igisl DE VRIES—OENOTHERA BIENNIS 193 



Before compressing the air in the seeds, these are thoroughly 

 soaked with water. Ordinarily they are exposed in small tubes, 

 half filled with water, to a temperature of about 30° C. during one 

 night. In the autoclave they remain from one to three days, at 

 the temperature of the room. My apparatus can accommodate 

 over 100 tubes at a time, each containing a different sample. After 

 leaving the reservoir, the seeds may be sown in seed pans for 

 cultivation or may be left to germinate in the same tubes, if it is 

 only intended to determine the proportion of seedhngs produced. 

 In order to do this the water is poured off through a small sieve, 

 the tube is closed by means of a cork, and the seeds are distributed 

 along the upper inner side of the tube, this lying horizontally. 

 In this way they get exactly the required amount of water and of 

 air for a vigorous germination. 



I will now give some figures to show the effect of this pump- 

 ing in of air into the previously soaked seeds. After pumping, 

 the degree of germination was determined by leaving the tubes 

 in a stove at 30° C. and counting the seedlings in samples of about 

 200 seeds each. Out of 18 capsules from self-fertilized flowers of 

 a spike of 0. Lamarckiana, 3400 seeds were counted, a separate 

 germinating tube being used for the contents of each fruit. Of 

 these seeds, 15 per cent germinated during the first two days and 

 only 3 per cent during the two following days, showing the normal 

 germination power to be almost exhausted. Then the seeds 

 remained three days in water under a pressure of 8 atmospheres, 

 after which they were brought back to the stove. The next two 

 days produced 22 per cent seedlings, and the four following ones 

 added only i per cent to this number. Then the remaining seeds 

 were tried with a needle. Only about 5 per cent contained 

 embryos, half of which at least were evidently in a decaying 

 condition. 



The total of germs was 46 per cent, leaving 54 per cent for those 

 with an undeveloped germ. From these figures we see that the 

 production of seedhngs from a sample of seeds may be more than 

 doubled by the pumping method, while all or almost all the healthy 

 germs may be made to germinate. Numerous similar instances 

 could be added. 



