454 Tricoiylous Races. 



fJicra {Kuciffia) glaiica in 1899, Ijut counted 5% tri- 

 cotyls, both on tlie phosphate and on the nitrate half. 

 I obtained the same result in ScropJiularia nodosa by 

 growing one half in ordinary garden soil and the other 

 in sand, both yielding 1% (1894). 



The two first experiments on the inliuence of manure 

 show that an increase in the number of tricotyls is as- 

 sociated with a diminution of the yield. The same result 

 can also be obtained if the harvest of the individual plants 

 is reduced in other ways. In Oenothera hirtella I adopted 

 two methods of doing this, late sowing and culture in 

 pots throughout the whole summer. Tricotylous seed- 

 lings of a single parent (with 66%) were planted out. 

 In one of the two groups the sowing had taken place in 

 March, in the other at the beginning of May (1898). 

 Some plants of the former sowing were kept in pots of 

 15 centimeters filled with well-manured garden soil 

 throughout the whole summer. Fertilization took place 

 in parchment bags and the individual crops, recorded and 

 evaluated in the usual way, gave the following result : 



HEREDITARY VALUES IN OENOTHERA HIRTELLA IN 1898. 



Percentage values 10 20 30 40 50 60 70 Mean 



Normal culture 

 After sowing in May 

 After culture in pots 



Number of plants experimented with 18, 36 and 19. 



We see therefore in this case a definite though small 

 increase in the number of tricotyls. The mean of the 

 harvest for the normal culture was Z.'b cc, for the late 

 culture 2.5, for the pot-culture 4.5 cc. This latter had 

 therefore not had the expected result of making the 

 yield smaller, and as, nevertheless, the proportion of 



