ately 1, and, for a given electric field, the same value of R is obtained for different gases over a large range of pressure; this indicates that there are no free electrons. In dry gases, the values of R are much smaller than in moist, for certain ranges of electric force and pressure. Also, in dry gases the divergence obtained with a given electric force increases as the pressure is reduced. Thus the value of k is large and increases as p is reduced. It can hence be shown that the mass of the charged particles must be small compared with that of a molecule, that is, electrons move freely in the gas. From a comparison of the divergences of streams of electrons in different gases at 20 mm. pressure, it has been shown that, with the same field and pressure, the energy of agitation of electrons is greater in helium than in nitrogen, and greater in nitrogen than in hydrogen. U is found to be (1-15×104xVk) cm./sec. The velocity W in the direction of the field has been determined experimentally in the same apparatus by superposing on the electric field a magnetic field at right angles, and measuring the deviation thus produced in the electron current. W is not proportional to Z, but is a function of the ratio Z/p. Agreement between the values of U and W for a given value of Z/p for a large pressure range indicated the absence of ions in nitrogen, carbon dioxide, carbon monoxide, argon, neon, and helium. Ions were detected in oxygen, nitrous oxide, and nitric oxide. The mean free path of the electrons in different gases for several values of Z/p has been deduced, and also the proportion λ of the energy of an electron lost by collision with a molecule. M. S. BURR.

Evaporation of electrons. W. SCHOTTKY (Z. Physik, 1925, 34, 645—675).-It is not the specific heat of the electrons in a metal but the change in the specific heat by loss of electrons which determines the variation with temperature of the thermodynamic potential of the electrons in the metal. This change of specific heat is related to the Thomson effect and the thermo-electric force between the same metal hot and cold; the latter can exist only when the effect of heat is to produce in the metal an internal state differing from the normal state.

E. B. LUDLAM.

Absorption coefficient for slow electrons in mercury vapour. R. B. BRODE (Proc. Physical Soc., 1925, 38, 77-79).-See A., 1925, ii, 1020.

Collisions of the second kind. H. D. SMYTH (Proc. Nat. Acad. Sci., 1925, 11, 679-682).-The electron affinity of iodine has been employed in an attempt at the direct verification of the existence of collisions of the second kind. If an electron comes within the field of a neutral iodine atom, and before the electron affinity of 3-5 volts is fully satisfied, i.e., before the process of binding the electron is completed, a second electron "collides" with the new

system, the second electron may gain kinetic energy by a collision of the second kind. Experiments have been carried out which, although not conclusive, support the idea. Preliminary work on the decomposition of ozone affords promising but not yet definitive results. R. A. MORTON.

Relative production of negative and positive ions by electron collisions. F. L. MOHLER (Physical Rev., 1925, [ii], 26, 614-624).—In an investigation of the ionisation produced by electron collision, current-voltage curves were obtained for mercuric chloride and hydrogen chloride which were similar to those for mercury and iodine, respectively, the negative ion current being small compared with the positive for electron currents of the order of 10-6 amp., and pressures of about 0.001 mm. The results do not support the theory that electron collisions dissociate polar molecules into positive and negative ions, but indicate that the primary effect is the production of a positive molecular ion. A. A. ELDRIDGE.

Free path of slow protons in helium. A. J. DEMPSTER (Nature, 1925, 116, 900-901).-Positively charged hydrogen atoms with velocities acquired by falling through 300-900 volts possess an unexpected range in helium and other gases, and they remain charged throughout the entire path. Apparently no free electrons are produced by protons of the velocity used; the result is anticipated in the case of helium from energy considerations. Singly-charged helium atoms disappear when the pressure is increased to 0-027 mm. of mercury; charged hydrogen molecules, which disappear at 0.07 mm., are probably dissociated on collision without alteration of velocity or direction. A. A. ELDRIDGE.

