440 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , These results are collected in Table YI , where it will be seen that the coefficient of viscous traction X is roughly three times that of viscosity .
Table VI .
A. fl. A. A/ a* .
4 -3*x 1010 1'4 x 1010 3-07 6 -7 x 109 2 -2 x 109 3*04 3 -6 x 1010 1 1 -0 x 1010 | 3-60 5 -9 x 106 2 -0 x 106 2*95 3 -3 x 101(1 J 3-30 9 -3 x 105 2 -8 x 106 3*25 12 -9 x 109 4 -2 x 10s 3-07 7 '6 x 105 2 -6 x 105 2-91 The Vertical Temperature Gradients on the West Coast of Scotland and at Oxshott , Surrey .
By W. H. Dines , F.R.S. ( Received November 10 , \#151 ; Read December 7 , 1905 .
) In a paper by Dr. Shaw and the author read before the Royal Society on May 14 , 1903 , * an account of an investigation into the conditions of the upper air over the sea in the neighbourhood of Crinan , on the West Coast of Scotland , was given .
Since that time two fresh series of observations in the same locality have been obtained , the results of which are now submitted .
In each case observations of temperature and humidity were made by self-recording instruments sent up by means of one or more kites , which were flown from the deck of a steam vessel .
Expenses .
The expense has been met by a grant of \#163 ; 200 made by the Government Grant Committee , a grant of \#163 ; 50 made by the British Association at the Southport Meeting , and of \#163 ; 40 at the Cambridge Meeting ; and also by an anonymous contribution of \#163 ; 25 by a Fellow of the Royal Meteorological Society .
These grants have not been used entirely for the observations at Crinan , but have afforded the means of carrying on experimental work at Oxshott ; by them , too , apparatus for a separate investigation carried out by Mr. G. Simpson on the North Sea has been provided . !
For the observations at Crinan in 1903 a tug was hired , and the Lords Commissioners ol the * ' Phil. Trans. , ' A , vol. 202 , pp. 123\#151 ; 141 .
t 'Met .
Soc. Quart .
Journ. , ' vol. 32 , No. 137 , pp. 15\#151 ; 25 .
1905 .
] Gradients in the West of Scotland and Surrey .
441 Admiralty , at the request of the Royal Society , kindly provided a very convenient vessel , H.M.S. " Seahorse , " for six weeks , commencing on June 19 , 1904 .
The Meteorological Council lent the necessary instruments , and bore the expense of maintaining a base station during both summers .
They also greatly assisted the work by sending a daily telegram with a forecast of the weather , and a statement of the magnitude of the barometric gradient .
Material Collected .
The results obtained consist of 16 fairly satisfactory traces at Crinan in August , 1903 , of 29 traces at Crinan in June and July , 1904 , and of 75 traces at Oxshott .
The average height for the first set ( Crinan , 1903 ) is 4500 feet , the second set ( Crinan , 1904 ) give 5340 feet .
The averages for Oxshott are 3300 feet for 1904 and 5200 feet for 1905 .
A meteorograph kindly supplied me by M. Teisserenc de Bort was used in 1903 ; since the beginning of 1904 a new form of meteorograph , which is described in the Quarterly Journal of the Royal Meteorological Society , * has been used .
With the exception of the months of July , August , and September , 1904 , when there has been sufficient wind ascents have been made at Oxshott on the days appointed by the International Aeronautical Commission , these days being , as a rule , the first Thursday in each month .
The other ascents at Oxshott have been made rather with a view of testing and , where possible , improving the apparatus than of obtaining definite information .
Owing to the inconvenient situation of the place , which is close to many thickly -populated suburbs , the rule was made of never using more than one kite .
This rule was rigidly adhered to until the commencement of the present year , but since it was found that increased experience and an improvement of the apparatus had led to a great diminution in the number of mishaps , the rule was broken through , and the number of kites extended to two .
It is still not considered desirable to use more than two , or to let out more than from 10,000 to 12,000 feet of wire .
It may be of interest to others engaged in similar work if I state that out of the 75 ascents at Oxshott , on five occasions the kite and meteorograph have broken away or fallen to the ground ; on one occasion this was due to the breaking of the wire near the kite under a pull of 200 lbs. ; on another occasion to the breaking of the front stick of a kite under a pull of 230 lbs. ; once to a careless fastening at the end of the wire , and twice to some unknown cause .
The loss of material has been one kite broken beyond repair , and about 2000 feet of wire .
The meteorograph has never been lost or damaged , since being secured in the middle of the kite it is well protected .
* Vol. 26 , No. 135 , pp. 217\#151 ; 227 .
442 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , On the three occasions when the kite has been detached from the wire it has fallen at a horizontal distance of about six times its vertical height , and the wire has been wound in undamaged almost to the end .
The average ratio of the vertical height to the length of wire employed has been a little over 6/ 10 .
Observations at Crinan in 1903 .
The results obtained in the summer of 1903 have been published in the Journal of the Royal Meteorological Society.* This set of observations were not so numerous or so consecutive as might be wished .
The steam tug employed was in many ways unsuitable , and the weather rendered kite flying difficult as well as unpleasant .
