204 The Occurrence of Ozone in the Upper Atmosphere .
By J. N. Pring , D.Sc .
, Lecturer on Electro-Chemistry in the University of Manchester .
( Communicated by Prof. Sir E. Rutherford , F.R.S. Received January 23 , -\#151 ; Read February 26 , 1914 .
) CONTENTS .
1 .
Introduction ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... .
2 .
Some Distinctive Chemical Properties of Ozone ... ... ... ... ... 3 .
The Action of Ultra-violet Light on Air ... ... ... ... ... ... ... 4 .
The Estimation of Ozone in the Atmosphere at High Altitudes ... 5 .
The Influence of Ozone in Contributing to the Blue Colour of the Sky 6 .
Summary ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... . .
Part 1.\#151 ; Introduction .
The question of the occurrence of ozone in the atmosphere has been for a long time a subject of much speculation , on account of the important bearing it has , both from physical and chemical considerations , in dealing with many atmospheric phenomena .
Many workers have shown the marked absorption effects that small quantities of ozone have on ultra-violet light.* The view has also been put forward that this gas plays an important part in contributing to the blue colour of the sky.*fThe active chemical properties of o ?
one would , moreover , if present , be expected to play a very important part in determining the salubrity of the atmosphere , on account of its intense bactericidal action , and its general property of destroying organic matter .
A'lthough a very large amount of work has been done with a view to ascertain the presence and to determine the quantity of ozone in the atmosphere , very few satisfactory attempts have been made to distinguish between this gas and other substances , such as oxides of nitrogen and hydrogen peroxide , which have very similar chemical properties .
The results obtained by different investigators are very discordant , and many of the methods used in extended measurements have since been found quite unreliable , and to give rise to misleading conclusions .
It appears now to be generally accepted that all the forms of test papers * Cf .
Kruger and Moeller , ' Phys. Zeit .
, ' vol. 13 , p. 729(1912 ) ; Fabry and Buisson , ' Count .
Bend .
, ' Vol. 156 , p. 782 ( 1913 ) .
t Ut infra .
Page ... 204 ... 205 ... 206 ... 211 ... 216 ... 218 The Occurrence of Ozone in the Upper Atmosphere .
205 and colorimetric methods which have been used so far are not applicable for quantitative results .
In some earlier work by the writer , in conjunction with W. Hayhurst , * a series of measurements on the quantity of ozone in the atmosphere were attempted .
By making use of kites and balloons , it was shown that definite small quantities of ozone occur at high altitudes , though no conclusive distinction between this gas , hydrogen peroxide , and oxides of nitrogen was made .
Thierryf had previously made measurements in the Alps at different altitudes .
The method used consisted in drawing measured volumes of air through a solution of potassium arsenite , which by reaction with ozone was oxidised to arsenate , and the remaining arsenite was estimated by titrating with a solution of iodine .
Some preliminary qualitative tests showed that there was no appreciable quantity of oxides of nitrogen or hydrogen peroxide present to complicate the reaction .
Definite quantities of ozone , which increased with the altitude , were found .
H. N. Holmes| carried out a series of qualitative tests by exposing papers impregnated with starch and potassium iodide .
Definite indications of ozone were found , the amount being greater in winter than in summer .
Part 2.\#151 ; Some Distinctive Chemical Properties of Ozone .
A large amount of preliminary work has been done to find some method of determining ozone chemically which would not be vitiated by other gases present in the atmosphere , and which would be simple in application .
1 % was necessary that the method should be capable of use in light apparatus which could be attached to small balloons .
The reagent which was considered most suitable in this application was an aqueous solution of pure potassium iodide .
It was found in these preliminary experiments that a concentrated solution reacts with ozone , even when this gas is present at high dilutions , with great rapidity .
This is also the case at | temperatures as low as \#151 ; 50 ' , when the gas is passed over the surface of the solid reagent .
As the reaction which takes place between ozone and potassium iodide is known to be very complex , it was hoped that an examination of the products obtained would enable a distinction to be made between ozone and other possible reacting gases in the atmosphere .
Reaction between Potassium Iodide , Ozone , Oxides of , and Hydrogen Peroxide.\#151 ; The measurements which were made to ascertain the reactions which take place between these gases and potassium iodide showed that , in * 'Chem .
Soc. Trans. , ' vol. 97 , p. 868 ( 1910 ) .
t ' Count .
Rend .
, ' vol. 124 , p. 460 ( 1897 ) .
t ' Amer .
