VOL. LXXXV.] PHILOSOPHICAL TRANSACTIONS. SSQ 



the intermediate ground near the Duke of St. Alban's Park. The following is 

 still more singular. " I observed," says Mr. Dalby, " what seemed to me a very 

 uncommon effect of terrestrial refraction, in April 1793, as I went from Fresh- 

 water Gate, in the Isle of Wight, towards the Needles. Soon after you leave 

 Freshwater Gate, you get on a straight and easy ascent, which extends 2 or 3 

 tniles; a mile, or perhaps a mile and an half beyond this to the westward, is a 

 rising ground, or hill; and it is to be remarked, that its top and the aforesaid 

 straight ascent, are nearly in the same plane: now in walking towards this hill, I 

 observed that its top, the only part visible, seemed to dance up and down in a 

 very extraordinary manner; which unusual appearance however evidently arose 

 from unequal refraction, and the up-and down motion in walking; but when the 

 eye was brought to about 3 feet from the ground, the top of the hill appeared 

 totally detached, or lifted up from the lower part, for the sky was seen under it. 

 This phenomenon I repeatedly observed. There was much dew, and the sun 

 rather warm for the season, consequently a great evaporation took place at that 

 time." Here, and also on Hounslow Heath, the rays of light passed near the 

 earth's surface a great way before they arrived at the eye; and it is more than 

 probable, that moist vapours were the principal cause of the very unusual re- 

 fractions: the truth of which conjecture seems to be verified by the following 

 circumstance. In measuring the base on Hounslow Heath, we had driven into 

 the ground, at the distance of 100 feet from each other, about 30 pickets, so 

 that their heads appeared through the boning telescope to be in a right line; 

 this was tlone in the afternoon. The following morning proved uncommonly 

 dewy, and the sun shone bright ; when having occasion to replace the telescope, 

 we remarked that the heads of the pickets exhibited a curve, concave upwards, 

 the farthermost pickets rising the highest; and we concluded that they were not 

 properly driven, till in the afternoon, when we found that the curve appearance 

 was lost, and the ebullition in the air had subsided. 



The new raised earth about the gun at King's Arbour, prevented a very ac- 

 curate measurement of the height of the instrument above the point of com- 

 mencement of the base; and therefore two opportunities only presented them- 

 selves for determining the actual terrestrial refraction; namely, at the ends of 

 the base of verification. From the depression taken at Beacon Hill, the re- 

 fraction was 38"; but the elevation of Beacon Hill, observed at the lower end 

 near Old Sarum, gives 50". These deductions perhaps cannot be deemed very 

 conclusive; because, as they depend on the difference in the vertical heights of 

 the ends of the base, every 2 inches of error in that difference will produce an 

 error of about I" in the computed refraction. We shall close this section with 

 the data whence those refractions were obtained. At Beacon Hill, the top of 

 the flag-staff near Old Sarum was depressed 42' 6'. At the other end of the 



4 p 2 



