1»J7 



(1) For tlio prodiu'tioii of milk a latiou cousistinj; of Lespedeza iiay auil raw 

 cottou seed is tlio clioapost. 



(2) For the iiroduetion of butter a ration of lA-spedo/.a or licnnuda hay and 

 steamed cottou seed is the most prolltahle. 



('A) When fed cotton seed, either raw, roasted, or steamed, cows will ;cain in Ih-sh 

 belter than when fed corn meal or cotton-seed meal. 



(i) Either liermnda or JA^apedeza hay is i)referal»le to timothy hay. 



(5) Coru meal is not an economical j^rain ration when fed with dry hay. 



(6) Steamed [cotton] seed will produce better butter than will eithi-r raw or roasted 

 seed. 



JI((j/Jor icorking mules (p. 4). — An cxpciiiiu'iit in which thioc; woi k- 

 inu' nailt's were, fod Bcrinudii hay and three others timothy hay (i<l 

 lihitidH tor 2 months, both kits receiving" tlie same amount ol' corn, 

 showed " practically no difference between the two rations." 



Comparative tests of machines A^'D ^methods for tiii: deter- 



:viINATION of fat IN MILK, L. G. PATTERSON (pp. 5-10, tigs. 7). — A 



des('ri[)tion is given of the Patrick brine-bath method, and the Bab- 

 cock and JU'imling centrifugal methods for determining the fat in milk, 

 together with the results of com])arative determinations by each of 

 these methods and the Adams gravimetric method. Five tests by the 

 Babcock method were spoiled by tin? belt becoming too loose. Of the 

 20 remaining tests by this method, the results of differed from the 

 gravimetric by 0.1 per cent or less, 8 by O.U per cent or over, and 2 by 

 over 0.;>'j per cent, the largest ditlcrence being 0.;38 per cent. Of ,14: 

 tests by the Beimling method 18 gave results differing from the gravi- 

 metric by 0.1 per cent or less, 4 by 0.2 per cent or over, and tlic great- 

 est ditlcrence was 0.24 per cent. Twelve of the 17 tests made by the 

 Patrick metliod differed from the gravimetric by 0.1 per cent or less, 2 

 by over 0.2 per cent, and the greatest difference was 0.27 per cent. 

 Tin? foHowing are the author's conclusions: 



The Beimling method requires less time than does either of the others, but tln^ scale 

 on the bottles is not as easy to read an<l no means are provided for kccpiuLC the fat in 

 a melted condition. 



The Babcock has the best bottles, but rcf[uires more time for its O])eration, and it 

 is absolutely necessary to have the si)ecitic gravity of the acid exactly 1.82. 



The I'atrick method is very convenient where a largo number of tests are to bo 

 made, but it requires considerable time to heat the bath, and the bottles are very 

 easily broken. 



Missouri Station, Bulletin No. 15, July, 1891 (pp. 16). 



Test of varieties of wheat and oats, 11. J. ^N'aters (pp. 

 -i-ll). — Tabulated data are given for 13 varieties of Avheat tested in 

 1880 and 55 in 1801, and for 11 of oats in 1880 and 13 in 1801. 



In the test of varieties of wheat in 1889, Fultz led in yield and gave the smallest 

 number of pounds of straw per bushel of grain; while in point of yield Red Chatf 

 Avas second and Hybrid ileditorrauean third. In 1891 the order was changed, giving 

 Extra Early Oakley the lead with a yield of 42.5 bushels, while Willits followed 

 with a yield of o9 bushels, and Coryell third, giving a crop of 36.9 bushels. In 1889 



