ELECTRIC WAVE TELEGRAPH Y—FLEMING. 189 
the earth. (See fig. 17.) Let a current be supposed to flow round 
it, in clockwise direction. Then it creates a magnetic field, the 
direction of which along the surface of the earth in a direction at 
right angles to the plane of the circuit, and at equal distances from 
the center, is toward the spectator on both sides. Suppose, then, 
that a wire equal in length to one side of the square is placed in 
contiguity to one vertical side, and that it carries a current oppo- 
site in direction to that in the side of the square (say, the right- 
hand side) to which it is in proximity. Then the magnetic field of 
this straight current is from the spectator at the right-hand side and 
to the spectator on the left-hand side. Accordingly, the total field on 
the right-hand side, due to the currents in the closed and open cir- 
cuits together, is less than that on the left, because the individual 
fields are added on one side and subtracted on the other. Since the 
two oppositely directed currents in the adjacent wires may be imag- 
ined to come so close as to annul 
each other, and since the parts of 
the remainder below ground may 
be considered to be removed with- 
out affecting the field above ground, 
we arrive at the conclusion that an 
antenna partly vertical and partly 
horizontal radiates most strongly 
in the direction opposite to that in 
which the free end points. 
Mr. Marconi discovered this fact 
experimentally, and made meas- me 
urements of the currents induced 200° 150° age 1792160" 
in receiving antenna placed at 
equal distances round this bent 
transmitter, and plotted the results 
in the form of a polar curve. (See fig. 18.) As a quantitative receiv- 
ing detector he made use of a Duddell’s thermal ammeter. In repeat- 
ing and confirming these experiments on a smaller scale last summer 
in the grass quadrangle of University College, I employed a form of 
thermal ammeter of my own design, made as follows: A vacuum 
vessel made like those which Sir James Dewar devised for storing 
liquid gases has four platinum wires sealed through the bottom of the 
inner test tube. One pair of these is connected in the vacuous space 
by an extremely fine constantin wire and the other pair by a fine 
tellurium-bismuth thermo-junction, with the junction resting on the 
fine wire. (See fig. 19.) When a galvanometer of suitable resist- 
ance is connected to the terminals of the thermo-junction and the con- 
stantin wire inserted in the circuit of the receiving antenna we have 
an arrangement which enables us to measure as well as detect the in- 
tensity of the electric waves incident on the antenna. This detector, 
Fie. 18.—Radiation in various azimuths 
from Marconi bent antenna. 
