actual condition of the weather from day to day at selected stations in different countries. It cannot be denied that the inquiry has not been attended with success. At vast expense millions of records of heat, rainfall, winds, clouds, barometric pressure, and so on have been secured; but hitherto no law has been recognized in the variations thus recorded, — no law at least from which any constant system of prediction for long periods in advance can be deduced.

this discovery had been made, that the | is confined to the record and study of the weather of particular days or even of weeks and longer periods could not, by its means, be predicted. A week in summer may be cold, and a week in winter may be warm; nor, so far as is even yet known, is there a single part of any year the temperature of which can be certainly depended upon, at least within the temperate zone. In certain tropical regions there are tolerably constant weather variations; but so far is this from being the case in the temperate zones of either hemisphere, that it is impossible to affirm On this point I shall quote first a recertainly, even that during a week or fort-markable saying of Sir W. Herschel's, night at any given summer season there which appears to me, like many such saywill be one hot day, or that during a cor-ings of his, to be only too applicable to the responding period in winter there will be present state of science. In endeavouring one day of cold weather. to interpret the laws of weather, “we are It became manifest also, at an early in the position," Herschel remarks, “of a epoch, that terrestrial conditions must be man who hears at intervals a few fragintimately involved in all questions of ments of a long history related in a prosy, weather, since the year in different coun- unmethodical manner. A host of circumtries in the same latitudes presents differ- stances omitted or forgotten, and the want ent features. Such differences are of two of connection between the parts, prevents kinds, — those which have a tendency to the hearer from obtaining possession of be constant, and those which are in their the entire history. Were he allowed to nature variable. For example, the annu- interrupt the narrator, and ask him to al weather in Canadian regions having explain the apparent contradictions, or to the same range of latitude as Great Brit-clear up doubts at obscure points, he ain, differs always to a very marked degree, though not always to the same degree, from that which prevails in this country here then we have a case of a constant difference due unquestionably to terrestrial relations. Again, when we have a hot or dry summer in this country, warm or damp weather may prevail in other countries in the same latitudes, and vice versa; differences of this kind are ordinarily variable, and in the present position of weather-science are regarded as accidental.

*

Hitherto, weather-science has depended solely on the study of these terrestrial effects as they vary under varying conditions. Modern meteorological research

I use this qualifying word, because some differences of the kind are more or less regular. Thus, when there is a dry summer in certain regions in the West of Europe, there is commonly a wet summer in easterly regions in the same latitude, and vice versa, the difference simply depending on the height at which the clouds travel which are brought by the south-westerly counter-trade winds. When these clouds travel high, they do not give up their moisture until they have travelled far iaiand or towards the east; when they travel low, their moisture is condensed so soon as they reach the western landslopes. It is not uncommonly the case again, that when we in England have dry summers, much rain falls on the Atlantic, and our drought is simply due to the fall of this rain before the clouds from the south-west have reached us. More commonly, however, drought in England is due to the delay of the downfall, in consequence of the clouds from the southwest travelling at a greater height than usual.

might hope to arrive at a general view. The questions that we would address to nature, are the very experiments of which we are deprived in the science of meteorology."

The late Professor De Morgan, indeed, selected meteorology as the subject on which, above all others, systematic observations had been most completely wasted, -as a special instance of the failure of the true Baconian method (which be it noticed is not, as is so commonly supposed, the modern scientific method).

66

There is an attempt at induction going on," says De Morgan, "which has yielded little or no fruit, the observations made in the meteorological observatories. This attempt is carried on in a manner which would have caused Bacon to dance for joy" (query); "for he lived in times when Chancellors did dance. Russia, says M. Biot, is covered by an army of meteorographs, with generals, high officers, subalterns, and privates, with fixed and defined duties of observation. Other countries, also, have their systematic observations. And what has come of it? Nothing, says M. Biot, and nothing will ever come of it: the veteran mathematician and experi mental philosopher declares, as does Mr. Ellis" (Bacon's biographer), “that no single branch of science has ever been fruit

fully explored in this way." A special is received on a summer day in the southinterest attaches, I may remark, to the ern hemisphere. Or instead of comparopinion of M. Biot, because it was given ing our summer heat with summer heat in upon the proposal of the French govern- the southern hemisphere, we may make ment to construct meteorological observa- comparison between the quantity of heat tories in Algeria. received by the whole earth on a day in June and on a day in December. Either way of viewing the matter is instructive; and I believe many of my readers will be surprised when they hear what is the actual amount of difference.

