Diurnal Variation of Temperature. 105 
_ From considerations analogous to the foregoing, we may 
easily perceive what is the amount by which the mean of deter- 
minations, made at any given hours of the day, will differ from 
the true daily mean. For example, if there be three daily de- 
b 
terminations at the hours h+m, h+n,h+p, their mean will be 
M+aq, sin (h+A+Q,) +09, sit (2h4+B+Q,)+cq, sin (3h+C+Q,) 
+ etc. 
in which 
7, 8in Q,=4 (sinm+sinn+sinp) ; g, sin Q,=} (sin 2m +sin 2n 
+sin 2p) 
7, cos Q, =} (cos m+cosn+cosp); g, cos Q,=+ (cos 2m + cos 2n 
+ cos 2p) ete. 
Supposing the hours of observation to be 7", 14", and 21%, 
as we have arranged them for this country, the corresponding 
quantities become 
g,=0'161, g,=0667; Q.=14", Q,= 8 
g,—0°244, g,=0°311; Q,=—16", Q.=10° 
9,=9°805, g,=0°333; Q,=18", Q,=12? etc. 
_ Were the hours of observation the somewhat more symmet- 
neal, although in the present case less advantageous, ones 7°, 
3", and 21", we should have 
9,=91738, g,=0°577; Q=125 Q.= 6 
97,=0°333, g,=0°644; Q.=18" Q =12" 
9,=0°745, g,=0°333; Q=16"77 Q.= 6 
€ smallness of the coefficients for the earlier variable 
valuable aid in attaining a knowledge of the principal terms 
of the formula, as will hereafter be seen. 
€t us now consider the problem in the form in which it 
Practically offers itself for those places in the Argentine Re- 
oye. at which we have been able to secure three daily obser- 
a pigs A : ; 
Stance which can help to bring us nearer to the truth, yet 
voiding illusory results which although derived from the data 
may have no foundation in the true law 
Which more than three daily observations have been made 