Chem. Journ. , ' vol. 47 , p. 497 ( 1912 ) .
VOL. XO.\#151 ; A. Q Mr. J. N. Pring .
the case of ozone , in accordance with an effect of mass action , the nature of the products given by this reaction depends on the total quantity of this gas which has been passed , and also varies in a high degree with the concentration of ozone per unit volume of gas , even when the same total amount of ozone has been passed.* It was found that , at a dilution of 1 part in 100,000 , the reaction with potassium iodide yields hypo-iodite and free iodine , but no iodate .
The products can be simply estimated by titration with standardised sodium thiosulphate solution , and the absence of iodate ascertained by then acidifying the solution .
Any iodate present would cause an immediate liberation of iodine .
Measurements made on the influence of temperature showed that , below \#151 ; 24 ' , when the reagent was solid , the above relation no longer applied , as , in this case , iodate was formed by ozone at all concentrations .
Oxides of nitrogen were found to react with potassium iodide to give mainly iodate.at all concentrations .
In this way a basis is provided for the distinction of ozone from nitrogen peroxide when the gases are present at high dilutions , and if the temperature is not below \#151 ; 24 ' .
Hydrogen peroxide at high dilutions when reacting with potassium iodide gave products similar to those obtained with ozone , so that this reagent does not furnish a distinction between these two gases .
However , the difficulty can be overcome by means of a separate test , using titanium sulphate dissolved in sulphuric acid , which gives a yellow coloration in presence of very small quantities of hydrogen peroxide , while ozone produces no effect .
It is seen below that , by the application of these tests to atmospheric measurements at high altitudes , it was established that there was no detectable quantity of hydrogen peroxide or oxides of nitrogen present , but in all cases definite quantities of ozone were found .
Part 3 .\#151 ; TheAction of Ultra-violet Light on Air .
The formation of ozone by the action of ultra-violet light on oxygen has-been demonstrated by a number of investigators and it has also been shown that this reaction takes place when air is used in place of oxygen .
It has been observed , on the other hand , that ozone is decomposed by ultraviolet light .
Eegenerf found that an equilibrium stage is obtained when with a certain concentration of ozone no further change is produced on exposing to the light .
* For further discussion of these reactions , see Pring , ' Chem. News , ' vol. 109 , p. 73 1914 ) .
t Regener , 'Ann .
d. Phys. , ' vol. 20 , p. 1033 ( 1906 ) .
The Occurrence of Ozone in the Upper Atmosphere .
207 It was found that this equilibrium varies with the temperature and with the nature of the light .
Measurements showed that only light of wave-length below 200 / ifj .
is effective in causing the production of ozone from oxygen .
This corresponds with the observation of Kreusler that the absorption point of light by oxygen occurs at 193 / a/ a and below .
It was also found by Eegener that the rays which are effective in causing the decomposition of ozone lie between 185 and 300 / a/ a , which agrees with the measurement made by Meyer , according to which ozone absorbs light of wave-lengths up to 257 / a/ a. Meyer draws attention , to the fact that the solar spectrum ceases at about 293 / a/ a , and attributes this to the absorption by ozone in the atmosphere .
Yon Bahr found that the rate of decomposition of ozone on standing increases rapidly as the pressure is lowered , and thus considers it improbable that any considerable quantity of this gas 'can exist in the upper atmosphere .
As is recalled below , very little experimental work has been conducted to ascertain the possibility of producing oxides of nitrogen and hydrogen peroxide by ultra-violet radiation .
An investigation was made in this work on the action of ultra-violet light on air , under different conditions of pressure , humidity , etc. , in order to obtain some idea of the concentration of ozone that can be reached in this way under the different conditions , and also to see if any other products , such as oxides of nitrogen and hydrogen peroxide , are produced by this action .
The apparatus designed for the exposure of the air is shown in fig. 1 , and at A , fig. 2 .
The main feature in the source of ultra-violet light used consisted in the application of the light , derived from a mercury arc , after passing through a quartz partition , directly to the air undergoing reaction .
In this way absorption of the rays by any intervening layer of air was avoided .
The quartz tube containing the mercury arc , B , was surrounded with a quartz bulb fitted with the inlet and outlet tubes e , e , through which the air was circulated .
The outside walls of the annular vessel were cooled by submerging the whole apparatus in a vessel of water .
The arc was started by means of an induction coil in the following manner .
A potential difference of 100 to 200 volts was connected to the terminals Ti , T2 , after first passing through the lamp resistances joined in series at E ( fig. 2 ) .