It is well known that our Astronomer Royal holds a similar opinion. De Morgan thus quaintly indicates his interpretation of one particular expression of Sir G. Airy's opinion:-"In the report to the Greenwich Board of Visitors, for 1867, the Astronomer Royal, speaking of the increase of meteorological observatories, remarks, 'Whether the effect of this movement will be that millions of useless observations will be added to the millions that already exist, or whether something may be expected to result which will lead to a meteorological theory, I cannot hazard a conjecture.' This is a conjecture, and a very obvious one; if Mr. Airy would have given, 2 3-4d., for the chance of a meteorological theory formed by masses of observations, he would never have said what I have quoted."

The simple combination of terrestrial considerations with the effects due to the sun's varying daily path having thus far failed to afford any interpretation of the varying weather from year to year, it is natural to inquire whether the variations in the sun's condition from year to year may not supply the required means of interpreting and hence of predicting weather-changes. We know that the sun's condition does vary, because we sometimes see many large spots upon the surface, whereas at others he has no spots, or few and small ones. We can scarcely doubt that these variations affect the supply of heat and light, as well as of chemical action and possibly of other forms of force; and hence we are certainly dealing with a vera causa, though whether this real cause be an efficient cause of weatherchanges remains yet to be determined.

We receive in fact, on June 30th, less heat and light than dwellers at our antipodes receive on December 30th, by the amount which would be lost if an opaque disc having a diameter equal to onefourth of the sun's,* came upon the sun's face as seen on December 30 at our antipodes. It need hardly be said that no spots whose effects would be comparable with those produced by such a disc of blackness have ever been seen upon the face of the sun. Spots are not black or nearly black, even in their very nucleus. The largest ever seen has not had an extent approaching that of our imagined black disc, even when the whole dimensions of the spot, nucleus, umbra, and penumbra, have been taken into account. Moreover, all round a spot there is always a region of increased brightness, making up to a great degree, if not altogether, for the darkness of the spot itself. So that unquestionably the summer heat in the southern hemisphere exceeds the summer heat in our hemisphere to a much more marked degree than the heat given out by the sun when he is without spots exceeds the heat of a spotted sun.

It is, however, rather difficult to ascertain what effect is to be ascribed to this peculiarity. It is certain that the Australian summer differs in several important respects from the European summer; but it is not easy to say how much of the difference is due to the peculIt may perhaps be as well to inquire, iarity we have been considering, and however, in the first place, whether any how much to the characteristic distincpeculiarities of weather can be traced to tion between the northern and southern another circumstance which ought to be at halves of the earth, - the great excess of least as efficient, one would suppose, as water surface over land surface in the any changes in the sun's action due to the southern hemisphere. It is worthy of spots. I refer to his varying distance | from the earth. It is known doubtless to all my readers that in June and July, although these are our summer months, the sun is farther away than in December, and this, not by an inconsiderable distance, but by more than three millions of miles. Accordingly, on a summer day in our hemisphere we receive less heat than

It is easily shown that such would be the size of from about 93 millions of miles to about go millions, or the imagined black disc. For the sun's distance varies in the proportion of 31 to 30. Hence the size of his disc varies in the proportion of 31 times 31 to 30 times 30, or as 961 to goo. The defect of the latter number 900 amounts to 61, which is about a sixteenth part of the larger number. But a black disc having a diameter equal to a quarter of the sun's would cut off precisely a

sixteenth part of his light and heat, which was the fact to be proved.

notice, however, that even in this case, it not unfrequently happens that the sun where we cannot doubt that a great dif- shows few spots for several weeks toference must exist in the solar action at gether, in the very height of the time of particular seasons, we find ourselves spot-frequency, while on the other hand quite unable to recognize any peculiari-it often happens that many and large ties of weather as certainly due to this spots are seen at other times. Nevertheless, this general law holds, that, on the whole, and taking one month with another, there is a variation in spot-frequency, having for its period an interval of rather more than eleven years.

with, though not one of perfect uniformity. Next to the systematic changes already considered, this alternate waxing and waning of spot-frequency might be expected to be efficient in producing recognizable weather changes. Assuredly, if this should not appear to be the case, we should have to dismiss all idea that the sun-spots are weather-rulers.

