Iowa City, together with the maps which he gives to illustrate them, point very clearly to line-squalls, associated with that class of V-depression in which the rain follows the trough. His maps do not exhibit the shape of the rain-area like the one we have just given, but they show the squalls sweeping across the State with a crescent-shaped front exactly like the Eurydice squall.

When we compare this class of squall with the pure and simple squall which we first described, it will be obvious that the two kinds have little in common except the name. The former kind seems to be simply a local intensification of a general sweep of wind. The latter, on the contrary, is associated with a very definite but complex phase of aerial circulation, which we shall understand better when we have described a precisely analogous class of thunderstorms.

THUNDERSTORMS ASSOCIATED WITH LINE-SQUALLS.

Squalls are rarely of sufficient importance to attract the notice of enough observers to enable the details of their shape and progress to be properly determined; but thunderstorms are such a striking manifestation of weather that they are much more easily traced, and an enormous amount of work has been done in late years in marking out the hourly advances and development of such disturbances.

Many, but not all, European thunderstorms have been found to be precisely similar to the line-squall which we have just described. Some of Bezold's diagrams of

Bavarian thunderstorms, which give the shape of the area covered by the storm at successive hours, show long narrow bands sweeping broadside on across the country exactly analogous to the squall-area which we drew in Fig. 50.

But the remarkable point is, that though some of these storm-bands are associated with the troughs of cyclones and V's, precisely similar bands are more often found either in front or in rear of the cyclone, where we can connect them (the bands) with no particular part of the cyclone, except that the front of the band is usually perpendicular to the line of progress of the depression in which it is formed.

We will first give an example of the storm and thunderstorm associated with the trough of a V-depression. On July 16, 1884, at about 6.15 p.m., the trough of a V-depression swept over Hamburg, and brought, as usual, a violent squall and heavy rain, with much thunder and lightning. This was only a section, as it were, of a line-thunderstorm. Dr. Sprung, by combining the section of barometric and other curves at Hamburg with the records of other observatories and the synoptic charts of Germany, at 8 p.m. the same evening, in the manner that we explained in our chapter on Meteograms, has built up the beautiful diagram of this storm, which we reproduce, with a few trifling modifications, so as to assimilate it with our other illustrations, in Fig. 51. There the isobars are given both in millimetres and their approximate equivalents in inches: t marks the line of the trough; the shaded liner the position and dimensions of the rainstripe; the long dotted arrow the direction in which the

755

757-5

760

29.7

29.8

29.9

whole system was being propagated; and the small solid arrows the direction and force of the wind across any section. The whole is evidently, in the main, one of those V's in which the rain begins just after the passage of the trough; but the curious projection upwards of the isobars just under the rain-stripe is unlike anything we have seen before; and the strong west wind, with four feathers on the arrow, is quite unconformable with the isobars according to our usual experience. From whence comes all this? First, for the upward projection of the isobars. All the barographs showed, about seven minutes after the passage of the trough, a sudden rise exactly similar to those marked b in Fig. 48 just as the heavy rain began, and this rise, as usual, was quite distinct from the general increase of pressure due to the rear of the V. If we were to superimpose a long, narrow, isolated ridge of high-pressure on the rear of a V, we should get an inverted V, or wedge, exactly like that which we find in the diagram under the rain-stripe; but we must not treat this like the ordinary wedgeshaped isobars which we have before described. The form is the same, but the cause is different.

FIG. 51.-Line-thunderstorm. t, Trough of V-depression; r, rain stripe.

Then for the wind-sequence. We find in front of the V a light south-east wind; then, just before and at the

commencement of the rain, a very violent squall (called "boe" in Germany) from the west; and, finally, a light south-west wind in rear of the whole disturbance. The westerly squall is nearly perpendicular to the isobars and to the lie of the rain-stripe; but, as the isobars here are not the lines of general atmospheric circulation, but are partly due to purely local causes, we need not be surprised that Buy Ballot's law does not hold.

Temperature was, as usual, very high in front; very low for the season in rear of the trough. If we could have drawn the isotherms, we should have found them running almost north and south, nearly parallel to the rain-stripe. It is impossible to determine at present how much of this cold is due to the mechanical transport of cold air with the heavy rain, and how much to the general descent of cold air in rear of the V as a whole.

The rain-stripe, we see, was in this instance a long narrow band, moving broadside on, and manifestly connected with the trough of a V-depression.

We will now give an example of line-thunderstorms which are not associated with the trough of either a V or a cyclone, though they also move broadside on, nearly perpendicular to the depression with which they are in some way associated. In Fig. 52 we give synoptic charts for France at 9 a.m. and 9 p.m. (Paris time) on August 21, 1879. The full lines are isobars, a few arrows show the general direction of the wind, and the one dotted line in each marks the mean position of the thunderstorms which were raging at that moment. In France the mean of the time between the first and last thunder is taken to give the position of the storm at a given hour. For instance,

if the first thunder was heard in a place at 8 a.m. and the last at 10 a.m., then 9 a.m. would be marked as the hour at which the storm passed the station. This method is manifestly inferior to that of noting the times of first and last thunder, and then plotting the shape of the storm on a chart. The barometric changes during the day were

really much more complicated than might appear from an inspection of the maps over the limited area of France. The large secondary cyclone whose centre lay over the west of France in the morning, appears to have crossed that country in a north-easterly direction, and to have merged in a complicated manner with a larger depression which lay to the north-west of Ireland in the early morning. By 9 p.m., however, the whole of the south-west of France was covered by an anticyclone, whose origin we

cannot trace.

In Fig. 53 we give a diagram of the positions at every alternate hour of two sets of thunderstorms that traversed