About this time Graham invented the Mercurial Compensation pendulum, which consisted of a glass or iron jar filled with quicksilver and fixed to the end of the pendulum rod, which, when heat lengthened the rod, expanded simultaneously the quicksilver, and made the centre of oscillation to continue at the same distance from the point of suspension. He afterwards conceived a notion, which John Harrison subsequently worked out, of making a compensation pendulum (or a pendulum that should in itself contain the power of equalizing its own action, whatever the change of temperature), forming it of various metals. In 1726 Harrison invented what is called the gridiron pendulum, composed of nine rods, five of steel and four of brass, which are so arranged that those which expand most are counteracted upon by those of less expansion. These two compensation pendulums, the gridiron and mercurial, are still in use, and with slight improvements are found to keep to time very accurately. The period had now arrived for the making of marine time-keepers sufficiently accurate for nautical use, and styled chronometers because they are most accurate time-measurers. Their value to navigators, and the immense impetus which would by such instruments be given to the science of navigation, had long been foreseen, but there were many great difficulties in the way of obtaining a perfect chronometer. The sailor, before the invention of this instrument, could ascertain the latitude of his ship at sea, by observation of the fixed stars. Supposing these stars to have first appeared to him in the zenith, and at his next observation to be one, two, or three degrees south of the zenith, he would know that he had sailed just so many degrees north of the place in which he first observed them. It was not, however, so easy for him to compute longitudes, because the diurnal revolution of the earth causes each meridian to pass successively under the same stars. It was necessary to have an accurate time-keeper, and to set it carefully to the solar time of some port in the kingdom, whose longitude was well known. The time-piece might then be carried out in a vessel sailing abroad, and the computations made by means of it would prove most wonderfully exact and important. By simply observing the moment at which the sun reached his meridian, when of course it would be 12 o'clock at noon, solar time, and then noting the difference between the solar time thus ascertained and the time of the chronometer, the mariner would be able by calculating 15 degrees to one hour of time, or 15 geographical miles to one minute, to make out his longitude. For example, if the time-piece had been set to time at the meridian of Greenwich observatory, and if it be one o'clock by the time-piece when it is mid-day, or meridian by the sun, then the place in which the longitude is taken must be in long. 15 degrees east of the meridian of Greenwich, and if it be eleven o'clock by the chronometer when the sun