NATURE 



I^May 3, 1906 



Lightning Flashes. 



In your issue of January 14, 1SS6 (vol. xxxiii., p. 245), 

 Mr. T. Mackenzie reported lightning from a banlv of cloud 

 to the clear sky, but, as it was quite dark, one cannot 

 be certain that there were no indistinct outliers. In 

 Hann's " Lehrbuch der Meteorologie " (ed. i, p. 632) 

 other cases of lightning from a cloud to the clear sky are 

 referred to. 



On the evening of March 26, at 6.30 p.m., before dusk 

 had set in, there was a large thunder cumulo-nimbus cloud 

 about eight miles north of Johannesburg. The summit of 

 this cloud was very sharp against a clear dark blue sky. 

 There was no false cirrus. Six flashes of lightning darted 

 from near tlie summit of the cloud into the clear sky. 

 The longest path was' about ten degrees. One flash re- 

 turned to the cloud, the others finished in the clear sky. 

 Before dusk set in this phenomenon ceased to occur. AH 

 the flashes were directed to that part of the sky from which 

 the cloud moved. 



In a well-known book on meteorology we read " it 

 is impossible to say whether a flash of lightning moves 

 from a cloud to the earth or in an opposite direction," and 

 further that the lightning is instantaneous. Hann does 

 not confirm these statements, and it is time that they were 

 modified in English text-books. Quite frequently I have 

 observed lightning flashes leaving a cloud for the earth, 

 but fading away before reaching it ; the opposite pheno- 



ging, 1903. 



menon has not been observed. The paths of lightning 

 shown by photographs taken in the Transvaal all indicate 

 discharges from cloud to cloud, and from cloud to earth. 

 The enclosed photograph, taken by Mr. T. N. Leslie at 

 V'ereeniging, is typical. Some flashes of lightning are 

 instantaneous, the majority are not, but I do not think 

 any exceeds a duration of a third of a second. The re- 

 volving wheel has been used, and shows that the duration 

 js often certainly much longer than i/4oth of a second. 

 Johannesburg, April 2. R. T. A. I. 



Diurnal Variation of lonisation in Closed Vessels. 



Until Messrs. Campbell and Wood give us some more 

 •definite information as to the magnitude of the daily 

 variation which they have found in the natural ionisation 

 of air in closed vessels (Nature, April 19, vol. Ixxiii., 

 P- 583). 't is somewhat premature to go into a detailed 

 discussion as to how this discovery will affect theories of 

 atmospheric electricity. Still, the letter in N.^ture of 

 April 26 (vol. Ixxiii., p. 607) on this question from Dr. 

 O. W. Richardson calls for some remarks. 



The facts are shortly : — (i) Messrs. Campbell and Wood 

 <liscover that the natural ionisation of air in a closed 

 vessel has a double daily period, the maxima being between 

 S a.m. and 10 a.m. and between 10 p.m. and i a.m., the 

 corresponding minima being at 2 p.m. and 4 a.m. ; (2) the 

 potential gradient in the lower atmosphere has, at most 

 places, also a double period, the maxima being at about 

 8 a.m. and 8 p.m., and the minima at about 4 a.m. and 

 midday. Thus, allowing for a certain amount of un- 

 certainty in the exact determination of the times of the 

 maxima and minima, we mav sav that the dailv variations 



of the natural ionisation and the potential gradient are 

 similar. 



In order to discuss a possible dependency of these two 

 factors. Dr. Richardson assumes that " the distribution of 

 the earth's field reduces itself to a case very similar to 

 that between two plane electrodes immersed in a gas and 

 maintained at a constant difference of potential." It is 

 more than questionable as to whether this assumption is 

 justifiable or not, for in atmospheric electricity we are 

 dealing with constant quantities of electricity, and not with 

 constant potentials. But, rather than follow up this objec- 

 tion, I would prefer to look at the problem from a different 

 point of view, and show that the exact contrary conclusions 

 can be deduced. 



In discussing this problem, it is usual to accept that 

 there is a negative charge on the earth's surface, and 

 that the corresponding positive charge is a volume charge 

 distributed in the atmosphere. Now all the measure- 

 ments which we have of the daily variation of potential 

 .gradient have been made within a few metres of the sur- 

 face. Within these few metres there can be, relative to 

 the charge on the earth, very little volume charge, so 

 what our measurements actually refer to is the charge on 

 the surface, the relation being dv/dh= —i^ttit. The point 

 to notice in this is that, with a given charge on the surface 

 and the corresponding charge in the atmosphere above, the 

 vertical distribution of the charge and the conducting 

 state of the upper atmosphere do not in the slightest affect 

 the potential gradient within a few metres of the surface. 

 If the potential gradient changes there it can only be by 

 a change in the surface charge on the earth. 



If there is a penetrating radiation which, besides 

 ionising the air in closed vessels, also ionises the air in 

 the atmosphere, we should expect from Messrs. Campbell 

 and Wood's experiments the ionisation of the air in all 

 parts of the atmosphere to have a daily variation. Thus 

 the air quite near the surface would twice a day be 

 exceptionally conducting ; one would expect that at these 

 times there would be a greater loss of the surface charge, 

 and so the remaining charge to be diminished, and with it 

 the potential gradient. The consequence would be a daily 

 variation of the potential gradient corresponding to the 

 variation of ionisation, but the inaxima of one correspond- 

 ing to the minima of the other. 



That such a relation does exist between the ionisation 

 of the lower atmosphere and potential gradient has been 

 shown by many observers situated in most parts of the 

 globe. Thus from Messrs. Campbell and Wood's results 

 one would expect minima of the potential gradient to occur 

 at about 8 a.m. and 10 p.m. : this is the exact reverse of 

 what reallv occurs. 



Thus it would appear as if Messrs. Campbell and Wood 

 have added one more to the many puzzling factors con- 

 nected with atmospheric electricity. 



Manchester University. George C. Simpson. 



Aueust Rainfall. 



According to Greenwich experience, August has been a 

 very dry month considerablv oftener about sun-spot 

 maxima than about minima. This fact may be of some 

 practical interest. 



Using Mr. Nash's table (from 1815), let us confine our 

 attention to the three years about tlie eight maxima and 

 the three about the eight minima, i.e. twenty-four years 

 in each division. 



The driest August in the miniyna division was in '55, 

 with 1-40 inches. But in the ma.xima division there are 

 ten cases of lower values, ranging from 125 inches down 

 to 0-4^ inch, viz. '38, '49, 'jq, '61, '6q, '71. '82, '83, '84, 

 '93. Since 1837 no three-year group of this division has 

 been without at least one such very dry August, two have 

 had two. and one three. 



The total August rainfall in those twenty-four-year 

 groups is. in the sun-spot maxima division. 502; inches, 

 in the minima division 66'50 inches, the higher value thus 

 showing an excess of 16-25 inches (nearly one-third of the 

 lower). 



The sun-spot maximum we are now near (190:;?) has not 

 been here considered, but I mav remark that in 1904 we 

 had one of those low August values (124 inches). 



Alex. B. MacDowall. 



NO. 1905, VOL. 74I 



