Appendix II:  Remarks on the Meteorological Observations at Framheim By B. J. Birkeland
The South Pole by Roald Amundsen

The First Account  |  Introduction, - Fridtjof Nansen  Chapters: I. The History of the South Pole | II. Plan and Preparations | III. On the Way to the South | IV. From Madeira to the Barrier | V. On the Barrier | VI. Depot Journeys | VII. Preparing for Winter | VIII. A Day at Framheim | IX. The End of the Winter | X. The Start for the Pole | XI. Through the Mountains | XII. At the Pole | XIII. The Return to Framheim | XIV. Northward | XV. The Eastern Sledge Journey - Lieutenant K. Prestrud | XVI. The Voyage of the "Fram" - First-Lieutenant Thorvald Nilsen
Appendix I: - The "Fram" - Commodore Christian Blom  Appendix II: Remarks on the Meteorological Observations at Framheim - B. J. Birkeland
Appendix III:
Geology By J. Schetelig Appendix IV: The Astronomical Observations at the Pole - A. Alexander, with Note by Professor H. Geelmuyden
Appendix V:
Oceanography - Professors Bjorn Helland-Hansen and Fridtjof Nansen

Appendix II

On account of the improvised character of the South Polar Expedition, the meteorological department on the Fram was not so complete as it ought to have been. It had not been possible to provide the aerological outfit at the time of sailing, and the meteorologist of the expedition was therefore left behind in Norway. But certain things were wanting even to complete the equipment of an ordinary meteorological station, such as minimum thermometers and the necessary instructions that should have accompanied one or two of the instruments. Fortunately, among the veterans of the expedition there were several practised observers, and, notwithstanding all drawbacks, a fine series of observations was obtained during ten months' stay in winter-quarters on the Antarctic continent. These observations will provide a valuable supplement to the simultaneous records of other expeditions, especially the British in McMurdo Sound and the German in Weddell Sea, above all as regards the hypsometer observations (for the determination of altitude) on sledge journeys. It may be hoped, in any case, that it will be possible to interpolate the atmospheric pressure at sea-level in all parts of the Antarctic continent that were traversed by the sledging expeditions. For this reason the publication of a provisional working out of the observations is of great importance at the present moment, although the general public will, perhaps, look upon the long rows of figures as tedious and superfluous. The complete working out of these observations can only be published after a lapse of some years.

As regards the accuracy of the figures here given, it must be noted that at present we know nothing about possible alterations in the errors of the different instruments, as it will not be possible to have the instruments examined and compared until we arrive at San Francisco next year. We have provisionally used the errors that were determined at the Norwegian Meteorological Institute before the expedition sailed; it does not appear, however, that they have altered to any great extent.

The meteorological outfit on the Fram consisted of the following instruments and apparatus:

Three mercury barometers, namely:

One normal barometer by Fuess, No. 361 . One Kew standard barometer by Adie, No. 889. One Kew marine barometer by Adie, No. 764.

Five aneroid barometers:

One large instrument with thermometer attached, without name or number. Two pocket aneroids by Knudsen, Copenhagen, one numbered 1,503. Two pocket aneroids by Cary, London, Nos. 1,367 and 1,368, for altitudes up to 5,000 metres (16,350 feet). Two hypsometers by Casella, with several thermometers.

Mercury thermometers:

Twelve ordinary standard (psychrometer-) thermometers, divided to fifths of a degree (Centigrade). Ten ordinary standard thermometers, divided to degrees. Four sling thermometers, divided to half degrees. Three maximum thermometers, divided to degrees. One normal thermometer by Mollenkopf, No. 25.

Toluene thermometers:

Eighteen sling thermometers, divided to degrees. Three normal thermometers-by Tounelot, No. 4,993, and Baudin, Nos. 14,803 and 14,804. Two torsion hair hygrometers of Russeltvedt's construction, Nos. 12 and 14. One cup and cross anemometer of Professor Mohn's construction, with spare cross. One complete set of precipitation gauges, with Nipher's shield, gauges for snow density, etc.

Registering instruments:

Two barographs. Two thermographs. One hair hygrograph. A number of spare parts, and a supply of paper and ink for seven years.

In addition, various books were taken, such as Mohn's "Meteorology," the Meteorological Institute's "Guide," psychrometric tables, Wiebe's steam-pressure tables for hypsometer observations, etc.

The marine barometer, the large aneroid, and one of the barographs, the four mercury sling thermometers, and two whole-degree standard thermometers, were kept on board the Fram, where they were used for the regular observations every four hours on the vessel's long voyages backwards and forwards.

