238 EXPERIMENT STATION EECORD. 



It was found that when the faulty and light seeds were taken out, the re- 

 mainder showed a much higher percentage of germination than samples from 

 the same lot which had not been separated. This was especially true of such 

 seed as tobacco, onion, and celery. 



Seed purity work, 1909, G. H. Chapaian {Massachusetts Sta. Rpt. 1909, 

 pt. 2, pp. 57-60). — The results are given in tabulated form of purity tests of 

 the seeds offered in the open market of timothy, red clover, redtop, oats, alfalfa, 

 alsike clover, while clover, alfalfa clover, orchard grass, Agrostis, Kentucky 

 blue grass, meadow fescue, millet, Italian rye grass, and yellow oat grass. The 

 most common impurities found in these seeds were plantain, rib grass, sheep 

 sorrel, and dock. 



HOETICULTTJEE. 



Experiments in breeding' sweet corn, 11. Pearl and F. M. Surface (Maine 

 Sta. Bui. 183, pp. 249-316, pis. S, figs. 7). — Experiments in breeding sweet corn, 

 extending over 3 years, 1907 to 1909, are reported, including the important data 

 and a brief account of the general features of the work. Further discussion 

 of the more technical results has been reserved for later publication. The 

 experiments w'ere conducted with a view of helping both the farmer and the 

 packer. 



Two types of corn, differing chiefly in regard to earliness of maturity and 

 fineness and depth of kernel, were dealt with in the experiments. Both types 

 are white in color and the corn designated as type 1 is a superior variety in 

 regard to the above named characteristics. Selection was conducted with both 

 types for improvements relative to earliness of maturity, yield both of ears 

 and stover, and the general conformation of the ear, especially with reference 

 to the shape and covering of the tip with kernels. Selection was made of de- 

 sirable plants in the field and between the ears harvested from these plants. 



A marked gain in earliness was observed after the first year's selection of 

 type 1 corn, but no further gain in earliness followed the selection practiced in 

 the two subsequent years. Corn selected from type 2 in 1907 was grown in a 

 new locality in 1908. The " new-place " effect appeared to offset the effect of 

 selection as far as the crop in 1908 was concerned. In general the first year's 

 selection has been followed by a marked improvement in respect to the confor- 

 mation of ear and the gain has been maintained where the corn has been grown 

 in localities to which it is well adjusted. 



Two years' ear-to-row test furnished no evidence that there is any close 

 association or correlation between the size or conformation of the seed ear and 

 the yield of corn obtained from it upon planting. The present experiments 

 point clearly to the conclusion that in any attempt to improve corn by selection, 

 the fundamental datum must be the performance of the row planted on the ear- 

 to-row system rather than the individual ear or plant. The authors, however, 

 do not advocate the isolation of a single pure line of homozygote strain as 

 the thing to be aimed at in practical corn breeding. The aim should be to 

 get rid of poor genotypes from the breeding plats as a whole and to permit 

 broad breeding between the rows of the selected seed. 



The experiments thus far give no evidence that there is a cumulative effect 

 of the selection of small fluctuating variations in sweet corn. A wide distribu- 

 tion of selected seed over the State in 1908 demonstrated the importance of 

 the factor of local adjustment in the improvement of the crop by breeding. 

 In 1908 type 1 corn did not show the same earliness throughout the State as 

 on the station plats. In 1910, however, the corn appears to have adjusted 

 itself to various localities and superior earliness was again in evidence. 



