November i8, 1909J 



NA TURE 



85 



sideration. It is therefore necessary to make such experi- 

 ments as those of Tammann at vastly higher temperatures 

 and pressures than those we have been considering, up to 

 probably more than 10,000 kilograms per sq. cm. (or 

 635 tons per sq. inch). 



In 1S93 some experiments were described by Parsons ' 

 in which carbon rods were heated by electricity under a 

 pressure usually of 15 tons per sq. inch, but rising in one 

 case to 30 tons per sq. inch. The pressure was obtained 

 by means of a hydraulic press, but no detail is given. 



I have been desirous for many years of making some 

 experiments at high temperatures and pressures, but for 

 a long time could think of no W'ay of ascertaining the 

 pressure at temperatures over a red heat except by the 

 use of compressed gases. In 1902 Sir Andrew Noble was 

 kind enough to have some drawings prepared for a wire- 

 wound steel pressure vessel to carry a pressure of 50 tons 

 per sq. inch. The pressure was to be supplied by a com- 

 pressed gas, and some details of the heating arrangements 

 ■were designed, when a calculation of the cost of the gas 

 ■rompressors, vessel and appurtenances, made it clear that 



190 



180° 

 170° 

 160° 

 150° 

 140 

 130° 

 120 

 110° 

 100 

 50^ 



ashamed of bringing them to your notice — I can only 

 say, in excuse, that everything must have a beginning. 



I believe, however, that the apparatus is sufficiently 

 simple, cheap, and effective to enable others with more 

 leisure at their disposal to make a beginning of an in- 

 vestigation of the properties of matter up to 100 tons per 

 sq. inch, and at temperatures up to about 2000° C. .At 

 present, however, it is not possible to infer with accuracy 

 the volume of the substance under these extreme conditions, 

 nor can its physical condition be more than approximately 

 and indirectly inferred — we must content ourselves with 

 the production of transformations which we can make 

 persist down to ordinary temperatures and pressures. 



If we refer again to the sulphur diagram, we shall see 

 how this possibility may arise. If sulphur is melted and 

 cooled slowly, monoclinic crystals are found — when the 

 temperature sinks below 98° C. these crystals undergo 

 spontaneous transformation to the rhombic form — but all 

 that we see is that the monoclinic crystals become opaque ; 

 the external form of the crystals is still monoclinic, but 

 they are merely pseudomorphs of the original crystals. 



1000 Kg/cm^ 



2000 



30OQ, 



the undertaking would be beyond my means. I then 

 endeavoured to find a simpler form of apparatus, and 

 finally was led to contemplate the substitution of graphite 

 for compressed gas. Spring having pointed out that crystal- 

 line graphite flows very easily at high pressures. A 

 simple trial made it clear that the graphite of Ceylon 

 does, in fact, possess the property of flowing like a liquid 

 under high pressure to a sufficient degree to allow of 

 pressure being transmitted by it. Graphite can be used, 

 with some reservations, to transmit a pressure just like 

 water or oil, though it is, of course, inferior in fluidity, 

 and, as I have now discovered, occasions a loss of " head " 

 which is not independent of the pressure itself. My former 

 statement in the Chemical Society's Journal, 1908, is 

 erroneous, though the results of the experiments are, I 

 Relieve, hardly, or not at all, affected by the mistake, for 

 a reason which will be clear later on. Aher several trials, 

 the apparatus which I have here to-night was evolved, and 

 ■some experiments were made with it. These experiments 

 are not of any great importance, and, indeed, I feel almost 

 1 Phil. Mag., xxxvi., ;o4. 



NO. 2090, VOL. 82] 



To obtain large octahedral crystals we may suppose that 

 we begin by melting sulphur and raising the temperature 

 and pressure until the former stands at 160° C. or over, 

 and the latter at not less than 1600 kg. /cm." (10-16 tons/sq. 

 inch). 



If we then slightly reduce the temperature or raise 

 the pressure, we shall have the crystallisation of the 

 sulphur in the rhombic form. By maintaining the pressure 

 as the mass cools, and when it is cold releasing the 

 pressure, we should finally extract rhombic crystals. To 

 this we may, of course, add that we need not expect 

 crystals of any size unless we cool at the proper rate. 

 It appears that there are at least two phenomena requir- 

 ing attention in relation to the production of crystals — 

 one is the relation between the amount of undercooling 

 necessary to induce spontaneous crystallisation, and the 

 other is the rate at which the crystals will grow when 

 they have once started. If we want large crystals we 

 must not have an e.\cessive number of points of spon- 

 taneous crystallisation, nor must we have too high a rate 

 of crystal growth; or the crystals will by all experience 



