In this diagram the vertical lines are drawn one magneton apart. The circles and arrows indicate by their position horizontally the number of magnetons in the molecule of the substance. It is to be noted how often the circles come on the vertical lines and how rarely they miss them.

A study of magnetite shows that it exists in five states in which the molecular moments are in the relation 4:5:6:8:10, but seem not to be exact multiples of the magneton. Even in the case of solid paramagnetic salts the atomic moment keeps close to whole multiples of the magneton.

From considerations based on its specific heat it seems essential to conclude that the oxygen molecule is rigid and has only a

single magnet. Its moment is found to be 14.014 magnetons. For the molecule of NO the moment is 9.039 magnetons.

The same atom seems capable of having very different magnetic moments according to its temperature and associates within the chemical molecule; e.g., near the absolute zero a nickel atom has a moment of three magnetons, above 400° it has eight, above 900° it has nine, while in solutions of its salts it has sixteen. Weiss says: "If we admit, what is extremely probable, that the magnetic moment resides in a material substratum, we can say that the magneton is a constituent element of a great number of magnetic atoms and probably of all. So far proof has been made for Fe, Ni, Co, Cr, Mn, V, Cu, U. The magneton

idea furnishes a new point of attack on magnetic problems. The parallelism of intensity of chemical properties with the number of magnetons raises new questions. What rôle do the magnetic phenomena play in chemical combination? Are chemical forces in certain cases the attraction of elementary magnets? Does valence stand in any relation to the magneton? . The magneton is a third constituent of matter which, like the electron and the alpha particle, belongs to a great number of molecules. Antiquity believed in the unity of matter. The alchemists in seeking to make gold endeavored to transform this belief into tangible reality. Now the discovery of constituents common to all atoms brings us anew to the conception of the Greeks and to the Unity of Matter." He suggests that his magneton may be one of the constituent elementary magnets which Ritz imagines arranged in neat rows with electrons revolving about them in order to explain the orderly series of spectral lines emitted by the atom.

It is a logical sequence which leads from the definition of the magneton as a quantity to its recognition as a separate entity. A study of the weight of Philadelphia chimneys would show that all the weights have a greatest common divisor, but it would not be long before the investigator would conclude that this particular weight is associated with some individual piece of matter, and the conception of a brick would be forthcoming.

Can the magnitude of the magneton be accounted for by supposing it due to one electron revolving in an atom? The magnetic moment of such an electron orbit can be calculated from a knowledge of three other quantities: the electric charge of the electron, the area of its orbit, and its time of revolution. When these are substituted in the formula for the moment a quantity is obtained equal to about one-tenth of the magneton. It is significant that they are not very different in magnitude.

Gans and Weiss have attacked magnetic problems from different sides. Weiss postulated a molecular field and arrived at the conception of the magneton. Gans, on the other hand, starts from a study of elementary magnets and their groupings and deduces a law for the susceptibility of ferromagnetic substances which is equivalent to that found by Weiss. It strengthens our confidence in this law that it has been derived independently by two different investigators following diverse methods.

Much interest attaches itself to investigations of magnetic properties at low temperature because the motion due to heat plays a less and less important part as the absolute zero is approached. The idea had occurred that those three metals, vanadium, chromium, and manganese, which lie next to the strongly magnetic group, iron, nickel, and cobalt, might, like these, be really magnetic, but have their Curie point so far below ordinary temperatures that they behave as paramagnetic, just as iron acts at high temperature. Onnes and Weiss investigated this possibility, but they found no trace of ferromagnetism in vanadium and chromium. Manganese was paramagnetic, but after having been fused it was ferromagnetic.

Gadolinum sulphate faithfully followed Curie's law that specific susceptibility is inversely as the absolute temperature even down to 17° absolute. At the melting-point of hydrogen it deviates slightly. At 290.3° absolute its specific susceptibility is proportional to 73, while at 13.91° absolute it has grown to 1468.

