Abstract of Baron Humboldt's Dissertation on Isothermal Lines, and the Distribution of Heat over the Globe. This highly valuable and extremely interesting Dissertation has imperative claims on the space of this Journal; but it occupies too much room to permit us to give it in detail: only those points relating especially to medical science will, therefore, be selected; but the extent of these even will render it necessary to divide our abstract into portions which may be inserted in the present and two or three subsequent Numbers.--Edit. THE distribution of heat over the globe belongs to that kind of phenomena, of which the general circumstances have been long known, but which were incapable of being rigorously determined or submitted to exact calculation, till experience and observation furnished data from which the theory might obtain the corrections and the different elements which it requires. The object of this memoir is to facilitate the collection of these data, to present results drawn from a great number of unpublished observations, and to group them according to a method which has not yet been tried, though its utility has been recognized for more than a century in the exposition of the phenomena of the variation and dip of the magnetic needle. As the discussion of individual observations will be published in a separate work, I shall at present limit myself to a simple sketch of the distribution of heat over the globe, according to the most recent and accurate data. Although we may not be able to refer the complex phenomena to a general theory, it will be of considerable importance to fix the numerical relation by which a great number of scattered observations are connected, and to reduce to empirical laws the effects of local and disturbing causes. The study of these laws will point out to travellers the problems to which they should direct their principal attention; and we may entertain the hope that the theory of the distribution of heat will gain in extent and precision, in proportion as observations shall be more multiplied, and directed to those points which it is of most importance to illustrate. As the phenomena of geography and of vegetables, and in general the distribution of organized beings, depend on the knowledge of the three co-ordinates of latitude, longitude, and altitude, I have been occupied for many years in the exact valuation of atmospherical temperatures; but I could not reduce my own observations without a constant reference to the works of Cotte and Kirwan, the only ones which contain a great mass of meteorological observations obtained by instruments and methods of very unequal precision. Having inhabited for a long time the most elevated plains of the New Continent, I availed myself of the advantages which they present for examining the temperature of the superincumbent strata of air, not from insulated data, the results of a few excursions to the crater of a volcano, but from the collections of a great number of observations made day after day and month after month in inhabited districts. In Europe, and in all the Old World, the highest points of which the mean temperatures have been determined are the convent of Peissenberg, in Bavaria, and the hospice of St. Gothard. The first of these is placed at 3264, and the second at 6808, feet above the level of the sea. In America, a great number of good observations have been made at Santa Fe de Begota and at Quito, at altitudes of 8,727 and 9,544 feet. The town of Huancavelica, containing 10,000 inhabitants, and possessing all the resources of modern civilization, is situated in the Cordilleras of the southern hemisphere, at 12,310 feet of absolute elevation; and the mine of Santa Barbara, encircled with fine edifices, and placed a league to the south of Huancavelica, is a place fit for making regular observations, at the height of 14,509 feet, which is double that of the hospice of St. Gothard. The mean temperature of the air at the convent of the Great St. Bernard, the height of which is 7,960 feet, is not determined. There are several villages in Europe placed at more than 5000 feet of altitude; for example, St. Jacques de Ayas at 5,479, and Trinita Nuova, near Grasfoncy, at 5,315 feet. These examples are sufficient to prove how much our knowledge of the higher regions of the atmosphere, and of the physical condition of the world in general, will increase, when the cultivation of the sciences, so long confined to the temperate zone, shall extend beyond the tropics into those vast regions where the Spanish Americans have already devoted themselves with such zeal to the study of physics and astronomy. In order to compare, with the mean heat of temperate climates, the results which Mr. Bonpland and I obtained in the equinoctial regions from the plains to the height of 19,292 feet, it was necessary to collect a great number of good observations made beyond the parallels of 30° and 35°. I soon perceived how vague such a comparison was, if I selected places under the meridian of the Cordilleras, or with a more eastern longitude; and I therefore undertook to examine the results contained in the most recent works. I endeavoured to find, at every ten degrees of latitude, but under different meridians, a small number of places whose mean temperature had been precisely ascertained, and through these, as so many fixed points, passed my isothermal lines, or lines of equal heat. I had recourse, in so far as the materials have been made public, to those observations the results of which have been published; and I found, in the course of this easy, but long and monotonous, labour, that there are many mean temperatures pointed out in meteorological tables, which, like astronomical positions, have been adopted without examination. Sometimes the results were in direct contradiction to the most recent observations, and sometimes it was impossible to discover from whence they were taken. Many good observations were rejected, solely because the absolute height of the place where they were made was unknown. This is the case with Asia Minor, Armenia, and Persia, and of almost all Asia; and, while the equinoctial part alone of the New World presents already more than five hundred points, the greater number of which are simple villages and hamlets, determined by barometrical levelling, we are still ignorant of the height of Erzeroum, Bagdad, Aleppo, Teheran, Ispahan, Delhi, and Lassa, above the level of the neighbouring seas. Notwithstanding the intimate relation in which we have lately stood with Persia and Candahar, this branch of knowledge has not made any progress in the last fifty years. We are not authorized, however, on account of the decrease of temperature in the upper regions of the atmosphere, to confound the mean temperatures of places which are not placed on the same level. In the Old World, good observations, which can alone be used for establishing empirical laws, are confined to an extent between the parallels of 30 and 70 degrees of latitude, and the meridians of 30° east longitude and 20° of west longitude. The extreme points of this region are the island of Madeira, Cairo, and the North Cape. It is a zone which is only a thousand nautical leagues, (one-seventh of the circumference of the globe,) from east to west, and which, containing the basin of the Mediterranean, is the centre of the primitive civilization of Europe. The extraordinary shape of this part of the world, the interior seas and other circumstances, so necessary for developing the germ of cultivation among nations, have given to Europe a particular climate, very different from that of other regions placed under the same latitude. The temperature of the atmosphere and the magnetism of the globe, cannot, like those phenomena which depend on one cause or on a single centre of action, be disengaged from the influence of disturbing circumstances, by taking the averages of many observations in which these extraneous effects are mutually destroyed. We must, however, guard against confounding, under the name of extraneous and disturbing causes, those on which the most important phenomena, such as the distribution and the more or less rapid development of organic life, essentially depend. Our object is to ascertain the quantity of heat which every point of the globe annually receives, and, what is of most importance to agriculture and the good of its inhabitants, the distribution of this quantity of heat over the different parts of the year, and not that which is due to the solar action alone, to its altitude above the horizon, or to the duration of its influence, as measured by the semidiurnal arcs. Moreover, we shall prove that the method of means is unfit for ascertaining what belongs exclusively to the sun, (inasmuch as its rays illuminate only one point of the globe,) and what is due both to the sun and to the influence of foreign causes. In distinguishing, as has long been done, between the solar and the real climate, we must not forget that the local and multiplied causes which modify the action of the sun upon a single point of the globe, are themselves but secondary causes, the effects of the motion which the sun produces in the atmosphere, and which are propagated to great distances. If we consider separately (and it will be useful to do this in a discussion purely theoretical,) the heat produced by the sun, the earth being supposed at rest and without an atmosphere, and the heat due to other causes regarded as disturbing ones, we shall find that this latter part of the total effect is not entirely foreign to the sun. The influence of small causes will scarcely disappear by taking the mean result of a great number of observations; for this influence is not limited to a single region. By the mobility of the aerial ocean, it is propagated from one continent to another. Every where in the regions near the polar circles, the rigors of the winters are diminished by the admixture of the columns of warm air, which, rising above the torrid zone, are carried towards the poles: every where in the temperate zone, the frequent west winds modify the climate, by transporting the temperature of one latitude to another. When we reflect, besides, on the extent of seas, on the form and prolongation of continents, either in the two hemispheres, or to the east and west of the meridians of Canton and of California, we shall perceive that, even if the number of observations on the mean temperature were infinite, the compensation would not take place. Raymond, Memoire sur la Formule Baromet. p. 108 and 113. It is, then, from the theory alone that we must expect to determine the distribution of heat over the globe, in so far as it depends on the immediate and instantaneous action of the sun. It does not indicate the degrees of temperature expressed by the dilatation of the mercury in a thermometer, but the ratios between the mean annual heat at the equator, at the parallel of 45°, and under the polar circle; and it determines the ratios between the solstitial and equinoctial heats in different zones. By comparing the results of calculation, not with the mean temperature drawn from observations made under different longitudes, but with that of a single point of the earth's surface, we shall set out with that which is due to the immediate action of the sun and to the whole of the other influences, whether they are solar or local, or propagated to great distances. This comparison of theory with experience will present a great number of interesting relations. In the year 1693, previous to the use of comparable thermometers, and to precise ideas of the mean temperature of a place, Halley laid the first foundations of a theory of the heating action of the sun under different latitudes. He proved that these actions might compensate for the effect of the obliquity of the rays. The ratios which he points out do not express the mean heat of the seasons, but the heat of a summer-day at the equator and under the polar circle, which he finds to be as 1·834 to 2·310. Phil. Trans. 