& PROCEEDINGS OF THE ROYAL SOCIETY OF LONDON. From April 23, 1896, to February 18, 1897 LONDON: HARRISON AND SONS, ST. MARTIN'S LANE, in ©rhinarg to |>w Paj MDCOCXCT1I. Q LONDON: HARRISON ANJ> SONS, PEINTEES IN ORDINARY TO HER MAJESTY, ST. MARTIN'S IANE. CONTENTS. -y Q. 442 JD.ioC Some Experiments on Helium. By Morris W. Travers, B.Sc 449 On the Gases enclosed in Crystalline Rocks and Minerals. By W. A. Tilden, D.Sc., F.R.S viii Page \ On Lunar Periodicities in Earthquake Frequency. By C. G. Knott, D.Sc., Lecturer on Applied Mathematics, Edinburgh University (formerly Professor of Physics, Imperial University, Japan) 45' No. 367. Meeting of February 11, 1897, with List of Papers read ...- 466 The Oviposition of Nautilus macromphalus. By Arthur Willey, D.Sc., Balfour Student of the University of Cambridge. Communicated by Alfred Newton, M.A., F.R.S., on behalf of the Managers of the Balfour Fund 467 On the Regeneration of Nerves. By Robert Kennedy, M.A., B.Sc., M.D. (Glasgow). Communicated by Professor McKeiidrick, F.R.S. 472 Meeting of February 18, 1897, with List of Papers read 474 On the Iron Lines present in the Hottest Stars. Preliminary Note. By J. Norman Lockyer, C.B., F.R.S 475 On the Significance of Bravais' Formulae for Regression, &c., in the case of Skew Correlation. By G. Udny Yule. Communicated by Pro- fessor Karl Pearson, F.R.S 477 Mathematical Contributions to the Theory of Evolution — On a Form of Spurious Correlation which may arise when Indices are used in the Measurement of Organs. By Karl Pearson, F.R.S., University College, London 489 Note to the Memoir by Professor Karl Pearson, F.R.S., on Spurious Correlation. By Francis Galton, F.R.S 498 Report to the Committee of the Royal Society appointed to investigate the Structure of a Coral Reef by Boring. By W. J. Sollas, D.Sc., F.R.S., Professor of Geology in the University of Dublin 502 The Influence of a Magnetic Field on Radiation Frequency. Commu- nication from Professor Oliver Lodge, F.R.S 513 The Influence of a Magnetic Field on Radiation Frequency. Commu- nication from Dr. J. Larmor, F.R.S 514 Obituary Notices : — Hermann Kopp i John Rae v Franz Ernst Neumann viii Sir Joseph Prestwich xii Sir George Johnson , xvi Henry Newell Martin , xx Brian Houghton Hodgson xxiii William Crawford Williamson xxvii Sir George Henry Richards, K.C.B xxxii Index xxxvii Erratum '„, xlvii PROCEEDINGS OF THE ROYAL SOCIETY. April 23, 1896. (Meeting for Discussion.) Sir JOSEPH LISTER, Bart., President, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. The following Paper was read for the purpose of opening the discussion : — "On Colour Photography by the Interferential Method." By G. LIPPMANN, Professor of Physics, Faculty of Sciences, Paris. Communicated by Sir JOSEPH LISTER, Bart., P.R.S. April 30, 1896. Sir JOSEPH LISTER, Bart., President, in the Chair. The Right Hon. Sir Richard Temple, Bart., a Member of Her Majesty's Most Honourable Privy Council, was admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " Note on Photographing Sources of Light with Monochromatic Rays." By Captain W. DE W. ABNEY, C.B., D.C.L., F.R.S. VOL. LX. B 2 Proceedings. April 30, 1896 — continued. II. " On the Determination of the Photometric Intensity of the Coronal Light during the Solar Eclipse of 16th April, 1893." By Captain W. DE W. ABNEY, .C.B., D.C.L., F.R.S., and T. E. THORPE, LL.D., F.R.S. III. " The Total Eclipse of the Sun, April 16, 1893. Report and Discussion of the Observations relating to Solar Physics." By J. NORMAN LOCKTER, C.B., F.R.S. IV. " On some Palaeolithic Implements found in Somaliland by Mr. H. W. Seton-Karr." By Sir JOHN EVANS, K.C.B., D.C.L., Treas. and V.P.R.S. Hay 7, 1896, Sir JOSEPH LISTER, Bart., President, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. In pursuance of the Statutes, the names of the Candidates recom- mended for election into the Society were read from the Chair as follows : — Murray, John, Ph.D. Clarke, Lieut.-Col. Sir George Sydenham, R.E. Collie, J. Norman, Ph.D. Downing, Arthur Matthew Weld, D.Sc. Elgar, Francis, LL.D. Gray, Professor Andrew, M.A. Hinde, George Jennings, Ph.D. Miers, Professor Henry Alexander, M.A. Mott, Frederick Walker, M.D. The following Papers were read : — I. " On the Liquation of certain Alloys of Gold." By E. MATTHEY. Communicated by Sir G. G. STOKES, F.R.S. II. " On the Occurrence of the Element Gallium in the Clay- Iron- stone of the Cleveland District of Yorkshire. Preliminary Notice." By Professor HARTLEY, F.R.S. , and H. RAMAGE. Pearson, Professor Karl, M.A. Stebbing, Rev. Thomas Roscoe Rede, M.A. Stewart, Professor Charles, M.R.C.S. Wilson, William E. Woodward, Horace Bolingbroke, F.G.S. Wynne, William Palmer, D.Sc. Proceedings. 3 May 7, 1896— continued. III. " The Electromotive Properties of Malapterurus ekctricus." By Professor GOTCH, F.R.S., and G. J. BURCH. IV. "The Occurrence of Nutritive Fat in the Human Placenta. Preliminary Communication." By Dr. T. W. EDEN. Commu- nicated by Dr. PYE SMITH, F.R.S. The Society adjourned over Ascension Day to Thursday, May 21. May 21, 1896. Sir JOSEPH LISTER, Bart., President, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " On the Changes produced in Magnetised Iron and Steel by cooling to the Temperature of Liquid Air." By Professor J. DEWAR, F.R.S., and Dr. J. A. FLEMING, F.R.S. II. " Note on the Larva and Post-larval Development of Leuco- solenia variabilis, H. sp., with Remarks on the Development of other Asconidse." By E. A. MINCHIN. Communicated by Professor LANKESTER, F.R.S. III. " Helium and Argon. Part III. Experiments which show the Inactivity of these Elements.'' By Professor RAMSAY, F.R.S. , and Dr. J. NORMAN COLLJE. IV. "On the Amount of Argon and Helium contained in the Gas from the Bath Springs." By LORD RAYLEIGH, Sec. R.S. . The Society adjourned over the Whitsuntide Recess to Thursday, June 4. B 2 4 Proceedings. June 4, 1896. The Annual Meeting for the Election of Fellows was held this day. Sir JOSEPH LISTER, Bart., President, in the Chair. The Statutes relating to the election of Fellows having been read, Professor Bonney and Mr. Salvin were, with the consent of the Society, nominated Scrutators to assist the Secretaries in the examin- ation of the balloting lists. The votes of the Fellows present were collected, and the following Candidates were declared duly elected into the Society : — Clarke, Lieut. -Col. Sir George Sydenham, R.E. Collie, J. Norman, Ph.D. Downing, Arthur Matthew Weld, D.Sc. Elgar, Francis, LL.D. Gray, Professor Andrew, M.A. Hinde, George Jennings, Ph.D. Woodwai Miers, Professor Henry Alexander, F.G.S. M.A. Mott, Frederick Walker, M.D. Thanks were given to the Scrutators. Murray, John, Ph.D. Pearson, Professor Karl, M.A. Stebbing, Rev. Thomas Roscoe Rede, M.A. Stewart, Professor Charles, M.R.C.S. Wilson, William E. Woodward, Horace Bolingbroke, Wynne, William Palmer, D.Sc. June 4, 1896. Sir JOSEPH LISTER, Bart., President, in the Chair. Professor Albert Gaudry, who was elected a Foreign Member in 1895, was admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " On the unknown Lines observed in the Spectra of certain Minerals." By J. NORMAN LOCKYEE, C.B., F.R.S. II. " On the Electrical Resistivity of Bismuth at the Temperature of Liquid Air." By Professor J. DEWAE, F.R.S., and Dr. J. A. FLEMING, F.R.S. Proceedings. 5 June 4, 1896 — continued. III. " On the Electrical Resistivity of pure Mercury at the Tempera- ture of Liquid Air." By Professor J. DEWAR, F.R.S., and Dr. J. A. FLEMING, F.R.S. IV. " The Hysteresis of Iron and Steel in a rotating Magnetic Field." By Professor F. G. BAILY. Communicated by Professor LODGE, F.R.S. V. " Observations on Atmospheric Electricity at the Kew Observa- tory." By C. CHREE. Communicated by Professor G. CAREY FOSTER, F.R.S. June II, 1896. Sir JOSEPH LISTER, Bart., President, in the Chair. Dr. J. Norman Collie, Dr. A. M. W. Downing, Professor Andrew Gray, Dr. G. J. Hinde, Dr. F. W. Mott, Rev. T. R. R. Stebbing, Professor C. Stewart, Mr. W. E. Wilson, Mr. H. B. Woodward, and Dr. W. P. Wynne were admitted into the Society. A congratulatory Address to Lord Kelvin, prepared for presentation to him on the occasion of the jubilee of his professoriate in the University of Glasgow, was read from the Chair and unanimously adopted. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " The Relation between the Refraction of the Elements and their Chemical Equivalents." By Dr. J. H. GLADSTONE, F.R.S. II. "On the Magnetic Permeability and Hysteresis of Iron at Low Temperatures." By Dr. J. A. FLEMING, F.R.S., and Professor J. DEWAE, F.R.S. III. " On certain Changes observed in the Dimensions of Parts of the Carapace of Carcinus mcenas." By H. THOMPSON. Communi- cated by Professor WELDON, F.R.S. IV. " On the Relation between the Viscosity (Internal Friction) of Liquids and their Chemical Nature." By Dr. T. E. THORPE, F.R.S., and J. W. RODGER. 6 Proceedings. June 18, 1896. Sir JOSEPH LISTER, Barfc., President, in the Chair. Lieut.-Colonel Sir Or. S. Clarke and Professor H. A. Miers were admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. Sir J. "W. Dawson exhibited new specimens of Carboniferous Batrachians. An oral communication was made by Professor J. A. Fleming, F.R.S., on behalf of Professor Dewar and himself, to the following effect :— In continuing our experiments on the electrical resistance of bis- muth at low temperatures and in magnetic fields, by the aid of a powerful electro-magnet, kindly lent to us by Sir David Salomons, we have observed the fact that a wire of electrolytic bismuth, when cooled in liquid air to a temperature of —186° C., has its resistance increased more than fortv-two times if it is at the same time trans- versely magnetised in a field of 14,000 units. The bismuth, when cooled in liquid air and thus magnetised, has its electrical resistance increased more than fifteen times, even when compared with its resistance at ordinary temperatures and not in a magnetic field. There is no reason to believe we have reached the limits of this increase. We reserve further details for a full communication to the Royal Society later. The following Papers were read : — I. " Etude des Carbures Metalliques." By M. HENRI MOISSAN. Communicated by Professor RAMSAY, F.R.S. II. " On Fertilisation and the Segmentation, 'of the Spore in Fucns." By J. B. FARMER and J. L. WILLIAMS. Communi- cated by Dr. D. H. SCOTT, F.R.S. III. " Complete Freezing-point Curves of Binary Alloys containing Silver or Copper together with another Metal." By C. T. HEYCOCK, F.R.S., and F. H. NEVILLE. IY. "Note on the Radius of Curvature of a Cutting Edge." By A. MALLOCK. Communicated by LORD KELVIN, F.R.S, Y. "A Magnetic Detector of Electrical Waves and some of its Applications." By E. RUTHERFORD. Communicated by Pro- fessor J. J. THOMSON, F.R.S. Angular Measurement of Optic Axial Emergences. 7 June 18, 1896— continued. VI. " Experimental Proof of van't Hoff's Constant, Dalton's Law. &c., in very dilute Solutions." By Dr. MEYER WILDERMAXN, Communicated by Professor FITZGERALD, F.R.S. VII. " On the Determination of the Wave-length of Electric Radia- tion by Diffraction Gratings." By J. C. BOSE. Communicated by LORD RATLEIGH, Sec. R.S. VIII. " The Effects of a strong Magnetic Field upon Electric Dis- charges in Vacuo." By A. A. C. SWINTON. Communicated by LORD KELVIN, F.R.S. IX. " On the Structure of Metals, its Origin and Changes." By M. F. OSMOND and Professor ROBERTS- AUSTEN, C.B., F.R.S. X. "Magnetisation of Liquids." By JOHN S. TOWNSEND. Com- municated by Professor J. J. THOMSON, F.R.S. XI. " Selective Absorption of Rontgen Rays." By J. A. MCCLELLAND. Communicated by Professor J. J. THOMSON F.R.S. XII. "On the Determination of Freezing Points." By J. A. BARKER, D.Sc. Communicated by Professor SCHUSTER, F.R.S. XIII. " The Menstruation and Ovulation of Macacus rhesus ; with Observations on the Changes undergone by the discharged Follicle. Part II." By WALTER HEAPE. Communicated by Dr. M. FOSTER, Sec. R.S. XIV. " Phenomena resulting from Interruption of Afferent and Efferent Tracts of the Cerebellum." By Dr. J. S. RISIEN RUSSELL. Communicated by Professor V. HORSLEY, F.R.S. The Society adjourned over the Long Vacation to Thursday, November 19. "Angular Measurement of Optic Axial Emergences." By WILLIAM JACKSON POPE. Communicated by Professor ARMSTRONG, F.R.S. Received February 7, — Read March 19, 1896. Crystals belonging to the monoclinic or anorthic systems are rarely obtained in which the optical orientation is such that a large crystal face is so nearly perpendicular to a bisectrix that the apparent optic axial angle as observed in air can be directly measured by means of the ordinary Fuess apparatus. It thus becomes 8 Mr. W. J. Pope. necessary to first grind plates of known orientation for optical examination ; this latter operation is by no means easily performed, especially in the case of brittle organic substances. Very usually, however, crystals belonging to the biaxial systems are obtained in which an optic axis apparently emerges into air through a particular face ; in these cases the accurate measurement of the angle between the apparent direction in air of the optic axis and the normal to the crystal plate becomes an important element in the determination of the optical constants of the crystal. The ordinary method of determining this angle is a direct one ; the crystal is adjusted in the optic axial angle apparatus and a read- ing is taken for the above emergence, after the position of the normal to the plate has been found by reflecting a beam of light down the telescope tube and turning the crystal until the shadow and reflected image of the crosswires coincide ; the angular differ- ence between the two readings is then the required apparent angle of emergence into air. This method of finding the position of the normal is, however, very tedious, and, unless the crystal plate pos- sesses a highly polished surface, very inaccurate. To remedy these defects a method has been devised of indirectly determining this angle by calculating it from the angle through which the optic axis is apparently refracted by an oil of high refrac- tive index. The crystal is mounted and adjusted in the optic axial angle apparatus in the ordinary way, and a reading is taken for the optic axial emergence in air; a parallel-sided glass cell containing a-bromonaphthalene or some other highly refractive liquid is then raised until it surrounds the crystal, and a second reading is taken of the apparent emergence of the optic axis. From the difference between these two angular readings the angle of emergence into air can be calculated, if the index of refraction of the oil is known. N Angular Measurement of Optic Axial Emergences. 9 In the figure, OA is an optic axial direction in the crystal, OB "is the direction of optic axial emergence into air, and OC is the direc- tion of emergence into a liquid of refractive index ^u; ON is the normal to the crystal plate. Then «, the angle of emergence into air, is NOB, whilst 0, the angle of emergence into the liquid is NOG and sin*/sin0 = ft-, it is required to calculate the angle «, from the observed value of a — 6. Then, since sin a/sin 9 = ^ - = sin{g~ Q— 0)} /* sin a __ sin ac cos (y. — 0) — cos a sin (a — 0) sin a = cos (a— 0) — cot a sin (a— 0) cota = cotfa— 0) -- J — 0) Or again, since sin a/sin 0 = a, sin a 4- sin 0 ;t — 1 sin a— sin 0 _ sin ^0 + 0) cos |(a— 0) , a — 0 whence tan - = - -- tan - ........... . (2) 2 yu— 1 2 a form more convenient than (1) for logarithmic calculation. To test the accuracy of the method, measurements have been made 011 biaxial plates of different optical properties, liquids of various re- fractive indices being used. The index of refraction of the liquid employed is conveniently determined with the Ptilfrich refractometer ; the refraction is so affected by differences of temperature and of purity that it is necessary to determine it for the liquid as actually used ; the liquid does not need to be specially purified. The measure- ments given in the two appended tables were made on plates of topaz, each of them cut perpendicularly to the acute bisectrix. By measurement of the optic axial angles, the apparent emergences into air for sodium light were found to be 53° 24' and 54° 42', respec- tively. These two sets of measurements suffice to show that the method possesses very considerable accuracy, although the values of a— 0 measured are not very large ; the numbers also seem to indicate 10 Prof. G. Lippmann. On Colour TABLE I. — Plate with angle of emergence in air 53° 24'. Liquid. ^D. a-e. a. A. 1 -6473 24° 15' 53° 26' + 2' a-Bromonaphthalene. . . . Benzene . ... 1 -5341 1 -4970 21 50 21 0 53 22J 53 27£ -H + 3£ Turpentine 1 '4726 20 20 53 20| — 3^ 1 -4673 20 12 53 21 — 3 1 -4634 20 10 53 28| + 4£ Chloroform 1 -4439 19 35 53 19| — 4i .Alcohol . . . . 1-3561 17 2 53 15 — 9 \^ater 1 -3327 16 21 53 22 — 2 TABLE II. — Plate with angle of emergence in air 54° 42'. Liquid. /*D. a-9. a. A. Carbon bisulphide 1 -6473 24° 58' 54° 38^' — 3i' a-Broinonaphthalene. . . . 1 -5341 1 -4970 22 35 21 41 54 44i 54 44£ + 2! + 2^ 1-4726 21 0 54 37 -5 Olive oil . 1 '4673 20 56 54 45 + 3 1 -4634 20 51 54 47£ + 5i Chloroform . 1-4439 20 17 54 42 0 -A.lcoh.ol . . . < 1 • 3561 17 45 54 48£ + 6£ Watey 1 -3327 16 54 54 37 — 5 that, as would of course be expected, the most accurate results are obtained with liquids of high refractive index, which give com- paratively large values of a — 9. By determining the values of a — 0 for each of two optic axes of a given crystal plate, it can easily be ascertained with what amount of accuracy the plate has been cut perpendicularly to the bisectrix. The principle of the method here described may very possibly be advantageously employed in other branches of optical investigation. " On Colour Photography by the Interferential Method." By G. LIPPMANN, Professor of Physics, Faculty of Sciences, Paris. Communicated by Sir JOSEPH LISTER, Bart., P.R.S. Received April 14,— Read April 23, 1896. Colour photographs of the spectrum, or of any other object, are obtained by the following method. A transparent photographic film of any kind has to be placed in contact with a metallic mirror during Photography by the Interferential Method. 11 exposure. It is then developed and fixed by the usual means em- ployed in photography, the result being a fixed colour photograph visible by reflected light. The mirror is easily formed by means of mercury. The glass plate carrying the film being inclosed in a camera slide, a quantum of mercury is allowed to flow in from a small reservoir and fill the back part of the slide, which is made mercury-tight. The plate is turned with its glass side towards the objective, the sensitised film touching the layer of mercury. After exposure, the mercury is allowed to flow back into its reservoir, and the plate taken out for development. The only two conditions necessary for obtaining colour, trans- parency of the film and the presence of a mirror during exposure, are physical conditions. The chemical nature of the photographic layer has only secondary importance ; any substance capable of giving, by means of an appropriate development, a fixed colourless photo- graph, is found to give, when backed by the mirror, a fixed colour photograph. We may take, for instance, as a sensitive film, a layer of albumeno- iodide of silver, with an acid developer; or a layer of gelatino- bromide of silver, with pyrogallic acid, or with amidol, as deve- lopers. Cyanide or bromide of potassium may be as usual employed for fixing the image. In a word, the technics of ordinary photo- graphy remain unchanged. Even the secondary processes of intensi- fication and of isochromatisation are employed with full success for colour photography. The photographic films commonly in use are found to be opaque, and formed, in fact, by grains of light-sensitive matter mechanically imprisoned by a substratum of gelatine, albumen, and collodion. What is here wanted is a fully transparent film, the light-sensitive matter pervading the whole of the neutral substratum. How can such a transparent film be realised? This question remained insoluble to me for ma,ny years, so that I was debarred trying the above method when I first thought of it. The difficulty, how- ever, is simply solved by the following remark. It is well known that the precipitation of a .metallic compound, such as bromide of silver, does not take place in the presence of an organic colloid, such as albumen, gelatine, or collodion. In reality, the metallic compound is formed, but remains invisible; it is retained in a transparent modification by the organic substances. We have only, therefore, to prepare the films in the usual way, but with a stronger proportion of the organic substratum ; the result is a transparent film. By mixing, for instance, a gelatinous solution of nitrate of silver with a gelatinous solution of bromide of potassium, no precipitate is formed, and the result is a transparent film of dry gelatine containing 15 and even 30 per cent, of the weight of bromide of silver. 12 On Colour Photography by the Interferential Method. The colours reflected by the film are due to interference : they are of the same kind as those reflected by soap bubbles or by Newton's rings. When a ray of definite wave-length falls on the sensitive plate, it is during exposure reflected back by the mirror, and then gives rise to a set of standing waves in the interior of the film, the distance between two successive loops being equal to half the wave-length of the luminous ray. This system of standing waves impresses its periodical structure on the film. The photographic deposit, therefore, takes the form of a grating, a continuous grating, perfectly adapted for reflecting the particular luminous ray which has given it birth. This theory can be subjected to experimental proof. If we ex- amine a photograph of the spectrum, or any other object by white light, we observe the following facts. (1.) Colours are seen in the direction of specular reflection, and are invisible in every other direction. (2.) The colours change with the incidence; the red changing successively to green, blue, and violet, when the incidence grows more oblique. The whole image of the spectrum is dis- placed, and gradually passes into the infra-red region. (3.) If the film be gradually moistened, the colour changes in the opposite direction, from violet to red. This phenomenon is due to the swelling up of the gelatine or albumen, causing the intervals between the elements of the grating to become larger. The smaller intervals, corre- sponding to violet and blue light, gradually swell up to the values proper to red and infra-red waves. A photograph immersed in water loses all its colours, these appearing again during the process of drying. For the same reason, a freshly prepared plate has to be dried before the correct colours can be finally seen. We have now to consider the case of compound colours, and to generalise the former theory, which is only applicable to the action of simple rays. I beg to subjoin an abstract of this generalised theory. It will be seen that if a compound ray of definite composi- tion impresses the plate, it gives rise during exposure to a definite set of standing waves, which impress their structure on the film, and impart to the photographic deposit a corresponding definite form. Though very complex, this can be described as made up of a number of elementary gratings, each corresponding to one of the simple rays which contribute the impressing light. When examined by white light, the reflected ray is shown to have the same composition as the impressing ray ; white light, for instance, imparts to the photographic deposit such a structure that it is adapted to reflect white light. The only a priori condition for the correct rendering of compound rays, is a correct isochromatisation of the film. This, again, can be practically effected by known processes, such as have been indicated by E. Becquerel, Vogel, Captain Abney, and others. Photographing witli Monochromatic Rays. 13 As a verification of this theory, I beg leave to project on the screen a series of colour photographs, representing natural objects : pictures on stained glass, landscapes from nature, flowers, and a portrait from life. Every colour in nature, including white, and the delicate hue of the human complexion, is thus shown to be reflected by a correctly developed photographic film. It is to be remarked that, as in the case of the spectrum, the colours are visible only in the direction of specular reflection. If I had tried to touch up these photographs by means of water colours or other pigments, these would be made apparent by slightly turning the photograph ; these pigments remaining visible under every in- cidence, they would thus be seen to stand out on a colourless back- ground. Thus the touching up or falsifying by hand of a colour photograph is happily made impossible. "Note on Photographing Sources of Light with Mono- chromatic Rays." By Captain W. DE W. ABNEY, C.B., D.C.L., F.R.S. Received March 31,— Read April 30, 1896. In a paper " On the Production of Monochromatic Light," com- municated to the Physical Society, and read on the 27th June, 1885, and which appears in the ' Philosophical Magazine ' for August in that same year, I stated that by the apparatus then described a monochromatic image of the sun could be thrown upon the screen. In the same periodical for June of the same year, Lord Rayleigh described a plan for obtaining a monochromatic image of an external object, in which a concave lens was placed behind the slit of a spectro- scope to produce an image of the object in monochromatic colour, the object being viewed through an aperture placed in the spectrum produced by the apparatus. I had been working independently at the subject at the same time, and my object was to get an image on a screen or photographic plate rather than to use the apparatus for visual observation. When a lens is placed behind the spectrum in the manner described in the paper above referred to, a white image of the prism can be obtained on a screen placed at some distance from the lens, and the size of the image can be increased or diminished according to the focal length of the lens, and its distance from the spectrum. Evidently, then, if an image of a luminous object can be cast on the surface of the prism, and a slit be placed in the spectrum, the image of the luminous object will be seen of the colour of the light passing through the slit. There are devices adopted at the present time for photographing the sun with light of various wave-lengths, but, as far as I am aware, they depend upon moving the image of the sun across the slit of the spectroscope, the 14 Photographing with Monochromatic Rays. plate moving across the slit in the spectrum at the requisite rate for the various impressions made by the different parts of the sun's image to coalesce. It had struck me some time since that the method thus indicated nearly eleven years ago might be more convenient than that adopted, but the time I had at my disposal prevented my carry- ing out a continuation of my experiments. Recently I have had occasion to take up this subject for a rather different purpose, and as the method seems to have been untried, I give it in more detail than I did then. My investigation called for a determination of the proportions of various rays emitted by the various parts of the carbon of the positive and negative poles of an electric arc light, and for this purpose the system of forming monochromatic images was found to be useful. The points of the electric light EL (fig. 1) were placed so that a beam of light passed through the slit S of the collimator on to the centre of the collimating lens L2. A convex lens L! of shorter focus than L2 was placed in the path of the rays, and so adjusted that a real image of the poles was formed on L2. These passed through the lens La as nearly parallel rays and struck upon the prism, and then passed through the remainder of the apparatus as sketched in fig. 2, where M is the prism, L3 a lens to bring the rays to a focus as a spectrum on ub after passing through a camera, A. L4 is a lens, shown in the figure connected with a camera, B, which brings the image of the prism arid the bright image cast on it to a focus at P. By placing a slit S2 in the spectrum, the image cast on P will be as monochromatic as the light coming through the slit. L: should be of such a focal length that it should be as near the slit as possible. With this" arrangement it is very curious to watch the variations in the brightness of the arc and of the flame which accompanies the movement of the slit through the spectrum, and as each variation can be photographed on a Cadett polychromatic photographic plate, we can obtain records of all that is Determination of Coronal Light during Eclipse. 15 occurring. Further, by using strips of lenses cut out at suitable distances from the axes (fig. 3), images of various colours can be placed side by side upon P, since a slit may be placed in the spectrum opposite each such strip of lens. Incidentally, I may men- tion that investigations into the cause of the variable nature of different flames can be carried out by this plan. For solar work, a long collimator appears to be a necessity, but the aperture need not be large. Suppose we determine to have an imacre of the sun on P (fig. 2) of 2 in. diameter, the image on M need not be more than 1 in. at most. For this purpose we must have a colli- mator 10 ft. long. Two lenses of this focal length can be fixed one at each end, and a slit in front of that lens which is presented to the sun's rays. The arrangements followed will be the same as those given for the electric light. There appears no difficulty in producing a monochromatic image of almost any size if the collimator be suffi- ciently long and the face of the prism sufiiciently large to take in the whole of the image cast on it.* I have replaced the prism by Hat refraction gratings with most satisfactory results. The gratings I employed had about 6,000 and 12,000 lines to the inch. The images were sharply defined, but, of course, weaker than when the prism was employed. For solar work this should not be an objection, since there is plenty of light to work with. I show some pictures taken by the plan I have described. For my purpose the images are sufiiciently sharp, although simple uncorrected lenses have been employed. * On the Determination of the Photometric Intensity of the Coronal Light during the Solar Eclipse of 16th April, 1893." By Captain W. DE W. ABNEY, C.B., D.C.L., F.R.S., and T. E. THORPE, LL.D., F.R.S. Received April 14,— Read April 30, 1896. (Abstract.) In this paper the authors give the results of the measurements of the intensity of the light of the corona, as observed at Fundium in Senegal, on the occasion of the solar eclipse of April 16th, 1893. The methods employed by them were practically identical with those used at Grenada, in the West Indies, during the eclipse of 1886, an account of which is given in the ' Phil. Trans.,' A, 1889, * It should be mentioned that to minimise diffraction the slits should be used fairly wide. Hence a long collimator such as described and a good dispersion will be necessary to obtain the best definition of the sun's image. — April 30. 16 Determination of Coronal Light during Eclipse. p. 363, with certain slight modifications suggested by their experience on that occasion. Two sets of observations were made : the first with a photometer equatorially mounted, and designed to measure the comparative brightness of the corona at different distances from the moon's limb, and the second with an instrument arranged to measure the total brightness of the corona, excluding as far as possible the sky effect. In both cases the principle of photometry was that of Bctnsen, the intensity of the coronal light being compared with that of a standard glow-lamp, according to the method of Abney and Festing. The measurements with the equatorial photometer were made by Dr. Thorpe, assisted by Mr. P. L. Gray, B.Sc., those with the second or integrating instrument were made by Mr. Jas. Forbes, jun., assisted by Mr. Willoughby, of H.M.S. " Alecto." The mean of ten concordant readings with the integrating photo- meter reduced to values of light intensity and expressed in Siemens* units was 0'026. The measurements with the equatorial photometer show that the visual brightness of the corona of the 1893 eclipse varied within comparatively wide limits, and that, at all events close to the moon's limb, there was marked variation in local intensity. If the several values taken in the direction of the poles and equator are grouped as in the former paper (loc. cit.), they are found to afford a curve almost identical in character with that already given, showing that the diminution in intensity from the moon's limb outwards is less rapid than accords with the law of inverse squares. The results are as follows : — Photometric Intensity. Distances in solar semi-diameters. Observed. Law of inverse squares. 1893. 1886. 1-6 0-060 0-066 0-066 2-0 0-048 0-053 0-042 2-4 0-038 0-043 0-029 2-8 0-030 0-034 0-022 32 0-024 0-026 0-016 3-4 0-018 0-021 0-013 These numbers would appear to show that the actual brightness of the corona was probably not very dissimilar at the two eclipses, the slight apparent diminution observed during the 1893 eclipse being, The Total Eclipse of the Sun, April 16, 1893. 17 in all probability, due to the haze, or opalescence, in the air which prevailed at the time. This haze, caused more by suspended and finely divided solid matter than by precipitated moisture, undoubtedly contributed to the general sky-illumination at the time of totality. The actual gloom during this phase of the eclipse at Fundium was certainly much less than at Grenada in 1886. It must not be for- gotten, however, that the altitude of the sun was very different on the two occasions. At Grenada it was only about 19° : the amount of cloud was from seven to eight (overcast = 10) at the time of to- tality, and much of the cloud was in the neighbourhood of the sun : whereas at Fundium the sun's altitude was t52°, and the sky was of a bluish- grey colour and practically free from cloud. The effect of these different conditions in the sky in the neighbour- hood of the disc is seen in Mr. Forbes' measurements when com- pared with those of Lieutenant Douglas, at Grenada. The ten fairly concordant observations at Fundium give, as already stated, an average value of 0*026 Siemens units at 1 ft. from the screen; and the value observed by Lieutenant Douglas, 15 seconds after totality, with the same photometer, although with a different lamp and galva- nometer, was 0'0197 light units. " The Total Eclipse of the Sun, April 16, 1893. Report and Discussion of the Observations relating to Solar Physics." By J. NORMAN LOCKYER, C.B., F.R.S. Received April 17, —Read April 30, 1896. (Abstract.) The memoir first gives reports by Mr. Fowler and Mr. Shackleton fis to the circumstances under which photographs of the spectra of the eclipsed sun were taken with prismatic cameras in West Africa and Brazil respectively on April 16, 1893. These are followed by a detailed description of the phenomena recorded, and a discussion of the method employed in dealing with the photographs. The coronal spectrum and the question of its possible variation, and the wave- lengths of the lines recorded in the spectra of the chromosphere and prominences, are next studied. Finally, the loci of absorption in the sun's atmosphere are con- sidered. The inquiry into the chemical origins of the chromospheric and prominence lines is reserved for a subsequent memoir. The general conclusions which have been arrived at are as follows : — (1) With the prismatic camera, photographs may be obtained with VOL. LX. 18 The Total Eclipse of the Sun, April 16, 1893. short exposures, so that the phenomena can be recorded at short intervals during the eclipse. (2) The most intense images of the prominences are produced bj the H and K radiations of calcium. Those depicted by the rays of hydrogen and helium are less intense, and do not reach to so great a height. (3) The forms of the prominences photographed in monochromatic light (H and K), during the eclipse of 1893, do not differ sensibly from those photographed at the same time with the coronagraph. (4) The undoubted spectrum of the corona in 1893 consisted of eight rings, including that due to 1474 K. The evidence that these belong to the corona is absolutely conclusive. It is probable that they are only represented by feeble lines in the Fraunhofer spectrum r if present at all. (5) All the coronal rings recorded were most intense in the brightest coronal regions, near the sun's equator, as depicted by the coronagraph. (6) The strongest coronal line, 1474 K, is not represented in the spectrum of the chromosphere and prominences, while H and K do not appear in the spectrum of the corona, although they are the most intense radiations in the prominences. (7) A comparison of the results with those obtained in previous eclipses confirms the idea that 1474 K is brighter at the maximum than at the minimum sun-spot period. (8) Hydrogen rings were not photographed in the coronal spec- trum of 1893. (9) D3 was absent from the coronal spectrum of 1893, and reasons are given which suggest that its recorded appearance in 1882 was simply a photographic effect due to the unequal sensitiveness of the isochromatic plate employed. (10) There is distinct evidence of periodic changes of the con- tinuous spectrum of the corona. (11) Many lines hitherto unrecorded in the chromosphere and prominences were photographed by the prismatic cameras. (12) The preliminary investigation of the chemical origins of the chromosphere and prominence lines enables us to state generally that the chief lines are due to calcium, hydrogen, helium, strontium, iron, magnesium, manganese, barium, chromium, and aluminium. None of the lines appear to be due to nickel, cobalt, cadmium, tin, zinc, silicon, or carbon. (13) The spectra of the chromosphere and prominences become more complex as the photosphere is approached. (14) In passing from the chromosphere to the prominences, some lines become relatively brighter but others dimmer. The same line sometimes behaves differently in this respect in different prominences. On some Paleolithic Implements found in Somaliland. 19 (15) The prominences must be fed from the outer parts of the solar atmosphere, since their spectra show lines which are absent from the spectrum of the chromosphere. (16) The absence of the Fraunhofer lines from the integrated spectra of the solar surroundings and uneclipsed photosphere shortly after totality need not necessarily imply the existence of a reversing layer. (17) The spectrum of the base of the sun's atmosphere, as recorded by the prismatic camera, contains only a small number of lines as compared with the Fraunhofer spectrum. Some of the strongest bright lines in the spectrum of the chromosphere are not represented by dark lines in the Fraunhofer spectrum, and some of the most intense Fraunhofer lines were not seen bright in the spectrum of the chromosphere. The so-called " reversing layer " is therefore incom- petent to produce the Fraunhofer spectrum by its absorption. (18) Some of the Fraunhofer lines are produced by absorption taking place in the chromosphere, while others are produced by absorption at higher levels. (19) The eclipse work strengthens the view that chemical sub- stances are dissociated at solar temperatures. " On some Palaeolithic Implements found in Somaliland by Mr. H. W. Seton-Karr." By Sir JOHN EVANS, K.C.B., D.C.L., Treas. and V.P.R.S. Received April 27,— Read April 30, 1896. Although some account of his recent discoveries in Somaliland (tropical Africa) has already been given to the Anthropological Institute by Mr. Seton-Karr, and has been published in their Journal,* these discoveries seem to me to have so wide an interest, and such an important bearing on the question of the originaJ home of the human race, that I venture to call the attention of this Society to them. In the course of more than one visit to Somaliland for sporting purposes, Mr. Seton-Karr noticed, and brought home for examination, a number of worked flints, mostly of no great size, which he laid before the Anthropological Section of the British Association, at the meeting last year at Ipswich. f Although many of these specimens were broad flat flakes trimmed along the edges so as to be of the "le Moustier type" of M. Gabriel de Mortillet, and although the general fades of the collection was suggestive of the implements being of palaeolithic age, they did not afford sufficient evidence to enable a satisfactory judgment to be formed whether they undoubtedly belonged to the palaeolithic period. * Vol. 25, p. 271. f Eeport, 1895, p. 824. C 2 20 On some Palaeolithic Implements found in Somaliland. Before returning to Somaliland, Mr. Seton-Karr visited my collec- tions, and studied the various forms of implements found in the river-gravels and Pleistocene deposits in different parts of the world, so as to become familiar with their leading features ; and on revisiting Somaliland during the past winter, he was fortunate enough to meet with a large number of specimens in form absolutely identical with some from the valley of the Somme and other places which he had seen in my collection. Of this identity in form there can be no doubt, and though at present no fossil mammalian or other remains have been found with the implements, we need not hesitate in claiming them as palaeolithic. They seem to be scattered all over the country, and to have been washed out of sandy or loamy deposits by the action of rain, or, in some instances, to have been laid bare by the wind. They appear also to occur most frequently in the neighbourhood of existing water- courses, which is at all events suggestive of the beds in which they occur having been in some manner the result of river-action. It is, however, at present premature to enlarge on the circumstances of their discovery. Their great interest consists in the identity of their forms with those of the implements found in the Pleistocene deposits of North Western Europe and elsewhere. Any one comparing the implements from such widely separated localities, the one with the other, must feel that if they have not been actually made by the same race of men, there must have been some contact of the closest kind between the races who manufactured implements of such identical forms. Those from Somaliland occur in both flint (much whitened and decomposed by exposure) and in quartzite, but the implements made from the two materials are almost indistinguishable in form. Those of lanceolate shape are most abundant, but the usual ovate and other forms are present in considerable numbers. Turning westward from Somaliland we meet with flint implements of the same character found by Professor Flinders Petrie afc a height of many hundred feet above the valley of the Nile. A few have been discovered in Northern Africa, they recur in the valley of the Manzanares in Spain, in some districts in Central Italy, and abound in the river-valleys of France and England. Turning eastward we encounter implements of analogous forms, one found by M. Chantre in the valley of the Euphrates, and many made %of quartzite in the laterite deposits of India ; while in Southern Africa almost similar types occur, though their age is somewhat uncertain. That the cradle of the human family must have been situated in some part of the world where the climate was genial, and the means of subsistence readily obtained, seems almost self-evident ; and that these discoveries in Somaliland may serve to elucidate the course by which human civilisation, such as it was, if not indeed the human On the Liquation of certain Alloys of Gold. 21 race, proceeded westward from its early home in the east is a fair subject for speculation. But, under any circumstances, this dis- covery aids in bridging over the interval between palaeolithic man in Britain and in India, and adds another link to the chain of evidence by which the original cradle of the human family may eventually be identified, and tends to prove the unity of race between the inhabitants of Asia, Africa, and Europe, in Palaeolithic times. " On the Liquation of certain Alloys of Gold." By EDWARD MATTHEY, F.S.A., F.C.S., Assoc. R.S.M. Communicated by Sir G. G. STOKES, Bart., F.R.S. Received April 14,— Read May 7, 1896. The molecular distribution of the metals in alloys of gold and of metals of the platinum group has been described by me at some length, in a series of papers which have already been published by the Royal Society.* New interest in the subject has, however, arisen in connexion with the extraordinary development in various parts of the world especially in South Africa, of certain processes which are now employed for extracting gold from its ores. Their use has been attended with the introduction into this country of a series of alloys of gold and the base metals which have hitherto rarely been met with in metallurgical industry. The base metals associated with the gold in these cases are usually the very ordinary ones lead and zinc, but their presence in the gold has given rise to unexpected difficul- ties, as the distribution of the precious metal in the ingots which reach this country is so peculiar, that it is not possible to estimate the value of the ingots by taking the pieces of metal required for the assay, by any of the well-known methods now in use. The grouping of the metal in these ingots presents much scientific as well as industrial interest, and the following is a brief state- ment of the facts which have been observed. r A. An ingot of gold weighing 3'545 kilograms was assayed with a view to subjecting it to the ordinary operation of refining. A piece of metal was, therefore, cut from the base of the ingot at the point marked A, and the following are the results of four assays made on this piece of metal : — Gold 1 665-8 2 663-6 3 662-4 4 658-0 Average 662'45 * ' Phil. Trans.,' A, vol. 183; p. 629, 1892. ' Boy. Soc. Proc.,' vol. 47, p. 180, 1890. 22 Mr. E. Matt hey. There was also 0'061 part of silver present in 1000 parts of the mass, the remainder being base alloy. Another set of assays from the same ingot, but from the opposite end, at the point marked B, gave the following results : — 1 429-9 2 459-5 3 439-0 4 . 429-0 Silver . 0'071 Average 439'35 The difference in the amount of gold between the two opposite ends of the ingot was, therefore, no less than 223*10 parts in 1000. The base metal present was proved by analysis to be chiefly zinc, Jead, and copper, as the following results will show on metal taken by a " dip," i.e., from the molten metal : — Zinc 15-0 Lead 7'0 Copper 6'5 Iron 2-2 Mckel 2-0 Silver 7'0 Gold (by difference) 60%3 100-0 B. Another ingot of alloyed gold weighing 12 223 kilograms gave at different parts of the ingot the following results by assay : — Four assays on a piece of metal cut at a — top of ingot — Gold. Silver. 1 664-0 0-090 2 662-5 0091 3 465-0 0-076 4 . 661-5 0-091 On the Liquation of certain Alloys of Gold. 23 Three assays at b — bottom of ingot^— Gold. Silver. 1 332-5 0181 2 652-0 0-095 3 410-5 0-057 And seven assays were made from a " dip," viz. — Gold. Silver. 1 622-0 2 574-4 0-072 3 653-5 0-011 4 623-2 5 580-0 0-138 6 603-3 — 7 . 562-3 — Average of the whole number of the assays made .... 576*2 0'090 It became evident, therefore, that the only method of determining the true quality of this ingot consisted in actually separating the gold and silver in mass, and this was effected by dissolving in nitro- hydrochloric acid, the silver being recovered as chloride and reduced to metallic silver, and the gold precipitated by iron chloride as pure metallic gold. The result of this operation yielded Gold 7-504 kilograms. Silver 0-928 which showed that the standard fineness of the ingot was Gold 614-0 Silver 75'8 and its true value £1,028 ; while the value, as calculated from the average of the assays previously made, Gold 576 Silv-r 0-090 would have been only £965. Analysis proved that the metals present other than gold were as follows : — 24 Mr. E. Matthey. Silver 8-1 Lead 16-4 Zinc 9-5 Copper 4*0 Iron 0-3 Gold (by difference) .. 61-7 100-0 The cause of the differences revealed by assays made from metal cut from various parts of the ingot was clearly due to liquation. ; but previous experience failed to afford any guide to the probable distribution of the precious and base metals in the ingot. C. Another instance, and on a mu.cn larger quantity of gold alloy than the two former examples, was afforded by an ingot weighing 39*625 kilograms, which showed such great variation in its gold con- tents at various points that the ingot was re-melted and cast into two separate ingots, from which portions of metal were removed for assay by drilling. by boring 709-0 by boring ror-s All these results fare the averages of assays made in triplicate, and a " dip " assay from the melted metal showed that it contained 701 parts of gold in 1000. The analysis of this metal gave — Zinc ,. 7-1 Lead 4'9 Copper 4'8 Iron 1'4 Silver 9-2 Gold (by difference). . 72'6 100-0 As in the former case, the gold and silver present were isolated in mass, and the actual yield of fine gold and silver so obtained was as follows : — On the Liquation of certain Alloys of Gold. 25 Gold 27-914 kilograms. Silver 3-568 which proved that the actual gold standard of the ingot was 703'9. The base metal in two similar ingots was found by analysis to be composed as follows : — (492 ) (494.) Silver 8'9 8'0 Lead 9'0 77 Zinc 4-8 8-5 Copper 5-2 3'2 Iron 0-4 V6 Nickel 0-8 1-8 Gold (by difference).. 7O9 69'2 100-0 100-0 from which it would appear that the presence of one or both of the metals — zinc and lead — bears in some degree upon these variations in quality — it being well known that gold will alloy, and be constant in quality, with either silver or copper or with both in almost any proportions. Advancing progressively, I now cite an instance of irregular dis- tribution in a much baser alloy of gold. An ingot of base gold alloy (P. 13) weighing 9'570 kilograms. Determinations from the top of this ingot gave results Point a — Gold. 265-0 378-4 383-0 Silver. 213 From the bottom, point 6 — 527-2 560-0 66 545-5 From a " dip " taken from the fused alloy - 561-0 618-5 75 683-0 differences which are too significant to need comment. 26 Mr. E. Matthey. In order to ascertain the effect exerted by these two metals — lead and zinc — in conjunction with gold, I prepared an alloy of 700 parts pure gold and 300 parts pure lead, and after mixing and casting into an open mould I cast the melted alloy into a spherical mould 2 in. in diameter, made of cast iron. Determinations made from different parts, after cutting the sphere into two halves, gave the following results, the assays being made in triplicate upon each portion of metal removed. (The weight of this sphere was a little over 2 kilograms.) FIG. 1. The result shows a decided tendency of the gold to liquate to the centre of the mass. In the next experiment gold was alloyed with lead and zinc in the following proportions : — Gold 75 parts. Lead 15 „ Zinc 10 „ adding the zinc when the alloy of the first two metals was thoroughly fluid, and after casting this into an open mould, the alloy was remelted and cast into the 2-in. spherical mould before mentioned. This alloy was extremely hard and very brittle. Portions removed from the different parts of the sphere, after cutting it across, gave the following results : — FIG. 2. On the Liquation of certain Alloys of Gold. 27 There is evidence of re- arrangement by liquation in this case which sends gold to the centre, but the result is complicated, as gravity appears also to send gold to the lower portion of the spherical mass. The foregoing mixture (No. 2) of Gold 75 parts. Lead 15 „ Zinc 10 „ was now further alloyed by the addition of 5 per cent, of pure copper and cast into a sphere which was very hard and brittle, and weighed about 2 kilograms. The following are the results at the points shown : — FIQ. 3. Here again, gravity appears to send gold to the lower portion of the sphere. The question arises, does the silver play any part in the distribu- tion of the baser metals, lead and zinc ? I therefore melted sphere No. 3 with 10 per cent, of silver, so that there were present : — Gold 63-4 (by difference) Silver 7'8 Copper 5'1 Zinc 8'8 Lead 14-5 Iron 0'4 100-0 and cast into an open mould, and subsequently into the spherical mould. The following were the results obtained of fine gold at the points indicated : — 28 Mr. E. Matthey FIG. 4. This sphere seems constant all over. In order to see what was the effect with pure gold alloyed with metallic zinc only, I cast an alloy of fine gold with 5 per cent, of zinc into a 3-in. spherical mould. The weight of the sphere was 3-438 kilograms. The results were as follows : — Fia. 5. (Five per cent. zinc. A slight but decided tendency of liquation of gold towards the centre. The same alloy, containing 95 per cent, of gold and 5 per cent, of zinc, was then alloyed with a further 5 per cent, of zinc and cast into the same sphere. This weighed 4*218 kilograms. The results were as follows : — On the Liquation of certain Alloys of Gold. FIG. 6. 638-8 (Ten per cent, zinc.) Feeling a little diffident about these results, 1 recast the foregoing alloy of gold with 10 per cent, of zinc, into the same mould. The results were as follows : — FIG. 7. (Ten per cent, zinc.) This shows that there is stiii a tendency in this gold alloy with 10 per cent, of zinc to become enriched towards the centre. This 10 per cent, alloy was then alloyed with a further 5 per cent, of zinc and cast into the same spherical mould. The weight of this sphere was 4'021 kilograms. The results were : — 30 Mr. E. Matthey. TIG-. 8. (Fifteen per cent, zinc.) It is abundantly evident therefore, that zinc alone will not account for the differences in the ingots of impure gold ; and the question arose, will the presence of a definite amount of silver in any way prevent the irregularity in composition ? To test this I alloyed the gold, which contained 15 per cent, of zinc so that it might also contain 7'5 per cent, of silver. This was cast into the 3-in. sphere and weighed 3'934 kilograms, and assays made on portions of metal cut from it gave the following results : — FIG. 9. dOi-4 (Fifteen per cent, zinc.) It was intended to contain — Zinc 15-0 Silver 7'5 Gold . 77-5 100-0 On the Liquation of certain A Hoys of Gold. 31 the extra richness of the gold over 77'5 being due to the volatilisa- tion of the zinc. This experiment appears to confirm, that on pp. 27, 28 (see results of fig. 4). The foregoing experiments show that lead is far more effective as a cause of liquation than zinc, and the question arises, do zinc and lead separate into distinct layers by gravity when they are simultaneously present in a mass of gold, as they are known to do when they (lead and zinc) are melted together and allowed to solidify slowly. If they do separate, are they respectively associated with precious metal ? Professor Roberts-Austen has given us a method of investigating such a problem. He has shown that it is easy to place a suitably protected thermo- junction in a mass of cooling alloy, and obtain by photography a record of the cooling of the mass,* a method which was employed by me for determining the temperatures at which the metals arsenic and antimony separate from bismuth. Applying this method to a mass weighing 44 grams of an alloy containing : — Gold , . . . 75-0 Lead 15-0 Zinc lO'O The following curve, No. I, is an autographic record of its solidifi- cation : — CTTEVE No. I. Cooling curve of Au,Ca,Zn,Pb. 73,° C. 655° C. ("main point) 407° C. *r°C. 206° C. Time . * See ' Boy. Soc. Proc./ vol. 52, p. 467. 32 Mr. E. Matthey. From this it will be evident, from the horizontal position (6) (of the curve No. I) that the mass solidifies as a whole at 635° C. ; bat there is a second break c in the curve at a temperature of 407° C. ; and there is yet a third break at d, 247° C. These latter points evidently are connected with the solidifying points of lead and zinc, but it is probable that these metals are, in solidifying, associated with some gold. The second curve, No. II, represents the cooling of the same mass of gold with 10 per cent, of silver added. It will be seen that the metal has still one main solidifying point 6, at 645° C. The lower point (c) of the former curve is entirely absent, but there is an indication of the lead point at 206°. The results clearly indicate that silver is a solvent common to both zinc and lead, which are not, as in the previous case (Curve I) free to separate from each other. Such a mass should be fairly uniform in composition, and assays from different portions of it proved it to be so. CURVE No. II. Time. Cooling of alloy of Au,Cu,Zn,Pb, (&'S$gA/') C.J.*/4°C. The latter curve (II) seems to change its direction at 767°, which is above the main solidifying point of the mass, and it remains to be seen whether this is of any significance. The inspection of the curves so obtained at once led me to infer that silver mast be a solvent for zinc and lead when these are present On the Liquation of certain Alloys of Gold. 33 in gold, and with the clear indication thus afforded I proceeded to make the following experiments : — The alloy- Zinc 11-0 Silver 7-5 Gold 81-5 100-0 and weighing 5'680 kilograms, was now alloyed by the addition of lead to produce a similar metal to P. 13 (see p. 25), say : — Zinc . . Lead. . Silver Gold.. 10 20 7 63 100 and this was cast into two spheres, a 2-in. sphere and a 3-in. sphere. This alloy was so hard and brittle that I was compelled to cut these spheres into two by sawing them. When so cut asunder it was evident that the upper portions of both these spheres had a marked white appearance, as compared with the lower portions, which possessed the yellow colour of gold. The 3-in. sphere weighed 3'484 kilograms. Portions removed from these two spheres at the points indicated showed the following results : — FIG. 10. And those from the 2-in. sphere, weighing 0'880 kilogram — VOL. LX. 34 On the Liquation of certain Alloys of Gold. Fia. 11. Very marked separation takes place in both instances, the differ- ences at various points of the sphere being very remarkable and forcibly illustrating the difficulties to which reference is made at the commencement of this paper. As, however, it appears, that when a certain amount of silver is present, the irregularity in composition disappears, I alloyed this mixture of — Zinc . . Lead. . Silver Gold., 10 20 7 63 with more silver, so that it contained 15 per cent, of silver (nearly half the united amounts of zinc and lead present in the alloy). This, cast into the 3-in. spherical mould, showed the following results at the points indicated. In appearance, the metal, when sawn in two, was homogeneous. The weight of the sphere was 3'459 kilo- grams. FIG. 12. Occurrence of the Element Gallium in Clay-Ironstone. 35 There is still evidence of liquation of gold towards the centre, but comparison of fig. 12 with those which immediately precede it' will show how greatly the arrangement of the alloy has been modified by the presence of the additional 8 per cent, of silver. The proportion of silver in this alloy was proved by assay to be 15'5 per cent. As there was still evidence of liquation, the metal was cast with .still more silver, making 20 per cent, of silver in all. The alloy, when cast into a mould, proved to be almost uniform in composition] the difference between the centre and the extreme portions being very slight. Liquation had practically ceased, a fact which proves incontest- ably that silver is the solvent for the base metals, zinc and lead, when they are alloyed with gold. Conclusions.— (1) Alloys of gold with base metals, notably with lead and zinc, now largely often met with in industry, have the gold concentrated towards the centre and lower portions, which renders it impossible to ascertain their true value with even an approximation to accuracy. (2) When silver is also present these irregularities are greatly modified. The method of obtaining "cooling-curves" of the alloys shows that the freezing points are very different when silver is present and when it is absent from the alloy. (3) This fact naturally leads to the belief that if the base metal present does not exceed 30 per cent., silver will dissolve it and form a uniform alloy with gold. (4) This conclusion is sustained by the experiments illustrated by figs. 9, 10, 11, 12, which, in fact, gradually lead up to it, and enable a question of much interest to be solved. 41 Ori the Occurrence of the Element Gallium in the Clay- Ironstone of the Cleveland District of Yorkshire. Prelimi- nary Notice." By W. N. HARTLEY, F.R.S., Professor of Chemistry, and HUGH RAM AGE, A.R.C.S.I., F.I.C., Assistant Chemist in the Royal College of Science, Dublin. Received April 13,— Read May 7, 1896. In the course of an investigation of flame spectra at high tempera- tures (« Phil. Trans.,' A, vol. 185, pp. 1029—1091 (1894) ) extended to the basic Bessemer process, the authors were occupied last July and August in observing the flames from the converters at the North Eastern Steel Company's Works, at. Middlesbrough-on-Tees. A large number of photographs were taken in series during the pro- gress of the "blow," and also of the "after blow," but these will D 2 36 Occurrence of the Element Gallium in Clay-Ironstone. form the subject of another communication dealing with the chem- istry of the process. Some of the photographs were remarkably fine in definition, and they extended from the less refrangible limit of the red rays to the ultra-violet, about wave-length 3240. It may be mentioned here, however, that every line and band in the different spectra was identified. Some of the photographs afforded evidence of very unusual constituents in the mixture of gases and vapours, which by their combustion and incandescence give the Bessemer flame. The identity of these could have been established only by means of very complete investigation of oxy- hydro°"en blowpipe spectra. Apart from all technical considera- tions which were kept in view, and of such purely scientific questions as were involved in similar previous researches carried out by one of us, the examination of these spectra was of great 'interest, more especially because of the proof of the rare element, gallium, being- present in the Bessemer metal, and in the roasted ore from which it was extracted. It was shown by very careful analyses that the gallium was concentrated in the iron, but all details of the operations involved in its separation and of the quantitative determinations are- reserved for a future communication. The evidence of the existence of gallium in the ore and in the metal rests on the measurements of the wave-lengths of the lines in a large number of photographed spectra and upon the relative strengths of the lines in the different materials examined and in the precipitates obtained therefrom. The following examples show the nature of this evidence : — • 1. Evidence from the Bessemer Flame Spectra. Seventy-six of the photographed spectra of the Bessemer flame contain a strong line with wave-length about 4171'5, which does not appear to be related to any other line in these spectra, and belongs, therefore, to some other element than those otherwise identified. 2. Evidence from the Spectrum of the " Mixer Metal " and of the different substances separated by its Chemical Treatment. The " mixer metal " heated in the oxy-hydrogen flame gives a spectrum of iron with a strong line having a wave-length of 417T6. The residue left after dissolving the iron by boiling with hydro- chloric acid also gives this line 4171*6 very strongly. • Precipitates obtained by boiling the solution of the iron with am- monium acetate give the line 41 71 '6 and also a weaker line, wave- length 40327. Electromotive Properties of Electrical Organ of Malapterurus. 37 The latter line is seen only in the absence of manganese, as it very nearly coincides with one of the group of strong manganese lines ; it is, therefore, obscured in the spectra of the Bessemer flame and of the crude iron. ^ The oxide of gallium was separated as far as possible from all other substances and heated in the oxy-hydrogen flame and the character- istic spectrum was then photographed from this oxide. 3. Evidence from the Roasted Ore, and substances separated therefrom. The roasted Cleveland ore was heated alone for thirty-five minutes in the oxy-hydrogen flame, it gave only a very faint indication of one line in the spectrum of gallium. The solution extracted from the ore by digesting it with warm dilute hydrochloric acid of double normal strength, when boiled with ammonium acetate gave a precipi- tate, the spectrum of which contained the line 4171'6 fairly strong. The silicious residue insoluble in strong hydrochloric acid, when decomposed by fusion with caustic potash and subsequent boiling with water, after concentration of the solution so as to retain the gallium, gave a spectrum containing both lines, 4171/6 and 40327. All other elements had been removed. The wave-lengths given are on Rowland's scale. The lines were measured on many plates and also repeatedly on the same plate, the results being the same in each case. Electromotive Properties of the Electrical Organ of Malapterurus electricm" By FRANCIS GOTCH, M.A. (Oxon.), F.R.S., and G. J. BuRCH, M.A. (Oxon.). Received April 2,— Read May 7, 1896. (Abstract.) The experiments were made upon six specimens of Malapterurus electricus, 12 to 15 cm. in length, brought from the River Senegal by Mr. A. Ridyard (ss. "Niger"), and generously placed at the. dis- posal of the authors by the Liverpool Corporation Museum Committee, to whom and to Dr. Forbes, the Director of the Museum, the authors desire to express their thanks. Three of the specimens were killed, in order to carry out experi- ments upon the isolated organ. The instrumental methods employed by the authors for determining for the first time the characters and time relations of the activity of the organ response were chiefly the following : — (a.) The record of the frog nerve muscle galvanoscope. (&.) The galvanometer connected with a suitable .rheotome. 28 Messrs. F. Gotch and G. J. Burch. (c.) The capillary electrometer, a large number (about 250) photo- graphic records being taken of the movements of the meniscus. Facsimile reproductions of typical records are given in the fuller communication. The electrometer was used either shunted by a resistance of from 80 to 100 ohms, or in connection with the outer plates of a special condenser, the inner plates of which were con- nected with the fish or its electrical organ. The organ responded to mechanical or electrical excitation of its nerves after removal from the fish, the response being unaffected by 1 per cent curare, or 1 per cent, atropine solution. No response could be evoked by such chemical agents as sodium chloride, gly- cerine, or weak acid, when applied either to the organ or its efferent nerve. The conclusions drawn by the authors from the experiments on the isolated organ and on the entire uninjured fish may be summarised as follows : — (1) The isolated organ responds to electrical excitation of its nerves by monophasic electromotive changes, indicated by electrical currents- which traverse the tissue from the head to the tail end ; this response commences from 0'0035" at 30° C. to 0'009" at 5° C. after excitation, the period of delay for any given temperature being tolerably constant, (2) The response occasionally consists of a single such monophasic electromotive change (shock) developed with great suddenness, and subsiding completely in from 0*002" to 0'005", according to the tem- perature ; in the vast majority of cases the response is multiple, and consists of a series of such changes (shocks) recurring at perfectly regular intervals, from two to thirty times (peripheral organ rhythm) ; the interval between the successive changes varies from 0'004" at 30° C. to O'Ol" at 5° C., but is perfectly uniform at any given tempera- ture throughout the series. (3) Such a single or multiple response (in the great majority of cases the latter) can also be evoked by the direct passage of an induced current through the organ and its contained nerves, in either direc- tion heterodromous (i.e., opposite in direction to the current of the response) or homodromous. (4) The time relations of the response are almost identical whether this is evoked by nerve-trunk (indirect) stimulation, or by the passage of the heterodromous induced current. (5) There is no evidence that the electrical plate substance can be excited by the induced current apart from its nerves, i.e., it does not possess independent excitability. (6) The organ and its contained nerves respond far more easily to the heterodromous than to the homodromous induced current, and the period of delay in the case of the latter response is appreciably lengthened. Electromotive Properties of Electrical Organ of Malapterurus. 39 (7) The peripheral organ rhythm (multiple response) varies from about 100 per second at 5° C. to about 280 per second at 35° C. (8) One causative factor in the production of the peripheral rhythm is the susceptibility of the excitable tissue to respond to the current set up by its own activity (self excitation). (9) In the uninjured fish mechanical or electrical excitation of the surface of the skin beyond the limits of the organ evokes a reflex response with a long delay (0'03" to 0'3") ; this reflex response con- sists of groups of shocks, each group showing the peripheral organ rhythm, but separated from its neighbour by a considerable interval of time (reflex or central rhythm). (10) In the uninjured fish electrical excitation of the skin over the organ evokes a response which may consist of a direct peripheral organ effect followed by such a reflex effect. (11) The minimal total reflex delay at 20° C. is 0'023", giving a central excitatory time of about O'Ol". (12) The reflex or central rhythm in our specimens showed a maximum rate of 12 per second and an average rate of from 3 to 4 per second. (13) The number of separate groups in the reflex response recurring at the intervals mentioned in the preceding paragraph was in our fish limited to from 2 to 5. (14) The E.M.F. of each single change in the organ response depends upon the number of effective plates with their nerves, and in 10 cm. of excited organ cannot possibly be less than 75 volts, and is probably much nearer 150 volts. As in our specimens the number of plates in series in 1 cm. of organ was 180, this gives a minimal possible E.M.F. of 0'04 volt, and a probable E.M.F. of 0*07 volt for each plate. The authors further conclude that, since each lateral half of the organ is innervated by the axis cylinder branches of one efferent nerve cell, and has no independent excitability, the specific characters of the reflex response of the organ express far more closely than those of muscle the changes in central nerve activity, and are pre- sumably those of the activity of a single efferent nerve cell. The single efferent nerve cell, the activity of which is thus for the first time ascertained, shows — (a.) A minimum period of delay of O'OOS" to O'Ol". (6.) A maximum rate of discharge of 12 per second. (c.) An average rate of discharge of 3 to 4 per second. (d.) A susceptibility to fatigue showing itself in the discharge failing after it had recurred from two to five times at the above rates. 40 Dr. T. W. Eden. 4i The Occurrence of nutritive Fat in the Human .Placenta. A Preliminary Communication." By THOMAS WATTS EDEN, M.D., M.R.C.P. Communicated by Dr. PYE SMITH, F.R.S. Received April 23,— Read May 7, 1896. (From the Laboratories of the Conjoint Board of the Royal Colleges of Physicians (Lond.) and Surgeons (Eng.)). Recently, while examining specimens of ripe placentae for fatty degeneration, I was struck by the regularity of the occurrence of fat in this structure, and especially by the nature and extent of its dis- tribution. I was then led to examine a series of specimens taken at different periods of gestation, with the result that a free deposit of fat was found in ten different placentae, all of which I believe to be non-pathological, and ranging practically through, all periods of gestation, from the sixth week up to term. The method employed for the demonstration of this fat, was to take slices from different parts of the placenta, and harden them for a few days in Muller's fluid; then to transfer thin strips, not exceeding one- third of an inch in thickness, to Marchi's fluid (1 per cent, solution of osmic acid 1 part, Muller's fluid 2 parts) for a week. The pieces were then embedded in paraffin, cut with a rocking microtome, and stained lightly with saffranine, eosine, or logwood and cosine, pr mounted unstained. By this process the fat is completely blackened, while the other tissues retain their normal staining reactions, so that the outlines of the fat-containing cells can be distinctly made out. By this method I have been able to demonstrate the constant occurrence of fat in certain well-defined regions of the human placenta. In the young human placenta, the epithelial covering of the villi consists of two layers, a superficial, nucleated, plasmodial layer, and a deep cellular layer. In a six weeks' ovum I found fat in the form of minute droplets in both these layers, but much more abundantly in the former than in the latter. These fat droplets show comparatively little variation in size, and they remain discrete, showing little or no tendency to form larger droplets by fusion ; they are confined to the perinuclear protoplasm, and are never found in the nuclei, which remain unaltered in number, form, and arrangement. The stroma of these villi contains here and there a trace of fat, but it is apparently healthy, and is furnished with well-formed wide capillaries filled with blood. The villi are, in fact, to all appearance healthy. Every villus does not show this deposit of fat, but it is present in very large numbers of them ; in every field of the microscope several villi • The Occurrence of nutritive Fat in the Human Placenta. 41 containing fat may be found. The amount of fat also varies con- siderably. In a young ovum the plasmodial layer of the villi shows great pro- liferative activity ; it throws out numerous club-shaped processes or buds, which represent the first stage in the development of new villi. These buds very frequently contain large numbers of minute fat droplets. I believe that this is a point of very great importance, showing, as it does, that the deposit of fat occurs in actively growing tissues of undoubted vitality. ]n the ripe placenta the proliferation of the plasmodial layer has ceased, and degenerative changes are present in scattered regions. But, of course, the great majority of the villi retain their vitality, and in these villi a free deposit of fat is present, showing the same distribution and characters as in the young placenta. I have also found a similar deposit of fat in the serotina. The six weeks' ovum, above referred to, showed very many decidual cells containing minute, discrete droplets of fat in the perinuclear proto- plasm. A placenta of the sixth month also showed an abundant fat deposit in' the same region. At term, the serotina shows many degenerative changes, and although it contains fat, it may well be doubted whether, at this period, this is a physiological deposit. The placenta, indeed, appears to be a storehouse of nutritive fat, just as is the liver. This appears to throw some light on what has long been one of the problems of foetal physiology, viz., the source frcm which the foetus obtains its supplies of fat. Diffusible substances such as sugar, salts, peptones, &c., were supposed to pass by osmosis from the maternal blood in the inter- villous spaces, to the foetal blood in the villi. But this could not be assumed of indiffusible substances such as fat. The truth would seem to be that fat is deposited from the maternal blood in the epithelium of the villi, and stored up there by the foetal tissues for their use. No great accumulation of fat occurs, as it appears to be from time to time absorbed and disposed of by the foetal circulation. It is, however, not altogether clear how si deposit of fat in the decidual cells can be made available for the purposes of foetal nutrition. Since finding this fat deposit in the human placenta, I have begun a series of comparative observations upon the placentae of other mammals. Up to the time of writing, I have examined two rabbits' placentae, one from an early, and the other from a late, period of gestation. In both there was a marked deposit of fat, chiefly in the superficial glandular layer of the maternal placenta, but also, though to a less extent, in the processes of the chorionic mesoblast, which form the homologues of the villi of the human placenta. The process appears to correspond closely to that observed by Mr. George Brook, in the transmission of fat from the yolk to 42 Mr. E. A. Mincliin. Note on the Larva and the the segmenting germinal area, by the parablast of mesoblastic ova.* I was under the impression when these observations were made, that fat had never been found, in this form, in the placenta before. I find that I am to some extent anticipated by a paper in the * Archiv fiir Gynaekologie,' February, 1896. t One of the authors (Aschoff) wished to examine a malignant uterine growth, which he believed to be of the nature of Deciduoma malignum. Before doing so, he examined several specimens of young human ova, in Older, as he says, to learn something of the struciure of growing chorionic villi. Some of the specimens he hardened in Memming's solution, and in all of these he found fat in the plasmodial layer of the villi. Aschoffs description of the fat deposit agrees very closely with that already given of my own specimen. " An den Flemmingschen Praparaten ist das Syncytium dadurch ausgezeichnet, dass es in seiner Bandzone eine dichte Anhaufung feinster Fetttrb'pfchen tragt. Dieselbe sind bald sehr fein, bald grobkornig, aber in den betreffenden Abschnitten des Syncytiums stets von gleicher Grosse Die Fetttropfcben iiberall sich finden, wo Chorionepithelzellen, in directesten Stoffwechselaustausch mit den Intervillosenraumen treten" (p. 531). Aschoff scarcely appreciates the physiological importance of the observation, but there can be no doubt that his observations and my own are mutually confirmatory. *' Note on the Larva and the Postlarval Development of Leucosolenia vanabilis, H. sp., with Remarks on the Develop- ment of other AsconidsB." By E. A. MINCHIN, M.A., Fellow of Merton College, Oxford. Communicated by Professor E. RAY LANKESTER, F.R.8. Received April 25,— Read May 21, 1896. Introductory Remarks. Through the kind hospitality of Professor de Lacaze-Duthiers, I was able to spend the spring and summer of last year at the marine laboratories of Banyuls-sur-Mer and Roscoff, where I was chiefly engaged in studying the embryology of the Asqons. In Banyuls I obtained the larvae of Leucosolenia cerebrum, H. sp., in June, and of L. reticulum, O.S. sp., in July. In Roscoff I found the larvae of L. varialilis, H. sp., all through August and the early part of September, * " Formation of the Germinal Layers in Teleostei," ' Eoy. Soc. Edin. Trans./ 1896. f " Ueber bosartige Tumoren der Chorionzotten," Apfelstedt und Aschoff. Postlarval Development of Leucosolenia variabilis, H. sp. 43 and of L. coriacea, Mont, sp., in September. Owing to the inexperi- ence with which I approached the difficult task of rearing these larvae, my results are not so complete in all details as T could wish, but in the case of L. variaUlis I was able to obtain a more or less perfect developmental series, and in the other three species I was able to make out satisfactorily the main points in (he metamorphosis, especially the important question of the relation between the cell- layers of the larva and those of the adult. I hope to bring my inves- tigations to completion during the present year, but, in the meantime, the results obtained seemed to me of sufficient importance to form the subject of a preliminary note. The material which I collected and preserved was further studied at Munich, in the laboratory of Professor Richard Hertwig, to whom I am indebted for much kind help and advice, as well as hospitality. The Development of Leucosolenia variabilis (Ascandra variabilis, H.). The larvae of L. variabilis are of the so-called amphiblastula type, but in many respects more primitive than the amphiblastula larva hitherto described in other Calcarea. The minute larvce (70 — 80 /*, in length, 50 — 60 fi in breadth) leave the mother sponge by the osculum, and at once rise to the surface of the water, where they swim for about twenty-four hours. They then sink to the bottom, where, after swimming about slowly for twelve to twenty-four hours more, they fix themselves and undergo metamorphosis. The larval life thus lasts for thirty-six to forty-eight hours. The oval larva (figs. 1 and 2)* is divided into an anterior region composed of ciliated cells and a posterior region composed of non- ciliated granular cells. The centre of the transparent larva is occu- pied by a conspicuous mass of yellowish-brown pigment. The ciliated cells are slender and elongated, reaching from the pigment to the surface of the body. Each cell bears a single flagellum, and the body of the cell is divided into an internal refractile portion and an external granular portion. These two portions of the cell are so distinct in the living object that a superficial examination gives the impression of an internal layer of refractile cells covered by an external granular layer, but by more careful investigation it is easy to make out that these two apparent layers are merely parts of a single layer of cells. The ciliated cells situated more posteriorly entirely lack the retractile inner portion, and appear granular throughout. They are also slightly broader, and have more convex outer surfaces than the other ciliated cells, forming an equatorial zone of intermediate cells, not very distinct in the living object. The * Figs. 1 — 6 represent the development of L. varialilis, x 1000 diameters. All but 1 and 2 are semidiagrammatic and combined from different preparations. 44 Mr. E. A. Minchin. Note on the Larva and the FiG. 1. — Newly hatched larva. region of the intermediate cells is generally marked by a slight con- striction, giving a waist, as it were, to the larvae. The granular cells are much fewer in number than the other elements, and are also of much larger size, but there are gradations in this respect, those placed at the posterior pole being much larger than those which border upon the intermediate cells. During the free-swimming larval period, considerable changes take place in the relative proportions of the different parts of the larvae. In the newly hatched larva (fig. 1) the anterior ciliated region is relatively large, with a very broad granular border to the cells, and the posterior granular cells are few in number. The number of granular cells now increases at the expense of the ciliated cells. Some of the ciliated cells, by absorption of the internal refractile portion of the cell, become intermediate cells, and these, in their turn, absorb their flagellum, increase in size, and become granular cells. This process goes on pari passu with a decrease in the granular border of the ciliated cells. In the larva of about twenty-four hours (fig. 2), the granular cells form a mass equal to that of the ciliated cells, and the latter have now a very narrow granular border. In Postlarval Development of Leucosolenia variabilis, H. sp. 45 FIG. 2. — Larva of second day. short, granular cells are formed during larval life by modification of ciliated cells, the intermediate cells being a stage in this process. Sections of larvae confirm and amplify the results obtained from a study of the living object (fig. 3). The inner portion of each ciliated cell, which in life appeared refractile, is seen to contain a series of vacuole-like structures, containing granular masses suspended in their interior. At the junction between the internal vacuolated and external granular portions of the cell is situated the opaque and deeply staining nucleus, which has a form like an onion, and is con- tinued externally into the flagellum. Often the inner side of the nucleus is indented by the vacuole beneath it, sometimes to such an extent that the nucleus has the form of a crescent in section. The intermediate cells are very distinct in sections, and by some methods of preservation and staining, e.g., osmic acid followed by picrocarmine, their protoplasm takes up the stain in a remarkable manner, so that larvae treated in this way appear to have a brightly coloured equatorial zone. They lack the vacuolated inner portion, characteristic of the 46 Mr. E. A. Minchin. Note on tJie Lama and the Ffa. 3. — Longitudinal section of larva. ciliated cells proper, and their nuclei are larger and paler with one or two nucleoli. The nucleus of the first intermediate cell frequently presents a curious appearance, being swollen out into a large vesicular structure containing two or three chromatin masses. This condition is apparently in connexion both with a process of rearrangement of the chromatin and with the absorption of the vacuoles. The granular cells are arranged in a single layer, and have large pale nuclei with nucleoli. Often the nucleus of the cell nearest the intermediate cells has a pointed outer end, evidently indicating the former connexion with the flagellum. Sections reveal a remarkable set of structures in connexion with the central pigment, which is now seen to have the form of a tube, open in front and behind, and enclosing a rounded, lens-like body, apparently a gelatinous mass filling the central cavity, the remnant, doubtless, of the segmentation cavity. Behind these bodies are a number of cells with coarse granules and small, very opaque, deeply staining nuclei.* One of these cells is placed in the longitudinal axis * Cf. Dendy's account of the larva of Crrantia la'byrinthica for similar cells, " On the Pseudogastrula stage in the Development of Calcareous Sponges," ' Roy. Soc. Yictoria Proc.,' 1889, pp. 93—101. Postlarval Development of Leucosolenia variabilis, H. sp. 47 of the larva, and its nucleus is usually, but not always, elongated in the same direction, so as to have a rod-like form. The whole struc- ture, with pigment, lens-like body, and central granular cells, gives strongly the impression of a primitive, light-perceiving organ. The pigment itself is lodged in the inner ends of the ciliated and inter- mediate cells, and is, no doubt, the same pigment as that observed by Metschnikoff* and Schulzet in the inner ends of the ciliated cells in the larva of Sycandra raphanus. As the intermediate cells pass into the condition of granular cells, they leave the pigment behind, so that the pigment is thickest in the region of the intermediate cells, at the sides of the lens-like body. The larva is thus composed of four kinds of cells, which may be termed the ciliated, intermediate, granular, and central cells. Since the intermediate cells are merely a transitional form between the ciliated cells proper and the granular cells, we have to reckon with three classes of cells only in the fully developed larva. The fixation takes place by the anterior pole of the larva, and the granular cells grow round the ciliated cells. The metamorphosis is complete in a few hours. Sections of fixed stages of the first day of fixation (fig. 4) show them to be composed of two very distinct cell Fia. 4.— Section of larva shortly after fixation, the metamorphosis not quite complete. layers : (1) a compact central mass of cells, easily recognisable, by their opaque, irregularly shaped nuclei and vacuolated cell protoplasm, as the former ciliated cells, surrounded by (2) a single layer of flattened epithelial cells, the former granular cells of the larva. . No trace is to be found of the central cells, which appear to be thrown out together with the pigment, at the metamorphosis. The inner mass is" the future gastral layer of the sponge, the outer epithelium the future dermal layer. * «Zur Entwicklungsgeschichte der Kalkschwamme," ' Zeitschr. f . Wiss. Zool,' V°f " Ueber den Bau u'nd Entwicklung von Sycandra raphanus," ib., vol. 25, suppl., pp. 247-280, Taf. XVIII-XXI. 48 Mr. E. A. Minchin. Note on the Larva and the The two component layers very soon begin to undergo changes of form and structure, which are best described separately, since the two layers develop more or less independently of one another, and a given stage in the development of one layer is not always found combined with one and the same stage in the development of the other. The dermal layer becomes divided (fig. 5) into two kinds of cells r (a) cells which retain the original form and characters and remain on the surface, and (b) cells with smaller nuclei, which sink below the outer epithelium and form a scattered layer between it and the FlG. 5.— Section of stage about twenty-four hours after fixation. The left side is represented as slightly in advance of the right side. gastral cells. The former (a) secrete each a single monaxon spicule, which appears first on the inner side of the nucleus, but soon grows out and projects free from the surface. The latter (6) unite into groups and secrete the triradiate spicules. The monaxons appear first, as in Sycandra raplianus* and begin to appear about twenty-four hours after fixation, the triradiates about twelve hours later. The dermal layer has thus become divided into two parts, which gradually assume the adult characters. I have not observed the origin of the pores. The gastral layer, at first a compact mass with no definite arrange- ment, soon begins to form a cavity (fig. 5). The cells assume a radiate arrangement, and a split-like lumen appears in the centre. Sometimes two or more such lacunar spaces arise, at first quite independent of one another, but later fusing to form a single gastral cavity, which soon becomes very large, causing the larva to increase considerably in size as a whole. At first the cavity is surrounded on all sides by gastral cells, but as it increases in size a spot appears where gastral cells are wanting, and the cavity is limited only by dermal cells (fig. 6). This is the region of the future osculum, and the dermal cells at this spot form the future oscular rim, where collar * Metschnikoff, loc. cit. Fostlarval Development of Leucosolenia variabilis, H. sp. 49 ¥10. 