In point of principle, too, nothing prevents our retaining the electrostatical units in the domain of galvanic electricity and in measuring, for example, the strength of a current by the number of electrostatic units which flow per second through its cross-section. But this would be in a double aspect impractical. In the first place, we should totally neglect the magnetic facilities for measurement so conveniently offered by the current, and substitute for this easy means a method which can be applied only with difficulty and is not capable of great exactness. In the second place our units would be much too small, and we should find ourselves in the predicament of the astronomer who attempted to measure celestial distances in metres instead of in radii of the earth and the earth's orbit; for the current which by the magnetic C. G. S. standard represents the unit, would require a flow of some 30,000,000,000 electrostatic units per second through its cross-section. Accordingly, different units must be adopted here. The development of this point, however, lies beyond my present task. FOOTNOTES: [Footnote 26: A lecture delivered at the International Electrical Exhibition, in Vienna, on September 4, 1883.] [Footnote 27: If the two bodies were oppositely electrified they would exert attractions upon each other.] [Footnote 28: The quantity which flows off is in point of fact less than _q_. It would be equal to the quantity _q_ only if the inner coating of the jar were wholly encompassed by the outer coating.] [Footnote 29: Rigorously, of course, this is not correct. First, it is to be noted that the jar _L_ is discharged simultaneously with the electrode of the machine. The jar _F_, on the other hand, is always discharged simultaneously with the outer coating of the jar _L_. Hence, if we call the capacity of the electrode of the machine _E_, that of the unit jar _L_, that of the outer coating of _L_, _A_, and that of the principal jar _F_, then this equation would exist for the example in the text: _(F + A)/(L + E) = 5_. A cause of further departure from absolute exactness is the residual charge.] [Footnote 30: Making allowance for the corrections indicated in the preceding footnote, I have obtained for the dielectric constant of sulphur the number 3.2, which agrees practically with the results obtained by more delicate methods. For the highest attainable precision one should by rights immerse the two plates of the condenser first wholly in air and then wholly in sulphur, if the ratio of the capacities is to correspond to the dielectric constant. In point of fact, however, the error which arises from inserting simply a plate of sulphur that exactly fills the space between the two plates, is of no consequence.]