Contributions to Astro -Meteorology. 39 



entirely from meteors seen in evening hours. But the pheno- 

 menon of radiation caused by parallelism of absolute motions 

 should be more distinctly evident as the radiant is nearer the 

 meridian. This occurs in general in the morning hours. 



While, then 5 the existence of rings or streams is a priori pro- 

 bable, and so the existence of radiants for very short periods is 

 to be looked for, yet the series now proposed will doubtless un- 

 dergo essential changes as we accumulate observations, or else 

 some other cause than ring-formations be found to account for 

 the radiation. 



There should in any case be a tendency to a radiation, both 

 from the zenith and from the point to which the earth is moving — 

 hence from the region lying between these points, i. e. from the 

 north-east quarter of the heavens. 



2. Influence of the August and November meteors upon the 

 temperature of the atmosphere. 



Erman early asserted the existence of periods from the 5th to 

 the 12th of February, and from the 10th to the 13th of May. 

 These he attributed to the influence of the August and Novem- 

 ber meteors, assuming that they passed at those epochs between 

 the earth and the sun. To the same cause he attributed certain 

 dark days, and other appearances, said to have occurred in the 

 years a.d. 1106, 1206, 1208, 1706, and 1547. 



In a series of papers read before the Paris Academy of Sciences, 

 and published in the Comptes Rendus*, M. Ch. Ste. -Claire 

 Deville has given the results of an elaborate investigation of the 

 alleged abnormal changes of temperature in these months, as 

 well as those alleged for corresponding days of August and No- 

 vember. 



M. Faye, in response to the first of these papers, shows con- 

 clusively that the dark days &c. adduced by Erman cannot be 

 referred with any probability to the meteors as their cause. 



In his earlier papers, M. Ste. -Claire Deville undertakes to show 

 that there are periodic variations of the temperature of the cri- 

 tical days in February, May, August, and November that corre- 

 spond to secular maxima and minima of the August and No- 

 vember meteors. For the August meteors he assumes a maxi- 

 mum in 1847 or 1848, relying upon the assertion of M. Coulvier- 

 Gravier. 



The existence of a maximum for the meteors in or near those 

 years is in itself exceedingly doubtful, and the evidence adduced 

 by M. Deville to prove corresponding changes of temperature 

 is also very far from satisfactory. 



* Vol. lx. pp. 577, 655, 696, 909 ; vol. lxi. pp. 5, 61, 350 ; vol. lxii. pp. 

 1054, 1 149, 1209 ; and lxhi. p. 1030. 



