Skip to main content

Full text of "Transactions of the American Society of Civil Engineers"

See other formats












m 




^ 



TRANSACTIONS 



AMERICAN SOCIETY 



CIVIL ENGINEERS 



(INSTITUTED 1852) 



VOL. LXV 
DECEMBER, igog 



Edited by the Secretary, under the direction of the Committee on Publications. 

Reprints from this publication, which is copyrighted, may be made on condition that 

the full title of Paper, name of Author, and page reference are given. 



NEW YORK 

PUBLISHED BY THE SOCIETY 
1909 



"^GZO.G 



Entered according to Act of Congress, in the year 1909, by the American Society of 
Civil Engineers, in the Office of the Librarian of Congress, at Washington. 



Note.— This Society is not responsible, as a body, for the facts and opinions advanced in 

any of its publications. 



CONTENTS 



PAPERS 

No. PAGE 

11 IH CAISSON DISEASE AND ITS PREVENTION. 

By Henry Japp 1 

Discussion: 

By J. S. Haldane 24 

Frederick Tj. Kbays 26 

Albert J. Loomis 29 

William L. Saunders 30 

Walton I. Aims 32 

Seward Erdman 32 

T. Kennard Thomson 33 

Henry Japp 34 

1119 COPYRIGHT IN DRAWINGS OF A TECHNICAL CHARACTER. 
By D. A. Usina 38 

1120 THE SIXTH STREET VIADUCT, KANSAS CITV. 
By E. E. Howard 42 

Discussiou: 

By Daniel Bontecou 95 

Victor H. Cochrane 96 

O. E. Mogensen 98 

N. T. Blackburn 100 

George H. Pegram 101 

E. E. Howard. 101 

^ 1121 SOME EXTENSIVE RAILROAD SURVEYS, AND THEIR COST PEI^ MILE. 

V By W. S. McFetridge 105 

^ Discussion: 

By C LINTON S. BiSSELL 131 

■^ F. Lavis 133 

0) E. W. Lewis 142 

George L. Dillman , ; 143 

,..- W. S. McFetridge 144 



% 



g: 



P 



O 1122 COMPUTATION OF STRESSES IN OPEN=WEBBED ARCHES WITHOUT 
^ HINGES. 

By C. W. Hudson 145 

Ml23 HYDRO=ELECTRIC POWER IN CANADA. 

By Cecil B. Smith 154 

Discussion: 

By H. HoLGATE 194 

v^ E. J. Beugler 199 

(J) Cecil B. Smith 200 



\3<saG> 



IV 

No. PAGE 

1134 TESTS OF BUILT=UP STEEL AND WROUOHT=IRON COMPRESSION 
PIECES. 

By Arthur N. Talbot and Herbert F. Moore 202 

Discussion: 

By John C. Moses 246 

Arthur N. Talbot and Herbert F. Moore 247 



1135 IMPURITIES IN SAND FOR CONCRETE: AN INFORMAL DISCUSSION. 

By Sanford E. Thomp.son 250 

R. W. Lesley 258 

C. P. Howard 260 

Richard L. Humphrey 260 

G. S. Davison 262 

S. Whinery 363 

Charles M. Mills 363 

T. F. Richardso.n 365 

E. G. Haines 267 

Sanford E. Thompson 371 



1136 FIRE.RESISTANT CONSTRUCTION OF BUILDINGS: AN INFORMAL 
DISCUSSION. 

By Herbert M. Wilson 274 

R. W. Lesley 290 

Richard L. Humphrey 391 

S. Whinery 393 



1137 THE SEWER SYSTEM OF SAN FRANCISCO, AND A SOLUTION OF THE 
STORM=WATER FLOW PROBLEM. 

By C. E. Orunsky 394 

Discussion: 

By W. C. Hammatt a8& 

Robert G. Dieck 385 

Charles E. Gregory 390 

E. Kuichlino 395 

Kenneth Allen 400 

Walter N. Frick.stad 405 

C. E. Grunsky 412 



1128 WALNUT LANE BRIDGE, PHILADELPHIA. 

By George S. Webster and Henry H, Quimby 423 



1129 HAS EXPERIENCE DEMONSTRATED THAT THE OILING OF ROADS IS 
THE MOST SATISFACTORY OR ECONOMICAL METHOD OF PRE= 
VENTING DUST AND PRESERVING THE ROAD SURFACE? AN INFOR- 
MAL DISCUSSION. 

By S. Whinery 462 

G. S. Davison 466 



No. PAGE 

1130 CONCRETE PILES. 

By Howard J. Cole 467 

Discussion : 

By Morton L. Tower 488 

L. J. Mensch 489 

Charles H. Higgins 491 

Maxwell M. Upson 494 

Thomas C. Desmond 498 

R. D. Coombs 502 

Homer A. Rkid 503 

KuNLEY Abbott 507 

Howard J. Cole 509 



MEMOIRS OF DECEASED MEMBERS 



George Adgate, M. Am. Soc. C. E 514 

William Price Craighill, Past-President, Am. Soc. C. E 517 

Charles Hermany, Past-President, Am. Soc. C. E 525 

Cabell Breckinridge, M. Am. Soc. C. E 539 

Walter Lorton Dusenberry, Jun. Am. Soc. C. E 533 



PLATES 



plate paper 

I. Portions of Profile and Plan. Sixth Street Viaduct, Kansas City. . . 1120 

n. Views of Sixth Street Viaduct aud Kaw River Bridge 1120 

HI. Concrete Piles, aud Pile Drivers 1120 

IV. Pedestal, and Tower Span, etc.. Sixth Street Viaduct 1120 

V. Detail of Pier 2, Kaw River Bridge 1120 

VI. Details of Columns and Bracing, Sixtli Street Viaduct 1120 

VII. Gallows Frame for Erection of Long Girders; and Concrete Floor, 

Sixth Street Viaduct 1120 

VIII. Concrete Plant for Laying Floor of Sixth Street Viaduct; and 

F'inished Floor of Approaches 1120 

IX. Map Showing Lines of Little Kanawha R. R., in West Virginia 1121 

X. Map of the Dominion of Canada 1123 

XI. Map of Part of the Province of Quebec 1123 

XII. Application of Instruments for Measuring Stresses, etc., in Com- 

pression Pieces; and Test Train on White Heath Bridge 1124 

XIII. Plan of Part of San Francisco, Cal., Showing North Point Main 

Relief Outlets 1 127 

XIV. Walnut Lane Bridge, Philadelphia 1128 

XV. General Plan and Elevation of Walnut Lane Bridge and Ap- 

proaches 1128 

XVI. Falsework and Main Arch Centering, Walnut Lane Bridge 1128 

XVII. Plan and Elevation, and Arch Details, Walnut Lane Bridge 1128 

XVIII. Construction of Arches, Walnut Lane Bridge 1128 

XIX. Cross-Sections, Walnut Lane Bridge • 1128 

XX. Moving Centering, and View of Voussoirs, Walnut Lane Bridge. . . 1128 

XXI. East Abutment, Walnut Lane Bridge 1128 

XXII. Floor Details, Walnut Lane Bridge 1128 

XXIII. Brackets and Piers, Walnut Lane Bridge 1128 



PAGE 
45 

53 

59 
61 
67 
71 

77 

81 
109 
157 
199 

209 

371 

427 

429 
431 
433 
435 
437 
441 
445 
447 
449 



VI 



PLATE PAPER 

XXIV. End of Centering, and View Between Ribs, Walnut Lane Bridge. . 1128 

XXV. Floor System and Deck View, Walnut Lane Bridge 1128 

XXVI. Views of Walnut Lane Bridge 1128 

XXVII. Centering. Walnut Lane Bridge 1128 

XXVin. Stress Diagram, Walnut Lane Bridge 1128 

XXIX. Approach Spans, Walnut Lane Bridge 1128 

XXX. Timber Piles Damaged in Driving, and Reinforced Concrete Piles, 

Type A, Made-in-Plaee Class 1130 

XXXI. Reinforced Concrete Piles, Type A, Made-in-Place Class, and 

Moulded Piles 1 130 

XXXII. Vertically-Moulded Piles of Reinforced Concrete for Boardwalk. 

Atlantic City, N. J 1 130 

XXXIII. Reinforced Concrete Piles, Type B, Made-in-Place Class 1130 

XXXIV. Reinforced Concrete Piles, Type B, Made-in-Place Class, and 

Type 3, Moulded Class 1130 

XXXV. Reinforced Concrete Piles, Type B, Made-in-Place Class, and 

Type 3, Moulded Class 1 ] 30 

XXXVI. Reinforced Concrete Piles, Type 3, Moulded Class 1130 

XXXVII. Views of Concrete Pile with Enlarged Base 1130 

XXXVIII. Method of Driving, and Pulling Out Concrete Pile Casing 1130 



PAGE 

451 
453 
455 
457 
459 
461 

409 

473 

477 
479 

481 

483 
485 
507 
509 



Portrait of the late William Price Craighill, Past-President, Am. Soc. C. E.. 516 

Portrait of the late Charles Hermany, Past-President, Am. Soc. C. E 534 



AMERICAN SOCIETY OF CIVIL ENGINEERS 

iNSTITUTED 1852 



TRANSACTIONS 



Paper No. 1118 

CAISSON DISEASE AND ITS PREVENTION.* 

By Henry Japp, M. Am. Soc. C. E. 



With Discussion by Messrs. J. S. Haldane, Frederick L. Keays, 

Albert J. Loomis, William L. Saunders, Walton I. 

Aims, Seward Erdman, T. Kennard Thomson, 

AND Henry Japp. 



The prevention of caisson disease has not received as much atten- 
tion as its cure, the general opinion being that it cannot be prevented. 
In forming an idea as to the possibility of its prevention, some- 
thing may be learned from Nature's method of carrying out work by 
the aid of living organisms under air pressure. 

First let us inquire whether Nature has a sufficient range of air 
pressure to make it necessary for a capacity to sustain variations with- 
out injury. The highest altitude on this earth is Mount Everest, 29 000 
ft. above the level of the sea, and the deepest mine is Tamarack No. 3 
Shaft, of the Calumet and Hecla ]\lanes, 4 407 ft. below sea level. 

The curve on Fig. 1 shows the absolute air pressures at various 
altitudes between these extremes. The total difference of 12J lb. per 
sq. in., or from 9| lb. below atmospheric pressure at sea level (viz., 
14.7 lb.) up to 2f lb. above normal, is considerable. 

Mount Everest has never been ascended, but an elevation of 
20 000 ft., has been reached, and men can live in comfort at an altitude 
of 7 000 ft., as in Mexico City, which has an air pressure of 3^ lb. 

* Presented at the meeting of June 2lI, 1909. 



2 CAISSON DISEASE AND ITS PREVENTION 

per sq. in. less than at the coast. Such a change of pressure is quite 
noticeable to those whose Eustachian tubes are closed, and, although the 
range of pressure is insufficient to produce caisson disease, yet travelers 
have complained of discomfort, not unlike a very mild attack, after a 
rapid journey by rail from Vera Cruz to Mexico City. 

Eagles are known to go to great heights beyond the range of 
human vision, fish are capable of going to great depths in water, and 
beasts of prey roam from the mountain side to the valleys far below. 
With each descent these creatures have undergone compression, and 
with each ascent decompression. 

It will be noted that Nature has no time limit for immersion under 
increased pressure, and the application of pressure is a much speedier 
and easier operation than that of decompression; and, while the com- 
pression is down-hill and attained with little eifort, the decompression 
can only be attained by the continuous effort of climbing, and, in 
long ascents, with frequent rests or by stages. This is surely a forecast 
of the stage-decompression theory. 

From Nature then, our best teacher, we learn that the time taken 
to enter the air-chamber may be short, the time spent in the air- 
chamber may be long, and the rate of coming out must be slow and 
accompanied by exercise. 

In early compressed-air work engineers were confronted with cais- 
son disease, and have long been striving for a remedy. Men of observa- 
tion noted first that lock tenders who made frequent entry to the air- 
lock for short visits were much more immune from the malady than 
workers who stayed for a full shift, and from this argued that shorten- 
ing the time in the air-chamber would be a safeguard, and, as higher 
pressures were used, the hours continued to be shortened until it was 
thought that half an hour at a maximum of 50 lb. gauge pressure 
was the limit of human endurance. 

As the number of compressed-air works increased, the workmen 
gained in experience, and it was noted that workmen of experience 
who suffered from the painful forms of caisson disease relieved the 
pain by re-entering the air-chamber. 

About thirty years ago Paul Bert advanced the theory that caisson 
disease was caused by the nitrogen of the air being dissolved in the 
blood. He advocated slow decompression, suggesting 30 min. for 2 to 3 
atmospheres (15 to 30 lb. gauge pressure) and 60 min. for 3 to 4 



CAISSON DISEASE AND ITS PREVENTION 



140 000 
120 000 
100 000 
80 000 
60 000 
40 000 






DIAGRAM OF ELEVATION AND PRESSURE 
OF ATMOSPHERE 

Plotted to the annrnximnte fnrmiils) 




T7i„,,„+; PAoir... T._. Higher Barometer reading 

Ele^ ation 60 346 x Log . ^^^^^. Barometer reading 


















































20 000 


— Mt. Everest— 
Mt. McKinley 


,29 000 ft. 
\20 461 ft. 










16 000 






\ 
















\ 










8 000 






\ 


\ » «„„ ,.. 








4 nnn 


— Mexico'eity— 








\ 


















"\^ 




Sea Level 
















^""--.^ —4 407 


ft. 






Tamarack 


fo.3 


Sliaft Calu 


met 


& Heela 




~~^-^^ 






















^"---^.^ 




















~~- 


__ 



10 



15 
Pounds 



20 



2.5 



Pig. 1. 



4 CAISSON DISEASE AND ITS PREVENTION 

atmospheres (30 to 45 lb. gauge pressure). Since his time many 
writers and experimenters have followed in his footsteps, but it was 
not until Dr. Smith, of New York, suggested that a recompression 
chamber would act as a remedy, and E. W. Moir, M. Am. Soc. C. E., 
of London, without knowledge of Dr. Smith's prior suggestion, actually 
built a medical air-lock, or recompression chamber, in 1890, on the 
old Hudson Tunnel, New York, that any reliable cure for the trouble 
was discovered. Fig. 2 shows the medical air-lock designed by Mr. 
Moir, of which six were used on the East River tunnels of the 
Pennsylvania Tunnel and Terminal Railroad. 

It was then found that if a man suffering from caisson disease in 
any of its many forms, varying from unconsciousness and paralysis to 
an acute pain in the limbs, was quickly recompressed and allowed to 
decompress slowly, in many cases a cure resulted, but not always. 

This went to prove that the disease was a mechanical one, and was 
caused by the dissolved air in the blood and tissues becoming liberated 
in the body in the form of expanding bubbles which tore the tissues, 
caused pressure on the brain, injured the spinal cord, or frothed up 
the blood and stopped the circulation and heart action. 

Much has been done since the introduction of the medical air-lock. 
The percentage of carbon dioxide in the air of the working chamber 
has been kept within safe limits, moisture and oil have been extracted 
from the air before delivery to the chamber, and the workmen have 
been well cared for; but still caisson disease claims its victims, and 
engineers consider this inevitable. 

The erratic manner of the disease puzzles doctors and engineers 
alike. Men who have worked for months in high pressures for the 
regulation shift without suffering have been suddenly attacked and 
died; others, working only half a shift, have been paralyzed; while 
exceptions have worked for 12 hours at a time and have suffered only 
slight pains which have quickly passed away. 

The writer, as Managing Engineer for S. Pearson and Son, Inc., 
on the East River Tunnels contract for the Pennsylvania Tunnel and 
Terminal Railroad Company, New York, with as many as ten tunnel 
headings under compressed air simultaneously, has had many oppor- 
tunities of studying this strange disease. 

When these tunnels were commenced, in 1904, the knowledge of 
compressed air at the disposal of the contractor was quite extensive. 



CAISSON DISEASE AND ITS PREVENTION 




6 CAISSON DISEASE AND ITS PREVENTION 

and included the long experience of Alfred Noble, Past-President, 
Am. Soc. C. E., on caisson work in the United States, and Mr. E. W. 
Moir's experience on the Forth Bridge caissons, the old Hudson Tunnel, 
and the Blackwall caissons and tunnel. 

In the light of past experience, the general conduct of the work 
was framed on a few established rules which when condensed amount 
to the following: No workman was allowed to enter the air-chamber 
without a physical examination by the qualified medical officer of the 
contractors. Sound physique was the chief requirement. The men 
were cautioned not to enter the air on an empty stomach, to wear warm 
clothing on coming out, and to drink hot coffee. 

The time worked in the air-chamber was limited to 8 hours with 
half an hour off for lunch, up to 32 lb. gauge pressure, and two spells 
of 3 hours each with 3 hours rest between for pressures from 32 to 
42 lb., and two spells of 2 hours each for pressures greater than 42 lb, 
with 4 hours rest between, with no limitation as to decompression. 
Two medical air-locks were installed on each side of the river, well- 
warmed dressing rooms were provided for the workmen, and there were 
covered gangways for access to the shafts. 

The air was cooled before delivery to the tunnels, and samples 
were taken in the tunnels by Mr. Noble's engineers and analyzed 
daily. The air was regulated so that the carbon dioxide did not exceed 
10 parts in 10 000, and the tunnels were kept in a sanitary condition. 

Owing to the grade of the tunnels being so deep on the Manhattan 
side of the river, the air pressure very quickly rose to 36 lb. 

Practically no cases of bends occurred until the pressure reached 
29 lb., and then, within a few days of each other, two men died. These 
men had entered the air-chamber without being passed by the doctor. 
Then it became necessary to post outside each air-chamber a guard 
whose duty it was to keep out of the tunnel men who had no doctor's 
pass. At this time, reliance could not be placed on the tunnel foremen, 
as they were likely to be absent from work, and new men had to be 
selected each day. 

Eor many months after the work started, while the men were 
being seasoned, the tunnel gangs and foremen were in a state of 
change, owing to the difficulty of getting good men and the frequent 
absences due to caisson disease, and it was a long time before an 
efficient organization was built up. 



CAISSON DISEASE AND ITS PREVENTION 7 

As the tunnels were driven deeper beneath the East River the 
pressure quickly rose, and ultimately reached 36 lb., gauge pressure, 
with only one set of air-locks in operation; but even the change at 
32 lb. from one shift of 8 hours to two shifts of 3 hours each gave 
no relief, and cases of bends, sometimes fatal, continued all the time. 

It was not long after 27 lb. was reached that the more sensitive 
members of the staff found that it did not pay to come out quickly, 
and at 30 lb. pressure it became a custom to take about i min. for 
each pound. After one or two additional fatal cases occurred, it was 
decided to limit the workmen to approximately the same rate of decom- 
pression, or actually 15 min. for 35 lb. pressure. 

Many of the men complained that taking so long to decompress 
gave them caisson disease, and it was difficult to compel them to take 
long enough. The guards at the entrance to the tunnels had now to 
record the time taken to decompress, and, as the workmen frequently 
used the lower muck locks as well as the upper man lock, it was 
impossible to tell when decompression commenced owing to the noise 
of exhausting air. Therefore it became necessary to run a small ^-in. 
pipe from the exhaust pipe of each lock to the cabin in which the 
guard was stationed, a whistle was attached, and a small ball of 
cotton was suspended by a light string over each pipe. The clerk, 
noting when the ball was puffed off for each lock, booked the workmen 
off as they left the lock. The material locks were fitted with inner 
material valves and, in addition, man valves of smaller size, and a 
pressure gauge and a clock were fixed in each air-lock. 

As the guard had already booked the men as they entered, he could 
tell if any were exceeding the regular working shift, and his record 
was valuable for checking the time-keepers. This rough method of 
checking the duration of the decompression was quite good enough for 
the purpose. An attempt to improve it was made when a 12-in. 
Crosby recording gauge was installed on one air-lock, but it was in- 
effectual because the air-lock was often sent out or decompressed with 
no one inside, and this complicated the record, which was much too 
small, and involved considerable trouble in locating the record of 
decompression for individual gangs. No doubt a suitable recording 
instrument could be devised for this special purpose. 

The effect of lengthening the decompression period to 15 min. 
reduced the number of cases of bends, and no doubt prevented many 



8 CAISSON DISEASE AND ITS PREVENTION 

fatal ones, but they still occurred. As the tunnels had many months 
to run at high air pressures, the question was : What else can be done 
to prevent them ? 

The writer, on coming out of the tunnel with the workmen, observed 
that the rate of decompressing was most irregular. One lock tender 
would allow the air to escape slowly until the 15 min. was almost 
exhausted, and then, by opening the valve, would let off the remaining 
pressure very quickly. Others would reverse the process, and exhaust 
quickly, and then keep the men under 2 or 3 lb. until the time expired. 
To avoid this, a simple decompressing valve was designed (Fig. 3) 
which gave a uniform decompression from 35 lb. to atmosphere in 
15 min. A somewhat similar one (Fig. 4) was designed for the 
medical air-lock, for 1 hour decompression for 35 lb., with an auto- 
matic ventilator attached. 

These valves certainly improved conditions, but still fatal cases 
occurred. After the first valve was under operation, the writer's 
attention was called to "Modern Tunnel Practice," by D. McN. 
Stauffer, M. Am. Soc. C. E., wherein a description is given of a 
needle decompression valve used on the air-locks at the Kiel Dry 
Dock Works, in Germany. A similar valve was also used for com- 
pression. On entering the lock the air was admitted at the rate of 
1.5 lb. per min., and was decompressed at the rate of | lb. per min., 
or, for 35 lb. gauge pressure, 23 min, for entering the air and 46 min. 
for leaving. This needle-valve was often frozen up, but otherwise 
worked well. 

Such speeds seemed altogether too slow, and Mr. Stauffer in his 
book states that "these rates would be deemed excessively slow in 
American caisson practice." No date is given for the Kiel Dry Dock 
work, but presumably it was under way in 1904. 

It was not thought advisable to increase the time of decompression 
at that time, but preliminary tests under air pressure in the tunnels 
for 14 hours were then tried for green men, followed by a second 
medical examination after decompressing in 15 min. A few men were 
eliminated by this test, and one case of permanent paralysis resulted 
from the test in 34 lb. gauge pressure. Fresh starters were made to 
stay in the tunnel for one-quarter of a shift for the first half, and if 
no caisson disease followed they were allowed to work for the second 
half. This proved a good safeguard. 



CAISSON DISEASE AND ITS PREVENTION 




10 CAISSON DISEASE AND ITS PREVENTION 

On November 8th, 1906, a second bulkhead was put in operation 
in one of the tunnels, and the pressure between the two bulkheads was 
reduced to 15 lb. gauge. The number of cases of disease was very 
small for that tunnel, and as soon as possible additional bulkheads 
were placed in all four tunnels. The result was to have been expected 
from the experience in other tunnels where two bulkheads were used; 
the exercise of the long walk between bulkheads, at low pressure, seemed 
to assist in driving off the bubbles of air from the blood. 

The workmen were allowed to decompress from 35 to 15 lb. as 
quickly as they pleased. They then walked for 500 ft. along the 
tunnel under 15 lb., taking at least 5 min., and then decompressed 
from 15 lb. to atmosphere in 10 min., so that in all about 16 min. 
were occupied in decompressing. Wlien the inner pressure was less 
than 32 lb. an 8-hour shift was worked, with i-hour interval for lunch, 
between bulkheads in low air pressure. 

Just when it looked as if the double bulkheads with stage decom- 
pression had eliminated fatal and severe cases of caisson disease, two 
deaths occurred in physically perfect subjects. 

In order to discover, if possible, the connection between the cases 
of disease and other things, charts were plotted by the medical staff 
for some months, showing the rise and fall of air pressure, hours 
worked, humidity, temperature, percentage of carbon dioxide, and 
number of green men in the tunnel, along with barometer readings, 
condition of weather, direction of wind, and number of cases of bends. 
The results were not very encouraging, but it was noted that the 
number of green men, the height of pressure in the tunnel and the 
number of cases of bends varied together. 

The percentage of cases in air pressure of 31^ lb. for 8-hour shifts 
was no more than the percentage in 324 lb. for two 3-hour shifts — 
in fact, it was, if anything, less for the longer shift. The decrease in 
length of shift added one extra gang of men, and probably many of 
these men being green accounted for this. 

In November, 1906, Mr. Moir became acquainted in part with 
Dr. Haldane's work of investigation for the British Admiralty on 
"Deep Sea Diving," and ordered the decompressing valves changed so 
that the decompression was accelerated at the commencement of the 
operation, and slowed off toward the end. In October, 1906, the 
writer, endeavoring to find out the effect of sudden decompression. 



CAISSON DISEASE AND ITS PREVENTION 



11 




O 
m 
O 
O 

■V 

:o 
m 

CO 
CO 



a 



12 CAISSON DISEASE AND ITS PREVENTION 

fitted an ordinary ^lass siphon bottle with a valve and pressure gauge. 
This bottle was partly filled with water, and put under 36 lb. air 
pressure for 12 hours. On suddenly releasing the pressure, no effer- 
vescence whatever took place, and only a very few minute bubbles were 
visible. The same thing was tried with the bottle partly filled with 
bullock's blood, with like results. Arguing that the placid surface 
of blood in a stationary vessel offered no such opportunity for dissolving 
air as occurs in the lung surface with the blood circulating, the siphon 
was again charged with 36 lb. pressure, but was rotated for 24 hours 
under this pressure. The same result as before was observed on 
suddenly releasing the pressure; but, on closing the valve, the pressure 
recorded on the gauge was about 1 lb. per sq. in. in less than 1 hour, 
and in 6 days a pressure of 13 lb. was recorded, showing that the air 
dissolved in the blood came off very slowly. 

If the cork of a bottle of aerated water is drawn quickly, it effer- 
vesces so violently that the water froths out of the neck of the bottle; 
but, if the gas is allowed to escape slowly, only a mild bubbling results. 
This in general is the theory on which slow decompression has been 
advocated in the past. The violence of the escape of carbon dioxide 
gas from aerated water, as compared with the escape of air from water 
released from pressure, is due to the greater solubility of carbon 
dioxide than air. 

It was difficult to know what else could be done at this time in 
the way of eliminating caisson disease on the East River tunnels, and 
the cases of disease were certainly less frequent and less severe, 
although at long intervals fatal cases still occurred. The workmen, 
by this time, had become seasoned, and more seasoned men were now 
available from other works. 

Dr. Leonard Hill read a paper on deep sea diving and caisson work, 
on September 26th, 1907, and exhibited the web of a living frog under 
a pressure of 20 atmospheres. On suddenly decompressing, bubbles 
were plainly visible in the blood vessels, multiplying quickly as they 
traveled along until the circvilation was choked; on recompression the 
bubbles disappeared and circulation was resumed. He stated that he 
and Mr. Greenwood had been under an air pressure of 92 lb., and had 
suffered no ill effects. 

Dr. J. S. Haldane, on November 29th, 1907, read a most enlighten- 
ing paper before the Society of Arts, in London, in which he gave 



CAISSON DISEASE AND ITS PREVENTION 13 

the results of his investigations for the Britisli Admiralty on " Deep Sea 
Diving," in which he was assisted by Dr. Boycott and Lieutenant Daniant. 

He set forth very clearly the question of the solution of gases by 
the blood, fat, and tissues of the body, and attempted to determine the 
time taken to saturate the blood with the nitrogen of the air. He 
proved concltisively that the quantity of carbon dioxide adjusts itself 
to a constant percentage inversely proportional to the pressure in the 
alveoli, or air cells. 

In atmospheric pressure, the percentage of carbon dioxide in the 
alveolar or expired air is 5.6%, and at a pressure of 2 atmospheres, 
absolute, this is reduced to 2.8% ; so that the question of the percentage 
of carbon dioxide in the air of the working chamber is not important 
unless it approaches the percentage of carbon dioxide in the air cells 
of the lungs. 

For instance, if the air-chamber is under an air pressure of 30 lb., 

or 3 atmospheres absolute, the percentage of carbon dioxide in the air 

cells is 5.6 divided by 3, or 1.86% ; and, if the percentage of carbon 

dioxide in the air-chamber does not exceed 1%, no ill effects will arise. 

This is ten times as much as is generally specified, namely, 10 parts in 

10 000, and greatly reduces the amount of compressed air necessary 

per man per hour, which can be calculated approximately from the 

following simple formula: 

.,,.., , 80 cu. ft. 

Cubic leet per man per hour = 

Percentage COg permitted 

Thus, if 0.04% is the CO2 in the atmosphere, and the percentage 

in the tunnel is allowed to go up to 0.10%, the air required per man per 

hour = --— = 1 333 cu. ft. 
0.06 

Dr. Haldane has completely removed all difficulty of respiration 
and discomfort of divers at great depths through the application of this 
theory regarding the percentage of CO, by supplying the diver with 
the same volume of compressed air per minute at different levels 
instead of the same volume of free air per minute. 

The question of ventilating compressed-air tunnels seldom arises, 
as the volume of air generally far exceeds the amount required for 
purity, on account of the tremendous leakage at the face, so that this 
is not nearly as important as the solubility of air by the blood and 
tissues of the body. 



14 CAISSON DISEASE AND ITS PREVENTION 

After demonstrating that the COg is dealt with automatically, he 
shows that nearly all the oxygen dissolved by the blood is taken up 
chemically by the hemoglobin, but the nitrogen remaining is incapable 
of combination with the system, and is dissolved by the blood and 
tissues. These, therefore, become saturated with nitrogen during 
immersion in compressed air. 

If saturation takes place quickly it will not matter how long a 
workman stays under air pressure; but, if slowly, then shortening of 
the working shift will be a safeguard. 

The rate of saturation is the same as the rate of desaturation if 
compression and decompression are instantaneous; so that, by altering 
the time for decompression, it is possible to determine by experiment 
the rate of saturation and desaturation. 

Dr. Haldane experimented with men and goats at high air pressures, 
with varying lengths of stay under air pressure, and with different 
speeds of decompression. From these he concluded that certain parts 
of the body, where the circulation is rapid and the number of blood 
vessels high for the mass of the part supplied, would be half saturated 
or desaturated in 5 min., while other parts, with slight circulation for 
the mass, would require 75 min. for 50% saturation, especially the 
fatty parts, as fat is found to dissolve about six times as much 
nitrogen as blood. 

Fig. 5 shows the rate of saturation, on the basis of half saturation 
in various times from 5 to 75 min., and curves of desaturation, if the 
decompression is instantaneous after varying times of immersion from 
i hour up to 8 hours. 

It will be seen that 90% saturation takes place in the slowest parts 
of the body in about 4 hours, and complete saturation for the quickest 
parts in about 40 min., so that after 40 min. immersion there is 
danger in too rapid decompression. 

Dr. Haldane has noted that no serious cases of caisson disease have 
occurred for rapid decompression at a working pressure of 19 lb. gauge 
pressure, or 2.3 atmospheres, absolute. Therefore, he concludes that 
it is always safe to decompress rapidly to half the absolute pressure, or 
even to the absolute pressure divided by 2.3, without bubbles being 
liberated. Further, if the decompression for the remaining pressure 
is continued at such a rate as will keep pace with the rate of desatura- 
tion, then, the absolute pressure of nitrogen in the blood never exceed- 



CAISSON DISEASE AND ITS PREVENTION 



15 







'^^S 


==:::; 








— 


- 


n 












Jl 


_r 


Lili 


fs;, 


Eii 


^ 


-. — 1 


U ;>' 


i' -5 




^ 




^""^^- 


>^S>^^ 




-- 


~^>^ 








^ 






— .j»go"'| 


'"lf( 


■^. 








/ i?<c 


^]^ 
~ 


r- 




' 




-^ 


:^^ 


~ 


"■ 


1 "■ 


^ 


\ 


- 


^ 






7 


'A 




■r^ 


^ 


^ 




-- 


^ 




■^ 


-^„ 




\ 


( 






/^ 


\$i 


Ni" 




/■ 


/ 




> 


^<^ 




s.. 


^ 


^ 


^" 


id 


<; 


\^.-^ 


N 


\ 


^ 








^«- 






/ 




/J 


*i 


^ 






■^ 




Vj 




\ 






. 


K° 


s 


Iv 


^ 






A' 






f' •* 














— 


, .7^ 


^ 


^ 


S; 


-^ 


5- 


-^^s^T"^ 




\, 


t« 






li 




■^ 


H 


^ 








-^ 


^ 


^ 
« 


— 


\ 


- 




\ 


\ 







/^ 


r 






^. 




^ 


i»» 










E 




3 


<f^^ 


v^? 




v\ 








g:;; 






I& 
/» 


} 


^'/^^ 




y 












f:; 


^ 


r 












s 


4W\ 


\ 


\' 


CO 






/ 


^/ . 


/^ 








^ 
y" 


^ 


s*^ 




















\ ' 


v 





\ 




z 


/ 




/P 




^/ 


^ 
























^ 


3: 


^ 


^ 


H 


T 


VI 




3 


/ 


//■J 


<^ 


% 














[t: 


^^^ 




s 


^ 


^ 








\ 


\ 


\\ 


u 


*• 


5 

to 




/ 


/ 

; 


// 


'/ 














^ 




^ 


















\ 




u 








/ 


/ 










y 


y 


^ 
















. - 


- 


^ 


^^ 


^ 


y 1 


Tl 










/ 






/, 


K 














^^ 




J 


i^ 


^ 


^ 










\ 


\\ 














/ 




















^ 


















V 


11 




O 

>- 

o 
:a 
i» 

- 3> S 

-n 
H c/) 

" ^ =c 
^ O 

<:$ 

--^^ 
n ^ 'XI 

S "D m 
. -0 m 
PD :o O 
m ^ -n 

CO o „ 
-rn t« m 

O 3> 

- i* -n H 

- O) 

S H 

50 
CO ^ 


z -n 

ro 
r- 







/ 


/ 










y 


-^ 


.^ 














^ 


g 


^ 


*^ 


-^ 


i 


1 






/ 






/' 


y 


^ 














:; 




^ 








1 1 


O 05 

c 








/S 














::; 


^ 






















I 






/^ 


[p- 










^ 


^ 


^ 


















^=^ 


^ 




V 






/ 








/ 


^ 


y 












^ 




^ 


^ 


^ 










1 










/i. 


/ 
> 












^ 


^ 




■-' 


^ 
















1 








/^ 


^^^ 












^ 


















^ 


_^^ 


^^ 




1 






/ 










^ 


y 












- 




:?= 


^ 


- 










CC 








/ 


l^ 
Q 














^ 


/ 


" 


^ 
























V 












^ 




^ 
















— -? 


^ 


^ 


"^ 






/ 










y 


/ 












^ 


- 


^ 




t5» 


j^' 


,-^o'\ 
















/ 


^r 














^ 




V 


^5. 


A>' 


^X^V^ 




















/^^. 




- 


■^ 






/ 




3^-^"^ 




























/ 




/ 


/ 




y 


^ 




>>>^ 


































/ 


/ 


/ 






































S 






/ 


/ 




3 


5 







i5 





-0 


3 


£3 


c^ 




5 






iS 










/ 






s 


i«- 




*i 






e 


i 




g 




•^ 











P 












1 


-b^. 




o"^ 


c 


s. 


sa 


rT 


5: 


^ 










-#. 
















s 






8 


1-1 




^ 


w*- 


£3 




« 






8 




















a 




F 




F 




p- 


cr 


cr 


p- 


p- 




p- 






p 





16 



CAiiSSON DISEASE AND ITS PREVENTION 



ing 2.3 times the absolute pressure in the air-lock during decompression 
or after returning to the atmosphere, no symptoms will occur. 

Table 1 shows the rates of decompression advocated by Dr. Haldane 
for caisson workers : 

TABLE 1. — Dr. Haldane's Rate of Decompression in Caisson and 

Tunnel Works. 



Working pressure, in 

pounds per square 

inch. 


Number of MiNtJTES for Each Pound of Decompression 

AFTER THE FlRST RaPID StAGE. 


After first 3 hours' 
exposure. 


After second or third 

3 hours' exposure 

showing an interval 

for a meal. 


After 6 hours or 

more of continuous 

exposure. 


18-20 
21-24 
25-29 
30-34 
35-39 
40-45 


2 
3 
5 
6 

7 

7 


3 

5 

7 
7 
8 
8 


5 
7 
8 
9 
9 
9 



If the pressure is 40 lb., gauge, decompress rapidly from 40 lb. to 

40-1-15 

— IT = 27i lb., absolute, or 12 J lb., gauge, in 3 min., and then 

take 7 min. for each remaining pound, viz., 12^ lb., or 87^ min. plus 
3 min. = 90J min., or \\ hours in all for a 3-hour immersion; or, 
for a 6-hour immersion, 115 min. in all. 

The upper diagram on Fig. 6 shows this graphically. 

It will be seen that, if this rule is followed, the condition of the 
absolute pressure, being half the pressure in the blood, is obtained. 
If uniform decompression for the same time were adopted, the decom- 
pression would be too slow for the earlier part, would very much 
retard the desaturation, and would give a greater tension in the blood 
than 19 lb., on coming out of the air pressure. 

It is obvious that while the air-lock has any pressure in it the 
desaturation takes place only in relation to that pressure, and, if one 
adopts Dr. Haldane's idea of keeping a portion of the tunnel at half 
the absolute pressure for a dressing room to act as a purgatorial 
chamber, where the men moy wash and change while desaturating down 
to 19 lb., we must make a new curve for desaturation and increase 
the time of decompressing, as will be seen by the second diagram 
on Fig. 6. 



CAISSON DISEASE AND ITS PREVENTION 



17 




W}4 Minutes 



40 



30 



20 



10 



\ 


\ 


\ 


% 


% 
















DECOMPRESSING FROM 40 LB.GAUGE PRESSURE 

TO ATMOSPHERE AFTER 3 HOURS IMMERSION. 

DR.HALDANE'S PURGATORIAL CHAMBER METHOD 












^ 


^^ 


^^ 
















Pres. 


uie ii 


I PlUj 


ratorial Cb< 


miber 


.491b. 
















1 






A 


u 


UOSJll 


lere 








\ 






^ 



40 



30 



!20 



10 



r 


'^ 


\ 


\ 
















DECOMPRESSING FROM 40 LB.GAUGE PRESSURE 

TO ATMOSPHERE AFTER 3 HOURS IMMERSION. 

METHOD USED IN EAST RIVER TUNNELS OF THE 

PENNSYLVANIA TUNNEL AND TERMINAL RAILROAD 






\ 


\ 


[\ 


b. 
















V 


\ 
















At 


111 


osph€ 


^ 

re 


\l8M 


inutes 














51015 30 45 1 
Minutes Hour 



Fig. 6. 



18 CAISSON DISEASE AND ITS PREVENTION 

From this it appears that the tension in the blood does not reach 

19 lb. until a period of 123 min. has been passed in the purgatorial 
chamber after 3 hours' immersion. 

It will be noticed that Dr. Haldane bases his theory of decom- 
pression on the fact that no cases of caisson disease are chronicled for 
men working in gauge pressures up to 19 lb., or 34 lb., absolute, and 
if it is safe to decompress suddenly from 34 lb., absolute, to 15 lb., 
absolute, it is safe to decompress suddenly from 50 lb., absolute, to 
22 lb., absolute, or from 35 lb., gauge pressure, to 7 lb., gauge pressure. 

The question arises at once: Are the times suggested by Dr. 
Haldane practical? 

Many caissons of small dimensions are sunk under air pressures of 
35 lb. and to cramp men in a small air-lock for 73 min. is out of the 
question, and the fitting of large compartments in small vertical air- 
locks, as used in building foundations in New York City, would be 
difficult. In tunnel work it would be easier, but the present method 
of rapid decompression is so radically different from Dr. Haldane's 
suggestion that it quite appals one to think of taking so long. 

The experience of most compressed-air works is that up to 27 lb. 
gauge pressure there is very little trouble, and, using this as the safe 
limit, the times required to reduce the saturation in the blood to 27 lb. 
is not so excessive for pressures up to 50 lb. 

The horizontal lines on Fig. 5 show the equivalent points for 27 
and 19 lb. in the blood for 50, 45, 40, 35, 30, and 27 lb., gauge pressure. 

It will be noticed that the slowly saturated parts of the body are 
only partly saturated in the first few hours of immersion, and there- 
fore the desaturation is proportionately fast. 

From Fig. 5, Tables 2 and 3 have been compiled; they show the 
risks taken in rapid decompression after long immersion. 

If the decompression is slow, these curves will be flattened out, and 
a longer time will elapse before the saturation of the blood falls to the 
equivalent of 19 lb. or 27 lb., as the ease may be. 

Fig. 7 shows the curves of decompression from 27, 30, and 32 lb., 
gauge, after 8 hours' immersion, and for 32, 35, 40, and 42 lb., gauge, 
after 3 hours' immersion, and 42, 45, and 50 lb., gauge, after 2 hours' 
immersion, on the basis that it is safe to decompress from 27 lb., gauge 
pressure, in 9 min. This reduces the saturation in the blood to the 
equivalent of 25 lb. on reaching atmospheric pressure. 



CAISSON DISEASE AND ITS PREVENTION 



19 




24 25 33 35 37 

20 30 40 

Mluutes 

Fig. 7. 



20 



CAISSON DISEASE AND ITS PREVENTION 



TABLE 2. — Times (in Minutes) Eequired for Pressure in Blood to 
Fall to the Equivalent of 19 Pounds Gauge Pressure with 
Instant Decompression after Various Periods of Immersion in 
Compressed Air. 



No. of hours of 


Gauge Pressures : 


immersion. 


27 1b. 


30 lb. 


351b. 


401b. 


451b. 


501b. 




Minutes. 


Minutes. 


Minutes. 


Minutes. 


Minutes. 


Minutes. 


1^ hour 


4 
7 
10 
14 
23 
28 
33 
35 
37 
38 


5 
10 
16 
21 
30 
39 
44 
47 
48 
50 


8 
15 
24 
32 
45 
.56 
60 
63 
64 
66 


10 
20 
32 
42 
59 
69 
74 
76 
78 
79 


14 
26 
40 
52 

72 
82 
86 
89 
90 
92 


16^ 
33 




1}4 hours 


48 
62 


3 " 


82 


4 " 


93 


5 " 

6 " 


98 
100 


7 " 

8 " 


102 
103 







The times in Table 3, being for instant decompression, must be suitably extended for 
slow decompression. 

TABLE 3. — Times (in Minutes) Eequired for Pressure in Blood to 
Fall to the Equivalent of 27 Pounds Gauge Pressure with 
Instant Decompression after Various Periods of Immersion in 
Compressed Air. 



No. of hours of 


Gauge Pressures: 


immersion. 


271b. 


301b. 


351b. 


40 1b. 


45 1b. 


501b. 




Minutes. 


Minutes. 


Minutes. 


Minutes. 


Minutes. 


Minutes. 


^ hour 




1 

2 
3 
5 
6 
8 
9 
10 
11 


3 
5 
6 
9 
15 
19 
22 
25 
27 
.28 


4 
8 
12 
17 
25 
32 
37 
40 
41 
43 


6 
13 
18 
35 
34 
45 
50 
53 
54 
5F 


8 


1 '• 


16 


15^ hours 


34 


2 '• 


33 


3 ■' 


45 


4 " 


57 


5 " 


61 


6 •' 


64 


7 " 


65 


8 '■ 


66 







The times in Table 3, being for instant decompression, must be suitably extended for 
slow decompression. 

Table 4 is taken from Eig. 7, and in part is confirmed by the 
experience in the East River tunnels as comparatively safe. Caisson 
disease will result, but no fatal or severe case should be experienced 
with physically sound men if these times are adhered to. 



CAlSaON DISEASE AND ITS PliEVENTION 



21 



TABLE 4. — Decompression Table Based on 9 Minutes Being Safe 
FOR 27 Pounds Gauge Pressure. 



Gauge pressure, 


Reduce pressure 


Total time in air- 


Total lime in air- 


Total time in air- 


in 3 minutes 


lock after 8 hours' 


lock after 3 hours' 


lock after 2 hours' 


in pounds. 


to: 


work: 


work: 


work: 


27 


6 lb. 


9 






30 


'>}4 " 


34 






32 


8^2 " 


33 


25 




35 


10 '• 




35 




40 


123^ '• 




48 




42 


is^ '• 




51 


■67. 


45 


15 •' 






42 


50 


171^ " 






48 



Some time after reading Dr. Haldane's paper and studying his 
theory, it became necessary to raise the pressure in the tunnels to 
40 lb., gauge. It was possible to make the workmen pass through three 
sets of air-locks on leaving the tunnel. The inner chamber was kept at 40 
lb., the intermediate chamber at 29 lb., and the outer chamber at 12^ lb. 

The men were ordered to take 5 min. in the iirst lock, 8 min. in 
the second lock, and 15 min. in the third. There was a distance of 
approximately 1 000 ft. between each pair of locks. Walking this 
distance and gathering in the stragglers generally required 10 min. 
to each chamber, so that, in all, 48 min. were taken to decompress from 
40 lb. to atmosphere. No severe or fatal cases resulted, and little time 
was lost by the men through caisson disease, the cases being only slight. 
Under this pressure 330 men were employed for 36 days, working 
3 hours on, 3 hours off, and 3 hours on. It is true that no green men 
were used on this work, as there were plenty of experienced air men 
available at that time. 

The third diagram on Pig. 6 shows the desaturation curve and rate 
of decompression, leaving the saturation of the blood equivalent to 
27 lb. on reaching atmospheric pressure. This result bears out in part 
the periods of decompression of Table 4, and, if a more rapid passage 
had been made through the first and second locks, probably a better 
result would have been attained, as the equivalent pressure of the 
blood would have reached 25 lb. on coming out, as shown by Fig. 7, 
instead of 27 lb. 

In caissons or tunnels with but one lock, it is a difficult problem 
to allow workmen as long a time as Table 4 indicates, as no one can 
enter the air-lock during decompression. One method of overcoming 
this difficulty would be to provide a lock with two small end chambers 



22 



CAISSON DISEASE AND ITS PREVENTION 



and a larger center chamber, four doors in all being necessary. Any- 
one making a short visit to the caisson could pass through without 
disturbing the pressure in the middle decompressing chamber. Fig. 8 
is a diagram of such a lock with connections. 

Some one has suggested, for diving bells and caissons, a detachable 
chamber like a boiler which the men could enter, and thus take as long 
as necessary to decompress. 
AIR-LOCK WITH MIDDLE DECOMPRESSING CHAMBER 



Pressure 



;^t-::.vji:<)- 



Fig. 8. 

The men on the East River tunnels rebelled against 15 min. for 
decompression, but, after putting the responsibility up to the foremen, 
in time they found that it was a safeguard and voluntarily lengthened 
the time to 20 min., and gladly submitted to 48 min. for 40 lb. 

The death rate due to caisson disease was comparatively small, 
averaging ^^ of 1% for the whole of the compressed-air work, and, 
from the experience gained, it would in all probability have been much 
higher if the decompression had not been lengthened. 

The fact that the only recognized cure for caisson disease is recom- 
pression in a medical air-lock followed by slow decompression, is a 
powerful argument in favor of slow decompression, and where it is 
at all possible, in future works, regulated decompression will in all 
probability be adopted. 

In caissons with small air-locks, the volume of air remaining when 
the lock is full of men is very slight, and very rapid decompression 
takes place. The workmen have a good opportunity to become seasoned 
in a caisson, as the pressure begins at 1 or 2 lb. and gradually increases 
day by day as the caisson sinks, and the highest pressures are required 
for but a few days. 



CAISSON DISEASE AND ITS PREVENTION 23 

On the East River tunnels two caissons were sunk to a final pressure 
of 33i lb., and very few cases of caisson disease occurred, none of which 
was fatal, although the decompression was rapid. On the other hand, 
on account of the tunnels on the Manhattan side starting out at a 
high pressure, the men had no chance to get seasoned, and many 
cases occurred, though the time of decompression was regulated, but not 
to such an extent as Table 4 indicates. 

Thus far, practically, no grandfatherly legislation hampers engi- 
neers in the United States in compressed-air work, but such legislation 
is threatened, and engineers should be prepared to guide it. The 
engineer is confronted with conflicting testimony on all sides, and, 
though his experience shows that some men are capable of rapid decom- 
pression without injury, he also knows that many men are injured. 
Because a few men can for a time rashly take the risk, shall he elimi- 
nate good men from compressed-air work by putting them to such a test 
without endeavoring to make it safe for all who are physically sound ? 

Prevention is better than cure, but absolute safety is impracticable, 
and some risk must be taken. In the writer's opinion, the hours of 
labor and time of decompression are matters that should receive the 
attention of engineers. The effect of legislation formulated by men 
ignorant of the conditions might greatly embarrass construction work, 
and, although it is contrary to the custom of this Society to draw up 
rules and regulations, yet it might render valuable aid if a committee 
of engineers, contractors, and doctors could be convened to study this 
subject and collect data, and ultimately draw up a table which would 
be a guide to the Profession. Such a table would be no doubt a com- 
promise between the time worked in the air pressure and the time 
taken to decompress. It would be a statement of a body of men in 
high repute; if acted upon, it would relieve construction companies 
of liability to a great extent, and, if carefully formulated, it would 
greatly lessen the dangers of compressed-air work. 

The writer is indebted to Alfred Noble, Past-President, Am. Soc. 
C. E., Chief Engineer of the East River Division of the Pennsylvania 
Tunnel and Terminal Railroad ; and to E. W. Moir, M. Am. Soc. C. E. 
Vice-President of S. Pearson and Son, Inc., for permission to include 
such information as is taken from the work of the East River tunnels; 
and to Dr. Haldane for permission to make use of Table 1, and, 
generally, for the benefit of his theories, in writing this paper. 



24: DISCUSSION ON CAISSON DISEASE 

DISCUSSION 



Dr. Haldane. Dr. J. S. Haldane, F. R. S. (by letter). — After very carefully read- 
ing Mr. Japp's paper, the writer considers it to be an extremely valuable 
contribution to a practical solution of the prevention of caisson disease. 
He clearly points out the difficulties involved in adopting without 
modification for all compressed-air work the plan of "stage decom- 
pression" worked out by Dr. Boycott, Lieutenant Damant, and the 
writer, which has been in successful use since 190Y, for divers of the 
British Navy. A full account of the experimental evidence on which 
this plan is based will be found in the Journal of Hygiene, 1908, pp. 
342-443. It will be seen from this paper that most of the experiments 
were made at very high pressures, such as 75 lb., and that they were 
carried out with a view to deep diving. It was for many reasons 
desirable to leave a good margin of safety; and, therefore, this margin 
was left, particularly as practical experience in deep diving was very 
limited, on account of other causes which the writer and his associates 
had just succeeded in removing. 

Work in tunnels and caissons, however, is conducted at much lower 
pressures; practical experience is much more wide; and, thanks mainly 
to Mr. Moir, recompression is usually readily available. It becomes a 
question, therefore, whether the margin of safety which was allowed 
in decompression cannot be narrowed without appreciable danger. The 
new evidence which Mr. Japp has brought forward certainly seems to 
indicate that this is the case. He records the fact that in the construc- 
tion of the East River Tunnels 330 "seasoned" men were employed for 
36 days in two 3-hour shifts at a pressure of 40 lb., and were decora- 
pressed by a modified form of stage-decompression in 48 min., without 
a single serious case of caisson disease occurring. Thus there must 
have been about 24 000 individual decompressions without any mishap. 
On the principles laid down in the writer's paper, Mr. Japp calculates 
that the maximum air-saturation left in any part of the bodies of these 
men at the end of decompression, corresponded to a pressure of 27 lb. 
The writer's calculations agree with this estimate. In view of these 
results the author proposes to shorten stage-decompression in such a 
way that in place of the 19 lb. of residual saturation which the writer 
has proposed, 25 lb. should be left in the case of men who are 
"seasoned" to the work. This would greatly shorten stage-decompres- 
sion, and render it a much more easily practicable process, although 
the pains in joints, etc., which are the most easily produced and least 
serious symptoms of caisson disease, would not be prevented. 

The reason for the proposed 19 lb. is that occasional serious cases 
seem to occur at working pressures down to 20 lb., and that a fatal 
case is recorded at 23 lb. On the other hand, there can be little 



DISCUSSION ON CAISSON DISEASE 25 

doubt that for the great majority of men a pressure of 25 lb. would Dr. Haldane. 
not cause appreciable danger. Provided, therefore, that susceptible 
men can be excluded, Mr. Japp's proposal would seem to be consistent 
with reasonable safety. 

Susceptible individuals can doubtless be excluded, to some extent, 
by a careful medical examination, with special regard to the state of 
the heart and blood vessels, and to any general debilitating cause, 
among which age must be included. Another very important factor 
is fatness. Dr. Boycott and Lieutenant Damant* have recently shown 
by experiments on animals that when the exposure to compressed air 
is prolonged, fatness increases the susceptibility to caisson disease. 
They decompressed 70 guinea pigs at a dangerous rate after exposure 
to a pressure of 100 lb. Exactly half of the animals died of caisson 
disease, and the other half were killed. The percentage of fat was 
then determined in the bodies of the two sets of animals, and was 
found to be much greater in those which had died of caisson disease, 
the average difference being in the proportion of 225 to 100. The 
caisson disease simply picked out the fatter animals. In consequence 
of these experiments, which confirmed in a most striking manner the 
conclusions which had been previously reached on theoretical grounds, 
all British naval divers are now disqualified at the medical examina- 
tions from deep diving if they are inclined to fatness. 

While a medical examination will probably exclude most of the 
more susceptible individuals, it still seems very doubtful whether all 
can be excluded in this way. With new men, therefore, it is desirable 
that preliminary trials should be made with short shifts, as was the 
practice in the East Eiver Tunnel work. These men should be care- 
fully watched after decompression, and promptly recompressed if they 
show any threatening symptom; and any man who has shown threat- 
ening symptoms should be rejected. 

With these precautions, in addition to having a medical air-lock 
and keeping the men within reach of recompression for about an hour 
after they come out, the writer thinks that in all probability Mr. 
Japp's plan would prove successful. For casual visitors, the duration 
of stay could he limited, so that on coming out the maximum saturation 
in their bodies should not exceed 19 lb. 

The conditions vary so widely in different kinds of work in com- 
pressed air that it seems hardly desirable to lay down by legislation or 
otherwise any hard-and-fast general rules as to decompression. The 
writer, however, thinks that for any particular undertaking in com- 
pressed air, special rules, suitable to the particular circumstances, 
should be drawn up and strictly enforced. If these rules could also 
have some authoritative sanction, as, for instance, in the case of 



* Journal of Hygiene, 1908, p. 445. 



36 DISCUSSION ON CAISSON DISEASE 

Dr. Haidaue. "Special Rules" under the Coal Mines Regnlation Act in England, 
engineers and contractors would be relieved to a large extent of what 
at present must be a very uncomfortable responsibility. 
Dr. Keays. Dr. FREDERICK L. Keays* (by letter) . — The prevention of com- 
pressed-air illness is, of course, most important. It is likely that with 
the modern demands of engineering, even greater compressed-air under- 
takings will be projected in the future than in the past. Any means, 
therefore, which will add to the safety of the workmen, will not only 
benefit humanity at large, but will, also permit of more extensive 
undertakings. 

The three great factors in the causation of compressed-air illness 
are: The degree of pressure, the time under pressure, and the decom- 
pression period. The degree of pressure is, of course, the chief factor; 
and it may be stated that, other factors remaining the same, the higher 
the pressure, the greater will be the dangers of compressed-air illness. 
The degree of pressure which is to be used in any undertaking is not 
a matter of choice, but depends on the depth to which the work must " 
be carried. The only means of preventing illness, then, which are left 
to the engineer are the regulations of shifts and of decompression 
periods. In the past it has been thought that the longer one stays 
under pressure, the greater will be the dangers of compressed-air 
illness. This is true, up to a certain point, namely, that at which there 
is complete saturation of the blood and body fluids with the gases of the 
air. Dr. Haldane has made a valuable contribution to the subject of 
compressed-air illness by his work on the question of body saturation, 
and Mr. Japp has worked out Dr. Haldane's theory in a plain and 
satisfactory manner. 

The body fluids are practically saturated at any given pressure in 
about three hours. This belief is based on the results of the work in 
the Pennsylvania East River Tunnels. From a study of the records, 
it was found that men working for periods of less than 3 hours were 
less likely to have "bends," and much less likely to have a serious 
illness than those who worked for 3 hours or longer. Furthermore, it 
was found that in comparing the results of 6-hour shifts — which, since 
they were divided into two 3-hour periods with an interval of 3 hours 
at normal pressure, really meant 3 hours' exposure — with the results 
of 8-hour shifts, at almost equal pressures and with corresponding rates 
of decompression, there were fully as many cases of illness with the 
former shifts as with the latter, and certainly more serious and fatal 
cases. A point which has not been sufficiently emphasized in the past 
is that the real danger in compressed-air work comes with or after 
decompression. For this reason the practice, which has been almost 
universally followed, of dividing the working shift into two periods, 

* Late Medical Director for S. Pearson and Son. 



DISCUSSION ON CAISSON DISEASE 27 

is a bad one, for it doubles the number of decompressions. The results Dr. Keays. 
would be better in tunnel and caisson work, if continuous shifts of 
4, 6, or 8 hours were arranged to suit the pressures; say, 4 hours for 
pressures of from 42 to 50 lb. +, 6 hours for pressures of from 32 to 
42 lb. +j and 8 hours for pressures below 32 lb. +, and if the decom- 
pression periods were increased to reasonably safe lengths of time. 

The question of the decompression period is a difficult one to settle. 
In the first place it must be admitted that certain men have ' in- 
dividual susceptibility to the effects of compressed air. Just what the 
reason for this susceptibility is has never been learned. The fact 
remains, however, that while one man may work even at a high pressure 
and decompress rapidly without any symptoms of illness, another, 
under even more favorable conditions of pressure, length of shift, and 
time of decompression, may succumb to sudden death from syncope 
and collapse, and the autopsy will show large quantities of free gas 
in the circulatory system. Among practical compressed-air workers 
there is a decided difference of opinion as to the necessary length of 
decompression periods, some even claiming that quick decompression 
is preferable. Nearly all men, however, who have dealt with the sub- 
ject from the physiological and medical standpoint, have advocated 
fairly long times for decompression. The writer believes that lengthen- 
ing the time of decompression is the most efficient means of reducing 
the number of cases of compressed-air illness, as well as the number 
of serious cases. 

On account of the common occurrence of cases of delayed onset 
of symptoms, it is not believed that lengthening the decompression to 
periods claimed by Bert, von Schrotter, Hill, and Haldane, to be 
safe, would prevent all cases of illness, or even serious and fatal cases. 
In the writer's experience a comparatively large number of men first 
begin to have symptoms of compressed-air illness several hours after 
decompression, in some cases after as many as 12 or 15 hours, and in 
a fairly large number, after 3 hours. In two fatal cases symptoms 
began to appear between 2 and 3 hours after decompression. Why 
these men should not have had symptoms, even if they had taken as 
long as 1 hour for decompression, the writ-^r cannot see. He believes, 
however, that reasonably long decompression periods are efficient in 
reducing the number of cases of illness, disability, and death. 

Just what are reasonably safe lengths for decompression periods 
have never been determined. In the work on the Pennsylvania East 
River Tunnels after April, 1906, a rate of decompression of at least 
1 min. for each 2 lb. pressure was required, and quite generally fol- 
lowed. This seemed to cause less illness and a reduction in the number 
of serious cases, as compared with previous months when there were 
no rules as. to the rate of decompression. Even this rate of decom- 
pression, which was longer than has been required in any compressed- 



28 DISCUSSION ON CAISSON DISEASE 

Dr. Keays. air work in the United States, did not prevent serious and fatal cases. 
For this reason the writer thinks that even slower rates of decom- 
pression should be tried, for pressures above 28 lb. +, below which, 
this rate of 2 lb. per min. appears to be safe. 

Mr. Japp has set forth Dr. Haldane's method of "stage" decom- 
pression, which is claimed by the latter to be more efficient in pre- 
venting illness than uniform rates of decompression of even longer 
times. The writer does not think that enough experimental work has 
been done to prove the efficiency of "stage" decompression; but its use 
in diving has proved very satisfactory, and experience with the extra 
bulkhead in tunnel work, which necessitates an interrupted decompres- 
sion roughly corresponding to Dr. Haldane's method, would argue 
favorably for "stage" decompression. 

If it were possible, in accordance with Dr. Haldane's suggestion, to 
arrange a large decompression chamber in which a pressure equal to 
one-half the actual tunnel pressure could be maintained, where the men 
could pass a long decompression period in comfort, and use the time 
for bathing and dressing and, perhaps, for eating, it would probably 
prove a valuable means of preventing illness, and, at the same time, 
permit of working with comparative safety under higher pressures 
than could now be attempted. 

Medical supervision and examinations are important and are instru- 
mental in preventing serious results from the effects of compressed 
air. Men with organic disease are certainly less able to withstand the 
effects of compressed air than normal individuals. New men are 
certainly more susceptible to the effects of compressed air than old 
workmen. When high pressures are being used, a preliminary test of 
about li hours under tunnel pressure for all new men, followed by a 
second medical examination, and short shifts of from 1| to 2 hours 
for the first working period, are effectual means of guarding against 
serious results among new workmen. These means proved very valu- 
able in the work on the Pennsylvania East River Tunnels. 

Much remains to be learned in regard to the . prevention of com- 
pressed-air illness. This knowledge can be gained only by experimental 
work with animals, supplemented by records of compressed-air under- 
takings. By adopting new methods which appear to be improvements 
upon the old, we may hope to arrive, finally, at a more definite knowl- 
edge of what is reasonably safe in the lengths of shifts and decom- 
pression periods. 

Compressed-air work in tunnels and caissons, where pressures above 
30 lb. + are used, will always remain a dangerous occupation in spite 
of all reasonable precautions, for one can hardly hope to prevent all 
cases of illness. Those cases due to small quantities of free gas are 
almost sure to happen, and a small proportion of them is likely to 
result in paralysis, which may cause permanent disability or death. 



DISCUSSION ON CAISSON DISEASE 29 

Until the reason why certain men are susceptible to the effects of com- Dr. Keays. 
pressed air, can be learned, and such individuals can be recognized, 
cases of sudden death from large accumulations of free gas are almost 
certain to occur occasionally. 

Dr. Albert J. Loomis.*^ — The author deserves great praise for the Dr. Loomis. 
clear and lucid manner in which he has stated the ideas of those 
who have investigated and written on the subject of compressed-air 
diseases. 

But, after a pretty wide experience, extending back over twenty 
years or more, with pressures ranging from 15 to 42 lb., gauge 
pressure, the speaker cannot agree with him, or his authorities, as to 
the cause of the disease under discussion. While admitting the 
"saturation theory" as laid down by Bert, von Schrotter, Haldane, and 
others, the speaker does not think that it is the cause of the so-called 
caisson disease, for the following reasons: 

First. — As to the formation of bubbles of gas or air in the blood- 
vessels themselves: They have only been demonstrated in exceedingly 
high pressures — 15 or 20 atmospheres — which, at the present time, are 
considered impracticable as working pressures. If these bubbles should 
occur in the blood-vessels of the men, working as they do and coming 
out as they do, there would never be any use for hospital locks, as the 
men would all die of suffocation before they could be gotten to them, 
owing to these minute bubbles becoming entrapped in the very small 
vessels or capillaries of the lungs, and thereby preventing the oxygena- 
tion of the blood, a condition which operating surgeons frequently 
meet, to their great regret. 

Second. — If the symptoms were due to any change in the blood 
itself, there would be a general set attacking all parts of the body to 
a greater or lesser extent, which is the very thing that does not 
happen, but, on the contrary, they show themselves in a very few 
localities, and most of these localities are in the lower extremities, 
situations which any surgeon who is at all proficient in locating 
nervous lesions by their symptoms, would place without hesitation in 
the spinal cord, and in the lower portion of the cord. 

On the other hand, there is practically but one set of symptoms, the 
so-called "staggers," which would lead physicians to look for a cause 
or lesion in the brain, a condition which is quite rare when compared 
with the great number of cases occurring in the lower extremities. 

Should there be any change in the composition or condition of the 
blood, the brain would naturally be the first organ to show such change, 
as it is without doubt the most delicate organ of the whole anatomy. 

In reference to the CO2, the engineers of the North River Division 
of the Pennsylvania Tunnels have prepared a chart on which all the 

* Chief Medical Officer, North River Division, Pennsylvania Tunnels and Hudson Com- 
panies. 



30 DISCUSSION ON CAISSON DISEASE 

Dr. Looniis. cases of compressed-air sickness have been plotted, and this chart 
shows that, on that work, the greatest number of cases occurred at a 
time when the COg was only 1 part in 1000, with a pressure of from 
20 to 25 lb.; and, likewise, when the pressure was 28 lb. and the CO, 
was 2 parts in 1 000, there were no cases of caisson disease. 

While the speaker is not prepared to place much stress on the CO2, 
other than as an indication of the general ventilation, he does think 
that weather changes play a very great part in the causation of the 
disease, as do also the habits of living of those engaged in the work. 

When it comes to the examination of men, the medical officer should 
fit himself for the work by a close study of the conditions under which 
the men are to work, and his power of rejection should be absolute. 
The selection of the men should he the duty of one man who should 
be responsible for their welfare while they are at work under pressure. 

In substantiation of these opinions, the speaker desires to state 
that, in the work of the Hudson Companies, in the construction of the 
Hudson and Manhattan Tunnels, about 38 000 men have been ex- 
amined, since 1902, for pressures of from 15 to 42 lb., with three 
deaths, and in the construction of the North Eiver tubes of the 
Pennsylvania Tunnels about 7 000 men have been examined for 
pressures of 15 to 38 lb., with no deaths. 
Mr. Saunders. WiLLiAM L. Saunders, M. Am. Soc. C. E. — This is a Subject about 
which very little of practical value has been written, and the speaker 
is glad to note that Mr. Japp has gone into the matter in so much 
detail, both from theoretical and practical standpoints. 

Caisson disease, or the "bends" as it is commonly called, does not 
differ from other diseases, in that doctors are very apt to disagree 
about it. Heretofore the subject has been treated in print mainly 
from a medical standpoint, various theories being advanced as to 
the cause of this disease, that is, whether it is due to mechanical or 
chemical action. The speaker remembers when the most plausible 
theory advanced for the cause of caisson disease was that the blood, 
which had been compressed and held in a state of tension, was 
suddenly released by decompression, creating a pain when coursing 
through the veins. This was illustrated by the statement that if a 
string is wrapped tightly around one finger and then suddenly un- 
wound, the finger is temporarily in pain. The subject has been dis- 
cussed so fully by Mr. Japp, however, and the theory of caisson 
disease has been so well establislied by the experiments to which he 
refers that it is now generally accepted that bubbles are the cause of 
the "bends." 

Compressed air is the speaker's hobby, and in his early experience in 
seeking information about it he has soared, like the eagle mentioned 
by Mr. Japp, to a height of 3^ miles, where he has felt the earmarks of 
the "bends," and has descended to great depths in mines and in a 



DISCUSSION ON CAISSON DISEASE 31 

diving apparatus under the sea, where similar sensations have been felt. Mr. Saunders. 

Personal experience in a matter of this kind always fits one to 
judge better the causes as well as the effects, and it may add some- 
thing to the interest in this matter to recite an experience which the 
speaker had some fourteen years ago in the old Hudson tunnel, now 
called the McAdoo tunnel. This tunnel was started by Mr. D. C. 
Haskin a little before, or about the time of, the Centennial Exposition 
in Philadelphia, in 1876. Mr. Haskin conceived the idea of driving 
a tunnel under the Hudson River by maintaining a chamber of com- 
pressed air to hold up the face. He used no shield, and as the material 
through which he worked was more or less unstable, he met with con- 
siderable difficulty, and some lives were lost. He sank his entire 
fortune in the enterprise, the work being taken up later by English 
capitalists, who contracted with Messrs. Pearson and Son to finish the 
tunnel. At that time E. W. Moir, M. Am. Soc. C. E., was in charge of 
the work, and the experience which he gained there has no doubt 
helped him to design and carry through the recent tunnel construction 
under the East River, which Mr. Japp has so ably referred to and with 
which he has been connected in so important a position. During the 
administration of the Pearsons, the late R. P. Rothwell, M. Am. 
Soc. C. E., at that time editor of the Engineering and Mining 
Journal, and the speaker were invited to examine the tunnel for the 
purpose of making a report thereon. About an hour was spent under 
some 35 lb. absolute pressure. Coming out of the locks rather rapidly, 
on reaching the washroom, the speaker was taken with a severe attack 
of the "bends," the effect being like that of a paralysis of the lungs. 
A horizontal tank, resembling a common horizontal boiler, was at 
hand, and the patient was transferred to a bench of straw, the door 
was closed, compressed air was turned on, and the relief was instan- 
taneous. A pressure valve, provided for the purpose of admitting and 
discharging pressure from the tank, was used to enable the prisoner 
to escape, but it was noticed that when the pressure fell only two or 
three points the "bends" returned, and it was only by the gradual expan- 
sion of the pressure, occupying about h hour, that the patient was able 
to open the door and walk out cured. The necessity for reducing the 
pressure gradually, as shoA;vn very graphically by the return of the 
"bends," indicates the value of the theory of gradual decompression. 
It also bears out the idea that time must be given to permit the escape 
of the little bubbles which accumulate in the body. 

An interesting incident was noted in connection with this experi- 
ence. When the pressure in the tank was lowered, through the dis- 
charge of air, a fog formed instantaneously, and as soon as air was let 
in and the pressure increased, this fog was as promptly dissipated. 
The speaker experimented with this for some time — manufacturing 
and annihilating fog. The lowering of pressure, of course, resulted in 



32 DISCUSSION ON CAISSON DISEASE 

Mr. Saunders, cooling the air, and, vice versa, increasing the pressure heated the air. 
As the air was cooled, fog was precipitated immediately when the tem- 
perature reached the dew point, the capacity of air for moisture de- 
pending entirely on volume and temperature, that is, a certain fixed 
volume of compressed air or free air can hold only a certain quantity of 
moisture at a certain temperature. If the temperature is raised, it 
will hold more water; if the temperature is lowered, the moisture 
must be precipitated first in fog, and later, in the form of rain. 

Mr. Aims. Walton I. AiMS, M. Am. Soc. C. E. — The literature on caisson 
disease is quite extensive. The causes of the disease now seem to be 
well determined, but difficulty is found in conforming practically with 
the medical rules and regulations for its prevention. 

The age of the compressed-air worker has an important bearing 
on his ability to withstand the effects of too rapid decompression, but 
youth generally lacks experience, and, in difficult work, some com- 
promise has to be made in drawing the age limit for entering the 
compression chamber. Dissipation acts strongly in undermining a 
man's resistance to the effects of compressed air, especially among 
workers of mature years, and even some temporary indisposition may 
occasion an attack in one who has previously been pronounced physi- 
cally fit by a medical examiner. 

In nearly all compressed-air work there is difficulty in compelling 
the workmen to take a reasonable time in the air-lock while decom- 
pressing. Believing that this is due largely to the extremely un- 
comfortable conditions in the ordinary air-lock, a method for relieving 
the lock of the cold and fog when decompressing was used at the 
Battery-Brooklyn tunnel by introducing a constant flow of heated dry 
air while the lock was being exhausted. In addition to removing the 
cold and fog from the lock, this heated air also provided for ventilation 
while decompressing, a feature not found in the ordinary compressed- 
air lock. The comparative freedom from caisson disease at the 
Battery-Brooklyn tunnel, where practically no age limit was estab- 
lished, seemed to indicate the value of thus heating and ventilating 
the air-lock during the decompression period. 

Dr. Erdman. Dr. Seward Erdman.* — The Speaker wishes to acknowledge his 
appreciation of Mr. Japp's instructive contribution to the literature 
of caisson disease, and feels that the general adoption of the sug- 
gested "Step Decompression Method" will aid much in robbing com- 
pressed-air work of its gravest dangers. 

The reduction, to terms and curves of mathematical precision, of 
the time and pressure elements, as shown on his charts, is most 
interesting; but it seems to the speaker that there will always be 
another factor to be reckoned with, one which cannot be reduced to 

* Formerly a Medical Officer on the East River Tunnels Contract. 



DISCUSSION ON CAISSON DISEASE 33 

mathematical terms, and yet which will continue to render uncertain Dr. Erdinan. 
the result of exposure to compressed air in each individual case. 

This indeterminate factor may be called the "personal equation," 
and seems to be dependent on the functional efficiency of the circulatory 
and respiratory apparatus. 

In the experience of all who have seen much of compressed-air 
work, cases are not infrequent where men, who are seasoned air 
workers, will work daily, for long periods of time, at fixed conditions 
of pressure, length of shift, and decompression time, without suffering 
any ill effects, but who come to work the next day, under exactly the 
same conditions, only to develop a severe or even fatal attack of "bends." 

The decompression time, indicated on Mr. Japp's charts, can be 
controlled by valves from the air locks ; but, for the individual, de- 
saturation of the body tissues and the blood must take place through 
the lungs; therefore the lungs act as the "escape valve" for the 
individual. 

In explanation of such accidents as are above cited, it is likely 
that this "escape valve" is temporarily clogged. 

Desaturation of any given part of the body will be impeded by 
causes which slow the blood current in that area. 

Thus passive congestion of the lungs, or the presence of mucus in 
the air spaces (as from a cold) may act as clogs to the "escape valve," 
thus increasing the time necessary for desaturation, and, under such 
conditions, the body does not rid itself of the gases in the average time, 
and an attack of "bends" results. 

T. Kennard Thomson, M. Am. Soc. C. E. (by letter). — The writer Mr. Thomson, 
has read this valuable paper with much interest, and notes that the 
author does not refer to the old-fashioned remedy — the electric bat- 
tery — probably because it is not nearly as reliable as recompression. 
The writer once cured himself of an attack of bends by electric 
shocks, after he had failed to get any relief from recompression; and 
the next time the battery failed to cure him and recompression did — 
and so it goes. He has had the bends after being under an air 
pressure of 23 lb., and was not affected by a pressure of 35 lb. on the 
same job. The purity of the atmosphere and the physical condition of 
the man at that particular time have much to do with the case. 

The writer, having no ear drums, has never had much trouble 
in keeping his ears from getting blocked, and, as a rule, he has been 
affected much less by the air pressure in vertical caissons than in 
tunnels; this he ascribes partly to the fact that the air in caissons is 
renewed more frequently than in tunnels and therefore is fresher, and 
partly to the increased time he was obliged to remain in the tunnels 
to get to the headings. 

Under the Harleai River, when under comparatively light pressure, 
while sinking through the foul river bottom, his men have been much 



34 DISCUSSION ON CAISSON DISEASE 

Mr. Thomson, troubled by the bends, but have experienced very little inconvenience 
nnder much greater pressure while the caissons were sinking through 
the clean clay below the river-bed. 

In an emergency, the writer has seen men work for 18 or 20 hours 
in pressures up to 25 lb. without any ill effects. He has gone in and 
out of the caissons repeatedly for 86 hours (without going to bed), 
under pressures up to 35 lb., without inconvenience — but would never 
be guilty of such hours again. 

When he had the bends, the pain disappeared while he was in a 
perspiration induced by drinking hot coffee, only to return when he 
had cooled off. He then tried a very hot bath and felt well until he 
had dressed. 

As stated by Mr. Japp, slow decompression is the best preventive 
known, and recompression is not always reliable. As the author 
also states, hot coffee should always be taken on coming out. 

Mr. Japp. Henry Japp, M. Am. Soc. C. E, (by letter). — The discussion, thus 
far, seems to indicate that the general opinion is that the cause of 
caisson disease is the presence of gas bubbles in the blood and tissues, 
the only one taking exception to this being Dr. Loomis, who states 
that bubbles have only been demonstrated under exceedingly high 
pressures — 15 or 20 atmospheres. 

The writer cannot agree to this, as he has seen the results of three 
autopsies on men who have died on coming out of pressures between 
30 and 35 lb. gauge, and in each case bubbles were plainly visible to 
the naked eye in the blood vessels; and he has read reports of many 
more, and in almost every case similar bubbles have been seen. In 
one case Drs. McWhorter and Erdman were able, at the autopsy, to 
collect and analyze the air which escaped from a large bubble in the 
heart. It may be thought that these bubbles are not present during 
the life of the subjects, but the writer has also seen what had every 
appearance of being little bubbles passing along in the blood vessels 
of the wrist of a member of the East River Tunnels staff about half an 
hour after he came out of the air pressure. 

Dr. Loomis' extraordinary success, in having so few deaths as 
he mentions, is highly commendable, but there is no doubt that if a 
larger number of fatal cases had passed through his hands he would 
have made similar observations. 

The fact that Dr. Leonard Hill has exhibited the formation of 
bubbles and the stoppage of the circulation in the web of a living frog 
by sudden decompression, and the disappearance of the bubbles and 
the resumption of the circulation by recompression, should of itself 
show that, although this was obtained by high pressures for the pur- 
pose of rapid demonstration to an audience, the same thing can readily 
occur with a lower pressure. 



DIWCUSSION ON CAISSON J)1SJ0A8K 



35 



Wliilo JJr. Looinis docs not accept the theory that bubbles of gas Mi 
cause bends, he presents no alternative. 

The writer is much indebted to Dr. Ilaldane for checking the cal- 
culation for stage decompression from 40 lb. pressure, and is glad to 
have it confirmed. 

Dr. Haldane suggests that as the conditions vary so widely in 
different kinds of work, it is undesirable to lay down by legislation or 
otherwise any hard-and-fast rules as to decompression, and in the 
paper the writer calls attention to the danger of legislation if it is not 
guided by the Profession. 

Since this paper was Ma-itten, however, the New York State Legis- 
lature has brought forward an act, now signed by the Governor, and 
taking effect January 1st, 1910, which limits everything in a very 
hard-and-fast manner. The hours of labor and the rates of decom- 
pression are regulated according to Table 5. 

TABLE 5. 



Japp. 



Gauge 
pressure, 
in pounds. 


Time under pressui'e. 


Interval between spells. 


Uniform 

decompression , 

in minutes. 


0-38 


8 hours, less interval 

2 spells of 3 hours each . . . 
2 spells of 2 hours each . . . 
2 spells of 1}4 hours each. 
2 si)ells of 1 hour each 


30 consecutive minutes spent in the 
open air 


18?4 


28-35 99 


At least 1 hour . 


34 


36-41.99 


At least 2 hours 


42 


42 45 99 


At least 3 hours 


46 


46-49 99 




50 









No employee shall be permitted to work in pressures exceeding 50 lb. per sq in. except 
in cases of emergency. 

The decompression shall be at the rate of 3 lb. every 2 min., unless the pressure shall 
be over .36 lb., in which event the decompression shall be at the rate of 1 lb. per min. 

This law is a great step in advance of anything attempted privately 
in the United States, and the Commissioner of Labor is to be congratu- 
lated, but if these rates are compared with the suggested Table 4 of 
the paper, it will be found that the new law proposes a greater degree 
of safety for the lower than for the higher pressures, and inasmuch as 
the law requires that the decompression shall be uniform, the final 
pressure in the blood on coming into the atmosphere will be much 
higher than would result from stage decompression. 

In order that this may be seen more clearly, the writer has calcu- 
lated the pressure in the blood for various pressures, decompressing 
uniformly as required by the law, and also by stage decompression, 
and gives the results in Table 6. 

It will be noted that, whereas the stage decompression proposed 
fixes the pressure in the blood on emerging from the air-lock at a 
constant of 25 lb. per sq. in., the uniform decompression periods re- 
quired by the law give pressures in the blood on emerging varying 



3G 



DISCUSSION ON CAISSON DISEASE 



Mr. Japp. from 25.70 up to 32.50 lb., such a result being obtained in the case of 
50 lb. with 17 min. longer for uniform decompression than stage de- 
compression, and when it is noted that, for a pressure as low as 9 lb. 
per sq. in., the law requires 2 min. for every 3 lb., or 6 min. for de- 
compression, it will be seen how inconsistent are the requirements. 
In other words, the new law requires generally more time for uniform 
decompression than is needed for stage decompression, while giving 
less safety. 

TABLE 6. 



Tunnel 
pressure, 
in pounds 

(gauge). 


Time worked, 
in hours. 


Uniform de- 
compression, 
in minutes. 


Pressure in 

blood on 

emerging, 

in pounds. 


Stage de- 
compression, 
in minutes. 


Pi essure in 

blood on 

emerging, 

in pounds. 


28 

36 

41.99 

45.99 

50 


8 
3 
2 

1 


42 
46 
50 


25.70 
30.25 
31.25 
32.00 
32.50 


14 
36 
37 
35 
33 


25 
25 

25 
25 
25 



It is doubtful whether Dr. Keays is correct in stating that the body 
fluids are practically saturated at any given pressure in about 3 hours. 
Fig. 6 shows that in 3 hours the slower saturated parts are only 81% 
saturated, but the writer is of the opinion that much could be done 
toward ascertaining the rate of saturation by further experiments. 

With regard to Dr. Keays' remarks on "delayed caisson disease" 
(that is, cases which occur many hours after decompression), and his 
deduction, from those which occurred after too rapid decompression, 
that the lengthening of the decompression to periods suggested by 
Dr. Haldane would not prevent all cases of illness, or even serious 
and fatal ones, this cannot be accepted as more than an expression of 
opinion, inasmuch as no cases had occurred in Dr. Haldane's ex- 
perience when his rules were adopted, and it is doubtful if any delayed 
cases would occur or would have occurred had Dr. Haldane's rules been 
adopted, as the sole object of slow-stage decompression is to prevent 
the formation of injurious bubbles. 

Delayed caisson disease after too rapid decompression can be ex- 
plained by the accumulation of gas in tissues, which, though strong 
enough to stand the pressure of the accumulated gas for some hours, 
ultimately break down under the strain. Liberated gas collected in 
pockets cannot very readily be absorbed again excepting by recom- 
pression. 

IMr. Saunders' description of the effects of recompression and 
decompression in his own case, when he was paralyzed, is an excellent 
illustration of the bubble theory. 

Dr. Erdman's idea of the lungs acting as an escape valve is very 
ingenious, and is quite in line with the general mechanical theory of 



DTSCUSSTON ON CAISSON DISEASE 37 

caisson disease. An investigation on the lines of his suggestion Mr. Japp. 
would be very valuable. 

Mr. Aims' reference to the effect of dissipation on the resistance 
of men to the effects of compressed air and to the effect of the general 
comfort of the men coming out of the air-lock is generally accepted as 
important, but with slow decompression the discomforts of fog and 
cold are not felt in air-locks, and therefore heating is not as important 
as in the case of more rapid decompression. 

Dr. Loomis' suggestion that the weather changes play a very 
great part in the causation of disease is not borne out by observations 
made by the medical staff on the East River Tunnels. 

Mr. T. Ivennard Thomson's remarks are valuable, coming from a 
man with so much actual experience in air pressure, and his acceptance 
of slow decompression as being the best preventive known — coming as 
it does from an engineer, the bulk of whose experience has been in 
caissons with small air-locks, where slow decompression is not an easy 
matter — leads one to hope that, even under such disadvantages, prac- 
tical men will ultimately arrange for a slow decompression. 

The reason why recompression is not always reliable, as Mr. Thom- 
son says, is that the bubbles having torn the tissues and done the dam- 
age, no amount of recompression which merely removes the bubbles 
will heal the torn tissue, but if recompression can be resorted to 
promptly, before any permanent injury is done, then relief will be 
obtained. 



AMEEICAN SOCIETY OF CIVIL ENGINEEES 

JNSTITUTED 1 853 



TRANSACTIONS 



Paper No. 1119 

COPYRIGHT IN DRAWINGS OF A TECHNICAL 
CHARACTER. 

By D. a. Usina, Assoc. Am. Soc. C. E. 



The copyrig-lit law, enacted Marcli 4th, 1009, to go into effect July 
1st, 1909, contains provisions which may be of great importance to 
engineers, in that it provides specifically for copyright on "Drawings 
(or plastic works) of a scientific or technical character," and in that 
the securing of the copyright is made an extremely simple matter. 

Prior to the passing of this law there has been no certain protec- 
tion against the pirating of designs for engineering structures. Where 
a design though new has involved only the skill of the engineer, it 
has been unpatentable, and, unless the work could be classed under 
the fine arts, there has been no reasonable hope of protection by copy- 
right. In many cases designs involving a high degree of engineering 
skill and gotten up at great expense have been copied by others with 
no moral justification whatever. National, State and city governments 
and private parties have asked for bids to be accompanied by original 
designs, and have appropriated the design of one bidder and let the 
contract to another — a species of piracy distinguished legally but not 
morally from the piracy of works of literature or the fine arts. 

It is hoped that the new law will put a material check on such 
practices by reserving to the proprietor of such drawings the sole 
right of printing or copying them. It is doubted if the law can be 
invoked to prevent the building of the structures shown in such draw- 
ings, with the possible exception of works which might be classed in 



COPYRIGHT IN TECHNICAI. DRAWINGS 39 

the fine arts. The limitations which can be put upon the use of the 
drawings, however, should afford a substantial degree of protection, 
and the ease with which copyright can be secured makes it undoubtedly 
worth trying for whatever protection can be secured. 

The language of the Act relative to technical drawings is as 
follows : 

"That the application for registration shall specify to which of the 
following classes the work in which copyright is claimed belongs: 
* * * (i) Drawings or plastic works of a scientific or technical 
character." 

Taken in connection with the remainder of the law, there is little 
or no doubt that this language was intended to render such drawings 
copyrightable. The view of the Register of Copyrights is : 

"that the language of the Act would authorize the deposit and regis- 
tration in this ofiice for copyright protection of drawings of a scientific 
or technical character such as under the present law would not be in- 
cluded in the term 'drawing,' which seems to be confined to artistic 
drawings. Examples of such works would be drawings for machinery, 
for engineering construction, for architectural works, etc. While these 
drawings have to some extent been registered in the Copyright Office 
heretofore, it has always been a question whether the law authorized 
the registration and whether copyright protection would follow." 

Protection is extended to "all the copyrightable component parts 
of the work," such for example as the drawings relating to any com- 
plete and mentally segregable part of the whole structure. 

Copyright is secured by publication with the required notice affixed 
to each copy; or, where "not reproduced for sale," by depositing a 
photograph or other reproduction with the Register of Copyrights. 
But engineering drawings are usually offered for sale (as part of the 
bidder's work), and, in such cases, the copyright is secured by affixing 
a notice of copyright; and after thus securing the copyright there 
must be promptly deposited in the Copyright Office two complete copies. 

Several permissible forms of notice are provided. It is sufficient 
if the word "Copyright" be applied to each sheet of drawings, together 
with the name of the proprietor and the date. It is customary now to 
apply the name and date to drawings; and the addition of the word 
"Copyright" is the simplest matter in the world. For what it is worth 
(and apparently it will sometimes be worth much) the word "Copy- 
right" should be added to all original drawings and copies thereof. 



40 COPYRIGITT IN TECHNICAL DRAWINGS 

In the case of drawings of especial value, it is advisable also to 
deposit copies promptly with the Register of Copyrights and to secure 
a certificate of such deposit, so as to obtain the apparent sanction of 
the Government, and to indicate such fact on the face of the drawings. 
For drawings of less value, the depositing of copies may be delayed 
until it is thought necessary to threaten or bring suit for infringement. 
The securing of copyright is probably effected by the mere notice. 
The deposit of copies is an additional formality which, though required 
to be made "promptly," and though a necessary preliminary step before 
a suit for infringement can be brought, does not seem necessary to the 
securing of the right. 

Copyrighted works may be sold or furnished to others with any 
reasonable limitation upon their use. For example, a bidder may 
furnish an original design with a notice of copyright and a special 
notice that the copyright and the copies submitted constitute parts 
of the bid and remain the property of the bidder in case his bid is not 
accepted, and that the design is not to be copied or submitted for bids 
from others. 

An interesting feature of the new law is the apparent sanctioning 
of actions against the United States and its officers; but much weight 
cannot be placed on this construction, in view of the broad prohibition 
in the general law against actions "sounding in tort" against the 
United States. Congress has provided no remedy for wrongs com- 
mitted by the National Government, except for violations of contract, 
express or implied. The provision of the new law bearing on this 
subject relates specifically to the recovery of costs in suits for infringe- 
ment and states that in such suits, 

"except when brought by or against the U. S. or any officer thereof, 
full costs shall be allowed, and the court may award to the prevailing 
party a reasonable attorney's fee." 

It is unnecessary to consider the question here at length, but it 
may be stated generally that after a tort is conmiitted by the United 
States there is no adequate remedy, but an injunction may be secured 
against a repetition of it; and perhaps, if one has sufficient notice of 
an intention to infringe the copyright, such infringement can be 
enjoined in the first instance. Possibly the framers of the law had 
such injunctive proceedings in view when they framed the above 



COPYRIGHT IN TECHNICAL DltA WINGS 41 

section exempting the United States and its officers from the payment 
of costs. 

The Act provides also that "any person who willfully and for 
profit" infringes a copyright may be punished by a fine and imprison- 
ment; and an infringer who had received such a special notice as is 
above referred to would be infringing willfully and for profit, and 
would hesitate to run the danger of fine and imprisonment. 

Under the old copyright law there were only a few cases in which 
it was attempted to secure a copyright upon a purely technical draw- 
ing. It was necessary to enter the drawings in the office of the 
Librarian of Congress, and the Register of Copyrights permitted such 
entry only under protest and was of the opinion that such drawings 
were not copyrightable. The validity of the copyright was generally 
doubted. The new law, which specifies such works particularly as one 
of the classes in which copyright may be claimed, and permits the 
securing of the copyright by applying to the drawings a notice that 
copyright is claimed, is a material advance toward the protection of a 
kind of intellectual property of far greater dignity and usefulness than 
circus posters and similar matters, to which the previous law had 
extended protection. To quote the language of the Constitution, upon 
which is based the power of Congress to make a copyright law, techni- 
cal drawings do more "to promote the progress of science and useful 
arts" than many of the classes of work heretofore provided for. 

The practical value of the new law to engineers cannot be predicted. 
We can only hope. It cannot be used to protect functional equiva- 
lents of the thing represented by the drawings. That is the purpose 
of the patent law, and protection thereunder runs for seventeen years 
only, while copyrights may last fifty-six years. It is quite possible for 
the courts to construe the new law so narrowly as to destroy its value. 
But it will probably be construed so that it will protect the engineer 
against the copying by others of the substance or of any material part 
of his original designs. 



AMERICAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1120 

THE SIXTH STREET VIADUCT, KANSAS CITY.* 

By E. E. Howard, Assoc. M. Am. Soc. C. E. 



With Discussion by Messrs. Daniel Bontecou, Victor H. Cochrane, 

(). K. MoGENSEN, N. T. Blackburn, George TI. 

Pegram, and E. E. Howard. 



The Sixth Street Viaduct, recently built, is a steel structure, about 
8 000 ft. long, providing an elevated roadway and two car tracks across 
the broad, flat valley of the Kaw River to the high land on each side. 
It was designed to accommodate, and, by improved conditions for 
comfort, safety, and speed, to facilitate street-car and vehicular traffic; 
thus contributing to the unification of the social and commercial 
interests of Kansas City, Missouri, and Kansas City, Kansas. 

The main viaduct and those portions of the side laterals or ap- 
proaches now finished include, in general figures, 1 900 lin. ft. of road- 
way 38 ft. wide, 6 100 lin. ft. 30 ft. wide, 1 000 lin. ft. 24 ft. wide, 
and 1000 lin. ft. 20 ft. wide, a total of 10 000 lin. ft. of roadway; 
also 8 500 lin. ft. of double-track railway. In this construction there 
have been used more than 120 000 lin. ft. of piles, 23 000 cu. yd. of 
mass concrete, 13 000 tons of steel, nearly 10 000 cu. yd. of reinforced 
concrete, 34 000 sq. yd. of asphalt pavement, besides minor and mis- 
cellaneous material. On one lateral, or side approach, where an incline 
could not be conveniently placed, there were installed two large electric 
traffic elevators of sufficient capacity to raise any vehicle and team. 
Figs. 1 and 2, Plate H, are general views of the viaduct and the bridge. 

♦Presented at the meeting of April 7th, 1909. 



THE SIXTH STREET VIADUCT, KANSAS CITY 



43 




44 THE SIXTH STREET VIADUCT, KANSAS CITY 

Kansas City is located on the south side of the Missouri Kiver, 
below one of its great bends, and lies in the States of Missouri and 
Kansas, with the crooked Kaw Eiver flowing through it to join the 
greater stream. While commercially and socially possessed of common 
interests, the community has political and governmental divisions, 
due to its position on the border line of two States, and is divided into 
two distinct cities with their various suburban settlements. 

The valley of the Kaw River, popularly called the "West Bottoms," 
is about a mile wide and is quite flat; it is skirted on each side by 
high blufi^s and hills, while the channel of the river meanders through 
it, in its lower reaches lying close to the western bluff. The Missouri 
River has been gradually pushed northward, through the agency of 
dikes and mattresses, and has deposited, in the last fifteen years, a 
strip of land, about one-quarter of a mile wide, extending from side 
to side of the valley of its tributary. While practically the entire 
residence portions and the principal retail business districts of both 
cities are on the hills, there are to be found, crowded into the Kaw 
Valley, many commercial industries, packing houses, stock yards, 
manufacturing plants, wholesale warehouses, the freight stations of 
nearly all the thirty-four railroads entering the city, several large 
terminal freight yards, and the present union passenger station. As 
an essential adjunct, there is, throughout the bottoms, a complicated 
network of railway tracks, occupying streets and alleys as well as 
private rights of way. 

The transaction of the business of the cities involves an enormous 
traffic into this valley from both sides, and all traffic from city to 
city must pass through it. A very large portion of the freight con- 
signed for local consumption must be hauled out of the bottoms; some 
up-town distributing houses haul goods out, store them, and then 
haul them back; the packing houses, factories, and other industries 
are continually conveying their products to both cities; and, too, a 
great deal of the merchandise consumed in Kansas City, Kans., is 
hauled from Kansas City, Mo. 

All this traffic is confronted by very inadequate passageways and 
very heavy grades from valley to hills, not less than 5% on the east 
side, and not less than 6% on the west side. Furthermore, on the east 
side there are from the northern part of the city only two possible 
entrances to the bottoms; and, on the west side, giving access to 



PLATE I. 

TRANS. AM. SOC. CIV. ENGR8. 

VOL. LXV, No. 1120. 

HOWARD ON 

HE SIXTH STREET VIADUCT, KANSAS CITY 




THE SIXTH STREET VIADUC'J', KANSAS CITY 45 

Minnesota Ave., the principal business street of Kansas City, Kans.. 
there is only one passageway. In addition to the steep grades to be 
contended with, the vehicular route to every freight station is over 
dangerous grade crossings frequently congested with traffic, while the 
interurban travel is compelled to follow circuitous routes. 

Not only does the highway traffic to and from the bottoms operate 
under adverse conditions, but the street railway facilities are inade- 
quate, there being only two through lines to handle the entire between- 
city travel. One of these is a very circuitous surface line, operated at 
great discomfort because of the many grade crossings, the steep 
grades, and the congestion of streets in the bottoms. The other line 
is elevated for a considerable part of the way, but the route is indirect. 
The need of additional rapid transit across the bottoms was long keenly 
felt by the street railway interests and the public, especially the latter. 

Various measures have been devised from time to time to expedite 
this travel, and a viaduct has often been advocated. Shortly after 
the disastrous flood of 1903, which, by the invmdation of the entire 
west bottoms, for a time completely cut off all intercommunication, a 
plan was proposed for the two cities to build jointly a street-car and 
vehicular viaduct, but it soon became apparent that no money was 
or would be available for such a purpose, and in any case the diffi- 
culties of two city governments jointly handling a structure located 
in two cities, two counties, and two States, seemed almost insuperable. 

The proposition to build a toll viaduct by private capital for an 
investment was then broached, and investigation was undertaken to 
determine the feasibility of such a plan. A preliminary survey showed 
that a viaduct could be built across the accreted lands near the Missouri 
River, then unoccupied, which would shorten the distance from center 
to center of business districts by street car 4 150 ft. (26%) and 18 
min. (60%) ; would give easy grades and uninterrupted passageway 
from city to city for highway travel; and, by a judicious location of 
laterals and approaches, could give convenient access, with elimination 
of all grade crossings, to each of the various freight stations in the 
bottoms. 

Careful investigation, extending over a number of months, was 
then undertaken to determine: First, the amount of time to be saved 
by the patrons of a viaduct; second, the deterioration of equipment 
to be lessened; third, the additional loads to be carried per vehicle; 



IG THE SIXTH fSTKEET VJADUCT, KANSAS CITY 

fourth, the actual money to be saved by the user; fifth, the amount 
of toll the user could afford to pay and still profit by the use of a 
viaduct; sixth, the probable volume of traffic to patronize such a 
structure; seventh, the effect of the viaduct on the distribution of 
population, and the consequent increase of travel; eighth, the probable 
growth of the cities and the increase of business; ninth, a minute and 
thorough estimate of the cost of construction; tenth, the annual cost 
of operation; eleventh, the annual income; and tvpclfth, the earnings 
to be realized. 

For many of these questions there was no precedent, and some 
conclusions were based on unsatisfactory data. The determination 
of the number of vehicles passing to and from the bottoms required 
merely an extended series of actual counts, secured from day to day, 
principally by detectives of the Pinkerton service. These counts were 
made at various times from October, 1903, to June, 1904 ; and numerous 
independent tallies agreed closely with a total of about 12 000 vehicles 
per day to and from the bottoms. 

As a further verification of these figures, statements were obtained 
from leading transfer men and large haulers as to the estimated use 
they would make of the viaduct, the total sum of which showed an 
entirely satisfactory figure. 

A study of comparative statistics showed Kansas City to have 
exceptional business prospects, for, while at present twenty-second in 
population, it ranks as ninth in bank clearings, twelfth in total postal 
business, second in railroads, second in grain business, and it is 
situated in the center of a rich and productive territory. 

Neither did the estimates of cost of construction and maintenance 
offer particular difficulty; but the determinations of the individual 
profit to be secured and the toll which could be reasonably expected 
were but approximations. 

The matter of the utilization of the viaduct for street cars was 
taken up early with the company operating the cars in the city, and 
with certain suburban lines, with the result that tentative agreements, 
later duly confirmed by contracts, were made for traffic arrangements. 

The conclusion of the preliminary investigation was that the struc- 
ture as then planned, reaching from city to city, with laterals at prop- 
erly selected points giving uninterrupted access to the freight houses. 



THE SIXTH STKIiET VIADUCT, KANSAS CITY 4? 

was feasible as an investmeiit, could be built for about $2 275 000, 
could be operated for about $180 000 per annum, and could earn from 
5% upward on the capital stock, depending on the future growth of 
the city. 

The Kansas City Viaduct and Terminal Railway Company was 
organized to build and operate such a viaduct, and the Common 
Councils of the two cities granted to the Company the necessary fran- 
chises. The principal provisions of these franchises are : 

The franchise is to continue in force for thirty years; 

A schedule of maximum tolls is fixed; 

A clause provides that the City may at the end of ten years pur- 
chase the viaduct at any time, by notifying the Company one year 
in advance, at the actual cost and a stipulated varying interest rate; 

A provision is made for the Company to pay the City 2% of its 
gross revenue. 

As soon as certain necessary legislation was obtained from the 
Missouri Legislature, financial arrangements were completed, right of 
way was secured, and the construction, for which general contracts 
had already been made, was begun. 

It had been the intention to commence the Mulberry Street Ap- 
proach first, to finish that approach and the east end of the main 
structure from the approach to the eastern terminus, and have this 
portion in operation before the completion of the main viaduct; but 
the Company's attorneys discovered many difiiculties in getting right 
of way for the approach, so the construction of the Main Viaduct was 
proceeded with until the legal complications could be untangled. 

The following pages contain a discussion of the design and con- 
struction in the following order: 

1. — Superstructure, design, manufacture, and inspection. 

2. — Substructure, design, and construction. 

3. — Engineering field methods. 

4. — Erection of superstructure metal. 

5. — Floors, concrete floor, paving, street-car deck, hand-rail, and 

lighting system. 
6. — Approaches and laterals. 
7. — Final costs. 



48 



Tiiii SIXTH ,sti;l':1':t viaduct, Kansas city 



1. — Superstructure. 
The design provides a roadway 30 ft. wide over a part, and 38 ft. 
wide over the remainder of the structure, to carry a loading varying 
according to span from 50 to 100 lb. per sq. ft., or a 15-ton road roller; 
and, separated from the roadway by a substantial hand-rail, a double 
track for electric cars, to support a continuous line of cars on each 
track, each car weighing 40 tons and having a length of 43 ft. 3 in. 
over all. Each of these assumed loads is increased by an impact per- 
centage as noted hereafter. 



GENERAL CROSS-SECTION 

6TH STREET VIADUCT 
KANSAS CITY 




SECTION OF MAIN VIADUCT 
through 38-ft. roadway 

Fig. 2. 

Before adopting the layout for the spans, an extended economic 
study was made by preparing estimates for different arrangements of 
girders and columns. The varying amount of shopwork, and the unit 
cost of material, as well as the total weight of metal per linear foot, 
were duly considered, the resultant variation of substructure cost with 
span length was estimated, and the most economical girder length, 
floor arrangement, and tower spacing for the confronting conditions 
were definitely determined. 

This led to the selection of a two-column structure with one column 



THE six'i'ii ,s'J'i;ki':t viaduct, kansas city 



49 



about under the center of the roadway, and the other about under the 
center of the motorway, with longitudinal girders riveted into the 
columns, cross-beams riveted into the girders and columns, and string- 
ers laid on top of the cross-beams. 



I Lep'accdabout \ljj>'"°="'^-'"^-" ;i4/ Bars.OiaStV U ' K ' For 30 p„Rua* 

I*—"! lOO'apart ^-^-" 2i^i^^:2' P "^ ^s'o 

One iW'Raa pipe conduit i rT'2"1 Expanded Metal i 



TRANSVERSE SECTION OF HIGHWAY PAVEMENT 
FOR 30-FT.ROADWAY 




J^ Cor Sar.C long over int. stringers 
at each floor beara^ 



Si^feSd «Oo.Ba.,.Ne„S... 



'^' Aflphalt 

h- Fiue t'tone 



I Roadway 



E 0-8 Bolls 1 _ 
Bpaced S o'Venters ^-'^ 



Irou;h PlB. 




DETAILS OF 
EXPANSION JOINT 



CROSS-SECTION OF ROADWAY FLOOR 
AND DETAILS OF HAND-RAIL 

6TH STREET VIADUCT 
KANSAS CITY 



f — 10^! 
DETAILS OF HAND-RAIL POST 

Fig. 3. 

Considerations of possible increase of railway traffic resulted in 
the decision to arrange the viaduct so that at any time there could 
be added two additional tracks, each to carry a continuous line of cars, 
each car weighing 70 tons and having a length of 43 ft. over all. This 
is to be done by adding a third row of columns with girders, and 
splicing the cross-beams to the original motorway cantilever beams. 

The two columns are 33 ft. from center to center transversely, and 
spaced so as to be loaded equally when the final loading is applied. 



50 THE SIXTH STREET VIADUCT, KANSAS CITY 

The tlaird column will be 31 ft. from the original motorway column 
and at mid-point between the two added tracks. The transverse brac- 
ing is arranged so that tracks may be laid on the ground between the 
columns, and occasional turnouts are allowed for by spans of special 
length. 

Longitudinally, the structure is divided into sections composed of 
one span of 30 ft., and seven spans of 45 ft. The 30-ft. spans have 
longitudinal bracing to care for the entire thrust of the 345-ft. sec- 
tion, thus imposing very few obstructions to free transverse passageway 
below, a stipulation of certain river-front property holders. The mid- 
girder between the towers is supported at one end in a pocket or shelf 
so as to slide and allow for expansion. 

The highway floor is of concrete, enveloping the steel stringers, 
and supporting a pavement of asphalt. Hand-rails are placed on each 
side of the highway. The car tracks are of ordinary rails with the 
usual creosoted pine ties and guard-rails. The maximum grade on 
the main viaduct is 1.5%, and the minimum is 0.5%, this latter being 
maintained so as to drain the asphalt pavement properly. 

The following working intensities were used in the design : 

Tension Stresses. 

Pounds per 
square inch. 

Flanges of plate girders (counting in one-eighth of the 

web) 14 000 

Kolled shapes 16 000 

Bending Stresses. 

Extreme fiber of rolled sections 16 000 

Extreme fiber of timber beams 2 000 

Compression Stresses. 

Top chords 18 000 - 70 

r 

Inclined end posts 18 000 — 80 ^. 

'*• 

Intermediate posts and diagonals 16 000 — SO 

r 

Lateral struts and bracing 16 000 — 80 

'/■ 

Columns 16 000 — 60 ' 

r 



TIIK SIXTH STRliliT VfADtKJT, KANSAS CITY 51 

Shearing Stresses. Pounds per 

square inch. 

Webs of plate girders 10 000 

Rivets 10 000 

Impact, for motorway loading 7 = — 

i + .500 

. ,.,,,. ,400 

Impact, for highway loading Z = ^ , ^ ^,-, 

Li -j- 1 50 

Under extreme temperatures and full live load, the shortest columns 
were computed to have an extreme fiber stress of 24 000 lb. per sq. in., 
but this was not considered excessive for such unusual conditions. 

All calculations were based on the use of equivalent uniform live 
loads instead of actual wheel concentrations. Because of the very 
considerable labor involved in computing anew, for each slight change 
in panel length, the cross-beam reactions, the results from a series of 
panels of different lengths were plotted and a curve was drawn con- 
necting all points. This diagram proved of great service in designing 
the many special girders, for a principal part of the calculations was 
saved by taking off directly the amount of the imposed loads. 

Trestle Construction. 

The various special girders and columns were detailed in con- 
formity with the standard portions of the structure, and may be 
generally included in the following description. 

Stringers. — Both the highway and the motorway stringers consist 
of I-beams set on top of the cross-girders and cantilevers, and con- 
nected thereto by rivets or bolts. Small shims are used under the 
highway stringers to provide a crown for the roadway. All stringers 
are continuous for the corresponding girder length, either 30 ft. or 45 
ft., and are joined by small splice-plates. At expansion joints, string- 
ers are halved and reinforced with bearing angles so that they slide 
upon one another. Connecting the motorway stringers, there is a rigid 
system of bracing, composed of single-angle diagonals, and of cross- 
struts of either one angle or one channel. The connections for all 
this bracing are made by lug-angles to the webs of the beams so as to 
leave their upper flanges clear for convenience in laying the ties. 

Cross-Girders and Cantilevers. — Cross-girders are spaced at 15-ft. 
centers, thus placing two intermediate beams in the 45-ft. girder spans 
and one in the 30-ft. spans. They are riveted to the webs of the main 
girders or to the columns, with their upper flanges level with the tops 



52 THE SIXTH STREET VIADUCT, KANSAS CITY 

of the main girders. These beams are 52 in. deep, with f-in. webs, 
and flanges composed of two angles. The intermediate stiffeners, which 
are pairs of angles placed at each concentration point, are crimped, 
while the end stiffeners have fillers on the web. The cantilevers at 
each end of each cross-beam follow the same detail, that on the high- 
way side varying in depth from 52 to 12 in., though the motorway 
cantilever is of full depth to its end, where provision is made for splic- 
ing on a future cross-beam. Splice-plates extend over the girders or 
columns, and connect the top flanges of the cross-beams and canti- 
levers. At columns, the thrust at the lower flange is carried into the 
cross-beams with cast-iron thrust-blocks bearing against the channels 
of the posts and the webs of the main girders. These blocks were cast 
to a driving fit, were driven to place after the field riveting was 
completed, and were also supported by bolts through the column webs. 

Main Girders. — The standard 45-ft. girders are 72 in. deep, with 
|-in. webs and duplicate flanges of two angles and cover-plates. The 
general details are the same as for the cross-beams, except that the 
intermediate stiffeners are spaced 5 ft. apart. 

Columns. — The columns, throughout, are of two rolled or built chan- 
nels joined by a built I, with the webs of the channels in general 
parallel to the longitudinal axis of the structure. Stay-plates connect 
the flanges of the channels. Columns extend clear to the top of the 
main girders and cross-beams, providing for them full riveted connec- 
tions. Lug-angles and brackets are riveted to the columns for the con- 
nection of the bracing, and shelf-angles are provided for convenience 
in girder erection. The shoe is built of side-plates and angles, fastened 
to the channels, and distributing the load to two heavy base-plates. 
Curved plates, 18 in. high, are fastened to each side, as sleeves for the 
anchor-bolts, and as bearings for the anchor-nuts. 

At expansion points the columns have on one side pockets made of 
plates riveted to the channels supporting a bottom or base on which 
the girders slide. The tops of the columns are tied back to the adja- 
cent span in a manner to relieve the main connection rivets of tension 
due to eccentric loading. 

Bracing. — Lateral bracing is used only on curves, and in each 
tower span, it being considered that elsewhere the concrete floor would 
give all needful lateral rigidity. Where used, it consists of two 
diagonals of two angles each, with sufficient area to take the stress in 
either tension or compression. These angles are placed just below 



PLATE II. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXV, No. 1120. 

HOWARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 




Fig. I.— The Completed Viaduct, Looking from the Eastern End Toward Kansas. 




Fig. 2.— The Completed Bridge Over the Kaw River. 



THE SIXTH STllKET VIADUCT, KANSAS CITY 53 

the cross-beams, are riveted to the column beams and to the main 
girders, and also to each other beam they touch. 

Longitudinal bracing at the towers is made up of two diagonals 
and a lower horizontal strut, all composed of four angles, laced, 
IE -shaped, and of dimensions to fit into the columns. At the upper 
points the diagonals are attached also to the longitudinal girder. 

Transverse stiffness is provided by knee-braces of two angles fast- 
ened to the column webs, and to the lower flanges of cross-girders, each 
column being braced in this way. The ends of pairs of girders resting 
in the expansion pockets are connected by a light bracing frame. 

Truss Spans. 

At the Kaw River are used two truss spans, each 300 ft. lOJ in. 
from center to center of end pins. They are of the double-intersection, 
deck, Warren type, with subdivided panels, there being ten main panels. 
The depth from center to center of chords is 37 ft. 6 in., and the trusses 
are spaced 33 ft. apart. These spans are riveted throughout, and atten- 
tion is called to their exceptional size ; each span weighs about 900 tons. 

The floor system is arranged just as on the trestle, with cross-beams 
and cantilevers riveted into the top chords and short vertical posts. 
Two small expansion joints for the floor are used on each span to 
avoid the distortion of the floor-beams as the top chord shortens. 

The top chords are of two web-plates, a cover-plate, and four angles,, 
with the upper angles placed inside. The various top chords are spliced 
only enough to assure alignment, and the contact of abutting webs 
is relied on to carry the compression. The bottom chords are of two 
built channels laced with angles. The mid-span chord has an area 
of 212 sq. in. 

Diagonals are alternately of built channels turned out, and of built 
I -struts, the latter passing through the former. Thus the channels 
rivet on the outside of the gusset-plates and the I's between them. 

The upper laterals are of two-angle diagonals placed in the plane 
of the lower edge of the chord. They reach the chords only at main 
panel points, and therefore intersect at the intermediate floor-beams, 
to which they are attached by plates passing through slotted holes in 
the webs, and are connected thereto by lug-angles. The lower laterals 
consist of diagonals, between main panel points, of laced, four-angle, 
I -struts, the depth of the chords. 

The upper laterals are made heavy enough to dispense with inter- 



54 THE SIXTH STREET VIADUCT, KANSAS CITY 

mediate transverse bracing between the trusses. Between the end posts 
there is cross-bracing of four-angle X -struts from the floor-beam 
above to a transverse strut below. This arrangement avoids the trans- 
ference of loads from truss to truss due to unequal deflections, and also 
the unsatisfactory details of transverse bracing fastened to other than 
vertical members. 

The shoes are built up of plates and angles, and are connected to 
the trusses by 12-in. pins. Ten segmental rollers, 8 in. high, support 
the expansion shoes, and cast-steel base-plates are used at each end. 

