106 Proceedings of Royal Society of Edinburgh. [sess. 
plied to the Registrar-General by the Edinburgh observers of the 
Scottish Meteorological Society. The observations were taken at 
Cumin Place until 1886, when the station was changed to Black et 
Place, where they are still carried on.* These two stations are 
only half a mile from each other and at the same height above 
mean sea-level, the readings being thus strictly comparable. In 
order to supplement the information given weekly to the Registrar- 
General, monthly means of the more important climatic elements 
have been computed for the seventeen years, thus enabling a 
tolerably complete representation of the climatic features to be 
given. Some of the monthly means have been compared with 
their averages for long periods, so that some idea might be obtained 
as to the departures from the normal during the period under con- 
sideration (see Table III.). Tables for each week, showing the 
cumulative variability of temperature, have been computed. These 
values are found by taking the day to day differences in the mean 
temperature f of successive days and adding up the seven daily 
values during the week. The variability of temperature is largely 
the result of changes in the direction of the wind. JSTo information 
is available for the period under review as to the height of the 
underground water or the percentage of gases, as carbonic acid gas 
and other deleterious substances in the underground air. This is 
a matter for regret, as it is generally believed that the state of 
public health depends in no inconsiderable degree on the chemical 
and bacteriological condition of the superficial layers of the earth. 
The curves which accompany this paper were prepared as 
follows : — The number of deaths from each disease having been 
extracted for each week during the seventeen years, the values 
were added up and divided by fifty-two in order to obtain the 
average weekly death-rate (see Table I.). The percentage excess 
or defect of each week’s mean from the average of the whole year 
was then calculated, and the values smoothed by Bloxam’s method. 
This process consists in assuming the average, for example, of the 
2nd week of January to be not the actual average of that week, 
* For some years after 1886, observations were also made for the Scottish 
Meteorological Society by Mr Buchanan at Oswald Road. 
t The mean temperature was assumed to be the arithmetical mean of the 
maximum and minimum readings. 
