Table of contents

The Paper describes experiments made with the object of developing suitable apparatus for the measurement of humidity at low temperatures. The three classical methods - the wet and dry bulb hygrometer, the dew-point apparatus, and the hair hygrometer - were studied and so modified that they could be used in a low temperature room inaccessible to the observer.

The common form of wet and dry bulb hygrometer is notoriousl unreliable, but the ventilated form is quite satisfactory provided the wind velocity past the bulb exceeds about 3 meters per second. A tubular form of wet and dry bulb hygrometer was developed with an electric fan for aspiring the air past the bulbs. A continuously-recording, resistance thermometer, type was also made, which had a cam device for rapidly dipping and withdrawing the thermometer bulb at periodic intervals from a reservoir of water.

The dew-point apparatus was modified into a form suitable for permanent installation; the cooling of the polished metal surface being effected by pipe connections to the cold brine circulation of the stores.

In another form a silver thimble, cooled by the evaporation of ether, was placed at one end of a long aluminium tube and a telescope at the other. By suitable arrangements of mirrors and lamps both the silver surface and the thermometer column were thrown into the field of the telescope. The end containing the thimble was inserted into the enclosure through a circular aperture. The instrument was portable and self-contained.

A distant reading type of dew-point apparatus was constructed, in which the formation of dew was indicated by the change of reflecting power of the cooled surface. A beam of heat radiation was reflected from the polished surface and concentrated on to a minute thermopile. The thermo-junctions were connected to a pivoted indicator situated at the observation station where the observer controlled the flow of cooling fluid by means of a solenoid valve. The formation of dew resulted in a sharp falling-off of the reflecting power and was indicated by the movement of the pointer of the instrument.

The hair hygrometer was studied under a variety of conditions and the changes in the calibration of this type of instrument, caused by exposure to low and high temperature, &c., were determined.

A very simple form of distant reading hair hygrometer was developed. In this the clockwork drum of the ordinary recorder was replaced by a tubular rheostat. The pointer was pressed into contact with the drum when a reading of the humidity was desired. The two portions of the resistance winding then constituted the two arms of a Wheatstone's bridge.

A number of other forms of hygrometers were also studied, notably the one utilising the heating effect on dry cotton when exposed to a humid atmosphere.

A convenient form of apparatus for the calibration of hygrometers over a wide range of humidities is described. It consists essentially of a drum fitted with a heavy glass front making an air-tight joint. The drum contains an electric fan which circulates the air over a dish of strong sulphuric acid. This method of reducing the humidity is more effective than bubbling and does not produce acid spray. The acid dish can be cut out of communication with the chamber by a ground glass cover operated by a lever projecting out of the drum.

Experiments are described showing that the formula used in reduction of data from the ventilated psychrometer applies at air speeds of 60 to 90 miles per hour without appreciable change of constants from those appropriate to a speed of 10 miles per hour.

An optical method is described for finding the concentration at various depths in a diffusing solution. The solution is contained in a closed vessel, the top and one side of which are of glass. On the glass side is a vertical scale. This vessel is immersed in another filled with water, which contains a mirror which can be rotated, and the position of which is read on a graduated scale. On a horizontal sliding stand is a telescope, which carries a horizontal wire illuminated by sodium light. The mirror is adjusted so that the image of the wire, after twice passing through the liquid, is seen on the cross wires of the telescope. The corresponding division of the vertical scale is also observed, which can be done to 0.02 mm. These measures enable the concentration to be determined at different depths, with an accuracy of about 0.05 gr. per litre.

The second part of the Paper details the method of calculating the coefficient of diffusion from a series of measures at different times. It is possible to obtain sufficient data for this in less than a day.

The results so far obtained are in good agreement with those obtained by other methods.

In 1916 (Proc. Phys. Soc., Vol. XXIX., p. 59) the author outlined a steady state null method for measuring the Thomson effect in short lengths of wire. The present Paper describes a convenient form of apparatus for carrying out measurements with rapidity and under good theoretical conditions. The short wire under test (S.W.G. 18) passes through electrical heaters which may quickly be brought to and maintained at steady temperatures differing by some 50°C. over the range 20°C. to 250°C. A short coil of the finest double silk-covered copper wire (S.W.G. 44) acts as the Thomson-Joule heat detector. Test experiments were performed with iron and constantan wires, yielding values of the Thomson effect at various temperatures over the range indicated. Many improvements in the method are described, chief among which is a greatly improved sensibility.