The trusses were calculated on the assumption of all loads being 
equally distributed between the two systems. Although the trusses 
are equally loaded when fully loaded, the resultant stresses vary some- 
what, because the larger part of the load on the highway truss is from 
the concrete floor, while on the motorway truss the live load is much 
greater. Both were calculated, and, in the make-up of sections, the 
chord stresses of the highway truss were combined with the web 
stresses of the railway truss and the two trusses were made identical. 

The one other truss bridge of the viaduct is a skewed span, 147 ft. 
long, over railway tracks. This span is of the same general type as 
that at the Kaw River, except that the intermediate vertical posts are 
absent, the floor-beams being supported on the gusset-plates. The depth 
of trusses is 15 ft. 6 in. At each end, owing to the skew, there is a 
diagonal girder, 62 ft. long, and the regular cross-beams are riveted 
into it. The span is supported on shoes of the usual type, that at the 
roller end resting on seven segmental rollers supported on a cast-steel 
base. There is no bracing, either longitudinal or transverse, in the 
columns under this span, and especially heavy sections were used. The 
columns supporting the fixed end are 42 by 32 in., having about 200 
sq. in. of section. Each of the four columns is anchored to its concrete 
pedestal by eight 2i-in. bolts 9 ft. long. 

Specifications and Manufacture. — The consulting engineers pre- 
pared complete detailed drawings, and from these the contractor pre- 
pared working drawings. These engineers' drawings do not give in 
full all minor details, such as the exact spacing of rivets, sizes of 
small plates, etc., but do fix the dimensions of all members and of all 
parts thereof, allowing the shops to fill in the lacking dimensions 
according to their preference. The shop drawings were then checked 
by the consulting engineers, who made the necessary corrections, and 
finally approved all details. As a representative, located at the shops, 



'J'llE SIXTH STRKET VIADUCT, KANSAS CITY 55 

gave all points prompt attention and decision, much time was saved 
and correspondence avoided. 

All metal is open-hearth, medium steel. The specifications provided 
that all plates should be rolled from slabs of at least six times the plate 
thickness, which slabs were to be made by rolling an ingot and cutting 
olf the scrap. The following maximum percentages of ingredients were 
fixed: 

Phosphorus, 0.04 to 0.06; sulphur, 0.04; silicon, 0.04; manga- 
nese, 0.70. The ultimate tensile strength on test pieces was to be 
from 60 000 to 70 000 lb., and the least allowable elastic limit, 35 000 lb. 
The least allowable elongations on specimens varied from 20 to 24%, 
with a corresponding reduction of area of from 44 to 36 per cent. 

First-class workmanship in every respect was specified. Sheared or 
hot-cut edges of plates were planed so as to remove i in. of metal. 
Rivet holes were sub-punched h in. less, and reamed to a diameter iV in- 
greater than that of the rivet, and it was required that the reamed 
holes be perpendicular to the metal surface. A modification in part 
of the foregoing was made, and the field holes for the stringer bracing 
and certain other minor parts were punched full size. 

The ends of all girders and abutting members were planed to 
make perfect contact. StifPeners on girders were ground so as to fit 
tightly against the flanges. This was strictly enforced in the case of 
stiffeners carrying concentrated loads, and in certain cases where 
the stiffeners did not quite touch they were removed and new angles 
riveted on. Edges of spliced web plates were planed and put in con- 
tact for their full length. 

It was specified that special attention be given to the cleaning and 
painting of the metal-work in the shop, where one coat of paint was 
applied. The metal was cleaned of rust, dirt, and loose scale before 
painting, and all surfaces of metal in contact were required to be 
painted before being put together. In a few instances it was necessary 
to have finished work cut apart in order to do this painting. 

A large portion of the work was laid out direct, without the use of 
templates. Gauge punches were used almost exclusively on pieces 
which duplicated several times. There was much duplication, of which 
could be mentioned 220 45-ft. girders, 480 floor-beams, 350 cantilever- 
beams, practically all stringers, and so forth. Although the columns 
are all of different lengths, groups were made together requiring only 
the details near the lower ends to be varied. 



56 THE SIXTH STREET VIADUCT, KANSAS CITY 

The rate of shop work varied considerably, but averaged about 900 
tons per month, finished and shipped. The Kaw Bridge spans, com- 
prising some 1 800 tons, were fabricated and shipped in five weeks. 

Careful inspection was made of all operations in the shop. Rivets 
were examined as driven ; there was a general supervision of laying out,, 
punching, shearing, assembling, etc. ; and the finished piece was checked 
before loading. The result of this careful inspection was apparent in 
the very few errors found in the field. 

Full records of the progress of each piece were kept, and weekly 
reports were made giving the status of work of all members and the 
percentage of the estimated total completed. The weights were all 
calculated from the drawings, and checked against the scale weights. 

2. — Design and Construction of Substructure. 

Surveys and borings showed bed-rock to be about 50 ft. below the 
ground surface across the entire valley except for a few hundred feet 
from each bluff, where it is very much lower. The material overlying 
bed-rock was found to vary from soft muck and running sand to clay, 
firm sand, and gravel. 

Because of the uncertain bearing capacity of this accreted land, 
and of a possibility of scour, piles were used under 276 of the 326 
pedestals of the viaduct. 

The first design contemplated the use throughout of a patented 
form of concrete pile, and the pedestals were designed accordingly. 
This patented pile was to have been built by driving into the ground 
a steel shell or mould having a removable point, and on reaching the 
required depth the shell was to be filled with concrete and gradually 
pulled out as the concrete was tamped into place. Further considera- 
tion and examination of driven piles made it seem to be undesirable 
and inexpedient to use this patented pile in the wet soil to be 
encountered. 

The manifest desirability of a concrete pile made the engineers 
loath to abandon that type of foundation, and led to the adoption of a 
built-up driven pile. A number of experimental built-up piles were 
made at once, and, after the feasibility of driving them was demon- 
strated, their general construction was immediately undertaken. 

Of the pedestals not supported on piles, some rest on rock, some on 
hard clay, and some on soft material. Thirteen pedestals near the east 



THE SIXTH STREET VIADUCT, KANSAS CITY 



57 



Olid of the viaduct are under the buildings of a machine shop or under 
raih'oad tracks. The material here is very soft, and, after experiments, 
an intensity on the soil of 1 ton per sq. ft. was allowed. In order to 
have the excavations shallow, the bases were reinforced. Plain steel 



bars were used 




i'm'-^ — ^ii'9— 

21'- 



Near the western end of the viaduct the pedestals rest on firm red 
clay, loading it to about 3 tons per sq. ft. The three Kaw Eiver piers 
are founded on rock, two being sunk by the pneumatic process, and one 
built by open excavation. 



58 THE SIXTH STREET VIADUCT, KANSAS CITY 

files. — The moulded concrete piles are all of one size, 30 ft. long, 
10 in. square at the lower, and 16 in, square at the upper end, with a 
diagonal 3-in. cut off each corner. Each is reinforced by four §-in. 
steel bars, extending from end to end, symmetrically placed as near the 
surface as practicable, and surrounded by a cylinder of heavy woven- 
wire fencing intended to occupy a position about 1^ in. from the sur- 
face of the concrete. In the later piles the wire was omitted. 

The forms or moulds for the piles were simple boxes, with bottom, 
sides, and ends of 2-in. plank suitably cross-braced. Lugs on the cross- 
braces engaged wooden wedges which tightened and held the parts 
together yet readily permitted them to be knocked down for moving. 
The diagonal corner moulding made the form almost water-tight. The 
upper end of the pile was shaped into a round head to facilitate 
driving, and the lower end was bluntly pointed. The several parts, of 
course, were made interchangeable. The interior surfaces were coated 
with oil before concreting. The piles were moulded lying on the 
ground, in groups of from 8 to 16, in position convenient to their point 
of driving. 

The specifications provided that the piles be made of concrete 
composed of "1 part of Portland cement, 3 parts of Kaw River sand, 
and 5 parts of clean hard broken stone to pass a 1-in. iron ring". The 
cement used was lola Portland, and cost, f. o. b. Kansas City, $1.60 
per barrel. The sand of the Kaw River is very clean, sharp, and well 
graded, testing about 52% retained on a No. 30 sieve; its cost, deliv- 
ered, was 75 cents per cu. yd. The stone was from quarries about the 
city, and cost $1.30 per cu. yd. delivered. There was considerable 
trouble in securing clean stone, for the limestone strata of this region 
are overlaid with clay, and, especially in a rainy season, it is impossible 
to prevent some clay from intermingling with the stone in the crusher 
bins. 

Some limited experiments were undertaken to determine whether 
the amount of clay present had a deleterious effect on the concrete. 
Test bars made of measured quantities of material, mixed in a uniform 
manner, were prepared with the stone as it was delivered from the 
crusher, and with stone carefully and thoroughly washed and cleaned. 
When these bars were tested and compared it was found in all cases 
that the unwashed clayey stone made stronger concrete than the 
clean washed stone. 



PLATE III. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXV, No. 1120. 

HOWARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 




Fig. 1. — Concrete Piles: Finished Piles Curing. 




Fig. 2.— Pile-Drivers, Showing Special Arrangement of Followers. 



THE SIXTH STREET VIADUCT, KANSAS CITY 59 

The concrete for the piles was made by hand on tight movable 
platforms. The practice was to mix together the stone, sand, and 
cement dry, and then add enough water to give a reasonably wet 
mixture. The material was distributed by wheel-barrows, and dumped 
into the prepared forms. After a 2-in. layer of concrete had been 
spread in the form, the reinforcing of wire and rods, already prepared, 
was laid in position, the remainder of the concrete was dumped in, 
tamped to place, and smoothed on top to a reasonably even surface. 
Sticks were used to tamp through the meshes of the wire and to hold 
the rods apart. The rods had a tendency to bunch together, and it 
required constant care and watchfulness to get them where they be- 
longed. Several methods were tried to obtain the desired results easily, 
but none was entirely satisfactory. The conclusion is, however, that 
stiif iron spacers, either as a ring or as an X, at intervals of about 3 ft. 
along the rods, would give superior results. 

Each pile was marked with the date made, and for several days 
thereafter was showered copiously with water. The side and end forms 
were usually removed in 48 hours, but the pile was not rolled from its 
base for about a week. 

Over-anxiety to re-use the base forms without delay caused the 
loss of several piles, for the heavy green piles were easily cracked by 
very small inequalities of the ground. Weather conditions, of course, 
affected the rapidity with which the piles hardened, but they were 
generally allowed to cure from 3 to 5 weeks before use, although some 
were driven 2 weeks after being made. JEach pile contains 1.4 cu. yd. 
of concrete. The finished piles are shown in Fig. 1, Plate III. 

For driving, or rather sinking, the piles, heavy turn-table drivers, 
equipped with a special kind of follower, were used. This follower was 
of timber, about 12 ft. long, with guides to fit the leads at its upper 
end and mid-point, and with a deep cylindrical collar at its lower end. 
A standard 3 000-lb. hammer rested on the follower and was fastened 
loosely to it. The usual "pile line" was carried down, through a steel 
sheave, then up and made fast to the top of the leads. To this sheave 
was fastened a 10-ft. length of steel cable having a slip-noose at its 
lower end, and a "pull-off line" attached to the slip-loop of the noose. 
A water-jet pipe, consisting of a 2-J-in. pipe 35 ft. long, with its lower 
end drawn down to a nozzle, completed the equipment. This jet pipe 
was handled by a line to a spool of the engine, and was supplied with 



60 THE SIXTH STREET VIADUCT, KANSAS CITY 

water at a pressure of about 120 lb. through a 3-in. standard fire-hose. 
To provide the water for the jets, a pumping plant, consisting of a 
high-duty 18 by 10 by 12-in. fire pump with suitable boiler equipment, 
was erected on the bank of the Missouri River about J mile from the 
structure, and a pipe line, from 3 to 4 in. in diameter, was laid from it 
along the entire viaduct. Fig. 2, Plate III, shows the pile-drivers and 
the special arrangements of the followers. 

In general, jjiles were driven before any excavating was done, and 
they were sunk, so that their upper ends were from 1 to 5 ft. below 
the ground surface, in the following manner : 

After the driver had been set to the correct position by the adjustment 
of the rollers on which it stood, the hammer and follower were lifted 
to the top of the leads. The jet was then played on the desired position 
of the pile, and, as the ground softened, it was gradually pushed down 
full length. In the meantime the slip-noose cable was placed around 
a pile some 4 ft. from its upper end and the pile was raised and swung 
into the leads where it was straightened up and the follower settled 
carefully on the protruding head. By means of the sheave-line, the 
pile was then churned up and down, the pile, the timber follower, and 
the hammer resting thereon all going up and down together, while the 
jet was kept in constant service. Drops of as much as 6 or 7 ft. were 
made, without apparent injury to the pile, and the churning process 
was continued until the required depth was secured. By the "pull-oflF" 
cable, the slip-noose was easily loosened and removed. 

The effect of running the jet down first was to cause the pile to 
keep its correct position, as it was found that otherwise each pile had 
a tendency to travel toward the adjacent ones previously driven, where 
the ground was still soft; and even then the pile tended to lean toward 
the side where the jet was played during sinking. It is believed that 
more accurate sinking could have been done with a pair of jets, one on 
each side. The simple slip-noose for handling the piles was developed 
only after the failure of several more or less elaborate schemes devised 
for the same purpose. 

The time consumed in the actual sinking of a pile varied somewhat, 
but, where no particular obstructions were encountered, from 2 to 5 
min. sufficed. However, the highest record for one gang in 10 hours 
was 36 piles. 

It was soon apparent that, owing to the delay in starting, cold 



PLATE IV. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXV, No. 1120. 

HOWARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 




Fir;. 1. — Pedestals: Form Removed. 




F[G. 2.— Tower Span, with Longitudinal Bracing. 



THE SIXTH STREET VIADUCT, KANSAS CITY Gl 

weather would probably set in and stop concreting before all the con- 
crete piles could be made, and it was necessary to devise a substitute 
so that the structure might be finished in the time assigned. Creosoted 
pine piles were, with reluctancy, decided upon; a rush order was given 
to pile contractors and creosoting plants, and, by constant attention, 
the piles were rapidly delivered. 
The specifications called for: 

"Piles cut from live, straight, long-leaf, yellow-pine timber; to be 
free from cracks or wind-shakes; to be so straight as to be at no point 
more than one-third of their sectional diameter out of a straight line 
joining the centers of the ends; to be of even, gradual taper from end 
to end ; and to be not less than 10 in. in diameter at the small end." 

The specifications for creosoting provided, after the usual require- 
ments for dressing, peeling, and artificial seasoning, that: 

"After the seasoning is entirely finished, pure, unadulterated dead 
oil of coal-tar shall be injected into the timber so that to each cubic 
foot of timber will be 12 lb. of creosote oil." 

Additional equipment for handling the timber piles was hastily 
provided. The drivers were of the ordinary light timber-framed type, 
braced by guy lines to the top of the leads, which were 60 ft. high. 
Each was mounted on two wooden rollers, 1 ft. in diameter and 20 ft. 
long, resting on timber sills laid on the ground. Lateral movement 
was secured by sliding the driver along the rollers. This roller arrange- 
ment proved good, permitting rapid moving and setting. Both drop- 
hammers and steam-hammers were used, the latter doing much more 
efiicient work. Water-jets, arranged about as previously described, 
were used. The efficiency of the jets was demonstrated in several 
instances when, because of a break-down of the pumps, attempts were 
made to drive without water, for these invariably resulted in the piles 
being injured and broomed, before they had reached their full depth, 
thus precluding further driving. 

On a portion of the structure, 50-ft. piles were driven, reaching to 
bed-rock; on the remainder, 35-ft. piles were used, resting, as do the 
concrete piles, on gravel and on clay. 

To determine the bearing capacity of a concrete pile, a test load 
was placed on one of the first piles sunk. The pile loaded was of 
standard size, was made on August 28th, driven on September 15th in 
25 min., and the loading was begun on September 24th. To support 



Chi TUK SIXTH STREET VIADUCT, KANSAS CITY 

the load, a timber platform, 12 by 4 ft,, was balanced on a single 
post resting on the pile head. Observations were made with an engi- 
neer's level, reading on a scale fastened to a piece of 1-in. pipe which 
was set into the pile a few inches and projected above the platform 
without being in contact with it. The load was of pig iron, each pig 
weighing about 66 lb., and was placed by hand at an average rate of 
6 tons per hour. 

During the first If hours the load increased from 4 tons (the weight 
of the platform) to 19 tons, with a very small settlement. This load 
was constant for the next 15^ hours, and the pile settled 0.012 ft. Dur- 
ing the next 4 hours the load increased from 19 to 35 tons, with an 
additional settlement of 0.038 ft. After an interval of 1^ hours, load- 
ing was resumed for 4 hours and 40 min., with a settlement of 0.097 ft. 
During the 17 hours and 40 min. this total load was kept on the pile, 
there was a further settlement of 0.062 ft. In the 44 hours of unload- 
ing, the load on the pile was reduced from 43.75 to 4 tons, and the 
pile rose 0.021 ft., leaving a net settlement of 0.198 ft. 

These results made necessary some revision in the pedestal design, 
as very much greater capacity had been claimed for the patented pile. 
Moreover, the material varied so much from place to place that this 
test was not believed to be representative of the entire structure, and 
some ready method for tests was desirable. It was clearly impracticable 
to load each pile to determine its bearing capacity, and as there is 
no way to estimate the capacity of piles sunk by jet, the following 
plan was adopted. 

Creosoted pine piles, 40 ft. long, were furnished and called "test 
piles." In cases where, by reason of special softness of the ground in 
driving adjacent pedestals, or from the data of the borings, there was 
uncertainty, one or two of these "test piles" were substituted for the 
concrete piles. These were driven by using the drop-hammer in the 
ordinary way, without a jet, and before any jetting had been done 
in the immediate vicinity. Observations were made of the drop of 
the hammer and the penetration of the pile from 25 to 30 ft. down, or 
about the position to be occupied by the concrete piles, and the re- 
sultant bearing capacity was calculated by the well-known formulas. 
In all cases these results gave a greater safe load than that actually 
on the piles, in fact, it was usually necessary to resort to the jet to get 
the timber piles to the full depth. 



TUK SIXT[1 S'I'RKET VIADIIOT, KANSAS CITY 63 

In the main viaduct there are 798 concrete piles, a total of 23 950 
lin. ft. in place; and 2 573 creosoted timber piles, aggregating 93 800 
lin. ft. in place ; a total of all piles of 117 750 lin. ft. 

Excavation and Concreting. — The excavation for pedestal bases was 
comparatively simple, and the work was done by hand. In nearly all 
instances where concrete and the shorter timber piles were used, no 
sheeting or bracing was needed, the sandy, clayey material standing 
to depths of 8 or 9 ft. For about 80 pedestals, most of which required 
pits some 13 ft. deep in very boggy, sticky ground, sheet-piles of 2-in. 
planli were driven by hand around timber frames, and these frames 
were dropped down to brace the planks as the excavation proceeded. 

As soon as the piles were sufficiently exposed they were sawn off 
at the required elevation. Before concreting they were cleaned of mud 
and filth by scraping with a hoe having a curved blade. Where con- 
crete piles were not settled to full depth they were cut off with chisels, 
and the reinforcing bars bent over. 

A number of pedestals, particularly those at the east end, under 
buildings and near tracks, required special attention in excavation 
with regard to bracing and shoring. One building, under which several 
were constructed, having been through two fires and a flood, was in a 
precarious condition, especially as it rested on soft flowing material. 
The base of one pedestal under it extends under the bed of an engine 
which was kept in operation by supporting it temporarily until the 
concreting was completed. For certain railway tracks carrying a heavy 
and frequent traffic, it was necessary to drive temporary trestles to 
permit the construction of adjacent pedestals. 

Forms. — In most cases no forms were used for the bases of pedestals, 
the pits being merely filled with concrete to the required elevation. 
The pedestal shafts are practically of uniform dimensions, and the 
forms were devised for repeated use. Each side was built as a single 
piece, of 1-in. horizontal, matched, dressed boards, with vertical posts 
of 6-in. square timbers. The four sides were held together by three 
square frames or hoops of 3-in. plank, bolted at each corner, bearing 
sidewise against the studs. Wedges were driven behind all posts not 
properly bearing. These shaft forms were easily adjusted to correct 
position by the insertion of small wedges under the bottom. Sheet-iron 
corner moulds were used to give the concrete a curved corner of 3 in. 
radius. For the vertical edges, probably the best results were secured 



04 THE SIXTIL STREET VIADUCT, KANSAS CITY 

by using sheet iron of No. 12 or 14 gauge, rolled to approximate radius 
and then nailed firmly to one side of the form and tacked to the adja- 
cent side only enough to keep the concrete from getting behind. No 
particular harm was done to this corner mould in pulling down the 
form. The upper corners were shaped by sheet-iron or wooden moulds, 
or by the use of a curved trowel. The latter, in the hands of a skilled 
workman, gave excellent results, but its use was deprecated because 
of apparent desire to use unskilled labor. Two wooden templates for 
supporting and holding in place the anchor-bolts, all of which were 
placed as the concreting proceeded, were fastened to the form. 

Concrete. — The concrete was specified to be mixed by machinery 
in the proportions: One part of Portland cement, three parts of clean, 
coarse, sharp sand, and five parts of hard, clean, broken stone to pass a 
2^-in. iron ring; the proportions to be determined by volume; all ingre- 
dients to be measured loose; and enough water to be used to form a 
wet concrete. 

In addition to the usual requirements for package delivery, etc., 
the specifications for the cement were: To be so fine that 97% in weight 
will pass a No. 74 sieve, and 90% pass a No. 100 sieve; to be of such 
tensile strength that briquettes of neat cement will show, for 1 day, from 
150 to 250 lb. per sq. in.; for 28 days, from 400 to 600 lb. per sq. in.; 
to show no drop in strength; to be slow-setting; preliminary set in not 
less than 30 min., final set in not less than 3 hours; to undergo boiling 
and steaming tests without failure. 

Sand was defined as "particles of hard, clean stone which will pass 
a :J-in. mesh sieve, and not less than 50% of which shall be retained on 
a No. 30 sieve." Both the Kaw and Missouri Rivers have great sup- 
plies of excellent sand, that of the former being somewhat cleaner, but 
both coming within the requirements. 

"Sound, hard rock, free from all dust and dirt," was called for, but. 
as has been noted, the output of local crushers, containing more or less 
clay, was used. It is difficult to state just how much clay was allowed 
to be in the stone, as a given load would seem very muddy when wet, 
yet clean when dry. The stone was generally accepted when there 
were no lumps of clay, or of clay and stone chips, even though many 
of the stones were muddy. Inspectors were stationed at the quarries 
to examine the material as loaded, yet, in spite of all care, Pome stone 
was used which contained small balls of clay. 



THE SIXTH STREET VIADUCT, KANSAS CITY 65 

Special detailed specifications in regard to mixing concrete were 
distributed among the various sub-contractors, to assure uniformity of 
product. 

The concrete was mixed so wet that no real "tamping" could be 
done, but it was deposited approximately in layers of from 9 to 15 in., 
and "thoroughly spaded" by thrusting a spade to the full depth of the 
blade at intervals over the entire area, special attention being given 
to spading near the surfaces of piles and forms, so that the mortar 
would have perfect contact. It was in general handled from the mixers 
in wheel-barrows, either on the ground or on elevated runs, as condi- 
tions required, and was thrown freely from the tops of the pits or forms. 
For other pedestals and for certain abutments, slightly more equip- 
ment was needed, and derricks with boxes of simple well-known types 
were used in handling materials. Anchor-bolts were usually placed 
after the lower half of the shaft of the pedestal had set, and they were 
very carefully adjusted to exact position. The upper foot of each 
pedestal was of rich concrete made with Joplin flint or granite chips. 
One of the pedestals, with the form removed, is shown in Fig. 1, Plate 
IV. The pedestals were finished from J to | in. low, to allow for grout- 
ing under the shoes. 

Various types of mechanical concrete mixers were used, both batch 
and continuous machines. The continuous mixers automatically meas- 
ured the sand, stone, and cement by means of worms, and were mounted 
so as to be easily portable. Although considerable pedestal concrete 
was made with these machines, the results were not as satisfactory as 
those produced by rotary batch mixers which mixed the concrete in 
revolving drums or boxes. The output of the continuous mixers was 
not uniform, even though they measured the materials accurately, 
especially where the vicissitudes of the work necessitated intermittent 
operation, with frequent stops and starts; the water was not easily 
regulated, and the mixture was not homogeneous. Moreover, after a 
time, numerous spots appeared on the surface of the concrete from 
these machines. It was found that these were caused by small balls of 
clay, from J to | in. diameter, embedded in the concrete, the results, 
of course, of clayey stone. Careful examination, however, proved that 
these clay balls were practically absent in the work of the batch mixers, 
showing that the churning and tumbling had been sufiiciently vigorous 
to pulverize and distribute all lumps of clay, with consequent harmless- 



66 THE SIXTH HTRKET VIADUCT, KANSAS CITY 

ness; while the mixing of the continuous machines had not been suffi- 
ciently vigorous to do so and therefore was inferior. 

Certain pedestals, abutments, and retaining walls had special forms, 
made of 2 by 6-in. matched, dressed flooring, laid on vertical studs 
spaced from 2 to 3J ft. from center to center. One side of each form 
was braced to the ground, and the other side was adjusted to it by 
numerous malleable-wire ties. A strut, fixing the correct width, was 
placed near each tie-wire. 

Vertical cleavage planes were made in abutments and retaining 
walls, separating those parts of the wall having different load inten- 
sities on the foundation. A cross-partition was built in the form, and 
the concrete was finished against it. Before building the adjoining part 
of the wall, the partition was removed, and tar roofing paper was nailed 
securely to the exposed surface, thus effectually preventing cohesion of 
the fresh material to the old. 

Kaw River Piers. — The three piers of the Kaw River Bridge are of 
concrete, of the same materials and proportions as used for the 
pedestals. 

The west shore pier rests on a ledge of rock about at low-water 
level, so that the construction was very simple, and the work was done 
in the dry. Light sheet-piling was driven around the pit, which was 
excavated by hand. Hand-pumps were sufficient to take away the 
seepage water, and the bed-rock was easily exposed, so that its surface 
could be cleaned, washed, and prepared for the concrete. The base of 
the pier was built up nearly to the ground surface. 

The other two piers also rest on bed-rock, and both were sunk by 
using pneumatic caissons. The caissons were similar ; that of the chan- 
nel pier was built on shore and floated to place, and that of the east 
shore pier was built in place. 

The caissons were of timber and had steel cutting edges. Timbers, 
12 in. square, of Oregon fir and Southern pine were used; and were 
thoroughly bolted and drifted together. The walls of the chamber were 
3 ft. thick, and the roof was 6 ft. thick. Every other stick of the upper 
course of roof timbers was omitted, to give bond for the concrete. 

The outer timbers were vertical, and each was fastened by at least 
two bolts extending into the chamber, besides drifts. Horizontal tim- 
bers were lapped at the corners, log-house fashion, thus avoiding exten- 
sive framing. The chamber was sheeted inside with 3-in. planks. 



PLATE V. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXV, No. 1120. 

HOWARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 





DETAIL OF CUTTING EDGE d X.ftT^'! ^ ^s"" 
13". 'i'l-rtSf/J^ Mji'-K'iB- 

^ ^^ 'H*- 8r3i ?i'-'^Jj'D„r,Boll I'lj I5. 



i 



-5-3p.-at 12' 5-0 « 2 






ffi ^ 



w^ 



5ii ► i^' DRIFT BCLT 



la'i tj^'yilte 



HALF END HALF 

&UVATION SEGTION 



^ u'a 4<j; — -^no"— ^^X 4io_4r ^^ ' ^-^HX'1«_3^'q'- 




CETAIL OF 4 X^'Z STRAP 



DETAILS OF PIER 2, KAW RIVER BRIDGE 

6TH STREET VIADUCT, KANSAS CITY. 



SECTION ON AB 



THE SIXTH STEEET VIADUCT, KANSAS CITY 67 

beveled for caulking, and fastened with boat spikes driven through 
bored holes. Every bolt, and the shanlc of every spike and nail project- 
ing into the chamber, was carefully wrapped with oakum before being 
driven entirely home or finally tightened up. On the outside all bolt 
heads were countersunk, thus leaving a smooth surface. Two cross- 
bracing trusses, with adjustable tie-rods, stiffened the walls of the 
chamber. 

A false bottom was put in the caisson for the channel pier, in order 
to float it to position; and, for this pier, a light tramway, on two pile 
bents supporting a standard-gauge track, was built from shore to pier 
site for conveying materials. Concrete was mixed on shore and carried 
out on push-cars in skip boxes. A derrick barge handled the boxes 
and the timbers for the crib. 

The crib, above the caisson proper, was of one thickness of 12-in. 
timbers sheeted outside with 2-in. plank nailed on vertically. 

The usual man and material shafts, with air pipes, blow-out pipes, 
etc., were installed. The locks used were very simple, being made by 
merely putting cast-iron diaphragms with drop-doors at each end of a 
section of ordinary shaft. The one objection to locks of this kind is 
the necessity of an inside lock tender, and this is not entirely a fault, 
as men are much less likely to be "plugged" by a lock tender with them 
in the lock than by a man outside who operates without definite knowl- 
edge of conditions inside. 

The material excavated was sand, with occasional deposits of silt, 
and very many boulders, so that most of the material was blown out 
dry. Numerous relics of the great flood were found in the river bed, 
indicating deep scour. The metal, notably some street car rails and 
some parts of steel bridges, was carried down and embedded in the 
concrete. 

Considerable blasting was necessary to prepare the bed-rock for 
piers. A ditch 3 ft. wide was blasted out around the entire perimeter 
of the cutting edge, so that the concrete is at no place less than 2.5 ft. 
below the bed-rock surface. The middle portion of the rock was not 
removed, but was washed clean, as also was the entire inside of the 
chamber. 

Concrete for filling the chamber was of standard proportions, and 
was mixed outside and locked into the pier through material shafts 2 ft. 
in diameter. It was placed in the dry, and carefully tamped to place, 



68 THE SIXTH STREET VIADUCT, KANSAS CITY 

especially against the ceiling, the boulders and pebbles carried down 
being mixed in. When the remaining space had been reduced to such 
small dimensions that the men could no longer work, the air was cut 
off, all doors were allowed to swing open, and into all shafts a thin 
grout was quickly poured from above until it was some feet above 
their lower ends. Ordinary concrete was deposited to fill them. 

Coffer-dams were not used, but the shafts were built up as soon 
as the crib was of the estimated required height, and from this point 
down the sinking was handled very carefully so as to keep the shafts 
in correct position. For the last 16 ft. of sinking for Caisson No. 2 
the shaft moved out of position only J in. at each end, and that in 
opposite directions. 

The sinking of Caisson No. 1 was complicated by the presence of a 
30-in. high-service water main lying so close to the pier site that the 
caisson had to be rotated to clear it. Pile bents were driven under the 
pipe for a support, but these went down, and a sling was put around it 
and passed up over an A-frame to a distant anchorage. All attempts 
to rotate this caisson to its correct position after it was below the 
obstruction failed. 

These caissons were models in construction; they were well lighted 
by electricity, and constant and adequate air pressure was maintained. 
Natural gas was available for fuel. 

For Pier 3 the shaft form was of 2 by 6-in. matched dressed floor- 
ing, which was placed horizontally on the sides of the pier and fastened 
to vertical studs, but on the curved ends it was placed vertically and 
drawn into position by f-in. rovmd iron hoops. These hoops were passed 
through holes in the vertical studs at tangential points and drawn tight 
by nuts threaded on their ends. Interior circular templates were placed 
near the bottom and at the top of each 10-ft. section of the end staves 
to hold them to position, but were removed as the concrete was built 
up. Ties of heavy malleable wire were put across between pairs of 
studs, and twisted tight to bring the sides to proper place. These ties 
were placed with such frequency as to hamper the placing of concrete, 
but could not be dispensed with. The building of a 10-ft. section of 
form and of filling it with concrete alternated, so that the pier was 
stripped below as it was built up above. Forms of this kind are not 
considered entirely satisfactory, being easily warped and distorted, and 
often delaying the concreting for readjustment. 



THE SIXTH STREET VIADUCT, KANSAS CITY 69 

The forms for the other two piers were made of the same planking, 
in 8-ft. lengths, but this was placed vertically all around the pier. 
Rigid frames, 4 in, thick, with curved ends of built-up plank nailed and 
bolted together, were spaced 4 ft. apart, and supported the sheeting. 
Five i-in. tie-rods held the sides of the frame together, and cross-brac- 
ing at each end and on each side provided against lateral displacement. 

The lower edges of the copings were formed to the required radius 
by sheet-iron corner moulds, and the upper edges by the manipulation 
of a curved trowel. The tops of copings were brought to a smooth level 
finish by floating with mortar before the concrete had set. 

3.— Field Engineering. 

The field engineering included surveys, pier location, inspection, and 
general supervision of all construction. 

The alignment of the viaduct was governed somewhat by property 
boundary lines. The structure lies in three tangents, aggregating about 
6 900 ft., and three curves, each of about 1 900 ft. radius, and totaling 
1 500 ft. in length. 

Exact surveys and final measurements were made to locate street 
crossings, railways, buildings, etc., before much work was done on 
detailed drawings. Base lines, paralleling the chosen center line 
tangents were carefully marked on the ground, and the angles at their 
intersections determined. Points on base lines, about 400 ft. apart, 
were marked by large wooden hubs driven down flush. In the boggy, 
quaky ground these were 4 by 6 in. and from 3 to 5 ft. long, but in the 
firmer ground not so large. A piece of sheet copper was nailed on 
top of each hub, and the exact point was indicated by the intersection 
of two scribed lines. Each main hub was carefully referenced by 
marking two intersecting lines. Exact distances from hub to hub 
were determined either by direct measurement or by triangulation. 
About one-third of the entire length was triangulated. For the 
systems of triangles, suitable base lines, usually intersecting with the 
main line, were marked and measured. Measurements were made as 
follows : Between two main hubs, 2 by 4-in. marking stakes were 
solidly driven down, about 98 ft. apart; between these, at about 6-ft. 
centers, small supporting stakes were placed. All these were sawn off 
or driven down to conform to the grade from main hub to main hub. 
When necessary a narrow ditch was dug to contain the stakes, it 



70 THE SIXTH STREET VIADUCT, KANSA.S CITY 

being desired to support the tape on them and not on the ground. 
A 100-ft. steel tape, which had been compared with a standard to 
determine its constants, was used. The required tension was secured 
with a spring balance, and the distance was marked by thrusting into 
a marking stake a knife blade with its shai-p edge at the desired 
graduation. The knife was left there and measured from in placing 
a second knife ahead, and so on until the second main hub was reached, 
and there the tape was read to an approximate thousandth. Thus any 
number of repetitional measurements could be made without the con- 
fusion resulting from making each intermediate point with a non- 
erasable mark. About six measurements were made for each hub dis- 
tance. The extremes varied from the mean about 1 in 80 000. The 
base line at the Kaw River was conveniently measured on the deck 
of an adjacent railway bridge, and the triangulation points there con- 
sisted merely of punch marks on the end floor-beams at fixed shoes. 

The measurement of angles was done by a method of repetitions 
and reverses which eliminated instrumental errors in the mean of each 
series of readings. Enough readings were taken at each point to 
make the triangles close within 1.5". The largest triangulated dis- 
tance was 1 131 ft., and was determined by two triangle systems, so 
that the independent calculated lengths checked within ^^^ i^- The 
difference in the elevation of the two end points was 65 ft. 

After all distances had been determined, the station and plus 
of each hub was calculated^ and locations for the pedestals were 
measured readily from the nearest hub. 

For the location of river piers, the true angular distance from each 
base line was calculated, and a temporary fore-sight was established 
by a single carefully-turned angle which was then measured exactly. 
The distance of the temporary fore-sight from the true line was 
calculated, and the point in the hub was then moved the necessary 
distance (f to li in.). Wlien the two reference lines and the bridge 
tangent were projected to the piers they intersected within a y^jy-in. 
circle. 

For pedestals, and for piers where possible, reference lines at right 
angles to the tangent were marked by several hubs well driven down. 
The longitudinal center lines of each row of pedestals were also suitably 
marked at intervals. Elevations along the entire line were determined, 
and numerous carefully checked bench-marks established. 



PLATE VI. 

TRANS. AM. SOC. CIV. ENQRS 

VOL. LXV, No. 1120. 

HOWARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 




DETAILS OF COLUMNS AND BRACING 

6TH STREET VIADUCT, 

KANSAS CITY 



THE SIXTH STREET VIADUCT, KANSAS CITY 71 

The inspection of construction embraced every detail, and daily 
records of the progress of each kind of work were prepared. For 
instance, not only did an inspector watch the placing of concrete, but 
a second man watched the entire process of mixing. There was an 
ample number of men to make inspection of stone, timber, cement, 
sand, pile-driving, forms, concreting, erection, riveting, painting, pav- 
ing, decking, etc., as has been noted previously. 

4. — Erection of Superstructure Metal. 

About 88% of the total length of the viaduct is of practically 
duplicate panels, and, in all details of construction, particularly in the 
erection, this characteristic feature permitted economic handling of 
materials and labor. 

Trestle Erection. — The traveler used for the ordinary structure con- 
sisted of a pair of derricks mounted on a frame arranged to run on 
track rails laid on the longitudinal girders, the masts thus being 33 ft. 
from center to center. The details of this traveler are clearly seen on 
Fig. 2, Plate IV. The masts were double, one stiff and one seated 
to turn. Each was a 14-in. square timber 40 ft. high, and the booms 
were built up of spliced timbers to 14 by 26 in., and 60 ft. long. The sills 
and cross-sills were 8 by 16 in. All connections were made by steel plates, 
and steel derrick fittings were used throughout. The equipment con- 
sisted of two 2-drum, 4-spool, hoisting engines of 20 h.p., each handling 
one boom with its lines and runners. In the ordinary structure the 
members varied in weight up to 10 tons, which one boom could easily 
handle. 

A service track was laid between the pedestals nearly the entire 
length of the viaduct, and the steel was unloaded alongside. The 
intention was to distribute immediately every piece to its proper loca- 
tion, but, owing to the operations of the substructure contractor, this 
could not be done, and it was necessary to store and redistribute them. 
The lighter metal was unloaded by a locomotive crane, the larger 
girders by yard derricks, or by hand upon cribs. 

It thus followed that a considerable part was erected from metal 
already distributed, but also a large part directly from the cars as 
they were run down from the yard. There was little difference in the 
rate of erection in the two cases, the former being retarded by the 
necessity of removing superposed pieces and hitching to material some- 



72 THE SIXTH STREET VIADUCT, KANSAS CITY 

what indiscriminately piled, and the latter by reason of the turbulent 
operation of service cars over the temporary tracks. Fig. 2, Plate IV. 

The pedestals were finished ^ to | in. low, and four 2 by 4-in. oak 
shims of suitable thickness were used to bring each column to the 
desired elevation. The anchor-bolts had been located so accurately 
that they sufficed to line up the columns, additional marks being 
unnecessary. 

The usual order of erection per panel was as follows : Two columns 
were set up at once, one by each boom, and braced up by the anchor- 
bolts; the connecting floor-beam was placed on its shelf-angles, and 
bolts were put in one-third of the open holes. The cantilevers were 
then put up and bolted, after which the main girders were raised 
singly, both booms being used for the heavier girders. The stringers 
were then laid on top, to be distributed later, and the traveler was 
immediately moved forward over the new panel, ready to set the next 
columns. The maximum day's run was six panels, about 275 tons. 

The light material in the deck was put in place in conjunction 
with the riveting, the stringers by using timber dollies, the bracing, 
etc., on a push-car running on the flanges of two adjacent stringers. 
However, as the traveler proceeded, it placed all sway bracing, tower 
bracing, and laterals. 

Certain long and heavy girders required some variation from the 
usual order of handling. Where the columns ahead were beyond 
reach of the booms they were raised by a gin-pole or a gallows-frame, 
which then served to lift one end of the girder, the other end being 
lifted by the traveler. The longest girder in the structure (107 ft.), 
weighing 62 tons, was erected in this manner, the gallows-frame con- 
sisting of a pair of 18-in. square timbers, 60 ft. high, connected with 
bracing and caps, equipped with 4-sheave steel blocks reeved with 
steel cable. Fig. 1, Plate VII. 

Because of buildings and railway tracks under the eastern seven 
panels, the girders could not be handled from below, but it was 
necessary to run the traveler back, free of obstructions, pick up a 
girder, set it transversely on the traveler track, move it forward by 
picking lap, booming out, and shifting the traveler, and finally to 
swing it around longitudinally to its final position. The columns in 
the buildings were dropped down through holes cut in the roof, with 
small trouble to the occupants, who, however, did not object to stepping 



THE SIXTH STRJ'^ET VIADUCT, KANSAS CITY 78 

outside for a few minutes while each girder was being manipulated 
above their office rooms. The girders immediately adjacent to Blufi 
Street extend through slots or pockets cut in the retaining wall which 
supports the street, and rest on pedestals 15 ft. back of, and thus 
entirely independent of, the wall. The tracks below these girders 
carry all the passenger traffic into the Union Depot, therefore unusual 
caution was necessary in erection to guard against possible interruption 
of train service. 

Erection of Truss Bridges. — The falsework consisted of 7-pile bents 
every 30 ft. under the panel points, and 6-pile bents midway between, 
the piles being well driven into the river bed. Four 8 by 16-in. 
stringers were placed under each truss, and a traveler track was sup- 
ported on two runs of 12 by 12-in. timbers. This traveler was com- 
posed of two derricks rigged together, and running on a track between 
the trusses in the plane of the bottom chords. Beside the traveler, also 
between the trusses, a standard-gauge track was laid for moving out 
the metal. 

After fastening down the fixed shoes, the bottom chords were laid 
on the camber blocking for the full length of the span. Beginning 
at one end, the web members, chords, floor-beams, and laterals, were 
erected in regular order. 

The parts of these trusses are heavy, and the connection plates large 
and stiff, so that fitting together the various members was tedious 
work. A clearance of tg in. was provided, but, in a member 22 in. wide, 
made of four |-in. angles and a |-in. plate, which had to be entered 
from 5 to 6 ft. into a second member equally stiff, slight irregularities 
would cause binding. Practically every diagonal and chord section 
had to be pulled to place with a runner. 

It was apparently impossible to do all the field riveting before the 
time of the expected washout of the falsework, and, in order to sup- 
port the spans in that contingency, drift-pins were tightly driven 
into one-third of the open holes, and fitting-up bolts into one-third of 
the open holes of all main connections. Fortunately, it was not neces- 
sary to trust to this expedient, as the expected high water did not come. 
It is evident the spans would not have fallen, but experience showed 
that there would have been a slight movement at each point, which, in 
the aggregate, would have caused trouble. 

The trusses were set carefully to correct the camber which was 



74 THE SIXTH STREET VIADUCT, KANSAS CITY 

maintained until the erection was complete. With the passage of the 
time required for riveting, the falsework settled at points, permitting 
the entire truss to settle. Attempts were made to jack up the trusses, 
but the only result was a pushing down of the piles under the wedges. 
A regular camber of about one-half as much as estimated was finally 
secured. With the complete dead load and such live load as has 
been on the spans, the bottom chords are just about straight. The 
serious consequence of this settlement, however, was in the tightening 
of the drift-pins. These were not taken out until the rivets had been 
driven in all the adjacent holes, and then it was found there had been 
enough movement to pinch them tightly in place, requiring much 
energy and effort to remove them. Most of these pins had been driven 
toward the inside, and the cramped quarters prevented efficient striking 
to back them out. Many had to be drilled out, and, as they were of 
very hard steel, it was difficult to keep the drill point on the pins and 
prevent it from running off into the softer metal. 

The lower chord splices of these spans were riveted first, and then 
the ends of the diagonals ; but the upper chord splices and the intersec- 
tion points of the diagonals were not reamed or riveted until the span 
had been swung. 

The similar trusses for the 147-ft. span were erected on a falsework 
of framed bents set on the ground. There was no trouble in main- 
taining the full camber in this span; and, as the members are much 
smaller than for the Kaw spans, their erection was quite simple. 

Riveting. — The contractor sought and received permission to ream 
all field holes in the field, thus making it unnecessary to assemble 
in the shops; and the metal as shipped out was merely sub-punched. 
Because of this, and since, for the unrestricted movement of various 
gangs, it was desirable to keep them separated, a very large amount of 
scaffolding wns placed and retained for some time, but it was of 
simple character. 

The joints were first thoroughly fitted up, and bolts were put in 
half of the holes; the other holes were then reamed, the bolts were 
shifted to the reamed holes, and the reaming was completed. 

Each joint was thus fitted twice before the rivets were driven, but 
trials of reaming and then riveting half a joint, and then reaming and 
riveting the remaining half resulted in the interference of the various 
gangs, and better progress was made by the procedure given. 



THE SIXTH STREET VIADUCT, KANSAS CITY 



75 



Holes for'Mo. Turned Bolts 



GENERAL DETAILS OF SHOES OF 

KAW RIVER BRIDGE 

6th STREET VIADUCT, 

KANSAS CITY. 







Draim hole 



CAST-STEEL BASE 

FIXED SHOE FOR WEST SPAN 
Fig. 5. 



76 THE SIXTH STREET VIADUCT, KANSAS CITY 

Neglect in fitting up caused the majority of cut-out rivets, for if 
the plates were not drawn tightly together small particles of reamer 
drillings would work in between them and cause successive rivets to 
loosen those previously driven. Natural conditions were seized upon 
to assist this work; for instance, the expansion joints were regularly 
rammed full of steel wedges, thus gaining the thrust of expansion to 
force together columns and girders. 

The reaming was done with Little Giant, No. 0, pneumatic drills, 
equipped with 4-flute, tapered reamers having barrels at least as long 
as the leng-th of the hole. Water, with a little soap, was used as a lubri- 
cant, and was kept constantly flowing on the bits. A knuckle-joint was 
tried for points where the reamer could not be held normal to the metal, 
but gave indifferent results; so, when unavoidable, holes were reamed 
on a slight bevel, but they were required to be perpendicular to the 
direction of stress. There was some trouble in making all holes of the 
same diameter, for the reamers had to be held very steady to accom- 
plish this. The ordinary run for two men was from 200 to 250 holes, 
§ in. in diameter, with a 3-in. grip, or equivalent, in 8 hours. The 
smooth, clean holes permitted the rapid driving of rivets without fur- 
ther drifting or fitting, and without the delay incident to entering hot 
rivets into imperfect holes. This was highly essential, especially in 
such cases as the lower chord splices of the Kaw River spans, where 
the rivets are 1 in. in diameter and have a 6-in. grip. 

Cleveland air hammers, with a 9-in. stroke, operating under a 
pressure of 110 lb. were used throughout. Grip dollies were required 
wherever they could be utilized, and hand-forges were used almost 
exclusively. With one heater and two riveters, 550 rivets, of l-in. 
diameter and 3-in. grip, were driven in 8 hours. 

The greater part of the riveting was open, easy work, although 
there were marked exceptions. The rivets in exi^ansion pockets re- 
quired special, crooked-jamb, joint dollies, in order to pass around the 
end stiffeners and rest on the next adjacent stiffener. 

The bracing of the motorway stringers was connected by |-in. 
rivets driven by hand without reaming — in fact, it was found that the 
lug-angles on the beams were set so high that the rivets were under 
the flanges, thus precluding the use of an air hammer. Very fair 
results were achieved, but it would have been much better to have had 
wider lug-angles. 



PLATE VII. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXV, No. 1120. 

HOV^ARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 




Fig. 1. — Gallows Frame used in the Erection of Long Girders. 




Fig. 3. — Concrete Floor-Trough and Cover Style of Form. 



THE SIXTH STREET VIADUCT, KANSAS CITY 77 

In the main structure there are about 300 000 field rivets. While 
accurate costs are not available, the following figures are a close 
approximation: Reaming cost 7 cents per hole, of which 5 cents 
was for labor and 2 cents for plant and tools; driving cost 8i cents 
per rivet, 8 cents for labor, and 4 cent for plant and tools. It is 
possible that too much of the plant cost is charged to reaming and too 
little to riveting. There were reported cut-outs of 3J%, but apparently 
all the rivets were not accounted for, and 5% would be nearer the true 
figure. 

Grouting Columns. — The space below each column base was filled 
with grout. A cement mortar of ordinary consistency was first pre- 
pared and a little daubed under the edges of the base plate around its 
entire perimeter and finished off neatly. After a day's time there would 
remain a tight dam all around, pierced, however, by a small hole on 
each side. The nuts and washers on the anchor-bolts were temporarily 
removed. Grout, diluted to a pouring thickness, was made in the pro- 
portion of about 1 of cement to 3 of fine sand containing a little clay. 
Several bucketfuls of the mixture were poured, as quickly as possible, 
into the anchor-bolt holes and the central hole until it ran out of all 
the "air holes," and then these were promptly daubed up. Mortar was 
added to fill entirely the side pockets; the foot of each column was 
filled with concrete finished off with mortar. 

This method was highly satisfactory, as it gave a full and even 
bearing under the entire surface of the base plate, as was proven by 
lifting up columns which had been grouted. 

For large solid castings, as under the Kaw spans, grout could not 
be used. The bases were set on steel shims about f in. above the 
concrete. A stop was put at one side, and small quantities of almost 
dry 1 : 1 mortar of cement and sand were put under the opposite side, 
pushed clear across, and rammed with a piece of i-in. plate. The 
entire space was filled from one side, thus insuring the absence of 
voids. 

Had holes through the casting been provided, the former method 
could have been used with superior dispatch and ease. 

Painting. — The steel received one coat of paint in the shops, and 
presented a good even coating when unloaded at the site. As it was 
erected, all abutting parts and surfaces inaccessible after erection 
were coated with a heavy paint cement daubed on in generous measure. 



78 THE SIXTH STREET VIADUCT, KANSAS CITY 

Upon the completion of the riveting of a section of the viaduct, the 
entire surface of the metal was examined, carefully cleaned vpith wire 
brush and scraper where necessary, and the heads of field rivets and 
all areas where the shop coat was scraped off or unduly marred were 
touched up with paint. Reaming borings added considerably to the 
ordinary work of cleaning. When these spots had dried, the first field 
coat (of green paint) was applied. The paint was delivered ready 
mixed, and was of rather stiff consistency, requiring, for proper applica- 
tion, small amounts thoroughly rabbed on by the vigorous use of a stiff- 
bristle brush held almost normal to the surface. 

When this coat had dried, requiring a varying time, but determined 
when it became impossible to "rub off" the dried surface skin and 
expose wet paint below, the final coat was put on. The same care 
was used in application, and the entire structure was finished up 
progressively. 

The inspectors then examined the work carefully, and remained 
with the finishers until the painting was acceptable. Small hand- 
mirrors were used to reflect the sunlight and illuminate dark corners 
and shadowed areas, thus exposing neglected patches almost impossi- 
ble to locate by any other means. In many places, too, as, for instance, 
inside a column, it is much more convenient to observe the reflection 
in a mirror than to attempt to view the parts directly. 

No painting was done in wet, or wet and freezing, weather, but the 
work was never stopped because of low temperature, provided the sur- 
face to be coated was dry and free from sleet or ice. 

5. — Viaduct Floors. 

The general details of the reinforced concrete floor are shown on 
Fig. 3. The average slab thickness is 6 in. above the stringers, the 
stringers being entirely encased, both to protect them from the smoke 
of locomotives below and to secure the added strength of the filled 
haunches. The amount of reinforcement is considerably in excess of 
actual load requirements, in order to provide for the temperature 
stresses of the concrete. It consists of transverse and longitudinal 
bars, near both the upper and lower surfaces of the slab, of extra bars 
over the cross-beams, of numerous vertical bars in the curbs, and of 
wire netting around each I-beam stringer. To support the hand-rail 



THE SIXTH STUEET VIADUCT, KANSAS CITY 79 

posts, bolts project through tlie sides, the lower bolts passing throvigli 
the I-beams. 

In order that the usual paving could be continued without inter- 
ruption over the points of expansion, a flexible connection was de- 
vised consisting of U-shaped plates bolted side by side. Thus a sort 
of bellows trough is formed, about 6 in. high and 2 ft. wide. One such 
bellows extends across the roadway at each expansion joint of the 
stringers, and is riveted to each stringer on both sides of the opening. 
The plates are thin enough to spring easily under the movement of the 
steel. The concrete is finished up to each side of the bellows, and 
fastened to it by bolts. These troughs were to be filled with a rather 
soft asphaltic material, the flow of which would accommodate it to the 
varying position of the containing plates. 

The specifications provide for concrete made of cement to be sup- 
plied by the company; of clean, hard, broken stone "free from all dust, 
dirt, shale, rotten stone, or other objectionable material", to pass a 1-in. 
ring ; and of Kaw River sand in the proportions 1 : 3 : 5 by volume, loose. 
To provide a bond for the contemplated wearing surface of asphaltum 
(which was to be laid without the usual binder course) it was required 
that clean stones be sprinkled on the surface of the wet concrete, and 
be tamped into it enough to embed them partially, yet permit them to 
project a little. Other principal specifications were: That the entire 
thickness of the floor was to be deposited in one continuous operation 
so as to avoid laminations or dry horizontal planes between consecutive 
layers; that all surfaces were to be finished neatly to the exact required 
dimensions; that all bolts and embedded castings were to be set to 
exact position; that reinforcing bars were to be spaced accurately; 
and, finally, that the contractor was to clean all stains and concrete 
droppings from the metal work. 

The motorway stringers at the side of the roadway were used as a 
runway for the concrete mixing equipment. A standard-gauge track 
was laid on these stringers for some distance, to accommodate the 
contractor's plant, which he placed on several railway flat cars. This 
plant consisted of two large elevated hoppers, connected by a cable 
carrying a conveyor bucket, and a belt conveyor for handling finished 
material, and similar appliances. Much time was consumed in the 
arrangement of this equipment which was soon found to be unsuitable. 
The company's engineers then designed a new and simple plant, the 



80 THE SIXTH STREET VIADUCT, KANSAS CITY 

old elaborate one was dismantled, and five of the smaller plants were 
quickly built and put into successful operation. 

The plant is shown in Fig. 1, Plate VIII; it consisted merely of a 
platform of timbers, mounted on wheels to run on the motorway 
stringers, supporting a mixer, a derrick, and an engine. The derrick 
raised the aggregates from below with a skip box, and a metal-lined 
chute extending over the roadway received the concrete from the 
mixer. The total cost of one plant, including all fittings except the 
mixer and the engine, was $700. 

Forms. — Several arrangements of forms were tried, and, as these 
made up a considerable part of the floor concrete cost, they merit special 
attention. 

The contractor first furnished all-metal forms, made of corrugated 
sheets pressed to shape and covered with light sheet iron. They were 
made in sections 5 ft. wide and about 5 ft. long, the longitudinal junc- 
tion of various sections being immediately under the stringers. Along 
each of these lower edges there was fastened a small angle having 
holes in its vertical leg. Keys were provided to pass through these 
slots to fasten adjoining forms together laterally. Provision was made 
for the forms .to lap about 1 in. longitudinally, but there was no way 
of fastening them together. To support these forms, three joists were 
placed from cross-beam to cross-beam, and three sections of form were 
laid thereon. Numerous imperfections were soon evident. The form 
had almost no lateral stiffness, each unit having an individual shape. 
At best, the surface presented was not true; the correct distance was 
not left on each side of and below the I-beams; there were openings 
and cracks at the transverse joints, as these could not be made to lap as 
intended; many pieces had to be held during adjustment; and, also, 
the forms were easily displaced by the ordinary walking about and 
working on them. 

Owing to the non-delivery of the metal forms, the contractor had 
already begun to build wooden forms of much the same type, that is, of 
sections to meet under the stringers and of such length that three 
were put between adjacent cross-girders. They were built of 1 by 4-in. 
frames spaced 18 in. from center to center, sheeted with J-in. boards, 
and covered with No. 14 sheet iron. These forms were hung on the 
stringers with two J-in. round, soft iron rods on each side. After ad- 
justing the form to position, the hangers were bent around the upper 



PLATE VIII. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXV, No. H20. 

HOWARD ON 

THE SIXTH STREET VIADUCT, KANSAS CITY. 




Fig. 1.— Arrangement of Concrete Plant for Laying Floor. 



* - -1 


7 






I 




Fig. S.— Finished Flooh of JIulberry and Hickory Street Approaches. 



THE SIXTH STREET VIADUCT, KANSAS CITY 81 

flanges of the I-beams. To remove the forms, the hanger rods were 
cut with heavy clippers close to the sheeting. These wooden forms 
were light and stiff, but they, too, showed objectionable defects. There 
were no means of really fastening the various units together, and there 
were presented, not only the transverse cracks of the metal forms but 
also a longitudinal crack under each stringer, and these latter were al- 
most inaccessible for inspection or adjustment. Nothing was found 
to prevent water and cement from leaking out. 

The company's engineers then designed a form, and the use of the 
others was discontinued as soon as they could be replaced. The finally 
acceptable form consisted of two parts: U-shaped troughs which were 
placed under the stringers; and covers which extended from trough to 
trough between stringers. They were of wood, 1 by 4-in. frames, with 
i-in. or f-in. sheeting, and covered with No. 16 or No. 18 sheet iron. 
The troughs were 13 ft. 10 in. long, or just enough to go easily be- 
tween the cross-beams; the covers were of various convenient lengths, 
from 4 to 6 ft.; and the outside curb faces were 18 ft. long. To sup- 
port the troughs, i-in. round iron hangers were used, a loop-shaped 
piece being put around the 2 by 4-in. sill on the bottom, and each 
leg being bent over the upper flange of the I-beam. Four or five ties 
were used for each section. To hold this trough in correct position and 
" prevent rotation, briquette spacers were used. These were made to fit 
the lower curve of the form, and had a shoulder on each side at the top 
to fit the lower flange of the I-beam. 

A nest of wooden moulds was provided, and briquettes of sand and 
cement were made in advance of the need, so as to be thoroughly hard- 
ened. One briquette was placed at each end of each trough (thus 
closing it against leakage) and one under each hanger. When the 
hangers had been drawn reasonably tight and bent over, the trough 
was so rigid that a man could walk along one edge. 

The cover forms, it was found, could be handled with a lap at one 
end, and, when the troughs were in position, could be rapidly laid, 
producing a continuous unbroken surface. The small remaining ver- 
tical hole (8 by 12 in.) at each floor-beam was covered with a piece of 
No. 8 sheet iron. The forms for the outside faces of the curbs were 
made of l-in. plank. These were stood on the already erected curb 
trough, adjusted to position, and nailed lightly. A horizontal hanger 
held them laterally, and they could be made truly vertical. 



82 THE SIXTH STREET VIADUCT, KANSAS CITY 

These forms proved to be highly satisfactory, presenting smooth, 
rigid, correctly-shaped surfaces, with very few chances of leakage. 

They were easily altered to meet conditions, and stood rough treat- 
ment well. Their erection and removal was simple, any part being 
handled by two men without extra scaffolding. Fig. 2, Plate VII. 

The briquette detail was one of the most advantageous features, 
and was developed by a prior use of wooden blocks, which gave trouble 
by breaking out the concrete as they were removed with the form. The 
cement briquettes remained tightly in place, and panel after panel of 
floor was built without a place to touch up below the beams. 

The cost of building 600 lin. ft. of these forms for the 30-ft. roadway 
was: 

Lumber $1 KiO 

Sheet iron 650 

Nails and sundries 165 

Labor 1 300 

Contractor's profit 325 



$3 600 



The first cost was $6 per lin. ft., or $0.20 per sq. ft. between curbs. 
Forms used five times cost 4 cents per sq. ft., or $1.30 per cu. yd. Many 
of these forms were used oftener than five times. The fact that they 
were built in great haste doubtless increased their cost somewhat. 

Making and Placing Concrete. — Owing to difficulty in securing a 
sufficient quantity of 1-in. stone from local crushers, the use of flint 
screenings from zinc mines was permitted. Although this material is 
clean, hard, and sharp, and is somewhat graded in size, needing but 
little sand, there are objections to it. A few moments after it had been 
deposited quite wet, the water appeared to drain away, leaving the 
upper stones clean and dry, without any mortar coating whatsoever, and 
the entire mass appeared to be inert and "dead". This appearance was 
evident even while the concrete was still green enough to be dug 
through with a stick. After several days the main body became hard 
and compact, but on top there was perhaps I in. of loose stones. Various 
mixtures were tried, to see if the concrete could be made to remain 
"quaky" for a time, but it could not be done. Yet experiments on the 



THE SIXTH STREET VIADUCT, KANSAS CITY 83 

tensile strength gave creditable results, and investigations by cutting 
through slabs thoroughly set showed the concrete to be very hard. About 
half the floor concrete was made of this flint, and the remainder of 
crushed limestone. The latter, of course, could be flushed with mortar, 
and gave normal results. 

All concrete materials were delivered on the ground under the via- 
duct. Skip boxes were filled with a batch of correctly proportioned 
aggregates and raised to the mixer above. The concrete was dumped 
into a trough and slid down to be distributed over the floor. To ex- 
pedite the filling of the small space below the I-beams, cement mortar 
was used to a height of about 2 in. above the lower flanges. 

The reinforcing -bars were spaced and held in position as follows : 
Eound |-in. iron rods were laid longitudinally on the forms, about 5 ft. 
apart, and on these the lower transverse bars were placed. The lower 
longitudinal bars were then laid in position and fastened at numerous 
intersections with wire ties. Wlaen some 3 in. of concrete had been spread, 
a rope was fastened to the ends of the |-in. rods and a strong pull 
slipped them out longitudinally from the finished work. A very minute 
inspection of the whole floor failed to show a single instance of a lower 
bar exposed. 

Combined supports for the upper bars and guides for finishing the 
upper surface consisted of 2-in. joists having small angles fastened on 
each side and a series of hooks in the lower edge. They were placed 
longitudinally with the structure and 10 ft. apart, and were held on 
small blocks resting on the stringers. The hooks supported the trans- 
verse upper bars in correct position. A gauge, sliding along the angles, 
gave the desired surface shape to the concrete. 

The concrete was first tamped with a large flat tool and, after 
hardening somewhat, was retamped with a corrugated tool. The latter 
was made of wood with parallel cleats, | in. wide, 1^ in. deep, spaced 
2 in. apart. The cleats were sharpened slightly. With proper manipu- 
lation this produced light corrugations, in varying directions, of -^j. 
to i in. variation, thus giving a satisfactory bond for the asphaltum. 
Scattering loose stone and tamping it in was tried, but the former 
roughening was superior. Neither method gave entire satisfaction 
where flint was used. 

The plants were moved forward as each section of about 15 ft. of 
roadway was completed. On an average, the five plants placed from 



84 THE SIXTH STREET VIADUCT, KANSAS CITY 

300 to 350 lin. ft. per day ; the best day's run for one gang was 98 lin. 
ft., or 90 cu. yd. 

The floor was finished level to the edges, and the curbs were built 
later, in the usual way. Bars projected upward into this curb every 
12 in. and bound it to the body of the work. 

With all types of forms, and to a pronounced degree with the first 
designs, there was considerable leakage from the wet concrete. This 
dripping water carried such a quantity of cement as to leave a whitish 
deposit on the metal where it drained, and in certain cases spattered 
down on the ground under the structure. Probably the actual loss of 
cement was small, but although it was easy to know how much cement 
went into each yard, or each day's run, no scheme was discovered to 
ascertain how much actually remained in the concrete. This cement 
was not cleaned from the metal promptly, and the cleansing later 
proved troublesome. The plan was to wash the metal with a strong 
stream of water immediately after the concrete was placed and until it 
ceased to drip. Adequate pumps and lines were not installed to do 
this, and finally much cleaning had to be done with wire brushes and 
scrapers, necessitating some repainting. 

The entire concrete floor was built in 67 working days, during which 
there was some freezing weather. The contract provided for heating 
any or all concrete materials, but none was heated other than some sand 
and this merely to thaw frozen lumps. However, wagon sheets were 
spread over the green concrete each night, and, when hard weather 
was expected, these were covered with stable sweepings. No bad 
effects were found on the floor proper, but some portions of curb- 
ing were frozen and had to be torn out and replaced. Some of the con- 
crete was laid when the temperature was 31°, and the following days 
were about the same, yet in 10 days' time asphalt rollers were run over 
the work. 

The 30-ft. roadway contains 24.6 cu. ft. per lin. ft. and the 38-ft. 
roadway 33 cu. ft. per lin. ft., or an average for both of 0.84 cu. ft. 
per sq. ft. of area between curbs. The average amount of reinforcing 
is 2.67 lb. per sq. ft. 

The main viaduct contains about 8 000 cu. yd. of concrete. The 
contract cost, exclusive of cement and bars, was $6,085 per cu. yd. The 
cost of completed floor, not including paving, was about 36J cents per 
sq. ft. 



THE SIXTH STREET VIADUCT, KANSAS CITY 85 

Paving. — The paving is of Bermudez asphalt. The following re- 
sults are from average analyses : 

Specific gravity 1.098 

Insolubles 2.0% 

Soluble in 88° naphtha 63.0 

Soluble in 62° naphtha 85.0 

Penetration at 77° Fahr 38° (Dow.) 

Fixed carbon on ignition 14.5% 

Ductility 20. (CM.) 

This hard asphalt was softened by the addition of asphaltic 
petroleum residuum which tested as follows: 

Specific gravity, Baume 11.25 

Boils 83° cent. 

Flashes 173.5° cent. 

From 18 to 20 lb. of this oil were mixed with 100 lb. of asphalt. 
The hardness of the resulting mixture at 77° Fahr. varied from 67 
to 78° (Dow). 

It is rather difficult to secure just the sand desired for asphalt in 
Kansas City, as the Missouri River sand is somewhat water-worn and 
the finer Kaw sands contain some loam. Pulverized limestone and 
cement were added to the sand to give the required 200-mesh ma- 
terial. 

The following mix was taken as a criterion, although the specifica- 
tions gave somewhat more latitude: 

Bitumen 9.5% to 11.5% 

Sand passing 200 6 % " 16 % 

100 8 % " 16 % 

80 10 % « 16 % 

50 14 % " 34 % 

40 

30 

20 

10 

The paving material was given close inspection, both at the plant 
and on the work. Samples, taken daily, were analyzed and sifted, and 
daily directions were given to the mill. On the viaduct, samples 



9 


% " 16 


% 


6 


% " 10 


% 


3 


% " 7 


% 


2 


% " 4 


% 



86 THE SIXTH STREET VIADUCT, KANSAS CITY 

were taken from 1 in every 5 to 15 loads, depending on the rapidity 
of delivery, etc. 

The rolling was done carefully with 6 and 8-ton rollers. Where 
necessary, the surface was ironed, witli especial attention to securing 
adequate drainage. The entire pavement was placed in 20 days. 

The iron trough expansion joints were filled with the ordinary 
mixture. At some of these the wearing surface cracked during cold 
weather, and in the later work thin separators were inserted to make 
these openings straight across the roadway. These joints are acting as 
expected, and are considered very satisfactory. 

Asphalt was not laid on the new embankments, but a Macadam 
pavement on a very heavy base was substituted and finished with oil. 

The contract cost for the asphaltum was $1.05 per sq. yd., and for 
the Macadam, $1 per sq. yd. Thus the concrete floor with pavement 
cost about 48 cents per sq. ft. 

Street-Car Deck. — The ties of the street-car deck are of creosoted, 
short-leaf, Southern pine, 6 by 8 in., laid flat, and spaced 13 in. from 
center to center. Every fourth tie extends across both tracks, thus 
serving to brace the whole deck together and support a possible run- 
way. Guard timbers are dapped 1^ in. on the ties and are set 11 in. 
beyond the gauge. On curves, the ties are on edge, the variation in the 
depth of the daps giving the outer-rail elevation. 

All the timber was surfaced on four sides, and finished to full size. 
The general creosoting specifications were the same as already given 
for piles, from 10 to 13 lb. of pure dead oil of coal-tar being used per 
cubic foot of timber. The dapping, boring, and cutting were done before 
creosoting. The deck timber cost delivered $45 per 1 000 ft., B. M. 

A small derrick car, equipped with a gasoline hoist running on the 
stringers, raised the material to the decks from its distributed position 
below. The ties were set to approximate place and the guard rails 
laid. Wooden rams were used to drive the guard rails down on the 
ties, to shift the ties to fit the guard-rail daps, and to settle them 
on the stringers. Then the hook-bolts and guard-rail bolts were in- 
serted, and the deck was tightened to place with socket wrenches on 
auger handles. 

The rails are of 70-lb., Am. Soc. C. E. section, 30 and 33 ft. long, 
laid with broken joints, fastened to every tie with two ordinary 5^-in. 
spikes. Six-hole angle-bars were used. 



THE SIXTH STREET VIADUCT, KANSAS CITY 87 

The tracks are bonded by two copper bonds at each joint, and set 
up with a screw compressor. For each bond, two |-in. holes were 
drilled through the rail web. Some of these holes were drilled with 
hand-ratchets, but as the rails were very hard and the drilling was 
slow, a machine drill was substituted. This consisted of a 1^-h.p. gaso- 
line engine, set on a push-car, operating a flexible shaft, which was at- 
tached to an automatic-feed drill working in a yoke that litted over the 
rail. This machine cost only $100, in addition to the engine, and gave 
creditable results. It would drill the four holes at a joint and move 
forward in 12 min. A very weak solution of sal soda was used for 
lubrication. 

The cost of laying deck timber was about $5 per 1 000 ft., B. M. ; 
for laying rails about 4 cents per lin. ft. of single track; the work of 
bonding cost about 50 cents per joint. 

The usual overhead trolley is used. The poles are of iron, from 
6^ to 5 in. in diameter and 20 ft. high above the ties. They are placed 
between the two tracks, about 90 ft. apart, and support cross-brackets 
for the wires. 

Hand-Rail. — The hand-rail was designed for strength, simplicity, 
and architectural harmony with the structure. It consists of square, 
cast-iron posts, 9 ft. apart, supporting horizontal rails of ordinary pipe, 
the diameter of the upper one being 3 in. and the other four 2 J in. 
Each post is fastened by two bolts embedded in the concrete floor, the 
upper of which has two nuts, in order to provide for the adjustment of 
the post. The pipe rails pass through holes in the posts, fitting rather 
loosely, and are joined by standard unions and made continuous. 
At each post, and for each pipe, two small steel wedges are driven 
into seats, gripping and pinching the pipe and holding it tightly in place. 

Thus the hand-rail could not be erected until the floor was finished, 
and then the posts and rails had to be put up simultaneously. A small 
two-wheeled truck was used to set up the posts. The forward end of 
this truck had a cantilevered fork on which a post was supported as it 
was drawn on to the projecting bolts. The railing pipes were inserted 
consecutively and jointed together as the successive posts were erected. 

When several hundred feet of railing was in place the posts were 
adjusted to points set by a transit, thus bringing the upper rail to a 
true line. The cost of the hand-rail is included in the general schedule 
price for superstructure metal erected. 



88 THE SIXTH STREET VIADUCT, KANSAS CITY 

Lighting System. — The annoyances of a trip through the bot- 
toms are aggravated by darkness, and an adequate lighting system 
for the accommodation of night travel was considered highly desirable 
for the viaduct. Natural gas is available, and was considered, as also 
were other plans for lighting, but an incandescent . electric system 
was selected. The results are exceedingly satisfactory; the illumina- 
tion is uniform and not glaring, yet so bright that a newspaper can be 
read anywhere on the roadway. 

Every fourth hand-rail post is extended above those intermediate, 
and supports a 16-c.p. Tantalum lamp, set in a vertical, moulded, mica 
socket, surrounded by a glass globe 10 in. in diameter. There is thus 
a light every 36 ft. along each hand-rail. The lights are uniformly 8 ft. 
6 in. above the pavement, and are staggered on the two sides, making 
one lamp for each 18 lin. ft. of roadway. 

The wiring is of the three-wire, 250 to 500-volt system, with two 
lamps in series, connected in three circuits so that any third of the 
lamps may be used without changing the factors of symmetry and dis- 
tribution. Certain main posts have lamps wired on the three circuits so 
as to show light regardless of which third of the system is in use. 

In a central office is installed a switch-board, a motor generator, 
and certain other machinery. From this office are run two sets of three 
circuits, each with a common neutral for each set, to furnish current 
for the lamps of half of the structure, and two similar sets of three 
circuits each to furnish current for the lamps of the other half. The 
two outside wires of each circuit, together with its series wires, are 
run in one metal conduit, and the neutral for each set of three circuits 
is run on glass insulators. 

The conduits and wires are supported on short wooden cross-arms 
fastened on the top of each floor-beam between the two outermost 
stringers on each side. They are thus well sheltered, on each side and 
above, by the concrete floor. 

From the office to the structure the four neutrals are combined into 
two wires, each of solid copper. No. 10, B. & S. gauge. Elsewhere, each 
neutral consists of a solid copper wire, No. 12, B. & S. gauge. The 
wires connecting two lamps in series are of the same character, and of 
No. 14 gauge. Where placed on glass insulators, the neutral wire is 
of single-braid, Okonite insulation, and all feeders and branches run 
in metal conduit have double-braid insulation. Joints in the wires 



THE SIXTH STREET VIADUCT, KANSAS CITY 89 

are soldered and double-taped with Okonite and friction tapes. The 
conduits are of loricated metal, and where branches are taken off, 
Crouse-Hinds condulets are inserted. 

The feeder wires of the lamp are led down through the post to a 
fuse-box fastened near its bottom, in which is a 6-ampere cartridge 
fuse. To reach the fuse-boxes from below the floor, a 2-in. pipe was 
placed in the concrete at each lamp-post, and through this is passed 
a light |-in. metal conduit, bent to prevent the entrance of water, and 
bushed at each end. 

All branch wires running under feeders from the conduit are en- 
eased in continuous lengths of Alphaduct. Corresponding with the 
variation of the size of the wires, the conduits are varied, the six 
carrying main feeders being 1^, 1, and | in. in diameter, respectively. 

There was serious trouble in entering the several wires into the 
conduits, because of their comparatively small size; slightly larger con- 
duits would have simplified the task. 

The company furnished the lamps, globes, and certain castings, but 
the contractor installed them, and furnished and installed all other 
equipment for $28 000. There are about 500 lights on the main viaduct. 

6. — Approaches and Laterals. 

As has been stated, legal complications prevented the main lateral 
on Mulberry Street from being built as at first planned, and it be- 
came impossible to effect a landing at 12th and Liberty Streets. This 
lateral was always considered of vital importance to the success of the 
enterprise as a whole, and when the original plans had to be modified 
it was decided to build to 12th Street — two blocks east of the former 
corner — and there install electric elevators of sufficient size and 
capacity to raise to the deck of the structure any vehicle with its horses. 
Further delays of a legal nature arose, however, and at present, De- 
cember 1st, 1908, the Mulberry Street lateral is built to the corner 
of Mulberry and 9th Streets, and two elevators are installed there. 
The Hickory Street branch of this lateral divides from the main branch 
about 300 ft. from the main viaduct, runs down on a 3% grade, turns 
under the former, crosses Mulberry Street, and ends at 9th and Hickory 
Streets. This branch is shown on Fig. 2, Plate VIII. 

These laterals provide only for the highway loading, the Mulberry 
Street portion having a roadway 24 ft. wide, and the Hickory Street 



90 THE SIXTH STREET VIADUCT, KANSAS CITY 

branch a roadway 20 ft. wide. The pavement floor, hand-rail, and de- 
tails are the same as for the main viaduct. The detailing of the steel- 
work is very similar, but of very much lighter construction ; it is a two- 
column structure with longitudinal girders, transverse beams, and 
longitudinal stringers. No towers or longitudinal bracing are used; 
the columns are fastened down with four anchor-bolts which provide 
the necessary stiffness in every direction. 

The foundations are principally of small concrete pedestals resting 
on sand. They are placed at a sufficient depth below the street grade to 
allow for cellar construction. Two of the elevators have been in- 
stalled. Each has a counterweighted moving platform, 11 by 32 ft., 
lifting a total distance of 40 ft. 9 in. They are designed to carry 
22 000 lb. at 50 ft. per min., or 11000 lb. at 100 ft. per min. One 
50-h.p. motor is used for each elevator. Automatic gates are provided 
above and below and on the platforms. 

These two approaches were finished and opened to traffic a few 
months after the completion of the main structure. 

7. — Conclusion. 

The distances saved by tlie viaduct are shown by the following 
computation : 

Distance between Sixth and BlutT Streets and Fourth Street and 
Minnesota Avenue. 

By Metropolitan Street Kail way line 11 950 ft. 

By Sixth Street Viaduct 8 400 " 

Saving 3 550 ft. 

Distance between Eighth and Wyandotte Streets and Eifth Street 
and Minnesota Avenue. 

By Metropolitan Street Eailway line 16 000 ft. 

By Sixth Street Viaduct 11 850 " 

Saving 4 150 ft. 

The entire construction of the main viaduct was completed in 18 
months; in fact, the work was done with such dispatch that several 
contracts were finished ahead of their allotted times. The substruc- 
ture was ready for steel in 7 months. The steel erection occupied 10 



THE SIXTH STREET VIADUCT, KANSAS CITY 91 

months, the concrete floor was in construction 3 months, and the asphalt 
was laid in 1 month. 

It is not possible, without lengthy explanatory detail, to give exact 
distributed costs of the complete structure. These figures have been 
compiled and comj)ared minutely with the preliminary estimates. When 
due allowance is made for various changes, additions, and deductions, 
which originated after the first estimates were prepared, a very close 
agreement is seen. The greatest variation between actual and esti- 
mated cost is to be found in the items "right of way" and "legal 
expenses", both of which are materially in excess of the estimates. 

Contracts were made on the basis of unit-price payments for quan- 
tities, and provision was made for all anticipated contingencies. The 
extras on all contracts were performed at actual cost plus 10 per cent. 
The total amount of such bills on the two principal contracts was about 
j^^ of 1% of the total final estimates. 

The total amounts paid to the contractors, and the total cost to the 
company, of the principal physical features of the structure are given 
below, with the unit-prices for the work. 

The costs refer to the Main Viaduct, which includes the Kaw 
River Bridge. 

Substructure : 

All pedestals, abutments, earth work, etc $283 000 

Three piers in the Kaw River 98 000 

$381 000 

Superstructure : 

Steel erected and painted $785 000 

Concrete floor 95 000 

Pavement 35 000 

Street-car deck and trolley 86 000 

Lighting system 28 000 

1 029 000 

Right of way 227 000 

Total $1 637 000 

Unit-Cosis of AYorl-. — 
Concrete in pedestals, including ex- 
cavation, forms, etc $10 per cu. yd. 

Concrete in shafts of river piers. ..... 11 " " " 



92 THE SIXTH STREET VIADUCT, KANSAS CITY 

Mass of foundations of pneumatic 
piers. (This comprises the gross 
volume of base, including timber, 
iron, concrete, and cost of sink- 
ing) $18 per cu. yd. 

Anchorage metal in place 4| cents per lb. 

Creosoted piles furnished at site 46.40 cents per lin. ft. 

Driving creosoted piles, measured be- 
low cut-off 30.25 to 31.35 cents per lin. ft. 

Concrete piles in place 88 cents per lin. ft. 

Structural metal delivered, erected and 

painted 3.31 cents per lb. 

Construction of concrete iloor (cement 

and bars furnished) $6,085 per cu. yd. 

Asphalt pavement 1.05 per sq. yd. 

Laying ties for street-car deck (all 

material furnished) 5.00 per 1 000 ft., B. M. 

Laying, bolting, and spiking rails 4 cents per lin. ft. of single 

track. 
Costs per Linear' Foot. — 

Viaduct with 38-ft. Eoadway, and Two Car Tracks. 

Substructure $38.15 per lin. ft. 

Steel 92.60 " 

Floor and pavement 18.95 " " 

Street-car deck 10.75 " 

Lighting system 3.50 " " 

Total cost $163.95 per lin. ft. 

Viaduct with 30-ft. Eoadway, and Two Car Tracks. 

Substructure $38.15 per lin. ft. 

Steel 88.75 " " 

Floor and pavement 15.35 " " 

Street-car deck 10.75 " " 

Lighting system 3.50 " " 

Total cost $156.50 per lin. ft. 



THE SIXTH STREET VIADUCT, KANSAS CITY 93 

Kaw River Bridge, 30-ft. Roadway and Two Car Tracks. 

Substructure $163.00 per lin. ft. 

Steel 198.50 " 

Floor and pavement 15.35 " " 

Street-car deck 10.75 " 

Lighting system 3.50 " " 

Total cost $391.10 per lin. ft. 

The small difference in cost between the 30-ft. and the 38-ft. road- 
ways is accounted for by the fact that the former is built strong enough 
in all details to allow for widening it to 38 ft. by lengthening the 
roadway cantilevers. 

Considering both the roadway and the total width of both street-car 
tracks as upper surface area, the costs per square foot of this area 
were: 

For 38-ft. roadway portion $2.66 per sq. ft. 

For 30-ft. roadway portion 2.93 " " 

Average for viaduct proper 2.835 " " 

For Kaw River Bridge 7.32 " " 

Personnel. — This viaduct has been a work of many minds and 
many hands, and any list of those associated with it must neces- 
sarily be incomplete. The structure was designed by, and con- 
structed under the supervision of, the firm of Waddell and Hedrick, 
composed of J. A. L. Waddell, and Ira G. Hedrick, Members, Am. Soc. 
C- E., Consulting Engineers. (This firm has since dissolved.) 

Mr. J. H. Thompson, of New York, was Chief Engineer, and had, 
as his local representative, Mr. E. S. Wliitney. V. H. Cochrane, Assoc. 
M. Am. Soc. C. E., was Engineer in Charge of Shop Inspection. L. R. 
Ash, Assoc. M. Am. Soc. C. E., was identified throughout with the 
designing and preparation of the plans. The writer served as Resident 
Engineer for the Consulting Engineers, and was in general charge of 
the construction. 

The general contract for the substructure was in the hands of Mr. 
James F. Halpin, of Kansas City, under whom were varioiis sub- 
contractors. The pneumatic piers were built by Kahmann and Mc- 
Murray, of Kansas City. The concrete piles and most of the timber 
piles were driven by The Foundation Company, of New York, under 



94 THE SIXTH STREET VIADUCT, KANSAS CITY 

the personal direction of the late George Adgate, M. Am. Soc. C. E. 
The steel was furnished and erected by the Eiter-Conley Manufactur- 
ing Company, of Pittsburg, with Mr. John P. Wagner as local repre- 
sentative. 

The concrete floor was built by The Expanded Metal and Corru- 
gated Bar Company, of St. Louis, under George R. Heckle, Assoc. M. 
Am. Soc. C. E. The asphalt paving was laid by the Parker Washington 
Paving Company, of Kansas City. The laying of the street-car deck, 
certain substructure, and other various work was done by H. C. 
Lindsly and Son, of Kansas City. The lighting system was designed 
by Messrs. Weeks and Kendall, Consulting Engineers, and installed 
by The Squire Electric Company. 

The writer remembers with appreciation the efforts of a number of 
assistant engineers and inspectors, who labored zealously and efficiently. 
Grateful thanks are due to many gentlemen for aid in collecting the 
foregoing data, particularly to Messrs. Waddell, Pledrick, and Cochrane. 



discussion: sixth street viaduct, KANSAS CITY 95 

DISCUSSION 



Daniel Bontecou, M. Am. Soc. C. E. (by letter). — This paper Mr. Bontecou. 
presents many points of interest, and is well worth careful reading. 

Since the author gives some details of the preliminary investiga- 
tion of probable profits, it would be of interest to know more of the 
assumptions made as to the proportion of the total team traffic which 
could be depended on to use the viaduct, and of the extent to which 
they were realized. The estimates of cost and the operating expense 
given would mean that the gross earnings of the completed structure 
would need to be about $900 per day. The revenue from street-railway 
traffic could be determined fairly well, but could hardly be estimated 
at more than, say, $125 per day, leaving a very large proportion of 
the earnings to be derived from team traffic. In view of the risk in 
estimating the use which would be made of a utility of this kind, where 
it is optional to pay a toll or to take another and slightly less con- 
venient route, it is a fair question whether so expensive a structure 
should have been built otherwise than by the communities served; or, 
whether a cheaper structure should not have been built to serve the 
street-car traffic only, in which case no operating expense except main- 
tenance would be involved. 

Since all the cars, and probably most of the other vehicles, must 
meet grades steeper than 1^%, it would seem that the viaduct roadway 
might have been placed with advantage nearer the surface, and still 
have afforded the proper clearances. 

The shore line of the Missouri River at the site of the trestle on 
the accretion was about 600 ft. south of the structure only some 
twelve years before the latter was built, and has been regulated by 
protection work a few miles above, and also near the mouth of the Kaw. 
The design of the trestle foundations, therefore, involved a careful 
consideration of the permanence of the river regulation. The experi- 
ence of the 1903 flood seems to the writer to justify the choice of pile 
foundations, but the conditions of soil and loading appear to call for 
piles longer than the concrete piles described. 

In building the foundations for some large engines in similar 
ground near the viaduct, tests on a surface of 2 sq. ft. at a depth of 
about 15 ft. showed a supporting power of 1 340 lb. per sq. ft., and a 
settlement of 0.3 ft. under a load of 2 800 lb. per sq. ft. It was neces- 
sary to use oak piles only, 22 to 24 ft. long, and 300 were driven in an 
area of 2 100 sq. ft. — the intention being to depend on the combined 
resistance of the piles and the compressed sand, silt, and clay. After 
allowing for the frictional resistance of the supported concrete, the 
load amounted to 36 000 lb. per pile, or 4 480 lb. per sq. ft. of founda- 
tion, the value of the piles, by the Wellington formula, being 30 000 lb. 



96 DISCUSSION : sixth street viaduct, KANSAS CITY 

Mr. Bontecou. This foundation was deeply submerged in 1903 without the occurrence 
of settlement. 

In cases where piles are dependent on friction, in somewhat plastic 
clay, there is probably an element of safety in driving them with a 
heavy hammer and low fall, rather than jetting them to place, and in 
arranging them as far as practicable so as to confine and compact the 
soil. The concrete piles used in the viaduct had a surface area of 130 
sq. ft., and, as placed and tested, showed a resisting power of 38 000 lb., 
or rather less than might be expected from piles driven in a cluster in 
such material, and probably less than would be desirable if the roadway 
were actually loaded to 100 lb. per sq. ft. 

To determine the effect of clay and dirt on the strength of concrete 
a good many tests have been made, with varying results ; and, although 
the writer has often permitted an admixture of 5% of earthy material, 
he has preferred usually to wash dirty stone in barrows having open- 
mesh bottoms until the water ran clear. It is difficult to understand 
why the presence of clay should add anything to the strength of con- 
crete, as shown by the author's tests, and if stronger concrete can be 
made of well-mixed and graded materials with clay, than without it, 
the writer must confess to a great deal of misdirected energy in the past. 

The Kaw River spans, as described, are an interesting example of 
riveted construction. In view of the high ratio of dead to live load, it 
would seem to many that a smaller impact allowance was permissible 
for the motorway load, and the well-understood advantages of pin con- 
nections seem to apply to this case with special aptness. 

A comparison of the weight and cost of the adopted design with 
one based on three pin-connected trusses for each span, with cross-beams 
carried entirely above the top chord would be an interesting addition 
to the paper. 

Ever since the flood of 1903 there has naturally been a feeling of 
apprehension regarding the Kaw River, and the requirement by the 
War Department that bridge piers be carried to bed-rock has been 
cheerfully accepted, regardless of the fact that, while fifteen local 
bridges were destroyed by that flood, there was no case in which a 
well-constructed pile foundation failed. It is not apparent, therefore, 
why it should have been thought necessary to excavate the bed-rock 
to receive any part of the pier construction. 
Mr. Cochrane. Vtctor H. CorHRANE. Assoc. M. Am. Soc.'C. E. (by letter). — The 
writer left the office of the Consulting Engineers to take charge of 
the shop inspection. In some respects the work was unique, and a brief 
account of it, in addition to what the author has given in his unusually 
complete paper, may be of interest. 

There were three inspectors on the work, two employed by the 
Consulting Engineers, and one by the manufacturers of the paint 
specified for the structure. While in the office, the writer, in the 



DISCUSSION : SIXTH STREET VIADUCT, KANSAS CITY 97 

capacity of Chief Draftsman, had assisted in designing the main Mr. Cochrane, 
portion of the viaduct, and was familiar with the locality where the 
structure was to be built; consequently, he was well acquainted with the 
theoretical and practical considerations affecting the design. Accord- 
ingly, he was given authority to approve all shop drawings. As the 
author remarks, much time was saved by this arrangement, the draw- 
ings being put in the shops, when desired, almost immediately after 
being turned over to the writer for approval. The chief saving in time, 
however, resulted from the advice and assistance the writer was able 
to give to the force engaged in preparing the shop drawings. It was 
necessary in several instances to make considerable modifications in the 
design, and these changes were made without having to await the 
approval of the Consulting Engineers. The writer's experience in this 
case convinces him that the inspector in charge of a piece of work 
of this magnitude should be quite familiar with the design, in both its 
theoretical and practical aspects. 

The squad of draftsmen assigned for the shop detailing seemed to 
consist largely of men inexperienced in making drawings for such a 
structure; and, although the engineers' plans were quite complete, the 
first shop drawings finished and checked were found to have so many 
errors that they had to be practically re-checked before being approved. 
After the men became more experienced in the work, it was not neces- 
sary to check the drawings so much in detail. Altogether, there were 
more than 600 sheets. From time to time, after a number of drawings 
had been approved, the weight of each member was computed and 
entered in a quadrille-ruled notebook. 

The paint used on the work was guaranteed by the manufacturers 
for a period of ten years, provided they were allowed to employ in- 
spectors to watch the work in all stages. Their representative in the 
shops succeeded in getting excellent work, although there was much 
friction at first. The paint used in assembling was so heavy and stiff 
that there was considerable difficulty in applying it. The painting 
was done under contract by Greek laborers, and, at first, it was diffi- 
cult to get them to clean the metal as thoroughly as desired, since the 
standard adopted was much in advance of anything to which they 
had been accustomed; the quality of work required, however, was 
clearly 'and forcefully stated in the specifications, and, in cases of 
dispute, the paint inspector was nearly always sustained by the repre- 
sentative of the engineers. 

The punching, assembling, and reaming was carefully watched 
throughout. Each member after being finished — but before being 
painted — was inspected for loose rivets and other defects, and all 
leading dimensions and open-hole spacings were carefully verified. In 
some kinds of members many pieces were nearly alike. In such cases 
lists were prepared and used instead of the unwieldy drawings. The 



98 DISCUSSION : sixth street viaduct, KANSAS CITY 

Mr. Cochrane, dates of acceptance were noted on these lists, or on the prints in case 
they were used. Many pieces had to be sent back for correction, as 
was to be expected, but no very serious errors were found, and little 
material was rejected on account of surface defects. The provision 
for field reaming greatly facilitated the shopwork, but increased the 
labor of checking up field connections, as no assembling was done, and 
consequently there was no way to insure the matching of rivet holes 
except by measuring each connection. 

In cases of errors the inspectors endeavored to remedy the defect in 
such a way as to cause the least possible expense to the shops without 
impairing the strength of the piece. When the shop employees learned 
that the inspectors could be depended on to assist them in correcting 
mistakes, little or no attempt was made to conceal them, and they were 
usually brought to the attention of the inspector as soon as discovered. 

Each day the inspector entered in a scratch-book, without regard 
to order, the marks and number of pieces of all material inspected 
during the day. As soon as the invoices of shipments were received, 
the material listed on the invoices was checked off the scratch-book; 
at the same time the scale weights were compared with the calculated 
weights. By this means many mistakes were discovered in the invoices; 
some were due to misreading the marks painted on the pieces; in many 
cases, the scales were incorrectly read; and, in some instances, the 
number of members shipped was in error. The scratch-book was not 
intended as a permanent record, but each Saturday afternoon a list 
was made of imfinished work then in the shops, a column headed 
"Inspected" being filled out from the scratch-book. Unfinished pieces 
remaining in the shops from week to week were not re-entered in this 
book unless advanced in construction, in which case they were checked 
off in a column headed "Remarks." Thus the book showed at a glance 
the exact state of unfinished work at any time. 

A record of shipments was made from the corrected invoices once 
each week in a book, a column headed "Number Previously Shipped" 
giving the numbers taken from a previous record-book. Owing to the 
errors in the invoices, previously mentioned, the shop's records of ship- 
ments, after a time, became very unreliable, and it was necessary to 
depend on the inspector's records altogether. 

In preparing the record-books mentioned, the headings of the 
various columns were not repeated on each page, but were placed near 
the beginning and end of the book, the tops of the intervening leaves 
being cut away. 

The difference between the estimated and scale weights for the 

main viaduct was only three-tenths of 1 per cent. 

Mr. Mogensen. O. E. MoGENSEN, M. Am. Soc. C. E. — The author deserves credit 

for describing so fully the construction and design of the Sixth 

Street Viaduct; it is particularly interesting to know the problems and 



DISCUSSION : SIXTH STREET VIADUCT, KANSAS CITY 



99 



difficulties which arose during the construction, and how they were Mr. Mogensen. 
overcome. 

One of the points brought out in the paper relates to the trouble 
experienced in securing clean crushed stone for the concrete, and the 
fact that a comparative test showed that the unwashed clayey stone 
made stronger concrete than the clean washed stone. Mr. Bontecou 
takes exception to this last statement, which reminds the speaker of 
tests conducted some years ago by W. A. Rogers, M. Am. Soc. C. E., 
with a view of ascertaining whether cement with varying proportions 
of loam had a detrimental effect on concrete, and, if so, to what extent. 

With Mr. Rogers' permission, the results of these tests are given 
in Table 1. 

TABLE 1.^ — Tensile Tests of 1-in. Briquettes of Portland Cement 
WITH Varying Proportions of Loam. 





Proportions: 




Tensile Strength, in Pounds per Square Inch. 


Percentage of 










loam. 


















Cement. 


Sand. 




28 Days. 


3 Months. 


6 Months. 


1 Year. 









r 


503 


523 


546 


563 


2 






439 


504 


509 


577 


5 


j- 1 


2 




356 


445 


454 


549 


10 






1 


346 


465 


464 


566 


20 






1 


344 


368 


418 


533 











393 


361 


384 


430 


2 






357 


390 


418 


458 


5 


I 1 


3 


348 


315 


353 


491 


10 






1 


243 


335 


379 


506 


20 






299 


334 


368 


515 




Neat 




825 


841 


713 





The proportions were determined by volume; the sand was obtained by screening 
gravel from Hammond through a No. 4 sieve; the sand was washed, 0.031 foreign matter 
being eliminated; the briquettes were kept 34 hours in air, and were then immersed in 
water until broken. 

It will be noted that there is actually a tendency toward an increase 
of strength for the briquettes after six months and one year; these 
were mixed in proportions of 1 part cement to 3 parts sand. 

In giving these results, it is not by any means the object to 
encourage an admixture of clay or loam with concrete, but the 
records of various tests made from time to time show that the engi- 
neer is not justified in condemning rock or sand off-hand if a small 
amount of clay or loam be present. E. C. Clarke, M. Am. Soc. C. E., 
reports that 10% of loam, used with clean sand and Rosendale cement, 
did not decrease the strength after six months or one year. He found 
that 350 briquettes of Portland cement were not weakened by clay 
in moderate quantities, nor were their weathering qualities impaired 
in 2i years.* Nor is the engineer justified in compelling the con- 

* Transactions, Am. Soc. C. E., Vol. XIV. Additional information may also be found in 
Engineering Nev>s (August 10th, 1905, p. 140) in an article by W. C. Hoad, Assoc. M. Am. 
Soc. C. E., regarding tests made at the University of Kansas, on the effect of clay and loam 
on the strength of Portland cement mortars. 



100 DISCUSSION : SIXTH STREET VIADUCT, KANSAS CITY 

Mr. Mogensen. tractor to go to the extra expense of cleaning the stone, if a limited 
amount of clay or loam is present, particularly if a batch mixer is 
used. Sticky clay adhering to stone cannot be removed effectively 
by playing a hose on the stone in a barrel. It is far more effective 
to run the clayey stone in a batch mixer long enough, and to make 
use of the grinding effect in removing the clay from the surface of 
the stone. 

Another point which has been brought out is the fact that built-up 
driven piles were better adapted for the wet soil encountered in the 
work than the contemplated patented pile. It would be interesting to 
know why the use of the patented pile was undesirable and inex- 
pedient, even if the ground were wet. The moulding of a concrete 
pile in place without subjecting it to the impact of the hammer, is 
unquestionably superior to driving it or sinking it by jetting, especially 
when the jetting has a tendency to reduce permanently the carrying 
capacity of the pile by loosening the surrounding ground. 

No mention has been made of stations. To an outsider not 
familiar with the local conditions, it seems that a few stations on the 
viaduct near the steam railroad stations, would have added, not only 
to the utility of the viaduct, but also to the earning capacity of the 
electric railroad and viaduct. 

Mr. Blackburn. N. T. Blackburn, Jun. Am. Soc. C. E. (by letter). — This project 
seems to have been carefully planned and well thought ovit, from every 
point of view, and the paper presents it very clearly. 

Evidently, both the concrete and pine piles, under the pedestals, 
were sunk to full depth by jetting alone, no blows being struck by 
the hammer. Jetting is used almost exclusively in sinking piles in 
Galveston, Tex., and all along the Gulf Coast, where the soil to be 
penetrated is sand. Under that portion of the Galveston sea-wall con- 
structed by the Government, the piles were driven by a steam hammer, 
the jet being used only through the sand. It was specified that the 
last 2 ft. which entered the underlying clay should be driven by the 
hammer alone. The piles were of green pine with the bark on, and of 
such length that they penetrated the sand and went about 2 ft. into 
the clayey subsoil, their length averaging about 33 ft. The average 
number of piles driven per 8-hr. day was 62, a No. 2 Vulcan steam 
hammer being used. However, there was probably less shifting of the 
driver than was required for the pedestals of the Sixth Street Viaduct. 

At Aransas Pass, Tex., during 1908-09, piles for the trestle used in 
building the jetty were driven without difficulty to a penetration of 
15 ft. by steam hammer alone, the hammer striking at the rate of 
about 80 blows per minute, and without a jet. The material penetrated 
was the hard sand of the seashore. 

The author's results of tests made to determine the effect of a 



DISCUSSION : SIXTH STREET VIADUCT, KANSAS CITY 101 

small amount of clay in the broken stone for the concrete, seems to Mr. Blackburn, 
bear out what the writer has been able to gather from the experiments 
of others. 

George H. Pegram, M. Am. Soc. C. E. — This paper is very inter- Mr. Pegram. 
esting as describing a modern structure designed to meet unusual 
conditions. It brings to mind a structure designed by the speaker, and 
built, in 1886, in Ninth Street — a parallel street. This was the first 
elevated railroad built outside of New York City. The design is 
described in a discussion of Dr. Waddell's paper on elevated railroads.* 

The structure was designed for steam-train trafiic, and the steel- 
work weighed 500 lb. per ft., or less than half the lightest New York 
railroad. The distinguishing characteristic was the support of the 
track rails in the chords of the trusses which were composed of two 
channels forming a trough, the channels acting as guard-rails for the 
wheels. 

The trusses were pin-connected, and the chords were connected by 
angle-iron diagonal lateral bracing. No wooden cross-ties were used. 
The objects were to reduce the noise to a minimum and avoid darken- 
ing the street under the structure, both of which were accomplished. 

The history of such structures is interesting. Happening to be in 
Kansas City about twelve years ago, the speaker observed that wooden 
cross-ties had been placed on the chords, the reason given being 
that street cars, with small wheel flanges, had replaced the steam 
trains, and the cross-ties made a safer floor in case of derailment. The 
speaker was recently informed that the pin-connected truss bracing 
has been replaced by riveted bracing, which is in line with experience 
in such construction. 

The foundations were the ordinary concrete piers, resting on piles. 

E. E. Howard, Assoc. M. Am. Soc. C. E. (by letter). — The detailed Mr. Howard, 
data of traffic and income of the viaduct, requested by Mr. Bontecou, 
are unfortunately not available. In general, the writer's information 
is that the inter-city highway traffic has been slowly increasing, since 
the structure was opened, and begins to approximate the estimated 
volume. Five-sixths of the total highway reveniie was expected to 
come from the laterals into the West Bottoms, and, as explained, 
these have not yet been built in accordance with the original plan; 
so that the present structure can hardly be fairly considered in con- 
nection with the prospective income of the original estimates. Only 
two street-car companies are now operating over the structure, but, 
from the present indications of increasing traffic, the revenue from this 
source will in time materially exceed the estimated amount. 

Referring to Mr. Bontecou's suggestion that the structure could 
have been placed lower, it may be noted that 425 ft. from the east end 

* Transactions, Am. Soc. C. E., Vol. XXXVII, p. 4S8. 



102 discussion: sixth street viaduct, Kansas city 

Mr. Howard, it is down to the clearance required for surface railway tracks, and 
is at the required height above flood water at the Kaw River. No very 
great saving would have resulted by increasing the grades, and 
although, as he remarks, there are grades much in excess of those on 
the viaduct immediately at each end, the reason for the light grades 
may be readily found in the original conception of providing a long 
uninterrupted roadway over which rapid speed could easily be made. 
His suggestion that it would have been wiser to have built to provide 
only for street cars finds an answer, as far as the traveling public is 
concerned, in the flood of 1908 when for a time the viaduct afforded 
the only means for the passage of vehicles from Kansas City, Mo., 
to Kansas City, Kans. With respect to the excavation of bed-rock 
in jthe Kaw River, it might be explained that the bed-rock was found 
to be on a decided slope, necessitating leveling; and, as the surface 
was somewhat shaly, it was thought best to prepare the ditch, as 
described, around the entire perimeter of the caisson. 

The figures given by Mr. Mogensen with regard to clay in stone 
for concrete are of particular interest. The writer regrets that he can- 
not give similar detailed figures of experiments with the various 
materials used, instead of the general conclusions from a few tests, 
but extensive experiments could hardly have been undertaken at the 
time. The general observations lead to a concurrence in Mr. Mogen- 
sen's conclusion, that probably better results are obtained by more 
vigorous and extra turning of clayey stone in the concrete mixer, than 
by washing the stone in wheel-barrows with a hose, as suggested by 
Mr. Bontecou. It would seem that if stone is dirty enough to demand 
washing it should be washed by some more certain method, as, for 
instance, revolving it in a drum under a water-jet. 

While it may be possible in one experimental trial to pour enough 
water into a wheel-barrow to wash the contained stone entirely clean, 
still, as work goes forward, and with the carelessness of laborers 
ordinarily employed, such washing on a large scale for an entire job 
would hardly be of the same thoroughness or of sufficient thoroughness 
to give good results; or, in other words, if such washing is sufficient, 
none is probably needed. 

The reason for abandoning the patented pile was that sample piles, 
driven in similar soil, when subsequently examined, were found to be 
discontinuous. Apparently, the soft soil had pushed in, as the shell 
of the pile was withdrawn, and occupied some of the space supposed 
to be filled by concrete. That is to say, the ground had pinched off 
the pile, so that when excavated it was seen to be merely discontinuous 
chunks of concrete, which, of course, immediately fell over. 

Mr. Mogensen states: 

"The moulding of a concrete pile in place without subjecting it to 
the impact of the hammer, is unquestionably superior to driving it or 



discussion: sixth street viaduct, KANSAS CITY 103 

sinking it by jetting, especially when the jetting has a tendency to Mr. Howard, 
reduce permanently the carrying capacity of the pile by loosening the 
surrounding ground." 

This is subject to question. As piles, or, in the case of moulded-in- 
place concrete piles, forms or cylinders for the piles, are driven 
rapidly in succession, the ground is very much compressed. This is a 
familiar phenomenon, seen always in driving piles. It would be of 
much interest to know what experience has shown to be the effect of 
such compression on moulded piles which are still green. If the con- 
crete is placed immediately after the moulds are driven, the material 
would be subject to the compression and vibration of the ground as 
successive piles were driven, about as the concrete would be taking 
its set. This reference, of course, is not to a single pile driven by 
laboratory methods, but as ordinary work would go forward. It seems 
that the advantages of a made-up pile are very definite, in that the 
pile can be closely examined just before sinking, so that it can be 
reasonably well known what is left in the ground. 

The latter part of the statement, regarding the reduction of carry- 
ing capacity of a jetted pile because of the loosening of the surround- 
ing ground by jetting, may also be questioned. The method given 
in the paper for approximating the bearing capacity of jetted piles 
was the only one that the engineers could evolve. If there is any more 
accurate method, a description would be much appreciated. Mani- 
festly, it is not possible to put a test load on every pile driven for a 
structure, and it would be interesting to know how the data were 
arrived at concerning the reduction of the carrying capacity of jetted 
piles. All engineers who have driven piles in swampy ground have had 
experience with cases where the pile drives very easily, and, yet, when 
left to set, can hardly be started again by the same hammer that drove 
it. Does not a similar action occur in the case of jetted piles, and 
is not the ground which settles back around the pile at least as solid 
as in its original condition, and probably more solid and dense? If 
this is not true, why is it that a hole is often left around the perimeter 
of a jetted pile where the ground has sunk away? The writer would 
very much appreciate any data concerning the bearing capacity of 
jetted piles or the method of determining it. 

Replying to Mr. Mogensen's inquiry regarding stations, it should 
be explained that the conditions at present existing make it unneces- 
sary to have intermediate street-railway stations, as there is practically 
no pedestrian traffic below the structure, it being paralleled on one side 
by railroad yards, and on the other side by unoccupied river bottom. 

Mr. Pegram's comparison of this structure with the Ninth Street 
Viaduct, built on a parallel street, is pertinent; and an examination 
and comparison of the two structures on the ground is interesting, as 
illustrative of the increase of live loads to be provided for and the 



104 DISCUSSION : sixth street viaduct, KANSAS CITY 

Mr. Howard, general change in details of construction. The pin-connected trusses 
of that structure were replaced by small riveted trusses in 1905, but 
the old top chords, each consisting of two channels, are retained, and 
now act as stringers; thus the columns, the cross-bracing, and the top 
chords of the original structure are still in use, and are operated over 
by a frequent and heavy modern street-car service. 



AMEEICAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 1853 



TRANSACTIONS 



Paper No. 1121 

SOME EXTENSIVE EAILROAD SURVEYS, 
AND THEIR COST PER MILE.* 
By W. S. McFetridge, M. Am. Soc. C. E. 



With Discussion by Messrs. Clinton S. Bissell, F. Lavis, E. W. 
Lewis, George L. Dillman, and W. S. McEetridge. 



The following paper is not a theoretical discussion of railroad 
location. It is intended to give a general description of some extensive 
railroad surveys, a brief outline of methods and results, and the 
cost per mile. The writer does not remember having seen a similar 
statement covering so many miles, viz., 1 400 miles of preliminary 
lines and 600 miles of location; and therefore trusts the paper may 
prove of value. The surveys were completed some time ago, at which 
time the tables of mileage and cost were prepared, but not heretofore 
published. 

Introductory. — Early in 1902 the Little Kanawha Syndicate began 
surveys for the extension of its lines, eastward from Palestine, W. Va., 
to Belington, and westward from Parkersburg, W. Va., to Zanesville, 
Ohio. About one and one-half years later it also took up the location 
and construction of a line running northward from Belington to the 
Pennsylvania-West Virginia State line. 

The surveys were conducted under the following charters: Zanes- 
ville, Marietta and Parkersburg Railroad, in Ohio; Parkersburg Bridge 
and Terminal Railroad, from the Ohio- West Virginia State line to 



* Presented at the meeting of May 19th, 1909 



106 RAILROAD SURVEYS 

Parkersburg (this division included a bridge over the Ohio River a 
few miles belov^' Parkersburg) ; Little Kanawha Railroad, from Parkers- 
burg to Burnsville; Burnsville and Eastern Railroad, from Burnsville 
to Belington; Buckhannon and Northern Railroad from Belington to 
the Pennsylvania- West Virginia State line; in all, some 328 miles of 
main-line location, exclusive of branch lines. 

Fig. 1 (Ohio) and Plate IX (West Virginia) show the country 
traversed and the main survey lines, many of the short lines not being 
shown. The lines shown give the general layout. The termini, as 
usual, were fixed; physical conditions also fixed the Little Kanawha 
River as the only outlet to the Ohio. These points decided in a general 
way the proposed route. Owing to local conditions, it was also believed 
that the heavier traffic would be west-bound, and therefore that every 
effort should be made to get as low a ruling grade as possible for 
this traffic. 

The desired results may be briefly stated as follows: 

1. — Easiest grades possible, especially against west-bound traffic; 

2. — ^Lightest curvature; 

3. — Shortest line; 

4. — Occupy to as great an extent as possible critical and strategical 
points, so as to block the route to other lines; 

5. — Reach certain definite places previously determined on; 

6. — And, taking into consideration the naturally heavy and ex- 
pensive work on any line, the total cost should compare 
favorably with the cost of any other line, many miles longer, 
with heavier grades and curves, but avoiding some of the most 
expensive work. 

All roads previously built through the adjoining regions have long 
stretches of 1.5% grades, and curves up to 12 and 14 degrees. The 
first surveys, therefore, were of a preliminary nature, in order to 
determine what grades and curves could be secured. 

After a number of surveys, locations, and explorations had been 
made, it was found that the following grades and curves were possible; 
in Ohio, 0.5%. grades, 4° maximum curve; Little Kanawha Division, 
0.3% grades, 8° maximum curve; Burnsville and Eastern Division, 
1.0% grades against east-bound and 0.5% grades against west-bound 
traffic, 8° maximum curves; all grades compensated for curvature at 
the rate of 0.04' per degree. 



RAILROAD SURVEYS 



107 




^^l ZANESVILLE, MARIETTA & \^^^^--^P ^)^-3v^ 
'^'"- PARKERSBURG R. R. N^* „ l::^W^- f^ t>©\^^ \ 

FROM vr^l ^M}'"'^ ^S ^ rflN 

\lbau<^i """"^ ^'^'^ '"^^f' """O ZANESVILLE 
fY OHIO 



Scale of Miles 



10 



3. Baltimore & Ohio R. R. J 

4. " " " S.W. R.R. 

5. Bellaire, ZanesvUle & Cincinnati Ry. 38. Mai-ietta, Columbus & Cleveland R. R. 
18. Columbus, Sandusky & Hocking Ry. 41. Wheeling & Lake Erie Ry. 

34. Cincinnati & Muskingum Valley Ry. 42. Ohio & Little Kanawha-R. R. 
37. Toledo & Ohio Central Ry. 



Fig. 1. 



108 



RAILEOAD SURVEYS 



These results were kept in view in continuing the surveys, and were 
obtained in each case. It was desired to avoid all momentum grades, 
and only in one case was it found necessary to use them. This case 
occurred at Mile 20 on the Burnsville and Eastern Division, where 
such a grade was introduced in order to avoid a long detour combined 
with some exceptionally heavy work. At some future time this grade 
can be taken out if desired. In the meantime, for any ordinary reason 
of operation, a train need never stop there, and thus get stalled; 
there is ijo station stop, and a water-station or siding is not feasible, 
on account of local conditions, so that trains can always "run for the 
hill." Fig. 2 is a profile at this place. 




1130 

M.P.18 

•<— West 



M.P.21 
East — 



M.r.l9 M.E.20 

West-Boimd 

Train loaded for 0.5?* grade and able to maintain 

uniform.speed of 10 miles per hour. 

BURNSTILLE & EASTERN R.R. 

Profile at M.P. 20, showing Virtual Profile for West-Bound Traffic. 

Fig. 2. 



The grades desired were very easy for parts of the country, and 
required some rather long continuous grade lines, the longest being 
on the Burnsville and Eastern Division, where there are 1.0% grades, 
7 miles and 7i miles long, respectively, and a 0.6% grade 14 miles long, 
all against east-bound traffic. Fig. 3 shows a grade-line profile of this 
division. 

To obtain these grades it was necessary to make a thorough study 
of the country and a detailed examination of several routes. 



PLATE IX. 

TRANS. AM. SOC. CIV. ENQR8. 

VOL. LXV, NO. 1121. 

