April II. 18S9] 



NATURE 



573 



\ 



marked end a north pole is greater than that caused by an 

 equal force in an opposite direction. Again, if such a bar be 

 held horizontally east and west (to avoid terrestrial influences), 

 and tapped with a mallet, the marked end at oace becomes a 

 north pole. A similar effect follows if the rod be warmed in 

 the flame of a spirit-lamp. Lastly, if it be placed inside a coil 

 and subjected to the action of a series of rather feeble magnetic 

 forces, of equal strength but alternating in direction, the marked 

 end will generally become a north pole, even though the last 

 of the alternate forces may have tended to induce the opposite 

 polarity. 



A rod treated as above described appears to be remarkably 

 sensitive to the action of light. When such a rod is placed 

 behind the mj^netometer, and illuminated by an oxyhydrogen 

 lamp about 70 cm. distant, there occurs an immediate deflection 

 of from 10 to 200 scale-divisions,^ the magnitude of the effect 

 varying in different specimens of iron. As the action of the 

 light is continued, the deflection slowly increases. When the 

 light is shut off, the magnetometer instantly goes back over a 

 range equal to that of the first sudden deflection, then continues 

 to move slowly in the backward direction towards zero. 



The first quick movement I believe may be due to the direct 

 action of radiation, and the subsequent slow movement to the 

 gradually rising temperature of the bar. With a thick rod (l 

 cm. in diameter) the slow movement is barely perceptible, 

 extending over only one or two scale-divisions in the course of a 

 minute, the spot of light becoming almost stationary after the 

 first sudden jump. With a thin rod the sudden effect is gener- 

 ally smaller, while the slow after-effect is greater, and may 

 continue until the spot of light passes off the scale. 



As a general rule the magnetic effect is such as to render 

 the marked end of the rod a north pole : occasionally, however, 

 it becomes a south pole, but in such cases I have always found 

 that the polarity is comparatively feeble. It may even happen ^ 

 that the marked end becomes north when certain portions of 

 the rod are illuminated, and south when the light acts upon 

 other portions. This is probably due to irregular annealing and 

 a consequent local reversal of the direction of maximum suscep- 

 tibility : it indicates that the light effect is local, and is confined 

 to the illuminated surface. In one remarkable specimen, which 

 happens not to have been annealed at all, the sudden effect and 

 the slow effect are in opposite directions. When the light is 

 turned upon this rod, there is at first a sudden deflection of 

 twenty magnetometer-scale-divisions to the left, the spot after- 

 wards moving slowly and steadily towards the right. When the 

 light is shut off there occurs at once a jump of twenty divisions 

 further towards the right before the spot begins to move back in 

 the zero direction. 



Some attempts have been made to repeat the experiments 

 with light polarized by means of a Nicol's prism ; but, either 

 because the largest prism at my disposal was too small (its 

 aperture being barely 2 cm.), or because too much of the 

 radiant energy was absorbed by the spar, I failed to get any 

 magnetic effects whatever with the prism in either position. 



[Prof. Silvanus Thompson has quite recently been kind 

 enough to lend me a very large and excellent Nicol's prism. 

 From a few experiments already made with this instrument it 

 appears tbnt the action of the light is quite independent of the 

 plane of polarization.] 



There can be no doubt whatever of the reality of the effects 

 here described : they are perfectly distinct, and are at any time 

 reproducible with certainty. The only question is how much of 

 them is primarily caused by the action of light, and how much 

 by mere incidental change of temperature. But, taking all the 

 circumstances into consideration, I think the evidence is in 

 favour of the conclusion that the iustantanecus magnetic change, 

 which occurs when a prepared iron bar is illuminated, is purely 

 and directly an effect of radiation. 



Physical Society, March 23.— Prof. Reinold, President, in 

 the chair. — Prof. J. V. Jones read notes on the use of Lissajous's 

 figures to determine a rate of rotation, and of a Morse receiver 

 to measure the periodic time of a reed or tuning-fork. In deter- 

 mining resistance absolutely by the B.A. or by Lorenz's method, 

 it is important to know the speed of rotation at the instant when 

 the deflection of the galvanometer needle is observed. To deter- 

 mine this, an arm carrying a mirror is caused to oscillate by a 

 pin placed eccentrically in the end of the spindle, and a Lissajous's 



' The magnetometer mirror was 1 metre distant from the scale, and each 

 divioion = 0*64 mm. (,^ of an inch). 



