incessant, they are not entirely without salt, may be taken as evidence in proof of a system of currents or of circulation in the sea, by which its waters are shaken up and kept mixed together as though they were in a phial.
407. Currents of the Atlantic.—The principal currents of the Atlantic have been described in the chapter on the Gulf Stream. Besides this, its eddies and its offsets are the equatorial current (Plate VI.), and the St. Roque or Brazil Current. Their fountain-head is the same: it is in the warm waters about the equator, between Africa and America. The former, receiving the Amazon and the Orinoco as tributaries by the way, flows into the Caribbean Sea, and becomes, with the waters (§ 103) in which the vapours of the trade-winds leave their salts, the feeder of the Gulf Stream. The Brazil current, coming from the same fountain, is supposed to be divided by Cape St. Roque, one branch going to the south under this name (Plate IX.), the other to the westward. This last has been a great bugbear to navigators, principally on account of the difficulties which a few dull vessels falling to leeward of St. Roque have found in beating up against it. It was said to have caused the loss of some English transports in the last century, which fell to leeward of the Cajoe on a voyage to the other hemisphere; and navigators, accordingly, were advised to shun it as a danger.
408. The St. Roque current.—This current has been an object of special investigation during my researches connected with the Wind and Current Charts, and the result has satisfied me that as a rule it is neither a dangerous nor a constant current, notwithstanding older writers. Horsburgh, in his East India Directory, cautions navigators against it; and Keith Johnston, in his great Physical Atlas, published in 1848, thus speaks of it: "This current greatly impedes the progress of those vessels which cross the equator west of 23° west longitude, impelling them beyond Cape St, Roque, when they are drawn towards the northern coast of Brazil, and cannot regain their course till after weeks or months of delay and exertion." So far from this being the case, my researches abundantly prove that vessels which cross the equator five hundred miles to the west of longitude 23° have no difficulty on account of this current in clearing that cape. I receive almost daily the abstract logs of vessels that cross the equator west of long. 30°, and in three days from that crossing they are generally clear of that cape. A few of them report the
