548 



MR. W. A. RICHARD SOX : A MICRO-METRIC fvol. lxxix, 



used in the conversion of the determined volumes into weights. 

 The reduction of averages and routine calculations were made 

 on a 10-inch slide-rule 



Now, in all work of this kind the problem of the effects of 

 sampling is prominent ; hut here the problem is complicated in 

 several ways. We are studying the point-to-point variation of a 

 granite, which, though by no means uniform, presents no startling 

 quantitative differences, and yet is at the same time coarse- 



Fic.l. OisTRiBUTion or Samples. 



% 



'■^•ea Hensbarro" . STAuste 



2 miles 



Fig. 2. Quartz. 



vtiveSicte 



±— t_+_ |o Order G 



Positive 



Points conformable to both 

 jpmQs Cjroupintfs 



Exceptions to I s - 1 order 

 ... s „.., v ® Do to 2" J oider 



%MtM 2 n - d Order Croupwob. © Do tc bo! h groupings 



grained, and, therefore, highly susceptible to the vagaries of 

 sampling. Moreover, if five minerals are present, one only may be 

 really increasing or decreasing, while the others, maintaining then- 

 proportions relatively one to the other, yet as a whole decrease 

 or increase antipathetieally with the varying constituent. Some 

 criterion is consequently necessary to separate significant variations 

 from these effects combined with those of sampling. Further- 

 more, when many localities are examined, considerations of time 

 and expense impose a limit to the number of slides available from 



