THE ROLE OF CONCENTRATION FLOWS AT THE STAGE OF CRYSTALLIZATION DIFFERENTIATION OF MAGMA

A. M. KUZMIN

In the article on the nature of concentration flows during crystallization [20], the author, on the basis of experimental studies, considered the role of concentration jets in the transfer of matter in solutions and in the phenomenon of crystal growth under conditions of mass crystallization. In the present work, materials are presented on the application of the experiments just indicated to the explanation of phenomena associated with the development of magma at the stage of crystallization differentiation, when concentration flows are one of the main factors determining the movement of matter both within the magma itself and its removal from the magma into the thickness of the earth’s crust.

In substantiating the role of concentration flows in the development of a number of phenomena closely connected with crystallization differentiation as a new factor, the author of the article does not intend to oppose his views to the views of A. G. Betekhtin [2, 3] and L. Greighton [10, 30], which, it must be supposed, also have a place in the formation of mineral associations.

The nature of movements in magma at the stage of crystallization differentiation

On the basis of a number of experimental studies of mass crystallization at many levels in the field of terrestrial gravity [15, 16, 18], it may be confidently considered that the formation of magma first passes through the stage of magmatic differentiation, and then the stage of crystallization. At the stage of magmatic differentiation the components making up the magma (the complexes \(K + M + Q\)), obeying the force of gravity, are regularly distributed in it in accordance with the hypsometric law of Laplace

\[ h = \frac{RT \ln p_0/p_h}{Mg} = \frac{RT \ln c_0/c_h}{Mg}. \]

This law in physical chemistry makes it possible to relate the magnitude of the pressure \(p_0\) at a given level, \(p_h\)—the pressure at height \(h\), \(M\)—the molecular weight of the substance of the complexes, \(T\)—the absolute temperature, \(R\)—the universal gas constant, \(c_0\)—the concentration at the given level, \(c_h\)—the concentration of the substance at level \(h\).

The regularity of the distribution of substances by density that follows from this can be illustrated at least by the distribution diagram of the principal components¹ that make up magmatic rocks, i.e.

Fig. 1. Diagram of the regular distribution of the principal (9) components: $\mathrm{SiO_2}$, $\mathrm{TiO_2}$, $\mathrm{Al_2O_3}$, $\mathrm{Fe_2O_3}$, $\mathrm{FeO}$, $\mathrm{MgO}$, $\mathrm{CaO}$, $\mathrm{Na_2O}$, $\mathrm{K_2O}$, $\mathrm{P_2O_5}$, $\mathrm{H_2O}$, composing intrusive rocks in the field of the Earth’s gravity.

In the field of the Earth’s gravity, magma is, as it were, stratified, and its total density increases from level to level from top to bottom (Fig. 1), a statement that requires no special explanation.

After the completion of magmatic differentiation, as a result of further cooling, the process of crystallization begins in the differentiated magma: an infinitely large number of centers of crystallization appear in it—the embryos of crystals—which for a certain time grow and increase in size. The appearance of a large number of crystals in cooling magma gives rise to complex forms of motion, the justification of which we find below.

In article [20] it was assumed that a dissolved substance in a solvent, according to the investigations of D. I. Mendeleev and his followers, generally forms a complex $K_0 = (K + M + Q)$, in which $Q$ is the heat

dissolution. A magmatic solution (melt) may be represented as consisting of a sum of complexes.

If, in a given supersaturated solution, crystal nuclei appear, then each of them creates around itself a diffusion region of nourishment (a crystallization courtyard) [7, 20], within which, throughout the entire process of crystal growth, there is continuous motion of the supersaturated solution toward the crystal. If we also assume that \(P(K+M+Q)\) of substance has diffused to the growing crystal, then from this amount \(qK\) will go to the construction of the crystal, and the remaining part \(\{(p-q)\cdot(K+M+Q)+q(M+Q)\}\), being lighter than the solution itself, rises above the crystal in the form of a vortex concentration flow [7, 8, 9, 17, 19, 20]^1). Similar torches [20] rise above each crystal separately.

The phenomena of crystallization diffusion [20] and concentration flows throughout the entire crystallization process are interrelated and mutually conditioned; we can eliminate them only by destroying the growing crystal itself [7, 9, 20]. From the moment of its appearance, this latter is the principal source giving rise to various kinds of disturbances in the solution, which before that had been in relative rest.

Experimental studies of concentration-flow phenomena [19, 20] have shown that, under conditions of mass crystallization, they play an extremely important role that should not be overlooked for the following reasons. First, as crystals grow, the concentration flow rises above each of them to a height more than 25 times greater than its transverse dimension. With the growth of a crystal, both the strength of the flow and the height to which it rises increase in accordance with its dimensions. Second, the crystallization flow rises from one crystal to another, a third, etc., located higher up, washes over them and merges with their flows and, as observations show, gradually increases its upward velocity. Third, the concentration flow, as a vortex, is capable in its motion of drawing in (sucking in) the supersaturated solution and thereby increasing its concentration. This circumstance fully provides more favorable conditions for enhanced growth of the crystals washed by the flows.

Fourth, crystallization at many levels, with constant and additional feeding of the saturated solution by dry salt, has shown [19, 20] that this salt passes into the solution, and the crystals of the upper levels increase in size more rapidly than crystals of equal size located below them. Such observations indicate that the concentration flows existing in the solution at the bottom of the crystallizer create the necessary conditions for dissolution of the added salt and, carrying it along, transport it upward and deposit it on crystals along the path of their motion.

Finally, concentration flows, by their motion, are capable of performing considerable work under conditions of mass crystallization; for example, they stir the solution: some of the streams move against gravity, while others descend downward. Nevertheless, the upward motion and transport of substance still predominate.

After this brief necessary digression, one may proceed to consider magma that is in the stage of crystallization differentiation. Magma, as stated above, is a sum of various—

^1) The release, during crystallization, of heat from the complex—\(Q\)—causes the appearance of convective motion, which enters as an obligatory part into the concentration flow.

of the chemical composition of complexes, which in general can be expressed as \((K+M+Q)\). In this complex, \(K\) denotes the substances corresponding to the composition of the rock-forming minerals; \(M\) represents the silicate part of magma of any composition, the mineralizers, and, finally, water, which, taken together, reacting with \(K\), lower its melting temperature and mutual solubility and make it possible, to some degree, for the complexes to remain in a liquid and mobile state.

