,,.' "**'"" " 'j|. 
 WATER SUPPLY. 
 
 W, H, CORFIELD, ESQ., I, A,, M, D, 
 
 Y ( (OXON.) 
 
 PROFESSOR OF HYGIENE AND PUBLIC HEALTH AT UNIVER- 
 SITY COLLEGE, LONDON ; MEDICAL FELLOW OF 
 PEMBROKE COLLEGE, OXFORD; AND MEDI- 
 CAL OFFICER OF HEALTH AT ST. 
 GEORGE'S, HANOVER SQUARE. 
 
 SECOND AMERICAN EDITION. 
 
 NEW YOEK: 
 
 D. VAN N08TRAND COMPANY, 
 
 23 MURRAY AND 27 WARREN STREET. 
 
 1890. 
 
UHITBRSITY 
 
 PREFACE 
 
 TO 
 
 SECOND AMEEICAN EDITION. 
 
 The first edition of the abstract of 
 these Lectures *appe&rd in Yan Nos- 
 trand's Engineering Magazine in 1875, 
 and subsequently were incorporated in 
 the Science Series, for the reason that 
 at that time there was a demand for 
 monographs on this special subject, the 
 literature at that period being somewhat 
 scanty. 
 
 The original Lectures were delivered at 
 the School of Military Engineering, 
 Chatham, England, in the autumn ses- 
 sion of 1873, and were subsequently 
 printed for private circulation, and at- 
 tracted by their clearness and concise- 
 ness in detail, a very wide attention. 
 
The principal point in view by Prof. 
 Corfield was public health, and in this 
 respect he did not enter into engineering 
 matters to any great extent, not more so 
 than the subject absolutely required, but 
 he went very thoroughly over the whole 
 ground, treating it in a practical, com- 
 mon-sense view. For these reasons, 
 and from the fact that a very great many 
 publications bearing upon this subject 
 have appeared since these Lectures were 
 printed, it has been deemed advisable, as 
 the publishers desire to keep the differ- 
 ent numbers of this Series available to 
 the public, to reprint this particular one 
 without change or addition, the fact 
 being that comparatively little if any 
 changes or additions could add to the 
 permanent value of the matter presented 
 in these pages. The Lecture form has 
 been preserved throughout. Prof. Cor- 
 field's further Lectures on " Sewerage 
 and Sewage Utilization" form No. 18 of 
 this Series. W. H. F. 
 
 NEW YORK, December, 1889. 
 
WATER AP WATER SUPPLY. 
 
 It will be our purpose to discuss, in 
 the first place, the sources and the kind 
 of water that are required for large com- 
 munities the kind in the first place, the 
 quantity in the next, the places to get it 
 in the third, and then the ways to con- 
 vey it to the community. 
 
 It is only one part of the fuel of a 
 community that we have to consider. 
 We shall then consider what are the 
 wastes from a large community, and 
 whether, although useless for the pur- 
 pose for which the original fuel was sup- 
 plied, they can be made useful for other 
 purposes, and if so, how? Whether 
 again there is any necessity of getting 
 rid of them, and if so, how this can be 
 
,6 
 
 clone most effectually, and most cheaply, 
 rind w.itTiout; prejudice to other commu- 
 "nities. 
 
 Now then as regards water. Water is 
 required in a large community for a 
 great variety of uses. These uses were 
 divided by the Eomans, and they have 
 been divided ever since, into public and 
 private uses. The public uses are such 
 as for cleaning streets, extinguishing 
 fires, for fountains, for public baths, and 
 so on. The private uses are for drinking, 
 washing, cooking, etc. Thus water you 
 see at once from the mere examination of 
 its uses comes to the community to be 
 soiled. It comes in order that the com- 
 munity may be supplied with one of the 
 necessities of life. It comes to wash 
 communities, places, and habitations. It 
 comes, I repeat, to be soiled. It is, 
 therefore, generally, when soiled, useless 
 for the purpose it was originally wanted 
 for. It has either to be purified or got 
 rid of. A community requires pure water 
 for some purposes, and those are especial- 
 ly for drinking and cooking. Pure water 
 
I do not mean chemically pure, but we 
 shall see directly what is meant hygieni- 
 cally by pure water is not necessary for 
 every purpose, such as for washing the 
 streets, extinguishing fires, etc. How- 
 ever, practically speaking, only one kind 
 of water can, as a rule, be supplied to a 
 community, and so it becomes necessary 
 for us to know where we can get this 
 sufficient supply of water of a certain 
 quality, viz., sufficiently good for drink- 
 ing. 
 
 Now, roughly speaking, a drinking 
 water should be, in the first place, trans- 
 parent. In the second place, it should 
 be transparent to white light ; that is to 
 say, it should be transparent and with- 
 out color. It must be without taste and 
 without smell, and it must deposit no 
 sediment on standing, and have no par- 
 ticles suspended in it. Those are the 
 rough qualities of water which anybody 
 can examine for himself ; the best way to 
 look at it is to look through about a foot 
 or 18 inches of it in a long glass cylinder, 
 placed on a piece of white paper. It 
 
8 
 
 must be aerated to be fit for drinking, 
 and cool. Now, if the water you are 
 examining does not fulfill these condi- 
 tions, it must be rejected at once, or 
 brought to satisfy them. We have to 
 consider how these conditions are to be 
 fulfilled, and we ought to satisfy them 
 on a large scale. But a water may com- 
 ply with all these conditions, and yet not 
 be a safe water to drink. It may contain 
 substances which you cannot tell in* any 
 of these ways, and, practically speaking, 
 all waters do. Substances whether in 
 solution or suspension may be hurtful, or 
 they may be harmless; and now I want to 
 tell you how, if you have a chemical 
 analysis of a sample of water before you, 
 you can tell whether that water is suita- 
 ble for your purpose or not. That is a 
 thing you do not generally find in engi- 
 neering books. It is necessary for you to 
 know it, because if a report is brought 
 up upon a particular water, you ought 
 to be able to know whether that will be 
 a satisfactory water or not. 
 
 Natural waters contain dissolved (in 
 
the first place especially) carbonic acid 
 gas. They contain all the constituents 
 of air in solution, but the gases are not 
 in the proportion in which they are in 
 atmospheric air. There is often a great 
 quantity of carbonic acid gas, and oxygen, 
 being more soluble than nitrogen, is 
 generally in larger proportion than in 
 atmospheric air. Now the carbonic acid 
 gas is the one that I am going to speak 
 of first. Water containing carbonic acid 
 in solution has the property of holding 
 in solution quantities of certain salts that 
 it would not dissolve otherwise, or only 
 in much smaller quantities, and the 
 chief of these is carbonate of lime. 
 Natural waters often contain, then, in the 
 first place, salts of lime, especially the 
 carbonate, dissolved in carbonic acid. 
 They contain often the sulphates of lime, 
 soda, magnesia, iron, and so on in fact, 
 different salts of these and other bases. 
 Phosphates they all contain, and also 
 chlorides and nitrates. All natural 
 waters contain the latter in certain pro- 
 portions even rain water. Almost all 
 
10 
 
 of them contain salts of ammonia. The 
 question arises which of these may be 
 allowed in water, and which may not, or 
 which, at any rate, may not be allowed 
 above a certain quantity, and what is 
 the quantity? Beyond those simple 
 characters for pure water which I 
 gave you a few minutes ago, there is a 
 property of natural waters which can be 
 easily ascertained by any one, and which 
 constitutes one of the best known differ- 
 ences between various specimens of 
 water, and that is the quality of hardness. 
 What does that mean? Hardness is 
 tested in this way. Pure water dissolves 
 soap, which is a combination of soda 
 with some of the fatty acids. Pure water 
 dissolves soap perfectly and forms a 
 lather at once. Now water containing 
 certain salts in solution, and notably 
 salts of lime, magnesia, and iron, does 
 not do so, because these salts form in- 
 soluble precipitates with the soap. That 
 is what is meant by the water being hard. 
 If a water, instead of lathering with soap 
 immediately, takes a great deal of 
 
11 
 
 trouble to make a lather, does not do it 
 till after some time, and causes a curdy 
 precipitate, then it is a hard water. 
 That is, of course, a very rough way of 
 putting it ; but the amount of soap that 
 is required before a water will lather, 
 gives a test of the amount of salts which 
 cause the hardness of the water, and the 
 chemist takes a standard solution of 
 soap and tries how much of this solution 
 is required before he can get a lather 
 with water, and he says that the water 
 has so many degrees of hardness. What 
 is meant by a degree of hardness 1 That 
 each gallon of the water contains in so- 
 lution an amount of salts which will pre- 
 cipitate as much soap as a grain of car- 
 bonate of lime would precipitate. What 
 is the importance of this? Hard water 
 is as a general rule less wholesome than 
 soft, and often much less so, it is not so 
 good for household purposes, nor for 
 use in engines, and it entails an enormous 
 waste of soap. It is therefore objection- 
 able, even if the hardness is caused by 
 the presence of harmless salts. The 
 
12 
 
 total amount of hardness, the degree of 
 hardness of a water before anything is 
 done to it, is called the " total hardness," 
 and if the total hardness of a water is 
 greater than six degrees on what is called 
 " Clark's Scale " (the value of a degree 
 of which I have already explained) it is 
 called a hard water ; if less, it is known 
 as a soft water. Now hard water (sup- 
 posing you have only got hard water, 
 and cannot get a supply of soft water) is 
 made softer, in the first place, by boiling. 
 That can be done on a small scale. If 
 you boil hard water, of course the car- 
 bonic acid is driven off, and the salts 
 held in solution by it, especially carbon- 
 ate of lime, are precipitated. There is 
 another way of rendering hard water 
 soft, and this can be applied on a large 
 scale ; it is known as " Clark's process." 
 The carbonate of lime is held in solution 
 in the water by carbonic acid ; you can 
 precipitate it by boiling, or prevent its 
 being held in solution by causing the 
 carbonic acid to combine with something 
 else, as with more lime, and Clark's pro- 
 
13 
 
 cess, which is now used on an extensive 
 scale (and ought to be used very much 
 more than it is) consists in adding to the 
 hard water milk of lime. This milk of 
 lime combines with the excess of car- 
 bonic acid, forming carbonate of lime, 
 which falls down as precipitate together 
 with the carbonate of lime that was pre- 
 viously held in solution, thus leaving the 
 water softer. If you boil water, and 
 then determine the hardness that remains, 
 that is called the " permanent hardness''; 
 an extremely important matter. The 
 importance of it consists in this, that it 
 cannot be removed at any rate on a large 
 scale, and, in the second place, that it is 
 due to salts several of which are injuri- 
 ous, so that a large degree of permanent 
 hardness indicates a bad water. Now 
 this permanent hardness (the hardness 
 that is lost by boiling is called u tempo- 
 rary hardness") is due chiefly to the sul- 
 phate of lime and chloride of calcium, 
 and to magnesian salts. These are all 
 objectionable in a water. Let me give 
 you some examples of degrees of hard- 
 
14 
 
 ness of various specimens of water so as 
 to give you a definite idea of hard- 
 ness. 
 
 The hardness of the Thames water 
 above London is 14 degrees of Clark's 
 scale. That is a hard water. The hard- 
 ness of the New Kiver water is 15^ de- 
 grees. That, too, is a hard water. The 
 water of Bala Lake has only J of a de- 
 gree of hardness, and of course that is 
 an exceedingly soft water. I must tell 
 you before going on that it is now very 
 usual to express hardness in another way. 
 That is to say, instead of saying so many 
 grains per gallon, as is done in Clark's 
 scale, hardness is now very generally ex- 
 pressed by parts in 100,000, and I men- 
 tion this at once, because the results of 
 most of the analyses that we shall have 
 to refer to during the lectures are given 
 in parts per 100,000. Of course, if you 
 are given the hardness of water in parts 
 per 100,000, you can convert it into de- 
 gress of hardness in Clark's scale by 
 multiplying by seven and dividing by 
 ten, because Clark's scale gives the re- 
 
15 
 
 sul s in grains per gallon; a grain per 
 gallon is one part in 70,000. On this 
 new scale, as an example, the hardness 
 for the last week of last year of the five 
 Thames companies was about 20 de- 
 grees, that is to say, about 14 degrees 
 by Clark's scale. 
 
 The hardness, again, of the water sup- 
 ply which is derived from deep borings 
 in the chalk was 29.4 on this scale, or 
 20.58 of Clark's scale. Of course that 
 is a very hard water indeed. But the 
 hardness of these two waters is quite 
 different, because the permanent hard- 
 ness of Kent water is very little indeed. 
 The total hardness of that water is al- 
 most entirely due to the carbonate of 
 lime, whereas, much of the hardness of 
 the water supply to London by the 
 Thames companies is due to salts other 
 than carbonates, especially to sulphates. 
 Therefore, you get much information 
 about the quality of water by its hard- 
 ness. If you know water has a high de- 
 gree of permanent hardness, you know 
 it has a very good chance of being a bad 
 
16 
 
 water. It contains probably sulphate of 
 lime and chloride of calcium, and per- 
 haps magnesian salts. The latter are 
 especially objectionable to water, and 
 any water which gives even a small 
 amount of salts of magnesia is to be re- 
 jected. Water containing these salts 
 causes diarrhoea when drunk, and it ap- 
 pears to be from the presence of these 
 salts in drinking waters that the swelling 
 of the neck known as goitre is produced 
 in Switzerland and other countries. 
 
 The next thing to which I wish to 
 draw your attention with regard to sub- 
 stances dissolved in water, is, the amount 
 of chlorides that may be present. I may 
 say broadly, that if you see in a report 
 on the quality of a water that it contains 
 much chlorine, or much common salt 
 (chloride of sodium), you may at once 
 put it down as a suspicious water, and 
 you will see why in a minute. Where 
 do you get chlorides in a water ? They 
 may come from an infiltration from the 
 sea. They may come again from strata 
 containing a quantity of common salt. 
 
17 
 
 Bat another source of chlorides in a 
 water is pollution by sewage. All sewage 
 contains a considerable proportion of 
 common salt. This is one of the necessi- 
 ties of life, it is contained in many of 
 our foods, and in excretal matters, 
 especially in the urine, and so sewage 
 contains it. The average amount in the 
 sewage of water-closeted towns is ten 
 parts of chlorine in the 100,000. Pure 
 natural waters contain less than a grain 
 of chlorine in a gallon, or about 1 part 
 in 100,000. So, if in a sample of water 
 for which you get the analysis sent, you 
 see more than a grain in a gallon of 
 chlorides, you must at once know the 
 reason why. London drinking water 
 contains 2 parts in 100,000. That is not 
 very bad water, and as it is got from the 
 Thames we know that it has been pol- 
 luted by sewage. The water derived 
 from the chalk the Kent water actu- 
 ally contains more than that, but we 
 have a very good reason for not object- 
 ing to it on that account, inasmuch as 
 we know that it is not rendered impure 
 
18 
 
 by sewage. The well waters of London 
 mostly contain more chlorine than sew- 
 age; they are, in fact, a concentrated 
 form of sewage which has gone through 
 certain alterations. I am not here allud- 
 ing to the Artesian Wells, but only to 
 those which are supplied by the subsoil 
 water above the London clay. The 
 amount of chlorine is a very good test 
 of the purity of a water, except that 
 you must always allow for the possibil- 
 ity of chlorides being present in the soil 
 through which that water has gone. 
 
 Nitrates and nitrites are given you in 
 the Registrar General's Reports as the 
 test for what is called "pre&iou* sewage 
 contamination." What does that mean ? 
 It means that the nitrates, etc., that are 
 dissolved in water come in a great major- 
 ity of cases (if not in all) from the oxyda- 
 tion of organic matter at some time or 
 other, or in some place. Now to show 
 you how plain it is that water must not 
 be rejected merely because it contains 
 nitrates, I must tell you that there are 
 nitrates and nitrites in all waters, even 
 
19 
 
 in small quantities in rain water. What 
 amount of nitrates may be found in 
 water without giving a suspicion of pre- 
 vious contamination? Allowing that 
 they are not injurious in themselves, yet, 
 inasmuch as they at once make you sus- 
 pect that the water containing them in 
 solution has, at some time or other, been 
 contaminated with organic matters to a 
 large extent, which organic matters have 
 been oxydized, the result being the pro- 
 duction of nitrates and nitrites inas- 
 much as that is the case, if you get 
 much nitrates, etc., represented in a 
 water, you must at once see if that water 
 is derived from a source where it is likely 
 to get contaminated with refuse matters; 
 because if it is, although the nitrates 
 are harmless, and although it is very 
 desirable that these matters should be 
 oxydized to that state, still you are al- 
 ways liable to its happening some day 
 that the water is contaminated by the 
 solution of these organic matters in their 
 crude un oxydized form, in which case 
 they are very often, if not always, danger- 
 
20 
 
 ous. Let us see what amount of nitrates 
 is found in various waters. In the drink- 
 ing water we get in London from the 
 Thames there are about 2 parts in a mil- 
 lion (or 0.2 in 100,000). In the New 
 River Water (North London water) a 
 little more than 3 parts; and in the 
 Kent water 4 parts in a million, so that 
 the deep chalk waters (which we know 
 must be very pure) contain more nitrates 
 than the others do, a sufficient proof 
 that the presence of nitrates is not of 
 itself a sufficient reason for rejecting a 
 water. The waters from the Cumber- 
 land Lakes contain very much less. To 
 give you an example of a water contain- 
 ing a great deal, I may cite the instance 
 of a well at Liverpool which was found 
 to contain more than 8 parts in 100,000 
 of nitrates and nitrites, which were all 
 derived (or in all probability derived) 
 from the oxydation of sewage that had 
 traversed the ground round that well. 
 If nitrates be present in large quantities 
 it must be regarded as a suspicious cir- 
 cumstance, unless you have good reason 
 
21 
 
 to know that the water comes from a 
 source which is beyond the suspicion of 
 contamination. There are quantities of 
 nitrates in many soils. The presence of 
 nitrates in water got from such soils 
 would not justify you in having the 
 water condemned as a source of supply 
 if there were no other reason. 
 
