Sept. 17, 1885 | 
NATORE 
489 
the principles of estimating the magnitude of the stresses on the 
different members of a structure, but not so in proportioning 
the members to resist those stresses. The practical result is 
that a bridge which would be passed by the English Board of 
Trade would require to be strengthened 5 per cent. in some 
parts and 60 per cent. in others before it would be accepted by 
the German Government or by any of the leading railway com- 
panies in America. This undesirable state of affairs arises from 
the fact that in our own and some other countries many en- 
gineers still persistently ignore the fact that a bar of iron may 
be broken in two ways—namely, by the single application of a 
heavy stress or by the repeated application of a comparatively 
light stress. An athlete’s muscles have often been likened to a 
bar of iron, but, if ‘‘ fatigue” be in question, the simile is very 
wide of the truth. Intermittent action—the alternative pull and 
thrust of the rower, or of the labourer turning a winch—is what 
the muscle likes and the bar of iron abhors. Troopers dismount to 
rest their horses, but to relieve a bar of iron temporarily of load 
only serves to fatigue it. Half a century ago Braithwaite cor- 
rectly attributed the failure of some girders, carrying a large 
brewery vat, to the vessel being sometimes full and sometimes 
empty, the repeated deflection, although imperceptibly slow and 
wholly free from vibration, deteriorating the metal, until, in the 
course of years, the girders broke. These girders were of cast- 
iron; but it was equally well known that wrought-iron was 
similarly affected, for in 1842 Nasmyth called the attention of 
this Section to the fact that the ‘‘alternate strain” in axles 
rendered them weak and brittle, and suggested annealing as a 
remedy, he having found that an axle which would snap with 
one blow when worn would bear eighteen blows when new or 
after being annealed. 
So important a matter as the action of intermittent stresses 
could not escape the attention of the Royal Commissioners 
appointed in 1849 to consider the application of iron to railway 
structures, and some significant and sufficiently conclusive ex- 
periments were made by Capt. Douglas Dalton and others. Cast- 
iron bars 3 inches square and 13 feet 6 inches span between 
the supports were deflected, both by the slow action of a cam 
and the percussive action of a swinging pendulum weight. When 
the deflection was that due to one-third of the breaking weight, 
about 50,000 successive bendings by the cam broke one of the 
bars, and about 1000 blows from the pendulum another. When 
the deflection was increased from one-third to one-half, about 
500 applications of the cam, and 100 blows, sufficed to rupture 
two of the specimens. Slow-moving weights on bars and ona 
small wrought-iron box girder gave analogous results ; and the 
deduction drawn by the experimenters at the time was that 
“iron bars scarcely bear the reiterated application of one-third 
the breaking weight without injury, hence the prudence of 
always making beams capable of bearing six times the greatest 
weight that could be laid upon them.” 
Although these experiments were entirely confirmatory of all 
previous experience, they would appear to have little influenced 
the practice of engineers, since Fairbairn, more than ten years 
later, in a communication to this Section, said that opinions 
were still much divided upon the question whether the con- 
tinuous change of load which many wrought-iron structures 
undergo has any permanent effect upon their ul'imate powers 
of resistance. To assist in settling the question he communicated 
to the Association the results of some experiments carried out 
by himself and Prof. Unwin on a little riveted girder 20 feet 
span and 16 inches deep. Once more the same important but 
disregarded facts were enforced on the attention of engineers. 
About 5000 applications of a load equal.to four-tenths of the 
calculated breaking load fractured the beam with the small ulti- 
mate deflection of three-eighths of an inch, and subsequently, 
when repaired, the beam broke with one-third of the load and a 
deflection of but a quarter of an inch, which sufficiently indicated 
how small a margin the factor of safety of four, when currently 
adopted, allowed for defective manufacture, inferior material, 
and errors in calculation, Still nothing was done, and the 
general practice of engineers and the Board of Trade regulations 
continued unaltered. 