Ions from hot platinum. H. A. ERIKSON (Physical Rev., 1925, [ii], 26, 625-628).-Measurements of mobility and of current-voltage relations indicate that the single negative ion and the final positive ion from platinum are air ions; the initial positive ion may also be an air ion (cf. ibid., 1923, 21, 720; A., 1925, ii, 6). A. A. ELDRIDGE.

Mobility of the ions of the active deposits of thorium and radium. H. A. ERIKSON (Physical Rev., 1925, [ii], 26, 629–632).—In each case, as with actinium, two positive, active substances differing in mobility were found; probably one is singly, and the other doubly charged (cf. A., 1925, ii, 79).

A. A. ELDRIDGE.

Chemical action of gaseous ions produced by a-particles. VI. Reactions of oxides of carbon. S. C. LIND and D. C. BARDWELL (J. Amer. Chem. Soc., 1925, 47, 2675-2697; cf. A., 1919, ii, 210; 1924, ii, 11, 12, 840).-In many reactions produced by a-particles from radon, the ratio M/N of molecules decomposed to a-particles is a few units greater than unity. This is ascribed to the formation of clusters of a primary ion and one or more neutral molecules. When a gas like oxygen is present, which has a great affinity for free electrons, these are trapped to form O2, which can then react and may also form clusters with neutral molecules, thus further

increasing the ratio M/N. When this occurs, oxidation replaces all reactions that would take place in its absence, and it is assumed that final chemical action does not result until electrical neutralisation is brought about by recombination of the positive and negative ions or clusters. The chemical reactions taking place in a number of gases mixed with radon have been followed by the pressure changes and by chemical analysis of the final gaseous products; in some cases, pressure measurements have also been made at sufficiently low temperatures to freeze out one or more components. In the decomposition of carbon monoxide, the ratio - M/N is about 2, but falls as the reaction progresses. The products are carbon dioxide, carbon, and a substance which is deposited on the wall of the vessel and is probably a suboxide of carbon; the ratio of carbon monoxide lost to carbon dioxide formed is 3: 1. In the presence of radon, carbon monoxide is oxidised by oxygen and the ratio - M/N is 6; it is concluded that clusters CO-O, CO with single positive or negative charges are formed from CO+, O2+, or O2, but not from CO-, and that the final reaction is (CO-OCO)++ (CO.O, CO)=4CO2. The rate is practically unaffected by the carbon dioxide formed. At the temperature of liquid air, the velocity was half that at 25°, proving that moisture is unnecessary for the ionic reaction. Under a-radiation, carbon monoxide and hydrogen combine to form a white solid, which may be a polymeride of formaldehyde, and the ratio -M/N is 3. Carbon dioxide is not decomposed, but it is probable that ionisation and complex formation take place, the complexes re-forming carbon dioxide when neutralised. Hydrogen and carbon dioxide react, forming water, a very little carbon monoxide, probably carbohydrates, but no methane; the ratio - M/N is 1.7. In the various mixtures of carbon monoxide, carbon dioxide, hydrogen, and oxygen, the ions of both reactants are active.

Revision of the atomic weight of germanium. II. Analysis of germanium tetrabromide. G. P. BAXTER and W. C. COOPER (J. Physical Chem., 1925, 29, 1364-1378).-The material from the analysis of the tetrachloride (A., 1924, ii, 690) provided the germanium used. Germanium tetrabromide, prepared by heating the metal in a current of pure nitrogen charged with bromine, was purified by thirteen fractional distillations in a vacuum. Neglecting the first four fractionations, the most volatile and least volatile portions were then analysed separately. The tetrabromide used was free from bromine, hydrogen bromide, arsenic, and silicon tetrabromide. Thirty-two analyses, in all, gave the ratios GeBr1: 4Ag and GeBr1: 4AgBr, from which the atomic weight of germanium is found to be 72.60 (Ag 107-880, Br 79-916), a value identical with that obtained from the analysis of the tetrachloride (loc. cit.). This agrees with Aston's results (Phil. Mag., 1924, [vi], 47, 394) on the isotopes of germanium. L. S. THEOBALD.