The onset of crystallization, indicating a disturbance of the previously existing equilibrium state, depending on changes in pressure and chiefly in temperature, is accompanied by the successive decomposition of the complexes of the magmatic solution. This process may be represented schematically as follows:

\[ p(K+M+Q)\to\{(p-p_1)K+(p-p_1)(M+Q)+p_1(K+M+Q)\}\to \]

\[ \{(p_1-p_2)K+(p_1-p_2)(M+Q)+p_2(K+M+Q)\}\to(p_2-p_3)K+ \]

\[ +(p_2-p_3)(M+Q)+p_3(K+M+Q)\to\{(p_3-p_4)K+ \]

\[ +(p_3-p_4)(M+Q)+p_4(K+M+Q)\}\to\ldots \]

In this scheme of the course of crystallization, from the quantity \(p(K+M+Q)\) of magma complexes, a certain part \((p-p_1)K\), \((p_1-p_2)K\), \((p_2-p_3)K,\ldots\), upon decomposition of the complexes, separates into the solid phase and will successively favor the appearance and growth of nuclei—crystals of rock-forming minerals; \(p_1(K+M+Q)\), \(p_2(K+M+Q)\), \(p_3(K+M+Q)\ldots\) are successive residual magmatic solutions, from which the growing crystals draw the material necessary for their growth and from which the process of separation of the solid phase successively continues; \((p-p_1)(M+Q)\), \((p-p_2)(M+Q)\), \((p_2-p_3)(M+Q)\ldots\) are liquid phases which, together with the part unused in the crystallization of the magmatic solution, rise above each crystal. The number of the latter in the crystallizing magma is arbitrarily large; hence it is natural to expect that the concentration streams are capable, from the moment of crystallization to the moment of its completion, of accomplishing an enormous and very complex work of moving the substance still present in the liquid phase. The continuing decrease in temperature leads likewise to the formation of a gaseous phase, which, in its upward tendency, will join the concentration jets, increase, one must suppose, their velocity, and maintain their directionality.

It follows from what has been said that crystallization in a magmatic solution under conditions of cooling will tend to eliminate the continually arising contradictory state in the supersaturated solution by means of a relatively intensified growth of crystals and the separation of \((M+Q)\), carried along in the form of a concentration stream. It is natural to expect that the growth of an infinitely large number of crystals at the stage of crystallization differentiation can create a very long-acting, powerful movement of concentration jets and thereby contribute in the same direction to the transport of matter present in the liquid state.

In the movement of the magmatic solution at the stage of crystallization differentiation, two following aspects should be distinguished. The first of them is determined by the fact that the system may be closed. In this case, experimental studies have shown that the concentration streams, reaching the surface of the solution, spread along it in all directions. In experiments on mass crystallization, the concentration-

currents, for example, from eleven columns, the crystals of which were suspended one above another, likewise reached the surface, partly spread along the surface, partly descended downward between the columns and the concentration jets rising around them, and all together created complex vortex motions within the solution. This complexity of the motion of the solution was due to the fact that the relative strength of the concentration flows created by each column was not the same, and therefore the intensity of motion within the solution in all its parts was also not the same.

If we imagine that the concentration jets reach the contact of the crystallizing magma with the enclosing rocks, or the contact of still-liquid magma with the already solidified apical crust of the same magma, then the motion in the latter, created by the crystallization process, will have the same character as in the experiment, and can be schematically illustrated by Fig. 2, a.

Fig. 2. Concentration flows and the motions created by them at the stage of crystallization of magma:
a — motions of the magma in the case of a closed system; b — motions of the magma in the case of the appearance in the roof of various kinds of fractures.

In studying porphyritic granites and granodiorites, it is not infrequently necessary to note that porphyritic segregations of orthoclase are arranged in a series of parallel jets and form a flow that clearly indicates the character of currents in the forming magmatic rock before its complete solidification. In Fig. 3 an example is schematically shown of the jet-like distribution of porphyritic crystals of orthoclase in the Belokurikha granite massif (Gorny Altai)¹. We have an entirely analogous picture also in the rounded stock of the Atturkol rapakivi, 14 sq. km in area (Gorny Altai). This small massif was recently exposed from beneath the cover of enclosing rocks, and in this case there is least reason to speak of a general horizontal flow of magma. The jet-like distribution of crystals of orthoclase and oligoclase occurred with the participation of spreading jets near the surface of contact with the enclosing rocks. On vertical sections of the rapakivi stock there are also visible jet-like—

¹ This schematic sketch was obtained from Docent S. F. Dubinkina.

distributions of porphyritic segregations, indicating upward-directed movements in the magma. Similar examples are found in H. Cloos, E. Cloos, Balk, and their followers, who studied many granite massifs of Europe and America.

The second factor in the movement of magma at the stage of its crystallization is determined by the fact that, in the roof of the enclosing rocks and in the apical solidified part of the magma, there are fractures and fracture zones that may serve as pathways for the movement of silicate and hydrothermal solutions (Fig. 2, б). Concentration flows in the boundary region magma—enclosing rocks, on the one hand, bring about complex

Fig. 3. Schematic drawing of the distribution of porphyritic orthoclases in one of the sectors of the Belokurikha granite massif (Altai).

movement and mixing, and on the other hand, part of the upward-moving magma and the emanations associated with it, being under pressure and subject to the difference of pressures in the fracture zones, move upward along these zones and fractures.

Turning to the very composition of the magma at the stage of crystallizational differentiation, it is clear that this composition changes with time, as is also shown by the considerable experience in studying the history of formation of dike bodies in intrusions and deposits closely connected with the development of magmatic formations.