 Salts of ammonia. These, too, are 
 contained in natural waters in exceed- 
 ingly small quantities. They do no 
 particular harm in themselves, but they 
 frequently come directly from sewage. 
 The numbers in a drinking water repre- 
 senting salts of ammonia ought to be in 
 the third place of decimals for parts in 
 100,000, or, if in the second place of dec- 
 imals, ought to be small. Now the water 
 supply of London, filtered Thames water, 
 contains .001 to .005 parts in 100,000. 
 That is pretty good. The water at Bala 
 Lake con tains .001 parts, and rain water 
 contains the same amount ; so that we 
 may expect salts of ammonia to be con- 
 tained in all natural waters. Sewage 
 contains about 6 parts in 100,000. Well 
 
22 
 
 water often contains large quantities, 
 four parts, for instance; the pump water 
 in London contains nearly one part in 
 100,000, and the water of the Thames at 
 London Bridge 0.1 part in 100,000; these 
 are all bad waters, so that when you see 
 ammonia mentioned in an analysis of 
 water in greater quantity than is repre- 
 sented on the second place of decimals in 
 parts per 100,000, you may always safely 
 condemn it, for on looking further you 
 will find what I am now going to speak 
 of, namely, organic matters. 
 
 Now the actual organic matters pres- 
 ent in a water may be in suspension or 
 solution. If there are organic matters 
 in suspension a water may be safely con- 
 demned, because they may even by agi- 
 tation pass into solution, and so the fact 
 of your trying to separate them may 
 cause more of them to get into solution. 
 Organic matters you will find in analysis 
 represented in two different ways. In 
 one, as, for instance, in the analysis given 
 by the Registrar General, you will find 
 organic matters represented in this way : 
 
23 
 
 so much organic carbon, and so much 
 organic nitrogen in the 100,000, and the 
 Eivers Pollution Commissioners have 
 given this as a standard, not of drinking 
 water, but of a water that shall be con- 
 sidered to pollute any watercourse to 
 which it is turned. Two parts of organic 
 carbon in 100,000 or 3 parts of organic 
 nitrogen in 100,000. What does the 
 London drinking water contain again? 
 From 3 to 4 in 100,000 of organic car- 
 bon and about .05 of organic nitro- 
 gen. Now we shall see at once the 
 difference between drinking water de- 
 rived from such a source as the Thames 
 and filtered, and drinking water derived 
 by boring into deep strata into the 
 chalk. The chalk water only contains 
 .06, that is the fifth of the quantity of 
 organic carbon, and .01, a fifth of the 
 quantity of organic nitrogen that the 
 water supplied by the Thames Compa- 
 nies contains ; so that when you come to 
 organic matters, you see the difference 
 at once between a water that is derived 
 from a pure source, and one from an 
 
24 
 
 impure. The other method that I have 
 to mention to you, which is used for ex- 
 pressing the amount of organic matters 
 in water, is called "Wanklyn's method/'* 
 from the chemist who discovered it. 
 This method consists in the conversion 
 of the nitrogen contained in the organic 
 matter in the water, or a considerable 
 part of it, into ammonia, and then it is 
 estimated as so much ammonia. I dare 
 say you all know that the test that 
 chemists have for ammonia is perhaps 
 the most delicate test with which we are 
 acquainted. This ammonia you will see 
 mentioned in the records of analysis as 
 "albuminoid ammonia," and to a certain 
 extent it does represent the amount of 
 organic matter in the water. This albu- 
 minoid ammonia in a drinking water must 
 not be allowed to be above the third 
 place of decimals. If it appears higher 
 than the third place of decimals in parts 
 of 100,000, if in the second place, or if 
 in the first, it is bad. If in the first 
 place it is decidedly bad water, and con- 
 
 * See Wanklyn's Water Analysis. 
 
25 
 
 tains a considerable amount of organic 
 matter in a state of solution. You may 
 consider that the albuminoid ammonia 
 represents about ten times its weight of 
 dry organic matter, and about forty times 
 its weight of moist organic matter. So 
 that .05 of albuminoid ammonia in 100,- 
 000 represents about 2 parts of moist 
 organic matter in the water. You see 
 that, when you have an analysis of water 
 before you, you must consider the differ- 
 ent things together. The nitrates help 
 to condemn a water with much organic 
 matter in it. The ammonia does the 
 same, and the chlorides especially so, 
 and chlorides are to be regarded as a 
 suspicious indication in water, if you 
 have not good reason to suppose that 
 they come from some other source than 
 the one I have indicated. The danger 
 of organic matter in drinking water con- 
 sists in this fact (of course organic mat- 
 ters are necessary to us for our food, and 
 it is not the mere fact of its being organic 
 matter that renders it dangerous,) that it 
 is organic matter in a state of rapid 
 
26 
 
 change, in a state of putrefactive 
 change; and then that it may contain 
 and often does contain (especially if it is 
 derived from excremental matter) the 
 poison of specific diseases, which may 
 be distributed in the drinking water to a 
 population and cause an outbreak of 
 cholera, typhoid fever, etc. We know 
 now what sort of water must be got for 
 drinking. The above are its character- 
 istics, and the water supply must either 
 comply with these conditions, or be 
 made to do so artificially. 
 
 Now how much of it is wanted? 
 You can look at this in two ways. You 
 can get to know by experience how 
 much bodies of men and towns always 
 have wanted. The amount, of course, 
 varies immensely with the use of baths, 
 whether they are public baths or not, 
 with the amount used for washing the 
 streets, and for manufactures, and also 
 with the amount of waste, because that 
 is a very important item. Now, for 
 washing, drinking, and domestic pur- 
 poses generally, you may put it down 
 

 27 
 
 (if there is reasonable amount of bath- 
 ing) at about ten gallons a head a day, 
 and then you must add nine or ten 
 more for flushing the sewers and washing 
 the streets. Much of this will be added 
 through the water-closets. Thus you 
 may say 20 gallons a day without waste 
 may be taken as a kind of average. 
 For trades you must allow 10 gallons 
 more as a rule. If there are public 
 baths, and where there are many ani- 
 mals, as horses, which require about 12 
 or 15 gallons a head for washing and 
 drinking, you must make a greater al- 
 lowance. You will see that about 30 
 gallons a head a day is the least, even 
 where there is no extra demand, and 
 that is about the amount provided in 
 London, and that is about the least that 
 you should aim at. Professor Kankine 
 tells you that 35 gallons is the greatest 
 amount necessary. However, they don't 
 think so everywhere. New York man- 
 ages to get through 300 gallons, and 
 does not find it too much. In ancient 
 Rome (to show you that these matters 
 
28 
 
 have been thought of a long time ago) 
 they had nine aqueducts to bring water 
 to the city. They thought it of so much 
 importance that several of these aque- 
 ducts were from 42 to 49 miles long, 
 and one of them, the Marcian, was 54 
 miles long. Frontinus, who was the 
 superintendent, and who wrote a most 
 excellent work about them, giving accu- 
 rate descriptions and measurements of 
 them, tells us the two most recent were 
 made because the seven already in exist- 
 ence " seemed scarcely sufficient for pub- 
 lic purposes and private amusements." 
 Now the sectional area of the water 
 supply to Rome by these aqueducts was 
 1,120 square feet, and it is pretty sure 
 that there were not more than 332,000,- 
 000 gallons daily brought to Rome by 
 them. I suppose there were not more 
 than a million people ; that gives you 
 about 332 gallons a day that they found 
 necessary.* 
 
 Now, let me give you one or two 
 
 * Mr. James Parker on the " Water Supply of 
 Rome." 
 
29 
 
 points about the measurement of water 
 that you will find useful. The measure- 
 ment of water you will often find given 
 in cubic meters. A cubic meter is 35J 
 cubic feet, or 220 gallons. T % hat is to 
 say, a cubic meter of water is 220 gal- 
 lons, and as a ton of water contains 224 
 gallons, a cubic meter of water is almost 
 exactly equal to a ton by weight (or tun 
 by measure). A cubic foot is rather 
 more than 6 gallons, and 100 gallons 
 are just about 16 cubic feet. Let me 
 just give you an example of this. Lon- 
 don, during December, 1872, was sup- 
 plied daily with 100 millions, nine hun- 
 dred thousand, and something odd, gal- 
 lons of water. That is to say 458,577 
 cubic meters, or about the same amount 
 of tons by weight or tuns by measure ; 
 that is, 201.8 gallons to each house, or 
 father less than a cubic meter to each 
 house, and 28.4 gallons to each person. 
 I told you it was 30. Well, it varies a 
 little. It is a little under 30 very often. 
 Of the total amount of water supplied 
 to a place, you may take it as a general . 
 
30 
 
 rule that 80 or 82 per cent, is required 
 for domestic purpose^, so that during 
 that month of December in London 
 there were about 23 gallons used for 
 domestic purposes. Hence the conclu- 
 sion about the quantity is, that the 
 least you must endeavor to get is 30 
 gallons a head a day without any very 
 extra demands. Of this about 80 per 
 cent, \vill be required for domestic and 
 the rest for public purposes. 
 
 So much for the quality of drinking 
 water, and the quantity to be supplied. 
 We have now to go on to consider the 
 places where water of this quality and 
 in sufficient quantity can be procured. 
 The main sources of water are rain, and 
 the sources that are subordinate to rain- 
 fallwells, springs, streams and rivers. 
 Some other sources which are used occa- 
 sionally, and which are of very little use* 
 for a great supply, are such as the dew, 
 ice, snow and distilled water. These 
 latter we may dismiss with a word or two, 
 as only of exceptional utility. Dew has 
 , been used in deserts and at sea. Ice and 
 
snow furnish enormous quantities of 
 water in certain,, places where they 
 abound. Ice furnishes an exceptionally 
 pure water, because in freezing the salts 
 are separated out, and the gases too; 
 such water, therefore, requires aeration. 
 Snow and ice if used should not be col- 
 lected near to dwellings, because of the 
 risk of contamination. Distilled water 
 is an important water supply now, espe- 
 cially at sea. Its chief fault is that it 
 requires aeration. To give it this, Nor- 
 mandy's apparatus may be used, or it 
 may be allowed to fall from one vessel 
 to another like a shower. It has been 
 said that cases of lead poisoning have 
 occurred at sea " partly from the use of 
 minium in the apparatus, and partly 
 from the use of zinc pipes containing 
 lead in their composition." (Dr. Parkes.) 
 ' So much for the subordinate sources, 
 which are all of little importance to us. 
 We now come to rain, which is the 
 original source of all great supplies. 
 Kain, which we are going to consider, 
 is, of course, caused by the fact that, 
 
32 
 
 when two air currents come together, 
 both saturated with moisture, one having 
 a lower temperature than another, the 
 mixed air, though it has a mean tempera- 
 ture, has not the mean capacity for 
 water, but a capacity less than the mean, 
 and so some of it falls as rain. Is rain, 
 as it falls, sufficiently pure to be used as 
 a source of drinking water? In the 
 first place it is very soft. In the second 
 place it is well aerated. It dissolves 
 especially carbonic acid and oxygen from 
 the air the former being about three 
 per cent, of the total dissolved gases 
 and the latter from 30 to 40 per cent. It 
 contains nitrates and nitrites, especially 
 during thunder storms. It contains salts 
 of ammonia, which render it more alka- 
 line when collected in the country. Near 
 towns it contains most of the impurities 
 that are found in the air of towns, and 
 especially it becomes acid instead of 
 alkaline, absorbing a large amount of 
 the sulphuric acid that is in the air. It 
 contains organic matter, and this in in- 
 creased amount near towns. Eain half 
 
33 
 
 a mile from the extreme southwest of 
 Manchester, although the wind was 
 blowing from the west, tasted flat, in- 
 sipid, oily and nauseous deposited or- 
 ganic matters, and even organized bodies 
 in considerable quantities, and left a 
 clear water above, containing more than 
 two grains of organic matter in the gal- 
 lon. Dr. Angus Smith, who examined 
 this water, makes the following remarks : 
 " It becomes clear from the experiments, 
 that rain-water in town districts, even a 
 few miles distant from a town, is not a 
 pure water for drinking ; and that if it 
 could be got direct from the clouds in 
 large quantities, we must still resort to 
 collecting it on the ground in order to 
 get it pure. The impurities of rain are 
 completely removed by filtration through 
 the soil ; when that is done there is no 
 more nauseous taste of oil or of soot, 
 and it becomes perfectly transparent." 
 He is therefore of opinion that rain col- 
 lected directly from the air cannot, at 
 any rate near to towns, afford a proper 
 water supply. However, since rain i$ 
 
34 
 
 lie source of all the supplies that we 
 get, it becomes necessary, and of great 
 importance in estimating the amount of 
 water that can be got in a district, to 
 measure the rain-fall of that district. 
 Now the depth of the rain-fall of a dis- 
 trict has extraordinary varieties, both as 
 to place and time. For instance, as re- 
 gards time, the tropical rain-fall is almost 
 all at one part of the year. With us it 
 is variable. The rain-fall is measured 
 in England by its depth in inches. The 
 rain-fall is greater in mountainous dis- 
 tricts, and on the leeward side of mount- 
 ains, if they are not high enough to 
 penetrate the clouds; but if they are, it 
 is on the windward side, because the 
 clouds do not get over the tops of the 
 mountains. Now, for the supply of 
 water, the important points to be known 
 about the rain-fall are these : The first 
 is the least amount of rain that has ever 
 been known to fall in a year in a district; 
 the minimum annual fall. Then it is im- 
 portant to know the distribution of the 
 rain throughout the year, and especially 
 
35 
 
 the longest drought, because you have 
 got to provide for that time as well as 
 for any other time, and the observations 
 on the rain-fall of a district should ex- 
 tend over not less than 20 years. Of 
 course it is not often that you can get 
 observations at any locality that have 
 been maintained for 20 years, and so we 
 shall have to consider in an instant or 
 two how we are to get over that diffi- 
 culty. 
 
 The machine used for measuring the 
 depth of rain-fall is called a rain gauge. 
 It is essentially a funnel, the area of the 
 top of which is known very accurately. 
 The top of this funnel is provided with 
 a vertical rim to catch the splashings, so 
 that none may be lost. Below the fun- 
 nel there is a glass vessel placed to re- 
 ceive the water. The height of the 
 water in it may be indicated by a float, 
 or its quantity may be ascertained at 
 given intervals of time by measuring or 
 weighing it, and that is the best plan. 
 Of course the number of cubic inches of 
 water, which is the same as the number 
 
 
36 
 
 of square inches of the area of the f un- 
 nel, gives you one inch of rain over that 
 area. Suppose the area of your funnel 
 is 20 square inches, 20 cubic inches of 
 water will obviously be the result of one 
 inch of rain- fall over that 20 inches. It 
 is most convenient to measure the water, 
 and the measuring glass is constructed in 
 the following way : At the place where 
 that amount of cubic inches of water 
 stands which is equal to the number of 
 square inches in the area of your funnel 
 a line is drawn, and this represents one 
 inch of rain-fall. If the area of your fun- 
 nel is 20 square inches, then you take 20 
 cubic inches of water which you have 
 weighed or measured accurately, place it 
 in your glass vessel and mark one at the 
 level where it stands, because that 
 amount is equal to a depth of one inch 
 of water over the area you are observing. 
 One cubic inch of water weighs 252^ 
 grains, almost exactly. That one inch is 
 divided into tenths and hundredths; and 
 with this vessel you are able to measure 
 the amount of rain that has fallen 
 
37 
 
 through the funnel in a given time. The 
 top of the gauge must be placed nearly 
 level with the ground ; the instrument 
 must, in fact, be sunk. It must be 
 placed in an open situation, and a fence 
 put around it if necessary. One is very 
 frequently placed at a height above the 
 ground, and one on the ground, to show 
 the difference in the amount of rain that 
 falls at the two levels. The amount of 
 rain that falls at the level of the ground 
 (leaving hills out of the question) is 
 always greater than the amount that 
 falls at any height above the ground. 
 If you have got records of the rain-fall 
 of a district for a considerable number 
 of years your work is to a great extent 
 done, because then you have merely to 
 take out the facts that you want. If 
 you have not, the only way to do it 
 (with a limited time) is to place rain 
 gauges at convenient situations, and as 
 many as possible all over the district you 
 are examining, and if there are any hills 
 in or near the district some of them 
 ought to be placed on their tops, and 
 
38 
 
 each of these rain gauges ought to be 
 carefully and regularly examined at cer- 
 tain fixed times. Then you must com- 
 pare the records of all these gauges with 
 the results given by the nearest rain 
 gauge that has been observed for a con- 
 siderable number of years, to get a kind 
 of relation between the rain that falls at 
 these different stations on your district, 
 and the rain that falls at the nearest 
 place from which you can get any re- 
 liable data, and from this comparison 
 you must calculate what will probably 
 be the longest drought in your district, 
 and what is probably the least annual 
 rain-fall. Now, the average in different 
 parts of England is from 22 inches to 
 100, or even 120 per annum; in some 
 countries, as Burmah, 180 to 220, and it 
 is even said to be as much as 600 inches 
 in one place. This useful rule was given 
 by Mr. Hawkesley (and certainly the 
 tables show that it is a very accurate 
 rule) that if you take the average rain- 
 fall of a place for 20 years, and subtract 
 a sixth from it, that will give you 
 
39 
 
 the average annual rain-fall of the 
 three driest years during that period. 
 If you take the average annual rain -fall 
 for 20 years, and take a third part from 
 it, that will give you the amount of rain 
 in the driest year of these 20 almost 
 exactly, and if you take the average of 
 20 years, and add a third to it, then that 
 will give you pretty nearly the amount 
 of rain in the wettest year. 
 