The observer at Malin Head , the reporting station nearest to Crinan , reported a gale to the Meteorological Office on 20 occasions during August , 1903 , and the captain of the tug often feared to take out his vessel , which was old , on account of the weather .
The following is a brief summary of the results:\#151 ; The average decrease of temperature for the first 500 metres was 20,2 C. , for the second 500 metres 3'*4 , and for the third 3'*6 , the total decrease for 1500 metres being 9'*2 , or at the rate of 0'*6 C. per 100 metres .
The winds were nearly all between south-west and north-west .
The temperature observed at the kite when at the level of Ben Nevis was always above the temperature of the summit at the same time , the differences ranging between 1'*6 C. and 6'*1 C. , and the average excess was 3'*3 C. One trace of especial interest was obtained : a thunderstorm broke out without previous warning while a kite was flying at a height of 4400 feet ( 1300 m. ) , and a meteorograph was drawn in through the actual region of the electrical disturbance , but the trace showTed no peculiarity of any kind .
For convenience of reference Table A , which appeared in part in the previous paper and , excepting the values for Oxshott , has been published in the Journal of the Royal Meteorological Society , is here reproduced .
Observations at Crinan in 1904 .
The results obtained from H.M.S. " Seahorse " are exhibited graphically in fig. 1 .
The diagram ( fig. 1 ) has been drawn on the same plan as before .
The isothermal lines for each degree Centigrade are put in as fully as the observations will permit , and where these have failed the isotherms are connected by dotted lines as a guide to the eye .
The letters S , M , D , and YD refer to the humidity , S denoting a value between 100 and 95 per cent. , M between 95 and 80 , D between 80 and 60 , and YD below 60 per cent. No great degree of accuracy is claimed for these values .
* Quart .
Journ. , ' vol. 25 , No. 130 , p. 155 .
1905 .
] Gradients in the West of Scotland and Surrey .
Table A.\#151 ; Table of Average Temperature Gradients in Degrees Centigrade for 100 Metres for Vertical Columns of different Heights .
Height of column .
500 1000 1500 2000 2500 500 to to to to to 1000 .
1500 .
2000 .
2500 .
3000 .
0-50 0-50 0*51 1 *10 0*80 0-71 0*68 0 *56 C. per 100 metres approximately 1 ' F. per 300 feet .
V 140 met res , 0 *7 0 .
0*52 0*54 0*53 0-62 0*56 0*56 0*52 0*50 0*48 0*46 0*44 0*68 0*73 0*77 0*53 0*34 0*49 0*51 0*77 0*57 0*37 0*30 Berlin balloon ascents ... ... ... ... ... ... ... ... . .
Kite ascents , U.S ... ... ... ... ... ... ... ... ... ... .
Average gradient for mountains ... ... ... ... ... ... .
Average gradient for Ben Nevis , July and August Adiabatic gradient for saturated air , initial temperature 12 ' C ... ... ... ... ... ... ... ... ... ... ... . .
Average summer gradient over sea , west J ^9 ?
' ' coast of Scotland ... ... ... ... ... ... ... . .
j Average gradient near London ( Oxshott ) , 1904 and 1905 ... ... ... ... ... ... ... ... ... ... ... ... ... ... Fig. 2 is taken from the official publications of the Meteorological Office , and shows the movement of the depressions for June and July , 1904 .
The heights at which clouds appeared are somewhat uncertain .
In determining cloud heights by an ascent of a kite there are three possible contingencies ; the kite may become indistinct but not entirely vanish during the passage of an opaque cloud , and in this case the height of the kite certainly agrees with that of the lower surface of the cloud .
This case is unusual .
In the second case the kite disappears into or emerges from a thick cloud sheet ; here again the height of the lower surface is perfectly determinate , but it seldom happens that the disappearance and the emergence , perhaps an hour later , occur at anything like the same height .
The third case is the most usual , when the kite disappears in or behind an opaque cloud , or becomes indistinct owing to a semitransparent cloud .
All that can be said here is that the cloud level is below the height of the kite , and an inspection of the humidity trace often shows that the kite at the time was in very dry air , and therefore not in the cloud at all .
If it is a large thick cloud that is in question , the kite need not have been above its upper level , but merely , from the observer 's point of view , behind some part of it .
The wind direction , wind force on the Beaufort scale , and weather each day are shown below in the diagram ( fig. 1 ) .
Where the wind direction at the level of the kite differed much from the surface direction , it has been put in at the corresponding height , and the absence of any letters on the diagram indicates that the surface and upper wind were substantially the same .
The 444 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , barometer curve shown at the top is copied from a Negretti and Zambra 's barograph that was kept at Crinan .
The period under review , June 20 to July 28 , 1904 , had weather of a somewhat exceptional character for the locality .
The prevalent summer wind is certainly west , i.e. , between south-west and north-west , but while Fig. 1 .
June 30*50 in , 20 , 21 25 24 25 26 27 ' 28 29 30 29*50 . .
W SSW SW SSW W W W SE NW sw THE fflTTE red r c b cr r rain .
t thunder .
b blue sky .
c detached clouds .
o overcast .
the " Seahorse " was at Crinan the wind with few exceptions came from between east and south-west .
