of the atmospheric pressure of the globe, owing to the comparatively large number of places from which barometric observations have been obtained, and to the circumstance that fewer places are required to represent the pressure over any portion of the earth's surface than are required to show the temperature, the winds, or the rainfall.

Since the information conveyed by isobarometric charts forms the basis of Meteorology, it has been necessary, in preparing this edition for the press, to recast the greater part of the work. The chapters on "The Distribution of Atmospheric Pressure over the Globe," and "The Relation of Atmospheric Pressure to the Temperature," are wholly new, and the Chapter on "Winds" is almost altogether new. The intimate connection of temperature with pressure is illustrated by the exceptional weather of last year (1867), the chief instances adduced being the mild weather of February; the unprecedentedly cold weather of March, which proved so injurious to vegetation; the cold weather of July, so disastrous in its effects on the crops; and the singularly fine weather of November. The great frost of Christmas 1860, the warm weather of September 1866, and the cold weather in the south of Europe in January 1868, are considered in the same relations. Attention is thus directed to the proximate causes of exceptional weather, and remarks are made with the view of suggesting how this knowledge may be turned to practical account. The importance of Iceland as the key to the climates of the greater part of North America and Europe is pointed out.

The subject of STORMS is more specially discussed, as being, from the practical application of the results of the inquiry to the prediction of storms, and consequently to the saving of life and property, a question the im

portance of which it would be difficult to overestimate. As the subject has been considered from the observational side, it follows that the law of storms, so far as enunciated, is to be regarded not as mere theory, but as the result of generalisations of observed facts. In the few cases where theoretical opinions are advanced, their theoretical nature is distinctly stated. The proximate cause of high winds is illustrated by the Great Hurricane which was so severely felt at Edinburgh on the 24th January 1868. The observations made by observers of the Scottish Meteorological Society at the time, supplied the materials for placing the relation of wind-force to differences of pressure in a very clear light. A new chapter has been added on " Weather and Storm-Warnings."

Other features of storms yet remaining to be investigated, and many undetermined points, in reference particularly to the diathermancy and distribution of the aqueous vapour of the atmosphere, solar and terrestrial radiation, the temperature of the soil, and ocean meteorology, are stated, in order that the attention of meteorologists may be directed to the most fruitful fields of research which the science presents.

Meteorology has been too long treated with neglect, both in popular and in liberal systems of education. In the popular education of this country it cannot be said to hold a place; and even in a University course, though classed among the subjects embraced by Natural Philosophy, its position in the curriculum of studies is little more than nominal. The causes of this neglect are not far to seek :-the chief being the absurd pretensions of weather-prophets and other prognosticators; the free and bad use long made of electricity and other imperfectly understood agents to explain the causes of atmospheric disturbances; the fewness of the observed

facts in comparison with the vastness of the weatherchanges they were adduced to explain; but above all, the real difficulty of the subject, owing to the manifold influences in operation. But now that Meteorology has discarded all pretensions and theories, except in so far as the latter are the legitimate result of observation, it has acquired rapid development, and established its claim to be regarded as the youngest of the sciences.

In consequence of this, a great and growing interest on the subject is spreading among all classes, as is evinced by the number of meteorological instruments purchased, by the many zealous and self-denying observers in all parts of the civilised world, and by the still more numerous class who volunteer their services as observers. But since a knowledge of the elementary principles of Meteorology is not yet very generally diffused, much time and labour are often wasted from the use of imperfect instruments and imperfect methods of observation. To meet this want, the author has particularly described the various instruments in use, given directions for placing them in situations and in positions best suited for observation, pointed out how they are liable to get out of order, and explained the methods by which they may be put right when they happen to go wrong. The methods of reducing Meteorological Observations are explained at length, and tables are given for facilitating the work.

DIRECTIONS TO METEOROLOGICAL OBSERVERS IN REGARD TO THE PUBLICATION OF THEIR STATISTICS.-The height of barometers above the level of the sea, accurately ascertained, is a serious desideratum in not a few published abstracts of Meteorological Observations. The labour which, in preparing this edition, the omission, or

rough and incorrect statements, of heights has occasioned, has been excessive. In many cases it was necessary to ascertain the heights, or check those given, by the method of the barometric measurement of heights-viz., by selecting a number of days during which the barometer remained steady over the region in which the place was situated, and then calculating the height from the pressure and temperature at the place as compared with places in the vicinity whose heights were known. When the place was isolated from other stations the annoyance was still greater. Thus, since the height of Santiago de Chile, usually given at 2600 feet, is evidently 500 or 600 feet too high, the observations at this important station had reluctantly to be omitted in the construction of the Isobarometric Charts. Meteorological societies and observers should lose no time in rectifying this flagrant omission.

In publishing monthly or other Abstracts of Barometric Observations, the following information ought to be given:-1. The hour or hours of observation; 2. The height of the cistern of the barometer above the sea; 3. The latitude and longitude; 4. The barometer reduced to 32° F., and corrected for instrumental errors, but not corrected for daily range; 5. The mean temperature of the air. A column of the observations reduced to sealevel may be added; but in such cases the reduction to 32° only should also be given. If these simple directions were observed, much labour and annoyance would be saved, and the value of the observations would be correspondingly enhanced. What meteorologists ask for, are simply the results of actual observations. It would be difficult to name any abstracts which excel, in these respects, those issued by P. Secchi.

Since storms and other important questions of Mete

orology can only be discussed by Synchronous Charts which represent the weather at specified hours, it is indispensable, in publishing daily observations, to state the hour or hours at which they were made. When this is not done, much confusion and unsatisfactoriness is the result. Sometimes daily barometric means alone are published; that is, if three observations are made a-day, none of these three observations, but a mean deduced from them, is published,—a figure which in discussing storms is in all cases of little use, and in cases when the storm is passing the place, absolutely of no use; and it cannot be used in determining the amplitude of the barometer fluctuations accompanying storms and other weather-changes. By this mode of publishing daily observations, a considerable proportion of the observations published in the Annales de l'Observatoire Physique Central de Russie,' the greatest storehouse of meteorological facts in the world, are unfortunately deprived of a great part of their value. Indeed, if instead of daily barometric means, one daily observation at a specified hour had been given at the stations, it might have been possible to prepare synchronous charts of the greater part of the northern hemisphere. The observations published by the Meteorological Institute of Norway may be referred to as models of daily Meteorological Records. Uniformity in the methods of observation practised by Meteorologists is also most desirable. A SECOND BRUSSELS CONFERENCE is perhaps the only means by which the advantages of uniformity of observation and of publication of results may be secured.

June 1868.