This communication deals with the limitations and irregularities of the Sprengel pump, resulting from the presence of air skins upon the interior surfaces. An elaborated form of the pump is described; also a plan whereby the usual gas skin may be so greatly reduced as to become a negligible quantity. The efficiency of the new pump is markedly superior to that of the older forms, and appears to remain practically constant. Two useful adjuncts to the pump are illustrated and described.

Tables have been compiled giving the corrections that have to be applied to the observed radium content of sealed platinum and silver tubes to obtain their true radium content. Two cases have been considered, namely, (1) that in which the active deposit is uniformly distributed throughout the volume of the tube such as in a full tube of salt, and (2) that in which the active deposit is uniformly distributed over the inner wall of the tube, such as in a tube containing a minute quantity of highly concentrated salt.

The results show that when the wall thickness is kept constant, the correction increases with the external diameter of the tube. Also for the same increase of external diameter, the increase of correction is more pronounced for the "empty" than for the full tube.

The methods of X-ray analysis have been applied to ice by Ancel St. John and by D. M. Dennison. The former refers the structure to a lattice composed of right triangular prisms of side 4.74 A.U. and height 6.65 A.U.; the latter to a similar lattice of dimensions 4.52 and 7.32 respectively. The actual arrangement of the atoms is not found in either case.

On certain suppositions, the arrangement can be found independently of direct X-ray analysis in the following way. Let it be assumed that each positive ion is surrounded symmetrically by negative ions, and vice versa; and in view of the low density of ice, let the number of neighbours be in each case as small as possible. The crystal is to be hexagonal, and is to have the right density. The result is that each oxygen atom is at the centre of gravity of four neighbouring oxygens, from each of which it is separated by a hydrogen atom. The dimensions of the structure agree with Dennison's figures.

The conclusion is supported by a comparison between the calculated intensities of reflection and the observed intensities as recorded by Dennison.

It is known that a plate placed close to a flanged orifice from which a stream of air or liquid is issuing is attracted towards the orifice. If the plate be mounted as a diaphragm it can be excited to strong vibration by a suitable blast, and a loud sound is produced with high efficiency.

It is known that the expression for the nth product of a series of radio-active transformations is properly represented as the sum of n terms of the type a_ne^-λ_nt. When n is known, the coefficients a_n and λ_n are determinable with comparative ease. The present Paper finds a criterion for determining n, the number of transformations, when this is not known from a priori considerations. This consists in the successive evaluation of a system of simple determinants easily constructed from the observations. The value of n is at once found from the order of the particular member of the system that vanishes.

It is shown that Fresnel's Formulæ for Reflection in Transparent Media may be treated by a simple graphical method which allows the effects of variations of the angle of incidence and of the refractive indices of the media to be traced readily.

The pump is based on a modified Sprengel action. It works automatically, the mercury being removed from the lower to the upper reservoir mixed with a current of dry air which is sucked through a side tube by a filter pump. The defects of design of former types of pump are discussed, their inefficiency at low pressures explained, and suitable simple remedies suggested. The introduction of an intermediate reservoir in the middle of the fall tube, kept automatically exhausted by the Sprengel action in the lower fall, allows the upper half to exert a positive exhaustion for every pellet of mercury falling down, even at the lowest obtainable pressure. The absence of compression in the first fall enables the maximum bore to be used for the fall tube; and hence speeds and efficiencies of exhaustion comparable to those of a Gaede pump are obtainable, though less than a pound of mercury is required to operate the pump.

The results of the many experiments on screening described in the accompanying Paper show that the most complete method of screening a valve set is to enclose it in a box made of metal of suitable thickness for the frequency used, and to seal hermetically all joints in the box and its lid. The smallest crack or hole is sufficient to allow of the escape of a detectable amount of the high-frequency energy, and it is shown that iron is far more effective than copper of the same thickness in preventing direct penetrational of radio-frequency magnetic fields through the metal.

While a practicable means of using an oscillator inside an hermetically sealed screening box is described for work of the highest precision, a much more convenient arrangement for the use of continuous-wave wireless direction-finding is described, with some results Which have been obtained with an experimental sample in actual use for this work.

The experiments herein described were carried out at the National Physical Laboratory in connection with investigations undertaken for the Radio Research Board.

The Paper consists of an investigation of the corrections applicable to the determination of the viscosity of liquids by flow through tubes, due to the abnormal flow at the ends.

General expressions for the end-corrections are obtained by the method of dimensions. These are employed in plotting the results obtained from experiments on the flow of mixtures of glycerine and water through pairs of tubes of the same diameter, but of different lengths.

The conditions that the flow at the ends may be purely viscous, and the value of the end-correction in this case, are obtained from the present experiments, and also from experiments on an orifice described in a previous Paper.

