]\gt ; Mr. A. Mallock .
Note the Each angle arrived at is the result of about ten photographs , either side of the position of balance .
In the case of air , for example , four photographs were taken at and six at For the former , the differences of density , in scale divisions of the galvanometer , were .
For the latter , Each of these exposures lasted two hours in the case of air .
* The numbers given for air well illustrate the degree of concordance obtained from different photographs .
It will be noticed that a rotation of the nicol of shows the transition from too low to too high a reading very definitely .
It is hoped that the final interpolated value is not more than in error .
I am confident that hydrogen and are really in the order given .
The value of the ratio of intensities given in the second column is , of course , subject to the error of calibration , which may in some cases conspire with the error of setting .
I shall be disappointed , however , if future investigation proves the results here given to be in error by much more than 5 per cent. Note on the Influence of Temperature on the Rigidity of Metals .
By A. MALLOCK , ( Received June 9 , 1920 .
) In these ' Proceedings , ' I described some } ) eriments on the influence of temperature on the value of Young 's Modulus for various metals .
The results showed that the more fusible the metal , the greater was the variation of the modulus with temperature , and suggested that , roughly , the decrement of the modulus for a given rise of temperature was equal to the ratio of the modulus at absolute zero to the melting temperature and a constant ( i.e. Since Young 's Modulus is a complex constant , involving both rigidity and volunle elasticity , it seemed worth while to examine the temperature effect on rigidity alone , and with this object in view I have recently carried out a further series of experiments on most of the metals previously tested .
The apparatus used was a torsion-balance , shown diagrammatically in fig. 1 .
A vertical rod , , is suspended by a long fine wire , , and test For hydrogen six hours were necessary to get adequate density .
vol. , pp. 429-437 .
At the Davy-Faraday Laboratory .
Influence of Temperature on the Rigidity of Metals .
piece , , in the form of a wire or narrow strip of plate , is clamped to the lower end of , and also to the fixed suppo1t , D. The whole of this part of the balance can be immersed in a bath of fluid at any required temperature .
, carlies a horizontal arm , , and also the two light arms , G. A -shaped wire at the end of makes an electrical contact veen drops of mercury the cups , as the balance passes its zero position , and ) means of an electromagnetic ement , not shown in the figure , the two small weights , , are allowed to drop , and acting through the arms , at either end of , produce a couple tending to accelerate the rotation of A. It is arranged , by means of a relay , that this action should only take place once in each complete oscillation , and not at each passage of the -shaped wire through the mercury .
Each passage , however , is on the chronograph , as by so doing it can be seen whether the nlaintaining couple is symmetrically applied .
The natural period of the balance on the suspension wire , , is long ( about 150 seconds ) , but with the test piece in position the periods were generally of the order of from 5 to 10 seconds according to the thickness and length of specimen tested .
If is the period of the balance on , is the period of the balance with the test piece at temperature is the period of the balance with the test piece at temperature 5 2 Mr. A. Mallock .
Not the ratio of the idities of the material tested at and is .
in practice never differs much from unity .
If for example , , and , its value is about , so that the correction to ( which would be the value of were infinite ) is hardly sensible .
In making a set of experiments the oscillations of th balance were mainained throughout , the only mtity varied being the temperature of the bath in which the test piece immersed .
The conditions thus identical at all temperatures ( except as regards the small expansion or coilfraction of the test piece , which can be allowed for ) , it was hoped that correct measures of would be , and this would have been the case had the idity been the only quantity affected by of temperature .
It was soon found , however , that all the metals had considerable internal viscosity , and that the viscosity was even more ected by temperature than the idity .
This was particularly noticeable in the pure metals used in these experiments .
The lower the temperature the less the viscosity , and hence the greater the amplitude of the oscillations with a constant maintaining couple : and although in any one set of experiments the periods as measured at each temperature seldom differed from one another by more than two or three parts in ten thousand , any alteration of adjustment gave equally concordant but different values .
It eenled worth while , therefore , to examine in some detail the causes which operated to produce these differences .
In the first place , if the motion is resisted by a force proportional to the velocity , the period is dependent on and being the coefficients of Tigidity and resistance ) , and not on alone .
Also , in most , if not all , well-annealed metals , is not a constant for a definite temperature , but depends also on the absolute time for which the distortion lasts .
This is conspicuous in the case of pure rhodium , platinum , gold , aluminium , and zinc , and the same property llay be observed in many non-metallic substances .
A small tuning-fork or bar of hard pitch , for instance , will at ordinary temperatures give a definite note , showing that several vibrations have occurred before the motion is extinguished , but it is only when the vibrations are very rapid that this happens .
Thus , even were it possible to measure with suffi.cient accuracy the period of a free but rapidly decaying oscillation , the nature and magnitude of would have to be knotyn before could be determined .
