THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CAMBRIDGE UNIVERSITY PRESS EonDon: FETTER LANE, E.G. C. F. CLAY, MANAGER (Biinburgfj: too, PRINCES STREET Errlin: A. ASHEK AND CO. i-cipjis: F. A. BROCKHAUS ilrfo got*: G. P. PUTNAM'S SONS Bombay anto Calcutta: MACMILLAN AND CO., LTD. All rights reserved THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER A STUDY OF SOUNDINGS WITH PILOT BALLOONS BY C. T. P. CAVE, M.A. J .v Nonne vides etiam diversis nubila ventis Diversas ire in partis inferna supernis ? LUCRETIUS, v. 646 Cambridge : at the University Press 1912 PRINTED BY JOHN CLAY, M.A. AT THE UNIVERSITY I'RESS CONTENTS CHAPTER PAGE LIST OF ILLUSTRATIONS ............ vi INTRODUCTION ............... vii I. THE STRUCTURE OF THE ATMOSPHERE AS DISCLOSED BY THE OBSERVATIONS OF PILOT BALLOONS AT DITCHAM ........... 1 II. THE METHODS OF OBSERVING; (a) TRIGONOMETRICAL METHOD WITH TWO THEODOLITES; (6) ONE THEODOLITE WITH ASSUMED UNIFORM RATE OF ASCENT; THE WORKING UP OF THE OBSERVATIONS. BALLOONS. THEODOLITES ....... 10 III. CHECKS ON THE ACCURACY OF THE METHODS OF REPRESENTING THE RESULTS OF THE OBSERVATIONS .............. 18 IV. THE RATE OF ASCENT OF BALLOONS THEORETICAL AND OBSERVED; LEAKAGE OF GAS FROM BALLOONS; THE RELATION OF THE VERTICAL MOTION OF BALLOONS TO THE GROUND CONTOURS .............. 25 V. SUMMARY OF RESULTS AND THE RELATION OF THE WIND TO THE SURFACE PRESSURE DISTRIBUTION .............. 32 VI. CHANGES OF THE WIND DURING THE DAY AND DURING CONSECUTIVE DAYS ... 54 VII. THE WIND IN THE STRATOSPHERE ........... 61 VIII. RATE OF INCREASE OF WIND VELOCITY NEAR THE SURFACE ...... 66 IX. GENERAL RESULTS; RELATION OF VERTICAL WIND DISTRIBUTION TO SURFACE PRESSURE DISTRIBUTION . 69 261602 LIST OF ILLUSTRATIONS PAGE Fig. 1. Model representing the vertical wind distribution on May 5, 1909, 6.43 p.m. . 4 * Sept. 1, 1907, 10.22 a.m. 5 3. Oct. 1, 1908, 4.20 p.m. . 6 11 4. May 7, 1909, 6.29 p.m. . 7 5. Nov. 6, 1908, 10.59 a.m. 8 6. April 29, 1908, 3.57 p.m. 8 7. July 29, 1908 . 9 Fi^s 8, 9 12 J. '{,0. v-/, 10, 11, 12 13 V J Fig. 13. Trajectory of pilot balloon, Feb. 22, 1909, 4.52 p.m. ..... 15 14. Pilot balloon ready for an ascent . . 16 15. Balance used when filling balloons .......... 17 16. Theodolite for observing balloons (Bosch) ......... 17 17. Feb. 26, 1908, 10.28 a.m. Height-wind diagrams: comparison of one and two theodolite methods ........... 18 18. Feb. 26, 1908, 11.5 a.m. Height- wind diagrams: comparison of one and two theodolite methods ........ 19 19. Feb. 18, 1909, 4.43 p.m. Height-wind diagrams: comparison of one and two theodolite methods ........ 20 20. Horizontal trajectories of balloon, June 3, 1908. ....... 21 21. Feb. 18, 1909 22 22. balloons, Feb. 26, 1908 22 23. Part of trajectory of ballon sonde of Aug. 5, 1909 . 23 24 23 25. 24 ,, air , 26. Relation of height of balloon to ground contours, Feb. 26, 1908 .... 29 27. )) V .... 29 28. ,, ,, June 3, 1908 .... 29 29. July 31, 1908 .... 30 30. Feb. 19, 1909 30 31. Relation of wind velocity to height in classes (a), (b), and (c) . 35 32. Feb. 1st, 1907, 6 p.m. a. Surface isobars. 6. Computed isobars at 2 kilometres 36 33. Feb. 20th, 1909 3 37 34. Diagram showing gradient and surface wind directions in Class (a) " Solid current " 37 ;, 35. May llth, 1907, 6 p.m. a. Surface isobars, b. Computed isobars at 3 kilometres 38 36. Diagram showing gradient and surface wind directions in Class (6). Increasing velocity 39 37. Feb. 2nd, 1908, 6 p.m. a. Surface isobars, b. Computed isobars at 3 kilometres . 40 38. Diagram showing gradient and surface wind directions in Class (c). Decreasing velocity 40 39. May 14th, 1907, 6 p.m. a. Surface isobars, b. Computed isobars at 2 kilometres 43 40. May 21st 3 43 41. Relation of Westerly winds to the distribution of pressure at sea level for ascents in Class (6) .............. 70 42. Relation of Northerly winds to the distribution of pressure at sea level for ascents of Class (b) 71 43. Relation of Southerly winds to the distribution of pressure at sea level for ascents of Class (b) 72 44. Relation of Easterly wind to the distribution of pressure at sea level for ascents in Class (c) 74 45. Diagrammatic map showing supposed relations of winds to the isobars . . 75 46. .......... 77 47. Relation of winds to the distribution of pressure at sea level in Class (e 2) 79 Diagrams : Relations of Wind Elements to Height . . . . . . 109 141 INTRODUCTION THE investigation of the wind currents of the air above the surface layers is one of the greatest importance in the study of meteorology ; one reason for the slow advance made by this science in the last fifty years is to be found in the fact that until quite recently meteorologists only took note of that part of the atmosphere that was close to the surface of the earth, and beyond some cloud observations and a few isolated records, such as those obtained by Glaisher, nothing was known of the conditions existing in the free air. The recent rise of aviation and its probable extension in the near future make it more than ever necessary to investigate the nature of the currents in the free air above the surface of the earth. During the last few years the conditions of temperature, humidity, and wind have been investigated by means of kites carrying self-recording instruments to very considerable heights. Free balloons carrying lighter instruments have continued these records to still higher regions, heights of 25 kilometres and more having been reached. The motion of such a balloon if accurately observed gives a record of the wind currents traversed by it in its ascent through the atmosphere. Such records may also be obtained by small balloons that carry no instrument when they are followed by means of a theodolite during their ascent. The following pages give some account of the investigation of the upper air by means of such observations, some of the records having been obtained from balloons carrying instruments and others from small free balloons carrying nothing beyond a stamped label to be posted if the burst balloon should be found after it reaches the earth. An account is given in the first chapter of the general types of structure disclosed by the observations, and figures are given of models prepared to show the sequence of wind velocities and directions met with during the ascents on occasions when the different types of structure were found. An account follows of the methods of observing the balloons and of the theodolites employed for this purpose, together with an account of how the observations are worked up to give the horizontal trajectory of the balloon, and the method of measuring the wind velocity and direction at different heights from the trajectory. In Chapter III will be found a discussion of the accuracy of the Vlll INTRODUCTION methods employed, and a comparison of the trajectory determined by the observations of two theodolites at opposite ends of a base line with that determined by the observations of one theodolite only, and the assumption that the balloon ascends with a uniform or at any rate with a known velocity. Following on this is a discussion of the rate of ascent of rubber balloons which it is of great importance to determine as accurately as possible ; in this connection the results of observations and of theoretical considerations by investigators in this country and on the Continent are given. The relation of surface air currents to the configuration of the ground is also touched on ; this is a point of great importance for aviators and it is one that should be gone into more fully with balloons that ascend more slowly than those that have been used in these investigations. A general summary of the results obtained is given in Chapter V in which certain types of structure in the atmosphere are recognised, and the different types are considered in their relation to the wind at the surface, the gradient wind, and the general distribution of pressure and temperature in the region. Five types are described : (a) wind in the upper air steady with no increase in velocity with height ; (6) wind in the upper air increasing, sometimes to several times the gradient value, but remaining more or less steady in direction ; (c) wind in the upper air decreasing in velocity ; (o?) reversals or great changes in wind direction in the upper air ; (e) wind in the upper air blowing away from centres of low pressure. In the types represented by these five classes the wind in the upper air has been compared with that on the surface. A consideration of the higher ascents has shown that the strongest current is as a rule to be found in the region just below the stratosphere. This rapidly moving current must be associated with a corresponding pressure distribution in that region. Recent researches 1 have tended to show that it is there that changes of pressure originate, and from this point of view the layer just below the stratosphere must be regarded as controlling the conditions throughout the atmosphere beneath. The transference of the supposed seat of action from the surface to the region of nine kilometres suggests that variations in the currents in the layers beneath might with advantage be referred to the conditions prevailing at the time at the nine kilometre level instead of to those at the surface. This method of looking at the results of the ascents was suggested to me by Dr W. N. Shaw when this book was already in type. An examination of the cases which are represented by diagrams at the end of the volume shows that the method would greatly simplify the systematic representation of the atmospheric stratum between the surface and the region in question. Starting with a strong Westerly wind under the stratosphere we find almost without exception that the Westerly wind falls off in the lower levels, and the falling off may proceed continuously to such an extent that the direction 1 See W. H. Dines, F.R.S., "Statical Changes of Pressure and Temperature in a Column of Air that accompany Changes of Pressure at the bottom," Quart. Journ. Royal Meteorological Society, vol. xxxvni. p. 41; and letters in Nature, vol. LXXXVIII. p. 141 by Dr W. N. Shaw, F.R.S., and p. 175 by Mr W. H. Dines. INTRODUCTION IX of motion is reversed at some point in the intermediate layers, so that near the surface an Easterly wind is shown instead of the Westerly one of the upper regions. Even if the intermediate layers themselves provided no variation in the distribution of pressure that would affect the velocity we should expect the strength of the current to be diminished in the lower layers because the density there is greater than in the higher regions, and the velocity corresponding to the pressure gradient transmitted from above would be less in the inverse proportion of the density ; on this ground alone the wind velocity near the surface would be reduced to about one- third of the velocity at nine kilometres. But in the cases considered it will be evident that the diminution in the Westerly wind is at a greater rate than the increase of density, and the additional decrease must be due to pressure distribution accruing in the lower layers. Actual reversals are accounted for by representing the additional decrease as a superposed Easterly wind originating from pressure distribution in the lower layers which is sometimes so great as to show a wind at the surface in a reversed direction. The gradual modification of the gradient with increasing depth below the nine kilometre layer could easily be accounted for by a distribution of temperature in the layers underneath such that the air to the North is always colder than the air to the South, that is to say by assuming a distribution of temperature that corresponds with the latitude. The more rapid reversals on special occasions would thus be accounted for by a fall of temperature from South to North greater than the average. It will be seen that from this point of view the reversal of the air current implies no discontinuity in the atmosphere below the nine kilometre level, the change from West to East taking place gradually throughout the whole thickness. This result is of general application in all the high ascents either as a decrease of the Westerly wind as described, or in some ca,ses as an absence of decrease of an Easterly wind if such should exist at the nine kilometre level, when the decrease of velocity due to the density as the surface is approached is balanced by the increase due to the pressure distribution in the lower layers. This effect of the lower atmosphere in producing an Easterly component of the wind which is stronger the nearer to the surface is quite in accord with the calculations of M. Teisserenc de Bort of an average Westerly circulation in the four kilometre level modified as regards the lower layers by the distribution of temperature. Similarly regularity is not apparent as regards the winds from North and South, and the recognition of this fact has led, on Dr Shaw's suggestion, to an examination of a number of the ascents by the analysis of the wind at each level into a West- East component and a South-North component. This process has simplified the classification of the ascents in a remarkable manner. It appears that the structure of the atmosphere as disclosed by all the high ascents can be represented as regards the West-East component by the gradual development of an East- West component increasing continuously as the surface is approached, and doubtless due to the temperature distribution in latitude. As regards the South-North X INTRODUCTION component the effect of the lower layers is to alter the velocity by the con tinuous addition of a component which may be from the North or from the South according to circumstances. The South-North component shows a decrease of intensity as the surface is approached but there is no differentiation between the effect of the lower layers such as that shown by the West-East component. Thus the variation in the Northerly and Southerly winds depends on meteorological conditions which may show effects in opposite directions on different occasions. The effect of the layers beneath the nine kilometre level may be seen in the ascent for Nov. 6th, 1908, when the West-East component at nine kilometres was 11 metres per second; the effect of the superposed East- West component due to the lower layers was to reduce the velocity fairly regularly till at 3 '5 kilometres the West-East component was balanced by the East-West component ; at lower layers there was a reversal and at one kilometre above sea level the East- West component was 13 metres per second, or perhaps it is clearer to say that the West-East component was 13 metres per second. At the same time the South-North component had increased from 5 to +9 metres per second. On Oct. 1st, 1908, the West-East component decreased from +11 metres per second at 9 kilometres to 1 at 4 kilo- metres, below which however the decrease was not maintained ; the South-North component decreased from 18 to 13 metres per second, the decrease continuing down to the ground level. One more example may be given, Sept. loth, 1911, an ascent not elsewhere discussed in this book ; the West-East component decreased from + 32 metres per second at 9 kilometres to 8 at 1 kilometre while the South-North component decreased from + 12 to 10 metres per second. The gradual increases which are here described may be distinguished from the occasional increases of velocity locally at different times at various levels which appear as protruberances on the curve of relation of velocity of the several components with height. With these localised disturbances may probably be grouped the remarkably rapid variations with velocity shown in the lower layers on some of the occasions when the balloon was lost to sight on account of clouds at a comparatively low level. For these disturbances no explanation is offered for the present. The foregoing considerations which did not suggest themselves till this book was already in type should be born in mind in Chapter IX which deals with the relation of vertical wind distribution to the distribution of pressure at the surface. The subject of the wind in the stratosphere forms the subject of a separate chapter. It is quite clear that when a balloon enters this region it meets with winds of much smaller intensity than those traversed below this level. With Westerly, Northerly, and Southerly winds the stratosphere wind as far as has been observed remains more or less the same in direction as the winds in the lower strata, though with greatly decreased velocity ; but when Easterly winds are found in the layers immediately below the stratosphere the wind in that region exhibits curious fluctua- tions ; the balloon trajectory traces out loops as though spiral motions were met with. Since the observations dealt with in this book were concluded several more balloons INTRODUCTION XI have been observed till they were well within the stratosphere, and these observations fully bear out what is herein recorded. When the balloon enters the stratosphere the West-East component decreases, as also however do the South-North or North-South components; a decrease in the West-East component would be occasioned if the air at this level were colder to the South than to the North, that is if there were a temperature gradient in latitude in the reverse direction to that at the surface ; this is probably the case ; observations in low latitudes by M. Teisserenc de Bort and Professor A. Lawrence Rotch have shown that at heights above nine or ten kilometres the temperatures in the low latitudes are lower than the temperatures at corresponding heights in higher latitudes. At the end of the book will be found two tables ; the first gives a list of the 200 ascents in order of date with the greatest height to which the observations were carried in each instance, to what class each ascent belongs, and the distance of the point of fall where this is known. The second table commencing on page 84 gives the wind velocity and direction for each ascent for every half kilometre of height, and at the beginning of each ascent will be found the gradient velocity and direction in all cases when the gradient was sufficiently definite for this to be calculated. I have to thank Mr R. Corless and Mr R. G. K. Lempfert, members of the staff of the Meteorological Office, for kindly giving me the necessary information about the gradient wind at the times of the balloon ascents. After the tables will be found 24 diagrams giving the wind velocity and direction plotted against the height for certain typical or interesting ascents, together with weather maps showing the isobars and the velocity and direction of the wind at the surface, information which was taken from the Weekly Weather Report of the Meteorological Office. Diagrams showing the variation of the wind with height have been prepared for all of the 200 ascents, but it was not found practicable to reproduce more than those that appear at the end of this book. Throughout the work metres and kilometres are employed for heights and distances, and metres per second for wind velocities. The direction of the wind is given in degrees from the North point, so that an East wind is 90, a South wind 180 and so on. The use of metric units has been adopted because they are used by the International Commission for Scientific Aeronautics, it being of great importance that observers in different countries should use the same units. In the case of atmospheric pressure the English unit of inches of mercury has been retained because the information concerning the pressure distribution at the times of the ascents has been taken from the publications of the Meteorological Office. The investigation of the upper air by means of pilot balloons is a somewhat lengthy process, and involves a considerable amount of tedious calculation, which can however be much lightened by the use of the slide rule and mechanical calculators. Apart from all other work the plotting of the trajectories of the 200 ascents has involved the solution of some 8000 triangles. Xll INTRODUCTION The investigations were undertaken and this book was written at the suggestion of Dr W. N. Shaw, F.R.S., to whom my grateful thanks are due for introducing me to a most interesting field of study and for his invaluable help both in the course of the investigations and in the writing of this volume. He also kindly supplied me with the diagrams of surface pressure in Figures 32, 33, 35, 37, 39, and 40. I must also express my indebtedness to Mr W. H. Dines, F.R.S., who has helped me in a number of ways ; without his unfailing assistance I should hardly have begun researches on the upper atmosphere. For the preparation of the diagrams that appear at the end of the volume I have to thank Miss Humphreys, a member of the staff of the Meteorological Office. C. J. P. C. DITCHAM PARK, PETEKSFIELD. 