6. — Section of stage of about the fourth day of fixation. cells are lacking. The gastral cells are at first elongated, but later become shorter, and take on the characteristic appearance of collar cells. I have not been able to make out whether all the gastral cells become collar cells, or whether some of them do not become the wandering cells of the adult, which seems very probable. The osculum appears about the sixth day of fixation. The Development of Leucosolenia cerebrum, H, L. reticulum, 0. S., and L. coriacea, Mont. These three species have larvae of the type with which we are familiar from the descriptions of Metschnikoff * and Schmidt, f namely, oval ciliated blast ulee, in which an inner mass is formed by immigration of cells into the interior. The process is most easily followed in the more transparent larva of L. reticulum (fig. 7), where the modification of ciliated cells into granular cells, and their sub- sequent immigration, takes place at the posterior pole. When the larva is ready for fixation, a considerable quantity of granular cells has been formed, though the cavity is far from being obliterated. In the opaque larvae of L. cerebrum and coriacea the process is more * " Spongiologische Studien," 'Zeitschr. f. Wiss. Zool.,' vol.32, p. 362, Taf. XXIII. f " Das Larvenstadium von Ascetta clatkrus und Ascetta primordialis" ' Arch. f. Mikr. Anat.,' vol. 14, pp. 249—263, Taf. XV, XYI. YOL. LX. B 50 Mr. E. A. Minchin. Note on the Larva and the FIG. 7. — Optical section of larva of L. reticulum, first day, x 500. difficult to follow, but in both immigration appears to take place from any point on the surface. In L. cerebrum and L. reticulum the larva swims for about twenty-four hours at the surface, and as long at the bottom, and fixes on the third day. L. coriacea, on the other hand, is remarkable for its abbreviated larval period as compared with the two Mediterranean species, since the larva fixes in a few hours, a fact doubtless in con- nexion with its life between tide marks, where the violent currents to which it is exposed renders a very sheltered, and therefore limited, habitat necessary for so delicate an organism. After fixation, the larva undergoes changes whereby the ciliated cells become surrounded by the formerly internal granular cells, so that the ciliated external layer of the larva represents the gastral layer of the adult, while the inner mass becomes the dermal layer ; the reverse of what was supposed by Metschnikoff and Schmidt (loc. cit.) to take place. In L. cerebrum I was able to observe the first appearance of the spicules. As in variabilis, the complete metamorphosis results in a stage in which the gastral cells form a compact internal mass, snr- Postlarval Development of Leucosolenia variabilin, //. Sp. 51 rounded by a single layer of dermal cells. Some of the cells of the dermal epithelium then form themselves into groups, usually of three cells, and each cell of such a group secretes the ray of a spicule. The first spicales are usually triradiate, but quite irregular in form, and at their first appearance they are quite superficial, their secreting cells forming part of the general epithelium, but later they become covered by the remaining epithelium, so that the dermal layer becomes divided into an internal connective tissue layer and an external flat epithelium. The process is essentially similar to that occurring in variabilis, except that in the latter the cells of the flat epithelium secrete each a monaxon spicule, which in cerebrum is not the case. General Considerations. The larva of L. variabilis is of interest as affording a transition from larvae such as that of L. reticulum, to the amphiblastula larva of the Sycons. The larva of reticulum (fig. 7) is composed of (1) ciliated cells, comparable to those of the amphiblastula, of which some (2) at the hinder pole are undergoing modification, and may be compared with the intermediate cells, and of (3) internal granular cells comparable to the posterior granular cells of the amphiblastula. To obtain a larva like that of variabilis from the type represented by reticulum, we must suppose the large cavity of the latter reduced to the extent to which this has occurred in the former. Then the granular cells which are formed at the posterior pole must remain where they are, since the cavity is too small to contain them, and, as more ciliated cells are continually being modified arotmd them, we get a larva with the three kinds of cells arranged as in variabilis. The central cells of variabilis — on the origin of which I have no observations to bring forward — are probably to be regarded as con- stituting a larval organ, [a, special adaptation of no importance for the postlarval development. The development of both reticulum and variabilis points to an early stage in which the larva is composed entirely of similar and equi- valent ciliated cells. I have not seen such a stage in any species, and doubt if it actually occurs in nature ; it is more probable that the process of cell differentiation, begins before the larva is hatched in all cases. In the absence of segmentation stages, it is impossible to decide this question; nevertheless, the facts seem to me to indicate, as the primitive larva in ascon phylogeny, a blastula composed of indif- ferent ciliated cells, in which a second type of cells (the future dermal layer) is formed by modification of certain of the cells. The collar- cell layer of the adult is derived directly from the primitive ciliated cells of the blastula. Comparing, now, the larva of variabilis with that of Sycon raphanus, 52 On the Larva and Development of Leucosolenia variabilis. as described by Schulze, it is obvious that the development is essen tially similar in both, the chief difference being with regard to the periods at which the various events take place. In both the granular cells increase greatly in number, but in raphanus this takes place while the larva is still in the maternal tissues, ' as is obvious from Schulze's figures,* and the larva is hatched in a condition similar to that of variabilis when about to fix. In variabilis the granular cells do not surround the ciliated cells until after fixation ; in raph- anus this process is begun while the larva is still swimming, and the granular cells may even give rise to spicules (monaxons) during the free swimming period (Metschnikoff, loc. cit.). It is obvious that in Sycon we have before us a hastening and shortening of the development, and, allowing for these embryological adaptations, we are able to understand how, from a larva such as that of reticulum, there has arisen a type of development apparently so different as that of the Sycon amphiblastula. The most important event in the post-larval development is the differentiation of the dermal layer into the outer epithelium and the inner connective tissue layer. This might seem at first sight to be a process comparable to the formation of a new layer, a mesoderm ; so that from this period onwards the sponge would be a three-layered organism. I do not, however, take this view, for the following reason. The immigration of cells from the epithelium to form the layer of triradiates is not an event, like the formation of a germ layer, which takes places once and for all in the life cycle of an individual, but it goes on whenever new triradiates are formed. In adult ascons I have found that the triradiates and the basal rays of the quadriradiates arise from cells of the outer epithelium which migrate inwards and arrange themselves into groups to form spicules, each ray being secreted by one cell or by cells derived from the division of a single cell. In the adult also the nuclei of the spicule secreting cells diminish in size after quitting the epithelium. Hence in the develop- ment of the sponge also, I regard this process as one not of blasto- genetic, but of histogenetic significance. The fact that in variabilis the epithelial cells also secrete spicules is to my mind a decisive proof of the unity of the dermal layer.f » ' Zeitschr.f. Wiss. Zool.,' vol. 25, suppl., Taf. XX and fig. 3, Taf. XIX. Schulze refers this increase in the number of the granular cells to their multiplication by cell-division, but as the granular cells do not at the same time decrease in size, it seems more probable that their increase is due, as in variabilis^ to their numbers being recruited from the clear (ciliated) cells. f Schulze has also figured very clearly the relation of the dermal cells to the monaxon spicules, one epicule to each cell, in the young fixed stages of Sycon raphanus (' Zeitschr. f . Wiss. Zool.,' vol. 31, pi. XIX, figs. 10, 11), although he states in the text that the spicules arise in the hyaline substance between the two layers. Helium and Argon, their Inactivity. 53 " Helium and Argon. Part III. Experiments which show the Inactivity of these Elements." By WILLIAM RAMSAY, Ph.D., F.R.S., and J. NORMAN COLLIE, Ph.D., F.R.S.E. Received April 22,— Read May 21, 1896. To chronicle a list of failures is not an agreeable task ; and yet it is sometimes necessary, in order that the record of the behaviour of newly discovered substances may be a complete one. It is with this object that we place on record an account of a number of experiments made to teat the possibility of forming compounds of helium and argon. It will be remembered that in our memoir on Argon,* Lord Rayleigh and Professor Ramsay described numerous experiments, made in the hope of inducing argon to combine, all of which yielded negative results. Two further experiments have been since made — again without success. 1. The electric arc was maintained for several hours in an atmo- sphere of argon. The electrodes were thin pencils of gas carbon, and, previous to the introduction of the argon, the arc was made in a vacuum, and all gas evolved was removed by pumping. Argon was then admitted up to a known pressure, and the arc was again made. A slow expansion took place ; one of the electrodes di- minished in length, and the bulb became coated with a black deposit. The resulting gas was treated with caustic soda and with a solution of ammoniacal cuprous chloride, and, on transference to a vacuum- tube, it showed the spectrum of argon along with a spectrum resembling that of hydrocarbons. Having to leave off work at this stage, a short note was sent to the * Chemical News ' on a Possible Compound of Argon. On resuming work after the holidays, the gas was again investigated, and, on sparking with oxygen, carbon dioxide was produced. Bat it was thought right again to treat the gas with cuprous chloride in presence of ammonia, and it now appeared that when left for a sufficient time in contact with a strong solution, considerable contraction took place, carbonic oxide being removed. There can, therefore, be no doubt that, although apparently all gas had been removed from the carbon electrodes before admitting argon, some carbon dioxide must have been still occluded, probably in the upper part of the electrodes, and that the prolonged heating due to the arc had expelled this gas and converted it into monoxide. It was, indeed, inexplicable how an expansion should have taken place unless by some such means; for the combination of a monatomic gas must necessarily be accompanied by contraction. It appears, therefore, certain that argon and carbon do not combine, even at * < Phil. Trans.,' vol. 186, A. 54 Drs. W. Ramsay and J. Norman Collie. the high temperature of the arc, where any product would have a chance of escaping decomposition by removing itself from the source of heat. It is hardly necessary to point out that such a process lends itself to the formation of endothermic compounds such as acetylene, and it was to be supposed that if argon is capable of combination at all, the resulting compound must be produced by an endothermic reaction. 2. A product rich in barium cyanide was made by the action of producer gas on a mixture of barium carbonate and carbon at the intense temperature of the arc. This product was treated by Dumas' process so as to recover all nitrogen ; and, as argon might also have entered into combination, the nitrogen was absorbed by sparking. All the nitrogen entered into combination with oxygen and soda, leaving no residue. Hence it may be concluded that no argon enters into combination. For the successful carrying out of these experiments we have to thank Mr. G. W. MacDonald. 3. A mixture of argon with the vapour of carbon tetrachloride was exposed for several hours to a silent discharge from a very powerful induction coil. The apparatus was connected with a gauge which registered the pressure of the vapour of the tetra- chloride and of the argon of which it was mixed. Careful measure- ment of the pressure was made before commencing the experiment, and after its completion. Although a considerable amount of other chlorides of carbon was produced, no alteration of pressure was noticeable; the liberated chlorine having been absorbed by the mercury present. Here again the argon did not enter into the reaction, but it was recovered without loss of volume. The remaining experiments relate to attempts to produce com- pounds of helium. The plan of operation was to circulate helium over the reagent at a bright red heat, and to observe whether any alteration in volume occurred — an absorption of a few c.c. could have been observed — or whether any marked change was pro- duced in the reagent employed. As a rule, after the reagent had been allowed to cool in the gas, all helium was removed with the pump, and the reagent was again heated to redness, so as, if a com- pound had been formed, to decompose it and expel the helium. Every experiment gave negative results ; in no case was there any reason to suspect that helium had entered into combination. A short catalogue of the substances tried may be given. 4. Sodium distilled in the current of gas, and condensed in drops with bright metallic lustre. The glass tube in which it was heated became covered with a coating of 5. Silicon, which caused no absorption. 6. A mixture of beryllium oxide and magnesium, yielding metallic beryllium, was without action. Helium and Argon, their Inactivity. 55 7. Zinc and, 8, cadmium distilled over in the current of gas. 9. A mixture of boron oxide and magnesium dust, giving ele- mental boron, produced no absorption. 10. Similarly, a mixture of yttrium oxide and magnesium dust had no effect. 11. Thallium was heated to bright redness in the gas, retaining its metallic lustre. 12. Titanium oxide mixed with magnesium dust was heated to bright redness, and caused no absorption. 13. Similar absence of action was proved with thorium oxide and magnesium powder. 14. Tin and, 15, lead, were heated to bright redness in the current of gas, and remained untarnished. 16. Phosphorus was distilled in the gas, and caused to pass through a length of combustion-tube heated to softening. Some red phos- phorus was formed, but no alteration of volume was noticed. 17. The same process was repeated with elemental arsenic. 18. Antimony and, 19, bismuth, at a bright red heat, retained their metallic lustre. 20. Sulphur and, 21, selenium, were treated in the same way as phosphorus ; no action took place. 22. Uranium oxide, mixed with magnesium dust, was heated to bright redness in helium. No change, except the reduction of the >xide, took place. The mixture was allowed to cool slowly in the irrent, and the helium was removed with the pump till a phos- )horescent vacuum was produced in a vacuum tube communicating rith the circuit. The mixture was re-heated, and no helium was rolved— not even enough to show a spectrum. The vacuum remained [impaired. It had been hoped that elements with high atomic weight, such as thallium, lead, bismuth, thorium, and uranium might have effected )mbination, but the hope was vain. 23. A mixture of helium with its own volume of chlorine was exposed to a silent discharge for several hours. The chlorine was contained in a reservoir, sealed on to the little apparatus which had the form of an ozone apparatus. ISTo change in level of the sulphuric acid confining the chlorine was detected after the temperature, raised by the discharge, had again become the same as that of tlie room. Hence helium and chlorine do not combine. 24. Metallic cobalt in powder does not absorb helium at a red heat. 25. Platinum black does not occlude it. 26. It is not caused to combine by passage over a mixture of soda- lime and potassium nitrate heated to bright redness. This was hardly to be expected, for it resists the action of oxygen in presence of caustic soda, even when heated by the sparks which traverse it. 56 Lord Rayleigh. On the Amount of Argon and 27. A mixture of soda-lime and sulphur consisting of polysulph- ides causes no change of volume in a current of helium passed over it at a bright red heat. 28. Induction sparks in an ozone apparatus passed through a mix- ture of helium with benzene vapour in presence of liquid benzene for many hours, gave no change of volume. The benzene was, of course, altered, but the sum of the pressures of the helium and the benzene- vapour remained as at first. Had helium been removed, contraction would have occurred. This ends the catalogue of negative experiments. Any compound of helium capable of existence will probably be endo thermic, and the two methods of producing endothermic compounds, where no simul- taneous exothermic reaction is possible, are exposure to a high tem- perature, at which endothermic compounds show greater stability, and the influence of the silent electric discharge. These methods have been tried, so far in vain. There is, therefore, every reason to believe that the elements, helium and argon, are non-valent, that is, are incapable of forming compounds. "On the Amount of Argon and Helium contained in the Gas from the Bath Springs."* By LORD RAYLEIGH, Sec. R.S. Received April 30,— Read May 21, 1896. The presence of helium in the residue after removal of nitrogen from this gas was proved in a former paper, f but there was some doubt as to the relative proportions of argon and helium. A fresh sample, kindly collected by Dr. Richardson, has therefore been ex- amined. Of this 2500 c.c., submitted to electric sparks in presence of oxygen, gave a final residue of 37 c.c., after removal of all gases known until recently. The spectrum of the residue, observed at atmospheric pressure, showed argon, and the D3 line of helium very plainly. The easy visibility of D3 suggested the presence of helium in some such proportion as 10 per cent., and this conjecture has been con- firmed by a determination of the refractivity of the mixture. It may be remembered that while the refractivity of argon approaches closely that of air, the relative number being 0'961, the refractivity of helium (as supplied to me by Professor Ramsay) is very low, being only 0*146 on the same scale. If \ve assume that any sample * I am reminded by Mr. Whitaker tliat helium is appropriately associated with the Bath waters, which, according to some antiquaries, were called by the Eomans Aqua Soils. t 'Boy. Soc. Proc.,' vol. 59, p. 206, 1896. Magnetised Iron, $c., cooled to Temperature of Liquid Air. 57 of gas is a mixture of these two, its refractivity will determine the proportions in which the components are present. The observations were made by an apparatus similar in character to that already described, but designed to work with smaller quan- tities of gas. The space to be filled is only about 12 c.c., and if the gas be at atmospheric pressure its refractivity may be fixed to about 1/1000 part, By working at pressures below atmosphere very fair results conld be arrived at with quantities of gas ordinarily reckoned at only 3 or 4 c.c. The refractivity found for the Bath residue after desiccation was 0*896 referred to air, so that the proportional amount of helium is 8 per cent. "Referred to the original volume, the proportion of helium is 1P2 parts per thousand. " On the Changes produced in Magnetised Iron and Steels by cooling to the Temperature of Liquid Air." By JAMES DEWAR, LL.D., F.R.S., Fullerian Professor of Chemistry in the Royal Institution of Great Britain, and J. A. FLEMING, M.A., D.Sc., F.R.S., Professor of Electrical Engineering in University College, 'London. Received April 25, — Read May 21, 1896. The action of the low temperature produced by liquid air upon the magnetic moment of steel magnets was studied by one of us in a few cases in a preliminary research made some time ago.* We have re- cently returned to the subject and made further investigations on the influence of the low temperatures thus obtained on magnetised iron and steels of very various compositions, with the object of de- termining the nature of the changes which take place in the magnetic moment of small magnets constructed of these metals, when cooled gradually or suddenly down to the lowest temperature obtainable by the use of boiling liquid air. The arrangements adopted in this investigation were as follows : — A reflecting magnetometer consisting of three small magnetised needles of watch-spring steel, cemented to a concave glass mirror, suspended by a single cocoon fibre, was placed in a tube so as to be free from disturbance by draughts of air. The small magnets were 8 to 10 mm. in length. The image of a portion of the filament of an incandescent lamp was reflected by the mirror on to a divided scale placed at a distance of 70 cm. from the mirror. The edge of the very sharp image of the filament, focussed upon the scale, * Friday evening discourse at the Koyal Institution, "On the Scientific Uses of Liquid Air," by James Dewar, LL.D., F.E.S., January 19, 1894. VOL. LX. F 58 Profs. J. Dewar and J. A. Fleming. Changes produced in enabled any angular displacement of the magnetometer needle to be easily determined. The position of this magnetometer needle was regulated by the field produced by an external controlling magnet. The small magnet, the behaviour of which at low temperatures was to be studied, was placed behind the magnetometer, with its centre at a distance of 1 to 10 cm. from the centre of the magnetometer needle and its axis in a direction passing through the centre of the magneto- meter needle, and at right angles to the direction of the undis- turbed magnetometer needle. The magnet to be examined was fixed to a brass wire, held in a wooden support in such fashion that the magnet under examination could be easily removed from its position behind the magnetometer, and restored to it again exactly. A large number of samples of steel and iron were then prepared in the form of small needles, generally 15 mm. long and about 1 mm. in diameter. These steels comprised nickel steels, with various percentages of nickel; chromium steels, with various percentages , of chromium; aluminium steels, with various percentages of aluminium ; tungsten steels, manganese steels, silicon steel, ordinary carbon steels in various states of tempering, soft-annealed transformer iron, soft- iron wire, and the same irons hardened by hammering. For most of these samples of steels we were indebted to Mr. R. A. Hadfield, of Sheffield, who kindly furnished them* to one of us in the form of wires. These short steel magnets were then all magnetised to " satura- tion " by placing them for a few moments between the poles of a powerful electro -magnet. One by one they were then placed .in position behind the magnetometer, and the deflection produced on the magnetometer needle observed. In any particular case this deflection may be taken as approximately representing the intensity of magnetisation of the sample, although, owing to the varying sizes of the sample and distance from the magnetometer, the deflections in the case of different magnets are not comparable with one another, and cannot be taken as indicating the relative intensities of mag- netisation of two different samples. This, however, was not impor- tant, as our object was not to compare the absolute values of the magnetisation of different classes of steels, but to observe the mode of variation of the magnetisation of any one sample when cooled from ordinary temperatures down to the temperature of liquid air. The method of proceeding was then as follows : — Having adjusted the image of the lamp filament to the zero of the scale, the small magnet under observation was placed behind the magnetometer, and the deflection of the magnetometer needle observed. A small vacuum-jacketed cup, filled with liquid air, was then brought up underneath the sample, and by its aid the magnet cooled suddenly in situ to a temperature in the neighbourhood of —186° C. In the Magnetised Iron, $-c.t cooled to Temperature of Liquid Air. 59 many cases this sudden cooling immediately deprived the magnet of a considerable percentage of its magnetisation, and the magnetic moment was reduced. This, however, was not universally the case. In some cases, as in that of the chromium steels, the first effect of this sudden cooling was an increase in the magnetic moment of the magnet ; in other cases hardly any change in the magnetic moment at all. The vessel of liquid air was then removed, and the magnet allowed to heat up again, which it very quickly did, to the tempera- ture of the room, or rather to a temperature at which the deposit of snow formed upon the needle immediately on coming out of the liquid air, fully melted. This was taken to be afc about 5° C. It was found that each magnet had certain peculiarities of its own. Taking first the ordinary carbon steel (a sample of knitting-needle steel) we observe the following facts : — Knitting-needle Steel (a) Tempered Glass Hard. — -The first effect of cooling this magnet was to diminish the magnetic moment by 6 per cent. On allowing the magnet to heat up again to the ordinary temperature, the magnetic moment still further dimin- ished by about 16 per cent. On cooling again the magnetic moment increased 10 per cent., and from and after that time cooling the magnet always increased the magnetic moment, and allowing to heat up again to ordinary temperature always diminished the magnetic moment, the magnetic moment at — 185° C. being about 10 per cent, greater than the magnetic moment at 5° C. The first effect, therefore, of the cooling was to permanently diminish the magnetic moment, but after a few alternations of heating and cooling, the magnet reached a permanent condition in which its moment, when cold, was greater than its moment when warm. These changes of magnetisation may be best represented as in the diagram in fig. 1, in which the firm lines represent to some arbitrary scale the moment of the magnet when at its ordinary temperature of 5° C., and the dotted lines represent to the same scale the moment of the magnet when cooled to -185° C. Knitting-needle Steel (b) Medium Temper. — The same general results were obtained with knitting-needle steel tempered to a medium temper. The first effect of the cooling to the low tempera- ture was to diminish the moment of the magnet. On allowing it to heat up again the moment of the magnet diminished still more. The next cooling caused an increase of magnetic moment, and from and after that time the steel settled down into a permanent condition in which the magnetic moment was greater at — 185° C. than at 5° C. by nearly 20 per cent, of its value at 5° C. (see fig. 2). Knitting-needle Steel (c) Annealed Soft.— The same general course of events was noticed in the case of the knitting-needle steel when made soft by heating to a red heat and allowing it to cool very F 2 60 Profs. J. Dewar and J. A. Fleming. Changes produced in doo- 7oo 600- 5oo 400- 3oo 200 loo FIG. 1. — Knitting-needle steel (glass hard). slowly. In this case, however, the first diminution of magnetic moment was still greater. On first immersion in the liquid air the magnet lost about 33 per cent, of its moment. On allowing it to heat up again to 5° C. it still further diminished in moment, and from and after that point it arrived soon at a permanent condition, in which its moment, when cold, was greater than its moment when warm by 30 per cent, of its moment at 5° C. These Magnetised Iron, $c., cooled to Temperature of Liquid Air. Gl Zoo loo* FIG. 2. — Knitting-needle steel (medium temper). changes of the medium- and soft-tempered steel are represented by the lines in the diagrams 2 and 3, in which the firm lines are proportional to the magnetic moment of the magnet at 5° C., and the dotted lines proportional to the magnetic moment at — 185° C. It will be seen that,, in the case of this carbon steel, the effect of softening the steel is to make more pronounced the effect of the final temperature changes ; the change of moment caused by cooling from the ordinary temperature to the temperature of liquid air. when the permanent condition has been reached, being in the case of the glass- hard steel an increase of magnetic moment of about 12 per cent. ; in the case of the same steel with a medium temper about 22 per cent., and in the case of the same steel tempered very soft about. 33 per cent, (see fig. 3). Chromium Steels. — Observations were then made with the magnets of chromium steel, having respectively 0'29 per cent., 1'18 per cent., 5'44 per cent., and 9'18 per cent, of chromium. In all these cases the first effect of cooling the magnet was to cause at once an increase of magnetic moment, and the subsequent heating up again to the ordi- nary temperature caused a decrease of magnetic moment. These 62 Profs. J. Dewar and J. A. Fleming. Changes produced in 600- 2oo 100 FIG. 3. — Knitting-needle steel (tempered soft). 60- FIG. 4. — Chromium steel. Cr C Si Mn = 0-29 = 0-16 = 0-07 = 0-18 Fe = 99-30 o- FIG. 5. — Chromium steel. Cr = 1-18 C =0-27 Si = 0-12 Mn = 0-21 Fe = 98 -22 loo 50- FIG. 6. — Chromium steel. Cr = 5-44 C =0*27 Si = 0 '50 Mn = 0-61 Fe = 92-68 ICO So- o- FIG. 7. — Chromium steel. Cr = 9'18 C =0-71 Si = 0-36 Mn = 0-25 Fe = 89 -50 Magnetised Iron, # 76 Profs. J. Dewar and J. A. Fleming. On the Electrical " On the Electrical Resistivity of Pure Mercury at the Tem- perature of Liquid Air." By JAMES DEWAR, LL.D., F.R.S., Fullerian Professor of Chemistry in the Royal Institution, and J. A. FLEMING, M.A., D.Sc., F.R.S., Professor of Electrical Engineering in University College, London. Received May 19,— Read June 4, 1896. Although the electrical resistivity of mercury at ordinary tem- peratures has been carefully examined by many observers, and accu- rate determinations made of the specific resistance* and temperature coefficient, and in addition an examination made of the variation of resistivity in mercury when cooled to temperatures as low as 100° C.,f we considered it would be of interest to examine the behaviour of pure mercury in respect of change in electrical resist- ivity when cooled to the temperature obtained by the employment of boiling liquid air. With this object we prepared a sample of very pure mercury in the following manner : Ordinary distilled mercury was shaken up with nitric acid in the usual manner to free it from other metals, and then carefully dried. It was then introduced into a bent glass tube formed of hard glass. This bent tube had both ends sealed, and a side tube connected in at the bend, by which it could be con- nected to a mercury vacuum pump. Two or three hundred grammes of the mercury was then introduced into one bend, and a high vacuum made in the tube. The side tube was then sealed off from the pump, and the mercury distilled over from one leg into the other. For this purpose, one leg of the bent tube was placed in ice and salt, and the other submitted to a gentle heat just sufficient to make the mercury distil under reduced pressure without ever bringing it into active ebullition. In this way the mercury is distilled over at a very low temperature, and the portion condensing in the cooler limb of the bent tube is entirely free from any contamination with silver, lead, zinc, or tin. By performing this distillation two or three times suc- cessively on the same mercury, a small quantity of mercury is at last obtained in an exceedingly pure condition. A glass spiral tube was then formed of lead glass, consisting of a tube having an internal diameter of about 2 mm., and a length of about 1 metre. This tube was bent into a spiral of about twelve close turns, each turn being nearly 2'5 cm. in diameter, and the ends of this spiral provided with enlarged glass ends formed of wider tube. The spiral, * " The Specific Resistance of Mercury," by Lord Rayleigli and Mrs. Sidgwick (Phil. Trans. R. S., Part I, 1883). See, also, Mr. E. T. GUazebrook (Phil. Mag., Oct., 1885), for other values. f Cailletet and Bouty (Compt. Rend., 100, 1188, 1885). Resistivity of Mercury at the Temperature of Liquid Air. 77 after being cleaned, was then very carefully filled with the purified mercury, and by running the mercury through a spiral several times, all air bubbles and air film were finally removed. Into the wider ends of the spiral, amalgamated copper electrodes were introduced, consisting of copper wire 4'4 mm. in diameter ; the wider terminal ends of the spiral were then closed by paraffined corks to keep the copper electrodes in position. This spiral, full of mercury, was placed in a test-tube, and paraffin wax cast round it so as to enclose it entirely, leaving only the copper electrodes protruding. In order to determine the temperature of the mercury in the glass spiral tube, a platinum wire, the resistance of which was known at all tempera- tures down to the temperature of liquid air, was also embedded in the paraffin wax closely in contact with the glass spiral, and proper electrodes brought out to enable the resistance of this platinum wire to be determined. This mass of paraffin wax was then cooled down in a vacuum vessel kept filled up with liquid air until the whole mass reached the temperature of the liquid air. The glass spiral and thermometer enclosed in wax was then removed from the bath of liquid air and placed in a vacuum-jacketed test-tube, in order that it might warm up with extreme slowness to the ordinary temperature of the air. Having in this manner cooled the mass of paraffin enclosing the glass spiral filled with mercury and the platinum resistance wire entirely to the temperature of liquid air, a series of observations were taken with the aid of two observers, one measuring the resistance of the mercury by a Wheatstone's Bridge, while at the same time the other observer at another slide wire bridge measured the resistance of the platinum wire, these observations being taken quite simul- taneously, and continued whilst the mass heated up from — 197'9° (platinum temperature) to 0°. All proper corrections were then applied to correct for the resistance of the connecting wires and the bridge temperature ; and the observed resistance of the platinum wire employed was corrected to determine from its resistance tem- peratures in terms of the standard platinum thermometer employed by us in our investigations on the thermo-electric power of metals and alloys (see Dewar and Fleming, 'Phil. IVIag.,' July, 1895, p. 95). This standard thermometer has always been denoted by the letter Px. The following table shows the corrected resistance of the mercury column and the corresponding platinum temperatures, as also the specific resistance of the mercury calculated from the accepted re- sistivity at 0° C. :— 78 Profs. J. Dewar and J. A. Fleming. On the Electrical Resistivity of Pure Mercury in C.Gr.S. Units at various Tempera- tures in Platinum degrees. Platinum temperature, pt, in terms of tlie standard platinum thermometer Pi- Observed and corrected resistance of mercury in lead glass spiral in ohms. Resistivity of mercury in glass in C.G.S. units. -197-9 0 '0551 6970 -197-8 0 -0551 6970 -197-5 0-0551 6970 -196-9 0-0566 7160 -195-2 0-0581 7350 -191-2 0-0601 7600 -182-7 0-0641 8100 -173-2 0 -0721 9120 -168-4 0 -0761 9620 -165-1 0 -0781 9870 -157-4 0 -0836 10570 -149-7 0 -0886 11200 -143-0 0 -0931 11770 -131 -9 0-1011 12780 -128-3 0-1041 13160 -122-9 0 -1081 13670 -117-5 0-1121 14170 -108-4 0-1191 15060 -103-7 0 -1231 15560 - 97-0 0 -1281 16200 - 91-1 0 -1331 16830 - 85-0 0 -1381 17460 - 79-1 0-1432 18100 - 73-1 0 -1482 18740 - 67-4 0-1532 19370 - 63-2 0-1582 20000 - 57-6 0 -1632 20630 - 52-5 0-1682 21270 - 48-9 0 -1753 22160 - 47-0 0-1833 23180 - 46-0 0 -1883 23810 - 44-9 0 -1933 24440 - 44-2 0 -1983 25070 - 43-5 0 -2033 25700 - 43-0 0-2183 27600 - 42-4 0 -2283 28860 - 42-1 0 -2383 30130 - 41-9 0 -2484 31410 - 41-2 0 -2584 32670 - 40-8 0 -2784 35200 - 40-6 0 -2884 36460 - 40-4 0-3184 40260 - 39-7 0-3585 45330 - 39-5 0 -3885 49120 - 39-4 0 -4185 52920 - 39-3 0-4385 - 55440 - 39-1 0 -4785 60800 - 38-7 0 -5186 65570 - 38-5 0 -5486 69360 - 38-3 0-5786 73160 - 37-7 0-6086 76950 Resistivity of Mercury at the Temperature of Liquid Air. 79 Platinum temperature, pt, in terms of the standard platinum thermometer PL Observed and corrected resistance of mercury in lead glass spiral in ohms. Resistivity of mercury in glass* in C.G.S. units. - 37-6 0 -6387 80760 - 37*2 0-6587 83280 - 36-7 0 -6787 85810 - 36'0 0 -7087 89600 - 35-2 0 -7208 91140 - 33-7 0-7228 91380 - 31-2 0-7248 91640 0 0 -7440 94070 + 13-1 0-7518 95060 + 16-3 0-7540 95330 + 35-4 0 -7653 96760 Adopting the value for the specific resistance of pure mercury at 0° C., which has been recommended by the Board of Trade Electrical Committee, viz., 94,070 C.G.S. units, we have reduced the observed resistances of the mercury column at various temperatures to their equivalents in resistivity in absolute units, and placed these numbers against the observed resistances in the table above. As the specific resistance of mercury has been so carefully observed by many observers, we did not, for a moment, consider it necessary to attempt a further determination of this constant. On plotting out these values of the resistivity of mercury in the form of a curve in terms of the corresponding platinum temperatures, we find the resistivity curve has the form shown in fig. 1. It will be noticed that the resistivity of the mercury decreases gradually from the point at which the observations finished, viz., at +35° C., to the temperature — 36° on the platinum scale. At this point the resistivity rapidly decreases to about one-quarter of its value in falling from —36° to — 50°, and this sudden change all takes place within the range of about 14° of temperature. At the temperature of —50° on the plati- num scale the resistivity curve again changes direction, and con- tinues downwards in such a direction as to show that if produced along the same line from the lowest temperature actually observed, viz., —204° on the platinum scale, it would pass exactly through the absolue zero of temperature on this scale, which is — 283° pt. It is also interesting to note that the part of the curve which corresponds to the mercury in the liquid state is almost exactly parallel to that part of the curve which corresponds to mercury in the solid condi- tion, although, owing to the difference in the absolute values of the resistivities at these parts, the temperature coefficients as usually defined are very different. In the solid condition between the tem- peratures of —197-9° and —97°, the mean increase in resistivity is 80 Resistivity of Mercury at the temperature of Liquid Air. FIG. l. 700,000' 13? -Eoo? -loo° o° +/oo g &£ & $0,000' 80,000- 1 . 70,000- *-* 60,000- i 4 1 lOjOOO- 0- -A J ^ ^ ^^''' ~*7.. ' 1° -£oo*° —loo-0 o-° +/oc Temperature in Platinum Degrees. S3'14 C.G.S. units per degree rise of temperature on the platinum scale ; between — 108'4° and — 57'6° the mean increase in resistivity in C.Gr.S. units per degree is 109*6 ; in the liquid condition between the temperature — 35'2° and 0° the mean increase in resistivity in C.Gr.S. units per degree is 83*2; temperature measurement being on the platinum scale as above denned. It may be stated here that tem- peratures defined by this platinum scale do not differ by more than about 0'5° from the Centigrade scale down to temperatures of —100°, but that the temperature of boiling liquid oxygen which, on the Centigrade scale is denoted by —182°, is, on the platinum scale Magnetic Permeability, fyc., of Iron at Low Temperatures. 81 derived from our standard thermometer, denoted by — 196'7°. This would show, therefore, that the temperature coefficient as usually defined is O000884 between —35° and 0°.* These observations are specially interesting as giving additional proof that in the case of a metal of known purity the variation of resistivity, as the metal is continuously cooled, is such as to indicate that it would in all probability vanish at the absolute zero of tem- perature. In the case of mercury, we are able to obtain a metal in a state of almost perfect chemical purity, and which, when continuously cooled, passes into the solid condition under circumstances which are entirely favourable to the prevention of stresses in the interior of the metal, due to cooling. These measurements, therefore, afford a further confirmation of the law which we have enunciated as a deduction from experimental observations, that the electrical resis- tivity of a pure metal vanishes at the absolute zero of temperature. "" On the Magnetic Permeability and Hysteresis of Iron at Low Temperatures." By J. A. FLEMING, M.A., D.Sc., F.R.S., Professor of Electrical Engineering in University College, London, and JAMES DEWAR, LL.D., F.R.S., Fnllerian Professor of Chemistry in the Royal Institution, &c. Received May 27,— Read June 11, 1896. Although considerable attention has been paid to the changes produced in the magnetic properties of iron, particularly its magnetic permeability and hysteresis, at ordinary and at higher temperatures, but little information has been obtained up to the present on the behaviour of iron and steel as regards magnetic properties when cooled to very low temperatures. By the employment of large quantities of liquid air we have been able to conduct a long series of •experiments on this subject, the results of which we propose here briefly to summarise, leaving for a future communication fuller details and discussion of the results. The experimental work has consisted in making measurements, chiefly by ballistic galvanometer methods, of the permeability and hysteresis loss in certain samples of iron and steel, taken in the form of rings or cylinders. The first experiments were concerned with the variation of the magnetic permeability of .soft iron under varying magnetic forces, the iron being kept at a constant low temperature, obtained by placing it in liquid air in a state of very quiet ebullition in a vacuum vessel. * This is in close agreement with the values obtained by Guillaurae, Mascart, and Strecker for temperatures between 0°C. and +30° C. 82 Profs. J. A. Fleming and J. Dewar. On the Experiments on Annealed Swedish Iron. A cylinder of iron was formed by winding up a sheet of Saukey's best transformer iron (Swedish).* The width of the strip was 4*895 cm., the thickness O0356 cm. ; three complete layers of the sheet iron were used in forming the core. The area of cross-section of the side of the cylinder so formed was 0'5229 sq. cm. The mean diameter of the cylinder was 3*612 cm. This cylinder of iron was placed in a clay crucible packed with magnesia, the lid luted on with fire-clay, and the crucible then raised to a bright red heat in a forge, after which it was allowed to cool very slowly. The iron cylinder was thus carefully annealed out of contact with air or any material con- taining carbon. This soft annealed iron ring was then wound over with silk ribbon, and two windings of silk-covered copper wire placed upon it ; the first or primary circuit consisted of 131 turns of No. 26 double silk-covered wire ; the secondary circuit consisted of 112 turns of No. 36 silk-covered copper wire. The magnetising force to which the ring is subjected when a current is sent through the primary coil is measured by the value of 4<7r/10 x the ampere-turns per unit of length of the mean perimeter of the ring, and this, in the case of the present ring, reduces to the number 14'507 times the ampere current. The magnetising force in absolute units is therefore very closely given by the number obtained by multiplying the current flowing through the primary coil in amperes by !4'5. The resistance of the primary coil at about 15° C. was 0'92 ohm, and the resistance of the secondary at the same temperature 8'98 ohms. The secondary circuit of this ring coil or transformer was then connected through appropriate resistances with a ballistic galvanometer, having a resistance of 18 ohms. The primary circuit was connected through suitable resistances and a current reverser with a circuit of con- stant potential. By these arrangements it was possible to reverse a definite current passing through the primary coils, and by observ- ing the throw produced by the ballistic galvanometer, to calculate the induction in the iron core. The galvanometer was calibrated by reversing a known current passing through a long solenoid, in the centre of which was placed a secondary coil of known turns and dimensions, which was always kept in series with the secondary coil of the transformer. In this manner a series of observations was taken with gradually increasing magnetising forces. Before com- mencing each series of observations, the ring was carefully demagnet- ised by passing through the primary coil an alternating current, which was gradually reduced in strength to zero, the ring coil being thus brought into a magnetically neutral condition. An increasing * This sheet iron was kindly given to us by Mr. K. Jenkins, to whom our thanks are due. Magnetic Permeability, CJT., of Iron at Low Temperatures. 88 series of primary currents was successively passed through the primary coil and reversed, the throw of the ballistic galvanometer being noted in each case. In the first set of observations the ring was kept at the ordinary temperature of the air, 15° C., and in the second set it was immersed in liquid air, and the following table shows the results, both for the high and for the low temperature observa/tions. After taking a complete magnetisation curve at the ordinary tem- perature, the ring was immersed in liquid air, bringing its tempera- ture down to about —185° C., and a complete series of observations taken again in the same manner, previously having first carefully Table I.— Magnetisation Curve of Annealed Soft Iron (Sankey's Transformer Iron). At 15° C. At -18G°C. (in liquid air). Magnetising force. Induction. Permeability. Magnetising force. Induction. Permeability. H. B. p.. H. B. 1 fj.. 0725 1000 1379 0-841 1000 1189 0-971 2000 2060 1-174 2000 1704 1-174 3000 2555 1-407 3000 2132 1-378 4000 2903 1-595 4000 2508 1 -595 5000 3135 1-886 5000 2651 1-840 6000 3261 2-145 6000 2797 2-10 7000 3333 2-440 7000 2869 2-58 8000 3101 2-99 8000 2675 3-35 9000 2687 3-83 9000 2350 4-47 ]0000 2237 5-08 K'OOO 1968 6-27 11000 1754 7-05 11000 1560 8-99 12000 1335 9-72 12000 1234 12-35 13000 1053 13-11 13000 992 17-22 14000 813 17-90 14000 782 22-1 14400 652 21-35 14300 670 demagnetised the ring as described by an alternating current. The ring was then taken out of the liquid air, allowed to warm up again to the ordinary temperature, and another complete set of observations taken at the ordinary temperature. In this manner a series of eighteen complete sets of observations were taken, about half of them being at lo° C. and half of them at — 185° C. In cooling the ring in liquid air, it was found to be important to cool it slowly by holding it some time in the dense gaseous air lying over the liquid air. If suddenly plunged into liquid air the iron becomes hardened. It was found that after the first five sets of observations, which were some- 84 Profs. J. A. Fleming and J. Dewar. On the what variable, the annealed iron ring was brought into a completely stable condition, in which the curve of magnetic induction plotted in terms of magnetising force taken at the low temperature was different from that taken at 15° C. by a perfectly constant amount, the observa- tions at the low temperature always lying on one curve, and those at the higher temperature always lying closely on the other curve. In the diagram in fig. 1 the two magnetisation curves are shown, the firm line curve being the magnetisation curve at 15° C., and the dotted curve being the magnetisation curve taken at —185° C. in the liquid air. The figures in Table I are the mean values obtained from the curves plotted from the thirteen sets of closely consistent observa- tions. These curves show that the permeability of soft annealed iron is reduced when it is cooled to about 200° below zero, for the whole range of magnetic forces between zero and 25 C.G.S. units. The permeability curves for the two states are likewise similarly shown on the same chart. The maximum permeability for this iron corresponds with a magnetising force of about 2 C.Gr.S. units; the maximum permeability at the ordinary temperatures for this iron is 3400, being reduced to 2700 when the iron is cooled to the temperature of liquid air. The percentage reduction in permeability becomes less as the magnetising force is increased beyond or reduced below this critical magnetising force. These experiments were repeated, as above stated, many times very carefully with this ring of annealed soft Swedish iron, and also with a second ring of the same kind, and have invariably shown the same results, viz., that the permeability of soft annealed iron is decreased by being cooled to this low tem- perature within the range of magnetising forces from 0 to 25. It will be seen that the highest induction reached in the case of this iron is 14,500 C.Gr.S. units, corresponding to a magnetising force of 25. This iron is of very high magnetic quality, and is of the same character as that which is much used for the construction of alterna- ting current transformers in commercial use. * A series of experiments was then made with the same transformer, keeping the magnetising forces constant, but allowing the iron to rise gradually in temperature up from the temperature of liquid air to 15° C. In these experiments the transformer was embedded in a mass of paraffin wax with a platinum wire resistance thermometer also embedded in the same mass in close contact with the ring coil. The paraffin wax encasing the ring coil and thermometer having been cooled down to the temperature of liquid air by immersing it in a large bath of the liquid air, it was then lifted out and placed in a vacuum-jacketed test-tube, so as to heat up with extreme slowness, and a series of observations taken by reversing a constant magneti- sing force at intervals, and observing at the same instant the tem- perature of the ring coil as given by the platinum thermometer. Magnetic Permeability, §c., of Iron at Low Temperatures. 85 FIG. l. Magnetising force in C.G£. units. I 234567 69 10 II 12 15 14 15 16 17 IB 19 20 21 14,000 WOO 12000 11,000 10000 g 8000 D : 7,000 "Q ^ 6000 5000 4,000 3000 2000 1000 Permeability And Magnetic in Soft Iron at + 15 ° Centigrade. \ \ 0 500 1000 1,500 2,000 2,500 6,000 3,500 4,000 Scale of Permeability in C.G.S. units. 86 Profs. J. A. Fleming and J. Dewar. On the The results of these observations are given in Table II, and these observations are set out in the curve marked soft annealed iron in fiff. 2. Table II. — Variation of the Magnetic Permeability of Soft Annealed Swedish Iron with Temperature. Magnetising force = 177 C.G.S. Temperature measured in platinum degrees by standard thermo- meter PI. Temperature. Permeability. 0° 2835 - 20 2815 - 40 2770 — 60 2727 - 80 2675 — 100 2622 -120 2560 -140 2497 -160 2438 -180 2381 —200 2332 The results show that as the temperature rises up from —185° C., or — 200° on the platinum scale temperature, up to the ordinary tempera- ture, the permeability of the soft iron for the particular magnetising force selected increases perfectly uniformly, the curve of increasing permeability with temperature being nearly a straight line. In the next place, we have examined the hysteresis of the same soft iron ring at different temperatures and for different maximum induc- tions. These observations were carried out by taking a complete series of hysteresis curves with the ballistic galvanometer, gradually increasing the inductions from zero to 12,000. After the complete hysteresis curves were obtained, their areas were carefully integrated with an Amsler planimeter, and the values reduced so as to express the hysteresis loss in watts per Ib. per 100 cycles per second, and these values plotted in terms of the maximum value of the magnetic induction per square centimetre of the iron core corresponding to each particular hysteresis loss. Nothing would be gained by giving the full details of all the observations by which these hysteresis curves were obtained. They were exceedingly numerous, and the tedious nature of the ballistic observations made it a matter of pro- longed observation to secure the whole series necessary, but the final results are shown in Table III. The curve in fig. 3 represents the increase of hysteresis loss with induction, and the observations which Magnetic Permeability r, cj-c., o/ Iron at Low Temperatures. 87 Fia. 2. 3,500 Relation of Permeability to Temperature. -200-/90-I80-I70-I60-I50-I40-I50-I20-IIO-IOO-90-60-70 -60 -50-40 -50 -ZQ -10 0 Scale of Temperature (Platinum Degrees). 88 Profs. J. A. Fleming and J. Dewar. On the i $** $ / Hy a steresis anke/s bes 'oss in s t transforn oft iron, er iron) / / / / X ^ / 2000 4,000 6,000 6,000 Induction in C.G.S. units. t 0,000 12,000 Table III. — Hysteresis Loss in Soft, Annealed Swedish Iron in Watts per pound per 100 cycles per second for various maximum Induc- tions. I. At +15°C. II. At -185° C. (in liquid air). f Maximum induction. Hysteresis loss. B. W. 844 0-0397 4026 0-4957 6743 1-062 9687 2-070 11618 2-632 8593 1-545 5516 0-823 Maximum induction. B. 688 ' 3603 6185 9461 11916 Hysteresis loss. W. 0-02519 0-4246 0-949 1-907 2-658 were taken at ordinary temperatures are denoted by the small circles. The observations for hysteresis loss which were taken at the tempera- ture of liquid air are denoted by the crosses. It will be seen that substantially the circles and the crosses lie on the same curve. The results of these observations, therefore, show that there is practically no change in the hysteresis loss in soft iron by cooling it to the tern- Magnetic Permeability, Magnetising' force, 20'39. 361-0 - 20 84-0 332-5 - 40 81-0 299-5 - 60 79-0 271-5 - 80 77-0 246-5 -1<)0 74-0 220-0 -120 71-5 193-0 -140 68-5 174-3 -160 67-0 163-0 -180 66-0 153-0 -200 64-5 144-0 We propose to continue the examination of the anomalous behaviour so presented by iron in different states of hardening by examining in the same way the changes of permeability in the case of several iron rings of the same dimensions formed in the one case of soft annealed iron, and in another case of the same quality of iron hardened, and in the remaining cases using steel of known composition at different states of temper. We desire to add that in the conduct of this research we have been under great obligations to Mr. J. E. Petavel for rendering us very efficient assistance in taking the exceedingly tedious ballistic galvanometer observations, and in reducing them when taken. Magnetic Permeability, $c., of Iron at Low ; Temperatures. 95 Table VI. — Variation of Permeability with Temperature. Pianoforte Steel. Temperature measured in Platinum degrees by standard thermo- meter Pj. Permeability. Temperature. - 0° - 20 - 40 - 60 - 80 -100 -120 -140 -160 -180 -200 We propose to continue the examination of the anomalous behaviour so presented by iron in different states of hardening by examining in the same way the changes of permeability in the case of several iron rings of the same dimensions formed in the one case of soft annealed iron, and in another case of the same quality of iron hardened, and in the remaining cases using steel of known composition at different states of temper. We desire to add that in the conduct of this research we have been under great obligations to Mr. J. E. Petavel for rendering us very efficient assistance in taking the exceedingly tedious ballistic galvanometer observations, and in reducing them when taken. Magnetising force, 7*50. N Magnetising force, 20-39. 86-0 361-0 84-0 332-5 81-0 299-5 79-0 271-5 77-0 246-5 74-0 220-0 71-5 193-0 68-5 174-3 67'0 163-0 66-0 - 153-0 64-5 144-0 VOL. LX. Dr. C. Chree. Observations on Atmospheric " Observations on Atmospheric Electricity at the Kew Observa- tory." By C. CHREE, Sc.D., Superintendent. Communi- cated by Professor G. CAREY FOSTER, F.R.S. Received May 11,— Read June 4, 1896. TABLE OF CONTENTS. PAET I. The Measurement of Potential in Theory and Practice. §§ PAGE 1 Historical and descriptive , 96 2 — 3 Interpretation of electrograph records 98 4 Selection of stations , 99 5 — 6 Comparison of results at the different stations 100 7 — 9 Eatio of readings at different stations, at different times, and under different meteorological conditions 102 10 Comparison of water-dropper and portable electrometer 106 11 Defects in instruments 108 12 Checks recommended 109 PAET II. Application of Results to Theories of Atmospheric Electricity. , 13—15 Theories of Exner and of Elster and G-eitel 110 16 Method of treating Kew observations 112 17 — 19 Anticipation of some objections : want of uniformity in conditions as to wind, and cloudiness; proximity to London 112 20 Tables of results, including particulars as to potential, vapour density, humidity, sunshine, temperature, barometric pressure, and wind velocity 114 21 Analysis of preceding tables according to voltages at base station .... 123 22 Further tables, each containing analysis according to magnitude of some one meteorological element. . . . , 125 23 Discussion of possible influence of vapour density 128 24 relative humidity 128 25 sunshine 128 26 temperature 129 27 barometric pressure 129 28 wind velocity 129 29 — 30 General summary of bearing of results on theory 130 PAET I. The Measurement of Potential in Theory and Practice. § 1. An electrograph belonging to the Meteorological Office has been in operation at Kew Observatory, with interruptions, since 1861. The results obtained in the early years of its existence were dis- Electricity at the Kew Observatory. 97 cussed in 1868 by Professor Everett,* and the results obtained in 1880 were discussed in 1881 by my predecessor, Mr. Whipple.f Nearly two years ago, with the approval of the Kew Observatory Committee and the Meteorological Office, I commenced an investiga- tion intended as preliminary to a consideration of the expediency of further publication of the electrograph records. My first object was to find out whether definite quantitative measurements of potential could be derived from the electrograph curves. To aid in this investigation observations have been made at several spots near the Observatory with a portable electrometer, by White, of Glasgow, whose scale value was determined at Uni- versity College by the kind assistance of Professor Carey Foster. To render intelligible the bearing of these observations on the question, a brief description is required of the nature and position of the electrograph. J It consists essentially of a water-dropper and a quadrant electrometer. The water is held in a can, some 14 inches high and 15 inches in diameter, supported on three insulators of the Mascart pattern. From the can a tapering tube, resting on a fourth insulator, projects through a hole in a window facing the west. The end of the tube whence the water issues is 4^ feet from the west wall of the Observatory, and 10 feet above the ground. The stream of water is regulated by two taps in the long tube. From the water- dropper an insulated wire passes to the needle of the quadrant electrometer. One pair of quadrants are kept at a given positive potential, the other pair at an equal negative potential, by means of a battery of 60 cells in series whose centre is to earth. The needle suspension carries a mirror, and light reflected from it produces a curve on photographic paper which is wound round a cylinder driven by clock-work. The position of the base line answering to the earth's potential — treated as zero — is obtained by putting the electrometer needle to earth, twice at least for each curve. Of late years the value of the curve ordinates, in volts, has been obtained from time to time by connecting the electrometer needle and one terminal of the portable electrometer, and varying their joint potential by means of ^n electrophorus. Simultaneous readings are taken of the curve ordinate and the portable electrometer. If the ideal were attainable, the stream from the water-dropper should break up exactly at the end of the tube, and be always sufficiently copious to ensure the immediate picking up by the can and the electrometer needle of the potential existing in the air at the spot in question. * ' Phil. Trans.' for 1868, p. 347. t ' B. A. Keport,' vol. 51, p. 443. .J (July 28.) Some alterations have been effected since the above was written i 2 98 Dr. C. Chree. Observations on Atmospheric Interpretation of Electrograph Record. § 2. The first question is : supposing the apparatus perfect, does the electrograph supply information as to the potential anywhere except at the spot where the stream of water breaks into drops ? To answer this question, one has to consider the influence of the environ- ment, notably the proximity of a lofty building. An investigation into this point was made ten years ago by Pro- fessor Exner, of Vienna, who found the equipotential surfaces near a building much deflected from horizontally. His results indicated apparently that for practical purposes the whole building might be regarded as possessing the earth's potential. Whilst it was antici- pated that Exner's conclusions would hold good of Kew Observatory, it appeared prudent as a check to take observations with the portable electrometer, at a series of points in a vertical plane perpendicular to the west wall near the water-dropper. Observations were taken at heights of 3, 6, and 9 feet from the ground, which possesses, it may be explained, a slope away from the building. The base line, starting at the Observatory wall, terminated 57 feet away in a parallel wall 11 feet high, belonging to a much lower building. The observations were repeated on several days, but one example will suffice. The potential measurements are in volts, the distances from the Observa- tory wall in feet. Table I. Observations on November 6, 1894. Mean Distance from wall 3 6 12 18 24' 30 36 42 48 54 potential. Potential at height 3 feet .. 4 6 18 38 48 46 34 24 16 6 26 6 „ .. 8 18 40 58 88 84 76 68 52 22 56 „ 9 „ .. — 28 44 76 102 120 120 108 68 36 78 In forming the means in the last column the results at 3 feet from the Observatory wall were omitted. The readings were uncorrected for variations of potential during the interval occupied by the obser- vations. So far as they go, the results are clearly confirmatory of Exner's. They show that the influence of a tall building in pulling down the potential extends to a considerable distance. § 3. The large dependence of the electrograph records on the im- mediate environment of the water jet complicates matters, but this need not prove a serious obstacle if the conditions allow us to regard the problem as one of statical electricity, in which influencing bodies are either stationary or at a distance. On this hypothesis, simulta- neous potentials at any two neighbouring points would stand to one Electricity at the Kew Observatory. 99 another in a practically constant ratio, a function only of their geometrical coordinates. If once this ratio were determined, one could deduce the potential at either point from that observed at the other. Regarding the spot where the water jet breaks up as one of these points, and selecting for bhe other a spot sufficiently distant from the building, one could deduce the potential gradient in the open, i.e., the increase in voltage per unit of height above the ground. This point of view was apparently acted upon by Exner,* and by Elster and Geitel.j In both instances the existence of corroborative evidence is referred to, but I am not aware that particulars have been published. It would also appear that Exner and Elster and Geitel directed their attention mainly, if not exclusively, to clear quiet days. There being no limitation to the use of the Kew electrograph, it appeared advisable not to restrict the investigations to days of a special kind, or to a particular season of the year. Selection of Stations. § 4. It appeared desirable to compare the potential at more than two stations, so as to ensure a sufficient variety in the surroundings. I shall distinguish the stations selected by the letters A, B, C, D, E, F. Of these A is the flat top of a stone pillar, 3 j feet high, in the Observatory garden, about 56 yards from the Observatory; it is surrounded by a frequently mown grass lawn. B is the top of a temporary wooden stand, 6f feet high, and only 3£ feet from the west wall of the Observatory. C is the centre of a flat plank supported 3J feet above the ridge of a wooden building, situated about 100 feet to the south-west of the Observatory ; it is 18 feet above the ground. D is on the south side of a stone parapet, 2£ feet high, encircling the flat roof of the Observatory ; it is 37 feet from the ground. E is the top of a camera stand, 5-g- feet above the Observatory roof, and 17 feet to the east of the central dome. F is the top of a stand on the roof — used for testing anemometers — level with the cups of the standard anemometer, from which it is distant about 17 feet to the north; ii is 57 ft. et above the ground. The observations were taken with the portable electrometer, and, as the burning end of the fuse was at a height of some 12 to 16 inches above the base of the electrometer, an addition of, say, 1£ feet requires to be made to the altitudes of the several stations to get the height from the ground of the spot whose potential was measured. A was the only station that could be regarded as practically unin- fluenced by the neighbourhood of a building, and even in its case we * ' Wien. Sitz.,' vol. 98, 1889. f ' Wien. Sitz.,' vol. 101, p. 703, 1892. 100 Dr. C. Chree. Observations on Atmospheric have the influence of a massive stone pillar some 2J square feet in section. A calculation of the potential gradient which regards the observations at A as referring to a spot 60 inches above the ground in the open is certain to give an under- estimate. As it is impossible, however, to dispense with a support of some kind, and the presence of the observer is also a disturbing influence, no exact allowance can be made for this. There have been four principal series of observations. In the first, occupying part of November and December, 1894, observations were taken, when practicable, once a day at stations A, B, C, D, and latterly at E also. In the second series, during part of March and April, 1895, observations were usually taken about 10.30 A.M. and 4.30 P.M. at each of the stations except F. The third series, during part of June and July, 1895, closely resembled the second; and the only material difference in the fourth was the substitution of station F for station D. No observations were taken on Sundays or on Saturday afternoons. The observations were taken in a fixed order, and, thanks to the skill of the observer, Mr. E. G. Constable, a complete set of readings occupied only some seven or eight minutes. The time scale of the electrograph curves is far from open, and for this and other reasons I have judged it best not to attempt to reduce the readings with the portable electrometer to a common instant. Comparison of Results at the different Stations. § 5. I have taken A as base station, and have found the ratios borne to the individual readings there by the corresponding readings at the other stations. Let rA, rE represent corresponding readings at A and B, and let _ 1 n B/A where 2 denotes summation for a series of n observations. Then r maybe called the mean value of the ratio for the series of observations. Also let us apply the term percentage deviation of the ratio from its mean to the quantity X100, B/A in which the terms in the numerator are taken irrespective of sign. Table II gives the extreme and mean values of the ratios during each series of observations, excluding three or four occasions when negative potentials were met with. Electricity at the Kew Observatory. Table II. 101 Series of Number of observa- observa- tions* tions. rs/r^. , — — *- — ^ rclr\. rD/r^. Max. Min. Mean. Max. Min. Mean. Max. Min. Mean. I. 25 0-38 0-17 0-26 3-05 I'll 2-22 3-33 1-46 2-41 II. 45 0-54 0-16 0-29 2-32 1-40 1-78 4-52 1-46 2-28 III. 31 0-50 0-17 0-27 2-29 1-00 1-70 3-67 1-11 2-14 IY. 23 0-41 0-09 0-22 2-86 1-33 1-92 Series of Number of observa- observa- tions. tions. **fi/rA. n?/rA. Max. Min. Mean. Max. Min. Mean. I. 25 4-95 2-46 3-12 , II. 45 6-30 1-74 2-68 __ __ III. 31 4-33 1-11 2-51 IV. 23 4-73 2-05 2-87 8-34 2-71 4-53 In series I there were only twelve observations at station E. In series III the mean ratios for the higher stations are depressed by one abnormally low reading. The means in the different series vary, but the differences are too small fco warrant any positive con- clusion. They indeed suggest the possibility of the potentials at the higher stations being relatively somewhat higher in winter than in summer, but this may arise from a slight want of uniformity in the procedure followed at the different seasons. The departures of the maxima and minima in Table II from the means are considerable, but the number of instances in which the departures from the mean are large is in reality small. This will be seen by reference to Table III, which gives the percentage deviations of the ratios from their means, treating each series of observations separately. Table III. Percentage Deviations from the Means. Series of observations. fn/r* rc^. r»rA I. 14 14 15 15 II. 19 10 19 21 — III. 19 11 20 20 — IV. 28 13 — 16 20 102 Dr. C. Chree. Observations on Atmospheric The irregularity in rB/rA may be due in part to the slightly unsteady character of the stand forming station B. The potentials at B were also much the lowest, so that errors of reading were there of most importance. At the highest station, F, the variations occurring in the potential sometimes made accurate measurements difficult. § 6. To give a clearer idea of the degree of uniformity shown by Table III, I give in Table IV the extreme and mean readings at the several stations, omitting, as in Table II, occasions of negative potential. Table IV. Readings in Volts at the several Stations. Series of observations. A. Max. Min. Mean. 264 104 158 708 50 206 174 27 100 830 29 249 B. C. Max. Min. Mean. 66 22 40 215 15 58 45 6 27 115 12 50 Max. Min. Mean. 552 120 352 1320 81 364 306 27 171 1452 52 476 I. II. III. IV. Series of observations. I). E. F. r ^ Max. Min. Mean. 648 152 385 1464 122 455 354 30 210 Max. Min. Mean. 776 264 524 1688 ' 182 531 498 30 246 1785 93 662 Max. Min. Mean. 2362 122 1032 I. II. III. IV. On one exceptional day the potential at A varied from —1200 to + 1290 volts in less than forty minutes; at station F it varied from —2424 to over +4000 volts in about the same time. Constancy of Ratios during the Day. § 7. Table V gives the mean values of the ratios for the forenoon and afternoon observations, treated separately, during those days when there were readings at both 10.30 A.M. and 4.30 P.M. The days available numbered 17, 10, and 9 respectively in the second, third, and fourth series of observations. The headings " A.M." and " P.M." distinguish the forenoon and afternoon observations. In each case treated in table V the mean value of the potential for the forenoon was considerably higher than that for the afternoon. Thus, at station A the ratio of the forenoon to the afternoon mean potential — for those days only on which there were both forenoon Electricity at the Kew Observatory. Table V. Forenoon and Afternoon Ratios. 103 Series of observa- tions. »*B/A.' **C/A. f * •> **D/A. 'E/A. !5_A' A.M. P.M. A.M. P.M. A.M. P.M. A.M. P.M. A.M. P.M. II. ! 0-29 0-28 1-82 1-83 2-24 2-21 2 '62 2-57 III. j 0-24 0-33 1 '75 1 -52 2 -16 1 -87 2-57 2-17 IV. 0-22 0-23 1 -96 1 -88 — — 2 -76 2 -87 3 -99 4 -46 and afternoon observations— was T37 in series II, T23 in series IU, and 1*48 in series IV. The difference between the mean potentials at the two hours on the specified days being so large, we may reasonably suppose that if any two other hours had been selected results would have been obtained showing a degree of accordance similar to that in Table V. The degree of accordance in the case of series II is truly remarkable, and in series IV, considering the smaller number of observations, it is but little inferior. If series III stood alone, we might suspect that in the afternoon the potential fell off more at the higher stations than at the lower, and this may of course be a true phenomenon of the season, midsummer, to which that series belongs. Possible Dependence of Ratios on the Weather. § 8. It is conceivable that under one regular set of climatic condi- tions the potentials at the higher stations might relatively to the lower be either abnormally high or abnormally low. To test this point, the observations in each series Kave been divided into sets, according to the value of such a ratio as rE/rA. Attention has been confined to series II, ill, and IV, as in series I the times of observa- tion were less regular ; but the forenoon and afternoon observations in series II and III have been considered separately. Supposing the number of measures of, say, rE/rA available in any one instance to be 2n or 2n + l, the n cases in which the ratio is largest form one set, the n cases in which it is smallest the other. For each of these sets the corresponding mean values of certain meteorological elements have been calculated, the data for the indi- vidual times of observation being derived from the self-recording instruments employed in the Observatory. The figures as to aqueous vapour and humidity have been deduced from, the thermograms, with the aid of a modification of Glaisher's table, compiled by the Meteoro- logical Office. 104 Dr. C. Chree. Observations on Atmospheric By " sunshine in hours " is meant the number of hours of sunshine measured by the Campbell-Stokes recorder up to the time of observa- tion. The data under this head have been limited to the most sunny series of observations, viz., II and III. The results are exhibited in Table VI, which shows also the maxima and minima values of the meteorological elements observed during the several sets of n observations. There is in Table VI no uniform and conspicuous connexion between the value of rE/A, or rF/A, and the corresponding value of any one of the meteorological elements considered. In the case alike of barometric pressure and temperature the second mean — answering to the n lowest values of rE/A or rF/A — is higher than the first in five instances out of six. The differences between the two means are generally, however, so small that the phenomenon may be purely accidental. In the afternoon observations of series II there is a somewhat conspicuous association of a low value in rE/A with a high value of previous sunshine ; but in series III there is no trace of such a phenomenon. The question whether there may not be certain occasional types of weather, whose influence is masked in such a table as VI, which are associated with either a high or a low value of the ratio rE/A, remains, I think, open. Evidence is in my hands which leads me to believe that during a low ground fog the potential gradient as a rale is decidedly higher near the ground where the fog is thick than higher up where the fog is slight. Summary of Results at Different Stations. § 9. The conclusion I am disposed to draw, though I regard it as only a probability, is that such general phenomena as diurnal or annual variation of potential near the ground in the open may be deduced with fair accuracy by applying a constant factor to the records of a portable electrometer, employed regularly at a fixed point on the Observatory roof or near its walls. It must be remem- bered, however, that all six stations were comparatively close together, and that the equipotential surface passing through the highest station would be in the open perhaps only 14 or 15 feet above the ground. There is thus no evidence to warrant the deduction of conclusions for a spot a mile or two away or a few hundred feet above the ground. On the trustworthiness of individual results deduced by means of a constant factor, one would not, after inspecting Tables II and III, be disposed to place much reliance. This question can hardly, how- ever, be settled satisfactorily unless one have apparatus for taking the observations at the different stations absolutely simultaneously. The largest departures from the means in Tables II and III are Electricity at the Kew Observatory. 105 II ' .S'S «o a §0 rH jgrHrH rH 10 OS TO CO CO CO r— I OS O »O iO rH rH rP X (N 10 OS O rP TO X OS OS l> *"H rH rH rH rH rP O rHrP lO TO rPrP rH O rHO O — I l> TO TO rP (M t>. QO OS OS rH rpcq !N!M rHO OO t>. TP OS , OS CO X O COCO rH ib 8.9 W gQO C^ xp X OS N IO CO 1O CO rP 10 X CO X CO JOS X OS X TO rH CO X C^l J>» O W C^J O TO -t** OS rH O rP rfl t» CO 9 T1 OS X •Jl a § I Foren an afterno 106 Dr. C. Chree. Observations on Atmospheric doubtless due in great part to changes occurring whilst the observa- tions were in progress. Possible Influence of Pattern of Instrument. § 10. The conclusions in the previous paragraph refer as yet only to the portable electrometer. They can be extended to the electro- graph records only if we are able to show that a fairly uniform ratio exists between the potential obtained with the water-dropper at a fixed station and that obtained with the portable electrometer at one or other of the stations A to E. The position of the water-dropper was maintained undisturbed, barring accidents, throughout the observations. It thus suffices to compare the curve readings with the corresponding ones with the portable electrometer at station A. The curves were accordingly measured at the mean times of each set of observations. The ratios of the individual readings to those at station A were calculated, and results obtained analogous to those in Table II. It will suffice for our present object to consider the results analogous to those in Table III. Table VII. Percentage Deviations from the Means (Electrograph/Portable). Series of observations. *!. II. III. IT. Percentage deviations 28 30 35 28 The spot where the jet breaks up resembles B more closely than any other station, and shares its low potential. Further, the electro- graph curves are read to the nearest 5 volts only, so that uncertainties in the reading are even more important than with the portable, read to the nearest 1 or 2 volts, at station B. Thus, the results in Table VII are, at least, not conspicuously worse than those in Table III. As a matter of fact, the results in Table VII were, I believe, somewhat prejudiced by a variation in the water jet throughout the day (see § 11). Supposing this defect removed, the evidence points to the conclusion that the diurnal, and possibly the annual, variations got out with the water-dropper situated in the Observatory, and the portable electrometer at station A, may be expected to be in good accord, assuming the conditions under which each instrument works to be maintained uniform. Attention was also directed to the possibility of the two different patterns of instrument being differently affected by the same climatic conditions. Each series of observations — the forenoon and afternoon observations of series' II and III being treated separately — was Electricity at the Kew Observatory. 107 arranged in descending order of some one meteorological element. Sup- pose there to be 2n or 2n + I observations in the series (or half series) « the mean values of the ratios borne by the electrograph readings to the corresponding ones with the portable electrometer at station A were calculated for the first n and the last n instances separately. Supposing r1} r2, and r to denote the mean ratios for the first n, the last n, and the whole 2n (or 2n + l) observations, then i{0-i-ra)/r}xioo may be regarded as the average percentage deviation of the two groups from the mean. Table VIII gives the value of this quantity in the case of the three meteorological elements from which a differen- tial effect was most feared. Table VII T. Value of M0-i-r2)/r} x 100. Meteorological element considered. Times of observations. observations. Vapour density. Sunshine. Wind velocity. I Day -15 - 8 " 1 Forenoon Afternoon -12 + 1 + 8 4- 4 + 8 + 12 m { Forenoon Afternoon -20 -15 + 14 + 15 + 22 -16 IV Day -15 < OJ- A plus sign denotes that when the meteorological element in question was above its mean the water-dropper was more than usually effective, relative to the portable electrometer ; a minus sign implies the contrary. The evidence in the case of wind velocity is so contradictory that we can safely assume that no uniform differential action exists. In the case of the two other elements the evidence is more consistent, and it is possible that a small differential action may exist. It looks as if much moisture, when not counterbalanced by a contrary action of sunshine, tends slightly to pull down the reading of the water- dropper relatively to that of the portable electrometer. The pheno- menon, supposing it to exist, might be ascribed to a loss of efficiency in a water jet when the vapour in the air increases, and a similar loss in a flame collector during bright sunshine. But an influence at least as likely is that of moisture, during damp weather, on the insulation of the electrograph. 108 Dr. C. Chree. Observations on Atmospheric Defects in Water-dropper and Portable Electrometer. § 11. Both instruments aim at communicating the potential at a fixed point in the air to an insulated conductor by detaching from a mass in electrical connexion with the conductor a continuous succes- sion of small elements. It is at least doubtful whether either instru- ment can ever fully accomplish its object. If the object were so far accomplished that a constant fraction of the true potential were recorded, the deficiency of this fraction from unity would hardly be of primary importance in dealing with diurnal or annual variation ; but if the fraction has itself a diurnal or secular variation it is a very different matter. In the water-dropper a uniform state of insulation of the water- can, electrometer needle, and connecting wire is not easy. Absolute insulation, when a voltage runs up to hundreds, is a somewhat ideal state of perfection. When the insulation is indifferent, the recorded may fall far below the true potential. The water jet, so to speak, is running up the potential, leakage from the can, wire, &c., running it down. The resultanfc effect depends on a variety of things, e.g., the rate at which the air potential is changing and the supply of water particles. Unless the potential is unusually steady, and the insula- tion exceptionally good, one may expect higher potential records with a copious jet than with a restricted one. In the portable electrometer there' is similarly some ground for expecting the potential recorded to be influenced by the rate of com- bustion of the fuse. The uniformity of the disintegrating mass may also be of import- ance. With a water-dropper there ought not to be much uncertainty on this ground, but as electrometer fuses are articles of commerce uniformity in their material and condition is less easily ensured. There is a final source of uncertainty common to the two instru- ments as commonly used. With the water-dropper the spot where the jet breaks up is apt to be slightly influenced by variations in the water pressure. When the issuing jet makes as usual an angle with the wall of a building, the consequences, as appears from Table II, are likely to be appreciable. It was a recognition of this fact that led to the taking of the observations with the portable electrometer at two nearly fixed hours, the afternoon one when the can was nearly full, the forenoon one when it was about half empty. The corresponding defect with the portable electrometer is the burning down of the fuse. When the fuse is used in a vertical position, the height of the spot whose potential is being measured diminishes as the fuse burns, and with the height the potential falls off. No direct comparison of the readings of the two instruments at one Electricity, at the Kew Observatory. 109 and the same spot was attempted during the observations, as it seemed undesirable to interrupt the continuity of the electrograph records. All that § 10 shows is that during any one series of observa- tions the fractions of the true potential picked up by the two instru- ments stood to one another in a fairly constant ratio. The presump- tion, certainly, is that neither fraction altered much throughout the few weeks covered by any one of the four series of observations. It is, however, I regret to say, perfectly certain, from the data on which § 10 is based, that one at least of the two instruments varied very considerably in the course of a year and possessed a,n appre- ciable diurnal variation. § 12. On the discovery of these defects it became not only justifi- able but necessary to subject the water-dropper itself to direct experiments. These have led to my proposing certain alterations which are now in process of execution. They aim at bringing the water-can and electrometer close together, and at maintaining a more uniform water pressure than heretofore. It appears also desirable to check the working of the apparatus in some way jinvolving the arrival at exact numerical results. The following operations A, B, C will, it is hoped, prove sufficient. The operation C need not be performed so frequently as A or B. A. Charge the quadrant electrometer needle to a high potential, and observe the rate of leakage over a fixed range by timing the motion of the spot of light across a scale with — (1) the wire connexion to the water-can complete, but the jet not flowing ; (2) the wire connexion broken at the can ; (3) the wire connexion broken at the electrometer. B. As a substitute, or as subsidiary to A. Connect a portable electrometer to the water-can, and, with the jet flowing, observe the potential recorded by the portable, when — (1) the can is connected as usual to the quadrant electrometer; (2) the connexion is broken at the quadrant electrometer ; (3) the connexion is broken at the can. C. Take a sufficient number of observations at a fixed station out- side with a portable electrometer, at or near two fixed hours a day, so chosen that at one hour the can is almost full, whilst at the other it is at least half empty. The use to be made of the results is obvious. I should also recommend any one using a portable electrometer to test its scale value from time to time by comparison with an absolute electrometer or a large battery of constant cells. It is well to lay in a new stock of fuses before exhausting one's supply, and to compare the old and new fuses by taking observations in rapid succession with samples of the two at a fixed station. 110 Dr. C. Chree. Observations on Atmospheric PART IT. Application of Results to Theories of Atmospheric Electricity. § 13. It seemed desirable to consider what bearing the special experiments might have on the general facts and theories of atmo- spheric electricity. In this investigation special attention has been given to the possible influence of aqueous vapour on electrical potential, on account of the important researches of Exner, and of Elster and Geitel. Theories of Exner and of Elster and Geitel. § 14. Exner has advanced the view that the potential gradient in the open, dV/dn in his notation, and the density q0 of aqueous vapour simultaneously present in the atmosphere, are connected by a formula dV/dn = constant -4- (l-f &gr0), where k is apparently a constant, the same at all places and at all seasons of the year. Exner, I believe, limited his observations, and presumably the application of the formula, to days comparatively quiet and free from clouds. To test the formnla he arranged his observations in groups, according to the amount of vapour present, and compared the mea.n vapour density — measured in grams per cubic metre — with the mean potential gradient, measured in volts per metre of height above the ground. In the ' Wien. Sitz.,' Bd. 99, p. 618, he gives a table including results from Vienna, Wolfenbiittel, St. Gilgen, and India, in which the vapour densities vary from 1*7 to 23'5. The table unquestionably shows a diminishing mean potential gradient accompanying an increasing mean vapour density. For values of q0 from 12'4 and upwards, however, — including all the Indian and most of the St. Gilgen observations — the change in dV/dn is somewhat small and irregular, An earlier, and somewhat similar, but less extensive table by Exner will be found on p. 434 of ' Wien. Sitz.,' Bd. 96. For information as to Elster and Geitel's work I am mainly indebted to a long paper by them in the ' Wien. Sitz.,' Bd. 101, p. 703, 1892. During 1888-91 they took an extensive series of observations on quiet days at Wolfenbiittel. If I follow their explanations, they took eye observations some ten times a day with an electrometer, in which flame from a lamp acts as collector, and deduced the mean value of the potential gradient dV/dn for the day. They compare these potential gradients grouped according to the value of the vapour density with Exner's formula, taken to be (dV/dn) (in volts per metre) = 1410/(l-fl-15g0), Electricity at the Keiv Observatory. Ill where q0 is measured as above in grams per metre. They give an abstract of the results on their p. 742, in the shape of a table which I reproduce. Elster and Geitel's Table III (loc. cit., p. 742). ?o = 1-6 1-9 2-5 3-7 4-6 5-6 6-5 7-6 8-4 9-4 10-6 13-5 dVldn (observed) = 502 430 400 318 252 137 184 148 112 115 118 121 „ (calculated) = 496 442 364 268 224 189 166 145 133 119 107 85 It would appear that Elster and Geitel, like Exner, found large departures from the mean dV/dn of a group amongst its individual members. § 15. Elster and Geitel next proceed to investigate a possible con- nexion between the potential gradient and the intensity of that species of solar radiation which dissipates a negative charge on an insulated sphere of polished zinc. If I understand them rightly, they measured the mid-day intensity J of this radiation, and compared the potential gradient with several formulae in which the variable was either J or J/, where / is a " Beleuchtungsf actor," equal apparently to (possible hours of sunshine) /12. Taking a formula dV/dn = 110 + 360a~J, where log a = O'OIOO, they give the following comparison of the results of observation and theory : — T 2'9 5-8 9-1 21-4 58'8 77-1 113-7 121-9 181-3 194-5 268-4 dVldn (observed) = 447 430 368 325 198 181 138 126 120 106 102 „ (calculated) = 447 425 402 330 203 171 136 132 116 114 111 The agreement seems better than in the case of Exner's formula, and Elster and Geitel seem strongly inclined to regard ultra-violet radiation as the direct cause of variations of potential on normal qniet clear days. They consider apparently that there are only two defective links in the chain of evidence, viz. : — (1) absolute proof that the earth is electrified negatively; (2) proof that there is a sufficient supply at the earth's surface of materials susceptible to the influence of ultra-violet light. There are, of course, numerous other theories of atmospheric electricity, but none, so far as I know, admits of numerical com- parison with observation. VOL. LX. 112 Dr. C. Chree. Observations on Atmospheric Method of Treating Kew Observations. § 16. In discussing the Kew observations 1 have in general em- ployed a method differing from the grouping system of Exner and Elster and Geitel, and have also treated the several series separately. It is clear from data mentioned by Exner that the potential gradients for individual members of his groups varied in some instances largely from the mean ; and it was soon obvious that the same phenomenon would present itself if any similar treatment were applied to the Kew results. This is undesirable, because by varying the limits of the groups the accordance of the results with a par- ticular formula may be much improved, or the reverse. However impartially, so to speak, the lines may be drawn, there is undeniably a risk of introducing some fictitious result ; and no critic can feel that he is in a position to judge of the results until he has examined for himself the circumstances of the grouping, a labour he naturally shrinks from. Again a wide range of such an element as vapour density can be obtained at a particular place only by combining results from all seasons of the year. This brings us to a second question. Electrical potential gradient has like vapour density, sun- shine, and temperature, a large annual variation, only, unlike these elements, it is highest in winter. It is thus obvious that when obser- vations from all seasons of the year are treated promiscuously, there is almost sure to be a marked association of high potential with low vapour density, little sunshine, and low temperature ; and a judi- ciously selected formula which makes potential diminish as any one of these elements increases is certain to show some approach to agreement with observation. It is thus desirable to compare together observations from a limited portion of a year, or, even better, from the same season of a series of years. Similar considerations show an advantage in treating separately results from different hours of the day. The isolation of particular seasons and hours has the dis- advantage of reducing the number of observations compared together. This is, however, partly compensated for by the greater homogeneity, of the material. It also enables one in some cases to compare readily the mean potential gradients which answer at different seasons or hours to like values of some one meteorological element (see § 23). Anticipation of some Criticisms. § 17. The Kew observations were not limited to quiet, compara- tively cloudless days, in the same way as the observations of Exner and Elster and Geitel seem to have been. It may thus not unlikely be supposed that the Kew results are affected by a variety of dis- turbing causes, which diminish their intrinsic value and their suit- Electricity at the Kew Observatory. 