6 ROYAL SOCIETY OF CANADA 
The difference for each day is obtained by averaging the twenty- 
four hours values read off from the trace at hourly intervals. The 
daily differences for the month are then averaged, giving the value to 
be found in the table; the values are therefore each a mean of about 
720 readings. 
TABLE. 
1903-04. 1904-05. 1905-06. 1906-07. 
SL ureter ai ian ee) —2.6 
AUS UT) SA AU —2.5 
September .. —5.8 —4.4 
October... DR ae Pei 
November, ..7 uj.) 24 2... | 4.6 —5.7 
December sos —8.2 —3.0 —8.2 
January.. .. —11.2 2.2 —6.3 
DORE CIE EN à NS M RP PA —2.7 —1.6 —4.3 
Mare are ie MS nin = Oe —2.0 —2.5 —2.9 
DPI Men oat agit ORO —3.1 —3.7 —4.6 
JE AO OR OS RNR RE bay —?.6 
SCG Ney CNE Mier eet pe —2.2 —2.5 
It will be at once apparent on inspecting the table that the values 
for 1903-04 are much greater than anything which followed up to Sep- 
tember, 1906. After this date the differences increase again. This 
means that on the average the top of the mountain was maintained 
much colder in comparison with the Observatory during the year 
1903-04 than for the two following years. During the latter part of 
1906 and the months of 1907 so far as obtained there appears to be a 
recurrence of the lower values. 
There is a certain amount of resemblance between the magnitude 
of these differences and the deviation of the mean monthly difference 
at the Observatory from the mean for the previous thirty years. That 
is to say, we observe that a greater average negative difference between 
the Observatory and the top of the mountain is accompanied or followed 
by a monthly temperature below the previous yearly averages, or in 
other words we have cool weather conditions. We observe also that a 
small difference in temperature between our high and low level stations 
is accompanied or followed by warmer temperature conditions. 
This may best be illustrated by the following curves. In Figure 1 
we plot the difference for each month from the table. On the same 
scale we plot the differences obtained by subtracting the monthly aver- 
age temperature from the average for that same month over a period 
of thirty or thirty-two years. In these curves the break in continuity 
