THE EVOLUTION OF PETKOLOGICAL IDEAS. 307 



tion of the earlior-formed minerals; * * * i\^q characteristic of 

 all [such occurrences] is that tlie sev(>ral constituents are concentrated 

 in a detinite order, which is identical with the order in which they 

 crystallize out from the mao-ma." 



All theories which depend on diffusion or molecular How liavt^ been 

 criticized by Mr. Becker on the ground that the rate of ditiusion is 

 too sh)w to produce the results attributed to it in any reasonal)le time. 

 He shows that, in the case of a column of water resting upon a layer 

 of copper sulphate, th(^ lapse of 1,000, 000 years would be recjuired to 

 produce sensible discoloration at a height of 850 meters, or semisat- 

 uration at a height of 84 meters; and he considers that the molecular 

 How of any compound in a silicate magma w^ould probably ])e at least 

 tifty times less rapid, so that a mass of lava 1 cubic kilometer in vol- 

 ume ''would not have had time to segregate into distinctly dili'erent 

 rocks l)y molecular flow if it had been kept melted since th(> close of 

 the Archican period." 1 am by no means averse to making h(>a\ y drafts 

 on the bank of time for geological purposes, but unless some elective 

 answer to Mr. Becker's arguments can be found 1 think that we shall 

 liave to give up unaided molecular flow as an important factor in the 

 origin of petrographical species. 



Mr. Becker has not, however, simply confined himself to destructive 

 criticism. He has proposed a theory of differentiation dei)endent on 

 '•fractional crystallization." During the cooling of a mass of molten 

 matter in a dike or laccolite, convection currents will be established. 

 These will act as stirrers, and, aided by diffusion, will tend rapidly to 

 restore homogeneity in the liquid mass after it has been destroyed l)y 

 the deposition of the first-formed crystals on the walls of tlu^ cooling- 

 surfaces. He compares a laccolite, in which the marginal parts are 

 different from the center, to a barrel of cider which has l)ecn frozen 

 from the outside. During the earlier stages nearly pure ice is foiined 

 on the walls, while the alcohol is concentrated in the central portion; 

 from this a licjuor, gradually increasing in strength, may be drawn off 

 as consolidation progresses. Here we see a furtlier dc\-eiopnient of 

 the idea originated ))y Darwin. 



Ail forms of the dilfei-cntiation theory Uiko as Ihcir starling point a 

 homogeneous magma, and then proceed to deri\'e from it the dillci'ent 

 varieties of igneous rocks as we now see them l)y magmatic or some 

 other form of differentiation. Are we justified in taking this view '^ 

 As ap})licd to certain districts, and (\specially to the ('hi'istiania dis- 

 trict which Professoi' I^i'cigger has done so nuich to (^lucidat(\ it has 

 j)roved of great value. But if we look at the general ([uestion, there 

 are many facts which should gi\<> us })ause. The earth's crust is cer- 

 tainly heterogeneous, and if magmas are, in any case, formed by the 

 refusion of solid'rocks,- it is probable, as Mr. Beck(M- has })ointed out, 

 that such magmas would be hetci'ogeneous at t\w start. K\'cn the 



