August 25, 1904J 



NA TURE 



401 



northward advance, and fretjiiently, when it reiu'hes lat. 

 12° to 16° N., it concentrates into a cyclonic storm. Such 

 a storm almost invariably marks the commencement of the 

 monsoon in the Bay of Bengal, and in about two out of 

 five years in the Arabian Sea. The advancing humid 

 currents in the rear of these initial cyclonic storms or 

 waves of disturbance march over the sea areas in a few 

 days, and thence cross the coasts towards which they are 

 determined by the low-pressure regions in the land areas 

 of Southern Asia, where they produce an almost complete 

 reversal or transformation of the weather conditions, the 

 result of which is that moderately high temperature and 

 small diurnal range of temperature, great humidity fre- 

 quently approaching saturation, much cloud, and frequent 

 rain obtain for the ne.xt three months over the greater 

 part of India, until, in fact, the middle or end of September. 



The reverse change — viz., the withdrawal of the humid 

 south-west currents — then commences, and is a slow 

 process, requiring usually from two to three months for its 

 completion. 



This is due to a gradual decrease of strength, and hence 

 to a fairly continuous contraction of the field of the current, 

 and also of its elevation or thickness. The current first 

 withdraws from North-Western India, being replaced by 

 light, variable, or north-westerly land winds. These land 

 winds increase in extension and volume with the continued 

 contraction of the south-west monsoon current. The more 

 important phases of the contraction and withdrawal of that 

 circulation from India are of especial interest. The first 

 phase, the retreat of the current from North-Western India, 

 accompanies a rise of pressure over the Persian area and 

 Xorth-Western India, with a shift of the trough of low 

 pressure from W.N.W. to N. or N.E. and corresponding 

 change of direction of the average tracks of the storms of 

 the period. This is followed after a short period of rain 

 in North-Eastern India and Burma by a rise of pressure 

 in -^ssam, Upper Burma, and Bengal, and the withdrawal 

 of the monsoon current from those areas. The current then 

 recurves over the centre of the Bay, in the same manner 

 as during the monsoon proper over the north of the Bay 

 and Bengal, and is directed or determined to the west or 

 Madras coast of the Bay, which hence receives frequent 

 rain during a short period of about two months — the rainy 

 season of the eastern and southern parts of the peninsula 

 south of Orissa and Ganjam. 



These rains were formerly described as accompanying 

 the setting in of the north-east monsoon on the Madras 

 coast. That, however, is a misnomer, as the true north- 

 east monsoon winds are dry land winds, and the rain-giving 

 winds of this period in Madras are those of the south-west 

 monsoon in its retreat or contraction down the Bay. The 

 period during which this rainfall occurs is hence now usually 

 termed the retreating south-west monsoon. 



The year in India may hence be divided into two mon- 

 soons of nearly equal length, viz. : 



(n) The north-east or dry monsoon. 



(6) The south-west or wet monsoon. 



The first terms are based on the general direction of the 

 air movement in the Indian seas during the periods, and 

 the second on the most prominent feature of the weather 

 in India itself. Of an average annual total rainfall of 

 41 inches (according to the most trustworthy calculation), 

 at least 85 per cent, falls during the wet season, and only 

 15 per cent, during the dry season. 



The dry monsoon in India is subdivided into — 



(i) The cold-weather period. 



(2) The hot-weather period or transitional period of pre- 

 paration for the south-west monsoon. 



The wet monsoon is divided into — 



(i) The south-west monsoon proper, or the period of 

 general rains. 



(2) The period of the retreating south-west monsoon and 

 gradual slow establishment of the dry monsoon. 



Each of these periods practically covers three months. 



One of the most noteworthy features of the meteorology 

 of India not referred to in the previous statement is that 

 the storms of each period — viz., the cold-weather period, 

 the hot-weather period, and the wet monsoon — are 

 characteristic and special to the period. Thev are all in 



NO. 18 I 7, VOL. 70] 



the broadest sense of the word cyclonic in character ; but 

 they originate under different conditions and exhibit very 

 different features in each of those periods. 