Oxygen is of peculiar importance for magnetic investigations, because as a gas it is paramagnetic, while most other gases are diamagnetic, and, besides, it can be worked with in the solid, the liquid, and the gaseous state. Curie showed that from 20° C. to 452° C. gaseous oxygen follows his law. Onnes and Oosterhuis found that it continues to do this down to about 170° absolute. Liquid oxygen does not follow Curie's law that the product of absolute temperature and specific susceptibility is constant. It does, however, admit the relation that the square root of the

absolute temperature multiplied by the susceptibility is constant. This relation holds from 90° to 65° absolute.

Solid oxygen has a much lower susceptibility than the liquid. When the liquid freezes the susceptibility drops to about one-third, then as the temperatures reaches - 240° C. another drop to one-half occurs in the susceptibility.

Any explanation of magnetic phenomena at low temperatures is complicated by the approach of the molecules to each other and by our ignorance as to whether any energy remains in matter at the absolute zero.

Let me call your attention to a curve showing the connection between specific susceptibility and atomic weight. Notice the periodic change from positive to negative susceptibility, the change, that is, from paramagnetic to diamagnetic. Take the sodium group, lithium, sodium, potassium, rubidium, cæsium, with atomic weights 7, 23, 39, 85, 133, and note their position. There is a fall in susceptibility with increasing atomic weight. While the curve is by no means smooth and regular, it certainly indicates a connection of a periodic character between the two quantities it represents.

At this point I cease my discussion of modern investigations in magnetism. It is not possible to give any impression save that of incompleteness, for the work is by no means finished. The problem of the reality of the magneton must be decided, and this conception must either be relegated to the limbo where dwell such cast-off terms as caloric, the N rays, hot ice, and the emission theory of light, or, with the electron, the alpha, beta, and gamma rays, be accepted as members in good standing of physical society. Who would want to serve a science that is finely rounded and complete, with no questions lacking an answer, with no phenomena waiting for explanation? Such a science might be beautiful, symmetrical, logically perfect, but it would also be dead.

(NOTE. The writer desires to make his acknowledgments to a number of authors for diagrams and other material employed in the above paper.)

NORTHEAST HIGH SCHOOL,

Philadelphia.

STANDARDIZED COLORED FLUIDS.1

BY

H. V. ARNY, Ph.D.,

Professor of Chemistry, Columbia University College of Pharmacy,
Member of the Institute,

AND

C. H. RING.

THIS is the fifth of a series of papers 2 on the subject of color standardization, the work being begun by one of us in connection with his duty as chairman of a sub-committee of the Committee on Revision of the National Formulary, to which was entrusted the task of obtaining reliable and definite standards for the various coloring agents-notably cudbear and caramel-for which recipes are provided in that book. In pursuing this investigation, there was devised a plan of color standardization which afforded a simple and practical way of establishing color standards.

As fully explained in the paper read before the Eighth International Congress of Applied Chemistry, the idea of preparing any shade of color by blending permanent standardized colored fluids of red, yellow, and blue appealed to him; and, as reported at that congress, the first set of such colored fluids was prepared by making half-normal solutions of cobalt, ferric iron, and copper. From these solutions were prepared the 88 blends, representing all the possible combinations in a total volume of 12 Cc. of fluid, where the amount of each of the three ingredients ranged from I to 12 Cc., fractional parts of the cubic centimetre being avoided.

The idea of using solutions of colored metallic salts for matching tints has already been employed in water analysis by Crookes, Odling and Tidy, and by Hazen; but the plan mentioned above has the following points of originality:

1 Communicated by Dr. Arny.

2" Color Standards," by H. V. Arny, American Druggist, 60 (1912), 35. "International Standards for Colored Fluids," by H. V. Arny, Proceedings of the Eighth International Congress of Applied Chemistry (1912), 26, 319. "Weiteres über Normal Farben," by H. V. Arny, Deutsch-Amerikanische Apotheker Zeitung, 33 (1913), 165. "The Problem of Color Standardization," by H. V. Arny and E. G. Pickhardt, Druggists' Circular, 57 (1914), 131.

VOL. CLXXX, No. 1076-15

199