EVOLUTION 



II 



fore changed the name of the latter to Pleonandras, so as to cover Neumedia, and thus 

 altered the main character of the group, all the genera of which have flowers like 

 Cjpripedium, and totally unlike those of the Apostasiaceas (PI. 2, figs. 6 and 7). 



The Apostasiaceas have regular flowers,/?w anthers (in Nemviedia versatile), a long 

 Jree style with three very small stigmas at the tip, a trilocular ovary with axile placenta- 

 tion, and dry dust-like pollen. There is nothing comparable with the column of the 

 Orchidaceas. 



The Diandric' have very irregular flowers, sessile anthers filled with viscid fluid in 

 which the sticky pollen-grains are embedded, tliree large confluent stigmas, and uni- 

 locular ovaries (except Selenipedium) with parietal placentation. 



The Apostasiace^ conform with the structural plan so frequent in non-orchidaceous 

 families, of flowers fertilised by insects wliich carry off the dusty pollen entangled 

 in their hairs. They have little in common with the Diandric except the two anthers 

 of the genus Apostasia, and even these are of diiferent shape and position. The whole 

 conception of the flower is diflerent from that of the Diandrx. 



Starting with the genus Nemviedia, which but for the three anthers and the seeds* 

 might have belonged to the Liliaceas, we find that it combines the only fertile anther 

 (^S Text-fig. 2) of the Monandras with the two {a^ and ^2) of the Diandra;. The 

 first step towards the Diandra; is shown by the genus Apostasia (see Text-fig. 3, p. 12), 

 in which the anther A"^ is reduced to a staminode in some species and suppressed 

 altogether in others. The two other anthers have closed together and embraced the 

 style. Next comes a big gap to the genus Selenipedium with the anthers and stigmas of 

 the Diandras, but the trilocular ovary of Apostasia. The next step is to Cjpripedium^ 

 in wliich the ovary is unilocular with parietal placenta. 



On the path of evolution towards the Monandry no intermediate stage has yet 

 been discovered, unless the closing up of the anthers a^ and a^~ around the style in 

 Apostasia be regarded as a step towards their fusion into a column. The stamens and 

 style of Neuwiedia adhere for a very short distance at the base, which may be a first 

 step towards their consolidation into a column, though in no wise as yet comparable 

 with the column of the Monandras. Just as the Diandras may have been derived from 

 Nemviedia by the conversion of the anther ^i into a staminode, so also the Monandry 

 may have arisen from the same source by the reduction to staminodes of the anthers 

 a^ and a^, and by the conversion of the upper stigma into a rostellum. 



It may be objected that European orchids could hardly have been derived from 

 wholly tropical plants peculiar to the Indo-Malayan region. According to Geikie, 

 however, the flora of the mouth of the Thames in the Eocene period was "the most 



' For a fuller account vide O.R. p. 355, Dec. 1932. 



* Wallich, who found only fruiting specimens of Neuwiedia Curtisii, put it down as a LiUaceous 

 plant as, like Asparagus, it has a red berry. 



