NOTES ON EECENT KESEAllCH. 



895 



tribution, worthy of careful study both by the practical breeder and the 

 student of variation and heredity. The actual experiments detailed by 

 the author are mainly agricultural, though the general principles deduced 

 iire of equal interest and value to horticulture. In view of recent writings, 

 it is decidedly refreshing to read Prof. Hays' opinion in the introduction, 

 that " in European countries much more attention is given to the improve- 

 ment of plants than in America." And again : " The seedsmen of America 

 have not kept pace with European seed firms in variety formation, nor 

 even in keeping up and improving old forms." But, judging from Prof. 

 Hays' paper; it is possible that the " great system of American experiment 

 stations " more than makes up for these shortcomings. 



Examples of Results of Breeding. 



As an example of the results obtained by breeding, Prof. Hays men- 

 tions that the experiments in Wheat-breeding carried out at the Minnesota 

 Experiment Station for ten years have resulted in the production of a 

 new variety, which produces nearly 25 per cent, increase in yield over the 

 older varieties from which it has been bred, together with other advan- 

 tages, such as increased power of resisting "rust." The author also 

 quotes the European case of Sugar Beets, in which " the amount of sugar 

 in the juice of the roots has been "increased probably 100 per cent, by 

 the rigid scientific methods, first started by Vilmorin in France, and now 

 practised on a large and extensive scale by European seed-growers." 



The Value of Large Numbers in Breeding Experiments. 



Prof. Hays rightly lays stress on the value of large numbers in 

 breeding experiments, with rigid selection of the best forms only. 



In regard to this, a point of great importance is noted, and that is 

 that the best individual does not ahvays produce the best progeny : not 

 that the worse produce better^ but that a few only of the "best " will 

 produce improved offspring, owing to the constant tendency to regress 

 towards the average of the ancestry. Much testing has, therefore, to be 

 done in order to ascertain which of the " best " are really the " best " for 

 continuing and improving the breed. This fact has also been noted in 

 animals : the best performers are not always the best at the stud. "What 

 is required, therefore, is a careful selection of the selected for breeding 

 purposes. 



General Facts concerning Heredity. 



Prof. Hays sums up the following principles to be observed in improving 

 plants : — 



" (1) The individual plant produced from a seed is the important unit 

 in plant breeding. The ' bud unit,' though of much consequence in case 

 of marked bud variation, is usually of minor importance. 



" (2) Heredity, centripetal-like, enables us to produce from certain choice 

 plants many descendants which on the average, quite resemble their 

 parents. 



" (3) Variation, centrifugal-like, causes the production among the 

 descendants, along with very many average plants, of a few very good 

 individuals, and a few very poor ones. 



