THE SCINTILLATION OF THE STARS. 223 



within a feAv seconds; an approximation of a degree or half a degree 

 suffices. 



2. Rediidion of ohservations. — It is well known that, all things be- 

 sides being the same, the scintillation is so much the more feeble in 

 proportion as the stars are nearer the zenith, Avhence it would seem 

 impossible to compare observations if they have not been all made at 

 the same height. 



But, as I shall explain in a subsequent note, in comparing a great 

 number of observations made under highly favorable conditions, inas- 

 much as there had been no apparent atmospheric perturbation, neither 

 for some days preceding nor following, I have ascertained that the 

 scintillation really decreases v.hen the star aproaches the zenith, and 

 that for any height whatever the scintillation is sensibly proportional 

 to the product obtained by multiplying the thickness of the stratum 

 of air which the luminous ray traverses by the astronomical refraction 

 for the height under consideration. 



Let us designate this product by P. By representing the height 

 of the atmosphere by 1, and counting the refractions by sexagesimal 

 seconds, we find that for djfferent heights the values of P are: 



These numbers sufficiently well represent the gradation of the nor- 

 mal scintillation at Merges, when the height of the star above the 

 horizon varies from 20° to 75°. Below 20° the calculated values do 

 not correspond with the observations, but the proximity of the hori- 

 zon affords a sufficient explanation of this deviation, while as regards 

 the stars situated at a greater height than 75°, their scintillation is 

 in general so feeble, that the smallest error in the appreciation of it 

 modifies very considerably the relation of the numbers. 



In this way, when we have observed at a height of 60° a scintilla- 

 tion of 1.6, and would know the scintillation of the same star under 

 the same conditions if it had been but at 45" above the horizon, we 

 should have : 



. . .„ . ,-. 1.6 X 81.8 

 Scintillation at 4o^ =:: ^oTi -— ^^^• 



3. Errors to he avoided. — Often, from one day to another, the scin- 

 tillation varies considerably, but it augments or diminishes in a 

 proportional manner for all the stars, except perhaps for those which, 

 near the horizon, have always a strong scintillation, and aside also 

 from accidental causes which may momentarily modify it. Among 

 these accidental causes we may cite first the crepusculum which 

 almost always considerably augments the scintillation, and secondly 

 the neighborhood of clouds. It was M. Keemtz, I believe, who first 

 observed that the scintillation is greater when there are clouds in the 



