First person account of the disaster, scientifically viewed
Through the courtesy of the Engineering News of New York I am permitted to insert the following extracts from its columns. The proprietors sent a corps of reporters, consisting of expert constructing and topographical engineers, who made exhaustive examinations, took photographic views and prepared accurate maps. Its resultant articles in four of its issues were by far the most complete and satisfactory account of the nature and causes of the disaster, as viewed from the standpoint of engineering science.
The construction of the dam, which in its later reconstructed form has just failed, was first authorized in 1836, but it was not till 1839 that $70,000 was appropriated for it, and Wm. E. Morris, Principal Assistant Engineer, and an able and experienced man, was placed in charge of it, and also of the dam near Hollidaysburg, on the east side of the mountains. The west dam, which has now failed, raised the water 62 feet, was 850 feet long on top, covered 4 acres of ground, and stored up 480,000,000 cubic feet of water, the estimated cost being $188,000, one year being required for construction. The engineer, Mr. Morris, in his report, dated Johnstown, November 1st, 1839, said:
The western division of the Pennsylvania Canal is supplied by water taken by feeders from the Little Conemaugh River and Stony Creek, which form a junction at this place. The valley of the Stony Creek is not well suited for a reservoir, as it has a steep descent, and in time of floods the stream is far too large and unmanageable.
The main branch of the Conemaugh has but one reservoir site on it, and that would flood the village of Jefferson and the railroad. Of the other branches, the south (the one adopted) is the only one that drains sufficient country to furnish a certain supply. This branch, when gauged in September, after one day’s rain, discharged in 24 hours 60,000,000 cubic feet of water. At the same time there were flood marks two feet higher, and a moderate estimate of full discharge in 24 hours is 160,000,000 cubic feet. The best site for a dam is two and a quarter miles from the mouth. A dam 62 feet high and 850 feet long on the top is suggested.
The valley is narrow at the dam and widens immediately above into an extensive basin. There is solid rock at both ends of the dam in which channels may be cut for flood water discharge. Drifts and shafts were sunk to insure in advance good foundations.
Two plans for the dam were presented: (1) a crib dam of timber and stone with weir on top to pass over freshet water; and (2) a mound of stone and earth, made perfectly water tight, raised 10 feet above the surface of the pool, having a waste weir in solid rock at one or both ends of dam.
Mr. Morris, however, pointed out that there were serious objections to the first plan, especially difficulty in uniting the body of the dam with the embankment to prevent breakage under the pressure, danger of undermining base, perishable nature of material constant repairs etc., and concluded that the second plan promised a permanent and durable dam, maintainable at small expense. He then presented estimates of quantities of material and labor needed, showing a total cost of $188,000.
In the annual report of the Canal Commissioners ending November 30th, 1852, it is stated:
The western reservoir will be entirely finished in a very short time . . . . It has been constructed in a most substantial manner, reflecting great credit on the contractors and engineers in charge. The sluice gates were closed in June, 1852 ; by August the water was 40 feet deep.
There had been some intermediate delay due to legal proceedings; but the dam appears, from the reports, to have been executed precisely according to the original plans.
After the abandonment of the State canal, shortly after its purchase by the Pennsylvania Railroad in 1857, a breach developed in this dam that involved no serious disaster. The gap was about 150 feet across at the top and extended down nearly to the bottom of the dam, yet still leaving enough of the old dam intact to retain a small remnant of the original reservoir. This gap is stated to have been closed and the damage repaired in 1879, at a total cost of $17,000, a sum hardly adequate to properly repair even a much smaller gap, not to speak of raising the original level; which the latter would not only have been a rash and dangerous procedure, without reconstructing the dam completely, but would have involved a heavy expenditure of money, however rashly done.
The engineer in charge of the reconstruction was General James N. Morehead. The primary cause of failure lay in no part of the work which he reconstructed, but in lack of sufficient spillway. The considerable leakage reported from the dam, however, would indicate that the reconstruction was none too secure, as would also its low cost, and it may well be that its lack of the substantial solidity of the old structure aggravated the disaster by aiding the dam to go out all at once instead of gradually.
