metabolism of the normal child is first to be known and should be the standard of reference for all others. If allowances are to be made for excessive weight or for decreased weight, such allowances can come later with a better understanding of the subject. Consequently, while we freely admit that other methods of prediction may be (for the time being, at least) of more help to the clinician and more practical from his standpoint, we still feel that the closest approximation to the real basal metabolism is that which should be reported. Hence in this connection we have given our recommendations for the closest approximation to basal, frankly stating that, if girls are found to be 10 per cent. above this standard, the work in the New York and Chicago laboratories definitely proves that, measured under the conditions ordinarily obtaining, this increase of 10 per cent. is without any significance whatsoever. If the increase is found to be 20 per cent. above basal, before concluding that the metabolism represents a pathological condition, the child should be studied after a night's repose in bed.

SUMMARY.

Based upon an analysis of the Nutrition Laboratory series of measurements of infants, young children, Girl Scouts, and adult women, it is found that a most important factor correlating with basal metabolism measurements is height. The importance of introducing some index of the state of nutrition, such as Pirquet's pelidisi, and the fact that the total 24-hour basal heat production divided by the pelidisi referred to height shows a straight line relationship from 3 months to adult life further justifies the emphasis laid upon. the height element in basal metabolism predictions. For new-born infants, both males and females, from 11⁄2 to 6 days old it is recommended that the formula involving length, previously published by Benedict and Talbot, be employed in predicting the metabolism. For girls from 1 week to 12 years of age a new method of prediction involving height alone is recommended. From 12 to 20 years of age the metabolism of girls may at present be best predicted from the curve based upon a series of experiments with Girl Scouts, showing the heat production per kilogram of body weight per 24 hours referred to age. For adult women 21 years and over the formula of Harris and Benedict for women, involving age, weight, and height, is retained.

SOME NEW EXPERIMENTS IN GRAVITATION.

FOURTH PAPER.

BY CHARLES F. BRUSH.

(Read April 26, 1924.)

The writer outlined "A Kinetic Theory of Gravitation" before the American Association For The Advancement of Science in December, 1910.1 Since that time he has presented several papers on the subject of gravitation, the last one at the general meeting of this Society in April, 1923.

That paper was chiefly descriptive of apparatus and method for comparing the velocities of freely falling bodies in two aluminum containers, alike in size, shape, and smoothness of surface, and dropped simultaneously, side by side, through e.ractly the same distance (about 122 cm.). The description was freely illustrated with plates and diagrams.

Each container, at the end of its journey, breaks an electric circuit. But the breaks of both containers are in series in the same circuit, so that the break which occurs first produces a bright spark. while the belated break gives no spark because its circuit is already open.

When the containers are equally loaded with the same metal, there is no visible spark at either break or a very feeble spark at one or the other indifferently. But when they are equally loaded with certain different metals, one container persistently produces a bright spark, though the containers are always reversed in position for each trial. From this it seems clear that the container giving the spark falls a little faster than the other. This sparking condition is clearly manifested when the faster container reaches the end of its free path as little as .0125 mm. (.0005 inch) in advance of its neighbor. This indicates a time difference about 1/400,000 second. (During their half second of falling the containers acquire a velocity about 4900 cm. per second).

1 Science, March 10, 1911; Nature, March 23, 1911.

2 Proc. Am. Phil. Soc., Vol. LIII., No. 213, January-May, 1914; Vol. LX., No. 2, 1921; Vol. LXI., No. 3, 1922; Vol. LXII., No. 3, 1923.

Some of the metals show very many times greater difference in falling velocities than the near minimum indicated above (about 1/100,000) and the paper referred to describes a method of approximate measurement of the lag of the slower container.

Following the presentation of the last paper referred to, many comparisons of metals and other substances were made with every precaution that could be thought of to avoid error, and in the light of much experience with the apparatus. The general results are embodied in the "Table of Relative Falling Velocities" herewith.

Aluminum, Magnesium, Carbon, Silicon, Boron, Water.

