professor of eugenics in the child-welfare research station of the State University of Iowa. DR. RALPH F. SHANER, for several years connected with the department of anatomy of the Harvard Medical School, has entered on his work as assistant professor of anatomy in the University of Alberta.

DR. D. BURNS, Grieve lecturer on physiological chemistry in the University of Glasgow, has been appointed professor of physiology in the University of Durham College of Medicine, Newcastle-upon-Tyne, in succession to the late Professor J. A. Menzies.

DISCUSSION AND CORRESPONDENCE THE CAUSES OF WHITENESS IN HAIR AND FEATHERS

My attention has recently been called to a statement by W.D. Bancroft1 to the effect that white hair and feathers owe their color to the entrance of air into their structure. Similar statements have appeared elsewhere at various times, and this conception appears to be widespread.

No one, to my knowledge, has ever presented any real evidence that either hair or feathers have any more air in them when white, than when colored. Furthermore it is quite unnecessary for them to have more air. I have never been able to see any difference in the structure of white hair and feathers as compared with colored hair and feathers, except for the presence or absence of pigment.

In 1904, I made the statement, in an address, that hair and feathers are white for the same reason that powdered ice or glass and other transparent substances in a fine state of division appear white.2

Hair consists of numerous cornified epithelial cells more or less incompletely fused together. In the case of human hair, most of the structure is cortical. These cells furnish a vast number of external and in

1 Applied Colloid Chemistry, 1921, p. 198.

2 See abstract in Biol. Bull., 1904, Vol. VI., No. 6, p. 311, for remarks about white feathers. See also Anat. Rec., 1918, No. 1, p. 52, for discussion of white hair.

ternal reflecting surfaces, as can be seen easily by placing a white hair on the microscope stage with no mounting fluid. When pigment is present, the incident light is more or less extensively absorbed, according to the amount of pigment, before reaching the deeper cells. The amount of undispersed light reflected, of course depends on the number of internal reflecting surfaces not screened by pigment. There is always some reflection of undispersed light by the hair cuticle, no matter how much pigment is present.

The white of feathers is produced mostly by the barbules which are of microscopic size and consist of single columns of cells.

Hair and feathers have many times the surface, external and internal, provided by small bodies of similar mass but less intricate structure. According to a well-known law, the surface of a cube varies relatively to the volume inversely as the diameter. Thus a cuboidal cell one tenth of a millimeter in diameter has ten times as much surface, relatively, as a body one millimeter in diameter. Furthermore, the amount of reflecting surface is increased by the irregular contour of the hair and feather elements. The total area of the vast number of facets in a single, unpigmented hair or feather which are in a position to reflect light to the eye is relatively very great.

White in hair and feather structures is due to failure or absence of pigment formation in the follicle before cornification takes place. I know of no critical evidence that either hair or feather structure can become white in any other way. The process is therefore slow, and the time required for a change to white is determined by the rate of growth.

Similar views are expressed in an article by Stieda where a discussion of the origin of the notion that hair may suddenly become white is discussed in detail.

R. M. STRONG LOYOLA UNIVERSITY SCHOOL OF MEDICINE, CHICAGO, ILL.

3 Verh. der Gesellsch. Deutscher Naturforsch. und Aerzte., 1910, Bd. 81, S. 222-224; also Anat. Hefte, 1910, Bd. 40, H. 2.

SIDEWALK MIRAGES

TO THE EDITOR OF SCIENCE: A number of communications, published in SCIENCE during the past year, on "Sidewalk mirages" having recently come to my attention, I would like to add my experience with this phenomenon to those which have been related. I have driven over a stretch of road, part asphalt and part concrete, daily for the past two years, and have looked for mirages under every condition of the weather. Over the distance of the three miles of roadway I have marked every spot where the mirage occurs.

