A series of very interesting experiments, fully detailed in this memoir, which were instituted for the purpose of ascertaining the extent of the diminution of electrical action in instances of imperfect or irregular conducting surfaces, led him to the general conclusion, that a very small quantity of the imperfect or fluid conductor was sufficient to transmit the electrical power, or to complete the chain. This induced him to try whether copper, if nailed upon wood, and protected merely by zinc or iron on its _under_ surface, or on that next the wood, might not be defended from corrosion: a question of great practical moment with regard to the arrangement of protectors. For this purpose, he covered a piece of wood with small sheets of copper, a nail of zinc of about 1/100 part of the surface having been previously driven into the wood: the copper surface remained perfectly bright in sea-water for many weeks; and when the result was examined, it was found that the zinc had only suffered partial corrosion; that the wood was moist, and that, on the interior of the copper there was a considerable portion of revived zinc, so that the negative electricity, by its operation, provided materials for its future and constant excitement. In several trials of the same kind, iron was used with similar results; and in all these experiments there appeared to be this peculiarity in the appearance of the copper, that unless the protecting metal below was in a large mass, there were no depositions of calcareous or magnesian earths upon the metal; it was clean and bright, but never coated. The copper in these experiments was nailed sometimes upon paper, sometimes upon the mere wood, and sometimes upon linen; and the communication was partially interrupted between the external and internal surfaces by cement; but even one side or junction of a sheet seemed to allow sufficient communication between the moisture on the under surface and the sea-water without, to produce the electrical effect of preservation. This last experiment of Davy is of greater importance than may at first appear, in showing what a small proportion of conducting fluid will complete a circuit, and in thus explaining phenomena, as I shall presently show, which might not otherwise be suspected to have an electrical origin.