176 T. G. BONNE Y [.September 



rock was a true breccia. That opinion was not altered by the study of 

 Professor Lewis's manuscripts, but I thought it possible that his 

 Kimberlite might be represented in certain very compact fragments of 

 serpentinous aspect, the nature of which I had been unable to deter- 

 mine, owing to the want of definite characters and to my own ignorance 

 of what a serpentine formed from a glassy or very compact peridotite 

 would be like. Apart from this possibility, my views on the main 

 question differed from those put forward by my friend. It was, 

 however, my obvious duty to keep the difference of opinion as far as 

 possible in the background, and to endeavour to act as a simple channel 

 for the publication of the views of one who was no longer able to speak for 

 himself. Not long after the book had been published, Sir W. Crookes 

 allowed me to examine a piece of breccia which had been obtained at a 

 depth of 1320 feet, and was in even better preservation than any 

 which I had hitherto seen. About the same time Sir J. B. Stone 

 forwarded to me another set of specimens which he had received 

 from Kimberley. Among these were two or three blocks, in almost as 

 good a condition as that just named, and from an even greater depth, 

 viz. 1400 feet. After study of these * I was more than ever convinced 

 that the Kimberlite was a true breccia, formed by the explosive de- 

 struction of some coarsely crystalline rocks, such as eclogites and peri- 

 dotites (including representatives of the sedimentary rocks of the region). 

 I was also able to ascertain the true nature of those fragments which 

 hitherto I had thought might possibly be serpentine of an exceptional 

 character ; they proved to be in reality nearer to argillites, but to have 

 undergone certain alterations, in all probability partly from contact 

 action, and partly from water, perhaps at a rather high temperature, 

 and no doubt at a later time. Thus I arrived at the conclusion, that 

 the so-called Kimberlite was not an altered peridotite, but a breccia, in 

 which the diamond, like the olivine, pyroxenes, garnet, etc., was not 

 authigenous, but a derivative from some older rock. This I thought 

 very probably was a peridotite, for an a priori argument, as we may 

 call it, which Professor Lewis had used seemed valid, even though 

 he mi"'ht have misunderstood the nature of the Kimberlite, and 

 his idea that a very basic rock would be the birthplace of 

 diamonds was confirmed by their occurrence in meteoric iron (Canon 

 Diablo 2 ) and their manufacture by Moissan through the intervention 

 of that metal. 



Two suggestive discoveries must next be mentioned, of which, 

 however, I was ignorant till within the last few months. A diamond 

 had been obtained in 1892 embedded in a garnet (pyrope); and in 

 another specimen no less than six diamonds occurred closely associated 



1 See Gcol. Mag. 1897, p. 448. 



2 Another occurrence of diamond (not very pure) in a meteorite which fell at Novo 

 Urei, Russia, Sept. 22, 1886, is mentioned hy Professor Kuntz, Eighteenth Ann. Report 

 of the U.S. Gcol. Survey, Part V. p. 1195. 