The secondary circuit of the coil made connection with one of these terminals at Tx and with a wire W , wrapped round the centre of the annular jacket of the lamp .
On working the induction coil by means of a tapping key K , a glow discharge was made to pass through the lamp , which started an arc between the reservoirs of mercury at the terminals .
When the lamp was immersed in a bath of water , this method of starting was still more 208 Mr. J. N. Pring .
Fig. 1 .
The Occurrence of Ozone in the Upper Atmosphere .
209 effective than when in air , though with the water cooling a higher voltage ( 200 ) was necessary to ensure continuous working .
A Topler pump ( P ) was connected with the apparatus by means of the two-way tap ( S ) for the purpose of exhausting and drying .
The tube M could be connected with a water pump to aspirate gases at low pressures .
The reagent was placed in the wash-bottle E. When estimating ozone , the solution used consisted of equal weights of potassium iodide and water .
The water cooling of this apparatus did not of course suffice to maintain a uniform temperature of the contained air .
In addition to this factor of temperature , other conditions prevailing in the upper atmosphere could not be reproduced in these experiments .
The nature of the ultra-violet light obtained from the mercury arc was of course limited by the transparency of the quartz , and would thus be expected to differ from the case of solar radiation acting on the uppermost layers of the atmosphere .
However , it was possible to obtain conclusive results on the relative formation of ozone , nitrogen peroxide , and hydrogen peroxide , with the particular light used .
Equilibrium Value of Ozone.\#151 ; In measuring the equilibrium value of ozone produced by circulating the air at very low speeds , it was found that the temperature of the lamp quickly imposed a limit to the concentration of ozone .
The highest yield of this gas was obtained when the arc was formed intermittently for a very short interval every half minute , thus minimising the rise of temperature .
A confined volume of air in the annular space of the lamp was exposed to the radiation .
The arc was formed for 1 second every 30 seconds , and the ozone estimated after different intervals in a series of experiments .
It was found that the quantity of this gas formed increased rapidly at first and afterwards slowly until , after 40 seconds total exposure , a maximum was practically reached containing 0*15 per cent , ozone .
A similar result was obtained when the experiment was conducted with pure dry oxygen , in this case the maximum amount of ozone formed was 0'2 per cent. E. Regener ( Joe .
cit. ' ) , when investigating the equilibrium in the production of ozone from oxygen at different temperatures , found this to be 3*4 per cent , at 20 ' , and 2-7 per cent , at 54 ' .
The discrepancy is probably due to differences in the types of apparatus used , as the equilibrium value would be determined by the relative transparency of the quartz for light of different wave-lengths .
Influence of Water Vapour on the Formation of Ozone.\#151 ; It was found that a more favourable yield in the formation of ozone by ultra-violet 210 Mr. J.N. Pring .
light is obtained with dry air or oxygen than in the ease of the moist gas.* The smaller yield of ozone in the case of moist air is probably due to the formation of traces of hydrogen peroxide , which reacts with ozone according to the equation H2O2 + O3 = H2O + 202Effect of Pressure on the Formation of Ozone.\#151 ; The apparatus used for this measurement was arranged as shown in fig. 2 .
The reaction vessel and connecting tubes were exhausted by means of the Topler pump P , containing phosphorus pentoxide at O. The air was stored in the gas-holder Z , where it could be dried by leaving in contact with phosphorus pentoxide for several days .
By means of the tap S , which was lubricated with phosphoric acid , the gas , after passing through the radiation vessel , could be led through the reagent placed in the wash-bottle E. In the comparative experiments 200 c.c. of air was passed in equal intervals of time , but at pressures varying from 760 to 30 mm. It was found that a marked fall in the ozone formation occurred with reduction of pressure .
Thus , at 760 mm. , the concentration reached was 0*01 per cent. , and at 30 mm. 0*0014 per cent. Formation of Hydrogen Peroxide.\#151 ; The action of ultra-violet radiation on water has been investigated by Kernbaum , f who found that a slight decomposition slowly takes place according to the equation 2H20 = H2O2 + H2 .
ChlopinJ investigated the formation of hydrogen peroxide from moist air when submitted to the radiation from a mercury arc lamp constructed of quartz .
The air was circulated through the reaction vessel at a speed of 35 litres per hour , and then passed through a solution of titanic acid dissolved in dilute sulphuric acid .
After passing the gas for 10 minutes this reagent assumed a slight yellow colour , indicating the presence of traces of hydrogen peroxide .
