I reckoned on being back at Buenos Aires by the beginning of September, and on getting, if possible, one station a day.  The distance, according to a rough calculation, was about 8,000 nautical miles, and I laid down the following plan:  To go about east by north with the prevailing northerly and north-westerly winds to the coast of Africa, and there get hold of the south-east trade.  If we could not reach Africa before that date, then to turn on July 22 and lay our course with the south-east trade for St. Helena, which we could reach before August 1; from there again with the same wind to South Trinidad (August 11 or 12); on again with easterly and north-easterly winds on a south-westerly course until about August 22, when the observations were to be concluded, and we should try to make Buenos Aires in the shortest time.

That was the plan that we attempted.  On account of the fresh water from the River La Plata, we did not begin at once to take samples of water, and with a head-wind, north-east, we lay close-hauled for some days.  We also had a pretty stiff breeze, which was another reason for delaying the soundings until the 17th.

For taking samples of water a winch is used, with a sounding-line of, let us say, 5,000 metres (2,734 fathoms), on which are hung one or more tubes for catching water; we used three at once to save time.  Now, supposing water and temperatures are to be taken at depths of 300, 400, and 500 metres (164, 218, and 273 fathoms), Apparatus III. (see diagram) is first hung on, about 20 metres (10 fathoms) from the end of the line, where a small weight (a) hangs; then it is lowered until the indicator-wheel, over which the line passes, shows 100 metres (54 fathoms); Apparatus II. is then put on, and it is lowered again for another 100 metres, when Apparatus I. is put on and the line paid out for 300 metres (164 fathoms) —­ that is, until the indicator-wheel shows 500 metres (273 fathoms).  The upper Apparatus (I.) is then at 300 metres (164 fathoms), No.  II. at 400 metres (218 fathoms), and No.  III. at 500 metres (273 fathoms).  Under Apparatus I. and II. is hung a slipping sinker (about 8 centimetres, or 3 1/4 inches, long, and 3 centimetres, or 1 1/4 inches, in diameter).  To the water-samplers are attached thermometers (b) in tubes arranged for the purpose.

The water-samplers themselves consist of a brass cylinder (c), about 38 centimetres (15 inches) long and 4 centimetres (1 1/2 inches) in diameter (about half a litre of water), set in a frame (d).  At about the middle of the cylinder are pivots, which rest in bearings on the frame, so that the cylinder can be swung 180 degrees (straight up and down).

The cylinder, while being lowered in an inverted position, is open at both ends, so that the water can pass through.  But at its upper and lower ends are valves, working on hinges and provided with packing.  When the apparatus is released, the cylinder swings round, and these valves then automatically close the ends of the cylinder.  The water that is thus caught in the cylinder at the required depth remains in it while it is being heaved up, and is collected in bottles.  When the apparatus is released, the column of mercury in the thermometer is broken, and the temperature of the water is read at the same depth as the water is taken from.