In the list of publications cited appear those of more than thirty of his Filipino colleagues and students. By such institutional activity the task depicted in Dean Baker's foreword has been carried on.

E. B. COPELAND

Emile Berliner, The Maker of the Microphone. By FREDERICK WILLIAM WILE. New York, N. Y. Bobbs Merrill Co., 1926.

FREDERICK WILLIAM WILE has been known to the American public as a newspaper correspondent and radio broadcaster, but in this book he will make himself known as the interpreter of the difficult and abstruse technical aspects of certain features of our civilization. He has done for Emile Berliner, "The Maker of the Microphone," what Michael Pupin did for himself in "From Immigrant to Inventor," and in some ways it is better.

The book is full of the romance of science, weaving as it does the conditions of the times that influenced the inventor. In particular the story of his work with the Bell Telephone Company is most readable. The story of how Bell was rewarded with success in the use of the electro-magnetic coil in making the first telephone is probably new to nearly every American. If the contact point on the coil had not welded so that it did not operate as intended, Bell might never have been known as the author of "the telephone, America's greatest invention." While Bell made use of induced electricity he did not use an "induction coil," as such is known and often termed "transformer."

Likewise Berliner by almost an accident learned that a variable contact would produce an undulating current, and from this was able to produce an undulating current corresponding to the sound pressures developed in a telephone transmitter.

The history of the telephone is fascinating, not only because of the personalities involved, but also because it shows how so many masters of many arts are involved in a great invention.

The story of the gramophone as it evolved by and about Berliner is almost as interesting as that of the telephone. The public generally does not know that the common disc record in use to-day is the invention of Berliner.

This book is commended to those who like to read a good biography, to those who like to learn how important inventions are made, and to those who like to learn once again that America is a land of opportunity.

The introduction by Herbert Hoover is a gem to be prized in any library.

BUREAU OF STANDARDS

F. C. BROWN

SPECIAL ARTICLES

NOTE ON THE PIEZO-ELECTRIC EFFECT IN ROCHELLE SALT CRYSTALS

ALTHOUGH the piezo-electric effect in Rochelle salt crystals cut normal to the a crystallographic axis has been studied,1 the only work on plates normal to the b and c axes consists of some old measurements by Pockels at room temperature. Because of the characteristic changes in piezo-electric activity at -20° C. and at 25° C. observed in plates cut normal to the a axis, it was thought to be of interest to obtain corresponding observations on crystal plates of other orientations. Plates were cut from a large crystal perpendicular to each of the three crystallographic axes. Electrodes of tinfoil were attached by means of Canada balsam and the pressure was applied at 45° to the axes in the plane of the faces on which the charges produced were measured. The results of these measurements are shown in Fig. 1. The charges were measured by means of an electrometer null method, except for the last three points on the curve for the a direction, which were obtained by

X

138 × 10-8 and 28.3 × 10-8 e.s.u. per dyne at 0° C. The changes in the values for the b and c directions per degree rise in temperature are, respectively, + 6.8 × 10-9 and 3.1 10-10 between 60 and +30° C. X The figure shows that the ions which produce the piezo-electric polarization move much more freely in the a direction in the crystal, especially in the temperature range from -20° C. to +25° C. Measurements of the dielectric constants also give results of the same nature. It was found, moreover, that the electrical anomalies, such as fatigue, hysteresis and residual charges, are relatively much smaller in the band c directions. Although the crystal cut normal to the b direction is much less active than one cut normal to the a direction, the absence of these irregularities and of the large temperature coefficient just above ordinary room temperature would make the crystal cut in this way more useful in practical applications of the piezo-electric effect. It is still about twenty times as active as quartz.

UNIVERSITY OF MINNESOTA

JOSEPH VALASEK

THE INFLUENCE OF EPINEPHRIN AND OF
THE SYMPATHETIC SYSTEM ON
SKELETAL MUSCLE FIBERS
AND CAPILLARIES

It is possible to observe skeletal muscle fibers and their circulation under the high power of the microscope, by a method which will be described later. Cats have been used in these observations.

Epinephrin in small doses (.2 to 4 cc, 1: 100,000) causes muscle capillaries and venules to dilate. New capillaries open while those already open may dilate further. The field becomes brighter, individual fibers show more clearly and the striations of the fibers appear or become clearer if already visible.