In the present work experiments were made to detect the formation of hydrogen peroxide by passing 60 litres of moist air through the apparatus during two hours , and leading through a solution of titanic acid in sulphuric acid and contained in a small glass spiral washer .
No change in colour was observed .
A comparative test made by taking hydrogen * Cf .
Warburg and Leithauser , ' Ann. d. Phys. , ' vol. 20 , p. 751 ( 1906 ) .
t ' Le Radium , ' vol. 6 , p. 225 ( 1909 ) .
X *Zeit .
anorg .
Chem. , ' vol. 71 , p. 198 ( 1911 ) .
The Occurrence of Ozone in the Upper Atmosphere .
211 peroxide solution showed that it is possible to detect with certainty the presence of 0'1 x 10-6 grin , equivalent of this compound with the above reagent .
In .
60 litres of air this would correspond to a volume of l SxlO-6 per cent. The amount formed under the conditions of the above experiment must therefore be below this value , which is very small compared with the ozone .
As hydrogen peroxide is decomposed by ozone , it is doubtful whether any appreciable quantity would be permanently stable in presence of the latter gas .
Formation of Oxides of Nitrogen.\#151 ; It has not yet been established whether oxides of nitrogen are formed by ultra-violet light , though Chlopin ( cit. ) obtained evidence of the formation of traces of this gas by applying metaphenylenediamine as a colorimetric test .
This method is extremely sensitive and open to doubt when applied in this connection .
In the present work , use was made of a method devised by Fischer and Marx* for distinguishing nitrogen peroxide from ozone , which depends on the fact that , when a mixture of the gases is passed into liquid air , the former separates as a white solid while the latter dissolves .
Measurements were made in the present case with purified air which was circulated through the radiation apparatus during intervals of several hours , and then passed into liquid air .
After filtering off the solid particles , these were allowed to evaporate and pass into a gas-holder .
No nitrogen peroxide could be detected in this gas by applying potassium iodide or tetramethyl base paper .
These results show that , as in the case of hydrogen peroxide , if any nitrogen peroxide is produced by ultra-violet light , as used in these experiments , the quantity is negligibly small when compared with the ozone Part 4.\#151 ; \#151 ; .Atmospheric Measurements at Sigh Altitudes .
A. In the Alps , at Altitudes between 2000 and 3500 metres.\#151 ; The absorbing vessel which was used to contain the reagent was the same as that employed in the earlier work of Hayhurst and Pring ( loc. cit. ) , and consisted of a spherical glass bulb of 5 cm .
diameter ( B , fig. 3 ) , which was provided with an inlet and outlet tube 1 cm .
internal diameten terminating in conical funnels ( C , C ) on the outside , and projecting , on the inside , for some distance inside the bulb .
In this way a quantity of reagent could be placed inside the bulb and secured against loss when the apparatus was placed in any position .
A free circulation of air through the vessel took place , when this was placed horizontally and exposed to a wind .
The glass * ' Ber .
, ' vol. 39 , p. 2557 ( 1906 ) .
Mr. J. N. Pring .
funnels C , C , were extended by means of paper ( D , D ) , and the whole enclosed in a cage of hard spring wire ( BB ) for protection .
The glass vessel was blackened on the outside and also covered with black cloth to protect the reagent from light .
By placing a fuming liquid in the vessel and exposing to a wind , the air in passing through was seen to assume a rapid circulatory motion inside .
The efficiency of the absorption of ozone during this passage was tested by taking a vessel of similar construction , placing the potassium iodide reagent inside , and then joining the vessel in series with a wash-bottle containing the same reagent .
On passing ozonised air through the two vessels in succession , and varying the speed of circulation , it was found that even at higher velocities than those subsequently given by the atmosphere , more than four-fifths of the ozone was in every case absorbed by the first vessel .
Volume of Air Circulated through the Reaction Vessel during a meant .\#151 ; An approximate calibration of the volume of air circulated , which was assumed to be directly proportional to the velocity of the wind , was made by placing some pure benzene in the vessel , and after exposing for definite intervals to a wind of known velocity , noticing the loss in weight .
The vessel was first calibrated in terms of the readings given by a portable anemometer kindly lent by the Meteorological Office , and then by comparison with a standardised anemometer at the Manchester University Observatory ; the readings were reduced to miles per hour .
Results of Calibration .
Temperature of benzene .
Units , anemometer scale .
Loss of weight of benzene .