1693, p. 878. In detailing the actual state of our knowledge on the distribution of heat, I have shown how dangerous it is to confound the results of observation with theoretical deductions. The heat of any point of the globe depends on the obliquity of the sun's rays and the continuance of their action, on the height of the place, on the internal heat and radiation of the earth in the middle of a medium of variable temperature; and, in short, upon all those causes which are themselves the effects of the rotation of the earth and the inequal arrangement of continents and seas. Before laying the foundation of a system, we must group the facts, fix the numerical ratios, and, as I have already pointed out, submit the phenomena of heat, as Halley did those of terrestrial magnetism, to empirical laws. In following this method, I have first considered whether the method employed by meteorologists for deducing the mean temperature of the year, the month, and the day, is subject to sensible errors. Assured of the accuracy of the numerical averages, I have traced upon a map the isothermal lines, analogous to the magnetic lines of dip and variation. I have considered them at the surface of the earth in a horizontal plane, and on the declivity of mountains in a vertical plane. I have examined the increase of temperature from the pole to the equator, which is inequal under different meridians; the distribution of the same quantity of heat over the different seasons, in the same isothermal parallel, and under different latitudes; the curve of perpetual snows, which is not a line of equal heat; the temperature of the interior of the earth, which is a little greater towards the north, and in high mountains, than the mean temperature of the atmosphere under the same parallel; and, lastly, the distribution of heat in the ocean, and the position of those bands, which may be called bands of the warmest waters. As the limits of this extract will not permit me to enter, in a detailed manner, upon these different discussions, I shall confine myself solely to the principal results. It was formerly the custom to take the maximum and minimum of temperature observed in the course of a year, and to consider half the sum as the mean temperature of the whole year. In order to diminish the errors of the method of annual extremes, it was perceived, though very late, that it was necessary to subdivide the curve which expresses the variation of temperature. Twenty-four extremes divided among twelve months of the year, give an annual mean more exact than the two extremes of all the observations. The ordinates do not increase uniformly and uninterruptedly up to the maximum of the year, and there are partial inflexions sufficiently regular. The more we subdivide, and the more we know the terms in the series, the more will these terms approximate, and the less error will there be in the supposition of an arithmetrical progression, and in that of the equidistance of the different maxima and minima of temperature. These considerations enable us to appreciate the three methods according to which observations are at present made. 1. Observations are made three times a-day, at sunrise and sunset, and at two o'clock in the afternoon. This was done at Geneva, during the three years 1796, 1797, and 1798. In the observations, the hour of mid-day was preferred to that of sunset. 2. Observations are made twice every day, at the two periods which are supposed to give the maximum and the minimum,--namely, at sunrise and at two o'clock in the afternoon. 3. Observations are made once a-day, at an hour which, in different seasons, has been found to represent the mean temperature of the day. It is thus that Mr. Raymond, by a judicious induction, has proved that the height of the barometer at mid-day gives, in our climates, the mean atmospherical pressure, corrected for the diurnal variation. In calculating a great number of observations made between the parallels of 46° and 48°, I have found that a single observation at sunset gives a mean temperature, which differs only some tenths of a degree from that which is deduced from observations made at sunrise and at two o'clock. The deviations of different months do not exceed 1·8, and they are very regularly positive or negative, according to the order of the seasons. Mr. Arago has examined for seven years the observations of noon. They give for Paris 5·4 more than the mean temperature of the whole year. Upon high mountains in the temperate zone, the difference is scarcely 1·8. By the application of coefficients, variable according to the latitude and the elevation, we may deduce the true mean temperatures from observations made at any particular period of the day, nearly in the same manner as we can ascertain the latitude of a place from altitudes of the sun, taken out of the meridian. De la Formule Barom. p. 213. The mean of the observations at noon at Paris, was 56·84; at Clermont in Auvergne, (elevation, 1348 feet,) 56·30; at Strasburg, (elevation, 453 feet,) 55·22.--Bulletin de la Soc. Philom. 1814, Oct. p. 95. At the hospice of St. Gothard.--Ephem. Soc. Pal. 1785, p. 47. If we do not stop at two observations of the maximum and minimum, but add a third observation, we commit an error more or less serious, if we divide simply by three the sum of the observations, without attending to the duration of the partial temperatures and to the place which the third observation occupies between the last terms of the series. Experience proves that the mean temperatures of the year, obtained by two or three observations, do not differ sensibly, if the intermediate observation is sufficiently distant (four or five hours) from the observation of the maximum and minimum. Whenever, therefore, we do not take into account the duration of the intermediate temperatures, we should prefer the two observations of the extreme temperature, which is the method most generally adopted. We shall content ourselves with pointing out the errors to which it is liable. In our climates, the two extreme terms do not divide the series of twentyfours into two equal parts. The maximum is an epoch nearly fixed: the rising of the sun retards or hastens it three hours. As we ought to take into account the duration of the partial temperature, in order to find the quantity of heat divided between the night and the day, we must couple the maximum of one day with the minimum of the day following, and not be content with taking half the sum of all the maxima and minima of a month. In the ordinary method, we determine only the mean temperature of the part of the day comprehended between the rising of the sun and two o'clock in the afternoon; and we take it for granted that the mean temperature is the same from two o'clock to sunrise next day. This double error, of want of equidistance and of the coupling of observations, does not in general produce errors of more than some tenths of degrees, sometimes in excess, and sometimes in defect, since the warm and cold days are mixed. Example.--On the 13th June, at 4h in the morning, 46·4; at 2h in the afternoon, 55·4; and at 11h in the evening, 50°, (erroneously, 46·4, or 8° centig. in the original.) In calculating by the duration, we have The true mean of which is 51·0. The mean of the three observations, as commonly taken, is 50·6. If we stop at the two extreme temperatures, we shall have for their half sum 50·45. 50·9, the mean for 10h of interval, = 509·0 52·7 9 = 474·3 48·2 5 = 241·0 Example.--At sunrise at 6h, 50°; at 2 o'clock in the afternoon, 62·6. At sunset, 51·8; at 2h, 66·2; at sunrise, 50°. The true means will be, for the first 24 hours, 56·9; and for the second, 59·0; for we shall have The method commonly employed gives 1/2 (50° + 62·6) = 56·3, and 1/2 (66·2 + 51·8) = 58·1. The errors being -- 0·6 and + 0·9, sometimes positive and sometimes negative. For 8h, 1/2 (50·0 + 66·2) x 8 = 450·4 for 8h = 472·0 16h, 1/2 (51·8 + 62·6) x 16 = 915·2 for 16h = 929·6. The error disappears when days of equal temperature succeed one another. It amounts to 1·8, if the mean temperatures of two successive days differ from 7° to 9°, which however very rarely happens. All the calculated results will err in defect, if the 365 ordinates through which the curve of the year passes do not express an arithmetical progression, and if the partial irregularities do not sensibly compensate one another. It is only on this supposition that we can judge by the extreme terms of the series, of the sum of the terms; that is, of the partial temperatures. It is very obvious, that near the maximum the increase ought to be more slow than in other points of the curve and that this increase in the temperature of the air ought to depend on the sine of the sun's altitude, and on the emission of the radiant heat of the globe. It appeared to me very important to establish, by observations made at every hour, at different periods of the year, and under different latitudes, the degree of confidence that can be placed in those results which are called mean temperatures. I have selected from the registers of the Royal Observatory at Paris clear and calm days, which offered at least ten or twelve observations. Under the equator, I have spent whole days in determining the horary increments and decrements of temperature, in marking the thermometer both in the shade and in the sun, and also the progress of evaporation and humidity; and, in order to avoid calculation, I measured with a quadrant the altitude of the sun at each observation. I chose days and nights completely calm, and when the heavens were entirely free from clouds, because the mass of vesicular vapours interrupts the radiation from the earth. The result of these experiments has been very satisfactory, and proves what had already been deduced from the coincidence between the temperature of the earth and the mean of daily observations, and from the regular progress of the mean temperatures of months in different years, that the effects of small disturbing causes may be compensated by a great number of observations. I have obtained analogous results by taking, for several months, the mean of 9 o'clock in the morning, of sunrise and midnight. I have computed the temperatures by the distance of the maximum expressed in time, and on the supposition of an arithmetical progression. I have found that, under the Torrid Zone, the morning curve, from sunrise to the maximum, differs very regularly from the evening curve. In the morning, the true mean temperature, such as we find by taking the duration into account, is a little greater than half the sum of the extremes. In the evening, the error is in a contrary direction, and the series of temperatures approaches more to a progression by quotients. The differences do not in general exceed half a degree, and calculation proves that the compensation is regular. It would be curious to examine the effect which the radiation of the earth has on these horary effects, as the changes of temperature at the surface do not follow the geometrical progression, in so far as they take place in a medium of uniform temperature. On the 3d and 4th September, 1811, lat. 48° 50'. The three last days show an equality of temperature which is very surprising, and which does not appear but in the true means. Sum of the temperatures during 24 hours. True mean temperature, taking into account the duration. Half sum of the two extreme temperatures. 625·71 Fahr. 57·92 Fahr. 58·28 Fahr. 672·49 59·90 61·88 834·67 66·74 65·12 834·67 66·74 68·00 835 37 66·74 63·50 63·61 Mean. 63·35 Mean. Example.--Latitude 10°25'. Calculation of a true mean by the duration. Supposition of an arithmetical pregression. Before the maximum, 11th September 1799, 70·52 Fahr. 6944 Fahr. 14th 69·26 68·00 18th 71·24 70·34 After the maximum, 18th August, 68·72 69·80 20th 70·16 71·24 27th 68·72 69·26 Before the maximum, 17th August, 69·26 68·00 After the maximum, 17th August, 65·48 66·02 Total effect, 17th August, 67·37 67·01 In order to avoid the use of an arbitrary measure, astronomers express the diameters of the planets by taking that of the earth for unity. In like manner, I express the mean temperatures, not in parts of the equatorial heat, but by the arithmetical ratios which subsist between this heat and that of the other parallels. This method frees us from the want of uniformity, which arises from the use of different thermometers. Instead of saying, that in Europe, under the parallel of 45°, the mean temperature is 13·4 Centigrade, or 56·12 of Fahrenheit, we say that it is = 1·0°,487, and in lat. 55° = 1·0°,29. These arithmetical ratios inform us of what is most interesting in the theory of the distribution of heat, that, in thermometers whose zero is the point of melting ice, the mean temperatures under the latitude of 45° and 55° are, in our regions, the half and the third nearly of the equatorial temperature, which I suppose to be 81·5. Having discussed the method of taking averages, and of reducing temperatures to general expressions, we shall now proceed to trace the course of the isothermal lines on the surface of the globe, and at the level of the sea. From a slight attention to the difference of climates, it has been remarked, more than a century ago, that the temperatures are not the same under the same parallels; and that, in advancing 70° to the east or the west, the heat of the atmosphere suffers a sensible diminution. In pursuance of our method, we shall reduce these phenomena to numerical results, and show that places situated under the same latitudes do not differ, in America and Europe, by the same number of degrees of temperature, as has been vaguely stated. This assertion would make us suppose that the isothermal lines are parallel in the temperate zone. Lat. Mean Temp. I. Parallels of Georgia, of the State of Mississippi, of Lower Egypt, and Madeira. Natchez .......... 