" FERRODOR " IS A NATURAL STEEL GREY METALLIC POWDER found in a very fine state of division, the Chemical Analysis of which shows it to contain ahout 95 PER CENT. OF RUSTLESS PEROXIDE OF IRON. "FERRODOR" is unaffected by Air, Water, Steam, Nitric and Sulphuric Acids, ALKALIES, Sulphur Ammonia and Cvanocen Gases, &c. &c. Franklin Institute Library PHILADELPHIA Class(> 91 7 Booktt ^ kS Accession48 081 Free from all substances that can affect metal. Impervious to external influences. The "Ferrodor" cannot absorb or impart oxygen, so that the oil in the paint is not destroyed, as is the case with Red Oxides. Elasticity that allows for expansion and contraction of metal under varying temperatures. (ABOVE IS A MOST IMPORTANT POINT.) Resists strong acid solutions, &c. SOLE MANUFACTURERS - GRIFFITHS BROS. & CO. MACKS ROAD, BERMONDSEY, LONDON, S.E. j THE CORROSION of METAL STRUCTURES CAN BE PREVENTED BY THE USE OF "FERRODOR " PAINT, known in Germany, &c, as " SCHUPPENPANZERFARBE." Manufactured from NATURAL METALLIC STEEL GREY RUSTLESS PEROXIDE OF IRON and PURE LINSEED OIL. "FERRODOR" PAINT possesses all the requirements of a Scientifically Perfect Protective Coating FOR METAL Structures. " FERRODOR " PAINT for interior and exterior Iron and Steel Ship Work, and all Metal Structures, surpasses Oxides of Iron, &c, and is SUPERIOR TO RED LEAD, SOLE MANUFACTURERS- GRIFFITHS BROS. & CO. MACKS ROAD, BERMONDSEY, LONDON, S.E. ADVERTISEMENTS. PAINTS ^"^'^^i. PAINTS BRIDGES. GASHOLDERS. Contractors to Her Majesty's Government, the Indian Government, the Natal Government, the Turkish Government, the Netherlands Railway, the London County Council, &c. &c. FOR ALL METALLIC STRUCTURES USE CROSBIE'S PAINTS. GUARANTEED GENUINE AND FREE FROM ALL ADULTERATION. Crosbie's Genuine Oxide of Iron Paint is made from pigment containing ninety per cent, of peroxide of iron unadulterated and extra finely ground in the very best genuine linseed oil. These Paints will successfully withstand the hottest sun and sea-air, and are largely used in tropical climates. They are also largely used in gas-works, as the are unaffected by heat and gaseous fumes. The covering power and durability of these Paints far exceed that of the ordinary oxide of iron paints. Messrs. Crosbie are Contractors to Five Govern- ments and to no less than 130 Corporations and Public Authorities. CATALOGUE AND TESTIMONIALS ON APPLICATION. ADOLPHE CROSBIE, Ltd., COLOUR WORKS, WOLVERHAMPTON. Also Makers of Genuine Lead Paints, Compositions for Ships' Bottoms, &e., &c. ADVERTISEMENTS. 4 Contractors to Her Majesty's Government, The India Board, The Crown Agents for the Colonies, The Board of Trade, &c. &c, PEACOCK & BUCHAN, INVENTORS AND MANUFACTURERS OF Anti-Fouling Compositions, FOB IRON, STEEL AND SHEATHED VESSELS. "COPPER PAINT" FOB WOOD BOTTOMS OF YACHTS, FISHING SMACKS, BOATS, &c. READY-MIXED, SANITARY, NON-POISONOUS PAINTS, For House and Ship Painting, Internally and Externally. Packed in LEVER-LID Tin Cans of 1 lb., 3 lbs., 7 lbs. and 14 lbs. each. Manufactory : SOUTHAMPTON. N.B —Peacock and Buchan's "Protective Undercoat" and also the "Anti-corrosive Metallic Paint" (Light or Dark) is specially adapted for all Metallic Surfaces, and is largely used by the Royal Mail Companies for Ships' Bunkers, Holds, Tank Tops, Steel Decks, and all exposed surfaces. 5 ADVEETISEMENTS. TORBAY « • PAINT. "WOLSTON'S" and " GALLEY'S TORBAY." For use on Metallic Structures in exposed positions. UNEQUALLED PRESERVATIVE & COVERING POWERS. Particulars and Shades on application to The Torbay Paint Company, Proprietors: STEVENS & CO. (WITH WHICH IS AMALGAMATED THE TORBAY & DART PAINT CO., LTD.), 26, 27 & 28 Billiter Street, London, E.O. 8 India Buildings, Fenwick St., Liverpool. Works & Mines : BRIXHAM & DARTMOUTH, DEVON. [To face Half -Title. METALLIC STEUCTUEES CORROSION AND FOULING, AND THEIR PREVENTION 6 ADVEETISEMENTS. THE MANUFACTURE OF HIGH-CLASS EEADY-MIXED PAINTS FOE ALL PUEPOSES ON METAL IS A SPECIALITY OP SUTER HARTMANN & RAHTJEN'S COMPOSITION COMPANY, Limited, 18 BILLITER STREET, LONDON, E.G. THEIR SHIPS' BOTTOM COMPOSITION IS SUPPLIED ANNUALLY TO 872,000 Tons of British and Foreign Navies and 5,436,000 Tons of Mercantile Marine. S. H. & It. C. Co., Ltd., will be pleased to advise upon the most suitable paint for different metallic structures, upon receiving details of situation and kind of structure. [To face Title Page. METALLIC STRUCTURES: CORROSION AND FOULING, AND THEIR PREVENTION a practical &tt43oofc TO THE SAFETY OF WORKS IN IRON AND STEEL, AND OF SHIPS; AND TO THE SELECTION OF PAINTS FOR THEM. By JOHN NEWMAN ASSOC. M. INST. C.E., F. IMPL. INST. AUTHOR OF EARTHWORK SLIPS AND SUBSIDENCES ' ; 'NOTES ON CONCRETE AND WORKS IN CONCRETE 1 ' SCAMPING TRICKS AND ODD KNOWLEDGE OCCASIONALLY PRACTISED UPON PUBLIC WORKS ' * NOTES ON CYLINDER BRIDGE PIERS, AND THE WELL SYSTEM OF FOUNDATIONS' 'FOUNDATIONS IN SAND ' (LECTURE, ROYAL ENGINEERS' INSTITUTE, CHATHAM) ETC. ETC. ETC. Honfcon: E. & F. N. SPON, 125 STRAND SPON & CHAMBERLAIN, 12 CORTLANDT STREET 1896 PBEFACE. This book has been written, as it is believed none has appeared for a long period, to help those who may be intrusted with the design, erection, or maintenance of any- kind of iron or steel structure, and to call to remembrance the salient points requiring attention, in order to prevent or arrest corrosion and fouling. The author, having had his attention specially directed, during many years, to the subject of corrosion and fouling and their prevention, has placed his experiences on record, and that of others in all parts of the world, to whom he offers his acknowledgments, and also to the Editors of the various Technical Journals mentioned. Chemical analyses of the composition of air, water, iron, and steel of various kinds are only referred to to illustrate the text where necessary ; however, no one can more appre- ciate the incalculable value of chemical analysis than the author, and the great help and guide it is to the engineer ; but analyses of air, rain, water, iron, and steel, are to be found in the various able and exhaustive works which relate to them, and therefore are not repeated ; the aim here being to present, as concisely as possible, information which may be useful to engineers, constructors, architects, students, owners, and those having the care of iron or steel structures and ships, Vlll COKKOSION AND FOULING. and to indicate some preventive and remedial measures that can be adopted for their preservation from corrosion and fouling. It should always he rememhered that the preservation of a metallic structure is only second in importance to that of its design and erection, and, unless corrosion and fouling are prevented or repressed, the time must come when the original strength will be so impaired as to be dangerous. Fouling is distinct in many of its features from that of corrosion, it has, therefore, been treated in Part II. of this book, and much information will be found in it relating to the fouling and corrosion of ships, pile, bridge, promenade, and landing piers, pontoons, and similar works of construction, and anti-corrosive, and anti-corrosive and anti-fouling paints and compositions: products of very great influence in the preservation and protection from corrosion and fouling of metallic structures, whether submerged or unsubmerged. The causes of fouling and corrosion in sea water are briefly examined, also the qualities required in an anti- corrosive and anti-fouling paint, and in other coatings used with the object of preventing the fouling and corrosion of ships and any submerged or partly submerged work, and corrosion in all. The subject being of a comprehensive nature, the book is not intended to be an exhaustive treatise, but a kind of vade mecum written to supply concise practical information on a subject of much importance to all having to do with the design, construction, or maintenance of metallic structures, whether floating or fixed. J. N. London: November 19, 1895. CONTENTS. PART I. THE SAFETY AND PRESERVATION OF IRON AND STEEL STRUCTURES. CHAPTER PAGB I. The Importance op the Efficient Maintenance op Metallic Structures, and the Prevention op Corrosion .. .. 1 II. Some Causes op Corrosion described .. .. 11 IIiXthe Character and Quality of Iron and Steel, and its Relation to Corrosion .. .. .. .. 30 IV. Notes on the Corrosion op Cast Iron and Steel .. 43 V. Notes on the Relative Corrosibilitt of Steel, Wrought Iron and Cast Iron, and Selection of the Metal for a Structure .. .. .. .. 59 VI. Notes on Rapid Corrosion, with Examples .. .. 67 VII. The Corrosive Influence of Soils, Vegetation, Situa- tion, Climate, Rainfall and Water .. * 84 VIII. Galvanic Action and Corrosion 100 J IX. The Influence of the Scale on Cast Iron, Wrought Iron, and Steel, as regards Corrosion .. .. 124 CORROSION AND FOULING. X. The Serviceable Life of Metallic Structures, and some Examples of Corrosion .. .. .. .. 131 XI. Notes on the Corrosion of Metal embedded in Con- crete, Brickwork, or Masonry ... .. 152 XII. The Influence of Design and Workmanship with regard to Corrosion \ t 157 XIII. Corrosion in Piles and Columns ... 175 XIV. Notes on Corrosion in Bridge Floors, Koadways, and Platforms .. .. .. .. ># _ 190 XV. J Corrosion in Water Pipes, Sewers, etc. .. .. 199 XVI. Corrosion by Contact with Wood, and in Girder- Seats and Beds .. .. 210 PAET II. THE PREVENTION OF FOULING AND CORROSION IN SUBMERGED STRUCTURES AND SHIPS. CHAPTER PAGE I. Fouling by Scum, Mud, and Marine Plants.. .. 223 II. Fouling by Molldsca, Coral, and Marine Animals .. 237 III. Fouling, and the Temperature and Saltness of the Sea 247 IV. Notes on Corrosion in Sea Water, and the Injurious Effect of Fouling .. .. ., 252 V. Preparation of the Surface of Metal before Painting #i 257 CONTENTS. XI CHAPTER VI. Notes on some Oils, Gums, and Eesins used in Paint VII. Notes on Lead and Cheap Paints, and the Require- ments of an Anti-cobeosive Paint VIII. Notes on Paints for Metallic Structures IX. Notes on Varnish, Tar, and Bituminous Paints X. Notes on Electric, and Superheated Steam Coatinqs, the External Corrosion of Boilers, and Explo- sive Paints .. XI. Anti-Fouling Paints, Compositions, and Fluids XII. Notes on Paint Specifications XIII. Scamping Tricks and Painting PAGE 268 277 291 302 312 320 359 366 INDEX to Parts I. and II. 369 CORROSION AND ITS PREVENTION. PART I. THE SAFETY AND PRESERVATION OF IRON AND STEEL STRUCTURES. CHAPTER I. THE IMPORTANCE OF THE EFFICIENT MAINTENANCE OF METALLIC STRUCTURES AND THE PREVENTION OF CORROSION. In nature there is no stability or permanence, for decay and reconstruction are everywhere to be seen, and nothing is known which does not undergo a change of state when subject to the action of intense heat, and most substances are more or less affected by water. Although no marked alteration of form or appearance may occur or be appreci- able in the initiatory stages, a continual change proceeds in almost every material used in construction from the time it is exposed to any of the elements, or when in contact with many other substances, for the power of decomposition is exercised by different substances on others. The com- bustion of decay has commenced and cannot be entirely prevented, but it may be much lessened, and, in some cases, completely controlled. The process of oxidation has become visible on metals when they have become tarnished, but a metamorphic condition will have occurred before its effects are apparent, and under the surface it may be proceeding rapidly and spreading around, such decomposing action frequently having nuclei from which its power springs. B 2 CORROSION AND ITS PREVENTION. The maintenance and present condition of iron and steel used in various forms in engineering and building construc- tion are likely to demand increasing attention, as either from fatigue, vibration, corrosion, decomposition, or general deterioration, the metal may have become either changed in its characteristics, or be so impaired as to be no longer the same metal as when erected, or of sufficient dimensions to sustain the load it originally was well able to support. The number of metallic structures requiring either strengthening or renewing must increase, and, in thoroughly opened-up countries, it is probable their restoration will soon be an important branch of engineering science. It is one in which special skill is necessary, not only to ascertain the real condition of a structure, but also to arrest any elements of decay, and to restore it to its original strength. Metallic railway and road bridges, and public buildings of importance have more or less continued attention bestowed upon them, but the care is usually, and not unnaturally, commensurate with the size of the bridge or building, and its public posi- tion and importance, notwithstanding that in the smaller structures the surfaces may be more exposed, and corrosion greater in comparison with the sectional area of the metal. Private warehouses and buildings are usually left either to a tenant to keep in repair, which generally means a coat or two of cheap oil paint every few years, or are under the supervision of some local builder whose knowledge of the circumstances which cause corrosion or depreciation is not, to say the least, too profound. Structures and buildings supported by iron or steel columns and pillars should be occasionally examined by an expert engineer, for in many cases, although the ironwork may be well designed so far as regards strength to carry any load, and with the view to easy erection, the instances are comparatively few in which their preservation from corrosion has been specially considered. However, metallic bridges will have to be rebuilt unless care is regularly taken to preserve all their parts ; for corrosion, if allowed to proceed, MAINTENANCE OP METALLIC STRUCTURES. 3 will, sooner or later, have increased to such an extent that, apart from the question of the fatigue of the metal and additional requirements of traffic, the reduction of area will be so considerable as to demand attention. Mr. Ewing Matheson in his paper ' Steel for Structures,' * wrote, " The preservation of iron from rust is not in this country sufficiently con- sidered." Steel being now so much used, and the sectional area reduced in many instances as compared with iron, the preservation of the metal in its original section is of even more importance than formerly. Corrosion arises, and is promoted or accelerated by chemical, voltaic, and mechanical action. It therefore varies, for almost every metal is differently affected, even by mere mechanical action, for it is less on the surface of hard metals than soft, and the resistance to penetration by vapour or moisture is also augmented. Chemical and voltaic action will vary according to the composition and homogeneousness of the metal. This indicates the direction in which an explanation may be sought for any exceptional and peculiar corrosion, and may lead to the cause of it being discovered. Metals, unlike Portland cement, which to a certain limiting period increases, or should increase, in strength with age, suffer a diminution of strength, however slow it may be, almost from the time they are used. To attempt to prevent corrosion without knowing the cause of such action can only by chance be successful, for the rapidity, and therefore the power of corrosive influences, depends upon the conditions and circumstances in which the metal is placed. Sir B. Baker has concisely declared that " it is the deviation from the average which really is so important in the design of engineering works." In a few instances it has happened that the reports of the most eminent professors of chemistry and the results obtained in engineering practice have not agreed, and yet undoubtedly both have been correct. The conditions under which the substances or liquids have been used has been the cause of the dissimilarity. In the * • Minutes of Proceedings, Inst. C.E.,' vol. Ixix. B 2 4 CORROSION AND ITS PREVENTION. laboratory the examination is conducted with the greatest minuteness, and time is allowed for important action, and attention is specially directed to discover everything that can he detected regarding the information desired. If all the conditions are not clearly stated under which a sub- stance or liquid is to be used, a report will most probably be either too favourable or unfavourable, although perfectly correct so far as probable results under laboratorial circum- stances are concerned. It is necessary for an engineer to fully explain to an analytical chemist the corrosive influences to which any substance will be exposed, and the circum- stances in which it will be used. Somewhat amusing instances have occurred because of the want of such concerted action, for it was said Portland cement was most injuriously affected by sea water, and so it may be in certain experiments, and it may be said so would any other substance, but good Portland cement, properly mixed and deposited, can be thoroughly relied upon under the conditions in which it is used in concrete structures in sea water, and has been most successfully and economically for many years; however, the elaborate dis- integration and action caused by minute analysis would probably generally cause an adverse view to be formed of its adaptability for such purposes. The different circumstances in which the material was used being alone responsible for the dissimilarity of action, for in analysis the Portland cement was minutely subdivided, whereas in a structure only the surface of a mass is exposed. Similarly, sea water for watering streets was reported upon unfavourably, but it is adopted, and with success, either for sewers or streets at several fashionable watering places, where it is almost a necessity of the town's existence that everything should be so arranged as to attract and be approved. The apparent discrepancy in the reports of the analytical chemists and the results found by engineers was caused by the different circumstances of the tests and the actual use in roads, &c. In the case of road-watering, the MAINTENANCE OP METALLIC STRUCTURES. 5 action of the sea water was merely superficial, and by reason of the road being hardened by its binding action and caked, there was less dust to be blown about ; but macadamised roads or paving setts, if frequently watered by sea water, became slippery. When used for sewer flushing, if done quickly, and with a sufficient quantity of sea water, and if time is not given for fermentative action to develop, no- thing unpleasant or injurious occurs. The circumstances of practical use modified the expected results determined by experiments in a laboratory, and yet while apparently impugning them, confirmed their accuracy. From these examples it can be readily judged, care should be taken to give full information to an analytical chemist in any special case in which the probable intensity of corrosive influences is desired to be known. Although mechanical tests are usually of more value to the engineer than chemical analyses, still the latter are always valuable, and so far as corrosion is concerned its probable progress cannot be determined except from the probable chemical action. It is gradually being considered a matter of vital im- portance in all countries to adopt some means of ascertaining the extent to which a metallic bridge or structure has deteriorated or is deteriorating ; and the paramount necessity of preserving the full powers of the metal is acknowledged in order that the required strength may be retained for the longest possible period. Corrosion and general deterioration, if allowed to proceed unchecked, must culminate in failure. In order to be able to form a correct judgment of the magnitude of the corrosion of any structure, it is necessary to give instructions in case a personal inspection cannot be made, so that correct data are collected. Some of the chief points requiring answers are : — 1. Situation. 2. Date of erection of the bridge. 3. Name and address of the bridge builder. 4. Materials used in the various parts. 6 CORROSION AND ITS PREVENTION. 5. State if any alteration in the character of the neigh- bourhood since erection, or if anything has occurred since erection or last inspection likely to produce a more corrosive atmosphere. 6. Whether any structural alterations have been made. If so, state them. 7. If any alteration in the load or traffic since erection, whether increased weight of engine or rolling stock, speed or traffic. 8. Thickness of the timber bearers between rail or chair base and girder flange, and whether it has been lessened or increased. 9. Permanent deflection. 10. Deflections on loading tests. 11. State whether any parts have been renewed. 12. Exact distance apart of the top and bottom flanges. 13. Length of the lattice bars or diagonals. 