In the first stages of crystallization of magma, when the temperature is still relatively high and capable of maintaining the silicate part of the magmatic solution in a liquid and mobile state, the component part \(M\) of the system \((M+Q)\) is essentially silicate and contains nonsilicate products to a lesser degree. I refer to the fact that as early as Yu. V. Wulff [7, 9] asserted that the solution feeding a growing crystal gives only part of the dissolved substance to the crystal, while the remaining part of the solution, remaining somewhat supersaturated, rises above the crystal in the form of a flow, i.e., a part of the unused complexes \((K+M+Q)\) of the magma is added to the system \((M+Q)\). Further, concentration flows, as follows from experimental data, are capable of additionally absorbing magmatic solution. At this stage such flows have an essentially silicate character, and, penetrating into open fractures in the enclosing rocks and into the solidified apical part of the magmatic mass, they favor the formation of vein bodies composed of magmatic formations, mineralogo-

the petrographic composition of which depends on the depth of occurrence of the source of the concentration flows.

At the subsequent stages of magma development, as the temperature in the complexes of the still-liquid part decreases, nonsilicate constituents—including the ore substance—will play an ever greater role; these will be deposited in the form of vein bodies or will penetrate through the system of fissures into the enclosing rocks and modify them accordingly.

In the case of complete solidification of the magmatic chamber, crystallization differentiation is completed, the activity of the concentration flows ceases, and the movement of ore-bearing and other fluids caused by them comes to an end.

The Role of Concentration Flows in the Formation of Intrusive Rocks

In order to visualize the role of concentration flows in the formation of magmatic rocks, let us consider a number of phenomena in whose development, during the process of crystallization, they take an indisputable part.

Porphyritic phenocrysts in intrusive rocks are a very common phenomenon. Porphyritic segregations of feldspars, represented by minerals of the orthoclase subgroup, occur especially often.

The study of the form of crystal growth in the case of mass crystallization has shown that growing crystals of one and the same composition, under one and the same conditions, often display unequal rates of growth. In some cases the faces of crystals build up unevenly; in others, alongside the general growth of crystals, there are clear traces of dissolution of the entire face or of part of it; in still others, the overall rate of growth of the crystal and the growth rate of the crystal faces change from one period of growth to another; in fourth cases, crystals that, in comparison with others, are in relatively favorable or more or less favorable conditions of nutrition show a relatively high rate of growth, and in this case they attain relatively large sizes and acquire a more regular form.

Thus, one of the crystals in an experiment on mass crystallization, initially having, in comparison with the others, relatively small dimensions, had by the end of the experiment increased considerably and, in size, was among the largest crystals. Conversely, crystals that were in variable growth conditions increased their dimensions relatively slowly.

Finally, let us recall here once more that in our experiments crystals of the upper levels generally show a tendency toward better growth than crystals of the lower levels. Here it is necessary to decipher the “tendency toward better growth.” At each level of crystallization, along with crystals of enhanced growth, we have crystals with a reduced growth rate; that is, the former are in conditions relatively favorable for growth, the latter in less favorable ones. As the experiments showed, favorable and less favorable conditions are determined by the concentration flows developing above each crystal under conditions of mass crystallization.

If we turn to porphyritic phenocrysts, for example in granites and granodiorites, then among them we find phenocrysts of various sizes. In some cases the phenocrysts, in their dimensions, differ only slightly from the dimensions of the orthoclases that are part of the composition

of the main granular mass of the rock. In other porphyritic rocks, the porphyritic phenocrysts differ sharply in their size and form from the same minerals of the groundmass.

F. Yu. Levinson-Lessing’s assertion that porphyritic phenocrysts form the component that, in a given melt, is present in excess relative to the eutectic is not confirmed in the study of porphyritic igneous rocks. In order not to be unfounded, I shall cite a sharply contradictory example. In the porphyritic Atturkol rapakivi (Gorny Altai), the porphyritic formations are represented by two minerals—orthoclase and oligoclase. It is curious to note that if, in some areas, the porphyritic segregations are represented chiefly by alkali feldspar, then in the granular groundmass the grains of plagioclase sharply predominate over orthoclase, the amount of which is comparatively small. If oligoclase predominates among the porphyritic phenocrysts, then in the groundmass oligoclase is numerically inferior to orthoclase. And finally, there are areas in the Atturkol rapakivi in which orthoclase and oligoclase occur together in porphyritic phenocrysts and in the groundmass of the rock in approximately equal quantities.

N. P. Semenenko, in the field of migmatites and granites of the Ingul River, in the granite massif of the Verblyuzhka River, in coarse-grained porphyritic granites, notes that microcline (13.6%), despite the presence in the rock of 44.8% oligoclase and 24.7% quartz, forms only porphyritic segregations and is almost absent from the groundmass.

One could cite a number of other examples which are likewise in clear contradiction to F. Yu. Levinson-Lessing’s assertion.

In accordance with our experiments, porphyritic phenocrysts may be those crystals which, as indicated above, are under favorable conditions of supply delivered simultaneously by concentration currents and crystallization diffusion.

The porphyritic phenocrysts of porphyry-granites attract particular attention. Here too, the porphyritic phenocrysts are represented predominantly by alkali feldspars. When studied in thin section, they prove to be distinctly zonal. The inner part of such crystals, as is to be expected, has regular outlines. The outer rim appears partly mosaic and poikilitic. Its boundaries are irregular and sometimes festooned. This outer rim of the phenocrysts is an aggregate of intergrown grains-crystals, unequal in size, of alkali feldspar of the groundmass, directly surrounding the porphyritic phenocrysts and oriented spatially in the same way as they are. The grains of such a rim continue their growth with the participation of concentration currents and, coalescing, form a continuous zone which, in essence, is polycrystalline. During formation of the rim, grains of quartz and grains of plagioclase are captured.

The sharply expressed overgrowth of the porphyritic crystal of alkali feldspar by a rim indicates that the porphyritic crystals of the granite-porphyries began to form in the deeper part of the chamber, and then, as the magma was squeezed out, it carried them into higher, sharply hypabyssal conditions, where they are relatively rapidly overgrown by a polycrystalline rim, with the simultaneous formation of an uneven-grained groundmass.