Soon after the introduction of wrought-iron bridges on railways, 
the testimony of practical working was added to that of experi- 
ments. In 1848 several girder bridges of unduly light propor- 
tions were erected in America, and one of 66 feet span broke 
down under the action of the rolling load in the same manner as 
Fairbairn’s little experimental girder. Again, in early American 
timber bridges the vertical tie-rods were often subject to stresses 
oscillating between 1 ton and fo tons per square inch and up- 
wards. Many of these broke, as did also the suspension bolts 
in platforms subjected to similar stresses. In my own ex- 
perience, dozens of broken flange-plates and angle-bars, and 
hundreds of sheared rivets, have been the silent witnesses of the 
destructive action of a live load. Like evidence was afforded by 
early constructed iron ships deficient in girder strength. Under 
the alternating stresses due to the action of the waves weak- 
nesses not at first apparent would, in the course of time be 
developed, and additional strength, in the way of stringers and 
otherwise, become imperative. 
If none of the preceding evidence had been forthcoming, the 
results of the historical series of experiments carried out by 
Wohler for the Prussian Ministry of Commerce would alone be 
conclusive. For the first time a truly scientific method of 
investigation was followed, and an attempt was made to deter- 
mine the laws governing the already proved destructive action 
of intermittent stresses. In previous experiments the bar or 
girder was alternately fully loaded and wholly relieved of load. 
Wohler was not satisfied with this, but tested also the result 
of a partial relief of load. The striking fact was soon 
evidenced on testing specimens under varying tensions, that the 
amount of the variation was as necessary to be considered as 
that of the maximum stress. Thus, an iron bar having a tensile 
strength of 24 tons per square inch broke with about 100,000 
applications of a stress varying from 7/ to 21 tons, but resisted 
4,000,000 applications of the 21 tons when the minimum stress 
was varied from zz/ to 11} tons. The alternations of stress in 
the case of some test pieces numbered no less than 132,000,000 ; 
and too much credit cannot be bestowed by engineers upon 
Wohler for the ingenuity and patience which characterised his 
researches. As a result, it is proved beyond all further question 
that any bar or beam of cast iron, wrought iron, or steel may be 
fractured by the continued repetition of comparatively small 
stresses, and that, as the differences of stress increase, the 
maximum stress capable of being sustained diminishes. 
Various formulz based upon the preceding experiments have 
been proposed for the determination of the proper sectional 
area of the members of metallic structures. These formulz differ 
in some essential respects, and doubtless many experiments are 
still required before any universally accepted rules can be laid 
down. Probably at the present time the engineers who have 
given the most attention to the subject are fairly in accord in 
holding that the admissible stress per square inch in a wrought- 
iron girder subject to a steady dead load would be one and a 
half times as great as that in a girder subject to a wholly live 
load, and three times that allowable in members subject to 
alternate tensile and compressive stresses of equal intensity, such 
as the piston-rod of a steam-engine or the central web-bracing 
of a lattice girder. If the alternations of stress to be guarded 
against are not assumably infinite in number, but only occasional 
—as in wind bracing for hurricane pressures, or in a vessel 
amongst exceptionally high waves—then the aforesaid ratio of 
3, 2, and 1 would not apply, but would more nearly approach 
the ratios 6, 5, and 4. 
Hundreds of existing railway bridges which carry twenty 
trains a day with perfect safety would break down quickly under 
twenty trains per hour. This fact was forced on my attention 
nearly twenty years ago by the fracture of a number of iron 
girders of ordinary strength under a five-minute train service. 
Similarly, when in New York last year I noticed, in the case 
of some hundreds of girders on the ‘‘ Elevated Railway,” that 
the alternate thrust and pull on the central diagonals from trains 
passing every two or three minutes had developed weaknesses 
which necessitated the bars being replaced by stronger ones after 
a very short service. Somewhat the same thing had to be done 
recently in this country with a bridge over the Trent, but the 
train service being small the life of the bars was measured by 
years instead of months. If ships were always amongst great 
waves the number going to the bottom would be largely 
increased, for, according to Mr. John, late of Lloyd's, ‘‘many 
large merchant steamers afloat are so deficient in longitudinal 
strength that they are liable under certain conditions of sea to 
be strained in the upper works to a tension of from 8 to 9 tons 
per square inch, and to a compression of from 6 to 7 tons— 
stresses which the experiments already referred to proved would 
cause failure after a definite number of repetitions. Similarly, 
on taking ground or being dry-docked with a heavy cargo on 
board, it has been shown that vessels are liable to stresses of 
over II tons per square inch on the reverse frames, but no 