McFETRlDGE ON 
RAILROAD SURVEYS 




RAILROAD SURVEYS 



109 



The topographical sheets of the United States Geological Survey 
were found of great value in making a broad, general study of the 
country. At that time the published sheets covered only a small por- 
tion of the lines, but, fortunately, this was the central part of West 
Virginia, where they were of particular value. It may be pertinent 
to state that these sheets were usually found quite accurate in regard 
to main summits, large rivers, and towns, but intermediate details were 
always more or less inaccurate. 




15 20 25 30 35 10 45 Miles from 

B"™8ville GRADE PROFILE, Bumsville 

BURNSVILLE & EASTERN R.R., 
M. P. 5 TO M.P. 45. 

Fig. 3. 

A large number of maps of small scale (1 in. to 1 mile or even 
smaller) were compiled and traced from various State, county, and 
road maps, on which the several survey lines could be indicated. Such 
maps are of particular value, and should be the first ones prepared— 
if possible before any surveys are made. 

General Layout of Surveys.- — The general direction of the survey, 
except along the Little Kanawha Division, was almost directly across 
the general drainage of the country. 

In Ohio a direct line between termini was first examined, but was 
found to be impracticable. This line marked the north and east limits 
of the country to be examined; certain conditions prevented the 
location of a line north of it. A systematic examination toward the 
southwest was then made, and finally a satisfactory line was developed, 
over which a final location was made. All the streams here lie in 
deep, narrow valleys, and are exceptionally crooked. The only feasible 



110 RAILROAD SURVEYS 

way to traverse much of the country was to get up out of the valleys 
and stay out. Such a method necessitated crossing about 100 ft. above 
several streams, and running short tunnels between the water-sheds; 
it also gave the shortest line, the easiest grades, and the lightest curva- 
ture. Any attempt to avoid the high crossings or tunnels involved long 
detours, excessive curvature, and heavier grades. On this division more 
miles of preliminary lines were run, in proportion to the location, 
than on any other. 

The main problem on the Parkersburg Bridge and Terminal Rail- 
road was the determination of the location for a bridge over the Ohio 
River. The Government regulations required 90 ft. clear head-room 
above low water and no piers in the main channel, which necessitated 
a 700-ft. span. Several feasible locations were surveyed, and complete 
data were obtained at each place. The location finally adopted is 
about 5 miles below Parkersburg, and is believed to be the shortest 
and cheapest railroad bridge crossing the Ohio between Pittsburg 
and the Mississippi, the 700-ft. span practically clearing the entire 
channel, the eastern approach being on the only bottom lands in that 
region which are not subject to overflow, and the western end being 
against a steep bank rising rapidly from the river. The bridge and 
viaduct approaches are but little more than half as long as the present 
bridge at Parkersburg. 

The Little Kanawha Division, in general, followed the Little 
Kanawha River. The hills rise abruptly from the river banks, the 
slopes often being as steep as 1^ to 1. The river is very crooked, and 
to follow it gave a long line with much curvature. Much distance 
could be saved by cutting through the country at various points, but 
the work was very heavy. The bottom lands are narrow and subject 
to overflow several times a year, the river sometimes rising 20 ft. in a 
day. A river line would have to be built along the steep side-hills 
for most of the way (this being expensive), or else be subject to 
inundation during high-water seasons. The hills are cut through in 
all directions by numerous tributary streams. The grade of the streams 
is from 50 to 80 ft. per mile, their length varying from 1 or 2 to 20 or 
30 miles, and all of them terminate against abrupt hills. Many 
previous surveys of this river had been made, and practically all of 
them followed the river, abandoning any serious attempt to shorten 
the line materially by cutting across country. The first Little Kanawha 



RAILROAD SURVEYS 111 

survey also followed the river, and a complete location vs^as made. This 
line was run principally for information, and, shortly afterward, was 
abandoned. The main problem on this division was to locate such a 
line that its total cost would not greatly exceed the cost of a river line 
equally well built (though no definite amount was ever fixed for this), 
that its grades and curves would be as light as by river, and that it 
would be as short as possible. This problem required the most minute 
study of the land, a thorough knowledge of its character, the running 
of many lines, and the comparison of many estimates. On a first 
examination of this country one was inclined to say offhand that it 
was impracticable to leave the river for any material distance because 
the cost would be prohibitive ; but when a thorough study began to show 
the length of line that could be saved by a mile or two of enormous 
work, the question had a different aspect. The only way to get sufficiently 
accurate data, on which to base final conclusions, was to make the 
actual surveys of the different routes. 

One item which was always more or less an unknown quantity, but 
which was the deciding point between two lines, in more than one case, 
was the treacherous character of some of the side-hills. In a number 
of places (the same trouble was also encountered on other parts of the 
lines) the side-hills were very likely to slide. It often happened that 
quite a large area, and sometimes from 15 to 20 ft. deep, would move 
bodily down hill for several feet when disturbed by any construction 
work. To avoid these places as much as possible, and to estimate their 
probable cost when comparing lines, were difficult problems. On loca- 
tion, a plan sometimes used was as follows : Where the grade was near 
the foot of a hill, the location was thrown away from the hill, in 
order to avoid cutting the slopes; this gave a fill where it was not 
apparently necessary, but it was cheaper in the end. When the grade 
was high up on the hillside, the opposite was done, for the purpose of 
getting the roadbed on solid ground. In many cases, however, such places 
could not be avoided in any way, and it was often found that ground, to 
all appearances perfectly firm, would slide after construction started. 
In comparing the various lines, this tendency to slide had to be taken 
into account, but the estimate of cost was of necessity more or less an 
approximation. These slides, throughout this country, are frequently 
a very large item of cost in building, and are usually underestimated. 

The line, as finally located, is a combination of river and cross- 



112 RAILROAD SURVEYS 

country line. It is 31 miles shorter than the river, in a total distance of 
100 miles. There are eight tunnels, usually short, the longest being 4 000 
ft. There are seven river crossings, with main spans from 100 to 300 ft. 

The Burnsville and Eastern Division is in the central mountain 
part of the State. The highest altitude reached is 1 725 ft. above sea 
level. At the junction with the Little Kanawha line, the altitude is 
750 ft. The lines generally run east; the mountains or hills, north and 
south; the drainage, generally north. There were several intermediate 
summits to be crossed. It was about as difficult to find the 0.5% grade 
west as to find the 1.0% grade east. The first line examined was up 
the head-waters of the Little Kanawha, up Fall Run, and through the 
head-waters of French Creek. The country was very rough and broken, 
and supporting ground for grades could not be found. This line also 
developed the fact that, owing to the rapid rise of the land at the 
eastern end, it would be necessary to use the maximum grade at once 
for getting out of the Little Kanawha water-shed, and then follow 
around the head-waters of the streams to the north, in order to get 
supporting ground. The line finally located was outside of the Little 
Kanawha water-shed. 

The Buckhannon and Northern Division followed the river for 
about two-thirds of its length, the other third being cross-country. 
The general problems of location were very similar to those on the 
Little Kanawha and Burnsville and Eastern Divisions. 

Methods Used. — Field parties were made up as follows: 

Monthly Salary. 

Assistant Engineer in charge $125 to $150 

Transitman 85 " 100 

Levelman 75 

Rodman 65 

Head chainman 50 

Rear chainman 45 

Rear flagman 40 

Stakeman 35 

Axemen (from two to five) 30 

Topographer 65 

Tapemen (two) 45 

Draftsman (part time) 60 



RAILROAD SURVEYS 113 

Camp outfits were not used. The parties boarded at liouses along 
the line. This was often a disadvantage, on account of difficulty in 
getting quarters, especially for a full corps; but, on the other hand, 
the party could frequently make its headquarters at some town and 
drive to and from the work, so that probably this method served just 
as well as furnishing camp outfits. 

Each party was given from 40 to 60 miles of line to cover, depending 
on local conditions. 

The following abstracts from the "General Instructions" give an 
idea of the work required to be done by each party : 

"The Assistant Engineer will receive instructions as to when and 
where preliminary lines will be run, grades proposed to be used, what 
the line is intended to develop, and the general information necessary 
in regard to the same. He will then be responsible for the amount and 
character of the work done by his party, and for the proper develop- 
ment of the country over which he works. 

"Topography will be taken on lines as instructed." 
(Topography was taken on practically all lines except in Ohio.) 
"After preliminaries are run, the Assistant Engineer will be in- 
structed over which line to locate. 

"All location will be first projected on topography sheets, etc. 
"No location will be assumed as final until examined and revised, 
if necessary, by the Assistant Chief Engineer, and also approved by 
the Chief Engineer." 

The Assistant Engineer was not expected to spend all his time with 
the party, but to be with it enough to see that work was going on 
satisfactorily, and that lines were being run over the proper routes; the 
remainder of his time was spent in a thorough study of the country, 
picking out routes to be examined, and looking after his necessary office 
work and correspondence. On location and in particularly difficult 
country he was to be with the party almost constantly. 

After the first route to be examined had been chosen, a preliminary 
line was run through; then the alternate routes were run, all surveys 
being tied together; and finally the lines required for a thorough 
development of all possible routes were run. Although the first pre- 
liminary line might give the grades and curves desired, and possibly 
be the line over which the final location was made, it was not finally 
determined on until the subject was fully investigated. 

Sometimes all the preliminary data required over a certain route 
may be obtained quickly by stadia methods, but some sort of a definite 



114 RAILROAD SURVEYS 

line is required from which reliable conclusions can be drawn. Especi- 
ally is this true in heavily timbered, hilly country, where the range of 
vision is very limited, and traveling on foot is excessively wearisome, 
and unless such a method is used some feasible route is very likely to be 
overlooked. Many times on these surveys the Assistant Engineer was 
sent back over parts of the line to hunt out something better, quite 
often with success. 

In locating long grades, it was preferable to start at a summit 
and run down hill. With a little experience, the Assistant Engineer 
could make a sufficiently close estimate of the amount to allow for 
compensation for curvature, and could run his line accordingly. In 
the mountainous part of the country here described this compensation 
amounts to about 6 ft. per mile, equal to 0.12% grade and preliminaries; 
for a 1.0% compensated grade, run on an 0.88% straight grade, gave 
the desired information. 

In case of the choice of two apparently equal routes, a location was 
made over each, and estimates were prepared for comparison before 
choosing a line. In following the larger watercourses, it was usual 
to locate a line on either side for purposes of comparison, and in order 
to determine the advisability of crossing from one side to the other 
either to get a better line or to block the country against rivals. 

Some branch lines were located, and quite an extent of line was 
located and afterward abandoned by reason of change of plans, but at 
the conclusion the miles of line actually located footed up one-third 
more (this is practically the correct proportion, but it varied a little 
more or less on different divisions) than would be required on the 
entire layout. This extra location was partly for comparison, but 
largely for reasons mentioned later. In addition to this, many miles 
of paper location, not located on ground, were made for the same 
reasons. 

It may appear to some that there was much unnecessary location 
and running of preliminary lines, but in rough country like this, and 
on work of this magnitude (in 220 miles of this line there were twenty- 
one tunnels, the longest being 4 000 ft., five viaducts from 400 to 1 000 
ft. long, and more than 100 ft. in height, besides numerous other 
bridges), it is time and money well spent. In no other way can the 
exact data be gotten, and it leaves no question as to the available 
routes and the grades obtainable. 



RAILROAD SURVEYS 115 

The locating engineer, or others on the ground, may feel certain 
that a line is not feasible, but it is hard to furnish proof, both to him- 
self and to others who are interested, but who will never be over the 
ground, that such is the case, except by map and profile. Again, an 
apparently hopeless line may show up much better than expected, and 
vice versa. In this the writer must not be misunderstood to mean 
that anyone can get the best location if he only runs lines enough; 
such is far from the case; lines can be run indefinitely without secur- 
ing the desired results, unless the proper judgment and knowledge are 
combined with them, the location being usually the most difiicult 
problem of raih'oad work. What is meant is that it is necessary to run a 
sufficient number of lines, preliminary and location, to arrive at correct 
conclusions and to get the requisite exact data to prove the conclusions. 

Topography, showing contours, houses, roads, watercourses, etc., 
etc., was taken on practically all lines. This was taken on 12 by 18-in. 
sheets. The transit line was plotted each day on the requisite number 
of sheets; a light pencil line, at right angles to the center line, was 
drawn through each station, for ease in plotting the topography; eleva- 
tions were marked at each station; the stations where contours crossed 
the center line were determined from the profile and marked; connec- 
tions to other sheets were shown, and then the sheets were ready for 
field use. The lines were plotted to a scale of 200 ft. to 1 in. The 
topography was plotted in the field. A hollow drawing-board, 18 by 
24 in., was used. The sheet in use was tacked to the board, and the 
additional sheets were carried inside. A strap around the shoulders 
of the topographer served to carry the board, and formed a support 
while plotting (Wellington's method). 

This method was preferred to any other; it is quicker; saves much 
copying and plotting; the work can be plotted better in the field, where 
everything can be seen at the time of plotting; and at night the 
Assistant Engineer has a finished map to look over and study. The 
topography was taken accurately by using a metallic cloth tape for 
distances and a hand-level for elevations. Only in this way can one 
get a projected location to correspond closely with the actual one. The 
topography was ordinarily taken for 300 ft. on each side of the center 
line; at particularly difficult summits or similar places a strip from 
1 000 to 2 000 ft. wide was often shown, the necessary topography being 
obtained by auxiliary lines. The sheets were inked in each night. 



116 RAILROAD SURVEYS 

To obtain a large general map showing all lines, the lines were 
carefully plotted on tracing cloth, the small sheets were fitted so as to 
make the center line on each sheet fit the center line on the tracing, 
and then the topography was traced. By this method any error in 
plotting or joining the small sheets was eliminated from the large map. 
This tracing, from which blue prints were made as required, was 
retained in the office, the sheets being used for rough work and 
field use. 

The location was projected in pencil on sheets in the usual manner, 
care being taken to determine all angles between tangents by the 
calculated courses from each sheet, not by using a protractor, or by 
measuring intersection angles, although either might be used as a 
check. After the line was located on the ground, the location was 
accurately tied to the preliminary line and replotted on sheets; the 
pencil projection was then erased, if so desired. 

In staking out the location, the aim was to get a profile to correspond 
with the projection, and not to get the lines in exactly the same 
relative positions shown by the projections. The lines always varied 
more or less in their relative positions. Also, it was often found 
desirable to change the location at places, giving a corresponding 
change in the profile ; for instance, it might be decided in the field that, 
owing to surface conditions, certain hills could be hit harder or 
perhaps avoided. The most frequent change was to put the line harder 
into steep side-hills. When projecting lines along such places, there is 
nearly always a tendency to fit the line too closely to the surface and 
not to allow enough cutting to put the roadbed on firm ground. Owing 
to these facts, it is usually better to have the first location projected 
and run in by the field engineer, who is most familiar with actual 
conditions; after that the revisions and necessary changes can be taken 
up by the higher officers. Projections made by anyone not thoroughly 
familiar with the ground should be used with caution. The best loca- 
tion cannot be obtained without the topography and a projected line; 
also, it cannot be obtained on the sheets alone, without a thorough 
and complete knowledge of the character of the country; the two must 
be studied together, the field engineer at all times corresponding freely 
and fully with the Chief Engineer and getting his ideas and in- 
structions. 

After the first location had been made, it was studied further in 



RAILROAD SURVEYS 



117 




118 RAILROAD SURVEYS 

the Chief Engineer's office; if any changes were desired they were 
taken up with the Assistant Engineer, usually by the Assistant Chief 
Engineer and the Assistant Engineer going over the ground together 
and there studying the question. After settlement, the location was 
taken as final for that route, and the maps and plans were brought 
up to date. 

Standards. — All curves of 3° or more had spiral approaches. These 
were allowed for in cross-sectioning by offsetting the slope stakes the 
required distance. The spiral was in standard form, so that, when the 
degree of curvature was known, the offset from the simple curve and 
the length of the spiral were obtained from tables. For simplicity and 
ease, all records, profiles, etc., were kept on simple curve data. When 
spiral curves came in tunnels, a special plan was made for each case, 
showing the offsets from the tangent and the simple curve at every 
10 ft. on the spiral, the alignment being kept on the tangent and the 
simple curve, and allowing the required offset in giving the widths for 
the tunnels. All grades were compensated 0.04' per degree of curvature. 

Vertical curves were inserted at all places when the change of grade 
was more than 0.1 ft. in 100 ft., a standard form being used in which the 
length of the vertical curve varied directly with the change of grade. 

The usual weekly reports, maps, and profiles were sent in by the 
Assistant Engineers. Standard forms were used for all notes, maps, 
profiles, plans, and reports. The standard rules were first issued as 
typewritten copies, and standard forms were adopted from time to 
time as needed; later, they were compiled and issued in book form, 
being blue-printed from tracings, making 22 pages of general rules 
and information, together with 24 standard forms for notes and plans. 
It was found to save much work and time in the Chief Engineer's 
office to have all data in uniform shape; much correspondence was 
likewise avoided, as employees could readily find sizes, scales, etc., for 
plans, and also other information. Therefore they were not com- 
pelled to write to headquarters, and thus avoided the delay of two or 
three days required for letters to go and come. 

A few of the standards used extensively on location are here shown. 

Fig. 4 shows the standard topography sheet, the same form being 
used on preliminary and construction lines. After the line was finally 
determined, sheets were prepared showing that line only, and the title 
was made to suit the construction work shown. 



RAILROAD SURVEYS 119 

The standard titles for profiles are shown by Fig. 5, and the standard 
titles for field books, by Fig. 6. The standard forms for transit and 
level notes are shown by Figs. 7 and 8, respectively. Fig. 9 shows the 
standard form for a location profile, and Fig. 10 shows the standard 
form for a situation survey for a bridge. When first prepared the 
structure was shown in pencil, and was not inked in until the detailed 

J^rofi-le of ) J-'ro/'ecieA ZjocolOoti/. 
I ^" JLocatiori' 
\JBevLsed, JLocaiioriy. 

^orrv io 

^ta, to (S'tcc- 



-W-, - ,. ^ , ..M)cme of Snqineer. 



• Profile- of Sevieeol^ JLocoiUon ^ 

QJ2^ 

SziholivisioTV C7^. TDiv. kTVo 

SiVi, jro (3fou 

JDuie 

Chief ^Tt^'r^. J^esCden^^Ti^h . 



Froo /ress ^rofCle . 

JBeJiaeerv stot. <5q Sfoc. 

^ivvr >Ag. ^SubdiynJ^- 

^ect'n ._ f-Q inc . 

JDocfe 

CNote — Tiller io ee^ui in. -upper- 2efir hocrtoi cornef of profiles. 
Fig. 5. 

design had been made by the Consulting Bridge Engineers, A. P. 
Boiler and H. W. Hodge, Members, Am. Soc. C. E. The standard 
form of report of openings required is shown by Fig. 11. 

Gosi. — The greatest number of miles of preliminary line run in one 
day hy one party was 7, and of location, 4^. The location averaged 
slightly more than 1 mile per day per party, except on the Burnsville 
and Eastern and on the Buckhannon and Northern lines, where it 



120 RAILROAD SURVEYS 



JL . H^. 7?. IS. 
_LeJii 7 ^^otes orb T^rel. line Ji. from 
Groifton to CA^oat^jville, <J^toL.O io PSO. 

J)ooie yi^sl-E'riyr 

or 
2L evelmoLTh. 



JSook JVb^. 



JL.IC.Jd.Id. 

J or L riotes oTh Ji, J_,ocaZtoyv xttou J\, 

J^rel. frorrv Orocftoru ioj^oalsville^^ <:yicc.O 

JDoLie Assf: I7rvjy 

or 

-TD 7 ^^ ^ -Lj GV eZirVOoTO, 

Jd oo/c JnTo 



JL.IC.Id.R. 
Trarisii) -noles on, r-evi^eol loccbtvorh o?7^ 

^lohdivD'J^. DiviiJ^. K^tcx..^ 

to ^S^too 

JDcLte 

-Booh JVo. -r, ., X ir ' 

Fig. 6. 



RAILROAD SURVEYS 



131 




123 



RAILROAD SURVEYS 







^ 

^ 


i 


te 
^ 










3 








r 




3 
Cn 
TV 


CD 


-*- 

o' 

3 


r 

o 
o 

p 

3 




































E9 






















C) 












^ 




S 






















^ 












O 




V5 


Z. 






























■fc 






Q 






































3 




































-^ 




















«^ 
















iz- 


» 




















Oo 












o 




? 























cb 












^ 




r 


1 




















N> 












:^ 




rr 


sr 








•^ 


*. 


9ft 


^ 


■^ 


O 




^ 


>o 


t^ 


?^ 








I 










<^ 


ON 


"i 


<W 


l\i 


db 




^ 


^ 


^ 


■^ 






\5 










1 




5o 










< 


1 




ON 


1^ 

Ob 




1_ 


1 


3' 
•+5 








^ 


~3 


?7 


W 




>c 




Ts, 


''O 


^ 




> 








3 








>5 


>9 


vo 


VO 


«) 


CO 








<o 


«o 


«0 

















cb 


90 


9o 


03 


Ot> 


Oo 






^ 


^ 


•N 









-^ 








^ 


S^ 


^ 


o 




6 






^ 


i 




Q 
Q 






Q- 



1 








>» 


>r 


> 


>1 


>l 


cr> 






fT 


p 


n 


p 






0? 










^1 


■?^ 


09 


■^ 


O 


^ 






$ 




o 








r.TT 


-♦■ 








o\ 


■?;> 


M 


"a 


''N 


^ 






Q 






o ■ 







































fj$ 




1-^ 


































k' 




-A 


































s =< 







































^! 




< 


































^o 




3 


































^" 




























\ 


^ 


\ 


\ 




:03 




























V 


^ 


\ 


X 




^ 




U) 
























1 


^ 


|\ 


\ 


0> 







P 
























'^ 


J 


1 


\ 


i^ 






























9 

A. 


II 





































1, 


:& 






-t 




























^ 




II 


1 






-T 
P 



r 

< 



RAILROAD SURVEYS 



123 





Cl'eari 


ng 




Ti'mbc 


r ) 




Cle 


■iring 


X 


: 


"im 


)er 


X 




Cleavii 


g 


X 






ii 


imber 




X 
















































( 


22. 




i 


^ 




















































V 














<? 














S-? 










oo 
















" "5 


















t- 














|S 


Si 




rT 
















t ^ 


















II 












s. 


II 1 




y 






II II 






BIG 


BE 


ND 




S :g 


















% 












S 


1- 




S: 






>^ 


































i 














i 




-a" 






-a 
















P 


^_ c 


















M 












5 

s 


« 




1^ 






1 = 








J 








i 


'^-a 


1 (§ 








750 


00 




a 












^ 


^ 




?% 






>1 






1 


N 










u 


T 














(ii 












c 






o 






§^ 






I 














1 














+ 












>-) 


InS 




gs 












/ &oli( 












g 




























+ + 




+ + 






f4 




/ 


ifdB^ 












"^ 


" 










































/ 


outuro 


\ 
























































Tuiiiw 


1 




















































/ 




N 


0. ' 


















































A 


\r 


J 






















































. 


/ 










~~ 


"■ 














































1 
















i 


















700 


00 
























1 








<-L 


eng 


th: 


70'> 


\ 












































1 
















1 




































Le 


rel 






f 
















1 












+( 


^ 





\ 
























/ 


















\ 




















\ 








































\ 
















/ 






\ 




















/ 




















\, 














/ 








\ 


\^ 


N 


































^ 


v^ 


\ 








J 














s 


"^ 


\ 


































\. 


'> 




/ 


V 
















































K 


.w 


60 


^' 


/^ 


y 


















































I 


,.w 


05' 


.i 










650 


00 

















__ 


r 


cai 


H. 


w. 


i51. 


■' 




M 


















o 
































W 


oob c 

aplt-Tv 


o. 




C 


1 


PI 


:;kry 


:;o. 




1 
























/• 




^ 














(>i 


la 










-; 


V. 
























'' 




\ 


H 


L.W.6 


38.2 








T 


1 










°?l 


D€ 


nni 


sEi 


ant 




G» 


o. > 


V. f 


mith 


1 ,\ <• 










( 


Ihas 


.P. 


Kn 


igh 






_^ 


s 










1^ 




S 








i^ 






<%^^ 

V"/'?' 










»; 












o J 


S 










?* 




1 




+ 




^ 




joo" 


oC^ 


y'o 






3 




8 


^° 


+ 






















f 




TX 


^'^^ 


?-," 




+ 




X 


A 








\ 


•ll 


?(S 


k 


S 


>reli 


n. ' 


A" 




-H H 




^' 








B 




A 


^\r4^'-: 


s 




'V 


w 




io"! 


'e. 


^K 


ih 


ill 


? 


1^1"- 


k^ 


?'= 


0°3C 


.\ 


g~ 




V 


\ 




c.= 


8°L. 






;: 


^ /N.36i12'e. 


*" 








r 




5^^^[4l 


1 


1 




j 


T.= 


42"'i)6'V 
339;3' \ 


k 










A = 
T.= 


srap , 

617.0 






i 1 


600 


00 




\'^ 






^1» 


A 




'Jh 


C.= 
K.= 


647.7', ^ 
881.05' 










L 


R.= 


1018. ?; 

716.(78 


1 


1 


hn^ 


80 




, 










3 
(I 




•A 


12 


90 


















13 


00 





















FiQ. 9. 



124 



RAILROAD SURVEYS 



averaged |- mile. Stakes were set every 100 ft. on tangents, and every 
50 ft. on curves. Special pains were taken with the instrument work 
and measurements, in order to avoid the chance of serious errors in 
the center line after construction commenced. The speed of location 
parties was usually limited by the amount of clearing that could be 
done, but the number of curves and the rough character of the ground 
were also large factors in limiting the speed. 

Each party cost from $35 to $40 per day, being allowed all expenses 
in addition to salaries. 

Table 1 gives the cost per mile of the completed surveys. It is to 
be noted that this is the total cost, .and includes office rent, purchase 
of instruments and supplies, general expenses, all salaries, field ex- 
penses, and the preparation of final maps, plans, profiles, and estimates, 
with everything in readiness to make contracts for the line. 

TABLE 1. 









Amount 
spent. 


Miles op Surveys: 


Average 

cost per 

mile. 


Average 


Company, 


Prelimi- 
nary. 


Location. 


Total. 


cost per 
mile of 
location. 


(!) 

L. K. R. R 


(2, 
$25 076.83 

19 812.77 

20 466.68 
6 651.98 

19 249.94 


(3) 

428.19 
509.03 
241.75 

84.56 
162.51 


(4) 

193.85 
105.23 
113.70 

38.17 
151.29 


(s) 

623.04 
614.26 
355.45 
122.73 
313.60 


(6) 
$40.31 
33.25 
57.58 
54.20 
61.34 


(7^ 
$129.36 


Z. M. & P 






188.28 


B.&E.R.R... 
P. B. & T. R. R 
B. &N. R.R... 






180.00 
174.28 
127.23 


Totals 


$91 258.20 


1 426.04 


603.24 


2 038.28 


$45.00 


$151.53 







Column 7 gives the cost per mile of actual location, including 
preliminary lines. Columns 3 and 4 show that there were from 2 to 5 
miles of preliminary lines run for each mile of location, except on the 
Buckhannon and Northern line. Table 1 also includes 302 miles of 
check levels, the cost being distributed among the various accounts. 
The data for the Parkersburg Bridge and Terminal line include surveys 
and soundings for the Ohio River Bridge. The cost per mile includes 
the topography on practically all lines, except on the Zanesville, 
Marietta and Parkersburg line, where it was taken only on the located 
lines. 

The cost shown in Table 1, being the total charge against engineer- 
ing from the inception of the project to the beginning of construction, 



RAILROAD SURVEYS 



125 



^|2280+90Expansion 

^B,R.= 832.97 

: Rail 832.92 
280+30 




126 



RAILEOAD SURVEYS 



contains a few items which might well be charged to other accounts 
than location. Instruments purchased could be a credit; some elaborate 
property surveys and bridge surveys could be charged to construction, 
but they probably are not large enough to have much effect on the 
cost per mile. If taken into account, they would reduce the cost. The 
cost on the Little Kanawha and on the Burnsville and Eastern Division 
was increased considerably owing to much work being done during a 
bad winter, when the weather was very unfavorable. The cost on the 
Parkersburg Bridge and Terminal line was increased by a large amount 
of property surveying in the city, and by the surveys for the bridge. 

The cost on all the West Virginia lines was increased by the 
immense amount of chopping and clearing necessary. When mountain 
laurel was encountered, all the axemen that could be worked could not 
keep a location party moving. 

The influence of the weather is a large item in the cost of these 
surveys ; a line run in the middle of winter may easily cost one-quarter 
more than if run during more favorable weather. 

The average cost of one mile of preliminary or location survey, 
determined from a detailed study of the daily reports of field parties, 
of office work done, and similar data, is shown by Table 2 (to the 
nearest dollar), and is believed to be very close to the actual figures. 

TABLE 2. 





Average Cost of One Mile: 


Company. 


Of preliminary. 


Of location. 


Of location, in- 
cluding prelimi- 
nary. 


L. K. R. R 


$25 
23 
35 
31 


$74 
79 

105 
94 


$99 


Z. M. &P. R. R 

B. &E. R. R 


102 
140 


B. &N. R.R 


125 







The figures in Table 2 include all expenses, as in Table 1. 

Table 1 shows a large variation in the cost of surveys on different 
divisions, the cost varying from $128 to $188 per mile, with an average 
of $151. On the assumption that lines located for comparison or 
similar purposes should be included in the average, one-third should 
be added to these amounts, as previously noted; the cost would then 
be as follows : 



RAILROAD SURVEYS 



127 



t^^ 



^M^l 



i^ "^ S" 



f^S-"'' 



"Ill 






s;^:! 
^ 







is 



^ ^ ^ ^- ^. 

V 't- ^ H ^ 



^ a? 












5^ 



i*! 



J^ ^ ^ 

^ ^ s> 

<^ o ^ 









vJ vj V 

Ci <:> ^ 






?tt ^ ^^ ?^ * 

0^ ^ Ci Ci "2> 



r 



ci ^ (); 



til. 

«! ^ ^ 






III? 



Oft- ^^^^ ^ 



Co 

1- 



oj^'N 









(DO 



I 



0) 

I 

Hi 

I 



r/3 

> 
"A 

> 
O 

w 



o 

w 
o 

M 
O 

O 

w 

o 

H 
& 

w 
o 



128 RAILROAD SURVEYS 

Low $171 per mile. 

High 251 " " 

Average 202 " " 

Throwing out of the account the mileage of abandoned lines, branch 
lines, etc., and charging the entire cost to the main line, terminus to 

Cni OKQ Of) 

terminus, would give ~^ — = $278.23 per mile, which would be 

rather expensive. This, however, is not a fair assumption, and should 
not be considered, because many miles of lines not needed to determine 
the main line were located for other reasons and purposes. Therefore, 
the plan of throwing out only duplications, for comparisons, as shown 
in the preceding paragraph, gives the correct average cost per mile 
for the development of the country, including actual comparative 
locations where needed. It should also be borne in mind that a large 
proportion of this duplication was necessary, owing to the laws of West 
Virginia, which require an actual line, located on the ground, and a 
complete map and profile of that line to be filed with the Secretary of 
State, and at the county seat, before a railroad company has any 
rights, of priority or otherwise, to that route or line. This required 
complete locations for all proposed branch lines, a large number of 
which were located, and also a complete location over any route for 
which it was desired to obtain rights. For these reasons, the lines 
located account for the excess of the mileage over the actual length of 
the main line. 

On the basis of Table 2, it may be assumed that, where the route 
has been previously determined within such narrow limits that the 
preliminary and location lines are of equal length, the surveys will 
cost from $100 to $140 per mile. This is borne out by the results 
on the Buckhannon and Northern line where the location and pre- 
liminary lines were practically equal and the cost was $127 per mile. 

These two statements may be combined and put in the following 
form: 

To locate one mile, including an equal length of pre- 
liminary lines, cost from $100 to $140; average $115 

To locate one mile, final location, including from two to 
five times as great a length of preliminary lines, cost 
from $128 to $188 ; average 151 



RAILROAD SURVEYS 129 

To locate one mile, final location, including from two to 
five times as great a length of preliminary lines, and 
one-third of a mile of location for comparison, cost 
from $171 to $251 ; average $202 

A tabulation of the mileage of the Buckhannon and Northern line, 
with reference to the actual length of line to be built, and showing 
how the results agree with the averages deduced from Table 1, is as 
follows, the Buckhannon and Northern line being used because the 
conditions there make it the best average of "all conditions" encountered 
on the various lines. 

Total miles located 151.29 

Miles of main line contracted for 80 

Miles of main line not contracted for 4 

Miles of connecting line located, but which 

may or may not be built, about 26 110.00 

Making actual miles 110 

Leaving duplications, comparisons, etc 41.29 miles. 

110 miles cost $19 249.94 = $175 per mile. 

Results. — The results obtained by studying and surveying the 
neighboring country, will amply justify the initial cost, and, without 
doubt, the final results will prove that it is cheaper in the end. A few 
thousand dollars judiciously spent on surveys will be more than made 
up for in construction, not only in the actual cost, but in a better line. 

That the desired results seem to have been obtained, may be shown 
briefly as follows: 

1. — Easiest grades: 0.5% west, 1.0% east (all the latter in the 25 

miles near the eastern end), other roads using 1.5 per cent. 
2. — Curvature: 8° maximum, as against 12° on other roads. 
3. — Short line: This is seen best on the map. The Little Kanawha 

line is 31 miles shorter than the river, the only water route. 

The Ohio line is 16 miles shorter than the existing lines 

between termini. 
4. — Occupation of critical points : Another line, since built, through 

the Burnsville and Eastern Division (the Burnsville and 



130 RAILROAD SURVEYS 

Eastern line never having been constructed) and a few miles 
on Little Kanawha Division, had to locate on and occupy said 
lines for about 15 miles, although using 1.5% grades and 12° 
curves. It could not have been built for its present cost 
without making arrangements with the Burnsville and Eastern 
and Little Kanawha Railroads. 

5. — The definite places previously determined on were, of course, 
reached. 

6. — It being decided, after many surveys — by which the advantage 
of this line proved so apparent — to build the line as described, 
no complete estimate of the cheapest possible line in first 
cost was prepared, but a suflficient number of estimates were 
made to show that the line as located would certainly fulfill 
this requirement. 

All the work was carried out under S. D. Brady, M. Am. Soc. C. E., 
then, as now. Chief Engineer of all these companies, who gave much 
personal attention to the work, and to whom the writer is indebted for 
assistance and information in preparing this paper, and for permission 
to publish the maps and costs. 

The writer was Assistant Engineer on the Little Kanawha Division, 
in charge of location; on its completion he became Assistant Chief 
Engineer of all the companies, both on location and construction. 



DISCUSSION ON RAILROAD SURVEYS 131 

Discxjssioisr 



Clinton S. Bissell, M. Am. See. C. E. (by letter). — The author Mr. Bisseii. 
states specifically that the paper is not a theoretical discussion of 
railroad location, and later he cites as the "desired results": 

1. — Easiest grades possible, against west-bound traffic; 

2. — Lightest curvature; 

3. — Shortest line, etc. 

He also mentions ruling grades of 0.3% on the Little Kanawha 
Division, and 0.5% on the Burnsville and Eastern Division. The 
momentum grade at Mile 20 on the latter division is introduced 
"to avoid a long detour" ; and, in fact, the reader is impressed through- 
out the paper by the evident great advantage of the "shortest line." 

To the writer, a railroad is simply an industrial plant, instituted 
to furnish transportation, and the "desired results" to be realized, are 
embodied in the one problem, viz., to obtain the line of greatest effi- 
ciency. The author has mentioned no measure of efficiency, nor has 
he indicated any limitations as to the duty required of the line. With- 
out these, the writer is at a loss to understand the basis upon which 
the location was made, unless it was merely to get a tolerably direct 
line, with gradients as low as good direction would allow. The author 
has either disregarded or withheld one element which Wellington 
rightly cites* as of "overwhelming importance," namely, the probable 
amount of traffic. The second important element — that of easy 
gradients — the author has observed without correlation to the probable 
traffic. The other two elements — curvature and distance — are, within 
reasonable limits, minor details of location, and yet apparently the 
author has given them great attention, to the negligence of one of the 
two more important elements. 

Since nothing can well be further removed from the theoretical 
side of the problem of location than the actual dollars and cents which 
the road can either earn or save, let us consider such part of the profile 
as the author has shown in Fig. 3, between Burnsville and Buckhannon. 

From Buckhannon westward it is noticeable that the two stretches 
of ruling 0.5% grade amount to about 3 miles in 33, or only 10% of the 
distance, neglecting the momentum grade which, according to the 
author, can be removed. 

Assuming at random that the line will carry annually 3 000 000 
tons of freight westward, with heavy locomotives and modern high- 
efficiency cars, the 0.5% ruling grade permits a train load of 1 700 
tons of lading; and the same train, returning with about one-third 
as much lading, can surmount the 1% grades going east. Taking the 

* "The Economic Theory of Railway Location," by the late A. M. Wellington, M. Am. 
Soc. C. E. 



132 DISCUSSION ON RAILROAD SURVEYS 

Mr. Bisseii. train-mile cost at $1, for convenience, the annual cost of operation 
can be roughly estimated. 

It is rather surprising that the two ruling 0.5% grades were not 
made 0.3% to accord with those on the Little Kanawha Division. For 
the sake of example, it will be assumed that this could have been done 
at the expense of 4 miles more of distance at a gross cost of $50 000 
per mile. For the 0.3% grade the train load going west would be 
2 420 tons of lading ; and the same train would return east with about 
one-fifth as much lading, against the 1% grades. Hence, the following 
figures may be deduced: 

3 000 000 ^ 1 700 X 33 X $1 west-bound $58 240 

Same trains returning east-bound 58 240 



Annual operating expenses on 0.5% grades $116 480 

3 000 000 -4- 2 420 X 37 X $1 west-bound $45 870 

Same trains returning east-bovmd 45 870 



Annual operating expenses on 0.3% grades $91 740 



Difference $24 740 

Less interest at 4% on cost of 4 miles of line 8 000 



Annual saving $16 740 

At 4%, this saving justifies a present expenditure of $418 500 to 
reduce the ruling grades from 0.5% to 0.3%; this reduction increases 
the efficiency of the line by a saving of 14% in expenses; and probably 
there would be enough capital left to take out the momentum grade at 
Mile 20. 

Unless the matter of location be viewed from some such standpoint 
as this, the writer does not understand how the "shortest line" practi- 
cable on the "easiest grades possible" and on the "lighest curvature" 
can be decided. The intimate correlation of these three elements does 
not permit of such a decision, except upon some basis other than 
simply the contour of the country. 

A few years ago the writer made several surveys and locations along 
the south shore of Nova Scotia between Cape Sable and Shelburne. 
The drainage, flowing down to the sea, crossed the direction of survey 
about at right angles. The route along the seashore was too irregular 
to justify a low line for any considerable distance, and a high line 
several miles inland touched but few towns, because these, for the most 
part, were situated along the shore. The final finished location was a 
compromise between these two lines. The whole route was wooded, 
especially on the slopes of the ridges where the maximum grades 
occurred. The specifications of the Dominion and the Provincial 



DISCUSSION ON KAILEOAD SUEVEYS 133 

Governments fixed the maximum grade at 1.5% and the maximum Mr. Bisseii. 
curvature at 6 degrees. 

In conducting these surveys, the writer found it necessary to take 
contours only on the ruling grades and a few other points on the line. 
The grades were first cut through, and the profile taken, after the 
manner mentioned in this paper, from the summit downward. To 
accomplish this object, the transit was frequently carried several 
miles ahead of the measured line and set on the summit or other 
governing point, and the profile run from a temporary bench-mark, 
the elevation of which was assumed or estimated with a barometer. 
The measured line was then brought forward, connected, and run 
again over the maximum grade. Particular attention was then given 
to the contours and to the topography of the stream crossing which 
was usually found at the foot of the grade. The great advantage in 
this method is the fixing of the proper point of connection at the foot 
of the grade with the assurance of passing through the governing point 
ahead. The writer believes that notwithstanding the necessity of 
running over the grade twice, time is saved by avoiding much backing 
up and subsequent alteration of the line back of the foot of the grade. 
It gives a continuous chainage, saves much contouring, and produces 
a preliminary map which lies close to the location at all the critical 
points. 

The author's figures on costs are very interesting and valuable; but 
the maximum rate of progress mentioned, 7 miles of preliminary line 
and 4i miles of location per day, has never been equalled by the writer. 
He considers it excellent work for one party to accomplish even half 
of these distances in one day. 

F. La VIS, M. Am. Soc. C. E. — This paper is a valuable addition to Mr. Lavis. 
the somewhat scanty literature relating to the subject, and, undoubt- 
edly, will be of considerable interest to those connected in any way 
with the conduct of surveys for the location of railroads. It is 
especially so to the speaker, as the methods used agree very closely 
with those developed by him between 1898 and 1902 on the Choctaw, 
Oklahoma and Gulf Railruad.* Those surveys resulted in the con- 
struction of more than 800 miles of railroad, and, probably, at least 
as many more miles of location were actually staked out on lines which 
were not built. The speaker has no exact figures at hand, but it is 
probable that more than 5 000 miles of preliminary were run during 
that time. 

During the last ten or fifteen years these general methods, with 
slight rcinor variations to suit local conditions and individual ideas, 
have become fairly well recognized as standard on most of the larger 
railroad systems of the country; at any rate, those engaged to any 

* Transactions, Am. Soc. C. E., Vol. LIV, p. 104 (1905). 



134 DISCUSSION ON RAILEOAD SURVEYS 

Mr. Lavis. extent in building new lines and extensions, recognize the absolute 
necessity of such methods, in order to determine without doubt that 
the line finally adopted is such "that no other line can be built through 
the same country, with the same or better ruling grades, with less 
expenditure, at the same unit prices."* 

The data relating to costs show close agreement with the costs 
presented by the speaker in the paper referred to, and, although, in 
both instances, they are somewhat higher than a great many locating 
engineers think their surveys have cost, it will generally be found that, 
when these low costs are given, the fact has been lost sight of, that 
on any survey for the location of a railroad, the costs, both of the 
field parties, and a proper proportion of the expenses of the main 
and division offices, covering all work from the reconnaissance to the 
completion of the final maps and profiles, ready for the contractors to 
bid on the construction, and the real estate agents to purchase right of 
way, should be included. 

The surveys described by Mr. McFetridge were made in 1902 and 
later; the costs previously given by the speaker were for surveys made 
during the latter half of 1902, so that the time agrees fairly closely. 
During 1903, of course, it was easier to get men; and salaries, as a 
rule, were comparatively low, so that the costs, during the latter part 
of the time when these surveys on the Little Kanawha were made, 
would tend to be low rather than otherwise; this is shown also by the 
salary list. 

Comparing the costs with those presented by the speaker,t it may 
be noted that the costs of the field work of the preliminary and location 
lines agree very closely. The average cost of the lines, as given in 
Table 2, shows the preliminary to have been about $28, and the location 
$88 per mile, whereas the costs given by the speaker were $26 and $71, 
respectively. The costs of the completed location, as described in the 
paper, vary from $127 to $188 per mile, averaging $151, on the basis 
of the mileage of located line staked out; or from $171 to $251, aver- 
aging $202, if the mileage of adopted location be taken, which latter 
may be compared with the cost of $192 per mile of adopted location 
as given by the speaker. The costs of field parties per day are given 
by Mr. McFetridge as between $35 and $40, whereas those described 
by the speaker cost from $42 to $65, and averaged $46. The somewhat 
lower cost per day of the field parties on the Little Kanawha, although 
the cost per mile of line was higher, is probably due to the difference in 
organization. On the Choctaw, although all lines were subject to 
revision, of course, by the Chief Assistant Engineer and the Chief 
Engineer, the Locating Engineers, whose whole expense, in addition 
to that of an Assistant Locator, was charged to the field party, were 

* " Railroad Location Surveys and Estimates," p. 8. 
t Transactions, Am. Soc. C. E., Vol. LIV, p. 133. 



DISCUSSION ON RAILROAD SURVEYS 135 

expected to be able to decide questions which, on the Little Kanawha Mr. Lavis. 
Railroad, were apparently not left to be settled by the Assistant Engi- 
neer in charge in the field. 

Judging from the description of the country and the number and 
length of preliminary lines run, it would appear that the two surveys 
are fairly comparable. The country described by Mr. McFetridge is 
in many cases rougher in detail than that through which the surveys 
described by the speaker were made, though, in the latter case, the 
character of the country required the detailed examination of a much 
wider range in order to determine the best location. This is shown 
by a comparison of the mileage of preliminary with that of location; 
on the Little Kanawha lines there were 2.37 miles of preliminary to 
1 mile of location, whereas on the lines described by the speaker, the 
average was 3,14, a little greater than 30% more. The effect of the 
rugged detail is shown in the considerably increased cost of staking 
out the located lines on the Little Kanawha. 

The generally prevailing idea as to the cost of railroad surveys is, 
that in the United States it should be about $100 per mile; and in a 
review* of the speaker's "Railroad Location Surveys and Estimates," a 
well-known railroad engineer made the statement that "much excellent 
location has been made in America in not easy country for less than 
$50 per mile of located line, and this cost includes the needed pre- 
liminary for that located line." He adds, referring to the supposed 
high costs of the Choctaw surveys, "We must conclude * * * that 
the parties were not made up of veterans." Veterans or not, the 
speaker does not believe that any line, outside of a tangent across a 
prairie, and sometimes not even that, can be located and proper maps 
completed, for any such sum as he mentions, and that for any ordinary 
country a rough estimate of $200 per mile would be much nearer correct 
for properly conducted surveys and the preparation of a proper set of 
plans ready for construction, these latter, of course, not including 
details of structures. The trouble with most of the published costs 
of location surveys is that they do not include anything but the cost 
of staking out a location on the ground ; and supervision, office expenses, 
equipment, costs of preparation of proper maps, profiles, etc., are 
omitted. 

There is considerable difference, also, in the amount and kind of 
information developed by surveys. As an instance of a not uncommon 
form presented to a contractor as the location profile. Fig. 12 shows 
a part of a profile of a line in the Southwest, which was 
actually sent to a New York contractor, and on which he was asked 
to make his bid, this bid to be a lump sum for the whole line. Eig. 13 
shows a section of a typical profile of located lines such as is 
usually required by most of the railroads in the West. The material 

* Technical Literature, April, 1907, p. 175. 



136 DISCUSSION ON RAILROAD SURVEYS 

Mr. Lavis. in the cuts is explored by making test pits and soundings, as are also 
the foundations for structures, when necessary, and the information 
obtained is shown on the profile. The quantities in cuts and fills are 
given, with the approximate length and location of the haul; the 
quantities in structui'es are estimated and shown. In short, the profile 
contains all the information necessary to prepare a close estimate of 
cost, and gives the contractor all the information, as to the nature of 
the work, which he requires in order to make an intelligent bid. 

The profile shown by Mr. McFetridge in Fig. 9, as the standard 
form for a location profile, seems to the speaker to be defective, inas- 
much as it shows no quantities, gives little indication of the material 
to be excavated, or the amount or direction of haul, and no indication 
of the form or location of bridges, culverts, or other structures, or the 
quantities of material required to build them. The information 
required for the proper design of the larger bridges is shown in good 
shape on the situation plan, Fig. 10, but it is very desirable to show 
on the profile the type of structure and the space it occupies. 

The speaker believes it to be of the greatest importance, not only 
to show the distribution of quantities on the location profile, but also 
to insist on their being worked out on the profiles of all projected loca- 
tions. There is nothing which gives a man as completely the intimate 
knowledge of the line which it is really necessary for him to have as to 
work out the distribution, and get a clear idea of what he is to do 
with the excavation, and where he is to get material for the embank- 
ments. 

One very important rule in regard to the final location, the speaker 
believes, should be made inflexible: After the profile of the located 
line has been platted, a temporary grade line fixed, tentative determina- 
tions of bridges, culverts, etc., made, and preliminary quantities calcu- 
lated and distributed, this latter being most important, the Locating 
Engineer should take the profile into the field, and, walking over the 
line, should study carefully the situation on the ground in the light 
of the information which he then has, which, by that time, should be 
almost complete. This final review, if conscientiously done by a com- 
petent man, will often save many thousands of dollars in the construc- 
tion, and avoid costly mistakes. 

Although, in general, the methods described by Mr. McFetridge 
correspond closely to those described in the speaker's paper, previously 
referred to, a few minor differences may be noted. In keeping level 
notes the speaker prefers to keep the rod readings on the turning points 
entirely separate from the rod readings for elevations. It seems hardly 
necessary to have columns for grade and cut-and-fill in the location 
level notes, these belong more properly in the level notebook of the 
Resident Engineer, in which he records the check levels run on assum- 
ing charge of his residency. 



DISCUSSION ON RAILROAD SURVEYS 



137 



Mr. La vis. 



1 ) 1 1 1 1 1 


t? A r. tI jf -lir J ' 1 1 1 




1 ^ O ^ 


1 / "^l*" 1 


_j jL" --J 1 rt L' 








~r - ^ -S °- 




f^ inji.'^ ^^ I -nr 




u 11 


V" 


\ _1 


1 I 


nT" *■ 


T> ,*«^ -lir 




^""^N. ■ i 


w J*^ J ' fe S 






^ 


\ 


t ^ 


I . p_ 








c2 ■»on r 


K 4280- V 


/ 


r 


A 


M 


v- ' 


^ it 






cS S L ^^ 


j^ Y 


# ^ o 


e 13 ^ " \ WT * *^ «* 




"r ~oS- 


^ ■ s ^ 


■■ " -J Ah W 


'^ ** *■ H^ O ° 


tn ^ ^ 


M \[) ^ 


to iff "D^ 1^-^ ^^1 IST J ' « 




ly 


tL 


N 


A O TJ iV -^^ 




\ 


j 1 


S JOAA 




ij] r 


I^V 


X ^4 


M 




Si 


i t 


^ A ^ 




Ot S § rt 


5 


L 


\ 


^ 






^1 


I\t 




g ■ 


^ 




± t 


i 


"T 4 


T 








3 It fe 


U. :=St 


-K 3 


t .. ± _ 35 ± 



138 DISCUSSION ON RAILROAD SURVEYS 

Mr. Lavis. The titles of profiles are generally more convenient wlien placed 
on the outside at each end; even if put inside they should be at each 
end rather than in the upper left-hand corner. Profiles accumulate 
rapidly on location, and with the title inside and only at one end, the 
chances are that two-thirds of them will have to be entirely unrolled 
to find the one sought. 

The question of the advantages of separate sheets as compared with 
rolled maps has been discussed too often for the speaker to hope to 
say anything new at this time. Mr. McFetridge, however, although he 
evidently prefers the separate-sheet method, recognizes the necessity of 
having a large rolled map of the whole line in order to get a compre- 
hensive view of the whole situation. The speaker takes exception to 
his very positive statement that the separate-sheet method of obtaining 
topography is quicker and better than any other; to say the least, this 
is open to argument. The speaker, who has had some little experience, 
prefers the method described in his paper. 

For small general maps of the country, on which — in addition to 
the principal roads, trails, streams, etc. — may be shown the various 
preliminary lines as well as the final location, the speaker prefers a 
uniform scale of 5 000 ft. to the inch. These maps are useful, not only 
to the field parties and in the office of the Chief Engineer during the 
location, but also in discussions and conferences with the officials of the 
road, and others who only require general information. They are also 
of convenience to contractors in bidding on the work, in order to en- 
able them to form a better idea of its accessibility, etc. They should 
always be prepared at the beginning of the survey, and information 
should be added as the survey progresses. 

The speaker hardly sees the necessity of staking out a location for 
the purpose of choosing between two apparently equal routes, or, as 
was done in this case, for comparing the lines on either side of the 
larger watercourses. If the preliminary lines have been run with any 
degree of skill, and the topography has been properly taken, there 
should be no difficulty in getting, from a projected location on the map, 
and an inspection of the ground where this projected line would lie, 
all the information necessary for the comparison of any line or lines, 
except under the most exceptional circumstances. 

The salaries named seem to be somewhat lower than those usually 
paid. In the West it is not uncommon to have a locating engineer 
in charge, with an assistant who looks after the details of the field 
work, leaving the former free to explore the country thoroughly, and 
devote more attention to the office work and the projection of various 
locations than he otherwise could. With a conscientious locating 
engineer this is a very good arrangement, and allows a much more 
careful threshing out of the country than would otherwise be possible. 
The author states that, on the surveys described: 



DISCUSSION ON EAILKOAD SURVEYS 



139 



Sj 






f.T. 197294- 75.45 | 


















































\ 




\ 1 






























































% 






-* 


1 


1 




I 












w 


















































g-s 




f 


I 










to" 




















































\ 




i 






H 


















































-II 


cnO[ 


ce 


. 




^ 






























































G 


O 




'W 


M 




•^ 




- 


g-- 


^ 




















































,- 


--' ' w 




X 








iV" 






- 
















































w 

p 


? 


H^ 




- 








■~ 
















































c 




c 


.-■ 




1 




^s 




























































/ 






1 


































































/ 








1 
































































/ 










1 














P.O. 1971741 


52.12 


































9;50- 


rfi 


/| 






















757 












































D: 




f^^ 






' 


^ 










































o 
















i" 




- 






<* § I 


ir; « 


'or 


i(g; 






\ - 






































S 


^ 




/ 




1 








J J" 


u ,..„ 






^^ 






































o 




s 






1 












1 






Y 






% 






































;^ 


» 


t3- 






1 


?/ 










1 






A 






4 


\ 






































S^ 








J/ 












i 








I-' 




" 


\ 








































J* 


1^ 








1 


















































i't^g^-rS,!/ 








irea 






























Summary IMile 373 
Excavation Embankment 
Pay Quantity 43430 
>i " borrow 3C59 
SweU 3264 
From Mile 374 101217 151570 
151570 












|_&S-i\' g \ 




inJkeJ 












































r 




--1 


vJ 












































*-* 










B- 


























•n 


























o 


D 
















;> 


















i 


























k! 
















1 


Di 


. ArPH 1ft An 


1 












.■^ 


















f^~ 


J 


-"^ 






§ 


1 


24"Vit, 


pipe 




1 






:« 






-" 




















o 








^ 


..^ 




o 


























- 






















5' 












\ 


yf^ 




















^ 






















s 
p 


'' 




o 








Ap 




















s 






J'i^ 




















§ 


1 


1 


i 


^. 


^3 


nfi 


~-> 


















^- 


' ' t;'=''' 






























>> 






1 






































tB'O' W 










*1 






^ 


1 


























- 












y "=■ ■■ 1 


/ 




, 


.4- 


..^ 


■^ 




1 




































o 


M- ^•-r-T' 








1 




















- "i 


:|^ 
























e:^ 




< 






















































fe^ 


P' 










a 














































SI 
■5 




- 


!Kf 






■| 




















^ 


























\ 




1 
















g 










































t 






;; 


lit;? 


\' 


















(1 














































o -^ o 


J- 


1 


^ 


> 
















10( 




































































W 










































^ 


_J-^ 


h 
































































/ 






1 


y 






















3 






































/ 




























< 

p 

a 








=^" 




































^ 




































g 






























/ 










































s. 




























^ 




bd 














Dfi 


.re 


a 68 Ac. 
















2> 




















o 






/ 










„ 










o ,'«" 


C.I 


Pipes 
























g 






-« Si 


















1 








% 


!»^ 


U- 




— 


i_ 


_. 


_ 


—62000 


— 


































!^ 




































































g 


S? 












OS 






















I 


.m!. 84l.i9 




























m 


'^ 








^ 
























Onl 


ul) li{678+30 












































V 
























































g 


" 
















\ 
























1 
































Q 


;x 


















^ 






















j 
































k| 


.^' 
















y 
























1 


















































^ 


■s 






















1 














--k' 


> 
































g 
























1 








■^ 




































o 


hWS^ 
















1 














? 












s 
























u 


1/ 


i^ 


y 
















j 






Y 






y} 












1 
























^ 


^ 


■T 


























































^ 




.-- 


t^ 
















































S' 


e 














\r'\ 




y 


<- 


7 


-- 


T- 


j- 


— 


— 


- 


T-384'87 


— 


L' 






R 






«H 












s- 




1 ' /^ " £ 




o 














1 




1 1 










p 






^ 












5 


Z- 


— '-^-^^ 


r- 


— 


f- 




^ 






Dr. Area 1905 Ac. 
















■^ ? 


; 










"ilc- 


-7 ^ 


xr 






^ 


i^ 




1 210 ft. Pile c"!! Frame Tr 


estie 1 


(a 






k 


/ 










4^1 


1/ 




^ 


1 i~ 


:^ 


slzr* 


]v 


aJ \ 














1 






^ 


- 




- 


— ,M- 


Z7 












fc 


1=^ 
















o 




1 






















\'\& 


>?> 














»i 


\\ 


s 














s 




1 




















n r^ 





Mr. Lavis. 



140 DISCUSSION ON RAILROAD SURVEYS 

Mr. Lavis. "The Assistant Engineer was not expected to spend all his time 
with the party, but to be with it enough to see that work was going on 
satisfactorily, and that lines were being run over the proper routes; 
the remainder of his time was spent in a thorough study of the country, 
picking out routes to be examined, and looking after his necessary 
office work and correspondence." 

The man who is responsible for the work of the party in the field 
should be with it most of the time, and if this devolves on the man in 
charge, it keeps his nose too close to the grindstone to allow of a 
proper perspective from which to view the problem as a whole, to say 
nothing of taking up practically his whole time. A proper, careful, 
study of the many possible projections cannot be made at night by a 
man who has put in already 10 or 12 hours' hard work in the field, 
although it is really surprising how much work men on locating parties 
manage to get through and thrive on. 

The salary of the topographer is given at $65 per month, and, with- 
out intending in any way to cast any reflections on the young men who 
filled these positions, this seems to be altogether too little for good men. 
Taking topography has always seemed to the speaker to be a very 
simple operation, and it really is, if one has any sort of instinct for 
it, but it is remarkable how few men seem to grasp the idea, even after 
considerable experience. The speaker believes the topographer should 
be paid only slightly less than the transitman. 

The fundamental idea of the proper principles governing the stak- 
ing out of the located line is stated very concisely by the author in 
the third paragraph on page 116, and this idea cannot be impressed 
too strongly on all connected with work of this class. "In staking out 
the location, the aim was to get a profile to correspond with the pro- 
jection, and not to get the lines in exactly the same relative positions 
shown by the projections." In other words, there is to be no slavish 
following of the projected location as a line on the map, but rather 
to reproduce, on the ground, the line which the studies of the map and 
the general situation have shown is there, and is the best. 

As the speaker has already pointed out elsewhere,* it is just here 
that modern methods differ from the old. It is neither the old paper 
location, made without reference to conditions in the field, nor the 
work of the Genius, described by Mr. Whineryf in his discussion of the 
speaker's paper, who threw his curves at the hills, as the cowboy throws 
his rope at a steer, and sometimes hit them and sometimes missed. If 
he missed, so much the worse for the railroad. The modern method 
collects all the available data affecting the location, eliminates defi- 
nitely the unavailable lines, projects on the map the line which best 
fulfills all the requirements of the situation, and then uses all the 

* " Railroad Location Surveys and Estimates," p. 125. 
t Transactions, Am. Soc. C. E., VoL LIV, p. 143. 



DISCUSSION ON RAILROAD SURVEYS 141 

intelligence of the Genius in reproducing that line on the ground. Mr. Lavis. 
No extraordinary mathematical calculations are required to lay out 
the located line from the projection, only plain common sense and 
locating skill. 

No amount of accuracy in making the original surveys and platting 
the map can compensate for lack of intelligence in transferring the 
projection to the ground, and extreme refinement of accuracy is 
wasted, if proper intelligence is used. 

The statement on page 111, contains, in a very small compass, the 
essence of the requirements which should govern the conduct of all 
location surveys; a "most minute study of the land, a thorough knowl- 
edge of its character, the running of many lines, the comparison of 
many estimates," are absolutely necessary in every case. The Locat- 
ing Engineer, and by that is meant the man really responsible for the 
location, should be thoroughly steeped in the details of the country, 
and the word steeped is used advisedly, in the sense that there should 
be time errough in all cases to allow the information acquired to be 
taken in, digested, and to permeate him through and through. 

One other point brought out in the paper, which is most important, 
is that actual surveys of all possible routes were made; as the author 
points out, this does not mean that an indiscriminate topographical 
map of the whole country was made, or was necessary. In spite of all 
that has been written about eliminating certain routes by eye, and per- 
haps by a hand-level, there is no question that absolute proof, by actual 
survey, should be furnished to the people putting up the money, or, 
at any rate, to those responsible for its expenditure, that there is no 
doubt that the location adopted is the best possible under the given 
governing conditions. The ability and so-called "eye for country," 
ascribed to the older generation of locating engineers, which many of 
them undoubtedly possessed, are just as necessary to-day, but must be 
supplemented by scientific methods, adequate surveys, and much hard 
work; and this is necessary, not only in order to be able to furnish 
proof that the line adopted is the hest, but generally in order to find 
this line. As stated also in the paragraph referred to, "On a first 
examination of this country one w^as inclined to say offliand that 
it was impracticable to leave the river * * *; but * * * a 
thorough study * * * gave the question a different aspect." This 
experience has been duplicated many, many times, and in many cases 
the study was not made, and the railroad paid the bill later. 

Not many years ago it was not an uncommon thing to hear some of 
the Geniuses with "eye for country"— and not much of anything else 
^talk about the foolishness of trying to get a 0.5% line through a 
1.0% country. The results obtained by these surveys are only another 
instance of the possibility and practicability of finding these lines, if 



143 DISCUSSION ON RAILROAD SURVEYS 

Mr. Lavis. One knows how to look for them. This description, of work actually 
accomplished, should be borne in mind by all who have to obtain money 
for such surveys, and, when necessary, should serve to stiffen their 
backbone to demand all that is necessary to carry out such surveys in 
an adequate manner. 

The author is to be congratulated on the very clear and concise 
manner in which the information is presented, and his recognition 
and presentation of the many vital points which differentiate proper 
methods of making location surveys and getting results from the 
methods too often adopted, of simply stringing together a lot of 
tangents and curves over which a train can be run somehow or other, 
if the stockholders' money holds out long enough and no one else builds 
a line alongside over which trains can be operated at half the expense. 
Mr. Lewis. E. W. Lewis, M. Am. Soc. C. E. — This excellent paper is of much 
interest to engineers engaged in railroad location. The methods 
outlined by the author are very nearly the same as those used by the 
speaker in making many surveys in widely scattered sections of the 
country. The make-up of field parties is identical with that used by him 
on the Northern Pacific Railway, and quite similar to general practice. 

The method of plotting topography in the field, while advantageous 
in many respects, would be a difficult performance in a rainy country, 
such as the Pacific Northwest, where it rains almost every day for 
long periods during the winter season. 

The speaker prefers taking the notes in a topography field book. 
These books being ruled like cross-section paper, the contours can 
' be plotted and the exact measured distance from the center line given 

for each 5- or 10-ft. contour, making it very easy to plot on a large 
map in the office. This plotting may be done in the evening, but, 
ordinarily, it is left until the next day. With a draftsman regularly 
employed to keep up the map, each day's field work may be mapped 
on the succeeding day. 

The method of taking the contours, by measuring from the center 
line with a tape, and obtaining the difference of elevation with a 
hand-level, is in accord with the speaker's practice as being the best 
when an accurate contour map is desired. He has learned by sad 
experience that the saving of a little time in taking topography by 
pacing distances, etc., is often the cause of a much greater loss of 
time when the paper location is run out. 

In regard to spiraling curves, if the line is likely to be a final 
location, particularly if the prospect is good for immediate construc- 
tion, the speaker prefers to run in the spirals with the location, 
although, unquestionably, it saves considerable time not to do so. 

The costs of surveys, given in the paper, are interesting as showing 
the wide differences on various lines. The approximate costs of some 
surveys with which the speaker has been connected in various parts 



DISCUSSION ON RAILROAD SURVEYS 143 

of the country may be of interest. On the north bank of the Mr. Lewis. 
Columbia River, in Washington, over very rough, difficult ground, 
largely on steep slopes, with numerous short tunnels, where, in some 
instances, it was necessary to lower men down the cliffs with ropes in 
order to get the line staked, the cost for a stretch of 32 miles was 
about $325 per mile. In Massachusetts, through an average country, 
largely wooded, the cost of 20 miles of location was about $150 per 
mile. In North Dakota, 60 miles of location over a rolling prairie, and 
crossing one deep, broad valley, cost $110 per mile. In another section 
of the same State, over rolling prairie, the cost was $80 per mile. 
These figures are for the total cost of the final location, including all 
preliminary, trial, and abandoned lines. 

George L. Dillman, M. Am. See. C. E. (by letter). — The writer is Mr. Diiiman. 
impressed by the absolute lack of value of this paper to the Profession, 
the uselessness of the conclusions drawn, and the pernicious possibility 
of these conclusions being applied to some other case of railroad 
location. No two locations are alike, and hardly any two are com- 
parable in time, cost, or value. 

The writer was transitman on two locations which are worth noting 
as being extremes in his practice : The first was in Cottonwood Canon, 
on the Grand River in Colorado, between Dotsero and Glenwood 
Springs. Afterward the Denver and Rio Grande built its line on the 
opposite side of the canon. There was a great deal of clearing, and 
the topography was such that direct measurement was often impossible, 
so that a great deal of triangulation for distance was necessary. This 
location cost more than $600 per mile. 

The second location was on the North Platte River, in Nebraska, 
extending more than 100 miles northwest from North Platte. The 
country was open, there was no clearing, the tangents averaged about 
9 miles, and the curve angles between them were small. This location 
cost between $17 and $18 per mile. 

These locations were each fully staked on the ground. The costs 
include maps, profiles, and computed estimates, as well as all prelimi- 
nary work. The parties had about the same number of men and 
equipment, except that more axe-men were added to the Grand River 
party. Each party worked efficiently and sometimes for long hours. 
If there was any difference, the cheaper location was the more carefully 
made. Tracings of each map were filed with the Interior Department, 
accompanying applications for rights of way across Government land, 
and each location was ready for the construction crews. 

These locations are probably not extremes; there may have been 
cheaper locations made; undoubtedly, there have been more expensive 
ones. What is desired is to bring out the fact that there is no possible 
connection between the costs of railroad locations; and that a full 
knowledge of the cost in one case cannot be applied to any other case. 



144 DISCUSSION ON RAILROAD SURVEYS 

Mr. McFet- W, S. McFetridge, M. Am. Soc. C. E. (by letter). — The writer 
appreciates the friendly criticism and discussion of his paper, but 
does not think necessary to add to it, except in reply to two or three 
questions. 

In reference to Mr. Bissell's criticism that no measure of efficiency, 
probable amount of traffic, etc., were given, these are about the first 
matters which must be determined, and must be known before the 
final details or "desired results" of the surveys are decided on by the 
officials. It was so in this case, and, while a discussion of them might 
have proved even more interesting than the matters mentioned in the 
paper, they did not come under its scope, and were intentionally 
omitted, as were the reasons for adopting such grades, etc. 

In regard to Mr. Lavis' criticism of the final profile, the distribution 
of material, etc., for estimating and for the contractors, was compiled 
on a profile almost in the form he gives. This profile was made in the 
Chief Engineer's office, with the assistance of the Locating Engineer. 
It seemed to work well, but in cases where headquarters could not be 
easily reached, the method suggested by Mr. Lavis, showing all this 
information on field profiles, would be better. 

Mr. Dillman apparently has made some wrong assumptions in 
regard to the paper, as the writer knows of no statements contained 
therein from which he could assume the meaning indicated by his 
conclusions. No one who knows anything of surveys would expect any 
two to be identical in method or cost, but the same general principles 
will apply to the conduct of any up-to-date survey; and there can be 
no better way of judging what a survey should cost, than to know what 
a similar one has cost, and what method and organization were used. 

The writer does not claim that any iron-clad rules should be 
followed, and believes that each survey is a separate study by itself. 
A knowledge of how diiferent surveys have been conducted, and what 
they have cost, however, should be of value to engineers engaged on 
them^ and likewise to the people who are paying for them, and who 
often think a survey is a matter of minor cost and importance. Know- 
ing from previous experience that it is -sometimes impossible to get 
enough money to make any kind of a decent survey, the writer hoped 
to help a little toward bringing about a just appreciation of the needs 
of such a survey. 



AMERICAN SOCIETY OF CIVIL ENGINEEES 



INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1122 



COMPUTATION OF STRESSES IN OPEN-WEBBED 
ARCHES WITHOUT HINGES. 

By C. W. Hudson, M. Am. Soc. C. E. 



The principles underlying the following method are often used 
in finding the stresses in certain classes of statically indeterminate 
structures built of elastic material, but, as far as the writer is aware, 
they have never been applied to finding the stresses in an open-webbed 
arch rib without hinges. 




If such an arch rib be considered cut, at the crown, by a vertical 
plane, it is divided into two parts which are each statically determinate. 

For the purpose of illustration, assume an arch rib as shown in 
Fig. 1. Consider the rib cut at the center by a vertical plane; this 
divides the rib into two parts, as shown in Figs. 2 and 3, the vertical 



U6 



STRESSES IN OPEN-WEBBED ARCHES 



member at the center being considered divided into two equal parts, 
one of which is taken with Fig. 2 and the other with Fig. 3. 

It is perfectly clear that a load at any point on the left half 
produces stresses in the truss members of that half which may be 
determined by statics when each half acts as an independent canti- 
lever truss, and that no stresses can be produced in the right half by 
this load when the two halves are not connected. 

h 




If we suppose the two halves to be connected by joining the points, 
a and h^ then, under any condition of loading, these two points have 
the same motion, as they are the same point. Therefore, for a load on 
the left half it is clear that the stress in any member of that half is 
that due to the given load when the half is considered as a separate 
structure modified by the stress acting between a and h caused by 
connecting these points, and the stress in any member of the right 
half is that due to the stress acting between a and h caused by con- 
necting these points. No matter what the nature and direction of the 
force produced by joining a and h, it may be considered replaced by a 
horizontal and a vertical component. 

If we suppose the two halves to be further connected by joining 
the points, c and d, then a and h have the same motion, and c and d 
have the same motion under any condition of loading. Therefore, 
for a load on the left half and the right and left halves connected at 
a and h and at c and d, the stress in any member of the left half is 
that due to the load on that half, when the half acts as a simple 
cantilever modified by the forces produced at a and c by connecting 
a to & and c to d, and the stress produced in any member of the right 
half is that due to the forces at h and d caused by connecting a to & and 
c to d. It hardly need be mentioned that the forces, due to connecting 
a to & and c to d, at a and c, are equal and opposite to those at h and d, 
respectively. 



STRESSES IN OPEN-WEBBED ARCHES 



147 



He 



No matter what the amount and direction of the forces acting 
between a and h, and c and d, they may each be replaced by their hori- 
zontal and vertical components. For the two halves connected and sub- 
jected to any loading, the forces produced by thus joining them may 
be represented by the four forces of Fig. 4, Ea, Va, He, and Vc, the 
direction of which may be as shown or the y^ 

opposite and the magnitude of which is to be 'Ha 

determined. Knowing that for elastic struc- 
tures deflections are proportional to the loads 
that produce them, four equations between 
the unknown forces. Ha, Vat He, and Vc, and % 

certain easily determined deflections for the ^^°- ^• 

half arch acting independently, may readily be written. The solution 
of the four simultaneous equations will determine the unknown forces 
acting at the crown of the arch for each half. These equations of 
condition for finding the unknown forces at the crown will now be 
written for a symmetrical arch. The general method is equally 
applicable to an unsymmetrical arch, however. 

Let a load of 1 lb., at any point, x, cause the point, a, to take a 
new position, a', and the point, c, a new position at c', as shown in 
Fig. 5. A, 





Fig. 5. 

Then A^ = vertical deflection of a due to a vertical load of unity 

at any point, x; 
Ag = vertical deflection of c due to a vertical load of unity 

at any point, x; 
A3 = horizontal deflection of a due to a vertical load of 

unity at any point, x; 
A4 = horizontal deflection of c due to a vertical load of 

unity at any point, x. 



148 



STRESSES IN OPEN-WEBBED ARCHES 





Fig. 7. 



Let a vertical load of 1 lb. at a cause the point, a, to take a new 
position, a', and the point, c, a new position at c', as shown in 
Fig. 6. 

Then d^ = vertical deflection ^'^ 

of a due to a 
vertical load of 
unity at a; 
cZg = vertical deflection 
of c due to a 
vertical load of 
unity at a; 
dg = horizontal deflection of a due to a vertical load of 

unity at a; 
d^f^ = horizontal deflection of c due to a vertical load of 
unity at a. 
Let a vertical load of 1 lb. at c cause the point, a, to take a new 
position, a', and the point, c, a new position at c\ as shown in 
Fig. 7. 

Then d^ = vertical deflection of a due to a vertical load of 
unity at c; 
d^ = vertical deflection of c due to a vertical load of 

unity at c; 
d^j^ = horizontal deflection of a due to a vertical load of 

unity at c; 
d^^ = horizontal deflection of c due to a vertical load of 
unity at c. 
Let a horizontal load of unity at a, as 
shown in Fig. 8, cause the points, a and c, to 
take the new positions, a' and c', respectivel5^ 
Then cZ, = horizontal deflection of a 
due to a horizontal load 
of unity at a; 
cZg = horizontal deflection of c due to a horizontal load of 

unity at a; 
cZj^g = vertical deflection of a due to a horizontal load of 

unity at a; 
d^^ = vertical deflection of c due to a horizontal load of 
unity at a. 







Ycif—dii, 



Fig. 8. 



Fig. 9. 



STEESSES IN OPEN-WEBBED ARCHES 149 

Let a horizontal load of unity at c cause the points, a and c, to 
take the new positions, a' and c', respectively, as shown in Fig. 9. 
Then d^ = horizontal deflection of a due to a horizontal load of 
unity at c; 
dg = horizontal deflection of c due to a horizontal load of 

unity at c; 
c?j5 = vertical deflection of a due to a horizontal load of 

unity at c; 
d^Q = vertical deflection of c due to a horizontal load of 
unity at c. 
The four simultaneous equations will now be written, from which 
the four unknown forces, Va, Cc, Ha, and He, acting at the crown and 
due to a load of unity at any point, x, may be determined. 

For the points, a and h, and the points, c and d, connected, the 
motion of a must equal the motion of h, and the motion of c must 
equal the motion of d. 

The downward motion of a = J^ — V^ d^ — ^^c^h + ^a ^hs + H^d^^, 
and 

the downward motion of 6 = + F^ (^i + V^ d^ + H^ d^^ + H^ d^^, 
and these must be equal, as they are for the same point. 
Therefore, 

^1 - ^a ^h - n ^h + H, rZ,3 + H^ d,, = V, d, + F, d, + fl-„ d,, + 

and V,d,+ V^d,=^ ( 1 ) 

The downward motion of c = J^ — V^a'^-z — ^^c '^h + -^a ^u + ^c '^i6' 
and 

the downward motion of d = -\- V^ d.^ -\- V^ d^ -\- 11^ d^^ + He '^^i6' 
therefore, 

^2 - T^a ^k - n ^\ + Ha f^4 + ^c ^^16 = T^a '^2 + V, d, + if„ d,, + 

and ^a t^2 + T; ^^4 = ^ (2) 

The motion to the right ot a = /l^ — F^ fZp — T^^ cZjj + H^^ d^ + If^ cZ^, 
and the motion to the right of 6 = — V^ dg — V^ cZjj — H^ d^ — H^ d^, 
therefore, 

^3 - V, d, - V, d,, + H^ d, + H^ d, = - y^ d, - V, d,, - H^ d, 
- H^ d,, 
and H^d, + H^d,==—^ ( .3 ^ 



150 STRESSES IN OPEN-WEBBED ARCHES 

The motion to the right of c = — ^4 + V^ d^^ + V^ d^.^ + ff^ d^ + 

and the motion to the right of cZ = + V^ rtj^ -f- V^ d^.^ — H^ d^ ■ — H^ (Zg, 
therefore, 

- ^4 + ^a ^10 + V, d,, + 11^ d, + H^ d, = F, d,, + V, d,, - H^ d, 

— He d^, 

and H^d, + H^d, = ^- (4) 

Solving these four equations, the following values for the unknown 
forces are found, in terms of twelve easily determined deflections: 

d. do 

■ V = + /i, ^ — J., ^ 

"' ^'^{d^d^—d^d^) ^2{d^d^ — d^d^) 



H„ = — ^ 



2(d,d,-d,d^) '2id,d, — d^d^) 

' 2 (d, fZg - d^ d^) ' 2 (rfg (Zg — dg fZy) 



de. cZk 



'2(fZgCZ8 — (Ze^y) *2((Z5(Z8 — cZgcZy) 

If the vakies of A^ to //^, and rZj to fZg, as found by actual computa- 
tion, are in the direction heretofore assumed, their numerical value 
is to be inserted in the foregoing expressions without regard to sign, 
if they are found to be in a direction opposite to that assumed, they 
must be inserted with a minus sign. 

From Maxwell's Theorem, it is known that: 

^2 ^^ ^3' ^'^^ ^6 "^^ ^T 

For any particular configuration of members in the arch rib, very 
much simpler expressions for Y^ and Vc may be written. 

For the case assumed, to illustrate this paper, a load at any panel 
point other than a, 

makes J^ =■ A^i ^^^ ^h = ^^4' ^^^'^ therefore F^ = 0. 

A load at panel point, a, 

makes ^^ = //g + (^*'i — ^2)' ^^^^ ^3 ^'^ ^^^^^ ^ ^^41 therefore F^ = - . 
For any panel point of the rib loaded except a 

F=A. 
' 2fZ, 

And for panel point, a, loaded: 

Y, = 0. 



STRESSES IN OPEN-WEBBED ARCHES 151 

Restated, for any panel point loaded other than a 

And for panel point, a, loaded: 

For such a configuration at the crown as indicated in Fig. 10, that 
is, v/here the diagonals adjacent to the crown meet at the top of the 
crown vertical, d^ = d^, and d^ and d^^ are no longer equal. 



a b a b ?L^ 




c d - ^ c d 



Fig. 10. Fig. 11. Fig. 12. 

Then, for any panel point loaded other than c, 

z/j = //g, and therefore, 

And for panel point c, loaded, Jj = z/j — (d^ — (Z3), and J.^ = d^, 
and therefore, 

F, = 0, and F, = ^ . 

For such an arrangement of members at the crown as shown in 
Fig. 11, it is readily observed that 

F„ = 0, and F, = ^^ 

for any panel point loaded. 

For such an arrangement at the crown as shown in Fig. 12, it is 
seen that 

V. = ^. and n = 

for any panel point loaded. 

The expressions for the horizontal crown forces remain the same 
for all the cases selected. 

Letting 2 (drd^ — d^^) = n, and remembering that d^ = d^, we 
have, for the value of the horizontal thrusts, 



Ha 


= 


-4 


n 


— A 


d. 


and 


^c 


= 


+ 4 


n 


+ ^4 


^5^ 

n ' 





152 



STRESSES IN OPEN-WEBBED ARCHES 




The above expressions for the unknown crown forces have been 
written from assumed values of the twelve deflections of the half arch 
acting as a statically determined structure. For any actual problem, 
these deflections, of course, should be determined accurately, both as 
to direction and magnitude. The expressions for the value of the 
unknown crown forces will always be of the same general form as the 
above, but may differ in some of the signs of the terms. 

The resulting plus values for Va, Vc, Ha, He, indicate that they 
act as assumed, and the minus values that they act in the opposite 
direction. 

TeTnperature Stresses. — Let 
Fig. 13 show the motions of a 
and h, and c and d, for a rise 
in temperature when each half 
of the arch is considered as a 
separate structure, the full lines 
representing the normal figure 
and the dotted lines the figure 
changed by a rise in temperature. 

Let Fig. 14 show the motions of a and h, and c and d, for a 
corresponding fall in temperature under the same conception as 
before. 

Then S^ = vertical deflection of a and h due to a certain change 
in temperature; 
82 = vertical deflection of c and d due to a certain change 

in temperature; 
S3 =^ horizontal deflection of a and h due to a certain change 
in temperature, 
and 84 = horizontal deflection of c and d due to a certain change 

in temperature. 

Four equations for the unknown crown forces due to connecting the 
two halves may now be written for both a rise and a fall in temperature, 
using the notation as previously defined. 

For a rise in temperature: 



Fig. 13. 



Fig. 14. 



The upward motion of a = + 5j + V^ fZj + V^ c\ — H^ d^^ — H^ d^^, 

^a f^3 — -He f^i5' 



^'a^h-Vc^h 



and the upward motion of 6 = + <5, 
for the two halves connected. 

These motions must be equal; therefore, V^ cZj + V^ d^ =: (5) 



STEESSES IN OPEN-WEBBED ARCHES 153 

In an entirely similar manner, the three following equations may 
be written: 

VJ^ + Vad^ = (6) 

Had, + H,d, =-8, (7) 

Had, + Hed, = -8, (8) 

For a fall in temperature : 

Vad^ + 7e^3 = (5a) 

Vad^ + V^d^ = (6a) 

HJ, + H,d, = 8, (7a) 

Had, + H,d^ = 8, (8a) 

An examination of Equations (5), (6), (5a), and (6a), shows that 
Va and Vc are zero ; therefore, the only crown forces due to temperature 
are horizontal. 

For a rise in temperature, remembering that fZg = f?^, and making 
(d^ d^ _ d,^) = g : 

da , ^ df. do dn 

(7g (Zg f7g cZj 

« = - ^'(d, d, - df)^^^(d, d, - cV) = - ^'1,'^^'J- 
For a fall in temperature, 

^- = + '= (MT?^ - '' (ip. - <) = + *4' ~ '' '^ ^"" 

(Zg CZg fZg fZg 

The foregoing gives a simple method for finding the stresses in 
every member of the arch due to a load at any point, or for any 
desired change in temperature. Under this method, the work may be 
easily divided into several sets of independent and easily understood 
operations at the very beginning of the computations, thereby leaving 
only the small and easily managed theoretical part for the person in 
charge. 



AMEEIOAN SOCIETY OF CIVIL EN&INEEES 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1123. 

HYDRO-ELECTKIC POWER IN CANADA.* 

By Cecil B. Smith, M. Am. Soc. C. E. 



With Discussion by Messrs. H. Holgate, E. J. Beugler, and 
Cecil B. Smith. 



Introductory. 



To view justly the progress which has been made in Canada, in 
developing water-powers, up to the present time, certain premises must 
be established, climatic conditions realized, the trend of population, 
habits, and occupations of the people pointed out, and, above all, the 
peculiarly favorable topography of the country described. 

The population of Canada is familiarly spoken of as agricultural 
in character, but this is only relatively true, as large and important 
industries flourish throughout the Eastern Provinces, and, indeed, it 
is painfully evident that even in this young and sparsely populated 
country the increase in population is almost entirely urban in char- 
acter, except on the great wheat plains of the Middle West. 

Industries are based on labor, raw materials, and power, and only 
in the last named would the ultimate conditions be unfavorable to 
continued progress if coal were the only source of power to be relied 
on, because Canada is very peculiarly situated as regards its coal areas. 

* Presented at the meeting of September 1st, 1909. 



HYDRO-ELECTEIC POWER IN CANADA 155 

Four areas are at present known, as follows : 

The Nova Scotian, which is on the extreme Atlantic seaboard, and 
produces a good bituminous coal, available as far westward as 
Montreal ; 

The Souris, which is in Southwestern Manitoba, and pro'duces a 
low-grade lignite of limited area and value; 

The Alberta, which covers a very large area extending from the 
United States boundary to north of Edmonton and from the plains 
well up into the Rocky Mountains, the quality ranging from lignite 
to anthracite, and, as regards both quantity and quality, probably one 
of the most valuable coal deposits in the world; and 

The Vancouver, which is very extensively distributed over the 
island and up some of the mainland valleys, chiefly developed for 
export to the Pacific Coast cities of the United States. 

Thus it appears that coal is available on the extreme west and 
east, but for 2 000 miles, covering much of the most populous and 
fertile center of the country, it is necessary to import coal from the 
United States for heat and power. 

From a hydraulic point of view, however, the country has been 
peculiarly favored, as a glance at the map will show. 

The Hudson Sea is deep set into the heart of the country, and the 
height of land between the waters flowing to this sea and to the 
St. Lawrence drainage system is from 1 200 to 1 500 miles long, and, 
which is of more importance, it is only from 1 000 to 2 000 ft. above 
the sea, is not of a mountainous character, but is a perfect network 
of lakes, swamps, rock, and forest, and affords a good reservoir and 
regulator for hundreds of large rivers, which, as they reach the parent 
St. Lawrence, descend over precipitous Laurentian rapids, admirable 
as sources of hydraulic power. This condition obtains all over 
Canada as far west as Eastern Manitoba, and, as the rainfall averages 
from 25 to 40 in. per year, with no marked tendency to drought, and 
as the long winters store up so much moisture that the low-water 
period often merges into the period of autumn rains, the pinch of low 
water is not in many cases a serious menace, except in March of severe 
winters. On the great western plains the topography and climate both 
militate against hydraulic values, as will be dealt with later. 

In British Columbia there are two diverse regions: the eastern 
and central portions have a low rainfall, and the glaciers feed the 



156 HYDRO-ELECTRIC POWER IN CANADA 

rivers throughout the summer, so that the low-water period is in the 
late summer and autumn. 

Along the Pacific Coast there is an excessive rainfall of from 
100 to 150 in. per year, but it is distributed unequally throughout the 
seasons, and only by making use of storage lakes can extieme low- 
water conditions be avoided. 

Until twenty years ago, hydraulic power often formed the basis 
of growing industries and determined frequently the location of the 
centers of population, but the larger centers of trade were developed 
by rail and water transportation, and in many cases were quite distant 
from such sources of power. 

With the advent of electric power transmission, however, these 
cities offered a favorable market for the sale of light and power from 
distant sources, and private companies, and in some cases municipali- 
ties, have invested large sums in hydro-electric constructions and trans- 
mission systems to serve the adjoining districts. 

While much has been accomplished, much remains to be done, and 
the discussion of present accomplishments which is herein given is 
intended to mark a mile-stone on the road. 

Of the many great rivers of Canada, the Mackenzie, Nelson, and 
St. Lawrence are by far the greatest, having drainage areas of 660 000, 
370 000, and 565 000 sq. miles, respectively ; but, outside of these, there 
are three districts worthy of notice from a hydraulic standpoint, 
namely, the Atlantic Coast, the Pacific Coast, and that great array of 
rivers which, in addition to the Nelson, flow into the Hudson Sea. 

Atlantic Coast. 

On the Atlantic seaboard the configuration of the Provinces of 
Prince Edward Island and Nova Scotia does not provide large drainage 
areas, and, also, in the presence of extensive coal measures and coal 
mining, the incentive to hydro-electric development has been lacking; 
indeed, a recent installation of steam-electric equipment having a 
capacity of 600 kw. at the pit mouth, from which electric energy is 
transmitted at 11000 volts to Amherst, N. S. (distant 6.5 miles), may 
be but the pioneer of much more extensive installations of a similar 
character in other parts of America. 

Farther westward in New Brunswick, there exists the condition 
more usual in Canada, namely, rivers having large drainage areas. 



PLATE X. 

TRANS., AM. SOC. CIV. ENGRS. 

VOL. LXV, NO. 1123. 

SMITH ON 

HYDRO-ELECTRIC POWER IN CANADA 




HYDRO-ELECTRIC POWER IN CANADA 157 

and having their sources in swamp, forest, lake, and muskeg, in 
which snow and frost linger late in April, or even until May, and 
feed these rivers throughout the dry months of August and September; 
indeed, it is usual in Canada to find the extreme low-water period in 
March, before the frost has ceased to hold the moisture from flowing. 
The Miramichi, Restigouche, and St. John are great rivers, and 
extensive hydraulic constructions may be expected for grinding pulp 
alone, in addition to serving near-by towns and cities by transmitted 
electricity. 

A most interesting and extensive development has been commenced 
at Grand Falls, N. B., by the Grand Falls Power Company, where the 
St. John River, vnth its sources in hundreds of lakes of Northern 
Maine, falls abruptly 115 ft. over a slate barrier. 

For these hydro-electric works the type of construction which has 
been selected is that of vertical, 10 000-h.p., 3-phase, 25-cycle, 11 000- 
volt units, operating under a head of 130 ft., and the designs, under 
which contracts for the construction of the works are about to be let, 
provide for a hydraulic capacity of 80 000 h.p., with a present ma- 
chinery installation of 40 000 h.p., which will generate current for 
ferro-manganese smelting, pulp grinding, and for transmission to the 
cities of Woodstock (70 miles), Fredericton (125 miles), and St. John 
(165 miles). An interesting feature of the design is that the gener- 
ators are to be placed in the wheel-pit immediately above the turbines, 
with a provision for the proper ventilation of this subterranean chamber 
as well as careful water-proofing to ensure satisfactory electrical 
conditions. The tunnel tail-race will be placed entirely below the 
level of the tail-race water, and the draft-tubes from each unit, there- 
fore, will be brought into the top of the tunnel from one side of the 
wheel-pit; this, it will be noted, is a variant on the methods adopted 
by the Toronto and Niagara Power Company in its works at Niagara 
Falls, Ont. 

Pacific Coast. 

On the Pacific Coast, the hydraulic conditions are very different 
from those in other parts of Canada, for the great mountain ranges 
are not good reservoirs, and, although the glaciers feed the rivers 
until late in the summer, yet at other seasons, owing to the light rain- 
falls east of the Coast Range, the rivers become very much reduced 



158 HTDEO-ELECTEIC POWER IN CANADA 

in volume; thus the excessive floods of the Fraser Eiver, for instance, 
make it unsuitable for hydraulic development for power purposes. 

Near the Coast, however, the rainfall is excessive, and the existence 
of small lakes has facilitated the construction of several interesting 
developments, serving Vancouver and the adjoining territory. As in- 
stances, those of the Stave Lake Power Company and the Vancouver 
Power Company may be mentioned. 

The works of the Stave Lake Power Company, now under construc- 
tion, are 35 miles east of Vancouver, on Stave River, where a head 
of 90 ft. is to be vitilized. The drainage area, estimated at 360 sq. 
miles, provides a large volume of water, owing to the excessive rain- 
fall, which varies from 100 to 150 in. per year. The flow in the 
river is estimated to vary from 2 000 to 50 000 cu. ft. per sec, making 
the river regulation somewhat difiicult. 

The equipment, as at present designed, will consist of five units, each 
of 5 000 h.p., and it is considered that the electrical output from this 
installation has a good prospective market in Vancouver and West- 
minster to which the current is to be transmitted at an ultimate 
voltage of 60 000. 

The Vancouver Power Company, Limited, has a most interesting 
development at Butzin Lake, 18 miles northward from Vancouver. 
This has been made possible by diverting the water of Coquitlam 
River at Lake Coquitlam into Butzin Lake by a 2^-mile tunnel 
having a section of 73 sq. ft.; thus the water required for operating 
the plant is conveyed from the Coquitlam drainage area, estimated at 
100 sq. miles, the run-off varying from 300 to 5 000 cu. ft. per sec. 
under an excessive rainfall of 150 in. per year. At Butzin Lake 
the water is taken from the head-works by steel pipes to a power- 
house, 1 800 ft. distant, and located at sea level, 400 ft. lower. 

The generating equipment is typical of the Pacific Coast, being 
originally composed of four units, each of 3 000 h.p. ; two Pelton wheels 
for each unit drive a 3 000-h.p., 2 300-volt generator at 200 rev. per 
min., in addition to which there is also a recent installation of a 
10 000-h.p. generator driven by four Pelton wheels, thus making an 
eqviipment having a total generating capacity of 22 000 h.p. 

The electric current from this statiorf is stepped up to 23 000 
volts and transmitted to Vancouver, New Westminster, Burnaby, Lulu 
Island, and elsewhere, and is used for operating electric railways, 



HYDRO-ELECTRIC POWER IN CANADA 



159 



salmon canneries, car shops, saw-mills, and for general municipal and 
industrial purposes, thus serving the requirements of an urban popula- 
tion of about 90 000. 

The transmission line consists of double towers each carrying two 
circuits of 3-phase, No. 2, B. & S., copper cable, and crosses navigable 




CANADIAN POWER TRANSMISSION SYSTEMS 
VANCOUVER DISTRICT 
VANCOUVER POWER COMPANY 

SCALE OF MILES 



Fig. 1. 

water at one place where there is a span of 2 Y50 ft. suspended 150 ft. 
above the water. 

The total single 3-phase circuit length of lines is 135 miles, and 
the transforming capacity at present in service is 14 650 kw. at the 
generating station and 16 400 kw.. at the various sub-stations ; all 



160 HYDRO-ELECTRIC POWER IN CANADA 

these transformers are air-cooled, and are of different sizes up to a 
maximum of 2 500 kw. per phase. 

In the interior of British Columbia the sparse population — almost 
entirely south of the main line of the Canadian Pacific Railway — is 
engaged either in agriculture or mining, and thus it happens that, 
aside from small developments at Revelstoke and Nelson for municipal 
purposes, the only interesting hydro-electric plant in this district is 
that of the West Kootenay Power Company, built to serve the mines 
and smelters at Rossland, Trail, and in the Boundary District. It is 
located at Bennington Falls, on the Kootenay River, 9 miles west of 
Nelson, B. C. For the new power-station— ^which supersedes a smaller 
adjoining station previously built — the design is of the vertical type, 
owing to the considerable variations in tail-race level, caused by 
gorging below the falls in high-water periods. 

The hydraulic head varies from 65 to 55 ft., being least at high 
water, and the minimum capacity of the river is estimated at about 
6 000 cu. ft. per sec, being the run-off from a mountainous drainage 
area of nearly 10 000 sq. miles, which, owing to the glacial sources 
of the river, gives a low-water period in winter and a high-water 
period in summer. The station is a very interesting example of com- 
plete concrete construction, including even the intake channels and 
draft-tubes of the turbines, and is constructed for an equipment of 
four units, each of 6 000 h.p., nominal rating, vidth the turbines of 
30% overload capacity. At present, two units are installed and a third 
has been ordered, in addition to which the old station has an equip- 
ment of 2 700 kw., and, when required, serves Rossland and Trail over 
a 31-mile, cedar-pole line at 20 000 volts on two No. 0, B. & S., 
3-phase circuits. 

At the larger station, the triplex, high-pressure, oil-thrust pumps, 
etc., are developed very much along the lines of the Niagara works, 
and are designed to supply oil to the thrust-bearings at 250 lb. per 
sq. in.; the governors were built by the I. P. Morris Company, of 
Philadelphia, adapted and modified from the well-known Escher- 
Wyss type. 

The step-up transforming station supplies 60 000-volt current for 
the new cedar-pole transmission circuits to Phoenix Mines and Grand 
Forks Smelters, and to Greenwood Mines, distant 79, 69, and 83 miles, 
respectively; and its present equipment consists of six single-phase. 



HYDKO-ELECTRIC POWER IN CANADA 



]C1 




162 HYDRO-ELECTEIC TOWEli IN CANADA 

water-cooled transformers, each of 1 850 kw., nominal rating, each 
three-phase bank being in a fire-proof compartment; there is also a 
bank of three single-phase, 1 250-kw. transformers held in reserve for 
the Rossland and Trail Districts at 20 000 volts in the event of a shut- 
down of the older station. 

The uses for which the electric current from this station is adopted, 
in this district, are quite varied, consisting of mine-hauling by 
electric locomotives, hoisting by induction and synchronous motors, 
lead smelting, handling of various machinery, pumping, air com- 
pression, and general industrial and municipal purposes. 

The total transforming capacity in seven sub-stations is 31 000 kw., 
and the active connected load is 21 000 h.p. 

Immediately east of this hydro-electric development are located the 
great Crow's Nest Pass coal areas, but the competition is not effective, 
in so far as service in the mining district is concerned. 

It is also anticipated that some of the adjacent mountain railways 
will become electrified in the near future, as the operation of rack- 
railways on 4:i% grades has not proven satisfactory or economical. 

Perhaps the most interesting feature of these works is the satis- 
factory maintenance of a 60 000-volt transmission line over a heavy 
mountainous country on a substantial but ordinary cedar-pole line. 
This is in sharp contradistinction to the prevailing tendency in the 
more level districts of Eastern Canada, where it is considered good 
practice to use steel towers with long spans for high-voltage trans- 
mission, but doubtless the decision was influenced by the inaccessible 
nature of much of the country traversed, the plentiful supply of large 
cedar timber in British Columbia, and also the high freight rates on 
structural steel, to the interior, from the Pacific Coast. 

The adjoining zinc and lead smelters at Nelson, B. C, will be 
supplied by these works or by the municipal hydro-electric works 
of the City of Nelson, on the other side of the river from the large 
power-station just described. 

This municipal power-station has now a capacity of 5 000 h.p. and 
an equipment of 1 000 h.p., which is about to be increased to 2 500 h.p. 
by the installation of a second vertical unit. The general construction 
has been carried out on a basis of four units having a total capacity 
of 5 000 h. p. 



hydlio-electric power in canada 163 

The Great Plains. 
Immediately east of the Rocky Mountains, in Canada, there are 
two great river basins, with many subdivisions, but of these the 
Mackenzie River, emptying into the Arctic Sea, lies in a practically 
uninhabited region; however, it is interesting to note that emigration 
and settlement will soon change all this, and, undoubtedly, in a few 
generations populous communities will occupy the upper water-sheds of 
this river, and will require all the tools of civilization, and among 
them the electric current from hydro-electric stations. Flour mills are 
already being established on the head-waters of the Peace and Atha- 
basca Rivers. ^^^^ ^^^^^^ ^^^^^ -^^^^^ 

The Nelson River drains an extensive water-shed which is a country 
in itself. All the inhabited portions of the Provinces of Manitoba, 
Saskatchewan, and Alberta, except a small area tributary to the 
Missouri and Mississippi Rivers, are drained by the various tributaries 
of this river. 

This water-shed is divided into three main drainage areas, the 
Saskatchewan, the Red and Assiniboine, and the Winnipeg Rivers. 
From a hydraulic point of view, each of these rivers exhibits distinct 
and dissimilar characteristics. 

The Saskatchewan River, of 125 000 sq. miles drainage area, rising 
in the eastern slope of the Rocky Mountains by many sources, fed 
from glaciers, and descending to the plains through many mountain 
gorges, offers, near its sources, reasonable opportunities for hydraulic 
development, and two companies are already active in the preliminary 
stages of contracting to supply Calgary, Alta., with electric power from 
hydro-electric stations to be erected on the head-waters of the Bow 
River. Doubtless, as population in the foot-hills increases, similar 
undertakings will be carried out to meet the requirements of near-by 
towns and cities. 

Eastward of the foot-hills, however, the conditions change, and 
the main branches become chiefly dependent on the precipitation on the 
great plains, which is rather light and either evaporates or passes 
off in spring or autumn floods. The character of the country also 
changes; the streams occupy deep, vdde gorges cut down into the 
plains, and have light grades and shifting silt and sand bottoms where 
dam construction would be difficult and expensive. 



164 HYDRO-ELECTRIC POWER IN CANADA 

Speaking generally, the main valley of the Saskatchewan, eastward 
from the foot-hills, is not adapted to hydro-electric constructions, and, 
as there is coal over a large proportion of this vast area or near it, it is 
not probable that such constructions will ever be numerous, extensive, 
or profitable, except at a point on this river immediately before it 
enters Lake Winnipeg. 

At this place, 250 miles north of the City of Winnipeg, the river 
passes through a series of small lakes and then, at Grand Rapids, 
drops quickly to the level of Lake Winnipeg. The change in level is 
112 ft. in a few miles, and there is thus afforded the opportunity for 
an enormous development of at least 300 000 h.p. whenever the demand 
arises for a power development at that point. 

The Red and Assiniboine branches of the Nelson River, with a 
drainage area of 105 000 sq. miles, are still more unfavorably located 
for hydraulic purposes, as the drainage basin is entirely in a plains 
country, where there is light rainfall and very little lake storage, thus 
the minimum run-off is extremely small in proportion to the areas 
drained, as a consequence of which, with the exception of a few quite 
insignificant opportunities, they may be written off the list of hydro- 
electric districts, not only for the foregoing reasons, but because the 
flood levels are usually bank full, and sand and silt bottoms, over 
wide flats, present expensive conditions for dam constructions. It 
may be remarked in passing that this drainage area is devoid of 
coal, except for the lignites of the Souris Valley; and the price of 
fuel and power is excessive, except where the coal supply is local or 
transmitted electricity from the eastward is possible. 

The Winnipeg branch of the Nelson River drains an area of approxi- 
mately 55 000 sq. miles in the State of Minnesota and in the Provinces 
of Ontario and Manitoba, and, from its peculiar topographical fea- 
tures, affords one of the most favorable locations for the development 
of hydraulic power in Canada, excelled perhaps only by the Niagara 
and the St. Lawrence and superior in some respects to the Ottawa 
River. 

This condition is somewhat, but not greatly, offset by the moderate 
rainfall, which is only from 20 to 25 in. per year over the main area, 
but, aside from this, there is a most extensive network of lakes (of 
which the location, throughout the northern portion, has been, in 
many cases, unmapped until recently). There are thousands of lakes 



HYDRO-ELECTRIC POWER IN CANADA 



165 




16G HYDRO-ELECTRIC POWER IN CANADA 

of all sizes, culminating in the Lake of the Woods. There are 1 600 
sq. miles of water surface on the Rainy River branch and Lac Seul on 
the English River branch, westward of which the two rivers unite 
near the eastern boundary of Manitoba and flow thence into Lake 
Winnipeg. 

On the English River branch of the Winnipeg River no large 
hydraulic constructions have yet been considered seriously, except 
near Dryden, as the country is inhabited only at a few points along 
the line of the Canadian Pacific Railway, which crosses the various 
smaller branches between the height of land and the Town of Kenora, 
but the Transcontinental Railway, now being constructed, will pass 
much nearer the more extensive water-power locations of this river. 

These locations, being only the more extensive ones of the English 
River branch, are rated by the Hydro-Electric Power Commission of 
Ontario at a total of 100 000 h.p., minimum dry-weather flow, 24-hour 
service, and it may be considered that the commercial capacity is at 
least 200 000 h.p. under ordinary conditions. It seems evident that 
the requirements of railway operation, flour milling, pulp and paper 
mill work, and possibly smelting will demand a careful study of the 
specific value of these powers in the near future, as soon as the Trans- 
continental Railway is in operation. 

The Seine and Rainy River branch of the Winnipeg River also 
possesses good hydraulic opportunities, and the Koochiching Falls, at 
Fort Frances, is now being developed on a basis of 14 000 h.p., for 
which the dam construction is now completed and power construction 
is under way. 

The Winnipeg River proper commences at Kenora, where the river 
leaves the Lake of the Woods, and from this point to the mouth of 
the river there is a chain of falls and chutes which, as sources of 
economical power, is difiicult to surpass or equal. 

The two conditions worthy of comment are that at each power 
location the rocky barrier creating the fall is admirable as a founda- 
tion for a dam, thus enabling constructions to be low in first cost, 
and at the same time creating considerable local storage to meet 
daily peak-load requirements. Also, the enormous lake storages, pre- 
viously referred to, very much reduce the flood discharge and increase 
the minimum. The latter may be taken at 13 000 cu. ft. per sec. at 
Kenora and 20 000 cu. ft. per sec. below the junction of the English 



HYDRO-ELECTRIO POWER IN CANADA 



167 



River, the flood discharge being not more than four times these 
amounts. 

Table 1 is a summary of the power sites on the Winnipeg River 
proper, between Kenora and the mouth at Fort Alexander. 

TAELE 1. — Power Sites on the Winnipeg River. 



Location. 


Available 

head, 

in feet. 


Minimum 

electric 

horse-power; 

24-hour 

rating. 


Average 

electric 

horse power; 

24-hour 

rating. 


Remarks. 


Kenora 


18 
45 
46* 

30* 
30* 

22 
13 
30 
10 
15 
21 
13^ 


31 000 
56 400 
83 600 

54 500 
54 500 

40 000 

22 300 
56 000 
18 500 

38 000 

39 100 

23 300 


42 000 
112 800 
309 000 

136 200 
136 200 

100 000 
55 750 

140 000 
46 200 
70 000 
97 800 
58 300 


5 500 h.p. installed. 


Island Falls 




Point du Bois 


20 000 h.p. being in- 
stalled. 


Slave Falls 


Upper Seven Portages 

Lower Seven Portages 

Mc Arthurs Falls 


18 700 h.p. installed at 
Lee Channel. 


Grand Bonnet 

Little Bonnet 




White Mud Falls 




Silver Falls 




Pine Portage Falls 








Totals 




497 200 


1 204 250 











* Including raising the water level by head-works. 

Of the sites mentioned in Table 1, Island Falls, Point du Bois, 
Slave Falls, Upper Seven Portages, and Grand Bonnet afford excellent 
construction opportunities. All but the first two in Table 1 are within 
easy transmission distance of the City of Winnipeg, and all are 
within short distances of two transcontinental railway systems. 

The situation at Kenora illustrates the great value of water-power 
as applied to flour milling. The town has little else to recommend it 
as an industrial center, not being at a competing railway point, nor 
at a bulk -breaking point; neither is there any surrounding population 
or any fertile areas; but, in spite of these apparently unfavorable con- 
ditions, three large flour mills are located there, and use about 4 000 
h.p., with a capacity of 10 000 bbl. per day. Two of these plants 
operate privately-owned hydraulic equipments, and the third pur- 
chases 1 000 e.h.p., at a flat rate of $10 per e.h.p.-year, from the munic- 
ipal power-station. The latter has been built recently at the eastern 
outlet of the Lake of the Woods with dimensions to receive machinery 
having a capacity of 4 000 e.h.p., one-half of which has already been 
installed. 



168 HYDRO-ELECTKIC POWER IN CANADA 

At the main or western outlet of the Lake of the Woods, at Kenora, 
there has also been built for some years, a stone dam from which 
water could be drawn to develop approximately 18 000 e.h.p. at low- 
water, but no further constructions have been commenced. 

At Point du Bois the City of Winnipeg (population 120 000) has 
commenced the construction of a large generating station involving 
a 75-mile transmission line to the city, and the designs contain some 
interesting features. A short (1 200-ft.) head-canal, concrete overflow 
walls, and a rock-filled dam in the main river will raise the water 
13 ft., creating a total head of 46 ft. and forming a lake 5 sq. miles 
in area, thus producing uniform and quiet hydraulic conditions. The 
dimensions of all control works, dams, etc., provide for an ultimate 
equipment of 60 000 h.p. The remainder of the work is to be a solid 
reinforced concrete structure involving an enclosed forebay room, 
power-house, and transformer-station equipped primarily with five 
5 300-e.h.p. units, which will consist of horizontal, double-runner tur- 
bines (draft-chests only) in reinforced concrete wheel-pits and directly 
connected through a breast wall to 3-phase, 60-cycle, 6 600-volt, gen- 
erators making 164 rev. per min. 

The transformation will be to 77 000 volts, by water-cooled, oil- 
insulated transformers having a capacity of 3 750 kw. per phase. 
Similar transformation at Winnipeg will be from 72 000 volts down to 
11 000 volts in a main terminal station, from which 11 000-volt feed- 
ers will distribute to three or more sub-stations for local distribution 
at suitable voltages; the transformers at the terminal station, how- 
ever, are to be oil-circulating, and oil-cooled. The system of remote 
control will be fully applied at both stations. 

The transmission lines, of No. 000 copper, or aluminum equivalent, 
will be carried on double-circuit steel towers spaced 600 ft. apart, each 
alternate tower being of elastic type with two legs only. All tower 
foundations are to be on concrete. 

The city utilities are already numerous, and include Artesian water 
supply, pumped electrically; and the desire to be able to offer cheap 
power to industries, present and prospective, which would otherwise 
labor under a great disadvantage in a city which cannot offer bitu- 
minous coal of a good quality for less than $6.50 per ton, while anthra- 
cite coal is $10.50 per ton retail, appears to be ample justification for 
this ambitious project for a city of 120 000 inhabitants. 



HYDRO-ELECTRIC POWER IN CANADA 169 

The Winnipeg Street Railway Company and its allied companies 
have in operation a hydro-electric station on the Lee Channel of the 
Winnipeg River, and a transmission line 67 miles long supplying 
power to a terminal station in the city. 

This generating station was selected apparently without proper 
investigation and without an eye to the future, and the company is 
already expending large additional sums on the tail-race channel in 
order to insure sufficient hydraulic head and flow of water at low 
water to operate the machinery now installed, which makes a com- 
pleted scheme incapable of further expansion and consisting of nine 
units having a combined rated capacity of 14 000 kw. 

The scheme of development consists of a rock-filled dam across 
the Winnipeg River, which raises the normal water level 4 or 5 ft. 
and forces a flow of water through a 14-mile high-water channel, from 
which already approximately 500 000 cu. yd. of rock have been exca- 
vated. 

The power-house is about 5 miles from the intake of this channel, 
and was designed for an operating head of 42 ft.; this, however, has 
not yet been realized, and it is understood that, when delivering 15 000 
e.h.p. at Winnipeg, the generating station is only operating under a 
head of 36 ft., and is at its present limit of capacity, hence the further 
excavations now in progress. 

The generating station has an open forebay, steel-flume wheel-pits, 
and units of five 2 000 kw., and four 1 000 kw., each operated by two 
twin horizontal turbines; the generators are for 3-phase, 60-cycle, 
2 300-volt current. 

The transformation is to 50 000 volts, and the transformers are 
oil-insulated, water-cooled, of 1 800- and 830-kw. capacity per phase, 
each being in a separate compartment. The total transformer capacity 
at the generating station is 19 500 kw. The transmission system is 
on double-circuit, steel towers, spaced 500 ft. apart, with protruding 
pieces of angle iron on each upper corner of each tower which act 
as lightning arresters. It is understood that this protection is very 
satisfactory, there having been only one shut-down during the first 
year's service (which commenced in 1906). The capacity of the 
terminal station in Winnipeg is equal to that of the transforming 
station at the generating works. 

The utilities served consist of the city and suburban street rail- 



170 HYDRO-ELECTRIC POWER IN CANADA 

ways, the city house-lighting, and current for the present municipally- 
operated street lighting. In addition to this, there is a small distribu- 
tion of power for miscellaneous purposes, and two large blocks of 
power for flour milling. The capacity of the generating station is 
already fully taxed during the low-water winter season, and no effort 
is made to encourage large users of power by offering attractive rates; 
this power-station cannot be looked upon as a serious competitor to 
the municipal power-station now being built. 

The three tributaries of the Nelson River, which have just been 
discussed hydraulically, converge in Lake Winnipeg and there form 
an enormous reservoir of 9 300 sq. miles at an elevation of 710 ft. 
above the Hudson Sea. 

When it is considered that the drainage area feeding the outlet 
of Lake Winnipeg into the Nelson River proper is nearly 350 000 sq. 
miles, and that the water supply comes from three different and dis- 
similar climatic and topographical river basins, one can appreciate 
what a large minimum flow there is; and, as the route of the river 
to the Hudson Sea is rock-ribbed, in cascades, pools, and waterfalls, 
all that need be said here is that, should electro-chemical or other 
demands require very extensive water-powers, where relation to habita- 
tion is immaterial, these, to the extent of at least 4 500 000 h.p., can 
be readily secured on the Nelson River between Lake Winnipeg and 
the Hudson Sea. 

Other Hudson Sea Rivers. 

Of the many extensive rivers flowing into the Hudson Sea and 
James Bay, there is very little accurate information, and hydraulically 
these rivers are almost unknown except in a general way. 

It can be appreciated, however, that, from the physical conforma- 
tion, many enormous water-powers exist, because the table-land sources 
of the rivers abound in lakes at an elevation of from 600 to 1 500 ft. 
above the sea, and this, taken with a good rainfall and the rocky 
nature of the country, warrants this certain assumption. It is gen- 
erally known, however, that near the sea coast most of these rivers are 
alluvial, and that the water-powers of value lie considerably inland, 
in general, at least 200 miles from the mouths. 

