30 PHILOSOPHICAL TRANSACTIONS. [aNNO \yQ3. 



the ingress at both places, by the method of M. Pingre above mentioned. There- 

 fore the difference of longitude between the Cape and Tobolsk is very exactly 

 settled, so exactly, that Mr. S. was persuaded that the error does not amount to 

 5 or 6 seconds. Therefore the error in the parallax arising from the error of the 

 difference of longitude is extremely small, scarcely amounting to -^ part of a 

 second. Therefore we are certain that the error in the sun's parallax, arising 

 both from the error of observation and the error of longitude, does not exceed -; 

 of a second in the comparison of the observations of the internal contact at To- 

 bolsk and at the Cape, even though we had no more observations to determine 

 the sun's parallax ; but the mean of a great many more must bring it very near 

 the truth. 



Mr. S. now proceeds to determine the parallax of the sun from the total du- 

 rations observed at different places. If therefore we compare the durations ob- 

 served at Tobolsk, Cajaneburg, Abo, and Tornea, with the durations observed at 

 Madras, Grand Mount, and Tranquebar, which give the greatest differences, the 

 results of the sun's parallax will be as follow. 

 Sun's Par. Dif. of du. 

 Tobolsk and Madras . . = 9".6'l 

 G. Mount = 8.33 

 Tranquebar = 8.5C 



Cajaneburg and Madras = lO.Oiyr. 

 G. Mount = 8.00 

 Tranquebar = 8.33 



The mean of these 12 results gives the sun's parallax = 8 '.68 ; and if we reject 

 4 of them, which are marked with the letter r, and which differ the most from 

 the rest, the mean of the remaining 8 gives the sun's parallax = 8".6l . This de- 

 termination of the sun's parallax cannot be depended on to any great precision, 

 because of the small differences between the durations compared, the greatest of 

 which amounts only to 2™ 50% and also because of the small number of compa- 

 risons. It serves only to show nearly what is the quantity of the sun's parallax. 



We are now to determine the limits of the error in the determination of the 

 sun's parallax by the durations observed at two different places. The greatest 

 difference of duration is between Tobolsk and Madras, which amounts onlv to 

 2*" 50'. If therefore an error of 20' in time is committed in the observations of 

 the ingress and egress, at both the places compared, this error of .<y in time will 

 cause an error of I" in the result of the sun's parallax ; and in the comparisons 

 of these places where the difference of duration is less, will occasion a greater 

 error; and therefore the determination of the parallax, by this method, cannot be 

 depended on to any great exactness, because of the small differences of the du- 

 rations compared. In this method however we are free from the uncertainty 

 arising from the difference of longitude not being exactly known. 



