In the teaching of photography to students the tendency has been to lay great emphasis on the chemistry of the subject while the physics of photography, which is at least as important as the chemistry, has too often been ignored. Dr. Roebuck has approached the subject from the standpoint of the physicist rather than from that of the chemist, with the result that in this book there is given a clear and valuable exposition of the elementary principles of sensitometry, that is, of the properties of photographic material and its behavior during exposure and development.

The chemistry of the book is distinctly weak, there is practically no discussion of the chemistry of development, and the few equations given for the action of developers are very much open to question. There are also a few obvious errors in chemistry such as the statement that Stas was a German, or that hydrochloric acid can be added to silver nitrate in order to produce an acid emulsion.

In the portion of the book dealing with general theory the author commences with a brief chapter on the historical development of the subject and then deals with the sensitometry of the gelatine dry plate. A short chapter then discusses the subject of color sensitiveness, and another, theories of the latent image. Further chapters deal with negative defects, a very practical chapter indeed, positive processes, lenses, color photography, and the general principles of composition.

The second part of the book consists of a laboratory manual containing a series of experiments to be performed by the student. This will be very valuable to any teacher arranging a course in photography and a student who has worked thoroughly through the course, repeating the more elementary portions several times, will have had a good training in the elements of the subject.

On the whole the book forms a valuable addition to the scanty list of modern works on photography and is to be recommended to all those who are interested in the scientific side of the subject.

ROCHESTER, N. Y.

C. E. K. MEES

THE PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES

THE first number of Volume 4 of the Proceedings of the National Academy of Sciences contains the following articles:

The Basal Katabolism of Cattle and Other Species: Henry P. Armsby, J. August Fries and Winfred W. Braman, Institute of Animal Nutrition, the Pennsylvania State College. The results show that the basal katabolism of different species is substantially proportional to their body surface.

The Location of the Sun's Magnetic Axis: F. H. Seares, A. van Maanen and F. Ellerman, Mount Wilson Solar Observatory, Carnegie Institution of Washington. In extension of the work of George E. Hale, a large number of observations were undertaken to determine the position of the sun's magnetic axis, which is found to lie near the axis of rotation at an inclination of about 6°, and to revolve about the axis of rotation in about 32 days.

Resonance and Ionization Potentials for Electrons in Cadmium, Zinc and Potassium Vapors: John T. Tate and Paul D. Foote, University of Minnesota and Bureau of Standards. The results agree within the limits of experimental error with the values as calculated from the quantum relation hy=eV, where is the frequency of the single radiation in the case of resonance potentials or the limiting frequency of the series of radiations in the case of ionization potentials.

The Validity of the Equation P=dv/dT in Thermo-Electricity: Edwin H. Hall, Jefferson Physical Laboratory, Harvard University. The equation is known to be unverified experimentally. The author gives a brief, critical discussion of the validity of some theoretical proofs by which the equation has been deduced.

On the Equations of the Rectangular Interferometer: Carl Barus, Department of Physics, Brown University. A discussion under the under the headings of: Auxiliary Mirror, Rotating Doublet, Ocular Micrometer, Collimator Micrometer.

The Resolving Powers of X-Ray Spectrometers and the Tungsten X-Ray Spectrum: Elmer Dershem, Department of Physics, University of Illinois. The theory of resolving power is given with the results of experiments on tungsten, in which the endeavor was made to obtain as high a resolving power as possible.

Note on Methods of Observing Potential Differences Induced by the Earth's Magnetic Field in an Insulated Moving Wire: Carl Barus and Maxwell Barus, Department of Physics, Brown University. A simple apparatus is described, and an elementary estimate first given. The apparatus was then modified, producing intensification, and new observations were made.

Dependence of the Spectral Relation of Double Stars upon Distance: C. D. Perrine, Observatorio Nacional Argentino, Cordoba. There is an indication that some external cause is operating in more or less definite regions of our stellar system upon the conditions which produce spectral class.

Hypothesis to Account for the Spectral Conditions of the Stars: C. D. Perrine, Observatorio Nacional Argentino, Cordoba. The spectral condition of a star depends chiefly upon its size and mass and the external conditions of density of cosmical matter and relative velocities of star and matter.

National Research Council: Minutes of the thirty-fourth, thirty-fifth and thirty-sixth meetings of the Committee; war organization of the National Research Council.

