20 MASS. EXPERIMENT STATION BULLETIN 339 



rank of white Dutch clover, the high soluble ash and phosphorus content of 

 orchard grass, the high calcium content of the bent grasses (Rhode Island bent 

 and red top\ the low rank of sheep fescue, and the rather low rank of Kentucky 

 bluegrass. 



(2) The application of a complete fertilizer lowered appreciably the dry 

 matter content of the fresh grass, and the crude fiber and calcium in the dry 

 matter; it materially increased the nitrogen in the dry matter and considerably 

 increased the phosphorus. Smaller increases were noted in ether extract, solu- 

 ble ash, and magnesium. These are average effects for all seven species studied. 

 Individual species showed some variation in their response to the fertilizer 

 treatment. 



(3) There is a very significant positive correlation between moisture content 

 of the fresh grass and content of nitrogen, soluble ash, calcium, and phosphorus 

 in the dry matter of the grass. There is an even more significant negative 

 correlation between moisture content and crude fiber content. It is concluded, 

 therefore, that the amount of water in a grass is a good indication of the nutri- 

 tive value of its dry matter. 



The Vitamin Content of Certain Pasture Grasses. (W. S. Ritchie and 

 J. G. Archibald.) This project was continued this year, samples of seven 

 species of grass being assayed in the green state in June for vitamins A and D, 

 and again in September for vitamin A. It is desired to repeat thie work another 

 season before publishing results. 



Performance Prediction from Pedigrees. (V. A. Rice.) The materials 

 for this study were published during 1935 and 1936 in the Guernsey Breeders' 

 Journal in the form of comparisons in production between a bull's daughters 

 and the dams of those daughters. Only those bulls with ten or more daughter- 

 dam comparisons were included. 



Intermediate indexes were computed for each of about 150 bulls. A study 

 revealed about 35 bulls with complete pedigrees from the standpoint of pro- 

 duction. These bulls were by sires with indexes and the grandsires also had 

 indexes. The dams of these bulls had records, the two grandams had records, 

 and the indexes were available on the sires of the two grandams. This material 

 was put together into one composite pedigree for the 35 bulls. 



The dam's record was averaged with her sire's index as the transmitting 

 ability of the cow. The same procedure was followed also in the case of the 

 two grandams. Six-tenths of the average of the sire's index and the dam's 

 transmitting index was added to four-tenths of the average of the grand- 

 parents' transmitting indexes to yield a figure known as the pedigree index for 

 these 35 bulls. This figure was 13,726 pounds of milk testing 5.09 percent. 

 These bulls were bred to 610 cows whose 1,045 records averaged 12,859 pounds 

 of milk testing 4.99 percent butterfat. The daughters should have produced, 

 theoretically, half way between the average production of the dams and the 

 pedigree index of the bulls. In other words, we should have expected them to 

 produce 13,293 pounds of milk testing 5.04 percent. They actually produced 

 13,312 pounds of milk testing 5.09 percent. In terms of total energy (milk 

 and fat combined) this is an error of about 0.7 percent. 



The material was then treated individually for each of the 35 bulls, and the 

 daughters missed expectations from 0.3 percent up to 16 percent, the general 

 average discrepancy being 4.5 percent. The error for 27 out of the 35 bulls 

 (77 percent) was 4 percent or less. 



It appears therefore that the transmitting ability of bulls may be told fairly 



