SOLAli EOLIPSE, MAY 0, 1SS,3. 



57 



The foUowiiif; tabic coiitaitis the reduced readiiij;s for May (I, and a coiiii)aii.soii w itli values 

 iiitcr|)olatcd IVoiii tlu' iiicau values above given. 



BAROMETKIC PKESSUKE 



The above tables show tbat on all days but May 6 there was a steady fall iu pressure, and 

 therefore the observed rise on that (biy \vas peculiar to the day. The last column of the second 

 table indicates a more rapid fall than the average during the fust partial phase, and also tliat the 

 time of the maximum difference from the average course occurred at about 12.10 p. m., or thirty- 

 three minutes after the close of the total i>hase. The highest reading was recorded at 11.50 a. ni., 

 or tliirtcen minutes after the end of totality. 



While the rise in pressure was small, being at its maximum only 0.019 above the average 

 value, it is yet greater than the errors of observation. The instrument used can be read accu- 

 rately within 0.005, and as it was kept at nearly uniform temperature during the period of obser- 

 vation, the possible uncertainty from failuie of the attached thermometer to record the tempera- 

 ture of the instrument can scarcely amount to 0.005. On the supposition that the errors have the 

 same sign, there still remains 0.01 inch which is not accounted for, and the observed progression 

 depends upon six or more observations. It is not difUcult to imagine also that the inflow of air 

 towards the path of the shadow might be snIUcient to cause an increased pressure suscei)tible of 

 measurement l\y a sensitive barometer. 



TEMPERATURE. 



The eft'ect of the eclipse in reducing the temperature of the air may be shown by comparing 

 the thermometric readings with tliose of other days. The mean values of the observed tempera- 

 ture, April 25 througli INIay 5, omitting the rainy day, May 4, arc as follows : 



The following table contains a comparison between the observed readings on Jlay G and 

 values interpolated from the above averages. It will be seen from the column of ditferences that 

 at 9.57 a. m., before the beginning of the eclipse, the observed reading was 1^.9 below the aver- 

 S. Mis. 110 8 



