ON THE TOTAL SOLAR ECLIPSE OF MAY 17, 1882. 257 
meridian passage of the sun observed with a small meridian circle. The results obtained 
were : 
Latitude : 26° 33' 21" 
Longitude : 1° 57' 40" east of Paris, or 
2° 07' 01" east of Greenwich. 
M. Trepied considers these values as provisional only. 
I observed the first contact through the finder belonging to Mr. Lockyer’s equatoreal 
(aperture 3 inches). As I was, however, not accustomed to use that instrument, 
I forgot that there was a reflecting mirror in the eyepiece, and therefore looked for 
the contact at the wrong place. I first noticed the moon on the solar disc at 
7 h 20 m 14 s (L.M.T.). 
The first contact was observed by M. Trepied at 
7 h 20™ 09 s (L.M.T.). 
The observation for the last contact was, of course, much better. I gave the signal 
at 
9 h 54 m 46 s (L.M.T.). 
M. Trepied gives for this contact 
9 h 54 m 57 s (L.M.T.). 
At the beginning and end of totality the whole party was naturally so much occupied 
with their special work that nobody was available to mark down the time of the 
signals. Mr. Buchanan, however, by means of a stop-watch, measured the time inter¬ 
vening between my two signals, which meant respectively that I saw the corona and 
that totality had begun. That time was 4 seconds, and though with suitable in¬ 
struments the corona might, perhaps, be seen sooner, I feel confident that with the 
naked eye my observation gives a fair estimate of the time of visibility of the corona 
before totality. The time agrees well with that determined by me during the eclipse 
of 1878 in Colorado, where I had measured it as 5 seconds. Mr. Buchanan also 
started a pendulum clock when I gave the sginal for totality, and stopped it again at 
the end of totality. The time of duration of the eclipse was thus determined, and I 
have considerable confidence in the accuracy of the results to within one or two seconds. 
Owing to the brilliancy of the chromosphere and the inner parts of the corona it is by 
no means easy to fix the beginning of totality with absolute accuracy, especially if the 
observer has never previously witnessed an eclipse. The experience gained during 
previous eclipses gives me confidence, however, as to the duration on this occasion. The 
calculated time of totality was 72 seconds, the interval between my signals was 
7 4 seconds, being probably too long, though not by more than one or two seconds. 
As the question of duration of totality has some interest, and as different observers 
differed considerably in their estimate, it may be well to give the opinion of some 
MDCCCLXXXIY. 2 L 
