PHANEEOGAMIA : FLOWERS, 363 



the epidermis (here termed the peach, bursicula) (figs. 209. c. 211. /*.), 

 one or two small groups of cells which become filled with viscine, and 

 are in part dissolved into it (figs. 211.*. 210. c.). The membrane covering 

 them is gradually decomposed, and they then lie free, and are termed 

 the retinacula (fig. 210. b. c.) ; if the epidermal membrane becomes de- 

 composed very early, they are termed retinacula nuda. In the last case, 

 the cellular tissue which separates the point of the caudicula from the 

 retinaculum is also decomposed at the same time, and thus caudicula 

 and retinaculum come into union (this also, therefore, happens before 

 the anther opens). In the first case, on the contrary, the two are fre- 

 quently separated, but so placed, that as the anther opens, the very 

 slightest movement of the pollen mass brings the point of the caudicula 

 into contact with the then always bare retinaculum, so that they adhere 

 together. It is perfectly easy to follow this process in the last- mentioned 

 case, in Orchis militaris, and especially easy in the very long rostellum 

 of the Neottiece. Gymnadenia albida and conopsea offer good ex- 

 am pies of the other case. The Orchidece diandra and the Aposta- 

 sies exhibit none of these remarkable peculiarities, having their anthers 

 regular, and pollen granules which do not adhere together. 



I have never yet succeeded in penetrating the very first stages of the 

 formation of the flower ; the little I have seen in Orchis latifolia and 

 Cypripedium Calceolus gives me ground to presume that only three 

 stamens are ever formed in rudiment, of which, in Cypripedium, one is 

 developed in a foliaceous form, whilst in the rest of the Orckidacece, two 

 are abortive, or appear as two little fleshy scales, which, in consequence 

 of the one-sided, excessive development of the upper floral envelopes 

 (the labellum), are pushed to the side of the single developed stamen. 



The most incomprehensible structure of the anther, if I have seen it 

 in the true light, occurs in Caulinia. Here, both in male and female, a 

 bracteole is developed into a pitcher-shaped organ, which in the female 

 is two-lobed above, resembling a germen with two stigmata ; in the male 

 splits up in the upper part on one side, imitating a perianth. 



On the little conical body which is embraced by each bracteole, an 

 envelope is formed, in both sexes, in the manner hereafter to be de- 

 scribed in the seed-bud ; and at this time it is impossible to determine 

 whether mas or "femina is being formed ; but then they begin to deviate, 

 and in the femina the seed-bud produces a second integument, and becomes 

 inverted, whilst in the mas the cone grows up into a large nucleus ; and 

 while this becomes invested by the envelope until only gradually a little 

 canal remains at the summit, it is entirely (?) converted into pollen, which 

 then finds an outlet by that opening at the apex. Finally, Brosimum Ali- 

 castrurn also appears to possess a most aberrant formation of anther. 

 The beautiful representations of this in our Botanical works look strik- 

 ingly like elegant, newly-turned chessmen ; and without knowing the 

 structure in nature, one may assert that the pictures have little resem- 

 blance thereto. 



generation, in the direction of the arrow (#, y) in fig. 209. a, b, Lower part of the left 

 half of the anther ; c, epidermis ; d, parenchyma of the walls ; e, pollen mass ; f, cau- 

 dicula ; g, point of the rostellum, here very short, h, epidermis (bursicula) ; i, reti- 

 naculum; k, part of the bursicula, which is subsequently dissolved, so that the retiniculum, 

 now free, comes in contact with the candicula, also set free by tho bursting of the 

 anther; /, loose, easily separable cellular tissue; i, outer surface of the conducting 

 cellular tissue (the stigma) ; w, parenchyma of the disc, forming a style. 



