A CONTRIBUTION TO THE GEOLOGY OF SOUTHERN MAINE 529 



metamorphism was mainly of the nature of intense crushing, 

 without chemical addition or subtraction. In most cases there 

 were occasional less altered patches, where the original texture 

 and mineralogy could at least be inferred, if not actually 

 observed. It was found that the rocks were invariably originally 

 holocrystalline, usually of granitic texture. The authors of the 

 new system propose various prefixes for the description of 

 texture. These are used in the following pages when the original 

 texture could be observed. When a rock is thoroughly schistose 

 or gneissic the prefix "meta" is employed. Where it is meta- 

 morphic, but shows indications of what the texture was before 

 metamorphism, the prefix "meta" is used and also a textural 

 prefix. Thus, " meta-grano-sitkose " is a metamorphic rock 

 in which a former granitic texture is evident, while "meta- 

 auvergnose, " is a hornblende schist in which all trace of original 

 texture is lost. In those cases where both metamorphic and 

 non-metamorphic portions were found, the metamorphism is 

 ignored in the nomenclature. 



Only one rock in the region was seriously altered by weathering. 

 This was the most basic of the types, being an almost pure 

 olivine rock. It contains serpentine and other alteration 

 products, and in the analysis a notable amount of water and of 

 carbon dioxide was determined. The analysis as a whole could 

 not be recast in terms of the new system because of these ex- 

 traneous substances. Nevertheless it was perfectly possible to 

 classify fhe rock, since the alteration products were found on 

 microscopic examination to be all replacements of the olivine. 

 The composition of the olivine could be determined from the 

 analysis, and the alteration products could be ignored in classi- 

 fying. The only possible chance for error in such a case lies in 

 the doubt about the original proportions between olivine and 

 magnetite, but the mathematical grouping is sufficiently broad 

 for classification to be made in this respect with reasonable 

 certainty. 



The tedious labor involved in the production of "superior" 

 analyses necessarily limits the number of them. In this case all 

 of the principal types were analyzed, but many subordinate 

 variations were of necessity examined only microscopically. 



