9.
10.
Leslie,
.100
. 81
.«4
. 49
Tredgold.
1C6
125
83
42
25
16
9
4
Wood.
125
83
63
52
42
36
31
28
25
From the above table, it appears that,
according to Wood, at 4 miles per hour
a horse can pnly draw half his load at 2
miles ; at 8 miles, only a quarter, etc.
Sir John Macniel estimates his power
at 60 lbs., moved 8 miles per diem at
same velocity (Gillespie). Wood's Prac-
tical Treatise on Railroads contains an
interesting chapter on horse- power. He
made many experiments. He quotes an
interesting memorial to the House of
Commons, May 3, 1830, from the pro-
prietors of various (33) stage coaches
running out of Liverpool, employing 70^
horses. These horses traveled an aver-
age distance of 13 miles daily, at a speed
not exceeding 10 miles per hour, and the
stock had to be renewed every three
years.
Tredgold assigned 37 pounds as the
power that a horse should exert over a
distance of 10 miles in a day at a velocity
of 10 miles per hour, or one hours' work^
This was founded upon his experiments
on stage coach horses. They endured
this service only three years.
The speed of North Chicago City rail-
way cars is 6 miles per hour, including
I stoppages, and the average time of ser-
' vice is reckoned at five years for each
j horse, traveling upon selected cobble
stone pavement. Before the cobble stone
was adopted the average railway service
was four years per horse. The chief
street railways of the United States esti-
mate the railroad Hfe of their horses at
from three to five years.
May 17th, 1881, 1 had the honor of ad-
dressing you on "The Best Pavement
Digitized by
Google
PROPELLING STREET CARS.
135
for Horse Railroad Tracks." Permit me
to quote from that paper. " I recently
made the following teste of the force re-
quired to start car 110 of the North j
Chicago City Railway Co., and to keep it
in motion after it was under way, using
a Fairbanks dynamometer. The track |
has a grade of two-tenths of a foot per |
hundred. (This grade is up and down, |
changing, say, each 250 lineal feet, and is {
compensated, as the observations were j
taken upon up and down grades.) The
track was not free from sand. Between '
Chicago Avenue and North Avenue, on
Clark, Division and Clyboum Avenues,
88 tests with an average of 148 passen-
gers, weighing (estimated at 140 lbs.),
with car, 6,772 lbs. The force required
to keep the car in motion at an average
speed of five miles per hour, including
stops, averaged 109.5 lbs., or per ton,
32.3 lbs. This is on an old, worn-out
iron rail. Between Chicago Avenue and
Madison Street, on Clark, on new steel
rail, 53 tests, with an average of 20.9
passengers, gave 59.8 lbs. as the force
required to keep the car in motion. This
is an average of 15.6 lbs. per ton. The
car made 17 starts on this track with an
average of 18.7 passengers. Average
force exerted to start, 426.5 lbs. Aver- 1
age per ton, 116.5 lbs. On the first men-
tioned track, 30 tests, with an average of
18.1 passengers, -gave an average force of
487 lbs. ; average per ton, 134.6 lbs.
Recapitulated, the force exerted per
ton was, in pounds :
On good track, to start, 116.5 ;
to keep in motion, 15.6.
On bad track, to start, 134.6 ;
to keep in motion, 32.3.
These tests indicate the great loss of
power entailed by bad track, and also the
great loss in starting ; and the better the
track the greater the relative loss in
starting. On the poor track 134.6 lbs.
per ton was exerted to start, and this is
4.1 times the force required to keep the
car in motion. On good track 116.5
pounds was the force required to start,
but this is 7.1 times the force required
to keep the car in motion.
Upon the North Chicago Citry Rail-
way the average weight of car and its
load is 7,740 pounds, or in short tons,
3.87. Passengers averaged at 140 pounds.
Our track is now all good. The average
force, therefore, exerted in propelling one
car is 3.87x15.6=60.372 pounds when
the car is in motion, and 3.87x116.5=
450.855 pounds force to start. The
horses average 137.97 minutes service
per diem. One hundred and three tests
upon 17 different cars, open and close,
on various Hues, with different drivers,
made by me on different days and hours,
give the following average for the horses:
Time consumed in stopping, during
which no power is exerted by the horse,
13.22 minutes. Time from starting .until
average speed is reached, 7.88 minutes.
Now, the horses average as per above
137.97 minutes daily service.
Deducting time they are not exerting
force, 13.22 minutes daily service.
Leaves actual work, 124.75 minutes
daily service.