14. If any lateral distortion in the girders. 15. Whether the vibration is excessive or not. 1 6. The state of the joints and the riveting. 1 7. The state of the paint. 18. Whether all the parts are free from corrosion, and mention those corroded, if any. 19. Quantity and weight of rust, if any, taken from structure. Top and bottom flanges, web, bars, bracing and wind-ties, flooring to be kept separate. 20. State if anything is observed influencing the corrosion or decay of the structure, and is not previously mentioned. A few of the preceding questions are only applicable to bridges, but those for metallic structures generally can be easily separated. The amount of the rust removed, if carefully collected under the superintendence of an expert, would periodically give an idea as to the relative rate of corrosion, and any exceptional initial quantity or increase would show that the durability of the structure was seriously affected ; also, if the amount of rust taken from each part was compared MAINTENANCE OF METALLIC STRUCTURES. 7 with that removed from the same part at the previous inspection, and the rate of corrosion of each part computed per unit of area, some indication, it could hardly be exact, would be obtained of the places especially liable to corrosion. In considering the data it is necessary to bear in mind that much of the strength of metallic bridges depends upon the metal retaining its elasticity, original sectional area, freedom from corrosion and buckling, absence of any loose or weak joints, or change in the fibre or the structure of the metal. If a bridge cannot be tested by loading, consequent upon the exigencies of the traffic, much can be gathered as to its condition by examination by an expert engineer ; for the permanent way men can easily detect rottenness in any timber, but cannot be expected to know much about the strains brought upon girders or their maximum safe deflec- tion under a passing load, or whether a permanent set has taken place, which, if it be permitted to continue, will probably ultimately be disastrous. The importance of frequent periodical examination of existing metallic bridges and structures cannot be over- estimated, and also their freedom from corrosion when first properly coated with a really preservative and anti-corrosive paint. In the United States of America many engineers favour periodical inspection by a skilled engineer who is an expert in bridge construction, and that the inspection should be compelled by legislative authority although not undertaken by it, but signed and certified reports sent in to be examined by a committee of experts, as occasionally bridges have proved by failure that they were " all wrong," when reported by foremen to be " all right." The number of old bridges, which is gradually being lessened, receiving the necessary attention, subject to in- creased loads, additional traffic, and at an accelerated speed, is not inconsiderable, and it is the smaller bridges rather than those of considerable span that perhaps more especially require inspection periodically, as they can hardly receive the same amount of care as any important structures, and 8 CORROSION AND ITS PREVENTION. yet the consequences of failure may be very serious indeed. The examination should he made by experts in metallic construction and repairs, or those trained to make such reports, and the time may come when there will be specially appointed engineers to make such examinations and reports, suggest precautionary and remedial measures, and carry them into execution. Not many lines of a specification are generally reserved for protection against corrosion, and some elaborate specifica- tions in all other respects may refer to painting in something like the following few words," all iron surfaces to be painted with two coats of metallic paint and oil, and with an addi- tional coat of lead and oil when the structure has been erected, the time of painting the last coat to be determined by the engineer." Thus the preservation of the metal from corrosion is only indirectly referred to, the covering of the metal with some substance being alone mentioned. Such a specification may be regarded as one which simply cares little for the maintenance of the original strength of a structure so long as it is erected. In the case of any metallic structure which can only be occasionally inspected by an expert engineer, or when it has to be erected abroad, special provision should be made against corrosive influences. Although calculations of the strains, alterations from the original design during construction, inferior material and workmanship, increased rolling load, the greater speed of the trains, unequal subsidence of the piers or abutments, or other special influences, are all important in their bearing on the serviceable life of a metallic bridge, that of proper maintenance is equally so, for if it be neglected, the strength of a structure is reduced, and cannot be restored except by additions or alterations, such as extra flange plates, stirf eners, web plates or bars, cross girders, and bracing. The metal may be said to be deteriorating no matter how slowly, from change of stress, vibration, and corrosion, and the chief aim is to reduce these deteriorating influences to a minimum. MAINTENANCE OF METALLIC STRUCTURES. 9 The determination of the gradual encroachment of molecular changes in metal due to frequent alteration of strain is not fully known, but it is accepted that any such alteration must affect corrosion. There are phenomena that will probably be discovered in the behaviour of iron and steel which require time to determine ; for instance, whether cast iron especially, and wrought iron and steel in a less degree, which has been undisturbed from its original position for a long series of years, becomes crystallised or not so as to be almost brittle on being disturbed, or broken with a light blow. Any alteration in the composition or texture of metals will affect the intensity of the corrosion. The metal in a bridge is most probably, if not always, more or less deteriorating from the effects of variable strain, oxidation, vibration, and even, it is said, from a structure being fixed and strained more either from expansion or contraction on one side than the other. It seems to be generally agreed that there are metallic bridges in use in various parts of the world which were not calculated to be strained as they are. The additional stress may be considered to arise from increase of load, speed, traffic, fatigue of the metal, loosening of the parts, the rails being fixed nearer the girders and without so thick an elastic medium as timber, and corrosion consequent upon inattention to the painting and its maintenance so as to prevent rust, or some of them ; the influence of corrosion being all-important as being not only superficial, but also as reducing the sectional area of the metal, loosening the parts, and therefore augmenting the strains, or, it may be, altering their nature, particularly in bridges of small span. The special danger of corrosion is when it from various causes produces exceptional local weakness or deterioration till the metal at such a point is strained beyond its limit of elasticity, then, although other portions of a metallic struc- ture may not be nearly so severely strained, the whole may fail, owing to want of structural stiffness, which it is equally as important to regard as mere strength. Some parts re- 10 CORROSION AND ITS PREVENTION. quiring special attention are the girder beds, bed and wall plates, tbe riveting, all junctions of parts, joints, and any- place where water or dirt can accumulate, and places where different metals or materials meet, such as wrought iron with steel or cast iron, and any metal with timber ; also, after years of service, the platforms of bridges, the abutments, parapets and wingwalls, whether of brickwork, masonry or concrete, may become in an unsatisfactory condition and aid in forming receptacles for the promotion of corrosion. It has been said the causes of the deterioration of iron bridges are : (1) faulty design ; (2) faulty execution ; and (3) careless maintenance. In some countries the State regulations require : (1) a thorough test before opening for traffic ; (2) the structure to be kept in good condition, which is understood to mean periodical inspection at intervals not greater than one year, annual cleaning, examination of all parts, removal of rust, and preservation by paint ; (3) a record kept of the strains in the structure ; (4) a repetition of the original tests at stated periods. The simple rolling load is not always commended, because the platform and permanent way may be so made as to alter the distribution of the strains, and a sudden application of load or a dynamic test has therefore been suggested. The formation of rust, it is considered, must be particularly guarded against by good painting, and by making all the parts easily accessible. The magnitude of the deflection, although perhaps indi- cating the general condition of a girder, is regarded as in- conclusive, not only because it has been observed that in many cases the centre of a girder has risen, but in those of comparatively large span the changes caused by temperature are important, and alter the form of a girder to an appreciable extent. 11 CHAPTER II. SOME CAUSES OF CORROSION DESCRIBED. In considering the various causes and influences that in some way or other produce corrosion, it is well to remember that probably in the whole range of the science and art of construction there is nothing more difficult to contend against than the decay of materials, or anything requiring so much constant care and diversified treatment, consequent upon the different nature of the materials, the various objects to which they are applied, and the changeableness of the influences that cause deterioration and decomposition. Almost from the moment of their manufacture, the metallic portions of a structure may be said to be subject to the combustion of decay which, it is decreed, must sooner or later overtake them ; the aim of the engineer is to reduce to a minimum every deteriorating influence. Strictly speaking, submerged structures would be considered to be thor.e constantly im- mersed ; however, there are but few structures which are not occasionally partly covered with either rain, fresh or salt water, and therefore submerged and unsubmerged work, so far as corrosion is concerned, is almost a distinction without a difference, except that metals constantly immersed are not subject to such severe corrosive influences as those alternately wet and dry. Metal having signs of corrosion or even the appearance of being " weathered," is in a condition of deterioration, and cannot be considered as equal in strength to the same iron fresh from the rolls or moulds ; and as a rule the nearer rust is to the surface of iron the more oxide of iron there is in it. 12 CORROSION AND ITS PREVENTION. Experiments have shown that corrosion, not exteriorly aided by galvanic action, on steel plates increases progressively. Those of Mr. Andrews, F.K.S., showed it to be 50 per cent, more the second year compared with the first. Formerly it was sometimes thought that rust on the surface of iron acted as an efficient protection, but when it is on a plate it causes the metal beneath it to corrode under certain conditions, and any film that may be regarded as a protective covering is not of uniform thickness or character and is liable to peel and fall, for the rust on iron in the form of hydrated peroxides does not adhere to the surface upon which it is formed, but continues to accumulate until at length the flakes of rust become too heavy to be held together. It should not be considered a protection, but as the opposite, for although it acts temporarily as a shield to some extent upon the metal on which it has formed, it has a corrosive action of its own. If it were an effectual protection, corrosion would be arrested, whereas everywhere there is evidence that corrosion may continue until there is no iron left except what is contained in the residual dust. In addition, iron, in becoming oxidised, increases in size according to the quantity of oxygen taken up, and so it expels, having no adhesion to the metal, the film of paint, and the latter in time drops to the ground. Eepetition and convenience have almost caused it to be regarded by some as an axiom that so long as iron is placed in salt-water mud, the metal will be free from any deleterious corrosive effect, because a film of rust will form which will protect the iron from further corrosion, and because there is no constant renewal of the water which reaches the iron. Even if fresh supplies of water were pre- vented from approaching the surface, which is hardly the case, although temporarily it may be a kind of shield to the solid metal, it will fail from the causes stated. On leaving the rolls oxidation may be said to have com- menced. This can, of course, be arrested if the metal be dipped in boiled oil or tar-asphaltum preparation ; but if it has to be afterwards worked or manipulated, any protective SOME CAUSES OF CORROSION DESCRIBED. 13 films or portions of the preparation that have been imbibed by the pores may be disturbed, and then these form nuclei ajound which corrosion proceeds. Why does corrosion occur ? The oxidation of iron being complex, requiring conjoint action or influences, and not being a simple chemical combination of two elements, the question cannot be answered in a short sentence. It is advisable to remember that there are two kinds of chemical affinity ; one in which substances are capable of combining, one of the substances being in any proportion whatever to the other ; and the other in which substances only unite in certain definite proportions. Solutions of solid substances in fluids are instances of the former, iron-rust and water that of the latter. " Neither bright iron nor steel will rust in pure water or in pure air. The presence of carbonic acid, or some similar agent, seems necessary, although the final product may be destitute of carbon. Even when oxygen, moisture, and carbonic acid are all present, rusting will not, it appears,* take place unless the moisture condenses on the surface of the metal. When rusting does take place under ordinary circumstances, the first stage appears to be the formation of ferrous carbonate. This carbonate is then dissolved in carbonic acid water to form ferrous bicarbonate, which latter is then decomposed in presence of air and moisture to form hydrated ferric oxide, magnetic oxide being formed as an intermediate product." Four compounds of iron with oxygen are known. Iron. Oxygen. Colour. Protoxide 1 1 Green hue, changing to red brown. Sesquioxide (peroxide) . . 2 3 Blood red. Black, or magnetic oxidef 3 4 Ferric acid 1 3 Deep crimson. * See ' Engineering,' February 9, 1894. * The black oxide scales which form on wrought iron when heated to whiteness, as in the operation of forging, are chiefly composed of this oxide. 14 CORROSION AND ITS PREVENTION. What is the cause of iron becoming corroded when exposed to moist air ? The necessary moisture being present in the air brings into action the affinity of oxygen for iron or steel, the oxygen of the air combining with the surface of the metal and producing two kinds of oxide of iron, the ferrous oxide, which quickly becomes converted into the ferric oxide. Its durability, therefore, chiefly depends upon its power to resist combination with oxygen. Dry air at a high temperature is not encountered in the majority of engineering structures, and therefore need not be considered at length, although even dry air oxidises iron at a high temperature, as is proved by the familiar example of a red-hot poker being found to be rusty when it cools. It has been found in warming buildings by steam that if air is inside an iron vessel or tube in which steam is condensing, iron corrodes rapidly. If the air is excluded it does not. Then external corrosion has to be guarded against more than internal decay. Lead slowly absorbs oxygen and carbonic acid in moist air, and copper slowly oxidises in a moist atmosphere, and a green carbonate is formed on its surface. Experiments have shown that mercury, like many other metals, does not undergo oxidation either in dry air or in aqueous vapour, but only when aqueous vapour is present in the air, then it corrodes rapidly and oxide of mercury is formed over the surface. This, it is considered, points to an electrolytic explanation of the oxidation. Although water is decomposed by contact with iron, it is chiefly the oxygen contained in water which causes corrosion, and it is accelerated as the temperature increases, and by absence of light. Some moisture is necessary, although its decaying power is dependent on other coincident circum- stances, such as climate and character of the air, but moisture is an all important element in the production of corrosion. For instance, steam pipes coated with mineral wool or slag wool are kept free from rust so long as the wool is entirely dry, but where moisture is present it permeates the wool and reaches the pipe, and its surface becomes corroded. SOME CAUSES OF CORROSION DESCRIBED. 15 [Professor Egleston (U.S.A.) found that when moisture is at ■all constant in blast-furnace wool there will he a decom- [position of the slag, and an attack on the iron by sulphuric [acid set free, or by organic acids if the material comes from the drainage of the soil. Oxidation which may be comparatively slow in a moist atmosphere is very greatly accelerated in air or water con- taining acids or other corrosive agents, consequently it is the climatic conditions which principally govern the rate of corrosion, and account for such variableness in the service- able life of structures in which steel or iron are employed. The appearance of corrosion on the surface of iron is familiar, although it assumes various forms, as a skin of green-brown, brown, or blood-red, or deep crimson rust, nearly circular raised lumps having a centre of energy underneath which will be found depressions in the plate, or bunches or flakes of red-brown rust, which can be easily detached. That of steel is not so well known, thus the lilac tinge in steel is produced by partial oxidation, and corrosion has commenced. Why does the process of oxidation continue and ultimately destroy an iron or steel plate, and reduce it to a condition of practically being mere dust ? Without attempting to explain the chemistry of the matter, it may be stated that in iron the film of oxide of iron on the surface being converted into ferric oxide, it becomes an active corrosive influence, inas- much as part of its oxygen is transmitted to the surface beneath to form ferrous oxide, and this process does not cease, but is repeated until the whole of the iron is oxidised, I and then soon becomes a mass of mere flakes of fragile scale, which do not adhere, and ultimately become dust. There- fore, the importance of preventing any oxidation is indisput- able, more particularly as reliable experiments have shown ; that the corrosion of plates of wrought and cast iron, and i especially of steel, rapidly increases with time. In the very earliest years of modern civil engineering, : Smeaton wrote on corrosion, " I had observed that when iron once gets rust, so far as to form a scale, whatever coat of 16 CORROSION AND ITS PREVENTION. paint or varnish is put over this, the rust will go on pro- gressively under the paint." It may be regarded as an axiom that when rust has commenced it will proceed. The hygroscopic character of the hydrated ferric oxide causes an absorption of moisture from the atmosphere, and so under ordinary circumstances the chief necessary element is present, and it may be aided by the rust, which being electrically connected with the iron, will result in galvanic action. Oxygen alone does not cause corrosion, as will be gathered from the experiments of Dr. Grace Calvert, F.R.S., who, in describing some he made with perfectly cleaned blades of steel and iron exposed for four months to the action of different gases, in order to determine whether tlie oxidation of iron is due to the direct action of the oxygen of the atmosphere, or to the decomposition of its aquous vapour, or whether the very small quantity of carbonic acid which it contains determines or intensifies the oxidation of metallic iron, found that the blades showed the following results : — Dry oxygon : no oxidation. Damp oxygen : in three experiments, one blade onl; was slightly oxidised. Dry carbonic acid : slight appearance of a white precipitate upon the iron, found to be carbonate of iron. Dry carbonic acid and oxygen : no oxidation. Damp „ „ ; oxidation very rapid. Dry and damp oxygen and ammonia : no oxidation. These experiments tend to show that carbonic acid, and not oxygen or aqueous vapour, is the agent which determines the magnitude of the oxidation of iron in the atmosphere. Therefore, carbonic acid in a damp atmosphere is a powerful and active agent in causing corrosion, and it "was further shown by experiments that when iron was inmersed in water containing carbonic acid it rapidly oxidised ; that it was not due to the fixation of the oxygen dissolved in the water, the occurrence of hydrogen collected above the liquid in the test -bottle sufficiently proved, but it Vas due to SOME CAUSES OF CORROSION DESCRIBED. 