47 2 Kidd and West . — The Controlling 
While it is thus clear that no change under the influence of carbon dioxide 
has occurred in the mechanical resistance of the testa sufficient to prevent 
the embryo from swelling to its normal size, and while we must conclude 
that sufficient water is present for the growth of the radicle to proceed, a 
further possibility remains* An increase in the breaking strain of the testa 
during the primary period of inhibition by carbon dioxide, although 
insufficient to produce any measurable effect on the physical swelling of the 
embryo, may yet prevent the growth of the radicle and so account for the 
non-germination of the seed. 
Muller ( 6 ) draws a conclusion for certain seeds that the breaking strain 
of the testa is nominally only slightly less than the expansive force of the 
embryo, both lying in the region of three and a half atmospheres. The 
conclusions of Crocker (3) with regard to Amaranthus retroflexus may also 
Text-fig. 4. 
be cited in this connexion. The seeds of this plant will not germinate 
immediately after gathering, but they will do so after a month or more of 
dry storage. The reason for this is that the colloids of the seed-coats by 
a slow process of hysteresis in the dry condition take on new characters 
which give them a much lower elasticity and breaking strain. When the 
seeds are soaked after this change has occurred, they are able to germinate 
by rupture of the seed-coats. If, however, germination is now prevented by 
keeping the wet seeds at subminimal temperatures, the colloids of the seed- 
coats in their saturated condition recover their former strength by a reverse 
process of hysteresis. The embryo is then no longer able to rupture the 
seed-coat when the temperature is raised, and a prolonged period of 
secondary dormancy ensues. 
In view of these results of Crocker and of Muller, the following direct 
experiments were carried out to compare the breaking strain of the seed- 
coats of Brassica alba in the case of normally germinating seeds and in the 
■case of seeds showing secondary dormancy after C0 2 -treatment. 
The apparatus used is represented in Text-ftg. 4 . 
