ous gases, sealed up in Geissler tubes, have been experimented upon, the discharge from a Rubmkorff coil being allowed to traverse the gas. Changes occur in the appearance of the luminous discharge where the magnet is excited; these changes are accompanied by a change in the resistance of fered to the current by the gas. Thus a tube containing hydrogen permitted the passage of a current marking twentyfive degrees on the galvanometer when the magnet was not excited, but when excited the galvanometer reading was forty degrees. It seems to be a law that the augmentation in the intensity of the current is greater with a gas which is a good conductor than with one which is a bad conductor.-12 A, XI., 19.

NEW SOURCE OF MAGNETISM.

M. Donati Tommasi is authority for the statement that if a current of steam at a pressure of from five to six atmospheres is passed through a copper tube of two to three millimeters in diameter, which is spirally coiled about an iron cylinder, the latter is magnetized so effectually that an iron needle, placed at the distance of some centimeters from the steam magnet, is strongly attracted, and remains magnetic so long as the steam is allowed to pass through the copper spiral.—6 B, XV., 1875.

MAGNETIC PERMEABILITY OF IRON, NICKEL, ETC. Mr. Rowland, of Troy, New York, in a paper on the magnetic permeability of nickel and cobalt, states that the views of the English and German philosophers as to the nature of force have given rise to different ways of looking upon magnetic induction. Thus, the Germans would say that this action was due in part to two causes-the attraction of the coil and the magnetism induced in the iron by the coil; the English, following Faraday, on the other hand, would consider the substance in the helix as merely conducting the lines of force, so that no action would be exerted directly on the compass needle by the coil; but the latter would only affect it in virtue of the lines of force passing along its interior, and so there could be no attraction in a perfectly vacant space. According to the first theory, the magnetization of the iron is represented by the excess of the action of the

electric magnet over that of the coil; while by the second
theory, when the coil is very close around the iron, the whole
action is due to the magnetism of the iron. The natural
unit of magnetism to be used in the first theory is that quan-
tity which will repel an equal quantity at a unit's distance
with a unit of force. On the second theory, it is the number
of lines of force which pass through a unit of surface when
that surface is placed in a unit field perpendicular to the
lines of force. As the result of his novel and very important
researches on the effect of heat on magnetism, Rowland states
that if it were possible for the magnetism of substance to
attain a maximum value, the co-efficient of magnetism by in-
duction would become, first, zero, and then negative, and the
substance would then become diamagnetic for very high mag-
petizing forces. This principle, announced independently by
Maxwell and Rowland, lacks as yet the confirmation of ob-
servation, although not contrary to our experience. Our
principal hope of confirming it by observation consists in
heating some body, and then subjecting it to a very high
magnetizing force, for Rowland has shown in the case of iron
and nickel the maximum of magnetization of nickel and of
iron decreases as the temperature rises, at least between the
limits of zero and 220° Centigrade. He finds from observation
that if nickel is heated from 15° to 220° Centigrade, the mag.
netization will increase if the magnetizing force is small, but
will decrease if it is large. In general, as the magnetizing
force is increased, the resistance of iron, nickel, and cobalt to
magnetization decreases, until a minimum is reached when
the metals have attained a magnetization equal to from 24 to
30 per cent. of their maximum of magnetization, and after
that resistance increased indefinitely.-7A, XLVIII., 32.

IMPROVEMENTS IN THE GRAMME MAGNETO-ELECTRIC MACHINE.

The magneto-electric machine invented by Gramme, which has within the past two years become quite famous, las received an important improvement in that ordinary magnets have been replaced by the plate magnets invented by Jamin, which give it a great advantage, not only because of the greater force for the same weight, but because of the extreme facility of their construction. These plate magnets can be built up and taken apart in a few minutes, an ex

tremely valuable feature when one is obliged to experiment in order to determine the strength of the current necessary for accomplishing a certain work, and one equally valuable to the physicist who may desire to elucidate obscure points in the theory of the machine. The Gramme machine has been still further improved by combining with its peculiar features the construction due to Wild and Ladd, by which an immense magnetic power is developed from a very slight initial movement of magnetism; by this means an instrument has been produced by which the same electric tension is attained with a velocity one half as great as that originally necessary. In the course of the numerous improvements that Mr. Gramme has made in his original machine, his latest construction seems to leave nothing to be desired. The number of electro-magnets and of coils is now reduced, from six and twelve respectively, to two and four. The ring is virtually doubled, giving far more facility in the application of the same machine to very different objects, such as galvanoplasty, lighting, heating, etc.; in the machines, as originally constructed with a simple ring, each one was only convenient for use for the immediate purpose for which it was designed and proportioned.—13 B, III., 139.