 So you get with a considerable amount 
 of accuracy the quantity of rain ; the 
 least amount of rain you are likely to 
 get, and the greatest as well. Then, of 
 course, you want to know the area of the 
 district, and besides the actual amount 
 of the rain-fall, you must also know the 
 amount which is available. In the first 
 place, a great deal of the rain-fall is lost 
 by evaporation and absorption. Evapo- 
 ration from the surface, and absorption 
 by plants, etc. Then, if the ground is 
 very porous to a great depth, a consider- 
 able amount will be absorbed so fast that 
 you cannot collect it. Most of the rain- 
 fall is at once available from or near the 
 
40 
 
 surface in steep countries, and especially 
 those which are formed of primitive and 
 metamorphic rocks, as granite, clay-slate, 
 etc., and generally from impervious 
 rocks that are steep-sided. Almost all 
 the rain-fall in these cases is available at 
 once. It runs off the surface and col- 
 lects in lakes, and is available directly. 
 And then, on hilly pasture lands in lime- 
 stone and sandstone regions, something 
 like two -thirds of the rain may be con- 
 sidered available, and on flat pasture 
 countries something like one-half. For 
 instance, on the green sand, Mr. Prest- 
 wich estimated that from 36 to 60 per 
 cent, is available. On chalk and loose 
 sand there is very little indeed available. 
 Now one of the most important things, 
 if not the most important thing to know, 
 is the geological character of the rocks 
 of the district you are examining, be- 
 cause that will tell you a very great deal 
 about the amount of available water, 
 and about the way to get it. We are 
 told that in chalk countries the rivers 
 and streams carry off at once about a 
 
41 
 
 fifth of the rain-fall ; that the evapora- 
 tion and absorption by vegetables and 
 animals amounts to as much as a third, 
 and that the remainder (i. 6., the greater 
 part of the total rain-fall) sinks into the 
 ground. In less absorbent strata you 
 may put down that it is about equally 
 divided that one-third is carried off by 
 the streams, etc.; another third absorbed 
 by plants and animals, or lost by evapo- 
 ration, while a third sinks into the 
 ground. 
 
 Well now, let us consider what means 
 have been taken to get at this water that 
 sinks into the ground. Of course it is 
 got at by digging down, and now we 
 must consider in what strata we are likely 
 to be successful in digging wells or mak- 
 ing borings to get underground water. 
 In the first place, wells in sands lying 
 over impervious strata, over clays es- 
 pecially, if they are not deep, do not, as 
 a rule, afford much water. They may, 
 however, afford a fair supply as to quan- 
 tity, but very often afford a bad supply 
 as to quality. For instance, the wells 
 
42 
 
 sunk into the sands and gravels over the 
 London clays afford a very impure water. 
 If water of this description has come di- 
 rectly from the surface, and especially in 
 the neighborhood of towns, it is contam- 
 inated in all sorts of ways. The water 
 in these wells never overflows or spouts 
 up. Wells, on the other hand, sunk 
 through impervious strata to pervious 
 ones below, generally, though not always, 
 supply excellent water. At any rate, 
 they have much greater chance of sup- 
 plying excellent water, because they sup- 
 ply the water that has come from the 
 high grounds at a considerable distance. 
 For instance, the borings that are made 
 through the London clay down to the 
 chalk, supply some of the best water in 
 London. The Kent water is still better, 
 and is supplied in large quantities by 
 borings which pass through the chalk, 
 through the upper green sand, and 
 through the gault (an inpervious stratum) 
 into the lower green- sand. These wells 
 are known as Artesian wells. The water 
 rises up a considerable height in them, 
 
43 
 
 and may overflow. It is often thought 
 that Artesian wells always overflow, but 
 they don^t. The water rises up to a cer- 
 tain height, which height is of course 
 determined by several considerations 
 for instance, by the height it came from 
 originally. Of course the water that 
 you get from under Kent is the water 
 that has fallen upon the outcrop of the 
 green-sand at a very considerable dis- 
 tance round the London basin. 
 
 Mr. Prestwich, who has paid the great- 
 est attention to the water supply of Lon- 
 don, and to the arrangement of the strata 
 around London, has calculated that, from 
 the lower green- sand underneath the 
 London basin, there is to be got an enor- 
 mous supply of water for the metropolis, 
 that is to say, on the presumption that 
 this lower green-sand is continuous under- 
 neath London. It would not be fair if I 
 did not tell you here that the lower 
 green-sand does not appear to be con- 
 tinuous underneath the London basin. 
 Some of the older strata are brought into 
 contact with the chalk, so that the lower 
 
44 
 
 green-sand is missing, probably, under- 
 neath a great part of the district. This 
 we know from deep borings which have 
 been made at several places. Of course 
 the chalk and also the green sand are 
 merely instances. You want to know 
 the alternation of the strata right away 
 down the whole geological series, so as 
 to be able to say, if you go into a coun- 
 try and study the maps and sections for 
 a short time, " If we make a well here 
 and bore down, we shall probably go 
 through a band of clay into a pervious 
 stratum, and get a supply of water." 
 You want for this purpose to study the 
 geological maps, and to have ample time 
 to do it. If we go below the chalk into 
 the oolitic series, we have similar alter- 
 nations of pervious and impervious 
 strata. When we go below this we 
 come to the new red sandstone, and I 
 mention this, because there is an im- 
 portant point connected with it. The 
 new red sandstone is (to a great extent) 
 a pervious stratum. It contains enor- 
 mous quantities of water, but the cau- 
 
45 
 
 tion about it is, that in many countries 
 it holds immense salt deposits. It is in 
 the new red sandstone of Worcester- 
 shire (for instance) that the salt deposits 
 of Droitwich are found ; so that borings 
 in the new red sandstone (although it is 
 true that some towns are supplied from 
 that stratum), are frequently found to 
 give a brackish water. Below this come 
 the Permian strata, in which you have 
 the magnesian rocks, that I mentioned 
 last time, and it is a mischievous thing 
 to bore into these strata, because you 
 may get water containing large amounts 
 of magnesian salts. Towns which are 
 placed upon these strata are best sup- 
 plied (like Manchester) from older for- 
 mations, such as mountain limestone, 
 and so on, which generally afford excel- 
 lent water. The best supplies are ob- 
 tained from them, not by boring or 
 by wells, but from springs. There is 
 one thing I must mention, before I leave 
 the wells, and that is, that the sinking 
 of deep wells may lower the level of the 
 water in the country above considerably, 
 
46 
 
 and that is a point that has often to be 
 taken into consideration. For instance, 
 Mr. Clutterbuck showed that wells at a 
 considerable distance from London have 
 been seriously affected by the pumping 
 of the green- sand water below London. 
 He showed that the level of the water in 
 these wells was affected so much, that 
 you could tell by the levels of the well 
 waters at a considerable distance from 
 London, whether the pumping had been 
 going on in London on the previous day 
 or not. There is another thing that re- 
 quires to be known, especially about 
 borings in the chalk, and that is, that 
 some of the borings will give an inex- 
 haustible supply of water, practically 
 speaking, while borings close by will 
 give you next to none. This Mr. Prest- 
 wich accounts for, by stating that the 
 water in chalk runs chiefly through 
 crevices, and does not infiltrate through 
 the mass of rock. Before I say a few 
 words to you about the construction of 
 wells, I have something to say about 
 springs, and the amount of water they 
 
47 
 
 supply. Now springs occur where you 
 have an impervious stratum cropping 
 out from beneath a pervious one, and 
 this may happen in various ways. 
 
 The water in springs, and also that in 
 wells, varies very much in quality accord- 
 ing to the place that it is taken from. 
 Spring water differs from rain water in 
 that it has passed through certain rocks, 
 and dissolved more or less considerable 
 quantities of substances on its way. 
 Spring water resembles rain water in 
 containing a considerable amount of car- 
 bonic acid in solution. This has the 
 property of dissolving many substances, 
 one of the chief of which is the carbon- 
 ate of lime. The water then passing 
 through the rocks dissolves carbonate 
 and sulphate of lime, salts of iron, etc. 
 It is important to know this for many 
 reasons. In the first place, some of 
 these waters dissolve, in mountain lime- 
 stone districts for instance, so much car- 
 bonate of lime as to become what is 
 known as petrifying springs. Of course 
 if you take a petrifying spring and bring 
 
48 
 
 it along an aqueduct, under certain con- 
 ditions your supply is stopped up ; and 
 one of the aqueducts at Rome is to be 
 seen to this day perfectly closed for a 
 considerable length with a deposit of 
 carbonate of lime and other salts, be- 
 cause the contractor took in a spring 
 that he was not told to tap a mineral 
 spring. Now the purest spring water 
 you can get comes from the igneous, the 
 metamorphic, and the older stratified 
 rocks. Many of these hard rocks yield 
 a very pure water without a great deal 
 of salts in solution. The mountain lime- 
 stone, the oolitic limestones, and the 
 chalk rocks also yield a good supply, 
 and these waters are fit for drinking so 
 long as they do not contain any quantity 
 of magnesian salts. Water from sand- 
 stones, especially the new red sandstone, 
 I have told you, often contains common 
 salt. Waters in clay countries very often 
 contain considerable quantities of the 
 sulphate of lime. The waters of the 
 London and Oxford clays do, as also the 
 water of the lower lias clay. These are 
 
49 
 
 bad waters. They are permanently bard 
 and unwholesome. Well waters have 
 partly the same qualities, unless they con- 
 tain additional impurities from the causes 
 I have mentioned before. River water 
 is often purer than spring water; that is 
 to say, it often contains less total solids 
 in solution. The permanent hardness is 
 generally greater. It contains less sub- 
 stances in solution, because much of the 
 carbonic acid has escaped, and the sub- 
 stances it held in solution have been de- 
 posited. River water very often con- 
 tains much more organic matter, espe- 
 cially near towns. 
 
 Wells sunk in hard rocks may require 
 no lining at all; if they pass through 
 sandy strata they require a lining of 
 brickwork, and sometimes part or the 
 whole of it must be set in cement. For 
 an artesian well, ao ordinary well is dug 
 first of a tolerable breadth and depth, 
 and then a boring is made which varies 
 from twenty down to three or four 
 inches in depth. As soon as an impervi- 
 ous layer is bored through, and a pervi- 
 
50 
 
 ous stratum reached, the water rises 
 through the boring into the well (which 
 acts as a sort of cistern), and has to be 
 pumped up, or it may rise so high as to 
 overflow. 
 
 The ordinary atmospheric lifting pump 
 is seldom used, but a kind of lifting pump 
 with a solid piston and metallic valves is 
 often used. In fact, the cylinder in 
 which the solid piston slides is connected 
 with the space between the valves above 
 the piston instead of below it. So that 
 when the piston is raised the water is 
 lifted through the upper valve, and when 
 it is depressed water is drawn from the 
 well into the body of the pump through 
 the lower valve. Forcing pumps are 
 also used. They are driven by engines, 
 and the water is pumped into air vessels, 
 by which the pressure on the mains is 
 equalized so that it does not come in 
 jerks. Let me mention one or two ex- 
 amples of artesian wells, and the amounts 
 of water got from them in different 
 strata. From the well of Grenelle, near 
 Paris, in 1860, there were about 200,000 
 
51 
 
 gallons daily. This well when first sunk 
 yielded 800,000 gallons daily, so that 
 you see the supply has considerably 
 diminished with time, which is an im- 
 portant thing to take note of. The bor- 
 ing of this well of Grenelle began at 
 twenty inches in width, and ends at 
 about eight or somewhat less. It is 
 1,800 feet deep (being one of the deepest 
 borings ever made), and more than 1,700 
 feet of it is lined with copper tubing, 
 which was placed there instead of some 
 wrought-iron tubing, with which it was 
 originally lined. The copper tubing be- 
 gins at 12 inches in diameter and goes 
 down to 6. The temperature of the 
 water in this well at about 1,800 feet is 
 as much as 82 F., and you may put it 
 down that as a rule, the temperature of 
 the water increases 1 F. for every 50 
 feet below the surface. Of course there 
 are certain places where it increases very 
 much more (about Bath for instance), 
 but these are exceptional cases. The 
 boring in the well at Trafalgar square is 
 sunk 384 feet from the surface into the 
 
52 
 
 chalk, and it yields 65 cubic feet in a 
 minute, or more than 580,000 gallons in 
 the 24 hours. There is a well in Wool- 
 wich in the chalk 580 feet deep, which 
 yields 1,400,000 gallons in 24 hours, and 
 the last I am going to mention in the 
 chalk is a well near London the Am- 
 well hill well close by the source of the 
 New River. That is only 171 feet deep, 
 and it is said to yield very nearly 2 
 million gallons in the 24 hours.* As all 
 this water underneath the London basin 
 comes originally from districts at some 
 distance from London, it is not to be 
 wondered at that the pumping at London 
 lowers the level of the water in the wells 
 in those districts. These are examples 
 of successful borings. Now, a word with 
 regard to the new red sandstone wells of 
 Liverpool. These wells you will find 
 described in the twelfth volume of the 
 proceedings of the Institution of Civil 
 Engineers. One of them, called the 
 " Bootle Well," has many points of in- 
 terest about it. Its maximum yield was, 
 *Hughes' " Water-works." 
 
53 
 
 in 1853, about 1,100,000 gallons in the 
 24 hours. A curious point about it is 
 that at the bottom of the well instead of 
 there being one boring there are 16 or 17. 
 These 16 or 17 borings are of very differ- 
 ent depths, and it became very interest- 
 ing to know whether the whole of them 
 were of any use, and Mr Stephenson 
 thought of blocking them up, all but 
 one. He did so, and found that one 
 yielded very nearly as much water as the 
 16, so that a very considerable amount 
 of capital had been wasted in the boring 
 of these holes. That is worth knowing. 
 There are six other public wells at Liver- 
 pool in this new red sandstone, and the 
 ordinary yield was about 4 million gal- 
 lons daily from them all. This was in 
 Is50. Eighteen years afterwards, evi 
 dence was given before the Commission- 
 ers on the Water Supply for the Metrop- 
 olis of a falling off in the water supply 
 of these wells. In fact, the continual 
 pumping had diminished the supply. In 
 1854, these wells in the new red sand- 
 stone at Liverpool were pronounced fail- 
 
54 
 
 ures by Mr. Eawlinson, as also were 
 others in England and America, and Mr. 
 Piggott Smith, in a report on the water 
 supply of Birmingham, confirmed this, 
 and it is a fact that they have had to be 
 supplemented by a supply of much su- 
 perior water from a distance. Mr. 
 Stevenson estimated the cost of a pump- 
 ing station for one of those Liverpool 
 wells, including shafts and steam engines, 
 at 20,000, and the annual cost per mil- 
 lion gallons a day at 1,324, this being 
 without interest or compensation, but in- 
 cluding depreciation^ Generally, well 
 waters are liable to vary in amount from 
 month to month, and from year to year, 
 as witnessed by the amounts pumped 
 from these Liverpool wells, and by the 
 amounts pumped year after year from 
 the Cornish mines. 
 
 After wells, the next thing we have to 
 consider is the way in which water can 
 be collected from springs and streams 
 over a large area, called a drainage area. 
 That is one method of supply, and the 
 other method, of course, is pumping 
 
55 
 
 from rivers. We tell the amount of 
 water that can be got from a large sur- 
 face of land, in the first place, by a way 
 I spoke to you about before, viz.: by 
 estimating the amount of available rain- 
 fall on it. Then we can tell it in another 
 way, by correctly measuring the amount 
 of water that is brought down by 
 streams and springs ; so that we have to 
 consider the methods used for gauging 
 springs, streams, etc. The gauge most 
 commonly in use is the one known as the 
 Weir gauge. Weir gauges are made by 
 damming up the stream, and making it 
 all pass over a sharp ledge or through 
 an orifice or notch, or row of notches, on 
 a vertical board. Then from formula 
 you can, by means of tables, calculate 
 the amount of water that passes through 
 the notch, or over the orifice of the weir 
 in a given time. You determine the 
 height of the still water by means of a 
 scale, the zero of which is level with the 
 base of the notch, and you do it in this 
 way. A stick is planted in the bed of 
 the stream, its top at some little distance 
 
56 
 
 from the weir, and so that its top is level 
 with the base of the notch, or row of 
 notches, in the weir, and then you 
 measure by the scale from the top of 
 this stick to the level of the water from 
 the orifice. That is one way. The next 
 plan is by calculating from the declivity. 
 This is only applicable to regular chan- 
 nels, like the New Kiver for instance, 
 and if the stream is small you can make 
 the whole of it pass through a trough, 
 and then calculate the velocity from the 
 declivity. Another way is by measur- 
 ing the maximum surface velocity, which 
 is done by means o'f floats of any sort, 
 or by means of fan wheels, and various 
 little instruments for measuring the 
 surface velocity of streams. You take 
 the maximum surface velocity, and about 
 three-fourths of this will represent the 
 mean velocity of the section. The dis- 
 charge of springs is estimated by the 
 time taken to fill a vessel of known 
 capacity. A word about the permanence 
 of springs and streams, which is an ex- 
 tremely important point. In the first 
 
57 
 
 place you must try and get evidence 
 from maps and trustworthy sources gen- 
 erally. At the bases of hills springs are 
 usually permanent. In flat countries 
 you may put it down that the reverse is 
 generally the case. Springs in limestone 
 countries are very permanent. Springs 
 in very permeable strata are very gener- 
 ally variable, unless they are tapped at 
 a considerable distance from the surface, 
 and then they often give an enormous 
 yield. 
 
