Table of contents

The Paper gives an account of experiments which have been made upon the electrical conductivities, the corrosion products and tensile properties of high purity aluminium and certain light aluminium alloys, which have been exposed to London atmosphere for a period of 24 years. The elements concerned are copper, nickel, manganese and zinc in varying amounts up to a few per cent. The variation of electrical conductivity, the growth of the corrosion product and its composition have been studied, and the tensile properties before and after exposure have been compared. Finally, there is a note on the corrosion products of high conductivity copper.

The study of optical spectra of positive rays is continued with the type of tube and Merton wedge described previously. The direction of observation is here longitudinal, in order to isolate the Doppler effect in the "moving" spectrum: accordingly a six-prism spectrograph is substituted for the échelon used with a narrower line in Part I. The results are added to those of Stark, Paschen, Wilsar and Krefft on the same spectrum, and the uncertainties in identifying atomic and molecular carriers are analysed. The distribution of velocity among the particles contributing to the "moving" spectrum is compared with the distribution of velocity in positive rays measured by the electromagnetic method and with some investigations of Retschinsky and Döpel. The appearance of the many-lined spectrum of hydrogen in the positive rays, and the ratio of intensity of the "moving" and "resting" spectra are also considered. The inverse square law of probability of electron capture, and some consequences of the work of Wien and Rüchardt, are the most likely controlling factors in the several phenomena.

Thin Films. - The method of making thin films of collodion by drying weak solutions in ether and alcohol on the surface of clean mercury, and their subsequent removal, is described. Flexible films are also made in the same way, using a modified formula, producing very robust films which withstand great distortion, and will stand being punctured without fracture. Both types can be produced rapidly and cheaply down to a thickness of one wavelength of light.

Sputtering. - Two novel types of sputtering apparatus are described, together with working data. A table is given containing sputtering data for 25 metals, nine of which appear for the first time. Notes relating to the effects of Shape and Age of Kathode, Shape of Anode, Form of Dark-Space and Kathode Temperature appear, together with novel modifications in the design of the Wehnelt break. Approximate temperature coefficients of resistance are assigned to eight new metals and more accurate values to three others, that relating to gold being new. The apparatus used for these determinations is described. Notes and curves on ageing appear regarding six metals and resistances for films of standard size for ten metals. A curve is given throwing, perhaps, some light on the formation of a sputtered film.

Bolometer. - Apparatus is described by which rings are moulded for supporting the thin collodion films and providing a reliable contact for overlying sputtered metallic films. The method of making the bolometers, together with holders of various types, and apparatus for blackening, are detailed. The comparative sensitivity of these new bolometers is discussed, and curves and an empirical formula given from which an estimation of the speed of the instruments can be determined.

The new type is roughly 400 per cent. faster than a representative bolometer of the Lummer and Kurlbaum type. The thickness of the component metal film is estimated.

The variation of refractive index of glycerine solutions in equilibrium with air of various humidities has been studied; the time for equilibrium to be reached, when thin films of glycerine are used, has also been investigated.

It is concluded that this property may be used very conveniently in a hygrometer.

Changes in the density and resistivity of tungsten during swaging and drawing have been investigated.

The densities of pure and thoriated tungsten (0.63 per cent. thoria) from the sintered bar down to wires 0.02 mms. in diameter are given. The experimental advantages of the use of benzene as the displaced liquid are mentioned.

Measurements of the densities of specimens cleaned in successive stages by etching reveal the presence of low density surface layers.

Variations of heat treatment likely to occur in practice are shown to have no appreciable effect upon the density of worked tungsten.

The density rises rapidly during swaging to a maximum value within 0.5 per cent. of the density of the perfect tungsten crystal. Further working produces a steady fall in density, which becomes more marked in the finest sizes.

Measurements have been made of the resistivity of the same specimens.

The resistivity falls rapidly in the early stages of working and reaches a minimum when the density is a maximum. It then increases at a uniform rate, which is approximately an exponential function of the diameter

This Paper gives the results of determinations of liquid transport produced by passing an electric current through aqueous solutions of copper sulphate divided perpendicular to the flow of electricity by a diaphragm of powdered glass. The probable factors that determine the liquid transport are investigated, over a wide range of concentrations, by an examination of the dependence of the flow at constant applied voltage and of the electric charge of the diaphragm on the electrolyte concentration, and also by a study of the relation between the flow per faraday and the dilution of the electrolyte.

It is shown that accurate measurements of input admittance (or of input impedance) under various conditions can be made by means of the Schering Capacity Bridge. The input circuit is regarded as being equivalent to a condenser with a definite phase angle, ϕ, or "loss angle," δ=90° - ϕ, and the results are expressed by stating the effective capacity and value of tan δ for each set of experimental conditions.

