No. 443] MUr.lHiKX l.\ I'LAMS. ybg 



may be a hybrid plant, whilst the mutations obtained by deXries 

 may be merely partial or complete reversions to the orij^inal 

 ancestors ot the plant."' It is quite possible, and even probable 

 th It CE. LunAixkiaw.x may have been ori^inall) derived from the 

 same type as iE. biennis as noted above, but to designate it as 

 a "garden variety." and as such ineli^nble as research material 

 is simple evasion. The plant in question has been under more 

 or less continuous observation tor a lumdietl and fifteen years 

 during which period it has been constant in its characters, and 

 has shown no evidence b\- anatomical similarity or physiological 

 behavior of being anything but an indej^endent species. With 

 what species could biinuis hybridize t(» jDroduce lamarckianal 

 The genus comprises a comparatively small number of types, all 

 natives of America, and none of which were available as a 

 hybrid mate to biennis at the time of the origin of lamarckiana. 

 The conjecture in question is totally unsupported after the most 

 rigid searcii for evidence upon the matter. 



Again to consider the mutants as reversions to the original 

 ancestors of lamarekiana is impossible, since the mutant forms 

 exhibit qualities not possessed by any other known members of 

 the genus, including biennis. 



Tin- pninl raised bv Bateson and Saunders (Reports to the 

 Kvolution Committee. Royal Society. I. p. 153, London. 

 1902) that the pollen of lamarckiana izon\.2i\x\':-, deformed grains, 

 which points to its origin by crossing, is without significance. 

 since the author has found that the stamens t)f plants of biennis 

 growing in the vicinity of New York exhibit a much larger pro- 

 portion of deformed pollen than that of the specimens of lamar- 

 ekiana cultivated in the New York Botanical Garden. 



It has been impossible so far to assign mutations to definite 

 causes, or to forecast the frequency, or occurrence of the phe- 

 nomenon. 'Ihese phases of the subject constitute the most 

 important problems of the subject, which await investigation. 

 Theoretical evidence upon such a subject can have but limited 

 value, and conclusions of any satisfactory degree of finality may 

 be expected only from direct experimental research under cir- 

 cumstances in which the probabilit\ of error is reduced to a 

 minimum. 



