92 



NATURE 



[October 4, 1917 



caespitosa, some of the finer grasses, and other plants. 

 Messrs. Cole and Imms offer suggestions as to possible 

 causes of the outbreak, and both reports agree in 

 recommending the cutting of trenches across the grass- 

 land as a preventive measure. The effectiveness of 

 spraying measures is also discussed. 



The cutting off of supplies of potassium salts from 

 the German deposits has forcibly directed attention to 

 other sources hitherto neglected. Of the many waste 

 products investigated fevk' appear to offer better pros- 

 pects of economic utilisation than the flue-dust of blast- 

 furnaces. That soluble potassium salts are present in 

 these flue-dusts is no new discovery, but only of late 

 have they received serious consideration. According to 

 tests by Mr. H. T. Cranfield, published in the August 

 issue of the Journal of the Board of Agriculture, the 

 potash-content of these flue-dusts is extremely variable, 

 the total (acid-soluble), potash ranging in the twelve 

 samples quoted from 297 to 1589 per cent. K2O, whilst 

 the water-soluble potash ranged from 123 to 925 per 

 cent. The flue-dusts vary greatly in colour, and, gener- 

 ally speaking, the lighter-coloured materials are richest 

 in jiotassium salts. Potassium sulphate is the principal 

 of these salts, the chloride being also present in smaller 

 proportion. It is suggested that the total annual out- 

 put of potash in these flue-dusts is probably not fewer 

 than 15,000 tons, of which quite one-half is soluble in 

 water. These data furnish adequate justification for 

 the Order recently issued by the Ministry of Munitions 

 whereby the sale and treatment of blast-furnace dust 

 are brought under control. 



In continuation of the experiments on the tempera- 

 ture-gradient in the lavas of Kilauea, referred to in 

 Nature of June 28 (vol. xcix., p. 352), Mr. T. A. 

 Jaggar, jun., records that bright lines in the lava-lake 

 give temperatures of about 1020° — that is, about 250" 

 above those of the lake magma 3 ft. below 

 the surface (Bull. Hawaiian Volcano .Observa- 

 tory, March, 1917, p. 34). The same author contri- 

 butes an article on "The Thermal Gradient at Kilauea" 

 to the Journal of the Washington Academy of Sciences, 

 vol. vii., p. 397, in which he further emphasises the 

 generation of heat at the surface " through completion 

 of the reaction between rising unstable gas mixtures 

 and through union with atmospheric oxygen." The 

 liquid lava in the lake is 14 metres deep, and rests on 

 a seemingly pasty bottom. The lower 5 metres of 

 the lake have a temperature of 1120° to 1170°; this is 

 attributed to the release of air from foundered blocks, 

 which reacts with the volcanic gases and produces 

 reheating. 



The mean monthly temperatures of the surface 

 waters of the Atlantic Ocean north of lat. 50° N. are 

 the subject of a paper by Dr. C. Ryder which appears 

 as one of the publications of the Danish Meteorological 

 Institute. In 1892 the institute published the isotherms 

 for six months of the year calculated from fourteen 

 years' observations. The present paper marks a great 

 advance, for it comprises all months of the year, is 

 based on forty years' observations, and extends to 

 lat. 50° S. Most of the observations are from Danish 

 vessels, and unfortunately data are lacking for the 

 sea a few degrees east of Iceland in most months. A 

 chart is given for each month of the year, based on the 

 mean temperatures calculated for stations of 1° 

 squares. The information is also tabulated in mean 

 values for the four decades of the period covered. 

 This arrangement was desirable for many reasons, 

 not least because the transition from steam to sail 

 resulted in certain areas being more frequented in some 

 decades than in others. Perhaps the most instructive 

 chart is that on which the isotherm of 9° C. has been 

 NO. 2501, VOL. 100] 



drawn for all months of the year. The January, 

 February, and March isotherms almost coincide. In 

 April the northern trend is marked in the east, and 

 in succeeding months the isotherm swings north 

 until it touches the north-west and south-east coasts 

 of Iceland in August. Then again it withdraws south- 

 ward. In the west there is far less divergence be- 

 tween the relative monthly positions of the isotherm, 

 for the cold southward current is maintained through- 

 out the year 



The problem of temperature measurement and the 

 pyrometric control of furnace-casting and ingot-teeming 

 temperatures in steel manufacture is one the import- 

 ance of which it would be difficult to overrate. Hitherto 

 on account of its supposed difficulty its solution has 

 not been attempted. Publication No. 91 of the Tech- 

 nologic Papers of the Bureau of Standards is therefore 

 to be welcomed in that it takes up this problem, and 

 the conclusion reached by Dr. Burgess, the author, is 

 that it does not reilly present serious difficulties or 

 uncertainties. Observations have been taken in 

 several steel plants. The most satisfactory instrument 

 to use is an optical pyrometer using monochromatic 

 light, and permitting observation from a distance of 

 streams of metal. It is shown that the necessary cor- 

 rections to the observed readings for emissivity of 

 metals and oxides to give true temperatures are suffi- 

 ciently well known, but there may be uncertainty in 

 the case of liquid slags. For streams of liquid iron 

 or steel the most probable value of emissivity to take, 

 with a pyrometer using red light of wave-length 

 A — 0-65 fx, is e = o-4o, corresponding to a correction of 

 139° for an observed temperature of 1500" C. The 

 value of e for liquid slags is usually about 0-65, but 

 varies with the composition of the slag. It appears 

 from the author's results that the temperatures of the 

 roof of an open-hearth furnace bear no necessary rela- 

 tion to that of the metal bath, which again it is shown 

 may have zones of considerable difference.? in tempera- 

 ture, depending upon the operation of the furnace. The 

 temperature of the roof of an open-hearth furnace, 

 depending upon the firing practice, may vary verv 

 rapidly, and within wide limits, from i55o°-i75o° C. 

 That^ of the bath is usually kept between 1600° and 

 1670° C. There appears to be a remarkable degree of 

 uniformity in casting temperatures actually acquired 

 by the melters in practice. Thus for nineteen con- 

 secutive Bessemer heats the teeming temperatures of 

 the_ ingots were all between 1500° and 1555" C, and 

 a similar degree of concordance was found^ m the open- 

 hearth practice of several mills. 



In^ view of the importance of Fourier's series in 

 physical applications, much interest attaches to a paper 

 by Prof. H. S. Carslaw on "A Trigonometrical Sum 

 andtheGibbs Phenomenon in Fourier's Series" {Amer- 

 ican Journal of Mathematics, vol. xxxix.. No. 2, 1917). 

 In this paper Prof. Carslaw gives a proof of the 

 property first noticed by Gibbs, namelv, that when a 

 function becomes discontinuous the sum to infinity 

 of its Fourier expansion does not always merelv change 

 at an infinitely steep gradient from the initial to the 

 final value of the function, but that in certain cases it 

 may, in the neighbourhood of the discontinuity, fluc- 

 tuate between a maximum and a minimum value out- 

 side the limits of value of the function itself. In other 

 words, the maximum and minimum values of the sum 

 of a finite number of terms of the expansion, just 

 before and after the discontinuity, mav be outside the 

 limits of value of the function itself^ the maximum 

 exceeding the larger value of the function, and the 

 minimum being less than the smaller value bv. amounts 

 which remain finite, even if the number of terms be 

 increased indefinitely. The proof is well illustrated bv 

 the diagrams at the end of Prof. Carslaw 's paper, not- 