If anything can be said to be clear about this catastrophe, it is that the dam’s poor structure as we must now believe that it was failed just when and as it did merely for lack of a little more sectional area in the spillway. For nearly half a century this spillway has sufficed; it may have been at times severely tried, but it has never before caused water to actually run over the top of the dam, and it probably would not have done so on this occasion had there been slightly more provision for normal discharge; because the area of the impounded water was not very large, not over 450 acres at most, to judge by the old reports, and the water took many hours to rise to the crest level.
The original estimates were for half earth and half rock and slate spoil. We can readily understand how this could be, and yet the dam be made practically all of earth. A rotten, slaty shale rock lies close beneath the surface, which for a foot or two disintegrates into earth almost completely after excavation. It was probably contemplated to excavate half the material from this soft rock, and a good deal was so excavated, but the large borrow pits on each side at the bottom of the reservoir indicate that a good deal of material was taken from lower down on the valley slopes, where there was more earth and perhaps a rottener rock. Be this as it may, there was hardly a piece of rock as big as a man’s fist is any of the central part now exposed to view. The lower edge of the northeast side of the gap showed an interior coating of spalls; but while these may be the outer evidences of a similar lining extending along the old work, inside the heavy riprap, it is more likely that it is a mere exterior coating of a few stones, which chanced to lodge there in the washout or fall from above. There is no similar evidence of interior stone in the southwest end.
The excessive phenomenal rainfall, the second contributing cause of the disaster, began Thursday night, May 30th, after several days of moderate prior rains, and continued almost till the dam gave way. The lake, although discharging its best through the contracted spillway, gradually rose at the rate of a foot an hour for several hours before the break, implying that the crest of the dam was 7 feet or so above the normal level, as the old records indicate until, at 2:30 P.M., of Friday, May 31st, water began to run over the crest.
The dreaded catastrophe was then certain, and could not be long delayed, no earth dam being capable of sustaining such a discharge over it. It took half an hour for the increasing current to gradually cut away the earth support from the lower side of the rubble heart wall, and then, at 3 P.M., the dam gave way. The breach once made was instantly enlarged to nearly its final dimensions, 250 feet across. In fact, some of the eyewitnesses state that it “burst with a report like thunder,” which it is quite possible would have been the effect, even if the real course of events was gradual but very rapid disintegration of the embankment in the torrent.
The foregoing account corresponds most correctly with that given by intelligent eyewitnesses of the break. A workman, who seems to have some grudge against the fishing company, absurdly declares that the dam had been leaking badly for weeks, and that on the morning before the disaster ”jets squirted out for thirty feet from the face of it,” and that these leaks rapidly increased until the whole gave way, without any running of water over the top. This is so far consistent with other facts that the dam is known to have been leaking for several years past, and has been the subject of considerable apprehension, especially since the spring floods of 1888.
The very remarkable fact that the dam is stoutly claimed by the company officers themselves to have been “inspected twice a month,” by some engineers as yet unknown, tends to show that it was in a dubious condition, for a dam in good condition has no need of such frequent inspection; and it is also specifically reported that the dam was inspected by Robert L. Haliday, Superintendent of the Lewiston and Sunbury Division of the Pennsylvania Railroad, “some years ago,” and declared unsafe. But on the whole, the evidence is decided that it was the flow of water over the top of the dam, and not in any sense inability of the dam to sustain the static pressure of water, which caused the disaster.
The dam stands about 450 feet above the town of Johnstown, the valley connecting it being about twelve miles long, counting all the bends of the channel, or about ten miles long by the railway. In the valley were situated the following places, to which we add also the census populations of 1880 and the estimated populations of I889.
It appears altogether probable that the population on the day of the disaster was fully 37,000, including Cambria City, which was below the fine stone bridge, which withstood the flood, and was speedily so choked by drift and human bodies as to form a second dam to drown out Johnstown.
The fates which have befallen the above towns may be thus briefly summarized, according to the best information now at hand.
