215 
made with an invariable pendulum. 
not so fit to be compared with those made in England, as 
with others made on a similar volcanic soil. 
The range in the temperature in 24 hours was from 74 0 to 
91 0 , and as we were obliged to place the instruments in a 
tent, the temperature rose in the day time, and fell at night, 
but without any uniformity. On the first day of observing 
coincidences, a set was taken after breakfast, and another 
before dinner ; but as it was soon seen that this would be to 
confine the whole of the observations to the hot period of the 
day, it was determined in future to take one set as soon after 
sun-rise as possible, in order to have a result which should be 
influenced by the whole night’s continued low temperature ; 
and another set towards the close of the day, in order to have 
a result partaking in like manner of the influence which the 
whole day’s high temperature might have on the length of 
the pendulum. I also endeavoured so to arrange things, that 
I should catch a sufficiently long period of uniform tem- 
perature during the interval of each set, that it might be 
taken with an unvarying thermometer ; hoping that by these 
arrangements, although no one experiment could produce 
strictly correct results, the opposite errors of the morning 
and evening observations would counterbalance one another ; 
that is, that the mean, between observations taken both in 
the hot and in the cold periods of the day, would probably 
give a just result ; or at least such a result as would fairly 
be entitled to stand by the side of rates deduced from transits 
of stars, the intervals between taking which, in like manner, 
included the same extremes of temperature. 
It should be borne in mind that the real desideratum, as far 
as respects rate, is not to know what is the aggregate loss or 
gain of the clock in twenty-four hours, but the actual rate at 
