Germination of Forest Trees 19 
of conclusions based upon one species of plant apparently may 
or may not hold for others. Shull’s investigations (14, 15, 18) 
are based mostly on Xanthium seeds. In his experiments he finds 
no evidences of the diffusion of oxygen through an absolutely dry 
seed coat. This is significant in that it shows an important role 
of water in preceding oxygen in penetrating seed coats. Ex- 
perimenting with Crataegus mollis Davis and Rose (16) find that 
seeds treated dry or those placed under water do not go through 
the process of after-ripening. Here again is evidence that both 
water and oxygen are necessary. These investigators, working 
on the effects of temperature upon the period of after-ripening, 
conclude that favorable moisture conditions and temperature con- 
ditions shorten the period. Atwood (19) confirms almost all of 
the conclusions drawn by Eckerson although working on Avena 
fatua. Crocker and Davis (20, 21) worked with water plants 
and their results totally different than those described for land 
plants need not be given. 
Unfortunately these conclusions are not based upon forest tree 
seeds. Such investigations have not been undertaken. This 
phenomenon will probably explain many of the cases of delayed 
germination which are well known to foresters. It is reasonable 
to assume that the conclusions based on Crataegus would also 
hold for such fatty seeds like the birches, spruces, hard maples, 
etc. It is also reasonable to suppose that most tree seeds pass 
through this period of after-ripening during the winter months ; 
if this is true it explains why it is often impossible to germinate 
certain tree seeds immediately after they have been gathered. 
Method of Attacking the Problem 
There are two general methods of determining the causes in- 
fluencing the behavior of seeds or plants growing under natural 
conditions. These are the observational and experimental 
methods. In the observational method we observe the kind of 
vegetation produced in response to a certain complex of physical 
factors and seek to find constant relations of one to the other in 
order to draw conclusions. In the experimental method we may 
