474 SEMI-CENTENNIAL OF ToRREY BorANICAL CLUB 
The six legitimate crosses which may take place between the 
six sets of stamens and the three different pistils are indicated by 
the dotted lines and arrows in FIGURE 11. | 
The pollen grains from the different stamens show differences 
in size similar to those reported by Darwin for the water-hyacinths 
and for the Pontederia found by Fritz Müller in the interior of 
Brazil; those of the two longest sets of stamens from mid-and short- 
styled flowers are largest, indicating that they are adapted to 
pollinating the long-styled pistil; those of the very short stamens 
from long- and mid-styled flowers appear about a quarter as large 
Long-styled HMid-styleà Short-styled 
mid-stamens longest stamens longest stamens 
Shortest stamens shortest stamens mid-stamens 
О e. c 
Fic. 12. Pollen grains from the three flower forms, to show comparative size, 
and similarity of those taken from stamens of similar length in the different forms 
jJ. 
( X 200 
(their actual volume averages only one seventh as great) indi- 
cating that they would pollinate only the short-styled pistil; while 
those of the mid-length stamens of the long- and short-styled 
flowers are intermediate in size, indicating adaptation to polli- 
nating the mid-styled flower (FIG. 12). Unfortunately for con- 
venience of exact comparison, the pollen grains of our species 
usually are not spherical, as intimated by Darwin for the species 
studied by him, but they are here rather ellipsoidal or lemon- 
shaped; in only ten out of three hundred measured have I found 
isodiametric grains. 
In order to obtain as definite a record as possible, five flowers 
of each form were selected, each flower from a different spike, 
