1913] MARTIN—POLLEN OF TRIFOLIUM PRATENSE 119g 
0.731 and 0.8772 volume-normal solutions of sucrose or in 1. 3888 
and 1.666 volume-normal solutions of levulose and properly dried 
gave as good results as those soaked in distilled water. This 
shows that these sugars have no toxic effect on the pollen. The 
efficiency of the membrane did not depend upon the fats or salts 
contained, for pieces extracted 5 days in alcohol and ether in a 
Soxhlet’s extraction apparatus or boiled for 16 hours in changes of 
distilled water did not lose their efficiency, although their physical 
properties were so changed that the requisite amount of soaking 
and drying had to be found again by experimentation. Another 
series of trials with parchment paper showed it to be as effective 
as the bladder and that the previous failure was due to insufficient 
drying. In these tests most attention was given to the pollen of 
T. pratense, although the pollen of T. hybridum was investigated 
sufficiently to discover that it germinated readily on the membrane 
and that its germination would permit more variation in the water 
content of the membrane than the pollen of T. pratense. 
The nature of the germination of the pollen of T. pratense on 
the membrane needs some discussion. Germination was not uni- 
form. On some portions of the membrane the percentage of 
germination was high, while on other portions there was no ger- 
mination. These different regions were usually quite definitely 
marked off, and the germination in a region was usually good or 
none at all. This lack of uniformity was mainly due to a difference 
of texture, composition, or thickness of these regions. A difference 
in these properties would make a difference in the amount of water 
supplied to the pollen in the different regions. Some variation no 
doubt exists between pollen from different anthers and between 
pollen grains from the same anther in respect to the water supply 
requisite for germination. But in mounting the pollen, the keel 
was sprung with a scalpel, and as the pollen was thrown from the 
anthers, it was collected on the instrument and then spread on the 
membrane. By this method of collecting and mounting, the pollen 
was well mixed, and the variation of the pollen would not account 
for all of the lack of uniformity in germination. 
The percentage of germination determined by taking into ac- 
count all the pollen on a membrane when any germination occurred 