31° 28' 64° 8' Funchal .......... 32 37 68 7 Orotava .......... 28 25 69 8 Rome ............ 41 53 60 4 Algiers .......... 36 48 70 0 Difference .... 7 0 4 1 II. Parallels of Virginia, Kentucky, Spain, and the South of Greece. Williamsburg ...... 38 8 58 0 Bourdeaux ........ 44 50 56 5 Montpellier ...... 43 36 59 4 Rome ............ 41 53 60 4 Algiers .......... 36 48 70 0 Difference .... 7 0 7 7 Lat. Mean Temp. III. Parallels of Pennsylvania, Jersey, Connecticut, Latium, and Romelia. Philadelphia ...... 39°56' 54° 9' New-York ........ 40 40 53 8 St. Malo .......... 48 39 54 5 Nantes .......... 47 13 54 7 Naples .......... 40 50 63 3 Difference .... 7 0 9 5 Ipswich .......... 42 38 50 0 Cambridge (Amer.) 42 25 50 4 Vienna ............ 48 13 50 5 Manheim .......... 49 29 51 3 Toulon .......... 43 7 63 1 Rome ............ 41 53 60 4 Difference .... 6 30 11 0 IV. Parallels of Canada, Nova Scotia, France, and the South of Germany. Quebec ............ 46 47 41 9 Upsal ............ 59 51 41 9 Padua ............ 45 24 57 7 Paris ............ 48 50 51 4 Difference .... 13 0 12 6 V. Parallels of Labrador, the South of Sweden, and Courland. Nains ............ 57 0 26 4 Okak ............ 57 20 29 8 Umea ............ 63 50 33 3 Enontekies ........ 68 30 27 0 Edinburgh ........ 55 58 47 8 Stockholm ........ 59 20 42 3 Difference .... 11 0 17 1 This table indicates the difference of climates, expressed by that of the mean temperature, and by the number of degrees in latitude which it is necessary to go northward in Europe, in order to find the same quantity of annual heat as in America. As a place could not be found in the Old World whose mean temperature was 48°, the same as that of Williamsburg, I have supplied it with an interpolation between the latitudes of two points whose mean temperatures are 56·5 and 59·4. By an analogous method, and by employing only good observations, I have found that See my Prolegomena de Distributione Geographica Plantarum, secundum Coeli temperiem et altitudinem montium, p. 68. 1. The isothermal line of 32° (0° centig.) passes between Uleo and Enontekies in Lapland, (lat. 66° to 68°; East long. from London 19° to 22°,) and Table Bay in Labrador, (lat. 54° 0'; W. long. 58°.) 2. The isothermal line of 41° (5° centig.) passes by near Stockholm, (lat. 60°, East long. 18°,) and the Bay of St. George in Newfoundland, (lat. 48°, and long. 59°.) 3. The isothermal line of 50° (10° centig.) passes by Belgium, (lat. 51°, East long. 2°,) and near Boston, (lat. 42° 30', West long. 70° 59.) 4. The isothermal line of 59° (15° centig.) passes between Rome and Florence, (lat. 43° 0', East long. 11° 40',) and near Raleigh in North Carolina, (lat. 36° 0', and West long. 76° 30'.) The direction of these lines of equal heat gives, for the two systems of temperature which we know by precise observations,--viz. part of the middle and west of Europe, and that of the coast of America, the following differences: Latitude. Mean Temp. of the West of the Old World. Mean Temp. of the East of the New World. Difference. 30 70·52 66·92 3·60 40 63·14 54·50 8·64 50 50·90 37·94 12·96 60 40·64 23·72 16·92 If we call the mean equatorial temperature 1, we shall have the half of this temperature in the Old World at 45°, and in the east of the New World at 39° of lat. The mean temperatures decrease Latitude. Temp. Temp. From 0°--20° In the Old World, 3·6 In the New World, 3·6 20 --30 7·2 10·8 30 --40 7·2 12·6 40 --50 12·6 16·2 50 --60 9·9 13·3 0 --60 40·5 56·5 In both continents, the zone in which the mean temperature decreases most rapidly is comprehended between the parallels of 40° and 45°. Observation here presents a result entirely conformable to theory, for the variation of the square of the cosine, which expresses the law of the temperature, is a maximum towards 45° of latitude. We have traced the direction of the isothermal lines from Europe to the Atlantic provinces of the New World, We have seen them approach one another from parallelism towards the south, and converge towards the north, particularly between the thermometric curves of 41° and 50°: we shall now endeavour to pursue them to the west. North America presents two chains of mountains, extending from N.E. to S.W. and from N.W. to S.E. forming almost equal angles with the meridian, and nearly parallel to the coasts which are opposite to Europe and Asia,--viz. the chain of the Alleghanys and the Rocky Mountains, which divide the waters of the Missouri and the Columbia. Between these chains stretch the vast basin of the Mississippi, the plains of Lousiana and of the Tenesse, and the states of Ohio, the centre of a new civilization. It is generally believed in America that the climate is more mild to the west of the Alleghany mountains, than under the same parallels in the Atlantic states. Mr. Jefferson has estimated the difference at 3° of latitude; and the gleditsia monosperma, the catalpa, and the aristolochia sypho, and other vegetable productions, are found so many degrees farther to the north, in the basin of the Ohio, than on the coast of the Atlantic. Mr. Volney has endeavoured to explain these phenomena by the frequency of the south-west winds, which drive back the warm air of the Gulf of Mexico towards these regions. A series of good observations, made, for seven years, by Colonel Mansfield at Cincinnati, on the banks of the Ohio, and recently published by Mr. Drake, in an excellent treatise on American Meteorology, has removed the doubts which obscured this point. The thermometrical means prove that the isothermal lines do not rise again in the regions of the west. The quantity of heat which each point of the globe receives under the same parallels is nearly equal on the east and west of the Alleghany range, the winters being only a little milder to the west, and the summers a little warmer. The migrations of vegetables towards the north are favoured, in the basin of the Mississippi, by the form and the direction of the valley which opens from the north to the south. In the Atlantic provinces, on the contrary, the valleys are transverse, and oppose great obstacles to the passage of plants from one valley to another. See my Essai sur la Geographie des Plantes, p. 154. Natural and Statistical View or Picture of Cincinnati and the Miami Country. 1 vol. 8vo. Cincinnati. The following comparison of the mean temperatures has been deduced with great care. I have taken for Philadelphia the means between the observations of Coxe and Rush. I have also referred, for correction, to the observations made by Mr. Legaux at Spring-Mill, upon the Schuylkill, to the north of Philadelphia. As Cincinnati is 512 feet above the level of the sea, its mean temp. is 1·4 too low. Cincinnati. Lat. 36° 6', West long. 84° 24'. Winter, 32·9 Fahr. Spring, 54·1 Summer, 72·9 Autumn, 54·9 Mean, 53·7 Philadelphia. Lat. 39° 56', West long. 75° 16'. Winter, 32·2 Fahr. Spring, 51·4 Summer, 73·9 Autumn, 56·5 Mean, 53·5 If the isothermal lines remain parallel, or nearly so, to the equator, from the Atlantic shores of the New World to the east of the Mississippi and the Missouri, it cannot be doubted that they rise again beyond the Rocky Mountains, on the opposite coast of Asia, between 35° and 55° of latitude. Through 122° 40' of west long. the isothermal line of 50° Fahr. appears to pass almost as in the Atlantic part of the Old World, at 50° of lat. The western coasts of the two worlds resemble one another to a certain point. But these returns of the isothermal lines do not extend beyond 60°. The curve of 32° Fahr. is already found to the south of the Slave Lake, and it comes still farther south in approaching Lakes Superior and Ontario. On account of the influence of west and south-west winds. See Dalton's Meteor. Observ. p. 125. In advancing from Europe towards the east, the isothermal lines again descend, the number of fixed points being few. We can only employ those which are made in places whose known elevation allows us to reduce the mean temperatures to the level of the sea. The few good materials which we possess have enabled us to trace the curves of 32° and 55·4. We know even the nodes of the latter curve round the whole globe. It passes to the N. of Bourdeaux, (lat. 45·46, W. long. 0·37,) near Pekin, (lat. 39·54, E. long. 116·27,) and Cape Foulweather to the S. of the embouchure of the Colombia, (lat. 44·40, W. long. 104°.) Its nodes are distant at least 162° of longitude. We have here pointed out only the empirical laws, under which are ranged the general phenomena and the variations of the temperature which embrace at once a vast extent of the globe. There are partial inflexions of the isothermal lines, which form, so to speak, particular systems modified by small local causes; such as the strange inflexion of the thermometric curves on the shores of the Mediterranean, between Marseilles, Genoa, Lucca, and Rome, and those which determine the difference between the climate of the western coast and the interior of France. These last depend much less on the quantity of heat received by a part of the globe during the whole year, than upon the unequal distribution of heat between winter and summer. It will one day be useful to have, upon particular charts, the partial inflexions of the isothermal lines, which are analogous to the lines of soundings or of equal declivity. The employment of graphical representations will throw much light upon phenomena which are deeply interesting to agriculturists. If, in place of geographical charts, we possessed only tables containing the co-ordinates of latitude, longitude, and altitude, a great number of curious facts relative to the configuration and the superficial inequalities of continents would have remained for ever unknown. In comparing places from the west to the east, and nearly under the same parallel, we find, West. Lat. Mean Temp. St. Malo, 48·39 54·5 Amsterdam, 52·21 53·4 Naples, 40·50 63·3 Copenhagen, 55·41 45·7 Upsal, 59 52 41·9 East. Lat. Mean Temp. Vienna, 48·13 50 5 Warsaw, 52·14 48·6 Pekin, 39·54 54·9 Moscow, 55·46 40·1 Petersburgh, 59·56 38·8 The elevation of Pekin is inconsiderable; that of Moscow is 984 feet. The absolute temperature of Madrid, to the west of Naples, is 59°; but the city is elevated 1978 feet above the level of the sea. Lat. Mean Temp. Bologna, 44·29 56·3 Genoa, 44·25 60·6 Lat. Mean Temp. Marseilles, 43·17 58·8 Rome, 41·53 60·4 We have already found that, towards the north, the isothermal lines are neither parallel to the equator nor to one another; and it is on account of the want of parallelism that we have, in order to simplify such complicated phenomena, traced round the whole globe the curves of equal heat. The position of the line of 32° acts like the magnetic equator, whose inflexions in the South Sea modify the inclinations at great distances. We may even believe that, in the distribution of climates, the line of 32° determines the position of the curve of greatest heat, which is as it were the isothermal equator; and that, in America and Asia, through 78° of west and 102° of east longitude, the torrid zone commences more to the south of the tropic of Cancer, or that it there presents temperatures of less intensity. An attentive examination of the phenomena proves that this is not the case. Whenever we approach the torrid zone below the parallel of 30°, the isothermal lines become more and more parallel to one another and to the earth's equator. The great colds of Canada and Siberia do not extend their action to the equatorial plains. If we have long regarded the Old World as warmer between the tropics than the New World, it is, 1st, because, till 1760, travellers used thermometers of spirit of wine, coloured, and affected by light; 2d, because they observed it either under the reflection of a wall or too near the ground, and when the atmosphere was filled with sand; and 3d, because, in place of calculating the true mean, they used only the thermometric maximum and minimum. Good observations give, Old World. Lat. Mean Temp. Senegambia, 15· 0 79·07 Madras, 13· 5 80·42 Batavia, 6·12 80·42 Manilla, 14·36 78·08 New World. Lat. Mean Temp. Cumana, 10·27 81·86 Antilles, 17· 0 81.05 Vera Cruz, 19·11 78·08 Havannah, 23·10 78·08 The mean temperature of the equator cannot be fixed beyond 811/2°. Kirwan values it at 84°; but only two places of the earth were known, viz. Chandernagor