Energy liberated by radium. R. W. LAWSON (Nature, 1925, 116, 897-898).-The total heat development of 1 g. of radium, free from its disintegration products, is calculated to be 23-28 or 25-47 cal. per hour, according as the number of atoms of radium disintegrating per second per 1 g. of the element is taken to be 3.40 or 3.72 × 1010, respectively. Hess' experimental value is 25.2 cal./hour, but when cognisance is taken of the amount of y-radiation unabsorbed under the experimental conditions employed, it is computed that the corrected value cannot greatly differ from 25.5 cal./hour. Support is thus given to the essential correctness of the value Z= 3.72 × 1010, in harmony with Kovarik's work. The data employed in the calculations are quoted. A. A. ELDRIDGE.

deposit of long life from radium. (MLLE.) I. Extraction and purification of the active CURIE (J. Chim. physique, 1925, 22, 471-487).—The method of recovery, from tubes which have contained radon, of the active deposit containing radium-D, radium-E, and polonium is described, together with their separation, and in particular the isolation of the polonium. The methods employed for measuring their activities are also described.

A. E. MITCHELL.

Theory of the range of a-particles. L. LOEB and E. CONDON (J. Franklin Inst., 1925, 200, 595607).-According to previous calculations of the ranges of a-particles by applying the concept of quantised energy exchanges to the classical dynamical treatment, assuming initially free electrons at rest (cf. Bates, A., 1924, ii, 813), the calculated stopping power of the inert gases for a-particles is from 0.6 to 0.8 of the observed stopping power. An error in these calculations has now been indicated, but on correcting for this the discrepancy becomes still greater. Thus the energy of the a-particle is spent in processes other than ionisation and radiation, or else the energy is spent more freely in ionisation than due to the assumption that the electron is initially the theory employed permits. The error may be at rest and responds to the a-particle as if absolutely free, whereas the velocity of the electron in its orbit is probably of the same order as that of the a-particle. M. S. BURR.

Absorption of B-rays by matter. (MME.) J.-S. LATTÈS and G. FOURNIER (Compt. rend., 1925, 181, 855-856). The mass coefficient of absorption, x, of bands of secondary B-rays is given by the relation

=a+bN, where N is the atomic number of the absorbent, a=5.73 and b=0.0547. The ratio, ab, is the same as for the B-rays of radium-E, which also obey the linear law. For an explicit emission obey the linear law. potential, V, the relation x=(a+bN). f(V) is proposed; the nature of f(V) is not given.

S. K. TWEEDY.

Number of particles in the B-ray spectra of radium-B and radium-C. R. W. GURNEY (Proc. Roy. Soc., 1925, A, 109, 540-561).-The origin of the "continuous B-ray spectrum of a radioactive material is still open to discussion. It has been uncertain whether it consists of the electrons ejected from the nucleus during disintegration or whether it

is merely a secondary effect. This problem has now been attacked by measurements both of velocitydistribution in the spectrum and of the total number of particles which constitute the spectrum. The spectrum was resolved by the usual magnetic method, and the particles were focussed on a slit opening into a Faraday cylinder, by which the charge associated with any part of the spectrum could be measured. It is concluded that both radium-B and radium-C possess a genuine continuous B-ray spectrum, quite distinct from the line spectrum which is superimposed on it. The continuous spectrum contains rather more than one electron for each atom disintegrating, which is to be expected if it is formed of the electrons ejected from the nucleus. The continuous spectrum of radium-B rises to a maximum at 170,000 volts, and has an upper limit at 650,000 volts. The spectrum of radium-C rises to a maximum at about 400,000 volts, and has a definite upper limit at 3,150,000 volts. The line spectrum is attributed to the conversion of y- into B-rays, the probability of which is about 1 in 7. The heating effect of the B-rays of radium-B+C is 5-6 cal. per hour per g. Of this, 1-3 cal. are ascribed to radium-B, and 4.3 cal. to radium-C. S. BARRATT.

