The Periodicity of Progressiz'c Mi(tatiu)is. 653 



our species, have arisen suddenly, these changes must 

 have been distributed more or less regularly over tiie 

 whole line of ancestors of the Oenothera. \h)\\ many 

 steps are combined into a single period of mutation can- 

 not he determined, and the question is obviously of sec- 

 ondary interest only. The available evidence seems to 

 indicate that only one step in the same direction occurs 

 at one time ; but obviously this does not exclude the possi- 

 bility of periods in which more numerous changes occur. 



In order to apply the results obtained with our prim- 

 roses to earlier hypothetical periods of mutation, I will 

 repeat the empirical pedigree of the first volume (p. 224), 

 but in a somewhat different form. I will indicate the 

 lateral branches which arise from the main stem in suc- 

 cessive years, that is to say the new species, in the form 

 of radiating groups (Fig. 148). Each group denotes the 

 mutations in a single generation. The main stem contin- 

 ues unchanged and successively produces the individual 

 groups. Together, however, they obviously belong to one 

 and the same period, inasmuch as each of them mainly 

 consists of the same species and in approximately ecjual 

 proportions. 



In order to compare this period with previous ones 

 the whole figure may be compressed to a single group. 

 This has been done in the upper part of Fig. 14^. The 

 lateral branches do not arise here from a single point, and 

 this is intended to indicate the fact that the figure em- 

 braces a series of generations in which the variations were 

 repeated. 



As stated above, we will now assume that the ances- 

 tors of our Oenothera have not always been mutable. 

 Therefore our group must have a limit below, and nuist, 

 so to speak, be borne by a stem without lateral brandies. 



