248 Dr Dickson on Diplostemo^ious Flowers, dec, 



carpels is unaffected by such disappearance. This, of course, 

 is only what might have been expected, if the outer stamens 

 are viewed as merely accessory parts. Thus, in the Ericacese, 

 we have in most of the species of Ledum the apparent dip- 

 lostemony which is frequent in the family, while in Ledum 

 latifoUum the younger stamens disappear ; and yet in this 

 species the carpels are superposed to the petals just as in 

 the others.* In Epacris, so nearly allied to the heaths, we 

 have also an absence of the stamens superposed to the 

 petals ; and yet the carpels have the same position as in the 

 Ericaceae. Contrasted with this, it is striking to observe 

 the consequence of the disappearance of the younger starnens 

 in a group of plants exhibiting what I believe to be a true 

 diplostemonous arrangement — one where the younger 

 stamens are the more internal. I have already mentioned, 

 as an example of this arrangement, a Biittneriaceous plant, 

 Lasiopetalum corylifolium, the organogeny of which has 

 been given by Payer. Here, the fertile stamens, which 

 form the outer and older whorl, are superposed to the 

 petals ; the inner and younger whorl consists of staminodes 

 alternate with the older stamens ; and the carpels alternate 

 with the staminodes, and are thus superposed to the petals.f 

 In Hermannia, on the other hand, where only the fertile 

 stamens superposed to the petals exist, the carpels are no 

 longer superposed to the petals, but are now found super- 

 posed to the sepals, occupying, apparently, the place of the 

 missing staminodes. In the Dombeyese, Baillon has de- 

 scribed in Astrapcea a single whorl of staminal groujDS, with 

 the carpels similarly superposed to the sepals. J 



* I am indebted to a friend for the facts regarding Ledum. 



t In the Biittnerese, moreover, Baillon has shown [Adansonia, II. p. 168) 

 that in Buttneria, &c., the fertile stamens superposed to the petals are, as 

 in Lasiopetalum corylifolium, older than the staminodes which alternate with 

 the petals, and that the carpels are also, as in that plant, superposed to the 

 petals. He has not stated which of these whorls is the more internal ; but I 

 can scarcely doubt that further investigation will show that the staminodes 

 are the more internal. 



X Payer has observed (Organogenic, p. 45) that in Melhania the carpels are 

 similarly superposed to the sepals, and I have been able to confirm his observa- 

 tion. I have also found that in Pentapetes the carpels occupy the same position. 

 It is probable that this arrangement obtains in the Dombeyese generally. 



The genns Mdhania exhibits the simplest form of andrcecium that occurs in 



