Page:Proceedings of the Royal Society of London Vol 69.djvu/75

Rh {{c|The Effect of the Types of Weatlier (Cyclonic or Anticydonic} on Temperature during fJte Five Years 1876-1880. }}

The period from 1876 to 1880 was first examined. With the object of classifying the days of this period, the atmospheric conditions pre- vailing on each day were recorded on a separate card, to be subse- quently grouped in any manner that might be thought desirable. The conditions recorded were

1. The mean temperature.

2. The direction of the wind.

3. The weather.

4. The direction of the gradient (high to low), and

5. The type of the weather prevailing (cyclonic or anticyclonic). For this purpose weather was set down as cyclonic if the isobar on

the 8 A.M. chart was concave to a region of low pressure, and as anti- cyclonic if the isobar was concave to a region of high pressure. The first classification was made according to the type of weather. It was thought possible that this second-order temperature effect might be a result of periodic variations in the relative frequency and effect of cyclonic and anticyclonic weather at different periods of the year. With the object of ascertaining whether this was the case,* the mean temperatures and frequencies of occurrence of cyclonic and anticyclonic weather, during each month of the year, were tabulated for the five years under observation. For the purpose of examining the results, curves of temperature were drawn whose ordinates are proportional, not to the absolute mean temperature for the month, but to the difference between the mean monthly temperature and the tempera- ture which would be represented by the mean of the ordinates of the first-order curve during the month ; so that the resulting curves show that part of the temperature variations which is represented in the harmonic analysis by a summation of all the curves of higher order than the first. Diagram 2, fig. 1 (p. 68), shows these curves of tem- perature for the cyclonic and anticyclonic days respectively. It will be seen that only during the three summer months, and to a slight extent during March, is the cyclonic weather cooler than the anticyclonic. But both the curves show the main characteristics of the second-order curve, viz., maxima in February and August, and minima in May and November. Moreover, the curve of difference of temperature between cyclonic and anticyclonic weather (fig. 2) shows three maxima in the year. Fig. 3 shows the differences in frequency of the two types of weather during each month, expressed as percentages of the whole number of days in the month ; here again the curve is irregular, having

given some reasons against regarding cold as definitely associated with anticyclonic weather in winter.
 * Mr. W. H. Dines, ' Quart Journ. Eoy. Met. Soc.,' vol. 23, p. 237, has already