OF ARTS AND SCIENCES. 291 



current, but enabled us in some measure to trace the deviations in 

 its course. At three o'clock, the ascent of a column of cold air was 

 still further proved, by holding a thermometer in the opening 'of the 

 well, the bulb being at the level of the ground, at a point where 

 surely mere radiation from the ice, some thirty feet below, should be 

 fully counterbalanced by the heat radiated to the mercury from the 

 curb and windlass that had been all day exposed to the hot sun. 

 The thermometer there indicated 43°.5 F., the temperature of the 

 external air being 78°. About five feet below the mouth, the ther- 

 mometer stood at 43°, and twelve feet down, at 40°. Water drawn 

 up from the bottom — without stopping to cool the bucket — was at 

 34°. Water drawn up at other times contained lumps of ice detached 

 from the coating of ice lining the well to the height of some five feet 

 above the surface of the water. When a candle was lowered down, 

 this ice appeared to be worn away on the northwest side, as though 

 small rills of warmer water were coming in from that direction. 



We visited the well at different times during the two following days, 

 and found the current as decided as when first perceived. In the 

 cool of the morning it expelled the dandelion pappus as readily as at 

 any other time, though then, from the want of contrast with exter- 

 nal heat, it was less felt by the hand or face. Calling the atten- 

 tion of the owner and one of his neighbors to the sensible upward 

 flow of air, we were told by both, that " cold air blows out " of the well 

 constantly, and has always done so. A remark expressed in such 

 terms was somewhat surprising, since the proprietor had assured us 

 on our first visit that no curi'ent existed in the well, as had been 

 repeatedly proved by lowering down a candle, which burned steadily in 

 any part. We finally repeated the candle experiment, and found, as 

 might have been expected from the gentle upward current filling the 

 whole circumference of the well, that the flame was not positively 

 deflected, except just below the edge of the opening in the marble 

 slab capping the wall. This hole, being of less diameter than the 

 well itself, compels the air, rising at the sides, to sweep towards the 

 centre, and accordingly, at all points within reach of the influence 

 of this centripetal contraction in the column, the flame of the candle 

 was turned inwards. This blowing of the candle towards the middle 

 from all parts of the circumference of the orifice, is worthy of remark, 

 as showing — what, indeed, the appearance of the ground outside 

 would indicate — that the dirt is so compacted around the marble 



