THALAMIFLOE.E. 211 



ILLUSTRATIVE GENERA : Cheiranthus, R. Br. ; Arabis, L. ; Alyssum, 

 L. ; JErophila, DC. ; Cochleariq, L. ; Iberis, L. ; Cakile, Tournef. ; Ery- 

 simum, L. ; Capsella, Vent. ; Lepidium, R. Br. ; Brassica, L. ; Crambe, 

 Tournef. ; Raphanus, Tournef. ; Subularia, Adans. ; Sehtzopetafon, Hook. 



Affinities, &c. The relationships of this Order to Papaveracese and Fu- 

 mariaceae, and thence with the apocarpous Orders standing near, have 

 been dwelt on already. With the Capparidacese the agreement is still 

 closer, in the general character of the flower and in the seeds ; but when 

 that Order has so few as six stamens they are not tetradynamous an 

 almost universal condition in Cruciterae. The true nature of the plan 

 of the anomalous flower of this Order is a subject of considerable con- 

 troversy. The numerical arrangement isS2+2P4A2 +4 G 2, disposed 

 thus: S.. + ; P ; : A., -f- ; : G.. . The two lateral sepals are attached 

 higher up than the other two, the four petals are in a single whorl. Lindley 

 and many others regard the six stamens as belonging to two circles of 

 four, the outer of which has always two stamens abortive, while the inner 

 pairs should normalty stand singly before the four petals. If the glands 

 in the receptacle are to be regarded as abortive stamens, which is 

 plausible, this structure really does exist in JErysimum Perojfskianiim, 

 where the two glands stand opposite two sepals and form a circle with 

 the two short stamens ; but there appears to be as many as six glands in 

 some Crucifers, which involves the existence of three staminal circles : as 

 a polyandrous form (Megacarpcea polyandra) has been met with, we may 

 even admit this. De Candolle, Moquin-Tandon, and Webb regarded 

 the stamens as normally four, the pairs being formed by chorisis. As to 

 the ovary, it is explained as being composed of two carpels with a 

 spurious dissepiment: the stigmas opposite the placentas, like those 

 of Papaver, being double and composed of a half from each carpel, just 

 as the placentas are. The fact that four carpels are sometimes found 

 in monstrosities, and are constant in the genus Tetracellion, merely indi- 

 cates a return to a symmetrical condition, ordinarily interfered with by 

 suppression of two carpels. The most recent views as to the structure of 

 the flower are those of Eichler, who affirms the existence of chorisis in the 

 long stamens. His notion is that the flowers of Crucifers consist of two 

 antero-posterior sepals, two lateral ones, four petals crossing the lateral 

 sepals diagonally, two lateral short stamens, two antero-posterior long 



stamens split into two, and two lateral carpels ; S 2+2 P 4 A 2-f- 2 G 2. 

 These views are adopted after an examination of the development of the 

 flower, in which there are first to be seen two tubercles for the fore and 

 aft sepals, then two for the lateral sepals ; the four petals originate simul- 

 taneously ; the two short stamens arise opposite the lateral sepals, the 

 two other staminal tubercles are subsequently developed at a higher level 

 than the preceding and decussate with them j these two after a time be- 

 come notched, the notch gradually deepens and ultimately forms two 

 distinct stamens. The position of the carpels is sometimes antero- 

 posterior (;) sometimes literal ( ..). The glands of the disk also vary in 

 number and position, and may, as suggested by Mr. Worthington Smith, 

 be the representatives of abortive stamens or pistils, which would, if 

 fully developed, render the flower symmetrical and isomerous. Van 

 Tieghem considers the gynaecium to consist of four carpels, two of which 

 are ovuliferous and terminate in stigmas, while the other two are sterile. 



P 2 