17 I oxygen liberated from the water by galvanic action. Caustic alkalies were also found to much retard corrosion. Although oxygen cannot alone be considered as the active agent of corrosion, for a piece of clean iron or steel can be exposed in a dry atmosphere for a long time without appreci- able corrosion, and this notwithstanding oxygen is present, its active corrosive power, i. e. the force of affinity it has for the metal, is great, but is governed by a variety of circum- stances. However, the conditions that render it the active agent generally more or less prevail, especially the ever- \ potent one of damp or moisture ; therefore in very dry : localities, as Siberia, or in the arid plains and deserts of the world, corrosion is very much slower, even to being almost imperceptible, than in countries near the sea and subject to an annual heavy rainfall or comparatively damp atmosphere. If the atmosphere is excluded from a bottle of distilled water [a smooth clean piece of iron can be immersed without ! deterioration, as also in salt-water from which the air is excluded, if 30 grains of quick or caustic lime to each ounce of water is added. It is a characteristic of oxygen, i. e. acid generator, that |lt is exceedingly active and will unite in some way or other with any substance that can contain it, and may alone supply the acid requisite for corrosion, and is the supporter of ordinary combustion and the vital part of the atmosphere ; jfor, as they have been created, neither animal nor vegetable life could be sustained without oxygen being present in the atmosphere, nor could there be water or even atmospheric air, aior clay, flint, potash, lime, magnesia or soda. Various com- plications are involved in considering the mutual influence of animals and plants on each other and on climate. The food (of plants may be said to be chiefly carbonic acid, water and j&mmonia, and there is, as it were, a constant interchange of jfood between the animal and vegetable creations, the former helping to supply carbonic acid to the plants, and the latter oxygen from the under surface of the leaves. In fact, so long as man exists on this earth in its present state, it is c 18 COREOSION AND ITS PKEVENTION. decreed that oxygen and carbonic acid shall be present. As carbonic acid (C0 2 ), which is always in the atmosphere in varying quantity, in presence of moisture, is so active a corrosive agent, what are the chief causes of it? The breath of animals, the decomposition of vegetable and animal matter (hence the importance of no vegetable substance resting upon a painted surface), limestone, chalk and all calcareous stones in which it exists in a solid form, the presence of an acid setting it free. It is well known that carbonic acid is most likely to accumulate to a noxious extent from the fermentation and putrefaction of decaying vegetable and animal matters, and that by throwing fresh slaked lime into such places it prevents the accumulation of carbonic acid by absorbing it, and produces carbonate of lime, and that when the lime is slaked all its carbonic acid is driven off by the heat, and that, therefore, it is free to combine with more, and that heavy rain also prevents its accumulation by dissolving it. It also renders water in which it is dissolved slightly acrid, and it is a narcotic poison, and although gases diffuse themselves through each other, carbonic acid gas is heavier than air, and therefore is present in some places in fatal quantities near the ground and bottom of wells ; and although a dog will be at once affected by it, his mouth not being sufficiently high above the ground to be beyond its noxious influence, a man may walk over the same place and suffer no inconvenience. Obviously, there is more carbonic acid present in crowded cities and manufacturing towns than in the open country. The quantity in the atmosphere is however very small, it being approximately, according to Dr. Angus Smith, in the open parts of London, 3 parts by volume in 10,000 parts of air. In the hills of Scotland 3*3 per 10,000, and in London streets 3*8 per 10,000. In the streets of Manchester during fog 6*8 per 10,000. In close buildings, it averages 16 per 10,000, and usually from 7 to 10 parts in houses. There are other agents, whether caused by smoke, the steam of locomotives, or heat, that accelerate corrosion, and SOME CAUSES OF CORROSION DESCRIBED. 19 among them sulphuric and sulphurous acid ; others proceed from water or the atmosphere, for neither rain nor dew are pure water, as they contain acids and salts in solution, chlorine, and ammonia, the latter being especially found wherever organic decay is in progress, as in farm-yards, stables, sewage, &c. Any substances which on decomposition cause free acids to separate from them act as corrosives, and if they are subject to steam pressure the deleterious effect is much accelerated. A carbonaceous deposit on iron or steel also tends to assist corrosion by acting as a nucleus to retain moisture and acids, and also by condensing gases in its pores, and by inducing galvanic action, carbon being electro- negative to iron. The influence of local circumstances upon the durability of materials is demonstrated in the Houses of Parliament at Westminster. Great care was taken in selecting the kind of stone so that it might be durable, and a magnesian limestone was chosen as it was known a very large number of buildings had been erected with that material in the middle ages, and that they were in an excellent state of preservation. No doubt the stone would have been almost proof against corrosive influences if the Houses of Parliament had been built in the open country, or in a country district, but to the atmosphere of large towns it had not been subject in the old buildings, and therefore decay occurred, which, however, has been arrested by treatment and a coating. In the case of iron or steel, corrosive influences would also be much increased, for it may be said the magnitude of the corrosion of the same metal is chiefly dependent upon the position and circumstances in which a structure is placed, and in great measure any difference in chemical action is so caused, hence the importance of a thorough consideration of the location and peculiarities of any district in which a corrosible engineering structure is to be erected. Corrosion is affected by the quality of the coal used. In London, the coal burnt is of much better quality than in manufacturing towns ; thus, in Manchester, it is said, the c 2 20 CORROSION AND ITS PREVENTION. coal has about 2 per cent, of sulphur in it, or about twice that of the coal usually consumed in London, and therefore the amount of sulphurous and sulphuric acid is greater, and such corrosive influences are increased, and are particularly powerful in manufacturing towns, whereas in the open country it is hardly felt ; and is an addition to the other corrosive agents, such as carbonic acid, moisture, and the usual deteriorating elements present almost everywhere. Coal and coke of very good quality contain about 1 per cent, of sulphur. Iron or steel immersed in water are not subject to such severe decaying influences as when they are placed in a damp atmosphere, or where there is little evaporation. Dampness is not mere moisture, or one approaching saturation, and for corrosive purposes may be defined as being that state in which moisture and air promote vegetable and other life conducive to rapid decomposition, especially when assisted hy a warm temperature and absence of light ; the result being a continual combustion or fading away of the substance by chemical action. All plant life, whether microscopic or not, has a powerful decomposing influence. The chief aim is to prevent the spores of such life being able to find a con- genial soil or one upon which they can germinate, and also to prevent the particular conditions under which they each exist. Slime, or dust in a damp state, is frequently a film of microscopic or other plant life. The capillary action that occurs in materials has an important influence on their durability, and within its limits decay and corrosion will be considerable. A contributory cause of corrosion is vibration, as it reduces the remaining strength of any weak place, greatly tends to produce crystal- lisation, opens any crack in and shakes loose any paint not closely and firmly adhering, and it is a general deteriorating influence. Vibration is not only experienced in structures exposed to a rolling load, but also in manufactories, warehouses and more or less every structure, and its probable amount and SOME CAUSES OP CORROSION DESCRIBED. 21 effect should be considered. The fact that steamships which are subject to constant vibration from engines, shocks of waves, &c, corrode quickly does not support the idea that vibration lessens corrosion, although the inside of vessels are not exposed to drying airs, and saline particles which get into the holds do not evaporate, and so the plates are much more severely tried than any rails. Bails in sidings corrode much quicker than those under frequent traffic as in a main line, but this is probably caused by their not being subject to concussion, and the rolling, abrading and cleansing action of the tires on the heads, to the dirtier or unclean state of the rails, to the corrosive influence of the general debris and deposits of a goods siding, the comparatively undrained state of the ballast, its somewhat necessarily soiled condition, and, it may be, the less durability and strength of the rails used. Consideration of some of the causes of corrosion and decay is useful in order to ascertain the reason of its occur- rence. Combustion may be described as a chemical action in which the union of one body with another is attended with development of heat, and, under ordinary circumstances, with an evolution of light. The experiment of the combustion of iron and steel wire in oxygen gas is well known, and is an example of this. Iron by some authorities is not considered to decompose water in the absence of air at ordinary temperatures without contact with some substance electro-negative to it. Professor Tyndall's experiments and labours showed that air is laden with myriads of germs and agents of decomposition ready to settle down and develop upon matter suitable to their growth. Fermentation has been described as a change in the elements of a body composed of carbon, oxygen and hydrogen without nitrogen, but putrefaction as a change effected in the elements of a body composed of carbon, oxygen, hydrogen and nitrogen. Dr. Brewer has briefly and well explained that " the carbon, oxygen, hydrogen and nitrogen of the original substance, being separated by decomposition, reunite in the following manner : (1) carbon and oxygen unite to form carbonic acid ; 22 CORROSION AND ITS PREVENTION. (2) oxygen and hydrogen unite to form water ; (3) hydrogen and nitrogen unite to form ammonia. When bodies contain- ing sulphur and phosphorus putrefy, the sulphur and phos- phorus unite with hydrogen, and form sulphuretted and phosphuretted hydrogen gases." Yapour is an elastic aeri- form fluid, which may readily he converted into a liquid or solid merely by change of temperature. A gas is an elastic aeriform fluid, which cannot be made to change its state, except by the application of artificial pressure and intense cold. The reason of some things being solid, gaseous, or liquid is that the particles are closest in the solid, furthest apart in the gaseous, and the others are liquid ; heat changing a solid, like ice, first into a liquid, and then into a gas, because it causes the particles to become further apart, thus ice melts into a liquid, and additional heat changes it into steam. Corrosive action increases with the temperature, and at freezing point Fahrenheit it is very little. The effects of temperature are very considerable in chemical action, and a certain degree of heat or cold is necessary to produce the greatest activity. Experience in Sweden indicates that it is the constant changes of temperature from heat to cold, or cold to heat, that weakens the iron in rails. Mr. Sandberg * states that this explains why the breakages of rails generally take place in the autumn or spring, and not when the metal is for some days at a constant temperature. Repeated changes of temperature, even from 65° F. to 212° F., produce, it has been shown by Mr. H. Tomlinson, F.E.S., a considerable amount of internal friction between the constituent mole- cules of the metals, which tend to modify some of its physical properties. 1 Steel is more sensitive to this influence than wrought iron. Mr. Andrews, F.E.S., has suggested that this may probably be the cause of the greater number of breakages of steel axles in India and Canada, &c, where variations of temperature are extreme. In the very large number of experiments of Mr. Andrews, F.E.S., the metal * See ' Engineering,' January 13, 1888. SOME CAUSES OF CORROSION DESCRIBED. 23 was kept at 0° F. for cold tests, and 100° F. for warm, and the appearance of fractures of the warm test was of a fibrous character, the cold had a more fine-grained and crystalline aspect. Statistics have clearly shown that railway axle fractures are much more likely to occur during very cold weather than any other. In Eussia, about 50 per cent, more steel axles then fail. Experience on the Grand Trunk Kail- way confirms this. However, increased rigidity of the permanent way no doubt contributes materially to the general results, and not merely the metallic change. Sudden changes of temperature will assist any corrosion which has occurred. In another way, alternations of mild and very cold weather aid corrosion, for they will cause water of con- densation to drip. In a similar manner dew is formed, it being the vapour of the air condensed by coming in contact with bodies colder than itself, and so is rain, for it is cold condensing the vapour of the air when near the point of saturation which causes water to fall. The variations of temperature have been registered to be not the same on both sides of a metallic bridge. This may cause a condensation of vapour ; for instance, at an arched bridge at Liege it was found on one occasion, between 8 a.m. and noon, that the range was about 29° F. on one side, and 16° F. on the other. The face or outside girders or ribs of an arch being the most exposed to the sun, wind, rain, &c, are also subject to greater variations of temperature. In some experiments on the arched steel ribs of the Morand and Lafayette bridges over the Ehone at Lyons, it was found the maximum temperature in the metal directly exposed to the sun occurred two hours earlier than the maximum shade temperature of the air. Experience has proved that very small quantities of carbonic, sulphuric and sulphurous acids will cause rapid corrosion of iron or steel. Where the better qualities of coal are consumed, as in non-manufacturing districts, there is not so much sulphur in the air, and consequently the amount of sulphurous or sulphuric acid to be brought down by rain, 24 CORROSION AND ITS PREVENTION. and so deposited upon any metallic surface, is reduced ; how- ever, sulphur set free from any substance in the presence of moisture will act corrosively and deleteriously on iron or steel. The smoke, vapour and gases escaping from ordinary locomotives have been shown by analysis to contain carbonic acid, which is considerably heavier than air ; carbonic oxide, which is a little lighter than air; vapours and sulphurous or sulphuric acid in greater or less degree according to the quantity of sulphur in the coal burnt, and in addition chlorine and ammonia ; the ordinary oxidising agents being also more or less present. Chlorine, which is heavier than air, is often employed as an oxidising agent, as, owing to its great affinity for hydrogen, it has the power of decomposing water, and setting free its oxygen, which, at the moment of its liberation, will combine much more readily with other elements than it will when perfectly isolated, and as the atmosphere of tunnels or underground ways is usually in a damp state, the conditions necessary for its most active attack are present, as with certain other elements which are es- pecially energetic in their nascent condition. Chlorine has very powerful affinity for most metals, and moist chlorine is more energetic in its action than the dry gas. Thus in tunnels, subterranean ways, tubular bridges, and wide over- bridges on railways, all the influences required for rapid corrosion are present, for the atmosphere is then usually more or less surcharged with moisture from the aqueous vapour escaping from the locomotives, and condensation is rapid, and the tendency of the aqueous vapour to take up carbonic acid and sulphurous gases has also to be considered. Dr. Grace Calvert's experiments showed that carbonic acid and moisture produce rapid corrosion, and in the case of the structures mentioned must be added sulphuric acid and chlorine, two very energetic corrosive agents. Therefore it is specially necessary to constantly protect iron or steel when placed under such conditions, and also to consider whether it may not be better to employ cement brickwork or Portland cement concrete, but not masonry, unless the stone employed SOME CAUSES OF CORROSION DESCRIBED. 25 will resist the acids, in preference to a metallic system of construction, and any increase in the first cost may be justified "by the much longer life of cement brickwork or concrete, and the very little attention required to maintain their surfaces in good condition. It may be well to remember that disinfectants, as tbey are used in many structures, generally act by means of their primary ingredient, whether it is chlorine, sulphur, or ni- trogen, producing by their decomposition in presence of moisture, hydrochloric, sulphuric, or nitric acid, all of which corrode iron or steel. A very careful examination by Mr. W. Thorner of the corrosion of the permanent way in the tunnels of the Weilburg and Nassau Eailway * was made in 1887. Eight samples of the rust, drainage water, mud and calcareous deposits in the tunnels were taken from all the localities and carefully ana- lysed. The results showed that except where calcareous infil- trations occurred, the composition of the rust in the tunnels was substantially similar from all localities. Chemical analysis proved the essential constituent to be ferric oxide, partly present as hydrate, with variable, but always small, quantities of silica, alumina, lime and magnesia ; carbonic acid, when present, being in connection with the two latter bases. The abnormal constituent was sulphuric acid, or more correctly an oxidised sulphur compound, which occurs in all proportions from 0-3 to 7-9 per cent., the actual state of combination being doubtful, but most probably in the form of basic ferric sulphate. Nearly all the samples showed a slightly acid reaction, but only gave a very small amount of sulphuric acid when exhausted with water. No rust contained iron as sulphide. The origin of the sulphuric acid may be most readily ascribed to the oxidation of sulphurous acid derived from sulphur in the coal and dissolved in the condensed steam from locomotives. This, trickling down the walls, is ab- sorbed by the ballast and slowly oxidised by atmospheric air, * ' Stahl und Eisen,' vol. lx., 1889. 