Zonal plagioclases. The next example, vividly illustrating the role of concentration currents in mineral formation, will be zonal plagioclases, which are extremely widespread in rocks of magmatic origin.

Among zoned plagioclases one may distinguish simple zoned formations (single-stage) and complex ones (multistage).

The simple single-stage zoned plagioclases include plagioclases whose chemical composition changes smoothly, without visible abrupt jumps, from the central part, relatively rich in the anorthite molecule, to the periphery, toward which the content of the anorthite component in the successive zones gradually decreases, while the albite content correspondingly increases. To illustrate what has been said, let us give at least two examples.

I. The center of the plagioclase corresponds to No. 37, the first intermediate zone corresponds to No. 33, the second to No. 27, and the outer zone to No. 22.

II. The center of the plagioclase is No. 33, its first intermediate zone is No. 27, the second is No. 22, and the outer zone is No. 18.

The complex multistage zoned plagioclases include plagioclases in whose structure several single-stage zoned zones, superposed one upon another, take part. The complexity of such zoned plagioclases is extremely varied even within one and the same thin section. Let us give the following examples:

I: No. 47 (in the center of the grain) | No. 52—47 | No. 62—47 | No. 62—42 | No. 57—42 | No. 52—37 | No. 47—33 | (at the periphery) No. 42—27 (Fig. 4, a).

II: No. 52 (in the center of the grain) | No. 67—52 | No. 65—52 | No. 62—47 | (at the periphery) No. 57—42.

III: No. 37—33 (in the center of the grain) | No. 45—37 | No. 45—33 | No. 42—33 (at the periphery) No. 37—27 (Fig. 4, b).

From the examples of complex zoned plagioclases given, it is clear that the central part of the plagioclase—the core—is surrounded by a series of successively superposed zoned shells; the zoned zones of each of them change smoothly from a higher plagioclase number to a lower one. The transition from the core to the zoned shell surrounding it and from one

Fig. 4. Zoned plagioclases.

the shell—for example, No. 62–47; toward the adjacent one, No. 62–47, etc., separated by a vertical line, it clearly has a step-like character.

The boundary between zonal stages in zoned plagioclase is sharply expressed, since the transition from one zonal stage—the inner one—to another—the outer one—is caused by the transition from a zone with a lower plagioclase number to a zone with a higher plagioclase number: the greater the difference in the content of the anorthite molecule in adjacent zones, the more sharply the transition (jump) from one stage to another is expressed. The distinctness of the replacement of one zonal zone by another is also emphasized, to a greater or lesser degree, by the dissolution of each preceding zone before the deposition of a new zone with an increased content of the anorthite molecule. In Fig. 4, a, it is clearly seen that, in a multistage zoned plagioclase, its core is composed of plagioclase No. 47; on its dissolved surface a zone of plagioclase No. 62 is deposited, upon which zones of plagioclases Nos. 59, then Nos. 55, 52, and 47 are superposed with a gradual transition—that is, the composition of the superposed zones changes in the direction of a gradual decrease in the content of the anorthite molecule. In the next stage, a zone of plagioclase No. 62 is deposited on the “blurred” surface of zone No. 47. After zone No. 62, zones Nos. 57, 52, 47, etc. are deposited successively.

In essence, zonal-sequential growth of crystals is characteristic of all growing crystals, but this feature cannot always be established by ordinary methods without resorting to etching methods. In individual cases, however, this phenomenon is quite obvious. For example, porphyritic orthoclases (microclines) in Transbaikalian porphyritic granites in a number of cases reveal one or two, and sometimes even three, zones of accretion. In this case, before the deposition of a new zone, the surface of the preceding one undergoes partial dissolution, and on the dissolved surface plates of biotite, biotite flakes, and quartz crystals are deposited, or small oligoclase crystals are attached to them. On the dissolved surface of orthoclase and on the foreign minerals lying on it, subsequent zones of alkali feldspar are superposed, as a rule not disturbing the general homogeneity of the included crystal.

In the Atturkol rapakivi-like granite massif, before the deposition of new orthoclase zones, prismatic crystals of oligoclases grow onto its dissolved surface.

Zoning of quartz is rather often observed in its head, owing to repeatedly recurring partial dissolutions of the rhombohedron faces and the growth of new zones upon them.

The zoning of titaniferous augite in gabbroic and diabase rocks is clearly distinguished by the intensity of coloration, which in a number of cases permits one to say that each successive zone is deposited on a partially dissolved preceding surface of crystal growth.

The zoning of plagioclases is usually explained and experimentally confirmed in accordance with Roseboom’s thermal diagrams and the experimental data of A. L. Day, E. G. Allen, and N. L. Bowen. But since in magma, at the stage of crystallization differentiation, its liquid part is in constant motion, in which concentration flows take a large part—flows which, as stated above, necessarily develop above each crystal separately.

To explain the formation of zoned plagioclases, let us imagine two levels of their formation, which are characterized by the physico-chemical conditions proper to them: \(P_1T_1C_1\) and \(P_2T_2C_2\). Within these

levels in the field of the Earth’s gravity the overall density of magmatic solutions decreases from bottom to top, and, consequently, the composition \(C_1\) and \(C_2\) of the magma changes regularly from one level to another. Hydrostatic pressure increases with depth, \(P_2>P_1\). The temperatures \(T_1\) and \(T_2\) at the given levels may differ little from one another. On the whole we may consider that the general physico-chemical conditions at one level and at the other are somewhat different and are capable of favoring the deposition, in qualitative and quantitative respects, of different mineral associations. Among them, it seems to me, under the given conditions plagioclases will be especially sensitive.

Under the conditions of crystallization just indicated, at the lower level, from \(K_{02} = (K_2 + M_2 + Q)\), plagioclases of a definite composition will precipitate in accordance with the experimental data of A. L. Day and E. G. Allen. At the upper level, with the corresponding physico-chemical data, from the complex \(K_{01} = (K_1 + M_1 + Q)\), in comparison with what has just been said, plagioclases with a lower content of the anorthite molecule will be deposited in the solid phase. The plumes of concentration currents rising above each crystal of the lower level carry upward an albite-molecule-enriched solution from the crystallization courtyard; part of \((K_2 + M_2 + Q)\), not completely used by the crystal for its growth, and during its motion the concentration current draws in a certain volume of supersaturated magmatic solution, whose composition is now somewhat enriched in albite molecules. The averaged substance is deposited upon the nuclei of the upper level in the form of a zone with a lower anorthite content.