113; ability for comparison with other results free from these extraneous effects. As to intrinsic value, there are, at least in England, seasons of the year when a nearly cloudless day is exceptional. For instance, during November and December, 1894, in ten days out of eighteen, on which observations were taken a little before noon, no bright sunshine was recorded up to the hour of observation. At such a season, if one- confined one's attention to nearly cloudless days, hardly any data would be obtained, and they might not unreasonably be regarded as abnormal. As to the disturbing action of clouds, this is no doubt in some cases very large ; but with clouds of this character the influence may be considerable when they cover only a small fraction of the sky, and probably, in some cases, even when they are below the horizon. Thus on one occasion at Kew, when part of the sky was covered by a thundercloud — so distant that only one or two faint lightning flashes were detected — sudden changes of potential of thousands of volts from negative to positive and back again were observed on the roofr whilst the sun shone at intervals. The sudden alternations of potential doubtless accompanied flashes of lightning, but no rain was falling anywhere near, and possibly an observer a few miles away might have regarded the day as an ideal quiet one. Again, there are other forms of clouds whose influence seems not unlikely to be much less than that of invisible vapour in motion nearer the ground. The mere interception of sunlight by cirrus clouds or detached masses of cumulus, if we may judge from some few experiments at Kew, has little if any effect. It should also be borne in mind that wind velocity and amount of cloud must both have varied appreciably from day to day, and even throughout the individual days of Exner's experiments. Some one — I forget who — defined a "quiet " day as one in which the flame of Exner's electrometer was not blown out. All the days of the Kew observations satisfied, of course, this definition, if one is allowed to- substitute the portable electrometer for Exner's, yet on one occasion the anemometer was recording a mean velocity of forty miles an hour. If aqueous vapour, as Exner supposes, is the sole, or even the- dominant, agent in producing changes in potential, its activity can- hardly be confined to days when there is little cloud, and the wind is low. § 18. As regards Elster and Geitel's theory, the data available for criticism are, I admit, defective, inasmuch as no measurements are taken at Kew of the dissipative effect of sunlight on negative elec- tricity. I presume, however, that bright sunshine — such as the Campbell- Stokes instrument records — always possesses this power, K 2 114 Dr. C. Chree, Observations on Atmospheric though doubtless in very variable degrees at different seasons. Solar radiation occurring after an observation is taken, clearly cannot affect it. Thus the data got out as to the amount of bright sunshine recorded prior to the observations must, I think, bear fairly directly 011 Elster and Geitel's theory. If it be true, the potential gradient must, I think, fall conspicuously as the number of hours of previous sunshine increases. § 19. An objection of a different kind is the proximity of the Kew Observatory to London. This objection has already been urged against Greenwich by investigators* whose theories do not harmonise with the results obtained there. A weekly period exists, they say, in the Greenwich electrograph curves, and this, they assume, can arise only from a weekly fluctuation in the amount of smoke, due to our insular habits of keeping Sunday. If, for a moment, we suppose the phenomenon and explanation both true — a pretty large assump- tion— there seems a wide step to the conclusion that results so affected are useless. I do not myself see that they need lead to •erroneous conclusions, unless one is dealing with a cycle whose period is seven days, or a multiple thereof, which a lunation, for instance, is not. In the present instance I would point out that the prevailing winds during each one of the series of observations were from directions included between N.N.W. and S., and that as Kew Observatory is some miles to the west of London, while the manufacturing districts are mainly in the east, it is difficult to see how London smoke could affect the results. The Observatory, I should add, is situated in a large open park to the immediate west of the extensive Kew Gardens. Even if the prevailing winds had been easterly, I question whether smoke would have exerted an appreciable influence. The analysis above mentioned of the electrograph results for 1880, by the late Mr. Whipple, seems to show that if any relation existed then between electric potential and wind direction, it varied with the season of the year; this would hardly have occurred if smoke present in east winds had an appreciable effect. Tables of Results. § 20. In discussing the observations, I have decided to commence vby incorporating the actual details in a series of tables. This will enable any one to judge for himself whether the conclusions finally arrived at are in accordance with the facts. The first eight tables give full particulars of the results. The arrangement is not * See pp. 42—43 of offprint of paper by Ekholm and Arrhenius in ' Bihang till i. Svenska Vet.- At ad. Handlingar,' Band 19, Afd. 1, No. 8, Stockholm, 1894. Electricity at the Kew Observatory. 115- 'U 1 li II 11 « s s g.2 03 f£'g 00 X O O •£ 11 X CO OS X l> OS X , ig &-§ j Ol> X TP (N t> CO CO ''T1 X tH C<1 •*?! O oT l& CO »0 -^ t> CO CD 0 lO 1 ' "H 1 | 1 1 I 1 i 8 00 C 1 •^ CO "# CD O O (N 1 -2 ^ I ^illsasl 3 i 9 Electricity at the Kew Observatory. 117 118 Dr. C. Chree. Observations on Atmospheric I ™ Oi 1 1 J^ I I g Si li 11 II &$ s"n o S S I OOrHC rH /• o T-J ?o ^ I ^ CO XO Oi iM O POlM COrH L^ 1 2 oo * Electricity at the Kew Observatory. 119 II o 3 W 2 s-§ i< O rH iH r-l O iH i-H i iH • rH O O -t>* CO CO I CO ° l^ f> L^^HrHM^ Electricity at the Kew Observatory. 121 -p pi I I I > io~ M 8 I-H O 1 s 21 I I o O I 5 H5 I S 11 11 1 * iHOOrHOOOr-l U s 'S OOOrHOlNOOOiHiHOO afh8 § rH i—i i— 1 iH L^§§ 2 12:2 Dr. C. Chree. Observations on Atmospheric I § I •8-S ii & i s 0 < ^f- OiOStNlO Electricity at the Kew Observatory. 123 chronological, but in descending order of the voltages observed at station A. In all the tables the humidity of saturation is taken as 100. § 21. I have divided the results in each of the previous eight tables into two groups, according to the order of voltages at station A. The two groups are equal in number of constituents, except as regards Table XV, where, following a marked line separating the voltages, I have included six in the first group, and seven in the second ; and Table X, where I have included four in the first group, and three in the second. In tlie last-mentioned case, the best line of demarcation is doubtful, and, on account of this, and the small number of constituents, little weight can be attached to the results. It may seem arbitrary to determine the groups by reference "to station A exclusively. It is, however, the station least influenced by buildings, and best fitted for accurate readings, while B is the worst. Also it will be seen that if one had adopted either C, D, or E as the standard station, or had taken a mean from all the stations, whilst the order of the constituents would in some tables have been considerably affected, the groups themselves would have suf- fered little or no change. Table XVII gives the mean potentials at station A for each group in the several series of observations, with the corresponding mean values of the meteorological elements. 124 Dr. C. Chree. Observations on Atmospheric 00 •* rf< CD CO 00 "* XO (M s co o os gs oo co ON x OS O COrH CO^ 1O1> ~o <§ II S O rH OO OfH OrH §rHrH XOVO (MCO -^CO ^rHOO OSOO >HOS rHOO m * * * ' ' * * rHrHO OO rHO rHO II 1*1 *-% gCOO COI> rHOS 1OOO § 00 00 CO J>- OT C^ ^1 iO O -t^ 00 rH rfl rff rp O CO CO £? rH CO CO t* rHOO OCO (NlO O(M SI *£% SS OS CO OS CO OS O 00 00 *> CO 1O CO CD CO OS CS •$ 05 CO Oi ^ CO CO OS O rH «O 1> rH CO CO ss IO CO II 11 ill 02 O "^ CO *>• ^^O rHC. 00^ CQ co ^w cqco «H CO (M rH rH I 2^ Electricity at the Kew Observatory. 125 § 22. A discussion might be based on the previous nine tables alone. Partly, however, to satisfy those who prefer a grouping system like that of Exner, I add further tables, in which the observations are arranged in groups, according to the magnitude of some one meteoro- logical element. In dealing with vapour density, barometric pressure and wind velocity, separating lines have been drawn at fixed values of the element considered. In the case of vapour density the limits required to be altered with the season. In the case of barometric pressure and wind velocity it was deemed sufficient to draw only one separating line, which answered, it will be seen, very nearly to the mean value. In dealing with sunshine and temperature,* the division has been into equal, or as nearly equal, groups as possible. After the tables follows a discussion, which embraces Tables IX to XVII as well. Table XVIII. Arrangement according to Vapour Density. Forenoon. Afternoon. Series Vapour Number Mean ^ Mean Number Mean Mean of obser- vations. grams per metre. of obser- vations. vapour density. potential at A. of obser- vations. vapour density. potential at A. , >6 7 7-71 191 4 6-61 141 I <6 11 5-27 147 3 5'13 143 r >7 8 8-12 185 6 8-35 140 II < 7 to 6 8 6-39 170 6 6-45 168 I <6 8 5-40 333 8 5-46 219 r >10 4 12-30 104 6 11-11 92 III \ 10 to 9 4 9-39 117 5 9-66 122 I <9 8 8-12 93 3 7-92 96 r >6 6 7-58 266 4 7-38 286 iv t <6 7 4-83 279 6 5-37 171 * The n + 1th constituent in the forenoon observation of Series IV is omitted, as it was doubtful whether to group it with the first n or last n. 126 Dr. C. Chree. Observations on Atmospheric Table XIX. Arrangement according to Hours of Sunshine. Series of obser- vations. Sunshine. Forenoon. Afternoon. Num ber of obser- vations. Mean sunshine hours. Mean potential at A. Number of obser- vations. Mean sunshine hours. Mean )otential at A. <{ Most Least 8 10 1-94 0 158 169 4 3 1-03 0 154 124 »{ Most Least 12 12 1-90 0-14 256 203 10 10 6-09 0-99 151 209 m{ Most Least 8 8 4-62 0'70 93 111 7 7 9-50 3-61 101 106 irf Most Least 7 6 1-06 0 237 315 5 5 5-60 1-56 168 266 Table XX. Arrangement according to Temperature. Series of obser- vations. Tempera- ture. Forenoon. Afternoon. Number of obser- vations. Mean tempera- ture. Mean potential at A. Number of obser- vations. Mean tempera- ture. Mean potential at A. i { Highest Lowest 9 9 47-2 39-9 170 159 4 3 45-7 37-1 154 124 n { Highest Lowest 12 12 51-4 43-5 191 267 10 10 56-1 46-4 145 216 m { Highest Lowest 8 8 69-6 63-6 95 109 7 7 74(6 68'0 81 125 XV { Highest Lowest 6 6 48-9 36-8 266 297 5 5 49*4 42-7 265 169 Electricity at the Kew Observatory. Table XXI. Arrangement according to Barometric Pressure. 127 Forenoon. Afternoon. TJ 4- ' of obser- vation. pressure, in inches. Number of obser- Mean pres- Mean potential Number of obser- Mean pres- Mean potential vations. sure. at A. vations. sure. at A. 1 { >30 <30 12 6 30-20 29-69 178 136 4 3 30-21 29-71 119 172 r >30 9 30-20 279 6 30-20 184 1 <30 15 29-64 199 14 29 -58 179 m { >30 <30 8 8 30-23 29 -81 108 96 6 8 30-27 29-75 103 104 TV •[ >30 6 30-35 315 5 30-33 233 <30 7 29-69 238 5 29-64 201 All >30 35 30-23 212 21 30-25 160 com- •< bined L <30 36 29-69 173 30 29-65 162 Table XXII. Arrangement according to Wind Velocity. Forenoon. Afternoon. Series of obser- vations. Wind velocity, miles per hour. Number of obser- vations. Mean velocity. Mean poten- tial. Number of obser- vations. Mean velocity. Mean poten- tial. T r 10 or > 10 13 16-6 174 3 16-7 142 1 1 < 10 5 4-2 140 4 3/2 ]41 r 10 or > 10 11 17-7 159 13 17-8 179 1 < 10 13 5-1 289 7 5-9 183 r 10 or > 10 6 15-0 85 11 15-6 101 III | < 10 10 6-3 112 3 5-3 111 r 10 or > 10 5 16-2 155 1 14-0 197 IV | < 10 8 4-4 347 9 6-8 219 All com- 4 bined *• 10 or > 10 < 10 35 36 16-6 5-1 151 232 28 23 16-7 5-7 145 180 VOL. LX. 128 Dr. C. Chree. Observations on Atmospheric Vapour Density. § 23. In Tables XVII and XVIII the forenoon observations of series IV, and both forenoon and afternoon observations of series II, support Exner's theory to a certain extent, inasmuch as they decidedly, on the whole, associate higher potential with lower vapour density. The forenoon observations, however, of series I lead in both tables to exactly the opposite result. Also in Table XVII, in five cases oat of eight, the higher potential is associated with the higher vapour density. In some instances, e.g., the afternoon observations of series III and IV, Tables XVII and XVIII lead to diametrically opposite conclusions. The following are instances of corresponding means of vapour density and potential, culled from the several tables. In Table XVIII, 8'12 occurs with both 185 and 93 ; in Table XT, 6-64 with 229 ; in Table XII, 6'62 with 180 ; in Table XVII, 6'57 with 204, 6'35 with 342, and 6'31 with 237 ; in Table XVIII, 6'6l with 141, 6-45 with 168, and 6'39 with 170. Again, in Tables XV and XVI we find 6'10 associated with 273, and 6'17 with 217. Lastly, in Table XVIII we have the following combinations, 5'46 with 219, 5-40 with 333, 5'37 with 171, and 5'27 with 147. In the face of such results, ifc seems difficult to believe in any intimate and uniform connexion whatsoever between potential gradient and vapour density. Relative Humidity. § 24. No special table is devoted to this. In Table XVII no less than six sub-cases out of eight associate higher relative humidity with higher potential. It will be noticed, however, that in three out of the six sub-cases referred to the differences between the mean humidities answering to the two groups are smaller than in either of the two sub -cases which associate higher relative humidity with lower potential. With the exception of the forenoon observations of series I, and the afternoon observations of series III, the differences between the mean relative humidities in the two groups are very small. Thus, on the whole, the evidence in favour of any distinct association of higher relative humidity and higher potential is insuffi- cient. Sunshine. § 25. There is in both the Tables XVII and XIX a balance of evidence in favour of a connexion of low potential with long previous sunshine. Out of eight sub-cases in each table, five favour this con- nexion in Table XVII, and six in Table XIX. The only sub- case in which the tables agree in associating higher potential with longer- previous sunshine is the afternoon observations of series I, and, for Electricity at the Kew Observatory. reasons already mentioned, this is not an important exception. There is thus a certain amount of general support to Elster and Geitel's- theory. An examination, however, of numerical details does not seem favourable to any such intimate connexion between sunshine and potential, as the formula suggested by them would imply. Taking, for instance, Table XIX, we notice in series III that, in the afternoon, a mean potential of 106, answering to a mean of 3*6" hours' sunshine, falls only to 101 when the hours of sunshine rise to 9'5. Again, in the forenoon observations of the same series, the mean hours of previous sunshine increase fully six times, while the poten- tial falls only from 111 to 93. The afternoon observations of series II are a striking illustration of the diverse conclusions to which the different methods adopted in Tables XVII and XIX may lead. Temperature. § 26. The forenoon observations of series IV, and both forenoon and afternoon observations of series II associate high potential with low temperature in both Tables XVII and XX ; and the balance of evidence is unquestionably in this direction. The only sub-case in which the two tables agree in associating higher potential with higher temperature is the afternoon observations of series I, which, as already explained, is the least important of the eight instances. On the whole, the evidence in favour of a connexion of high poten- tial with low temperature is just about as strong as that in favour of a connexion of high potential with little previous sunshine. Barometric Pressure. § 27. Higher potential is associated with higher pressure in the forenoon observations of each of the four series both in Tables XVII and XXI. In the afternoon observations, however, the higher- potential is associated with the lower pressure in three cases out of four in Table XVII, and in two cases out of four in Table XXI. The phenomenon, in short, is an apparently clear association of high potential and high barometric pressure in the forenoon, and an appa- rent absence of any connexion in the afternoon. Wind Velocity. § 28. A somewhat striking similarity exists here to the phenomena observed in the case of barometric pressure. In both the Tables XVII and XXII there is in the forenoon results a conspicuous association of high potential with low wind velocity. In Table XXII, it is true, series I observations form an exception, L 2 130 Dr. C. Chree. Observations on Atmospheric but it is rather apparent than real. For if, instead often, we adopt eleven miles an hour as limiting value for the velocity, we get in that instance two equal groups with the following results : — Group. Mean velocity. Mean potential. 1st 19-6 153 2nd 6-8 175 Higher potential is here associated with lower velocity, and, as the groups are equal, the result is presumably a fairer representation of the facts than that afforded by Table XXII. Whilst the association of high potential with low wind velocity in the forenoon seems thus conspicuous, there is in the afternoon no certain evidence of any such connexion. Thus, in Table XVII, higher potential is associated as often with higher as with lower velocity ; and in Table XXII, whilst higher potential is associated with lower velocity in three sub- cases out of four, the differences between the mean potentials for the first and second groups are small. In series III observations the difference is also very uncertain. If, for instance, we divide these observations into two equal groups, by taking 15 as separating value for the velocity, we obtain for each group identically the same mean voltage, 103, though the mean velocities for the two groups are respectively 18'7 and 8*1. In Table XXII the figures obtained by combining all four series of observations, afford an excellent example of what may happen when results, from all seasons of the year, are treated promiscuously. The individual series, as we have seen, show no clear association of high potential with low velocity in the afternoon observations, but, when the four series are combined, such an association seems conspicuous. The phenomenon, in reality, is mainly due to the comparatively large number of instances in which the velocity happened to be high •during the season when the potential was at its minimum. General Summary of bearing of Results on Theory. § 29. A comparatively small number of observations may be suffi- cient to disclose defects in an existing physical theory, and yet be inadequate to warrant the promulgation of a positive opinion as to the true theory. This is the most satisfactory point of view from which to regard the facts presented here. They are, in my opinion, sufficient to show the incompleteness of any theory which assumes simultaneous values of potential and any single meteorological element to be so intimately connected that the value of the one can be deduced, as a rule, from that of the other without taking into account other important influences. On the other hand, they are not sufficiently varied to justify the conclusion that the connexions Electricity at the Kew Observatory. 131 traced in §§ 25 to 28 between low potential and long previous sun- shine, high temperature, low barometric pressure, and high wind velocity constitute the normal state of matters at every station, irrespective of the hour or the season. Provisionally 1 should prefer to regard these associations as possibly accidental, even at Kew, but believe they indicate the lines on which more exhaustive inquiries might profitably proceed. Another possibility indicated by these associations, viz., that the potential tends to be higher during anticyclonic than during cyclonic weather seems also worthy of attention. An attempt was indeed made in the present instance to check this conclusion directly by reference to the weather reports of the Meteorological Office. The published data relate, however, to 8 A.M. and 6 P.M. ; so that, on a considerable number of occasions the nature of the isobars at the hours of the observations was uncertain. Taking the remaining instances, I calculated the mean potential for the cyclonic and anti- cyclonic conditions separately for each one of the four series, treating the forenoon and afternoon observations apart, except in the case of the first series. In live cases out of the seven thus presented, the mean potential for the anticyclonic group exceeded that for the cyclonic. There is thus something to be said for the hypothesis. It should be mentioned, however, that individual occurrences of high potential in cyclonic weather and of low potential in anticyclonic weather were not infrequent. § 30. The results of the present inquiry are, I believe, irreconcile- able with Exner's theory, in so far as it connects simultaneous individual values of potential and vapour density. The question remains open whether the annual variations of potential and vapour density may not be related through a formula of Exner's type — where A and B are constants for a given station, dV/dn and q0 representing monthly means of potential gradient and vapour density near the ground. Whilst the data available are far too limited for drawing a final conclusion, I think it worth while to add in Table XXIII a compari- son of the results at station A — regarded as 60 inches above the ground — with those deduced from Elster and Geitel's special form of the equation dV/dn = The figures are the arithmetic means of the values for the forenoon and the afternoon hours of observation. Observed. Calculated. 153 269 205 249 103 176 245 267 132 Observations on Atmospheric Electricity at Kew. Table XXIII. Potential at Station A. Series of observations. I. II. III. IV. The density of aqueous vapour is a quantity having but a small diurnal variation,* and it would appear, from a table published by General Sabinef that the calculated mean potential for the day — taken as the mean of the calculated values for the 24 hours — would differ but little from that answering to only the two times, 10.30 A.M. and 4.30 P.M. Thus the calculated values in Table XXIII may be regarded as close approximations to the true calculated means for the seasons of the four observations. On the other hand, according to the table of diurnal variation of potential in the paper by Mr. Whipple already referred to, the true means obtained from observations at every hour of the day might be expected to be on an average some 10 per cent, higher than the observed values in Table XXIII. It ought, further, to be remembered that, as explained in § 4, the potential at station A must fall short of the true potential at a point 60 inches above the ground in the open, also the fraction of the existing potential picked up by the portable electrometer may be appreciably less than unity. Thus the fact that the calculated values in Table XXIII are so decidedly larger on the average than the observed is perhaps rather in favour of the formula than otherwise. If we may judge, however, from the few data in the table, there seems some ground for the suspicion that the formula will prove to give too narrow a range. Before concluding, I have much pleasure in acknowledging the ready and valuable help I have received from Mr. E. Gr. Constable, Senior Assistant at the Kew Observatory. Mr. Constable took all the electrical observations and the measurements of the meteoro- logical curves, and gave me in addition much useful information derived from his long experience of the working of the electrograph and portable electrometer. * A fact difficult to reconcile with the general form of Exner's theory, f ' Roy. Soc. Proc.,' vol. 18, 1869, p. 8. On the unknown Lines in the Spectra of certain Minerals. 133 •• On the unknown Lines observed in the Spectra of certain Minerals." By J. NORMAN LOCKYER, C.B., F.R.S. Received May 16,— Read June 4, 1896. In the first note of the series " On the New Gases ohtained from Uraninite," by the distillation method, I remarked* " I have already obtained evidence that the method I have indicated may ultimately provide us with other new gases, the lines of which are also associated with those of the chromosphere." In a subsequent paper " On the Gases obtained from the Mineral Eliasite," I gave a list of several lines unknown to me, and suggested that they might indicate the existence of a new gas or gases in that mineral, and I addedf " Although the evidence in favour of a new gas is already very strong, no final verdict can be given until the spectra of all the known gases, including argon, have been photo- graphed at atmospheric pressure, and the lines tabulated. This part of the inquiry is well in hand." The inquiry above referred to has now been completed and in the following manner : — Photographs were taken of the spectra at atmospheric pressure of nitrogen, oxygen, chlorine, carbonic anhydride, coal gas, sulphuric anhydride, phosphoretted hydrogen, and argon, these being the gases which, from the experience thus far acquired, are likely to be associated with those given off by minerals. In addition to these, the lines of mercury, potassium, and platinum, were also photo- graphed. The lines of platinum are always present in the spectra for the reason that the spark is passed between platinum poles, while the lines of mercury or potassium frequently appear according as the gases are collected over mercury or potash. The spectroscope employed has a collimator and camera with object glasses of 3 in. aperture, and focal lengths of 5 ft. and 19 in. respectively. Two prisms of 60° were used, giving a length of spectrum of about 1*75 in. between K and D. In order to facilitate the reduction of the photographs, the solar spectrum was photographed under exactly similar instrumental con- ditions. Micrometric measures were made of H and K, and other well-known lines throughout the spectrum, and by means of these and Rowland's wave-lengths, a curve was carefully constructed. It may be incidentally mentioned that in the photographs of the spectra of gases at atmospheric pressure, H and K are generally present as pole lines, being probably due to an impurity of calcium in the platinum poles. * ' Eoy. Soc. Proc.,' vol. 58, p. 70. t ' Roy. Soc. Proc.,' vol. 59, p. 3. 134 Mr. J. Norman Lockyer. On the unknown In each case in which K was present, the micrometer scale was set to the reading for this line, and the photograph to be reduced then adjusted until the K line was under the cross-wires of the micro- meter. Each line in the spectrum was then in turn brought under the cross-wires, and the micrometer readings noted. The corre- sponding wave-lengths were then read off from the curve, and in this way, lists of the wave-lengths of the lines in the various spectra were compiled. These lists were then all thrown together into one table, giving the wave-lengths and intensities of all the lines recorded, and the spectra in which they appear. For the wave-lengths thus obtained no greater accuracy than one indicated by four figures is claimed. It was my intention in the first instance to give five figures from the more elaborate tables of some of the elements given by other observers, but this had to be abandoned in consequence of the considerable variations found in the tables between the results as given by different observers. First, as regards the gas from eliasite. The following list gives the lines obtained in the complete inquiry after the lines due to the old gases have been eliminated. It should be stated, however, that several of the lines have wave-lengths very near those of the old gases ; these have been retained when the more intense lines of the old gases are absent from the spectra. These cases are pointed out in the table. In the case of some of the lines in the visible part of the spectrum, more accurate wave-lengths have been recorded "by means of a four prism Steinheil spectroscope. These lines are indicated by (s). Attempts have been made to concentrate the eliasite gas by the process of sparking with oxygen over potash, but the quantity of gas remaining is so small, and so largely admixed with helium and argon, that a new research, using very much more material, is essential. It should be remarked that the list of lines which have been observed and photographed in the spectrum of the gases from eliasite represents the results of several experiments which have been made with different samples of the mineral. Some of the lines have only been seen once, while others have been noted several times. This suggests that the origins of the lines are very diverse, and it seems probable that some constituents of the mixture of gases obtained are absorbed by the potash in the process of sparking. Next, with regard to the other minerals already examined. As it is impossible for me to go on with this research for the next few months, it seems desirable, in the interest of other workers, to give at the same time a complete list of the unknown lines, so far as the observations have at present gone, indicating their mineral origins, and whether or not lines nearly coincident in position have been observed in any celestial body. Lines observed in the Spectra of certain Minerals. "1 I bfi £ *r --* a G t~'o ** co co co co co co co co ^o co co co co °! 2 S W) S 'S^ I 3 lol I d 5 S r? 136 Mr. J. Norman Lockyer. On the unknown i 11 O 0 ' r O ^ 4115 * Orionis 4128 '6 a Cygni, Ei 4131'4aCjgni,Ei Is isf 2.Sg gig CO CO CO CO 00 rH Ci rfi\n) CifH CO CO XO 1> O O O O ^ ^* ^* ^* 0 . iH . O ... O OJ Tp CO O \O J> ,-| rjt CO rH 00 O5 i~^ C^l iN (M 00 •' p cp us . T1 y *° T1 9s O5 CO CO * GO CO 1 0 .K nT, 1C .A Srf. 35*45 0 "282 or 0 '302 1 0 pfJ nr» 1 (} *7 39 '11 0 '205 8'0 40 '0 O.OKO in *i 48 '0 0-sa2 OK .n 51 '4 0.401 94.' 7 ? 52 •! O.on« -i e .4, 55 '0 0 "208 n'K. 56 '0 0 '209 or 0'355 11 *7 or 19 *9 Nickel 58 '7 0-186 ll'O Cobalt .... ... 59 »5 0.1 WQ ifk.q Copper. . 63 '6 O '184 n'7 Zinc 65 '3 0*151 9'fl 69 '0 0*214 14 "75 75 '0 0*200 15 *0 79*0 0 -339 &c 26 "8 &c 79*95 0-190 or 0'213 15 *2 or 17 -0 Rubidium 85 '5 0*133 11*4 87*66 0-152 13*3 Yttrium. 89 '1 0*197 17 "6 90-6 0-242 21 "9 103 '0 0 *232 23 "9 9 Palladium Silver 106-5 107 '92 0-213 0-121 22-7 13'1 1]2'0 0*124 13-9 113 '7 0-153 17*4 Tin 119 '0 0*232 or 0*161 27'6 or 19'2 120-0 0*204 or 0*200 24-5 or 24-0 126 '85 0 *192 or 0 *214 24*4 or 27*2 132 '9 0*117 15*6 137 '43 0*117 16*1 138'2 0*143 19*8 Cerium , 140 '2 0-143 20-0? Iridium 193 *1 0*165 31 *9 ? 195 -0 0*172 33*5 Gold 197*3 0*127 25*1 200-0 0*107 or 0-099 21*5 or 19*8? Thallium 204-0 0*106 21*6 Lead 206 -95 0-129 or 0*119 26 *7 ? or 24 *5 208-0 0-154 32*0? 232-6 0*123 28*7 142 Dr. J. H. Gladstone. The Relation between the The most notable change from previous tables is the increase in the value of hydrogen and the decrease in that of carbon, but the necessity of this has been gradually recognised by the principal workers on the refraction of organic bodies. This in no way affects the well-determined value CH2 = 7'6. It should be borne in mind that the specific refraction cannot claim a constancy equal to that of the atomic weight. The latter is generally believed to be identical under all circumstances, though the element may be capable of combining with another in two or more multiple proportions. On the other hand, several of the elements, as oxygen and iron, exhibit two or more specific refractions, which are not in multiple proportion, but depend upon the manner of combination. The best recognised of these are given in the third column, and the existence of others is indicated by an " &c." Beside these well-marked differences, there are many smaller variations, scarcely, if at all, beyond the limits of experimental error, which depend upon differences of physical condition or chemical structure. The numbers given in column 3 are therefore subject to an uncer- tainty, which may in some instances amount to 5 per cent. Where there is a greater divergence among the values observed, or where the deductions have been made from only one specimen, it is indi- cated by a query. PART II. — The Relation between the Specific Refraction and the Combining Proportion of the Metals. In the paper " On the refraction equivalents of the elements " previously referred to, it was shown that if the metallic elements be arranged in the order of their specific refractions, they are roughly in the inverse order of their combining proportions. In the lecture at the Royal Institution, I showed that this inverse order followed an approximate law, namely, that the " specific refractive energy of a metal is inversely as the square root of its combining proportion." This generalisation was proved for uni- valent metals, the figures showing (with the exception of sodium) a practically constant value for the product of the specific refraction and the square root of the combining proportion. By the aid of the table in the first part of this communication, the generalisation can now be tested throughout the whole range of the metallic elements. Refraction of the Elements and their Chemical Equivalents. 143 Univalent Metals. Metal. Specific refraction. ^ Combining proportion. Product. 0-514 2 -65 I.Otf Sodium • • • • 0-202 4 .on Potassium . . . . . 0'205 6 .OK 1.90 Rubidium • 0-133 q -94. 1.90 Silver 0-121 10 '3 1 .no 0 *117 11 "5 1 '^ 0-107 14*1 1 = the number of bands between the flat glass and either side of the blade in a distance DB from the edge ; BD = a and HPI = 0. 166 Note of the Radius of Curvature of a Cutting Edge. FIG. 3. £' We have Whence D'F+GE' = 2 a sin 10, nearly; FK+LG= KL = In the case of the razor on which these measures were made N = 85 % = 3 nz = 2 aje = 8a = 0*00405 in. sin \\Q — | and since for soda light \\ — O'OOOOllG in. nearly KL = 0-0000116 x 88-0-00405 x 0'25 = 0-00102-0-00101, nearly. Thus K L is not greater than O'OOOOl, and if it is assumed that the actual edge has the curved cross-section, indicated by the dotted line in fig. 3, the radius of curvature cannot be greater than 1/200,000 of an inch. A well sharpened razor will cut a hair, when merely pressed against it at about an eighth of an inch, or rather more, from the place where the hair is held. Human hair taken from the head has a circular cross-section, and varies in diameter in different individuals from 0'002 to 0*004 in. With a hair of 0'0025 in, diameter, fixed at one end and free at the other, it was found that half a grain acting at an eighth of an inch from the fixed end, bent it through an angle of about 30°. On the Determination of Wave-length of Electric Radiation. 167 The razor applied at the same distance from the fixed end would sometimes cnt through the hair before it had bent it as much as 30° ; and this shows that a force of half a grain must make the pressure per unit area at the place of contact sufficient to cause crushing or disruption of the material even when the edge has entered the hair to a distance comparable with the radius of the latter. If we assume that the thickness of the edge is 1/100,000 in. and that it has entered the hair until the length of the edge engaged is 1/1,000 in., the area in contact will be about 1/100,000,000 of a square inch and the pressure per square inch rather more than 3 tons, if the total force over the area of contact is half a grain. It is difficult to get any direct measure of the pressure required to destroy by crushing or shearing the material of which hair is com- posed, but horn which is of the same nature requires a much larger pressure than 3 tons per square inch to crush it. A rough experiment showed that a cylindrical steel punch with a flat end, began to sink into a block of horn when the pressure was between 12 and 16 tons per square inch. It would seem, therefore, that although the optical method shows that the thickness at the edge cannot be greater than 1/100,000 inch, the real thickness judged by the pressure per unit area necessary to cause the edge to cut in the way it actually does, must be considerably less than this. " On the Determination of the Wave-length of Electric Radia- tion by Diffraction Grating." By JAGADIS CHUNDER BOSE, M.A. (Cantab.), D.Sc. (Lond.), Professor of Physical Science, Presidency College, Calcutta. Communicated by LORD RAYLEIGH, Sec. R.S. Received June 2,— Read June 18, 1896. While engaged in the determination of the " Indices of Refraction of various Substances for the Electric Ray " (vide ' Proceedings of the Royal Society,' vol. 59, p. 160), it seemed to me that the results obtained would be rendered more definite if the wave-length of the radiation could at the same time be specified. Assuming the rela- lation between the dielectric constant K and the index JJL as indicated by Maxwell, to hold good in all cases, it would follow that the index could be deduced from the dielectric constant and vice versa. The values of K found for the same substance by different observers are, however, found not to agree very well with each other. This may, to a certain extent, be due to the different rates of alternation of the field to which the dielectrics were subjected. It has been found in general that the value of K is higher for slower rates of alternation 1 08 J)r. J. 0. Bose. On the Determination of the and the deduced value of ^ would therefore be higher for slow oscil- lations, the longer waves being thus the more refrangible. The order of refrangibilities would in such a case appear to be some- what analogous to that in an anomalously dispersive medium like iodine vapour. With exceedingly quick ethereal vibrations which give rise to light, there is an inversion of the above state of things, i.e., the shorter waves are generally found to be the more refrangible. It would thus appear that there is a neutral vibration region for each substance at which this inversion takes place, and where a trans- parent medium produces no dispersion. It would be interesting to be able to determine the indices of refraction corresponding to different wave lengths, chosen as widely apart as possible, and plot a curve of refrangibilities. A curve could thus be obtained for rock salt, which is very transparent to luminous and obscure radiations, and fairly so to electric radiation. Carbon bisulphide, which is very transparent to all but the ultra- violet radiation, would also be a good substance for experiment. For the construction of a curve of refrangibility for electric rays, having different vibration frequencies, the indices could be deter- mined by the method of total -reflection referred to above. The determination of the corresponding wave-lengths, however, offers great difficulties. Hertz used for this purpose the method of inter- ference, the positions of nodes and loops of stationary undulation produced by perpendicular reflection being determined by means of tuned circular resonators. Sarasin and De la Rive subsequently repeated these experiments with different sized vibrators and resonators. They found that the apparent wave-length depended solely on the size of the resonators. The wave-length found was approximately equal to eight times the diameter of the circular resonator. From these experiments it was supposed that the radiator emitted a continuous spectrum consisting of waves of different lengths, and that the different receivers simply resonated to vibrations with which they happened to be in tune. If this supposition be true the emitted radiation should, by the action of a prism, or better still, a^diffrac- tion grating, spread out in the form of a continuous spectrum. If, on the contrary, the radiation is monochromatic, the spectrum should be linear. The experiments to be described below may throw some light on this question. Professor J. J. Thomson, referring to the above case, is of opinion that the hypothesis of a continuous spectrum is highly improbable. It is more likely that, owing to the oscillation being of a dead-beat character, the resonator is set in vibration by the impact of incident electric waves. Each resonator vibrating at its particular free period, Wave-length of Electric Radiation by Diffraction Grating. 169 measures its own wave-length. There is, however, one difficulty in reconciling the theoretical value with that actually obtained. According to theory, the wave-length should be equal to twice the circumference, or 2?r times the diameter of the circular resonator. The value actually obtained by Messrs. Sarasin and De la Rive is, as has been said before, eight times the diameter of the circle. Rubens, using a bolometer and Lecher's modification of the slide bridge, determined the nodes and loops in a secondary circuit in which stationary electric waves were produced. A curve obtained by representing the bolometer deflections as ordinates and the distances of the bridge from one end as abscissae, shows the harmonic character of the electric disturbance in the wire. It was found that the wave-length obtained by this method did not depend on the period of the primary vibrator; the wave-length measured was merely that of the free vibration started in the secondary circuit by the primary disturbance. Hertz's method is therefore the only one for the measurement of electric waves in air, and the result obtained by this method is vitiated by the influence of the periodicity of the resonator. It was therefore thought desirable to obtain the wave-length of electric radiation in free space by a method unaffected by any peculiarity of the receiver. I have succeeded in determining the wave-length of electric radiation by the use of curved gratings, and the results obtained seem to be possessed of considerable degrees of accuracy. Rowland's method of using the curved grating for obtaining diffraction light spectra was also found well suited for the production of pure spectra of electric radiation. The focal curve / in this arrangement is a circle, having as a diameter the straight line joining the centre of curvature C with the apex M of the grating. FIG. 1. Gr, the grating ; M, its apex ; f, the focal curve. A source of radiation situated on this curve will give a diffracted spectrum, situated on the same curve defined by the equation (a+ 6) (sin i ± sin 6) — n\ 170 Dr. J. C. Bose. On the Determination of the where a + b is the sum of breadths of strip and space in the grating, i = angle of incidence, 6 = angle of diffraction. The sign of 9 is taken positive when it lies on the same side of the normal as the incident radiation. In the above equation there are two interesting cases : — (1) When the receiver is placed at C, 0 = 0° (a + 6) sini = n\. (2) When the deviation is minimum i •= 0 2 (a + 6) sin i = n\. Arrangement of the Apparatus. The grating, which is cylindrical, is placed vertically on a wooden table, with its centre at C, occupied in the diagram by the spiral spring coherer S. With the radius, which joins the centre to the apex of the grating, as a diameter, a circle is engraved on the table — the focal curve — on which the radiator and the receiver are always kept. A pin is fixed immediately below the apex, and a graduated ring sunk in the table with this pin as the centre. The graduated FIG. 2. The radiator, R, and the receiver, S, revolve round a pivot vertically below the apex of the grating, along the focal curve. The angles aro measured by the graduated circle, D. circle is used for the measurement of the angles of incidence and diffraction. Two radial arms revolving round the pin carry the radiator and the receiver. The ends of the arms near the pin have Wave-length of Electric Radiation by Diffraction Grating. 17 i narrow slits, through which the pin projects. The slits allow the necessary sliding for placing the radiator and the receiver on the focal curve. It would be better to have the sliding arrangement at the free ends of the arms, the pin passing through the central ends, acting as a pivot. The circle is graduated into degrees, but one- fourth of a degree may be estimated. Description of the Apparatus. The Radiator. — Electric oscillation is produced between two metallic beads and an interposed sphere 0'78 cm. in diameter. The beads and the interposed sphere were at first thickly coated with gold, and the surface highly polished. This worked satisfactorily for a time, but, after long -continued action, the surface of the ball became roughened, and the discharge ceased to be oscillatory. After some difficulty in obtaining the requisite high temperature, I succeeded in casting a solid ball and two beads of platinum. There is now no difficulty in obtaining an oscillatory discharge, and the ball does not require so mnch looking after. As an electric generator, I at first used a small Ruhmkorff's coil, actuated by a battery. I, however, soon found that the usual vibrating arrangement is a source of trouble ; the contact points soon get worn out, and the break becomes irregular. The oscillation pro- duced by a single break is quite sufficient for a single experiment, and it is a mere waste to have a series of useless oscillations. But the most serious objection to the continuous production of secondary sparks is the deteriorating action on the spark balls. Anyone who has tried to obtain an oscillatory discharge knows how easily the discharge becomes irregular, and the most fruitful source of trouble is often traced to the disintegration of the sparking surface. In my later apparatus I have discarded the use of the vibrating interrupter. The coil has also been somewhat modified. A long strip of paraffined paper is taken, and tinfoil pasted on opposite sides ; this long roll is wound round the secondary to act as a condenser, and appropriate connexions made with the interrupting key. This arrangement Fm. 3. The Radiator. VOL. LI. 172 Dr. J. C. Bose. On the Determination of the secures a great saving of space. Two jointed electrodes carry the two beads at their ends ; the distance between the beads and the interposed ball can be thus adjusted. This is a matter of importance, as the receiver does not properly respond when the spark-length is too large. Small sparks are found more effective with the receiver used. After a little experience it is possible to tell whether the discharge is oscillatory or not. The effective sparks have a smooth sound, whereas non- oscillatory discharges give rise to a peculiar cracked sound, and appear jagged in outline. The wires of the primary coil are in connexion with a small storage •cell through a tapping key. The coil, a small storage cell, and the key are enclosed in a tinned iron box. It must be borne in mind that a magnetic disturbance is produced each time the primary FIG. 4. The Radiating Box, one-fifth natural size. circuit of the induction coil is made or broken ; a sudden variation of the magnetic field disturbs the receiver. The iron box in which the coil is enclosed screens the space outside from magnetic disturbance. On one side of the box there is a narrow slit through which the stud of the press-key projects. In front of the box is the radiator tube, which may be square or cylindrical. The radiating apparatus used in the following experiments has a square tube 1 sq. in. in section. The apparatus thus constructed is very portable. The one which I often use is 7 in. in height, 6 in. in length, and 4 in. in breadth. To obtain a flash of radiation it is merely necessary to press the key and then release it. The break is made very sudden by an elastic spring. The Spiral Spring Receiver. — The receiving circuit consists of a spiral spring coherer in series with a voltaic cell and a dead-beat galvanometer of D'Arsonval type. An account of this form of re- ceiver has already been given (vide " On the Indices of Refraction of Wave- length of Electric Radiation by Diffraction Grating. 173 various Substances for the Electric Ray," ' Roy. Soc. Proc.,' vol. 59, p. 163). The receiver is made linear by arranging bits of steel spiral •springs side by side, the sensitive surface being 3 mm. broad and 2 cm. in length. An electrical current enters along the breadth of the top spiral and leaves by the lowest spiral, having to traverse the intermediate spirals along the numerous points of contact. The resistance of the receiving circuit is thus almost entirely concentrated FIG. 5. The Spiral Spring Coherer. at the sensitive contact surface, there being little useless short cir- cuiting by the mass of the conducting layer. When electric radia- tion is absorbed by the sensitive surface, there is a sudden diminution of the resistance, and the galvanometer in circuit is violently de- flected. By adjusting the electromotive force of the circuit the sensitiveness of the receiver may be increased to any extent desir- able. The receiver at each particular adjustment responds best to a definite range of vibration lying within about an octave. The same receiver could, however, be made to respond to a different range by an appropriate change of the electromotive force acting on the circuit. Very careful adjustment of the E.M.F. of the circuit is necessary to make the receiver respond at its best to a particular range of electric vibration. The Cylindrical Grating. — The source of radiation — the spark gap — being a line, the curved diffraction grating is made cylindrical. The spark gap is always kept vertical ; the grating is made of equi- distant metallic strips, which are vertical and parallel. A piece of thin sheet ebonite is bent in the shape of a portion of a cylinder and kept in that shape by screwing against upper and lower circular guide pieces of wood. Against the concave side of the ebonite are stuck strips of rather thick tinfoil at equal intervals. Five different o 2 174 Dr. J. C. Bose. On the Determination of the FIG. 6. The Cylindrical Diffraction G-rating. gratings were thus made with strips or spaces equal to 3 cm., 2*5 cm., 2 cm., 1*5 cm., and 1 cm. respectively. The diameter of the cylindrical grating is 100 cm. It would perhaps have been better to use a grating with a less curvature, but it must be remembered that the intensity of radiation is very feeble, and I was apprehensive of the receiver failing to respond when placed at too great a distance. I find from the sensibility of the receiver used that it would be possible to increase the diameter of the cylinder to about 150 cm., and this size I intend to use in the con- struction of my next grating. The aperture of the grating is in the following experiments reduced to the smallest practicable limit. Account of the Experiments. The receiver being placed at a suitable position on the focal curve, the radiator is moved about on the same curve till the diffracted image falling on the receiver produces response in the galvanometer. The procedure adopted is as follows. The receiver is placed, say, at the centre of the grating (0 = 0°). The electric ray at first falls on the grating at a large angle of incidence. A series of flashes of electric radiation are now produced by manipulating the key, and the angle of incidence gradually decreased till the receiver suddenly responds. The angle of incidence corresponding to the zero angle of diffraction is thus determined. The receiver is then placed at a new position on the focal curve, and the corresponding angle of incidence determined as before. In this way a series of angles of incidence, with their corresponding angles of diffraction, are found for each grating. Wave-length of Electric Radiation ly Di/r action Grating. 175 It should be remarked here that numerous difficulties were encountered in carrying out the experiments. The reflections from the walls of the room, from the table, &c., were at first sources of considerable trouble. By taking special care, I succeeded in elimi- nating these disturbances. The radiating balls were placed about 1 cm. inside the square tube. This prevented the lateral waves acting on the receiver. The receiver was provided with a guard tube, which stopped all but the diffracted radiation reaching the sensitive surface. The insulated wires from the ends of the receiver were protected by thick coatings of tinfoil, and led to the galvanometer, which was placed at a considerable distance. The cell and the galvanometer were enclosed in a metallic case with a narrow slit for the passage of light reflected from the galvanometer. In spite of all these precautions, I was baffled for more than six months by some unknown cause of disturbance which I could not for a long time account for. It was only recently, when nearly convinced of the futility of further perseverance, that I discovered the mistake in supposing sheets of tinned iron to be perfectly opaque to electric radiation. The metal box which contains the radiating apparatus seems to transmit a small amount of radia- tion through its walls, and if the receiver happens to be in a very sensitive condition it responds to the feeble transmitted radia- tion. I then made a second metallic cover for the radiating box, which precaution was found effective, provided the receiver was riot brought very close to the radiator. The receiver is still affected if placed immediately above the radiator tube, though two metallic sheets be intervening. For this reason I had to postpone taking the reading for minimum deviation till I had made a radiation-proof box. A soft iron box (to prevent escape of magnetic lines of induc- tion), enclosed in a second enclosure of thick copper, would, I expect, be found impervious to electric radiation. With the second protective enclosure, all difficulties were prac- tically removed. As a test for the absence of all disturbing causes, I observed whether the receiver remained unaffected when the grating was " off." There is a further test for the absence of external dis- turbances. The response, if only due to the diffracted beam, depends on the position of the radiator on the focal curve. If this angle of incidence is decreased, there should then be no action on the receiver. I found the positions of the radiator on the focal curve producing action on the receiver, to be well denned, and I experienced no further disturbance due to stray radiations. The grating is fixed vertically on the table, so that its centre is at the same height as that of the middle of the receiving and radiating tubes. A small mirror is fixed at the middle of the central strip. The observer, placing his eye at the same height as that of the 176 Dr. J. C. Bose, On the Determination of the radiator, levels the grating till the image of the eye is seen reflected by the mirror. I first obtained an approximate value of the wave-length with a 2-cm. grating, and then took careful and systematic readings with the different gratings. By different gratings is meant the same carved piece of ebonite, on which strips of different breadths were successively applied. The grating was found fairly adjusted, and the readings taken on the right side of the grating agreed well with the corre- sponding ones on the left side, I did not, therefore, think it necessary to take double readings, but took the various readings alternately on the right and on the left side. In one case only I found the grating on one side giving slightly better reading than the other. When the incident angle is too oblique, the diffracted image is not sharp, and I therefore did not extend the reading beyond 40° of incidence. Spectra of the tirst order only were observed. The response in the receiving circuit was somewhat feeble when 1 cm. or 1*5 cm. grating was used. But a 2-cm. grating gave stronger indications. With 2*5 and 3 cm. gratings the response was very energetic and the definition of the diffracted spectrum very sharp. For example, when the receiver was kept fixed, and the angle of incidence gradually varied, there was an abrupt and strong response produced in the receiving circuit, as soon as the angle of incidence attained the proper value. A slight varia- tion of this angle, even of less than a quarter of a degree, produced displacement of the diffracted image, and there was then no further action on the receiver. Had my graduated circle permitted it, I could have got more accurate readings. The radial arms carrying the receiver and radiator were of too primitive a design to make it worth while to attempt greater accuracy. I give below the readings of the angles of incidence and the corresponding angles of diffraction obtained with the different gratings, and the wave-length deduced from them. Grating A. — Breadth of strip = 1 cm. i. •• \. Mean \ for A. 38-0° 18° 1-849 35-0 37'0 20 19 1-831 1-854 1 -843 38-75 17 1-837 Wave-length of Electric Radiation by Diffraction Grating. 177 Grating B.— Breadth of strip = T5 cm. i. e. A. Mean for B. 38-0° 0° 1-847 26-0 10 1-836 1-844 28-5 8 1-849 Grating C. — Breadth of strip = 2 cm. i. e. A. Mean for C. 27-5° 0° 1 -846 22-0 5 1-847 1-849 20-0 7 1-855 Grating D. — Breadth of strip = 2'5 cm, i. e. A. Mean for D. 21-5° 0° 1-832 29-5 33-0 - 7 -10 1-852 1-854 1-845 34-0 -11 1-841 Grating E. — Breadth of strip = 3 cm, i. e. A. Mean for E. 18-0° 0° 1-854 23-25 25-5 - 5 - 7 1-845 1-851 1-848 31-0 -12 1 -843 It would thus be seen that the different values of wave-length obtained from the above experiments are concordant, the mean value being 1*846 cm. I then carefully removed the electrical vibrator, and measured approximately the size of the sparking balls. The radiator, it must be remembered, was placed vertically inside a square tube, each of whose sides is 2'5 cm. The radiator was about 1 cm. inside from the free end of the tube. 178 On the Determination of Wave-length of Electric Radiation. The diameter of the central ball = O78 cm. „ each side bead = O3 „ Distance between the outer surfaces of the beads = 1'5 cm. ?J „ inner (sparking) surfaces ,, = 0'9 „ The wave-length, 1*84, is almost exactly equal to twice the distance between the sparking surfaces of the beads. Without further ex- periments with different sized radiators, it is difficult to say whether the above simple relation is accidental or not. The following rough determinations, made with a second radiator, may be of some interest in connexion with the above. I took off the central sphere from the radiator used in the last experiment, and substituted a larger ball. The distance between the inner sparking surfaces is then 1'2 cm. Breadth of Strip = 3 cm. k. e. \. Mean. 23-0° 0° 2-34 29-0 -5 2-38 2-36 34-5 -10 2-36 The wave-length found is approximately equal to 2*36 cm., and twice the distance between the sparking surfaces is 2*40 cm. Conclusion. — The experiments described above seem to prove that the diffracted spectrum is not continuous, but linear. The method of determining the wave-length of electric radiation by diffraction grating is seen to give results which are concordant. The deter- minations are not affected by the periodicity of the receiving circuit, the receiver being simply used as a radioscope. With a better mounting and a finely graduated circle, it would be possible to obtain results with a far greater degree of accuracy. I hope to send, in a future communication, the results obtained with a better form of apparatus, with which I intend to study the relation of the wave- length with the size of the radiator, and the influence of the enclosing tube on the wave-length. I shall at the same time send an account of transmission gratings. Effects of strong Magnetic Field upon Discharges in Vacuo* 179 14 The Effects of a strong Magnetic Field upon Electric Dis- charges in Vacuo." By A. A. C. SwiNTON. Communicated by LORD KELVIN, F.R.S. Received June 10,— Read June 18, 1896. As is well known, when the lines of force of a magnetic field cut the path of the cathode rays in a vacuum tube, the rays are deflected in one direction or another, according to the polarity of the lines of force. If, on the other hand, the relative positions of the vacuum tube and the magnet are such that the lines of force and the cathode rays are parallel, the rays are not sensibly deflected. Under certain circumstances, however, I have found that with the rays and lines of force. parallel, other phenomena occur both in regard to the appearance of the discharge and in connexion with the internal resistance of the tube. The apparatus employed consisted of a Crookes tube of the form illustrated, supported vertically over one pole of a straight electro- magnet. The tube, which was excited by means of a 10-inch Ruhm- korff coil, working much below full power, was about 11 inches in length. The cathode terminal consisted of an aluminium plate at one end of the tube, and the anode a similar plate at one side. The tube was exhausted to a degree that gave considerable green fluor- escence of the glass, with a very slight trace of blue luminescence of the residual gas in the neighbourhood of the cathode and anode. The magnet ^mployed had a soft iron core 12 inches in length and 1|- inches diameter. It was wound with 2376 turns of No. 18 S.W.Gr. copper wire, which, when supplied with continuous electric current at 100 volts pressure, allowed from 13 to 14 amperes to pass, and magnetised the iron core practically to saturation. When the Ruhmkorff discharge passed through the tube, the magnet not being excited, the general appearance was as shown in ftff. 1, the walls of the tube showing everywhere green fluorescence, which was especially strong all over the rounded end of the tube opposite the cathode. A very small amount of blue luminescence could also be faintly seen just below the cathode, and also in the vicinity of the anode. With the tube and magnet placed as in fig. 2, as soon as the magnet was excited, the whole appearance of the discharge in the tube was found to alter immediately to what is shown in the illus- tration. Excepting for a very little at the top of the tube near the cathode, and a very bright spot at the bottom immediately over the magnet pole, all the green fluorescence of the glass disappeared, while extending from near the cathode to the bright spot at the 180 Mr. A. A. C. Swinton. The Effects of a strong bottom of the tube, a very bright cone of blue luminescence with a still brighter whitish blue core, made its appearance. When under these conditions the tube was slightly moved sideways, the bright spot at the apex of the cone, and the cone itself moved, the spot and apex always maintaining a position exactly over the centre of the magnet pole. At the same time the minor blue luminescence pro- ceeding from the anode terminal, due probably to the " make " current of the Ruhmkorff coil, was bent downwards towards the magnet as shown, and deflected sideways one way or another accord- ing to the polarity of the magnet, which polarity, however, did not affect in any way the vertical cathode stream. The internal resist- ance of the tube, as measured by an alternative spark gap on the Buhmkorff coil, was also found to be very greatly diminished while the magnet was excited. With the magnet not excited, the alterna- tive spark would leap from 1^ to If inches, while, when the magnet was excited, the gap had to be reduced to about j- inch before the sparks would pass. As soon as the current from the magnet was cut off, the appearance of the tube immediately reverted to what is shown in fig. 1, and its internal resistance increased to what it had been before. Magnetic Field upon Electric Discharges in Va«uo. 181 Experiments were also tried with the tube reversed as shown in fig. 3. In this case the internal resistance was affected by the magnet just as it had been previously. The appearance of the tube was also altered by the diminution almost to vanishing point of the green fluorescence, the presence of very bright blue luminescence on the under side of the cathode next the magnet, some less bright blue fluorescence near the anode, and a considerable amount of faint blue luminescence throughout the remainder of the tube. In this case, as in the other, the tube reverted to its normal appearance as soon as the magnet was demagnetised, and the appear- ance was the same whether the pole of the magnet next the tube was north or south. 182 Mr. F. G. Baily. The Hysteresis of — Fiq.S. — Further experiments with the tube placed horizontally so that the magnetic lines cut the cathode rays produced the usual deflection of the latter, but did not seem to have any appreciable effect on the internal resistance of the tube. "The Hysteresis of Iron and SteeJ in a Rotating Magnetic Field." By FRANCIS G. BAILY, M.A. Communicated by Professor LODGE, F.R.S. Received April 9, — Read June 4, 1896. (Abstract.) That the hysteresis of iron varies with the conditions of magnetic change has been ascertained in some instances, notably those in which the attractions between the molecular magnets of the Weber-Max- well-Ewing theory are diminished by super-imposed vibrations in the Iron and Steel in a Rotating Magnetic Field. 183 molecules. By deduction from this theory it lias been surmised that the hysteresis in magnetic metals under the influence of a constant rotary magnetic field will be less than that in an alternating field in which the magnetising force passes through a zero value. As familiar practical examples of the two conditions may be instanced : the armature core of a continuous current dynamo, and the iron cir- cuit of an alternating current transformer or choking coil. It is supposed that residual magnetism is due to the combination of molecular magnets in stable magnetic arrangements, and that the energy dissipated in any magnetic change corresponds to the work done in breaking up these arrangements. This energy is rendered kinetic by the movement of the magnets to form new combinations, the magnets either oscillating about the new position or moving to it aperiodically, according to the amount of damping to which they are subject. It is further suggested that the damping is of an elec- trical or electro- magnetic nature rather than of africtional character, being produced by the effect of rapid oscillations of the magnets on the surrounding particles or medium. Hence any movement of the molecular magnets during which the formation of new combinations is checked or prevented will take place with considerable reduction in the energy loss due to this cause. Such a condition is realised when the magnetic substance is sub- jected to a rotary magnetic field of sufficient strength to force the molecules to maintain a direction parallel to that of the field. If hysteresis is due only to the formation of new combinations and not to mechanical restraint, then under these conditions it will vanish altogether. Experiments were carried out to verify this deduction. A finely laminated cylinder of iron was suspended on its axis between the oles of an electro-magnet which was capable of rotation about the axis of suspension of the cylinder, thus producing a magnetic field rotating in a plane at right angles to this axis. The cylinder, though otherwise free to rotate, was restrained from continuous rotation by a spring, and the angle of rotation and consequent restoring force of the spring was indicated by a beam of light reflected from a mirror on the cylinder. The speed of the electro- magnet and the exciting current could each be varied. On rotating the magnet, the armature was dragged round until the restoring force of the spring equalled the force due to hysteresis, and the value of the latter could be obtained from the observed deflexions. The result showed that the value of the hysteresis under these con- ditions was very different from that obtained in an alternating field. At first the value was higher for corresponding inductions, but at an induction of about 16,000 in soft iron and 15,000 in hard steel the hysteresis reached a sharply defined maximum and rapidly dimin- 184 Mr. E. Rutherford. A Magnetic Detector of ished on more complete magnetisation, until at an induction of about 20,000 it became very small with every indication of disappearing altogether. Soft iron and hard steel gave very similar curves, and in both the curve of hysteresis-induction cut the curve obtained from the values in an alternating field at a point just before the maxi- mum. The result fully bears out the deduction from the theory, and proves in addition that hysteresis is not sensibly due to anything of the nature of mechanical restraint of the molecules. The form of the curve also gives clear indications of the three stages of molecular movement, the first stage giving a slowly rising curve, the second a straight rapid rise, and the third a straight and much more rapid descent. Further experiments were carried out on the effect of speed of rotation. In an alternating field the speed of reversal has been shown to be without sensible effect on the hysteresis, and theory points to this result as a natural deduction. The above apparatus was well adapted for testing the matter, since the hysteresis per reversal could be read at each instant independently of the speed. From an ex- tremely slow speed up to 70 revolutions per second no definite change was found in the value of the hysteresis. At the same time several small modifications were noted, produced by rapid variations in the speed of rotation or magnetising force. The effect lasted through many revolutions, but ultimately the same steady condition was arrived at. At a,nd near the maximum value the hysteresis was very variable. The effects were much more marked in soft iron than in hard steel, as would be anticipated from the theory of their constitution. The experiments in their verification of an untried deduction form a strong proof of the validity of the molecular theory of magnetism, and throw some light on the nature of the molecular complex and of the interactions which take place therein. "A Magnetic Detector of Electrical Waves and some of its Applications." By E. RUTHERFORD, M.A., 1851 Exhibition Science Scholar, New Zealand University, Trinity College, Cambridge. Communicated by Professor J. J. THOMSON, F.R.S. Received June 11,— Read June 18, 1896. (Abstract.) The effect of Leyden jar discharges on the magnetisation of steel needles is investigated, and it is shown that the demagnetisation of strongly magnetised steel needles offers a simple and convenient means for detecting and comparing currents of great rapidity of alternation. Electrical Waves and some of its Applications. 185 The partial demagnetisation of fine steel wires, over which is wound a small solenoid, was found to be a very sensitive means of detecting electrical waves at long distances from the vibrator. Quite a marked effect was found at a distance of over half a mile from the vibrator. Detectors made of very fine steel wire may be used to investigate waves along wires and free vibrating circuits of short wave-length. Fine wire detectors are of the same order of sensitiveness as the bolometer for showing electrical oscillations in a conductor. This detector also has the property of distinguishing between the first and second half oscillations of a discharge, and may be used for determining the damping of electrical vibrations and the resistances of the discharge circuit. A method of experimentally determining the period of oscillation of a Leyden jar circuit by the division of rapidly alternating currents in a multiple circuit is explained. The capacity and the self-induct- ance of the circuit for high frequency discharges may also be deduced, so that all the constants of a discharge circuit may be experimentally determined. In the course of the paper the following subjects were investigated. (1) Magnetisation of Iron by High Frequency Discharges. — The effect of the Leyden jar discharge on soft iron and steel is fully examined. Steel needles which had been placed in a solenoid and subjected to a discharge were examined by dissolving them in acid. It was found that there was apparently only evidence of two half oscillations in the discharge, and this effect is due to the demagnetising force exerted by the needle on itself during the discharge. The effect of continued discharges on the demagnetisation of mag- netised steel needles was investigated, and also the effect of varying the length and diameter of the steel needles. When a discharge is sent longitudinally through a magnetised steel wire the magnetic moment of the needle is always decreased, due to the circular magnetisation of the wire by the current through it. This " longitudinal " detector, when of thin steel wire, was found to be a sensitive means of detecting electrical oscillations of small amplitude. Both the "longitudinal" and " solenoid al " detectors may be readily used for comparing the intensities of currents in multiple circuits when traversed by currents of the same period. (2) Detection of Electrical Waves at Long Distances from the Vibrator. — A compound detector needle was composed of fine steel wires and a small solenoid wound over it. When this detector was placed in series with the wires of a receiver, the electrical oscillations set up in the circuit tended to demagnetise the magnetised detector needle. By this method electrical waves from a Hertzian vibrator were 186 Mr. J. S. Townsend. detected for long distances. An effect was obtained at over half a mile from the vibrator. (3) Waves along Wires. — The uses of fine steel wires for examining the distribution of currents along wires are explained. (4) Damping of Oscillations. — A method of determining the damp- ing of discharge circuits is investigated. The absorption of energy in spark gaps is deduced, and the apparent resistance of the air break to the discharge determined. (5) Resistances of Iron Wires. — Quantitative results are given for the resistance of iron wires for very rapid alternations. The value of the permeability of the different specimens is deduced, and it is shown to vary with the diameter of the wire and the intensity of the discharge. (6) Absorption of Energy by Conductors. — The absorption of energy of iron and non-magnetic cylinders placed in solenoid through which a discharge passed were determined. Iron cylinders were found to absorb much more energy than copper ones of the same diameter, and the permeability of the iron for the discharge is deduced. (7) Determination of the Period of Oscillation of Leyden Jar Dis- charges. — A method of accurately determining the period of oscillation is based on the division of rapid alternations in a multiple circuit, one arm of which is composed of a standard inductance, and the other of a variable electrolytic resistance. The value of n, the number of oscillations per second, when the currents in the branches of the multiple circuits are equal, is, under certain conditions, given by — R where B = resistance of electrolyte to the discharge, W = value of the standard inductance. The value of the self-inductance and capacity of the discharge cir- cuit for very rapid oscillations may also be experimentally deduced. " Magnetisation of Liquids." By JOHN S. TOWNSEND, M.A. Dub. Communicated by Professor J. J. Thomson, F.R.S. Received June 11,— Read June 18, 1896. (Abstract.) The experiments on the coefficient of magnetisation of liquids were made with a sensitive induction balance. Both circuits were com- muted about sixteen times a second, so that very small inductances could be detected by the galvanometer in the secondary circuit. The principle of the method consisted in balancing the increase of the Magnetisation of Liquids. 187 mutual induction of the primary on the secondary of a solenoid arising from the presence of a liquid in the solenoid against known small inductances. Thus, if the sum of the inductances be reduced to zero, as shown by the galvanometer in the secondary giving no deflection, the balance will be disturbed to the extent 47T&M, due to the insertion of a liquid into the solenoid whose coefficient of mag- netisation is &, and the galvanometer in the secondary circuit will give a deflection when the commutator revolves. An adjustable inductance is then reduced by a known amount, m, till the deflection disappears ; so that we get 47T&M = in .'. k = m/4 7r\f, where m and M are quantities easily calculated. Since the formula does not contain either the rate of the rotation of the commutator nor the value of the primary current, no particu- lar precautions are necessary to keep these quantities constant. In all the determinations the magnetising force was varied from 1 to 9 centigram units, and in no case was there any variation in k. The densities of the salts in solution were also varied over large ranges, and showed that the coefficient of magnetisation for ferric salts in solution depended only on the quantity of iron per c.c. that was present, giving the formula 107fc = 2660 W— 7*7 for ferric salts, where W is the weight of iron per c.c., the quantity — 7'7 arising from the diamagnetism of the water of solution. A similar result was obtained for ferrous salts, the corresponding formula being 107& = 2060 W— 77, the temperature being 10° C. The following table shows the coefficient of magnetisation for the different salts examined, w being the weight of the salt per c.c. of the solution : — 10? &. Fe2Cl6 916w-7-7 Fea(S04)3 .... 745^-7-7 Fe,(NO,)6..... 615 w- 7-7 FeCl2 908w-7'7 FeSO4 749*0-7-7 The effect of temperature was also estimated, the results of the experiments being shown by means of a curve (fig. 1), the x ordinates of which denote the temperature, and the y ordinates are proportional to the coefficient of magnetisation, a length corresponding to 50 being subtracted from each for convenience of representation. VOL. LX, p 188 Prof. J. B. Farmer and Mr. J. LI. Williams. The first is drawn from results of experiments performed on ferric chloride containing 0'086 gram of iron per c.c., the second from ferrous chloride containing O148 gram of iron per c.c., the third from ferric sulphate containing 0'105 gram of iron per c.c., and the fourth from an alcoholic solution of ferric chloride. The curves all show about the same temperature coefficient at points corresponding to the same temperature. " On Fertilisation, and the Segmentation of the Spore, in Fucus." By J. BRETLAND FARMER, M.A., Professor of Botany at the Royal College of Science, and J. LI. WILLIAMS, Marshall Scholar at the Royal College of Science, London. Communicated by D. H. SCOTT, M.A., Ph.D., F.R.S. Received May 21,— Read June 18, 1896. The object of the present communication is to give an account of the chief results of an investigation into the processes connected with the formation and fertilisation of the oospheres and the germination of the spore in Ascophyllum nodosum, Fucus vesi- culosus, and Fucus platycarpus. The more obvious details of development have been especially studied by Thuret, and later by Oltmanns. But neither of these writers paid any special attention to the behaviour of the cell-nuclei, nor did they succeed in observing the actual process of fertilisation. Behrens has com- municated an account (* Ber. d. Deutschen Bot. Gesel.,' Bd. IV) of some researches made by himself on the fertilisation of the oospheres, but we are unable to accept his conclusions for reasons shortly to be recounted. The material for these investigations was obtained in London from Bangor, Plymouth, and Jersey, but it was compared with other material collected and fixed at the seaside at Bangor, Weymouth, and Criccieth. Furthermore, all the growing apices and eon- ceptacles for sectioning were collected by one of us directly at the three last named places. Some samples were gathered between the tides, and fixed at once, others were first kept for a time in salt water ; the best results, however, were obtained from plants collected in a boat about two or three hours after the tide had reached the plant, and also from other plants taken a short time before they were left exposed by the ebb tide. In order to study the fertilisation and germination stages, male and female plants were kept in separate dishes, and were covered over so as to prevent drying up. This method gave far better results than those more usually advocated. On the appearance of the On Fertilisation, and the Segmentation of the Spore in Fucus. 189 •extruded sexual products, the female receptacles were placed in sea- water, and after the complete liberation of the oospheres, a few male branches with ripe antherozoids were first placed in a capsule of sea water until it became turbid owing to their number. If on •examination the antherozoids proved to be active, smalt quantities were added to the vessels containing the oospheres. The latter were then fixed at intervals of five minutes during the first hour, and then at intervals of fifteen minutes, up to six hours after the addition of the antherozoids. After that, samples were killed at longer intervals up to three days, and this was continued till we had material fixed at all stages for the first fortnight. At first we used sea water in which to keep the embryos growing, but a proper solution of Tidman's sea salt was found to answer quite as well. For fixing, we tried the following reagents — chrome alum, picric -alum, Mann's picro-corrosive, corrosive sublimate, and acetic acid ; these were all dissolved in sea water, absolute alcohol, Flemming's and Hermann's solutions, and the vapour of osmic and formic acids. The Flemming's (strong formula) and Hermann's solutions were •diluted with equal parts of sea water. The first three fixatives were unsuccessful, acetic-corrosive yielded fair nuclear figures, but the material proved very brittle, and the spores were somewhat dis- torted. A portion of the cytoplasm was disorganised and the polar radiations were not preserved. Absolute alcohol fixed the oospheres and newly fertilised spores without distortion, but was useless for all other stages. Vapour fixing with osmic acid succeeded better than any of the preceding reagents but was greatly inferior to either Hermann's or Flemming's solutions in preserving the protoplasmic structure in an unaltered state. After the material had been fixed it was dehydrated and passed in the usual way into paraffin, the temperature of which was not allowed to exceed 50° C., and it was then cut with the microtome. The sections were stained with Heidenhain's iron-heematoxylin, with Flemming's triple stain, and a large number of other dyes. The results, which were compared carefully, led us to rely chiefly on the two staining processes mentioned, but at the same time we often obtained valuable preparations with other staining reagents as well. In spite of repeated attempts, we have not succeeded in observing the first nuclear division in the oogonium, but the later ones have been seen both in Fucus vesiculosus and in F. platy carpus, in which eight oospheres are formed. Oltmanns asserts that in Ascophyllum, in which only four oospheres are commonly formed, eight free nuclei occur at an earlier stage, but that four of these ultimately abort, and do not become centres of cell formation. Our observations tend to confirm him in this respect, but we found that in some cases a fifth oosphere, smaller than the rest, was occasionally differentiated, P 2 190 Prof. J. B. Farmer and Mr. J. LL Williams. and that when freed from the oogonium it exerted an attraction on the antherozoids just like its larger sister oospheres. When an oogonial nucleus is about to divide, it first becomes slightly, then very much, elongated so as to resemble an ellipse. Fine radiations are seen to extend from the two ends into the surrounding cytoplasm. The latter is at first tolerably uniformly granular, but as the radiations around the polar areas increase, these regions become cleared altogether of the granules which then become massed outside them. The nucleus rapidly becomes more spindle- shaped, and its chromatic elements are chiefly grouped near each pole, leaving a clear space about the equator in which the nucleolus. is situated. In this respect the nuclei of Fucus offer a striking con- trast to those of Pellia epiphylla already described (* Annals of Botany,' vol. viii, p. 221) by one of us. In the latter plant the chromatic portion of the nucleus assumes an equatorial position at the corresponding stage in division, whilst the polar regions are clear. The polar radiations continue to increase and the nucleus to- lengthen, until the entire structure recalls the figure of a dumb-bell, in which the nucleus answers to the handle, and the radiation areas to the knobs. If the radii be traced outwardly, they are seen to terminate either in the frothy protoplasm, on the angles where the foam walls meet, or on the large granules which surround the cleared areas and are embedded in the foam. This point is one of considerable importance, and we shall revert to it further on. No structures were seen which could certainly be identified as centrosomes, although bodies suggestive of them were often observed ; but these proved to be so variable in size and position, as well as in number, that we feel unable to attach any special significance to them. The next stage in the mitosis is that in which the interpolar spindle arises, with the chromosomes disposed upon its equator. The spindle is very remarkable inasmuch as it is entirely intranuclear, somewhat resembling that described by Fairchild for Valonia, or by Harper for Peziza. The nuclear wall can be distinguished until quite late in karyokinesis, and it is possible that no complete mingling of the cytoplasm with the contents of the nucleus takea place here. The spindle is extremely clear, and in several prepara- tions, owing to a fortunate contraction during manipulation, the ends of the nuclear part of the spindle also had broken away from the cytoplasmic poles, and were visible as clean conical structures forming the poles of the nuclear spindle. The chromosomes were too minute to admit of their development being satisfactorily studied, but in all the oogonial spindles their number was estimated at ten when seen arrayed on the spindle equator. They were only seen in profile, and consequently it was difficult to be sure whether there On Fertilisation, and the Segmentation of the Spore in Fucus. 191 were really ten or twelve, but the absolute number is not of im- portance as all the nuclei were compared from the same aspect. Remains, more or less preserving the original form, of the nucleolus were sometimes visible at this and even in a later stage. No division- planes are formed in the oogonium until the full complement of nuclei are produced; after this the positions which they will ultimately occupy are indicated by the heaping up into lines (or rather plates) of the cytoplasmic granules above referred to. These seem to be repelled equally from all the nuclei, thus effecting a symmetrical division of the entire oogonium. After the complete delimitation of the oospheres within the oogonium, we observed, as an occasional circumstance, that one of the oospheres might contain two, or even three, nuclei, a fact also noticed by Oltmanns. When the oospheres are extruded, and come to lie free in the water, they grow in size, and are turbid with granules, which are very abundant in the cytoplasm. The chromatophores early become distinguishable from the other constituents of the cell, and the nucleus occupies a central position. It is itself sur- rounded by a dense layer of cytoplasm, which later on becomes very strongly marked. About five minutes after the mixing of the sexual cells, the antherozoids are found to have slipped into many of the oospheres. We failed to observe the act of penetration, but found a number of cases in which the antherozoid could be recognised within the oosphere, before its final fusion with the nucleus of the latter. It is a roundish, densely staining body, and, unlike the majo- rity of animal sperm cells as yet described, it imports into the egg no system of radiations along with it. Judging from the short period of time elapsing between its penetration of the surface of the oosphere and its arrival at the exterior of the female nucleus, it must pass through the intervening cytoplasm with great rapidity. It then becomes closely appressed to the nucleus, and is about as large as the nucleolus of the latter. It rapidly spreads over a part of the female nucleus as a cap, and it presents a less homogeneous aspect than before. Both it and the female nucleus assume a granular condition, which is probably to be interpreted as representing a coiling and looping of the lining of the respective nuclei. Finally the two nuclei coalesce, and the original components can no longer be distinguished. Complete fusion may be effected in less than ten minutes after addition of the antherozoids to the water. These results are in striking accordance with those described by Wilson in connexion with the fertilisation of the eggs of echinoderms in his recent " Atlas of Fertilisation." A delicate pellicle is meanwhile formed around the periphery of the oosphere, which is thus easily distinguished from the unfertilised oospheres, in which such a membrane is wanting. The texture of the 192 Prof. J. B. Farmer and Mr. J. LI. Williams. cytoplasm also changes, and tends to assume a more definitely radiat- ing character, the lines starting from the nucleus as a centre. We observed, not unfrequently, rather large cells in which two- nuclei of equal size were lying in close juxtaposition. These cells, with their nuclei, answer exactly to the description given by Behrens of the fertilisation stage in plants examined by him. We are unable however, to accept his interpretation, for, in the first place, the series of fertilisation stages which we have observed, and have briefly described above, in no way correspond with the appearances described by him, and secondly, because these large cells (Behrens himself emphasises their size) are seen in material to which no antherozoids have had access. Furthermore, the average size of the young oospores is not obviously greater than that of the oospheres. themselves. We regard the bodies in question as representing abnormal developments of oogonial cells, and not as being in any way concerned with fertilisation. Moreover, we have occasionally observed one cell in the divided oogonium, much larger than the rest, to contain two, or even sometimes three, nuclei, and these nuclei are then always close together. These facts have led us to reject Behrens' account of the process. A very large number of experiments were made, in order to deter- mine, if possible, the time which elapsed between the addition of the antherozoids to the oospheres and the first division of the spore. A short summary of different sets of observations on Ascophyllum is given in the subjoined tables. SERIES I, — Observations on Ascophyllum conducted at the Seaside, (a) The antherozoids were added to the oospheres at 10 o'clock A.M. Lot 1. Fixed 23 hours after the addition of antherozoids. Nucleus preparing for division. „ 2. „ 24 ,, „ „ Nucleus divided, rhi- zoid rudiment present, no dividing wall. ,, 3 & 4 „ 32 „ „ „ Nucleus divided, no rhi- zoid, dividing wall pre- sent. „ 5. „ 36 „ „ „ Spore divided into about six cells. (6) The antherozoids added between 11 and 12 P.M. Lot 1. Fixed 24 hours after the addition of antherozoids. Nucleus divided, a few with rhizoid rudiments and division wall. „ 2. „ 25 „ ., „ Same result. „ 3. „ 25 „ „ „ Not beyond spindle stage. „ 4. „ 28 „ „ „ Nucleus divided, no rhi- zoid or dividing wall. On fertilisation, and the Segmentation of the Spore in Fucus. 