BREEDING TIMOTHY AT SVALOF 



Hernfrid Witte 

 Director of the Forage Plant Section of the Swedish Seed Association 



MY INTENTION in presenting 

 the following report is to give a 

 short account of the methods 

 of timothy breeding at Svalof, 

 the multifariousness of forms studied, 

 and the results gained from this branch 

 of breeding up to the present time. 



The Meadow culture in Sweden is of 

 high importance, about one-third of the 

 open area being occupied by meadows, 

 most of which are temporary. Sweden's 

 annual want of clover and grass seed 

 amounts to about 11 millions of kilo- 

 grams, of which not less than 4.8 mil- 

 lions is Timothy seed. In the year 

 1907, when grass breeding work was 

 taken up at Svalof, timothy was the 

 first to be made an object of investiga- 

 tion; but also orchard grass (Dactylis 

 glomerata) , meadow fescue {Festuca pra- 

 tensis), tall oat grass {Avena elatior), 

 perennial ryegrass (Lolium perenne), 

 field bromegrass (Bromus arvensis) and 

 several others were added. 



THEORY AND METHODS OF BREEDING 

 TIMOTHY 



Timothy normally being cross-pol- 

 linating and exhibiting manifold varia- 

 tions in nearly all its characters, appears 

 in an infinite number of different forms, 

 though only a very few of them are 

 fully constant, giving a uniform pro- 

 geny after self-fertilization. Suppose, for 

 instance, that twenty-five different pairs 

 of characters (surely a far too low num- 

 ber) are represented in timothy, this 

 certainly means that (2)" or Z2> million 

 constant combinations may exist, but 

 the number of possible combinations, 

 amounting to {2^^Y or about 1.126 

 milliards, means only one perfectly con- 

 stant individual in d>3 millions. It is 

 therefore practically impossible to base 

 timothy breeding on selection of con- 

 stant individuals. Instead one has to 

 select a great number of individuals, to 



study the practical value of the progeny 

 of each of them — ^if possible, in several 

 generations. It may be necessary to 

 repeat the selection before at last intro- 

 ducing to agriculture the variety or 

 varieties that have proven to be of the 

 highest practical value. 



The practical method used at Svalof 

 in breeding perennial grasses appears 

 from the following scheme. From na- 

 ture, or from cultures at Svalof, pedi- 

 greed plants are selected and vegeta- 

 tively multiplied in plots for the pur- 

 pose of attaining the truest possible 

 estimate of different characters desired. 

 The best of these plots are then re- 

 increased vegetatively into larger plots, 

 situated as solitarily as possible. This 

 is marked on the chart "vegetative 

 multiplication isolated." With seeds 

 from the last-mentioned plots compara- 

 tive trial plots are laid out, and when in 

 several trials a sort has proven prac- 

 tically superior to the very best com- 

 mercial seed, and also has afforded 

 evidence of practical uniformity, it 

 may be multiplied and distributed on a 

 large scale. Uniformity is studied in 

 special pedigree plots where the plants,, 

 raised from seeds obtained in artificial 

 isolation, are growing in like distances, 

 each way, generally 40 cm. In these 

 pedigree trials new selections are made — 

 individual as well as mass selections. 



VARIATION AND HEREDITY OF DIFFERENT 

 CHARACTERS 



The length of stem varies remarkably,, 

 from 20 to 30 cm. on the one hand to 

 110 to 120 cm. on the other (see Figs. 

 1 and 5), and displays a series of 

 hereditary gradations, though at times 

 it may be difficult distinctly to state the 

 hereditariness of the separate gradations, 

 since modifications from the environ- 

 ment are rather considerable. 



291 



