and on the Variations of the Barometer. 1$ I 



occurred the 7th of the fecond (lunar) month at the time of 

 the lafl quarter *. 



The greateft depreffion of the mercury that had been ob- 

 ferved at Paris for forty years happened January 20, 1791J 

 on the clay following the full moon. 



This coincidence appears to be mod; regular in fair and 

 moderate weather; and, in general, when the barometer fell 

 during the interval between the new or full moon and the 

 quarter, an evident perturbation in the atmofphere accom- 

 panied. I have remarked the fame anomalies in the diurnal 

 variations of the barometer; they are regular in fettled wea- 

 ther, and are difturbed in changeable. 



Mr. Howard draws this general conclufion, that the gra- 

 vity of the atmofphere as indicated by the barometer may be 

 fubjecl to certain periodical changes, effected by a caufe more 

 it.eady and regular than either change of temperature, cur- 

 rents, or folution and precipitation of water, to which the 

 whole variation has been heretofore attributed. 



On comparing ihe mean height of the barometer at each 

 phafe of the moon with the general mean of the year, he 

 found that the mean at the new and full moon was 'conftantly 

 lower, and that at the quarters higher than the general mean, 

 The fame refult attended the fame examination of the ten 

 years obfervations made at London. The barometer, ac- 

 cording to the latter, is depreffed at leaftf four-tenths of an 

 inch (Englifh) while the moon paffes from the two quarters 

 to the conjunction and opposition, and riles in the fame pro- 

 portion in her return to the quarters; an effect which the 

 author attributes to the attraction of the fun and moon for' 

 the matter of the atmofphere, which ought to have its tides 

 as well as the ocean, but lefs confiderable in proportion to 



* Tr appears that the author has here miftaken the numerical diftinc- 

 tion of the month?, as if it were applied to the lunar and not to the calen- 

 dar roont 'S, as the chart lhows it is. As the elevation here mentioned is, 

 perhaps, almoil without a- parallel in the records of fcience, it may not be 

 uninterefting to add a few particulars refpccling it, of which wc were in- 

 formed by Mr. Howard. The latter part of the preceding month had 

 been ftormy, and Iblar halos had appeared on two luccelfive days, followed 

 by rain, fnow, and hail, amidft which the barometer began rifmg rapidly" 

 on the d ly before the full moon. This elevation went on for forty-eight 

 hours, with the wind moitly fouth-weft, and blowing very hard. It af- 

 terwards apparently went round by north to eaft, with frotly nights. The 

 higheft temperature obferved on the day of the maximum was 39 , with 

 a It i 11 atmofphere and a fog. The deprelhon which followed began with 

 the wind at fouth-caft, and thence it veered to the weftward, molt proba. 

 bly by. fouth, as the froft went off with rain. — Edit. 



f In the original it is one-tenth, which appears by reference to the 

 author's i'tatement to be correct — Edit. 



k 3 K* 1 



