06 



NATURAL niSTORY OF PLAKTf^. 



carpels, each superposed to a petal,' which appear to cohere for a certain 

 distance along the inner angle, where each ovary contains a vertical 

 placenta bearing two parallel rows of ascending anatropous ovules ; 

 the micropyle looks downwards and inwards. 

 The plant is frutescent, and gives off linear 

 exstipulate leaves. The flowers are terminal, 

 and usually solitary. 



In certain species of P/cnra/K/rn, the sta- 

 mens nearest the perianth are reduced to 

 staminodes ;" and in others, which ofler a 

 good transition between these and the cy- 

 clandrous 1 libber t'lce, there are not only 

 staminodes intermixed with the fertile sta- 

 mens,^ but on the other side of the gy- 

 na^ceum, whose organization remains unaltered are some of these 

 sterile filaments which never become fertile.* Now so nu- 

 merous are the transitions between the species of Hibbcrfia which 

 possess a circular and perfect androceum, and those which possess 

 unilateral stamens, either all fertile or all sterile, that after 

 studying all the species it appears impossible to split them up into 

 sufficiently distinct generic groups.' 



Hihbertia anguslifolia. 

 Fig. 139. 

 Diagram. 



1 We have observed {loc. cit.) how the 

 gynseceum first appears as two carpellary leaves 

 sapcrposed to the petals which alternate with sepal 

 5, and how the a])parcntly alternipetalous dis- 

 sepiment is formed merely hy the floral axis 

 drawn out into a wedge and receiving the inser- 

 tion of the bases of the carpellary leaves on its 

 very oblifjuc fiiccs. 



- This occurs not only in Pleurandra pro- 

 perly so called, but also in UevuMemma (ex. 

 DC, Syxt.. i. 412; Prodr. i. 71; Deless.. 

 Icon., t. 71-77 ; — Km)I-., (Jen., n. 4757;— 

 Walp., Ann., i. 16, some of which are Ocea- 

 nian, while others come from Madagascar — 

 the latter often i)os!;essinp oi)posite or nearly 

 opi)Osite leavi-s. They were collected and stiidied 

 for the first time" by Commkksov and by 

 NoROMiA, who, according to DurKTiT-TnouARS 

 (den. Madat/tmc, IS), gave them the nunio 

 Anidjn. In 11. Cuntmernonil DC, there are 

 flowers without Htcrilo stamens. Kach carpel 

 contains two ov«iK-s. Tlie gynareum is the 

 game in IT. deallmlu H. \Ui., when' tlie insertion 

 of the styles is much bent outwanls. Tlie fertile 

 stamens have long erect linear ititrorsc anthers, 

 the external staminodes ore much shorter. On 



Hemistemma, see also Hook. F., ITooJc. Journ. x. 

 48 ; and F. Mttelleb, F/nqm. i. 151. This last 

 observer has also clearly shown that species like 

 //. spicafa serve as a passage between Pleurandra 

 and llemixtemmn {Fraijtn. ii. 1). 



^ Hemijdeurandra HKNTn.& HoOK. {loc. cit.). 

 " Stamina unilateralia ; staminodia ad utrttmque 

 latus staminnm sita r. in tota peripheria. In 

 Hemistepho Ducmm. & Harv. (Hook. Joum., 

 vii. 51) peduncvli vnilateraliler (X-Jlori et 

 staminodia nonnuUa eliam .fiib niaminihus oh- 

 servantvr." Ilemistephus limaris Drimm. & 

 Harv. was to F. Mueller Hemistemma linen re 

 {Fraffm.,\. 162). The sjime author has projxised 

 a section Diphurandrn for his //. a.<tperifulia. 



* In Jl. angu.st>foHa Hextu. (Fl. Auatr., i. 

 21), the diagram of which is given in fig. 139, 

 we often see two bundles of fertile stamens, with 

 a bundle of staminodes between them — one on 

 each side, and a fourth the other side of the 

 gynieceum. 



* " Gtnus e stnmininn indole commode in 

 nectiows \ diriditiir, fjuarum nonnuUtt ab aur- 

 toribut pro (jeneribux hahentur. yimit tamen 

 artijicialet tunt, ♦•«• habitu contonant." (H. H., 

 loc. cit.) 



