STAMENS-SEEDS. 



v\ 



calyx. Beyond this point it follows the curvature in succession of the tube, of the hood, 

 and, if it be present, of the beak. As a rule the style is of uniform thickness, but in 

 one species (P. superba) it increases in thickness towards the stisrma. 1 The stigma is 



e -«~ ~"c 



sub-capitate, generally (though rarely very far) exserted, but sometimes distinctly included 

 (P. lachnoglossa). The enlarging fertilised ovary detaches the corolla from the disc and 



at the same time splits the tube upwards from the base in the middle line behind. It 

 is rare, however, for the corolla to be immediately caducous (P. ezcelsa); as a rule it 

 becomes gradually extruded by the elongating capsule till the base is clear above the 

 mouth of the calyx tube. Then as a rule it either falls off or is blown away, but it is 

 by no means unusual to find the tip of a ripened and dehiscing capsule still crowned 

 by a withered corolla. 



The capsule varies a good deal in the genus as a whole, but is very constant in 

 size and form within specific limits and often throughout natural groups of species. The 

 relative length of capsule and calyx affords a more useful character than that of < 

 and corolla tube. This is not alone because the capsule is much more constant than the 

 corolla tube, which is, as has been seen, rather variable, but because the teeth, to which 

 the variation of the calyx is principally confined, usually become in the fruit shrivelled 

 and less conspicuous, and because the enlarging capsule by expanding the calyx tube 

 renders the limbus more easily discernible between the bases of the teeth. The usual 



shapes of capsule are lanceolate or ovate-lanceolate, rarely sub-rhomboid (P. rex), and in 

 these cases with the apex oblique and in front of the middle line; not infrequently, however, 

 the capsule may be ovate (P. trichoglossa, pycnantha), or rarely sub-globular (P. fragihs), 

 and in these cases with the apex erect and central. Its surface is smooth and variously 

 reticulated, and (except in P. fragilis, where the capsule is transparent and almost herba- 

 ceous) its walls are of a firm consistence; it varies in colour (in herbarium specimens) 

 from pale tan to black. The capsule is persistent, being always found attached to the 

 stems of a former season where these happen to be collected. The dehiscence is loculi- 

 cidal, but is confined to the upper third of the anterior and the upper half of the posterior 

 suture, the valves nearly always remaining coherent (even in specimens showing stems 

 of a second season) from the base to opposite the upper limit of the placental surfaces. 



The seeds vary greatly throughout the genus in number, size, shape, colour, and 

 surface markings ; in individual species, on the other hand, these characters are very 

 constant. In number they vary from 3-4 in each locule (P. gibbera) to 18 or 20 

 (P. dolichorrhizay, in size from 0-6 X 0'6 mm. (P. fragilis) or VZ X 0-7 mm. (P. carnosa) 

 to 3-5 X 2'0 mm. (P. Clarkei) or 425 X 2-0 mm. (P. pycnantha)- in shape from spherical 

 (P. fragilis), ellipsoid (P. Clarkei), egg-shaped and narrower at the lower extremity 

 (P. longiflora), egg-shaped and narrower towards the apex (P. gibbera), to nearly fusiform 

 (P. curves). In transverse section the seeds are usually rounded or ovoid, rarely tr, 

 quetrous (P. schizorrhyncha). The hilum is indicated by a furrow, the upper portion of 

 which is often filled by a spongy crest which, though of different appearance and lighter 

 colour, is in reality only a modified portion of the testa developed from the coats of 



, i JL m„A fka rmVrnnvle At the lower extremity of the seed there 



the ovule where these surround the micropyie. ***> wo j 



i • • „u„* .nnnora +o be a distinct process. But this also merely 



is also in many species what appears to d« <*■ u«* f j 



consists of testa, not even modified, and has no special structure. It ,s wort obv,ou> 

 where the testa is thin and closely adpressed to the albumen, and u not viable at all .f 



l Maiimowicz, Mel. Biol, sii 874. 



