DErARTMKNT OF E.\ I'KRl M KNlAf. PlANT-BrEEDING. H 



Second, experiments with clover to determine the variation in 

 hardiness and the possibiHty of breeding improved races which will 

 be less susceptible to winter killing. These experiments at the 

 present time, are conducted on a limited scale to obtain further in- 

 formation with a view, ultimately of conducting more careful 

 studies of the subject. Some forty or fifty varieties of clover are 

 now under cultivation in small plats. 



Third, a limited number of selections have been made with brome 

 grass to determine what variations occur in this forage plant, and 

 also to determine what advance can be made by the selection of 

 improved strains. 



Fourth, cultivations are being made of different varieties of al- 

 falfa in order to test their comparative adaptability and value as 

 foundation stocks, with a view, ultimately, of taking up careful 

 breeding investigations with those wdiich are found to be best suited 

 to New York conditions. 



Fifth, some experiments have been undertaken in the breeding 

 of mangels, particularly with reference to securing single germ 

 varieties, and also to determine the best methods of breeding such 

 crops, and what can be accomplished in the production of strains 

 giving a high yield of dry matter per acre. The relation of specific 

 gravity to percentage of dry matter has been studied during the last 

 two seasons and it is expected that the data on this subject will be 

 ready for publication during the next year. 



Adams Act Research. 



The work conducted under the Adams Act is required by the text 

 of the law to be limited to investigations of fundamental scientific 

 nature. The aim, therefore, in all of the investigations outlined 

 under this heading is to extend our knowledge of the fundamental 

 laws of breeding, rather than to produce improved varieties. 



Studies 0)1 variation. — One of the problems oi greatest im- 

 portance to plant-breeding at the present time is to determine the 

 factors influencing or causing variations, and the extent to which 

 it is possible to artificially produce or induce variations of value 

 in breeding. Much of the investigation thus far conducted by the 

 department has been on this problem, and the experiments may be 

 grouped under the following heads : 



I. Experiments to test the so-called Knight's Law that high feed- 

 ing increases the range of variation. An extended series of ex- 

 periments has been conducted on this subject in the last two years 



