PLANTS JAVANICjE llARIORES. 623 



be either transverse, that is, having its radicle distant from 

 the umbihcus, or approximated to it. 



The want of Albumen in this tribe is also a character of 

 considerable value, but its absence is not quite so frequent 

 as has been supposed, there being some room for doubt 

 w^here its union with the cotyledons is most intimate. 



The deviations from ordinary arrangement of the Anthercs 

 may be considered as next in value to those already enume- 

 rated ; of these modifications, that of the greatest importance 

 is where the antherse are disposed in a simple series and 

 equidistant ; this occurs I believe only when they are ten 

 in number, or in the very rare case where they are reduced 

 to five ; the mere number, where they are not equidistant 

 nor closely approximated, seems to be of much less im- 

 portance. 



The degrees of development of stamina in the female 

 flower can never be employed in the formation of genera : 

 in several, perhaps in many cases, the antherae contain 

 pollen, but, on comparing it with that of the male flower, 

 I am disposed to believe it to be imperfect or effete. It 

 would seem at first sight, that, in the numerous cases where 

 the style is deflected, the stigmata are ^o closely approxi- 

 mated to the antherae, that impregnation by their pollen is 

 at least probable ; but by this deflection of style the stig- 

 mata are equally exposed to the influence of the antherae of 

 the male flowers, which are generally lateral, while the 

 female is terminal, the exposure being nearly equal, whether 

 the inflorescence is pendulous or erect. 



The modifications of calyx in Sterculiecs, especially in 

 the depth of division and direction of its segments, have 

 hardly more than specific or sectional value; and even 

 those processes which occur in several species, either in the 

 segments or tube, seem to be of no greater importance. 

 The most remarkable anomaly of all is the in duplicate 

 aestivation found in a single species belonging to New Hol- 

 land and in Sterculia villosa of Roxburgh. This mode of 

 aestivation readily passes into the valvular in the corolla of 

 several families, and is therefore of no great weight. 



It is singular that so great a difference as that between 



