tions than those that are still in operation in London. What

has been the practical issue of this prolonged and wide

experience 1 Every medical witness that has appeared before

Digitized by

Google

90 WATER SUPPLIES

us, whether his general feeling was favourable or unfavour-

able to the water, has told us unhesitatingly that he knows

of no single instance in which the consumption of this water

has caused disease. This is the unanimous testimony of the

medical officers of health, of the water analysts, and of the

bacteriological experts, — of all, in short, whose attention has of

necessity been directed to the subject." The Commissioners

therefore think that the risk of disseminating disease, even

by admittedly polluted river water, is, under conditions

similar to those which obtain in the Lea and the Thames,

and where the water is equally carefully collected and

filtered, so small as to be negligible. The serious outbreaks

of typhoid fever in the Tees valley in 1890-91, which were

investigated by Dr. Barry, a Local Government Board inspector

of great experience, were attributed by him to the pollution

of the river Tees by sewage. The Medical Officer to the

Local Government Board, in his introduction to this Report,

says, " Seldom, if ever, has the fouling of water intended for

human consumption, so gross or so persistently maintained,

come within the cognisance of the medical department, and

seldom, if ever, has the proof of the relation of the use of

water so befouled to wholesale occurrence of enteric fever

been more obvious and patent." These outbreaks were

carefully considered by the Metropolitan Commissioners, and

they concluded that Dr. Barry's evidence connecting them

with the polluted Tees water was not conclusive.

Amidst such a conflict of opinion it is safest to suspend

one's judgment ; but even the most ardent advocate of the

use of river water will admit that it should receive as little

sewage as possible, and that the sewage should be previously

subjected to the most effective system of purification. Storage

reservoirs also should be provided, sufficiently large to allow

of the average daily supply being furnished without taking

in any part of the flood-water, and the filters should be kept

in a thoroughly efficient condition. That the neglect to

maintain these conditions might result in an outbreak of

Digitized by

Google

RIVER WATER 91

typhoid fever or cholera seems possible if not even probable,

and the fact that a town using polluted water has remained

free from such epidemics for a series of years is no proof

that such immunity will be perpetual. In the section

treating of "Diseases disseminated by Potable Waters"

many examples will be quoted in which polluted river water

has been proved, so far as actual proof is possible, to have

been the cause of serious outbreaks of both typhoid fever and

cholera.

The amount of water which can be taken from a river for

supplying a town varies according to (a) the area of the

watershed, (b) the topography and geological character of

the ground, (c) the average rainfall, and the rainfall during

a consecutive series of dry years, ((f) the distribution of the

rainfall throughout the year, (e) the amount of water which

must be supplied for " compensation " purposes, and (/) the

facilities for obtaining storage.

The available watershed, of course, includes only that

portion of the whole watershed which feeds the river above

the point at which the water will be abstracted. This can

only be ascertained by actual measurement, though approxi-

mate estimates may be made from hydrographical maps on

which the river basins are defined.

The contour of the ground surface also affects the supply,

for upon this depends greatly the rapidity with which the

rainfall, especially when heavy, will flow over the surface into

the stream. The character of the surface and of the subsoil

will also affect the amount which will flow directly into the

river, and the amount which will percolate and pass into the

river at a lower level. All the above also will be factors in

determining the amount of evaporation, or, in other words,

of determining the available rainfall. The surface drainage

area does not always correspond with the true drainage area,

since there may be springs within the surface area fed from a

source without that area ; and, on the other hand, rain which

falls on the surface area may pass by underground channels

Digitized by

Google

92 WATER SUPPLIES

beyond the limits of the watershed. All these possibilities

have to be borne in mind, and the locality carefully examined

to ascertain whether such conditions exist, and to what extent

they will affect the water supply.

The way in which the rainfall in any particular district

can be ascertained has already been described. The minimum

rainfall for a year, or a series of years, can only be determined

from records continuously taken for many years ; but it is

found that, under ordinary circumstances, the maximum

rainfall exceeds the average by one-third, whilst the minimum

falls short of the average by the same amount. The mean

rainfall during the three driest consecutive years is usually

about one -fifth less than the average. Thus, where the

average rainfall for a series of years is 30 inches per annum,

the maximum will be about 40 inches, the minimum 20 inches,

and the mean for the three driest consecutive years 24 inches.

