well as by men, particularly by a man in a passion, or on a low scale of civilisation. But there is no language, even among the lowest savages, in which the vast majority of words is not rational. If, therefore, Mr. Darwin (p. 35) says that there are savages who have no abstract terms in their language, he has evidently overlooked the real difference between rational and emotional language. We do not mean by rational language, a language possessing such abstract terms as whiteness, goodness, to have or to be; but any language in which even the most concrete of words are founded on general concepts, and derived from roots expressive of general ideas.

There is in every language a certain layer of words which may be called purely emotional. It is smaller or larger according to the genius and history of each nation, but it is never quite concealed by the later strata of rational speech. Most interjections, many imitative words, belong to this class. They are perfectly clear in their character and origin, and it could never be maintained that they rest on general concepts. But if we deduct that inorganic stratum, all the rest of language, whether among ourselves or among the lowest barbarians, can be traced back to roots, and every one of these roots is the sign of a general concept. This is the most important discovery of the Science of Language.

Take any word you like, trace it back historically to its most primitive form, and you will find that besides the derivative elements, which can easily be separated, it contains a predicative root, and that in this predicative root rests the connotative power of the word. Why is a stable called a stable? Because it stands. Why is a saddle called a saddle? Because you sit in it. Why is a road called a road? Because we ride on it. Why is heaven called heaven? Because it is heaved on high. In this manner every word, not excluding the commonest terms that must occur in every language, the names for father, mother, brother, sister, hand and foot, &c., have been traced back historically to definite roots, and every one of these roots expresses a general concept. Unless, therefore, Mr. Darwin is prepared to maintain that there are languages which have no names for father and mother, for heaven and earth, or only such words for those

objects as cannot be derived from predicative roots, his statement that there are languages without abstract terms falls to the ground. Every root is an abstract term, and these roots, in their historical reality, mark a period in the history of the human mind-they mark the beginning of rational speech.

What I wish to put before you as clearly as possible is this, that roots such as đã, to give, sthā, to stand, gā, to sing, the ancestors of an unnumbered progeny, differ from interjectional or imitative sounds in exactly the same manner as general concepts differ from single impressions. Those, therefore, who still think with Hume that general ideas are the same thing as single impressions, only fainter, and who look upon this fainting away of single impressions into general ideas as something that requires no explanation, but can be disposed of by a metaphor, would probably take the same view with regard to the changes of cries and shrieks into roots. Those, on the contrary, who hold that general concepts, even in their lowest form, do not spring spontaneously from a tabula rasa, but recognize the admission of a co-operating Self, would look upon the roots of language as irrefragable proof of the presence of human workmanship in the very elements of language, as the earliest manifestation of human intellect, of which no trace has ever been discovered in the animal world.

It will be seen from these remarks that the controversy which has been carried on for more than two thousand years between those who ascribe to language an onomatopic origin, and those who derive language from roots, has a much deeper significance than a mere question of scholarship. If the words of our language could be derived straight from imitative or interjectional sounds, such as bow wow or pooh pooh, then I should say that Hume was right against Kant, and that Mr. Darwin was right in representing the change of animal into human language as a mere question of time. If, on the contrary, it is a fact which no scholar would venture to deny, that, after deducting the purely onomatopoeic portion of the dictionary, the real bulk of our language is derived from roots, definite in their form and general in their meaning, then that period in the history of language which gave rise to these roots, and which I call the

Radical Period, forms the frontier-be it broad or narrow-between man and beast. That period may have been of slow growth, or it may have been an instantaneous evolution: we do not know. Like the beginnings of all things, the first be ginnings of language and reason transcend the powers of the human understanding, nay, the limits of human imagination. But after the first step has been made, after the human mind, instead of being simply distracted by the impressions of the senses, has performed the first act of abstraction, were it only by making one and one to be two, everything else in the growth of language becomes as intelligible

as the growth of the intellect; nay, more SO. We still possess, we still use, the same materials of language which were first fixed and fashioned by the rational ancestors of our race. These roots, which are in reality our oldest title-deeds as rational 'beings, still supply the living sap of the millions of words scattered over the globe, while no trace of them, or anything corresponding to them, has ever been discovered even amongst the most advanced of catarrhine apes.

