~).~)2 ME. W. A. EICHAEDSOK : A MTCEOMETBTC [vol. lxxix, 



Space does not permit of each case being argued separately. 

 But it will be understood that variations were compared with 

 standard deviations in the drawing of boundary-lines, etc. 

 Furthermore, since rock-changes proceed upon some definite plan, 

 smaller variations than the limits of sampling may become signifi- 

 cant when each locality does not stand alone. If the variations 

 from point to point are merely due to the sampling, then their 

 distribution over the area will be haphazard. If, on the other 

 hand, they group themselves in a definite way, it is clear that 

 such arrangement could not be regarded as caused by sampling, 

 but must be sought for in real differences in the material. In 

 such an event, the standard deviation is a test which greatly 

 assists the evaluation of exceptions to the general grouping. 



Table IV (facing this page) sets forth the results of the measure- 

 ments, reduced to percentages by weight, for the fifty localities 

 within the main granite-mass. The difference between the weights 

 and volumes involves only a small correction, and the volumes are 

 omitted in order to save space. There are several species of mica 

 present, but they are grouped here as ' coloured ' and ' colourless.' 

 The nature and distribution of the types will be considered below. 



The porphyritic character of the granite can be dealt with 

 where large surfaces suitable for measurement are available, as in 

 the Gready Quarries. Nevertheless, some difficult}- exists even 

 here, for the phenocrysts and ground-mass crystals of orthoclase 

 are not sharply separated as they are, for example, at Shap. There 

 appears to be a series in which every size is present from the 

 largest phenocryst to the normal ground-mass grain. It is, there- 

 fore, difficult to make sure that slides represent ground-mass only. 



This question of the porphyritic orthoclase chiefly affects the 

 Luxullyan area. The amount of phenocrystic orthoclase was 

 measured at Golden Point (No. 5) and found to be 18"3per cent. ; 

 and at Gready (No. 4) and found to be 19 - 8 per cent. Round 

 figures of IS and 20 per cent, respectively were taken, and the 

 compositions given in Table IV for these localities have been 

 corrected for this amount. It is certain that Helman Tor (No. 6) 

 should also be corrected. The phenocrysts here seem smaller, and 

 less numerous ; but, on the scattered lichen-covered boulders, 

 satisfactory measurements could not be made, and no correction 

 has been applied. Towards the margins in this area the pheno- 

 crysts either disappear or are small in amount, but easterly 

 exposures are neither numerous nor good. This district, therefore, 

 was sampled by one section, along which good exposures were 

 available. The results of this section Avill probably give a better 

 result for the bulk composition of the area, than if a number of 

 specimens from unsatisfactory localities had been included. 



In the other areas the porphyritic orthoclase is less troublesome. 

 The phenocrysts are much smaller and generally quite sporadic. 

 At Foxholes a measurement indicated about If per cent. ; another 

 at Yondertown not more than 5 per cent. These figures are well 

 within the limits of slide-sampling, and corrections were not made. 