26 COREOSION AND ITS PREVENTION. producing sulphuric acid, which, penetrating the rails by capillarity, attacks them, forming ferrous and ferric sulphates. This, however, does not account entirely for the result, for rust obtained from ironwork exposed in a garden at Osna- briick gave 0*8 to 5*5 per cent., and these the sulphurous acid in the exhaust steam of the locomotives could hardly contribute. A further series of experiments showed that nitrous acid and nitrites are formed by the direct action of iron upon air or water, and these substances are well known to act most energetically in the conversion of sulphurous into sulphuric acid. The existence of free sulphuric acid in the exhaust steam of a locomotive was determined experimentally on a heavy goods engine travelling at a speed of 30 miles per hour. The compositions of the incondensable chimney gases were found to be : — „. . „ . First Second Experiment. Experiment. Carbonic acid 5*4 6*1 Oxygen 13-3 13-0 Nitrogen 81-3 80-9 Carbonic oxide — — Sulphurous acid — — 100-0 100-0 Free sulphuric acid, ammonia and ferric oxide were found, but no nitrous oxide in the water absorption apparatus used for collecting and obtaining the composition of the gases. The amount of sulphuric acid evolved per hour by the engine was estimated, after deducting that required to saturate the small proportion of ammonia present, at very nearly 5 lbs., a quantity which under appropriate conditions is likely to act very destructively upon ironwork. The action is likely to be strongest in tunnels where the rock is not very wet, and the exhaust steam is without means of rapid escape, and especially in those where the rock is poor in carbonate of lime. Where the ground is very wet and provision is made for drainage, the soluble gases are taken up by the water, and removed in a very diluted form, before they have much SOME CAUSES OF CORROSION DESCRIBED. 27 chance of doing harm. Carbonate of lime acts beneficially by directly neutralising the sulphuric acid and converting it into gypsum. It was observed that the heaviest rusting took place between rail and sleeper when both were of iron ; the acid water being introduced by capillary attraction, forms layers of rust which by continual accretion attains a thickness of 0*39 to 0*59 inch. Wooden sleepers, on the other hand, protect the iron on account of their low con- ductivity, which prevents the precipitation of the acid gases, if the surface of the sleepers is sufficiently covered by the ballast. The methods suggested for the prevention of corrosion in tunnels after consideration of the results of these experiments are: (1) Covering the ironwork as far as possible with heavy or so-called carbonised tar, not ordinary gas tar, or asphalt applied to the metal in a properly clean condition and renewed at intervals. (2) The precipitation being greatest at the coldest points in the tunnel, i.e. on the ironwork, the latter should be covered with ballast, leaving only the rail head exposed, small limestone being the best ballast for this purpose. (3) Effective drainage of very wet tunnels. (4) In rock tunnels, when the rock is poor in carbonate of lime, the ground should be completely covered with small limestone fragments, or when that is not sufficient, the roof and walls should be lime-washed from time to time. Milk of lime sprinkling is not recommended. (5) The engines should make as little exhaust steam as possible in tunnels, and dispense with its use wherever it can be done. These, experiments, being conducted especially with the view to the prevention of rust under such conditions, are particularly valuable. The rock, &c, tunnels were in coarse- grained dolerite, Devonian limestone, and shales. In sound greenstone, containing no carbonate of lime, the rusting was the strongest. In the cretaceous marls containing 80 per cent, of carbonate of lime, corrosion was very slight. The Devonian schists were poor in lime. One sample of rust was from the Underground Railway, London. 28 CORROSION AND ITS PREVENTION. The principal impurity in the Mont Cenis tunnel is carbonic acid, and there is an excess of aqueous vapour, and corrosive influences are considerable. The carbonic acid sinks to the lower part of the tunnel, and therefore the air the workmen breathe is affected, and not that of passengers in the carriages. An analysis of rust taken from a bridge on the Pennsylvania Eailroad, made by Mr. Kent, of the laboratory of the Stevens Institute, U.S.A., showed that in the water solution, by which method the tests were made, it was found that iron, ammonia, sulphuric acid and traces of sulphurous acid and chlorine were present, and carbonic acid in considerable quantity. These are readily formed by the fuel and steam of locomotives, and when they are dissolved in moisture, an acid and saline solution is formed having a powerful corrosive effect. Some experiments made by Mr. Kent indicated that " sulphurous acid is rapidly changed into sulphuric acid in the presence of iron and moisture, and that the iron is thereby rapidly corroded." The sulphurous acid escaping from locomotives must therefore be considered one of the most active agents of the corrosion of railway bridges. The weight of vapour which will saturate a cubic metre (35-317 cubic feet) of air, at different temperatures, is as follows : — Centigrade. Fahrenheit. Grammes. Lbs. At avoirdupois. - .10° or 14° 2-302 or •0051 - 5° 55 23° 3-406 )5 •0075 - 0° )5 32° 4-915 55 •0108 » + 5° )5 41° 6-845 •0151 + 10° 'J 50° 9-445 55 55 •0208 >• + 15° » 59° 12-860 •0283 » + 20° )5 68° 17-311 •0381 + 25° 55 77° 23-067 55 •0508 To absorb the aqueous vapour without sensibly altering the hygrometric condition of the air, a difference in the tempera- ture inside and outside the Mont Cenis tunnel of 15° C. in winter and 5° in summer is sufficient, and as this always SOME CAUSES OF CORROSION DESCRIBED. 29 exists, no additional quantity of air is required for its absorption. The gases from the locomotives passing through the Arlberg tunnel, although great care is taken in the selection and combustion of the fuel, have, it is stated, caused it to be necessary to renew the whole of the ironwork after being ten years in use. The traffic is not considered to be great. The atmosphere contains nitrogen, oxygen and numerous other gases, such as carbonic acid, nitrous, sulphurous, and sulphuretted hydrogen, ammoniacal and hydrochloric gases, varying very considerably in different localities, and in the same, according to the height above the ground. Some gases are especially to be found in certain localities, for instance, carburetted hydrogen in and around marshes, nitrous acid in districts subject to storms. Generally, town air contains more ammonia than country air, except in such special places as farm-yards, &c. ; and night air more than day air. The air of a sea-shore has generally a marked proportion of hydrochloric acid in it. 30 CORROSION AND ITS PREVENTION. CHAPTEE III. THE CHARACTER AND QUALITY OF IRON AND STEEL, AND ITS RELATION TO CORROSION. It is generally agreed that the magnitude and rate of corrosion much depends upon the quality and homogeneous- ness of the iron, for if films of oxide are present in the metal they come in contact with the iron, and then local action is set up and corrosion is uneven. It has been found in cable work that annealed iron of good quality does not deteriorate in sea water nearly so much as the inferior makes do, and that a less mass of the best iron is a better protection against corrosion than a much heavier one of inferior quality. It has also been found that hard cast iron with an even close grain resists corrosion to a considerable extent, but that soft foundry cast iron quickly corrodes, and that the whiter descriptions of cast iron corrode slower ; thus white cast iron of good quality resists corrosion better than grey cast iron, and the latter better than soft foundry iron, but grey cast iron is to be preferred for other reasons. ' The Admiralty have for many years recognised that corrosion varies with the quality of the material, and have fixed certain chemical tests for steel or iron, which, if the metals satisfactorily sustain, are found to be safeguards against accelerated corrosion, and the metal then requires only ordinary care and the adoption of the usual means of pre- servation to prevent corrosion. In fact, want of uniformity in the composition of iron or steel resulting in laminations and imperfections in wrought iron, and heterogeneousness in steels, is now acknowledged to have a decided influence in CHARACTEK AND QUALITY OF IKON AND STEEL. 31 promoting local corrosion. Also impurities in those metals within reasonable limits, that is, so as not to interfere with other necessary qualities, such as working and tenacity, are considered to affect their durability. It sometimes occurs that all the plates of a structure seem to be corroded equally ; it also happens that one plate may be more corroded than another, although apparently subject to precisely similar influences and conditions ; the only reason for such variable- ness in the magnitude of the corrosion is considered to be the want of homogeneousness of the metal, variation in its electric condition, the proportions of carbon, &c, being dis- similar, and one plate or bar having been manipulated, whether by hammering, bending, or in other ways, more than another, for the effect of bending iron or steel is to compress the particles or fibres on one side of the neutral axis, and to elongate them on the other. Examination of the bared iron of a number of painted plates that had been submerged, has shown that the varia- tion in the degree of corrosion was almost unaccountable if it did not arise from want of uniformity in the quality of the plate, for while some plates are equally affected, or almost free from corrosion, others will be very variably attacked ; local action, whether galvanic or not,^ having taken place to promote corrosion. In boilers, for instance, hardly any two similarly corrode, some rusting quicker than others. Constant care in the manufacture in order to main- tain the same quality and purity of the metal is, therefore, one of the chief factors in preventing corrosion, and the greater the purity, homogeneousness, and freedom from foreign matters of the metal, other conditions being similar and in accordance with specified requirements relating to strength, the less will be the corrosion. It is well to remember that what may be called com- mercially manufactured metals used in construction are not generally in an absolutely pure state. This arises not from intention on the part of the maker, but from other causes, such as the process of manufacture and contact CORROSION AND ITS PREVENTION. with the vessels and the necessary plant used in its pro- duction. A difference of surface caused by methods of production of the metal will lead to a variation in the rate and character of the corrosion, not only because of the alteration of the surface or skin, but also because of the difference of texture. The efficiency of the processes of manufacture from smelting to rolling, whereby hard and soft particles in the mass may be avoided, will affect the homogeneousness of the metal and the rate of corrosion, and complete fusion or melting are necessary to obtain it. The extreme quantity of metallic iron in commercial bar iron is about 99-8 per cent., but ordinary commercial iron contains about 98 to 98-3 per cent, of pure metal, the alloying elements being phosphorus, silicon, sulphur, carbon, manganese, &c. In ordinary cast iron, the pure metal would be about 93-5 per cent., and there would be about 2 • 5 per cent, of graphite. The various compounds do not corrode equally with pure iron, for a dif- ferent material is affected, hence the composition of a metal will influence the rate of corrosion. Care in the selection of the most suitable mixture to obtain homogeneousness is required so that metal is not used that may have been compounded without much regard to uniformity of texture and character. Analytical chemists of the highest authority are agreed that the presence of other metals in small quantities, such as iron, tin, zinc, arsenic, and the like, promotes the formation of insoluble scale on copper. In the Minutes of Proceedings of the U.S. Naval Institute, 1886, an account is given of the corrosion of the copper of the Juniata. Several plates, it was found, on the vessel being put in dry dock in October 1882, were so corroded as to be nearly or completely per- forated, the immersed surface having become covered with a pale green, earthy-looking film, which blistered, many of the blisters split, and the coating flaked off to a great extent. It was considered that the copper beneath the spots of scale was softer than the surrounding copper, and made the CHARACTER AND QUALITY OF IRON AND STEEL. 33 difference required for setting tip the corrosive action, and that the corrosion was principally due to the presence of spots of oxide of copper on the surface of the plates at the time they were put on, and that the action arose from unsound cakes of copper heing taken for rolling, which contained cavities or air cells such as are occasionally pro- duced, and that by rolling, the cavities are extended in the direction of the length of the bar, and when the bars are rolled into sheets, in the direction of the width of the sheet. In granular iron, the degree of homogeneity, purity, and density appears to be considerable, and the smaller the crystals, the harder it is, and corrosion is likely to be reduced. In fibrous iron, the cellular and imperfectly amalgamated condition is considered to be caused by lower- ing the temperature to such an extent that imperfect welding results, as the bar is drawn out of the rolls in too cold a condition for the particles to unite ; they consequently slide on each other, the result being fibrous iron, whereas it is declared granular iron requires a high temperature to produce it, and then the iron is homogeneous, or, in any case, much more so. It has been affirmed that iron should be granular, as it is the essential property of welding and amalgamation, and that all other textures are the result of defective and imperfect welding in the process of manu- facture. However, other matters may influence the result, such as the composition, rapidity of rolling, and pressure. In a fibrous metal, rust seems to eat its way into the laminations or sandwich of iron and dirt, as Dr. Siemens termed it, and forms a way for the entrance of corrosive influences. In a homogeneous metal, corrosion appears to be more even and regular over the surface, and therefore more of a granular nature, but even homogeneous metal is subject to pitting from corrosion ; in fact, in any compound substance, corrosion will generally not be even and regular, and it can hardly be otherwise ; however, the more a metal partakes of a homogeneous character so is corrosion likely to be uniform. Bad welding, unequal working, hardness, D 34 CORROSION AND ITS PREVENTION. roughness, or polish of the surface will also cause a difference. In steel, uniformity can be approached more than in iron ; however, the cinder in iron and the scale of steel are different. The former is a dielectric glassy substance and has no corrosive influence, whereas the scale produced in rolling steel has a very deteriorating effect, as it is a magnetic oxide negative to the steel, and therefore corrosion is rapid, as also if such scale is rolled into steel plates. Hard and crystalline irons are generally considered to be less oxidable than ductile or fibrous iron. Professor Kick, of Prague, after many experiments, which he only regards as indicating the probable general rule, not as proving it, with nitric, sulphuric, and hydrochloric acids and their combina- tions, with mordants composed of the salts of copper, &c, found that soft and fibrous iron of very good quality, also fine-grained iron, when attacked by the acid, is uniformly so, and with a limited elimination of the carbon, the surface retaining a dull lustre, a few incised specks and cinder-like holes being only observable. Coarse-grained iron and hot- short iron were found to be more powerfully attacked. In 10 minutes the surface became black, in half an hour, a black muddy deposit appeared, which could be washed off, and a number of small holes were distributed over the surface. This iron was not uniformly attacked, some portions being more deeply affected than others. Malleable iron or annealed iron rusted quicker than wrought iron, the action of the acid being very powerful and irregular. In Bessemer steel and cast steel, the action of the acid produces very fine fissures. In cast iron, the attacked surface presented a tolerably uniform grey colour. Imperfect welding some consider is due to the defective mobility of the mass, the consequence of insufficiently high temperature to admit of the union of the particles. "Welding in a non-oxidising atmosphere may be regarded as a step towards perfection. The electric welding of angle, T-irons and similar sections, pipes, is another ; and, perhaps, the weak- ness of joints made in such sections of iron may be so CHARACTER AND QUALITY OP IRON AND STEEL. 35 lessened, and collecting places for corrosive influences be thus reduced ; but there has hardly been time to determine the effect of electric welding on the corrosibility of the metal. It is well known that there is a diminution of cohesive strength consequent upon welding. Some experiments with welded bar-iron have shown this to be about 20 per cent., and that resistance to sudden strain and to the action of accumulated work is, as a rule, even more impaired by welding than the resistance to constant pressure. Material that is almost identically composed frequently has its powers of durability affected by its crystallisation. Quartz and flint, marble and chalk, are examples of this action. Any operation that deteriorates the coherence or density of a substance aids corrosion, because it not only affects the coherence of the mass, but separates the molecules and exposes them to the active agents of decay. In making up a bloom by ordinary piling, if the piles are not of uniform quality the finished iron cannot be uniform, and therefore corrosion will be unequal, and while the edges may be irregularly corroded, the surface of a plate will be pitted, and bars may become furrowed ; but even assuming the piles are of uniform quality, the surfaces of the bars which make the bloom, in being heated, have scale formed upon them, and although some of this skin or scale is removed by subsequent processes, such as by hammering and pressing, a portion remains in the iron after it is rolled and in its finished state, the dark streaks when iron is broken, notched, or manipulated in a lathe showing this. Apart from any moisture being confined upon the surface and the blistering of the paint, which are causes that may induce local corrosion distinct from any inherent properties of the paint, the unequal corrosion of wrought iron may be said to principally arise from electric action taking place between the streaks or veins, or portions or particles of scale incorporated with the metal in the processes of manufacture, which causes the mass to be one possessing varying electro- d 2 36 CORROSION AND ITS PREVENTION. chemical properties, consequently electric action can "be generated. As every metal is electro-positive to its own oxide, and as the positive metal is more quickly acted upon than the negative metal, it follows that the former must be deteriorated, and that decomposition will not be uniform in consequence of impurities causing local action ; and further, there is nothing to prove that this internal action can be arrested or suppressed ; in fact, reliable experiments support the belief that when electric action is set up, it will continue for a long time, and may not cease from various causes, and may even change in character ; it cannot be said, however, that so far as the question of films in wrought iron are concerned and their continued action on the mass, it has been determined whether the electric action ultimately ceases or not, although it becomes less.* Iron in cooling does not crystallise in the same form, for its structure varies to some extent according to its composi- tion, therefore corrosion will not be exactly the same. The microscope shows that crystals, streaks, and what may be called needles, spots, regular shaped figures and cavities can be discerned. Corrosion can, therefore, hardly be considered as likely to be uniform. It is known the segregation of substances held in solution requires time. If it were possible to solidify liquid metal in a moment, it would be more homogeneous than when it is gradually cooled. In situations exposed to the atmosphere, water, rain, &c, and the corrosive influences which may be said to almost always prevail in engineering structures, dense, compact materials of uniform texture should be used, and those composed of elements, in any case, not readily soluble, and if practicable, insoluble in water, should be adopted. Although iron and steel are apparently compact, it is known there are pores in the mass, and gases condensed in them have been extracted. Mr. Thorner at Osnabriick made some experiments with iron, and they showed that the ■* Vide Chapter VIII., on "Galvanic Action and Corrosion." CHARACTER AND QUALITY OF IRON AND STEEL. 