and the decree of correlation is much less in the mutants tlian 

 in the i)arent. This is seen by inspection to be true in one 

 species during the first year of its existence, and is confirmed by 

 the exact observations on other forms a dozen years after their 

 mutative origin. Consequently the features in question may not 

 be taken to be in any way the result of selection but are in them- 

 selves new qualities. 



Lamarck's evenin<^-primrose offers such striking and easily 

 recognizable examples of discontinuous variation, and has been 

 the object of so much detailed study that we are in danger of 

 giving way to the supposition that the mutation -theory rests 

 upon the facts obtained from this plant alone. It is to be said, 

 however, that if it and all of its derivatives were destroyed, the 

 results of experimental studies which have been made upon 

 mutations in other species, upon the behavior of retrograde and 

 ever-sporting varieties, the occurrence of systematic atavism, antl 

 of taxonomic anomalies, pelories and other morphological fea- 

 tures would furnish ample support for the conception of unit- 

 characters, and serve to establish the fact that mutations have 

 occurred in a number of species representing diverse groups. 



It is now becoming plainly apparent that the phenomena of 

 hybridization, by the opportunities afforded for the study of the 

 included unit-characters in a segregated condition, for the anal- 

 ysis of complex characters, and of the various principles gov- 

 erning the transmission, activity, dominancy, latency and reccs- 

 sivity of characters, promise to yield results of the first magnitude 

 concerning the mechanism of descent and heredity. The pos- 

 sibilities of crosses between species comparatively widely different 

 in morphological and physiological constitution among plants 

 indicate that the ultimate generalizations upon hybridism will 

 find broader exemplification in plants than in animals. 



It is pertinent to point out in this connection that the un- 

 guarded use of the terms " variation " and " mutation " to desig- 

 nate phenomena of segregation and alternative inheritance when 

 races or species are thrown together in a hybrid strain is bound 

 to result in much confusion, especially in dealing with plants, 

 since it is well known that direct mutants of either parent occa- 

 sionally occur in such mixed strains. 