The nature of the road surface seems immaterial, but the effect of a "water surface can be obtained wherever the level of the eye approaches that of the road surface. The mirage is not visible in cold winter weather and it is best during the very hot days in July and August. I believe that the intensity of the effect is unquestionably a function of the temperature of the road surface and the air immediately above it. That one observes a true mirage in this phenomenon and not a simple reflection can be demonstrated by the fact that an object "mirrored" on one of these surfaces will show an angle of incidence of probably 45° or greater, whereas the angle of reflection is, as stated previously by another observer, very small, approximating a few degrees only.

Mirror-like effects on asphalt roads are common, but have not the clarity of the images seen in a mirage, nor can mirror effects, due to reflection simply, be seen on a concrete road, so far as I have observed.

The position of the sun is of no influence, as mirages have been observed at the same spot at all times of the day.

CARNEY'S POINT, N. J.

ALLAN F. ODELL

DISCOVERY OF A PREHISTORIC ENGRAVING REPRESENTING A MASTODON

TO THE EDITOR OF SCIENCE: It may be of interest to you to learn of the recent reexamination of Jacobs' Cavern, a prehistoric rock-shelter located in extreme southwest Missouri, some three miles from Pineville, county seat of McDonald County. This

cavern was examined by Dr. Charles Peabody and Mr. Warren K. Moorehead, of Phillips Academy, in 1903, report of their examination appearing in 1904 in Bulletin No. 1, "Exploration of Jacobs' Cavern."

Subsequent periodical and amateur investigations carried on by the writer, who now owns the land upon which this cavern is located, have resulted in the discovery of a number of very interesting artifacts. Chief among these are bone and horn awls, flint implements, engraved and polished implements of stone, and shaft straighteners and smoothers. Portions of an adult human skeleton, accompanied by an engraved sandstone pipe, have also been found.

The latest discovery was made on April 17, 1921, when the writer and Mr. Vance Randolph exhumed several engraved, perforated, and otherwise ornamented bones. These were apparently firm and sound but as a precautionary measure pen drawings were made immediately. Nevertheless, upon being examined a few weeks later, it was found that the bones were rapidly disintegrating. Immediate preservative treatment was resorted to but was so limited by local conditions that it was found impossible to save more than the most important specimen.

In many respects this bone is very interesting. One side bears an engraving which prominent archæologists have agreed seems to resemble a mammoth or mastodon. The reverse side bears two rows of parallel zigzag lines, lengthwise of the bone, the design corresponding closely with those found on the sandstone pipe. This design is also accompanied by another evidently intended to represent some member of the deer family.

The writer felt that Phillips Academy was naturally entitled to priority rights of reexamination of the cavern. However, Mr. Moorehead found it impossible to visit the cavern and recommended that Dr. Clark Wissler, of the American Museum of Natural History, make the examination. Dr. Wissler is now on the ground for that purpose.

Photographs of the most important specimens are in process of preparation and a

FOR Some time the writer, when photographing fossils, has used the whitening process contributed by Professor S. H. Williams, but, with many others, he has found it not altogether satisfactory. In order that the whitened specimen should contrast with a white background it has been necessary to over-expose or over-develop the prints. Because of this, many of the minor details of fossils have been lost in reproduction, and the pictures, as a rule, have seemed flat and "lifeless." In addition, it is usually the practise to opaque the background of the negative as an aid in determining how far to carry the development of the print. This process is painstaking and slow at best.

Some time ago, the writer, with the assistance of Mr. Parke Bryan, developed a slight variation in the photographing of whitened fossils that seems to be a decided improvement. The time required is materially shortened, in that the negative requires no opaquing, and the results are so gratifying in the way of improved reproductions that it seems worth while to outline briefly the method.