Where careful daily gaugings of a stream have been made for

a few years, the proportion of the rainfall finding its way into

it can be ascertained, and by calculation the amount which

would pass into the river, with the minimum rainfall, can be

approximately determined. The following table, compiled

from the 22nd Annual Report of the State Board of Health of

Massachwetts, shows the rainfall received and collected during

a series of years on the Sudbury River watershed.

During the sixteen years, 1875-90 inclusive, the average

rainfall was 45*8 inches. The calculated maximum rainfall

on this area is 61*1 inches, and the minimum 30*5 inches.

The observed maximum and minimum were 57*9 and 32*8

inches respectively. The mean rainfall for the three driest

years (1882-84) was 38*8, whilst the calculated mean is 36*6

inches, so that doubtless the calculated amounts will closely

approximate to the truth when the records for a much longer

period of years are available. The percentage of rainfall col-

lected does not vary directly with the rainfall, and neither

the smallest nor largest proportion collected corresponded

with the lowest and highest rainfalls ; but the results do not

Digitized by

Google

RIVER WATER

93

vary to such an extent as to render it difficult to determine

approximately the minimum amount available. The cause

of this variation is due in part to the seasonal variation in

the rainfall, and in part to the variation in the amount

evaporated.

Year.

1875

Rainfall.

Raiufall collected.

Per cent collected.

45*49

20*42

44-9

1876

4956

23-91

48-2

1877

4402

25*49

57*9

1878

5793

30-49

52-6

1879

4142

1877

45-3

1880

38-18

12-18

31-9

1881

4417

20-56

46-6

1882

39-39

18-10

45-9

1883

32-78

11-19

341

1884

47*13

23-78

30-5

1885

43*54

18*92

43-4

1886

4606

22*82

49-5

1887

42-70

24-23

56-7

1888

57*46

35-75

62-2

1889

49-95

29-06

58*2

1890

53-00

27-00

50-9

Mean for 16 yrs.

4580

22-67

49-5

A knowledge of the seasonal rainfall and the seasonal

variation in the flow of the stream is also absolutely neces-

sary, since upon these factors depend in a great measure the

amount of storage which will be required to collect the

water during periods of abundance for use during periods of

drought. During the sixteen years' records of the Sudbury

River, the mean daily flow during the month when the river

was lowest was only 60 gallons per acre of the watershed ;

during the driest three months it was 148 gallons ; during

the driest twelve months, 777 gallons ; whilst the mean daily

flow for the whole period was 1686 gallons. From these

records the reporters to the Massachusetts State Board of

Health have calculated a table showing the "amount "of

storage necessary to make available different quantities of

water per day from each square mile of watershed, where the

Digitized by

Google

94

WATER SUPPLIES

conditions are similar to those which exist at Sudbury River."

To obtain 100,000 gallons per day per square mile, the storage

reservoir must be capable of holding 2,200,000 gallons per

square mile of watershed; to obtain 1,000,000 gallons per

day, the reservoir must hold 540,000,000 gallons. For

intermediate quantities the original table must be consulted. 1

Of course these results can only be used where the conditions

which obtain resemble somewhat those of the watershed

under consideration. The following table for the river

Thames is calculated from data given in the Report of the

Royal Commission on Metropolitan Water Supply : —

Year.

Rainfall.

Rainfall collected.

Per cent collected.

1883

28-4

13-3

46-8

1884

22*9

70

30*8

1885

29-15

8-3

28-5

1886

31-1

11-1

35-7

1887

213

8-2

38-5

1888

28-45

8-9

31-3

1889

25-6

9-1

35-5

1890

22-8

5-7

250

1891

333

9-8

29-3

Average of 9 yrs.

27*0

9-05

335

If this table be compared with the corresponding one for the

Sudbury River, it is evident that a considerably larger pro-

portion of the rainfall is available from the watershed of the

Sudbury than from that of the Thames.