If the process of growth of plagioclase crystals proceeds under calm conditions, then upon the initially formed plagioclase crystal of a definite number there is later superposed, as a result of the activity of concentration currents, the next zone of plagioclase substance with a somewhat lower content of the anorthite molecule; upon it, in turn, another, a third, a fourth, etc., zones are deposited, in which the content of the anorthite molecule successively and smoothly decreases and, conversely, the content of the albite molecule increases. As a result we obtain a single-stage zonal plagioclase, in which the chemical composition changes regularly and smoothly from the center to the periphery.

An entirely different picture in the structure of zonal plagioclases will be obtained in the case when their formation develops in an environment in which periods of quiescent crystallization are repeatedly interrupted, and the gradual change of composition from the center to the periphery, characteristic of zonal plagioclases, is disturbed and becomes more complex, multistage. The reason for the disturbance of the smooth development of the process of formation of zonal plagioclase lies in the fact that the conditions of growth of such plagioclases change discontinuously and intermittently.

As we have seen, the deposition of each successive zonal stage is preceded by uneven dissolution of previously deposited zones. This circumstance is connected with the fact that the temperature of the solutions washing the growing crystals rises, which corresponds at the same time to solutions with a more elevated content of anorthite substance in them. The appearance of new, qualitatively different solutions at a new stage of growth of zonal plagioclases ensures the deposition of a new zone of a higher plagioclase number in comparison with the preceding one.

Such an abrupt change in physico-chemical properties should be explained as a consequence of cooling that develops in jumps. This circumstance also generally causes, from time to time, the appearance of additional supersaturation, which ensures additional acceleration of crystallization and an increase in the rate—

of the motion of concentration streams. These latter cause disturbance in deeper horizons of the magma and the entrainment upward of magmatic solutions containing plagioclase material with a greater concentration of anorthite than at the given level.

Each such disturbance in the general course of magma formation causes concentration streams to carry into higher horizons of the rocks being formed solutions rich in anorthite molecules, which then ensure deposition, on the dissolved surface of zoned plagioclase, of a zone with a higher anorthite content.

Albite rims and albitization. In granites and granodiorites the phenomenon of albitization can be observed quite often. Usually albite in the indicated rocks appears around plagioclase grains in the form of an outer zonal rim. But in a number of cases the albite rim is sharply separated from the rimmed grains of rock-forming minerals, including plagioclase. In the Atyurkol rapakivi-like granite, albite rims often simultaneously completely or partially surround adjacent grains of plagioclase, or of plagioclase and orthoclase.

In the case of cataclasis, albitization of alkali feldspars and plagioclases may be observed, as a consequence of which such rocks as granites and syenites may acquire the character of albitites. In the Zavitinsk deposit, for example, pegmatites, after their crushing, undergo albitization to varying degrees and diverse mineralization associated with it.

Fields of development of skarns and associated iron and copper deposits in the Kuznetsk Alatau, as a rule, are accompanied by a broad development of metasomatic hornfelses of quartz-albite composition, which are formed before the development of the skarn formations proper and which in a number of cases lead to the creation of pseudoporphyritic rocks (diorite-porphyrites of the Kaz iron deposit).

Finally, quartz-albite veins in the Kuznetsk Alatau, and in particular in Gornaya Shoria, are widespread formations.

All this, taken together, indicates that from the depths of the magmatic rocks being formed into the apical parts of the solidified intrusion and into the enclosing rocks, along a system of fractures and zones of crushing, concentration streams carry out, in enormous quantities, complexes of albite or albite-quartz, which in the part of the Earth’s crust accessible to us, exposed by erosion, manifest themselves in diverse forms of occurrence.

Reaction rims. This group of formations includes rims, for example, surrounding olivine or rhombic pyroxenes. Olivine is surrounded on all sides or partially by a rim: a) of rhombic pyroxene; b) of rhombic pyroxene and the augite pyroxene overlying it; c) of rhombic pyroxene and the hornblende overlying it; d) of augite and amphibole; e) of amphibole; f) of amphibole and biotite. In some cases the formation of the rims just indicated is explained as a consequence of the interaction of two adjacent minerals in the region of their contact, which leads to the formation of new minerals. In other cases it is assumed that similar reaction rims arise as a consequence of the interaction of a mineral with a liquid magma.

It cannot be said that the existing explanations of rim formation are fully acceptable, for the following reasons. First, alongside minerals bearing rims, there are the same minerals lacking such rims, despite the fact that both minerals occur under identical conditions of formation. Second, often in one and the same schlieren, rims, for example, around olivine grains enclosed in

in the plagioclase, have different compositions: a) rhombic pyroxene, b) rhombic pyroxene and biotite, c) rhombic pyroxene and augite, d) augite pyroxene and amphibole and biotite, e) biotite. Thirdly, at the contact with the rims the plagioclases themselves show no changes: the boundary between the plagioclase and the rim is usually sharp.

The contradictions just noted make it possible to propose somewhat different explanations for the formation of rims.

From the existing views that the formation of rims occurs at the stage of crystallization differentiation, it is natural to suppose that at this stage concentration currents play a considerable role in the growth of crystals; as noted above, these carry to higher horizons not only \((M+Q)\), but also the \(K_0\) present in the solution—\((K+M+Q)\). In the present case \(M\) corresponds to partly dissolved silicates of magnesium, iron, calcium, and potassium and contains various mineralizers. Concentration currents, rising upward, wash, for example, olivine along their path; with a relatively high content of \((\mathrm{Mg}, \mathrm{Fe})_2[\mathrm{SiO}_4]\) in the solution, the medium will favor its growth, but if the composition of the washing solutions proves enriched in silica, then, naturally, a rim of rhombic pyroxene will be deposited on the olivine, as in zonal plagioclases. Further, if in the incoming solution, besides an increase in the concentration of silica, there is also an increased content of aluminum and calcium, since the growth of the plagioclases has not yet ended (is continuing), then in such a case the rim of rhombic pyroxene is replaced by an augite rim, and the supply of potassium, aluminum, and water ensures the formation of a rim of amphibole and biotite.