As will be seen, the shore party was thus left without mercury sling thermometers, besides having no minimum thermometers; the three maximum thermometers proved to be of little use. There were also various defects in the clockwork of the registering instruments. The barographs and thermographs have been used on all the Norwegian Polar expeditions; the hygrograph is also an old instrument, which, in the course of its career, has worked for over ten years in Christiania, where the atmosphere is by no means merciful to delicate instruments. Its clockwork had not been cleaned before it was sent to the Fram, as was done in the case of the other four instruments. The barographs worked irreproachably the whole time, but one of the thermographs refused absolutely to work in the open air, and unfortunately the spindle pivot of the other broke as early as April 17. At first the clockwork of the hygrograph would not go at all, as the oil had become thick, and it was not until this had been removed by prolonged severe heating (baking in the oven for several days) that it could be set going; but then it had to be used for the thermograph, the mechanism of which was broken, so that no registration was obtained of the humidity of the air.

The resulting registrations are then as follows: from Framheim, one set of barograms and two sets of thermograms, of which one gives the temperature of the air and the other the temperature inside the house, where the barometers and barograph were placed; from the Fram we have barograms for the whole period from her leaving Christiania, in 1910, to her arrival at Buenos Aires for the third time, in 1912.

Of course, none of these registrations can be taken into account in the provisional working out, as they will require many months' work, which, moreover, cannot be carried out with advantage until we have ascertained about possible changes of error in the instruments. But occasional use has been made of them for purposes of checking, and for supplying the only observation missing in the ten months.

The meteorological station at Framheim was arranged in this way: the barometers, barograph, and one thermograph hung inside the house; they were placed in the kitchen, behind the door of the living-room, which usually stood open, and thus protected them from the radiant heat of the range. A thermometer, a hygrometer, and the other thermograph were placed in a screen on high posts, and with louvred sides, which stood at a distance of fifteen yards to the south-west of the house. A little way beyond the screen, again, stood the wind-vane and anemometer. At the end of September the screen had to be moved a few yards to the east; the snow had drifted about it until it was only 2 1/2 feet above the surface, whereas it ought to stand at the height of a man. At the same time the wind-vane was moved. The screen was constructed by Lindström from his recollection of the old Fram screen.

The two mercury barometers, the Fuess normal, and the Adie standard barometer, reached Framheim in good condition; as has been said, they were hung in the kitchen, and the four pocket aneroids were hung by the side of them. All six were read at the daily observations at 8 a.m., 2 p.m., and 8 p.m. The normal barometer, the instructions for which were missing, was used as a siphon barometer, both the mercury levels being read, and the bottom screw being locked fast; the usual mode of reading it, on the other hand, is to set the lower level at zero on the scale by turning the bottom screw at every observation, whereupon the upper level only is set and read. The Adie standard barometer is so arranged that it is only necessary to read the summit of the mercury. It appears that there is some difference between the atmospheric pressure values of the two instruments, but this is chiefly due to the difficult and extremely variable conditions of temperature. There may be a difference of as much as five degrees (Centigrade) between the thermometers of the two barometers, in spite of their hanging side by side at about the same height from the floor. On the other hand, the normal barometer is not suited to daily observations, especially in the Polar regions, and the double reading entails greater liability of error. That the Adie barometer is rather less sensitive than the other is of small importance, as the variations of atmospheric pressure at Framheim were not very great.

In the provisional working out, therefore, the readings of the Adie barometer alone have been used; those of the normal barometer, however, have been experimentally reduced for the first and last months, April and January. The readings have been corrected for the temperature of the mercury, the constant error of the instrument, and the variation of the force of gravity from the normal in latitude 45°. The reduction to sea-level, on the other hand, has not been made; it amounts to 1.1 millimetre at an air temperature of -10° Centigrade.

The observations show that the pressure of the atmosphere is throughout low, the mean for the ten months being 29.07 inches (738.6 millimetres). It is lower in winter than in summer, July having 28.86 inches (733.1 millimetres), and December 29.65 inches (753.3 millimetres), as the mean for the month, a difference of 20.2 millimetres. The highest observation was 30.14 inches (765.7 millimetres) on December 9, and the lowest 28.02 inches (711.7 millimetres) on May 24, 1911; difference, 54 millimetres.

Air Temperature and Thermometers.

