STl'DIES OX THE C.EXETIC3 OF FLOWER-COLOURS, ETC. 95 



Wliiit iiiiikes tlie breediug of oiu" plaut very diffii-ult is the pjor germination 

 of seeds. Thuugli each pj'xis contains a large number of them many do not 

 come very often to germination ; fre€j[uently I have met ^^■ilh the cases where 

 even no one fi-om one pyxis had germinated, so that some exparimenfs 

 described in this jmper are based on ratlier few iudindnals. Many experi- 

 ments have been earned on to overcome tliis difficulty. Sinca, according to 

 my vievr, the poor germination of seeds wuiiLl be chiefly due to the difficulty 

 of water passage through seed-coats, one of my methods was to rub out seeds 

 h'ghtly with coarse quartz sand, so as to injure shghtly seed-coats and make 

 them easily permeable to water ; this ti-eatment has however given no good 

 results whatever. Then I have tried the method of high pi-essure adopted by DE 

 Vries :' I have used an autoclave-like apparatus specially made for the 

 pm-pose, and obtained the high pressure hy the use of an h'(jn receptacle 

 containing oxygen under high pressure (150 atmospheres when full) instead of an 

 automobile pump, because when the receptacle is connected -with the apparatus, 

 the high pressm-e is attainable instantaneously in the latter. Seeds soaked 

 in water dming one night and placed in it were subjected by tliis means to 

 the pressure of 8 atmospheres or less during 24 hom-s, and then sown as 

 usual. The following method was made use of in some cases : seeds were 

 first treated with 60%" or less concentrated sulphuric acid dming 30 minutes, 

 and iifter repeated washings, they were soaked in water during one night 

 and then sown. The two latter methods above described seem to promote 

 the germination to a certain degi'ee, but not so much as might be wished for. 



Seeds fi'om chiferent plants, and later those fi-om different flowers of one 

 individual were sown separately-. Since seeds of PoHulaca are very fine, if 

 we pom- down water to seed-pans fi-om above as usual, seeds might often be 

 thrown from one pan to anothei", thus causing the mixtm-e of seeds fi-om 

 different pai-ents. To avoid this possibility, no water has been given from 

 above : a number of seed-pans were placed in a rectangiilar vessel of wood, 

 180 cm. long to 90 cm. wide and 10 cm. deep, partly filled with water, so 

 that the latter may go up gi-adually. 



My study on some of the cix)3ses came already to a certain definite end, 

 and I like tf) publish here its results, because they will, as I think, enable 

 ' Bot. (Jaz. Vol. 59, 1915, pp. 192-193. 



