ARTIFICIAL CLASSIFICATION. 93 



This necessity seems to be most urgent in regard to the 

 sexual parts, as the principle of the classes from the twenty- 

 first to the twenty-third. The later scholars of Linnaeus as- 

 sure us, that he wished to leave out the twenty-third class as 

 altogether unnecessary. It was reserved, however, for Sir 

 James Edward Smith, the worthy inheritor of the Linnaean 

 treasures, to propose the happiest alteration* This consists in 

 recognising the difference of sex as then only essential, when 

 it manifests itself by an actual difference in the structure of 

 the female, male, and hermaphrodite flowers ; w hen thus the 

 male-blossoms, as in the Oak, stand in catkins, but the fe- 

 male flowers are insulated. This difference, not of sex alone, 

 but also of forms, is called by Smith didinia ; and he accord- 

 ingly rejects from the twenty-first, twenty-second, and twenty- 

 third classes, all the genera, in which no such correspondence 

 between the difference of sex and the difference of form takes 

 place. Hence among others, the genera Acer, Vtratrnm, 

 Hydrocharis, Stratiotes, Sagittaria, and so forth, are placed 

 much rather according to the number of their filaments, tlian 

 according to their difference of sex. 



139. 



Less fortunate attempts to improve the Linnaean systemi 

 have been made by Thunberg, Suckow, llebentisli, and 

 lately by Claude Richard. According to Thunberg's plan, 

 the Gynandria is removed, and is brought, with peculiar im- 

 propriety, under the second class. He also rejects Monoccia^ 

 Dioecia, and Polygamia, whilst he classes all these plants ac- 

 cording; to the numbers of their anthers. Otliers have thrown 

 away the Monadelphia and Diadelphia, and have acknow- 

 ledged the numerical proportions as the only principle of clas- 

 sification. But by these means the difficulties have only 

 been made greater, and some classes have been overloaded 

 with genera, so that their investigation has been rendered 

 much more troublesome. 



Richard's later proposal is founded partly on a pretended 

 better division of the Polyandria and Icosandria, from wliich 

 he separates a new class, Hysterandria, in which the fihuneIlt^* 



