depth, by which the water which flows in from the Atlantic is restored to that ocean." The explanation of the Gibraltar surface-current that has found most general acceptance is that first offered by Dr. Halley; who attributed it to the excess of evaporation from the surface of the Mediterranean, caused by the combined action of the powerful sun of that latitude and of the hot winds of Africa, over the whole amount returned to its basin, either directly by rainfall, or by the rivers which discharge themselves into it; whereby its level would be progressively reduced, if it were not maintained by an inflow from the Atlantic. The obvious objection to this explanation, taken by itself, is, that as the water which passes off by evaporation leaves all its salt behind it, and as the water which enters through the Strait brings its own salt with it, there must be a progressive increase in the density of the water of the Mediterranean, until it reaches the point of saturation. And although the water at and near the surface shows no such excess, it has been seriously maintained that the waters of its abyssal depths may be so highly charged with salt as to become saturated with it, and even to be depositing it at the bottom of the basin. The condition of the Red Sea is even more peculiar than that of the Mediterranean. For whilst the evaporation from its surface is enormous -- annually amounting, it has been estimated, to a layer of not less than eight feet in thickness. scarcely any of the fresh water thus withdrawn is returned to its basin either by rain or rivers. But the level is kept up by a strong current that continually sets inwards from the Arabian Gulf through the Strait of Babelmandeb, and the water thus flowing in is continually adding more salt to that which the basin already contains. Now, if there were no outflow of the water which is concentrated by evaporation, there would be a rapid accumulation of salt in the Red Sea; which, so far as can be judged from the surface-water, does not prove to be the case. Such an outflow, as Captain Maury has well argued,* can only be by means of an undercurrent; and he applies the same doctrine to the case of the Mediterranean, getting over the difficulty of the ridge" or "reef" by citing numerous cases in which a deep current is brought up, by the con 66 formation of the bottom, to comparatively near the surface. This, then, was the state of the question we were charged with investigating; and the result of our inquiries has not only cleared up the whole difficulty of this particular case, but has furnished a definite clue to the investigation of a subject of very much larger importance-the general Oceanic circulation. I shall present these results, not in the order in which they were obtained, but in that which will best enable me to explain the rationale of the under-current, whose existence they have demonstrated. We found, by the double evidence of chemical analysis and specific gravity, that there is a certain small excess of salt in the waters of the Mediterranean over that of the Atlantic, and that this excess is somewhat greater in its bottom-water than in water taken near its surface; whilst in the Atlantic, where any difference presents itself, it is rather in the contrary direction. But the density of none of the numerous samples of bottom-water which we took up at various parts of the Mediterranean basin, from depths varying between 207 and 1,743 fathoms, was such as to give any sanction to the notion that there is a progressive increase in the quantity of salt it contains. On the contrary, the results of our enquiries seemed to prove that while there is an excess of evaporation from the surface of the Mediterranean above the quantity of fresh water returned into its basin, the augmentation of density thus occasioned is in some way restrained within a very narrow limit. We found no difficulty in distinguishing Mediterranean from Atlantic water by the excess of salt which it contained; and the testimony to "character” thus given by the hydrometer and the test-tube stood us in good stead, not only by affording valuable confirmation to the direct evidence of motion which we obtained by other means, but by supplying the want of that evidence where circumstances prevented us from obtaining it. The Atlantic entrance of the Strait of Gibraltar, bounded by Cape Trafalgar on the north and by Cape Spartel on the south is about twenty-two miles wide; but there is a gradual narrowing within as far as Capes Tarifa and Alcazar, between which the width is only about nine miles; whilst between the "Pillars of Hercules," which bound the Mediterranean entrance, the distance is about twelve miles. The eastern end of the strait is the deepest the mid-channel opposite Gibraltar having 1 a depth of more than 500 fathoms. From character. And the very curious fact was this the bottom gradually but irregularly determined (two sets of observations, slopes upwards towards the western end, made