There were strong winds from some southerly point on several occasions until July 15 , but after that date very light winds from the east and south-east were the rule .
On three days no ascent could be obtained , and it may be remarked that never before when employing a vessel did we fail to get some sort of daily ascent .
1905 .
] Gradients in the West of Scotland and Surrey .
After July 19 the barometer trace becomes almost a straight line , and it was found that the winds failed entirely at the height of a few thousand feet .
This was most noticeable on July 20 , the only day after the 15th on which there was a good breeze .
Although at the surface the force was 6 on the Beaxtfort scale , a " strong breeze " and three kites were tried , not one Fig. 1\#151 ; continued .
July u u t ; / y 1/ c t/ w c *A closed curve thus Q denotes a temperature inversion .
30*50 in .
50*00 . .
\#163 ; \lt ; D \#163 ; 9*50 - o \#163 ; 9'00 2,000 m. 1,000 m. could be got to rise above 2000 feet ( 670 metres ) .
Fig. 2 shows how few depressions passed near Crinan during the six weeks of observation , and the absence of west and north-west winds to some extent defeated the object of the investigation , namely , a determination of the oceanic temperature gradient , but on the other hand the unusual prevalence of land winds have afforded the means of comparison with the preceding years when different conditions prevailed .
446 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , The table of gradients and the diagram indicate a departure from the preceding results in the steeper gradient for the first 500 metres , and in the comparatively numerous temperature inversions between 1000 and 2000 metres .
A steep gradient in the daytime in summer is a characteristic of continental observations , * and it is not surprising that the prevalence Fig. 2 .
, 2 .
MOVEMENTS OF DEPRESSIONS .
/ .
3rd to 11th , irregular t about Spain , arh Jbai to our southern , distru jVo tracks ca\gt ; v 2\gt ; e s7u June , 1904 .
of land winds should produce a steeper gradient .
Such marked inversions of temperature have not been observed at Crinan in the preceding years of observation , and the inference certainly is that temperature inversions between 500 and 2500 metres are unusual off the Scotch coast with westerly * The gradient 1T0 for the first 500 metres of the United States Weather Bureau ascents ( see table ) is beyond the adiabatic rate for dry air .
These were chiefly , if not entirely , day-time ascents in the summer .
1905 .
] Gradients in the West of Scotland and Surrey .
44 winds .
It is noteworthy that if we take the average gradient up to 1500 metres ( 5000 feet ) it has been practically the same in each of the three summers ; although the observations at Crinan , or rather over the sea in the neighbourhood of Crinan , in 1902 and 1903 were , on account of the prevalence of westerly winds , equivalent to observations over the North Atlantic .
Fig. 2\#151 ; continued .
Z. MOVEMENTS OF .
DEPRESSIONS .
July , 1904 .
Comparison with Ben Nevis .
Table B shows the relation between the temperature on the summit of Ben Nevis and the temperature of the free air about 60 miles to the southwest at the same time .
The temperature given by the kite is the mean of those obtained on the ascent and descent , and the Ben Nevis temperature 448 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , is that of the exact hour that stands nearest to the middle of the kite ascent .
The figures are in striking contrast with those relating to the two preceding years .
For the first time the temperature at the kite was occasionally found to be below that on the summit of the mountain .
Owing to the use of larger kites and a lighter meteorograph , it was possible to get observations on days on which in the summer of 1902 it would have been impossible , on account of the lightness of the wind , to reach the level of Ben Nevis , but it is obvious that this is not a sufficient explanation , since the days on which Ben Nevis was the warmer were not days of exceptional calmness .
Table B.\#151 ; Difference of Temperature between Ben Nevis and a Kite at the same level about 60 miles to the south-west .
Date .
Kite .
Ben Nevis .
* Difference , B.N. \#151 ; kite .
W eather under kite .
Dry .
Wet .
June 21 , 12 * 2*4C .
1 *2 1 *2 -1 *2 W. , 4 yb 22 , 12 3*3 2*5 2*5 -0*8 W. , 4 , c 24 , 12 5*5 3*2 3*2 -2*3 S.E. , 2 , c 28 , 12 6*0 8*2 5*8 2*2 E.S.E. , 1 , c 29 , 12 7*5 10 *5 7*6 3*0 S.E. , 3 , b 30 , 2 p.m 8*0 11*1 6*6 3*1 S.E. , 4 , c July 1 , 3 " 3-6 2*8 2 *8 -0*8 S. , 6 , r 2 , 12 *1 .
2 *4 2*4 2*4 0*0 S.W. , 3 , r 4 , 11 A.M 2 *4 1*7 1*7 -0*7 S.S.W. , 3 , r 5 , 12 I. 5 *0 2 *3 2*3 -2*7 W. , \gt ; 2 , c 6 , 11 A.M 8*8 7 *8 7*8 -1 *0 S.S.W. , 7 , r 7 , 12 1 *4 0*6 0*6 -0*8 S.W. , 6 , c 8 , 11 A.M 4-3 3*9 3 *9 -0*4 S.S.W. , 7 , r 13 , 2 p.m 4*9 7*1 7*1 2*2 S.S.W. , 4 , b 15 , 11 A.M 6*7 6*6 6*6 -0*1 S.S.W. , 5 , c 16 , 3 p.m 1.5*0 3*2 3 *2 -1 *8 W.S.W. , 3 , b 19 , 12 8*8 12 *3 8*7 3*5 S.S.E. , 2 , b 21 , 12 7*5 7*5 7*5 0*0 S.8.E .
