136 Development of the Natural System under [BOOK!. 



idea is not improved by the fact that he ascribes to the organs 

 physiological characters which they do not really possess ; thus 

 he regards the vessels as the most important organs of nutri- 

 tion, which they are not in fact, and upon this double error 

 he builds his primary division of the whole vegetable kingdom 

 into vascular and cellular plants, and then by a third mistake 

 believes that this division coincides with the division of plants 

 into those which have and those which have not cotyledons. 

 The already established division into Monocotyledons and 

 Dicotyledons, which rests upon a leading and purely morpho- 

 logical mark, is spoilt by De Candolle through his following 

 Desfontaines in ascribing to the Dicotyledons a different mode 

 of growth in thickness from that of the Monocotyledons, and 

 characterising the one as exogenous, the other as endogenous. 

 But this notion is utterly incorrect, as von Mohl showed twelve 

 years later ; and if it were correct, it would still be unimportant 

 in a systematic point of view, because it appeals to a mark 

 which is morphologically of quite subordinate importance. 

 The worst consequence of these mistakes was, that the 

 Vascular Cryptogams were introduced into the same class 

 with the Monocotyledons, a decided step backwards, if we 

 compare De Candolle's system with that of Jussieu. In spite 

 of these grave defects in the primary divisions of the whole 

 vegetable kingdom De Candolle's system deserved the fame 

 which it acquired and long maintained ; it had this advantage 

 over Jussieu's system that in the class of Dicotyledons, the 

 largest division of the whole kingdom, larger sub-divisions 

 appeared, and these served to unite families that were in 

 many points essentially related ; the Dicotyledons were in 

 fact divided first of all into two artificial groups according to 

 the presence of two floral envelopes or one; the first and 

 much the larger of these was again broken up into a series of 

 subordinate groups, which pointed in many ways to natural 

 affinities. That these groups, which have only quite recently 

 been materially altered, did to a very considerable extent take 