Mr. Beardmore calculates that during summer, the Thames,

Severn, Loddon, Medway, and Nene, which flow over a variety

of geological strata, only carry off less than one-eighth of the

rainfall, whilst the Mimram and Wandle, which arise in and

flow through chalk districts only, yield nearly half the total

rainfall. Certain rivers are much more constant in their flow

than others, the result depending chiefly upon the conforma-

tion of the watershed and the character of the subsoil. If

the stream be fed chiefly with surface water the variation

1 State Report 1890, page 342.

Digitized by

Google

RIVER WATER 95

will be very considerable, whilst if fed chiefly from the sub-

soil the flow will be comparatively uniform. All these factors,

therefore, have to be taken into consideration when estimating

the available supply and the amount of storage necessary.

Where there are riparian owners having a right to the use

of the water for any purpose, as for manufacturing, or as a

motive power, further complications are introduced. Sufficient

water must be allowed to pass down the river to satisfy all

their reasonable requirements. Only the amount in excess of

this can be appropriated, and as during seasons of drought

they may require the whole flow of the river, the impounding

reservoirs must be large enough to store water during seasons

of abundance sufficient to tide over these periods when none

can be collected.

The quantity of water which must be stored to equalise

the supply during the longest period of drought which may

possibly occur can only be determined when the average daily

demand is approximately known, and the whole of the con-

ditions above referred to have been carefully investigated.

The number of days* storage required varies in this country

from 120 to 300; the smaller quantity only being required

on the western side, where the rainfall is heavy and the

number of rainy days considerably above the average. In

the eastern counties, where exactly the opposite conditions

obtain, about ten months' storage may be necessary.

The amount of storage required may be calculated from

the rainfall statistics only, or from the stream gaugings, but

both must be considered if the result is to be reliable. The

gaugings may be effected by various methods : (a) by means

of sluices ; (b) by aid of current metres ; (c) by means of

weirs ; (d) by gauging the surface velocity. Where a rough

approximation only is desired, a straight portion of the stream

may be selected which is tolerably uniform in width and

section, and where the water flows smoothly, or where by a

little labour such uniformity may be produced. By plumbing

the depth at different points across the stream and measuring

Digitized by

Google

96

WATER SUPPLIES

the width, the cross section can easily be calculated. The

length of the selected portion, 20 yards or more, must be

marked off, and the time noted which it takes a chip or float

to traverse this length in mid-stream on a calm day. The

mean velocity of the whole body of the water may be taken

as 75 that of the surface velocity. These data are sufficient

to give the volume required.

For example, the area of a section of a stream is found to

be 45 square feet, and the time taken by a float in traversing

a distance of 60 feet is 80 seconds. Required the flow in

gallons per day.

45 x 60 x 75

80

= 25 "3125= flow in cubic feet per second.

25*3125 x 60 x 60 x 24 x 6*25 = 13,668,750 gallons per 24 hours.

The ratio of the mean to the surface velocity is not a constant,

and its value is variously estimated by engineers from the

results of actual experiments. It varies with the rapidity of

flow, the nature of the channel, depth of water, or form of

cross section, but the first named is probably by far the

most important factor. Mr. Beardmore adopts the formula

U = V + 2*5 - \/5V where U equals the mean, and V the surface

velocity per minute. This formula gives the following values

for U :—

Surface Velocity

Value of IT

in Feet

per Minute.

Mean Velocity.

in terms of V.

5

2-5

•5

10

5 5

•55

20

12-5

•625

50

36*5

•73

100

80*2

•802

200

170*9

•855

Where greater accuracy is required and the stream is large, a

current metre may be employed.

" Having fixed on the station where the cross section of a

Digitized by

Google

RIVER WATER 97

large river is to be taken and the velocities ascertained, take

a number of soundings across the stream, at 8, 10, or 12

points, according to the breadth. These lines of sounding

divide the section into a number of trapezia, and the area of

each of these is to be calculated. Then, at a point half-way

between each of the two lines of sounding, is to be fixed a

small boat containing the current metre (Fig. 10), by means

of which 5, 6, or 7 velocities are to be determined in the same

vertical line. The arithmetical mean of these is then to be

Fio. 10.

multiplied by the area of the trapezium to which they apply.

The sum of these products is evidently the discharge of the

river — it is equivalent to the total sectional area multiplied

by the mean velocity " (Hughes's " Waterworks," quoted from

D'Aubuisson's Traite d } Hydraulique a Vusage des Ingenieurs).

