742 



THE AMERICAN NATURALIST. [Vol. XXXVII. 



within the last three years and it will not be profitable to 

 rehearse the details at this time. Briefly stated deVries's inves- 

 tigations may be embodied in the following paragraphs. 



1. Observations were chiefly concerned with a large num- 

 ber of plants growing wild and under cultivation, of the type of 

 CEjiotlicra lamarckiana. The identity of the parent form was 

 found by comparison with the original description of the plant 

 made a century earlier, and by comparison with a type specimen 

 in the Museum d'Histoire Naturehe in Paris collected in 1788. 

 The actual name of this plant in the revised nomenclature is a 

 matter of minor importance in the present connection. 



2. raim^-:rs of individuals of the parent type, as a result of 

 cross- and self-pollination indifferently, constructed seeds which 

 developed into independent forms, constant and self-maintenant, 

 which differed in habit, structure, stature, appearance and prop- 

 erties from the parent type. 



3. The aberrant or mutant forms might be divided by 

 characters as sharp and numerous as most of the so-called 

 minor species of the systematist. 



4. No forms intermediate between the mutants, or between 

 the mutants and the parent type were found. 



5. That the mutant forms were really groups of phylogenetic 

 value was proven by their behavior when crossed with one another, 

 with the parent form, and with other species in the same genus. 

 The hybridization experiments with these forms has yielded 

 some exact evidence as to the preponderance of phylogenetically 

 older characters by reason of the fact that the mutants are 

 forms the exact ages of which are known. Of the crosses of 

 (E. lata and ffi. nanella with the parent form, from a half to 

 three-fourths were found to be of the parent type, and the 

 remainder of the mutant type form. The crossing of mutants 

 with each other produces a generation many of which show 

 reversionary characters. The mutation hybrids are constant in 

 succeeding generations. The separation of antagonistic charac- 

 ters in the first generations is weighty evidence in support of 

 the theory of elementary characters, and for the mutation theory. 



6. The new types were either constant from the beginning, 

 or if weak, inconstant or perishing, showed no tendency to revert 