M. Gautier, of Geneva, and later MM. Arago and Barratt made a series of researches into the tabulated temperature at several stations, and for many successive years. They arrived at the conclusion, that on the whole, the weather is coolest in years of spot-frequency.

But recently the matter has been more closely scrutinized, and it has been found that the effects due to the great solar spot period, although recognizable, are by no means so obvious as had been anticipated.

These effects may be divided into three classes, those affecting (1) temperature, (2) rainfall, and (3) terrestrial magnetism.

As respects the first, it has been discovered that when underground temperatures are examined, so that local and temporary causes of change are eliminated, there is a recognizable diminution of temperature in years when spots are most frequent. We owe this discovery to Professor C. P. Smyth, Astronomer Royal for Scotland. The effect is very slight indeed, barely recognizable. I have before me, as I write, Professor Smyth's chart of the quarterly temperatures from 1837 to 1869, at depths of 3, 6, 12, and 24 French feet. Of course,

Now, from the first discovery of spots, it was recognized that they must, in all probability, affect our weather to some degree. It was noticed, indeed, that our auroras seemed to be in some way influenced by the condition of the sun's surface, since they were observed to be more numerous when there are many spots than when there are few or none. Singularly enough, the effect of the spots on temperature was not only inquired into much later (for we owe to Cassini and Mairan the observation relating to auroras), but was expected to be of an opposite character from that which is in reality produced. Sir W. Herschel formed the opinion that when there are most spots the sun gives out most heat, notwithstanding the diminution of light where the spots are. He sought for evidence on this point in the price of corn in England, and it actually appeared, though by a mere coincidence, that corn the most remarkable feature even at had been the cheapest in years of spot- the depth of 24 feet, is the alternate frequency, a result regarded by Herschel rise and fall with the seasons. But it as implying that the weather had been is seen that while the range of rise and warmer on the whole in those years. It fall remains very nearly constant, the was well pointed out, however, by Arago, crests and troughs of the waves lie at that "in these matters we must be care-varying levels. After long and careful ful how we generalize facts before we have a very considerable number of observations at our disposal." The peculiarities of weather in a single and not extensive country like England, are quite insufficient to supply an answer to the wide question dealt with by Herschel. The weather statistics of many countries must be considered and compared. Moreover very long periods of time must be dealt with.*

* When Herschel made his researches into this subject, the law of spot-frequency had not been discovered. He would probably have found in this law, as some have since done, the explanation of the seven years of plenty and the seven years of famine typified by the fat kine and lean kine of Joseph's dream. For if there were a period of eleven years in which corn and other produce of the ground waxed and waned in productive ness, it would be not at all unlikely that whenever this waxing and waning chanced to be unusually marked, there would result two series of poor and rich years apparently ranging over fourteen instead of eleven years. We have seen, above, that the waves of spot-waxing and spot-waning are not all alike in shape and extent. Whenever then a wave more marked than usual came, we should expect to find it borrowing, so to speak, both

scrutiny, I find myself compelled to admit that I cannot find the slightest evidence in this chart of a connection between underground temperature and the eleven years' period of sun spots. I turn, therefore, to the chart in which the annual means are given; and noting in the means at the lesser depths "confusion worse confounded" (this, of course, is no fault of Professor Smyth's, who here merely records what had actually taken place), I take the temperatures at a depth of 24 French feet. Now, neglecting minor features, I find the waves of temperature thus arranged. They go down to a little more than 46 1-2 degrees of the common thermometer in 1839-40; rise to about 47 3-4 degrees in 1847; sink to 47 1-4 degrees in 1849; mount

in trough and crest from the waves on either side. It would require but a year or so either way to make the wave range over fourteen years; and observed facts even during the last half-century only, show this to be no unlikely event.

