274 The Pedigree Families. 



Let us now examine the origin of the various new 

 species in this family a httle more closely. We will begin 

 with the 1889 crop. It comprised about 400 specimens of 

 Lamar ckiana and two rosettes and one annual plant of 

 O. lata. But as I already had this species under cultiva- 

 tion I did not save them. Besides this, there were two ro- 

 settes of O. elliptica and two others of a new form, O. 

 spathiilata, which however I did not succeed in winter- 

 ing. Dwarfs were fairly plentiful in this crop; two of 

 them developed stems but did not succeed in flowering; 

 ten remained in the rosette stage but only two of them sur- 

 vived the winter. These formed the stock of -a nanella- 

 family of which I have cultivated five generations, which 

 will be described later (§ 18). One of the dwarfs that 

 appeared in 1889 had the narrow leaves of 0. elliptica 

 but was a nanclla in every other respect. It remained in 

 the rosette stage but died during the winter. 



Besides the mutations we have already named there 

 occurred in the 1889 crop two examples of an entirely 

 new form, 0. ruhrinervis. One of them stayed in the ro- 

 sette stage and could not be wintered ; the other devel- 

 oped a fine stem very early, bore a profusion of flowers 

 which were however not enclosed in bags, and set a 

 quantity of seed which was harvested on the 8th of Oc- 

 tober 1889. 



These seeds were sown on the 5th of May 1890 on 

 a small bed. The red-veined individuals were distinctly 

 recognizable very soon after the seeds came up, and the 

 rest were destroyed. In September I had forty speci- 

 mens of ruhrinervis of which nine flovv^ered and set seed. 

 The rest were either annual but not adult, or rosettes. 



The seed saved in 1890 was used in 1891 partlv for 

 a culture on almost barren sand and partly for a control 



i 



