Agricultural Chemistry — Sheep- Feeding and Manure. 295 



in the field ; but as almost every animal lost, and some very con- 

 siderably, during the first 8 days, and then gained within a week 

 or two very large amounts, it was thought that the average results 

 would be overstated if the first week were not taken into ac- 

 count. This is accordingly done ; and as the clover-chafF was not 

 weighed during the first eight days, it is supposed to have been 

 taken at the same rate as the mean of all the other periods. With 

 this exception, and the replacement of No. 5 sheep in Pen 1 by 

 a fresh one, as already noticed, the results of the Table are 

 exactly as obtained by experiment. On inspection of the Table 

 it will be seen that there is not a single animal which does not 

 indicate a loss of weight at some, generally several periods of 

 the experiment ; whilst at others there is frequently during one 

 week, or for several weeks together, an increase far above the 

 average. Indeed such is the apparent generality of this fluctua- 

 tion, which was so prominent also in the case of the first series, 

 that the plan frequently adopted of deciding upon the quality of 

 different foods by putting animals for a week or two on one, and 

 then a week or two on another, and comparing the results, would 

 seem on this account alone to be sufficiently condemned. 



The average weekly gain or loss placed side by side will show 

 how far there is any uniformity as to fluctuation throughout the 

 several pens at the different periods of weighing. 



SERIES II. 



Table 3. — Showing the Weekly Average Gain or Loss per Sheep in 



each Pen throughout the course of the experiment. 



