THE ARTIFICIAL PRODUCTION OF DIAMOND. 
91 
of temperature might still be found necessary to concentrate the reactions. It 
seems however probable that the rate of cooling might be so much prolonged as to 
obtain much larger crystals and a larger total yield. 
The presence of crystals of silica, alumina and magnesia and the spinels and 
pyrope associated with diamond in rapidly cooled iron alloys, and also when oxidized 
by steam and some other gases, appears to have a bearing upon the presence of 
similar crystals usually found in association with diamond, and to be compatible with 
the conclusions of Bonney that eclogite is the parent-rock of the diamond in South 
Africa. It seems probable that both the eclogite and the diamond may have been 
crystallized nearly simultaneously from an iron alloy. 
Moissan, after a recital of the geological conditions existing in the South African 
pipes (see ‘ Four Electrique,’ p. 115), came to the conclusion that diamond was not a 
vein mineral, but must have been evolved in the midst of a plastic mass ; and he 
concludes that iron at high pressure must have been the matrix. Our experiments, 
however, seem to show that bulk pressure on the metal does not play a part, but 
that the previous heat treatment, the impurities in the iron and the condition of 
the gases within the metal, are the important factors. 
It is interesting to note that in the best experiments the yield of diamond in 
rapidly-cooled iron has reached 1 -1-20,000 of the weight of iron, whereas the weight 
of diamond obtained from the blue ground of the Soutli Africaii mines is otdy 
1 -F 5,400,000. This comparison appears to be confirmed by the relative rarity ol 
microscopic diamonds we have found in the many analyses we have made of Ijlue 
ground and of the conglomerate from Brazil. 
Thus in cooled iron there may be more than 270 times as much diamond as exists 
in the bulk average of blue ground. 
In conclusion, I desire to express my obligations for kind assistance and advice to 
Sir Dugald Clerk, Prof Jeans, Mr. Stanley Cook, Mr. Campbell Swinton, and 
to many other friends, as also to Mr. H. M. Duncan. 
From 1906 to 1908 inclusive, the late Mr. Trevor Cart assisted me in the 
arrangement of the experiments and was responsible for most of the analyses until 
the time of his death. 
From January, 1911, to August, 1914, Mr. H. M. Duncan acted as my assistant 
and analyst and has given valuable help in the collection and tabulation of the whole 
of the work. During the preceding and intervening periods the analyses were made 
in the laboratory at my house. 
YOL. CCXX.—A. 
o 