 Springs in primary strata and in 
 granite countries are very often very 
 permanent indeed, and it is in these 
 countries you have some of the large 
 lakes which are used ' for supplies of 
 water. Jn clay basins the water supply 
 is variable as a rule, being very great in 
 the winter, when there are often floods, 
 and very small in summer. In chalk 
 countries the springs are more perma- 
 nent, for the reason that they draw from 
 considerably beyond the actual basin. 
 Intermittent springs sometimes occur, 
 especially in the chalk ; they are due to 
 
58 
 
 the gradual collection of water in sub- 
 terranean hollows, which, when filled 
 above a certain level, empty themselves 
 by means of a syphon- shaped outlet ; 
 it is obvious that they must not be re- 
 lied on as sources of a supply of water. 
 This will end our consideration of the 
 merits of different localities from the 
 water supply point of view. 
 
 Having found a sufficient supply of good 
 water, or a sufficient supply of water 
 that can be purified on a large scale by 
 filtration a subject which we shall con- 
 sider further on or by means of Clark's 
 process, which I have described to you, 
 or by both combined, we come to the 
 modes of collection and distribution, 
 which vary very much as to the site, 
 sources, etc. One of the oldest plans, 
 and for all that one of the best, is the 
 eastern or Roman plan, if you like so to 
 call it, which is that of tapping natural 
 springs at their sources, or lakes, above 
 the places to be supplied, and conduct- 
 ing the water by channels or aqueducts 
 above or below ground, or alternately 
 
59 
 
 above and below, as occasion may re- 
 quire ; collecting it in large cisterns, al- 
 lowing the sediment to settle, and then 
 distributing by means of gravitation. 
 
 In later times we can adopt the same 
 plan, and distribute either by gravitation 
 or by steam power, as we choose. Per- 
 manent springs at a distance may be con- 
 veyed by the Roman plan through chan- 
 nels across the country, covered the 
 whole way right up to the distributing 
 reservoirs or tanks. The conduits may 
 be built of masonry and cement, like the 
 Roman aqueducts, embedded in puddle, 
 or they may be earthenware pipes, in 
 which case they must be laid in water- 
 tight trenches, and jointed securely, or 
 the water may be contaminated in vari- 
 ous ways, and much of it may be lost, 
 or the pipes may be of cast iron, and 
 this should be the case where deep val- 
 leys have to be crossed by means of in- 
 verted syphons. 
 
 Earthenware pipes are not strong 
 enough to be used as inverted syphons. 
 The rule is, that if the fall is greater 
 
60 
 
 than 1 in 300, then cast iron pipes should 
 be used. The fall of these conduits 
 should be 5 feet in a mile, if they are of 
 something like 2 feet in diameter, which 
 is of a small size. If larger, it may be 
 less, down to 1 in 10,000, or 6 inches in 
 the mile. That is the fall of the New 
 Kiver conduit that supplies part of the 
 north of, London with water. 
 
 The velocity of the water should not 
 be less than one foot in the second, so 
 that it may move at a sufficient rate, nor 
 greater than four feet in a second, for 
 fear it should wear away the course by 
 carrying down stones, etc. As an opin- 
 ion about this plan, which I am going to 
 describe to you at greater length, I may 
 mention that Mr. Rawlinson stated, in a 
 discussion on the water supply of Mel- 
 bourne, which you will find reported in 
 Vol. 18 of the proceedings of the Insti- 
 tution of Civil Engineers, that "he 
 thought the plan of gathering spring 
 water in Great Britain, by means of 
 earthenware pipes to some common 
 storage reservoir, was one that might be 
 
61 
 
 favorably looked at ; the modern means 
 of making earthenware pipes offered 
 many facilities ; and where springs were 
 at a sufficient elevation ani tolerably 
 permanent, the water might be collected 
 and brought into a covered reservoir on 
 the Eastern plan. There were situations 
 where that plan might be preferable to 
 making an impounding reservoir." 
 
 Now, I should like to give you a short 
 account of some of the points which are 
 to be observed in the Roman aqueducts 
 at Eome ; and afterwards I propose to 
 give you an account of some extremely 
 remarkable Roman aqueducts which are 
 very little known, and which have been 
 very seldom described, to wit, the aque- 
 ducts with which the town of Lugudu- 
 num, now called Lyons, was supplied, 
 which aqueducts have some very inter- 
 esting and instructive points about them, 
 as you will see directly. 
 
 As I think I told you before, Rome 
 was supplied by nine aqueducts. The 
 first two were built entirely underground 
 for their whole length, because the water 
 
62 
 
 supply might otherwise have been cut 
 off in case of invasion. The more an- 
 cient of these two, the oldest of all the 
 nine aqueducts, ran for a distance 
 of about 11 miles. I need not say any- 
 thing more about that one. When the 
 Romans built the third aqueduct they 
 were, it appears, no longer afraid of its 
 being destroyed by enemies, and so they 
 built it partly above ground, and partly 
 underneath the ground. By the direct 
 road to the place from which they took 
 the water was 39 miles from Eome. 
 Three thousand men were set to work 
 at it under the Praetor Marcius, and so 
 it has been called the Marcian aqueduct. 
 This aqueduct was made so strong that 
 the two succeeding ones were built on 
 the top of it, so that you have the 
 three channels one above another. The 
 size of the channel of the Marcian aque- 
 duct was about 5 Roman feet high by 
 2 wide.* The thickness of each of the 
 sides was a foot. You can see this aque- 
 
 * The Koman foot was equal to about 11. 65 English 
 inches. 
 
63 
 
 duct outside one of the gates of Rome 
 at the present day. 
 
 On these aqueducts there were venti- 
 lating shafts. There were also what are 
 known as piscinae, or small settling res- 
 ervoirs. These piscinae I shall describe 
 to you a little further on. Then the 
 base of the channel was broken up by 
 inequalities, partly to help to break the 
 very considerable fall, and likewise to 
 aerate the water by agitation. 
 
 I may now say a word or two about the 
 water supply of the Roman town of Lug- 
 udunum, in Gaul. In the first place, I 
 must remind you that those aqueducts 
 which supplied Rome with water were 
 carried across no deep valleys; they had, 
 it is true, often to be supported on high 
 arches,because they pass over low ground, 
 and the Romans have over and over again 
 been blamed for not using syphons ; it 
 has been said that the Romans were not 
 acquainted with the properties of water, 
 in that they did not use syphon sin these 
 aqueducts. We shall see directly wheth- 
 er that is true or not. 
 
64 
 
 The town of Lugudunum (Lyons) was 
 supplied by water by means of three 
 aqueducts. The first of them was built 
 in the first century before Christ, and 
 here is the description of it in a few 
 words. It had two branches, which unit- 
 ed at a particular place. It passed over 
 a large plateau in a straight line ; then 
 went underground. Emerging from be- 
 neath the ground, it descended, by means 
 of inverted syphons, into a deep valley, 
 and was received at the bottom of that 
 valley on a supporting bridge of arches. 
 It was thus carried across the valley, and 
 ascended the other side into a reservoir. 
 So you see in the course of this aque- 
 duct, which was built in the first century 
 before Christ, there was a large and 
 deep valley crossed by means of invert- 
 ed syphons, by the very method which 
 we employ now ; and this shows you 
 that the Romans then certainly under- 
 stood and perfectly well appreciated the 
 properties of the syphon. 
 
 I will now give you a description of 
 the second aqueduct by means of which 
 
65 
 
 Lugudunum was supplied with water. 
 It was underground the whole way, 
 and it carried the water to a greater 
 height than the other. The reason that 
 it was constructed at all was, because 
 the water was required to be carried to 
 a greater height than the former aque- 
 duct brought it. It was very nearly the 
 size of the Marcian aqueduct. It was 
 built of cubical stones placed together, 
 as I may tell you a great many of these 
 aqueducts were built. The stones were 
 placed together without cement, and 
 they fitted so accurately that some aque- 
 ducts built in this way are not even lined 
 with cement. This aqueduct is in all 
 probability intact at the present day for 
 three-fourths of its length. Now we 
 come to the third, which is the most im- 
 portant of the three, and which is, per- 
 haps, the most remarkable Roman aque- 
 duct of which we have the remains any- 
 ; where. The two former ones did not 
 bring the water to a sufficient height. 
 There is at Lyons an abrupt hill (Four- 
 vieres), on which several Roman palaces 
 
66 
 
 were built, and it was necessary to bring 
 water to these. The Emperor Claudius, 
 who was born at Lugudunum, and who 
 lived there, determined to bring water 
 onto this hill. He had already made an 
 aqueduct for Rome (the Claudiun aque- 
 duct), and so he knew something about 
 it. He had not used inverted syphons, 
 however, in his aqueduct at Rome, and 
 for the simple reason, as you will pres- 
 ently see, that it was practically impos- 
 sible ; but he comes and orders a new 
 aqueduct to be built for the city of 
 Lugudunum, and it is that one which we 
 are now going to consider, as briefly as 
 possible. 
 
 This aqueduct descended in the first 
 place into three or four valleys on its 
 way. The aqueduct was 52 kilometers 
 long, including the syphons. It had 17 
 or 18 bridges of arches to carry it over 
 low grounds, and four bridges to carry 
 the syphons across the valleys. 
 
 And now I may tell you the size of 
 the two more important of these valleys. 
 The valley of the river Garon, which is 
 
67 
 
 the second one it had to cross, is 120 
 meters deep, and 800 meters broad. 
 
 The valley of Bonan , which is 
 
 the next, and which is the place at which 
 I examined the aqueduct very carefully 
 some time ago , is 139 me- 
 ters deep, and 1,060 meters across, be- 
 tween the two reservoirs, which are 
 placed one on each side of the valley. 
 So you see these are two very consider- 
 able valleys that had to be crossed. 
 
 And now, how did the Komans man- 
 age to effect their purpose? Bridges 
 were out of the question, although we 
 know that they built splendid aqueduct 
 bridges, where possible, in such situa- 
 tions, as witness the well-known Pont 
 du Gard, near Nismes, which had three 
 rows of arches one above another, sup- 
 porting the channel, and which is even 
 now so perfect that it is about to be 
 utilized for the purpose for which it was 
 originally built. 
 
 They used inverted syphons. I told 
 you that earthenware pipes will not do 
 for syphons. Cast iron pipes need to 
 
68 
 
 be employed for large syphons. The 
 Romans could only work iron on a 
 small scale, and so used leaden syphons. 
 One thing they did, and which it is im- 
 portant to note in this : the water was 
 brought up along the single channel of 
 the aqueduct the specus, as it was 
 called which in this particular one is 
 about 2 Eoman feet broad by 6 high, 
 into a reservoir. This reservoir had 
 some such dimensions as 5 yards by 
 nearly 2, and the walls were about a 
 yard thick ; there was an opening in 
 the roof for the purpose of cleansing, 
 and on the front side of the reservoir 
 (the one facing down the valley), there 
 were several holes into which the leaden 
 pipes were fixed. Now one of these 
 valleys had 8 leaden syphons, another 
 9, and another 10 ; and the object, of 
 course, of dividing the water in this 
 way was that they might get pipes that 
 would resist the enormous pressure, and 
 if a pipe burst the rest might remain 
 sound, so that only part of the water 
 would be lost. Delorme, I should tell 
 
69 
 
 you, has calculated that this single aque- 
 duct brought 11 millions of gallons of 
 water into the place in 24 hours. It is 
 hardly likely that it brought so much as 
 that, but it certainly brought a consider- 
 able amount. 
 
 The interior of the channels was usual- 
 ly constructed of very small stones care- 
 fully placed, and generally laid in ce- 
 ment. There was in this particular one 
 and probably it was so generally a 
 layer of cement along the walls of the 
 watercourse, and another layer, a consid- 
 erably thicker one, along the base of the 
 channel. The arches of the bridges 
 were built of enormous rectangular 
 blocks of stones, and the pillars broken 
 at certain intervals by layers of brick- 
 work buried in cement. The whole of 
 the exterior of this was covered over 
 with the work known to engineers as the 
 " opus reticulatum," which is made of 
 cubical pieces of stone, fitted carefully 
 together. 
 
 There is another curious thing to ob- 
 serve, and that is that the syphons were 
 
70 
 
 provided with little tubes, or valves, to 
 let out any air that might be carried 
 down from the height by the water, and 
 which might otherwise break the pipes. 
 In the smaller valleys there were small 
 leaden tubes, which rose up from the 
 lowest part higher than the reservoirs, 
 and in the larger ones weighted valves 
 were used for the same purpose. But 
 what I want you to see in this is, that 
 by the time the Romans constructed 
 even the earliest of these aqueducts at 
 Lugudunum, they knew perfectly well 
 the properties of water. They knew 
 perfectly well they could make it travel 
 up to the top of a hill if it had come 
 down a slightly higher hill on the other 
 side of a valley. Now I just wish to 
 give you the height of the reservoir on 
 the one side of the valley of Bonan, the 
 deepest of fhem all. The height of that 
 reservoir above the level of the Saone at 
 Lyons, is 151 meters, or something over 
 that. At the other side of the valley 
 into which the water was received the 
 reservoir was 143 meters above same 
 
71 
 
 level, that is to say, the difference in 
 height between those two reservoirs was 
 only eight meters. In another case, it 
 was 9 meters. Not only then did they 
 understand these matters so well as that, 
 but they actually lessened this amount 
 by causing the syphons to enter nearest 
 reservoir the one nearest the place to 
 be supplied high up close to its roof, so 
 that they actually thus diminished the 
 pressure by at least a meter. I have 
 given you this description at such 
 length, because it shows how much we 
 have to learn from what has been done a 
 very long time before our own age, and 
 also because there are so few descrip- 
 tions of these splendid aqueducts. 
 
 We now come to the next plan, that 
 of having a large drainage area, and of 
 collecting the water from that area into 
 an impounding reservoir. Before I be- 
 gin to describe this, I will give you a 
 brief account of one or two important 
 impounding reservoirs. The first one 
 will be that of the Bivington Pike reser- 
 voir, which now supplies the town of 
 
72 
 
 Liverpool with most of its water. This 
 Rivington Pike reservoir is calculated to 
 supply 21 millions of gallons of water 
 per day to Liverpool, and it has 481 
 million ^cubic feet of contents, with a 
 drainage area of 16^ square miles ; its 
 embankment is 20 feet high. You will 
 see from that, that it is calculated to 
 contain 150 days' supply. 
 
 Then there is a reservoir which was 
 made to supply Melbourne with water, 
 the particulars of which are given in 
 the volume from which I quoted to you 
 before, namely, Vol. 18 of the Proceed- 
 ings of the Institution of Civil Engi- 
 neers, in a paper by Mr. Bullock Jackson. 
 It is. called the Yan Yean reservoir. The 
 description runs thus : 
 
 " The Yan Yean reservoir was formed 
 by throwing 'an embankment across a 
 valley between two spurs of hills ; thus 
 retaining the rain-water which falls on 
 the natural basin, as well as the flood- 
 water which is led into it in winter from 
 the Upper Plenty Eiver ; the river itself 
 and the artificial watercourse forming,. 
 
73 
 
 in the latter case, a vehicle for its con- 
 duction. The ' area of this reservoir, 
 when full, is 1,303 acres ; the greatest 
 depth is 25 feet 6 inches, and the aver- 
 age depth not less than 18 feet. Its 
 contents measure nearly 38,000,000 cubic 
 yards, or upwards of 6,400,000,000 gal- 
 lons. The area of the natural catchwater 
 basin, independent of the reservoir, is 
 4,650 acres ; so that, including the area 
 of 600 acres drained by the watercourse, 
 there is a direct drainage into the reser- 
 voir of 5,250 acres. . . . The origi- 
 nal surface of the ground at the site of 
 the Yan Yean reservoir consisted of a 
 stiff retentive clay; the site was, there- 
 fore, admirably adapted for a reservoir. 
 Prior to the commencement of the 
 works, about two-thirds of the whole 
 area were densely timbered with large 
 specimens of eucalyptus, which were 
 taken up and burnt. The sides of the 
 reservoir, excepting in two parts, rise in 
 a steep slope. The embankment is 1,053 
 yards in length at the top, and 30 feet 9 
 inches in height at the deepest part ; the 
 
74 
 
 width at the top is 20 feet ; the inner 
 slope is 3 to 1, and the outer slope 2 to 
 1. The inner slope is pitched with rough 
 stones from 15 to 20 inches deep. Along 
 the center is a puddle bank and puddle 
 trench, with an inner apron and check 
 trench. The puddle trench and bank 
 are unusually thick, because, in the first 
 place, almost the whole of the material 
 used in the construction of the bank was 
 clay, so that it entailed little extra ex- 
 pense ; but principally, because previous 
 to the works being commenced, the site 
 of the embankment was occupied by 
 trees of a gigantic size, with long strag- 
 gling roots, which were all grubbed up, 
 and which it was feared might leave 
 clefts in the soil." 
 
