HAZEN: TRIMORPHISM AND INSECT VISITORS OF PONTEDERIA 475 
except in the case of the short-styled form, of which only three 
spikes had been preserved, and from each flower five pollen grains 
from the longest stamen and five from the shortest were selected 
at random and measured by means of the eyepiece micrometer. 
The twenty-five measurements from each of the six stamen types 
were then averaged. Some weeks later a duplicate set of twenty- 
five measurements was made in the same manner, and while they 
averaged uniformly slightly greater than those of the first set, the 
difference was only such as might be expected because of the 
personal equation, though in each of the three forms of the second 
set it was noticeable that one or two of the flowers were especially 
vigorous, as shown by a considerable increase in size of all grains, 
both larger and smaller, above the average for similar anthers in 
other flowers. Nevertheless, when the second set of measurements 
was combined with the first, the ratios between the different sized 
grains diverged from those obtained from the first set by only a 
negligible amount, indicating the substantial reliability of the 
work. The fifty measurements of each type of pollen give the 
following result: 
? Long гер ps from median пас stamens. oousi 37-44 X 33-79 microns 
from shortest stamens.. -<< 0555... 23.04 X 21.69 
ima e кш from longest stamens.......... EAD ад 46.33 X 41.61 47 
m shortest stamengé аеро, 23.95 X 20.95 
Shore syed то. from longest stamens. ооу. 45.84 X 41.32 
rom median mid-length stamens........ 36.94 X 33.01 
It wil! be seen that the pollen grains from the two sets of longest 
and shortest stamens correspond most closely, the mean diameters 
of the shortest sets differing by less than half a micron, but 
even the divergence between the mean diameters of the two 
mid-length sets amounted to only seven tenths of a micron. 
There is a much closer correspondence here than in the 
smaller number of measurements made by Francis Darwin 
for his father on the Brazilian plant. It does not appear 
whether Darwin selected pollen grains from more than one 
anther of each type, and there is considerable variation in different 
plants. The extremes found in Pontederia cordata may be of 
interest. The grains from the median (i. e., longest) mid-length 
stamen of the long-styled flowers showed such averages in dif- 