2 April, 1912. CHAPTER I THE STRUCTURE OF THE ATMOSPHERE AS DISCLOSED BY THE OBSERVATIONS OF PILOT BALLOONS AT DITCHAM THIS book gives the results of 200 observations of pilot balloons or ballons sondes. The process of observation consists in watching the progress of small balloons as they rise through the air and are carried along by the winds. The majority of the observations were made at Ditcham in Hampshire, on the southern slopes of the South Downs. Fourteen ascents in May, 1907, were made at Totland Bay in the Isle of Wight, and two, on July 1st, 1907, at Chobham Common. From the observations the height and horizontal distance of the balloon are computed, generally from minute to minute, by methods which will be described and discussed. The purpose of the investigation is to determine whether the wind in the upper air is the same in direction or velocity as that at the surface, and to form a numerical estimate of the deviations that are observed. If we regard the air as composed of a series of horizontal layers or strata each with its own special wind velocity and direction at the time of observation, it is evident that the balloon rising with approximately steady upward motion will be carried along first by the surface current, and subsequently by the currents in the successive layers. We may suppose without inaccuracy that very little time is lost in adjusting the velocity of the balloon to the flow of the air current in which it happens to find itself at any moment, because the whole mass of the balloon is quite trifling compared with the forces that would be required to hold it against a flowing current of air ; and we are therefore justified in assuming that the horizontal distance traversed in any minute represents the average velocity of the layer passed through by the balloon in its journey aloft during that minute. Thus the horizontal velocities of the air in successive layers may be regarded as given by the observations. The thickness of the layers traversed in successive minutes depends upon the rate at which the balloon is ascending. In many cases this has been assumed to be uniform, and reasons will be given for considering that this assumption is sufficiently nearly justified for the results to be relied upon as giving at least a general representation of the true motion of the atmosphere in successive layers. For the sake of uniformity the heights in the atmosphere will be given in kilometres, and for general purposes we shall consider the atmosphere as made up of a pile of layers each a kilometre in thickness. The highest of the observations deal with a pile of 18 such kilometre layers, which took 80 minutes or more to c. 1 2 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER observe, but the large majority were much less thick. In some ascents only one or two kilometre layers were traversed, but a large number included five kilometre layers. On each occasion the balloon was watched until one or other of the following events happened ; either the balloon became invisible because it entered the clouds, or it was seen to burst and begin a precipitate descent, or it became so small that the observer was no longer able to identify the speck in the telescope. Sometimes this last event happened in consequence of the gradual diminution of the speck beyond the power of the eye to identify it, and on other occasions the observer after taking his eye from the telescope could not again find the speck. Sometimes in hazy weather the balloon was lost to sight when, had it been clear, its diameter would have been easily discernible. A large number of the ascents were made in the evening, as at that time the convection movements due to the heating of the ground by the sun's rays no longer interfere much with the uniform ascent of the balloon ; moreover the haze, which was found to interfere with the observations during cloudless days, was less at that time than at an earlier hour ; it was also found that near the time of sunset, when the sky was becoming less bright, the balloon illuminated by the sun's rays became very easy to see at long distances ; in some cases a balloon was watched for many minutes after the sun had set on the earth's surface, the balloon shining brightly and looking like a planet seen through a telescope ; on one occasion a balloon was seen to burst under such circumstances at a horizontal distance of about 40 miles. The time of sunset was also convenient because it corresponded nearly with one of the hours of observation for the Daily Weather Report of the Meteorological Office. The heights reached, and in many cases the immediate cause of the termination of the experiment, are given in the general table of results. It will be understood from this description of the mode of procedure that from each ascent one can form a mental picture of the successive currents of air traversed by the balloon in its journey. It is true that the geographical position of the balloon will be different for each minute, but these differences are so small compared with the extent of the atmospheric current that the variation in position may be disregarded ; and as a general rule we may regard the position of the successive currents as applicable to the atmosphere immediately above the observer at the time of the commencement of the ascent. The mental pictures thus obtained of the succession of air currents which constitute the structure of the atmosphere over the observing station are of the most varied and sometimes of the most complicated character. In order to enable the reader to carry with him an idea of the sort of structure which may be disclosed by the observations of a pilot balloon the various features have been classified according to some prominent characteristic which is easily recognised in the diagrams repre- senting the results of the ascents. A further note must be made before proceeding to consider the different types of structure which have been disclosed by the observations. The wind close to the THE STRUCTURE AS DISCLOSED BY THE OBSERVATIONS surface is influenced by the shape of the ground and other obstacles from which the upper layers are free. Hence the variations close to the surface are often of a specially complicated character having little relation to the structure of the atmosphere as a whole. Generally speaking the wind increases from its surface value in direct proportion to the height above sea level (see Chapter ix), with some little veer in direction until a height of between half a kilometre and a kilometre is reached. o Thereafter the effect of the surface may be regarded as no longer applicable. The surface wind is accordingly a very unsatisfactory datum to which to refer the variations in the upper air. Generally speaking the wind gradually approximates to that computed as the " gradient wind " (see page 32) from the distribution of pressure at the surface. In many ways the gradient wind is a better datum than the observed surface wind and it has been noted in the tables and marked on the diagrams whenever a reasonably satisfactory computation could be made. In some cases however the distribution of pressure in the neighbourhood of the station is too irregular and too ill-defined for a satisfactory computation of the gradient wind to be made. With this explanation we proceed to refer to some of the principal types of structure of the atmosphere which a study of the diagrams has disclosed. It must be remembered that they are necessarily limited to the occasions when balloons can be followed with a telescope. These are generally occasions of clear weather. During rain, or when there is fog or low cloud, observations are not possible. The figures that follow are taken from cardboard models prepared to show the distribution of wind direction and velocity with height. Each card shows by its direction and length the wind direction and velocity at each kilometre of height 1 . In general the light coloured cards represent winds from between 300 (W.N.W.) through North to 120 (E.S.E), that may be supposed to come from polar regions; the dark cards winds from 120 through South to 300, that may be supposed to come from equatorial regions. This classification is only approximate, since it is evident that, for example, a Northerly wind at the station may be a current of air that has been drawn from an equatorial region, but has curved round and passes over the station from a Northerly direction. (a) " Solid " current. The first characteristic type is that which we may call the " solid " current, that is to say that after the interference of the surface has been passed, and the gradient velocity approximately reached the wind remains steady both in direction and velocity in the upper layers. This case is illustrated by the diagram representing the result of the ascent on May 5th, 1909 (Fig. 1). (6) Continued increase of velocity beyond that of the gradient wind. In some cases it is possible to explain the increase of wind velocity by the change in the distribution of pressure in the upper layers without any discontinuous change in the air supply by reference to the distribution of pressure and temperature on the surface. Often however no such explanation is evident. These cases are illustrated by the results of ascents on Sept. 1st, 1907 and Oct. 1st, 1908 (Figs. 2 and 3). 1 The arrow-head flies with the wind. 12 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER (c) Decrease of velocity in the upper layers. In this case after the gradient velocity has been reached the velocity falls off showing that the regime indicated by the surface pressure is over, and a new distribution commences, arising so far as we know from causes unrelated to the surface conditions. Sometimes the new regime is itself indicated by the observations at higher levels ; not infrequently the observa- tions had come to an end before the conditions in the higher levels were disclosed. This class is illustrated by the result of the ascent on May 7th, 1909 (Fig. 4). (d) Reversals, or great changes of direction in the upper layers. Reversals of a b FIG. 1. Model representing the vertical wind distribution on May 5, 1909, 6.43 p.m. Class (a), a. From South side. b. From West side. direction are nearly always preceded by the falling off of the velocity to a light air. These cases may therefore perhaps be regarded as a continuation of a,scents similar to those of class (c) if it had been possible to continue the observations. In these cases we see the superposition of distinct systems of currents without any specific relation between them that can be accounted for by a knowledge of surface conditions. This class is illustrated by the result of the ascent of Nov. 6th, 1908 (Fig. 5). (e) Upper wind blowing out from distant low pressure centre ; frequent reversals in the lower layers. A number of cases present themselves in which the THE STRUCTURE AS DISCLOSED BY THE OBSERVATIONS upper wind is completely at variance with the gradient wind, either in direction, in velocity, or in both. In many of these cases it seems as though the upper wind were blowing outwards from the region overlying a low pressure system. The wind in the upper layers is often far in excess of the gradient value, and increases as in class (6) ; there are cases in which the structure might be classified with either class. This class is illustrated by the result of the ascent of April 29th, 1908 (Fig. 6). (f) The wind in the Stratosphere. In several ascents balloons have been followed to great heights and have been kept in sight after they have entered the region of the stratosphere. It has been found from the results of the ascent of ballons sondes that after a certain height is reached, a height that varies from day to day and from place to place, the ordinary diminution of temperature with height ceases, and that the column of air above this height remains at approximately the a b FIG. 2. Model representing the vertical wind distribution on Sept. 1, 1907, 10.22a.m. Class (b). a. From South side. b. From East side. same temperature as far as the observations have continued. This upper region of the atmosphere has been called the Stratosphere or Isothermal Layer. It has been found that at a height corresponding roughly with the commencement of the strato- sphere the wind velocity usually decreases, sometimes in a very marked way. This class is illustrated by the results of ascents of July 29th and Oct. 1st, 1908 (Figs. 3 and 7). So far as observations of the movements of balloons are concerned the region thus identified is a region of lighter winds as compared with those of the strata beneath. Mr W. H. Dines, F.R.S., informs me that he has had a few cases of no appreciable decrease in velocity up to 16 km. It will be noticed that in many cases equatorial winds have polar winds above 6 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER them. The frequency is too great for this to be merely coincidence, but it is quite possible that this superposition represents the condition of clear weather necessary for the observation of balloons. It may be argued that in the converse case when an equatorial wind is above a polar one we should get conditions favourable for the a b FIG. 3. Model representing the vertical wind distribution on Oct. 1, 1908, 4.20 p.m. Class (b). a. From East side. b. From South side. formation of clouds and rain, and consequently conditions unfavourable for the observation of balloons, and some evidence will be brought forward to show that this is sometimes the case. THE STRUCTURE AS DISCLOSED BY THE OBSERVATIONS a b FIG. 4. Model representing the vertical wind distribution on May 7, 1909, 6.29 p.m. Class (c). a. From South side. b. From West side. THE STRUCTURE OF THE ATMOSPHERE IX CLEAR WEATHER . a. b Fir;. "}. Model representing the vertical wind distribution on Nov. ("I, 1'JOS, 10.59 a.m. Class (d). a. From South side. b. From East side. Fir;, li. Model representing the, vertical wind (listrit)ution on April 29, 1!)08, .'5.~>7 p.m. Class (). , the altitude from a is given by and (see Fig. 12), or aO = AOcos0, AO tan (j> cos 6 = AO tan ft, tan < = tan 3 sec .(3). 14 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER Putting for /3, and b cos X for 6 in (l) we get OB = -. , , r (b cos X sin y h cos y) (4). sm(<-y) v With the aid of a Brunsviga calculator a table was constructed showing what should be added to ft to give < for values of /3 between 5 and 60, and for values of 6 from 1 to 16; another table was constructed giving the numerical value of ( b cos X sin y h cos y) where 6 = 2680 metres and h = 105 metres for values of y from 5 to 25 and for values of X from to 10. With the aid of a slide rule it was then possible to calculate the height of the balloon when it was nearly in line with the two stations as long as there was sufficient difference between the altitudes and y. The use of these tables was a help in checking the heights when the balloon went in a Northerly or North Westerly direction, as was frequently the case ; but in such cases it was found better to plot the trajectory on the one theodolite method ; using the heights as calculated by the tables as a check on the accuracy of the other method. In cases where it was useless to calculate the positions of the balloon from the two sets of observations, or in cases where the balloon had been lost at one of the stations, the remaining observations were dealt with as in the one theodolite method. (b) The one theodolite method. It is often desirable to send up a pilot balloon at very short notice, as when the sky clears for a short time on a cloudy day ; or when some particular phenomenon is occurring, and there would not be time for an observer to go to the out station. In such cases the rate of ascent of the balloon must be known, and where this is the case the height is a function of the time that has elapsed from the start of the balloon. In the case of moderate heights the rate of ascent of the balloon may be taken as uniform. Theodolite observations are taken at each minute from the start and the position of the balloon is obtained from the bearing, the angular altitude and the assumed height. The problem of the rate of ascent of the balloon, and of the effect of rising or falling currents of air, which certainly occur in the lowest strata of the atmosphere will be referred to later. It is obvious that the one theodolite cannot give such accurate results as the two theodolite method when the base line is fairly large compared with the distance of the balloon, but it probably gives a very fair approximation to the conditions of wind velocity and direction ; with respect to direction it is probably superior to a kite. The great advantage of the one theodolite method is that it can be used almost at a moment's notice, whereas the two theodo- lite method requires more preparation ; the working up of the results of the obser- vations is far less laborious, and three or four sets of observations could be worked up for the one theodolite method in the same time that one could be done from the double observations. Evidence will be given later to show that when the balloon is in a bad direction for observation or when it gets fairly distant with respect to the length of the base line the one theodolite method is actually superior to the two theodolite method. THE METHODS OF OBSERVING In the case of either of the methods of observing we obtain a series of positions which gives the horizontal trajectory of the balloon. This is drawn out to scale (Fig. 13) and from the length of the minute runs is measured the wind velocity during that minute ; that is assumed to be the wind velocity corresponding to the mean height of the balloon during that minute. The wind velocities are read off the diagrams by means of a scale which gives the values of the wind velocities directly. Most of the diagrams have been drawn on a scale of 1 to 30480. This scale, 5-0 N IKm. -5-28 (5850) FIG. 13. Trajectory of pilot balloon, Feb. 22, 1909, 4.52 p.m. Observed with two theodolites. The cross lines show the position of the balloon at each minute; the figures in brackets are the heights in metres above the sea level. which is 1000 feet to 1 centimetre, was adopted in the first instance because English units were used in the measurements but the diagrams were drawn on millimetre paper ; with the use of metric units a more convenient scale might be found, but as a number of diagrams had been made before this change was made in the units it was thought better to keep the same scale ; it offers no difficulties except that a special scale had to be made to measure oft' kilometres on the diagrams. If the run of the balloon is very long the trajectory is drawn out on half the above scale. 