Loss of weight per unit of anemometer scale .
Vessel 1 18 -0 ' 135 grm. 5 -30 grm. 0 0392 " i 18-0 95 4-70 0*0495 " i 16 -5 70 2*67 0 -0382 " 2 17*0 81 3 -55 0-0440 " 2 17 -0 97 4*35 0 0449 Mean 17 *0 \#151 ; \#151 ; 0-0433 Since the vapour pressure of benzene at 17 ' is 66-5 mm. and its molecular weight 78 , it follows that 1 grm. of the saturated vapour will occupy The Occurrence of Ozone in the Upper Atmosphere .
213 3*29 litres .
The quantity displaced during a reading of 1 unit on the anemometer , viz. , 0*043 grin .
, will therefore correspond to 0*141 litre .
The calibration of the anemometer showed that 1 unit corresponded to a horizontal flow of air of 0*027 mile , so that a flow of wind corresponding to a displacement past the apparatus of 1 mile corresponds to a circulation of 5*12 litres in the vessel .
Results of Measurements made in the Alps.\#151 ; The tests , which lasted several days , were made near Scheidegg at an altitude of 2l30 metres and at a point near the Jungfraujoeh of 3580 metres altitude .
Hydrogen Peroxide.\#151 ; This was tested for by exposing titanic acid in presence of sulphuric acid to the wind in an apparatus similar to that used for the ozone estimation .
The colour of this reagent remained quite unchanged after exposing for two days at the different altitudes and under different conditions of weather , thus showing that there is no appreciable quantity of hydrogen peroxide in the atmosphere .
It was noticed , however , that freshly fallen snow or hail gave a very marked coloration with the reagent .
No tests were made with glacier water , though this would be expected to retain the hydrogen peroxide associated with the snow .
Nitrogen Peroxide.\#151 ; In the experiments tabulated below , where estimations have been made of ozone , it was found that in no case was any potassium iodate formed on exposing the iodide in the reaction vessel to the atmosphere .
Time of measurement .
Altitude .
Units , anemo- meter .
Litres of air circulated .
Total volume of ozone ( at N.T.P. ) .
Volume of ozone per unit volume of air .
Aug. 22-23 , 1913\#151 ; 7.0 p.m. to 8.0 a.m metres .
2130 110 15 *2 c.c. 0 05 0*000003 Aug. 23\#151 ; 8.0 A.M. to 7.0 P.M 2130 813 113 0-31 0 *0000027 Aug. 23\#151 ; 24\#151 ; 7.0 P.M. to 8.0 a.m 2130 3176 438 0-87 0 *00000197 Mean value at 6970 feet ( 2130 metres ) 2 *5 x 10~6 Aug. 24\#151 ; 6.0 to 7.30 p.m 3580 1195 165 0 62 0-0000038 Aug. 25\#151 ; 8.50 a.m. to 5.45 p.m. ... 3580 10694 1470 7*1 0-0000048 Aug. 26 \#151 ; 6.0 a.m. to 6.0 P.M. * 3580 1520 210 1*18 0 *0000056 Mean value at 11690 feet ( 3580 metres ) ... ... ... ... ... . .
4 *7 x 10-6 _____ \#166 ; ______ . .
r .
\#166 ; 1 ______ 214 Mr. J. N. Pring .
As pointed out above this shows the absence of any appreciable quantity of oxides of nitrogen .
Ozone Estimations.\#151 ; These were made by placing about 7 c.c. of a mixture of equal weights of potassium iodide and water in the reaction vessel and placing on the ground horizontally in the direction of the wind .
The anemometer was erected in the neighbourhood .
The mean quantity of ozone , viz. , 2*5 x 10-6 at 2130 metres and 4*7 x 10~6 at 3580 metres , are in close agreement with the results obtained by de Thierry ( loc. cit. ) , which gave 1*6 x 10-6 at 1050 metres and 4*4 x 10-6 at 3020 metres .
B. Measurements made with Free Balloons.\#151 ; For these measurements use was made of the hydrogen balloons which are used in meteorological investigations at the Manchester University .
These balloons with the attached instruments rise to an average height of about 10 miles and then burst .
The deflated skin retards the rate of fall of the instruments to the ground .
A knowledge of the height attained and the temperature is obtained by a recording baro- and thermograph .
The reaction vessel for the ozone tests was of the same form as that used in the tests at ground level , and was suspended vertically from the balloon together with the other instruments .