A series of measurements made on an R valve is recorded, and it is shown that the results are in good agreement with the theoretical investigations of Miller and Nichols. The theoretical investigation has been extended to allow for the effect of dielectric losses in the valve, since these were found by experiment to be rather large, and to have an appreciable effect on the capacity and phase angle of the input circuit.

It is shown that the effective capacity may vary from about 10μμF. to 100μμF. for an R valve, and the phase angle may vary from about 80° leading to 126° leading, depending mainly on the load in the anode circuit. Values of phase angle greater than 90° correspond to a negative resistance, or negative power factor, and occur when the load in the anode circuit is inductive. The variations of input capacity and phase angle with filament voltage, anode voltage, input voltage, and frequency are also investigated.

A brief résumé is given of the procedures which have been developed for determining the distribution of direct current in systems of linear conductors. In this connection reference is made to practically all the laws, theorems, principles and procedures generally considered to pertain to this particular field of investigation. Consideration is then given to a principle which when employed usually leads more directly to the solution of problems than does any of the procedures commonly used.

This principle applies to systems of linear conductors in which the currents are proportional to the impressed electromotive forces; the electromotive forces may be any function of time, and may be distributed in any manner throughout the system; and the branches may contain resistance, inductance, capacitance or any two or all of these in series, may be so arranged as to move with respect to a permanent magnet, thus developing counter electromotive forces, and may be connected by contacts or mutual inductances or both of these. For such a system of conductors the current in any branch is that which would result if all impressed electromotive forces were replaced by a single impressed electromotive force, located in the particular branch and equal to the drop in potential which originally would have appeared across the break had this branch been opened. While this principle is a logical consequence of well-known laws, it has been used but very little and seems to be practically unknown. It is shown here that it may be used to advantage in all or practically all cases in which the conductors form a series-parallel combination or a network which may be changed to a series-parallel combination by opening the branch in which it is desired to determine the current.

A bridge is described by which quick measurements can be made of capacitances covering a range of from 1μμF up to 30μF, the power factor also being indicated. The unknown capacitance C is put in parallel with a resistance P, and the effective self-inductance of the combination, which is approximately equal to -P²C, is read on a mutual inductometer forming part of the bridge. By giving P a series of suitable values scale multipliers providing for a very wide range of capacitance are obtained. The inductometer used is of a new type, having a circular scale extending to about 260°, the percentage accuracy of reading being almost constant over the greater part of the range. This novel scale system allows the lower readings to be taken with good accuracy. A small rheostat allows the power loss in the condenser to be balanced, and enables the power factor to be deduced.

The Paper gives an account of experimental work which has led to an extension of our knowledge of the second spark spectra of In and Ga, into the region of long wave-lengths. After making a careful study of the spark spectra of the elements in hydrogen and air under different conditions of excitation the author has identified the lines corresponding to the second series in the spectra of doubly-ionised Indium and Gallium. The principal features of the spark spectra are now known, and all these structures are in beautiful accord with the quantum theory of spectral line emission.

To help in the detection of series in In II attempts have been made to find regularities in Sn III. Visual observations under varying conditions of excitation revealed a strong triplet in the region λ5100. The detection of two more triplets, particularly a diffuse multiplet, confirmed the above identification. A member of the first fundamental series has also been found with the aid of the corresponding members in In II and Ga II. A possible triplet of the first sharp series is suggested, and evidence of its correctness is sought by correlating it with corresponding members of related spectra.

The Paper describes measurements of the secondary E.M.F.s displayed by selenium cells and the apparently abnormal effect of illumination on the corresponding secondary currents. Both are shown to be probably due to an invisible water film in parallel with the selenium. That this film is the principal seat of the polarisation seems to be proved by the fact that the polarisation disappears almost completely with prolonged drying. On this basis a satisfactory quantitative explanation of the various effects is given. An important practical consequence of the desiccation of the cell is the large increase obtained in its sensitivity to light.

The design of a chronograph which will give a time trace subdivided into units and fractional units requires the association of a rotating system and a vibrating or time-keeping system.

It a low-frequency vibrating system be employed, the rotating system, although keeping good average time, may show exaggerated errors in unit subdivisions due to acceleration and deceleration between periods of control.

The Paper discusses the association of a high-frequency synchronous alternating-current motor with a high-frequency valve-maintained tuning-fork.

It deals with the design of valve-maintained motors and condenser-tuned valve-maintained tuning-forks and gives expressions for the frequency of the maintained system for motors coupled to tuning-forks both without and with tuning condensers.