16 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER The wind directions are read off the trajectory by means of a transparent celluloid square ruled as a protractor. The wind direction is given in degrees from the North point; East being 90, South 180, West 270 and North 360 or 0. When the values of the wind velocity and direction have been read off and tabulated a diagram is constructed giving the relation between the height and the wind elements. In the case of ordinary pilot balloon ascents the balloons used are rubber balloons weighing from 28 to 30 grammes (Fig. 14) ; they are filled to lift from 85 to 90 grammes; and their rate of ascent is about 152 metres (500 feet) per minute. They are coloured dark red as it is found that if they are so coloured they are more visible than if they are white. An uncoloured balloon is perhaps more easily seen against a blue sky at the beginning of the ascent but as the balloon expands the rubber gets thinner and more transparent. At first a smaller balloon was used, but FIG. 14. Pilot balloon ready for an ascent. it has been found convenient to use rather a larger size as stated above and always to use the same kind of balloon with the same lift in order to get the same rate of ascent. Many of the observations for wind velocity were made on large balloons carrying instruments (ballons sondes) which are sent up from Ditcham on the days arranged by the International Commission for Scientific Aeronautics for combined observations. These balloons are much larger than the pilots and many different sizes have been used. The most usual size is one weighing about 250 grammes ; they are filled with hydrogen to lift 200, 300 or more grammes ; at the 300 gramme lift they have a diameter of nearly one metre. The height attained before bursting both by the large balloons and the small ones varies greatly ; it depends principally on the quality of the rubber, and on how the balloons have been kept. It is well not to keep them long as the rubber deteriorates rapidly. It is best to keep them in a tin case with tissue paper to THE METHODS OF OBSERVING 17 prevent them from being in contact with the metal ; the case should be kept in a warm place, and the balloon may be well and evenly warmed just before inflation. In any case they should be kept in the dark as light has a deteriorating effect on rubber. While.it is being filled the balloon is attached to a balance (Fig. 15). By means of this instrument the lift of the balloon can be balanced against weights, and in this manner the balloon can be given a lift which can be measured without detaching it from the pipe by which it is filled. The observations were taken by the special theodolite designed by Dr de Quervain for observing balloons. Two such theodolites, Fig. 16, made by Herr Bosch of FIG. 15. Balance used when filling balloons. FIG. 16. Theodolite for observing balloons (Bosch). Strassburg were used until the spring of 1909. After that date most of the one theodolite observations were taken with a somewhat similar instrument made by Messrs Gary Porter. The telescope of these theodolites has a reflecting prism which reflects the light at right angles in such a way that the observer is always looking in a horizontal direction whatever the altitude of the balloon. The arm of the telescope that carries the eyepiece passes through the centre of. the vertical circle, so that the observer has merely to take his eye from the eyepiece to read the vertical angle, instead of having to read it from the side as in the ordinary pattern of theodolite. The horizontal circle is also arranged so as to be easily read from the same position. The Bosch theodolites have a telescope with an object glass 5 cm. in diameter ; with the eyepiece supplied the magnification is about 20 diameters. The Gary Porter instrument has an object glass of 6'1 cm. diameter, and has three eyepieces. c. CHAPTER III CHECKS ON THE ACCUBACY or THE METHODS OF REPRESENTING THE EESULTS OF THE OBSERVATIONS CONSIDERABLE doubt has been thrown on the accuracy of the one theodolite method 1 and it is evident that if there are up and down movements in the atmosphere the values for wind velocity obtained from the observations of angular altitude will be more or less at fault. A check on the method can be obtained by comparing the trajectory drawn by the two theodolite method with one drawn on the one theodolite Velocity: metres per second 10 20 Direction 300* 360 I 3 I, w V FIG. 17. Feb. 26, 1908, 10.28 a.m. Height- wind diagrams : comparison of one and two theodolite methods. One theodolite. Two theodolites. method from one only of the sets of observations used in determining the trajectory in the first case. Figs. 17 and 18 show the height- wind diagrams plotted from trajectories drawn out on both the method of one and of two theodolites. The wind velocities were fairly large above two kilometres, and the balloon was rising faster than the normal rate at the end of the observations, possibly owing to an 1 The Free Atmosphere in the Region of the British Isles, W. H. Dines, F.R.S., p. 28. CHECKS ON ACCURACY 19 upward current caused by hills ; up to two kilometres the agreement between the two diagrams is tolerably close. Fig. 19 shows the height-wind diagram drawn out by both methods for an ascent on Feb. 18th, 1909 when there was a reversal at about 3 '7 km. The two methods of plotting the trajectory give remarkably concordant results in this case ; the wind directions are in disagreement to a certain extent in some cases, particularly when the velocities are small ; this is only to be expected, for in the graphical method employed it is not easy to measure accurately the direction of a very short line ; and a small error in the position of the points representing the positions of the balloon at two consecutive minutes causes a large error in the direction line when the points are close together. I W Velocity: metres per second 10 20 Direction 300 360 C !\ FIG. 18. Feb. 26, 1908, 11.5a.m. Height-wind diagrams: comparison of one and two theodolite methods. - - One theodolite. --------- Two theodolites. Fig. 20 shows a pair of trajectories for June 3rd, 1908. It is I think clear from this diagram that when the balloon is at a great distance compared with the length of the base line the one theodolite method is the more accurate ; the irregularities between the positions at 7.45 and at 7.54 are certainly due to small errors in the bearings ; it is extremely unlikely that the wind between adjacent layers in the atmosphere should have fluctuations such as those shown on the two theodolite trajectory ; they would also entail large fluctuations in the vertical velocity ; a large vertical velocity being accompanied by a large horizontal velocity and vice versa ; while in the case where the balloon appears to reverse its direction its height would have to decrease at the same time. The one theodolite method in this case is at any rate not open to these objections as will be seen from the figure. When therefore the balloon gets into a bad position with regard to the base line, or when its distance becomes very great compared with the base, a combination of 32 20 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER the one and two theodolite methods is advisable, a rate of ascent being adopted which best fits in with the heights deduced from the observations. In the case of very high ascents which have been observed with two theodolites Velocity: metres per second 10 300 Direction 360 1QO C 2OO FIG. 19. Feb. 18, 1909, 4.43 p.m. Height-wind diagrams : comparison of one and two theodolite methods. One theodolite. . _ Two theodolites. it has been found that the assumption of a uniform ascensional velocity will not hold ; a nearer approximation would be to suppose that the balloon ascended with a uniform acceleration ; this assumption while not strictly true is probably sufficiently near the CHECKS ON ACCURACY 21 7-54 7-45 IKm. T S FIG. 20. Horizoptal trajectories of balloon, June 3, 1908. SS Calculated on one theodolite method. >v/ TT On two theodolite method. 22 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER 5-5 IKm FIG. 21. Horizontal trajectories of balloon, Feb. 18, 1909. SS One theodolite method from 4.43 to 5.16 p.m. TT Two theodolite method from 4.43 to 5.16 when balloon was lost from station A. One theodolite method from 5.16 to 5.31 p.m. FIG. 22. Horizontal trajectories of balloons, Feb. 26, 1908, 10.28a.m. and 11.5a.m. SS Trajectory calculated on one theodolite method. TT two CHECKS ON ACCURACY 23 truth for the practical purpose of constructing diagrams giving the relation between the height and the wind elements. An acceleration is chosen which best fits in with the heights as determined by the observations. Fig. 21 shows part of a trajectory so determined ; the balloon was at a horizontal distance of between 11 and 12 kilometres, and at a height of between 12'5 and 15 kilometres during this part of its trajectory. The two trajectories are in extremely close agreement in most places, and it is interesting to see the loop just after 7.56 making its appearance in both diagrams. 7-47 )< Two THEODOLITES o ONE THEODOLITE \ Kilometre / FIG. 23. Part of trajectory of ballon sonde of Aug. 5, 1909 ; comparison of the one and two theodolite methods. The position of the balloon at 8.0 p.m. was 11 '87 kilometres W. 13 S. of the starting point. The one theodolite method has been criticised on the grounds that in particular cases the rate of ascent of a balloon may differ widely from the normal. This may be the case particularly when the balloon is near the ground. If the balloon ascends with say double the normal velocity it is evident that the real wind velocities will be double those deduced from the assumed and erroneous rate of ascent. But it is uncommon to find the rate of ascent differing widely from the normal after the balloon has ascended a kilometre or thereabouts above the surface. Consider what effect the increase or decrease of height, due to an increase or decrease of ascensional velocity at the B FIG. 24. 24 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER beginning of the ascent, will have if subsequently the balloon ascends with the normal velocity. Let A be the station (Fig. 24), the position of the balloon after n minutes, and 0' its position after n + 1 minutes ; the wind velocity is measured from the length of BB' ; let OB = H and O f B' = H+h where h is the mean rate of ascent of the balloon ; let the altitude of = a and of 0' = a!. Then BB' = AB'-AB = (H+ h) cot a'-H cot a r-. sin (a a!} , = H - '-,+heQta 1 ........................ (l). sin a sin a If the value of H is in error and the real value is (H + d), then BB' becomes TT sin (a a') , sin (a a') , , H - -- : - , + -- -- : 7 . + Acota' .................. (2); sin a sin a sin a sin a subtracting (1) from (2) the difference is -' sm a sin a This expression is large if a and a' are small and if a differs considerably from a! ; but in actual observations if the altitudes are small it means that the n A balloon is a long way off and in such a case the altitudes will not vary much between consecutive minutes ; on the other hand in those cases where the altitude does vary much between consecutive minutes the balloon is near the station and a and a! are large ; hence in practice during the later part of an ascent the expression (3) is small. To take a numerical case suppose the balloon to be 1000 metres higher than the assumed height (probably a very extreme case) and let a= 15 30' and a! = 15 ; the true value of BB' will be about 126 metres more than the value calculated from the erroneous height, and the wind velocity deduced from the latter will be about two metres per second too large. With a long base line the observed wind velocities for the first one or two minutes will be liable to error. If A and B be the stations (Fig. 25) and OA the horizontal trajectory the length OA will be proportional to sin OB A ; hence a small error in the angle OB A will cause a large error in the calculated length OA. The effect of this will be referred to in Chapter ix. CHAPTER IV THE RATE OF ASCENT OF BALLOONS THEORETICAL AND OBSERVED; LEAKAGE OF GAS FROM BALLOONS ; THE RELATION OF THE VERTICAL MOTION OF BALLOONS TO THE GROUND CONTOURS THE RATE OF ASCENT OF BALLOONS THE rate of ascent of balloons in still air is important in connection with the one theodolite method. With a rubber balloon 1 the free lift remains constant except in so far as it is affected by the tension of the rubber which increases the pressure inside the balloon to some extent. When the balloon has attained a steady rate of ascent, which it does in a few seconds after the start, the resistance of the air must equal the free lift and the rate gf ascent is given by the equation L = kpr*v* where k is a constant, p the density of the air, r the radius and L the free lift of the balloon. If p, is the density of the gas inside the balloon f^r 3 is its mass which is constant, except for leakage. may be taken as constant and therefore p varies as 1 v 2 ~ , and as L is constant or v*pk is constant, and v varies as p~b. Mr Dines gives the following table showing approximately the corresponding values for height, density, and ascensional velocity, taking into account the fall of temperature with height : Height p v Surface I'OO 1-00 3000 m. -73 1-05 6000m. -53 MO 9000 m. -39 1-17 12000 m. -26 1-25 In the ascents considered in these pages there were three in which the balloon remained in a direction and at a distance favourable for calculation by the two theodolite method, and in which the balloon was observed from both stations up to heights exceeding 12 kilometres. The calculations were made by combining the 1 The Free Atmosphere in the Region of the British Isles, W. H. Dines, F.B.S., p, 27. c. 4 26 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER methods of one and two theodolites, as explained in Chapter n, and the following values were obtained : Vertical velocity of balloon at different Heights. v v v at 12 j at 4 km. at 12 km. v at 4 metres per min. Dines value 1-168 1908 Oct. 1 186-5 222 1-185 0'91 2 202 253 1-253 1'52 1909 Aug. 5 152 188 1-237 0'75 Professor Hergesell has made a large number of observations of the rate of ascent of pilot balloons, both in the free air, and in the Cathedral and in the library of the University at Strassburg 1 . He finds that the rate of ascent of small rubber balloons is a function of the weight and the free lift of the balloon. If A is the free lift and B the weight in kilogrammes, Q the cross section is proportional to (A + J3)*. If R is the air resistance, R A = Q .f( V] where V is the rate of ascent, and and W=0) 1 Proc. Roy. Soc. A. Vol. LXXX. p. 530. 4 _ 2 28 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER F- w I. W the ratio of the upward accelerations at pressures p and p is ^ _ w or (if -=? =A * \ -*o MA ~2 which must be equal to the ratio of the resistances experienced, viz. -L ~~* - i Now pi , r' I /n+l\% - = and = - Po m r 2 m t\m + n) v hence - = m - r \m+l/ \l-4 Thus the velocity of the balloon at first increases as the one-sixth power of the ratio of the density of the air at the elevation attained to the density at ground level and when n is large (that is when the elastic compression is small) the upward velocity reaches its maximum not far from the greatest elevation to which the balloon can attain. The balloon will cease to ascend when F= W and this leads to the following expression for the limiting value of m : rp, ,, 760 ihe pressure is then - - mm. m The balloon is of course losing hydrogen as it ascends. A balloon of the quality used for the majority of the ascents and filled to lift 85 grammes loses 27 / of its free lift in four hours ; or at the end of this time its free lift would be about 7 1 grammes. According to Professor Hergesell's formula it would then have a rate of ascent of 140 metres per minute instead of about 150. The loss of buoyancy in respect of time is practically linear for four hours, and the diminution of the rate of ascent in one hour is negligible. But it must be remembered that the balloon is expanding as it ascends and the loss of gas would be greater as the balloon's diameter increased. Connected with the subject of the rate of ascent of pilot balloons is the question of vertical currents in the atmosphere due to the wind blowing over irregularities on the ground. It is obvious that a hill would produce an upward current in a wind blowing against it, but what the amount of such an upward current would be has not been accurately determined. Captain C. H. Ley made a number of observations of pilot balloons, measuring the angular diameter of the balloons to ascertain the distance. From observations made in Herefordshire 1 he came to the conclusion that inequalities in the ground caused a marked effect in the upper air, and that rising currents due to hills were transmitted upwards to great heights, to as much as 20,000 feet in some cases. In a later paper however Captain Ley says that the 1 Quart. Journal R. Met. Soc. Vol. xxxiv. p. 27. THE RATE OF ASCENT OF BALLOONS 29 200 metres 100 metres / 50 metres above sea, 100 metres above sea, 50 metres above sea, iKm. FIG. 26. Relation of height of balloon to ground contours, Feb. 26, 1908. FIG. 27. Relation of height of balloon to ground contours, Feb. 26, 1908. FIG. 28. Relation of height of balloon to ground contours, June 3, 1908. 30 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER uj 05 O 05 O5 i I (V f E I a o co THE RATE OF ASCENT OF BALLOONS 31 effects of hills are "noticeable up to 1500 feet or more 1 ." Balloonists have noticed an increased vertical velocity when going over hills especially when near the ground. In order to examine these effects the figures 26 to 30 have been prepared ; they show the height of the balloon in relation to the height of the ground contours ; the dotted horizontal line represents the height at which the balloon would have been had it ascended with uniform velocity ; the other line shows the distance of the balloon above or below this point for each minute from the beginning of the ascent. On Feb. 26th, 1908 there was a considerable rise as the balloon passed over a hill; there was a strong wind which was blowing nearly at right angles to Compton Down, a long hill, with a steep face on what was the windward side at the time. There were two balloon ascents, one closely following the other, and both show the rise on passing over the hill, both balloons were about 2 kilometres high at the end of the ascent 2 . On June 3rd the height of the balloon appears to have little connection with the ground contours ; on this occasion the wind was very light and there are indications of a downward current between two thunderstorms. On July 31st, 1908 there was a rise over a hill and a drop in the vertical velocity behind it. On Feb. 19th, 1909 with a fairly strong wind a little above the surface the ground contours had no effect on the height of the balloon. Further observations on this point would be useful, but I do not consider that the effect of ground contours would vitiate the results of one theodolite observations, though in the first one or two kilometres it might cause some errors in the calculated wind velocities. It is known that considerable upward currents take place under cumulus clouds, but as most of the pilot balloon ascents here considered took place either in clear weather or with uniform sheets of cloud, there is no evidence on this point from these observations. In the ascent at 6.54 p.m. on June 3rd, 1908 the balloon ascended with considerably less than the normal velocity ; there is evidence of a downward current of 10 metres per minute between ground level and 0'5 kilometre ; it is of course possible that the balloon was leaking and was therefore not ascending with normal velocity ; at the time however there was a considerable quantity of false cirrus in the north, associated with a thunderstorm that had taken place in the afternoon, while to the south there was a large cumulo-nimbus in the distance, which produced a thunderstorm during the night ; it seems likely therefore that between the two disturbances there was really a downward current which caused the balloon to ascend with a decreased velocity. 1 Quart. Journal R. Met. Soc. Vol. xxxv. p. 18. 