statics, and the pressure increases with the depth, and a very smal vertical column of iron will produce a great pressure throughout the liquid mass, a pressure proportional to its height. If, then, the column is kept in a state of fusion while the metal in the flask is cooling and setting, the particles will be packed more closely together, the density will be increased, and a stronger bar obtained. In an economical view this is very important, for a very inferior iron will by this means give as strong a bar, or stronger, than a much better, and therefore dearer, iron.

Experiments have been made only on a small scale as yet, but we purpose to extend them in number so as to generalize on the subject. and ascertain whether, by a proper depth of runner head, we cannot make bars of a mixture of Gartsberrie II. and old fire-bars of the worst scrap iron which shall be as strong as the best irons will give. I had hoped that I should have had time to write a short account of the researches we have made on the strength of timber, but I have found it impossible for the present,-I will only draw your attention to this point. In ascertaining the strength of timber, these two points must be kept quite distinct, viz: the constant expressing the relative strength, and the constant to be used in the formulæ, such as those given in Turnbull, which we may call the constant representing the absolute strength.

If a hundred experiments are made on a particular kind of fir, the breaking weight in each case being taken under exactly similar circumstances, the 100th part of the sum of all these weights represents fairly enough the mean strength of that particular wood, i. e., it will do to compare with a number determined in the same way for another kind of fir, and the greater this mean weight is, the stronger the wood. So far, then, as scientific research is concerned, this method is fair,but the same constant will not do for the practical man. I say that for this we must take the least weight that broke the beam, and not the mean, for the workman wants to know not what the average strength of the wood is, but how much he may trust to a particular specimen of it; and therefore he wants to know what is the least weight that broke a fair, uniform, sound beam under given circum

stances.

The first set of constants will guide him in the choice of his material, but if he begins to calculate the dimensions from the formula, as in Turnbull, he wants a different constant, viz: the one I have above pointed out.

My attention was drawn to this subject by reading the following passage in Professor Barlow's book on Strength of Materials, page 26. "Practical Rule.-Since the strength of direct cohesion must necessarily be proportional to the number of fibres, or to the area of the section, it follows, that the strength of any iron rod will be found by. multiplying the number of square inches in its section by the corresponding tabular number as given above.

"This, however, gives the absolute strength, or rather the weight that would destroy the bar; and practical men assert that not more than one-fourth of this ought to be employed. I have, however, left

more than three-fourths of the whole weight hanging for twenty-four or forty-eight hours without perceiving the least change in the state of the fibres, or any diminution of their ultimate strength."

It will be seen that Professor Barlow uses the expression "absolute strength" in the way in which I have proposed to use "relative strength."

I say that if he proposed the smallest weight instead of the mean, the result would be one which practical men ought to trust. As to the arbitrary "one fourth," it only shows that we have not yet succeeded in proving to practical men that we can do them any good, and Mr. Barlow justly intimates that there is a great sacrifice of the powers of the material in this very arbitrary reduction.

I trust to offer some further remarks on this interesting subject,— as well as to be able to lay before your readers at a future time results of experimental enquiries made by Mr. Ranger, our lecturer and professor of general construction. I hope we shall be able to point out some very important features in the case, to which as yet but little attention seems to have been given.

College, Putney, July, 1845.

M. CowIE, Principal.
Civ. Eng. & Arch. Jour.

Comparison of English, French, and Belgian Railways. The Fares charged, the Expenses and the Profits.

While so many railways are in the course of construction, it is interesting to know as exactly as possible every item of expenditure connected with them. An able engineer, M. Julien, under whose direction are the works on the line from Paris to Lyons, has undertaken to give an idea of these expenses. The reports of the railways of France and abroad supply him with materials which he has admirably digested and arranged.

Average Cost of Railway per mile.-The first thing that strikes us in M. Julien's Memoire is the observation that the average of 300,000 fr. per kilometre (£19,200 per mile,) generally admitted as the probable cost of construction of the main line in France, is certainly too small. Of this M. Julien has given proofs. This is an important consideration at a time like the present, when speculation is so eagerly directed towards railways. On many of the English lines the expense has been from 700,000 to 800,000 fr. (£44,800 to £50,200 per mile. In France we cannot expect the cost to be less than 350,000 to 400,000 fr. on the average (£22,400 to £25,600 per mile.)