The problem that remains to be solved in our last Lecture is the origin of those roots. —Fraser's Magazine.

THE prediction of the seasons, for any considerable period in advance, is of course

* See The Times of April 15th.

A Lecture delivered at the Royal Institution, Feb. 14.

the problem whose solution must affect the most important social interests, inasmuch as all the operations of agriculture are necessarily dependent on the varying character of the weather. Recently, in order to afford some practical information as to the:

for five-day periods we find the same principle to be true, for if a cold five-day period sets in after warm weather the odds are two to one that the next such period will be cold too; but if the cold has lasted for two months, the odds are nearly eight to one that the first five days of the next month will be cold too. The probability of change is, however, greater for the fiveday periods than for single days.

Similar results follow for the months, but here again the chance of change shows an increase. If we revert to the instance we first cited, that of rain, the result is not that if it once begins to rain, the chances are in favor of its never ceasing; all that is implied is that the chances are against its ceasing on a definite day, and that they increase with the length of time the rain has lasted. The problem is similar to that of human life: the chance of a baby one year old living another year is less than that of a man of thirty living till thirty-one.

The practical meaning of all this is, that although we know that a compensating anomaly for all extraordinary weather exists somewhere on the earth's surface, e. g. the very common case of intense cold in America, while we have a mild winter, which was most strikingly true last January (especially on the 22nd, the day of the snowstorm in Minnesota), there is no reason as yet ascertained to anticipate that this compensation will occur at any given place in the course of a year. In other words, when definite conditions of weather have thoroughly established themselves, it is only with great difficulty that the courses of the atmospheric currents are changed.

Attempts have not unfrequently been made to predict the seasons for a long period in advance, but without much success hitherto. One great cause for failure is that accurate meteorological records do not extend beyond the beginning of the century at more than a few stations, and for these we cannot eliminate the local influences altogether. Thus it is hardly possible to say what has been the approximate temperature of these islands for more than twenty years, a period far too short for the definite recognition of a cycle. The shortest of these cosmical cycles which has been determined is the sun-spot period, of II years, according to Wolf, and there are indications of far longer periods, such as 33 years, or even 693 years, according to Hornstein.

At the last meeting of the British Association at Brighton, a very interesting paper by Mr. Meldrum, of the Mauritius, was read, in which he showed that the cyclones, for which that district of the Indian Ocean enjoys an unenviable notoriety, have been more frequent in some years than in others, and that these years of maximum frequency occurred at intervals of about eleven years, coinciding with those of maximum sun-spot frequency. This agreement is most important, and it has been abundantly corroborated by an examination of the rainfall at such stations in the southern hemisphere as are available. This has been carried out by Mr. Lockyer in Nature, as well as by Mr. Meldrum. The results, for the comparatively short period to which they refer, are most striking, as they are sufficient to show that a periodicity is traceable in the weather of the Indian Ocean, which is eminently suggestive of a close relationship between the changes which take place on the sun's surface, and the phenomena of our own atmosphere. It will at once be asked, why has not this periodicity, if it exists, been detected long ago by an examination of European records, which are far more complete than any existing for the Indian Ocean? The answer to this is twofold: in the first place we are pre-eminently in the district of the variable winds, and our storms are not nearly so regular in their character as those of the Mauritius, where almost the sole type of storm is the true tropical cyclone, with its concomitant rainfall. It is next to impossible in this country to keep an accurate record of all the several storms which pass over us. The existence of congregate storms is not unfrequent; two, and even three, systems of disturbance being traceable at the same time within the area of the United Kingdom. Are these one single storm or several, and how should they be counted in a catalogue? Rain also cannot be taken as a sign of the frequency of storms in a year; for, although we know that warm winters are invariably wet and stormy, it cannot be asserted that the all but constant rainfall of last summer was in any way related to storms. There is, however, in the second place, a far deeper reason for the non-discovery of these cycles in any chance series of rainfall records. The sun passes through phases of greater and less activity, and the terrestrial phenomena corresponding to the epochs

