January 15, 1920] 



NATURE 



505 



typical fossil of the intertrappean beds of the 

 Deccan, named in honour of the well-known 

 secretary of the Asiatic Society of Bengal, 

 James Prinsep. Numerous photographic views, 

 diagrams, and maps add to the interest of the 

 work, most of which are taken with due 

 acknowledgment from the publications of the 

 Geological Survey of India. Of the author's 

 own views, two have been selected for re- 

 production, one showing a typical landscape 

 in the crystalline area of the peninsula, and the 

 other that unique feature in one of the great 

 peninsular rivers — the falls on the Narbada, near 

 Jabalpur. T. H. D. L. 



METEOROLOGY IN THREE DIMENSIONS.^ 



TN 1916 Mr. W. H. Dines put together in a 

 •^ concise report the information then available 

 about the pressure, temperature, and density of 

 the atmosphere up to heights of 15-20 kilometres. 

 His report is now published, and should prove 

 extremely useful and informing both to the new 

 generation of meteorologists and to the wider 

 circle whose interest in the atmosphere is non- 

 professional. 



The first nine sections deal with the methods 

 and places of observation, the averages and 

 seasonal variations of pressure, temperature, 

 and density, and the stratosphere and tropo- 

 sphere ; short accounts of humidity and atmo- 

 spheric motion are also included. 



The tenth and eleventh sections are concerned 

 with the results of the statistical treatment of the 

 original data ; the interpretation of these results 

 will provoke much discussion. First, the cor- 

 relation coefficient between (i) the mean tempera- 

 ture of a vertical column extending from a 

 height of I km. to a height of 9 km. and 

 (2) the pressure at the top of the column is 095. 

 The hydrostatic equation connecting variations of 

 pressure at i km. and 9 km. with variations of 

 the mean temperature of the column of air is 



SA_¥,^5T 

 A A T2 • 



From this it follows 



Vo 



that if ^ 

 Pi 



or 



is (i) zero 

 correlation co 



(2) proportional to -^", then the 



efficient between pg and T is unity. 



The first condition is not fulfilled in temperate 

 latitudes ; the second condition would be satisfied 

 if the isobars at 9 km. were parallel to those at 

 I km. — i.e. if ftie wind-directions at these levels 

 were identical. But in the troposphere convec- 

 tion is always tending to make the direction of 

 the wind the same at all levels, so that the magni- 

 tude of the correlation coefficient found by Mr. 

 Dines may be due to the effectiveness of con- 

 vection in regulating the wind. It would be in- 

 teresting to know the differences from parallelism 



1 Meteorological Office. Geophys'cal VTemoirs, No. 13. " The Charac- 

 teristics of the Free Atmosphere." (London: Meteorological Office, 1919.) 

 Price 2J. net. 



NO. 2620, VOL. 104] 



permitted by the 005 by which the actual co- 

 efficient falls short of unity. 



Secondly, if Tq, Pq; Tj, Pj, etc., are the tem- 

 peratures and pressures at heights of o, i, 2, . . 

 13 kilometres, then the correlation coefficients 

 between corresponding T's and P's, beginning 

 with Tq, Pq, are as follows : -ii, -42, -66, •j'^, -84, 

 .85, .86, .86, .86, .71, .32, -.19, -.36, -.28. It 

 follows that pressure and temperature go up and 

 down together with great regularity at all heights 

 between 3 km. and 9 km. Presumably the same 

 would hold for the surface were it not for the 

 effects of radiation and of the surface water of the 

 ocean upon the surface temperature of the air. 



Two outstanding deficiencies in the information 

 available call for comment. There are no records 

 from the United States, India, Australia, South 

 Africa, South America, and Japan. This is no doubt 

 partly due to the difficulties of recovering records 

 in these countries if the ordinary European method 

 of investigation is used ; but it is also due to the 

 defects of pre-war international meteorological 

 organisation in which no place was found for an 

 active permanent bureau. Further, the informa- 

 tion about atmospheric motion is hopelessly in- 

 adequate. This arises less from leek of original 

 records than from the absence of any proper 

 arrangements for summarising the results of pilot- 

 balloon ascents. A young meteorologist seeking 

 a field of independent research might do worse 

 than turn to the statistical treatment of vectors. 



Before the war the investigation of the free 

 atmosphere was, broadly speaking, pure research; 

 the work had no direct application in forecasting 

 or climatology, and the means of investigation 

 were slight and relatively expensive. During the 

 war a knowledge of the actual conditions of the 

 atmosphere at least up to 20,000 ft. (6 km.) became 

 essential for heavy artillery and for aviation, and 

 their importance for actual daily forecasting began 

 to be dimly recognised. Now that artillery opera- 

 tions are over and aviation is practically restricted 

 to low levels, there is a great risk of the investi- 

 gation at higher levels by aeroplanes and kite- 

 balloons being neglected ; and instead of informa- 

 tion being available an hour or two after it was 

 obtained, records would again creep in months 

 or years out of date, with no possibility of imme- 

 diate practical usefulness. E. Gold. 



SIR THOMAS R. ERASER, F.R.S. 

 XA/HEN, in 1877, and then in his thirty-sixth 

 * » year, Thomas Richard Eraser was called to 

 succeed Sir Robert Christison as professor of 

 materia medica in Edinburgh University, it could 

 scarcely have been anticipated how closely he was 

 to rival his great master in his length of tenure 

 of the chair and in the distinction with which he 

 was to fill it. In his varied spheres of action 

 Fraser attained a commanding position as a 

 physician, as an investigator, and as a professor. 

 Gifted with acute senses and a fearlessly logical 

 mind, and trained in the habits of accurate ob- 

 servation and experiment in the laboratory, P'raser 



