I N May, i92i, the writer observed a curious plant of Cheiranthus cheiri growing in the garden of Merion Lodge, Hagley, Worcestershire. On examination, the plant was found to exhibit interesting deviations from the usual structure. Though the inflorescence was of the normal racemose type the general appearance of the plant was quite unlike an ordinary wallflower. This was partly due to the,inconspicuous nature of the flowers. The petals were greenish-yellow, simple in shape, showing no differentiation into limb and claw, somewhat concave, and much reduced in size. Indeed, they were smaller than the sepals and therefore almost enclosed by the calyx (Figs. I, I a). The sepals were of the typical shape, size and number (2 + 2; 2 saccate), but were inclined to be petaloid. Their margins and apices were usually yellowish. The flowers gave the impression of unopened flower-buds rather than mature specimens. Sepals and petals had dropped from the oldest flowers, but no fruits appeared to be forming. This circumstance may probably be correlated with the absence of an attractive corolla. Selfpollination could not occur for reasons which follow. The most striking feature of the flowers was the entire absence of functional stamens and their replacement by in various stages of development, a stigma and minute ovules being formed in many cases. The characteristic number (6) was everywhere preserved, and often the usual arrangement (2 + 4) could be observed. Another curious feature was the tendency of the rudimentary to fuse with the central gynaecium forming a composite structure (Figs. 2, 3, 6, II, I3)'. This fusion was partial or complete, both as regards individual staminate carpels and the whorl as a whole. Where such fusion occurred the seam was usually well marked. The car-