 According to Mr. Hawkesley, the con- 
 siderations that you have to take into ac- 
 count in constructing impounding reser- 
 voirs are these : In the first place you 
 have to consider the extent of the drain- 
 age area. In the second place, the 
 amount of rainfall. And in the third 
 place the quantity of rainfall which can 
 
75 
 
 be collected into any reservoir which it 
 is practical to make in the district. The 
 size of these reservoirs must be propor- 
 tioned to the population to be supplied, 
 their area often requiring to be -fa of the 
 area of the water- shed. Mr. Hawkesley 
 stated in a discussion, that he considered 
 on an average of years that 30 inches of 
 rainfall out of a rainfall of 48 inches, 
 could be collected in an impounding res- 
 ervoir. It is usually considered that 
 one-sixth part of the total rainfall must 
 be put down as lost every year by floods 
 that you cannot store. The water that 
 you cannot collect is, of course, lost by 
 evaporation from the surface of the 
 ground, absorption by plants, and so on. 
 Now as to the site of the reservoir. 
 In the first place steep-sided valleys are 
 the best situations. In the next place, 
 it is necessary, of course, that the place 
 for collecting and storing water should 
 be sufficiently high above the place to 
 be supplied, so as to enable you to sup- 
 ply water by gravitation, and necessary 
 also, that it shall not be too high above 
 
76 
 
 it, so that you may not have too great a 
 rush of water. 
 
 Then besides the situation, the incline 
 of the rocks must be considered. It is 
 especially important in limestone that 
 the dip of the strata shall be in the 
 direction in which the water is running, 
 because if the dip is against it you very 
 often have immense quantities of water 
 lost, disappearing between the strata 
 and running away in another direction. 
 Stiff impervious clay or compact rock 
 affords the best situation. Trial shafts 
 or borings require to be made at various 
 places, it being better to make shafts 
 than borings, to see if .you have a suffi 
 ciently impervious material for the bed 
 of the reservoir, and a sufficient depth 
 of it. It is only with small reservoirs, 
 as a rule, that you can safely puddle the 
 whole of the bottom, or that it is done, 
 and for this reason in small reservoirs 
 the site is of less importance, as you can 
 puddle the whole of the bottom, and 
 carry it under the embankment of the 
 puddle wall. 
 
The embankment should have con- 
 structed what is called a puddle wall 
 down the center of it. I shall do well 
 to give you some rules about this. Mr. 
 Rawlinson lays it down, that the puddle 
 wall is to be a foot thick at the surface 
 of the ground for every three feet in 
 height of the embankment, that is to 
 say, that in an embankment 100 feet 
 high, the puddle wall should be about 
 33J feet thick at the base. Then it 
 slopes up to the top so as to be about 
 four feet broad at the top. Having de- 
 cided the thickness that you are going 
 to make the puddle wall by the height 
 that you are going to make the embank- 
 ment, according to that rule, you have 
 then to dig what is called a puddle 
 trench. This is dug down to a con- 
 siderable depth into the impervious bed 
 that will be the bottom of the reser- 
 voir. The trench is usually sunk with 
 sides sloping towards one another, 
 though this is considered by some 
 authorities to be an insecure plan. It 
 would involve a considerable amount of 
 
78 
 
 extra work, which would to a great ex- 
 tent be unnecessary, to sink the puddle 
 trench with sides diverging from one 
 another, as you would expect it ought to 
 be, and so it is sometimes recommended 
 to sink the puddle trench with perpen- 
 dicular sides. If it is very wet at the 
 bottom of a puddle trench, it is usual to 
 begin filling it with Portland cement 
 concrete, and then to go on with the 
 puddling. For puddling only the stiffer 
 kinds of clay are used. On each side of 
 the puddle wall a masonry wall is built, 
 about equal to it in thickness. The ex- 
 ample I gave you was one in which the 
 embankment slope on each side of this 
 puddle wall was pretty correct, namely, 
 three to one inside, and two to one out- 
 side. This embankment is made of such 
 materials as can be obtained in the 
 neighborhood, and the whole embank- 
 ment must be made in very thin layers, 
 and should be trampled in as much as 
 possible. The inner slope of the em- 
 bankment is shingled up to a little short 
 of the water-mark, and from that point 
 
79 
 
 it is pitched with blocks of stones. It is 
 sometimes necessary to make minor em- 
 bankments across valleys that may join 
 with the one you are going to make into 
 a reservoir. Now a reservoir requires a 
 waste weir for the storm waters. This 
 is generally made round the end of the 
 embankment, or cut into the hillside. 
 The water is carried from this point 
 down to the old stream-course, and the 
 channel is puddled until you are well 
 clear of the embankment. 
 
 I have one or two words to say about 
 the reason for the existence of these im- 
 pounding reservoirs, and also about the 
 size which it is necessary to make them, 
 and the rules that are laid down for the 
 amount of water that they should hold. 
 In the first place, they are necessary 
 where a sufficiently copious and perma- 
 nent supply cannot be got from a river 
 or large stream, or from artesian wells, 
 in order to secure a constant supply of 
 water throughout the year, and they do 
 this by storing the extra supply of water 
 during floods, so that it may be saved 
 
80 
 
 for use in times of drought ; secondly, 
 they allow a settling to take place ; and, 
 in third place, they are necessary to pre- 
 vent damage to the lower lands by 
 floods, for great damage is occasionally 
 done by the floods, even of such rivers 
 as the Thames and the Severn, and, of 
 course, great quantities of water are 
 wasted. 
 
 The size must depend upon the 
 amount of water required, and upon the 
 permanence of the supply ; we have 
 reckoned the requisite supply at thirty 
 gallons per head per day. Impounding 
 reservoirs should, according to the opin- 
 ion of many engineers, hold a six months 
 demand. "You can tell how much that 
 is, if you will lay down the amount of 
 gallons which you intend to supply per 
 head, and the population to be supplied. 
 If possible, the gathering-ground that 
 supplies these reservoirs should be so 
 large that the least available annual 
 rainfall is sufficient for the supply ; and 
 then the reservoir should contain an ex- 
 cess of six months demand over six 
 
81 
 
 months least possible supply ; that is to 
 say, supposing the least possible supply 
 at any time during the year is zero, then 
 the reservoir must contain six months 
 demand. The reservoir must be (to put 
 it in Mr. Hawkesley's words) "suffi- 
 ciently large to equalize all the droughts 
 and floods to which the country was 
 subject. Occasionally, but not very fre- 
 quently, there might be a great excess 
 of downfall, resulting in floods as large 
 as three or four hundred times the mini- 
 mum volume." Now the minimum vol- 
 ume is only about an 18th or 20th part 
 of the mean volume, so it follows, that 
 the floods are only 15 or 20 times the 
 mean volume. 
 
 Now with regard to compensation. 
 It is necessary in many instances to 
 compensate owners, mill- owners, and 
 others, people who are interested in the 
 streams that you are going to impound, 
 and, on an average, it is found that in 
 England one-third of the amount of 
 water requires to be given as compensa- 
 tion to these people, and, therefore, 
 
82 
 
 two- thirds remain for the use of the 
 town. This compensation, of course, 
 must be considered in determining the 
 size of the reservoir. Sometimes it has 
 been arranged that the amount given to 
 the owners on the banks should be the 
 average summer discharge, minus the 
 floods, and sometimes special compensa- 
 tion reservoirs have been built to collect 
 the water from a certain portion of the 
 drainage area, these compensation reser- 
 voirs being entirely under the control of 
 the persons who are to be compensated. 
 However, you may take it as an aver- 
 age, that about one-third in England 
 generally goes to them. 
 
 The culverts have been commonly 
 built through the embankment in the 
 made earth. This is stated to be a bad 
 plan. Mr. Rawlinson says they should 
 always be built in the rock or in the 
 solid ground, and not in the made earth. 
 The water tower is generally built just 
 inside of the embankment, and the dis- 
 charge or outlet pipes open into it with 
 valves, which valves ought to be inside 
 
83 
 
 the embankment, and not outside of it. 
 What are called " separating weirs ' 
 have been constructed in some reser- 
 voirs. They are ingenious contrivances 
 by which the water, when at its ordinary 
 height, flows over the weir into the cul- 
 vert to be taken away to the town. 
 When it is in flood, the force with which 
 it comes enables it to pass over the 
 opening leading to the culvert, and to 
 get away into the old watercourse. 
 "Feeders" for divtrting streams into 
 the reservoir are also sometimes necessa- 
 ry. It is often found to be necessary to 
 cut a new course for the stream that 
 runs down the valley, especially if it be 
 a very large stream, or if it be a stream 
 that is liable to floods. 
 
 I see that I forgot to mention one 
 point, which I should have stated at the 
 beginning of the lecture, with regard to 
 the situation of these reservoirs. The site 
 must not be too low, for if it is, the res- 
 ervoir is necessarily too shallow, and 
 shallow reservoirs are very bad, in that 
 the water cannot possibly be kept pure, 
 
84 
 
 it being perfectly impossible to store it 
 and keep it pure in shallow reservoirs. 
 If the ground is too high, and no other 
 suitable place can be got, then it is nec- 
 essary to make what are called " balanc- 
 ing reservoirs," so that the force of the 
 water may be broken by its being kept 
 in a series of reservoirs at different levels. 
 
 I do not profess to have given you the 
 engineering details, as you will plainly 
 see. All I have tried to do is to give 
 you some of the most important points, 
 according to the best authorities that I 
 have been able to find. 
 
 The channels are generally made of 
 masonry or brickwork. The water-way 
 is, according to Rankine, best semi-cir- 
 cular, or a half square, or a half hexa- 
 gon. These channels are usually made 
 cylindrical ; they require ventilating 
 shafts after the custom of the Romans. 
 Occasionally they are made with an 
 egg-shaped section, like large sewers. 
 
 Channels require to be curved at their 
 junctions, or at any rate they require to 
 be joined at very acute angles. 
 
85 
 
 With regard to aqueducts, Mr. Kaw- 
 linson tells us that "aqueducts of iron 
 will probably be cheaper than masonry 
 or brickwork constructions. " They have 
 been made self-supporting by Mr. Simp- 
 son, by constructing them in the form 
 of tubular iron girders. 
 
 Now, with regard to the fall of these 
 channels, I gave you one or two points 
 before, when considering the pipes con- 
 veying the streams. In the discussion 
 on the water supply of Paris, in the 
 25th Volume of the Proceedings of the 
 Institute of Civil Engineers, Mr. Bate- 
 man gave the following example with 
 reference to the Loch Katrine aqueduct 
 of the Glasgow Water Works: "The 
 fall was 10 inches to the mile through- 
 out, except where the water was carried 
 by syphon pipes across deep valleys, 
 which, in one instance of a hollow of 
 250 feet, was done for a distance of 3^ 
 miles, and in these cases there was a fall 
 of 5 feet per mile, to economize the size 
 of the pipes." 
 
 This aqueduct, I believe, is about the 
 
86 
 
 largest that has been constructed. The 
 channel is cylindrical, and about 8 feet 
 in diameter. Mr. Rawlinson said in the 
 same discussion, that "the fall of an 
 aqueduct must be in proportion to the 
 depth and volume of water which it had 
 to deliver. The fall of the New "River 
 in London was 1 in 10,000, or 6 inches 
 to the mile, but with so large a volume, 
 and an unpaved channel, it was necessary 
 to form a weir, and give the water a 
 vertical fall of a few inches at certain 
 points of its course. He found that 
 plan was adopted in the East. In laying 
 out a line of aqueduct two principles 
 were involved. If it were graded, as 
 the Romans graded some of theirs, from 
 5 to 15 feet per mile, there would be 
 difficulty in stopping the water at any 
 point. It was practicable, however, to 
 grade an aqueduct^ having a fall of 15 
 feet or 20 feet per mile, if vertical falls 
 were introduced at intervals, alternately 
 with level or nearly level lengths. This 
 mode enabled an engineer to fix the 
 velocity, so as to prevent undue wash- 
 
87 
 
 ing. The vertical falls tended to aerate 
 the water, and this in itself constituted 
 an additional advantage. All covered 
 aqueduct conduits should be abundantly 
 ventilated, and there should be side en- 
 trances, stop gates, overflows and wash- 
 out valves." Sometimes in aqueduct 
 bridges the sectional area of the channel 
 is diminished, and the gradient made 
 steeper. This, of course, gives greater 
 velocity to the water, and a smaller 
 amount of material is required, and so 
 less expense incurred in constructing the 
 bridges. So much as to the masonry. 
 
 Now as to pipes. Earthenware pipes 
 are made up to about 3 feet in diameter. 
 If they are of compact glazed earthen- 
 ware, they are very tough and strong, 
 but they will not bear shocks, either the 
 shocks of water or v anything else, and 
 they cannot be jointed so as to resist a 
 great pressure, and so are not suitable 
 for syphons. We will not say anything 
 more about lead pipes, because they are 
 not now used for this purpose. Cast 
 iron pipes, Kankine says, should be of a 
 
88 
 
 uniform thickness; and he lays down the 
 following rule for the minimum thick- 
 ness : " The thickness of a cast iron pipe 
 is never to be less than a mean propor- 
 tioned between its internal diameter and 
 one forty eighth of an inch." But, he 
 adds, " it is very seldom indeed that a 
 less thickness than f of an inch is used 
 for any pipe, how small soever." Large 
 cast iron pipes are liable to burst, and 
 there are some instances on record of it; 
 one in the water-works for the supply 
 of Melbourne, which I have already 
 mentioned once or twice, in which case 
 the pipe was 33 inches in width, was 
 laid through the embankment of the res- 
 ervoir and burst. Now this is what Mr. 
 Hawkesley said in a discussion on the 
 subject at the Institution of Civil En- 
 gineers, about the bursting of cast iron 
 pipes : "Cast iron in the shape of a 
 pipe would stand little unequal pressure 
 externally, although such a pipe would 
 bear an enormous pressure when equally 
 distributed, whether applied externally 
 or internally, and most in the former 
 
case, as the metal then would be under 
 compression," and he went on to say 
 that, at u the Rivington Pike reservoir 
 of the Liverpool water-works, two lines 
 of pipes were carried through an em- 
 bankment 20 feet high, at a distance of 
 16 feet from the top of it. They were 
 cast iron pipes, each pipe being made in 
 10 or 12 pieces, and they are the largest 
 pipes that have been laid, each pipe 
 being 44 inches in diameter. Now, out 
 of these two lines of pipes, fully one- 
 third of the pipes so placed, which were 
 excellent castings, were broken, although 
 they had borne a pressure of 300 feet in- 
 ternally. The fractures invariably oc- 
 curred at the top and bottom, and not at 
 the two sides, as might have been expect- 
 ed. The pipes being flattened and dis- 
 torted by the pressdre of the earth, were 
 subjected to a strain at the top and bot- 
 tom greater than at the sides, and were 
 undoubtedly broken by compression. 
 This fact convinced him that pipes in 
 that position were very insecure. Com- 
 monly, in similar cases, there was a 
 
90 
 
 pressure of water on the inside, and a 
 pressure of earth on the outside ; and it 
 was a usual arrangement for the valve 
 which shut off the water to be placed 
 under the embankment" (that is a point 
 I have referred to as one of considerable 
 importance), " so that if a pipe became 
 ruptured when in use the water would 
 escape into the embankment, and if it 
 found its way to the back of the puddle, 
 the embankment would be torn down, 
 and the whole of the water in the reser- 
 voir set free. It was not, therefore, de- 
 sirable that large pipes should be laid 
 under an embankment, where they would 
 be subject to a considerable pressure of 
 earth." 
 
 When a pipe of that magnitude breaks 
 it usually does great damage. In one of 
 these pipes that I have just mentioned 
 to you, sixteen million gallons of water 
 were capable of being discharged daily, 
 and if an accident occurred there would 
 be a column of 44 inches in diameter, 
 acting with perhaps 200 or 300 feet of 
 pressure to be dealt with. Another 
 
91 
 
 thing about these large pipes is, that 
 there is a considerable difficulty in re- 
 pairing them. One length of these 
 weighs about 4 tons, so that they cannot 
 easily be dragged about or taken up. 
 Now, cast iron pipes are said often to 
 break from the pressure of the air. 
 Whenever air gets driven in along with 
 the water, and especially so in syphons 
 where valleys are crossed, these pipes 
 are broken (it is said) by the collection 
 of compressed air. 
 
 Mr. Hawkesley tells us that he con- 
 siders that they are broken when the air 
 is let out ; that it is the shock caused by 
 the running together of the two sepa- 
 rated parts of the water that causes the 
 breakage of these pipes, when the com- 
 pressed air that is collected in them is 
 let out too suddenly; and he recom- 
 mends, and has practiced in the case of 
 those large mains at the Liverpool water- 
 works, the adoption of valves with an 
 aperture of only f of an inch ; through 
 these the air rushes out, but they do not 
 permit the columns of water to come 
 
92 
 
 together very suddenly ; there should be 
 one of these at each place throughout 
 the channel where the pipe is higher 
 than the theoretical line, or than the 
 line of the fall. At each one of these 
 places air is liable to accumulate and to 
 become compressed, and, perhaps, to 
 burst the pipe. At each one of these 
 places, therefore, there should be means 
 of letting out the compressed air, and 
 even with regard to this precaution we 
 were, as I showed you before, forestalled 
 in the aqueducts of the ancients. 
 