B-Ray spectrum of the natural L-radiation from radium-B. D. H. BLACK (Proc. Camb. Phil. Soc., 1925, 22, 832-833).-Fifteen lines of a comparatively low-energy B-ray spectrum with energies extending

from 4290 to 12670 volts have been obtained from a source of radium-B in equilibrium with radium-C. Under the same experimental conditions, radium-C gives no such lines, which are therefore attributed to radium-B. It is suggested that the electrons are ejected by L X-rays, from atoms of atomic number 83, acting on the M, N, and O levels of the same atom. A. E. MITCHELL.

Natural X-ray spectrum of radium-B. (SIR) E. RUTHERFORD and W. A. WOOSTER (Proc. Camb. Phil. Soc., 1925, 22, 834-837; cf. preceding abstract). By the original method of Rutherford and Andrade (A., 1914, ii, 408, 698), with a rock salt crystal and a radon tube as source, the separation of the strongest lines a1 and B1 of the L-spectrum of radium-B has been measured. The results are in accord with the separation deduced with the aid of Coster's data for an element of atomic number 83. The results were confirmed by similar measurements using a calcite crystal. It is certain therefore that the earlier measurements of Rutherford and Andrade were in error and that the L-spectrum has its origin in an atom not of atomic number 82 but of atomic number 83, or that in this type of disintegration the emission of the y-ray follows instead of precedes the escape of the electron. A. E. MITCHELL.

Analysis of the B-ray spectrum due to the natural L-radiation of radium-B. D. H. BLACK (Proc. Camb. Phil. Soc., 1925, 22, 838-843; cf. preceding abstracts).-By assuming that the B-ray spectrum from radium-B has its origin in the action of rays corresponding with L X-rays, generated within the atom, on various other atomic levels, it is shown that eleven of the fifteen lines obtained correspond

with an atom of number 83, whilst with an atom of number 82 it is possible to account for only six cf them. The results support the hypothesis that electrons are ejected by y-rays after the nuclear change in the atom has taken place. Similar measurements with thorium-B show that its B-ray spectrum is practically identical with that of its isotope radium-B, and it is suggested that in this case also the y-rays are emitted subsequently to the disintegration. A. E. MITCHELL.

Atomic number of a radioactive element at the moment of emission of the y-rays. C. D. ELLIS and W. A. WOOSTER (Proc. Camb. Phil. Soc., 1925, 22, 844-848; cf. preceding abstracts).—The energies of the three main y-rays from radium-B and the strongest of radium-C-have been measured to 1%. The results are in agreement with those calculated when the emission follows the nuclear disintegration and confirm those of Black and of Rutherford and Wooster. A. E. MITCHELL.

B-Ray type of disintegration. C. D. ELLIS and W. A. WOOSTER (Proc. Camb. Phil. Soc., 1925, 22, 849-860; cf. preceding abstracts).-In view of the results of Black, of Rutherford and Wooster, and of Ellis and Wooster, it is considered established that in the ẞ-ray type of disintegration the emission of the from the nucleus. The mechanism of y-ray and y-ray takes place after the ejection of the electron disintegration electron emissions is discussed. The view of Meitner that the y-rays are emitted during the reorganisation of the nucleus after the disturbance caused by the disintegration electron has been extended to the idea that the y-rays are emitted by the electronic system of the nucleus. The similarity of electronic systems in successive nuclei of the same radioactive family is stressed. The results are interpreted in terms of level systems. Meitner's proposals for the explanation of the inhomogeneity of the velocities of disintegration electrons are discussed, and it is concluded that the electron is actually emitted from the nucleus with a varying velocity, but no explanation of this is advanced.

A. E. MITCHELL.

Nuclear structure of radioactive atoms and the emission of y-ray spectra. J. THIBAUD (Compt. rend., 1925, 181, 857-859).-On the assumption that the nucleus of a radioactive atom consists of a condensed kernel, surrounded by positive corpuscles revolving in quantised circular orbits, it is shown that the observed y-ray spectra may be explained by the passage of a positive corpuscle from one orbit to another of higher energy level causing the emission of one quantum of y-radiation. In certain cases, the existence of fine structure within the nucleus must be supposed. S. K. TWEEDY.

