in connection with Figs. 33 to 36. This action continues until side B is heavier than side B', when the reversed unbalanced condition of the car body causes a reversed precessional movement of the gyrostat wheels. This reversed precession continues steadily and unhastened so long as the heavy side B is balanced by the contact of the idle wheel 'with the table L', that is, until the projecting ends of the axles of spin A and A' are brought forwards (in the figure) into the plane of the paper. Then the continuation of the reversed precession brings the axle A' upon the table H', the reversed precessional motion is then hastened, and this hastened precession raises the side B and lowers the side B' of the car-frame, thus bringing the car-frame into its initial unbalanced condition (side B' heavier than side B). The above-described action is then repeated, and so on.

The stability of the Brennan car is due to the hastened precession which is caused by rolling action of one or the other of the projecting axles of spin upon the tables H and H', while the axles of spin are departing from a line at right angles to the length of the car, and to the steady and unhastened precession, while the axles of spin are moving towards a line at right angles to the length of the car. The hastened precession on the one hand quickly alters the condition of balance of the car so as to limit the departure of the axles of spin from a line at right angles to the length of the car, and the steady and unhastened precession, on the other hand, insures the complete return of the axles of spin to a line at right angles to the length of the car.

The hastened precession is accomplished with great friction losses by the rolling axles A and A' in Fig. 37, and it is reported that Brennan is working upon an automatic motor-driven mechanism to produce the hastened precession without exhausting the energy of the gyrostat wheels.

Two devices like Fig. 37 with their rocker-axles at right angles to each other would hold a one-legged body in equilibrium; indeed, such a double mechanism would make it possible to use a one-wheeled car, but the wheel would have to have a deep double flange to make it roll along a rope or rail. Such a one-wheeled car, a sort of hyper-wheelbarrow car, would be of no value for practical use, and, indeed, most of us believe that Brennan's two-wheeled car is nothing more than a scientific toy.

CALCULATION of Torque-REACTION DUE TO PRECESSION

Let n be the revolutions per second of a spinning wheel, P the revolutions per second (or the fraction of a revolution per second) of the axis of spin due to the precession, and K the moment of inertia of the spinning wheel in pound X feet squared. Then the torque reaction is equal to 42nPK poundal-feet or nPK pound-feet.*

See Franklin and MacNutt's "Elements of Mechanics," p. 150.

JOSIAH WILLARD GIBBS AND HIS RELATION TO MODERN SCIENCE. III

BY FIELDING H. GARRISON, M.D.

ASSISTANT LIBRARIAN, ARMY MEDICAL LIBRARY, WASHINGTON, D. C.

Catalysis, Colloids and Chemical Purity.-When chemical change can be produced in a system by the mere presence of small quantities of another substance which itself usually remains unchanged at the end of the process, such an effect is called catalysis and the agent employed a catalytic agent. Of the varied aspects of catalytic processes we have different examples in the decomposition of substances by the presence of finely divided metals like platinum or colloidal nickel, in the rapid evolution of oxygen from potassium chlorate when a small quantity of manganese dioxide is present, in the solution of insoluble chromic chloride through the mere presence of chromous chloride, in the inversion of cane sugar by acids, in the saponification of fats and esters, in the synthesis of indigo by oxidation of naphthalin, in the standard manufacture of sulphuric acid in the leaden chambers and the later improvements of the method through the presence of platinum or ferrous oxide, in catatyptic photography without light, in the reversible physiologic and therapeutic action of the animal and vegetable ferments and enzymes, in the synthesis of nuclein during the development of the embryo, and in the pathologic effects of poisons, venoms and the toxins of disease. Many theories of catalytic action have been advanced, of which the earliest and most original is that of Leibig. Liebig supposes catalysis to be due to the fact that the catalytic agent has power, like that of a tuning fork, to set up sympathetic molecular vibrations in the substance acted upon, producing chemical change. This theory has been proscribed by Ostwald because, being a figment of the mind, it is neither capable of proof nor susceptible of refutation, leading the subject into a blind alley, from which further scientific advance is impossible.101 It has therefore remained, like Hamlet's father, "quietly inurned," as a beautiful, imaginative hypothesis which we can neither prove nor disprove. Of other theories of catalysis the most important is that of Ostwald himself, summed up in his famous definition: A catalytic agent is one which modifies the velocity of a chemical reaction without appearing in its final process. This statement introduces two new ideas, the notion of infinite swiftness and infinite slowness in chemical change and the fact that catalytic change may be brought about by a series of intermediate reactions. It will be seen that Ostwald's definition is elastic enough to include as 101 Ostwald, "Ueber Katalyse," Leipzig, 1902.