EDWIN BIDWELL WILSON MASS. INSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASS.

SPECIAL ARTICLES

TERNARY SYSTEMS AND THE BEHAVIOR OF
PROTOPLASM
I

In order to define more accurately the nature of certain changes which are observed in protoplasm (its normal water content, edema, cloudy swelling, fatty degeneration, necrosis) we have been continuing our study of the be

havior of various simple colloids so far as their powers of hydration and dehydration are concerned under the influence of changes in their surroundings. Since the chemistry of the proteins is rather complicated, we have turned to a study of the colloid behavior of the chemically simpler soaps, for these show close analogy in their processes of hydration and dehydration to certain proteins. The soaps,

however, behave in their turn much like mutually soluble systems of the type phenolwater-salt, and so we have passed from a study of the soaps to a study of these simpler physico-chemical systems. From these we have then built backwards through the soaps to the proteins and from these to the properties of living cells. The study as a whole makes clearer, we think, the nature of various changes which are observed in living matter. Many of the "vital" phenomena of cells may be interpreted in the terms of the behavior of simple hydrophilic colloids. These in turn, may be interpreted as expressions of the changes to be observed in systems of mutually soluble materials (like two liquids and a solid, a liquid and two solids, etc.) more particularly the changes incident to their "separation" in their critical realms" with the accompanying changes in viscosity, in light transmission, in state of solvent " or "dissolved" substances, etc.

66

66

II

Our studies on soaps not only corroborate the work of various well-known authors (Hofmeister, Lewkowitch, Krafft, Merklen, Goldschmidt, Botazzi, Victorow and Leimdörfer), but amplify their studies in that we worked with pure (salt-free) soaps and with longer series of such while subjecting them to more widely varying external conditions than is the case in most of the investigations thus far reported.

We began with the preparation of equimolar amounts of various salt-free soaps in the presence of a definite volume of water. For this purpose we neutralized (at the temperature of boiling water) the proper fatty acid with an equivalent of the proper alkali in a unit volume of water. When not otherwise speci

fied, our standard soap mixtures contain the proportions represented by a mol of the fatty acid neutralized by the gram equivalent of the proper metallic hydroxide, oxide or carbonate in the presence of a little water.

As long known from empiric practise, the different soaps bind totally different quantities of water. We first determined the absolute amounts of water that are absorbed by equimolar amounts of different oleates when prepared as described above. If that capable of holding most water is named first, the order in which these different soaps absorb water is about as follows; potassium, sodium, ammonium (?), lithium, magnesium, calcium, lead, mercury. Under the conditions of our experiments the first four bind all the water offered them (several hundred per cent.). Magnesium, however, holds but sixty per cent. its weight of water, and calcium oleate but forty. Even lower figures (about 10 per cent.) are obtained for the oleates of mercury and lead.

This general order in which the oleates with different basic radicals hold water is repeated by the palmitates, margarates and stearates. If the amount of water used in the preparation of the molar equivalents of soap is sufficiently reduced (to one fourth that stated above) then this same order may also be discovered in the case of the caprylates. These general findings seem therefore to justify the conclusion that a first factor in the determination of the amount of water held by any soap resides in the nature of the basic radical combined with the fatty acid.

We tried next to determine the effect of combining the same basic radical with different fatty acids of the same series. In these experiments we again neutralized one mol of the fatty acid with an equivalent of the necessary base (sodium or potassium hydroxide) in the presence of a constant volume (one liter) of water. The absolute amount of water taken up by a mol of any of these salts, as determined by discovering the maximum amount of water which such will take up at room temperature and yield a stiff jelly, increases progressively with the increase in the

molecular weight of the fatty acid used. The absolute amounts of water absorbed vary enormously. From the lower members of the series (from the formates through the caproates) no colloid jellies at all can be obtained. The crossing line is well marked by sodium or potassium caprylate. These soaps form clear (molecular) solutions in twice their weight of water but they form jellies with once their weight of water. The amount of water that will be thus taken up and yield a jelly increases progressively as acids above caprylic are used so that by the time stearic acid is reached, one part of soap will easily take up a hundred or even two hundred times its weight of water and form a solid mass. Experiments with fatty acids beyond stearic are not yet completed. Obviously then, with a given base, a second element in the amount of water held by a soap depends upon the nature of the fatty acid contained in the soap and its height in the series.