 Pipes are also sometimes burst by the 
 pressure of the water when a valve on 
 the main is closed ; this difficulty has 
 been overcome by a plan mentioned by 
 Mr. H. Maudslay at the Institute of 
 Civil Engineers : " In some instances 
 there had been a small valve and pipe, 
 so placed at the side of the large main 
 as to join the main both before and be- 
 yond the large valve, in order that the 
 whole body of water might not act like 
 a water-ram on the closing of the large 
 valve. This plan has been adopted in 
 
93 
 
 the Neptune fountain at Versailles, and 
 also, he believed, in the mains supplying 
 the fountains at the Crystal Palace. On 
 shutting the large valve, the main flow 
 was stopped, but the small pipe permit- 
 ted a continuous flow of the smaller 
 quantity, and thus the danger of burst- 
 ing was avoided. The second valve was 
 afterwards closed gradually. He thought 
 that this was the most simple plan that 
 could be adopted, and perhaps the least 
 costly, while it was certainly very effect- 
 ive." Another plan is that described by 
 Mr. Hawkesley, as follows : " The valve 
 upon the main at Liverpool was divided 
 into three openings, each of which was 
 provided with a separate screen, so that 
 by raising or lowering each of these 
 slowly in succession the water was either 
 admitted or turned off very gradually. 
 The object of dividing the valve into 
 three apertures was to enable a work- 
 man to operate with facility on any one 
 of the screws. In large pipes, where the 
 pressure was great, it was necessary, in 
 order that the brass pieces, upon which 
 
94 
 
 the valve acted, might not be abraded, 
 that only a certain amount of pressure 
 should be put upon them, and that the 
 friction under that pressure should not 
 be greater than a man could overcome, 
 by simply turning a handle, without 
 stripping the thread of the screw. As a 
 further provision the center valve was 
 made very narrow ; the side valves were 
 first closed and then the center one, so 
 that concussion was prevented. In ad- 
 dition there were branches at various 
 points, upon which equilibrium valves, 
 with a piston underneath, were placed, 
 and others had double beat valves. But 
 as these valves required to be heavily 
 weighted, the inertia of the weight 
 -would, if other means were not taken, 
 prevent the valve from rising so rapidly 
 as was desirable. Therefore, between 
 the weight and the valve there was a 
 spring, the action of which was inde- 
 pendent either of the valve or of the 
 weight, so that, instead of the valve 
 waiting for the large weight to rise, the 
 spring immediately yielded under it and 
 
95 
 
 the water was discharged instantaneous- 
 ly. When these valves were used not 
 the slightest shock was experienced. If 
 there had been, the pipe would undoubt- 
 edly have been ruptured, for the length 
 of the column, and the velocity, upon 
 which the force of concussion was de- 
 pendent, were both very great. That 
 was another reason he preferred a small- 
 er pipe. There were still, however, other 
 precautions. Powerful disk- valves, made 
 by Sir W. Armstrong & Co., and which 
 acted in a similar way to the cataract 
 apparatus of a small power steam engine 
 were placed upon the main. They were 
 made to close slowly, being let go by a 
 trigger. As a hundred million gallons 
 might pass through the main in twenty- 
 four hours, if a pipe burst, without any 
 provision being made to stop the flow, a 
 great deal of mischief would ensue. 
 Supposing, however, a fracture to occur 
 when the disk valve was open, then the 
 valve would gently close in about 
 two minutes, and arrest the discharge. 
 These valves cost 300 each. He had 
 
96 
 
 found them to act, on various occasions, 
 extremely well, and but for them the 
 country would have been flooded on 
 several occasions." 
 
 Now, we have considered the Roman 
 plan and also the plan of collecting 
 water by drainage areas into large im- 
 pounding reservoirs and conveying it by 
 channels to the place that wants it, the 
 place where it is to be distributed. 
 When it comes there, it is collected in 
 what are called service reservoirs. The 
 most ancient examples of these service res- 
 ervoirs are those very piscinae, upon the 
 Roman aqueducts, which I have spoken 
 of, and you can see examples of them in 
 Rome at the present day. The best I 
 ever saw was at a place called Bona in 
 Algeria, where is to be seen a set of the 
 most magnificent service reservoirs. The 
 plan was to have four compartments. 
 The water was first let into one of the 
 two upper ones ; it then fell from that 
 into one below, possibly over a waste- 
 pipe. The water then passed, possibly 
 through strainers, into another compart- 
 
97 
 
 ment on the same level, and it then rose 
 through the roof of that compart- 
 ment into a third at the level of 
 the first one, out of which it went 
 onwards, and considerable settling 
 took place. Now, there were means 
 of scouring out these two lower 
 compartments, which could be shut off 
 from the upper ones so that the mud 
 might be got out of them. The water, 
 when it is brought to these reservoirs by 
 either of these two methods, or when it 
 is got into them, as it very often is now 
 for the supply of large towns, directly 
 out of the river, very often requires to 
 be filtered, as mere settling is not enough 
 for it. We have then to consider what 
 materials are used for filtering the water, 
 what size the filter beds require to be, 
 and what effect is produced on water by 
 filtration. 
 
 Now, in the first place, the materials 
 that are commonly used for filtration are 
 sand and gravel. The different merits 
 of sand and gravel and also of charcoal 
 I shall have to consider in the next 
 
 
98 
 
 lecture, but I must conclude this lecture 
 by telling you that the effect of filtration 
 of water, even by sand and gravel, is not 
 merely the mechanical effect of remov- 
 ing the suspended substances that the 
 water may contain, but that, at the 
 same time, there is a chemical action 
 going on. This is on account of the air 
 that is contained between the little par- 
 ticles of sand, which air is so brought 
 into contact with the finely divided 
 water that any substances in the water 
 that are capable of oxydation do become 
 oxydized, and a considerable amount of 
 the organic matters in the water are thus 
 oxydized, and transformed into innocuous 
 matters. That is the first important point 
 to understand about filters, whether in 
 filtering water for drinking purposes, or 
 with regard to a filter about which we 
 shall have to say more after a while a 
 filter to purify sewer water. 
 
 I have shown you that it was a fallacy 
 to suppose that the Romans did not 
 understand the principle of the syphon, 
 but that they constructed most admir- 
 

 99 
 
 able ones on the aqueducts that brought 
 water to Lyons. It so happens, by a 
 curious chance, that I have recently 
 seen some plans and sections of the 
 Koman aqueducts which supplied Jeru- 
 salem with water, and on one of those I 
 find a syphon, not made with lead pipes, 
 but a syphon made of stone. It is made 
 of blocks of stone with a hole through 
 each; the blocks are put together so as 
 to form a continuous pipe. Each piece 
 is cut at the end so that around the pipe 
 itself, the aperture in the stone, there is 
 a ring left projecting on the face of the 
 stone, and that ring fits into a groove on 
 the next stone. That made a sufficiently 
 tight syphon to convey the water, with- 
 out any great amount of leakage, to a 
 considerable vertical depth and up again. 
 The depth, as far as I can judge from 
 the plans, is about 100 feet from the 
 highest point to the lowest. Well, now, 
 we get up to the point where the water 
 has reached the town, and there I told 
 you it is almost necessary, certainly 
 usual, to construct a service' reservoir. 
 
100 
 
 The Romans constructed them under the 
 name of piscinae ; and I told you, I 
 think, in two words, how those were 
 made ; I now want to give you a rather 
 longer account of their construction. 
 The water that was brought by one of 
 the aqueducts to Rome was taken direct 
 from the river Anio, and the result of 
 taking the water direct from the river 
 was that after the heavy rains it was 
 charged with mud, and though large 
 cisterns were provided, in which, by an 
 ingenious arrangement, much of the 
 sediment was caught, still it was not 
 considered satisfactory by Frontinus, 
 who was the engineer, and who, there- 
 fore, under his patron the Emperor 
 Nerva, altered the source. Still the 
 water that came to Rome required to 
 have settling tanks, as described by Mr. 
 Parker in a paper I quoted before, and 
 from which I again quote : 
 
 " The building consisted of four cham- 
 bers two beneath and two above. Sup- 
 posing, for the sake of illustration and 
 in the absence of a diagram, the letters 
 
101 
 
 , ' 
 
 represent the fou'r chambers'. The 
 
 CD.' v ' " . '.,'.'.' 
 
 channel of the aq'jectoftfc; . ceding " 'fr'otn". 
 the east, at a tolerably high level enters 
 the chamber B. Thence the water 
 passed (possibly over a large waste pipe) 
 into the chamber beneath, D. Between 
 D and C there were communications 
 through the wall (possibly provided 
 with fine grating). Through the roof 
 of C there was a hole, and the water 
 passed upwards, of course, finding the 
 same level in A as in B, whence it was 
 carried off into another stream. By the 
 aid of sluice gates the water could be 
 transferred direct from chamber B to 
 chamber A, and access was obtained by 
 an opening to the chambers beneath, 
 and the mud was from time to time 
 cleared out." 
 
 Just the same thing was the case at 
 Lugudunum (Lyons). Large settling tanks 
 have been found on the hill of Four- 
 vieres, consisting of two reservoirs with 
 vaulted roofs, thus described : One of 
 them was 48 feet long by 44 feet broad, 
 
102 
 
 , and 20 Test high, with two conduits to 
 ^ admit the water, and several round holes 
 -<fn j:ht- roof frprh t which it could be 
 drawn. The walls were 3 feet thick, 
 lined with very hard cement. A second 
 was 100 feet long, 12 feet broad, and 15 
 feet high, divided by a wall into two 
 chambers. A third was a large one, 
 of which five of the supporting arches 
 remain, and the discharge conduit, 1^ 
 feet broad, which distributed the water, 
 by means of leaden pipes (of which a 
 specimen has been found) to the palaces, 
 gardens, etc. In some cases similar con- 
 structions formed public reservoirs from 
 which the people drew the water. In Rome 
 " there were 591 open reservoirs (lacus) 
 for the service of all comers. * * * 
 These reservoirs were what we usually 
 speak of as fountains ; and some hun- 
 dreds are in use to this day, many proba- 
 bly on the site of the older ones. There 
 were very stringent laws respecting 
 their use. Heavy penalties were inflicted 
 upon any one dipping a dirty bucket or 
 vessel into the reservoir. There were 
 
103 
 
 also laws respecting the 'overflow,' as 
 the fountains, of course, were constantly 
 running ; these were the most important 
 to keep in order, as all the poorer classes 
 depended entirely upon them for their 
 supply of water." 
 
 Now let us consider the Service Res- 
 ervoirs as they are made now. Service 
 reservoirs must either be placed at a 
 low level, so that the water has to be 
 pumped from them, or high up, which is 
 better, so that the water, if not brought 
 to them at that level, is pumped into 
 them, as at Lyons, on the Rhone, where 
 the water is brought to them at the 
 highest point. They are made to con- 
 tain a few days' supply. In the first 
 place, they must always be covered ; 
 even the Roman ones were. The reason 
 of their being covered is that, if not 
 covered, the water becomes impure, for 
 the impurities of the air dissolve in the 
 water, and the growth of confervse is 
 also, of course, very much aided by 
 light. If they are at the level of the 
 ground, they are built of masonry. 
 
104 
 
 Mr. Rawlinson says, " The ground ex- 
 cavated for the foundation of a tank 
 should be made perfectly water-tight. 
 The bottom may be covered with clay 
 puddle and the side walls be backed or 
 lined with clay puddle. The thickness 
 of the puddle should not be less than 12 
 inches. If the site selected for a tank is 
 sand, gravel, or open jointed rock, great 
 care must be taken to give the puddle a 
 full and even bearing over the whole 
 surface area ; open joints in rock must 
 be cleaned out and then filled up with 
 concrete. In gravel, large stones must 
 be removed and the entire surface 
 brought to a level, smooth, and even 
 plain. Clay puddle will only resist the 
 pressure of water when it rests solidly 
 on an even bed, so as to prevent the 
 water forcing holes through it, which 
 wil] be the case if there is a rough, 
 uneven surface and open space beneath."* 
 
 The roof is supported on piers with 
 arches between the n, and across some- 
 
 * Suggestions as to the preparation of plans as to 
 Main Sewerage and Drainage and as to Water Supply. 
 
105 
 
 times iron columns are placed in rows 
 supporting the girders which carry the 
 arches. The supply pipe has one or 
 more exits, a v aste and a wash-out, 
 which may be connected by valves so 
 that the supply can be directly connect- 
 ed with the exit independently of the 
 tank. 
 
 The water is received in a sort of well 
 or tower through which it passes into 
 the tank, and after settling has taken 
 place it passes out through a valve into 
 the exit pipe. When the supply is too 
 great it is carried off by an overflow to 
 which the wash-out pipe may be jointed. 
 
 Well now I should like to give you a 
 more detailed description of such reser- 
 voirs, and I take as an instance, and 
 that for several reasons, the description 
 of some reservoirs with supply tanks : 
 
 "The reservoir of Passy is intended 
 to receive the waters pumped from the 
 Seine at Chaillot, and those furnished 
 by the Artesian well of Passy when 
 disposable ; it is composed of three com- 
 partments, two of which are covered by 
 
106 
 
 a second range of arches, the third, in- 
 tended as a reserve in case of fire, being 
 deeper than the rest, and only of one 
 story ; the two upper ranges of arches, 
 also, are to be made to hold a supply of 
 water, one- of them being covered and 
 the other not. The united capacity of 
 these various compartments is 9.227,097 
 gallons, and their levels above the Seine 
 are respectively arranged at 150 feet, 
 and 163 feet, above zero of the scale of 
 the bridge of la Tournelle. The capacity 
 of the separate reservoirs is, for those 
 nearest to the ground, respectively 
 2,232,800 and 2,344,984 gallons; these 
 are covered with reservoirs of the ca- 
 pacity of 1,282,792 and 1,495,729 gal- 
 lons ; and the uncovered side portions 
 of the reservoir are devoted to the re- 
 maining 870,792 gallons. These build- 
 ings are formed on the ' tuf du calcaire 
 lacustre,' which afforded a hard, resist- 
 ing foundation, and did not require any 
 particular precautions to prevent the 
 subsidence of the piers, or to secure the 
 water tightness, or the impermeability 
 
107 
 
 of the bottom. The external walls have 
 been in consequence carried down to the 
 depth of 8 ft. 4 in. and have a width of 
 8 ft. 8 in. all around. The floors are of 
 masonry, 1 foot thick in meuliere and 
 cement, covered with a rendering coat of 
 1 inches of the same cement worked to 
 a fine face. This is covered with a 
 range of cylindrical vaults of 10 feet 
 opening, springing from pillars 2 ft. 8 
 in. square - upon the top, gradually en- 
 larging to 5 ft. at the bottom. It is cal- 
 culated that in no case does the weight 
 brought upon a square inch of this 
 masonry exceed 152 Ibs. The thickness 
 of the arch forming the roof of the first 
 tier, and the floor of the second division, 
 is about 1 ft. 2 in. on the crown ; that of 
 the roof of the upper division is only 4J 
 inches, executed in two courses of tiles 
 bedded in cement, and 'rendered' with a 
 coating of that material and covered with 
 concrete. 
 
 <l The reservoirs of Menilmontant are 
 considerably larger than those of Passy, 
 and being founded upon the upper mem- 
 
108 
 
 bers of the Paris Basin, special precau- 
 tions were required to insure that the 
 ground should not yield under the com- 
 bined pressure of the masonry, and the 
 29^ million gallons of water intended to 
 be stored. The marls covering the gyp- 
 sum of which the mountain of Menil- 
 montant is composed, were not consid- 
 ered to be able to withstand that weight. 
 The foundations of the piers were there- 
 fore carried lower down, and thence 
 built in a description of rough rubble of 
 menliere set in hydraulic lime. The 
 bottom floor of the reservoir is arched 
 over these piers, and the upper tier of 
 arches rests upon this floor.'' 
 
 It is only fair to tell you that some 
 engineers, and among others, Mr. Raw- 
 linson, considered the plan of building 
 two-storied reservoirs as a bad one, and 
 not to be imitated ; but it is necessary to 
 know that there is such a plan, and the 
 description applies, to a great extent, to 
 all reservoirs. 
 