catalytic agencies such physical forces as light, electricity, extremes of heat or cold or the action of living tissues, and from this point of view the explosion of a cartridge or a charge of dynamite by percussion, the decomposition of water by electrolysis and its synthesis by the electric spark, the effects of light in photography and in healing disease, the wonderful thermodynamic effects of Henri Moissan's electric furnace, the occasional changes of food in cold storage, are further examples or analogues of catalytic action, and this is all we know of its physical nature. As to a dynamic explanation of how catalysis takes place, we have not got beyond the familiar jest of the laboratories: "Q. What is catalysis? A. Action by contact. Q. What is action by contact? A. Catalytic action." Gibbs's treatment of the subject is interesting as affording a mathematical criterion of what catalysis is and what it is not. It will be remembered that when the entropy of an isolated chemical system, say a bar of steel, has attained a maximum or its free energy a minimum value, the final state of the substance in question has been called by Gibbs a "phase of dissipated energy,' implying that it has become physically and chemically inert, so that its equilibrium will not be sensibly disturbed by the presence of other substances or by such small physical agencies as an electric spark. But when the proportion of the proximate components of the substance in connection with its pressure and temperature is such that it does not constitute a phase of dissipated energy, the contact of a very small body or physical agency may produce energetic changes in its mass which do not stop short of complete dissipation. This is catalysis, and Gibbs's definition of a catalytic agent-one capable of reducing a substance to a phase of dissipated energy without limitation as to their relative proportions—is characteristic of the mathematician. A chemical system at constant temperature has several states of equilibrium corresponding to different minima of its isothermal potentials, and on the solid diagrams of Gibbs these minima are valleys at the bottoms of sloping curves. The effect of a catalytic agent on the diagram is to obliterate the ridge between two depressions representing different states of equilibrium on the free energy surface. This means that a system disturbed by a catalytic agent may pass from a higher to a lower minimum of free energy, but never from a lower to a higher unless acted upon by external forces of considerable magnitude. When the lowest minimum of free energy, indicated by the lowest depression on the diagram, has been attained, the substance can no more leave the final phase of dissipated energy than an inert body can be made to go up a hill without the intervention of external forces. On Gibbs's showing, the phase of dissipated energy is the criterion of catalytic action, the condition for which is that the substance acted upon should not have attained such a phase, while the forces operating flow, as in other mechanical, thermal, chemical or electric happenings, from higher to lower potentials. The accuracy of this reasoning is

