is not an artifact, i.e. due to the displacement of an ordinary spindle, is indicated by the fact that the whole mitotic figure is asymmetrically developed, in conformity with its diagonal position (D and E).

Of course it is well known that the longitudinally dividing cells of the cambium form an exception to Sachs' law of 'rectangular intersection of successive division-planes' and Hertwig's modification of Sachs' hypothesis, but it has been assumed by Giesenhagen and others that, in the cambium, the karyokinetic figures lie with their polar axes perpendicular and their equatorial planes parallel to the long axes of the cells.

The formation of a cell-plate starting from one of these obliquely placed spindles is a very interesting phenomenon. The spindle becomes greatly extended laterally by the addition of peripheral fibers and gradually assumes the curved form shown in (E). As more peripheral fibers are successively added the remains of the central fibers disappear from about the cell-plate, leaving two separate aggregations of fibers which are connected by the first formed portion of the cell-plate (F). These aggregations of kinoplasmic fibers, which may be called kinoplasmasomes, have a very characteristic form and structure. They extend across the cell-at right angles to its longitudinal axis-from one radial wall to the other (C), and are located in the centre of the protoplast midway between its tangential surfaces (F). In sectional view (F), they have a somewhat wedge-shaped outline, bluntly convex in front and tapering to a point at the rear along the cell-plate. The kinoplasmasomes move in opposite directions towards the ends of the cell (B and C). As they move forward, the cell-plate is extended until it eventually reaches the two ends of the cell, thus dividing the protoplast into two similar portions each of which contains one of the daughter nuclei. The latter remain close together near the centre of the cell during the process of cell-plate formation. The writer has been unable to demonstrate any visible connection between the daughter nuclei and the kinoplasmasomes or their constituent fibers. Except at the beginning of cell-plate formation the daughter nuclei are in a semi-'resting condition' and are provided with a clearly defined nuclear membrane and numerous nucleoli. Not infrequently the distance traversed by the kinoplasmasomes, in passing from the vicinity of the daughter nuclei to the ends of the protoplast, may be from one to several millimeters.

This type of cell division, in which the process of cell-plate formation is so greatly extended--both as regards space and time-and so clearly dissociated from the usual phenomena of karyokinesis, promises, upon further analysis, to be of some significance in the discussion of the dynamics of cell division.

1 Schacht and Russow claimed to have seen several nuclei in the cambial cells of Pinus. ? The writer is dealing with the normal longitudinal divisions of the cambial initials.

ON THE POSSIBLE FORM OF THE EQUATION OF STATE OF

POWDER GASES

By A. G. WEBSTER

BALLISTIC INSTITUTE, CLARK UNIVERSITY, WORCESTER, MASSACHUSETTS

Communicated May 14, 1919

It has been customary for ballisticians to make use of the equation proposed by Clausius,

in the simplified form, suitable for the high temperatures concerned,

(1)

(2)

At the same time it is customary to make use of the experimental results of Mallard and le Chatelier and of Berthelot and Vieille on the specific heats which state that C, is a linear increasing function of the temperature. While apparently no experiments have been made on C, it is assumed that the difference of the specific heats is constant, as in the case of an ideal gas.

It has occurred to me to examine the question of the most general form possible for the equation of state that shall permit of variability of the specific heats, but maintain the constancy of their difference. This question does not appear to have been treated,

By an application of the two laws of thermodynamics we obtain the wellknown equation

If we use the usual letters for differential equations, putting x for v, y for p, z for T divided by C, C, supposed constant, and as usual p for dz/dx, q for dz/dy we have the very simple partial differential equation,

This may be very simply integrated by Cauchy's method, which consists in integrating the system

where the capital letters represent the derivatives of F with respect to the corresponding small letters, and u is an extraneous parameter. Having found

Contribution from the Ballistic Institute, Clark University, No. 5.

five integrals, with five arbitrary constants xo, yo, zo, Po, go we make the latter functions of a second parameter v satisfying the equations

Instead of adopting Cauchy's form for the introduction of the arbitrary function, we will attempt to pass the integral surface through the plane zo = const., representing an isothermal. We put

If we adopt the Clausius equation for the form of one particular isothermal, we may put

=

2a

(8)

so that we have the parametric equation of the surface. It may be noted that putting u = 0, zo T we fall back on the ordinary Clausius equation (1) as a particular case, with (2) and the ideal gas equations as still more. particular.

In order to obtain the expression for the energy for such a gas, we make use

We have now to make use of equations (6) in which, replacing the usual thermal notation, and now using x and y for the arbitrary parameters,

I have also integrated the equation for the case that the difference of the specific heats is a linear function of the temperature, but this seems not necessary in the light of present experimental data.

To

❤'(y)

(11)

THE RELATIVE ADSORPTION OF MIXTURES OF OXYGEN AND NITROGEN IN COCOANUT SHELL CHARCOAL1

BY HARVEY B. LEMON AND KATHRYN BLODGETT

RYERSON PHYSICAL LABORATORY, UNIVERSITY OF CHICAGO

Communicated by A. A. Michelson, May 19, 1919

It is a well known fact that gases are adsorbed in charcoal with rates and in total amounts that vary in a manner closely related to the boiling temperatures of the gases. A mixture of gases may accordingly have its proportions entirely altered by adsorption. This is the principle of the method of Gehloff for the isolation of atmospheric neon.2 Dewar has mentioned it as a convenient means of extracting a high percentage of oxygen from the air.3 A quantitative knowledge of the manner in which the presence of one gas to situration affects the adsorption of another is of great importance since these are the conditions of use under which charcoal has sprung into prominence in modern warfare.

The experiments herein described deal with the relative adsorption of mixtures of oxygen and nitrogen in varying proportions by a highly activated charcoal prepared in the manner described in a previous report by one of the writers. Relatively large amounts of charcoal are employed with respect to the quantity of gas used so that saturation is in all cases reached in the course of thirty minutes or so. The charcoal which weighed 6.5 grams when saturated with dry air at 20° and 750 mm. pressure was contained in Pyrex glass bulbs which could be outgassed by a diffusion pump. Outgassing was for four and a quarter hours at 582°C. After outgassing the tubes were immersed in liquid air of definite age and temperature. While immersed they were exposed to the given gas mixture contained in a constant volume of 975 cc. The initial pressure of the mixture was 73.95 cm. and observations were taken of it at intervals while the adsorption was going on and until it had ceased to fall. A barometer and McLeod gauge formed a part of the above mentioned volume for this purpose.

In figure 1 is given the data in graphical form where log p in cm. is plotted against log t in minutes after the adsorption began. The same sample always is included in this record, two others were used as controls. It is noted that the logarithm of the pressure reached by saturation is almost in a linear relation with the percentage of oxygen in the mixture.

This is shown in figure 2 as the line NO. Here log final pressure is plotted against percentage of oxygen.

The other lines of this diagram NN' and OO' show the final pressures of amounts of pure nitrogen and pure oxygen equal to those existing in the mixture at corresponding values of abscissa but adsorbed separately.