466 
Saunders.—A Reversionary Character 
function. The two solid carpels alone, in these cases, are prolonged upwards 
to form a style, consequently the lobes bearing the stigmatic surfaces are 
naturally centred, as is particularly well seen, e. g., in the Wallflower 
(CheirantJms Cheiri , L.) (Fig. 40), over the two so-called commissures— 
now proved to be not true commissures, but consolidated carpels. 1 The 
few remaining genera fall under two heads : (a) those in which valve carpels 
as well as solid carpels are prolonged to form the style, and therefore pre¬ 
sumably form part of the capitate stigma, as in Biscutella (valves separating 
but indehiscent) (Fig. 37) and Guiraoa arvensis , Coss. (indehiscent) (Fig. 41) ; 
(1 b ) those with the style short or wanting, in which the valves, but little 
shorter than the solid carpels, bear the stigmatic papillae, as e. g. Matthiola 
incana , Moricandia , and possibly Loncophora. Hence, we need no longer 
subscribe to the time-hono ( ured morphological fiction of the commissural 
stigma. In place of it, we have clearly shown to exist a morphological 
dimorphism sufficiently elastic to permit of a considerable degree of physio¬ 
logical interplay without change of ground-plan. It was the serious 
difficulty involved in the acceptance of a genuine commissural stigma which 
led Lindley nearly a hundred years ago to suggest that the Cruciferous 
placentae are in reality independent carpels (see above, p. 455). He was 
led to this view through a comparison of the typical siliqua, with an ovary, 
with four stigmas of Eschscholzia californica , Cham., which latter type of 
gynoecium he conceived might arise if the structures represented by the 
Cruciferous valves were contracted to mere threads producing partially 
aborted stigmas, and if the placentae, on the other hand, were widened to 
become the placentiferous valves of Eschscholzia , bearing well-developed 
stigmas (Figs. 38, 39). Though Lindley thus arrived at a correct interpre¬ 
tation of the siliqua, his evident conception that similarity of function 
indicated morphological identity caused him to miss the true homologies, 
and to propound his solution in a way which introduced a new difficulty 
almost as serious as the one it removed, for his scheme involves a shift of 
the orientation through 90°. Moreover, the replum is not double, as he 
supposed. Hampered by this idea of morphological equivalence, although 
he considers the possibility that his conjectured four carpels in Eschscholzia 
might produce four placentae and thus account for the multiple rows of 
ovules in this genus, he inclines against this view. It must further be noted 
that these theoretical considerations were unsupported at the time by any 
actual evidence, and that they still left unexplained other apparent con¬ 
tradictions. Thus, the unsoundness of the premises caused the correctness 
of Lindley’s conclusion regarding the Cruciferae to remain unperceived, and 
botanical opinion has perforce acquiesced in the highly unsatisfactory 
alternative, the commissural stigma and the spurious partition. Only by 
1 It must be through some error that CheirantJms is instanced by Eichler as a case where the 
stigma lobes alternate with the placentae (Bliithendiagramme, ii, p. 204). 
