three-eighths thick. The only difference between the two is the distance of the stays, the first being in squares of 25 inches area, and the other in squares of 16 inches area.

From 995 to 1295 lbs., the increase of the swelling or bulge on the side was inappreciable.

Failed by one of the stays drawing through the iron plate after sustaining the pressure upwards of 13 minute.

From the above experiments, it will be observed that the weakest part of the box was not in the copper, but in the iron plates, which gave way by stripping or tearing asunder the threads or screws in part of the iron plate at the end of the stay marked a, fig. 13.

The mathematical theory would lead us to expect that the strength of the plates would be inversely as the surfaces between the stays; but a comparison of the results of these experiments shows that the strength decreases in a higher ratio than the increase of space between the stays. Thus, according to the mathematical theory, we should have

Ult. strength 2nd plate per sq. in.

=strength 1st plate × 25

=815 × 25

=1273 lbs.

Now this plate sustained 1625 lbs. per square inch, showing an excess of about one-fourth above that indicated by

the law.

This is an excess of the force required to strip the screw of a stay 11ths of an inch in diameter, such as those which formed the support of the flat surfaces in the exploded boiler.

It will be found that a close analogy exists throughout the whole experiments, as respects the strengths of the stays when screwed into the plates, whether of copper or iron; and that the riveting of the ends of the stays adds to their retaining powers an increased strength of nearly 14 per cent. to that which the simple screw affords.

The difference between a fire-box stay when simply screwed into the plate and when riveted at the ends is therefore in the ratio of 100: 114, nearly the same as shown by experiments in the Appendix.

It is desirable that we should determine, by mathematical investigation, the strain exerted on each stay or bolt of the fire-box.

Let A, B, C, D, E, F represent the ends of the bolts or stays; 01, 02, 03, 04, the centres of the squares formed by the bolts. Suppose a pres

sure to be applied at each of the points 01, 02, 03, 04, equal to the whole pressure on each of the squares, then the central bolt A will sustain fourth of the pressure applied at O1, also one-fourth of the pressure applied at O2, and so on; so that the

one

HO

003

FIG. 14.

902

до

Ов

C

OB

E

whole pressure on A will be equal to the pressure applied to one of the square surfaces. Hence we have

Strain on the stay of Table I. =

Strain on the stay of Table II. =

815 × 25

= =9 tons.

2240 1625 × 16

2240

= 11 tons.

The stay in the latter case was 13ths of an inch in diameter; hence the strain upon one square section would be about 13 tons, which is considerably within the limits of rupture of wrought iron under a tensile force.*

In the experiments here referred to, it must be borne in mind that they were made on plates and stays at a temperature not exceeding 50° of Fahrenheit; and the question naturally occurs, as to what would be the difference of strength under the influence of a greatly increased

*For the remainder of the experiments see Appendix No. II.

temperature in the water surrounding the fire-box, and that of the incandescent fuel acting upon the opposite surface of the plates.

This is a question not easily answered, as we have no experimental facts sufficiently accurate to refer to; and the difference of temperature of the furnace on one side, as compared with that of the water on the other, increases the difficulty, and renders any investigation exceedingly unsatisfactory. Judging, however, from practical experience and observation, I am inclined to think that the strengths of the metals are not much deteriorated. My experiments on the effects of temperature on cast iron* do not indicate much loss of strength up to a temperature of 600°. Assuming, therefore, that copper and wrought-iron plates follow the same law, and taking into account the rapid conducting powers of the former, we may reasonably conclude that the resisting powers of the plates and stays of locomotive boilers are not seriously affected by the increased temperature to which they are subjected in a regular course of working. This part of the subject is, however, entitled to further consideration; and I trust that some of our able and intelligent superintendents will institute further inquiries into a question which involves considerations of some importance to the public, as well as to the advancement of our knowledge in practical science.†

After these results, and the investigations given previously on the form and strength of Boilers, it will not be necessary to pursue this part of the subject further

*Vide Transactions of the British Association for the Advancement of Science, vol. vi. p. 486.

† Since the above was written I have completed a series of experiments on wrought-iron plates and rivet-iron at various temperatures, from 30° under the freezing-point to red heat. These experiments are the more satisfactory as they exhibit. no diminution of strength from 60° to 400° of temperature; but an increase of heat from that point to a dull red heat

than to observe, that in every construction of vessel calculated to generate steam of high elastic force, we should preserve a large margin of strength as regards the working pressure and the ultimate power of resistance. Six times the working pressure is not too much to provide for contingencies, and vessels so constructed are better calculated to avert danger from explosion than those whose powers of resistance verge upon that of the bursting pressure.

shows a considerable reduction of strength and a great increase of ductility, the plates being in the ratio of 20-3: 15.5 tons per square inch, as regards strength, and the rivet-iron as 35:16.

The experiments on temperature are, therefore, quite conclusive as to the effects produced on wrought-iron whenever it approaches a red heat. At that temperature nearly one-half of the strength is lost: it becomes exceedingly ductile, and is drawn considerably in the direction of the strain before its cohesive powers are destroyed. In this respect we may, however, remark, that it suffers little or no diminution in its powers of resistance up to a temperature of 500°.

154

LECTURE VII.

ON STEAM AND STEAM BOILERS.

WHEN last I had the honour of addressing you, it was upon the strength and form of vessels calculated to hold steam of high elastic force; and in these enquiries I endeavoured to show the necessity of attending to two things:

1st. The material.

2nd. Its distribution, and the form in which it should be applied to effect the greatest powers of resistance.

In the consideration of these subjects, I had reference to vessels subjected to internal pressure, having a tendency to force them open, or to tear the parts asunder. This force or tendency to rupture from a tensile strain is probably the most common to which boilers are subjected; but we have other forces to guard against, such as that which arises from collapse, which, I regret to say, has not in some constructions received that attention which the importance of the subject demands.

It is well known to persons conversant with the construction of the steam-engine, that a very considerable proportion of our boilers have internal flues, some of them of large diameter; and these, being surrounded by water, are compressed in every direction with the same force as the steam which acts upon its surface.

Now, it is evident that an internal flue, such as we have described, unless it is made perfectly cylindrical, is liable to lose its shape, and to become a flattened instead of an arched or cylindrical surface. Whenever this takes