Table of contents

The absorption of radiation from the sun in an atmosphere varying exponentially with the height is considered, as in a former paper; but here the earth's curvature, and that of the level layers in the atmosphere, is taken into account. The values of this absorption as previously calculated are valid so long as the sun's zenith distance does not exceed 75°, but for greater zenith distances the necessary corrections are of importance. It is shown that the absorption, and resulting ionization or dissociation of the air, should begin to increase before ground sunrise, the interval varying from about 10 minutes at the equator to about an hour at 60° latitude.

An account is given of a linear neon-lamp time base by means of which cathode-ray oscillograph traces due to recurring electrical or magnetic phenomena can be accurately superimposed, even though the train frequency be subject to wide and irregular variations. The underlying principle of this time base lies in the use of a glowing neon lamp which is extinguished by energy furnished either by the phenomenon under investigation or by some other phenomenon associated therewith.

A description is given of an X-ray crystallograph designed on the principle used by Seeman and Bohlin in which the slit, specimen to be analysed and photographic film are disposed along the circumference of a circular camera. Typical crystallograms obtainable are shown.

The cubic-crystal analyser is described and some examples demonstrating its use are given.

A new and convenient form of apparatus is described for the measurement of liquid-transport produced by the application of an electric field across a diaphragm of parchment paper. Results obtained with solutions of copper sulphate over the concentration range 0.0005-normal to 1-normal are compared with corresponding results obtained previously, by the author, with a diaphragm of powdered glass. The liquid-transport per faraday with the parchment diaphragm decreases strongly with increasing concentration up to 0.2-normal, but maintains practically a constant value over the concentration range 0.4-normal to 1-normal. This value is compared with that deduced for the electrolytic water-transport per faraday from the values of the experimental Hittorf transport number; and evidence is adduced indicating that there is no increase in the electrolytic water-transport per faraday with decreasing concentration from 0.4-normal. The data for the solutions over the concentration range 0.0005-normal to 0.2-normal appear to be consistent with the hypothesis of a composite effect produced by a constant electrolytic water-transport per faraday (of the same magnitude as the observed transport per faraday over the concentration range 0.4-normal to 1-normal) and by a specific (electro-endosmotic) action of the parchment diaphragm. The determination of electrolytic water-transport is utilized to correct the values of the experimental Hittorf transport number obtained from the data of Hittorf and Metalka. The true transport number in the case of copper sulphate solutions decreases with decreasing concentration from 1-normal, but the values at concentrations 0.125-normal and less appear to be little different from that at infinite dilution.

A study of the spark spectrum of antimony under varying conditions of excitation has led to an extension of the existing classification of the spectrum of Sb IV. Many new lines are classified as secondary members of the series. A second P term has been established for the Sb V spectrum.

It is shown that the previously reported intensity-modifications, found to take place in the high-frequency electrodeless discharge in mercury vapour at low pressures (0.001 mm.), can be completely explained in terms of recently determined electron-impact excitation curves. The modified spectrum arises from the fact that the mean free path is large, enabling the electrons to attain high velocities which result in the strengthening of lines involving upper singlet levels. The high-frequency spectrum at a pressure of 3 mm. is found to be arc-like, supporting the above conclusion.

The fine structure of the unclassified line λ 6123 (Hg I) is determined with a Fabry-Perot interferometer and found to have eight components. Four of these form a multiplet and may arise from one isotope only. The line λ 6123 probably involves an upper singlet level. The unclassified spark line λ 4797 (Hg II) shows four fine-structure components, widely separated and of similar intensities.

A considerable number of new heads in the known system of bands in the extreme red, due to BaF, have been observed. A new system, also due to BaF, has been discovered in the near infra-red, and assigned to the transition ²Π → ²Σ.

The refractive index of gaseous n-pentane is found to be 1.001683 for the green mercury line, λ 5461, the result being expressed in relation to the density of the gas. The dispersion over the range λ 4358-λ 6708 is represented by the Sellmeier equation: (μ-1) = 14.605 × 10²⁷/(8978.4 × 10²⁷ - ν²), where ν is the frequency of the incident light. Revised values for gaseous chloroform are: μ = 1.001448 for λ 5461 and (μ - 1) = 15.391 × 10²⁷/(10933 × 10²⁷ - ν²).