193 SERIES II. — Observations on Ascophyllum carried on in the Laboratory. Antherozoids added between 5 and 7 P.M. Lot 1. Fixed 22£ hours after the addition of antherozoids. Nucleus divided, no rhi- zoid or dividing wall. „ 2. „ 23 „ „ „ Nucleus preparing for division. „ 3. „ 23 „ „ „ Same as 1. „ 4. „ 24£ „ ,, ,, Nucleus divided, rhizoid present, no dividing wall. The above observations prove that there is no essential difference between the behaviour of material examined in London and at the seaside respectively. After fertilisation, the cells rest for a long interval of time — com- monly about twenty -four hours, as shown in the foregoing table — before they begin to segment. The principal changes which occur during the interval are, first, in the rapid increase in the thickness of the peripheral cell wall, and, secondly, in the more regular arrange- ment of structure exhibited by the protoplasm. The alveolar, or foam character is extremely clear, and the chromatophores, which by this time have become very prominent, are noticed to be situated in the angles formed by the convergence of the foam walls ; they are often bent and otherwise distorted, and so accommodate themselves to the structural condition of the foam. Other granules, which stain deeply, and probably represent food reserve of a proteid nature, are also abundantly scattered through the cytoplasm. The first segmentation-division resembles, in a general way, the oogonial nuclear divisions already described, and the polar areas become similarly cleared of granules. The achromatic threads form- ing the polar radiations are very clearly seen to be attached to the foam-like structure of the cytoplasm, and, indeed, in some cases, insensibly to pass into it. At other times fibrils end on granules (or, perhaps, on the protoplasmic lining of the granules), and sometimes again a fibril may fork, and its branches end either on granules or on the foam angles. The inference to be drawn from these facts seems to be that the radiations are the result of a change — a differ- entiation— in the protoplasm as it already exists, and that they do not owe their origin to the presence of any special " spindle-forming sub- stance," by virtue of which they may be supposed to develop and "grow" as new structures in the cell. We propose, however, to discuss the general bearings of our observations on this and on other questions of theoretical interest in a future memoir, in which the evidence for our views will be set forth in detail. When the achromatic nuclear spindle appears, it also, as in the 194 On Fertilisation, and the Segmentation of the Spore in Fucus. oogonial mitoses, is intranuclear, and it is often separated from, the well-defined persistent nuclear wall by a clear space. The chromo- somes, when assembled on the spindle, at the equator, are seen to be twice as numerous as in the oogonial nuclei, i.e.., seen in profile we counted them as twenty in number. We were unable to distinguish any such grouping of the chromosomes as would lead to the conclu- sion that the chromosomes of the mate and female nuclei respectively had so far preserved their original identity as to appear in the form of two separate groups. The long interval of time which, in Fucus, elapses between fertilisation and the first nuclear division possibly may admit of a more thorough mingling or fusion of the parental chromosomes than would seem to be the case in some animals, e.g., the Copepoda as described by Riickert and by Hacker. During the diaster stage the connecting achromatic fibres are at first very distinct, but they soon become fainter, and no cell-plate is formed across them. The two daughter nuclei gradually pass into the state of rest, each being first hemispherical, with crenate projec- tions on the flattened side turned towards its sister nucleus. Only after nuclear division is complete does the first cell wall appear. The cell is sometimes spherical when this happens, and then it is divided into two similar hemispheres. Further divisions may then appear, whilst the general contour of the embryo still remains more or less spherical. These cases occurred most frequently when the germinat- ing spores were illuminated on all sides. But most commonly the first cell wall cuts the spore into two dissimilar halves, one of which grows out and forms a rhizoid. Often this projection is already apparent even before the first nuclear division occurs, and in any case one of the two daughter nuclei always passes down into the protuberance. The immediately succeeding divisions have been sufficiently de- scribed by Thuret and others, but we may remark that the division of the nuclei in all cases precedes the formation of a cell plate, which is not formed in connexion with the achromatic connecting fibrils as in the higher plants. The doubled number of the chromosomes is retained during the vegetative divisions of the thallus, and is constant throughout the somatic cells of the mature Fucus plant. Hence it follows that the reduction in the number of the chromosomes (in the female plants), is associated with the differentiation of the oogonium — the mother cell of the sexual products. Thus Fucus, in this respect, approximates more closely to the type of animal oogenesis than to that which obtains in those higher plants in which the details of chromosome reduction has been followed out. Regarded from the standpoint of the number of its chromosomes, the Fucus plant resembles the sporophyte of the higher plants, whilst Changes in the Dimensions of Carapace of Carcinus moenas. 195 the gametophyte of the latter, with its reduced number of chromo- somes, finds its analogue merely in the maturing sexual cells of Fucus. But until we know more of the nuclear changes as they occur in other Algae, and especially in the more primitive forms, it seems unadvis- able to go further than to indicate the possibility that we may require to revise our present ideas on the comparative morphology of the higher and lower groups of the vegetable kingdom. Even if we regard the reduction in the number of the chromosomes as a fact which is primarily of physiological importance, we may safely conclude, from the universality of its occurrence, that it is also intimately connected with the phylenogenetic development of living forms, and hence it must meet with due recognition on the part of the morphologist who is engaged in comparing the life-history of one group of organisms with that of others. " On certain Changes observed in the Dimensions of Parts of the Carapace of Carcinus mcenas." By HERBERT THOMPSON. Communicated by Professor W. F. R. WELDON- F.R.S. Received May 19,— Read June 11, 1896. In making some measurements of young male Carcinus mamas from Plymouth, corresponding to those made by Professor Weldon on young females of the same species, and published by him in the Report of a Committee for conducting statistical inquiries into the measurable characteristics of plants and animals (' Roy. Soc. Proc.,' vol. 57, p. 360), some interesting facts were observed as to changes taking place in the relative dimensions of certain parts of the carapace of these crabs in the space of the last three years. The carapace of the adult male crab, measured in the median antero-posterior line is, roughly, from 40 to 60 mm. long. Now, of young male C. mcenas collected at random at Plymouth in the year 1893, I had, for the purposes of measurement, 3,077 specimens, ranging between 10 and 15 mm. in length of carapace, and on these, besides the carapace length, as above defined, two other measure- ments were taken, viz. (1) "frontal breadth," the distance in a straight line between the tips of the two teeth which form the outer "boundaries of the orbits, and (2) the "right dentary margin," measured in a straight line from the tip of the first to that of the last lateral tooth on the right side of the carapace. The measurements were made in the way described in the Report above mentioned (ibid., pp. 361 — 2) : and owing to the rapid growth and alteration of proportional dimensions in the young crabs, they were sorted into groups, the members of each of which differed by less than 0'2 mm. in carapace length, thus giving five groups for 196 Mr. H. Thompson. On certain Changes observed in the every 1 mm. of growth in carapace length, or twenty-five groups for the whole range of 10 — 15 mm. carapace length. The numbers contained in the separate groups ranged from seventy-two in the smallest group to 178 in the largest group. The arithmetical mean and mean error in each group is set out in Table I infra. Similar measurements were made in the case of 1,957 young male C. mcsnas from Plymouth of the year 1895. These were likewise divided into groups differing by 0*2 mm. of carapace length : and the numbers contained in the twenty-five groups between 10 and 15 mm. carapace length ranged from thirty-four in the smallest one to 111 in the largest. The arithmetical means and mean errors are given in Table I infra. On comparing the two sets of measurements (expressed in terms of the carapace length which was taken as the unit) it appears, as regards the " frontal breadth," that in every one of the twenty-five groups without exception the average size of the frontal breadth in the 1893 crabs exceeded that of the 1895 crabs of corresponding size. Seeing how small the groups are the result is a striking one, and is given in greater detail in the following Table : — - C. mcenas. — Frontal Breadth. Carapace length in millimetres. Average excess of 1893 crabs over 1895 crabs. In thousandths of carapace length. In millimetres. 10—11 11—12 12—13 13—14 14-15 6-30 7-29 6-73 5-26 3-53 0*07 0'08 0'08 0-07 0-05 On the other hand, if the species in 1895 has a smaller average frontal breadth, it compensates for the deficiency by having a larger right dentary margin. This was found to be the case in twenty- three out of the twenty-five groups, the two non-conformist groups lying one near each end of the range. The arithmetical means and mean errors are given in Table I infra, and the results, tabulated in a corresponding form to those of the frontal breadth measurements, are as follows : — Dimensions of Parts of the Carapace o/Carcinus moenas. 197 C. moenas. — Right Dentary Margin. Carapace length in millimetres. Average excess of 1895 crabs over 1893 crabs. In thousandths of carapace length. Tn millimetres. 10-11 11—12 12—13 13—14 14—15 1-39 2'09 1-87 1-56 1-42 O'Ol 0-02 0-02 0-02 0-02 As these results seemed to indicate that a change in regard to these dimensions was taking place in the species, it was desirable to compare similar measurements in the adult. Fortunately Professor Weldon was able to supply me with 254 specimens of male G. mo&nas with a carapace length ranging between 40 and 63 mm., taken at Plymouth at random in 1892-3 : and for comparison he procured 496 individuals collected at Plymouth in January of the present year and corresponding in size. Measurements similar to those made on the young ones gave the following results : — In frontal breadth the 1892-3 crabs exceeded the 1896 crabs on an average by 8'85 thousandths of their carapace length, which for an average length of 50 mm. is equivalent to 0*44 mm., while in the right dentary margin the 1896 crabs exceeded those of 1892-3 on an average by 3'1 thousandths, or an equivalent of 0'16 mm., thus fully confirming the results arrived at in the young ones. Whether these results indicate a permanent change in the species at Plymouth in respect to these particular dimensions of the carapace, tending to the establishment of a new variety, or whether it is a mere oscillation such as, for all we know, may be constantly going on in the relative dimensions of the various parts of the members of al) species, can only be decided by further measurements, which, it is hoped, may be continued on the same species after another interval of two or three years. Meanwhile, the persistence with which the same tendency asserts itself in the twenty-six groups into which we have divided these crabs of 1892-3 and 1895-6 is remarkable, and may perhaps induce others to take measurements of other animals at definite intervals, and establish similar comparisons. I wish to add my hearty thanks to Professor Weldon for suggest- ing the line of investigation and furnishing material and ever-ready help. 198 Changes in the Dimensions of Carapace of Carciims moenas. Interruption of Afferent and Efferent Tracts of Cerebellum. 199 "Phenomena resulting from Interruption of Afferent and Efferent Tracts of the Cerebellum." By J. S. RISIEN RUSSELL, M.D., M.R.C.P., Research Scholar to the British Medical Association, Assistant Physician to the Metro- politan Hospital, and Pathologist to the National Hospital for the Paralysed and Epileptic, Queen's Square. Com- municated by Professor VICTOR HORSLEY, F.R.S. Received June 17,— Read June 18, 1896, (From the Pathological Laboratory of University College, London.) (Abstract.) The research was undertaken in the hope of obtaining evidence in support of or against the view that the cerebellum exercises a direct influence on the spinal centres, as opposed to any indirect influence exerted through the agency of the cerebral cortex. The inferior peduncle of the cerebellum was accordingly divided on one side, the organ itself and its other peduncles being otherwise left intact, and the results obtained by this procedure were controlled by experiments in which the lateral tracts of the medulla oblongata were divided on one side without injury to the pyramid on the one hand or to the posterior columns and their nuclei on the other. Further control experiments consisted in dividing transversely the posterior columns and their nuclei a few millimetres above the calamus scriptorins, on one side, without including the lateral tracts of the medulla in the lesion. The results obtained by these different experiments were supple- mented by others in which the electrical excitability of the two cere- bral hemispheres was tested and compared, immediately after division of one inferior peduncle of the cerebellum, and at some later period, such as three weeks, after the section of the peduncle ; also after partial hemisection of the medulla in which all the structures on one side were divided, with the exception of the pyramid which was left as far as possible intact. Other experiments consisted in observing the ways in which con- vulsions, induced by the intravenous injection of the essential oil of absinthe, were modified by division of one inferior peduncle of the cerebellum, by partial hemisection of the medulla in which the pyramid was the only structure left intact on one side, and by transverse section of the posterior columns and their nuclei, on one side, a few millimetres above the calamus scrip torius. Considered in conjunction with results previously obtained by the author and others after ablation of one lateral half of the cerebellum, and after intracranial section of the auditory nerve, the results now 200 Dr. Russell. Phenomena resulting from Interruption obtained afford valuable information with regard to many of the functions of the cerebellum ; but they are not claimed as supplying definite information on the important question as to whether the cerebellum exercises a direct downward influence on the spinal •centres or not. Many of the results obtained suggest the possibility of such a downward influence ; but most of the effects can as readily be explained by supposing that they are the result of the interruption of afferent impulses passing from the periphery to the cerebellum. The direction of rotation was towards the side of the lesion after division of one inferior peduncle, or in other words if, as was always the case, the left peduncle was divided, the animal rotated like a right handed screw entering an object. The direction of rotation was thus the same as after intracranial section of the auditory nerve, and the reverse of what results on ablation of one lateral half of the cerebellum. The bulk of the afferent impulses, whose interruption is responsible for this phenomenon, probably reach the inferior peduncle from the auditory nerve, but that all the impulses are not derived from this source was shown by the fact that lateral section of the medulla below the auditory nerve and its nuclei may result in similar rotation. The disorders of motility which followed division of one inferior peduncle corresponded exactly with those observed after ablation of one lateral half of the cerebellum. In view of the results obtained by Claude Bernard, and by Mott and Sherrington, as regards im- pairment of movement after section of sensory spinal roots, it is suggested that the defects of movement which result from section of one inferior cerebellar peduncle may be due to the interruption of such afferent impulses passing to the cerebellum, rather than to the cutting off of efferent impulses from the cerebellum to the spinal centres, The way in which the sensory defects correspond in dis- tribution to the motor, and the fact that recovery of sensory conduc- tion commences before any improvement in motor power can be detected, are held to support this view. Cutting off of some afferent impulses can alone be considered responsible for the ocular displacements met with. These displace- ments correspond with those which are the result of ablation of one lateral half of the cerebellum, the displacement of the globes being downward and to the opposite side from the lesion. The displace- ments following lateral section of the medulla were the same; but after division of the posterior columns and their nuclei on one side, the displacement of the globes was downward and to the side of the lesion. Spasm, which was easily detected in the back and neck muscles on the side of the lesion, causing incurvation of the vertebral axis to that side, alone furnished any satisfactory information in support of the of Afferent and Efferent Tracts of the Cerebellum. 201 possible control which the cerebellum may exert on the spinal centres. The state of the knee-jerks afforded no satisfactory information on this point. The blnnting of sensibility met with is held to be further proof that the cerebellum is concerned with sensory as well as motor pro- cesses, as was contended by the author in a former paper, Faradic excitability of the opposite cerebral hemisphere was found to be less than of that on the side of the lesion, both when the in- ferior cerebellar peduncle was divided, and when partial hemisection of the medulla was performed, leaving the pyramid intact. The most satisfactory explanation of this phenomenon appears to be that the removal of some afferent inhibitory influence from one half of the cerebellum allows this half of the organ to further inhibit the cortex of the opposite cerebral hemisphere ; an explanation in keeping with that offered when the results of ablation of the cere- bellum were under consideration. This view is strengthened by the remarkable results obtained by the intravenous injection of absinthe in animals in whom the same lesions had been previously produced, for with the pyramidal system absolutely intact on both sides, there was an entire absence of con- traction of the muscles of the anterior extremity on the side of the lesion, and diminution of contraction of the muscles of the posterior extremity on this side, as compared with those of the opposite limb. Such was the result obtained when the convulsions were induced soon after the lesion, but when induced at some remote period, such as three weeks after, the muscles of the anterior extremity on the side of the lesion contracted, though the contractions were much less powerful than were those of the opposite anterior extremity, and were often largely tonic in character. Transverse section of the posterior columns, and their nuclei alone on one side, did not alter the character of the absinthe convulsions in such a remarkable manner as did division of the peduncle and lateral section of the medulla. After such a lesion the muscular contractions in the anterior extremity on the side of the lesion were less power- ful than were those in the opposite anterior extremity, and there was more tonus and less clonus than in the contractions on the opposite side. Both these characters were evident in the early convulsion's of a series, but became much more pronounced in the later convulsions. The author contents himself with recording these facts, and makes no attempt to speculate as to their probable significance. The paper is illustrated by tracings obtained of the muscular con- tractions resulting from excitation of the cerebral cortex with the induced current, and from the convulsions evoked by the intravenous injection of absinthe, and demonstrate the points alluded to in that part of the text which deals with these phenomena. 202 Mr. W. Heape. " The Menstruation and Ovulation of Macacus rhesus" By WALTER HEAPE, M.A., Trinity College, Cambridge. Com- municated by Dr. M. FOSTER, Sec. R.S. Received June 15,— Read June 18, 1896. (Abstract.) The specimens used in the following investigation were collected in Calcutta in 1891. Anatomy of the Cervix. — A valve-like structure is formed in the canal of the cervix by means of three strong folds, one of these folds fits into a recess formed by the two other folds, and forms a valve which persists throughout life. It is unlike any other structure of the cervix with which I am acquainted. Breeding. — A definite breeding season for Macacus rhesus seems to be proved, but it is equally certain that in different parts of the Continent of India the breeding season occurs at different times of the year. Menstruation. — A congestion of the skin of the abdomen, legs, and tail, a swelling and congestion of the nipples and vulva, and flushing of the face, are all prominent external signs of menstruation. A regular menstrual flow occurs consisting of a viscid, stringy, opaque white fluid filled with granules, and containing also red blood corpuscles, pieces of uterine tissue, both stroma and epithelium, and also leucocytes. The following classification of the various stages passed through is adopted : — A. Period of rest. Stage I. The resting stage. B. Period of growth. Stage II. The growth of stroma. Stage III. The growth of vessels. C. Period of degeneration. Stage 1Y. The breaking down of vessels. Stage Y. The formation of lacunae. Stage YI. The rupture of lacunae. Stage VII. The formation of the menstrual clot. D. Period of recuperation. Stage VIII. The recuperation stage. The surface of the uterine mucosa, which is smooth and semi- transparent during Stage I, becomes swollen and opaque during Stage II. and flushed during Stage III ; it then becomes highly con- gested, Stage IV, and dark red spots, due to the formation of lacunae, appear on the surface in Stage V ; when Stage VI is reached, free The Menstruation and Ovulation of Macacus rhesus. 203 blood is found in the uterine cavity ; the menstrual clot is formed during Stage VII, and the torn mucosa is healed in the final, Stage VITI. Histology. — The uterus consists of an internal mucosa and external muscular layers ; the mucosa is composed of uterine and glandular epithelium, blood vessels, and stroma. The uterine epithelium lines the surface of the stroma, the glandular epithelium lines pits in the stroma and is continued into branches of those pits which extend from their lower end into the deeper part of the stroma. The stroma itself is a delicate connective-tissue-like layer ; the internuclear protoplasm is drawn out into delicate processes which form a continuous network, and there is no intercellular substance. The histological changes which take place during the menstruation of Macacus rhesus are very similar to those which I have already described in a former paper, *' The Menstruation of S&mnopithecus entellus ('Boy. Soc. Proc.,' vol. 54, and 'Philosophical Transactions,' vol. 185). Work similar to that which I have already described for S. entellus has been undertaken for Macacus rhesus, and the phenomena compared step by step. While it has been thought advisable to note the points of similarity and of difference which occur in the menstrua- tion of these two species, and to point out the fact that the results arrived at by the study of the menstruation of Macacus rhesus entirely confirm the results which my examination of 8. entellus led me to publish, I have purposely avoided all unnecessary repetition and have been obliged in consequence to assume some knowledge of the details given in my former papers. It is all the more important to publish this account, as the results which I have arrived at differ in some important particulars from the accounts of menstruation which have been generally accepted. Stage I. — The mucosa of Macacus rhesus is thicker and the proto- plasmic network denser, the glands more numerous and more branched than is the case in 8. entellus. I find no radial fibres. Stage II. — There is a great increase in the number of nuclei by amitotic division and fragmentation. Hyperplasia occurs. The mucosa becomes much swollen. Stage III. — The vessels increase in number and size, and they are congested. There is an increase of leucocytes. Stage IV. — Hypertrophy of the walls of the vessels in the super- ficial part of the mucosa, followed by degeneration, occurs ; the small vessels break down and extravasation of blood takes place. There is no sign of the migration of leucocytes. Stage V. — Lacunaa are formed at first some distance below the epithelium, but they gradually displace the intervening tissue and come to lie directly below the uterine epithelium. VOL. LX. Q 204 Mr. W. Heape. Stage VI. — The uterine epithelium degenerates and ruptures, and the blood contained in the lacunae is poured into the uterine cavity. Stage VII. — Denudation follows, and the formation of the mucosa menstrualis takes place in the same way and to the same extent as in S. entellus. Stage VIII. — The recuperation takes place as in 8. entellus. With regard to the new uterine epithelium I find fresh evidence in support of my contention that it is formed, not solely from epithelial elements which already exist, such as the torn edges of glands, but also directly from elements of the stroma tissue. Ovulation in Macacus rhesus. — Only one case has been met with in which it can possibly be supposed that ovulation and menstruation have occurred simultaneously ; this is the only case in which a recently discharged follicle was found in the ovary of a menstruating Macacus rhesus ; it does not follow that ovulation in this case was brought about by menstruation ; indeed, the absence of any sign of the recent bursting of a follicle in any other of the seventeen cases examined is in itself strong presumptive evidence that the two pro- cesses are distinct. This result may be confidently asserted for Macacus rhesus during the non-breeding season ; at the same time it must be remembered that I have not investigated Macacus rhesus during the pairing season ; probably at that time ovulation may be more frequent, and may more often be coincident with menstruation ; but, however that may be, menstruation occurs in Macacus rhesus regularly with- out ovulation taking place, and my former views are confirmed, namely, that ovulation does not necessarily occur during each men- strual period, and that it is not necessarily brought about by menstruation. I feel warranted in going further than this and asserting that the regular occurrence of menstruation without ovulation, even though it be in the non-breeding season, is sufficient evidence that ovulation is a distinct process, and that it depends upon a law or laws other than the laws which govern menstruation. The Discharged Follicle. — The changes undergone by the discharged follicles of Macacus rhesus during the non-breeding season are of interest. Very shortly after rupture the follicle is pear-shaped, and the place where rupture took place is to be seen in sections. The wall of the follicle is composed of branched cells which, along the inner edge of the follicle, are longitu dinally disposed and form a denser layer sharply defining the wall from the central cavity. The cavity contains a network of densely granular material and no blood clot. The Menstruation and Ovulation of Macacus rhesus. 205 Hypertrophy now takes place, the wall becomes much thickened and folded, and a growth of cells takes place from the wall into the cavity of the follicle, the sharply marked boundary of the wall is lost, and the long protoplasmic processes of the cells within the cavity are continuous with the cells of the wall. The vessels of the wall now become enlarged and increased in number. Hypertrophy is no longer evident ; the tissue is denser and shrunken, and the whole structure is reduced in size. Gradually the cavity of the follicle is also reduced in size, and the tissue contained therein becomes denser until it is hardly to be distinguished from that composing the wall. * Finally the whole of the cellular remains of the follicle consist of a comparatively small mass of cells with no trace of the follicle wall and no central cavity, a nearly solid mass of tissue, in the midst of which a few blood vessels run. The cells which compose this mass now scarcely differ from the ovarian stroma cells ; they have gradu- ally undergone the change, and instead of branched cells they now appear as polyhedral cells or multinucleated polyhedral protoplasmic masses with intermediate finely branched connective tissue elements bounding them. This structure is surrounded by a layer of fine nucleated fibres ; but soon these disappear, and the remains of the follicle are no longer distinguishable from the rest of the ovarian stroma. Throughout, no trace of a blood clot within the follicle was seen, and therein these ruptured follicles differ from what is usually de- scribed as a normal ruptured follicle in the human female. This difference between two animals, both of which undergo menstruation, is remarkable and worthy of special attention. I have some reason to believe the difference may be due to the presence or absence of the breeding season in monkeys, and to periods in the human female, which are in the one case favourable, and in the other case not favourable, to conception. If this be true, the period of the human female which is unfavour- able to conception would be comparable to the non-breeding season of monkeys, and the period favourable to conception with the breeding season of monkeys. It is not maintained that among civilised peoples at the present day there are definite breeding and non-breeding times, but the com- parison is in harmony with the view that at one period of its exist- ence the human species had a special breeding season. VOL. LX. 206 Drs. W. Ramsay and J. Norman Collie. "The Homogeneity of Helium and of Argon." By WILLIAM RAMSAY, Ph.D., F.R.S., and J. NORMAN COLLIE, Ph.D., F.R.S. Received July 21, 1896. Preliminary. It was pointed out by Lord Rayleigh and one of the authors that it is a legitimate conclusion to draw, from the found ratio between its specific heat at constant pressure and that at constant volume, that argon is a monatomic element (* Phil. Trans.,' 3895, A, p. 235). A similar deduction can be drawn regarding helium (' Chem. Soc. Trans.,' 1895, p. 699). And as the molecular weight of hydrogen is accepted as twice its atomic weight, and as the density of helium is approximately 2, and that of argon approximately 20, the molecular weights of these elements are approximately 4 and 40 respectively. If, however, the molecule is identical with the atom, then the atomic weights must also necessarily be 4 and 40. Bat argon, with an atomic weight of 40, finds no place in the periodic table of the elements, if, as is usual, it is contended that the elements must necessarily follow each other in the numerical order of their atomic weights. Certain suppositions may be made which would obviate this diffi- culty. First, the evidence from the ratio of the specific heats may lead to a false conclusion. But it is inconceivable that any struc- ture, except one of the simplest kind, should transform all energy communicated to it as heat, into kinetic energy of translation. Still, before a final decision on this point is arrived at, it would be well to a.ctually determine the specific heat of argon, and this will shortly be done. It may, however, be mentioned, that preliminary experiments have shown it to be much lower than that of hydrogen, air, or carbon dioxide, volume for volume. Second, helium and argon may consist of a mixture of monatomic with diatomic molecules. The perfectly normal expansion of these gases appears to negative this supposition ('Phil. Tra'ns.,' loc. cit., p. 239, and « Roy. Soc. Proc.,' vol. 59, p. 60). Even at a tempera- ture of — 88° there appears to be no marked tendency towards association. It is true that the ratios of the specific heats do not quite reach the theoretical number I1 66 7. That found for helium was 1'652, and that for argon T659, with the most carefully purified samples. Assuming (what there seems good ground to doubt) that the last decimal place may be trusted, helium can be calculated to contain nearly 7 per cent, of diatomic molecules, and argon rather more than 3 per cent. If this calculation be permitted, the atomic weight of helium would become 4'02, taking its found density at The Homogeneity of Helium and Argon. 207 2*15, and of argon 38'62. This would place argon below potassium, the atomic weight of which is 39'L However, it must t>6 acknow- ledged that such refinements in calculation are far from trustworthy. Third, helium and argon may each consist of a mixture of two or more elements. This view has been expressed with regard to helium by Professors Bunge and Paschen (« Sitzungsber. d. Akad. d. Wissensch.,' Berlin, 1895, pp. 639 and 759), on the ground that the lines of its spectrum can be shown to belong to two distinct series. The question whether argon is a mixture or not was discussed in the memoir by Lord Rayleigh and one of the authors (Zoo. cit.9 p. 236). It is with this possibility that the present communica- tion has to deal. Two methods suggest themselves as suitable in order to ascertain whether argon and helium are mixtures of two or more elements, or are single elements. The first is fractional solution in water; the second fractional diffusion. The second method is obviously the better calculated to yield the desired data ; for if these gases contain constituents of different density, diffusion is an infallible means of separating them. Description of Diffusion Apparatus. After a number of trials, the stem of an ordinary tobacco-pipe was found to yield the best results. Plaster of Paris is too porous, and various forms of graphite tried did not effect so rapid a sepa- ration of two known gases as unglazed clay. In fact, nothing could have been more satisfactory than this apparatus. It consists of a reservoir for the gas, A, into which projects a piece of the stem of a tobacco-pipe, B, sealed at the lower end in the flame of an oxy-hydrogen blowpipe. When the stop-cock C is open, and D and E shut, the gas in A must pass through the pipe-clay tube on its way to the reservoir of the pump F. The fall of the mercury in the tube G, read on the scale H, is timed, about 8 cm. fall being taken as sufficient for the purpose. The mercury rises in A, and falls in the reservoir I during the diffusion. When the experiment is finished, the gas is pumped out of the reservoir F, and collected in tubes similar to that depicted at K, and stored in a frame resembling a miniature umbrella-stand. The stop-cock D is then opened, and the clip L is shut, and the less diffusible portion of the gas is pumped out and collected in other tubes, and set apart. The purity of the gas is ascertained by means of the vacuum tube M. After all gas has been removed, the stop-cocks C and D are shut; a new charge of gas is introduced at N, the stop-cock B being opened, and the operation repeated. After a sufficient amount of the first diffusate has been collected, it is. again introduced into the reservoir A, and the process repeated. E 2 208 Drs. W. Eamsay and J. Norman Collie, When towards the end only a small amount of gas is available, the process may be modified by raising the reservoir I, and so dimi- nishing the volume of A. The clip L is then closed, and the gas is allowed to diffuse as before, but the volume in A is kept constant. The rate of diffusion can be compared with that of hydrogen under precisely similar circumstances. In all the experiments the temperature did not alter by more than a degree or two ; as the object was to effect a separation, and not to make accurate determinations of the rates of diffusion of gases, careful regulation of temperature was unnecessary. Determination of the Ratios of Diffusion of Gases of known Purity. (a) Hydrogen. — The time required for the column of mercury in H to sink through 8 centimetres, starting always from the same level, was found in three experiments to be (1) 433", (2) 420", and (3) 437" ; the mean is 430". The average rate per millimetre is 5'37". (6) Oxygen. — The time which pure oxygen, made from permanga- The Homogeneity of Helium and Argon. 209 nate, took to diffuse to the same extent was 1719", giving an average rate per millimetre of 2T49". (c) Acetylene. — The gas was prepared from pure calcium carbide by the action of water. It dissolved completely in alcohol. The time required for diffusion was 1550", giving a rate per millimetre of 19-37". Assuming the times for the diffusion of these gases to be pro- portional to the square roots of their densities, we have — For oxygen _ _ 2l-39". Found 21-49". \/l-0082 For acetylene 5'37"* ^13'008 = 19-29". Found 19-37". A/I -0082 This process may therefore be trusted to give fairly accurate results when applied to test the rates of diffusion of gases of known purity. The Separation of a Mixture of Gases. To ascertain whether a separation could be easily effected, experi- ments were made (a) on a mixture of oxygen and carbon dioxide, and (6) on a mixture of hydrogen and helium. (a) Oxygen and Carbon Dioxide. — The original mixture contained 36 per cent, by volume of carbon dioxide. It was split into two approximately equal portions ; each of these was again split into two. The most diffusible part contained 30'2 per cent, of carbon dioxide, and the least diffusible part 41 '0 per cent. (6) Hydrogen and Helium. — The original mixture contained 50 per cent, of each gas, and its volume was 38 c.c. 19 c.c. were diffused ; this was again halved, 9*5 c.c. being passed through the pipe ; and finally another diffusion of the 9'5 c.c. yielded 4*12 c.c. of mixed gases. The hydrogen was removed by explosion with oxygen. This mixture now consisted of 67 per cent, of hydrogen and 33 per cent, of helium. From these experiments it is seen that a partial separation of such gases is easily carried out. The Fractional Diffusion of Argon. Four hundred c.c. of argon, newly circulated over red-hot magne- sium until spectroscopic traces of nitrogen were carefully removed, was diffused according to the subjoined scheme : — 210 Drs. W. Ramsay and J. Norman Collie. More diffusible. I Less diffusible. The densities were determined by weighing. These numbers show that no important separation has been effected. The difference in density of the two portions may possibly be attributed to experimental error. When the density of the heavier portion was taken the weather was damp, and we have found it difficult to obtain concordant results under such circumstances, owing doubt- less to the uneven deposition of moisture on the surfaces of the bulb and its counterpoise. But as it stands, the difference is an extremely minute one, and it may, we think, be taken that any separation of argon, if effected at all, is very imperfect. The Fractional Diffusion of Helium. Two hundred c.c. of helium from fergusonite of density 2'13 were separated into two nearly equal portions by diffusion. The rate of diffusion was 7*14" per millimetre as a mean of two experiments, giving 7'13" and 7'15" respectively. The most diffusible portion of this gas gave the rate 7'12" per millimetre. The more diffusible half of this gas had the rate 7'48", and the least diffusible of the remainder 7 '38", the temperature being lower. A second specimen of helium from mixed sources, samarskite, fergusonite, broggerite, &c., which showed the nitrogen spectrum strongly, gave a rate for the first portion of 8'29". This half on rediffusion had the rate 7'64", and the residue of 8'39", showing that a separation was being effected. The heavier residue of the remainder from that portion which showed the rate 8 "39" was too small to make it possible to diffuse it by the usual method. A second method was therefore resorted to, and it was directly compared with hydrogen under the same circum- stances. While hydrogen took 12'14" per millimetre, the residue took 21*00", and calculating its density from these rates, we have — 21-00")2x 1-0082 QAO (12-14'T This would correspond, if it be granted that the impurity is nitro- gen, to a percentage of 8'5 of that gas. This residue showed a The Homogeneity of Helium and Argon. 211 strong nitrogen spectrnm ; and the nitrogen was removed by sparking with oxygen in presence of soda, until the spectrum attested its absence. (It will be remembered that 0*01 per cent, of nitrogen is still visible under moderate pressures, * Eoy. Soc. Proc.,' vol. 59, p. 265.) The rate was again measured against that of hydrogen under pre- cisely similar conditions, and it was found that while hydrogen took 20*00" for diffusion, this specimen of helium took 28*28". And calcu- lation shows its density to be now 2*015. These experiments were sufficient to show, we think, that while it is possible to separate nitrogen from helium, even although the former is present in only small amount, we had not succeeded in separating helium itself into two portions of different densities. If, then, helium were a mixture, its constitutents must possess nearly the same density. In no case was any alteration of the spectrum to be noticed ; the diff usate and the residue were similar, and showed all the well known lines of helium with the usual intensity. But it was deemed advisable, in view of the importance of the matter, to undertake a much more elaborate set of experiments. The helium was carefully purified from hydrogen and nitrogen by circulation over magnesium, copper oxide, phosphorus pentoxide, and soda lime, until a small quantity admitted into a vacuum tube in connection with the circulating apparatus showed no spectrum either of hydro- gen or nitrogen, even at a comparatively high pressure, when these gases are more easily detected. The helium was then fractionated in a manner exactly similar to that shown in the graphic scheme for argon (p. 210). The rates of diffusion of the two samples of gas were then measured. More diffusible portion — Time of diffusion reduced to 0° 662*5" Hydrogen 492*3" Density, calculated from rate ....,,.. 1*826 Less diffusible portion — Time of diffusion 654'9" Hydrogen, at same temperature 484*4" Density, calculated from rate 1*842 The density of hydrogen was taken as 1*0082, on the standard, oxygen = 16. These samples were next weighed. More diffusible portion — Volume of globe 16*2*843 c.c. Pressure at filling 668*5 mm. Temperature 19*20° Weight 0*02450 gram Density 2'049 212 Drs. W. Ramsay and J. Norman Collie. Less diffusible portion — Volume of globe 162*843 c.c. Pressure at filling 663'8 mm. Temperature 19'93° Weight 0-02902 gram Density 2'452 Tlie less diffusible portion was next subjected to the process of removing nine-tenths, the remaining tenth being collected apart. This process was repeated three times, so that any portion of gas less diffusible than the main bulk should thus be left as a residue. From the more diffusible portion nine-tenths was also diffused out. The more diffusible portions were then mixed, and the density was again determined. Volume of globe 162'843 c.c. Pressure at filling 765' 7 mrn. Temperature 20'98° Weight 0-02801 gram Density „ . 2*057 This number is practically identical with that previously obtained, viz., 2-049. It was of interest to follow the less diffusible gas, so as to ascertain what impurity caused its higher density. Another set of fractiona- tions was therefore carried out, and after five separate processes, in each of which a residue was left, and that residue further diffused, so as to separate all light gas as completely as possible, a few c.c. of gas were collected, in which the spectrum of argon was strong. Now we are certain that at no stage in the operations was any con- siderable quantity of air admitted by leakage. It may safely be said that the total amount of air could never have exceeded 5 c.c. And inasmuch as the density of samples of helium from various sources, which had undergone very little handling, differed by small amounts, varying between 2*114 and 2'181, this must be ascribed to contami- nation with argon, contained in the mineral from which the helium had been obtained. Every effort was made to detect any unknown lines in the spectrum of the residue, but in vain. With the jar and spark-gap, the blue spectrum of argon was visible, and was compared with that from a standard tube. If thus the increased density is due to argon, it is possible to calcu- late the proportion of the latter ; first, in the lightest gas of density 2*117 found in samarskite ; second, in the residue in which the argon had been concentrated, possessing the density 2*452, on the assump- tion that helium possesses the density 2'042. The first must contain 0*42 per cent, of argon ; the second, 2'28 per cent. The Homogeneity of Helium and Argon. 213 The rate of diffusion of the gas of density 2*057 was determined .finally, so as to afford a check on its density. It took 657*9" for a quantity to diffuse ; while the same volume of hydrogen under pre- cisely similar circumstances took 492'3". Reducing these numbers to density, if hydrogen be taken as 1*0082, the helium possesses the density 1*801, which compares very favourably with the number already found, 1*826. As a final check on these results, a sample of helium from an entirely different source, samarskite, was so diffused, that first nine- tenths were removed by diffusion ; from the residue nine- tenths was again removed, and the process was repeated a third time. The more diffusible portion was tested as regards rate; while hydrogen took 492*3" to diffuse, this sample took 652*6". Stated as density, ths number is 1*771. The actual density was next determined, with the following result : — Volume of globe 162*843 c.c. Pressure at filling 691*6 mm. Temperature 19*85° Weight 0-02567 gram Density 2*080 This number closely coincides with the density of the previous specimen, freed from argon by diffusion ; and in this case it must be remembered, no systematic process for separating two possible con- stituents was carried out, but the heavier portion only was removed. The heavier gas separated by diffusion was examined for argon, and it was possible to see the green group of five lines, but not the red lines. And with a jar and spark-gap, argon could just be detected. The rate of diffusion of this gas, which, stated as density, gives the number 1*8, differs from the density determined by weighing, viz., 2*08, or thereabouts. This might be caused (1) by a lighter portion passing over first during diffusion, leaving a heavier portion behind •. or (2) by the hypothesis that the rate of diffusion of helium is ab- normal ; and helium has already shown such very remarkable pro- perties in relation to refractivity for light, and conductivity for electricity, that the hypothesis is not unwarrantable. The first supposition, however, is the more probable, and was put to the test in the following manner. A smaller apparatus was made for measuring the rate of diffusion of 10 to 20 c.c. of gas ; and the rates of the sample of density 2*08, and of the less diffusible residues from this sample were determined. Both the hydrogen and the helium were carefully measured and diffused under precisely similar conditions. While the hydrogen took 181" to diffuse, the helium of density 2*08 took 246*6", implying a 214 Drs. W. Ramsay and J. Norman Collie. density of T871 ; and the residue diffused in 266' 6", which corre- sponds to a density of 2%187. In each of these experiments about half the helium passed through the porous plug. The denser portion of this gas was again diffused five times, lighter portions being removed. This corresponds to a residue of 30 c.c. from 400 c.c. of the original gas. The rate of diffusion of this sample compared with that of hydrogen was almost identical with the last, namely 208" to 143", and corresponds to a density of 2*133. The gas is therefore not increased in density by this process. The lighter gas was submitted to a similar fractionation, and the ratio of its diffusion-rate to that of hydrogen was 24675" to 181 -0", as a mean of several closely concordant experiments. This corres- ponds to a density of 1*874. We have accordingly : — Density. " Heavy " portion 2*133 " Light " portion T874 Not content with this, we pushed fractionation still further ; the helium was divided into seven portions (by fractionation) and then submitted to methodical fractional diffusion, in which the heavier portions were transferred to the " denser " side, and the lighter portions to the " lighter " side. This process was repeated four times, and the end portions were each divided into two ; the lighter portion of the "lighter" was collected separately, and its rate determined. It took 258*5" to diffuse, compared with 189*5" for an equal volume of hydrogen ; its density calculated from these rates was 1*876. It is clear, therefore, that the limit has been reached in purifying the lighter portion by diffusion. It should have been mentioned that the portion of 2*133 density as well as that of T874 density had been sparked with oxygen in presence of potash, and in a vacuum tube showed mere traces of hydrogen, every other gas being absent. The spectrum of hydrogen is still visible, even when 0*01 per cent, of that gas is present. At various times during the attempt to separate helium, the spec- trum has been carefully examined. The very first portions of the lightest gas gave an identical spectrum, seen with a hand-spectro- scope, with the very last portions of the heaviest gas. Professor Ames, of the Johns Hopkins University, has however kindly undertaken to photograph the spectra using a dispersion -grating ; so that if any difference can be detected, it will ere long be made known. Lord Rayleigh was so kind as to measure the refractivity of these extreme portions of the fractionated gas. His process has been described in the ' Proceedings,' vol. 59, p. 202. For the sample of helium sent him in July, 1895, he found the number 0*146. The lighter portion of the fractionated gas of density 1*876 had a refrac- The Homogeneity of Helium and Argon. 215 tivity, compared with air as unity, of O1350 ; the heavier portion, of 0*1524. The ratio of these numbers is very nearly that between the densities of the gases, viz. : — 0*1350 1-876 , 1-876 00824 = 2018' m8tead °f MM ' Conclusion. It must be remarked that the rate of diffusion of helium is too rapid for its density measured by weighing. There can be no doubt, we think, that the density of the lighter portion, instead of being 1-874, would be, if actually weighed, 2'05 or 2*08. And the heavier portion has doubtless a proportionately higher density. But, assum- ing that the densities calculated from the diffusion-rates are correct, the densities of the two gases, supposing that two exist, are T871. and 2*133, respectively. Also, we must not omit to state that careful experiments were made with the more rapidly diffusing gas to prove that the first portions passing over did not diffuse at a more rapid rate than the later por- tions, no difference in diffusion rates, compared with those of hydrogen under the same circumstances having been detected. That helium, then, consists of a mixture of two or more distinct gases is one solution of the problem, probably the one which recom- mends itself at first sight. But there is another, so revolutionary in its character that much must be done before it can be regarded as even worthy to be entertained. So much has, however, been lost to science by what may be termed scientific incredulity, that we regard it as well worth putting to a rigorous proof. It is that a separation has been effected of light molecules from heavy molecules ; that, in fact, a gas — in this case helium — is not constituted entirely of molecules of the same weight, but that the mixture of molecules which we term helium have weights which average 2"18, or whatever the density of ordinary undiffused helium may ultimately be found to be. The same supposition would, of course, be applicable to oxygen, nitrogen, or any gas. In separating such molecules from each other a practical limit must necessarily be reached, and this limit appears to have been reached with helium. There is negative and positive probability in favour of this sug- gestion. First, no gas has been submitted to methodical diffusion with a view to effect such a separation, argon excepted ; and here, too, there is faint evidence of a similar kind. It is proposed to carry out similar experiments with gases of undoubted homogeneity according to the usual views ; and till such experiments have been made, it is impossible to decide the point definitely. Second, Mr. E. C. C. Baly's experiments on oxygen appear to 216 Prof. W. N. Hartley. On the Spectrum of Cyanogen point to a similar conclusion ; although no great alteration in density has been produced, yet there is a sign that a kind of separation is being effected electrically. There is also in favour of the supposi- tion the unlikelihood that two or more gases, so like one another as the constituents of helium, should exist with densities so near each other; and the probability that some separation should have been detected by aid of the spectroscope. Lastly, the refractivities of both gases, if there be two, appear to be equally abnormal ; now, different gases have different refrac- tivities in no known relation to their densities, as, for example, hydrogen O5, oxygen nearly 1. But the refractivities of the dif- ferent portions of helium are proportional to their densities ; a statement which is true of any one gas, inasmuch as refractivifcy is directly proportional to pressure, i.e., mass in unit volume. The refractivity of helium, also, is so small that it totally differs in this respect, as, indeed, it does in most of its physical properties from every other gas, and it is moreover a monatomic gas. Tt is therefore permissible to seek for an explanation of its remarkable properties in framing any hypothesis which admits of being put to the test. "On the Spectrum of Cyanogen as produced and modified by Spark Discharges." By W. N. HARTLEY, F.R.S., Royal College of Science, Dublin. Received July 13, 1896. The Production of Cyanogen in the Electric Arc. — The very careful and numerous experiments of Liveing and Devvar* have very generally been accepted as affording evidence sufficient to establish the existence of an emission spectrum of cyanogen as distinct from that of carbon in the electric arc. Kayser and Runge,f though at first disinclined to accept such a conclusion, obtained additional evidence by experimenting with the arc in air, and in carbon dioxide. They found that the ordinary carbon spectrum and that of cyanogen appeared with rapidity alternately in the arc in air, though there could be no difference in temperature sufficient to account for the production of two different carbon spectra. With the poles immersed in carbon dioxide no such changes were seen, the carbon spectrum alone being visible, which evidence led them to concur in the views of Liveing and Dewar. The chief evidence of the existence of a cyanogen spectrum rests on the fact that this substance is actually synthesised in the arc when nitrogen is present, and because without * ' Roy. Soc. Proc.,' vol. 30, pp. 152—162, 494—509 : vol. 34, pp. 123—130 and pp. 418—429. f " Ueber die Spectren der Elemente. Z welter Abschnitt. Ueber die im galva- trischen Liclitbogen auftretenden Bandenspectren der Kolile." ' Abh. K. Preuss. Ak. Wiss.,' 1889, p. 9. as produced and modified by Spark Discharges. 217 nitrogen, elementary carbon does not yield the same spectrum, no matter what the temperature may be ; and lastly, that cyanogen gas burns with a flame of which the banded spectrum is known as that uf cyanogen by reason of the foregoing facts. Furthermore, I have found by recent experiments that when a condensed spark is passed between electrodes of gold in an atmosphere of cyanogen, the same spectrum is photographed. If we admit that under conditions favourable to synthesis from its elements, cyanogen is capable of emitting a spectrum of its own, this emission should occur only at the moment of its formation, but while giving consideration to this view we are met by the difficulty that the flame of cyanogen burning in oxygen would less probably emit a spectrum of the compound substance itself, which is being burnt, than a spectrum of the products of its combustion, or of the separated elements of which it is composed, which are nitrogen and carbon ; and for this reason, that the process it is passing through is not a synthetical but an analytical one. Indeed it has been shown by Liveing and Dewar* that when cyanogen is exploded with oxygen it gives a bright continuous spectrum, but no cyanogen spectrum, or carbon bands, or carbon lines. I shall have to refer to these facts and adduce later evidence of the existence of the cyanogen spectrum in the latter part of this paper. Evidence derived from their Spectra, of the progress of Chemical Changes in Flames. — In support of the view that the flame of burning cyanogen ought to exhibit the spectrum of carbon, I may mention the following facts which have been recorded during a very careful examination of a number of photographs of the spectra of flames which were obtained by burning gases under normal atmospheric conditions. The majority of these photographs were taken in 1882. The Combustion of Compound Sub- Components of the Spectra photo- stances, graphed. Hydrocarbons in oxygen. Carbon bands, cyanogen bands, water- vapour lines.*!" Sulphuretted hydrogen in air and in Sulphur bands and water-vapour lines, oxygen. Ammonia in air. Water-vapour lines. Carbon disulphide in air. Sulphur bands only. Carbon disulphide and nitric oxide. Sulphur bands only. Carbon monoxide and oxygen. Continuous spectrum of carbon mon- oxide. Faint lines due to carbon, very few in number. * l Roy. Soc. Proc.,' vol. 49, p. 222. " On the Influence of Pressure on Flames." f When nitrogen is present, Liveing and Dewar have observed the formation of N02 (loc. cit.). 218 Prof. W. N. Hartley. On the Spectrum of Cyanogen By the combustion of ammonia in oxygen, water vapour lines are produced, and new bands and groups of lines attributed by Eder and Valenta to ammonia. Some of these are, however, due to a compound other than ammonia. It will be observed that compounds during combustion as a rule show the spectra of one or other of their constituents, or of both. In the case of hydrogen compounds they show the product of the com- bustion of hydrogen, which is a substance of great stability, and can therefore exist at a high temperature. In the nitric oxide and carbon disulphide spectrum, the sulphur bands, which are very strong, probably obscure those of carbon. There is a strong continuous band of rays which would likewise serve to obscure them. C. Bohn* has examined the spectra seen in a Bunsen burner of the form devised by Tecluf (which is simply a modification of that described by Smithells), and compared the spectra with that obtained by Swan, and with the discharge in Geissler tubes containing various hydrocarbon gases. He concludes that it is impossible to define a carbon band spectrum, as the differences observed were greater than could be accounted for by alterations in temperature and pressure. He also states that sulphur, hydrogen, and carbon disulphide, also carbon monoxide, were burnt, but that all these flames yielded con- tinuous spectra. This statement is incorrect, or at least inaccurate. J Bohn's observations were evidently made on too limited a region of the spectrum, and without the aid of photography. On Bohn's paper Eder has made some observations, quoting both his measurements in the visible and ultra-violet spectrum, which he observes must have been unknown to Bohn.§ He describes in what manner and by what causes the edges of the carbon bands are altered in position or in character. The observations of Eder on the spectra of hydrocarbon flames are quite in agreement with those previously communicated by me to the Royal Society on the oxyhydrogen flame spectrum and the oxy-coal gas spectrum. On certain Chemical Changes occurring in the Spark and in Flames. Though it is now accepted as a fact that the arc in air yields the spectrum of cyanogen, and that the evidence of this is, first, the identity of certain bands observed in the flame of burning cyanogen * ' Zeitschrift fur physikal. Chemie,' vol. 18, p. 219, 1895. f ' J. prak. Chemie ' [2], vol. 52, pp. 145—160, 1895. J " .Flame Spectra at High Temperatures " (' Phil. Trans.,' A, vol. 185, pp. 161 —212, 1894). § " Ueber Flammen und leuchtende Q-ase " (' Zeitschrift fur physikal. Chemie, vol. 19, p. 1, 1896). as produced and modified by Spark Discharges. 219 with those seen in the arc ; second, that these bands cannot be due to the effect of an alteration in temperature, giving rise to a second spectrum of carbon ; nevertheless, as I have elsewhere pointed out,* cyanides in a condensed spark do not produce this spectrum, no matter whether they are extremely stable cyanides, such as that of potassium, or those of the most easily decomposable character, such as mercuric cyanide. This appeared to me to mark the inadequacy of the facts derived solely from observations on the arc, to establish the existence of a definite cyanogen spectrum. Moreover, it was shown that lines somewhat resembling the edges of cyanogen bands are seen when graphite poles are moistened with water and the spark is passed through air; these lines are intensified and developed into bands when the water contains ammonium chloride, calcium chloride, or zinc chloride, and the bands become stronger as the solu- tion used is more concentrated. If the lines observed are the edges of bands belonging to the cyanogen spectrum, by what means do the chlorides give rise to their production ? No one has yet supplied the answer to this question, neither has it been proved that these lines in the spectrum of graphite are the edges of cyanogen bands, though Ederf and Valenta state that they are such because the wave-length measurements are approximately the same. I believe that I am now able to offer an explanation of the action of the concentrated solutions of chlorides, and to prove in addition, that the bands and lines are really due to cyanogen and not to ele- mentary carbon. If hydrochloric or any other mineral acid be carefully tested, it is found to contain ammonia. The only ammonia- free acid is sulphur- ous acid freshly prepared by passing sulphur dioxide gas into water, carefully freed from ammonia and from any possible contamination with it. If from the usual samples of so-called pure mineral acids, salts of calcium or zinc be prepared, the ammonia salt present is not eliminated, but it goes into solution and crystallises out with such calcium or zinc compound, or, if the salt does not crystallise, it remains in solution, and, as a consequence, the salt will show in its solution the effect of a larger proportion of ammonium salt, accord- ing to its degree of concentration. Hence if the bands, said to be cyanogen bands, are due to the nitrogen of the ammonia present, the spectrum of the graphite poles will exhibit the bands more strongly, as there is less water in the solution. But this does not account for * ' Phil. Trans.,' vol. 175, p. 49, Part I, 1884, and ' Boy. Soc. Proc.,' vol. 55, p. 344, " On Variations observed in the Spectra of Carbon Electrodes, and on the Influence of one Substance on the Spectrum of another." t 'Wien, Akad. Wiss. Denkschriften,' vol. 60, 1893, "Line Spectrum ,of Elementary Carbon." 220 On the Spectrum of Cyanogen produced by Spark Discharge. the fact that the spark does not show the- cyanogen bands when cyanides are submitted to its action. In this case it is possible that the temperature is too high, and that the cyanogen is decomposed, possibly by oxidation, for there can be no doubt that such condensed sparks are at a higher temperature than that of the arc. We know, too, that several metals are oxidised when volatilised in the spark, if not entirely at least partially.* But by using gold electrodes with the cyanides we do not obtain even a carbon spectrum. Here again, possibly, the carbon is oxidised, and we know that carbon dioxide in carbonates yields no spectrum of carbon, nor any lines peculiar to carbon dioxide. I have sought in every direction for a reasonable explanation of that which, up to the present, has proved inexplicable, in order that by working on some hypothesis one might devise a means of putting the matter to experimental proof. This has now been accomplished in the following manner. An almost saturated solution of pure crystallised potassium cyanide was put into a tube fitted with graphite electrodes in the manner described in a previous communication.f The apparatus was fitted into a horizontal wooden tube with a window of quartz at one end, and carbon dioxide was passed into the tube until filled. The spark was then passed for five minutes, and again for ten minutes, a photograph being taken of the two spectra. The instrument used gave a dispersion equal to four quartz prisms. A glass tube with a similar window of quartz was fitted with gold electrodes and filled with cyanogen gas, and another spectrum was photographed. A fourth spectrum was then obtained by passing cyanogen into the wooden tube containing the graphite electrodes ; after the carbon dioxide had been expelled by air and replaced by cyanogen, the Ll'tube was filled up with the solution of potassium cyanide. In all four cases the principal group of the cyanogen bands was obtained, but it was not very strong. A flame of cyano- gen was then photographed with exposures varying from one to two, five, and ten minutes. A beautiful series of spectra was obtained, and the lines belonging to the edges of bands constituting the prin- cipal group were found to coincide exactly with those photographed from the potassium cyanide solution when the spark was passed in an atmosphere of carbon dioxide and in cyanogen, also when the spark was passed between gold electrodes in cyanogen. These appear to be the bands referred to by Eder and Valenta, which were described as carbon lands% when graphite electrodes were used with the spark * ' Boy. Soc. Proc.,' TO!. 49, p. 448, " On the Physical Characters of the Lines in the Spark Spectra of the Elements." f ' Phil. Trans.,' vol. 175, p. 49, 1884. J Hartley and Adeny, ' Phil. Trans.,' vol. 175, p. 63, Part I, 1884. Variation in Portunus depurator. 221 in air. From the modification of their appearance, and the measure- ments originally made from them, their identity was not quite apparent, although probable. It thus appears that, with the spark, the cyanogen spectrum is nothing like so strongly marked, as is the case with the flame of the gas, only one group of bands being represented, and that when the spectrum is taken in air the cyanogen does not appear, because in all probability the substance is oxidised. I have already stated that the formation of cyanogen which yields the characteristic spectrum is a synthetical operation, that compound substances, when burnt in flames, do not, as a rule, emit the spec- trum of the compound, but the spectrum of one or more of the elements of which it is composed, or that of one or other of its products of combustion. How then are we to account for the cyanogen spectrum in the flame of burning cyanogen ? The conditions under which combustion takes place are these : there is an excess of the gas, the temperature of the flame is exceed- ingly high, and the gas within it is not in contact with a solid sub- stance, hence immediate decomposition does not occur, and the gaseous compound is heated to incandescence. " Variation in Portunus depurator" By ERNEST WARREN, B.Sc., Demonstrator of Zoology at University College, London. Communicated by W. F. R. WELDON, F.R.S. Received July 1, 1896. The following measurements were undertaken at the proposal of Professor W. F. R. Weldon, and to him I am greatly indebted for many suggestions, and for the kindly help he has always so readily given me. The crabs were obtained from the Biological Station at Plymouth, and sent at intervals during a period of about two years, dating from the autumn of 1893. Only males were measured. Seven measure- ments were made on each individual, corresponding to those made by Professor Weldon on the female of Garcinus moenas (' Roy. Soc. Proc./ vol. 54). 1. Carapace length, AB (fig. 1). 2. Total carapace breadth, CO'. 3. Frontal breadth, DD'. 4. Right antero-lateral, AC. 5. Left antero-lateral, AC'. 6. Right dentary margin, CD. 7. Left dentary margin, C'D'. VOL. LX. R 222 Mr. E. Warren. The total number of crabs measured was 2300. The determina- tions were made with compasses and a 3-decimetre ivory scale divided into half millimetres. The measurements were recorded to the tenth of a millimetre. As a test of accuracy, fifty crabs were indiscrimi- nately taken out of a large number which had previously been measured; these were remeasured, and the results compared with those before obtained. It was found that the mean difference between any two measures of the same dimension was 0'107 mm. FIG. 1. — Portunus depurator -l CO CO CO IP II 1 s i I iO •* -H O CO O ,-H CO ^ O O -H .2 if s § ? ? i I II ' S c^ I J III -« ' tt i— 1 10 0 ia s CD CO £ £ S S8 S5 — 05 £ i 0 £ •—i i— i 10 I—I 05 of carapace length. a 1 I X 1 lO i 0 i s 2 ci 1 CD i i I g CD I I X I X 1 Total breadth. rf oo CO 0 i-H 05 GO * CD »o * CO * 1— 1 O iH (M CO ; 10 CO t- X 05 2 1370—1367 17 1 2 17 1366—1363 16 16 1362—1359 15 1 1 15 1358—1355 14 2 1 14 1354—1351 13 1 1 1 13 1350—1347 12 1 2 i 1 2 1 12 1346—1343 11 1 1 i 2 4 1 1 11 1342—1339 10 1 3 2 2 5 1 2 1 10 1338—1335 9 1 1 4 3 5 8 4 1 9 1334—1331 8 1 3 4 5 8 5 5 1 2 i 8 1330—1327 7 2 1 2 6 8 7 12 4 1 7 1326—1323 6 I 2 6 11 15 10 4 6 2 6 1322—1319 5 1 3 4 5 10 10 15 7 7 2 1 5 1318—1315 4 5 15 11 12 17 14 7 5 1 4 1314—1311 3 1 6 14 11 17 29 10 8 4 2 1 1 1310—1307 2 1 1 2 3 8 11 14 22 21 9 10 6 2 2 1306—1303 1 1 2 8 6 15 8 13 14 14 12 8 5 3 1 1 1302—1299 0 3 2 8 16 26 18 21 7 4 3 0 1298—1295 1 1 2 1 7 7 19 16 23 22 20 10 3 1 1294—1291 2 1 3 4 11 9 22 18 16 5 1 2 1290-1287 3 1 2 3 3 7 13 15 16 13 15 6 5 1 3 1286—1283 4 4 5 15 12 12 9 13 4 2 4 1282—1279 5 1 1 4 4 6 8 17 11 7 2 5 2 1 5 1278—1275 6 3 3 7 9 14 6 7 5 2 6 1274—1271 7 4 4 6 9 9 4 2 1 1 7 1270—1267 8 1 2 3 5 1 7 5 2 2 1 1 8 1266—1263 9 2 1 4 2 4 5 2 2 9 1262—1259 10 2 2 1 2 1 4 1 10 1858—1255 11 3 1 1 1 2 3 11 1254—1251 12 1 1 1 1 1 12 1250—1247 13 2 1 13 1246—1243 14 1 14 1242—1239 1.5 15 1238—1235 16 1 1 16 1234—1231 17 17 1230—1227 18 1 18 0 05 GO ^ CD 10 ^ 00 w ,_, 0 i— i N co >* IO CD ^ X 05 o r-> 1— { 236 Mr. E. Warren. II. Correlation Surface of Total Breadth and R. Dentary. 1432 Individuals, Measurements in thousandths of carapace length. i 1 O5 ! ? I 2—455 g I 1 T X 7 05 •~o T 1 rH (N T CO g ! rH rH VO 2—515 01 rH cc g T § i X rt 1 3 1 i % J§ ® "3 1 § * 3 3 § § § § § rH lO rH 10 o 3 GO O5 O rH rH rH 2 co rH 1370—1367 17 1 2 17 1366—1363 16 16 1362—1359 15 1 i 15 1358—1355 14 2 1 14 1354—1351 13 1 1 1 13 1350—1347 12 i 1 3 i 2 12 1346—1343 11 1 2 3 1 1 2 1 11 1342—1339 10 2 3 1 1 3 3 2 2 10 1338—1335 9 1 3 3 3 5 1 5 3 1 2 9 1334—1331 8 1 3 2 6 3 3 2 4 8 3 8 1330—1327 7 i 3 4 4 2 8 6 4 2 3 1 7 1326—1323 6 1 1 3 2 7 10 10 5 6 5 5 1 1 6 1322—1319 5 2 i 1 4 2 4 10 12 9 7 3 5 1 3 1 5 1318—1315 4 rj 5 8 7 12 13 12 12 9 2 2 2 1 i 4 1314—1311 3 i 3 5 10 8 26 13 14 9 8 2 1 1 1 1 3 1310—1307 2 1 2 i 4 6 18 13 17 13 10 12 8 4 1 2 1306—1303 1 1 1 5 5 7 7 16 6 15 11 8 7 5 3 4 2 2 1 1302—1299 0 7 6 7 21 7 17 14 7 11 6 4 1 0 1298—1295 1 1 4 4 9 7 10 17 22 17 17 8 5 5 4 1 1 1294—1291 2 3 3 6 10 12 12 14 17 5 5 2 1 2 1290—1287 3 1 1 1 5 5 12 10 8 18 11 10 5 5 5 1 2 3 1286—1283 4 1 2 6 3 11 12 7 7 9 8 4 2 2 2 4 1282—1279 5 1 1 3 6 7 12 9 6 9 4 3 2 2 3 1 e tj 1278—1275 6 3 4 4 3 711 8 7 5 3 1 6 1274-1271 7 1 4 9 5 4 5 5 4 1 1 1 7 1270—1267 8 1 1 2 2 5 5 2 1 2 4 2 2 1 8 1266—1263 9 1 1 1 2 4 3 4 1 1 1 1 1 1 g 1262—1259 10 1 3 2 3 1 1 1 1 10 1258—1255 11 1 1 2 1 3 2 1 11 1254—1251 12 1 1 1 1 1 12 1250—1247 13 1 1 1 13 1246—1243 14 1 14 1242—1239 15 15 1238—1235 16 1 1 16 1234—1231 17 17 1230—1227 18 1 18 0 I— 1 05 00 * CD 10 * co <0 0 T* CO rH U5 CO iH p OT8T— Z08T Z rHrH COrH rH rH T* CO *> CM COCOCOCMrH rH rH n 908T— 808T T rH rH CM CO rj< 5M CM (N •* (M CM rH -j- 308I-663T 0 rHCM I-HCMINCMIO rH rH O5 CO rH rH rH rH Q 86ST— 96&T T rH r^ CO CM CM COCMCM rH -,- *63I— I6ST Z rH iHrHrHrH^T}ICO JCM CO (N rH CM rH rH n 06ST— Z8SI 8 CM CM CM CO CM *Q rH CM CO CM CM o 0 9831— 88ST * rH O CO -^ rH Tjl CO rH CO (N CO ^ 38KT— 6Z3I 9 CM rH rH CO CO CO CM CM CO rH g 8Z3T— 9ZZT 9 rH rH CM (M CO *> X X X 05 O5 o 2 o rH Frontal breadth. s3 3U X GO X *• CO * * CO oq 3 0 H * CO ; * CO *- X O5 747—744 12 1 12 743—740 11 1 11 739—736 10 10 735-732 9 1 1 1 1 9 731-728 8 1 1 1 2 1 i 8 727—724 7 1 2 1 4 1 1 2 7 723—720 6 1 3 3 4 3 i 1 6 719—716 5 1 1 2 4 3 4 3 5 715—712 4 1 1 1 3 4 5 5 1 1 1 4 711—708 3 2 3 4 3 2 5 10 5 5 2 1 1 1 3 707—704 2 1 1 3 4 5 10 6 5 9 3 3 5 2 2 703—700 1 3 3 8 4 12 7 8 5 7 3 2 1 699—696 0 2 4 5 7 5 13 3 8 5 2 4 1 0 695—692 1 3 2 1 1 8 6 8 7 2 8 4 2 1 691—688 2 2 3 1 5 1 3 7 8 2 2 2 2 687—684 3 4 1 5 1 2 4 2 3 3 2 3 683—680 4 1 1 2 1 2 2 3 1 1 1 4 679—676 5 5 3 1 1 1 1 5 675—672 6 2 3 3 1 6 671 -668 7 1 1 1 7 667—664 8 8 663-660 9 1 9 659—656 10 10 655—652 11 11 651—648 12 1 12 00 *• CO lO * co * iH 0 iH * CO * » CO fc- X O5 Variation in Portumis depurator. 239 V. Correlation Surface of Frontal Breadth and R. Dentary. 460 Individuals. Measurements in thousandths of carapace length. f 1 152—455 I 160—463 § 1 1— 1 172—475 cc | 1 i M i I 1 I i— i rH 1 10 AH Frontal breadth. rf *- « o 4 cc * i— ( 0 — * CO « o cc t> 00 05 747—744 12 1 12 743—740 11 i 11 739-736 10 10 735—732 9 1 1 1 1 9 731—728 8 3 2 1 1 8 727—724 7 2 1 1 2 1 3 1 1 7 723—720 6 1 1 2 4 1 2 2 i 1 1 6 719-716 5 1 5 4 1 1 2 2 2 5 715—712 4 1 1 3 4 ] 5 4 2 1 ] 4 711—708 3 1 1 1 4 3 4 6 2 5 5 3 6 1 2 3 707—704 2 2 4 4 3 4 6 9 6 4 4 6 2 2 1 2 703—700 1 1 1 6 2 1 6 7 7 3 7 6 7 2 2 1 2 1 1 699—696 0 3 3 3 6 3 6 8 5 8 4 3 4 3 0 695—692 1 1 3 3 4 4 2 - 8 4 7 3 4 1 2 1 1 691—688 2 1 1 1 8 1 2 4 2 4 5 3 4 3 1 1 2 687—684 3 1 1 3 1 1 2 3 4 4 2 2 1 1 1 3 683—680 4 1 1 1 2 1 1 2 3 2 1 4 679-676 5 1 3 2 4 2 5 675—672 6 2 ,1 1 2 2 'l 6 671—668 7 1 1 1 7 667—664 8 8 663—660 9 1 9 659—656 10 10 655—652 11 11 651—648 12 1 12 ** CO 0 * CO CO lO « CO « r- 1 0 * (M CO * « CO *> Variation in Portunus depurator. 241 VII. Correlation Surface of R. Antero- lateral and L. Antero-lateral. 1432 Individuals. Measurements 1 in thousandths •f £ rH CO 10 co CO % § rH O S 01 O CO CD cb ^H £ £ S3 So rH Cl S % 1 1 rH rH 10 01 rH of carapace length. o i I ft el CD O? I i 3 CM 10 1 s § 1 3 J> el 1 ^ 4< 1 oo 00 1 Cl I I I ! rH I 30 B. an tero- lateral. Hi co rH rH rH rH 0 Oi 00 *- 0 0 ; CO 00 01 0 242 Mr. E. Warren. VIII. Correlation Surface of R. Antero -lateral and B. Dentary 1432 Individuals. Measurements 1 in thousandths of carapace length. 1 1 cc T 1 T T 'N 456—459 1 1 r-H ! 5 476—479 1 1 rH 1 1 I 1 1 rH g 1 H s 1 CD rH 520—523 Si \a i B. Antero-lateral. rH rH s 05 00 * CO o ; CO « - O rH « CO « 10 CO 1> CO O5 o rH 3 IN rH 819—816 10 2 10 815—812 9 1 1 1 1 9 811—808 8 2 2 2 8 807—804 7 1 3 1 2 ? 803—800 6 1 3 5 11 7 6 4 2 6 799—796 5 1 2 g 8 11 8 13 10 10 1 l 5 795-792 4 6 11 11 20 16 12 5 3 1 4 791—788 3 3 5 5 16 22 23 21 10 8 3 1 3 787—784 2 1 i 4 16 18 33 41 32 25 10 4 2 783-780 1 1 1 5 12 21 24 42 32 17 13 9 5 2 1 779—776 0 1 5 10 17 32 36 36 16 11 6 1 1 775-772 1 1 .1 3 6 16 22 36 27 32 13 i 2 1 1 771—768 2 3 8 22 24 30 30 20 13 6 2 2 767—764 3 6 17 13 19 16 14 8 5 2 3 763—760 4 5 6 5 14 17 8 • 2 1 4 759—756 5 2 3 9 11 6 c 3 1 5 755-752 6 2 2 2 4 9 5 6 751—748 7 1 1 3 1 1 1 7 747—744 8 ; 2 2 1 1 1 8 743—740 9 9 739—736 10 1 10 rH o rH 05 00 * CO 0 « CO •M rH 0 rH aq CO 4 0 CO «> CO 05 O rH rH rH (M rH Variation in Portunus depurator. 243 IX. Correlation Surface of R. Antero-lateral and L. Dentary. 1432 Individuals. Measurements in thousandths £» X $ £ o 2 oo o in o | oo cq CO 0 rl 00 N CD a •* 50 CO 0 * CO 0, rH o * H CO * 10 CO «> X Oi o rH rH rH 819—816 10 1 ! 10 815—812 9 2 1 1 9 811—808 8 8 1 1 1 £ 807—804 7 1 ] 1 3 1 7 803—800 6 1 2 6 8 11 3 6 2 6 799—796 5 2 1 4 11 12 i 10 i 12 2 1 1 5 795—792 4 1 4 12 11 12 20 16 L 3 2 4 791—788 3 1 ] 2 c 10 12 24 24 23 8 5 1 1 £ 787—784 2 1 '2 8 9 14 25 35 35 19 27 8 2 2 783—780 1 1 1 "L 4 8 18 19 40 28 22 15 14 f 2 1 1 1 779—776 0 a 3 4 17 21 32 32 28 20 12 4 0 775-772 1 1 j 8 9 13 20 36 28 25 12 8 9 1 771—768 2 4 b 9 23 30 31 21 12 9 K 4) 1 c 767—764 3 ] 1 3 0 15 16 19 15 13 * / 0 1 3 763—760 4 ] 2 5 1 6 11 11 14 5 *: ^ 759—756 5 1 1 - 2 2 10 11 t 5 1 ] i K 755—752 6 ] ] 1 2 1 9 J 3 1 6 751-748 7 4^) o '•i l 747—744 8 l 1 - ] O 8 743_740 9 i) 739—736 10 1 10 r-l f some interest, and was readily determined. Archegonia were present close to the sporangia, and at the same level on the process. When the process, after producing sporangia, had con- tinued its growth, archegonia and antheridia were present on the portion beyond the sporangia, as well as on the older part, and, in cases in which more than one group of sporangia had developed, the intervening region bore sexual organs. Rhizoids are also produced abundantly from the shaded side of the process, and, so far as exter- nal appearance is concerned, there is no reason to doubt the pro- thallial nature of the region on which the sporangia are situated. The tissue underlying the .sporangia, however, presents peculiarities in structure which may modify this conclusion to some extent. Beneath the single sporangia developed on the edge of the prothallus a few tracheides, which agree in every respect with those present in apogamous prothalli, were always to be found. Similar elements were always present in the tissue beneath the groups situated on the process. It is possible that here, as in the case of the sporangia upon the prothallus edge, the first tracheides are developed before the young sporangium can be recognised. All that can be stated with certainty is that they are already present beneath very young sporaugia. The tracheides may become connected together into a Development of Sporangia upon Fern Prothalli. 253 band, resembling a rudimentary vascular bundle, and suggesting a comparison with the vascular supply of a sorus. The development of the sporangium could not be followed in detail in the material obtained as yet, but a sufficient number of stages have been found to make it clear that there is no difference of importance from the well known course of development of thd same member on the sporophyte. In the youngest stage seen the apex of the sporangium was occupied by a tetrahedral cell, the cells destined to form the lateral portions of the wall having already been cut off from a large, dome-shaped terminal cell, the limits of which were clearly recognisable. This was borne upon a stalk cell. A tetrahedral archesporium is formed, from which tapetal cells are cut off. The tapetum subsequently becomes two-layered, and the central cell developes into a group of sporogenous cells. From these, in the most mature sporangia found, a number of dark brown spores had developed, while the tapetum was represented by numerous granules between the spores. The number of spores appeared to be the same as was contained in a sporangium developed on the sporophyte. The sporangium wall was perfectly developed ; the cells of the annulus showed the characteristic thickening of their walls, which were of a dark brown colour, and a well formed stomium was present. When tested with dehydrating agents, the mechanism of the annulus was found to be perfect. The stalk consisted of four rows of cells. JSTo sporangia have been found in which the spores were ripe, but in .view of the advanced stage of development in those observed, there is every probability that some may be obtained. It will be interesting to ascertain if the spores are capable of germination, and if the prothalli produced show any peculiarities. The spores seen already possessed a thick wall on which indications of sculpturing were appa- rent, and a single nucleus was present in each. When the unnatural conditions under which they developed are borne in mind, it is not surprising that many imperfect sporangia were found. Such sporangia were in fact the more numerous. Some- times the arrest of development had taken place before the tapetum had originated from the archesporium, but more commonly the double layer of tapetal cells was present surrounding a sporogenous cell which had become highly refractive, the nucleus being indistinguish- able. The annulus could be made out, but its cells were thin walled and colourless, and the whole sporangium was pale and more flattened than one of the same age in which sporogenous tissue had formed. No evidence has yet been obtained of the production of sporo- phytes, showing vegetative organs upon the cylindrical process, but one example was seen in which a group of sporangia, situated on the apex of the process, was surrounded by ramenta. VOL. LX. u 254 Mr. W. H. Lang. Preliminary Statement on the Scolopendrium vulgare, I/., var. ramulosissimum, Woll. — The cultures of this fern were made in the manner already described for Lastrcea dilatata. The spores were obtained from a plant grown in the open air in the Royal Gardens, Kew. The prothalli were at first heart-shaped, and on many of them normally produced embryos developed. No further changes ensued in those on which young plants were present, and they soon became colourless and died. In those which had remained unfertilised, how- ever, the apex continued directly into a cylindrical process,* which was of considerable thickness, and in some cases attained a length of 5 mm. The lateral portions of the prothallus showed no further growth, and became in time brown or colourless appendages to the base of the cylindrical process. On the process were numerous archegonia, and its prothallial nature was still further shown by the presence, in. some instances, of thin lobes of tissue, which generally bore antheridia. Sections through the process in this stage show that the archegonia are normally formed, and reach almost to the apex, and that tracheides are absent from the tissue. The archegonia are capable of fertilisation, for in some instances normally produced embryos were found. After the process has in this manner attained a greater or less length, its tip becomes yellowish, contrasting with the deep green colour of the region behind. Near the apex ramenta develope, which soon completely clothe the tip of the process and render it white and conspicuous. Archegonia are present to just below the ramenta. Longitudinal sections at this stage show that one or two small eleva- tions corresponding to the rudiments of the apex of the stem, and the first leaf of the sporophyte have been formed. Beneath the broad tip a flat mass of small meristematic cells extends ; the meristematic tissue is continuous with that of the stem and leaf apices, but, on passing away from these, is separated by several layers of large, non- meristematic cells from the surface. In a slightly older stage the stem apex has become conical, and a number of leaves have formed which are circinately curved, and form a bud clothed with ramenta. In the meristematic mass numerous tracheides have been developed. One large group is central in position, and extends to the limit between prothallial and sporophytic tissue, while others are found beneath the bases of the leaves, and are in continuity with their pro- cambial strands. The apex of the stem is occupied by an initial cell, the relation of which to the initial cell or cells of the apex of the process has not yet been traced. The young sporophyte appears to be a direct continuation of the process. It is possible that some of * Prothalli of Scolopendrium, which from the brief description given of them appear to have borne similar processes, are mentioned by E. J. Lowe, in the ' Gard. Chron.,' November 10, 1895. They were not investigated further. Development of Sporangia upon Fern ProthallL 255 the cases of apogamy recorded by Stange* were of this nature, but in Doodia caudata, R. Br., which is the only one of his species yet investigated in detail ,f the elevations, from which sporophytes de- veloped, were situated on the under surface of the prothallus. This case appears to be intermediate in character between Scolopendrium and the species investigated by De Bary.J Several prothalli were found bearing sporangia ; these were grouped together in large numbers, usually upon the upper surface of the cylindrical process, but sometimes both above and below. Archegonia were situated close to the groups of sporangia. In the region of the prothallus, underlying the group, a strand of tracheides was found ; in one instance this was connected with a spherical mass of tracheides developed to all appearance within the venter of an archegODium whose neck had not opened. The tissue upon which the sporangia are inserted is thin walled, and its cells have granular contents ; it contrasts sharply with the cells of the prothallus which have a large vacuole and walls which stain much more deeply with hsematoxylin. As in the case of Lastrcea dilatafa, the stages seen render it prob- able that the sporangia follow the usual course of development. Two layers of tapetal cells are formed which surround a considerable mass of sporogenous tissue. Many of the sporangia fail to attain full development; they remain colourless, and in time wither. A few have been found, however, with a well developed annulus of a dark colour; these contained spores which have not, however, been examined in detail. In one case two ramenta overarching a group of sporangia were seen. At first sight it seemed possible that they might correspond to an indusium, but, when taken in connexion with another example in which a cylindrical process, which bore sporangia laterally, termi- nated in an apogamously produced biid, another explanation appears more probable ; this will be referred to again below. It is worthy of note that another variety of this species has been found to produce young plants, the first fronds of which bore numerous prothalli while still in connexion with the stem.§ The prothalli on which these plants appeared had been subjected to repeated subdivision, a process which in other species || has been found to induce apoganious development of the sporophyte. Unfor- tunately nothing is known of the manner in which these peculiar plants of Scolopendrium were produced, but it is possible that they arose apogamously. The case of Scolopendrium would then be com- * ( Ber. der Gesellsch. f. Bot.,' Hamburg, 1880, p. 43. f Heim, ' Flora,' 1896, p. 329. J Loc. cit. § In a paper by Mr. E. J. Lowe, read at the Linnean Society, February 20, 1806. || Stange, loc. cit. 256 Mr. W. H. Lang. Preliminary Statement on the parable to that of Trichomanes alatum* in which apogamy and apospory co-exist. Prothalli have been found to arise directly from the older fronds of another variety of Scolopendrium.'f An attempt will now be made to bring the peculiar modification of the life-history cycle of these ferns into relation with previously recorded cases of apogamy, and to estimate its theoretical bearing. A full consideration of these points must be deferred until more extended observations have been made. There seems no reason to doubt the prothallial nature of the cylin- drical process : its origin, the character of its cells, the presence of functional sexual organs, the development of rhizoids, and the direct transition to an ordinary flat prothallns apex sometimes met with, are sufficient grounds for this conclusion. The distinction between its origin as a direct continuation of the prothallus, and the cases in which it arises behind the apex which has lost its meristematic cha- racter, is not an essential one. Both forms occur in Lastrcea dilatata ; in the latter case the process may be compared with the numerous elevations which appear on the under side of old prothalli of Doodia caudala,$ and are capable of apogamous development. The forma- tion of such processes by prothalli which have attained a considerable size without having been fertilised, appears to be of not infrequent occurrence, and is usually associated with apogamy. It is recorded in Todea pellucida, Carm., T. rivularis, Sieb.^ and Athyrium filix- fcemina, B&rnh.,\\ and the writer has found in Aspidium frondosum, Lowe, as many as six apogamous buds, formed from the tips of cylindrical processes, which arose from the anterior margin of a prothallus. The term cylindrical process^" has been used to avoid confusion with the middle lobe developed in aborting prothalli of Pteris cretica and Aspidium filix-m as. This, as De Bary has shown, may be regarded as corresponding to some extent with the first leaf of an apogamous sporophyte.** A structure comparable with this middle lobe has been found in prothalli of Lastrcea dilatata, which had also produced a cylindrical process ; usually one or more sporangia were borne upon it. Tracheides were always present in the tissue beneath sporangia, * Bower, 'Annals of Botany,' rol. 1, p. 300. f Druery, 'Linn. Soc. Jouru.,' vol. 30, p 281. J Heim, loc. cit., p. 340, fig. 12. § Stange, loc. cit. || Druery, ' Gard. Chron.,' November 10, 1895. ^[ It is impossible to determine whether the structure to which Wigand (' Bot. Zeit.,' 1849, p. 106) applied this name, and which he inclined to consider as a rudimentary axis, was of tho same nature or was a true middle lobe, but the latter appears the more probable conclusion. ** Loc. cit., p. 464. Development of Sporangia upon Fern Prothalli. 257 and the question arises whether their occurrence is to be regarded as of morphological significance. They have been found in the pro- thalli of a number of species of ferns, and, in every case investi- gated, were associated with apogamy. In the case of Pteris cretica the differentiation of the tracheides in the prothallus precedes the origin of the bud.