It has always, however, been subject to certain faults. To commence with, it points, not to the geographical north, but to the "magnetic pole," a point some distance from the geographical pole, and one, moreover, which is not quite permanent. The fact that the magnetic pole varies its position is impressively shown by the fact that a special department at Greenwich Observatory is continually employed, by the aid of delicate self-recording instruments, watching and setting down its fluctuations. And the premier observatory of the world, it should be remembered, exists primarily, not in the interests of pure science, but as a department of the British Admiralty in order to study matters of interest to navigation. Thus we have testimony to the importance of these little vagaries on the part of the magnetic compass. But in addition to these inherent faults there is a new source of error in the magnetic compass which man has introduced himself by making his ships of iron instead of wood. Every ship of the present day is a huge magnet. A piece of iron left in the same position for a length of time becomes polarised, which is to say that it acquires the properties of a magnet; and two magnets always exert an influence upon each other. Consequently the ship, after lying for perhaps a year in one position, during the period of building, becomes itself magnetic and interferes with its own compass. Then, again, our methods of ship construction aggravate this trouble. It is believed that every molecule of iron is itself a minute magnet with a north and south pole of its own. These lying in confusion in the mass of unmagnetised iron neutralise each other, so that the mass, taken as a whole, does not exhibit any magnetic power. But if by some means the whole of the millions of millions of molecules can be set the same way--with all their north poles in one direction, and their south poles in the opposite direction--then they will all act together. Instead of neutralising each other they will then help each other, and under those conditions the mass of iron will possess that peculiar power which is distinctive of a magnet. So long as a piece of iron is left in the same position the magnetism of the earth is thus acting upon the molecules. Just as it tends to place the compass needle north and south, so it does with every molecule in the iron mass. And if, while lying still, the iron be hammered, the shaking of the molecules due to the hammering loosens them as it were and assists the earth's power in pulling them into position. One has only, then, to watch the riveting up of a ship, and to see the vigorous way in which the riveters wield their hammers, to realise that when the thousands or even millions of rivets have all been finished the material of that ship will have had the very best possible chance of becoming magnetic.