160 
Philippine Journal of Science 
1920 
dark — is to postpone. The year’s experience shows that the 
more favorable the conditions, the greater the computed depres- 
sion of the sun. Fig. 1 attempts to show some of the relations 
graphically. The abscissas are the sidereal times corresponding 
to the standard (east 120°) times of observation; the ordinates 
are, for curve 1, minimum distances of the ecliptic from the 
zenith of Manila, expressed crudely by measuring in millimeters 
on the star map accompanying the American Nautical Almanac ; 
for curve 2, distances of the Galaxy east of the zenith of Manila, 
taken from Winslow Upton’s Star Atlas, and expressed roughly 
in hours of right ascension. The plotted points show the depres- 
sion of the sun in degrees. Since the axis of the Zodiacal Light 
lies near the ecliptic, the ordinates of curve 1 express roughly 
the distance of this from the zenith. It was found that greater 
distances of the Zodiacal Light, and clear weather free of cirrus 
clouds, gave large values of depressions; the nearness of the 
Galaxy had not so much importance. 
These large depressions, over 14°, were quite unexpected; in 
fact, before computing in two or three cases I had supposed my 
recognition of the blue color to be tardy. For all these reasons 
I consider them more probable than any others, and the largest 
of them, 14° 46.7', as the most probable. 
In applying this result to the question of atmospheric height, 
a value of horizontal refraction at sea level must be used, as 