at an interval of six weeks, giving where what has been spoken of as the identical results) that the water at 250 "reef" or "ridge" presents itself. These fathoms is considerably denser than the terms, however, give a very erroneous idea water at 500 fathoms. Now this fact, of the real nature of the bottom. The taken by itself, would serve to indicate water along the northern or Spanish side that the middle stratum is in motion over of the strait is there comparatively shallow, the stratum below it, since its superposithe 100-fathom line running at about tion could not otherwise be maintained; twelve miles' distance from Cape Trafal- and as it unquestionably consists of Medgar; but it should be remembered that iterrean water, that motion must be from even this depth is equal to that of the east to west, or in opposition to the surfacegreater part of the North Sea and the current. This inference was fully borne British Channel, which scarcely anywhere out by the experiments made with a "curexceeds 100 fathoms. The southern or rent-drag" devised by our excellent comAfrican half of the channel is considerably mander, Captain Calver, whose large hydeeper; the 100-fathom line running at drographical experience and practical two miles from Cape Spartel, whilst be- ability were here peculiarly valuable. tween this and the opposite 100-fathom This "current-drag was so constituted boundary there is an irregular rocky chan- as to present a resisting surface so much nel, some parts of which are comparatively larger than that of the boat from which it shallow, whilst in others there is a depth was suspended that, although the latter of nearly 200 fathoms. From this ridge, would tend to move in the direction of the which constitutes a kind of marine surface-current in which it was allowed "watershed" separating the basin of the freely to float, this tendency would be Mediterranean from that of the Atlantic, more or less counteracted by the action the bottom slopes downwards on the At-of any under-current which the currentlantic side, at nearly the same inclination drag, when lowered to any given depth, which it has towards the Mediterranean; might there meet with; and thus the moso that at about forty miles westward the tion of the boat would be retarded, comdepth is equal to that of the deepest part pletely checked, or even reversed, accordof the channel between Gibraltar and ing to the strength of such under-current. Ceuta. Now, taking this depth at 500 In our first set of experiments, in which a fathoms, or 3,000 feet, and taking the aver-westerly wind combined with the surfaceage depth of the channel between the two current to carry the boat inward, the force 100-fathom lines at the western end of the of the under-current at a depth of 250 Strait to be 150 fathoms, or 900 feet, the fathoms did not quite suffice to check its difference between these viz., 350 fath- motion; whilst with the "current-drag" oms, or 2,100 feet is the elevation of the at 400 fathoms the inward drift of the so-called "ridge," which has been sup- boat was more perceptible. Still its reposed to prevent the outflow of an under- tardation was such as to indicate that the current. But as the length of the incline water beneath could not be stationary, leading up to this "ridge" is 35 miles, the but must be moving in a direction opposed rise is only 60 feet per mile, or 1 foot in to that of the surface-current; since its 88-an extremely moderate gradient. action on the "current-drag" so nearly And further, as the average depth of water neutralized not merely the action of the on this " ridge" is fully double the aver- wind and surface-current on the boat from age depth of Dover Strait, and as there which it hung, but that of the whole upper are channels across it whose depth reaches stratum on the suspending rope. This 1,200 feet, the impossibility of the passage inference was fully justified by the result of an under-current from the Mediterra- of our second set of experiments; for the nean into the Atlantic appears to have wind being then from the east, the rate of been somewhat hastily assumed. the surface-current was reduced by it, whilst it so far neutralized the action of that current on the boat, as to leave it more free to be influenced by the "current-drag." We then had the satisfaction of seeing our boat carried outwards, or towards the Atlantic, against the powerful stream in which it floated, by the invisible agency of a current flowing at a depth of Our own investigations commenced at the eastern extremity of the Strait; and we found that whilst both the temperature and the density of the surface-water distinctly indicated its Atlantic source, the temperature and density of the water at between 200 and 500 fathoms' depth no less distinctly indicated its Mediterranean Having thus settled the facts of the case, we were led to consider their rationale; and found reason to agree fully with Captain Maury in the explanation he offered. As the water