and Pondicherry, to which old travellers attributed annual temperatures above 811/2°. At Chandernagor, in latitude 21·6, the mean temperature, according to Cotte, is 91·9; but the Jesuit Boudier marked only the days when the thermometer was above 98·6 and below 57·2: and at Pondicherry, in latitude 11·55, the mean temperature, according to Cotte, is 85·3, and according to Kirwan, 88°; but M. de Cossigny observed with a spirit-of-wine thermometer. The distribution of heat over different parts of the year differs, not only according to the decrease of the mean annual temperatures, but also in the same isothermal line. It is this unequal division of the heat which characterizes the two systems of climate of Europe and Atlantic America. Under the torrid zone, a small number of months are warmer in the Old World than in the New. At Madras, for example, according to Dr. Roxburgh, the mean temperature of June is 89·4; at Abusheer, 93·2; but at Cumana I have found it only 84·6. With respect to the temperate zone, it has long been known that, from the parallel of the Canary Isles to the Polar Circle, the severity of the winter augments in a progression much more rapid than the summers diminish in heat. It is also known that the climate of the islands and the coasts differs from that of the interior of continents, the former being characterized by mild winters and less temperate summers. But it is the heat of summer particularly which affects the formation of the amylaceous and saccharine matter in fruits, and the choice of the plants that ought to be cultivated. As the principal object of this memoir is to fix, after good observations, the numerical relations between the unequal quantities of heat distributed over the globe, we shall now compare the mean temperatures of three months of winter and summer under different latitudes, and show how the inflections of the isothermal lines modify these relations. In following the curves of equal heat from west to east, from the basin of the Mississippi to the eastern coasts of Asia, through an extent of 4000 leagues, we are struck with the great regularity which appears in the variations of the winter temperature. I. Differences of the Seasons from the Equator to the Polar Circle. Cisatlantic Region. Long. 1° W. and 17° E. Transatlantic Region. Long. 58°--72° W. Isothermal Lines of Mean Temperature. Mean Temperature. Winter. Summer. Diff. Winter. Summer. Diff. 68° 59·0 80·6 21·6 53·6 80·6 27·0 59 44·6 73·4 28·8 39·2 78·8 39·6 50 35·6 68·0 32·4 30·2 71·6 41·4 41 24·8 60·8 36·0 14·0 66·2 52·2 32 14·0 53·6 89·6 1·4 55·4 54·0 This table shows the increase of the difference between the winters and summers from 28° and 30° to the parallels of 55° and 65°. The increase is more rapid in the Transatlantic zone, where the isothermal lines of 32° and 50° approach one another very much; but it is remarkable that, in the two zones which form the two systems of different climates, the division of the annual temperature between winter and summer is made in such a manner that, upon the isothermal line of 32°, the difference of the two seasons is almost double of that which is observed on the isothermal line of 68°. Cisatlantic Region. Long. 31° E. and 22° W. Places. Latitude. Mean Temperature. Whole Year. Winter. Summer. (Pondicherry) 11·55 85·3 77·0 90·5 Cairo, 30·02 72·7 57·7 84·7 Funchal, 32·37 68·7 63·9 72·5 Rome, 41·53 60·4 45·9 55·2 Bourdeaux, 44·50 56·5 42·1 70·7 Paris, 48·50 51·4 38·3 66·2 Copenhagen, 55·41 45·7 30·7 64·6 Stockholm, 59·20 42·3 25·5 61·9 Drontheim, 63·24 39·9 24·7 61·3 Umeo, 63·50 33·3 12·9 54·9 Transatlantic Region. Long. 69° E. and 99° W. Places. Latitude. Mean Temperature. Whole Year. Winter. Summer. Cumana, 10·27 81·9 81·7 83·7 Havannah, 23·10 78·1 71·2 83·3 Natchez, 31·28 64·8 48·6 79·2 Cincinnati, 39·06 53·6 32·9 72·9 Philadelphia, 39·56 54·9 32·2 73·9 New York, 40·40 53·8 29·8 79·2 Cambridge, 42·25 50·4 34·0 70·5 Quebec, 46·47 41·9 14·2 68·0 Nain, 57·10 26·4 0·6 48·4 Fort Churchill, 59·02 25·3 6·8 52·2 If, instead of the temperatures of the seasons, we consider, I do not say the days of the maxima and minima of the year, which are the ordinates of the concave and convex summits of the entire curve, but the mean temperatures of the warmest and coldest month, the increase of the differences becomes still more perceptible. We request the reader to compare, in the following table, only the places which belong to regions bounded by the same meridians, and consequently to the same system of climate; as, for example, to the region of Eastern America to that of Western Europe and that of Eastern Asia. We must also attend to the changes of temperature produced by the monsoons in a part of the equinoctial regions, and distinguish, under the temperature zone, between the climate of the interior, or the continental climate, and that of islands and coasts. Places. Lat. Mean Temperature. Difference. Observations. Coldest Month. Warmest Month. Cumana, 10·27 80·1 84·4 4·3 Uninterrupted trade-winds. Pondicherry, 11·55 76·1 91·4 15·3 Monsoons. Radiat. of sands. Manilla, 14·36 68·0 86·9 18·9 Monsoons. Vera Cruz, 19·11 70·0 81·7 11·7 North winds in winter. Cape Francais, 19·46 77·0 86·0 9·0 uninterrupted trade-winds. Havannah, 23·10 70·0 83·8 13·8 North winds in winter. Funchal, 32·37 64·0 75·6 11·6 Insular climate. Natchez, 31·28 46·9 78·8 31·9 Transatlantic region. Interior. Cincinnati, 39·6 29·6 74·4 44·8 Same system of climate. Pekin, 39·54 24·8 84·2 59·4 Region of Eastern Asia. Philadelphia, 39·56 29·8 77·0 47·2 Transatl. region. Eastern coasts. New York, 40·40 25·3 80·8 55·5 Idem. Rome, 41·53 42·1 77·0 34·9 Cisatlantic region. Milan, 45·28 33·8 55·2 21·4 interior land. Buda, 47·29 27·7 71·6 43·9 Idem. Paris, 48·50 35·1 69·8 34·7 Nearer the western coast. Quebec, 46·47 14·0 73·4 59·4 Transatl. region. Eastern coast. Dublin, 53·21 37·6 60·3 22·7 Region of the West of Europe. Insular climate. Edinburgh, 55·58 38·3 59·4 21·1 Idem. Warsaw, 52·14 27·1 70·3 43·2 Interior land. Petersburg, 59·56 8·6 65·7 57·1 East of Europe. North Cape, 71·0 22·1 46·6 24·5 Climate of coasts and islands. We may conclude, in general, that, for any given place in the curves which express the annual temperatures, the ordinates of the concave and convex summits differ the more from one another, as the temperatures diminish. In the New World, under 40° of latitude, we find a greater difference between the warmest and coldest months of the year than in the Old World, at Copenhagen and Stockholm, under 56°- 59° of latitude. At Philadelphia, the thermometer descends to 50° or 59° below the freezing-point, while, under the same parallel in Europe, it descends scarcely 30·6 below it. On high mountains in islands of little extent, and along the shores, the lines of annual temperature take nearly the same form as in warm climates, having only a less degree of curvature. The difference between the seasons, too, becomes smaller. At the North Cape, in 71° of latitude, and in the isothermal line of 32°, it is almost 11° greater than at Paris, in 49° of latitude, and in the isothermal line of 50°. The sea-breezes and the fogs which render the winters so temperate, diminish at the same time the heats of summer. The characteristic of any climate is not the difference between the winters, expressed in degrees of the thermometer: it is this difference, compared with the absolute quantities indicated by the mean temperature of the seasons. Leopold von Buch's Travels in Lapland, tom. ii. II. Difference between the Winters and Summers, in following the same Isothermal Line from West to East. The differences between the seasons of the year are less great near the convex summits of the isothermal curves, where these curves rise again towards the North Pole, than near the concave summits. The same causes which affect the inflexion or the greatest curvature of the isothermal lines, tend also to equalize the temperatures of the seasons. The whole of Europe, compared with the eastern parts of America and Asia, has an insular climate; and, upon the same isothermal line, the summers become warmer and the winters colder, in proportion as we advance from the meridian of Mont Blanc towards the east or the west. Europe may be considered as the western prolongation of the old continent; and the western parts of all continents are not only warmer at equal latitudes than the eastern parts, but, even in the zones of equal annual temperature, the winters are more rigorous and the summers hotter on the eastern coasts than upon the western coasts of the two continents. The mean temperature of the year being equal to the fourth part of the winter, spring, summer, and autumnal temperatures, we shall have upon the same isothermal line of 53·6 (12° cent.) At the concave summit in America, 74·40 West long. [Formel] At the convex summit in Europe, 2·20 West long. [Formel] At the concave summit in Asia, 116·20 East long. [Formel] This analogy between the eastern coasts of Asia and America sufficiently proves that the inequalities of the seasons, of which we have endeavoured to fix the numerical relations, depend on the prolongation and enlargement of continents towards the pole; of the size of seas in relation to their coasts; and on the frequency of the N.W. winds, which are the vents de Remous of the temperate zone; and not on the proximity of some plateau, or elevation of the adjacent lands. The great plateaus of Asia do not stretch beyond 52° of latitude: and, in the interior of the New Continent, all the immense basin bounded by the Alleghany range and the rocky mountains, and covered with secondary formations, is not more than from 656 to 920 feet above the level of the ocean, according to the levels taken in Kentucky, on the banks of the Monongahela, at Lake Erie. Drake's Nat. and Statist. View of Cincinnati, p. 63. The following table indicates, for all the habitable temperate zone, the division of the same quantity of annual heat between the two seasons of winter and summer. The numbers which it contains are either the result of direct observations, or of interpolations between a great number of observations made in neighbouring places and situated under the same meridian. We have followed each isothermal curve from west to east, giving the preference to places situated near the summits of the curve, as presenting at the same time the greatest differences in the distribution of the annual heat. The longitudes are reckoned from the observatory of Greenwich. Isothermal Lines from 32° to 68°. Long. Lat. Mean Temperature. Winter. Summer. Isoth. Line of 68°. 82·10 W. 29·30 Florida, 53·6 80·6 16·56 W. 32·37 Madeira, 63·5 72·0 3· 0 E. 36·48 North Africa, 59·0 80·6 Isoth. Line of 63·5. 89·40 W. 32·30 Mississippi, 46·4 77·0 14·11 E. 40·50 Italy, 50·0 77·0 Isoth. Line of 59°. 84·10 W. 35·30 Basin of the Ohio, 39·2 77·9 3°--4°E. 43·30 Middle of France, 44·6 75·2 Isoth. Line of 54·5. 84·40 W. 38·30 America, West of Alleghany, 34·7 75·2 74·10 W. 40· 0 America, East of ditto, 32·5 77·0 1·32 W. 47·10 West of France, 39·0 68·0 9·20 E. 45·30 Lombardy, 34·7 73·4 116·20 E. 40· 0 East of Asia, 26·6 82·4 Isoth. Line of 50°. 84·20 W. 41·20 America, West of Alleghany, 31·1 71·6 71·10 W. 42·30 America, East of ditto, 30·2 73·4 6·40 W. 52·30 Ireland, 39·2 59·5 0·40 W. 53·30 England, 37·4 62·6 2·20 E. 51· 0 Belgium, 36·5 63·5 19· 0 E. 47·30 Hungary, 31·1 69·8 116·20 E. 40· 0 Eastern Asia, 23·0 78·8 Isoth. Line of 45 5. 71· 0 W. 44·42 America, East of Alleghany, 23·9 71·6 2·10 W. 57· 0 Scotland, 36·1 56·5 12·35 E. 55·40 Denmark, 30·3 62·6 21·20 E. 53· 5 Poland, 28·0 66·2 Isoth. Line of 41°. 71·10 W. 47· 0 Canada, 14·0 68·0 9·20 E. 62·45 West of Norway, 24·8 62·6 17·20 E. 60·30 Sweden, 24·8 60·8 24·20 E. 60· 0 Finland, 23·0 63·5 36·20 E. 58·30 Central Russia, 22·1 68·0 Isoth. Line of 36·5. 71·40 W. 50· 0 Canada, 6·8 60·8 18· 5 E. 62·30 West coast of Gulf of Bothnia, 17·6 57·2 22·20 E. 62·50 East coast of ditto, 16·5 59·0 Isoth. Line of 32°. 57·40 E. 53· 0 Labrador, 3·2 51·8 19·50 E. 65· 0 Sweden, 11·3 53·6 25·20 E. 71· 0 North Extremity of Norway, 23·9 43·7 The following table shows the oscillations, or the maxima and minima, observed in the division of the heat between the seasons. I have added the means of the winters and summers found at different degrees of longitude, and under the same isothermal line. Isoth. Lines. Degrees of Long. examined. Oscillations observed in the Means. Means calculated. Winters. Summers. Winters. Summers. 