TAXONOMY AND CLASSIFICATION. 83 



iscovery of the hermaphrodite flower. . . . The flower, with its 

 reat stamens 10 cm. long in some species, must have been 

 )Ioration of the perianth . ,. mofc te u no ^ 



- of the carpels 

 id ovules in some recent cycads suggests the probabih 



The more elaborate flowers 

 >/>hyid." 

 215. In this case, again, there is difficulty in tracing any homology 

 Jtween the sphorophylls of the gynseceum and the carpels 



the stalks of the 



isted at that tii 

 i i ium. A-, to the rest of 



•perms, and as the gvnaeciu] 

 apocarpous, there is an extiani.lmarv -'nulmlv i 

 '■ ['■-, and -m-h a,-- one -. 



oi flower among the Ranales, especially the Marpwl 



by Newell Arbe°r aid 7ohn\^Jk^\heNymphwacel, 

 other polycarpica among Dicotyledons,' ,l/;.«„„/r,,r, 

 mmmtm among Monocotyledons are taken as i 

 primitive features, while the Piperahs, Amentifer 

 with very simple flowers, are regarded as derived 



'*" < <»mpli. at,,! ones by a process of reduction. ' 

 generally exhibit such a network of cross alliances 



and partly from a stock allied to the ( 

 ' -ivatives of one stock f rorr 



ne to be in favour of the theory o 



lany cases the derivatives of one stock from t 

 « is to m« 



Inglenan view of 

 uglandacece, etc., 



led to the Ri 



Jtropous ovuh 



as. The last 



Oosperms are derived from a s 

 ars to meet moi 

 ' it is interesting to note that A 

 .v,. ia „, action of the flower of a hemi-ai 

 witwi.* 81 * 68 almost exactly as regards the structure of the stan 

 n the subsequently discovered Williamsonia mexicana of VViel 

 anH +t ^ there are striking analogies between the Angiosp< 

 ™ Wie Cycadeoids is undeniable. It is also true that the analc 

 ° 0me accentuated if we take into consideration the older and i 



amphitropous ovules. The 



217. On the wl 

 J*}, the view t: 

 toe facts, an d Sc( 