, 2 , b 22 , 11 A.M 9*0 7*5 7*5 -1 *5 S.K. , 1 , c 25 , 12 10*1 8*5 7*5 -1 *6 E. , 1 , b 26 , 12 1.5*0 7*4 6*7 2*4 E. by S. , 3 , c 28 , 12 1.11 *5 11 *4 8*6 -0*1 E. , 1 , c * Denotes a temperature inversion at about the height of Ben Nevis .
Explanation of letters :\#151 ; b. Blue sky .
c. Detached clouds , o. Overcast , r. Bain .
With the exception of July 13 , on each day on which the mountain was the warmer the air on its summit was dry , and mostly very dry ; with the same exception also the winds were south-east .
It is noticeable too that the differences were small on the five rainy days given in the table .
In the preceding paper it was suggested that the lower temperature of the mountain was due to the adiabatic cooling of the air as it was forced up the mountain slope by the prevailing westerly winds , the temperature gradient so produced being greater than the ordinary gradient in the 1905 .
] Gradients in the West of Scotland and Surrey .
449 free air ; and the results obtained in 1904 seem to me to support this conclusion .
Oil rainy days the gradient in the free air , at least throughout the vertical region in which rain is forming , must be the adiabatic one , for the dynamic cooling of an ascending current is the only admissible cause of any but the lightest rain , and on such days the adiabatic cooling should produce much the same temperature on the mountain and in the free air .
With a fairly clear sky and a south-east wind the air would be warmed , by contact with , the ground before reaching Ben Nevis , and it is probable that any excess of moisture it may have had originally would be deposited on the slopes of .the mountains lying to the east and south .
A westerly wind , on the other hand , could not be dry , coming from the sea , and , as is usually the caser would produce clouds and a saturated condition on Ben Nevis .
Different instruments were used in each year and the question arisen whether these differences are due to instrumental errors .
An error may arise from an incorrect base line or from an incorrect scale .
The former error is excluded by the fact that in each year the instruments were compared almost daily with verified thermometers , and an error exceeding half a degree must have been detected .
Some error from the second cause is certainly possible , but it could not have been large .
The usual range of temperature at Crinan during the summer is small , the temperature during the daytime seldom exceeding 20 ' C. or falling below 12 ' C. , so that an error in the scale amounting to 1 in 16 may , perhaps , have escaped detection .
No means existed of obtaining low temperatures artificially , and the meteorograph used in 1902 was never compared with a thermometer at temperatures outside the ordinary range of the district .
That used in 1903 was obtained from M. Teisserenc de Bort , he sent me particulars as to the scales , and I do not doubt that they were correct .
I tested the meteorograph used in 1904 over the whole range of likely temperatures before going to Crinan , and although the zero of these instruments is liable to alteration , the scale is not .
Hence , on the whole , , I consider that while an instrumental error of something under lu C. is possible , it is hardly likely that the differences are due to this cause .
The fact that temperature inversions were far more prevalent in 1904 than in the previous years shows that the average weather conditions were different .
Temperature Gradients and Weather Classification .
The temperature gradients over the sea on the west coast of Scotland have now been ascertained on 68 days , and it seems desirable to separate and arrange them according to the weather in which they were observed .
450 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , The classification of weather is that suggested by Dr. Shaw in a paper read before the Scottish Meteorological Society in December , 1904.* South-easterly Type .
A pressure distribution favourable for winds between east and south for the 24 hours to which the observations refer .
South-westerly Type .
Favourable for winds between south and west .
North-westerly Type .
* Favourable for winds between north and west .
North-easterly Type .
Favourable for winds between north and east .
Variable Anticyclonic .
Variable winds during the prevalence of an anticyclone .
Variable Cyclonic .
With sequence of winds incidental to the passage of a cyclone .
A cyclonic region is taken as one in which the isobars are concave to the low-pressure region and probably form closed curves round the low-pressure centre .
An anticyclone is one in which they are concave to the high-pressure region .
In cases where the isobars are approximately straight the region is classed as intermediate , and it has been necessary to add the term " transitional " for cases where for the interval in question the type is changing .
The results are shown in Table C , next page .
It must be borne in mind that these gradients refer to the daytime and to the summer .
Inasmuch as the daily temperature variation on the sea is very trifling , or perhaps even non-existent away from the ship which carries the thermometers , we may take the conditions as being those prevailing , without reference to the hour , during the summer months .
There is no reason to suppose they would be different in the winter , but in the absence of direct information no definite assertion can be made .
The observations from strata over 2000 metres have been omitted , as they are not sufficiently numerous to lead to any conclusion , and , indeed , the figures given in the table are such , that when considered in connection * 'Journal of the Scottish Meteorological Society/ Third Series , Nos. 20 and 21 .
1905 .
] Gradients in the West of Scotland and Surrey .
