W. Harkness—The Solar Parallax. 379 
that anything has yet been published which noe to show 
how accurately they will furnish the solar paralla 
Photographic observations.—For observing the lane transit of 
Venus there were used at least two kinds of photoheliographs, - 
constructed upon widely different principles. In what follows 
I shall consider only the results yielded by al dic be a of the 
kind get by the United States Transit of Venus ies 
reductions of the United States cain of Venus 
“Seiad are not yet quite completed, it is impossible to 
say exactly what degree of accuracy the paeioeenDys will give; 
but fortunately the same instruments which were used in De- 
cetnber, 1874, to observe the transit of peat at Kerguelen 
Island, Hobart Town and Peking, were used in May, 1878, to 
observe the transit of Mercury at Cambridge, Mass., Washing- 
ton, D.C. and Ann Arbor, Mich.; and as the transit of Mercury 
photographs are completely reduced, Rear Admiral John Rodg- 
ers, Superintendent of the Naval agian) has kindly 
authorized me to make use of the results. They are as follows: 
e total number of plates measured was 119, "of which 25 
were made at Cambridge, 30 at Washington, and 64 at Ann 
Arbor. Each plate was measured by two different persons. 
The errors to be considered are of four different kinds, name 
constant and accidental errors in measuring oe plates, and con- 
stant and accidental errors peculiar to each station. 
Kach plate having been measured in duplicate, if the posi- 
tions of Mercury upon the sun’s disk given by the measures of 
the first observer are subtracted from those given by the meas- 
ures of the second observer, the mean of all the residuals thus 
obtained will be the constant error due to personal equation in 
reading. Its amount for each station is 
In altitude. Tn azimuth. 
Cambridge.) = 5-2 —0"°10 —0” 
Washington... Rog bint —0°09 +0:08 
hes eee Wiehe i +0°15 aie 02 
Thus it appears that, for the mean of the three stations, the 
constant error of reading i is practically zero. 
If the mean of the readings by the two observers is accepted 
as the truth, the probable error of the position of Mercury upon 
the sun’s disk, as determined from a single set of readings by 
one observer, is 
In altitude. In azimuth. 
Cambridge... 6. sss 5. +0"°18 +07:20 
Washingwu ee ee +0°19 0°18 
Axn Arbors te ovise ccs +0°24 +0°28 
The locus of the average probable error of reading therefore 
lies wibis a circle whose radius is 0-21. 