* This is the case also with the single sporangia formed on the edge of the prothallus, and probably holds good for the groups of sporangia borne on the process. But tracheides may occur in the prothallus at a distance from the place of origin of buds or sporangia. Putting aside the case of the middle lobe, the prothallial nature of which is open to doubt, a large bundle of tra- cheides was found in the substance of a fleshy prothallus of a variety of Scolopendrium vulgare, which bore numerous archegonia on the sarfaces immediately above and below the traeheides. Elongated cells, which resemble sclerenchyma fibres, occur in the midrib of cer- tain frondose liverworts. f A still more instructive example is afforded by the presence of tracheides in the massive endosperm of certain cycads. J This latter case shows clearly that such elements may be formed in the gametophyte to meet a physiological need. Ib seems inadvisable, therefore, to lay stress on the presence of tra- cheides as a means of distinguishing between the two generations, and the more so since their occurrence in a portion of the prothallus which is about to bear a bud or sporangia can be recognised as a physiological advantage. Such means of procuring a sufficient water supply maybe a necessary preliminary to the development of a young sporophyte or a group of sporangia. Lastly, it remains to consider the view to be taken of the presence of the characteristic reproductive organs of the asexual generation upon the gametophyte, and to consider its bearing upon the nature of alternation of generations in the archegoniatse. Since the dis- covery that in certain cases the one generation could arise directly from the other without the intervention of the proper reproductive organs, such cases have been used in support of the view that the alternation in the Archegoniatse was homologous. § On the other hand, it has been maintained, both on grounds of the exceptional nature of these cases of apospory and apogamy, and of comparative phylogeny, that the distinction between the two generations was a much deeper one ; that the alternation was not homologous, but anti- thetic. || So far no case has been recorded in which the proper reproductive organs of the one generation were situated upon the * Farlow, loc. cit., p. 269. f Goebel, « Outlines,' p. 145. J I am indebted to Professor Bower for this unpublished fact. § Pringsheim, ' Jahrb. f . Bot.,' bd. 9, p. 43. || Bower, ' Annals of Botany,' vol. 4, p. 347. 258 Mr. W. H. Lang. Preliminary Statement on the other without the intervention of the vegetative organs. At first sight such appears to be the case in the prothalli of the two species described ; sporangia were present in close proximity to the sexual organs, the vegetative organs of the sporophyte being, at most, repre- sented by a mass of cells underlying the group of sporangia, and even this distinction may not be recognisable beneath the single sporangia on the edge of the prothallus. Several reasons may be adduced, however, against regarding these phenomena as evidence that the alternation of generations found in the ferns is not antithetic. In the first place, it is to be noted that the two forms in which sporangia have been observed upon the gameto- phyte are highly variable species, and that the varieties studied were well-marked crested forms. Further, the conditions under which the prothalli existed were in several respects unnatural. Among them the fact that fertilisation was prevented by not watering the cultures from above, and that a prolonged growth of the unfertilised prothalli was thereby induced, is of special interest, for it appears that apogamy is liable to occur under such conditions in ferns which, as a rule, reproduce sexually. While these considerations do not of themselves preclude deductions being made from these peculiar forms of repro- duction, they necessitate especial caution in their use in the discusoion of broad morphological questions. Further, a number of reasons exist for considering the production of sporangia on the prothallus as a special case of apogamy. In Scolopendrium vulgare a sporophyte may develope from the tip of the cylindrical process. This may happen after a group of sporangia has been developed. In one case two ramenta were present, one on .each side of a group of sporangia; they were in every respect similar to the ramenta which develope on the tip of the process when it is being transfprmed into the apex of a bud. Whenever a group of very young sporangia was seen it was situated upon the apex of the lobe, and the sporangia were in a more advanced stage of develop- ment the farther the group to which they belonged was removed from the apex. This has been most clearly seen in the case of Lastrcea dilatata in which no buds with vegetative organs have as yet been seen, although in one case ramenta were associated with the sporangia, but it also holds for Scolopendrium. The explanation of these facts, which appears most probable, is that each group of sporangia had occupied the apex of the process when very young, and had become farther removed from this position as the process continued to increase in length. It is uncertain whether this growth is by direct continuation of the original growing point of the process, or whether the development of a group of sporangia at the apex necessitates the formation of a new growing point ; possibly both forms occur. If the latter be the case a process on which several Development of Sporangia upon Fern Prothalli. 259 -groups of sporangia are present must be looked upon as a sympodium. Some probability is lent to this view by the fact that the first- appearance of the process in Lastrcea is usually as a sympodial con- tinuation of the axis of a prothallus whose true apex has developed one or more sporangia. Since the group of sporangia and the tissue of peculiar character on which they are seated are developed in the place of an apoga- mously produced vegetative bud, they may be looked -upon as con- stituting a very reduced sporophyte. The drain upon the resources of the prothallus entailed by the production of this reduced bud, which is incapable of further growth, is much less than when a vegetative bud is formed. This explains why a number of such sporangial groups can be produced and supported by a single pro- thallus. The occurrence of a number of vegetative buds on a single prothallus is the exception, but may happen, as the case of Aspidium frondosum, before mentioned, shows. It is probable that it is in the constitution of the nuclei that a means of distinction between cells of the oophyte and the sporophyte must be looked for in these cases in which the two generations are in intimate connection with each other.* The complete life history of the fern is in these cases still further shortened than in the ordinary cases of apogamy ; not merely the formation of a zygote by the fusion of antherozoid and ovum, but the formation of an embryo, in which any differentiation of the vegeta- tive organs can be detected, is omitted, and the sporophyte is reduced to a mass of tissue which may be compared to a placenta bearing sporangia. The occurrence of single sporangia upon the edge of the prothallus may, in the light of the series of stages described, be con- sidered as a still further case of reduction of an apogamous sporo- phyte. While this does not altogether prevent the explanation of the presence of sporangia upon the prothallus from the point of view of the supporters of the homologous nature of the two generations, it brings the present case into line with other exceptions to the normal life-history cycle, whose bearing on the nature of alternation has been discussed by Bower, f The present case, although more striking in its appearance, seems, so far as it has been investigated, to afford no sufficient reason for dissenting from the conclusion at which he arrived. It is of interest to note the additional evidence, were such needed, which these observations afford of the generalization made by Goebel,J that the sporangium is to be regarded as an organ sui generis. * Bower, ' Trans. Bot. Soc. Edinb.,' vol. 20. t ' Annals of Botany,' vol. 4, 1890, p. 347. t ' Bot. Zeit.,' 1881, p. 707. VOL. LX. X 260 Prof. G. B. Grassi. The Reproduction and From the staff of the Royal Gardens, Kew, I received ready assistance in many practical matters in the conduct of the cultures ; my thanks are especially due to the curators, Mr. Watson and Mr. Nicholson. November 19, 1896. Sir JOSEPH LISTER, Bart., President, in the Chair. Dr. Francis Elgar was admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. In pursuance of the Statutes, notice of the ensuing Anniversary Meeting was given from the Chair. Mr. Shelford Bidwell, Professor Bonney, and Mr. Horace Brown were by ballot elected Auditors of the Treasurer's accounts on the part of the Society. The Secretary read the Titles of the Papers received since the last meeting, which, under the new Standing Orders, had been published (see ' Proceedings,' No. 362). The following Papers were read : — I. " The Reproduction and Metamorphosis of the Common Eel (Anguilla vulgaris)." By G. B. GRASSI, Professor in Rome. Communicated by Professor E. RAY LANKESTER, F.R.S. II. " Total Eclipse of the Sun, 1896. — The Novaya Zemlya Observa- tions." By Sir GEORGE BADEN-POWELL, K.C.M.G., M.P. Communicated by J. NORMAN LOCKYER, C.B., F.R.S. III. " Preliminary Report on the Results obtained with the Prismatic Camera during the Eclipse of 1896." By J. NORMAN LOCKTER, C.B., F.R.S. " The Reproduction and Metamorphosis of the Common Eel (Anguilla vulgaris)" By G. B. GRASSI, Professor in Rome. Communicated by Professor E. RAY LANKESTER, F.R.S. Received October 19, 1896. Read November 19, 1896. Four years of continual researches made by me in collaboration with my pupil, Dr. Calandruccio, have been crowned at last by a success beyond my expectations, that is to say, have enabled me to Metamorphosis of the Common Eel. 261 dispel in the most important points the great mystery which has hitherto surrounded the reproduction and the development of the Com- mon Eel (Anguilla vulgaris). When I reflect that this mystery has occupied the attention of naturalists since the days of Aristotle, it seems to me that a short extract of my work is perhaps not unworthy to be presented to the Royal Society of London, leaving aside, how- ever, for the present, the morphological part of my results. The most salient fact discovered by me is that a fish, which hitherto was known as Leptocephalus brevirostris, is the larva of the Anguilla vulgaris. Before giving the proofs of this conclusion I must premise that the other Muraenoids undergo a similar metamorphosis. Thus, I have been able to prove that the Leptocephalus stenops (Bellotti), for the greatest part, and also the Leptocephalus morrisii and punctatus belong to the cycle of evolution of Conger vulgaris ; that the Lepto- cephalus hceclteli, yarrelli, bibroni, gegenbauri, kollikeri, and many others imperfectly described by Facciola, and a part of the above- named Leptocephalus stenops of Bellotti, belong to the cycle of evolu- tion of Congromurcena mystax ; that the Leptocephalus tcenia, in- ornatus, and diaphanus belong to that of Congromurcena balearica • that under the name of Leptocephalus kefersteini are confounded the larvae of various species of the genus Ophichthys ; that the Lepto- cephalus longirostris and the Hyoprorus messanensis are the larvae of Nettastoma melanurum, and that the Leptocephalus oxyrhynchus and other new forms are larvae of Saurenchelys cancrivora, and that finally a new little Leptocephalus is the larva of Muroena helena. The form known as Tylurus belongs to Oxystoma, of which we unfortunately know nothing more than a figure by Raffinesque. I have not been able to find the Leptocephalus of Myrus vulgaris, of which I have had only a single young individual, in which the trans- formation was already far advanced. Neither have I found the Lepto- cephalus of Chlopsis bicolorj a very rare form, which is related to Murcena and to Murcenichthys. As the result of these observations, the family of the Leptocephalidae has been definitely suppressed by me ; the various forms of that family are, in fact, the normal larvae of the various Mura3noids. In regard to the greater part of the above-named species, the con- trol has been threefold, namely : — Firstly, anatomical. I have compared the various stages in all their structures, and have made the due allowance for the changes brought about by the metamorphosis at the close of larval life. Secondly, natural. I have found in nature all the required transi- tional stages. Thirdly, experimental. I have followed, step by step, the meta- morphosis in aquariums. x 2 262 Prof. G. B. Grassi. The Reproduction and Therefore, the hypothesis of Giinther that the Leptocephali are abnormal larvas, incapable of further development, must be rejected. All this is related by myself at length, with all historical details which concern the question, in a large memoir which is about to appear in the Journal edited by Professor Todaro. Until now all these facts have been unknown because normally they can only be observed in the abysses of the sea at a depth of at least 500 metres. Fortunately, along a part of the coast of Sicily strong currents occur, which must be ascribed to the tide, producing very large displacements of the water in the narrow Strait of Mes- sina. I shall give further details concerning these currents in my large memoir. In consequence of the strong currents, sometimes — I say sometimes, because there is no regularity, and one may have to wait for a year without obtaining any material — not only many deep-sea fishes, but also all stages of the development of the Murae- noids are met with in the surface-water. To these currents we owe all the captures of Murcena Jielena with ripe eggs, which is in accord- ance with what I had already argued from other facts, namely, that the reproduction of the Muraenoids takes place at great depths of the sea. Before I proceed to speak of the Common Eel, I must premise that Dr. Kaffaele has described certain pelagic eggs as belonging to an undetermined species, putting forward the suggestion that these eggs belong to some Muraenoid. This matter has been investigated by myself, and I have shown that the newly hatched larvae (called " praa-larvae " by me) derived from these eggs have essentially the character of Leptocephali. The life history of the Muraonoids, leaving aside for the present the Common Eel, is as follows : — Females can only mature in very profound depths of the sea, that is to say, at least a depth of 500 metres. This fact I established by finding well-known deep-sea fishes together with Leptocephali, ripe Muraenae, and quite ripe eels {see below). The females of those species which do not live at this depth must therefore migrate to it. The male, however, can mature at a smaller depth, and therefore they migrate into the greater depth when they are already mature. Fertilisation takes place at great depths ; the eggs float in the water ; nevertheless they remain at a great depth in the sea, and only exceptionally, for unknown reasons, some of them mount to the surface. From the egg issues rapidly a pras-larva, which becomes a larva (Leptocephalus) with the anus and the urinary opening near the tip of the tail. The larva then becomes a hemi-larva, the two aper- tures just named moving their position towards the anterior part of the body, which becomes thickened and nearly round. By further -change the hemi-larva assumes the definitive or adult form. The Metamorphosis of the Common Eel. 263 larva, as well as the hemi-larva, shows a length of body much greater than that exhibited by the young Muraenoid of adult form into which they are transformed. By keeping specimens in an aqua- rium, I was able to establish a diminution of more than 4 cm. during the metamorphosis. With regard to the greatest length which the larva can attain in a given species, -and the amount of diminution which accompanies metamorphosis, there are great individual varia- tions. The history of the Common Eel, to which I am now about to refer, is very similar to that given above for the other Mursenoids. The Common Eel (Anguilla vulgaris) undergoes a metamorphosis, and before it assumes the definitive adult form it presents itself as a Leptocephalus, which is known as Leptocephalus Irevirostris. This Leptocephalus was discovered in the Strait of Messina many years ago. A specimen was also captured by the " Challenger," and another specimen was taken by the Zoological Station of Naples in the Strait of Messina. This form is occasionally carried to the surface by currents. By exception, in the month of March, in the year 1895, we captured several thousands of them in one day, but the best way to secure this Leptocephalus (and a very easy one) is to open the intestine of the Orthagoriscus mola, a fish which is common in the Strait of Messina, and in it one is certain to find a very large number of specimens. It must be observed that Orthagoriscus mola is a deep-sea fish. The specimens of Leptocephalus brevirostris found in the intestine of Orthagoriscus are more or less altered by digestion. Those specimens of Leptocephalus brevirostris which are taken near the surface in the open sea are in a better state of preser- vation, but, unfortunately, these also frequently have the epidermis injured so that they cannot maintain their life in an aquarium for more than a few days; they live long enough, however, to allow us to observe that it is their habit to conceal themselves in the sand or in the mud as the Common Eel (Anguilla) does. Here it is to be noted that the various forms of Leptocephali have habits resembling those of the Mursenoids to which they belong, i.e., they dig into the sand or abstain from doing so according as the adult form has or has not this habit. I now pass on to the characters of Leptocephalus brevirostris. I give them here in the same order as I shall use in my larger memoir. The length varies from 77 — 60 mm., the same extent of variation as observed in other Muraenotds. The caudal fin tends to assume the form which it has in the Elver* or young Anguilla. It is to be noted that in other Leptocephali the caudal fin also tends always to exhibit the adult form. The lower jaw projects sometimes more than the * The word " Elver " is used in this paper in its strict sense, viz., for the young form of Anguilla vulgaris as taken when ascending rivers in vast numbers. 264 Prof. G. B. Grassi. The Reproduction and upper jaw, as in Anguilla. The margin of the month is wide, as in Angnilla. The tongue is free, as in Anguilla. On the other hand, the youngest elvers which I have observed, have smaller eyes than Leptocephalus brevirostris, and this need not surprise us since we know that in other species of Mura3noids the diminution of the eyes occurs during the metamorphosis. T,he nostrils are separated from one another, the anterior tubes are relatively at a considerable distance from the tip of the snout and from the rim of the mouth. They are in a position in which they are observed in many other Leptocephali, which are destined to transform themselves into adult forms having the anterior nostrils in nearly the same position as in the Common Eel. The posterior nostrils, on the contrary, are not tube-like, and are in the same position as those occupied in the adult Anguilla. It is worth remarking that in other Leptocephali also the posterior nostrils have already assumed the adult position when the anterior ones are still far removed from it. In L. brevirostris I find a larval dentition, which resembles that of the other Leptocephali. In correspondence with the small size of Leptocephalus brevirostris the number of larval teeth is small. Researches fonnded, firstly, on the enumeration of the myomeres ; secondly, upon the enumeration of the dorsal and ventral arches of the vertebrae of the caudal extremity (hypnrals) ; and, thirdly, upon the enumeration of the posterior spinal ganglia, lead with great certainty to the conclusion that the Leptocephalus brevirostris is the larva of a Muraenoid, the number of whose vertebrae must lie between 112 and 117, most probably 114 or 115. Such a Mursenoid is the Anguilla vulgaris. The Mura3noid indicated cannot be any other of those occurring in the Mediterranean, because they all have a number of vertebras higher than 124.* Counting the myomeres in Leptocephalus brevirostris one finds generally only 105 complete, five others incomplete, and all the others in a state of transparency and incomplete formation. These latter are fortunately a,t the posterior extremity, where other criteria come to our assistance, namely, the spinal ganglia and the vertebral arches. To show how I arrive at the number of vertebrae which must be possessed by the adult individual, corresponding to a given Leptocephalus brevirostris, I quote the following example : — I assume that three vertebras develop themselves in correspondence to the first four incomplete myomeres, and that 105 must develop themselves in relation to the 105 complete myomeres, that is to say, between the fourth and fifth myomeres, between the fifth and sixth, and so on, until we reach the 105th vertebra, lying between the 104th and 105th myomeres. I * Muroenesox savanna is said to have 109 vertebrae, but it is doubtful whether it really occurs in the Mediterranean. The position of its nostrils and the number of its branchiostegal rays render its association with Leptocephalns brevirostris impossible. Y Metamorphosis of the Common Eel. 265 further conclude that seven other vertebras are developed at the caudal extremity, as indicated by the number of vertebral arches and the spinal ganglia in that region. We count, therefore, in all 115 vertebrae, and this is the number which can be easily seen in many specimens of Anguilla vulgaris. Here I must particularly insist that I have .ascertained in an absolute manner that during the metamorphosis of the Mureenoids, the number neither of the myomeres nor of the vertebral arches, nor of the spinal ganglia is subjected to any change. The hypurals of Leptocephalus brevirostris are precisely 'the same as in the elver of Anguilla vulgaris. The last hypural which is fused with the urostyle may present itself as a single piece, or may be more or less cleft. These are variations which are met with also in the elver. Just as in the elver, the last hypural but one is always extensively cleft, or, if the expression is preferred, doubled. To the last hypural corre- spond five rays, whilst four correspond to the last but one, and one to the last but two, the whole structure being identical with that found in the elvers of Anguilla vulgaris. Of these ten rays, the eighth, seventh, and sixth are bifid, both in Leptocephalus brevirostris and in the elvers of Anguilla vulgaris. In the pectoral fin of Leptocephalus brevirostris the definitive rays can be observed, and these are of the same number as in the elvers of Anguilla vulgaris. Leptocephalus brevirostris is transparent, and has colourless blood. The red cor- puscles are wanting, but there are present so-called " blood-plates " (" Blutplattchen " in German) similar to those of the inferior vertebrates. The bile is also colourless. This fact is observed in all the other Leptocephali. Leptocephalus brevirostris is, however, the only one which is free from all pigmentation. Correspondingly, the Common Eel is the only species of Muraenoid which at the close of metamorphosis is devoid of all trace of larval pigmentation. It was this observation which first led us to the discovery of the relations between Leptocephalus brevirostris and Anguilla vulgaris. In making transverse sections of Leptocephalus brevirostris, I found other characters which confirm the relation between it and the Com- mon Eel ; for instance, the branchiostegal rays are ten to eleven in number, as is also observed in the elvers of Anguilla vulgaris. In the Common Eel the well-known lateral branch of the fifth pair of the cranial nerves exists. It is also found in Leptocephalus brevirostris. This lateral branch could not be found by Dr. Calandruccio in the other common Murasnoids of Sicily, and is wanting also in the other Leptocephali. The mucous-canal-system (sensory canals) in the head are already developed, partially, in Leptocephalus brevirostris, and are incom- pletely developed in the elver. As in the elver, so in Leptocephalus brevirostris, the pyloric cceca are wanting. The blind extremity of 2(i6 Prof. G. B. Grassi. The Reproduction and the stomach and the incompletely developed swim-bladder, which is as yet free from contained gas, are present both in Leptocephalus brevirostris and in the elver of Anguilla vulgaris. The pronephros is in active function as in the other Leptocephali. The Malpighian glomerules of the kidney (mesonephros) are lobed as in the eel, and their number corresponds with that observed in the Helmichthys stage, of which I will speak further on. The genital gland, not yet sexually differentiated, is almost identical with that of the same stage. In short, it may be said that the whole organisation of Leptocephalus brevirostris corresponds with the organisation of the Common Eel, if we make allowance for those changes, which are observed in the matamorphosis of the other species of Mureenoids, such as reduction of the pancreas and of the liver, disappearance of the proto-skeleton, complication of the musculature, increase in size of the cerebellum, loss of the larval teeth, development of the defini- tive teeth, &c. From the description of these Leptocephali I must pass on, briefly, to speak of the stages nearer to the condition ol the elver. I am, however, obliged to leave a break in the series, which, however little its significance, yet certainly will make some impression on the minds of those who do not realise with what caution I have formed my con- clusions. I must confess that since I have learnt how difficult it is to procure an entire series of the development of a Mura3noid, I am more astonished at being able to recognise a single stage in the- development of a given species than at not finding the whole series. I in nst point out that the break in my series of the development of Anguilla vulgaris would have been much smaller if I could have persuaded myself to kill and preserve one of the hemi-larvae which I happened to meet with at the end of the year 1892. They were really transitional stages between Leptocephalus brevirostris and that stage which I shall describe further on. I published this fact in a preliminary note in the month of May, 1893. They were transparent with almost colourless blood, without any trace of pigmentation, except at the eyes, and had lost all the larval teeth, whilst they possessed already very few and very minute teeth of the definitive series. The body was thickened, and already showed the cylindrical form. They measured little less than 8 cm. In short, they were Leptocephalus brevirostris on the way to transformation into Anguilla vulgaris. As a matter of history they actually did transform them- selves in my aquarium with the usual diminution in their dimen- sions, and subsequently proceeded to increase in bulk.* The meta- morphosis took place as usual without the animal taking in any * The fact that I actually hare obtained in an aquarium the transformation of L. brevirostris into Anguilla vulgaris is of prime importance. The time occupied was one month. Metamorphosis of the Common Eel. 267 nourishment whatever. The resumption of growth was accompanied by a resumption of feeding. Unfortunately, I had no other indi- viduals of this stage. The stage which I now pass on to describe can be obtained during the winter in the sea. I have never found them at the mouths of rivers. The length varies from 54 to 73 mm. Most individuals measured about 65 mm. The body is relatively longer than in the elver. It is also relatively deeper, as in Leptocephalus. We are reminded of Leptocephalus also by the pigment of the eye, the vitreous transparency of the body, the swim-bladder being indis- tinguishable in the living animal, and the absence of all larval pig- mentation. The blood is slightly coloured, and the bile is already green. Slight pigmentation can be seen along the central nervous system, and at the middle part of the caudal fin. This commence- ment of the definitive or adult pigmentation in the regions named before it occurs in any other part is also seen in other Muraenoids. The definitive teeth are very minute, and few in number. The intestine contains no food. After what I had observed in the other Mursenoids, the simple observation of the barely indicated teeth, and of the absence of aliment in the gut, would have been sufficient to convince me that the stage now under notice must be preceded by a Leptocephalus phase. Indeed, if we did not admit such a preceding history, we could not understand how this little fish could have attained such a size without acquiring well developed teeth, and with- out nourishing itself. In conclusion, no one would hesitate, even not knowing Lepto- cephalus brevirostris, to refer the stage now under discussion to a Mureeiioid about to complete its Leptocephalus metamorphosis, were it not for the fact that there has been so much question concerning the reproduction of the Common Eel, and that so many capable observers have failed in dealing with it, that every new observation is received with scepticism. The stage of which I am now speaking, in the hands of a pure systematist, would probably be described as a Helmicthys, a genus established for certain forms of Leptocephali far advanced in transformation. The next forms to which I have to refer are captured in the course of migration from the sea into fresh water. When kept in an aquarium they assume the characters of the elver, diminishing more or less in volume, and without nourishing themselves. The elvers of the Common Eel can present themselves in stages differing little from that last described, as well as in a form which has already developed the full pigmentation of the adult. Even those which most resemble the preceding stage always have a character which distinguishes them easily, namely, the presence of definitive pigment, more or less superficially placed on the head, and not to be 268 Prof. G. B. Grassi. The Reproduction and confounded with the pigment round the posterior extremity of the brain, which latter is already present in the preceding stage. In specimens taken at the mouths of rivers this more or less superficial pigment was, so far as I could ascertain, always present. As the pigmentation develops itself, the little eel gradually under- goes a diminution in all its dimensions. It results from my measure- ments, that the fully pigmented elver has an average length of 61 mm., while for the more or less colourless elver the average length is 67 mm. I found pigmented elvers which were reduced in length to 51 mm., a size which I never observed in those elvers in which the development of pigment had not taken place. The facts which I have stated demonstrate that the eel goes through a metamorphosis, and that Leptocephalus brevirostris is its larva. Some further considerations remain to be given, although I believe that zoologists will not consider the question still an open one after the record of facts given above — facts, which anyone may verify by examining the material which is preserved in my hands. Many to whom I have related my discovery of the history of the Common Eel have objected that eels are found almost everywhere, whilst Leptocephalus brevirostris is limited to Messina. In reply, I must say that, first of all, it is not true that Leptocephalus brevirostris is limited to Messina ; secondly, that at Messina there are special currents, which tear up the deep-sea bottom which everywhere else is inaccessible ; thirdly, although it is true that on the coasts of many countries where Anguilla vulgaris is found, no one has ever seen a Leptocephalus brevirostris ; it is also true that in no country, not even in those where eels are abundant, has anyone ever seen an eel of less than 5 cm. in length. Since it has to be admitted that no one knows the eel before it arrives at the length of 5 cm., there is no greater difficulty in supposing that during this unknown period the eel passes through a Leptocephalus stage than in supposing that it does not do so. The critical study of the literature of this subject, and a great many continued observations, have occupied me for many years, and have been undertaken just in those places where young eels are to be found. They enable me, from my own studies, to affirm with assurance that young eels with the definitive adult form do not exist of less than 5 cm. in length. From the study of the memoir of Raffaele on pelagic eggs, I have come to the conclusion that the eggs of his undetermined species No. 10, having a diameter of 2'7 mm. and differing from all the others in the absence of oil globules,* must belong to the Anguilla * Kenewed researches have convinced me that this egg is that of Anguilla vulgaris. There is, however, another egg belonging to an undetermined Muraenoid which is devoid of oil-drops, and can easily be confused with the true eggs of Anguilla. Metamorphosis of the Common Eel. 269 vulyaris, because from them Dr. Raffaele obtained prse-larvse which had only forty-four abdominal myomeres. T endeavoured for two years in vain to study these eggs at the Zoological Station of Naples. I found only a few of them, and these died prematurely. In another point my researches have yielded a very interesting result. As a result of the observations of Petersen, we know now that the Common Eel develops a bridal coloration or " mating habit," which is chiefly characterised by the silver pigment without trace of yellow, and by the more or less black colour of the pectoral tin, and finally by the large eyes. Petersen inferred that this was the bridal coloration from the circumstance that the individuals exhibiting it had the genital organs largely developed, had ceased to take nourishment, and were migrating to the sea. Here Petersen's observations cease and mine begin. The same currents at Messina which bring us the Leptocephali bring us also many specimens of the Common Eel, all of which exhibit the silver coloration. Not a tew of them present the characters described by Petersen in an exaggerated condition, that is to say, the eyes are larger and nearly round instead of elliptical, whilst the pectoral fins are of an intense black. It is worth noting that in a certain number of them the anterior margin of the gill slit is intensely black, a character which I have never observed in eels which had not yet migrated to the sea, and which is wanting in the figures and in the originals sent to me by Petersen himself. Undoubtedly the most important of these changes is that of the increase of the diameter of the eye, because it finds its physiological explanation in the circumstance that the eel matures in the depths of the sea. That, as a matter of fact, eels dredged from the bottom of the sea have larger eyes than one ever finds in fresh- water eels, I have proved by many comparative measurements, made between eels dredged from the sea bottom and others which had not yet passed into the deep waters of the sea. Thus, for instance, in a male eel taken from the Messina currents and having a total length of 34 J cm., the eye had a diameter, both vertical and transversal, of 9 mm., and in another eel of 33 J cm., the same measurement was recorded. In a female eel, derived from the same source and purchased in the market, whose length was 48 \ cm., the vertical diameter of the eye was 10 mm., and the transversal diameter rather more than 10 mm. These are not the greatest dimensions which I observed, and I conclude from these facts that the bridal habit described by Petersen was not quite completed in his specimens, and that it becomes so only in the sea and at a great depth. In relation 'to these observations of mine stands the fact that the genital organs in the eel taken in the Messina currents are sometimes more developed than in eels which have not yet entered the deep water. Thus it has happened that male individuals have occurred showing 270 The Reproduction and Metamorphosis of the Common Eel. in the testes here and there knots of spermatozoa. These spermato- zoa are similar to those of the Conger vulgaris, and must be con- sidered as ripe. As is well known, so advanced a stage of sexual maturity has never before been observed in the Common Eel. This appears to be due to the fact that the males hitherto examined had not yet migrated into the deep water of the sea. Eels with big eyes taken from the depths of the sea were, before the above facts were known, described as a distinct species under the name of Anguilla bibroni (Kaup) and of Anguilla kieneri (Kaup), not to be confounded with Anguilla kieneri (Giinther), which is a synonym of Lycodes 'kieneri. In certain cloacae of ancient Rome which to-day are disused and contain pure water, remarkable eels are found of a length of from 20 — 30 cm. of a grey colour, without trace of yellow, of male and female sex, with enormous eyes and with more or less rudimentary genital organs. They are individuals which, confined in a place without light, have acquired prematurely one of the characters of the bridal habit without a corresponding development of the genital organs. These individuals are probably incapable of ulterior de- velopment, as the condition of their genital organs seems to demon- strate. Under the name Anguilla kieneri (Kaup) there have probably been included some individuals which had acquired big eyes under con- ditions similar to those described for the eels of these Roman cloacae. From these and similar observations it clearly results that all the European eels must be included under a single species, and this is an important fact from another point of view, namely, that it destroys an objection which might be raised against my conclusion with regard to the development of Anguilla vulgaris from Leptocephalus brevirostris, namely, the objection that Leptocephalus brevirostris belongs not to Anguilla vulgaris, but to Anguilla kieneri, or to Anguilla bibroni. To sum up, Anguilla vulgaris, the Common Eel, matures in th& depths of the sea, where it acquires larger eyes than are ever observed in individuals which have not yet migrated to deep water, with the exception of the eels of the Roman cloacae. The abysses of the sea are the spawning places of the Common Eel : its eggs float in the sea water. In developing from the egg, it undergoes a metamorphosis, that is to say, passes through a larval form denominated Leptoce- phalus brevirostris. What length of time this development requires is very difficult to establish. So far we have only the following data : — First, Anguilla vulgaris migrates to the sea from the month of October to the month of January ; second, the currents, such as. those of Messina, throw up, from the abysses of the sea, specimens which, from the commencement of November to the end of July,. Eclipse of the Sim, 1896. — Novaya Zemlya Observations. 271 tire observed to be more advanced in development than at otber times, but not yet arrived at total maturity ; third, eggs, which according to every probability belong to the Common Eel, are found in the sea from the month of August to that of January inclusive ; fourth, the Leptocephalus brevirostris abounds from February to September. As to the other months, we are in some uncertainty, because during them our only natural fisherman, the Orthagoriscus mola, appears very rarely; fifth, I am inclined to believe that the elvers ascending our rivers are already one year old, and I have observed that in an aquarium specimens of L. brevirostris can transform themselves into young elvers in one month's time. " Total Eclipse of the Sun, 1896.— The Novaya Zemlya Observations." By Sir GEORGE BADEN-POWELL, K.C.M.G., M.P. Communicated by J. NORMAN LOCKYER, C.B., F.R.S. Received November 19, — Read November 19, 1896. (Abstract.) The author gives an account of the circumstances under which it became desirable to fit out an expedition to observe the eclipse in Novaya Zemlya, and the arrangements made to convey it by his yacht " Otaria." Details are given of the observing station, the erection of the dif- ferent instruments, and the scheme of work. The valuable spectroseopic results obtained are still under process of being worked out ; but the coronagraph results are reported in detail, and copies of the chief photographs are appended. The meteorological and other conditions during the eclipse are duly recorded. ^Preliminary Report on the Results obtained with the Pris- matic Camera during the Eclipse of 1896." By J. NORMAN LOCKYER, C.B., F.R.S. Received November 17, — Read November 19, 1896. (Abstract.) The author first states the circumstances under which Sir George Baden-Powell, K.C.M.G,, M.P., with great public spirit conveyed an eclipse party to Novaya Zemlya in his yacht " Otaria," to which party was attached Mr. Shackleton, one of the computers employed by the Solar Physics Committee. The prismatic camera employed, loaned from the Solar Physics 272 List of Officers and Council nominated for Election. Observatory, was carefully adjusted before leaving England, and a programme of exposures was drawn up based upon the experience of ]893. As the station occupied lay at some distance from the central line, this programme was reduced by Mr. Shackleton. Two of the photographs obtained are reproduced for the informa- tion of other workers, as some time must elapse before the discussion of all the results can be completed. This discussion and Mr. Shackle- ton's report on the local arrangements and details of work, are promised in a subsequent communication. The lines photographed in the " flash " at the commencement of totality — happily caught by Mr. Shackleton — the wave-lengths of which lines have been measured by Dr. W. J. S. Lockyer, show interesting variations from those photographed by Mr. Fowler in the cusp during the eclipse of 1893. ' With the exception of the lines visible in the spectra of hydrogen and helium, and the longest lines of many of the metallic elements, considerable differences of intensity from the lines of Fraunhofer are noticeable. The coronal rings have been again photographed, and the results of 1893 have been confirmed. November 26, 1896. Sir JOSEPH LISTER, Bart., President, in the Chair. Dr. George Murray and Professor Karl Pearson were admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. In pursuance of the Statutes, notice of the ensuing Anniversary Meeting was given from the Chair, and the list of Officers and Council" nominated for election was read as follows : — President. — Sir Joseph Lister, Bart., F.R.C.S., D.C.L. Treasurer.— Sir John Evans, K.C.B., D.C.L., LL.D. Secretaries. — / Professor Michael Foster, M.A., M.D. I Professor Arthur William Biicker, M.A., D.Sc. Foreign Secretary. — Edward Frankland, D.C.L., LL.D. Other Members of the Council. — Professor William Grylls Adamsr M.A.; Professor Thomas Clifford Allbutt, M.D. ; Professor Robert Mathematical Contributions to the Theory of Evolution. 27;} Bellamy Clifton, M.A. ; William Turner Thiselton Dyer, C.M.G. ; Professor James Alfred Ewiiig, M.A. ; Lazarus Fletcher, M.A. ; Walter Holbrook Gaskell, M.D. ; Professor Alfred George Greenhill, M.A • William Huggins, D.C.L. ; Professor Charles Lapworth, LL.D. \ Major Percy Alexander MacMahon, R.A.; Professor Raphael Meldola', F.C.S.; Professor William Ramsay, Ph.D. ; The Lord Walsinglmm' M.A. ; Professor Walter Frank Raphael Weldon, M.A. ; Admiral William James Lloyd Wharton, C.B. The following Papers were read: — I. "Mathematical Contributions to the Theory of Evolution. On • Telegony in Man, &c." By KARL PEARSON, F.R.S., University College, with the assistance of Miss ALICE LEE, Bedford College, London. II. " On the Magnetic Permeability of Liquid Oxygen and Liquid Air." By J. A. FLEMING, M.A., D.Sc., Professor of Electrical Engineering in University College, London, and JAMES DEWAK, LL.D., F.R.S., Fullerian Professor of Chemistry in the Royal Institution. " Mathematical Contributions to the Theory of Evolution. On Telegony in Man, &c." By KARL PEARSON, F.R.S., Uni- versity College, with the assistance of Miss ALICE LEE, Bedford College, London. Received August 27, — Read November 26, 1896. (1) The term telegony has been used to cover cases in which a female A, after mating with a male B, bears to a male C offspring having some resemblance to or some peculiar characteristic of A's first mate B. The instances of telegony usually cited are (i) cases of thoroughbred bitches when covered by a thoroughbred dogt reverting in their litter to half-breds, when they have been previously crossed by dogs of other races. Whether absolutely unimpeachable instances of this can be produced is, perhaps, open to question, but the strong opinion on the subject among dog-fanciers is at least remarkable; (ii) the case of the quagga noted by Darwin (see * Origin of Species,' 4th edition, p. 193), and still more recently (iii) a noteworthy case of telegony in man cited in the 'British [edical Journal' (see No. 1834, February 22, 1896, p. 462). In this latter case a very rare male malformation, which occurred, in the male B, was found in the son of his widow A, by a second msband C. Here, as in the other cases cited, a question may always- raised as to the possibly unobserved or unknown occurrence of the 274 Prof. Karl Pearson. characteristic in the ancestry of either A or C, or again as to the chance of the characteristic arising as a congenital sport, quite inde- pendently of any heredity. It seems unlikely that the observation of rare and isolated cases of asserted telegony will lead to any very satisfactory conclusions, although a well-directed series of experi- ments might undoubtedly do so. On the other hand, it is not impos- sible than an extensive and careful system of family measurements might bring to light something of the nature of a telegenic influence in mankind. If such a telegenic influence really exists, it may be supposed to act in at least two and, very possibly, more ways. (a) There may be in rare and isolated cases some remarkable change produced in the female by mating with a particular male, or some remarkable retention of the male element. (6) There may be a gradually increasing approximation of the female to the male as cohabitation is continued, or as the female bears more and more offspring to the male. It is extremely unlikely that any system of family measurements would suffice to bring out evidence bearing on (a). On the other hand, a closer correlation between younger children and the father, and a lesser correlation between younger children and the mother, as compared with the correlation between elder children and their parents might, perhaps, indicate a steady influence like (&) at v\rork in mankind. Shortly, such measurements might suffice to answer the question as to whether younger children take more after their father and less after their mother than elder children. Without hazarding any physiological explanation as to the mode in which telegonic influence can or does take place, we may still hope to get, at any rate, negative evidence as to a possible steady telegonic influence by an investigation of suitable family measurements. (2) Unfortunately, the collection of family data is by no means an easy task, and to procure those head-measurements, which, I think, would be most satisfactory for the problem of heredity, would require a large staff of ready assistants, and could only be undertaken on the necessary scale by the action of some scientific society or public body. The data concerning 800 to 900 families which have been recently collected for me deal only with stature, span, and arm-length, which are measurable with more or less accuracy by the untrained observer, and are only suitable for more or less rough appreciations of hereditary influence. The numbers in each family measured were strictly limited, in order to remove the influence of reproductive selection from the determination of the correlation between parents and children, and the result of this limitation has been that compara- tively few couples of elder and younger brothers, and of elder and younger sisters are available. They were, indeed, collected in the Mathematical Contributions to the Theory of Evolution. 275 first place with a view to the problem of heredity in the direct line, and with no thought of their throwing any light on the problem of telegony. That steady telegonic influence might be deduced from such family data has only recently occurred to me, and I should now hesitate to publish any conclusions on this subject, based on some- what mixed and sparse returns, did I not consider that it may be a long time before more extensive returns are available, and that the publication of this method of dealing with telegony may induce others to undertake the collection of a wider range of material. My own 800 family data cards did not provide a sufficiently large number of either brother- brother or sister-sister couples to give a strong hope of a difference between the correlation coefficients sufficiently large as compared with its probable error to base any legitimate conclusion upon. I, therefore, again borrowed from Mr. Galton his 200 family data returns, and from these 1,000 families was able to select 385 brother-brother pairs and 450 sister-sister pairs. In these statistics each individual is only included in one pair, and the difference in age between the elder and younger mem- bers of each pair differs very widely from pair to pair. In some cases there may be several years between the ages and several intervening children ; in others the members of the pair may be successive children following each other in successive years. In each case all we can say is, that if there be a steady telegonic influence, the rela- tion of the elder member to the parent will weigh down the same scale, and in the final result we ought to find a distinctly greater or less correlation, as the case may be. I think a more serious objection to the data than the variation in the number of years between fraternal pairs is the mixture I have made of data collected at different periods and in somewhat different manners. My own data are drawn, I think, from a wider class of the community than Mr. Galton's. They are not exclusive of his class, but, I think, cover his class, and go somewhat further down in the social scale. They suffice to show that the means and variations change considerably from one social stratum to another, and what is still more important that the Galton-Functions or coefficients of correlation for heredity are far from being constant even within the same race, as we pass from one rank of life to a second. Thus, my means for stature in the case of both fathers and mothers are upwards of ^ in. less than Mr. Galton's, but my means agree fairly well with his results in the case of both sons and daughters. There are also good agreements and somewhat puzzling disagreements not only in the variations, but, above all, in the coefficients of correlation for heredity. I reserve for the present the full discussion of my heredity data, but I wish it to be quite understood that my conclusions in this paper are based, not upon the best possible data, e.g., measurements made on one class of the com- VOL. LX. Y 276 Prof. Karl Pearson. munity under one system, but upon all the data which, for some time to come, appear likely to be available. These data are neither quan- titatively nor qualitatively ideal, but, on the other hand, they must be given a reasonable amount of weight in considering whether, at any rate in the case of one organ — stature, — any steady telegenic influence can be traced in man. The reduction from the family measurement-cards, the formation of the eight correlation tables, and the calculation of both variation and correlation coefficients have been undertaken by Miss Alice Lee of Bedford College, — a task requiring much labour and persistency. I have independently verified, and in some minor points corrected her calculations, as well as added the probable errors of the con- stants determined. (3) The following are the means and standard-deviations with their probable errors for the various groups. Table I. — Stature of Families in Inches. Class. Number. Mean. Standard deviation. Fathers of sons . . • 385 68 -5740 ±0-0878 2 -5554 ±0-0621 Elder sons . . . . . . 69'1494±0'0913 2 -6550 ±0-0645 69 -1948 ±0-0933 2 -7128 ±0-0659 63 3078 ±0-0854 2 '4848 ±0-0604 Fathers of daughters . • • . . 450 68 -3344 ±0-0878 2*7605 ±0-0621 63 -9244 ±0-0823 2'5878±0-0582 64 -2200 ±0-0794 2 -4985 ±0-0562 Alothers of daughters . . 63 -1794 ±0-0758 2 -3827 ±0-0536 All the quantities have here been calculated precisely as in my third memoir on the mathematical theory of evolution (see ' Phil. Trans./ A, vol. 187, pp. 270 — 271). In this case, however, no child is included twice as a child, and parents are not weighted with their offspring. Thus reproductive selection is not allowed to influence the results. It will be seen that the probable errors of the means and standard deviations are, as in the former paper, too large to allow of absolutely definite conclusions when those conclusions are not sup- ported by a continuous change of values, or directly verified by the numbers of the earlier memoir. But one or two such conclusions may be drawn, and I will note them before passing to correlation. (i) The law of sexual interchange referred to in my former paper (p. 274) is confirmed with greater uniformity. Fathers of sons are sensibly less variable than fathers of daughters, and mothers of daughters are sensibly less variable than mothers of sons. In other Mathematical Contributions to the Theory of Evolution. 277 words, to judge from stature, the exceptional parent tends to have offspring of the opposite sex. (ii) Younger sons are taller and more variable than elder sons, and elder sons are taller and more variable than fathers. This conclusion, although less markedly, appears in the results on pp. 270 and 281, of my former paper. It might be accounted for by : (a) A secular change going on in the stature of the population, and even noticeable in the difference between the stature of younger and elder sons. (6) A further growth of sons, and an ultimate shrinkage, which will leave them at the age of their fathers with the same mean height and variation. (c) Conditions of nurture on the average less favourable, and on the whole less varied in the case of elder than in that of younger children.* (d) Natural selection. The difference between younger and elder sons and between elder sons and fathers represents the selective death rate in man due to causes correlated with stature in the years between youth and manhood, and man- hood and age. The difference is thus to be accounted for by a periodic and not a secular change. Possibly (a), (6), (c), and (d), may all contribute to the observed' results. It cannot be denied that (d) has a special fascination of its own for the student of evolution, but prolonged study of the laws of growth must precede the assertion that we have here, or in any similar case, real evidence of an actual case of natural selection. (iii) Younger daughters are taller than elder daughters and elder daughters than mothers. This is in complete agreement with the result for fathers and sons. Further : Daughters, as a class are far more variable than mothers, but while in the earlier memoir younger daughters were sensibly more variable than elder daughters — and thus exactly corresponded with sons — elder daughters are in this case more variable than younger. I have been unable to find any slip in the tables or calculations, which might account for this divergence. It exceeds considerably the probable error of the observations, and is not in accordance with the general law connecting the variation of parent and offspring evi- denced for both sexes in the earlier, and for sons in the present memoir — e.g., the variation — whether it be due to growth- change, * Mr. Francis Galton suggests this as a possible cause. It has, I think, to be taken in conjunction with a greater amount of parental experiment, not only in the birth, but in the nurture of the elder children. y 2 278 Prof. Karl Pearson. or to selective death-rate, or to secular evolution — diminishes with age. (4) The following are the coefficients of correlation (r) and the coefficients of regression (B) for parents and sons : Table II. — Inheritance of Stature by Sons. Father and elder sons Father and younger sons .... Mother and elder sons Mother and younger sons . . . r. 0-4120 ±0-0264 0-4170 ±0-0262 0'4094± 0-0265 0-4111 ±0-0264 R. 0 -4281 0-4427 0 -4374 0 -4488 If we measure, as seems reasonable, the hereditary influence of parentage by the magnitude of the coefficient of correlation between parent and offspring, then several important conclusions may be drawn from this table. (i) There is no sensible difference between the influences of the father on younger and on elder sons, and no sensible difference between the influences of the mother on younger and on elder sons. If we pay attention to such slight differences as exist, there would appear, not to be an increase of paternal and a decrease of maternal influence on younger children, but an extremely slight increase of both. In other words, so far as stature in sons is concerned, judged by correlation : No steady telegonic influence exists. (ii) There is a very slight prepotency of the father over the mother in the case of both younger and elder sons ; a prepotency which will be slightly magnified when account is taken of the abso- lute stature of the two parents. But the great prepotency of paternal inheritance noticed in the former memoir is not confirmed. The co-efficients of maternal in- heritance have been increased by more than 30 per cent, (from O293 to 0'410), while those of paternal inheritance (0'396 as compared with 0*414) have remained almost stationary. This result seems to show the want of constancy of the Galton's functions for heredity within the same race. An explanation on the ground that the present statistics embrace a wider range of the community than the earlier, and possibly a more closely correlated class,* fails, at any rate in part, owing to the sensible constancy of the paternal correla- tion. The main difference of course between the present and the former statistics is the exclusion of the influence of reproductive * I have pointed out (loc. cit., p. 284) that working and lower middle class families appear to be more closely correlated than those of the upper middle class. Mathematical Contributions to the Theory of Evolution. 279 selection, but why should this be expected to influence only the mother ? The father of many children remains equally influential, but the mother's relation is weakened when we give weight to the quantity not the relative ages of her children. This is not a steady telegonic influence, but a correlation between fertility and heredi- tary influence in mothers, which if it could be verified by further observation, would undoubtedly be of high significance. I would accordingly suggest as a possible law of heredity, deserving careful investigation, that : Hereditary influence in the female varies inversely as fertility. In my paper on " Reproductive Selection," (' Roy. Soc. Proc.,' vol. 59, p. 301), I have pointed out the important evolutionary results which flow from a correlation between fertility and any inheritable characteristic. If a law of the above character should be established after further investigation, it is conceivable that it may act as an automatic check on the extreme effects of reproductive selection. (iii) The above results give us for practicable purposes a quite sufficiently close value of the correlation between parents and sons, when the influence of reproductive selection is excluded. Judging from stature the correlation between sons and parents is very closely given by 0-41 ±0-03. The J, adopted by Mr. Galton, may, I think, safely be increased by 25 per cent., and further, the assumption that collateral heredity is twice as strong as direct heredity must, I hold, be finally discarded, for no determination of the former has given such a high value as 0-82. (5) Hitherto we have regarded only the coefficients of correlation, and considered them to measure the strength of the hereditary in- fluence, but it must be remembered that the means of elder and younger sons are not the same, and that there is another way of looking at the problem. We may ask : Do younger or elder sons differ most from the stature of their father, and is the order altered in the case of the mother ? If we neglect the influence of sexual selection (see " Contributions to Math. Theory of Evolution," 111, pp. 287 — 8) we have, if hf and hm be deviations of father and mother from their means, and M, and My be mean heights of corresponding fraternities of elder and younger sons in inches : M, = 69-1494 + 0-4281fc/+0-4374/i,B. My = 691948 + 0-4427 V+0-4488fc». Now the ratio of the mean heights of parents is 68'5740 : 63'3078 = 280 Prof. Karl Pearson. 1-0832,* while the ratios of 0*4374 to 0-4281 and 0*4488 to 0*4427, are only T0219 and 1*0139 respectively, thus there is still a slight prepotency of paternal influence on stature to be recorded. (See § (4) (ii).) Confining our attention to the differences in stature for fathers and sons corresponding to all mothers whatsoever, we have, if Def be the difference in stature between father and corresponding fraternity of elder sons, D^ between father and fraternity of younger sons : Def= 0-5754-0-5719^. Dyf= 0-6208- 0-5573 hf. Hence the difference betwen the father and fraternity of younger sons will be greater than the difference between the father and the corresponding fraternity of elder sons unless the father be 3'110 inches less, or 1*059 more than the average. But 3*11 is about 1*2 and 1*059 about 0*415 times the standard deviation of the stature of fathers, or, fraternities of younger sons are nearer in stature to their father than fraternities of elder sons in about 46 per cent, of cases. Similarly if Dgnt, ~Dym represent the differences of stature of mothers and fraternities of elder and younger sons respectively, we have in inches Dem = 5*8416- 0-5626/4. -Dyrn = 5-8870-0-5512AOT. Thus fraternities of younger sons are always more divergent than fraternities of elder sons from the stature of their mothers, unless the mother be 3*982 inches less, or 10*53 inches more than the average. These are 1'6 and 4*24 times the standard deviation in stature of mothers ; or, only in about 5*5 per cent, of cases are fraternities of younger sons nearer in stature to their mothers than elder sons. Now, it is difficult to read into these results any evidence for a steady telegenic influence. It is true that the case of younger sons being more like their parents than elder sons occurs in eight times as many cases with the father as with the mother, but the broad fact remains that in more than half the cases, judged by difference of stature, the elder son is more like the father than the younger son. In fact, examined in this way by difference of stature — not an un- . natural manner of first approaching the problem — the true closeness of parent and offspring appears to be quite obscured by some secular, or, at any rate, periodic (see § 3) evolution in stature between successive generations — an evolution which even makes itself felt in the interval between younger and elder sons. * 13/12 = 1*0833; thus these returns again confirm Mr. Galton's selection of this fraction for the sexual ratio for stature. Mathematical Contributions to the Theory of Evolution. 281 (6) Turning to the results for daughters, we have the following table for the coefficients of correlation and regression : — Table III. — Inheritance of Stature by Daughters. Fathers and elder daughters Fathers and younger daughters Mothers and elder daughters Mothers and younger daughters . . . 0-4829 ±0-0220 0-4376 ±0-0236 0-3953 ±0'C250 0-4542 ±0-0230 E. 0 -4528 0 -396L 0 -4293 0 -4763 These results, more numerous than those for sons, are, for reasons which I am unable to explain, much more divergent. We may note the following points : — (i) There is a sensible difference between the coefficients of corre- lation for either parents with younger and elder daughters. Thus, the difference of the coefficients for fathers with elder and younger daughters is 0*0453, and the probable error of this only 0*032 ; while for mothers the corresponding difference is 0*0589, and the probable error of the difference only O0328. The difference, however, is in the opposite sense. We are thus face to face with an increasing maternal and a decreasing paternal influence on the stature of daughters. In other words, our statistics are entirely opposed to any steady telegonic influence on the sfcature of daughters. If such a thing were conceivable, we should be confronted with the case of the mother influencing the father, the reverse of telegony. (ii) The mean correlation of fathers and daughters is very slightly higher than that of mothers and daughters (0*4602 as compared with 0*4247) . Thus, to judge by the mean coefficients of correlation, the father is slightly more prepotent than the mother in heredity. The mean coefficients of regression are for fathers 0*4244, and for mothers 0'4528, or in the ratio of 1 : T067, but the ratio of the paternal to the maternal stature is T083, or this slight prepotency is still pre- served if we judge the matter by regression coefficients. Again, we notice an immense increase (0*2841 to 0'4247) in the correlation between mothers and daughters when we compare the present results with those of my earlier memoir. As an explanation of this, I have already suggested the possibility of a law exhibiting a relation between fertility and hereditary influence in mothers (§ 4 (ii) ). (iii) The mean coefficient of correlation in stature between either parent and a daughter may be taken to be — • 0*44±0'02. Mathematical Contributions to the Theory of Evolution. Thus, it does not differ very widely from the value suggested (0'41) for sons, but is even further removed from the value (0'33) at first determined by Mr. Gralton. The greater correlation between sons and both parents noticed in my first memoir is not borne out by the present statistics ; the advantage is now — it is true to a much less extent — with daughters. On the whole, I am not well satisfied with these results for daughters. I can see no persistent source of error in the method of collecting the observations, nor can I find any mistake in the calcu- lations. I can only trust that more elaborate returns and measure- ments of other characteristics may some day throw light on what now appear to be anomalies. (7) Finally, I may just notice what conclusions are to be drawn, if we pay attention to the absolute difference in stature between parents and daughters. Let Sem and dym be the differences in stature between elder daughters and mothers, and younger daughters and mothers respectively, then in inches we have for the corresponding arrays : cem = 0-7450-0-5707^. fy* = 1-0406-0'5237A«. Thus, arrays of younger daughters differ more from their mothers in stature than arrays of elder daughters, if the mothers be more than 6'29 in. below the mean or more than 1*63 in. above the mean, or if their deviations are not within the limits of about — 2'64 and 0'68 times the standard deviation of mothers. This gives us about 74 to to 75 per cent, of elder sisters nearer in stature to their mothers than younger sisters. If fye, Sfy be the stature differences for fathers and daughters, we have ; ty = 4-4100-^0-5472/y. fo = 4-1144-0-6039/i/. Here, so long as the father lies between 5'21 in. less and 7'41 in. more than the average, the array of younger daughters will more nearly approach him in stature than the array of elder daughters. These limits correspond to 1'89 and 2'68 times the standard devia- tion of fathers. Accordingly, about 90 to 97 per cenfc. of younger sisters are closer in stature to their fathers than elder sisters. Thus, if we had started the discussion of the problem from a consideration of the relative nearness in stature of daughter to father and mother, we should have found that a great majority of younger sisters were nearer to their fathers than their elder sisters, and a considerable majority of elder sisters nearer to their mother than their younger sisters. We might then have concluded that there were substantial Magnetic Permeability of Liquid Oxygen and Liquid Air. 283 grounds for inferring the existence of a telegonic influence. But it is clear that if there be anything of the nature either of a periodic or of a secular change in stature going on, then since men are taller than women, any group of younger women will appear closer to their fathers than to their mothers, when compared with a group of elder sisters. Thus, no legitimate argument as to a telegonic influence can be based on such a result. I have purposely considered this method of approaching the problem, because it is the method whioK first occurred to me, as it probably may do to others. It can very easily, however, lead to our mistaking for a real telegonic influence an effect of periodic or secular evolution, or, indeed, of different con- ditions of nurture. (7) In conclusion, we may, I think, sum up the statistics dis- cussed in this paper as follows : — (i) So far as stature is concerned there is no evidence whatever of a steady telegonic influence of the male upon the female among mankind. (ii) It is improbable that the coefficients of correlation which measure the strength of heredity between parents and off- spring are constant for all classes even of the same race. For stature in the case of parents and offspring of both sexes, the value 0'42, or say 3/7, may be taken as a fair working value, until more comprehensive measurements are made. This makes heredi- tary influence in the direct line stronger than has hitherto been supposed. (iii) The divergence between the results of this memoir and that of the former memoir on " Regression, Heredity, and Pan- mixia " would be fairly well accounted for, if there be a hitherto unobserved correlation between the hereditary influence and the fertility of woman. t; On the Magnetic Permeability of Liquid Oxygen and Liquid Air." By J. A. FLEMING, M.A., D.Sc., F.R.S., Professor of Electrical Engineering in University College, London, and JAMES DEWAR, LL.D., F.R.S., Fullerian Professor of Chemistry in the Royal Institution, &c. Received Novem- ber 20,— Read November 26, 1896. The remarkable magnetic properties of liquid oxygen were pointed out by one of us in a communication to the Royal Society in 1891,* * ' Eoy. Soc. Proc.,' December 10th, 1891, vol. 51, p. 24. See a letter to the President by Professor James Dewar, F.E.S. 284 Profs. J. A. Fleming and J. Dewar. On the and were subsequently described to the Royal Institution in a lecture delivered in 1892.* We have for some time past directed our attention to the question of determining the numerical values of the magnetic permeability and magnetic susceptibility of liquid oxygen, with the object of determining not only the magnitude of these physical con- stants, but also whether they vary with the magnetic force under which they are determined. Although a large number of determinations have been made by many observers of the magnetic susceptibility of different liquids taken at various temperatures, difficulties of a particular kind occur in dealing with liquid oxygen. One method adopted for determining the magnetic susceptibility of a liquid is to observe the increase of mutual induction of two conducting circuits suitably placed, first in air, and then when the air is replaced by the liquid in question, the susceptibility of which is to be determined. A second method con- sists in determining the mechanical force acting on a known mass of the liquid when placed in a non-uniform magnetic field. Owing to the difficulty of preventing entirely the evaporation of liquid oxygen, even when contained in a good vacuum vessel, and the impossibility of sealing it up in a bulb or tube, and having regard to the effect of the low temperature of the liquid in deforming by contraction and altering the conducting power of coils of wire placed in it, it was necessary to devise some method which should be indepen- dent of the exact constancy in mass of the liquid gas operated upon, and independent also of slight changes in the form of any coils of wire which might be used in it. After many unsuccessful preliminary experiments the method which was finally adopted as best complying with the conditions introduced by the peculiar nature of the substance operated upon is as follows : — A small closed circuit transformer was constructed, the core of which could be made to consist either of liquid oxygen or else immediately changed to gaseous oxygen, having practically the same temperature. This transformer consisted of two coils, the primary coil was made of forty-seven turns of No. 12 S.W.G. wire, this wire was wound into a spiral, having a rectangular shape, the rectangular turns having a length of 8 cm. and a width of 1'8 cm. This rectangular-sectioned spiral, consisting of one layer of wire of forty-seven turns, was bent round a thin brass tube, 8 cm. long and 2-| cm. in diameter, so that it formed a closed circular solenoid of one layer of wire. The wire was formed of high conductivity copper, doubly insulated with cotton, and each single turn or winding having a rectangular form. The turns of covered wire closely touched each other on the inner circumference of the toroid, but on the external circumference were * See 'Roy. Inst. Proc.,' June 10th, 1892, "On the Magnetic Properties of Liquid Oxygen." Friday evening discourse, by Professor J. Dewar, F.R.S. Magnetic Permeability of Liquid Oxygen and Liquid Air. 285 a little separated, thus forming apertures by which liquid could enter or leave the annular inner core. The nature of this transformer is shown in Fig. 1. FIG. 1. Diagram of the Closed Circuit Transformer used in the Experiments. The mean perimeter of this rectangular-sectioned endless solenoid was 13 J cm., and the solenoid had, therefore, very nearly 3*5 turns per cm. of mean perimeter. When immersed in liquid oxygen a coil of this kind will carry a current of 50 amperes. When a current of A amperes is sent through this coil the mean magnetising force in the axis of this solenoid is, therefore, represented by 4*375 times the current through the wire, hence it is clear that it is possible to produce in the interior of this solenoid a mean magnetising force of over 200 C.Gr.S. units. This primary coil had then wound over it, in two sections, about 400 or 500 turns of No. 26 silk-covered copper wire to form a secondary coil. The primary and secondary coils were sepa- rated by layers of silk ribbon. The exact number of turns was not counted, and as will be seen from what follows it was not necessary Profs. J. A. Fleming and J. Dewar. On the to know the number. The coil so constructed constituted a small induction coil or transformer, with a closed air-core circuit, but which when immersed in a liquid, by the penetration of the liquid into the interior of the primary coil, became changed into a closed circuit transformer, with a liquid core. The transformer so designed was capable of being placed underneath liquid oxygen contained in a large vacuum vessel, and when so placed formed a transformer of the closed circuit type, with a core of liquid oxygen. The coefficient of mutual induction of these two circuits, primary and secondary, is therefore altered by immersing the transformer in liquid oxygen> but the whole of the induction produced in the interior of the primary coil is always linked with the whole of the turns of the secondary coil, and the only form-change that can be made is a small change in the mean perimeter of the primary turns due to the con- traction of the coil as a whole. In experiments with this transformer the transformer was always lifted out of the liquid oxygen into the cold gaseous oxygen lying on the surface of the liquid oxygen, and which is at the same temperature. On lifting out the transformer, the liquid oxygen drains away from the interior of the primary coil, and is replaced by gaseous oxygen of very nearly the same tem- perature. The vacuum vessel used had a depth of 60 cm. outside and 53 cm. inside, and an internal diameter of 7 cm. It held 2 litres of liquid oxygen when full ; but, as a matter of fact, 4 or 5 litres of liquid oxygen were poured into it in the course of the experiment. Another induction coil was then constructed, consisting of a long- cylindrical coil wound over the four layers of wire, and a secondary circuit was constructed to this coil, consisting of a certain number of turns wound round the outside of the primary coil, and a small adjusting secondary coil, consisting of a thin rod of wood wound over with very open spirals of wire. The secondary turns on the outside of the primary coil were placed in series with the turns of the thin adjusting coil, and the whole formed a secondary circuit, partly out- side and partly inside the long primary cylindrical coil, the coefficient of mutual induction of this primary and secondary coil being capable of being altered by very small amounts by sliding into or out of the primary coil the small secondary coil. This last induction coil, which will be spoken of as the balancing coil, was connected up to the small transformer,, as just described, as follows : — The primary coil of the small transformer was connected in series with the primary coil of the balancing induction coil, and the two terminals of the series were connected through a reversing switch and ammeter with an electric supply circuit, so that a current of known strength could be reversed through the circuit, consisting of the two primary coils in series. The two secondary coils, the one on Magnetic Permeability of Liquid Oxygen and Liquid Air. 287 the transformer and the one on the balancing induction coil, were con- nected in opposition to one another through a sensitive ballistic galvanometer in such a manner that on reversing the primary current the galvanometer was affected by the difference between the electromotive forces set up in the two secondary coils, and a very flue adjustment could be made by moving in or out the adjusting coil of the balancing induction coil. The arrangement of circuits is shown in fig. 2. FIG. 2. /WW Ww Arrangement of the Circuits of the Transformer and Induction Coil. For the purpose of standardising the ballistic galvanometer employed, the primary coil of the balancing induction coil could be cut out of circuit, so that the inductive effect in . the ballistic galvanometer circuit was due to the primary current of the closed circuit transformer alone. A resistance box was also included in the circuit of the ballistic galvanometer. The resistance of the ballistic galvanometer was about 18 ohms, and the resistance of the whole secondary circuit 30'36 ohms. The experiment then consisted in first balancing the secondary electromotive forces in the two coils exactly against one another, then immersing the transformer in liquid oxygen, the result of which was to disturb the inductive balance, and in consequence of the magnetic permeability of the liquid oxygen core being greater than unity, a deflection of the ballistic galvanometer was observed on reversing the same primary current. The induction 288 Profs. J. A. Fleming and J. Dewar. On the through the primary circuit of the small transformer is increased in the same proportion that the permeability of the transformer core is increased by the substitution of liquid oxygen for gaseous oxygenr and hence the ballistic deflection measures at once the amount by which the magnetic permeability of the liquid oxygen is in excess over that of the air or gaseous oxygen forming the core of the trans- former when the transformer is lifted out of the liquid. As a matter of fact it was nfiver necessary to obtain the inductive balance pre- cisely. All that was necessary was to observe the throw of the bal- listic galvanometer, first when the transformer was wholly immersed under the surface of liquid oxygen, and, secondly, when it was lifted out into the gaseous oxygen lying on the surface of the liquid, the strength of the primary current reversed being in each case the same. In order to standardise the galvanometer and to interpret the meaning of the ballistic throw, it was necessary to cut out of circuit the primary coil of the balancing induction coil, and to reverse through the primary circuit of the small transformer a known small primary current, noting at the same time the ballistic throw pro- duced on the ballistic galvanometer, this being done when the transformer was underneath the surface of liquid oxygen. It will be seen, therefore, that this method requires no calculation of any coefficient or mutual induction, neither does it involve any know- ledge of the number of secondary turns on the transformer, nor of the resistance of the secondary circuit ; all that is necessary for a successful determination of the magnetic permeability of the liquid oxygen is that the secondary circuit of the transformer should remain practically of the same temperature during the time when the throw of the ballistic galvanometer is being observed, both with the transformer underneath the liquid oxygen and out of the liquid oxygen. If then the result of reversing a current of A amperes through the two primary coils in series when the secondary coils are opposed is to give a ballistic throw, D, and if the result of reversing a small current a amperes through the primary coil of the transformer alone is to produce a ballistic throw, d, then if p is the magnetic permeability of liquid oxygen, that of the gaseous oxygen lying above the liquid and at the same temperature being taken as unity, we have the following relation : — "T — = P- — !> -d a which determines the value of /JL. Deferring for a moment the correction to be applied to determine the value of the magnetic permeability of liquid oxygen in terms of that of a vacuum, the following are the results of observation : — Magnetic Permeability of Liquid Oxygen and Liquid Air. 289 OBSERVATIONS ON MAGNETIC PERMEABILITY OP LIQUID OXYGEN. Throws of Ballistic Galvanometer. Induction Coils balanced. { 4 '0 mm. to left T The transformer in liquid oxygen. 4 '2 „ „ V Primary current = 37 '8 amperes reversed 4 *3 „ „ J through primary coils. I "] The transformer lifted out of liquid oxygen 17 -0 mm. to right | int,n 00ia gaacous oxygen at the same tem- Exp. II. Current = 28 '8 amperes through primary 4-2 , J coils. 290 Exp. X. Profs. J. A. Fleming and J. Dewar. On the right f I 12 -| 12 -0 j 12*2 ^ ""I The transformer lifted out of liquid oxygen '0 mm. to right into cold gaseous oxygen at the same tem- perature. Current = 28 '1 amperes reversed through J primary coils. The transformer in liquid oxygen. {1 e ransormer n qu oxygen. I Current = 28 '1 amperes reversed through " " J primary coils. Exp. XII. The transformer in liquid oxygen. The transformer lifted out of liquid oxygen into cold gaseous oxygen at same temperature. Current reversed in primary coils, in amperes. Ballistic throw in millimetres. Deflection to the right. Current reversed in primary coils, in amperes. Ballistic throw in millimetres. Deflection to the right. 58-8 50-2 50-2 10-5 15-0 17-0 50-2 50-8 50-0 47-0 48-5 49-0 The above table shows the results of the observations made with the small transformer alternately placed underneath the surf ace of liquid oxygen, and then lifted up into the cold gaseous oxygen lying above the surface of the liquid oxygen. It will be noticed that the ballistic throws in each set of observations are not constant, but that there is a tendency, usually, for the throw to increase if repeated, whilst the transformer is still maintained in the same condition, This is in all probability due to the fact that the continued passage of the primary current heats the primary circuit of the balancing induction, coil, and hence heats, also, by radiation, the secondary coil of the balancing induction coil, and, therefore, by enlarging the area of the adjusting coil, continually breaks down the inductive balance. It was found necessary, therefore, to take the observations in groups at equal intervals of time. First, a group of three observations was taken, the transformer being in liquid oxygen, the balance being, as nearly as possible, obtained. Then the transformer was lifted out of the liquid oxygen, and the ballistic throws again taken, reversing the same primary current ; next again immersed in liquid oxygen, and finally once more taken out of the liquid oxygen. Taking the sets Magnetic Permeability of Liquid Oxygen and Liquid Air. 291 of observations marked I, II, III, IV, the mean of the means of the three observations in Sets I and III, corrected for the variation in the primary current, were taken as the result of the measurement in liquid oxygen, and this result was then compared with the ballistic- throws in Set II. Again, the mean of the means of sets of observations II and IV,, properly corrected for variation of primary current, were compared with the mean of the observations in Set III, and the result is to give the data for calculating the permeability of the liquid oxygen for a primary current through the primary coil of the transformer of about 37 amperes, corresponding very nearly to a mean magnetising force of 166 C.Gr.S. units. The sum or difference of these means of the throws, taken in the liquid oxygen and out of the liquid oxygen, depending on whether they are on the opposite or the same side of the zero of the scale, gives us the value of the quantity denoted by D in the Table I below, and in the formula for the value of /*. The above sets of observations, I, II, III, and IV, refer to a primary current of about 37 amperes ; but similar sets of observa- tions were taken with a primary current of about 8 amperes, 28 amperes, and 50 amperes respectively, and the result's of all these- observations, which are included in the sets of observations, I to XII, above given, have been reduced in Table I below to show the mag- netic permeability of the liquid oxygen corresponding to different megnetising currents. The set of observations marked Experiment Valid Experiment VI in the above table of results, gives the observa- tions for standardising the ballistic galvanometer. In the first case- the primary coil of the balancing induction coil was cut out, and a primary current, having a value of 0*1145 ampere, was reversed through the primary coil of the transformer alone, and gave ballistic- deflections as stated in the observations in Set V. These observations serve to standardise the galvanometer and interpret the meaning of the throw obtained when the large current is reversed through the primaries of the two induction coils, the secondaries of which are opposed. It will be noticed that one important advantage of the- above- described method is that the quantity which we desired to know, viz., the amount by which the presence of the liquid oxygen increases the magnetic permeability of the core of the transformer, ia the quantity which is measured directly, and that any error in the measurement of this quantity does not affect the permeability to anything like the same proportional extent. An error of about 10 per cent, in the measurement of the ballistic throw would only affect the fourth place of decimals in the number representing the perme- ability of the liquid oxygen. The results of all the above observations, when reduced, are com- prised in the following table : — VOL. LX, z 292 Profs. J. A. Fleming and J. Dewar. On the Table I.— Table of Results of Observations on the Magnetic Permeability of Liquid Oxygen. A - Total ballistic Ballistic throw primary current, in amperes, passing through primaries of the transformer and balanc- Correspond- ing mean magnetising force in C.G-.S. units in primary circuit of transformer. throw which would be produced if primary current of A amperes were reversed through primary of trans- former alone *±a. of galvanometer resulting from immersion of the transformer in liquid oxygen. Transformer and balancing induction coil being opposed V- = permeability calculated from '-'-5 a ing coil. = D. 8*037 35-2 1734 4-33 1 -00250 28-13 123-0 6068 14-9 1 -00246 37-8 165-4 8153 21-18 1 -00260 36-8 161-0 7938 23-57 1 -00297 50-5 220-9 10894 32-98 1 -00304 The values of the permeability given in the foregoing table are not all of equal weight. The calculated value of JJL — 1 depends upon the observed ballistic throw, and this cannot be read to a high degree of accuracy when the throw is as small as 4 millimetres. We consider that the best result is obtained by taking the mean of the values for the primary currents, 37'8, 36'8, and 50'5 amperes, and these values give /* = 1*00287, with a probable accuracy of + 0'0002. This value of the permeability of the liquid oxygen corresponds to a magnetising force lying between 166 and 220 C.Gr.S. units. It will be seen that this method is best applicable to the determination of the permeability under large magnetising forces ; and that these observations do not, in them- selves, allow us to state whether the permeability is a constant for all forces, or is a function of the value of the force. In the next place the value is a relative one. The number 1*00287 is the ratio of the magnetic permeability of liquid oxygen to that of the gaseous oxygen nearly at the same temperature resting upon the surface of the liquid. We were not able by this method to detect the difference between the permeability of the cold gaseous oxygen lying on the surface of the liquid oxygen when in quiet ebullition, and which has a temperature of about —182° C., but a density of at least three times that of oxygen at 0° C., when compared with that of gaseous oxygen at ordinary temperature, and under the normal pressure. In a very valuable memoir on the determination of magnetic suscepti- bilities, M. P. Curie* has examined the susceptibility of gaseous *.' Theses presentees a la Faculte des Sciences de Paris pour obtenir le grade de Docteur es Sciences Physiques,' par M. P. Curie, Paris, 1895.' This memoir is of remarkable interest in many ways. Magnetic Permeability of Liquid Oxygen and Liquid Air. 293 oxygen at different temperatures, and shown that between the limits of 0° C. and 452° C. tbe magnetic susceptibility of oxygen (K) per unit of mass is a function of the absolute temperature T, such that 106 K = 33700/T, and that the value of K (per gram) at 0° C. is, therefore, 123/106. The mass of 1 c.c. of oxygen gas at 0° C. and 760 mm. is O0014107 gram, and, reciprocally, the volume of one gram is 708'9c.c. at 0° C. and 760 mm. Hence the magnetic susceptibility of gaseous oxygen at 0° C. and 760 mm. per unit of volume (one c.c.) would be 123 x 0'00141 x 10~6 = 0'173 X 10~6, which is not very different from that obtained by other observers.