247 



record the opinion, founded on repeated observations upon the early- 

 bud, that this, the simpler view, is sustained by the facts. 



Taking a raceme of Dicentra spectabilis, in which the lowest 

 flowers had fairly opened, while the upper were still minute buds, I 

 traced all gradations from the perfectly-developed stamens down 

 through earlier stages until the filaments were no longer visible. 

 The smaller flower-buds of a still younger raceme showed only 

 three small united excrescences before each of the outer petals ; 

 these were the only indications of the future stamens. I then placed 

 a flower-bud, about one-hundredth of an inch in diameter, on a glass 

 slide, and, after removing the sepals, subjected it to a magnify- 

 ing power of fifty diameters. The petals were too small to meet at 

 the top and cover in a central portion where the organs of genera- 

 tion were to be developed. Hence I was able to see the nascent 

 organs exactly in situ. The three excrescences on each side had 

 dwindled to a single, slight protuberance, and very close examination 

 was requii'ed to discover the lines at which the division into three sta- 

 mens was ultimately to begin. 



The relative size of the slight lobes of the protuberance was in 

 proportion to the future anthers, i. e. the middle lobe, which repre- 

 sented the two-celled anther, was fully twice as large as those on 

 either side of it which were to have but one cell each. Before the 

 inner petals there was no sign of any nascent stamen. Evidently, 

 the one-celled stamens do not originate there. Here, then, we have 

 the case obviously resolved into one of chorisis. The three stamens 

 occupy the place of one. The relative position of the stamens in 

 the full-grown flower is the one with which they start, and in which 

 they continue during the entire period of their growth. 



In the distribution of the fibro-vascular bundles going to the sta- 

 mens, everything points to the same conclusion. 



Divide the flower in halves at right angles to its flat surface, and 

 continue the division straight on through the middle of the pedicel ; 

 then, through the base of, and in the same plane, with the stamineal 

 phalanx, make a section thin enough to be transparent ; continue 

 this section until it runs out into the split surface of the pedicel. 

 Two primary fibro-vascular bundles will be seen going up through the 

 pedicel. After giving off" branches for the ovary, sepals and petals, 

 two large trunks run on into the middle stamen. The trunk on the 

 right sends off" a branch to the half-stamen on that side, and the half- 

 stamen on the other side receives a branch from the corresponding 

 main trunk. Stating the facts with direct reference to the mor- 

 phology of the stamens, we see no more primary divisions of fibro- 

 vascular bundles going to the phalanx than would go to one stamen. 

 The lateral stamens are supplied from the trunk which supplies the 



