634 BEPOJRT— 1892. 



with a positive volume electrification. No corresponding phenomena are ohserved 

 ■when the current passes from the narrow to the wide part. 



The fact that in all cases experimented upon positive volume electrifications 

 are observed hut never similar negative electrifications is surely of significance. 



Some of the results recently brought to light by investigations on the dis- 

 charge of electricity have interesting cosmical applications. Thus it is found that 

 such a discharge through any part of a vessel containing a gas converts the whole 

 gas into a conductor.^ The dissociation which we imagine to take place in a 

 liquid before electrolytic conduction takes place must be artificially produced in 

 a gas by the discharge itself. We may imitate in gaaes which have thus been 

 rendered conductive many of the phenomena hitherto restricted to liquids : thus I 

 hope to bring to the notice of this meeting cases of primary and secondary cells in 

 which the electrolyte is a gas. There are other wa^s in which a gas can be put 

 into that sensitive state in which we may treat it as a conductor, and we have 

 every reason to suppose that the upper regions of our atmosphere are in this state. 

 The principal part of the daily variation of the magnetic needle is due to causes 

 lying outside the surface of the earth, and is in all probability only an electro- 

 magnetic effect due to that bodily motion in our atmosphere which shows itself 

 in the diurnal changes of the barometer. A favourite idea of the late Professor 

 Balfour Stewart will probably be thus confirmed. The difference in the diurnal 

 range between times of maximum and times of minimum sun-spots is accounted 

 for by the fact that the atmosphere is a better conductor at times of maximum 

 sun-spots. 



The mention of sun-spots raises a point not altogether new to this section. 

 Careful observations of celestial phenomena may suggest to us the solution of 

 many mysteries which are now puzzling us. Consider, for instance, how long it 

 would have taken to prove the universal property of gravitational attraction if the 

 record of planetary motion had not come to the philosopher's help. And surely 

 the most casual observation of cosmical effects teaches ua how much we have yet 

 to learn. 



The statement of a problem occasionally helps to clear it up, and I may b& 

 allowed, therefore, to put before you some questions, the solution of which seems 

 not beyond the reach of our powers. 



1. Is every large rotating mass a magnet ? If it is, the sun must be a powerful 

 magnet. The comets' tails, which eclipse observations show stretching out from 

 our sun in all directions, probably consist of electric discharges. The effect of a 

 magnet on the discharge is known, and careful investigations of the streamers of 

 the solar corona ought to give an answer to the question which I have put.' 



2. Is there sufficient matter in interplanetary space to make it a conductor of 

 electricity? I believe the evidence to be in favour of that view. But the con- 

 ductivity can only be small, for otherwise the earth would gradually set itself to 

 revolve about its magnetic pole. Suppose the electric resistance of interplanetary 

 space to be so great that no appreciable change in the earth's axis of rotation could 

 have taken place within historical times ; is it not possible that the currents in- 

 duced in planetary space by the earth's revolution may, by their electromagnetic 

 action, cause the secular variation of terrestrial magnetism ? There seems to mo 

 to be here a definite question capable of a definite answer, and as far as I can 

 iudge without a strict mathematical investigation the answer is in the affirmative. 



3. What is a sun-spot ? It is, I believe, generally assumed that it is analogous 

 to one of our cyclones. The general appearance of a sun-spot does not show any 

 marked cyclonic motion, though what we see is really determined by the distribution 

 of temperature, and not by the lines of flow. But a number of cyclones clustering 

 together like the sun-spots in a group should move round each other in a definite 



' An experiment by Hittorf (Wied. Ann., vii. p. 614) suggested the probability of 

 this fact, which was proved independently by Arrhenius and myself. 



" The efforts of ]\Ir. Bigelow have a bearing on this point, also some remarks 

 which I have made in a lecture before the Royal Institution {Proc. Boy. Inst., 1891), 

 but nothing decisive can be asserted at present. 