32° 83 3·2 to 24·8 51·8 to 53·6 14·0 52·7 41 107 14·0 24·8 62·6 68·0 19·4 65·3 50 200 23·0 37·4 62·6 78·8 30·2 70·7 59 87 39·0 44·6 75·2 77·0 41·9 75·2 68 84 53·6 59·0 71·6 80·6 56·3 77·9 The deviations round the mean,--that is, the inequality of the winters on the same isothermal line,-- increase in proportion as the annual heat diminishes, from Algiers to Holland, and from Florida to Pennsylvania. The winters of the curve of 68° are not found upon that of 51°, and the winters of 51° are not met with on the curve of 42°. In considering separately what may be called the same system of climate, --for example, the European region, the Transatlantic region, or that of Eastern Asia,--the limits of the variations become still more narrow. Wherever, in Europe, in 40° of longitude, the mean temperature rises To 59·0 The winters are from 44·6 to 46·4 and the summers from 73·0 to 75·2 54·5 36·5 41·0 68·0 73·0 50·0 31·1 37·4 62·6 69·8 45·5 28·4 36·1 57·2 68·0 41·0 20·3 26·8 55·4 66·2 In tracing five isothermal lines between the parallels of Rome and Petersburg, the coldest winter presented by one of these lines is not found again on the preceding line. In this part of the globe, those places whose annual temperature is 54·5, have not a winter below 32°, which is already felt upon the isothermal line of 50°. If, in place of stopping at the most rigorous winter which each curve presents, we trace the lines of equal winter temperature, (or the isocheimal lines,) these lines, instead of coinciding with the lines of equal annual heat, oscillate round them. As the isocheimal lines unite points placed on different isothermal lines, we may examine to what distance their summits extend. In considering always the same system of climates, --for example, the European region,--we shall find that the lines of equal winter cut isothermal lines which are nine degrees distant. In Belgium, (in latitude 52°, and in isothermal latitude 51·8,) and even in Scotland, (in latitude 57°, and isothermal latitude 45·5,) the winters are more mild than at Milan, (in latitude 45·28,) and in isothermal latitude 55·8,) and in a great part of Lombardy. Farther to the north, in the Scandinavian peninsula, we meet with three very different systems of climate:--viz. 1, the region of the west coasts of Norway to the west of the mountains; 2, the region of the eastern coasts of Sweden to the east of the mountains; and, 3, the region of the west coasts of Finland, along the gulf of Bothnia. No where, without the tropics, is the division of the annual heat among the seasons more equal. In the temperate zone, under parallels nearer to our own, Ireland presents an example still more striking of the union of very mild winters with cold and moist summers. Notwithstanding a difference of four degrees of latitude, the winters there are as mild as in Britain, while the mean temperature of the summers is three degrees less. This is the true maritime climate. The month of August, which on the same isothermal line, in the east of Europe, in Hungary, has the temperature of 71·6, reaches only 60·8 at Dublin. The month of January, whose mean temperature at Milan, and in a great part of Lombardy, is only 35·6, rises in Ireland to 5·4 and 7·2. On the coasts of Glenarm, also, (in north lat. 54·56,) under the parallel of Konigsberg, the myrtle vegetates with the same strength as in Portugal. It scarcely freezes there in winter, but the heat of summer is not capable of ripening the vine. Wahlenberg Flora Carpath. p. 90. Irish Transactions, tom. viii. p. 116, 203, 269. These examples are sufficient to prove that the isocheimal lines deviate much more than the isothermal lines from the terrestrial parallels. In the system of European climates, the latitudes of two places that have the same annual temperature cannot differ more than from four to five degrees, while two places whose mean winter temperature is the same may differ more than nine or ten degrees in latitude. The farther we advance to the east, the more rapidly do these differences increase. The lines of equal summer, or isotheral curves, follow a direction exactly contrary to the isocheimal lines. We find the same summer temperature at Moscow, in the centre of Russia, and towards the mouth of the Loire, notwithstanding a difference of eleven degrees of latitude. Such is the effect of the radiation of the earth on a vast continent deprived of mountains. It is sufficiently remarkable that the inflexions of the isothermal lines, and the division of lands and seas, are such upon the globe that every where, in North America, in Europe, and in Eastern Asia, the mean temperature of the summers does not denote more than 36° in the parallels of from 45° to 47°. The same causes which, in Canada and in the north of China, sink the curves of equal annual heat, where the isothermal lines (those of 51·8 and 53·6,) corresponding to the parallels of 45° and 47°, tend to raise the lines of equal summer, or the isotheral curves. Continuation of our Abstract of Baron Humboldt's Dissertation on Isothermal Lines, and the Distribution of Heat over the Globe. AFTER what has already been stated respecting the limits between which the annual heat divides itself on the same isothermal curve, it will be seen how far we are authorized to say, that the coffee-tree, the olive, and the vine, in order to be productive, require mean temperatures of 64·4, 60·8, and 53·6, Fahr. These expressions are true only of the same system of climate,--for example, of the part of the Old World which stretches to the west of the meridian of Mont Blanc; because, in a zone of small extent in longitude, while we fix the annual temperatures, we determine also the nature of the summers and the winters. It is known, likewise, that the olive, the vine, the varieties of grain, and the fruit-trees, require entirely different constitutions of the atmosphere. Among our cultivated plants, some, slightly sensible of the rigors of winter, require very warm, but not long, summers; others require summers rather long than warm; while others, again, indifferent to the temperature of summer, cannot resist the great colds of winter. Hence it follows, that, in reference to the culture of useful vegetables, we must discuss three things for each climate: the mean temperature of the entire summer,--that of the warmest month,--and that of the coldest month. I have published the numerical results of this discussion in my Prolegomena de Distributione Geographica Plantarum, secundum Coeli Temperiem; and I shall confine myself at present to the limits of culture of the olive and the vine. The olive is cultivated in our continent between the parallels of 36° and 40°, wherever the annual temperature is from 62·6 to 58·1, where the mean temperature of the coldest month is not below from 41·0 to 42·8, and that of the whole summer from 71·6 to 73·4. In the New World, the division of heat between the seasons is such, that, on the isothermal line of 58·1, the coldest month is 35·6, and that the thermometer sometimes sinks there even, during several days, from 14° to 10·4. The region of potable wines extends, in Europe, between the isothermal lines of 62·6 and 50°, which correspond to the latitudes of 36° and 48°. The cultivation of the vine extends, though with less advantage, even to countries whose annual temperature descends to 48·2 and to 47·48; that of winter to 33·8, and that of summer to 66·2 and 68°. These meteorological conditions are fulfilled in Europe as far as the parallel of 50°, and a little beyond it. In America, they do not exist farther north than 40°. They have begun, indeed, some years ago, to make a very good red wine to the west of Washington, beyond the first chain of mountains, in the valleys which do not extend beyond 38° 54' of lat. On the continent of Western Europe, the winters, whose mean temperature is 32°, do not commence till on the isothermal lines of 48·2 and 50°, in from 51° to 52° of latitude; while, in America, we find them already on the isothermal lines of from 51·8 to 53·6, under from 40° to 41° of latitude. If, instead of considering the natural inflexions of the isothermal lines,--that is to say, those that propagate themselves progressively at great intervals of longitude,--we direct our attention to their partial inflexions, or to particular systems of climates occupying a small extent of country, we shall still find the same variations in the division of the annual heat between the different seasons. These partial inflexions are most remarkable: 1st. In the Crimea, where the climate of Odessa is contrasted with that of the S.W. shores of the Chersonesus, sheltered by mountains, and fit for the cultivation of the olive and the orange-tree. 2dly. Along the Gulf of Genoa, from Toulon and the Hieres Isles to Nice and Bordighera, (Annales du Museum, tom. xi. p. 219,) where the small maritime palm-tree, chamoerops, grows wild, and where the date-tree is cultivated on a large scale, not to obtain its fruit, but the palms or etiolated leaves. 3dly. In England, on the coast of Devonshire, where the port of Salcombe has, on account of its temperate climate, been called the Montpellier of the North, and where (in South Hams) the myrtle, the camellia Japonica, the fuchsia coccinea, and the buddleia globosa, pass the winter in the open ground, and without shelter. Knight, Trans. Hort. Soc. vol. i. p. 32.--In 1774, an agave flowered at Salcombe, after having lived twenty-eight years without being covered in winter. On the coast of England, the winters are so mild, that orange-trees are seen on espaliers, which are sheltered, as at Rome, only by means of a matting. 4thly. In France, on the western coasts of Normandy and Brittany. In the department of Finisterre, the arbutus, the pomegranate-tree, the yucca gloriosa and aliofolia, the erica Mediterranea, the hortensia, the fuchsia, the dahlia, resist in open ground the inclemency of a winter which lasts scarcely fifteen or twenty days, and which succeeds to a summer by no means warm. During this short winter, the thermometer sometimes falls to 17·6. The sap ascends in the trees from the month of February; but it often freezes even in the middle of May. The lavatera arborea is found wild in the isle of Glenans; and opposite to this island, on the continent, the astragalus bajonensis and the laurus nobilis. Bonnemaison, Geogr. Botan. du Depart. du Finisterre, (Journal de Botan. tom. iii. p. 118.) From observations made in Britanny for twelve years, at St. Malo, at Nantes, and at Brest, the mean temperature of the peninsula appears to be above 56·3. In the interior of France, where the land is not much elevated above the sea, we must descend 3° of latitude in order to find an annual temperature like this. It is known, from the researches of Arthur Young, that, in spite of the great rise of the two isothermal lines of 53·6 and 55·4 on the western coast of France, the lines of culture (those of the olive, and of the maize and vine,) have a direction quite opposite, from S.W. to N.E. This phenomenon has been ascribed, with reason, to the low temperature of the summers along the coast; but no attempt has been made to reduce to numerical expressions the ratios between the seasons in the interior and on the coast. In order to do this, I have chosen eight places, some of which lie under the same geographic parallels, and others in the prolongation of the same isothermal line. I have compared the temperatures of winter, of summer, and of the warmest months; for a summer of uniform heat excites less the force of vegetation, than a great heat preceded by a cold season. The terms of comparison have been along the Atlantic; the coasts of Brittany, from St. Malo and St. Brieux to Vannes and Nantes; the sands of Olonne; the Isle of Oleron; the embouchure of the Garonne and Dax, in the department of the Landes: and, in the interior, corresponding to the same parallel, Chalons sur Marne, Paris, Chartres, Troyes, Poitiers, and Montauban. Travels in France, vol. ii. p. 91. The line which limits the cultivation of the vine, extends from the embouchure of the Loire and of the Vilaine, by Pontoise, to the confluence of the Rhine and the Moselle. The line of the olive-trees commences to the west of Narbonne, passes between Orange and Montelimart, and carries itself to the N.E. in the direction of the Great St. Bernard. Decandolle, Flor. Franc. 3d edit. tom. ii. pl. viii. xi. Lequinio, Voy. dans le Jura, tom. ii. p. 84--91. Places in the Interior. Latitude. Mean Temperature of the Year. of Winter. of Summer. of the Warmest Month. Fahr. Fahr. Fahr. Fahr. Chalons sur Marne ........ 48·57 50·5 36·1 66·6 67·5 Paris .................... 48·50 51·1 38·7 65·3 67·5 Chartres ................ 48·26 50·7 37·0 64·6 65·7 Troyes .................. 48·18 52·2 38·3 67·3 68·4 Chinon .................. 47·26 53·4 38·7 69·1 70·2 Poitiera .................. 46·39 54·3 39·7 67·1 69·3 Vienne .................. 45·31 55·0 38·7 71·6 73·4 Montauban .............. 44·01 55·6 42·6 69·3 71·4 Places on the Coast. St. Malo ................ 