108 THE SWEDISH GENETIC INSTITUTE IN AKARP 



not increase the nematodes in the soil, the yield of other cereals, especially of 

 oats, may be much higher than else. 



In other cases the valuable hereditary differences are not easy to find out 

 because of the fact that they manifest themselves only certain years, sometimes 

 with long intervals. The various (and often very important) hereditary different 

 ces in wheat with regard to the susceptibility of the attacks of Cicadula sexnotata 

 ear only to be seen about every fifth year (during the period 1900—1921 the 

 years are: 1902, 1912, 1915, 1918). Another disease in wheat, which also shows 

 considerable hereditary differences in different varieties and lines, was seen especi* 

 ally in the years 1907 and 1920. In other years the grains of different varieties 

 show very differing resistance against bad weather during harvest, and so on. 

 One of the main points of future breeding may therefore be, I think, the coUec* 

 tion through continued and systematic crossings of all such hereditary good qua* 

 lj.ties discovered by and by, even if they separately only seldom play a practical 

 role. In this manner the mean yielding of a series of years may surely be increased. 



A purpose of common interest to the plant breeding is the working out of 

 the best and simplest methods, for example, with regard to the selection in the 

 progeny of a cross, when the segregation of quantitative characters is very com» 

 plex. It is, then, necessary to have a very large F2«generation. If the parent* 

 varieties are rather dissimilar in characters a progeny of 20,000 plants is not much. 

 As it for practical reasons in general is impossible to raise 20,000 plots in the 

 F3«generation the strongest possible selection must take place in F2, although 

 selection with regard to quantitative character is very difficult. The question as 

 to the most eflfective method to be used in the F2<selection now becomes of im» 

 portance. In the last year the writer has selected only a very small percentage of 

 plants (in wheat), so that in the assumed case about 200 plots would be grown 

 in F3. Now a more definitive selection between these plots can be made. The 

 main point, then, is to work out the best methods for determining the yielding 

 power in small plots, as it would be an altogether too large a task to determine 

 the yieldingspower of a great number of cross^descendants from many different 

 crosses grown in sort*trials of ordinary size. 



A good deal of the work of the institute, therefore, aims at stating the facts 

 concerning the best and at the same time the most scientific methods to be used 

 in plant breeding. The results of this work shall be published later. 



With regard to teaching lectures on theoretical and applied genetics are gi* 

 ven at the University of Lund in the winter months, November— March. In the 

 summer the field experiments are demonstrated at different times, and the students 

 may take part for shorter or longer time in the current scientific work of the institute. 

 They also dispose of ground sufficient to run field^experiments of their own. 



The building contains seven working rooms, intended for examination of 

 plants and for ordinary breeding work. The material is stored in a spacious attic; 

 there are also cellars, and a neat barn connected with the main building and in» 

 tended for plant breeding work. For microscopic investigations and the like the 

 institute maintains a laboratory in the botanical institute at the University of Lund. 

 The institute disposes of about 3 hectares fine sandy loam^soil for the field experiments. 

 A suitable rotation of crops is maintained. The area may be easily increased if necessary. 



