618 BOWMAN— ECOLOGY AND 



tion of the chambers and later of the pollen from the very young 

 parenchymatic tissue. These are first that the two pollen sacs fuse in 

 the upper part of the anther where there is no bilateral arrangement 

 by a median line, but a line of chambers occurs in the middle plane 

 of the apex; the second is that not all the cells of the young anther 

 parenchyma or endothecium become sporogenous as they do in 

 other anthers, but some cells become the alveolar walls. Warming 

 further remarks that in his opinion this is not an old phylogenetic 

 condition but a recent adaptation and is seen in not only the man- 

 groves, Rhizophora, JEgiceras, etc., but in other families as the 

 Onagracese above mentioned and the Orchidacese (Phams and 

 Bleteia, etc.), as well as Viscum of the Loranthacese. 



The mechanism of the dehiscence, however, is just as interesting 

 as the formation of these peculiar anthers and their pollen. This 

 feature was brought out in examination of the cellular structure of 

 the bud. As the other workers on the species have shown, the 

 anthers in transverse section are triangular, or obovate-triangular 

 with the dehiscing faces introrse and the back or outer side of the 

 anther is a broad expansion of the connective (Fig. 2, PL VI.). 

 This connective area, as well as the partitions of the pollen loculi 

 contain a peculiar kind of cell. All the previous investigations 

 have overlooked these cells. They happened to be brought out in 

 sections which had been double-stained in safranin and methyl- 

 green to contrast the lignified walls of the idioblasts. While exam- 

 ining these sections there was noticed in the outer cells of the con- 

 nective area of the anther a layer of cells which contained peculiar 

 lignified, transverse ring thickenings inside the cellulose wall (see 

 Figs. 3 and 4, PI. VI.). In these anthers this reinforced area ex- 

 tends clear to the tip and the cells composing it are rather elongated. 

 According to our interpretation, these cells play an important me- 

 chanical part in the dehiscence of the pollen. As the pollen ripens 

 in the loculi, the thin exothecium shrinks and while this is taking 

 place the strain produced on this thin-walled cell layer, particularly 

 along the middle line of the pollen loculi, by the rigidity of the areas 

 composed of the reinforced cells, a rupture occurs at the weakest 

 places, i. c, at the middle line where the partitions are thinnest. 

 When the split has occurred all along the line the exothecium falls 



