spring, however, which from its simplicity has been | commonly used, is liable to this objection, that it diminishes in elasticity by frequent compression, and thus the scale by which its force is ascertained must be constantly varying. To remedy this defect there has sometimes been substituted for the spring, a bag of air communicating with a glass tube in the form of a lengthened U, containing a liquid which is depressed in one leg and raised in the other in proportion to the compression of the air in the bag, thus affording a measure of the compressing force. Leslie's anemometer depended on the principle that the cooling power of a stream of air is equal to its velocity. Another instrument depended on the evaporation of water, the quantity evaporated in a given time being proportional to the velocity of the wind. The raising of a column of fluid above the general level of its surface is the principle of Dr. Lind's anemometer, Fig. 41. It consists of two glass

Fig. 41.

tubes about 9 inches long andths of an inch in diameter, connected at their lower extremities by another tube of glass onlyth of an inch in diameter. To the upper extremity of one tube is fitted a thin metal cap bent at right angles, so that its mouth may receive the current of air in a horizontal direction. Water is poured in at the mouth till the tubes are nearly half full, and a scale of inches and parts of an inch is placed between the tubes. When the wind blows in at the mouth of the cap, the column of water is depressed in the tube below the cap, and elevated to a similar extent in the other tube, so that the distance between the surfaces of the fluid in each tube is the length of a column of water, the weight of which is equal to the force of the wind upon a surface equal to the base of the column of fluid. The object of the small tube which connects the two larger ones is to prevent the oscillation of the fluid by irregular blasts of wind. The absolute velocity of the wind is deduced from the height of the column of water, or it may be ascertained from the tables constructed for the purpose. Thus, according to Dr. Lind, a column of water 0.025 inch, or a fortieth of an inch high, exerts a pressure of rather more than 2 ounces 1 drachm upon a square foot of surface, and balances the effect of a gentle wind moving at the rate of about 5 feet in a second, or not quite 4 miles an hour. When the column of water is 1 inch high, the force of the wind on a square foot is nearly 5lbs., its velocity 32 miles an hour, and its character a high wind. When the column marks 3 inches, the force is upwards of 15 lbs. on the square foot, the velocity above 564 miles per hour, and the character a storm. At 9 inches the force on the square foot is stated to be 46lbs. 14 oz.; the velocity 97 miles an hour, producing a most violent hurricane. Thus, it will be observed that in the greatest storms, the difference between the atmo

spheric pressures on the windward and leeward sides of any object does not amount toth of the pressure of the leeward side; for that is capable of supporting a column of about 33 feet of water.

Of late years the most common forms of anemometer are those by Dr. Whewell and Mr. Osler. The general arrangement of Whewell's anemometer will be understood from Fig. 42, in which it will be

Fig. 42.

S SE

SE

seen that by means of a vane, a windmill fly is constantly presented to the wind in whatever direction it may blow, and the fly of course revolves with greater or less rapidity according to the velocity of the current. An intermediate train of wheels, set in motion by the fly, causes a pencil to descend over a fixed cylinder, having thereon a trace of variable length, according as the wind is more or less strong. 10,000 revolutions of the fly cause the pencil to descend only of an inch. The surface of the fixed cylinder is japanned white, and is divided into 16 or 32 equal parts by means of vertical lines, the intervening spaces corresponding to 16 or 32 points of the compass, and a mark left by the pencil upon one or more of these spaces, shows the direction of the wind. The pencil has two motions, the first from above downwards, and this increases in rapidity as the wind blows more strongly, and by the extent of its depression registers the whole amount of wind that has been blowing. The second motion depends on the changes in the direction of the wind; and the pencil and its frame being carried round by the vane, the direction is registered by this cross movement. In this arrangement, therefore, the vane, the windmill fly, the intermediate train of wheels and the pencil, all obey the direction of the wind; while the cylinder which marks the points of the compass remains fixed, so that the pencil in descending and moving about with the wind thus traces an irregular line on the cylinder. If the fly revolve in the simple proportion

by the pencil is proportional to the space which would be described by a particle of air in a given direction in a given time, such as one day, taking into account the strength of the wind and the time for which it blows.

The line marked by the pencil upon the cylinder is not a single line, but a broad irregular path. This is occasioned by the wavering of the wind. The vane is in almost constant motion, swinging to and fro through an arc often not less than a quarter of a circle; but the middle of the line which gives the mean direction can readily be ascertained, while the length of the line is in proportion to the product of the velocity of the wind and the length of time during which it blows in each direction, which product is called its integral force.

The amount of friction in this machine is very considerable, arising from perpetual screws working in toothed wheels, for the purpose of converting the rapid motion of the fly into a slow, descending, vertical motion, again carried out by a thread turning within a movable nut. There is also the friction of the pencil attached to this nut pressing with sufficient force so as to leave a trace on the fixed cylinder. Hence when the force of the wind is small the fly would experience a greater amount of comparative retardation than with strong gales.

