At the same period, 9 a. m., the minimum of the preceding night, (by a self-registering minimum thermometer) is noted and put down; also the temperature of the dew-point, ascertained by Daniell's hygrometer. This does not take more than five minutes. Again, at eleven p. m., or just before going to bed, the observation is repeated; at this time noting the maximum (by a self-registering maximum thermometer) of the preceding day.

At each observation, the degree at which a thermometer (which should always hang close to the barometer) stands must be recorded and placed just below the barometric height, as shewn above, so that the correction for temperature may at any time be made.-Miscellaneous remarks can be scribbled down at the same time.In this way, all things necessary for the purposes of any comparison are recorded, and may be made available at any future period of leisure. At the evening observation, the dew-point is not essential, and we have generally omitted it; for the ascertaining it, is a little experiment not readily performed by candle-light.

The Meteorological Table of the Royal Society is now published monthly in the Athenæum; and a register kept in the way recommended above, would enable any person to institute, with the greatest correctness-by the assistance of Daniell's Meteorological Essays -a comparison on the following daily particulars: Barometer, thermometer, and hygrometer at 9 a. m.; the height of the former by aid of the temperature placed below the height at which it stands, can be reduced to 32. The maximum and minimum of the 24 hours can also be compared, together with the state of the wind and weather.

By devoting an hour of leisure to the purpose, the mean results for each month may be deduced in the usual manner, and all necessary correction duly estimated and applied.

With respect to the instruments which should be employedThe Barometers commonly met with are by no means applicable to observations pretending to any accuracy. In the shops of the best manufacturers and opticians no two barometers agree: the height of the mercury is never actually measured in them-they are graduated one from another, and their errors thus unavoidably perpetuated. Few of them have any adjustment for the change of

level of the mercury in the cistern; no neutral point is marked upon them, nor is the diameter of the bore of the tube ascertained; and in some the capacity of the cisterns is perpetually changing from the stretching of a leathern bag.

We have endeavoured, as far as possible, to obviate these difficulties in the following manner—

1st. Procure one or two of the best barometer tubes, and a sufficiency of clean mercury; take care that the interior of the tube is quite clean and dry; heat it gradually before the fire;-strain the mercury through a fine clean and dry silk handkerchief, and then heat it quite hot, or what is better, just allow it to boil; when cool enough, strain it again through the silk handkerchief into a dry warm 2oz. glass apothecary's measure: 2ndly, then fill the warm dry tube with it; whatever air-bubbles are seen, may be easily removed by passing the largest up and down the tube.— Again, 3rdly, fill the tube quite full, and placing the finger firmly over the bore, plunge it into a cup of clean warm mercury; on removing the finger, the mercury falls a certain distance from the top. Before taking the tube out of the mercury, place the end of the finger again firmly over the bore, and then remove it. The space above the mercury is now occupied by very rarefied air, and by passing this three or four times very carefully up and down, keeping the finger close on the open end, all the air sticking in any part will be removed. Then allow the mercury (the finger being still upon the open end of the tube) to run slowly down to the closed end, taking care that no air-specks are any where visible; then remove the finger, fill the tube quite full, and repeat the operation twice or thrice. Lastly, carefully fill the tube, and remove by a very slender piece of whalebone or strip of quill every speck of air that appears at the junction of the two portions of mercury, or elsewhere, above it. Again fill the tube quite full, and seal it securely with sealing-wax-placing a piece of gummed paper over the sealing-wax.

Now take a quire of long, smooth, clean, white paper, that has been exposed some little time to the air,-in order that it may not contract by drying-and, drawing a pencil line along it, measure, very carefully, 14 inches, which, by a pair of compasses, may be

taken from the scale you intend to apply to your barometer,-measure it two or three times, to be perfectly correct; put a minute speck of ink at each end of the measured distance; with a fine file make a mark upon the tube, about an inch or an inch and a half from the end now closed with sealing-wax; place this in exact opposition with one of the ink marks upon the paper, and laying the tube along the pencil line, make a little mark upon it by a touch of the file, exactly corresponding with the other ink mark upon the paper. By repeating this operation, in the most careful manner, you will have 28 inches accurately measured upon the tube. Then 1st. by adapting an ivory point to the cistern, the extremity of which must exactly correspond to the lower file-mark upon the tube,-2nd. fixing the 28th inch division of the scale opposite the uppermost file mark,—and 3rd., keeping the surface of the mercury in the cistern always just touching the ivory point, the indications of your barometer may be fully relied on.* The construction of the cistern is described in the second volume of "The Analyst,” p. 217.

After the cistern has been fixed, and before entirely closing the bottom of it, of course, the sealing wax and gummed paper must be removed from the end of the tube.

With regard to Thermometers. For the purposes before stated three will be necessary: one-a self-registering, of Rutherford's construction-for the minimum; another self-registering for the maximum; and a third to hang close by the barometer. These should be purchased of some philosophical instrument maker, on whom reliance can be placed; where, also, the Hygrometer should be obtained.

The corrections for the temperature of the mercurial column will be found in a table, p. 372, and for the bore of the tube, p. 363, of Daniell's Essays. And when these are applied, the constant difference between the mean of any two barometers, within short distances, is attributable to elevation.

72

ON THE NON-PERMEABILITY OF GLASS BY

WATER.

THE following experiments were made in the North and South Atlantic and Indian oceans, by Mr. Rudder, late of Birmingham, during his passage to New South Wales, with the view of ascertaining whether water passed through the pores of glass bottles, sunk at different depths in the ocean. The result of the experiment with the bottle, No. 4, prepared after the manner described in the annexed diagram, and sunk 150 fathoms, clearly proves the

No. 1.

E, Piece of wood supporting the cork H at A, and resting against the inside of the bottle at B.-H, cork.-G, sawdust and pitch.-I, wine.-J, canvass passing up the neck of the bottle to the outside, and covered with boiling pitch. C, wooden plug, driven in with melted pitch upon the body J. -D, iron nail passed through the plug, C, over which a piece of canvass was passed, and that with the canvass at J, pitched over and firmly secured to the neck of the bottle with string, likewise pitched over, as at F. No. 4.

1, Piece of wood supporting the cork, 4, at 2, and resting upon the bottom of the bottle at 3.-5, layer of melted pitch. 6, cork driven in with melted pitch, and afterwards covered over with a thick layer of pitch.

7, cap of sheet lead, driven upon the top of the cork and pitch, 6, and bound down with copper wire and string at 8.-9, outside coating of pitch, extending from the rim of the lead, covering over the neck of the bottle.

non-permeability of glass,-a fact which is corroborated by a series of experiments subsequently made by Mr. Wickenden, and described, by that gentleman, in a very lucid manner, on reading the Paper -of which the following is an abstract-before the members of the Birmingham Philosophical Institution.

A wine-bottle was prepared as described in the annexed diagram, No. 1; another was likewise thus secured :-a piece of wood was

placed in the bottle to support a cork, over which was inserted sawdust and melted pitch; a wooden plug, dipped in hot pitch, was driven in, and through the top of the plug a nail was passed transversely, secured with twine; the whole was then coated over with melted pitch. Mr. Benson, the surgeon of the ship, Princess Victoria, also prepared a bottle as follows:—a cork was tightly driven in and covered over with bladder and leather, after which it was pitched and a wooden cap placed closely over the mouth and neck, secured with a piece of canvas. The three bottles were then sunk to the depth of 58 fathoms; when hauled up, the bottle, No. 1, contained about an ounce of water; the other two bottles were perfectly free from any liquid. The bottles were again lowered in 83 fathoms without any addition to the water, and a third time in 112 fathoms; when, on being carefully examined, No. 1 was found one-fourth full, and the water discoloured with the wine; the second bottle contained about an ounce of water: Mr. Benson's was, also, two-thirds full, with the cork driven in and floating on the surface, and the water much agitated, having the appearance of champagne when the cork is first drawn, the air producing a loud hissing in its effort to escape. In making this experiment the lines were impelled towards the west, by a strong current setting in from the east, which rendered it necessary to back the boat to the west, with considerable rapidity, in order to sink the lines perpendicularly. On the bottles being lowered a fourth time to the same depth, viz., 112 fathoms, Mr. Benson's bottle, when brought up, was quite full of water, and the cork replaced in its original position; the water in bottle No. 2 was not increased, but the nail, passed through the plug, was very considerably bent inward, and the top of the cork and pitch brought nearly on a level with the glass. The bottle No. I had received an additional quantity of water, making it one-third full, but there was no perceptible change in the outer coverings.

Another experiment was subsequently made by sinking a bottle, prepared as described in figure 4, in the subjoined diagram, to the depth of 142 fathoms, sustaining a pressure of about 426 pounds on the inch. When drawn up, it was found that no water had penetrated the bottle, and the only perceptible alteration was an indentation of about one-fourth of an inch in the lead, produced by the excessive pressure.