Comparison of Fares charged in England, Belgium, and France. -On the first view, one would imagine the difference between one railway and another to be so great that it would be in vain to attempt to reconcile them and deduce general laws. The fares, the number of passengers, their distribution among the classes of carriages, would seem liable to infiníte variation according to the frequency of trains, the gradients, the curves, and every thing else which has an influence on the expense and profits. M. Julien has, however, in the general

results, collected some very singular coincidences. What can be more dissimilar than the rate of fares in England, France, and Belgium? The fare per mile actually received appears on the average

Then, again, the manner in which the passengers are distributed into the three classes differs very considerably. Of 100 passengers the average proportion in each kind of carriages will be

In England 18 passengers in the first class, 46 in the second, 36 in the third. In France

15

66

66

30

27

46 66

66

55
63 66

The total return per mile is in England £2,236, in Belgium £1,160. It is nevertheless a curious fact that the traffic is apparently the same in Belgium as in England, and it is the traffic which affords the criterion of public utility. The Birmingham railway now produces £7,208 per mile; the Orleans line returns are stated at less than £3,200 per mile; M. Bartholomy in a recent work has stated the amount at £2,816; but the Orleans railway will have numerous extensions, and will probably produce one half the return of the Birmingham, with a rate of fare half less, and on the supposition of an equal traffic.

M. Julien has also pointed out another resemblance. The English lines, on the average, derive two-thirds of their revenues from passengers and one-third from luggage, and the proportion is found to be as nearly as possible the same in Belgium.

Comparison of the Actual Cost of Conveyance in the Three Countries. What is most singular is that, on the railways of all countries, the total amount of general expenses corresponding to the running of a train over one kilometre of ground is as nearly as possible 3 fr. (about 3s. 10d. per mile,) and this expense is divided into two equal parts, of which one represents the force of locomotion and the wear of material, the other the expense of management by clerks and superintendents and the general working of the railway.

It is singular to observe in the above table that the fare of a third class passenger in England is that of a second class passenger in France and of a first class passenger in Belgium. In the above calculation we have reckoned the centime = 1-10 of a penny, but its real value is at present rather less.-Ed.

+ We were much struck with the low charges of railway traveling in Belgium, as compared with those of England. From Antwerp to Brussels, for example, a distance of 24 miles, the fares are,-diligences, 3 francs 50 cents; chars-a-bancs, 2 francs 25 cents; and wagons, 1 franc 75 cents; that is 2s. 9 d., 18. 74d., and 1s. 44d.; whereas the same distance upon any of the great lines in England would be above 6s., 4s, and 2s. Upon an average we found it to be only about one-half the expense of similar traveling at home. The consequence is that far more people travel by railway in Belgium than in England. We were astonished at the crowds that went with us in every direction, but more particularly between Brussels and Antwerp, and towards Ghent and Bruges. The increase in the receipts of the different lines has been great, as the advantages have been developed. The lines now in operation were completed in October, 1843; since,which time their receipts have increased in an extraordinary proportion. During the past year (1844) the traffic of heavy merchandize was nearly double the amount of the preceding year, amounting to about 500,000 tons. The increase of passengers was still more extraordinary, that traffic alone having realized something not far short, it is said, of 10,000,000 francs.-Ramsay's Belgium and the Rhine.

