p.irt 4] 



STUDY OF THE ST. AUSTELL GTtASITE. 



549 



each locality, and some means of evaluating the observations is 

 the more urgent. 



To afford material for a statistical study of the problem of 

 sampling, eight slides were taken from a granite of Dyer's Quarry, 

 Meledor (No. 48). Seven of these were cut from two specimens, 

 selected at a distance of about 40 feet from the contact of granite 

 and slate. An eighth slide was obtained from the Geological 

 Survey, and is, therefore, independently selected. The volumes of 

 the minerals obtained by measurement are recorded in Table I, of 

 which the lowest row gives the number of fields measured as an 

 indication of the area of the slides. All locality averages stated 

 hereafter are weighted according to the areas of the slides involved ; 

 but the general averages for whole districts are the unweighted 

 means of the localities. 



Table I. — 1'ekcentage Mineral Composition bv Volume of Eight Slides 

 from Dyer's Quarbt, Meledor (No. 48). 



Mineral. 



A. 



15. 



C. 



I). 



E. 



F. 



G. 



II. 



Quartz 



.1 337 



28-8 



38-2 



27-8 



20-5 



34-0 



267 



44-0 



Ortlioclase .. 



35-3 



49-0 



3T5 



45-0 



45-3 



42 1 



49-9 



28-6 



Plagioelase 



. 20-6 



117 



137 



18-5 



19-4 



11-9 



137 



151 



Biotite 



37 



32 



77 



2-0 



3-2 



47 



3-2 



5-0 







Muscovite 



n. f. 



0-3 



02 



O'l 



n. f. 



0-2 



n. f. 



n. f. 



Apatite 



. n.f. 



n. f. 



1-2 

 tr. 



tr. 



n.f. 



n. f. 



n.f. 



o-i 



Magnetite ... 



o-i 



0-2 



tr. 



n.f. 



tr. 



tr. 



tr. 



Tourmaline .. 



4-0 



3-6 



57 



36 



2-5 



1-8 



37 



3-3 



Topaz 



1-3 



3-0 



3-8 



2-1 



2-9 



2-2 



2-4 



3-3 



Fluorite 



07 



0-2 



n.f. 



n. f. 



0-2 



o-i 



0-4 



o-i 



No. of Fields.. 



14 



18 



18 



20 



13 



20 



16 



14 



The addition of the eighth slide hardly affected the mean 

 appreciably, and the final mean of the eight slides may lie 

 considered as giving a close approximation to the true composition 

 of the rock. Averages were now calculated for all combinations of 

 the slides, taken one at a time; two at a time ; and so on, up to 

 seven at a time. In this way suites were obtained, each contain- 

 ing large numbers of averages, involving the worst conditions of 

 sampling in the material. The criterion of sampling adopted was 

 the standard deviation (that is, the error of the mean square). 

 Accordingly, the standard deviations of the minerals of each suite 

 ( that is, for one, two, three, etc. slides) were calculated ; these 

 will be found in Table II, and are plotted in Hg. 3, p. 550, 