* If then it could be supposed that gaseous oxygen followed the same law down to —182° C., and taking the gas in a condition when the density is nearly 0'00423, the volume susceptibility (&) at —182° C. would be 1*6 x 10~6, and hence the permeability (/*)> where should be 1-00002. It is, however, certain that the susceptibility per unit of mass will not continue to increase in accordance with the hyperbolic law, because this would imply that at the absolute zero of temperature the susceptibility would be infinitely great, and hence the above number 1 '00002 gives a superior limit for the permeability of the gaseous oxygen at — 182° C. lying on the surface of the liquid oxygen.t The conclusion is that the correction to be applied to the above observed value of ^ for the liquid oxygen, viz., T00287, to refer it to a vacuum taken as unity, is altogether masked by the unavoidable errors of experiment, and hence, pen ding further more exact measure- ments, this may be taken as the value of the constant. We have, however, at the present time, arranged a method which will enable us we hope to determine directly the magnetic susceptibility of liquid * .Faraday, ' Experimental Researches,' vol. 3, p. 502, gives a value for the sus- ceptibility of gaseous oxygen at 60° F., referred to an equal volume of water as unity, which, when reduced to absolute values by taking the magnetic susceptibility -of water as 0'79 x 10~6, gives the value of the susceptibility as 0'143 x 10~6. Becquerel found a value not very different. f The critical temperature of oxygen is —118° C. The corresponding absolute temperature is 155°. If we then put T = 155, in Curie's formula, 106K = 33700/T, we get 106K = 217'4, as his deduced extrapolated value for the sus- ceptibility per unit of mass. Since the density of liquid oxygen, as determined by one of us (J. Dewar) is l'137o, our value for the susceptibility per unit of mass of the liquid oxygen is 228/l'1375 = 2007. These figures show that the hyperbola does not represent the value of the susceptibility per unit of mass below -the critical temperature. 294 Profs. J. A. Fleming and J. Dewar. On the oxygen with, far greater accuracy. This method consists in observing the mechanical force which acts upon a vacuum bulb or mass of matter of known and very low susceptibility when it is suspended free from gravity in a vessel of liquid oxygen, and in a variable mag- netic field. Under these conditions a vacuum bulb of very thin glass would behave like a strongly diamagnetic body, and if the mag- netic susceptibility of the vacuum bulb or test mass is &15 and that of liquid oxygen is &3 for equal volumes, then the apparent diamagnetic susceptibility of the mass will be — (&2 — ki), and the actual para- magnetic susceptibility of liquid oxygen may be deduced fro in a knowledge of &i and — (&2 — &i)« By this method we hope to be able to determine whether the permeability of liquid oxygen is a function of the magnetising force. The latest experimental results and measurements made with solutions of iron salts, such as those made recently by Mr. J. S. Townsend,* appear to^show that the magnetic permeability of solutions of these iron salts is a constant quantity at least for a range of magnetic forces varying from 1 to 9 C.G.S. units. The value, viz. 1*00287, as determined by us for the magnetic permeability of liquid oxygen, shows that the magnetic susceptibility (&) per unit of volume is 228/106. It is interesting to compare this value with the value obtained by Mr. Town send for an aqueous solution of ferric chloride, and which he states can be calculated by the equation 10° & = 91-610— 0'77, where w is the weight of salt in grams per cubic centimetre, and k the magnetic susceptibility. Even in a saturated solution, w cannot exceed O6, hence, from the above equation, we find the value of the magnetic susceptibility of a saturated solution of one of the most para- magnetic iron salts, viz., ferric chloride, is 54/106 for magnetic forces between 1 and 9. This agrees fairly well with other determinations of the same constant. On the other hand, the magnetic suscepti- bility of liquid oxygen for the same volume is 228/106, or more than four times as great. The unique position of liquid oxygen in respect of its magnetic susceptibility is thus strikingly shown. It is, how- ever, interesting to note that its permeability lies far below that of certain solid iron alloys generally called non-magnetic. The 12 per cent, manganese steel of Mr. B. A. Hadfield is usually spoken of as non-magnetic, yet the magnetic permeability of this last substance has been shown to be 1'3 or 1*4. We have applied the foregoing method also to the determination of the magnetic permeability of liquid air. Since liquid air which * See 'Phil. Trans.,' A, rol. 187, 1896, "Magnetisation of Liquids," J. S. Townsend, M.A. Magnetic Permeability of Liquid Oxygen and Liquid Air. ^95 has been standing' in a vacuum vessel for any length of time has a composition which varies with the time and which may contain an much as 75 or 80 per cent, of oxygen, it was not to be expected that very closely consistent results could be obtained in the case of air. The following figures show, however, the observational results : — PERMEABILITY OF LIQUID AIR. Throws of Ballistic Galvanometer. Induction Coils balanced. Exp. I. Exp. II. Exp. III. Exp. IV. Later. Exp. Y. Exp. VI. Exp. VII. Exp. VIII. 1 *5 mm. to right 1-2 17 -0 mm. to left 17-5 , 0*3 mm. to right 0-3 , 17 '0 mm. to left 17'0 „ 17'3 , 2-8 mm. to left 2-8 „ 18 -8 mm. to left 19-2 „ 19-4 „ 19-8 „ 3 '5 mm. to left 3-4 , 22 '0 mm. to left 22-0 „ 22-0 , The transformer in liquid air. Current = 38 -0 amperes reversed through primary coils. The transformer lifted out of liquid air into cold gaseous air at the same temperature as before. 37 '5 amperes reversed. The transformer in liquid air. Current = 37 amperes reversed through primary coils. The transformer lifted out of liquid air into cold gaseous air, and at the same tempera- ture as before. Current = 37 amperes reversed through primary coils. The transformer in liquid air. Current = 367 amperes reversed through primary coils. The transformer lifted out of liquid air into cold gaseous air, at the same temperature as before. Current = 37 amperes reversed through primary coils. The transformer in liquid air. Current = 36*7 amperes reversed through primary coils. The transformer in liquid air. Primary circuit of balancing coil cut out of circuit and O'lllS ampere reversed through primary of transformer to standardise the ballistic galvanometer. The results of these observations, when reduced, show that corre- sponding to a primary current of 37'5 amperes, or a mean mag- netising orce of 164 C.G.S. units, the apparent magnetic permea- bility of liquid air in terms of gaseous air of the same temperature is 1-00240. 2 A VOL. LX. Anniversary Meeting. At the time of these observations the liquid air used had probably become almost entirely liquid oxygen by the evaporation of the nitrogen. The figure, however, serves to check approximately that of the liquid oxygen. In conclusion, we desire to express our thanks to Mr. J. E. Petavel for the assistance he has given to us in the above work. We hope shortly to be able to make a further contribution to this portion of the investigations on which we are engaged, on the electrical and magnetic constants of liquid oxygen, and which will include a deter- mination of the dielectric constant of liquid oxygen, made with the object of determining the extent to which this substance obeys Maxwell's law connecting magnetic permeability, dielectric constant, and optical refractivity. November 30, 1896. ANNIVERSARY MEETING. Sir JOSEPH LISTER, Bart., P.R.C.S., D.C.L., President, in the Chair. The Report of the Auditors of the Treasurer's Accounts, on the part of the Society, was presented as follows : — " The total receipts on the General Account during the past year, including balances carried from the preceding year, amount to £8,928 Is. 3d, and the total receipts on account of Trust Funds, including balances from the preceding year, amount to £5,009 Os. 2d. The total expenditure for the same period amounts to £7,287 12s. 3dL on the General Account (including £300 on loan to the Coral Boring Committee), and £3,347 11s. *ld. on account of Trust Funds, leaving a balance on the General Account of £1,605 9s. 4c£. at the bankers (which includes £1304 17s. 3d. on deposit — Dr. Ludwig Mond's gift, £54 10s. Publication Grant Account, and £29 11s. lOd. Water Research Account), and in the hands of the Treasurer a balance of £34 19s. Sd. ; leaving also at the bankers a balance on account of Trust Funds of £1,661 8s. 7d." The thanks of the Society were voted to the Treasurer and Auditors. Lists of Fellows deceased and elected. The Secretary then read the following Lists : — Fellows deceased since the last Anniversary (Nov. 30, 1895). On the Home List. 297 Chambers, Charles. Childers, Right Hon. Hugh Cul- ling Eardley, F.R.G.S. Erichsen, Sir John Eric, Bart., F.R.C.S. Green, Alexander Henry, M.A. Grove, Right Hon. Sir William Robert, D.C.L. Harley, George, M.D. Hind, John Russell, LL.D. Humphry, Sir George Murray, M.D. Johnson, Sir George, M.D. Martin, Henry Newell, M.A. Mueller, Baron Ferdinand von K.C.M.G. Prestwich, Sir Joseph, D.C.L. Reynolds, Sir John Russell, Bart., M.D. Richards, Sir George Henry, Admiral, K.C.B. Richardson, Sir Benjamin Ward, M.D. Sharp, William, M.D. Trimen, Henry, M.B. Verdon, Hon. Sir George Frederic, K.C.M.G. Walker, James Thomas, General, R.E., C.B. On the Foreign List. Daubree, Gabriel Auguste. Fizeau, Hippolyte Louis. Gould, Benjamin Ap thorp. Kekule, August. Newton, Hubert Anson. Withdrawn. Bateman, James, M.A. Fellows elected since the last Anniversary. Clarke, Lieut. -Colonel Sir George Sydenham, R.E. Collie, J. Norman, Ph.D. Downing, Arthur Matthew Weld, D.Sc. Elgar, Francis, LL.D. Gray, Prof. Andrew, M.A. Hinde, George Jennings, Ph.D. Miers, Prof. Henry Alexander, M.A. Mott, Frederick Walker, M.D. Murray, John, Ph.D. Pearson, Prof. Karl, M.A. Stebbing, Rev. Thomas Roscoe Rede, M.A. Stewart, Prof. Charles, M.R.C.S. Temple, Sir Richard, Bart., G.C.S.I. Wilson, William E. Woodward, Horace Bolingbroke, F.G.S. Wynne, William Palmer, D.Sc. 298 Anniversary Meeting. On the Foreign List. Grandly, Albert. Heim, Albert. Kohlrausch, Friedrich. Langley, Samuel Pierpont. Lie, Sophus. Metschnikoff, Elias. Mittag-Leffler, Gosta. Schiaparelli, Giovanni. Lippmann, Gabriel. The President then addressed the Society as follows :— Nineteen Fellows and five Foreign Members have been taken from the Royal Society by death since the last Anniversary Meeting. The deceased Fellows are — John Russell Hind, December 23, 1895, aged 73. The Right Hon. Hugh Culling Eardley Childers, Japuary 29, 1896, aged 69. General James Thomas Walker, February 16, 1896, aged 69. Charles Chambers, March, 1896, aged 61. William Sharp, April 10, 1896, aged 91. Sir John Russell Reynolds, May 29, 1896, aged 68. Sir George Johnson, June 3, 1896, aged 78. Sir Joseph Prestwich, June 23, 1896, aged 84. The Right Hon. Sir William Robert Grove, August 2, 1896, aged 85. Alexander Henry Green, August 19, 1896, aged 64. The Hon. Sir George Frederic Verdon, September 13, 1896, aged 62. Sir John Eric Erichsen, September 23, 1896, a^ed 78. Sir George Murray Humphry, September 24, 1896, aged 76. Baron Ferdinand von Mueller, October 9, 1896, aged 71. Henry Trimen, October 18, 1896, aged 53. George Harley, October 27, 1896, aged 67. Henry Newell Martin, October 28, 1896, aged 44. Admiral Sir George Henry Richards, November 14, 1896, aged 76. Sir Benjamin Ward Richardson, November 21, 1896, aged 68. The Foreign Members are — Gabriel Auguste Daubree, May 29, 1896, aged 82. August Kekule, July 13, 1896, aged 66. Hubert Anson Newton, August 12, 1896, aged 66. Hippolyte Louis Fizeau, September 18, 1896, aged 77. Benjamin Apthorp Gould, November 27, 1896, aged 72. Although biographical notices of nearly all will be found in the * Proceedings,' there are some to whose labours I may make brief reference to-day. President's Address. 299 Sir William Grove presented the rare spectacle of steady and dis- tinguished devotion to science in spite of the claims of an exacting profession. Grove was an eminent lawyer. Called to the bar in 1835, he was for some time kept from active work by ill health ; but he subsequently acquired a considerable practice, and becoming a Queen's Counsel in 1853, was for some years the leader of the South Wales Circuit. His practice was mainly in patent cases, and the reputation he obtained in that field led to his being appointed a member of the Royal Commission on the Patent Laws. His work as an advocate was, however, by no means confined to such matters ; he was one of the counsel — Serjeant Shee and Dr. Kenealy being the others — who defended the Rugeley poisoner, William Palmer, and he was engaged in many other causes celebres. The eminent position to which he had risen at the bar led to his appointment in November, 1871, as a Judge of the old Court of Common Pleas, a post which in 1875 was converted by the Judica- ture Act into that of a Judge of the High Court. This office he held until his retirement in 1887, when he became a member of the Privy Council. Throughout the greater part of his long and distinguished legal career, Grove's love of science impelled him to devote a large share of his energies to its pursuit. It is remarkable that his first paper, which was communicated to the British Association in 1839, and which also appeared in the ' Comptes Rendus,' and in Poggendorff's ' Annalen,' contained a description of the " Grove's cell," which was afterwards used in every physical laboratory in the world. This was succeeded by a long series of memoirs, chiefly on electrical sub- jects, among which one of the best known is that on the gas battery. In 1842 he delivered, at the London Institution, an address which was, in the following year, developed into the celebrated series of lectures : " On the Correlation of Physical Forces." In these he dis- cussed what we should now call the transformations of energy ; and, though Professor Tait, in his " Historical Sketch of the Science of Energy," * assigns precedence in calling " attention to the gener- ality of such transformations " to Mrs. Somerville, there can be no doubt that Grove was an independent and very advanced thinker on that subject. For many years Sir William Grove took a very prominent part in the affairs of the Royal Society, and was one of the most active pro- moters of the reform of its constitution, which took place in 1847. It is largely to his efforts that we owe our present system of electing only a specified number of Fellows in each year. He was also one of the founders of the " Philosophical Club." He was President of the British Association in 1866, and, in the * ' Thermodynamics/ p. 58. 300 Anniversary Meeting. course of his address, observed : " The Kew Observatory, the petted child of the British Association, may possibly become an important national establishment ; and, if so, while it will not, I trust, lose its character of a home of untrammelled physical research, it will have superadded some of the functions of the Meteorological Department of the Board of Trade, with a staff of skilful and experienced observers."^ Although the British Association long ago handed over the care of its " petted child" to a Committee appointed by the Royal Society, the Society and the Association have lately appointed a joint Committee to urge the Government to supply the funds for converting the Kew Observatory into a " national establishment " similar to the Reichsanstalt at Charlottenburg. We are thus striving to realise to-day the suggestion thrown out, thirty years ago, by Grove. In Sir Joseph Prestwich we have lost almost the last link that remained which connected geologists of the present day with the founders of the science in the first half of this century. To him we are indebted, not only for the first comprehensive classification of the tertiary beds of this country — to several of which he assigned the names by which they will henceforth be universally known — but, also, for their correlation with the strata of the Paris Basin. To him, also, is due the credit of having been the first to establish the authenticity of the remains of human workmanship found in the drift-deposits of the valley of the Somme, and of thus having laid secure foundations on which arguments as to the extreme antiquity' of man upon the earth may be based. In France his name was known and respected as much as in England, and it would be hard to say how much of the advance in geological knowledge during the last sixty years was not due to his unintermitted labours, which extended over the whole of that period. The earliest scientific investigation of Armand Hippolyte Louis Fizeau was 011 the use of bromine in photography, and was published in 1841. He will always be remembered as the first who carried out experiments designed to measure the velocity of light produced by a terrestrial source, and travelling through a comparatively small dis- tance near the surface of the earth. These observations, made in 1849, were very difficult ; but the value of the method employed is attested by the fact that a quarter of a century afterwards it was adopted by M. Cornu, and- that with the improved apparatus employed by him it gave results of the highest accuracy. A few years afterwards Fizeau performed another classical experi- ment by which he measured the change in the velocity of light pro- duced by the motion of the medium in which it travels. * ' Correlation and Continuity.' Fifth Edition, 1867, p. 278. President's Address. 301 He also devised an extremely delicate method (based on the inter- ference of light) of determining the coefficients of thermal expansion of small bodies, such as crystals. The instrument he designed has been carefully studied by the Bureau International des Poids et Mesures, with very satisfactory results. On account of these and other researches, M. Fizean has, for nearly half a century, occupied a conspicuous position among European physicists. He was awarded the Rumford Medal in 1866, and became a Foreign Member of the Royal Society in 1875. Our distinguished Foreign Member, Professor Hubert Anson Newton, Senior Professor of Mathematics at the Yale University, New Haven, died at his home in New Haven on the 12th of August last. He was born at Sherbourne, in the State of New York, in 1830 ; studied at Yale College, where he graduated in 1850, and was called to the Chair of Mathematics in the University at the early age of twenty-five. On the organisation of the Observatory of the University in 1882, Professor Newton was appointed Director ; and though he resigned this position in 1884, the whole policy and success of the Observatory ever since, and, indeed, its very existence, are in no small measure due to his warm interest and untiring efforts. Professor Newton's name will ever remain associated with his important researches on Meteor Astronomy, beginning as early as 1860, and with his inquiry into the possible capture of comets by Jupiter and other planets. His historical investigations, and discus- sions of the original accounts, showed that the phenomena of meteor showers are of a permanent character, and come within the range of Celestial Dynamics, and that predictions of returning meteoric displays are possible. Professor Newton was President of the American Association for the Advancement of Science in 1885, and was for many years an Associate Editor of the 4 American Journal of Science.' He was a man of noble character, held in universal esteem, and greatly beloved by all those to whom he was persqnally known. The death of August Kekale will be felt as a severe loss to chemical science all over the world. Not only did his great activity in original research enrich organic chemistry with many new and interesting compounds, bu^ his announcement of the tetradic valency of carbon, and, especially, his theoretical conception of the benzene ring, gave an impulse to the study of structural chemistry which has introduced order into the vast array of organic compounds, both of the alcoholic and aromatic types, and has not, even yet, expended itself. In recognition of his life-long work, the Council of the Royal Society awarded Professor Kekule the Copley Medal in 1885. Another Foreign Member who has passed away from us during b02 Anniversary Meeting. the year is the distinguished mineralogist and geologist, M. Daubree. After leaving the Ecole Poly technique in 1832, he was sent on a mission to investigate the modes of occurrence of tin-ore in Cornwall and on the Continent. His reports showed such ability that he was appointed Professor of Mineralogy and Geology at Strasburg, at the age of 25; afterwards (1861-2)' he became Professor of Geology at the Musee d'Histoire Naturelle at Paris, and at the same time Pro- fessor of Mineralogy at the Ecole des Mines ; in the same year he succeeded to the Chair at the Institut vacated by M. Cordier. From 1872 to 1884, when the rules of the Service made retirement by reason of age compulsory, he acted as Director of the Ecole des Mines. M. Daubree was the leader in France in experiments for the synthetic reproduction of minerals and rocks, and his laboratory furnace was the first to yield crystals of oxide of tin having the lustre, colour, and hardness of the mineral cassiterite; his memoir on the zeolites and other minerals, produced since Roman times through the action of the hot springs of Plombieres on the bricks arid concrete, has been of general interest both to mineralogists and geologists. Other important experiments led him to infer that circulating water, rather than heat or vapours, has been the essential agent in all phenomena of rock transformation. M. Daubree gave much attention to the description and classification of meteorites, and made numerous experiments relative to the reproduction of material having similar characters. Tiie Council was much occupied during the earlier part of the session with the consideration of the proposed " Standing Orders " relating to the conduct of the meetings, and to the Publications of the Society — a subject which has engaged the anxious attention of previous Councils. In framing these Standing Orders two principal objects were kept in view. Firstly, to increase the interest of the meetings by giving greater freedom in the conduct of them, and by enlarging the opportunities for discussion ; and secondly, to obtain a more secure, and, at the same time, more rapid judgment as to the value of communications made to the Society ; so that, while the high standard of the 'Philosophical Transactions' is retained, or even raised, greater rapidity in the publication of these and of the ' Proceedings ' may be attained. To secure these latter objects, the Council has called to its aid, in the form of Sectional Committees, a number of Fellows much greater than that of the Council itself, to whom will be entrusted the task of reviewing the communications to the Society, and of making to the Council such recommendations with respect to them as may seem desirable. It is further probable that by using the special knowledge of the several Sectional Committees in the detailed consideration of special questions, the Council will have more time at its disposal than it has at present President's Address. 303 to consider the matters of larger policy which are so frequently brought before it. It soon became evident that no satisfactory Standing Orders securing these advantages could be drawn up which would not be in some way or other inconsistent with the Statutes at present in opera- tion. It was accordingly resolved to modify the Statutes ; and this has been done by giving to certain Statutes a more general form tlian that in which they have for a long time appeared, so that such alterations of detail as may from time to time seem desirable may be effected by changes in the Standing Orders only, without inter- fering with the Statutes. I gladly avail myself of this opportunity of acknowledging the great help which the Council received from Mr. A. B. Kempe, in respect to the many legal points which arose in connection with the change of Statutes. A copy of the Statutes, as amended during the present session, as well as of the Standing Orders adopted, will be found in the Year-book, which has been instituted by one of the new Standing Orders, and which will be pub- lished each year, as soon after the Anniversary Meeting as possible. The International Conference called to consider the desirability and possibility of compiling and publishing, by international co- operation, a Complete Catalogue of Scientific Literature, was duly held ; and the Society may be congratulated on the successful issue of a meeting, to the preparations for which a special International Catalogue Committee, appointed by, and acting under the authority of, the Council, had devoted much time and labour. The Conference met in the apartments of the Society on July 14, 15, 16, and 17, under the presidency of the Bight Hon. Sir J. Gorst, Vice-President of the Committee of Council on Education, and was attended by forty- one delegates, representing nearly all countries interested in science. The Society was represented by the Senior Secretary, Professor Armstrong (Chairman of the International Catalogue Committee), Mr. Norman Lockyer, Dr. L. Mond, and Professor Riicker. Four other Fellows of the Society, General Strachey, Dr. D. Grill, Professor Liversidge, and Mr. R. Trimen were among the delegates appointed by the Indian and Colonial Governments. The Conference resolved that it was desirable to compile and publish a catalogue of the nature suggested in the original circular issued by the Royal Society, the administration being carried out by a Central International Bureau, under the direction of an Inter- national Council, with an arrangement that each of such countries as were willing to do so, should, by some national organisation, collect and prepare for the Central Bureau all the entries belonging to the scientific literature of the country. It was further resolved that the language of the catalogue should be English, and a proposal that the Central Bureau should be placed in London was carried by 304 Anniversary Meeting. acclamation. The Conference finding itself unable to accept any of the systems of classification proposed, requested the Royal Society to form a committee which shonld consider this and other matters which were left undecided by the Conference. The Council are already taking steps to perform the duties thus entrusted to them by the Conference. The delegates of the Society reported that the whole proceedings of the Conference were carried on with remarkable good feeling, and even unanimity, and that the confidence felt and expressed by the various delegates in the fitness of the Royal Society to complete the work begun by the Conference was most gratifying. In connection with the fact that the proposed International Cata- logue is to be in part arranged according to subject matter, it may be stated that the Council, acting upon a resolution of the International Catalogue Committee, have taken steps towards the practice of append, ing subject indices to the papers published by the Society, and have recommended th'e same practice to other Societies. The work connected with the Society's own Catalogue is progressing. Vol. XI, the last of the decade 1874-83, has been published, and the preparation of the Supplement, which has been found necessary for this and preceding decades, is being pushed on. For the Subject Index to the Catalogue, slips have been prepared, and the Catalogue Committee will soon have to advise the Council as to the system of classification to be adopted. . The Grant of £1000 in aid of publications, which My Lords of the Treasury promised last summer to place upon the Estimates of this year, has been sanctioned by Parliament, and a moiety of it has already been paid to the Society. The Council have already felt the great advantage of having this money at their disposal, and have framed Regulations for its administration which they trust will be found to work satisfactorily. The Council have made some small changes (which have been approved by My Lords of the Treasury) in the Regulations for the administration of the Government Grant of £4000 in aid of Scientific Inquiries, directed chiefly towards more effectually securing that Grants made should be expended for the purpose for which they were given, and that objects of permanent interest obtained by Grants should be properly disposed of. The only two Grants made this year which call for special mention are that of £1000 to the Joint Permanent Eclipse Committee of the Royal and Royal Astronomical Societies, for observations of the Solar Eclipse of August, and that of £800 for boring a coral reef in the Pacific Ocean, administered by the Committee appointed by the Royal Society, both drawn from the Reserve Fund. The Expedition to bore the Coral Reef received valuable assistance President's Address. 305 from My Lords of the Admiralty, who directed H.M.S. " Penguin " to carry the observers from Sydney, N.S.W., to Funafuti, the seat of the boring, and to render the Expedition all possible help during the whole of the operations. T desire to express on behalf of the Society our recognition of this renewed token of the willingness of My Lords of the Admiralty to further scientific inquiry. Though the full Report of the Expedition has not yet reached the Council, informa- tion has been received fco the effect that the boring operations had to be suspended when a depth of only 75 feet had been reached ; a layer of sand and boulders presenting obstacles which the experts employed were unable to overcome. It js much to be regretted that an undertaking, which promised scientific results of very great value has thus so far failed. The appeals of the Council to H.M. Minister for Foreign Affairs and to My Lords of the Admiralty for assistance to the Eclipse Expeditions met with most cordial and effective response, for which -we would express our gratitude. We also desire to acknowledge the courtesy shown and help afforded to the observing parties in Norway and Japan by the respective Governments of those countries, and to record our high appreciation of the enthusiastic and effective aid given to those under the direction of Mr. Norman Lockyer, at Vadso, by Captain King Hall and the Officers and crew of H.M.S. " Volage " ; to Dr. Common, also in Norway, by Commodore Atkin- son, of H.M.S. " Active " ; to the Astronomer Royal's party, in Japan, by the Officers of H.M.S. " Humber," " Pique," and " Linnet," kindly detached by Admiral Sir A. Buller to convey the various members of the expedition to and from Yezo, and to aid them during the observa- tions. Both in Norway and in Japan unfavourable weather rendered to a large extent nugatory the elaborate preparations which had been made for observing the eclipse. But British astronomy was splendidly saved from failure on this important occasion by the munificence and public spirit of Sir George Baden Powell, who fitted up, at his own expense, and accompanied an expedition in his yacht " Otaria " to Novaya Zemlya. The instruments employed were pro- vided by our Fellows, Mr. Lockyer and Mr. Stone, of the Radcliffe Observatory, Oxford; and the observations were entrusted to Mr. Shackleton, one of the computers employed by the Solar Physics Committee. In brilliant weather photographic observations were made, which promise to yield novel results of a highly important character. At the request of the President of the Board of Trade the Council nominated, in March, Professors Kennedy and Roberts- Austen as two members of a Committee to investigate the loss of strength in steel rails. So far as I am aware, the Committee has not yet made 306 Anniversary Meeting. its report. More recently, in July, the Council, at the request of H.M. Secretary for Colonial Affairs, appointed a Committee to con- sider, and if necessary to investigate, in conjunction with Surgeon- Major Bruce, who has made important researches in the matter, the disease caused in cattle in Africa by the Tsetse Fly. The Committee is still engaged on the inquiry. We believe that the Council, in cordially responding to requests like the above, and in freely placing at the disposal of H.M. Govern- ment its scientific knowledge and its acquaintance with scientific men, is performing one of its most important functions. The Council of the Royal Society is again and again called upon to approach H.M. Government on behalf of the interests of science, and when it does so always meets with a cordial reception and a respect- ful hearing, even on occasions when public necessities prevent a favourable reply being given to its requests. In return, the Council believes it to be its duty (when called upon to do so), not only to place its own time and labour ungrudgingly at the service of H.M. Government, but also to ask for the co-operation of other Fellows of the Society, or even other scientific men not Fellows of the Society, feeling confident that whenever the matter in hand has practical bearings beyond the simple advancement of Natural Knowledge, the value of a scientific man's time and energy will be duly considered. Some correspondence has taken place with the War Office relative to resuming the borings in the Delta of the Nile, which were carried on for a time some years ago, and which, though not completed, yielded valuable results. The Expedition to the Soudan has, how- ever, prevented anything being done. The Council learn with pleasure that the old borings, undertaken for a purely scientic object, have indirectly been a valuable means of supplying certain districts of the Delta with sweet water. If anything had been needed to justify the meetings for discussion recently established, it would have been supplied by the brilliant success of that held during the present session on Colour Photo- graphy. On that occasion, M. Lippmann gave us a demonstration of results of unprecedented beauty, obtained by extremely simple means, though based on profound mathematical reasoning. Such meetings can only prove fruitful when they are held in consequence of some theme needing such a discussion as is afforded by a special meeting ; and their occurrence must therefore be uncertain and irregular. The purpose for which they were instituted would be frustrated if they were held at times fixed in any formal way, irre- spective of whether they were needed or no. Three of the informal gatherings recently instituted, limited to Fellows of the Society, have been .held during the session, and were judged to be very successful. President's Address. 307 The Council has had occasion during the past session fco present an address of condolence to Her Majesty, the Patron of the Society, on the lamented death of Prince Henry of Battenberg, and to the Royal Academy on the occasion of the death of their President, Lord Leighton. In the absence of Council, during the recess, I sent another message of sympathy 011 the death of Sir J. Millais. I had the privilege of presenting on behalf of the Council, an address of congratulation to our late President, Lord Kelvin, on the occasion of his Jubilee, nobly celebrated in Glasgow last summer, by a very remarkable concourse of scientific men from all parts of the world, assembled to do him honour. Addresses were also sent to our Foreign Member, Professor Can- nizzaro, on the celebration of his seventieth birthday, and to the University of Princeton, New Jersey, U.S.A., on the occasion of its Sesquicentenary Anniversary. Under the guidance of the Scientific Relief Committee, the Council has during the year granted £100 to assist scientific persons or their relatives in distress. The Council desires to call the attention of the Fellows to the fact that, during the year, as during past years, the income of the fund has exceeded its expenditure, and that more aid could be given -than has been given. With the view of increasing the usefulness of the fund, the Council has added to the list of those who can make representations to the Council concerning relief the Presidents of the Mathematical, Physical, and Entomological Societies. I cannot but give expression to my deep regret, shared, I am sure, by every Fellow, that Lord Rayleigh, whose tenure of office as Secretary has been marked as much by faithful devotion to the in- terests of the Society as by scientific brilliancy, has thought it right, in consequence of increasing pressure of other engagements, to retire. But I rejoice that the Council can submit to your suffrages a man well qualified to wear the mantle laid down by Lord Rayleigh. The Fellows will be pleased to learn that Mr. Rix, who was com- pelled by the condition of his health a year ago, to resign the position which he had held for many years with such great advantage to the Society, has much improved under the lighter labour of the Clerkship to the Government Grant Committee. As his successor in the office of Assistant- Secretary, the Council, out of eighty-four candidates, unanimously selected Mr. Robert Harrison, who entered upon his duties on the 24th of April last. The scientific work of the Society during the past year has been full of deep and varied interest. Early in the session the announce- ment of Rontgen's great discovery burst upon the world. Its won- derful applications to medicine and surgery attracted universal attention to it ; and physicists everywhere have since been engaged 308 A nniversary Meeting. in investigating the nature of the new rays. Perhaps no outcome of such inquiries has been more remarkable than the fact observed by our Fellow Professor J. J. Thomson, that the rays have the power of discharging electricity, both positive and negative, from a. body surrounded by a non-conductor ; a mass of paraffin wax, for example, behaving in their path for the time being like a conductor of elec- tricity. It appears that Lenard had before observed the discharge of both kinds of electricity through air by the rays with which he worked. Lenard's rays, however, differ from Rontgen's in being deflectable by a magnet, implying, in the opinion of most British physicists, that they are emanations of highly electrified particles of ponderable matter, while Rontgen's are regarded as vibrations in the ether. The question naturally arises whether Lenard, in the observations referred to, may not have been working with a mixture of Rontgen's rays and his own. While points like these are still under discussion by experts, we cannot but feel that the letter X, the symbol of an unknown quantity, employed originally by Rontgen to designate his rays, is still not inappropriate. I have before referred to Lippmann's beautiful demonstration and discussion of colour photography in one of our meetings. Very important researches have been made both by Lord Rayleigh and by Professor Ramsay into the physical properties of the new substance, helium, discovered by Ramsay in the previous session. Among their most striking results is the fact ascertained by Rayleigh that the refractivity of helium is very much less than any previously known, being only O146 ; between three and four times less than that of hydrogen, the lowest that had before been observed, although helium has more than twice the density of hydrogen. And equally surprising is Ramsay's observation of the extraordinary distance through which electric sparks will strike through helium, viz., 250 or 300 mm. at atmospheric pressure, as compared with 23 mm. for oxygen and 39 for hydrogen. Such properties appear to indicate that in helium we have to do- with an exceedingly remarkable substance. The density of helium appears to be really slightly different according to the mineral source from which it is obtained ; and this circumstance seemis to give countenance to the opinion arrived at by Lockyer and also by Runge and Paschen, from spectroscopic investi- gation, that helium is not a perfectly pure gas. But whatever other gas or gases may be mixed with it, they must be as inert chemically as the main constituent ; for all Ramsay's elaborate attempts to induce it, or any part of it, to combine with other bodies have entirely failed. Professor Roberts- Austen, in the Bakerian lecture, brought before President's Address. o09 us astonishing evidence that metals are capable of diffusing into each other, not only when one of them is in the state of fusion, but when both are solid. We learned that if clean surfaces of lead and gold are held together in vacuo at a temperature of only 40° for four days, they will unite firmly and can only be separated by a force equal to one-third of the breaking strain of lead itself. And gold placed at the bottom of a cylinder of lead 70 mm. long thus united with it, will have diffused to the top in notable quantities at the end of three days. Such facts tend to modify our views concerning the mutual relations of the liquid and solid states of matter. Such are a few samples of the many highly interesting communica- tions we have had in physics and chemistry. On the biological side also, there has been no lack of important work. Of this I may refer to one or two instances. Professor Schafer has given us an account of the well devised experiments by which he has conclusively established that the spleen is on the one hand capable, like the heart, of independent rhythmical contractions, and, on the other hand, has those contractions controlled by the central nervous system acting through an extraordinary number of efferent channels. Professor Farmer and Mr. Lloyd- Williams made a very beautiful contribution to biology in the account they gave of their elaborate investigations on the fertilisation and segmentation of the spore in Fucus. Especial interest attached to this communication, from the fact that it described in a vegetable form exactly what had been established by Oscar Hertwig in Echinodermata, viz., that out of the multitude of fertilising elements that surround the female cell, one only enters it and becomes blended with its nucleus. Lastly, I may mention the very remarkable investigation into the development of the Common Eel, which was described to us a fortnight ago by Professor Grassi, to which I shall have occasion to refer in some detail when speaking of his claims to one of the Society's medals. These, as I have before said, are but samples of what we have had before us ; but I think they are in themselves sufficient to justify the statement that, in point of scientific interest, the past year has been in no degree inferior to its predecessors. COPLEY MEDAL. Professor Gad Gegenbaur, For. Mem. U.S. The Copley Medal for 1896 is given to Carl Gegenbaur, Professor of Anatomy in Heidelberg, in recognition of his pre-eminence in the science of Comparative Anatomy or Animal Morphology. Professor 310 Anniversary Meeting. Gegenbaur was born in 1826, and a few weeks ago his 70th birthday was celebrated by his pupils (who comprise almost all the leading comparative anatomists of Germany, Holland, and Scandinavia) by the presentation to him of a " Festschrift " in three volumes. Gegen- baur is everywhere recognised as the anatomist who has laid the foundations of modern comparative anatomy on the lines of the theory of descent, and has to a very large extent raised the building by his own work. His ' Grundziige der vergleichenden Anatomie ' was first published in 1859, when he was 33 years old. In the second edition, published in 1870, he remodelled the whole work, making the theory of descent the guiding principle of his treatment of the subject. Since then he has produced a somewhat condensed edition of the same work under the title of ' Grundriss ' (translated into English and French), and now, in his 71st year, he is about to publish what will probably be the last edition of this masterly treatise, revising the whole mass of facts and speculations accumu- lated through his own unceasing industry and the researches of his numerous pupils during the past quarter of a century. Gegenbaur may be considered as occupying a position in morph- ology parallel to that occupied by Ludwig in Physiology. Both were pupils of Jahannes Miiller, and have provided Europe with a body of teachers and investigators, carrying forward in a third generation the methods and aims of the great Berlin professor. Gegenbaur's first independent contribution to science was published in 1853. It was the outcome of a sojourn at Messina in 1852, in company with two other pupils of Johannes Miiller, namely Albert Kolliker (still professor in Wiirzburg) and Heinrich Miiller, who died not long afterwards. These young morphologists published the results of their researches in common. Gegenbaur wrote on Medusae, on the development of Echinoderms, and on Pteropod larvae. A long list of papers on the structure and development of Hydrozoa, Mollusca, and various invertebrata followed this first publication. The greatest interest, however, was excited among anatomists by his researches on the vertebrate skeleton (commenced already in 1849 with a research, in common with Friedreich, on the skull of axolotl). In a series of beautifully illustrated memoirs he dealt with and added immensely to our knowledge of the vertebral column, the skull, and the limb- girdles and limbs of Vertebrata, basing his theoretical views as to the gradual evolution of these structures in the ascending series of vertebrate forms upon the study of the cartilaginous skeleton of Elasmobranch fishes, and on the embryological characters of the cartilaginous skeleton and its gradual replacement by bone in higher forms. His method and point of view were essentially similar to those of Huxley, who independently and contemporaneously was engaged on the same line of work. President's Address. 311 For many years Gegenbaur was professor in Jena, where he was the close friend and associate of Ernst Haeckel, but in 1875 he accepted the invitation to the Chair of Anatomy in Heidelberg, and in view of the increased importance of his duties as a teacher of medical students, and therefore of human anatomy, though still con- tinuing his researches on vertebrate morphology, he produced a large treatise on that subject, which has ran through two editions. In this work he made the first attempt to bring, as far as possible, the nomenclature and treatment of human anatomy into thorough agreement with that of comparative anatomy, and to a very large extent the changes introduced by him have influenced the teaching of human anatomy throughout Europe and America. There is probably no comparative anatomist or embryologist in any responsible position at the present day who would not agree in assigning to Gegenbaur the very first place in his science as the greatest master and teacher who is still living amongst us. He is not only watching in his old age the developments of his own early teachings and the successful labours of his very numerous disciples, but is still exhibiting his own extraordinary industry in research, his keenness of intellectual vision, and his unrivalled knowledge and critical judgment. ROYAL MEDAL. Sir Archibald GeiJcie, F.R.S. One of the Royal Medals is conferred on Sir Archibald Geikie, on the ground that of all British geologists he is the most distinguished, not only as regards the number and the importance of the geological papers which he has published as an original investigator, but as one whose educational works on geology have had a most material influence upon the advancement of scientific knowledge. His original papers range over many of the main branches of geological science. His memoir upon the ' Glacial Drift of Scotland ' (1863) is one of the classics in British geology. His work on the ' Scenery of Scotland, viewed in connection with the Physical Geology ' (1865) was the first successful attempt made to explain the scenery of that country upon scientific principles, and is still without a rival. His papers on the " Old Red Sandstone of Western Europe " (1878-79) gave for the first time a clear and convincing picture of the great lake period of British geology, founded upon personal observation in the field. His many original contributions to the Volcanic History of the British Isles form a succession of connected papers, crowded with important observations and discoveries, and brilliant and fertile generalizations respecting the abundant relics of former volcanic VOL. LX. - 1J 312 Anniversary fleeting. activity in the British Isles from the earliest geological ages to Middle Tertiary times. In the first series of these papers — commencing* with the " Chrono- logy of the Trap Rocks of Scotland" (1861), and ending with the " Tertiary Volcanic Rocks of the British Isles" (1869), abundant original proofs were advanced of the activity of volcanic action in the Western Isles of Scotland, and of its long duration in geological time. The second series (1871-88) was especially distinguished by the publication of his remarkable paper on the " Carboniferous Vol- canic Rocks in the Basin of the Firth of Forth," our earliest, and, as yet, oar only monograph on a British volcanic area belonging to a pre-Tertiary geological system. The third series (begun in 1888) commenced with his memoir on the " History of Volcanic Action during the Tertiary Period in the British Isles," a paper which is by far the most detailed and masterly contribution yet made to the subject, and for which the Brisbane Medal was awarded him by the Royal Society of Edinburgh ; and this succession of papers has been followed by the publication of others of almost equal importance. Sir Archibald Geikie has also written many papers and memoirs bearing upon geological processes arid their effects, which have become permanent parts of oar scientific literature. While carrying out this highly important original work in Geology, Sir Archibald has most materially contributed to the advancement and diffusion of scientific knowledge by his many educational works upon Geology and Physical Geography. His ' Elementary Lessons on Physical Geography' has passed through several English and Foreign editions ; his ' Outlines of Field Geology ' is now in its fifth edition ; and his article on Geology — originally contributed to the ' Encyclopaedia Britannica ' in 1879 — was afterwards expanded by him into his well-known ' Text- book of Geology,' which has become the acknowledged British standard of Geology in general. ROYAL MEDAL. Professor C. V. Boys. The other Royal Medal is awarded to Professor Boys, who has given to physical research a method of measuring minute forces far exceeding in exactness any hitherto used, by his invention of the mode of drawing quartz fibres, and by his discovery of their remark- able property of perfect elastic recovery. Professor Boys has himself made several very important researches in which he has employed these fibres to measure small forces. Using a combination of a thermo- junction with a suspended coil in a galvano- meter of the usual D'Arsonval type, a combination first devised by D'Arsonval himself, Professor Boys developed the idea in the micro- Presidents A ddress, 313 radiometer, an instrument rivalling the bolometer in the measurement of small amounts of radiation. Its sensitiveness and accuracy were ob- tained in part by the use of a quartz fibre to suspend the coil, in part by the admirable design of every portion of the instrument. Professor Boys was the first to show its value in an investigation into the radiation received from the moon and stars. In his great research on the value of the Newtonian constant of attraction, Professor Bovs used quartz fibres to measure the gravitation forces between small bodies by the Mich ell- Cavendish torsion method. He redesigned the whole of the apparatus, and, calculating what should be the dimensions and arrangements to give the best results, he was led to the remarkable conclusion that accuracy was to be gained by a very great reduction in the size of the apparatus. This conclusion he justified by a determination of the value of the New- tonian constant, which is now accepted as the standard. Professor Boys has also made some remarkable studies by a photo- graphic method of the motion of projectiles, and of the air through which they pass. All his work is characterised by the admirable adjustment of. the different parts of the apparatus he uses to give the best results. His instruments, are, indeed, models of beauty of design. RUMFORD MEDAL. Professor Philip P. Lenard and Professor W. C. Rontgen. In tlr^ case of the Rumford Medal, the Council have adopted a course, for which there are precedents in the awards of the Davy Medal, but which is, as far as the Rumford Medal itself is concerned, a new departure. They have decided to award the Medal in dupli- cate. It has often happened in the history of science that the same discovery has been made almost simultaneously and quite indepen- dently by two observers, but the joint recipients of the Rumford Medal do not stand in this relation to each other. Each of them may fairly claim that his work has special merits and characteristics of its own. To day, however, we have to deal, not with points of difference, but with points of similarity. There can be no question that a great addition has recently been made to our knowledge of the phenomena which occur outside a highly exhausted tube through which an electrical discharge is passing. Many physicists have studied the luminous and other effects which take place within the tube ; but the extension of the field of inquiry to the external space around it is novel and most important. There can be no doubt that this extension is chiefly due to two men — Pro- fessor Lenard and Professor Rontgen. 2 B 2 314 Anniversary Meeting. The discussion which took place at the recent meeting of the British Association at Liverpool proved that experts still differ as to the exact meaning and causes of the facts these gentlemen have dis- covered. No one, I believe, disputes the theoretical interest which attaches to the researches of both ; or the practical benefits which the Rontgen rays may confer upon mankind as aids to medical and surgical diagnosis. But whatever the final verdict upon such points may be, the two investigators whom we honour to-day have been toilers in a common field, they have both reaped a rich harvest, and it is, therefore, fitting that the Royal Society should bestow upon both of them the Medal which testifies to its appreciation of their work. DAVY MEDAL. Professor Henri Moissan. The Davy medal is given to Professor Henri Moissan. Notwithstanding the abundant occurrence of fluorine in nature, the chemical history of this element and its compounds has until recently been scanty in the extreme, and, as far as the element in the free state is concerned, an entire blank. And yet from its peculiar posi- tion in the system of elements, the acquisition of a more extended knowledge of its chemical properties has always been a desideratum of the greatest scientific interest. The frequent attempts which have been made from time to time to clear up its chemical history have been constantly baffled by the extraordinary difficulties with which the investigation of this element is beset. Thanks to the arduous and continuous labours of M. Moissan, this void has been filled up. He has effected the isolation of fluorine in a state of purity, and prepared new and important compounds, the study of which has placed our knowledge of the chemical and physical properties of this element on a level with that of its imme- diate allies. During the last few years M. Moissan has turned his attention to the study of chemical energy at extremely high temperatures, and by the aid of the electric furnace, which he has contrived, he has succeeded in obtaining a large number of substances whose very existence was hitherto undreamt of. It is impossible to set bounds to the new field of research which has thus been opened out. The electric furnace of M. Moissan has now become the most powerful synthetical and analytical engine in the laboratory of the chemist. On studying the accounts which Moissan has given of his re- searches, we cannot fail to be struck with the originality, care, perse- verance and fertility of resource with which they have been carried President's Address. 315 on. The Davy Medal is awarded to him in recognition of his great merits and achievements as an investigator. DARWIN MEDAL. Professor Giovanni Battista Grassi. The Darwin Medal for 1896 is awarded to Professor Grassi, of Rome (late of Catania), for his researches on the constitution of the colonies of the Termites, or White Ants, and for his discoveries in regard to the normal development of the Congers, Muraense, and Common Eels from Leptocephalus larvas. From a detailed examination of the nature and origin of the colo- nies of the two species of Termites which occur in the neighbour- hood of Catania, viz., Termes lucifugus and Callotermes flavicollis, he was able to determine certain important facts which have a funda- mental value in the explanation of the origin of these and similar polymorphic colonies of insects, and are of first-rate significance in the consideration of the question of the share which heredity plays in the development of the remarkable instincts of " neuters," or arrested males and females, in these colonies. Professor Grassi has, in facb, shown that the food which is administered by the members of a colony to the young larvae determines, at more than one stage of their development, their transformation into kings or queens, or soldiers or workers as the case may be, and the value of these researches is increased by the observations which he has made on the instincts of the different forms, showing that they do not in early life differ from one another in this respect, and are all equally endowed with the potentiality of the same instincts. These do not, however, all become developed and cultivated in all alike, but become specialised, as does the physical structure in the full-grown forms. A very different piece of work, but having a no less important bearing on the theory of organic evolution, is that on the Lepto- cephali. These strange, colourless, transparent, thin-bodied creatures, with blood destitute of red corpuscles, had been regarded as a special family of fishes, but have been proved by Grassi's patient and long- continued labours to be larval forms of the various Mureenoids. The most astonishing case is that of the Common Eel, Anguilla vulgaris, the development of which had been a mystery since the days of Aristotle. It had been long known that large eels pass from rivers into the sea at certain seasons, and that diminutive young eels, called in this country Elvers, ascend the rivers in enormous numbers. But, although the species is very widely distributed, no one in any country had been able to discover how the elvers were produced. Grassi has shown that, large as the eels are that pass into the sea, they are not perfectly developed fish, but only attain maturity in the depths of the 31fi Anniversary Meeting. ocean. There they in due time breed, and from their eggs are hatched the young Leptocephali, which, after attaining a certain size, cease to feed, and assume the very different form of the elver. The possibility of establishing these remarkable facts depended on the powerful oceanic currents that prevail about the Straits of Messina, bringing up occasionally to the surface the inhabitants of the depths of the sea. Grassi was thus able to obtain from, time to time both adult eels with fully developed sexual organs and their larval progeny, and he actually observed in an aquarium the develop- ment of a Leptocephalus brevirostris into an elver. Such highly meritorious contributions to evolution are fitly recog- nised by the award ot the Darwin Medal. The Statutes relating to the election of Council and Officers were then read, and Professor Liversidge and Dr. Common having been, with the consent of the Society, nominated Scrutators, the votes of the Fellows present were taken, and the following were declared duly elected as Council and Officers for the ensuing year : — President.— Sir Joseph Lister, Bart., F.R.C.S., D.C.L. Treasurer.— Sir John Evans, K.C.B., D.C.L., LL.D. Secretaries — { Professor Michael Foster, M.A., M.D., D.C.L., LL.D, 1 Professor Arthur William Riicker, M.A., D.Sc. Foreign Secretary. — Edward Frankland, D.C.L., LL.D. Other Members of the Council. Prof. William Grylls Adams, M.A. ; Professor Thomas Clifford Allbutt, M.D.; Professor Robert Bellamy Clifton, M.A.; William Turner Thiselton Dyer, C.M.G.; Prof. James Alfred Ewing, M.A. ; Lazarus Fletcher, M.A. ; Walter Holbrook Gaskell, M.D. ; Prof. Alfred George Greenhill, M.A. ; William Huggins, D.C.L. ; Prof. Charles Lapworth, LL.D. ; Major Percy Alexander MacMahon, R.A. ; Prof. Raphael Meldola, F.C.S. ; Prof. William Ramsay, Ph.D.; The Lord Walsingham, M.A. ; Prof. Walter Frank Raphael Weldon, M.A. ; Adml. William James Lloyd Wharton, C.B. The thanks of the Society were given to the Scrutators. Financial Statement. 317 318 Financial Statement. CO CS CO CO rH C5 O 00 rH t> O5 CO CO t> O O CO C5 00 00 rH CO rH 10 N rH 00 (N O O CO iH CO rH CO ^ : rQ : fl J— ^ |y ^ ^ ? ft Financial Statement. 319 O I s" 2 ;j3 * o. O i- 00 ?! s I •^ •£ . 09 r? § a ^ a g >: ca B^ a -a 4 SS-illJlSl!* SS s3« § s s M § a •s § - fn ^3 c3 «S ^ ^ • •*• . «srvesa*ia?;'»'qi!;3l«d ^ 3 -T cs b^, -SaS'^>^!r^Q5 aoPH^-s -3 -I Q J S <=« 9$ r -s -3 , S i a fl ^--. S 'o S 5- -o & - O rn|W o a o ^ - 1 i - £ I I «. 8 re Ch 13 d., 2f per Ce a ^3 a - s ° -3 a ^ OQW '0. T3 T3 » O 00 320 Financial Statement. Trust Funds. to OS 00 3 S a, 111 § ,® c^ 10 II * A, PQ • £9109 <3 co fe •« -* 6 1 -a ^ * I 1 I M C m T; £ •S o t" O OnS S JH eJOOO5W5(MO JJJI 'S *• 8 ** 'oo 9 •*3 322 Trust Funds. "to "* "* ll "e •" oc a 05 I] ^Ja > IO 71 «s <- ^ I ** r" 0 C i^ 05 X O5 r™ "* I 0 V g 1 ^ 1 w £ £ 1 i*-^> S ^ ^E « a -««1 c 2*« ? J'? * •^ 4; W ^ - ^ ll 1 2 1 K <2 •» ~ 0 rAi P. — — CO 0 I i * o Hi P;M ,o r. ^ Trust Funds. 323 •*'« a ij 00 00 -» 8 -* || ^ coco cs -^0005 as | to O -> rH * "-0 S 2 *>' ° ° S 2 1 c< 9 8 ® II ^ w t* as ^1 rH ^fi C£ ^ ,8 6 £ 4 G Ct ^ w .S ^ "> ^ r- r« * § Cq 0 -7 o' t3 ° ^ *" • S "s o ua ' "g1^ 1 ^ 1 **• 1 I] i is ! 11 : 1 Ss 05 ;§ ^ '' 3^0).. -3 ' 1-^^ 1^ 5S 1 ^« r^ |^ Q=S*o Ps"SS -2 ,!i 111 ti rl -il 11 1 ' * • co HMW sUWPR ^§ 02pH pq S X "" -^ ^ ^ ^ S-t t*^ ^ '"^ § S -IM § — p, ^ ~* O .pq : ** ••c ^-^ = "§ ^ i 98 J Sp? ^ g- § ^ ^^g^'^^^J ^ |i .|r^^(N^cooi| i^p5 ^n'S 3-^d ?;_2 III III 11 || ^o ^ - .2 s 5 o > ~ H H. " « ^ 326 Trust Funds. ** 3 211 ^c 5 P "c'^ 3 ! «0 i • ^ P r? 1 .j C 0 cc'i !>• «o „ r* 1 r- i r-l c> <*l* r || «i B | ^2 = l| i db 3 *3 1 M 1 0 £ OQ : OQ «• 1 i i! ! - 1 I Z^ f— § cr "•^ X "S g |6 1 « O 4« £ % f>i CO ^ •I « W JS J - hQ •is o I& 1 h «*'<*« O J> ; O § rjj ^ C o o il 1 1 •*•« 0 o II J CO ^' g^j C o ^ ^ ^y- Tf 0 || ^ t> «o 11 «rt | & : 1 s Tjl «~T | a ^H r s J 11 I O (f. 0 £ . i: 1 pq 11 1 P •v 0 £ ; *-° £ 5 Income and Expenditure Account. 00 (N i— i O 895 'ovembe Cl Ci C O 22 3S3£8 C5 W •* ^M •* r-i CO CO O O-i -i O O CO Ci O r-l 1~^ rJ -^ O -f O 10 !M O (M CO Salaries Publica li alogue of Scien Index to ditto : S ill. to 5 ElisiUiL J ^ Z-s ® -S 2O «3 fl^^^ *lWllli*?^ 5 Ir.lllll CO I •s OOOOOOOOOCi VOL. LX. 2 c 328 Account of Grants from the Donation Fund. Professors Albert Heim, Gabriel Lippmann, G. Mittag-Leffler, and G. Schiaparelli were, at the meeting on the 26th of November, balloted for and elected Foreign Members of the Society. The following Table shows the progress and present state of the Society with respect to the number of Fellows : — Nov. 30, 1895 . . Since Elected Since Compounded Since Deceased . . Withdrawn Nov. 3P, 1896 Patron \ and Foreign. Eoyal. 4! + 9 45 Corn- pounders. 141 4- 2 4- 1 - 6 138 £4 yearly. £3 yearly. 200 + 13 - 1 - 3 209 Total. 498 -f 25 — 24 - 1 498 Account of Grants from the Donation Fund in 1895-96. £ s. d. Dr. Gamgee, in aid of his Researches 011 the Behaviour of Haemoglobin, &c., toward Ultra-violet Rays 50 0 0 Coral Reef Committee, towards the Purchase of Dia- monds for Boring a Coral Atoll in the Pacific Ocean .... 150 0 0 Dr. M. Foster, for Dr. W. Poole, Medical Officer of the British Central African Protectorate, for the Purchase of a Microscope to aid him in his Researches 21 211 Sir A. Geikie, in aid of Mr. Reid's Geological Borings at Hoxne 30 0 0 Sir A. Geikie, to assist him in Excavations at Hitchin 50 0 0 Profs. Fleming and Dewar, in aid of their Researches on the Diamagnetic qualities of Metals at Low Temperatures 50 0 0 Prof. Burdon Sanderson, in aid of his Investigations in relation to Tuberculin 60 0 0 Dr. Yaughan Harley, in further aid of his Researches on Absorption from the Alimentary Canal 25 0 0 Dr. J. G. Stoney, for Calculations of the Positions of the November Meteors 15 0 0 Professor Sherrington, to aid him in his Researches on the Nervous System 50 0 0 Marine Biological Association, towards the Purchase of a Steam Yacht for trawling . , . , . , 100 0 0 £601 2 11 Professor Hermann's Theory of the Capillary Electrometer. 329 December 10, 1896. Sir JOSEPH LISTER, Bart., F.R.C.S., D.C.L., President, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. The President announced that he had appointed as Vice-Presi- dents — The Treasurer. Professor Clifton. Mr. Thiselton Dyer. Dr. Huggins. The following Papers were read : — I. " On Professor Hermann's Theory of the Capillary Electro- meter." By G-EORGE J. BURGH, M.A. Communicated by Professor BURDON SANDERSON, F.R.S. II. "An Attempt to determine the Adiabatic Relations of Ethyl Oxide." By E. P. PERMAN, D.Sc., W. RAMSAY, Ph.D., F.R.S., and J. ROSE-INNES, M.A., B.Sc. III. " The Chemical and Physiological Reactions of certain Synthe- sised Proteid-like Substances. — Preliminary Communica- tion." By JOHN W. PICKERING, D.Sc. (Lond.). Communi- cated by Professor HALLIBURTON, P.R.S. IY. " An Experimental Examination into the Growth of the Blastoderm of the Chick." By RICHARD ASSHETON, M.A. Communicated by ADAM SEDGWICK, F.R.S. "On Professor Hermann's Theory of the Capillary Electro- meter." By GEORGE J. BURGH, M.A. Communicated by Professor BURDON SANDERSON, F.R.S. Received July 21, —Read December 10, 1896. I have received, by the courtesy of Professor Hermann, a copy of his paper* on " Das Capillar-Electrometer und die Actionsstrome des * * Archiv fur die Ges. Physiologic,' vol. 63, p. 44C. 2 C 2 330 Mr. G. J. Burch. On Professor Hermanns Muskels," in which he discusses the analyses of certain electrometer curves of muscle variation described by Professor Burdon Sanderson.* His first statement demands an explanation on my part. He says, " Bevor ich auf Sanderson's Versuche und Schliisse eingehe, mochte ich zeigen dass der von Burch und von Einthoven aufgestellte, das Capillar-Electrometer betreffende Satz, welcher der Construction zu Grunde liegfc, auch aus meiner Theorie des Instruments unmittelbar folgt, was beide Autoren, obwohl sie meine Arbeit erwahnen, nicht bemerkt haben. Da beide ihren Satz empirisch gewonnen haben, so kann derselbe als eine schone Bestiitigung meiner Theorie betrachtet werden." As a matter of fact, I did not know of Professor Hermann's paper until after I had formed my own theory. In my second paperf on the subject I mentioned that it had also been treated by him, " mainl}- from a mathematical standpoint," and implied that, in my opinion, his data were insufficient. I still think so, and cannot admit that my experimental results prove the correctness of his views. That a mathematical formula, based upon a certain hypothesis, should agree with observed facts may be strong evidence in its favour, but is not necessarily a proof of the soundness of the hypo- thesis. For instance, the equation p — E . e~rt may represent the discharge of a Ley den jar through a circuit of no inductance, or the swing of a pendulum in treacle. That it happens to be also the expression for the time-relations of the capillary electrometer does not of itself imply that the same causes are at work in all three cases, but simply that the forces concerned are so related that the movement is dead-beat. Professor Hermann, starting from Lippmann's polarisation theory, assumes the simplest conceivable relation between the rate of polarisation and the acting P.D., namely, that they are proportional to one another. Putting i = the intensity of the current, and p = the amount of polarisation at the time t, he gets dp/dt = hi, in which li is an instrumental constant. Writing E for an electromotive force, which may be constant or variable, and w for the resistance of the circuit, he arrives at the differential equation * ' Journal of Physiology,' vol. 18, p. 117. f "Time-Relations of the Capillary Electrometer," 'Phil. Trans.,' A, vol. 183, p. 81, 1892. Theory of the Capillary Electrometer. 331 My position in relation to the problem was very different. I wanted to make a capillary electrometer from the description given in Lippmann's Theses. In order to get better results, I determined by actual experiment what were the conditions of sensitiveness and rapidity, and in doing this found out so much about the instrument that the "einfachste denkbare Annahme," referred to by Hermann, would not have commended itself to me. My paper on the " Time-Relations of the Capillary Electrometer " was a condensed account of a small portion of the work done by me. For various reasons I did not then enter into my views as to the theory of the instrument, and will confine myself here to a statement of them, which must be regarded as preliminary. Professor Hermann speaks of my theory having been empirically obtained. I demur to that expression as open to misconstruction. My working formula may rightly be called empirical, since it neglects certain tsrms of the complete expression, which I have found to neutralise each other in a suitably selected instrument, but my theory of the time-relations of the capillary electrometer was founded upon first principles and verified by experiments. My starting point was the fundamental fact that in the capillary electrometer a mechanical effect is produced by an electrical cause. But there are several links between the cause and the effect, and a strong probability that each of them involves a time-function. They are shown in the following scheme : — I. II. III. IV. A difference of A change in the Presumably giving And does work in potential (the establishment of which is delayed constant of capillarity at two interfaces between rise to polarisa- tion at the afore- said interfaces. moving a column of mercury, against the force of gravity by the (varying) mercury and an (with more or less internal ohmic electrolyte. rapidity according resistance of the to the (varying) electrometer) amount of fluid produces friction in the tube). Poiseuille showed in 1846 that the flow of a liquid through a capillary tube varies directly as the pressure. Of this I was not aware till later, but it leads to precisely the same differential equa- tion as that adopted by Hermann. Writing Q for the quantity of electricity, C for the constant of capillarity, P for polarisation, and W for the work done, the sym- bolical expression of the problem is — /(Q/, C,, Pt) = 0(W,). 332 Mr. G. J. Burch. On Professor Hermanns Hermann has passed over C, and omitted to take W into account, confining himself to the theoretical relation between Q* and P/. But we know very little about polarisation, save in the case of solid electrodes. The term polarisation, as frequently used, includes two phenomena, which ought to be kept distinct, viz. : — (a) That condition of the interface between two conductors, of which one at least is an electrolyte, in which the molecules are under a stress not greater than they are capable of supporting without chemical change. . (&) A deposit upon the surface of a solid, or in the contiguous liquid, of the products of actual electrolysis. If one of the conductors is a solid, the inevitable local differences of condition or of composition enable actual electrolysis to take place, even with a P.O. smaller than that proper to the chemical change implied. But if both conductors are liquid and perfectly pure, the stress is so far equalised that no electrolysis is possible until the E.M.F. reaches a certain value, more sharply defined in proportion as the materials are pure. I hold that with differences of potential which do not reach this limit, the electromotive force is transmuted without electrolysis into mechanical force, and manifests itself as kinetic energy, until by the motion of the meniscus it becomes transformed into potential energy. The locus of transformation from electrical to mechanical force must clearly be the two interfaces mercury-acid and acid-mercury, and it is upon these that the stress acts. The resistance is distributed along the tube, and is partly electrical, but to a far larger extent mecha- nical. Is it reasonable, therefore, to assume that the sole cause of delay is the " Polarisations-geschwindigkeit " of the meniscus ? I believe that in the case of an interface between two liquids, the rate of polarisation is to be measured in terms of the vibration-period of a molecule, rather than in decimals of a second. Actual electrolysis is another matter, and I hold that it does not take place in a properly working electrometer. We do not assume electrolysis when two pith balls repel each other after receiving a charge, nor when a closed coil is slipped over a magnet. But the coil cannot be got off again, nor can the balls fall together without the generation somewhere of a current. , I cannot see why we should assume electrolysis in the case of the capillary electrometer. The marvellous rapidity of the action to which I have not yet found a limit, is against it, as is also the fact that the substitution for the acid, or the addition to it, of any substance which can be electrolysed by a smaller electromotive force, reduces the range of potential dif- ference for which it can be used. Theory of the Capillary Electrometer. 333 The presence of even a trace of impurity is soon manifested by the blocking of the capillary, and if this block is removed by electrolysis, the instrument behaves for some time in an abnormal way. It shows signs of a residual charge, like that of a Leyden jar, the mercury rising again after the short-circuiting key is opened, instead of simply ceasing to fall. This I ascribe to polarisation of the kind met with between solids and electrolytes, and to this the term "Pol arisations-geschwindig- keit " would be applicable. But no good electrometer will show it, except with electromotive forces greater than ought to be employed. I have held from the first that the capillary electrometer acts by transforming electrical into mechanical energy without any chemical interchange, and that this is possible because at the interface between two liquids which do not diffuse into each other the stress is so evenly distributed that no one molecule can be strained to a degree sufncient to detach any part of it until the stress is intense enough to break down all similar molecules simultaneously. But if by polarisation is meant this condition of the interface, then I maintain that it must precede the movement, and must be deve- loped with almost inconceivable rapidity. In order to investigate the form of curve produced by recording the motion of the meniscus when the electrometer is acted upon by an electromotive force varying with the time according to some known law, e.g., the pulsating or alternating current of a dynamo, Professor Hermann puts his equation into a somewhat different form, namely : dpjdt+rp—rof(t) = 0, where r and e are constants, and cf(t) = E is the electromotive force represented as a function of the time. But this is simply my own formula for the estimation of the E.M.F. expressed as a differential equation. For dp/dt is, in the polar curves taken with my machine, merely the subnormal N", and rp is identical with &Ar, whence dpjdt + rp rcf(t) is identical with K + *Ar - / («) volt, which being interpreted signifies fThesub-1 f A constant mul- 1 f The I normal I , J tiple of the dis- I f A constant! I E.M.F. at 1 to the (^ 1 tan ce from the f " \ multiple of /] time t (m curve. J zero-line. J I volts). 334 Mr. G. J. Burcli. On Professor Hermanns Professor Hermann finds the complete primitive of this differential equation, and then, introducing various values of r and the function E = e/(0> draws, by a process which is indeed laborious, the curves of the corresponding excursions. My own method gives a good deal of the information so obtained in a much simpler manner. Adopting the letters used by him, when/ vanishes we have dp,'dt + rp = 0, that is to say, whenever the E.M.F. falls to zero the reduced values of the subnormal and the radius vector are equal, but of opposite sign, and the curve, therefore, can never come back to the zero line under the action of a current which pulsates but does not alternate (see figs. 2 and 4 in Hermann's paper). When the meniscus crosses the zero line, rp = 0, and dp/dt = ref(t), i.e., the impressed E.M.F. is then directly proportional to the subnormal. This involves the further fact that the crossing of the zero line by the meniscus must always lag behind the change of sign of the E.M.F. If dpjdt vanishes, as it does at the apex of a spike or the bottom of a notch, the instantaneous value of the impressed E.M.F. is directly proportional to the distance of the meniscus from zero. The curves drawn by Professor Hermann are for the most part, so far as the eye cau judge, similar to those obtainable under like condi- tions with the capillary electrometer. I have photographed and analysed many such, using rheotomes and dynamos of various kinds, both alternating and direct current, as sources of E.M.F. I have proposed, in a paper which has been in the publisher's hands since last November, that this method should be used to determine the characteristic current curves of dynamos.* All the confusing influence of the lag vanishes when such curves are analysed — there is no need to trouble about the equation to the curve, since each several term of its differential equation at any given point is found at once by my mode of analysis. But I must point out that an error has crept into Professor Hermann's rendering of the curve given in fig. 6 — or, rather, as it only pretends to be an approximation, that it is not equally accurate throughout. The por- tion c'd'f which corresponds to a diminishing negative (below zero) potential is represented as rising with 'increasing velocity instead of falling more slowly, as it should do. Yet, when this negative poten- tial ceases, the curve commences to fall from d' to e' along the logarithmic curve of discharge. This is impossible. When e f(t) is negative, the algebraic sum of dp/dt and rp must be negative also if the fundamental equation holds good. Probably the straight line cd has been placed too far to the right. * ' The Electrician/ July 17, 1896, et sey. Theory of the Capillary Electrometer. 335 Professor Hermann questions the accuracy of my method of analysis when applied to steep curves. My answer is that I do not employ it in such cases, preferring to take photographs of sudden changes upon plates moving with suffi- cient rapidity to suitably develope the curves. Thus in Professor Burdon Sanderson's paper,* figs. 1, 2, 3, and 4, on Plate 1, and figs. 3, 4, and 5, on Plate 3, were intended to show within the limits of a page the entire course of certain phenomena. I did not analyse them, but simply measured the times of the maxima and minima. The remaining curves, viz., figs. 5 and 6, Plate 1, figs. 1 — 7 on Plate 2, and figs. 1 and 2 on Plate 3, are all suitable for analysis, with the exception of the first phase of fig. 1, which is almost too steep. I have done some thirty or forty of this kind. As regards the further criticisms, so far as the physical interpre- tation of the curves is concerned, I can only say that cases did occur in which the maximum E.M.F. of the second (positive) phase exceeded the maximum E.M.F. of the first (negative) phase of the same response. With respect to curves, like those in figs. 3 and 4, Plate II, the part referred to by Professor Burdon Sanderson as the " hump," is not merely the curve of discharge. The actual negatives which I measured show a rise of the meniscus after its rapid downward movement has ceased, and while it is still above the zero line, and a similar rise is plainly visible to the eye after every one of the "spikes " in figs. 1, 2, and 4, Plate 1, which were photo- graphed with the machine moving more slowly. It is impossible for the mercury, under these conditions, after approaching the zero line, to recede from, without crossing it, except under the influence of a negative Acting P.D. That is to say, the Impressed E.M.F. must be of the same sign as the charge already in the instrument, but must be of higher potential difference. In some negatives this second rise in followed by a descent more rapid than that of the curve of discharge, and therefore indicating a small positive Acting P.D. I first noticed and called attention to it in connection with the curves illustrating my paperf on the " Time Kelations of the Capillary Electrometer," but refrained from discus- sing its physiological significance. * ' Journal of Physiology/ vol. 18, p. 117. f < Phil. Trans.,' A, vol. 183, p. 104* 336 Attempt to determine the Adiabatic Relations of Ethyl Oxide. " An Attempt to determine the Adiabatic Relations of Ethyl Oxide." By E. P. PERMAN, D.Sc., W. RAMSAY, Ph.D. F.R.S., and J. ROSE-INNES, M.A., B.Sc. Received November 6,— Read December 10, 1896. (Abstract.) The wave-length of sound in gaseous and in liquid ethyl oxide (sulphuric ether) has been determined by the two first-mentioned of the authors, by means of Kundt's method, between limits of temperature ranging from 100° C. to 200° C., and of pressure ranging from 4000 mm. to 31,000 mm. of mercury, and of volume ranging from 2'6 c.c. per gram to 71 c.c. per gram. Making use of the same appa- ratus throughout, the results obtained are to be regarded as com- parative, and, by careful determination of the pitch of the tone transmitted through the gas, it is probable they are approximately absolute. The sections of the complete memoir deal with (I) a description of the apparatus employed, (II) the method of ascertaining the weights of ether used in each series of experiments, (III) determinations of the frequency of the vibrating rod, (IV) the calculations of th£ adiabatic elasticity and tables of the experimental results, and (Y) a mathematical discussion of the results. The last section is due to Mr. Rose-Innes. As the theoretical results are of interest, a brief outline of them may be given here. It will be remembered that one of the authors, in conjunction with Dr. Sydney Young, showed that for ether, and for some other liquids, a linear relation subsists between pressure and temperature, volume being kept constant, so that p = bT — a. It has been found that a similar relation obtains between adiabatic elasticity and temperature, volume, as before, being kept constant ; so that, within limits of experimental error, where E stands for adiabatic elasticity, E = jT-A, g and h being functions of the volume only. Between these two equations, we may eliminate T, and so express E as a linear function of p, volume being kept constant. The coefficient of p in such an equation would be g/b, and this fraction, on being calculated from the data 'available, proves to be nearly constant. For working pur- poses it is assumed that g/b may be treated as strictly constant, and Reactions of certain Synthesised Proteid-like Substances. 337 it is shown that this assumption does not introduce any serious error within the limits of volume considered. We tnen find it possible to integrate the resulting differential equation, and the complete primi- tive enables us to draw a set of adiabatic curves. We believe that this is the first time adiabatic curves have been obtained for any substance except perfect gases. A mathematical discussion is added as to what extent the equations E = gT-h and yjb = constant, can be considered as strictly true, and not merely approximate. The experimental results for liquid ether form an appendix to the paper. "The Chemical and Physiological Reactions of certain Synthesised Proteid-like Substances. Preliminary Com- munication." By JOHN W. PICKERING, D.Sc. (Lond.). Communicated by Professor HALLIBURTON, F.R.S. Re- ceived November 10, — Read December 10, 1896. The experiments of Professor Grimaux,* made more than ten years ago, have until recently attracted but little attention amongst English physiologists, although that investigator has synthesised a series of colloidal substances which, in their chemical characteristics, show striking similarities to proteids. Working alone, and in collaboration with Professor Halliburton, If have shown that three of the substances synthesised, viz., the " Colloids amidobenzoic A and B," formed by the interaction of phosphorus pentachloride and meta-amido-benzoic acid at 125° C., according to the details described in Grimaux's papers, and the " colloide aspartique " formed by the passage of a current of dry gaseous ammonia over solid aspartic anhydride heated to 125° C., not only give the leading chemical reactions of proteids, but when intra- venously injected into dogs, cats, or pigmented rabbits, cause extensive intravascular coagulation of the blood, in a manner indis- tinguishable from the physiological action of nucleo-proteids. When injected into the veins of albino rabbits or into the vascular system * G-rimaux, ' Comptes Kendus,' vol. 93, p. 771, 1881 ; ibid., vol, 98, p. 105, 1884 ; ibid., vol. 88, p. 1434 and p. 1578. t Pickering, ' Journ. Pliysiol.,' vol. 14, p. 341, 1893 ; ' Comptes Eendus,' vol. 120, p. 1348, 1895; 'Physiol. Soc. Proc.,' Feb. 1.6, 1895 ('Journ. Physiol.,' vol. 17) ; 4 Journ. Physiol.,' vol. 18, p. 54, 1895; Hid., vol. 20, p. 171, 1896; ibid., vol. 20, p. 310 ; Halliburton and Pickering, ' Journ. Physiol.,' vol. 18, p. 285, 1895. 338 Dr. J. W, Pickering. The Chemical and Physiological of the Norway hare (^Lepus variabilis), during" its albino condition, these substances fail to induce intravascular coagulation of the blood, although they hasten the coagulation of the blood when drawn from the carotids, in a precisely similar manner to nucleo-proteids. Taking these facts as the basis of my investigations, I have en- deavoured to synthesise substances which will approach more nearly in their chemical and physiological reactions to proteids than those briefly described above ; and to further investigate the properties of Grimaux's colloids. I. General Description of Experiments. I have up to the present synthesised seven different colloidal sub- stances, by the interaction of either phosphorus pentachloride or pentoxide on certain well-known derivatives of proteids, and the details of their preparation, physical properties, chemical and physio- logical reactions are described below. Colloid a. — Prepared by the interaction of equal parts of meta- amido-benzoic acid, biuret, and three times its weight of phosphorus pentoxide at 125° C. in a sealed tube. The best results are obtained by continuing the heating for about six hours, although a similar substance is obtained by heating for half an hour at 130° C. The product of the reaction is a pinkish-grey friable powder, which is insoluble in cold water, and almost insoluble in boiling water. This substance should be repeatedly washed until all traces of phosphoric acid a.re removed. When heated with Millon's reagent it fails to give the reaction characteristic of tyrosine and proteids ; it also does not give the well-known colour reactions with the salts of copper, nickel, cobalt, and caustic potash. It gives the typical blue reaction associated with the name of Frohde* when heated with sulphuric and molybdic acids, as well as the xantho- proteic reaction. If the amount of biuret exceeds the amount of meta-amido-benzoic acid, then the excess of biuret left over gives its typical colour reaction with copper sulphate ancl potash. The pinkish-grey powder, obtained by the reaction described above, should be dissolved in ammonium hydrate, and the resulting solution evaporated down at the temperature of the atmosphere in vacuo, when the resulting product appears as a number of translucent yellowish plates, which are tasteless and inodorous, and closely resemble in appearance both Grimaux's " collo'ides amido-benzoique and aspartique " and dried serum-albumen. These plates are with difficulty soluble in cold water, but readily pass into solution on warming. The solution obtained does not coagulate on heating, but * Frohde, ' Annalen der Chemie,' vol. 145, p. 376. Reactions of certain Sunthesised Proteid-like Substances. 339 if a trace of a soluble salt of either barium, strontium, calcium, magnesium, or sodium be added, a pronounced coagulum is obtained 011 heating. This point will be returned to you in a subsequent- section, but the similarity to dialysed serum-albumen may be pointed out, as that substance is stated not to coagulate when heated.* The solution does not coagulate spontaneously on standing, neither will the addition of "fibrin ferment (i.e., a nucleoproteidf) induce coagulation. It gives a typical xanthoproteic reaction, a violet with copper sulphate and potash, a dark heliotrope-purple with cobalt sulphate and potash, and a faint yellow with nickel sulphate and potash. It also gives Frdhde's sulpho-molybdic reaction ; I may, however, remark that I found that several substances chemi- cally allied to proteids yield this reaction, which is therefore not diagnostic of proteids alone. An alcoholic solution of alloxan gives with the solid plates a brilliant red coloration (Krasser'sJ reaction) similar to that produced with plates of serum-albumen. Negative results were obtained with the reactions associated with the names of Liebermann,§ Adamkiewicz,|| and Millon.^f The solution is neutral and laevorotatory (aD = —52), and if treated with pepsin and a O2 per cent, hydrochloric acid, or by an alkaline solution of trypsin, for several days at 38° C. it does not peptonise. Qualitative analysis shows that this substance does not contain phosphorus in its molecule. It is precipitated from solution by mercuric chloride, silver nitrate, and lead acetate. These precipitates yield the same colour reactions as the original substance. The precipitate formed by the addition of lead acetate, like that obtained by the addition of this substance to a proteid solution, redissolves on the passage of a current of sulphuretted hydrogen through the solution in which it is suspended, and judging by chemical tests alone, the nature of the substance is unchanged by the processes of precipitation and redissolving. Its physiological action is, however, markedly changed, as will be shown later on. The original solution is readily precipitated by trichloracetic, phosphotungstic, phosphomolybdic acids, and by acetic acid and potassium ferrocyanide, as well as by salicylsulphonic acid ; the pre- cipitate formed by this last substance is coagulated by heating in a manner similar to the coagulation produced by heating the pre- cipitate resulting from the addition of this substance to a proteid * Schmidt and Aronstein, ' Pfluger's Arcliiv,' vol. 8, p. 75, 1874. f Vide Halliburton, ' Journ. Physiol.,' vol. 18, p. 306, 1895. J Krasser, ' Monat. i'iir Chem.,' vol. 7, p. 673 ; ' Muly's Jahresb.,' vol. 16, p. 1. § Liebermann, ' Maly's Jahres.,' vol. 18, p. 8. || Adamkiewicz, 'Ber. d. deut. Chem. Gresell.,' vol. 8, p. 761. If Millon, ' Comptes Kendus,' vol. 28, p. 40. 340 Dr. J. W. Pickering. The Chemical and Physiological solution. I may here mention that salicylsulphonic acid does not precipitate disintegration products of proteids like leucine, tyrosine, xanthine, or hypoxanthine. All the precipitates cited above give the colour reactions charac- teristic of the original substance. If the original solution is saturated with either magnesium sul- phate, ammonium sulphate, or sodium, chloride, the whole of the colloid rises to the surface of the liquid, and may be skimmed off. On placing this scum in an excess of distilled water, it rapidly redissolves, forming a pale yellow opalescent solution, which gives all the chemical reactions characteristic of the original substance. If the amount of neutral salt be insufficient to produce precipita- tion, the passage through the liquid of a current of carbon dioxide or of sulphur dioxide will effect the same result. Neither of these gases will, however, cause precipitation in the entire absence of salts. The following experiments illustrate the results produced by the intravenous injection of this substance into dogs, rabbits, and cats. The procedure adopted was identical with that described in the previous papers published by Professor Halliburton and myself,* on the intravascular injection of Grimaux's colloids. In all cases the animal was ana3sthetised by a mixture of chloroform and ether, an excess of the latter substance being used when the subjects were dogs. 