30.5 



so Giseke, setting out on erroneous principles, has 

 presented us with a rumbling waggon-load of new 

 hard-named genera, dismembering the old ones, not 

 only by insufficient characters, but by characters 

 that do not exist, and establishing new ones with as 

 little scruple or success. 



"Mr. Roscoe first suggested a method of reducing 

 the genera of the Scita?ninece to regular order, by 

 essential characters derived from the structure of 

 the stamen, particularly its filament. 



"This principle is found to be the only one which, 

 while it is clear and precise in defining technical 

 essential characters, leads to the establishment of 

 natural genera. 



" It must be observed, that the learned author of 

 the Prodr. Nov. Holl. (Mr. R. Brown,) follows the 

 hypothesis of Jussieu in not allowing a corolla to 

 these plants, inasmuch as they are monocotyle- 

 donous. 



" Mr. Roscoe's ideas of a Scitaminean flower are 

 exactly consonant with Mr. Brown's, except that 

 he considers the inner perianth of the Thalia as a 

 corolla with a double limb, which is the most obvi- 

 ously natural mode of considering it, and in which 

 I without hesitation concur. 



" The Scitaminea as well as the Cannes are pro- 

 perly placed in the Monandria Monogynia of the 

 Linnaean sexual system." (See Sir J. E. Smith's ar- 

 ticle Scitami?iece in Rees's Cyclopaedia.) 



In one of the latest letters addressed to Mr. Roscoe 

 by Sir James, in 1827, he congratulates him on the 



VOL. II. x 