South Fork “No very serious results; a few drowned.”
Mineral Point “Entirely wiped out.”
Conemaugh “Almost depopulated,” and every building swept away.
Woodvale….The same.
Conemaugh Borough The same, with more escapes to the hill.
Johnstown, Cambria– The average opinion of a number of leading citizens is that 10 per cent of the population is lost.
It is generally estimated that one-third of the bodies, at least, will never be recovered at all, being burned up, buried under sand and detritus or carried down the river. We do not see how it can ever be expected to determine the loss much more accurately than by the rude process we have just used, unless it may be by a census of the remaining population combined with an estimate based on the vote of the last Presidential election.
That the chances for life of any one in the way of the flood after the dam had once given way were very small, is evident when we remember that, as in all such cases, the flood advances, not with a comparatively shallow advance guard, but with a solid wall in front, which strikes a house or human being with terrible velocity, and directly downward rather than from the side. All accounts agree that the water did in fact advance “like a wall 30 or 40 feet high and it is wholly in accordance with physical laws that it should do so, in fact, it cannot well do otherwise. The water which first flows out, being retarded by the rough surface, trees, rocks, houses, animals and other obstacles, speedily loses its own velocity, but furnishes an almost frictionless surface over which other water can slide, like ice down a plank, with almost the full theoretical velocity due to the fall. Thus the top of the flood is continually moving much faster than the bottom, and falls over the end of it when it reaches it, to be itself retarded and to surrender a large part of the energy due to its velocity in tearing up the ground and beating down, not driving forward, any unfortunate creature or structure which stands in its paths. The bottom of the flood is relatively stationary, the top has its full theoretical velocity due to the fall, and the average velocity of advance hence becomes but slightly more than half the top velocity; the water, in fact, rolling over itself at the front very much as a wheel rolls, except that the lower part of the flood-wheel never rises again when it once strikes the ground. The fall from the reservoir to Johnstown having been about 450 feet, the actual time taken by the flood to reach Johnstown corresponds very closely with that which this theory requires. The top of the water would have nearly the theoretical velocity due to a fall of 450 feet, which is about 100 miles per hour, that due to the 160 feet fall of Niagara Falls being about 70 miles per hour. Allowing for all the frictional losses, a man who sustained the direct impact of this torrent had about the same chance of resisting it and escaping alive as he would in sustaining the impact of Niagara Falls itself. The tendency which such a flood would have to bury many bodies is also evident.
The reservoir, at its normal water-level, held 480,000,000 cubic feet of water, but at the break it contained nearly 640,000,000 cubic feet, or say, 20,000,000 net tons. How vast a body of water this is will be better appreciated by comparing it with Niagara Falls The discharge over the falls is in the neighborhood of 18,000,000 cubic feet per minute. It would therefore take nearly thirty-six minutes for Niagara Falls to discharge an equal body of water. The reservoir was emptied, all but the last harmless drippings, in just about that time, note that during its continuance a body of water substantially, equal to the vast flood of Niagara Falls was pouring through the 429 feet gap in the dam.
Had this body of water struck Johnstown, there would have been even less of the city left than there is, but this did not occur. The whole valley above it had likewise to be filled with water. This valley is somewhat irregular in width, varying from 300 to 2,500 feet; but independent and closely agreeing estimates place its average width at 750 feet for a depth of 20 feet, giving a cross section of 15,000 square feet. The distance from the dam to the Johnstown bridge is by the river channel 18 miles. The railway is about two miles shorter, owing to cutting across bends. It would therefore require a total volume of about 14×5,280×15,000=1,108,800,000 cubic feet to fill the whole valley 20 feet deep, or about óo per cent more than the reservoir contained. As nearly as we can ascertain it took the flood fifteen minutes to cover the 12 miles between South Fork Station and Johnstown or say twenty minutes from the break until the jam at the Johnstown bridge occurred. At the latter moment, therefore, the channel could not possibly have been 20 feet deep throughout, although it was probably deeper than that at the lower end.
