subsides, and mercury rises in F until it stands at A, a height of thirty inches above the surface of the mercury in the vessel G.

Many of the vacuum experiments described in the Proceedings of the Society were made in the Torricellian vacuum; but the air-pump was also used. Some of the experiments of this nature relate to the behaviour of birds and animals in rarefied air. Fishes were placed in a receiver, in a vessel of water; on exhausting, they swelled out considerably, turned over, and quickly died. When dissected the air-bladder was found to be empty. An eel lived for some time in vacuo, but was found to be dead at the end of an hour; the air-bladder, as in the case of the fish, was empty.

Pascal found that a bladder partially inflated at the level of the sea became wholly inflated on the summit of a mountain. Robervall made the same experiment with a carp's bladder, which burst when placed in the Torricellian vacuum. This suggested to some of the members of the Academy the idea that the expansion of air, "when at absolute liberty in any place," might be determined by annulling the atmospheric pressure by the weight of a column of mercury below the air, the expansion of which it is desired to determine. Accordingly, two similar tubes, open at one end, and terminated at the other by a bulb, were filled with mercury. One was inverted and opened under mercury in the usual way of performing Torricelli's experiment; into the other a small quantity of air was introduced before the tube was inverted. By calculation it was found that the air had expanded to 173 times its original bulk, but the experiments were not very concordant.

I may mention, in passing, that the experiment proving the incompressibility of water, by causing it to force its way through the pores of a metal, which is universally attributed to the Accademia del Cimento, was performed many years earlier by Bacon,* who used a sphere of lead for the purpose, while the academicians employed a sphere of gold.

In the next paper we shall consider the early labors of the Royal Society and of the Academie des Sciences relative to the subject under discussion.

(To be continued.)

The Economy of Pumping Engines Practically and Commercially Considered.

From the London Artizan, August, 1865.

One of the earliest applications of the dynamic power of steam was in raising water, as is shown in the writings of the oft-quoted Marquis of Worcester, whose method was subsequently improved by Captain Savery; but it was not until the invention of Newcomen's engine that anything like practical success was achieved; and even that, ingenious as it was, considering the age in which it was produced, presented an ungainly aspect, and yielded but little work for the fuel consumed. *See Novum Organum." Book 2. Aph. 45.

Different, indeed, is the present Cornish pumping engine, improved, as it has been, by the powerful intellects brought to bear upon the first crude form, by Stephenson, Watt, Wicksteed, and others, both of the past and present time.

Our object, however, at present is not to occupy space with matter merely historical, but to establish some criteria as to the most suitable forms in which pumping engines may be applied in the different localities and under the diverse circumstances which present themselves to the hydraulic engineer.

If the question resolved itself merely into the relation of the fuel consumed and duty done, it would at once be settled by reference to certain well established experiments, or, more correctly, results of working. Thus, two double-acting engines at Vauxhall, in 1843-44, which were constructed by Messrs. Fenton, Murray & Wood, consumed 11.5 lbs. of coals per horse power, whereas two Cornish engines at the East London Water-works, averaging seven years to 1850, yielded more economical results, consuming only 2.99 lbs. of coal per horse power per hour. In these cases the circumstances were so similar that decision in favor of engines of the Cornish class was arrived at, and on that principle the machinery at the Vauxhall Water-works was subsequently designed. On the other hand, it must be remembered that Cornish engines are costly and large in proportion to their power, their movements being comparatively slow and action single; hence, in cases where the power required is small, and space an object, it may sometimes be found that the advantages of the Cornish engine are outweighed by its disadvantages.

Direct-acting or bull engines are usually stated not to be nearly so economical in their working as the Cornish engines, but the cause of this may not at first sight appear clear. It is, however, by some supposed that it is due to the equalizing influence of the heavy beam upon the motion of the engine; if so, this can, of course, be obtained by attaching a vibrating wheel to the machine, but we are not inclined to attribute what inferiority may exist to this cause, but rather to a less perfect arrangement of the working gear.

The economy of the Cornish engine is, doubtless, due to two peculiarities; first, the high degree of expansion employed; and second, the mode of applying the dynamic force of the steam to propel the water. The advantages of applying steam expansively, so well known now to engineers generally, appear to have been early appreciated by the Cornish men, whether accidentally or by scientific research, as far as we are concerned, it matters not; but these are available to a greater extent at the opening of a mine than afterwards, as an engine of greater power than at first requisite, will usually be erected, and, of course, the steam can then be cut off very early in the stroke; but as the works extend, and more water flows to the pump well, more steam is needed, and, consequently, so high a degree of expansion as was commenced with is not maintained.

The experiments tried on the Cornish engine at the East London Water-works showed the following effects: When there was no ex

pansion, the duty done being taken as 100; it was when the steam expanded through 0.397 of the stroke, 162-6, and when through 0.687 of the stroke, 224. While speaking of trials, it is desirable to refer to a source of error existing in some of the reports of duty in Cornwall, that is, the inefficiency of pump valves. In the case of the Holmbush engine, the water actually raised was 14.7 per cent. less than the calculated quantity adopted in reporting, and this loss appears to represent the quantity of water which ran back through the valves of the pump while they were closing; thus the duty represented by the steam power was 231, 486, 192 foot pounds per 112 pounds of coals, and the useful effect with Welsh coals only 122, 376, 128 foot pounds. This duty is high, but it must be mentioned that when the trial was made, the engine had not long been erected, and was doing but light workthe diameter of the cylinder being 50 inches, and stroke 9 feet 1 inch, while the horse power was but 26-5 horses. Thus the area of piston being 1963-5 square inches, the allowance per horse power would be 74 square inches, and the steam expanded through 0-83 of the entire stroke, or nearly five times its initial bulk, its pressure varying from 24.98 lbs. per square inch, down to 4 lbs. per square inch.