2 Both the balloons unfortunately burst prematurely. CHAPTER V SUMMARY OF RESULTS AND THE RELATION OF THE WIND TO THE SURFACE PRESSURE DISTRIBUTION SUMMARY OF RESULTS IN considering the 200 results obtained from the ascents it may be useful at the outset to have before us a note of what will be regarded as novel or unexpected, as distinguished from that which our previous knowledge would lead us to expect. We may start from the consideration of the gradient wind as computed from the distribution of pressure. Recent work on the comparison of the observed wind at the surface, or in the lower sections of the upper air by means of kites, with the distribution of pressure at the surface have led to the general conclusion that the direction of the isobar is frequently an accurate representation of the direction of the surface wind. There is, however, usually a deviation of the surface wind from the isobaric line through some angle between and 45 towards the low pressure side, and there are occasions of wider divergence which deserve further investigation. As regards velocity, we may calculate it from the measured distance of two isobars between which the station lies, by means of the formula y = 2o)/3 V sin X where y is the gradient, &> the angular velocity of the earth, X. the latitude, V the velocity and p the density of the air. This gives a result which is certainly related to the actual wind, but which may require some modification on account of the curvature of the path traversed by the air, and is in excess of the observed velocity at the surface on account of the friction between the moving air and the ground. We cannot tell precisely what correction ought to be made for curvature of path because we do not know the curvature : we know however that if the curvature of the path is counter clockwise, as in a cyclonic depression, the velocity necessary to balance the gradient is less than that required for a straight path, and vice versa if the curvature of the path is clockwise like the isobar of an anticyclone then greater velocity is required to balance the gradient than would be required if the isobars were straight. Neither do we know what correction ought to be made in any special case on account of friction. Our only means of dealing with such a question is a prolonged study of the relation of the observed surface wind to the gradient, which has not yet SUMMARY OF RESULTS 33 been undertaken. Consequently in comparing the surface wind with the gradient wind we must be prepared to make allowance for corrections on account of these known causes, before we come to consider whether the observations disclose other causes, or other evidences of departures hitherto unrecognised. We know further that the distribution of pressure within the first kilometre of height is not likely to be very different in shape from that at the surface. A change in the distribution in height may be supposed to occur in consequence of the differences of density of columns of air over adjacent areas, due to differences of temperature, humidity or pressure, but for one kilometre these differences of density are small within a limited area. Hence as a first approximation (and doubtless with some exceptions) we may regard the distribution of winds indicated by the pressure at the surface as holding for the first kilometre or thereabouts. Above the surface the effect of friction will certainly be less marked, and in the upper air the comparison between the computed and actual winds will become freed from that influence. The agreement therefore should become closer above the ground level. This anticipation has been borne out by the observations recorded herein, and supported by others derived from kite ascents. From these we conclude that the gradient velocity will probably be reached within about one kilometre and the attainment of the gradient velocity marks the first stage in the structure of the atmosphere. The transition from the surface velocity to the gradient velocity may also, as a first approximation, be regarded as a linear variation proportional to the height of the point above sea level (see Chapter IX). We may therefore regard the first kilometre as the section within which the gradient velocity ought to be attained in normal cases. The cases in which the gradient velocity is not attained either as regards direction or velocity, or both, will be regarded as displaying some feature which should be the occasion of further investigation. We have next to consider the part of the structure above the first kilometre. Here we have little to guide us. We cannot fairly assume that the gradient above is the same as at the surface, because the pressure distribution may be seriously affected by differences of density in different columns owing to differences in the temperature at the base, or to the temperature gradient of the column, or to other causes. We are dependent therefore on the cases that actually present themselves, and we can only note that uniformity in direction and velocity of the wind above the first stage requires the lines of temperature and pressure to be similar and according to a certain law. Uniformity of direction requires isobars and isotherms to be parallel. Uniformity in velocity requires that there should be uniformity of temperature in each horizontal layer (i.e. distance between isotherms should be infinite). The surface gradient would certainly become intensified in the upper air if the lines of temperature distribution were similar to the lines of pressure distribution ; it would be attenuated if the lines of temperature distribution were opposite to those of surface pressure, supposing in both cases that the vertical temperature gradients are identical. c. 5 34 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER In a few cases maps have been constructed to show the pressure distribution at some particular height above the surface, the pressures being calculated from the surface pressures and temperatures alone. It is obvious that such maps must be approximations only to the real pressure distributions at the heights considered ; but there are cases where particular vertical wind distributions can be explained by supposing that the pressures in the upper air are due to the surface pressures and temperatures only. The difference of pressure at any height h above the stations s l and s 2 is given by the formula pi pi =2\p l \ Q1 2 (3 inches per kilometre), 4r 7 -L where p., and p 1 are the pressures and t* and ^ the temperatures at the stations s 2 and $!, at the ground, and p.! and pf the pressures at the heights under consideration above the stations. The pressure above Portland Bill, the nearest station to Ditcham for which values are given in the Daily Weather Report, is taken as unity, and isobars are drawn for every tenth of an inch above and below the value at Portland Bill. Since the shape of the isobars is given by the relative pressures the numerical value of the pressures has not been calculated. A reversal or a considerable deviation of the upper wind from the gradient wind, after it has once been reached, involves a change in the density conditions of the structure of the upper air which requires investigation. Hence the different cases will be classified according to the transformation that takes place above the first kilometre, or wherever the nearest approximation to the gradient wind is reached. It will be desirable here to indicate the method which has been adopted to represent the results. It will be remembered that the upper currents may show deviations from the surface current either in magnitude, or in direction, or in both, and in consequence we have to represent variation with height both in velocity and in direction. There is unfortunately no satisfactory method of representing these related variations in a single diagram. The most direct method of representing the results of an ascent is to make a plan of the path of the balloon as in Fig. 13, showing the trajectory for Feb. 22, 1909. In this case the orientation of the line shows the direction of the wind in each minute, and the distance of consecutive points marked on the diagram shows the magnitude of the velocity. The heights can also be written against each point, and thus all the information can be put on a single diagram, but it is a very extensive one, and the comparison of velocities by comparing distances does not easily present itself to the eye. We have therefore adopted the rather more recondite method of using two adjacent diagrams showing the variation with height of velocity and of direction respectively. A representation on this plan of the results of typical ascents, with weather maps showing the corresponding meteorological situation, follows the account of the general results. As regards the direction of the upper currents it is a matter of ordinary experience that sounding balloons sent up under various conditions of weather are SUMMARY OF RESULTS 35 generally, though riot by any means always, found at some point to the East of the starting point, arid the idea of a dominant or persistent westerly current in the upper air forming part of the general circulation round the pole has been kept in mind, but the occasions upon which a westerly upper current has declared itself as characteristic of the highest strata reached are comparatively few. The 200 ascents have been divided into the following classes : (a) 1. "Solid" current; little change in velocity or direction; the wind reaches the gradient value and does not increase very much at greater altitudes. 2. No current up to great heights. Considerable increase of velocity; gradient value reached and sur- passed ; increase often accounted for by surface temperatures. (b) (c) Decrease in velocity in the upper layers. (c) Reversals or great changes of direction. (e) Upper wind blowing outward from centres of low pressure; frequently reversals at a lower layer. 1. Upper wind between West and North. 2. Upper wind between South and West. Classes (a), (b) and (c) are represented diagrammatically in Fig. 31. There a,re many cases in which the wind does not reach the gradient velocity at all, but since many of these cases exhibit in other respects the same features as cases of classes (a) and (c) they have been included in these and will be dealt with under these classes. The con- sideration of the wind in the surface layer and in the Stratosphere will be dealt with in separate chapters. There are a certain number of ascents which cannot be classed with any of the above. It must also be remembered that the ascent may terminate or the balloon be lost to sight at a low level when, had it been 11 i i Wind Velocity observed to a greater height, the ascent Pia 31 Relation of wind velocity to height might have been included in another class. in Classes (a), (b), and (c). Class (a). " Solid " current. In this class (34 cases), the wind remains the same both in direction and in velocity in the upper layers. In order that this condition may be fulfilled the gradient and the shape of the isobars must be the same in the upper air as at the 52 36 surface, and for this to be the case the temperature distribution on the surface must be fairly uniform, and this is found to be the case in the examples under consideration. There are some exceptions, Feb. 1st, 1907, and Feb. 20th to 24th, 1909, for example, when there was a definite temperature gradient, high temperatures being found to the West and low to the East ; a map of the isobars at 2 kilometres on Feb. 1st, 1907 (Fig. 32) shows a gradient for northerly winds, but the gradient is not steep enough to cause any great increase in the wind velocity. A map of the isobars at 3 kilo- metres for Feb. 20th, 1909 (Fig. 33), shows a pressure gradient less than that at the surface ; on this occasion the gradient velocity was not reached, the wind remaining uniform at a velocity slightly below the gradient velocity. In most of the cases of this class the temperature distribution is either very uniform or irregular with no definite gradient. FIG. 32. Feb. 1st, 1907, 6 p.m. a. Surface isobars, b. Computed isobars at 2 kilometres, showing relative differences of pressure from that over Portland Bill. The direction of the wind, both gradient and observed, has been tabulated for each case (see Fig. 34) \ It will be noted that there is no recorded surface wind between 135 and 235 ; there are several cases in which the gradient wind is within these limits ; in five of these cases the surface wind is backed and in one case it is veered with regard to the gradient wind direction. It seems probable that the absence of surface winds from the South in this class is really due to the configuration of the ground at the station; the line of the South Downs runs from East to West ; 1 The two following cases have not been included in the diagram : June 23, 1908, gradient direction 330, surface direction 60; April 17th, 1907, gradient direction 40, surface direction 355. SUMMARY OF RESULTS 37 one would expect a wind from the East of South to be deflected more to the East and a wind from the West of South to be deflected more to the West on the southern slopes of the Downs. The deflection to the East is more marked than the deflection to the West, as should theoretically be the case ; an upper wind should be veered in respect to a surface wind owing to the Earth's rotation, apart from the local configuration of the ground. FIG. 33. Feb. 20th, 1909, 6 p.m. a. Surface isobars, b. Computed isobars at 3 kilometres showing relative differences of pressure from that over Portland Bill. Gradient Surface - -S-. FIG. 34. Diagram showing gradient and surface wind directions in Class (a) "Solid current." The cases of this class do not seem to be associated with any particular configu- ration of isobars. Those cases when there is no wind to speak of up to great heights may be considered to belong to a subclass of case (a). The cases sometimes occur in anti- cyclonic weather, and what little wind there may be varies from height to height, so that winds of various directions are found superposed one on the other. The three ascents of Aug. 5th, 1909, are typical of this class which is an uncommon one. 38 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER Class (?>). Considerable increase of velocity. In this class the wind velocity increases rapidly with height ; it usually reaches the gradient velocity at 0*5 to 1 kilometre, and the increase goes on rapidly to double the gradient velocity or even more. April 2nd, 1907, is a typical case ; the gradient velocity of 10 metres per second is reached at 0'5 kilometre ; while at 1'6 kilometre the observed velocity is double the gradient. The typical cases are associated with cyclonic systems of some intensity to the West or North, and there is often a strong temperature gradient along the surface, the high temperatures being found in the regions of high pressure and vice versa ; this in itself would increase the wind velocity in the upper layers, but it is FIG. 35. May llth, 1907, 6 p.m. a. Surface isobars, b. Computed isobars at 3 kilometres, showing relative differences of pressure from that over Portland Bill. not possible to calculate accurately what effects temperature would have in this direction without knowing the vertical temperature gradients over the area, in- formation which is not forthcoming. In many cases it is possible, however, to account for the increased velocity solely by differences of surface temperatures. May llth, 1907, Fig. 35, is a case in point ; the surface pressures are rather irregular with a gradient for light winds ; a map of the isobars at 3 kilometres shows a very strong pressure gradient over the English Channel which quite accounts for the strong wind at that height. Several cases are included in this class which differ markedly from the typical cases, as for instance May 18th and Feb. 5th, 1908. In these cases a high pressure area is over or close to the station and there is a strong temperature gradient in the right direction for increased velocity in the upper air. SUMMARY OF RESULTS 39 In comparing the gradient and the observed surface wind directions (see Fig. 36) it will be seen that the majority of cases of this class are from a direction between 135 and 180, and from 270 to 360 ; in other words the case is associated with advancing depressions or with depressions that have passed to the northward of the station. The phenomenon of a gradient wind from a direction East of South becoming a surface wind from a more easterly point, and a gradient wind from a point West of South becoming a surface wind from a more Westerly point is again brought out in this class. One or two cases require special mention. On Feb. 2nd, 1908, with a northerly wind at the surface, the wind at 4 kilometres reached a velocity of 30 metres per second, or three and a half times the gradient velocity. A V-shaped depression appeared over these islands on the following day, the main depression moving towards Scandinavia. A map of the isobars at a height of 3 kilometres, Fig. 37, shows that the gradient for northerly winds at that height is considerably steeper than on the surface. Gradient Surface FIG. 36. Diagram showing gradient and surface wind directions in Class (b). Increasing velocity. On Jan. 12th, 1909, is another case of extremely rapid increase in wind velocity with height ; the temperature gradient is in the right direction for an increase of westerly winds in the upper air ; it should be noticed in this case, however, that a deep depression moved down to the neighbourhood of this country by Jan. 14th and the case therefore has some resemblance to those in class (e). Class (c). Decrease in velocity in the upper layers. This class is almost entirely connected with easterly winds on the surface. In a typical case, such as Jan. 2nd, 1908, the surface wind is high but considerably below the gradient velocity ; the wind increases rapidly with height, often quite as rapidly as in Class (b) ; a maximum is reached at about 1 kilometre, above which there is a rapid falling off; the maximum, which is usually very well marked, some- times exceeds and sometimes falls short of the gradient velocity ; the latter is usually high, being often 20 metres per second or more. The diagram (Fig. 38) showing the wind direction, both gradient and observed, at the surface shows the easterly character of winds that fall off in the higher layers, THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER With few exceptions the gradient winds fall between 50 and 150, and the observed winds fall between and 100, and in every case save one the observed surface wind is backed with regard to the gradient wind. In general it may be said that in this class the pressure is high to the North or North-East and relatively low to the South. It has long been held that on the polar side of a depression the isobars in the higher layers no longer form closed curves ; in FIG. 37. Feb. 2nd, 1908, 6 p.m. a. Surface isobars, b. Computed isobars at 3 kilometres, showing relative differences of pressure from that over Portland Bill. Gradient Surface FIG. 38. Diagram showing gradient and surface wind directions in Class (c). Decreasing velocity. the upper layers the depression is a southward extension of the polar low pressure system ; the slackening of the wind velocity in the upper layers on the polar side of a cyclone quite corroborates this view. With regard to the wind direction in this class it has been pointed out above that the surface wind is nearly always backed with regard to the gradient wind. In some cases the gradient direction is never reached ; but it is reached or the nearest SUMMARY OF RESULTS 41 approach to it is reached a little above the point of maximum velocity. At higher levels there is sometimes a backing towards a Northerly or North-Easterly point, and sometimes the wind remains Easterly up to the highest point reached. In a few cases there is considerable veering in the upper layers ; for instance on May 16th, 1909. Examples of an Easterly wind being found at great heights occur on May 6th and 7th, 1909, on Aug. 5th, 1909, and on March 3rd, 1910. On May 6th and 7th, 1909, the direction remained from East to South-East up to the level of the Stratosphere ; on Aug. 5th, 1909, it remained between North and East up to 8 kilometres, when there was a discontinuity, the wind above being South-East up to 12 kilometres, when all real wind seems to have ceased. On March 3rd, 1910, the wind after being South-East at 3 kilometres slowly backed to North at 8 kilometres and remained between North and North-East to 11 kilometres when the peculiar wind conditions of the stratosphere were reached. Class (d). Reversals or great changes of direction in the Upper Layers. Cases of reversal of wind direction are very fairly common, but they are probably not so common as the number of cases recorded in these observations would lead one to expect ; for very often when these conditions were suspected a balloon was sent up when it would not have been otherwise. The explanation of the reversal is not always easy to determine and probably different causes are at work in different instances. In some cases when the wind velocity is small, as for instance on April 9th, 1907, the change through 360 may be explained by some local eddy, or even possibly by the spiral motion of the balloon during its ascent through a calm atmosphere. In some cases differences of surface temperatures may occasion a reversal of the pressure gradient in the upper air. The cases are considered in detail below. On Jan. 25th, 1907, a westerly wind was found at a height of one kilometre, the surface wind being from 115. On this occasion there was a ridge of high pressure over the station, between a deep depression over Scandinavia and a shallower one over Southern Spain. Temperatures were low over the Continent and relatively high over our Western coasts. The warm westerly current due to the Northern depression was evidently rising over the cold current due to the Southern depression. As will be found in further examples this rising of a moisture laden current results in precipitation, and in this example snow showers fell in most places in our islands. On March 30th, 1907, a surface wind from 250 veered to 55 at a little over one kilometre. There are no indications on the weather map of gradients for North- Easterly winds ; this case may possibly be due to the outflow of air from above the depression situated in the neighbourhood of