A rough calculation of the amount of air which would pass through the vessel during an ascent and descent was made from the following data:\#151 ; From the calibration made above it was seen that the exposure of the vessel to a flow of wind corresponding to a displacement of air of ,1 mile caused the passage of 5*12 litres .
Since 1 mile = 1*6 x 105 cm .
, it follows that a displacement of 1 cm .
at atmospheric pressure will cause the circulation of ( 5*12 x 108)/ ( 1*6 x 105 ) = 0*032 c.c. On the assumption that the volume circulated is proportional to the displacement through the air , it follows that during an ascent and descent the mass of air passed through in grammes is given by 2 ( px \#151 ; p2 ) x 13*6 x 0*032 , or 0*87 ( pi\#151 ; p2 ) , where pi is the atmospheric pressure in centimetres of mercury at ground level , p2 that at the highest level reached , and 13*6 the density of mercury .
The volume circulated in litres ( measured at KT .
P. ) is therefore 0*675 ( px\#151 ; pz ) .
The values for the pressure at different altitudes have been taken from meteorological tables , as reproduced on p. 215 .
At a height of about 6000 metres the temperature is always below the freezing point of the reagent ( \#151 ; 24 ' ) so that reaction must take place with the solid .
It was seen in Part 1 that under these conditions the method applied did not enable a distinction between ozone and oxides of The Occurrence of Ozone in the Upper .
215 nitrogen .
However , in all measurements made up to 6500 metres , the latter gas was found not to be present in any appreciable quantity .
Altitude .
I Pressure .
| Altitude .
Pressure .
Altitude .
Pressure .
metres .
mm. metres .
mm. metres .
mm. 1000 673 4000 458 401 8000 267 1500 631 5000 9000 233 2000 586 6000 353 11000 168 3000 522 7000 308 20000 39 .
, 1 The reagent placed in the vessel consisted of 5 to 7 c.c. of the 50-percent .
potassium iodide solution as used in all the other measurements .
After being returned by the finder through the post the amount of solution remaining usually amounted to about 2 c.c. The values given in the tables are calculated on the original quantity of reagent .
Time of ascent .
Mean direction of wind .
Height attained .
Estimated | volume of air circulated .
Weight df iodine liberated in total reagent .
Potas- sium hydrate formed .
Mean concentration of ozone in 1 volume air .
Free I2 ( + hypo-iodite ) .
Combined i3 ( as iodate , etc. ) .
j Aug. 3 , 1909\#151 ; kilom .
litres .
mgrm .
mgrm .
mgrm .
9.0 p.m W.S.W. 16 -0 46 -6 0*12 0*13 4 *8 x 10 " ' i March 18 , 1910 \#151 ; 12.0 NIGHT N. 13 *5 43 -8 0*21 0-13 0-004 6-8xl0-7 9 P.M ST .
8*5 34 7 1 02 0*70 Nil 4-4xl0-6 !
| March 19\#151 ; 1 2.0 A.M N. 19 -5 48 ; 1 0 *40 1-12 2-8x10-6 | May 18\#151 ; I \#166 ; . . . . .
9.40 p.m S. 17-0 47 5 0*76 0-76 2 -8 x 10-6 | | May 19\#151 ; 2.10 A.M S. 12-0 43-8 0*13 0-34 trace 9-5x10-7 6.30 a.m S. 20-0 48 *3 0*21 0*40 1-1x10-6 !
| March 1 , 1911\#151 ; 5.30 p.m N.W. 19*0 48*2 0*303 0 682 Nil 1 -8 x 10~6 : 1 May 4\#151 ; 1 7.15 P.M N.W. 20*0 48*3 0-612 2-36 Nil 5-4x10-6 1 -i Aug. 6 , 1913\#151 ; 8.0 p.m N.N.E. 6*5 29*8 0-416 Nil 1-2x10-6 1 Mean ... ... 2 -1 x IQ-6 i It is seen that , except in one case , a large part of the reaction had resulted in the formation of iodate .
The one case in which this compound had not * formed was where the ascent had only reached an altitude of 6*5 kilometres , where the temperature indicated was \#151 ; 31*5 ' .
Since the cryoscopic point of the reagent is \#151 ; 24 ' , complete solidification had probably not occurred .
Mr. J. N. Pring* The fact that , as in the case of all measurements at lower altitudes , there was no iodate formed in this experiment , indicated the absence of oxides of nitrogen , and the formation of iodate in experiments at greater altitudes is presumably due to the reaction of ozone on the solid reagent , as was established in laboratory experiments .
