STAMENS. 89 



th. They occur, however, in certain species of Ornithogalum (e.g. Ornithogalum 

 itans and cMoranthum), in Allium rotundum and sphcBrocephalum, and in the 

 onkshood (Aconitum). Occasionally such staminal stipules are modified as honey- 

 jreting glands at the base of the stamen, e.g. Boryphora (cf. figs. 214 ^^ and 214 ^''). 



It sometimes happens in monstrous flowers that the stamens are transformed 

 to carpels, or we may find here and there an isolated stamen, which is partly so 

 Ddified and partly still poUiniferous. In such monstrosities it usually happens 

 at it is the upper part which forms pollen, and the lower part which produces 

 ules (cf. figs. 213 ^ and 213 ^). From this and other facts it has been inferred that 

 e ovary corresponds really to the sheaths, the style to the petioles, and the stigma 

 the laminae of the floral-leaves concerned. The monstrous flower of a Saxifrage 

 gs. 213^^ and 213^^) shows that anthers and ovules can be produced from the 

 me part of the leaf -stalk. This flower (213 ^^) produces at the periphery five 

 pals and five narrow, green petals; in the centre two carpels (shaded dark in fig. 

 .3 ^^) as in normal Saxifrage flowers. Between the petals and carpels, i.e. where the 

 amens are usually found, there are ten structures which, whilst resembling both 

 rpels and stamens to some extent, remind one forcibly of the excavated leaf- 

 chis of so many of the Pitcher Plants (cf. vol. i. pp. 125-133.) One of these 

 represented in fig. 213^^. Its free extremity consists of an irregularly serrated 

 ale, which may be compared either to a stigma or to the continuation of an 

 ither, and may be regarded as the metamorphosed lamina. The excavated portion 

 ilow may be regarded as the petiole. In its cavity are four rows of yellow 

 ■otuberances, which might at first sight be taken for ovules. Closer investigation 

 ows, however, that they contain pollen-mother-cells, each inclosing four pollen- 

 ains. Here, then, we find the petiole consisting partly of carpel and partly of 

 ithers, from which it may be concluded that that portion of the carpel which 

 oduces ovules corresponds entirely in position to the pollen-producing tissue. 



The parts of the anther which produce Pollen in special chambers are known 



Pollen-sacs, the tissue which binds these together as the connective. The 

 nnective is a direct continuation of the filament, and, like this, is penetrated 

 r a vascular bundle. The pollen-sacs may be arranged like niches around the 

 lumnar connective, which itself terminates in a sort of little shield, as in the 

 3W Tree (cf. fig. 234 ^), or they may be situated symmetrically right and left of it. 



the latter case the pollen-sacs may lie at the edge of the connective in one 

 ace, as in the Juniper (figs. 214^^ and 214^*), or they may be in pairs, i.e. two 

 lien-sacs to the right and two to the left of the connective (fig. 214 ^). This 

 ;ter form is by far the most frequent, and occurs in certainly 90 per cent of 



Phanerogams. It must be pointed out that the two pollen-sacs of each pair 

 e separated from one another by a partition-wall only in the young anther, 

 lis disappears later on, and in the mature anther one finds, instead of four, only 

 'O sacs filled with pollen. Sometimes all four pollen-sacs run together in this 

 ty, by the breaking down of the parti-walls, as in Sundew (Drosera), Moschatel 

 doxa), Monotropa, and especially in Globularia (cf. figs. 216^' and 216^). In 



