The sphere can be raised and lowered by means of clockwork, an electric motor, or a water motor. The laboratory has a small water motor, which was used for this purpose by attaching an arm to the circumference of the wheel and fastening the end of the string to this arm. When the arm is in a certain position of its revolution, the sphere rests lightly on the bottom of the cup. When the arm has turned 180° from this position the sphere touches the plate.

The sphere, cup, and plate must be of the same metal. As the quadrants of the electrometer are of brass, we made these of brass to avoid all contact electricity.

Comparative observations have been made with this collector and the water-dropper for a month. The changes seem to be similar, but the deflections of the water-dropper are the larger.

It seems possible, with some mechanical improvements, to make this form of collector superior to any other.

OBSERVATIONS.

The observations show that

The potential of the air was generally low and positive, seldom as high as 25 or 30 volts.

The potential usually fell before precipitation, storms, or when the relative humidity increased.

The potential during precipitation, with a very few exceptions, was always low and positive.

Almost all the negative electricity, except that which was followed by precipitation, occurred during west to northwest gales, or during cold waves.

Low clouds sometimes seemed to affect the observations, but high clouds seemed to have no influence.

There was very slight variation with altitude, at least, between two and ten meters above the ground.

There was no appreciable variation between collectors placed on different sides of the building.

ON OBTAINING THE ELECTRIC POTENTIAL OF THE UPPER AIR.

On the morning of May 6th, the potential of the air at a point ten feet above the ground and three feet from the walls of the laboratory, obtained by the usual water-dropping method, was, reduced to volts, 0.5. A paper kite covered with cloth and tinfoil, with its longest axis about four feet, was flown, the connecting string being heavy English

twine, previously soaked in a mixture of glycerine and water. The end of this string was connected to a wire well insulated, which in turn was connected to one set of the quadrants of the Trowbridge electrometer, the other set of quadrants being connected with the ground. The needle was connected to the positive pole of a Beetz solid battery of 200 cells (200 volts). The needle was at once deflected to its limit, indicating a high positive potential for the air at an elevation of less than 300 feet. Remaining for a few seconds at this high positive, it would suddenly change to an equally high negative, sometimes without the least warning. It was, without doubt, extremely variable. The high positive indications seemed to be more prevalent. Connecting the kite-string with the multiple quadrant electrometer, described in this paper, the following results were obtained. The connection and charge of the needle were as in the other instrument. A fine index-pointer records the deflections in this instrument, and the mirror, scale, and dark room are dispensed with. A Daniell cell gives a deflection of half a degree. The deflection given by the kite was at times over 25 degrees in a positive direction, or equivalent to over 100 volts. The index-hand was seldom still, as in the previous case evidencing an extreme variableness of the electrical condition of the air at that place and time. The wind was from the east, steady and light, the pressure 30.061, the temperature 49° F., the relative humidity 77, and the sky covered with a low pallium of stratus clouds moving from the east slowly.

On the next day, May 7, the kite was again flown, this time reaching an altitude of about 500 feet. The potential of the air at a point ten feet from the ground, obtained by a water-dropper, reduced to volts, was 0.4.

The table on the following page shows the deflections for short intervals. These deflections were comparatively steady, and had not the variableness of those on the preceding day. The wind was east, and had now been blowing from that quarter for nearly thirty hours. The sky was covered with stratus clouds, having the unusual appearance of billows with the crests pointing to the earth. The pressure was 30.040, the temperature 45° F., the relative humidity 75.

The experiment demonstrates that it is comparatively easy to obtain some indication, even if it be only a relative one, of the potential of the air at high altitudes. The method is simple and direct, and with the exception of the original cost of electrometer and charging battery, quite inexpensive. A series of simultaneous observations of this character would doubtless be of value in meteorology.

INVESTIGATIONS ON LIGHT AND HEAT, MADE AND PUBLISHED WHOLLY RIN PART WITH APPROPRIATION FROM THE RUMFORD FUND.

XXII.

THE EFFECT OF TEMPERATURE ON THE MAGNETIC PERMEABILITY OF IRON AND COBALT.

BY JOHN TROWBRIDGE AND AUSTIN L. MCRAE.

Communicated May 13, 1885.

THESE experiments were made with a view, –

1. To repeat Professor Rowland's experiments.*

2. To determine the change of the magnetic permeability with

a. The change of temperature.

b. The variation of the size and thickness of the rings.

c. The temper of the substance.

3. To determine whether the outer layer of the substance shielded the inner layers from magnetic influence.

The method of experiment was the same as Rowland's, except that the small coil around the horseshoe magnet used to bring the needle to rest was taken out of the circuit of the induced current, and a thermal junction of copper and german silver was placed in the lamp used with the mirror galvanometer, and the circuit made through the galvanometer and a key. This circuit was independent of the induced current circuit. By closing the key, a thermal current could be sent through the galvanometer. This arrangement served the same purpose as the sliding coil on the magnet, and was more convenient.

The measurements are made in the C. G. S. system of absolute units.

The symbols used are:

Hintensity of the magnetizing force.

B = magnetic induction, or the number of lines of force passing through a unit of area.

M = magnetic permeability.

* Phil. Mag., 1873, 1874.

Ttemporary part of B.

N

α

total number of coils on the ring magnet. = mean radius of the ring.

= radius of cross-section of the ring.

γ = strength of the current.

Nnumber of coils on the helix.

N' number of coils on the earth inductor.
Rradius of earth inductor.

C

h

h'

V

deflection of earth inductor.

deflection of the helix when the current is reversed.
deflection of the helix when the current is broken.
vertical force of the earth's magnetism.

8 = deflection of the tangent galvanometer.

The magnetic potential inside of a ring solenoid is

=-2 Ny0,*

where is the azimuth angle about the axis of the ring. Differentiating, we have

the distance from the axis; hence the magnetic force

To find the total number of lines of force passing through any area, we find an expression for the force at a single point, and then integrate this expression all over the surface between the proper limits.

If we take a section of the ring and divide it into bars of length BC and width d x, we shall have BC= 2 √ r2 — x2, and we shall find the total number of lines of force through the section to be equal.

+ 2 Noy M 2 √r2x2dx=4π N ̧y M(a-√a2 — r2)

since in the rings used the parenthesis differs but little from unity.

*Maxwell's Elect and Mag., sect. 681.