of the Mediterranean basin is continually losing by evaporation a larger amount than is returned to it by rain or rivers, an inflow from the Atlantic must take place to keep up its level; as was long since maintained by Dr. Halley. Now, if this inflow consisted of fresh water, the total quantity of salt in the Mediterranean basin would remain the same; the density of the Mediterranean water would consequently undergo no increase; and a column of any given depth on the Mediterranean side of the Strait would be in equi 250 fathoms, or more than a quarter of a mile refer more particularly), it had gone up beneath the surface. And when the "cur- and down vertically nearly a hundred miles; rent-drag" was let down to a depth of 400 and had so well agreed with its fellow fathoms, or nearly half a mile, the westerly throughout, that the difference between movement of the boat still gave distinct them never amounted to a degree (Fahr.). evidence of an outflow, although its rate When, therefore, we found ourselves anwas reduced. chored by our sounding-line, and no maThe existence of an under-current, car-noeuvring of our vessel proved able to rerying forth the Mediterranean water ren- lease it, nothing remained but to free dered denser by evaporation, to discharge ourselves by putting a "breaking-strain it into the great Atlantic basin, thus be- upon the line (which was almost like a tug came a demonstrated fact; and the whole at our heart-strings), leaving our lead, wadifficulty of the case vanished, save that ter-bottle, and thermometer to be a puzzle which had been thought to arise out of the to the fishes of the present, and a prize to obstruction created by the "ridge" at the the geologist of the future. opposite end of the strait. It was obvious, however, that this difficulty was rather apparent than real. For not only had the greatest strength of the under-current in the deepest part of the channel been found to lie at a depth very little exceeding that of the "ridge," but the water of which it consists had been found to be of the densest anywhere found in the Mediterranean. Feeling it right, however, to do all we could to satisfy objections raised by such authorities as Sir C. Lyell and Professor Buff, we proceeded to the so-called "ridge," with the view of testing the under-current in the methods of which the results had proved so satisfactory at the other end of the strait. By means of our water-bottle and thermometers, we ascer-librium with a column of the same depth tained that the water of the deeper chan- on the Atlantic side. But the upper curnels intersecting the "ridge" was no less rent of saline Atlantic water brings in a distinctly Mediterranean than that of the certain quantity of salt in addition to that under-current near Gibraltar; and since it which the Mediterranean basin previously is obvious that this could not have contin- contained; the density of the Mediterraued to maintain itself against the inflow nean water is consequently augmented; of Atlantic water, unless it had an outward and a column of it reaching to any given movement of its own, the existence of an depth, is rendered heavier than an equivaunder-current over the deeper portion of lent column of Atlantic water. Hence the the ridge cannot be a matter of reasonable excess of downward pressure will displace doubt. We very much regretted that the the lower portion of that column, which uneven nature of the bottom prevented us will flow outwards as an under-current; from here obtaining the more direct demon- and this will not be checked by the inclistration which the use of the "current-nation of the bottom, which the motion drag" had afforded us at the opposite end kept up by a very small constantly-acting of the Strait; but it would have been out excess of pressure will enable it to surof the question to attempt it. For we mount. The withdrawal of a portion of here lost, by their becoming entangled among rocks, not merely a sounding-lead, but also a pair of thermometers and a water-bottle; and a like entanglement having taken place in another sounding, it was only after considerable expenditure of time and trouble that the instruments were recovered. One of the lost thermometers had become quite a "pet" with us; for in the important series of deep temperaturesoundings taken in the Porcupine expedition of 1869 (to which we shall presently the lower stratum will produce a renewed reduction of the surface-level, and this will occasion a further inflow of Atlantic water, which will in turn undergo concentration whilst adding to the quantity of salt in the Mediterranean basin; so that there will be always, on the one hand, a tendency to the restoration of the level lowered by excess of evaporation, and, on the other, a tendency to the restoration of the equilibrium disturbed by excess of pressure; and the inflow and the outflow will so far keep each strong surface-current is always setting outwards through these straits into the Egean; and that