statistics regarding the actual occurrence of tornadoes, we find the certainty lessened by the fact that the present limited resources of the Signal Service result in defective reports or in none at all from sparsely settled regions. Lieutenant Finley found that of thirty-eight predictions that tornadoes would occur, made in April and June, 1884, eighteen wore verified, and that of nineteen predictions made in June and July, 1885, fifteen were generally verified. In all cases there were violent storms, either tornadoes, hurricanes, or hail. Owing to the extremely local nature of tornadoes, their tracks at times being only a mile or two in length and a few hundred feet in width, it is obvious that many predictions must apparently fail, owing to the fact that the effects are not seen until long afterward, or not at all where there are vast stretches of treeless prairie. It is doubtless true that this failure, due to the vagueness and unsatisfactory nature of the reports, induced Professor T. B. Maury to maintain, as late as 1882, that the prediction of a tornado was a triumph not yet attained by the science of meteorology, though doubtless he believed that success would be achieved at no very distant day. In order that the reader may see some of the reasons for expected progress in this science, let us examine, first, the methods in use by Lieutenant Finley for tracing the movement of air masses, and second, the movement of the air-currents in the tornado cloud, as seen by hundreds of observers.
It is well known that, owing to frequent telegraphic reports, the pressure, temperature, cloud-formation, extent, and movement of immense masses of air are permanently recorded. The conditions favorable to tornadoes are positive and noticeable. The areas of warm southerly and cold northerly winds are well defined, uniform, of large extent, and reach well to the north and south. High contrasts of humidity, abnormal variations in dew-point, the location of areas of barometric minima and maxima, with their lines of actual and probable progressive movement, and especially the velocity and direction of the wind, must be considered and mapped out on special charts. The temperatures are thrown out of their usual equilibrium and normal distribution over an extent at times of two thousand miles of territory. The cold air encroaches far into the Southern States, and the warm air of the South at such times may stream northward during a week or ten days. The movement in readjusting the equilibrium is like two pendulums thrown far apart which swing toward their common center with a force proportioned to the extent of their displacement. But this simple simile only fits the case roughly, because the questions of wind-direction, the location of the moving center of low pressure, and especially the inequality of the displacement of the air-masses north and south, make the problem very complex. Lieutenant Finley says that "the departure from normal conditions of temperature in case of tornado development is from 15° to 50°, but with this abnormal condition of temperature there must be abnormal conditions of humidity,
