38o 



NATURE 



[February i6, 1899 



limits of existence of the following hydrates within -33°'6C. 

 and 186* C. (the temperature at which water begins to decom- 

 pose magnesium chloride) have been investigated : — 



MgCI.,.i2H,0, MgCI.,.8aH.,0, MgCl., 8«II.,0, MgCl., 6H.,0, 

 MgCl.,.41 1,.0. MgCl.,. 2H,0. 



It is impossible within the limits of this notice to discuss the 

 details of the investigation, but the brief indication here given 

 of the nature and scope of the inquiry may serve to direct 

 attention to a research which is obviously of wide interest. 



THE NATURAL HISTORY OF CORDIERITE 



AND ITS ASSOCIATES} 

 'T'HE last quarter of the present century has witnessed an 

 extraordinary outburst of petrological activity, due, in a 

 large measure, to the application cf precise mineralogical 

 methods to the study of the constituents of rocks. The petro- 

 logist, and through him the geologist, owes, therefore, an 

 enormous debt of gratitude to the mineralogist ; at the same 

 time, the benefits have not been wholly one-sided. Mineralogy 

 is becoming something more than a mere catalogue of the 

 crystallographic, chemical and physical characters of museum- 

 specimens, and this is largely due to the influence of petrology. 

 It may end in breaking down the artificial systems of classifica- 

 tion which are in vogue, and introducing others more in accord- 

 ance with genetic principles. 



A good illustration of the advantage of studying minerals 

 from the natural history point of view may be obtained by 

 considering some facts relating to the modes of occurrence and 

 origin of corundum, spinelle, sillimanite and cordierite — four 

 minerals which are so frequently found together that they 

 have been called the "faithful companions." Corundum is 

 crystallised alumina (Al.,03), true spinelle is an aluminate of 

 magnesia (MgO.AUO.,), sillimanite is the silicate of alumina 

 (Al203.Si02), and cordierite is a silicate of alumina and 

 magnesia (2MgO.2Al.jO3.5SiO.>). The mutual replacing pro- 

 perties of ferrous oxide and magnesia, and of ferric oxide and 

 alumina complicate the composition of the spi nelles and cordi- 

 erite. All the minerals contain alumina, and it is this fact 

 which determines their paragenesis. They occur, usually in 

 combinations of two or more, under the most diverse geological 

 conditions : — 



( 1 ) As the constituents of foliated crystalline rocks of more or 

 less doubtful origin. 



(2) As the products of conlact-metamorphism round plutonic 

 masses. 



(3) As the constituents of inclusions in plutonic rocks, dykes, 

 lavas and agglomerates. 



(4) As the direct products of the crystallisation of igneous 

 magmas. 



(5) As the direct products of the crystallisation of artificial 

 silicate-magmas. 



Cordierite-gneisses are found in many parts of the world in 

 association with biotite-gneisses and other foliated crystalline 

 rocks. \'arious views have been expressed as to their origin. 

 Some petrologists are content to refer them to the Archaean 

 system ; others regard them as due to the contact or thermo- 

 dynamic metamorphism of ordinary argillaceous sediments ; 

 and others as rocks of mixed origin, containing both igneous 

 and sedimentary material. The last view, as applied to certain 

 members of the group but not to all, derives support from the 

 fact that where cordierite-rocks occur as contact products, they 

 always belong to the inner zone, and sometimes give distinct 

 evidence of the intimate intermixture of igneous and sedimentary 

 material. 



Cordierite-rocks, often containing sillimanite and a green 

 spinelle, have been recognised, iluring the progress of the 

 Geological Survey, at many points in the Southern Highlands 

 of Scotland, in the counties of Aberdeen, Banft', Forfar and 

 Argyle, and quite recently corundum has been detected in some 

 of these ; so that the list of the " faithful companions" is now 

 complete so far as Scotland is concerned. It is doubtful at 

 present whether all the Scottish cordierite-rocks are of the same 

 age and mode of origin. Some are contact-rocks, but others 

 may, for the present at least, be more safely classed with the 

 older crystalline-schists. All are undoubtedly the result of the 

 metamorphism of highly aluminous rocks. 



* Abstr.ict of (lie presidential address delivered to the Geologist:,' 

 ■ n, by J. J. H. Teall, F.R.S., on Februar)- j. 



A very interesting case of the occurrence of all four minerals 

 in rocks due to contact-action has been described by Salomon, 

 It occurs in the southern part of the Eastern .•Mps round the 

 great mass of tonalite, of which Monte Adamello forms the 

 culminating point. 



