TRANSACTIONS OF SECTION A. 323 



the author to draw further consequences of the hypothesis that certain auroral 

 forms are caused by a -rays. 



It is found, from the relation between range and velocity, that a-rays will 

 get down to heights varying between 70 and 300 km., which is the interval 

 actually observed for most auroral forms. Further, calculations show that 

 o-rays will strike the atmosphere at an angular distance from the magnetic axis 

 of about 17°, which gives the right position of the auroral zone. 



The question is treated whether a positive charge of the radiation is con- 

 sistent with the diurnal distribution of aurora. The distribution found for a 

 number of polar stations shows two distinct night maxima, and their existence 

 is found to be in accordance with mathematical theory. The theory, however, 

 in its present state is incomplete for a quantitative comparison with observa- 

 tions, and it cannot give any definite answer regarding the sign of charge. The 

 explanation of thin drapery bands given by Stormer, however, is in favour of a 

 positive radiation. 



In order to explain the thin drapery bands a homogeneous radiation is neces- 

 sary, and homogeneity is in fact a property of a-rays from a single radio-active 

 substance. 



The peculiarity of the draperies of showing series of equidistant bands is 

 explained by assuming the radiation to consist of homogeneous groups starting 

 under the same initial conditions. Such radiations necessary for the formation 

 of parallel bands will be present if the source on the sun contain a radio-active 

 substance and its various disintegration products. The parallel bands will give 

 a kind of magnetic spectrum of a-rays characteristic of the radio-active sub- 

 stances in the source. 



The author finds that the a-ray hypothesis, as far as our present knowledge 

 goes, explains in a satisfactory way the properties of the definite auroral forms, 

 while the assumption of a radiation of the fi-ray type meet's with eerious diffi- 

 culties. 



2. A Friction Permeameter. 1 By W. H. F. Murdoch, B.Sc, M.I.E.E. 



This instrument is a development of those described in a paper by the author 

 viz., 'Magnetic Testing of Iron.' 3 



The object of the arrangement is as follows : To keep the magnetic circuit 

 invariable during a test, to keep the coefficient of friction constant, to have uni- 

 directional sliding of the moving portion and no necessity to reset the speci- 

 men at each reading. To arrange matters so that the magnetic induction over 

 the gap is as uniform as possible, so that the law of traction is fulfilled. To 

 eliminate errors in H, the resistance of the magnetic circuit should be reduced 

 to a minimum ; the instrument also should be direct reading and portable. 



The instrument consists of a specimen of cylindrical section, magnetised by 

 a solenoid excited by an electric current, and entirely surrounded by a hollow 

 cylindrical yoke of iron. This yoke is divided symmetrically at the middle and 

 the lines of induction act on this gap together with other attractive forces. If 

 now the top portion of the yoke is rotated about the specimen as a vertical axis, 

 and slides steadily on the lower portion of the yoke, then the torque may be 

 measured, or the angular displacement between the zero and pointer be read 

 while sliding is occurring. To eliminate the friction pull previous to magnetisa- 

 tion the pointer is adjusted to read zero under mechanical friction only. Conse- 

 quently when the specimen is magnetised the magnetic pull is taken up on the 

 surfaced faces and the angular displacement is due to the magnetic induction 

 only. Therefore the magnetic induction in the test specimen can be obtained by 

 multiplying the square foot of the reading by a constant, Owing to the com- 

 pleteness of the magnetic circuit, tests at high inductions can be made and the 

 results agree closely with ballistic tests besides possessing the advantage of great 

 rapidity. ° 



1 Electrician, vol. 67, p. 930. * Proc. Inst.E.E., vol. 40, p. 137, 



y 2 