48·39 55·5 42·4 66·9 67·5 St. Brieux ................ 48·31 52·3 41·7 64·4 67·1 Vannes .................. 47·39 51·8 39·7 64·4 65·8 Nantes .................. 47·13 54·7 40·5 68·5 70·5 La Rochelle ............. 46·14 53·1 40·3 66·6 67·1 Oleron .................. 45·56 58·1 44·6 68·5 72·1 Bourdeaux .............. 44·50 56·5 42·1 70·9 71·4 Dax .................... 43·52 54·1 44·4 67·3 68·9 Farther south, from 441/2° of lat., the comparisons become incorrect, because France, locked between the Ocean and the Mediterranean, presents, along its last basin, in the fine region of the olives, a system of climate of a particular kind, and very different from that of the western coast. These results are deduced from 127,000 observations, made with sixteen thermometers, of, no doubt, unequal accuracy. In supposing, on the theory of probabilities, that, in such a number of observations, the errors, in the construction and exposure of the instruments, and in the hours of observation, will in a great measure destroy one another, we may determine, by interpolation, either under the same parallel or upon the same isothermal line, the mean winters and summers of the interior and of the coast of France. This comparison gives-- Mean Winter. Mean Summer. I. Isothermal Lines of 52°·7 Coast ........ 40·6 65·1 Interior ...... 38·5 68·0 54°·7 Coast ........ 41·4 67·3 Interior ...... 39·2 68·4 Annual Temp. I. Parallels of 47° to 49° Coast .... 41·0 66·7 53·0 Interior .. 37·8 66·6 51·6 45° to 46° Coast .... 42·3 67·8 55·8 Interior .. 39·2 69·3 54·7 As the isothermal lines rise again towards the western coasts of France,--that is to say, as the mean temperature of the year becomes there greater than under the same latitude in the interior of the country,--we ought to expect that, in advancing from east to west under the same parallel, the heat of the summers would not diminish. But the rising again of the isothermal lines, and the proximity of the sea, tend equally to increase the mildness of the winters; and each of these two causes acts in an opposite manner upon the summers. If the division of the heat between these seasons was equal in Brittany and in Orleannois, in the climate of the coast and the continental climates, we ought to find the winters and summers warmer in the same latitude along the coast. In following the same isothermal lines, we readily observe, in the preceding table, that the winters are colder in the interior of the country, and the summers more temperate upon the coasts. These observations confirm, in general, the popular opinion respecting the climate of coasts; but, in recollecting the cultivation and the development of vegetation on the coasts and in the interior of France, we should expect differences of temperature still more considerable. It is surprising that these differences between the winter and the summer should not exceed 1·8, or nearly a quarter of the difference between the mean temperature of the winters or the summers of Montpellier and Paris. In speaking of the limits of the cultivation of plants upon mountains, I shall explain the true cause of this apparent contradiction. In the mean time it may be sufficient to remark, that our meteorological instruments do not indicate the quantity of heat which, in a clear and dry state of the air, the direct light produces in the more or less coloured parenchyma of the leaves and fruits. In the same mean temperature of the atmosphere, the development of vegetation is retarded or accelerated according as the sky is foggy or serene, and according as the surface of the earth receives only a diffuse light during entire weeks, or is struck by the direct rays of the sun. On the state of the atmosphere, and the degree of the extinction of light, depend, in a great measure, those phenomena of vegetable life, the contrasts of which surprise us in islands, in the interior of continents, in plains, and on the summit of mountains. If we neglect these photometrical considerations, and do not appreciate the production of heat in the interior of bodies, and the effect of nocturnal radiation in a clear or a cloudy sky, we shall have some difficulty in discovering, from the numerical ratios of the observed summer and winter temperatures of Paris and London, the causes of the striking difference which appears in France and England in the culture of the vine, the peach, and other fruit-trees. Young's Travels in France, vol. ii. p. 195. When we study the organic life of plants and animals, we must examine all the stimuli or external agents which modify their vital actions. The ratios of the mean temperatures of the months are not sufficient to characterize the climate. Its influence combines the simultaneous action of all physical causes; and it depends on heat, humidity, light, the electrical tension of vapours, and the variable pressure of the atmosphere. It is the last cause which, on the tops of mountains, modifies the perspiration of plants, and even increases the exhaling organs. In making known the empirical laws of the distribution of heat over the globe, as deducible from the thermometrical variations of the air, we are far from considering these laws as the only ones necessary to resolve all the problems of climate. Most of the phenomena of nature present two distinct parts,--one which may be subjected to exact calculation, and another which cannot be reached but through the medium of induction and analogy. Continuation of our Abstract of Baron Humboldt's Dissertation on Isothermal Lines, and the Distribution of Heat over the Globe. HAVING considered the division of heat between winter and summer on the same isothermal line, we shall now point out the numerical ratios between the mean temperature of spring and winter, and between that of the whole year and the warmest month. From the parallel of Rome to that of Stockholm, and consequently between the isothermal lines of 60·8° and 41°, the difference of the months of April and May is every where 10·8 or 12·6, and all the successive months are those which present the most rapid increase of temperature. But, as in northern countries, (in Sweden, for example,) the month of April is only 37·4, the 10·8 or 12·6 which the month of May adds, necessarily produces there a much greater effect on the development of vegetation than in the south of Europe, where the mean temperature of April is from 53·6 to 55·4. It is from an analogous cause that, in passing from the shade to the sum, either in our climates in winter, or between the tropics on the back of the Cordilleras, we are more affected by the difference of temperature than in summer and in the plains; though in both cases the thermometrical difference is the same,--for example, from 5·4 to 7·2. Near the polar circle, the increase of the vernal heat is not only more sensible, but it extends equally to the month of June. At Drontheim, the temperatures of April and May, like those of May and June, differ not 10·8 or 12·6, but 14·4 or 16·2. In calculating for Europe, from 46° to 48° of lat. for ten years, the mean temperatures of every ten days, we find that the decades which succeed one another differ near the summits of the annual curve only 1·41, while the differences rise in antumn from 3·6 to 5·4, and in spring from 5·4 to 7·2. In distinguishing upon the same isothermal line the places which approach its concave or convex summits, in the same system of climates in the northern and southern regions, we shall find,-- 1st. That the increase of the vernal temperature is great, (from 14·4 or 16·2, in the space of a month,) and equally prolonged, whereever the division of the annual heat between the seasons is very unequal, as in the north of Europe, and in the temperate part of the United States. 2dly. That the vernal increase is great, (at least above 9° or 10·8,) but little prolonged, in the temperate part of Europe. 3dly. That the increase of the vernal temperature is small, (scarcely 7·2,) and equally prolonged, wherever there is an insular climate. 4thly. That in every system of climates, in the zones contained between the same meridians, the vernal increase is smaller, and less equally prolonged, in low than in high latitudes. The isothermal zone from 53·6 to 55·4 may serve as an example for confirming these different modifications of spring. In Eastern Asia, near the concave summit, the differences of temperature between the four months of March, April, May, and June, are very great, and very equal, (15·7, 13·3, and 13·9.) In advancing westward towards Europe, the isothermal line rises again; and, in the interior of the country, near the convex summit, the increase is still greater, but little prolonged: that is to say, that; of the four months which succeed one another, there are only two whose difference rises to 13°; they are 9·4, 13·3, 4·1. Farther west, on the coasts, the differences become small and equal,--viz. 3·6, 6·5, 5·6. In crossing the Atlantic, we approach the western concave summit of the isothermal line of 53·6. The increase of vernal temperature shows itself anew, and almost as great and as much prolonged, as near the Arctic concave summit. The differences of the four months are 10·4, 13·9, and 10·8. In the curve of annual temperature, the spring and autumn mark the transitions from the minimum and the maximum. The increments are naturally slower near the summits than in the intermediate part of the curve. Here they are greater, and of longer continuance, in proportion to the difference of the extreme ordinates. The autumnal decrease of temperature is less rapid than the vernal increase, because the surface of the earth acquires the maximum of heat slower than the atmosphere; and because, in spite of the serenity of the air which prevails in autumn, the earth loses slowly, by radiation, the heat which it has acquired. The following Table will show how uniform the laws are which I have just established: Names of Places. Lat. March. April. May. June. Differences of Temperature of the Four Months. Mean Temp. of the Year. I. Group,--Concave Summits in America. Natchez .............. 31·28° 57·9° 66·2° 72·7° 79·5° 8·3° 6·1° 7·2° 64·8° Williamsburg ......... 37·18 46·4 61·2 66·6 77·7 14·8 5·4 11·2 58·1 Cincinnati .............. 39· 0 43·7 57·4 61·2 70·9 13·7 3·6 9·7 53·8 Philadelphia ............ 39·56 44·1 53·6 62·1 72·3 9·5 8·5 10·3 53·6 New York ............ 40·40 38·7 49·1 65·8 80·2 10·4 16·7 14·4 53·8 Cambridge ............ 42·25 34·5 45·5 56·8 70·2 11·0 11·3 13·3 50·4 Quebec ................ 46·47 23·0 39·6 54·7 63·9 16·6 15·1 41·2 41·7 Nain .................. 57· 0 6·8 27·5 37·0 43·3 20·7 9·5 8·1 26·4 II. Group,--Convex Summits in Europe. 1. Continental Climate: Rome .................. 41·53 50·4 55·4 66·9 72·3 5·0 11·5 5·4 60·4 Milan .................. 45·28 47·8 51·1 65·1 70·5 7·7 9·5 5·4 55·8 Geneva ................ 46·12 39·6 45·5 58·1 62·2 6·1 12·4 4·1 49·3 Buda .................. 47·29 38·3 49·1 64·8 68·4 10·8 15·7 3·6 51·1 Paris .................. 48·50 42·3 48·2 60·1 64·4 8·5 11·9 4·3 51·1 Gottingen .............. 51·32 34·2 44·2 57·7 62·2 10·1 13·5 4·5 46·9 Upsal .................. 59·51 29·5 39·7 48·7 57·9 10·3 9·0 9·2 41·9 Petersburg ............ 59·56 27·5 37·0 50·2 59·4 9·5 13·1 9·2 38·8 Umeo .................. 63·50 23·0 34·2 43·7 55·0 11·2 9·5 11·3 33·3 Uleo .................. 65· 0 14·0 26·2 41·0 55·0 12·2 14·8 14·0 33·1 Enontekies ............ 68·30 11·5 26·6 36·5 49·5 15·1 9·9 13·0 27·0 2. Climate of the Coast: Nantes ................ 47·13 50·0 53·6 60·1 65·7 3·6 6·5 5·6 54·7 London ................ 51·30 44·2 49·8 56·5 63·1 5·6 6·7 6·7 51·6 Dublin ................ 53·21 41·9 45·3 51·8 55·6 3·4 6·5 4·0 48·4 Edinburgh .............. 55·57 41·4 47·3 50·5 57·2 5·8 3·2 6·7 47·8 North Cape ............ 71· 0 25·0 30·0 34·0 40·1 5·2 4·0 6·1 32·0 III. Group,--Concave Summit of Asia. Pekin .................. 59·54 41·4 57·0 70·3 84·2 15·7 13·3 13·9 54·9 In all places whose mean temperature is below 62·6, the revival of nature takes place in spring, in that month whose mean temperature reaches 42·8 or 46·4. When a month rises to, 41·9, the peach-tree (amygdalus persica,) flowers, 46·8, the plum-tree (prunus domestica,) flowers, 51·8, the birch-tree (betula alba,) pushes out its leaves. Cotte, Meteorologie, p. 448.