SURFACE WATERS OF THE NORTH ATLANTIC OCEAN. 
109 
intensity of the Avhole of the Atlantic anticyclone, and of the cyclonic area to the 
north of it being taken into acconnt. The movements of water set up directly Ijy 
these systems are modified by, firstly and chiefly, the configiii'ation of the land, and 
secondly by the etfects of melting of ice. 
The precise effects of variations in the atmospheric systems, which must in the first 
instance be regarded as the independent variables, can ])e ascertained to some extent 
by comparing the circulation in the two years 1896 and 1897. The changes in 
circulation required to account for the observed differences in the distril)ution of 
temperature and salinity have already been suggested, hut it seems desirable to state 
them in a more general form before attempting to discuss their causes. 
The principal point to be considered in the early part of 1897 is the weakness of 
the drift circulation compared to 1896. The south-easterly drift from the north- 
u'estern area is weaker, and the surface waters generally are therefore warmer and 
salter. Again, the easterly drift towai'ds Europe is weaker, there is less “ banking 
up ” of water on the lAiropean coast, and the outflow to the north-east and to the 
Irminger current is weaker ; the main easterly drift appears, in fact, to be further 
south than last year, it consists more exclusively of Gulf Sti'eam water, and its 
course is more towards the African coast. 
The differences are of tlie same general type until August, when the large 
delivery southward in the Azores region becomes most strongly marked. The efiect 
of the decreased strength of the northward streams during spring and early summer 
appears in the diminution in the supply of Polar water ; the melting of ice has 
obviously gone on more slowly, and the increase in the fresh-water streams is smaller 
and occurs later. The difference is least marked in the case of the Labrador current, 
which depends least on the warm streams for its supply. 
In the autumn the movement l)ecomes more easterly and northerly, and the direction 
of the easterly drift is more towards the land in the soutli-west of Europe, causing 
more “ banking-up ” and conse(piently stronger noitherly streams than at the corre¬ 
sponding period of 1896. The change, however, comes too late for these .streams to 
produce the enormous melting of ice and consecpient outlk)w of Polar water observed 
in the previous year, and the phase quickly gives way to tlie drift circidation of 
winter. The characteri.stic “ north and south ” feature becomes strongly developed, 
owing to the form of the autumn distribution just noted, and to the absence of 
Polar water spreading over the surface. 
The construction of charts showing the distribution of atmospheric pressure and 
temperature during individual months is a matter of great difficulty. The discussion 
of material obtained from ships’ logs is beyond the resources of the private investi¬ 
gator; the only means of getting at the information required is to utilise the 
monthly averages of observations made at coast stations surrounding the area, or on 
islands situated within it, and to eke out the information obtained with the excellent 
general summaries puljlished in the ‘ U.S. Pilot Chart,’ and the ‘ Bulletin mensuel du 
