In telegraphy, as well as in the question of lightning rods, attention has been but incidentally paid to the improvement of ground conductors, and this point has not been the object of that careful study that has been bestowed upon the establishment of aerial lines. It is only recently that the interest created by lightning rods has given rise to new forms of conductors differing from those formerly used. The publications of the Prussian Academy of Sciences of from 1876 to 1880 contain some information of special importance in regard to this. It is stated therein that the effect of ground conductors may be notably increased by the division of the earth plates and the use of metallic rods, without necessitating a greater output of material. These facts, however, have not as yet been put to profit in practice for the reason, perhaps, that the considerations, which have remained general, have not at once permitted of obtaining forms what could be employed with perfect knowledge of the results. This is what led Mr. Ulbricht, of Dresden, to make calculations for a few forms of conductors, and to test their approximate values. The results of these researches are printed in the _Elektrotechnischen Zeitschrift_ for 1883 (p. 18). [Illustration] The equations found show, in the first place, that there exist three means of obtaining a considerable effect, as regards the ground conductor, with a slight expenditure of material: The cylindrical electrode may be drawn out into the form of a bar or wire; the plate may be rendered narrow, and elongated in the form of a ribbon; and, besides, the annular plate may be enlarged in lessening the metallic surface. Finally, a short, open cylinder with a vertical axis may be formed by curving a narrow plate or ribbon. It is not necessary to see the formula to recognize the fact that this cylinder must behave like a ribbon and a flat ring. The radius increasing, and the surface remaining constant, the resistance of the earth here likewise approaches zero. As the resistance of the earth is inversely proportional to the diameter of the plates, the zero resistance can also be reached by dividing a plate _ad infinitum_. As the parts of the plate may be brought quite close to each other without perceptibly interfering with the action, a _network_ has finally been reached by a division carried very far, yet limited, and by connecting the parts with one another by conducting cylinders.