418 NATURAL SCIENCE [December 



been to the petrologist. Each is regarded as an unmitigated nuisance, 

 interfering with the observation of facts for which the enquirer is- 

 searching ; and as the stratigrapher is expected to map solid rock 

 through its drift covering, the petrological specialist, when consulted 

 as to the character of a rock, is expected to say, not what it is, but 

 what it may once have been. 



Professor Merrill, however, has endeavoured, in the work under 

 review, to make the processes and results of weathering as interesting 

 and full of real importance as the American and English glacialists 

 are making the drift. For it is the weathering of rocks which forms 

 soil, the link between the dead earth-crust and the living plants and 

 animals upon it. Geologists have studied rocks in all their various 

 kinds, their origin, differentiation, and metamorphism, their birth, 

 growth, and life ; we have now to study their death, and the earliest 

 parts of the process which culminates in their resurrection. 



In order to make his book self-contained, Professor Merrill devotes 

 several chapters to considering the chemical and mineralogical com- 

 position structures, mode of occurrence, and various types of rocks ; 

 this is a fairly useful summary, but we are inclined to doubt whether 

 the geologist will need, or the lay-reader understand it. Many of the 

 illustrations in this part are admirable, and the abundant analyses are 

 particularly valuable for reference later on. To many of these no 

 references are appended, and we may conclude that they are due to 

 the author himself, who must have expended a great deal of time and 

 labour on the analyses throughout the work. 



Part III. is devoted to the weathering of rocks, each of the 

 chemical and mechanical agencies being taken in turn. Several 

 valuable instances of the effect of alternations of high and low 

 temperature are given, and the effect of cold rain on highly heated 

 surfaces is referred to. While stress is laid on the effect of hydra- 

 tion, the work of carbonated rain-water, and of the humic, ulmic, 

 and crenic acids naturally comes in for a lion's share of consideration. 

 The action of the first of these solvents upon many silicates can be de- 

 tected within ten minutes, while forty-eight hours' digestion will obtain 

 from some amphiboles, epidotes, felspars, etc., quantities of lime, mag- 

 nesia, iron, alumina and silica, amounting to from 0*4 to 1 per cent, 

 of the mass. Hornblende is more easily acted upon than felspar, and 

 even magnesian silicates are attacked, so that serpentine cannot be 

 considered a final product of decomposition. Increasing the pressure 

 on the solvent has much more effect than prolonging the time of its 

 action. Daubrce's experiments on attrition are referred to, and the 

 work of plants, bacteria, termites, and marine animals on the sea-bed, 

 is not neglected. 



Special cases of weathering are next treated in detail and illus- 

 trated by full analyses. Mere bulk analyses of the fresh and 

 weathered rock are misleading, as thev do not show all that has 

 actually occurred. It is necessary to ascertain which constituents 

 are least liable to be leached out, and to recalculate the analyses on 

 the assumption that they remain constant. Alumina and iron oxides 

 are least liable to this, and the analyses of acid rocks are worked out 

 on this assumption. An example will show the value of this method ; 

 the one chosen illustrates, in addition, that while only 30 per cent, of 