The first development, north of the height of land which lies 
between the St. Lawrence River and the Hudson Sea, will probably be 
at Iroquois Falls, on the Abitibi River. The construction of the 



HYDRO-ELECTRIC POWER IN CANADA 171 

Ontario Government Kailway, parallel to the river, and that of the 
Dominion Government Transcontinental Railway east and west im- 
mediately north of Lake Abitibi, suggest the early development of 
Iroquois Falls for purposes of pulp grinding, or power for mines, or 
for general distribution to the junction city which will immediately 
grow up in that vicinity. 

The St. Lawrence Basin. 

The St. Lawrence River drainage basin is the chief seat of popu- 
lation in Canada, and naturally the hydro-electric developments of 
this region are most extensive. 

The total drainage area is 565 000 sq. miles, but of this about 
90 000 sq. miles comprise the water surface of the Great Lakes, which 
form enormous compensating reservoirs, their regulation and control, 
international in character, presenting a very delicate physical and 
political question, additionally complicated by the rights and needs 
of navigation; therefore it is advisable to discuss the water-powers 
of the main river and of its many important branches separately. 

Sault Ste. Marie. — The first water-power location on the main 
river is at Sault Ste. Marie, where Lake Superior, with 31 000 sq. 
miles of water surface, empties into Lake Huron, falling approxi- 
mately 18 ft. and discharging about 60 000 cu. ft. per sec. at low 
water, thereby affording opportunity for about 90 000 e.h.p., after 
making allowances for various hydraulic losses. The present develop- 
ments have been somewhat international in character, and have been 
carried out chiefly by the Lake Superior Power Company, which, with 
allied companies, disposes of 6 500 h.p. of hydraulic and electric power 
on the Canadian side of the river (out of a designed total of 20 000 
h.p.) to rail mills, pulp mills, and for industrial and municipal pur- 
poses in the Canadian city. 

On the United States side of the river the same company con- 
structed a canal and power-house, under unfavorable physical con- 
ditions, designed to obtain 45 000 h.p. ; but machinery for only one- 
half of this, or less, has ever been installed, and the chief customer, 
outside of municipal and small industrial users, is a carbide company 
having works in the vicinity; a third company, on the ITnited 
States side, has a small hydro-electric development, but it is probable 
that no further concessions will be granted there, or, if granted, 



172 HYDEO-ELECTRIC POWER IN CANADA 

will be under the control of the International Deep Waterways Com- 
mission, so that the needs of navigation, both as regards the main- 
tenance of the level of Lake Superior and the water required in 
increased quantities for operation of the three locks at this point, 
may be fully conserved. 

It is worthy of more than passing comment that, although these 
canals and locks serve an enormous commerce, their existence has 
been of little local value, whereas the life of the two adjacent cities 
of Sault Ste. Marie depends chiefly on the creation and operation of 
hydro-electric power, and the two Governments interested will be 
well advised to use every effort to facilitate rather than hinder the 
power use of all the water passing from Lake Superior, which is not 
required for locking purposes. It is evident that control works at 
the head of the rapids can be constructed so as to compensate for 
any quantity of water which may be diverted and used for power 
purposes; and, geographically, as an assembly point for raw materials 
by water and rail, Sault Ste. Marie has many advantages. 

Niagara Falls. — Continuing seaward from Sault Ste. Marie, the 
next location which affords hydraulic opportunity is spoken of 
familiarly as Niagara Falls. 

Lake Erie flows into Lake Ontario by a river route which, from 
Chippewa to Queenston, is a series of rapids and the cataract, but 
the cataract of Niagara Falls is barely one-half of the total difference 
in elevation, which is 327 ft., and that the first great hydro-electric 
developments have been made there with effective heads of from 136 
to 175 ft. can only be looked upon, in the future, as an economic 
blunder, because there are potential developments which can be made 
at several locations in the United States and Canada using a net head 
as great as 300 ft., which, for as large an installation as 200 000 h.p. 
or greater, would not be as expensive as those now constructed, and 
the deduction is here offered that when industry makes the demand 
neither international understandings nor the makeshifts of opportunist 
politicians will prevent capital from carrying out the construction of 
these undertakings for the benefit of mankind. 

This question of utilizing the hydraulic power of the Niagara 
escarpment is, in reality, a world problem, not a local one, because 
there is there latent 5 000 000 continuous electrical horse-power, con- 
sidering only the low-water flow of 180 000 cu. ft. per sec. at an 



HYDRO-ELECTKIC TOWER IN CANADA 



173 




*^' 



<J^ 



Of a^^& 



174 HYDEO-ELECTRIC POWER IN CANADA 

effective head of 300 ft., and, estimating even so little as 2i lb. of 
coal per horse-power-liour, this would be equivalent to a saving of 
55 000 000 tons of coal per year, should this great force of Nature 
ever become so fully utilized ; and why should it not be used in increas- 
ing amount as required? There will still be available for scenic pur- 
poses all the surplus water, which will still pass over the falls, even 
after the minimum has all been used for power purposes. Consider 
the question otherwise; this location is unique, as it is on one of the 
greatest highways of commerce in the world, both water-borne and by 
rail, and it is in the midst of a fertile and populous region, which, 
for 150 miles in all directions, is becoming more thickly populated 
year by year. The steady growth in the electric load carried by the 
present stations leads to the inevitable conclusion that we are only 
in the infancy of the utilization of this great gift of the Creator. 

The level of Lake Erie can be readily controlled by compensating 
works at the head of the rapids, and might be designed so as to con- 
strict the natural channel at the same rate as the artificial channels 
were created, either near the river or by other routes from Lake 
Erie to Lake Ontario, and this arrangement appears to meet and 
answer the only serious criticism which has tended to array the ship- 
ping interests in opposition to further utilization of the water for 
power purposes; and, indeed, diminution of natural flow presents cer- 
tain positive advantages, as it will check that rapid erosion of the 
Horseshoe Falls which is now taking place. 

Recurring to the actual constructions which have been made since 
the first station of the Niagara Falls Hydraulic Power and Manufac- 
turing Company was completed, on the United States side of the river, 
in 1881, it may be briefly stated that there are six large generating 
stations, two on the United States side and four on the Canadian side 
of the river, engaged in producing electrical energy for commercial 
purposes, and, in addition to these, there are numerous small stations, 
one at Niagara Falls, Ont., and several along the route of the Welland 
Canal, which produce a total of probably 15 000 h.p. for miscellaneous 
electrical and hydraulic purposes. 

It is unnecessary to describe in detail the diverse constructions, 
distributions, etc., which have been described so often and so minutely 
by various engineers, but certain broad facts will be given by which 
the scope of present accomplishment may be gauged. 



HYDRO-ELECTRIC POWER IN CANADA 175 

On the United States side, the Niagara Falls Hydraulic Power and 
Manufacturing Company has the distinction of being first in the field, 
and, from diminutive beginnings in 1872, it has expanded its works 
until at the present time it has 77 500 e.h.p. installed, and building- 
construction is under way for additional equipment having a capacity 
of 70 000 e.h.p. In addition to this, water is supplied to various 
industries developing 8 000 h.p. of direct hydraulic power for pulp, 
paper, flour, and other mills. 

The power from these works is used for street railway and various 
industrial and electro-chemical purposes, but chiefly for the manu- 
facture of aluminum, by The Aluminum Company of America, which 
company has here its largest works. It is understood that this com- 
pany contracts for nearly 60 000 h.p. 

The head canal for this development is 110 ft. wide and 14 ft. deep 
at normal water level. It is constructed through the City of Niagara 
Falls, N. Y., for a distance of 4 400 ft., and terminates at a f orebay 
immediately overlooking the river. The water is used under a total 
head of 210 ft. and a net head of about 200 ft., and is led down the 
side of the cliff in steel penstocks to the power-stations at the lower 
river level, where is installed a heterogeneous collection of direct- 
current and alternating-current generating units composed of various 
sizes from 150 to 1 000 kw. in the older station and from 3 500 to 
6 500 kw. in the newer station, which facts well illustrate the growth 
of the art and the magnitude of the business of this station, which is 
entirely local, involving no transmission lines of any importance. 

The Niagara Falls Power Company has two generating stations 
operating under an average effective head of 140 ft., one on each side 
of a short head-canal, and the common tail-race is a brick-lined tunnel 
passing to the lower river, 7000 ft. distant. 

These stations, heralded at the time as pioneers, have long since 
been eclipsed by the stations on the Canadian side of the river, and 
also suffer from certain engineering defects which, no doubt, were 
unforeseen at the time of construction and have since been cumulative. 

The ice difficulties, after years of strenuous effort, have been 
largely overcome, by improvements in the intake works and by the 
addition of an ice shoot to the tail-race tunnel, but the first station, 
having a capacity of 50 000 h.p., has unprotected ice racks and is 
equipped with free, outward-discharge, double-runner, vertical tur- 



176 HYDEO-ELECTRIC POWER IN CANADA 

bines. The ten generators are each 5 000-h.p., 2-phase, 2 200-volt, with 
revolving armatures. The second station, having a capacity of 
60 000 h.p., has additional submerged-arch entrances and an enclosed 
forebay for further protection from ice. The turbines are of the 
single-runner, vertical type, with draft-tubes; the eleven generators 
are each 5 500-h.p., 3-phase, 2 200-volt, and are partly of revolving-field 
and partly of revolving-armature types. The tail-race tunnels are 
led from the power-stations, by very sharp curves, to a junction, and 
this, together with the back-piling in Power-house No. 1 due to the 
use of free-discharge wheels, makes it impracticable to run both 
stations to full capacity. It is understood that 83 000 h.p. is about all 
that is generated, although the market in Buffalo is being largely 
abandoned to the Canadian Niagara Power Company, its station being 
several miles nearer and belonging virtually to the same financial 
interests. 

The power from these stations is now distributed in large blocks, 
chiefly to local electro-chemical consumers, and the policy of the 
company in creating a manufacturing district, with railway switching 
facilities, building sites, etc., and giving long-term power contracts, 
with guaranties of continuity, under penalty, has created a large and 
stable industrial center consuming more than 60 000 h.p. continuously 
from those two stations. The power transmitted to Buffalo, Tonawanda, 
etc., is 22 000-volt, 3-phase, 25-cycle. The pole lines are of ordinary 
wooden type and carry four 3-phase circuits, two on each line, 
some circuits being of copper and some of aluminum. Until the load 
was transferred to the Canadian station, all the varied electrical 
utilities of Buffalo, to the extent of about 40 000 h.p., were operated 
from these United States stations, as are still the various interurban 
and scenic electric railways converging at Niagara Falls. 

On the Canadian side of the Niagara River four large generating 
stations have been constructed recently; three of these are at Niagara 
Falls, and the fourth is at a point 12 miles westward. 

The Dominion Power and Transmission Company (formerly the 
Hamilton Cataract Power Company) first commenced to deliver power 
to Hamilton, Ont., in 1897, from a small station 12 miles westward 
of Niagara Falls at the foot of the Niagara escarpment, and drawing 
its water supply from the Welland Canal and thus from Lake Erie. 

This hydro-electric development utilizes a hydraulic head of 267 ft.. 



HYDEO-ELEGTRIC POWER IN CANADA 177 

and is a very economical one, as excessive expenditures for head-works 
were not required. This enterprise has a progressive management, and 
a growing market for power has expanded it until, at the present 
time, the machinery equipment, including transformers, has a nominal 
rating of 38 400 h.p. 

It is an interesting feature of this generating station, that the 
turbines are entirely of European manufacture and design, being 
supplied by Italian and German makers, as the earlier experiences 
with manufacturers in the United States were not satisfactory. 

The electric current is generated at 2 400 volts, 66§ cycles, and is 
transmitted to the surrounding towns and cities at various voltages. 
The chief market is in Hamilton, 32 miles westward, to which three 
pole-line, 3-phase circuits, two of copper and one of aluminum, carry 
power at 40 000 volts. In addition to this, however, the company 
serves a network of street and interurban railways which it owns or 
controls, and supplies power and light in Brantford, Welland, St. 
Catharines, Dundas, etc., with a total of 132 miles of transmission line. 

This company, owing to its favorable investment and strongly 
entrenched position, is able to distribute with profit at low rates for 
power, and has thus built up a large business very rapidly. It is under- 
stood that the peak load in Hamilton alone, which is a city of only 
60 000, is considerably more than 20 000 e.h.p., which includes current 
for operating street and interurban railways owned by this company, 
and also cotton mills, knitting mills, flour mills, rolling mills, and 
machine shops; in fact, every phase of a busy manufacturing city is 
provided for. 

This company also owns small hydro-electric generating stations at 
St. Catharines and Brantford, and may be said to control, almost 
exclusively, the business of supplying electricity in the Niagara Dis- 
trict west of Niagara Falls and Thorold. 

The generating station of the Canadian Niagara Power Company was 
constructed between 1901 and 1905. This company is practically 
identical with that of the Niagara Falls Power Company, and, 
although it distributes some 2 000 to 3 000 h.p. locally to various 
industries in Niagara Falls, Ont., it has not entered into the long- 
distance transmission field, no doubt considering the local sale to be 
ultimately more profitable. The generating station is of the wheel-pit 
type, with a tail-race tunnel 2 200 ft. long, and operates under an 



178 HYDRO-ELECTRIC POWER IN CANADA 

effective head of only 140 ft. The water is taken from the rapids 
immediately above the cataract, and although protecting works, sub- 
merged-arch entrances, enclosed forebay room, and sluiceways have 
been constructed to meet the difficult ice conditions, they are only 
moderately satisfactory, and a continuous winter service is maintained 
only by strenuous efforts and with considerable difficulty. 

The generating works are complete for eleven units, each having a 
nominal rating of 10 000 e.h.p., and five of these machinery units 
are now installed. 

The turbines, of Swiss design, are of the vertical type, with draft- 
tubes and double runners, and the generators are 12 000-volt, 3-phase, 
25-cycle. 

This station is inter-connected with those of the Niagara Falls 
Power Company by a large number of 11 000-volt underground cables, 
thus warranting great certainty of service; but the chief market for 
power is in Buffalo. 

The company has a transformer station, adjoining the generating 
station, with an equipment of fifteen single-phase, 1 250-kw. trans- 
formers, all mounted in one room; and although these are arranged so 
as to be operated, if desired, at various voltages up to 60 000, the 
pressure used at present is 22 000. 

The transmission system consists of steel poles, carrying two 
3-phase circuits of aluminum cable mounted on cellulose insulators, 
and a third circuit is now being built. The capacity is 12 500 h.p. per 
circuit at 22 000 volts. At Fort Erie this transmission system has two 
spans of 1 600 and 2 200 ft., respectively, suspended 130 ft. above the 
water of the Niagara River, and it is understood that considerable 
difficulty was experienced in cable breakage until elastic counter- 
weighted suspensions were made at the end towers with suspended 
insulation at the center tower. 

Under the franchise conditions existing between this company and 
the Provincial Government, and. having in view the water limitations 
now being reported on and recommended by the International Deep 
Waterways Commission, it would appear to be inevitable that this 
company will soon be forced to enter actively into exploitation of the 
Canadian market, in order to enable it to increase its output sales 
beyond the present, which not only amounts to about 40 000 h.p. from 
its own station, but to this must be added some 5 000 h.p. which is 



IIYDltO-KLKCTKlC POWER IN CANADA 179 

being obtained from the Electrical Development Company and 
exported. 

It is interesting to note that the combined output of the three 
stations of these allied companies, i. e., The Niagara Falls Power 
Company and the Canadian Niagara Power Company, is more than 
500 000 000 kw-hr. per year, and that the load factor is more than 70%, 
owing to the large number of electro-chemical customers. 

Between 1902 and 1906, the Ontario Power Company constructed, 
immediately below the cataract, a station which is equipped at present 
with six horizontal units, each of 10 000 or 12 000 e.h.p., generating 
3-phase, 25-cycle, 12 000-volt current, at 1874 rev. per min., under a net 
head of 178 ft. The turbines are of German design and manufacture. The 
water is brought to the station by an 18-ft. steel pipe, 6 500 ft. long, 
which is embedded in concrete just below the hydraulic grade line, and 
buried in Queen Victoria Park a few feet beneath the surface. 

. This company selected a favorable location for its entrance works 
and forebay, which are located at the head of the rapids and are 
completed for a capacity of 180 000 h.p., and, as far as ice troubles are 
concerned, it may confidently expect to give a continuous service. 

In several other features these works are interesting, particularly in 
respect to the methods of control. To handle successfully an 18-ft. 
column of water, 6 500 ft. long, moving at 15 ft. per sec, under ordi- 
nary commercial electrical conditions, was no mean task, but it is 
understood that, by means of a spiral overflow structure, this has been 
accomplished. The electrical control is centered in the transforming 
and distributing station, which is 600 ft. from the generating station 
and on the high blufi overlooking the cataract. At this point all the 
current from the generators to the high-t-ension lines is controlled. 
The transforming station is equipped with eighteen single-phase, oil- 
insulated, water-cooled transformers, each of 3 000 kva., each bank 
being in a separate compartment. This company has created several 
allied companies which distribute 62 500-volt and 12 000-volt current 
in the United States and Canada. Its extreme eastern delivery is at 
Syracuse, N. Y., 161 miles distant, and it distributes current westward 
to St. Catharines, Welland, and Port Colborne; but in view of the 
contract it has now entered into with the Hydro-Electric Power Com- 
mission of Ontario for the supply of from 25 000 to 100 000 h.p. at 
from $9.00 to $9.40 per h.p. per year for 12 000-volt continuous power, 



180 HYDKO-ELECTEIC POWER IN CANADA 

it appears evident that its sphere of influence in Southwestern Ontario 
is in the near future to become very considerable, and the output of 
its station increased to at least 50 000 h,p. 

At present its maximum load is 25 000 h.p. in the United States 
and 5 000 h.p. in Canada, sales being at the rate of 110 000 000 kw-hr. 
per year. 

This company's transmission lines are very extensive and varied 
in character, and comprise several hundred miles of double-circuit and 
single-circuit, aluminum lines, chiefly mounted on steel towers carrying 
one circuit each. Some of the branch lines and 12 000-volt lines, 
however, are on timber, A-frame towers or cedar poles. The various 
constructions have been carried out substantially and scientifically, 
but are subject to the criticism of transmitting at only 60 000 volts 
for distances up to 160 miles, whereas, at the present time, 80 000-volt 
transmissions are in use and 100 000-volt projects are under construc- 
tion; however, it is true that the suspended insulator has not yet been 
fully developed to unquestioned satisfaction, and, at 100 000 volts, air 
emanations from conductors, line troubles, etc., are certain to be 
considerable. 

In 1903 the Toronto and Niagara Power Company was formed by 
parties interested in the Toronto Street Railway Company and the 
Toronto Electric Light Company, and, after negotiations to buy 
electric current from other Niagara companies had failed, a lease was 
secured from the Provincial Government to build a generating station 
in Queen Victoria Park at Niagara Palls at a site lying between those 
of the Canadian Niagara Power Company and the Ontario Power 
Company, the works of both these companies being then under con- 
struction. The location chosen was notable in two particulars, the 
intake being almost wholly in the surging rapids above the cataract, 
while the tail-race tunnel was located so as to emerge beneath the 
cataract itself. The construction of these works proceeded from 1903 
to 1907. 

Preliminary to construction, it was most difiicult to construct the 
coffer-dams and keep them water-tight, and, together with the exten- 
sive overflow dam, etc., afterward constructed, made a heavy invest- 
ment necessary. The wheel-pit is constructed for eleven units, each 
of 8 000 kw., and four of these units are now installed. The 
operating head is about 140 ft. and the station is equipped with 



HYDRO-ELECTKIC TOWER IN CANADA 181 

Francis double-runner turbines, constructed by United States manu- 
facturers on Swiss models. In various items, however, these turbines 
differ from previous constructions. The draft-tubes are carried out- 
ward to twin, parallel, tail-race tunnels, and emerge from underneath 
so as to maintain the water seal. The penstocks are alternately right 
and left, in groups of two, thus economizing in wheel-pit length. The 
supporting structures for the vertical main shafts of the units are 
concrete arches instead of steel girders. Altogether, these works give 
evidence of advance in the art of wheel-pit construction. The power- 
house is a most ornate structure; indeed, this company has carried 
on all its operations with a lavish hand, and a resulting large invest- 
ment. Its generating company is known as the Electrical Develop- 
ment Company, and the 3-phase, 25-cycle, 12 000-volt current generated 
is delivered outside of Queen Victoria Park to a transformer station 
whence the Toronto and Niagara Power Company transmits it at 
60 000 volts to Toronto, 80 miles distant (population 300 000), over a 
double-circuit steel-tower line. This line is carried across Hamilton 
Beach and Canal and the Welland Canal by pairs of very high and 
heavy towers; otherwise the transmission line passes through a level 
cultivated country without serious transmission difficulties, although 
including various spans across small valleys of more than 1 000 ft. 
each. 

The transforming stations are built apparently with a great effort 
to attain safety, each single-phase transformer being in a separate 
compartment, and there is very extensive concrete cell construction for 
bus-bars, lightning-arrester compartments, etc. The Toronto terminal 
station also has an auxiliary water-tower for supplying water to trans- 
formers in addition to the water service from the city mains. The 
current delivered at this station at 55 000 volts is stepped down to 
11 000 volts and is then carried by underground feeders to various sub- 
stations in the city where it is transformed into 600-volt, D. C, and 
various voltages, A. C; also, the frequency is changed from 25 to 
60 cycles. Altogether, from 20 000 to 22 000 e.h.p. are used by the two 
subsidiary companies, in addition to sales at Niagara Falls to the 
Canadian Niagara Power Company for export, as before mentioned, 
and there are also subsidiary companies formed to distribute eastward 
in the United States. 

This brief description of the capacities of the various Niagara 



182 HYDRO-ELECTEIO POWER IN CANADA 

generating stations — amounting to 450 000 e.h.p. in Canada, when the 
stations described are fully equipped — and the fact that several other 
companies own charters and water rights, and are anxious to enter 
the field if given opportunity and encouragement, at first, would sug- 
gest that electric energy can now be obtained in this favored district 
at very low rates, and, comparing the rates offered with the cost to 
any given consumer to produce power or light by steam-driven prime 
movers, certainly the rates do improve these conditions somewhat, as 
is shown by the steady growth of electrically-operated industries in 
Hamilton, Ont. Moreover, from $20 000 000 to $25 000 000 have already 
been invested by these four companies in generating stations and 
transmission systems which must find a return by the sale of the 
electrical output, but, for the last ten years, the people of Southwestern 
Ontario have steadily asserted that, to create a great industrial area 
in that part of the province, still cheaper power ought to be and could 
be obtained. 

Boards of trade began to agitate in 1900, and in 1902 a committee, 
from various cities and towns, was formed to take concrete action, 
which, by pressure on the Provincial Government, induced the passage 
of an Act in 1903, enabling two or more municipalities to appoint a 
Commission and, after investigation, to carry out, if desired, the con- 
struction of the necessary works at Niagara or elsewhere for generat- 
ing, transmitting, and distributing electric energy. 

This municipal commission was appointed, made a very exhaustive 
investigation, and reported in 1906 to the seven municipalities then 
represented in the movement, but, owing to the limitations evident 
in the working of the Act, and to a larger movement having com- 
menced, there has been no specific outcome from this report. 

In 1905 the Provincial Government, becoming more fully alive to 
the importance of this question, appointed a Government Commission, 
which was instructed to study and report on the hydraulic possibilities 
in the Province, the industrial demand then existing for electric 
energy, and the prime cost of generating and delivering the same to 
the various industrial centers and customers. The reports arising 
from the investigations were presented, in due course, to the Legisla- 
ture during the session of 1905-06, and, although the facts elicited 
furnished a comprehensive idea of the industrial condition then 
existing, and the cost of satisfying the same, and also, by comparison. 



HYDRO-ELECTRIC POWER IN CANADA 183 

set forth the rates offered by distributing companies, it has only been 
after a subsequent protracted struggle that definite results have been 
obtained and the evident will of the people carried out. 

The vested interests immediately attacked the reports and the 
bona fides of those who drew them up, but the Provincial Government, 
recognizing the demand made upon it, passed a very comprehensive 
Act, appointing a permanent Government Commission. In this com- 
mission was vested the rights of the Crown, and it was authorized to 
act in response to the demands of the municipalities and in their 
behalf. 

This Commission, after 2i years' existence, has justified itself; and 
a group of sixteen municipalities has made a demand upon the Com- 
mission to secure electric power, to construct transmission lines from 
Niagara, to transform the electric energy thus transmitted at the 
various sub-stations, to administer the construction, and to operate 
the works for a period of 30 years, over which period a sinking fund 
will be created out of the rates to repay the investments. The amount 
of power to be distributed immediately is approximately 30 000 e.h.p., 
but it is confidently expected that this will increase very rapidly to 
probably 100 000 e.h.p. in 5 years. 

The various municipalities have voted the money necessary for 
local distribution, and the Government has recently contracted for the 
construction of 300 miles of transmission lines, and this work is to be 
followed immediately by the construction of the necessary sub-stations 
and local distributions. 

This popular movement has yet to be justified by capable and 
honest administration, deficiency in which is the weakness of many 
municipal movements in America, and it is to be hoped that the aegis 
of the Government Commission will render the inherent merits of the 
project effective. 

The large area of country to be covered and the necessary anticipa- 
tion of a still wider area as soon as the first installation has justified 
itself, have forced on the Government engineering staff a considera- 
tion of higher voltage, and, according to present plans, the ultimate 
voltage of transmission now being provided for is 100 000. This 
advanced practice, however, has been adopted by reason of the neces- 
sary provision for further extensions of the Hydro-Electric Commis- 
sion's system to points still more distant from the generating station 



184 HYDllO-ELECTllIC POWER IN CANADA 

than London and St. Thomas (140 miles). The world will, doubtless, 
watch closely the results obtained in carrying out this huge municipal- 
provincial distribution of electric energy. 

St. Lawrence Biver.- — The St. Lawrence River, between Lake 
Ontario and Montreal, may be considered as having its navigation 
interests paramount, and, although attempts have recently been made 
to obtain permission from the Canadian Government to devote the 
whole river to power development, it is not likely that such permission 
will ever be granted. 

In the meantime, along the ship-canal system, many small installa- 
tions have been permitted, but they are chiefly local in chai'dcter, and 
are devoted to operating mills and factories. They amount roughly 
to 10 000 h.p., and, in addition to this, hydro-electric stations having 
a capacity of about 5 000 h.p. supply the various near-by towns with 
light and power, and light the ship canals during the navigation 
season. 

The whole river, if available, would be a great source of power, 
as it is capable of developing, at low water, 1 500 000 h.p. of electric 
energy at the various rapids. 

Montreal.- — In the vicinity of Montreal (population 400 000), 
several interesting constructions have been made at various times by 
various companies, but finally all these interests have been merged into 
the Montreal Light, Heat and Power Company, which controls the 
situation at present, although a rival company has recently been 
organized to develop power on the abandoned Beauharnois Canal of 
the St. Lawrence system and transmit to Montreal and distribute in 
opposition to the Montreal Light, Heat and Power Company. The 
latest construction by the latter company is at Soulanges, on the 
Soulanges Canal of the St. Lawrence River, 27i miles west of 
Montreal, where 40 000 cu. ft. per sec. are available under a head of 
54 ft. The equipment now being installed consists of three units, 
each of 3 750 kw., which will generate 3-phase, 60-cycle current at 
4 000 volts. The transmission will be on one circuit at 44 000 to 
40 000 volts, and the transformation will be made by banks of single- 
phase, water-cooled transformers having a rated capacity of 2 500 kw. 
per phase. 

At present, this company owns two water-power stations and four 
steam stations, the latter having a capacity of 9 000 h.p. The com- 



HYDRO-ELECTRIC POWER IN CANADA 



185 



pany also purchases 16 000 h.p. delivered in Montreal by the Sha- 
wenegan Water and Power Company. 

The station at Lachine, on the St. Lawrence River, is an example 
of daring but mistaken construction, carried out 12 years ago, when 
the transmission problem had not been as fully worked out as at 



CANADIAN POWER TRANSMISSION SYSTEMS 

' MONTREAL AND QUEBEC DISTRICT 

PROVINCE OF QUEBEC 

SCALE OF MILES 



10 



10 20 30 40 50 00 iO 
& rowei-Stntions 




.. f-' c 

S^' UNITED STATES Ol? 
Fig. 5. 



present. The works are really constructed as a part of the river, 
and therefore, in operation, they are subject to troubles from frazil 
in the head- and tail-races; but, after excessive expenditures, made for 
the purpose of creating quiet conditions above the power-house and 
protecting the tail-race from back-water, the station can be reasonably 
depended on to give continuous winter service. 



186 HYDRO-ELECTEIC POWER IN CANADA 

The dependable hydraulic head is only 14 ft., and the eight genera- 
tors, each having a rated capacity of 750 kw., are operated by groups 
of six turbines each, driving jack-shafts. 

The current generated is 3-phase, 60-cycle, at 4 400 volts, and is 
transmitted over eleven circuits at this voltage direct to Montreal (6^ 
miles), where it is stepped dovpn to 2 400 volts for distribution. The 
station v^as designed for a capacity of 16 000 h.p., but it has never 
been equipped with machinery beyond 8 000 h.p. 

At Chambly, on the Richelieu River, a branch of the St. Lawrence, 
18 miles southeast of Montreal, the company owns a generating station 
which has also had a history interesting to engineers. The concrete 
dam, constructed to create a head of 34 ft., failed soon after comple- 
tion, and the first year's operation was much disturbed by frazil, but 
a new dam has since replaced the faulty structure, and the station now 
operates normally. 

The equipment consists of eight units, each of 1 800 kw., driven 
by four turbines per unit, the total generating capacity of the station 
being 20 000 e.h.p. ; but the average load carried is 15 000 e.h.p. The 
current, which is 2 200-volt, 2-phase, 66-cycle, is stepped up to 25 000 
volts, 3-phase, by 2 750 kw., single-phase, air-blast transformers, and is 
delivered over four circuits to Montreal, where it is transformed into 
the city distribution systems. 

The cojnplexity of distribution at Montreal may be appreciated 
when it is noted that the current is of three different frequencies, 
namely, at 30 cycles from Shawenegan, at 60 from Lachine, and at 66 
from Chambly, to which must be added the ordinary conversions to 
D. C, and to various distributing voltages, A. C. 

The total distribution of power in the city from the various stations 
of the company varies from, say, 30 000 h.p. in the summer to as much 
as 53 000 h.p. at peak load in the winter, and this amount will be sup- 
plemented soon by delivery from the Soulanges station now under 
construction. 

St. Lawrence River Branches. — The St. Lawrence River Basin is 
traversed by many large tributaries which will be discussed briefly in 
respect to such developments as have been carried out, but it may be 
stated briefly that, although many small towns and mining and milling 
centers have hydraulic and electric stations of modest proportions 
satisfying present needs, there yet remains opportunity for developing 



IIYDUO-ELECTRIC POWER IN CANADA 187 

millions of horse-power on such rivers as the Ottawa and its tribu- 
taries, the St. Maurice, the Saugenay, the Trent, Spanish, Nepigon, 
etc., and projects innumerable of greater or less merit are being studied 
carefully by engineers and financiers with a view to meeting industrial 
needs as they arise, and creating new centers and industries to con- 
sume the available power. 

The Kaministiqua Power Company has constructed a generating 
station at Kakabeka Falls, on the Kaministiqua River, and transmits 
power to the City of Fort William (population 15 000) at the head of 
Lake Superior, where grain elevators, flour mills, foundries, and 
municipal utilities consume some 6 000 or 7 000 h.p. at the present time. 

The development has a hydraulic head of 175 ft., and carries the 
water through two 10-ft. concrete tubes a distance of 6 500 ft. to a 
compensating reservoir, and thence by steel pipes to the power-house. 
The equipment consists of two 4 000 kw., 3-phase, 60-cycIe, 4 000-volt 
units, driven by German-built turbines, and the current thus generated 
is raised to 25 000 volts by two banks of single-phase, air-cooled trans- 
formers of 1 500 kw. per phase ; it is transmitted 19 miles over two 
circuits of No. 00 copper on wooden poles to Fort William. 

In 1904 the Canadian Copper Company (otherwise the Interna- 
tional Nickel Company) constructed a generating station on the 
Spanish River at Turbine, from which current is delivered to its mines 
and smelters at Copper Cliff near Sudbury, which lies north of 
Georgian Bay. 

The construction of a large head reservoir created good operating 
conditions, and increased the hydraulic head to 85 ft. Two 2 000-kw. 
units are installed and a third is now under construction. The current 
as generated is 3-phase, 25-cycle, at 2 400 volts, and is stepped up to 
35 000 volts for transmission. It is carried 29 miles to Copper Cliff, 
by two 3-phase circuits of No. 1 copper mounted on a double-pole line. 

The company has various sub-stations where the current is dis- 
tributed for service in the mines and at the smelters, which it is under- 
stood turn out more than one-half of the world's consumption of 
nickel. 

The Mond Nickel Company is also installing a 1 500-h.p. hydro- 
electric station for similar purposes on the Vermilion River, and the 
current is to be transmitted 11 miles to the Victoria mines and 
smelters, 25 miles southwest of Sudbury. 



188 HYDRO-ELECTRIC POWER IN CANADA 

The Town of Orillia (population 6 000) has constructed a power- 
station on the Severn River, and transmits some 1 600 to 1 800 h.p., at 
22 000 volts, a distance of 20 miles. It distributes the current for 
industries as well as for pumping and lighting; and, by offering low 
power rates, has established a stable industrial center. 

The City of Ottawa (population 75 000) is in the midst of many 
developed and undeveloped water-powers, and the latter, in the event 
of the construction of the Georgian Bay Ship Canal, would become 
readily available at low investment costs. In addition to the Ottawa 
River itself, many of its larger branches, especially in the Province of 
Quebec, are very great latent sources of hydraulic power, and it may 
be stated conservatively that within a radius of 75 miles of Ottawa 
City there can be developed on the Ottawa River and its main tribu- 
taries 500 000 h.p. at low water, which amount would be increased 
materially by the construction of the ship canal. 

Near or in the City of Ottawa there are ten hydraulic and hydro- 
electric developments operating under various heads, but chiefly on the 
Chaudiere Falls at from 22 to 35 ft. These developments form the 
industrial center of the city, and operate very extensive pulp and 
paper mills on both sides of the river. There are also cement mills, 
carbide furnaces, and other industries, and power for street railways, 
lighting, and general city use is also produced. About 35 000 h.p. is 
now developed, and, as many conflicting interests were contending 
for their respective shares of the flow of the river, it became evident 
that some improvement was imperative, so that an agreement to con- 
struct a proper dam at Chaudiere Falls has been entered into and the 
work is in progress. 

This dam will increase the hydraulic head somewhat, and will 
drown out the rapids immediately adjoining, thereby lessening the 
very serious present difiiculties in respect to operating in the winter 
season, when the water is full of frazil, and chiefly to conserve the last 
drop of water in low-water periods, thereby increasing the available 
power at such seasons. However, in the near future, more power will 
be required, and it is proposed that the city itself develop the splendid 
powers quite near it on the Gatineau River, where a head of more than 
80 ft. can be secured at a point within 8 miles of the city, and where 
50 000 h.p. can be developed at one location. Considering the vast 
timber and mineral resources tributary to Ottawa, and its position as 



HYDRO-ELECTRIC POWER IN CANADA 189 

Capital of the Dominion, it may be considered certain that the avail- 
ability of unlimited hydro-electric power assures the industrial future 
of the city. 

Quebec. — The Ottawa is only one of many great rivers flowing into 
the St. Lawrence from the north between Montreal and the sea, and, 
as all these rivers rise in unbroken forests on the Laurentian high- 
lands, and are all well supplied with splendid storage facilities, they 
are all valuable sources of power. However, the sparse population has 
been a deterrent, and the only large works north of the St. Lawrence, 
and devoted to the generation of electric energy for transmission pur- 
poses, is on the St. Maurice Eiver at Shawenegan Falls, where the 
Shawenegan Water and Power Company has created a large industrial 
town and also generates a large amount of power for transmission to 
Montreal, Sorel, Three Eivers, Thetford, etc., and will doubtless soon 
enter the field at Quebec in opposition to the present companies. 

The hydraulic conditions at this place are most favorable, as the 
head-works consist of a natural lake, which prevents trouble from 
frazil in the winter. The drainage area is some 18 000 sq. miles, and 
a flow of 10 000 cu. ft. per sec. is considered the minimum. This, at 
a hydraulic head of 130 ft., provides more than 100 000 e.h.p. at all 
seasons. The works have been carried on partly for the generation of 
electric energy and partly to operate local industries hydraulically. 
The products of these industries are pulp, paper, aluminum, carbide, etc. 

The electric station has three 4 000-kw. and two 7 000-kw. units, 
generating 2-phase, 30-cycle current at 2 200 volts, and this is trans- 
formed to 50 000 volts by banks of 1 100-kw. and 2 200-kw. single- 
phase, water-cooled transformers having a total capacity of 32 000 kw. 

The transmission circuits are of aluminum cable on wooden poles. 
The Montreal service (85 miles) has been in operation for the past 10 
or 12 years, and has given good satisfaction. 

The delivery in Montreal is by two circuits, and amounts to 
16 000 h.p.; besides which, the company serves the asbestos and other 
mineral districts on the south side of the St. Lawrence at Danville, 
Thetford, etc. The accomplishments of this company within the last 
ten years have been considerable; indeed, its works form the basis of 
many large industries, and it can be confidently expected that the 
whole 100 000 h.p. available at this one water-power will be made use 
of in the near future. 



190 HYDRO-ELECTRIC POWER IN CANADA 

While Shawenegan is an instance of the water-power being devel- 
oped as a foundation for industries and to supply distant markets, 
the developments in the vicinity of Quebec City (population 90 000) 
have been made strictly to supply existing needs. The installation 
on the Chaudicre Eiver, south of the St. Lawrence, has a hydraulic 
head of 114 ft., and is laid out for 6 000 h.p., but the equipment, thus 
far, is for not more than one-half this amount. The current is 3-phase, 
66-cycle, and is generated at 11 000 volts. It is transmitted to Levis 
and also to Quebec City by insulated cable laid on the bed of the river. 

Two developments north of the St. Lawrence, on the Montmorency 
River and on the Jacques Cartier River, supply current for city 
lighting, street railway operation, cotton mills, etc. 

The Montmorency works, having a capacity of 4 000 h.p., have five 
units operating under a head of 220 ft., and generate 2-phase current 
at 5 000 volts, which is transmitted direct over four circuits to the city 
sub-stations. 

The Jacques Cartier works have 2 000 h.p. installed, and the 
current is transmitted 18 miles at 20 000 volts over two 3-phase cir- 
cuits to the city limits, where the company has a steam auxiliary 
station of about 2 000 h.p. to carry peak loads. 

Some Conclusions. 

A careful consideration of the electrical data contained herein 
will indicate the trend of activity in Canada in hydro-electric invest- 
ments and the reasons therefor. In general, it may be said that on 
the Atlantic Coast and over the western and central portions of the 
great plains very little benefit is to be derived from such constructions, 
but over all the Provinces of New Brunswick, Quebec, Ontario, East- 
ern Manitoba, Western Alberta, and British Columbia, the continued 
construction of additional hydro-electric stations and the distribution 
of electric current to the practical limits of transmission may be 
confidently predicted. The resources of Canada, in minerals, timber, 
and agriculture, are so varied and so immense that the growth of 
industry as a legitimate outcome of the development of these resources, 
and the corresponding increase in population, may be looked on as a 
foregone conclusion, and it is as certain that, in tliis industrial growth, 
low-priced power will form a large factor. 

At present the hydro-electric developments serve to a greater or 



HYDRO-ELECTRIC TOWER IN CANADA 101 

less extent such industrial centers as Montreal, Quebec, Ottawa, Peter- 
boro, Toronto, Hamilton, Winnipeg, and Vancouver, besides a great 
number of smaller ones; but only a beginning has been made. Prob- 
ably the 500 000 h.p. now hydraulically developed would include only 
stations of 1 000 h.p. or more, and there are literally many millions of 
horse-power available for development within easy reach of the present 
population; but, aside from the increased industrial demand, there are 
specific demands of great importance. Already, as noted, several 
mining interests are served by hydro-electric power, and, with the 
growth of mining investments, the importance of such service will 
increase. Then again, pulp and paper (with the depletion of United 
States supplies) will become a much more important industry than at 
present, and consequently much hydraulic and electric power for 24- 
hour service will be thus consumed. 

Particular attention, however, is directed toward the enormous 
opportunities for electro-chemical industry. At Niagara Palls, N. Y., 
a large group of such industries is now well established, and in 
Canada a modest beginning has been made at Shawenegan, Ottawa, 
and Niagara Palls, Ont. ; but Canada has in its great rivers sources of 
power as cheap as any in the world, except that the labor item may 
increase slightly the investment per horse-power beyond that of a 
similar construction in Europe. The whole question rests on trans- 
portation and markets; the latter, artifically restricted by the United 
States tariff, must seek an outlet in Great Britain, and, as it is well 
demonstrated that $10 per horse-power-year, continuous power, at the 
low-tension switch-boards, is a remunerative price for large installa- 
tions, and will cover all fixed charges, depreciation, and attendance, 
there would appear to be a good field for investment. 

The financial condition in Canada at present, as regards hydro- 
electric enterprises, is worth considering. Some 10 or 15 years ago 
many of the brighter minds in the country seized the idea that, by 
securing valuable long-term franchises, on the one hand, for electrical 
service, and on the other for water privileges on near-by rivers, a 
splendid opportunity existed for profitable exploiting. 

The net result at present is that, in almost all the large towns and 
cities, these services are administered by companies whose charges for 
power and light are only regulated by two things : either the exigency 
of the occasion, or the cost of a similar service produced by the 



192 HYDRO-ELECTRIC POWER IN CANADA 

customer from coal usually imported from the United States; and as 
a good quality, run-of-mine, bituminous coal costs from $3.50 to $4.00 
per ton in Ontario and Quebec and $6.50 per ton in Manitoba, it is 
evident that such competition is harmless, and the electrical companies 
were and are able to do business at very profitable rates. 

In recent years, however, the municipalities have become active, 
and several smaller towns and cities have constructed municipal hydro- 
electric stations. The movement is very active at present in the 
Province of Ontario, and to a less extent in other provinces. 

The struggle is bound to be prolonged and bitter; indeed, much 
bad feeling and resort to disreputable tactics have already been 
exhibited. Municipal and provincial credit has been attacked in the 
London market, and all the influence which heavy vested interests can 
exert is and will be directed against further municipal action. 

The ultimate result is evident; the public mind will become more 
and more inflamed by the obstruction of the electrical companies, and, 
undoubtedly, in the end, the municipalities will carry out the ex- 
pressed wish of the majority; but in the meantime public credit is un- 
settled, the manufacturers are playing a waiting game, and no one 
is benefited because the expansion of the market, and consequent 
advantage to every one, including the electrical manufacturing com- 
panies, has not taken place. 

It would appear advisable to consider the appointment of a Public 
Service Commission by the Dominion Government to regulate the rates 
which may be charged by companies already in operation, so as to 
secure that benefit to the public which the presence of abundant water- 
powers warrants, because it is not probable that, at the most, more 
than a majority of the industrial centers will be served municipally, 
and, judging by the results obtained by the Dominion Railway Com- 
mission, much satisfaction might be obtained and the public mind be 
set at rest if a rate-regulating commission for electric utilities, with 
wide powers of action, should be appointed. 

Concurrently, however, with the distribution of electric power for 
utilization in cities, etc., it is to be anticipated that very large con- 
structions will be undertaken in the near future to furnish power 
for the operation of railways over heavy grades, where coal is ex- 
pensive or traffic dense, and that at the same time the natural re- 
sources of the country will develop to an increasing extent, and that 



HYDRO-ELECTRIC POWER IN CANADA 193 

in mining work, pulp and paper manufacture, electric smelting, and 
general electro-chemical service will be found the most attractive fields 
for the enormous and not fully appreciated sources of cheap and 
widely-distributed power which is now wasting in so many districts 
of Canada. 



194 DISCUSSION : hydro-electric power in CANADA 

DISCUSSION. 



Mr. Hoigate. H. HoLGATE, M. Am. Soc. C. E. (by letter). — This subject is one 
of the great features of present-day development and advancement, 
and is of growing importance and interest to the community. 

The well-knovpn phrase "Conservation of Natural Resources" can 
probably be applied to hydraulic power development along lines similar 
to those which apply to forests and their preservation or destruction. 
A forest may be wasted or burned, but the land, in most cases, will 
reproduce something of value. In developing hydraulic power, a per- 
manent condition of affairs may be created, so that the full poten- 
tiality of the river at the development site can never be realized, thus 
causing a permanent waste of power. Of course, there are commercial 
features which sometimes limit the capital outlay and prevent the 
building of works to produce maximum results, but, it often happens 
that permanent works are built for present requirements only, and are 
afterward extended or improved to meet the growth of business, gen- 
erally resulting in a greater outlay than would have been necessary 
originally, and also, even then, owing to the limits imposed by initial 
design, failing to be able to utilize all the power which might be avail- 
able, or even may be required. 

No water-power proposition should be undertaken before it has 
received the fullest possible study; and investors would be consulting 
their own interests if they secured sufficient information to establish 
the total physical possibilities, and the cost of the development. It is 
of first importance to establish the full potentiality of the stream, 
making use of all favorable physical features; then any modifications 
which business prudence may dictate may be made, but one should 
not lose sight of the fact that at a future time it may become neces- 
sary to expand and carry out an ideal form of development; therefore, 
the earlier works should be built to fit in with future work, as far as 
practicable, and thus avoid unnecessary expense and costly interrup- 
tion of business. 

It is not possible to say what is to be in the future, but lack of 
confidence in the future progress of Canada has stamped itself on 
many a hydraulic power development, and the time has come when 
all doubts and fears for the great expansion of trade in the Dominion 
may be dismissed; Canada may now be said to have emerged from the 
"hobbledehoy" stage of growth and to have entered upon good sturdy 
progresssive manhood, with growing confidence in the future. 

In granting water-power rights, it will be proper for the Govern- 
ment to insist that the initial works shall be built with due regard 
to the possible maximum development in the future, so that all possible 
power can be made available when conditions arise which may demand 
its use. 



DISCUSSION : HYDKO-ELECTKIC POWER IN CANADA 195 

Though Mr. Smith's paper does not contemplate going into details, Mr. Holgate. 
yet he has mentioned two plants which were designed and built by the 
writer and are used extensively for mining and smelting purposes. 
The first is the development at Bonnington Falls, on the Kootenay 
River, in British Columbia, and the secend is the development for the 
International Nickel Company at High Falls, on the Spanish River. 
The latter serves all the works of the International Nickel Company in 
what is known as the Sudbury District, in Ontario. 

As these two plants are radically different in design, and represent 
different methods of development, their cross-sections, Figs. 6 and 7, 
may prove interesting. 

The water-powers of Canada are distributed over almost the whole 
country, but the district east of Port Arthur is particularly favored, 
and, owing to the accessibility of the powers south of the height of 
land, those on rivers tributary to the St. Lawrance will naturally be 
developed before those on rivers flowing into Hudson Bay (errone- 
ously called Hudson Sea). 

In particular, attention is called to the possibilities of the St. 
Lawrence, and to the rivers tributary to it east of Quebec, as this 
territory, in so far as the north shore of the St. Lawrence is concerned, 
is not referred to in Mr. Smith's paper, nor has it received much 
attention from any one. 

In regard to the St. Lawrence River, the mean level of Lake 
Ontario is 245 and the mean level of the St. Lawrence at Montreal is 
18, thus creating a fall of about 227 ft. The minimum flow of the river 
is about 187 000 cu. ft. per sec, so that, on this basis, the minimum 
energy of the river in this stretch in about 100 miles is about 
3 500 000 h.p. 

The fall in the river is overcome by a canal system affording a 
depth of 14 ft., with locks 255 by 45 ft. The various canals are: 

Canal. Lockage. Length. 

Galops 17.81 ft. 7.60 miles. 

Morrisburg 14.81 " 3.70 " 

Farren's Point 10.78 " 0.75 " 

Cornwall 46.82 " 11.50 " 

Soulanges 84.13 " 14.00 " 

Lachine 47.62 " 8.50 " 

Totals 221.97 ft. 46.05 miles. 

This leaves stretches of river navigation aggregating about 54 
miles, so that nearly one-half of the navigation distance from Lake 
Ontario to Montreal is through the canals. 

The present depth of the canals will not carry the larger Lake 
boats, and the question of deepening the canals and enlarging the 



196 DISCUSSION : IIYDRO-ELECTRIO power in CANADA 

Mr. Hoigate. locks is being discussed. This would be a very heavy undertaking, and 
would result practically in rebuilding all the canal works, as the new 
depth should be at least 22 ft. 

An exhaustive investigation into the feasibility of damming the 
St. Lawrence at the foot of the Long Sault Rapids, a few miles west 
of Cornwall, has been carried on by J. W. Rickey, M. Am. Soc. C. E., 
during the past three years, and the writer has examined all the data 
gathered, and has no doubt of the possibility of carrying out the 
work. The result of building dams at the sites selected will be the total 
elimination of floods above the rapids, caused now by ice jams, and 
the improvement of flood conditions due to ice below the rapids, and, in 
addition to this, the present Cornwall Canal could be eliminated by 
placing a single lock at the dam. 

Ice conditions on the St. Lawrence are annually a source of trouble, 
and, of course, this would be prevented by treating the rapids in this 
manner. The writer is convinced that a careful study will show that 
it is possible to dam the river at the foot of each of the rapids, thus 
providing for navigation throughout the whole distance, and using 
the present canal system as an alternative route. In doing this, the 
necessary depth could be obtained, and navigation conditions would be 
decidedly improved by providing a deeper channel and by permitting 
vessels to move with increased speed, as they must now proceed through 
long stretches of canal with numerous small lockages, making their 
progress very slow, both up and down. 

Of course, the tourist will complain that he cannot "shoot the 
rapids," and that the natural beauty of the river is marred, but when it 
is considered that carrying out such a work will provide a vast amount 
of power, and also will improve a great international water highway, 
the argument of the lover of wild Nature can have little weight. 

In some cases the construction of such works would not be justified 
if the result were only the production of power, but coupled with this 
there is the possible improvement to navigation, and the two results 
will justify at least an exhaustive study of the river, with this end in 
view; particularly is this so when the increasing value of power is 
considered, and also the increasing impoi*tance of this water route. 

The relations between the Nations are happily such as to render 
it possible to bring about a solution of the problem, and if they co- 
operate, the St. Lawrence may be made a gigantic power producer, and 
a more important water highway than it is now. 

Attention is next drawn to the possibilities offered for the develop- 
ment of water-power east of Quebec, on the rivers flowing south into 
the St. Lawrence. These begin with the Saguenay and may be said to 
end with the Natisquahan River, although there are gigantic water- 
powers on the Hamilton River, which, at present, are commercially 
inaccessible. 



DISCUSSION : HYDRO-ELECTRIC POWER IN CANADA 197 



POWER DEVELOPMENT 

FOR 

HURONIAN COMPANY 
CROSS-SECTIONS THROUGH GENERATING STATION 

i ' ■ ■ ' SCALE 6f FEET ' 



Mr. Holgate. 




^|^^^^:f^f^^y^ 



Fi(i. 6. 



198 



DISCUSSION : HYDRO-ELECTRIC POWER IN CANADA 



Mr. Hoigate. Large developments are possible on the Saguenay, Bersemis, God- 
bout, Outarde, Manicouagan, Manitou, Eomaine, St. John, and Natis- 
quahan Rivers. A development of powder on St. Marguerite River, 
made by the writer, is used for the manufacture of wood pulp. 

An unique situation occurs about 200 miles east of Quebec, where 
the Outarde and Manicouagan Rivers enter the St. Lawrence about 
5 miles apart. There, within a radius of 5 miles, the writer has 
examined three possible developments, the combined horse-power of 
which would be 600 000. 




POWER DEVELOPMENT 

FOR 

WEST KOOTENAY POWER 

AND LIGHT CO. 

SECTION ON CENTER LINE OF 

No.2 MAIN UNIT 

SCALE IN FEET 



Fig. 7. 



All the water-powers north of the St. Lawrence are close to tide 
water, the back country for a long distance is covered with spruce, and 
iron exists in large bodies, so that it is only a matter of time when 
these powers will be commercially utilized. When smelting by electric- 
ity becomes a pronounced possibility, these water-powers will become 
particularly useful; at present they offer inducement principally to 
the paper-making industry. 

Winter conditions, of course, are severe in this district, but not so 
severe as to affect operation materially, and the larger the river, the 
less trouble will be experienced in the operation of the power. 

Plate XI, a map of this district, shows the location of the rivers, 




MAP OF PART OF THE 

PROVINCE OF QUEBEC 



DISCUSSION : IIYDEO-ELECTRIC TOWER IN CANADA 199 

and it should be noted that the distance from these points to Europe is Mr. Hoigate. 
much less than from any of the Atlantic ports. 

The water-power possibilities of Canada are widespread and very 
great, and under wise direction, with the rapid growth of the country, 
it is reasonable to expect a great advance in the utilization and 
development of hydraulic power. Though many sections of the 
country where important developments are in operation have not been 
mentioned, it will be seen that much has yet to be done before the 
resources of Nature are even partially made use of. 

Electro-chemistry, requiring large quantities of power at low cost, 
will find a great field in the water-powers of Canada, and will no 
doubt assist in the early advancement of power development. 

E. J. Beugler, M. Am. Soc. C. E. (by letter). — Mr. Smith's paper Mr. Beugier. 
forms a comprehensive account of the development of Canadian water- 
powers. 

This development appears to have followed the natural laws of 
supply and demand, and is in accordance with the general policy that 
water-power plants are constructed only when the power can be gene- 
rated profitably, and not before, unless the project is developed pre- 
maturely. The latter condition, unfortunately, is the case with several 
developments in the United States which, thus far, have failed to pay 
dividends. 

In districts where cheap coal or other fuel can be secured for the 
generation of power, the hydraulic plant, unless very favorably situated 
as regards cost of construction, cannot be developed profitably. Many 
water-power projects that could not now be operated to advantage will 
ultimately prove valuable. It may be news to some to know that the 
first cost of a complete hydro-electric development, and consequently the 
fixed charges or interest on this cost, is, as a rule, greater per horse 
power generated and delivered than the same elements in a steam plant. 

It is also of interest, in view of the general disposition toward 
preserving "water-power resources" in the United States, to consider 
the fallacy of certain ideas regarding conservation. This matter has 
been taken by many to mean that our water-powers, coal beds, and 
forests should be saved for future generations at the expense of 
present needs and economies. To the writer's mind this view is entirely 
wrong. Our resources are for our judicious use whenever demanded, 
and while wanton waste is also wrong, the policy of discouraging 
capital from investing in water-power developments by threats of oner- 
ous taxes, laws, and regulations simply results in a loss to the public 
from the delayed use of these resources, some of which, if not vised 
immediately, are irrevocably lost. 

The construction of water-power plants will not economize our 
supply of coal so long as the coal can be bought for a price which will 
produce a certain amount of power for less money than that for which 



200 DISCUSSION : hydro-electric power in CANADA 

Mr. Beugler. it can be produced in a water-power plant, nor would it be sound busi- 
ness policy to develop such plants under this condition. 

Water-power plants should not be subject to extraordinary govern- 
ment taxes. Taxing a water-power company for the use of water, 
where the company has already paid a good price for riparian rights 
and necessary lands, simply shifts the line of tax payment. The power 
company must add this item to its operating costs, and consequently 
must sell its power for a higher figure, so much higher, in many cases, 
that the water plant would be unable to compete with coal power; 
hence, any additional tax upon the development of water-powers may 
have the effect of preventing such development and cause a continued 
use of a failing supply of coal, an effect which is diametrically 
opposite to that desired by the devotees of conservation. 

Subsidy with governmental regulation, the latter as advocated by 
Mr. Smith, would appear to be more logical than additional taxation. 

The point the author makes regarding favorable topographical con- 
ditions is an important one, especially as to the beneficial effect of 
swamp and lake areas on minimum stream flow. There are anomalies, 
however, which give opposite results from what might be expected and 
cannot be explained satisfactorily by the difference in the rainfall and 
weather conditions. For example, a comparison of the minimum run- 
off (as recorded by the Ihiited States Geological Survey) for New 
England streams and for those of the southern Appalachian district 
shows that for the northern district, with many lakes, the minimum 
ranges from 0.2 to 0.4 cu. ft. per sec. per sq. mile on water-sheds of 
from 1 000 to 2 000 sq. miles, while in streams of the southern district, 
with practically no lakes or swamps, it ranges from 0.4 to 0.6 cu. ft. 
per sec. per sq. mile for water-sheds of the same area. The comparison 
is made for equal areas of water-shed, as the general law appears to 
be that the amount of run-off is in an inverse ratio to the water-shed 
area, that is, the smaller the water-shed, the greater the minimum run- 
off per square mile to be expected for any particular stream. 
Mr. Smith. Cecil B. Smith, M. Am. Soc. C. E. (by letter). — Mr. Holgate's dis- 
cussion, suggestive, as it is, of the widespread and extensive latent 
hydraulic powers capable of development in various parts of Canada, 
might do well for the subject of future papers, as even the Engineering 
Profession itself is not well informed on this subject; the writer's 
intention, however, was to outline broadly the present achievement. 

In Sweden and Norway a large block of power is devoted to 
nitrogen fixation and other electro-chemical uses, and one may look 
confidently to the same action as regards some of the larger available 
powers on the St. Lawrence and Atlantic seaboard; in the interior, 
however, one may expect that industrial demand, keeping pace with 
population and manufactures, will, as in the past, be the determining 
feature. Too much attention, however, cannot be given to those im- 
pediments, partly physical, partly political, which have ended in many 



DISCUSSION : IIYDRO-ELECTRIC POWER IN CANADA 201 

constructions being financial failures. The physical obstacles are indi- Mr. Kmith. 
cated by Mr. Ilolgate, in that the design which will give the best 
future results is often attempted prematurely, making the fixed invest- 
ment too large for the revenue obtainable during the first few years 
after operation has commenced; or, on the other hand, very frequently 
cheap temporary construction is adopted for the purpose of lessening 
the capital cost, with the result that just as the plant begins to secure 
a large market, the works must be rebuilt in order to meet the current 
demands. 

It is a common fallacy, which finds expression in the public mind, 
that power companies are making excessive profits, but the public does 
not fully recognize that a heavy original investment must often wait 
for profits and take the risk. 

The logical deduction is that where, as in Canada, the water-powers 
are still largely vested in the Crown (or State), the wise policy, looking 
to the best industrial interests of the country, would be to grant power 
rights to responsible parties on very generous terms, but with a strict 
requirement as to actual tise, and a tight rein as to rate regulation by 
a commission appointed for the purpose. As long as the investing 
public retained confidence in the fairmindedness of such a federal 
court or commission, there would be every encouragement to develop 
on liberal lines, and this is the present policy of the Dominion Govern- 
ment. Mr. Beugler goes a step farther and suggests subsidy, and, as 
there is a strong demand on the part of the public for electric power 
at low rates, this is hardly compatible with speculative risk and long 
shots for large future profits, or, in other words, large stock issues in 
proportion to the bonds. 

Canada has had many years of subsidy land grant, and now has a bond 
guaranty for railways, and might it not be a fair balance of interests 
if a city, province, or state, or even the Federal Government, would, for 
large undertakings involving heavy initial outlays, guarantee the bond 
interest, thereby lowering the rate of interest and doing away with the 
excuse of large stock issues. 

Combining bond guaranty with rate regulation would place hydro- 
electric investments on a sound low-rate basis, thereby assuring the 
public of not only a low rate for light and power, but a steadily 
decreasing rate as the volume of business increased. 

From residence in the Southern Appalachian districts the writer 
would conclude that the minimum flows were much less than those 
quoted by Mr. Beugler, in fact, there is good authority for a minimum 
on these rivers of less than 0.1 cu. ft. per sec. per sq. mile of water-shed. 
As regards area of water-shed, the writer's experience has been that 
the smaller the area, the less the minimum run-off per square mile. 
This accords also with the law of probability, as a small water-shed 
would be much more apt to suffer from an acute drought than a 
large one. 



AMERICAN SOCIETY OF CIVIL ENGINEERS 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1124 



TE8T8 OF IJUILT-Ur vSTEEL AND WROUGHT-IRON 
COMPKE8SI()N riEOES.* 

By Arthur N. Talbot, M. Am. Soc. C. E., and Herbert F. Moore, Esq. 



With Discussion by Messrs. John C. Moses, and Arthur N. Talbot 
AND Herbert F. Moore. 



The principal tests described in this paper were made on the follow- 
ing compression pieces: (a) a steel column (called Column No. 1) 
built up of angles, plates, and lattice bars, all the parts being light; 
(h) four wrought-iron bridge posts which had seen long service in a 
bridge truss; and (c) three posts and a top chord in a railroad bridge 
under service. The tests of (o) and (h) were made in a testing 
machine; for (c) a locomotive and cars formed the load. Auxiliary 
tests were made on lattice bars and other parts. 

These tests were taken up with a view of determining experi- 
mentally: (1) something of the variation of stress, both longitudinal 
and lateral, throughout the channel members of the compression 
piece; (2) something of the amount and distribution of stress in the 
lattice bars of columns, and the action of similar bars under separate 
test; and (3) the general relation between the component parts and 

I 
the column as a whole. The slenderness ratio of the columns, —, ranged 

from 38 to 66. It will be seen that, instead of following the more 
* Presented at the meeting of June 2d, 1909. 



TESTS OF COMPRESSION PIECES 203 

common effort to determine the effect of length of column, emphasis 
was placed on measuring the distribution and range of stress over 
the various parts of the column. 

It is unnecessary to state that built-up compression pieces (whether 
long or short) are not perfect, the natural imperfections of the com- 
ponent parts being increased in the process of fabrication. To non- 
homogeneity of structure and lack of straightness in the component 
angle or channel are added such further imperfections as kinks and 
eccentric connection of parts, which go to increase the opportunities 
for local flexure in the component parts and for flexural stresses in the 
column as a whole. An attempt has been made in these tests to measure 
the deformations in the presence of such conditions, and to find the 
general distribution of stress. In view of the many limitations sur- 
rounding such tests, the results are to be taken as suggestive and 
qualitative, and not as exact determinations. 

Results of Tests. 

The following observations on the results are given here with a 
view of showing the trend of the tests : 

/. Variation of Stress Throughout the Length and the Cross- 
Section of the Channels and of the Column as a Whole. — Channel 
members show evidence of considerable local flexviral action, such as 
may be produced by lack of straightness or by any method of applying 
the load eccentrically. This is especially true in the flimsier column. 

The condition of flexure varies markedly throughout the length of 
the member, in some cases the maximum compression in one cross- 
section being at the extreme fiber on one side of the channel, and in a 
near-by section the other side of the channel showing the excess of 
stress. 

There were indications of sudden changes in the relative amount 
of stress carried by the two channels at near-by sections, indicating 
general flexure of the column. 

The measurements made indicate in a number of cases stresses in 
the extreme fiber from 40 to 50% in excess of the average stress, and in 
some cases even higher. 

The amount of eccentricity necessary to account for the increase 
of stress foiuid in individual channels, based on lack of straightness 
and the ordinary theory of flexure, is relatively small. 



204 TESTS OF COMPRESSION PIECES 

//. Stress in Lattice Bars. — The amount of deformation observed 
in lattice bars is relatively small. 

The measurements indicate a stress in the lattice bars which 
would be produced by a transverse shear equal in amount to 1 to 3% 
of the applied compression load, or to that produced by a concentrated 
transverse load at the middle of the column length equal to 2 to 6% 
of the compression load. 

///. Tests of Lattice Bars. — Tests of lattice bars for load-carrying 
capacity show that the usual form of bar is a very inefficient com- 
pression member when loaded eccentrically through a riveted con- 
nection. 

Under common conditions of loading, the maximum fiber stress may 
be as much as three times the average stress. 

In the compression tests of single lattice bars, the ultimate strength 
was in no case as much as one-half of the elastic limit of the 
material. 

IV. Relation hetween Component Parts and the Column as a 
Whole. — It is a question whether the component members of a built-up 
column act together to form an integral compression piece, especially 
in resisting oblique distortion. 

The distribution of stress under working loads, and even up to 
incipient failure, may be quite different from that which exists when 
the column becomes crippled. This is due to the yielding of the 
more strained parts after the yield point is reached at any fiber, and 
the consequent redistribution of stress. 

No relation has been found between the stresses actually observed 
and the stresses computed by column formulas. 

The sudden failure of a test column at a relatively low load by 
buckling of lattice bars is accounted for when the amount of trans- 
verse shear developed in other test columns and the strength of lacing 
found in lattice-bar tests are taken into consideration. 

Laboratory Tests of Columns. 

Description of Columns. — One steel column and four wrought-iron 
columns were tested. The steel column was built especially for the 
purpose of this test by the American Bridge Company at the Lassig 
plant. The wrought-iron columns were halves of bridge posts taken 
from an old bridge of the Chicago, Burlington, and Quincy Railroad. 



TESTS OF COMPRESSION PIECES 
COLUMN N0.1 



205 



o o o o o o. 




1 
J 


;— 134/— > 


r 

L 




o 0000 o 



o 000 00 




Fig. 1. 



WROUGHT-IRON COLUMNS 

15'l0" 




F\.a% I H X i2)i 



o o o^ 
o o cy^ 



L ^ciK- 


c 


PI 17 I 13 J 1^ 


^ 


^0 




t '^oy' 


xoy-- 


^ 



CROSS-SECTIONS OF COLUMNS 



n 


< 13M" > 


r 






L 



Column Xo.l 
2 Plates 20 x?^" 
4 Aiigles2°x2°xM"' 
Area= 18.76 sq.in. 




Wrought Iron Tests 

Nos. 2.3,1, and 5 

2-10"ClianiieIs 

Old Model 

Area= 17.61 sq.in. 



Posts 

White Heath Bridge 

2-12 "Channels 

Old Model 

Ai-ea = 12.02 sq.in. 



Upper Chord 
White Heath Bridge 

2 Plates U'x H/iG 

4 AngIes3}^x3HxJIo 

1 Cover-Plate 21 x^ 

2 Flats 4x5^' 

Area =18.67 sq.in. 



Fig. 3. 



206 



TESTS OF COMPRESSION PIECES 

TABLE 1.— Data of Columns. 





Column 
No. 1. 


Wroc ght-Iron Posts. 


White Heath Bridge. 


Column designation. 


Nos.2,3,4, 
and 5. 


Retests 
2oand 
4a. 


Posts 
U^ L^ N.. 

U:, 1,3 S. 


Upper 
chord 

south. 


Area of column section, in inches 

Length, center to center 

End conditions 


18.76 
21ft. 
Pin 
parallel to 
lacing. 
37.8 

37.2 

593 

37.7 

1 by J in. 

and 

l|by Tcin. 

Single. 

63° 30' 


17.64 
15 ft. 10 in 

Pin 

parallel to 

lacing. 

43.5 

41.2 

400 

33.7 
2i by i in. 

Double. 
45° 


17.64 
14 ft. 7 in. 

Pin 

parallel to 

lacing. 

40.1 

38.0 

367 

33.7 
2^ by i in. 

Double. 
45° 


12.02 

25 ft. 
Lower end 
pin, upper 
end riveted 

66.1 

41.0 

416 

22.2 
2i by g in. 

Double, 
riveted at 
crossing. 

45° 


48.67 

19 ft. 10 in. 

Riveted. 


, axis parallel to lacing 

; , axis perpendicular to lacing. 

of each flange member, axis 

'" perpendicular to lacing for 

full length of column 

of each flange member, axis 
** as before, for distance be- 
tween adjacent lacing rivets. 
Lattice bars, dimensions of sec- 
tion, in inches 


40.7 
29.6 

2J by g in.. 


Kind of lacing 


on bottom. 
One cover- 


Angle of lattice bar with axis 
of column 


plate. 

45° 
on bottom. 



Table 1 gives the dimensions of the columns. The details are shown 
in Figs. 1 and 2 ; and Fig. 3 shows the cross-sections. The steel column 
(Column No. 1) was designed with relatively thin component parts 
and relatively large outside cross-section dimensions. It will be seen 
that this column is not so stocky as the usual bridge column. This 
section was chosen because it seemed to offer better opportunities 
under the conditions of the test for the study of the distribution of 
stress, longitudinal and lateral, and especially of the stresses developed 
in the latticing, than a less flimsy column. In the earlier tests, the 
lattice bars of Column No. 1 were fastened in place with turned bolts 
in reamed holes so that one set of bars could easily be replaced by 
another. After several tests were made with the lacing thus fastened, 
the bars were riveted in place for the later tests. The turned bolts 
in reamed holes apparently held the column together in the column 
tests as well as did the rivets. The proportions of the wrought-iron 
bridge columns were in no way unusual, and the posts represented 
good practice at the date of construction of the bridge. The columns 



TESTS OF COMPRESSION PIECES 207 

became available through the replacing of the bridge with a heavier 
structure; they were apparently in good condition. 

All the columns were pin-ended, the pin being parallel to the plane 
of the lacing. The details of the ends are shown in Figs. 1 and 2. 
The wrought-iron posts were cut in two. One end was left as used 
in the bridge, and bearing plates and batten plates were bolted to the 
other end. 

Procedure of Tests. — The method used in studying the distribution 
of stress was to measure at various parts of the column, by means of 
extensometers, the deformation over short distances, and its variation 
over a section perpendicular to the axis of the column. The extensom- 
eters were placed outside of the flanges of one of the channels in 
such a way that the shortening at any point of the cross-section of the 
member could be determined on the hypothesis that in this part of the 
channel a plane section before loading remained a plane section after 
the load was applied. It is seen that this hypothesis is not dependent 
upon the integrity of the column as a whole. A similar arrangement 
was made in observing deformations in lattice bars. The position of 
the instruments is shown by Fig. 1, Plate XII. Two sets of instru- 
ments were used, and sometimes more. Generally, they were placed 
on the front and back flanges of one channel and then on the other, 
but sometimes on the front or hack flanges of two channels at the same 
time. This procedure necessitated the removal of the instrimients to 
new positions after several applications of the load, and, in the complete 
test of any column, involved several hundred applications of the load. 

In all the tests to determine the distribution of deformations, care 
was taken not to exceed the elastic limit at any point. Under these 
conditions, it is considered, for the purposes of the discussion, that the 
stress developed at any point is proportional to the deformation. The 
ratio of the intensity of the stress to the deformation per unit of length, 
therefore, will be taken to be equal to the modulus of elasticity of the 
material. 

Extensometers. — In the observations of deformation in the channel 
members of the columns, the extensometers used in most of the tests 
were Ames test gauges mounted on suitable frames. These gauges 
were graduated to ^ vois ^^-i ^^^^^ were read by estimation to yo-fynn ^^^ 
These instruments magnify the change of length by means of a clock- 
work device rotating a hand on a dial. 



208 



TESTS OF COMPRESSION PIECES 



ATTACHMENT OF EXTENSOMETERS 
TO CHANNELS 




Fig. 4 shows the method of attachment of these extensometers to a 
column. The gauged length was usually 8 or 9 in. As will be seen, 
the extensometers measured the change of length between points 
slightly outside of the members tested. From the actual instrument 
readings were computed the deformations and stresses at the extreme 
fibers of the flanges of the channels by the ordinary method, which 
assumed a rectilinear distribution of deformation and stress. 

The type of instrument used is light, simple, and convenient, and, 
under very severe usage in other tests, it had shown itself to be 
durable. The limits of accuracy of 
the instrument were fairly well 
determined by careful calibration, r^ ^ 
While the maximum errors pos- V^ 
sible are large, it gives fairly 
trustworthy results under careful 
manipulation, and any available 
instrument of greater precision 
would be too bulky or too likely 
to have its parts deranged under 
the conditions of column tests, 
especially in field tests, to be en- 
tirely satisfactory. 

The accuracy of all these instruments was tested by comparison 
with a Brown and Sharpe micrometer acting through a 10 to 1 lever. 
Basing judgment on the maximum deviation observed in the calibra- 
tion, and on the smallest deformations measured in any test of 
columns, it is felt that the error in stress determination for the 
channel members is generally less than + 10%, and that it is well 
within this limit for most of the determinations in the laboratory 
tests. This general limit of accuracy is corroborated by a comparison 
of the average stresses on the center of gravity of the flange members 
with each other and with the known average stress. To those accus- 
tomed to greater precision in calciilations and to the greater refine- 
ment of most laboratory tests, this limit of accuracy may seem crude. 
Having in view that the purpose of the tests was qualitative rather 
than quantitative, and considering the great variation in the distribu- 
tion of stress found in the columns, and the general consistency of the 
results, it may be considered that the instruments were satisfactory. 





PLATE XII. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXV, No. 1124. 

TALBOT AND MOORE ON 

TESTS OF STEEL AND WROUQHT-IRON COLUMNS. 



Pl _ i 


/v'^'v^^^^^^^^^^^H 






^-i^j^ 





Fig. 1.— Extensometers in Place. 



Fig. 3.— Attachment ok Instruments in 

Cross-Bending Test of 

Column No. 1. 




Fig. 3. —White Heath Bridge and Test Train. 



TKSTS Oi'' OOMrUESSION TIECES 



209 



111 several of the tests, both in the laboratory and in the field, 
extensometers of the Johnson type were also used, but it was found 
that the results were not trustworthy, and no account of these measure- 
ments is given here. 

The Ames test gauges were also used to measure deformations in 
the lattice bars of all columns but No. 1. On account of the very low 
stresses in the lattice bars, it is felt that the stresses determined in 
them in this way may be in error by ± 20%, or even more. In Column 
No. 1, a Ewing extensometer, reading to ^(t^oo in. was used, and the 
stresses in the lattice bars of this column are judged to have been 
determined with an error in any case of not more than ± 10 per cent. 

Testing Machine. — The testing machine used was the Eiehle, 
600 000-lb., vertical, screw-power 
testing machine in the Laboratory 



of Applied Mechanics of the Uni- 
versity of Illinois. This machine 
has a clear space of 36 in. between 
screws, and will take in compres- 
sion specimens 25 ft. long. The 
speed of head used in these tests 
was in nearly all cases 0.4 in. per 
minute. This machine is equipped 
with massive guide frames which 
take the very heavy side thrust 
which occurs on the cross-head 
when a specimen is tested under 
oblique load. These guide frames 



METHODS OF LOADING 





Central 

Loading 

Column No,2a and 

Column No.l for 

Tests 11 and 14 

Fig. 5. 



Loading 

Column No.2a and 

Column No.l for 

Tests 12,13, and 15 



are entirely independent of the weighing mechanism. 

Method of Loading. — Two methods of loading were used, central 
and oblique. In all cases the load was applied at the ends of the 
column through the pins, and in a plane at right angles to the web 
and passing through the nominal axis of the column. In central 
loading, the pin was adjusted to an even bearing on the machine, the 
object being to apply the load equally to the two webs of the column. 
In oblique loading, the pin was supported on a narrow block, as 
shown in Fig. 5, in such a way as to secure a given eccentricity. The 
center of the block was taken as the point of application of the load. 
This is approximately true, the error probably not being greater than 



210 TESTS OF COMPRESSION PIECES 

i in. In two tests with Column No. 1, the point of application of 
the load is uncertain, as by an oversight the two bearing pieces used 
on each pin were placed unsymmetrically with respect to the axis of 
the column. It seems probable that the loading was nearly central. 

Routine of Tests for Stress Distribution. — In' a test for stress dis- 
tribution, the column was placed in the machine and a light initial 
load was applied. The extensometers were then attached in position 
to measure the deformation at some point of the column, and an 
initial reading was taken. A known load was applied by the testing 
machine and the instruments were read again. The load was then 
released to its initial value and another application of the load was 
made. If the second readings did not exactly check the first, further 
applications of the load were made. In cases where the observed 
deformations were large or seemingly abnormal, the test was repeated 
at another time, and in some cases as many as ten observations were 
made on the same gauged length. In some of these cases the instru- 
ments were reset, their places being exchanged. The instruments 
were next attached in a new location, and the process was repeated. 
Thus the stress distribution in various parts of the column was finally 
determined. The load generally used in the laboratory tests was 
10 000 lb. per sq. in. of section of the column in excess of the initial 
load. 

Results of Tests for Stress Distribution in Channels. — Tables 
2 and 3 give results of the tests to determine stress distribution and 
variation in the flange members found in twelve of the column tests. 
The stresses given are calculated from the observed deformation, 
using for the modulus of elasticity 28 000 000 lb. per sq. in. for steel 
and 26 000 000 lb. per sq. in. for wrought iron, these values checking 
closely with the total shortening of the columns and with the average 
deformations observed throughout their length. As heretofore de- 
scribed, the stress noted is the average over a space of 4 or 44 in. 
on either side of the point indicated. Any lack of agreement between 
the average stress on the center of gravity of the flange members and 
the average stress for the load applied is probably due principally to 
instrumental errors. 

Figs. 6, 7, and 8 show graphically the stress distribution and varia- 
tion. The full line gives the stress at the east side (front) and the 
dotted line at the west side (back). 



TKSTS OF COMPRESSION PIECES 

TABLE 2. — Stresses in Column No, 1. 



211 



North Channel. 


South Channel. 


West side. 


East side. 


West side. 


East side. 


"3 
a 




Q 03 

5^ 


II 


1^ . 




i| 




o>; 

-2 > 


II 


6« 


a =« 





Test No. 1. 



1 


8100 


10 200 


10 600 


12 200 


10 000 


9 700 


6 500 


8 400 


8 600 


9 600 


10 900 


11300 


a 


9 700 


10 100 


10 100 


8 100 


9 600 


10 000 


11 900 


11500 


11 600 


8 900 


10 500 


11 100 


3 


7 500 


10 200 


11 000 


11 800 


10 000 


9 700 


11500 


9 800 


9 800 


9 000 


10 300 


11 000 


4 


10 500 


10 000 


10 000 


10 500 


11000 


11 000 


9 000 


8 600 


8 600 


lOSOO 


10 100 


10 500 


5 


9 600 


9100 


9 100 


10 900 


9 700 


9 500 


10 300 


8 500 


9 700 


8 800 


10 100 


10 600 


6 


9 900 


11000 


11 100 


10 000 


9 600 


9 700 


9 400 


8 400 


8 100 


9 000 


9 200 


9 100 


7 


10 000 


7 600 


6 900 


11 000 


10 000 


9 900 


9 000 


9 400 


9 9CK) 


9 200 


10 000 


10 400 


8 


8 too 


8 900 


9 000 


10 800 


11300 


11 500 


12 700 


10 500 


10 400 


12 200 


10 100 


10 300 


9 


10 000 


11000 


11 100 


9 400 


11 000 


11000 


14 500 


10 000 


9 000 


9 000 


9 200 


9 200 


10 


14 0(W 


10 000 


9 300 


14 300 


11800 


11 000 


5 400 


8 900 


9 300 


7 200 


9 600 


10 000 


11 


8 600 


9.500 


9 600 


10 000 


9 800 


9 600 


9 000 


8 800 


8 400 


9 800 


11000 


11 400 


12 


13 300 


9 200 


8 700 


9100 


9 400 


9 800 


10 800 


9 300 


9 000 


10 700 


9 600 


9.500 


13 








8 700 


8 800 


9 000 


7 800 


9 000 


9 200 









Test No. 2. 



6 


11500 


12 100 


13 100 


11500 


10 000 


9 500 


9 100 


9 000 


9 300 


5 800 


7 500 


7 700 


7 


12 200 


12 600 


12 700 


14 600 


13 500 


13 300 


5 900 


7 500 


8 000 


8 300 


9 500 


9 500 


8 


11 300 


12 500 


12 800 


12 400 


12 500 


10 600 


7 800 


8 200 


8 300 


7 500 


10 300 


10 700 


9 


14 800 


13 300 


12 700 


9 500 


10 500 


10 600 


11 800 


7 300 


7 000 


10 .300 


8 400 


8 300 


10 


9 000 


10 200 


10 400 


16 200 


10 500 


9 600 


7 800 


10 500 


11 400 


9 000 


8 400 


8 500 


11 


15 200 


10 200 


9 700 


13 700 


11 600 


11 500 


7 600 


8 100 


8 300 


5 000 


6 300 


8 6(J0 


12 


10 300 


11900 


12 300 


12 700 


11600 


11 500 


9 400 


8200 


7 700 


7000 


7 500 


9 700 


13 


13 200 


11800 


11700 


9 900 


10 500 


10 800 


7 500 


8 600 


9100 


6 300 


6 800 


8 400 



Test No. 3. 



1 








13 800 


10 400 


9 500 


7 500 


8 800 


8 900 








2 


8 900 


10 600 


10 900 


8 800 


10 400 


10 600 


8 300 


9 300 


9 400 


10 000 


9 700 


9 700 


3 


10 500 


9 800 


9 800 


8 700 


9 000 


9 200 


10 400 


8 700 


8 700 


11900 


10 500 


10 300 


4 


9 800 


10 800 


11 000 


11 100 


10 700 


10 700 


8 300 


8 300 


8 500 


13 300 


12 600 


12 200 


5 


10 500 


10 000 


9 800 


11 200 


10 700 


10 700 


9 000 


10 000 


10 100 


9 300 


9 500 


9 400 


6 


11300 


13 300 


12 500 


10 900 


10 500 


10 700 


10 400 


9 700 


9 500 


7 500 


9 400 


9 600 


7 


10 400 


10 700 


10 700 


11 900 


11 000 


10 800 


9 200 


10 000 


10 000 


8 400 


8 800 


8 800 


8 


12 100 


10 200 


10 200 


12 100 


13 000 


13 100 


8 500 


9 300 


9 400 


7 300 


9 000 


9 400 


9 


7 600 


9 900 


10 200 


10 500 


10 900 


10 800 


14 400 


9 300 


8 400 


10 900 


10 600 


10 700 


10 


10 500 


12 000 


13 100 


14 200 


11 000 


10 500 


7 200 


10 400 


10 900 


7 800 


8 700 


8 700 


11 


14 200 


10 500 


9 700 


7 500 


8 500 


8 600 


8 300 


9 400 


9.S00 


7 000 


9 400 


9 800 


12 


10 400 


12 000 


12 300 


6 500 


7 500 


7 800 


11 000 


9 300 


9 000 


7 900 


9100 


9 400 


13 


12 600 


9 800 


9 500 


8 300 


8 400 


8 200 


7 200 


9 800 


10 200 


11 000 


10 800 


10 900 



313 TESTS OF COMPEESSION PIECES 

TABLE 2. — Stresses in Column No. 1 — (Continued). 



North Channel. 


South Channel. 


West side. 


East side. 


West side. 


East side. 


04 


si 


a ?3 


1 j^' 
Is 




5^ 






CM 


II 




CM 

a g 

6^ 


i| 



Test No. 4. 



1 

2 


11300 


10 800 


10 700 


10 500 


11 100 


11 200 


6 100 


3 000 


2 500 


8 900 


10 200 


10 600 


3 


10 000 


1 1 600 


11 600 


10 100 


11 100 


11 100 


8 400 


8 500 


8 500 


12 000 


11 300 


11 100 


4 


11 200 


13 300 


12 500 


9 900 


10 900 


11 000 


7 600 


9 400 


9 700 


7 900 


9 300 


9 5U0 


5 








12 800 


11 100 


10 700 








10 700 


10 200 


10 100 


6 


10 100 


5 400 


4 500 


12 800 


10 700 


10 300 


6 500 


8 900 


9 500 








7 


8 000 


10 600 


11 (X)0 


10 200 


9 700 


9 700 








8 700 


8 500 


8 600 


8 


9 800 


laooo 


12 400 


9 700 


10 000 


10 100 


9 0U0 


9 400 


9 400 


10 300 


11300 


11400 


9 


11500 


10 600 


10 400 








8 500 


10 700 


10 900 


7 900 


11 100 


11500 


10 


8 600 


8 300 


8100 


6 800 


8 700 


9 100 


10 800 


7 500 


7 000 


9 300 


10 700 


11 000 


11 














9 500 


9 000 


8 900 


13 000 


12 700 


12 600 


la 


10 900 


10 400 


10 300 


10 700 


9 900 


9 600 














la 


11000 


10 700 


10 700 


11100 


10 600 


10 500 


7 300 


9 800 


9 600 


12 700 


11900 


11800 



Test No. 5. 



V6 


12100 


10 400 


9 300 


11000 


9 500 


8 700 


9 700 


10 500 


10 900 


10 400 


10 800 


10 900 


1 


9 800 


11 100 


11 800 


10 000 


10 200 


10 500 


12 200 


10 400 


9 300 


13 5O0 


11 300 


10 400 


1V4 


7 900 


8 500 


8 600 


5 800 


8 200 


8 900 


7.S00 


8 400 


8 900 


7 400 


7 900 


8 300 


2 


11800 


11 400 


11 100 


10 900 


10 600 


10 600 


8 500 


8 600 


8 700 


9 800 


10 200 


10 400 


2^ 


10 700 


11 000 


11 100 


10 800 


10 200 


10 200 


11800 


11 900 


11800 


9 700 


9 700 


9 600 


3 


5 800 


7 500 


8 500 


7 800 


8 000 


8 500 


8.300 


8 500 


8 000 


7 700 


9 100 


9 700 


3^ 


8 000 


8 100 


8 500 


7 400 


8 100 


8 500 


7 800 


8 600 


9 000 


7 500 


7 600 


7 700 


4 


15 400 


13 000 


12 400 


13 500 


13 000 


12 800 


11 600 


11800 


11800 


9 900 


9 900 


9 900 


4U. 


12 200 


10 700 


10 300 


12 200 


10 200 


9 200 


9 900 


9 900 


9 800 


9 800 


9 800 


9 700 


5 


9 400 


9 500 


9 600 


11000 


10 100 


9 600 


9 700 


9 500 


9 500 


8 900 


9 800 


9 900 


5^ 


9 90<3 


10 700 


11 100 


10 500 


10 900 


11 000 


10 900 


9 100 


8 300 


8 700 


8 800 


8 900 


6 


7 400 


7 300 


7 200 


8100 


9 200 


9 700 


13 800 


11900 


11 100 


11 100 


10 600 


10 200 


fii^ 


10 900 


10 200 


9 700 


10 200 


10 300 


10 500 


8 600 


8 900 


8 900 


8 500 


8 600 


8 600 


7 


12 100 


11700 


11400 


11400 


10 900 


10 400 


8 300 


8 8O0 


8 900 


5 900 


7 900 


8 700 


7^ 


9 900 


9 200 


8 900 


9 600 


9 400 


9 500 


6 500 


7 600 


8.300 


4 600 


6 400 


6 900 


8 


9 800 


8 600 


8 300 


8 100 


9 700 


10 400 


11 400 


9 500 


8 500 


12 7(X) 


10 800 


10 100 


8H 


9 200 


9 700 


10 000 


9 700 


9 500 


9 300 


20 200 


15 400 


13 100 


15 400 


14 000 


13 200 


9 


6 500 


7 600 


8 300 


9 700 


9 700 


9 900 


12 900 


13 200 


13 100 


9 400 


9 900 


10 200 


91^ 


10 000 


10 400 


10 600 


12 900 


9 500 


8 100 


4 600 


7 800 


8 900 


4 600 


7 300 


8 500 


10 


16 400 


12 900 


11 200 


irooo 


14 400 


12 900 


8 000 


8 600 


8 700 


6 800 


8 000 


8 600 


10W 


11 300 


10 700 


10 400 


9 900 


10 000 


9 700 


10 500 


11 100 


11 300 


7 400 


8 900 


9 600 


11 


7 000 


8 400 


8 900 


8 100 


9 100 


9 600 


12 600 


11 100 


10 800 


9 800 


10 800 


11000 


IIU 


9 700 


8 900 


8 600 


11 900 


10 500 


9 700 


8 200 


11 100 


11700 


8 900 


9 700 


9 800 


12 


14 000 


12 400 


11600 


11900 


11 000 


10 400 


6 900 


7 500 


7 800 


8 100 


8 800 


9 400 


12}^ 


9 700 


8 900 


8 600 


10 500 


12 100 


12 700 


10 800 


11800 


12 400 


11 900 


11000 


10 800 



TESTS OF COMPRESSION PIECES 313 

TABLE 3. — Stress in Wrought-Iron Bridge Posts. 



North Channel. 


South Channel. 




West side. 


East side. 


West side. 


East side. 


Ph 


^1 
5^ 


o ^ 

t-. 
o 






t4 
o 




2 5j 


Center 
of giavity. 

Inner 
fiber. 




o 


II 



Column No. 2, Test No. 6. 



I 


4 600 


6 500 


7 300 


6 400 


7 900 


8 500 


4 300 


8 300 


9 700 


5 000 


6 800 


7 400 


2 


7 300 


7 3U0 


7 300 


8 :iOO 


8 400 


8 400 


9 800 


9 000 


8 800 


9 000 


8 000 


7 600 


3 


5 500 


5 900 


6 000 


7 100 


7 80J 


8 000 


7 700 


8 800 


9 200 


6 700 


7 600 


7 900 


4 


6 200 


6 100 


6 200 


6 100 


6 900 


7100 


8 400 


8 800 


9 000 


8 200 


8 000 


8 000 


5 


7 20(J 


6 500 


6 100 


8 800 


8 600 


8 500 


7 500 


8 800 


9 300 


8 700 


7 800 


7 500 


6 


6 800 


6900 


6 9O0 


8 400 


8 400 


8 400 


7100 


7 600 


7 800 


6 900 


7 200 


7 400 


7 


7 500 


7 900 


8 000 


7 100 


7 400 


7 500 


7 500 


8 200 


8 500 


7 500 


7 400 


7 800 


8 


6 400 


7 400 


7 800 


6 200 


7 200 


7 600 


7 700 


7 900 


8 000 


6 600 


6 400 


6 100 



Column No. 3, Test No. 7. 



1 


9 300 


9 000 


9 000 


8 400 


9 500 


10 000 


10 000 


9 000 


10 100 


8 200 


8 000 


7 900 


2 


10 200 


10 000 


9 900 


10 200 


11 500 


11900 


8 400 


9 100 


10 000 


9100 


8 700 


8 500 


3 


8 800 


8 700 


8 600 


9 800 


10 000 


10 000 


10 000 


9 700 


9 600 


9 400 


9 600 


9 600 


4 


10 500 


9 100 


8 600 


8 500 


9 500 


9 900 


8 600 


9 800 


10 200 


9 300 


7 500 


6 900 


5 


10 200 


10 400 


10 500 


8 800 


9 600 


9 900 


10 100 


9 800 


9 700 


10 400 


10 200 


10 200 


6 


10 600 


10 000 


9 800 


10 300 


10 000 


10 000 


10 400 


10 400 


lf> 400 


10 000 


9 900 


9 900 


7 


10 300 


9 900 


9 700 


9 600 


9 900 


10 000 


8 700 


9 500 


9 700 


10 900 


10 000 


9 700 


8 


9 800 


9 600 


9 500 


10 200 


10 900 


11 000 


7 600 


9 600 


10 300 


7 800 


9 900 


10 600 



Column No. 4, Test No. 8. 



1 


9 000 


9 800 


10 000 


10 000 


10 700 


11 000 


8 200 


10 000 


10 800 


9 000 


9 200 


9 300 


2 


8 700 


9 300 


9 500 


10 100 


10 800 


11 100 


10 100 


11 600 


12 100 


8 600 


9 400 


9 700 


3 


9 300 


9 800 


9 900 


10 800 


11 200 


11 400 


10 000 


10 200 


10 200 


7 400 


8 400 


8 700 


4 


9 500 


S700 


9 800 


11 100 


10 200 


10 000 


9 700 


11 100 


11600 


7 100 


8 200 


8 500 


5 


10 300 


10 500 


10 600 


11 900 


12 400 


12 600 


8 400 


9 400 


9 800 


7 700 


8 800 


9 200 


6 


11 600 


12 000 


12 100 


12 400 


13 600 


13 900 


9 500 


10 800 


11100 


7 400 


9 500 


10 200 


7 


12 800 


12 500 


12 400 


13 400 


14 200 


14 600 


7100 


8 300 


8 700 


6 400 


6 700 


6 700 


8 


8 600 


11600 


12 700 


13 000 


11300 


10 700 


6 600 


8 500 


9 200 


6 200 


7 800 


8 400 



Column No. 5, Test No. 9. 



li 


14 COO 


11700 


10 800 


12 600 


11700 


11400 


7 400 


8 100 


8 400 


7 400 


7 900 


8 000 


1 


11300 


12 000 


12 300 


11 100 


10 000 


9 600 


6 700 


8 500 


9 200 


5 100 


6 800 


7 500 


m 


UHOO 


11 100 


11 100 


11 100 


10 800 


10 700 


5 700 


6 200 


6 300 


6 600 


6 900 


7 100 


2 


13 100 


12 600 


12 400 


12 900 


13 400 


12 200 


8 000 


8 000 


8 000 


7 800 


8 200 


8 300 


2.V^ 


12 700 


12 000 


11 700 


13 100 


11 500 


11 100 


9 300 


9 200 


9 200 


8 700 


8 400 


8 300 


3 


11800 


12 000 


12 200 


11 400 


12 200 


12 400 


8 200 


8 400 


8 400 


6 300 


8 000 


8 600 


S]4 


13 500 


10 5C0 


9 300 


11600 


10 700 


10 400 


8 200 


8 200 


8 200 


8 400 


8 300 


8 200 


4 


14 .300 


13 100 


12 700 


14 300 


12 300 


11 500 


7 300 


7 400 


7 500 


7 400 


8 700 


9 200 


4U 


11600 


11 600 


11 500 


12 300 


12 300 


12 300 


9 000 


9 000 


9 000 


8 700 


8 700 


7 800 


5 


12 600 


12 400 


13 300 


11700 


10 400 


10 000 


8 300 


8 400 


8 400 


7 400 


8 500 


8 100 


5W. 


13 700 


12 100 


11500 


13 500 


13 100 


11 400 


8 300 


8 300 


8 300 


8 400 


8 400 


7 600 


6 


12 600 


13 700 


12 700 


13 500 


13 .500 


12 200 


7 600 


7 600 


7 600 


7 300 


7 000 


7 100 


6)^ 


13 100 


11 900 


11500 


11 400 


11400 


11500 


8 400 


8 400 


8 400 


10 500 


9 000 


8 500 


7 


11 100 


11 300 


11500 


11 200 


12 300 


13 600 


7 900 


8 000 


8 000 


6 500 


8100 


8 600 


7U 


10 600 


10 100 


9 900 


nooo 


10 300 


9 900 


8 000 


8 200 


8 200 


5 900 


7 900 


8 700 


8 


10 300 


12 300 


12 800 


11300 


12 300 


13 400 


8 500 


8 600 


8 600 


10 800 


11800 


12 200 



^14 



TESTS or COMPRESSION PIECES 

TABLE Z~ (Continued). 



North Channel. 


South Channel. 


West side. 


East side. 


West side.. 


East side. 


i 




a =8 

O 


11 


II 


O 


11 




O 


S3C 


Outer 
fiber. 

Center 
of gravity. 


II 



Column No. 4a, Test No. 10. 



1 

2 
3 
4 
5 

e 

7 

7)4 



8 100 


.50f) 


10 200 


8 400 


8100 


8 000 


11 700 


10 800 


10 .500 


9 900 


1 
11 200 1 


8 300 


10 200 


10 800 


500 


7 200 


7 500 


12 100 


10 900 


10 400 


11 100 


12 500 


8400 


9 400 


9 700 


7 400 


8 200 


8 40(J 


12 100 


11 400 


11 100 


11 000 


11 400 


8 700 


8 800 


9 000 


6 200 


7 200 


7 500 


10 80;) 


10 500 


9 700 


9.300 


11 700 


8 300 


8 800 


9 000 


7 800 


6 900 


6 800 


11 800 


11500 


12 100 


10 900 


10 400 


8 700 


9 700 


9 900 


6 700 


500 


6 40(1 


13 300 


11 100 


10 400 


10 400 


11 100 


8 300 


8 000 


8 000 


7 300 


7 300 


7 4110 


13 000 


11 600 


11 100 


13 000 


12 400 


8 200 


9700 


10100 


6 800 


6 90O 


OKdil 


14 300 


13 500 


12 700 


13 100 


11 100 


7 500 


8 600 


9 100 


6 300 


7 200 


7 500 


12 300 


11 100 


10 800 


10 300 


15 200 


75a) 


8 400 


8 800 


7 000 


8100 


8 500 


13 000 


13 300 


13 500 


9 000 


11400 


6 300 


7 200 


7 500 


5 900 


6 100 


6 200 


12 700 


12 300 


12 100 


12 300 


10 700 


6800 


7 800 


8 200 


5 300 


6 800 


7 200 


14 200 


14 200 


14 200 


11 000 


11 200 


6 600 


7 800 


8 200 


7100 


7 100 


7 100 


15 400 


13 700 


12 900 


13 500 


n .500 


6 700 


7 100 


7 2TO 


4 900 


6100 


6 700 


1 1 900 


13 600 


14 600 


11 100 


12 100 


6100 


7 600 


8 300 


4 800 


8100 


9 400 


12 300 


12 400 


12 400 


9 300 


16 200 


6000 


8200 


9000 


8 400 


6 800 


6 200 


10 700 


12 200 


12 800 


10 000 


16 300 



11 800 
13 100 

11 600 

12 600 
10 300 
11400 
12 400 
10.500 

9 400 
12 300 

10 100 

11 300 

11 100 

12 400 
11 100 
11100 



Column No. 2a, Test No. 11. 



)4 


11600 


10 900 


10 400 


11400 


10 500 


10 000 














1 


7300 


9.300 


10 000 


7 300 


9 700 


10 400 


8 400 


8 800 


9 000 


10 200 


10 200 


10 200 


1)4 


11600 


9900 


9 600 


11 600 


10 900 


10 000 


8 500 


10 000 


10 600 


11 400 


12 000 


12 2TO 


2 


8 900 


8 700 


8 600 


10 500 


10 300 


10 000 


10 200 


9 000 


8 600 


11 500 


11 200 


11000 


2)4 


11 200 


9 700 


9 200 


10 900 


10 300 


10 000 


9 900 


10 000 


10 200 


11 700 


10 300 


9 900 


3 


9 500 


10 500 


10 700 


12 200 


10 800 


10 000 


11 800 


10 2(K) 


10 800 


11 4(J0 


10 600 


10.300 


3)4 


10 500 


11000 


11000 


12 100 


10 600 


10 000 


10 400 


8 600 


8100 


9 700 


9 000 


9 400 


4 


9 800 


9 500 


9 400 


10 600 


9 900 


9 600 


9 800 


9 400 


9 200 


10 400 


10 600 


10 500 


4)4 


10 300 


9 600 


9 300 


11 200 


10 200 


10 000 


8 700 


9 000 


10 400 


10 400 


10 000 


10 000 


5 


9600 


9 800 


9 900 


10 400 


10 800 


11 000 


10 300 


9 200 


8 800 


10 400 


10 200 


10 200 


5)4 


12 200 


11600 


11200 


12 000 


11 100 


10 700 


12 500 


10 400 


10 000 


12 200 


11900 


11700 


6 


10 100 


9 700 


9 700 


9 900 


9 700 


9 700 


9 100 


10 400 


10 900 


10 300 


10 700 


10 700 


6^ 


11300 


11000 


10 800 


10 600 


9 900 


9 700 


10 800 


9 900 


9 700 


11 000 


11 TOO 


lOOTO 


7 


9 300 


9 600 


9 709 


9 100 


9 300 


9 200 


9 800 


9 800 


9 800 


9 800 


10 8TO 


10 TOO 


7^ 


10 200 


10 900 


11400 


9 100 


9400 


9 500 


9 300 


9 600 


9 800 


9 900 


10 70(J 


10 800 


8 


9 900 


9 800 


9 700 


11100 


10 000 


9 600 


8 800 


10 300 


10 900 


10 800 


10 700 


10 800 



Column No. 2a, Test No. 12. 



H 


10 600 


10 000 


9 8TO 


10 4TO 


10 600 


10 600 


11 5TO 


10 6TO 


10 000 


13 200 


12 600 


12 300 


1 


8 4TO 


8 400 


8 500 


9 7TO 


10 800 


11200 


9 800 


10 700 


10 900 


lOOTO 


11 3TO 


11700 


1^ 


8 2TO 


8 000 


7 900 


9 2TO 


10 600 


10 900 


9 2TO 


9 600 


9 7TO 


8 800 


10 000 


10 200 


2 ■ 


9 200 


9 800 


10 000 


9 900 


10 900 


11 300 


11400 


10 000 


9 400 


12 200 


11400 


llOTO 


2^ 


10 800 


9 900 


9 600 


10 8TO 


9 600 


9 100 


10 900 


10 900 


10 900 


11 400 


11 600 


11600 


3 


9 100 


10 800 


11200 


11000 


10 6TO 


10 4TO 


10 600 


11200 


11 300 


10 4TO 


10 200 


10 100 


3^ 


10 7TO 


10 900 


10 800 


12 000 


lOfiOO 


lOOTO 


10 600 


10 900 


10 9TO 


10 2TO 


10 600 


10 600 


4 


10 000 


8 800 


8 400 


11 400 


9 400 


9 4(10 


8 9TO 


10 0TO 


10 3TO 


8 800 


10 300 


10 900 


4)4 


10 600 


10 000 


9 900 


10 200 


lOOTO 


9 9TO 


lOOTO 


10 100 


10 200 


8 600 


9 500 


9 800 


5 


lOlTO 


10 300 


10 400 


11300 


llOTO 


11000 


11800 


10 200 


9 600 


10 2TO 


9 700 


9 7TO 


5)4 


10 9TO 


11300 


113TO 


11400 


10 600 


10 300 


13 000 


11000 


10 400 


11 000 


9 500 


9 000 


6^ 


11500 


10 800 


10 5TO 


9 500 


9 000 


9 000 


9 500 


9 900 


10 000 


9 700 


9 500 


9 500 


(i}4 


n 7TO 


9 8TO 


9 200 


12 700 


10 300 


9 600 


10 000 


10 000 


lOOTO 


9 600 


10 3TO 


loeoo 


7 


9 900 


11 300 


11 6TO 


9 900 


11 100 


10 600 


10 500 


9 800 


9 3TO 


9 600 


10 000 


10 0TO 


7^ 


8 000 


9 700 


10 3TO 


15 400 


15 400 


15 4TO 


78TO 


lOOTO 


10 6TO 


6 800 


8 5TO 


9 TOO 


8 


10 000 


9 8TO 


9 900 


11700 


lOlTO 


9 900 


8 800 


10 4TO 


11100 


9 400 


9 4TO 


9 400 



TESTS OF COMPRESSION PIECES 

TABLE S— (Continued). 



215 







JSoRTH Channel. 








South Channel 






West side. 


East side. 


West side. 


East side. 






>. 






>» 












!>. 




'S 


^C 


*■> 




n-.^- 




«S"' 




H 


fes-; 


^"^fe 


41 ^ 


ife 


a 


0fi 


1-1 

o 






o 


si 




O 






o 


0$ 



Column No. 2a, Test No. 13, 



1 


4 5IX) 


7 000 


8 600 


7 000 


8 600 


9100 


9 800 


10 800 


11100 


10 300 


12 300 


13 000 


!J 


7 600 


9 000 


9 500 


10 300 


10 400 


10 600 


10 900 


10 100 


9 700 


12 800 


12 600 


12 600 


3 


7 000 


8 6U0 


9 100 


9 400 


10 100 


10 300 


11 300 


11 300 


11300 


11 600 


12 600 


13 000 


4 


8100 


9 100 


9 500 


10 400 


10 800 


10 900 


9 300 


10 400 


10 700 


9 600 


11600 


12.500 


5 


8 200 


9 500 


10 000 


10 000 


10 200 


10 200 


10 900 


8 900 


8100 


10 600 


10 700 


10 800 


6 


10 000 


10 400 


10 COO 


10 500 


10 000 


10 000 


8 500 


9 GOO 


10 000 


9 6C0 


10 400 


10 800 


7 


9 800 


9 800 


9 900 


10 200 


10 300 


10 300 


9 500 


8 800 


8 700 


9 900 


10 000 


10 100 


8 


8 000 


7 800 


7 700 


12 300 


10 300 


9 200 


7 700 


8 700 


9100 


8 700 


10 000 


10 500 



Table 4 gives a number of the most marked deviations from 
average stress. The excess of the maximum fiber stress is given as a 
percentage of the average stress. 

TABLE 4. — Maximum Observed Fiber Stresses in Flange Members 

OF Columns. 



u 

5s 



2 

3 

4 

5 

4a 

3a 

3a 

2a 






Features of Lacinp. 



U by j\ in., bolted 

I by I in., bolted.., 
1 by ;| in , 



Method of Loading. 



Central. 
Slightly eccentric. 

Central. 



Oblique arm 
Oblique arm. 



1 in. (Fig. 5). 
3 in. (Fig. 5). 



Percentage of excess 
of maximum fiber 
stress over average 

stress. Highest five 
values. 



42, 41, 
68, 64, 
50,41, 
35. 29, 
67, 55, 
31, 23, 

20, 17, 
41. 29, 

43, 43, 
53, 49, 

21, 16, 
43, 42, 
35, 34, 



39, 31, 23 
50, 49, 35 
35, 32, 19 
28, 27, 26 
49, 29, 27 

23, 21, 17 

12, 11, 9 

24, 32, 19 
38, 35, 35 
42, 40, 37 

13, 12, 11 
24, 23, 20 
23, 31, 21 



In most cases the maximum stress was in the outer fiber of the 
channel; sometimes very high stresses were found in the inner fiber. 
Generally, the stress in the opposite channel was correspondingly less. 



21Q TESTS OF COMPRESSION PIECES 

STRESS DISTRIBUTION IN CHANNELS OF COLUMN NO. 1 

NORTH CHANNEL SOUTH CHANNEL '^°«^'?,C„^4'J^^^ ^op '""^Center"- 

n f^.^^^rinner^oP Inner '^'^of Outer Outer ^'^"f Inner ^ ^ner o£.^ Outer 




stress in Lb. per Sq. In. Column No.l Test No.l 

^. SOUTH CHANNEL 

NOnTH c-""'^' ~ - 



NORTH CHANNEL 



Column No.l Te3t No.2 

SOUTH CH "■ 



LICSO iil IJ". J--- -^-l- TnilTH CHANNEL NORTH CHANntl. T„„ t OUU 1 H l-n/. i... t .- 

NOntH CHANNEL □TomT„„°f<,.nipr Outer Outer Center Inner ni£fii Inner Center Outer 
Outer Center Inner^Inner Center Gutter Outer ^ _ ._„. fF4 _pi,^,_^ of-^Fiber 




Coluniu No.l Test No.3 



Bottom 



Bottom 



Fig. 6. 



TESTS OF COMPRESSION PIECES 



217 



COLUMN NO. 3. TEST NO. 7 

NORTH CHANNEL m _ m SOUTH CHANNEL 



Outer Center of Inner 

Fiber Gravity Fiber 



Jl Top 



i i g 



_ East Side 
-West Side 



Inner Center of Outer 

Fiber Gravity Fiber 



I Bottomj 
COLUMN NO. 5. TEST NO. 9 

NORTH CHANNEL SOUTH CHANNEL 



Outer Center of Inneu 
Fiber Gravity Fiber 


Jjropjy 


Inner Center of Outer 
Fiber Gravity Fiber 


10 000 

13 000 

14 000 
10 000 

12 000 
14 000 
10 000 

13 000 

14 000 






'^A 


%%%%%%%%% 




', / 








/■ 


' 








^?^ 


/i\ ~ 


- -IT 


, 


^^ rr^^ r^=' 












v 








y 








^^ 


?;'- =7-= 




', 








\ 








'^N 




' . 


~\ 




" 1 ^ I 




n 








■'N> 


. 








>. 


N/ ^ 






' i it 




K 


















\J 


Ai\ - 






3 '^-^ 






^,, 
















1/ 


/ i \ 


',1 




- -S^ -^^ 




n; 


















V 


. . 






J^ ' ^ 






W, 








\ 










\/ 


/f 




'-^^ -.2 






' 








1 






























/ 








/' 




V, 




V ^^ 






// 






< 












\ 




3 












/ l\ 






/ 




V ~ 


\ 




3 7^ 




\ 


,' 






'■ ' 






N 




/k\ 


.', 




" r s!' 






\ 






K 








>, 








- I 






^ 






■^ 








\^ 


' 


{ 




-■• V' 






■^ 








ff- 








l\ 


\/ 


/ 




tl ^ 






y 


' 
















/4\ 


\ ) 




~ ii '^K 




















'7 






\r 




" T" 5 




1 








% 








.'s 


• 








ir z 




h 


^^ 






% 












V ' 


'L 




Ji 7 1 
























/^ 


/ 




\T ^x"*. 






















7*, 








E "^^ 






> 
















N^ 


. 


1/ 




7 2 














^1 








\>^ 


\/ 


■ 




- ^'" ,2 






// 








r 








/' 




y 




-r y 




/■ 








.f 










'/ 




.'/ 




-^ e2 










/ 










/ 


> 


' , 


■\ 




5 2 




1 








w 










' 


\/ 






I-W^- -^ V- 




i; 


,, 




s 










\ 


, 




H'' 










\ •~. 




1 


\' 








^ 






_ aL 




rii ^3 


i§§i§iii§ 


_n n 






1- -1 


°°.H« ^^"SS 



-West Side 
. Average 



PIQ. 7. 



218 



TESTS OF COMPRESSION PIECES 



Stress in Lattice Bars. — Table 5 gives the results of tests to 
determine the average stress in the various lattice bars of the columns. 
Tests 14 and 15 were tests on the lattice bars only. The distribution 
of stress over the cross-section of the bar is discussed in another place. 
The average stresses in the lattice bars are computed from the observed 
deformations, using a modulus of elasticity of 28 000 000 lb. per sq. in. 
for the steel column and 26 000 000 lb. per sq. in. for the wrought-iron 
columns. As might be expected, from the irregular variation of stress 
along the flange members of the columns, the stress in the lattice bars 
was found to vary greatly. 

Table 6 gives the largest stresses observed and the corresponding 
transverse shear. The transverse shear given in this table is that 
which would cause a stress in the lattice bars equal to the maximum 
stress observed in any lattice bar, and was computed by doubling the 
transverse component of the maximum load observed on a lattice bar. 
In the case of obliquely loaded columns, the transverse component of 
the load was computed on the assumption that the load was applied 

TABLE 5. — Total Stress, in Pounds, on Lattice Bars Under Load on 

Columns of 10 000 lb. per sq. in. 

Column No. 1. 



Lattice 


Test No. 5. 


Test No. 14. 


Test No. 15. 


bar. 


Front. 


Back. 


Front. 


Back. 


Front. 


Back. 


1 
2 

a 

4 

5 

e 

7 
8 
9 

10 

11 
12 
13 
14 
15 
16 
17 
18 
19 
20 
21 
22 
23 
24 
25 
26 


lOJc 

loot 



loot 

100c 
201)^ 
100c 
200f 




100c 


100c 

loot 




200c 
200* 
500c 
200^ 


100c 
200f 
200c 
200t 
200c 


4U0c 
600 f 
600c 
400< 
100c 
300f 
200c 
200^ 
300c 
200^ 
300c 
lOOf 
200c 
200f 
300c 
400t 
200c 
100* 
200c 
300f 
600c 
2 000< 
900c 
800* 
800c 
700* 


SCO* 

1 bOO* 

800c 

300* 

' SOOc 
900* 
200c 
80<)* 
500c 
700* 
SOOc 

8noc 

SOOc 


300c 

200* 


400* 


400* 
400c 
400* 
100c 

■ ■ ■ " 

900* 

1000* 



300* 

1000* 

200* 

200* 

600* 

200* 

SOOc 

800* 



.300c 


2 500* 
2 100c 
1 600* 
1900c 
700* 

1 .500c 
1400* 
1500c 

700* 

1300c 

600* 

2 700c 
400* 


1 m'c 

' m'c 
1 Vob*' 

' 700*" 



300* 

■ 4bbc 

■ 4bbc 



TESTS OF COMPRESSION PIECES 



219 



TABLE ^—(Continued). 
Column No. 2a, Test No. 11. 