Osler's anemometer traces the direction of the wind and its pressure on a given area, together with the amount of rain, on a register divided into 24

Fig. 43.

D

the quantity of rain. The whole length of the paper is divided, for the hours of the day, into 24 parts by lines crossing the former at right angles. A new register paper is placed on a board, and accurately set every day, and the board is carried along by means of a clock mechanism, behind three pencils A, B, C, which may be considered as the fingers or indexes of the machine. The board, which is placed upright as in the figure, moves on friction-rollers, and is thus moved along as the time advances.

The pencil в is the index of direction. The method by which the instrument is turned so as to obey the direction of the wind will be seen in Fig. 44. A set of vanes or sails revolve vertically in a plane at right angles to that of the pressure plate P, and drive a cog-wheel, which by rolling on a fixed cogged circle, turns all the rest of the apparatus round till the vanes are presented edgewise to the current, so as not to be turned by it either way, when the pressure plate being at right angles to the vanes is acted on with full effect. As the vanes turn in the direction of the wind, a spiral worm on the shaft near its lower end raises or lowers the nut 1, Fig. 43, which does not turn round, and from which hangs the arm carrying the middle pencil в, which thus traces a mark on one of the long lines of the register if the wind be blowing towards one of the cardinal points, or a mark between these lines, if it be blowing from intermediate points, such as NN.W., N.W. &c., which may be represented by fainter lines parallel

with the others.

The method by which the pressure plate P, Fig. 44, is always made to face the wind, has been already described. This plate is suspended by means of 4

hour are measured, and the pressure of the wind in violent gales also recorded. The motion of the plate is communicated to the register below by a wire connected with the bell-crank c, with another wire descending through the hollow upright shaft and kept stretched by a spiral spring. To this wire is attached the upper pencil c, which thus descends lower the more the pressure plate P is pushed back, and returns to the top of the paper when the pressure ceases. The distance to which the pencil is thus depressed, represents, by means of a scale of parallel lines, shown in Fig. 43, the pressure of the wind in pounds on the area of 1 square foot, or its velocity in miles per hour.

The pencil a registers the rain in a similar manner. The rain after falling into the vessel D on the roof flows into G, one of the two divisions of a gauge balanced on an axis and supported by a second balance. As the water accumulates, this second balance begins to descend, and so raises the upright rod to which the lever FA is attached. This lever FA carries the pencil, which by this action is raised, showing upon the lowest set of parallel lines the quantity of water collected in the gauge. When this quantity becomes equal to a certain depth of rain or to a certain number of cubic inches on a foot square, the small gauge oversets, the water is discharged, and the other compartment H of the gauge is brought under the pipe. The pencil then returns to its first position at the bottom of the paper, and begins to rise on the scale as the rain is collected. In a trace of this kind it will be seen that the more rapidly the rain falls, the sharper will be the angles formed by the trace of the pencil; but if the rain be slow and gradual, the elevating or diagonal lines will be drawn out into a considerable length; and when no rain falls a horizontal line will be drawn, as shown in Fig. 43, from VI to after VIII, and from x to I.

It will be seen, then, from this arrangement, that as the register is constantly and hourly moved along behind the three pencils, a continued record or trace of the direction and pressure of the wind, together with the amount of rain, is left on the paper. Figs. 43 and 44 are intended to convey a general idea of Mr. Osler's arrangement in the Royal Exchange, London, where the register paper is made to last a week. When the Editor visited the Meteorological Observatory at Greenwich a few years ago, he noticed that the register was placed horizontally on a table, and being on a larger scale, it was changed every day. By means of such an instrument we ascertain the direction, the duration, and the force of the wind. It is necessary, however, for the purposes of science that the integral of the wind be deduced for each point of the compass; that is, to ascertain the entire quantity of wind which has blown from each point during a given period. Now if the force of the wind were constant, the integral would be obtained by multiplying the length of time that the wind is blowing, by the velocity with which it moves.

The integral of the wind, or the total quantity as measured by its intensity and duration jointly, may

be thus illustrated. If the intensity or force of one wind be to that of another as 2 to 3, and if the former wind last 6 hours, and the latter only 2, the integral of the former is double that of the latter, for it has blown twice as much air over the place; for supposing the second wind to be opposite to the first, it must blow for 4 hours before it will carry back all the air which the first had blown over. The integral, therefore, is proportional to the product of the mean intensity, or velocity, and the duration multiplied together.