A new spectro-photographic method of measuring absorption coefficients is described. A portion of the slit of a spectrograph is covered by the absorbing medium, and in front of the photographic plate or of the slit, a neutral wedge is placed. The resulting spectogram consists of two parts, one due to light which has passed through the filter and the wedge, the other to light which has traversed the wedge only. It is shown how, from a knowledge of the distance between two points, one in each part of the spectogram, which are of the same density, the absorption coefficient of the filter at any wavelength may be determined. The source of light need not be constant.

The thermal separation obtained in gas mixtures containing nitrogen is compared with that obtained in mixtures containing carbon monoxide. A similar comparison is made of mixtures containing carbon dioxide or nitrous oxide. The gas analysis required in the measurement of the effect is made by means of the Shakespear katharometer. The behaviour of nitrogen is found to be similar to that of carbon monoxide. The effect given by carbon dioxide is generally a little greater than that given by nitrous oxide.

The pairs of gases examined provide a special case for the application of the Enskog-Chapman theory, as in each pair the molecular weights and mean collision areas are the same. It can thus be deduced that the molecular field of nitrogen is similar to that of carbon monoxide, and that the field of carbon dioxide differs little from that of nitrous oxide.

The Paper gives an account of an investigation on the time for which the continuous discharge of a neon lamp may be interrupted without putting out the lamp. This time, which has been found to be of the order of 50 micro-seconds, depends on the voltage across the lamp and also on the current passing through it.

1. Observations on electrification of various dusts have been made by using a large volume of air at a pretty high speed. The climatic conditions are totally different from those under which other similar observations were made.

2. The number of substances tried is considerably increased. In some cases the sign of electrification is the same as that observed by Rudge.

3. The factors governing the phenomena of electricity of dust storms are stated as follows: (1) Material of the dust. (2) Its size. (3) The gas raising the cloud. (4) The velocity of the gas. (5) The temperature. Account of the first two factors is given in the present communication.

4. It appears that the electricity developed is of frictional nature, which itself is, of course, unknown.

5. Some quantitative estimates of electrification of chlorides and nitrates of sodium and potassium are made. It is seen that chlorides of sodium and potassium are equally efficacious, whilst sodium nitrate is about 4½ times as efficacious as potassium nitrate.

6. Dusts of various roughly uniform sizes were obtained with the help of a few graded sieves. The magnitudes of the various sizes were obtained after very laborious observations under microscope. It is shown graphically that, other things being the same, the number of volts developed by blowing increases very rapidly as the size grows less.

In this Paper some preliminary experiments are described with a mechanical device vibrating a piston at one end of a tube so that a sound wave is emitted at the other. The particle velocities in the sound wave are measured by a Rayleigh disc, and resonance curves are drawn. It is shown how the energy in the sound wave can be calculated from the results obtained. The experiments are directed towards obtaining a standard source of sound, and the results are encouraging, but a considerable amount of work remains to be done.

A description is given of apparatus employed for measuring coefficients of sound-absorption by the stationary-wave method. The apparatus differs from that used by earlier workers in the use of (1) a small tuned hot-wire microphone for determining relative pressure-amplitudes in the sound-waves; (2) the employment of a steady valve-driven source of sound with arrangements for maintaining the strength at a constant value; (3) the screening of source and experimental pipe from disturbances due to the movements of the observer. By the employment of a certain procedure the relation between the response of the microphone and the amplitude of the pressure-variation in the sound-wave is eliminated. Some examples of the employment of the apparatus for determining the coefficients of absorption at normal incidence of acoustic plasters and hair-felt are given.

The Paper describes a robust form of the Ewing Ball and Tube Flowmeter. suitable for the metering of gases or liquids under pressure, as, for example, the ammonia in a refrigerating plant. The necessary pressure-tight joints for connecting the conical tube to the circuit are described, and also a device for cutting off the flow should the tube fail. An investigation is included of the behaviour of the instrument under practical conditions of pulsating flow.

The instrument described is based on a principle which has not hitherto been utilised in the design of gas analysis instruments. A quartz crystal is maintained in vibration piezo-electrically, and stationary waves are set up in the gas between the flat surface of the crystal and a movable reflector. The position of the nodes is recognised by the reaction on the quartz crystal, resulting in an increase of the current in the maintaining circuit. The distance from node to node is a measure of the composition of the gaseous mixture, assuming it is composed of two gases which do not react.

An account is given of experiments which have been carried out using a "balance method" to obtain the "J" discontinuity which Barkla has found with heterogeneous X-rays scattered by elements of low atomic weight.

In a series of experiments, extending over several months, no such discontinuities have been found.