Thus, on the basis of what has been said, it seems that the foregoing makes it possible, more simply and more naturally, proceeding from the process of crystallization, to explain the formation of the very frequently encountered rims around particular minerals. Of course, one cannot deny the manifestation of reactions at the contact of two media, but this question, for the magmatic rocks themselves, cannot be considered clear.

The structures of magmatic rocks, chiefly intrusive ones, in a number of cases are more easily explained from the positions that we used to explain other phenomena developing during the period of formation of magma.

Usually the sequence of separation of minerals in the process of formation of intrusive rocks is determined on the basis of the relative idiomorphism of the rock-forming minerals, that is, the earlier a crystal separated, the more perfect its idiomorphism and the more sharply it is subjected to corrosion by minerals that separated later. However, on the question touched upon there are two opinions. F. Yu. Levinson-Lessing, illustrating the sequence and conditions of crystallization in the case of formation of hypidiomorphic-granular structural relations of hornblende, plagioclase, orthoclase, and quartz, says that “here it is only a matter of the sequence that determines the beginning of crystallization of each mineral; in reality there are moments when two or even three minerals crystallize simultaneously” [22].

A. N. Zavaritskii, on the same subject, says that the generally accepted statements about the order of separation of minerals in intrusive rocks “cannot be accepted without reservations. Idiomorphism does not indicate the order of the beginning of crystallization, but only partly indicates the order of the end of crystallization of individual minerals. Inclusions of one of the minerals in another are older only than the outer part of this mineral, but its inner part could have formed even earlier than these inclusions.”

What has just been said is explained by A. N. Zavaritskii with the following considerations: “Inclusions of an idiomorphic primary mineral \(A\) in mineral \(A\) indicate that mineral \(A\) began to crystallize earlier than the surrounding outer part \(B\), but the inner part of mineral \(B\) could also have formed earlier. Crystals \(A\), however, may continue to be released later” [12, 13].

Vogt also speaks of the possible simultaneous presence in the magma, during crystallization, of several nuclei of rock-forming minerals.

If it is assumed that supercooling of the magma before the beginning of crystallization is its usual state, then it is natural to suppose that its principal substances are in a supersaturated state and that, under these conditions, they are capable of forming in it almost simultaneously nuclei which will then grow. The rate of their growth will depend on the difference between the melting temperature and the temperature of the magma at the stage of crystallization. Further, the rate and duration of crystal growth also depend on the amount of one or another substance in the melt, on the degree of subsequent cooling of the magma, and therefore on the degree of supersaturation, diffusion, and concentration currents that ensure the growth of crystals.

We have mentioned more than once that the decomposition of the complexes \(K = (K + M + Q)\), composing the magma, on the one hand promotes the growth of crystals, and on the other—the nucleation of new centers of crystallization. The complexes are delivered to the crystals as a result of crystallization diffusion, and, together with the decomposition products \((M+Q)\), the concentration currents also bring the complexes themselves.

It is not superfluous to recall that the composition of the complex changes in character depending on the temperature state of the crystallization stage of the magma. At the first stage of crystallization, when the temperature is high and sufficient for the silicate constituents of the magma to be in solution, the composition of \(M\), besides mineralizers and other products, contains a relatively high percentage of silicates and silica. With a decrease in temperature the composition of \(M\) changes toward a decrease in the content of the silicate part and an increase in the solution of silica, carbonates, carbon dioxide and water, ore and other components.

At the level of formation of granitic rocks, crystallization diffusion and concentration currents will favor the simultaneous growth of crystals of plagioclase, orthoclase, and quartz. A certain idiomorphism in the clearly expressed prismatic form of the plagioclases indicates that diffusion and concentration currents exhausted the substance necessary for the growth of plagioclase in solution before the completion of the growth of orthoclase and quartz, whose growth continued. In some granites it is quite possible to see that this growth continues with partial dissolution of plagioclase, and sometimes also with partial replacement of plagioclase by orthoclase material.

Quartz, as a rule, is xenomorphic; its growth ends somewhat later than that of orthoclases, which, like plagioclases, are partially dissolved, giving way to silica. The movement of concentration currents ceases as soon as the intergranular spaces are filled with the substance of the rock-forming minerals. Only subsequent fracturing can open paths for the possible movement of gases and solutions released from deeper parts of the magma, where the crystallization process is still continuing.

Another example of structural features that may attract attention is the formation of a poikilitic or poikilophi-

of the structure of gabbroic and diabase rocks. A. N. Zavaritskii explained its formation as the result of recrystallization of the rock during its further cooling. It seems to us that such structures are much more naturally explained with the participation of crystallization phenomena in the magma, when concentration flows in places bring in more pyroxene material than plagioclase material. Poikilitic structure in gabbroic and diabase rocks develops, as investigations show, locally, and all gradual transitions can be found from a typically gabbroic to a typically poikilitic structure within one and the same hand specimen, a circumstance that does not permit support for A. N. Zavaritskii’s opinion on this question.

Acquaintance with the traps (dolerites) of the Central Siberian Platform and Kuzbass shows that prismatic grains of basic plagioclases often form a framework structure through which the remaining part of the liquid magma is able to move. This latter ensures the filling of the free interstices in the framework: a) with a granular aggregate of titaniferous augite, b) with a single grain of the same pyroxene, c) one pyroxene grain, as it grows, fills several adjacent interstices in the framework structure and acquires the character of a poikilite. In this latter case the poikilitic “inclusions” of plagioclase in pyroxene often appear as fragments of the framework.