To express by a clear formula the expense of working a railway, M. Julien has drawn a relation between the goods carried and the passengers who form the principal item of the receipts. The basis of this relation is perfectly plausible. He has accordingly expressed the whole amount of traffic, both for goods and passengers, by a certain number of passengers. His unit of calculation is one passenger carried one mile, and all he has to find is the cost of this unit including every expense. He has found that it is

4.8 66

.738

66

On the English lines, exactly, The last amount however includes the English Government tax, and is accounted for by the smaller number of passengers in each train, for the English companies, correctly or erroneously, prefer a few passengers paying well to a great number paying little: they also have very frequent trains, persuaded justly that what most concerns. them is not to spend comparatively little but to obtain a large gross return. If the passengers of an average train were the same in number in England as in France, and if the government tax were equal also, the cost of each passenger per mile would be the same as with us. Merchandize.-According to M. Julien's comparison a ton of merchandize carried by slow trains costs rather more than 5 centimes in France and Belgium, (.8 of a penny per mile,) and in England 7 centimes (1.12 d. per mile.)

Proportion of Expenses to Profits in the three countries.-With us the Government often, and the committees of the chamber of Deputies always, have admitted as an absolute rule, that on a railway the expense represents 45 per cent. of the returns whatever their amount, from which it is concluded that the net profit is 55 per cent. of the receipts. This opinion has had the unfortunate effect of greatly exaggerating the net profit of poor railways. M. Julien has not had much trouble in proving the inaccuracy of this opinion, all accredited as it is. The more railways are frequented, or, the traffic remaining the same, the more the fares are raised, the greater proportion will the profit bear to the total returns. For there are a large number of constant expenses which have to be incurred in every case. On the English railways, which in 1842 gave altogether a receipt of 111 millions of francs, or (£4,625,000,) and an expenditure 48 millions of francs, or (£2,000,000,) the expense was not more than 43 per cent. of the total receipts; consequently the net profit was 57 per cent.; and it would have been as much as 62 per cent. if the English Government tax were as moderate as our own. On the Belgian lines

however the expenses are 60 per cent. instead of 43, and on the Strasburgh and Basle 73 per cent.

In the expenses as stated above, the interest of capital is not included; all that is reckoned is the actual cost of working the lines: otherwise we should have to add in France to the expense a sum of 17,500 fr., or 20,000 fr. per kilometre (£1,120 to £1,280 per mile,) representing an interest of 5 per cent. on the cost of constructing a railway with a double line of rails. And since we can scarcely expect in France an average total return much larger than that of the Belgian lines, which is about 20,000 fr. per kilometre, this item alone swallows up the total receipt.

Reserve Fund for renewal of rails.—In the accounts of railways an item ought to be presented which has not hitherto been introduced into them, because the railways are at present new, but which will soon have to find a place: and that is an annual reserve to collect the capital necessary for a future renewal of rails. How long will the rails last? Twenty years, perhaps, and the crossings and sidings ten or twelve. A double line costs about 80,060 fr. per kilometre (£5,120 per mile.) To form this capital M. Julien proposes to reserve annually 4,000 fr. per kilometre of the line (£256 per mile.) M. Julien is right, and the commissioners have made a great mistake in omitting this consideration in their calculations. The capitalists, if they be wise, and seek, not adventurous speculations, but sound investments, will not commit the same error. On the whole the perusal of M. Julien's Memoir is of the nature to abate any very exalted notions on the subject of railways, and will give valuable information to those about to embark their fortune in those enterprizes.

Jour. Des Debats.-Ibid.

Prevention of Incrustation in Boilers-Dr. Ritterbandt's process— Mr. Johnston's Patent Boilers.

Sir,-From your remarking on Dr. Ritterbandt's patent, "this has every appearance of being an eminently useful invention, we have never before seen the great evil which it is intended to obviate, grappled within so scientific and effectual a manner," you are evidently predisposed in favour of his invention; yet, from the impartiality you have displayed towards all controversial writings which I have perused in your Magazine for a series of years, I am satisfied that you will find room for the insertion of the following remarks on boiler incrustations.

Scientific men have long labored under the delusion that the deposits in boilers are insoluble, owing to their being chiefly composed of sulphate of lime.

I am much gratified to find that Dr. Ritterbandt and others are now convinced that it is "the heat employed to generate steam causing the lime which exists in the water, in the form of soluble bicarbonate of lime, to be converted into an insoluble carbonate of lime; and (that) in marine boilers, incrustation is generally promoted by the carbonate of lime set free by the heat, which, as it floats in the water