# 



^6 DR. O. DICKIE ON THE GEKUS CAXXA. 



when tlie pollen is nearly mature; and fig. 2 shows the exter- 

 nal appearance (front view) about the same period. 



At the time of fecundation, the face of the anther and that of 

 the flattened style are in close contact ; and the pollen is shed 

 before the flower opens, and may be seen adhering to the upper 

 part of that organ : it is, in fact, swept out of the anther by the 

 stigma ; the face of the anther and upper part of style and stigma 

 are at first firmly braced together. As already stated, close in- 

 spection is necessary in order to see that the shrivelled staiheu 

 in the newly-expanded flower retains, though rather indistinctly^ 

 the two-celled structure, which is obvious enough at an earlier 

 period 



The next question has reference to the nature of the petaloid 

 organ to which the stamen adheres. The relation of the two, at 

 an early stage, is seen in fig. 2, from w^hich it is evident that, 

 while the anther is very fully developed, the appendage is rudi- 

 mentary. But, further, an important point to be noted is this : 

 tlie petaloid part adheres by one edge to the back of the anther, 

 along the line which corresponds to the connective ; the other 

 margin is free at the upper part, below it is partially adherent 

 to the style, the entire appendage embracing rather more than 

 one-half of the anther and part of the style, its function, ap- 

 parently, being to keep those parts in close contact when the 

 pollen is shed ; fig. 3 represents a plan, in ti^sverse section, 

 illustrative of this. There may be several conjectures as to the 

 nature of this petaloid appendage. First, it may be a portion of the 

 anther, some holding the upper part to be the functionless cell, 

 become petaloid, as well as the filament ; its coexistence with two 

 cells in the anther shows that this view must be abandoned: 

 or, secondly, it may be a wing-like appendage of the back of the 

 anther, which is not probable j thirdly, it may be considered to 

 represent one of the stamens of the same series as the fertile 

 one, adherent to this latter and petaloid. A. Kichard, while he 

 regarded the anther as one-celled, stated, moreover, that the pe- 

 taloid filament ought to be held as formed by the " union of two 

 stamens, one producing pollen, the other abortive, which is re- 

 presented by the petaloid lamina, on one of the sides of which 

 the anther is inserted." 



This, undoubtedly, is the correct view, and appears to ve 

 indicated by the distribution of the vasciJar bundles, althougn> 



* In the 'Botanical Magazine,' under Canna lutea, fig. 1085, there are re- 

 marks which, in part at least, are near the truth. 



