August 24, 1905] 



NA TURE 



409 



it be present in exces: . the silica itself. It has also been 

 supposed that the orde: .nay be one of the fusibility of the 

 various minerals under the conditions of their formation ; 

 the least fusible mintr::ls being the earliest to crystallise, 

 and the most fusible the latest. Interesting speculations 

 concerning the melting point of quartz at high pressures, 

 and its consequent order of crystallisation, have, for 

 example, been published recently by Sironieyer and 

 Cunningham. 



It is not necessary, however, to regard the molten 

 magma as a mere mixture of fused minerals which solidify 

 more or less independently and consecutively; it is more 

 reasonable to regard the whole magma as a solution in 

 which the various minerals are dissolved, and from which 

 they crystallise as it cools. Now the temperature at which 

 a substance separates from solution is generally far below 

 its melting point, and the order in which the constituents 

 of a mixed solution will crystallise is the order of their 

 solubilitv in it, and bears no direct relation to their fusi- 

 bility or to their chemical composition. 



Teal! in iqoi, after discussing the controversies and 

 the evidence on which they are based, came to the con- 

 clusion that rock-magmas are solutions, and that the 

 order in which the minerals consolidate depends upon the 

 nature of the constituents and their properties, and is not 

 by any means the order of their freezing points. As to 

 the particular minerals which crystallise, he thought that 

 the molecular grouping in the magma is determined by 

 mass action and by the mutual affinities of the bases, the 

 silica, and the alumina. Concerning future research he 

 ventured to predict that the next advances were to be 

 made by experiment controlled by the modern theory of 

 solutions. 



Thirteen years earlier Teall had himself contributed a 

 valuable suggestion based upon Guthrie's work on cryo- 

 hydrates. When a mixture of nitrate of lead and nitre is 

 fused and allowed to cool, the constituent which is in 

 excess will crystallise out as from a solvent until the pro- 

 portions left in the liquid state are 47 of lead-nitrate to 

 53 of nitre, and this mixture will then solidify at 207°, not 

 as a imiform compound, but as an intimate mixture of the 

 two salts, the entcctic, which crystallises at the lowest 

 possible temperature, and is the only mixture which has 

 exactly the same composition as the liquid from which it 

 solidifies. Teall made the illuminating suggestion that 

 micropegmatite is an eutectic consisting of quartz and 

 felspar, and represents in certain rocks the final mother- 

 liquor from which the other minerals have crystallised out. 

 Eutectics in metallic alloys have been much studied during 

 recent years : in the Address of igoi Teall was able to 

 strengthen his case by showing that spherulitic and micro- 

 pegmatitic structures found in obsidian and other acid 

 rocks are paralleled by similar structures developed in 

 eutectic alloys, according as tliev have been rapidlv or 

 slowly cooled. 



In the following year appeared a theoretical paper by 

 MeyerholTer concerning the ideal case of a molten mixture 

 of two substances, a and b, which do not suffer double 

 decomposition, nor form a double salt, nor an isomorphous 

 mixture. 



Let a diagram be constructed, with temperatures as 

 ordinates and composition of the magma as abscissse, 

 giving by a curve the nature of the magma which is in 

 equilibrium with either solid a or solid b. The curve has 

 the form of a V ; one arm represents the temperature and 

 constitution of the liquid which can be in equilibrium 

 with o, and the other that of the liquid which can be in 

 equilibrium with b ; and the lowest point corresponds to 

 the eutectic, which is in contact with both. 



Let a point above the curve represent the temperature 

 and constitution of the liquid magma containing excess of 

 b : as the magma cools this point descends to the b branch 

 and travels along it while b is crystallising out, until the 

 eutectic point is reached, when a and b both crystallise 

 out together at a temperature below the melting point of 

 either. The order of crystallisation is therefore determined 

 solely by the composition of the magna as compared with 

 that of the eutectic. If, however, the liquid be cooled 

 slowly, crystallisation may be postponed until it has become 

 supersaturated with regard to one constituent or the other, 

 or both : a state of affairs represented by a prolongation 



NO. 1869, VOL. 72] 



of the amis of the V below its lowest point, and then 

 the order of the crystallisation may be inverted. 