We tried next the effects of different alkalies, of different neutral salts and of different non-electrolytes upon the hydration capacity of different soaps (caprylates, laurates, oleates, palmitates, margarates and stearates of sodium and potassium). Our conclusions under this head may be summed up as follows:

1. The addition of any alkali to a "solution" of any of these soaps at first increases its viscosity or (in a limited volume of water) leads to its gelation; with higher concentration of the added alkali, there follows a decrease in viscosity ("liquefaction ") which change is succeeded, at sufficiently high concentration of the alkali by complete separation of the soap from the dispersion medium as a dry mass floating upon the "solvent." When equimolar solutions of the different soaps are compared it is found that the effects of an added alkali vary with (a) the fatty acid in the soap, (b) the base combined with the soap and (c) the basic radical of the added alkali. The lowermost members of the fatty acid series neither gel nor come out of "solution" upon addition of an alkali. The caprylates gel and come out easily while the higher soaps show these changes in increasingly marked

degree. When potassium and sodium soaps are compared, it is found that an added alkali will produce the series of changes earlier in a sodium soap than in a potassium soap. Similarly, if the effects are compared of adding equinormal solutions of potassium or sodium hydroxide to a given soap the former is found not so effective (in other words, a higher concentration is demanded to produce the series of changes noted above) as the latter. When solutions of the hydroxides of the bivalent or trivalent metals are used, the effects of the metallic radicals and the formation of metallic soaps with their low hydration capacity dominate the picture. Such hydroxides, therefore, lead uniformly only to decrease in viscosity and separation of the slightly hydrated soap from the dispersion mediums.

2. The addition of salts of the bivalent and trivalent metals to potassium, sodium, ammonium and lithium soaps leads to a clouding of the mixtures, a decrease in viscosity and a decrease in power to gel. The picture is, again dominated, in other words, by the production of the metallic soaps with their low hydration capacities. A more careful study of the hydration and dehydration of the soaps of the alkali metals under the influence of various salts is therefore, limited to the salts of the alkali metals. As generally known in technological practise, these salts lead to a "salting out" of the soap, or, when used in smaller amounts, to a "gumming" or "stringing " of the soap. We were able to confirm and amplify here the investigations of other workers in this field which have shown that such gumming and ultimate salting out are dependent upon the concentration and the chemical nature of the salt used. With rare exceptions (more particularly those salts which in aqueous solutions are not "neutral") all the ordinary salts of potassium, sodium, lithium, etc., at first increase the viscosity of a potassium or sodium soap to a point where at proper concentration a soap jelly results, be•yond which further increase leads to a fall in viscosity (liquefaction) until, in still higher concentrations of the salt, the soap begins to

separate from its clear dispersion medium, at first as a cloudy jelly and then as a (practically dry) dehydrated soap mass swimming upon the clear "solvent."

The intensity with which these successive changes are brought about again varies, at the same concentration of salt, with the fatty acid in the soap, the nature of the basic radical in the soap and the basic radical of the salt used. Potassium salts, for example, are less effective in bringing about the series of changes than the corresponding sodium or lithium salts.

The acid radical (fluoride, chloride, bromide, iodide, nitrate, sulphocyanate, sulphate, acetate, tartrate, citrate) in the series employed by us seems to influence the end results so little as to come within the limits of experimental error. In other words, with salts of a given base the acid radical is practically of immaterial importance.

When an alkali and a salt are together added to a soap, the action of the two is found to be algebraically additive. An alkalinized soap may be salted out by adding a neutral salt and at a concentration of the latter which would not by itself have proved effective. Vice versa, a partially salted soap may be completely dehydrated by adding an alkali to a concentration at which the alkali alone would have produced no such effect.

It is also of interest that all these effects of alkali, of salts, etc., are largely reversible. A soap dehydrated by an alkali or a salt can be rehydrated by merely adding water; a soap partially dehydrated by a sodium salt can be rehydrated by substituting a potassium salt, etc. Most interesting, however (and physiologically important), is the fact that magnesium, calcium and even iron and copper soaps can, through the addition of the proper salts or hydroxides of the alkali metals, be slowly brought back into the more highly hydrated soaps of these alkali metals.

3. The non-electrolytes (alcohol, glycerin, dextrose, saccharose, lactose, urea) as compared with the electrolytes have at the same concentration relatively little effect upon the hydration and dehydration of soaps. They