Temperature Tables, by Prof. W. H. Dove, Cor. Mem. of the British Association with Introductory Remarks by Lieut.-Col. EDWARD SABINE, General Secretary.

THE following pages contain the numerical elements of an investigation on which Professor Dove has been employed for several years past; they consist of the mean temperatures of the different months of the year, of the seasons, and of the year itself, at above 800 stations on the surface of the globe, with the differences between the summer and winter seasons, and the hottest and coldest months, and a notice of the number of years during which the observations have been continued in each case, and of the hours of observation. The original memoir, in which this assemblage of the results of observation will be embodied, will appear in the volume of the Transactions of the Berlin Academy of Sciences, which will be published in the course of 1848; in that memoir the temperatures will be expressed in degrees of Reaumur's scale, which is more commonly used in Germany than either Fahrenheit's or the Centigrade; the transformation of these into Fahrenheit's scale, in which they appear in the subjoined tables, is Professor Dove's own work; a labour undertaken and executed by him expressly for the convenience of the members of the British Association, and of those who profit by its publications. Whether we regard this sacrifice of time, on the part of a person who is undoubtedly in the first rank of those who are advancing by the sure path of induction into hitherto uninvestigated regions of meteorological science, simply as an evidence of his desire to promote the general advancement of science; or, as perhaps we may permit ourselves to do, as an evidence of that general desire, of which he has given so many proofs, strengthened in this particular case by the recollection of the respect and regard with which his presence was welcomed at the Cambridge meeting of the British Association in 1845,-we are bound, in either case, to express our acknowledgements for a labour which certainly augments greatly the value of the tables to British readers.

The mean temperatures have not been corrected for the diurnal variation corresponding to the particular hour or hours at which the observations were made, excepting in a few instances, which are marked red. (for reduced) in the column showing the hours of observation; but nine tables have been added, containing in degrees of Fahrenheit the corrections to the true mean quantities for nine stations differing widely in geographical circumstances, computed by Bessel's formula

is prefixed, and north when there

tx=u+u'(sin x+U')+u"(sin 2x+U")+u'''(sin 3x+U'''). The latitudes are south when the sign is no sign. The longitudes are east when there is no sign. The heights are in British feet. the column showing the hour of observation, it signifies that the results have been computed by the formula

N.Y.=a+2b+2c+a'

6

is prefixed, and west when Where N.Y. occurs in

in which a is the observation at 6 A.M., b at 3 P.M., c at one hour after sunset, and a' at 6 A.M. the following day. D, in the final column, signifies that the mean temperatures of the several months at the stations so distinguished, are taken from one or other of four memoirs published by Professor Dove in the Transactions of the Berlin Academy, entitled 'Über die nicht periodischen Änderungen der Temperaturvertheilung auf der Oberflache der Erde.' The seasons are the usual meteorological seasons, winter consisting of December, January and February, and so forth.

Besides the correction already noticed for the diurnal variation of the temperature corresponding to the particular hour or hours at which the observations at a station have been made (which has only been applied in a few specified instances in the subjoined tables, as it appeared preferable to give generally the original determinations independent of the corrections), the mean temperatures derived from the particular year or years during which the observations were continued at each station, will require in each case to be corrected for the non-periodic variation of the temperature in the year or years in question. On this subject the reader is referred to the preface and to pp. 34 to 59 of Professor Dove's fourth memoir (entitled as above) in the Transactions of the Berlin Academy for 1847. From the mean temperatures corrected for the diurnal and for the non-periodic variations, Professor Dove proposes to construct isothermal curves for each month of the year, and has expressed his intention of taking a future opportunity of communicating the curves to the British Association; but he is desirous that in the mean time a few of the more striking phænomena which have been brought into notice by this extensive assemblage and intercomparison of the results of observation should be briefly indicated.

1. The most novel at least, if not the most important of these, is the remarkable cosmical fact which they establish, of the existence of an annual variation in the aggregate mean temperature at the surface of the globe, having its period of maximum during the sun's northern declination, and of minimum during his southern declination.

If we compare the temperatures at stations in the southern hemisphere with those of stations of corresponding latitude in the northern hemisphere, we find that it is generally the fact, that the sum of the simultaneous mean temperatures at the two stations in the months of June, July and August (or of those months which constitute the northern summer and the southern winter), is greater than the sum of the simultaneous mean temperatures in December, January and February, or in the months which constitute the northern winter and southern summer. Let us take for example the comparison of the mean temperatures in the seasons at Toronto, latitude 43° 40′ N., and Hobarton, latitude 42° 53′ S., derived in each case from five years of hourly observation at the magnetical and meteorological observatories at those stations; we have from 37,296 observations at Toronto, and 37,320 at Hobarton (Sundays not being days of observation) the mean temperatures of the different seasons as follows:

We perceive by this comparison that the mean temperature of the months of December, January and February at Toronto + the mean temperature of the same months at Hobarton, exceeds by 22°.7 the mean the months of June, July and August at Toronto + the mean of the same months at Hobarton.