As has already been stated, minimum thermometers and mercury sling thermometers were wanting. For the first six months only toluene sling thermometers were used. Sling thermometers are short, narrow glass thermometers, with a strong loop at the top; before being read they are briskly swung round at the end of a string about half a yard long, or in a special apparatus for the purpose. The swinging brings the thermometer in contact with a great volume of air, and it therefore gives the real temperature of the air more readily than if it were hanging quietly in the screen.

From October 1 a mercury thermometer was also placed in the screen, though only one divided to whole degrees; those divided to fifths of a degree would, of course, have given a surer reading. But it is evident, nevertheless, that the toluene thermometers used are correct to less than half a degree (Centigrade), and even this difference may no doubt be explained by one thermometer being slung while the other was fixed. The observations are, therefore, given without any corrections. Only at the end of December was exclusive use made of mercury thermometers. The maximum thermometers taken proved of so little use that they were soon discarded; the observations have not been included here.

It was due to a misunderstanding that mercury thermometers were not also used in the first half-year, during those periods when the temperature did not go below the freezing-point of mercury (-89° C.). But the toluene thermometers in use were old and good instruments, so that the observations for this period may also be regarded as perfectly reliable. Of course, all the thermometers had been carefully examined at the Norwegian Meteorological Institute, and at Framheim the freezing-point was regularly tested in melting snow.

The results show that the winter on the Barrier was about 19.° C. (21.6° F.) colder than it usually is in McMurdo Sound, where the British expeditions winter. The coldest month is August, with a mean temperature of -44.5° C. (-48.1° F.); on fourteen days during this month the temperature was below -50° C. (-58° F.). The lowest temperature occurred on August 13: -58.5° C. (-73.3° F.); the warmest day in that month had a temperature of -24° C. (-11.2° F.).

In October spring begins to approach, and in December the temperature culminates with a mean for the month of -6.6° C. (+2O.l° F.), and a highest maximum temperature of -0.2° C. (+31.6° F.). The temperature was thus never above freezing-point, even in the warmest part of the summer.

The daily course of the temperature -- warmest at noon and coldest towards morning -- is, of course, not noticeable in winter, as the sun is always below the horizon. But in April there is a sign of it, and from September onward it is fairly marked, although the difference between 2 p.m. and the mean of 8 a.m. and 8 p.m. only amounts to 2° C. in the monthly mean.

Humidity of the Air.

For determining the relative humidity of the air the expedition had two of Russeltvedt's torsion hygrometers. This instrument has been accurately described in the Meteorologische Zeitschrift, 1908, p. 396. It has the advantage that there are no axles or sockets to be rusted or soiled, or filled with rime or drift-snow.

Fig. 1.

Fig. 2.

Fig. 3.

The two horsehairs (h, h') that are used, are stretched tight by a torsion clamp (Z, Z', and L), which also carries the pointer; the position of the pointer varies with the length of the hairs, which, again, is dependent on the degree of humidity of the air. (See the diagrams.) These instruments have been in use in Norway for several years, especially at inland stations, where the winter is very cold, and they have shown themselves superior to all others in accuracy and durability; but there was no one on the Fram who knew anything about them, and there is therefore a possibility that they were not always in such good order as could be wished. On September 10, especially, the variations are very remarkable; but on October 13 the second instrument, No. 12, was hung out, and there can be no doubt of the correctness of the subsequent observations.

It is seen that the relative humidity attains its maximum in winter, in the months of July and August, with a mean of 90 per cent. The driest air occurs in the spring month of November, with a mean of 73 per cent. The remaining months vary between 79 and 86 per cent., and the mean of the whole ten months is 82 per cent. The variations quoted must be regarded as very small. On the other hand, the figures themselves are very high, when the low temperatures are considered, and this is doubtless the result of there being open water not very far away. The daily course of humidity is contrary to the course of the temperature, and does not show itself very markedly, except in January.

The absolute humidity, or partial pressure of aqueous vapour in the air, expressed in millimetres in the height of the mercury in the same way as the pressure of the atmosphere, follows in the main the temperature of the air. The mean value for the whole period is only 0.8 millimetre (0.031 inch); December has the highest monthly mean with 2.5 millimetres (0.097 inch), August the lowest with 0.1 millimetre (0.004 inch). The absolutely highest observation occurred on December 5 with 4.4 millimetres (0.173 inch), while the lowest of all is less than 0.05 millimetre, and can therefore only be expressed by 0.0; it occurred frequently in the course of the winter.


