390 



NATURE 



[February 27, 1896 



elusion from the figures there given, except that this wire, either 

 from its quality or its situation, behaved in a different manner 

 from any one of the many specimens I have examined during 

 the past five years. The value of 8R/5/ decreases with increase 

 of temperature in a most phenomenal manner.^ In cases in 

 which I have observed this phenomenon in a lesser degree it has 

 indicated a breaking down in the insulation, consequent on rise 

 in temperature. I am unable to find any evidence that the 

 insulation was tested at high temperature during these experi- 

 ments. 



(4\ Any expression of opinion by Messrs. Holborn and 

 Wien necessarily carries weight ; nevertheless, I would venture 

 to suggest that (considering the amount of experimental work 

 previously performed by those who advocate the methods of 

 platinum thermometry) the examination of only two wires, one 

 of which was admittedly exposed to the action of furnace gases, 

 affords insufficient grounds for the adverse conclusions arrived 

 at by the authors. 



(5). Ueycock and Neville's determinations of the freezing 

 points of copper, gold, and silver are admitted by Holborn and 

 Wien to be in "good agreement" with their own. The nature 

 of this agreement is shown in the following table. 



Heycock and Neville. 



Copper ... io8o'5 

 Gold ... 10617 

 Silver ... 9607 



Holborn and Wien. 

 1082 

 1072 

 971 



It is worthy of notice that copper is the only metal of which 

 Holborn and Wien used large quantities, comparable with the 

 ma.sses experimented on by Heycock and Neville. Also in this 

 case Holborn and Wein determined both melting and freezing 

 points. Their results (using practically the same thermometer 

 throughout) range from 1076° to 1093°, whereas Heycock and 

 Neville's values, when using six distinct platinum coils possessing 

 very dissimilar constants, range from io79°'o to io8i''7, a very 

 different order of agreement. 



In the case of gold and .silver, Holborn and Wien used small 

 quantities, determining their results by observations of the 

 melting points, and thus the method of experiment adopted 

 renders it probable that the temperatures observed would err 

 on the side of excess. The close agreement in the case of 

 copper, and the higher values found by Holborn and Wien in 

 the experiments on gold and silver, are therefore significant. 



Thus, when the conditions are similar (as in the case of 

 copper), we may regard the results obtained by the different 

 observers as practically identical. Such agreement would be 

 impossible if platinum resistance thermometers ordinarily 

 underwent, at high temperatures, changes of the nature 

 of those observed in the wires studied by Holborn and Wien. 

 The differences would then be measurable not by units, but by 

 tens and hundreds ! and these discrepancies would be found not 

 only when different methods were used, but also when the 

 same observations were repeated with different platinum 

 thermometers. 



If Tables XL, XII., and XIII. of Heycock and Neville's 

 paper ^ are examined, it will be found that although in each case 

 from 6 to 8 different platinum thermometers were used in which, 

 for example, the value of 5 varied from i -495 to 2*04, the extreme 

 resulting temperatures differ by a smaller quantity than the 

 differences obtained by Holborn and Wien when repeating an 

 observation without change in the conditions, by means of the 

 same thermo-couple. 



Finally, I assert that Holborn and Wien have produced 

 no evidence sufficient to support the somewhat sweeping 

 conclusions given by them on p. 394 of their paper. I have 

 .shown that in their experiments on only two samples of wire, 

 they have neglected the precautions insisted upon by those 

 who have devoted years of study and experiment to the investi- 

 gation of the platinum thermometer, and this portion of their 

 work is only useful in so far as it emphasises the validity of the 

 conclusions arrived at by those who preceded them. 



I fully appreciate the great value of Holborn and Wien's 

 direct determinations of high temperatures by means of the 

 thermo-couple and the air thermometer, and I admit it is 

 probable that for temperatures exceeding 1400° C. or so, the 

 thermo-couple is the more convenient, and possibly the more 



1 This extraordinary behaviour of Holborn and Wien's Sample II. (for 

 reasons previously given the behaviour of Sample I. is of no significance) 

 is noticeable in the numbers attained by them at loiu, as well as at high, 

 emperatures. 



