18 - BULLETIN 353, U. S. DEPARTMENT OF AGRICULTURE. 
Extreme care in sampling has kept the probable error very low on 
the single sample, so that it is nowhere excessive, but replicating the 
sample three times reduces the probable error 51.5 per cent, while a 
replication five and six times reduces the probable error over single 
samples 55.1 per cent and over three replications only 7.4 per cent. 
It does not seem necessary, therefore, in practice to replicate more 
than three times. Single samples, however, can not be considered 
safe when there is wide variation within the plat unless extreme care 
is used to make the sample composite and representative of the entire 
area. 
MOISTURE PERCENTAGES IN GREEN FORAGE AND IN FIELD-CURED FORAGE, AS SHOWN 
BY SAMPLES. 
Farrell, in an article in the American Journal of Agronomy,’ sug- 
gests the desirability of expressing alfalfa-hay yields in terms of green 
weight. In the article referred to above, he reports 76.5 per cent of 
moisture lost in air drying, which would be approximately equivalent 
to 79.5 per cent of total moisture. The average percentage of moisture 
in the 23 analyses of green alfalfa reported by Jenkins and Winton ? 
was 71.8. 
At Arlington Farm, Va., green alfalfa averaged 75.2 per cent of 
moisture in 20 samples. This percentage is probably near the aver- 
age for moderately thrifty alfalfa grown without irrigation in the 
Central and Eastern States. Alfalfa grown under irrigation and cut 
when one-tenth in bloom at Chico, Cal., averaged in 1914, 76.9 per 
cent of moisture. In 1911 McKee? found at this station as the 
average of 28 determinations in alfalfa not quite in bloom 85.8 per 
cent of moisture. The 1914 results indicate that the condition of 
crowth affects the moisture content very decidedly. Owing to excessive 
heat and scarcity of water, the alfalfa used for the 1914 samples was 
less vigorous than that of 1911 and correspondingly less succulent. 
These differences indicate very clearly the danger of basing yields on 
the green weight, as suggested by Farrell, or of using some arbitrary 
percentage of moisture in making corrections on the green weight. 
Samples should always be taken in experimental work when the crop 
is harvested and the amount of moisture in the forage at that time 
determined from them. Field-cured alfalfa at Chico in 1914 had 22.3 
per cent of moisture, while Jenkins and Winton ” report as the average 
of 21 analyses only 8.4 percent. The samples of Jenkins and Winton 
had probably dried out to some extent after being brought into the 
laboratory. 
1 Farrell, F. D. Basing alfalfa yields on green weights. In Jour. Amer. Soc. Agron., v. 6, no. 1, p. 42-45, 
Gio E. H.,and Winton, A. L. A compilation of analyses of American feeding stuffs. U.S. Dept. 
Agr., Office Exp. Stas. Bul. 11, p. 22-75, 1892. 
*’ McKee, Roland. Arabian alfalfa. Jn U.S. Dept. Agr., Bur. Plant Indus. Cir. 119, p. 25-30, 1913. 
