Sorting and blending of unit characters. 195 



Therefore in the summer of 1912 reciprocal crosses between Oe. 

 nutans and pycnocarpa were made. Seed of the two parents and of 

 the reciprocal crosses was sown in seed pans in March 1913. When 

 the seedlings each bore a few leaves they were transplanted to flats. 

 From here they were transplanted to the garden in May. The spring 

 and summer of 1913 was very dry, thus being unfavorable for the 

 Peronospora of which not a trace was seen. 



Of the 1913 cultures of Oe. pycnocarpa, in one lot (from a 7-car- 

 pled pod) less than 2*^/0 came into flower during the summer of 1913 

 (2 plants in about 140). These flowered in September, and were too 

 late to ripen seed before frost. The others formed autumnal rosettes. 

 In a second lot (from normal pods) all flowered as annuals, the con- 

 ditions of the seedling culture in flats probably being such as to fit 

 them for early maturity. About 5°/o (6 out of 120 plants) of Oe. nutans 

 flowered as annuals in time to ripen seed before frost. The remaining 

 ones formed autumnal rosettes. 



The hybrids of the Fi generation of the 1913 cultures showed a 

 higher percentage of annuals than did the parents. Of the annuals 

 there were two types easily distinguished by the inflorescence, foliage 

 and coloration. The two types were not distinguished until stem for- 

 mation was quite well advanced, when the coloration of the stem and 

 certain leaf characters were observed which indicated two forms. ^Tien 

 they came into flower the evidence of two hybrid types was very clear. 

 They began flowering early in Jul}^ Even at this period I was not 

 able at that time to sort out the remaining young rosettes. But as the 

 autumnal rosette leaves began to appear and attain maturity, there were 

 three hybrid types in the Fi generation. Two of these could readily be 

 assigned to one or the other of the types represented by the annual 

 flowering individuals. The third one proved to be a distinct type with 

 no annual flowering individuals (in the 1913 culture). 



The segregation of two or three distinct hybrid types in the first 

 generation of a cross is a rather unusual phenomenon. De Vries (1907, 

 1909, 1913) has applied the terms "twin hybrids" and "triplet hybrids" 

 to such Fl segregates^). 



*) Triplet hybrids (or as de Vries calls them "'Drillinge" (251, 1913), or "triple 



hybrids" (Bot. Gaz. 47, 1 — 8, 1909) are diflFerent from the "Tripelbastarde" of Lidforss 



(see Lidforss, B., Resume seiner Arbeiten über Rubus, Zeitschr. ind. Abstamm.- u. 



Vererb.-Lehre 12, 1 — 18, 1914; also Bot. Centralbl. 12(5, 290, 1914) who uses the term 



to denote a combination through crossing of three parents, according to the following 



scheme; Rubus af finis X {acuminatus X caesius), and its reciprocal. 



13* 