Effect of sunlight on the radioactivity of lead and uranium. (MLLE.) S. MARACINEANU (Compt. rend., 1925, 181, 774-776; cf. A., 1925, ii, 348).— When exposed to sunlight, ordinary inactive lead acquires radioactive properties, which may persist for some months after the insolation. The activated lead affects a photographic plate, causes scintillations on a screen of zinc sulphide, and has slight

ionising properties. After exposure to sunlight, the activity of uranium oxide shows marked variations which may amount to as much as 50% of the total activity. If the uranium oxide is exposed simultaneously to sunlight and to bombardment by the a-particles from polonium, the variations after exposure are more frequent than for sunlight alone. W. HUME-ROTHERY.

Possible significance of tetrahedra numbers in the natural system for atomic mass and atomic structure. H. STINTZING (Z. Physik, 1925, 34, 686-697).-All the atomic masses, including isotopes, can be represented by a scheme of tetrahedra. One element passes into the next higher by the addition of two hydrogen nuclei about a new axis of symmetry. Each fourth element requires a special structure, the carbon type; each eighth element reverts to the noble gas type. When iron is reached, a type of closer packing is introduced, and in the sixth period, rare earths and platinum group, the packing is still closer. The increase in the number of isotopes with increasing atomic weight is represented by the increased possibility of adding protons centrally to the tetrahedra; more are predicted than have yet been found. E. B. LUDLAM.

Magnetism and the structure of atoms and molecules. B. CABRERA (J. Phys. Radium, 1925, [vi], 6, 273-286; cf. A., 1925, ii, 1107).-The variation of paramagnetism with atomic number provides a means of fixing the position in the atom of the last electron acquired by the atom or cation concerned. For the iron and rare-earth groups, Stoner's scheme of atomic levels is in accord with experiment. An explanation is given of the differences in magnetic behaviour exhibited by the tervalent ions of chromium and cobalt and also of differences within the group of rare earths. The constant A of the Curie-Weiss equation, Xm(T+A)=Cm, is still obscure from the point of view of physical significance. Its importance is, however, emphasised, and the suggestion is made that it depends on the interaction of paramagnetic atoms with neighbouring atoms.

R. A. MORTON.

Physical structure of the elements. C. G. BEDREAG (Bull. Acad. Sci. Roumaine, 1925, 9, [9-10], 8-14; cf. A., 1925, ii, 363).-The alteration in valency and in spectral multiplicity (multiplets) with atomic number is discussed. J. S. CARTER.

Nuclear numbers. D. DE BARROS (Compt. rend., 1925, 181, 719-722). By insertion of the nuclear numbers, 4 [the nearest integer to the value (M-N), where M is the atomic weight and N the atomic number], in a rearranged periodic table, in which the elements are classed in sixteen columns of five rows each, a table is obtained almost exactly the same as that found experimentally by Mayer for the arrangement of magnetised needles round a magnetic pole. For each element in any one column, the relation M-2N-A1 holds, where x indicates the row in which the element occurs. Some doubtful atomic weights, and also those of undiscovered elements, are calculated by this relationship.

S. K. TWEEDY.

Doublet and triplet separations in optical spectra as evidence whether orbits penetrate into the core. D. R. HARTREE (Proc. Camb. Phil. Soc., 1925, 22, 904-918).-From an analysis of doublets and triplets in optical spectra by means of the formula of Landé (A., 1924, ii, 711), it is concluded that, except for lithium-like atoms, the p-terms of all known spectra correspond with orbits which penetrate into the core. This is not in agreement with Bohr's assignment of quantum numbers for the p-terms of sub-groups. The evidence for d-terms is not so the spectra of neutral atoms of the copper and zinc complete, but suggests that at least in Cs I and TII ment with Bohr's quantum number assignment. these correspond with penetrating orbits not in agreeWhen the terms belong to a Ryberg sequence the in multiplet systems. Landé formula holds approximately for separations A. E. MITCHELL.

