300 PROFESSOR FORBES ON THE CLIMATE OF EDINBURGH. 
59. The preceding table (pp. 348—49) contains the synopsis of the whole thermo- 
metric observations described in Section 2, arranged under each day of the year, 
and divided into four decennial periods. From these partial results, the great 
length of period required to determine the mean temperature of any given day 
may be estimated. I have farther added the extremes of the observed mean 
temperature of the day during forty years, and the difference of these, which I 
call the “fluctuation.” This uncertainty as to the mean temperature of a given 
day amounts occasionally in winter and spring to 30° Fahr. or more. 
60. The mean daily temperatures in the preceding Table are projected on a 
large scale in Plate XVIII. Notwithstanding the casual sinuosities of the curve of 
temperature from day to day, the interpolating curve derived from the formula 
obtained in Art. 56 evidently represents the course of temperature with general 
exactness, and the temperature of a given day may be estimated from it with far 
more accuracy than by the observed mean of that day only, even if derived from 
forty years’ observations. 
61. Nevertheless, it is interesting and important to ascertain whether there are 
not partial inflections of the annual curve, subject to recur, and which cannot be 
satisfactorily represented by the usual periodic series. These are fitly termed by 
M. QUETELET “ periodic anomalies.” * 
62. With this object in view, as well as to facilitate a comparison of the Edin- 
burgh temperatures with observations made elsewhere, I caused the jive-day 
means to be taken from one end of the year to the other (derived from the collec- 
tive observations of 40 years), in the same way that was done by the earlier 
meteorologists, and more lately by M. Dove, who has published a considerable 
collection of such results.+ By this process the sinuosities of the larger curve 
are very much reduced, and anything like a systematic deviation from symmetry 
is more clearly shown. 
63. The results are contained in the following table, and are represented in 
Plate XIX. 
* Memoire, &c., p. 4. 
+ Dove, Fiinftigige Mittel, &c. Folio, Berlin; and in the Tables of the Prussian Statistical 
Bureau. Folio, 1858, 