Inclusions, derived either from a contact-zone or from the 

 crystalline-schist formation, containing two or more of the 

 minerals in question, have been observed in igneous rocks 

 occurring under the most diverse conditions in many parts of the 

 world. They have been found, for example, in the tonalite of 

 Monte .-Xviolo ; in the kersantite-dyke of Michaelstein in the 

 Hartz ; in the andesitic lavas of the Eifel, the Siebengebirge and 

 the south-east of Spain ; and, finally, amongst the ejected 

 blocks of the Laacher See and .\sama Varna in Japan. There 

 is evidence, moreover, that in most cf these cases the minerals, 

 or some of them, occur not only as constituents of the inclusions, 

 but also as the direct products of crystallisation from the igneous 

 magmas. Thus, in the mica-andesite of Hoyazo(Cabo di Gata) 

 cordierite occurs in two forms: (I) as irregularly bounded 

 grains up to the size of a hazel-nut, and (2) as sharply defined 

 idiomorphic crystals in a glassy base. The former are inclusions ; 

 the latter are crystals which have separated from the magma. 

 Rock-fragments, consisting very largely of a cordierite-gneiss 

 from which the isolated grains of cordierite have been derived, 

 are also very common in this andesile. Osann, who has 

 described this very interesting case, points out that the abun- 

 dance of indigenous cordierite, coupled with the presence of 

 numerous inclusions of cordierite and cordierite-gneiss, points to 

 the conclusion that portions of the foreign rock have been 

 dissolved, and that a magma of exceptional composition has 

 thus been formed, out of which cordierite has crystallised. 

 Many other cases are known in which the solution of foreign 

 aluminous material ha_s so modified a magma that members of 

 the group under consideration have crystallised out of it. 

 Moreover, it is not necessary that the minerals should be present 

 in the foreign material. It is sufficient that the necessary 

 chemical constituents should be present. Thus a basalt from 

 Kollnitz in Carinthia has involved fragments of an argillaceous 

 rock, and partially dissolved them. The normal basalt is holo- 

 crystalline, but in the neighbourhood of the inclusions it becomes 

 glassy, and crystals of spinelle and cordierite, which are absent, 

 both from the basalt and the inclusion, occur. The partial 

 solution of the fragments evidently modified the composition of 

 the basalt, so that it cooled as a glass after cordierite and spinelle 

 had separated out. It is interesting to note, in [passing, that the 

 addition of alumina to the basaltic magna has tended to prevent 

 crystallisation. This effect of alumina is well known to glass- 

 makers. 



The formation of corundum in an igneous rock as the con- 

 sequence of the solution of argillaceous material is well illus- 

 trated by the ca.se described by Prof Busz. The mineral occurs 

 round inclusions of clay slate in a felsite from South Brent. 

 Many cases of the presence of corundum in igneous rocks under 

 conditions which prove that it must have crystallised out of the 

 magma, are now well known ; and amongst the most interesting 

 are those recently found in Hastings County, Can,ada, where the 

 mineral occurs in dykes of syenite. In these, however, there 

 appears to be no evidence that the excess of alumina is due to 

 the solution of argillaceous rocks. 



The remarkable synthetic experiments ot Dr. Morosewicr 

 give a complete and satisfactory account of the chemical and 

 physical conditions under which corundum, spinelle, sillimanite 

 and cordierite separate out of alumo. silicate m;igmas ; and, 

 therefore, of many of the natural occurrences above referred to. 

 .Vlumina is soluble in magmas agreeing in composition with 

 albite, nephelinc and anorthite, or with mixtures of these, and 

 crystallises out as corundum on prolonged cooling at high 

 temperatures. If both silica and alumina are present in excess 

 of that necessary to form felspar, sillimanite is formed until the 

 excess of silica is used up, and then the remaining excess of 

 alumina crystallises out as corundum. The presence of magnesia 

 determines the formation of spinelle, or of cordierite, or of 

 both, according to the excess of alumina and silica above that 

 necessary to form felspar with the soda, potash and lime 

 present. All these phenomena may be verified within the 

 range of temperature \n a Siemens' furnace, such as that iised in 

 glass-works. The minerals obtained are in every way similar, 

 except as regards size, to those which occur. in nature. 



It thus appears that the " faithful companions " may be 

 formed either by the metamorphism of sedimentary depo.sits, or 



NO. 1529, VOL. 59] 