'Experiment 1. — Fox terrier (weight 27 Ibs. 10 oz.) ; 25 c.c. of a 0*75 per cent, solution of the colloid a was injected, and proved fatal. Pronounced exophthalmos and dilatation of the pupils, and typical stretching movements were observed. Post-mortem examination made immediately after death revealed pronounced clots in the jugular vein, inferior vena cava, and portal vein, and a slight clot in the left ventricle and in the pulmonary artery. Experiment 2. — Large black cat (weight 9 Ibs. 6 oz.) ; 40 c.c. of the colloid proved fatal, with similar symptoms as above. Immediate post-mortem examination showed pronounced clots in the left ventricle, right auricle, inferior vena cava, portal, and jugular veins. The remainder of the blood was fluid, but coagulated very rapidly after withdrawal. Experiment 3. — Black rabbit ; 38 c.c. of the same substance pro- duced a similar result. Experiment 4. — Albino rabbit ; 42 c.c. proved fatal. Death was accompanied by pronounced exophthalmos and dilatation of the pupils and stretching movements of the limbs. Post-mortem exami- nation showed the blood throughout the vessels to be fluid. It, how- ever, rapidly coagulated after withdrawal from the vessels, and the coagulability of samples of the blood taken from, the carotids during • Op. cit. Reactions of certain Synt/iesised Proteid-li/se Substances. 341 the injection of the colloid was also hastened ; thus after 20 c.c. of the colloid had been injected, the time of complete coagulation of blood withdrawn from the carotids was hastened by 2 minutes, after 30 c.c. by 3f minutes, and after 35 c.c. by 4 minutes. It will be evident that the results recorded above are similar to, if not indistinguishable from, those produced by the intravenous injection of a nucleoproteid. When slowly introduced into the circulation of dogs, and to a much lesser degree of rabbits, in minute quantities, the effect pro- duced on the coagulability of the blood is the converse of that resulting from the introduction of larger quantities. This effect is more pronounced than that obtained by the intravenous injection of Grimaux's colloids, and more resembles Wooldridge's* " negative phase," which is characteristic of a nucleoproteid, but is not so pro- nounced as the result obtained with that substance. This result is illustrated by the following experiment : — Experiment 5. — Large black mongrel. Anaesthetic, ether and morphia (weight, 60 Ibs.) ; 1 c.c. of a 0'025 per cent, solution colloid a was injected very slowly, the injection being distributed over half an hour, at the end of which time the retardation of the time of coagu- lation of blood withdrawn from the animal's carotid was found to be 8 minutes 30 seconds. A second dose of 1 c.c. of the same solution injected and distributed over 20 minutes caused a further retardation in the time of coagulation of the carotid blood of 2 minutes ; but a third injection distributed over a similar period of time hastened the coagulability of the blood that had been previously retarded, so that the retardation, as compared with the time of coagulation before the injection of the colloid, was only 1 minute 30 seconds. After a still further injection of the colloid, the blood coagulated more rapidly than in the normal condition, and finally, when the dose was pushed, intravascular coagulation of the animal's blood occurred, and death resulted. If the colloid is separated from the solution by saturation with magnesium sulphate, sodium chloride, or ammonium sulphate, as before described, and the scum redissolved in distilled water, the opalescent solution obtained will, when intravenously injected into pigmented rabbits, produce typical intravascular coagulation. Repetition of the process of precipitation and redissolving however, destroys the physiological activity in a manner similar to the result produced with both nucleo-proteids and Grimaux's synthesised colloids. If the solution formed by the passage of a stream of sulphuretted hydrogen over the precipitate formed by the addition of lead acetate to the colloid is injected intravenously into pigmented rabbits or * Wooldridge, p = 106, q.p., so that cT 1Q1U Hence, unless the latent heat of carbon be enormously great com- pared with that of other substances, cT/T will be considerable. If X be as great as the latent heat of vaporisation of carbon given by Trouton's law, i.e., about 4000 calories, or 16'8 X 1C10 ergs, £T/T would be about -fr, and £T would be nearly 220° C. for each atmo- sphere, and a change of pressure of about 18 atmos. would raise the temperature of the crater to that estimated for the sun. The corre- sponding increase of radiation would be very great, for the radiation varies, at least approximately, as the fourth power of the absolute temperature. This would lead one to expect that the radiation would be nearly doubled for each 4 atmos. added. Such an increase as this certainly does not take place, so that we may conclude that either the temperature of the crater is not that of boiling carbon, 382 Effect of Pressure on temperature of Crater of Electric Arc. or else that the latent heat of volatilisation of carbon is very con- siderably greater than that calculated from Trouton's law. Even though this latent heat were as great as the heat of combustion of C to C02, i.e., 7770, there would be an increase of about 70 per cent, in the radiation for an increased pressure of 6 atmos. Such an enormous latent heat is unprecedented, and yet our experiments would, almost certainly, have shown such an increased radiation as this. So far, therefore, the experiments throw considerable doubt on the probability that it is the boiling point of carbon that determines the tempera- ture of the crater. It might be questioned whether there is energy enough in the current to do all this work, but upon an extravagant estimate of the amount of carbon volatilised in the crater, it appears that there is more than a hundred times as much energy supplied by the current as would be required for volatilising the carbon, even though its latent heat were as great as the heat of combustion of C into CO2. There is another considerable difficulty in the theory of the tem- perature of the crater being that of boiling carbon arising from the slowness of evaporation. The crater on mercury is dark, but then it volatilises with immense rapidity and the supply of energy by the current being more than 100 times that required merely for evapora- tion, there seems very little reason why even a considerable difference in latent heat should make any sensible difference in the rate of evaporation of mercury and carbon, especially as, at the same tem- perature, the diffusion of carbon vapour is nearly three times as fast as that of mercury vapour and the temperature immensely higher. We would, in conclusion, call attention to a cause of opacity in the solar atmosphere that is illustrated by the effect of convection currents in the long tube we were observing at high pressures ; these convection currents behaved just like snow, or any other finely divided transparent body immersed in another of different refractive index. Light trying to get through is reflected backwards and forwards in every direction, until most of it gets back by the way it came. The con- sequence was that even the electric arc light was unable to penetrate the tube at high pressure, when these convection currents were active. The only light that came out of the tube was the feeble light outside, which was returned to us by reflection at the surfaces of these con- vection currents. In a similar manner we conceive that any part of the solar atmosphere which is at a high pressure, and where convec- tion currents, or currents of different kinds of materials, are active, would reflect back to the sun any radiations coming from below, and reflect to us only the feeble radiations coming from interplanetary space. In his paper on " The Physical Constitution of the Sun and Stars " (' Roy. Soc. Proc.,' No. 105, 1868), Dr. Stoney called attention to an action of this kind that might be due to clouds of transparent Influence of Temperature upon Electrotonic Currents. 383 material, like clouds of water on the earth, but in view of the high solar temperature it seems improbable that any body, except, perhaps, carbon, could exist in any condition other than the gaseous state in the solar atmosphere ; so that it seems more probable that sun-spots are due, at least partly, to reflection by convection streams of gas, rather than by clouds of transparent solid or liquid particles. " Influence of Alterations of Temperature upon the Electro- tonic Currents of Medullated Nerve."* By AUGUSTUS D. WALLER, M.D., F.R.S. Received December 14, — Read December 17, 1896. (Abstract.) The effects of a rise of temperature upon electrotonic currents may be briefly stated as follows : — 1. The ordinary electrotonic currents, A and K, are temporarily diminished or abolished at about 40°. 2. At about 30° of a rising temperature the K current is increased without notable alteration or with actual diminution of the A current. 3. On returning from 40° towards the normal (15° + 2°) tempera- ture, the A and K currents reappear. K is increased and A is diminished, so that the previous normal inequality A > K is diminished, or actually reversed to A < K. In all cases the quotient A/K is diminished ; in some cases it actually falls below unity. [The negative variation is temporarily abolished at about 40° ; a positive gives place to a negative variation in consequence of a raised temperature to 40°.] The above three statements are illustrated by Experiments 2366, 2322, and, from the examination of their records, it will be clear that there is here no question of the effects being due to alterations of resistance. A and K are tested for alternately, and the deflection by O'OOl volt is taken at intervals of about ten minutes. [Other examples of a similar character are given in the * Proceedings of the Physiological Society' for November, 1896, and a record of temporary diminution of the negative variation is given in fig. 12 (Experiment 777), * Phil. Trans.,' 1897.] * In all the experiments referred to in this communication, the polarising cur- rent is by one Leclanche cell (the resistance in its circuit being about 100,000 ohms). The nerve lies upon four unpolarisable electrodes fixed at intervals of 12 mm., serving as leading-in electrodes to the polarising current and leading-out electrodes to the electrotonic current. On the galvanometer records, the anelectro- tonic deflection A reads upwards, the katelectrotonic deflection K reads downwards ; aiter-anelectrotonic and after-katelectrotonic deflections A' and K' read respectively downwards and upwards (there being under the conditions of experiment no marked homodromous after-katelectrotonic deflection). 384 Dr. A. D. Waller- Influence of Alterations of Exp. 2322. — Influence of raised Temperature upon Anelectrotonic and Katelectrotonic Currents. Time. Tempera- ture. A. A'. K. K'. TO'TTO volt. i i Omin. 17° _ 1 _ _ __ 9 1 5) + 12 -2 — — 2 J> — — — trace + 2 5 >> + 12 -2 — — 6 >> — — — trace + 2 fio 21 + 12-5 -2-5 11 — — — — trace + 3 15 30 -5 -t-3-5 •§' j 16 + 11-5 -5 — || 20 38 + 3 — — — . 21 39 — — i — 25 39 — — -2 — L26 38-5 + 3 — — — 30 35-5 + 5 -1-5 . 31 35 — — -3-5 +0-5 40 28 + 8-5 -2-5 — — 41 — — — -4-5 +1 50 24 + 8-5 -2 — — 51 — — — -4-5 +1 52 ~ _— — 9 The K current is very small, the K' after-current is comparatively large. In consequence of heating to 39'5°, K is increased, A and K' are diminished. The quotient A/K is diminished. m&w * Temperature upon Klectrotonic Current*. 385 Exp. 2366. — Influence of raised Temperature upon An electro tonic and Katelectrotonic Currents. Time. Temperature. A. K. T5U VOlfc. 0 min. 16° _ _ 8 1 V + 11-5 — 2 ?> — -4-5 7 + 11 -5 — 8 » — -4 f!2 19 — — 8 114 22-5 + 12 — 15 24 — -4-5 16 26 -12 — Hi 17 fills 28 29-5 n-11 -5 119 31-5 — — 20 33 + 10 — 21 35 — -5 125 40 + 2 — 26 40 1-5 30 36 + 4 — 31 35 — -4 32 34 — — 10 33 33 + 4 — 34 32 — — 7 42 26-5 — — 7 43 26 _ -8-5 44 25-5 + 6-5 52 22-5 — — 6-5 53 22 + 6 — 54 22 — -6 After heating to 40° the A current is diminished, the Jv current is increased, and well-marked A' after-current has developed. The quotient A K is diminished; 386 Dr. A. D. Waller. Influence of Alterations of Electrotonic after-currents, A' and K', when present to any marked degree, are opposed to the previous electrotonic currents A and K. Designating A and K respectively as positive and negative, the after-currents A' and K' are respectively negative and positive. Such after-currents are in general modified by previous rise of tempera- ture, which gives rise to an evident A' (negative) in a nerve which previously gave no marked A', and abolishes a K' (positive) that may previously have been present. Experiment 2366 exhibits the development of an evident negative A' subsequent to heating of the nerve. Experiment 2322 exhibits the abolition of a positive K', evident previous to heating of the nerve. A fall of temperature causes an increase of the A current and, in less degree, of the K current; by reason of the diminution of resistance that takes place with lowered temperature, the increase of A is more marked than is apparent upon the record, and the smaller increase of K is quite masked by the diminution of resistance. The quotient A/K is augmented. At a temperature of —4° to —6° both currents are somewhat suddenly abolished ; this abolition may be complete and final, no recovery taking place, or it may be temporary, being succeeded by imperfect recovery as the nerve temperature returns towards normal. It is noteworthy that the A and K currents are not abolished at 0° suddenly, and all but finally abolished at — 4° Temperature upon Electrotonic Currents. 387 to —6°, probably by reason of the nerve having been frozen at this temperature and thus cut to pieces. It is evident that little stress is to be laid upon an apparent decrease of K with falling temperature (2417) and increase of K \vith rising temperature (2366). On the other hand, a diminished A with rising temperature (2366) and an increased A with falling temperature (2417) are not open to doubt. Exp. 2334-5.— Influence of lowered Temperature upon Anelectrotonic and Katelectrotonic Currents. Time. Temp. A. K. ToW v°lfc- 0 min. 17° __ __ 9-0 1 + 11-5 — 2 — -2-5 7 + 11-5 — 8 M — -2-5 r 9 ] + 11-5 — 10 16-5 — -2-5 17 15 + 11-5 — 18 ]4'5 — -2-5 27 8 + 12 28 7'5 — -2-5 2 J 37 3 + 12 0 I38 2-5 — -1-5 47 -0-5 + 10-5 — 48 -0-5 — -1 56 -3 — -0-5 57 -3 + 6-5 58 -3-5 — — 3-5 .67 -4 0 — 68 -4 0 77 + 1-5 0 — 78 + 2 — 0 80 4 — — 3-5 87 6 •4- 2 — 88 6 — — trace 98 9 + 3-5 99 9 -5 -0'75 108 11 + 3-5 — 109 11 "5 — -1 116 12 — — 4-5 180 14 + 4 -2 6-0 i Influence of Temperature upon JZlectrotonic Currents. 389 Expt. 2417.— Effect of Cold on A and K, Time. Temperature. A. K. A/K. ToVo volt. 0 15° _ 7-5 1 — -4 2 j> + 13 — 3-25 7 5» — -4 — 8 '5 + 13-5 — . 3-37 11 14 — ^ 12 — + 14-5 3-62 13 — . — -4 14 12-5 + 14-5 — 3 62 15 — — -3'5 16 11 + 15 4-28 17 — _ -4 18 9 + 15-5 — 3-87 21 6 — — 5*5 23 4 — -3-5 24 3 + 16 5 . . 4-71 27 2 — -3-5 28 1-5 + 16-5 4-71 29 — ! — -3 _ 30 0-5 4 17 — 5-66 32 — _ 4-5 33 -0-5 + 16-5 — 34 — — -3 — 35 1 + 17 — 5-66 36 — -2-5 37 -1-5 + 17 - — 0-8 38 — -2-5 39 — 2 + 17'5 — 7 40 — — -2-5 41 -2-5 + 17 — 6-8 42 — — -2-5 — 43 -3 + 17 — 6-8 44 — — -2 — - 45 -3 + 16-5 — 8-25 48 — — -2 — 49 -3-5 + 16-5 8-25 52 L — 4 The A effect obviously increases with fall of temperature (increasing resistance) j the K effect apparently diminishes, but actually increases a little, the increase being masked by increased resistance. The A/K quotient is obviously increased. The voltage calculated from the data of this experiment is : — At 15° A = 0-00173 volt. K = 0-00053 volt. 10 5 0 A = 0-00244 A = 0-00285 A = 0-00360 K = 0-00059 K = 0-00064 K = 0-00070 Dr. A. U. Waller. Influence of Alterations of Exp. 2344. — Influence of Alterations of Temperature upon the Electrical Resistance of Nerve. The following experiment (2344) made to test the effect of rising and falling temperature upon the electrical resistance of nerve, and the value attaching to observations of a standard deflection by constant E.M.F. as an indication of altered resistance, shows very Temperature upon Electrotonic Currents. 391 clearly that such standard deflection gives measure not only of the electrical resistance, but also — due reservation being made of the effect of drying in the course of a prolonged observation at raised temperature — is itself available in measure of the alteration of tem- perature of the nerve. Exp. 2344.— Deflections by a small constant E.M.F. (O002 volt) through a .Nerve at rising and falling Temperature and through two Galvanometers. Time. Thermometer. Demonstrating galvanometer, G-J. Recording galvanometer, OK. 1 min. 16-5° 18-5 c m. 14 -0 mm. 5 16*5 18-5 14-0 10 18-0 19'5 15'() 15 24-0 21-5 17 0 20 30-5 25-5 20-0' 25 35-5 28-5 23-0 30 39-0 30-0 23-5 35 40-0 30-0 24-5 40 38-0 29-0 23-0 45 33-0 26-0 20-0 50 28-0 22-0 17-0 55 25-0 20-5 16-0 60 23-0 19-0 15-0 40" 35° J0° « /ncremente of GemperaCurs - increments of o'ef/ecC/on read upon Gf. n >• " measured from the record of G2 . G, 50 JO 15 20 25 JO 65 4O 50 55 60 mm. [Experiments on the comparative effects of acids and bases upon the A and K currents, have shown that within a certain moderate range of concentration (soakage of the nerve in n/15 to nj^O solution for one minute) acid favours the K current and disfavours the A 392 Influence of Temperature upon Electrotonic Currents. Re experi in 1 1 1|o 1 s i .s | - M i^ iH O »0 t- N t- N CC 00 00 CO 1> O rH CO W rH ,H I I I I I I I I I I III I I I I I 7 i 10 ^t O rHIMrH + + + S 3 * § g o g »B «8 O «4H **H y zinc, C. These were iron, copper, silver, a trace of lead, also some sodium and potassium. There was also a trace of chromium, and this, like the trace of iron, was probably pre- cipitated as basic chloride or as hydroxide. The wave-lengths of the lines of iron, silver, and copper need not be recapitulated. Lead 4057 3682 and 3639 Chromium 4289 4274 „ 4253 The metals precipitated by zinc after addition of lead acetate, F. — The metallic deposit yielded a complex spectrum containing the lines already mentioned of the following elements : iron, chromium, copper, silver, gallium (a trace), potassium, sodium, and, of course, lead, as this had been added. The lead here appears as a banded spectrum, the edges of the bands seen being those at wave-lengths : — 5675 5460 4980 4824 4657 4597 4370 4314 4225 4140 4061 4057 3985 and 3954 Nickel lines. ... 3525 „ 3415 The copper lines were very strong, the silver weak. The potassium lines were the following : — 4047-1 4043-5 strong, then much fainter — 3447-5 and 3446'5. On the Occurrence of Gallium in Clay -ironstone. 397 The precipitates of Phosphates and basic acetates D, E, and G. Precipitate E. Chromium 5206 4289'0 4274'0 4253'0 3606 3594 0 3579*0 Gallium strong 4171'6and 4032'7 (the latter somewhat weaker). Calcium weak 4226'8 Potassium strong 4047'0 4043'5 Sodium strong 5893'0, faint 5635, and 3303'0 Precipitate D, The chromium line 5206 did not appear in the spectrum of this precipitate. Both the gallium lines were very distinct, 4171*6 and 4032-7. It is remarkable how very generally the spectrum of potassium appears along with that of the precipitated substances, whether metals or basic acetates. Precipitate of basic acetates, G. This contained iron, chromium, lead, gallium, potassium, ami sodium. The lines were those which have already been particu- larised. The Residue left by Zinc, H. — This was heated with aqua regia, when all but a very small quantity of silica with a trace of a metallic oxide dissolved. The liquid was filtered and the filtrate evaporated with excess of hydrochloric acid to remove nitric acid. It was diluted with water, when it showed a green colour. It was saturated with sulphuretted hydrogen and filtered to separate the precipitate. The precipitate was partially soluble in sodium hydrogen sulphide, yielding a sherry-coloured solution ; the constituent causing. this colour was not identified, the quantity present being very small. The residue, insoluble in alkaline sulphide, contained copper and a. trace of lead, but no mercury, bismuth,' or cadmium. The filtrate from zinc and precipitated metals J, was diluted and heated to boiling. It gave a precipitate, and therefore ammonium acetate was added to the hot liquid, and after boiling for several minutes it was filtered. The filtrate became turbid immediately ; it was then boiled and more ammonium acetate added and then filtered ; the filtrate again became turbid. This precipitate was filtered off and heated in the oxyhydrogen flame. It contained no gallium, but the spectrum gave lines of iron, copper, sodium, potassium, and a .trace of lead. It is evident that all the gallium was extracted by the repeated additions of ammonium acetate solution and boiling. 398 Prof. W. N. Hartley and Mr. H. Ramage. The various precipitates of basic acetates were mixed, with the exception of that from 7, which contained no gallium. In order to separate phosphoric acid, the precipitates were fused with about three times their weight of mixed carbonates. Some potassium nitrate was added towards the end of the fusion, to convert chromium into chromates. The heavy metals were left as oxides or carbonates, the phosphoric acid going into solution. After extraction with hot water, the solution was filtered. Filtrate L. Residue M. Coloured greenish by man- Dried and fused in a silver dish ganates, boiled -with a few drops with caustic soda to dissolve of alcohol to separate manganese gallium hydroxide. Extracted as hydroxide. Solution, after with water and filtered. Residue again filtering from manganese, not examined further. Solution : was yellow from chromates. acidified with HC1 and ammo- nium chloride and ammonia added. The precipitate was fil- tered off, dissolved in HC1, and sparked to observe its spectrum. These gallium spectra showed that there were still traces of chromium in the gallium chloride, and from this the gallium was purified completely by precipitation in a strongly acid solution with potassium ferrocyanide and subsequent removal of the iron by treat- ment with sodium hydrate, according to the method of Lecocq de Boisbaudran.* The foregoing description of the analytical details proves the presence of gallium in the metal, and gives a clear indication of how it may be separated by a simple process. In subsequent operations on the blast-furnace metal, the ferrous chloride was mixed with calcium carbonate, and the gallium was found to be all precipitated and capable of easy separation 'from the calcium salt.f Latterly it was found i.o be more convenient to boil the acid solution containing gallium with an excess of the iron under examination, and thus the gallium is concentrated in the residue T?hich remains un dissolved.;]: It became necsssary to consider what was the source of the gallium contained in the iron. Was the gallium concentrated in the metal ? Or did it pass into the slag of the converter? Was it originally con- tained in the ore, the lime, or the fuel ? Was it easily volatilised, so as to pass off with fume or with flue dust ? * ' Comptes Rendus,' vol. 94, p. 1228. t Loc. cit., p. 1629. | ' Comptes Rendus,' vol. 49, p. 1625. On the Occurrence of Gallium in Clay-ironstone. 399 On February 10th we received from Mr. C. R. Ridsdale, the Chemist, at the North Eastern Co.'s Steel Works, at Middlesbrough, samples of the following materials : — 1. " Mixer metal," i.e., mixed blast-furnace metal. 2. Roasted Cleveland iron ore. 3. Flue dust. 4. Tap cinder. 5. Manganese ore. 6. Lime. On February 12th, photographs of the oxy hydrogen flame spectra of these substances were obtained. The following are the particulars of this examination : — 1. The roasted Cleveland ore contained iron, sodium, potassium, manganese, chromium, nickel, copper, gallium, lead, and calcium. 2. The blast-furnace metal contained iron, sodium, potassium, manganese, nickel, copper, gallium, and lead. 3. Flue dust contained iron, sodium, potassium, manganese, chromium, nickel, copper, silver, gallium (doubtful), lead (strong), calcium, and rubidium. Rubidium was identified by the lines 4202 and 4216. (Thalen.) Calcium by line 4226, in the blue. It is evident now that gallium is contained in the ore and is con- centrated in the metal. 1. The manganese ore (a 15 per cent. Spanish ore) contained iron, sodium, potassium, manganese, copper, silver, lead, indium, and calcium. The lines by which the indium and the silver were identified are as follows : — Indium 4510'2 4101-3 Silver 3383-5 3282-1 The occurrence of indium is remarkable, as hitherto it has been found only in zinc blendes. 2. Tap cinder contained iron, sodium, potassium, manganese, copper, and lead. 3. Lime contained calcium, magnesium, potassium, and sodium, a trace of iron, and a trace of manganese. The lime showed the following bands, characteristic of lime* :— - Band in the orange from 6253 to 6116, degraded towards the more refrangible side. Band from 6075 to about 5900. * ' Phil. Trans.,' vol. 185, p. 182. 400 Prof. W. N. Hartley and Mr. H. Ramage. Very strong band from 5598 to 5485. Band of continuous rays with other bands discernible in it. Less refrangible edge of band 5445. Band in the same at 5422, 5390, 5359, 5341, 5322. The more refrangible edge of band 5304. Very narrow band in the blue, more like a very strong broad line from 4222 to 4215. The magnesium oxide was identified by three bands, more or less connected by diffused rays. 1st. From 3929 to 3856 2nd. „ 3834 „ 3805 3rd. „ 3805 „ 3682 On these bands were seen ten iron lines, six in the first principal group and four in the second, all very faint, but with apparently the following wave-lengths, which correspond wiih the lines seen in oxyhydrogen flame spectrum of ferric oxide. They are also closely in approximation to, and probably identical with, the following arc lines, measured by Kayser and Runge in iron. 3860-03 3856-49 3826-04 3824-58 3758-36 3748-39 3745-67 3737-27 3735-0 3722-69 3720-07 Roasted Cleveland Iron Ore. Process for the Extraction of Gallium. This ore is a complex substance, and contains elements which render the complete extraction of the gallium very difficult. It is in great part soluble in strong hydrochloric acid, but the iron goes into solutions as a ferric salt, and difficulties arise in attempting to reduce it to the ferrous, state. Zinc and iron are both liable to contain gallium, and, without a very careful examination of a quantity of the metal, it would be wrong to use them as reducing agents, seeing that the quantity of metal required in the process is large in com- parison with the sample treated. Sulphurous acid and kindred sub- stances yield sulphates which cause a quantity of the alkaline earths to separate as sulphates, and, as these precipitate in faintly acid solu- tions, there is a risk of basic gallium sulphate being carried down with them. Dilute hydrochloric acid yields a solution poor in iron, but the dis- solved matter is richer in gallium than the original ore. A large proportion of silicic acid is, however, contained in the solution. Experiments were made on quantities of 50 grams of the ore, and the spectra from the sesquioxide metals were carefully compared with the spectra from the similar products from the metal, and we find that, as in the comparison of the original samples of ore and On the Occurrence of Gallium in Clay-ironstone. 401 metal, the gallium lines are decidedly stronger in the spectra of the substances extracted from the metal. One kilo, of finely powdered ore was mixed with dilute hydro- chloric acid of double normal strength, measuring about 1250 c.c. Some carbon dioxide was disengaged and an insoluble residue left .which was removed by filtration. The filtrate was then heated when a gelatinous separation of silica occurred. After evaporation to dryness, a further addition of hydrochloric acid yielded a solution which was not highly coloured, and, presumably, did not contain much iron. The silica rendered insoluble was removed by filtration, and to the filtrate ammonium chloride and ammonia were added. The precipitate thus formed was dissolved in hydrochloric acid, reduced with sulphur dioxide, nearly neutralised, and boiled with sodium thiosulphate. The precipitate was dissolved in hydrochloric acid and again precipitated by ammonia. This precipitate was examined for gallium. The insoluble residue was also examined, and a comparison of the two spectra showed that a larger quantity of gallium remained in the insoluble residue than was extracted by the acid. It was found that gallium could be .extracted from this by fusion with caustic soda and lixiviation with water, and that the residue, after such treatment, contained no gallium. Operations on this particular ore were suspended until other samples had been examined. , The following ores from the collection in the Royal College of Science, Dublin, were examined: — 1. Yorkshire clay ironstone from near Middlesbrough. 2. Clay ironstone from Grosmont, Whitby, Yorkshire. 3. Northamptonshire ore (clay ironstone). 4. Black band ore, Mount Melville mine, St. Andrews. One kilo, of each was reduced to fine powder, and 100 grams of Nos. 1, 2, and 3, and 500 grams of No. 4 were extracted with dilute hydrochloric acid as in the previous case. In each sample gallium was found, but the proportion was very small in the Northampton- shire ore, and still more minute in the black band. Without operat- ing on several hundred grams it would have been scarcely possible to detect the gallium in the Mount Melville ore. These ores had not been roasted, and in this they differed from the sample received from the North Eastern Steel Works. The effect of roasting is the same as increasing the proportion of gallium in the ore. Estimation of Gallium in the Blast Furnace Metal from Middlesbrough. The sample weighing 575 grams consisted of 155 grams of fine powder and 420 grams of coarse powder. The latter portion was heated with hydrochloric acid until the acid was nearly neutralised, when the liquid was decanted and filtered. 402 Prof. W. N. Hartley and Mr. H. Ramage. Residue A. Solution B. The residue A was heated with hydrochloric acid to which a small quantity of nitric acid was added from time to time ; the solution was diluted and filtered. Residue C. Solution D. Residue G. — Dried and heated 0*5 gram in the oxyhydrogen flame. The lines of gallium, chromium, nickel, and iron are strong, and lines of sodium, manganese, potassium, copper, and silver are also present. Solution J3. — Boiled for two hours with part of the finely powdered sample added gradually to neutralise all the free acid, so that the gallium in the solution might be precipitated as a basic salt.* The solution was decanted and filtered. The residue was boiled with solution D, to which the remainder of the finely powdered sample was slowly added ; after boiling for several hours the solution was filtered, and the residue F washed with water. The filtrate was mixed with that from solution J>, the mixture forming solution G, which should be free from gallium. This solution was boiled with freshly precipitated copper hydrate,! and the precipitate examined spectrographically for gallium. It contained none. Residue F. — Boiled with an excess of hydrochloric acid, diluted, filtered, and washed, Residue H. Filtrate I. Residue H. — Dried, powdered, and mixed with residue C. Gentty heated, the mixture decomposes and expels hydrocarbons, causing the mass to ignite and evolve some white fumes. The substance was thus seen to be very inflammable, and the temperature was reduced as quickly as possible. When cold, it was covered with aqua regia and heated on the water bath for several hours, then diluted and filtered. Filtrate added to J, forming solution K. Residue L. Residue L. — A small quantity of it was heated in the oxyhydrogen flame. The gallium line is strong. 45 c.c. of strong sulphuric acid was heated in a porcelain basin until it gave off white fames; the residue was then added forming a pasty mass which was kept hot for about three hours ; white fumes being emitted during the whole time. Water was then added, and the liquid filtered. Filtrate N. Resi- due M. A portion of the latter was heated in the oxyhydrogen flame. The gallium line is still present. Besides the small quantity remaining in the residue M, the gallium should now be in the solutions K and N. Solution K was evaporated nearly to dryness to expel the excess of acid, then diluted, saturated with sulphur dioxide, nearly neutralised with ammonia, and boiled to * ' Comptes Rendus,' vol. 93. p. 818. See also a complete account, ' Separa- tion du Gallium d'avec les autres elements,' par M. Lecocq de Boisbaudran. Paris, Oautbier- Villain. 1884. Reprinted from the ' Annales de Chimie,' 6. Serie, t. 2. f ' Comptes Rendus,' vol. 94, p. 1154. On the Occurrence of Gallium in Clay-ironstone. 403 reduce the iron to the ferrous state. This operation was unsuccessful, a quantity of iron remaining in the ferric state. The solution N was, therefore, added and the mixture evaporated that the moro volatile acids might be expelled by the sulphuric acid. On adding- water to the residue a small quantity of matter remains undissolved ; it was removed by filtration. Residue M?. Up to this stage no reagent had been used which was likely to contain gallium, and we had to consider which of the processes known to separate gallium would be suitable under the conditions of our analysis. The simplest would have been to boil with iron or zinc, but gallium is found associated with both of these metals, and it was decided not to use them. Precipitation by barium carbonate would have been easily effected if sulphuric acid had not been present in such quantity. Bat, to avoid inaccuracy, the best — although more troublesome process — seemed to be the precipitation of the phos- phates of the sesquioxide metals in an acetic acid solution, there being phosphoric acid already in the liquid. The precipitates should contain all the gallium, chromium, and aluminium as phosphates and some phosphate of iron. The gallium is easily separated from chromium and iron by fusion with caustic soda, and from phosphoric acid, aluminium, and chromium by precipitation with potassium ferrocyanide. The iron was first reduced by passing sulphur dioxide into the solution until it became strongly charged, and heating to boiling, with addition of ammonia, to neutralise the excess of free acid. The addition of ammonia was continued until the white precipitate which formed remained undissolved after boiling for two or three minutes. Boiling water was then added to make the volume of the solution about four litres; this dilution caused a large quantity of light coloured precipitate to form. Ammonium acetate was added, and the liquid, after boiling for several minutes, filtered. Residue 0. — The filtrate was boiled and ammonium carbonate added until a quantity of pale, greenish -coloured precipitate was deposited. More ammonium acetate was added, and the liquid, still acid with acetic acid, was filtered. Residue P. The process just described was repeated with the filtrate, the pre- cipitate R being slightly darker than P. Filtrate Q, Small quantities of the three residues, 0, P, R, were examined spectrographicaily. The gallium lines are strongest in R. The filtrate Q was again boiled with addition of ammonium carbonate to neutralise some of the excess of acid, and the precipitate S, small in quantity and of a dark green colour, was removed by nitration. It contained only a trace of gallium. The precipitates 0 and S, containing a much larger proportion of iron than P and R, were dissolved in hydrochloric acid, and the 404 Prof. W. N. Hartley and Mr. H. Ramage. gallium, &c., precipitaied, after reducing the iron to the ferrous state. First precipitate, U. The second contained some gallium ; the third, very dark in colour, was free from that metal. The precipitates P, R, and U were dissolved in hydrochloric acid , and the solutions filtered to remove a small quantity of insoluble matter which was added to residue M. Two drops of violet-coloured filtrate were tested with potassium ferrocjanide, and so marked was the reaction that it was decided to repeat the process of reduction and precipitation to remove as much iron as possible. The first precipitate W contained nearly all the gallium ; the second contained a small quantity, and the third contained none. The first and second precipitations Z7, whose spectra are seen in 1341 and 1343, contain a small quantity of gallium. They were redis- solved, reduced, and boiled with excess of ammonium acetate, and the precipitate collected. A second precipitate was free from gallium. The former was fused with caustic soda, extracted with water and filtered. The filtrate was acidified wihh hydrochloric acid, and boiled with ammonia for some time, and the gallium phosphate thus pre- cipitated was collected. This precipitate was added to W. Residue W, Sfc. — This contained principally gallium and chromium phosphate with some iron phosphate. It was dissolved in hydro- chloric acid, and the solution made to contain about one-fourth its volume of strong hydrochloric acid. Potassium ferrocyanide was added, but not an excess, and the bulky precipitate collected. An excess of the reagent was added to the filtrate, which, after standing twenty-four hours, was filtered. Very small quantities of the two precipitates were examined spectrographically ; the second is decidedly richer in gallium than the first. Residues M and M'z with the small Residues added to them as described. — Ignited at a red heat to burn combustible matter. The mass became grey and weighed, when cold, 8 grams. It was very bulky, and consisted largely of silica. Fusion with fusion mixtures converted the silica into alkaline silicates, which were removed by solution in water, leaving a black residue. This was fused with caustic soda and sufficient nitre to oxidise the graphite, &c. Water dissolved all of this, excepting a small quantity of red oxide of iron, part of which was examined for gallium. None present. The filtrate was acidified with hydrochloric acid, evaporated to dry ness, and dried at 120° C. to dehydrate silicic acid. The dry residue was digested with strong hydrochloric acid, and water added. It was then filtered to remove some silica, which was found to have retained only a trace of gallium. The filtrate was mixed with a small excess of ammonia, and boiled •for some time; the gallium being precipitated probably as phosphate. The filtrate in this and in all similar cases was again boiled, after On the Occurrence of Gallium in Clay-ironstone. 405 adding a few drops of ammonia; if any precipitate was produced it. was collected and added to the other precipitate. The precipitate in this case was added to ferrocyanide precipitates obtained from the residue W. The paper, after being scraped to remove the residue as far as possible, was burnt in the oxyhydrogen flame. The gallium lines are strong. The ferrocyanide precipitates with others rich in gallium were ignited at low redness to decompose the cyanides, and then fused with pure caustic soda. The produce was extracted with water and filtered. Residue from Fusion. — Dissolved in hydrochloric acid, expelled the excess of acid, added water, reduced the ferric salt, and filtered. Residue remaining contained only a trace of gallium. Filtrate. — Boiled with an excess of ammonium acetate and filtered off the precipitate. The filtrate was mixed with sodium phosphate and boiled, thus yielding a second precipitate. The filtrate from this was again boiled, and ammonium carbonate added until a third precipitate was produced. Very small portions of these three precipitates were burnt in the oxyhydrogen flame. The first two were rich in gallium, while the third contained only a trace. Ignited the first and second precipitates, heated the residue in a platinum crucible with hydro- chloric and sulphuric acids, expelled the former acid by heating until the white fumes of sulphuric acid were evolved, and then fused the residue with caustic soda. Extracted with water and filtered. After a second fusion the residue was practically free from gallium. The alkaline filtrates were acidified with hydrochloric acid, and the gallium precipitated by boiling with ammonia until the excess ot ammonia was expelled. Filtered and tested the filtrate by repeating the process of boiling with ammonia until no further precipitate resulted. The precipitates of gallium hydrate and phosphate, obtained as described, were dissolved in hydrochloric acid and, after adding one- fourth the volume of the solution of strong hydrochloric acid, an excess of potassium ferrocyanide was added. After standing for one day the precipitate was collected, washed, and ignited. It weighed 0-0670 gram. This residue was dissolved by heating with strong sulphuric acid in a platinum crucible, some water being added, after heating strongly, and then an excess of caustic soda prepared from sodium. The crucible was then heated until the water was expelled, and the residue retained in the fused caustic soda. The process was repeated on the residue which remained after adding water and filtering. The second residue was practically free from gallium. The filtrates were collected in a platinum basin, made faintly acid with hydrochloric acid, and saturated with sulphuretted hydrogen. 406 Prof. W. N. Hartley and Mr. H. Ranmge. A brownish coloured precipitate was removed by filtration. It con- tained copper, lead, and silver, but no gallium. The nitrate was boiled to expel sulphuretted hydrogen, and the gallium precipitated with ammonia as described above. The precipitate was collected and ignited. It weighed 0*0288 gram. This residue possessed a very light yellow colour. One milligram was burnt in the oxyhydrogen flame ; its spectrum shows the two gallium lines very strongly. Lines of sodium, potassium, iron, calcium, and lead are present, but those of the last three are exceedingly weak. The remaining 0*0278 gram of residue was fused with hydrogen potassium sulphate ; water and sulphuric acid were added, and the crucible heated until fumes of sulphuric acid were evolved. Water was again added, and a small residue removed by filtration. This residue weighed 0*0040 gram. The gallium was separated from the iron by two extractions with caustic soda solution. The ferric hydrate was dissolved in. hydro- chloric acid, and reprecipitated by ammonia. The ferric oxide weighed 0*0022 gram. The gallium in the nitrate was then reprecipitated and weighed as oxide. It weighed 0*0213 gram. A few drops of sodium phosphate were added to the filtrate, and sufficient ammonia to make it turn red litmus paper blue. After boiling for a few minutes the liquid was filtered, the paper being dried and burnt in the oxyhydrogen flame. The gallium lines are present in its spectrum, but are very weak. The oxide of gallium now possessed a scarcely perceptible, faint yellow colour. It does not. represent the whole of the gallium present in the sample, as a small quantity was removed and lost in testing the precipitations and residues. We are able, however, to estimate this quantity by comparing the lines in the different spectra with lines and spectra obtained by heating weighed quantities of gallium oxide. In this way we estimate the total quantity of gallium to be as follows : — Pure oxide _____ ........ ....... . 0*0213 gram. In 0*001 gram of impure oxide ---- 0*0008 „ In residue insoluble in HKS04 ---- 0*0004 „ In other substances ........... . ', . 0*001 „ Gas03, total ........ 0'0235 gram. 0*0235 gram of pure Ga203 contains 0*0175 gram of gallium, equal to °'°175 * 1Q° = 0*00304 per cent. 575 One part of gallium is contained in 33,000 parts of ornde iron. On the Occurrence of Gallium in Clay-ironstone. 407 An estimation of the gallium in the " mixer metal " had been at- tempted in the spring of this year, but the separation was not as complete as in the process just described. The figure obtained, however, is so closely in accord with the above that we will briefly describe the process and record the result. The sample, weighing 340 grams, was boiled with hydrochloric acid until the latter was nearly neutralised ; the solution was then decanted, and fresh acid added to the residue. When the acid ceased to have any marked action the whole liquid was filtered, and the residue A washed, dried, and treated separately for the separation of gallium. Filtrate B. — From this filtrate gallium was precipitated by calcium carbonate, but phosphates and sesquioxide metals, including chromium, rendered the precipitate a too complex mixture, and we had recourse to the ferrocyanide method. The gallium was separated from the iron by pure sodium hydrate, and finally precipitated as hydrate and ignited. The oxide weighed 0'0149 gram, and this amount corresponds to 0'0033 per cent, of gallium in the sample or one part in 30,000 of the iron. We know, by the spectrographic examination of the residues, &c., that the whole of the gallium was not obtained, and that the oxide weighed was not quite pure gallium oxide, but with the experience gained in this estimation we were able to make the more exact analysis already described. In conclusion, we may state that this blast furnace metal contains more gallium than the richest source of that element hitherto known. The mineral referred to is a zinc blende from Bensburg on the Rhine, about eight miles from Cologne ; it is found in the Franzisca adit of the Luderich mine. MM. Lecocq de Boisbaudran and Jungfleisch extracted 62 grams of crude gallium from 4300 kilograms, or nearly 4^ tons of the ore ; this is in the proportion of 1 in 72,000, but they believed the actual quantity present to be about 1 part of gallium in 50,000 of the ore. We have recently discovered other sources of gallium, bat cannot include the details of our later work in the present communication. 408 Prof. C. S. Sherrington. Examination of the Peripheral January 21, 1897. Sir JOHN EVANS, K.C.B., D.C.L., LL.D., Vice- President and Treasurer, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. The Right Hon. Sir John Eldon Gorst, a member of Her Majesty's Most Honourable Privy Council, was admitted into the Society. The following Papers were read : — I. " On Cheirostrobus, a new Type of Fossil Cone from the Cnlci- ferous Sandstones." By D. H. SCOTT, M.A,, Ph.D., F.R.S. II. " Experiments in Examination of the Peripheral Distribution of the Fibres of the Posterior Roots of some Spinal Nerves. Part II." By C. S. SHEEEINGTON, M.D., F.R.S., Holt Pro- fessor of Physiology, University College, Liverpool. III. " Cataleptoid Reflexes in the Monkey." By C. S. SHERRJNGTOX, M.D., F.R.S., Holt Professor of Physiology, University College, Liverpool. IV. " On Reciprocal Innervation of Antagonistic Muscles. Third Note." By C. S. SHERRINGTON, M.D., F.R.S., Holt Professor of Physiology, University College, Liverpool. " Experiments in Examination of the Peripheral Distribution of the Fibres of the Posterior Roots of some Spinal Nerves, Part II." By C. S. SHERRINGTON, F.R.S., Holt Professor of Physiology, University College, Liverpool. Received November 12, 1896,— Read January 21, 1897. (Abstract.) This paper is in continuation of one brought before the Society in 1892, and published in 'Phil. Trans.,' B, vol. 184. In that commu- nication the peripheral distribution of the sensory nerve-roots of the sacro-lumbar and the thoracic regions was examined. In the present the examination is extended to the cervical and brachial sensory Distribution of Fibres of Posterior Roots of Spinal Nerves. 409 roots, and to the skin distribution of the cranial nerves. The com- munication is divided into four sections. In Section I the field of peripheral distribution of each root is described from the Yth cervical to the lower end of the brachial region. The description given is taken in each case from one particular experiment, which has proved a typical one for the root in question, and then deviations from this type are appended to it in the form of annotations. Particular attention was paid to the question of the skin-fields of the several divisions, ophthalmic, maxillary, and inandibular of the cranial Vth, in order to see if the fields possessed the characters of segmental skin-fields, or those of peripheral nerve-trunk skin-fields. They were found to conform with the latter, not with the former. A curious relation of the posterior edge of the field of the Vth to the external ear is found to exist, indicating that the position of the visceral cleft is still adhered to as a boundary line for the field of the trigeminus. The sense of taste as well as of touch is foand to be destroyed in the anterior two-thirds of the tongue after intracranial section of the Vth ; this makes it extremely doubtful whether the corda tympani can have gustatory functions in the monkey, as has been believed in some cases in man. No loss of eye-movements, or inter- ference with them, has been found to result from intracranial section of the Vth. The results obtained on the various successive nerve-roots cannot well be abstracted. The glossopharyngeal field on the tongue has been successfully delimited. After cranial Vth and all the upper cervical posterior roots have been severed, there still persists a small field of sentient skin, which includes the external auditory meatus and a part of the pinna. This field, although not corresponding to the situation given by anthro- potomists to the distribution of the auricular branch of the vagus*, comes either from it or the glossopharyngeal. It presents interest as being the only field representing the whole cutaneous distribution of an entire nerve, which does not conform with the rules of zonal distribution holding good in the case of each of the other nerve-roots examined, and these now include the whole series. The posterior root of the 1st cervical nerve has a skin-field in the cat which includes the pinna. The posterior root of the same nerve in Macacus has no skin-field at all, its skin-field having apparently been included in the Ilnd cervical of Macacus, not in the cranial Vth. The root fields con- tributing to the surface of the brachial limb are Illrd, IVth, Vth, Vlth, Vllth, and VHIth cervical, and 1st, Ilnd, and Illrd thoracic. Of these, the VIII th cervical is the only one which includes the whole of the surface of the free apex of the limb ; its distribution in this respecb closely resembles that of the Vlth lumbar sensory root in the pelvic limb. VOL. LX. 2 I 410 Distribution of Fibres of Posterior Roots of Spinal Nerves. The Ilnd section of the communication deals with the degree of conformity between the distribution of the spinal ganglion fibres in the skin and their distribution in the underlying deep tissues of the limb. It is shown that, although the skin fields of the ganglia are in the middle of the limb region dislocated from the median line of the body, the distribution of the fibres of the root ganglion is never- theless, when its deep distribution is taken into account, to a com- plete ray of tissue extending in an unbroken fashion from the median plane of the body oat along the limb to (in the case of the nerves, extending farthest into the limb) the very apex of it. This distribu- tion conforms, therefore, with that shown in a previous paper to be typical of the distribution of the ventral (motor) root. The distinc- tion is not, therefore, as between afferent and efferent, but as between cutaneous and muscular. A detailed analysis of the distribution of the deep sensory fibres is in this paper carried out for the YIth lumbar spinal ganglion of Macacus rhesus ; this ganglion was chosen because its skin-field, occupying the free apex of the lower limb, is one as far dislocated from the median line of the body as any in the whole spinal series, and presents, therefore, the greatest apparent discrepancy between the distribution of its afferent and efferent roots. A com- parison of the distribution of the afferent and efferent roots in this (Vlth lumbar) nerve was made by means of the Wallerian method ; the results show the peripheral distribution of the two to be minutely similar. From this, and from other observations given, the rule is put forward as a definitely established one that the sensory nerves of a skeletal muscle in all cases derive from the spinal ganglion (or ganglia) corresponding segmentally with that (or those) containing the motor cells, whence issue motor nerve-fibres to the muscle. The reflex arc, in which the afferent and efferent nerve-cells innervating a muscle are components, need not, therefore, as far as anatomical composition is concerned, involve irradiation through more than a single spinal segment. Section III deals with general features of arrangement recognisable in the distribution of the roots ; for instance, the determination of the position of the primary dorsal and ventral lines of the limbs, the examination of the asserted rotation of the limbs and of the asserted torsion of the limbs, and of the asserted homologies between muscles, &c., of the brachial and pelvic limbs respectively, by the criteria for re- examination of such questions provided by the facts elicited in the course of the work ; the cross-lapping of the skin- fields across the median line of the body, the overlapping of com- ponent parts of a single field, the serial overlapping of adjacent fields, the degree of overlapping in different regions of the body, the degree of overlapping in peripheral nerve-trunk fields, the amount of over- lapping of spinal ganglion-fields compared with that of peripheral Cataleptoid Reflexes in the Monkey. 411 nerve-trunks, the comparison of sensory overlapping with motor overlapping, the relation of overlapping to acuteness of sensation : individual variation, its extent and frequency, as far as can be judged from the skin-fields. Comparison between the human bracbial plexus and that of Macacus is made, and it is pointed out that the human plexus is slightly prefixed, as compared with that of Macacns. Finally, in Section IV, " shock," and various spinal reactions are examined, especially with reference to their effects upon the size and other features of the areas of the root-fields, &c., and the results collated and discussed. " Cataleptoid Reflexes in the Monkey." By C. S. SHERRINGTON, M.A., M.D., F.R.S., Holt Professor of Physiology, University College, Liverpool. Received December 29, 1896, — Read January 21, 1897. A phenomenon came under my observation in the course of experi- ments upon monkeys at the commencement of the present year which seems sufficiently interesting to merit record here. Its occur- rence, so long as certain conditions of experiment are maintained, appears regular and predictable. Although the character of the movements executed by the skeletal muscles when excited reflexly through the medium of the isolated spinal cord is variable, one feature common to them is their compara- tive brevity of duration. Many of them are, as pointed out by Fick and by Wundt years ago, hardly distinguishable in several particu- lars from the simple twitches elicitable from an excised muscle, so brief and local and inco-ordinate do they appear to be. Others are more prolonged, and, as I have described in a paper recently com- municated to the Society, exhibit various forms of sequence or "march" (Hughlings Jackson). Without recapitulating the con- clusions there drawn from the data given in that paper, I wish here to merely point out that of movements due to purely spinal reflex action, although some are fairly extensive, most are quite short- lasting, and not so prolonged as the longer of those that can be elicited under appropriate conditions from the cortex cerebri ; also that if prolonged they, like the final phase of prolonged movements initiated from the cortex, tend to become clonic, or to exhibit that kind of action which in the paper referred to above I have desig- nated "alternating." The reflex movements, the subject proper of this note, are, on the contrary, of extremely prolonged duration, and absolutely devoid of clonic character and of alternating character. If the cerebral hemi- 412 Prof. C. S. Sherrington. spheres bo carefully removed, e.g., from a monkey, with avoidance of haemorrhage and of fall of body temperature, and if sufficient time be allowed to elapse for subsidence in the animal of what may be called immediate shock, movements can be evoked remarkably different from those I have ever seen elicitable as purely spinal or as cerebral reactions. If a finger of one of the monkey's hands bo stimulated, for instance, by dipping it into a cup of hot water, there results an extensive reflex reaction involving movement of the whole upper limb. The wrist is extended, the elbow flexed, the shoulder protracted, the upper arm being drawn forward and somewhat across the chest. The movement occurs after a variable and usually prolonged period of latent excitation. The movement, although it may be fairly rapid, strikes the observer each time as perfectly deliberate ; it is of curiously steady and "smooth " performance. Sometimes it is carried out quite slowly, and then, as a rule, the extent of it is less ample. The most striking feature of the reflex is, however, that when the actual movement has been accomplished the contraction of the muscles employed in it does not cease or become superseded ly the action of another group, but is continued even for ten and twenty minutes at a time. The new attitude assumed by the limb is maintained, and that too without clonus or even tremor. In the instance cited, namely, that of the fore limb, the posture assumed suggests the taking of a forward step in quadrupedal pro- gression, and in that posture the animal will remain for a quarter of an hour at a time. The degree of, for instance, flexion assumed in the new posture seems much dependent on the intensity and duration of the stimulus applied. If the degree is extreme, the attitude of the limb may not be maintained to its full extent for the time mentioned ; thus, the • elbow, at first fully flexed, will in the course of a minute or so be found to have opened somewhat. This opening can be often seen to • occur per saltum, as it were, but the steps are quite small, and recur- rent at unequal intervals of between perhaps a quarter of a minute and a minute. After some relaxation from the extreme phase of the posture has taken place, the less pronounced attitude, e.^.,semiflexion at the elbow, may persist without alteration obvious to inspection for ten minutes or more. Apart from the occasional step-like relaxations, the contraction of the muscles is so steady as to give an even line when registered by the myograph. A renewed stimulation of the finger excites further flexion, which is maintained as before in the way above described. The posture can be set aside without diffi- culty by taking hold of the limb and unbending it ; the resistance felt in the process of so doing is slight ; the posture thus broken down is not reassumed when the limb is then released. Analogous results aro obtainable on the hind limb. Hot water Cataleptoid Reflexes in the Monkey. 413 applied to a toe evokes always, so far as I have seen, flexion of ankle and knee ; usually of hip also. This movement is " deliberately " executed, and always institutes a maintained posture. If finger (or toe) of both right and left limb be placed together in the hot water, there results symmetrical reflex movement of both the right and the left fore limbs (or hind limbs), leading to assumption of a fairly symmetrical posture by the right and left limbs respec- tively, the posture being similar to but duplicate of that evoked in the one limb only on excitation of that limb. This may appear a self-evident sequel to the observation given earlier, but is not so when an observation immediately to be mentioned is taken into con- sideration. Not the least interesting part of the reflexes under consideration is a remarkable glimpse which they allow into the scope of reflex inhibition as regards the co-ordinate of movements of the limbs. Although the posture taken up by the right fore limb consequent upon excitation of a finger is symmetrically duplicated by the left limb when both hands are simultaneously stimulated, the effect of excitation of the two hands does not lead to symmetrical posture if the excitation be not synchronous but successive. If when the right arm has already assumed its posture in response to an excitation of the right hand, the left hand be stimulated, there results, while the left arm in obedience to the excitation is lifted and placed in the flexed posture, an immediate and, if the stimulus be at all more than slight, complete relaxation of the right arm. The right arm drops flaccid while the left is raised and maintained in the raised attitude. Similarly, excitation of the right foot breaks down the posture assumed by the right arm, and conversely, and even more easily, stimulation of the right hand breaks down a posture assumed by the right leg. Again, a nip of the right pinna causes relinquishment of a posture assumed by the righb arm or by the right leg. If the right pinna is pinched when both arms are in this cataleptoid posture, com- plete inhibition can be readily exerted on the right arm, but usually only partial relinquishment can be induced in the left arm. To exert complete inhibition upon the posture of the left arm, the pinna pinched must be that of the left side. Similarly the posture reflexly evoked by appropriate stimulation of either hind limb can be inhibited by excitation of either pinna or of either fore limb, but predominantly by pinna and fore limb of the same side as the limb to be inhibited. The inhibition of the hind limb is much more easily elicited from the opposite hind limb than from the opposite fore limb or opposite ear. I have never yet seen it obtained diagonally upon the fore limb from the opposite hind limb. The movements obtained in the limbs by exciting the limbs them- selves are only cited above as examples to illustrate the general 414 Prof. C. S. -Sheningtou. characters of the condition. The details of the results will be given in a fuller paper dealing with the subject. I was prevented from, inquiring thoroughly into the phenomenon when it was first met with ; but in the course of the present summer and autumn the investigation has been systematically undertaken. I will conclude this preliminary note by adding that throughout the observations the animal's respiration remains apparently unaffected by the stimuli effective to produce the various reflexes and inhibitions such as above described. The respiration is tranquil, rather deep, regular, and often somewhat frequent. The animal in all niy experiments has been completely blind, but a sharp conjunctival reflex exists. The knee jerks are elicitable but are not exaggerated. Thetonus of the sphinc- ters appears about normal. The pulse is full, regular, and fairly frequent. I have not at present succeeded in evoking the cataleptoid reflex by simply placing the limb in the desired posture. In applying the term cataleptoid to these reflexes, I do so because the reflexes recall, in some respects, strikingly certain phases of hypnotic condition, by some writers distinguished as cataleptic, and because the strict significance of the prefix implies a steady main- tenance of possession subsequent to seizure, and is therefore peculiarly applicable here, whether these reflexes be or be not allied to hypnotic catalepsy. " On Reciprocal Innervation of Antagonistic Muscles. Third Note." By 0. S. SHERRIXGTON, M.A., M.D., F.R.S., Holt Professor of Physiology, University College, Liverpool. Received December 29, 1896,— Read January 21, 1897. In a former number* of these c Proceedings ' attention was drawn to a particular form of correlation existing between the activity of antagonistic muscles. In it, one muscle of an antagonistic couple is, it was shown, relaxed in accompaniment with active contraction of its mechanical opponent. The instance then cited was afforded by certain of the extrinsic muscles of the eyeball, but I had previously noted indications of a like arrangement in studying the reflex actions affecting the muscles at the ankle-joint of the frog,f and it seemed probable that the kind of co-operative co-ordination demonstrated for the ocular muscles might be of extended application and occurrent in various motile regions of the body. The observations to be men- tioned below do actually extend this kind of reciprocal innervation * Vol. 52. April, 1893. Sherrington. t Foster's ' Journ. of Phyeiol.,' TO!. 13, 1892. On Reciprocal Liner vation of Antagonistic Muscles. 415 to the muscles of antagonistic position acting about certain joints of the limbs. If transection of the neural axis be carried out at the level of the crura cerebri in, e.g., the cat, there usually ensues after a somewhat variable interval of time a tonic rigidity in certain groups of skeletal muscles, especially in those of the dorsal aspect of the neck and tail and of the extensor surfaces of the limbs. The details of this condi- tion, although of some interest, it is unnecessary fco describe here and now, except in so far as the extensors of the elbow and the knee are concerned. These latter affect the present subject. The extensors of the elbow and the knee are generally in strong contraction, but alto- gether without tremor and with no marked relaxations or exacerba- tions. On taking hold of the limbs and attempting to forcibly flex the elbow or knee a very considerable degree indeed of resistance is experienced, the triceps brachii and quadriceps extensor cruris become, under the stretch which the more or less effectual flexion puts upon them, still tenser than before, and on releasing the limb the joints spring back forthwith to their previous attitude of full exten- sion. Despite, however, this powerful extensor rigidity, flexion of the elbow may be at once obtained with perfect facility by simply stimulating the toes or pad of the fore foot. When this is done the triceps enters into relaxation and the biceps passes into contraction. If, when the reflex is evolved, the condition of the triceps muscle is carefully examined, its contraction is found to undergo inhibition, and its tenseness to be broken down synchronously with and indeed very often accurately at the very moment of onset of reflex contraction in the opponent prebrachial muscles. The guidance of the flexion movement of the forearm may therefore be likened to that used in driving a pair of horses under harness. The reaction can be initiated in more ways than one, electrical excitation of a digital nerve or mechanical excitation of the sensory root of any of the upper cervical nerves may be employed ; I have seen on one occasion a rubbing of the skin of the cheek of the same side effective. Similarly in the case of the hind limb. The extensor muscles of the knee exhibit strong steady Don-tremulent contraction under the appropriate conditions of experiment. Passive flexion of the knee can only be performed with use of very considerable force, the quad- riceps becoming tight as a stretched string. The application of hot water to the hind foot then elicits, nevertheless, an immediate flexion at knee and hip, during which not only are the flexors of those joints thrown into contraction, but the extensors of the knee joint are simultaneously relaxed. Electric excitation of a digital nerve or of the internal saphenous nerve anywhere along its course will also initiate the reflex. The same relaxation of existing contraction in the extensors can 416 On Reciprocal Innervation of Antagonistic Muscles. be obtained by electrical excitation of the tract in the crura cerel-ri, when, as sometimes happens, that excitation evokes flexion at elbow or at knee. This and the previous fact which evidences that the result is obtainable after complete removal of the whole cerebrum bear out the view arrived at in my former paper that for this reciprocal and, as I believe, elementary co-ordination, it is not essen- tial that " high level " centres (Hughlings Jackson) be employed. I incline to think, however, that this kind of co-ordination at elbow and knee is probably largely made use of in movements initiated via the cerebral hemispheres as well as in the lower reflexes, on the observation of which the present Note is based. This conclusion is indicated by its occurring in response to excitation of the pyramidal fibres in the crura. In the case of the reciprocal in nervation of antagonistic ocular muscles I was able to prove that it took place even in " willed movements." It seems, in view of what has been shown above, legitimate to extend that result to the additional examples afforded by elbow-joint and knee. Regarding the innervation of the triceps brachii and quadriceps extensor cruris, it is interesting to note that these muscles, which are of all among the limb muscles particularly difficult to provoke to action by local spinal reflexes, are the very ones which, when the level of the transection is pontial or prepontial, exhibit tonic contraction the most markedly. The well-known and oft-corroborated Sanders-Ezn phenomenon of inaccessibility of the extensors of the knee to spinal reflex action has, as I have recently shown, certain limitations, but at the same time so long as the transection is spinal — even when carried out so as to isolate not merely a portion of, but the whole, spinal cord entire from bulb to filurn terminale — does apply very strictly to excitations arising in its own local region proper. And the spinal reflex relations of the triceps brachii in this respect, as pointed out elsewhere, somewhat resemble those of the distal portion of the quadriceps extensor of the leg. Alteration of the site of tran- section from infrabulbar to suprabulbar levels works a curious change in this. The Sanders-Ezn phenomenon then becomes subject to strik- ing contravention. I have, after the higher transection, several times seen excitation of the hind foot itself provoke unilateral ideolateral extension of knee, a result incompatible with the Sanders-Ezn rule even under the limitations of ideolaterality, &c., which I consider must be attached to it. And similarly with the triceps at the elbow. The difference between the accessibility of the quadriceps to reflex action after infrabulbar and after suprabulbar transection may, how- ever, be less abrupt than it appears at first sight, and a superficial rather than a fundamental distinction. When extensor rigidity has ensued at elbow and knee after suprabulbar transection, the reflex excitability of triceps brachii and quadriceps cruris seems in a man- On Cheirostrobus, a new Type of Fossil Cone. 417 ner as difficult as in the presence of exclusively spinal mechanisms. The reflex inhibitions the subject of this Note show, however, that the accessibility is not really greatly or even at all altered ; the nexus is maintained, but the conduction across it is signalised by a different sign, minus instead of plus. The former, to find expression, must predicate an already existent quantity of contraction — tonus, to take effect upon. It seems likely enough that even when the transection is infrabulbar and merely spinal mechanisms remain in force, the same nexus obtains, but that then that background of tonic contraction is lacking, and that lacking the play of inhibitions remains invisible, never coming within the field of any ordinary method of observation. Under the conditions adopted in my experiments, various other reflex actions, that seem probably examples of this same kind of co- ordination, can be studied, for instance, a sudden depression arid curving downward of the stiffly elevated and tonically up-curved tail which can be elicited by a touch upon the perineum. But with these and also with other details regarding the reflexes at elbow and knee T hope to deal more fully in a paper to which the experiments re- corded here are contributory. " On Cheirostrobus, a new Type of Fossil Cone from the Calci- ferous Sandstone." By D. H. SCOTT, M.A., Ph.D., F.R.S., Hon. Keeper of the Jodrell Laboratory, Royal Gardens, Kew. Received December 29, 1896 — Read January 21, 1897. The Peduncle. The first indication of the existence of the remarkable type of fructification about to be described, was afforded by the study of a specimen in the Williamson collection, from the well-known fossili- ferous deposit at Pettycur, near Burntisland, belonging to the Calci- ferous Sandstone Series at the base of the Carboniferous formation. This specimen is a fragment of stem, of which seven sections are pre- served in the collection.* Its discoverer thought it might possibly belong to the Lepidostrobus found in the same bed. " If so," he ad.ds, " it has been part of the axis of a somewhat larger strobilus than those described." f A detailed examination of the structure of this specimen convinced me that it is essentially different from any Lepidodendroid axis, and is, certainly, anew type of stem.J * The cabinet-numbers are 539 — 545. t Williamson, " Organisation of the Fossil Plants of the Coal-measures." Part III. ' Phil. Trans.,' 1872, p. 297. J A short account of this specimen was given by me before the Botanical Section of the British Association at the Liverpool meeting, 1896. 418 Dr. D. H. Scott. On Cheirostrobus, a new Type As it was the examination of this fragment of stem which first },ut me on to the track of the new cone, it may be well shortly to describe its chief characteristics, reserving all details for a future paper. The specimen, which is about 7 mm. in diameter, bears the bases only of somewhat crowded leaves, the arrangement of which, though not quite clear, was most probably verticillate, with from nine to twelve leaves in a whorl, those of successive whorls being superposed. Each leaf-base consists of a superior and an inferior lobe, and each lobe is palmately subdivided into two or three segments. The leaf-traces, which are single bundles where they leave the central cylinder, subdivide in both planes on their way through the cortex, to supply the lobes and segments of the leaf. The central cylinder is polyarch, the strand of wood having from nine to twelve prominent angles, with phloem occupying the furrows between them. With the exception of the spiral protoxylem-elements at the angles, the tracheae have multiseriate bordered pits, thus differ- ing conspicuously from the scalariform tracheae of the Lepidodendrere. The interior of the stele is occupied by tracheas intermingled with conjunctive parenchyma. There is a well-marked formation of secondary tissues by means of a normal cambium.* The Strolilus. Mr. B. Kidston, F.Gr.S., kindly informed me that he had in his possession sections of a fossil cone from Burntisland having certain points in common with the Williamson specimen. On inspecting these sections with Mr. Kidston I was soon convinced that this uude- scribed cone really belonged to the same plant as the fragment of stem in the Williamson collection, and that the latter might well be the peduncle of the former. At the same time, I satisfied myself, and Mr. Kidston agreed with me, that the whole organisation of his cone is fundamentally different from that of any Lepidostrobus, the deci- sive point being that the new cone has compound branched sporo- * The general structure of this axis, including the course of the bundles and the subdivision of the bracts, is correctly described by Williamson, loc. cit., p. 297. -As regards the latter point, he says " peripherally the bark breaks up into main or primary bracts, which again subdivide, as in the transverse section, into secondary ones, demonstrating that each primary bract does not merely dichotomize, but sub- divides, both horizontally and vertically, into a cluster of bracts — a condition corre- sponding with what T have already observed in the smaller strobili described." These smaller strobili are those of the Burntisland Lepidostrobus, to which, by a strange coincidence, Williamson, loc. cit., p. 295, erroneously attributed the same character, as regards subdivision of the bracts, which actually exists in the new cone. The only explanation appears to be, that Williamson interpreted the strucbure of the Lepidoatrobus in the light of that of the peduncle, which, as we shall see, really belonged to a totally different fructification. of Fossil Cone from the Calciferous Sandstone. 4J9 phylls, eauli of which bears a number of sporangia. Ifc became evident that this cone must be placed in a new genus, and the con- clusion arrived at from the study of the peduncle was thus confirmed. Mr. Kidston most generously handed over his sections to me for examination and description, and also obtained for me from the owner the remains of the original block, from which I have had a number of additional sections prepared. Only a single specimen of the cone is at present known. Before cutting sections, the piece, which includes the base but not the apex of the sfcrobilus, was about 2 inches long. It was found at Pettycur, near Burntisland, in 1883, by Mr, James Benuie of Edinburgh. The specimen is calcified, and its preservation is remarkably perfect, so that the whole structure is well shown, though the complexity of its organisation renders the interpretation in some respects difficult. The cone in its present somewhat flattened condition measures about 5 cm. by 2'3 cm. in diameter. The diameter in its natural state would have been at least 3'5 cm. That of the axis is about 7 mm., exactly the same as that of Williamson's peduncle. Thus the extreme length of the sporophylls, which have on the whole an approxi- mately horizontal course, is about l-4 cm. The sporophylls are arranged in somewhat crowded verticils, fourteen of which were counted in a length of an inch, 2'5 cm. There are twelve leaves in each whorl, and the members of successive whorls are accurately superposed, a fact which is shown with the greatest clearness in tangential sections of the cone. This is evi- dently a point of great significance in considering the affinities of the fossil. The sporophylls themselves have a remarkably complex form. Each sporophyll at its insertion on the axis, consists of a short basal portion or phyllopodium ; the bases of the sporophylls belonging to the same verticil are coherent. The sporophyll branches immediately above its base, dividing into a superior and an inferior lobe, which lie directly one above the other in the same radial plane. Almost at the same point, each of the lobes subdivides in a palmate manner into three segments, which assume a horizontal course, whereas the com- mon phyllopodium has an upward inclination. It is probable that sometimes, especially at the base of the cone, there may be two instead of three segments to each lobe. As a rule, however, each sporophyll consists of six segments, of which three belong to the superior (ventral or posterior) and three to the inferior (dorsal or anterior) lobe. The segments are of two kinds — sterile and fertile. Both alike consist of a long, straight, slender pedicel, running out horizontally, and terminating at the distal end in a thick laminar expansion. The sterile segments are the longer, and in each the lamina bears an 420 Dr. D. H. Scott. On Oheirostrobus, a new Type upturned foliaceous scale as well as a shorter and stouter downward prolongation. Each of the fertile segments ends in a fleshy laminar enlargement not unlike the peltate scale of an Equisetum or a Galamostachys. These fertile laminae, which are protected on the exterior by the overlapping ends of the sterile segments, bear the sporangia. Four, perhaps in some cases five, sporangia are attached, by their ends remote from the axis, to the inner surface of the peltate fertile lamina. Each sporangium is connected with the lamina by a somewhat narrow neck of tissue into which a vascular bundle enters. The sporangia are of great length, and extend back along the pedicels until they nearly or quite reach the axis. The sterile and fertile segments alternate regularly, one above the other, in the same vertical series. So much is evident, but the ques- tion which segments are fertile and which sterile, has presented great difficulties, owing to the fact that the same segment can scarcely ever be traced continuously throughout the whole of its long course, and that the pedicels of sterile and fertile segments present no constant distinctive characters. For reasons, however, which will be fully given in a subsequent paper, I think it highly probable that in each sporophyll the segments of the lower lobe are sterile, and those of the upper lobe fertile, constituting the sporangiophores. The sporangia and pedicels are all packed closely together so as to form a continuous mass. The external surface of the cone was com- pletely protected by its double investiture of fertile and sterile laminae. The spores are well preserved in various parts of the cone, and, so far as this specimen shows, are all of one kind, their average dia- meter being 0'065 mm. At the base of the cone, where macrospores, if they existed, might naturally be looked for, the spores are of the same size as elsawhere. So far, then, there is no evidence of hetero- spory. The spores are considerably larger than the microspores of the Lepidostrobi. Those of the Burntislaud Lepidostrobus, for example, are barely 0*02 mm. in diameter. The spores of our plant approach in size those of Sphenophyllum Dawsoni, or the microspores of C 'alamo stacliys Gaslieana. The sporangial wall, as preserved, is only one cell in thickness ; it bears no resemblance to the palisade-like layer which forms the wall of the sporangium in Lepidostrobus, but has the same structure as that of a C alamo stachys* The sporangial wall of Sphenophyllum Dawsoni is similar. The anatomy of the axis of the cone agrees closely with that of * See Weiss, " Steinkohleu-Calamarien," yol. 2, 1884, Plate XXIV, figs. 3, 4, and 5 ; Williamson and Scott, " Further Observations on the Organisation of the Fossil Plants of the Coal-measures," Part I, ' Phil. Trans./ 1894, PL 81, fig. 31. of Fossil Cone from the Calciferous Sandstone. 421 the peduncle above described, except for the absence of any secondary tissues. The wood has twelve prominent angles, at which the spiral tracheae are situated, so its development was, no doubt, centripetal. The inner tracheae have pitted walls, and are intermixed with scat- tered parenchymatous cells, imperfectly preserved. The phloem has entirely perished. The most interesting anatomical feature is the course of the leaf- trace bundles, which can be followed with the greatest exactness on comparing sections in the three directions. A single vascular bundle starts from each angle of the stele for each sporophyll, and passes obliquely upwards. When less than half way through the cortex, the trace divides into three bundles, one median and two lateral. The lateral strands are not always both given off exactly at the same point. A little farther out, the median bundle divides into two, which in this case lie in the same radial plane, so that one is anterior, and the other posterior. The median posterior bundle is the larger, and before leaving the cortex this, in its turn, divides into three. There are now six branches of the original leaf-trace, three anterior, and three posterior, which respec- tively supply the lower and upper lobes of the sporophyll. The three segments of the lower lobe are supplied by the two lateral bundles first given off, and by the anterior median bundle, while the upper segments receive the posterior median bundle and its two lateral branches. In the base of the sporophyll, all six bundles can be clearly seen, in tangential sections of the cone, three above and three below. As the segments become free, one bundle passes into each, and runs right through the pedicel to the lamina. In the fertile lamina the bundle subdivides, a branch diverging to the point of insertion of each sporangium. One of the longitudinal sections passes through the base of the cone, so as to show part of the peduncle in connection with it. In this peduncle secondary wood is present, just as in the separate specimen belonging to the Williamson collection. Higher up in the axis of the cone, where the sporophylls begin to appear, the secondary wood dies out. This evidence materially confirms the conclusion that the Williamson peduncle really belongs to our strobilus. Diagnosis. It is evidently necessary to establish a new genus for the reception of this fossil; the generic name which I propose is Cheircstrobus, intended to suggest the palmate division of the sporophyll-lobes (xeipi hand). The species maybe appropriately named Pettycurensis, from the locality where the important deposit occurs, which has yielded this strobilus and so many other valuable specimens of 422 Dr. D. H. Scott. On Cheirostrobus, a new Type palaeozoic vegetation. The diagnosis may provisionally ruu as follows : — Cheirostrobus, gen nov. Cone consisting 9f a cylindrical axis, bearing numerous compound sporophylls, arranged in crowded many-membered verticils. Sporophylls of successive verticils superposed. Each sporophyll divided, nearly to its base, into an inferior and a superior lobe; lobes palmately subdivided into long segments, of which some (probably the inferior) are sterile, and others (probably the superior) fertile, each segment consisting of an elongated stalk bearing a terminal lamina. Lammas of sterile segments foliaceous ; those of fertile segments (or sporangiophores) peltate. Sporangia large, attached by their ends remote from the axis, to the peltate laminae of the sporangiophores. Sporangia on each sporangiophore, usually four. Spores very numerous in each sporangium. Wood of axis polyarch. C. Pettycurensis, sp. nov. Cone, 3—4 cm. in diameter, seated on a distinct peduncle. Sporo- phylls, twelve in each verticil. Each sporophyll usually sexpartite, three segments belonging to the inferior, and three to the superior, lobe. Sporangia densely crowded. Spores about 0'065 ram. in diameter. Horizon: Calciferous Sandstone Series. Locality : Pettycur, near Burntisland, Scotland. Found by Mr. James Bennie, of Edinburgh. Both generic and specific characters are manifestly subject to alter- ation, if other similar fossils should be discovered. In the mean time the above diagnoses are given, in order to facilitate identification. Affinities. Any full discussion of affinities must be reserved for the detailed memoir, which I hope to lay before the Royal Society in a short time. At present only a few suggestions will be offered The idea of a near relationship to Lepidostrobus — so specious at firsb sight — is negatived by ascurate investigation. There may have been a certain resemblance in external habit, as there is in the naked-eye appearance of the sections, but this means nothing more than that the specimen is a large cone, with crowded sporophylls and radially elongated sporangia. The only real resemblance to Lepidostrobus is in the polyarch strand of primary wood, but even here the details, as, for example, the structure of the trachea?, do not of Fossil Cone from the Calciferous Sandstone. 423 agree. In other respects the differences from any Lepidodendroid fructification are as great as they can be. I do not doubt that the genus with which Cheirostrobus has most in common is Sphenophyllum. The chief points of agreement are as follows. 1. The superposed foliar whorls. This certainly agrees with the vegetative parts of Sphenophyllum, and, according to Count Solms- Laubach, the superposition holds good for the bracts of the strobili also.* 2. The deeply divided palmatifid sporophylls agreeing with the leaves of various species of Sphenophyllum, e.g., 8. tenerrimum. 3. The division of the sporophyll into a superior or ventral, and an inferior or dorsal, lobe, agreeing with the arrangement in 'Sphewophyllum Dawsoni, or 8. cuneifolium, according to M. Zeiller's interpretation.f 4. The differentiation of the sporophyll into sterile segments (bracts) and fertile segments (sporangiophores). The comparison with Sphenophyllum is much strengthened if, as I believe to be the 43ase, the segments of the inferior lobe in Cheirostrobus are sterile, and those of the superior lobe fertile. 5. The repeated subdivision of the leaf-trace vascular bundles, in passing through the cortex of the axis,J as in Sphenophylhim Stephanense. 6. The attachment of the sporangia to a laminar expansion at the •distal end of the sporangiophore. As regards this point, comparison should be made with the Bowmanites Romeri of Count Solms-Lau- bach (loc. cit.). 7. The structure of the sporangial wall. I think that the sum of these characters, to which others might be added, justifies the suggestion that Cheirostrobus may be provisionally placed in the same phylum, or main division, of Pteridophyta, with Sphenophyllum, though indications of possible affinities in other directions are not wanting, and will be discussed on another occa- sion. Cheirostrobus, even more than Sphenophyllum itself, appears to •combine Calamarian with Lycopodiaceous characters, and might reasonably be regarded as a highly specialised representative of an •ancient group of plants which lay at the common base of these two series. It appears likely that in Cheirostrobus one of those additional forms * ' Bowmanites Romeri, eine neue Sphenophylleen Fructification,' 1895, p. 242. t " Etude sur la constitution de 1'appareil fructificative des Sphenophyllum." •* Mem. de la Soc. Q-eol. de France, Paleontologie,' Mem. 11, 1893, p. 37. J Cf. Renault, ' Cours de Botanique fossile,' vol. 2, PL 14, fig. 2 ; PL 15, fig. 3, Tol. 4, p. 15. VOL. LX. 2 K 424 Proceedings and List of Papers read. of Palaeozoic Cryptogams, allowing of comparison with SphenopTiyllum, lias actually been brought to light, the discovery of which Dr. Williamson and I ventured to anticipate at the close of our first joint memoir.* January 28, 1897, Sir JOSEPH LISTER, Bart., RR.C.S., D.C.L., President, followed by Sir JOHN EVAN'S, K.C.B., Treasurer and Yice- President, in the Chair. Mr. John Eliot and Dr. Edward Charles Stirling were admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. The Treasurer offered the congratulations of the Society to the President on his elevation to the peerage. The following Papers were read :— I. " On the Capacity and Residual Charge of Dielectrics as affected by Temperature and Time." By J. HOPKINSON, F.R.S., and E. WILSON. II. " On the Electrical Resistivity of Eleptrolytic Bismuth at Low Temperatures and in Magnetic Fields." By JAMES DEWAR,, M.A., LL.D., F.R.S., Fullerian Professor of Chemistry in the Royal Institution; and J. A. FLEMING, M.A., D.Sc., F.R.S., Professor of Electrical Engineering in University College, London. III. " On the Selective Conductivity exhibited by certain Polarising Substances." By JAGADIS CHUNDER BOSE, M.A., D.Sc., Pro- fessor of Physical Science, Presidency College, Calcutta. Communicated by Lord RAYLEIGH, F.R.S. - * Williamson and Scott, " Further Observations on the Organisation of the Fossil Plants of the Coal-measures," Part I, ' Phil. Trans./ B, 1894, p. 946. On the Capacity and Residual Charge of Dielectrics. 