W. D. Alexander on the Crater of Haleakala. 45 
IV. After allowing for the previous errors, amounting in the 
aggregate to less than ™! ‘05, the remaining discrepancies of ob- 
servation of invariable stars must arise from variations of the 
apparent brightness of the stars from atmospheric obscuration. 
I —— this into two parts: 
The general and permanent obscuration depending on the 
i Sola of the star. 1 have ascertained that this follows the 
law of refraction; but the coefficient varies in amount on dif- 
ferent evenings, sometimes equalling "50 for an altitude of 20°, 
and at other times not exceeding one-third that amount. I ap- 
ply the correction for obscuration to each star, in reducing the 
observations, so as = eliminate this; and I therefore regard it 
as no part of the error of observation, The correction may be 
easily’ applied atena directly determining the altitudes, by a 
prepar 
2. The temporary and local obseuration, from various atmos- 
pheric causes. By more than 400 observations of 18 stars of the 
2nd to the 4th magnitude, « Persei and 7 Arietis = ap 6x: 
tremes, I found that thea error ofan 
all causes, amounted to ™ ‘ll: “so that the mean error from the ne 
porary and local obscuration alone is on the aver. rage™10, There 
is a great difference between different days; the average mean 
error of all the stars being sometimes as low as ™-07 for a whole 
evening, and at other times for an evening ciate equally 
clear, as high as ™-20, se difference between the stars is no 
less conspicu ous. The stars « and 7 Persei are near each other, 
and nearly of the same he ibe Yet the mean error of: an 
observation of 7 Persei, assuming it to = an invariable star, is 
nearly three times as great as 0 of ¢ Pers 
It is evident therefore that if the pestis I have reached are 
even approximately correct, the errors of Sinsevedion and of | 
the j instrument, if'ordinary care is re a re of no material con- 
sequence ; and accurate results can only he obtained by multi- 
plying observations on different evenings, so as to eliminate as 
far as possible the errors arising from the variations of the sky. 
Art. V.—On the Crater of Haleakala, Island of das pea 
Group; by Prof, W. D, ALEXANDER. (From a letter to one 
of the Editors.) ‘ ‘ 
I HAVE just been in ummer vacation on. Mani, and 
in the anos of it Pagid survey of the great crater of 
Haleaka During the vacation I went three times to the sum- 
mit. The first time I rode up from Makawao before sunrise, and 
mt about seven hours in collecting mineral specimens and 
