July 1, 1891.] 



K N O ^A/^ L E D G E . 



125 



upon us that the great-star-belt is a genuine girdle of stars 

 in space ; in which, also, the foundations of the sidereal 

 system are laid, the MilUy Way being an appendant to it, 

 of lesser rank. In short, the most noteworthy arrange- 

 ment in the architecture of our universe seems to consist 

 of a great ring of large stars intersecting an equal ring of 

 small ones at the extremities of a common diameter. Let 

 us recapitulate the evidence : To start with, we have the 

 probability — shall we say certainty '? — that the Galaxy is a 

 ring-shaped structure, having, as we sought to show, no 

 great branches in its own plane. Next, we have the 

 significance of an almost entirely isolated, symmetrical 

 belt of bright stars (stars singularly uniform in magnitude 

 and distriliution) encircling the whole heavens, and cutting 

 the Milky \\'ay in two exactly opposite parts. Then we 

 have the striking suggestiveness of the disturbed state of 

 the Milky Way in these parts, coupled with its evenness 

 both in outline and aspect elsewhere. Lastly, we have the 

 evidence derived from the aflfiuity between the Milky Way 

 and the stars in the belt ; the galactic off-set in Perseus 

 lying along the direction of the belt, the stellar off-sets 

 from Cygnus to Aquila, and from Crux to Sagittarius, 

 lying along the Milky Way ; the galactic streamers, more- 

 over, in Ophiuchus and Scorpio, nay, even the main 

 stream of the Milky Way, turning aside to the star-belt 

 just where the greatest angular distance separates them. 



The double-ring structure enunciated above dovetails in 

 with all these points ; indeed, it seems the only logical 

 deduction from them. Both Sir William Herschel's 

 hypothesis of an even distribution of stars throughout our 

 stellar system, and Proctor's spiral theory, fail utterly to 

 account for the fact of a zone of bright stars associated 

 with, but differentiated from, a zone of small ones in the 

 manner observed. Further than this, the theory of a 

 double-ring furnishes us with a rational explanation of the 

 conspicuous absence of streamers round the interior face 

 of the Galaxy, for it tells us where a powerful extraneous 

 force is to be found counteracting altogether the action of 

 the Milky Way upon itself. 



THE EXPERIMENTAL METHOD IN GEOLOGY 



By Vaughan Coknish, B.Sc, F.C.S. 



THE record of the investigation of a geological 

 problem may generally be divided" into two parts, 

 the descriptive and the explanatory. A rock, for 

 instance, is described according to its mode of 

 occurrence, structure and mineralogical composi- 

 tion ; then follow deductions as to the epoch at which 

 it was formed, and the mechanism of the actions by which 

 its particular characters have been produced. This, as a 

 rule, marks the limit of the geologist's investigation of 

 such a problem ; seldom, far too seldom, are the conclu- 

 sions submitted to the decisive test of experimental 

 methods. This lack of the confirmatory evidence of 

 experiment makes a large part of the literature of Geology 

 very unsatisfactory reading, the deductions being too often 

 either indefinite or inconclusive. In the present article 

 \vc give a sketch of some of the efforts which have been 

 made to raise Geology to the rank of an experimental 

 science. 



In the last years of the eighteenth century a controversy 

 raged between the schools of Huttoii and of Werner as 

 to whether heat or the action of water had been the 

 dominating influence at work in the formation of the 

 rocks of the earth s crust. By wliat agency, for example, 

 had chalk been converted into limestone or marble ".' How 



can this have been effected by heat, said the school 

 of Werner, since heat decomposes carbonate of lime, 

 expelling the carbonic acid ? The answer to this question 

 was furnished by the experiments of Sir James Hall, 

 " ()n till' iictii/n of livdt IIS iiiiiilijicil III/ pressure." Chalk 

 was heated in a gim-barrel, the end of which was 

 firmly closed. Under these conditions, the pressure in- 

 creasing as the temperature is raised, the carbonic acid is 

 not driven off from the carbonate of lime, the change 

 induced being not chemical but physical, the powdery 

 non-coherent chalk being converted into a compact crystal- 

 line mass, having all the characters of limestone, or of 

 marble. Hall also investigated another problem connected 

 with the same controversy. Hutton maintained the purely 

 igneous origin of those rocks which have characters 

 similar to the modern lavas. It had, however, been 

 noticed that if a piece of a crystalline rock were melted in 

 a crucible it was not reproduced on cooling, but that an 

 uniform glassy mass was formed. By a judicious com- 

 bination of the methods of observation and of experiment, 

 Hall obtained important evidence as to the conditions of 

 crystallisation of rocks. He observed during eruptions of 

 lava that a great part of the crystallisation of the con- 

 stituent minerals took place slowly, and by degrees, during 

 the gradual cooling of the mass of molten rock. Basing 

 a method on this observation, he melted various rocks in 

 graphite crucibles, and maintained the materials in a state 

 of fusion for a long time, taking care that the temperature 

 should be somewhat above that necessary to melt the 

 glassy mass. Crystals gradually formed, and a crystalline 

 rock was reproduced, of which the melting point was 

 higher than that of the glass formed in previous experiments, 

 where the cooling had been rapid. Similar experiments 

 were conducted about the same time (1804) by Gregory 

 Watt. They were on a larger scale, a reverberatory furnace 

 being employed in place of a crucible. The molten material 

 was only allowed to cool with extreme slowness. From 

 time to time samples were withdrawn and examined after 

 solidification. Those in which the anueaUng process had 

 1 continued longest were the most perfectly crystalline, and 

 i possessed the highest specific gra%'ity, just as a natural 

 crystalline rock, such as granite, is denser than a glassy 

 rock (('.;/. obsidian) of the same chemical composition. 

 These early experiments elucidated several important 

 points with regard to the processes which have taken place 

 in the formation of the eruptive rocks. The products 

 obtained were, however, at most very imperfect reproductions 

 of the natural rocks, and the methods for the determination 

 of mineral species were at that time too rough to allow of 

 the identification of the small and imperfect crystals 

 obtained. Before the date (18(56) of the next important 

 experimental research on the formation of rocks by igneous 

 fusion, the application of the microscope in petrological 

 work had effected a revolution in this respect. A slice of 

 rock, so thin as to be transparent, reveals to the microscope 

 the outline, and even the internal structure, of the minute 

 crystals which form its groundworli or /'"«•. The crystal 

 of each mineral species shows its characteristics of form, 

 the particular angles at which its faces are inclined to one 

 another, and the lines developed in the process of grinding 

 the thin section which indicate the directions of cleavage. 

 Not less important in identification are the optical 

 characters which determine the tints which different parts 

 of the field of view assume according as the polarised light 

 passes through the plate of one or other of the minerals of 

 which the rock is composed. The refinements of optical 

 analysis enable the identification of the species to be made 

 with certainty even in crystals of microscopic size. .\t the date 

 to which we have referred, M. Daubree published his experi- 



