are not at all agreed, and the observations made on its phenomena (made at different stations) do not accord in a satisfactory manner. In fact, it appears as if the indications of the instruments are due to local causes, so that they* do not lend themselves to any useful generalizations. When a thunder-storm is actually raging in the neighborhood of a station, the indications of electrometers thereat are most erratic and violent, but it can not be said that any electrometer enables us to perceive the approach of a storm one whit earlier than we are able to do by careful watching of the clouds. As regards forecasting thunder-storms, this can be done in a general sort of way; but it is not practicable to predict which villages or parishes, or even counties, will be visited. When the daily weather charts are drawn, if we find that there is an unevenness in the isobaric lines—that is, if these are wavy, or bulge out irregularly—we know that thunder-storms are likely to burst somewhere or other over the country, but that is all we can say. At each station the barometer is unsteady—the mercury moving up and down in the tube—during the actual continuance of the storm; but this oscillation of the mercurial column has nothing to do with the irregularity in the isobaric lines above mentioned. Forecasting these storms is, therefore, always an uncertain and a thankless task, for local success is rarely attained.
Among the earliest symptoms of the approach of a thunderstorm is the appearance on the western horizon of a line of cumulus ("wool-pack") clouds, exhibiting a peculiar turreted structure. I say on the western horizon, for most of our changes of weather come from that quarter, and it has been proved that thunderstorms, like wind-storms, advance over the country, generally, from some westerly point. This bank of clouds moves on, and over it appear first streamers and then sheets of lighter upper cloud—cirrus, or "mare's-tail"—which spread over the sky with extreme rapidity. The heavy cloud mass comes up under this film, and it is a general observation that no electrical explosion or downfall of rain ever takes place from a cloud unless streamers of cirrus, emanating from its upper surface, are visible when the cloud is looked at sideways from a distance.
Thunder-storms are generally accompanied by falls of hail as well as rain, and these hailstones assume the form of lumps of ice—some even as large as hens' eggs, and weighing several ounces, having been known to fall. The stories of masses of hailstones, weighing many pounds, having been found after storms, are explained by the fact that the hailstones, after they have fallen, may have frozen to each other and formed a solid lump on the ground. Large hailstones are composed of alternate layers of clear crystalline and white porous ice, and many of them consist of an aggregate of smaller hailstones which have attached them-