The refractive index of gaseous ethyl bromide in relation to density is found to be 1.001261 for the green mercury line, λ 5461. The Sellmeier formula representing the dispersion over the range λ 4358-λ 6708 is: (μ - 1) = 12.407 × 10²⁷/(10138 × 10²⁷ - ν²), ν being the frequency of the incident light.

In Heydweiller's modification of the Carey Foster bridge as ordinarily used (with variable mutual inductance) the capacitance of the condenser can be read directly, but the power-factor has to be deduced by calculation. To facilitate direct reading of both capacitance and power-factor the author has developed two bridge systems, A and B, based on that of Carey Foster. In both of these a fixed mutual inductance M is used and there is no added resistance in the condenser arm.

The capacitance C eq, falling dots M/PR, where P and R are the resistances of the other two arms. To give direct reading for C, P can be varied in simple steps (e.g. 10, 100, 1000, 10,000), giving range multipliers, while R consists of a conductance box reading in millimhos. In system A a fourth arm Q of variable low resistance is added to the bridge. The power-factor is given by Q/Mω and can be read directly (at given pulsatance ω) if a slide wire is used for Q. In system B the power loss in C is balanced by the addition of impurity to the mutual inductance M by means of a closed-loop circuit variably coupled to both the primary and secondary coils of M. When the resistance of this loop circuit is set proportional to the frequency, the scale of the double inductor which varies the couplings can be graduated to read the power-factor directly. When a simple amplifier is used in the detector circuit, a capacitance range of 100μμF up to 10μF can be obtained and a power-factor range from 0.0001 to 0.01 with high accuracy of reading.

The magnetostriction of unannealed wrought iron is measured, by means of an optical lever of length 0.12 mm., up to an intensity of magnetization of 985 e.m.u. The longitudinal extension and the intensity of magnetization are measured for a cycle of magnetization, both being found to show hysteresis. The extension is approximately proportional to the square of the intensity but seems to depend slightly on the previous magnetic history as well as on the intensity of magnetization. This hysteresis indicates that the extension is a consequence of the magnetization, in the same sense that the magnetization is a consequence of the applied field.

A development of Callendar's continuous-mixture method for the measurement of the total heat or specific heat of a liquid is given. It consists of the determination of the amount of heat extracted by a stream of water from a hot stream of the liquid. A discussion of the heat losses and the method of reducing and allowing for these is included, as well as experimental tests of the theory of the method.

A plate method for the determination of thermal conductivity and its variation with temperature between 0 and 100° C. is described, Although specially adapted for medium conductors (of the order 0.001 to 0.02 c.g.s. units) such as rocks, it can be used for substances of lower conductivity.

This paper describes an investigation of the attenuation of radio waves, of wave-lengths between 5 and 10 metres, when transmitted directly along the earth's surface. A brief description is given of the transmitters employed, one being a fixed installation used with an input power supply of the order of 500 watts, the other a transportable set operated with an input supply of about 50 watts obtained from batteries.

The observations of field-intensity were obtained by measurement of the audio-frequency voltage across the telephones of a simple two-valve loop receiver, a brief description of which is given. Measurements were carried out at both Slough and Teddington of the field-intensity at various distances from the transmitter up to 700 metres. In some of the experiments a negative attenuation effect was observed in the radiated field over a distance of about 4 wave-lengths. Later measurements carried out under more favourable conditions did not show this effect, which was, therefore, attributed to the interference of waves reflected from trees and buildings in the neighbourhood of the transmitter. Some qualitative observations were made at distances up to 20 miles with a single-loop direction-finder. These observations showed that the signal intensity on such short wave-lengths depends to a great extent upon the existence of obstacles in the path of transmission. The signal-intensity at a distance of 20 miles over a direct air line was of the same order as that obtained at 4 miles for transmission along the ground.

A comparison has been made between the experimental results and those calculated from a simple wave-attenuation theory, the electrical constants of the earth being taken into account. As a result of this comparison, the value of the conductivity of the earth appears to lie between 5 × 10⁸ and 30 × 10⁸ e.s.u. for the frequencies of 30 to 60 mega-cycles per second employed. These values correspond to resistivities of from 1800 to 300 ohm/cm.³ The most suitable value of the dielectric constant of the earth is about 10, although the experimental method does not enable this to be obtained with any great accuracy.

A brief description of some experiments carried out with a single-loop direction-finder on the wave-lengths under consideration is appended to the paper.