o64> MR. W. A. RICHARDSOK : A MICROMETEIC [vol. lxxix, 



It becomes evident that topaz only occurs in quantity when the 

 minerals of the contact-group are practically absent. Moreover, 

 the amounts of contact-minerals and the topaz closely correspond. 

 It is clear, therefore, that negative correlation is practically 

 perfect. 



Secondly, it will be recalled that, in the primary grouping 

 of the minerals on the map of the district, the distribution of 

 topaz differed from that of other minerals in being clearly related 

 to the margin of the granite. And, if the topaz is parti}" the 

 result of assimilation of sediment, then we should naturally expect 

 it to increase somewhat towards the margins of the mass. Conse- 

 quently, on the whole, the topaz may be regarded as largely a 

 contact-mineral representing the andalusite of other districts, as 

 Dr. Flett maintained. 



V. Structure and Grain. 



When speaking of the ' grain ' of a rock, it is far better to have 

 some quantitative value in mind than to rely on such vague 

 expressions as 'coarse' and 'tine' grain. In choosing such a 

 measure it is necessaiy to decide whether all the minerals shall 

 be measured, or one only. If all the minerals are taken into 

 account, in what way are we to express the result ? And, if one 

 mineral is to be used, which is likely to give the most useful 

 result '? No accepted method seems to exist, but the matter is 

 helped by Mr. J. A. Howe's l conclusion that, if one mineral is 

 large, all the others tend to be so too. In order to see whether 

 this statement could be applied to the rocks under investigation, 

 the principal minerals in a few slides from different parts were 

 measured. It was found that the mean diameters of quartz, 

 orthoclase, and plagioclase generally did correspond, but that the 

 size of mica-flakes remained more or less constant throughout. 



In view of this result quartz was finally chosen as the mineral 

 upon which to base measurements for the grain, because among 

 the principal minerals it gives the nearest approach to an equi- 

 dimensional unit. The following conditions were observed : — 



(a) Where the quartz occurred in the aggregates described above (p. 553) 



the individual grains were measured (between crossed nicols). and 

 not the size of the aggregates. 



(b) The mean of the greatest and least diameters was taken. 



(c) The measurements were tabulated and counted at millimetre-intervals : 



that is, all lying between 1 and 2 mm. were counted as 1'5 mm., and 

 so on. For sizes less than 8 mm., however, smaller intervals were 

 iised. 

 (d). The interval, or grade, containing the greater number was adopted as 

 the grain for the locality : in other words, the mode of the measure- 

 ments was adopted, and not their arithmetical mean. The mode has 

 many advantages for the purpose. First, a certain amount of time is 

 saved by avoidance of routine calculations ; secondly, the most nume- 

 rous size is a better measure of grain than mean size ; and, thirdly, if 



1 ' The Geology of Building-Stones ' 1910. p. 49, 