By considering these results together with those made on ground level at altitudes up to 3*5 kilometres , the conclusion may be drawn that there is no detectable amount of hydrogen peroxide or oxides of nitrogen in the higher atmosphere , but that there is a definite small quantity of ozone .
The amount of this last gas increases with the altitude up to 3'5 kilometres .
The mean values estimated in the measurements made in the Alps were 2-5 x lO-6 in one part of air at 2-5 kilometres altitude , and 4*7 x 10-6 parts at 3-5 kilometres .
In the measurements made with the balloons in this country , the mean volume of ozone between ground level and altitudes up to 20 kilometres gave an average value of 2T x 10"6 .
Even after allowing for the absence of this gas at lower levels on account of organic impurities , this value would indicate that there is no very great increase in the amount of ozone in the atmosphere between altitudes of 4 and 20 kilometres .
The amounts measured , however , are known to be sufficient to exhibit marked absorption and chemical effects , and thus play an important part in the permeability to radiation and in the chemical purification of the atmosphere .
Part 5.\#151 ; The Part played by Ozone in Determining the Blue Colour of the Sky .
The blue colour of the sky has been ascribed to a number of operating causes .
Firstly , the turbidity of the air or presence of small particles of dust has been shown to exert a selective reflection on light of different wavelengths .
Secondly , it has been pointed out that the molecules of the air are themselves able to scatter radiation in a similar manner to the action of small particles of dust.* Thirdly , the presence of ozone has been put forward by a number of investigators to account for this colour.f The coloration has also been attributed to fluorescence phenomena produced by this gas .
J * Of .
Rayleigh , ' Phil. Mag. , ' vol. 47 , p. 375 ( 1899 ) ; Schuster , ' Theory of Optics , ' p. 328 ( 1909 ) ; 'Nature , ' vol. 81 , p. 97 ( 1909 ) ; King , 4 Phil. Trans. , ' A , vol. 212 , p. 375 ( 1913 ) .
t Of .
Chappuis , ' Comptes Rendus , ' vol. 91 , p. 985 ( 1880 ) ; Hartley , * Chem. Soc. Trans. , ' vol. 39 , p. Ill ( 1881 ) ; Spring , ' Bull , de l'Acad .
Beige ' ( 3 ) , vol. 36 , p. 504 .
J Cy Lallemand , 'Comptes Rendus , ' vol. 75 , p. 707 ( 1872 ) ; Hartley , 'Nature , ' vol. 39 , p. 474 ( 1889 ) . .
The Occurrence of Ozone in the Upper Atmosphere .
217 It has been found* from spectrophotometric measurements on skylight that the distribution of intensity observed could not be altogether accounted for by the assumption of an atmosphere conforming to Rayleigh 's formula nor of a turbid medium containing coarser particles .
These , and the results of other investigators , indicate that there is a blue absorption colour of the air .
In the present work the results of the approximate measurements made on the quantity of ozone in the atmosphere at altitudes ranging up to 20 kilometres supply data which enabled a rough comparison to be made in the laboratory of the depth of colour given by ozone of equivalent amount .
For this experiment a glass tube of 2'8 metres length and 4 cm .
diameter was taken .
The walls were provided with side tubes , one near each end , to enable the passage of the ozonised gas through the tube .
The two ends of the main tube were covered by thin plates of glass , which were cemented by sodium silicate solution so as to make air-tight connection .
The outside of the tube was wrapped with black paper , and a white paper disc placed over one of the end plates .
On illuminating this by daylight and sighting through the other end , the intensity of coloration produced on admitting ozone of known concentration could be observed .
The ozone for .
this purpose was prepared from oxygen by passing through a number of annular glass tubes where it was exposed to the silent electric discharge produced by an induction coil .
After time had been allowed for the composition of the gas in the sighting tube to become uniform , an analysis of the gas was made by connecting the outlet side tube to a vessel containing acidified potassium iodide solution , and allowing a known volume of the gas to pass .
It is seen from the atmospheric measurements that , in this country , an approximate concentration of two parts of ozone in one million of air was found as the average amount of this gas between ground level and an altitude of 20 kilometres , while the amount found in the Alps at an altitude of 3| kilometres was five parts in one million .
Taking this second value , and allowing 8360 metres as the height which would be occupied by the atmosphere if at N.T.P. , this concentration of ozone in a vertical section of the atmosphere is equivalent to a layer of the pure gas of a thickness of 4'2 cm .