In artificially constructed channels of uniform cross section,

such as canals, culverts, and pipes (the two latter may be

running full, but must not be under pressure), various formulae

have been devised for estimating the flow from the fall per

mile and the hydraulic mean depth. 1 Beardmore's modification

of Eytelwein's formula is the one usually employed —

1 The hydraulic mean depth is the sectional area of the water divided

H

Digitized by

Google

9 8 WATER SUPPLIES

U = 55V2HH,

where U equals the mean velocity in feet per minute, R the

hydraulic mean depth, and H the fall in feet per mile.

Example. — In a circular channel of 2*5 feet diameter,

having a fall of 5 feet per mile, and running exactly half full

of water, what is the flow in cubic feet per minute ?

R=|. H = 5.U = 55V2x5xg = 137'5.

2*5 2 x *785

The area of a section of the water is " - « ~ = 2*453 feet.

This, multiplied by the velocity, 137*5, gives a yield of

337*3 cubic feet per minute.

Streams of any magnitude are usually gauged by engineers

by the aid of artificially constructed weirs. Theoretically

the velocity with which the water passes over the weir is that

which a body would acquire in falling through a distance

equal to the difference between the surface level of the water

above the weir and the surface of the weir itself. A body

falling from rest acquires at the end of one second a velocity,

g, which is approximately 32 feet per second. The mean

velocity at the end of any number of seconds, t, will be

^ = -^, the space traversed, *, in that time will be -X

and the velocity at the end of that period tg. Eliminating t,

we find that v 2 = 2sg = 2 x 32 x s, therefore

v=$\/s.

Theoretically, therefore, the velocity with which water passes

over the actual surface of the weir is eight times the square

root of the difference in level above referred to. But this is

the lowermost stratum of the water only, the strata above

having a less velocity, decreasing upwards as the square root

by the wetted perimeter. In circular pipes running full, S'lid equals the

wetted perimeter, and <f J, 785 the cross section of the water ; R therefore

equals Jrf.

Digitized by

Google

RIVER WATER 99

of the depth from the surface level The mean velocity of all

the strata will be that of the particles at § the depth of the

lowermost, therefore

Unfortunately friction has to l>e taken into account, and as

this varies with the shape of the weir, its width, etc., the

above formula has little more than theoretical interest.

Numberless experiments have been recorded and many

formulae deduced therefrom for weirs of different kinds.

Here, however, it is only necessary to refer to the one most

frequently employed, that derived from Mr. BlackwelPs

experiment made on the Kennet and Avon Canal on

the flow of water over 2 -inch planks. Let Q equal the

quantity of water flowing over the weir in cubic feet per

minute, then

Q = cw V* 3 .

Where w = the width in feet, * the depth of water in inches,

and c = a constant multiplier, found by experiment and

given in the following table (quoted from Slaggs* Water

Engineering) : —

Depth

$=1 inch

Value of

c = 3'50

>>

= 2 inches

>>

= 4 25

i»

= 3 „

>>

= 4-44

,,

= 4 „

>>

= 4-44

>»

= 5 „

»>

= 4-62

>»

= 6 „

>>

= 4-57

>>

= 7 „

,,

= 4*61

,,

= 8 „

»>

= 4*48

> >

= 9 ,,

>>

= 4*44

For depths of 3 inches and upwards c may evidently be

taken as 4 5. As an example, it is required to calculate the

flow over a weir of 5 feet in width, the level of which is 6

inches below the even surface of the water.

Since s=6, c=4*5 and w=5

Q=4*5x5x \Zg3

Q = 333 cubic feet per minute.

Digitized by

Google

ioo WATER SUPPLIES

Under certain circumstances, as where a lock gate and

sluice are available, the flow may be determined from the area

of the sluice and the vertical distance between the centre of

the sluice and the level of the water in the stream. Theoretic-

ally the velocity of the water passing through the sluice would

be 8 Js, but from friction and other causes it is always less

Fig. 11.

than this. With very small sluices of from 1 to 16 square

inches area, Poncelet and Lesbros' factor, '62 may be taken

as approximately correct. If therefore the area of the sluice

A be known, the flow per second will be —

Q = A x *62 x S\/s = approximately 5 A\fs.

If A and s be expressed in feet, Q will be the flow in cubic

feet per second.