THE FRICTION AND THERMAL CONDUCTIVITY OF GASES. In a memoir by Messrs. Kundt & Warburg on the friction and thermal conductivity of gases for heat, the authors endeavor to investigate the accuracy, at high temperatures and low densities, of the laws deduced by Maxwell, Meyer, Loschmidt, Stefan, and Boltzmann, which for ordinary temperatures and densities hold good in gases; they find, first, that the co-efficient of sliding friction between moving gas and a fixed plane has a determinate value dependent on the nature of the gas, so long as this is present in layers thicker than fourteen times “the mean length of path of the molecules' as defined by the kinetic theory of gases; the co-efficient is also inversely proportional to the pressure. Second, the absolute value of the co-efficient of sliding friction is found to be 0.7 xl, on the assumption that the molecule of gas is reflected from the moving surface used in the apparatus with velocities of translation equal to those of the surface itself. For air at 760 millimeters, 1=0.000083 millimeter,

H

therefore the co-efficient of friction should be 0.000058 X pressure; but actual observations give a result very nearly twice as great. Hence it is concluded that in the striking of the molecules against the walls, their velocities are not completely equalized. The absolute co-efficient of friction for the air is given by these authors at 0.000189, being exactly midway between the four previous determinations made by Graham, Maxwell, Meyer, and Puling. The co-efficients of friction for hydrogen and for carbonic-acid gas were determined by them to be respectively 0.488 and 0.806 (that of the air being 1), agreeing closely with the values deduced from the observations of Graham. The co-efficient of friction for pure steam at a temperature of 15° Centigrade resulted about one half of that of air. The investigation into the dependence of the co-efficient of friction on the density or barometric pressure of the gas shows that the diminution of friction with pressure is greater the rarer the layer of gas. Further experiments bearing upon the kinetic theory of gases were made by Messrs. Kundt & Warburg in that they attempted to determine the co-efficient of conductivity for heat. Their approximate result for the atmosphere is one eleventh less than that deduced a few years ago by Stefan; and from these same observations there resulted also the value of the radiating power of glass, which agreed nearly with that of Lehnebach. The variation of the radiating power with the temperature does not seem to them to have been reliably determined in the classical work of Dulong and Petit.-Monatsbericht der K. Akademie von Preussen, Berlin, 1875, 160.

THE CONNECTION BETWEEN FLUORESCENCE AND ABSORPTION.

Dr. Sorby, President of the Royal Microscopic Society, states that he has been surprised to find that some of those who have paid considerable attention to such subjects have so far misunderstood the question as to suppose that the light of fluorescence consists of rays which are, as it were, reflected by this solution, and do not penetrate through it, so that the spectrum of the fluorescence would show a bright band in the same place as some dark bands seen in the spectrum of the transmitted light. This is certainly an error, and his own observations agree more nearly with Lubarsch,

who shows that of eight different substances the spectrum of the light of fluorescence extends some distance on the red end side of the principal absorption band in the spectrum of transmitted light; so that the spectrum of fluorescent substances can never contain rays which are more refrangible than those which are most readily absorbed by a very dilute solution. This, although a very general rule, yet has some decided exceptions. In some substances, under strong illumination, the light of fluorescence does contain rays of greater refrangibility than those most readily absorbed by a dilute solution, and extends from the red end a little beyond the centre of the main absorption band. A number of little known and interesting fluorescent solutions are quoted by Sorby in illustration of his remarks.- Monthly Micr. Journal, p. 161.

THE ISOCHRONISM OF THE BALANCE SPRING. William D. Glasgow, in a short article in the Horological Journal, on balance springs, states that the isochronism of the balance spring of a watch is a subject bristling with controversy. There are some who say that every spring must be isochronized; others that every length of spring has its isochronous point of suspension; others that mere length has absolutely nothing to do with isochronism. Mr. Glasgow holds that length has every thing to do with it, as shown by his own experiments. Too short a spring, whatever may be its form, will make the short arcs of the balance's vibrations to be performed in a less time than the long arcs. Thus a spring with ten turns may be too short, and will lose in the short arcs and gain in the long arcs. A spring of two turns will be too long, and will describe its longer arcs in too short a period. The best length for a flat spring is, he finds, fourteen turns; but a flat spring, although the most common, is also the worst form, as it does not expand and contract properly. It will assist the action in this spring if it is always a little small, as this gives more freedom to the portion of the coil next to the stud. The Breguet spring, although differing very little in form from the flat spring, is essentially different in action and principle, having perfect freedom to expand in a circle all around. From twenty to twenty-five turns is, he finds, the