temperature of temperature

If a similar comparison be made of stations in South America, South

Africa, New Holland, New Zealand, or the Falkland Islands, with stations in corresponding latitudes in the northern hemisphere, we find a similar result, though varying, as might be expected, in the amount of difference at the two periods of the year; but all testifying that the aggregate mean temperature at the surface of the globe is greater in June, July and August, than in December, January and February. This is a result of the unequal distribution of land and sea in the two hemispheres, and of the difference in the effect produced by the sun's heat according as it impinges on a solid or on a liquid surface. It is well-known that Britain is nearly the middle point of an hemisphere distinguished by the preponderance of land at the surface, and New Zealand the middle, or nearly so, of an hemisphere distinguished by the preponderance of an aqueous surface. The greater proportion of land is consequently in the northern hemisphere as taken with reference to the sun's path; whilst the southern hemisphere has the greater proportion of ocean surface. The existence of a periodical variation of the aggregate mean temperature derived from the sun, might thence have been (and possibly may have been) anticipated, but it is now made known to us as an experimental fact, by the modern practice of systematic observation, and of the intercomparison of mean numerical results so obtained.

2. The differences between the hottest and the coldest months of the year, which form a part of the results of Professor Dove's investigation, are a very important element in judging of climatic relations.

If the intercomparison of these differences be extended so as to include places which differ in latitude as well as in other circumstances by which climate is influenced, we may find in these tables striking examples of the great diversity which takes place in different parts of the globe in the distribution of the mean temperature of the year into its several months and seasons. Thus if we refer to Commewine in Guiana, and Jakutsk in Siberia, we find the difference between the hottest and coldest months to be 2°.2 of Fahrenheit at Commewine, and 114°4 at Jakutsk; and if we limit the comparison to places which correspond in latitude, to stations for example in Northern Asia compared with others in Europe or in North America, we shall still find that the diversity is very great in this respect.

The terms of Insular and Continental climates have been employed to characterise respectively those climates in which the difference of the mean temperatures of the hottest and coldest months is very small, and those in which it is extremely large.

The distinction thus established and the designations so employed have been understood to refer exclusively to the differences between the extreme temperatures either of months or of seasons, and to be irrespective of the absolute mean temperature of the year, season or month. It is obvious however that circumstances might exist, which would cause the mean temperature of the coldest month, for example, at a station to be extremely low when viewed in reference to what might be regarded as the norinal temperature of that month corresponding to the parallel of latitude, whilst other circumstances might be wanting which would occasion the mean temperature of the hottest month to be excessive in the other extreme, when compared in like manner with a normal temperature corresponding to the latitude; or vice versa, that the temperature of the hottest month might be excessive in respect to heat, whilst that of the coldest month might not suffer a corresponding depression. In such cases the mere difference between the temperatures of opposite seasons or of the extreme months, irrespective of the difference of the temperature of either from a normal temperature, would fail to render manifest and to distinguish the peculiarity which prevails during one portion

of the year only. By the assemblage of a great body of observations from a great variety of places, normal values may now be assigned for the mean temperature of months and seasons as well as of the year, and thus abnormal differences occurring at any period or during any portion of the year may at once be recognised. From the progress already made in this branch of the inquiry, Professor Dove is enabled to show, in respect to the continents of the northern hemisphere, that northern and central Asia have what may be termed a true Continental climate both in winter and in summer, viz. a cold winter and a hot summer; that Europe has a true insular or sea climate in both seasons, viz. a mild winter and a cold summer; and that North America inclines to a continental climate in winter and to a sea climate in summer. Consistently with these distinctions, the isothermals representing mean temperatures of the year are found in a lower latitude in America than in either Europe or Asia: the differences between the extremes of temperature in the year are greater in northern and central Asia than in North America, and greater in North America than in Europe; and the lemniscate form of the winter isothermals around two centres, one in Asia and the other in America, passes in the summer isothermals into a circular form around an American centre only. The lemniscate form of the lower isothermals in the northern hemisphere has been supposed by some philosophers to indicate a causal connexion with the magnetic isodynamics of highest value, from the circumstance that these latter have also the form of lemniscates. The systematic and progressive secular change which the isodynamics undergo has always, however, constituted a strong objection to this hypothesis of causal connection; and it now further appears that the lemniscate form itself of the isothermals subsists only during the winter season; that it is in fact merely a periodical phase, whereas the isodynamic lemniscates undergo no such periodical variations.

Many local meteorological phænomena previously known as remarkable but insulated facts, are found to arrange themselves in due subordinate connexion with the comprehensive and systematic views which result from so extensive a generalisation: amongst these Professor Dove mentions-the diminution of atmospheric pressure in Asia in summer; the transformation of the trade-winds into monsoons; the change in the direction of greatest cold in the thermic wind-rose; the absence of the subtropical rains at the northern limits of the monsoons; the small annual variations of the meteorological phænomena in tropical America compared with the great periodical variations in tropical Asia; and the late occurrence (usually in February) of the minimum of temperature in North America.

I need not add how valuable would be a discussion of the special causes of these and of other cognate phænomena from the pen of Professor Dove himself in the volumes of the Association, accompanying the curves which he has so kindly promised to communicate.

EDWARD SABine.