Of this power is exerted to maintain
motion, 116.87 minutes daily service.
And extra power is exerted during 7.88
minutes daily service.
The horse power, therefore, exerted in
propelling a North Chicago railroad car
with its average load by a team in its
average day*^ work is :
450.855 X 311.5 X 7.88
33.000
starting.
60.372 X 623 X 116.87
= 33.53 H. P.
= 133 22 H. P.
33000
maintaining motion.
Total, 166.75 H. P.
This is used during 137.97 minutes,
average per minute, 1.208 H. P. per
team ; or for each horse, 604 H. P.
Upon a poor track my previously quot-
ed experiments show that this power
would be about doubled, or 2.4 H. P.
would be used per average car. About
I of the horse power is used in starting
the car (20.1 per cent.). Mr. Angus Sin-
clair experimented upon the Third Ave-
nue Elevated Railroad, New York, and
estimated that the average pull on the
draw- bar was five times greater than it
would have been if the motion of the
train could have been continuous. See
National Car- Builder.
A. M. Wellington found, by his experi-
ments, that the initial friction in starting
trains of loaded cars was 5.47 times that
required to keep them in motion at a
speed of 10 to 15 miles per hour. See
Digitized by
Google
136
VAN NOSTEAND'S ENGINEERINU MAGAZINE.
Trans. A. S. C. E., December, 1884.
Charles E. Emory, Ph. D., found 11.8
pounds per ton of 2,000 pounds to be the
resistance on straight and level track in
New York. This is less than my average,
but his tests were probably made on a
center-bearing rail, the usual rail in New
York, and this we know offers less resist-
ance to progress, as the head is compara-
tively clean, while the step-rail head
upon which I experimented was level with
the adjoining outside pavement^ and con-
sequently covered more or less with sand
and dirt.
D. K. Clarke, in his work on tramways,
states that H. P. Holt found the resist-
ance per gross ton on straight, level track
varied from 15 to 40 lbs. ; Henry Hughes,
26 lbs. — often much more, occasionally
less ; M. Tresca, 22.4 lbs. per ton. Sub-
sequently M. Tresca removed two flanged
wheels on one side of the car, and then
found the resistance 15.25 lbs. Mr.
Clarke assumes 20 lbs. per ton, and says
at times it is 40 lbs. per ton. " An aver-
age of 30 lbs. per ton may be taken for
the calculation of the ordinary tractive
force." In his second volume he states :
" The average resistance (30 lbs. per ton),
already in the first volume adopted for
calculation, may be re-adopted, although
an occasional maximum of 60 lbs. per
ton may be reached, and, on the contrary,
a minimum of 15 lbs. per ton when the
rails are wet and clean, straight and new.
Mr. Clarke's remarks refer to grooved
rails, which offer greater resistance than
the step rail. General Gilmore estimates
this resistance at 16f lbs. per short ton
with track in average condition for United
States. Mr. C. B. Holmes, President
and Superintendent Chicago City Rail-
way, stated that his cable railway required
for ordinary operations, engines of 477
horse-power ; of that it took 389 horse-
power to move the cable and machinery.
Eighty-eight horse-power (18^ per cent.)
was used for the propulsion of 240 cars,
weighing 6,000 lbs. each, and carrying
each 5,u00 lbs. of passengers. The aver-
age speed was 9 miles per hour, or 792
feet per minute. This statement would
indicate that only ^(f^^SQl horse-power
per car was required, while my experi-
ments would give : as 3.87 (my average
load) is to 5.5 (his average load), so is
1.208 horse-power (used by me) to 1.71
horse-power required.
There must have been some mistake in
his test, for .367 horse-power =12, 111
foot-pounds. A*s his speed is 792 feet
per minute, the tractive force exerted
would be orily 15.29 pounds for 5.5 tons,
a resistance of less than 3 pounds per
ton (2.78 pounds), which is impossible
upon a step rail.
Our fellow-member, D. J. Miller, M.E.,
while employed upon the above men-
tioned cable railway, made experiments
upon the horse-power used. He found
that at an average speed of 6.85 miles
per hour or 602.8 feet per minute, 1
horse-power was required for each ton of
cable and machinery and .2 of a horse-
power for each ton of car and its passen-
gers. For my average load of 3.87 tons,
this would equal .774 horse-power, in-
stead of 1 2 horse-power as estimated by
me. Mr. Miller's .2 horse-power =6,600
foot pounds. His average speed being
602.8 feet per minute, his resistance to
traction could have been only 10.95
pounds, including starting the cars. This
is 3.94 times the resistance found by Mr.
