230 XIV. CEUCIFEE^. 



double stamen. This theory does not acconnt foT the lower position of the two antero-posterior stamens 

 relatively to the lateral sepals, nor for the situation of the short stamens opposite to the carpels, which 

 would be contrary to the laws of alternation. 



Later botanists (Lestiboudois, Kunth, Lindley, and then J, Gny, Schimper, Wydler, Krause, 

 Duchartre, Chatin, Godron) advocate an entirely different theory. They do not admit the doubling of the 

 long stamens ; they aifirm, contrary to the organogenic observations of Payer, that in the very young 

 flower each group of twin stamens springs from two distinct protuberances, separate from each other and 

 exactly opposite to the petals. They consider the androecium to be composed of two quaternary whorls : 

 1st, the lower whorl, represented by the two lateral stamens only, and which is an imperfect one, from 

 the constant arrest of two stamens which should be developed in front of the antero-posterior sepals ; 

 2ndly, the upper whorl, composed of the four large stamens which were originally developed opposite to 

 the petals, but which approach each other afterwards, so as to form two pairs. As to the pistil they 

 consider it to be normally formed of four carpels opposite to the four sepals, an arrangement which is 

 observable in the genus Tetrapoma. Thus the original plan of the flower may, according to them, be 

 formulated as follows : four sepals, four petals, four outer stamens, of which two are never developed, four 

 inner stamens, and four carpels, of which the two antero-posterior are developed in Tebapoma only ; all 

 these whorls exactly alternating with each other. 



More recently, A. G. Eichler has published (in 1865), in the ' Flora of Brazil,' the result of his 

 researches. He afiirms, with De Gandolle, that each pair of long stamens results from the splitting up, 

 or chorisis, of a single stamen, because, according to his organogenic researches, the protuberance from 

 which each pair of stamens springs is originally simple, and only divides afterwards. With regard to 

 the two-celled anthers of the twin stamens, which, according to the partisans of the non-development 

 theory, ought to he one-celled, Eichler avers that this objection is valueless ; that the question here is 

 not of a doubling, which divides an entire organ in two halves ; but of a chorisis, which results in a 

 sort of multiplication of the organs ; and that, further, in the genus Atelanihei-a the long stamens are 

 constantl}' one-celled. 



With regard to the polyandry observable in some species of Megacarpaa, in which the androscium 

 is composed of 8-16 stamens, we may, according to Eichler, admit that it results from an unusual 

 multiplication of the long stamens, and that the chorisis has been extended to the lateral stamens. We 

 must, besides, remember that this tendency to multiplication (which is really exceptional, and not found 

 in all the species of Megacarpma) also appears in CleomeiB, a tribe of Capparidem, a family closely allied 

 to Cruciferm ; their androecium, which is normally hexandrous, and arranged lilte that of Crudferm, 

 presents in some species of Cleome four stamens, and in Polanisia eight or oo, collected in antero- 

 posterior bundles, the two lateral remaining solitary, or very rarely being represented by stamens. 

 According to Eichler's organogenic observations, the anterior sepal appears first, and then the pos- 

 terior, after which the two lateral sepals appear together ; the four petals then appear simultaneously, 

 and occupy four points, diagonally crossing the lateral sepals. The androecium first appears as two large 

 obtuse gibbosities, opposite to the lateral sepals, which remain simple, and become the short stamens. 

 Soon after their appearance the two similar antero-posterior gibbosities appear, inserted higher than the 

 preceding, larger and more obtuse ; and these, enlarging more and more, gradually divide into two 

 protuberances, which finally become tvro long stamens. Never, says Eichler (who energetically maintains 

 his opinion against that of Duchartre, Chatin, and Krause), never are these stamens, when young, exactly 

 opposite to the petals ; they are, on the contrary, then nearer to the median line, an arrangement which is 

 still more obvious in some hexandrous Capparidete. He affirms, contrary to the observations of Chatin, 

 that in Cruciferce with fewer than six stamens {Lepidium mderale, latifolium, virginianum, &c.) the lateral 

 stamens are inserted lower than the two antero-posterior, which proves that the latter do not belong 

 to a lower whorl, as the incomplete-development theory demands. Eichler consequently regards the 

 theory of chorisis as true, but his application of it ditfers from that of De Candolle, in that the latter 

 makes three tetramerous whorls (androecium, corolla, and calyx), while Eichler only admits one, i.e. 

 corolla, and assigns a binary type to the androecium and calyx ; and his view of the composition of the flower 

 would be : — two antero-posterior sepals, two lateral sepals, four petals diagonally crossing the lateral sepals, 

 two short lateml stamens, two antero-posterior stamens (each doubled), and two lateral carpels valvately 

 juataposed. 



The family of Cniciferce is closely allied to Capparidem, Papaveraeem, and Fumariacece (see these 



