798 REPORT — 1898. 



of Europe, and how the anomalies which these curves show may contribute in a 

 large degree to the discovery of their origin. 



In the diagram which was exhibited to the Section are plotted down six annual 

 curves for temperature, air-pressure, rainfall, magnetic declination, and for the 

 vertical and horizontal components of the earth's magnetic force. A portion of these 

 curves are for Greenwich, others for two different stations in the Netherlands. 

 It is now shown that all these curves, however dissimilar in their general direc- 

 tions, are exactly alike in their anomalies. Except in a very small number of 

 instances, where the data at hand were not yet sufficient to make the phenomenon 

 appear, every maximum or minimum in one curve seems to have its exact counter- 

 part in every other curve, be it meteorological or magnetic. This is even the case 

 for very small accidents of a curve, so that it is probable that it will ultimately 

 prove to be true for every single feature of these curves, however insignificant. 



These facts, in the first place, show once more how all the phenomena on the 

 earth must be, to a large extent, governed by one and the same potent cause. But 

 they also seem to give a valuable method for discovering, if not always the cause 

 of meteorological phenomena, yet of the centre on the globe from which such a 

 cause emanates. 



It is also shown how the cause of a certain minimum in the temperature 

 curves, indicating a sudden cooling in the last days of June all over the British 

 Isles and part of Western Europe, what ever it may be, must have its seat to the 

 west or north-west of the coast of Scotland, and at no great distance. 



9. The Classification of Poly diurnal Weather Types in relation to the Prolon- 

 gation of the Daily Forecast in Western Europe. By Douglas Archi- 

 bald, M.A., F.R.Met.Soc. 



The results of modern meteorological investigation, whether official or amateur, 

 statistical or synoptic, have shown the existence of specific types of weather em- 

 hriicino- wide areas and intervals of time varying from several days to months, and 

 seasons which tend to produce persistence or recurrence ofthemore ephemeral changes 

 connected with the passage of the smaller, temporary, and movable cyclonic and 

 anticvclonic systems. 



These large weather types may be roughly classified as : — (1) Seasonal or x 

 monthly, and (21 y daily. 



The former appear most clearly in tropical countries, such as India, where the 

 y daily changes are small and where the summer season is always characterised by 

 the formation of a permanent cyclonic area over Persia and North-Western India, 

 with subsidiary low pressure troughs over the Ganges basin, round which the so- 

 called S.W. monsoon circulates with a complete reversal of pressure conditions at 

 the opposite season. 



Tliese normal conditions, expressed thus in general terms, are evidently the 

 result of seasonal actions due to the direct influence of the sun on the Asiatic con- 

 tinent and Indian Ocean, and, stated thus, may be predicted to recur and form a 

 laro-e proportion of the seasonal forecast every year. 



When, however, we descend from the general to the particular we find each 

 vear's S.W. and N.E. monsoon differ from that of every other year. 



Superposed on the regular normal type is what may be termed the yearly 

 seasonal abnormal type, corresponding to which both the intensity and shape 

 of the seasonal cyclone and anticyclone and the accompanying weather vary. Such 

 type, however, once initiated at the critical comnjencing month of the season, is 

 found to persist more or less all through. This is the practical basis of the seasonal 

 or .r monthly weather forecast of the Indian Service, so ably worked by Mr. Eliot, 

 r.R.S., at Simla. 



In Europe the seasonal type, though still manifest, is not large enough in com- 

 parison with the y daily changes to enter in as a specific factor in the forecast. 



On the other hand, the restriction of the forecast to 24 or 36 hours is unneces- 

 sarily arbitrary in view of the study of y daily types foreshadowed by the late 



