170 



PROCEEDINGS OF THE AMERICAN ACADEMY 



It is apparent from the residuals of Table II. that the standard catalogue, 

 depending wholly on modern observations, is for the most part free both from 

 accidental and periodic errors. 



We must therefore conclude, that if the periodiciti/ is inherent in the observations 

 themselves, it must be produced by a cause common to all instruments and all obseiTers. 

 It would be dilBcult to assign any pliysical cause for the fact, that the zero point 

 of errors depending on the Kight Ascension varies but little from o'', and on the 

 Declination but little from +12°, for all observers and all instruments. No ex- 

 planation of this periodicity is sufficient which does not account for these observed 

 facts. It is a less violent supposition that the errors in question have been entirely 

 transferred from the provisional Right Ascensions on which the clock errors de- 

 pend. Professor Newcomb has shown, tlicoretically, that the errors of single period 

 are eliminated in successive revisions by about l-5th of their entire amount. The 

 errors aie simjily differently distributed by applying to the standard catalogue the 

 corrections given by observations, but without changing ttie zero points. 1 shall 

 now attempt to show why they can never be wholly eliminated by the present 

 method of observation. I begin by comparing with the standard catalogue, the 

 provisional catalog" ''s on which the clock errors depend, for the observatories 

 under consideration. Greenwich, Oxford, and Edinburgii, employ tlie same star 

 places for clock errors, the places being communicated eacli year from Greenwich 

 to Oxford and Edinburgh. Tlie Paris observations depend on Leverrier's places 

 of the fundamental stars, together with a large number of additional stars whose 

 places are given by the observations themselves. Washington II. depends 

 Aviiolly on tiie places of the British Nautical Almanac for 1860. Brussels 

 depends also mainly on the Nautical Almanac. Melbourne depends on the 

 same authority, plus corrections given by the Melbourne observations. Wash- 

 ington I., from 1S612 to 1867, depends on tlie places of the American Ephemeris. 

 From 1867 to 1869, the clock errors depend on the same authority, plus correc- 

 tions given by tiie observations from 1862 to 1865. For 1870 they depend oa 

 Newcomb's catalogue, found in the volume for 1867. Harvard College depends on 

 the "Pulkowa Hauptsterne." The residuals are given in the following table : — 



TABLE III. 



