82 



Museum, Cambridge (see Fig. V., Plate VII.) shows that more than 

 one-half of the rock is composed of olivine in the form of more or less rounded 

 grains. The interspaces between the olivine grains are occupied as a rule 

 either by a deep brown ("basaltic") hornblende or a nearly colourless 

 pyroxene. The bisilicates, hornblende and pyroxene, polarise uniformly over 

 large areas ; but they do not show definite crystalline outline. Biotite and 

 chloritic alteration products are sparingly present, but there is no definitely 

 recognisable felspar. Here then we have a rock which answers to TSCIIERMAK'S 

 definition of picrite, so far as the amount of olivine is concerned ; a fact of 

 some importance, as many of the rocks termed hornblende-picrite do not 

 answer to this definition. 



In subsequent papers Professor BONNP;Y has extended his term, 

 hornblende-picrite, to rocks which occur as boulders in Anglesea and South 

 Wales and also as dykes in Anglesea, the Lake District and the Lleyn peninsula. 

 All these bear a close macroscopic resemblance to each other in consequence 

 of the presence of numerous black cleavage surfaces which show lustre- 

 mottling ; but they vary considerably as to the relative proportions of the 

 different constituents. Felspar comes in and sometimes plays a not inconsider- 

 able part in their composition, either in the fresh or altered condition. 

 Olivine and its pseudomorphs decrease in amount and, indeed, are as a rule 

 less numerous than the definition of TSCHERMAK requires. 



It was stated that the term hornblende-picrite was introduced on the 

 assumption that the Schreisheim rock was a true picrite, i.e., an olivine- 

 diallage rock. Professor COHEN, who was responsible for this identification, 

 has since shown that he was mistaken and that the supposed diallage is in 

 reality hornblende. (1) He considers that this removes the rock in question 

 from the picrite group altogether, and proposes the term Hndsonite, on the 

 ground that a similar rock is extensively developed on the Hudson river. 

 Mr. G. H. WILLIAMS points out (2) that this term has already been 

 appropriated in America for a variety of pyroxene, and adds to the confusion 

 by proposing Cortlandtite. In a still later paper, however, he lays no stress 

 on the term Cortlandtite and simply calls the rocks hornblende-peridotites. (3) 



The conclusions which the present writer would draw from this discussion 

 on a question of nomenclature are that petrographers should definitely 

 recognise that rocks are simply mineral aggregates without any well 

 characterised individuality ; and that they should content themselves in most 

 cases with describing the varieties they meet with, and refrain from inventing 

 a new " ite " for every local variation in composition and texture. Having 

 regard to the nature of rocks it seems to him that less harm is often produced 

 by the extension of old terms than by the invention of new ones. 



If, then, we accept Professor BONNET'S term we may say that the picrites as a 

 group may be divided into two sub-groups the augite-picrites and the 

 hornblende-picrites. The Inchcolm rock may be taken as the type of the 



(1) N.J., 1885, Band I., p. 242. 



(2) N.J., 1885, Band II., p. 175. 



(I) Professor DANA had previously described allied rocks as chrysolitic hornblendite. G.M., 

 1881, p. 69. 