The method is a combination of the common lighting arrangement used in portrait photography, and the whitening process of Professor Williams. The specimen is mounted on a slender stick with modeling clay and then coated with a thin film of white. A dull white background, placed some distance behind the specimen, is turned at an angle such that it receives the full light but does not reflect it toward the camera. After the photographing table is orientated so as to give the conventional light direction and the desired lightshade contrast to the relief features, a screen is placed between the specimen and the source of light so as to intercept the direct rays. The screen consists of one or more thicknesses of

cheesecloth sewed on a wire frame, the number of thicknesses depending on the intensity of the light. Every feature of the fossil now shows clearly on the ground glass of the camera, although the specimen appears dark against a pure white back.

It has been found that the shadows on the under side and away from the light source are more intense than the image on the ground glass indicates, and except in the case of relatively flat specimens it has been necessary to use a slight back reflection. A sheet of dull finish white cardboard held at the proper angle has in every case been sufficient for this purpose. If an actinometer is used to determine the time of exposure, it is obviously the light of the shaded specimen that is to be tested.

DEPARTMENT OF GEOLOGY, UNIVERSITY OF MISSOURI

MAURICE G. MEHL

SCIENTIFIC BOOKS Vitamines: Essential Food Factors. By BENJAMIN HARROW, Ph.D. New York, E. P. Dutton & Co., 1921. Pp. 219. Price $2.50. The author of this book has been at great pains to popularize a subject which the laity will certainly be glad to have so clearly presented. About half the volume is preliminary to the specific topic; it is a general account of nutrition and the story is well told. One is disposed to wonder whether readers who require such a very elementary introduction will appreciate the later chapters which are of necessity more difficult. However, a rare degree of order and simplicity is maintained to the end. The writer has a judicial attitude; he does not assert opinions of his own but quotes others with fairness and has evidently been in correspondence with the leading investigators that he may accurately express their views.

Of course not much space can be devoted to controverted matters in a book of this character. But a dogmatic tone is avoided. It should be plain to the reader that many problems await solution. Among the questions not fully settled may be mentioned the

following: whether rickets is due to lack of Fat Soluble A, whether there is an antiscorbutic vitamine (Water Soluble C), and in what sense pellagra may be rated as a deficiency disease. All the material is handled in a cautious and modest way with the result that no encouragement is given to faddists of any kind.

PERCY G. STILES

EXPERIMENTS ON THE RECORDING AND REPRODUCTION OF CARDIAC AND RESPIRATORY

SOUNDS

We have recently conducted experiments at the Bureau of Standards in which permanent records of cardiac and respiratory sounds have been made and reproduced by the use of a telegraphone. The records have also been made audible throughout the room with the aid of audion amplifiers and a loud-speaking telephone.

A carbon telephone transmitter of ordinary type with a rubber adapter substituted for the mouthpiece was used for the stethoscope. The currents from the telephone transmitter were amplified by means of a five-stage audion amplifier which was connected to the recording element of a steel wire telegraphone. The magnetic records of the cardiac and respiratory sounds thus obtained were made audible by connecting telephone receivers to the telegraphone in the usual manner. The telegraphone currents were also amplified by means of a three-stage audion amplifier which was connected to a loud speaking telephone. In this way the sounds were made audible throughout the room.

This method of obtaining permanent records of cardiac and respiratory sounds and of reproducing them offers interesting possibilities in the study of normal and pathological conditions of the heart and lungs and their demonstration to an audience for purpose of instruction.

SPECIAL ARTICLES

THE SEPARATION OF THE ELEMENTS CHLORINE AND MERCURY INTO ISOTOPES IN SCIENCE of March, 1920, Harkins and Broeker reported that they had obtained a separation of chlorine into isotopes by diffusing hydrogen choride gas. The separation at that time amounted to an increase of atomic weight equal to 0.055 unit, or a change of density amounting to 1,550 parts per million. This separation has been definitely confirmed by Dr. Anson Hayes and the writer, who have secured an increase of 0.04 unit of atomic weight in a larger quantity of material. Elaborate purifications have been resorted to, and definite evidence has been secured to show that the increase in density found is actual, and not due to impurities. The details of this work were supposed to have been printed in the August number of the Journal of the American Chemical Society. However, since the date of publication of this number is doubtful on account of the printers' strike, it seemed advisable to answer here the considerable number of inquiries as to whether we have secured definite evidence of the separation.

