33 feet of sea. A column of sea-water 33 feet long (or 10 metres, or 5 fathoms) and square inch area, weighs 15 lbs. The atmosphere, mass for mass, is just equivalent to 5 fathoms of sea, but 33 feet depth is a mere fraction of our sca-water. And again, a quantity of water is of much greater capacity for heat than the same quantity of air, the thermal capacity of air being very much less than that of water. Briefly then we may say that in the transport of heat over the whole solid globe the air goes for nothing. The sea is the great carrier. The air, of course, has an

enormous

indirect effect in the shape of gales, moderate winds, or trade winds, and in moving the water. It moves the surface of the water just as it moves the ships over the water. So that wind, indirectly helps the sea as a heat carrier. The wind is the

distributer, the sea is the carrier-the great

carrier-in the transport of heat

ON THE RATE OF A CLOCK OR CHRONOMETER AS INFLUENCED BY THE MODE OF SUSPENSION

[Being a Paper read before the Institution of Engineers in Scotland, February 27, 1867.]

IT is well known that the rate of a chronometer, a clock, or a watch may be altered by altering its mode of support. On land, clocks ought to be fixed in as solid a manner as possible, so as to prevent vibration, either by their own action or from extraneous causes, from being communicated to the supports of the pendulum. Even the best astronomical clocks hitherto made are not well arranged in this respect.

A marine chronometer or watch exhibits in a very striking manner the effects of varying the mode of support. A watch which keeps very good

time when carried in the pocket, or laid on a soft pillow, will go at a different rate if laid on a marble slab, or on a hard board. These variations of rate are not due to any imperfections of the balancewheel or mechanism of the watch or chronometer, but arise from reaction due to the motion of the

moving parts. A well-balanced watch will go equally well whether supported in a vertical or horizontal plane; and a well-made watch will, I believe, not be subject to uncertainty of above a quarter of a second per day, if carried about in the pocket all day and put under the pillow at night. This I can testify from experience of a good pocket-watch which I have tried now for nearly two years; indeed, a good pocket-watch, if well treated, is comparable in its performances with the best marine chronometer.

I was very much struck some time ago by a remark made to me by Mr. Archibald Smith, of Jordan Hill, regarding a demi-chronometer, with detached lever and compensated balance, presented to him by the Admiralty for the voluntary assistance he had given them in working out methods

for adjusting the compasses of iron ships. Mr. Smith found that this watch was going well, until one day he observed it had gained fifteen seconds, the reason of which he could not explain until he had recollected that instead of its having been put under the pillow as usual, it had been hung up in a suspended watch case.

The question now arises, What is the cause of these variations, and how on dynamic principles are they to be explained? The dynamics of the subject are indeed very simple, and can be easily reduced to a well-known general problem.

A simple pendulum when it vibrates through a very small arc, vibrates according to the law of simple harmonic motion. Take a spiral spring, with a heavy weight hanging by it, stretch it a little and let it go, and it vibrates according to the same law. The vibrations of a tuning-fork, or any other instrument giving a similar musical sound, are also according to the law of simple harmonic motion. A case of roughly approximately simple harmonic motion we have when the piston moves to and fro in a cylinder, the head of the piston-rod

being guided by a cross-head and slides, and the crank and fly-wheel making one revolution for every backward and forward movement of the piston. The balance-wheel of a watch, vibrating to and fro through a certain angle, performs very approximately a simple harmonic motion. The longer the hair-spring is, the more nearly it will approach to simple harmonic motion, and it will keep time the more accurately.

Now, against every change of motion of a body there is a certain reaction, and every motion to and fro of the balance-wheel of a watch or chronometer reacts upon the case of the watch or chronometer; and if the case is so suspended as to be free to vibrate, the motion of the balance-wheel will generate a vibration of the whole, so that we have two motions to consider-one, that of the balance-wheel inside the watch; the other, that of the whole watch except the balance-wheel. Upon the mode of suspension of the watch or chronometer will depend the nature of the vibration which it takes up and the resultant effect upon the The rate is accelerated or retarded according

rate.