498 ON FINDING THE TIME ACCURATELY 
justified in this conclusion, by the results of many hundred trials, 
verified by calculating as simple altitudes those from which I 
had deduced the time as equal altitudes. Indeed I am rather 
disposed to give the preference to the simple altitudes for ac- 
curacy, if a considerable change of temperature has taken place 
between the morning and evening observations, as the conse- 
quent change of refraction is seldom taken into account in 
the calculation of the equations to equal altitudes, though it 
may produce a sensible effect on the determination of the 
time. To overcome these difficulties, has been my sole.object, 
and my only wish, in making this communication, is to enable 
every observer to put this method to the test, in the hope that 
he may derive the same satisfaction from it that I have done. 
In the morning, when the sun has nearly 10° of altitude, 
and farther from noon than two hours, if far advanced in the 
season, but otherwise, the nearer the prime-vertical the bet- 
ter, I observe eleven successive altitudes of the sun’s lower 
limb, reflected from quicksilver, where the situation will admit 
of it; where it did not, I have employed with equal success 
pure limpid oil covered by one of Troughton’s plate-glass an- 
gular roofs ; but in most situations it may be used in a room, 
without any cover; always taking care, in the morning obser- 
vations, to set the index, with the utmost accuracy, to an even 
10’ or 20’ greater than the sun’s altitude, and then to wait the 
contact. If I use a chronometer, I have an assistant, who 
counts the seconds aloud, which I direct him to note, with the 
fractions, &c. at each observation. In the afternoon I pro- 
ceed in the same way, only setting the sextant 10’ or 20’ less 
than the sun’s altitude, and carefully noting the barometer and 
thermometer, for each series of observations, in order to cor- 
rect the mean refraction. 
I 
