or electric changes in the molecular structure of the iron, caused by friction in the bearings and great velocities; and in his opinion it was probable that the continual strains and percussions to which the crank axle is subjected would account for the changes in the molecular constitution of the iron.

Mr. Hodgkinson was certain, from the results of his experiments, that a succession of strains, however slight, would produce a permanent deterioration of the elasticity of the iron. Mr. Fairbairn had been told by the engineer on the Leeds line, that he considered all crank axles to be constantly deteriorating from percussions, strains, &c., and that they should be removed and replaced by new ones periodically, to avoid danger of fracture.-A discussion arose as to whether the crystallized appearance observed in fractured axles arose from defects in the manufacture, in the quality of the iron, or from the effects of working, either by percussions, strains, or magnetic action.-Mr. Grantham, although a manufacturer of cranked axles, admitted that straight axles were less liable to break. Cranked axles, from the way in which they were welded together and shaped, were rendered weak and liable to fracture. On other grounds, however, he believed that the cranked axles were preferable, as they produced a steadier motion, and much heat was saved.-Mr. Garnett believed that more straight axles had broken than cranked ones. -Professor Willis showed the effect of vibration in destroying molecular arrangement, by reference to the tongues in musical boxes, &c.

Mr. Nasmyth believed that the defects in axles, &c. arose in the manufacture, especially from cold swaging and hammering, and also from over-heating in welding, all of which causes injured the toughness of the iron. In small articles, he found great advantage from annealing; and he believed that axles might be annealed very cheaply, and would be more serviceable. He disliked the fashion of referring all unaccounted phenomena to magnetism and electricity, although he was convinced that very singular electric phenomena accompanied the transit of locomotives and the rapid generation of steam. With this was connected the nonoxydization of rails, where the traffic was in one direction, and the rapid oxydization when the same rails were travelled over in both directions, as in the Blackwall railway. He had also observed that brasses, in some cases, had from friction entered into cold fusion,that is, at a heat not perceptible to the eye, a complete disintegration of the molecular structure had taken place, and he had seen the brass spread as if it had been butter or pitch. He had no doubt that this arose from electricity, but had not ascertained the fact

from experiment.-Mr. Fairbairn stated, that in hand-hammered rivets the heads fre

quently dropped off, and presented a crystallized appearance, while those compressed by machine were sound. He found that repeated percussions, from the rivetting, hammering plates, &c., induced magnetism in iron boats.-Mr. Vignoles could not, from his experience, subscribe to Mr. Nasmyth's theory of the oxydization of rails by single traffic, as the railway from Newton to Wigan had been single for a long time, and was as bright as the Manchester and Liverpool. The Blackwall railway was not an analogous case, as no locomotives were employed.— Mr. Roberts disbelieved the deterioration of axles by work; he would rather trust an old axle than a new one. He believed cold swaging and hammering to be the chief causes of mischief. In fact, if axles were sent out sound and well manufactured, they would rather improve by working.-Athe

næum.

INSTITUTION OF CIVIL ENGineers.

MARCH 22, 1842. "Results of a Trial of the Constant Indicator upon the Cornish Engine at the East London Water-works." By Professor Moseley, F.R.S., &c.

The object of this communication is to exhibit and explain the results given by the author's indicator during a continuous registration from the 28th January to the 25th February 1842, the engine during that time making 232,617 strokes. The numbers registered by the counter of the engine and the indicator were noted each morning and evening, and are recorded in a table appended to the paper. The differences between each two consecutive numbers registered by the counter, giving the number of strokes made between each two observations, are contained in one column of the table, and in another column are the differences between the successive registrations of the indicator. These are followed by the mean registrations of the indicator at every stroke of the engine, being the quotients of the numbers in the lastnamed column divided by the corresponding numbers of the preceding column. The paper after thus stating the numbers registered daily by the indicator during the period of trial, proceeds to explain the formula to which they are to be applied, in order to determine the work done daily by the engine. The formula, when reduced from the general one by the introduction of the numerical values of the constants dependant upon this engine, is

