452 



SCIENCE. 



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



of them are monoecious or dioecious. Even 

 if there is gametic purity in such plants, 

 the probability is that the fact can be dis- 

 covered only by a long line of scientific 

 experimenting for that particular purpose, 

 and not by the work of the man who de- 

 sires only to breed new plants. (9) A 

 cultural variety, in any true acceptation 

 of the term, is a series of closely related 

 plants having a pedigree. It runs back 

 to one individtial plant, from which propa- 

 gation has been made by seeds or asexual 

 parts. Now, one can never predict just 

 what combination of characters any plant 

 will have, even though it be strictly Men- 

 delian. A person might have a thousand 

 plants of peas of which no one plant shows 

 any of the characters in the proportion of 

 3 to 1, let alone all the characters as 3 to 

 1; and yet the total average numerical re- 

 sults might conform exactly to the Men- 

 delian law. Mendel's law is a law of av- 

 erages. The very fact that one must 

 employ such large numbers to secure the 

 numerical results shows that we can not 

 predict as to individuals. For example, 

 in ten plants of pea, Mendel found the 

 following ratios in respect to seed-shape 

 and seed-color: 



Mendel reports one instance in which the 

 ratio in seed-shape was 21 to 1, and an- 

 other of 1 to 1. He also reports instances 

 of seed-color of 32 to 1 and 1 to 1. It has 

 been said that, because of Mendel's work, 

 we shall be able to produce hybrid varieties 

 with the same certainty that we produce 



chemical compounds. Now, a plant is 

 made up of many combinations of many 

 units, and these combinations are the re- 

 sults of mathematical chance or probability. 

 Chemical compounds are specific entities, 

 in which the parts combine by mathemat- 

 ical definiteness. The comparison, as it 

 appeals to me, is fallacious and the conclu- 

 sions unsound. 



We must remember that there are whole 

 classes of cases of plant-breeding that do 

 not fall under hybridization at all. Grant- 

 ing the De Vriesian view that selection is 

 incompetent to produce species from indi- 

 vidual fluctuations, it is, nevertheless, well 

 established (and admitted by De Yries) 

 that very many of our most useful cultural 

 varieties have been brought to their present 

 state of perfection by means of selection; 

 and by selection they are maintained in 

 their usefulness. Selection will always be 

 a most important agency in the hands of 

 the gardener— none the less so now that we 

 have challenged its role in the evolution of 

 the plant kingdom. For the time being, 

 the new discussions of hybridization are 

 likely to overshadow all other agencies in 

 plant-breeding ; but selection under cultiva- 

 tion is as important now as it was in the 

 days of Van Mons and Darwin. 



rV. INTERPRETATION OF HYBRIDISM. 



I believe that the clearest insight into 

 this whole new question of hybridization is 

 to be got by following the work of De Vries. 

 The concluding parts of the second volume 

 of his 'Mutationstheorie, ' a volume devoted 

 wholly to hybridization, is on the press at 

 this moment. The Mendelian laws are 

 fully discussed in this volume, but the sum- 

 mary conclusions may be presented here. 

 De Vries had been working at hybridiza- 

 tion long before he discovered Mendel, 

 and had arrived at practically the same re- 

 sults; he had also arrived at other results 



