308 
MR. GLAISHER ON THE THERMOMETRICAL OBSERVATIONS 
same times, I determined two mean values, the one from the morning observations,, 
and the other from the afternoon observations, and compared the difference between 
them, with the difference as exhibited in my tables ; in nearly all cases the values 
thus found were nearly alike ; hence it appeared that the corrections apply equally 
well to all the years since 1774. 
Having ascertained this fact, I felt I had the means, to a certain extent, of deter- 
mining the quality of the instruments which have been used, as, for the most part, 
different instruments were used in the morning and in the afternoon observations, 
and also the correctness of the observations generally ; and thus the means of ascer- 
taining the value of the results which could be deduced from them. 
I need not mention all the tests to which I have subjected the observations, but 
briefly state that the results were, a conviction on my mind that the instruments had 
been uniformly good, that the observations had been faithfully recorded as read from 
the instruments, and that very great care had been taken in reading at the times 
stated ; the latter circumstance was most satisfactorily proved from the fact of the 
results being the same when the times of observation have been such that the changes 
were rapid, and consequently a small error in the time of reading would have entailed 
a considerable error in the results. 
I found however that during the time the maximum and minimum thermometers 
were in use, their readings were frequently in defect or excess respectively as com- 
pared with those of the other thermometers made during the day, and this was found 
to be more frequently the case in the later than in the earlier observations. On ex- 
amining farther and bearing in mind that a self-registering thermometer, whose read- 
ing is taken once a day only, merely registers the extreme reading which has taken 
place in the preceding twenty-four hours, many of the apparent discrepancies vanished, 
yet still some remain for which I cannot account. I know it has been said that at 
times the sun has shone upon, or its reflected rays have impinged upon the maximum 
thermometer ; but, if this has been the case, I feel certain that it was of rare occur- 
rence, and therefore it is not sufficient alone to warrant us in rejecting a long series 
of observations : possibly no journal of the weather has been kept for any length of 
time, where an attempt has been made to have the instruments properly exposed, 
that such accidents have not happened. 
There being thus three independent methods of determining the mean temperature 
of the air, viz. first, from the morning observations by one thermometer; secondly, from 
the afternoon observations by another thermometer ; and thirdly, from the observa- 
tions made by the maximum and minimum thermometers, I had every means of ascer- 
taining whether the one or the other of these methods was bad. Having satisfied myself 
that the observations were well worth any amount of labour bestowed upon them, I 
became anxious to reduce them to a useful and accessible form ; but the amount of 
work required to reduce observations extending from the year 1774 to 1843, with the 
exception of five years, from 1781 to 1786, during which interval no observations 
