obtain decisive results as to secular periodic weather-changes, even though these observations were to be carried on for a hundred years or more. The terrestrial atmospheric currents, in their passage over the same portion of the earth's surface (whether this passage be periodic or not periodic), never take the same route, or have the same limits, and consequently the localities over which they pass on their return will happen to be at one time in the centre of the current, at another time more or less near to its northern or its southern edge.
Under these circumstances the state of barometer and of thermometer, the amount of precipitation, the force of the wind, etc., upon which these researches are based, must yield conflicting results. Some, however, have supposed that-better results might be obtained by combining these quantities, and taking the mean of all the readings. This method quite does away with anomalies, it is true, but then the result has no specific value whatever, though the aim of all such researches must always be to determine the weather specifically. The periodicity of weather phenomena can only be determined by means of investigations carried on according to the geographical method, i. e., by studying these atmospheric phenomena in their continuity both as regards time and space.
As has been already said, the track of the storm of November l7th-23d layover nearly the same regions of the northern hemisphere in 1873 and 1854. Had these air-currents taken a course only a few degrees more to the north or to the south, their existence and their identity would have been so ill determined by the barometer of a single locality that they might easily have been overlooked.
On the 28th of November, 1873, at 10 p. m., the barometer at Emden stood at 756.4 millimetres, and then kept on falling till the 30th, at 8 a. m., when it was 744.2 millimetres; after this it rose till at 10 p. m. it was 762.6 millimetres. This not very considerable fall of the barometer would not have deserved special notice, were it not followed on the evening of the 29th by a storm of some violence, which through the night became a hurricane. In 1854, at the same period of the year, the barometer underwent a similar change, only much greater. On November 27th, at 10 p. m., the height of the barometric column was 757 millimetres. It then fell, till on the 29th, at 2 p. m., it was 727.8 millimetres, when it began rising till it reached 757.9 millimetres at 8 a. m., on the 30th. The quantitative difference of barometric pressure at Emden on November 27th and 28th, between the years 1873 and 1854, is simply the result of the difference in the tracks of the storms, and the consequent distance of the place of observation from the storm-centre. At Thorshavn, between November 25th and 28th, in 1873, the barometric changes were precisely the same as had been observed at Emden on the same days in 1854. In 1873, at 9 p. m. of the 25th, the mercurial column at Thorshavn was 757.3 millimetres; on the 26th, at 8 a. m., it was 746.9 millimetres; on the 26th at 9 p. m.,