Iceland, and as such should be classed with Class (e), but it is not a typical case. On this occasion there was no rain over these islands except a little in the extreme South- West of Ireland. On April 1st, 1907, the surface wind was South and it was South again at 4 kilometres, but the intermediate wind was about 115; the velocity was 3 to c. 6 42 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER 4 metres per second but it decreased to practically a calm at 3 kilometres when there was a sudden change of direction from 155 to 220 ; during this day and the next a Southerly wind was replacing an Easterly wind, the Southerly wind spreading from the West ; it would seem from this ascent that the Easterly wind still persisted for a time in the middle layers when it had been replaced by the Southerly wind at the surface and also at a considerable height. On April 9th, 1907, with extremely small wind velocity there was a backing of the wind completely round the compass between the surface and 1*5 kilometre; there was a centre of a cyclone almost over the station, and the temperatures over this country were extremely uniform. On April 15th, 1907, at 10.50 a.m. it would seem that the shallow depression that caused the surface wind from 35 failed to make its influence felt above one kilometre, and that the wind above was influenced by the low pressure system to the North-West ; the wind in this case was very light in the upper layers. It seems probable that the wind system on the South-East of the Atlantic low was bringing warm air from the South, and this current would rise over the North -Easterly current which came from colder regions ; in this case one would expect the warm and moist current from the Atlantic rising above the cold current to produce rain, and this was the case, rain having fallen over Southern England and over France. The ascent in the evening of the same day showed a somewhat similar veering of the wind, though at the highest point observed, T5 kilometre, it seemed to have backed to a more easterly point. On May 14th, 1907, there was a surface wind of 30 which gradually veered to 115 at a height of a little less than one kilometre ; at the same time the velocity fell nearly to zero ; above this the wind direction changed to 325 and then backed to 250 at 2 kilometres with some increase in velocity. The surface wind was probably due to the shallow low whose centre was over Brest, the upper wind being due to the depression over the North Sea. The temperature distribution was very irregular but it seems likely that in this case, as usual, there was a warm westerly current flowing over a colder easterly one ; we find also that there was a belt of rain from Spain to the North of England with thunderstorms in places. A map of the isobars at 2 kilometres, Fig. 39, shows that at this height the pressures in the neighbourhood of the North Sea were slightly lower than those near Brest. On May 17th, 1907, a light surface wind of 305 backed to 240 r at 2 kilometres; in this case the upper current certainly came from a warmer region. There was rain on our East coasts. On May 21st, 1907, a surface wind of 135 backed to 85 at 07 kilometre, and then veered to more than 250 at 2 kilometres, after which it backed to 215 with great increase in velocity ; the explanation of this curious wind distribution is probably due to the low pressure system situated over Brest being obliterated at the height of about one kilometre; above this height the wind seems to be influenced by the larger depression whose centre was between Scotland and Norway. A map of the isobars at 3 kilometres, Fig. 40, distinctly shows that surface temperatures would SUMMARY OF RESULTS FIG. 39. May 14th, 1907, G p.m. a. Surface isobars, b. Computed isobars at 2 kilometres showing relative differences of pressure from that over Portland Bill. FIG. 40. May 21st, 1907, 6 p.m. a. Surface isobars, b. Computed isobars at 3 kilometres showing relative differences of pressure from that over Portland Bill. 6-2 44 account for the obliteration of the shallow low. In this case again we have conditions theoretically favourable for rain, and we find that rain fell over almost the whole of these islands with thunderstorms in several places. On May 25th, 1907, there was a case somewhat similar to the last one, a North- Easterly wind veering to 200 at 2 kilometres ; the same explanation possibly applies ; there is again a low pressure system of small intensity over Brest with relatively high temperatures which would cause the depression to be obliterated in the upper layers. A map of the isobars at 2 kilometres shows that the Brittany depression has disappeared, or has been moved more to the North- West at this height. In this case also the Southerly current came from warmer regions than the north-easterly surface current ; there was rain with thunder in most districts. May 27th, 1907, is a very remarkable case ; the very shallow southerly breeze may have been a local phenomenon, the remains of a day breeze on to the land ; the Northerly wind seems to be caused by the general gradient due to the Anticyclone over Iceland and the depressions to the South ; but the upper South- Westerly current is difficult to explain ; it may possibly be the outflow from a depression which spread over this country on the morning of the 30th, the map giving some indications that such a depression existed over the Atlantic on the date of this ascent ; it moved very slowly after reaching our coasts and pursued an irregular course. This ascent should be compared with those in Class (e) and might perhaps have been included in that class. As in other cases of inversions there was rain with thunderstorms in places. On June 29th, 1907, a surface wind of 340 backed to 175 at 2*5 kilometres ; the velocity was 12 metres per second a little above the surface but rapidly fell off above; it would appear that surface temperatures would explain the obliteration of the small low in the upper layers. In this case the Continental temperatures were higher than those to the North-West, and we see the Southerly current above the Northerly one: rain and thunder resulted from these conditions. On July 1st, 1907, a ballon sonde was watched to the cloud level at 2 kilometres with a North wind ; as the balloon fell to the North-North-East there must have been a reversal at some point above the cloud level. The surface temperature distribution does not seem to explain the change of pressure gradient in the upper air. Whatever the explanation may be the Southerly current came from a warmer region than the surface Northerly current ; again in this case we find rain and thunderstorms over the district. On July 22nd, 1907, a South wind on the ground level became a West wind at a little above one kilometre. In the morning a shallow low was indicated over Kent; during the day the maximum temperatures were fairly high over the North coast of France, parts of Ireland, and the West of this country, while they were low over the East coasts ; it is probable therefore that the change of wind is associated with this temperature distribution. Rain and heavy thunderstorms were reported over the country. The case of Aug. 5th, 1907, is further discussed under the heading of " Changes SUMMARY OF RESULTS 45 occurring during the day." The remarkable backing of the wind at 2.24 p.m. is no doubt due to the pressure distribution ; a V-shaped depression was passing away to the East and influenced the surface wind, while the upper current was probably due to the low pressure system over the Hebrides. Temperatures over the Continent being much higher than those to the West we find the Southerly wind above the North-Westerly one, and as usual we have rain arid thunderstorms over the district. On Aug. 23rd, 1907, the surface wind of 155 differed by 175 from the gradient direction ; it is just possible that in this case a mistake had been made in the setting of the theodolite, though reference to the original note book does not support this idea ; the falling off of the wind above 2 kilometres and the sudden backing at 3 '5 kilometres seems to point to an inversion at a greater height ; there were intermediate clouds from West-North-West moving very slowly, with an upper cloud sheet moving more quickly. On Aug. 28th, 1907, at 11.38 a.m. a wind of 115 at the surface suddenly veered to 235 at 2 '5 kilometres. In the evening the surface wind which had become very light came from 165 and backed through East and North to 225 at 1'5 kilometre, remaining very steady above. The temperatures on the whole were higher to the South, but the case is not very clear. The upper wind is due no doubt to the low that was passing up our western coasts. In the morning observations the upper wind is strongly curved outwards from the surface isobars. There was rain in many places and lightning was reported in the Channel at night. On Oct. 14th, 1907, there was a very remarkable case of inversion ; at the surface the wind was from 340, and it rapidly rose to more than 10 metres per second at a height of about half a kilometre ; at a height of 0*8 kilometre it had fallen to a calm, but rapidly rose again to more than 10 metres per second at one kilometre, with sudden veering through 200 to direction 190, the gradient direction being about 215. A squall had just passed by with heavy rain, and the ascent was made because scud was seen to be moving from opposite directions in different layers. The 6 p.m. map shows a small centre of low pressure over the South-Eastern counties : this disturbance had evidently just passed the station when the ascent was made. On March 3rd, 1908, the wind veered considerably in the first half kilometre, going from 165 to 235 ; the direction then slowly backed to 215 at 2'5 kilometres when there was a sudden veer to 295 followed by a slow backing to 225 at 3*5 kilometres. The sudden change is without any adequate explanation ; the centre of a shallow low pressure system lay over the North of England on the morning of this day. On March 12th, 1908, a surface wind from 215 veered to the North, the change being nearly complete at 0'9 kilometre. In this case the lower temperatures on the Continent account for the obliteration of the shallow low over our East coasts and intensify that over Europe in the upper air. On March 1 3th the surface temperatures were again favourable to gradients for North- Westerly winds in the upper air. 46 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER In both cases the warm winds from the Atlantic seem to rise over the colder current from the Continent, and the result was that there was rain or snow in many places. On March 21st, 1908, the surface wind of 135 veered to 200 at 2 kilometres ; the upper wind which blew outwards from the surface isobars was no doubt due to the advancing low off the West coasts ; the warmer air from the Atlantic is here again seen to be flowing over the colder air from the Continent ; there were showers of rain and hail over the western districts of our islands. On April 9th, 1908, the surface wind was from 135 ; it was very light and proved to be a shallow current, only half a kilometre thick ; the upper wind was due to the general gradient. The temperatures were rather irregular but on the whole were higher in the West. There was rain over most parts of the British Islands. On June 2nd, 1908, the surface wind of 145 veered to 240 in the first kilo- metre ; the gradients are slight and irregular, and the temperatures over the Continent being relatively high the pressure in the upper layers would be lowest to the North-West ; this would occasion South-Westerly winds over the station in the upper air. June 2nd is an exception to the general rule that the upper current comes from the region of warmer temperatures ; the upper wind was from a little West of South while the surface wind came from the South-East where the temper- atures were the highest ; it is possible however that the surface air really was drawn from the region of lower temperatures over the North Sea, and that the current only became a South-Easterly wind near the station. On this day rain and thunderstorms were recorded. On July 27th, 1908, there was a curious case of a light Westerly wind veering to about 200 ; the change was complete at 1*5 kilometre, above which the direction remained extremely steady. In the upper layers this case corresponds to Class (b), the wind velocity increasing to 25 metres per second at 8 kilometres. The high temperatures over the Continent and the relatively low ones near the low to the North-West would increase the pressure gradient in the upper layers, and would also result in the Southerly wind flowing over the Westerly ; rain occurred but in this case thunderstorms were not reported. The two ascents on the morning of Sept. 30th, 1908, show a remarkable phenomenon ; after a slight veering up to 1 kilometre there is a great backing of the wind accompanied by a complete falling off of velocity ; above the calm layer there was a discontinuity and the direction above is not very different from that at the surface. At still greater heights the wind behaves like a Southerly wind of Class (b), the velocity increasing up to the greatest heights reached. It would appear that the lower ' Southerly wind is related to the high pressure system over the Continent, while the upper Southerly wind seems to belong to the system of low pressure over Iceland. On November 3rd, 1908, with an Easterly surface current of small velocity there was a steady backing of the wind direction from 130 at half a kilometre to 300 at SUMMARY OF RESULTS 47 4'5 kilometres. This was possibly the commencement of the Westerly wind that was found in the upper layers on Nov. 7th, 8th, and 9th. The velocities were small but were increasing slowly above 4 kilometres. On Nov. 6th and 7th, 1908, the conditions were very remarkable; a strong Easterly wind on the surface became a North- Westerly wind at 4 kilometres, there being a layer of calm air between 3 and 4 kilometres on Nov. 6th ; on the 7th the wind changed at a lower level, and the upper wind was from the West. On both days there were low temperatures over the Continent and high temperatures on the West coasts which to a certain extent would account for the reversal of the wind in the upper layers. A study of the weather maps for these days shows an Easterly or North-Easterly wind over Western Europe and a Westerly or North- Westerly wind over Eastern Europe ; it seems as though the upper Westerly wind descended at the region of highest pressure, part of it going on to make the Westerly wind of Eastern Europe, part of it flowing back to make the Easterly wind of Western Europe. On the 6th there was some rain on the East coasts of England and Scotland, and on the 7th over South- Western Erance ; on both days there were clear skies over Germany where presumably there was a current of descending air. On the following day, Nov. 8th, somewhat similar conditions were maintained ; the strong Easterly wind at the surface fell off from 25 metres per second at 1 kilo- metre to less than 5 metres per second at 2'5 kilometres; at the same time the direction which was 90 at 2 '5 kilometres backed steadily to 300 at 4'5 kilometres. Neither surface pressures nor temperatures would lead one to expect an upper Westerly current. These three cases are very instructive as they show the wind conditions that occurred during the passage of a well marked but not very deep depression to the South of this country. On the 7th there was a depression over the Western Mediterranean; this moved eastward during the day and decreased much in intensity, but it was followed by another which came from the Atlantic and was off the North- West coast of Spain by the evening of the 7th ; by the evening of the 8th its centre lay over the Southern coast of France. On all three days there is evidence of the upper wind descending and dividing into two parts, one going onward, the other flowing back under its own course in the upper air. On June 1st, 1909, we have a case of a Northerly surface wind changing to a Southerly wind at a height of 2 kilometres ; in this case it appears that the Northerly wind was forcing its way under a warm Southerly current and producing rain as shown on the weather chart for that date. The upper wind seems to have belonged to the system of the Continental Anticyclone ; the lower to the advancing Atlantic high pressure system which caused a flow of cold air from the Arctic regions. In practically all the cases of this class we find that the upper wind is flowing from the warmer regions, and the lower one from the colder ; so that the upper wind is, if not absolutely at any rate potentially, warmer than the lower. In almost every case we find rain occurring as a result, as would be expected if a warm current of air rose over a cold one. In most cases too we find thunderstorms occurring. In fact it 48 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER seems possible that a reversal or a great change of wind direction in the upper layers may be necessary for the production of lightning ; it is hard to suppose that a thunderstorm would go on for more than a short time unless electricity were being conveyed to the region of the thunderstorm from elsewhere ; if one current were at a different electrical potential from the other the supply might be maintained. In connection with the subject of reversals it must be noticed that when there is a discontinuity or sudden change of wind direction the velocity always falls off to a very small value ; even with a sudden veering or backing of the wind this is the case. This must be borne in mind in what has been stated above about one current flowing over another. The old idea of one current flowing close over another, and producing waves, as wind produces waves over the sea, receives no support either from theoretical considerations or from actual observations. It is obvious that the wind at any height must be the result of the pressure gradient at that height, and it does not seem possible that the pressure gradient should change abruptly between adjoining layers of the atmosphere. Calm layers mark the boundary between com- paratively sudden changes of wind direction. Class ( Q where V h is the velocity at the height h above sea level, h the height of a well exposed anemometer above sea level, and V s the velocity recorded by the anemo- meter. This linear increase of velocity with height near the ground is called the "regular" increase in the following pages. At Ditcham there is no anemometer and the horizontal velocity of the balloon in the first minute is taken instead for the purpose of discussing the variation of wind near the surface. Out of 174 ascents 61 per cent, show the regular increase. The following table shows the percentages of regular to non-regular increase for ascents observed with one and two theodolites. Increase Increase No. of regular not regular cases per cent. per cent. One theodolite 124 69 31 Two theodolites 50 42 58 Total 174 61 39 The number of cases of regular increase is much smaller for the two theodolite method than it is for the one theodolite method, and this might seem at first sight to cast some doubt on the reality of the phenomenon. But it must be remembered that the first observation in the two theodolite method is the least satisfactory, both from the large apparent motion of the balloon, owing to its being close to the stations, and from the fact that the triangle made by the balloon with the two ends of the base is not adapted for accurate calculation. It must also be remembered that an error of one or two metres per second would often transfer an ascent from the regular to the not regular class. When this is considered it is a very striking fact that so many of the ascents show this linear increase of velocity with height. 