16,2 Fisher: Low-Sun Phenomena in Luzon 159 
in the isothermal region and higher, and they catch light sooner 
than anything at 2 kilometers. 
The importance of this sort of observation did not occur to 
me at first. But the following notes were made and are here 
recorded in Table 3. The times given are watch times, un- 
corrected. 
TABLE 3. 
— 
Date. Time. Remarks. 
1919, . Me BS 
oat PIS 54 45 | Zenith blue. 
= St ieee 5 00 40/| Could see a pale cirrus about 30° elevation about halfway in 
the W. side of the Square of Pegasus. 
OP Diss a 4 34 80]/ Zenith pale blue. 
|| See ppeecaee 4 51 00/ Cirrus evident as pale spot in east at estimated 45° altitude. 
oh eee AES 4 88 20] Zenith pale blue. 
ae Saige 50 00 | Cirrus visible in east up to estimated 45° altitude. 
MOT Ose ic 3 4 84 15} Zenith pale blue. 
DO swiscscceun 4 37 30 Cirrus visible in east up to estimated 20°. 
May 3055.0 2... 4 29 40/ Zenith pale blue. ; 
Doe 4 87 00 | Cirrus streak between a and Y Cassiopeie. Altitude computed 
80° 237. 
beasehe se ee 27 20) Zenith pale blue. A little before this (perhaps a minute, my 
attention was on the zenith), the high overflow in the east 
from a cumulo-nimbus was dimly visible as brighter than the 
sky at one edge, about 2° lower than @ Arietis, or about 18° 
computed altitude. 
It would seem then that cirrus and cirro-stratus clouds, whose 
average height at Manila is 10.9 kilometers (extremes, 4 and 
18 kilometers), become visible as brighter than the sky at al- 
titudes of 20° to 40° in the east from 6 minutes to 16.5 minutes 
later than the blue appears at the zenith; but that the very 
dense overflow from a cumulo-nimbus may become visible as 
brighter than the sky at about the same altitudes nearly simul- 
taneously with the zenith blue. Hence I conclude that probably 
the blue of the zenith comes from direct sunlight, not from 
twilight reflected from air much higher than cirrus clouds, and 
certainly not from the hazy lower air. 
_ fe mean of all the values of the sun’s altitude given in Table 
2is 18° 38.1". Ido not, however, consider this the most probable 
value, as the effect of most elements involved in a morning ob- 
Servation—such as fall of the time ball, comparison with the 
Watch, hesitation over the appearance of the blue tint, delay in 
“acing the watch, the nearness of the Galaxy and the Zodiacal 
Light, the existence of thin clouds or haze unobservable in the 
169667—_4 
