286 BOTANICAL GAZETTE [ocroBER 
PFEFFER (9) says: “The conditions which lead to certain seeds 
resting under the soil for as long as fifty years and germinating when 
dug up have not as yet been determined.’ This, as well as the 
sudden appearance of weeds in forests after fires and in meadows 
of many years standing immediately upon plowing, is probably 
explained by a few simple facts. Weed seeds are produced in great 
abundance. Because of variation in seed coat characters or in some 
cases of embryo characters, a given crop distributes its germination 
over a period of years. Seeds deep in the soil germinate less readily 
because of lack of oxygen or water, and those that do grow exhaust 
the stored food before reaching the surface. Bringing such seeds 
to the surface greatly increases their germination and removes the 
danger of exhaustion of the stored food. The plants of meadows and 
forests keep the water supply reduced and thereby cut down the 
chances for the germination and later growth of the weed seeds present. 
With the destruction of the plants of the forest or meadow comes 4 
great increase in the germination of the weed seeds and a removal of 
the opposition to their future growth. These phenomena, then, will 
probably all be explained by a study of the germinative characters 
of the seeds such as is described in the experimental portion of this 
paper, along with certain other well established facts on germination. 
It is undoubtedly true that many of the tests that have been made 
for the vitality of weed seeds are untrustworthy, because the signifi- 
cance of the seed coats has been overlooked. This is clearly shown 
by the results of Duvet (3b) and WALDRON (12), who have carried 
on extensive experiments to determine the length of time weed seeds 
must be buried in order to lose their vitality, In column I of 
Table X is shown the percentage germination determined by 
me with seed coats broken and with favorable temperatures. In 
column II appear Duvet’s results, in which he uses what he terms 
the “most favorable temperatures,” but overlooks of course the seed 
coat effects, The figures quoted from Duvet are from the column 
“original samples,” which means fresh seeds, as were the seeds for 
determining the percentages of column I. 
The effect of rupturing the seed coats, as is shown in this table, 
is very evident, although Duvet has in part overcome the seed coat 
effects by high temperatures. The average percentage of germina- 