Lattice 


Front Side. 


Back Side. 


bar. 


Under. 


Over. 


Under. 


Over. 


1 

2 
3 
4 
5 
6 
7 
8 


700c 

100c 
3 000f 

100c 
3 000/ 

200/ 
3 700/ 



700c 
100/ 


100c 


200c 


800c 


1 300c 
200/ 

1 000/ 
800c 
800c 


200/ 
800c 


100c 
200/ 

200c 
300c 
300c 
200c 
300c 



Column No. 2a, Test No. 12. 



1 


1500c 


2100/ 


300c 


200c 


2 


200c 


400/ 


2 300/ 


200c 


3 


800/ 


200/ 


2 750/ 


400c 


4 


700c 


500/ 


2 750/ 


300c 


5 


1 600/ 


200/ 


2 750/ 


300c 


Q 








2 800/ 


800c 


7 


1 400/ 


800c 


4 100/ 


300c 


8 


100/ 


800c 


300/ 


800c 



Column No. 2a, Test No. 13. 



1 


1000c 


700/ 


800/ 


500c 


2 


.^lOOc 


700/ 


2 650/ 


4G0c 


3 


700c 


900/ 


2 750/ 


300c 


4 


500c 


1000/ 


2100/ 


200/ 


5 


800c 


600/ 


3150/ 


1000c 


6 


1 000c 


700/ 


4 100/ 


700c 


7 


700c 


900/ 


1 500/ 


iOOc 


8 


leooc 


200c 


2 000/ 


1400c 



through the center of the bearing blocks. This transverse component 
was then subtracted from the amount of shear which had been calcu- 
lated from the deformation of the lattice bars as before noted, and 
the remainder has been tabulated under the heading "Transverse shear 
in column due to nominal central load." 

Tests to Failure. — After the wrought-iron bridge posts had been 
tested for stress distribution under working loads, they were loaded to 
failure. Deformations were measured in the flange members of that 
part of the column on which the previous test had given the heaviest 
stress. Table 7 gives the result of the tests to failure. For all the 
tests of wrought-iron bridge posts, whether loaded centrally or eccen- 
trically, the failures were very gradual. Final failure occurred near 
the middle or at the end. In the former case, high stresses in one 
channel had been shown by the deformation measurements at working 



220 



TESTS or COMPRESSION PIECES 



COLUMN NO. 2 TEST NO. 6 

NORTH CHANNEL Top„ SOUTH CHANNFL 

Outer Center o£ InnerrLllnner Center of Outer 
tt Gravity Fiber ^ Fiber Gravity F.ber 



UPPER CHORD, L^gUi. OF WHITE 
HEATH BRIDGE 




= 113] 1 I 
=» •= °" Bottom ^ 

East Side roi UMN NO. 4 TEST NO. 8 

WestSide CULUMIN '^'^^J__^ gOUTH CHANNEL 

NORTH CHANNEL p!^^ j^„ Center of Outer 

""1'ber ^^GrlTit"/ Xr t^d! Fiber Gravity^ Fiber 




-East Side 

-West Side Bottom 

WEST ELEVATION 
Fig. 8. 



TESTS OP COMPRESSION PIECES 



221 



s == s 

gain- 

« .- Cr^ 

2 83 (t 

CD O 

— O 

0* 






£5 



^ o o 

S c^n »3 2 o 



p,o-a 01 " o 



H,d o 



go. 

£ m „ tK B 

C S 0_ O CO 



; T g o rs 3 c 



S3' 



•B § B 1 
Or<5 g,*S m 



^ _3 & 



-£-5 » 



3P -1 
? D B 

o 



mn^ 



'•o' 



o K ft 

O »! 

o> &m 









OB 3 
P-= B 
2,3 B 

»r- (Ti B 



W £- 



-JQ0 00 5O 



OOOO 

o o o o 

OO — 1-' 

COtDOO 



2 o 






(6 O'M'^ B c 

aj§§ p. 



B ^ 









O 

txi 

M 
W 

<: 

D 
< 



Oi03 ^ CD CD 



03 



O 



a°r.f83B 



t?4 






OOOO 



en ce CO 






o 

(I 

B i-iKI 



td 



OOOO 
OOOO 

ocjooSco 



2 



o 

o 



tz^ 



OOOO o s« 
OOOO b g 



10 



O 



g^H:^q5-p5§' 



232 



TESTS OF COMPRESSION PIECES 



loads. In the latter case, a bending in one channel at working loads 
was noted by the instruments at the panel nearest the end of the 
column. In two of the three columns in which failure took place in 
the end of the column, as the instruments did not show over-stress 
in the laced portion of the column, the injured ends were removed, new 
end connections were put on, and the columns were retested as Nos. 
2a and 4a. 

TABLE 7. — Eesults of Tests to Failure, 



Column Number.. 


2 


3 


4 


5 


2a 


4a 


Test number 


6 


7 


8 


9 


13 


10 


Method of loading 


Central. 


Central. 


Central. 


Central. 


Oblique arm. 


Central. 


Load at failure, in 










2 in. (Fig. 5). 




pounds 


466 000 


480 000 


452 000 


492 000 


475 000 


475 000 


Average stress at 




failure, in pounds^ 














per square inch.. 


26 400 


27 200 


25 700 


27 900 


20 900 


26 900 


Average stress a' 














failure in tests 














of short pieces of 














flange members 














in pounds pei 














inch 


36 800 


86 800 


m 8(X) 


36 800 


36 800 


36 800 


Percentage of ex 














cess of ultimate 














strength of short 














pieces over 














column strength. 


39.5 


35.5 


43.5 


32.0 


37.0 


37.0 


Load at first sign 














of yielding, in 














pounds 


317 000 


264 000 


256 000 


264 000 






Average stress at 




first sign of yield- 














ing, in pounds 














per square inch . 


18 000 


15 000 


14 500 


15 000 






Method of failure. 














and remarks 


End 


End 


End 


Binved at 


Buckled near 


Bowed at 




failed. 


failed. 


failed. 


middle in 
plane per 
pendicular 
to lacing. 


bottom in 
plane paral- 
lel to lacine. 


middle in 
plane per- 
pendicular 
to lacing. 



Failure hy Budding of Lattice Bars. — In one of the tests of 
Column No. 1, the lattice bars failed by buckling suddenly and with- 
out warning. As tested, the column was fitted with the light lattice 
bars (1 by | in.) riveted in place. The test had in view the trial for 
stress distribution under a slight obliquity, which was not carefully 
determined. No measuring instruments were in place. A preliminary 
load was being applied. Wlien the load reached 150 000 lb. (8 060 lb. 
per sq. in. of cross-section), the alternate lattice bars in the upper 
half of the column buckled. A failure of this kind was quite un- 
expected at such a low load. Although an observer was watching the 
column, the failure was so sudden that he was unable to follow the 
movement of the parts. In this respect it was quite in contrast to the 



TESTS OF COMPRESSION PIECES 



233 



failure of the other columns. The machine was at once stopped. 
Little damage was done to the column, except to the lacing bars. 
The webs were easily straightened, new lacing bars put on, and the 
column was used in another test. 

Tests to destruction under compression had previously been made 
on lattice bars like those used in this column, and the results, in 
the absence of other data, may be useful in estimating the load carried 
by the lattice bars at failure. Under conditions of loading similar 
to the conditions found in column lattice bars, these sample bars 
failed under an average load of 2 100 lb. Assuming that the bar in 
this column which first failed was carrying 2 100 lb. when failure 
occurred, the transverse shear in the column may be computed. The 
following is a tabulated statement of the results of this test: 



Column. 


Compressive 
load, in 
pounds. 


Lacing. 


Manner 

of 
loading. 


Probable 
maximum 

load on 
lattice bar, 
in pounds. 


Correspond- 
ine shear 
in column, 
in pounds. 


Ratio of 
transverse 

shear to 

compression 

load. 


No. 1 


150 000 -j 


Single 63° 30' 

IbyM, 
riveted. 


"Very 

slight 

obliquity. 


t 2100 


3 760 


0.0251 



The column, now riveted up with the heavier lacing bars (1^ by 
^j. in.), was loaded obliquely, as shown in Fig. 5, with 300 000 lb. 
(15 900 lb. per sq. in. of cross-section), and the stresses in the lacing 
bars were measured. No sign of failure was apparent, and, from the 
data given in a succeeding paragraph, it appears that under this load 
the lattice bars must have been stressed to about three-quarters of their 
ultimate strength in compression. 

Cross-Bending Tests of CoZwmns.— Cross-bending tests were made 
on one of the wrought-iron columns and on Column No. 1. The tests 
were made in an Olsen 200 000-lb. testing machine fitted for testing 
beams 20 ft. long. The columns were supported at the ends and 
loaded at the center with a light transverse load. The column was 
placed first with the plane of the lacing perpendicular to the load, and 
then with the plane of the latticing parallel to the load. The lattice 
bars used in the tests of Column No. 1 were IJ by xe in. in cross- 
section ; in one test they were bolted in place and in another they were 
riveted. The deflection at various points along the beam was measured 



23-L 



TESTS OF COMPRESSION PIECES 



NORTH CHANNEL 
Outer Center of Inner 

Fiber Gravity Fiber 



COLUMN NO. 4 A. TEST NO. 10 

SOUTH CHANNEL 
Inner Center of Outer 

Gravity Fiber 

ill 

S 2! S S 




COLUMN NO. 2 A. TEST NO. 11 

wrvHTH rHANNEL SOUTH CHANNEL 

Ou.er"°''\"nlfor'Snner \]^^^ Inner C.nter of Outer 

Fiber Gravity Fiber l^h^TZ Fiber Gravity Fiber 




S 5J=° S " S rn 

EaslSWc UL lU 

West Side Bottom 

Pig. 9. 



TESTS OF COMPRESSION TIECES 



225 



COLUMN N0.2il TEST N0.12 

NORTH CHANNEL |i^" SOUTH CHANNEL 

Outer Center of Inner ..I Inner Center of Outer 

Fiber Gravity^ _ Fiber yL^BJU F'ber Gravity Fiber 

D o o o o o 




T ' -/T- A-^ 


±2 X ^ zk 


\' -^v ^:j 


-1/ -J-^ ' ^v 


^^ ^ift ' r. 


. -5 t -- 


7^ 7- , 


^ ^ 


t^ V C. 


^t . ::. "^ % 


i^ 4 ' ' ^/ ^ 








t J t' 


^ J.^ X- 


4 t \ 


t^i %. 


HI ,7^ E 


U- (i U 


^ ^ % 


. IS ^ j>. 


''t H \ 


-^ ' ' t ^- 


i7 Jv^ ^'' 


, 'C '^ -.< 


'~1\ llv ^ 


^ -V X 


^ 1 iN ^^ 


T-nii \ -^^^v 



g I 8 g g g 1 

^SS o»^^o.^^^ fgotton 

East Side >^|t 

West Side II 



COLUMN N0.2^TEST N0.13 

NORTH CHANNEL |l| SOUTH CHANNEL 

Outer Center of Inner y>4 Inner Center of Outer 
Fiber Gravity Fiber ixza^lD Fiber Gravity Fiber 













g 

o 


00 


o 


§ 




00 


O 


F 


t 

X 

/5\ 

X 
X 


] o 

s 


g 


l§ 




1 


g 

o 


§ 




















1 




1 ^ 


\ 








' ^/ 






]j 


^ \ 
























\ 


1 / 


\ 




\i \ 








1 






i' 


\ 


^ 






















\ 


1 


\ 




\ ' 






i 












T 






















\ 


1 1 


i\ 


1 




















u 
























1 






1 


1 
















\ 




/ 
























1 








1 










', , 






\ 




' 
























1 








1 




^ 






\'\ 






\- 


I 


- 




— 


















1 








ll 




\ 






V 






















/ 




1 






'1 
















p\ 


























1 












/ ' 










\ 






















,' li 




1 










f' 


1 












~^\_ 






























1 










I 














\ 




























1 


1 




1 


















\ 


/ 






























\ \ 










^ 


^ 


\ 






\ 




ll 




























' 


\\ 












^, 


\ 








\ 






























;, 






1 








^, 




\ 






\ 


// 




























J 


h 


ll 










- 


\ 






, \ 






















, 


\ 




1 


h 


ll 


! 








1 1 


>i 






















T 




/ 










1 










1 




1 1 


, \ 
























\ 




















^1 








', \ 


A 






















\ 




















1 












f^ 


— 




















\ 




1 






1 


































I 












4- 


w 


~ 


1 


/ 












^ 




r 
















1 




\ 










1 


1 








T 
















/ 




/ 


1""" 






t 


/_ 


1 


; 










1 




^ 














/ 






' 






\ 








1 


x_ 


/ 














/' 
















/ 








\ 








I 


' 




















/ 






\i 


A 


1 


\ 










\ 






1 1 


,/ 






1 1 1 i 1 1 1 


i 1 §§ § 1 1 1 




^ 




^ 




J 








- 


— 


- 


_-\ 


►Vest Side 


l' 







Pig. 10. 



226 TESTS OF COMPRESSION PIECES 

with Ames test gauges, and the actual curve assumed by the column 

under transverse load was thus determined. The theoretical elastic 

curve was computed from the common theory of flexure, not counting 

the lattice bars in the calculation of the moment of inertia. Fig. 2, 

Plate XII, shows the deflectometers and extensometers on Column No. 1 

under the cross-bending test. Fig. 14 shows the deflection curves given 

by the column under transverse load and also the computed elastic 

curves. 

It will be noted that when tested with the lacing vertical, Column 

No. 1 shows much greater deflection than that computed from the usual 

beam formula, while the stiffer wrought-iron column shows a much 

closer agreement with the curve, the heavy lacing apparently adding 

stinHGSS 

Field Tests of Columns. 

Description of Bridge. — The bridge on which the field tests were 
made is an eight-panel single-track Pratt truss which spans the 
Sangamon River near White Heath, Illinois. The bridge is on the 
line of the Illinois Central Railroad between Champaign and Clinton, 
Illinois. The span measured 158 ft. 6 in. Fig. 11 gives a diagram of 
the bridge and Fig. 3, Plate XII, is from a photograph of the bridge. 

Loads. — The test load applied to the bridge consisted of a mogul 
locomotive and tender followed by a loaded coal car and a caboose. 
Fig. 11 shows the test train, with dimensions and weights. This train 
was furnished through the courtesy of the railroad. 

Memhers Investigated. — The members studied for stress distribu- 
tion were Posts U^L^ South, U^L^ South, U^L^ North, and upper chord 
U^U^ South, The bridge diagram, Fig. 11, shows the location of the 
members. The posts were made up of two steel channels, double laced, 
while the top chord was made up of two built-up channels with a cover- 
plate on top and lacing across the bottom. The cross-section of the 
members tested is shown in Fig. 1. Fig. 12 shows the post and 
the chord. 

Measurement of Deformation. — Ames test gauges were used as 
extensometers, and the method of attachment was the same as in the 
laboratory tests of columns. The method of reduction of instrument 
readings to stresses at the extreme fibers of members was also the same. 

Routine of Tests. — As in the laboratory tests, the stress distribution 
in the channels and the lattice bars was studied. The method of 



TESTS OF COMPKESSION PIECES 



227 



DIAGRAM OF WHITE HEATH BRIDGE, 
tr, U^ Ui Us U., U, 




Pig. 11. 



POST AND UPPER CHORD OF WHITE HEATH BRIDGE 
ICov. ri. Sl'x'/i'a x2lV' 

-l'2-> 



o o oo o o ooVo'-olS a Bio . 
foil O/Kolo • 

[ojo ©Wolo • 

o o o o o o o o|o„c-|o„4^e'r5' • 




338 



TESTS OF COMPRESSION PIECES 



procedure was as follows. Instruments were placed on some portion 
of the column to measure the deformation over a short gauge length, 
and a reading was taken. The test train was then run upon the bridge 
to a given position (one approximating the maximum load on the 
member under test), and the instruments were read again. The train 
was then run off the bridge, and the instruments were again read. 
This procedure was repeated several times, at least three applications 
of the load being made and frequently several more. The instruments 
were then moved to another part of the column, and that part was 
tested. Observations were made on both flange members and lattice 
bars. The tests covered a period of eight days. 



TABLE 8. — Stresses in Posts of White Heath Bridge. 



North Channel. 


South Channel. 


East side. 


West side. 


East side. 


West side. 


i 


ll 


^60 


5« 












§1 

5« 




a ss 

6^ 


P 



Post U^L^ South. 



1* 


5 000 


5 300 


5 400 


2 600 


3 700 


2 700 


4 900 


5 000 


5 000 


2 300 


2 500 


2 600 


2 


3 400 


3 500 


3500 


1900 


1600 


1500 


3 300 


3 400 


3 400 


1600 


1800 


1800 


3 


3 500 


3 800 


4 000 


1 800 


1900 


2 000 


2 600 


2 900 


3 100 


2 600 


2 200 


2 100 


4 


3 700 


3 400 


3 300 


2 800 


3 000 


3 000 


1 600 


3 100 


3 600 


2 400 


2 800 


2 9C0 


5 


3 500 


3 900 


4 000 


2 900 


2 900 


2 900 


2 800 


3 000 


3 100 


2 200 


2 600 


2 800 


6 


3 600 


3 600 


3 600 


3 400 


3 800 


3 900 


3 400 


3 400 


3 400 


4 100 


3 400 


3 200 


7 


2 600 


3 200 


3 400 


3 300 


3 100 


3 000 


2 400 


3 000 


3 300 


2 600 


2 700 


2 800 


8 


3 300 


3 200 


3 200 


3 900 


3 600 


3 500 


3 600 


3 400 


3 300 


3 700 


3 900 


3 900 


9 


3 200 


3 100 


3 000 


3 600 


3 300 


3 200 


1 700 


2 500 


2 800 


2 900 


3 300 


3 400 


10 


2 800 


2 700 


2 700 


3 100 


3 300 


3 300 


2 700 


2 400 


2 300 


3 300 


3 700 


3 900 


11 


8 300 


2900 


2 800 


4 500 


8 400 


3 100 


2 900 


2 000 


1700 


3 500 


3 900 


4 100 


12 


1400 


2 000 


2 300 


3 100 


3 200 


3 200 


1800 


1800 


1 800 


4 800 


4 700 


4 600 


13 


1300 


1500 


1.500 


2 800 


2 700 


2 600 


2 600 


2 900 


3 100 


5 200 


4 900 


4 800 



Post TJ^L^ North. 



1 


4 900 


4 900 


4 900 


2 600 


2 500 


2 500 


4 800 


4 900 


4 900 


2 100 


2 700 


2 800 


^H 


4 800 


4 700 


4 700 


2 900 


2 800 


2 700 


8100 


4 900 


5 700 


900 


1 500 


1 700 


9. ■ 


4 800 


4 400 


4 800 


2 600 


2 500 


2 500 


4900 


4 700 


4 600 


2 700 


1700 


1 300 


2H 


4 000 


3 900 


3 900 


2 700 


2 800 


2 900 


4 300 


3 700 


3 500 


2 300 


1800 


1600 


3 


3 900 
3 900 


4 200 
3 900 


4 300 
4000 


2 800 
2 400 


2 900 
2 500 


2 900 
2 500 














3^ 


3 200 


3 800 


4100 


2 400 


2 500 


2500 


4 


3 400 


4 600 


5000 


2 700 


2 700 


2 700 


3 900 


3 700 


8 700 


8 500 


3 200 


3 100 


f" 


3 800 


3 400 


3 200 


2 500 


2 300 


2 200 


3 500 


3 500 


3 400 


3 100 


3 000 


3 000 


3 900 


3 700 


3 600 


3 700 


3 700 


3 700 


3 100 


3 600 


3 800 


1400 


2 100 


2 400 


5% 


4 ICO 


3 600 


3 400 


4 000 


3 400 


3 200 


3 500 


3 900 


4 100 


3 500 


2 900 


2 800 


fi^ 


3 600 


8 200 


3 100 


3 4«) 


3 300 


3 200 


8 700 


3 500 


3 400 


3 900 


3 600 


3 500 


6^ 


3 300 


3 300 


3 300 


3 600 


3 500 


3 500 


4 500 


4 700 


4 800 


2 000 


2 400 


2 600 



* No explanation for the high values in Panel 1 has been found. Five determinations of 
stress were made, including the removal and re-attachment of instruments. 



TESTS OF COMPRESSION PIECES 
POST L^Ls NORTH 

NORTH CHANNEL ^ ^ I.ner^°^'''' ''''^''''^'- Outer 

8 8 § Center of 8 S if Jf S § i Center of S S § 
1 ! S Gravity o S S llkdilU S 2 1 Gravity % % Z 



239 



l-U-= -§ JfJ-i-- 4/ — 


^ -<^r" - :^___ 


— j V— 


I 1 — 1 — * — ' — 


"t/r:";i;^+v^ 


::4"ii iti-f? 


— fA f 1 




-- ^-V-T^-y— T--'-^ 


::i:;:=ii:^:ii:t4i 


fflfflNWWIfl 



8 i 



N Z^ S' S^-S4 .4LZ- 


V ^t i/^i i ^ "(< 


7t r^ 


^^ 4 


7^7 ^• 


7 ''^ 


a1 ?:: „ ^, :: 




r " 1 " r 


; 1 


^3 "■ JJ 


^ ^ 1 it 


T t /__ t\ 


' - ^4 


-; -J ^ - n 


^44- t f 


iT ^ 4 ^ A^ X 


> "J" S 4 V 1 A 


I ^.- ^ 4 v^ / 


4 I 4 il Z Z 


w / ^^ i: \ ^ 


, ±A _.^ A ± ^^ ^ 



. S S o 

-Etist Side 
-West Side 



Bottom 

NORTH CHANNEL ^^^'^ tlg^s SOUTH g^^^^ CHANNEL ^ ^ 
Outer Inner Inner Outer 

Fiber Fiber H ^°P I Fiber Fiber 

5 S Center of S § ol i i i Center of 

M ^ as Gravity o s<! ^ 




« to Bottom 

Fiber Stress 

East Side 

West Side 

Fig. 13. 



Fiber SU-ess 



230 



TESTS OF COMPRESSION PIECES 



Results of Tests for Stress Distribution in Channels. — Table 8 
gives the results of the tests to determine the stress distribution and 
variation in the channels of the bridge posts, and Table 9 gives those 
for the top chord. The stresses given are calculated from the observed 
deformations, using a modulus of elasticity of 30 000 000 lb. per sq. in. 
The conditions of measurement of deformation were much the same 
as in the laboratory tests. The stress noted is the average stress over 
a space of 4^ in. on either side of the point indicated. 

TABLE 9. — Stresses in Upper Chord of White Heath Bridge. 



Lower Side (Laced). 



Upper Side (Cover-plate) . 



North Channel. 



"3 


Outer 


Inner 


* 


fiber. 


fiber. 


Oi 






1 


4 300 


8 900 


a 


4 900 


4 300 


3 


2800 


5500 


4 


2 800 


5 100 


5 


5 700 


5 400 


6 


5 500 


4 800 


7 


5 '.toe 


4 700 


8 


6 200 


5 300 


9 


6 000 


5 300 


10 


5 900 


5100 



South 
Channel. 



Outer 
fiber. 



5 000 
4500 

3 400 

4 200 
3 900 

5 1(10 
3 400 

3 300 

4 700 
4 400 



Inner 
fiber. 



5 500 
4800 

4 500 

5 300 
4 800 
4 200 

3 600 

4 400 

3 800 

4 900 



Distance 
from end, 
in inches. 



33 
57 

81 
105 
129 
153 
177 



North 
edge. 



5 500 



5 200 
5 000 

4 800 

5 400 
5 000 



Over. 

north web 

plate. 



6 000 



5 40O 

6 100 
5 900 
5 700 
5 000 



Over 

south web 

plate. 



5200 
5 200 
5 900 
5 700 
5900 
5 500 
5 700 



South 
edge. 



5 700 
5 500 

5 30O 

6 000 
5 900 
5 600 
5 600 



TABLE 10. — Maximum Observed Fiber Stress in Flange Members 
OF Columns in White Heath Bridge. 



Column number 

Test number 

Lacing 

Maximum observed compressive stress 
in an extreme fiber, in pounds per 
square inch 

Percentage of excess maximum observed 
stress over average, highest five values. 



ZJsLaSouth 

F 1 

Double, 45", 

riveted at 

crossing. 



5 200 
73* 
63 
60 
57 
47 



ZJaLsNorth 

F 2 

Double, 45°, 

riveted at 

crossing 



5 700 
64 
48 
41 
31 
31 



173^74 South 

F 3 

Cover -plate on top. 

Double, 45" on 

bottom. 



6 200 
20 
19 
17 
17 
17 



* No values from Panel 1 have been included in this table, as no explanation is known 
of the high stresses indicated in that panel. 

Figs. 8 and 13 show graphically the stress distribution and varia- 
tion. In these figures the full lines give the stresses at the west side 
(front), and the dotted lines the stresses at the east side (back). 

In Table 10 are given a number of the highest observed fiber 
stresses. The excess of the maximum fiber stress is given as a per- 



TESTS OF COMPRESSION PIECES 



231 



Deflection, in Inches under Central Load of 1000 lb. 



o 


o 2 


o 














^ 


-^ 










/ 


o 

8 


























^ ' 










/ 






,/ 






/ 










/ 


K 


1 

1 










/ 










/ 




,^ 


/ 








A 


V 




1 


f 










/ 




o 




// 


S 


/ 


/ 


^ 








y 






1 
I 




f 






/ 




t 


o 




// 

1 




// 




<; 




^ 










1 
1 




t3 


— 




/ 










\ 

1 


\\ 




S 














1 






\1 




3 


^3 




\ 


£. 


\ 


s\ 


p. 








s\ 


H. 


\ 




E- 






\ 




£- 


h*' 




^ 


» 




\ 


S 








^^r^z 
















m 






\ 






\ 










■*'■ 


^ 


^ 


\ 






































f 


'e'>S 


^^ 












\ 


o 


Culuni 


u No.4 
















^ 


\ 










\ 


° 














Cc 


)lur 


an 


No. 


1 












S 



& o 

M m o 




p ^ 






5 o S 























_ 








^--r 


-5- 


^ 


s. 




M 




— 




r^ 










.,._ 


cc 


_. 


^ 


Avera 


sre Stress 




a 








































r 


1 


m 




IP 








■ 






^^ 




m 


m 


^ 


i 
















Li 












a cr 








_^ 


•-.^ 
















*^L ! \ 








_,' 


-— 


M 




-^.N 




^ 


k-^ 








'^•^ 


^■ 


— - 


"""/ 






C 








<. 


^^ 








233 TESTS OF COMPRESSION PIECES 

centage of the average stress. At most sections the maximum stress 
was in the outer fiber of the channel, but in some cases it was found 
at the inner fiber. 

In the tests of the bridge posts an attempt was made to determine 
the stresses in a few of the lattice bars. These stresses were very small, 
and the precision of the extensometer was not sufficient to measure 
them with any great degree of accuracy. It should be noted that the 
lacing of the posts in this bridge was double, and the bars were riveted 
together at their intersection. In several cases it was found that a 
lattice bar under load bent in the shape of a very flat S curve, the 
point of attachment to another lacing bar, at the middle, being a point 
of inflection. 

Special Tests on Bridge Columns. — Tests were made on the batten 
plates at the top of one of the posts, and under load a slight bending 
of the plates between channels was found. The bending took place in 
a horizontal plane. 

In one post the change of stress was observed as the locomotive 
and train moved across the bridge. Extensometers were placed at 
aa, hh (Fig, 12), and on the floor beams, and the changes in reading 
were noted as the train moved across the bridge. In the inner channel 
of the post, tension was set up as the locomotive came opposite the post. 

Tests of Lattice Bars^ Small Columns, and Column Material. 

Compression Tests of Lattice Bars. — Many of the lattice bars in a 
column, as they transmit stress from one flange member of the 
column to the other, are under compression. To study the action of 
lattice bars under compression, a series of tests on single lattice bars 
was made. Fig. 15 shows the arrangement of the apparatus. The 
lattice bar was tightly bolted to the blocks, B^ and B^. The upper 
block, B^, was fastened to the cross-head of a testing machine, and the 
block, B^, was pressed against the weighing table of the testing 
machine. A spherical-seated bearing block was used, to insure an even 
bearing. Ames test gauges, E, mounted on suitable frames, were 
attached to the lattice bar over a short gauged length. From the read- 
ings of these gauges, the deformation of the extreme fiber of the bar 
was computed. 

In this test of lattice bars, the load was applied with an eccentricity 
approaching that to be expected in a column for the lattice bars which 



TESTS OF COMPRESSION PIECES 
Average Stress, in Thousands of Pounds per Square Inch. 



233 



8 


* 


33 It- Ki O CO C5 *. ! 


-« c 




















f 






























<^ 


/ 


fl 


1 












c^^ 


4iy 




7 












A 




/ 














# 


/// 




i 










X 






/ 




sin 








/ 






//// 


i 










X 








V \ 










/" 










1 \ 




















\ 
















1 




\ 






CO o 












II 


\ 






CO o 














\ 






s 

3 




















!2! 

p 




























7 

' 1 












S 








r/^ 


















^r 


*/ 




1 














A// 


I "^ 


/^' 














1 


/ 




p' 












/ 


7 


/ 


/ 














/ 


/ 1\ 


/ 














/ 


^ 1 


/ 















o 

-n 



IS 
1^ 

Is 

Ul l/J 

i" m 

cr Z 
o 

?o O 

g" CO 
g CO 

B. o 

05 I 

•-! ^ 

o^ CO 

i CO 

CD 

& o 



> 

H 
H 

o 
m 

CD 

> 

CO 



234 TESTS OF COMPRESSION PIECES 

are next to the flange member (here designated "under" bars). The 
lattice bars outside of these "under" bars are here designated "over" 
bars. The stress distribution across the section of the "over" bars, 
which are under compression, is probably more uneven than the stress 
distribution found in these tests. However, these tests give some idea 
of the relative behavior of lattice bars of various proportions, and of 
the large eccentricity of loading of all lattice bars. 

Lattice bars of the following cross-sections were tested: Flat bars 
li by i in., 1 by f in., I by tV in. ; angles IJ by li by J in. ; channels 
li by I by J in. Several channel and angle lattice bars were tested 
with ends flattened and ribs turned inward, to minimize the eccentricity 
of loading. Bars of the following lengths between centers of rivet 
holes were tested: 8J in., 13^ in., and 20 in. The rivet holes were 
i in. in diameter. All bars were tested in a Philadelphia Machine 
Tool Company's 100 000-lb. testing machine, and loads and extensom- 
eter readings were taken to failure. 

Observations were also made on the behavior of a lattice bar in a 
column under load, with a view to determine the distribution of stress 
over the section. Column No. 1 was loaded obliquely. The instru- 
ments were placed on an "over" bar which had been found to carry 
a high compressive stress, and readings were taken to determine the 
distribution of stress across the section. 

When Column No. 1 was under cross-bending test, observations 
were made to determine the stresses transmitted by lattice bars and 
their distribution over the section of the bars. Extensometers were 
placed successively on most bars under compression on one-half of the 
column, and on some bars which were under tension. In both of these 
tests the bars were IJ by yg in., and were riveted. 

Results of Tests of Lattice Bars. — The results of the tests of 
single lattice bars are given in Figs. 17 and 18, and in Tables 11 and 
12. Fig. 17 shows the ratio of maximum to average stress in the bars 
13i in. long between centers of rivet holes. It also gives the result 
of the test of stress distribution in a lattice bar of Column No. 1. 
Table 11 gives the stresses at failure of the various bars tested singly. 
The average stress on the various bars, which corresponds to a maxi- 
mum fiber stress of 40 000 lb. per sq. in., as taken from these tests, 
has been noted and is given in Table 12. The results of the tests to 
failure are shown graphically in Fig. 18. The angle and channel bars 



TESTS OF COMPRESSION PIECES 



235 



tested with flattened ends and ribs turned inward failed in the flattened 
part at loads no greater than the bars fastened with ribs turned 
outward. 

TABLE 11. — Compression Tests op Lattice Bars. 
Average of two specimens. 




♦The values in this column were obtained by multiplying the observed stress 

** -I V 40 000 

at fadure by . , , . , . ^ -. — ^r-j r — —• 

yield point determined from tests 

TABLE 12. — Average Stress in Lattice Bars Which Corresponds to 
A Maximum Fiber Stress of 40 000 lb. per sq. in. 



Section of bar, 
in inches. 



; by x^, flat 

1 by i, flat 

U by A, flat 

H by Uby J, angle .. 
IJ by j by |, channel 



Distance 


Corresponding average 


from center to center 


stress, in 


of rivet holes, 


pounds per square 


in inches. 


inch. 


13^ 


11600 


13| 


14 000 


13| 


14 900 


i.sl 


15 900 


13i 


17 500 



Table 13 gives the results of the test for stress distribution in the 
lattice bars of Column No. 1 as it was stressed in cross-bending. 

Tests of Small Columns. — Tests of two small compression pieces 
were made in order to study the effect of slight bends and kinks in the 
column upon the distribution of stress. The deviation from a straight 
line, in these nominally straight pieces, was measured before the load 
was applied. The deformations on two opposite faces for a given load 
were measured. The extensometer was similar to that used on the 



236 



TESTS OF COMPEESSION PIECES 



single lattice bar tests. The instrument was shifted from one position 
to another along the column. The columns were finally loaded to 
failure. One of the columns was a flat piece of steel, 3 by 0.72 in. in 
cross-section, and 40 in. long. It was held at the upper end by wedge 
grips in the cross-head of the machine and at its lower end rested on 
a spherical-seated block. The second compression piece was a 4-in. 
channel 40 in. long. The ends were planed square; the upper end bore 
on a flat compression plate in the iron head of the machine, and the 
lower end rested on a spherical-seated block. 

TABLE 13. — Stress in Lattice Bars of Column No. 1 under Cross 

Bending. 

Column tested as a beam centrally loaded over span of 19 ft. 8 in.; 
lattice bars 1^ by ^ in. ; 17.89 in. center to center of rivet holes ; 
rivets J in. in diameter. 



Bar. 



Maximum fiber stress, 
from Ames dials. 



Averajre stress, from 
EwinK extensometer. 



Ratio, 
maximum to average. 



"Over" Bars in Compression. 



UE 


6000 


2 600 


2.31 


16J57 


8 300 


2 000 


4.15 


IBE 


4.S00 


1800 


2.39 


WE 


7 700 


1 400 


5.50 


22E 


7000 


2 900 


2.42 


24E 


8 900 


2 200 


4.04 




Average, 3.47 



"Under" Bars. 



15 TF 
25B 
17E 
25W 
19 fV 



4 300c 
3 800^ 

3 QUOt 
8 900c 

4 500c 



2 500c 

3 100f 
3 000? 
2 800c 
2 900c 



1.72 
1.23 
1.30 
1.39 
1.55 



Average, 1.44 



"Over'' Bars in Tension. 



16TF 


7 000 


3 000 


2.33 


18Tr 


7 000 


2 200 


3.19 


20 TF 


4 500 


2 400 


1.87 


22 IT 


5 700 


2 200 


2.59 


24 PT 


7 000 


2 700 


2.59 


26 W 


3 800 


2 200 


1.73 




Average, 2.38 



TESTS OF COMPRESSION PIECES 



237 



Fig. 16 gives the results of tests of the small columns. The 
dotted line represents the maximum fiber stress computed by consider- 
ing the eccentricity of loading at any cross-section to be equal to the 
deviation of that section from a straight line connecting the ends of the 
column. The deflections were slight, and were neglected in the 
calculations. The solid line represents the stresses on the two sides 
of the column, as determined from the extensometer readings. 
TABLE 14. — Tension Tests of Material from Columns. 



Test piece from: 


Material. 


Number of 
te.st pieces. 


Average stress 

at yield point, 

in pounds per 

square inch. 


Average stress 
at ultimate, 
in pounds per 
square inch. 


Average 
Elonga- 
tion. 
Percent- 
age. 


Column No. 1 
Angles 


Steel. 

Steel. 

Steel. 
i Wrought ( 
1 iron, f 


2 

2 

1 
3 


43 300 

41) 700 

43 400 
30 700 


61 600 

58 100 

62 000 
46 800 


37 


/g by Xl-in. 
lacing bar 


42 


1 by |-in. 
lacmg bar 


38 


Channels of wrought- 
iron columns 


17 



Tests of Column Material. — Table 14 gives the results of the tensile 
tests of samples of material from various parts of the flange members 
of the columns, and also of tension tests of lattice bars. Table 15 
gives tension tests of lattice bars like those used in the compression 
tests of lattice bars. 

TABLE 15. — Tension Tests op Lattice Bars. 



Shape. 


Dimensions of 

cross-section 

of whole bar, 

in inches. 


Number of 

specimens 

tested. 


Average stress 

at yield point, in 

pounds per 

square inch. 


Average stress 
at ultimate, 

in pounds per 
square inch. 


Percentage 

of elongation 

in 2 in. 


Channel. 
Angle.. . . 

Flat 

Flat 

Flat 


libyiby J 

U by U by i 

R by /g 

Ibyg 

n by i 


1 

I 

2 
2 


43 000 
45 000 
38 700 
42 400 

44 600 


57 600 
50 300 
57 000 
61700 
60 800 


36.5 
24.5 
45.5 
45.8 
42.8 



Comparison and Discussion. 

The Action of Built-up Compression Pieces.— In analytical dis- 
cussions of column action, the stress is usually assumed to vary uni- 
formly from a minimum on one side of the cross-section to a maximum 
on the opposite side, and the whole cross-section of the column is 
considered to act as a unit. The longitudinal axis of the column is 
also considered to take a definite elastic curve under load. In the 



238 



TESTS OF COMPRESSION PIECES 



derivation of most column formulas, it is assumed that the amount of 
deflection of the elastic curve from the original position of the axis 
is an important element in fixing the maximum stress in the column. 
Although these assumptions are generally used as the basis of column 
formulas, it may be well to consider whether conditions may not exist, 
in columns of ordinary form and dimensions, which will render doubt- 
ful the general applicability of some of these assumptions, and will 
dwarf the effect of others. At any rate, it seems worth while to 
consider the effect of other conditions in a built-up member. It must 
be borne in mind that the built-up column is subject to imperfections 
of fabrication, and that some crookedness and eccentricity must exist. 

RESULTS OF COMPRESSION TESTS TO FAILURE 
OF LATTICE BARS • 



S 20 000 







.L 


ik 


xl>i 


^H 




















^ 






























"^ 


<< 


• 1 


J 


H^ 


X'^i 


i 


















K 


b„- 


•^ 


























"'•* 




v^ 


1 
























*oo 


"4" 


o 
























r 




■^ 




"V, 






























^ 














































































































1 



100 200 

( Length ) -|— ( Kadius of gyration ) 

Fig. 18. 

The component parts of the column may be relatively slender and 

flimsy. Whether there is integrity of cross-section under load, is a 

question. In the tests herein described, the amount of deflection from 

the original axis, for loads up to a point somewhat below incipient 

failure, was found to be slight (generally between 0.04 and 0.1 in.), 

much smaller than necessary to account for the stresses observed in the 

columns. 

The action of short columns at failure may be expected to be 

different from that of longer columns, although the stresses up to 

incipient failure may be the same. Granting that the conditions of 



TESTS OF COMPEESSIOlSr PIECES 239 

non-straightness are such that the distribution of stress over the cross- 
section is the same for the two lengths of column, and that the 
deflection of the column is so slight as not to affect materially the 
stresses developed, the longer column will be in more danger of im- 
mediate and sudden collapse after the yield point of the material 
in any fiber has been reached, and the total load carried before complete 
failure will, in general, be less. This is because, in a ductile material, 
after the stress at one side of the column has passed the yield point, 
the total resistance of the section to compression will increase, while 
the resistance to cross-bending may not. Under the conditions named, 
the bending moment due to eccentricity will be the same until the 
yield point in some fiber is reached. After yielding begins, the 
greater deflection in the longer column rapidly increases its relative 
eccentricity, and more rapid failure may be expected than with the 
shorter column. 

Indications of Data. — It will aid in the interpretation of the data 
of the distribution of stress over the channel members of the columns 
to point out a few simple indications. Reference may be made to the 
diagrams in Figs. 6, 7, 8, and 13, and to Tables 2, 3, 8, and 9. 

1. — Any lack of agreement between the average load stress and the 
average of the stresses given for the four centers of gravity of channel 
flanges may be ascribed to errors of observation. 

2. — If the stress at the center of gravity of one channel is above 
the average stress throughout the length of the column, and the 
corresponding stress for the other channel is similarly below the aver- 
age stress, there must be an eccentricity in the application of the load 
at the two ends. If the stresses at the center of gravity of one channel 
member forms in the diagram a straight line which crosses the line of 
average stress, and that for the other channel crosses in the opposite 
way, the eccentricity of the load application must be oblique. 

3. — If the stress at the center of gravity of a channel in near-by 
points is greater first in one channel and then in the other, the change 
may be due to crookedness of the column throughout that part of the 
length. 

4. — If, in one channel or in one channel flange, the stress at the 
center of gravity remains constant and that of the extreme fiber 
varies, the change may be due to local crookedness of this channel and 
there will be a lateral bending of this member. 



340 TESTS OF COMPRESSION PIECES 

5. — If the front side of a channel has a higher stress than the 
back side, there must be bending action through its web, and 
vice versa. 

6. — Changing stresses in the diagonally opposite corners of a chan- 
nel may indicate twisting of the channel, and another combination of 
stresses may indicate a twisting or oblique distortion of the column as 
a whole. 

An inspection of the diagrams shows that all these indications are 
found in the tests. 

Does the Built-up Column Act as a Unitf — Engineers have often 
expressed doubt as to whether the parts of a built-up column act as a 
unit, although column formulas assume this unity of action. The 
tests throw some light on the question of the integrity of cross-section 
under load. The individual channel, of course, acts as a unit to resist 
bending action, though there are indications of twisting. The in- 
tegrity of the whole section with reference to a plane parallel to the 
lacing seems probable, except as twisting action exists. With refer- 
ence to a plane through the axis perpendicular to the plane of the 
lacing, this unity of action is not so certain. The tests on the dis- 
tribution of compressive stress and likewise the cross-bending tests 
indicate that these built-up columns did not in all cases act as a unit 
but rather as two members not fully restrained by the lacing. The 
stresses in two channels at points in the same cross-section do not give 
the regularity of variation which would exist if the column bent as a 
unit. The elastic curve assumed by Column No. 1 under cross-bending 
load, shown in Fig. 2, Plate XII, differs from the computed elastic 
curve, though that for the wrought-iron column gives little difference. 
In the case of the posts of the White Heath Bridge, however, there 
is much closer agreement and a seemingly closer approach to unity 
of action. 

Effect of Non-straigJitness of Built-up Columns Upon Distribution 
of Compressive Stress. — The effect of crookedness or other irregulari- 
ties of a constituent member of -a built-up column may be realized if 
a rough analysis of the case be made. Consider a part of one of the 
channels forming a column, taking the length between the connections 
of two adjacent lattice bars. This member is under compression. 
Owing to non-straightness or to the non-homogeneity of the material, 
the load on this short piece is not evenly distributed over the section; 



TESTS OF COMPRESSION PIECES 



241 



that is, it is not centrally loaded, but may be considered to have an 
eccentricity with respect to the gravity axis. Call this eccentricity e 
(Fig. 19). Neglect any deflection of the piece under consideration 
due to the load. Call the compression load coming on this piece 
P; A its area of cross-section; I its moment of inertia about YY, and 
r the corresponding radius of gyration; and c the distance from YY 

ECCENTRICITY IN CHANNEL MEMBER 



-^Y— 




Fig. 19. 



to the remotest fiber. Then the bending moment due to the eccentricity 
is Pe. The maximum stress will be 



P c P / e cx 



The excess of the stress in the extreme fiber of the piece over the 



average stress, produced by the eccentricity, e, is then 



„ , and hence 



the term, — ^1 gives the proportionate excess of stress in the extreme 

fiber. This value is applicable to the channel, or to one flange of 
the channel, or it may be applied to the column as a whole by using 
the properties of the whole section. In the single channel under con- 
sideration, c is relatively large and r is relatively small, and the 
excess of maximum stress for a given eccentricity, e, may be expected 



242 TESTS OF COMPRESSION PIECES 

to be relatively large. It will be seen that for an excess of 50% in 
the extreme fiber of a channel of Column No. 1, e, by this formula, 
would be 0.045 in., and, in the wrought-iron columns, 0.057 in. A 
slight variation from straightness in a channel will account for con- 
siderable increase of stress. 

Excess of Maximum Fiber Stress over Average Stress in Channel 
Members. — The diagrams and data show that the compressive stress 
is unevenly distributed over the cross-section of the columns tested, 
and also that there is great variation in this distribution at the vari- 
ous sections along the length of the column. It will be noted that in 
a number of sections the excess of stress was from 40 to 50 per cent. 
In one test of Column No. 1, an excess of 67% was found, and in the 
White Heath Bridge an excess of 73 per cent. Possibly these values 
were unusual, or the observations were erratic, but the indications 
of a fiber stress of from 40 to 50% in excess of the average stress were 
not uncommon. 

It may be seen that among the causes to which the high fiber stress 
may be attributed are (a) non-straightness of the column as a whole, 
(b) non-straightness of the component channels, or eccentricity in the 
delivery of stress to them by the lacing, and (c) unknown eccentricity 
in the application of the load. It would be of interest to know how 
much of the increase of stress may be due to any one of these condi- 
tions. A study of the tests of Column No. 1 shows that generally only 
a small amount may be said to be due to non-straightness of the 
column as a whole. In but few cases is it found to be more than, 
say, 5 per cent. In four places it seems that the excess attributable to 
this may be estimated to be between 20 and 25 per cent. The efPect 
of non-straightness of the individual channels seems to be greater. 
At several points the excess of stress attributable to this cause appears 
to be from 30 to 50 per cent. As already noted, a kink in the 
channels of 0.045 in. would give, by the analysis made, an eccentricity 
sufficient for a 50% increase in stress. Not all of this crookedness 
need be between adjacent rivet points, as the stress may not reach 
normal for some distance on either side. The eSect of the third 
condition, eccentricity of application of the load, will vary with the 
construction. In Column No. 1 the effect of undetermined eccentricity 
of application of load appears to be not nearly as great as the effect of 
non-straightness of the component channels. 



TESTS OF COMPRESSION PIECES 243 

In the wrought-iron columns, which are much stockier, the lack 
of straightness in individual channels has less effect, seemingly less 
than 15%, and much the larger part of the high fiber stresses appears 
to be due to general column eccentricity or to eccentricity of loading. 

The resvilts for the posts of the White Heath Bridge are of interest 
in this respect. It is evident that the effect of non-straightness of 
channels was not very large, and also that the effect of non-straightness 
in the column as a whole was relatively small. There is, however, an 
evident bending in the direction of the web of the channels. For 
example, in UJ^r^ South, the back side of the channels has the maxi- 
mum stress at the top and the front side at the bottom. The bending 
moment producing this may be due to obliquity of end pressures or to 
a bending by the connecting floor-beam and top chord. A twisting 
action is also apparent. Post UJLi^ North gave quite similar results. 

Compressive Strength of Lattice Bars. — In the discussion of stress 
developed in column lacing, the stress considered was the average over 
the bar. As usually attached, there is considerable flexure in the bar, 
and the ability of the bar to carry this eccentric load should be con- 
sidered. The bars are most likely to fail in compression, since they 
act as long columns eccentrically loaded. 

The tests of individual lattice bars (Fig. lY and Table 12) show 
that the maximum fiber stress may be several times the average stress. 
It is also seen that even in a short lacing bar the maximum load 
carried is only about one-half the yield point of the material. The 
necessity of using very low working loads on lattice bars appears to be 
important. It will be noted that at low stresses there is similarity of 
distribution of stress in the slender bars and in the thicker bars, but 
the slender bars fail at smaller computed fiber stress. 

The results of tests to destruction of individual lattice bars are 
fairly well represented by the formula : 

P I 

-7 = 21 400 — 45 - 
A r 

where P = load at failure, in pounds, A = area of cross-section, in 
square inches, I is the distance, in inches, from center to center of 
rivet holes, and r is the radius of gyration, in inches, of the cross- 
section of the lattice bar. These results may be considered to be 
applicable to "under" lattice bars. For "over" bars it seems probable 
that the average stress at failure would be considerably less. 



244 TESTS OF COMPRESSION PIECES 

Effect of Cover-Plates and End Connections. — In the tests of the 
White Heath Bridge, the effect of the cover-plate seems striking. 
The upper chord, U .JJ ^, composed of two built-up channels with one 
cover-plate, gave an excess fiber stress of 20% at the worst section, 
while the posts, composed of two channels laced on both sides, gave a 
maximum of 73 per cent. The high value in the posts may be due to 
other causes, but it seems reasonable to expect that the cover-plate will 
act to reduce the irregularities in fabrication. 

The connections of the ends of the posts evidently exerted a very 
noticeable effect on the stress distribution. In one of the posts tested, 
the stress was greatest at one corner of the post at the top and at the 
diagonally opposite corner at the bottom. It will be remembered that 
the posts were riveted to the top chord, and were connected with the 
lower chords by pins. The floor-beams were riveted to the sides of 
the posts, and this connection affects the stress distribution. Readings 
of deformations taken on the floor-beams and posts show that the 
loaded beam was partly restrained at the ends by the post and that 
there was an appreciable bending in the post. 

Stresses in Column Lacing. — If the load carried by one channel 
of a column were the same throughout its length, no stress would be 
carried by the lattice bars. Such stress is developed whenever there 
is a change in the relative amount of loads carried by the two channels. 
If at the section, AB (Fig. 2), there is an equal division of load 
between the two channels, and also at the section, CD, and if at some 
section, EF, the division of load is unequal, it is evident that the 
lattice bars must be called into action to transmit this stress, and that 
transverse shear exists in the interval. In general, the conditions 
producing this must be complex, rendering analysis unsatisfactory, 
except in so far as the shear may be due to a known eccentricity of 
loading. 

It is evident from the tests that the relative stress in the two 
channel members varies considerably from end to end and that the 
stress in the lattice bars also varies. It seems probable that the transverse 
shear developed may be traced largely to irregularities in outline, or at 
least that these irregularities may be expected to cover up other causes 
of stress in the lacing of centrally-loaded columns, if we include in 
such irregularities all unknown eccentricity. 

The amount of transverse shear necessary to produce the maximum 



TESTS OF COMPRESSION PIECES 245 

observed lattice-bar stress (given in Table 6) is of interest, though 
of course it cannot be taken to be conclusive. The measurements were 
generally made at working loads. So far as observations were made on 
columns tested to failure, the distribution of stress remained much the 
same up to incipient failure. The values given in Table 6 indicate 
maximum average stresses in the bars such as would be caused by a 
transverse load ranging from 2 to 6% of the central compression load 
or of a transverse shear one-half as great. 



246 DISCUSSION ON TESTS OF COMPRESSION PIECES 

DISCTJSSIOlSr 



Mr. Moses. JoHN C. MosES, M. Am. Soc. C. E. (by letter). — This investigation 
is to be welcomed as a forerunner of many more in which examinations 
of the actual conditions under working loads will be made. Tests to 
destruction, or to any point beyond the elastic limit, are of very doubt- 
ful value, and the practice of making them the basis for designing has 
been justified only by the supposed impracticability of making such 
tests as are described by the authors. Certainly nothing can be more 
illogical than to find that load, the first application of which will cause 
destruction, and then to assume that any arbitrary fraction of that 
load can be carried indefinitely with a proportional factor of safety. 
The absurdities of this method have become especially apparent when 
applied to such materials as reinforced concrete, but are none the less 
existent when applied to homogeneous materials like steel. 

The difiiculty of making the proper kind of tests does not alter 
the facts, and the proofs now presented that such tests are not im- 
practicable are very encouraging. They will need to be many in 
number and to be made on structures when first erected and again on 
the same structures after they have been in use a while. They will 
need especially to be made on details of all kinds in order to eliminate 
the very large element of guesswork which now enters into this very 
important part of designing. Some of the authors' conclusions bear 
very significantly on this subject. They find "no relation * * * 
between the stresses actually observed and the stresses computed by 
column formulas," and suggest that the design of details and condi- 
tions of manufacture may be such as to make void the usual assump- 
tions entering into these formulas. 

The whole subject of detailing is connected intimately with that of 
manufacturing. It is of essential importance economically, but has 
not received its proper share of attention. At present, working stresses 
are commonly limited to one-half the elastic limit, principally because 
of the uncertainty which exists in regard to the proportion of the 
strength of the main members, which can be developed by the details. 
With more exact knowledge of the actual distribution of stresses, and 
with the advance in the art of designing details which will accompany 
this advance in knowledge, it may become apparent that higher cost 
of manufacture will be much more than offset by reduced expense 
for material. Better methods of manufacture will also make possible 
the use of stronger steels, with all the possibilities such a change fore- 
shadows. The recent rapid increase in the variety of sections rolled 
shows that this study of detailing need not be confined to the sections 
which have been standard for the past thirty years. 

The authors show by example that a bend of less than ^6 in. in the 
channels of Column No. 1 would account for an increase of stress of 



DISCUSSION ON TESTS OF COMPRESSION PIECES 247 

50 per cent. Shop experience shews that, aside from kinks due to Mr. Moses, 
handling during shipment from the mills or about the shops, the 
actual punching of angles, plates, and channels will curve the separate 
pieces sufficiently to cause such eccentricities as that cited, although 
the parts are apparently straightened by the lattice. Channels will 
not only bend, but will also twist, when punched in the flanges. Their 
flanges will perhaps be kinked at every hole by the blow of the punch. 
Plates will show buckles between the rivets after driving. Twisters 
must often be applied to columns during riveting to keep them from 
assuming a spiral form. Designers insist that lattice bars of I-shaped 
columns made of four angles, shall intersect on a common rivet, and 
that the tie-plates shall be of the same thickness as a single bar, expect- 
ing the angles to be pulled down by the riveting. They are pulled down, 
with a resulting kink in the outstanding legs as well as in the riveted 
leg. Numerous examples of this kind could be cited in support of the 
authors' statement as to conditions of manufacture. 

The authors' experiments on lattice bars show the desirability of 
more adequate provision for staying the open sides of columns, when 
such columns are used; and the frequent bending of lattice bars in 
shipment and erection emphasizes the argument to those acquainted 
with the facts. A return to closed sections with separate joint pieces 
is among the possibilities. 

The accumulation of further data is to be earnestly desired, and 
will certainly have a marked effect on the future development of 
structural work. 

Arthur N. Talbot, M. Am. Soc. C. E., and Herbert F. Moore, Messrs. Talbot 
Esq. (by letter).— It seems to the writers that the comments of Mr. ^° ^°^^' 
Moses, which are from the vievppoint of the practicing engineer, are 
in accord with the results of the tests given in the paper. Many 
structural engineers have recognized that the shop and field processes 
used in the fabrication of structural members produce imperfections 
which give conditions varying widely from the ideal condition tacitly 
assumed in the analyses at the basis of the formulas used in design. 
How much the imperfections can be reduced, with better methods and 
greater care, and what conditions must be expected to be present 
under the best practicable methods of manufacture, can only be matters 
of conjecture. At different times, much interest has been manifested 
in questions connected with the action of the details of structural 
members while under stress; but, it may be doubted whether engineers 
generally appreciate the large variation in the distribution of stress 
over the sections commonly used in compression members. 

The need of experimental work, to determine the stresses which are 
developed in compression members built under ordinary conditions of 
fabrication and erection, must be apparent. That it is actually practi- 
cable to make observations on structural members which will give 



248 DISCUSSION ON TESTS OF COMPRESSION PIECES 

Messrs. Talbot trustworthy information seems evident from these tests. It is hoped 
and Moore. ^^^^^ interest in this problem will be awakened among engineers and 
manufacturers, and that some institution or organization will take up 
the subject in a thorough and comprehensive manner. Such tests 
involve extreme care, and they are expensive, with regard to time and 
labor, whether done in the field or in the laboratory. A full study of 
the action of the compression piece at loads which do not stress the 
material beyond the elastic limit should be included. The expenditure 
involved is far beyond that of tests to destruction alone. It is not 
necessary to state that, to be of value, such an investigation must be 
accompanied by a careful study and analysis of the tests and results, 
though this has not always been appreciated. A programnae of tests 
need not involve a large number of test pieces, but, to be really useful 
for the purpose in view, the time devoted to the test and stijdy of each 
piece must be ample, and the total cost of even a fairly comprehensive 
investigation will be very large. It may be expected, however, that the 
value of the results would repay many times the cost of the work, and 
the expense would be justified by the added security and, perhaps, by 
the economy of metal which might result from the investigation. 

Very recently the writers' attention was called to articles* by 
Professor Lilly, of Trinity College, Dublin, in which he shows, by ex- 
periments on compression pieces formed principally of rolled tubing, 

that, for columns having the same slenderness ratio, — , the strength. 

varies with the relation which the thickness of the metal bears to the 
diameter of the tube, and he discusses a wrinkling action which, up to 
certain thicknesses, governs the strength of the column, irrespective of 
its length. The tests are quite suggestive. Professor Lilly did not 
investigate the loss of strength due to lack of straightness, to eccen- 
tricity of parts, and to the lacing together of component shapes to form 
one member. The conditions in built-up compression pieces make the 
irregularities of stress due to the form of the section, to the manner 
of connection of the component parts, and to the imperfections in- 
herent in manufacture, even more marked than the variations to be 
found in simple rolled sections. 

That the column formulas in common use have limitations, has been 
understood, but the effect which the condition of the component parts 
of a compression member exerts on the distribution of stress over the 
section has not been appreciated, nor has that of the eccentricty of con- 
nection of latticing, and of the possible non-integrity of section. It 
would seem quite probable that, for columns of the same length and 
containing the same amount of metal, one which is of stocky form and 
in which the metal is distributed so as to resist local flexural and tor- 

*"The Strength of Columns," Proceedings, Inst. Mech. Engrs., June, 1905. "The 
Design of Struts," Engineering (Londoji), January 10th, 1908. 



DISCUSSION ON TESTS OF COMPEESSION PIECES 249 

sional action will be much stronger and more satisfactory than a Messrs. Talbot 
column of more flimsy form, which has its metal spread in thinner 

sections, even although the slenderness ratio, , » of the former may 

be considerably more than that of the latter. It seems reasonable to 
expect that a form of section which resists lateral bending, torsional, 
and collapsing stresses, will be much more satisfactory than a more 
flimsy type of column, for the lengths most common in ordinary bridge 
construction. For the longer lengths, the slenderness ratio must exert 
a stronger influence. As far as concerns the strength of the component 
angle, channel, or other structural shape used in a built-up compression 
piece, many engineers have been satisfied with the provision that the 
slenderness ratio of the component member may be less for the length 
between the points of attachment of lacing than the slenderness ratio 
for the column as a whole, and have given little attention to the possible 
non-integrity of the section, or to the probable effect of imperfections 
of manufacture. Fortunately, the large influence of the slenderness 
ratio in column formulas has given sections with which failures have 
not occurred. Whether a column formula should include a factor 
depending on the form of the section and the relative thickness of the 
metal, or whether the allowable stresses for any form of column 
should be based on experimental data for the section used, will depend 
on developments. In either event, full information as to the properties 
of various types of columns and various thicknesses of metal will be 
of great service. The same tests would determine the effect of the 
usual imperfections of manufacture. It would seem, too, that the 
reliability of any analysis of secondary stresses based on the assump- 
tions of the common column formulas may be questioned. With refer- 
ence to the effect of the form of column, Mr. Moses' suggestion in 
reference to the closed section and the form of lacing is pertinent at 
this time. His remarks on the imperfections in shop work and the 
frequency of their occurrence are also to the point. 

It is evident that imperfections of manufacture need consideration 
in the design of columns. Wliether better methods may be used, or 
whether the effect of necessary shop conditions shall be covered in the 
allowed working stresses, is worth discussion. Changes may be made 
so that lattice bars, although they take small stresses, may become more 
efficient compression members. Though column formulas have kept 
down the allowed working stresses, by reason of the large influence of 

the slenderness ratio, --, they seem (for short and medium lengths) to 

be deficient in putting so much weight on the slenderness ratio and so 
little on the shape and probable integrity of the cross-section. 



AMEEICAN SOCIETY OF CIVIL ENGINEERS 

INSTITUTED 1853 



TRANSACTIONS 



Paper No. 1125 

IMPURITIES IN SAND FOR CONCRETE. 

An Informal Discussion at the Annual Convention, July Sth, 1909. 



By Messrs. Sanford E. Thompson, R. W. Lesley, C. P. Howard, 

Richard L. Humphrey, G. S. Davison, S. Whinery, 

Charles M. Mills, T. F. Richardson, E. G. 

Haines, and Sanford E. Thompson. 



Mr. Thompson. Sanford E. THOMPSON, M. Am. Soc. C. E. — During the last two or 
three years attention has been called to several cases where concrete 
failed to harden, and these failures could not be attributed to the 
quality of the cement. Some of these cases have been investigated 
thoroughly, one in particular to a point where quite definite con- 
clusions were reached, and this will serve as an illustration of at least 
one of the causes of poor concrete. 

It has been proved conclusively that the sizes, and the gradations 
of sizes of particles of sand, affect the density, strength, and perme- 
ability of the mortar, and definite laws governing these relations have 
been framed, notably by Mr. Feret,* in France, which show that with 
sand having grains of known sizes the strength of the mortar may be 
estimated. These laws have been corroborated in a general way by 
tests in the United States and elsewhere. 

In the United States, also, further tests have shown the effect of 
scientific methods of proportioning or grading the aggregates of the 
concrete upon its density and strength.f 

However, one frequently finds a sand which absolutely fails to obey 
the laws of gradation of sizes, or of density. Not only may the 

* Annates des Fonts et Chaussees, IV, 1892. 

f'Laws of Proportioning Concrete," by William B. Fuller and Sanford E. Thompson, 
Members, Am. Soc. C. E., Transactions, Am. Soc. C. E., Vol. LIX, p. 67. 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 351 

mortar or concrete made from such sand fail to show the strength Mr. Thompson, 
which would be expected, but it actually fails to harden, or the 
hardening which should take place within a week is delayed for perhaps 
two or three months. 

Frequently, the sand which produces such bad results looks all 
right, and even those who have had much experience in concrete con- 
struction may be deceived by its appearance. Its mechanical analysis, 
that is, the gradation of sizes, may be good, and it may appear clean, 
and yet the quality may be such that it has to be absolutely prohibited 
from use in concrete, and, if used without previous testing, it will 
cause failure. 

This does not disprove the laws referred to, but simply indicates 
that there is something further; that the laws of density apply to a 
clean sand, and that frequently it contains some other material which 
affects its combination with cement, either mechanically or chemically, 
or perhaps both. Such results, such failures of the mortar and con- 
crete to set, show the absolute necessity, not merely of a careful ex- 
amination of a sand which is to be used, not merely of a mechanical 
analysis to determine the sizes of the particles, but of a laboratory 
test, as thorough a test, in fact, as would be given to the cement itself. 

In the special case referred to, a two-story machine-shop with con- 
crete walls was being erected by local contractors, with no engineering 
advice except in the preparation of outline plans. The walls of the 
basement were 12 in. thick, those of the first story, 10 in. thick, and 
those of the second story, 8 in. The interior of the building was mill 
construction with timber beams and plank floors. One night, the 
walls being just above the level of the second floor, during a severe 
windstorm the building collapsed. An examination by the speaker, 
who was called on to make an investigation, immediately after the 
accident, showed that the concrete, although mixed in the proportions 
] : 2^ : 5, had failed to harden, even in the basement. 

The concrete of the entire building was so soft that, although parts 
of it had been laid for at least two months, a knife blade could be 
thrust into it, and it was even difficult to pick from the wall a piece 
hard enough to carry away as a sample. The appearance of the con- 
crete was dark and dirty. There was a thin, hard skin on the outside, 
which had helped to deceive the contractors into believing that the 
material would eventually harden, although they were by no means 
satisfied with it. 

Investigation showed that the concrete had been proportioned 
1 : 2^ : 5, and mixed in a satisfactory manner. It was hand-mixed, but 
apparently well done, and tests of the actual proportions by analysis 
were considered unnecessary. 

It was evident that the cause of the trouble lay in the materials. 
The last of the carload of cement had been used, only the empty bags 



252 DISCUSSION ON IMPURITIES IN SAND FOR -CONCRETE 

Mr. Thompson, remaining; and because representative samples of this cement could 
not be obtained, the other materials had to be tested much more care- 
fully than otherwise would have been necessary. 

The sand and gravel were taken from the site of the building when 
the cellar was excavated. The gravel was ordinary New England 
gravel, ranging from fine, that is i-in., up to particles perhaps 2 in. 
in diameter. It was fairly clean. Some of the pieces were slightly 
coated with dirt, but not more so than is almost always found in a 
gravel bank. 

The gravel screened contained about 12% of sand finer than the 
^-in. size, that is, the screening was imperfect. However, this is 
almost always the case with screened gravel, so that in gravel concrete 
there is apt to be an excess of sand over the nominal proportions 
specified. In screening, a part of the sand, especially on wet days, is 
carried down with the larger stones. 

The gravel was washed as gravel is usually washed in hand-mixed 
concrete. A hose was turned on the pile before the mixing was begun, 
and the fine material was washed down to the bottom of the pile and 
shoveled into a wheel-barrow with the rest. Then the gravel from the 
barrow was dumped on the mixing platform, and was again washed 
with the hose, the dirt simply flowing with the water from the surface 
to the bottom of the pile to be shoveled up and mixed with the gravel 
in the concrete. This is the ordinary but inefi^ective way of washing 
gravel in hand-mixing, unless special apparatus for washing is used. 

The sand appeared to be good. In some places it was rather dark 
colored, but the most of it would pass an ordinary inspection, and 
would be called a sand of fairly good quality, certainly good enough 
for the work which was being done. The results of the mechanical 
analysis of the sand, given in Table 1, were above the average. Three 
per cent, by weight passed a No. 100 sieve, and about 25% was caught 
on a No. 8 sieve, but a closer examination of the bank showed con- 
siderable variation in the sand; in some places it was rather dark and 
reddish in color, while piles which had dried out looked "dead." 

The representative of the cement company suggested a practical 
test which had never before come to the speaker's attention. Taking 
a double handful of moist sand from the pile, he allowed it to run 
between his hands as they were held with the thumbs up about 1 in. 
apart, at the same time moving his hands back and forth. Repeating 
this operation several times, always taking naturally moist sand from 
the interior of the bank, between the fingers of both hands there was 
collected a dark slimy substance which contained scarcely any grit. 
Some of this, scraped from the fingers and afterward tested by ignition, 
was found to consist almost entirely of vegetable matter. 

A further examination of the bank showed that, in places where 
it had been cut to a vertical face for an excavation or for screening out 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 



253 



the sand, the rain had washed down from the surface soil a material Mr. Thompson, 
similar to that which had collected between the fingers in the foregoing 
test, which formed a scum on the vertical face of the bank. 

The surface soil was an ordinary, medium quality of loam, with, 
however, one or two dark, almost black, streaks in it, from ^ in. to 
I in. thick. 

All indications seemed to point to the cause of the trouble being 
vegetable matter in the sand, and vegetable matter which had ap- 
parently been washed down by the rains from the surface soil into 
the porous sand and gravel underneath, gradually coating the grains. 

Tests of Sand. — Samples of the sand and stone were subjected to 
thorough tests. The mechanical analysis is shown in full in Table 1. 

TABLE 1. — Mechanical Analysis. 



Analysis No 


A-92. 




A-96. 




Description 


Average sand from excavation. 


Reservoir sand. 


Date of collection. . 


May 5th, 1908. 




1907. 


" " analysis ... 


May 8th, 1908. 


May 9th, 1908. 


Percentage of 












moist ur 


e 


2.6 before drying. 













Size of sieve. 
















Total 
weight 


Total 
percentage 


Percentage 


Total 
weiglit 


Total 
percentage 


Percentage 










Inches. 


No. 


passing. 


passing. 


14-in. sieve. 


passing. 


passing 


Mill, sieve. 


1.50 


]i4in. 


97.0 


100.0 


100.0 


100.0 


100. 


100.0 


1.00 


1 in. 


97.0 


100.0 


luo.o 


100.0 


100.0 


100.0 


0.50 


14 in 


97.0 


100.0 


100.0 


100.0 


100.0 


100.0 


0.25 


^in. 


9d.7 


95.6 


100.0 


96.2 


96.2 


100.0 


0.16 


5 iu 


81.8 


84.4 


88.4 


88.9 


88.9 


92.5 


0.0583 


12 in. 


56.5 


58.2 


61.0 


68.2 


68.2 


71.0 


0.0335 


20 in. 


39.5 


40.7 


42.6 


48.2 


48.2 


50.1 


0.0148 


40 in. 


14.6 


15.0 


15.8 


17.0 


17.0 


17.7 


0.0055 


lOa in. 


2.6 


2.7 


2.8 


2.3 


2.3 


2.4 


0.00.30 


200 in. 


1.8 


1.9 


2.0 


1.1 


1.1 


1.1 



Very fortunately, for purposes of comparison, the speaker happened 
to have in his laboratory a large sample of another sand which had 
been used satisfactorily in the construction of a reservoir* near Boston. 
This reservoir, because of its size and shape^lOO ft. in diameter and 
nearly 50 ft. high — required special tests of the materials. The 
mechanical analysis of this good reservoir sand, shown in Table 1, 
was found to be almost identical with the sand used in the building 
which failed; moreover, the chemical composition was also practically 
identical, each containing about 75% of quartz. 

Notwithstanding this apparent similarity, 1 : 3 mortar from the 
reservoir sand gave a strength of 272 lb. in 7 days and 382 lb. in 28 

* The Engineering Record, January 12th, 1907, p. 32. 



254 DISCUSSION ox IMPURITIES IN SAND FOR CONCRETE 

Mr. Thompson, days, while the poorest sample of the sand in question gave an average 
of 20 lb. in 7 days and 75 lb. in 28 days. 

Volumetric Test. — To be sure that the bad quality of the sand was 
not due to the shape of the grains, which might prevent proper com- 
pacting, and, consequently, an excess of voids in the mortar, volumetric 
tests were made, and the density was found to be normal. The density 
of 1 : 3 mortar made with the sand in question was 0.679, and that of 
mortar made from the reservoir sand in the same proportion was 0.683. 

Mica. — A little mica was observable in the samples. A careful 
examination, however, indicated that this was not sufficient to cause 
trouble. 

Clay. — It has been sometimes claimed that, from a chemical stand- 
point, clay matter is injurious to sand. An examination of the chemi- 
cal analysis showed, however, that the amount of clay, although large, 
was approximately the same as in the good reservoir sand, and, there- 
fore, could not be considered as the cause of the poor quality. 

Tensile Tests. — The tensile strength of the mortar from the sand 
in question, as already stated, averaged 20 lb. per sq. in. in 7 days and 
75 lb. in 28 days. 

The tests in different series using sand from different parts of the 
bank ranged, in the 7-day test, from 11 to 40 lb. To be sure that the 
cement used in testing or the laboratory methods were not at fault, 
tests were made in two different laboratories and with three different 
cements. Three different proportions were also used, namely, 1 : 2J, 
1:3 and 1:3J. Specimens, for comparative tests in air, were stored 
in a moist closet, and in water. The results all checked so closely as to 
eliminate the question of cement or of manipulation. 

Comparison of Moist and Dry Sand.— A comparison of 1:3 mortar 
made with the moist sand as it came from the natural bank and with 
the same sand after drying, showed some relative increase in strength 
due to the drying. In this series the average strength at 7 days of the 
sand direct from the bank was 8 lb., and of the dried sand mortar 23 lb. 

Comparison of Natural Sand and the Same Sand after Washing. — 
A comparison of the sand in question before and after washing was 
as follows: 

Unwashed 40 lb. at 7 days, 102 lb. at 28 days. 

Washed 97 " "7 " 196 " " 28 " 

Standard 200 " "7 " 275 " " 28 " 

The grains of the washed sand were not thoroughly clean. 

Microscopical Examination. — The microscopical examination of the 
grains of sand as obtained from the bank showed them to be covered 
with a dark brown coating which did not readily brush or wash off. 
The particles of the Waltham sand, on the other hand, were clean. A 
strong magnifying glass was found efficient in examining the grains 
for coating. 



DISCUSSION ON IMrURITIES IN SAND FOR CONCRETE 



255 



Washing Tests. — To examine and test the character of the silt in Mr. Thompson, 
the sand, it was found necessary to remove the fine matter by washing. 

To compare the results by washing and by screening out the silt 
through a No. 100 sieve, tests were made in both ways with the follow- 
ing results: 

Worst sample of defective sand: 1.60% silt by washing; 1.14% 

by screening. 
Average sample of defective sand: 2.46% silt by washing; 

2.68% by screening. 
Keservoir sand: 1.16% silt by washing; 3.66% by screening. 

Although there is but slight difference between the results from 
washing and from screening, the chemical analyses indicate that the 
washing removes much more of the deleterious vegetable matter than 
the screening, so that this process should be followed when testing 
silt in sand. 

It is noticeable in the above results that the poor sand gave a larger 
percentage of silt by washing than by screening, while the good sand 
gave a much smaller percentage by washing, indicating that in the 
latter case a large proportion of the very fine material was of heavier 
mineral origin. 

The method used in washing to remove the silt was a very simple 
one, but was found to be more effective than several other more 
elaborate methods which were tried. The sand was placed in a large- 
mouthed quart bottle about half full of water, and shaken thoroughly. 
The dirty water was poured off, and the operation repeated several times 
until the water was nearly clear. The wash-water was then evaporated, 
the residue thoroughly dried, and the loss on ignition, which determined 
the quantity of organic matter, was found. 

Before igniting, the silt was passed through a No. 100 sieve to 
remove large particles of dirt which evidently would not be injurious 
if distributed through the mortar. 

Chemical Analysis of Silt.— ^oth the poor sand in question and the 
good reservoir sand were washed, and a chemical analysis was made 
of the silt, including a test of the residue by ignition. The analyses 
of the silt were as shown in Table 2, 

TABLE 2. — Analyses of Silt. 





Percentage 

in reservoir 

sand. 


Percentage in 
average sand, 
B, from site. 


Percentage in 
worst sand, 
G, from site. 




7.80 

54.13 

23.10 

9.00 

2.07 

3.50 

45.0 


8.24 
45.32 
25.13 
6.71 
1.93 
11.86 
50.0 


13.07 




34.94 




20.80 


Oxide of iron 


5.90 








26.33 




35.0 







256 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Thompson. Inspection of these analyses shows but one notable difference: The 
organic matter in the silt from the reservoir sand was 2.50%, while in 
the silt from the poor sand it was very high. In one of the samples of 
poor sand the organic matter was 11.9% and in the other 26.3 per cent. 
When ignited, the organic matter in the poor sands gave off a peculiar 
odor of woody fiber, or leaf mould, thus indicating its probable vegeta- 
ble origin. The good reservoir sand, on the other hand, gave no 
appreciable odor, and showed a reasonably small percentage of organic 
matter. 

Reference has been made to the comparison of screening and wash- 
ing. Tests by ignition show that, by washing, frequently double the 
percentage of organic matter is obtained than by simply screening. 

It is noticeable that the silica is low, but it must be remembered 
that this is the analysis of the silt and not of the total sand. 

Introducing Silt Into Standard Mortar. — To confirm still further 
the conclusion that the silt was the cause of the trouble, a mortar was 
made up of 1:3 standard sand with an addition of li% of silt based 
on the weight of the sand. The resulting tensile strength of the mortar 
was 29 lb. per sq. in. in 7 days and 107 lb. in 28 days, whereas the 
normal strength of the standard sand mortar was about 200 lb. in 7 
days and 275 lb. in 28 days. 

A similar test was made by introducing 1^% of silt into neat 
cement, and the resulting strength was about one-half that of the neat 
cement without silt. 

Concrete Tests. — Specimens of concrete were also made, sand from 
different parts of the bank being used in different tests. The cement 
used in these tests was of a well-known brand, which was carefully 
tested to see that it was normal. The poor sand and gravel from the 
site of the building were used in the test as well as the good reservoir 
sand and gravel. The blocks were broken at the age of about 18 days 
at the Watertown Arsenal, with the results shown in Table 3. 

Hardening of Specimens. — Careful watch was kept of the specimens 
of concrete and of the mortar as they hardened, and pats of the con- 
crete were also made. The mortar and concrete made with the good 
reservoir sand set up hard within 24 hours with a light gray color, 
while the other specimens remained so soft that they could not bear 
the pressure of the thumb nail for several days. 

Conclusions. — As already stated, the tests indicated conclusively 
that the trouble with the sand was due to the vegetable matter which 
it contained, and that subsequent tests of other sand, and examinations 
of structures, indicate this to be a common cause for poor mortar 
or concrete. In many cases no failure results, but the concrete does not 
harden properly and never becomes as strong as it should. 

The percentage of silt given in the chemical analysis of the silt 
appears to be large, being 11.9% in one sample and 26.3% in another. 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 



257 



When given in terms of the total sand, however, it is a very small Mr. Thompson. 

percentage, because there is so small an amount of silt in the sand. 

Based on the weight of the total sand before washing, therefore, the 

percentage of silt as it comes from the bank is 0.27% in one case and 

0.39% in another, an extremely small amount, and one which would 

not be indicated by any test of settlement in water or rubbing in the 

palm of the hand. 



TABLE 3, — Compressive Strength of Concrete, in Pounds 
PER Square Inch. 

Age, 18 days. Proportions, 1 : 2^ : 5. 



Aggregate. 


Cement K. 


Cement A. 




Individual. 


Average. 


Individual. 


Average. 


Sand A. . and Gravel E 

" MA + ^B " " K 

" HA.-\-%B " " E 

" W " " E 

" W " " W 


'818 
692 


■755 


200 
625 
749 
653t 
1 505 


200* 

'687 

I'SOS* 



Notes.— Sand A— Worst sand from site of building. 

Sand B.— Average sand from site of building. 

Sand W.— Good reservoir sand. 

Gravel E. — Screened gravel from site of building. 

Gravel W.— Screened reservoir gravel. 
* Only one specimen, but a good one. 
t Only one specimen, and possibly slightly defective. 

These and other tests indicate that there are two percentages of 
vegetable matter, which appreciably affect the quality of a sand : First, 
the percentage of vegetable matter in the silt, and second, the per- 
centage of vegetable matter in the sand. Although the tests made thus 
far are too few to enable one to draw definite quantitative conclusions, 
it would appear that, in order to be injurious, the organic matter in the 
silt must be more than 10%, and, at the same time, the organic matter 
in the total sand must be more than one-tenth of 1 per cent. 

Both these conditions are necessary, because it appears from tests 
that, in certain cases, the vegetable matter in the silt may be greater 
than 10%, but there may be so little silt in the sand that the organic 
matter in the total sand will be less than one-tenth of 1%, and the 
sand will pass the tensile test. 

Whether the cause of the results is due entirely to chemical action 
or whether it may be due in part to mechanical action, the organic 
matter surrounding the grains so that the cement will not adhere, has 
not been determined. Most probably the cause is chiefly chemical, but 
in a small measure also mechanical. The subject appears to be of 
sufficient importance to warrant further and thorough investigations. 



258 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Thompson. In Conclusion, special stress should be placed on the necessity, in 
concrete work, of thorough tests of the sand. It is not merely necessary 
to examine sand with the eye; it is not sufficient to test it by rubbing 
it in the hands; it is not enough to make a mechanical analysis and to 
determine the sizes and the gradations of the particles, but, in every 
case, unless the sand is from a bank of known good quality and has 
been previously tested, as careful tests as are required of the cement 
should be made. Probably the best one is the ordinary tensile test 
advocated by the Society's Special Committee on Concrete and Rein- 
forced Concrete, which requires that: 

"Mortars composed of one part Portland cement and three parts 
fine aggregate by weight when made into briquettes should show a 
tensile strength of at least 70% of the strength of 1 : 3 mortar of the 
same consistency made with the same cement and standard Ottawa 
sand." 

In case a sand must be used immediately, with no time to make 
tensile tests, or in case special investigations are needed to determine 
the causes of poor quality, the washing test and determination of the 
organic matter are of special value. The mechanical analysis, which 
shows the proportions of the grains of different size, is also of great 
value as indicating the comparative value of different sands which are 
free from organic matter. 

There is one further thought which might be brought out. These 
impurities have been spoken of as dirt and matter which might be 
washed out. One sample of sand which was sent to the speaker from 
Philadelphia, had a great deal of dirt in it, that is, it had much more 
fine material than the defective sands which were investigated, and 
yet washing it increased the strength of the mortar instead of decreas- 
ing it. The fine material was found to be, not organic matter, but of 
mineral composition. 

Mr. Lesley. R. W. Lesley, Assoc. Am. Soc. C. E. — Mr. Thompson's contri- 
bution is very interesting. In the light of the recent report of the 
Special Committee on Concrete and Reinforced Concrete, of this 
Society, it rather emphasizes the principle of "locking the stable door 
after the horse has been stolen;" in other words, the building had 
fallen before the exhaustive examinations of sand were required. On 
the other hand, the speaker believes that the report contains, in one 
paragraph, what is more germane to the subject of concrete and rein- 
forced concrete than any other lesson, namely, that the sand shall 
be tested in the same way that the cement is tested. The mortar must 
be a standard mortar. This report provides a means to determine that 
question. 

At the present moment there is no doubt that in every large cement 
manufacturing plant in the United States there are many inspectors 
who are testing cement prior to its shipment to various parts of the 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 259 

country. The speaker happens to know that, to-day, in one mill, Mr. Lesley, 
there are no less than twenty or twenty-five men engaged in testing and 
inspecting cement. One part of that cement is to be used to three parts 
of sand, and yet the speaker does not know of anyone anywhere who 
is testing the sand to be used in the same manner that the cement is 
being tested. 

The engineer requires that cement shall be tested in a dozen 
different ways, to show that it meets the requirements of the specifica- 
tion, and no cement manufacturer, under existing conditions of inspec- 
tion at the mill, can produce a cement that will not be actually and 
scientifically right; and yet that cement may be used, in the propor- 
tion of one to three, with a sand that no one has tested, no one has 
examined, and about which nothing is known. 

Mr, Thompson's paper brings us face to face with a state of affairs 
which has been going on for many years, and, by making a require- 
ment that all sand shall be tested in the same manner that cement is 
tested, the Committee on Concrete and Reinforced Concrete has cer- 
tainly taken a great step forward. 

In the recent report of the Building Code Commission of the 
Board of Aldermen of New York City, a step farther has been taken. 
Whereas, to-day, cement only is tested, under the report of the Com- 
mittee on Concrete and Eeinforced Concrete, both the cement and the 
sand will be tested. But there is still another step: Suppose one has 
good cement and good sand, and yet the building fails. That failure 
may be due to the fact that the contractor has used seven or eight, or 
even ten parts of sand, and again there is an opportunity ''to lock the 
stable door after the horse has been stolen." In this particular report 
of the Building Code Commission, it is provided that the sands 
represented in the mortar shall be inspected from time to time, and 
that the mortar from the mortar box shall have 70% of the strength 
of the same cement, used with standard Ottawa sand. In other words, 
it places the responsibility for the mortar which binds the concrete, or 
which binds the masonry, on the men who are actually providing that 
mortar. 

The cement is to be tested, the sand is to be tested, and the mortnr 
is to be tested, and the speaker believes that this is a most important 
step forward. He also believes that, this country being one of the 
most progressive in the use of concrete, the day is not far distant 
when the American, instead of whittling timber with a knife, will 
whittle concrete with a shovel; that the same cleverness which, 
heretofore, has enabled American genius to make its way forward, will 
enable it in the future to whittle with a shovel, and to do things with 
this plastic material which to-day seem impossible. 

Recently, at a meeting of one of the scientific societies, one man 
read a paper on the use of concrete with nails as reinforcement. 



260 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Lesley. Another engineer stated that he had made some concrete and had rein- 
forced it with jackstones; and that statement was again made ridicu- 
lous by a man who had used old hoop-skirts as a reinforcement, and 
if American genius has gone so far in the use of concrete, there seems 
to be no limit to what it will do in the future. 
Mr. Howard. C. P. HowARD, M. Am. Soc. C. E. — The Speaker has recently seen 
some specifications in which there was a provision that the stone used 
should be free from oil, and understood that the parties using these 
specifications went so far as to refuse stone from certain quarries on 
account of the oil which it contained. Such a precaution would 
probably not occur to most engineers. It emphasizes the importance 
of the most careful examination of all three ingredients, namely, 
cement, sand, and stone, and also of actual tests of concrete briquettes 
made from the same materials used in the structure. 
Mr. Hum- RiCHARD L. HUMPHREY, M. Am. Soc. C. E. — There has been a 
^ '®^' decided change in the requirements for sand, and this is necessary, 
because, to-day, the character of the sand used in cements, mortars, and 
concretes is of the highest importance since it has a vital effect on 
their strength. Specifications which require the sand to be clean and 
sharp are wholly inadequate, and yet there are many engineers who 
pass on the quality of sand by its appearance, thus complying with 
this specification. The fallacy of thus judging it is readily illustrated 
in an examination of sands from different parts of the country, such 
an examination showing how easily the eye can be deceived as to its 
quality. Sands have been sent to the speaker from various parts of 
the country, and he has had the opportunity, in a number of cases, 
of noting how easily it is possible to be deceived. As an instance, two 
sands from the Isthmian Canal may be given. The one which com- 
plied with the specification by being clean and sharp, and apparently 
the better of the two, proved by test and careful examination to be 
inferior to the sand which was not sharp and appeared to be dirty. 
The physical tests were the means of determining the relative quality 
of these two sands. As a matter of fact, the poorer sand was the 
cheaper as it was more readily accessible for the work in question. 
This, the speaker believes, is one of the many illustrations of the 
value of making physical tests of sand before passing on its quality. 
Elaborate tests of cement are made, and just as thorough tests of the 
sand should be made, because its effect on the strength of concrete 
is as great as, if not greater than, the cement which binds it together. 
Many sands which appear to be sharp and of a fine quality are of 
uniform size and contain so much fine material that they produce 
very slow-hardening mortars, thus interfering with the stability of the 
structure in which they are used. 

There is no doubt that a great deal of the inferior sand used to-day 
is the cause of failures. It may be stated that the use of this sand 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 261 

is in a large measure due to a reasonable ignorance on the part of Mr. Hum- 
engineers and others in charge of the work. It is highly important ^ '^^^' 
that the engineer should be alive to the necessity of testing the sand 
just as rigidly as he tests the cement, thereby securing a better grade 
and eliminating many troubles which may arise from the use of 
improper material. The report of the Society's Special Committee 
on Concrete and Reinforced Concrete, in requiring a mortar-box test, 
marks an important step forward. 

The fine material found in sand, which is generally referred to as 
loam, should be properly called silt, because it generally consists of 
very finely-divided material of the same character as the sand itself. 
In the investigation of sands from all parts of the country, by the 
Structural Materials Testing Laboratories of the United States Geo- 
logical Survey, it has been found that the presence of this fine material, 
if it is of a granular character, proved beneficial, as it increased the 
density of the mortar or concrete, and consequently its strength. In an 
effort to increase the impermeability of mortars and concretes, so- 
called colloidal clays have been used. A fine sand is always used 
with the idea of filling voids, and the speaker believes that its character 
is immaterial, if it is granular and is sufficiently fine to fill the voids 
properly. Indeed, where the so-called colloidal clays seem to produce 
a greater density, it has been found that a very finely-divided sand 
would produce the same effect. On the other hand, there are sands 
and gravels which contain an excess of finely-divided material which 
it is necessary to screen out in order to get the requisite density and 
strength. In the study of sands referred to, an appreciable quantity 
of organic material is rarely found. Where sand or gravel is not well 
graded, or concrete is not properly proportioned to secure maximum 
density, a finely-divided material or a little extra cement itself may 
be added with beneficial effect. Where the sand or gravel naturally 
possesses the requisite amount of fine material to fill the voids, the 
addition of more material of this character would tend to reduce the 
strength of the mortar and prove otherwise detrimental. Again, a 
large percentage of finely-divided material will produce mortars and 
concrete which set very slowly, and this is a decided disadvantage. In 
this discussion, material finer than will pass a No. 200 sieve is referred 
to as silt. Many natural sands contain 6 or 8% of a very fine material, 
which, by reason of filling the voids, adds materially to the strength of 
the resulting mortar and concrete, and gives rise to the impression that 
small percentages of finely-divided material are beneficial to the mortar 
or concrete. If the mortar or concrete is granular, and the fine material 
does not occur as a coating on the particles of sand, and is not in too 
great excess, there is no doubt that it will be beneficial. On the other hand, 
fine material which is of a flaky character and coats the particles of 
sand or gravel, has a decided weakening action, and is objectionable. 



262 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr, Hum- The Lise of clays of various kinds is also objectionable, because in 

P fey. Qp(jgj. iQ i^^ix them with the cement properly, they should be thoroughly 

ground and dried to an impalpable powder. Otherwise, they will have 

a tendency to ball up, and the result will be little nodules of clay in 

the mortar or concrete. 

Mr. Davison. G. S. Davisok, M. Am. Soc. C. E.- — The importance of using proper 
sand for cement mortar has long been recognized. Some fifteen or 
sixteen years ago the speaker drew up specifications for cement mortar, 
to be used in the foundations of a large industrial establishment, and 
great care was taken to specify that the sand should be sharp river 
sand, such as is found in abundance in the neighborhood of Pittsburg. 

This sand was shipped to the point of construction by car, and at 
considerable expense to the contractor for freight. After a few days 
it was discovered that the mortar was not setting properly, and, on 
making a series of investigations, it was decided that the trouble was 
caused by the sand. It was just as specified, but the mortar made from 
it would not stand the tests. As a temporary expedient, it was agreed 
that, for a few days, and in some unimportant foundations of the 
building, the contractor would be permitted to use, in the mortar, 
bank sand which was being excavated for the foundations for the 
building. Much to everybody's surprise, the mortar made with this 
bank sand was satisfactory in every particular, 
Mr. whinery. S. Whinery, M. Am. Soc. C, E. — The word "sharpness" is fre- 
quently used in speaking of the qualities of sand, and perhaps it is 
worth while to call attention once more to the fact that it is an 
indefinite and even deceptive term. This, the speaker believes, has 
been pointed out a number of times in the Transactions of the Society. 
The test usually applied to determine the sharpness, that of grinding 
the sand between the fingers, is, to say the least, not decisive. The 
rounded and polished grains of a beach sand, if of a certain size, seem 
by this test to be decidedly sharp. If these same grains be crushed 
to a comparatively fine powder and rubbed between the fingers, they 
will give the impression that they lack sharpness, although the indi- 
vidual grains are angular, with very sharp edges. 

It is undoubtedly true that the character of the different kinds of 
sand has not been investigated sufficiently in connection with the 
making of mortar and concrete. There may be great differences in 
the character of sands, which, in most respects, appear to be alike. 
Some have a greater degree of what has been called the quality of 
absorption, and the cement seems to take a stronger hold of them 
than it does of other sands which are apparently just as good. This 
quality of the surface of sand grains deserves more careful investiga- 
tion than it has yet received. Until engineers are able to drop the 
indefinite term "sharpness," in specifications for sand, and to substitute 
for it some accurate scientific description of the sand wanted, un- 
satisfactory results may be expected sometimes. 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 263 

Charles M. Mills, M. Am. Soc. C. E. — One of the sands tested Mr. Mills 
by Mr. Thompson in his studies subsequent to the special investigation 
to which he refers was taken from the excavation for the subway on 
Market Street, Philadelphia, Pa., from the intermediate and lower 
strata, the cleanest being selected for the work. 

The large proportion of silt was regarded with suspicion, and its 
use in concrete, except in unimportant details, resulted only after 
thorough tests, extending over 90 days. The material was hauled to 
a yard, passed through f-in. screens, and used without washing; it 
contained a good proportion of coarse particles, well graded from 
coarse to fine. Daily tests were made, in order to control the quality, 
the proportion of silt varying from 6 to 29 or 30%, which was 
determined by the matter not settling in 45 sec. when a sample was 
agitated in water, the sample being successively agitated in fresh water 
until the water was clear. The proportion of silt averaged about 21^%, 
the upper limit of 30% being unusual, and representing the approxi- 
mate limit for rejection. The silt was practically impalpable, and 
analysis showed that it consisted usually of fine sand and disinte- 
grated mineral matter in about equal proportions. The sample fur- 
nished to Mr. Thompson contained 0.4% of organic matter which did 
not seem to affect the strength in mortar briquettes. 

Briquettes in the proportion of 1:3 gave better tensile strengths 
than those obtained from any other of the coarse sands or fine gravels 
offered for use, except in the case of a certain river-washed silicious 
sand which was remarkably clean and well graded from coarse to fine. 
With this exception, the tensile strength of 1 : 3 briquettes made with 
the excavated gravel was in excess of that obtained from briquettes 
made with all other materials, except mixtures of stone dust and stone 
grit. A few typical tensile tests, given in Table 4, show the results 
with standard quartz sand, the river-washed sand referred to, and the 
excavated gravel, in which the content of silt is stated. 

The tests in Table 4 are in general agreement with many others 
made during the progress of the work, both in the values obtained and 
in the vagaries with respect to the effect of the varying proportions of 
silt, and, in considering the latter, the constituents of the silt should 
be borne in mind. 

The silt comprised the fine material so desirable in producing 
dense and water-tight concrete, and, at the same time, the material in 
which it occurred made tenacious mortars. 

The tests on 6-in. cubes revealed the same vagaries with respect to 
the effect of more or less silt in the excavated gravel, the tests and 
experience with the concrete in which this gravel was used confirming 
the favorable conclusions reached from the tensile tests respecting the 
value of the gravel for use as the fine aggregate. The presence of so 
much fine material made it necessary to exercise special care in 



264 



DISCUSSION ON IMPUEITIES IN SAND FOR CONCRETE 



Mr. Mills. 





W 


R) 


H 


H 


w pg 


CO 




1 

1 


H 


M 
o 

P 


tD 2 3 




k 


^ p 


■0 


^ P £ 




p 

^ 


1 a 


D 


g. p. ^ 


p 


CD 


5 01? 01? 


CD (Yo Ci 
p. 'q P 


p. 


& ~ 


t-i 


CD 


=^ ^ < 


P 
< 


'^ < N 


p' 


JO 


p i 


i 


5 2. M 






<i 




£ : P 




£. 


2. : 












: : P 












-1 








Q "So "co 


CO 


CO CO 


CO 


iO CO GO 


& 


s^ ^ 


-^ ^-^ ^ 


— ' 


.— ' ^- ^^ 


P 


t& 


■jS CO 


w 


iS to tS 


■-^ 


Ol 










>4^ 


M h-i 


"" 


05 1— CO 










^ ^ ^ 


to 






»-L CO CO 


00 


s 


CO 05 


w 


^-3 3 


& 


*. 


4i CO 


CO 


<(^ CO to 


p 




*i » 


^ 


O 05 --! 

c;i *. oc 


W 


— CO 


35 


C/! 










CO 














I-' to to 


3 

o 


w 


CC CO 


CO 


-J? ^ "* 


Ol 


lb. 4^ 


>u 


lb- it^ OS 


B 




CO cn 


2 


-J to to 


01 


-1 


us CO 




to ~i -J 










o: 












CO 


CO w 


cc 


o oc 00 


3 
o 


lt>. 


J^ .1^ 


«^ 


>c> iC^ CO 


D 


iC^ 


CO CI 


en 


*» -> to 


c- 


~J 


r> 50 




CO or O 










g 
















p 










JO 






m X 






4^ 
o 




-a 


J^ .CO 

o o 




^rs 






o 




o 


o o 




p! 










5 










^__, 














H 






H 




J"* 












CD " 


2 


C5 




f5 no 




^5=3 


CD 


o 




CD o 




c 










3 










> 












j_i * 


"^ 5 












^ CO • 


-°53 












Ci CO 


• p 












to o ; 


M 










CD 








-^ 


<! 








<J 




E. 




P 
















r'-c' 




t9^ 






ETCD 








•-! » 




ii » OJ 






$ 




CD 






(DOBS 




§ CM 






0-5' 




?S 5' 






!•§ 




s|s 


CD 




1^ 




SJ?P 


3 
P 




cc P 




Hip 






• < 
a 




§:! 






p 




►O^ 






n 




dCrq 






CD 




S ^ 






tn 










Ms 




CD •"•= 










cn 5 






d 




O 








3 






5-3 







DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 265 

cleansing the scum from the surfaces of previously deposited concrete Mr. Mills, 
in making joints. Briquettes made with cement and the pulverized 
dried scum gave very small tensile strengths. 

It is desired, incidentally, to emphasize the necessity for the ex- 
amination of the properties of silt in any given case, since it may either 
be a source of dangerous weakness in mortar, or a valuable con- 
stituent in gravel for concrete, when water-tightness is desired. 

One purpose in introducing the foregoing is to invite a discussion 
of the following question, which is asked with full appreciation of the 
laboratory researches and the experiments with colloidal clay and 
other materials in reducing the permeability of concrete : When water- 
tightness, or the reduction of permeability in concrete is necessary, 
and the available materials for concrete do not contain in themselves 
a sufficient quantity of the fine material required for this result, is it 
wise to introduce a fine material in the intermediate aggregate, or is it 
better to increase the quantity of cement? 

In considering the foregoing, it is desired that the practical features 
of the case be considered from the standpoint of the engineer who 
aims to control the quality of the materials, and to see that such 
materials are properly used. It is also understood that the most 
judicious proportioning of the coarse and fine aggregates will be made 
in any given case and that the proportion of cement will produce the 
desired structural strength. It is obvious that local considerations 
of available supplies will affect the decision, but it is desired to assume 
that the requisite materials are available at reasonable, or what may 
be termed balanced, prices. It is also to be assumed that the materials 
are mixed by machine, according to approved methods, and that the 
question is intended to apply to natural or ordinary commercial 
material, and not to patented or secret compounds. 

T. F. EiCHARDSON, M. Am. Soc. C. E.— The statement has been made Mr. Richard- 
that fine sand or silt tends to increase the density of mortar. This may ^°°' 

be true for lean mortars where 1 part of cement is used to 4 or 5 
parts of sand, but the speaker does not think that it is true for richer 
mortars such as are commonly used in American practice. It certainly 
is not true for mortar of 1 part of Portland cement to 2 parts of sand. 

During the construction of the Wachusett Dam of the Massa- 
chusetts Metropolitan Water-Works, extensive experiments were made 
to determine the effect of using sands of different sized grains, and 
various combinations of sand, on percolation and on the tensile strength 
of mortars. Three grades of sand, called fine, medium, and coarse, 
were used in the experiments ; the fine sand all passed a No. 100 
sieve, the medium sand all passed a No. 30 sieve and was held 
on a No. 100 sieve, and the coarse sand all passed a No. 8 sieve and 
was held on a No. 30 sieve, these sieves having 10 000, 900, and 64 
meshes, respectively, to a square inch. All three sands had practically 



266 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Richard- the Same percentages of voids, ranging from 41 to 42%, and the 
fine sand had practically the same percentage of voids and the same 
mechanical analysis as the cements used in the experiments, which 
were standard American cements. 

Experiments to determine the percolation through, and the tensile 
strength of, the mortars, were made with the fine, medium, and coarse 
sands, and with nineteen mixtures of these sands, the mortars experi- 
mented on consisting of 1 part of cement to 1, 2, 3, and 4 parts of 
sand by measure. A mixture of 75% of coarse sand and 25% of fine 
sand gave a sand having 31% of voids. 

As illustrating the density of the mortars, it was fovmd that a 2: 1 
mortar, made from two barrels of coarse sand and one barrel of 
cement, gave 2.4 barrels of mortar, while, if two barrels of fine sand 
were used, 3.0 barrels of mortar were obtained. All mixtures of the 
fine, medium, and coarse sands gave more mortar than the coarse sand, 
the 75% of coarse and 25% of fine sands giving 2.53 barrels of mortar 
for the 2: 1 mixture. This was to be expected, as the 41% of voids 
in the coarse sand, being large, were filled by the cement, while the 
42% of voids in the fine sand were not filled by the cement, as the fine 
sand and the cement had grains practically of the same size. 

In testing for percolation and for the tensile strength of the 
mortars, it was found that mortars of coarse sand with no admixture of 
fine or medium sands gave practically the least percolation and the 
greatest tensile strength, with 1 part cement to 1 part or 2 parts sand. 
For mortars of 1 part of cement to 3 parts of sand, the mortar made 
of 50% coarse, 25% medium, and 25% fine sands gave slightly better 
results than that made with 100% of coarse sand, both as to percola- 
tion and tensile strength, and this was also true to an increasing 
extent for mortar of 1 part of cement to 4 parts of sand. The mortars 
of fine sand, whatever the proportions of cement might be, in all 
cases gave the greatest percolation and the least tensile strength. 

The sands used in the above experiments were almost entirely silica, 
and contained little or no alumina. The pressure was 75 lb. per 
sq. in. — equivalent to a head of about 170 ft. 

In coarse sand the particles themselves are impervious to water, 
and the fine cement fills the interstices between the grains for mortars 
consisting of 1 part of cement to 2^ parts of sand or richer. For 
leaner mortars, consisting of 1 part of cement to 4 or 5 parts of sand, 
such as were experimented with in the French tests by M. Feret, 
which have been referred to, a mixtvire of fine or medium sand is 
desirable to help out the cement. Such mortars, however, are rarely 
used in American practice. M. Feret's tests indicated that the best 
sand was a mixture of coarse and fine sands, without any medium 
sand. The experiments referred to by the speaker indicated that 
medium sand mixed with coarse sand is more desirable than fine sand. 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 2G7 

E. G. Haines, Assoc. M. Am. See. C. E. (by letter). — This sub- Mr. Haines, 
ject reminds the writer forcibly of a remark made to him several 
years ago, with some heat, by a contractor, to the effect that it was "a 
mighty easy thing to sit at a mahogany desk and write specifications 
for sand, but quite another matter to provide it on the work." As a 
result of that remark, the writer made an examination of the sands 
found in some 25 sq. miles of country, and came to the conclusion 
that the contractor was right. 

There are often sections of country, of quite large extent, in which 
work is required to be done, and in which no pit sand can be found 
fit for use in its natural state on account of the large quantity of 
other material present. On one piece of railroad construction, in 
West Virginia, of which the writer had charge of eight miles, it was 
necessary to obtain the sand from the bottoms of the streams. All such 
material had to be washed, to free it from mud and clay, and in 
order to obtain the necessary material, every bar in more than 8 
miles of the Monongahela River and Ten-Mile Creek was explored, and 
work often had to be suspended on account of high water. To have 
obtained the sand elsewhere would have required hauling it a number 
of miles by rail, and several miles by wagons, over roads which, at 
certain seasons, were almost impassable. 

With the exception of a limited territory near the coast and the 
mouths of the larger rivers or their branches, the problem is often, not 
to obtain a perfect sand, but any at all, worthy of the name, within 
a cost consistent with economical construction; and it is a fact that 
in and near New York City much sand is being rejected and con- 
demned, which would be hailed with delight by both contractors and 
engineers on many important pieces of work outside of that territory. 
The reason for this seems to be, that all the sand used near New 
York City, is compared with, and expected to be equal to, that known 
as "Cow Bay" sand, from the immense sand plants on the north shore 
of Long Island. 

This sand occurs as a bluff, from 50 to 150 ft. high, throughout 
nearly the entire length of the north shore, and, except for a few feet 
of loam and top soil, constitutes nearly all of Long Island. In the 
case of the north shore, near the bluff, even this small amount of 
soil is non-existent, as it has been eroded by the winds and moved 
inland, so that for some distance hardly any soil exists, and the only 
vegetation consists of scattered clumps of grass and wild vines. The 
bluff is usually within 100 ft. of the shore line, with a quite steep face 
and a nearly level beach. The water deepens rapidly to a depth suffi- 
cient to float small vessels and barges, and, where no piers have been 
provided, small vessels have often been loaded by "laying on" the 
beach just after high water, and "hauling off" at the next high tide, 
the vessel being loaded in the interim by wheel-barrows or wagons. 



268 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Haines. The sand itself is nearly pure quartz, a light yellow, and very uniform 
in color, and would be accepted almost anywhere on its general ap- 
pearance alone. It is not to be wondered at, therefore, that large 
sand and gravel plants have been established at which these materials 
are screened and shipped, principally by boat, for work near the tidal 
waters about New York City; or, that it is held up as a standard of 
comparison with other sands. 

The familiarity of engineers in this .section with Cow Bay sand, 
however, has largely obscured the fact that it is purely a local com- 
modity, obtainable, except at great cost, in only a limited section ; and 
this has tended to raise the standard of requirements to a point not 
consistent with economy. It is not uncommon to find in specifica- 
tions the requirement that sand "shall not contain more than 2% by 
volume, of loam (silt), when tested by shaking in a closed vessel con- 
taining water, and allowing same to settle in the vessel;" or, "meas- 
ured after pouring off the water containing the loam, and allowing 
same to settle." Even Cow Bay sand will hardly fulfill this require- 
ment, and the writer has generally found the percentage to run much 
nearer 5% than 2% with this sand. It has also caused the rejection 
of much excellent material, obtainable at less cost, which would have 
usually answered every requirement of the work. 

To attempt to enforce such a requirement is a genuine hardship 
in many sections, and answers no useful purpose. Nevertheless, the 
attempt is often made, but usually it is given up before the end of the 
work. Most bank* of pit sand or gravel are overlaid with a few feet 
of -fine material, differing in no way from that below except in the 
fact that it is much finer; and to remove the same before removing 
the sand is both costly and of no advantage, as it will usually make a 
better graded material if it is allowed to mix freely in the excavation 
before screening out the gravel, particularly if the material is strat- 
ified in the bank, as is often the case, and provided the amount of fine 
material is not too great. 

It is generally possible to find deposits of sand within a reasonable 
distance of most works of construction, but they are quite often over- 
laid with silt or soil to such a depth, that, if the bank be small, or the 
quantity required be small, the stripping of the top is necessary in 
order to avoid an excess of the fine material; and an inferior sand is 
often brought from a distance, in order to avoid delay and the cost 
of stripping the bank. Under these circumstances, the writer has 
always been unable to understand why more artificial sand is not pre- 
pared and used, as he has often seen excellent material for it close 
at hand, while inferior natural sand was being hauled from some dis- 
tance for use on the work. It would seem that a small portable pul- 
verizer, or set of rolls, might be used to advantage, especially where a 
crusher is in use on the work. There seems to be a deep-rooted pre- 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 269 

judice against the use of stone dust for sand, and on one piece of Mr. Haines, 
work with which the writer was connected, at least 1 000 cu. yd. of 
crusher dust from trap rock boulders was allowed to go to waste, and 
natural sand bought, although tests gave results for the dust equal to 
or better than the sand. 

The writer has been engaged in engineering construction for more 
than eighteen years, and during that time he has had occasion to 
pass judgment on a great many thousand yards of sand, but, while he 
has made many tests by water and screens, most of the sand has been 
accepted or rejected on the old-fashioned tests of inspection and rub- 
bing in the hand; and he does not hestitate to say that these are the 
best tests of fitness which are known to-day. No two banks, no two 
shipments from the sand bank, and no two loads from the snme ship- 
ment of natural sand are precisely alike; and, from the nature of the 
material, it cannot be expected. As a rule, a bank of sand cannot 
be passed or accepted as a whole, for there are nearly always some 
portions of it unfit for use, and the best that can be done is to accept 
or reject the sand as delivered. For obvious reasons, this requires 
that it be done, generally, while it is still on the wagon or car. A 
decision, therefore, must be given at once, and this can be done, be- 
cause any great amount of silt or top soil is readily apparent to the 
eye, and whether it is "sharp," silicious sand, whether it contains a 
large quantity of small shale pebbles or whether it is coated with clay, 
is equally apparent on rubbing it between the fingers or palms of the 
hands. 

The term "sharp" as here used does not refer to the angularity 
of the grains, and, while the term has been copied blindly for forty 
years or more, it was not originally understood in that way, nor is it 
now among mechanics using the sand. It refers to the peculiar prop- 
erty of sound produced when particles of quartz are rubbed against 
each other. A coated sand, or one containing much shale or other ma- 
terial, fails utterly under this test, and there is as much difference in 
the sound as between two different tones from a musical instrument. 
There is really little reason why sand should be angular, as the great- 
est volume for the surface to be coated, would be obtained if each 
grain were a perfect sphere. Engineers might just as well require 
cubical gravel. For years the same arguments against round sand 
were advanced against the use of gravel for concrete, but, for various 
reasons, it is now generally preferred to crushed stone. The very 
action which has produced sand has also tended to make it spherical, 
and, as a rule, the most angular sand contains also the most dirt or 
clay. 

There is one thing which cannot be determined by the rough tests 
already mentioned, and that is the amount of vegetable or organic 
matter contained in the silt, or in the sand itself. However, this can 



270 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Haines, be determined once for all for any bank by taking samples over the 
area at different points. It is properly a laboratory test, however, and 
should be made before the use of any sand from the pit is allowed. 
The amount of organic matter found by Mr. Thompson in the sand 
which caused failure appears to have been remarkably small to have 
produced such results ; the writer has used sand which he knew to con- 
tain a much greater amount without any retardant or weakening ac- 
tion being apparent. He noticed such an action once, however, when, 
for mixing mortar, he was obliged to use water from a small marsh 
where the water was stagnant and had a decided color and an oily ap- 
pearance on the surface; he concluded that it was due to some oil 
or acid from the decaying vegetation. It is possible that some similar 
action took place in the case mentioned by Mr. Thompson. It might 
also be added that the specifications of the Board of Water Supply of 
New York City, for the great Catskill Aqueduct, allow in sand "not 
more than 3% by weight of vegetable matter," and, while the amount 
seems rather large, it is probably justified on account of the character 
of the country from which some of the sand must be obtained. While 
it is doubtful if the above amount of vegetable matter has ever been 
reached on this work, some has been apparent, but no injurious action 
has been noticed by the writer. 

With concrete laid in the usual manner, with a considerable excess 
of water, a sand which contains much more silt than would be well 
to allow in sand for mortar may be safely used, as the agitation of 
the water, from depositing the concrete, keeps much of the finer silt 
in suspension, and it will pass off over the top of the forms if provision 
is made for it. If this is not done, it is deposited as a slimy mass on 
the top of the concrete, sets very slowly, if at all, and should be re- 
moved and thoroughly cleaned off before resuming work. For the 
same reason, it is practically useless to add finely-powdered dry clay 
to concrete in order to increase its density, and it is also a mistake to 
carry the grinding of cement to too fine a point, as a large quantity of 
it will be carried in suspension and deposited on the top of the work, 
wherever work is suspended. 

Mr. Lesley has called attention to the fact that while cement is 
tested in many ways, sand is vised in a proportion of three to one, 
often without test. This very difference in the volume, as well as in 
the character of the materials, accounts for the fact. A sufBcient 
quantity of cement must, of necessity, be provided to cover quite a 
period of time, and, from its nature, it must be stored under cover; 
while, in the case of sand, it is often difllicult to store any large 
amount. In fact, on large works in New York City, like the Eapid 
Transit Subway, or the Atlantic Avenue Improvement of the Long 
Island Railroad, in Brooklyn, of which the writer was in charge of 
sections, where materials have to be stored in the streets, it is often 



DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 271 

difficult to store even a day's supply, and teams are kept constantly Mr. Haines, 
hauling. Even in open country, on the Catskill Aqueduct, the writer 
has often had a large number of teams hauling to a single shaft, and 
that from different pits, in order to provide for the current day's 
work. For this reason, any extensive tests are impossible, and the best 
that can be done is to take test samples from the pit as a whole, and 
then see that the delivery conforms as closely as possible to the standard 
adopted. Test samples should be taken at times from the material de- 
livered, as well as from the mortar or concrete used; but, as the ma- 
terial will probably be in the work before the result of any extensive 
tests can be known, their main value is as corroborative evidence as to 
the quality of the materials actually used. 

