November 5, 19 14] 



NATURE 



26' 



being determined from a photograph which showed 

 the comet with a tail. Its direction of movement is 

 given as north-north-west. 



A Little-known Pioneer of Spectrum Analysis. — 

 \\'hile we owe the birth of spectrum analysis to Sir 

 Isaac New'ton (1642-1727), the first use of the slit to 

 WoUaston (1802), and the first accurate mapping of 

 the solar spectrum to Fraunhofer (1814), the name 

 of the young pioneer, Thomas Melville, who followed 

 closely after 5s^ewton is nearly forgotten. Although 

 Melville died in 1753, at the age of twenty-seven, he 

 was the first to employ the prism in laboraton,- re- 

 search, for he undertook the examination of various 

 flames, introducing sal ammoniac, potash, alum, etc., 

 into burning spirits. He was undoubtedly the first to 

 observe the yellov.- soda flame and notice its definite 

 frequence in flames and its position as regards other 

 colours. His name is briefly mentioned in Sir Henr}' 

 Roscoe's " Spectrum Analysis," and Miss Gierke's 

 "History of Astronomy," gives him full credit for his 

 work. It was in reading the latter book that the 

 attention of Mr. John A. Brashear was directed to 

 the reference to Melville. Through Mr. Carnegie and 

 Dr. Hew Morrison, the latter being principal librarian 

 of Edinburgh, Mr. Brashear was enabled to obtain 

 copies of Melville's essays read before the Medical 

 Society of Edinburgh on Januan,- 3 and February 7, 

 1752, on observations on light and colours, in which 

 appears the remarkable observation of the peculiar 

 sodium light. These papers are now reprinted in the 

 current number of the Journal of the Royal Astro- 

 nomical Society of Canada (vol. viii., No. 4, July- 

 August), and a ver}- good service has been done by 

 making them available to everybody. 



Displacement of Spectrum Lines at the Sun's 

 Limb. — Messrs. J. Evershed and T. Royds describe in 

 Kodaikanal Observatory Bulletin, No. 39, some further 

 researches they have been carrjing on with regard to 

 displacements of the spectrum lines at the sun's limb. 

 Two explanations have been suggested to account for 

 these shifts, one view being that they are due to 

 pressure and the other due to motion in the line of 

 sight. In this paper the authors state that the 

 pressure theon,- presents a much more rational ex- 

 planation of the phenomenon than the motion theory, 

 but hold that there are difficulties in accepting the 

 former, which have not been lessened, but largely 

 increased by further research. The authors consider 

 that the relative shifts of different lines at the limb 

 have no particular meaning when determined by refer- 

 ence to lines at the centre of the disc, for these latter 

 have shifts peculiar to themselves; the measures they 

 have made show that the absolute shifts of the lines 

 at the limb referred to a terrestrial standard show 

 no relation to pressure shifts, and, further, the abso- 

 lute shifts do not increase with the wave-length. In 

 the investigation now published they determine the 

 limb shifts by combining measures of limb minus 

 centre shifts, with the centre minus arc shifts, the 

 algebraical sum representing the absolute or lirnb 

 minus arc shifts. They deal with the limb shifts in 

 relation to the intensity of the lines, the relation be- 

 tween limb shifts and pressure shifts, limb shifts in 

 relation to wave-length, and the shifts of the cyanogen 

 bands. The conclusions they arrive at are that both 

 the so-called cyanogen bands and the iron lines show 

 shifts which they attribute to a movement of recession 

 from the earth.' While the view that the solar gases 

 are actually repelled by the earth receives some support 

 from other lines of evidence, they consider an alter- 

 native hypothesis, namely, that the sun's gravitational 

 field affects the wave-length of the light emitted in 

 accordance with Einstein's theory- of relativity. 



NO. 2349, VOL.. 94] 



CLIMATOLOGY. 



A PAPER by Dr. J. H. Garrett (.Medical Officer of 

 Health) on "The climate and topography of 

 Cheltenham and its near neighbourhood " is published 

 in the Proceedings of the Cotteswold Naturalists' 

 Field Club for 1913 (vol. xviii., part ii.). The tables 

 for thirty-five years (1878-1912) are derived from the 

 records contained in annual health reports, or from 

 the summarised values issued by the Meteorological 

 Office. The figures show, as the author points out, 

 that as regards the mea. monthly maximum and 

 minimum temperatures, the alternations from one 

 season to another are very gradual. For the years 

 1892-1912 these were, respectively : — January, 43-6°, 

 33'6°; July, 71-6°, 534°; absolute maximum, 93° in 

 September, 191 1, minimum 6-5° in February, 1895. 

 The annual rainfall shows considerable variation : 

 19-5 in. in 1892, 379 in. in 1882, normal for thirty-five 

 years, 268 in. The prevalent wind directions are 

 S.-VV., those from N.-E. are comparatively few, while 

 calms are very frequent. The average yearly sunshine 

 (1903-12) was 1553 hours. Dr. Garrett remarks that 

 quite a different climate is experienced on the Cottes- 

 wold Hills, within four miles of the centre of the town. 

 An elaborate discussion by Dr. V. Conrad of the 

 climate of Carinthia, being part vi. of the valuable 

 climatology of Austria, has been published by the 

 Central Meteorological Institute of Vienna. This 

 mountainous province covers an area of 10,327 km-, 

 of w'hich 537 per cent, is above the level of 1000 

 metres. Like other Alpine countries it partakes of 

 the sea climate of western Europe and of the land 

 climate of eastern Europe, and is subject to abrupt 

 contrasts within short distances. Tropolach, for in- 

 stance, in the Gail \'alley, with an eastern aspect, 

 has a mean January te'- iperature of 18-7° F. and a 

 winter temperature of 223°, while Pontafel, 75 miles 

 S.S.W., with a south exposure, has a January tem- 

 perature of 286° and a winter temperature of 30-6°. 

 The mean absolute jearly range of temperature in 

 some of the valleys amounts to above 50° C. (90° F.); 

 the absolute extremes recorded were below —22° F. 

 and above 95° F. The yearly distribution of rainfall 

 is well shown by a tinted map, referring to the period 

 1876-1900. With few exceptions there is a tendency 

 towards increase of rainfall from north to south. The 

 driest parts are the Noric Alps and neighbouring dis- 

 tricts,, with 31-35 inches of rain; the wettest parts 

 are in the south-west, with 58-^8 inches. The province 

 possesses several long series of observations, e.g., 

 for Klagenfurt since 1813, but there is only one sun- 

 shine recorder, which is at the latter station. 



Dr.* Filippo Eredia. has recently contributed to the 

 Rendiconti of the Accademia dei Lincei (vol. xxiii.) 

 two interesting memoirs on the distribution of cloud 

 in Italy. The papers show the seasonal and annual 

 distribution ^or 132 cities, and the orographic influence 

 on the monthly mean values at a number of stations 

 in the valley of the Po. In the latter region, on the 

 Adriatic, central and lower Tyrrhenian, and Ionian 

 slopes, winter is the season of most cloud. On the 

 upper Tyrrhenian slope and on those near the Alps, 

 spring is th cloudiest season. At Alpine regions and 

 those under the direct influence of mountain systems 

 there is ge^i 'rally least cloud in winter and in summer 

 at other stations. The mean annual distribution and 

 amplitude are shown by diagrams. 



Another paper by Dr. Eredia, on the climate of 

 Southern Italian Somaliland, is published by the 

 Colonial Ministry of Rome as No. 14 of Rapporti e 

 Mcnografie coloniali. The obser\'ations include inter 

 alia those made by the Italian expeditions and at 

 stations afterwards established, and are carefully 