Where the river is of considerable dimensions, and it is

Digitized by

Google

RIVER WATER 101

desired to record the variations in the flow automatically, a

tide-gauge may be used (Fig. 11).

By aid of such an instrument the rise and fall of the float is

recorded on a revolving cylinder, so that not only the extent

of the variations, but the exact time at which they occurred

is registered.

Where the amount of water to be abstracted from a river

is very small compared with its volume, of course all these

elaborate investigations are unnecessary. In such cases also,

storage will only be required to supply the town during periods

when the river is in flood and the water turbid.

In exceptional cases only can river water be abstracted

at a point sufficiently high to supply a town by gravitation.

Usually the water is pumped into storage reservoirs, from

which it flows on to the filter beds, and it may again require

to be pumped after filtration into service reservoirs at such an

elevation as to permit of the water supplying the town by

gravitation. Service pipes *may be attached to the rising

main if houses have to be supplied en route. When pumping

is going on the flow will be from the pumping station to the

houses, but when the pumping ceases the flow will be in the

contrary direction, from the service reservoir. The water

taken from the Thames and Lea for the supply of the

metropolis is all pumped into service reservoirs in order to

obtain the necessary pressure, the height to which it is lifted

being on an average about 200 feet.

Limited supplies of water can be obtained from streams

having a good fall, by aid of rams, turbines, or water-wheels,

when the place to be supplied is at too great an elevation to

be supplied directly by gravitation. These automatic pumping

machines will be described in a later section.

A large number of towns in England derive their water

supplies from rivers. In the Tees valley, Darlington, Stock-

ton, Middlesborough, and several smaller towns are supplied

from the Tees ; Durham is supplied from the Wear, Carlisle

from the Eden, Ripon from the Ure, York from the Ouse,

Digitized by

Google

102 WATER SUPPLIES

Knaresborough from the Nidd, Leeds from the Wharfe and

Washburn, Doncaster from the Don, Wakefield in part from

the Calder, Ely from the Ouse, Newark from the Trent, 1

Leamington from the Leam, Shrewsbury, Worcester, and

Tewkesbury from the Severn (Gloucester also occasionally),

Plymouth from the Mew, Sandown (Isle of Wight) from the

Yare, etc. On account of the prevalence of typhoid fever in

certain of these towns (Stockton, Darlington, Middlesborough,

York, and Newark, for example) the possibility of obtaining

water supplies from other sources is being discussed. On the

other hand, certain towns are contemplating improving their

present supplies by resorting to rivers. Cheltenham, for

example, is completing works for augmenting its present

supply by drawing from the Severn at Tewkesbury. It is

now supplied in part by private wells, of which there are

over 2000, in part by spring and surface water collected in

reservoirs belonging to the town (this water when stored

has a tendency at certain seasons of the year to acquire a

disagreeable odour from the growth of vegetable matter,

chiefly Cham), and in part by the head waters of the Chelt,

which is also impounded in a reservoir. This reservoir will

hold 100,000,000 gallons, and is usually full to overflowing

about the end of March; it then loses water pretty continuously

until November, when again the feeders exceed the draught.

100,000 gallons a day have to be turned down the Chelt as

compensation water. The closing of surface wells, and the

increasing demand for water for water-closets and for flushing

sewers, and other municipal purposes, has on several occasions

run the reservoirs so low as to cause considerable anxiety.

There is within five or six miles of the town a perennial supply

of pure water from springs, which form the head waters of

the Thames, but Parliament has refused to allow them to be

diverted for the use of the town. In 1881 powers were

obtained for bringing water from the Severn at Tewkesbury,

and for supplying that town and the villages en route. The

1 Vide Chapter IX.

Digitized by

Google

RIVER WATER 103

severe drought of last year (1893) caused these works to be

proceeded with. The Medical Officer of Health says that

the water is wonderfully good, and the volume magnificent.

That it receives the sewage of several towns along its course

is acknowledged, but that there is any evidence of this pollu-

tion at Tewkesbury is denied. Worcester has taken its supply

from the Severn for forty years, and although the filtration

is said to be far from perfect, it has suffered nothing. This

town, however, pours its sewage into the river at a point

seventeen miles above the Cheltenham intake, and a mandamus