About six months after our notice of the separation of chlorine into isotopes had been published, Bronsted and von Hevesy published a notice in Nature indicating that they had separated mercury into isotopes. However, since the extent of the density change reported by them was only about one thirtieth of that previously obtained by us in the case of chlorine, it seemed to us that the evidence for this separation of mercury was inconclusive, since a change of 50 parts per million in density might be due to minute amounts of impurities. In order to see if they could confirm these results, Dr. R. S. Mulliken and the writer have vaporized mercury at low pressures. The mercury was carefully purified by five fractional distillations in air at low pressures, and one in a high vacuum, after initial purifications with nitric acid. The increase in density obtained amounts to 69 parts, and the decrease to 64 parts or a total

change of density of 133 parts per million, or 0.027 unit of atomic weight.

The evidence that a separation has actually been obtained rests in the quantitative agreement between our results and those of Bronsted and von Hevesy, with respect to the rate of separation (efficiency of process). If we consider the efficiency of our more ideal apparatus as 100 per cent., that of the other investigators is 75 per cent. while that of our less ideal apparatus used in the greater part of the work in order to save the expense of carbon dioxide as a cooling agent, was 93 per cent. when the vaporization was slow, and as low as 80 per cent. for a rapid vaporization. We have obtained evidence that there is a slight separation of isotopes produced when mercury is distilled slowly at a sufficiently low pressure.

The rate of separation of two isotopes varies as the square of the difference of their atomic (or molecular) weights, and the product of their mol fractions, as the logarithm of the cut, and inversely as the atomic (or molecular) weight.

A diffusion coefficient has been calculated to represent the relative separation of isotopes attained in terms of the atomic weight change, when a definite cut is made. The values are 0.00843 for neon, 0.00868 for magnesium, 0.00450 for lithium, 0.00758 for nickel, while the experimentally determined coefficient for mercury is 0.00570. For chlorine the coefficient is 0.00950 for hydrogen chloride, 0.00690 for methyl chloride, 0.00494 for chlorine, 0.00413 for methylene chloride, 0.00295 for chloroform, and 0.00229 for carbon tetrachloride.

It is of interest to note that there are 9 isotopic forms of MgCl, (or more if there is a chlorine of atomic weight equal to 39), 7 of CCl, and if mercury consists of 6 isotopes, there are 63 isotopic forms of Hg,Cl2. In addition to this most of the isotopic forms of C.Cl, consist of a number of space isomers.

UNIVERSITY OF CHICAGO, August 30, 1921

6

WILLIAM D. HARKINS

AN ARTIFICIAL NERVE

PHYSIOLOGISTS are keenly interested in all attempts to discover an explanation or an analogy for the passage of the nerve stimulus. Most enlightening suggestions have recently been presented by Lillie1 in his studies of passivity phenomena in pure iron wires. It seems that the transmission of the momentary wave of activity which occurs in a passive iron wire on activation in 70% nitric acid is closely analogous both chemically and electrically to the passage of the nerve impulse.

The general similarity of the two phenomena was apparently first noticed by Wilhelm Ostwald and subsequently elaborated by his student Heathcote.2 In a paper published in 1907 under the caption "Transmission along a nerve" (p. 909) Heathcote writes as follows:

In 1900, then, Prof. Ostwald called our attention to the possibility of nerve transmission being a process akin to the transmission of activity. . . . It is to be expected . . . that transmission of aetivity would be slower immediately after the first transmission owing to products of reaction around the iron. This has been confirmed by direct experiments in the case of iron in nitric acid. An effect of this kind in a nerve would explain the nature of "fatigue" so far as it concerns nerves.

After discussing the small amount of energy consumption in both transmissions Heathcote summarizes his conclusions as follows: