April 3, 1908] 



SCIENCE 



539 



terval of time without being accompanied by 

 tremors which were recorded at Cheltenham. 

 When the center of a depression passed over 

 the coast line near Cheltenham the tremors 

 were much more pronounced. These tremors 

 also occurred, but with less intensity, when a 

 pronounced high area passed over the coast 

 line. They also occurred when the barometric 

 changes were such as to cause sudden pres- 

 sure changes over a large extent of coast line. 

 No tremors accompany barometric depressions 

 or sudden changes which take place wholly 

 over land, even when comparatively near Chel- 

 tenham. The period of these minute waves 

 is about 3.3 seconds and has no definite rela- 

 tion to the periods of the pendulums them- 

 selves, which varied between 18 and 28 seconds. 

 Two cases were noted when the period was 

 5.0 seconds. 



It was pointed out that a barometric de- 

 pression when over land ought to raise the 

 earth's surface on account of the reduced 

 pressure, and when over the ocean the water 

 should rise so that the pressure on the ocean 

 bed would be practically unchanged. Any 

 load applied to, or removed from the earth's 

 crust by a barometric change would have an 

 abrupt margin at the shore line. 



At the conclusion of Mr. Burbank's paper 

 Mr. W. J. Humphrey presented two papers; 

 the first paper being entitled " Anode and 

 Cathode Arc Spectra." 



In the case of direct current arcs the 

 spectral analysis of the light from the regions 

 of the two poles gives very different results. 

 When the carbons contain only small amounts 

 of metals or their salts the metallic lines are 

 practically confined to parts of the arc near 

 the negative pole, while the carbon or cyan- 

 ogen bands are most pronounced near the posi- 

 tive pole. This difference has been ascribed 

 by some observers to a kind of electrolysis in 

 the arc, causing an accumulation of the 

 metallic particles on and about the negative 

 pole. Others have considered it due to a 

 similar accumulation of the metallic particles, 

 due not to electrolysis, but to distillation, and 

 to convection. 



The author does not accept any of these 



theories as being both necessary and sufficient 

 to fully account for the phenomena observed. 

 He accepts the theory, largely due to J. J. 

 Thomson, that the arc consists mainly of nega- 

 tive corpuscles moving with great velocity 

 from the negative to the positive pole, together 

 with an approximately equal number of posi- 

 tive ions moving much slower in the opposite 

 direction. Ionization takes place mainly 

 next the positive pole, and the positive ion or 

 " rest-atom " drifts under the voltage of the 

 arc towards the negative pole, where pre- 

 sumably the corpuscles are most numerous 

 and have their greatest velocity. 



The shocks then of these " rest-atoms " by 

 the swiftly moving corpuscles is supposed to 

 be the cause of the spectrum lines, which are 

 concentrated about the negative ptole simply 

 because this is the place where the corpuscles 

 are most numerous and most energetic. 



Mr. Humphrey's second paper was devoted 

 to the subject " The Luminous Particle a 

 Strong Magnet." 



Attention is called to the fact that, so far 

 as we know, a magnetic field can act only 

 on some other magnetic field; that an electric 

 current is accompanied by a magnetic field; 

 and that a moving electric charge is an elec- 

 tric current. From this it is argued that the 

 luminous atom, whose spectral lines are 

 changed by a magnetic field, must have a 

 magnetic field of its own, due to negative 

 corpuscles in some sort of orbital motion. 



Attention is also called to the fact, shown 

 by Langevin, that a ring of electrons or cor- 

 puscles when acted on by a changing magnetic 

 field will correspondingly change its angular 

 velocity, but not its orbital radius. This sup- 

 posedly leads to a fixed self-induction for the 

 ring, and makes it possible to calculate the 

 electromotive force generated in the ring by 

 any given change in the magnetic field in 

 which the atom chances to be placed, and also 

 the resulting current. 



If the atom is constructed in general as the 

 above experimental facts would lead us to be- 

 lieve, then the change in wave-length of a 

 spectral line, when produced in a magnetic 

 field, will bear the same relation to the undis- 