Phenomena of autometamorphism in igneous intrusive rocks develop during the period of their formation and constitute changes in the composition of rock-forming minerals associated with the conditions under which the intrusive rock itself comes into being. The essence of autometamorphism is reduced to reactions between previously formed minerals composing the rock in the apical part of the intrusion and gases and vapors released from the magmatic melt in the process of continuing crystallization. But this process, under the influence of volatile gases and fluids rising from below upward out of the magmatic chamber at the stage of continuing crystallization differentiation, has a metasomatic character and leads to partial or complete alteration of individual areas of intrusive rocks and their minerals by transformation in the presence of additional material brought out of the magma. Such autometamorphic transformations must include, for example, silicification, albitization, chloritization, sericitization, epidotization, biotitization, amphibolitization, and other phenomena of the same order.

According to F. Yu. Levinson-Lessing, the phenomena of autometamorphism enter into the cycle of processes of which the genesis of a given rock is composed during the period of formation of intrusive bodies as a whole.

The mechanism of the development of autometamorphism in magmatic rocks consists of the following points:

1. This phenomenon is, as a rule, preceded by deformation, expressed in varying degrees of crushing.

2. The removal of gases and fluids into the apical, solidified but still hot part of the intrusive body from the underlying depths of the magmatic chamber, in which the process of crystallization continues.

3. Reactions proceed between the minerals of the rock and the gases and fluids being carried out, leading to partial or fundamental transformations of the substance of the former minerals.

4. Alongside this, the process of filling cracks proceeds, and a metasomatic partial or complete replacement of the primary composition of intrusive rocks by new mineral formations develops to varying degrees.

There is no need to say that the concentration flows and the motions of gases and solutions produced by them through the hot, but already solid, apical part of the intrusive play a leading role in the development of the phenomenon of autometamorphism.

Hydrothermal Solutions

There is much evidence that hydrothermal solutions are genetically closely connected with intrusive rocks; this evidence may be found in L. Greighton, A. G. Betekhtin, F. I. Wolfson, and many others [2, 6, 10, 25].

It was indicated above that, at the stage of crystallization differentiation, concentration flows carry into the upper horizons of the magma $(M+Q)$ and the complexes $K_0=(K+M+Q)$ present in solution. The composition of $M$ depends on the stage of formation of the intrusion. As was stated above, in the first stages, when the temperature of the melt in the magma still remains relatively high, silicate components play a large role in the composition of $M$; these are capable of being in a dissolved state. In addition, mineralizers and the constituents of ore solutions are present in the composition of $M$. The concentration flows $(K+M+Q)$ entrained and drawn from the magma at these same first stages of formation are also essentially silicate.

At later stages, when the process of formation takes place at relatively lower temperatures, the role of silicate formations in the composition of the $M$ complexes is considerably reduced. Instead, silica, carbonates, haloids, carbon dioxide, sulfur, and ore components acquire predominant significance in the solutions carried out by concentration flows. The composition of the complexes $(K+M+Q)$ themselves also changes in the same direction.

In confirmation of the considerations expressed, one may cite examples of the close connection of quartz-tin and quartz-tungsten deposits with intrusions contaminated by tin and tungsten. Such examples are well known for the Far East, Transbaikalia, Kazakhstan, and other territories of the USSR. In the present example we see that, at the first stages of formation of the intrusion, tin and tungsten enter into the composition of the $M$ complex, are carried into the upper horizons, and, during crystallization of the apical part of the magmatic chamber, are captured by the forming minerals or form independent minerals in the form of cassiterite and wolframite, usually occurring in a dispersed state.

At later stages the tin and tungsten carried out by the concentration flows, entering zones of fractures, unite into common flows moving toward the surface of the earth (Fig. 2, б) and contribute to the formation of quartz-tin and quartz-tungsten veins.

The very formation of fractures and fracture zones and their repeated renewal, as well as their repeated filling by mineral associations, indicates that the movement of emanations along them is created not only by concentration flows and not only by the necessity conditioned by the field of terrestrial gravity, but that this movement of solutions does not occur: a) without the participation of the pressure produced by the movement of the roof, which tends to eliminate the disturbances in hydrostatic relations between the magma and the roof that arise as a result of crystallization differentiation; b) without fluctuations in the stresses, temporarily created in the magma by internal pressures as a result of the disintegration of complexes during crystallization; c) without external pressures, now increasing and now weakening, created by folding stresses on the part of the enclosing rocks. Thus, the process of formation of hydrothermal formations is complex, and the causes of the movements of emanations, con-

connected with the process of crystallization of the magma, and by the force of gravity, are superimposed processes that cause the appearance, from time to time, of pressures capable of accelerating and renewing the movement of emanations from the crystallizing magma toward the surface along the cracks and fracture zones created by them both in the magmatic rocks themselves and in the enclosing rocks.

As for the place of localization of hydrothermal deposits in the field of magma formation, this question appears to us as follows. As is clear from practice, on the surface of an intrusive body there are domes of various size and form. In the first stages of emanation, those rising upward as a result of the action of concentration currents may accumulate in greater mass in dome-shaped swellings (Fig. 2,a), and the cracks arising here will facilitate their movement in the enclosing magma of the rocks to higher horizons (Fig. 1,b). At the next stage, in the apical part of the magmatic chamber, a shell of intrusive rocks of granitic composition is formed. The continuing process of crystallization releases new portions of emanations, which likewise will accumulate in dome-shaped protrusions, while the cracks arising in the apical part of the intrusion will provide paths for the movement of solutions and for the formation of fields of greisenization, ore concentrations in the form of veins, and other, more complex bodies.

Conclusion

In the preceding pages the author has attempted, on the basis of experimental studies of mass crystallization, to consider from new positions a number of questions relating to the formation of magma at the stage of crystallization differentiation. The study of mass crystallization from supersaturated solutions under rapid cooling convinces us that, before the beginning of crystallization, the supersaturated solution rapidly, as it were, stratifies, and its density increases from top to bottom [15, 16, 18]. This gave firm grounds for assuming that, by the beginning of crystallization differentiation, the magma had already undergone magmatic differentiation and that its component constituents had been regularly distributed in it in the way that follows from the figure given above (Fig. 1). Movement at this stage was completed mainly by the distribution of the substances composing the magma according to the hypsometric law.

