The working pressure of the flue is found therefore to be, under the limiting formula, 70 lbs. per square inch, and this is the pressure that would be allowed.

Now suppose that the construction of the superheater was such that the heat or flame impinged upon it at nearly a right angle to its surface, and the working pressure would be reduced, under the rules the constant would be reduced to 22400 instead of the 30000 used for the shell in the previous calculation, and the calculation under this new condition would be thus:

PART IV.

EXPERIMENTS* UPON THE STRENGTHS OF RIVETTED JOINTS, OF STAYED FLAT SURFACES, AND OF BOILER FURNACES.

The experiments given in the following pages are valuable, not only from the eminent authority from which they emanate, but also because the data they give bears evidence on every hand of the soundness of the Board of Trade Rules bearing upon the respective subjects.

Summaries only of the above experiments are given, as they are sufficient for the above purpose. While Mr. Kirkaldy's tables in full would be valuable, they occupy more space than can be given to them in a work of this kind.

The experiments upon rivetted seams were made by Mr. David Kirkaldy under direction of the Engineer Surveyor in Chief of the Board of Trade and his assistants.

The object of these experiments was to ascertain, first, the different effects produced upon the strengths of steel plates, by drilling, by punch ing and afterwards boring out the holes, and by punching the plates and then annealing them; and secondly, to test the strengths of rivetted joints in which the holes had been made in each of the above ways.

Beginning with the tests of plates, we have in summary 1 the dimen sions, and in summary 2 the ultimate stress (or stress at which the plates severed) per square inch of metal left between the holes.

* From a 66

memorandum for the information of Board of Trade Surveyors" issued by the Board of Trade in 1885.

ULTIMATE STRESS OF PERFORATED PLATES PER SQUARE INCH.
OF GROSS AREA AT HOLES.

The superiority of the drilled plates over the others is clearly seen by the larger stresses sustained, whilst the punched plates, by the smaller stresses they bore, are evidently inferior to all the other specimens.

The calculated percentage value of the perforated plates as compared with the unperforated is found by the formula,

and is given by Mr. Kirkaldy in the last line of Reports D, E, F, G, and H* as "Width of plates less holes percentage".

The values obtained by this formula differ from the percentages deduced from the results of tests. These, which are given in Mr. Kirkaldy,s report under the heading of "Ratio of holes to solid," are found by the formula:

(2.)

Percentage of strength of plate at holes compared with solid plate

Stress per square inch of gross area at holes

Stress per square inch of solid plates

X 100.

་

The values given by the above formula are the actual percentages of the strength of the perforated plates as compared with the solid plate, and for the lengthway specimens, they and the calculated values are to be found in the summary 3.

The differences between the results as found by formula (1) and (2) are given in summary 4 for the lengthway specimens.

It will be understood, that in summary 2 the stresses are calculated with refernce to the gross area of the perforated plates, i. e., including the metal removed by the holes. In the cases of the specimens of the same thickness in which the diameter of the holes and the distance between them are identical, this is by far the best system of comparison. In cases, however, where the percentage of metal removed by perforation differs, it is obvious that the stresses borne by the different perforated plates when expressed in terms per square inch of the gross area at holes, cannot properly be compared. A comparison can, however, be made by means of the stresses borne per square inch of net section between the holes. A knowledge of the stresses per square inch of the metal *The reports here referred to are those which are here summarised.