451 with the number of observations , it would not be unreasonable to say that the differences are of such an order that they may be purely chance ones .
Table C.\#151 ; Gradients , in metres , at Crinan for various Types of Weather .
No. of observation .
0 to 500 .
500 to 1000 .
1000 to 1500 .
1500 to 2000 .
Mean .
Cyclonic .
28 0 *48 0 *60 0-46 0 -54 0 *52 Intermediate .
31 0 *76 0-54 0*42 0 *52 0 *56 Anticyclonic .
4 ... ... 0 *66 Observations wanting .
Transition .
5 0 *54 0-56 0*50 0 -40 0 *50 North-westerly type .
24 ... ... .
0*66 0*62 0*44 0*32 0*51 South-westerly .
21* ... ... .
0-72 0-60 0-40 0 '58 0 '58 South-easterly .
11 ... ... .
0 *58 0 *48 0 *26 0 '58* 0 *44 North-easterly .
Not sufficient observations .
Variable .
7 ... ... .
0 *46 0 *60 Observations wanting .
0 *53 # Only two observations and therefore not included in the mean .
The noticeable points are the small gradient from 0 to 500 metres in cyclonic weather compared with the high value for the same strata for the intermediate type , and the low value from 1000 to 1500 for the southeasterly type .
This latter result is due to the temperature inversions occurring in 1904 , but it is not possible to say definitely whether a small gradient is the usual accompaniment of the south-easterly type .
It will be necessary to obtain more observational information before forming a definite opinion on the whole question .
The absence of observations for the higher strata with the anticyclonic and variable types shows plainly that these types are accompanied by very light winds at a small height , since the want of observations can only be due to the want of sufficient air motion to raise the kites .
The one safe conclusion to be drawn seems to be that the different types of weather are not characterised by any peculiarity of gradient , neither is there any reason Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 for expecting that they should be , except in special cases .
Anticyclonic conditions , it is well known , produce during a winters night on land a sharp inversion in the lower strata , and the same conditions during a summers day produce the steepest gradient that the conditions of equilibrium will admit of , but the two results tend to cancel each other and give on the whole an average gradient .
It is almost certain , too , that neither result occurs over the open sea .
The gradient in the precise region in which rain is forming , or , more strictly , in which vapour is being condensed , must in general be the adiabatic gradient for saturated air , but , compared with the volume of the whole mass of the atmosphere , such regions are very small indeed .
Let us call such a region of ascending air A. Whether this region be the rainfall area of a large cyclonic disturbance , or merely that of a small local shower from an isolated cloud , there must be associated with it a compensating region , B , in which the air is descending .
Ferrel has pointed out that owing to the latent heat of condensation the air in and over the region A must be relatively warm , and that this warmth will suffice to produce the circulation .
% But there is a point which may be easily overlooked .
The adiabatic rate for dry air , such as will be found in B , is far greater than for the saturated air which is found in A , and if the passage from A to B be a purely adiabatic one , the temperature in B will be far higher than in A , a condition which would immediately reverse the direction of the circulation .
Hence the circulation cannot occur , unless it be a forced one , and not a convectional effect .
Where A is a mountain slope up which a wind is blowing , we have a forced circulation , with a foehn wind in the region B , but in the free atmosphere it is very difficult to see what source of power there can be to produce a forced circulation .
It seems more likely that the change from A to B is not adiabatic , and that the region B is very extensive when compared with A and perhaps some considerable distance from it , also that the air which is descending in B , to replace in the lower strata that which is rising in A , is air that ascended some time since and has had time to become cool by admixture with other air and perhaps also to some extent by radiation .
If this is the case there will be no special gradient produced by the condensation of vapour anywhere but in the small and limited regions where the condensation is occuring .
In support of the above suggestions I may quote the results* obtained by unmanned balloons , which show that over cyclonic regions at * " Uberdie Temperaturabnahme mit der Hohe bis zu 10 km .
, nach den Ergebnissen der internationalen Ballonaufsteige ; " ' Sitzungsberichte der Kaiserl .
Akademie der Wissenschaften in Wien , Mathem.-naturw .
Klasse,5 vol. 113 , Abt .
IIa , May , 1904 .
1905 .
] Gradients in the West of Scotland and Surrey .
453 great heights the air is relatively warm , and over anticyclonic regions relatively cold .
The temperature inversions are easily explained on the same hypothesis .
The almost invariable accompaniment of an inversion of temperature has been , both at Crinan and Oxshott , a large decrease in the relative humidity , and this leads to the conclusion that the warmth is due to the compression of descending air rather than to the presence of a horizontal current of air from a warmer quarter .
Inasmuch as descending air is warmed 1 ' C. for every 100 metres of its descent , it must soon become specifically lighter than the air around it and will tend to spread out as a sheet of warm air ; just as wine .or spirits may be made by careful pouring to form a separate layer on the surface of water .
Hence each ascending current that is accompanied by much condensation of vapour should tend to form an inversion elsewhere .
No doubt temperature inversions may be formed by a warm current of air coming from a warm quarter and overlying a cold stratum , but from my own experience I do not think this method to be of frequent occurrence in England .