2 Chetn. Soc. Trans., 1895, pp. 188-190. 



NO. 1374, VOL. 53] 



accurate instrument. Below such temperatures, however, I 

 consider that the weight of evidence is in favour of the accuracy 

 of the platinum thermometer. In any case, such evidence is 

 in no way weakened by the experiments of Holborn and Wien. 



E. H. Griffiths, 



Earth Tremors, 

 In Prof. Milne's article in Nature of December 26, he states 

 that earth tremors are more frequent during the winter than 

 during the summer, that they are frequent with a low barometer, 

 and still more frequent when the locality of observation is 

 crossed by steep barometrical gradients. In the North-West 

 Himalayas, throughout the winter months, slight earth tremors 

 are exceedingly frequent, and occur, so far as can be judged 

 without instrumental records, more frequently by night than 

 by day. This may be in part due to the fact that during the 

 day most people would be moving about in downstairs 

 rooms, while at night the same people would be in 

 upstairs rooms, and both they and their surroundings 

 perfectly quiet ; but, whatever may be the day and 

 night relation, there can be no doubt that during the winter 

 months in Simla peculiar little earth tremors are remarkably 

 frequent. My experience has been that these tremors are not 

 so much connected with areas of low barometer as with the 

 co?Jimencement of a sudden and large change in atmospheric 

 pressure from a high to a low, a reduction of pressure which 

 need not necessarily be accompanied with steep barometric 

 gradients or high winds at or near the earth's surface. In the 

 case of earthquakes, also, I have noticed subsequent large 

 changes in atmospheric pressure. Thus at about midnight on 

 January 15-16, 1896, a (for these regions) rather severe earth- 

 quake occurred, which lasted from im. 20s. to 4m. in differeni 

 localities. On the plains the most severe shocks were felt at 

 midnight, 15th, at Simla, at oh. 30s. a.m. on the i6th, 

 and at Srinagar at i a.m. on the i6th. Times for other 

 places in the Punjab were published in the newspapers, 

 but I have omitted to keep them. The above, however, show 

 that the shock was felt at Lahore at midnight, at Simla half an 

 hour later, and at Srinagar an hour later. The barometric 

 records show that for the forty-eight hours from 8 a.m. on the 

 1 6th to 8 a.m. on the i8th, pressure changed as follows : — 



Srinagar (Kashmir) 



Astor ,, 



Murree 



Lahore 



Simla 



Quetta 



-0-187 

 -0062 

 -0165 

 -0-I44 

 - 0*140 



-|-0*022 



From the above figures, it appears that a considerable decrease 

 of barometric pressure occurred between the morning of the 

 1 6th and the morning of the i8th, and that this fall was central 

 over Srinagar ; while the times of occurrence of the earthquake 

 show that the movement of terrestrial disturbance was directed 

 towards this central area of diminishing pressure. It has always 

 appeared that the atmospheric changes which ordinarily occur in 

 tropical and subtropical countries would be a wholly inadequate 

 cause to account for the considerable earthquakes which at times 

 occur ; but I have undoubtedly noticed that very slight earth 

 tremors constantly take place when a sudden and large decrease 

 in atmospheric pressure coin/nenccs after a considerable period of 

 high pressure. W. L. Dallas. 



Simla, January 29. 



"Roches moutonnees." 



Some ten years ago, I came across in an old memoir a rational 

 explanation of the term roches moutonnees ; but I made no note 

 at the time, and have been unable to trace the reference. How- 

 ever, my scepticism was fortified, and I proceeded to search 

 French dictionaries, which made it clear that mo utomie meant 

 " frizzled like sheep's wool," and not " sheep-like." Yet M de 

 Lapparent tells us (" Traite de Geologic," 3me ed., p. 281) that 

 these glaciated rocks ' ' produisent une impression analogue a celle 

 d'un troupeau de moutons endormis, d'ou le nom de roc/ies 

 moutonnees " ; and who shall question this precise .statement of 

 a French author interpreting his own language ? It is the ex- 

 planation that has been taught to all of us, though I know of 

 only one field-geologist who seriously maintains that roches 

 moutonnees might be taken for a flock of sheep. Agassiz states 