In a similar manner the area of a rectangle is proportional to its length and breadth taken jointly. Now if we have such a figure whose length represents the duration of the wind, while its breadth represents the force or velocity, as this force is constantly varying, the breadth of the figure must also vary, in order that its area may still represent the integral of the wind correctly. It is the object of Osler's anemometer to describe such a figure. In Whewell's instrument, on the contrary, the integral is represented simply by the depth which the pencil descends. In this instrument no attempt is made to record the time during which each wind blows, the times of its changes, or its force at any given moment, but merely the order of its changes of direction, and the integral or entire quantity that blows from each point, or rather from each rhumb or division of 11 or 221 degrees. This is known by the length of pencil trace described in each vertical division of the cylinder measured vertically, not following the windings of the track. These windings must be neglected as far as they are confined to one rhumb or division, the centre of which corresponds to one of the 16 or 32 points of the compass. All winds therefore not deviating more than 53° from any one of these points if 32 be used, or 1110 if 16 be used, are regarded as blowing exactly from that point. This is a defect common to all anemometers now in use, but by increasing the number of subdivisions, the results will be more accurate.

W

Fig. 45.

E

Having obtained by means of Whewell's instrument the integrals, or lengths of line described by the pencil in each division during a certain period, we may lay down these lengths, or proportionate ones, in their proper order and direction, so as to form a continued crooked line, which expresses all the quantities and changes of the wind, and is called the type of the wind for that particular place and period. Fig. 45 represents the type of all the wind that blew over Plymouth Spi during the month of August and part of September, 1843. Such a line may be regarded as the path that would have been described by a vessel drifting upon a still lake during that period,

Aug

provided it moved with a speed always bearing a constant ratio to that of the wind. If the two ends of this line be joined by a straight line, this will show the direction of the resultant, or average effect of all the winds felt at Plymouth during that period; which in this case is N. 23° E., or about equivalent to a SS.W. wind. This average direction is not the prevailing direction of the wind, or that in which it most commonly blows; for the prevailing winds may be very gentle, and the greater force of those from the opposite quarter may more than compensate for their shorter duration, so that the average direction as regards the integral, or time and intensity taken jointly, may be very different, or even opposite to the average direction if time alone be considered. In this country, however, both these averages have nearly the same direction; the latter, or time-average, being equivalent to a wind blowing from some point between S. and W.; and the former, or true average, though apparently very variable when the resultants of different months or seasons are compared, yet in the type of a whole year, its general direction is found invariably to run northward, and mostly eastward from the starting point. In the present state of the inquiry there is some discrepancy between the results obtained by different instruments; Whewell's placing the mean direction for three years nearer N. than E., while Osler's makes it nearer E. than N. The latter is more likely to be correct, because in Whewell's instrument the velocity of the windmill fly does not bear a constant ratio to that of the wind, but is more than proportionally faster in a quick than in a slow wind, so that the distance which the pencil descends being proportional to the revolutions of the fly, cannot correctly represent integrals of wind; that is to say, the spaces through which it descends during two successive periods do not necessarily bear the same ratio to each other, as do the quantities of wind that have passed over the instrument during these two periods. This objection is surmounted in Osler's instrument, which is driven by a clock, and merely directed or regulated by the wind.

But if Osler's instrument is more correct than Whewell's, it is more difficult to represent the results in the useful form above described. If the instrument be in perfect order, the upper trace made on the paper by the pencil c should be such that its ordinates are proportional to the velocity of the wind; that is, the ordinates at any two different moments should bear the same ratio to each other, as did the velocities of the wind at these two

moments. In this case the total amounts of wind passing over the instrument during different periods, will be proportional to the areas of the portions of curve traced during those periods; that is to say, the spaces contained between the curve, the base line, and the two ordinates at the beginning and end of each period. It is only by measuring

(1) The ordinate of any point of the curve, is its least, or perpendicular distance from the axis, or base line, which in this case is the top of the paper.

and comparing these areas that we can obtain the proportion of the integrals of wind during different periods of time.

To lay down a type of the wind similar to Fig. 45, we must divide the periods in such a manner, that during each period in which the direction of the wind may have been constant, or confined within certain limits, such as two rhumbs, or 2210, or one rhumb, 1110. For this purpose that part of the register paper which registers direction must be divided by 16 or 32 longitudinal lines, such that when the vane points to any one of the 16 or 32 principal points of the horizon, the pencil в may rest midway between two of these lines. We must then note all the points where the pencil track intersects these lines, and from every such intersection raise a perpendicular to the top of the paper; these perpendiculars will evidently divide the upper curve, or that of force, into portions, each of which may be regarded as belonging to one wind only, for during its description the wind did not deviate more than 50, (if we use 32 points, or 1140, if we use only 16,) on either side of a certain point. By ascertaining the areas of these different portions, and drawing lengths of line proportional to those areas, placing those lines in their proper directions, and in their proper order, we may obtain a type of the wind more correctly than by the method before described.

