THE EVOLUTION OF LIVING BEINGS. 37 



these are invariably thrown by dominants and not 

 by recessives, which makes it very probable that 

 the dominants throwing them were impure, because the 

 fact that, ,,dominants" throw ^mutants" and recessives 

 do not, can also be expressed in this way : that, WHENE- 

 VER THE PURITY OF A FORM is VISIBLE (as in recessives) 



IN REGARD TO CERTAIN CHARACTERS, IT THROWS NO 

 ,,MUTANTS", INVOLVING THESE CHARACTERS, WHILE A 

 FORM, FROM WHICH IT CAN NOT BE DETERMINED AT 



SIGHT, WHETHER IT is PURE OR NOT (as in the case of a 

 mixture of pure dominants and impure dominants) 



MAY THROW ,,MUTANTS". 



The only legitimate conclusion which we can draw 

 from the regrettable fact, that it is so difficult to make 

 out with certainty that a certain form is specifically 

 pure in all respects, is that we must require the severest 

 possible test from him who wants to prove the exis- 

 tence of mutations. Knowing how difficult it is, to show 

 that a given form is free from recessives, we must dis- 

 qualify,apriori,all claims of having proved the existence 

 of mutations, based on the demonstration that a cer- 

 tain form has thrown recessives, no matter in how fee- 

 ble proportions. In the second place we must refuse to 

 accept as evidence all cases in which the reciprocal F! 

 generations, obtained by crossing the supposed mutant 

 with the form from which it arose, are either not-uni- 

 form or dissimilar. 



In the third place we must refuse to accept as evi- 

 dence all cases in which the numerical proportions in 

 the uncontaminated Fj generation of such a cross 

 deviate from the normal mendelian ones, or can only 