37 porosity of volume per cent, of mass in ordinary cast iron was 1*41 ; in different kinds of Bessemer steel, 0*41 to 1*20 ; in a locomotive tire ingot, 0 • 57 ; in a locomotive tire ingot intended for making hard forgings, 0* 97, the remainder being rail ingots. One described as basic ingot iron, 1*95; another as basic steel, 1 • 22 and 2*17. It is considered the experiments showed the porosity of iron and steel is subject to considerable variation in the different kinds of steel, that although strength and porosity are generally in direct relation to each other, the stronger being as a rule the denser, several notable exceptions have been observed ; and that these latter are consequent upon the irregular distribution of the pores through the mass. The pores were found to be of micro- scopic fineness, and must not be compared with visible blow-holes.* These experiments confirm the statement that corrosive influences may penetrate a mass of iron or steel. It seems to be agreed that iron and steel absorb at high temperatures gases which are partly disengaged when the temperature is lowered, and that the diseng'agement may be accompanied by an appreciable change in the chemical composition. Some experiments of Mr. Parry (1872), and. afterwards of Herr Muller, show that gases are enclosed in iron and steel; for instance, hydrogen, nitrogen, carbonic oxide, are present in them, and the gas by volumes in percentage of that of metal varies considerably, depending not only upon the nature of the metal but, as might be expected, upon the method of manufacture, for the cavities contain gas which is a poor conductor of heat, and so in cooling inequalities in texture and strength result and corro- sive action follows. In addition it is well to remember that moisture and air will penetrate the microscopic cavities in any metal. Flaws in plates occur owing to defective rolling, &c. Air, vapour, water, steam, &c, may penetrate into these from the ends of the plates or otherwise, and increase corrosion. Castings are affected by cooling, contraction and * See ' Stahl und Eisen,' vol. vi. , 38 CORROSION AND ITS PREVENTION. " drawing " of the masses, and various defects in moulding, such as the sand being too compact, and so not allowing the gases to escape from the metal through the sand, thus causing scabbing and honeycombing; also from it being too loose, a rough, uneven surface being the result. This is often the case in greensand moulding ; the weight of a green- sand casting is usually some 10 per cent, more than when dry sand moulds are used, consequent upon " swelling " all over owing to looseness of the sand and its capability of being compressed by the metal. If the whole mass were everywhere in absolute molecular contact it would be air- proof, hence anything that is done to consolidate it and press it together and cause greater cohesion tends to reduce corrosion. Unequal corrosion arises from a plate or any portion of a plate being more porous or less impervious to moisture or air than another ; and without entering into the question whether it is possible or impossible for a liquid to pass into the interior of a piece of iron, it would seem that it is, for gases are known to be diffused through iron, and the sweating of the cylinders of hydraulic presses has been considered as a proof that such action is possible. It has been found when the natural skin or surface left in casting has been removed from cast iron, water under a pressure of about 3 to 3£ tons per square inch will pass through the pores of the iron. It would permeate the upper surfaces and into the interior at much less pressure. In some substances, as wood, sponge, the absorption of water takes place with great force, and the power of capillary attraction has to be considered. The different processes of manufacture of iron or steel from the rough mass converted by hammering, pressing, or rolling, &c, show that cavities and interstices are present in iron and steel, and that attention to the uniform good quality of the metal and care taken in the manufacture can alone reduce them to a minimum. It may be accepted generally that the more porous and less carefully manufactured iron or steel will not only become corroded the quickest, but will Toe so CHARACTER AND QUALITY OF IRON AND STEEL. 39 more or less throughout the mass ; whereas in a thoroughly- homogeneous and very dense metal the corrosion would be superficial. If any moisture in the surface metal to be coated was extracted by the paint, and the paint hermetically sealed it from air and moisture and gases of all kinds, corrosion, if any, could only proceed from gases or moisture in the pores of the metal, aided, as it might be, by galvanic action being set up consequent upon want of homogeneousness in the metal and the presence of impurities. The texture of wrought iron will to some extent indicate the quality of the iron, for if it be good it will be fine, uniformly close-grained, and of a silvery grey colour ; if common and inferior iron, it will have a coarse granular fracture and look something like cast iron, but the quality cannot be reliably ascertained from a mere inspection, not even if aided by the microscope, for mechanical and chemical tests are necessary. The colour of iron in its crude state is chiefly dependent upon the proportion of carbon in it, and therefore in appearance it is either of greyish tint, mottled grey, very pale grey merging into white, or nearly so. By many it is considered carbon is regarded as affording protec- tion against corrosion. Wrought iron when bent is com- pressed on the inner and extended upon the outer side, the result being that any cavities on the inner are lessened while on the outer side they may be elongated, unless the reduction of area neutralises that action, which it will generally do to a small extent. The metal by bending is thus caused to be of unequal texture, and the more open portion will be liable to the most corrosion, if not carefully and uniformly protected. In some recent (1894) French experiments it was found that acids applied to irons or steels when the latter are strained above their elastic limit corrode the surface with greater rapidity along the lines of deformation while the metal is subjected to the action of a disturbing force, and that there are zones of deformation between which the metal is unaffected. This is probably caused by the disturbance 40 CORROSION AND ITS PREVENTION. and opening of the structure and grain of the metal and by- its increased porosity. There is no doubt that a specification which demands a metal of very good and homogeneous quality requires one which is much more able to resist corrosion than if the common kinds of iron were allowed to be used. However, although this is something gained, even the best qualities of iron or steel should be protected as much as possible against corrosive influences, and every care taken to ascertain that the paint or composition* does not contain anything that can corrode or otherwise injure the metal, such as acids, which oil, necessarily so much used in oil paints, frequently does. Mr. R. Mallet, in the early experiments made on the corrosion of different kinds of iron in pure and foul atmospheres, and in clear, foul, acidulated, and salt waters (see his original reports), found that when the metal was exposed to water holding air in combination, the surface corroded uniformly or in patches, either by rust or conversion of the original iron into plumbago, and that the extent of the corrosion depends upon the want of homogeneousness of the surface or metal, or its density or hardness, or in the combination of the carbon with the iron. Mr. Andrews' recent experiments have shown that it is important to guard against accidental galvanic action in structures, and that bars, plates, rivets, bolts should be as near as practicable of a similar material, temper, and composition, &c. "With regard to the method of manipulation and uni- formity of the texture of iron and steel, hydraulic forgings are to be preferred to hammered forgings, for work done by the press is more accurate, destructive vibrations are not caused, and the force of the press is transmitted in its entirety to the ingot, whereas the useful effect of the hammer is not more than a comparatively small percentage of the theoretical effect, because of the elasticity of the blow. Press-formed ingots are usually finer towards the centre, * Vide Part II. of this book, on ' The Prevention of Fouling and Corrosion in Submerged Structures and Ships.' CHARACTER AND QUALITY OF IRON AND STEEL. 41 those produced by the steam-hammer are poorer at the centre. In hydraulic forgings of large size, the action of the press causes the pressure to be more equal without jarring, and it is brought upon the centre of the mass, whereas ordinary steam hammering causes it to be chiefly on the surface, and then not always equally so, and the centre of the mass is perhaps hardly affected. Compressing the liquid metal in the mould immediately after casting, and substituting an hydraulic press for the hammer in the subsequent forging of the metal, makes it of closer and more uniform texture, and therefore it is less liable to corrosion. A heated ingot may resist the blow of a hammer, but the hydraulic press continues to act until it has forced the particles into closer contact, and uniformly so to the eye, whereas the effect of a blow is but for a certain distance and it then ceases, the action of steam-hammering being superficial as compared with that of hydraulic power. Similarly in the case of riveting, a light riveting hammer will not set up the centre of a rivet, or fill the rivet hole. A plying hammer may, and frequently does, but hydraulic riveters do ; however, they cannot always be used. Uniform continued pressure is always to be preferred to any kind of concussion. As any voids in rivet holes will admit air and moisture, bad riveting is sure to result in accelerat- ing corrosion, even if it does not cause it to appear. Taking into consideration that a girder may have been badly riveted, the corrosion and consequent reduction of strength will continue until at last its weakened condition compels its removal, which probably might have been averted if the riveting had been done by hydraulic power, and the metal carefully painted with a really preservative paint. It has been stated that corrosion may vary in strained and un- strained metal. As in the pin system of connections the strain is localised upon certain parts of a structure, whereas in the riveting method of connection it is distributed ; there is an advantage, apart from others, in the rivet as compared with the pin system as being less liable to promote corrosive influences, always provided the riveting is properly done. 42 CORROSION AND ITS PREVENTION". Eolled joists, and iron of similar shape, are of compara- tively soft iron that will lengthen without cracking or tearing under the unequal strains it receives from the grooves in the rolls, and the quality is inferior to that used m ordinary bridge girders, and there is usually an appreciable quantity of phosphorus and sulphur in it. Such iron is generally more easily corroded than the better kinds of iron plates, bars, &c, used in bridge construction. Some unexpected results have been noticed on copper where a thin strip of the surface has been compressed by marking with a tool or been polished, as when so treated it has not corroded in sea-water, although the untouched surface has. This effect is considered to be caused by the compressive action of the tool on its being drawn across the metal, which is thereby made more dense, and by the polished or smoother face preventing lodgment of water upon or against any minute inequalities of the surface and its penetrating the metal. It may be useful to here give a scale * of hardness of metals that was used in the laboratory of the Technical High School at Prague, composed of 17 metallic substances, arranged in ascending order from the softest to the hardest. The tests were made by drawing a cylindrical piece with a conical point six times along a polished surface of the metal to be tested : (1) pure soft lead; (2) pure tin; (3) pure hard lead; (4) pure annealed copper; (5) cast fine copper; (6) soft bearing metal, copper 85, tin 10, zinc 5; (7) cast iron, annealed; (8) fibrous wrought iron; (9) fine-grained light grey cast iron; (10) strengthened cast iron, melted with 10 per cent, of wrought turnings; (11) soft ingot iron, with 0-15 per cent, carbon, will not harden ; (12) steel, with 0-45 per cent, carbon, not hardened; (13) steel, with 0-96 per cent, carbon, not hardened ; (14) crucible cast steel, hardened and tempered, blue; (15) crucible steel, hardened and tempered, violet to orange-yellow; (16) crucible steel, hardened and tempered, straw-yellow; (17) crucible steel, glass hard. Vide ' Technisclie Blatter des Deutschen Polytechnischen Vereines in Bokincn,' vol. xiv. 43 CHAPTEE IV. NOTES ON THE CORROSION OF CAST IRON AND STEEL. Cast iron is a very heterogeneous compound, and care should be taken to ascertain that castings have not been made with the sole object of causing them to appear externally sound enough to pass examination. Iron of one kind, even if very good, is not found to be equal to metal made from judicious mixtures. Cast iron being crystalline, and wrought iron fibrous, corrosion is not similar. If cast iron is hard, not to brittleness or so as to decrease its working strength, of an even, close grain, with the carbon combined and not in the form of graphite, it resists corrosion to its greatest extent. If impure soft foundry iron, or the commonest mixture that has been called cast iron, but which is a mass composed of sand, cinders, scoria, or recrement, and soft, uneven, and open-grained metal having large crystals is used, it will corrode very quickly. On the other hand, close-grained, homogeneous grey iron will not readily corrode, nor will white iron of good quality and even texture. Cast iron when it is very rich in carbon is soft, like plumbago, and can easily be broken. Hence, when by corrosion it is becoming decomposed into such a state, its strength is gradually being destroyed. The process of corrosion in cast iron is different in some respects to wrought iron, and in steel it cannot be said to be entirely the same as either. Cast iron slowly decomposes when it is placed in sea-water, the iron becoming dissolved or extracted from the mass, the remaining substance apparently occupying the same bulk as the original cast iron, there 44 COEEOSION AND ITS PREVENTION. being no easily recognised reduction or increase of section as in wrought iron. The cast iron becomes of a graphitic nature, ultimately closely approaching the condition of plum- bago or black lead, the material being subject to a molecular and chemical change more regular than wrought iron, and corrosion does not occur in similar streaks, grooves, or fur- rows, and is generally considered to be superficial and comparatively slow, but being one of decomposition by a chemical change of the mass, thus altering the character and composition of the metal, the depth of which, at present, cannot be ascertained in a structure without analysis, it should not be concluded because wrought-iron piles have become quickly corroded that all that is required is to substitute cast iron for the wrought iron, and that, because to ocular demonstration, the cast iron appears to be in as good condition as when placed in the work, no deterioration and decomposition have taken place. In fresh water, the corrosion of metal of good quality should be very slow and superficial, although by the ordinary laws of nature some decomposition will take place. The process of corrosion being so different to that of wrought iron, or ordinary oxidation, which is apparent on the surface and ultimately reduces the size of the metal, it is necessary to be very cautious in determining the magni- tude of the deterioration or decomposition, for a considerable loss of strength and cohesion may have taken place without any alteration of volume, or even appearance. This gradual deterioration of cast iron should demand regular inspection of the metal at intervals, and it is advisable that pieces should be submerged under the same conditions as the permanent structure so that they can be analysed and examined. In wrought iron, the reduction in strength may be said to be proportional to the measured corrosion, but in cast iron no deteriorating effects may be visible, and yet exist to an important extent ; for instance, it was found by a high chemical authority in India, Dr. Lyon, on some pieces of cast iron being examined after 4£ years' immersion in CORROSION OF CAST IRON AND STEEL. 45 Bassein creek, which water is pure undiluted sea water having a specific gravity of 1 ■ 028, and contains 3000 grains of solid matter per gallon, of which 1605 consist of chlorine in combination with sodium, magnesium, &o., that they had undergone a change to a depth of ^ inch from the surface, resulting in a small solution of the iron forming this portion of the piles. In commenting on several such experiments, Dr. Lyon stated that in his opinion the action of sea water on cast iron consists in a gradual solution of the cast iron, leaving the carbon undissolved, and as the percentage of iron in cast iron varies within certain limits in different specimens, it follows that, unless the percentage of iron be reduced below these limits, such change must remain undetected unless the iron has been analysed before immer- sion. He considered cast iron so immersed will deteriorate in process of time, and it is a question whether it would not be advisable to guard the iron from corrosive action of the water by coating it with some protective material, or by rendering it electro- negative under the influence of some more oxidable metal, such as zinc. Dr. Grace Calvert, F.K.S., made a series of experiments in acidulated water by immersing grey cast iron, 0*39 inch cubes, made of Staffordshire cold-blast iron. Specific gravity, 7 '858. Weight of cubes, 237 grains. One cube only was placed in a corked bottle holding 31 cubic inches of much diluted sulphuric, hydrochloric, acetic, and phosphoric acids. The action of the acids on the iron was slow, but at the expiration of three months, although the appearance of the cubes of cast iron had not changed, some of the tubes, especially those immersed in acetic acid, had become soft to the extent that a knife-blade would penetrate the cube 0*11 to 0*16 of an inch. The solutions of the acids men- tioned were replaced by a fresh supply in each bottle every month during two years, when changes had taken place in all the cubes, acetic acid having the most powerful decom- posing effect, then hydrochloric and sulphuric acid ; phosphoric acid not showing such a result. The acids had so acted 46 CORROSION AND ITS PREVENTION. upon the metal as to change its nature, without alteration of its bulk or the appearance of its surface. The weight of a cube was after two years' immersion but 54 grains, and the specific gravity only 2-751, instead of 7-858. It was found that the iron had either been dissolved or extracted from the mass by the action of the acid, and in its place was a carbonaceous compound of less specific weight, having very little cohesion, the bulk or dimensions of the cubes being the same as that of the original cast iron. Analysis of the same cube Analysis of cube of grey after two years' cast iron as described, immersion in acetic acid, before immersion. resulting in its becoming a carbonaceous substance. 95 413 79-960 2 900 11-070 0 790 2-590 0 478 6-070 0 179 0-096 0 132 0 059 0 108 0-155 100 000 100-000 These experiments, although made under more severe con- ditions than those which occur in ordinary practice, are most instructive ; for the acids mentioned are often present either in the air, ready to be brought down by rain, or are in the waters of rivers or the sea, and it may be said that either the acids, or the elements from which they may be formed by decomposition, are more or less present in all the waters and situations in which engineering structures have to be erected, therefore, all surface water should then be regarded as acidulated. As under similar conditions conflicting statements have been made as to the corrosion of cast iron in sea water, perhaps the difference may have been due to disturbance of the COKROSION OF CAST IRON AND STEEL. 