Larger doses (.8 cc to 3 cc, 1:100,000) cause constriction of capillaries, some of which entirely disappear. Stronger doses than these cause closing of many more capillaries and much more marked constriction.

Epinephrin caused twitching of the muscle fibers sometimes even with the small doses, although the larger doses cause more marked and more lasting effects. The twitching is at right angles with the longitudinal axis of the fibers.

Epinephrin also causes the intestinal villi to become more transparent. It has a similar action upon the bladder.

Stimulation of the lumbar sympathetic with weak induction shocks causes the field to become more opaque. Many new capillaries and venules open (skeletal muscle) and those already open dilate. As the stimulation is increased the field becomes more

THE INDIANA ACADEMY OF SCIENCE THE Indiana Academy of Science held its fortysecond annual meeting at Ball Teachers College, Muncie, Indiana, on December 2, 3 and 4. A total of 80 papers was presented. These were distributed among the various sections as follows: General meeting, 8; Botany, 22; Zoology, 19; Chemistry-PhysicsMathematics, 15; Geography-Geology, 12, and special meeting on the Teaching of Science in the High School, 4.

The officers in charge of the meeting were:

President, Dr. W. M. Blanchard, De Pauw University, Greencastle; Vice-president, L. J. Rettger, Indiana State Normal, Terre Haute; Secretary, Dr. Ray C. Friesner, Butler College, Indianapolis; Assistant Secretary, Dr. W. P. Morgan, Indiana Central University, Indianapolis; Acting Treasurer, Dr. H. E. Enders, Purdue University, Lafayette; Editor, J. J. Davis, Purdue University, Lafayette.

The public lecture of the academy on the evening of December 3 was given by Professor W. H. Hobbs. of the department of geology, University of Michigan, who spoke on "The First Greenland Expedition of the University of Michigan." This talk was illus trated by two reels of motion pictures of the expedition.

Preceding the regular meetings of the academy, the entomologists of Indiana held their annual informal round table and discussed the scientific and economic problems of current interest in the state.

The newly elected officers for the year 1927 are as follows:

President, Frank B. Wade, Shortridge High School. Indianapolis; Vice-president, Fred J. Breeze, Ball Teachers College, Muncie; Secretary, Dr. Ray C. Friesner, Butler College, Indianapolis; Assistant Secretary, Dr. W. P. Morgan, Indiana Central University, Indianapolis; Treasurer, Dr. Marcus W. Lyon, Jr., South Bend; Editor, J. J. Davis, Purdue University, Lafayette; Assistant Press Secretary, Dr. J. A. Nieuwland, Notre Dame University, South Bend. Notre Dame University was chosen as the place for the next annual meeting.

HARRY F. DIETZ, Press Secretary

Scientific Events:

Concilium Bibliographicum; Opening of the Botany School of the University of Sydney; The Institute of Chemistry of the American Chemical Society; The Ella Sachs Plots Founda tion for the Advancement of Scientific Investigation

Scientific Notes and News

University and Educational Notes

Discussion and Correspondence:

The Increase in Scientific Periodicals since the Great War: ALICE C. ATWOOD. Hooke's Law Again: PROFESSOR JOSEPH O. THOMPSON. Seymour Sewell on "Salps of Indian Seas": DR. MAYNARD M. METCALF. Storm Damage at Long Beach, N. Y.: DR. HENRY S. SHARP Scientific Books:

A History of our Times: PROFESSOR T. D. A. COCKERELL

Scientific Apparatus and Laboratory Methods:

A Quick Method of Preserving Cats for Dissection: HORACE E. WOOD. A Culture Medium for Free-living Flagellates: JAMES B. LACKEY Special Articles:

Concerning the Protoplasmic Currents accompanying Locomotion in Ameba: PROFESSOR HERBERT W. RAND and S. Hsu. The Increase in the Calcium of Hens' Blood accompanying Egg Production: J. S. HUGHES, R. W. TITUS and B. L. SMITS Science News

245

SCIENCE: A Weekly Journal devoted to the Advancement of Science, edited by J. McKeen Cattell and published every Friday by

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y. New York City: Grand Central Terminal. Annual Subscription, $6.00. Single Copies, 15 Cts.

