406 DE. M. T. MASTERS OlS* THE FLOBAL 



stamens, and carpels, while ni, n2, ns are in a line with the three 



petals and with the three inner stamens. jfi 



In fact, the median nerve !Ni, after leaving ni ni 



the ovarian wall, divides radially into two ^^ 



divisions, one of which passes into the median ^^ ^ 



sepal Si, where it soon breaks up laterally iu^sSraTthe 



into numerous subdivisions, while the other nerves at the base 



passes into the staminode Ax . ^^^^^ ^^^^^• 



The nerves N2, Ns pass into the corresponding lateral sepals 

 without dividing radially, so that the cords corresponding to the 

 stamens A2, A3 are, in this case, entirely wanting. The inter- 

 mediate nerves ni, n2 each divides radially, one division passing 

 into a lateral petal, the other into a fertile stamen. The median 

 nerve of this series, na, passes into the lip, Ps, where it speedily 

 trifurcates by lateral subdivision, but does not divide radially ; 

 for, as the inner median stamen as is not developed, there is no 

 vascular cord required for it *. The three cords for the gynsecial 

 row are numerically as complete above as they are below the 

 perianth, the primary nerves Ni, 1^2, Ns supplying the three 

 carpels Gi, G2, Gs. 



A difficulty, as before remarked, arises as to the large and 

 apparently single stigmatic lobe which is placed in the median 

 line. It might readily be assumed that the two lateral styles 

 were in this case absent, while the median one, Gi, was alone 

 present. But this inference does not accord with the position 

 of the placentas, two of which are placed laterally and one in the 

 median liae superposed to the lip. The placentas of course alter- 

 nate with the styles, the styles in fact being prolongations of the 

 dorsal sutures, while the placentas correspond to the united 

 margins of three carpels. An examination of the vascular 

 bundles in the style above the emergence of the stamens shows 

 the three bundles Ni, N2, N3 very distinctly alternating with the 

 3-lobed mass of conducting tissue continuous below with the 

 placentas. Into the stigmatic lobe itself, which is a triangular 

 cushion-shaped body with the apex of the triangle in the middle 

 line, it has appeared to me that the median bundle, Ni, does not 

 always enter, though it does so in some cases. The other two 

 bundles, N2, Ns, corresponding to G2, Gs, however, are always 



Darwin's view that the nerves corresponding to A2, A3 are in any degree 

 incorporated with the lip is, I believe, an error ; the nerves figured by him at 

 the side of the lip result merely from the lateral branching of the bundle nj. 



M- J ' \ 