47 skin on the cast iron or to the different quality of the metal, or to some local cause not noticed, and if observed, not thoroughly investigated. In one place cast iron is pointed out as being uninjured after about half a century's submersion or erection. On the other hand, some cast iron has un- doubtedly had its nature changed, and become another substance, and been made quite soft and somewhat similar to plumbago. In any case it is a material to view with some doubt, because any corrosion cannot be seen as it can be on the surface of wrought iron or steel. It may be perfectly sound, and it may not be. It is not agreeable to be answer- able for so unreliable a material ; and it is one not to be too quickly adopted without full consideration of the consequences of failure. The shrinkage strains in castings affect the homogeneous- ness of the metal, for the particles near and about the surface are able by the softness of the interior metal to get closer to each other in setting than those in the more central portion which become cooled more slowly ; thus when the surface metal has become corroded, the interior iron will be affected much more quickly. Eibs and lugs cast on a cast-iron plate, as on a pile joint flange or bearing plate, have a tendency to curve the plate. If the ribs and lugs be thinner than the plate they will become cooled first, and then, in resisting the shrinkage of the plate, probably make it curve upwards. Where the greater mass of metal is, cooling and shrinkage will be last, therefore the thicknesses of a casting should be regular, and mouldings on columns are better omitted, or they can be attached to the columns as ornaments. The shrinkage of a casting takes place while it is changing from a red to a black heat. Homogeneousness in a casting is neces- sary to ensure uniform resistance to corrosion. The metal in castings cooling first near the surface has a close and equal grain, and is harder than the internal portion, therefore it is not homogeneous throughout the mass, for the outer and inner portions have not the same strength. In very thick castings the interior metal will approach a cellular or honey- 48 CORROSION AND ITS PREVENTION. comb form. Mr. Barlow's experiments on beams showed that the strength of the skins of bars of British cast iron was about two and a half times greater than the interior metal. The importance of preserving from corrosion the outside skin of columns or castings is great, for if it becomes corroded or a change occurs in the character of the metal, such deterioration can only be regarded as very serious, as it is the strongest metal, and even if one side only apparently has so altered, the corrosion and deterioration will soon penetrate and en- velop the outer skin. In thinner castings, which, however, are liable to be more brittle, say those having a thickness not exceeding about f inch, this inferior interior metal is not so evident. Columns which have the most unequal thickness will have the greatest tendency to crack, and when a column is cracked, continued and increased pressure is sure to enlarge the fissures. Corrosion will soon occur internally and in all the cracks. In some experiments at the Eoyal Gun Factories, Wool- wich, it was found that for columns vertical casting was the best system, as the scum of the iron is then at the top, and not on one side, as in horizontally cast pillars. An extra length, equal to about the diameter of the column, can be cut off in the lathe. If the casting is thick, the extra length should be a little more than equal to the diameter of the column. The decided weight of practical opinion and ex- perience is in favour of casting columns vertically under the pressure of a high column of metal in order to ensure their solidity. The greater the solidity, the less the corrosion. Even when a column is cast on end, and a head of metal equal to the external width or diameter of the pillar is cast on, and afterwards cut off in the lathe, any projections or thicker portions of the pile may be hollow or porous. To increase the soundness of castings a plan has been adopted by M. von Riet of fixing a separating chamber above the flask. It is divided into three circular compartments. The molten metal is poured into the largest of these in such a way as to give it a whirling motion, which causes the heavy CORROSION OF CAST IRON AND STEEL. 49 metal to keep to the sides, the light scoria remaining in the centre. A communication to the second compartment is made at one side of the first, and here a further separation is effected in a similar manner, the purified metal escaping finally into the third compartment, in the middle of which is the pouring hole to the flask. On cooling, the first compartment is found to contain large lumps of cinder, in the second the iron left is of a spongy texture. The castings, however, are stated to be very sound and solid. If a girder is cast on its side, the scum of the iron is on one side. On the other hand, when it is cast upright, that is with the bottom flange lowest, the scum is in the top flange, where it affects the strength the least, and can be provided against by a slight increase in the thickness of the top flange. Unequal corrosion of a girder can thus be prevented, whereas when the scale is on one side of the web and one end of the flanges, corrosion, if a girder be exposed, will be unequal, and galvanic action may be set up by nuclei being formed in the iron, and a much greater surface of the cast iron has not the same naturally protective silicious scale. In the process of casting, the molten metal fuses the sand upon the surface of the mould and causes a hard silicious skin to be on the iron, which, if not perforated, broken, or altered, but covered with a really anti-corrosive paint, is a decided protection, and acts as a shield against corrosive influences. If the skin be removed, broken, or injured, the natural protective coat is destroyed, and corrosion or change of condition of the iron may ensue, unless the paint is sufficient to prevent it and is periodically maintained in that condition. What is technically called " scabbing " occurs sometimes in castings, i. e., the sand rises from the under surface of the mould during the time the molten metal is running into it. Metal mixed with sand cannot be as strong as solid metal, therefore there is local weakness and inequality. Great care should be taken to procure uniformity of the nature of the metal throughout a column, or any corrosion will be unequal. As the kind and quality of moulding sands vary, and different E 50 CORROSION AND ITS PREVENTION. systems of mixing and preparing the material are used, and as it is generally considered that the silicious scale is a pre- servative, as undoubtedly it is if kept intact, it is well to remember that in good sand the proportions of silica and alumina are usually about 94 and 5 per cent, respectively, with traces of magnesia and oxide of iron. Sand containing much metallic oxide, especially lime, is not desirable for moulding. It is evident that the resulting scale on the cast iron cannot always be of the same naturo, nor have identical anti-corrosive powers, nor will any paint adhere or otherwise have similar resistance to decaying influences on different material. It is probable that the character of the scale on cast iron may have a not inconsiderable effect in determining the corrosion, and may partly be the cause of in one case cast iron apparently being perfectly sound after a structure has been erected many years, and in another quite soft or seriously deteriorated in strength. Inequalities in the scale will also occur unless the moulding sand be of uniform character, perfectly mixed and thoroughly incorporated. Such defects may be and have been caused by the sand in a mould being rammed too tightly in one place and too loosely in another, and by want of homogeneousness in the sand. They will produce heterogeneousness in the scale, and also in the surface of the iron, and perhaps in the iron. Moulding sand should be equally tough and porous. Mr. Hodgkinson re- corded that some cast-iron girders he tested had a permanent set when loaded with from ^ T to ^ of the breaking weight, and that any weight, however small, injures the elasticity of cast iron, consequently any corrosion or decomposition of the metal may have serious consequences. Eecently in the U.S.A., and, it is said, in some instances in this country, thin sheets of wrought iron have been in- serted in the centre of the mould before casting, It was first done in the case of plates for cooking stoves, and is stated to render them unbreakable by fire. Large iron pipes have also been so cast. The wrought iron perforated sheets sometimes used are No. 27 wire guage, and it is said that CORROSION OF CAST IRON AND STEEL. 51 with such, a plate inserted as described, a ^-inch cast-iron plate is equal in strength to that of a plate 1 inch in thick- ness. The contact of metals of different composition is not, however, to be recommended from the point of view of the prevention of corrosion, as galvanic action maybe generated. Colonel Yolland's report, September 16, 1884, on the fall of a cast-iron railway bridge, declared that "it is perfectly well known that cast iron cannot be relied upon." For the girders of railway bridges it is now practically an obsolete material. Cast iron is being abandoned in the best designs even for arched bridges. Its use in engineering construction is likely to be confined to cylinders, columns and piles statically strained, inferior parts of structures, pipes, &c, and probably will ultimately hardly be adopted for anything having to sustain any considerable or varying strain, or in any position in which a sudden failure would be disastrous or even in- convenient. Defects in the composition of cast iron, air-holes, cold shorts, cracks, cinders, &c, and corrosion are not to be easily detected, and can only now be ascertained by me- chanical tests and chemical analysis. Cast iron, unless very great care is exercised in its composition and in the method of casting, should be regarded as an insidious material that may have the appearance of solidity and durability, and be really a more or less honeycombed and decaying mass, no longer possessing reliable strength, compactness, or durability. The importance and necessity of its being well and carefully manufactured throughout the mass has been recognised by the greater skill and attention given to promote its homo- geneousness and good quality in every respect, but steel and wrought iron cannot but be considered as more reliable materials both as to strength and known durability. It is not here intended to condemn the use of cast iron or any other material, for each has its use, but in most engineering structures great care is desirable in its manufacture, for local defects are more serious in their results than in wrought iron or steel, therefore motives of prudence demand cast iron should only be used for those parts of a structure in statical E 2 52 CORROSION AND ITS PREVENTION. compression, or when subject to very slight other strains, and unless these conditions are complied with sudden failure'may result, the combined effect of overstrain and deterioration from decomposition of the mass consequent upon corrosion. With regard to the corrosion of steel, that of wrought iron and iron generally being examined in Chapter II., steel is a somewhat indefinite substance, for it is known structural iron is steely in some degree. Dr. Siemens suggested that the following definition might be found to answer all re- quirements, namely, that steel was a compound of iron with any other substance which tended to give it superior strength, and stated there are substances, compounds of iron, man- ganese, and silicon sold for steel, which corrode very rapidly. It is the improper selection and use of the material which causes any excessive corrosion, for under proper conditions steel is quite as durable as iron. The structure of steel depends upon its composition and the method of manufacture. It has been shown by means of microscopical examination that corrosion will be variable unless homogeneousness is attained, for if it is not, there will probably be crystals, insoluble and other substances, minute cavities caused by contraction and other matter in such steel, and galvanic action may be set up internally. Examination by the eye, and especially with the aid of a microscope, shows that the fluid-compression process as applied to steel ingots causes homogeneousness in the structure of the metal, hence it is a decided preservative against corrosive influences, but whether the same uniformity and solidity of texture can be or is pro- duced by other means cannot be referred to here ; however, anything that conduces to such an end is to be regarded as having a decided anti-corrosive effect. Unless the composition and processes of manufacture of either iron or steel are declared, its probable uniformity of texture cannot be judged. Dr. Siemens has said that " mild steel was really iron of the best character. It was produced, not like puddled iron, in small quantities to be welded together with the chance of enclosing foreign matter, and CORROSION OF CAST IRON AND STEEL. 53 producing irregular results ; but it was produced in large masses, 10 or 12 tons of fluid substance, and there was every probability that such a material was uniform to the utmost degree. Practice had fully substantiated the fact that there was no material more uniform than very mild steel, but it is not specially adapted for girders, a somewhat harder steel being better adapted for the purpose, and the selection of the best kind of steel for a structure is a matter requiring con- sideration in each case." If the section becomes small, as in continuous girders, it would seem to be advisable to use a milder steel than in ordinary girders, so as to have metal of the very greatest reliability, and that will bear almost half its breaking strain without apparent permanent set, and that will bend rather than break. In cold climates experience has shown that it is advisable to use steel possessing very considerable tenacity and softness, whereas in mild or warm climates almost any kind of approved steel can be employed. The effect of quality on corrosion is a somewhat disputed question. Some consider mild steel for railway bridges should be of great purity, have no foreign matter other than carbon and manganese, that the proportion of carbon and manganese should never exceed a certain limit, the metal have a silky and not a granular fracture, however the appearance of fracture may be altered by the nature and manner of the strain applied, and that apart from questions of greater suitability of the steel for bridge work, corrosion is then likely to be less active. But it is to be remembered Bessemer tire steel, for instance, should contain, to insure soundness consequent upon the process of manufacture, about 1*25 per cent, of manganese. To show the futility of com- paring the corrosion of a piece of steel with that of wrought iron without due consideration of the composition of each metal and the manner in which it is manufactured, careful analysis of a number of experiments conducted by different skilled operators demonstrates that in common steel having the scale on it, and made without much care, corrosion commences from a number of nuclei and not equally, 54 CORROSION AND ITS PREVENTION. whereas in steel of the most approved composition and carefully manufactured so as to be homogeneous, and the scale removed, the corrosion is more uniform and rather less than that of wrought iron of a similar class of metal. If what may "be termed had or inferior steel is compared with good wrought iron, the latter will appear to be better able to resist corrosion. If bad steel and bad wrought iron, or good and good respectively were compared, the steel would most probably be found to be the better able to resist corro- sion, one reason being the absence of defective welding, which is a not unimportant cause of corrosion in wrought iron. It has not yet been proved, although it has been stated that steel in the atmosphere of towns rusts quicker than iron, but effective painting would prevent it, and the design of a structure should, if possible, enable it to be readily accessible for that purpose. It is well known mechanical tests in every respect similarly made on metal chemically identical and from the same bar of steel will give different results. This can hardly be attributed to the heterogeneous chemical composition of the material. It may be to the extreme sensitiveness of steel of great strength to physical change, hence the tendency to use principally the milder steels in engineering con- struction. In fact, it has been stated that without annealing, it is impossible to get even in one bar an absolutely homo- geneous molecular structure, consequently corrosion may be greater at one place than another, and corrosive streaks and pitting may be promoted in conjunction with other causes. It seems to be proved that, owing to strain, injurious mole- cular change may be very local. The results of experiments appear to indicate that annealed steel is much more soluble in sea- water than tempered or hardened steel ; and that when they are in contact, the softer metal is rather rapidly corroded. Some recent experiments on the Netherland State Eailways, extending for about five years, 5000 trains passing over them yearly, tended to show that mild steel rails lost through wear and rust about 25 per cent, more than hard, CORROSION OF CAST IRON AND STEEL. 55 "but that rails of great hardness were considered as more liable to fractures ; the lighter rails, 60 lbs. per yard, ai e specified to have a tensile strength of 33 tons per square inch, and the heavy rails, 80 lbs. per yard, 36*8 tons per square inch. As yet it cannot be said whether soft or hard steel is the more liable to corrosion, but it would seem that soft or mild steel is rather more corrosible than hard, "but the different composition of the metal must always influence the results. Usually, the hardness of steel depends chiefly upon the amount of carbon in it, and the finer the^grain the harder will be the metal and the more carbon in it. Con- siderable inequalities exist among different specimens of steel and iron, and therefore no fixed rate of corrosion can be established. The power of the corrosive influences will also seldom be precisely the same. The reliable and recent practical experiments of Mr. Thos. Andrews, F.E.S., M. Inst. C.E., of the Wortley Iron- works, on simple corrosion, i.e. the test plates not being subject to galvanic action, in sea water taken from Filey Bay, and changed monthly, showed that as regards soft steel plates, the lower the percentage of combined carbon the less the corrosion, the best scrap iron corroding during a period of 110 weeks much less than the steels for the same time. The wrought iron, which contained roughly about double the percentage of phosphorus and manganese, corroded more than the best scrap iron, but less in the total period than any of the steels, with the exception of soft Bessemer. An excess of manganese in any of the metals is liable to produce increased corrosion, because it tends to induce local action from its uneven distribution. Weighing the test plates proved the correctness of the conclusions, and the results of galvanic action confirmed them. Further, limited to the question of corrosion only, it may be surmised that a body possessing a fine crystalline texture like steel, and having a complex chemical composition, would be ultimately more liable to ordinary corrosive disintegration than a fibrous substance such as wrought iron, and that in the construction 56 COREOSION AND ITS PREVENTION. of iron or steel ships, or any structures submerged in sea- water, or any kind of marine work, it is always desirable, in order to reduce internal causes of disintegration, to endeavour to have the bars, plates, rivets, bolts, as near as practicable of a similar material, temper and composition. A certain percentage of carbon makes steel hard, but steel having comparatively little carbon is not so. Hardened steel has a very much finer texture on fracture than before it underwent that process, when it might be coarse. The question of variation of strength in the same chemically composed bar is one that has not yet been satisfactorily solved, although many reasonable inferences have been deduced, and at present it seems to be in the direction of change of structure that the solution of the problem is to be sought. It has been shown in practice that hardness alone does not always in rails necessarily produce great power of resistance to wear, and it may be considered as proved that hard and soft metal in other respects of the same character will not be similarly affected by corrosion. Careful and continued microscopical examination has shown that the crystallisation of steel does not necessarily proceed equally, and that in cooling down it undergoes a molecular change. Any variation in the metal may aid corrosion either by the formation of nuclei for such influences or by the production of galvanic action. The Societe Cockerill, of Seraing, arranged their steel in four classes. Extra mild steel, carbon, 0 • 05 to 0 • 20 per cent., used for boilers, ship and girder plates. Mild steel, carbon, 0-20 to 0-35 per cent., used for railway axles, tires, rails. Hard steel, carbon, 0 • 35 to 0 • 50 per cent., used for rails, special tires, pieces subject to friction. Extra hard steel, carbon, 0-50 to 0-65 per cent, used for delicate springs, files, saws and cutting tools. As aluminium is being tried and used in iron and steel manufacture, it may be well to here state that experiments have shown it is affected by acid and alkaline liquids and also by sulphuric and nitric acids, and therefore not to be particularly desired where considerable corrosive CORROSION OF CAST IRON AND STEEL. 