But Gilbert is astonishing when he goes on to make thick boxes of gold, glass and marble; and, suspending his needle within them, declares with excusable enthusiasm that, regardless of the box which imprisons the magnet, it turns to its predestined points of north and south. He even constructs a box of iron, places his magnet within, observes its behavior, and concludes that it turns north and south, and would do so were "it shut up in iron vaults sufficiently roomy." In this, he was in error, for experiments show that if the sides of the box are thin, the needle will experience the directive force of the earth; but if they are sufficiently thick--thick as the walls of an ordinary safe--the inside of such a box will be completely screened; none of the earth's magnetic lines will get into it so that the needle will remain indifferently in any position in which it is placed. Some years ago, the physical laboratory of St. John's College, Oxford, was screened from the obtrusive lines of neighboring dynamos by building two brick walls parallel to each other and eight inches apart and filling in the space with scrap iron. A delicate magnetometer showed that such a structure allowed no leakage of lines of force through it, but offered an impenetrable barrier to the magnetic influence of the working dynamos. Gilbert's greatest discovery is that the earth itself acts as a vast globular magnet having its magnetic poles, axis and equator. The pole which is in our hemisphere, he variously calls north, boreal or arctic. Whilst that in the other hemisphere he calls south, austral or antarctic. He sought to explain the magnetic condition of our globe by the presence, especially in its innermost parts, of what he calls true, terrene matter, homogeneous in structure and endowed with magnetic properties, so that every separate fragment exhibits the whole force of magnetic matter. He is quite aware that his theory is a grand generalization; and admits that it is "a new and till now unheard-of view," and so confident is he in its worth that he is not afraid to say that "it will stand as firm as aught that ever was produced in philosophy, backed by ingenious argumentation or buttressed by mathematical demonstration." In developing his theory of terrestrial magnetism, Gilbert fell into certain errors, chiefly for want of data, but partly also by reason of his adherence to the view that the earth exactly resembled his terrella in its magnetic action. Accordingly, he believed that the magnetic poles of the earth were diametrically opposite each other and that they coincided with the poles of rotation, whence it followed that the magnetic meridian everywhere coincided with the geographical, and that the magnet, unless influenced by local disturbances, stood true to the pole.