\ 



^: 



J, 



i 

 i 



68 MR. J. G. BAKEIl*S aYSTEMA IRIDACEAKUM. 



Iridese the three inner segments are permanently convolute, — all 

 arrangements prohably in some way connected with fertilization, 

 the import and mode of action of which have still to be worked out. 

 Stamens. — In the anther throughout the order there is no ma- 

 terial difference in struetui^e ; neither can we regard the differ- 

 ence between the epigyuous and perigynous insertion of the sta- 

 mens as here of much account, as one slides into the other very 

 gradually. The three stamens, whatever may be their direction 

 in the expanded floAver, are always inserted opposite the outer 

 segments of the perianth. In the Australian genus Diplarrhena 

 the three filaments are unequal in length, and one of the anthers 

 is abortive. In one solitary instance, for a typical Iridaceous 

 plant {EleiitJierine anomald) Herbert has recorded the presence of 

 six stamens ; but it has never been seen again, and may perhaps 

 have been a monstrosity of the common American EletUherine 

 plicata. Mr, Bentham regards as an anomalous member of Iri- 

 dacese the Australian genus Campynema, which has always six 

 stamens. I have spoken already of the different direction of the 

 stamens in the expanded flower, and how in this respect we get 

 two well-marked types correlated with the shape of the perianth. 

 The three filaments, when free, never become dilated andpetaloid ; a. 



but in a considerable number of Iridacea? the stamens are more or 

 less decidedly monadelphous ; and for this reason, in the old books, 

 the order is separated out under tw^o Linnean classes far removed 

 from one another. In a small number of genera, such as Tigridia^ 

 Tieiisseuxia^ and Qalaxia^ the filaments are completely and 

 thoroughly monadelphous ; and in a large number, such as Crocus 

 and GladiohtSj they are completely and unmistakably free. But 

 in other genera, well marked by habit, inflorescence, and perianth- 

 characters, such as Sisyrinclimm and Libertia,vie get very various 

 degrees of union in plants thoroughly uniform in other points. 

 Txia^ a well-marked triandrous generic type, embracing wide spe- 

 cific diversity, has two monadelphous representatives, Ixia mona- 

 delpha and curta. Bomuleay a genus with 36 triandrous species, 

 with a habit peculiarly its own, has a monadelphous analogue in the 

 single species on w^hich Sweet founded his genus SpatalantTius. 

 So that, as a whole, we must say that triandrous pass gradu- 

 ally into the monadelphous Iridacese along several lines of 

 junction. 



PistiL — In the ovary there is no material diftercnce in the order, 



i 



i 





% 



t 



} 



L 





