gradually rose to a maximum with the 12th, beyond which to the 18th they were progressively reduced. The crushing strain was reduced to its first minimum with the 4th melting, and attained its maximum with the 14th. Taking the iron of the 1st melting as a standard, the maximum increase in strength due to re-melting was 32 per cent. with the tensile ; 41 per cent. with the transverse, and 135 per cent. with the crushing strain. It should be observed that Mr. Fairbairn's experiments did not include the tensile strain: col. 4 in Table 3 has been calculated (499) from the transverse and crushing strains in cols. 2 and 3; the results in col. 4 are very regular among themselves, despite the irregularities in col. 3.

TABLE 3. Of the EFFECT of RE-MELTING on the STRENGTH of

NOTE.-Maximum and minimum results marked *.

(7.) It would appear from all this, that the method of obtaining increased strength by re-melting cast iron is very

uncertain; it will also be very expensive in fuel, labour, and waste of metal. With iron such as that in (5), where the mean tensile strength was increased from 1 to 18.265·6 = 3·26 at the 4th melting, it would no doubt be commercially advantageous: in such a case experiments should be specially made on the iron intended to be used (87).

(8.) By maintaining cast iron in a state of fusion for lengthened periods, the tensile strength is greatly increased: thus with iron twice re-melted and kept in fusion for

15,861

20,420

2

24,383

3 hours

25,733 lbs.

per square inch. In another set of experiments, the time

being

=

=

(9.) Cold-blast iron is considerably stronger than hot-blast iron; taking the former = 1.0, that of the latter was found at Lowmoor 831; at Dowlais, 835 at Ystalyfera, 802. The deterioration in strength appears by the American experiments to be proportional to the temperature of the blast; thus the strength of cold-blast iron being 1.0, it is reduced to ⚫865 with the blast at 150°, and to 807 at 250°, &c.

WROUGHT IRON AND STEEL.

(10.) The strength of wrought iron increases, as might be expected, with repeated working in the fire and under the hammer. Mr. Clay found that the strength of a puddled bar being 1.0, it becomes 1.36 when piled three or four times, and 1.41 when piled six times; beyond that point, however, its strength declines, and is reduced to that of a puddled bar when piled twelve times.

The same authority has shown that with steel, the strength of a puddled bar being 1.0, it becomes 1.253 at the fourth piling, after which it declines and is reduced to 0.94 at the seventh piling.

(11.) "Welded Joints."-The strength of wrought-iron welded joints appears by the experiments of Kirkaldy to be very variable, the mean from eighteen experiments on bars from 11 to

inch diameter = 8066, the strength of a solid bar being 1.0; in extreme cases it is as low as 562, or little more than half, in others as high as 974, the great difference being due no doubt to imperfect workmanship.

With steel the loss of strength by welding is still more considerable; the same authority shows that the strength of welded steel joints varies from 55 to 404 of that of a solid bar. The strength of steel is also affected considerably by hardening, tempering, annealing, &c., as is shown by Table 1: when heated and quenched in oil, Mr. Kirkaldy obtained the extraordinary strength of 96.1 tons per square inch, which is exceptional and anomalous. The same steel made as hard as possible by being highly heated and quenched in water, gave 40.2 tons only: the mean for ordinary rolled or tilted bars being 47.84 tons per square inch: see cols. 3, 9 of Table 2.

(12.) "Screwed Bolts.”—There are two ways of measuring screwed bars, namely by the diameter at the top of the thread, or that of the plain bar before screwing, and by the diameter at the bottom of the thread: the former is the most convenient, and will be followed here. Mr. Kirkaldy obtained some curious results; he found that when the thread was chased in the lathe, or cut by new dies, the strength was nearly proportional to the diameter at the bottom of the thread as might be expected, and varied with different sizes, between 67 and 82 per cent. of that of a plain bar, the mean being 72.5 per cent. But when old dies were used, the metal seemed to be compressed rather than cut, and the strength was much greater than with new dies, varying from 77 to 89 per cent. of that due to a plain bar, the mean being about 85. It will be the safest course to reckon the strength as due to new dies, or 72.5 per cent. of plain bar; see cols. 5, 6 of Table 2.

(13.) "Plate-iron and Steel."-Rolled plates of iron and steel are rather weaker than the same materials in the form of bars, as shown by Table 1; the ratio happens to be nearly the same for both: thus taking Kirkaldy's results, with wrought iron we

?

have 21.6 25.7 = .84, or 84 per cent., and with steel 38.447.880, or 80 per cent.

Effect of the Grain."-Experiments have shown that the tensile strength of both wrought iron and steel, lengthways of the grain, is greater than that crossways, as shown by Table 1: thus with wrought iron we have 22.6 ÷ 20.6 1.097, or 9.7 per cent.; and with steel 40 136.6 1.096, or 9.6 per cent., being practically the same for both.

=

In arranging the plates for girders, &c., those subjected to tension should be cut so that the strain is in the direction of the grain, and in boilers, where the circumferential strain is double the longitudinal (71), the direction of the grain should be arranged accordingly.

(14.) “ Effect of Annealing.”—It has been found by experiment that the effect of annealing wrought iron in the bar, plate, and chain form is to reduce the tensile strength: this is the more remarkable, being just the reverse of the effect on steel plate, which is to increase the strength as much as 55 per cent. (38).. By hammering cold the strength of wrought iron is much reduced, but by annealing it is partially restored experiments at Woolwich show the effect of both processes on bars of different sizes: thus bars

per square inch. The mean tensile strength of ordinary bar iron is 25.7 tons per square inch by Table 1, hence the loss by cold-hammering 18

32

36

13

per cent.: the mean being 27 per cent. After annealing, the strength became

22.1

24.6

23.5 tons.

per square inch. Hence after both processes, there still remains a loss of

14

4

8.6 per cent.