450 



SCIENCE. 



[N. S. Vol. XVII. No. 429. 



have been found by various workers in dif- 

 ferent fields, -in this country notably by 

 Spillman in hybrid wheats. Mendel was 

 able to discover the numerical law because 

 he eliminated nearly all of the confusing 

 contingencies. In the discussion of every 

 bold new hypothesis, we are in danger of 

 becoming partisans, taking a stand either 

 for it or against it. The judicial attitude 

 is also the scientific one. We want to 

 know. 



Two processes are now going forward in 

 the discussion of Mendel's law— one the 

 explaining away of 'exceptions,' the other 

 the endeavoring to find the true place of 

 the law in the scheme of evolution. The 

 one is primarily an effort to uphold the law ; 

 the other is primarily a desire to adjudge 

 it. One is an effort to apply it universally ; 

 the other to determine whether it is uni- 

 versal. Already so many adjustments 

 have been made of the Mendelian prin- 

 ciples that it is becoming difficult to de- 

 termine what Mendelism is. These cases 

 are typical of the discussions on almost 

 every vital question connected with evolu- 

 tion. At the hard places we make a sup- 

 position and modify the hypothesis in the 

 face of a fact. We can prove anything 

 by supposing. 



The results of Mendel's work have two 

 important bearings on current evolution dis- 

 cussion: (1) on the part that hybridization 

 plays under natural conditions in the evo- 

 lution of the forms of life, and (2) the part 

 that it plays in plant-breeding. In the 

 former category, Mendel's work gives a 

 hint of definiteness to the role of hybrid- 

 ization in the origination of new combina- 

 tion-forms. In the latter category, it is 

 difficult as yet to measure its importance, 

 since extended applications to practice have 

 not been made and since, also, the Men- 

 delian principles have been so much ex- 

 tended and redefined within the past two 

 years that it is difficult to determine just 



what is Mendelism and what is an endeavor 

 to make the Mendelian suggestions fit our 

 present-day knowledge. In discussing the 

 application of Mendel's work to plant- 

 breeding, I desire to keep in mind the work 

 that he did with peas, upon which' the 

 'Mendel law' chiefly I'.ests. 



III. APPLICATION TO PLANT-BEEEDING. 



The wildest prophecies have been made 

 in respect to the application of Mendel's 

 law to the practice of plant-breeding, for 

 the mathematical formulte express only 

 definiteness and precision. Unfortunately, 

 the formulae can not express the indefinite- 

 ness and the unprecision which even Mendel 

 found in his work. My own feeling is that 

 the greatest benefit of Mendel's work to 

 the plant-breeder will be in improving the 

 methods of experimenting. We can no 

 longer be satisfied with mere 'trials' in 

 hybridizing: we must plan the work with 

 great care, have definite ideals, 'work to a 

 line,' and make accurate and statistical 

 studies of the separate marks or characters 

 of plants. His work suggests what we are 

 to look for and new ways of attacking dif- 

 ficult problems. 



Beyond this, I do not see how the orig- 

 inal Mendelian results will greatly modify 

 our plant-breeding practice. The best 

 breeders now breed to unit characters, for 

 this is the significance of such expressions 

 as 'avoid breeding for antagonistic char- 

 acters,' 'breed for one thing at a time,' 

 'know what you want,' 'have a definite 

 ideal,' 'keep the variety up to a standard.' 

 In certain classes of plants the Mendelian 

 laws will be found to apply with great 

 regularity, and in these we shall be able 

 to know beforehand about what to expect. 

 The number of cases in which the law, or 

 some modification of it, applies is being 

 extended daily, both for animals and plants 

 (see, for example, Bateson and Saunders' 

 report to the Royal Society on heredity) ; 



