results and that refraction, which always makes the sun appear a little too high, throws the shadow a trifle toward noon at all times. That is, the time is a little too fast in the morning and too slow in the afternoon. More than that, a correction is always necessary in order to find civil, or clock, time.

The simplest form of sun-dial is the best, and as a regulator of clocks the dial is good within one or two minutes. The "noon mark" is simply a north-and-south line marked on a horizontal plane and the style is any object fixed to

the dial and slanted so as to point to the north pole. On four days of the year the sun is right with mean time and the shadow mark may be set on those days, or on other days the noon mark may be set by consulting the table in the almanac which shows the variation of the sun from civil time in even minutes. Thus on October 10, 1909, the noon mark could be made by the shadow of the style at 11.47 by the clock and it would be right for all time to

come.

A device less dependent upon the climatic conditions was the water-clock, or clepsydra. It is said that this instrument was in use among the Chaldeans and ancient Hindus. Sextus Empiricus says that the Chaldees

used such a vessel for finding their astrological data, but remarks that the unequal flowing of the water and the alterations of atmospheric temperature rendered their calculations inaccurate.

In this instrument the water, which falls drop by drop from the orifice of one vessel into another, floats a light body that marks the height of the water as it rises against a graduated scale and thus denotes the time that has elapsed. As a measure of hours of the day in countries such as Egypt, where the hours were always equal and thus where the longer days contained more hours, the water-clock was very suitable, but in Greece and Rome, where the day, whatever its length, was always divided into twelve hours, the simple water-clock was as unsuitable as a modern clock would be, for it always divided the hours equally and took no account of the fact that by such a system the hours in summer were longer than in winter. In order, therefore, to make the water-clock available in Greece and Italy it became necessary to make the hours unequal and to arrange them to correspond with unequal hours in the Greek day. This plan was accomplished by placing a float upon the water in the vessel that measured the hours, and on the float stood a figure made of thin copper, with a wand in its hand. This wand pointed to an unequally divided scale. A separate scale was provided for every day in the year, and these scales were mounted on a drum which revolved so as to turn round once in the year. Thus as the figure rose each day by means of a cogwheel it moved the drum round one division or one-threehundred-and-sixty-fifth part of a revolution. By this means the scale corresponding to any particular day of winter or summer was brought opposite the wand of the figure, and thus the scale of hours was kept true. In fact, the water-clock, which kept true time, was made by artificial means to keep untrue time, in order to correspond with the unequal hours of the Greek days. One of the more complicated forms of the water-clock was probably

invented by Ctesibus of Alexandria. In the Athenian courts a speaker was allowed a certain number of amphoræ of water for his speech, the quantity dependent on the importance of his suit. Both the simple and more elaborate forms of clepsydræ were introduced into Rome in the second century B.C.

second century B.C. A Chinese water-clock, reputed to be over 3,000 years old, consisted of four copper jars, on ascending steps, with small openings and filled every morning. The purpose of the series was to obviate the irregularity in dropping which would be caused by the greater weight in the first jar at the beginning of the day.

The running of fine sand from one vessel into another was found to afford a still more certain measure of time, so the hour-glass came into being. This instrument consists of two bulbs of glass united by a narrow neck; one

of the bulbs is nearly filled with dry sand, fine enough to run freely through the orifice in the neck, and the quantity of sand is just as much as can run through the orifice in an hour, if the instrument is to be really an hour-glass; in a minute, if a minute-glass. It is said that King Alfred observed the lapse of time by noting the gradual shortening of a lighted candle.

The pendulum is the mechanical basis of modern clocks and was first scientifically investigated by Galileo in the latter half of the sixteenth century. The story runs that while he was praying one day in the cathedral at Pisa his attention was arrested by the motion of the great lamp which, after being lighted, had been left swinging. Galileo proceeded to time its oscillations by the only watch in his possession-namely, his own pulse. He found the times, as near as he could tell, to remain the same, even after the motion had greatly diminished. Thus was discovered the isochronism of the pendulum. Later experiments carried out by Galileo showed that the time of oscillation was independent of the mass and material of the pendulum and varied as the square root of its length.

Galileo's invention did not become generally known at that time, and fifteen years later, in 1656, Christian Huygens independently invented a pendulum clock which met with general and rapid appreciation. The honor of this. invention belongs, therefore, to both Galileo and Huygens.

Wheel-work had been known long before the time of Galileo and had been skilfully applied by Archimedes. When therefore some sort of wheel mechanism was needed to keep the pendulum oscillating, the mechanical means were at hand. Galileo saw that if the pendulum could be kept swinging, a timepiece could be constructed which would be mathematically perfect. There must be some reservoir of force such that when a pendulum comes back and touches it the touch shall allow some pent-up power to escape and to drive the pendulum forward. An arrangement of this kind was contrived by Galileo. He provided

a wheel with a number of pins around it. The pendulum had an arm attached to it and there was a ratchet with a projecting arm which engaged with the pins. This arrangement is called an escapement.

The type of escapement invented by Galileo was, for practical purposes, full of imperfections, and it was left for later inventors to modify his ideas and to improve on

The

them until an accurate timepiece was achieved. balance-wheel was invented, which does the work of the pendulum, and various escapements, such as the crown or verge escapement, the anchor-and-crutch escapement, the dead-beat escapement and the gravity escapement, have all taken their place in the development of the timepiece. The prime requisite of a good escapement is that the impulse communicated to the pendulum be invariable, notwithstanding any irregularity or foulness in the train of wheels. The compensating balance-wheel is a balance