APPENDIX NO. 8. TEST OF A TRANSIT MICROMETER. 
469 
In June, 1896, the relative personal equation C. H. S.—J. F. H. was determined 
at Washington with a key, using the half-transit method on tw r o nights and on the 
remaining night by simultaneous time sets with tw r o instruments. It w r as found to be 
C. H. S.—J. F. H. = + o s .238 db o s .oo7. 
In this test with the transit micrometer the relative personal equation of the same 
observers (see Table III) is found to be 
C. H. S.—J. F. H. = — o s .o22 db o s .oi4. 
Accuracy as shown by the Residuals in Each Time Set. 
To secure a standard with which to compare the transit micrometer observations, 
19 primary longitudes observed by Coast and Geodetic Survey parties in the interval 
1895-1899 were examined. Nine observers, nearly all having long experience in this 
class of work, were concerned in these 19 determinations. They represent, therefore, 
average work with the key by expert observers. 
The probable error of a single observation of an equatorial star (weight unity)* 
was computed in each of these longitude determinations from the residuals of each star 
from the mean for the time set. Its minimum value for any group of ten nights was 
d=o s .o22, its maximum value =ho s .o72, and ité mean value d=o s .036. If these values be 
reduced in accordance with the scale of weights to correspond to the average probable 
error of a single observation of a star, for stars of such declinations as were used in the 
test of the transit micrometer, thus making a direct comparison possible with the 
values given below in Table V, they become rbo s .025 for the minimum, ±o s .o82 for 
the maximum, and =bo B .04i for the mean. 
In the transit micrometer test all stars were given equal weight regardless of their 
declination. The average value of the probable error of a single observation of a star 
for each observer with the transit micrometer, is given below in Table V for comparison 
with the average d=o s .04i, shown above for key observations. 
* For the scale of weights used, see p. 291 of Appendix 7 of the C. & G. S. Report for 1898, 
“Determination of Time, Longitude, Latitude, and Azimuth.”