( at N.T.P. ) .
The results given in the tables below record the observations made with the sighting tube when filled with oxygen containing different concentrations of ozone .
The thickness of the layer of pure gas which is equivalent to this concentration is also given .
* Nichols , * Phys. Review , ' vol. 26 , p. 497 ( 1908 ) .
218 The Occurrence of Ozone in the Upper Atmosphere .
Percentage concentration of ozone in oxygen .
Equivalent thickness of layer of pure ozone .
Colour observed .
0-20 cm .
0-55 Colour uncertain .
0-36 1*0 Faint bluish-green .
1*7 4-7 Marked blue colour .
2-8 7 *8 Deep steel blue colour .
It is difficult to compare the colour of the gas in a tube of this nature with that of the sky on account of a large influence exerted by the nature of the illumination .
The results obtained , however , indicate that the presence of ozone is a contributing factor to the blue colour .
This work is being extended with a view to determining more precisely the quantity of ozone in the upper atmosphere , its variation from time to time and with altitude , and its influence on the constant of solar radiation .
Summary .
A method has been devised which enables an estimation to be made of ozone when at very high dilutions , and at temperatures down to \#151 ; 60 ' .
With very dilute gases , the reagent used , consisting of an aqueous solution of potassium iodide , enables a distinction to be made between ozone and oxides of nitrogen , if the temperature is not below the freezing point of the reagent ( -24 ' ) .
By applying a source of light given by a mercury arc in a quartz vessel it was found that ozone is formed up to a certain equilibrium stage .
The rate of formation of the gas is only slightly lowered by the presence of water vapour , but decreases rapidly with falling pressure .
By exposing air to the source of light used , no formation of oxides of nitrogen or hydrogen peroxide could be detected .
Estimations made of the quantity of ozone in the atmosphere at high altitudes in this country showed an average amount of 2'5 x 10-6 in one volume of air , and that no very large increase occurs between 5 kilometres and the highest altitude reached , viz. 20 kilometres .
In the Alps , a mean value of 2'5 x 10-6 was measured at an altitude of 2100 metres , and 4*7 x 10-6 at 8580 metres .
No detectable amounts of oxides of nitrogen or hydrogen peroxide were found in either of the last two cases .
It was found by colorimetric measurements in the laboratory that this last amount of ozone gives a marked blue colour , thus showing that ozone must be regarded as a contributing factor in determining the blue colour of the sky .
Diffraction of Light by Spheres of Small Relative Index .
219 In conclusion I wish to express my thanks to Sir Ernest Eutherford for the kind interest he has taken in this work , to Miss M. White for the loan of meteorological balloons , and to the Director of the Jungfrau Eailway for travelling facilities he generously extended .
Part of the incidental expenses of this work have been defrayed by a Government Grant from the Eoyal Society .
On the Diffraction of Light by Spheres of Small Relative Index .
By Lord Eayleigh , O.M. , F.E.S. ( Received February 4 , \#151 ; Read February 26 , 1914 .
) In a short paper " On the Diffraction of Light by Particles Comparable with the Wave-length , " * Keen and Porter describe curious observations upon the intensity and colour of the light transmitted through small particles of precipitated sulphur , while still in a state of suspension , when the size of the particles is comparable with , or decidedly larger than , the wave-length of the light .
The particles principally concerned in their experiments appear to have decidedly exceeded those dealt with in a recent paper , f where the calculations were pushed only to the point where the circumference of the sphere is 2'25 The authors cited give as the size of the particles , when the intensity of the light passing through was a minimum , 6 to 10 fi , that is over 10 wave-lengths of yellow light , and they point out the desirability of extending the theory to larger spheres .
The calculations referred to related to the particular case where the ( relative ) refractive index of the spherical obstacles is 1*5 .
This value was chosen in order to bring out the peculiar polarisation phenomena observed in the diffracted light at angles in the neighbourhood of 90 ' , and as not inappropriate to experiments upon particles of high index suspended in water .
I remarked that the extension of the calculations to greater particles would be of interest , but that the arithmetical work would rapidly become heavy .
There is , however , another particular case of a more tractable character , viz. , when the relative refractive index is small ; and although it may not be * * Roy .
Soc. Proc. , ' A , vol. 89 , p. 370 ( 1913 ) .
t ' Roy .
Soc. Proc. , ' A , vol. 84 , p. 25 ( 1910 ) ; ' Scientific Papers , ' vol. 5 , p. 547 .