It is probably a mistake to suppose that our present instruments and methods can detect changes of temperature which may well suffice to set in motion large masses of the atmosphere .
The vertical component of the velocity in the ordinary cyclone is extremely small\#151 ; a few hundred feet per hour perhaps .
This may be inferred from the rate of rainfall , and also from the velocity and incurvature of the winds .
Now on a non-rotating earth such an updraft would certainly require only a very small excess of temperature to produce it .
In actual practice the temperature depends chiefly upon the direction of the wind , the northerly winds of the cyclone being cold and the southerly warm , and the changes so produced may well mask all other effects .
Gradients at Oxshott .
The temperature gradients at Oxshott are given in Table D. It has been previously stated that the majority of these ascents were not made with the definite idea of obtaining information , but rather with that of testing the apparatus .
In consequence the kites have generally been allowed to rise as fast as they would , and have also been drawn in rapidly .
Under these circumstances while the average gradient is perfectly clear , the gradient for short steps is not so easy to determine .
For this reason it is given in the table in steps of 1000 instead of in steps of 500 metres .
But while this applies to the individual ascents , it is not likely that there is much error in the average for each step of 500 metres , and these values have therefore been inserted in Table A for the sake of comparison .
VOL. LXXVII.\#151 ; A. 2 K Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , Table D. Date .
O\#151 ; IOOO .
Wind .
Below .
Direction .
Force .
Above .
Direction .
Type of weather .
1904 .
Feb. 19 ... 0*40 22 ... 0*90 Mar. 10 ... 0*60 18 ... 0*68 22 ... 1 *04 23 ... 0*95 30 ... 1 *08 Apr. 4 ... 1 -10 6 ... 0*85 7 ... 0-75 8 ... 1*10 13 ... 0*43 28 ... 0*84 29 ... 0*33 May 13 ... 1*03 Sept. 30 ... 0*64 Oct. 3 ... 0*73 5 ... 0 '32 6 ... 0*73 7 ... 0*60 8 ... 0*70 Nov. 3 ... 0*28 8 ... 0*74 30 ... 0*48 Dec. 1 ... 0*40 5 ... 0*84 13 ... 0*53 16 ... 0*58 17 ... 0*72 29 ... 0*42 1905 .
Jan. 4 ... 0*63 5 ... 0*66 28 ... 0-44 30 ... 0*29 Feb. 3 ... 0 61 4 ... 0*56 10 ... 0*62 17 ... 0*70 Gradient .
1000\#151 ; 2000 .
0*00 0*48 0*46 0-46 0*22 S.W. 6 N.N.W. 3 N. by E. 5 S.W. 3 s.w. 4 N. by E. 6 S.W. 3 w.s.w. 7 w.s.w. 3 W.N.W. 6 W.S.W. 6 s. 7 w.s.w. 3 s.s.w. 4 w.s.w. 6 s.s.w. 4 N.E. 3 W.S.W. 4 w.s.w. 6 N.W. 5 N.W. 6 W. 3 w.s.w. 7 w.s.w. 4 w.s.w. 3 s.w. 6 N.W. 5 s.w. 6 s.w. 7 w. 5 w.s.w. 6 w. 5 w. 2 w.s.w. 4 w.s.w. 7 S.W. by W. 5 W.S.W. 3 t w. 5 W. W. W. by S. __# N.N.W. 8 .
by W. w.s.w. s.w. S.W. by W.t W. by N. at 700J W. by N. N.W. W. N.W. N.W.S N. by W. W.S.W. w.s.w. W.N.W. N.W. || N.W. by W.T N.W. W.N.W. ff W.N.W.JJ W.N.W. * * * S ** T. N.W. , I N.E , A S.W. , A N.W. , I N.W. to N.E. , I N.W. , I N.W , I N.W , I N.W , I N.W , I S.W , I S.W , A S.W , I S.W , I S.W , I N.E , A S.W , C N.W , I N.W , I N.W , T. I N.W , I N.W , I S.W , I S.W , I N.W , C S.W , I s.w. I N.W , A N.W , I N.W , C **N.W , A N.W , I N.W , I N.W , I N.W , I N.W , I * Slight inversion at 900 m. f Very dry at 1500 m , 30 per cent. J W.S.W. at 1000 .
Inversion 4 ' C. at 1200 m. , 20 per cent , humidity .
S Inversion 2 ' C. at 900 m. || Wind very strong at 800 m. , light at 1700 m. T Inversion 7 ' C. at 900 m. ; very dry .
** Unusually high barometer , 30'90 and over , ft Wind very strong at 600 m. JJ Inversion at 1800 m. , 40 per cent , humidity .
1905 .
] Gradients in the West of Scotland and Surrey .
Table D\#151 ; continued .
Wind .
Date .
Below .
Above .
Type of weather .
0\#151 ; 1000 .
1000\#151 ; 2000 .
Direction .
Force .
Direction .
1905 .
