Page:Proceedings of the Royal Society of London Vol 69.djvu/86

78 cases the higher orders are entirely negligible. At Valencia and Aberdeen, however, the amplitude of the third-order curve approaches one-third of that of the second-order curve, and therefore becomes more important ; but its amplitude amounts to only one-third of a degree, and the second-order effect tlms remains the only one of any practical importance.

Mr. A. A. Eambaut* has harmonically analysed the curve of the monthly mean temperatures for a single year (1899) at Oxford with a similar result, and in this case, as in the analysis of the 1884 Kew temperatures, the second-order curve is comparatively very large. It will be seen that all the stations hitherto mentioned have been on or near the sea-coast, and indeed it may be said that all British stations are comparatively near to a sea-coast. It was therefore thought ad- visable to make an investigation of the temperatures at a Continental station, where the effects due to the proximity of the ocean \vould be eliminated. For this purpose Vienna was chosen. Dr. Hann, in a paper on the " Temperature of Vienna, " based on observations for 100 years, gives the harmonic analysis of the curve of mean daily temperatures (Table I, Vienna). The amplitude of the second-order curve is only -^th of that of the first-order curve, and that of the third-order curve is approximately of the same magnitude. Moreover, the weather at Vienna has been investigated for the years 1876-1880 in the same way as the weather at Kew, and curves of temperature difference from the first order curve value have been drawn for each wind and each type of weather. No definite periodic curves were obtained similar to those obtained for Kew, all the curves being dif- ferent from one another, and in themselves irregular : the only universal characteristic was that the cyclonic and anticyclonic curves for the

DIVERGENCE FROM FIRST ORDER CURVE.

VIENNA 1876 TO 1880. DEC. JAN. FEB. MAR. APR. M AY J UN. JUL. AUQ. SER OCT. NOV. DEC.

Thick. During cycionic weather. Thin. asiticycionic.

whole year (Diagram 7), and moreover for each wind, showed a re- markable parallelism, the cyclonic being almost always a little warmer


 * 'Rov. Soc. Proc.,' vol. 67, p. 221.