Any attempt to measure the quantity of precipitation -- even approximately -- had to be abandoned. Snowfall never occurred in still weather, and in a wind there was always a drift that entirely filled the gauge. On June 1 and 7 actual snowfall was observed, but it was so insignificant that it could not be measured; it was, however, composed of genuine flakes of snow. It sometimes happened that precipitation of very small particles of ice was noticed; these grains of ice can be seen against the observation lantern, and heard on the observer's headgear; but on returning to the house, nothing can be discovered on the clothing. Where the sign for snow occurs in the column for Remarks, it means drift; these days are included among days of precipitation. Sleet was observed only once, in December. Rain never.


The figures indicate how many tenths of the visible heavens are covered by clouds (or mist). No instrument is used in these observations; they depend on personal estimate. They had to be abandoned during the period of darkness, when it is difficult to see the sky.


For measuring the velocity of the wind the expedition had a cup and cross anemometer, which worked excellently the whole time. It consists of a horizontal cross with a hollow hemisphere on each of the four arms of the cross; the openings of the hemispheres are all turned towards the same side of the cross-arms, and the cross can revolve with a minimum of friction on a vertical axis at the point of junction. The axis is connected with a recording mechanism, which is set in motion at each observation and stopped after a lapse of half a minute, when the figure is read off. This figure denotes the velocity of the wind in metres per second, and is directly transferred to the tables (here converted into feet per second).

The monthly means vary between 1.9 metres (6.2 feet) in May, and 5.5 metres (18 feet) in October; the mean for the whole ten months is 3.4 metres (11.1 feet) per second. These velocities may be characterized as surprisingly small; and the number of stormy days agrees with this low velocity. Their number for the whole period is only 11, fairly evenly divided between the months; there are, however, five stormy days in succession in the spring months October and November.

The frequency of the various directions of the wind has been added up for each month, and gives the same characteristic distribution throughout the whole period. As a mean we have the following table, where the figures give the percentage of the total number of wind observations:

N. N.E. E. S.E. S. S.W. W. N.W. Calm.

1.9 7.8 31.9 6.9 12.3 14.3 2.6 1.1 21.3

Almost every third direction is E., next to which come S.W. and S. Real S.E., on the other hand, occurs comparatively rarely. Of N., N. W., and W. there is hardly anything. It may be interesting to see what the distribution is when only high winds are taken into account -- that is, winds with a velocity of 10 metres (32.8 feet) per second or more. We then have the following table of percentages:

N. N.E. E. S.E. S. S.W. W. N.W.

7 12 51 10 4 10 2 4

Here again, E. is predominant, as half the high winds come from this quarter. W. and N.W. together have only 6 per cent.

The total number of high winds is 51, or 5.6 per cent. of the total of wind observations.

The most frequent directions of storms are also E. and N.E.

The Aurora Australis.

During the winter months auroral displays were frequently seen -- altogether on sixty-five days in six months, or an average of every third day -- but for want of apparatus no exhaustive observations could be attempted. The records are confined to brief notes of the position of the aurora at the times of the three daily observations.

The frequency of the different directions, reckoned in percentages of the total number of directions given, as for the wind, will be found in the following table:

N. N.E. E. S.E. S. S.W. W. N.W. Zenith.

18 17 16 9 8 3 8 13 8

N. and N.E. are the most frequent, and together make up one-third of all the directions recorded; but the nearest points on either side of this maximum -- E. and N.W. -- are also very frequent, so that these four points together -- N.W., N., N.E., E. -- have 64 per cent. of the whole. The rarest direction is S.W., with only 3 per cent. (From the position of the Magnetic Pole in relation to Framheim, one would rather have expected E. to be the most frequent, and W. the rarest, direction.) Probably the material before us is somewhat scanty for establishing these directions.

Meteorological Record from Framheim.

April, 1911 -- January, 1912.

Height above sea-level, 36 feet. Gravity correction, .072 inch at 29.89 inches. Latitude, 78° 38' S. Longitude, 163° 37' W.

Explanation of Signs in the Tables.

SNOW signifies snow.

MIST ,, mist.

AURORA ,, aurora.

RINGSUN ,, large ring round the sun.

RINGMOON ,, ,, ,, moon.

STORM ,, storm

sq. ,, squalls

a. ,, a.m.

p. ,, p.m.

I., II, III., signify respectively 8 a.m., 2 p.m., and 8 p.m.

° (e.g., SNOW°) signifies slight.

2 (e.g., SNOW2) ,, heavy.

Times of day are always in local time.

The date was not changed on crossing the 180th meridian

Appendix III: Geology By J. Schetelig