4l6 FOURTH GROUP.— SEED-PLANTS. 



Increase in the number of the members of a floral whorl may occur in different ways, 

 as the following examples show. According to Eichler's ^ exhaustive investigations the 

 flowers of the Fumariaceae may be referred to a type, in which there are six decussate 

 pairs of members : 



two median sepals, 

 two outer lateral petals, 

 two inner median petals, 

 two lateral stamens, 

 two median (always abortive) stamens, 

 two lateral carpels. 

 The two lateral stamens are however represented in some Fumariaceae {Dicentra, 

 Corydalis) by two groups each of three stamens ; each group consists of a middle 

 stamen and two on each side of it ; the former has a quadrilocular 

 (entire) anther, the latter have each a bilocular (half) anther, and 

 Eichler explains this by assuming that the two lateral stamens 

 are stipular formations, branches therefore from the base of the 

 middle stamen. In Hypecoum Eichler assumes a cohesion of 

 each pair of opposite stipular stamens, so as to produce a 

 staminal whorl apparently of four members. 



According to the same author the flowers of the Cruciferae and 

 Cleomeae (a tribe of Capparideae) may be derived from a type 

 gram'^fFlmariacere. Tf- which is represented in Fig. 346 A, and appears as the empirical 

 ter Eichler. diagram in Cleome droserifolia, in species of Lepidhim, in 



Senebiera and Capsella. This typical flower consists of — 

 two outer median sepals, 

 two inner lateral sepals, 

 four diagonal petals in one whorl, 

 two outer lateral stamens, 

 two inner median stamens, 

 two lateral carpels. 

 Deviations from this type are caused by two or more stamens appearing in place of 

 one of the inner stamens, in the Cruciferae usually two (Fig. 347), in the Cleomeae 

 sometimes two, sometimes more (Fig. 346 B). Such a substitution of two or more 

 stamens for one is called by Payer dedoiibleinent (doubling, duplication), by Eichler and 

 others collateral chorisis, and may apparently be regarded as branching which takes 

 place at a very early period ; this is supported in the present case by the fact that in 

 AteJafithera, one of the Cruciferae, the median stamens are only split and each branch in 

 the split stamens has only a half anther, while in Crambe each of the four inner stamens 

 puts out a sterile lateral branch, which may be explained as the beginning of a still further 

 increase of the stamens, such as actually occurs in the Cruciferous genus Megacarpaea 

 and in many Cleomeae. Though we may be unable to give a mechanical explanation 

 of the increase in the typical number of the stamens of the inner whorl and the history 

 of their development may be obscure, yet it seems certain that the instabihty of the 

 number of the members of the androecium points to the view, that a deviation from the 

 original typical number of two members has arisen in this part of the flower of the 

 Cruciferae and Cleomeae, while the other floral whorls maintain a striking constancy : 

 in Tetrapo7na and Holargiditcm only amongst the Cruciferae the gynaeceum so far 

 varies as to have two median carpels in addition to the two lateral, and there is thus a 

 four-valved ovary. 



The theory of the duplication of an originally simple stamen has been applied in 

 cases where it is inadmissible. In many cases this supposed simple formation is not to 



'I 



1 Ueber d. Bliithenbau d. Fumariaceen, Cruciferen u. einiger Capparideen (Flora, 1S65, Nr. 

 28 35, u. 1869, p. i), and Bliithendiagramme, p. 195, where other works are cited. The flowers 

 of the Fumariaceae however admit of a different explanation (see on page 41S;. 



