Cardtius and Cnicus *which appear to he diceciotis. 15 



stance : it forms a part of his charactei- of the genus Baccharis^ 

 which Richard and Jussieu had previously proposed to Hmit 

 to such species as were dioecious, and which, thus compre- 

 hendmg Molina of the Flora Peruviana, contains many spe- 

 cies. The plants forming two of the new genera there pro- 

 posed [Petrobium and Brachylcena) he has ascertained to be 

 dioecious ; another genus, Piptocarpha^ he suspects to be so ; and 

 the dioecious Gnaphaliums (to which he shows that margarita- 

 ceum must be added) are also thrown into a separate genus. 



It will be observed, that the greater part of the genera men- 

 tioned belong to orders which have floiets of different sexes in 

 the same capitulum; in such the prevalence of one sort of floret 

 in all the capitula of a plant, to the exclusion of the other, is a 

 circumstance not so unexpected as m the order Syngenesia 

 jEqimlisy where all are hermaphrodite ; to this, however, Pe- 

 trohium and Brachyl(sna are referable ; and Mr. Brown's de- 

 scription (in the same paper) of the separation of the sexes in 

 Serrattda tinctoria, led me to notice the same circumstance in 

 Serratula, or, as it is now most frequently called, Cniais arvensis, 

 and in some other species of the genera Cardiius and Cniais, ail 

 of which were supposed to have hermaphrodite flowers only. 



So long ago as the year 1807 I had observed that there were 

 many plants of Saratula tinctoria in which the antherae were 

 entirely abortive; but finding others in which all the oro-ans 

 were apparently perfect, it did not occur to me that there vi'as 

 any separation of sexes. 



On re-examining this plant, in consequence of Mr. Brown's 

 obsen-ations, the striking difference between the male and fe- 

 male flowers, which had formerly induced me to look for some 

 specific difference between the plants bearing them, appeared 

 to point out a very ready mode of examining the nearly allied 

 species by the external appearance of their capitula without 

 the labour of a minute dissection. 



Looking at Cniais arvensis with this view, I soon found that 

 different ])atches of it had flowers which presented differences 

 similar to those of the Serratula tinctoria, and dissection con- 

 firmed the external appearances ; by the examination of very 

 many specimens, I ascertained that some plants bore flowers 

 the anthera? of which were invariably abortive, and that in others 

 the ovaria as invariably withered without producing seeds. 



A more detailed account of the differences between the male 

 and female flowers is as follows : 



Tlie female florets are somewhat shorter and smaller tlian 

 the male, particularly the laciniae and dilated part of the tube 

 of the corolla ; liencc the male capitulum, when hi flower, ap- 

 pears 