The spacing of the lines indicates approximately the relative velocity differences found. Differences, if any, between substances arranged in the same line were too small to be appreciable.

The observed difference in the falling velocities of platinum and sulphur (lag of the sulphur) is approximately .02 cm. in the 122 cm. of fall; or about I part in 6000.

The quantitative value of zinc in the table is somewhat uncertain, for reasons indicated in former papers, and later in the present paper.

It seems significant that the fast-falling metals, platinum, gold, tungsten and lead all have high atomic weight, and density. Lead, though less dense than the others, has higher atomic weight. Yet bismuth, with virtually the same atomic weight as lead, and little less density, is considerably slower.

Much lower down in the scale of velocities we find the elements of comparatively very low atomic weights and small densities, aluminum, magnesium, carbon, silicon, boron (and water) with no appreciable difference in behavior. This also seems significant.

Sulphur and selenium are in a class by themselves, being much slower than the light elements last referred to. Much care was taken to make sure of this. They were always melted in their containers, and each was checked repeatedly, and very satisfactorily, with aluminum and carbon.

It was rather expected that tellurium (also melted in its container) would, on account of its chemical similarity, behave much like sulphur and selenium. But it did nothing of the sort. On the contrary it behaved very much like zinc and tin. Probably this was due to its comparatively high atomic weight and density. This again seems significant.

In general then, it appears that the metals of very high atomic weight and great density have the highest falling velocities. And elements of very low atomic weight and small density have the lowest falling velocities; while elements of intermediate atomic weights and densities (zinc, tin and tellurium) have intermediate falling velocities or mass-weight ratios.

Allotropic condition seems not to affect mass-weight ratio. Thus, Acheson graphite and electric light carbon showed no difference in falling velocities. And the same was true of ordinary yellow sulphur and its gutta percha-like modification obtained by melting at a much higher temperature; though much heat energy must have been absorbed and rendered latent during the change.

In a former paper it was shown that "fusible alloy " behaved the same as its constituent metals mixed but not alloyed.

Only one experiment has been made involving chemical combination. Lead sulphide (galena) was found a little slower than metallic lead. As nearly as could be estimated, it behaved like a

mixture of lead and sulphur in the proper proportions to form the sulphide. Apparently, in this case, chemical combination does not affect mass-weight ratio. The writer thinks this is probably true as a general proposition, though of course a single isolated experiment is but a small step toward establishing it.

In the hope that some close relationship might be found between the observed falling-velocities of the elements tested, and their physical or chemical properties or combinations of properties, a table of all the common elements was prepared, arranged in the order of their densities. In parallel columns were placed the corresponding atomic weights, atomic numbers, specific heats and thermal conductivities both in terms of weight and volume. But a careful study of this Table failed to shed any new light on the subject.

It was thought that something of value might be learned by comparing the transparency (or opacity) to very hard X-rays of equal weight-thicknesses of some elements. To this end discs or cylinders of 24 elements, including all those shown in the Table except boron, were prepared of thicknesses inversely proportional to their densities; so that the X-rays must pass through equal weights of the several elements before reaching the photographic plate below them. Thus the platinum disc was only .117 cm. thick, while the magnesium cylinder was 1.439 cm. high. All were mounted in a circle of holes in an aluminum plate or carrier, placed directly over the photographic plate with only the usual orange and black papers intervening. The photographic plate was backed by a thick lead plate, and the whole outfit was revolved in a horizontal plane by clockwork, around the center of the circle of specimens; thus securing equal exposure to all. Several revolutions were made during an exposure.

The hard X-rays were furnished by a Coolidge tube with tungsten target, and excited by 200 Kv. The target was placed 1 meter above the photographic plate, and the rays were filtered through a 3/4 mm. copper plate and an aluminum plate. 15 or 20 seconds exposure was found suitable.

Prints from the negatives were made through a wide range of exposures under a tungsten lamp, while the negatives were revolved as before, to secure uniform exposure.

The prints are beautiful and extremely interesting in several ways;