1 See Advisory Committee for Aeronautics, Reports and Memoranda, No. 9, by Dr W. N. Shaw, F.R.S. RATE OF INCREASE OF WIND VELOCITY NEAR THE SURFACE 67 If the ascents are classified according to the direction of the surface wind the following percentages are found, taking the North, East, South and West quadrants. Increase Increase No. of regular not regular Quadrant cases per cent. per cent. North 47 64 36 East 46 70 30 South 36 44 56 West 33 58 42 Total 162 60 40 In every case there is a preponderance of regular increase except in the South quadrant. If the North-East, South-East, South-West, and North-West quadrants are taken the following percentages are found. Increase Increase No. of regular not regular Quadrant cases per cent. per cent. North-East 50 62 38 South-East 44 64 36 South-West 26 46 54 North- West 41 66 34 Total 161 61 39 If the ascents are classified according to wind velocity the following percentages are found. Wind Increase Increase velocity No. of regular not regular metres per second cases per cent. per cent. to 4-9 82 67 33 5 to 9-9 84 58 42 Total 166 63 37 It thus appears that the strength of the surface wind does not make much difference to the regularity or otherwise of the initial increase of velocity ; but there are fewer cases of regular increase with winds from a Southerly or South- westerly direction than from other directions. At Ditcham a Southerly wind is one which strikes the line of the South Downs at right angles ; the Downs heing the first elevated ground met with by such a wind ; this would cause irregular eddies over the tops of the hills. The ground also rises 30 to 40 metres to the North of the station from which the balloons are sent up, and the ground to the North is covered with trees some of which are 20 metres high ; this must perceptibly affect the air in which the balloon is rising during the first minute of the ascent in the case of Southerly winds. The formula of increase of velocity with height 92 68 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER can only be considered correct for the particular locality under consideration. It is probable that the velocity increases in height by a factor and not by a constant addition, and a formula put forward by Dr Shaw 1 as a likely one is T7 H+ a rr where H is the height above the ground, V the velocity there, T 7 the anemometer velocity and a a constant which at Ditcham would be 167 metres. 1 Advisory Committee for Aeronautics, Reports and Memoranda, No. 9, p. 8. CHAPTER IX GENERAL RESULTS ; RELATION OF VERTICAL WIND DISTRIBUTION TO SURFACE PRESSURE DISTRIBUTION IN considering the various ascents described in the foregoing pages the following facts appear. The wind distribution of class (a), when the velocity remains fairly constant with regard to height, is found at times when the temperature gradient over the area is not very pronounced. Class (6) on the other hand is found when there is a considerable temperature gradient, the area of low temperatures being more or less coincident with the area of low pressure. This distribution of pressure would, apart from other things, cause an increased pressure gradient in the upper air. Winds of this class are found when cyclonic areas are centred to the North or to the West of the station. This is perhaps the normal condition of affairs when a cyclonic system is passing to the North of these islands, especially in winter, for warm air is drawn from the South- West from the Atlantic, causing a considerable temperature gradient. Winds of class (6) are found (l) to the South of a cyclone that is passing in an Easterly direction to the North of the station, the wind direction being West (Fig. 41), (2) on the North-East of an anticyclone with a depression far to the North- West, the wind direction being Northerly (Fig. 42) ; (3) on the West side of an anticyclone with a depression far to the North- West, the wind direction being Southerly (Fig. 43). The last type of pressure distribution does not however always give a vertical wind distribution of class (6) ; it is frequently found associated with class (e 1), the upper wind being North- Westerly. Of the 200 ascents described in this volume 50 are included in class (6), which thus seems to be the commonest form of wind distribution in this country. In fact it seems possible that if observations were carried high enough it would be found that in all cases, save in some cases of Easterly winds, the wind velocity would increase to several times the gradient value before the stratosphere was reached. This leads to the consideration as to whether the classification adopted is a valid one. It must be noted however that in most cases of class (6) the excess of velocity is met with in the first few kilometres. Those cases in which the increase does not take place till a height of 8 to 10 kilometres is reached might perhaps be included in class (a) ; and it might be better perhaps to differentiate the classes by the behaviour of the wind in the first 5 or 6 kilometres. 70 surface wind upper wind FIG. 41. Relation of Westerly winds to the distribution of pressure at sea level for ascents in class (b). - surface wind FIG. 42. Relation of Northerly winds to distribution of pressure at sea level for ascents of class (b). 72 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER surface wind FIG. 43. Relation of Southerly winds to the distribution of pressure at sea level for ascents of class (b). RELATION OF VERTICAL WIND DISTRIBUTION TO SURFACE PRESSURE DISTRIBUTION 73 Class (c), consisting of winds falling off in velocity with height is restricted almost entirely to Easterly winds, and in general the pressure is high to the North or North-East, and low to the South (Fig. 44). The dividing line between this class and the reversals is not sharply marked; a surface wind from the East often underlies a wind from a Westerly or South- Westerly direction, but this is by no means always the case, and both these classes must be looked at in conjunction with class (e) the upper winds that blow away from centres of low pressure. Indeed the different classes into which the varying forms of vertical wind distribution have been divided in the foregoing pages must be regarded to a certain extent as artificial ; in Nature they must be looked at as being in reality only different parts of the complicated structure of the Atmosphere. When there is no low pressure system for a long way to the West of the station the Easterly wind seems to be maintained, though- with diminished velocity, through the whole thickness of the troposphere ; but if there is a low pressure system to the West the type is usually of class (cZ) or (e) with a reversal to a Southerly or to a Westerly wind. A consideration of the cases in class (e) will I think make it clear that from centres of low pressure we may have vast currents of air flowing in the upper regions to great distances from the centres of the disturbances. The positions in which these conditions occur have already been indicated on the North-West, North, and North-East of an anticyclone and on the East, South-East, and sometimes on the North-East of a depression. Sometimes on the West of an anticyclone the dis- tribution is of class (6), sometimes of class (e) y the difference probably depending on the positions of neighbouring low pressure areas. Further investigation is needed on this and indeed on many other points. Air rises near the centres of depressions ; the air drawn into the cyclonic area and rising near its centre must flow outward in some way, and it appears from these observations that the outflow does not take place symmetrically in every direction, but it forms currents which flow usually in such a direction as to cause Westerly or North -Westerly winds in the upper layers. It is of the greatest interest to know to what points these upper currents flow. They must ultimately end in some region where they flow downward again towards the surface, for the equivalent of the air which is raised in the depression must return to the surface to take the place of the air that has been raised. Certain parts of anticyclonic areas are presumably those in which the air reaches the surface, and some cases have been indicated, notably those of Nov. 6th ; to 8th, 1908, where there is distinct evidence of a downward flow of the upper current. It is probably not in the centres of anticyclonic areas that the descent of air takes place. Dr W. N. Shaw, F.R.S., and Mr R. G. K. Lempfert, in their Life History of Surf ace Air Currents, state that they have "failed to identify the central areas of well marked anticyclones as regions of origin of air currents," and that the areas of descending air seem to be the shoulders and protuberances of anticyclones and the regions between cyclonic depressions 1 . 1 The Life History of Surface Air Currents, p. 24. c- 10 74 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER surface wind upper wind FIG. 44. Relation of Easterly winds to the distribution of pressure at sea level for ascents in class (c). RELATION OF VERTICAL WIND DISTRIBUTION TO SURFACE PRESSURE DISTRIBUTION 75 Fig. 45 shows diagrammatically a possible way in which the upper winds flow from the centre of a low pressure system towards an anticyclonic area. If the upper wind finds its way to the surface in some part of this latter region it must flow in a clockwise direction round the high pressure centre and return towards the low pressure in the direction indicated on the diagram. Now the air that has risen over the low pressure will usually be cooled at the wet adiabatic rate, while the air that descends over the anticyclone will be warmed at the dry adiabatic rate. If the same air has accomplished the journey it will return towards the anticyclonic area with a higher temperature than that at which it started. This warm air will move towards the East or South-East side of the depression and being a warm current will tend to rise further to the East of the point from which it, or its surface wind upper wind FIG. 45. Diagrammatic map showing supposed relations of winds to the isobars. equivalent, had started. In other words the point at which the air is rising, that is the centre of the depression, will have moved somewhat in the direction from which the anticyclonic air is flowing to feed it. Is it possible that these con- siderations may throw some light on the progressive motion of a cyclonic system in certain cases ? These considerations may also explain the fact that cyclonic systems may sometimes remain stationary for considerable time. If the air from an anticyclonic region passes over a cold land surface it will be chilled before reaching the low, and will not therefore have a tendency to rise in front of the depression, in other words the depression will not advance as in the case supposed above. This condition may be found in winter or early spring with a high to the North or North- West of these islands and a low over the Bay. The air flowing from the North on the East side of the high is partly drawn from the arctic regions and is also cooled 102 76 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER as it flows over the cold regions of the Continent, and the front of the low over the Bay will be partly fed by this cold air ; a depression in such conditions may remain stationary for some days (Fig. 46). In considering rising or falling currents of air it must be remembered that such currents are probably not exactly of the character formerly supposed. It seems Hkely that the rising air of a cyclonic depression is largely the rising of a warm current of air and its flowing over a colder current from a different direction, and C5 ' is not merely the ascent of air warmer than its surroundings 1 . The old idea of a warm core above a depression, causing a vertical circulation must be discarded since the temperatures over cyclonic systems have been found to be lower than those over anticyclones. Mr W. H. Dines in a paper read before the Royal Meteorological Society says "the term ascending current of a cyclone... appears to be incorrect. The actual phenomena seem rather to be a bulging upward of the strata between 1 or 2 kilometres and the isothermal, a bulging downward of the strata above the isothermal, accompanied with a lateral contraction over a large area of the lower strata and a lateral expansion of the strata below the isothermal. I much doubt whether the actual vertical motion of any air particle involved in a cyclonic circulation often exceeds two kilometres 2 ." The North -Westerly or Westerly current which the observations of pilot balloons have disclosed as the usual forerunner of an advancing depression should be considered in any theory of the dynamics of a cyclonic system. Is this current composed of the same air as the Westerly or North- Westerly current of the rear of the system which flows over the top of the South-Westerly or Southerly wind of the surface ? This and other questions must be decided by future research. This current too is of some importance in view of the methods of forecasting of M. Gabriel Guilbert. M. Guilbert maintains 3 that divergent winds, that is winds which blow away from areas of low pressure, indicate the advance of a low pressure system. A divergent wind in the upper air appears to be the very frequent forerunner of an approaching depression, and it may well be that the upper wind may descend to the surface in places and give a divergent component to the surface wind, as observed by M. Guilbert. A word must be said of the frequency with which reversals or great changes of wind direction are associated with thunderstorms ; again and again as will be seen in the account of the ascents under classes (d) and (e) thunderstorms occur over some part of the country when these conditions are revealed. It is indeed possible that some such conditions are necessary for the production of a thunderstorm. It is difficult to see how the difference of potential between one layer of the atmosphere and another can be maintained unless wind currents from different directions are bringing masses of air at different potentials near together ; but if say a polar wind 1 Forecasting Weather, by Dr W. N. Shaw, F.R.S., pp. 211, 212. 2 "The Statical changes of pressure and temperature in a column of air that accompany changes of pressure at the bottom," Quarterly Journal Royal Meteorological Society, Vol. xxxvin. p. 41. 3 NouveUe Methode de Prevision du Temps. RELATION OF VERTICAL WIND DISTRIBUTION TO SURFACE PRESSURE DISTRIBUTION 77 FIG. 46. 78 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER is flowing under an equatorial and the air in the two regions is at different potentials a constant difference of potential will be maintained as long as the two opposite winds persist. A curious fact has been noticed at Ditcham, namely that with light Northerly winds, especially in summer, and when a reversal may be expected in the upper air the guns from ships firing in the Channel are heard far more plainly even than when there is a Southerly wind. The explanation is probably that the sound waves that travel upward in a Northerly direction are refracted x>n entering the upper Southerly current, and are finally refracted down to the surface. This perhaps accounts for the fact often noticed that guns arid fog signals may be heard at a great distance from the origin of the sounds, but not in intermediate places. It may also explain the popular superstition that the firing of heavy guns brings rain ; the conditions that cause the firing to be heard plainly at great distances, being also those that bring heavy rains of a thunderstorm type. It has been mentioned in Chapter I that an equatorial wind flowing over a polar one may be a condition favourable for the formation of cloud and rain, and therefore unfavourable for the observation of balloons. The very common type of weather of North-Easterly winds accompanied by cloud and rain is probably of this character. On several occasions ballons sondes have been sent up with a North-Easterly or Easterly wind in cloudy weather and have been found at places from North-West to North-East of the station. A good instance of this occurred during the ascents of June 3rd, 1909, when the balloons started to the South- West and were found near Norwich, Maidenhead, Didcot and Oxford. Another example is in the ascents of May 18th to 19th, 1910, when several balloons started in a Westerly direction but were recovered from the Midlands. On May 19th there was extraordinary audibility of the sounds of trains on the Portsmouth railway. It has been noticed at Ditcham on occasions when there is a Southerly or South -Westerly current flowing over a North -Easterly one that the trace of the microbarograph shows marked fluctuations of pressure. So much is this the case that with a North-Easterly wind with low clouds and rain and fluctuations of the microbarograph no hesitation is felt in sending up ballons sondes although the sea c,ast lies not many miles to the South ; in such cases the balloons almost always enter a reverse current and fall in the midland counties. Further observations of this class are needed, and observations of the upper clouds would be instructive when these can be glimpsed through temporary rifts in the lower cloud sheet. The conditions will often be found to be those of class (e%) (Fig. 47) with a low pressure system to the South- West, and an upper wind blowing from the centre of the depression ; this seems to have been the case on June 3rd, 1909. Further observations of pilot balloons are greatly to be desired, and if observa- tions could be made from several stations simultaneously much more might be learned about the movements of the atmosphere. Four stations in the British Islands, say one in the South of England, one in the North-East of England or RELATION OF VERTICAL WIND DISTRIBUTION TO SURFACE PRESSURE DISTRIBUTION 79 surface wind FIG. 47. Relation of winds to the distribution of pressure at sea level in class (e 2). the South-East of Scotland, one in the South- West of Ireland, and one in the West or the North- West of Scotland, at which observations of pilot balloons should be made on every day when conditions allowed would probably result in a very great addition to our knowledge of the structure of the_atmosphere. 80 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER GENERAL TABLE OF THE ASCENTS. The first column gives the date, the second the time of the ascent. The third column gives the greatest height observed in kilometres. The fourth column gives the class (see p. 35). The fifth column gives the point of fall in those cases in which the balloon was recovered in kilometres and the direction in degrees from North through East, South and West. Day and hour of ascent Greatest observed height. Kilometres Class Point of fall Distance. Kilometres Direction in degrees from N. 1907 Jan. 25 11.42 a.m. 1-5 d 37 90 Feb. 1 4.6 p.m. 1-9 a 2 11.45 a.m. 1-2 m 4 10.44 1-2 m 8 0.45 p.m. 0-9 el 14 2.9 1-2 m 23 3.43 1-3 m ,, 26 noon 1-2 c Mar. 28 5.55 p.m. 2-0 m 137 360 29 6.25 3-0 a 309 10 30 6.19 1-3 d Apr. 1 6.12 3-9 d 39 1 ,, 2 5.46 2-2 b 4 6.34 3-5 c 5 4.40 4-0 a 6 6.13 4-3 a 8 6.14 3-3 a , 9 6.20 1-6 d 14-5 160 , 13 5.56 5-9 b 169 300 , 15 10.50 a.m. 1-6 d 7 315 , 15 6.36 p.m. 1-5 d , 16 6.27 ., 1-7 c , 17 4.49 1-9 a 18 6.40 5 el 19 5.59 3 el 20 4.52 4 el May 3 2.33 1-2 b 10 10.22 a.m. 2-2 b 10 6.32 p.m. 3-3 b (150 22) 11 5.40 3-2 b 13 6.27 1-0 c 14 6.53 2-2 d (40 260) 16 7.17 4-2 a 17 6.43 1-9 d 18 6.53 2-4 b 20 4.49 , 1-5 m 21 6.8 , 3 d 24 7.3 , 7-0 b 25 7.13 , 2-7 d i 27 6.11 , 6-5 d 1 29 7.11 , 1-5 c 1 Totland Bay. TABLE OF ASCENTS 81 Day and hour of ascent Greatest observed height. Kilometres Class Point of fall Distance. Kilometres Direction in degrees from N. 1907 June 6 3.9 p.m. 2-2 b 6 6.38 3-7 b 8 7.15 5-5 b 17 7.15 1-6 m 20 6.40 2-4 b 106 55 29 7.11 2-5 d * 1 July 1 3.15 10 d (66 75) 1 1 4.7 2-1 d 22 0.47 1-8 d 24 3.7 3-5 e2 Aug. 5 2.24 1-2 d 35 31 5 7.19 2-7 b 23 7.5 3-5 d 28 11.38a.m. 2-5 d 28 6.23 p.m. 6-0 d Sept. 1 10.22 a. m. 4-8 b * 4 5.4 p.m. 2-0 el 6-5 73 23 2.17 ,, 7-8 el 9 75 27 4 38 ,, _! 1 t.OC ,, 4-3 b 46 309 Oct. 14 5.24 1-0 d Dec. 26 0.4 4-2 c 24-5 270 1908 Jan. 2 10.43 a.m. 4-2 c 2 3.47 p.m. 2-2 c 3 10.49 a.m. 1:6 c 64 268 3 11.19 2-7 c 61 267 3 3.59 p.m. 5-3 c 4 11.2 a.m. 1-4 c 11 3.38p.m. 6-3 el 10-5 240 17 2.25 2-8 a Feb. 1 4.32 2-0 a 2 4.6 3-7 b 4 4.48 2-5 b 5 10.20 a.m. 2-0 b 8 11.31 2-9 a 8 4.25 p.m. 3-4 a 12 2.35 6-1 b 55 358 13 3.8 3-8 b 14 4.38 5-2 el 26 10.28 a.m. 2-3 b 18-5 130 26 11.5 2-2 b 14-5 112 Mar. 3 10.12 3-6 d 13 53 5 11.30 3-4 el ,, 5 5.5 p.m. 2-0 el 11 60 11 5.28 2-6 b 12 5.52 3-2 d 8-5 164 13 5.40 2-0 d 14 10.37 a.m. 1-7 c 3 315? 21 5.50 p.m. 5-0 d 24 10.19 a.m. 4-7 el 8-5 112 27 11.4 2-5 el 27 5.41 p.m. 3-3 c 28 5.40 3-0 el * Ballon sonde. 1 Chobham Common, C. 11 82 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER Day and hour of ascent Greatest observed height. Kilometres Class Point of fall Distance. Kilometres Direction in degrees from N. 1908 Mar. 30 11.29 a.m. 1-1 m 9 65 Apr. 1 5.50 p.m. 2-0 el 1 6.13 1-3 el 4 4.14 2-0 m 8 10.39 a.m. 4-2 a 8 6.46 p.m. 4-7 a 9 11.8 a.m. 3-9 c 11-5 180 9 5.53 p.m. 4-8 d 294 120 16 11.12 a.m. 2-5 c 89 268 29 3.57 p.m. 6-2 el May 2 0.48 6-7 e2 21 35 11 0,37 3-9 b 18 6.37 7-0 a 114 80 21 7.18 3-0 b 23 11.0 a.m. 3-0 a 27 10.16 36 b 30 210 29 8.12 2-5 c 30 7.2 p.m. 7-0 a 56 300 June 1 7.32 4-6 b 97 333 2 7.15 6-3 d 66 21 3 10.19 a.m. 4-9 e2 3 6.54 p.m. 11-6 e2 4 7.0 1-8 c 5 6.54 5-2 b 10 7.15 6-4 el 22 6.36 13-0 el 45 212 23 7.0 . 