57 influences exist, although on high authority it has been otherwise declared. In some recent experiments of the Alloys Eesearch Committee by Professor Eoberts-Austen, C.B., F.B.S., samples of alloys were kept for some months before being analysed. During this time those which con- tained from 40 to 60 per cent, of aluminium had spontane- ously disintegrated, and had fallen to powder. The powder was not oxidised, but consisted of clean metallic grains, probably resulting from chemical changes which had gradu- ally taken place in the solid alloys. Whether the iron and aluminium were in a state of solution or were chemically combined when molten, there can be little doubt that they are so combined in the metallic powder, as attempts to remelt this powder have proved unsuccessful, which points to the formation of an infusible compound. Solutions are considered to be dissociated chemical compounds. These two metals, iron and aluminium, may have been too hot to unite when in the molten state ; but at a lower temperature long- continued proximity at last effected their chemical union. It will hardly be supposed that this is an isolated case in the study of alloys. M. Le Chatelier, in a recent paper read before the Academic des Sciences, Paris, stated that equal parts of aluminium and copper were fused together in a crucible. The resulting ingot was then placed for 24 hours in a solution of common salt and lead chloride with a view to dissolving out the un combined aluminium. No apparent change could, at the end of this time, be observed in the ingot. It was accordingly removed from the saline bath, washed and dried. At the end of 12 hours the whole mass was found to be reduced to the state of powder from the spontaneous oxidation of the alloy. A similar ingot, not immersed in the solution sodium and lead chlorides, was unchanged at the end of a month. It is understood to have been already found advantageous to paint the bottoms of aluminium-built vessels. In any case, they can hardly be considered incorrodible. The French Government have used, and are using, 58 CORROSION AND ITS PREVENTION. aluminium for torpedo boats. At the Norfolk Navy Yard, U.S.A., two aluminium plates were immersed in sea water 45 days. One was pure aluminium, the other contained 6 per cent, of copper. The pure metal was then only slightly affected, the other was much roughened by corrosion. They were again immersed and left for three months. The pure metal had suffered little, and was only slightly rough on the surface, and had few barnacles upon it, but the other was covered with them, and was more corroded. It is re- ported that the three small aluminium boats used by Mr. Wellman in his 1894 Polar Expedition, soon after being brought back could be easily crumbled in the hand. It would appear sufficient time has not elapsed for any absolute decree to be made as to the resistance of aluminium to corro- sion in sea water. The purity, homogeneousness, and method of manufacture no doubt affect the results. Some consider- able research has resulted in it being here expressed that it would be prudent, in now using it in sea water, to consider it to be, so far as regards corrosion in constructional engineering work, in a more or less tentative condition, and to protect it similarly to iron or steel when submerged in sea water. When cold steel or steel at a low temperature is punched, the metal surrounding the hole is deleteriously affected, but if the metal is afterwards annealed the injurious effect is considerably reduced. As what may be called damaged metal is more quickly corroded than that which is sound, it has a doubly pernicious influence. The density of steel and the molecular condition are affected by treatment. Any kind of mechanical work upon steel has the effect of altering its physical character. Eolling down at a high temperature increases its specific gravity by compression of the molecules into intimate contact. The effect of phosphorus on steel is to make the metal brittle, liable to red shortness, and difficult to roll. Phosphoric steels are known to be untrustworthy when subject to severe sudden strains. They are less ductile and, it would appear, less homogeneous than ordinary steels, and therefore more liable to corrosive influences. 59 CHAPTER V. NOTES ON THE RELATIVE CORROSIBILITY OF STEEL, WROUGHT IRON AND CAST IRON, AND SELECTION OF THE METAL FOR A STRUCTURE. In determining whether to rise steel, -wrought iron, or cast iron in any metallic structure, its relative corrosibility should be considered. The preceding chapters have referred more especially to the processes of corrosion in steel, wrought iron and cast iron, and should be read in deliberating upon the relative corrosibility, which has to be considered under the circumstances in which the material is placed, apart from that of cost, weight, size, strength, reliability and adaptability for the purpose required. It may be affirmed that a material which possesses the necessary elements of resistance is to be preferred to one requiring protective works, but, on the other hand, the cost of a structure made of a certain material may equal, if it does not exceed, the expenditure required by the adoption of a material requiring protective works. This more particularly applies to river and marine structures, which should be able to withstand blows from shipping, barges, trunks of trees and logs, ice and debris aided by the force of flood water. The corrosion of wrought iron has been referred to in Chapter II., and although it rusts somewhat more quickly than cast iron, it has the great advantage that the corrosion can be readily measured and the degree of deterioration ascertained ; nevertheless, when cast iron is used a greater thickness is generally required, and considered from the mere point of view of exposed surface to thickness, it has an 60 CORROSION AND ITS PREVENTION. advantage, as comparatively little of the metal may be decomposed if protected by paint; in fact, it has been preferred to wrought iron simply because of this, the idea being that by increasing the thickness of the cast iron the molecular charge would be reduced. This may or may not be so. It has not been proved, but may logically be deduced ; however, it is well to remember that other matters require to be considered, for increased thickness is not necessarily a proportional addition of strength. In the engineering structures herein more particularly considered, it is not advisable to use thicker castings than about 1\ inches, nor thinner than about f to 1 inch, the latter more particularly on account of the scale, the former because of the difficulty in obtaining homogeneousness, equal cooling, and accord in the order and direction of crystallisation ; for it has been found wherever the order of crystallisation is disturbed there will be weakness and unequal corrosion, hence the importance also of as few abrupt bends, sharp angles and variations of thickness as practicable. Solid forged wrought-iron piles are better able to with- stand sudden blows than cast iron, but are considerably dearer. For tropical climates, general opinion seems decidedly to be in favour of cast-iron piles for submerged or the alternately wet and dry portions of a structure, as it has been found they appear to be in good condition, whereas the wrought-iron bracing, bolts and nuts occasionally submerged soon become deteriorated ; however, the surfaces of the latter are greater as compared with the sectional areas, and they are more exposed generally. For the piers of lofty viaducts it is open to question whether cast iron is so well adapted as wrought iron, for other reasons than the mere relative corro- sibility of the metals, as wrought-iron columns may be of sufficient diameter to allow of internal inspection, but, if they were of cast iron it may not be so, and the severe lateral strains on such a pier require to be met by a very complete system of bracing. In considering the deterioration of the metal, it is always well to be sure galvanic action is not the RELATIVE C0RR0SIB1LITT OF STEEL, ETC. 61 chief cause of any accelerated corrosion, and that it does not proceed from relative simple corrosion, if that expression can be used in regard to such a complex matter, for the electro- positive metal will become oxidised more rapidly than the electro-negative metal. It is very important to search for any interfering conditions in endeavouring to ascertain the relative corrosion of steel, wrought iron, or cast iron in any structure. To draw general conclusions from an isolated case is unadvisable, and has destroyed the apparent value of experiments and examinations conducted without regard to interfering conditions ; however, if any experiments are conducted without such interfering conditions they are entitled to be considered as reliable, and are applicable to metal placed under the same circumstances in structures, but if otherwise, they may not be so. Those fully competent to judge have affirmed that by a judicious selection by an expert of steel and iron test pieces according to their chemical com- position, so as to take every precaution that they will show what is desired to be demonstrated, the steels might be declared by experiment to resist corrosion better than wrought iron or vice versa, almost at the will of those conducting the experiments. The following question has to be answered in determining, solely as regards corrosibility, whether wrought or cast iron should be used for any purpose. Is it better to rely upon a material, such as wrought iron, in which the amount of corrosion can be seen, or upon a material, such as cast iron, in which the degree of decomposition cannot be measured except by chemical analysis and mechanical tests, although experience appears to demonstrate that such deterioration of strength and decomposition of the mass are both slow and superficial ? No concise answer can be given to such a question, the various circumstances and conditions require to be carefully considered in all their different aspects. In submerged work, when the metal has simply to act as a column sustaining an insistent weight, and is not subject to variations of load or blows, cast-iron hollow columns are to 62 COREOSION AND ITS PREVENTION. "be preferred to wrought-iron solid columns, their interior "being properly preserved from corrosion. Many cast-iron columns are still apparently unaffected although they were erected from forty to fifty years ago in sea, estuary and fresh water ; however, there is no absolute proof by chemical analysis, or the usual tests for strength, that a portion of the metal has not "become decomposed, or that decomposition is not proceeding, and the strength reduced and gradually diminishing, and a prudent course is to allow for such decomposition by increased thickness, within certain limits, and additional to any thickness required to meet the strains. Solid wrought-iron piles are not often adopted for the columns of sea or river piers, and expert opinion indicates that for such structures cast-iron piles are to be preferred ; in fact, instances have occurred in which it has been con- sidered advisable to substitute cast-iron for wrought-iron piles erected in sea-water in consequence of the corrosion of the wrougbt-iron piles having been very great, amounting, to as much as 2 inches in a column originally of 5 inches diameter, in from ten to fifteen years, the piles being exposed to the air every tide, and the coast open.* This, however, it is advisable to consider as an exceptional case, and it is unfortunate that full particulars are not to be obtained of the reason of this severe corrosion, such as analysis of the iron, sea-water, whether galvanic action had been caused by juxtaposition of other metals, whether the scale was or was not removed, and whether there were any exceptional circumstances tending to promote corrosion or not, such as the constant discharge of sewage into the sea near the piles. The late Sir J. W. Bazalgette once frankly remarked he had built a wrought-iron aqueduct for conveying sewage, and found it was not a wise selection, and had since adopted cast iron for similar structures, because, at any rate, a thicker material was thus provided at a less cost. The question of durability in cast iron Mr. Mallet defined as resolving itself, * See Chapter VI., Notes on Rapid Corrosion, with Examples. RELATIVE CORROSIBILTTY OF STEEL, ETC. 63 as regarded cost, into one of so much increased thickness at the outset as amply to allow for it in reference to a pre- determined period of endurance. For columns that are to be in the earth or submerged, and that consequently cannot be seen after erection, because of the greater thickness and somewhat slower rate of apparent corrosion, the statement seems to be justified that cast iron is a better material to use than wrought iron or steel • but if proper care can be taken in its preservation, wrought iron or steel is a more reliable metal, therefore the chief point for consideration is whether the wrought iron or steel can be made almost permanently secure by care in maintenance. The best material may become dangerous when neglected and allowed to waste away ; but in works which permit of inspection and are carefully protected, wrought iron or steel are generally much to be preferred to cast iron, owing to their reliability and the extent of the corrosion being visible, and this even if the corrosion be somewhat quicker in action bearing in mind that decomposition in cast iron may not be merely superficial or apparent, or capable of being measured until examination, analyses and test are made of a piece or pieces taken from the structure to prove that corrosion is merely superficial. The durability of wrought iron or steel, so far as regards freedom from corrosion, may be said to be almost governed by the care taken in its preservation. Why is this the case more than with cast iron ? Apart from relative corrosibility, the form in which the material is used is the governing j condition. In cast iron, not only are the exposed surface ■ areas very much less than is the case with a mass of plates, j bars, angle, T, and other small sections of iron, rivet, and ■ rivet-heads, but the silicious skin received in the sand mould ■ can in great measure, by care, be preserved throughout, and I the surface is also unbroken by rivet holes, projections and |joints, around or in which moisture can collect and cause comparatively hidden oxidation of the surfaces or edges of fjthe plates and other parts of a structure ; therefore, although 64 CORROSION AND ITS PREVENTION. in the mass of metal there is not any change other than can he seen, there may be rapid corrosion between the plates, or underneath the angle T, channel irons, bars, cover-plates, and around the rivet-heads. Briefly, the relative exposed surface to the sectional area is much greater in a wrought- iron than in a cast-iron column or girder, consequently corrosion is increased, unless the metal is protected. With regard to the relative corrosibility of wrought iron and steel, has experience shown that steel properly protected corrodes quicker than wrought iron under similar conditions and circumstances ? So far as regards the safe employment of steel, it may be considered no fear need exist in using it. Occasionally a few alarming reports are disseminated to the detriment of steel, but upon investigation it is generally found either the circumstances were such that wrought iron would have been similarly affected, or the experiments were made under interfering conditions, such as galvanic action, which unfortunately render any conclusions that might be drawn from their results either right by good fortune, or j misleading from the manner in which the tests were con- ducted. The original bloom or scale on test pieces has been proved to vitiate experiments. Several years ago the steel \ plates of a small vessel of very shallow draught were found to become rapidly pitted in the waters of the Irrawaddy, and a kind of frightful example was produced to show thei inability of steel to resist corrosion ; but it was shown that all iron vessels rapidly corroded in that river owing to peculiarities in the water. Admiralty experiments demonstrate that there is practi- cally no difference in the rate of corrosion between iron and the mild steels used by the authorities, provided, in the case of] steel plates, they are divested of oxide on the surface. When under the same conditions as to corrosive influences and painting, if the scale was thoroughly removed, iron and steel corroded at about the same average rate, but steel more* uniformly than iron. The scale being much harder tci remove by chipping or scraping, after many trials, the] EELATIVE COEEOSIBILITY OF STEEL, ETC. 65 immersion of plates and bars in dilute acid baths similar to those used in zincing iron proved to be the best method of clearing away the scale. It may be said all experts agree whether steel is used for plates, even when they are as thin as ^ inch, boilers, or general shipbuilding or other purposes, the corrosion of the two metals is considered to be not very different, and any advantage is on the side of steel. A few qualify this declaration by stating that where steel is entirely immersed in water no more rapid deterioration occurs than in iron, but that in ships, between the light and load line, or wind and water, steel requires a little closer attention to keep the plates coated with anti-corrosive composition, in order to prevent its rather more rapid corrosion. In the U.S.A. some engineers consider steel to be much more durable, and less liable to corrode than wrought or cast iron. In this country perhaps not so decided an opinion would be expressed, although it would be in the same direction. Some experiments made in France have shown that steel produced from pure iron, having hardly any manganese, carbon, phosphorus, or silicon, was less liable to corrosion than ordinary wrought or cast iron in the proportion of 0'4 to 1*4, being much in favour of a pure metal, such as very mild steel. Mr. Mangold, chemist to the Cologne- Miisen Mining Company, recently made * some experiments to ascertain the comparative corrosibility of iron and steel sheets, 1 millimetre = 0*039 inch in thickness, with the following constituents : — When pickled in strong aqua regia (3 parts hydrochloric acid to 1 of nitric acid, specific gravity, 1 • 4) the steel was energetically and uniformly attacked. With the iron the action began slowly, becoming more rapid with the duration Carbon .. Manganese Silicon .. Iron. Per cent, 0-16 0-24 0-72 Steel. Per cent. 0-06 0-25 0-00 * ' Stahl und Eisen,' vol. xy. v 66 CORROSION AND ITS PREVENTION. of the experiment, the surface produced was rough, showing signs of irregular welding. When immersed in spring water, the loss by two days' exposure was 0-04 per cent, in the iron, and 0-078 per cent, in the steel. The loss by ten days' exposure, the water being changed at intervals, was 0*09 per cent, in the iron, and 0*240 per cent, in the steel. When heated in a muffle at a bright red heat without access of air, the loss in 4£ hours was 18-32 per cent, in the iron, and 32 • 20 per cent, in the steel ; and the loss in two hours more, 14-62 and 28-07 per cent, respectively. When exposed to an oxidising flame for three days both kinds were completely destroyed. When treated with a 1 per cent, salt solution, the loss in 24 hours in the iron was 0-037 per cent., in the steel, 0-128 per cent. The loss in three days was 0*085 and 0-220 per cent, respectively. When exposed to carbonic acid, the waste gases of a spathic ore-calcining kiln, and air, the loss in two days was 0 * 84 per cent, in the iron, and 0 • 94 per cent, in the steel, at four days, 1 * 60 and 1 * 40 respectively ; and at six days, 1*95 and 1*90. When exposed to air and water vapour, the loss in twelve days was 0*29 per cent, in the iron, and 0 * 68 per cent, in the steel. When galvanised, wrought-iron sheets are more durable than those of mild steel. The latter can be more cheaply produced, and from the more uniform character of the metal have a better surface, but the zinc coating is thinner, and when it becomes unsound the plate is more rapidly corroded than is the case with sheet iron. Provided the ordinary precautions are taken, similar to those adopted with wrought iron, experience shows that in considering whether steel or iron shall be used in any structure, their corrosibility may be taken as practically the same, but much depends upon the quality and texture of the metals, and therefore the strength, uniformity, and other qualities of steel are unbalanced advantages in its favour; however, the value of a protection from corrosion is greater in the case of steel than iron, because of the less section of metal required. 