162 



S A. mn— Tornadoes and UaiUlnrms of April [No. 2, 



The figures in brackets are comijuted by the above logarithmic for- 

 mula) for the actual distribution of vapour up to the highest observing sta- 

 tions, but as they exceed the pressures of saturabioa at the probable tem- 

 peratures of the air they do not rojjresent actual facts. The differeuoea 

 between them and coi-responding pressures of saturated vapour re- 

 present in fact the vapour which must have been condensed. It thus 

 appears that on the afternoon of the 30th April, and probably throughout 

 the greater part of the following day, the distribution of temiserature 

 was such as to cause vertical movements of the atmosphere above the 

 level of 5,000 or 6,000 feet, and that the cloud level, even supposing no 

 such movements to be in progress,* was probablj^ at the comparatively 

 low elevation of 8,000 or 10,000 feet. The conditions were therefore 

 favourable for the development and intensification of any disturbances 

 casually occurring in the currents set up. It now remains to inquire 

 whether any relation can be traced between the places of occurrence 

 and tracks followed by the tornadoes and the distributions of pressure 

 at or near the cloud level where they in all probability commenced. 



For this purpose, the 4 p. m. observations of pressure on each of 

 the two days of the storms have been reduced to a level of 9,000 feet 

 above the sea, and the resulting values are laid down on the last two 

 charts. These high-level pressures have been computed on the as- 

 sumption that the temperature decreased each day on ascending accord- 

 ing to the scale above shown — an assumption which is probably not far 

 wrong as regards the Himalaya stations, and the line of nearest stations 

 on the plains, vig., Ludhiana, Roorkee, Barielly, and Grorakhpur, but 

 perhaps rather wide of the truth as regards more distant places. If the 

 decrease of temperature above the more distant stations was loss rapid, 

 as is probable, the effect of the assumption made must be to diminish 

 the pressures towards the south of the area covered by charts, making 

 the gradients for westerly winds less steep, though in other respects 

 producing little change. Thus the pressures as computed for Bareilly 

 and Agra are almost identical on the 30th April, and differ by only 

 0-217" on the Jst May ; whilst, if we assume for Agra the rate of decrease 

 of temperature found by Glaisher in his balloon ascents in clear 

 weathei-, the excess of Agra over Bareilly at the 9,000 feet level is 

 0-24:7" on the 30th and 0-333" on the 1st. The computed pressures at 

 9,000 feet are given in the following table and have been laid down on 

 charts V and VI : — 



* In an ascending current, since there is no diminution of the quantity of vapour 

 until condensation commences, cloud will be found wherever tiio expansion due to 

 upward motiou has reduced the temperature to the dew point. 