SCIENCE is the official organ of the American Association for the Advancement of Science. Information regardIng membership in the Association may be secured from the office of the permanent secretary, in the Smithsonian Institution Building, Washington, D. C.

Entered as second-class matter July 18, 1928, at the Post Office at Lancaster, Pa., under the Act of March 8, 1879.

THE GRAPHIC REPRESENTATION OF RELATIVE VARIABILITY1

IT has been the generally accepted biometric practice to use the coefficient of variation as the measure of the relative variability or scatter of frequency distributions. This constant is

V =

100 (standard deviation) Mean

It gives the standard deviation of the distribution in terms of the mean value of the varying character. By expressing the scatter of the distribution in this way it becomes possible to compare the relative variabilities of characters measured in different absolute units.

But the coefficient of variation has never been an entirely satisfactory constant to biologists, at least. While formally correct enough, within the limits of its definition, it does not readily or instantly call up in the mind an adequate picture of the real degree of scatter of the distribution. This is, in part at least, because two things, the mean and the standard deviation, are involved in it. When one reads the value of the standard deviation of a particular distribution he recalls that roughly three times this quantity on either side of the mean includes the entire frequency and this gives at once some concept of the biological extent and meaning of the variation, in the particular

case.

There would seem to be a place of usefulness for an adequate graphical method of depicting relative variability for comparative purposes, so that one may see the difference or likeness in the variation of a man and a mouse, for example, in respect of body-weight. It is the purpose of this paper to describe such a graphic method, and to illustrate its applications.

The method may best be approached through a concrete illustrative example. We have lately been studying in this institute the normal variation and correlation of the relative cell volume of human blood, in relation to age, body-weight and stature. present situation regarding the measurement and graphical depiction of variation in these four charac

The

1 From the Institute for Biological Research of the Johns Hopkins University.

2 Cf. Pearl, R., and J. R. Miner. "A Biometric Study of the Relative Cell Volume of Human Blood, in Normal and Tuberculous Males." Johns Hopkins Hospital Bulletin. In press.

FIG. 1.

Histogram showing variation in body-weight in a group of 272 normal males.

ters, in a series of 272 normal males, is fairly exhibited in Table I and Figs. 1 to 3.

Plainly the diagrams tell nothing whatever about the relative or comparative variability in this group of males in respect of the three characters, bodyweight, stature and relative cell volume. They are correctly plotted histograms, but the unit of abscissal measure is different in each case and direct comparison is impossible.

From Table I we learn, through the coefficients of variation, that the group is from three to five times more variable relatively in respect of age and bodyweight than it is in respect of stature or relative cell volume. But what does this mean translated into

Standard deviation

5.22.15 yrs.

19.95.58 lbs. 2.45.07 in. 2.47.07%

Coefficient of variation (per cent.)

17.06.51

13.16.39

3.60.10

5.42.16

terms of distribution of frequency? A simple, direct and easily interpreted answer is not forthcoming.

Suppose now we decide to express the age, the body-weight, the stature and the relative cell volume of each of these 272 individuals as a percentage of their respective mean values, the mean of each character being taken as 100 per cent. And further suppose we express the frequencies as respectively so much per one per cent. of the mean of each character. These are simple and entirely permissible transformations of the original data.

The data in their original form and after the transformation described are shown in Table II.

If now the figures in the columns headed A and B in Table II be plotted on arithmetically ruled coordinate paper we shall have a true picture of the relative variability of the four characters considered. This is done in Fig. 4. Each of the four frequency polygons has the same area, as a result of the transformations effected in the B columns.

This method of plotting superimposes the different polygons of variation on a common Cartesian coordinate grid, with the mean value for each of the compared variables at the same abscissal point. It constitutes no new method of measuring biological variation, but merely visualizes effectively what the coefficient of variation measures.

The method of plotting used in Fig. 4 shows at a glance that the 272 men of this group differ among

ABSOLUTE AND RELATIVE FREQUENCY DISTRIBUTIONS FOR VARIATION IN (a) AGE, (b) BODY-WEIGHT, (c) STATURE, AND (d) RELATIVE CELL VOLUME OF THE BLOOD IN 272 NORMAL MALES

Observed absolute

frequency

Per cent. which mid-point of class is of mean

Absolute frequency per

one per cent. of mean