ON THE ORIGIN OF SPECIES. 491 



ble period is passed, and in that case retains its old characters. The 

 new species make their appearance in several, probably in numerous, 

 individuals, and during each year of the mutation period. If they do 

 not do this, their chances of life are exceedingly small, but in the 

 other case their chances are sufficient, even if the new species are not 

 in any regard superior to the mother species. The weaker ones among 

 the new forms disappear of course very early. 



The real struggle for life, in which natural selection must decide 

 whether the young types shall continue to exist or not, only comes later 

 on; it is not a war between species, but against other organisms, and 

 against climate and soil. 



In 1886, when I was preparing to write my Intracellular Pangene- 

 sis, the above mentioned considerations were only partly known to 

 me. De Bary's studies on Draba verna only appeared in 1889, Bate- 

 son's book in 1894, Scott's article a little later, etc. But what was 

 known at this time was sufficient to convince me that the formation 

 of species should lend itself to experimental investigation. This was 

 certainly directly opposed to the reigning opinion and especially the 

 conception of a slow and gradual origin was not in favor of my view. 

 It was thought that sudden transitions were limited to the so-called 

 varieties, that they occurred in agriculture and horticulture only, and 

 besides so rarely that an actual study of the problem was not to be 

 thought of. 



I then began a more systematic study of the variability of plants, 

 a subject which always had possessed a great attraction for me. It 

 very soon became apparent that observations in nature and in the 

 garden could not lead to the desired goal. Even if one pays constant 

 attention to the same individuals and the same localities, visiting them 

 in various seasons and in different years, the observations remain too 

 incomplete. This is but natural, since mutability commences with 

 the seed and in nature but comparatively little seed, after germinat- 

 ing, attains its full development. I therefore decided to have recourse 

 to sowing-experiments and for this purpose collected as much seed 

 as possible from wild growing plants. 



This seed was sown in my experimental garden, in some cases on 

 quite a large scale. Besides I sowed seed gathered from some speci- 

 mens of wild plants growing isolated in the garden. It was of course 

 my aim to try to find among them one or more species which were pass- 

 ing through a mutation period. Among the seed sown was for in- 

 stance that of Verbascum thapsiforme, Thrincia hirta, Crepis bien?iis, 

 Centaurea nigra, Capsella Bursa pastoris, Bidens cernua, Aster Tripo- 

 lium, Cynoglossum officinale, Sisymbrium Alliaria, Daucus Carota, and 

 a number of other wild plants. As far as possible I allowed myself 