The total energy communicated to the water, which had to expend itself in some way before the water could come to rest, and which was in fact nearly all expended between the dam and Johnstown bridge, was the inconceivably vast aggregate represented by 20,000,000 tons falling 400 feet.
The conclusion of the whole matter, as reached by the expert examination of the Engineering News, is thus summed up:
The original dam was designed and built, as already stated, by the late Wm. E. Morris, Principal Assistant Engineer Pennsylvania State Canals, in charge of the Western Division. He was an able and experienced engineer, and the dam, under his supervision, was thoroughly well built by the late General James N. Morehead. It had no central core of masonry as the preliminary estimates indicated, but it was built in horizontal layers, thoroughly watered and rammed, riprapped on both slopes, and provided with an ample spillway through rock, and an arch culvert underneath it, through which ran five two-feet cast-iron pipes for discharging the water during the dry season into the South Fork, from which the water ran down in the river channel to feed the State canal at Johnstown, 14 miles below by the river, which was the head of canal navigation. The canal, and the dam with it, were abandoned as public structures in 1857-1858.
The first break in it, which occurred in July, 1862, was caused by a defect in the foundations of the culvert, through which the five two-feet discharge pipes were carried. This break did comparatively little damage, the reservoir having been only half full, and the discharge having been quite slow, wholly from the bottom, and choked from time to time by fall of material from above. So far as we can determine, it carried out only about half as much material from the dam as the last break, or about 50,000 cubic yards. There is great difficulty in determining with exactness the quantity of material carried out in the first break. Most of those who ought to know say, ‘150 to 200 feet wide on top,’ but then most of the accounts of the last break have called the gap only 200 feet, whereas it is actually more than twice that. No photographs taken after the first break are discoverable, but it is probable that the gap was smaller than the present one, especially on the upper side, where the remarkable benches of old work, still remaining after two break-aways, testify to the care and thoroughness with which the original embankment was constructed, and rammed in regular layers.
At the bottom of the old break, also, enough of the material remained to make a little pond about 8 feet deep above the dam, which remained in this condition, unused, until, in May, 1875, the property, consisting of something over 500 acres, was sold to Congressman John Reilly. The lake itself was about 400 acres in size, not 700 as has been reported. After holding the property unused till 1879, Mr. Reilly offered it for the sum of $2,000 to the late Colonel B. F. Ruff, an old and successful railroad and tunnel contractor, and the originator of the South Fork Fishing and Hunting Club. Colonel Ruff interested two other Pittsburgh gentlemen in the project, and stated to them that the dam could be reconstructed for a sum not to exceed $1,500, and that ‘he would take a contract to do it for $1,700.
On this basis the club was organized, and for some time these three gentlemen were its only members. Not one of them is now connected with it. Colonel Ruff’s idea had been to reconstruct the dam much lower only 40 feet high; but it soon appeared that to cut down the rock spillway would cost more than to reconstruct the dam to its original height, and by this time this had been done the total expenditure, as shown by the pay rolls, had been slightly over $10,000, or about twenty cents per cubic yard. There still remained to be done the riprapping of the slopes and other miscellaneous work, as to which our information is less precise, being only that it ‘may’ have cost $7,000, but not more, bringing the total cost up to the very small figures of $17,000, which have been given on other authority in newspaper dispatches. This work was all done in the summer of 1880. The original dam was estimated to cost $188,000, and actually cost nearly $240,000.
Colonel Ruff engaged as foreman and superintendent for this work a Mr. Edward Pearson, of Pittsburgh. It is a general impression in the vicinity of Johnstown and Pittsburgh, among those who know anything about it, that Mr. Pearson was the ‘engineer’ of the repairs, but this is incorrect. He is not and never has been an engineer, but after 1880 was employed in the local freight department of the Pennsylvania Railroad at Pittsburgh, until he formed his present connection, which is with the firm of Haney & Co., general teamsters for the Pennsylvania Railroad freight department.