" This has been observed in two specimens. 



figure is obtained by using another mirror mounted on a vibrating 

 reed driven electrically. Equality of period is obtained when 

 the resulting elliptical figure is permanent, and the frequency of 

 the reed is determined subsequently. In making the experi- 

 ments it is found convenient to control the speed of the disk by 

 braking it either with the finger or a piece of wood. The reed 

 consists of a rectangular steel rod, 100 centimetres long, and 

 section i'5l x o-6o centimetres, and the length of the vibrating 

 segment can be altered by sliding it through a clamp. To per- 

 mit of this change without altering the electrical contacts, the 

 latter are formed by two independent springs, one of which is 

 always in contact with the rod. For determining the frequency 

 of the reed a second pair of contacts are operated by the vibrator 

 and the currents recorded by a Morse receiver, whilst simul- 

 taneously the paper receives marks from a pen controlled by a 

 standard clock. The limit of accuracy of this part of the ex- 

 periment is found to be yV percent., due to changes in the 

 reed's frequency. This is a serious defect, and the author of the 

 paper asked for advice as to the precautions necessary to obtain 

 reeds of constant pitch. Prof. Ayrton, whilst recognizing the 

 extreme importance of determining speed accurately, suggested 

 that the inconstancy of the reed may be due to the impulse being 

 given at the end instead of the middle of its swing, and recalled 

 an experiment, shown before the Society by Prof. Perry and him- 

 self, in which the pitch of an electrically driven fork was varied 

 greatly by altering the adjustment of the contact screw. Refer- 

 ring to Dr. Thompson's modification, where two tuning-forks 

 drive each other, it was pointed out that the method requires 

 the synchronism of the two forks to be very exact. Mr. Blaikley 

 inquired whether any doors were opened or closed during the 

 experiments, as the pitch of a reed is affected by the size of its 

 resonance chamber. He also stated that the pitch of reeds 

 driven pneumatically could be maintained constant to I part in 

 10,000, and mentioned that two forks nearly in unison influence 

 each other's period when near together. Referring to the two 

 forks mentioned by Prof. Ayrton, Prof. S. P. Thompson said it 

 was advantageous to mount such forks on sounding boxes, for 

 when placed at a suitable distance apart they then exert consider- 

 able mutual control. The sketch put on the board by Prof. 

 Jones led Dr. Thompson to suppose that a perfect method of 

 driving forks had been devised, for two springs were .shown 

 touching opposite sides of the bar, and such an arrangement might 

 be used to complete the circuit, only when the reed is in the 

 middle position. He also believed that forks give greater con- 

 stancy than single reeds, and mentioned some recent improve- 

 ments, in which one prong of an electrically driven fork is made 

 of phosphor bronze. Mr. T. H. Blakesley, reasoning from 

 ideas suggested by Mr. Stroh's experiments on vibrating mem- 

 branes, concluded that the periods of forks, placed at \ wave- 

 length apart, would not influence each other. In reply to a 

 question from Mr. F. J. Smith, Prof. Jones thought there could 

 be no "creeping" of the reed through the clamp. He also 

 stated that he had been led to use a reed from the results ob- 

 tained in Delaney's system of telegraphy, and the fact that Lord 

 Rayleigh considered electrically driven forks satisfactory. — Dr. 

 Hofford read extracts from the following papers :— On the Clark 

 cell as a source of standard currents, by Prof. R. Threlfall and 

 Mr. A. Pollock. The authors find, contrary to ordinary ideas, 

 that Clark cells can be used to send currents of considerable 

 magnitude without the E.M.F. being appreciably changed, and 

 have constructed cells which give o-ooi amperes steadily for half 

 an hour. This result has been obtained by increasing the size 

 of the cell so that each electrode is about 5 square inches in 

 area, and the internal resistance is about 6 ohms. For the 

 ordinary small test-tube cell, the resistance of which may be 

 about 1 500 ohms, the current ought not to exceed O'ooooi. On 

 closing the circuit the P.D. (potential difference) drops almost 

 instantaneously to its steady value, and when the circuit is 

 opened rises equally rapidly to very nearly the original E.M.F. 

 The cells completely recover in time. If the current sent be too 

 large, the P.D. falls for a time, and afterwards rises and tends 

 towards a fixed value. In this respect Clark's cells are 

 greatly superior to large Daniell's, sending currents through the 

 same resistance. The paper contains several tables and curves, 

 as well as valuable results respecting the close agreement between 

 the E.M.F.'s of a great number of different cells. — On the 

 application of Clark's cell to the construction of a standard 

 galvanometer, by Prof. R. Threlfall. A large cell, as above men- 

 tioned, together with a known platinoid resistance, are used to 

 standardize a reflecting galvanometer, constructed with a single 