 To take an example nearer us, there is 
 Mr. Simpson's elevated reservoir on Put- 
 
109 
 
 ney Heath ; that contains ten millions of 
 gallons altogether. There is there a 
 double covered reservoir to contain 
 filtered water for domestic use, and a 
 smaller open one to contain unfiltered 
 water for the streets, and to supply the 
 Serpentine, and so on. So that you see 
 it is usual in these cases to build several 
 reservoirs together. This covered reser- 
 voir that has to contain water for do- 
 mestic purposes, is double, or constructed 
 in two halves. Each part has an area of 
 310 feet by 160 feet, and a depth of 20 
 feet. The sides all round have a slope 
 of one to one. This gives a mean area 
 of 290 feet by 140 feet, and a capacity 
 of about 5,075,000 gallons for each res- 
 ervoir, exclusive of the space occupied 
 by the piers. "Hence the whole ca- 
 pacity may be taken " as stated by Mr. 
 Simpson in his evidence, " at 10,000,000 
 gallons. The sides of the reservoir are 
 cut out in the form of steps, which are 
 filled up with concrete to a uniform 
 slope of one to one ; and a bed of con- 
 crete one foot in thickness is also laid 
 
110 
 
 over the whole bottom ; each half of the 
 reservoir is covered with eight brick 
 arches, averaging rather less than 20 
 feet span, the arches being each 20 feet 
 span, and the others 18 ft. 8 in. Two 
 piers supporting these arches are built 
 lengthways, and are each 310 feet long 
 at the top, and 270 feet at the base. 
 The arches are each one brick in thick- 
 ness, and are covered over with a layer 
 of puddle, the haunches being filled up 
 with concrete. The piers are carried 
 out 14 inches thick; but the division 
 wall between the two parts of the reser- 
 voir is rather more than four feet thick, 
 with a concrete slope of one and a half 
 to one on each side. The 14-inch piers 
 supporting the arch are built with large 
 circular hollows 17 J feet diameter. The 
 centers of these circular hollows are 40 
 feet apart, so that solid brickwork 23 
 feet long is left between the circular 
 hollows, supposing a horizontal section 
 taken through the centers of the hollows. 
 Each of the 23 feet spaces has a 14-inch 
 counterfort carried out at right angles. 
 
Ill 
 
 These counterforts occur at intervals of 
 26 feet and 13 feet alternately, and pro- 
 ject 6 feet wide at the base, on each side 
 of the pier, and run out to nothing at 
 the top, or springing of the arches." 
 " The versed sine or rise of the arches is 
 5 ft. 3 in., or rather more than one-fifth 
 of the span. Each arch is provided 
 with two openings in the center, com- 
 municating with a line of 12 inch earth- 
 enware tubular pipe, which passes 
 through the spandrels and communicates 
 with perforated iron tops in the division 
 wall between the two parts of the reser- 
 voir. By this contrivance the space 
 above the water in the covered reser- 
 voirs is effectually ventilated. The sup- 
 ply pipe from Thames Ditton is 30 in. in 
 diameter and comes into each part of 
 the reservoir at the level of top water, 
 which is a few inches below the spring- 
 ing of the arches. At this level a waste 
 weir, or overflow, is fixed to prevent the 
 reservoir being filled too full. The exit 
 mains to London consist of two 24-inch 
 pipes, and they pass from the bottom of 
 
112 
 
 the reservoir, which has an inclination 
 in one direction of 1 in 20, and a fall 
 across of six inches." * 
 
 Now a great reason for the existence 
 of these service reservoirs is, that the 
 hourly demand during the day varies 
 very much from the mean. It is some- 
 times so much as three times the mean 
 demand, during certain hours, so that by 
 this means it is not necessary for the 
 mains to be made inordinately large. 
 But otherwise the mains would have to 
 be made large enough to give the great- 
 est demand, instead of being only suffi- 
 cient for the mean demand. And this is 
 the case if the reservoir is only large 
 enough to contain half the daily demand. 
 In that case the distributing pipes need 
 only to be calculated to give the greatest 
 hourly demand. These you will recol- 
 lect are underground tanks. Elevated 
 tanks are sometimes made of cast iron, 
 or wrought iron plates bolted together, 
 and tied by wrought iron rods at the bot- 
 tom, to one another. The supply, exit, 
 and overflow pipes ought to be together 
 
 (*Hughes' "Waterworks.") 
 
113 
 
 in a corner of the reservoir, in a small 
 separate compartment. This separate 
 compartment is connected with the main 
 reservoir by a valve, so that the main 
 reservoir can be cleaned out and the 
 supply go on independently of it. Thus 
 you can shut out the supply, stop pump- 
 ing, open a valve, and let out all the 
 water from the large reservoir by the 
 supply pipe to the town. Then you can 
 close the valve and let the supply go on 
 through this little separate reservoir, 
 while the other is being mended or 
 cleaned* out. 
 
 The overflow pipe, or waste pipe, or 
 whatever you like to call it, ought to 
 open into an open channel, and not be 
 connected, as is very frequently the case, 
 with the nearest drain or sewer. It 
 ought to open above ground, because as 
 the reservoir is covered, if it does not do 
 so, the foul air from the drain will come 
 up that waste pipe, and be dissolved by 
 the water in the cistern, and so you will 
 render the water that you have taken so 
 much trouble to get pure, you will ren- 
 
114 
 
 der it impure, and that is what is con- 
 tinually done in all towns, and in houses, 
 as I shall tell you presently when speak- 
 ing of sewerage. 
 
 For distributing basins or tanks, Ran- 
 kine says that "the most efficient pro- 
 tection against heat and frost is that 
 given by a vaulted roof of masonry, or 
 brick, covered with asphaltic-concrete, to 
 exclude surface water, and with two or 
 three feet of soil, and a layer of turf.' r 
 Mr. Bawlinson says that " brick and 
 masonry tanks, if arched, may be cover- 
 ed in with sand, or fine earth, to the 
 depth of 18 inches, which will preserve 
 the water cool." 
 
 Up to the present time we have been 
 describing works connected with im- 
 pounding reservoirs ; now, with regard 
 to river works. With river works you 
 still more certainly require settling res- 
 ervoirs into which water may either flow 
 directly through culverts from the river, 
 as it does at Chelsea, or into which it 
 may be pumped. When the water flows 
 in from culverts, you require almost in- 
 
115 
 
 variably to have filter beds, which we 
 shall describe a little further on. Some- 
 times for river works it is necessary to 
 construct a weir right across the river, 
 in order to keep the water as near as 
 may be at constant level. The engine 
 power employed ought to be consider- 
 ably greater than that which is actually 
 wanted, one-third greater at any rate, 
 and, of course, there ought always to be 
 a reserve engine. At the Chelsea works, 
 to which I have before referred, the de- 
 positing reservoirs are made in London 
 clay, and the bottom and sides are mere- 
 ly lined with cement placed upon this 
 clay. From this the water passes direct 
 to the filter beds. 
 
 With regard to those cases in which 
 the water is taken from rivers, there are 
 certain things I want to tell you. I 
 want to tell you something about the 
 purification of river water. We know 
 that into rivers, especially in thickly 
 populated countries, an enormous 
 amount of refuse matter of all sorts is 
 thrown, and it is necessary to know 
 
116 
 
 whether this refuse matter is destroyed 
 in its passage along the rivers, that is 
 to say, whether the water, after running a 
 certain distance, becomes sufficiently pure 
 to be used for drinking. And now I 
 must quote to you from a book from 
 which I shall have occasion to quote a 
 great many times during the course of 
 the remaining lectures, a book entitled 
 "A Digest of Facts relating to the 
 Treatment and Utilization of Sewage." 
 
 " The evidence collected on this head 
 by the Koyal Commission on Water 
 Supply was very various. Dr. Frank- 
 land says: 
 
 c There is no process practicable on a 
 large scale by which that noxious mate- 
 rial (sewage matter) can be removed 
 from water once so contaminated, and 
 therefore I am of opinion that water 
 which has been once contaminated by 
 sewage or manure matter is henceforth 
 unsuitable for domestic use.' " 
 
 Now the results of experiments are 
 found to give the following facts : In 
 the first place it appears that in rivers 
 
117 
 
 that are well known to be polluted, and 
 the water of which has a temperature 
 not exceeding 64 Fahrenheit, a flow of 
 between eleven and thirteen miles " pro- 
 duces but little effect upon the organic 
 matter dissolved in the water." To re- 
 move all uncertainty from the " varia- 
 bility of the composition of the river 
 waters at different times of the day," ex- 
 periments were made by mixing filtered 
 London sewage with water ; 6i it was 
 then well agitated and freely exposed to 
 the air and light every day, by being 
 syphoned in a slender stream from one 
 vessel to another, ' falling each time 
 through three feet of air." The mixture 
 which originally contained in 100,000 
 parts .267 of organic carbon and .081 of 
 organic nitrogen was found to contain, 
 after 96 hours, .250 of organic carbon 
 and .058 of organic nitrogen ; and after 
 192 hours, .2 of organic carbon and 
 .054 of organic nitrogen. The tempera- 
 ture of the air during this experiment 
 was about 20 deg. Cent. (68 Fahrenheit). 
 "These results indicate approximately 
 
118 
 
 the effect which would be produced by 
 the flow of a stream containing 10 per 
 cent, of sewage for 96 and 192 miles 
 respectively, at the rate of one mile per 
 hour." They show, then, that at the 
 above temperature, during a flow of 96 
 miles, at the rate of one mile an hour, 
 the amount of organic carbon was re- 
 duced 6.4 per cent., that of organic 
 nitrogen 28. 4 per cent.; while during the 
 flow of 192 miles, at the same rate, 
 the amounts of these two substances 
 were only reduced 25.1 and 83.3 percent, 
 respectively. It is shown that the oxy- 
 dation of this organic matter is chiefly 
 affected by the amount of atmospheric 
 oxygen dissolved in the water, "such 
 dissolved oxygen being well known to 
 be chemically much more active than the 
 gaseous oxygen of the air." 
 
 It was found, however, that the action 
 of this dissolved oxygen was not really 
 anything like so quick or so perfect as 
 generally supposed, and that 62 per cent, 
 of the sewage was the maximum quan- 
 tity that would be oxydized during 168 
 
119 
 
 hours, even supposing that the oxydation 
 took place during the whole time at the 
 maximum rate observed, which was cer- 
 tainly not the case. 
 
 ; " It is thus evident, that so for from 
 sewage mixed with 20 times its volume 
 of water being oxydized during a flow 
 of 10 or 12 miles, scarcely two-thirds of 
 it would be so destroyed in a flow of 168 
 miles at the rate of one mile per hour, or 
 after the lapse of a week. . . . . . 
 
 Thus, whether we examine the organic 
 pollution of a river at different points of 
 its flow, or the rate of disappearance of 
 the organic matter of sewage when the 
 latter is mixed with fresh water and 
 violently agitated in contact with air, or 
 finally the rate at which dissolved oxygen 
 disappears in water polluted with 5 per 
 cent, of sewage, we are led in each case 
 to the inevitable conclusion that the 
 oxydation of the organic matter in sew- 
 age proceeds with extreme slowness, 
 even when the sewage is mixed with a 
 large volume of unpolluted water, and 
 that it is impossible to say how far such 
 
120 
 
 water must flow before the sewage mat- 
 ter becomes thoroughly oxydized. It 
 will be safe to infer, however, from the 
 above results, that there is no river in 
 the United Kingdom long enough to ef- 
 fect the destruction of sewage by oxy- 
 dation. 
 
 Now there were several scientific men 
 who gave evidence of another sort, and 
 who declared that practically speaking 
 water was sufficiently pure after even a 
 short flow. The answer to that state- 
 ment is found if we just go into a few 
 of the public health facts. Here is one. 
 This is gathered from Mr. Simon's re- 
 port on the cholera epidemics of London 
 in 1848-49 and 1853-54. "When the 
 Lambeth Company took its water from 
 the Thames near Hungerford Bridge, 
 the people who drank that water died at 
 the rate of 12.5 per thousand. When 
 the source of supply was moved to the 
 Thames at Thames Ditton, the mortality 
 was only 3.7 per thousand, while at the 
 same time, and in the same districts, the 
 mortality among the people who were 
 
121 
 
 supplied with water by the Southwark 
 Company from the Thames at Battersea 
 was at the rate of 13 per thousand." 
 
 I could give you any number of facts 
 of that sort to show you that water that 
 has been polluted is dangerous to drink. 
 I may just mention to you the opinion 
 which Sir Benjamin Brodie, the late Pro- 
 fessor of Chemistry at Oxford, has 
 given ; he said in his evidence before 
 the Kivers' Pollution Commissioners : 
 "I believe that an infimtesiinally small 
 quantity of decayed matter is able to 
 produce an injurious effect upon health. 
 Therefore if a large proportion of or- 
 ganic matter was removed by the pro- 
 cess of oxydation the quantity left 
 might be quite sufficient to be inj urious 
 to health. With regard to the oxyda- 
 tion we know that to destroy organic 
 matter the most powerful oxydizing 
 agents are required; we must boil it 
 with nitric acid and chloric acid, and the 
 most perfect chemical agents. To think 
 to get rid of organic matter by exposure 
 to the air for a short time is absurd." 
 
122 
 
 I give you those statements in order 
 to bring you to the conclusion to which 
 I wish you to come, namely, that we 
 should not take water for the supply of 
 villages and towns from a river that has 
 been contaminated at all, if it can pos- 
 sibly be helped ; that it has never been 
 proved that such water gets really pure 
 again ; and that at certain times there- 
 fore very considerable danger may arise 
 from drinking such water ; in fact, as 
 Mr. Simon said when examined before 
 the Koyal Commission on Water Supply, 
 "it ought to be made an absolute con- 
 dition for a public water supply that 
 it should be uncontaminable by drain- 
 age." 
 
 The water when taken from the river, 
 or even if it is taken from the gentle 
 slopes of cultivated lands, and also in 
 some other instances, requires to be fil- 
 tered as well as allowed to settle ; de- 
 position is not sufficient of itself. It is 
 important, also, to keep out inferior 
 waters, that is when there are several 
 sources ; and with this condition you 
 
123 
 
 may prevent the necessity of the water 
 all requiring to be filtered. 
 
 Mr. Parker says: "At last it may be 
 interesting to know what Frontinus did, 
 or rather, what he says with becoming 
 modesty, his patron, the Emperor Nerva, 
 accomplished on this score : ' But the 
 water of the Anio Novus often spoilt the 
 rest, for since it was the highest* as to 
 level, and held the first rank as to abund- 
 ance, it was most often made use of to 
 help the others when they failed. The 
 stupidity, however, of the Aquarii was 
 such that they had introduced this 
 water into the channels of several others 
 where there was no need, and spoilt 
 water which was flowing in abundance 
 without it. This was the case especially 
 as regards the Claudian, which came all 
 the way for many miles in it6 own chan- 
 nel perfectly pure, but when it reached 
 Borne and was mixed with the Anio it 
 lost all its purity. And thus it happen- 
 ed that many were not in fact helped at 
 all by the addition of the extra water, 
 through the want of care on the part of 
 
124 
 
 those who distributed it. For instance, 
 we found even the Marcian, the most 
 pleasant to drink on account of its 
 brightness and freshness, in use in the 
 baths, and by the cloth-fullers, and ac- 
 cording to all accounts employed for the 
 most base services. It pleased, there- 
 fore, the Emperor to have all these sepa- 
 rated, and for each to be s,o arranged 
 that first of all the Marcian should be 
 assigned to its own use, so that the Anio 
 Vetus, which from various reasons was 
 found to be less wholesome, as well as 
 being at a low level, should be employed 
 for the watering of the gardens in the 
 suburbs, and in the city itself, for viler 
 purposes.'" 
 
 So you see they had, even then, found 
 out that one water was more wholesome 
 than another, and when they had got 
 supplies from two or three sources they 
 knew it was better to keep them sepa- 
 rate, and so use the best one for drink- 
 ing purposes and the inferior ones for 
 other purposes. 
 
 Now when water containing substances 
 
125 
 
 in suspension is passed through a medium 
 provided with fine pores, it is, of course, 
 at least the purer by virtue of the re- 
 moval of all such matters as are unable 
 to pass through the pores. If that were 
 all that filtration accomplished, it would 
 be only a fine straining process. But 
 that is not all. If you take a large 
 quantity of porous material, for instance, 
 a large mass of sand, or gravel, or espe- 
 cially charcoal almost any porous ma- 
 terial and pass water through it, water 
 containing certain substances in solution, 
 and certain substances in suspension, 
 those in suspension will remain unless 
 they are fine enough to pass through the 
 pores of the material. But all these 
 porous substances contain an immense 
 amount of air between their pores, and 
 the water ^y being passed through them 
 is divided into an infinite number of ex- 
 ceedingly small rivulets, exceedingly 
 small streams, and so the substances in 
 solution in the water are brought into 
 the closest possible contact with the 
 oxygen of the air between the pores of 
 
126 
 
 the filtering material, and so when you 
 have passed the water through a filter, 
 a chemical action takes place, and not 
 merely a mechanical action. You have 
 a mechanical action first, and then you 
 have also a chemical action. That chemi- 
 cal action consists in the oxydation of 
 the substances held in solution in the 
 water that is, such substances as are 
 capable of oxydation, and these are the 
 ammonia and the putrescible organic 
 matters which are so dangerous when 
 left in drinking waters. 
 
 One of the best filtering substances, 
 that is, one which alters the substances 
 contained in water most in its passage 
 through it, is animal charcoal, and you 
 will find in the 26th and 27th vols. of 
 the Proceedings of the Institution of 
 Civil Engineers, a most important and 
 interesting discussion on this property of 
 animal charcoal, and other substances 
 sand, and so on upon the power of 
 these materials to cause the oxydation 
 of substances in water. I should tell 
 you that the paper itself to which I re- 
 
127 
 
 fer in that 26th vol. is not worth read- 
 ing, but the discussion afterwards is very 
 well worth careful study. The paper is 
 worthless, because it came to an entirely 
 erroneous conclusion on account of the 
 experiments being performed by a pro- 
 cess which is practically worthless. 
 
