NATURE 



[September i, 192 i 



modern measurements, that Cavendish and his 

 colleague froze out and separated the hydrates 

 HN03,3H20, HN03,2H20, and HoSO^.HgO; 

 determined accurately their compositions and 

 melting points, as well as those of the eutectic 

 mixtures in which these hydrates are concerned ; 

 and secured data which can be plotted with re- 

 markable accuracy on a modern freezing-point 

 diagram. 



In addition to the chemical papers and manu- 

 scripts, the second volume of the " Scientific 

 Papers " includes reprints of the remaining papers, 

 of which the most important describes the well- 

 known "Experiments to Determine the Density 

 of the Earth." In dealing with this section of 

 Cavendish's work the editor has obtained con- 

 tributions from Dr. Chree, who writes a note on 

 the determination of the height of the aurora, and 

 gives an account of Cavendish's magnetic work; 

 from the Astronomer Royal, who writes on 

 Cvivendish's astronomical manuscripts; from Sir 

 Archibald Geikie, who writes on Cavendish as a 

 geologist ; and from Sir Joseph Larmor, who 

 adds a note to a manuscript on "The Refraction 

 on a Mountain Slope," and gives an account of 

 Cavendish's mathematical and dynamical manu- 

 scripts. 



It is a tribute to the work which has been ex- 

 pended on these two volumes that only sixty-six 

 out of 452 pages of the first volume, and 220 

 out of 496 pages of the second, are occupied by 

 reprints of the papers from the Philosophical 

 Transactions. The Cambridge University Press 

 has produced a worthy memorial of the work 

 of one of the most distinguished of Cambridge 

 men, and no student of the history of science in 

 England can afford to ignore or to neglect these 

 volumes. T. M. L. 



Paris Weather Statistics. 



Atlas MeteoroJogique de Paris. By Joseph Levine. 

 Pp. vi + 83-i-g plates. (Paris: Gauthier-Villars 

 et Cie, 192 1.) 20 francs. 



MUCH more will be found in this atlas than 

 is to be inferred from the title. The 

 author promises to set out graphically the annual 

 values of meteorological elements for Paris from 

 1700 to 1920, with monthly values from 1761. 

 This is shown in a series of plates. He also gives 

 complete monthly and annual tables for several 

 elements from 1874 to 1920, with a column of 

 annual departures from average, and of varia- 

 tions from year to year. The wind tables are not 

 NO. 2705, VOL. 108] 



so full, as they date back only to 1890, and some 

 of the other tables do not begin until 1876 or 

 1878. In addition, there is a table of extreme i 

 barometer readings from 1809 to 1919 for each | 

 month and for the year, and of highest and lowest ' 

 mean monthly and annual readings from 1757 to 

 1919. The highest recorded barometer reading 

 at an altitude of 67 m. was 781-2 mm. in February, ; 

 182 1, and the lowest 713-5 mm. in December of ' 

 the same year. During the period from 1878, of 

 which fuller details are given, the highest read- 

 ings were 782-4 mm. on January 16, 1905, and 

 782-3 mm. on January 17, 1882, at an altitude of, 

 50-3 m. (corresponding to 780-7 mm. at an alti- > 

 tude of 67 m.), and the lowest 718-1 mm. on. 

 January 10, 1916. It is to be remarked that at j 

 Greenwich, in the same period, the highest read-' 

 ings — 782 mm. — were recorded on January 17,] 

 1882, and January 29, 1905. The latter was J 

 nearly a fortnight later than the Paris maximum,! 

 though the former was on the same day, indi-| 

 eating a very extensive anticyclone, with possibly | 

 an even higher reading at some intermediate point. 

 Naturally, no such accordance can be expected in 

 the minimum readings. 



The highest shade temperature at Paris was 

 38-4° C. (101-1° F.) on July 20, 1881, five days 

 after the Greenwich reading of 97-1° F., which 

 has been exceeded only by that of 100-0° F. on 

 August 9, 191 1, on which day the Paris reading 

 was 97-7°. The lowest shade minimum in the 

 same period at Paris was —25-6° C. (—14-1° F.) 

 on January 20, 1879, about 20° F. lower than 

 anything at Greenwich since 1841 ; but in spite 

 of the greater rigour of the Paris frosts, they 

 occur neither so early nor so late as at Green- 

 wich. The limiting dates at Paris are October 5 

 and May 13; at Greenwich, September 27 and 

 May 24. The corresponding limits for ground 

 frost at Paris are September 13 and June 9, but 

 the period covered by the table is only from 1902 

 to 1920. There are no real limits at Greenwich 

 for ground frost, for it has been recorded during 

 the same period in both July and August. 



The mean rainfall of Paris is about an inch 

 less than that of Greenwich. In the forty-six 

 years of the table 28 in. was exceeded at Paris 

 twice, and at Greenwich eight times. On the 

 other hand, in six years at Paris, and in only two 

 at Greenwich, did the annual total fall below 

 19 in. 



Unfortunately, there is scarcely any informa- 

 tion about duration of sunshine. The author re- 

 marks that the record is not homogeneous, and 

 gives only the figures for 1919. 



W. W. B. 