We were also told that Colonel Ruff was ‘the engineer,’ but this statement also is incorrect. So far as we can ascertain by diligent inquiry, he not only was never an engineer, but he had never been engaged, before this time, even as a contractor on water-works or dam construction. If he was ever so employed at all, it would appear that it must have been to an unimportant extent. In fact, our information is positive, direct and unimpeachable that at no time during the process of rebuilding the dam was ANY ENGINEER WHATEVER, young or old, good or bad, known or unknown, engaged on or consulted as to the work. The precautions taken against failure were only such as an experienced railroad contractor’s knowledge of hydraulic engineering indicated were admissible without further increasing a contemplated investment of $3,700, which had to be increased at best by over $10,000.
Information gathered for us by Mr. T. S. Miller, M. E., of the Lidgerwood Manufacturing Co., who assisted in our surveys, corresponds with that gathered by us from other reliable sources, that the work of reconstruction was done with the slight care which the preceding facts make probable. The old material, which had caved in, and so lost its compactness, was left untouched; the top of the dam was worked down on to it; the old pipes and culvert, which still remained in somewhat injured condition, were covered over with earth and permanently closed, a double row of hemlock plank sheet piling being driven across the old channel. The water during reconstruction was carried across the dam in a board flume, which was raised from time to time as the work progressed. There was no careful ramming in watered layers, as in the first dam, although some say there was some ramming. There was much leaking during the process, and some tons of hay and straw were filled in. The dam was finally made fairly tight, but there has always been some leakage at the bottom, and of late years this has been increasing. The truth as to the exact amount of leakage is very difficult to ascertain. The original crest height of the dam was decreased from one to three feet, and the spillway was shortly after obstructed with gratings to retain fish, and a trestle bridge was built across the opening.”
Negligence in the mere execution of the earthwork, however, if it existed is of minor importance, since there is no doubt that it was not a primary cause of the disaster; at worst, it merely aggravated it. The primary causes of the disaster were the lowering of the crest, the dishing, or central sag in the crest, the closing of the bottom culvert, and the obstruction of the spillway.
Of the final blow as it struck Johnstown, this may be said:
The main body of the flood rushed directly westward, through the very heart of Johnstown, sweeping it clean, and impinging directly against the mountainside. The bridge, whose ‘resistance of the torrent’ has been the matter of so much talk, was a noble four-track structure, just completed, fifty feet wide on top, thirty-two feet high above the water line, consisting of seven skew spans of fifty-eight feet each. It still remains wholly uninjured, except that it is badly spalled on the upper side by blows from the wreckage, but that it so remains is due solely to the accident of its position, and not to its strength, although it was and is still the embodiment of solidity. Had the torrent struck it squarely, it would have swept it away as if it had been built of cardboard, leaving no track behind; but fortunately (or unfortunately) its axis was exactly parallel with the path of the flood, which hence struck the face of the mountain full, and compressed the whole of its spoils gathered in a fourteen-mile course into one inextricable mass, with the force of tens of thousands of tons moving at nearly sixty miles per hour. Its spoil consisted of (1) every tree the flood had touched in its whole course, with trifling exceptions, including hundreds of large trees, all of which were stripped of their bark and small limbs almost at once; (2) all the houses in a thickly settled town three miles long and one-fourth to one-half mile wide; (3) half the human beings and all the horses, cows, cats, dogs and rats that were in the houses; (4) many hundreds of miles of telegraph wire that was on strong poles in use, and many times more than this that was in stock in the mills; (5) perhaps fifty miles of track and track material, rails and all ; (6) locomotives, pig-iron, brick, stone, boilers, steam engines, heavy machinery and other spoil of a large manufacturing town. All this was accumulated in one inextricable mass, which almost immediately caught fire from some stove which the waters had not touched. Hundreds of human beings, dead and alive, were caught in it, many by the lower part of the body only. Eyewitnesses describe the groans and cries which came from that vast holocaust for nearly the whole night as something fearful beyond all power of description.
Excerpt from David Beale, Through the Johnstown Flood, 1890. Pp.88 -99
