256 
POPULA.R SCIENCE REVIEW. 
influences or weathering on all surface rocks, that hand speci- 
mens so collected are not likely to turn out correct representa- 
tives of the rock mass as a whole. 
A single example of this may suffice ; in the last number of 
this Keview, p. 208, will be found an analysis of the rock from 
an eruptive dyke in the Penrhyn quarries. North Wales, de- 
scribed as a greenstone, and containing no less than twenty-nine 
per cent, of the carbonates of lime and magnesia along with 
sulphate of lime. It is well known that such eiupti\^e silicated 
rocks, when normal, do not contain any carbonates or sulphates 
whatsoever, and an inspection of the specimen showed that the 
carbonates found were really contained in cracks and pores in 
the rock, and had been of subsequent formation, proceeding evi- 
dently from the decomposition of the silicates of these bases 
caused by the infiltration of water containing carbonic acid. 
The analysis of this rock is not without value in showing the 
nature of the changes which do take place under such circum- 
stances ; but in order to prevent the tabulated results of this 
analysis leading to erroneous deductions, the rock itself should 
have been described as abnormal, or altered from its original 
condition. In the same way basalts and dolerites are often 
described as containing carbonate of lime, but this, as in the 
above case, is only a result of decomposition ; and in no case 
have the analyses of such rocks (taken in quarries or excava- 
tions beyond the reach of such external influences) shown the 
presence of any carbonates, unless, as happens in the lavas of 
Vesuvius, they may have been fragments mechanically entan- 
gled in the silicated mass. 
When it is proposed to make a chemical examination of any 
particular rock, it should first of all be carefully studied in the 
field, in order that a correct opinion may be formed as to the 
true nature of the rock substance itself, when uncontaminated 
or unchanged by external influences ; a specimen may then be 
taken which, in some measure, will represent the actual rock 
mass on the large scale, although this is attended with con- 
siderable difficulty and trouble, unless (as fortunately in England 
is generally the case) excavations or quarries have laid bare a 
face of rock, and so afforded facilities for obtaining the un- 
altered rock itself. 
The quantity required to be taken and pulverised, in order 
to obtain an average for analysis, must entirely depend upon 
■whether the rock is of a fine or coarse grained texture. In the 
latter case, a much larger quantity must naturally be employed ; 
it should not be pulverised in an iron mortar, as, when rocks are 
hard, it is found that the pow^der so produced wdll contain suf- 
ficient iron, derived from the abrasion of the mortar, to affect 
the accuracy of the results. The pieces should be broken up on 