borne out by Emil Fischer's researches in structural chemistry, which show that the intrinsic stability of chemical systems is usually such that it can not be disturbed by "intramolecular wobble," chemical change being brought about by extramolecular or catalytic influences. The mathematical treatment of catalysis gives us a deeper insight into phenomena which no one has as yet succeeded in explaining. "We have not," says Bancroft, "the first suggestion of an adequate theory of catalysis so essential to a better understanding of chemistry and of life itself. A true theory of catalysis will enable us to solve the problem of the transmutation of the elements, of which we have already had examples in the substances derived from radium, and the recent derivation of tellurium from copper by Sir William Ramsay. The action of animal and vegetable protoplasm is probably catalytic and the chemist can now make some vegetable substances, such as indigo or alizarine, more cheaply and purely than the plants themselves do. Could we substitute inorganic catalyzers for the vegetable enzymes and ferments in all cases, we might, as Bancroft points out, duplicate everything except the plant itself. Recently Loeb has interpreted the fact that some eggs can be developed by osmotic pressure alone, while others require fertilization, by the explanation that, in the former class the nuclein synthesis, which is necessary for segmentation, is started within the nucleus as a catalytic process, one of the products of the reaction being the catalyzer itself; while eggs requiring fertilization are such that the necessary nuclein synthesis must be started by some external catalytic agency.1 y.102 Again catalysis is the key to the causes and treatment of infectious diseases, the toxins and antitoxins of which are probably colloidal catalytic agents. A few drops of such a colloid as cobra venom will rapidly reduce a living animal body to a definite phase of dissipated energy, as far as its vital activity (or "free energy") is concerned, and such catalysts as colloidal metals, which Bredig has shown to act exactly like the ferments and enzymes, can themselves be "poisoned" or rendered inert by other substances, just as toxins, venoms and poisons can be neutralized by antitoxins or other antidotes. Gibbs did not discuss colloids explicitly, because substances of such indefinite or irregular formation do not admit of mathematical treatment as such, but the physics of what we know of their intimate structure is implicit in his chapters on chemical conditions obtaining at surfaces of discontinuity. Colloids are semi-solid substances, and colloidal solutions are "pseudo-solutions," being suspensions of minute, discrete particles of matter which are not true solutions, in that they obstruct the passage of light, while neither the freezing point nor the vapor tension of the solvent can be sensibly lowered. Graham thought of colloids as dynamic phases of matter, possessing internal energy, while crystalloids are static and inert. The former include reversible colloids like gelatine which, heated with warm water, will upon cooling solidify 102 Loeb, Science, 1907, N. S., XXVI., 425-37.

into a gel," and redissolve upon heating into a colloidal solution or "sol"; and irreversible colloids, which, when heated with warm water, will coagulate at once into an unchangeable precipitate. Living protoplasm, as Darwin has shown in his experiments upon Drosera and other plants,103 acts exactly like a reversible colloid. Dead protoplasm, such as a coagulated blood clot, is an irreversible colloid consisting of a fixed network, the meshes of which contain the "sol." There is no evidence of internal structure in living protoplasm, and Hardy supposes that structure in dead protoplasm is produced by submortem or postmortem changes associated with coagulation. Whether the phase rule can be applied to colloids is still an open question bound up with the complex nature of bodies of which we know so little. But recently Siedentopf and Zsigismony have shown that colloidal metals, organic ferments and enzymes are systems in two phases of vast surface tension consisting of suspensions of ultra-microscopic particles acted upon by chemical, thermodynamic and electric potentials. Of such suspensions animal and vegetable bodies are largely made up, protoplasm being a sort of microscopic emulsion, the physiological action of which seems to be bound up with chemical, thermal, electric and osmotic changes between its semi-permeable membranes and surfaces of discontinuity and the various surface tensions and surface energies derived from the free energy of chemical or electric change. If we conceive of colloidal solutions as made up in this way, each tiniest particle being an ultramicroscopic furnace, retort or battery in itself and carrying a definite charge of electricity, we can understand how Liebig's theory of sympathetic vibrations might be applicable to colloidal catalysis at least, and how finely divided metals, serpent venoms or the excretions of microorganisms can produce the extraordinary effects they do. In close connection with the theory of catalysis is the nature of chemical purity and the fact that chemical changes rarely proceed directly to their final product, but usually pass through a series of intermediate stages. For a long time chemists have noticed that absolutely dry or pure substances will not interact directly upon each other, but the cooperation of a third substance is necessary for chemical change. Dried chlorine does not of itself act upon copper and other metals, but the presence of a little moisture will cause it to act upon them at once. A mixture of carbonic acid and oxygen is not explosive when thoroughly dry, but the slightest trace of steam will cause an explosion. The rapid solubility of zinc in sulphuric acid depends upon impurities in the former. Ebullition depends largely upon gaseous impurities in the boiling substance. Absolutely pure or distilled water has no digestive value, but, by its absorptive power, acts as an irritant or poison to the lining membrane of the stomach. Traces of moisture or other impurities have therefore a marked catalytic effect, a theory of catalysis which was first advanced as early as 1794 by Mrs. Fulhame in her Essay 103 Darwin, "The Power of Motion in Plants," passim.