PRESSURE IN FORMATION OF ROCKS AND MINERALS 731 



pressure of i atm. — there is also the freezing-pressure at a definite 

 temperature; for example, mercury freezes at —38.85° under 

 I atm.; at 0° C. its freezing-pressure — the pressure which will 

 cause it to freeze — is about 7,600 atm. 



The purpose of the present paper is to discuss briefly the avail- 

 able experimental evidence in the light of the principles involved, 

 to indicate the conclusions which the somewhat scanty observa- 

 tions seem to justify, and to point out the Umitations to which such 

 conclusions are necessarily subject. In general it would seem that 

 the importance to geology of the effects of pressure upon changes 

 such as the melting of a single pure substance — those, namely, 

 usually regarded as physical changes — ^has been overestimated 

 relatively to that of the influence of pressure upon systems con- 

 taining more than one component — namely, upon chemical changes. 

 The influence of pressure upon chemical equilibrium is especially 

 marked if one or more of the components of the system are volatile, 

 and must therefore be taken into account in any discussion of the 

 behavior of the magma; indeed, the order of separation from a 

 complex solution containing a volatile component depends just as 

 much on whether, and how, the pressure changes as upon the mode 

 and rate of cooling. 



There has been very little direct experiment on the influence of 

 pressure upon the formation of rocks and minerals, by reason of 

 the technical difficulties in the way of making such experiments; 

 so that we are forced in large part to reason from analogy with those 

 substances whose behavior it has been possible to investigate. In 

 reasoning from analogy we must, of course, bear in mind the inhe- 

 rent limitations to which conclusions thus derived are subject. On 

 the other hand, in reasoning backward from present-day field evi- 

 dence, one must also proceed with caution; for, presuming even 

 that one is aware of all the factors in the net result as we see it, it 

 is hard to disentangle the effects of these several factors, especially 

 since the intensity of each is in general unknown. For example, we 

 can make only a very rough guess at the temperature and pressure 

 prevailing at the time when a given process was taking place, and 

 consequently cannot gauge satisfactorily the relative importance 

 of these factors in producing the result as we see it. Moreover, the 



