ON CERTAIN EPIGYNOUS MONOCOTYLEDONS. 15? 



alternately in two rows, enclosed in large bracts. The individual 

 flowers again stand in two rows, as shown in the diagram, upon 

 their almost horizontal support. It will be noticed that the 

 flowers do not stand in the middle behind the next bract, but 

 midway in front of another, and thus behind a third, which, 

 however, is separated from the flower itself by that lateral one. I 

 see only one way of explaining this strange arrangement of the 

 flower. The branch, like the rachis, represents a distichous 

 spike, with alternate flowers closely crowded, but is modified by 

 the two rows of flowers having approached near together late- 

 rally, at the side of the main axis, and the margin of each bract 

 having been interposed between the flower at its side and the 

 bract belonging to the latter. Under this supposition we must 

 assume the suppression of the bract belonging to the flower which 

 stands next to the axis. 



The development of this flower takes place in the following 

 manner. The first sign of a bud, where it can be called inde- 

 pendent, is a perfectly simple, -fleshy nodule, the second stage of 

 which is exhibited by its nearest, outer neighbour ; this second 

 bud, namely, already appears somewhat flattened at the top, and 

 almost as if a little depressed or excavated. After this we gra* 

 dually detect the parts of the flower, which however can only be 

 recognized by their position, since they present themselves as 

 very inconsiderable elevations. These parts are developed in 

 the order from without inwards, in which they subsequently lie 

 upon one another, but it is remarkable here that the stamen 

 which is ultimately to be abortive makes its appearance before 

 the rest, that is to say, at the same time as the inner petals, for 

 at one epoch of the development four parts may be distinctly 

 observed in the interior of the flower. The five other stamens 

 do not appear until afterwards. At this epoch and for a little 

 longer, the abortive stamen has twice or thrice the magnitude of 

 the rest. At the outset of the development, when the various 

 parts of the flower rise out of the formative mass, they are com- 

 pletely unconnected, and only begin to exhibit cohesion at the 

 lower part when the abortive stamen has already been left con- 

 siderably behind. I may observe, that it is unnecessary here to 

 assume that originally free organs subsequently become blended 

 together, in the true sense of the word ; for, since the points of all 



