the StcK 



"o varied so f; r fr- tual counts - 



shown when the stock is trensplnn r distributed - 



that it r if e rent scheme for making the 



estimate wre necessary, 



Tae actual amount of stock invariably fell below 

 the ietimate. This T; e chiefly to one thing: 



In estimating the .e or several 



assumed average b o estimated by taking a cotint 



from a foot square frame set in that part of the bed 

 supposed to be ontation of the whole bed, 



tfuet c. few of these counts were made and the estimate 

 of all that plot a these. 



the first pi ce, Ss^on is bound to favor 

 a better rather than orage bed, an" secona, the fraa* 



will be set in the best part of the bed. ven though 



this were not true in cases, it woul I - ible 



to pick out ti bed or sov beds from a plot of 100 that 



woul$ give ir average for all. When the stand i:, 



;- 10,000 per 48 , ifference of 1,000 one way 



or the other cannot be perceived b; >al glance. 



iso in the a, a hit or miss shot 



few of 1 to be average ones, will be 



the b,!Sis for the t , The tendency is to 



ost the work, or feel sure of itc result that the beet 

 row; 3ked as verege, 