In a rock-magma there are of course many other factors 

 to be taken into account as determining the order in which 

 the minerals separate ; for example, the formation of both 

 double salts and isomorphous mixtures, the possible pro- 

 duction of unstable solid compounds which may bfcome 

 converted into stable compounds or may be re-dissolved soon 

 after they have come into existence ; and also the relative 

 velocities of crystallisation, changes of temperature and 

 pressure, action of steam, &c. ; but the principle laid down 

 by Meyerhoffer must be that which controls the process. 



It might be objected that on this hypothesis the con- 

 solidation of every rock-mass ought to terminate with an 

 eutectic mixture, whereas this appears to be by no means 

 the case ; in fact, it is only among some acid rocks that 

 structures much resembling the eutectic mixtures of alloys 

 are to be found. On the other hand, if the conditions of 

 cooling are such that the magma becomes supersaturated 

 with one mineral after another, it will overshoot the eutectic 

 composition before each crystallises, and the final con- 

 solidation may be a well-marked sequence instead of a 

 simultaneous crystallisation. 



The controversies which have raged concerning the classi- 

 fication of rocks and their nomenclature appear to me to 

 contribute little to the real advancement of knowledge. 

 There are, I think, two more profitable lines of research 

 which should accompany each other. We may take the 

 facts as we find them and endeavour to explain them by 

 the known laws of solutions aided by the phase-rule, pro- 

 vided that we have good reason to believe that rock-magmas 

 behave like solutions, and we may make experiments upon 

 slags and fused silicates and ascertain how far they re- 

 semble natural rocks in their behaviour and their mineral 

 constitution. Some of the workers in this field have been 

 led to regard rock-magmas as undoubtedly similar to 

 ordinary solutions ; others hesitate to seek an explanation 

 for their features in the laws which govern the solutions 

 studied in the laboratory. The two views are represented 

 in the persons of the tw"0 men whose names are most 

 closely identified with recent experiments, Vogt of 

 Chrisliania and Doeller of Graz. 



UoL'lter's Work on Melting Points and Solubilities. 



The labours of Doeller and his pupils have been largely 

 devoted to the melting points of the rock-forming minerals 

 and their solubility in silicate magmas. From experiments 

 upon these minerals and their mixtures they have come 

 to the conclusion that in many cases the melting point of 

 the mixture is about the mean of the melting points of 

 the constituents, and that in such cases, therefore, there 

 is no evidence that the freezing point is lowered, or that 

 an eutectic mixture is formed ; so that it is not safe to 

 apply the theory of cryo-hydrates to fused mixtures of 

 silicates. 



Doelter is therefore led to regard the silicate-magmas 

 rather as mixtures of various constituents which may be 

 dissolved in each other, but which are not by any means 

 necessarily identical with the minerals which separate on 

 cooling. The whole process seems to him to be far too 

 complicated to be explained by any such simple principle 

 as the mere relative proportions of the various constituents 

 to each other and to their eutectic mixture ; the order of 

 crystallisation must be determined by a number of factors, 

 such as temperature, velocity of crystallisation, the interval 

 between the softening and fusing of each mineral (which 

 he finds to be considerable), viscosity, capillarity, the 

 presence of water and mineralising agents, and the ab- 

 sorption of adjacent rocks. 



To choose a simple example : minerals such as zircon, 

 corundum, and titanite separate for the most part early, 

 because they are less soluble. On the other hand, mag- 

 netite is one of the more soluble minerals, and yet it is 

 one of the first to separate; the same is to a crlain extent 

 true of augite, but not always. It is possible that in a 

 magma which still contains the iron of the magnetite 

 in solution plagioclase and augite mav be comparatively 

 soluble and magnetite comparatively insoluble, but that 

 when magnetite has already crystallised out from the 

 magma the plagioclase and augite may be comparatively 