The integrals of the wind have hitherto been referred to only as relative quantities, admitting of comparison only with each other. They have, however, absolute values easily comparable with our common standards or measures. If a pressure plate, acting on a spring, as in Osler's anemometer, be fixed to one extremity of a long beam, or some machine by which it could be moved through the air with any required velocity between 1 mile and 50 miles an hour, it matters not whether the path be straight or circular provided the plate always face the direction in which it moves. If the air be still, the effect on the plate is the same as if it were at rest, and received a wind of a known velocity upon its surface. By this means it can be discovered what velocity of wind is required to produce any amount of compression in the spring, such as may be obtained by placing weights of known value on the pressure plate. In this way a scale of the velocities or force of the wind may be constructed, and absolute values in miles may be assigned to all the lines which compose any type of the wind; and on measuring by the scale thus obtained, the length of the resultant or line joining the two ends of the type, we thus obtain not only the direction, but also the extent in miles of the entire movement of air produced by the combined effect of all the winds that have blown during the period for which the type was constructed. In this way it was ascertained that the resultant of all the winds that blew over Greenwich during the year 1841, was equivalent to the passage of 47,900 miles of air towards E. 28° 30' N. In 1842 the direction of the resultant was E. 27° N., and its length 36,750 miles. By dividing these numbers by

But the average velocity of the winds at Greenwich during the former year was 18.7, and during the latter, 18.3 miles an hour; for the whole integrals of wind for those years, as shown by the length of their type line, measured along all its windings, was in 1841 no less than 167,322 miles, and in 1842, 159,950 miles, showing that the whole movement of the air in this country is about 4 times as great as its resultant or effective movement. The more variable the wind may be at any place, the smaller the proportion of it that will be effective; and if these observations could be made on the open ocean within the range of the constant trade winds, the type would probably be a straight line, and the numbers expressing the total and resultant integrals of wind would be equal.

The direction or length of the resultant for any given period may be obtained more simply than by laying down such a type as Fig. 45, for as the lines, in whatever order they may be placed, will eventually lead to the same point, the figure may be simplified by collecting and summing up all the integrals that belong to the same wind or point of the compass, or that fall within the same angle of 1140 or 2240, but the smaller the angle the better, and then drawing lines proportional to the 16 or 32 sums thus obtained, which lines placed in any order, but in their proper directions, will give the same resultant as if the whole type were drawn. But as some of the lines thus drawn are parallel to others, it is not necessary to draw more than half of them, subtracting from

There are methods of finding the resultant by calculation, for which we must refer to other works, and especially to Sir W. Snow Harris's Report to the British Association in 1844, on the working of Whewell's and Osler's Anemometers at Plymouth in the years 1841, 1842, and 1843. The Editor has also to acknowledge the use he has made in this article of a small work of his, entitled "The Tempest," published in 1848, under the direction of the Committee of General Literature and Education, appointed by the Society for Promoting Christian Knowledge. The reader interested in the subject will do well to consult the Reports of the British Association, and also the Meteorological Reports of the Greenwich Observatory.

un

ANNEALING, (from an Anglo-Saxon word Analan, to heat or to burn.) From the constant tendency of the English to abbreviate words in common use, this term has been contracted by workmen into nealing. Annealing is a process used in the manufacture of GLASS and also of certain metals. In the glass manufacture it consists in placing the articles of glass while hot in an oven called the leer, where they are allowed to cool gradually. They would otherwise be too brittle for use; for the slightest scratch is often sufficient to cause annealed glass to break, and even to fly to powder, while it will often resist a considerable blow. The curious properties of unannealed glass can be exhibited by means of the Bologna phial and Prince Rupert's drops. The former is a rudely shaped phial 3 or 4 inches long, about an inch in diameter, and the glass is about or of an inch thick. If we drop into this phial a pistol bullet from the height of 2 or 3 feet it will not break it, or we may strike it a hard blow on the outside with the handle of a hammer without producing fracture, but if a grain of sand or a sharp bit of flint be allowed to fall into it, it will crack and fall into fragments. This effect is in some cases produced immediately, but in other cases, several minutes elapse before it takes place : and when we think the experiment has failed and are about to drop in another bit of flint, it will suddenly fly to pieces. This remarkable property seems to be destroyed by age, for among some apparatus which has come into the possession of the Editor, are a number of Bologna phials at least 50 years old, some of which have lost this property; but some specimens which will not break by a particle of flint, fly to pieces on scratching the interior surface with a file. Fig. 47 is a copy, one half of the size of the original of one of these phials which have lost this property. It would seem that during the long period in Fig. 47.

each one that which is parallel to it. So that whatever number of points be used, the number of their integrals can always be reduced to one-half, by subtracting each non-cffective wind, or each one that is