425 "On the Capacity and Residual Charge of Dielectrics as affected by Temperature and Time." By J. HoPKlNSOk; F.R.S., and E. WILSON. Received December 15, 1896,— Read January 28, 1897. (Abstract.) The major portion of the experiments described in the paper have been made on window glass and ice. It is shown that for long times residual charge diminishes with rise of temperature in the case of glass, but for short times it increases both for glass and ice. The capacity of glass when measured for ordinary durations of time, such. as l/100th to l/10th second, increases much with rise of tem- perature, but when measured for short periods, such as 1/106 second, , it does not sensibly increase. The difference is shown to be due to the residual charge, which comes out between l/50,000th second and l/100fch second. The capacity of ice when measured for periods of l/100th to l/10th second increases both with rise of temperature, and with increase of time, its value is of the order of 80, but when measured for periods such as 1/106 second, its value is less than 3. , The difference again is due to residual charge coming out during short times. In the case of glass, conductivity has been observed at fairly high temperatures and after short times of electrification ; it is found that the conductivity after l/50,000th second electrification is much greater than after l/10,000th, but for longer times is sensibly constant. Thus a continuity is shown between the conduction in dielectrics which exhibit residual charge and deviation from Max- well's law and ordinary electrolytes. '•' On the Electrical Resistivity of Electrolytic Bismuth at Low Temperatures, and in Magnetic Fields." By JAMES DEWAR, M.A., LL.D., F.R.S., Fullerian Professor of Chemistry in the Royal Institution; and J. A. FLEMING, M.A., D.Sc,. F.R.S., Professor of Electrical Engineering in University College, London. Received January 4, — Read January 28, 1897. In a previous communication to the Royal Society we have pointed out the behaviour of electrolytically prepared bismuth when cooled to very low temperatures, and at the same time subjected to transverse magnetisation.* During the last summer we have extended these * See ' Proc. Roy. Soc.,' vol. 60, p. 72, 1896. " On the Electrical Resistivity of Bismuth at the Temperature of Liquid Air," by James Dewar and J. A. Fleming. See also ' Phil. Mag.,' September, 1895, Dewar and Fleming " On the Variation in the Electrical Resistance of Bismuth when cooled to the Temperature of Solid Air." 2 K 2 426 Profs. J. Dewar and J. A. Fleming. On ike observations, and completed them, as far as possible, by making- measurements of the electrical resistance of a wire of pure bismuth, placed transversely to the direction of the field of an electromagnet, and at the same time subjected to the low temperature obtained by the use of liquid air. Sir David Salomons was so kind as to lend us for some time his large electromagnet, which, in addition to giving a powerful field, is provided with the means of easily altering the interpolar distance of the pole pieces, and also for changing from one form of pole piece to another. The form of the pole piece most frequently used was that of a truncated cone. The magnet was always excited by a constant current obtained from a constant potential circuit. To save the considerable labour of determining again and again the strength of the interpolar field, this was determined once for all, corresponding to various interpolar distances and a given exciting current. The field was measured by suddenly removing from it a small exploring coil of wire of known area, the same being connected to a standardised ballistic galvanometer. By this means a curve was constructed which showed at once the axial interpolar field at the central point in terms of the interpolar distances, the magnetising current being kept constant. This curve proved, as was to be expected, to be nearly a rectangular hyperbola. This being done the bismuth wire to be examined was formed into a narrow loop of a single turn, about 3 or 4 cm. in length, and the ends soldered to leading-in wires of copper. The loop was placed in a small glass vacuum vessel, with the plane of the loop perpendicular to the direction of the axial magnetic field of the magnet. The loop was placed at equal distances from the two pole pieces, and in a nearly uniform field of known strength. The vacuum vessel was then filled up with either liquid air, a solution of solid carbonic acid in ether, or else simply with paraffin oil. In a fourth case the vacuum vessel was closed, and liquid air having been placed in it, this liquid was caused to boil under a reduced pressure of 25 mm., thus giving a temperature falling as low as —203° C. In another experiment the vacuum vessel was dispensed with, the bismuth wire was simply wrapped in cotton wool, placed between two pieces of thin mica between the pole pieces, and by pouring upon the wrapping a copious libation of liquid air, the temperature of the bismuth wire was reduced to —185° C. In all cases great care was taken to avoid thermo-electric complica- tions, by providing that the soldered junctions by which the bismuth wire is connected to the copper leading-in wire were at exactly the 'same temperature, and to secure this the junctions were always kept well covered with the refrigerating solution. Electrical Resistivity of Electrolytic Bismuth. 427 The bismuth employed was electrolytic bismuth pressed into wire 0*5245 mm. in diameter, and its purity was confirmed by spectro- scopic examination. These arrangements being made, the observations consisted in measuring the electrical resistance of the bismuth at one tempera- ture, but when the transverse magnetic field had values varying from zero to nearly 22,000 C.G.S. units. In the following tables the results are collected. The electrical resistivity of the bismuth is stated for each temperature, and for the various transverse fields employed. As the specimens of the bismuth wire used in the various experi- ments had different lengths, the actual figures of observation are not given, but they have been reduced so as to give the volume resistivity of the bismuth, corresponding to a certain temperature and magnetic field strength. In the case of the experiment in liquid air boiling under a reduced pressure, on account of the size of the vacuum vessel necessary to contain the required initial volume of liquid air, the pole pieces of the magnets could not be brought very near together, and hence the field could not be raised to a very high value. Hartman and Brauns Pure Electrolytic Bismuth. Resistivity of Bismuth Transversely Magnetised at Ordinary Tem- peratures ( + 19°C.). Strength of field Yolume resistivity in (C.a.S. units). C.GLS. units. 0 116,200 1,375 118,200 2,750 123,000 8,800 149,200 14,150 186,200 21,800 257,000 Resistivity of Bismuth Transversely Magnetised at —79° C. Strength of field Yolume resistivity in (C.G.S. units). C.G-.S. units. 0 78,300 650 83,300 2,300 103,500 3,350 114,800 4,100 134,000 5,500 158,000 7,900 201,000 14,200 284,000 428 Profs. J. Dewar and J. A. Fleming. On the Resistivity of Bismuth Transversely Magnetised at —185° C. Strength of field (C.G.S. units). 0 1,375 2,750 8,800 14,150 21.800 Yolunie resistiyity in C.GKS. units. 41,000 103,300 191,500 738,000 1,730,000 6,190,000 Hartman and Braun's Pure Electrolytic Bismuth. Resistivity of Bismuth Transversely Magnetised at —203° C. Strength of field (C.G-.S. units). 0 2,450 Volume resistivity in C.G-.S. units. 34,300 283,500 Electrical Resistivity of Bismuth in C.G.S. units, transversely magnetised in a Constant Magnetic Field, but at variable Tem- peratures. Temperature In the magnetic field. nf thp Oiif nf tli A bismuth wire. magnetic field. Strength 2450 C.G.S. units. Strength 5500 C.G.S. units. Strength 14,200 C.G.S. units. -i- 19° C. 116,200 123,500 132,000 187,000 - 79,, 78,300 105,000 158,000 284,000 -185 „ 41,000 186,000 419,000 1,740,000 -203 „ 34,300 283,500 ~ ~~ It will be seen that the observations lead to the following conclu- sions. If the transverse field is zero, then cooling the bismuth always reduces its resistance. If then the bismuth is transversely magnetised, the resistance is increased, and for every temperature below the normal one (about 20° C.), there is some particular strength of trans- verse field, which just annuls the effect of the cooling, and brings the resistance of the bismuth back again to the same value it had when not cooled, and not in any magnetic field. Hence the curves showing the resistance at any temperature lower than the normal one (20° C.) as a function of the transverse field, cross the curve showing the resistance as a function of the field when taken at the normal tem- perature. These crossing points are, however, not identical for Electrical Resistivity of Electrolytic Bismuth. 429 different resistance-temperature-field curves. The lower the tem- perature the less is the strength of field which will bring the bismuth back to its original resistance when not cooled and not in the field. The observations have been set out graphically in the diagrams in fig8- 1> 2, and 3, and it will be seen that there are in fig. 1 four curves. Each of these curves corresponds to a different temperature, viz., that of liquid air (-185° C.), liquefying carbonic acid in ether (—79° C.), ordinary temperatures (20° C.), and a fourth shorter curve, which corresponds to a very low temperature of —203° C., obtained by . 1. $000,000 5,000000 4000000 I -§ 3,000,000- /,ooo,ooo- Ch&nge of of when O- » 3,000- /QOOO- /5,OOO- £O,000 -n 3trehg£h of Transverse Magnetic F/e/d /n C.G.S.u/7/te. evaporating liquid air under a reduced pressure. This last curve is only continued for a short distance. These curves show the mode of variation of the resistance of the bismuth at a constant temperature as a function of the transverse magnetic field ; and they show how 430 Profs. J. Dewar and J. A. Fleming. On tho FIG. 2. 600,000 • 700,000 • 600,000 • Cnange of Rests £/wty of E/ec£ro/y£/'c Bismuth when transverse/y magnetised. (£n/arged sda/e). O- 2OOO- 4000- 6,OOO •_ 8,OOO- ^ Strength of Transverse Magnet/'c f/'e/d m C.G.S. unite. remarkably the resistance is affected by such magnetisation. The* curve of resistance taken in liquid air, shows that by a transverse magnetising field having a strength of 22,000 C.G.S. units, the resistance of the bismuth is made 150 times greater than the resist- ance of the same wire in a zero field, but at the same temperature. The lower the temperature to which the bismuth is reduced the greater is the multiplying power of a given transverse field upon its electrical resistivity. Hence a still lower temperature than we have been able to apply would doubtless render the bismuth still more sensitive to transverse magnetisation. We have already shown that pure bismuth is no exception to the- generally observed fact that all pure metals continuously lose their electrical resistivity as they approach in temperature the absolute? Electrical Resistivity of Electrolytic Bismuth. 431 zero. Hence at this last temperature it should be converted into a non-conductor by a sufficiently strong transverse magnetisation. This result will have to be taken into consideration in framing any theory of electrical conduction. In this respect bismuth is a remarkable exception to other metals^ We have tried the effect of transverse magnetisation at low tem- peratures on zinc, iron, and nickel, but find no effect sensibly greater at low than at ordinary temperatures, although these metals- have their resistance affected by magnetisation to a small degree. Bismuth has an exceptional position amongst other metals, both in respect of its large coefficient of the Hall effect, and also in the degree to which its resistance is thus affected by transverse magnetisation, and in addition, as above shown, in the degree to* which cooling to low temperatures affects this ability to be so changed by magnetisation. Very small amounts of impurity in the metal reduce these remark* able qualities considerably. We may mention here that we have repeated the experiments we made some time ago* on certain specimens of chemically prepared bismuth, and for which we found the electrical resistance had a minimum value for a certain temperature. We have again verified this fact, both for the same and for a similar specimen. In the former experiments the bismuth wire used was embedded in paraffin wax during the cooling, and the suspicion had arisen that strains might thus have been produced which had affected the results. In the repetition of the experiments, we suspended the bismuth wire freely in liquid air, so that no strains could be produced ;. and, in addition, we tried the effect of mechanical stress on the resistance directly. We satisfied ourselves that the cause of the anomaly in the behaviour of the chemically prepared bismuth in respect of electrical resistance at low temperatures was not to be found in any effect due to strain. In fig. 3 a series of curves have been drawn showing the variation in resistivity of the electrolytic bismuth for certain constant trans- verse magnetic fields and varying temperatures. These curves were obtained by taking sections of the curves in figs. 1 and 2. The curves in fig. 3 are practically the continuation from 19° C. down to — 186° C. of curves which have been given by Mr. J. B. Heiiderson,f for a range of temperature lying above 0° C. They show that if a wire of electrolytic bismuth is placed trans- versely in a certain magnetic field, there is, for a wide range of field, * See ' Phil. Mag.,' September, 1895, p. 303. Dewar and Fleming " On the Variation in the Electrical [Resistance of Bismuth when cooled to the Temperature of Solid Air." t See ' Phil. Mag.,' 1894, vol. 38, p. 488. 432 On the Electrical Resistivity of Electrolytic Bismuth. Fia. 3. n \r 600,000- 300,000- 400,OOO- .-^ 300,000- «S | £00,000- ^ JO O,000- «- f Resist iw of Elect ro/y transfers — £00 - /50° -/00° — 50°- —0° +£0° Temperature //? decrees centigrade. a certain temperature at which the bismuth has a minimum electrical resistivity, and, therefore, a zero temperature coefficient, and that the temperature of this turning point is higher the stronger the transverse field. These curves also show that at a temperature of about 150° C., the bismuth would probably cease to have its resis- tivity affected by a transverse magnetic field.* In conclusion, we desire to mention the assistance we have received from Mr. J. E. Petavel in the work described above. * Drude and Nernst (' Wied . Ann.,' vol. 42, p. 568) found that, with a transverse field of 7000 C.GLS. units, the total percentage increase of resistance of electrolytic bismuth was 22'0, 8'0, TO, and 0'4 per cent, respectively at temperatures of 16° C., 100° C., 223° C., and 290° C. On the Selective Conductivity of Polarising Substances. 433 " On the Selective Conductivity exhibited by certain Polarising Substances." By JAGADIS CHUNDER BOSE, M.A., D.Sc., Professor of Physical Science, Presidency College, Calcutta. Communicated by Lord RAYLEIGH, F.R.S. Received January 14, — Read January 28, 1897. In my paper " On the Polarisation of Electric Rays by Double- refracting Crystals " (vide ' Journal of the Asiatic Society of Bengal,' May, 1895), and in a subsequent paper " On a New Electro-Polari- scope" (' Electrician,' 27th December, 1895), I have given accounts of the polarising property of various substances. Amongst the most efficient polarisers may be mentioned nemalite and chrysotile. Nemalite is a fibrous variety of brucite. In its chemical composition it is a hydrate of magnesia, with a small quantity of protoxide of iron and carbonic acid. This substance is found to absorb very strongly electric vibrations parallel to its length, and transmit those that are perpendicular to the length. I shall distinguish the two directions as the directions of absorption and transmission. Chrysotile is a fibrous variety of serpentine. In chemical composition it is a hydrous silicate of magnesia. Like nemalite, it also exhibits selective absorp- tion, though not to the same extent. The transmitted vibrations are perpendicular, and those absorbed parallel to the length. Different varieties of these substances exhibit the above property to a greater or less extent. I have recently obtained a specimen of chrysotile with a thickness of only 2'5 cm. ; this piece completely polarises the trans- mitted electric ray by selective absorption. The action of these substances on the electric ray is thus similar to that of tourmaline on light. It may be mentioned here that I found tourmaline to be an inefficient polariser of the electric ray ; it does transmit the ordinary and the extraordinary rays with unequal intensities, but even a considerable thickness of it does not completely absorb one of the two rays. In Hertz's polarising gratings, electric vibrations are transmitted perpendicular to the wires, the vibrations parallel to the wires being reflected or absorbed. Such gratings would be found to exhibit electric anisotropy, the conductivity in the direction of the wires being very much greater than the conductivity across the wires. The vibrations transmitted through the gratings are thus perpen- dicular to the direction of maximum conductivity — or parallel to the direction of greatest resistance. The vibration absorbed is parallel to the direction of maximum conductivity. As the nemalite and chrysotile polarised the electric ray by unequal absorption in the two directions, I was led to investigate whether 434 Prof. J. C. Bose. On the Selective Conductivity they, too, exhibited unequal conductivities in the two directions of absorption and transmission. Nemalite, unfortunately, is difficult to obtain, and the specimens I could get here were too small to make the necessary measurements. I have, however, in my possession two specimens which I brought from India ; of these, one is a perfect specimen of a fair size, and I obtained with it strong polarisation effects. The second piece is not as good as the first, and rather small in size. I cut from this latter piece a square of uniform thickness, the adjacent sides of the square being parallel to the directions of transmission and absorption respec- tively. The resistances of equal lengths in the two directions (with the same cross section) were now measured. The first specimen I gave to Messrs. Elliott Brothers for measure- ment. They informed me, on the 13th of October last, that the resistance in the direction of transmission was found to be 35,000 megohms, and that in the direction of absorption, only 14,000 meg- ohms. It will thus be seen that the direction of absorption is also the direction of greatest conductivity, and the direction of transmission is the direction of least conductivity. My anticipations being thus verified, I proceeded to make further measurements with other specimens. From the perfect specimen of nemalite in my possession, I cut two square pieces, A and B. The size of piece A is 2'56x2'56 cm., with a thickness of 1/1 cm. B is- 276 x 2-76x1-2 cm. For the determination of resistances I used a sensitive Kelvin gal- vanometer, having a resistance of 7000 ohms. With three Leclanche cells, 1/4 volt each, and an interposed resistance equivalent to- 55,524 megohms, a deflection of 1 division in the scale reading was obtained. The following table (p. 435) gives the results of the measurements which I carried out. The results given clearly show how the difference of absorption in the two directions is related to the corresponding difference in conductivity. I then proceeded to make measurements with chrysotile. The specimens I could obtain were not very good. I cut two from the same piece, and a third specimen was obtained from a different variety. The ratios of conductivities found in the three specimens were 1 : 10, 1 : 9, and 1 : 4 respectively. In every case the direction of absorption was found to be the direction of maximum conductivity. [A fibrous variety of gypsum (CaS04), popularly known as Satin- spar, also exhibits double absorption ; and in this case, too, the con- ductivity in the direction of absorption is found to be very much greater than in that of transmission. exhibited by certain Polarising Substances. 435 Resistance between Specimen A. Deflections. two opposed faces 2-56 x 1-1 cm. Ratio of the conduc- separated by tivities. 2 -56 cm. In the direction of transmission „ ,, absorption.. 26 360 2136 megohms 154 | 1 :13'8 Specimen B. Deflections. Resistance between two opposed faces 276 x 1-2 cm. separated by 2-76 cm. Ratio of the conduc- tivities. In the direction of transmission „ „ absorption . . 28 370 1983 megohms 150 „ | 1 : 13 '4 One of the strongest polarising substances I have come across is the crystal epidote. The crystal is very small in size, and I could not get with it complete absorption of one of the two rays. But it exhibits very strong depolarisation effect, even with a thickness as small as 0'7 cm. This is, undoubtedly, due to strong selective absorp- tion in one direction. I cut a square from this crystal 0'7xO-7 cm. with a thickness of 0'4 cm. Using an E.M.F. of 14 volts the deflections obtained (proportional to the two conductivities) were 105 and 20 divisions respectively. The conductivities in the two directions are, therefore, in the ratio of 5'2 : 1. With an E.M.F. of 100 volts and a -diminished sensibility of the galvanometer, the deflections were 205 and 40, the ratio of the conductivities being as 5'1 : 1. — January 28, 1897.] It would thus appear that substances like nemalite which polarise by •double absorption, also exhibit double conductivity. It is probable that, owing to this difference of conductivity in the two directions, each thin layer unequally absorbs the incident electric vibrations ; 3,nd that by the cumulative effect of many such layers, the vibrations which are perpendicular to the direction of maximum conductivity are alone transmitted, the emergent beam being thus completely polarised. [Owing to the great difficulty in obtaining suitable specimens, I have not been able to make a more extended series of determina- tions. The relation found, in the cases described above, between double absorption and double conductivity is, however, suggestive. 436 On the Selective Conductivity of Polarising Substances. It should, however, be borne in mind that the selective absorption exhibited by a substance depends, also, on the vibration frequency of the incident radiation. I have drawn attention to the peculiarity of tourmaline which does not exhibit double absorption of the electric ray to a very great extent. The specimen I experimented with is, however, one of a black variety of tourmaline, and not of the semi- transparent kind generally used for optical work. Though the experiments already described are not sufficiently nume- rous for drawing a general conclusion as to the connection between double absorption attended with polarisation, and double conductivity, there is, however, a large number of experiments I have carried out which seem to show that a double-conducting structure does, as a rule, exhibit double absorption and consequent polarisation. Out of these experiments I shall here mention one which may prove interesting. Observing that an ordinary book is unequally conducting in the two directions — parallel to and across the pages — I interposed it, with its edge at 45°, between the crossed polariser and analyzer of an electro- polariscope. The extinguished field of radiation was immediately restored. I then arranged both the polariser and the analyzer vertical and parallel, and interposed the book with its edge parallel to the direction of electric vibration. The radiation was found completely absorbed by the book, and there was not the slightest action on the receiver. On holding the book with its edge at right angles to the electric vibration, the electric ray was found copiously transmitted. An ordinary book would thus serve as a perfect polariser of the electric ray. The vibrations parallel to the pages are completely absorbed, and those, at right angles transmitted in a perfectly polar- ised condition. — January 28, 1897.] Proceedings and List of Papers read. 437 February 4, 1897. Sir JOSEPH LISTER, Bart., F.R.C.S., D.C.L., President, in the- Chair. A List of the Presents received was laid on the table, and thanks- ordered for them. The President stated that a paper had been received from Dr. Arthur Willey, Balfonr Student of the University of Cambridge, the recipient of a Government Grant, and now staying at the Loyalty Islands, to the effect that he had discovered the ova of Nautilus. The following Papers were read: — :r-r i.. I. " On the Condition in which Fats are absorbed from the Intes- tine." By B. MOORE and D. P. ROCKWOOD. Communicated by Professor E. A. SCHAFER, F.R.S. II. " The Gaseous Constituents of certain Mineral Substances and Natural Waters." By WILLIAM RAMSAY, F.R.S., and MORRIS W. TRAVERS, B.Sc. III. " Some Experiments on Helium." By MORRIS W. TRAVERSA B.Sc. Communicated by Professor W. RAMSAY, F.R.S. IV. " On the Gases enclosed in Crystalline Rocks and Minerals." By W. A. TJLDEN, D.Sc., F.R.S. V. " On Lunar Periodicities in Earthquake Frequency," By C. G. KNOTT, D.Sc., Lecturer on Applied Mathematics, Edinburgh University (formerly Professor of Physics, Imperial Univer- sity, Japan). Communicated by JOHN MILNE, F.R.S, 438 Messrs. B. Moore and D. P. Rockwood. On the ** On the Condition in which Fats are absorbed from the Intestine." By B. MOORE and D. P. ROCKWOOD. Commu- nicated by Professor E. A. SCHAFER, F.R.S. Received December 24, 1896,— Read February 4, 1897. (From the Physiological Laboratory of University College, London.) In 1858 Dr. W. Marcel* announced to this Society the discovery that bile possesses the remarkable property of dissolving to a clear solution large amounts of fatty acids, and mixtures of these, when heated above their melting points, and that, on cooling, these bodies are again thrown out as a fine precipitate or emulsion. We have repeated these experiments, and are able to confirm the accuracy of Marcet's observation. Thus we found that 6 c.c. of dog's bile at 62° C. dissolved completely 1'5 grams of the mixed fatty acidsf of beef suet, and similar solubilities were found in other -cases. No other observations than these have, so far as we are aware, been made on the effect of temperature on the solubility of fatty acids in bile ; although different writers have mentioned that fatty acids are soluble in bile, no measurements have been made of the extent of their solubility. AltmarinJ has recently surmised that fats are absorbed from the intestine as fatty acids, dissolved in the intestine by the agency of the bile, but has made no quantitative experiments on the solubilities of the fatty acids in bile. The for- gotten experiments of Marcet, mentioned above, led us to think that the fatty acids might possess, at tine temperature of the body, a fair amount of solubility in bile, and as the solubility at this temperature is that of most physiological interest, we have made a. series of deter- minations of the solubilities of oleic, palmitic, and stearic acids, and of natural mixtures of these in the proportions in which they occur in lard, beef suet, and mutton suet, in the bile of the ox, pig, and dog. Different methods were used in the determination of these solu- bilities : — 1. To a measured amount of the bile under experiment, kept at a temperature of 39° C., small weighed quantities of the fatty acid ainder experiment were added, until no more dissolved. 2. A quantity of bile was saturated at 39° C., with excess of the fatty acid, and filtered from the excess of undissolved acid through a * ' Eoy. Soc. Proc.,' 1858, vol. 9, p. 306. f Throughout this communication the expression " fatty acids " means the fatty acids present in fats, oleic, palmitic, and stearic acids. I 'Arch. f. Anat. u. Physiol.,' 1889, Anat-.-Abth., Suppl. Band, p. 86. Condition in which Fats are absorbed from the Intestine. 439 hot funnel, at this temperature ; the filtrate was cooled to about 0° C., and the precipitate collected, dissolved in ether, recovered therefrom, and weighed ; the weight, compared with the volume of the filtrate, gave a measure of the solubility. 3. To a series of equal volumes (10 c.c.) of bile in test-tubes, a rising series of weights of fatty acids was added (O05, 0*1, 0'15, 0'2, &c., grams), and those tubes noted, in which, after the lapse of a sufficient time at 39° C., complete solution did not take place. The following is a summary of our results. 1. Ox bile .... 2. Pig's bile .. Lard fatty acids.* Beef suet acids. Mutton suet acids. Oleic acid. Palmitic and stearic acids. 2 -5—4 p. c. 4 2-5— 3 p. c. 5-6 „ 1—2 -5 p. c 1—2-5 „ 4-5 p. c. Less than 0-5 p. c. 3. Dog's bile . . 6 '25 4-7 „ 2 — — The fatty acids are not dissolved as soaps, but probably as fatty acids, for the solution becomes strongly acid ; moreover, the material thrown out on cooling dissolves easily in ether, and, when recovered, saponifies at once with sodium carbonate. The solution is not entirely due to the bile salts, for mere removal of the " bile mucin " greatly diminishes the solvent power, although the "mucin'* redissolved in sodium carbonate solution has no solvent power, and, again, a solution of mixed bile saltsf stronger than bile has not nearly so much solvent power as the bile itself. . • Palmitic and stearic acids are very feebly soluble in bile at 39° C., and in mixtures are probably dissolved by the aid of the admixed oleic acid. Action of Filtered Intestinal Contents on Fats. The filtered intestinal contents contain both pancreatic juice and bile, and hence should both decompose and dissolve fats at body temperature if these are absorbed as dissolved fatty acids; this was experimentally found to be the case with filtered intestinal con- tents of the dog, which in different cases possessed a very variable * The numbers given are the minimum and maximum of a number of determina- tions in different samples of bile. t The solution used was a 9 per cent, solution of the bile salts of a sample of ox bile which dissolved 2'5 per cent, of the fatty acids of beef suet ; this solution oi bile salts only dissolved 1 per cent. VOL. LX. 2 L 440 Messrs. B. Moore and D. P. Kockwood. On the power, dissolving 1 to 5 per cent, of the fat of beef suet at 39° C» The solution becomes viscid, semi-fluid, or completely solid on cooling, arid redissolves'on warming again. With the filtered contents of the intestine of the pig and rabbit similar results were not obtained, but the fat became altered, being in part converted into fatty acids, and in part giving rise to a voluminous precipitate. Simultaneous Action of Pancreas and Bile on Fats. Finely minced, fresh dog's pancreas (1 gram) was added to bile (10 c.c.), and then the fat of beef suet (0*25 gram) ; the fat com- pletely dissolved in three hours at 40° C. ; on cooling, the solution became turbid, and finally semi-solid. In a control experiment, pancreas alone decomposed fat into fatty acids, but did not dissolve it. The solubilities stated above are quite sufficient to account for the removal of all the fat of the food from the intestine as dissolved fatty acid, since they exceed the concentrations found in the intestine of other materials, such as sugars and albumoses, which are removed in solution. Other experiments, however, on the reaction of the intestine during fat absorption, lead us to think that all the fat is not removed as dissolved fatty acids, but that these are replaced to a variable extent (in some animals, to. a very large extent or completely) lay dissolved soaps. Reaction of Intestinal Contents during Fat Absorption. We have determined the reaction of the contents of the dog's small intestine during fat absorption, from pylorus to caecum, to various indicators, litmus, methyl-orange, and phenolphthalein, and cannot agree with the statement of some other experimenters, that it is acid throughout.* In sixteen experiments on this animal we only once found the reaction acid to litmus up to the caecum, and this was an obviously poor experiment, in which the intestine was almost empty. The reaction to litmus at the pylorus is neutral, faintly acid, or faintly alkaline ; from here onwards the acidity increases, reaches a maximum about the middle of the small intestine, and then becomes less acid, to change to alkaline at a point situate two-thirds to three-fourths of the way along the intestine ; from this point on to the caecum the alkalinity increases. f The reaction to methyl- orange and phenolphthalein explains this ; the intestine is alkaline to methyl-orange all the way from pylorus to caecum, and equally com- * Cash, 'Arch. f. Anat. u. Physiol.,' 1881, p. 386 ; Munk, 'Zeitsck. f. Physiol. Chem./ vol. 9, 1885, pp. 572, 574. t There is usually a reversion to an acid reaction in the large intestine, in "which case the contents of the caecum are almost neutral. Condition in which Fats are absorbed from the Intestine. 441 pletely acid to phenolphthalein, showing that the acid reaction to litmus in the upper part is due to weak organic acids, while the alkaline reaction in the lower is due to fixed alkali, accompanied by dissolved carbonic acid. The alkaline reaction to methyl-orange in the upper part, where it is acid to litmus and phenolphthalein, shows that in that part there is an excess of bases, above that quantity necessary to combine with all the inorganic acids, which are combined with very weak organic acids (probably fatty acids), for methyl-orange is a stable indicator, and does not react to such acids, while litmus, and, still more so, phenol- phthalein, are indicators which are affected by these acids. In the lower third or thereabouts, where the reaction is alkaline to litmus, there cannot be any fatty acids present in solution. Any fat absorbed as free fatty acid in solution must, therefore, be taken up from the upper two-thirds or three-fourths of the intestine where the reaction is acid to litmus, but even here a considerable part is probably being absorbed in solution as soaps, as is shown by the reaction being at the same time alkaline to methyl-orange. In the lower part all the fat absorbed must be taken up as soaps. During fat absorption in the white rat,* the reaction of the con- tents of the small intestine is commonly alkaline to litmus from pylorus to ceecum, and is never acid for a greater distance than 2 or 3 in. below the pylorus ; in this animal, therefore, nearly all the fat must be absorbed in solution as soaps. We have not investigated the reaction of the intestinal contents in other animals during fat absorption, but in the rabbit, during carbohydrate absorption, it is strongly alkaline all the way, from pylorus to caecum, and in the pig the mixed contents during the absorption of a mixed meal (meal and oats) had a strong alkaline reaction. As already stated, the filtered consents in these animals do not perfectly dissolve fat, and the portion dissolved must be in the form of soap, because the reaction remains, alkaline to litmus after solution. In such animals it is probable that the greater part of the fat must be absorbed as soaps. The main objections which have been urged, against absorption of fats as soaps are, first, absorption in presence of an acid reaction in the dog, in which case it was supposed impossible that soaps could be present simultaneously in solution,f and, secondly, that the * In this animal the intestinal contents are usually semi-solid. Care was taken to mix them so as not to obtain the alkaline surface reaction sometimes described. On thorough mixing an alkaline reaction was obtaine.d. f The acid reaction is also commonly supposed to preclude the possibility of the formation of an emulsion, and Cash ('Arch. f. Anat. u. Physiol.,' 1881, p. 386), in experiments chiefly made to determine this point, failed to find any emulsion within the dog's intestine. In ten out of sixteen experiments we obtained more or less emulsion, and in fire of these, in almost the entire length, a perfect emulsion, con- taining immense numbers of minutest fat globules, and possessing a marked acid 2 L 2 442 Prof. Ramsay and Mr. Travers. The Gaseous amount of alkali required in the intestine for the absorption of all the fats of a fattj meal, as soaps, is out of all proportion to the amount actually present, being about twice the total alkalinity of the body.* The first objection has already been discussed ; it has been shown that the acid reaction is due to weak organic acids, and that an alkaline reaction can be obtained by the use of a proper indicator, due to a compound of these weak acids with bases ; in other words, to soaps. The second objection may be met by the supposition that the same quantity of alkali acts cyclically as a carrier in conveying quantity after quantity of fatty radicle, as soap, from the intestine. The soaps are, it is known, broken up in the intestinal cells, and formed into fats by the action of the cell ; in such a reaction alkali is set free, and there is no obvious reason why it should not be returned to the intestine and serve to carry a fresh portion of fatty radicle dissolved as soap into the epithelial cells. Such an action takes place in the acid secreting cell of the gastric gland, where sodium chloride is taken up from the blood, split into acid and alkali, and the alkali returned to the blood while the acid passes into the gland lumen ; it is not, therefore, unreasonable to suppose that a similar action can take place in the intestinal absorbing cell. We conclude that in certain animals, such as the dog, fats are absorbed partially as dissolved fatty acids,' and partially as dissolved soaps ; while in other animals, such as the white rat, fats are chiefly, if not entirely^ absorbed as dissolved soaps. " The Gaseous Constituents of certain Mineral Substances and Natural Waters." By WILLIAM RAMSAY, F.R.S., and MORRIS W. TRAVERS, B.Sc. Received December 30, 1896, —Read February 4, 1897. It is still uncertain whether helium is a single elementary gas or a mixture of two or more gases. If a mixture, it is probable that they should occur independently, and that the proportion of the con- stituent gases should vary in samples from different sources. During the past year the gases obtained from a large number of minerals and natural waters have been examined with a view to investigate this point, and, also, to determine whether any new gaseous element could be discovered. In every instance the results have been negative ; no reaction to litmus. Although fats are not absorbed in the form of an emulsion, it is evident that the formation of an emulsion in the intestine must enormously increase the surface exposed to the action of the intestinal fluids, and proportion- ately increase the rate at which the fats are decomposed and dissolved. * Munk, ' Virchow's Archiv,' vol. 95, 1884, p. 408. Constituents of certain Mineral Substances and Waters. 443. indication of the presence of any new element has been obtained, nor has any abnormality been observed in the spectrum of any of the examined. Fm. 1. Method of Examination of the Mineral Substance. The mineral was ground to fine powder in an agate mortar, and then mixed with about twice its weight of acid potassium sulphate. This mixture was placed in a hard glass tube, which was connected with a Topler pump, and, after exhaustion, heated to a red heat by means of a large Bunsen burner. The gases evolved were pumped off and collected over mercury in a tube containing a little caustic potash solution. In some instances, however, the mineral was heated alone ; the same result was obtained, but the evolution of gas takes place rather more slowly. In order to diminish any chance of leak- age of air into the apparatus, the hard glass tube was connected with the pump in the manner shown in fig. 1. The tube was drawn out to a neck at the point A. A piece of thick- walled rubber tube was fitted over the end of the tube B connected with the pump, and it was then forced tightly into the neck of the hard glass tube. By pouring a little mercury into the cup C the joint could be made absolutely air-tight. Examination of Minerals and Rocks. Several samples of fergusonite, monazite, and samarskite were first examined, and were found to give quantities of helium up to 1*5 c.c. per gram. Columbite (a variety of tantalite), an isomorphous mixture of niobate and tantalate of iron and manganese, gave 1*3 c.c. of gas con- sisting chiefly of helium. Pitchblende, containing zirconium, obtained by Dr. Hillebrand from Colorado, gave 0'36 c.c. of gas per gram, of which 0*3 c.c. was helium. Another sample gave 0'27 c.c. of helium per gram. 444 Prof. Ramsay and Mr. Travers. The Gaseous '- Malacone, ZrS04, from Hitteroe in Norway, was the only mineral in which argon was found. Five grams of the mineral gave 12 c.c. of gas unabsorbed by caustic soda. After explosion with oxygen, and absorption of the residual oxygen with phosphorus, about Ol c.c. of gas remained. The residue was introduced into a tube with aluminium electrodes which was sealed off from the pump and attached to a coil giving a discharge sufficiently powerful to heat the electrodes to a red heat. The nitrogen was quickly absorbed by the red-hot electrodes, and, as soon as the banded spectrum had dis- appeared, the lines of helium and argon became visible. The green line of the helium spectrum was very strong, and the glow in the- tube was distinctly green. The argon present was in too large quantity to be attributed to- accidental leakage of air into the apparatus ; but, in order to confirm this exceptional result, and also to determine whether the green effect in the tube was due entirely to the low pressure of the helium, the experiment was repeated with a larger quantity of the mineral. With 10 grams of the mineral a quantity of gas was obtained, which, after removal of nitrogen, gave a yellow glow in the vacuum- tube ; argon was again present, and its second spectrum could be brought out very strongly by means of a jar and a spark-gap in the- secondary circuit. The experiment was repeated a third time with the same result. This proved conclusively that inalacone contains both argon and helium. Cinnabar. — Five grams gay e 0*5 c.c. of gas, which consisted only of carbon monoxide. Cryolite. — 7*6 grams gave only a minute bubble of carbon mon- oxide, Apatite. — Six grams gave O5 c.c. of a gas consisting wholly of hydrogen and carbon monoxide. Baryta-celestine. — No gas was evolved; the pump remained at a phosphorescent vacuum. Serpentine. — This specimen was from the Riffelhorn, and has been analysed by Miss Aston ;* 5 grams gave 4 c.c. of gas which consisted wholly of hydrogen. Gneiss, from the Diamirai Glacier, directly below the peak of Nanga-Parbat, Kashmir, brought home by Dr. Collie : 3 grams gave 6 c.c. of hydrogen. Scapolite, a silicate of calcium, magnesium, and aluminium, gave no gas. Cobalt ore, containing a considerable quantity of manganese dioxide: — 3'2 grams of mineral, heated alone, gave 35 c.c. of gas- consisting wholly of oxygen. * ' G-eol. Soc. Journ.,' 1896, p. 452. Constituents of certain Mineral Substances and Waters. 445 Lava from Iceland : — Two specimens were examined ; in each case a little carbon dioxide was obtained. Some specimens from the Kimberley diamond field, obtained from Mr. Crookes : — Blue clay : — A considerable quantity of a mixture of hydrogen and carbon monoxide was obtained. After explosion with oxygen, no trace of gas remained. Coarse-grained gravel and so-called " carbon " gave the same result. Examination of Specimens of Meteoric* Iron. Specimens of meteoric iron were kindly sent for examination by Dr. Fletcher of the British Museum : — Greenbrier County meteorite : — Ten grams of metal gave a fairly large quantity of gas on heating, which consisted wholly of hydrogen. Toluca meteorite: — One gram gave 2'8 c.c, of pure hydrogen. Charca meteorite : — One gram gave 0*28 c.c. of hydrogen. Bancho de la Pila meteorite (' Min. Mag.,' ix, 153) : — One gram gave 0*57 c.c. of gas. It consisted of hydrogen. Obernkirchen Meteorite, from Schaumberg-Lippe, Germany, de- scribed by Wichs and Wohler (' Pogg. Ann.,' vol. 120, p. 509) :— One gram gave 2*6 c.c. of gas. The gases from these meteorites were exploded with oxygen, and were found to contain no trace either of argon or helium, or of nitrogen. The carbon compounds present were possibly produced by the decomposition of the oil, &c., with which the shavings of meteoric iron had become contaminated. It will be remembered that a previously examined specimen of meteorite was found to contain both argon and helium. Examination of the Gases held in Solution "by the Waters of certain Mineral Springs. Old Sulphur Well, Harrogate.— One carboy of water gave 650 c.c. of gas from which, after circulation and sparking, 45 c.c. of argon were obtained. Spectroscopic examination of the gas proved that it contained nothing but argon. Strathpeffer Wells. — One carboy of water gave 1 litre of a gas which, after sparking and circulation, gave 22 c.c. of pure argon. The gas was separated from these waters by the method described by Lord Rayleigh ('Phil. Trans.,' A, vol. 186, p. 220). Mineral Springs of Cauterets.— The mineral springs of the Hautes Pyrenees, particularly those containing sulphides, have long been known to contain considerable quantities of nitrogen. Dr. H. C. Bouchard, of Paris, has recently (' Compt. Rend.,' vol. 121, p. 392) pub- lished an account of his examination of gases obtained from the wells 440 Prof. Ramsay and Mr. Travers. The Gaseous at Cauterets, which he has found to contain a considerable quantity of a mixture of argon and helium. He appears to have made a rough spectroscopic examination of the gases, and has stated in his paper- that some of the lines in the red end of the spectrum do not belong to the spectrum either of argon or of helium. The author, a medical man, has dealt with the matter from a purely clinical standpoint, and his paper contains no data with regard to the supposed new lines. To obtain samples of these gases, it was necessary to make a journey to Cauterets, and to visit the wells personally. Taking advantage of the Easter holidays, we left England provided with twelve tin cylin- ders, each with a capacity of 2 litres, for the purpose of collecting samples of gas from as many of the wells as we could obtain admis- sion to. The management of the baths and wells granted us permis- sion to visit the actual sources from which the baths, &c., are supplied, and courteously gave us every assistance, placing at our disposal the services of men connected with the different establishments. We were able to obtain samples of gas from four of the springs close to the town, but, on account of the deep snow, some of the more distant " sources " were quite inaccessible. The " sources " are for the most part situated at the end of tunnels driven for some distance into the hill-side. The water rises from below into tanks beneath the floor of the tunnels, and is conducted through pipes to the baths. Circular holes, about 9 inches diameter, in the floor formed the only means of inspecting the interior of the tanks. The gas appeared to rise with the water from natural springs in the bottom of the tanks ; it was this gas that we collected for our investigation. The apparatus employed is shown in the accompanying figure. A piece of rubber tube B is fitted on to the lower tap of the cylinder A, which was then sucked full of water. The taps were then closed, and the cylin- der fixed in a vertical position, the rubber tube hanging down into the tank. A second piece of rubber tube, C, was fitted on to the funnel D, which was lowered into the tank. Water was then drawn up into the rubber tube, which was immediately slipped over the nozzle of the upper tap on the tin cylinder. The taps were then opened, and the funnel brought over some point on the floor of the tank, from which gas was escaping. The gas rising into the funnel rapidly replaced the water in the cylinder which escaped back into the tank by the lower tube. In some of the wells a large quantity of gas could be collected in a short time, but in others the bubbles rose only very slowly. Name of " source." Temp. Time required to fill vessels. Raillere 39'5° C. One tin in two hours. Des CEufs 51*0 ,, Three tins in 30 minutes. Caesar 46'0 „ One tin in four hours. Espagnol 46'0 „ Three tins in about 15 minutes. Constituents of certain Mineral Substances and Waters. 447 FIG. 2. Floor of Tunnel. We proceeded with the examination of the gases immediately on our return to London. The gases were transferred to a glass gas- holder containing potash solution, and circulated over red-hot mag- nesium and copper oxide. The residual gas was pumped out of the circulating apparatus, and sparked with oxygen over potash to remove final traces of nitrogen. Preliminary Spectroscopic Examination of the Gases. Raillere. — Argon and helium, helium strong. Des CEufs. — Argon, with less helium. Espagnol. — Argon, with helium ; the yellow and green helium lines very distinct, with jar and spark-gap. Caesar. — Argon, with a little helium. The tubes were carefully compared with normal argon and helium tubes, but no new lines could be detected. An attempt was made to separate the gas into its constituents by taking advantage of their relative solubilities. A measured quantity 448 Gaseous Constituents of certain Mineral Substances, fyc. of the gas was confined over a large quantity of boiled water, and the residue taken for examination. Raillere 3'7 c.c. taken, 1*0 c.c. residue. Des CEufs 8-5 „ 4'0 Csesar 2'2 „ 0'5 „ Espagnol 8*0 ,, (not measured). The residue showed the helium lines rather more strongly. The Des CEufs gas was submitted to fractional diffusion by the method described in the following paper. The gas was divided into two portions by diffusion through a. porous plug. These two fractions were then diffused separately, the light fraction of the heavy gas, and the heavy fraction of the light gas forming an intermediate fraction, This was again separated by diffusion into a heavy and a light portion, which were mixed with the heavy and light fraction obtained in the second stage. The process was repeated four times, and the resulting fractions, after sparking with a little oxygen, were rediffused so as to obtain the lightest sixth of the light fraction, and the heaviest sixth of the heavy fraction. In a Pliicker tube, the helium line, D3, appeared somewhat stronger in the light gas, but the difference was not so marked as might have been expected. Neither of the tubes showed any lines other than those of the argon or helium spectrum. The other samples of gas were not submitted to the diffusion pro- cess, as it did not seem probable that any results of value would be obtained. In another paper it is shown that separation of helium from argon can be effected by taking advantage of the absorption of that gas by the platinum splashed on to the walls of the tube during the passage of the discharge. The gas is made to circulate at about 3 mm. pressure through a vacuum -tube with platinum electrodes, and kept cool by a water-jacket. The helium, together with any nitrogen or carbon compounds that may be present, is absorbed by the platinum, and may be liberated by heating the tube with a Bunsen's burner. The heavier fraction of the Des CEufs gas, and some of the gas from the Raillere were treated by this process, and the gas liberated from the platinum on heating was in each case introduced into a vacuum- tube with aluminium electrodes. The tube showed a banded spectrum which disappeared as the nitrogen was absorbed by the heated aluminium, leaving only normal helium at low pressure and a trace of argon. If any other gas, other than argon and helium, be present in the residue from the gas evolved from these various springs, after removal of the nitrogen, the methods employed have totally failed to bring it to light so far. It certainly cannot be present in any measurable quantity. Some Experiments on Helium. 449 " Some Experiments on Helium." By MORRIS W. TRAVERS, B.Sc. Communicated by Professor W. RAMSAY, F.R.S. Received December 30, 1896,— Read February 4, 1897. In July of last year Professors Runge and Paschen (' Phil Mag./ 1895, [ii], vol. 40, pp. 297 — 302) announced their discovery that the spectrum of the gas from cleveite indicated the presence of two ele- ments. They also stated that by means of a single diffusion through an asbestos plug, they had been able to effect a partial separation of the lighter constituent, which was characterised by the green glow which it gave under the influence of the electric discharge in a vacuum-tube, and which was represented in the spectrum by the series containing the green line, X = 5015'6. Subsequently, at the meeting of the British Association at Ipswich, Professor Runge exhibited a tube containing the so-called green constituent; the colour of the glow differed strongly from that of an ordinary helium tube, but the gas contained in it was evidently at very low pressure, as phosphorescence was jusfc commencing. Professor Runge has since acknowledged that the green effect in the helium tube may be produced by a change of pressure alone (' Astrophysical Journal,' January, 1896). During an exhibition of the spectrum of helium at -the soiree of the Royal Society on May 9, 1895, it was noticed that one of the Pliicker tubes which had been running for nearly three hours, had become strongly phosphorescent. The tube was fitted with platinum electrodes, and the helium had apparently been absorbed by the platinum sparked on to the walls of the tube. We observed the same phenomena to take place on several subsequent occasions, but only in the case of tubes with platinum electrodes.* Now, if helium is not a single gas, it must consist of a mixture of two or more monatomic gases, capable of mechanical separation, and it is possible that one of its constituents might be absorbed by the platinum faster than the other. At the end of September, 1895, I commenced some experimental work on this subject, with the view of separating the two or more possible constituents from one another. The results were negative. I employed in these experiments a piece of apparatus figured below (fig. 1). A large Pliicker tube, bent into a U -shape, has two side-tubes, A and B. The electrodes are of platinum, and project far into the tube ; the straight parts, which are of thick wire, and about 30 mm. * So far as I know, this phenomenon was first recorded by Professor Norman Lockyer (< Eoy. Soc. Proc.,' 1895, vol. 58, p. 193). 450 Mr. W. Travers. FIG. l. long, are protected by a sheath of thin glass tube, the spirals at their ends being of thin platinum wire. The side-tube A is connected, by means of a tube containing pentoxide of phosphorus, with an appa- ratus for the introduction of gases into vacuum-tubes (' Trans. Chem. Soc.,' 1895, p. 686). The tube B is connected with a tap on the Top- ler's pump. The apparatus was first thoroughly exhausted and heated by a Bunsen's flame, and then, after closing the tap on B, helium was introduced at about 3 mm. pressure. The electrodes were connected with the secondary terminal of a coil, and the cur- rent was turned on, making a the cathode. A deposit of platinum quickly appeared on the walls of the tube round a, and the following changes took place in the colour of the glow : — 1. Yellow, with slight tinge of red. 2. Bright yellow. 3. Yellowish- green. 4. Green ; green line very strong. 5. Green, with phosphorescence. 6. Phosphorescent vacuum ; spark passed between electrodes out- side the tube. The tube was then connected with the pump by opening the tap on B, but, as might have been expected, no trace of gas could be re- moved. The tap was again closed, and the tube was warmed care- fully with a Bunsen's burner. The gas was slowly given off from the platinum, and on passing the discharge, colour- changes were observed to take place in the glow, from green to yellow. From this experiment, it was obvious that the whole of the helium would be absorbed by the platinum splashed off, but it yet remained Some Experiments on Helium. 451 to be proved that the change in colour in the glow was not due to the absorption of the yellow constituent more quickly than the green one. The vacuum-tube used in the last experiment was again filled with helium to about 3 mm. pressure, and the discharge was passed till the glow had become green, and the green line had reached its maxi- mum intensity. Now, if any separation had taken place, the gas which had been absorbed by the platinum should contain a large pro- portion of the yellow constituent of helium, and should give a yellow glow in a vacuum-tube, even at low pressure. The remaining gas in the tube was, therefore, removed by pumping, and after closing the tap on B, the gas was driven off from the platinum, by warming with a Bunsen's flame. The current was then turned on, and a glow appeared of the green colour invariably shown by helium at low pressure. The change of colour in the tube during absorption of the helium is, therefore, to be entirely attributed to the lowering of the pressure. In describing these experiments I have used the term absorption in its general sense, as it is impossible to say at present whether we are dealing with a case of simple occlusion or not. The platinum, when it is deposited, is black and non-metallic in appear- ance, but, on heating, it assumes the colour and general character of ordinary platinum, and sometimes breaks away from the tube in thin scales. The change is probably the same as that which takes place when platinum-black is heated. In a few of my experiments, I used helium containing traces of hydrogen, nitrogen, and carbon compounds. In these cases I found that not only was the helium absorbed, but also the other gases, to a greater or less extent. Hydrogen is readily absorbed, and next in order come carbon compounds and nitrogen. Argon is taken up only in very small quantity ; in fact, this process serves as a method of separation of helium from argon, even when the helium is present to the amount of only 2 per cent. To carry out this separation, the gas is made to circulate at about 3 mm. pressure, through a vacuum-tube of the type used in the last experiment. To effect this, the Topler's pump is replaced by a Spren- gel's pump, arranged as shown in fig 2, to deliver the gas removed from the vacuum-tube back into the tube C. To regulate the supply of gas entering the apparatus, the tap F was carefully turned, till the gas bubbled slowly through the mercury contained in the small tube D. The tap E served as a by-pass during the preliminary pnmping- out of the apparatus, and was closed during the experiment. By carefully regulating the quantity of gas which entered the apparatus, and the rate of flow of mercury in the Sprengel's pump, ifc was possible to maintain a constant pressure in the apparatus for a long time. 452 Some Experiments on Helium, FIG. 2. To facilitate the absorption of the gases during the experiment, the vacuum- tube was kept cool by a water-jacket, G, closed at the bottom by a cork fitting tightly round the tube. When it was necessary to heat the vacuum-tube, the jacket could be loosened from the cork, and slipped up the side- tube B, which was bent round, and extended vertically for about 10 inches in a straight line with the vacuum-tube. The gas was made to circulate for about six hours, and at the end of that time the tap F was closed, the tap E was opened, and the apparatus thoroughly exhausted. The jacket Gr was then raised, and the gas expelled from the platinum by heat was pumped off. From mixtures containing very little helium, a small quantity of that gas was separated, mixed with a trace of argon. On the Gases enclosed in Crystalline Rocks and Minerals. 453 Kayser and Friedlander (' Chem. Zeitung,' vol. 9, p. 1529) have stated that in a vacuum-tube fitted with platinum electrodes, and containing atmospheric argon, the argon became absorbed by the deposited platinum, and the tube then showed certain of the helium lines. I have never been able to absorb argon to more than the very slightest extent, and though I have often had argon-tubes, which have become black, owing to the deposition of platinum, through which a powerful discharge has passed for many hours, I have never noticed any marked absorption. A specimen of argon, the lightest fraction obtained from Professor Ramsay's diffusion experiments, was treated in the manner just described. After several hours' circulation it was found that the gas absorbed by the platinum consisted only of argon, and no trace of helium could be detected. This process has also been applied to the analysis of the gases from certain mineral springs ; the results of these experiments form the subject of another paper. "On the Gases enclosed in Crystalline Rocks and Minerals." By W. A. TILDEN, D.Sc., F.R.S. Received December 19, 1896,— Read February 4, 1897. It has long been known* that many crystallised minerals contain gas enclosed in cavities in which drops of liquid are also frequently visible. The liquid often consists of water and aqueous solutions, occasionally of hydrocarbons, and not unf requently of carbon dioxide, the latter being recognisable by the peculiarities of its behaviour under the application of heat. The liquid supposed to be carbon dioxide has been found in some cases to pass from the liquid to the gaseous state, and therefore to disappear, and to return from gas to liquid at temperatures lower by two or three degrees than the critical point of carbon dioxide. This seems to indicate the presence of some incondensable gas, and as H. Davy found nitrogen in the fluid cavities of quartz, it seemed probable that the alteration of the critical point was due to that gas. My attention was drawn to this subject by the observation that Peterhead granite, when heated in a vacuum, gives off several times its volume of gas, consisting, to the extent of three-fourths of it» volume, of hydrogen (' Roy. Soc. Proc.,' vol. 59, p. 218). * The chief literature of this subject is contained in the following papers : — Brewster, < E. S. Edin. Trans.,' 1824, vol. 10, p. 1 ; ' Edin. J. Science,' vol. 6, p. 115 ; Simmler, < Pogg. Ann.,' vol. 105, p. 460; Sorby and Butler, 'Koy. Soc. Proe.,' vol. 17, p. 291 ; Yogelsang and Geissler, ' Pogg. Ann.,' vol. 137, pp. 56 and 257 : Hartley, ' C. S. Trans.,' 1876, vol. 1, p. 137, and vol. 2, p. 237, also 1877, vol. 1, p. 241. 454 Dr. W. A. Tilden, On the Gases enclosed Since this observation, I find that the presence of hydrogen in crystalline rocks has been recognised by other observers, notably by A. W. Wright (' Amer. J. Sci.,' Ser. 3, vol. 12, p. 171). In the course of a study of the gases from meteorites, Wright obtained from a certain " trap " rock, the origin and character of which is not stated, at a low red-heat, " about three-fourths of its volume of mixed gases, which were found to contain about 13 per cent, of carbon dioxide, the residue being chiefly hydrogen. Another specimen of trap containing small nodules of anorthite was examined at the request of Mr. G. W. Hawes, who had observed gas cavities in a thin section of the mineral prepared for microscopic examination. This gave off somewhat more than its own volume of gas, which was found to contain some 24 per cent, of carbon dioxide." Professor Dewar and Mr. Ansdell have also examined one or two rocks in the course of their researches on meteorites (' Roy. Inst. Proc./ 1886). They found that both gneiss and felspar, containing graphite, yield gas, which, upon analysis, was found to have the composition stated below. Occluded gas in volumes of the rock. , CO2. CO. H2. CH4. N2. Gneiss 5«32 , 82'38 2'38 13*61 0'47 1-20 Felspar T27 94'72 0'81 2'21 0'61 1-40 Dewar and Ansdell remark that " the small quantity of marsh gas^ no doubt, comes from the disseminated graphite, but the presence of the hydrogen is more difficult to explain, and requires further inves- tigation." I have lately been following up this question, and have obtained results which present some points of considerable interest. For materials I have been indebted chiefly to my colleague, Professor Judd, who has also supplied information as to the probable geological age of the specimens of rocks and minerals tested. All that I have examined yield permanent gas when heated in a vacuum. This gas varies in amount from a volume about equal to that of the rock or mineral to about eighteen times that volume. It usually consists of hydrogen in much larger proportion than that found by the observers just quoted, together with carbon dioxide and smaller quantities of carbon monoxide and hydrocarbons. Every specimen has been examined by the spectroscope for helium, but in no case could D3 be recognised, or any other line which would lead to a suspicion of the presence of this substance. The gas is very frequently, but not always, accompanied by water in notable quantities. The gas is apparently wholly enclosed in cavities which are visible in thin sections of the rock when viewed under the microscope, but as they are extremely minute, very little gas is lost when the rock is in Crystalline Rocks and Minerals. 455 reduced to coarse powder, and as a result of experiment in one or two cases, I find that practically the same amount of gas is evolved on heating the rock whether it is used in small lumps or in powder. In the first series of experiments undertaken with the object of a rapid survey of the materials, the gases were not completely analysed. They were collected, measured, the carbon dioxide removed by potash, and the residue examined by the spectroscope. When ignited in the air it always burned with a pale flame resembling that of hydrogen. The table (p. 456) shows the results of these experiments. A selection of these was then subjected to more careful and exact analysis. For this purpose fresh masses of the rock were taken, and the gas extracted in the usual way. The following are the results : — C02. CO. CH4. No. H2. Granite from Skye • 23-60 6-45 3-02 5 -13 61 '68 Grabbro from Lizard 5 '50 2-16 2-03 1-90 88-42 Pyroxene gneiss, Ceylon 77 '72 8-06 0-56 1-16 12-49 Gneiss from Seringapatam Basalt from Antrim . . 31-62 32-08 5-36 20-08 0-51 10-00 0-56 1-61 61-93 36-15 To account for the large proportion of hydrogen and carbonic oxide in these gases, it is only necessary to suppose tha,t the rock enclosing them was crystallised in an atmosphere rich in carbon dioxide and steam which had been, or were at the same time, in contact with some easily oxidisable substance, at a moderately high temperature. Of the substances capable of so acting, carbon, a metal, or a protoxide of a metal, present themselves as the most probable. The reduction of carbon dioxide or of water vapour by carbon gives rise to the formation of carbon monoxide, and if carbon had been the agent the proportion of this gas in the mixture must have been greater than is found to be the case. It is, of course, well known that carbon dioxide and water vapour are both dissociated at moderately high temperatures, but the greater part of the liberated oxygen recombines at lower temperatures, though a small portion may remain free in the presence of a large quantity of an indifferent gas or vapour. No free oxygen has been found in any of the gases analysed. Direct experiments, made with ferrous oxide (obtained by gently heating pure chalybite) and with magnetic oxide of iron, show that while the former, at a red-heat, decomposes both steam and carbon dioxide quite freely, liberating hydrogen and carbon monoxide, and becoming itself oxidised into m agnetic oxide ; the latter has no action VOL. LX. 2 M 456 On the Gases enclosed in Crystalline Rocks and Minerals. i f Volume of per volu of rock ip 9 co ^ o i> o » Q^ H/* i^j 5» CO 00 O •!>• CD Oi C^J 00 ^O -t"*» I?* i>» rH GO 00 00 C^ T"i 1Q O X>* CO O CO ® 00 GO "^ ^-O CO J>» O 00 O ***? O O ^O O ^p-l^rHtM ^2^*GO OiOCOXCOOCO^GOr-irH ^Tjt CO 00 CO 1O* O C^ ^}* CO CO O Ci 00 CO 00 (M iO CO GO *O (^J CO ^O ^1 !>• iHC l>.^ i^- Q 1> iH CO J : : ^ ''• 1 ^ : ::!,::::::: . ... 55 • 5 1 .§ ~ t» 5 " o • ... •g s s e8 - ^ ss o - . -g 5^ .§ erj ^2 - ^42 * ^ § c3 ^ .2" g -2 1 .2" g 5g Ksjfl gr^ss^s . s '' " « is " s II = ^£H ^. O tiJ • 9 — , 'CD nH • cs - -- S :-e.a s§ | : rtw : '• j§ : : : iir^'gosS'^^ «^l* •2 g J '« g .® -2 § 3 N -g ^ S £ -"s S : | a"1^-^^ g^gj- jS^'Sea^'oSSs"'' cj^^T'S CScS w?^ ?H JHSnCSj^Q f^ G S-! HJ i; . ^ Q ^ pp O*O O *£ ?i ^b G^fn Q DH cb ^O'PPH On Lunar Periodicities in Earthquake Frequency. 457 at all upon either steam or carbon dioxide. Magnetic oxide of iron is the final product of the action of steam or of carbon dioxide at a high temperature upon metallic iron : — 3Fe + 4H20 = Fe304 + 4H2. 3Fe + 4C02 = Fe304+4CO. Now, metallic iron has been detected in basalts and some other rocks by Andrews (' Brit. Assoc. Rep./ 1852, Sections, p. 34), and by other observers (e.g., G. W. Hawes, * Amer. J. Sci.,' Ser. 3, vol. 13, p. 33), and I have verified this observation in the case of the gabbro of Loch Coruisk. But it must be remembered that both the reactions indicated in the equations just given are reversible, and therefore the presence of metallic iron along with the magnetic oxide in such rocks cannot be taken by itself as final proof that the oxide and the associated gases, hydrogen and carbonic oxide, are the pro- ducts of the action of steam and carbon dioxide upon metallic iron. The presence of marsh gas in these rocks and the production of large quantities of hydrocarbonous gases, as well as liquid petroleum, in many parts of the earth's surface, tend to support the view, which is apparently gaining ground, that in the interior of the earth's crust there are large masses, not only of metal but of compounds of metals, such as iron and manganese, with carbon. Assuming the existence of such material, it is easy to conceive how, by the action of water at an elevated temperature, it may give rise to metallic oxides and mixtures of hydrogen with paraffinoid and other hydrocarbons. This view was put forward some years ago by Mendelejeff (" Principles of Chemis- try," Translation by Kamensky and Greenaway, vol. 1, 364 — 365), and it has lately received further support from the results of the study of metallic carbides, which we owe especially to Moissan (' Roy. Soc. Proc.,' vol. 60, 1896, pp. 156—160). " On Lunar Periodicities in Earthquake Frequency." By C. G. KNOTT, D.Sc., Lecturer 011 Applied Mathematics, Edinburgh University (formerly Professor of Physics, Imperial University, Japan). Communicated by JOHN MILNE, F.R.S. Received November 4, 1896,— Read February 4, 1897. (Abstract.) 1. Introduction. — The paper is a discussion of Professor Milne's Catalogues of 8331 earthquakes, recorded as having occurred in Japan, during the eight years 1885 to 1892 inclusive. These catalogues, forming vol. 4 of the ' Seismological Journal of Japan,7 458 Prof. C. G. Knott. are unquestionably the most complete ever constructed for an earth- quake-disturbed country. The discussion is really a working out of certain lines suggested in a paper on " Earthquake Frequency," communicated by me in May, 1885, to the Seismological Society of Japan, and published in vol. 9 of the ' Transactions ' of that Society. In that paper I pointed out the importance of subjecting earthquake statistics to some strict form of mathematical analysis, and gave a simple arithmetical process for separating the annual and semi-annual periods in earth- quake frequency. The results then obtained have been fully corro- borated by Dr. C. Davison in his paper " On the Annual and Semi- annual Seismic Periods " (' Phil. Trans.,' vol. 184, 1893) ; and my suggestion that the annual period is connected with barometric pressure is also strongly supported by Dr. Ferd. Seidl in his pam- phlet 'Die Beziehungen zwischen Erdbeben und Atmospharischen Bewegungen' (Laibach, 1895). The semi-annual period, which was first clearly brought into evidence in my earlier paper, does not admit of a very ready explanation. In my paper of 1885 I also considered in some detail the various tidal actions which might reasonably be supposed to have a determin- ing influence on earthquake frequency. From lack of material it was not possible at that time to make a satisfactory search for lunar periodicities ; but the remarkable fulness of information con- tained in Professor Milne's latest catalogues tempted me to under- take the labour involved in (first) tabulating the statistics in terms of lunar periods, and (second) analysing harmonically the tables so prepared. 2. The Lunar Daily and Half-daily Periods.. — In one of the cata- logues the earthquakes are classed according to district. Districts 1 to 6 lie on the N.E. and E. coasts of Japan, reckoning from the north; districts 6 to 11 on the S. coast; and 12 to 15 on the W. coast. Districts 6 and 7 are the most important, the former being the region including Tokyo and Yokohama, and the latter the region including Nagoya, which was the scene of the destructive earthquake of October 28, 1891. The investigation into a possible lunar daily period is conveniently based upon this classification into districts. Had that not been done by Professor Milne the labour involved in taking into account differences in local time would have been enormous ; for, to compare the time of occurrence of a recorded earthquake with the immediately preceding meridian passage of the moon, it was necessary to apply corrections for longitude and local time. The statistics for each district were, in the first instance, separated out and tabulated according to time of occurrence, estimated in hours after the immediately preceding passage of the moon. The method On Lunar Periodicities in Earthquake Frequency. 459 is explained in full in the paper. To lighten in some measure the labour of the harmonic analysis, certain districts were thrown together to form a district group. Table I contains the number of earthquakes in each district or district group, which formed the material for discussion. Table I. Number of Description of District. earthquakes. district. 1 397 ISTemura. 2—5 627 E. coast. 6 1432 S.E. corner. 7 3632 Nagoya, &c. 8 245 Kii Channel. 9—10 335 E. and S. of Kyushn. 11 384 W. of Kyushu." 12 112 i" W. coast of Main 13 U8 ( Tl A 14-15 145 J Of the tabulated numbers for each district or district group, over- lapping means of every successive five were taken, and these were divided by the mean of all. The numbers so obtained represent relative frequencies throughout the lunar day, and are given in Table II, which also contains a like series for all the earthquakes taken in combination. The most important are the frequencies for districts 6 and 7, and also for all combined. They are shown graphically in the figure (p. 461). Each series of numbers was then discussed by harmonic analysis in accordance with the Fourier expansion x = 1000 -f £ c« sin » ( — --- |- a, » = 1 \ 25 where x is 1000 times the relative frequency at time t, estimated in hours after the meridian passage of the moon, and where the amplitude c» and the phase an are to be calculated. The amplitudes and phases for the first four harmonics are given in Table IY. There is a tendency for the second harmonic amplitude to be greater than the first, while in half the number it is the greatest of all. As regards the times of occurrence of the maxima for the different harmonics, there is no regularity except perhaps in the case of the second harmonic. In four (1, 6, 7, 8) the maximum of the second harmonic falls within two hours of the half time between the upper and lower meridian passage of the moon. In the others it falls within two hours of the times of upper and lower meridian passage. 2 M 2 460 Prof. C. G. Knott. On Lunar Periodicities in Earthquake Frequency. 461 462 Prof. C. G. Knott. Table IV. — The Coefficients c and a, the amplitudes and phase- coefficients. District. DU 62 < &'«» VOL. LX. 2 o 482 Mr. G. U. Yule. On the Significance of Bravais9 Formula1 where &i&2, &'i&'2 are the regressions in the two cases. To which distribution are we, in such a case, to attribute the greater corre- lation ? Bravais' coefficient solves the difficulty, we may say, in one way, by taking the* geometrical mean of the two regressions as the measure of correlation. It will still remain valid for non-normal correlation. But there are other and less arbitrary interpretations even in the general case. Suppose that instead of measuring x and y in arbitrary units we measure each in terms of its own standard deviation, Then let us write X- = fy~ ......... ............. (5), and solve for p by the method of least squares. We have omitted a constant on the right-hand side, since it would vanish as before. We have, at once, That is to say, if we measure x and y each in terms of its own standard deviation, r becomes at once the regression of x on y, and the regression of y on x. The regressions being, in fact,, the funda- mental physical quantities, r is a coefficient of correlation because it is a coefficient of regression.* , Again, let us form the sums of the squares of residuals in equations (1) and (5). Inserting the values of 6l5 62, and />, we have — (7). Any one of these quantities, 'being the sum of a series of squares, must be positive. Hence r cannot be greater than unity. If r be equal to unity, or if the correlation be perfect, all the above three sums become zero. But can only vanish if x y -- = 0 's. In the general case, the first expression may be interpreted as the mean standard deviation of the ^-arrays from the line of regression, and the second expression as the mean standard deviation of the y-arrays from the line of regression. Otherwise we may regard — r2 as the standard error made in estimating x from the relation x = %, and as the standard error made in estimating y from the relation y = M, these interpretations being independent of the form of the correla- tion. (2.) Case of Three Variables. Let the three correlated variables be Xj, X2, X3, and let #l5 a?2, #3 denote deviations of these variables from their respective means. Let us write, for brevity, NV, S(>2a) = 2 o 2 484 Mr. G. U. Yule. On the Significance of Bravais Formula? Our characteristic or regression-equation will now be of tlie form 613 and 613 being the unknowns to be determined from the observations by the method of least squares. I have omitted a constant term on the right-hand side, since its least-square value would be zero as before. The two normal equations are now — or replacing the sums by the symbols defined above, and simplify- ing — = 612 ris = r12 = 0, r13= r« = 0, r13 = r12 = r13= ±r r13 = r13 = ± v/05= 0'707 r12 = + v/0'5 r12 .= — V 0 +1 — 1 0 1 and 2rz— 1 2rz— 1 and — 1 0 and 1 0 ., —1 One is rather prone to argue that if A be correlated with B, and B with C, A will be correlated with C. Evidently this is not necessary. A may be positively correlated with B, and B positively correlated with C, but yet A may, in general, be negatively correlated with C. Only, if the coefficients (AB) and (BC) are both numerically greater thanO'707, can one even ascribe the correct sign to the (AC) corre- lation. It is evident that one would, in general, expect to make a smaller standard error in estimating x\ from the two associated variables #2 and a'3, than in estimating it from one only, say a°2. But it seems desirable to provevthis specifically, and to investigate under what conditions it will hold good. The necessary condition is — ri22 + r138— 2r12r23r13 2 for Regression, fyc., in the case of Skew Correlation. 487 that is, 2— 2r12r13r23 > r12a-r122r132, > 0. ; or But (r13— ri2r23) is the numerator of /»H, the net coefficient of corre- lation between x\ and #3. Hence the standard error in the second case will be always less than in the first, so long as p13 is not zero. The condition is somewhat interesting. To take an arithmetical example, suppose one had in some actual case • > r<*> 'Jo $ • • r12= +0-8 >i t- r23= + 0-5 r13= +0-4. One might very naturally imagine that the introduction of the third variable- with a fairly high correlation coefficient (0*4) would con- siderably lessen the standard deviation of the x^- array ; but this is not so, for 0-4— (0-5X0-8) />13~ -/0-75XO-86" :°' sb the third variable would be of no assistance. III. Case of Four Variables. This case is, perhaps, of sufficient practical importance to warrant our developing the results at length as in the last. If a?i, %2, it's, a'4, be the associated deviations of the four variables from their respective means, the characteristic equation will be of the form (14). The normal equations for the fr's are, in our previous notation, Hence r12 r,3 r24 r13 1 rsi r24 r23 1 (15), and so on for the others, b^ 613, &c., we may call the net regressions of xi on ajz, ajj on a?3, &c., as before. By parity of notation^we have 488 On Bravais' Formula in the case of Skew Correlation. 12 ?*23 ?*24 1 fsi ^34 1 and we may again call tlie net coefficient of correlation between Xi and #2. Expanding the determinants, we have, in fact, ........ (16). There are six such net coefficients, />12, />13, /314, p^ pUt pu. The above values of the regressions are again those usually obtained on the assumption of normal correlation.* The net correlation pn becomes, on that assumption, the coefficient of correlation for any group of the %i wz variables associated with fixed types of #3 and #4. If we write we have, after some rather lengthy reduction, where 1 4) J In normal correlation, o-!-/! — ^i2 is the standard deviation of all ajr arrays associated with fixed types of xz, »3, and #4. In general corre- lation, it is most easily interpreted as the standard error made in estimating 0*1, by equation (14), from its associated values of x2, #3, and x^ As in the case of three variables, the quantity R may be considered as a coefficient of correlation. It can range between +1, andean only become unity if the linear relation (14) hold good in each indi- vidual instance. We showed at the end of the last section that the standard error made in estimating x1 from the relation *' Professor Pearson, " Eegression, Heredity, and Panmixia." ' Phil. Trans.,' > Yol. 187 (1896), p. 294. Mathematical Contributions to the Theory of Evolution. 489 was always less than the standard error when only xz was taken into account, unless />13 = 0. We may now prove the similar theorem that when we use three variables, xzi «3, *i, on which to base the estimate, the standard error will be again decreased, unless Pli = 0. The condition that S(?r), in our present case, shall be less than S(r2) in the last, is, in fact, 22 + r132 + rM»- n^-r^Vu2 -r13V242 -| Wl— J 132— 2r12r13r23)(l--r232— r242— ?' This may be finally reduced to — 0, that is />142 > 0. The treatment of the general case of n variables, so far as regards obtaining the regressions, is obvious, and it is unnecessary to give it at length. We can now see that the use of normal regression formulae is quite legitimate in all cases, so long as the necessary limitations of inter- pretation are recognised. Bravais' r always remains a coefficient of correlation. These results 1 must plead as justification for my use of normal formulas in two cases* where the correlation was markedly non-normal. " Mathematical Contributions to the Theory of Evolution. — On a Form of Spurious Correlation which may arise when Indices are used in the Measurement of Organs." By KARL PEARSON, F.R.S., University College, London. Re- ceived December 29, 1896,— Read February 18, 1897. (1) If the ratio of two absolute measurements on the same or different organs be taken it is convenient to term this ratio an index. If u =/!(#, y) and v =/2(^, y) be two functions of the three variables a/*, 2/, 0, and these variables be selected at random so that there exists no correlation between #,?/, y,z, or z,x, there will still be found to * ' Economic Journal,' Dec., 1895, and Dec., 1896, " On the Correlation of Total Pauperism with Proportion of Out-relief." 490 Prof. Karl Pearson. exist correlation between u and vt Thus a real danger arises when a statistical biologist attributes the correlation between two functions like u and v to organic relationship. The particular case that is likely to occur is when u and v are indices with the same denominator for the correlation of indices seems at first sight a very plausible measure of organic correlation. The difficulty and danger which arise from the use of indices was brought home to me recently in an endeavour to deal with a consider- able series of personal equation data. In this case it was convenient to divide the errors made by three observers in estimating a variable quantity by the actual value of the quantity. As a result there appeared a high degree of correlation between three series of abso- lutely independent judgments. It was some time before I realised that this correlation had nothing to do with the manner of judging, bat was a special case of the above principle due to the use of indices. A further illustration is of the following kind. Select three num- bers within certain ranges at random, say #, y, z, these will be pair and pair uncorrelated. Form the proper fractions xfy and z\y for each triplet, and correlation will be found between these indices. The application of this idea to biology seems of considerable importance. For example, a quantity of bones are taken from an 088uariumt and are put together in groups, which are asserted to be those of individual skeletons. To test this a biologist takes the triplet femur, tibia, humerus, and seeks the correlation between the indices femur / humerus and tibia / humerus. He might reasonably conclude that this correlation marked organic relationship, and believe that the bones had really been put together substantially in their individual grouping. As a matter of fact, since the coefficients of variation for femur, tibia, and humerus are approximately equal, there would be, as we shall see later, a correlation of about 0'4 to 0'5 between these indices had the bones been sorted absolutely at random. I term this a spurious organic correlation, or simply a spurious correlation. I understand by this phrase the amount of correlation which would still exist between the indices, were the absolute lengths on which they depend distributed at random. It has hitherto been usual to measure the organic correlation of the organs of shrimps, prawns, crabs, Ac., by the correlation of indices in which the denominator represents the total body length or total cara- pace length. Now suppose a table formed of the absolute lengths and the indices of, say, some thousand individuals. Let an " imp " (allied to the Maxwellian demon) redistribute the indices at random, they would then exhibit no correlation ; if the corresponding absolute lengths followed along with the indices in the redistribution, they also would exhibit no correlation, Now let us suppose the indices not to have been calculated, but the imp to redistribute the abso- Mathematical Contributions to the Theory of Evolution. 491 lute lengths j these would now exhibit no organic correlation, but the indices calculated from this random distribution would have a correlation nearly as high, if not in some cases higher than before. The biologist would be not unlikely to argue that the index correla- tion of the imp-assorted, but probably, from the vital standpoint, impossible beings was " organic." As a last illustration, suppose 1000' skeletons obtained by distribut- ing component bones at random. Between none of their bones will these individuals exhibit correlation. Wire the spurious skeletons together and photograph them all, so that their stature in the photo- graphs is the same ; the series of photographs, if measured, will show correlation between their parts. It seems to me that the biologist who reduces the parts of an animal to fractions of some one length measured upon it is dealing with a series very much like these pho- tographs. A part of the correlation he discovers between organs is undoubtedly organic, but another part is solely due to the nature of his arithmetic, and as a measure of organic relationship is spurious. Returning to our problem of the randomly distributed bones, let us suppose the indices f emur/humerus and tibia/humerus to have a correlation of 0'45. Now suppose successively 1, 2, 3, 4, &c., per cent, of the bones are assorted in their true groupings, then begins the true organic correlating of the bones. It starts from 0'45, and will alter gradually until 100 per cent, of the bones are truly grouped. The final value may be greater or less than 0'45, but it would seem that 0*45 is a more correct point to measure the organic correlation from than zero. At any rate it appears fairly certain that if a biologist recognised that a perfectly random selection of organs would still lead to a correlation of organ-indices, he would be unlikely to accept index-correlation as a fair measure of the rela- tive intensity of correlation between organs. I shall accordingly define spurious organic correlation as the correlation which will be found between indices, when the absolute values of the organs have been selected purely at random. In estimating relative correlation by the hitherto usual measurement of indices, it seems to me that a statement of the amount of spurious correlation ought always to be made. (2; Proposition L — To find the mean of an index in terms of the means, coefficients of variation, and coefficient of correlation of the two absolute measurements.* Let a?!, a^, a?3, a?4 be the absolute sizes of any four correlated organs ; mlt Wz, Wa, m4 their mean values ; 22 + v32— V232)/(2v2v3) = —0'0543. Height and breadth ; r21 = (y + v?— V212)/(2i71v2) = 0'1243, . This is the first table, so far as I am aware, that has been published of the variation and correlation of the three chief cephalic lengths.f It shows us that there is not at all a close correlation between these chief dimensions of the skull, and that a small compensating factor for size is to be sought in the correlation oE height and length, i.e., while a broad skull is probably a long skull and also a high skull, a high skull will probably be a short skull, and a low skull a long skull. Without substituting the values of v1} v2, t'3, ri2, ?'i3, r23 in (v), we can find />, or the correlation between breadth/length and height/length indices from ; P = (V132-hV232-Y122)/(2Y13V23). This follows at once from the general theorem given in my memoir on " Regression, Panmixia, and Heredity," ' Phil. Trans./ vol. 187, A, p. 279, or by substitution of the above values of rt2, ri3, r& in (v), we find : ; P == G'4857, If we calculate from (vi) the correlation between the same cephalic indices on the hypothesis that their heights, breadths and lengths are distributed at random, i.e., that our "imp "-has constructed a number of arbitrary and spurious skulls from Professor Ranke's measurements, we find : PQ — 0-4008. It seems to me that a quite erroneous impression would be formed of the organic correlation of the human skull, did we judge it by the magnitude of the correlation coefficient (O4857) for the two chief * All the absolute measures given are in millimetres, and the coefficients of variation are 'percentage variations, i.e., they must be divided by 100 before being used in formulae (i), (ii), and (iii). f I hope later to treat correlation in man with reference to race, sex, and organ, as I have treated variation. 496 Prof. Karl Pearson. cephalic indices, for no less than 0'4008 of this would remain, if we destroyed all organic relationship between the lengths on which these indices are based. Example (d). To find the spurious correlation lettueen the indices femur jliumerus and femurj tibia. The following results have been calculated* from measurements made by Koganei on Aino skeletons. (See ' Mittheilungen aus der medicinischenFacultat der K. J. Universitat, Tokio,' Bd. I. Tables.) I have kept the sexes apart although there are but few of each. 3 Skeletons. Number = 40 to 44. Measurements in centimetres. Femur, F : ml = 40'845, E. BtNDIN OCT1 Q a L718 v.60 Physical & Applied Sci. Serials Royal Society of London Proceedings PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY