Tricotylous Half Races. 381 



Very small proportions of tricotyls were yielded by 

 Silene hirsnta, which only produced 3 in 80,000 seed- 

 lings. The following species produced from 1-2 speci- 

 mens in every 10,000 seedlings: Argemone grandiflora, 

 Aster tenelhis, Clarkia elcgans, Godctia amoena, Hyos- 

 cyarnus pic t us, Silene Aruieria, and others. I observed 

 no tricotyls at all in sowings of the same extent of 

 Argemone mexicana, Datnra laevis, Hyoscyanius albus, 

 Nigella damascenaj Phacelia texana etc. I tested 800 

 seeds of each of 8 species of Cerinfhe, and only obtained 

 a single tricotylous plant in C. bicolor, C. gymnandra 

 and C. major. 



The seeds mentioned were all obtained from the 

 nursery of Messrs. Haage & Schmidt in Erfurt. It is 

 not unlikely that if the seeds were bought from other 

 nurseries, different results could be obtained, especially 

 from firms who do not exchange seed with the nurseries 

 at Erfurt. 



Similarly the seeds of wild species occasionally pro- 

 duce tricotyls, but, as it seems, only in very small quan- 

 tities, and they have hitherto given no promise of yield- 

 ing an intermediate race. As instances I mention Ra- 

 phanus Raplianisfrnni and Epilobinm hirsiitum, of which 

 species I found 1-2 tricotyls in large crops. If the seeds 

 of wild species come from botanical gardens the pro- 

 portion of tricotyls is sometimes greater; for instance 

 in Silene nocti flora (1892) it was about 20 in 10,000 

 seedlings. Amongst trees I have hitherto found tri- 

 cotylous seedlings abundantly in Acer Pseudo-Plafanus 

 and Fagns sylvatica and also in Robin i a Pseud-Acacia 

 and Ulmus campestris. 



In order to test the hereditary capacity of the cotyle- 

 don variants in such crops, I have frequently planted 