the water which this current carries off contains a large quantity of salt. It is just as evident, therefore, in the case of the Black Sea as it is in that of the Baltic, that all the salt would be washed out of its basin within a limited time (a period of three hundred years has been assigned) if it were not for a reflux of salt water by an under-current from the Egean. And as the water of the Black Sea is found to contain a pretty constant proportion of salt-about one-third of that contained in ordinary sea waterthe existence of such an under-current may be safely predicted on the double ground of à priori and à posteriori necessity. other in cheek, that neither the lowering | forcing back a portion of its lower stratum of the level nor the increase of density into the Baltic, thus tending to restore the will ever exceed a very limited amount. equilibrium of the two columns, the disBaltic Current.· The truth of this ra- turbance of which is as continually retionale can be tested in a way that would newed by the influx of fresh water from scarcely be anticipated, viz., by its applica- the surrounding land. bility to the precisely converse case of an There can be no doubt that the same is inland sea, of which the surface-evapora- true of the Black Sea, which is constantly tion, instead of being greater (as in the receiving from the Danube, the Dnieper, Mediterranean), is less than the amount of and the Don a much larger amount of fresh water brought into it by rain and fresh water than it loses by surface evaprivers. Such a condition exists in the oration; so that its level would progresBaltic, which receives the drainage of one-sively rise if a means of escape for the exfifth of Europe, whilst its loss by evapora- cess did not exist in the Bosphorus and tion is comparatively trifling; so that its the Dardanelles. It is well known that a level would be gradually raised, if it were not for the outflow which is continually taking place through the Sound (as well as through the Great and little Belt) into the North Sea. Now, as this outflow carries off a large quantity of salt, whilst all the water which the Baltic receives from rain and rivers is fresh, it is obvious that every particle of salt would in time be washed out from the Baltic basin, if it were not for an under-current which brings back into it the salt water of the North Sea. The existence of such an under-current, at no great depth beneath the surface, was demonstrated two hundred years ago by an experiment resembling that which we made last summer in the Strait of Gibraltar: a boat having been carried along, in opposition to the surface-current, by the action of the General Oceanic Circulation. But the under-current against a basket weighted Physical theory, which has proved itself with shot that was let down into it from adequate to account for these two sets of the boat. This under-current runs so phenomena, has a much wider and more near the surface that, as we are told by important range of application. For it Professor Forchhammer, its action of the leads us to affirm that such a reverse under-current upon the lower part of the movement of the upper and under strata immersed hull of large vessels is some- of water must take place in any case in times strong enough to overcome the con- which there is at the same time a disturbtrary action of the surface-current upon ance of the relative levels and of the relatheir upper part, and to carry them along tive densities of two masses of water in free in its own direction; and he has ascer- communication with each other, provided tained that this under-current has the salt-only that the increase of density be on the ness of the water outside the Sound, whilst side of the reduction of level. We have the surface-current is readily accounted for by the Physical theory already applied to the case of the Gibraltor current; for whilst, on the one hand, the surface-current is always tending to reduce the level of the Baltic to that of the North Sea, the influx of fresh water into the Baltic and the outflow of a portion of its salt are always tending to diminish its density. The weight of a column of Baltic water, therefore, will always be less than that of a column of North Sea water of the same height; while the excess of downward pressure in the former will be continually seen that though the force which immediately acts in producing the double currents in the Strait of Gibraltar and the Baltic Sound is that of gravity, yet that their maintenance in both cases is dependent upon solar heat. For if, on the one hand, the effect of that heat upon the surface of the Mediterranean did not suffice to produce an excess of evaporation, with its consequent lowering of level and increase of density, there would be no Gibraltar currents. Whilst, conversely, if the rivers that discharge themselves into the Baltic basin were less copiously fed by the rains and snows whose source lies in the atmospheric vapours pumped up by