22 THE VOYAGE OF H.M.S. CHALLENGER. 



sufficient to enable us to arrive at a positive conclusion with regard to the mode of 

 formation of the rock. 



We shall first examine the question in its lithological bearing, which, indeed, in the 

 present instance is the most certain basis to work upon, for in this case we have no other 

 positive data than the lithological. Frankly admitting once more that lithological consti- 

 tution alone cannot always decide the question of origin, we find that this dubiety increases 

 in proportion to the number of new peridotic rocks discovered, for each fresh discovery 

 seems to upset views previously entertained. It may be of service here to cast a rapid 

 glance over the characters of the principal known peridotites, and on the received inter- 

 pretations of the mode of formation of olivine considered as a mineral. 1 It may be 

 admitted in a general manner that no objections can be raised a priori against the 

 igneous origin, pure and simple, of a peridotic rock. Olivine unquestionably belongs 

 to that class of magnesian silicates which can be crystallised artificially with the greatest 

 facility by dry fusion, as shown by the products of smelting furnaces, and by the experi- 

 ments of Daubrde. 2 The igneous origin of olivine is also proved by the presence of this 

 mineral in the lavas of active volcanoes and in older rocks universally admitted to be 

 pyrogenous. This fact is shown in a most conclusive manner by the basic volcanic 

 glasses dredged by the Challenger in the Pacific Ocean ; those glassy lapilli contain the 

 most beautiful olivine crystals, 3 and might in some cases be classed as olivine rocks, since 

 more than one-half of their whole mass is composed of crystals of olivine, which are so 

 compact and numerous as to seem to be present in greater quantity than the vitreous 

 ground mass. 



Not only does olivine, considered as a mineral, appear in conditions where its igneous 

 origin is unquestionable, but some peridotites if we are to judge from the investiga- 

 tions of competent observers, present positive characters of eruption. Hochstetter 

 admits that dunite 4 should be considered as an eruptive rock. According to Bonney, 



1 We do not pretend to say that olivine is an essentially pyrogenous mineral, but that name may be given it in 

 most cases, though it is occasionally found in crystalline limestone, in talc-schists, and in other rocks whose mode of 

 formation is still open for discussion, and whose origin cannot be considered as purely and simply pyrogenous. In this 

 respect, olivine is somewhat like amphibole and pyroxene. Though generally present in volcanic rocks, these minerals 

 are nevertheless sometimes the constituent elements of rocks which cannot be considered as having an igneous origin, 



3 Daubree, Comptes rendus, t. Ixii.; Bull. Soc. Geol. de Fr., 2^ me Serie, t. xxiii., 1866 ; Eapport sur les progres 

 de la Geol. Exper., 1867, p. 122. 



3 These crystals are perfectly identical with those described by Penck (Studien iiber lockere vulkanische Auswiirf- 

 linge, Zeitschrift d. d. geol. Gesell., 1878, p. 8) ; and by Van Werveke (Neues Jahrbuch fur Mineralogie, &c., 1879, 

 pp. 484, 485). 



4 The dunite here alluded to was discovered by F. Hochstetter in 1859 (Zeitschrift d.d. Geol., Ges., 1864, p. 342), 

 near Nelson, New Zealand, where it constitutes a large mass in Dun Mountain, with cliffs 4000 feet high. It is inter- 

 calated in an immense vein of serpentine from one to two miles long. The dunite is composed exclusively of granular 

 olivine, of a greenish-grey colour, possessing all the physical characters given in the foregoing descriptions of the speci- 

 mens examined by me. Chromic iron is also present in the dunite, but much more developed than in the specimens 

 from St. Paul. Prof. Hochstetter believes that the outflow of the dunite occurred during the mesozoic period. As no 

 microscopical description of this rock has yet been published, I may here give the principal characters presented under 

 the microscope. In the slide I have examined (which was kindly lent to me by Prof. Maskelyne) it appeared, in 



