An Experimental Study of Field Methods. 547 
of the ascending currents of air, caused by the radiation of heat 
from the various materials of the ground along the line of sight, 
a typical case of which is shown in Fig. 15. 
Each belt of sand, clay, loam, gravel or covering of each will 
absorb and give off heat to the superimposed air according to 
its own absorbing and radiating power, giving rise to ascend¬ 
ing air currents of different densities and indices of refraction. 
Rays of light, which traverse the lower part of such a lens, will 
be refracted the most, and though slight in any single lens, by 
successive refraction on long sights it amounts to considerable. 
On sights of 300 m. the writer has observed the lower line of 
sight to be refracted 5 cm. more than the upper. 
Again, the almost infinite variety of ways in which rays of 
light traverse a large number of such currents, when influenced 
by every gust of wind, is sufficient to explain the peculiar fluc¬ 
tuations of the line of sight, called primary and secondary vi¬ 
brations. 
THE EFFECT OF LENGTH OF SIGHT ON ACCURACY. 
The question of what length of sight will give maximum 
accuracy of measurement, was tested by 422 test sights, aggre¬ 
gating eighty-five miles, made on an accurately measured base 
line, during the months of July and August. Here also are the 
effects of “ differential refraction ” plainly seen. This work 
consisted of from twenty to thirty stadia readings on nineteen 
stations, 100 feet apart, distant from 200 to 2,000 feet from 
