102 THE JOURNAL OF BOTANT 



the flower is supposed to be held upright, as in Midler's figures. 

 This, of course, is not the natural position of the flower, which is 

 horizontal or drooping, but it is convenient for description. 



The stigma in latifolia is vertical, almost parallel with the axis of 

 the flower, facing forwards, its lower lip slightly turned upwards, and 

 is somewhat concave. On looking into a flower it is at right angles 

 to the line of vision, and its quadrangular shape is well seen. In 

 Midler's plant (Epipactis) it is horizontal, at right angles to the axis 

 of the flower, the upper portion thrust well beneath the base of the 

 anther, nearly reaching the back wall of the column, the lower part curv- 

 ing downwards ; it faces upwards, and is convex. In front view it is 

 parallel with the line of vision, and appears crescent-shaped, as 

 shown in Midler's figure. 



The rostellum in latifolia is large. Midler says that in his plant, 

 there is no rostellum. At first I thought this was the case with the 

 Thorenc plant, but later I found that if the anther is turned back on 

 its hinge, there is, at any rate in bud, a glistening conical projection, 

 which has the appearance of a not fully developed rostellum, in the 

 middle of the upper edge of the stigma ; if touched, weak strings of 

 viscid matter adhere to the touching object. It is entirely concealed 

 beneath the anther, for the stalk of which I at first mistook it. 

 Miiller's fig. 11 shows a rounded projection in exactly the same 

 position ; perhaps he failed to recognize it as a rudimentary rostellum, 

 for it may have been less developed in the Westphalian plant. 



A clinandrum, or shallow cup, occupies the summit of the 

 column in latifolia, bounded on the back and sides by the walls 

 of the column, and in front by the back of the stigma. The anther 

 leans forward over this cup, and when its cells open, the pollinia are 

 deposited in it, lying on their sides. In Miiller's plant there is no 

 clinandrum, the corresponding space being roofed over by the stigma. 

 The anther, in latifolia, seen from the side, is narrow at the base, 

 broad and truncate at the apex, appearing rhomboidal ; it is hinged to 

 the bade wall of the column, and its base is separated from the 

 stigma by the whole diameter of the clinandrum. In Miiller's plant, 

 however, it is shaped like a cap of liberty, its hinge is in front 

 of the back wall of the column, and its semi-circular base projects 

 well over the upper part of the stigma. On dissecting a 'very 

 young bud, expecting to find the pollinia in situ within their cells, I 

 found to my surprise, that they were already outside the anther 

 and enclosed in an extremely thiri tight-fitting and verv glossy 

 membrane. They were so much larger than the opening's of the 

 anther-cells that it was clear that they could not have passed through 

 them except at a still earlier period, when they were much smaller. 

 The membrane appeared to be continuous with the cell- walls. 

 Looked at from beneath, there was a funnel-shaped channel between 

 the anther and the pollinia. 



The pollinia in latifolia are club-shaped, their thick ends rounded, 

 their thin ends pointing forwards, firmly attached to the rostellum! 

 In Miiller's plant they are smaller, shaped like a sugar-loaf, with 

 flat circular bases and curved sides tapering to an obtuse point. 

 Even in bud they stand upright on their bases on the front surface 