Feb. 18 ... 0*46 -0*08 S.W. 6 W.N.W.* S.W. , c Mar. 2 ... 0-72 0*00 N.N.E. 6 N.E. N.E. , I 3 ... 0*90 \#151 ; N. 3 N. by E. N.E. , A 6 ... 0*69 0*00 W.S.W. 3 N.W. N.E. , C 8 ... 0*58 \#151 ; s.s.w. 5 S.W. S.W. , C 17 ... 0-67 \#151 ; s.s.w. 6 S.W. by W. S.W. , I 25 ... 0*64 \#151 ; s.w. 3 W.S.W. S.W. , c 28 ... 0*63 \#151 ; s.w. 5 W.S.W. S.W. , I 29 ... 0-73 0*32 s.w. 5 W.S.W. S.W. , c 31 ... 0 *67 0*67 W. by S. 5 W.N.W. N.W. , A Apr. 1 ... 0-90 0*23 S.S.W. 3 W.S.W. T. A. to C. S.W. 4 ... 0*74 0*47 w. 4 N.W. S.W. , C 7 ... 1 *06 0*65 N.W. 4 N.W. by W. y. C 13 ... 0*77 0*72 S.S.E. 5 \#151 ; S.E. , C 17 ... 0*28 \#151 ; N.E. 7 E.N.E. N.E. , C 20 ... 0*64 \#151 ; N.N.E. 5 \#151 ; N.E. , I 24 ... 0*48 \#151 ; W.N.W. 3 \#151 ; N.E. , c 25 ... 0*80 0*30 S.W. 4 W. by S. S.W. , 1 27 ... 0*73 0*00 S.W. 5 W.f S.W. , I 28 ... 0*63 \#151 ; S.S.W. 7 \#151 ; S.W. , c May 5 ... 0*58 0*32 N.N.E. 4 E.N.E. N.E. , I 8 ... 1 *00 \#151 ; N.N.W. 3 \#151 ; N.W. , I June 1 ... 0*85 0*70 S.W. by W. 5 W. by N. S.W. , A 14 ... 0*69 \#151 ; N.E. 4 \#151 ; S.E , I 19 ... 0*79 0*62 S.W. 3 W.S.W. S.W. , I Aug. 10 ... 0*93 \#151 ; S.W. 4 \#151 ; S.W. , c 11 ... 0*78 \#151 ; w. 3 \#151 ; N.W. , 0 16 ... 0*70 \#151 ; E.N.E. 5 \#151 ; N.E. , I 18 ... 1-00 \#151 ; S. 6 \#151 ; S W. , c 21 ... 1 *05 \#151 ; S.S.W. 3 S.W.t S.W. , I 26 ... 1 *10 0*07 s.s.w. 3 s.s.w.S S.W. , c 30 ... 0*53 0*50 N.W. 3 N.N.W. N.W. , c Sept. 2 ... 0-65 \#151 ; w. 4 \#151 ; N.W. , I 18 ... 1*00 0*00 E.N.E. 6 E. by N. || N.E. , c 23 ... 0*62 0*73 E.N.E. 4 \#151 ; N.E. , c 28 ... 0*95 0*54 E.N.E. 5 \#151 ; N.E. , c 30 ... 0*90 0*08 N. 4 N. by W.T N.W. , I The wind force is estimated from the pull and behaviour of the kites , and from the daily weather chart .
The anemometer is too sheltered to be reliable .
# Upper wind weak , t Inversion at 1100 m. X Gradient 10*5 m. to 900 m. , then inversion 4'*5 C. Humidity 40 per cent , at 1200 m. S Inversion with dryness 40 per cent , at 1800 m. || Inversion at 1100 m. Humidity 30 per cent. T Inversion 4'*5 C. at 1700 m. , No ascent of a less height than 666 metres is included in Table D. Where one gradient only is tabulated this is the average from the surface to the 456 Mr. W. H. Dines .
Vertical Temperature [ Nov. 10 , highest point reached , which may have been anywhere between 666 and 1500 metres .
It will be noticed that nearly all the ascents have been made with winds from some westerly point ; this is because it is difficult to start a kite with an east wind owing to a belt of trees and houses which lie to the west .
But few ascents are tabulated for the summer months , owing partly to my absence in 1904 , and partly to want of sufficient wind in 1905 .
In general there is not sufficient wind for kite flying in the south-east of England during the summer .
There are in the table a few cases in which the adiabatic rate for dry air has been exceeded .
This appears to me to be an impossibility , but it is unnecessary to discuss the question , since it has been fully gone into by Mr. H. H. Clayton , * who gives what seems to be a satisfactory explanation .
The gradients at Oxshott are steeper than those at Crinan for the first thousand metres , but less steep for the second thousand .
The first fact is easily explained , for the daily range of temperature is confined to the lower layers of air , it is larger inland and the Oxshott ascents have all been made in the daytime , when the surface temperature is near its maximum value .
It is somewhat remarkable that the average gradient taken up to 2000 metres is almost identical in each case , although the values for the separate steps differ considerably .
Can it be that there is some form of compensation , so that if in one step of 500 metres the gradient is increased , the same cause or set of causes tend to lower it in another step ?
We have the following figures :\#151 ; Average to 2000 metres at Crinan in 1902 )\gt ; 0*54 0*53 " " Oxshott , 1904 and 1905 ... 0*50 The last step is missing for Crinan on August , 1903 , but up to 1500 m. it was 0*61 .
