1914] CROCKER & DAVIS—DELAYED GERMIXATIOX 313 



28-30 mm., 1 120-1200 per cent; while in full air pressure, other 

 conditions the same, an elongation of 48-55 mm., or 1820-2200 

 per cent, occurs. In these conditions only the stored food of the 

 embryo is available for growth. It is evident that there is a much 

 more economical use of the food in the presence of oxygen than in 

 its absence, and still the growth is very considerable in its absence. 

 It is possible that the use of nutrient solutions would have induced 

 much greater growth, for it is certain that much leaching (35) of 

 the salts of the embryos must have occurred in these experiments. 

 Even in light, no leaf branches are ever formed in the seedling 

 grown in vacuo, nor does any chlorophyll develop. In weak diffuse 

 light at 2 5 C, the seedling did not develop chlorophyll even 

 after a month, if the air pressure was less than 5 mm. If the air 

 pressure exceeded 5 mm., chlorophyll generally developed in these 

 conditions. This shows that the oxygen pressure demanded for 

 greening is rather slight though easily measurable. Seedlings in 

 darkness under full air pressure all develop one or more leaves. 

 If the air pressure is reduced much below 5 cm., no leaf branches 

 develop even after two months. If the air pressure is more than 

 15 cm., there is an abundant formation of leaf branches. The 

 development of a primary root also seems to be dependent upon 

 oxygen pressure. Nearly every seedling growing in a drop culture 

 in a Van Tieghem cell shows the early development of a primary 

 root, as do seedlings growing on the water surface. Fewer seedlings 

 developing at the bottom of a considerable layer of water show 

 primary roots, and none in cultures with air pressure reduced 

 much below 5 cm. The growth of water seeds and seedlings in 

 vacuo stands in contrast to the facts found by Shull (30), Becker 

 (2), and At wood (1) for the seeds of land plants, where definite 



oxygen minima exist. 



IV. General discussion 



\\ ork to date has shown that delayed germination of seeds is 

 ured in a variety of ways: by almost absolute exclusion of water 

 seed coats, as in the " hard-seeded" legumes and species of 

 eral other families; by limiting the degree of swelling of the 

 bryo by surrounding structures, as in Alisma and probably 