--Wahlenberg, Flor. Lap. pl. 51. At Rome, it is the month of March, at Paris the beginning of May, and at Upsal the beginning of June, that reaches the mean temperature of 51·8. Near the hospice of St. Gothard, the birch cannot vegetate, as the warmest month of the year there scarcely reaches 46·5. Barley, in order to be cultivated advantageously, requires, during ninety days, a mean temperature of from 47·3 to 48·2. By adding the mean temperatures of the months above 51·8,--that is, the temperatures of those in which trees vegetate that lose their foliage,-- we shall have a sufficiently exact mean of the strength and continuance of vegetation. As we advance towards the north, vegetable life is confined to a shorter interval. In the south of France there are 270 days of the year in which the mean temperature exceeds 51·8; that is to say, the temperature which the birch requires to put forth its first leaves. At St. Petersburg, the number of these days is only 120. These two cycles of vegetation, so unequal, have a mean temperature which does not differ more than 5·4; and even this want of heat is compensated by the effects of the direct light, which acts on the parenchyma of plants in proportion to the length of the days. If we compare, in the following Table, Eastern Asia, Europe, and America, we shall discover, by the increase of heat during the cycle of vegetation, the points where the isothermal lines have their concave summits. The exact knowledge of these cycles will throw more light on the problem of agricultural geography, than the examination of the single temperatures of summer. Playfair, Edin. Trans. vol. v. p. 202.--Wahlenberg in Gilbert's Annalen, tom. xli. p. 282. In the system of European climates, from Rome to Upsal, between the isothermal lines of 59° and 41°, the warmest month adds from 16·2 to 18° to the mean temperature of the year. Farther north, and also in Eastern Asia and in America, where the isothermal lines bend towards the equator, the increments are still more considerable. Lines of Equal Heat. Names of Places. Latitude. Mean Temp. of the Year. Sum of the Mean Temp. of the Months that reach 51·8. Number of these Months. Mean Temp. of the Days which reach 51·8. Mean Temp. of the Warmest Months. Observations. Isothermal line of 59·0, Rome, 41·53 60·4 585° 9 64·8 77·0 Basin of the Mediterranean. Nismes, 43·50 60·3 593 9 65·8 78·3 Idem. Isothermal line of 53·6, Pekin, 39·54 54·9 499 7 71·2 84·2 Eastern concave summit. Poitiers, 46·34 54·3 426 7 60·8 69·3 Convex summit. Nantes, 47·13 54·7 438 7 62·6 69·8 Idem, coasts. St. Malo, 48·39 53·8 431 7 61·5 68·4 Idem. Philadelphia, 39·56 53·4 463 7 66·2 77·0 Western concave summit. Cincinnati, 39·6 53·8 458 7 65·5 74·3 Idem. Isothermal line of 50·0, London, 51·30 51·8 364 6 60·6 66·6 Insular climate. Paris, 48·50 51·1 381 6 63·5 69·8 Near the coasts. Buda, 47·29 51·1 323 5 64·6 72·0 Interior. Isothermal line of 48·2, Geneva, 46·12 49·3 311 5 62·2 66·6 Interior. Dublin, 53·21 48·7 282 5 56·5 60·8 Climate of the coasts. Edinburgh, 55·57 47·8 279 5 55·8 59·4 Idem. Isothermal line of 41·0, Upsal, 59·51 41·9 229 4 57·2 61·9 Convex summit. Quebec, 46·47 41·7 318 5 63·7 73·4 Western concave summit. Isothermal line of 32·0, Petersburg, 59·56 38·8 236· 4 59·0 65·7 East of Europe. Umeo, 53·50 33·3 118 2 59·0 62·6 E. Coast of Gulf of Bothnia. North Cape, 71·0 32·0 0 0 0 46·6 Interior climate. Enontekies, 68·30 27·0 116 2 58·1 59·5 Continental climate. As two hours of the day indicate the temperature of the whole day, there must also be two days of the year, or two decades, whose mean temperature is equal to that of the whole year. From the mean of ten observations, this temperature of the year is found at Buda in Hungary, from the 15th to the 20th of April, and from the 18th to the 23d of October. The ordinates of the other decades may be regarded as functions of the mean ordinates. In considering the temperatures of entire months, we find that, to the isothermal line of 35.6, the temperature of the month of October coincides (generally within a degree,) with that of the year. The following Table proves that it is not the month of April, as Kirwan affirms, (Estimate, &c. p. 166,) that approaches nearest to the annual temperature. Names of Places. Mean Temperature of the Year. of October. of April. Cairo ........ 72.3 72.3 77.9 Algiers ...... 69.8 72.1 62.6 Natchez ...... 65.0 68.4 66.4 Rome ........ 60.4 62.1 55.4 Milan ........ 55.8 58.1 55.6 Clincinnati .... 53.6 54.9 56·8 Philadelphia .. 53.4 54.0 53·6 New York .... 53.8 54.5 49·1 Pekin ........ 54.7 55.4 57.0 Buda ........ 51.1 52.3 49.1 London ...... 51.8 52.3 49.8 Paris ........ 51.1 51.3 48.2 Geneva ...... 49.3 49.3 45.7 Dublin ...... 48.6 48.7 45.3 Edinburgh .... 47.8 48.2 46.9 Names of Places. Mean Temperature of the Year. of October. of April. Gottingen .... 46.9 47.1 44.4 Franeker ...... 52.3 54.9 50.0 Copenhagen .. 45.7 48.7 41.0 Stockholm .... 42.3 42.4 38.5 Christiania .... 42.6 39.2 42.6 Upsal ........ 41.7 43.3 39.7 Quebec ...... 41.9 42.8 39.6 Petersburg .... 38.8 39.0 37.0 Abo .......... 41.4 42.0 40.8 Drontheim .... 39.9 39.2 34.3 Uelo .......... 33.1 37.9 34.2 Umeo ........ 33.3 37.8 34.0 North Cape .. 32.0 32.0 30.2 Enontekies .... 27.0 27.5 26.6 Nain .......... 26.4 33.1 27.5 As travellers are seldom able to make observations for giving immediately the temperature of the whole year, it is useful to know the constant ratios which exist in each system of climates, between the vernal and autumnal temperatures, and the annual temperature. The quantity of heat which any point of the globe receives, is much more equal during a long series of years than we would be led to believe from the testimony of our sensations, and the variable product of our harvests. In a given place, the number of days, during which the N.E. or S.W. winds blow, preserve a very constant ratio, because the direction and the force of these winds, which bring warmer or colder air, depend upon general causes, --on the declination of the sun,--on the configuration of the coast,--and on the lie of the neighbouring continent. It is less frequently a diminution in the mean temperature than an extraordinary change in the division of the heat between the different months, which occasions bad harvests. By examining, between the parallels of 47° and 49°, a series of good meteorological observations, made during ten or twelve years, it appears that the annual temperatures vary only from 1.8 to 2.7: those of winter, from 3.6 to 5.4; those of the months of winter, from 9°. to 10. 8. At Geneva, the mean temperatures of twenty years were as follow:-- Years. Mean Temp. 1796, 49.3° 1797, 50.5 1798, 50.0 1799, 48·7 1800, 50.5 1801, 51.1 1802, 50.9 1803, 50.4 1804, 51.1 1805, 47.8 Years. Mean Temp. 1806, 51.4° 1807, 49.3 1808, 46·9 1809, 48·9 1810, 51·1 1811, 51·6 1812, 47.8 1813, 48.6 1814, 48.2 1815, 50.0 Mean of twenty years,.... 49.67° If, in our climates, the thermometrical oscillations are a sixth part of the annual temperature, they do not amount to one twenty-fifth part under the tropics. I have computed the thermometrical variations, during eleven years, at Paris, for the whole year, the winter, the summer, the coldest month, the warmest month, and the month which represents most accurately the annual mean temperature; and the following are the results which I obtained: Observations of M. Bouvard. Mean Temperature of the Year. of Winter. of Summer. of January. of August. of October. Paris, 1803........ 51·1° 36.7° 67.6° 34.3° 67.6° 50.5° 1804........ 52.0 41.0 65.5 43.9 64.6 52.7 1805........ 49.5 36.0 63.1 34.9 64.8 49.3 1806........ 53.4 40.6 65.3 43.0 64.6 51.8 1807........ 51.4 42.3 67.8 36.1 70.5 54.3 1808........ 50.5 36.7 66.2 36.3 66.6 48.2 1809........ 50.9 40.5 62.4 40.8 64.2 49.6 1810 ....... 50.9 36.5 63.3 30.6 63.7 52.9 1811........ 52.7 39.2 63.1 26.6 63.7 57.6 1812........ 49.8 39.6 63.1 34.7 64.2 51.1 1813........ 49.8 36.1 61.7 32.5 62.6 53.1 Mean of these 11 years, 51.1 38.7 64.0 36.6 65.1 51.9 At Geneva, the mean temperatures of the summers were, from 1803 to 1809,-- Years. Mean Temp. of. Summers. 1803 -- 67·3° 1804 -- 65.0 1805 -- 62.2 1806 -- 65.7 1807 -- 68.2 1808 -- 62.9 1809 -- 63.0 Mean of seven years, 64.9 M. Arago has found that, in the two years 1815 and 1816, the last of which was so destructive to the crops in a great part of France, the difference of the mean annual temperature was only 2°, and that of the summer 3.2°. The summer of 1816, at Paris, was 59.9°,--4.7° below the mean of the former. From 1803 to 1813, the oscillations round the mean did not go beyond --2.9°, and +3.4°. In comparing places which belong to the same system of climates, though more than eighty leagues distant, the variations seem to be very uniform, both in the annual temperature and that of the seasons, although the thermometrical quantities are not the same. Years. Paris. Geneva. Paris. Geneva. Paris. Geneva. Mean Annual Temperature. Difference between mean Ann. Temp. and that for 12 years, 51.1. Mean Annual Temperature. Difference between mean Ann. Temp. and that for 12 years, 49.6. Mean Temperature of Winter. Difference with the mean Winter Temp. of 12 years, 38.7. Mean Temperature of Winter. Difference with the mean Winter Temp. of 12 years, 34.9. Mean Temperature of Summer. Difference with the mean Temperature of Summer for 12 years, 64.6. Mean Temperature of Summer. Difference with the Mean Temperature of Summer for 12 years, 64.9. 1803 51.1° 0° 50.4° + 0.8° 36.7° -- 2.0° 32.2° -- 2.7° 67.6° + 3.0 67.6° + 2.7° 1804 52.0 + 0.9 51.1 + 1.5 41.0 + 2.3 38.3 + 3.4 65.5 + 0.9 66.2 + 1.3 1805 49.5 -- 1.6 47.8 -- 1.8 36.0 -- 2.7 33.8 -- 1.1 63.1 -- 1.5 63.0 -- 1.9 1806 53.4 + 2.3 51.4 + 1.8 40.6 + 1.9 38.5 + 3.6 65.3 + 0.7 64.6 -- 0.3 1807 51.4 + 0.3 49.3 -- 0.3 42.3 + 3.6 35.8 + 0.9 67.8 + 3.2 68.2 + 3.3 1808 50.5 -- 0.6 46.8 -- 2.8 36.7 -- 2.0 33.8 -- 1.1 66.2 + 1.6 65.5 -- 1.4 1809 50.9 -- 0.2 48.7 -- 0.9 40.5 + 1.8 35.1 + 0.2 62.4 -- 2.2 65.1 -- 1.8 1810 50.9 -- 0.2 51.1 + 1.5 36.5 -- 2.2 63.3 -- 1.3 1811 52.7 + 1.6 51.8 + 2.2 39.2 + 0.5 65.1 + 0.5 1812 49.8 -- 1.3 47.8 -- 1.8 39.6 + 0.9 63.1 -- 1.5 1813 49.8 -- 1.3 48.6 + 1.0 36.1 -- 2.6 61.7 -- 2.9 Continuation of our Abstract of Baron Humboldt's Dissertation on Isothermal Lines, and the Distribution of Heat over the Globe. ALL the ratios of temperature which we have hitherto fixed belong to that part of the lower strata of the atmosphere which rests on the solid surface of the globe in the northern hemisphere. It now remains for us to discuss the temperature of the southern hemisphere. The southern hemisphere receives the same quantity of light; but the accumulation of heat in it is less, on account of the emission of the radiant heat which takes place during a long winter. This hemisphere being also in a great measure covered with water, the pyramidal extremities of the continents have there an irregular climate. Summers of a very low temperature are succeeded, as far as 50° of south latitude, by winters far from rigorous. The small quantity of land in the southern hemispheres, contributes not only to equalize the seasons, but also to diminish absolutely the annual temperature of that part of the globe. There is reason to believe that this want of dry land would produce an effect still more sensible, if the division of the continents was as unequal in the equinoctial as in the temperate zones. The dry lands in the two hemispheres are in the ratio of 3 to 1. The dry lands between the tropics are in the two hemispheres as 5 to 4, and without the tropics as 13 to 1. Theory and experience prove that the difference of temperature between the two hemispheres cannot be great near the limit which separates them. The differences of the two hemispheres become more sensible in the warmest months. Rio Janeiro. Mean Temp. Havannah. Mean Temp. June .......... 68.6° December ...... 71.8° July .......... 70.2 January ........ 70.2 January ...... 79.2 July ............ 83.3 February ...... 80.6 August ......... 83.8 The division of the heat between the different parts of the year gives a particular character to southern climates. In the southern hemisphere, on the isothermal lines of 46.4 and 50.0, we find summers which in our hemisphere belong only to the isothermal lines of 35.6 and 40°. The mean temperature is not precisely known beyond 51° of S. latitude. Navigators do not frequent those regions when the sun is in the northern signs, and it would be wrong to judge of the rigour of winter from the low temperature of the summer. The inequal temperature of the two hemispheres, which is less the effect of the eccentricity of the earth's orbit than of the unequal division of the continents, determines the limit between the N.E. and and S.E. trade-winds. But, as this limit is much more to the north of the equator in the Atlantic Ocean than in the South Sea, we may conclude that, in a region between 130° and 150° of W. longitude, the difference of temperature between the two hemispheres is less great than farther to the east in 20° or 50° of longitude. The low strata of the atmosphere which rest upon the aqueous surface of the globe, receive the influence of the temperature of the waters. The sea radiates less absolute heat than continents; it cools the air upon the sea, by the effect of evaporation; it sends the particles of water cooled and heavier towards the bottom; and it is heated again, or cooled, by the currents directed from the equator to the poles, or by the mixture of the superior and inferior strata on the sides of banks. With respect to the temperature of the ocean, we must distinguish between four very different phenomena. 