Optical constants. I. Optical behaviour of certain atomic models. II. Lateral scattering from a gas. C. G. DARWIN (Proc. Camb. Phil. Soc., 1925, 22, 817-823, 824-831).-Mathematical.

Structure of manganese. H. COLLINS (Chem. News, 1925, 131, 355-358).-Speculative.

Striated discharge in hydrogen. D. A. KEYS (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 143148). Single striations suddenly become double when the P.D. applied to the tube is raised to a certain critical value, the magnitude of which depends upon the pressure of the gas. Under certain conditions in a new discharge tube, the striations may be made to revolve rapidly by the application of a magnetic field at a certain region of the discharge and in a particular direction. The spectra of the light from the negative glow and from a striation show that there are more excited atoms than molecules in the negative glow and that the reverse is true of the striations. The spectra from the two sides of a single striation are not the same nor is the distribution of intensity in the spectral lines from different parts of a single striation that of the light from the corresponding part of a double striation. J. S. CARTER.

Secondary spectrum of hydrogen. O. S. DUFFENDACK (Physical Rev., 1924, [ii], 23, 107; cf. A., 1925, ii, 333). With a low-voltage arc in a mixture of hydrogen and mercury vapour the Fulcher bands and other strong lines, particularly groups at 46254634 and 4562-4580 A., appear at 13 volts. Ha appears at 11 volts and Hg at 12 volts. Hydrogen excited in a three-electrode tube gave a complete spectrum at 15 volts, but only faint traces of Balmer lines below 16.5 volts. Presumably they arise in the mixture by dissociation of hydrogen by impacts with excited mercury atoms. A. A. ELDRIDGE.

Characteristics and spectra of low-voltage arcs in hydrogen, nitrogen, and mixtures of hydrogen with mercury and nitrogen. C. T. KWEI (Physical Rev., 1925, [ii], 26, 537-560).-In hydrogen, the arc broke at 16-23+0.03 volts, in agreement with Duffendack's value for the ionisation potential of the hydrogen molecule; that of the hydrogen atom is 13.67+0.14 volts. In nitrogen,

the arc broke at 16.18+0.05 volts, but the lowest voltage at which the arc could be maintained without oscillations is 16.90±0.07 volts, which is regarded as the ionisation potential of the nitrogen molecule. With mercury in a tube appended to the discharge tube, two arcing and two breaking potentials could be obtained, the difference between the two arcing potentials (about 10-4 volts) corresponding with the ionisation potential of mercury; under other conditions, a single breaking potential was observed. Similar results were obtained with a mixture of hydrogen and nitrogen; there appears to be a critical potential of nitrogen at 22.7 volts. In hydrogen, the Balmer lines and the secondary spectrum always appeared together, the former being more intense; in mixtures of mercury vapour and hydrogen, the mercury lines and the Balmer series lines were much enhanced relative to the secondary spectrum after passing the second arcing potential. The results indicate that between the first and second breaks hydrogen molecules were dissociated into atoms. The ultra-violet band of ammonia, associated with the 22.5 volt critical potential, was observed with mixtures of hydrogen and nitrogen in any proportion, but not with either gas alone. It is considered to be due to a molecule NH,, where x is probably 3. The Schuster bands are ascribed to a molecule containing nitrogen, hydrogen, and oxygen, possibly NH,OH, which is extremely unstable, and is formed only at voltages above 70, the critical voltage for active nitrogen.

A. A. ELDRIDGE.

Band spectrum of nitrogen and its theoretical interpretation. R. T. BIRGE (Physical Rev., 1924, [ii], 23, 294-295).-It is concluded that a common final quantum state of the second and fourth band groups is also the initial state of the first group. Only one possible assignment of true vibration quantum numbers is indicated, leading to a possible explanation of various changes in physical appearance. Active nitrogen is apparently a metastable, energised, neutral N2 molecule, chiefly with vibration quantum number 11. The quantum structure of triplet bands is established; the moment of inertia of the N2 neutral molecule in the non-vibratory initial state of the second group is 17-2 × 10-40. CN and N, bands are not identical. A. A. ELDRIDGE.