At the stage of crystallization differentiation of magma, complex movements also develop, which are generated by the appearance of an infinitely large number of crystals. This circumstance must play an enormous and diverse role in the history of magma formation. The latter follows from those considerations that, first, crystallization diffusion is a necessity for crystal growth, providing nourishment for the growing crystal, which, however, does not completely use the solution material delivered to it. Secondly, another necessity for crystal growth is the emergence above each of them of concentration currents. My experiments have shown that concentration currents play an enormous role in the life of crystals and in mass crystallization; they are capable of carrying dissolved substance from below upward and, together with crystallization diffusion, ensure the growth of crystals, which in the process of crystallization takes place in the dynamic setting of complex movements of the magmatic solution [16, 18, 19, 20].

Experimental studies of mass crystallization thus made it possible to clarify the nature of movements in magma at the stage of crystallization differentiation, the role of the crystallization process

in this motion, in the transport of matter and in the formation of certain textural and structural features of rocks, the formation of porphyritic phenocrysts and zoned plagioclases, and a number of processes known under the general name of autometamorphic.

Of course, a whole series of questions connected with the formation of magma has remained unilluminated, but the author did not set himself this task in the present article. The author wished to draw the attention of researchers to the circumstance that the infinitely large number of crystals that have arisen in magma and are growing in it can perform enormous work, something that until now has been overlooked.

LITERATURE

1. Betekhtin A. G. Some considerations on the causes of motion of hydrothermal solutions. Trans. All-Union Mineralogical Society, issue 1, 1952.

2. Betekhtin A. G. On the genetic connection of hydrothermal formations with intrusions. Basic Problems in the Study of Magmatogenic Ore Deposits. Moscow, 1955.

3. Betekhtin A. G. On the causes of motion of hydrothermal solutions. Basic Problems in the Study of Magmatogenic Ore Deposits. Moscow, 1953. Second ed., 1955.

4. Bowen N. L. The Evolution of the Igneous Rocks. Moscow, 1934.

5. Bowen N. L. A brief general history of magmatic differentiation. Geology of Ore Deposits. Transactions of the United States, Moscow, 1937.

6. Wolfson F. I. Problems in the study of hydrothermal deposits. Moscow, 1952.

7. Wulff G. V. On the question of the rates of growth and dissolution of crystal faces. Warsaw University News. Warsaw, 1895—1896; Selected Works on Crystallography. Moscow, 1952.

8. Wulff G. V. How crystals grow. Priroda, 1915.

9. Wulff G. V. Crystals, Their Formation, Appearance, and Structure. Moscow, 1917.

10. Graton L. The nature of the ore-forming fluid. Moscow, 1946.

11. Daly R. O. Igneous Rocks and the Depths of the Earth. Moscow, 1936.

12. Zavaritskii A. N. Physicochemical Foundations of the Petrography of Igneous Rocks. Moscow—Leningrad, 1926.

13. Zavaritskii A. N. Igneous Rocks. Moscow, 1953.

14. Kuzmin A. M. Microcline (An attempt at a systematics of alkaline feldspars). Publ. TPI, vol. 74, issue 1, Tomsk, 1953.

15. Kuzmin A. M. Experiments on the simultaneous crystallization of a large number of crystals. Scientific-Technical Conference of the Tomsk Industrial Institute, 1940.

16. Kuzmin A. M. Mass crystallization at many levels. Publ. TPI, vol. 63, issue 2, 1950.

17. Kuzmin A. M. Mass crystallization and the associated idea of the physical state of solutions. Interuniversity conference in Tomsk on solid dielectrics and semiconductors, 1958.

18. Kuzmin A. M. Mass crystallization. I. Crystallization from supersaturated solutions at many levels under rapid cooling. Publ. TPI, vol. 95, 1958.

19. Kuzmin A. M. Mass crystallization. II. Crystallization at many levels with constant feeding. Publ. TPI, vol. 95, 1958.

20. Kuzmin A. M. The phenomenon of concentration flows during crystallization. Bull. Siberian Branch, Acad. Sci. USSR, No. 6, 1958.

21. Levinson-Lessing F. Yu. Works of S. F. Zhemchuzhnyi relating to problems of mineralogy and petrography. Selected Works of Acad. F. Yu. Levinson-Lessing, vol. III, 1952.

22. Levinson-Lessing F. Yu. Petrography. Selected Works of Acad. F. Yu. Levinson-Lessing, vol. IV, 1955.

23. Levinson-Lessing F. Yu. and Zhemchuzhnyi S. F. Porphyritic structure and eutectics. Publ. St. Petersburg Polytechnic Institute, 1906, vol. 5, issue 1—2; Selected Works of Acad. F. Yu. Levinson-Lessing, vol. I, 1949.

24. Lovits T. E. Experiments explaining various phenomena that occur during the preparation of salts. Technological Journal, vol. 2, part 1, St. Petersburg, 1805; “Classics of Science.” Selected Works on Chemistry. Moscow, 1955.

25. Lindgren V. Mineral Deposits, vol. I, Moscow, 1934.

26. Mendeleev D. I. Principles of Chemistry, vol. I, Moscow, 1947.

27. Mendeleev D. I. Solutions, vol. IV, Moscow, 1937.

28. Rosenbusch H. Descriptive Petrography. Moscow, 1934.

29. Semenenko N. P., Siroshtan R. I., Stepanets V. D. The field of migmatites and granites of the Ingul River, Kiev, 1954.

30. Graton L. C. Nature of the ore-forming fluid. Econ. Geolog., No. 2, 1940.

1. The diagram is constructed on the following basis. On the ordinate are plotted the mean densities of magmatic rocks, taken from G. Rosenbusch [28] and R. O. Daly [11]; on the abscissa line, at the level of a given density of rocks, segments are laid off proportional to the percentage content of the corresponding oxides. The sum of the segments on the abscissa line is equal to the sum $100\%$: $\%\,\mathrm{SiO_2} + \%\,\mathrm{TiO_2} + \%\,\mathrm{Al_2O_3} + \cdots = 100\%$. ↩