Apr., 1889. 
SEPARATION OF ROCK CONSTITUENTS. 
91 
* G. pyrenaicum, 88. (2nd and 3rd Editions. Under the hedge of 
a meadow by a footpath between Cotlieridge and Bransford 
Hoads, St. John’s, near Worcester.) 
G. molle, 33. 
* G. lucidum, 33. Very plentiful on the rocks. Ill. 
(To be continued.) 
THE SEPARATION OF ROCK CONSTITUENTS BY 
MEANS OF HEAVY SOLUTIONS.* 
BY T. H. WALLER, B.A., B.SC. 
In investigating the history and geological relations of the 
crystalline rocks, the microscopical examination will some¬ 
times show differences between masses closely similar in their 
average chemical composition—varying groupings of the 
elements being brought about in all probability by differences 
in the conditions of crystallisation and solidification. To 
take a simple example, the average composition of the 
ordinary Hebridean gneiss of the North-west of Scotland— 
what is generally termed the bulk analysis—agrees almost 
exactly with that of a normal andesite ; such, for instance, 
as that from Montserrat, which I have shown here on two 
or three occasions. 
In addition, therefore, to making a bulk analysis of a 
rock, it is frequently desirable to analyse the component 
minerals, so far as they can be separated. 
For the purpose of this separation two lines of procedure 
have been used. In the one the powdered rock is allowed to 
stream between the poles of a tolerably powerful electro¬ 
magnet, which retains those minerals which contain iron in 
any notable proportion, allowing the others to pass freely. 
This requires frequent repetition before the separation is 
complete, and then leaves the problem of the separation of, 
say, quartz and felspar untouched. Nevertheless, it is a useful 
auxiliary to the other method, and an ordinary bar-magnet is 
serviceable in removing magnetite from other minerals. 
The other course is that which I am to show to you this 
evening, namely, the use of heavy solutions. 
* Transactions of the Birmingham Natural History and Micro¬ 
scopical Society, April 17tli, 1888. 