 Here I must give you an example. 
 Dr. Frankland tells us that he filtered 
 New River water through animal char- 
 coal; that before filtration it contained 
 in solution about 18 grains in a gallon of 
 solid matters, that after filtration it con- 
 tained 11.6. Of course you are pre- 
 pared for a less amount of impurity after 
 filtration. Now the organic and other 
 volatile matters contained in the water 
 before filtration amounted to .37 of a 
 grain in a gallon, and after filtration the 
 amount was 15 ; that is to say that more 
 than one-half of these matters were re- 
 moved by filtration through animal char- 
 coal. After a month this charcoal re- 
 moved still more organic matter, and 
 some mineral matters as well, and even 
 a few months afterwards one-half of the 
 
128 
 
 organic and volatile matters 'only re- 
 mained after filtration. These experi- 
 ments show a very important thing, 
 which is perfectly true of a sand filter 
 as it is of an animal charcoal filter, and 
 that is, it is not by storing up these mat- 
 ters that a filter works, or else it would 
 be of no use whatever to make a filter. 
 You would have it choked up in a very 
 short time, and it would continually have 
 to be renewed, whether made of sand, of 
 gravel, of charcoal, or what not. It is 
 by oxydizing the substances that the 
 advantage is obtained, and the results of 
 oxydation you can find in the water 
 afterwards, and these results of the oxy- 
 dation are nitrates and nitrites, and car- 
 bonates. Of course these are harmless 
 matters, and that is the important action 
 which a filter has. Dr. Frankland stated 
 that he had passed the water supplied to 
 London by the Grand Junction Company 
 through a thickness of three feet of ani- 
 mal charcoal, at the rate of 41,000 gal- 
 lons per square foot per day of twenty- 
 four hours, under a head of water of 
 
129 
 
 thirty feet, the charcoal being in granules 
 like coarse sand, and that at that rate 
 a tremendous rate more than one-half 
 of the organic matter was removed. 
 He thought from these experiments on 
 animal charcoal that persons who had to 
 supply water to towns ought to use it, as 
 at any rate one of the media in the filter 
 beds. I must not pass from charcoal 
 without mentioning that vegetable char- 
 coal is agreed on nearly all hands to be 
 almost entirely useless for purposes of 
 filtration. In the first place it contains 
 enormous amounts of salts which are 
 soluble in water, so that the water be- 
 comes very much harder in passing 
 through it than before, and then it does 
 not purify water in the way that animal 
 charcoal does. 
 
 Well, now some of the effects of sand 
 filters, as employed by the Water Com- 
 panies, Wanklyn points out. He says- 
 that the Thames water at Hampton con- 
 tains fifteen parts of albuminoid am- 
 monia, or ammonia derivable from or- 
 ganic matter, in one hundred millions, 
 
130 
 
 that is to say, .15 in 100,000, which is the 
 way we have generally reckoned it, and 
 that after filtration by the company it 
 only contains 5 or 6; so that you see 
 water is capable of being purified that 
 is, the matters in solution are capable of 
 being altered in drinking water on an 
 immense scale. 
 
 Now what sort of things are these 
 filter beds, as they are made? because 
 laboratory experiments are all very well, 
 but you have practically to do it on a 
 large scale. Mr. Hawkesley has made 
 some large water- works, as you are most 
 of you probably aware, at Leicester, and 
 there, there is a reservoir of forty acres 
 in extent. There are also four filter 
 beds, each ninety nine feet long, and 
 sixty-six feet wide, and eight feet eight 
 inches deep from the ground. The water 
 comes in separate channels to these filter 
 beds, and it is passed downwards through 
 the following filtering materials: Two 
 feet six inches of sand, and then two 
 feet six inches of layers of gravel of 
 various sizes (from the size of beans up 
 
131 
 
 to eggs) to the drains below, and thence 
 by pipes into an octagonal pure water 
 tank. This tank, eight feet eight inches 
 deep, holds seven feet eight inches of 
 water, and is sixty- six feet from side to 
 side. That is the general plan. 
 
 The supply comes to the filter beds 
 from the reservoir at various points ; it 
 passes through two feet six inches of 
 coarse sand for, it must be observed, 
 fine sand will not do, as it gets choked 
 up by the suspended matters in the 
 water and then through two feet six 
 inches of gravel. The filtering beds 
 have sloping sides and are made of sand, 
 fine gravel, coarse gravel, then very 
 coarse gravel, with a drain at the bot- 
 tom. The filtered water is delivered 
 into an upright pipe in the tank, which 
 comes within two feet of the top, so 
 that the pressure of the water on the 
 beds from above can never be greater 
 than that due to a height of two feet. 
 It is essential that the pressure on the 
 surface of the beds should not be too 
 great. 
 
132 
 
 Well now from these filters six hun- 
 dred or seven hundred gallons per day 
 per square yard flow, and the proper 
 rate of vertical descent for the water, as 
 it is generally considered, is six inches 
 per hour, not more, or seventy-five gal- 
 lons per square foot in twenty-four 
 hours, and that you see is about the rate 
 at which it passes through these last 
 named works ; now the effect at this 
 particular place is that the water is clari- 
 fied, and a considerable proportion of 
 the organic matters in solution are re- 
 moved from it. The sand of the surface 
 of the filter beds requires scraping from 
 time to time, and also renewing. 
 
 At the Gorbals Filtration Works near 
 Glasgow, the filtering materials are 
 placed in vertical compartments with 
 passages between them, in each of which 
 the water rises to nearly its original 
 level and then flows over into the next 
 compartment and down through the 
 filtering material in it. There are two 
 other plans I must mention; at Black- 
 burn, for instance, there is no filtration. 
 
133 
 
 There they have a surface reservoir, and 
 they take the water out of it from the 
 top by a sort of process of decantation. 
 They let it settle, and then take only the 
 water from the top. Another plan is in 
 practice at St. Petersburg. There the 
 water is made to fall down a series of 
 steps, and then through wire gauze, and 
 lastly through sand niters, and by these 
 means the water which is generally very 
 impure is rendered tolerably pure and a 
 considerable amount of putrescible or- 
 ganic matters is collected from this wire 
 gauze. 
 
 Now we have to consider briefly the 
 ways in which water may be distributed 
 in towns. In the first place, as to the 
 mains : their size must be calculated ac- 
 cording to the supply required. 
 
 Mains are often made in towns on 
 both sides of the streets, in order that 
 the supply may not be entirely cut off 
 during repairs. There must be means 
 provided by which the water may be 
 stopped in a main in order that it may 
 be repaired. The bends and junctions 
 
134 
 
 should always be curved. There should 
 be no junctions made at right angles, 
 and there should be no angular junctions 
 if it can be helped. Mains should be 
 made of cast iron. They should be 
 greater than 3 inches in diameter. The 
 best service pipes for houses are f in., or 
 1 inch wrought iron service pipes that 
 screw together. They are better than 
 lead, and they are likewise cheaper than 
 lead. Wrought iron pipes are better 
 than lead for this reason, that certain 
 kinds of water act upon lead. Soft 
 water is apt to act upon lead. Fortu- 
 nately, hard waters, containing a con- 
 siderable amount of carbonic acid, act 
 very little on leaden pipes, and so it is 
 the practice very frequently to have 
 leaden pipes and cisterns made of lead, 
 and practically very little harm results. 
 If you refer to the 25th vol. of the 
 "Proceedings of the Institution of Civil 
 Engineers," you will find a discussion on 
 water supply, and there you will see 
 that Mr. Bateman gave it as his opinion 
 that even soft water acted very little 
 
135 
 
 indeed on leaden pipes, after a time. 
 It acts on them at first, but the leaden 
 pipe or cistern soon gets covered inside 
 with an insoluble coat of subcarbonate 
 of lead, and the result is that afterwards 
 the water acts very little on it. The 
 water of Loch Katrine, which is supplied 
 to Glasgow, acts very little on the lead- 
 en pipes and cisterns used. However, 
 there is no reason for having lead if dan- 
 ger be apprehended as likely to result ; 
 wrought iron will do just as well, and is 
 cheaper. 
 
 A town may be supplied in one of 
 two ways. These two ways are known 
 as the Constant and Intermittent sys- 
 tems. First, there is the Constant sys- 
 tem, in which, of course, the mains are 
 always full, and the water is brought 
 into the houses by pipes from the mains, 
 no cisterns being needed, as the water is 
 always in the pipes, and you have only 
 to turn a tap in order to get it. Second- 
 ly, there is the Intermittent system, in 
 which the water is only supplied for a 
 short time during the day, and in this 
 
136 
 
 Intermittent system it is therefore neces- 
 sary to have cisterns in the houses. Now 
 as to the relative advantages and disad- 
 vantages. Professor Bankine says : 
 " The system called that of Constant ser- 
 vice according to which all distributing 
 pipes are kept charged with water at all 
 times, is the best, not only for the con- 
 venience of the inhabitants, but also for 
 the durability of the pipes and for the 
 purity of the water; for pipes when 
 alternately wet and dry tend to rust, and 
 when emptied of water they are liable 
 to collect rust, dust, coal-gas and the 
 effluvia of neighboring sewers, which 
 are absorbed by the water on its re- 
 admission. In order, however, that the 
 system of Constant service may be car- 
 ried out with efficiency and economy, it 
 is necessary that the diameters of the 
 pipes should be carefully adapted to 
 their discharge, and to the elevation of 
 the district which they are to supply, 
 and that the town should be sufficiently 
 provided with town reservoirs. When 
 these conditions are not fulfilled, it may 
 
157 
 
 be indispensable to practice the system 
 of Intermittent service, especially as re- 
 gards elevated districts, that is to say, 
 to supply certain districts in succession 
 during certain hours of the day." You 
 see, therefore, that Professor Rankine 
 emphatically condemns the system of 
 Intermittent service as compared with 
 that of Constant service. 
 
 Now the great objection to the system 
 of Intermittent service is the necessity 
 of having cisterns, whatever they are 
 made of. Water becomes impure in cis- 
 terns, dust collects in them, and the cis- 
 terns require frequently to be cleansed. 
 If this is not done the water may even 
 become dangerous to drink. Where cis- 
 terns are necessary, slate cisterns are 
 the best. They require to be made with 
 good cement, or they are apt to leak, 
 and then you are liable to get red lead 
 or something of that sort used to fill up 
 the joints, and so you get the water 
 tainted. Iron rusts, and for that reason 
 cast iron mains require to be varnished 
 inside and out. Zinc has been used for 
 
138 
 
 cisterns and also for pipes, but zinc often 
 contains lead, and cases have been known 
 of lead-poisoning having resulted from 
 the use of zinc pipes or cisterns. There 
 have been plenty of ways proposed for 
 coating lead pipes so that the water may 
 not act upon them. Several of them are 
 absolutely objectionable ; one of the 
 methods, for instance, was the use of a 
 varnish containing arsenic ; and even 
 other varnishes, which do not seem to be 
 objectionable, are not now practically 
 used. 
 
 If you look in vols. 12 and 25 of the 
 Proceedings of the Institution of Civil 
 Engineers, you will see a great many 
 arguments for and against both the 
 " Constant " and " Intermittent " sys- 
 tems, and one argument against the 
 "Intermittent" system is always that 
 the amount of waste is enormous. It is 
 stated, as you will there find, that at that 
 time the amount of water wasted in Lon- 
 don was something like half the supply. 
 You find it alleged that there is great 
 waste also on the Constant system, be- 
 
139 
 
 cause, it is said, the mains are always 
 full and the taps are apt to be left 
 running. But this may be provided 
 against by having the taps placed inside 
 the houses, and then you will be quite 
 sure there is not much waste. Then, the 
 waste that has been observed with the 
 Constant system has been mostly caused 
 where the Intermittent system has been 
 changed for the Constant system, and in 
 that case you do sustain a loss of water; 
 a loss on account chiefly of faulty pipes, 
 \ and leaky fittings, for such as may do 
 y very well under the Intermittent system 
 are not good enough to be employed for 
 the Constant system. In Liverpool, at a 
 particular date, there were used 33,000,- 
 000 gallons of water a week, in the sup- 
 ply of which only 1,000,000 gallons 
 , were supplied on the Constant service, 
 and the whole of the remaining 32,000,- 
 000 gallons were on the Intermittent ser- 
 vice. For some weeks, as an experi- 
 ment, three sevenths of the town were 
 put on the Constant service, and then 
 the amount of water used rose from 
 
140 
 
 33,000,000 to 41,000,000 gallons per 
 week. But where there has originally 
 been sufficient attention to the fittings, 
 and where they are strong enough it is 
 otherwise. For instance, in the case of 
 Wolverhampton, at the same period, it 
 is stated that in that town there was a 
 saving effected by changing from the 
 Intermittent to the Constant system, a 
 saving of no less than 20 gallons per 
 head per day. (Vol. xii., p. 503.) 
 
 A disadvantage of the Constant sys- 
 tem is that the water supply sometimes 
 runs short in the higher parts of the 
 town, while in the lower parts there is a 
 sufficient supply; so that cisterns would 
 sometimes need to be provided, even 
 under the Constant system, in these high- 
 er parts of the town. 
 
 As a summary : With the Constant 
 system the waste of water is certainly 
 less than with the other if the fittings 
 are properly attended to, and if the fit- 
 tings, pipes, etc., have been originally 
 arranged for the Constant system. The 
 water in the case of the Constant ser- 
 
141 
 
 vice is purer and fresher, and at a lower 
 temperature in summer, and less subject 
 to frost in winter. The water is purer 
 because it escapes the impurities which I 
 have already pointed out, as collecting 
 in pipes, and it also escapes those im- 
 purities which the water gets by being 
 stored in cisterns. 
 
 The inconvenience from interruption 
 to the supply during repairs is never 
 actually experienced, as the interruption 
 need only be for a few hours. On the 
 other hand, the interruptions and the 
 waste caused by neglect of turn-cocks, by 
 the limitation of the quantity of water, 
 by leaky taps and cisterns, and in other 
 ways these inconveniences are absent. 
 Then the leakage from pipes is less. In 
 the Constant service the pipes are made 
 stronger, and practically there is much 
 less bursting. Mr. Hawkesley states 
 that the difference between the systems 
 is a question of pipes and fittings, and 
 that when the supply is well managed 
 the waste under the Constant system is 
 less. 
 
142 
 
 Then the water supply should always 
 be to the top of the house, and, if pos- 
 sible, to each story of the house. If cis- 
 terns are necessary, those used for drink- 
 ing water should always be separate from 
 any other cistern in the house. If, for 
 instance, there is a cistern for the water- 
 closet, it should be entirely separate from 
 the cistern used for the storage of drink- 
 ing water ; there should be two separate 
 cisterns. Then a chief point to attend 
 to with regard to the drinking water 
 system is that it should be covered. 
 Secondly That it should be easily ac- 
 cessible, so as to be readily cleaned out : 
 and thirdly and this is a most important 
 point that the waste pipe from it should 
 empty out into the open air either over 
 the surface of the yard or over a roof 
 or into a rain- water pipe, which itself 
 does not go down into a drain. The 
 waste pipe should on no consideration be 
 connected with any water-closet appa 
 ratus or with drains. This is almost in- 
 variably done, and that is why I insist so 
 much on the importance of this point. 
 
143 
 
 I may tell you that one of the most fre- 
 quent causes of typhoid fever in London 
 at this moment of this I have not the 
 slightest doubt is that the waste pipes 
 from the drinking water cisterns are con- 
 nected with some part of the sewerage 
 apparatus, and very often directly with 
 the sewers. The house drain, more fre- 
 quently than not, being unventilated, the 
 waste pipe of the drinking water cistern 
 becomes the ventilator of the house 
 drain, and the foul air of the house drain 
 goes up into the space between the sur- 
 face of the water and the lid of the cis- 
 tern, and is absorbed, and the result, in 
 many cases, as I have frequently observ- 
 ed, has been a severe attack of diarrhea 
 through the whole household, or else of 
 typhoid fever, and I have no doubt in 
 some cases of cholera also. 
 
 The overflow pipes from other cisterns 
 we need not be so particular about, be- 
 cause we do not require to drink the 
 water; but it is just as well that they 
 should empty in a similar way if pos- 
 sible. If not, they may be made to end 
 in what is called the D trap of the water- 
 
144 
 
 closet. I shall explain that, however, 
 more fully further on. 
 
 Now we have brought the water into 
 the house either into the cisterns, or it 
 may be merely into the pipes, which 
 are kept constantly full, and which have 
 taps at various levels inside the house. 
 When inside the house, it may be puri- 
 fied still further, if necessary, by house- 
 hold charcoal filters, or by boiling and 
 then being left to stand in stone vessels. 
 That is an excellent plan, and I must tell 
 you here that impure water may be puri- 
 fied to a very considerable extent by 
 making an infusion in it; for instance, 
 an infusion of tea. This is very import- 
 ant for you to know when you may have 
 to drink water in marshy countries. A 
 great deal of mischief is sometimes 
 done by drinking water in marshy coun- 
 tries, and this mischief may be prevented 
 by merely boiling it. That is a very 
 good thing, but still it is better, on the 
 whole, to make a weak infusion of some- 
 thing like tea, in it, and that is the sys- 
 tem which has been practiced for a thou- 
 sand ears in China. 
 
14 DAY USE 
 
 ETURN TO DESK FROM WHICH BORROWED 
 
 LOAN DEPT. 
 
 liis book is due on the last date stamped below, or 
 on the date to which renewed. 
 
UNIVERSITY OF CALIFORNIA LIBRARY