Holmes, but nearly 30 per cent, less than
my experiments would indicate. Mr.
Miller, however, assumed the weight of
passengers, having no count of their
number, and must have overestimated
the load and experimented with the track
unusally clean. My average of 15.6
pounds per ton, agreeing so nearly with
that of M. Tresca, 15.25, as above quoted,
confirms my opinion that it cannot be
far wrong. While it is true that M.
Tresca's experiment quoted was with
only one flanged wheel upon each axle,
yet that wheel traveled in a groove, and
the resistance could not vary much from
my two flanged wheels not in a groove.
The car wheels in Chicago are 30 inches
in diameter. The horses of the North
Chicago Railway weigh about 1,100 lbs.
each. The speed" at which they travel
upon the road averages 623 feet per min-
ute or 7.08 miles per hour. Theur aver-
age horse power developed being each
.604 horse-power equals 19.93:^ foot-
pounds. Divided by 623, the distance
per minute, gives 31.99 lbs. tractive force.
Leslie's Qstimate at 7 miles per hour was
25 lbs. Wood's estimate was 36 lbs., at
the same speed.
Our horses work daily 2 hours 17.97
minutes, but seven days in the week, un-
less prevented by some unforeseen cause.
Digitized by
Google
PROPELLING STREET OARS.
137
I have neglected extra resistance caused
bj coryes, because our lines are chiefly
tangents, and it is very difficult to meas-
ure the force exerted upon curves, for it
varies greatly between 400 and 1,000 lbs.,
upon the dynamometer with the same
car and load. My tests were so unsatis-
factory upon curves that I have thought
it best to omit them entirely. Then, too,
the horse walks around the curve, and
the lessened speed in a measure offsets
the increased resistance.
The greatest exertion of force upon a
tangent during my dynamometer experi-
ments occurred in starting a loaded car.
It was 1,500 lbs. average per ton 283.5
lbs. Passing through some slush, caused
by snow thrown upon the track, it equaled
75.6 lbs. per ton.
In estimating for any independent mo-
tor to propel a street car upon the North
Chicago. Railway, I would take the maxi-
mum load and resistance. I have known
of 120 passengers upon an open car.
Averaging them at 140 lbs. each, equals
16,800 lbs. ; car, 4,800 lbs. : total, 21,600
lbs., or 10.8 short tons. Speed in start-
ing, to 623 feet per minute ; average,
311.5
10.8 tonsX311.5 feet X 283.5 lbs.
33.000 ~
28.9 horse-power required ; a small por-
tion of this power would be constantly
employed, but it must be in reserve.
With the electric or cable system no such
allowance would be required, for the
reason that this excess of power is only
needed to start the car, and my experi-
ments indicate that the car is starting
only one-seventeenth of the time, while
it requires no power one tenth of the
time. For each 17 cars upon a line,
therefore, it would be necessary to fur-
nish power to start one car and to main-
tain sixteen cars in motion, less the power
when stationary, as it is not probable,,
nor is it necessary, that all should start-
at the same instant
During my experiments the car stopped
upon an average each 1,178 lineal feet..
We stop only at street intersections, or
at the center of blocks more than 500'
feet long.
The following returns are taken from
the Sixteenth Annual Beport of the Mas-
sachusetts Board of Railroad Commis-
sioners :
Name of Railroad.
Highland
Lynn <& Boston
Metropolitan. . .
Middlesex
South Boston..
TotaL
Number
of horses
owned.
909
608
8188
601
867
6158
Number
of Miles Run.
1,670,847
1,052,296
6,046,879
1,047,411
1,470,261
11,287,196
Number of
Passengers
Carried.
10,452,441
6,864,009
84,674,185
7,099,892
9,706,299
68,196,776
Average No.
of Passengers;
per Round
Trip.
48
60
88
46
41
Average.
48.4
From the above it appears that the
stable average daily distance traveled by
the above horses equals 10.04 miles, found
by dividing total number of miles run by
total number of horses, and this by 865
A i« 1 • u o 11,287,196x2
and multiplymg by 2 ^^33^3^5 =
10.04.
The average number of passengers for
these five raUroads per round trip being
43.4 per single trip, equals 21.7. Aver-
aging them at 140 pounds, equals 3,038
Vol. XXXV.— No. 2—10
pounds. Add weight of car, 4,8()'0 pounda
equals 7,838 pounds, or 3,919 short tons,
which is in excess of my average load of
3.87 tons.