of work (in lbs. raised one foot high), done upon each square inch of the piston through any given time, during which the number registered by the indicator is represented by N, and the space in feet which the piston traverses by L. The second term of the formula, which is very small as compared with the first, is a correction for the influence of the friction of the indicator on the number registered by it. The formula being then reduced by the substitution in it of the numerical values before alluded to, the whole number of units of work per square inch of the piston done between the 28th January and the 25th February is shown to have been 21 464 067-1727. From this is deduced the work done during the same time upon the whole area of the piston as well as the duty done upon the piston for each cwt. of coals. These calculations are followed by a comparison of the results given by the indicator with those previously obtained from actual experiment by Mr. Wicksteed; whence it appears, that with a necessary allowance for a difference in the lengths of stroke at the periods of the two experiments, the results of the two are almost coincident. The work per stroke upon every square inch of the piston, as obtained by experiment, is 120-574, whilst as shown by the indicator it is 119.338 lbs.

Professor Moseley exhibited the indicator and described its construction and action; it consists of two cylinders, each four inches long, communicating by pipes with the top and bottom of the cylinder of the steam-engine to which the instrument is applied. In each of these cylinders there works a solid piston four square inches in area; both being fixed upon the extremities of the same rod, which (when the indicator is in action) sustains in the direction of its length a pressure equal to the difference between the pressures upon the two indicator pistons, or equal to the effective pressure of the steam on four square inches of the piston of the engine. This pressure is made to bear upon a steel spring, connected by a link at each end with a similar spring, supported at its centre upon a projection from the frame of the instrument. The pressure of the piston-rod upon the lower spring, causes the two springs to separate from each other, and the separation produced is, by a well-known law of deflection, directly proportional to the pressure sustained, so long as the deflections are small. A peculiar form, first suggested (it is believed) by M. Morin, is given to these springs; one surface of the spring is plane, and the opposite surface is of a parabolic form, by which equal strength is given throughout every portion of its length. In

a spring thus formed, the deflection is distributed more equally throughout, and being thus diminished for a given separation of the springs at every point, the elastic limits are nowhere so soon exceeded.

By this connexion of the piston-rod with the springs, its position is made to vary directly as the effective pressure upon four square inches of the area of the piston of the steam-engine, so that every additional pound in that pressure will cause the piston-rod to alter its position by the same additional distance in the direction of its length.

A steel wheel (termed the integrating wheel) having the edge milled, turns upon the piston-rod as its axis, traversing with it also in the direction of its length. Through the arms of this wheel pass three rods, connected at their extremities by two pieces so as to form with them a rigid frame, which turns, in fixed bearings, upon hollow axes through which the piston-rod passes, so that the integrating wheel is free to traverse longitudinally upon the frame, but cannot revolve without carrying the frame with it. The integrating wheel is made to revolve by the rotation of a cone which is held in contact with it by a spiral spring, acting constantly against the extremity of the axis of the cone. A system of bevil-wheels communicates to this cone the rotation of a pulley, which is driven by a cord carrying a weight at one extremity and communicating by the other with the piston-rod of the engine, or with some point whose motion accords with it, but travelling through a less space. The circumference of the pulley moving precisely as the piston, the angle described by the cone in any period of time, must be exactly in proportion to the space described by the piston in that time. The circumference of the integrating wheel moving with that part of the cone, with which it is in contact, the portion of a revolution which it is made to describe in a given time, is dependent, first, upon the angle which the cone describes about its axis, during that time; and secondly, upon the distance of its point of contact from the apex of the cone at that time. If either of these two elements of variation remained always the same, then the portion of a revolution, made by the wheel, would vary directly as the other, whence it follows, by a well-known principle of variation, that when both these elements vary, it varies as their product; or that the portion of a revolution, made by the integrating wheel in a given time, varies directly as the product of two factors, one of which is the angle described during that time by the cone, and the other the distance of the point of contact of the wheel and cone, from the apex of the cone. The former of these factors

