G8 
MR. IT. N. DICKSON ON THE CIRCULATION OF THE 
because of the number of observations required and partly for the sake of comparison 
with existing charts of “ mean ” conditions and with meteorological charts) the 
“ period of observation” was extended to one month and the “ interval” restricted to 
the same period. Using data kindly furnished by the Danske Meteorologisk Institut, 
the British Meteorological Office, and other similar departments, I found it possible to 
construct, with a fiiir approximation to accuracy, charts showing the distribution of 
surface temperature over the North Atlantic during the months of May, August, and 
November, 1893, and February and May, 1894 (23). Tlie usefulness of these charts. 
notwithstanding their being altog 
O & O 
ether of the nature of a first attempt, has been 
shown by Pettersson (24) and Meinardus (25). The attempt to construct similar 
charts of surface sallnitv was unsuccessful, tiie material, consisting of rough hvdro- 
meter deteiinlnations made on ])oard ship, proving, naturally, perfectU worthless. 
(It wars not uncommon to find salinities of 32 and 39 j)ro mille within a few days and 
miles of each other in the middle of the Atlantic.) 
Tlie preliminary expeibnent being so far successful, it seemed worth while to make 
a continuous series of monthly temperature charts extending over a considerable 
period, and if possible to obtain material for adding charts of salinity for the same 
tiine. 
II. Material used in Constructing the Charts. 
Having decided to attem])t the preparation of charts showing the distribution of 
surface temperature and salinity for each month of the two years 189G and 1897, a 
study of existiiig charts of mean surface movement and of my 1893 and 1894 charts 
satisfied me that the parallel of 40'’ N. lat. should be retained as the southern 
Ijoundary. The Meteorological Council agreed to furnish me with extracts from logs 
of all temperature observations niade north of that line during the years as they were 
received. The copying and arranging lias Ijeen done throughout b 3 'Mr. W. G. James, 
of the Meteorological Office, and the observations extend over 600 sheets, roughly 
16,000 observations, forming the greater part of the temperature data utilised. In 
addition to the Meteorological Office oljservations the temperature values on the 
maps published monthly Iq^ the Danish Meteorological Department were inserted on 
the charts, along with occasional data, pidilished and in manuscript, kindh" sent to 
me 1)}^ the United States Hydrographic Department, the Bureau Central Meteoro¬ 
logy jue de France, and by Professor Petterssox. iMany of the temperature 
observations received directly from the observei's collecting samples of water were 
also sent to the Meteorological Gffice and came to me in duplicate, but a large 
number of additional observations was received and jJaced on the charts. In some 
cases verified thermometers were supplied to observers. 
Temperatures were all reduced to the Centigrade scale before charting. Apart 
from the necessity for making tlie \vork comparable with that of workers in other 
countries, 1 found the difference of temperature represented b}' 1° C. better adapted 