It is fair to assume that these horses
are worked to the best advantage and
that this is all that can be expected of|a
horse upon a tramway.
The stables of the North Chicago Rail-
way are located at or near one end of
each line. The horses are in excellent
condition. Their mileage could not be
Digitized by
Google
138
VAN nostrand's kjngineeeing magazine!.
increased, even if it was thought desir-
able, unless they were changed from car
to car upon the road, and this would
cause delay and inconvenience. They
now make two round trips and could not
make more without adding 50 per cent,
to the distance they now travel or chang-
ing upon the road.
FLAME CONTACT— NEW DEPARTURE IN WATER HEATING.
By THOS. FLETCHER, F.O.S.
From ** English Meohanio and World of Science.'^
It is my intention to prove to you, on
theoretical grounds, and also by experi-
mental demonstration, in such a manner
AS will admit of no possiMe doubt, that
the present accepted s^ytem of water
heating, by gaseous or other fuel, is a
very imperfect means for an end, and is,
both in theory and practice, essentially
faulty. My statements may appear bold,
but I come prepared to prove them in a
manner which I think none of you will
•question, as the matter admits of the
simplest demonstration. I will, in the
first place, boil a specified quantity of
water in a flat-bottomed vessel of copper ;
the time required to boil this you will be
able to take for yourselves, as the result
will be visible by the discharge of a
strong jet of steam from the boiler. I
will then take another copper boiler of
the same form, but with only one- half
the surface to give up its heat to the
water, and will in this vessel boil the
same quantity of water with the same
burner in a little over one-half the time,
thus about doubling the efficiency of the
burner, and increasing the effective duty
of the heating surface nearly fourfold, by
getting almost double the work from
one-hfjf the surface. The subject is a
comparatively new one, and my experi-
ments are far from complete on all points ;
but they are sufficiently so to prove my
case fully. As no doubt you are all
aware, it is not possible to obtain flame
contact with any cold, or comparatively
cold, surface. This is readily proved by
placing a vessel of water with a perfectly
flat bottom over an atmospheric gas
burner; if the eye is placed on a level
with the bottom of the vessel a clear
space will be seen between it and the
flame. I cannot show this space on a
lecture table to an audience ; but I can
prove its existence by pasting a paper
label on the bottom of one of the boilers,
and exposing this to the direct impact of
a powerful bui-ner during the time the
water is being boiled, and you will see
that it comes out perfectly clean and un-
colored. Now, it is well known that
paper becomes charred at a temperature
of about 400° F., and the fact that my
test paper is not charred proves that it
has not been exposed to this tempera-
ture, the flame being in fact extinguished
by the coohng power of the water in the
vessel. I need hardly remind you that
the speed with which convected or con-
ducted heat is absorbed by any body is
in direct ratio to the difference between
its own temperature and that of the
source of heat in absolute contact with
it ; and, therefore, as the source of the
heat taken up by the vessel is nothing
but imbumt gases, at a temperature
below 400° F., the rate of absorption
cannot, under any circumstances, be
great, and the usual practice is to com-
pensate for this inefficiency by an enor-
mous extension of surface in contact
with the water, which extension I will
prove to you is quite unnecessary. You
will see I have here a copper vessel with
a number of solid copper rods depending
from the lower surface ; each rod passes
through into the water space and is
flattened into a broad head, which gives
up its heat rapidly to the water. My
theory can be stated in a few words:
" The lower ends of the rods, not being
in close communication with the water,
can, and do attain a temperature suffi-
ciently high to admit of direct flame con-
tact, and as their efficiency, like that of
the water surface, depends on the differ-
ence between their own temperature and
that of the source of heat in absolute
Digitized by
Google
FLAME CONTACT.