varies directly as the space described by the piston of the engine, and the latter as the effective pressure then exerted by the steam upon the piston: therefore the portion of a revolution made by the integrating wheel, varies as the product of the space described by the piston of the engine during a given time, by the effective pressure of the steam upon it during that time; that is, it varies as the work or dynamic effect of the steam upon the piston during that time; whence it follows, that the number of revolutions or parts of a revolution made by the integrating wheel, during the stroke is proportional to the whole work, or dynamical effect of the steam upon the piston during the stroke.

By a number of toothed wheels the number of revolutions of the integrating wheel is registered to five places of integers, and to one place of decimals. The number registered is not diminished by the backward motion of the cone during each return stroke, because the integrating wheel ascends to the apex of the cone, and remains there during each return stroke, so that no number is registered during that interval.

In order effectually to guard, however, against any error which might arise from this reversed motion of the piston, a combination of wheels has been introduced, by which the revolution of the cone can be arrested during the return-stroke; and to adapt the instrument to register (if required) every stroke, a fourway cock has been constructed, by which one of the indicator cylinders may be made to communicate always with the steam end of the steam-engine cylinder, and the other to be acted upon by the vacuum end: in this case the movement of the cone should be constantly forwards.

The Professor then gave the mathematical formula by which the work is determined from the numbers registered by the indicator. He then described the difference between the instrument, and that of M. Morin for applying the principle of M. Poncelet, to consist,

First, in all those mechanical combinations which are peculiar to the instrument in its application to the steam-engine. M. Morin's instrument having been applied to measure the traction of horses.

Secondly; in the surface of a cone being substituted for the plane surface of a circular disc; by which arrangement the rapidity of the changes of velocity due to corresponding changes in the position of the integrating wheel is diminished in the same proportion in which the sine of one-half the angle of the cone is less than unity; and the force necessary to drive the integrating wheel being diminished in the same proportion, the chance of an error arising from the slipping

of the edge of the integrating wheel on the surface from which it receives the impulse is lessened in proportion.

Thirdly; in the separation of the registering apparatus from the integrating wheel; by which separation, whilst the springs are relieved from the effect of the momentum and the friction due to the weight of the registering apparatus, the latter being in a state of quiescence, the registration is legible whilst the indicator is in action.

Fourthly; in the variable appearance of the links connecting the springs together, by which variation the same series of deflexions may be obtained under different ranges of pressure.

In fact, that the indicator has nothing in common with the "compteur" of M. Morin, except the principle of Mr. Poncelet, and the springs under a modified form.

The

The amount of the friction of the pistons, was then examined, and the peculiar construction of their metallic packing explained: it was shown also, that instead of great difficulties arising from the friction of the integrating wheel upon the cone, or its slipping upon the surface, a very slight pressure of the spring produced sufficient adhesion to drive the registering apparatus. Professor then explained the advantages resulting from a registration of the duty of steam-engines generally, not during the time of a few isolated experiments, as with the common indicator, but extended over any given period, and through every stroke of the engine, displaying all the changes which had occurred during that time :-with this view it had been decided that the instrument should be attached to the engines of the Great Western steam vessel on her next voyage to America.

He then expressed his obligation to Mr. Wicksteed for the facilities afforded him for the experiments at Old Ford, and paid a well-merited compliment to Mr. Holtzapffel for the excellent construction of the indicator.

In reply to a question from Mr. Vignoles, he stated that the instrument was not under its present form adapted to locomotive engines, but that a grant of 1007. had been made by the British Association for the construction of such an instrument.