of the former character are excessive evaporation in some parts of the globe, and consequent excessive precipitation in others. We must therefore ascertain in what districts we are to look for the one and for the other of these phenomena respectively. The fact of the mutual compensation of anomalies which was mentioned when treating of M. Köppen's paper will show that it is not impossible that the years of maximum rainfall at the Mauritius may be years of great dryness here. In fact, we cannot yet say where we shall find the maximum solar effect produced.

The lesson that we are to learn from the fruitful researches of Mr. Meldrum has been pointed out by Mr. Lockyer, we must aim at attaining a thorough knowledge of the movements and changes of our own atmosphere, and then seek to establish a connection between them and other cosmical phenomena, such as terrestrial magnetism, the relation of which to the state of the sun's surface was pointed out by Sir E. Sabine twenty years ago.

We must now proceed to the subject of Weather Telegraphy and Storm Warnings, with which the name of Admiral FitzRoy will always be associated. Justice, however, compels us admit that this country was not the first to issue telegraphic weather intelligence to its seaports; for in the year 1860, when the possibility of introducing such a system was being discussed here in London, the step had actually been taken in Holland, at the instance of Professor Buys Ballot. At the present time there is not a single European country, except Greece, which has not its own meteorological organization. In most cases telegraphic weather reports are published in the newspapers, while the example set by Le Verrier about 1858, of the publication of a lithographed daily bulletin, has been followed by our own office, six hundred copies of whose charts are issued daily to subscribers and for exhibition at seaports; of late years Russia too has commenced the publication of a lithographed bulletin.

If we want to see weather telegraphy on its grandest scale we must cross the Atlantic, where, under the direction of Brigadier-General Myer, no less than three charts are issued every day by the chief signal office of the United States at Washington. This undertaking is rendered

possible by the fact that the whole organization is military, and that its efforts are almost entirely concentrated on the preparation of these reports, while the telegraphic system of the States is placed at the disposal of the Signal Office for a certain space of time every day. By this means it is rendered possible to publish the chart and report simultaneously in all the principal cities of the States. It is not unimportant to consider what this system costs, as compared with our own. We spend, at the outside, £4,000 a year on our weather telegraphy, when we appor tion to it its due share of the expenses of the office, including rent; while the vote for the Signal Service, or, to use its familiar designation in the New York Herald, for "Old Probabilities," is no less than 250,000 dollars-about fourteen times as much as our own expenditure under the same head.

The first thing noticeable about our system of reporting stations is, that we are entirely exposed on the west side, the direction from whence most of our storms come, and that we have little prospect of improving our condition in this respect. Five years ago a proposal was made to moor vessels off our western coasts, having telegraphic communication with the shore, and to use them as floating observatories. The trial that was made with H.M.S. Brisk, at the entrance of the Channel, resulted in a total failure. The plan has been recently resuscitated by Mr. Morse, the well-known American telegraphic engineer. His idea is, that buoys will ride easier at anchor in a heavy sea than ships, and that these buoys might be made of such a size that habitable turrets might be built on them, which might be used as observatories.

Latterly an announcement has been made that the Portuguese Government intend to establish a reporting system between the Azores and the mainland, and an application has been made to the Meteorological Committee to contribute towards the expense. I need scarcely say that they have at once replied that they would be ready to assist in this particular development of weather telegraphy to the extent of their power.

We have, in the meantime, instituted a comparison between the 8 A.M. reports from Angra do Heroismo and from Valencia for the last two and three quarter