174 ANNUAL REPORT SMITHSONIAN" INSTITUTION, 1919. 
object glass were shaded from direct solar radiation by a canvas 
screen; but even this was scarcely necessary, for the clouds before 
totality provided a still more efficient screen, and the feeble rays 
which penetrated could not have done any mischief. A heating of 
the mirror by the sun's rays could scarcely have produced a true 
alteration of scale though it might have done harm by altering the 
definition ; the cloud protected us from any trouble of this kind. At 
the Oxford end of the comparison the scale is evidently the same for 
both sets of plates, since they were both taken at night and inter- 
mingled as regards date. 
It thus appears that the check scale is legitimately applicable to 
the eclipse plates. But the method may not be so satisfactory at 
future eclipses, since the particular circumstances at Principe are not 
likely to be reproduced. As regards other sources of systematic 
error, our chief guaranty lies in the comparatively large amount of 
the deflection to be measured, and the test satisfied by the check 
plates that photographs of another field under similar conditions 
show no deflections comparable with those here found. 
Y. GENERAL CONCLUSIONS. 
39. In summarizing the results of the two expeditions, the great- 
est weight must be attached to those obtained with the 4-inch lens 
at Sobral. From the superiority of the images and the larger scale 
of the photographs it was recognized that these would prove to be 
much the most trustworthy. Further, the agreement of the results 
derived independently from the right ascensions and declinations, 
and the accordance of the residuals of the individual stars (p. 152) 
provides a more satisfactory check on the results than was possible 
for the other instruments. 
These plates gave — 
From decliuatious 1. 94" 
From right ascensions 2.06" 
The result from declinations is about twice the weight of that from 
right ascensions, so that the mean result is 1.98", with a probable 
error of about ±0.12". 
The Principe observations were generally interfered with by 
cloud. The unfavorable circumstances were, perhaps, partly compen- 
sated by the advantage of the extremely uniform temperature of the 
island. The deflection obtained was 1.61." 
The probable error is about ±0.30", so that the result has much 
less weight than the preceding. 
Both of these point to the full deflection 1.75" of Einstein's gen- 
eralized relativity theory, the S'obral results definitel}^, and the Prin- 
cipe results perhaps with some uncertainty. There remain the Sobral 
astrographic plates, which gave the deflection 0.93", discordant by 