Equations are given enabling viscosity to be measured by experiments on the flow through orifices or short tubes.

Focal variations of phase in the presence or spherical aberration are calculated directly from the axial intersection points of rays of known inclination, and the graphical method employed shows that the focus for which phase variations have a minimum value may be found and suitably interpreted when the finiteness of the wave-length is disregarded. The diagrams here employed for the representation of spherical aberration have advantages over the more usual diagrams in several important particulars, and it is suggested that the older forms should be abandoned.

The Paper describes a ready method of determining the stress-optical coefficients, the Hilger interferometer being employed. Young's modulus of elasticity and Poisson's ratio are determinable simultaneously.

The present Paper is the result of an experimental investigation to determine the viscous properties of (a) carbon dioxide and nitrous oxide, and (b) nitrogen and carbon monoxide. Direct comparisons have been made in each case by observing the time required by a mercury pellet to force an equal volume of gas through a capillary tube, at atmospheric and steam temperatures. For the above groups of gases the times of fall for each gas are equal at 15.0°C. and 100.0°C., and hence, independently of any knowledge of the absolute viscosity, it has been proved that the viscous properties are identical over the above temperature range. The absolute viscosity has been obtained by comparison with air, and the mean area of collision deduced by using Chapman's formula.

A new and simple apparatus for the measurement of the capacities of electrolytic cells has been developed, which allows corrections to be made for the leakage and self-depolarisation of the cell. It gives results consistent to one or two per cent., and of the same order magnitude as those obtained by other observers.

Starting from the inference, pointed out by O. W. Richardson on the basis of electron theory, that the magnetization of a substance implies a moment of momentum about the axis of magnetization, the consequences of applying a rotating magnetic field to a specimen of magnetic material are analysed.

It is shown that magnetization in a direction transverse to the plane of the rotating field should result, in sense related to the direction of rotation of the field by the left-hand screw rule; and proportional in intensity to the first power of the angular velocity of the field and the susceptibility of the material.

Experimental tests were made, using a magnetometer, but without success, disturbing effects having not as yet been eliminated.

The absolute values of the viscosity of the gas silicane have been determined for the temperatures 15°C. and 100°C. The data have been used to calculate the mean collision area presented by the silicane molecule in the gaseous state. Use is then made of the knowledge of this dimension in an attempt to elucidate the structure of the silicane molecule in relation to those of other gaseous hydrides whose molecules have equal numbers of extra-nuclear electrons.

The pressure gradient in a liquid that flows through a conical tube is considered as depending on three effects: the variation in kinetic energy with distance along the axis; the loss due to purely viscous flow; and the loss caused by eddying. Expressions are developed giving approximately the gradient due to the three effects independently. The independence of the effects due to varying velocity and to "losses" is assumed as a working hypothesis. It is hence shown that these two effects should be capable of determination by two sets of experiments on any chosen cone, the direction of flow being reversed in the second experiments.

The effect of the entrance conditions is discussed.

The results of experiments on three ambroid cones of semi-angles 15½', 1°51' and 4°2' respectively are given. The conicality and losses effects are deduced, and these are plotted, the values for all the cones being reduced to a common basis. The results are compared with the theoretical values, experiments on a parallel ambroid tube and Stanton's experiments on parallel tubes of drawn brass. Two interesting conditions are mentioned, and a peculiarity is found in the case of divergent flow at moderate speeds through the cones of larger semi-angle.

A brief description is given of the mechanism of Prof. E. T. Whittaker which compels all exchanges between the kinetic energy of electrons and radiant energy to conform to the quantum condition. A modified form of the mechanism is suggested, consisting of two magnetons, or two ring electrons, placed near together with their planes parallel. It is pointed out that the quantum mechanism gives the law of force postulated by Langmuir in his model of a static atom, which leads to results identical with those obtained from the circular orbits of Bohr's theory.

The Paper relates to an experiment in which a neon tube and a condenser in parallel are connected in series with a high resistance and source of current.

In these circumstances intermittent current is found to pass through the lamp, and the Paper discusses the conditions governing the frequency and duration of the resulting flashes.

An analysis of the effect of residuals and earth capacities in Anderson's Inductancecapacity bridge is made, and it is shown that if balances are obtained

1. by balancing the bridge with direct currents;

2. by making the alternating current adjustments by means of a small series resistance (s') and parallel condenser (C') in the condenser arm;

then the changes required in s' and C' to hold the balance at different frequencies are equal and opposite to the variations of the effective (series) resistance and capacity of the condenser with frequency.

