■28 ~ PHILOSOPHICAL TRANSACTIONS. [aNNO 1763. 



sun's parallax = 8''.57 ; and if we reject 8 of them, which are marked with the 

 letter r, and which differ more than tV of a second from the mean of the whole, 

 the mean of the remaining 13 results gives the sun's parallax 8". 57, differing only 

 one hundredth part of a second from that which was determined from the obser- 

 vation at the Cape, and agreeing in a most surprizing manner with what was 

 formerly determined by the comparisons of the observations at places on this side 

 of the line only, where the base was so small, as before said; a most convincing 

 proof of the great precision with which the parallax of the sun is determined by 

 the late transit of Venus. 



We shall now inquire into the limits of the error that may attend the determi- 

 nation of the parallax by the observation of the internal contact. An error of 

 jm jQs qP j^jjj^g j,^ j.j^g observation at Tobolsk, when compared with the 

 observation at the Cape, will produce an error of I" in the sun's parallax : and if 

 we suppose an error of 35' of time in the observation at Tobolsk, and an error 

 of the same quantity in the observation at the Cape, and both in contrary direc- 

 tions, this also will produce an error of only 1" in the sun's parallax. If there- 

 fore no greater error could be committed in the observations at Tobolsk and the 

 Cape, we aie certain that the comparison of Tobolsk and the Cape gives the 

 sun's parallax so exact, that the error does not exceed one second from the true 

 parallax. But this is too great an error to be supposed in the observations, be- 

 cause Mr. S. has shown, in the former paper, that an error of only 6' in time 

 was committed in the observation of the contact by persons observing even in 

 the same place; therefore if we suppose an error of 6' of time in the observation 

 at Tobolsk, and an error of the same quantity in the observation at the Cape, 

 and both in contrary directions, the error produced in the parallax by those 12% 

 will amount only to ^ of a second, even though we had only these two observa- 

 tions to determine the sun's parallax: but since we have a great number of very 

 good observations, made at other places, it follows that the mean of all these 

 must give the sun's parallax to a less error than -^ of a second, and consequently 

 very near the truth. 



In all places where the internal contact at the egress was observed, and where 

 there were more observers than one, we find a difference in the time of each ob- 

 server; the observation at Greenwich is an exception to this, as the 3 observers 

 all agree to the same second, in the observation of the contact of Venus with the 

 sun's limb; which is the more surprizing as they used telescopes of different 

 constructions and of different magnifying powers. This coincidence not only 

 surprized Mr. S., but also the Rev. Mr. Hornsby, now Savilian professor of as- 

 tronomy at Oxford. Mr. Hornsby went to Greenwich in the beginning of the 

 year 1 762, and on his return told Mr. S. that his surprize was at an end, for 

 he had been informed at Greenwich, that Mr. Green, the assistant observer 

 there, as soon as he judged that the internal contact was formed, called out now- 