1st. The temperature of the water at the surface corresponding to different latitudes, the ocean being considered at rest, and destitute of shallows and currents. 2d. The decrease of heat in the superimposed strata of water. 3d. The effect of billows on the temperature of the surface water. 4th. The temperature of currents, which impel, with an acquired velocity, the waters of our zone across the immoveable waters of another zone. Hitherto we have attended to the distribution of heat on the surface of the globe at the level of the sea. It only remains for us to consider the variations of temperature in the higher regions of the atmosphere, and in the interior of the earth. The decrease of heat in the atmosphere depends on several causes, the principal of which, according to Laplace and Leslie, is the property of the air to increase its capacity for heat by its rarefaction. If the globe was not surrounded by a mixture of elastic and aeriform fluids, it would not be sensibly colder at the height of 8747 yards than at the level of the sea. As each part of the globe radiates in every direction, the interior of a spherical envelope, which would rest on the top of the highest mountains, would receive the same quantity of radiant heat as the lower strata of the atmosphere. The heat, it is true, will be spread over a surface a little greater; but the difference of temperature will be insensible, since the radius of the spherical envelope will be to that of the earth as 1.001 to 1. Essay on Heat and Moisture, p. 11.; and Geometry, p. 495. Considering the earth as surrounded with an atmospherical fluid, it is obvious that the air heated at its surface will ascend, dilate itself, and be cooled, either by dilatation or by a more free radiation across the other strata that are equally rarefied. These are the ascending and descending currents, which keep up the decreasing temperature of the atmosphere. In comparing towns situated on elevated plains with those which are placed on the declivity of mountains, I have found for the first an augmentation of temperature, which, on account of the nocturnal radiation, does not exceed from 2.7 to 4.14. The following are the results which I have obtained from exact data in the temperate zone, from the plains to 1000 metres of elevation. Every hundred metres of perpendicular height diminishes the mean temperature of the year, by the same quantity that a change of one degree of latitude does in advancing towards the pole. If we compare only the mean temperature of summer, the first 1000 metres are equivalent to 0.81 Fahr. From 40° to 50° of latitude, the mean heat of the plains of Europe decreases in Europe 12.6 of Fahr.; and this same decrease of temperature takes place on the declivity of the Swiss Alps from 0 to 1000 metres of elevation. Differences of Latitude, Compared with Differences of Elevation. Mean Heat of the Year. Mean Heat of Summer. Mean Heat of Autumn. I. At the Level of the Sea. a. Latitude, 40° .......... 63.14° 77.00° 62.60° b. Latitude, 50° .......... 50.54 64.40 II. On the Declivity of Mountains, a. At the foot in 46° of latitude .... 53.60 68.00 51.80 b. At an elevation of 1000 metres .. 41.00 58.46 42.80 I shall now conclude this memoir by the enumeration of the most important results which have been obtained by Baron Von Buch, M. Wahlenberg, and myself, on the distribution of heat in the interior of the earth, from the equator to 70° of N. latitude, and from the plains to 3600 metres (11,808 feet) of elevation. The interior temperature of the earth is measured either by the temperature of subterraneous excavations, or by that of springs. This kind of observation is very liable to error, if the traveller does not pay the most minute attention to local circumstances which are capable of altering the results. The air, when cooled, accumulates in caverns, which communicate with the atmosphere by perpendicular openings. The humidity of rocks depresses the temperature by the effect of evaporation. Caverns that have little depth are more or less warmed according to the colour, the density, and the moisture, of the strata of stone in which nature has hollowed them. Springs indicate too low a temperature, if they descend rapidly from a considerable height upon inclined strata. There are some under the torrid zone and in our climate which do not vary in their temperature throughout the whole year more than half a degree, and there are others which show the mean temperature of the earth only by observing them every month, and taking the mean of all the observations. From the polar circle to the equator, and from the tops of mountains towards the plains, the progressive increase of the temperature of springs diminishes with the mean temperature of the ambient air. The temperature of the interior of the earth is, at Lat. Temp. Fahr. Vadso ....70°0' 35°96' Berlin ..52 31 49 28 Lat. Temp. Fahr. Paris ....48°50' 53°6' Cairo ....30 2 72 5 In equinoctial America, I have found it in the plains from 77° to 78.8°. The following are examples of the decrease of temperature from the plains to the tops of mountains. Zone of 30°--55°. Lat. Mean Temp. of Air, Fahr. Temp. of the Interior of the Earth. Cairo ............ 30°02' 72.68° 72.50° Natchez ........ 31 28 64.76 64.94 Charlestown .... 33 00 63.14 63.50 Philadelphia .... 39 56 53.42 52.16 Geneva .......... 46 12 49.28 50.74 Dublin .......... 53 21 49.10 49.28 Berlin .......... 52 31 47.30 49.28 Kendal .......... 54 17 46.22 47.84 Keswick ........ 54 33 48.02 48.56 Zone of 55°--70°. Calscrona ...... 56 06 46.04 47.30 Upsal .......... 59 51 41.90 43.70 Umeo .......... 63 50 33.26 37.22 Vadso .......... 70 00 29.66 35.96 When we consider what a large portion of the globe is covered with the sea, and examine the temperature of the deepest waters, we are constrained to admit that, in islands, along coasts, and perhaps even in continents of small extent, the interior heat of the earth is modified by the proximity of the strata of rocks on which the waters of the ocean rest. I have considered successively, in this memoir, the distribution of heat,--1, at the surface of the globe; 2, on the declivity of mountains; 3, in the ocean; 4, in the interior of the earth. In explaining the theory of isothermal lines and their inflexions, which determine the different systems of climates, I have endeavoured to reduce the phenomena of temperature to empirical laws. These laws will appear much more simple when we shall have multiplied and rectified by degrees the numerical elements which are the results of observation. In the following general Table of the distribution of heat, the temperatures are expressed in degrees of Fahrenheit; the longitudes are reckoned from east to west of the meridian of the observatory of Greenwich. The mean temperatures of the seasons have been calculated, so that those of the months of December, January, and February, form the mean temperature of winter. An asterisk (*) is prefixed to those places whose mean temperatures have been most accurately determined, and in general by means of 8000 observations. The isothermal lines have a convex summit in Europe, and two concave summits in Asia and Eastern America. Isothermal Bands. Names of Places. Position. Mean Temp. of the Year. Distribution of Heat in the different Seasons. Maximum and Minimum. Lat. Long. Height in feet. Mean Temp. of Winter. Mean Temp. of Spring. Mean Temp. of Summer. Mean Temp. of Autumn. Mean Temp. of Warm. Month. Mean Temp. of Coldest Month. Isothermal Bands from 32° to 41°. Nain ................ 57°08' 61°20' W 0 26.42° -- 0.60° 23.90° 48.38° 33.44° 51.80° --11.20° * Enontekies .......... 68 30 20 47 E 1356 26.96 + 0.68 24.98 54.86 27.32 59.54 -- 0.58 Hospice de St. Gothard 46 30 8 23 E 6390 30.38 18.32 26.42 44.96 31.82 46.22 15.08 North Cape .......... 71 00 25 50 E 0 32.00 23.72 29.66 43.34 32.08 46.58 22.10 * Uleo ................ 65 03 25 26 E 0 35.08 11.84 27.14 57.74 35.96 61.52 7.70 * Umeo ................ 63 50 20 16 E 0 33.26 12.92 38.80 54.86 33.44 62.60 11.48 * St. Petersburg ........ 59 56 30 19 E 0 38.84 17.06 38.12 62.06 38.66 65.66 8.60 Drontheim .......... 63 24 10 22 E 0 39.92 23.72 35.24 61.24 40.10 64.94 19.58 Moscow ............ 55 45 37 32 E 970 40.10 10.78 44.06 67.10 38.30 70.52 6.08 Abo ................ 60 27 22 18 E 0 40.28 20.84 38.30 61.88 40.64 -- -- Isothermal Bands from 41° to 50°. * Upsal .............. 59 51 17 38 E 0 42.08 24.98 39.38 60.26 42.80 62.42 22.46 * Stockholm .......... 59 20 18 03 E 0 42.26 23.52 38.30 61.88 43.16 64.04 22.82 Quebec .............. 46 47 71 10 W 0 41.74 14.18 38.84 68.00 46.04 73.40 13.81 Christiania .......... 59 55 10 48 E 0 42.08 28.78 39.02 62.60 41.18 66.74 28.41 * Convent of Peyssenburg 47 47 10 34 E 3066 42.98 28·58 42.08 58.46 42.98 59.36 30.20 * Copenhagen .......... 55 41 12 35 E 0 45.68 30.74 41.18 62.69 48.38 65.66 27.14 * Kendal .............. 54 17 2 46 W 0 46.22 30.86 45.14 56.84 46.22 58.10 34.88 Malonin Islands ...... 51 25 59 59 W 0 46.94 39.56 46.58 53.06 48.46 55.76 37.40 * Prague .............. 50 05 14 24 E 0 49.46 81.46 47.66 68.90 50.18 -- -- Gottingen ............ 51 32 9 53 E 456 46.94 30.38 44.24 64.76 48.74 66.38 29.66 * Zurich .............. 47 22 8 32 E 1350 47.84 29.66 48.20 64.04 48.92 65.66 26.78 * Edinburgh .......... 55 57 3 10 W 0 47.84 38.66 46.40 58.28 48.56 59.36 38.30 Warsaw ............ 52 14 21 02 E 0 48.56 28.76 47.48 69.08 49.46 70.34 27.14 Coire ................ 46 50 9 30 E 1876 48.92 32.36 50.00 63.32 50.36 64.58 29.48 Dublin .............. 53 21 6 19 W 0* 49.10 39.20 47.30 59.54 50.00 61.16 35.42 Berne .............. 46 05 7 26 E 1650 49.28 32.00 48.92 66.56 49.82 67.28 30.56 * Geneva .............. 46 12 6 08 E 1080 49.28 34.70 47.66 64.94 50.00 66.56 34.16 * Manheim ............ 49 29 8 28 E 432 50.18 38.80 49.64 67.10 49.82 68.72 33.44 Vienna .............. 48 12 16 22 E 420 50.54 32.72 51.26 69.26 50.54 70.52 26.60 Isothermal Bands from 50° to 59°. * Clermont ............ 45°46' 3°05' E 1260 50.00 34.52 50.54 64.40 51.26 66.20 28.04 * Buda ................ 47 29 19 01 E 494 51.08 33.98 51.08 70.62 52.34 71.60 27.78 Cambridge (U. S.) .... 42 25 71 03 W 0 50.36 33.98 47.66 70.70 49.82 72.86 29.64 * Paris ................ 48 50 2 20 E 222 51.08 38.66 49.28 64.58 51.44 65.30 36.14 * London .............. 51 30 0 05 W 0 50.36 39.56 48.56 63.14 50.18 64.40 37.76 Dunkirk ............ 51 02 2 22 E 0 50.54 38.48 48.56 64.04 50.90 64.76 37.75 Amsterdam .......... 52 22 4 50 E 0 51.62 36.86 51.62 65.84 51.62 66.92 35.42 Brussels ............ 50 50 4 22 E 0 51.80 36.68 53.24 66.20 51.08 67.28 35.60 * Franeker ............ 52 36 6 22 E 0 51.80 36.68 51.08 67.28 54.32 69.08 32.90 Philadelphia .......... 39 56 75 16 W 0 53.42 32.18 51.44 73.94 56.48 77.00 32.72 New York .......... 40 40 73 58 W 0 53.78 29.84 51.26 79.16 54.50 80.78 25.34 * Cincinnati ............ 39 06 82 40 W 510 53.78 32.90 54.14 72.86 54.86 74.30 30.20 St. Malo ............ 48 39 2 01 W 0 54.14 42.26 52.16 66.02 55.76 66.92 41.74 Nantes .............. 47 13 1 32 W 0 54.68 40.46 54.50 68.54 55.58 70.52 39.02 Pekin .............. 39 54 116 27 E 0 54.86 26.42 56.30 82.58 54.32 84.38 24.62 * Milan .............. 45 28 9 11 E 390 55.76 36.32 56.12 73.04 56.84 74.66 36.14 Bourdeaux .......... 45 50 0 34 W 0 56.48 42.08 56.48 70.88 56.30 73.04 41.00 Isothermal Bands from 59° to 63°. Marseilles ............ 43 17 5 22 E 0 59.00 45.50 57.56 72.50 60.08 74.66 44.42 Montpellier .......... 43 36 3 52 E 0 59 36 44.06 56.66 75.74 60.98 78.08 42.08 * Rome .............. 41 53 12 27 E 0 60.44 45.86 57.74 75.20 62.78 77.00 42.26 Toulon ............. 43 07 5 50 E 0 62.06 48.38 60.80 75.02 64.40 77.00 46.40 Nangasacki .......... 32 45 129 55 E 0 60.80 39.38 57.56 82.94 64.22 86.90 37.40 Natchez ............ 31 28 90 30 W 180 64.76 48.56 65.48 79.16 66.02 79.70 46.94 Isother. Bands from 68° to 77°. * Funchal .............. 32 37 16 56 W 0 6°.54 61.40 65.84 72.50 72.32 75.56 64.04 Algiers .............. 36 48 3 01 E 0 69.98 61.52 65.66 80.24 72.50 82.76 60.08 Isother. Bands above 77°. * Cairo ................ 30 02 31 18 E 0 72.32 58.46 73.58 85.10 71.42 85.82 56.12 * Vera Cruz ............ 19 11 96 01 W 0 77.72 71.96 77.90 81.50 78.62 81.86 71.06 * Havannah ............ 23 10 82 13 W 0 78.08 71.24 78.98 83.30 78.98 83.84 69.98 * Cumana ............ 10 27 65 15 W 0 81.86 80.24 83.66 82.04 80.24 84.38 79.16