primary than they had been in most portions of its front. By the fourth day the secondary had passed away, and then the typical weather of the rear of a cyclone was experienced.

RAIN WITH STEADY BAROMETER.

So far for rain with a rising barometer; now we must consider precipitation with a steady barometer. To Englishmen this is more perplexing than rain with a rising mercury.

In the latter case, we see at once that there is some disturbance going on; but in the former we often have a steady downpour for several hours, with an absolutely steady barometer. Rain of this class is much more common in continental Europe than in Great Britain, except in one very rare case, which will be mentioned hereafter.

The rain is always either non-isobaric, or of that kind which is associated with secondaries and not with primary cyclones. For this reason, the rain is never accompanied by a gale of wind, though there are often angry gusts at the beginning and end of the rainfall.

In Fig. 90 we give a photographic engraving of the author's barographic trace in London, on July 1 and 2, 1877. This, being absolutely untouched by hand, gives the minute irregularities of pressure in a manner which no hand-copied diagram can ever do. The horizontal lines represent differences of half an inch of pressure, the thickest one marking the level of 29.5 inches. The horizontal lines are drawn at six-hour intervals.

At first sight, there might seem to be little sign of any disturbance, for the actual changes of barometric level are insignificant, and the diurnal variation is more obvious than usual in Great Britain. If, however, we look carefully at the trace, we shall find that just before 6 a.m. on July 1 there is a very small dip of the barometer, and that then the trace is almost quite straight till about 4.30 p.m., when there is another small dip; after which the regular diurnal variation is absolutely undisturbed.

In London rain commenced at the first dip, and continued without intermission till the second, after which the sky cleared.

The charts for that day, which unfortunately the number of illustrations at our disposal does not admit of reproducing, show that this was all caused by the formation and passage of a small secondary over the north of France and the English Channel; and both

the rain and the barographic trace are most characteristic of this class of depression. A case of this sort shows,

more than any other, the superior value of a continuous trace over an intermittent barograph; for, though the latter permits of the tabulation of hourly values for the determination of diurnal variations, they entirely lose all chance of following the more minute alterations of pressure, which are often accompanied by great changes of weather. The most interesting point about secondaries is the contrast between the intensity of the weather which they induce and the apparently small disturbance of pressure. In primary cyclones the gradients are to a certain extent a good measure of the intensity. In secondaries, on the contrary, the rainfall has no relation whatever to the barometric disturbance. This, of course, makes it very difficult for the forecaster. All he can say when he sees a secondary is-rain; but he can give no estimate of the quantity of precipitation, as he can of the force of the wind in a primary cyclone.

Rarely in Great Britain, frequently in continental Europe, habitually in the tropics, we have purely nonisobaric rains, totally unconnected with any secondary. These are often indicated on the barographic trace by a sudden sharp rise of the type we illustrated in our chapter on Thunderstorms. This is probably a purely local effect of a heavy downpour pressing the air down by its own weight.

The other case of rain-this kind often with a gale of wind-with an apparently steady barometer, only occurs in very unsettled weather. In our chapter on WeatherTypes, we gave several examples of nearly stationary cyclones, which increased much in depth, while some of the adjacent anticyclones increased in height. As a

necessary consequence, there must be some station where no change of pressure would be observed; but on one side pressure would decrease, while it increased on the other; so by this means very steep gradients might come to lie over the station. The wind would rise to a gale, while the weather would conform to the shape of the isobars, but the mercury would remain stationary; we might, in fact, say that the station was "nodal" as regards the fluctuations of surrounding pressure.

It would be an extreme case when no change of pressure took place, and could only happen at a limited. number of places. But under the same conditions there will always be a number of stations where only a moderate fall of the barometer takes place, but a gale out of all proportion to the apparent depression is experienced. This illustrates the important difference between the fall of the barometer due to the passage of a well-defined cyclone, and that due to the rearrangement of the distribution of pressure round the station. As an example, we may turn to Fig. 93 in the next chapter, where we give two charts of North-Western Europe, on February 6, 1883, at 8 a.m. and 6 p.m. respectively. The position of the isobar of 304 ins. is practically the same in both maps; but between the morning and evening observations, pressure has fallen 04 inch in the west of Ireland, and risen 0.2 inch over Sweden. The shape of the isobars has not altered much, so that gradients have become steep, with little change of wind-direction. Thus many stations, near the nodal isobar, will experience an increase of wind with either a rising, stationary, or slightly falling barometer. For instance, at Aberdeen, marked A, the wind-arrow

shows that the wind had risen from a fresh breeze to a moderate gale; while the motion of the isobars does not indicate a fall of more than 0.1 inch in the ten hours which elapsed between the two sets of observations.

FINE WEATHER WITH LOW OR FALLING BAROMETER.

From the above, in which the weather is out of all proportion to the depression of the mercury, we readily pass to the converse case, in which the fall of the barometer is quite disproportioned to the severity of the weather which is afterwards experienced. In the North of Europe, during the winter months, and when the westerly type of weather prevails, the barometer will sometimes fall half an inch or more, and often below 28.5 ins., while no strong winds follow, and the general appearance of the sky is bright, with perhaps a little cumulus cloud. This also is readily explained by reference to our large Atlantic charts. In them we saw that when the Atlantic is covered by a persistent area of low pressure, the depth of the lowest point often suddenly decreases nearly an inch, and that the gradients near the centre are very slight. In some phases of that type of weather, the area of low pressure stretches over Europe, and the minimum of this area rises up and down exactly as when the centre lies over the ocean.

If, then, Great Britain, for instance, lay within that area, pressure might decrease a whole inch, and neither storm nor rain be experienced. The great fall of pressure would, of course, develop steep gradients, somewhere to the west of those islands; but as the depression was not