139
contact with them, we must, if my theory
is correct, obtain a far greater duty from
them. I do not wish you to take any-
thing for granted ; and although the sur-
face of the rods, being vertical, can only
be calculated for evaporating power at
one-half that of a horizontal surface, as
is usual in boiler practice, my margin of
increased duty is so great that I can
afford to ignore this, and to take the
whole at what its value would be as hori-
zontal surface, and still obtain a duty 50
per cent, greater from a surface which is
the same in area as the flat-bottomed
vessel on the fireside, but having only
one-third the surface area in contact with
the water. I do not, of course, profess
to obtain more heat from the fuel than it
contains, but simply to utilise that heat
to the fullest possible extent by the use
of heating surfaces beyond compaiison
smaller than what have been considered
necessary, and to prove not only that the
heating surface can be concentrated in a
very small area, but also that its effi-
ciency can be greatly increased by pre-
venting close water contact, and so per-
mitting combustion in complete contact
with a part of the heating sufface. 1
will now boil 40 oz. of water in this flat-
bottomed copper vessel, and, as you will
see, sharp boiling begins in three minutes
fifteen seconds from the time the gaUb is
lighted. The small quantity of steam
evolved before this time is of no import-
ance, being caused partly by the air
driven oflf from the water and partly
from local boiling at the edges of the
vessel owing to imperfect circulation. On
the bottom of this vessel is pasted a
paper label, which you will see is un-
touched by the flame owing to the fact
that no fl^e can exist in contact with a
cold surface. It may be thought that,
owing to the rapid conducting power of
copper, the paper cannot get hot enough
to char. This is quite a mistake, as I
will show you by a very curious experi-
ment. I will hold a small plate of copper
in the flame for a few seconds, and will
then hold it against the paper. You will
see that, although the copper must of ne-
cessity be at a temperature not exceed-
ing that of the flame, it readily chars the
paper. We can, by a modification of
this experiment, measure the depth of
the flameless space, as the copper, if
placed against the paper before it has!
time to be previously heated, will, if not
thicker than 1-40 inch, never become hot
enough to discolor the paper, showing
that the flame and source of heat must
be below the level of a plate of metal
this thickness. In repeating this experi-
ment I must caution you to use flour
paste, not gum, which is liable to swell
and force the paper past the limit of the
flameless space, and also to allow the
paste to dry before applying the flame,
as the steam formed by the wet paste is
liable also to lift the paper away and
force it into the flame. I will now take
this vessel, which has only one-half the
surface in contact with the water, the
lower part being covered with copper
rods, 3-16 in. diameter, i in. centers
apart and H in. long, and you will see
that with the same burner as before,
under precisely the same conditions,
sharp boiling takes place in 1 minute 50
seconds, being only 13 seconds more
than half the time required to produce
the same result with the same quantity
of water as in the previous experiment.
Although the water surface in contact
with the source of heat is only one-half
that of the first vessel and the burner is
the same, we can see the difference not
only in the time required to boil the 40
oz. of water, but also in the much greater
force and volume of steam evolved when
boiling does occur. With reference to
the form and proportions of the conduct-
ing rods, these can only be obtained by
direct experiment in each case for each
distinct purpose. The conducting power
of a metallic rod is limited, and the
higher the temperature of the source of
heat, the shorter will the rods need to be,
so as to insure the free ends being below
a red heat, and so prevent oxidation and
wasting. There are also other reasons
which limit the proportions of the rods,
such as liability to choke with dirt and
difficulty of cleaning, and also risk of
mechanical injury in such cases as ordi-
nary kettles or pans — all these require-
ments need to be met by different forms
and strengths of rods to insure perma-
nent service, and, as you will see further
on, by substituting in some cases a dif-
ferent form and type of heat conductor.
To prove my theory as to the greater
efficiency of the surface of the rods in
contact with the flame as against that in
direct contact with the water, I have an-
Digitized by
Google
140
VAN NOSTRA ND' 8 EN (ilN BERING MAGAZINE.
other smaller vessel which, inelading the
rods, has the same total surface in con-
tact with the flame, but only one- third
the water surface as compared with the
first experiment. Using again the same
quantity of water and the same burner,
we get sharp boiling in 2 minutes 10
seconds, being an increase of duty of 50
per cent., with the same surface exposed
to the flame. The rods in the last ex-
periment form two-thirds of the total
heating surface, and if we take, as I
think for some careful experiments we
may safely do, one-half the length of the
rods to be at a temperature which will
admit of direct flame contact, we have
here the extraordinary result that flame
contact with one-third of the heating
surface increases the total fuel duty on
a limited area 60 per cent. This really
means that the area in contact with the
flame is something like six times as effi-
cient as the other. In laboratory ex-
periments it is necessary not only to get
your result, but to prove your result is
correct, and the proof of the theory ad-
mits of ready demonsbration in your
own laboratories, although it is unfit for
a lecture experiment, at all events in the
only form I have tested it If you will
take two ordinary metal ladles for melt-
ing lead, cover the lower pjui; of one of
these with the projecting rods or studs
and leave the other plain, you will find on
melting a specified quantity of metal in
each that the difference in duty between
the two is very small. The slight in-
crease may be fully accounted for by the
difference in the available heating sur-