100 
MR. II. N. DICKSON ON THE CIRCULATION OF THE 
2^ ill the south-eastern area. Temperature has not fallen much in the north-east, 
1jut south-west of Iceland the 10° line has moved a long way to the south-east. 
This line keeps its position in lat. 50° N. on the 40th meridian, hut to the north 
temperature lias fallen in this longitude—note the movement of the o'" isothermal. 
The salinity ol)servations are, unfortunately, not very numerous for this month, 
being the last. They are sufficient, however, to show an immense extension of the 
fresh-water area east and south-east of Newfoundland, and some increase of salinity 
in the south-eastern and eastern areas. 
The most important facts are therefore the extension south-eastward of the cold 
and fresh area from Newfoundland, and the fall of temperature, and probably also of 
salinity, in the region south-east and south of Greenland. 
Except near tlie United States coast, temperature and salinities are lower in 1897 
than in 189G, and this difference is greatly exaggemted east and south-east of New¬ 
foundland, In the east and south-east temperature and salinity are higher than 
in 1896. The line of 8° runs along the east side of the Faeroe-Shetlaud Channel 
instead of tlie west, but the line of 7° is rather more to the north, and the salinity 
seems practically the same. 
V. The Movements uf the Surface Waters. 
The detailed description of the charts has shown that while considerable variations 
of an irregular type occur, there are nevertheless changes over the whole area 
covered which evidently represent a continuous secpience. The charts for each 
month differ from those for the montlis immediately preceding and following in such 
a way as to form a satisfactory intermediate step between the two : they tell, in fact, 
an intelhgil)le story about the distribution of l)oth temperature and salinity, showing 
progressive changes ])earing certain relations to each other, having certain seasonal 
phases wliich occur in l)oth years, and certain features in which the one year differs 
materially from the other. The Important jioint is accordingly established, that the 
metliod em})loyed is adequate to its purpose. Hence, without employing special ships 
or observers, a continuous survey (,)f tlie surface changes in the North Atlantic could 
be kept up with comparatively little trouble or ex}iense : a much larger number oi 
observations than I was able to deal witli could be olitained, worked up, and charted 
for about .£300 a year. 
In attempting to explain changes of temperature and salinity in the surface waters 
witliout a knowledge of those occurring in the layers below the surface, it is necessary 
to consider separately the })robal)le effects on these elements of seasonal changes 
without movement of water, and of horizontal or^ vertical movements with or with¬ 
out seasonal changes siqierposed. The most impoi-tant points ajipear to be these : — 
I. The annual range of temperature increases with the latitude, and the normal 
temperafure gradient northward is greater in winter than in summer. Temperature 