The assumptions made in obtaining the above conclusions are that the residual inductances and resistances of the "non-inductive" arms of the bridge are invariable with frequency and that the resistance of the inductive arm varies as the square of the frequenc y No knowledge of the absolute values of the residuals, &c., is required for the method.

The method is illustrated by results obtained with a condenser of capacity 0.5μF, and details are given showing how the chief experimental trouble, viz., drift in D.C. balance owing to temperature variations, may be overcome.

It is shown that an earth capacity acting at any point in the arm of a bridge may be replaced by two earth-impedances acting at the ends of the arm together with an impedance in series with the arm. By integration the result is extended to small distributed capacities.

Two methods are given for the elimination of the error due to the end impedances. Complete elimination can only be obtained by the use of shields connected to the ends of the bridge arm.

The Paper describes a method by which automatic voltage regulation to 0.15 per cent. may be obtained for such purposes as the operation of photometric standard lamps on an ordinary outside supply varying by as much as 10 per cent. The lamp is placed a cross an unbalanced Wheatstone bridge of which two opposite arms are composed of tungsten filament lamps. The increase of resistance of these lamps, due to the extra current passing through them when the outside voltage rises, causes a shift in the balance of the bridge such that the voltage across the photometer lamp remains unaltered if the values of the resistances in the arms be properly proportioned. The power taken is about 40 times that used in the regulated circuit.

A formula is obtained for the flow of a viscous liquid through a slightly conical tube by neglecting terms containing the square of the obliquity.

The approximation is shown to be justified by the agreement in the values of the viscosity found experimentally with tubes of differing degrees of conicality.

1. The absorption of γ-radiation in zinc sulphide has been measured by different methods, and in each case its value was found to be approximately the same.

2. No indication was observed of the occurrence of an abnormal absorption when the radioactive salt was mixed with the sulphide.

3. A table has been compiled giving the values of the ratio of the true to the apparent radium content for tubes of different diameters filled with radioactive luminous compound. This table applies to the case of a compound whose apparent density is 2.03 grammes per cubic centimetre.

4. The values of the absorption of γ-radiation from radium have been measured in different salts, particular attention being paid to the barium salts with which radium is often found in combination.

For many reasons the structure of crystals of organic substances invites examination by the methods of X-ray analysis; but their molecular complexity would seem to throw great difficulties in the way. It is possible, however, that the difficulties in the case of aromatic compounds may be surmounted by adoopting a certain hypothesis, viz., that the benzene or naphthalene ring is an actual structure, having definite size and form, and that it is built as a whole into the organic substances in which it occurs. Reasons can be given why this is a priori probable.

The examination of certain organic crystals has been made. The results are in general agreement with the hypothesis, and lead to various deductions of interest.

Measurements of the range of β-particles in various metals were made in 1910 by Prof. W. H. Bragg. The present Paper relates to a repetition of these measurements, employing 3 grams of radium bromide. The method consists in determining the ionisation in a lead chamber exposed to very penetrating γ-radiation, when various linings are inserted in it. The quantitative results obtained earlier are substantially confirmed by the present investigation.

The Paper deals with a method whereby the effects due to excessive damping in moving coil galvanometers are largely diminished. The circuit is opened after an interval of time which is small compared with the periodic time of the instrument. This gives rise to an increased first deflection, and consequently, greater sensitivity.

Analytical Treatment. - The theory indicates the greater sensitiveness of the galvanometer when the method described in the Paper is employed. For the measure ofQuantity it is equally as satisfactory and more simple in its application that the ordinary method.

Constant Time Interval of Closing the Secondary Cricuit and Constant Total Resistance. - The values of the Quantity of Electricity are directly proportional to the first deflection. The experiments demonstrate that this is true within the limits of error in observation.

Variable Resistance of Secondary Circuit. It is shown that the Quantity of Electricity, Q, is proportional to (deflection)/Rexp(at₁/R) where R is the total resistance,t₁ the time interval of closure of the secondary circuit, and α is a constant.

Constant Quantity and Variable Total Resistance. The observed first deflection is equal to β;/R exp(at₁/R) where α and β; are instrumental constants.

Experimental Confirmation of Theory. Experiments were made to test the accuracy of the theory, the time interval t₁ being determined (a) by a weight falling with uniform velocity, which strikes two levers in turn, the first lever interrupting the primary coil of a mutual inductance, and the second interrupting the secondary circuit; and (b) by a freely falling weight released by the operation of the primary switch.

The Testing of Iron. The method has been applied to the testing of iron when the magnetic force, H, was of the order 7 × 10^-6 C.G.S., and the corresponding magnetic induction, B, was of the order 1,000 × 10^-6C.G.S. The gain in sensitivity has been fruitful of greater accuracy in the results.