Cement is a manufactured product, and can and should conform 
to fixed requirements, while sand is a natural resource, and must be 
accepted, rejected, or modified to suit requirements. The latter can 
often be nicely done by varying the amount of cement, and for that 
reason the method adopted by the Board of Water Supply of New York 
City, namely, of paying for all cement used in permanent construc- 
tion at a certain price bid by the contractors, is to be commended. 
The writer would also recommend that, in the case of large 
works, the sand and gravel rights be acquired by the companies for 
whom the work is done, and the product furnished to the contractors 
at a price just sufficient to cover the money invested, at a reasonable 
rate of interest. 

The writer believes in tests, and he also believes in specifications; 
but, while a few years ago the latter were so general in character as 
to be of little use, they are now often so voluminous as to be cumber- 
some, and consequently little read. Also, at times, they appear to 
have been written by idealists, without any regard to the natural re- 
sources of the country from which the supplies must be obtained. 

As "David Harum" says: "A reasonable amount of fleas is good 
for a dog, it keeps him f'm broodin' on bein' a dog." To paraphrase, 
the writer would say that a reasonable amount of impurities in sand 
is good for the sand : It keeps it from being pure quartz, which, in the 
United States at least, is very hard to obtain. 

Sanford E. Thompson, M. Am. Soc. C. E. (by letter). — The dis- Mr. Thompson, 
cussion has brought out a number of the very interesting points which 
one encounters in the study of sand for mortar and concrete. The 
difficulty of judging the quality of the sand by the eye, and the conse- 
quent necessity for actual tests, is emphasized by Mr. Humphrey in his 
comparison of two sands, the tests of which proved the sand apparently 
poor to be the better. The comparatively small benefit from sharpness 
of grain, a point insisted upon by the writer for several years, is coming 
to general recognition. 

In connection with Mr. Mills' statement with reference to the tests 



272 DISCUSSION ON IMPURITIES IN SAND FOR CONCRETE 

Mr. Thompson, of the sand from the Philadelphia Subway, it may be added that while 
the total sand contained 0.4% of organic matter, the percentage of 
organic matter in the silt washed from the sand was 6%, or less than 
the 10% proposed as a maximum allowable limit in the opening dis- 
cussion. As there suggested, it appears that both the percentage of 
organic matter in the total sand and the percentage in the silt must 
be considered. Of course, it must always be borne in mind that the 
percentages of organic and vegetable matter referred to are based on 
ignition tests, and not on mechanical separation of the sand. 

Mr. Mills' question as to the practical advantage of introducing fine 
material into concrete for the purpose of making it more water-tight 
is an extremely important one. As he suggests, local conditions must 
aifect the conclusion, but the character of the structure also is a most 
important consideration. On a small job, such as a tank, the cost of 
forms, reinforcement, and incidental expense is so large that an increase 
in the quantity of cement will form but a small percentage of the total 
cost, and rich proportions can be used economically. On the other 
hand, on a large job like a dam or a tunnel, the cost of the cement is 
a very important factor, and the introduction of a small quantity of 
fine material for the purpose of using leaner proportions may be of 
great advantage. If the strength is of small consequence, and the pro- 
portions of cement can be reduced by the addition of some inexpensive 
material like fine sand, hydrated lime, or clay (provided it is pulverized 
so finely that it can be introduced without danger of forming balls in 
the concrete), so as to make the proportions 1:4:8 instead of 1: 2^: 5, 
nearly i bbl. of cement will be saved in every cubic yard of concrete. 

Mr. Richardson brings out the variations which may be produced 
by different grading of sand. It is hoped that he will publish shortly 
the results of the valuable series of tests made during the construction 
of the Wachusett Dam. 

Although the writer agrees with many of the points brought out by 
Mr. Haines, he must take very decided exception to the view that the 
best tests of fitness of sand known to-day are "the old-fashioned tests 
of inspection and rubbing in the hand." One example out of many 
which might be cited as indicating the incorrectness of this view is 
described in the opening discussion. This particular sand was in- 
spected and approved, not only by two or three practical builders, but 
also by two engineers acting independently, both of whom had had 
years of successful experience as specialists in concrete construction ; 
and yet the 1 to 3 mortar made from this sand had a tensile strength, 
at 7 days, ranging from zero up to 40 lb. per sq. in. The quality of 
the mixing water, as Mr. Haines suggests, occasionally affects the 
hardening of the concrete, but, in the case referred to, the concrete 
failed to harden on the job when mixed with water from the town 
water-works, and the mortar also failed to attain any strength in the 



DISCUSSION ON IMPUIUTIES IN SAND FOR CONCRETE 273 

laboratory when mixed with Newton City water, and made up with Mr. Thompson, 
three different brands of cement, while mortar of standard sand made 
with the same water and the same cement attained normal strength. 

From many such experiences as the one cited, it is difficult to avoid 
the conclusion already indicated, that for important concrete construc- 
tion, especially for reinforced concrete work, unless from a bank of 
known good quality which has been tested previously, tests of the sand 
are as necessary as tests of the cement. 

Finally, it may be well to call attention once more to the two 
qualities which are of the greatest importance in selecting sand, namely, 
the cleanness and the gradation of grains. From the preceding remarks 
it is evident that: 

(1). — Sand for concrete must not contain vegetable matter or other 
ingredients which will prevent or delay the setting and hardening of 
the concrete. 

(2). — Sand for concrete must be selected and the proportions of the 
mixture chosen with a view to obtaining the strength or the water- 
tightness required for the given structure. 

Tests such as those described by the writer in the opening discussion 
indicate that occasionally a sand may be graded properly in size of 
grains, and yet may be unfit for use because of the injurious matter it 
contains. On the other hand, a perfectly clean sand may be so fine 
and so poorly graded that its use is uneconomical or sometimes even 
dangerous. 

Comparative tensile tests of mortar made from the sand in question 
and of standard sand mortar mixed in the same proportions at the same 
time — as recommended by the Joint Committee on Concrete and Rein- 
forced Concrete — are of the greatest importance as showing in a 
general way whether or not the sand is safe to use. The tensile test 
will not distinguish the cause of any irregularities, nor will it show 
wherein the quality of the sand may be improved for strength or water- 
tightness. 



AMEEICAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1126 

FIRE-KESISTANT CONSTRUCTION OP BUILDINGS. 

An Informal Discussion at the Annual Convention, July 8th, 1909. 



By Messrs. Herbert M. Wilson, R. W. Lesley, Richard L. 
Humphrey, and S. Whinery. 



Mr. Wilson. HERBERT M. WiLSON, M. Am. Soc. C. E. (by letter). — The inventory 
of natural resources prepared by President Roosevelt's Conservation 
Commission has opened new lines of thought and investigation to the 
Engineering Profession as well as to other departments of the social 
fabric. 

Architects and engineers have realized in a general way the waste 
of property vakies involved in building and engineering construction 
not designed to resist such destructive agencies as fire and earthquake. 
A cursory examination, however, of the statistical results presented 
to the Mineral Section of the Conservation Commission shows the 
character of the waste by differentiation into its components of lumber, 
masonrjs steel, and miscellaneous building material. 

As many of the members of this Society well know, the Conserva- 
tion Commission was divided into four sections, charged with the 
study of the waste of natural resources, as follows: (1) Water; (2) 
Forests; (3) Land, and (4) Minerals. The Federal Commission was 
fortunate in having at its disposal the accumulation of more than a 
quarter of a century of research by Government Bureaus charged with 
the investigations of the problems under consideration, foremost among 
which were the Geological Survey, the Forest Service, and the General 
Land Office. 

The Secretary of the Mineral Section of the Conservation Cora- 
mission is Dr. Joseph A. Holmes, Expert in Charge of the Technologic 
Branch of the Geological Survey, and to the writer, as Chief Engineer 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 275 

of that branch, was assi^ed the inquiry into the waste of mineral Mr. Wilson, 
resources involved in building and engineering construction. He is 
indebted to Dr. Holmes and to the Director of the Geological Survey, 
Dr. George Otis Smith, for permission to present to this Society some 
of the results of this investigation. 

The Supervising Architect of the Treasury Department has under 
his control public buildings costing more than $200 000 000, and is 
spending each year more than $20 000 000 in the construction of new 
buildings. At his behest, in order that he may design these buildings 
60 as to render them most resistant to fire and earthquake at a minimum 
expenditure, inquiries are now' being conducted in the field by economic 
geologists skilled in the study of the building stones, gravels, sands, 
and clay deposits of the country which are most suitable and most 
cheaply accessible to the three hundred cities scattered throughout the 
United States in which public buildings are now under construction 
or projected. The samples are transmitted to the Pittsburg Laboratory 
of the Geological Survey for analysis and test as to their qualities 
either as aggregates for concrete, or building stone or brick. 

In co-operation with the Underwriters' Laboratory at Chicago, and 
independently at the Government Testing Laboratory at Pittsburg, in- 
vestigations are being conducted in reference to the rates of conductivity, 
and the fire-proofing and fire-resistant properties of various structural 
materials, including strength under load before and after subjecting to 
fire. Tests of the crushing strength of large samples of building stones, 
of beams, columns, and floor-slabs of concrete and reinforced concrete, 
and of building brick and tile, are likewise being conducted in order 
to determine which will give the most efficient results in building con- 
struction of the highest type. This feature of the inquiry has in view 
the fact that the Government should maintain as high a standard of 
efficiency in the construction of public buildings as is commensurate 
with economic design, since such buildings are intended to be of a 
permanent and enduring nature. The magnitude and importance of 
this work are such that the Government cannot afford to take any risk 
concerning methods of construction or materials to be used. The fact 
that the Government does not insure its buildings against loss by fire 
makes it especially necessary that it be provided against such loss by 
making them fire-proof. It is estimated that if the pxxblic buildings 
of the United States were insured, the cost to the Government would 
be more than $600 000 annually. 

A startling fact, developed in considering the value and fire risks of 
Government buildings, is that the total cost of fires in 1907 amounted 
to almost one-half the cost of new building construction in the country 
for the year. The cost of building construction in forty-nine leading 
cities of the United States reporting a population of less than 
18 000 000, amounted to $661 021 286, and the cost of building construe- 



276 DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 

Mr. Wilson, tion for the entire counti-y in the same year is conservatively estimated 
at more than $1 000 000 000. A preliminary inquiry had shown that 
the cost of fires to the people was nearly $500 000 000, for the year 1907. 

The preliminary estimate that nearly one-half the value of all new 
buildings constructed in the United States in one year was destroyed 
by fire, was sufficient warrant for a thorough statistical inquiry, vphich 
would develop the facts for the whole United States, in connection 
with the specific inquiry for the Government regarding fixe-proof and 
earthquake-resistant construction. This inquiry included the value of 
the property destroyed by fire, the maintenance of fire departments, the 
payment of insurance premiums less the benefits returned, protective 
agencies, and the additional cost of city water supplies, etc. It has 
developed the fact that the total cost of fires in the country in 1907 
amounted to more than $456 485 000, or about $1 250 000 daily. This 
cost was five times as much per capita as in any country of Europe, 
and was equivalent to a tax on the people exceeding the total value 
of the gold, silver, copper, and petroleum production. This fire cost 
was greater than the true value of the real property and improvements 
in any one of the following States: Maine, West Virginia, North 
Carolina, North Dakota, South Dakota, Alabama, Louisiana, or 
Montana. This means that the total destruction of the real property 
in any one of these States in a year would not represent a loss greater 
than that caused by fires in the whole United States. 

The actual fire losses due to the destruction of buildings and their 
contents amounted to $215 084 709, or a per capita loss for the United 
States of $2.51. 

These actual fire losses are greater than the true value of the real 
property in the States of Utah, Delaware, Florida, Idaho, Wyoming, or 
Nevada. If a storm or earthquake devastated any one of these States 
completely, there would be a great outcry. Assistance would be rushed to 
the stricken people; a mighty attempt would be made to find a remedy 
for the conditions, and the activity would not cease until the remedy 
had been found. But, with the fires scattered throughout the country 
and the loss distributed among hundreds of cities and towns, there is 
no appreciation of the seriousness of the conditions, and very little 
attempt to better them. In many cities a big fire loss for the year is 
looked upon complacently, as if it were a natural and unavoidable 
condition. The significance of these figures lies in the fact that the 
loss in the cities of the six leading nations of Europe amounted to 
33 cents per capita, or about one-eighth of the above. In addition to 
this waste in wealth and natural resources, 1 449 persons were killed 
and 5 654 injured in fires. As convincingly set forth, at the recent 
conservation meeting of the Engineering Societies in New York City, 
by Mr. Charles Whiting Baker, the buildings consumed, in the year, 
if placed on lots of 65-ft. frontage, would have lined both sides of a 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 277 

street extending from New York to Chicago. A person journeying along Mr. Wilson, 
this street of desolation would have passed in every thousand feet a 
ruin from which an injured person was taken; at every three-quarters 
of a mile in this journey he would have encountered the charred 
remains of a human being; and, when burned out, at the end of a 
year, the fire would begin all over again on an even longer street. 

An analysis of these fires losses, with a view to making the results 
of the structural materials investigations useful, not only to the 
Government, as the greatest consumer of such materials, but also to 
the various States and municipalities, and to the people of the country 
as a whole, showed that it would be desirable to point out, not only 
the most efficient methods of construction for each locality, but also 
how such improved construction might be brought within the cost 
limits of all the people. In this analysis, the fact stands out promi- 
nently that much of the fire waste is due to fires extending beyond 
the limits of the buildings in which they started. Exact figures as to 
the losses due to exposure were not obtainable, but the most conserva- 
tive estimate indicates that at least 27% of the losses resulted from 
fires extending beyond the building of origin. Thus nearly one-third 
of these losses are due to the inflammable construction of American 
buildings, for in Europe, where fire-resistant construction prevails and 
building regulations are designed and enforced with a view to restrict- 
ing the progress of a fire, there is practically no loss from exposure 
fires. And it is even more significant that only $68 000 000 of the 
loss from fires in the United States was in brick, concrete, and other 
slow-burning construction, while double that amount, or $148 000 000, 
was in frame buildings. 

The investigation into the actual cost of fires in the United States 
was quite thorough, and included a complete inquiry into the total 
cost of water-works systems and the part of the construction and equip- 
ment, the annual expense, depreciation, taxes, interest, and mainte- 
nance costs chargeable to fire service; also the total cost of fire depart- 
ments, the annual expense, depreciation, taxes, interest, and mainte- 
nance charges. The results showed $145 604 362 as the amount of fire 
premiums paid to the insurance companies above the losses paid by 
them ; $28 856 235 as the total annual expense of the water- works up- 
keep chargeable to fire service ; $48 940 845 as the total annual expense 
of fire departments ; and $18 000 000 as the total annual cost of fire 
protection. 

The results of this general inquiry into the causes of fires are 
summarized in Table 1. 

A careful examination of Table 1 tells the story to the engineer. 
In studying it, he should bear in mind that, in the last thirty years, 
the total fire waste, being only the value of property destroyed, 
amounted to $4 484 000 000. 



278 



DISCUSSION ON FIRE-EESISTANT CONSTRUCTION 



Mr. Wilson. ^^^ ^ 




00 

o 

8 

o 






00. 

at- 






00 











fo .. ^ 
B g C 
a: d I© 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 



279 



The general inquiry included not only the cost of fires in the United Mr. Wilson. 
States, but also the cost in Europe, including fire waste, excess of 
premiums over insurance paid, annual expense of city water-works and 
fire departments chargeable to fire service, and private fire protection 
in a number of countries. From the results obtained, it is evident that 
the United States is paying annually a preventable tax of more than 
$366 000 000, or nearly enough to build a Panama Canal each year. 
The estimate gives the total cost of fires in the United States, if build- 
ings were as nearly fire-proof as they are in Europe, as $90 000 000. 

The figures are set forth in Table 2. 

TABLE 2. — Annual Loss and Outlay in the United States, on 
Account of Fires; With Comparison Showing Probable Annual 
Loss and Expense, if Buildings Were as Nearly Fire-proof 
AS in Europe. 





United States, 
1907. 


Per 
capita. 


United States, 

if buildings 

were as neai-ly 

fire-proof as 

in Europe. 


Per 
capita. 


Total loss bv fire 


$215 084 709 
145 604 36a 

28 856 235 
48 940 845 
18 000 000 




$41 000 000 
28 000 000 

epoo 000 

10 000 000 
5 000 000 




Excess of premiums over insurance paid. . 

Annual expense of water-works, chargeable 

to fire service 




Annual expense of fite departments 

" " " private fire protection.. 




Total fire waste 


$456 486 151 


$5.34 


$90 000 000 


$1.05 






Total loss by fire 


$215 084 709 
241 401 442 


$2.54 
3.82 


$41 000 000 
49 000 000 


$0 48 


Annual expense of fire protection 


0.57 







In Table 2, and more particularly in the following discussion, it 
should be borne in mind that, while fires will continue to occur in 
buildings and destroy them and their contents to a greater or less 
extent, the fire which extends beyond the building of origin and to its 
neighbor, and thereby produces a conflagration, does so only because 
the building construction is of the most inflammable nature. The 
costs which are here recorded, and exceed in amount those due to fire 
waste, are preventable only in so far as they are an incident to con- 
flagration conditions as distinguished from fires restricted to the build- 
ing of origin. The statements set forth relative to the possible reduc- 
tions in the cost of fire departments, city water supply, and private 
fire protection charges, must be considered only in connection with 
those service charges which are due to combatting ever-present con- 
flagration conditions. 

The insurance and property loss in the destruction of the City of 
San Francisco by fire and earthquake in 1906 amounted to about 



280 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 



Mr. Wilson, 



$315 000 000. The area burned was 3 000 acres, embracing 520 city 
blocks, on which 25 000 buildings were destroyed, and it is notable that 
of these only 3 000 were of brick or stone, and 22 000 were of frame 
construction. 

Since 1866, the conflagrations of the United States have cost, in 
fire waste alone, $936 551 135. This loss, according to European stand- 
ards, is needless, and is preventable. If, as it has been shown, the 
actual cost of fire protection in all its phases is more than the fire 
waste, it is fair to assume that more than $1 000 000 000 has been ex- 
pended in this period in an attempt to take care of these conflagra- 
tions. This, also, is an unnecessary expenditure, if European standards 
are considered, and makes a total tax from conflagrations alone since 
1866 of more than $2 000 000 000. This is true because the conflagra- 
tion loss in any of the European countries is almost nothing. It is 
indeed rare when a fire extends beyond the building in which it starts. 

Notable conflagrations in the United States have been those of 
Chicago, in 1871; Boston, in 1872; Baltimore, in 1904; and San Fran- 
cisco, in 1906. The year 1908 has a conflagration record exceeding that 
of 1907, one such disaster alone, that of Chelsea, Mass., involv- 
ing an insurance loss, not taking into account the incidental losses, of 
$8 846 879. 

The analysis of the fire losses in the United States in 1907 is given 
in Table 3. 

TABLE 3. — Fire Losses in the United States for 1907. 

Statistics Gathered by the United States Geological Survey. 





Total. 


Urban. 


Rural. 


Total fire loss. 


$215 084 709 
109 156 894 
105 927 815 

68 425 267 
31 092 687 
37 332 580 

146 659 443 
78 064 207 
68 595 235 

165 257 
36 140 
129117 

2.51 


$107 093 283 
50 173 625 
56 919 658 

48 908 744 
19 816 474 

29 092 270 

58184 539 

30 357 151 

27 827 388 

105 406 

25 297 
80 109 

2.54 


$107 991 426 


Fire loss: buildings 


58 983 269 


Fire loss: contents 


49 008 157 


Total, brick, etc., fire loss 


19 516 523 


Brick, etc. : Are loss, buildings 


11 276 213 


Brick, etc.: fire loss, contents 


8 240 310 


Total frame flre loss 


88 474 903 




47 707 056 


Frame fire loss: contents 


40 767 847 


Total number of fires 


59&51 


Number of fires in brick, etc., buildings.. 
Number of fires in frame buildings 

Loss per capita 


10 843 
49 008 

2.49 







From Table 3 it appears that the loss is rather evenly divided 
between urban and rural population. The large losses in cities and 
villages are not surprising, for not only do they contain many inflam- 
mable buildings filled with valuable property, but their proximity 
renders them particularly subject to exposure risks. The equally large 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 



281 



losses in rural districts, wliere buildings are widely separated, can only Mr. Wiis 
be explained by the assumption that the remarkable efficiency of the 
city fire departments prevents even greater losses in urban fires, and 
that the buildings and contents in the rural districts are generally 
totally destroyed when fire breaks out. This is indicated in the total 
of the building losses, which amounts to more than $50 000 000 in 
cities, while in rural districts the loss in buildings, as distinguished 
from contents, is nearly $60 000 000. The contents loss in cities 
amounts to nearly $57 000 000, as against $49 000 000 in rural districts, 
though the' contents of city buildings amount in value to several times 
those in rural districts. The much heavier losses in masonry buildings 
in cities as compared with those in the country are due to the rarity 
of buildings of this character in rural districts. The losses in frame 
buildings in cities and villages, amounting to more than $30 000 000, 
as compared with losses in buildings of the same character in the 
country, amounting to nearly $48 000 000, can be attributed to the 
efficiency of the fire departments and the utter lack of such protection 
in rural districts. In European cities, the construction of frame 
buildings is absolutely prohibited in all municipalities, and few are 
erected in rural districts, owing to the relatively high price of timber. 
In the United States, the conditions are the reverse. Until recently 
timber has been the cheaper, as well as the more easily worked, mate- 
rial, as compared with brick, stone, and steel. More than two-thirds 
of the total loss in the United States in 1907 was due to frame build- 
ings. The exact losses were : $146 695 442 in frame buildings, and 
$68 425 267 in masonry and steel buildings. 

As shovpn in Table 4, in those States in which the timber supply is 
abundant, there is an increased per capita loss of 59 cents over the per 
capita loss in the comparatively treeless States. 

TABLE 4. — The Per Capita Fire Loss for 1907 in Eleven States 
Where Timber Is Scarce, and in Eleven States Where Timber 
Is Plentiful. 

Statistics Gathered by U. S. Geological Survey. 





Total 
population. 


Total 
Are loss. 


Per 

capita. 


Group 1, States having scarcity of timbers: Iowa, 
111., Okla.. CoDn., Del., N.'J., S. Dak., R. I., 
Kans., Nebr. . and N. Dak 

Group 3, Timbered States: Wash,, La., Tex., Miss., 
Wis., Ark., Mich., Pa., Minn., Ore., and N. C... 


16 785 460 
35 569 533 


$38 606 558 
73 895 950 


$3.30 
2.89 



Studying these fire losses by geographical divisions of States — a 
division generally used by the Census Bureau — as set forth in Table 
5, a remarkable feature is the large per capita loss in the Southern 



283 



DISCUSSION ON" FIRE-EESISTANT CONSTRUCTION 



Mr. Wilson. States, namely, $3.66, or more than $1.00 in excess of the per capita 
loss in any other division. The cause lies in the fact that the Southern 
States are well timbered, and, in addition, suffer from the handicap 
of inefficient fire protection in the cities and villages. 

TABLE 5. — The Per Capita Fire Losses for 1907 in the United 
States, as Shown by Geographical Divisions of the States. 

Statistics Gathered by U. S. Geological Survey. 



Geographical Division. 


Total 
population. 


Total 
fire loss. 


Fire loss 
per capita. 


North Atlantic: Me., N. H., Vt., Mass., R. I., Conn., 
N. Y., N. J., and Pa 


33 779 018 
11 574 988 
29 036j645 
16 368 558 
4 783 557 


$59 447 532 
25 349 323 
68 793 148 
59 908 922 
12 676 426 


$3.50 


South Atlantic: Del., Md., D. C, Va., W. Va., N. 
C, S. C, Ga., and Pla 


3.19 


North Central: Ohio, Ind., 111., Mich., Wis., Minn., 
Iowa, Mo., N. Dak.. S. Dak., Nebr., and Kans.. 

South Central: Ky., Tenu., Ala., Miss., La., Tex., 
Okla.. and Ark 


3.87 
3.66 


Western: Mont., Wyo., Colo., N. Mex., Ariz., Utah, 
Nev., Wash., Ore., and Cal 


2.65 







Statistics gathered by United States consular officers in Europe, at 
the instance of the National Board of Fire Underwriters, showing fire 
losses in six of the most prominent of the European countries, give 
the surprisingly low per capita loss of 33 cents. Had the United States 
such a per capita loss, instead of one of $2.51 for a total estimated 
population for 1907 of 85 532 761, the total fire waste would have 
amounted to only $28 623 290, or a saving of natural resources from 
fire, alone, to the extent of $186 461 419. 

These facts are set forth in Table 6, which shows that the average 
loss per capita is $0.33, as compared with $2.51 in the United States. 

TABLE 6. — Fire Losses in Six European Countries. 
Statistics Gathered by The National Board of Fire Underwriters. 



Country. 


Years. 


Annual aver- 
age 
fire loss. 


Population, 
1901. 


Loss per 
capita. 




1898-1902 
1901 

1900-1904 
1902 

1901-1904 

1901-1905 


$7 601 389 

660 924 

11 699 275 

27 655 600 

4 112 735 

999 364 


20 150 597 

2 588 919 
38 595 500 
.56 367 178 
32 449 754 

3 325 023 


$0.29 




0.26 




0.30 




0.49 


Italy 


0.13 




0.30 







Table 7 is a comparison of America's needless waste and Europe's 
prudence by grouping the per capita losses in cities of approximately 
the same size in the two countries. 



DISCUSSION ON FIRE-EESISTANT CONSTRUCTION 



283 



H t-i 



W 1^ 2 td 50 1^ 'tJ 



o" 3 



5" O 



M 05 Oi O O 

CO >-* 05 CO »-• 



Cn QD 05 



o ^:> 05 ex m 

CTT CO I-* CO CO 



to 10 ^ OS 



to 05 i-i 10 to 



-1 05 05 U» 

fo -J 00 en 



H O tsl CO (> g 6d 



? a> & 



63- S ^- -r 



^ '" 



2 22 



9 g I 



iTl 





tr' 





S3 





& 


c 


15 


w 


►d 


p 




^- 


p 







ff 



CO CO c;» 



M- en to trt to O 
COQ05DiOtOl-''-'tO 
lOOOJtOOlCnOSCO 

oc;»»ih*>to>— gjcji 



CO to -1 



Mr. Wilson. 



CO 



Q 



^ \-.- 



m 






^^ 



0) 


3 


1-j 




(C 


M 


p.. 





,q- 


M 


«<^ 


H 



c^ : 



• 


P> 





g 


ro 


M 





» 












iq 







!z! 


pa 




1 — ' 


> 


CO 


52) 

n 


p: 




:^ 


W 


,^5 


a 


^ 


w 







£0 


►ti 


b 


H 


&> 


> 


bd 


!2J 


i-t 





CD 


3 


p: 


M 




W 







l-b 





^ 
P 




t-f 


a 


d 


w 


I-+5 




P 


Ifi 





> 



w 



I-" to CO 



to 05 tn 



284 DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 

Mr. Wilson. Realizing that the loss from fires due to the destruction of property 
involves only a fraction of the total loss, a careful statistical inquiry 
was made into the additional cost of water supplies necessary for com- 
bating conflagrations. The inquiry developed the fact that in 1907 
the water-works systems of the United States were valued at 
$1 129 247 532, the cost chargeable to domestic service being accepted 
as $883 575 85C, and that chargeable to fire service $245 671 676, or 22% 
of the whole. 

In these systems the pipe used amounted to 7 097 800 tons, of which 
5 080 873 tons were used for domestic service and 2 016 927 tons for fire 
service. The total number of hydrants was 420 394, and the number of 
hydrants in fire service, 350 152. The cost of hydrant fire service was 
$29 761 400, and the value of established or contemplated high-pressure 
fire systems amounted to $22 191 388. 

The facts covering the property values involved in additional water 
supplies for combating fires were secured after much difiiculty, and, 
as any engineer must know, are neither as complete nor as satisfactory 
as could be desired. To engineers and superintendents of water-works, 
5 700 report blanks were mailed and an extensive personal correspond- 
ence was taken up in order to secure data which were not furnished on 
the first call in such condition as would make them useful in the inves- 
tigation. Only a small percentage of the replies were complete enough 
for use in tabulation. Under these circumstances, recourse was had 
to the total cost of water supplies, as secured by the census for cities 
having a population of 30 000 and more and owning water-works. Data 
for cities having a population of less than 30 000 were obtained from 
the "Spectator's Year-Book." These data were then segregated to 
show the total cost under five geographical divisions of the United 
States, and, finally, the total cost in each of these geographical divi- 
sions was subdivided in order to show the total cost in cities of four 
classes of population, as follows: Cities of less than 5 000; between 
5 000 and 30 000; between 30 000 and 100 000; and with more than 
100 000 inhabitants. 

By adopting the same plan of classification in tabulating the reports 
received from engineers and superintendents of water-works, who gave 
detailed information, it became possible to apply percentages to the 
total cost of water-works systems for the corresponding geographic divi- 
sions and classifications by sizes of cities, and, from these, the results 
given were obtained. The geographical divisions are arranged in 
accordance with the different requirements in the various localities, 
and are the result of personal inspection and interviews with hydraulic 
engineers and superintendents of water-works in a number of typical 
cities throughout the United States. In the prosecution of this in- 
quiry the writer received much valuable advice and assistance from 
many members of this Society throughout the United States, several 



DISCUSSION ON FIEE-RESISTANT CONSTRUCTION 285 

of whom spent long hours in going over their plans for distribution Mr. Wilson, 
in the cities under their charge or with the operation or design of 
which they were connected. In this manner, only, has it been possible 
to arrive at any reasonable approximation as to the additional cost of 
water supply and water distribution made necessary by having a reserve 
for fire fighting under conflagration conditions. 

This inquiry was not based on the idea that water-works are consid- 
ered for two purposes of equal importance, viz., domestic water service 
and fire-protection service, but on what it would cost for water-works 
systems for fire protection alone. The governing idea has been to ascer- 
tain, in any city having a modern water- works system, what it would have 
cost, in the matter of storage or pumping capacity, in the extent and 
dimensions of the distributing system, and in the number of hydrants, 
should the city design them anew with a view to supplying only 
domestic needs and such minor service as would permit fire fighting, 
with a view to confining the fire to the building of origin, with the 
latter built of fire-resistant materials. In the New England and Mid- 
dle States much consideration has been given to the subject of increas- 
ing water supplies which were not originally planned to meet the 
enormous growth in population which has taken place in the last two 
decades. Pumping service and distributing systems in the hearts of 
such cities have had to be entirely replaced with larger primary sup- 
plies, and in some have had to be duplicated. In a few cities a separate 
high-pressure fire service has been installed. On the other hand, through- 
out the Southern States, this feature of increasing existing supplies 
has been given little consideration. In the more modern cities of the 
Central States, the city service was originally designed in the light of 
the most modern needs and practice. In the Far Western States, 
especially in the small places, the requirements of domestic supply, on 
account of irrigation of gardens and city yards, preponderates to such 
an extent that practically none of the water supply has been considered 
as of account for fire protection, the domestic supply being of such 
magnitude as to take care, not only of ordinary fires, but also of con- 
flagrations. This is less true in cities with populations of more than 
30 000. In the humid regions the domestic requirements are so great 
in cities of 250 000 or more inhabitants, that but a small percentage 
of the water supply need be considered as of account for fire protection. 
For these reasons the subdivision was made by population as well as by 
geographical distribution. 

IVTuch careful consideration has been given these matters from time 
to time in the past especially by J. T. Fanning, M. Am. Soc. C. E. In 
a paper on "Distribution Mains and Fire Service" before the American 
Water Works Association in 1892, he considers that domestic con- 
sumption for all purposes may be treated uniformly as 66§ gal. per 
capita per day, as compared with the very much greater per capita supply 



286 DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 

Mr. Wilson, fumished in the large cities on account of fire service and domestic 
consumption. For a city of average population, say 100 000, lie adopted 
a rate of flow for fire service of 10 000 000 gal. per day, and for com- 
bined domestic and fire service, 16 666 666 gal., or about 60% of the 
water supply on account of fire service. In a later paper, before the 
same Association, at its Twenty-sixth Annual Convention, Mr. Fan- 
ning furnished some additional information on this subject, in which 
he found that the portion of the main required for the domestic flow 
at the times when there is also a fire flow, and for different populations, 
was as follows : for 5 000 population, 11.24% ; for 10 000 population, 
19.52%; for 25 000 population, 43.75%; for 50 000 population, 57.14%; 
and for 100 000 population, 88.89 per cent. 

The outcome of the investigation by the Geological Survey indicates 
that, on the average, for the whole United States, 22% of the total 
expenditure on behalf of public water supplies, or less than one-fourth 
of the total cost of water-works systems, is due to the additional sup- 
plies necessary for protection against fires of such magnitude as may 
be propagated beyond the building of origin. The range in this addi- 
tional cost has been found to be practically from zero up to 60% in 
certain classes of small modem cities. 

It will readily be seen that this great cost of water-works charge- 
able to fire protection could be obviated to some degree if fire-proof 
construction were as good in the United States as it is in Europe, and 
to that extent the additional expense and use of necessary materials 
are waste. 

The losses in natural resources due to fire waste are not the only 
preventable losses. Those resulting from earthquake are also great in 
amount and are equally preventable by the adoption of proper design 
and materials in the construction of public and private edifices. Re- 
cently, commenting on the most terrible earthquake disaster of modern 
times, that at Messina, Italy, the Italian Ambassador, Baron Edmondo 
Mayor des Planches, stated: 

"My Government has many things for which to thank the people 
of America since the earthquake in Southern Italy besides the great 
monetary assistance and the sympathy we received. We are grateful 
for the scientific information we have received and the invaluable hints 
which have been given us by the United States Geological Survey. 
This information will be of more lasting benefit to Italy than even 
the great sums of money which have been sent to our people." 

This statement of the Italian Ambassador was based on the in- 
formation given to his country in a bulletin relating to the San 
Francisco disaster, the work of Richard L. Humphrey, Frank Soule, 
and John S. Sewell, Members, Am. Soc. C. E., for the Structural 
Materials Division of the Geological Survey. 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 287 

Possible means of reducing the enormous waste due to fires, con- Mr. wiison. 
flagrations, and earthquakes, are: 

(1). — By tests and investigations to determine the relative fire- 
resisting properties of building materials, the relative rates of heat 
conductivity of such materials, and the development of systems of 
construction vphich offer the maximum resistance to fire. These tests 
should have in view the classification of the order of merit of the 
building materials, and the cheapening of cost of construction by the 
use of those best suited to the purpose, since cheaper materials are 
not often used through lack of knowledge of their availability and 
properties. 

(2). — By the dissemination of information regarding the less in- 
flammable materials of construction, their strength and durability, the 
methods of utilizing them in building construction, and the availability 
of the most suitable of these materials near the locality in which they 
are wanted. 

(3). — By tests and investigations at the sites of such disasters as 
those at Messina and San Francisco, and by dissemination of informa- 
tion relative to the non-inflammable building materials, and the best 
method of designing and assembling them to fit them to withstand 
such shearing and torsional strains as they may be subjected to by 
earthquake. 

(4), — By the enactment of building codes and the enforcement of 
the same with the view to securing more fire-resistant and more fire- 
proof construction. In European cities the erection of wooden build- 
ings is prohibited, and the oversight of brick, stone, steel, and cement 
construction is such as to diminish danger from fire due to defective 
flues, poor electrical wiring, etc., and to confine the fire to the building 
in which it originates. 

It is evident that something must be done to stop the unnecessary 
waste of structural materials. Certain of these materials, such as wood 
and iron, are not inexhaustible by any means, but are even approaching 
exhaustion. In order to obtain the best use from these materials in 
the future, they must be used with a less lavish hand. Waste means 
increased cost in the very near future. 

The known supplies of high-grade iron ores in the United States 
are estimated at 3 840 000 000 tons, and unless the present increasing 
rate of consumption is curtailed, they cannot last beyond the middle of 
the present century. There are, in addition, 59 000 000 000 tons, or 
nearly twenty times the amount of low-grade iron ore, which un- 
doubtedly will be used when the conditions of the market warrant it. 
To increase the life of these iron-ore supplies, it is evident that the 
people of the United States must soon turn to concrete- making ma- 
terials, brick, tile, and other clay products, and building stones, as 
substitutes for the more perishable timber and the more limited metal 



288 DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 

Mr. Wilson. Supplies, rrom the above a study of the causes of waste of structural 
materials is evidently of prime necessity. The first source of such 
waste has been shown to be fires. A second source, and one closely 
related to fire losses, is that due to waste of iron and steel placed 
underground in city water mains or in pumping plants, on account of 
fire and conflagration protection. A third source of such waste is that 
due to improper mining methods, whereby the whole product is not 
extracted; and an even greater source of waste is that due to improper 
or improvident methods of preparing the minerals for market both in 
shipment and in the metallurgical processes. Perhaps the largest waste 
of structural materials is that due to lack of knowledge of their 
strength. Engineers and architects adopt working stresses for con- 
crete and metal construction from one-fourth to as high as one-eighth 
of the supposed working strength of the material. This means that 
from three to six times the necessary quantities of such materials may 
be used in structures, and, as a result, present systems of building 
construction are wasteful. 

In order that this waste may be brought to the attention of the 
public and the law-makers, and in order that the discrepancy in cost 
between wood and less inflammable materials of construction may be 
reduced, thus encouraging the use of non-inflammable materials, the 
investigations on which the above data are based were undertaken. 
It is believed that through dissemination of information as to the local 
availability of cement-making materials, of gravel and sand suitable 
for concrete construction, of clay suitable for brick- and tile-making, 
and through tests and investigations which will show the most appro- 
priate method of mixing and proportioning these materials, and of 
designing them with the minimum amount of each material which may 
suffice its purpose, will the cost of construction be reduced and the use 
of such materials be encouraged. 

Within the past few years marvelous strides have been made in the 
substitution of iron and steel for wood, due to the investigations of 
engineers, physicists, and chemists into the properties of these materials 
and the great amount of attention given to their fabrication by manu- 
facturers and architects. More recently the engineering and technical 
professions have advanced to a great extent the uses of cement in con- 
crete manufactures, but in a vastly greater period practically nothing 
has been done toward ascertaining the physical and chemical properties 
and the better modes of manufacture and use of the products of clay 
and stone. With these objects in view, the Government, as the largest 
consumer of such materials, is undertaking such tests and investiga- 
tions as may develop the most suitable of these less perishable building 
materials for each particular use and locality. These tests have in 
view the establishing of the physical properties of these materials, the 
suggestion of improved methods of manufacture with a view to 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 289 

economy, improved methods of mining and marketing in order to Mr. Wilson, 
improve the quality, reducing the quantity and cost, and extending the 
life of such materials. The investigations include the assembling of 
information relative to the most fire-resisting and fire-proof forms of 
construction, the former for the prevention of conflagrations due to 
secondary or exposure fires and the latter for the prevention of the 
destruction of the buildings in v^hich the fires originate. The tests 
also include investigations to prevent the loss of materials by the action 
of salt vpater on the Isthmus of Panama and in government and muni- 
cipal sea-walls and piers, and also to prevent the loss of materials by 
the action of alkalies in the arid region, by electrolysis in cities, and 
by other destructive agencies. The average architect, engineer, or 
contractor does not take chances in the use of unfamiliar materials 
or designs, but confines himself to those of which he has knowledge 
as having proven moderately successful. Authoritative investigations, 
therefore, are in progress, with the idea of establishing the strength 
and properties of each of the materials under varying conditions in 
widely separated localities, and of pointing out the most economical 
material for each use. 

Such of these investigations as bear on the construction of build- 
ings to resist fire are divided into three groups. A series of such 
investigations is being carried on in the Structural Materials Division of 
the Geological Survey under the immediate supervision of Mr. Richard 
L. Humphrey, for cement, concrete, and reinforced concrete, and 
metal; under Mr. A. V. Bleininger for title, brick, and other clay 
products; and under Mr. E. F. Burchard for building stone. Under 
the Fuels Division, inquiries are being conducted, in co-operation with 
the National Board of Fire Underwriters, into the relative fire risk in 
the storage of such liquid fuels as gasoline, kerosene, and alcohol. In 
the Mine Accidents Division, an inquiry into the risks in storing and 
handling explosives is being conducted. 

The Structural Materials Investigations already undertaken, and, 
in part, reported on, include fire tests of forty large panels representa- 
tive of a section of a wall of a building. These panels were constructed 
of various building stones, brick, terra cotta, concrete building blocks, 
reinforced concrete, etc., and were subjected to the temperature to be 
expected in a conflagration. The concrete building blocks and the 
reinforced concrete panels were made of various grades of river and 
slag sands, various gravels, broken stones, cinder, etc. The brick 
specimens were of common, hydraulic pressed, and sand-lime brick. 
Careful record was made of the method of proportioning and mixing 
the mortar and the concrete materials. The panels were subjected 
to freezing and to hot water before and after the test and to drenching 
through the nozzle of a flre hose, when the maximum temperature had 
been reached. An effort was made to obtain the maximum tempera- 



290 DISCUSSION ON FIIIE-KESISTANT CONSTRUCTION 

Mr. Wilson, tures of conflagration conditions, namely, 1 700° Fahr. Thermo- 
couples were placed in the panels in order to determine the rate of 
temperature rise and the rate of heat transmission from the face to 
the rear of the panel. The temperature rise at the rear of the panel, 
in connection with the thickness of the latter, furnished information 
regarding the transmission of heat to the outer side of a building wall. 

The Avork of the National Fire Protection Association should be 
known to every member of this Society. This Association includes 
as active members, the American Institute of Architects, the American 
Institute of Electrical Engineers, the American Water Works Associa- 
tion, and a number of other national technical engineering societies, 
fire underwriters' associations, city fire department associations, etc., 
and the writer believes that it would be to the advantage of this 
Society to be a member of that Association, in order that closer co- 
operation might be had along the lines of work on which both are 
engaged. The associate membership of that Association and its 
officers are chosen from that rapidly increasing group of professional 
men known as fire-protection engineers, many of whom are civil engi- 
neers, members of this Society, mechanical engineers, electrical engi- 
neers, architects, testing engineers, and the more advanced members 
and officers of the National Board of Fire Underwriters. 

Through its committees and in co-operation with the Underwriters' 
Laboratories in Chicago, and the structural materials laboratories of 
the United States Geological Survey, valuable data are being collected 
and important reports issued on the best methods of fire-proof building 
construction. These reports are discussed at the annual meetings, and 
result in the preparation of specifications on such subjects as fire- 
proof construction, concrete and reinforced concrete in building con- 
struction, the proper materials to be used in the construction of gravity 
sprinkling tanks, the proper design of fire-proof window and door 
frames, elevator shafts, fire-walls, and other subjects of the kind on 
which the membera of this Society who have to do with building con- 
struction should be informed. 

The writer has trespassed on the time of the Society in bringing to 
its attention some of these matters, which may appear to be but 
distantly related to its functions, because it is his confident belief 
that only through an awakened and enlightened conscience, which can 
be most conservatively stimulated and wisely directed by engineers 
and architects, will the public, the producers, the manufacturers, and 
the governing bodies, embark on the reforms so sorely needed if build- 
ing construction is to be placed on a more enduring and economic basis. 

Mr. Lesley. R. W. Lesley, Assoc. Am. Soc. C. E. — There is one point in the 
fire risk to which the speaker would like to call attention, that is, the 
legislation in reference to fires in the United States and in Europe. 
In France, where he lived for some time, and was familiar with the 



DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 291 

law, the owner is responsible to the adjacent property holder for fire Mr. Lesley, 
extending beyond his own building; in other words, if an owner chooses 
to build an inflammable frame structure, he takes the risk of burning 
down his neighbor's property, and he is liable for the damage, and 
this responsibility probably has a great deal to do with the forms of 
construction adopted. This policy has been adopted in many other 
countries of Europe, the speaker believes, and is a part of the law. 

EiCHARD L. Humphrey, M. Am. Soo. C. E.- — All engineers, and Mr. Hum- 
particularly members of this Society, who have to deal with engineer- ^ ^^^' 
ing construction, should give careful consideration to the facts 
presented in Mr. Wilson's discvission. The summary of these facts 
points out clearly the numerous annual losses in the United States, 
which are largely or entirely preventable, and the dangers of the 
destruction of property by fire which exist throughout the country. 

The conditions which prevail in the United States are not unusual, 
and are the natural result in a country which has grown with great 
rapidity. The congestion of population has created problems which 
necessitate special legislation in order to correct the evils which have 
resulted from laxity and insufiicient municipal supervision in building 
construction. 

It is only after a city has outgrown the flimsy structures incidental 
and necessary to the settlement of new country that the congestion of 
population forces on the inhabitants the necessity for greater safety 
and provision against the dangers from fire. 

The development of the building laws must necessarily be extremely 
slow, particularly because of the great difficulty in securing legislation 
which will correct the evils of the building laws caused primarily by 
the commercial influences which effect such legislation. 

We are prone to look to Europe, to study the conditions there, and, 
in drawing comparisons, to calculate what our losses would be if we 
secured the same types of construction as obtain in European countries. 
This, of course, is very helpful, and is one of the means of drawing the 
attention of cities to proper legislation in a practical way. At the 
same time, it should be borne in mind that European cities are very 
old, and the congestion of population in the principal cities has been 
such as to necessitate, many years ago, such revisions of the building- 
laws as would insure proper safety in those congested centers. In 
other words, European cities have been many years in obtaining per- 
fection in their building laws. The United States is rapidly approach- 
ing, indeed it has approached, in many of its larger cities, such a condi- 
tion in the congestion of its population as to make it absolutely neces- 
sary that steps be taken to prevent types of construction which are so 
dangerous as to form a constant menace of a serious conflagration. 
Many of these structures are fire-traps, constructed so flimsily that it 
is almost criminal to permit their existence. 



292 DISCUSSION ON FIRE-RESISTANT CONSTRUCTION 

Mr. Hum- Anyone who was so fortunate as to investigate such great conflagra- 
^ ^^^' tions as those at Baltimore and San Francisco is overwhelmed at the 
absolute worthlessness of the types of building construction which are 
in vogue in this country to-day. It is true that, after a fire has gutted 
a building, the walls and skeletons are found to be standing, and that 
the building has passed the conflagration satisfactorily. As a matter 
of fact, what remains of the structure forms a very small percentage 
of its value, the greater part of the latter having been lost in the 
destrviction of the trimming and the various appurtenances which go to 
make up the complete building. In New York City buildings which 
are labeled fire-proof, and are apparently of reasonable fire-resistive 
construction, may be found to be surrounded by great areas of struc- 
tures poorly constructed, with unprotected openings of plain glass 
which cracks at the first touch of the flame and enables a fire from 
a neighboring property to spread into the building and destroy its 
contents. Indeed, the intensity of some of the fires which occur in 
such buildings is so great as to destroy adjacent buildings which are 
reasonably fire-proof. Even the heat generated by fire in a building 
which is reasonably fire-proof is sufficient to destroy the integrity of 
the structure, because the building does not possess the requisite fire 
resistance. Structures should be designed so as to offer resistance 
required for the conditions. It is manifest that a warehouse in which 
large quantities of inflammable materials are stored must have greater 
protection against fire than a building which contains only a small 
quantity of such materials. Even where buildings are of the best fire- 
resistive type of construction known to-day, they are often destroyed 
by the conflagration of the fire-traps which surround them. Eor this 
reason, the building laws should be retroactive, in that they should 
compel the removal of all structures which are not of reasonable fire- 
resistive construction, especially in the congested centers of large cities. 
In the speaker's opinion, buildings should not be permitted to exist 
which are known to be fire-traps, and particularly should this be the 
case where the structure is used as a place of assembly for women 
and children. 

In the matter of fixe-resistive construction, it is a regrettable fact 
that most deplorable conditions exist in schools and places of public 
assembly, and, within recent years, there have been such terrible 
catastrophes as to horrify the entire nation, and yet these are only 
sufficient to stir up public sentiment for a brief period of time. As 
soon as the interest begins to wane, the public drops back into the 
same condition of indifference, and can only be aroused by a similar 
holocaust. If the building laws are to be adequately revised within a 
reasonable time, it is absolutely essential that the foremost societies 
of the United States, of which this Society should be a leader, should 
take an active part in the campaign, in order to provide sufficient pro- 



DISCUSSION ON FIRE-EESISTANT CONSTRUCTION 293 

tection against fire and thereby reduce the enormous annual losses. Mr. Hum- 
This Society has taken an active part in the work of conserving the ^ ^'^^' 
nation's resources, and certainly there is not a more fruitful field 
for carrying on this work than in the attainment of structures of such 
a character as will offer adequate resistance to fire, thereby reducing 
to the smallest possible amount the enormous annual destruction of 
building material. 

S. Whinery, M. Am. Soc. C. E.— This subject is one to which the Mr. whinery. 
speaker has not given much attention, but it seems to him that great 
importance may be attached to the particular fact referred to by Mr. 
Lesley, namely, the absence of personal responsibility on the part of 
the owner. 

Doubtless, one reason why people are more careless in the United 
States than in any other country, is that, because of the very complete 
and efficient system of fire insurance, the losses through fire are not 
visited on the persons whose buildings are directly burned, but are 
practically shared by the whole community; and, while fire insurance 
is undoubtedly a very great benefit to the country at large, it may be 
a question whether it is an unalloyed blessing. 

If the owner could be made to feel that if he puts up a cheap 
wooden building — a fire-trap, as it has been called — and that building 
should be burned down, a large part of the loss would fall on him, he 
would undoubtedly take more care to make it reasonably fire-proof; 
but, as long as he knows that he can insure the building for nearly 
its whole value, and thus compel the community at large to share 
the loss by fire, he will naturally be careless, and will favor cheap 
construction. 

From this point of view, it might be wise, and it might also 
result in reducing the aggregate losses by fire, if insurance companies 
were prohibited from insuring any building for more than one-half 
of its value. 



AMEEICAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1127 

THE SEWER SYSTEM OF SAN FRANCISCO, 

AND A SOLUTION OF THE 

STORM- WATER FLOW PROBLEM.* 

By C. E. Grunsky, M. Am. Soc. C. E. 



With Discussion by Messrs. W. C. Hammatt, Egbert G. Dieck^ 

Charles E. Gregory, E. Kuichling, Kenneth Allen, 

Walter N. Frickstad, and C. E. Grunsky. 



This paper ought to have been wi'itten and presented to the 
Engineering Profession long ago. It is now offered in order that 
suitable record may be made of the steps taken by San Francisco to 
better the primitive conditions vphich have been maintained there too 
long, and to call attention to a novel solution of the storm-water flow 
problem which is an outgrowth of the study of San Francisco 
conditions. 

The writer was a member of the 'Board of Engineers' of 1892-93, 
generally referred to as the Sewerage Commission ; he was one of the 
two 'Engineers in Charge to devise and provide a sewer system for 
San Francisco,' of 1893 ; he was in 1899 the 'Engineer in Charge,' with 
two associates, of the design of the sewerage system, which, with minor 
modifications, is now being carried out, and he was for more than 
4 years, 1900-04, City Engineer of San Francisco, in which capacity 
he revised and re-submitted the plans of the sewer system, recast the 

* Presented at the meeting of May 19th, 1909. 



THE SEWER SYSTEM OF SAN FRANCISCO 



295 



F 



jr o 




296 THE SEWER SYSTEM OF SAN FRANCISCO 

cost estimate thereof, and, as funds permitted, directed the construc- 
tion of the first elements of the new system. This statement is offered 
in explanation of the free use which has been made in this paper of 
the reports on the sewerage problem of San Francisco, prepared or 
participated in by the writer from time to time. 

Topographic Features. 

San Francisco covers the entire northerly portion of a peninsula 
in which a spur of the Coast Range, coming up from the south between 
San Francisco Bay and the Pacific, terminates. The city is irregu- 
larly square in outline. On the west is the Pacific Ocean ; on the north, 
the ocean, the Golden Gate, and the bay; on the east, the bay; and on 
the south, Marin County, from which San Francisco is separated by an 
east and west line extending from the bay to the ocean. 

The area of the city is about 30 000 acres, of which about one- 
third is built-up territory. The built-up section covers the north- 
eastern quadrant of the city, and extends, in long narrow strips, west- 
ward, in the northern half of the city, to the ocean, and southerly 
along such principal lines of travel as the San Bruno and Mission 
Roads to the County line. Within recent years the growth of the 
section on the ocean slope, particularly the region just south of 
Golden Gate Park, has been rapid. 

The highest point in the city is practically in the city's geographical 
center. Here are the Twin Peaks which rise to an altitude of a little 
more than 900 ft. A short distance to the northwest of Twin Peaks 
is Blue Mountain, with a summit at practically the same height. 

The ridge which divides the city into two slopes, one easterly toward 
the bay, the other westerly toward the ocean, enters the city from the 
south, a little west of a median line, and holds a fairly direct northerly 
course to a termination at Fort Point. The lowest points on this 
ridge are more than 200 ft. high. Before the recent completion of the 
bay shore cut-off, by the Southern Pacific Company, its main line of 
railroad into the city crossed this ridge from east to west just south 
of San Francisco, and within the city crossed back over it from west 
to east to reach the east central portion of the city. Parts of Golden 
Gate Park and the Richmond District lie upon the flat top of this ridge 
northward from the Twin Peaks group of hills. 



THE SEWER SYSTEM OF SAN FRANCISCO 297 

The principal spurs within the city, from this peninsular backbone, 
which subdivide the eastern or bay slope, are, in their order from 
south to north: Hunter's Point ridge, terminating in a cape which 
extends far out into the waters of the bay in the southeastern portion 
of the city (south of this spur are two small valleys opening to the 
bay, of which the southernmost, Visitacion Valley, lies in part in 
San Mateo County) ; Bernal Heights, between two branches of 
Islais Creek; the Potrero Hills, which are barely connected with the 
main ridge by a slight swell in the ground; and the northern ridge, 
terminating in Kussian and Telegraph Hills on the northeastern bay 
front of the city. 

The western water-shed or ocean slope is subdivided, by the flat- 
topped Ocean House ridge and the Point Lobos spur, into three 
drainage basins, of which the northernmost is drained by Lobos 
Creek; the second embraces a region of drifting sand dunes extending 
well up on the western slope of the central group of hills; and the 
third is the northern part of the water-shed of Merced Lake. 

Some of the hills near Point Lobos are more than 300 ft. high; 
the Fort Point ridge, within the Presidio, has a height of more than 
400 ft.; the Presidio Heights, near where the northern ridge leaves 
the main ridge, have greatest elevations of 300 to 380 ft.; Lafayette 
Square rises to a height of 380 ft.; Clay Street Hill is 360 ft. high; 
Russian Hill, 370 ft., and Telegraph Hill, more than 290 ft. The 
lowest point on the spur which extends northwesterly from Russian 
Hill and terminates at Black Point is about 90 ft. high. On the Potrero 
Hills are points nearly 340 ft. high, and on the Hunter's Point ridge 
several hills rise to more than 250 ft. Railroad Avenue crosses this 
ridge at an elevation of 60 ft., and San Bruno Road at the same eleva- 
tion. The spur which extends eastward from the main ridge south of 
the Hunter's Point ridge near the south line of San Francisco, leaves 
the main ridge with an elevation of 550 ft. This ridge separates Bay 
View Valley from Visitacion Valley. It is crossed by the San Bruno 
Road at an elevation of 140 ft., and terminates with a hill the top 
of which is 400 ft. high. 

Bernal Heights rise to more than 490 ft.; Buena Vista Park is 
nearly 520, and Lone Mountain nearly 480 ft. high. 

Between the several ridges or spurs of hills just mentioned lie 
valleys or broad flat slopes opening out upon the bay or ocean. The 



298 THE SEWER SYSTEM OF SAN FRANCISCO 

ocean slope water-shed south of the Point Lobos ridge has no well- 
defined line of drainage. Irregular sand dumes, formed of sand blown 
eastward from the ocean beach, form a broad broken slope to the 
ocean. Some of the drifting sand has been carried over the main 
peninsula ridge just north of the central group of hills, and, separating 
into two parts, at Lone Mountain, has found lodgment in several 
broad belts extending eastward and southeastward across the peninsula 
to Mission Bay. In smaller quantity, some sand has also been carried 
across the peninsula south of Twin Peaks. 

Sand has also been blown inland from the bay beach lying on the 
northern front of the city to the westward of Black Point. The great 
sand dunes, now only seen in small remnants, that were formerly at 
and near Lobos Park and some distance up on the slopes to the east 
and north, had their origin on this beach. 

Where sand has thus drifted over portions of the site of the city 
it covers clays, shales, and serpentines. To some extent it encroached 
upon the waters of the bay on the eastern waterfront adding body 
to the marginal strips of swamp land and mud. Where the natural 
surface of the city is not covered by sand a clay soil predominates. 
On the hills a shaly or serpentine rock is frequently cut into when 
trenching for sewers, gas, or water pipes. 

The northern and eastern shore of the city in its original condition 
was deeply indented by arms of the bay. These have in part been 
filled in. The most notable of these filled-in areas is the former 
Yerba Buena Bay, which gave the original name to San Francisco. 
This extended as far west as Montgomery Street and from the base of 
Telegraph Hill at the north to the Rincon Hill region on the south. 
Mission Bay, too, has been cut ofF and filled in. The extent of this 
filling will be appreciated when it is stated that here, over an area of 
about 150 acres, all official street grades are at City Base (6.7 ft. above 
ordinary high tide). 

Islais Creek, south of the Potrero Hills, opens to the bay through 
broad marsh lands and mud flats, 500 acres of which are within the 
city limits. It is in this region that, fully two miles inland, official 
street grades as low as 1 ft. below City Base may be noted. 

South of Hunter's Point an area of from 700 to 800 acres within 
the city limits extends out into the bay over submerged mud flats 
which will probably some time be filled in to about City Base. 



THE SEWER SYSTEM OF SAN FRANCISCO 299 

On the northern front of the city the tidal area and bay surface 
that have been or that ultimately will be cut off from the bay by the 
established seawall line aggregate about 270 acres. 

The total area of originally salt-marsh lands and bay surface 
within the boundaries of San Francisco is nearly 2 000 acres. 

The built-up or improved sections of San Francisco fall naturally 
into districts of two characters: one class embracing portions of the 
city on sloping ground, hillsides, and hill-tops, at such elevations that 
they can be readily drained by gravity flow; the other embracing 
the low flat areas for which it is difiicult or impossible to provide satis- 
factory sewage collection without recourse to pumping. The several 
districts of each class, as the foregoing description of the topography 
indicates, do not lie contiguous to each other. The drainage and 
sewerage problem is thus rendered much more complex than if there 
were only one instead of many drainage basins. 

Water-shed Areas. — The extent of the several water-sheds, the run- 
off waters of which are to be cared for was in 1893 estimated as follows : 

Visitacion Valley water-shed 990 acres. 

Bay View water-shed 1 220 

*Islais Creek water-shed 3 900 

Precita Creek water-shed 1 730 

Mission Bay water-shed 5 860 

Yerba Buena Bay water-shed 430 

A North Beach water-shed 320 

Washerwoman's Bay water-shed 1 030 

Presidio water-sheds 930 

Lobos Creek water-shed 2 120 

Ocean Beach water-sheds 5 650 

*Laguna de la Merced water-shed 3 350 

In this list Precita Creek has been given independent rank. It is 
a tributary of Islais Creek, uniting with the latter at the margin of 
the salt marsh. Bernal Heights forms a prominent line of division 
between these two creeks. 

Visitacion and Bay View Valleys. — Both Visitacion and Bay View 
Valleys are still sparsely settled, but the latter at least is a section in 
immediate need of main sewers. 

* Only the area of these water-sheds within San Francisco is here noted. 



300 THE SEWER SYSTEM OF SAN FRANCISCO 

Islais Creek. — The main branch of Islais Creek heads to the south- 
ward of San Francisco. It descends rapidly in its westerly course to 
the low marsh lands which extend far up into the valley of this creek. 
The creek before emerging upon these marsh lands lies for some 
distance in a deep gorge which takes a course through private lands. 
The rectangular system of laying out streets has been adhered to in 
some places along this creek, where, without question, there should 
have been streets following lines of drainage. Where lands are not 
yet subdivided, it is to be hoped that this matter will receive proper 
attention. 

Islais Creek receives the run-off waters from a portion of the 
San Bruno Mountain, to the south of San Francisco, from the north- 
east slope of the Twin Peaks group of hills, and from the entire 
southern slope of Bernal Heights. The region which it drains, the 
south central portion of San Francisco, is a rapidly growing section 
of the city. 

By statutory enactment, Islais Creek is declared to be navigable 
"from Franconia Landing near Bay View turnpike to its outlet into 
the Bay of San Francisco, and thence easterly along the southerly 
line of Tulare Street to the city water front on Massachusetts Street, 
of the width of the channel of said creek." However, Kentucky Street 
has been carried across the creek on a solid fill and there is still some 
doubt as to whether the creek, which across the marsh is but a slight 
depression in the mud's surface, will be opened and kept open for 
navigation purposes. But at the mouth of the creek a large basin 
should be dredged to a depth of about 40 ft. to meet in part the 
needs of San Francisco's growing commerce. It is foreseen that a 
project for such a basin will ultimately be carried out. All the Islais 
Creek marsh has been laid out in streets, and official grades have been 
established. Sometimes this was done in the most absurd fashion. 
Thus, on San Bruno Road and on Fifteenth Avenue, grades are as low 
and even lower than on Kentucky Avenue, li miles nearer the bay. 

Precita Creek. — This creek drains a region which lies between the 
Islais Creek water-shed on the south and the Mission Bay drainage 
on the north. As a drainway, the creek has been replaced by Army 
Street sewer. The discharge of sewage is upon or through the marsh 
into Islais Creek, a short distance east of the eastern spur of Bernal 
Heights. 



THE SEWER SYSTEM OF SAN FRANCISCO 301 

Mission Bay Water-shed. — The drainage basin tributary to Mission 
Bay is of large extent, and is for the most part densely populated. 
It includes the heart of the city, the entire Mission and the Western 
Addition. All the sewage and storm-water run-off of this region, except 
a relatively small portion intercepted by Brannan Street sewer, finds 
its way into the open waterway of Channel Street. The large Channel 
Street sewer, and the sewers' on Seventh, Sixth, and other streets, dis- 
charge into this waterway. The discharge from these sewers represents 
more than half of the sewage and rain-water of the built-up portion of 
the city. During half the year there is no rain to dilute the sewage, 
and it ebbs and flows in Channel Street, creating a filthy nuisance 
beggaring description. There is no chance for sweeping it out into the 
bay because there is no circulation of water. It will continue to be 
offensive until the sewage is carried to some other point of outfall. 
But even under the contemplated arrangement to be hereinafter de- 
scribed the remedy will not be complete. There will still be some, 
though very dilute, sewage occasionally discharged into Channel Street, 
and this open water will remain under a disadvantage common to all 
long narrow inlets from a tidal bay, that of infrequent change of its 
water body. 

The principal sewer of the Mission Bay water-shed, as now in 
service, is the Channel Street sewer, 11 ft. wide and 8J ft. from invert 
to roof. Second in importance, if judged by size, is the Brannan 
Street sewer, 9 ft. wide and 7 ft. high in the clear. It intersects the 
sewers of crossing streets; and, by reason of a connection at its head, 
near Tenth Street, with the Channel Street sewer, is supposed to act 
in a measure as a relief outlet for it. The grade of the Brannan Street 
sewer does not conform closely to the grades of the sewers which it 
intersects. These intersected sewers are generally from 1 to 3 ft. 
higher than the bottom of the Brannan Street sewer, which thus 
should become a recipient of the sewage from the entering sewers. As 
a matter of fact, it is found that the Brannan Street sewer was con- 
structed so carelessly that, where in tunnel, under the south slope of 
Rincon Hill, there is an offset of 10 or 15 ft., and its bottom grade 
for several blocks slopes the wrong way. The accumulation of silt in 
the sewer obliterates the non-conformity of grade with the intersecting 
sewers and there is therefore a partition of flow at practically every 
such sewer crossing. 



302 THE SEWER SYSTEM OP SAN FRANCISCO 

The Brannan Street sewer discharges into the bay under the First 
Street wharf. 

Other sewers, on Mission, Howard, and Folsoni Streets, have inde- 
pendent outfalls, under the wharves or piers of the waterfront. 

The sewage from the northern slope of the Potrero Hills is dis- 
charged by small sewers upon the Mission Bay mud flats or, reaching 
the Potrero Street sewer, it is delivered into the bay in the Potrero 
near the Union Iron Works. 

Yerha Buena Bay Water-shed. — The name, Yerba Buena, as applied 
to any section of San Francisco is now practically lost. Yerba Buena 
Bay, the old harbor of Yerba Buena, the precursor of San Francisco, 
has long ago been filled in and built upon. The extreme westerly end 
of the bay, that is to say, the point of greatest indentation, reached 
Montgomery Street. The name has been used in the various sewer 
reports to designate the low, flat portion of San Francisco which lies 
north of Market Street, near the bay, and the water-shed includes 
the slopes of the hills to the north and the west which send their 
drainage through this flat area. As a sewer district, though of small 
extent, it is of particular interest and importance, first, because about 
120 acres thereof are of the filled-in character, second, because it em- 
braces the most important business section of the city. In this old 
part of San Francisco are to be found the most striking examples of 
sewer construction, without comprehensive plan. Almost every street 
and alley has its 3 by 5-ft. sewer (see Fig. 2), and at every street 
crossing there is the typical sewer intersection. North and south 
streets with level surfaces, that is, without gradient, are no exception 
to the rule. The street in which the standard brick sewer cannot be 
found is a welcome exception. Until about 1890 there was an outlet 
for each east and west sewer at the water front. Then, however, a 
number of these sewers were diverted on East Street from their direct 
courses and a common outfall at Washington Street was provided for 
Market Street sewer and the sewers as far north as Jackson Street. 
This interception of the sewage from the west, which includes the 
intersection by the lower Market Street sewer of all flow in the 
sewers from Geary Street to Sacramento Street, makes the Washington 
Street outlet a point at which a large quantity of sewage is discharged. 
This outlet is at the low-water line, under the wharf. 

The sewers of Vallejo and Pacific Streets have been turned into the 



THE SEWER SYSTEM OF SAN FRANCISCO 



303 



Broadway sewer, which thus becomes their common outlet. It, also, 
discharges under a wharf at the low-water line. It follows, from the 
conditions indicated by this description, that much slush and filth has 
found lodgment in the sewers of this flat area; and it may be added 
that the gradual subsidence of a part of the area, and the construction 
of sewers, a block or two at a time, which did not conform to the 
other sewers in grade, sometimes being several feet too high or too low, 
have greatly augmented the evil. 




PART OF 
SAN FRANCISCO 

SHOWING 
NETWORK OF BRICK SEWERS 



• 3x5 Brick Sewers 

Wooden Sewers 
Pipe Sewers 



Fig. 2. 



The conditions, as they were found when the first examination- of 
this district was made, in 1892, plainly pointed to the use of the exist- 
ing sewers, in part, at least, as storm-water conduits, and the con- 
struction of a new system for sewage only. 

North Beach Water-shed. — The North Beach region has a long 
frontage on the bay. It extends from Telegraph Hill to Black Point. 



304 THE SEWER SYSTEM OF SAN FRANCISCO 

The main difficulty to be overcome in this region results from the 
long bay frontage. Sewers now in use discharge into the bay wherever 
they happen to reach the water line. Sewers of this character are 
those on Larkin, Hyde, Mason, and Powell Streets. 

Washerwoman's Bay Water-shed. — From the Presidio Heights east- 
ward to Russian Hill and Black Point, the northern slope of the 
northern ridge is toward that part of San Francisco Bay known 
originally as Washerwoman's Bay. Here, again, the peculiarity of a 
steep descent from the ridge and the large extent of comparatively 
flat marginal land is to be noted. Some of the land along the bay 
remains to be reclaimed. 

The built-up area is here fast extending downward on the hill- 
sides toward the bay. Sewers for a long series of years terminated 
in land-locked basins or at the inshore limit of the tidal marsh. Some 
of the principal sewers have more recently been extended to the bay 
shore. The present outfall points of this district are at the bay shore 
on Pierce Street, on Steiner Street, on Fillmore Street, under a wharf, 
and on Webster Street. 

Lohos Creek. — Lobos Creek had its head originally in Mountain 
Lake. Though not fed by any well-defined natural lines of surface 
drainage, it includes within its water-shed all the Richmond District. 
The southeastern limit of its water-shed is at Lone Mountain. The 
plateau region northward from the Lone Mountain ridge, westward 
from First Avenue, and southwestward from the Presidio ridge is 
drained by this creek. For convenience of classification, all the 
ocean front drainage, from Fort Point to Point Lobos, may be in- 
cluded in the Lobos Creek water-shed. 

The principal sewer of this district is the Richmond main. Here, 
at variance with the rest of San Francisco, an orderly arrangement 
of the sewers, which are of comparatively recent construction, may be 
noted. The outfall of the Richmond main is just west of Baker's 
Beach, at the foot of Twenty-seventh Avenue. 

The Ocean Beach Water-shed. — This region fronts on the ocean, 
and extends from Point Lobos to Laguna de la Merced. It is without 
natural lines of drainage, except the small gullies on the western 
slope of the Twin Peaks group of hills which are soon lost in the 
deep irregular sand dunes which have drifted in from the ocean. 

Laguna de la Merced Water-shed. — Lake Merced lies in the south- 



THE SEWER SYSTEM OF SAN FRANCISCO 305 

western portion of the city, but a few feet above sea level. The lake 
has a surface area of about 300 acres. Of its total water-shed of 
7 500 acres, somewhat more than one-half lies south of the city in 
San Mateo County. 

The lake is still in use as a source of water for San Francisco, 
being fed by the waters which percolate through the sands covering 
a large portion of its water-shed. Some 2 700 acres, of which about 
2 300 lie within San Francisco, are owned by the Spring Valley Water 
Company, around the lake, for its protection against pollution. 

The surface run-off from the south is intercepted and turned 
through a tunnel into the ocean. Likewise, the storm -waters of the 
Ocean View region are brought within reach of the same tunnel by a 
long flume. The sewage of Ocean View is carried in a line of cast- 
iron pipe to an outfall into Merced Creek below the lake. 

Studies for a Sewer System and its Design. 

The need for a comprehensive system of sewers in San Francisco 
was pointed out, with increasing frequency and emphasis, by various 
city officials and committees, until in 1892 the Board of Supervisors, 
after conference with Frank Soule, M. Am. Soc. C. E., of the University 
of California, passed the following resolution, dated March 7th, 1892 : 

"Resolved, That Professor George Davidson of the United States 
Coast and Geodetic Survey, Col. CJeorge H. Mendell of the United 
States Engineers, and Irving M. Scott, Esq., be and are hereby em- 
powered to select and appoint two engineers, and when the same are 
so appointed that they, with the gentlemen named, shall constitute a 
Board of Engineers on and from September 1st, 1892, to devise and 
provide a system of sewerage for this city and county, as proposed in 
the report of the Committee on Streets, etc., of this Board, made 
January 25th, 1892." 

At a meeting on August 8th, 1892, the first three members of the 
Board of Engineers then created : 

"Resolved, That Marsden Manson* and C. E. Grunsky be appointed 
as the two civil engineers to complete the Board of Engineers, as 
contemplated by Resolution No. 6612, adopted by the Board of Super- 
visors, March 7th, 1892; provided, they agree to devote their entire 
time to this work after being officially notified by Professor Davidson." 

It quickly became apparent to the Board of Engineers thus created, 
that there were no adequate records of the sewers in use. The sewers 
* Member, Am. Soc. C. E., and now City Engineer of San Francisco. 



306 THE SEWER SYSTEM OF SAN FRANCISCO 

had for the most part been constructed piecemeal, a block or a street 
crossing at a time, and diagrams to illustrate the work had frequently 
been considered superfluous. In many cases no record had been made 
of the completed structures. It was necessary, therefore, to make an 
examination of the sewers by entering manholes and by exploring the 
larger sewers. In this way dimensions and conditions as to service- 
ability, if retained in use as a part of a comprehensive system, were 
ascertained. Several exploring parties were organized for this purpose, 
and the results of their work were at once mapped on a suitable scale. 

The preparation of a complete topographic map was also under- 
taken and this was subsequently found to be of great value in 
facilitating the studies. 

Attention was at once given by the Board to the selection of suit- 
able points of outfall for the city's sewage. It was assumed at the 
outset that only one method of disposal — discharge into the bay or 
ocean — would come under serious consideration. Therefore, plans 
were made to observe the direction and velocity of the bay currents. 

The work of the Board of Engineers had not been far advanced 
before its independent course gave offense in certain quarters and led 
to its dismissal (January, 1893). Its books and the data collected 
in the four months of activity were by the Board of Supervisors turned 
over to Mr. Marsden Manson who, however, declined an offer to con- 
tinue the work. A few days later, after repeated conferences with 
members of the Board of Supervisors, Mr. Manson and the writer 
were appointed as engineers in charge to devise and provide the sewer 
system. 

The work outlined by the Board of Engineers was continued, and 
has become the basis for all sewer work in San Francisco since that 
time. On July 1st, 1893, the existing sewers, 200 miles in length, 
had been examined and their general condition noted. All sewers 
through which it had been possible for a man to pass had been examined 
as to general condition. In some instances, where data could not be 
obtained otherwise, the sewers were uncovered. The ascertained facts 
were tabulated for convenient reference, and were platted on sheets 
of convenient size on a scale of 100 ft. to the inch. Sixteen of these 
sheets were at that time complete, ten were in progress, and forty more 
would be required to cover the built-up areas of the city. At this 
time, the general map of the city, on a scale of 600 ft. to the inch, 
had been nearly completed. 



THE SEWER SYSTEM OF SAN FRANCISCO 307 

Float observations at all stages of the tide were made in the bay 
from a number of points, principally Hunter's Point, Potrero Point, 
Center Street Pier, and Powell Street Pier. The purpose of these 
observations was to make sure that the discharge of sewage at the out- 
fall points, as finally selected, would, by reason of ample dilution, 
become inoffensive and harmless before any of it drifted inshore. 

The work of the Engineers in Charge, however, was not carried 
forward to the completion of a design of a system of sewers. They 
shared the same fate as the Board of Engineers— summary dismissal 
in the fall of 1893 — and nothing further was done until 1899 when the 
Board of Supervisors determined to continue the work under the same 
engineers but with the addition of the City and County Surveyor, the 
late Charles S. Tilton ; this addition being made as a matter of policy. 
The resolution authorizing the continuance of the work was passed 
in May, but, owing to some uncertainty regarding its legality, it had 
to be re-enacted in modified form. This took time, and it was not 
until late in September that it seemed certain that the funds set apart 
for the work would become available. In the meanwhile the work was 
under way and was rapidly pushed to completion before the close of 
the year. 

Before submission to the Board of Supervisors, the report of 1899 
on the sewer system was passed upon by Rudolph Hering, M. Am. Soc. 
C. E., and received his endorsement. 

Some Features of the Problem. 

The question of determining whether sewage and rain-water should 
be carried in the same or separate conduits was necessarily one to be 
determined for each of the many districts into which the city is 
divided by its topographic features. The method and place of disposal 
and the sewers already in service were elements requiring to be 
weighed in this determination. It was also necessary to make a careful 
study of rainfall records, in order to ascertain the volumes of storm- 
water flow that would have to be passed through the sewers to the bay 
or ocean. 

Before taking up the discussion of these matters some additional 
notes relating to the old sewers, to show the general lack of system 
in their arrangement, may not be out of place. 

The statement, that the plan for providing sewers seems to have 
been to construct egg-shaped brick sewers, 5 ft. high and 3 ft. wide, in 



308 THE SEWEK SYSTEM OF SAN FRANCISCO 

all streets and alleys, where property was valuable and could afford to 
pay for large sewers, as applying to most of the sewer work that had 
been done in San Francisco, is substantially true. The size of sewer 
required was frequently determined by the Superintendent of Streets, 
who was never a civil engineer, and the prescribed depth of the sewer, 
as defined by ordinance, was 10 ft. below the official grade of the street. 
Therefore, sewers are found, as already explained, where they are 
all but useless, and types of construction have resulted that are cer- 
tainly unique. These facts were set forth at some length in the 
Progress Report of 1S93 of the 'Engineers in Charge.' 

For years the ordinary improvement of street crossings included 
sewers across the streets in both directions and storm-water inlets 
at the curbs on the four corners. The sewers in such intersections 
were built of brick — four Avings from a central manhole. The invert, 
as required by ordinance, was placed 10 ft. below street grade, generally 
level, or, due to the intelligence of most of the sewer contractors, a few 
inches low at the down-hill side of the street intersection. The sewers 
in the intersection might connect with other brick sewers of like size, 
or with larger sewers, or with small pipe sewers, according to what was 
prescribed at some other time for the streets leading from the inter- 
section. After a time, there was a modification of this arrangement 
of sewers at intersections of streets. Where it was known that the 
property owners of outlying districts could not afford the large brick 
sewers, pipe sewers were prescribed for the crossings. The "Bobtail" 
crossing was introduced. This consists of a section of 5 by 3-ft. sewer 
16 ft. long placed along the axis of one street with pipe sewers from 
its ends and with pipe sewers along the axis of the cross street. A 
final modification for crossings of small sewers consisted in omitting 
altogether the brick sewer and putting in two practically horizontal 
pipe sewers along the center lines of both intersecting streets. 

There is no excuse for the many useless storm-water inlets that have 
been constructed on gutter summits. These occur in all parts of the 
city which had been improved to any extent before 1890. In 1893, in 
a central portion of the residence district of the city, a count was made 
of the storm-water inlets located at comers from which there was a 
down gradient in both directions, into which, therefore, no water can 
flow. The district covered by the count extended from Taylor Street 
on the east to Broderick Street on the west, and from Washington 



THE SEWER SYSTEM OF SAN FEANCISCO 309 

Street on the north to Hayes Street on the sovith. In it there are 
1 080 storm-water inlets, of which 149 are absolutely useless, and many 
more receive water from so small an area of street surface that they 
are to be classed as unnecessary. Each of these useless and unneces- 
sary storm-water inlets, or catch-basins, as usually called, cost the 
property owners about $90. 

In San Francisco, as in most of our cities, the improvement of 
streets and the construction of local sewers is paid for by the property 
which fronts on the improvement. The work may be done in districts, 
as was done in the Richmond District, when the main lines of drainage 
were put in, and in the case of the Army Street sewer; but, ordinarily, 
only short stretches of sewer were covered by the official order of 
construction or in advance of such order, under which there would 
be a public letting of the work to the lowest bidder, the property 
owners would be prevailed upon to enter into a private contract to do 
the work. The favored contractor was at times responsible for the 
size of the sewer, suggesting it and securing official approval; at other 
times the Superintendent of Streets and Sewers determined the char- 
acter and size, at any rate, this matter was subject to his approval. 
Only in the minority of cases was it possible to refer to a general 
plan covering some section of the city for sewer sizes. 

The city, preceding 1900, had a City and County Surveyor, whose 
principal duties related to lot surveying. This was an elective office. 
As the law also required that a City Engineer be appointed by the 
Board of Supervisors, it was customary for the Board of Supervisors 
to appoint the incumbent of the elective office to the office of City 
Engineer. For services rendered in the capacity of Surveyor or Engi- 
neer fees were collected, and when special work of any kind was 
required by resolution or order of the Board of Supervisors, the com- 
pensation to be made was usually named therein. 

Supervision of construction was in the hands of the Superintendent 
of Streets. The City Engineer staked out line and grade and issued 
a certificate, showing satisfactory construction or showing by diagram 
departures from prescribed alignment or grade. He was not held 
responsible for the acceptance of work, however great the deviation 
from a required position might be. 

This unsatisfactory method of prescribing and carrying out street 
and sewer work continued until the present city charter went into 



310 THE SEWER SYSTEM OF SAN FRANCISCO 

effect, in January, 1900. Since that time a Board of Public Works has 
had charge of street and sewer construction, and under a well-organized 
Bureau of Engineering, with a salaried City Engineer at its head, all 
matters relating to sewer design and construction receive competent 
and proper attention. 

A few diagrams were prepared in 1893 by the Engineers in Charge, 
to illustrate the lack of system, and peculiarities of the treatment of 
sewerage problems in San Francisco. Of these only one. Fig. 2, show- 
ing the almost universal use of 3 by 5 ft. brick sewers, is exhibited 
herewith. 

Rainfall in San Francisco. 

When the studies relating to the design of a comprehensive sewer 
system for San Francisco were commenced in 1892, an inquiry was 
at once made of the Weather Bureau for data relating to rates of 
precipitation. All that could then be learned was that there were no 
rains of an inch or more per hour, and that therefore the Weather 
Bureau had not undertaken to record the rate of rain. But it was 
agreed that the matter should at once be taken up, as it was quite 
as important for San Francisco to have positive evidence that the 
maximum rain in an hour was about 0.5 in. as for the cities of the 
Atlantic Slope to know that 2 in. or more per hour may be expected 
occasionally. 

Therefore, in 1899, when the matter was taken up again, enough in- 
formation was at command to enable satisfactory conclusions relating 
to maximum rain intensities to be reached. Mr. Thomas Tennent, a 
retired maker of nautical instruments, who had for nearly 50 years 
kept a rainfall record, was of assistance in the rainfall study. His 
records were placed at the writer's disposal* and were freely used. 
From them and from those of the Signal Service and United States 
Weather Bureau (commencing March 9th, 1871), it was found that 
there had been rain in excess of 2 in. per day as shown in Table 1. 

In order that the infrequency of the days on which there is rain 
in excess of 2 in., and their occurrence, unless as a rare exception, only 
in the months November to April, may be understood, some further 
notes on rainfall and climate may not be out of place. Very little rain 
is expected from May 1st to November 1st. There are no violent 

* At the writer's suggestion, these records were turned over by Mr. Tennent to the U. S. 
Weather Bureau. They were lost in the Are of 1906. 



THE SEWER SYSTEM OF SAN FRANCISCO 



311 



thunderstorms, as in the States of the Atlantic Slope, the Middle West 
and the Central Northwest. Even in the winter, rain storms accom- 
panied by thunder and lightning are of rare occurrence. 

TABLE 1. — Rainfalls, at San Francisco, in Excess of 
2 Inches per Day. 





o 

a 


a 
ft 


C 
M 

3.50 
2.46 
2.46 
3.64 
3.09 
3.98 
3.56 
2.17 
3.22 
4.28 
3.62 


1 
ft 


o 
n 


Date. 


to 

V 

a 


Jan. 39, 1850.. 
Dec. 17, 1853.. 
Dec. 25, 1852. . 
Jan. 7,1853.. 
Mar. 27, 1853. . 
Apr. 16, 1853.. 
Mar. 13, 1854.. 
Oct. 31, 1858.. 
Mar. 26, 1861.. 
Dec. 26, 1861.. 
Jan. 5,1862.. 




3.20 

3.00 

2.54 

2.06 

2.85 

3.45 

2.35 

3.06 ' 

2.53 

2.02 

3.67 1 

! 


Jan. 9,1863.... 
Jan. 10, 1863.... 
Jan. 16,1863.... 
Jan. 17, 1862.... 
Feb. 31, 1862.... 
Nov. 26, 1864.... 
Dec. 12, 1864.... 
Jan. 17, 1866. .. . 
Jan. 20, 1866.... 
Dec. 19, 1866.... 
Dec. 20, 1866.... 


Feb. 20, 1867. . . . 
Feb. 21, 1867.... 
Dec. 18, 1871.... 
Dec 19,1871.... 
Dec. 23, 1871 .... 
Jan. 8,1872.... 
Nov. 29, 1872.... 
Nov. 23, 1874.... 
Nov. 17, 1875.... 
Mar. 5, 1879. . . 
Jan. 39, 1881.... 


2.12 
2.22 
2.83 
3.13 
2.48 
2.30 
2.06 
3.98 
2.30 
2.73 
4.67 


Dec. 23, 1884... 
Nov. 23, 1885... 
Dec. 31, 1885... 
Jan 33, 1886... 
Feb. 4, 1887. . . 
Feb. 5. 1887... 
Mar. 13,1889... 
Feb. 15, 1891 . . . 
Dec. 39, 1891 . . . 
Nov. 23, 1896... 
Mar. 23, 1899... 


2.01 

2.58 

2.78* 

2.35 

2.22 

2.93 

2.54 

3.38 

2.31 

2.00 

2.15 



* In 11 hours. 

The cyclonic disturbances that pass over the country from west to 
east — usually from two to four per month — are generally of great 
extent. When they are accompanied by rain, there is ordinarily a 
steady downpour at a moderate rate until the area of low barometer has 
passed on to the eastward, when the change of wind direction from 
southeast to southwest and the fall in temperature may be accom- 
panied by heavy clearing-up showers. These showers, which are not 
frequent, are generally local. They may cover only a small portion 
of the area of San Francisco, and generally have sharply defined . 
limits. Even when the entire city is shower swept, the heavy fall of 
rain is not simultaneous in all parts of the city. Neither are the 
maximum rates of rain long sustained. The fact that the maximum 
rainfall in one hour is well below 1 in. is clearly established by all the 
records of rain that have been kept in San Francisco. On this point, 
under date of June 29th, 1899, Mr. Thomas Tennent says: 

"In reply to your inquiry concerning the rainfall in San Francisco, 
and what has been the greatest quantity which has fallen in a given 
time, say one hour, I will state that I have personally kept a record of 
the rainfall in this city from 1849 to 1896, measuring the quantity in 
each 24 hours at 8 a. m. On looking over my records [preceding the 
establishment of the U. S. Signal Service in March, 1871] I find that 
the heaviest rainfalls we have had in 24 hours occurred, viz.: 1864, 



312 



THE SEWER SYSTEM OF SAN FRANCISCO 



November 26th, 3.98 in., with the remark, 'Heavy rains in the morn- 
ing; rain and gale at midday; steady rain p. M. and evening.' 1866, 
December 19th, 4.28 in., 'heavy rain and gale in the morning; hail 
with thunder at midday, rain during the balance of the day and 
evening.' 1866, December 20th, 3.62 in., 'high wind and cloudy in the 
morning; then a steady rain all afternoon and evening.' I can state 
without hesitation, that we have never had a fall of rain of 1 in. 
in any one hour." 

Based on the records available in 1899, it was found that: 
The average number of rainy days per year is 66. There are to be 
expected no days on which it rains 5 in., and the probabilities are : 

1 day in 50 years with more than 4.60 in. 

" " 4.00 " 

" " 3.50 " 

" " " 3.00 " 

" 2.50 " 

" " " 2.00 " 

" 1.50 " 

" " 1.00 " 

" " 0.75 " 

" " " 0.50 " 

" « " 0.25 " 

less than 0.25 " 

The frequency of rain and the number of days on which rain fell 
in excess of the several amounts above named is shown by the dia- 
gram. Fig. 3, as determined by actual count for a period of 50 years, 
from 1849 to 1899. The close agreement between the fact and the 
probability, indicated by the curve, is noteworthy. 

From this study it was concluded that rain storms with excessive 
fall of rain are rare; that nearly two-thirds of the rainy days are 
days with less than i in. of rain; and that any day with rain in excess 
of 5 in. is highly improbable. 

It should be stated, too, that on the days that show light rain the 
precipitation is iisually well distributed, the rain is light and con- 
tinuous during the greater part of the 24 hours, more frequently than 
in the form of showers of short duration. 

The data secured by the U. S. Weather Bureau subsequent to the 
inquiry of the Board of Engineers in 1893, relating to rain intensity, 



1 " 


" 25 


1 " 


" 10 


1 " 


" 6 


2 days 


" 5 


1 day per year 


2 days 




6 " 




9 " 




16 " 




26 " 




40 " 


(( a 



THE SEWER SYSTEM OF SAN FRANCISCO 



313 



were carefully studied. As already stated, there had been a sufficient 

number of heavy rains to enable satisfactory conclusions to be reached. 

On January 20th, 1894, it rained 0.15 in. in 5 min. 



10 



" October 12th, 1899, " 


0.09 




" November 23d, 1896, " 


" 0.08 




" January 20th, 1894, " 


" 0.17 




" November 23d, 1896, " 


" 0.14 




" November 23d, 1896, " 


" 0.55 




" January 20th, 1894, " 


0.36 





1 hour 



70 Days 




0.50' 1.00' 1.50* 



2.00' 2.50' 3.00' 
Rainfall per Day 

Fig. 3. 



3.50' 4.00" 4.50' 5.00 



Using these and other data from the records, a curve of rainfall 
intensity was prepared, or rather a limiting curve to show the maxi- 
mum amount of rain to be expected in any time unit up to 24 hours. 
The lower portion of this curve was platted on an enlarged scale. 
This curve is shown by Figs. 4 and 5. 

Based on rainfall measurements made subsequent to 1899, a new 
curve has come into use, which shows somewhat greater rain inten- 
sities for short time-periods. The curve now in use is shown in 
dotted lines. 

The most important fact relating to rain intensity disclosed by 



314 



THE SEWER SYSTEM OF SAN FRANCISCO 



the U. S. Weather Bureau records since 1899* is that there was 
one rain storm in the spring of 1904 which exceeded the intensities 
indicated by the curve now in use, by about 10% for all periods of time 
up to one hour. Only two other rain storms were of greater intensity 
than those indicated by the curve, but the excess in one case was 
only for a 5-min. period, in the other for a 10-min. period. It is 
believed that the heavy rain in the spring of 1904 did not cover the 
entire city. 




10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 
Time, in Hours 
Fig. 4. 

The interpretation of the diagram prepared in 1899 established the 
following for San Francisco: 

The maximum amount of rain to be expected: 
In any 5-min. period is 0.15 in. 



10 " 

20 " 

30 " 

40 " 

50 " 

60 " 
2-hour 
3 " 
6 " 

12 " 

24 " 



" 0.22 " 

" 0.32 " 

" 0.40 " 

" 0.47 " 

" 0.54 " 

" 0.60 " 

" 0.90 " 

" 1.15 " 

" 1.83 " 

" 3.00 " 

" 5.00 " 



* From facts furnished by H. De H. Connick, Assoc. M. Am. Soc. C. E., Assistant City 
Engineer. 



THE SEWER SYSTEM OF SAN FRANCISCO 



315 



It was confidently assumed that these quantities of rain would 
rarely be exceeded, and, if exceeded, that the area covered by the 
excessive rain would probably be small; that therefore these quantities 
might safely be accepted in estimating maximum storm-water run-off. 



0.70 



O.GO 



0.50 



,j3 0.40 



3 0.30 



0.20 



O.IO 



















^ 


MAXIMUM RAINFALL 

AT 

SAN FRANCISCO 










^ 












#V^ 


/". 


) Nov.23,'96 












4> 


r 












^') 




^^ 




c 


Jan.20,'91 
' Oct.20,'99 






si 


pf 








( 


, Nov.16,'99 






1 (/Jan.i 

// 

// Oct 1 


) Jan.20,'94 
0,'94 
^Nov 23,'96 

> '00 




Nov.ie,'99 












/ / b Nov 
// '-'Nov. 
ll 


33,'96 
Le,'99 
















1 


a 


3 


4 


5 


G 


T 


8 


90 



Time, in Minutes 
Fig. 5. 



Bearing in mind the fact that a rain expressed in inches per hour 
is practically equivalent to the same number of cubic feet per second 
per acre, it follows from the above that it may rain in San Francisco 
as follows: 



Duration, 
in minutes. 

5 


Rate, in inches 
per liour. 

1.80 


Equivalent, in 

cubic feet per second 

per acre. 

1.80 


10 


1.32 


1.32 


20 


0.96 


0.96 


30 


0.80 


0.80 


40 


0.70 


O.YO 


50 


0.65 


0.65 


60 


0.60 


0.60 



316 



THE SEWER SYSTEM OF SAN FRANCISCO 



This is represented graphically in Fig. 6. The ratios above noted 
were used in estimating the required capacity of storm-water sewers. 
They have been modified somewhat since 1899, as stated, for the shorter 
time periods, so that at present, when a sewer for a small district is 
designed, somewhat larger values for storm-water flow are obtained. 

For comparison, the limiting rain curve now in use is shown, as 
already stated, in Fig. 5, and the corresponding maximum rainfall 
rates in Fig. 6. 



2.00 
1.50 
1.00 
0.50 


\ 

\ 

\ 

\ 

\ 

\ 










RAIN INTENSITY 

AT 
SAN FRANCISCO 




\ 
V)Jan. 

\ \ 
\ \ 


:0,'94 
















0ct.l2.'99 
o 

r 




*s . 














ONov. 

1 


23,'9G 
'Nov.23,'96 


^ 


^«^^^ 




Nov.23,'96 ( 


> 
















C 


) Jan.20,'94 








1 


2 


3 


4 


[) 


& 


C 


7 





8 





90 



Time, in Minutes 
Fig. 6. 

Table 2 is the rainfall rate table now in use. 

TABLE 2. — Rainfall Intensity at San Francisco, 
As in Use in 1908. 



Time, in 


Intensity, in 


Time, in 


Intensity, in 


minutes. 


inches per hour. 


minutes. 


inches per hour. 


5 


2.16 


20 


1.02 


6 


1.97 


30 


0.83 


7 


1.81 


40 


0.72 


8 


1.69 


50 


0.65 


9 


1.59 


60 


0.60 


]0 


1.50 


70 


0.56 


15 


1.20 







In order that the great disparity between the maximum rainfall 
rates to be expected on the Pacific Coast and those of the Atlantic 



THE SEWER SYSTEM OF SAN FRANCISCO 317 

Slope may become apparent, Fig. 4 shows several other limiting rain 
curves, in part taken from the report of the Board of Consulting 
Engineers for the City of Baltimore.* The limiting rain curve adopted 
by this board for Baltimore, and also a limiting curve determined by 
rainfall in Washington, D. C, in Philadelphia, Baltimore, and in the 
State of Maryland, are reproduced in the diagram, and also the limiting 
rain curves suggested by A. N. Talbot, M. Am. Soc. C. E., and by 
Emil Kuichling, M. Am. Soc. C. E. 

The rainfall intensities indicated by the San Francisco curve of 
1899 can be expressed by a fornuila of the type 

h 



1 = 



^* +«"•* (21) 



t + 60 

Here I is the greatest possible mean rate of rainfall (the maxi- 
mum intensity), expressed in inches per hour, when t represents the 
duration of the downpour, in minutes. The numerator, 6, is a constant 
for any locality, but will vary within wide limits for different localities. 

For San Francisco, h = 3.68. 

For points on the Atlantic Slope, h will generally have a value lying 
between 8 and 12. 

In Table 3 rain intensities are noted for several values of h, and, 
for comparison, the rain intensities as deduced from the limiting 
rain curve of 1899 are also noted. 

The Estimate of Storm-Water Flow. 
Method of 1899. 

Having determined the maximum rainfall rates, some method had 
to be adopted of estimating, from these rates, the maximum rate of 
flow to be expected at any point of the storm-water conduits. Such 
method must of course be applicable to areas of all manner of shapes 
and sizes, of varied surface topography, and of varied character so 
far as determined by the improvements, present and prospective. 

It is obvious that the maximum rate of flow at any point of a 
storm-water conduit will be less than the maximum rate of rainfall 
upon the tributary area. Some of the rain will be taken up by porous 
surfaces, some of it will evaporate, and at the time that the flow is 
a maximum much of the rain-water will still be scattered over the 

* Report of Sewerage Commission, City of Baltimore, 1897. 



318 



THE SEWER SYSTEM OF SAN FRANCISCO 



surface or will be in the conduits the filling of which retards the 
passage of and elongates the flood wave. 

TABLE 3. — Maximum Intensity of Rainfall, As Determined in 
1899 FOR San Francisco, and Based on the Formula: 

h 



2 t 



+ «' 





San Francisco. 


Atlantic Slope. 


Duration 
of 


From curve 


By formula: 


By formula: 












minutes. 


of 1899. 


b = 3.68 


& = 4 


6 = 8 
Inches per 


b = 12 
Inches per 




Inches per hour. 


Inches 


Inches 






hour. 


hour. 


hour. 


hour. 


5 


1 80 


1.80 


1.95 


3.90 


5.85 


10 


1.30 


1.32 


1.43 


2.86 


4.28 


15 


1.10 


1.10 


1.19 


2.38 


3.57 


20 


0.95 


0.97 


1.05 


2.10 


3.15 


30 


0.80 


0.81 


0.87 


1.75 


2.62 


40 


0.70 


0.72 


0.77 


1.55 


2.33 


50 


0.65 


0.65 


0.70 


1.40 


2.11 


60 


0.60 


0.60 


0.65 


1.30 


1.95 


90 


0.50 


0.51 


0.55 


1.10 


1.65 


120 


0.45 


0.15 


0.49 


0.98 


1.47 


180 


0.38 


0.387 


0.42 


0.84 


1 26 


240 


0.34 


0.350 


0.38 


0.76 


1.14 


1440 


0.20 


0.182 









Experience in other cities indicates that the combined effect of 
these causes, which contribute to make the rate of flow in the conduit 
less than the rain rate, can with some degree of approximation be 
expressed as a percentage based on the proportional part of the 
drainage area which is impervious. This fact has been pointed out 
by Mr. Emil Kuichling,* and also the further fact that density of 
population may be used as a guide in determining the correction 
factor to be used. 

It was entirely out of the question to make experiments in San 
Francisco to be used in the approximation of reduction factors. The 
assumptions made, therefore, were based on the experience in other 
cities. In San Francisco, as has been explained, the demarkation of the 
area tributary to any storm-water main is not possible without first 
modifying the sewer crossings. But, even if it had been possible to pre- 
pare one or more areas for observation, it would have been folly to do 

*Transactions, Am. Soc. C. E., VoL XX, p. 1. 



THE SEWER SYSTEM OF SAN FRANCISCO 



319 



this in the hope that some storm might occur within the short time 
available for preparing the report, particularly as results were wanted 
before the next wet season. It is to be remembered, too, that such 
results would have been of slight value, because they would have 
represented run-off rates for conditions as they prevailed in 1899 and 
not for those that will prevail in the future. 

The study of population distribution was based largely on the 
records in the Registrar's office. From the size of voting districts and 
the number of votes in each, with due allowance for the class of people 
inhabiting each, an estimate of population density was made. Based 
on this study, it seemed proper to regard all parts of the city having 
a population of 100 or more per acre as impervious. A population 
density was then assumed for each sewer district such that the aggre- 
gate in all districts would be 1 000 000. 

Table 4 shows the reduction factors finally used in estimating the 
required storm-water capacity from the rate of rainfall. 

TABLE 4. 



Population per acre. 


Reduction factor. 


Population per acre. 


Reduction factor. 


20 
30 
40 
50 
60 


0.30 
0.40 
0.50 
0.55 
0.60 


70 
80 
90 
100 


0.65 
0.70 
0.73 
0.75 



The procedure in determining the required storm-water capacity for 
any point of a conduit was substantially as follows: The time was 
approximated which would elapse before the water falling as rain ^in 
the most remote part of the district would reach the point at which 
the capacity was to be determined. Distance and fall between govern- 
ing points along drainage lines enabled a quick and sufficiently close 
approximation of this time to be made. From 3 to 5 min. were then 
added as the time allowance for the rain-water to collect in gutters 
and flow to the storm-water inlets. Although, within this elapsed time, 
there may have been shorter periods with higher rates of rainfall, it 
is to be assumed that these were practically coincident throughout the 
district, and that much of the water falling at the maximum rate upon 
nearby portions of the area had already passed before the water from 
more remote sections arrived. Consequently, the average rate of 



320 THE SEWER SYSTEM OP SAN FRANCISCO 

rainfall during the entire critical period was alone taken into account. 
This was the maximum rate of rain for the estimated time, as shown 
by the curve in Fig. 6. 

The reduction factor, determined by population density, or other- 
wise, was then applied. In this way the maximum run-off rate per acre 
was estimated from point to point for each main sewer. Where two 
main lines of sewers came together, independent estimates were made 
for each, and the results were combined, if the two districts were not too 
dissimilar in shape and extent. Otherwise, the time determination, as 
made for the larger district alone, was used in calculating the rate 
of run-off. This was done because the maximum flow in the drainway 
of the smaller district would occur earlier than in the larger district 
and the allowance for required capacity might be'come unreasonably 
large. 

It is to be understood that this method of calculating storm- 
water flow was applied with care. Cases are readily conceivable where 
the shape of a contributing area is such that it would be improper to 
apply to the entire area a rain rate determined from the time required 
by water to flow from the most remote point thereof. The area under 
consideration may have one or more finger-like projections, long and 
narrow, which, if taken into account, would make the time long and 
the rain rate small. In such cases it becomes necessary to examine 
the lower compact portion of the area by itself. The rain rate is 
determined in the usual way, and if its application to the partial 
area — particularly if, owing to surface character, a larger proportion 
of run-off is to be assumed — gives larger results than the treatment of 
the district as a whole, then this value determines the required conduit 
capacity for storm-waters. Sometimes a succession of trials is requisite 
to establish the maximum possible value. 

This method, in its entirety, however, is lansatisfactory, mainly 
for the reason that it leaves too much latitude to the judgment of 
the engineer. It has not found general acceptance. 

Similar and yet more serious objection may be urged against the 
use of the various formulas for storm-water flow that have from time 
to time been suggested by hydraulicians for application to urban areas, 
None of these, moreover, could have been applied to San Francisco con- 
ditions without special study relating to the determination of constants 
to fit them to local conditions. Even then there would have been 



THE SEWER SYSTEM OF SAN FRANCISCO 321 

grave doubt as to whether formulas predicated on rainfalls exceeding 
those at San Francisco from two to six-fold for time periods of one 
hour and less should be considered applicable. It is only necessary to 
glance at the rain curves in Fig. 4 to appreciate the force of this 
statement. 

In all the formulas that have come to notice the rate of rain, 
with some suitable exponent, generally unity, appears as a factor. 
The rain rate to be used in the formulas must be determined by the 
engineer, whose judgment is guided by the available rainfall data and 
his past experience. Thereupon a correction factor, based upon the 
character of the surface of the district, must be adopted. Even when 
rain data are ample and the constants have been determined with 
care the result may be rendered more or less doubtful if the district 
under consideration has an unusual shape, or has exceptional hypso- 
graphic features. 

These considerations led to the conclusion, in 1899, that it was 
not practicable to adapt any of such well-known formulas as those of 
Hawksley, Adams, McMath, or Biirkli-Ziegier, to the conditions pre- 
vailing in San Francisco. 

The method already described, appearing in 1899 to be the most 
rational then known, was therefore adopted. 

But the need was felt, and has not heretofore been met, of some 
more satisfactory method of translating rain rates into storm-water 
flow. The result of a further study of this subject can now be pre- 
sented with the hope that it may prove of some value to the Engineering 

Profession. 

New Method of Estimating 

Required Capacity of Storm-Water Conduits. 

Surface Run-off. — The rain which falls upon any area during the 
time in which the run-off rate is increasing is in part accounted for 
by infiltration into porous soils and by evaporation; in part it flows 
past the outfall point of this area, and in part it has swelled the water 
contents of the storm-water conduits within the area, or is still in 
transit across the surface to the storm-water inlets. 

It has been stated that the method of estimating sewer capacities, 
as used in the San Francisco problems, takes into account, after a 
fashion, the retarding influence of temporary water storage on the 
ground and in the conduits. The new method makes this storage, 



323 THE SEWER SYSTEM OF SAN FRANCISCO 

which is capable of approximation, an essential element in the calcu- 
lation of storm-water flow. 

This method is applicable wherever it is possible to determine the 
volumetric increase of the water actually in the conduits during the 
progress of a rain of the extreme type. The first step, again, is the 
determination of the maximum rain rates for increasing time inter- 
vals. Thereupon, with these rain rates as a guide, the rates can be 
determined at which the water will run from the surface of the 
ground. The final step is the modification of these surface run-ofi 
rates by the effect of water accumulation or storage on the surface 
and within all conduits which are located above the point at which 
storm-water capacity is to be determined. 

It is to be remembered that the percentage correction, heretofore 
noted as applicable to rain rates when storm-water flow is to be esti- 
mated by the method used in 1899, covers, not alone the reduction of 
flow due to water absorption and evaporation, but also the retarding 
effect due to temporary water storage. These correction factors, therefore, 
cannot be used when it is desired to estimate for a relatively short 
time during the progress of a heavy rainfall how much of the water 
reaches and enters the storm-water inlets. 

The greatest intensity of rain is rarely at the beginning of a 
storm. For the purpose of estimating conduit capacities it must 
be assumed to occur at a time when the ground is already wet and 
no longer capable of absorbing water as rapidly as at the beginning 
of the rain. The rate of water absorption by porous ground plus 
evaporation, though possibly several times greater at the commence- 
ment of a downpour than after the same has continued for some 
time, by no means keeps pace with the rain rate. It follows that it 
would not be correct to apply a reduction factor, ascertained from 
aggregate quantities of rain and of run-off produced by the storm, 
to the maximum rain rate in order to estimate therefrom the max- 
imum surface run-off rate. 

The amount of water which evaporates, during the short time 
covered by the critical periods involved in urban run-off problems, 
will be relatively small and, for all practical purposes, may be dis- 
regarded. 

The run-off, consequently, from an impervious area, the highest 
supposable type of urban area, during any critical period, measured 
on the ground where the rain falls, may be regarded as equal to the 



THE SEWER SYSTEM OF SAN FRANCISCO 323 

maximum rain intensity for that period; and the rate of inflow into 
storm-water inlets will be this run-off rate, modified only by the effect 
of changes in the volume of water temporarily stored on the surface 
of the area. 

It will be shown, hereinafter, that the amount of water actually 
lying upon or flowing over the exposed surfaces of the area under 
consideration cannot be materially different, at the end of a critical 
time period, from the amount thus covering the area at the beginning 
of the period, except when the critical time period is of short duration. 
It may be neglected whenever the critical time period exceeds 15 min. 

The amount of water in temporary storage on the ground's surface, 
that is, the water momentarily in transit, on roofs, on lawns, in gut- 
ters and otherwise, to the storm-water inlets, can be estimated if the 
time is known that it will take water during heavy rains to flow to 
the inlets from the outer portions of the small areas served by the 
individual storm-water inlets. This time is usually from 5 to 10 min. 
Make the unfavorable assumption that it is 5 min. 

The area tributary to each inlet may now be regarded as sub- 
divided into 5 concentric zones, from the innermost of which water 
will reach the inlet in the average time of i min., from the next in 
li min., and so on to the last one in 4^ min. 

At the end of any 5 min., supposing the rain to fall at a uniform 
rate during this time, there will still be on the ground, flowing toward 

4.5 3.5. ^ , 

the inlet, -r- in. of the rain which fell on the outer zone,-;:- m. of the 

rain which fell on the next smaller zone, and so on down to the 
innermost zone, on which there will still be one-half of the rain which 
fell during the last minute. (No correction has been made for the 
fact that the outer half, by time, of each zone is somewhat larger 
than the inner half.) 

If, now, the area which is served by an inlet be considered as 
circular, with the inlet in the center of the area, then the areas of 
the successive zones and the entire area will be to each other as 
4, 12, 20, 28, 36, and 100. 

Calling the total amount of rain which fell upon the area in 5 
min. lOOx, then the amount of water remaining on the ground at 
the end of this period will be 

/ 4 :36 100 , 196 , 324 \ 

iio + 10 + TT + ir + Tr/^ = 6^^- 



324 



THE SEWEE SYSTEM OF SAN FRANCISCO 



The water on the ground at the end of the 5-inin. period, there- 



fore, is 



6G X 



' or two-thirds of the total amount of rain which fell 



100 a; 
in that period. 

Generalizing, it may be said that what is true of an area with 
circular outline is true of other compact shapes, and that, therefore, 
two-thirds of the water which falls during any time period equal to 
the time required by water to reach the inlets from the outer portions 
of areas drained by each, will, at the end of that period, be still on 
the surface in transit to the inlet. 

For any limiting rain curve, therefore, the amount of water on 
the ground in a district with impervious surface can be estimated 
for any instant of time, if the time is known that will be consumed 
by the water in flowing from the limits of an inlet area to the inlet. 

For the inlet or entrance time equal to 5 min., and the rain rates 

indicated by the limiting rain curve of 1899 for San Francisco, the 

amount of water on the ground at the end of successive intervals of 

5 min. is shown in Column 5 of Table 5. 

TABLE 5. — Water on the Ground during Kainfalls of Ex- 
treme Intensity at San Francisco. 

Based on Rain Intensities as determined in 1899, using the formula: 

3.68 



I=. 



2 t 



t + 60 



+ « 











Water on 






Successive 


Greatest pos- 
sible total 


Rain during 


Intensity of 
rainfall, in 
inches per 
hour: or, ap- 
proximately, 
second-feet 
per acre. 


the 
ground 
at the 


Greatest 

possible 

total 


G J,— G7._^5 

cubic feet 
per acre. 


periods: 

Duration, in 

[minutes. 


rain to the 

end of the 

period, in 

inches. 


each succes- 
sive period, 
in inches. 


each 

period, in 

cubic 

feel per 

acre. 


rain to the 
end of eacb 
period, in 
cubic feet 
per acre. 


(1) 


(2) 


(3) 


(4) 


(5) 


(6) 


(7) 


0to5 


0.150 


0.150 


1.795 


360 


540 




5 to 10 


0.219 


0.069 


0.828 


166 


790 


195 


10 to 1.5 


0.275 


0.056 


0.672 


134 


1 000 


32 


15 to 20 


0.321 


0.046 


0.552 


112 


1 170 


22 


20 to 25 


0.364 


0.043 


0.516 


103 


1 320 


9 


25 to 30 


0.404 


0.040 


0.480 


95 


1 470 


8 


3C to 35 


0.441 


0.037 


0.444 


89 


1610 


6 


35 to 40 


0.476 


0.035 


0.420 


84 


1 730 


5 


40 to 45 


0.509 


0.033 


0.396 


79 


1830 


5 


45 to 50 


0.541 


0.032 


0.374 


75 


1 970 


4 


50 to 55 


0.571 


0.030 


0.360 


72 


2 080 


3 


55 to 60 


0.600 


0.029 


0.348 


70 


2 180 


2 


60 to 120 


0.906 


0.306 


O.:306 


61 


3 290 




120 to 180 


1.158 


0.252 


0.252 


50 


4 200 




18C to 240 


1.396 


0.238 


0.238 


48 


5 090 





THE SEWER SYSTEM OF SAN FRANCISCO 



325 



|g§gggg5;S„ 


Duration of the critical 
period, in minutes. 


1.00 
1.00 
1.00 
1.00 
1.00 
1.00 
1.00 
1.00 
1.00 
1.00 


All 
impervious. 


c 


oooooooooo 


75%- 
impervious. 


II 

coW 

• as 


poop p p p p p p 

'o" 'tn 'o> ca "aj '3: 'os ^ <I '00 


box 

impervious. 


II >^ 

P 3 
a 

118 



'en 


oopooopopp 
'to 'os ji. '>«^ 'ife. '*. 'c in 'o5 '-."! 

<!0005>ft.O5'CC0<l>t^ 


impervious. 


p p p p p p p p p p 
I-t ^ to Ml 'lo 'co CO '>*ik 'ui 'a 


None 
impervious. 




•g-g-o-g'g'g'ogb'S 


All 
impervious. 




p p p p p 

'oD OD 'oo bo cr '00 00 'qo bo 'cD 


75% 
impervious. 


m t^o- 


p p p p p p p p p 

OlO5O5OiO5O5*<J<}-^i0D 


50% 
impervious. 


1 

00^ 


p p p p p p p p p 

its. j^ 'rfi. '►^ Ul C« 'cjt 'o5 'oi -^l 
tilfc^50l-'05350 0'. CO 


25% 
impervious. 


*- 


oppppppppp 
'to io '10 CO 'co 'm *>. i(^ 'en 05 

COC;iCD'-^**-<lCO**4^rf^ 


None 
impervious. 




S'gbo'g'og'g'o'o 


All 
impervious. 




oooooooooo 


75% 
impervious. 




oooooooooo 


50% 
impervious. 


II II II 

Oi-'oo 


p p p p p p p p p p 

M-t005*.CnO55Ol-'CO-J 


35% 
impervious. 


to 


pppopppppp 
b5bscsoib5'3;--><j-QbD 
K)CO*.cn-i;DiOil^-Jto 


None 
impervious. 




ggo'g'o'go'ggo 


All 
impervious. 




p p p O' p p p p p p 

!P cp b CD "to *CD 'XJ '0 ':D ":D 


75% 
impervious. 


