with Remarks on terrestyial Rtfraclion. 213 



orists have consequently insisted upon the propriety of noting 

 the temperature at the two stations. The experiment has 

 been tried. At Pendle Hill, 507 feet higher than Rumbles 

 Moor, the temperature, June 19, 1821, was for several hours 

 4° or 5° the warmest. The air was nearly saturated with 

 moisture, and the refraction, contrary to theory, remarkably 

 small. In addition to this fact, the zenith distances of hills at 

 distances of 30 to 40 miles, although observed in different 

 years, rarely varied so much as 10". Hence it is probable 

 that the diminution of temperature, variable as it is without 

 doubt on the sides of mountains, may be constant in the greater 

 part of the path of the ray. If the temperature noted at the 

 base and sunmiit of the hill where the instrument is used, 

 should indicate a decrease of heat different to 1° for eveiy 

 200 or 400 feet, the observer must be prepared for an unusual 

 refraction, and will do well to repeat the observations the next 

 opportunity. This method is more convenient, and probably 

 of more real utility, than the preceding one. The degree of 

 refraction at the nearest extremity of the curve most sensibly 

 mfluences the estimate of the direction of the ray. 



We must next advert to the more considerable and dan- 

 gerous variations which, distorting the ray without the sliglrost 

 reference to its extent, have rendered the ratios of the refrac- 

 tion to the arc so inexplicably capi'icious. 



The diurnal variation of refraction, of which the effects have 

 been described in my last, is without doubt the principal yet 

 unforeseen cause of these apparent anomalies. Within the 

 range of my observations, I may describe it as being peculiar 

 to certain periods of the year, generally from the middle of 

 January to that of June. In the course of the three first 

 months of last year, strange to tell, it was onl}^ once percepti- 

 ble at the Observatory, the apparent elevation of Rumbles 

 Moor scarcely varying 8" during the whole period. The ex- 

 cepted observation was on a frosty day, nearly the only one 

 during the winter. At Rumbles Moor, June 12, 1821, tlie 

 variation was nearly at a maximum. May 28, 1822, the an- 

 gles were constantly the same from noon till within an hour 

 of sunset, when the refractions, as is usually the case, suddenly 

 increased 10". Unable as I am efficiently to account for the 

 cause of the diurnal variation, yet as its effects are greatest 

 upon angles nearest to the ground immediately around the 

 station, and as it appears to be more intense on a plateau 

 than on a peaked mountain, and in low situations ratlier than 

 ill more elevated ones, it may be supposed to proceed from the 

 radiation producing dew, known to be several degrees colder 

 llian the contiguous air. This explanation becomes more 



probable. 



