OTHER THEORIES OF SPECIES-FORMING. 347 



true for several of the other forms also. Only the two 

 species, 0. gigas and O. scintillans, appeared very rarely. 



"Thus de Vries had, in his seven generations, about fifty 

 thousand plants, and about eight hundred of these were 

 mutations. When the flowers of the new forms were arti- 

 ficially fertilised with pollen from the flowers on the same 

 plant, or of the same kind of plant, they gave rise to forms 

 like themselves, thus showing that they are true elementary 

 species.* It is also a point of some interest to observe that 

 all these forms differed from each other in a large number 

 of particulars. 



"Only one form, O. scintillans, that appeared eight times, 

 is not constant as are the other species. When self- 

 fertilised its seeds produce always three other forms, O. 

 scintillans, O. oblongata, and 0. lamarckiana. It differs in 

 this respect from all the other elementary species, which 

 mutate not more than once in ten thousand individuals. 



"From the seeds of one of the new forms, 0. Icevifolia, col- 

 lected in the field, plants were reared, some of which were 

 O. lamarckiana and others O. Icci'ifolia. They were allowed 

 to grow together, and their descendants gave rise to the 

 same forms found in the lamarckiana family, described 

 above, namely, O. lata, elliptica, nannclla, rubrinervis, and 

 also two new species, O. spatulata and leptocarpa. 



"In the lata family only female flowers are produced, and, 

 therefore, in order to obtain seeds they were fertilised with 

 pollen from other species. Here also appeared some of the 

 new species already mentioned, namely, albida, nannella, 

 lata, oblongata, rubrinervis, and also two new species, ellip- 

 tica and subovata. 



"De Vries also watched the field from which the original 

 forms were obtained, and found there many of the new 

 .species that appeared under cultivation. These were found, 



* 0. lata is always female, and cannot, therefore, be self-fertilised. 

 When crossed with O. lamarckiana there is produced fifteen to twenty 

 per cent, of pure lata individuals. 