Mr. Cowper, in compliance with the request of Professor Moseley, illustrated his description by setting the instrument in motion, showing that the registration depended upon the revolutions of the integrating wheel; he demonstrated the cases of motion without pressure, and pressure without motion; in the former case, the integrating wheel being stationary at the apex of the cone while revolving, does not receive

any impulse from the contact with it, and therefore does not register; in the latter case, the surface of the cone upon which the integrating wheel traverses, being at rest, does not communicate any rotative motion to it, and consequently no registration can take place; but when motion and pressure are combined, the cone revolving, and the integrating wheel travelling from the apex some distance towards its base, the exact product of the motion of the cone and the steam's pressure upon the piston would be registered by the amount of the revolution of the integrating wheel.

Mr. Wicksteed observed, that every facility had been afforded to Professor Moseley for applying his new indicator, for the purpose of ascertaining the duty performed by the Cornish engine at Old Ford, but that he had not at all interfered with the experiments, being desirous of ascertaining whether the results would correspond with his trials. That after the work of the engine had been registered while it was making about 179,000 strokes, the mean result, as stated by Professor Moseley was so nearly that arrived at by Mr. Wicksteed, that he had no doubt of the accuracy of the machine as a good indicator of the real duty performed by the engine; the difference in the result of the mean pressure of the steam, deducting the vacuum, or 0.73lb., was 0.12lb., namely, according to Mr. Wicksteed's experiments 12.94-0·73-12-21lbs, and according to Professor Moseley 12.09lbs. ; this difference might arise from a variation in the mean length of stroke during the two sets of experiments-from a slight variation in the point at which the steam had been cut offfrom a variation in the level of the water in the pump well, or other practical causes-the difference, however was so insignificant, that he would rely on the accuracy of Professor Moseley's indicator, and allow the possibility of a slight error in his own experiments.

Mr. Farey observed, that Professor Moseley's instrument must be influenced by variations in the length of stroke, for whenever the piston makes a long stroke, the cone and the train of registering wheels must be turned farther round, and would register a higher number than they would do in case of a shorter stroke, supposing the impelling force exerted by the steam to be always the same. If the instrument could be really made to give its results according to the actual length of all the varying strokes made during the time of observation, by truly aggregating these varying lengths into one sum, the results would be free from the usual uncertainty respecting an average length of stroke.

In the monthly reports of engines in

Cornwall the performance is reckoned according to some reputed length of stroke, which had been fixed upon for each engine, when it was first reported, and it is afterwards assumed that no departure from that reputed length has taken place, when in fact such departure does often occur.

It would be very desirable to have a moving card applied to the new instrument, in order to indicate the impelling force of the steam in the cylinder, by tracing curves on paper like those by the ordinary indicators. This, it appeared, might be done with the advantage of causing the paper on which the curve is drawn to travel onwards, and bring fresh paper into its place, so as to obtain a series of distinct curves for as many succeeding strokes.

The form of the springs of Professor Moseley's instrument would be a decided improvement if substituted for the spiral spring of ordinary indicators; Mr. Farey had applied to an ordinary indicator, a mode of exhibiting at a glance, whether the engine was exerting more or less force than its ordinary appointed task; the plan answered that purpose; but as it required the indicator to be always in action the spring of the indicator broke after working more than two days, he therefore abandoned it. The springs in the new instrument were proved by the trial at Old Ford to be capable of enduring continual exertion without breaking.

The Professor had stated that the scale of flexure of the new springs was found to be exactly, according to theory, equal divisions with equal forces; this might be expected, because the flexure of the springs was small, and the bending force acted in a direction nearly at right angles to the length of the springs. In ordinary indicators the scale should not always be equal divisions, because the wire of the spring being wound spirally into a screw of small diameter, the spiral obliquity of the thread of such screw becomes more oblique to the direction of the bending force, as the spring is stretched, and less oblique as the spring is compressed, and hence the scale of pounds per square inch, by which the curve should be measured for summing up the results, ought to be a scale of unequal divisions.