3-7 a 26 7.18 5-9 a *July 27 7.19 8-6 d 183 44 28 7.0 13-0 e l;f 145 170 29 9.27 a.m. 4-4 a 29 5.3 p.m. 5-3 a * 29 7.0 13-3 e 1; ,, 30 8.6 a.m. 5-4 el 30 0.36 p.m. 4-0 el * 30 7.0 11-8 el;f 31 7.0 13-4 e l;f Aug. 1 8.10 a. m. 3-4 el 1 5.53 p.m. 4-2 el 1 7.5 4-0 el 2 5.54 4-1 el 2 7.30 4-8 el Sept. 30 8.36 a.m. 3-7 d 30 11.17 6-7 d * 30 4.31 p.m. 16-0 b;f 105 40 Oct. 1 8.15 a.m. 3-6 b * 1 4.20p.m. 17-6 b;f 114 18 2 8.20 a.m. 3-9 b 45 11 * 2 4.20p.m. 16-2 b;f 57 10 3 10.45 a.m. 3-2 b 3 3.25 p.m. 4-2 b Nov. 3 0.47 6-5 d 6 10.59 a.m. 9-5 d 7 3.25 p.m. 6-1 d Ballon sonde. TABLE OF ASCENTS 83 Day and hour of ascent Greatest observed height. Kilometres Class Point of fall Distance. Kilometres Direction in degrees from N. 1908 Nov. 8 4.26 p.m. 4-5 d 40 266 16 10.47 a.m. 5-5 e 1 1909 Jan. 12 10.58 a.m. 0-9 b 12 2.22 p.m. 2-1 b * 12 3.54 3-2 b 15 11.40a.m. 3-0 b 19 10.34 2-0 m 19 0.33 p.m. 2-9 m 20 10.20 a.m. 2-9 c 30 3.21 p.m. 3-0 b Feb. 5 4.28 3-5 b 6 4.48 1-7 b 7 4.28 4-1 el 11-5 326 12 5.8 4-2 c 13 5.6 3-6 b 14 4.50 3-5 b 15 2.21 5-0 b 17 8.17 a.m. 6-0 el 18 4.43 p.m. 7-9 el 19 10.2 a.m. 2-6 c 39 309 19 4.44 p.m. 6-4 a 20 10.26 a.m. 5-0 a 56 352 20 4.46 p.m. 6-0 a 37 353 21 2.35 5-6 c 22 11.38a.m. 3-3 c 22 4.52 p.m. 5-8 a 35 231 23 4.30 3-7 a 24 4.45 5-5 a 36 225 Mar. 5 5.13 4-5 e 1 Apr. 19 8.15a.m. 4-7 b 39 13 21 7.58 5-2 a 26 2.41 p.m. 2-9 b 87 29 May 2 7.7 5-0 el ,, 4 8.15 a.m. 3-0 b 4 7.4 p.m. 5-5 c ,, 5 8.10 a.m. 2-4 c 20 282 5 2.54 p.m. 6-3 a * 5 6.43 10-2 a 94 292 6 10.33 a.m. 2-8 c 197 310 6 0.27 p.m. 8-0 a * 6 6.25 12-8 a;f 81 305 7 2.52 13-0 c;f 41 290 * 7 6.29 15-0 c;f 46 272 16 6.22 3-0 e2 31 0.13 4-3 b June 1 10.23 a.m. 2-4 d ,, 3 0.5 p.m. 1-2 c 3 0.59 3-6 e2 21 7.0 4-2 b Aug. 5 0.8 2-8 a 5 2.30 8-0 a 5 6.33 185 a;f 19 254 1910 *Mar. 3 4.30 p.m. 15-2 c;f 20 265 * Ballon sonde. 112 84 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER TABLE OF THE ASCENTS GIVING THE WIND VELOCITIES IN METRES PER SECOND AND THE WIND DIRECTIONS IN DEGREES FROM NORTH (THROUGH EAST, SOUTH, AND WEST) FOR EVERY HALF KILOMETRE. Figures in square brackets indicate that the height opposite the figures has been nearly but not quite reached. All ascents are from Ditcham and observed with one theodolite unless otherwise stated. The cause of cessation of observations when it has been recorded, is given by the words distance, clouds, etc., in the notes at the foot of the column of observations. CLASS (a). ''Solid" current; little change in velocity or direction, wind reaches the gradient value and does not increase very much at greater altitudes. Gradient Surface Feb. 1, 1907 4.6 p.m. vel. dir. Mar. 29, 1907 6.25 p.m. vel. dir. 6-0 180 April 5, 1907 4.40 p.m. vel. dir. 11-2 220 April 6, 1907 6.13 p.m. vel. dir. 16-0 270 April 8, 1907 6.14p.m. vel. dir. 11-2 280 April 17, 1907 4.49 p.m. vel. dir. 8-0 40 Gradient Surface 7-3 10 6-0 130 4-5 235 6-0 275 6-5 250 5-0 355 0-5 8-5 25 8-5 175 9-5 245 11-5 270 12-0 260 9-5 330 0-5 1-0 9-8 25 8-5 185 9-5 230 16-0 265 14-5 280 8-5 345 1-0 1-5 9-0 40 9-7 190 6-2 225 19-5 260 12-5 290 8-0 350 1-5 2-0 9-0 10 9-2 205 7-5 220 15-0 260 9-0 290 9-0 350 2-0 2-5 11-0 185 7-0 225 14-5 255 9-0 280 2-5 3-0 11-0 190 11-0 230 14-5 260 11-5 275 3-0 3-5 12-0 225 14-5 255 3-5 4-0 9-5 265 14-5 260 4-0 distance TABLE OF ASCENTS CLASS (a) continued. 85 May 16, 1907 7.17 p.m. Jan. 17, 1908 2.25 p.m. Feb. 1, 1908 4.23 p.m. Feb. 8, 1908 11.31 a.m. Feb. 8, 1908 4.25 p.m. April 8, 1908 10.39 a.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 9-0 350 22-0 240 12-0 355 9-5 315 8-0 320 Gradient Surface 4-5 360 6-5 350 5-5 315 4-5 305 6-5 10 0-5 9-0 355 25-0 235 10-5 350 7-0 315 6-5 330 6-5 25 0-5 1-0 9-0 335 21-0 230 14-0 350 6-0 325 8-5 335 7-0 35 1-0 1-5 9-5 340 14-5 230 16-0 355 8-5 325 7-0 320 5-0 15 1-5 2-0 8-0 320 23-0 220 16-5 360 9-0 325 11-0 305 7-0 25 2-0 2-5 8-0 320 23-0 220 9-0 320 1 11-5 330 10-0 40 2-5 3-0 10-0 310 [10-0 315] 10-5 320 7-5 40 3-0 3-5 9-5 315 7-0 60 3-5 4-0 13-0 310 Totland Bay 1 Decrease of velocity to 6*0 2 theodolites to 2-4 km. between 2 km cloud . and 2-5 km. 9-5 55 4-0 April 8, 1908 6.46 p.m. May 18, 1908 6.37 p.m. May 23, 1908 11.0a.m. May 30, 1908 7.2 p.m. June 23, 1908 7.0 p.m. June 26, 1908 7.18 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 12-0 270 6-0 90 6-0 330 10-0 90 Gradient Surface 5-0 30 3-0 270 5-0 315 3-0 80 4-5 70 4-0 85 0-5 7-5 25 6-0 300 7-0 320 4-0 90 4-0 30 9-0 60 0-5 1-0 8-0 20 9-5 325 5-5 355 5-0 130 6-0 355 9-5 35 1-0 1-5 9-5 25 5-0 310 5-5 355 5-0 95 7-5 360 10-0 55 1-5 2-0 9-5 20 8-6 270 6-0 5 4-5 110 7-5 15 12-0 60 2-0 2-5 8-5 360 10-0 270 7-0 355 6-0 105 5-5 5 7-5 60 2-5 3-0 7-5 25 9-5 260 [8-5 15] 7-5 105 5-0 20 7-5 70 3-0 3-5 6-0 360 9-5 255 7-5 135 6-0 15 9-0 75 3-5 4-0 3-5 350 9-0 255 5-0 130 13-0 80 4-0 4-5 6-0 350 10-0 270 4-0 145 12-5 85 4-5 5-0 10-5 255 4-0 140 12-0 80 5-0 5-5 13-0 240 6-0 120 11-0 85 5-5 6-0 11-0 245 8-0 135 [8-5 90] 6-0 6-5 9-5 240 7-0 135 6-5 7-0 12-5 255 [7-5 125] 7-0 2 theodolites to 3 km. 2 theodolites cloud 86 THE STRUCTURE OF THK ATMOSPHERE IN CLEAR WEATHER CLASS (a) continued. Julv 29, 1908 July 29, 1908 , Feb. 19, 1909 Feb. 20, 1909 Feb. 20, 1909 Feb. 22, 1909 9.27 a.m. 5.3 p.m. 4.44 p.m. 10.26 a.m. 4.46 p.m. 4.52 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient 30? 4-0 360 12-0 155 10-0 160 12-0 170 Gradient Surface 5 - 5 30 3-0 355 6-5 105 8-0 135 5-0 115 5 - 80 Surface 0-5 8-0 15 3-5 5 12-0 130 9-5 140 7-0 150 5-0 70 0-5 1-0 6-5 330 4-0 25 17-5 145 14-0 160 8-5 150 3-5 85 1-0 1-5 6-0 20 3-0 355 17-0 140 10-0 160 9-0 155 6-5 100 1-5 2-0 5-5 15 7-0 5 13-0 150 9-0 175 7-0 165 5-0 80 2-0 2-5 6-0 20 7-0 10 11-0 145 9-0 160 7-0 1GO 7-0 65 2-5 3-0 9-0 20 \ 6-5 10 10-5 145 9-5 175 8-5 165 7-0 35 3-0 3-5 8-0 5 6-0 20 10-0 145 11-0 175 9-5 160 8-0 40 3-5 4-0 5-0 15 7-0 360 10-0 145 11-0 155 10-5 185 8-0 40 4-0 4-5 6-5 355 10-0 145 10-5 170 10-0 195 9-5 35 4-5 5-0 8-5 360 8-0 145 11-0 170 8-0 185 9-0 30 5-0 5-5 7-5 140 9-0 190 8-5 15 5-5 6-0 8-5 130 9-0 185 [7-0 25] 6-0 6-5 [10-0 135] 6-5 2 theodolites 2 theodolites distance 2 theodolites to 5-1 I km. to 4-^ ' km. hurst] CLASS (a). Subclass no current. Aug. 5, 1909 Aug. 5, 1909 0.8 p.m. 2.30 p.m. vel. dir. vel. dir. Gradient Surface 1-0 60 2-0 210 0-5 1-5 110 3-0 175 1-0 2-0 50 5-0 40 1-5 2-0 60 2-0 70 2-0 3-5 50 4-0 75 2-5 3-5 30 4-5 40 3-0 4-0 30 3-5 3-0 70 4-0 3-0 55 4-5 2-0 40 5-0 3-0 50 5-5 2-0 80 6-0 3-0 GO 6-5 2-5 55 7-0 2-5 20 7-5 1-0 100 8-0 [1-0 160] TABLE OF ASCENTS 87 CLASS (a) continued. Gradient Surface Feb. 23, 1909 4.30 p.m. vel. dir. 11-0? 70? Feb. 24, 1909 4.45 p.m. vel. dir. 9-0 90 April 21, 1909 7.58 a.m. vel. dir. 9-0 170 May 5, 1909 2.54 p.m. vel. dir. 17-0 100 May 5, 1909 6.43 p.m. vel. dir. 19-0 120 May 6, 1909 0.27 p.m. vel. dir. 21-0 120 Gradient Surface 6-5 15 5-0 70 6-5 110 14-0 65 5-0 85 0-5 10-0 35 8-0 55 8-5 130 13-0 65 6-0 90 0-5 1-0 9-0 40 7-5 35 3-5 160 20-0 90 15-0 95 13-0 110 1-0 1-5 11-0 35 7-0 55 6-0 140 15-0 110 16-0 105 12-5 115 1-5 2-0 11-5 30 5-5 25 7-0 150 14-5 120 17-5 115 11-0 125 2-0 2-5 11-5 25 3-5 40 7-5 130 16-0 110 17-5 105 12-5 130 2-5 3-0 12-0 25 6-5 35 8-0 135 16-0 110 17-0 110 12-0 130 3-0 3-5 11-5 20 7-0 40 8-5 115 16-0 125 12-0 105 8-5 125 3-5 4-0 9-0 55 7-0 125 17-0 125 15-0 110 10-0 125 4-0 4-5 9-0 45 9-0 145 16-5 130 18-5 120 11-0 130 4-5 5-0 10-5 50 10-0 160 17-0 130 180 125 13-0 125 5-0 5-5 10-5 45 17-0 130 17-0 130 12-0 120 5-5 6-0 17-0 130 14-0 125 10-0 130 6-0 6-5 13-0 115 11-0 125 6-5 7-0 14-0 105 12-0 130 7-0 7-5 14-0 125 14-0 125 7-5 8-0 15-0 130 18-5 125 8-0 8-5 19-0 125 8-5 9-0 15-0 130 9-0 9-5 i 13-0 130 9-5 10-0 - 17-0 110 10-0 2 theodolites burst burst to 2-4 km. 88 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CLASS (6). Considerable increase in velocity. April 2, 1907 5.46 p.m. April 13, 1907 5.56 p.m. May 3, 1907 2.33 p.m. May 10, 1907 10.22 a.m. May 10, 1907 6.32 p.m. May 11, 1907 5.40 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 11-5 180 7-0 130 20-0 255 12-5 205 9-0 200 4-0 190 Gradient Surface 3-5 130 6-0 85 5-5 260 3-0 180 2-5 105 2-0 90 0-5 10-0 185 6-0 90 12-0 260 9-0 205 12-5 175 5-0 175 0-5 1-0 15-0 185 6-0 135 20-0 265 10-0 205 14-0 185 12-0 165 1-0 1-5 19-0 195 7-0 165 17-0 200 17-0 190 11-0 185 1-5 2-0 22-0 195 10-0 160 19-0 200 15-5 190 15-0 190 2-0 2-5 10-0 130 15-0 190 19-5 190 2-5 3-0 9-0 125 20-0 185 19-0 190 3-0 3-5 7-0 140 3-5 4-0 12-0 140 4-0 4-5 15-0 140 4-5 5-0 18-0 145 5-0 5-5 16-0 140 5-5 cloud 6-0 [17-0 140] Totland BayjTotland Bay Totland Bay 6-0 May 18, 1907 6.53 p.m. May 24, 1907 7.3 p.m. June 6, 1907 3.9 p.m. June 6, 1907 6.38 p.m. June 8, 1907 7.15 p.m. June 20, 1907 6.40 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 6-0 30 7-0 150 14-0 280 15-0 270 9-0 210 12-0 250 Gradient Surface 4-5 10 3-5 235 9-0 285 6-5 280 2-5 180 4-5 225 0-5 10-5 20 9-0 230 13-0 280 13-0 285 6-0 180 18-0 240 0-5 1-0 12-0 25 12-0 230 20-0 280 13-5 295 12-0 205 20-0 1 240 1-0 1-5 10-0 20 10-0 230 20-0 270 13-5 295 16-0 215 19-0 240 1-5 2-0 8-0 25 10-0 205 20-0 275 14-0 300 11-5 215 24-0 235 2-0 2-5 10-0 205 16-0 305 12-0 225 [30-0 235] 2-5 3-0 11-0 180 19-0 305 11-0 215 3-0 3-5 9-5 175 21-5 310 14-0 220 3-5 4-0 10-5 175 17-0 230 4-0 4-5 12-5 180 16-0 240 4-5 5-0 13-5 175 18-0? 235 5-0 1 vel. about 25'0 at 1'7 km. Continued on next page. TABLE OF ASCENTS 89 CLASS (6) continued. May 18, 1907 6.53 p.m. May 24, 1907 7.3 p.m. June 6, 1907 3.9 p.m. June 6, 1907 6.38 p.m. June 8, 1907 7.15 p.m. June 20, 1907 6.40 p.m. vel. dir. vel. dir. 5-5 15-0 175 17-0 240 5-5 6-0 14-0 180 6-0 6-5 16-0 180 6-5 7-0 18-0 190 7-0 Totland Bay cloud Totland Bay distance distance lost behind house Aug. 5, 1907 7.19 p.m. Sept. 1, 1907 10.22 a.m. Sept. 27, 1907 4.38 p.m. Feb. 2, 1908 4.6 p.m. Feb. 4, 1908 4.48 p.m. Feb. 5, 1908 10.20 a.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 8-0 310 4-0 360 9-0 100 7-5 355 11-0 5 11-5 360 Gradient Surface 2-0 245 5-5 25 6-5 90 5-5 350 7-0 355 2-0 360 0-5 3-0 290 4-5 20 11-5 115 8-5 355 13-0 355 5-0 15 0-5 1-0 2-5 315 4-5 290 15-0 130 9-0 30 13-0 10 10-0 25 1-0 1-5 3-0 240 5-0 300 14-0 140 9-5 25 15-0 20 10-0 25 1-5 2-0 12-0 215 7-0 295 13-0 140 17-0 15 19-5 20 13-0 25 2-0 2-5 13-0 205 8-0 290 12-0 145 22-5 25 2-5 3-0 13-0 285 13-0 140 27-5 30 3-0 3-5 16-0 280 15-0 125 28-0 25 3-5 4-0 16-0 270 16-0 130 4-0 4-5 19-5 265 4-5 cloud 2 theodolites to 1-8 km. lost owing to quick movement Feb. 12, 1908 2.35 p.m. Feb. 13, 1908 3.8 p m. Feb. 26, 1908 10.28 a.m. Feb. 26, 1908 11.5 a.m. Mar. 11, 1908 5.28 p.m. May 11, 1908 0.37 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 8-0 150 7-0 190 12-5 295 12-5 295 5-0 320 Gradient Surface 4-0 135 4-0 170 7-5 295 6-5 295 5-0 310 3-0 145 0-5 7-0 135 5-0 250 9-5 310 7-5 300 6-0 325 2-5 145 0-5 1-0 6-5 170 9-0 270 6-0 295 8-5 290 5-5 345 6-0 175 1-0 Continued on next page. c. 12 90 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CLASS (b) continued. Feb. 12, 1908 2.35 p.m. Feb. 13, 1908 3.8 p.m. Feb. 26, 1908 10.28 a.m. Feb. 26, 1908 11.5 a.m. Mar. 11, 1908 5.28 p.m. May 11, 1908 0.37 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. 1-5 5-0 145 6-0 230 9-0 295 9-0 285 8-0 350 6-0 235 1-5 2-0 6-0 200 7-5 210 15-0 330 15-0 320 10-5 340 7-0 220 2-0 2-5 9-0 185 11-0 185 [18-5 325] 11-0 315 9-5 215 2-5 3-0 9-0 180 15-0 210 10-0 205 3-0 3-5 8-5 170 17-0 190 10-0 195 3-5 4-0 8-0 185 [19-0? 195] 4-0 4-5 11-5 190 4-5 5-0 13-0 190 5-0 5-5 14-5 185 5-5 6-0 13-0 180 6-0 2 theodolites 2 theodolites 2 theodolites 2 theodolites 2 theodolites to 2-2 km. distance to 2-2 km. burst burst cloud cloud May 21, 1908 7.18 p.m. May 27, 1908 10.16 a.m. June 1, 1908 7.32 p.m. June 5, 1908 6.54 p.m. Oct. 1, 1908 8.15 a.m. Oct. 2, 1908 8.20 a.m. vel dir. vel. dir. vel. dir. vel. dir. vel dir. vel. dir. Gradient Surface 12-0 300 5-0 65 2-0 150 6-0 360 6-0 180 4-0 180 Gradient Surface 7-5 290 2-0 55 3-5 70 4-5 325 3-0 30 2-5 100 0-5 10-5 295 3-5 50 3-5 120 8-5 335 5-0 75 5-0 140 0-5 1-0 11-5 265 5-5 45 6-5 160 6-0 330 4-0 90 5-5 170 1-0 1-5 10-0 250 8-5 35 12-0 170 7-0 305 4-5 75 8-0 170 1-5 2-0 10-0 240 8-5 30 11-5 160 11-5 315 7-0 80 3-0 165 2-0 2-5 14-0 235 8-5 30 12-0 150 13-0 290 9-5 90 5-0 170 2-5 3-0 25-0 230 5-0 15 17-0 145 12-0 300 8-5 85 9-0 180 3-0 3-5 8-5 15 18-0 150 13-5 300 12-0 90 10-0 170 3-5 4-0 180 155 16-0 300 8-5 185 4-0 4-5 15-0 155 150 295 4-5 5.0 12-5 290 5-0 distance TABLE OF ASCENTS 91 CLASS (b] continued. Oct. 3, 1908 10.45 a.m. Oct. 3, 1908 3.25 p.m. Jan. 12, 1909 10.58 a.m. Jan. 12, 1909 2.22 p.m. Jan. 12, 1909 3.50 p.m. Jan. 15, 1909 11.40 a.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 7-0 165 5-0 180 12-0 285 12-5 290 14-0 270 Gradient Surface 2-0 115 2-5 125 8-0 270 8-0 290 6-0 295 4-5 270 0-5 7-0 130 5-5 160 16-0 280 10-0 295 10-0 290 7-5 275 0-5 1-0 12-5 155 9-0 175 [20-5 290J 15-0 305 12-0 300 14-0 285 1-0 1-5 9-5 150 10-0 165 20-0 295 17-5 300 14-0 275 1-5 2-0 11-5 155 7-5 175 29-0 300 24-0 300 14-0 270 2-0 2-5 8-5 160 8-0 175 26-0 300 17-0 280 2-5 3-0 7-0 155 7-0 175 24-0 295 26-0 270 3-0 3-5 8-5 170 3-5 4-0 12-0 165 4-0 distance 2 theodolites to 2-5 km. Jan. 30, 1909 3.21 p.m. Feb. 5, 1909 4.28 p.m. Feb. 6, 1909 4.48 p.m. Feb. 13, 1909 5.6 p.m. Feb. 14, 1909 4.50 p.m. Feb. 15, 1909 2.21 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 11-0 350 18-0 330 5-0 360 11-0 65 10-0 360 Gradient Surface 7-5 355 9-0 310 4-0 350 7-5 30 5-0 355 5-0 345 0-5 10-0 345 13-0 315 5-5 360 14-0 35 9-0 10 8-5 355 0-5 1-0 15-5 345 17-0 325 6-0 355 16-0 65 9-5 10 16-0 350 1-0 1-5 18-0 345 20-0 335 7-5 350 14-5 60 8-5 15 14-5 320 1-5 2-0 19-0 345 22-0 330 15-0 50 9-5 20 10-0 325 2-0 2-5 22-0 345 25-0 330 13-5 30 10-5 15 16-0 330 2-5 3-0 25-0 345 29-0 325 16-0 35 12-0 5 18-5 335 3-0 3-5 30-0 325 17-5 30 13-0 30 16-0 345 3-5 4-0 15-0 25 9-5 350 4-0 4-5 19-0 25 13-5 325 4-5 5-0 25-0 325 5-0 2 theodolites from 1'3 km. to 2-6 km. 2 theodolites to 2-4 kra. 2 theodolites burst 1 distance distance distance 122 92 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CLASS (b) continued. April 19, 1909 8.15 a.m. April 26, 1909 2.41 p.m. May 4, 1909 8.15 a.m. May 31, 1909 0.13 p.m. June 21, 1909 7.0 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 7-0 170 8-0 185 7-0 160 7-0 225 8-0 200 Gradient Surface 4-0 120 5-5 160 5-5 135 5-0 160 5-5 165 0-5 8-5 135 10-0 160 10-0 140 10-0 175 6-5 175 0-5 1-0 7-0 155 10-5 200 13-0 125 9-0 185 9-5 215 1-0 1-5 7-5 185 16-0 220 13-0 125 4-5 200 12-5 210 1-5 2-0 7-0 185 16-5 220 12-5 125 7-5 235 14-0 210 2-0 2-5 8-5 195 14-0 215 5-5 120 10-0 245 13-0 200 2-5 3-0 8-0 205 [15-5 215] 9-0 125 11-0 230 14-0 200 3-0 3-5 6-0 225 13-5 225 16-0 195 3-5 4-0 7-0 205 12-0 230 22-5 205 4-0 4-5 11-0 200 4-5 burst theodolite accidentally moved CLASS (c). Feb. 26, 1907 noon April 4, 1907 6.34 p.m. April 16, 1907 6.27 p.m. May 13, 1907 6.27 p.m. May 29, 1907 7.11 p.m. Dec. 26, 1907 0.4 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 4-5 90 6-0 50 9-0 345 11-0 140 11-0 135 Gradient Surface 2-5 30 4-5 80 3-0 360 4-5 320 3-0 70 7-5 80 0-5 6-5 35 6-5 80 8-5 10 11-0 340 8-5 70 10-0 100 0-5 1-0 8-0 45 6-0 70 5-0 35 0-5 360 4-5 125 17-5 110 1-0 1-5 [4-5 45] 5-5 65 4-5 25 1-5 80 10-0 115 1-5 2-0 4-0 70 5-0 90 2-0 2-5 5-0 50 3-5 70 2-5 3-0 4-0 360 2-5 355 3-0 3-5 3-5 5 6-0 55 3-5 4-0 5-0 60 4-0 cloud cloud Totlaud Bay cloud Totland Bay cloud distance TABLE OF ASCENTS 93 CLASS (c) continued. Jan. 2, 1908 10.43 a.m. Jan. 2, 1908 3.47 p.m. Jan. 3, 1908 10.49 a.m. Jan. 3, 1908 11.19 a.m. Jan. 3, 1908 3.59 p m. Jan. 4, 1908 11.2 a.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 19-0 105 16-0 105 20-0 105 20-0 105 16-5 100 19-0 90 Gradient Surface 9-5 55 10-0 50 7-0 55 8-5 55 9-0 50 7-5 70 0-5 10-0 65 16-0 60 11-0 70 17-5 65 16-5 60 15-0 75 0-5 1-0 22-5 80 14-0 80 16-0 85 20-0 85 21-5 85 14-0 95 1-0 1-5 7-0 85 12-0 90 11-5 115 20-0 105 13-0 105 14-0 100 1-5 2-0 10-0 65 10-0 90 8-0 120 11-0 100 2-0 2-5 9-0 50 10-0 95 10-0 80 2-5 3-0 8-0 35 7-5 110 3-0 3-5 10-5 40 6-5 105 3-5 4-0 10-5 30 8-0 100 4-0 4-5 9-5 100 4-5 5-0 distance lost behind 11-0 90 lost behind 5-0 tree tree Mar. 14, 1908 10.37 a.m. Mar. 27, 1908 5.41 p.m. April 9, 1908 11.8 a.m. April 16, 1908 11.12 a.m. May 29, 1908 8.12 a.m. June 4, 1908 7.0 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 8-0? 115? 13-0 230 25-0 90 20-0 70 9-0 360 Gradient Surface 6-0 100 5-0 205 2-0 25 17-0 55 9-0 40 12-5 25 0-5 5-5 125 8-5 215 4-0 65 18-5 65 12-0 50 14-5 25 0-5 1-0 2-0 115 11-5 235 7-5 10 20-5 80 12-5 70 6-5 20 1-0 1-5 0-5 65 10-0 250 5-0 5 20-0 90 14-5 80 3-5 330 1-5 2-0 8-5 250 5-0 350 14-0 90 11-5 80 2-0 2-5 7-5 260 5-0 25 13-0 65 2-5 3-0 7-0 280 3-5 355 3-0 3-5 4-0 325 3-5 4-0 [3-5 315] 4-0 2 theodolites burst 2 theodolites to 0-7 km. 2 theodolites to 2 -Okm. 94 THE STRUCTURE OF THE ATMOSPHERE IX CLEAR WEATHER CLASS (c) continued. Jan. 20, 1909 Feb. 12, 1909 Feb. 19, 1909 Feb. 21, 1909 Feb. 22, 1909 May 4, 1909 10. 20 a.m. 5.8 p.m. 10.2 a.m. 2.35 p.m. 11.38 a.m. 7.4 p.m. vel dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient "'( J? 30? 23-0 60 11-0 145 10-5 155 13-0 150 Gradient Surface Surface ~-. ') 5 18-0 30 7-0 100 4-0 115 5-0 80 7-0 80 0-5 12-5 30 19-0 30 11-0 115 10-0 145 9-0 105 13-0 90 0-5 1-0 11-0 15 24-0 50 15-5 130 11-0 155 7-5 110 10-5 120 1-0 1-5 12-5 10 27-5 70 15-0 130 7-5 155 8-0 110 1 1 -5 1 40 1-5 2-0 14-( ) 25 24-0 75 13-0 130 j 7-0 140 7-5 110 12-0 135 2-0 2-5 12-( ) 40 22-5 60 8-5 140 5-0 110 6-0 70 11-0 130 2-5 3-0 [7-( ) 45] 11-5 65 6-0 120 4-0 40 7-0 105 3-0 3-5 12-0 45 6-5 130 8-0 100 3-5 4-0 17-5 55 6-0 170 9-0 105 4-0 4-5 3-5 185 7-5 90 4-5 5-0 5-0 170 7-0 100 5-0 5-5 5-0 175 5-0 115 5-5 distance 2 theodolites distance haze May 5, 1909 ! May <>, 1909 June 3, 1909 8.10a.m. 10.33a.m. 0.5p.m. vel. dir. vel. dir. vel. dir. Gradient 14'0 130 24'5? 120 22'0 80 Gradient Surface 5'5 90 4'5 85 lO'O 40 Surface 1 0-5 j 12-0 100 8-0 95 14-0 45 j 0-5 1-0 ll'O S5 18-5 105 12-0 95 1-0 1-5 12-0 90 13-5 115 1-5 2-0 11-0 100 18-5 115 2-0 2-5 10-0 130 2-5 burst? distance cloud TABLE OF ASCENTS CLASS (d). Reversals or great changes in direction. 95 Jan. 25, 1907 11.42 a.m. Mar. 30, 1907 6.19 p.m. April 1, 1907 6.12 p.m. April 9, 1907 April 15, 1907 6.20 p.m. 10.50 a.m. April 15, 1907 6.36 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 3-5 180 9-0 70 9-0 360 Gradient Surface 6-0? 25 4-5 255 3-0 185 2-5 170 8-5 35 6-0 10 0-5 6-0 10 4-0 250 3-0 210 2-5 170 9-0 40 7-0 20 0-5 1-0 1-0 310 2-0 325 3-5 150 0-5 35 2-0 205 2-0 100 1-0 1-5 3-0 230 [3-5 55] 3-5 115 1-0 190 2-0 235 [2-0 70] 1-5 2-0 3-5 130 2-0 2-5 4-0 150 2-5 3-0 1-5 220 3-0 3-5 4-0 210 3-5 4-0 [5-0 190] 4-0 at 0-3 km. direction was 55. veering began at 0-7 km. a veer of 75 between 2 '8 and between 0*7 and 1 - 6 km. the wind dir. sudden change of direction at wind veered to 145 at 1 - 2 km. then altitude became too great for the theodolite used 3 km. went from 190 through E., N. and S. to 165. cloud 0-8 km. accompanied by decreased velocity, cloud changed suddenly to 70. cloud May 14, 1907 6.53 p.m. May 17, 1907 6.43 p.m. May 21, 1907 6.8 p.m. May 25, 1907 7.13 p.m. May 27, 1907 6.11 p.m. June 29, 1907 7.11 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 7-0 10 5-0 90 6-0 95 Gradient Surface 1-5 30 4-5 305 1-0 135 5-0 50 4-0 180 6-0 340 0-5 4-0 40 7-5 310 4-0 90 10-0 95 2-5 350 11-0 330 0-5 1-0 1-0 325 7-5 310 3-5 140 7-5 140 5-5 20 4-0 275 1-0 1-5 7-0 250 7-5 255 1-5 215 5-5 180 6-0 15 1-5 275 1-5 2-0 8-0 245 [8-0 240] 2-5 265 6-0 205 2-5 360 0-5 200 2-0 2-5 10-0 230 6-5 185 3-5 310 2-5 180 2-5 3-0 1-0 205 3-0 3-5 1-5 225 3-5 4-0 3-5 180 4-0 4-5 8-0 225 4-5 5-0 7-5 220 5-0 5-5 8-0 220 5-5 Continued on next page. THE STRUCTURE OK THE ATMOSPHERE IN CLEAR WEATHER CLASS ((/) continued. May 14, 1907 6.53 p.m. May 17, 1907 ; May 21, 1907 6.43 p.m. 6.8 p.m. May 25, 1907 7.13 p.m. May 27, 1907 June 29. 1907 6.11p.m. 7.11p.m. vel. dir. 6-0 7-5 230 6-0 6-5 6-0 240 6-5 Totland Bay cloud Totland Bay Totland Bay wind backed Totland Bay cloud Totland Bay wind veered to 75 at between 0-7 km. surface and cloud 0-5 km. haze July 1, 1907 3.15 p.m. July 1, 1907 July 22, 1907 4.7 p.m. 0.47 p.m. Aug. 5, 1907 2.24 p.m. AUK. 23, 1907 7.5 p.m. Aug. 28, 1907 11.38 a.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 8-0? 15?' 6-0 310 7-0 330 3-0 155 Gradient Surface 5-5 40 4-5 20 6-0 175 1-0 60 4-5 115 0-5 4-0 40 6-0 10 4-0 175 1-0 360 7-0 160 4-5 125 0-5 1-0 2-5 25 6-0 25 3-0 240 4-0 240 6-0 185 6-0 145 1-0 1-5 7-0 10 6-0 265 3-0 155 4-5 130 1-5 2-0 6-0 350 9-5 170 4-0 135 2-0 2-5 8-5 130 6-5 235 2-5 3-0 6-0 140 3-0 3-5 4-0 120? 3-5 Chobham Chobham Common Common 2 theodolites reversal reversal reversal above above. above. 