The disturbances of the cold weather are large shallow 

 depressions which originate in the upper humid return 

 current of the north-east monsoon circulation, chiefly in the 

 Persian plateau region, and which drift eastward with a 

 slight southing across Extra-tropical India. Storms do not 

 occur south of the Deccan or peninsula-dividing ranges during 

 this period. These storms are cliiefiy remarkable for the 

 frequent development of stationary secondary depressions 

 in the Punjab, usually of much greater intensity than the 

 primaries ; a feature of which, I believe, there is no parallel 

 elsewhere. They are of great importance, as they give the 

 main snow supply to the Western Himalayas and the light 

 but general occasional rain required for the wheat and 

 other cold-weather crops of Northern India. 



The storms of the hot weather are local disturbances of 

 very limited extent, usually in large areas of slight de- 

 pression, and are occasionally of remarkable intensity and 

 great violence. In the areas to which the local sea winds 

 of the period e.xtend (more especially Bengal and Assam)- 

 they occur chiefly as local thunderstorms with violent winds 

 and brief heavy downpours of rain, but sometimes as 

 tornadoes rivalling those of certain districts of the United 

 States in intensity and destructiveness. In the dry interior 

 they occur as dust-storms, usually without rain, and are 

 most violent in the driest districts, including Sind, the 

 Punjab, and Rajputana. Occasionally, when the convective 

 movement is especially vigorous, they develop into hail- 

 storms of great intensity. The rainfall accompanying these 

 hot-weather storms is of little general agricultural value 

 except in the tea districts of Assam and Bengal. 



Finally, the wet monsoon is characterised by the frequent 

 occurrence of cyclonic storms of every degree of intensity 

 and of very varying extent. The great majority of them 

 originate in sea areas of nearly uniform temperature as 

 disturbances in a massive current highly charged with 

 aqueous vapour and subject to large variations of intensity 

 and extension. The more prominent features of these 

 storms, more especially of the most violent, including the 

 hurricane winds, excessive rainfall, and the phenomena of 

 the central calm and the accompanying storm wave, are 

 too well known to require description. The chief import- 

 ance of these storms, of which an average of about ten 

 (of different degrees of intensity) occur every year during 

 this period, arises from the manner in which they modify 

 the distribution of the rainfall, discharging it abundantly 

 over the districts traversed by the storms at the expense 

 of the districts outside of their field. 



The most important and variable feature of the weather 

 in India from the practical standpoint is rainfall. Its value 

 depends upon its amount and occurrence in relation to the 

 needs of the staple crops. The measurement of rainfall is 

 carried out, on a uniform system, at upwards of 2^00 rain- 

 gauge stations. The average distribution of rainfall, month 

 by month and for each season, has been determined from 

 the data of about 2000 stations. It should, however, be 

 recognised that the probability that the rainfall will con- 

 form exactly to this distribution in any year is nil. Average 

 rainfall charts represent a distribution about which the 

 actual varies from district to district more or less con- 

 siderably, the local variation for prolonged periods being 

 practically compensatory. Such mean or normal data and 

 charts are undoubtedly of value, more especially for the 

 determination of rainfall anomalies and their relations to 

 pressure, temperature, and other anomalies. There is 

 apparently a tendency to assign a greater value to these 

 charts of mean rainfall distribution than they deserve. 

 Charts showing the amount and time distribution of the 

 rainfall best suited for the requirements of the staple crops 

 would — for India at least — be more interesting and valu- 

 able. This is a work that I regret has, for various reasons, 

 not yet been carried out by the Indian Meteorological 

 Department. 



In most regions in India a moderate variation (positive 

 or negative) in the amount of the rainfall is of comparatively 

 small importance, more especially if the precipitation occurs 

 in amount and at intervals suited to the requirements of 

 the crops. During the thirty-year period i874-iqo3 there 

 were six vears in which the distribution of rainfall affected 



