in all cases expressed according to a common standard of temperature), such a close conformity subsists between the salinity of the equatorial water of the surface and that of the polar waters of the bottom, as can scarcely be accounted for in any other way than by the continual and tolerably rapid ascent of the latter. Another indication of this ascent is given by the moder ation of the surface-temperature of oceanic water, even under the equator. If there were no ascent of colder water from beneath, there seems no reason why the constant powerful insolation to which equatorial water is subjected should not raise the temperature of its surface to the highest possible elevation. The limit to that elevation, which is obviously set by the cooling influence of evapora tion, is probably that which is met with in the Red Sea, where the monthly average for August rises to 86| and for September to 88, whilst the maxima rise much higher, temperatures of 100, 106, 100, and 96 having been noted on four consecutive days. Moreover, along the Guinea Coast, and especially in the Bight of Biafra, the surface- temperature is stated to range as high as 90. But in these cases there is no reduction of surface-temperature by the upward movement of polar water ; for this is altogether excluded from the Red Sea by the shallowness of the Strait of Babelmandeb, whilst the depth of the bottom along the Guinea Coast is too small to allow of its being overflowed by the glacial stratum. Now, over the deeper parts of the equatorial Atlantic the surface-temperature usually ranges between 75 and 80; and this is its ordi nary range in the Mediterranean during the months of August and September. That the temperature of an equa torial ocean should be thus kept down to that of a sea of which the greater part lies between the parallels of 40 and 35, can scarcely be accounted for in any other way than by the continual uprising of polar waters from beneath. The same principle, once admitted, fully accounts for that amelioration of the cold of north-western Europe, which (as already shown) cannot be fairly attributed to the Florida Current or true Gulf Stream. For it is obvious that a continual efflux of the lower stratum from the polar areas towards the equatorial must involve a continual indraught of the upper stratum towards the polar areas ; and this indraught will be much more marked in the Northern than in the Southern Atlantic, on account of the progressive narrowing of the former, whilst the latter progressively widens out. Of such a slow northerly set of a stratum of water, extending downwards to a depth of at least 600 fathoms, we have evidence in a comparison of the temperature-soundings taken in the " Porcupine " expedi-
[chart]
tions of 1869 arid 1870, between the coast of Portugal (34 N.) and the Faroe Islands (59^- N.), from which the section fig. 5, has been worked out. For it is there seen that, although the surface-temperature is reduced by the thinning-out of the superficial stratum, there is but a slight change in the position of the bathymetrical isotherms of 45 and 40 ; so that there is an obvious continuity of a stratum of many hundred fathoms thick ness between these two points, notwithstanding their sepa ration by 2o of latitude. The contrast between the position of the isotherm of 40 at 800 fathoms depth off the Faroes, and its position at less than 300 fathoms depth under the equator, is most remarkable. We have seen that the isocheimal in the latter area would not be below 75, and yet we find water colder than 40 lying at within 300 fathoms of the surface; whilst, on the other hand, the normal isocheimal at 59| N. would certainly be below 40 (probably no more than 35), and yet we find water above 43 extending downwards to 600 fathoms, and water above 40 D to 800 fathoms. Thus the vertical oceanic circulation carries a vast mass of water which is beloiv the normal off the coast of Portugal, into a region where it is above the normal, with very little loss of heat by the way, except in its surface-film ; and a little consideration will show that such a movement must be much more effectual as a heater than a corresponding move ment of a thin stratum of much warmer water. For the latter, when it passes beneath an atmosphere much colder than itself, will soon be brought down to a like standard, not having warmer water from below to take its place when it has been cooled down ; whilst in the former, each sur face-layer, when cooled below the temperature of the warmer stratum beneath, will sink and be replaced by it. Now since the true Gulf Stream, when we last know it, has been so thinned out that it could not long retain any excess of temperature, it seems inconceivable that it should exert any decided effect on the temperature of the Faroes and the coast of Norway, unless (as supposed by Dr Petermann and Professor Wyville Thomson) its thickness undergoes an increase from less than 100 fathoms to 600. But since the course of Dr Petermann s isotherms shows that the north ward flow extends across the whole breadth of the Atlantic between Newfoundland and the British Isles a distance of about 2000 miles we are required to believe that a rivulet (for such it is by comparison) of 60 miles breadth and 100 fathoms depth (see section, fig. 5), of which the greater part turns southwards round the Azores, and of which the remainder is flowing due east when we last recognise it, is able to impart a northerly movement to a stratum of 2000 miles in breadth, and at least 600 fathoms depth. On the other hand, the eastward set of this stratum, considered as a northward indraught into the polar area, is readily accounted for by the excess of easterly momentum which it derives from the earth s rotation, this being only half as rapid in lat. 60 as it is under the equator ; and since there is a still more rapid reduction in the rate of this rotation in yet higher latitudes, the con tinually increasing excess of easterly momentum will give to the northward flow a progressively stronger eastward set. On the other hand, the deficiency of easterly momentum in the cold underflow coming from the pole towards the equator will tend to produce a lagging-behind, or westward set of that underflow ; and this has been shown by the " Challenger" temperature-soundings to be the case, the cold deep strata of the Western Atlantic surging upwards along the slope of the North American coast-line, as is shown in fig. 6, where we see not only the bathymetrical isotherms of 60, 55, and 50, but the yet deeper isotherms of 45 and 40, successively rising to the surface as we approach the land ; while at a depth of only 83 fathoms, a temperature of 35 was encountered, which, at no great distance to the south, would only be found at a depth of 2000 fathoms. That the cold water should thus run up-
Ill 4