Perhaps this increased value is associated with the bad weather of that time .
Convection Currents .
It very soon becomes obvious to anyone engaged in flying kites that there are great irregularities in the direction of motion of the air , both vertically and horizontally .
The horizontal variation is shown beyond dispute by the changing azimuth of the kite , and the vertical variation may be inferred but cannot be predicated with absolute certainty by the changing altitude .
Thus , if a kite is flying with 5000 feet of line the altitude may vary from 20 ' to * 'Annals of the Astronomical Observatory of Harvard College/ vol. 58 , part 1 , pp. 14 , 15 .
1905 .
] Gradients in the West of Scotland and Surrey .
457 50 ' or even 60 ' , but these angles , at least up to 55 ' , may be dependent on the velocity only and not on any vertical component .
It has happened more than once that the angle of a kite with 3000 feet of wire has exceeded 70 ' , and these instances , which were all associated with the presence of a large massive cumulus cloud , show undoubtedly the existence of a strong vertical component .
It is seldom , however , that the angle exceeds one which might be produced naturally by a strong horizontal current at the level of the kite .
Hence on a day when the altitude is constantly changing through a large angle , although one feels certain that convection currents produce a large part of the change , it is not possible to say how large a part .
These currents are .
most likely produced by changes of temperature , and it is of interest to ascertain the precise difference of temperature between the falling and rising currents .
With this object in view I have been carefully over the records obtained on such days , but with one exception have failed to find any important difference in the temperature at a definite level when the kite was rising , and when it was falling .
This exception occurred on June 19 , 1905 , at 4.20 p.m. A kite was flying at the end of 8000 feet of wire , at a height of about 4500 feet , hidden behind a large cumulus cloud , the strain upon the wire being steadily maintained at 40 to 50 lbs. The tension in the wire rose rapidly but without jerks to 200 lbs. , and remained at this value for about 2\ minutes ; it fell back again in the same manner to 50 lbs. , at which value it remained steady .
On winding in the kite the temperature trace was found to be somewhat blotted at the critical point , as , unfortunately , too much ink had been put on the pen , but it was sufficiently distinct to give the following data .
In the course of a few minutes the kite had risen and fallen again through a vertical height of 1300 feet .
This change of height was associated with a rise of tempera- , ture of 4 ' C. , but the temperature was not symmetrical with regard to the height .
About three minutes before the updraft occurred the temperature was 5 ' C. at a height of 4500 feet , it then rose steadily to 9 ' C. at 5800 feet , the highest point reached , and declined to 7 ' C. as the kite fell back to its original level .
During the same period the relative humidity rose from 90 to 100 per cent. , and fell back to 95 per cent. In this particular instance there is little doubt that the kite was caught in a powerful ascending current .
The rise of temperature with elevation could hardly be due to the ordinary inversion , because it was associated with an increase of the humidity , whereas a temperature inversion is generally accompanied by great dryness .
It is probable that these currents are sometimes responsible for the breaking away of a kite , for this is not by any means the first time a similar phenomenon has been noted , but it is the Sir W. D. Niven .
[ Mar. 22 , first time that the kite has escaped undamaged and the trace been decipherable .
In general a steady and uniform wind is associated with steady temperature conditions , but when the temperature at a given height is subject to much fluctuation , so that the meteorograph registers different temperatures each time it passes through that height , the wind also is usually variable in direction and velocity , but these latter conditions are not necessarily accompanied by a steep temperature gradient .
The Calculation of Ellipsoidal Harmonics .
By Sir W. D. Niven , K.C.B. , V.-P.R.S .
( Received March 22 , \#151 ; Read March 29 , 1906 .
) 1 .
The object of this note is to show how ellipsoidal harmonics of the fourth , fifth , sixth , and seventh degrees may be calculated .
Some of the fourth and fifth degrees are easily found , depending as they do upon the solution of a quadratic equation .
When , however , the type of the harmonic of the fourth degree is \#169 ; i \#169 ; 2 where , if we employ G-reek letters for current co-ordinates , \#169 ; 1 = _P_ + _2 ?
1_+_\#163 ; __1 1 a2+01b2+0xc2 + t2 " 2 y2 @2 = - ?
---b H---5---1 , 2 a2+02 b2+02 e2+ and , in order that \#169 ; 1 @2 may satisfy Laplace 's equation , the quantities 0U 02 are to be found from a2+01 + b2+01 + c2+0 !
+0i-02 a ?
+02 b2+02 02\#151 ; 01 0 , 0 , ( 1 ) ( 2 ) the elimination of 92 leads to a sextic equation in 9\ .
Since , however , the roots occur in pairs as in ( 1 ) and ( 2 ) , if we put 0l + 02 = 2u , 0i \#151 ; 02 = 2v , ( 3 ) the equations for u and v derived from ( 1 ) and ( 2 ) will be of lower degrees than the sixth .
When the substitutions ( 3 ) are made in ( 1 ) the latter becomes ---- -----1----1------1----------f-- = 0 , ( 4 ) a2-\-u + v b2 + u + v c2-\-u + v v and equation ( 2 ) will be the same with the sign of v changed .