nearly to 47 3-4 degrees again in 1852-53; things prevailed there. Next we have the evidence of the Rev. R. Main, director of the Radcliffe Observatory at Oxford, who finds that westerly winds are slightly more common when sun-spots are numerous than at other times. And lastly, Mr. Meldrum, of Mauritius, notes that years of spot-frequency are characterized on the whole by a greater number of storms and hurricanes, than years when the sun shows few spots.

The association between the sun-spot period and terrestrial magnetism is of a far more marked character, though I must premise that the Astronomer Royal, after careful analysis of the Greenwich magnetic records, denies the existence of any such association whatever. There is, however, a balance of evidence in its favour. It seems very nearly demonstrated that the daily sway of the magnetic needle is greater when sun-spots are numerous, that magnetic storms are somewhat more numerous at such times, and that auroras also are more commonly seen. Now it has been almost demonstrated by M. Marié Davy, chief of the meteorological division in the Paris Ob

a general way by magnetic disturbances. So that we are confirmed in the opinion that indirectly, if not directly, the weather is affected to some slight degree by the great sun-spot period.

are at 47 degrees in 1856-57; are nearly at 48 degrees in 1858-59; then they touch 47 degrees three times (with short periods of rising between), in 1860, 1864, and 1867; and rise above 47 1-2 degrees in 1869. Now if we remember that there were maxima of spots in 1837, 1848, 1859-60, and 1870, while there were minima in 1843, and 1855-56, I think it will be found to require a somewhat lively imagination to recognize a very striking association between the underground temperature and the sun's condition with respect to spots. If many spots imply diminution of heat, how does it come that the temperature rises to a maximum in 1859, and again in 1869? if the reverse, how is it that there is a minimum in 1860? I turn, lastly, to the chart in which the sun-spot waves, and the temperature waves are brought into actual comparison, and I find myself utterly unable to recognize the slightest association between them. Nevertheless, I would not urge this with the desire of in any way throwing doubt upon the opinion to which Professor Smyth has been led, knowing well that the long and careful ex-servatory, that the weather is affected in amination he has given to this subject in all its details, may have offered ample though not obvious evidence for the conclusions at which he has arrived. I note also, that, as he points out, Mr. Stone, director of the Cape Town Observatory, Still I must point out that not one of and Mr. Cleveland Abbe, director of the these cases of agreement has anything Cincinnati Observatory, have since, "but like the evidence in its favour which had it is believed quite independently, pub- been found for an association between lished similar deductions touching the the varying distance of Jupiter and the earth's temperature in reference to sun-sun-spot changes. For eight consecuspots." All I would remark is, that the tive maxima and minima this associaeffect is very slight and very far from being obvious at a first inspection. Next as to rainfall and wind. Here, again, we have results which can hardly be regarded as striking, cept in the forcible evidence they convey of the insignificance of the effects which are to be imputed to the great eleven-year spot period. We owe to Mr. Baxendell, of Manchester, the most complete series of investigations into this subject. He finds that at Oxford, during the years when sun-spots were most numerous, the amount of rainfall under west and south-west winds was greater than But accepting the association between the amount under south and south-east weather and the sun-spot changes as winds, while the reverse was the case in demonstrated (which is granting a great years when spots were few and small. deal to the believers in solar weatherApplying corresponding processes to the prediction), have we any reason to bemeteorological records for St. Peters-lieve that by a long-continued study of burg, he finds that a contrary state of the sun the great problem of foretelling

tion has been strongly marked, and might be viewed as demonstrated,— only it chances unfortunately that for two other cases the relation is precisely reex-verscd; and in point of fact, whereas the period now assigned to the great sunspot wave is eleven years and rather less than one month, Jupiter's period of revolution is eleven years and about ten months, a discrepancy of nine months, which would amount up to five and a half years (or modify perfect agreement into perfect disagreement) in seven or eight cycles.