2^6 Gen. Roy’s Account of 
matter as may poffiblv hereafter lead to further inveftigatlons 
of this curious, but at prefent vague and indetermined fubjeft : 
for from thefe obfervations it will appear, that terrefirial re- 
fraction, inftead of being -ith of the comprehended arc, accord- 
ing to M. Bouguer, _^th according to Dr. Maskelyne, 
r * th according to M. Lambert, varies from \ to -J-th part 
of that arc; and perhaps, if it had been portable for us to have 
tried it on heights- confiderably more elevated, we fhould have 
found it almoft wholly to vanifh. 
Art. II. Relative heights . 
Before we proceed to give any account of the obferved angles 
* of elevation or depreffioti, at the Rations reciprocally, for try- 
ing the quantity of terreftrial refraction, it will be proper to 
call to remembrance, that, in the meafurement of the bafe on 
Hounflow Heath, the mouth of the pipe at Hampton Poor- 
houfe was (hewn to be elevated about 60 feet above low-water 
fpring tides at the fea, as far as could then be determined by 
referring it to the furface of high water at Ifleworth ; and that 
the extremity of the bafe near King’s Arbour was found by 
levelling to be higher than the former end by 31 feet 3 inches. 
The mouth of the pipe at the fouth-eaft end of the bafe of 
verification at High Nook near Dymchurch, in Romney Marflh, 
Lieut. Fiddes found by levelling to be above low-water mark 
at fpring tides 22.1 feet. 
The top of the parapet of the north turret of the Keep of 
Dover Caftle was found by Lieut. Hay, of the Royal Engi- 
neers (by levelling from the top of the cliff at Queen Eliza- 
beth’s gun downwards, and adding to that the height of the 
7 
