D. P. Todd—Solar Parallax from the Velocity of Light. 61 
dredth part. I think I shall not be regarded far wrong in 
assigning a probable error twice as great. is o in con- 
sideration of the very, unfavorable limitations under which the 
determination was executed, and quite independently of what 
has since become known 
Il. The first determination by Cornu, related in the Journal 
de I’Ecole Polytechnique, xliv cahier, vol. xxvii, 1874. The 
resulting velocity of light is 298,500*+-1,000. 
Ill. The s etermination by Cornu, related in the 
Annales de |’Observatoire de Paris, Mémoires, tome xiii, 1876. 
The resulting velocity of light is 300,400*+300. Helmert has 
given a rediscussion of these experiments in the Astronomische 
Nachrichten, vol. lxxxvii, 1876. His interpretation assigns the 
velocity 299,990 kilometers, the probable error of which I have 
estimated at 200 kilometers. 
he first determination by Master A. A. Michelson, U. 
S. Navy.* The resulting velocity of light is 300,100 kilo- 
meters. I consider this determination of equal weight with 
that of Foucault, and with the first determination by Cornu. 
V. The second determination by Mr. Michelson. The num- 
ber of this Journal for November, 1879, contains a brief recital 
of these experiments. I shall adopt, for the purposes of this 
paper, the results given on the “corrected slip,” for his second 
determination of this constant of velocity. 
The largest result for velocity of light, 800,142 kilometers. 
‘The smallest result for velocity of light, 299,692 kilometers. 
Giving equal weight to the one hundred separate determina- 
tions, the resulting velocity of light is 299,930 kilometers per 
ond. I find that the computed probable error of this deter- 
mination is no larger than six kilometers. But, in the deter- 
mination of almost no other astronomical or physical constant 
should we consider the computation of probable error of the 
final result from a corresponding number of observations quite 
so illusory. In consideration of all the sources of error, acci- 
dental and systematic, I think the probable error of this result 
may be estimated at 100 kilometers. 
All these several determinations of the velocity of light are 
now combined as follow :— 
I. 298000 + 1000 Weight 1 
IL 298500 + 1000 1 
Il. 299990 + 200 25 
IV. 800100 + 1000 1 
a 299930 + 100 100 
The resulting most probable velocity of light is 
299,920 kilometers = 186,360 miles per second. 
* Proceedings of the American Association for the Advancement of Science 
vol. xxvii, 1878. 
