232 EXPERIMENT STATION RECORD. [Vol.40 



seemed to be best adapted to heavy clay soils. On soils of this type an Increase 

 in yield of 30 per cent was obtained from heavy as compared with thin seeding, 

 while the corresponding increase on alluvial clay was only 5 per cent. It Is 

 pointed out that while the increases in yield from the heavy rate of seeding 

 are comparatively small ; thicker seeding has the advantage of tending to crowd 

 out weed growth. 



Experiments with Danish and foreign strains of clover and grass, III, 

 1914-1917, E. I.indhakd (Tidsskr. Plantcavl, 25 (1918), Xo. 1, pp. 117-175).— 

 Strains of orchard grass, perennial rye grass, timothy, meadow fescue, white 

 clover, alsike clover, field brome grass, kidney vetch, and black medic were 

 compared In different localities under a cooperative arrangement. 



mi olne strains of orchard grass, Olsgaard ranked first in productiveness and 

 Tystofte No. 2 second. Five strains of rye grass, including Irish rye grass 

 grown from commercial seed, wen- t> -ted. A strain known as Luudba?k gave 

 the highest yield, exceeding the yield of the lri<h strain by 10 per cent. Only 

 threi' strains of timothy, including a COmmerda] sample, were tested, and of 

 these the Leading one, known as •« Trifolinm " No. 12, produced in the first 

 year about 40 per cent more hay than was Becnred from the commercial strain, 

 but proved Less hardy, especially on lowlands. 



A strain of meadow fescue designated Fadleaforeningen No. was compared 

 with Dsehnfeldt No. 2 and P.-vlinfeldt N<>. .".. strains derived from American 

 stock. These three strains mixed with white clover on ordinary upland soil 

 yielded in two years a total of 9,810, 9,120, and 8,910 leg. per hectare (4.36, 4.0">. 

 and i tons per acre), respectively, while on moorland soil the corresponding 

 yields were 7,150, 8,000, and s.Oi kg. 



Tests of while clever strains grown mixed with meadow fescue showed that 

 the Danish strains StrynQ and IforsS, as compared with Bohemian white 

 clover, were much better adapted to the conditions un both upland and low- 

 land soils. Tystofte No. 17 was found to be later than Strytio and also t" 

 give higher yields, but the strain is not regarded as sufficiently tested for 

 commercial seed production. A Danish strain of alsike clover designated 

 Fsellesforenlngen gave in general better results than Swedish commercial 

 samples. 



Fiellesforeningen field brome grass gave an average of 5 per cent more hay 

 than was secured from an ordinary commercial strain. Two strains of kidney 

 vetch, Tystofte No. 8 and Tystofte No. 28. compared with commercial stock, 

 were found superior in productiveness. Tystofte No. 8 is described as blossom- 

 ing a week later than the commercial stock and yielding a heavier crop at the 

 first cutting. Fsellesforeningen black medic proved to be an early strain of 

 vigorous growth, and yielded in this test about 13 per cent more hay than 

 was obtained from an English strain. 



Methods used and results obtained in cereal investigations at the Cornell 

 Station. II. II. LOVS and \Y. T. Cbaxq [Jour. Amcr. Soc. Agron.. 10 (1918), No. 

 4, pp. 145-157, pi. 1, fig. 1). — The authors describe the methods employed in 

 cereal breeding work at Cornell University, with particular reference to the 

 rod-row system of studying plant selections. Briefly, the method of selection 

 is as follows : 



"The heads or plants are selected and tested for one year in head or plant 

 rows. The best rows are selected in the field, harvested, thrashed, and grown 

 the second year in rod rows repeated two or three times, depending on the 

 amount of seed available. Only the very poorest are eliminated the second 

 year. The rest are continued in rod rows repeated ten times for at least three 

 years. The best new strains are multiplied and tested in Increase plats. The 

 best ones are finally distributed to farmers for further comparison. In this 



