36 Some Anomalies in the Winds of Northern India. [Jan. 20 



anomalous hot winds is to be found, not in the distribution of pressure 

 at the level of the plains, but in an interchange between the lower 

 atmospheric strata and those at high levels, effected through the 

 medium of convection currents set up by the diurnal heating of the 

 earth's surface when the sun shines. This hypothesis was first put 

 forward by Koppen to explain the diurnal inequality of wind velocity, 

 and the author shows that in the dry season the vertical distribution 

 of temperature in India is such that convective action capable of 

 producing such interchange must occur. 



In the second part the diurnal variation of the wind velocity is 

 attributed to convective interchange, and it is shown that probably the 

 annual inequality may be explained in the same way, the barometric 

 gradients prevailing at high levels between the plains and the moun- 

 tains to the north of India being subject to an annual variation 

 dependent on the temperature. To verify the conclusions deduced 

 from the convection hypothesis, the midday pressures at 10,000 feet 

 above sea-level have been computed for the months of January, May, 

 July, and October from the observations of many years at forty 

 stations, and the resulting values have been laid down on a set of 

 charts, another set giving the distribution of pressure at sea-level 

 and the prevailing wind directions over India. The high level distri- 

 bution of pressure, as shown on the charts, is found to be exactly such 

 as would produce the observed anomalies in the wind direction and 

 velocity. The charts also furnish reasons for the particular paths 

 taken by the disturbances which bring the winter rainfall of Northern 

 India and by the cyclonic storms originating at different seasons in 

 the Bay of Bengal, the rule being that the storm centre follows the 

 line of ]owest pressure in a stratum of the atmosphere lying above all 

 local obstructions such as the mountain ranges in the interior of 

 India. 



The third part of the paper is devoted to proving that in years 

 when the summer rains fail the gradients for westerly winds at 

 10,000 feet over Northern India are intensified, in the first place by 

 the unusual cold over the North-west Himalaya, due to the previous 

 snowfall, and afterwards by the great heat of the plains, which have 

 not been cooled by the usual precipitations in June and July. The 

 evidence for this conclusion is not so clear as it might be ; but it is 

 shown that when the most trustworthy observations are compared 

 the gradients for westerly winds at 10,000 feet over the Grangetic 

 plains were very high in the remarkably dry years 1877 and 1880, 

 whilst they were very low, that is to say, there were gradients for 

 easterly winds over a great extent of the plain, in 1879 and 1884, 

 which were years with excessive rain. In the moderately dry year 

 1883 there was a considerable gradient for westerly winds, but not 

 nearly so great as in 1877 or 1880. 