In view of the repeated occurrence of the phenomenon in the laboratories of Prof. Barkla, it is suggested that some condition which has not yet been published is required for its production. The intention is expressed of further investigating this effect by the method which is described in the Paper.

This Paper extends the results obtained in a previous Paper to the case of rectifying valves working at low voltages with unsaturated electron currents. The most economic conditions are briefly discussed, and it is shown that, as far as the limited information with respect to the life of the modern valves is concerned, it is probable that long life should be provided for.

Previous work on this subject is reviewed, and the results of brightness measurements on zinc sulphide compounds containing radium are given for periods up to 4,000 days. From a comparison of the brightness curves of compounds made with the same luminescent material, but with different radium concentrations it is shown that the brightness-time relationship is of the form B=rf(rt), where r is the radium content.

The experimental facts available for developing and testing a theory of the luminescence are summarised, and a number of different theories, including a modified form of the recovery theory, are considered.

Finally it is shown that the observed brightness curves are in excellent agreement with Rutherford's original theory of the destruction of active centres, provided this be combined with a simple hypothesis as to the cause of the progressive increase in the light absorption of the material which has been found experimentally. This leads to the following form of the brightness-time relationship - log {B/(b+B)} +kt+a=0, where a, b and k are constants, of which the last two are proportional to the radioactive concentration for any given grade of luminescent material.

From the results obtained on two grades of compound it is concluded that (a) the rate of destruction of active centres is six to nine times that of the ionisation of inactive molecules, and (b) in the new material used for the compounds measured, 20 to 50 per cent. of the molecules are in the active state.

In the first section of the Paper the experimental arrangements employed to obtain accurate resonance curves are described. In the second section resonance curves with increasing exciting currents show increasing distortion until an unstable state of affairs is arrived at in which the amplitude for a given current over a certain frequency range can have two different values, depending upon whether the frequency has been approached from above or below. The indication here is that a decrease of resonant frequency takes place with increase of amplitude, and it further appears that an increase of damping also takes place. Experiments with a free fork showed that these effects were still present, though not to the same extent as with the driven fork. Static experiments showed a departure from the straight line law both in the case of the deflection of the fork prongs for given loads, and the flux change through the core for a given deflection of the prongs. In section III the effect of such departures on the equation of motion is considered mathematically; it is shown that the term depending on the cube of the amplitude is the most important and that its inclusion gives rise to resonance curves of the form observed. The effect of the static square term is shown to be in effect a dynamic cube term. In section IV the experimentally obtained resonance curves are examined in the light of the theory, and substantial agreement is found, while in section V a detailed calculation of the various constants is made. In section VI another effect of the non-linearity is found in the possibility of producing fundamental frequency vibrations in the fork by means of exciting currents of double frequency, and the effect of a first harmonic in an exciting current of fundamental frequency is considered.

In section VII distortions of a second type are shown to consist of "coupled circuit" effects: these at large amplitudes are modified by distortions of the first type.

The key employs mercury contacts and is enclosed in a glass casing. The two upper fixed contacts consist of downwardly projecting copper rods, the two lower fixed contacts of cups containing mercury, and the two movable contacts of downwardly projecting copper rods (which enter the fixed cups in one position of the switch) surmounted by cups containing mercury (which receive the fixed rods in the other position of the switch). The connections are those of a reversing switch, and the middle contacts can be raised or lowered by turning a vertical shaft which is retained in either position by a jockey spring.

The measurement of the effective inductances of "four-terminal" resistances or "shunts" is discussed. The method of measurement developed in the Paper is an application of the Kelvin Double Bridge, used with alternating current, the phase angle adjustment being obtained by condensers shunting the ratio arms. The procedure to be adopted for eliminating or evaluating corrections due to "leads" is described, and the advantages of the bridge in being free from stray fields, practically independent of frequency, capable of use with almost any desired current strength, and, moreover, very easy to work, are pointed out. Results of measurements made by this method on a series of typical shunts are recorded.

The Paper summarises the results obtained as to the diurnal variation of magnetic disturbance during 1912 and 1913 at Cape Denison (67° 0' S. lat., 142° 40' E. long.), the base station of the Australasian Antarctic Expedition. The results are compared with corresponding results from Eskdalemuir and Cape Evans, the base station of the British Antarctic Expedition. The Paper also discusses the relations between magnetic disturbance and aurora at Cape Denison.

The Paper gives particulars of regularities observed in the spectrum of tin corresponding to the trebly-ionised stage of the element. Pairs of sharp and diffuse series are discovered which converge to a common limit with a sixteen-fold value of the Rydberg constant. The fundamental and super-fundamental series doublets are also indicated. Term values are evaluated from the related system of members by assuming a value for the term "G."

