I "4 PHANEROGAMS. 



whether these facts are sufficient to estabHsh the axial character of these anthers; 

 and these cases may, therefore, be considered for the present as exceptions to the 

 foliar nature of stamens. But, besides, the morphological homology of the 

 separate parts of the ordinary stamens is not yet altogether determined, more 

 precise investigations into the history of development being still wanting in this 

 direction. Cassini and Roper consider the two anther-lobes as the swollen lateral 

 halves of the lamina of the stamen ; their loculi would therefore in that case be 

 mere excavations in the tissue of the leaf; the pollen-mother-cells become differ- 

 entiated inside the young tissue of the leaf, like the spore-mother-cells in the fertile 

 segment of the leaf of Ophioglossaceae. According to this view the furrow between 

 the two pollen-sacs of an anther-lobe (see Fig. 327, H) would correspond to the 

 margin of the staminal leaf; but this cannot be the case^, at least not always, 

 according to Mohl's observations. When the stamens become transformed into 

 petals (by the so-called ' doubling ' of the flower) as in the rose, poppy, Nigclla 

 damasceiia, &c., it may be observed with certainty that the anterior and posterior 

 loculi do not stand opposite one another, which would be the case if one belonged 

 to the upper, the other to the under side of the staminal leaf; but that both are 

 formed on the upper surface, the anterior loculus nearer the median line of the leaf, 

 the posterior one nearer its margin. It is further observable that in such cases the two 

 pollen-sacs of an anther-lobe do not always stand close to one another, but that they 

 are frequently separated by a tolerably broad piece of the leaf, and that this inter- 

 mediate piece contracts in the normal state into the partition-wall between the two 

 pollen-sacs. The greater stress must be laid on these observations of ]Mohl, because 

 in them the abnormal development only shows more plainly what can often enough be 

 seen in a horizontal section of the anther and connective of normal stamens, viz. that 

 the pollen-sacs of an anther-lobe evidendy belong to 07ic side of the stamen ; it appears, 

 however, that they must in some cases be referred to the under (Fig. 327, C, H), 

 in others to the upper side (Fig. 330 C). The origin of the pollen-mother-cells and 

 the development of the wall of the separate pollen-sacs calls to mind so vividly in all 

 essential features the corresponding phenomena in the sporangium of Lycopodiaceae 

 and even of Equisetaceae, that it may be assumed, until more exact observations 

 bring something different to light, that each pollen-sac {i. e. each loculus with its 

 wall) corresponds to a sporangium, and hence also to a single pollen-sac of Cycadeae 

 and Cupressineae ; and that therefore the anther usually consists of four pollen-sacs 

 springing side by side from the anterior or posterior side of a staminal leaf, the sacs 

 lying in pairs so close to one another right and left of the connective, that they 

 coalesce more or less laterally to form one anther-lobe. But before we pass on to 

 the consideration of the pollen-sacs and their contents, we must again recur to the 

 discussion of the entire stamen and androecium. 



The stalk of the anther (the filament with its connective) is either simple or 

 segmented. The simple filament may be filiform (Fig. 329) or expanded into the 

 form of a leaf (Fig. 328), sometimes even very broad, as in Asclepiadeae and 

 Apocynacese ; or it may be broad below (Fig. 332/") or above; it generally termi- 

 nates between the two anther-lobes, but is not unfrequently prolonged above 



11. V. Molil, Vermischte Schriflen, p. 42. 