II II II 


OOOOOOOOOO 

258g8?8gS82^S'g 


50% 
impervious. 


OOOOOOOOOO 


25% 
impervious. 




OOOOOOOOO p 
lliOICO-iOOO'-'COCHOD 


None 
impervious. 





< s^ 



+ 



+ 



o 



<.| 



H 

Si 



O 
o 

M 

o 

o 
o 



CO ^ 






W 
O 

M 

M 



326 THE SEWER SYSTEM OP SAN FRANCISCO 

The absorption of water by the soil will vary directly with the 
proportional part of the surface that is pervious, and will also vary 
with the character of the soil and subsoil. It will be largest in 
districts having deep soils of the sand or gravel type that take an 
indefinite quantity of water freely, and will be least in districts with 
heavy clay soil or with a thin soil cover on impervious rock forma- 
tions. 

The rate at which the water thus deflected from a course across 
the surface of the ground will sink into the soil should be assumed 
to be a gradually decreasing rate. Soil will take water somewhat 
more readily at the beginning of a downpour than after a heavy 
rain has continued for some time. 

This rate is determinable, and upon it will depend the relative 
amount of the rain which will from time to time, during the progress 
of a heavy rain, reach the storm-water inlets. 

Data have not been accessible for an entirely satisfactory deter- 
mination of reduction factors to be applied to rain rates to reduce the 
rain rates to surface run-off rates. The factors noted in Table 6, 
therefore, are to be considered as tentatively suggested. They are 
to be used until checked and corrected by further observations. They 
apply to soil conditions such that the rate of water absorption by 
the soil at the beginning of a critical period will be about | in. per 
hour, decreasing gradually to about i in. per hour at the end of 1 
hour. 

Notation. — The following notation is used in this paper: 

A = Area, in acres, of the district above the point at which 

storm-water discharge is to be estimated ; 
T = End of a j^eriod of heavy rainfall ; 
T^ = Time at which the storm-water discharge is a maximum ; 
B = Beginning of a period of heavy rainfall ; 
B' = Time, t minutes preceding jB ; 

t = Elapsed time, in minutes, from B to T; 
t^^ = Elapsed time, in minutes, from B to T^\ 
R = Maximum rainfall in one hour; expressed in inches; 
I = Maximum rain rate, in inches per hour (or, which is the 
same, second-feet per acre), for a period of time, t 
minutes in duration; 
Tb = Maximum rain rate, in inches per hour, during t min- 
utes preceding the time, B, during time, B' to B; 



THE SEWER SYSTEM OF SAN FRANCISCO 337 

Ut = The rate of inflow into storm-water inlets during the 
time, B to T, expressed in cubic feet per second per 
acre ; 
ttB = The rate of inflow into storm-water inlets during the 
time, B' to B ; 
C = Volumetric capacity for storm-water (storage capacity) 
of all conduits, above the point at which discharge is 
to be estimated, expressed in cubic feet per acre; 
Cs = Storm-water contents of the conduits at the time, B, ex- 
pressed in cubic feet per acre; 
Cj. = Storm-water contents of the conduits at the time, T, ex- 
pressed in cubic feet per acre ; 
Cj.— Cs = Therefore the augmentation of the water temporarily 
stored in the conduits during the time, B to T ; 
Gb = The water on the ground, on roofs, etc., and in the 
gutters, flowing toward the storm- water inlets, at the 
time, B, expressed in cubic feet per acre; 
Gt = The water on the ground, etc., flowing toward the storm- 
water inlets at the time, T; 
{ = The time, in minutes, required by water to flow from the 
outermost portions of the area served by a storm-water 
inlet to the inlet; 
Gj, + i=^ The water on the ground, on roofs, etc., flowing toward 
the inlets, i minutes after the close of a critical period — 
a continuation of maximum rainfall conditions being 
assumed ; 
r = The surface run-ofi: rate during the time, B to T, 
expressed in depth, in inches per hour, or in cubic 
feet per second per acre; 
r' = The surface run-off rate during the time, B' to -B, 
expressed in depth, in inches per hour, or in cubic 
feet per second per acre; 
a = The constant in the equation r = a I (and approximately 
r' = a Is). This constant will have different values 
for different areas. For areas with impervious surface 
throughout, a is unity ; 
& = A constant appearing in the formula for rain intensity; 
d = The mean discharge of storm-water, in cubic feet per 
second per acre, at the point for which storm-water 
discharge is to be estimated, during the time, B to T; 



328 THE SEWER SYSTEM OF SAN FRANCISCO 

dy= The storm-water discharge at the tmie, T, expressed in 
cubic feet per second per acre; 

d^ = The maximum storm-water discharge at the time, T^, 
exjjressed in cubic feet per second per acre; 

els ^= The storm-water flow at the time, _B, expressed in cubic 
feet per second per acre; 

tj) =^ The time, in minutes, required for water to flow from the 
most remote portions of a district to the point at which 
discharge is to be estimated (including a time allow- 
ance, i, for flow to inlets); 
S = The increase of the water quantity in temporary storage 
on the ground and in the conduits during the time, B 
to T; 

(S,^ = The increase of the water quantity in temporary storage 
on the ground and in the conduits during the time, B 

to 3;,; 
q = The total depth of rain, in inches, during the time, B to 

T; 
g' = The surface run-off, in inches, during the time, B to T\ 
e = A. constant, appearing in the special case when the 

limiting rain curve is a parabola. 

The Storm-Water Flow. — The maximum quantity of rain, ex- 
pressed in cubic feet, that may fall in the time, B to T, on one acre, 
will be 60 / t. The portion of this water which will reach the storm- 
water inlets is 60 r t. This is the surface run-off per unit area, and 
must be equal to the total flow during the time, B to T, at the point for 
which discharge is to be estimated, plus the augmentation of the water 
storage on the ground and in the conduits during the same time. 

60 r i = malt (1) 

mrt= m d t -\- {Ct — Cb) + {Gt — Gb) (2) 

^ _ .. (^r — C'jg) -f {Gt — Gb) fos 

"^ - ' 601 ^^ 

or d =al (^t-Cb) -h (Gr-GB) ^^^ 

bO t 
But if d is known, and if d^can be determined, it will be possible to 
find the value oi d^ on the assumption that 

2 1, 
d = ^dr + 3 «B (5) 

1 
From which fZ^ = 2 (3 (Z — cZ^) (6) 



THE SEWER SYSTEM OF SAN FRANCISCO 



339 



The value assumed for d in Equation 5 results from the hypothesis 
that the discharge-curve for the time, B to T, approximates in shape a 
parabola, tangent to a horizontal line at T and passing through the 
point B, as shovpn in Fig. 7. 

















T 












^ 


^ 






i 


\ 








B 


1 








h 




t 


^ 




1 


f. 








B' ^ 




-^ 




+ 










. 



DISCHARGE-CURVE 
Fig. 7. 

But the value of fZ^ can be approximated on the further assumption 

that only the rainfall in the t minutes preceding the time, 5, need be 

taken into account. For the greatest possible rain intensity preceding 5, 

the value of ds-, without too great error, may be called equal to the mean 

run-off rate during the t minutes preceding B. That is to say: 

Because, approximately, r' = a /g (7) 

Therefore, near enough, d^ =i r' =: a Is (8) 

1 
And from Equation 6, (Zy =: ^^ (3 cZ — (xIb) (9) 

It appears from the above that the value of d increases with 
increasing values of Cb- It increases in value, also, for decreasing 
values of Cf. (See Equation 4.) Therefore the rain which will give d 
maximum value must be so distributed to the time preceding and 
immediately following the time, B, that the conduits will be as full 
of water as possible at the beginning, and at as low a stage as possible 
at the end, of the critical period. With this fact the type of rain 
storm becomes known which will make the mean discharge through- 
out a critical period a maximum. 

The rain must evidently fall preceding the time, B, with constantly 
increasing intensity, and during the time, B to T, with constantly 
decreasing intensity. 

It follows that the greatest possible rain intensity, at a time instant 
immediately preceding B, will be the same as the greatest possible rain 
intensity at the time, T. Consequently, for the most unfavorable rain 
conditions, the value of Gr^ will be equal to Gj, _^_ ^ (the inflow time 
being i minutes). 



330 THE SEWER SYSTEM OF SAN FRANCISCO 

Equation 3 may now be written 

but S^Ct — Cb -\- Gt — Gt+ ,. (11) 

therefore d =^ r -— (12) 

The values oi G^ and Gj,+i can be calculated for any rain curve, 
as has already been shown. The value of d in Equations 3 and 10 and of dj, 
in Equation 6 are dependent, not alone upon the intensity of the rain, 
but also upon the duration of the extreme rain conditions, and upon 
the volumetric increase of the water in the conduits and on the ground 
during the critical period. 

It may be stated in passing that, for the inlet ends of the con- 
duits, Ct — Cb = 0. Therefore, from Equation 10, for the special 
case in which d = xi and c7y = Uj, 



callinof i = 5 



'' = — ""7,^ <»*' 



and from Equation 6 

1 
'fr =2 (3 w — r') (15) 

It may be inferred from Equation 9, 

1 
dr^ 2(3cZ — a/^), 

that the value of dr will follow closely the value of d; that, therefore, 
the same type of rain storm which will give the mean discharge during 
a critical period maximum value will also make the discharge at the 
time, T, maximum. It is reasonably certain that this is the case, as 
will be shown later in the presentation of a graphical solution of the 
problem. This fact is not demonstrated by Equation 9, because C^ 
is at maximum value when I^ is at maximum value, while d^ in- 
creases with the increasing value of C^ and with the decreasing value 
of Ib- The graphical illustration will show that in all ordinary cases 
the effect of changing values of Cb upon the value oi dr outweighs 
the effect of the corresponding changes in the rain rate, I^. But the 
introduction of the variable, Ib, into the calculation of dj, from d 
makes it apparent that the time at which d is maximum is not 



THE SEWER SYSTEM OF SAN FRANCISCO 331 

necessarily the time at which dx is maximum. These times, in fact, 
are not coincident. 

The least satisfactory part of the foregoing demonstration lies in 
the approximation of a value for d^ and in the assumption that the 
maximum discharge can be computed by Equation 9 from the mean 
discharge during the critical period and the discharge at the beginning 
of the period. 

Nor is it, as a matter of course, strictly correct, except in the case 
of impervious areas, to assume that the relation of / to i^ is the same 
as the relation of r to /. This assumption is probably near enough to 
the truth for all practical purposes. 

The true value of dxi therefore, may depart to some extent from 
the value determined, as explained, from d and dB- The room for 
such departure will decrease as the values of d]i and d approach each 
other; it will be greatest when (ijs ^ 0; that is, the opportunity for a 
departure from the correct value is greatest in the case of no rain 
preceding the critical period. It is quite possible that for this 
special case some other expression than Equation 9 can be found that 
will better express the relation between d, dn^ and c/y/ but the further 
investigation of this special case would only be of interest if it were 
desired to deduce formulas for application to types of rain storms 
other than those which produce the maximum discharge. 

To estimate the values of Ct and C^ let it be assumed that a first 
approximation of conduit sizes by any method has established a first 
value of C. The value of C when the conduits are to carry water 
other than that due to rain will not, as a matter of course, be the 
entire volumetric contents of the conduits, but that portion thereof 
which is available for the temporary storage of rain water. Call 
t^ equal to to- This fixes tentatively the position of T in the time 
scale. The limiting rain curve furnishes the value of / and of 
-Z/j, and, by application of the proper reduction factor, the values of 
r and of / are found from r ^ a I and r' = a 1^. 

It is now possible to estimate the maximum value of Ob- This 
will result, as already explained, if the rain storm is of such a charac- 
ter that, preceding the time, B, there is a rainfall of constantly in- 
creasing intensity. See Fig. 8. The flow at the time, B, can now 
be calculated; but the ratio of this flow to the flow at the time, 
T, may be, preliminarily, considered the same as the ratio of 1^ to 7. 



333 THE SEWER SYSTEM OF SAN ERANCISCO 

With the aid of this ratio, on the assumption that the space occu- 
pied by water in the conduits is proportional to the flow (the velocities 
being practically the same for half full and for full or nearly full 
stages) a first approximation of C ^ is possible. The first approximate 
value of Ct is C. That is, all conduits are assumed to be full at the 
time, T. 

The value of d resulting from this assumption will be too small, 
because, even under the most unfavorable conditions, Gr will always 
be somewhat less than C. 

The value of dj. follows by the use of Equation 9. 

The conduit sizes originally entering into the calculation can 
now be revised, and the values of Ob ^^^ ^V ^^^ again determined. 
It is now desirable to make the estimate of Ct with greater pre- 
cision than for first approximation purposes. This can readily be 
done, because, at the time, T, the conduits are full at the point 
for which storm-water flow is to be estimated and the rate of flow 
at all upper ends of conduits can be estimated from the rain rate 
which prevailed in the last i minutes preceding the time, T. 

It will ordinarily be found convenient to call 

Ct ^ KG (IG) 

where K \q o. coefficient to be ascertained for each area under con- 
sideration. The value of K will approach unity for small districts. 

From Equations 9 and 10, 

g^--y__ KJ^^-Qb Gt — Gj-_^^ 

GOt 60 t 

and d,^l\?Ar-^^-^^ - ^r-GT+5\_ 1 
^ 2 L V (50 i 60 « / J" 

For all cases in which f exceeds 15 min., Gj, — (r^-i-g will have so 
small a value that the term containing this expression may be 
neglected, therefore, for t > 15, 

^'^Wi«=HK'-^^^)-'-] (-) 

Numerical Illustrations. — Let it be required to determine the 
capacity of the outfall sewer for a business district in San Francisco, 
with impervious surface throughout, the area of the district being 
1 000 acres. Suppose that the district has such shape and such sur- 
face slope that it will take water 50 min., including a 5-min. entrance 



THE SEWER SYSTEM OF SAN FRANCISCO 333 

allowance, to flow from the most remote portion thereof to the point 
at which the conduit capacity is to be determined. Temporary storage 
of water on the ground's surface, because t > 15, may be neglected 
{Gt — Gj, f 5 = 0). 

The maximum rain rate in San Francisco for 50 min., when J = 

o no 

__ — '— (see Table 2), is 0.65 in. per hour, or 0.05 sec-ft. per acre 

For an impervious surface, a = 1.00; therefore r = J, and r' = 7^. 

A first approximation of the storm-water discharge can be made 
by the method heretofore described as in use in San Francisco. 

By this method, 

c\^ = 0.75 X 0. 05 = 0.488 sec-ft. per acre. 

Let it now be assumed that calculations for a number of points in 
the district have established conduit sizes having a volumetric capac- 
ity for storm-water of 1 500 cu. ft. per acre. 

Let it further be supposed that calculations already made have 
established K = 0.8. 

An examination of the limiting rain curve of the type recom- 
mended for use in San Francisco will show that during a period of 
time, t minutes in duration, immediately preceding or following a 
critical period of the same duration, the greatest possible rainfall will 
be, in round numbers, one-half of the rain during the critical period. 
For approximation purposes, therefore, the quantity of water in the 
conduits at the beginning of a critical period may be taken at one-half 
of the amount in the conduits at the close of the period. 

Approximately Cb = -^^ 

but Ct ^ K G = 0.8 G, 

therefore, Gt ~ G^ = 0.4 G = 600. 

3 6S 
The limiting rain curve for San Francisco, hased on J = 



2t 



t + 60 

Fig. 6, indicates that in the 50 min. preceding the time, T, tliere may 
be a rainfall aggregating 0.539 in.; that is, I = 0.047 in. per hour. The 
rainfall in 50 min. preceding the critical period cannot, as shown by the 
rain curve, exceed 0.272 in., because the total rain in 100 min. cannot 
exceed 0.811 in. 

The rainfall, 0.272 in. in 50 min., represents a rain intensity of 
0.326 sec-ft. per acre. 



334 THE SEWER SYSTEM OF SAN FRANCISCO 

Conse(iuently, r =^ al — 0.647 (a beiug unity) 
r' = als = 0.326 

From Equatiou 19, fZy = ^ [ 3 ("0.647 — ^^^^ — 0..326 "I 

= 0..508 sec-ft. per acre. 
A dr = 508 sec-ft. 

As this value is somewhat larger than the first approximation, it 
may now be necessary to correct the value of G. For the purpose of 
this illustration, it will be supposed that no correction is necessary. 

Other values of t are now to be tried, in order to find that value 
of i which will give dj^ maximum value. 

For t = 40 min. 
r = 0.711 
r' = O..3o3 

= 0.515 
A clj, = 515 sec-ft. 

For t = 30 min. 
r = 0.806 
r' = 0.392 

.,= ^[3 (0.800- J;°«,)- 0.392] 

= 0.513 
A dj, = 513 sec-ft. 

The critical period is evidently about 40 min., and the maximum 
storm-water discharge, or required conduit capacity, will be about 
515 sec-ft. 

As a second illustration, let it be required to ascertain the storm- 
water discharge from a district, 1 000 acres in area, in which only 
250 acres of the surface are impervious. Here, again, the time, tp, 
is to be approximated from probable conduit gradients. Suppose 
fj) = 50 min. 

The value of a, to be used in the equations r = a I and / = a I ^^ 
will be found in Table 6, that is a = 0.44. 

Suppose that, by preliminary estimate, = 800, and that K = 0.8. 

(J 
Again accepting the approximate value, (7^ = -5-, makes Cg = 320. 



THE SEWER SYSTEM OF SAN FRANCISCO 335 

For t ^= tj) = 50 mill. 

1=0.647 
and Is = 0.326 

r = 0.44 I = 0.285 

r' = 0.44 Ib = 0.143 

(U =-[s ("0.285 — ^?^^ — 0.143 1 
«r — 2 L V 8 000/ J 

= 0.196 

A dr =196 sec-ft. 

For t = 40 mill. 
1= 0.711 
Is = 0.353 
r = 0.44 I = 0.313 
r' = 0.44 Is = 0.155 

,,,- 3 [3 (0.313 -^•)- 0.15.5] 
«r — 2 L \ 2 400/ J 

= 0.192 

A (It =- 192 sec-ft. 

For t = 60 min. 
I = 0.599 
Zb = 0.306 
r = 0.44 1 = 0.264 
r' = 0.44 Is = 0.135 

,Z,-1 [3 (0.264--^) -0.1351 
"^ — 2 L V 3 600/ J 

= 0.195 

A (It = 195 sec-ft. 

The critical time is evidently about 50 min., and the required 
capacity about 196 .sec-ft. 

It must be remembered that the values of the reduction factor, a, 
noted in Table 6, are based on a specific rate of water absorption by 
the soil. They are to be suitably modified whenever, for any district, 
other rates of water absorption by the soil shall have been determined. 

The increase of water contents of the conduits, when conduits 
are already constructed, should be calculated from the estimated water 
stages whenever this is practicable, instead of using the above indi- 
cated approximation method. 

How the value of t^ may be found, when the storage increase 
during a critical period is known, will be explained later. 



336 THE SEWER SYSTEM OF SAN FRANCISCO 

Graphical Solution. — The method of estimating the reqixired capac- 
ity of storm-water conduits just illustrated becomes of importance, 
first, because it is based on sound principles and can be made appli- 
cable to districts of any shape, character, and size; and second, 
because it enables the problems to be solved by simple graphical 
methods, as will now be explained. 

With the mass curve of rainfall for extreme conditions of precipi- 
tation as a guide, there is to be constructed for any area under con- 
sideration a second integral or mass curve representing the maximum 
amount of surface run-off during the progress of a downpour of the 
extreme type. The new curve will be the rain curve modified by 
the appropriate reduction factor, a, to be found in Table 6. The new 
curve, for convenience, may be called the surface run-off curve. 

The surface run-off curve for areas with impervious surface 
throughout will coincide with the limiting rain curve, because, for 
the impervious surface, the factor, a, is unity and r = I. The San 
Francisco surface run-off curve for impervious surface, based on rain 
intensities expressed by the formula, 

3.68 



2 t 



i + 60 



(which will closely approximate the rain curve of 1899), is shown in 
Fig. 9. 

The time scale in this diagram indicates minutes subsequent to 
the time, B. The ordinates represent, in cubic feet per acre, the 
water which will ultimately reach the storm-water inlets. 

The total run-off during any time period is the difference between 
the ordinates at the beginning and at the end of the period. Some 
of this water, however, is still flowing across the surface toward the 
inlets at the end of the period. 

The surface run-off curve for an area having a surface which is not 
impervious throughout can be obtained from the curve for impervious 
areas by multiplying the ordinates by the proper reduction factors. 

Suppose now that the limiting surface run-off curve be platted in 
a system of co-ordinates with the point, B, at the beginning of the 
co-ordinates (Fig. 8). 

Let T be located on this run-off mass curve so that 5 to T on the 



THE SEWER SYSTEM OF SAN FRANCISCO 



337 



time scale is the actual critical period for the point of the conduit 
under consideration. 

Make T D' = S 

the value of S being S = Gj, — Cj^+ Gj, — Gj, ^ j (11) 



then 



d = 



D' F 



60 t 



The storm-water discharge at the time, B, as already explained, 
which will make d a maximum, will be the discharge due to a maxi- 
mum possible fall of rain in the time, t minutes, preceding B. The 
rainfall in this time must be one of increasing intensity. The run-off 
curve, therefore, will take the position as shown, from B" to B. 



GRAPHICAL SOLUTION 

OF THE 

STORM-WATER FLOW PROBLEM 

VALUES HERE SHOWN ARE 

BASED ON 

THE 

LIMITING RAIN CURVE 

AND THE 

SURFACE RUN-OFF CURVE FOR 

IMPERVIOUS AREAS 

AT 

SAN FRANCISCO 




t+(iO 
t =time, in minutes, 
Jj=; maximum intensity of rain 

during t minutes, in inches per hour; 
S = increase of water quantity on the ground 

and in the conduits during the time, B to 



Fig. 8. 

The discharge of storm-water at the time, B, may be taken, with- 
out material error, as equal to the mean surface run-off rate during 
the time, B" to B, provided, of course, that i be not at material varia- 
tion from tj) Therefore, 

The mass curve of storm-water discharge after the time, B, will 
therefore be represented by a curve to which the line, B" B, is tangent 
at B, and which will gradually rise to the point, D\ The inclination 



338 



THE SEWER SYSTEM OF SAN FRANCISCO 



of the tangent to this curve at the point, D', will represent the maxi- 
mum flow at the time, T. 

The shape of this curve will probably be such that its change of 
direction is somewhat more rapid at the time, B, at which the down- 
pour is most intense, than at the time, T, at which the flat top of 
the flood wave in the conduit has reached the point for which dis- 
charge is being estimated. It will be reasonable to assume, if 5" B be 



1.00 
0.80 



9 0.40 




f+60 ' 
f = time, in miDutes; 
/= maximum intensity of rain 

during t miuutes,in inches per hour; 
S^ increase o£ water quantity on the ground 

and in the conduits during the time, B to T 



50 60 70 

Time, in Minutes 
Fig. 9. 



120 



prolonged to the point, E', on the time ordinate, one-third of the way 
from B to T, that the line, £" D\ will be a close approximation to the 
inclination of the curve at the point, D\ This is in conformity with 
the assumption heretofore made that 



" - 3 



+ % 



(5) 



By assuming a new position for the point, T, with corresponding 
modification of the curve, B" to B, a new position of the point, E', 
will be found. It thus becomes possible, by determining successive 
locations of the points, E' and D', to find the particular value of i 



THE SEWER SYSTEM OF SAN FRANCISCO 339 

which will give the line, E' D\ its greatest inclination; which will, 
in other words, for the known value of S, give dx maximum value. 

The same result may be obtained in a much simpler way. Sup- 
pose the value of S to be scaled ofi upward from B in the diagram, 
Fig. 9, establishing the point, D. Let B' be the point on the run-off 
mass curve for a time period equal to 2 t. Then the line, B' T, of this 
diagram will be parallel to the line, B" B, of the diagram, Fig. 8. If 
this line, B' T, prolonged cuts the time ordinate lying two-thirds of 
the way from i? to T at E, then D E will be parallel to the line, E' D\ 
of the diagram. Fig. 8, and the inclination of the line, D E, will give 
the desired value of dr- 

For every possible location of T a new position of E will be found. 
The successive positions of E will lie upon a curve which, as will read- 
ily be seen, is always the same no matter what the value of S may be. 
Therefore, to find the maximum value of dr, it will only be necessary 
to draw a tangent from the point, D, Fig. 9, to the £'-curve. 

The point of tangency thus found will be distant from B on the 
time scale two-thirds of the time, imJ and the inclination of the tan- 
gent will be the greatest possible value of dr. 

For the limiting rain curve of San Francisco, based on 

.3.68 



1 = 



4- t "■* 



t + 60 

and impervious surfaces (a = 1), the resulting locus of the point, E, 
or the £'-curve, is shown in Fig. 9. 

Suppose, now, that a value of Cr — C« has been approximated. 
Add to this Gt — Gr+h, obtained from Table 5, on the supposition 
that i = 5 and that tm will be about equal to ^o- 

Scale upward from B the value of ;S' thus obtained, and draw D E 
tangent to the E-cutyq at E. 

Then d^, ^= inclination of line, D E, 
and t^^^ = one and one-half times the scaled time from B to E. 

It will be convenient to complete the diagram by platting to suit- 
able scale the differentials of the E'-curve. A third curve will thus 
be obtained which may be called the discharge curve. This curve 
will give at once the value of d^ for any value of t„^. 

By calling t^ ^= ijy, a. quick first approximation of the value d^ 
can thus be made. 

It follows from the preceding demonstrations that for every mass 



340 



THE SEWER SYSTEM OP SAN FRANCISCO 



curve of surface run-off, on the hypotheses as made, there is a definite 
inter-relation between the values of Sm, im> and dm- This inter-rela- 
tion can be determined from the diagram, and can be tabulated. 

Ordinarily, the fact most readily ascertainable for any district is 
the value of tj), which (in the case of urban storm-water flow prob- 
lems) may be accepted as a first approximation of the value of tm- 
In tabulating the information obtained by diagram it will, therefore, 
be convenient to note regular time intervals in the first column. 

The values noted in Columns 2, 3, and 4 of Table 7 are based on 
the limiting rain curve for San Francisco as determined by the 
formula, 3 gg 



1 = 



2 t 



+ «»•'' 



t + 60 
and are applicable to areas with impervious surface throughout. 

The corresponding columns in Table 8 contain values based on 
the same limiting rain curve with reductions due to water absorption 
by soil when only 25% of the surface is impervious. (For the reduc- 
tion factors used see Table 6.) 

TABLE 7. — Storm-Water Flow in Its Eelation to the Duration of 
THE Critical Time Period, and to the Augmentation of the 
Water Quantity Temporarily Stored in the Conduits. 

For areas with impervious surface throughout. For San Francisco. 

3.68 



Based on J = 



2t 



t-j- 60 



+ f 





Augmentation 

of storage 

in the conduits, 

t'y — C^, 

in cubic feet 

per acre. 

(2) 


By Diagram. 




Duration of 
critical period, 

in minutes. 

(0 


Augmentation of 

storage on the 

ground and in the 

conduits, 

s, 

in cubic feet 
per acre. 

(3) 


Discharge, 

in second-feet 
per acre. 

(4) 


Discharge, 
by method used 
in 1899, calling 

*»« = ^Z>' 

in second-feet 
per acre. 

(5) 


5 
10 
15 
20 
30 
40 
50 
60 
70 
80 
90 
120 


5 
270 
365 
440 
545 
635 
725 
800 
870 
980 
985 
1120 


200 
300 
385 
450 
555 
640 
7.S0 
800 
870 
930 
985 
1 120 


1.25 
0.93 

0.78 
0.69 
0.57 
0.51 
0.46 
0.43 
0.40 
0.38 
0.36 
0.32 


1.34 
0.99 
0.82 
0.71 
0.60 
0.54 
0.49 
0.45 
0.43 
0.41 
0.38 
0.34 



THE SEWER SYSTEM OF SAN FRANCISCO 



341 



TABLE 8. — Storm-Water Flow, in Its Relation to the Duration of 
THE Critical Period and to the Augmentation During Such 
Period of the Water Quantity Temporarily Stored in the 
Conduits. 

For areas with 25% of the surface impervious. For San Francisco. 

3.08 



Based on 7 



2t 



f +60 



+ i"-* 





Augmentation 
of the storage in 


By Diagram. 




Duration of 


Augmentation of 
storage on the 
ground and in 
the conduits, 

8. 




Discharge, 

by method used 

in 1899 


critical period, 

tm, 

in minutes. 


the conduits, 
in cubic feet per 


Discharge, 

dm, 
in second-feet 


caUing f,„= to, 
in second-feet 




acre 


in cubic feet 
per acre. 


per acre. 




(I) 


(2) 


(3) 


(4) 


(S) 


5 


5 


140 


0.89 


0.63 


10 


170 


190 


0.60 


0.46 


15 


205 


220 


0.45 


0.39 


20 


230 


240 


0.37 


0.34 


30 


265 


270 


0.28 


0.28 


40 


290 


295 


0.24 


0.25 


50 


320 


320 


0.20 


0.23 


60 


345 


345 


0.19 


0.21 


90 


375 


375 


0.14 


0.18 


120 


415 


415 


0.12 


0.16 



Referring again to Fig. 8, it remains to be shown that no rain condi- 
tion preceding the time, B, can be assumed which will give the line, 
E' D', greater inclination than resulting from that which has already 
been declared to be the most unfavorable. 

Assume the extreme case of no rain preceding the time, B. In 
this case there will be no storm-water in the conduits and no water 
on the ground at the time, B. Consequently, C ^ =; 0, and Ob = 0. 

From Equation 11, /S = Cj, + Gj,; 

but, as has already been shown, the maximum value of Cn for 
extreme rain intensity preceding the time, B, is about one-half of C^; 
therefore, approximately, for this condition, 

where C b represents volumetric water contents of the conduits due 
to maximum rainfall conditions prior to the time, B. 
For the special case shown in the diagram. 



for t = 


50, 


C,- 


-C« = 


730, 


therefore 






Oy ^ 


1 460. 


From Table 5, 






Gj, = 


75 


consequently 






8 = 


1 535. 



3-12 THE SEWER SYSTEM OF SAN FRANCISCO 

If the critical time were the same for the condition of no rain 
preceding the time, B, as for the condition of heavy rain preceding B, 
then it would be proper to scale this value of S downward from T in 
Fig. 9, and it would at once be apparent, because the inclination of 
£"' D'^ is less than the inclination of E' D\ that the most unfavorable 
assumption is that of heavy rain preceding the critical period. But 
the durations of the critical periods are not the same in the two 
cases. On the same hypothesis on which Equation 5 is based, the 
critical period for the case of no preceding rain is of nearly twice 
the duration as for the case of heavy preceding rain. For the special 
values above noted, one-third of the critical period would be about 
30 min., and the corresponding value of d'^ would be the inclination 
of the broken line (Fig. 9) rising from the 30-min. point of the 
base line. 

This, also, is less than that resulting from the originally assumed 
unfavorable conditions. The same result can be shown for any other 
value of t and for any other rain intensities preceding the time, B, 
less than those originally assumed as the least favorable. 

The conclusion relating to the most unfavorable distribution of 
rain heretofore announced appears therefore to be correct. The rain 
rates that should be assumed as prevailing before the beginning of 
the critical period are the greatest rates that are possible, the rain 
being such that it falls with gradually increasing intensity until the 
time, B. 

The comparison of the results by the new method of computation 
with those obtained by the method of 1899, as shown in Table 7, indicate 
reasonably close agreement for areas with impervious surface. 

The comparison in Table 8 for areas which are only in small part 
impervious indicates a greater range of unit values for the new 
method than result from the older method of calculation. The method 
of 1899 probably gives too small values for small areas which are only 
partly impervious. 

It may be repeated that water quantities in the conduits at the 
times, B and T, can ordinarily be calculated with a sufficient degree 
of approximation. At the time, T, the main conduit is flowing at 
capacity at the point for which storm-water flow is to be estimated, 
and the inlets on all the conduit branches are at the same time receiving 
water at the run-off rate which corresponds to the last few minutes of 



THE SEWER SYSTEM OF SAN FRANCISCO 343. 

the critical time period. It is possible, therefore, to determine to what 
proportion of their capacity the conduits are charged at their inlet 
ends. With this information for the upper ends of conduits, and 
the knowledge that the conduits are full near the point under con- 
sideration, a close approximation of the water stage at the time, T, is 
possible, and G^ may be calculated. 

By a similar line of reasoning, the stage of water in the conduits 
can be determined for the time, B, and the value of Cg can also be 
calculated. 

But for first approximation purposes it will be well to establish the 
values of K for increasing areas. Then call the ratio of the water 
stored in the conduits at the time, B, to the water stored at the time, 
T, equal to the ratio of the mean run-off during the ^d minutes pre- 
ceding B to the mean run-oft" in the 1 1, minutes after B. For San 
Francisco this ratio, as already stated, may be called as one to two. 

G, 



That is, 




but 


G^=K C, 


therefore 


P _KG 

^B 2 , 


and Cy - 


r ^G 



To the value of Gj, — 0^ thus or otherwise determined add the value 
of Gj, — Gj, ^ J, as found in Column 7 of Table 5 for i = 5. The sum 
will be the value of 8 to be used, as already explained. 

When this method of determining storm-water flow is applied to 
a conduit system already constructed, it may happen that, owing to 
excessive or deficient storage capacity in the conduits, there will be 
wide divergence in the values of t^ and tj) . 

The value of tm niay be considerably less than to if the conduits 
are too small, if they are surcharged at critical times; and it may be 
materially larger than ^^ if along the line of the conduits there is 
unusual reservoir space to be filled before the maximum stage is 
attained. It will be nearly twice as great as to for the special case 
of no rain preceding the critical period. 

The graphical method of determining the storm-water flow is 
applicable to any limiting mass curve of surface run-off. It is 
applicable therefore to any such curve obtained by applying a reduc- 



344 THE SEWER SYSTEM OF SAN FRANCISCO 

tion factor to the run-off rates from areas with impervious surfaces. 

But when the problem is presented of finding the storm-water flow 
for districts having surfaces which are not impervious throughout, 
it will not be necessary to construct new curves. The original curve for 
im^iervious conditions throughout may be used to solve any such problem. 

When the value of ;S = Gj. — C^ 4- Cr^ — Ctt + i ^^^^ been 
determined for an area of any surface character, divide this value by 

the factor, a (to be found in Table 6), and proceed with the modified 

c 
value of S thus obtained {S' = - ), as though the area had an 

a 

impervious surface throughout. There will thus he obtained a modified 
value of d'fy which may be called cZ'y, such that 

d' = ~^' 
^ a 

therefore, dj, = a d'j,. 

That is to say, the value thus obtained from the curves or from 
the tables is to be multiplied by the same reduction factor before used, 
to give the value of d^. 

New Formula and Discharge Tables. — A study of the limiting rain 
curve of 1899 for San Francisco has led to the adoption of a formula, 
as already stated, which, it is believed, will be found generally appli- 
cable for the determination of rain intensities for all periods of time 
that come under consideration in urban run-off problems. 

In its general form the formula may be written 

h 



1 = 



2^ , A4 C21) 



+ «" 



t + m 

Here & is a constant for any locality, and t is the duration of the 
rain, in minutes. 

This formula gives the value of I, in inches per hour, or, which 
is the same thing, in cubic feet per second per acre. 

For San Francisco, h = 3.68 will give values closely agreeing with 
those determined in 1899. (See Table 3.) 

For Atlantic Slope conditions, ordinarily, h = 8 to 12. 

Wherever the maximum rain intensity for 1 hour is known, the 
value of h can be determined. It will generally be well, however, 
provided data are adequate, to base the value of h on determinations 
from the maximum rain intensities for 10 min., 30 min., and 1 hour. 
In most cases the use of an average value from these three determina- 
tions should satisfy every requirement. 



THE SEWER SYSTEM OF SAN FRANCISCO 345 

The rain intensities resulting from h = 3.68, 4, 8, and 12, in this 

formula, have already been noted in Table 3. 

Whether the expression recommended, or any other expression in 

the general form -r . ..-, 

I = cf (t) (22) 

be used in estimating the maximum rain intensity during the time, 
t, the fact is noteworthy that for all possible values of c or of 6 in 
the other case, the corresponding ordinates gf the resulting mass 
curves will be proportional to the values of c or & that have been used. 

This fact makes it possible to establish by a single diagram (for 
impervious surface conditions), base values of storage augmentation 
and of storm- water flow, being for the special case of fe = 1 or 
of c = 1. 

From such base values the storage and discharge for other values 
of h or of c, and also for any condition of the surface, can then be 
found by multiplication with the factors, ah or aCj as the case may be. 

Table 9 contains the base values of S^^, d^, and (Cy — C^),that is, 
('S^) a = 1, (clj ^ = ,, and (Oy— 0^)^=1, for the type of rain 

6 = 1 6=1 6 = 1 

curve herein recommended, as determined by diagram. 

The values in the last column of Table 9 conform closely to a 
curve the equation for which is 

0.71 



y 

Z T, 

t, 



_1%^^, 0.4 (23) 

«™ + 60 

This is the value of dm for the special case in which o = 1 and 

6 = 1. For any value of a and of h, therefore, 

^^m- Q-^1^^ = 0.71 a J (24) 

«™ + 60 + - 

As it frequently will be desirable to compute the discharge directly 
from the maximum rainfall in 1 hour, the value of h may be found 
in terms of R and substituted. 

For t = 60, there will he I = R. 

1= ^ (21) 

2 i I ^ 0.4 

r+i5o"^ 

^ = ^4 (^-^ 

h = 6.14 E (26) 



346 



THE SEWER SYSTEM OF SAN FRANCISCO 



Therefore, from Equation 24, 
4.36 a R 



fL =- 



2 L 



-=4.36 



L 



«^(') 



.(27) 



^m + 60 

TABLE 9.— Base Values of (Z^ and Ct— 0^; 
When a = l and & =: 1; 
In the Formula for Surface Kun-Off: 

a b 



r — a I 



2 t 



+ t' 



« + 60 

All values determined by diagram. 
(Subject to correction.) 



Duration of the 


Total storage 
increase, 


Increase of storage 
in the conduits, 


Discharge, 


critical period, 


'" 6 = 1, 


(Cy - C^) a = 1 
6 — 1, 


(d„,)a = l 
6=1, 


in minutes. 


in cubic feet 


in cubic feet 


in second-feet 




per acre. 


per acre. 


per acre. 


5 


54 





0.340 


10 


82 


73 


0.252 


15 


105 


100 


0.212 


20 


122 


120 


0.187 


30 


151 


1.50 


0.156 


40 


174 


170 


0.139 


50 


198 


197 


0.125 


60 


218 


218 


0.116 


90 


268 


268 


0.097 


120 


305 


305 


0.087 



In the practical application of this formula, a should be considered 
constant for each area, as already explained. Its value is directly 
dependent upon the rate of water absorption by exposed soil surfaces 
during heavy rains. The value of a for any value of h and any value 
of t, therefore, can be determined by experiment. For surfaces im- 
pervious throughout, the value of a is unity. 

Some other values of a preliminarily suggested for use in this 
formula are to be found in Table 6. 

The foregoing discharge formula will be directly applicable to the 
determination of the maximum discharge when, for any district 
under consideration, the value of t which will make dr^ a maximum 
is known. 

As a first approximation, without great error, t,n may be called 
equal to tj^. Consequently, a first apiiroximation of dj, by the 
formula is possible whenever the time can be estimated that it will 



THE SEWER SYSTEM OF SAN FRANCISCO 347 

take water to flow from the most remote portions of a district to the 
point at which discharge is being estimated. 
It can readily be seen, when 

_ b 

^~ 2^ ,fOA (21) 

« + GO 
that (G,~G,^,)=^ah[G,-G,^,]^^^ (28) 

b = 1 

and that (O^ — G'^) = a & [Cj, — CJ„ ^ i (29) 

b = 1 

consequently, also, S ^ a h (S)^ _ ^ (30) 

b = 1 

From the values given for (S^^)^ _ ^ in Table 9, for the special case 

b =^ 1 

of i = 5, the relation between t^ and (S^) ^ — i ^^^ ^^ reduced to 

6 --- 1 

formula. It will be found that, with a sufficient degree of accuracy for 

all values of i^^ which are likely to arise in connection with urban 

problems, 

(«m)a^l = 27.5 VV (31) 

6 zz; I 

therefore S^ = 27.5 a 6 •^t^ (32) 

'•'-(27§^y (^=" 

From Equations 11 and 32, 

C,-C^^ 27.5 a h v/^- (Q^-G,^,) (34) 

C^—C^ = 27.5 a h sjt^— a h (G^ — G^ ^ X ^ i r35^ 

and when i = 5, 

G,-G^=^ah [27. 5 s/T^- {G.-G^^ ,), ,] ^^^^^ 

It has already been shown that 

6 = 6.14 R (26) 

Consequently, from Equation 32, 

S^ = 169 a B V^ (37) 

and from Equation 36, 

C^-Cj,^aB [l69 vV_ 6.14 ^G^-G^^ X -^ i] . (33) 

Likewise, from Equation 37, 

t = ( ^"^ V (39) 



348 



THE SEWER SYSTEM OF SAN FRANCISCO 



and, convenient for purposes of preliminary approximation, from Equa- 
tions 27 and 37, 

1 

( -li-^4-< 0-0 



S„ 



38.8 v^L 



f,„, + (50 



(40) 



TABLE 10.— Storm-Water Flow. 

b 
For b ^ 4, in the Formula, I =~2l 



t + GO 






All values in this table are based on 

0.71 a h 



d^ 



4.36 a li 



2«, 



n _J_ / 



Sm = 27.5 v/i; 
and Cr- Cs = S^ - (Gr - Gr ^ ,) 
& = 4 and i? = 0.65. 



2 t. 



Hk L t 



trr. -f 60 





All 


750/0 


500/0 


250/0 


None 




IMPERVIOUS. 


IlfPERVlOUS. 


IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 
























period, 


_^ 




.u 


4:> 


^ 




^ 


4J 


^ 


+3 


^nv 


CQ « * 


Oi <s 


«) S^ aJ 


(U a> 


an i- aJ 


0^ a; 


a?S<iJ 


oj 6 


«!?,&; 


01 3) 


in 
minutes. 


1 Bl 




j c8 




1 U cS 




1 c8 




:2fe 
1 .a 2 


^i? 




O go, 


11 


f^2 ^ 


i- 


E-i^ fe 


's, 


E-i-S S 


° S 


^,.0 a, 


It 




^ ga 


^ gs. 


ID 


gft 


CD a 

(A 


OgP. 


5 


.35 


1.38 


35 


1.36 


30 


1.13 


30 


1.01 


25 


0.88 


10 


310 


1.02 


280 


0.91 


250 


0.79 


210 


0.67 


165 


0.55 


15 


400 


0.85 


350 


0.74 


295 


0.63 


240 


0.51 


190 


0.40 


ao 


480 


0.74 


410 


0.63 


345 


0.53 


355 


0.42 


205 


0.33 


30 


600 


0.63 


500 


0.53 


410 


0.43 


315 


0..33 


230 


0.23 


40 


690 


0.55 


580 


0.46 


460 


0.87 


350 


0.38 


235 


0.185 


50 


780 


0.50 


650 


0.43 


510 


0.33 


380 


0.35 


240 


0.155 


60 


850 


0.46 


700 


0.38 


550 


0.30 


400 


0.215 


345 


0.135 


90 


1050 


0.39 


860 


0.33 


660 


0.245 


460 


0.170 


260 


0.098 


120 


1210 


0.35 


980 


0.28 


740 


0.215 


510 


0.145 


275 


0.080 



Wlien, therefore, the increase of water storage during a critical 
time period can be determined, the duration of this period can be 
found from Equations 33 or 39, and dm can be estimated by formula, 
Equations 24 or 29; or the value of d^ can be taken from Tables 10, 
11, and 12, or from the general table. Table 13. Or, the value of d^ 
may be estimated directly from the value of S„^ and an approximate 
value of tm by the use of Equation 40. In this last case the estimated 
discharge is to be treated as an approximation until the value of t^, 
has been verified. 



THE SEWER SYSTEM OF SAN ERANCISCO 

TAELE 11.— Storm-Water Flow. 

h 



349 



For 6 = 8, in the Formula, I == — ^ 



f + 60+ ^ 
All values in this table are based on 

0.71 a b 4.36 a B 



cl„ = 



2 L 



L 



2 L 



tm + 60 ■ ^ 
S^ = 27.5 ah^t^ 
and Cr- 0^= S^-(G-0,^,) 
6 = 8 and i? = 1.30. 



i„,, + 60 



+ t 





All 


750/0 


500/0 


250/0 


None 




IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 














1 










period, 

in 
minutes. 


1 o <S 


dm 

second-feet 

per acre. 


1 of 

fr^'2 43 
gft 


dm 

second-feet 

per acre. 


1 =s 

gft 


dm 

second -feet 

per acre. 


gft 


dm 

second-feet 

per acre. 


cubic feet 
per acre. 


dm 

second-feet 
per acre. 


5 


70 


2.77 


70 


2.65 


65 


2.53 


65 


2.40 


60 


2.27 


10 


620 


2.03 


590 


1.91 


550 


1.79 


520 


1.68 


480 


1.56 


15 


810 


1.70 


760 


1.58 


710 


1.48 


660 


1.37 


600 


1.36 


,20 


960 


1.49 


900 


1.38 


830 


1.28 


760 


1.18 


690 


1.07 


30 


1190 


1.24 


1 100 


1.14 


1010 


1.04 


920 


0.94 


830 


0.86 


40 


1380 


1.10 


1 270 


1.01 


1 150 


0.91 


1040 


0.83 


920 


0.74 


50 


1 550 


1.01) 


1 440 


0.91 


1330 


0.82 


1 170 


0.74 


1010 


0.65 


60 


1710 


0.93 


1560 


0.85 


1410 


0.76 


1360 


0.68 


1100 


0.60 


90 


3 090 


0.78 


1900 


0.71 


1700 


0.64 


1510 


0.56 


1310 


0.49 


120 


2 420 


0.70 


3190 


0.64 


1960 


0.57 


1730 


0.50 


1500 


0.43 



Tables 10, 11, 12, and 13 have been prepared by the use of the fore- 
going formulas, all based on 

1/0.4 

i-f 60 
i = 5 
and an absorption of water by the soil at the rate of about § in. per 
hour at the beginning of a critical period, decreasing gradually to a 
rate of h in. per hour at the end of 1 hour. 

In the case of conduits already constructed, neither the value of 
Oy — Cjj, nor the value of cZ„^, taken from the tables for t^^ = «^, 
should be considered final. 

To facilitate the use of the formula and of the reduction factors, 
a, Table 13 has been prepared. In this table the value of 

0.71 



t^. + «o 



+ <« 



350 



THE SEWER SYSTEM OF SAN FRANCISCO 



appears iu Column 3, and the value of 
[27.5 ^t^-(G,. 



G„ 



5) 0=1 

-•6=1 



in Column 2. The value of a & is given in Columns 4 to 18. 

To find the discharge, by this table, multiply the value found in 
Column 3 by the appropriate factor taken from Columns 4 to 18; and 
to find the increase of the water stored in the conduits, multiply the 
values found in Column 2 by the same factor. 

The table, as already stated, is based on the assumption that i = 5; 
and that, during the time, tm, the absorption of water by exposed soil 
surface is at rates decreasing gradually from about § in. per hour to 
J in. per hour at the end of 1 hour. 



TABLE 12.— Storm-Water Flow. 

h 



For h = 12, in the Formula, I = 

All values in this table are based on 

0.71 a b 



2 t 



t + GO 



fL 



-+« 



4.36 a B 



2L 



-hL 



2 L 



and (7, 



Kr + 60 • ™ t,, + GO 

S,^ = 27.0 a h ^t^ 

C/J = ^m — (G^T— ^T+ 5) 

h = 12 and B = 1.95. 



+ i. 





All 


750/0 


50% 


250/0 


None 




IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 


IMPERVIOUS. 
























period, 


05 (C (P 


V 01 


«5 S 6 


S) . 


J^t6 




!■ !- 


0} • 

0) ® 


,cqy ® 
























in 
minutes. 


1 of 

Sft 




cubic f 
per ac 


second- 
per ac 


1 a^ 

gft 


■B gt. 


1 .2 ^ 
ga 




1 0=8 
gP. 


(t Pi 


5 


100 


4.15 


100 


4.03 


95 


3.90 


90 


3.78 


90 


3.65 


10 


930 


3.05 


900 


2.95 


860 


2.84 


820 


2.72 


780 


2.60 


15 


1210 


2.54 


1 160 


2.44 


1 110 


2.34 


1060 


2.21 


1010 


2.11 


20 


1440 


2.23 


1?80 


2.14 


1310 


2.03 


1 240 


1.92 


1 170 


1.80 


30 


1790 


1.86 


1700 


1.76 


1610 


1.68 


1520 


1.58 


1420 


1.47 


40 


2 070 


1.66 


1960 


1.57 


1840 


1.48 


1730 


1.39 


1610 


1.29 


50 


2 330 


1.50 


2 200 


1.41 


2 080 


1.33 


1 930 


1.25 


1790 


1.15 


60 


2 560 


1.39 


2 410 


1.31 


2 260 


1.22 


2 110 


1.14 


1950 


1.06 


90 


3 140 


1.18 


2 950 


1.11 


2 750 


1.04 


2 550 


0.95 


2 350 


0.88 


120 


3 620 


1.04 


3 390 


0.98 


3150 


0.90 


2 910 


0.83 


2 670 


0.77 



THE SEWER SYSTEM OP SAN FRANCISCO 



351 



10 to o» en ^ 05 lo I-* I-* ' 

OOOOOOOCrtOOt" 



Duration of critical 
period, t , in minutes. 



M N) K) I-' "-"-'►- •:; 

cooooooooo_, 

O '>-' '" '-' 'i-^ '-' k) to M » 
-jasOT CHOC osroto *>••<! 

"o o "o o o "o o o o o*- 

oooooooooo— ' 

COMMCOOSOaOSOSMM,^ 

io Vo CO CO 'ie V. '*• en bi cr. W 

moiootnooioai in~^ 

10 »o to JO io 10 »o to c« cc^^ 

01 en 05 05-! "-} ''JD O V-i CO 0> 

ooocjiomocnoo^^ 

>_n— i_i>_.iOJOtS*OfO to^^ 
-^'•O ^ CD O ^ to >£.■ Ci O^ 

ocioc;>ooc;iocn o^^ 

OI-^P-"-'l— ■-'►-'^K) 10^_^ 

CDO^^^COCn-:!CO»--*C500 

oocnoioioooino^ 

OCOCCCODQDQCOOOOX CX!__ 

oooo'oo'o'oo'o'O 
ooooooooo o— 

-) ^J »1 -3 -7 -J -^ -I ») -J ^ 

iO to CO CO CO rx Ji^ t*^ UT -> O 
oioioocntnocno o^' 

CnCJ»C;TC3505*7COOOCO«- 

oomocjTOTOtncn o- — 

tn or cu en C5 05 05 C5 Ci C5-^ 
enwoi;oooocn enc- 

rf^cn en en enc 

CO o I-" io eo c 
cnooooc 

ioto joio tot 

ooooo'c 
oooooc 

coeocococo>i^4^eneno:cn 
ooooenoenoen oi> 

OOOOOOOOi-"-'^ 

en bt b» 05 -"! oc o co '— ' co S, 

ooenooooenenoli; 

COCOCOCOOOOOO o^* 
OS -<! bo «g'>-i to eo ji.'-5 covi 
oooe«oooeno o-— 

ODCOCOCDCOCDCOCDO O^^ 
CD O h-i Vo eo en *7 CD to CS QO 

ooooenoocnool_- 



cubic feet 
per acre. 



second-feet 
per acre. 



All 
impervious. 



impervious. 



50o/o 
impervious. 



25^ 
impervious. 



None 
impervious. 



All 
impervious. 



75% 
impervious. 



500/0 
impervious. 



250/0 
impervious. 



*>enencncne;TCnenca Oi^- 


None 


cooi-'ioeoejT-icD — cnc^ 
cnoooooaiocn en^^ 


impervious. 


iotojoiotototototo to— V 


All 


ooooooooo 0^ 


impervious. 



D3 



75% 
impervious. 



500/0 
impervious. 



25% 
impervious. 



None 
impervious. 



II II 






S3 0- 
II II 



&2 









P 



P 






+ 



+ 



+ 



+ 



o 






cl 

CO 

O 



352 THE SEWER SYSTEM OF SAN FRANCISCO 

The type of the formula for rain intensity, as here recommended, 
is the direct outcome of an attempt to make the resulting values follow 
closely the San Francisco limiting rain curve for all values of t up to 
24 hours. 

The parabola, as a limiting curve, was tried, but, for higher values 
of t, departed too far from the curve of 1899 to be considered entirely 
satisfactory. 

When all time periods that come into consideration are less than 
2 hours, as will be the case in nearly all urban problems, a parabola 
can frequently be found which will closely approximate the limiting 
rain curve, and may be used in its stead. 

In such case let the equation of the limiting rain curve be 

q ^ e \^t (total rainfall in inches) (41) 

The equation of the surface run-off curve will then be 

q' = a e Vt (42) 

The value of a, as has already been explained, is dependent upon 
the amount of impervious surface in any area, and varies, too, with 
the duration of the critical period. For impervious areas, a is unity. 

For t = 60, the value of q will be the maximum rainfall in 1 hour, 
therefore, from Equation 41, 

B = e sjm . . (43) 

'=77^6 <"> 

e = 0.129 i? (45) 

therefore q' = 0. 129 a i? \/« (46) 

The maximum intensity of the rainfall during any number of 
minutes can be found from 

J = ^ = ^ (47) 

and 1=1^^^ (48) 

\/t 

By reference to Fig. 10 it can be seen that for a curve the equa- 
tion for which is 

q' = ae\/t (42) 

the equation of the line, B' T, can be written 

0.414 a e 
y= J a; + 0.586 a e \/t (49) 

For X = -^1 this equation will give the ordinate of the point, E 



TPIE SEWER SYSTEM OF SAN FRANCISCO 



353 



The locus of the point, E, will be expressed by 

0.828 ae t . ^ „,,„ /, 

+ 0.586 a e wt 



3^/t 

w = O.SQ2ae\^r 

or w ^ 1. 056 a e \/x 

Therefore the tangent to the £'-curve at the point, E, 

dio _ 0.628 ae 0.647 a e 
dx 



s/s 



vr 



(50) 

(51) 
(52) 



(53) 




WHEN THE LIMITING RAIN CURVE 
IS A PARABOLA 

Fig. 10. 

The vahie of -^ as thus found will be inches in depth in 1 miu. 

d X 

Therefore d^, in inches in 1 hour, or in second-feet per acre, 

0.647 a e 38.8 a e 



d^ 



60 



And, by substitution, 
5.00 aB 



'^L 



d^ = 



s/l 



^^^:^ cu. ft. per sec. per acre. 



(54) 
,(56) 



w — 

3 



354 THE SEWER SYSTEM OF SAN FRANCISCO 

The value of S, expressed in inches of depth, will be 

2 t /^0M7 a eX 

from which it follows (writing t^^^ for t, and S^ for S): 

S^ = 3 630 [0.8G2 a e s/t^— 0.431 a e VQ (50) 

/S,„j — 1 565 a e \/t^ cu. ft. per acre (57) 

From Equation 57, t^ = ( f"^ ) ' (58) 

By substitution froni Equation 45, 

t = ( ^"^ ^ ^ (59) 

and S^^ = 202 a i? V^ (60) 

Finally, from Equations 54 and 58, 

, 60 700 a e^ .,,-,. 

f^m= ^ (^1) 

and from Equations 55 and 60, 

'--A: ("^^ 

The value of a, for use in the foregoing equations, may be taken 
from Table 6, for the soil absorption conditions on which that table 
is based. 

For San Francisco, the value of e to be used in these equations is 
0.0775, the value of B, the greatest possible rainfall in 1 hour, being 
0.60 in. 

For the rainfall conditions prevailing on the Atlantic Slope, the 
value of e will ordinarily lie between 0.15 and 0.25. 

In order that the importance of using good judgment in the con- 
struction of the limiting rain curve may become apparent, Tables 14 
and 15 have been prepared. 

In Table 14 corresponding values of 8,,, and d,n are given for the 
various values of tf„, estimated in the one case by the use of formulas 
based on limiting rain curves of the San Francisco type, and in the 
other by the use of formulas based on the parabola type of curve. 
This table is of general application. 

In Table 15 the comparison is made between corresponding values 
of tf„ and d^ for selected values of /8',„; but in this case a specific 
value of R was assumed : R = 0.60. The figures on the same lines are 
not those that would result from an adaptation of curves of the two 



THE SEWER SYSTEM OF SAN FRANCISCO 



355 



types to the rain records, because properly selected curves may not 
have coincident values of R. 

For the San Francisco type of curves: 

h 
1 = 




2L 







«^ + 60 ' "™ 


and 


S... 


= 169 a I? Vi^ 


For the parabola: 








I 


60 e 




^K. 


5 a B /S',„ 




V(„ "0.4 t„. 




s,. 


= 202 aB -J, 



TABLE 14.— Comparative Values of Storm- Water Flow Esti- 
mated FOR Tw^o Types of the Limiting Kain Curve. 

For impervious areas tliroughout. 





FoK A Limiting Rain Cueve of the 


When the Limiting Rain Curve 


Duration 
of the 
critical 
period, 


San Francisco Type. 


IS A Parabola. 


^m, 


dm. 


Sm, 


dm. 




cubic feet 


second-feet 


cubic feet 


second-feet 




per acre. 


per acre. 


per acre. 


per acre. 


5 


378 i? 


2.13 i? 


452 i? 


2.24 R 


10 


535 iJ 


1.56 i? 


640 i? 


1.58 J? 


15 


655 i? 


1.30 i? 


780 i? 


1.29 iJ 


20 


755 i? 


1MB 


905 H 


1.12 i? 


30 


930 i? 


0.96 i? 


1 010 B 


0.91 R 


40 


1 070 i? 


0.85 R 


1 280 7? 


0.79 i? 


50 


1 200 ie 


0.77 ti 


1 430 R 


0.71 i? 


60 


1 310 i? 


0.71 R 


1 570 1? 


0.65 i? 


90 


1 600 R 


0.60 i? 


1 920 i? 


0.53 i? 


120 


1 850 i? 


0.54 i? 


2 220 R 


0.46 i? 



Application of the New Method to Special Cases. — Should it be 
desired to know the extent to which a branch conduit contributes 
to the discharge of a main conduit at the time, T, two procedures are 
possible. The discharge estimate may be made for the main conduit 
for points just above and just below the entrance of the sub-main. 
The difFerence between the two results will be the required discharge 
of the sub -main. 



356 



THE SEWER SYSTEM OF SAN FRANCISCO 



TABLE 15. — Comparative Values of Storm- Water Flow Estimated 

FOR Two Types of the Limiting Kain Curve. 

For B = 0.60. 

For impervious areas throughout, 
a = 1, and B = 0.60. 





For a Limiting Rain Curve op the 


When the Limiting Rain Curve 


Increase of 


San Francisco Type. 


is a Parabola. 


storage, 






in cubic feet 


',, 


d,„, 


t„, 


dm. 


per acre. 


minutes. 


second-feet per 
acre. 


minutes. 


second-feet 
per acre. 


200 


3.89 


1.43 


2.72 


1.82 


250 


0.08 


1.17 


4.25 


1.46 


300 


8.75 


0.99 


6.11 


1.21 


400 


15.6 


0.77 


10.9 


0.91 


500 


24.2 


0.64 


16.9 


0.73 


600 


35.0 


U.54 


24.4 


0.61 


700 


47.5 


0.47 


33.4 


0.52 


800 


62.0 


0.42 


43.6 


0.45 


900 


78.6 


0.38 


55.0 


0.40 


1000 


97.3 


0.35 


68.1 


0.36 



Or, as may frequently be convenient, the sub-main is considered 
separately. In this event it must be remembered that the mass curve 
of surface run-off, which is of most unfavorable shape for the main 
district, is as shown in Fig. 8. During the first part of the critical 
period, the rainfall will have its greatest intensity. The run-off pro- 
duced by the rain of greatest intensity will have time to escape from 
the sub-district (supposed to be relatively small) before the main 
conduit flows at capacity. Therefore, in applying the graphical method, 
only that part of the run-off curve is to be used which lies nearest to T. 
But, for decreasing intensities of rain, in a small district, there will 
probably be less water in the conduits at the time, T, than i^ minutes 
before T (tj^ now applying to the sub-district). Consequently, instead 
of an augmentation of water quantity stored in the conduits there will 
be a depletion of this storage. The value of Oj, — G^ or of S, as the 
case may be, will be negative, and must be platted upward from the 
curve at T, instead of dovpnward. 

In the supposed case. Fig. 8, the sub-district time is 20 min. and 
the required discharge of the sub-main is represented by the inclination 
of the line, E^^ !>,. 

It may be of interest to note that the described method of estimating 
flow at any point of a water conduit when the rates are known at which 



THE SEWER SYSTEM OF SAN FRANCISCO 357 

water is supplied to it by its feeders, is applicable to rivers and canals* 
as well as to covered conduits, and also to conduits operating in part 
under pressure. 

A closed conduit serving throughout under pressure is an extreme 
case. There can then be no augmentation of the quantity of water 
stored in the conduit. The effect of an increase or decrease in the 
rate of supply will be instantaneous from end to end of the conduit. 
The time, i, will become zero. The points, B, T, and E on the curve 
will all be at the origin of the co-ordinates, and the maximum rate of 
flow at all points of the conduit will be equal to the maximum rate 
at which water is supplied, because, for ^ = 0, the £'-curve is tangent 
to the supply curve at the point, B. 

At the other extreme is the case of flow through large reservoirs 
or basins. In this case, the storage of water is large, the time, i, is 
relatively long, and the maximum rate of outflow is correspondingly 
small. 

In applying the method of estimating required storm-water capacity, 
as described, it is well to remember that a proper determination of the 
value of h eliminates the necessity for introducing into the calculation 
any further allowance for safety margin except such as may be covered 
by a selection of a value for the factor a. As a is unity for areas 
with impervious surfaces throughout and as the value of a is otherwise 
definitely determinable, it will be seen that the room for error is small. 

General Deductions. — It appears from the foregoing: 

1. — That the method as described of estimating storm-water flow 
is applicable to districts of any shape, size, or surface condition. 

2. — That if a conduit affords no reservoir space, the storm-water 
flow at all points thereof would be equal to the rate of inflow at the 
storm-water inlets, and the effect of an increase or decrease of this 
rate would be immediately manifest at all points of the conduit. 
(This is the case of a conduit under pressure.) 

3. — That the longer the time required for storm-water concentration, 
the larger will be the unit storage capacity of the conduits in any 
district, or vice versa, and the greater the influence of the conduit 
storage upon the rate of flow. 

* This is substantialJy the method of estimating river flow defcribed by the writer in a 
recent discussion of a paper on a California flood. Transactions, Am. Poc. C. F., Vol. LXI, 
p. 332. 



358 THE SEWER SYSTEM OF SAN FRANCISCO 

4. — That an outfall conduit, receiving no further accessions, below 
a certain point: 

a. — If occasionally or always submerged must be given uni- 
form capacity from that point to the point of outfall; 

h. — If not flowing under pressure, and if it have a free out- 
fall not obstructed by back-water, its maximum discharge 
will decrease somewhat with increasing distance below that point. 

If the shape of the limiting rain curve has been properly deter- 
mined from adequate rainfall observations, the secondary curves derived 
therefrom should be smooth curves. The value of discharge deter- 
mined therefrom will, in other words, follow a definite law of change. 

It is obvious that the graphical method of determining discharge 
may be applied to any rain storm for which a mass curve of the rain- 
fall can be platted. The proper location of the points, B and T, on 
the resulting surface run-off curve will be found by trial. 

The method of estimating storm-water flow as here described com- 
mends itself for use whenever the limiting rain curve is known : 

First. — 'Because the maximum fall of rain in all possible time in- 
tervals is taken into account. 

Second. — Because the element of uncertainty relating to the rate 
of rain-water inflow into storm-water inlets is confined within narrow 
limits, and the possible error in the selection of a reduction factor is 
correspondingly small. 

Third. — ^Because the essential element, the increase of water storage 
in the conduit system during a critical period of time, is subject to 
determination with reasonable accuracy. 

Fourth. — ^Because it is simple and generally applicable. 

Storm-Water Relief Outlets and Sewer Capacities. 

The ordinary principle of providing storm-water relief outlets from 
the sewers of the combined system, where opportunity offered, has 
been adhered to in designing the storm- water conduits for San Fran- 
cisco. The infrequency of rain in excess of i in. per day made it 
appear desirable to let the relief outlets come into service only when 
the rainfall rate exceeded this amount. In other words, the sewers 
were planned to carry the run-off due to a rain of i in. in 24 hours 
to the sewer outfall. The remainder of the flow during storms will pass 
through the relief outlets on the shortest course to bay or ocean. 



THE SEWER SYSTEM OF SAN' FRANCISCO 359 

The sewers of the combined system, therefore, were given a capacity 
above relief outlets to carry: 

The sewage proper; 

A certain amount of soil-water, entering by leakage; 

The run-off due to a rainfall of greatest intensity; 

Some allowance as a safety margin. 
Below relief outlets: 

The sewage proper; 

A certain amount of soil-water, entering by leakage; 

The run-off due to a rainfall at the rate of i in. in 24 hours. 
The sewer capacity thus determined for points below relief outlets 
is nearly four times as great as the mean flow of sewage proper. The 
storm-water reliefs, therefore, should not come into service with more 
than 25% of sewage in their outflow, and this amount only at times 
when their outflow is small. When in service at full capacity the 
sewage carried by them will be so dilute as to be quite inoffensive. 

Sewage Disposal. 
Dilution and Outfall Points. 

Consideration was given to the various methods of sewage disposal 
which have been developed and found satisfactory. 

Utilization by application to land was entirely out of the question, 
owing to San Francisco's location at the extreme point of a narrow 
peninsula, and the remoteness of lands which would be suitable for 
cultivation under sewage irrigation. 

Treatment in filter beds, to produce an inoffensive effluent, or chem- 
ical treatment for like purpose, would entail much additional cost 
and is not at present necessary for any part of San Francisco's sewage, 
owing to the great dilution that can be secured, and the favorable set 
of the currents from the selected outfall points away from the shore. 

Dilution in the waters of the bay and the ocean is without doubt 
for San Francisco the natural and proper method of sewage disposal. 

The questions to be considered, therefore, relate to the best points 
of outfall and the best method of conveying the sewage to those points. 

Without discussing at length the studies of tidal currents made 
in 1893 with floats, and the results of current measurements made 
from anchored boats by the United States Coast and Geodetic Survey 
some years before, the conclusions reached will be briefly stated. 



360 THE SEWER SYSTEM OF SAN" FRANCISCO 

The float observations did not cover the entire water front, nor all 
possible stages of the tide. The means were not at command to ac- 
complish this. Generally known facts or opinions relating to the 
direction and velocity of the bay currents led to the selection of 
certain points from which the floats were started at various stages of 
the tide and then followed day and night by boat until stranded or 
well started seaward. The float was of wood about 6 ft. long, weighted 
at one end and carrying a flag or lantern at the other. 

Float observations were made from points off Hunter's Point, Potrero 
Point, Center Street Pier, and Powell Street Pier. At each of these 
points floats were started at low tide, at high tide, and at intermediate 
stages. There were no float observations made oceanward from Golden 
Gate as these did not seem necessary. The current in and out is 
swift. Someone has computed that 1 cu. mile of water passes through 
Golden Gate each way every day. This vast volume, as it flows out- 
ward, entrains to some extent the water that lies off the beach, south- 
ward from Point Lobos. This is probably the reason why the ocean 
beach is exceptionally free from drift. 

All the Hunter's Point flood-tide floats passed up the bay, clear of 
tide marsh and shore line. The ebb-tide floats cleared Potrero Point by 
more than half a mile and passed east of Mission Rock. 

Floats started at flood tide from the Sugar Refinery wharf at 
Potrero Point took a course inside of a line drawn from Potrero Point 
to Hunter's Point; one was caught and held in the flood-tide eddy 
south of Potrero Point. At ebb tide all floats passed to the east of 
Mission Rock into the deep water of the bay. 

The floats started at flood tide from the wharf of the Pacific Rolling 
Mill, farther north on Potrero Point, all hugged the shore toward 
Hunter's Point. The ebb-tide floats showed equally unfavorable con- 
ditions. Their course was close to the end of Center Street pier, to the 
westward of Mission Rock. 

Floats started at flood tides from the end of Center Street pier, 
oil wharf, passed close to Potrero Point. Those at ebb tide moved 
toward Mission Rock. 

The floats started at the end of Powell Street pier on flood tides in 
part took a course somewhat off shore into deep water, in part they went 
gradually shoreward. The ebb-tide floats all went off shore more or 
less, most of them clearing Black Point by more than one-quarter 
mile. 



THE SEWER SYSTEM OF SAN FRANCISCO 361 

The conclusion reached from the foregoing observations was that 
sewage might safely be discharged into the bay or ocean on the northern 
frontage of the city, provided only that the outfall be arranged far 
enough off shore in deep water. 

The eastern frontage of the city, from near Market Street to near 
Hunter's Point, is not a desirable locality for disposing of sewage. 
It was recognized, however, that the discharge of sewage in limited 
amounts through submerged outlets in deep tidal waters is rarely 
objectionable, and, therefore, that small shore areas which cannot at 
once be made tributary to the contemplated main sewers, may, with 
proper restrictions, be allowed to discharge into the bay. 

Hunter's Point was found to offer exceptional advantages for a 
delivery of sewage into swift tidal currents with ample and thorough 
dilution. 

The following points of outfall were finally selected: 

Off North Point, with discharge in at least 36 ft. of water at 1 200 ft. 
or more from the water-front line. 

Off Hunter's Point, in a corresponding depth. 

Off the ocean shore just west of Baker's Beach, at the foot of 27th 
Avenue. 

Off the foot of Scott Street, in at least 36 ft. of water, and well off 
shore. 

Two other points were selected, though not to be utilized at once. 
These are Fort Point, where it may be found desirable after many years 
to concentrate all sewage reaching the northern city front, and Point 
Lobos, where it was foreseen that the sewage of the Sunset and other 
ocean-slope districts should ultimately be delivered. 

Of the areas tributary to the selected outfall points it may be said: 

The North Point main is to receive the sewage from an area of 
about 11 500 acres, which includes the greater portion of the residence 
district of the city and the entire business section. About 80% of the 
city's population was within this area at the time the report was 
written in 1899. 

The territory tributary to the Hunter's Point main comprises 2 800 
acres. 

The Point Lobos outfall will ultimately receive the sewage from the 
greater portion of the ocean slope of the city. 

The Scott Street outfall will serve the Harbor View district, until 
it becomes desirable to extend the North Point main to Fort Point. 



362 THE SEWER SYSTEM OF SAN FRANCISCO 

The conclusion having been reached that the best points for the 
discharge of sewage are the projecting points of the water front which 
are swept by swift tidal currents, it became necessary to give Fort 
Point special consideration. Fort Point, at the south side of Golden 
Gate, is without doubt the best location for an outfall, but, for the 
present, it is inaccessible, owing to remoteness from the built-up section 
of the city, and the great cost that would be involved in reaching it in 
advance of general water-front improvement. 

In the order of their desirability as points of outfall for the 
sewage of bay-slope districts there seems to be no question that Fort 
Point should be ranked first, though, as explained, not now available; 
North Point, second, and Hunter's Point, third. 

Consequently, the collecting system of sewers has been planned so as 
to make the greatest area possible tributary to the North Point outfall, 
and let Hunter's Point take care of the remainder. 

This is the fundamental reason why, as elsewhere explained, the 
upper portion of the Islais Creek drainage is to be cut oS from the 
Hunter's Point system and made tributary by gravity flow to the North 
Point main, by tunneling along San Jose Avenue under College Hill. 
By this arrangement, moreover, an immediate installation of the outfall 
sewers for an important outlying district becomes possible, and it 
has the further advantage of reducing the amount of sewage to be 
pumped over from the Hunter's Point main into the North Point sewer 
system if, in the future, it ever becomes necessary to abandon the 
Hunter's Point outfall. Furthermore, it appears unobjectionable to per- 
mit the temporary delivery of sewage in small amounts at other points 
of the bay front, provided the discharge be always effected well off 
shore in deep water. 

Sewerage Districts. 

Sewerage districts were thereupon laid out in conformity with these 
principles, and the plans for the systems of collecting conduits were 
completed. 

The areas noted in Table 16 and the estimated population, which is 
really the assumed distribution of population when San Francisco 
has reached the 1 000 000 mark, were used as guides in estimating, by 
methods already described, the quantity of sewage, soil-water, and 
storm-water to be cared for in the sewers. 



THE SEWER SYSTEM OF SAN FRANCISCO 363 

TABLE 16. — San Francisco Sewer Districts. 



Name of district and sub-district. 


Area, 
in 

acres. 


Estimated 

future 
population. 


North Point District: 


836 
.381 
246 
.579 
583 
177 
794 
149 
1125 
620 
660 

1 307 

27 
978 
334 

782 
41 
845 
126 
157 
279 
500 

11525 

122 

68 
91 
123 

488 
65 
77 

115 

■sm 

590 
249 

2 294 

816 

1200 

212 

1 658 

2 564 
783 


24 800 




8 500 




7 300 




21700 


Glen Park. 


17 500 


Silver Avenue West 


5 300 


Valley Street and UpTifir Armv Street 


31700 


West Bernal Heights 
Lower Islais Creek.. 


' 


6 000 




47 900 


West Potrero, Upper 
Eighteenth Street... 




31000 




33 000 


Fourteenth Street 


74 500 




1600 


Eleventh Street 


54 400 


East Potrero 


16 700 


Fifth Street, Sixth 


Street, Seventh Street and 


103 700 


Kearny Street 


6 300 


Mission Flats .... 


72 500 


North Rincon Hill 


15 700 


Portsmouth 


31 000 


Yerba Buena 


33 500 


North Beach 


30 000 




Totals 


674 500 


Hunter's Point District: 

College Homestead... 




3 700 




2 000 


St. Mary's 


3 600 




4 900 


University Mound 




14 600 


West Silver Terrace and Miscellaneous 

North Silver Heights 


2 000 

3 100 


Railroad Avenue 




4 600 


North Slope, Hunter' 


s Point 


12 200 


Bay View 




23 600 


South Slope, Hunter' 


s Point 


10 000 




Totals 


86 300 


Harbor View District 




48100 


Richmond 


60 000 


West Richmond 


19 300 


Upper Sunset 


64 200 


Lower Sunset 


98 600 


Ocean View 


29 800 







Population, Water Consumption, and Sewage. 

Before describing the district systems, a word is to be said relating 
to the prospective growth of San Francisco, and the water consumption. 
The population estimate, as made in 1899, is shown by the diagram. 
Fig. 11. For comparison, the population of several other cities of the 
United States is also shown. The conclusion was reached that about 
50 years would elapse before San Francisco would have a population 
of 1 000 000. Fifty years being a reasonable time for which to insure 
adequate service to all parts of the city, and fully realizing that the 



364 



THE SEWER SYSTEM OF SAN FRANCISCO 



forecast now made for a period so remote is to be regarded other than 
as a probability, this population was accepted as that on which esti- 
mates of water consumption and, therefore, volume of sewage proper 
and the like, should be based. 

The information presented in Fig. 11, and extended to include the 
results of the census of 1900, is shown in Table 17. 

1 200 000 



POPULATION 

OF 

SAN FRANCISCO 




y 



-1 1 100 000 
1000 000 
900 coo 
800 000 
700 000 
600 000 i 
500 000 \ 

p 

400 000 
300 000 
200 000 
100 000 



1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 

Fig. 11. 

TABLE 17. — Population of Cities from United States Census, and 

Estimated. 



Year. 


Washington. 


Cleveland. 


Cincinnati. 


San Francisco. 


Baltimore. 


Boston. 


1840 


23 364 


6 071 


46 338 




102 313 


93 383 


1850 


40 001 


17 034 


115 435 




169 054 


136 881 


1860 


61 122 


43 417 


161 044 


56 802 


212 418 


177 840 


1870 


109 199 


92 829 


216 239 


149 473 


267 354 


250 526 


1880 


177 624 


160 146 


255 139 


233 950 


332 313 


362 839 


1890 


230 392 


261383 


296 908 


298 997 


434 439 


448 477 


1900 ■ 


278 718 


381 768 


325 902 


342 782 


508 957 


560 892 


1910 








*480 000 






1920 








*590 000 






1930 








♦700 000 






1940 








*830 000 






1950 








*960 000 






1960 








*1 100 000 







* Estimates made in 1899. 



THE SEWER SYSTEM OF SAN FRANCISCO 365 

As a result of the destruction of the business section and a large 
part of the residence district of San Francisco by fire in 1906, there 
was at that time a sudden drop in the actual population curve. 

If recast at the present time there would, no doubt, be some de- 
parture from these estimates. 

From information available in 1899, it was found that, exclusive 
of water pumped from bay or ocean for the supply of bathing tanks, 
for use in street sprinkling, and for > use in condensers, about 70 gal. 
per capita per day were supplied to the city by the Spring Valley 
Water Company. The maximum daily per capita consumption at that 
time was 82 gal. 

It is believed that San Francisco's cool climate, particularly in the 
summer, the compact development of the residence districts, with 
restricted lawn and garden areas, and the natural difficulties in the way 
of bringing in a copious supply, will always compel more than ordinary 
economy in the use of water. It is not likely, therefore, that the 
present consumption of 75 gal. per day will ever be greatly exceeded. 
Allowing 20% as the excess of the maximum daily consumption over 
the mean, and making a small allowance for water taken from ocean, 
bay and private wells (4% was assumed), the conclusion was reached 
that 94 gal. per day should be taken into account in estimating the 
volume of sewage proper. 

About 75% of the water used for all purposes in the city will 
reach the sewers. But the rate at which the water is used and delivered 
to the sewers is not uniform. It was assumed, without special local 
studies to confirm the accuracy of this assumption, that the maximum 
flow would exceed the mean by 50%. 

Based on these premises, it was estimated that for each 1 000 inhab- 
itants the average quantity of sewage proper, on days of maximum 
water consumption, would be 0.109 sec-ft., and that the maximum rate 
at which this would reach and be carried off by the sewers would be 
0.163 sec-ft. This was the unit adopted for use in estimating the 
required capacity. 

Soil-Water. 

The quantity of soil-water entering the sewers by leakage depends 
on the care with which the sewers are constructed, their permeability, 
leaky Joints, and also the relative elevation of the water-table and 
sewer. When the water-table is below the sewer there may be leakage 
from the sewer into the ground, but there will be no reverse flow. 



366 THE SEWER SYSTEM OF SAN FRANCISCO 

Throughout a large portion of San Francisco the sewers will lie 
above the soil-water plane. If winter rains bring the water-table up 
to cover the sewer, in the high parts of the city, it will be for short 
periods only. Therefore, very little inflow from the soil into the sewers 
is to be assumed, except in the low marginal and down-town areas, 
where the plane of saturation lies above the sewers. It would be 
interesting to know whether this plane could not be permanently 
lowered, in view of the fact that the bay front of the city is, or was, 
skirted by mud flats which are to be considered nearly impervious. 
But this cannot be ascertained short of very expensive experiments. 

If there is a great leakage into the sewers, they may have some 
effect on the water-table, and the leakage will gradually decrease 
until equal to the ground-water replenishment, from whatever source 
this may come. In the absence of adequate information relating to 
ground-water depth in different parts of the city, a uniform allowance 
for an increment of the flow in the sewers due to this cause was made 
for all the high portions of the city and a larger allowance for the low 
areas. The ground-water reaching the sewers from high ground was 
introduced into the calculation of sewer capacity at 0.001 sec-ft. per 
acre, and that from the low parts of the city, where the sewers, for 
the most part, will lie below the natural soil-water plane, at 0.003 
sec-ft. per acre. 

The Tides of San Francisco Bay. 

The range of tides in San Francisco Bay is thus referred to in 
the Pacific Coast Pilot, by George Davidson, Hon. M. Am. Soc. C. E., 
who for many years was in charge of the work of the United States. 
Coast and Geodetic Survey on the Pacific Coast: 

"From the lower low water ('low water large') the tide rises for 
about 7h hours, say 4.4 ft. to the smaller of the two high tides (%igh 
water small'), then falls 1.4 ft. in less than 4* hours to the 'low water 
small,' which is higher than the preceding low water; then rises say 
2.9 ft. in 6i hours to the higher high water or Taigh water large;' 
it then falls again 5.8 ft. in over 7 hours to the lower low water, or 
'low water large.' 

"Instead of the above figures, the fall from high water small, or 
'half tide' to the 'low water small,' may range from 3J ft. at one 
position of the moon to 0.3 ft. at another; in the latter case there 
will be apparently a long stand of about 5 hours. 



THE SEWER SYSTEM OF SAN FRANCISCO 367 

"The average difference of the higher high and lower low waters 
of the same day is 6.2 ft., with a greatest observed range, noted for 
February 8th, 1876, of 9.93 ft." 

The "Tide Tables" for 1899 of the United States Coast and 
Geodetic Survey gave the following information: 

The expected mean range of tides is 4.5 ft. 
The expected mean range of spring tides is 5.4 ft. 
The expected mean range of neap tides is 3.5 ft. 
The expected mean range of great tropic tides is 7.2 ft. 
The expected maximum range of tides in 24 hr. is 7.8 ft. 
The expected minimum range of tides in 24 hr. is 3.1 ft. 
From these tidal ranges, and the area of the bay, together with 
the area and elevation of the adjoining svibmersible land, and with 
proper allowance for the difference in the time at which the high 
and low tides occur, the total volume of flow through the Golden 
Gate can be estimated. 

To the outflow through the Golden Gate, however, there is to be 
added the flow of the rivers tributary to the bay. This is not 
inconsiderable. 

It is the run-off from more than 60 000 sq. miles of tributary area. 
The two rivers, the Sacramento and the San Joaquin, contribute 
more than 4 000 sec-ft. at their lowest autumn stage, and more 
than 80 000 sec-ft. throughout the first six months of the year, often 
reaching from 150 000 to 200 000 sec-ft. for short periods of time. 

River flow through tlie bay alone would effect sewage dilution 
from 30 to 1 500 fold, yet the river water is only a trifle when com- 
pared with the great return flow of waters which have entered the 
Golden Gate on a rising tide. 

North Point Sewer District. 

The North Point Sewer District has an area of 11 525 acres. Its 
estimated population, when that of the city, of present area, is 1 000 000, 
will be 674 500. 

About 80% of the city's population in 1899 lived within this dis- 
trict. It includes the northern frontage of the city easterly from 
Black Point, the eastern frontage as far south as the hills on the 
south side of Islais Creek, and extends inland to the summit of the 
peninsula ridge and along the same, south to, and slightly beyond, the 
county line, including all the Islais Creek water-shed already described. 



368 THE SEWER SYSTEM OP SAN FRANCISCO 

Nearly all this district is to be sewered on the combined system. 
The old sewers are utilized to the greatest extent possible. 

The position of the main intersecting sewer was determined by two 
governing points, the outfall at North Point and the point on the 
old line of Mission Creek where the sewage of the Mission is to be 
intercepted. The sewer was placed low enough to serve the district 
near Harrison and Fourteenth Streets. At that point many street 
grades have been established without due regard to future require- 
ments, and to the permanent injury of adjoining property. Improve- 
ments of street surface and improvements on private land have been 
made to conform to these grades, many of which are too low. It is 
doubtful whether material correction of such grades will ever be 
possible. The North Point main, therefore, was planned to fit the 
grades as established, the aim having been to keep the hydraulic grade 
line at least 1 or 2 ft. below the street surface at the critical points. 
This was the best that could be done for this district which, under 
present conditions, is submerged whenever there is a heavy downpour 
of rain. In order that the reason for small safety margin, for the scant 
elevation at a few points of street surface above the hydraulic gradient 
of the full conduit, may be better understood, it may be stated that there 
were such points, as at Eighteenth and Division Streets, where the 
grades marked out a pot-hole arrangement. The official grade there 
at the time the sewer system was planned was 7.5 ft. above city 
base,* while at all street intersections around this point the official 
grades ranged from 10.5 to 13 ft. At Thirteenth and Harrison Streets 
there is a similar arrangement. Street grades of 6 and 6.5 ft. are 
entirely surrounded by official grades several feet higher. Official street 
elevation is similarly defective at the northern termination of York, 
Florida and Alabama Streets, and at the intersection of Fifteenth and 
Shotwell Streets; also at Sixteenth and Shotwell Streets; at the eastern 
termination of Fifteenth Street, on Folsom Street midway between 
Sixteenth and Seventeenth Streets, and at Fourteenth Street; and there 
are still other points where grades should be changed so as to eliminate 
completely the pot-hole feature. 

In the past this district has suffered one inundation after another. 
The proposed increased capacities of the sewers and storm-water drains 
will greatly ameliorate this condition, but the mischief done by the 

* The grade at this point was raised a few years after the report was^written. 



THE SEWER SYSTEM OF SAN FRANCISCO 369 

establishment of faulty street grades cannot be entirely undone by the 
construction of sewers, and it must be left to the property owners to 
find out how much better off they would be if the high-water line in 
the sewers could be kept 5 or 6 ft. below the street surface instead 
of rising almost to it. The only satisfactory relief from the local 
floods that may result from defective or choked storm-water inlets and 
from the annoyance of water in cellars, or, at least, of interrupted 
cellar drainage during severe rain storms, will result from raising 
the street grades, particularly at the points where pot-hole features 
are to be eliminated. The present property owners are often the 
innocent sufferers, from the desire of their predecessors to have official 
grades established at natural ground height or but little above it, in 
order to save the cost of grading. 

The head of the North Point main, if considered as an interceptor, is 
far out on the Mission Road, at the point, in fact, where this road 
crosses Islais Creek. From this point the sewer will be in tunnel for 
a distance of 4 500 ft. following the line of Mission Road and San 
Jose Avenue. 

At Army and Valencia Streets the sewage now flowing down Army 
Street sewer is to be intercepted and diverted northward. The present 
sewer eastward from this point will continue in service as a storm- 
water relief outlet, and as a main from which a mile farther east 
sewage will be pumped to the nearest point of the North Point main. 

The course of the North Point main is almost due north following 
the streets best adapted by reason of elevation to receive it. It will 
be on Twenty-sixth Street from Valencia Street to Treat Avenue, on 
Alabama Street from Twenty-fifth to Nineteenth Streets, and will 
reach Division Street at Eighteenth Street. It will follow Division 
Street to Harrison Street, paralleling the present Channel Street 
sewer; thence it will be on Harrison Street to Eighth, on Eighth to 
Howard, thence on Howard and New Montgomery or Second Street 
northwest to Market Street and north to the water front on either 
Montgomery or Sansome Street. All along its course it will intercept 
the sewers entering it from the west. 

At Eighteenth and Division Streets, at Fifteenth Street, and about 
on the line of Fourteenth Street, will be storm-water relief outlets of 
large capacity. The outlet arrangement at Eighteenth Street is practi- 
cally that of two parallel conduits with a common roof, separated from 



370 THE SEWER SYSTEM OF SAN FRANCISCO 

each other by a dividing overfall weir. The entire flow will be 
confined by this weir to the North Point main until the full capacity 
is reached, whereupon the excess flow will go over the weir into the 
storm-water conduit. 

Other storm-water outlets will be provided at Seventeenth, Six- 
teenth, Eleventh, Seventh, Sixth, and Fifth Streets, also at Mission 
Street and at Commercial and Jackson Streets, as well as a final one at 
the North Point screen-house, where the sewage will enter the outfall 
pipes. 

Large conduits will carry the water from each of the storm-water 
outlets in the most direct line to the open water in Channel Street, or 
to the bay. Of these conduits, the only one of special interest is the 
Division Street conduit which will have its head at Eighteenth and 
Division Streets, but will be the recipient also of the storm-water over- 
flow at Seventeenth, Sixteenth, Fifteenth, and Fourteenth Streets, and 
at several other points along its course to the head of the open water 
in Channel Street. 

This conduit, shown on Plate XIII, will be called upon to carry ulti- 
mately about 760 sec-ft. near Eighteenth Street, and about 1 YOO sec-ft. 
at its outfall end. It will be a conduit 9 ft. high and 11 ft. 9 in. 
wide from Eighteenth Street to Tenth Street, being paralleled by a 
second conduit from Alameda Street to Tenth Street. Thence to its 
outfall, for a distance of about 2 000 ft., the conduit will be constructed 
with two or three compartments. Three, each 9 ft. wide, will be 
required ultimately. It may be found advisable, however, to con- 
struct only two of these at the outset. Owing to the fact that the 
floor of the conduit must be kept as high as possible on account of the 
unfavorable character of the ground along its route, and because the 
street surface along the line of this conduit is lower than desirable, 
scant headroom is available, and the conduit has been planned with 
a flat top of reinforced concrete. 

It has been suggested by interested citizens that the main storm- 
water outfall should be at the foot of Sixteenth Street, and by some 
it is even contended that the sewage of the entire area that can be 
made tributary to the intersection of Sixteenth and Division Streets 
should be carried in a large conduit along Sixteenth Street to the bay. 

Taking Fourteenth and Harrison Streets as the initial point for 
such an outfall, it can readily be seen by a glance at any map of the 



PLATE XIII. 

TRANS. AM. 80C. CIV. ENQR8. 

VOL. LXV, No. 1127. 

GRUN8KY ON 

THE SEWER SYSTEM OF SAN FRANCISCO. 




THE SEWER SYSTEM OF SAN FRANCISCO 371 

city that the distance thence to Sixteenth Street and along Sixteenth 
Street to the end of Center Street wharf would be twice as great 
as the distance from the same point along Division Street to the head 
of the open water in Channel Street. Consequently, a conduit along 
Sixteenth Street would have less gradient than a storm-water outlet 
along Division Street. It would have to be of larger cross-section, and 
the velocities in it would be less than if located along Division Street. 
In other words, for conduits of similar dimensions, on the two routes, 
there would be more probability of surface inundation at and near 
Fourteenth and Harrison Streets if the conduit were located along 
Sixteenth Street than if located as planned. 

The foot of Sixteenth Street does not meet the requirements of a 
permanent outfall for the sewage of the area that can be made tributary 
to it. Nothing that can be delivered at North Point should be allowed 
to reach the water at such a central point on the eastern bay frontage. 

Attention may be called to the fact that, in line with this require- 
ment, the sewage of the eastern portion of the Potrero Hills is to be 
collected in a secondary interceptor which will skirt the northern base 
of these hills and connect with the North Point main at Ninth and 
Howard Streets. The storm-water overflow from this secondary East 
Potrero interceptor will flow in suitable conduits to the Division 
Street relief outlet. 

The North Point main will be in tunnel under Telegraph Hill for 
2 500 ft., if it follows Montgomery Street. If located on Sansome 
Street, the tunnel construction, if required at all, would be much 
shorter. 

In its largest section, the sewer was planned to be circular, 8 ft. 
in diameter, with a concrete invert and a brick arch. In doubtful 
ground it is to be supported on piles. Double sheet-piling on each 
side is to facilitate trenching. The piles used as guide piles for the 
sheeting are to be left in place as a part of the sewer foundation, and 
the sheet-piling on either side of the trench is to serve as a part of 
the form for the concrete invert. Longitudinal steel rods are to be 
placed in the concrete, and as many piles as may seem necessary for 
support are to be driven in the trench and capped with the concrete 
of the sewer invert. 

The hydraulic gradient prescribed for the sewer when flowing 
full will have an elevation of 7.00 ft. above City Base at Eighteenth 



372 THE SEWER SYSTEM OP SAN FRANCISCO 

and Division Streets, dropping to 3.50 ft. below City Base at the 
outfall. Its gradient is about 0.0005, or 1 in 2 000. The sewer invert, 
as originally planned, was to have gradients gradually changing from 
1 in 1 800 near Division and Eighteenth Streets to 1 in 2 000 at the 
outfall. 

The hydraulic grade line at the North Point screen-house will be 
0.50 ft. above extreme high water of the bay and 3.20 ft. above 
ordinary high water. The sewer invert at the same point will be about 
12.90 ft. below City Base, or a little more than 6 ft. below ordinary 
high tide. 

During the rising tide, therefore, the discharge of sewage will be 
retarded; during the falling tide it will be accelerated. The effect 
of this will be a reduced outflow into the flood tides and an increased 
discharge into the ebb, which is a desirable feature. 

The area of the city on the bay front which lies to the eastward of 
the North Point main will be treated in three subdivisions, from each 
of which sewage will be pumped into the North Point main. 

It is not proposed in this paper to enter upon a minute description 
of the system of sewers outlined for these and for the many other 
main and sub-districts, because the problems are in large part the same 
as are presented in all growing cities. Only those features of the 
problem will be referred to which appear somewhat unusual or may 
otherwise be considered of special interest. 

Yerha Buena and Mission Flats Sub-District. — The Yerba Buena 
sub-district, which extends from Telegraph Hill on the north to 
Rincon Hill on the south, and lies between the North Point main 
and the bay, is a low, flat, for the most part filled-in, section of the 
city. It is throughout an important business district. 

The storm-waters of this sub-district are to be delivered to the bay 
for the most part in the sewers now in use, but remodeled as far as 
practicable to give them continuous slope toward their outfall points. 

Sewage proper will be provided for in separate conduits leading to 
a pumping station between Commercial and Sacramento Streets, a 
short distance east of Drumm Street. The pumps will send the 
sewage westward into the North Point main. 

The Mission Flats sub-district, 845 acres in extent, which embraces 
the old Mission Bay region as far as it lies below or east of the 
North Point main, will also be sewered on the separate system. The 



THE SEWER SYSTEM OP SAN FRANCISCO 373 

pumping station for this district is to be located north of Channel 
Street, near Fifth Street. The sewage south of the Channel will be 
carried under the open waterway in an inverted siphon. The pumps 
will send the sewage northwestward along Fifth Street into the 
North Point main. 

In both these districts the sewers of the separate system will be 
ironstone pipes. The smallest lateral sewers are to be 8 in. in 
diameter, and the minimum gradients on which these are to be laid, 
with but few exceptions, will be 1 in 200. 

The principal objection to the use of the combined system in these 
districts is the fact that the surface of the districts lies low. Very 
much of it is at City Base. The efficient drainage of basements has 
not been possible with the combined system heretofore in use. Better 
service was demanded. It was necessary to choose between, on the 
one hand, a combined system of sewers leading to pumping stations, 
with storm-water relief outlets to the bay and intermittent pumping, 
and, on the other hand, the complete separation of the sewers from the 
storm-water conduits. In the former case there would have been times, 
during rain storms, when the sewers would have been full to a plane 
well above high tide in the bay, and back-flow into basements unguarded 
by check valves would have been inevitable. From this situation there 
was no remedy except either to provide pumping capacity on a 
combined system great enough to handle the maximum storm-water 
flow, or to introduce the separate system. The latter plan was adopted, 
but it, too, has its disadvantages. One is the duplication of conduits 
in streets already crowded with pipes for gas and water, and with 
conduits for telephone and telegraph wires and wires for transmitting 
electricity. It was found necessary in many places to provide sewers 
for each side of the street. 

The main drawback to this system is probably to be found in the 
fact that some of the streets of the filled-in areas are still in a condition 
of subsidence. On this point there was very little information avail- 
able in 1899. It was known that the lower end of Market Street, 
which had been paved many years before at official grade, was then 
several feet low, and the various buildings with their sidewalks, which 
had been constructed at grade, had settled measurably. Nothing was 
known concerning the continuance of the subsidence. The general 
opinion seemed to prevail that but little more was to be expected. 



374 



THE SEWEE SYSTEM OF SAN FRANCISCO 



The adopted system, under these circumstances, included provision for 
the support of the sewers on piles. It was proposed to place the 
sewers in a continuous, pile-supported bed of reinforced concrete 
wherever the ground was doubtful. 

In the light of later information, it now seems uncertain whether 
any type of construction can be devised that will hold the sewers at 
grade. 

As City Engineer, during 1900 to 1904, the writer made studies of 
street subsidence. Careful levels were taken at intervals of one year. 
The results of these studies for the two years, 1901 to 1903, are shown 
in brief form in Table 18. 

TABLE 18. — Subsidence of Streets in San Francisco. 





Subsidence, in Feet. 




1901-1902. 


1903-1903. 


Davis Street, Market to East Streets 


0.05 
0.03 
0.05 
0.05 
0.15 

I- '■'' 1 




Jackson Street, Montgomery to East Streets. .... 




Spear Street, i\tarket to Bryan Streets 




Mission Street, First to East Streets 


0.02 


Harrison Street, Fourtli to Seventh Streets 


0.19 




0.05 


Sixth btreet, Brannan to Channel Streets 


0.08 







Erom April, 1902, to April, 1903, there was no appreciable sub- 
sidence on Davis Street except a small amount at its intersection 
with Vallejo Street, where the subsidence during the preceding 12 
months had been 0.08 ft. 

Erom April, 1902, to April, 1903, there was no appreciable sub- 
sidence on Jackson Street except a small amount at Drumm and 
East Streets, where the subsidence during the preceding year had 
been 0.06 ft. 

Erom April, 1902, to April, 1903, there was subsidence on Spear 
Street only at Mission Street, where the subsidence during the pre- 
ceding 12 months had been 0.06 ft. 

The subsidence is not confined to the streets alone, but includes 
adjacent buildings; even those within the area under consideration 
that are supported on piles 90 ft. and more in length show measurable 
subsidence. 

The lower end of Market Street in 1903 was from 2 to 3 ft. below 



THE SEWER SYSTEM OF SAN FRANCISCO 375 

official grade from curb to curb, the street-car tracks supported on piles 
having settled with the remainder of the street. 

As subsidence of the character here referred to includes every 
structure in or adjacent to the street, no provision for the support 
of sewers would be adequate to prevent settling. If the sewers are 
laid to grade there will be sections dipping below grade after a few 
years, and the dip will increase until after a time the departure from 
the true gradient may be so great that reconstruction of some of the 
down-town sections will be necessary. 

The earthquake of 1906 has occurred since the subsidence studies 
were commenced. No information is at hand at this writing as to 
the effect which the violent shaking has had upon the general elevation 
of the street surfaces in the made-land districts. Local upheavals of 
a foot or more, and local depressions of the ground's surface have 
been noted in various parts of the city. Where these occurred they 
were more or less destructive to all conduits in the streets, and would 
have caused serious derangement of the sewer system as planned if it 
had been in service. It is probable that whatever changes have been 
produced in the street levels, by the earthquake, have brought them 
nearer to a permanent condition of stability. At any rate, the proba- 
bility of the recurrence of earthquakes of sufficient force to prove 
destructive to sewers is too remote to be taken seriously into account 
in planning the system. The location of the areas where a derangement 
is most likely to occur, however, has been pretty clearly indicated by the 
breaks in conduits of all kinds in the earthquake of 1906, and this 
knowledge may here and there lead to some modification of the loca- 
tion of a sewer or to some modification of design. 

Lower Islais Creek District. — A dual treatment has been planned 
for the Lower Islais Creek district. From this district sewage proper 
is to be pumped through a long force main into the North Point 
sewer. Where there are no sewers now in use the collecting system 
will be separate from the storm-water drainage; but where sewers have 
already been constructed to serve for sewage and storm-water com- 
bined these are to remain in service. The arrangement at the pump- 
ing station, which will be centrally located, will be such that excess 
storm-water will pass through a relief outlet to Islais Creek. When- 
ever the excess flow is long continued, which will be coincident with 
a full North Point sewer, the pumping from the sewers on the combined 



376 THE SEWER SYSTEM OF SAN ERANCISCO 

system will be stopped. The sewers of the separate system are to reach 
this pumping station at 15 ft. below City Base. This low grade is 
made necessary by the large tributary area with low-surface elevation. 
At a mean stage of water in the receiver at the pump the total lift, 
including friction in the discharge main, will be about 50 ft. 

Only those sewers of this, as of other, districts that lie within 
improved territory or that are at once required to reach pumping 
stations or outfall points were recommended for immediate construc- 
tion. There are some cases in which the sewers recommended for 
immediate construction lie along streets crossing the marsh and not 
now graded. In these cases it was assumed that such streets would be 
ordered graded as a matter of urgency, and that within a few years 
they would be ready to receive the sewers. 

The North Point Outfall. — The screen-house at the outfall of the 
North Point main is to be a neat structure of brick or concrete. There 
will be two settling basins of small diameter, to intercept coarse sand 
and the broken rock reaching the sewers occasionally with the storm- 
water from the macadamized streets of the outlying regions. These 
basins are to be arranged so that one or both may be in service. The 
removal of material from them will be by dredging, possibly by the 
hydraulic method. 

Alongside of the main sewer, just before it enters the screen- 
house, there is to be a long overfall weir, with crest at the hydraulic 
grade line, or a little below this, over which any excess flow of storm- 
water will drop into a low, flat conduit of brick or concrete leading 
directly to the bay. 

The screen-house will be equipped with screens for the interception 
of rags, cork, and other floating material that should be removed from 
the sewage before it is discharged into the bay waters. The outfall 
pipes are to be connected with the settling chambers, and are to be 
carried at least 400 ft. beyond the head of a solid pier projected for 
this point. They were planned to be of cast iron, 5 ft. in diameter. 
It may be found desirable to substitute reinforced concrete. They are 
to be placed in a trench excavated by dredging to about 50 ft. below 

low tide. 

Hunter's Point Temporary Outfall. 

In the case of Hunter's Point main sewer, it is not practicable, 
owing to the unimproved condition of the streets and their location, 



THE SEWER SYSTEM OF SAN FRANCISCO 377 

across marsh surface and points of hills, to carry the sewer at once to 
its final point of outfall. The improvement of the streets along which 
the sewer is to be built is too far in the future. Their grading for 
sewers is out of the question on account of cost. The main, therefore, 
will terminate for a time on the north shore of Hunter's Point, and a 
discharge will be effected through a submerged pipe well off shore. 

Harbor View Sewer District. 

The time may come when the demand for keeping the waters of 
the bay absolutely free from sewage will be so great that, whether 
offensive to the senses or not, the delivery of sewage on the eastern 
and northern bay frontage of the city must be stopped. Should this 
time ever come, the sewers as now planned will remain in service, but 
pumping will be necessary to send all sewage along the water front to 
an outfall at Fort Point. The arrangement can there be made to 
discharge practically all sewage of the bay slope of the city into the 
ebb tide. In this event the sewage from so remote a region even as 
Bay View would be delivered by pumping into the North Point mam, 
and would be rehandled by the main pumps located at or near the site 
of the proposed screen-house. Until this time comes, there will prob- 
ably always be more or less sewage sent into the bay at points less 
desirable than those selected as main outfall points. 

Until actual experience shows it to be undesirable to send limited 
amounts of sewage into the bay with delivery into deep water, at 
a few selected points, the great expense of concentrating all at Fort 
Point need not be incurred. 

It was planned, therefore, to deliver the sewage of the Harbor View 
District, lying on the northern front of the city to the westward of 
Black Point, into the bay at the foot of Scott Street. The sewers of 
this district, as at present, will carry both sewage proper and storm- 
water. Storm-water relief outlets are planned with a view to bringing 
existing sewers into use at their full capacity. The drainage system of 
the district was designed with special reference to the present improve- 
ments. The intercepting sewers follow streets along which immediate 
construction is practicable. When the reclamation of the waterfront 
property has here been completed, a second line of intercepting sewers 
of small capacity will be needed. The storm-water relief outlets will 
serve as local sewers below the upper intercepting line until they 



378 THE SEWER SYSTEM OF SAN FRANCISCO 

reach the waterfront, where their ordinary flow will follow the lower 
interceptor and be carried by the same to the main outfall point. 

This system of sewerage will have the disadvantage, for the low area 
close by the waterfront (as yet but partially improved), of affording 
only imperfect basement and cellar drainage. But the property owners 
will know the limitations, and can plan their improvements to meet 
the conditions. 

The discharge into the bay will be through a large pipe, about 
800 ft. long, terminating in water 40 ft. deep. The capacity of this 
pipe will be adequate for the ordinary sewage flow, with the usual 
included small amount of rain-water. All excess flow, as already 
stated, will go to the bay through storm-water relief outlets. 

EiCHMOND Sewer District. 

For the Richmond District a sewer project was adopted some 
years previous to the studies of 1899. The sewage and rain-water are 
carried in common conduits to an outfall just west of Baker's Beach. 
At the time this outfall was inspected no trace of sewage at the shore 
nor of offensive odor could be detected. The strong tidal currents 
sweeping this part of the shore warrant its use as a point of disposal. 
The later designs provide for an extension of the outfall farther off 
shore into deeper water and a completion of such mains as have not 
yet been constructed. 

Other Ocean-Slope Sewer Districts. 

The topography of the ocean slope of the city westerly and south- 
erly from Richmond District made a subdivision into four additional 
districts appear desirable. These are the West Richmond, Upper 
Sunset, Lower Sunset, and Ocean View Districts. The last named 
includes the Lake Merced region, and extends southward to the 
County line. 

The problem of handling the sewage of these districts, which for 
the most part will in time be densely populated, is complicated by 
the fact that the ocean beach, extending for several miles southward 
from the Cliff House near Point Lobos, is a pleasure ground of the 
people, and must be preserved undefiled. 

This consideration led the writer and his associate Migineers to 
recommend the separate system of sewerage for all these districts, 



THE SEWER SYSTEM OF SAN FRANCISCO 379 

and a delivery of all sewage proper into the ocean, off Point Lobos, 
where it will be so diluted as to disappear completely. From this 
point, moreover, the oceanward flow is away from and not toward the 
beach. 

Storm-waters were to be collected in separate conduits with out- 
falls into the ocean at selected points. The principal outfall points 
tentatively suggested for storm-waters were at the foot of J Street 
and at the foot of X Street. 

It is understood that the plan of treatment here outlined for the 
Ocean Beach drainages is not to be adhered to. It is now proposed to 
construct at once an interceptor for both sewage and storm-water which 
will discharge into the ocean at the Point Lobos outfall point already 
referred to. Sewage will be liberated in deep water well off shore in 
the vicinity of Mile Rock. 

The rapid improvement that is taking place on the ocean slope has 
made it appear undesirable to install the works for the temporary 
delivery of sewage from a West Richmond pumping station into the 
Richmond mains. 

It will be possible to place the proposed main sewer of the ocean- 
slope districts along V Street to Forty-fifth Avenue, to T Street, to 
Forty-sixth Avenue, to R Street, to Forty-seventh Avenue, to J Street, 
to Forty-eighth Avenue, and along Forty-eighth Avenue into the tunnel 
under Fort Miley Heights. 

A low-level sewer will follow the Great Highway northward from 
X Street to a junction with the main interceptor. Velocities in the 
low-level sewer are estimated at 2.8 ft. per sec. 

All the ocean-slope districts, except portions of Richmond, were 
but sparsely settled in 1899, but a rate of growth was anticipated 
that made it imperative to provide means at an early date for a 
disposal of the sewage from the improved areas. At that time there 
were improved areas of small extent at and near the beach, both to 
the north and to the south of the Park, and Upper Sunset was fast 
becoming popular as a residence district. At Ocean View, too, in 
the southwestern part of the city, a portion of the built-up area was 
over the divide on the ocean slope and needed attention. 

The ultimate point of outfall taken into consideration for this 
ocean-slope section of the city was Point Lobos, near Mile Rock. The 
small service required for a number of years, and the high initial 



380 THE SEWEK SYSTEM OF SAN FRANCISCO 

cost of tunneling under Point Lobos Heights, made it undesirable to 
attempt to reach this point with an outfall sewer at the outset. For 
this reason the following treatment of the ocean-slope region was 
recommended. This has since been modified, as already explained. 

Upper Sunset District. — The region lying south of Golden Gate 
Park, between the Park and the Twin Peaks and Blue Mountain group 
of hills, as far west as Twenty-ninth Avenue, has been designated 
Upper Sunset District to distinguish it from that part of Sunset 
which lies at lower elevation and nearer the ocean. 

The sewage of this district, under the plan of 1899, was to be 
carried northward across Golden Gate Park and delivered into the 
mains of the Richmond District, with outfall at the foot of Twenty- 
seventh Avenue, just west of Baker's Beach. The storm-waters were 
to be carried westward in separate conduits. 

At the present time both sewage and storm-water are delivered into 
the Park, as a temporary arrangement, and are being used for 
irrigation. 

Lower Sunset District. — The main western slope of the city, from 
Golden Gate Park southward, to near Lake Merced, is known as Lower 
Sunset District. In 1899 the greater portion of this region was 
covered by high, irregular sand dunes. Streets had long ago been laid 
out, but official grades remained to be established. At that time, 
therefore, it was premature to commit the city to a definite scheme 
of sewer alignment. Since then, however, this region has been one 
of rapid development, and no mistake was made in 1899 in recom- 
mending the immediate construction of works for sewage disposal for 
which there is growing need. In this district, as for Upper Sunset 
District, a complete separation of sewage from storm-water was 
recommended. This plan has now been modified, as already stated. 

West Richmond District. — The region known as West Richmond 
District is practically a part of the Lower Sunset District, separated 
from the latter by Golden Gate Park. The sewage of both this and 
Lower Sunset District was to have been led into the receivers of a 
pumping station located near the ocean on the north line of the Park. 
It was to be pumped from there temporarily into the mains of the 
Richmond District, ultimately through a tunnel under Point Lobos 
Heights into the ocean near Mile Rock. Under the modified plan, the 
pumping station is rendered unnecessary. 



THE SEWER SYSTEM OF SAN FRANCISCO 381 

Ocean View District.— Ocean View District is the area sloping 
westward from the peninsular ridge, in the southwestern part of the 
city, toward Lake Merced. It embraces parts of Ocean View, Lake 
View and Ingleside. It is a sparsely populated district, but needs out- 
fall sewers at once because its drainage is toward Lake Merced, which 
is still in use as a source of water for the city and will probably be 
used indefinitely as an emergency source. 

The storm-waters of this district will necessarily follow natural 
drainage lines westward to points where they can be intercepted for 
diversion past Lake Merced. Some temporary works for such diversion, 
with permanent tunnel outfall, have already been constructed by the 
Spring Valley Water Company. 

Sewage proper will be collected in separate conduits, and vTill be 
carried in a closed main to a discharge into the main sewers of the 
Lower Sunset District. 

Pumping Stations. 

The details for the pumping stations were worked out by Mr. H. C. 
Behr, Mechanical Engineer. Centrifugal pumps were prescribed for all 
pumping stations, although, for use at the West Richmond station, 
vertical, crank-driven, triple-plunger pumps were taken into account, 
to be used until the low delivery through the tunnel at Point Lobos 
could be effected. Under the modified plans there will be no West 
Richmond pumping station. 

The receiver capacities were fixed, in a measure, to fit the avail- 
able ground area. If necessary, they may be added to in the future. 
The receivers were planned circular in outline, in triplicate for each 
station, and it was assumed that all would have to be carried below 
the water-table, and that the pneumatic method of construction would 
be necessary. The receivers at each pumping station are to be inter- 
connected and arranged for use separately or in conjunction. The 
sewage, before entering the receivers, is to be screened. Ventilation is 
to be provided by a shaft or chimney, in which a small fan blower, 
with a multiple injector nozzle, is to create a draft. 

All pumps are to be belt-driven. Small units are suggested for the 
pumps in order to secure the greatest possible adaptability to all re- 
quirements. The motors were only tentatively selected for cost-esti- 
mate purposes, it being apparent that gas, oil, and electric motors are 



382 



THE SEWER SYSTEM OF SAN FRANCISCO 



all well suited for the work required. Steam engines require more 
space and a greater cost of attendance. They are objectionable on 
account of the smoke nuisance, and are not as well suited to the inter- 
mittent character of the work that may be required. 

The force or discharge main from each pumping station is to be 
of cast iron. Additional facts relating to the pumping stations are 
given in Table 19. 

TABLE 19. — Sewage Pumping Stations. 



Name of station. 




Total lift, 

including 

friction, 

in feet. 