solar radiation, there would be neither raising of level nor reduction of density in its contents, and the Sound currents would be suspended permanently as they often are for a time. constant interchange between Polar and Equatorial waters, through the seas of the temperate zone, may be predicted as a physical necessity. For the reduction in the temperature of the Polar column, the whole of which may be brought down, by the continued exposure of the surface to But a constant and decided difference in atmospheric cold almost to its freezingthis action of solar heat upon two bodies of point, must diminish its height whilst augwater communicating with each other, will menting its density; and thus the water have a much more direct influence in main- of the surrounding area must flow in to taining currents between them, by the maintain the level thus lowered. But difference in the temperatures which they when the column has been thus restored will respectively acquire. Let the Medi- to an equality of height, it will possess such terranean basin be supposed to be filled an excess of weight, that its downward with water of the same density as the At-pressure must force out a portion of its lantic, and up to the same level: and to be then cooled down below the freezingpoint of fresh water by the withdrawal of solar radiation, whilst the surface of the Atlantic continues to be heated, as at present, by the almost tropical sunshine of the Gibraltar summer. The cooling of the Mediterranean column, which reduces its bulk without diminishing its weight, would at the same time lower its level and augment its density.* An indraught of Atlantic water would take place to restore that level; but this indraught would augment the weight of the column, giving it an excess above that of the column at the other end of the Strait; and the equilibrium could only be restored by the escape of a portion of the deeper water of the Mediterranean as an undercurrent towards" tide-race") may show themselves in such the Atlantic, thus again reducing the sur- narrowed channels, when no movement is face-level of the Mediterranean. Now so discernible in the open waters from which long as the warm Atlantic water, which they lead; the contraction of the channel comes in to restore that level, is in its turn bringing the moving power, that was presubject to the same cooling, with conse-viously diffused, to bear upon a limited quent lowering of level and increase of density, so long would the vertical pressures of the two columns, which would speedily recover their equilibrium if both were subjected to the same heat or the same cold, remain in a state of inequality; and so long, therefore, must this vertical circulation take place between them. Now the case thus put hypothetically has a real existence. For the Mediterranean cooled down by the withdrawal of solar heat, let us substitute the Polar basin; and for the Atlantic, the Equatorial ocean. The antagonistic conditions of temperature being constantly sustained, a It may be well here to mention that, as was long since ascertained by Despretz, contraction of seawater with reduction of temperature continues down to its freezing point (which is from 5 to 7 Fahr. below that of fresh water), instead of giving place to expansion, as is the case with fresh water at 39 1-29. Sea-water thus acquires its greatest density when it is on the point of freezing. deeper water; and thus an outflow of icecold water will be occasioned from the Polar towards the Equatorial area, over the sea-bed of the deepest oceanic basins, whilst at the same time there will be a continual indraught of warmer surfacewater into the Polar basin, which can only be supplied by a general pole-ward movement of the upper stratum of Equatorial water. These movements will not have the character of currents; for it is only where the communication between the two bodies of water takes place through a narrow strait, that differences so inconsiderable can give rise to a perceptible movement between them. Every one is familiar with the fact that powerful currents (such as that of a "mill-race" or a quantity of water so that its rate is increased in proportion to the reduction in its amount. But the movement is not the less real when diffused, than it is when concentrated; and the same vertical circulation would take place between the two extremities, or between the centre and circumference, of the same continuous basin, under opposite conditions as to heat and cold, as would exist if they were connected by a comparatively narrow channel of communication. That the "vertical circulation" here predicted on theoretical grounds actually takes place in any mass of salt water of which one part is exposed to surface-cold and another to surface-heat, is capable of ready experimental proof. Let a long narrow trough with glass sides be filled withalt water, and let heat be applied at one end (the Equatorial) by means of a thick bar of metal laid along the surface, |