In comparison with the duty above mentioned of the Holmbush engine, we may take some of those reported in Cornwall for September, 1864. The following list comprises those out of thirty-four, which exceeded the average duty of 49,800,000 foot pounds per 112lbs. of coal:

To this list we will append the duties attained at various times by

It should here be noted that the highest duties of the Cornish and Wicksteed engines are calculated for Welsh coal, from its relative evaporative value to that used at the East London Water-works.

In tendering for the erection of certain pumping engines at the

Grand Junction Water-works, in 1844, the contractor was bound to guarantee each engine (having a 40-inch cylinder) during the first twelve months to do a duty of 73.8 millions.

This shows the extent of economy attained at the East London Water-works at that date to be very considerable, as it is highly improbable that any contracter would render himself liable to conditions involving a doubtful issue.

It is usually supposed that small engines are not so economical as large, but there are now some near London doing a duty of about 80,000,000 of foot pounds per 112 lbs. of coal.

We will now proceed to consider the second advantage of the Cornish engine, namely, the mode of applying the dynamic force derived from the steam in the cylinder.

The steam acts in raising a weight and drawing water from the well into the pump barrel, this latter item being small in proportion to the former; then the pressure being equalized on both sides of the piston, the weight which had been lifted falls, forcing the water through the outlet valve from the pump barrel, and drawing up the steam piston ready for another stroke. From Mr. Wicksteed's experiments* on the 80inch Cornish engine, we quote the following particulars, to show the distribution of the steam power, averaging during the stroke those quantities which vary: Preponderating weight, 55,401 lbs. or 11.037 lbs. per square inch of piston; water raised by engine, 4,125 lbs., or 0.821 Îbs. per square inch; cold water pump, 186 lbs., or 0.037 lbs. per square inch; hot water pump, 6 lbs.; air pump, 591 lbs., or 0.117 lbs. per square inch; friction 1,009 lbs., or 0.200 lbs. per square inch; imperfect vacuum, 3,664 lbs., or 0.730 lbs. per square inch; total 64,982 lbs., or 12.94 lbs. per square inch steam pressure on the piston. Now, the effect of the steam in raising the preponderating weight is evidently produced most conveniently, for if, at the commencement of the stroke termed the in-door or steam stroke, the engine runs a little fast, no shock occurs, but the extra momentum is quietly absorbed by a slight increase of speed upward of the weight being raised. Then when the out-door stroke begins, the preponderating weight by its own gravity quietly forces the water out of the pump to wherever it may be required, coming gradually to rest at the termination of the stroke. The superiority of this mode is at once evident when we consider what may be called the riotous movement of the water in a pump worked by an engine with a fly-wheel, the direction of the water's motion being in this case reversed without allowing the current time to come to rest, and producing, as it were, a series of blows, destructive alike to the machinery and its economy, and to any observer the effect of hydraulic shocks may be made evident by placing the hand upon a main leading from a pumping engine, as thereon the beat of the pump valves may be felt even a mile from the engine. This transmission of blows is, of course, to be traced to the comparative inelastic quality of water. Then, again, the gradual failing pressure of expanding steam is not favorable,

* See Wicksteed's "Experimental Enquiry."

if it be directly applied to the propulsion of water, which, being liquid, cannot so well carry the varying effect without loss of power.

With regard to rotary pumps, all we shall observe is this: A perfect rotary pump, worked by an uniform moment of pressure, would probably be an improvement on the Cornish pumping engine, if its motive power could be produced as cheaply; otherwise, except for purposes of trifling importance, we do not feel disposed to place much reliance upon them. One of the most important details of the pumping machinery rests in the valves, upon the construction of which the smooth working of the engine mainly depends. When the old clack valves were used, the vibration due to their closing was something enormous, in some cases shaking the buildings to such an extent that the engine could only be worked for a few hours at a time; but when Harvy and West's double beat valves were introduced, this difficulty and that arising from loss of water while the valves were closing were at once obviated; subsequently, valves closed by numerous balls, or small india rubber flaps were introduced, and also valves consisting of cylinders perforated on their peripheries, and surrounded by india rubber straps were applied, as also a variety of other contrivances, most of them ingenious and some useful. The double beat valves may occasionally be fouled, as once happened at the Ajax engine at the East London Water-works, when an eel came up the wind bore of the pump; but such cases are exceedingly rare. Recently, surface condensers have been applied to Cornish engines, both at the Scarborough and Kent Water-works, and certainly this description of steam machinery appears to afford them great facilities for working satisfactorily, as all the water passing from the main pump may, if it be desired, be allowed to flow through the condenser, thus insuring a good vacuum.

In entering upon the next branch of our subject-the question of finance-great caution is needed, as the means of obtaining correct information are very scarce, and, when found, give data rather perplexing to generalize, especially for comparison with those relating to double-acting engines. The figures, which will be quoted, are taken from actual practice, not mere estimates. In the first place, it will be necessary to come to some conclusion as to criteria of horse power of Cornish engines, as usually worked. For this purpose, various particulars have been gathered which tend to show that, taking a wide range of practice, there is, on the average, an allowance of about thirty square inches of piston surface per horse power; and this does not appear unreasonable, as it corresponds to fifteen square inches per horse power, in a double acting engine, the mean velocity of the piston on both strokes, and including stoppages at the end of each stroke, may be taken, for an engine working regularly, as being about one hundred feet per minute, but in different engines the speed varies very considerably. Taking the allowance of thirty square inches per horse power as granted, the mean cost of bright Cornish engines will be found to amount to about fifty-four pounds per horse power, exclusive of boilers. This appears very heavy, but it applies to engines of average dimensions, and includes the pump work and duplicates of the valves; and of this amount the main pump work costs about twenty-five per