Great efforts have been made in the determination of the wavelengths of the lines of the tin spark spectrum. The plates were finally measured from spectra taken in the second order of a 5-ft. concave grating and an accuracy to about 0.05 Å.U. is claimed.

The experimental method of sorting out the lines due to this high stage of ionisation (Sn IV) is the same as that adopted previously by the writer in the case of Sn, Ga, In.

The latent heat of evaporation of sulphur was determined by finding the loss of weight of a vessel full of sulphur when energy was dissipated in it at a known rate. Heat losses were prevented by immersing the vessel in the vapour of sulphur; a method of correction for the heat loss by having a dummy vessel with a smaller heating coil, and using the differences in energy and in weight evaporated, for the two vessels, was tried but gave unsatisfactory results.

The value found for the latent heat was 79, the accuracy being estimated at 2 per cent.

The refractive index of gaseous carbon tetrachloride has been found for the green mercury line (λ5461), the result being expressed in connection with the density of the gas, that is, so as to show the refractivity of the gas by the same number of molecules as 1 c.c. of hydrogen contains at N.T.P.

Adopting the value 1.001799 for the refractive index under these conditions, the dispersion of the gas has been studied over the range λ4800 to λ6700, and is represented by the expression

µ - 1 = 13.543 × 10²⁷/(7831.7× 10²⁷ - ν²)

ν being the frequency of the light.

A number of lines lying between λ7282 and λ3142 occurs in the condensed discharge spectrum of the more volatile gases of the air. These have been attributed to singly ionised neon, and the present Paper comprises an analysis which confirms this view.

This Paper is an extension of the investigation, started by the author and Mr. Vick, of the justification of employing Newton's Law of Cooling when estimating the external heat loss from the surface of a solid in a gaseous medium. In this case, the loss in a vertical system, containing in turn CO₂, oxygen and nitrogen, was examined, the apparatus consisting of compensated platinum wires mounted along the axes of ice-water cooled glass tubes and connected to the terminals of a Callendar-Griffiths bridge. The results show that, for tubes between 1 cm. and 3 cm. diameter, the law is applicable only as far as 10°C. for pressures between 1 atmosphere and 6 cm. mercury. The deviation above this temperature depends on the width of tube and the density of the gas, being greater for CO₂ than for oxygen or nitrogen.

The interference and diffraction effects of sound waves of frequencies above 4,000 per second are studied quantitatively, using as detectors both sensitive flames and the ear.

A telephone energised by oscillatory currents produced by a thermionic tube circuit is used as a source of sound. This source has distinct advantages over other high pitch sources such as the Rayleigh bird call and other forms of whistle. Many of the experiments can be done with the sensitive flame fed by gas at ordinary pressures.

Two forms of sensitive flame are employed, one responding to points of maximum air density variation, and the other to maximum motion of the air.

The experiments on interference show results analogous to the optical interference experiments of Lloyd, Fresnel and Newton, and very close agreement is obtained between measured and calculated positions of points on the hyperboloid fringe systems. An experiment is described in which results analogous to certain Heaviside layer effects are obtained.

Diffraction near various obstacles and a sound zone plate are studied.

Some observations are recorded on the fatigue of the ear and binaural audition.

This Paper deals with certain small errors attendant upon the use of the rider of a precision balance, and shows how such errors may be rendered negligible. Attention is also drawn to some disadvantages resulting from the employment of velvet-covered pan arrestors and a plan is described for eliminating them.

1. When one insulating solid is rubbed along and perpendicular to an identical one they acquire charges, -ve and +ve, respectively.

2. Continued rubbing so affects the surfaces of "rubber" and "rubbed" that their properties change, with reversal of sign to +ve and -ve respectively.

3. The effect 2 is permanent, unless the surfaces be restored by heat treatment.

4. Oblique impact of two identical solids generates on them unequal charges. The net charge is, in general, -ve.

The present Paper embodies an attempt to consider the problem of diffraction of light by a very thin plate of transparent material, bounded by a straight edge, with greater exactness than is attained in the usual elementary treatment on the Fresnel-Huygen principle. The method adopted, though not completely rigorous, bases itself on the electromagnetic theory of light, and seeks to express the disturbance in the field in the form of functions which are solutions of the equations of wave-propagation. The formulæ obtained indicate that the light diffracted by the edge should exhibit colour and polarization effects varying in a remarkable manner with the thickness of the plate and the direction of observation. Effects having the general character of those indicated by the theory have actually been observed in experiment. The theory, however, requires modification in the case of thicker laminæ, where further complications arise which are not here taken account of.