« o * 


pi 

fe-^ p. • 


Capacity 

of 
receiver, 
in gallons. 


.3 = 0) 

^5 


CM ^ 

^ C o 

|o.S 


Yerba Buena 

Mission Flats 

Lower Islais Creek. 
Bay View 


6 
15 
12 

5 


31 
35 
65 
30 


42 
120 
180 

35 


35 
80 
88 
14 


60 000 
60 000 
60 000 
40 000 


2100 
4 000 
4 700 
1 650 


16 

18 
18 
8 







Types of Sewers. 

Concerning the types of sewers to be used, but little need be said. 
Ironstone pipes are to come into use up to diameters of 2 ft. Sewers 
of the next larger size, and up to 3 ft. 6 in. by 5 ft. 3 in., are to be 
egg-shaped. Larger sewers are to be circular, except as stated in the 
case of the main relief outlet to Channel Street, which is to be flat-^ 
the lower end in two or possibly three compartments, as already ex- 
plained, each nearly rectangular, 9 ft. wide and SJ ft. high. 

The egg-shaped sewers and the large circular sewers were planned 
to have concrete inverts with reinforcement to give added strength and 
stability. They were designed without vitrified brick invert lining. 

It is understood that the plans have recently been modified some- 
what in the matter of sewer types. The use of reinforced concrete is 
to be extended. All sewers of greater diameter than 2 ft. are to be 
constructed of this material. Vitrified brick is to be used for invert 
lining of some of the sewers. 

It is to be stated in this connection that recent examinations by 
the City Engineer of concrete sewers in Sixth Street and H Street 
that were constructed about five years ago show that these sewers, 
except for a slight wear in the invert of the H Street sewer, are in 
excellent condition. No deleterious effect of sewage upon the concrete 
could be noted. 



THE SEWER SYSTEM OF SAN FRANCISCO 383 

Whether the tendency, so generally apparent throughout the 
country, to build the sewers with concrete shells of minimum thickness, 
which is sure to be attended with many failures, is to be followed or 
not, remains to be seen. 

The Bond Issue. 

The report of 1899 was accompanied by a cost estimate covering 
the construction of the sewers and storm-water conduits immediately 
required. The expenditure of $4 600 000 was recommended. 

At an election, held soon after the adoption of the report, bonds in 
this amount were voted to carry forward the work. Before these bonds 
were issued there was a modification of the form of government in San 
Francisco. Theretofore the city was subject to control by the State 
Legislature. Legislative enactment, in the early history of San Fran- 
cisco, had consolidated City and County. Amendments to the Consoli- 
dation Act were thereafter the convenient means of securing privileges 
in the city of laying out new streets, and of prescribing and limiting 
the powers of the Board of Supervisors and the other City and County 
officials. 

The charter which was adopted in November, 1899, modified all 
this, and, among other things, created a Board of Public Works, with 
usual duties; but this charter, which went into effect on January 8th, 
1900, made no provision for carrying to completion the proceedings 
relating to bond issues commenced under the Consolidation Act. The 
matter was carried into the courts, whose decision was to the effect 
that the sewer bonds already voted could not be issued. New pro- 
ceedings were necessary. 

During the first few years under the new charter such matters as the 
bonding of the city for improvements moved slowly. Each step had 
to be subjected to the test of the courts. It was not until 1903, there- 
fore, that a new order was passed by the Supervisors directing the City 
Engineer again to prepare plans for a comprehensive system of sewers. 

In accordance with this requirement, the plans of 1899, with but 
slight modification, were re-submitted by the writer, then City Engineer, 
under date of June 30th, 1903. The cost estimate was increased 
because it covered sewers not recommended for immediate construction 
4 years before, and because there had been considerable advance in the 
price of materials, and particularly in wages. The estimated cost of 



384 THE SEWER SYSTEM OF SAN ERANCISCO 

the sewers and accessory structures, then recommended for construc- 
tion, was $7 250 000. 

The question relating to the issuance of sewer bonds in this amount, 
together with bonds for other purposes, in the aggregate about 
$17 000 000, was submitted to the voters in November, 1903. The 
bonds for sewers were again voted. But the bonds as authorized were 
to bear only 3^% interest. This low rate of interest coupled with the 
fact that the government of the city had meanwhile passed into the 
hands of officials who did not command public confidence, made the 
sale of these bonds impossible. At length further attempts to sell 
bonds of this issue were abandoned, and the City Engineer was called 
upon for a new report. This was after the great calamity of 1906. 
Restriction of the sewer work to the main lines of sewers was de- 
manded, in order to keep the bond issue as low as possible. The 
latest recommendation, therefore, does not cover all the sub-mains 
that were included in the recommendations of 1903, nor even the full 
system of 1899. 

The final submission, in May, 1908, to the voters, of the proposition 
to expend $4 000 000 on the sewer system again resulted favorably by a 
vote of about 15 to 1. Under the bond issue thus authorized the 
systematic construction of the projected main sewers and inter- 
ceptors has now been commenced. 



DISCUSSION : THE SEWER SYSTEM OE SAN FRANCISCO 385 

D ISCU S SI ON. 



W. C. Hammatt, Assoc. M. Am. Soc. C. E. (by letter). — In the Mr. Hammatt. 
writer's opinion, there are few cases when it is advisable to handle 
storm-water and sewage together; and it is particularly objectionable 
when the discharge is below water. For a given district, variations 
of the quantity of sewage proper are slight, while the storm-water will 
vary from nothing for a great part of the year to a quantity many 
times in excess of the sewage on some occasions. 

In a main emptying below water level, therefore, and designed to 
carry both storm-water and sewage, the velocity would be so slight 
when sewage alone was being carried, that sedimentation would take 
place. Cases are conceivable where the sedimentation would be so 
great during prolonged dry seasons that the mass would be too great 
for the first storm to remove, and serious damage might be done to 
the outfall piping and the lower sewer districts. 

It might be claimed that the relief outlets would take care of all 
the storm-water if the main pipe were obstructed in this way, but if 
such a thing should happen, and the storm were of a low intensity, 
the sewage which would take the relief outlet would be diluted so 
slightly as to be offensive, and as the relief sewers are to empty in 
exposed places, this would be objectionable. 

The writer believes that sewage proper and storm-water should be 
handled in separate conduits and under separate systems. The sewage 
should be discharged below low-water level in such a manner as to 
produce proper dilution and removal by currents, as described by Mr. 
Grunsky. The storm-water should be discharged above tide level, 
when possible, at the most convenient beach or pier head. The out- 
fall for sewage proper should be of such cross-section that the velocity 
against the back-water would never fall below 0.5 ft. per sec. 

Robert G. Dieck, Assoc. M. Am. Soc. C. E. (by letter). — In con- Mr. Dieck. 
nection with the design of an improved sewerage system for San 
Francisco, it may be of some interest to state the method of design 
followed in the storm-water drainage system of Manila, P. T. To 
understand clearly the essential geographical features of the city, a 
few remarks are necessary. The city is situated at the mouth of the 
Pasig River, on the east shore of Manila Bay, on an alluvial plain 
formed of sand and gravel, mingled with loam, brought down from 
the range of mountains which forms the main drainage slopes of the 
Island of Luzon. In addition to the major stream, the river has two 
main branches which take off within the limits of the city, and serve 
as additional reliefs in times of high tide and floods in the river. One 
branch extends north and the other south, and both are of about equal 
importance. There are several smaller streams, which intersect these 



386 



DISCUSSION : THE SEWEK SYSTEM OF SAN FRANCISCO 



Mr. Dieck. branches and the main stream, forming a network of canals of more 
or less importance, which makes the system of waterways practically 
interdependent. There are also several minor connections to the bay, 
but these are of lesser importance. Except at extreme low water, 
practically all the streams are navigable for small river boats, and 
furnish a ready means of transportation. The larger streams have 
been dredged where shoaling has occurred. 



OUTLINE SKETCH OF 

WATERCOURSES 

MANILA, P.I 




Pig. 13. 

The ground surface rises from about 2 m. above mean low water 
in the southern and western portions of the city to about 5 m. above 
the same datum in the northeastern portion, where small foot-hills are 
encountered. Practically 75% of the area of the city lies within these 
elevations. As the average range of the tide does not exceed about 
1 m., water is encountered in portions of the town at a depth of 1 m. 
below the street surface at high tide. Because of this low elevation, 
the water-table is almost constantly near the surface in large areas of 
the town; and, for the most part, in the business and residence sec- 
tions, trenches more than 1.5 m. deep require to be pumped. As these 
sections required attention at once, it was soon seen that the shallow 
trenches made necessary by the high water-table would influence seri- 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 387 

ously the design of the sewers. During the rainy season, which occurs Mr. Dieck. 
from about June to late in October or even to early December, as a 
consequence of the torrential rains, the water-table rises, and, in some 
parts, almost coincides with the street surface, making dry trenching 
an impossibility without pumping. 

The river has a depth of about 14 ft. at the east end of the harbor 
(obtained by dredging), decreasing to about 6 ft. at extreme low water 
at the eastern limits of the city. The main branches have a maximum 
depth of about 8 ft. Gaugings of these streams and the river have not 
been made, but, in times of flood, when the current has a velocity of 
nearly 10 ft. per sec, the discharge of the river must be at least from 
50 000 to 60 000 sec-ft. For the most part, the esteros (as the branches 
are called) are sluggish and rather shallow, exposing large portions of 
their beds at low water. Their frequency, however, makes them 
admirable recipients of storm-water, and makes further subdivision 
into drainage districts unnecessary. 

Before it was possible to arrive at a satisfactory solution, it was 
evident that the entire question of street grades would have to be 
considered, and that some definite plan for these, in opposition to the 
former, haphazard plan, would have to be adopted. To this end, care- 
ful levels were taken throughout the city, and the data were marked 
on the city plans. Preliminary computations clearly showed that to 
avoid excessive fill, which, under local conditions, would have proven 
extremely costly, and to protect property interests against serious 
damage following violent changes in grade, a flat gutter grade was 
absolutely essential. After long discussion, this grade was fixed at the 
absolute minimum of 0.002. It had been observed that water would 
stand, on the most carefully constructed work laid to a grade less than 
this, and as all stagnant water is objectionable in the tropics, it was 
held that this feature should govern the work, even if in street con- 
struction the item of filling were large. As a further limiting factor, 
it was fixed arbitrarily that the street surface at no point should be 
less than 0.75 m. above the highest tide. The limits set varied for 
the various districts. Although it was known that all the low dis- 
tricts would be subject to overflow during high floods in the river, it 
was still held that some reasonable height of street surface should 
be fixed in all cases, in order to improve, wherever possible, a naturally 
bad condition. 

Another limitation was in reference to cover. After some observa- 
tion of 24-in. concrete pipes, under heaviest city traffic, the minimum 
cover was fixed at 0.60 m. Up to the present this cover has been suffi- 
cient. For reinforced sewers no limit was set except an amount suffi- 
cient to clear the paving. Under the worst conditions, therefore, with 
the further provision that the sewers should not be laid below mean 
low water, because of the likelihood of silting, the sewers had to be 



388 



DISCUSSION : THE SEWER SYSTEM OF SAN FEANCISOO 



Mr. Dieck. designed to fall entirely within from 1.15 to 1.25 m. It was essential, 
also, that the space occupied by the sewers should be kept at the lowest 
limit in order to provide for other structures in the narrow streets, 
these, in some cases, being only 8 m. between building lines, with a 
clear roadway of only 6 m. At every turn, therefore, certain definite 
restrictions were to be observed, and also the strictest economy in space 
and head-room. 

A study of the rainfall records was then made. Fortunately, the 
admirable records of the Philippine Weather Bureau were available, 
and were examined for maximum intensities for a period of at least 
16 years. The records clearly showed that intensities of more than 
3.00 in. per hour were not uncommon. Table 20 is a list of typical 
storms which were considered as probably the worst to be encountered. 

TABLE 20.— Typical Storms in the Philippine Islands. 



Date. 



Apr. 29, 1!)0.5 
May '21, 1893 
June 8, Ism 
" 15, 1891 
Sept. 18, 1887 
Aug. 26, 1>103 
July 19, 1899 
Sept. 15, 1891 
July 16, 1890 
Nov. 16,1891 
June 1, 1902 
Aug. 6, 1889 
June 30, 1H89 
Sept. 14, 1898 
Aug. 27, 18w8 
Sept. 17, 1888 
July 13, 1904 
Aug. 28, 1897 
Mar. 9, 1894 
July 12, 1901 
Sept. 6, 1896 
July 23, 1888 
Aug. 26, 1S86 
May 29, 1891 
Sept. 20, 1904 



Total 
rainfall, 
in inches. 



2.56 
2.36 
2.26 
2.16 
2.06 
2.05 
2.02 
1.98 
1.98 
1.96 
1.91 
1.89 
1.86 
1.86 
1.85 
1.85 
1.81 
1.81 
1.80 
1.75 
1.73 
1.68 
1.65 
1.65 
1.62 



Duration, 

in 
minutes. 



7 
20 
17 

35 
10 
16 

10 
15 

15 
10 
10 



Excessive 
rainfall, 
in inches. 



Hourly 
intensity, 
in inches 
per hour. 



0.79 

2.36 

0.98 

0.59 

1.57 

1.62 
Record imperfect. 

0.785 
Record imperfect. 
Record imperfect. 
Record imperfect. 

0.59 I 

1.18 

0.985 I 
Record imperfect. 

1.79 I 

0.73 

1.57 I 

Record imperfect. 

0.79 I 

0.985 I 

Record imperfect. 

0.59 

0.787 

0.71 



3.50 
4.72 
5.90 
4.43 
3.05 
3.24 

3.94 



5.51 
3.54 
3.48 

3.07 
4.37 
5.90 

4.72 
3.15 

2.36 
4.72 
4.25 



Storms having an intensity of 2 in. were common throughout the 
rainy season, and occurred when the ground was thoroughly saturated, 
as well as during the drier times. It was evident, therefore, that no 
relief could be expected from ground storage, but, as the runs in the 
sewers were short, it was decided that the maximum flow could not 
occur until about 10 min. after a severe rain had commenced. Table 
20 shows that ordinarily the very heavy rainfalls did not last more 
than from 20 to 25 min., so that if the maximum discharge occurs as 
stated, and provision is made for lesser rainfalls, there can be flooding 
only for 10 or 15 min. As there are no cellars or basements in the 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 389 

city, provision against back-water in the sewers was not required, and Mr. Dieck. 
it was considered that with safety and at the expense of only a 
temporary inconvenience, flooding of the roadways would be permis- 
sible during unusual storms. Under such cases the sewers would run 
under a head, which would increase their capacity. 

It was then decided to design the sewers under the assumption 
of a 2-in. rainfall, discarding the excessive storms from consideration, 
relying on the sewers under pressure to discharge the excess, and per- 
mitting the flooding of the roadways for 15 niin. In the design no 
account was to be taken of the increased capacity of the sewers due 
to this head, the excess being viewed merely as a factor of safety in 
the design. This compromise was made only after the closest prelimi- 
nary trials and after it was clear that the restricted head-room, flat 
grades, and the necessary economies in expenditures required some 
concessions to be made. It was also kept in mind that during severe 
storms the river and the esteros were subject to overflow and that this 
could not be prevented by any possible provision in the sewers. The 
results obtained under this assumption, while not satisfactory in all 
cases, appear to be justified, for no serious flooding had occurred in 
1907, two years after the first sewer was built under the general plan. 

At first, a run in the gutters of 120 m. had been fixed as the maxi- 
mum in the built-up sections, but later observations on the gutter flow 
showed that this was excessive, and the maximum was reduced to 
100 m. Sewers with inlets at this distance apart give better satisfac- 
tion, and flooding has been lessened. All inlets have direct connections 
to the sewers, and are provided with silt basins and grating covers. 

In the designs, Kutter's formula for discharge, with sewers run- 
ning full, was used, the values of n being taken as follows: 

0.011 for cement plaster finish, 

0.013 for clean brick and glazed pipe, 

0.015 for local stone. 

A velocity of 3 ft. per sec. was sought wherever possible, so as to 
lessen the likelihood of silting, which had always been the greatest 
difficulty in the old sewers built under Spanish direction. All drainage 
areas were kept completely separated, and the connection between the 
old sewers built by the Spaniards was stopped in order to make the 
direction of flow definite. To a great extent, the deposits were un- 
doubtedly due to this error in design, and the immediate betterment 
in the condition of the old sewers following the removal of the con- 
nections was made evident by the increased discharge. Wherever pos- 
sible the old Spanish sewers were utilized, the sides and inverts being 
made smoother by cement plaster. 

In the absence of any data on the character of the run-off, an 
assumption of 75% was at first made; but this was found to be ex- 



390 DISCUSSION : the sewer system of san francisoo 

Mr. Dieck. cessive in large sections and deficient in the closely built-up parts of 
the Walled City and in the business sections. The final run-ofi 
coefficients were fixed as follows: 

Heavily built-up 0.90 

Ordinary business and residential 0.75 

Suburban 0.50 

Undoubtedly, these high coefficients lessened the flooding of the 
streets and are ample for all future work. It is difficult to compare 
the assumptions with the results obtained by Emil Kuichling, M. Am. 
Soc. C. E., or with those by Mr. Grunsky, but it is clear that there 
has been no serious error in the designs. 

The esteros are in need of improvement, but their betterment would 
involve the city in such enormous expenditure that no definite plans 
to this end have yet been evolved. A large amount of dredging is 
necessary, and the streams must be confined to definite channels by 
walls. A preliminary section for a light shallow wall of reinforced 
concrete was designed, but the preliminary estimates of cost were so 
appalling that this could not be recommended. These esteros are the 
natural reliefs of the river, and their improvement is of the highest 
importance in the scheme of storm-water drainage. At times the 
velocity of the current does not exceed 1.0 ft. per sec. and no scour 
of the bed is possible. A deepening of the streams would result in 
increased flow, and their present unsanitary condition would be re- 
moved. It is quite probable that there must be some control of their 
currents, by gates or otherwise, if the reduction of deposits is to be 
secured, but, at present, there seems to be no necessity to consider that 
in the design of the confining walls. 

From these remarks it will be clear that the drainage problem of 
Manila has many nice points to recommend it to the attention of 
hydraulic engineers, and while no claim is made to any originality, it 
is believed that the solution is somewhat out of the ordinary. 

Charles E. Gregory, Assoc. M. Am. Soc. C. E. — This elaborate 
paper presents an unusually able theoretical solution of the problem 
of run-off in storm-water sewers, and is most valuable as an academic 
discussion of the subject. In the speaker's judgment, however, the 
author has not given enough weight to the influence of storage on the 
surface of roofs and streets. After making certain assumptions for a 
hypothetical case, this part of the problem is very ingeniously and cor- 
rectly solved for that case. The assumption is made that the rain 
reaches the sewer from the roofs and streets in maximum volume 
within 5 min., which is similar to the assumption often made by 
supposedly careful investigators and writers. The author, however, 
shows a much better appreciation of what actually occurs than they 
in that he takes account of the storage on the surface for the time 



DISCUSSION : THE SEWEE SYSTEM OF SAN FRANCISCO 391 

assumed. For many roofs and a few street surfaces this assumption Mr. Gregory. 

of 6 min. is no doubt approximately true, but a careful consideration 

of what should occur, theoretically, and a few observations of what 

actually does occur, will indicate a wide variation from the assumed 

5 min., and demonstrate that storage on the surface, particularly for 

short, intense storms, has a much greater influence on the run-off than 

is indicated in the paper. 



0.024 

0.023 

0.022 

0.021 

0.020 

0.019 

0.018 

0.017 

0.016 

^ 0.015 

1 0.014 

•-, 0.013 

1 0.012 
5 0.011 

2 0.010 
ft 009 


















1 








/ 




















1 








/ 


















1 










/ 


























/ 




















1 








/ 


















/ 








/ 


/ 


















/ 








/ 




















1 








/ 


























1 


















JL 


j 


























^h 










1 
















i 


.o^ 








/ 


















^ 








rh 


■ 
















a?/ 






y' 


/ 


1 


















f/ 


.^^ 






J 














0.008 
0.007 
0.006 
0.005 
0.004 
0.003 
0.002 
0.001 






i 


1 c 


y 




< 




















f 


# 






ff 


















^T} 


y 




.{ 


{ 


















*/ 


1/ 


























// 


' 




o^; 


y 




















/ 




■<^ 


f/ 




















/ 


1 




/ 






















/ 






/ 






















L 


^^ 



























012 34 5 07 

Time, in Minutes . 

Fig. 13. 

It is not the speaker's purpose to attempt to solve this very inter- 
esting and intricate problem completely, but simply to present a few 
studies of storage on catchment areas, which have been made for the 
solution of special problems, and to show their relation to the general 
subject. 



392 DISCUSSION : the sewer system of san francisco 

Referring to a paper by the speaker and the discussion thereon by 
A. Marston, M. Am. Soc. C. E.,* it was there brought out that: 

"First. — The water falling on the so-called 'impervious' area of an 
ordinary sewer water-shed does not all run off as fast as it falls, but 
part of it accumulates in increasing quantity on the surface during 
downpours of moderate length, such as cause the maximum discharges 
from sewer districts of ordinary size. 

"Second. — As the storm continues at the same rate, the ratio of run- 
off from the 'impervious' area to the rate of rainfall increases, owing 
to the increased depth and velocity of the surface flow toward the sewer, 
until finally, if the storm lasts long enough, the rate of run-off from 
the 'impervious' areas becomes equal to 100% of the rate of rainfall." 

Rough observations indicate that the depth of water which accu- 
mulates dvTring heavy downpours on an ordinary street, sidewalk, or 
roof, is greater than is generally assumed, and that the mean velocity 
of flow in the gutters is not as great as it is usually computed. 

The accumulation of rain on a street has been considered theoreti- 
cally in the following manner : 

The laws and coefficients of flow are not very well determined 
for very shallow depths, and the impact of the falling rain and other 
factors materially retard this flow. It was assumed, however, that 
Kutter's formula was applicable, and, to allow for these retardations, 
a high value of n was assumed. As an asphalt street, from which the 
greatest flow may be expected, has a low crown, it was assumed, as 
has been observed to be roughly true, that the accumulation of depth 
would be about uniform over its entire surface. From these assump- 
tions, the diagrams on Fig. 13 were deduced. Three curves are there 
shown. The mass curve of depths for the entire rainfall is for a rate 
of 4 in. per hour. Points for the other two curves were obtained from 

S (H„ — H.) 
the expression, t = , by plotting values of H.^ and Q^ in 

2 

terms of depths. 

t = increment of time, in seconds, for surface to rise from 

H^ to H^; 

S = area of unit length of half the street; 

H^ = depth at end of increment of time, t; 

H^ = depth at beginning of increment of time, t; 

I = rate of rain, in cubic feet per second, falling on the area, S; 

^2 = flow, in cubic feet per second, at edge of gutter, due to 

depth, H^; 

Q^ = corresponding flow at beginning of time, t, for depth, H^. 

As an extreme example of storage in streets, the following is 

presented : 

* " Rainfall, and Run-off in Storm- Water Sewers," Transactions, Am. Soc. C. E., Vol. 
LVIII, p. 498. 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 



393 



The run-off from, and the storage in, the gutter of a 24-ft. street, Mr. Gregory. 
1 000 ft. long, with a slope of 0.0025 in 1, is computed in the same 
general manner as for the street surface. The run-off curve is shown 
on Fig. 14. For a gutter, or a surface with a uniform slope, the 
formula is modified to take account of the varying depths, as follows : 
S (H, - H,) - (A, V, T,-A, r, T,) 



t = 



(Q2 + ^1) 



1.3 
1.2 

^ 1.1 
3 1.0 

f 0,9 






















































..4t^ 


















y^ 


















Wy\ 








|0.6 
S 0.5 

|0.3 

(§ 0.2 

0.1 












'€^ 


/ 
















s* 


.v>V 


















.^#^ 
















^f 


yy^ 


^^ 


















^^^' 
















."t5l 
















aoX'^iS 


%^ 


















<^^^ 























100"S' 
90^ 

80 I 

60 I 
50 o 

a) 

40 I 
30 I 
20 I 
10 




10 20 30 40 

Time, in Minutes, from Beginning of Uniform Rain at Rate of 4 In. per Hour. * 

Fig. 14. 

In this case, t^, H^, H^, Q^, and Q^ have the same meaning as 
before, but are applied to flow in the gutter. 

8 = average width of each increment of depth multiplied by 

1000 ft.; 
I =1 000 times the inflow from each unit of length, and is taken 

from the curve of run-off. Fig. 13; 
T^ = total time elapsed from beginning of uniform rain, in 

seconds, or until the depth, H^, is reached ; 
J. 2 = difference in area between the wet cross-section of gutter 
for the depth, H^, and the average wet cross-section for 
the upper reach of the gutter where the depths are 
variable ; 
V^ = average mean velocity in the gutter until the depth, H^, is 

reached ; 
1\, A^, and F^ bear the same relation to //j as T^, A„, and Y ^ bear 
to B.^. 

On Figs. 14, 15, and 16, the rate of the run-off curves is deduced 
by the same general formula for a rain at the rate of 4 in. per hour, 
but the storage and inflow curves are omitted, and a curve showing the 
percentage of rainfall running off is added. 



394 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 



Fig. 15 is for a street 60 ft. wide and 600 ft. long, with an asphalt 
pavement and a slope of 0.005 in 1. 

Fig. 16 is for a roof 25 ft. by 100 ft., sloping at the rate of 0.02 in 
1 toward a gutter across the 25-ft. side. 

There are a great many roofs and streets comparable to those 
assumed for the curves on Figs. 15 and 16. A steeper slope, if paved 
with stone blocks, would give results equivalent, for flow and storage, 
to the flatter smooth surface. The Sixth Avenue Sewer District 
gaugings have been reported by the speaker.* This district contains 
many 600- and 900-ft. blocks, also many flat roofs, 25 ft. by 100 ft., or 
larger. It is estimated that it takes 15 min. for water to flow through 
the sewers from the remote parts of this water-shed to the gauge. In 
a sewer, the rate of run-off at the gauge point in the sewer from such 



1. 
1. 

gi.; 

o 

-1.; 

lo.! 
o 
' cO, 

|a 

go, 

«o, 

0. 

0. 
0, 





, 
























































^ 


















Sf 



















'A' 
















4 


',i(? 1 
















th\ 














ifrv\ 
















f 


xH 


- 












S 
















*7 


j 
















/ 


i/ 


L 
















tl 


/ 


3 














4/ 


/; 
















f^ 


A 


j?^ 












^ 


r^ 


A 


'g 












■o«! 


y 


> 














-=-=^1 















100 « 
90 I 
805 
70 S 
60 I 

30 i 

iO^ 
30 
20 
10 



10 

Time.in Minutes 



Fig. 15. 



20 




01234 rj 6789 10 11 
Timu.iii Minutes 

Fig. 16. 



roofs and streets, as were assumed for Figs. 15 and 16, may be esti- 
mated by assuming that this catchment area is divided into five con- 
centric zones, the outside limits of each being 3 min., 6 min., 9 min., 
12 min., and 15 min., respectively, from the gauge. The area of those 
zones may be taken from the gauge-point outward in the ratio of 1, 2, 3, 
4, and 5. The average rate of run-off, expressed as a percentage of the 
rainfall, for the entire area at the end of 15 min. of uniform rain, at 
the rate of 4 in. per hour, is found by averaging the percentage of run- 
off for the average time length of each zone read from the percentage 
curves on Figs. 15 and 16, respectively. 

The averages for the street and roof areas are thus found to be 
11% and 48%, respectively. If the catchment area was of such shape 

* Transactions, Am. Soc. C. E., Vol. LVIII. 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 395 

that all the zones had equal areas, the percentages would be about Mr. Gregory. 
20 and 63, respectively. In a similar manner, a 5-min. time length 
gives a run-off of about 70% of the rainfall. It is not likely that a 
rate of 4 in. per hour would be maintained for more than 10 min., 
and the run-off from the nearby areas would be greatly reduced for 
the last 5 min. The actual gauged run-off from this area was about 
40% for a rain lasting 15 min. This observed percentage may be 
accounted for by assuming the area to be made up of roofs and streets 
varying from 5 to 20 min. in time length, taking into account the 
storage in the catch-basins and sewers, and making allowance for 
pre-maximum and post-maximum rain rates, the greater influence 
being exerted by the surface storage. For such rates of rainfall as 
occur in San Francisco, the velocities of the surface films, due to less 
depth, would be much less than for a 4-in. rate, and, consequently, the 
time lengths for similar areas would be greater and have a correspond- 
ingly greater influence on the run-off. The speaker wishes particularly 
at this time to call attention to the great influence exerted by the 
shape, size, character, and slope of each area, and by the frequency of 
inlets from the streets. 

E. KuicHLiNG, M. Am. Soc. C. E. — The author deserves the highest Mr. KuichUng. 
credit for having made a more complete analysis of a complicated 
hydraulic problem than has heretofore been published. The novelty 
in his procedure is the estimation and consideration of two factors, the 
first being the quantity of storm-water temporarily in transit on the 
surface toward the various sewer inlets; while the second is the 
quantity temporarily stored in the sewers themselves. Both these 
factors tend to reduce the rate of maximum discharge by making the 
outflow somewhat less than the inflow until the conduits become entirely 
filled. 

The first factor is obviously dependent on the extent and slope of 
the territory tributary to each sewer inlet or catch-basin. If the entire 
area were impervious and divided into a number of equal or equiva- 
lent hopper-shaped parts, each having a catch-basin in its middle 
point, and in which the time required for the water to flow at uniform 
velocity from the boundary line to the central basin is tg minutes, then 
the quantity of storm-water which will still be on the surface at the 
end of such time is shown by the author to be two-thirds of the precipi- 
tation on the total area during that time. 

It should be noted, however, that uniform velocity of flow over the 
surface of a hopper-shaped area involves a constantly reducing slope 
for the successive elementary zones, as the depth of the water is con- 
stantly increasing from vhe periphery to the central catch-basin. Such 
a condition of surface is rarely encountered in municipalities, and 
hence it must be assumed that the velocity increases steadily as the 
water passes successively from the larger exterior elementary zones to 



396 DISCUSSION : the sewee system of san francisco 

Mr. Kuichiing. the smaller interior ones on the same grade. In general, the velocity 
in a channel, or on a surface, is proportional to the square root of the 
depth; and as the depth increases very rapidly on the successively 
reducing elementary zones of like grade, it follows that much less than 
two-thirds of the precipitation on a uniformly graded component area 
will still be on the surface in transit to the catch-basin at the end of 
tg minutes. 

To illustrate this feature of the case by an example, assume a 
hopper-shaped area, 1 000 ft. square, having each of the four equal 
sides graded on a uniform slope to the central catch-basin. Conceive 
the area divided into five zones, each 100 ft. wide. Beginning at the 
basin the areas of these zones will be 40 000, 120 000, 200 000, 280 000 
and 360 000 sq. ft., respectively, or in the proportion of 1, 3, 5, 7, and 9. 
Next assume that the rain falls at uniform intensity for a sufficient 
length of time to allow the run-off from the outermost zone to reach 
the basin. Now, if it were possible for the velocity to be uniform, as 
well as the slope, the water flowing over the innermost zone would be 
about 25 times as deep as on the outermost one. With a uniform 
grade, however, the velocity will increase from zone to zone, and the 
depth in the vicinity of the basin will be materially less than the 
amount just mentioned, since the time of transit is greatly reduced. 
If the velocity in each zone varies with the square root of the depth 
of the flow, the time of transit from boundary to center will be less 
than three times as long as that across the outermost zone, and the 
quantity of water remaining on the surface at the end of such time 
will be considerably less than two-thirds of the precipitation during 
that time. 

In urban districts, the catch-basins or sewer inlets are generally 
much less than 1 000 ft. apart, on the average, and the time required 
for the run-off from the boundaries of the tributary areas to reach the 
inlets is usually not more than 10 min., as the water soon reaches the 
gutters, in which the flow is much more rapid than on a plane surface. 
During this period of time, the precipitation may be 0.5 in. in depth 
over the entire territory, and perhaps one-half the water may remain 
on the surface in transit to the inlets, but, at its termination, the sur- 
face storage comes to an end if the rain continues as before, and the 
run-off will thereafter proceed at its full rate to each inlet, since the 
storage on the surface is not cumulative when the intensity of the 
rainfall is uniform. For large districts, in which the time of passage 
through the longest line of sewers is 30 min. or more, during which 
time a precipitation of 1.5 in. may occur, the retention of a depth of 
0.25 in. on the surface during the first 10 min. of the downpour is not 
of much significance. 

The same may also be said with respect to the storage in the net- 
work of sewers, as the cubical contents thereof do not often exceed 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 397 

an average rate of 1 000 cu. ft. per acre, including house connections. Mr. Kuichiing. 
This capacity is equivalent to a depth of 0.28 in. of rainfall; but, as 
the whole of the precipitation does not reach the sewers, owing to 
various causes of loss, it may be assumed that the capacity is equivalent 
to the run-off from a precipitation of about 0.40 in. In heavy rain- 
falls of comparatively long duration, however, this storage capacity 
will usually become filled some time before the end of the storm, and 
surcharge of the sewers will ensue unless their discharging capacity 
is adequate. 

There can be no doubt as to the propriety of taking these two 
factors of surface storage and conduit storage into consideration for 
small districts and short rainfalls, as the aggregate effect may often be 
very appreciable ; but, in the case of large urban districts, having times 
of transit through the longest line of sewers of more than 30 min., it 
seems to the speaker that the aforesaid effect will be lost in great 
degree. 

The essential factor in the problem is the relation of the intensity 
of the precipitation to its duration. Usually, this is estimated more or 
less arbitrarily from inadequate data, and without an exact knowledge 
of the extent of territory covered by the rainfall at the recorded 
intensity. In most cases, also, the extreme observations are disre- 
garded, and the rainfall curve is adapted to the more frequent down- 
pours. If the sewers are proportioned closely to such a curve, there 
will be no margin of capacity, and it may be questioned whether a 
moderate reserve capacity, such as would be gained by leaving the 
above-mentioned two factors out of account, is not a desirable feature. 
It would certainly be interesting if the author would indicate quantita- 
tively how the full consideration of the said two factors would affect 
the design of a large sewer system in the Eastern States. 

In regard to the rainfall curve that should be used in designing 
sewers, it has long seemed to the speaker that only those durations 
should be taken into account for which the intensity or rate per hour 
was uniform, or nearly so. The object is to ascertain the probable 
maximum run-off from districts of different magnitude, or duration 
of flow through the longest line of sewer. Now, it is obvious that 
the run-off from a homogeneous surface will become largest when the 
rainfall is greatest and is likewise distributed uniformly over the 
entire area for a sufficient length of time to allow the water from the 
outermost boundary to reach the point of observation at the foot of the 
territory; hence it is of primary importance to deduce fi*om the rain- 
fall data the maximum uniform intensities for different periods 
of time. 

The next step is to determine whether this relation is limited to 
areas of any particular magnitude, or whether it is possible to obtain 
a larger rate of run-off at the foot of an extensive area by using 



398 DISCUSSION : the sewer system of san francisco 

Mr. Kuichiing. precipitations of varying intensity, but having an aggregate duration 
equal to the time of flow through the longest line of sewer. In general, 
it will be found that, if the relation between the maximum uniform 
intensity and its duration has been established correctly, the rate of 
run-off at the foot of a homogeneous territory of any size will be 
greatest when the entire area is covered by a rainfall of the highest 
uniform intensity corresponding to the length of time required for 
the water to flow to the point of observation from the outermost 
boundary. Deviations from this rule, however, may possibly occur 
when the area is not homogeneous, and in such cases one must be 
governed by the given or assumed conditions of the particular problem. 
An instance of non-homogeneous territory is afforded by the follow- 
ing example : Let A equal the number of acres at the foot of a district, 
with a run-off factor equal to f^, and the time of flow through the 
longest line of sewer equal to t^ minutes; B, f^ and t^, C, f^ and t^, 
etc., equal the same elements for the next succeeding component areas; 
also, let the maximum uniform intensity of the rainfall, in inches per 

a 
hour (or cubic feet per second per acre), be expressed by 7 = t A- b " 

where t is the duration, in minutes. The rates of run-off, in cubic 
feet per second, from the several component areas, at the end of the 
time r = ^1 + ^2 + ^3 + • • • will ^^^^^ be q^ = f^ A I, q^ = f^B I, 
q^ = f,^ C I, etc., and their sum at the foot of the district will be: 

Q = (q^ + q,+ Q,+ ...)=I (f,A + f,B + f,C+ ...) = 

If the district were entirely homogeneous, we would have fi = f2 = 
/^3 :=...= f^ and hence, also : 

Q^ = ly ( A -f- 5 + (7 + . - .) = y^^ (A-\-B+C-{-.. .)= ^^, 

where N denotes the total area of the district =^ (A -{- B -{- C -\- ...). 

It will be noticed that if the run-off factor, f, represents the true 
mean for the whole district, the values of Q and Q^ will be equal, thus 
indicating that it is immaterial whether the surface is homogeneous 
or not. 

To pursue the subject a little further, conceive the territory to be 

homogeneous and divided into equal squares along the sides of which 

the run-off from each square, or aggregation of squares, moves with 

equal velocity, Vq, in feet per second; also let the two consecutive sides 

of each square have an aggregate length of 60 Vq, corresponding to the 

path traversed by the run-off in 1 min. At the end of T minutes, the 

number of squares contributing to the discharge at the lower corner 

(T-\- 1) 
of the first square will be T ^ ; and, as the area of each square 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 399 

is m = (30 Vg)^ square feet, the total tributary area, in acres, will be Mr. Kuichiing. 
T(T±l)m afm TiT+l) T (T -\- 1) 

We thus have Q as a simple function of T, the form of which shows 
that Q increases directly with T and attains its greatest value only 
when T is greatest, or when the entire area contributes to the discharge. 
This result is due to the peculiar form of the expression for the factor, 
7. With some other form for I it may happen that Q will become a 
maximum for a value of T less than that corresponding to the time 
for flowing through the longest line of sewer in the district. 

From this brief analysis of the run-off problem, it appears that 
attention should be directed mainly to the proper form of expressing 
the maximum uniform intensity of the rainfall, I, in terms of its 
duration, t, in minutes. By the use of this factor, the indefinite ques- 
tion of seeking a maximum, Q, with rainfalls of variable intensity 
during the total time, T = (t^ -\- t^ + t^ + ...)> will be entirely 
avoided, as the former necessarily embraces all combinations of the 
latter. 

The next matter of importance is the investigation of the proper 
value of the run-off factor, f ; but, as the author has not given the 
results of his own experiments, there is no reason for discussing it 
here at considerable length. Few direct and reliable measurements of 
the run-off from a given territory and uniformly distributed rainfall 
appear to be available. In some cases the rainfall is observed at only a 
single point in or near the district, and no proof of uniform dis- 
tribution is offered, while, in other cases, both mean intensity and 
distribution are estimated. It is very desirable, therefore, to increase 
the number of accurate measurements for surfaces of different char- 
acter, and to include therein the actual time of passage through the 
longest line of sewer, as well as the time required for the water flow 
over the surface from the outermost boundary to the nearest catch- 
basin or inlet. 

To exhibit the range in the estimated values of the run-off factor, 
f, or the ratio of the run-off to the rainfall at the time of maximum 
discharge, the following data are submitted : 

For roof surfaces assumed to be water-tight. ..f = 0.70 to 0.95 

" asphalt pavements in good order 0.85 to 0.90 

" stone, brick, and wooden block pavements 

with tightly cemented joints 0.Y5 to 0.85 

" same with open or uncemented joints. .. . 0.50 to 0.70 
" inferior block pavements with uncemented 

joints 0.40 to 0.50 

" macadamized roadways 0.25 to 0.60 

" gravel roadways and walks 0.15 to 0.30 



400 DISCUSSION : the sewer system OE SAN FRANCISCO 

Mr. Kuichling. For viiipaved surfaces, railroad yards and 

vacant lots f = 0.10 to 0.30 

" parks, gardens, lawns, and meadows, de- 
pending on surface slope and character 
of subsoil 0.05 to 0.25 

" wooded areas or forest land, depending on 

surface slope and character of subsoil.. 0.01 to 0.20 

Others do not attempt to make close estimates of the different 
kinds of surface in an urban district, but content themselves with 
general average values of f, as follows: 

For the most densely built-up portion f =^ 0.70 to 0.90 

" the adjoining well built-up portion 0.50 to 0.70 

" the next adjoining residential portions, 

with detached houses and buildings 0.25 to 0.50 

" the suburban portions, with few buildings. 0.10 to 0.25 

Similar variations are found in the estimates of maximum intensity 
of rainfall for which provision is to be made in the sewers or other 
drainage channels, even in neighboring localities where the rainfall 
has the same character and average annual depth. In view of such 
large differences in practice in the choice of f and /, there is little 
encouragement for the adoption of refinements in the general theory 
of the subject. 
Mr. Allen. Kenneth Allen, M. Am. Soc. C. E. — It is a good thing to place 
on record the conditions governing the design, and a description of 
the essential features, of an important improvement of this kind. 

Of the fifteen cities of the United States having in 1906 a popula- 
tion of more than 300 000, all are situated near bodies of water of 
such size that up to the present their sewage is discharged into them 
in a crude condition. Baltimore, however, in order to protect the 
oysters of Chesapeake Bay, is constructing a purification plant, and 
in order to avoid an undue pollution of the water, Chicago, San 
Francisco, Boston, and Washington have spent large sums to secure 
more favorable points of outfall. Without question, in view of 
increasing populations and a growing sensibility to cleanliness and 
sanitation, the next two decades will see extensive improvements in 
most of the other cities in the list. 

Of these fifteen cities, four are on the Great Lakes, five (including 
Chicago) drain to the Mississippi or Ohio Rivers, and six deliver 
their sewage almost directly to the ocean. San Francisco, however, 
is especially favored in being able to locate her outlets — ultimately, at 
least — in strong currents of deep water which trend off shore and 
soon pass out to sea. New York and Boston are also located on great 
tidal bays, but the waters of Upper New York Bay oscillate between 
the Hudson and East Rivers and the Narrows in close proximity to 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 401 

a dense population for several days before being finally dispersed in Mr. Allen, 
the Lower Bay or ocean, while Massachusetts Bay is studded with 
islands, and the shores are lined with towns and resorts, so that great 
care has been required in selecting points of outfall to avoid ofiense. 

With a discharge into large bodies of water there should be suffi- 
cient dilution to prevent putrefactive conditions, and the problem 
resolves itself into one of rapid dispersion. This can be secured with 
a single outlet where the current is swift in comparison with the 
volume of sewage. At San Francisco, as at Boston, interception and 
a discharge at a few well selected points, after the removal of the 
grit and coarser floating matters, has been decided on as the best plan. 
To secure a more rapid dispersion, multiple outlets are provided in the 
Elbe at Hamburg, and have been proposed for the Passaic Valley Sewer 
in New York Harbor; but where cities are located on sluggish streams, 
it is sometimes preferable to maintain numerous independent outlets 
in order to afford opportunity for a more rapid mixing with the water. 
The question of interception to one or a few outlets, as at Buffalo or 
Louisville, or a direct discharge at numerous points, as at New York 
and Philadelphia, is a local one, and must depend on the circum- 
stances; but it may be said that the tendency is toward the former 
plan for large cities, so that it has generally come to be looked on as 
the proper procedure. 

The question as to the value of float experiments in locating sewer 
outfalls is sometimes raised. The speaker believes them to be of great 
value as indicating the direction in which sewage may be expected to 
flow under similar conditions. It is true that a surface float furnishes 
no measure of bottom velocities, which may be quite contrary in 
direction; that the course taken by one float may be quite different 
from that of one set adrift a few minutes later, although under 
conditions which appear to be identical; and that it is influenced more 
or less by the wind ; but in answer it may be said : 

First. — A surface float represents the motion of that portion of the 
current which comprises the greater part of the sewage, through the 
tendency of the latter to rise and spread out over the surface, diffusion 
in a vertical direction being comparatively slow; and if bottom or 
mean velocities are desired these may be determined, approximately, 
by the use of double or rod floats. 

Second. — Although the trace of a single float will not represent a 
true average result, yet it will show the path that sewage discharged 
at the starting point might take; and from a sufficient number of 
records the probability of its going in any given direction may be 
predicted with considerable certainty. 

Third. — A float is influenced by the wind; but this, with properly 
designed floats, is mainly through surface currents induced by the 
wind, and it is the trend of the surface currents which determines the 



402 DISCUSSION : the sewer system of san erancisco 

Mr. Allen, path that sewage would follow. The float, therefore, gives the desired 
information. 

In questions involving dilution, depending on the net discharge of 
tidal and upland waters, the speaker does not consider results from 
floats as reliable as those obtained by current meters, as it is difficult, 
by means of the former, to integrate the motion of the filaments 
properly in a vertical section. Even with the current meter it is a 
long and expensive operation, if reliable results are to be obtained on 
tidal streams. An ingenious apparatus for facilitating such work was 
devised some years ago by Mr. Luigi d'Auria, and used by him on 
some public work undertaken under the direction of the late William 
Ludlow, M. Am. Soc. C. E. This device consisted of a pole which 
could be lowered vertically to the bottom of the stream from the side 
of a pontoon. An endless corrl which passed through pulleys near the top 
and bottom of the pole was used to carry a 2-in. ball of half the specific 
gravity of water to the bottom. At a given signal the ball was released, 
and, rising to the surface at a speed which a computation showed would 
be uniform, and then being caught by a floating grillage, the mean 
velocity was obtained by dividing the measured horizontal distance 
traversed by the time taken to rise to the surface. 

Continuous records of single floats covering several tides are 
valuable in determining the range of tidal oscillation and the resultant 
movement seaward, but, to secure information regarding the probable 
dispersion of sewage, the plan used by the speaker, on Chesapeake 
Bay, of setting a large number of floats adrift at short intervals and 
locating them as frequently as practicable — each being identified by 
a number — is an excellent one to adopt, being ra\nd and eeononiica]. 

The importance of a careful consideration of local rainfall records 
is well illustrated in the case of San Francisco, where, if Eastern 
practice had been blindly followed, the size and cost of the sewers 
would have been greatly increased. In this respect it would be inter- 
esting to learn why the particular form of rainfall fornuila was adopted, 

a 
and whether the more usual and simpler form, I = , would not 

^ t + h 

have answered equally well. 

In forecasting the flow of sewage from a given district, the in- 
filtration of ground-water is often an important but uncertain factor. 
The daily volume has been stated variously in gallons per mile of 
sewer, per acre or square mile drained, per capita of population, per 
linear foot of sewer joint, and per square foot of surface. The 
rational way is to consider infiltration as a function of the sewer sur- 
face, if of continuous concrete or brick, and of the linear foot of 
joint in the case of pipe or other sectional forms of construction. 

The perviousness of the soil affects the rate of leakage to sewers. 
In New Orleans, George G. Earl, M. Am. Soc. C. E., Superintendent 
of the Sewerage and Water Board, found that : 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 403 

"In open and unpaved areas we are probably getting double the Mr. Alleo. 
amount estimated at times of very heavy rains, particularly when 
they follow a long dry spell, which has cracked open our clay soil, 
with deep and frequent fissures. 

"Wliether this flow is all seepage or partly due to flooded surfaces 
where manholes exist and where, even, some one may open a manhole 
cover occasionally, we cannot tell as yet * * *." 

The allowance for infiltration was 0.003 sec-ft. per acre or 31 800 
gal. per mile per day, and an estimate made after construction showed 
this figure— which is the maximum used for San Francisco — to be 
substantially correct. 

In studies made for the sewerage of Baltimore, in 189Y, Messrs. 
Hering and Gray, the Consulting Engineers for that work, recom- 
mended an allowance ranging from 0.00024 to 0.0018 sec-ft. per acre. 

J. N. Hazlehurst, M. Am. Soc. C. E., secured a somewhat excep- 
tional result* with an 8-in. vitrified pipe laid in 2i-ft. lengths, with a 
mortar joint of 1 part Louisville cement and 2 parts sand. This was 
kept covered for 6 hours with a damp cloth, and then immersed for 
24 hours. The infiltration was then found to be only 1 329 gal. per 
mile per day. 

X. H. Goodnough, M. Am. Soc. C. E., reported* that on 137 miles 
of pipe sewers in Massachusetts, varying from 8 to 30 in. in diameter, 
the infiltration was about 40 000 gal. per mile per day before con- 
nections were made, and that it amounted to 70 000 gal. per mile per 
day on 700 miles of sewer, but that 80 000 gal. per day is estimated 
in the extensions of the sewers of the Metropolitan District of Boston. 

In Maiden, Mass., the infiltration was 50 000 gal. per mile per day, 
70 gal. per capita per day, or 0.015 sec-ft. per acre. 

D. H. Maury, Jr., M. Am. Soc. C. E., mentions experiments made 
in Peoria, 111., on a 24-in. vitrified conduit laid in 2-ft. lengths, with 
1 : 1 Portland cement mortar joints, in which the infiltration found by 
numerous measurements was generally from 8 000 to 35 000 gal. per 
mile per day, although, in some instances, the figure exceeded 
60 000 gal. 

The amount of infiltration, of course, will depend largely on the 
ground- water level, 30 000 to 50 000 gal. per day for each mile of 
well-laid pipe sewer being probably a fair allowance under ordinary 
circumstances, but where the exclusion of ground-water is important, 
this may be reduced by special precautions, and where the sewer lies 
above the ground-water level there will be a leakage from instead of 
into the sewer. 

Alexander Potter, Assoc. M. Am. Soc. C. E., has secured good 
results by making joints with sulphur, sand, and pitch, and the 
speaker has done this by the use of tar pitch and cement, kneaded 

* Engineering News, August 27th, 1903. 



404 DISCUSSION : the sewer system of SAN FRANCISCO 

Mr. Allen, by hand to the consistency of dough, as has been the practice for a 
number of years in the construction of sewers in Atlantic City. Some 
20 years ago the speaker attempted to secure a water-tight joint with 
natural (Black Diamond) cement under furnished specifications, but he 
was unsuccessful. More recently, however, he has laid about i mile of 
6-.in. and 8-in. pipe, from to 9 ft. below ground-water level, under 
the usual specifications of Portland cement mortar with a yarn gasket, 
which, after repairing about a dozen small leaks, the contractor made 
practically water-tight. To secure impervious work in this way, how- 
ever, is more difficult and costly than it should be, and the speaker 
believes that some bituminous joint, poured like a lead joint, may 
eventually supersede cement^ — as it already has done to a certain 
extent abroad — and will furnish a more elastic joint, and, at the same 
time, a more economical result. 

In the case of brick sewers, infiltration varies greatly with the 
quality of the work and the depth below ground- water level, but there 
are few definite data to serve as a guide. 

Owing to the general substitution of concrete for brick in larger 
sewers, there is less infiltration than formerly, and this may be done 
away with practically by careful grading of the aggregate, wet mix- 
tures, and continuous work. In the speaker's experience the greatest 
difficulty is at the springing line, where the joint between arch and 
invert, in spite of the bonding key, offers the least resistance to the 
entrance of external water under pressure. Infiltration in concrete 
sewers has been prevented entirely by water-proofing laid between the 
concrete and a brick lining, but the extra cost of the water-proofing 
and the lining is hardly warranted in order to prevent the compara- 
tively small amount of leakage which would otherwise occur. In a 
reinforced concrete sewer built at Waterbury, Conn., the infiltration 
under a head varying from 4 to 12 ft. was found to be 2.08 gal. per 
day per square foot of surface. 

The poor alignment, size, and shape, mentioned by the author as 
existing in the old San Francisco sewers, have been similarly noted 
in Boston, New York, and Baltimore, and may be found in most cities 
large enough to have had sewers more than 50 years old. As a rule, 
smaller municipalities are freer from such absurdities; but, as an 
adopted standard design, the old San Francisco 3 by 5-ft. intersection, 
constructed uniformly 10 ft. below the surface, irrespective of 
topography or connecting grades, is unique. 

The author alludes to changes in the elevation of bench-marks. 
This, to the extent of 0.10 ft. or more, may be due to upheaval by 
frost, or by a lowering of the ground-water level on account of the 
construction of drains, the sinking of wells, etc., depending on the 
character of the soil, unless the foundation extends to bed-rock. Aside 
from this, variations have been known to occur, due to more general 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 405 

movement of the earth's crust. Large areas are known to be involved Mr. Alien. 
in these movements, such as the gradual sinking of the New England 
Coast, which, as John R. Freeman, M. Am. Soc. C. E., has pointed 
out,* is taking place at the rate of about a foot a century. 

But there is reason to believe that there may be other alterations 
in the relative elevations of points within such short distances of each 
other as to affect the bench-marks of a municipality. 

Some time ago, Horace Andrews, M. Am. Soc. C. E., brought a 
case in point to the attention of the speaker in connection with the 
"re-levelling of a number of points of the Bavarian precise levels 
which were originally determined in the interval between 1869 and 
1873 and the mean levels taken in 1897."t These were both levels 
of the highest order, with refinements as to accuracy of standard and 
permanent marking of bench-marks, which, unfortunately, are often 
disregarded. 

"The discrepancies revealed were about equally plus and minus 
and in only one or two cases did there seem to have been some gross 
change in the foundation of the building on which the mark was 
placed. 

"The differences in this Bavarian work were found by touching on 
nearly fifty points scattered over a large region. * * * Dig. 
crepancies of from 13 to 14 mm. seemed to be common, while several 
were as large as 15 to 17 mm. and two were of 41.5 mm. and one was 
86.7 mm. In addition to these was the gross change alluded to of 
233.5 mm. in one point." 

About 1894, precise levels were run in connection with a general 
topographical survey of Baltimore then being made by Colonel H. T. 
Douglas, and numerous bench-marks were established. Twelve years 
later the speaker had occasion to test a good many of these, and, 
although precise levels were not run, the discrepancies noted were 
checked so carefully that he was convinced that numerous changes 
had occurred during the intervening years. This opinion was con- 
firmed by results reached on another division of the same work, where 
it was observed that the greatest discrepancies occurred where the 
slopes were most abrupt. Discrepancies of this kind might appear to 
be due to a difference in the graduation of the rods, but as these had 
been compared carefully with a standard, the conclusion reached was 
that this was not the cause. The apparent errors of elevation noted 
varied up to several hundredths of a foot. 

It would be interesting to learn the general experience of municipal 
engineers who have had the opportunity of checking benches estab- 
lished by precise methods after the lapse of several years. 

Walter N. Erickstad, Assoc. M. Am. Soc. C. E. (by letter). — Mr. Frickstaci. 
During the past two years, Oakland, on the east shore of San Francisco 

* Report on the Charles River Dam, 1905. 

t Zeltschrift fiir Verniessungswesen, August 1st, 1900. 



406 DISCUSSION : the sewer system of SAN FRANCISCO 

Mr. Frickstad. Bay, has spent about $750 000 on culverts and sewers carrying storm- 
water, and additional sewers have been, or are about to be, designed. 
The general method of determining the capacity is adapted from one 
of Mr. Grunsky's reports on "A Sewer System for San Francisco," 
substantially outlined in this paper as the "Method of 1899." To those 
in the city engineer's office this seemed to be a clear, logical, and simple 
method of making the computations, after having determined certain 
preliminary assumptions, involving questions of expediency and policy. 
The claim, therefore, that an improved method has been evolved is of 
special interest to the writer, who is connected with the city engineer's 
office in Oakland. 

It must soon become apparent to anyone who attempts to design 
a practical and economical storm sewer that he has attacked a difficult 
problem, very unsatisfactory from a mathematician's point of view. 
There are inherent uncertainties which leave the final solution largely 
a result of the designer's judgment. The principal questions, all or 
several of which apply to any drainage area, are as follows: 

1. — In a growing community, a correctly proportioned storm sewer 
or culvert will not only serve existing conditions, but conditions as 
they will exist at some future date. Shall that date be ten, twenty, 
thirty, or a hundred years hence? A mathematical answer to this 
question, obtained from interest or sinking fund tables, is not satis- 
factory, because too many other considerations are involved. This is 
especially the case when funds must be obtained from a reluctant city 
governing board, or by the approval of two-thirds of the voters at a 
"bond election," or over the protests of a large proportion of the 
property owners to be specially assessed. 

2. — What will be the population of the district at the end of the 
period selected? Estimates can readily be made, but on final analysis, 
the answer is pure guesswork. 

3. — What is the relation between run-off and population, or, to 
make the question simpler, between impervious area and population? 
This can be approximated, but the inaccuracy is indicated by the fact 
that the figures are usually given as 25, 50, 75, or 100 per cent. 

4. — Assuming a certain population, or a certain proportion of 
"impervious area," how much of the area is truly impervious, as the 
term is used in sewer design? For example, what proportion of the 
roof area is connected directly to the sewers? To what extent are 
macadam streets and gravel walks actually impervious? 

5. — What will be the effect of pervious area on run-off? This will 
vary extremely within a single district, according to the character of 
the top soil, the subsoil, and the degree of saturation at the beginning 
of a critical period. Authorities differ widely in their statements of 
percentages; and, as a matter of fact, the percentage will not be the 
same for any two districts. This, however, for excessive storms over 
small areas, is one of the minor considerations. 



DISCUSSION : THE SEWER SYSTEM OF SAN FEANCISCO 



407 



6. — What will be the effect of varying grades within the district, Mr. Frickstad. 
both surface grades as they exist, and hydraulic grades as they vary 
from moment to moment ? Very little attempt has been made to answer 
this question. 

7. — What will be the effect of variations of rainfall from moment to 
moment during the limits of the critical storm? 

8. — What shall be assumed as the maximum rainfall rate? Few 
cities have records showing rainfall rates for short intervals covering 
the past forty or fifty years, which data are necessary before a satis- 
factory curve can be plotted. In default thereof, most cities do the 
best they can with incomplete records, allowing considerable weight 
to the records of cities of like conditions. Even if the records be com- 
plete, is it not fairly certain that at some time all known rainfall rates 
will be exceeded? 

Curves of total yearly rainfall, the proximity of the two cities, and 
general conditions would seem to indicate that rainfall rates in Oak- 
land should be the same as in San Francisco. On March 30th, 1906, 
however, P. F. Brown, M. Am. Soc. C. E., observed rates which were 
higher than any recorded for San Francisco. Table 21 is a comparison 
of his limiting rain curve, and the San Francisco curve. 

TABLE 21. 





Rates, in Inches per Hour. 


Duration, in minutes. 




Oakland (1906). 


San Francisco (to 1908).* 


5 


2.60 


2.16 


10 


2.00 


1.50 


20 


1.41 (Measured) 


1.02 


30 


1.10 (Measured) 


0.83 


40 


0.94 


0.72 


50 


0.81 


0.65 


60 


0.74 (Measured) 


0.60 



* From Mr. Grunsky's paper. 



9.- — Shall the sewer be designed for the extreme run-off, or shall we 
deliberately permit flooding of the streets, with resulting damage, once 
annually, or once in three, five, or ten years? This, which also may 
modify the answer to the first question, is partly a matter of engineer- 
ing economics, but is more closely related to the funds available. It is 
a very practical question, especially under the separate system and in 
suburban districts, and may modify the capacity of the sewer several 
hundred per cent. 

10. — What will be the cost of duplicating the structure, cutting off 
part of the district, or otherwise furnishing relief at some future time? 
The answer to this may modify materially the answer to the first 
question. 



408 DISCUSSION : the sewer system of san francisco 

Mr. Frickstad. 11. — What retardation of the water may be expected from the shape 
of the ground surface? This question is partly involved in the fourth, 
fifth, and sixth questions, but deserves a separate statement. Average 
ground, and the improvement thereof, consists of a series of hollows, 
runs, ridges, and hills. It may be said that the rain, as it falls, gathers 
in the low places and flows thence over a series of weirs. These weirs 
vary infinitely in shape and length, and, as the pond changes in depth, 
the depths over the weir or the velocity of approach, or both, vary 
extremely. This question is especially important in a small drainage 
area, and its influence on design can never be determined mathe- 
matically. 

12.^ — Where cold weather may be expected, what allowance should 
be made for a combination of frozen ground, a large body of snow, 
and a warm rain ? 

13. — What shall be the final factor of safety ? 

An examination of these uncertainties, even omitting those which 
might be settled by a careful census of the population and a physical 
examination, will show that the capacity of a sewer designed for the 
ultimate development of a district, with unlimited funds available, to 
provide absolute protection, might exceed by several hundred per cent, 
the capacity of a sewer sufficient for existing conditions, providing 
relief for all times of the year excepting a fraction of one hour. 

The proposed new method provides for definite procedure by intro- 
ducing the factor of storage in permeable soil. Before applying the 
method, however, it is necessary to make assumptions as to the degree 
of saturation at the beginning of the storm; and to observe or assume 
a rate of permeability — at best an average of an infinite number of 
rates which will actually occur in the given area. 

The new method also assumes that one must take into account one 
of the delays to which water is subject after it falls upon the earth. 
Omitting for the present, any doubt about the correctness of the theory, 
it should be noted that these delays arise from many causes, and, even 
as a whole, are not one of the largest factors in determining capacity. 

In two specific cases in Oakland the conduit capacity per acre 
drained has been examined. In the first case, called the "Grove and 
Jones Streets storm sewer," at the corner of these two streets, the 
conduit is a 4i-f t. circular concrete sewer, carrying 87 cu. ft. per sec. 
The ultimate area to be drained is 175 acres. The present conduit 
capacity is about 27 000 cu. ft., and the ultimate capacity, on the com- 
pletion of the upper end, will be about 50 000 cu. ft. Taking the 
author's assumption that the conduit will be half full at the beginning 
of the period, the volume to be filled ultimately will be 25 000 cu. ft., 
being 286 cu. ft. per acre, or a layer of 0.079 in. over the area. The 
probable rainfall rate, based on a 30-min. period, is 1,10 in. per hour, 
of which 0.55 in. would fall in the period. The increase of water in the 
conduits, therefore, would be about 14% of that which falls in the 



DISCUSSION" : THE SEWER SYSTEM OF SAN FRANCISCO 409 

district. This is a fair illustration of conditions in a residence district Mr. Fnckstad. 
with moderate grade. 

The second case has somewhat less population, and has steep grades. 
At the corner of Eleventh Avenue and East Nineteenth Street the 
conduit is a 24-in. pipe, carrying 32 cu. ft. per sec. The area drained 
is 67 acres. The present conduit capacity above the point mentioned 
is about 3 300 cu. ft., but ultimately, extensions at the upper end may 
double this. The area available for filling during a storm, therefore, 
may be taken as 3 300 cu. ft., being about 50 cu. ft. per acre. This 
would be a layer of 0.0138 in. over the area. The probable rainfall in 
20 min. was taken as 0.47 in., of which the available area in the con- 
duits, when extended, would take about 3 per cent. 

The engineer, by the exercise of judgment, or through force of cir- 
cumstances over which he has no control, must fix assumptions in 
answer to all the uncertainties enumerated. In view of this fact, undue 
refinements in any one of them, or in calculations based thereon, seem 
to be out of place. 

Admitting that conditions in a given locality might warrant refined 
calculations (as, for example, in a fully developed district), there are 
several points in the theory of the author's new method which seem 
fallacious to the writer. 

Of these points the principal one is the introduction of "conduit 
storage" as affecting the maximum rate of flow past a given point. 
This would certainly imply that portions of the water suffer a retarda- 
tion of velocity in some part of their course. This is not conceivable 
in a well-designed system where each segment has a free outlet, except- 
ing that it may be retarded by flatter grades on the lower portions. It 
is not believed, however, that the author considers this retardation in his 
deductions, as this is taken care of in his discarded "Method of 1899." 
As a matter of fact, there would be a slight acceleration of velocity 
during the critical period over that given by the sewer grade, due to a 
slight increase in the hydraulic grade during the rise of water in the 
sewer, according to recorded velocities observed before and after the 
crest of a flood in open streams. 

The true action of the run-off may best be explained by reverting 
to the author's illustration of a small area contributory to an inlet. 
After stating the assumptions of the illustration, he says: 

"The area tributary to each inlet may now be regarded as sub- 
divided into 5 concentric zones, from the innermost of which water 
will reach the inlet in the average time of i min., from the next in 
li min., and so on to the last one in 4^ min." 

******* 

"The water on the ground at the end of the 5-min. period, there- 

66 X 

fore, is , or two-thirds of the total amount of rain wiiicli fell in 

' 100 x' 

that period. 



410 DISCUSSION : the sewer system of SAN FRANCISCO 

Mr. Frickstad. "Generalizing, it may be said that what is true of an area with 
circular outline is true of other compact shapes, and that, therefore, 
two-thirds of the water which falls during any time period equal to the 
time required by water to reach the inlets from the outer portion of 
areas drained by each, will, at the end of that period, be still on the 
surface in transit to the inlet." 

A fvirther examination of this illustration shows that on the inner 

98a; 
area, at the end of the period, there will be *— -, being practically three- 

o 

tenths of the whole GGx. In the second area there will be 18a;, being 
about 27% of the whole. In the third there will be 23%, in the fourth, 
15%, and in the outer, 5 per cent. The average depths of water over 
the several areas will be as 98, 30, 15, 7^, and 2. 

All this increase of water appears in spite of the fact that it has 
opportunity to flow away freely. It is a misnomer, therefore, to call it 
"storage," for there is no retardation of velocity under the conditions 
of the illustration. 

Furthermore, the above figures, while showing an accession of 
water on the areas near the inlet, are not factors in determining the 
maximum rate of flow into the inlet at the end of the period, which may 
be truly called the "critical instant." To be sure, the rate can be 
calculated therefrom, b;it is, quite independent of all the above, rain- 
fall intensity times area. Or, in actual practice, intensity times area, 
in proper units, less storage (water delayed in transit) due to causes 
not appearing in the theoretical illustration. 

It may further be said that this illustration is in no way modified 
if the water be conceived to be wholly or partly concentrated in rivulets 
or gutters, instead of being spread uniformly over the ground. 

If this illustration be a correct conception of a small drainage area, 
it may properly be extended to cover a larger district, for example, a 
district which requires 60 min. for the water to flow from the outer 
zone to the inlet. The quantity of water on the surface at the "critical 
instant," for which the capacity of the inlet must be proportioned, may 
readily be determined ; but, independently, the flow into the inlet will 
be intensity of rainfall times area, less the quantity retarded by soaking 
into the ground, and less the quantity retarded in transit due to the 
pond-and-weir character of the average ground surface. Here, again, 
it would seem fair to conclude that the theoretical rate would not be 
modified if the water approaches the inlet in a uniform sheet, or in 
rivulets, or in artificial conduits. All such modifications, as well as 
those of varying grades, are cared for by the assumption that 60 min., 
no more and no less, are required to make all parts of the area tributary. 

At this point it should be emphasized that a storm-water system is 
merely a convenient underground concentration of water which other- 
wise would flow over the ground. For the purposes of this discussion, 
it is assumed to have a course as free as the open gi-ound. Its capacity 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 411 

at each point, therefore, must be at least as large as is indicated by Mr. Frickstad. 
the location of that point in the system of concentric rings described 
in the illustration. Hence the above statement, that conditions will in 
no way be changed if the water be conceived to be concentrated in 
channels, that is, the quantity of water within the ring, and the rate 
of flow across the area, will be unchanged. 

Referring again to the Oakland examples cited, it will be recalled 
that sewers have been built through the lower ends of the districts. 
Later, as convenience or necessity requires, the sewers will be extended 
farther into the upper ends. It is somewhat uncertain how the new 
method would be applied to these cases. The actual volumetric capacity 
of the conduits, as designed in each case, is about half the ultimate 
capacity. Following the new method blindly, it would seem that, the 
reserve volume being less, the sewer should now be larger than when 
finally completed. On the other hand, the water is at present flowing- 
over the ground, or on the streets, in courses having an infinite 
capacity. Consistent with the theory of the new method, this condition 
should exert a large influence on the rate of flow of the sewer at any 
point considered. In fact, since part of the system has an indefinitely 
large capacity, might it not be said that the flow in the constructed 
portion will be very small because of '^storage" in the open portion? 
As a matter of fact, this will be partly true, but it will be because of 
actual retardation on the ground and in the gutters. 

Furthermore, it would appear, from the theory of this method, 
that the larger the conduit capacity above a certain point the smaller 
may be the sewer at that point. Obviously, there must be some limit 
to this, or it might be concluded that an inlet may be very small because 
the surrounding gutters and earth have an infinite capacity; and the 
outlet of a system may be small because the upper end is grossly 
too large. 

The fundamental error, if there be such in the author's deductions, 
seems to be in the statement preceding Equation 1. That is, while the 
statement is strictly correct, it is not clear how the total quantity of rain 
falling on a given area, or the total quantity flowing therefrom in a 
given period, governs the computations for capacity. It is the maxi- 
mum flow, that at the "critical instant," at the end of the "critical 
period," which presumably is required. Of course, if an examination 
of the average or total run-ofl' is undertaken, for the purpose of 
determining the maximum, and that course is proved to be the simplest 
and most accurate, it should be adopted; but the relation of the average 
to the maximum, in Equation 5, depends on an assumed hydrograph, 
which is probably not correct for any actual storm. 

In brief, the wi'iter's whole doubt of the theory of the new method 
may be summed up in the question : Does not the "critical instant" 
come at the end of the "critical period" for a storm-sewer system, all 



413 DISCUSSION : the sewer system of SAN FRANCISCO 

Mr. Frickstad. portions of which have a free outlet; and can the condition of the 
conduit at any time previous thereto — or after — have any influence on 
the capacity at that instant ? 

Mr. Grunsky. C. E. Grunsky, M. Am. Soc. C. E. (by letter). — The conduit storage 
feature of the writer's solution of the storm-water flow problem will 
perhaps be made somewhat clearer by the use of a diagram which 
illustrates graphically what takes place in a drainage basin during a 
downpour having a duration of f^ minutes. 

Let the time period, tj^ (which must not be mistaken for the critical 
time, <„J, be divided into ten sub-periods, each 0.1 ^^. It will then be 
found that for values of tp, as ordinarily occurring in municipal 
problems, the total maximum rain, q, may be expected to fall in the 
sub-periods as follows: 

In one sub-period of 0.1 t^ minute's duration 0.30 q. 

a a u u a u u 14 O 

u a u u '.i a " Q in Q 

« " " " " '' " 0.085 g. 

" " " " " " " 0.075 Q. 

" " " « " " " 0.070 g. 

" " " " " " " 0.065 Q. 

" " " " " " " 0.060 g. 

« " " " " " " 0.055 g. 

« " " '< " " " 0.050 g. 

These values are based on the San Erancisco, and on the parabolic, 
types of rain curves. In order that their departure from a formula of 

a 
the general type, I = . , ,j , may be apparent, there are shown in 

Table 22, in parallel columns, sub-values as determined for a 60-min. 
period, by the San Erancisco formula, by the parabolic formula, and 

2 t 
by the Kuichling formula, the latter, q = , being based on 

t ~\~ Ji\j 

120 
~ t + 20' 

The sub-periods are arranged in Table 22 as they would appear in 
a rain of decreasing intensity. They may follow each other in any 
order. In other words, the rain curve, as established by a recording 
gauge, will lie somewhere between two curves, the one being the 
limiting rain curve, representing a rain of decreasing intensity, and 
the other, the same curve inverted, representing a rain of increasing 
intensity. 

The drainage basin or district under consideration can be sub- 
divided into ten sub-areas such that the time, including the inlet time, 
which it will take for the water to flow from the limits of the inner- 
most sub-area will be 0.1 i^, from the limits of the next 0.2 «^, from 
the limits of the next 0.3 f^, and so on. Let these sub-areas be desig- 



DISCUSSION" : THE SEWER SYSTEM OP SAN" FRANCISCO 



413 



Mr. Grunsky. 







— 




— 








aA 

Effective Area. 




^ 


l-l . 




^a L o _. 








. . _ 


p 
p" 


4^- 


\h. 


i 


\. 


> 


h > 










































-bi 


















^^^ ^^ 






i 


: 




: 








The shaded area is the 

Discharge at the Gauge Point 

during the time-increment, 0. 

which precedes the time, T. 


i 


f" 






6 


=> 


c 












^ 








f 


o 






y 












I 
f 


i 

i( 




X x^ 












p 

i5* 


1 

V 


/ 








RUN-OFF DIAGRAM 

FOR CONSTANT 

VELOCITIES. 








5^ 






y/ 














4/ 


A 














Al 


:/ 


a 














Y 


fc 




a 




















"8 























































414 



DISCUSSION : THE SEWER SYSTEM OF SAN" FRANCISCO 



Mr. Grunsky. nated, in their order from the iiinerinost to the outermost, A^, A^, A^, 

A^Q. A diagram, as shown in Fig. 17, can now be prepared. 

Here, the sub-areas, each modified by its proper reduction factor from 
Table 6, are platted up from the base line, and horizontal lines are 
drawn in such a way that the distances between the successive lines 
correspond to the effective area values, a« -4.„. 

TABLE 22. — Thf: Distribution of the Maximum Rainfall in 60 
Minutes to Sub-periods of 6 Minutes each. 

The sub-ppriods are arranged in the order of rain intensities. 



Number of the 


Rainfall in 0.1 tj^ Minutes, when tj^ = 60 Minutes. 


sub-period 
each 0.1 tj^ rain. 


San Francisco 
formula. 


Parabolic formula. 


Kuichling formula, 
q = 1.50. 


1 


0.29 q 
0.13 q 
0.10 q 
0.080 q 
0.075 q 
0.072 q 
0.067 q 
0.065 q 
0.U62 q 
0.057 q 


0.32 q 
0.13 q 
0.10 q 
0.085 q 
0.075 q 
0.067 q 
0.063 q 
0.058 q 
0.053 q 
0.050 q 


0.31 q 




0.19 q 


8 


0.13 q 
0.095 q 


4 


5 


0.073 q 


6 


0.056 q 


7 


0.047 (/ 


8 


0.039 q 


9 


0.032 q 


10 


0.027 q 







The greatest rain which may fall in any sub-period (0.1 t^^ minutes) 
is now scaled off along some convenient portion of the base line, and 
vertical lines are drawn from each end thereof. The rainfall in each 
of the other sub-periods is also laid off along the base line in such 
manner that the sum of the rain in the ten sub-periods is equal to the 
total rain in the time, t^. In the diagram, the sub-periods are shown 
in the same order as in the table. They may be arranged in any other 
order, as already stated. 

Assuming — though merely for the purpose of illustration — that, 
throughout the time i^, when gutters and conduits are carrying con- 
siderable water, the velocities remain constant in all parts of the 
system (being those assumed in calculating A^, A^, A^, etc.), then 
diagonal lines can be drawn through each rectangle in the diagram, 
each line representing the progress of the water through a sub-district. 
The flow of water through the whole district can now be traced by 
following the successive connected diagonals from the outer sub- 
district to and through the innermost. 

The area, D B T, Fig. 17, is the run-off from the district in the i^ 
minutes, B to T, which has passed the outfall point at the time, T. 
The area, D E T, is the water in temporary storage in conduits and on 
the surface, that is to say, the water which is still in transit. 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 



415 



The area between each two successive oblique lines, as drawn in the Mr. Grunsky. 
diagram, from the upper horizontal to the base line, represents the 
total run-off in the successive increments of time, each of which is 0.1 i^ 
minutes in duration. 

If the assumption of constant velocities were correct, diagrams of 
the character shown would lead directly to a determination of dis- 
charge values and to discharge curves. Fig. 18, but this is not a safe 
assumption. It has here been made merely as a basis for the prepara- 
tion of diagrams, such as the one shown, to aid in determining the 
probable time after the occurrence of the maximum rain intensity when 
the discharge will be a maximum. 











THE 


DISCHARGE-CURVE BASED 


ON FIG. 17, 




































sA 


















^ 


\ 






o 


a 

n 














^ 






\ 






I 

1^ 


a 
1 










^ 


^ 








\ 


\ 




> 
1 

"3 










^ 


1 


'he disc 
ifvelc 


harge- 
icities ] 


mrve a 
emaiu 


s it woi 
!d cons 


ildbe 
tant 


V 


^ 




g 

£ 




^ 






















s 


^0.1 <^> 


^0.1 «i,-> 


<-0.1 ttr- 


^-O.l t!r> 


"-0.1 ifl> 


■^0.1 to" 


<-0.1 tu^ 


^0.1 ls> 


*0.1 te* 


■^-0.1 to-> 




















f 







































Fig. 18. 



The diagram. Fig. 19, serves to illustrate what takes place under 
the assumption of rain of uniform intensity in the many small areas 
tributary to inlets. In the case of these areas, the successive sub- 
areas will increase in extent from the innermost to the outermost about 
as stated in the paper. This diagram for the progressively increasing 
sub-areas tributary to inlets, illustrates and confirms the conclusions 
already presented. 

By constructing diagrams, similar to those shown in Figs. 17 and 
18, for areas of various shapes, and for rains of various types, a 
number of discharge curves were obtained, each of which served to 
indicate, approximately at least, the time at which discharge would 
be at maximum value. It was found that, for the same value of <^, 
the shortest interval between maximum rain intensity and maximum 
discharge would obtain in the district in which the effective center 



416 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 



Mr. arunsky. (computed with due regard to the location and effective size of each 
sub-area) was nearest to the gauge point. 

Thus, for a fan-shaped district, having sub-areas which increase 
from the innermost to the outermost in the ratios (for effective areas) 
of 1 to 3 to 5, etc., the greatest discharge will take place about <^ 
minutes after the beginning of the downpour. In districts with uni- 
form effective sub-areas, this time would be about 0.90 tjj minutes, and 
in districts that are of broad-base triangular shape, the time would be 
about 0.40 tj^ minutes. 

RAIN OF UNIFORM INTENSITY FOR i MINUTES. 
I) Inlet Areas. 




Fig. 19. 



This fact suggests that the ratio of t^ to tj^ is a function of the 
shape of the district. The diagram, Fig. 17, is based on a fan-shaped 
district, the sub-areas increasing with distance from the gauge point. 

Taking the foregoing into account, and allowing for the fact that 
the actual discharge curves will depart somewhat in shape from those 
determined on the assumption of constant velocities, a formula has 
been framed in which weight is given to the distance (in time of flow) 
of the sub-areas from the gauge point. This formula, having proper 
regard for the permeability of each sub-area, may be written: 

^ (0.1 «! ^1 + 0.2 a, A.^ + 0..3 a^ A^ -j^ 0.4 a^ A^ -f 0.5 a^ A^ -f- 



«„. = 



a A 



0.6 ttg J.g + 0.7 tty A^ -f- 0.8 ttg J.g -f 0.9 a^ A^ -f a^p ^k,). 



(63) 



discussion: the sewer system of SAN FRANCISCO 417 

The formula, if written for five instead of for ten sub-areas, becomes : Mr. Grunsky. 
195/ 

t^ = ^^y (0.2 «! A^ + 0.4 ^2 A., + 0.6 ag A^ + 0.8 a, A^ + a^ A^) . . (04) 

which may be. written 

25 ^ 
t„, = ^^^(«^i ^1 + 2 a^ ^2 + 3 «3 ^3 + 4 «4 -"^4 + -5 «5 ^5) (65) 

For a district which has impervious surface throughout: 

«.=-^^(^i+2^2 + 3^3 + 4JL,+ oA,) (66) 

And, in special cases (all impervious areas) : 

1. — For A J to J. 2 to J. 3, etc., as 1 to 3 to 5, etc., 

tm ~~~ o.yo t j^, 
2. — For A^ = A^ = A^, etc., 

tm = 0.75 t^. 
3. — For Aj^ to ^2 to Ag, etc., as 9 to 7 to 5 to 3 to 1, 

^^ = 0.55 t^. 
In ordinary cases, the subdivision of a district into three sub-areas 
will suffice. In that event, the formula may be written: 

t,^^ = 0.40 — ^ (ttj J-i + 2 a, JL2 + 3 ag ^3) (67) 

<x A 

It has been shown in the paper that tm varies with the square of S 
(Equation 33) ; consequently, if the foregoing equation for t^ is cor- 
rect, the storage space, i. e., the conduit contents, should be greater per 
acre for a district that broadens with distance from the gauge point 
than for one with its greatest extent near the gauge point. As this 
is, in fact, the case, it may be regarded as confirming in a measure 
the conclusion which was reached by another line of reasoning. 

The formula for im (Equation 67), does not contain 8, and there- 
fore it becomes possible to estimate the duration of the critical rain 
and also the maximum discharge without first approximating the value 
of 8. This is as it should be, because the increase of water in transit 
during a critical period, whether the water flows on the surface or 
in underground conduits, varies with the conditions that modify its 
velocity of flow. These same conditions fix the value of /^, from which, 
as shown, i^ can be approximated. 

What shall be assumed as the maximum rainfall rate is very prop- 
erly asked by Mr. Frickstad. The available rain records are usually 
single-station records. When the rainfalls at several points in the 
same small drainage basin are combined, the combination should be 
made, not by averaging results, but by combining the rain curves ac- 
cording to actual time. It is evident that, if there were a number of 
station records in a district, the composite curve (long-time records 
being assumed), would show a smaller maximum rain rate than single- 



418 DISCUSSION : the sewer system of SAN FRANCISCO 

Mr. Grunsky. station records, because the station maxima will not always fall to- 
gether. This is particularly true for short-time periods. Consequently, 
when areas are of considerable extent, the probability that station 
maxima will exceed the maxima that should be used in constructing 
curves should be recognized, and a proper limiting curve may be drawn 
that will not fall beyond all the points of observation. On the other 
hand, a storm may travel down stream in the drainage basin in such 
a way that the time of maximum rain intensity is later at the gauging 
point than in the heart of the district. Under such conditions, there 
may be unusual concentration of run-off waters at the outfall point. 
This fact, which is of importance in large districts, should be allowed 
for by giving the rain curve a safe position for the long periods of 
time. It is well, therefore, to let few if any observed points for periods 
of 1 hour or more fall outside of the rain curve, while some records, 
for 5 and 10 min., and an occasional record for other periods less than 
an hour, above the line may not be objectionable. Conditions, how- 
ever, may be such that, owing to the high initial cost of a conduit 
system, it will be unwise to attempt to prevent all inundation. In such 
case, the rain curve may fall far short of being strictly a limiting 
curve. 

The statement will bear repetition that a reliable rain curve can be 
constructed for any locality if the maximum rain in some period of 
time, about one hour in duration, has been correctly determined by 
observation. 

The rain records cited by Mr. Dieck for Manila indicate that a safe 
limiting rain curve will be obtained by making & = 17 or e = 0.35. 
The resulting value of B is about 2.75 in. 

Mr. Allen asks why it is necessary to depart from the type of 
formula for rain intensity which has heretofore found general accept- 
ance. This is necessary because the older type does not, with sufficient 
accuracy, indicate rain rates or rain amounts for the longer time 
periods that must occasionally be given consideration in designing 
storm-water conduits. No formula which is based upon the assump- 
tion of a fixed uniform quantity of rain in varying times, or which 
approximates such a type, vsdll answer. Those who wish to cut loose 
from all formulas of rain intensity may assume a specific rain distribu- 
tion to sub-periods of time, as has already been pointed out. 

Mr. Gregory is undoubtedly correct in what he says relating to the 
great variation in the time that it takes water to reach inlets. When 
this time is assumed to be short — 5 min. — the time of concentration 
is cut down, and the rain rate entering into the formulas is increased. 
The assumption of a short inlet time was made by the writer in order 
that the resulting formula may to this extent be a safe formula; but, 
it is shovsra in the paper that some variation in the time of flow to 
inlets will make but little difference in the resulting formula, except 



DISCUSSION : THE SEWER SYSTEM OF SAN FRANCISCO 419 

only for very small areas. It is the increase or decrease of the water Mr. Gnmsky. 
in transit which counts, and not the amount thereof. When there is 
no increase or decrease, the run-off is 100%, as quoted by Mr. Gregory, 
and this is the essence of the writer's reasoning (see Table 5). 

The estimate of time, made by Mr. Gregory, at which the run-off 
from any impervious surface becomes equal to the rain rate (the rain 
rate being assumed uniform) is interesting and instructive, but of no 
apparent application to the theory of run-off presented in the paper. 
This time is not the time, tj^, which it will take water, during extreme 
rain conditions, to flow to the inlets. The point in connection with 
the whole problem which is particularly to be emphasized is that it 
matters very little whether the storage be within the conduits or on 
the surface, provided only that tj^ is estimated correctly, being the dis- 
tance to the inlets divided by the velocity of approach to the inlets (no 
matter whether this be on roofs or across lawns or in gutters), plus 
the time of flow in the conduits from the outlying inlets to the gauge 
point. 

Data are not as complete as desirable for a proper application of 
the new formula to the Sixth Avenue sewer in New York City, the 
flow of which, on August 4th, 1888, has been cited by Mr. Gregory. 
He states, in reference to this sewer, that it takes 15 min. for water 
to flow through the sewers from the remote parts of the district to the 
gauge. He also says that there are many 600 and 900-ft. blocks in 
the district, and many flat roofs. Assuming 10 min. as the allowance 
for the time required by water to reach the inlets, then ^^ = 15 + 10 
= 25 min., and, taking t^n = 0.80 tj^, makes t^ = 20 min. 

The downpour on August 4th, 1888, produced a maximum run-off 
of 1.16 sec-ft. per acre. 

The measured rain is reported to have been at the rate of 3.54 in. 
for 10 min. For 20 min. it was at the rate of 2.59 in. per hour.* 

Using the rate for 20 min. and the formula as developed for a San 
Francisco type of limiting curve, it will be found that : 

h = 9.89t 
R = 1.61 

The district is 90%" impervious; therefore, from Table 6, for i^ = 

25 min. : 

a = 0.98 

, , 4.36 a i? , „, , 
and cZ„j = — —-— — =-■ 1.71 sec-ft. per acre. 

This is 50% more than the recorded flow, estimated from the water 
stage in the sewer. The sewer was at the time surcharged. It is 
doubtfi;l, therefore, whether the value, 1.16 sec-ft. per acre, is the 

* Transactions, Am. Soc. C. E., Vol. LVIII. p. 464, Plate LXIX. 

tif thelO-min. rate of 3.54 in. had been used, the value of 6 would have been 9.91. 



430 DISCUSSION : the sewer system of SAN FRANCISCO 

Mr. Grunsky. real maximum that would have occurred if the sewer had been larger. 

It is possible, too, that the value of t^ used in the formula has been 

underestimated. It has been assumed that the measured rain rates 

in this case are reliable. 

The parabolic type of rain curve for the same elements above used 

would make: 

e = 0.193 

i? = 1.50 

-, -, 5aJ2 ^ „^ „^ 
and a,„ = — 3;^ = 1.65 sec-ft. per acre. 

No basis can be found for the opinion expressed by Mr. Kuichling, 
that the effect of storage upon discharge, which he admits for small 
districts, is lost in large districts. The diagram. Fig. 17, should afiord 
a complete answer to this view. The diagram is applicable without 
regard to the size of the drainage basin. 

The new formula may be regarded as a confirmation of the good 
judgment heretofore used by engineers, particularly by Mr. Kuichling, 
in adopting reduction factors to be applied to rain rates when run-off is 
to be estimated. As long as channel storage, in its effect upon the 
rate of discharge, was overlooked, it was natural to assume that absorp- 
tion of water was the main cause for a discharge value less than the 
rain rate; but, the reduction is, in part, sometimes altogether, as has 
been shown, due to the changing quantity of water in transit to the 
gauge point. It remains but to add that the new formula as written, 
or the formula with some modification of its numerical factor based 
on better knowledge of the shapes of the discharge curves from areas of 
all possible outlines, will leave small room for error in the case of 
impervious or nearly impervious areas. It is not necessary to assume 
any loss of water from such areas ; all is accounted for. 

A satisfactory concrete example for Eastern rainfall conditions, 
asked for by Mr. Kuichling, cannot be given without a study of rainfall 
data, for which time has not been available. 

Of the series of questions asked by Mr. Frickstad, nearly all re- 
lating to the run-off under known conditions of rain have been an- 
swered in the paper. The conduit capacity which has entered into the 
demonstration is the storage capacity of an adequate system of con- 
duits. Where conduits are inadequate, or are laid only in portions of 
a district, the value of S includes all increase of water on the surface 
as well as that in the conduits. Care should be taken that the storage 
effect be combined with the surface run-off, and not with the rain. The 
rain rate and the surface run-off rate are only then equal when the 
surface is impervious throughout. 

Oakland has a somewhat heavier rainfall than San Francisco. In 
applying the formulas based on the San Francisco type of rain curve, 



DISCUSSION : THE SEWEK SYSTEM OF SAN FEANCISOO 421 

the value of b should probably be placed somewhere between 4.00 and Mr. Grunsky. 
4.50. It is quite possible that the parabola will more nearly approx- 
imate the Oakland rain curve. 

The term, storage, as used in the paper, is fully explained. The 
quantity of water in transit is continually changing. It increases 
during a downpour. The case is exactly the same as though the 
water were flowing through a reservoir. If this view be taken of the 
problem, it will be clear that the rain preceding a critical period may 
have some effect on the flow during the critical period, and, therefore, 
on the maximum flow. Let it be supposed that there be sufficient con- 
duit capacity above the gauge point to give the conduit the character 
of a reservoir. It will then be plain that a maximum rain will pro- 
duce smaller run-off rates if the reservoir be empty at the beginning of 
the critical period than if it be full. In other words, the most unfavor- 
able critical period will be one preceded by heavy rain; but, on this 
point, enough has been said. 

In order to make clear the effect upon the formula of the assump- 
tion that the discharge curve has a shape during a critical period, sub- 
stantially as shown in Fig. 7, it may be noted that, if the discharge 
curve had been taken as a straight line from B to T, that is, if it had 
been assumed that the discharge increases at a uniform rate, then the 
resultant formula for dm would have contained a numerical factor, 
0.83, instead of 0.71. The probability is thought to lie with the lower 
value, for all ordinary problems. Special assumptions and resulting 
modifications of the formula may become necessary when the range in 
the sub-area sizes is out of the ordinary. 

The sinking or settling of certain areas in San Francisco alluded to 
in the paper, is a local phenomenon, due entirely to compression of the 
material on which parts of the city have been built. The writer, while 
City Engineer, commenced the work of precise leveling which is being 
extended to all parts of the city. In connection with this work, tests 
were made to find whether certain low streets, the surface of which had 
once been at official grade, were still settling, or whether they had come 
to rest. It was found that some of them were still sinking. Neither 
frost nor change of ground-water level has anything to do with this 
phenomenon. 

Recapitulation. 

Having determined a value of 1^ from the known extent and topo- 
graphic features of a district, and having estimated values of three 
or of five sub-areas, as explained, the following formulas may be used : 

a A = a^ A^ -^ a., A.^ + a.^ A^ acres. 

or a A = a^ A^ -\- a^ A^ -h a, A.^ + a^ A^ -{- a^, A^ " 

^™ ^ ^""^^ cT^l ^"i -ij + 2 a^ J.2 + 3 a, ^3) minutes. 



^,n, = I r.r, Z"^ , ) minutes. 

2 



422 DISCUSSION : the sewer system of san francisco 

Mr.Grunsky. or t,^^ = 0.25 -^ (a^ A^ -\- 2 a^ A^^ 3 a^Ag-\- 4 tt^A^-i- 5 a^A^) minutes. 

For the San Francisco type of limiting rain curve: 

r & 

-t = — ^ inches per hour. 

/,„ + 60 ^ ' 
r = a I " " " 

Ji = ;-— -7 inches in one hour. 

6.14 

'S',,1 = 27.5 ah ^J t^^^ cubic feet per acre. 

8„^ = imaBsjX, " " " a 

V27.5 a 6/ 

t - ( ^'" Y 
'" ~ Vl69 a bJ 

-J _ 0.71 a & ,. , 

c^OT n~^^ cubic feet per second per acre. 

^H I + 0.4 

«, + 60 "^ "' 

, 4.36ajB 

< = — 27 '^ " " " " " 

"t I / 0.4 

«» + 60 "*" '™ 
For the parabolic type of limiting rain curve: 

^ — — -^ inches per hour. 

r = a 1 " " " 

_B = 7.75 e inches in one hour. 

^m = 1 565 ae ^ t^,^ cubic feet per acre. 

S„, = 2(}2aR^i;^^ - " u u 

''"^ (rsfs^^)" "^^^^^*''- 

t = ( ^"' ^ ' 

V202ai2/ 

d,n = ""'^ cubic feet per second per acre. 

rt — oali ^^ ^^ ^^ jj ^j jj 

Values of a^ for both types of rain curve, dependent on the value of 
h, or e, and of t^, are noted tentatively in Table 6. The velocities 
which prevail when conduits are full should be used in estimating the 
value of tj)- 

In applying the new formula, all roof areas, street pavements, paved 
yards, macadamized roads, and the like, are to be considered impervious. 
The loss of water during short-time periods from such surfaces is 
small when compared with the quantity of rain. Lawns, open ground, 
dirt roads, and the like, are to be classed as pervious. 



AMERICAN SOCIETY OF CIVIL ENGINEERS 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1128 

WALNUT LANE BRIDGE, PHILADELPHIA.* 

By George S. Webster and Henry H. Quimby, Members, Am. Soc. O. E. 



The bridge on the line of Walnut Lane, Philadelphia, over Wis- 
sahickon Creek, is notable for the great size of its main arch, for 
the novelty of some of its features, the character of its material, and 
the method of its construction. It is hoped that a description of it; 
with a discussion of the principles of its design, the reasons for its 
special features, and the lessons learned in the course of its construc- 
tion will be of interest to the Profession. 

The ravine of Wissahickon Creek, in Fairmount Park, is narrow 
and deep. It separates two populous sections of the City — German- 
town and Roxborough — which are both on high ground, and hereto- 
fore most of the travel between them has been required to make a long 
detour, all the bridges being at a low level and the approaches to them 
too heavy in grade for business hauling. 

By Ordinance of July 13th, 1905, City Councils authorized the 
construction of a high-level bridge on the line of Walnut Lane, the 
point where the ravine is narrowest and most advantageous for the 
crossing. 

The required length of structure, of course, depended on the eleva- 
tion of the deck, and as considerable grading of approaches was neces- 
sary—cutting for one and filling for the other — the rate of grade to be 
obtained on the approaches, together with the estimated money value 
of each percent, of reduced street gradient, entered into the considera- 
tion of the problem of dimensions and design. The question was some- 

* Presented at the meeting of September 15th, 1909. 



424 WALNUT LANE BRIDGE, PHILADELPHIA 

what further involved by the fact that on the Germantown side the 
approach is from three streets. Fig. 1 exhibits the original profiles 
and the elevations and grades adopted. A portion of the filling for the 
west approach was obtained from the cut of the east approach, and was 
carted across the bridge. 

The bridge deck was given an ascending grade to the east — lh% 
to the middle of the main span and 1% from there to the east abut- 
ment. This change of grade was made in order to produce a slight 
crowning and relieve the elevation, or side view of the bridge, of the 
optical illusion of a dished appearance in the deck over the arch. This 
crowning is equivalent to a local grade of 1 in 400 on each side. 

The grades adopted give the floor at the middle an elevation of 
147 ft. above ordinary water level in the creek. As the depth of the 
structure at the crown is 14 ft., the soffit is 133 ft. above the creek and 
about 103 ft. above the drive, giving an imposing and majestic sweep 
of arch that is very impressive. 

As the two springing points are at the same elevation, the grade 
over the bridge, in making the spandrel piers on one side higher and 
therefore thicker, puts extra dead load on that side, and this is 
counterbalanced by making the walls of the spandrel piers on the 
other side 4J in. thicker. 

In considering the matter of type of structure and length of spans, 
it was regarded as very undesirable, for aesthetic reasons, that there 
should be any pier or column interposed between the creek and the 
pleasure drive which winds along the west bank at an elevation of 
about 30 ft. above the water. The situation clearly indicated an arch 
as the only suitable form of support for the roadway, and the necessity 
of clearing both the drive and the creek, while making reasonable pro- 
vision for the future widening of the drive, taken in consideration with 
the profile of the rocky bluffs, fixed the location of the abutment piers 
and the span of the main arch, which was made 233 ft. between the 
vertical faces of the piers. 

The extraordinary length of the main span gave to the structure a 
monumental character, and it became important that the remainder of 
the design and the material of which the bridge was to be constructed 
should be in accord with it. 

The lowest first cost would have been found in a steel viaduct with 
a wooden floor, but the character of the surroundings — natural park 



WALNUT LANE BRIDGE, PHILADELPHIA 



425 



,^ t^ t-* »-• l-i to 



o o o 




426 WALNUT LANE BRIDGE, PHILADELPHIA 

scenery of rocky and wooded slopes — would cause such a structure to 
appear inappropriate and inadequate, and the possibility of future 
neglect of proper maintenance enhanced the importance of selecting 
a construction of the most permftnent character. 

The development of the sphere of usefulness of concrete, and the 
city's extended experience in the use of it for arches, walls, and floors, 
with new and improved methods of manipulation and construction, 
whereby more nearly uniform character and value are obtained, and 
the adequacy and permanence of the material are assured, together 
with the great economy obtained through its use, influenced the adop- 
tion of concrete as the structural material of the bridge throughout, 
and it was designed so as to require steel reinforcement at only a few 
points. 

The most economical length of the approach spans was determined 
by trial to be 60 ft. from center to center of the piers, or 53 ft. clear. 
The conditions called for three of these spans on the west slope and 
two on the east slope of the ravine. 

The profile of the rocky banks naturally fixed the location of the 
springing points of the main arch, and gave a graceful proportion 
between the clear height above the creek and the length of the span. 

The type of structure selected is that of the stone arch bridge in 
Luxemburg, Europe, built in 1904, which is substantially two bridges 
side by side with the space between them floored over. The type is 
economical for such a structure, because large masonry arches of a 
width sufficient for their own independent stability have more than 
ample strength for any highway live load that may have to be carried, 
and any increase in width or strength would be utilized only in sus- 
taining the weight of the added material. Thus, in the main span of 
this bridge, the arch ribs themselves constitute one-half of the whole 
load to be carried — both dead and live — and, as the live load is less 
than one-twelfth of the dead load, a single barrel of the full width of 
the floor would clearly be wasteful because its thickness could not 
safely be made much less than that of the narrow ribs. The width of 
the floor is 60 ft., while that of each of the two ribs is 18 ft. — giving 
together 36 ft. of supporting arch, or only 60% of the floor width. 
The space between the ribs is 16 ft., leaving 4 ft. of floor to overhang 
on each side. 

Each of the two halves or ribs of the bridge carries two vertical 



PLATE XIV. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXV, No. 1128. 

WEBSTER AND QUIMBY ON 

V^^ALNUT LANE BRIDGE. 




Fig. 1 —Walnut Lane Bridge over Wissahickon Creek, Philadelphia. 




Fig. 2. — North Side, Walnut Lane Bridse. 



WALNUT LANE BRIDGE, PHILADELPHIA 427 

walls, 2 ft. 6 in. thick, which extend up to and carry the floor. These 
walls are set in 9 in. from the edges of the ribs, giving an actual over- 
hang of 4 ft. 9 in. to the sidewalk floors on the outside of the bridge. 

The three aisles thus formed, and spanned by the floor, are 20 ft. 
and 14 ft. for the middle and side aisles, respectively, from center to 
center of the walls. 

These longitudinal walls are full spandrels over the 53-ft. arches, 
but on the main arch they are pierced by four 20-ft. arches over each 
haunch, leaving 60 ft. blank over the crown. These 20-ft. openings 
are flanged 6 in., in order to give the barrel or rib of each arch a width 
of 3 ft. 6 in., and pilasters of equal relief, 2 ft. 10 in. wide, were made 
between the arches to afford angles for concealing the vertical joints 
in the wall. In the 53-ft. arches the flanging of the walls is 9 in., 
making the ribs 4 ft. wide, but a sheeting of concrete, 18 in. thick, 
connects the ribs into pairs constituting the barrels, 18 ft. wide. 

For architectural effect, the abutment piers at the ends of the main 
arch are widened to project 3 ft. beyond the general line of the struc- 
ture, and they provide for refuge bays in the sidewalks of the deck. 
The twin shafts of these piers are hollow, 6 by 10 ft., giving a thick- 
ness of wall of 3 ft., which is increased below the water-tables by the 
pilasters under the arches. They have a floored chamber under the 
sidewalk, in order to give access to the electrical conduits and the 
bases of the light poles. The 10-in. vertical wrought-iron drain pipes 
from the grate-top inlets in the driveway gutters are secured to the 
inside wall of these shafts, and pass the drainage out through 16-in. 
cast-iron pipes built into the body of the pier and outletting on the 
rocky banks of the creek. 

The floor structure consists of transverse rolled-steel I-beams, 6 ft. 
from center to center, with concrete jack-arches between them. 
The beams over the 20-ft. middle aisle are 20-in. 65-lb. per ft., and 
those over the 14-ft. side aisles are 15-in. 42-lb. per ft. The size of the 
floor members was determined by the assumed concentrated loads 
adopted as the standard for highway bridges in the City of Phila- 
delphia, which consist of a truck, weighing 40 tons, on four wheels 
spaced 20 ft. between axles, and 6 ft. gauge, or two street-railway cars 
side by side, weighing 20 tons on each truck. 

The concrete jack-arches can be depended on to distribute an axle 
load over two or more beams when the axles are more than 6 ft. apart. 



428 WALNUT LANE BRIDGE, PHILADELPHIA 

for a beam cannot deflect under its load without throwing some load 
on its immediate neighbors. The floor concrete is filled monolithically 
to 3 in. above the tops o£ the floor-beams. If the spring of the jack- 
arch be taken at the lower fillet of the beams, and a centroid of pressure 
be assumed li in. above the tops of the beams, there is constituted for 
distributing purposes an arch of 12-ft. span, having a rise equal to the 
depth of the beam, and an effective thickness of 3 in. at the crown. 
A further assumption is made that the embedded beam will distribute 
a concentrated wheel load of 10 tons laterally over a space equal to the 
gauge of the wheels — 6 ft. Each transverse foot of floor arch, there- 
fore, is required to sustain 3 333 lb. of live load, and if one-half of this 
be distributed to the adjacent beams, the horizontal force developed 
in the distributing arch will give, in the worst case — that of the 15-in. 
side aisle beams — a unit stress of only 112 lb. per sq. in. Thus it 
appears unnecessary to compute the beams for more than one-half 
of an axle load when the axles are 12 ft. or more apart, but the 
standard weight of the beams used gives value equivalent to five- 
eighths of the 20-ton axle load at a fiber stress of 15 000 lb., without 
allowance for the reinforcing value of the encasing concrete. As it 
is found by experiment, confirming calculations based on conservative 
assumptions of adhesion of the concrete to the steel, that the embed- 
ment of rolled beams in concrete adds one-third to their carrying 
capacity, the margin possessed by the floor to carry heavier live loads 
than those specified is so great that the live load may be doubled with- 
out stressing the steel beyond common practice. The construction 
proves to be very rigi.d under rolling loads. 

Careful estimates were made of the relative cost of floor construc- 
tion of this type and of reinforced concrete — concrete beams and slabs 
reinforced with steel bars. The latter appeared to be the lower in cost 
by about $4.60 per panel, or less than $1 per lin. ft. of bridge, but at 
that time experience with construction of that class, for the severe 
service of heavy street traffic, was limited, and it was felt that the steel 
beams with independent strength were less of an experiment, and 
besides, the beams would furnish support for the forms of the floor 
arches. As these forms were required to be carried high in the air, 
the support afforded by the beams, and consequently the lesser cost of 
scaffolding, doubtless quite offset the higher cost of the material. 

The 15-in. beams and the 20-in. beams in line with them were 





,1 . 




Roxborough End 



DATA: 
Main span, 233 ft. clear. 
Rise of main arch, 70 ft. 6 in. 
Approach spans, 53 ft. clear. 
Spandrel arches, 80 ft. clear. 
Total length of bridge. 580 ft. 
HeiKht above cccek, U7 ft. 



CONCRETE BRIDGE 

ON LINE OF 

WALNUT LANE 

OVER THE 

WISSAHICKON CREEK 




WALNUT LANE BRIDGE, PHILADELPHIA 439 

firmly connected by heavy angle irons with four tI'^^^- turned bolts, in 
order to constitute a tie between the two ribs of the arch, and prevent 
any possibility of a movement apart and split of the floor. 

All the beams are completely encased in concrete, which is 3 in. 
thick over the top surface and around the bottom flanges. At the 
latter point, i^-in. wire loops, 3 in. apart, are used to prevent the 
concrete from loosening. 

The floor over the middle aisle is depressed to permit the use of 
wooden ties and cinder ballast for two street-railway tracks, and to 
afford also, if it should ever be required, depth for an underground 
trolley conduit. 

The whole of the driveway floor surface below the paving is water- 
proofed with a coating of coal-tar roofing pitch, from J to i in. 
thick. 

The depression is filled with cinders, over which is laid the 6-in. 
concrete paving base. The driveway, 40 ft. wide between curbs, is 
paved with asphalt, and the two sidewalks, 8 ft. 3 in. wide, are grano- 
lithic with steel-edged curbs. 

The 2^-it. longitudinal walls are connected at intervals over the 
piers and ribs of the approach arches, and over the ribs of the main 
arch, by 24-in. cross-walls, which serve as braces for the walls and as 
diaphragms or transverse bracing to transmit to the arch ribs the wind 
.forces on the floor. The floor is separated by joints into sections to 
localize the contraction, and the interruption of its continuity prevents 
its action as a horizontal beam to transmit the wind stress to the main 
piers. Each section, therefore, is required to transmit its horizontal 
forces to the arch ribs. To provide the requisite stability in the long 
slender ribs of the main arch, they are flared 1 ft. 9 in. on each face, 
from the crown to the spring, increasing the width from 18 ft. at the 
crown to 21 ft. 6 in. at the spring. This flaring corresponds to a verti- 
cal batter on each face of 1 in 40 in a projected cross-section, and the 
edges of the ribs, therefore, are in planes parallel to and almost coin- 
ciding with the battered faces of the piers. It also causes a slight but 
not noticeable warp in the face surface, the ribs being made rectangu- 
lar in section at every point in a radial direction. The thickness of 
the ring is 9 ft. 6 in. at the spring and 5 ft. 6 in. at the crown. The 
ribs of the approach spans are not flared — the width being 18 ft. 
throughout — and their faces are plumb. 



430 WALNUT LANE BRIDGE, PHILADELPHIA 

The transverse walls do not extend over the space between the ribs, 
the latter being independent of each other, but the shafts of the piers 
at the ends of the main arch are connected by a transverse arch of 12 
ft. clear span and 10 ft. wide, having embedded in it twelve 1-in. 
square steel bars diagonally crossed to act as transverse bracing for 
the shafts. The transverse walls on the ribs are 2 ft. thick throughout 
the bridge, except those over the west half of the main arch where they 
were increased in thickness to 2 ft. 4J in., in order to provide sufficient 
weight to make the dead load on the arch symmetrical by counter- 
balancing the heavier load on the east half caused by the greater height 
of the spandrel piers due to the ascending grade over the bridge. 

The abutments were especially designed to economize material and 
at the same time avoid too utilitarian an appearance. The thrust of 
the end arch ribs is utilized to balance the earth pressure; and the 
curvature of the wings, besides embellishing the structure, has the 
effect of broadening the foundation and thus giving the equivalent of 
a much thicker wall to resist overturning. This effect is increased by 
the construction of a buttress, 4 ft. thick, at the extreme end of the 
wing in a direction radial to the curve, or at an angle of 45° with the 
axis of the bridge. At the junction of the wings and the front wall, 
steel rods are embedded for the double purpose of giving cantilever 
value to the projecting wings, and preventing vertical cracks in the 
sharp re-entrant angles of the corners. Some of these rods are con- 
tinued in a straight line past the middle of the curve of the wing and 
within 6 in. of the surface, in order to give, if it should be required, 
beam value to the wing between the abutment wall and the buttress at 
the end of the wing. 

The east abutment is 64 ft. high, from foundation to coping, about 
38 ft. of this being above the natural surface of the ground in front 
of it. The south wing of this abutment is built around a 30-in. water 
main, a hole 6 ft. in diameter being made in the wing to give free 
access to all sides of the pipe for repairs. The foundations of all the 
piers were required to be constructed over this pipe, and in each case 
an arch was formed over it to give 18 in. clear space around it, steel 
rods being embedded over the openings to resist any tendency to crack. 
The foundation of Pier 4 was extended because the pipe arch was near 
the end of the pier. 

All piers, abutments, and wings, were founded on solid rock, the 



PLATE XVI. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXV, No. 1128. 

WEBSTER AND QUIMBY ON 

WALNUT LANE BRIDGE. 





WALNUT LANE BRIDGE, PHILADELPHIA 431 

surface of which was found to be extremely rugged, requiring no 
stepping. In the foundations of each of the main arch piers two 
drill holes were sunk 8 ft. below the bottom to make sure of the 
solidity of the rock, and no crevices were found. The rock is a 
mica schist. 

The open character of the spandrels on the main arch, with their 
numerous vertical joints, presented a condition quite different from 
that commonly found with solid spandrels of rigid walls and earth 
filling, where the load is distributed and the tendency to buck up is 
met with concentrated resistance. Here, the loads, except immediately 
over the crown, are all concentrated, and there is no stiffness in either 
walls or floor to concentrate resistance to buckling. Each panel delivers 
its own individual load, dead and live, at the same point always, and 
resists to only the same amount. The long ring rib, therefore, acts as a 
column, square-ended, throated in the middle, subject to varying eccen- 
tric loads, and made up of blocks accurately fitted together but not 
adhering, and, consequently, it is dependent for its stability entirely 
on its sufficiency of thickness to accommodate the varying center of 
pressure. Therefore, it became specially important to determine the 
extreme eccentricity of the pressure line by trying every possible condi- 
tion of loading, both concentrated and distributed. The effects of six 
different conditions are shown by the stress diagram, Plate XXVIII. 
The shape adopted for the arch is an intermediate between the extremes 
of eccentricity, coinciding very nearly with the line of dead or con- 
stant loading, with a slight arbitrary allowance at the spring for 
possible initial stress in the intrados from any shrinkage or settlement 
which might take place after keying, or from being keyed at a tempera- 
ture above mean, as seemed likely to be the case. This shape was 
obtained by compounding segmental curves on three centers for each 
surface, as shown on Plates XVII and XXVIII. 

The arch was made square-ended, and continuous at the crown, in 
preference to hinging it. While pivotal bearings at the three points 
would have eliminated temperature stresses and reduced the likelihood 
of initial stresses from falsework settlement, the cost would have been 
greater, and far less graceful lines would have resulted. The line of 
center pressure, if confined to pivots at springs and crown, would 
depart far from the mid-line of the ring at the quarters and require 
considerably greater thickness there than is needed with the flat-end 



432 WALNUT LANE BRIDGE, PHILADELPHIA 

bearing. The continuous taper in thickness as well as width of rib, 
from the abutments to the crown, is an important element of beauty 
in the design. In further pursuit of lightness in the appearance of 
the ring, a simple moulding was constructed on the outer face of each 
rib, tapering in width in proportion to the thickness of the ring. 

The span of the arch, from center to center of bearings on the 
skewbacks, is 240 ft., and the rise of the mid-line is 70.5 ft. 

The maximum fiber stress appears to be at the spring, where there 
is the greatest theoretical eccentricity, because of the arbitrary allow- 
ance made for initial stress. This maximum is 380 lb. per sq. in., and 
the average at the same point is 240 lb. 

The temperature stresses were computed in two ways, the results 
corresponding fairly. One method was by using the ordinary deflec- 
tion formula for beams, and the other was by calculating the increase 
and decrease in length of extrados and intrados on the curvature 
resulting from the change in length of the ring and its confinement 
at the ends, the changes in length of fiber representing stress assumed 
to be in proportion to a modulus of elasticity of 2 000 000 in both com- 
pression and tension. Computations were based on a change of the 
general temperature of the arch of 40° Fahr., either way from the 
temperature at keying, which will change the length of the arch 1 in. 
and cause a change of 1 in. in the elevation of the crown. A rise 
increases the curvature at the crown, shortening its radius, and — 
because of the fixed ends — decreases the curvature of the haunches, 
lengthening the radius. This produces compression of the intrados 
at the crown— maximum at the