrete, and with Paronychias through Tillaea, 

 which has but two ovules, and agrees in habit with Illecebrum. 



Decandolle has exemplified the same method in his Mono- 

 graph of the family Melastomaceae ; but he has not thought 

 proper to follow it in later works, either from the difficulty of 

 presenting the characters in the same compendious form, or 

 from the marked natural distinctions between the family de- 

 scribed and other groups. 



One subject particularly attended to by Decandolle in all 

 these monographs was, the relative proportion of the mem- 

 bers of each family found in different regions of the globe. 

 By ascertaining in each instance this point, he hoped to ob- 

 tain, at length, the requisite data for perfecting a branch of 

 botanical science, in Avhich, from a very early period of his 

 life, he had taken a lively interest — I mean the geographical 

 distribution of plants. 



Already, in the year 1807, he had given, in his Flore 

 Fran9aise, a sketch of the several botanical regions into which 

 France might be divided ; and when, in 1817, Humboldt pub- 

 lished his Prolegomena on the Geographical distribution of 



