oct. — dec. 1857.] Cyclone at Nellore. 67 



peared, it became evident that the rivetment was fast falling, and 

 the earth work in rear washed away. This state of affairs continu- 

 ed for about one hour, during the whole of which time no object 20 

 yards distant had been visible, and the trees immediately in front 

 of the temple, which were too deeply embedded in the earth to be 

 uprooted, had been reduced to mere shapeless stumps. About this 

 time (J to 9) the wind began to shift a little to the eastward, and 

 we then knew that, if the bund could hold out a little longer, no 

 further serious damage would occur to the part on which we were 

 standing. The change in the direction of the wind rendered how- 

 ever our position, on the outer pial of the doorway, no longer ten- 

 able, and we sought refuge in the middle recess, with the Tahsildar 

 and the other natives. The wind continued to shift still more to 

 the eastward, and the houses abovementioned in a few minutes 

 again became visible. The moment the spray and wind had suffi- 

 ciently subsided, we quitted the Pagoda which by this time was by 

 no means a safe place, though probably deemed by us more inse- 

 cure than it really was, and made the best of our way against the 

 gale, until the shelter of some walls enabled us to get on a little 

 faster, and thankful were we when we again stood on the main land, 

 after crossing, with considerable difficulty, the piece of ground 

 and bridge over the surplus channel, which before had intervened. 



When I reached home, at J past 9 o'clock, I found the ba- 



4- 3 



rometer had risen ~ ths, and by 10 o'clock^ths more towards fair 

 weather had been attained. The hurricane rapidly subsided into a 

 strong gale, and from that to a moderate one, and by \ past 11 or 

 12 a fresh breeze only remained, blowing from the east by south 

 and south east, and from that point it has continued to blow ever 

 since. 



I cannot say exactly what the barometer stood at for the 48 

 hours preceding the setting in of this terrible hurricane, but I sub- 

 join the readings taken by me during the night of the first, and 

 morning of the second November ; they will give a fair idea of 

 the extent and character of the barometric fluctuations, and I sub- 

 join the reading at 12 noon this day, which will show the usual 

 state of this particular barometer in fine weather. 



