468 SEMI-CENTENNIAL OF TORREY BOTANICAL CLUB 
quently smeared with a secretion which behaves under reagents 
much like certain globules visible inside the cell; the other cells 
show very scant cytoplasmic contents. Many of the hairs, how- 
ever, particularly toward the tip of the perianth segments, have 
one cell notably distinguished from the others, often by its swollen 
ellipsoid form, but always by its strongly refractive, colorless con- 
tents contained in the vacuole which practically fills the cell. To 
anticipate a possible suspicion that these hairs are abnormal, it 
may be remarked that they also occur on the perianth of Pon- 
tederia montevideensis and of Piaropus azureus. On the upper 
part of the stamen filaments are hairs of apparently similar char- 
acter, though consisting of only three cells, a basal cell set in the 
epidermis, a terminal globular secreting cell, and between them a 
barrel-shaped cell with colorless refractive contents (FIGS. I, 4). 
This cell is perhaps slightly more resistant to reagents than the 
swollen cell of the perianth hairs, but in both the presence of 
tannins is indicated, though probably not associated with exactly 
the same other substances that may be found in the hypodermal 
idioblasts. These peculiar stamen hairs are developed early 
(FIG. 5) and in buds only three millimeters long the barrel-shaped 
cell occasions difficulty in sectioning, much more than the hypo- 
dermal cells. More thickly sprinkled over the upper part of the 
filaments, and also the upper part of the long- and mid-length 
styles (FIGS. 9, 10) are simpler hairs, consisting only of the globular 
secreting cell and a basal cell. LeMaout and Decaisne* figure 
the short-styled pistil as fringed on one side with numerous spread- 
ing hairs, but I have always found it almost entirely devoid of 
such structures. 
Growing plants of the form known as Pontederia cordata lanci- 
- folia, collected 23 March 1918 near Tampa, Florida, by Professor 
. and Mrs. К. A. Harper, arrived in New York in good condition a 
week later, with only the flowers withered. These somewhat 
dried flowers, when soaked out in water, showed a blue color in 
the idioblasts of the hairs on both filaments and perianth; the 
subepidermal idioblasts in the upper part of the perianth also 
showed a fine deep prussian blue color, though all the anthocyanin 
had disappeared from the epidermal cells, except for the solid 
* LeMaout et Decaisne. Traité général de Botanique. 607. 1868. 