2 theodolites cloud cloud Aug. 28, 1907 6.23 p.m. Oct. 14, 1907 Mar. 3, 1908 5.24 p.m. 10.12 a.m. Mar. 12, 1908 5.52 p.m. Mar. 13, 1908 Mar. 21, 1908 5.40 p.m. 5.50 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 1-5 l(;r, 1 8-0 205 7-5 340 7-0? 2-0 50 6-0 190 Gradient Surface 2-0 170 1-5 215 3-0 75 5-5 135 0-5 2-0 100 10-0 330 2-5 210 2-0 260 2-5 50 4-0 130 0-5 1-0 5-0 240 10-5 l j j> J j Mar. 28, 1908 5.40 p.m. April 1, 1908 5.50 p.m. April 1, 1908 6.13 p.m. April 29, 1908 3.57 p.m. June 10, 1908 7.15 p.m. June 22, 1908 6.36p.m. vel. dir. 6-5 6-0 25 6-5 7-0 6-0 30 7-0 7-5 6-0 45 7-5 8-0 5-0 40 8-0 8-5 7-0 50 8-5 9-0 8-0 50 9-0 9-5 7-5 40 9-5 10-0 7-5 55 10-0 10-5 8-0 55 10-5 11-0 6-0 50 11-0 11-5 5-5 55 11-5 12-0 5-5 60 12-0 12-5 5-5 15 12-5 13-0 - 8-0 30 13-0 2 theodolites 2 theodolites 2 theodolites 2 theodolites to 4-3 km. burst cloud 1 cloud cloud burst July 30, 1908 July 30, 1908 8.6a.m. 0.36p.m. Aug. 1, 1908 8.10 a.m. Aug. 1, 1908 5.53 p.m. Aug. 1, 1908 7.5 p.m. Aug. 2, 1908 5.54 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 6-0 330 5-0 20 5-0 20 Gradient Surface 2-5 270 3-0 300 5-0 355 4-0 350 6-5 345 4-0 175 0-5 4-5 280 3-0 285 7-0 5 5-0 350 9-0 355 2-5 360 0-5 1-0 1-5 275 3-0 285 6-0 355 4-5 345 5-0 10 7-0 320 1-0 1-5 5-0 330 3-0 320 12-0 340 10-5 5 9-0 5 10-5 320 1-5 2-0 5-0 350 4-0 330 11-0 340 13-0 355 12-5 360 13-5 330 2-0 2-5 7-0 10 5-5 320 11-0 345 14-5 5 13-0 360 13-5 325 2-5 3-0 8-5 360 7-0 320 13-0 350 13-5 355 14-5 350 13-0 320 3-0 3-5 7-5 360 8-0 320 [16-0 345] 15-5 350 16-5 355 14-0 320 3-5 4-0 7-5 345 5-0 310 18-5 360 18-5 360 17-0 310 4-0 4-5 5-0 350 4-5 5-0 8-0 350 5-0 5-5 [8-0 355] 5-5 102 1 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CLASS (e) continued. Gradient Surface Aug. 2, 1908 7.30 p.m. vel. dir. Nov. 16, 1908 10.47 a.m. vel. dir. Feb. 7, 1909 4.28 p.m. vel. dir. 11-0 185 Feb. 17, 1909 8.17 a.m. vel. dir. Feb. 18, 1909 4.43 p.m. vel. dir. 13-0 120 Mar. 5, 1909 5.13 p.m. vel. dir. 15-0 200 May 2, 1909 7.7 p.m. vel. dir. Gradien Surface 1-0 125 5-0 340 7-5 135 2-5 290 4-0 115 3-5 160 2-5 250 0-5 4-0 45 5-0 335 11-0 140 3-5 290 8-0 115 9-0 180 4-0 265 0-5 1-0 6-5 20 4-0 295 11-5 130 4-0 290 10-5 110 7-0 200 2-5 350 1-0 1-5 9-0 30 5-0 265 6-0 155 3-0 80 7-0 130 7-5 210 2-0 10 1-5 2-0 13-0 40 3-0 160 6-0 150 2-5 55 6-0 130 3-5 220 4-0 10 2-0 2-5 13-0 30 2-5 2-5 3-5 340 3-0 125 5-0 250 9-0 345 2-5 3-0 12-5 25 5-5 5 3-0 270 8-0 320 1-0 140 8-0 260 9-5 350 3-0 3-5 13-0 25 4-0 355 2-5 320 10-0 320 2-0 110 9-0 270 11-5 350 3-5 4-0 13-5 10 4-5 330 2-5 290 10-5 335 2-0 360 12-0 270 12-0 340 4-0 4-5 14-5 10 4-0 10 9-5 310 3-0 30 15-0 275 14-5 350 4-5 5-0 [18-0 10] 7-5 10 7-5 310 2-0 335 20-0 335 5-0 5-5 [10-5 5] 7-0 305 1-5 330? 5-5 6-0 7-5 295 3-0 310 6-0 6-5 5-0 305 6-5 7-0 6-0 320 7-0 7-5 8-5 325 7-5 8-0 [10-0 335] 8-0 2 theodolites burst 2 theodolites to 2-6 km. to 5 '6 km. TABLE OF ASCENTS 103 CLASS (e) continued. 2. Upper wind between South and West. Gradient Surface July 24, 1907 3.7 p.m. vel. dir. 6-0 125 May 2, 1908 0.48 p.m. vel. dir. June 3, 1908 10.19 a.m. vel. dir. 6-0 145 June 3, 1908 6.54 p.m. vel. dir. 5-0 40 May 16, 1909 6.22 p.m. vel. dir. 9-0 70 June 3, 1909 0.59 p.m. vel. dir. 22-0 80 Gradient Surface 4-0 140 4'5 190 2-0 60 6-5 50 4-5 50 15-0 35 0-5 4-5 120 4-0 165 4-0 60 6-0 80 9-0 50 16-0 45 0-5 1-0 2-0 50 2-0 120 5-5 80 4-5 110 8-5 70 10-0 95 1-0 1-5 2-0 150 2-0 110 6-0 105 3-0 110 4-0 120 7-5 110 1-5 2-0 1-5 195 2-5 120 5-0 110 4-0 160 6-0 150 7-5 100 2-0 2-5 2-0 220 1-5 120 2-5 130 6-0 155 5-5 190 9-0 105 2-5 3-0 3-0 215 2-0 125 2-5 170 5-0 150 6-5 185 8-0 120 3-0 3-5 3-0 220 1-5 185 6-0 180 5-0 160 7-0 140 3-5 4-0 3-5 270 9-0 175 5-0 165 4-0 4-5 6-0 270 8-5 180 5-0 160 4-5 5-0 6-0 290 [11-0 165] 5-5 180 5-0 5-5 7-0 250 4-0 195 5-5 6-0 9-0 240 5-0 170 6-0 6-5 10-0 245 3-5 170 6-5 7-0 7-5 150 7-0 7-5 8-5 155 7-5 8-0 6-5 170 8-0 8-5 4-0 170 8-5 9-0 4-0 250 9-0 9-5 4-5 230 9-5 10-0 4-0 200 10-0 10-5 2-0 170 10-5 11-0 5-0 180 11-0 11-5 ^0 205 11-5 2 theodolites 2 theodolites 2 theodolites 2 theodolites to 0-8 km. to 4-0 km. to 2-8 km. to 9-0 km. distance burst 104 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CLASS (f). The Stratosphere. Gradient Surface July 28, 1908 7.0 p.m. vel. dir. 9-0 40 July 29, 1908 7.0pm. vel. dir. 9-0 40 July 30, 1908 7.0 p.m. vel. dir. 7-0 320 July 31, 1908 7.0 p.m. vel. dir. Sept. 30, 1908 4.31 p.m. vel. dir. 6-0 180 Oct. 1, 1908 4 20 p.m. vel. dir. 4-0 160 Gradient Surface 35 5-0 360 5-0 355 4-5 150 3-0 150 0-5 5-0 360 6-5 280 5-0 355 4-5 175 3-0 150 0-5 1-0 4-5 10 8-0 285 5-0 360 4-0 150 4-5 155 1-0 1-5 4-0 5 8-5 290 4-0 360 3-5 160 7-0 175 1-5 2-0 2-5 330 5-0 30 6-0 285 6-0 340 3-5 185 8-5 175 2-0 2-5 5-5 330 5-5 25 8-0 270 6-5 330 3-0 175 7-5 175 2-5 3-0 6-5 330 6-0 15 10-5 275 10-0 330 2-5 180 10-0 185 3-0 3-5 6-5 340 6-5 360 13-5 290 19-0 325 5-0 220 11-0 180 3-5 4-0 7-0 335 7-5 5 13-5 305 19-0 320 10-0 230 135 175 4-0 4-5 8-0 340 10-0 10 9-5 300 21-5 315 9-0 210 14-0 175 4-5 5-0 11-5 350 11-5 5 9-5 305 21-5 310 9-5 210 13-0 185 5-0 5-5 15-0 350 10-0 5 12-0 310 20-5 310 10-0 215 12-5 195 5-5 6-0 15-0 355 8-0 5 12-5 310 21-5 310 9-5 220 12-0 210 6-0 6-5 15-0 350 8-0 5 12-5 305 22-5 310 10-5 215 16-0 215 6-5 7-0 16-5 350 10-0 5 12-5 295 20-5 305 12-5 205 15-5 210 7-0 7-5 20-0 350 12-5 20 11-5 300 19-0 310 11-5 205 15-5 210 7-5 8-0 23-0 360 12-5 20 10-5 300 24-5 320 12-0 215 17-0 210 8-0 8-5 23-0 360 14-5 15 11-0 300 27-5 320 12-0 220 19-5 210 8-5 9-0 23-5 350 15-0 10 13-0 310 30-5 320 14-0 220 21-0 210 9-0 9-5 22-0 355 15-0 10 12-0 320 29-5 310 16-5 220 22-0 215 9-5 10-0 22-0 350 18-5 10 12-0 305 29-5 310 17-5 230 23 5 205 10-0 10-5 20-5 340 19-0 5 14-5 305 31-5 300 20-0 230 24-0 210 10-5 11-0 20-5 340 20-0 5 15-0 315 33-0 305 20-0 230 24-0 210 11-0 11-5 230 340 24-0 5 13-0 320 35-0 300 23-0 225 25-0 195 11-5 12-0 16-5 340 22-0 5 33-0 300 25-5 225 26-0 195 12-0 12-5 9-5 340 18-0 10 30-0 305 25-0 225 26-0 190 12-5 13-0 4-0 310 13-0 350 27-0 315 25-5 230 23-0 190 13-0 Continued on next page. TABLE OF ASCENTS CLASS (f) continued. 105 July 28, 1908 7.0 p.m. July 29, 1908 7.0 p.m. July 30, 1908 7.0 p.m. July 31, 1908 7.0 p.m. Sept. 30, 1908 4.31 p.m. Oct. 1, 1908 4.20 p.m. vel. dir. vel. dir. vel. dir. 13-5 24-0 310 28-5 230 190 190 13-5 14-0 27-0 230 14-0 205 14O 14-5 23-5 235 10-5 210 14-5 15-0 19-0 225 9O 210 15-0 15-5 13-0 205 7-0 175 15-5 16-0 13-0 199 7-0 180 16O 16-5 9-0 170 16-5 17-0 8-5 180 17O 17-5 7O 180 17-5 2 theodolites 2 theodolites 2 theodolites 2 theodolites 2 theodolites 2 theodolites to 7'7 km. to 13 km. to 9 km. to 10 km. to 3-5 km. to 4 km. distance burst distance burst distance burst Oct. 2, 1908 4.20p.m. May 6, 1909 6.25 p.m. May 7, 1909 2.52 p.m. May 7, 1909 6.29 p.m. Aug. 5, 1909 6.33 p.m. Mar. 3, 1910 4.30 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 60 160 21-0 120 22-0 115 22-0 115 4O 90 Gradient Surface 2-5 130 15-0 12-2 70 8-5 115 0-5 5-5 125 15-0 70 2-0 45 9-5 105 0-5 1-0 4-0 140 12-0 100 18-0 95 4-5 70 10-0 125 IO 1-5 7-0 160 12-0 110 14-0 100 14-0 95 5-5 80 11-0 135 1-5 2-0 8-5 16() 11-0 115 10-0 105 9-0 100 5-0 40 12-0 140 2-0 2-5 90 160 10-0 115 11-0 100 9-0 100 5-0 80 9-0 145 2-5 3-0 12-5 155 10-0 120 10-5 100 13-5 100 6-0 85 8-0 145 3O 3-5 9-5 165 9-5 120 10-0 95 9-0 90 3-5 65 4-5 90 3-5 4-0 7-0 155 9-5 90 14-0 100 10-5 90 3-0 60 7-0 95 4O 4-5 8-5 150 10-0 110 9-0 100 15-0 90 4-0 55 5-0 50 4-5 5-0 10-0 155 10-0 120 8-0 115 11-5 90 5-0 60 6-0 70 50 5-5 12-0 170 12-0 130 7-0 120 6-5 85 4-0 45 5-5 65 5-5 6-0 10-5 180 12-5 140 7-5 100 4-5 80 2-5 25 5-5 55 60 6-5 8-0 185 HO 135 6-0 80 5-0 65 2-0 40 4-5 45 6-5 7-0 8-5 180 11-0 145 4-5 65 7-5 60 3-0 70 4-0 40 70 Continued on next page. c. 14 106 THE STRUCTURE OF THE ATMOSPHERE IN CLEAR WEATHER CLASS (f) continued. Oct. 2, 1908 May 6, 1909 May 7, 1909 May 7, 1909 Aug. 5, 1909 Mar. 3, 1910 4.20 p.m. 6.25 p.m. 2.52 p.m. 6.29 p.m. 6.33 p.m. 4.30 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. 7-5 7-0 190 9-0 155 7-5 80 7-5 80 3-0 80 3-5 25 7-5 8-0 9-5 190 13-0 140 8-5 105 6-5 115 0-0 5-5 25 8-0 8-5 10-5 200 16-0 135 9-0 120 2-0 100 5-0 30 5-0 20 8-5 9-0 10-0 215 17-0 130 7-5 125 5-5 110 9-0 25 5-0 10 9-0 9-5 9-5 225 16-0 130 6-0 155 4-5 100 5-0 30 5-0 15 9-5 10-0 11-0 215 18-0 130 5-5 155 3-0 100 5-0 40 6-5 45 10-0 10-5 11-0 215 20-0 130 7-5 140 3-5 135 4-0 30 5-5 30 10-5 11-0 12-0 215 20-0 135 9-0 130 4-5 140 3-5 35 4-0 20 11-0 11-5 12-0 220 15-0 145 6-0 145 6-0 125 3-5 40 5-0 80 11-5 12-0 12-0 225 14-0 140 5-5 130 4-5 1-0 1-0 190 12-0 12-5 12-5 220 12-0 115 2-5 165 4-0 2-5 350 5-0 265 12-5 13-0 12-5 220 [8-0 145] 2-0 210 2-0 1-0 330 3-0 250 13-0 13-5 15-5 205 2-0 3-0 280 2-0 13-5 14-0 15-0 200 5-0 3-0 1-5 85 14-0 14-5 10-0 215 3-0 5-5 3-0 270 14-5 15-0 10-0 215 0-5 5-0 8-0 275 15-0 15-5 10-0 205 5-0 15-5 16-0 8-5 210 2-5 16-0 16-5 4-0 16-5 17-0 1-5 17-0 17-5 3-5 17-5 18-0 3-0 18-0 18-5 3-0 18-5 2 theodolites 2 theodolites 2 theodolites 2 theodolites 2 theodolites 2 theodolites to 4 km. to 9'5 km. from 1 to to 15 km. to 15 km. burst light failed burst 2 km. above above above 14 km. the 12 km. the 11 -5 km. the direction is direction is direction is exceedingly exceedingly exceedingly variable variable. variable. burst light failed TABLE OF ASCENTS 107 Unclassified Ascents. Feb. 2, 1907 11.45 a.m. Feb. 4, 1907 10.44 a.m. Feb. 14, 1907 2.9 p.m. Feb. 23, 1907 3.43 p.m. Mar. 28, 1907 5.55 p.m. May 20, 1907 4.49 p.m. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. vel. dir. Gradient Surface 4-0 45 8-0 315? 7-0 25 5-0 201 Gradient Surface 2-5 360 4-0 355 3-0 315 3-5 10 2-5 135 4-0 40 0-5 5-0 50 5-0 350 8-0 325 6-0 20 1-5 70 5-0 40 0-5 1-0 8-5 65 8-0 300 3-5 360 7-0 30 5-5 70 3-0 10 1-0 1-5 [5-0 50] 9-0 80 4-0 10 1-5 2-0 7-5 95 2-0 cloud cloud cloud? distance Totland Bay Gradient Surface June 17, 1907 7.15 p.m. vel. dir. 6-0 270 Mar. 30, 1908 11.29a.m. vel. dir. 18-0? 270? April 4, 1908 4.14 p.m. vel. dir. 16-0 340 Jan. 19, 1909 10.34 a.m. vel. dir. 7-0? 315? Jan. 19, 1909 0.33p.m. vel. dir. 7-0? 315? Gradient Surface 5-0 250 9-5 235 6-5 305 4-0 330 7-5 325 0-5 7-5 260 9-0 235 8-5 320 8-5 340 10-5 335 0-5 1-0 6-5 235 11-0 240 13-0 330 10-0 340 8-5 340 1-0 1-5 6-5 235 12-0 340 13-0 305 12-0 330 1-5 2-0 11-0 335 5-0 290 10-0 325 2-0 2-5 9-5 250 2-5 3-0 [15-0 270] 3-0 cloud 2 theodolites 2 theodolites burst 2 theodolites above 1-1 km. 142 108 DIAGRAMS: RELATION OF WIND ELEMENTS TO HEIGHT The following diagrams represent graphically the relation of wind velocity and direction to height in a certain number of cases. The wind velocity is given in metres per second, and the direction in degrees from the North point through East, South and West ; an East wind is therefore represented by 90, a South wind by 180 and so on. Heights are given in kilometres. The short vertical lines on the diagrams show the gradient velocity and direction where this could be calculated. The weather maps show the pressure distri- bution and the wind at the surface at about the time of each ascent. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 109 APR.5. 1907 APR.6.1907 6 p.m. Barometer and Wind. 6 p.m. VELOCITY 10 20 30 N DIRECTION E 5 W N w> 4 3 2 \ 5 4 3 Z I c * APR 5. 4 H H. 40M. RM. K 7' 5 - > APR 6. 6 I3M. P.M. l 1 :> 4 ^^ / ^ ) 10 20 30 100 200 Class (a). 300 110 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT APR 18. 1907 APR 19. 1907. 6 p.m. Barometer and Wind. 8 a.m. VELOCITY 10 ?0 30 N DIRECTION E 5 W N J 4 3 ^ i c y. APR 18. 6 H. 40M F!M. 5 / ~7* ^* / i f <: ,| T N 3 10 eo 30 100 ZOO 300 Class (e 1). DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 111 MAY 14 1907 MAY 17. 1 907 6 p.m. Barometer and Wind. 6 p.m. VELOCITY DIRECTION J 4 3 2 1 5 4 3 2 1 n MAY 14 6 H. H. 53M. RM >, f r^\ y \ s^ $ / ^*^- * MAY 17.6 43M. P.M. ( / V - 4 V / y } 10 20 30 100 Class (d). Totland Bay. 200 300 112 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT MAY 21. 1907. MAY 25.1907 6 p.m. Barometer and Wind. 6 p.m. .0 VELOCITY 10 20 30 N DIRECTION E 5 W N > 4 3 2 1 5 4 3 2 1 n MAY 21.6 H. H SM. P.M. ^ X' S, > I J^ -r~" \ <;' -^ >: MAY 5.7 I3.M PM / ^ . / ' k ^ . s ^/ , 3 ^^ S 10 30 100 Class (d). Totland Bay. 200 300 DIAGBAMS : RELATION OF WIND ELEMENTS TO HEIGHT 113 MAY 7. 1907 MAY 28. 1907. 6 p.m. Barometer and Wind. VELOCITY ID 20 30 N 8 a.m. DIRECTION E S W N ID 9 8 7 6 5 4 3 2 1 n MAY 27.6 H. MM. P.M. \ j \ ^* 2 { J { L > I m ^ N -^ i \ t f S c 2 '0 10 30 100 Class (rf). Totland Bay. 200 300 15 114 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT AUG.28. 1907. 8 a.m. Barometer and Wind. 6 p.m. VELOCITY DIRECTION ,0 10 30 N E S W N 7 1 I.3E1M A.M. 7 6H.23MP.M 10 30 Class (d). 100 200 300 DIAGRAMS I RELATION OF WIND ELEMENTS TO HEIGHT JAN 2. 1908. 115 8 a.m. Barometer and Wind. 6 p.m. VELOCITY 10 20 30 N DIRECTION E 5 W N J 4 3 2 1 5 4 3 2 1 C ^ IO.H.4 3r 7 ,.A.K y 1 1 j '^ p . c S < ^ - ^. v < S? / QLJ 4 j n ~t wRM X } - / ] 10 20 30 100 200 300 Class (c). 152 116 DIAGRAMS I RELATION OF WIND ELEMENTS TO HEIGHT FEB.2.1908 FEB 4.1908. 6 p.m. Barometer and Wind. 6 p.m. VELOCITY DIRECTION J 4 3 2 1 5 4 3 2 1 FEB ^. A H. ( H. i GM. P.M. y . ) s* T "^=* i .1 FEB A. 4 48 M . P.M ' . ^ .S ^ ) > r D 10 20 30 100 200 300 Class (b). DIAGRAMS I RELATION OF WIND ELEMENTS TO HEIGHT MAY 2 1908 117 8 a.m. Barometer and Wind. 6 p.m. VELOCITY DIRECTION N IU 9 8 7 6 5 4 3 E 1 n i cH.4t P.M. >. 5 f C, X. 5 r ? -^ * ^' s ^ / ^ s ' X. 10 20 30 100 200 Class (? 2). Two theodolites to 4 kilometres. 300 118 DIAGRAMS I RELATION OF WIND ELEMENTS TO HEIGHT MAY 30. 1908. MAY 31. 1908. 6 p.m. Barometer and Wind. 8 a.m. VELOCITY DIRECTION JU 9 8 7 6 5 4 3 Z I n MAY 30.7 H. 2M. P.M. I \ / < i --^ I | > | > I 7 3 I / r" ^> 100 Class (a). Cloud at 7 kilometres. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT JUNE 3. 1908. JUNE 4. 1908. 119 6 p.m. Barometer and Wind. 8 a.m. ,0 VELOCITY 10 20 3D N DIRECTION s W N It. II 10 9 B 7 6 5 4 3 2 1 n ^ V JUNE 3. 6 H. 54 M P.M. / S < L. < > 'I ^ k \'. C -* -> i f -.c ^ I V ) i i i ) r-\ I y ) \ < ^ A < > ^~ '0 10 0. 30 100 00 Class (e 2). Two theodolites to 2-8 kilometres. 300 120 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT JUNE 2E.I90B. JUNE 23,1908. 6 p.m. Barometer and Wind. 8 a.m. DIAGRAMS : RELATION OF WIND ELEMENT.* TO HE 121 I5 [ 14 13 12 II 10 9 8 7 6 5 4 3 2 1 n ] ID 20 30 ^ H. J E S W h JUNE 22.6 3BM. P.M. . 7 \ < )- ; x . g > ). ( r 2 \ / i I . I - /. V Z ./ > 5 s ;. V ^ f V 5 _!) y" i ! *" ^-~ .j 10 20 30 JOO 200 300 Class (e 1). Two theodolites to 4'3 kilometres. c. 16 122 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT JULY2S.B00. Barometer and Wind. 6 p.m. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 123 VELOCITY DIRECTION I5 C 14 13 12 II ID 9 8 7 6 5 4 3 2 \ n 3 ID 20 3D r M. si E 5 vy N 7H.Q P.M. ^-, ^ v "i^c; ? 2 . [ 1 \. 3 ! j ! \ - .> p i / ./< j _ / I . i \ / 7 J \ \ \ s \ \ ID 20 3D 00 Class (^1); (/). Two theodolites to 7-7 kilometres. 300 162 124 DIAGRAMS: RELATION OF WIND ELEMENTS TO HEIGHT JULY 29. BOB. Barometer and Wind. 6 p.m. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 125 VELOCITY DIRECTION 15 14 13 12 II 10 9 8 7 6 5 4 3 1 n D 10 20 3D f M \ sj E 5 W h 7H.C P.M V * 2 [ 1 X / \ 2 2 7 s x / 7 1 / \ / \ ( 7 . : O 20 30 Q 100 200 3DO Class (el)-, (/). Two theodolites to 13 kilometres. 126 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT JULY31. 1908. 29 9 29-9 Barometer and Wind. 6 p.m. DIAUKAMS; RELATION OF WIND ELEMENTS TO HEIGHT 127 VELOCITY DIRECTION 15 14 13 12 II 10 9 8 7 6 5 4 3 2 1 n ] 10 20 30 \ M v| E 5 VV h 7h.O RM \ \ s \ N. x . ! ^ i 9 / ' I } \, / ) > -^ ; 7 ") 1 ( 7 \ ^ \ / \ jj / s / 10 20 30 100 200 300 Class (e 1); (/). Two theodolites to 10 kilometres. DIAGRAMS: RELATION OF WIND ELEMENTS TO HEIGHT AUG i. 1906. 7 a.m. Barometer and Wind. 6 p.m. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 129 4 3 2 I 5 4 3 4 3 Z I VELOCITY 10 10 20 DIRECTION 30 N BH.IC M AM. P.M. 7 H .5 M. P.M. 30 DO Class (e 1). W N 200 300 c. 17 130 DIAGEAMS I RELATION OF WIND ELEMENTS TO HEIGHT SEPT 30. 1908. 7 a.m. Barometer and Wind. 6 p.m. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 131 VELOCITY DIRECTION 10 20 30 N E 5 W N 7 6 5 4 3 2 I n 10 8K36MA.M. IIH.I7 M A.M 30 DO 200 300 Class (rf). 172 132 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT OCT I. BOB Barometer and Wind. 6 p.m. DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT 133 D 30 N DIRECTION E 5 W N L.U 19 IB 17 16 15 14 13 IE II 10 9 8 7 6 5 4 3 E 1 n 4 H .2 ON 1RN/ y > :/ ? . I t > \ ^ 1 | X ^ { y t / i . / x ' ' f ^ J ) 5 1 7. / I ? t P r \ ; 10 20 30 100 200 Class (6) ; (/). Two theodolites to 4 kilometres. 300 134 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT NOV6 1908. 7 a.m. Barometer and Wind. 6 p.m. VELOCITY DIRECTION ( 3 II 3 c 3 r si E S \y t n / OH. 5 9 /i.A.N 1 u R X I U ~7 1 , / I f c; f I 3 A / / ^ *r I v: ~ _ s I \ ^-^ ^ d 3 1 n ^- J> ' , / u c ) 1C } ft 3 3( 3 Clas C < ) d). D D 2D 30 D DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT NOV. 7 1908. 135 7 a.m. Barometer and Wind. 6 p.m. VELOCITY 10 20 30 N DIRECTION E S W N IU 9 8 7 6 5 4 3 2 \ C 3H.2 5i ^i.P.M ^ S | r J ' < ./ 1 , ^- 7 x. ( > ^ > f j D 10 20 30 100 200 300 Class (d). 136 DIAGRAMS : RELATION OF WIND ELEMENTS TO HEIGHT JAN 12. 1909 7 a.m. Barometer and Wind. 6 p.m. VELOCITY DIRECTION c I 3 2 1 4 3 I \ (. I OH.5 8r 2r Oh ^-* i. 7 2H.2 ^. RN/ > >=T ^ " ( ! i> *x X. { f \ 3H.5 /i.FW * > : . j . ^^ ^ 1 ^ X r k 3 10 20 30 100 200 300 Class (6). DIAGRAMS : RELATION OP WIND ELEMENTS TO HEIGHT FEBI8.I909. 137 7 a.m. Barometer and Wind. 6 p.m. ,0 VELOCITY 10 20 30 N DIRECTION E S W N u 3 8 7 6 5 4 3 2 1 n 4H.4: ^ .P.M. > r i I. / J \ \ (. 1 \ ^ ^^ 5 ^~-_ I ^ \ \ < ( < ) r- -S ^ \ 10 20 30 100 200 Class (e 1). Two theodolites to 5-6 kilometres. 300 c. 18 138 DIAGRAMS: RELATION OF WIND ELEMENTS TO HKKJIIT ADD 5. 1909. AUG6.I909. C p.m. Barometer and Wind. 7 a.m. DIAGRAMS: RELATION OF WIND ELEMENTS TO HEIGHT 13!) ,0 VELOCITY 10 20 3D N DIRECTION 5 W _u IS 18 17 16 15 14 13 12 II 10 9 a 7 6 5 4 3 2 1 D AUG 5. 6 H. 33M P.M. > 1 < P'- . , i : > ^^ . - , - - ^- - - V. X- ~^*^ 2 ) 5 j \ < c 5 ? >^ ^" fc ~- \ r \ t J \ i . I " ^ L 7 t / f ^ } ^ < L < \ < N r ) * ^ ? DIRECTION 5 VV N 30 M .RM. ID 20 30 100 00 Class (c); (/ ). Two theodolites to 15 kilometres. 300 INDEX Accuracy of methods, 18 et seq. Increase of wind velocity in upper air, 3, 38, 69 Ascent of balloons, rate of, 16, 25 et seq. International Commission for Scientific Aeronautics, Atmosphere, structure of the, 1 et seq. xi, 16, 57 Audibility of sounds, 78 Isothermal Layer, see Stratosphere Balance for filling balloons, 17. Leakage of gas from balloons, 28 Ballons d'essai, 10 Lempfert, R. G. K., 73 Ballons sondes, 10, 16 Ley, Capt. C. H., 28 Balloons, free lift of, 16 Low pressure, wind blowing away from, in upper leakage of gas from, 28 air, 4, 48, 73 ,, method of starting, 11 ,, rate of ascent of, 16, 25 et seq. Mallock, A., 27 ,, weight of, 16 Microbarograph, 78 Base line, 10 Models showing vertical wind distribution, 3 et seq. Bosch theodolite, 17 Observing, methods of, 10 et seq. Calculator, Brunsviga, 14 Cary Porter theodolite, 17 Pilot Balloon ascents, importance of, 78 Chobham Common, ascents at, 1, 81, 96 Pilot Balloons, vii, 1, 2, 10 et seq. Cloud, 6, 31, 78 Portland Bill, 34 Compton Down, 31 Precipitation with reversals, 41 et seq., 78 Contours of ground, effect of, on air currents, 29, Pressure distribution, relation of, to vertical wind 30, 31 distribution, 32 et seq., 54 et seq., 69 et seq. Convection Currents, diurnal, effect of on wind velocity, 60 Quervain, A. de, 17 Decrease of wind velocity in upper air, 4, 39, 73 Rain, 6, 41 et seq., 47, 78 Dines, W. H.. viii, xii, 5, 10, 18, 26, 76 Reversals, ix, 4, 41 et seq., 73, 76, 78 Ditcham, observations at, 1, 81 et seq. thunderstorms occurring with, 42 et seq., 47, 76 Gradient wind, 3, 32 Rotch, A. Lawrence, xi Guilbert, Gabriel, 76 Gun firing and rain, possible origin of super- Shaw, W. N., viii, ix, xii, 66, 68, 73, 76 stition, 78 Slide rule, 11 Guns, sounds of firing, 78 "Solid" current, 3, 35, 69 South Downs, effect of, on winds, 36, 67 Hergesell, H., 10, 26, 27 Strassburg, observations at, 26 INDEX Stratosphere, viii, x, 5, 01 et ,w/. temperature in, 61 .Surface, increase of wind velocity above. (Hi ,, wind, 3, 66 Teisserenc de Bort, L., ix, xi, 10 Temperature distribution, effect of, on winds, ix, xi, 33, 34, 36, 38, 41 et seq., 69 Temperature, vertical distribution of, 61 Theodolite for balloon observations, 17 ,, method of one and two compared, 18 et seq. ,, observations with one, 14, 66 two, 10, 66 Thunderstorm, 31 Thunderstorms occurring with reversals, 42 ft. s?q., 47, 76 Totland Bay, ascents at, 1, 80, 85, 88, 89, 92, 96, 107 Trajectory, horizontal, of balloons, 15, 21, 22, 23, 34 Units used in the work, xi, 15 Vertical currents in atmosphere, 27, 2^, 47, 51, 73, 75, 7