The indicators originally used by Boulton and Watt were of a large size, with a long and powerful spring curled into a cylindric form, as large in diameter as could be included in the cylinder of the indicator, and the motion allowed to the piston by the spring was very short; such indicators were judiciously proportioned, and they do not show any sensible inequality of divisions in their scale. But recently, indicators have been frequently made without the knowledge

of their true principle, and the rules of proportion are not observed, so that it will sometimes be found, on actual trial of such instruments with weights, that their scale of pounds per square inch is not in equal divisions, although it is usual to employ a scale of equal divisions for summing up the curves traced by them.

In Boulton and Watt's indicators the scale of pounds per square inch was formed from actual trial with weights; but such trials were made when the indicator was cold, and dismounted from its place upon the steam-engine.

A much better mode is to apply the weights on the upper end of the piston-rod when the indicator is placed on the cylinder of the engine, while it is hot, its piston being supplied with the same quantity of oil, and the spring being in the same state as when it is in use. The depression of the piston by the weights is recorded by drawing a line with the pencil of the instrument on the card itself in the same manner as the usual atmospheric line is drawn thereon.

A series of lines thus drawn with given weights, become so many original stages for subdividing between them, to form a true scale for summing up the curve described under the same circumstances and nearly at the same time.

Professor Moseley's instrument had two cylinders and pistons operating in concert on the same piston-rod, and springs of peculiar construction to indicate the unbalanced pressure exerted by the steam to impel the piston of the engine. The elastic force wherewith the steam acts above the piston (called the positive pressure or plenum) is shown by a common indicator, but the elastic force wherewith the uncondensed steam is at the same time reacting beneath the piston (called the negative pressure, or imperfect exhaustion or vacuum) is not shown; hence the observations are limited to two odd halves of the stroke made by the piston; those halves being commonly the plenum during the descent, and the exhaustion during the ascent of the piston; it is taken for granted that the other two odd halves are the same as those which are observed, although such assumed parity is not always the true state of the case.

In the new instrument the indication that it would make by drawing on a card, would be that of the difference subsisting between the plenum above, and the exhaustion beneath, the piston of the engine during its descent and ascent, wherefore it would indicate on one card as much as two ordinary indicators can do on two cards, if they are applied one to the top and the other to the bottom of the cylinder of the steam-engine;

in that case each indicator shows on its own card what the elastic force of the steam is during the plenum, and what it is during the exhaustion, but the required result (which is the difference between the two) must be obtained by combining together in the computation those distinct curved lines which are drawn on two separate cards. Professor Moseley's combined indicator pistons, acting on the same springs, would at once indicate such difference, by the curve which it would trace on the one card.

In answer to a question from Mr. Parkes as to whether the new instrument had been put to any other test than its apparent agreement with Mr. Wicksteed's estimate of the resistance overcome, and whether the common indicator had been applied to the engine at the same time, Professor Moseley said, that he had not compared the instrument with any other, but had subjected Mr. Wicksteed's calculations to a rigid investigation, and felt quite satisfied that they approximated closely to the truth. He relied upon them as corroborations of the accuracy of the instrument.

Mr. Parkes observed that it would have been more satisfactory to engineers to have been assured that every means had been taken to demonstrate the truth of the results recorded by an instrument which had such important functions in view. He wished to know in what manner the pressures denoted had been ascertained,-whether by weights or by comparing them with a mercurial column. He had found the latter mode more exact than weights, in verifying the scale of the common indicator, as the instrument being heated was then in precisely the same state as when it was in use. He had found that a certain amount of correction was frequently necessary, as both the spring and the amount of piston friction were affected by heat.

Professor Moseley replied that the instrument had not been compared with the mercurial column, but that the resistance of the springs, and the friction of the piston and instrument generally, had been ascertained by very accurate experiments, so that he had full confidence in the results.

Mr. Parkes said that notwithstanding the respect and deference he felt for Professor Moseley's attainments and ingenuity, his past experience would not permit him to place entire confidence in the results afforded by the instrument: indeed he considered them to be altogether fallacious as representing the force acting on the piston of the Old Ford engine. He could not admit that the apparent near identity between Mr. Wicksteed's computations of resistance, and the constant indicator's registration of forcr