Excitation potentials of the band spectrum of nitrogen. H. SPONER (Z. Physik, 1925, 34, 622— 633). The spectrum was observed when nitrogen was excited by electron impact in a tube which was first calibrated by using a neon-helium mixture. The excitation potentials for the zero band of the positive and negative groups were measured (1st, 2nd, and 4th positive, 9.3, 13.0, and 14.8 volts, respectively; 1st negative, 19-6 volts; ionisation potential, 16.5 volts). The suggestion is made that active nitrogen consists of atoms which undergo a threefold collision with 1 mol. of nitrogen, as a result of which the two atoms can unite, and the energy of dissociation which they lose is acquired by the molecule with which they collided, and the glow is due to the radiation from this excited molecule. This gives as the energy of dissociation of the nitrogen

ultra-violet. J. C. MCLENNAN and A. B. McLAY Absorption spectra of various elements in the Absorption spectra of gold, silver, copper, iron, (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 89—111). manganese, cobalt, nickel, and chromium have been obtained with a quartz spectrograph and of gold, silver, copper, iron, antimony, bismuth, arsenic, and manganese with the fluorite spectrograph. A large number of absorbed wave-lengths are tabulated. J. S. CARTER.

Flame spectra of carbon monoxide and water gas. II. F. R. WESTON (Proc. Roy. Soc., 1925, A, 109, 523-526; cf. A., 1925, ii, 928).-Previous results have been confirmed by experiments with an improved disposition of the apparatus. On the supposition that the direct burning of carbon monoxide in oxygen gives rise to the continuous and banded flame spectrum of carbon monoxide, whilst indirect burning by reactions with water molecules causes the emission of the "OH" bands only, it is concluded that for an equimolecular mixture of carbon monoxide and hydrogen, the combustion of the former gas is effected almost exclusively by the indirect process. The addition of this proportion of hydrogen almost eliminates the normal flame spectrum of carbon monoxide. S. BARRATT.

One-valency-electron emitters of band spectra. R. S. MULLIKEN (Physical Rev., 1925, [ii], 26, 561572).-Evidence is collected showing that the molecules BeF, BO, CO+, CN, and Na+, together with the molecules MgF, AIO, and SiN, are all ready emitters cules is indicated. of band spectra, and a close similarity of these molecules is indicated. The nature of multiplicity in band spectra is discussed. A. A. ELDRIDGE.

Ultra-violet absorption spectra of butenonitriles and butenoic acids. P. BRUYLANTS and A. CASTILLE (Bull. Soc. chim. Belg., 1925, 34, 261— 284). The absorption spectra of the following acids and their nitriles have been studied: crotonic acid, isocrotonic acid, their ẞ-chloro-derivatives, methylacrylic acid, cyclopropanecarboxylic acid, and Bhydroxybutyric acid. The results obtained show that the crotononitrile, b. p. 108° (Bruylants, A., 1922, i, 817, 924; 1924, i, 1053), possesses the same structure as solid crotonic acid, and that the nitrile, b. p. 121°, is identical in structure with isocrotonic acid. These results are in agreement with those obtained from measurements of the heat of combustion of the two isomerides. The isomerisation of vinylacetonitrile into crotononitrile (cf. Bruylants, loc. cit.; Auwers, A., 1923, i, 661) is incomplete in N/1000and N/10000-sodium hydroxide; even after 15 days, the curve is not identical with that of crotononitrile in alkaline solution. In very dilute solutions of sodium hydroxide (N/1000), only a single variety of crotononitrile is produced. With more concentrated alkaline solutions (N/100- or N/20-sodium hydroxide), the change is more marked, and the spectrum is displaced towards the visible region in comparison with that of isocrotononitrile, b. p. 121°, in N/10