When , however , is of on thoe Rigidit .
large , some form of maintenance is necessary in order that the period may be easured , and , when the maintenance is supplied , as in these experiments , by a constant couple acting for part of each oscillation , its effect on the period may be illustrated by the diagram in fig. 2 .
Let force and displacement be measured parallel to and velocity parallel to X. Also let the free oscillations be about the point O. Then the co-ordinates of the circle of radius OA will represent velocity , restitutive force and dis- placement ) an unresisted oscillation OA , where being the period .
Suppose , in the first place , that the motion experiences a constant resistance , , such as would be produced by solid friction .
If there is no nIaintenance , the result is an oscillation , still , with period , of amplitude OA\mdash ; , about from to , followed by a similar oscillatio1l of amplitude , about from to .
Thus , an oscillation with initial amplitude , OA , opposed by a constant force measured by , decreases by a constant quantity , , at each oscillation , and ceases when amplitude is so far reduced that the path cuts the circle of radius about O. At the point of the circle with radius , let a co.nstant maintaining 454 Note on Influence of Rigidity of Metatsforce , measured by , come into action .
The motion will then represented by a circle of radius about There will , of course , .
be no sudden ange of velocity at , but there is a sudden change of virtual amplitude from to , and a change of phase equal to .
Let the circle of radius be continued fro to , and there let the maintaining force cease while the resistance continues to operate .
The further motion , until changes at , is represented by a circle of radius about , and there is a change of ) has of at .
For the last half oscillation , the amplitude is If the maintained oscillations are to be repetitions of one another , B. must be equal to A. The time in each of the parts into which the complete oscillation is thus divided is , as before said .
Hence the total time between successive A is measured by the sum of the angles .
This is to when That is , the maintained period is unaltered and equal to when , and only when , the sum of the changes of phase introduced by the maintenance is zero .
If instead of a constant resistance the extinction force is proportional : the velocity the natural period is dependent on , and the circles in must be replaced by appropriate arithmic spirals , bu the above condition for no alteration of period by still holds .
All the differences found in the present experiments for values which should be identical ( differences which though small c.ompared to the periods themselves were often of the same order as differences due to variation of rigidity ) can be traced either to the unknown and varying viscous resistance which depends not only on the chemical purity of the metal , but also on its mechanical condition , on the temperature , and on the absolute period of the balance ; or to the effect of maintenance , which if adjusted for alteration as regards a certain amplitude ceases to be so when the amplitude varied .
The general conclusions , however , aslegards rigidity with those previously derived from the experiments on Young 's Modulus and show that the more infusible the metal the less is idity affected by of temperature , but I do not regard the numerical results obtained as being sufficiently accurate for publication , although the method by which they were reached and the difficulties met with are worth recording .
A few experiments were made on non-metallic substances , among which indiarubber was included .
Sir James Dewar called my attention to the great .
change which takes place in its rigidity when the temperature is low , and on Balmer of Hydrogen ectroscopic Phenomena .
455 trial I found that at the temperature of liquid air was between 800 and 1000 times as great as C. With ebonite on the other hand the difference was only about to 1 .
Preparations are being made for a further series of experiments by a method which will record resistance as well as the amplitude and period of the oscillation .
An Extension of the Series of Hydroge ?
oscoptc Phenomena long Vacuum Tnbes .
By R. W. WOOD , For .
Mem. ( Received June 21 , 1920 .
) [ PLATES 1 AND 2 .
The spectrum of hydrogen has probably received more expsrimental study than that of any other element , and yet in spite of the extended series of lines exhibited by solar hydrogen ( a Balmer series of 32lines ) , the vacuum tube has steadfastly refused to exhibit more than the first 12 members .
The solution of the problem looked rather hopeless at the start , for every possible method of electrical excitation , combined with variable pressure , various tube diameters , presence or absence of foreign gases and has been tried over and over again .
It will be remembered that by employing absorption as a method of the lines , I was able to extend the series of sodium from the eight members previously known in the emission spectrum , almost indefinitely ; 48 in the first , and 06 in a subsequent one made in collaboration with R Fortrat with the powerful quartz spectrograph of Prof. Weiss at uric further extension merely a matter of increased resolving power and length of the vapour column .
This method is impracticable , bowever , in the case of hydrogen , since this element shows only when subjected to electrical excitation , which of course introduces the same diffficulties which occur in the case of the study of the emission spectrum .
The first thing to determine , in attempting to add more lines to a series , is the factor which has heretofore fixed the limit of the series .
A somewhat careful study of the literature on the subject failed to yield any very definite information regarding this matter .
I could not , in fact , find any reproduction of a spectrogram of the series which showed more than 'Astrophys .
Journal , '

