TWIN HYBRIDS OF OENOTHERA HOOKERI T. AND G. 



285 



the Hookeri-likv strongly resembled the two narrow-leaved plants, 

 described above. There were 53 percent Hookeri-Wke. Among the 

 others I found three plants, which unfolded the characters described 

 elsewhere for the mut. lutescens (de Vries 1916 a, p. 7) and a slender 

 specimen with smaller leaves. After self-fertilization each of them 

 yielded a uniform progeny, which embraced 60, 57 and 39 plants 

 for the three lutescens and 39 for the deviating type. This latter was 

 a strikingly different combination of characters, having low, slender 

 and richly branched stems with smooth small leaves, slender spikes 

 and narrow flower buds of a pure green, large petals (3 cm.) and the 

 stigmata high above the anthers. The leaves reached in every direc- 

 tion about half the size of those of the rubiennis, the stems about 

 half their height. There was almost no individual fluctuation in this 

 group. 



In 1915 1 had self- fertilized 8 specimens of rubiennis. Their pro- 

 geny embraced about 45—65 plants for each of them and consisted 

 in almost every case of three types, viz., rubiennis, Hookeri-Uke and 

 tutescens. 



I found in August 1916 the following figures: 



Table 13 



This gives an average percentage figure for the Hooker /-like of 31 

 and for lutescens of 8. Besides these types there were found in No. 

 7 eleven plants of a new type (not counted in our table). This was a 

 pure aurea, with broad golden leaves sharply contrasted with the 

 remainder, even with the pale lutescens. They were transplanted in 

 June to another bed, where eight of them flowered. The golden 

 color diminished here, however, and the leaves became more mar- 

 kedly green, insuring the nourishment of fruits and seeds. But the 

 type remained an obviously new one. Its appearance in about 18 

 percent of the whole culture evidently points to a Mendelian recessive 