67 CHAPTEE VI. NOTES ON RAPID CORROSION, WITH EXAMPLES. With regard to the conditions that cause rapid corrosion, iron in a very finely divided state, when exposed to the atmosphere, may oxidise so quickly as to become ignited. It is recorded that a cast-iron shell, which it was known had laid under water for about 200 years, when brought to land was found to be honeycombed, and some of the metal was in so fine a state of division that it gradually steamed, and became red hot ; hence there may be danger in rapid corro- sion under such circumstances, for decay may be said to be a process of more or less continual combustion. This is an instance showing that corrosive agents are especially active if they have not previously come in contact with a substance for which they have an affinity. The smoke and vapour issuing from locomotives has a marked corrosive effect on girders under which trains pass, therefore, all hollows and depressions in which they can accumulate and hover about it is well to avoid, and to screen with a plate such portions of any girders. In sheltered cornices or nooks, or in ornamental work, or where pro- jections occur, or in any places which will collect soot and dust and retain moisture, corrosive influences will be very active, and considerably more so than if the surfaces were smooth and even. Exposure to acid or alkaline vapour or liquid, to damp, occasional submersion in sea or fresh water, sewage, or the waste liquids from chemical works, consider- able and varying heat and cold, constant or alternately occurring heavy rains with drying winds, hot sun, frost or v 2 68 CORROSION AND ITS PREVENTION. snow, produces rapid corrosion, and also severely tries any paint used as a protective coating. Heating apparatus in buildings which causes damp places in the walls, or a state of humidity probably ending in vapour, has to be regarded as an active corrosive influence ; and it is advisable to con- sider it when girders, bearers, columns, &c, are subject to such heat, perhaps for a few hours during the day, and to the ordinary temperature of the internal air at other times. Corrosion in fish-plates and metallic sleepers is generally most where the fish-bolts and fastenings occur. Inspection seems to indicate that the corrosion was much accelerated by loose bolts and fastenings, and it is difScult to determine whether the decay is more due to mechanical action and friction, or to simple corrosion, but there is no doubt such mechanical action greatly accelerates corrosion, even if it does not induce it. Mr. W. H. Cole, Assoc. M. Inst. C.E., states* that on the Sindh Sagar district of the North- Western Eailway of India nearly the whole line, some 300 miles, was laid in 1886 with steel sleepers in sand, with a stone or brick ballast topping, the soil of sand and clay being im- pregnated with saline matter, but the air is generally very dry. The portion of the line that runs westward between the Salt range and the right bank of the Jhelum Eiver towards the Indus is, however, for months exposed to inun- dation and is saturated by drainage from the hills. Sleepers weighing 148 lbs. each in 1886, averaged only 87 lbs. in 1890, or a loss of 61 lbs. in four years, or 41 per cent. Wooden sleepers were then substituted for steel sleepers. Guided by experiments with steel sleepers buried in sand, on the East Coast Eailway, the authorities decided not to use them within ten miles of the sea. It has, therefore, been considered that steel sleepers should not be adopted in brackish soil, especially if it be moist. A steel sleeper, being so thin, soon fails when attacked by rust. A cast-iron pot sleeper, being thick and apparently less liable to rust, although decomposition of the metal may * * Minutes of Proceedings,' Inst. C.E., vol. cxvii. NOTES ON EAPID CORROSION, WITH EXAMPLES. 69 be proceeding, will last longer in brackish, soil. It may be mentioned incidentally, as it may cause unequal corrosion, that trough sleepers, except when the permanent way has become consolidated in clean gravel or sand ballast, are difficult to pack on any embankment of soft or yielding earth, and their use is particularly undesirable at or near any points and crossings or junctions. While steel sleepers become soon rusted in some situations, it should not be for- gotten that the white ant will quickly destroy wooden sleepers in certain districts. From the effects noticed, some sewage water has been considered as corroding iron almost as fast as diluted vitriol. Cast-iron sewer penstocks, and the valve gearing, have been found in a few years to be rusted away through the chemical action of the sewer gas. A serious corrosive condition is always produced when raw sewage, or any sewage, or water made impure by human use, ammoniacal liquor, or chemical works, is discharged into river or other waters. Land fre- quently irrigated is in a condition to produce active corrosive influences, for it may be " Sated with exhalations rank and fell, The spoil of dunghills, and the putrid thaw Of nature." It is generally soils that are the most permeable, and easily dried or warmed, that afford examples of successful irriga- tion, hence they are caused to be decidedly corrosive in their action. The rainfall in an open district, as compared with a forest country, is much less, varying in different localities ; added to which it has been found that evaporation under trees is only about one-third of that in the open ground, and that a cover- ing of leaves will still further increase the difference. Metal- work in such localities will be subject to severe corrosive influences, in addition to any chemical action the rain or moisture may possess by extraction or impregnation with the juices of the trees and leaves ; such water generally having 70 CORROSION AND ITS PREVENTION. a deleterious action on paint. Metallic structures erected upon " Swampy fens, Where putrefaction into life ferments, And breathes destructive myriads ; or in woods, Impenetrable shades, recesses foul, In vapours rank and blue corruption wrapt," and upon marshes or bogs, will be liable to rapid rusting. It is therefore well to consider whether it is advisable to use some other and more suitable material than steel or iron. Streams flowing from forest or wooded lands, or peaty soils, or any land having much decomposed matter upon or in it, generally have considerable sediment in them, and promote corrosion. By observing whether the water causes vigorous vegetable growth, and abounds with life, fish, or mollusca, indication is obtained that it will be decidedly corrosive. Salt-water marshes, lagoons, and estuaries are also to be classed as actively corrosive. It is well to here remember that carbonic acid is thrown off in great abundance by decaying plants in jungles, and the wind not being able to reach them, it therefore settles and destroys animal life, as in the Valley of Death in Java, and in some of the jungles of Hindustan, &c. Grass, wood, and the leaves of plants, radiate heat very freely ; rough woolly leaves radiate heat much more freely than the hard smooth polished leaves. By radiating heat freely, they condense any vapour which touches them into dew, and so a damp corrosive atmosphere may be formed in their immediate proximity. The salts in sea-water are very hygroscopic, i. e. readily absorb and retain moisture, and have an energetic action on metals. Dust impregnated with sea-water may be blown about and appear to be quite dry, but when the air is charged with moisture it will become damp, and act as an active corro- sive agent. Its effect, however, on macadamised roads is to bind and solidify them, and on stone pitching it tends to make them slippery. The employment of salt to melt snow on roadways, &c, will cause the resulting water to be actively NOTES ON EAPID CORROSION, WITH EXAMPLES. 71 corrosive ; and if it reaches iron pipes it will have a dele- terious effect upon them. Cases have occurred in the same water, and under ap- parently the same conditions, in which cast and wrought iron have very quickly corroded, there being some peculiar in- fluence, circumstance, or character of the metal or water which caused such a result. A perusal of this book may indicate some causes of rapid or exceptional corrosion, or enable them to be discovered. It may be said all materials generally used in engineering works decay more rapidly when exposed to tidal action, or from a few feet above the ground, than they do when con- stantly submerged in ordinary sea or fresh water, or buried in earth in which foundations are usually placed. Anything which produces chemical action aids corrosion, as do fresh supplies of water or moisture, or the mixing of well-aerated fresh with salt water. In the latter case, corrosion in either cast or wrought iron or steel is very rapid. Mr. Beardmore mentioned an instance of this in the case of a sea-lock, in which soft water was " locked " down direct into very salt sea-water. When the lock at the head of a ship canal con- taining very soft water was pumped dry for repairs, after thirty-five years' existence, all the cast and wrought iron portions of the wooden gates, &c, were found to be equally destroyed, even the spikes of the platform planking had perished, though the timber was perfectly sound. Cast or wrought iron similarly exposed, and for the same length of time, in the Thames, he said, would be scarcely affected ; and he attributed the intense action of the water upon the iron to the mixture of fresh and salt water. This corrosive action would be intensified as the saltness of the water increased. Metal placed in a tank in which water is not changed will not corrode so quickly as if the water was constantly being emptied and renewed. The constant dripping of water on iron has also an active corrosive effect. Mr. Andrews's, F.K.S.,* * ' Minutes of Proceedings,' Inst. C.E., vol. cxviii. 72 CORROSION AND ITS PREVENTION. experiments have recently caused him to ohserve that, " Cor- rosion of metals is liable to occur in tidal streams, or under circumstances where the different parts of metallic structures, vessels, &c, may he exposed to the action of waters of dis- similar salinity. In tidal streams this state is brought about by the gradual rise and inward flow of salt water, and the outward flow of the fresh water. Hence, the upper and the lower portions of a metal structure or vessel, although composed throughout of the same metal, are exposed to electrolytic disintegration from the galvanic action set up by solutions of different salinity on the metal. Moreover, there are indications that magnetic influence tends to increase the corrosion of steel." Mr. Edwin Clark stated that chemical action was in- variably accompanied by electric action. It was, therefore, probable that differences in the electrical conditions, or rather the different electrical capacities of large and small masses, might explain any difference in their durability. He tested it by two pieces of plate 4 inches square and J inch in thick- ness. One was attached to one of the tubes of the Britannia bridge by rivets, and was in direct metallic communication and formed part and parcel of the tube, but the other plate was separated from the tube by a thick glass insulating rod. The difference in the oxidation of the two plates was evident, as, in the course of eighteen months, the plate in communi- cation with the tube was scarcely affected by rust, whereas the plate insulated by the glass arm, placed between it and the plate, was decaying rapidly. It was evident the more rapid decay of the small insulated plate was due to this insulation. Iron or steel in tubular railway bridges, and in most railway bridges in a lesser degree, depending upon circum- stances, girders in tunnels or underground railways, and- similar structures exposed to the smoke, steam, and heated gases of locomotives, are especially liable to become rapidly corroded unless frequently tended under skilled direction. Although the maximum quantity of carbonic acid found in the NOTES ON EAPID CORROSION, WITH EXAMPLES. 73 Metropolitan Kail way was under 13 parts, or not so ranch as in close buildings, in railway tunnels the extent of acceleration of the oxidising influences is not to be measured by any excess of the quantity of carbonic acid only, as other agents of corrosion are present, such as aqueous vapour, sulphurous acid (S0 2 ), which is produced whenever sulphur is burned in the air or in oxygen, and is heavier than air, and experi- ments have shown that it quickly changes into sulphuric acid (S0 3 ) in the presence of moisture and iron, and that it is a powerful corrosive agent, and one of the worst met with in railway tunnels or bridges; for the ordinary trade sulphuric acid, or " oil of vitriol," which consists of one atom of the anhydrous acid combined with one atom of water, has a strong affinity for water, dissolves almost all the metals, and is the most powerful acid known, being, perhaps, the most important of all chemical products, and very largely used for manufacturing purposes. It is open to much question whether the system of wrought iron or steel girders, with brick arches between them, is a preferable one to adopt in the construction of under- ground railways, because of the actively corrosive condition of such subterranean ways when worked by locomotives burning coal. Smoke and vapour will cling to and hover about the roof of a tunnel and the bottom flange of the girders, and any tie-rods are much exposed to corrosive influences. Sir B. Baker, in his paper "The Metropolitan and Metropolitan District Bailways," * remarked, " In places wrought-iron girders were used, but though more trustworthy than cast iron whilst new, they are exposed to greater risk from hidden oxidation. Experience has shown the trouble and cost of maintaining ironwork exposed to the atmosphere of an underground railway to be such as would justify a considerable increase in the first cost, by substituting brick- work and deep cuttings for ironwork and shallower con- struction. Where the depth was sufficient for an arch, brick covered ways were adopted in preference to iron-girder * ' Minutes of Proceedings,' Inst. O.E., vol. Ixxxi. 74 COKROSION AND ITS PREVENTION. constructions, on account of the smaller cost and increased durability." Sir John Fowler confirmed Sir B. Baker's views, and further said, " That with covered ways, and, indeed, all structures in London, it was wise to use brickwork when possible, and not ironwork, and even to incur some additional expense by so doing. Oxidation with iron and steel structures was always serious enough in London, but when hidden it might be dangerous." It will be noticed that no opinion is expressed, or experience recorded as to the relative corrosion of cast and wrought iron, but only that wrought iron is " exposed to greater risk from hidden oxidation." It may be written because of the severe corrosive in- fluences, general experience with ironwork in underground railways, in which ordinary and not electrical locomotives work, and in tunnels and similar subterranean ways, indi- cates that it is better to use brickwork in Portland cement, as being comparatively unaffected by the decomposing agencies which so seriously affect metallic structures. Wherever the atmosphere is of an exceptionally corrosive character, as is the case in some of the manufacturing districts, brickwork, and certain kinds of stone, not all, by any means, are to be preferred to iron or steel. As the air of London is not nearly so corrosive as some manufacturing cities and towns, such as Manchester, a stronger reason would prevail in any such localities for adopting some other more durable system of construction than that embracing iron or steel girders. Corrosion will always be considerable in tunnels until smoke, vapour, steam, and gases, that vitiate the atmosphere or cause it to be actively corrosive, are prevented from entering. Some will always be present, even if sufficient power can be obtained by other approved means than steam. Among the chief causes of the disturbances of the air in tunnels, are the different atmospheric pressures, the temperature, and the winds. With very few exceptions, in long tunnels the atmosphere is saturated with moisture. The aqueous vapour from the locomotives surcharges the NOTES ON EAPID CORROSION, WITH EXAMPLES. 75 atmosphere and condensation is rapid. The aqueous vapour has a tendency to take up a considerable portion of carbonic acid and sulphurous gas, and bring it down when the vapour assumes the form of water. In tunnels the corrosion of rails generally increases as the ventilation becomes less and the length of the tunnel greater. The loss of weight by corrosion some consider as twice as much as that which corresponds to the wear of the upper surface of the rail ; but it is to be remembered the number, tonnage, and speed of the trains, the character of the air and moisture, and the construction of the permanent way, and any particular circumstances will much influence results, and that no rule can be deduced. It is generally found that rails in a wet place in a tunnel wear quicker than when they are in a dry position. In the St. Gothard Tunnel the water that percolates, or is formed, and trickles down the walls, is found to contain sulphuretted hydrogen, sulphur dioxide, ammonia, and carbon dioxide, besides other products from the smoke of the locomotives. In the middle of the tunnel the tempera- ture is about 73° F. The corrosive and decaying influences are therefore bad, and for the telegraph cables it was found necessary to have an exceptionally strong protective coating. A St. Gothard Eailway official report has declared that " the life of rails in long and ill- ventilated tunnels is scarcely one-third of their duration in the open air." The weight of the rails has, with a view to mitigate this effect, been in- creased from 74*6 lbs. per yard to 88*7 lbs. per yard. In a tunnel in France, 984 yards in length, which receives a little water by percolation, the rails are on beech sleepers, which rest upon gravel ballast. 230,000 trains, at a speed of about 19 miles per hour, had passed over them when they were withdrawn after 11^ years of service. The double-headed steel rails were 78 lbs. per yard, and 8*74 yards in length ; 10 sleepers to each rail. The original weight of a rail length was 682 lbs. After use it was reduced to 518 lbs., the loss being 164 lbs., or 18*76 lbs. per yard. The most 76 CORROSION AND ITS PREVENTION. important influence was the rolling of the wheel on the top table, combined with the rust which formed during the time elapsing between the last train in the evening and the first in the morning. Abrasion and corrosion combine to reduce and wear away the area of the head of a rail, but the loss on the other por- tions of a rail is due almost entirely to corrosion. It has been found by careful experiment that by well covering wooden sleepers with clean ballast, the precipitation of acid gases to the iron rail flange is greatly prevented, and the passage of any acid water considerably hindered. At the eighth meet- ing of the German Railway Engineers' Association, held at Stuttgart, 1878, twenty-seven authorities were in favour of keeping the sleepers covered, and only seven to the contrary. It may be well to here note the recommendation of an Italian railway commission on permanent way, appointed in 1882, on ballast and sleepers. It was decided that though uncovered sleepers may be preferable in northern climates, it is necessary in Italy to completely embed them in ballast. It was considered by covering the sleepers, instead of the ballast being laid level with their upper surface, the timber was protected from the direct rays of the sun, sudden changes of temperature, moisture, dryness, and frost. Anything which prevents decay of the sleeper, or corrosive influences reaching it, will lessen corrosion of the rails. Mr. H. Footner, M. Inst. C.E., of the London and North- Western Eailway, examined * some steel rails of similar section in different districts, and, on the whole, the results seem to show that the character of the atmosphere much influences the life, and that steel rails laid in an agricultural district will not corrode so quickly as when they are laid in a railway through a manufacturing town and neighbourhood. In a district where there are many chemical or salt works, the wear and corrosion are very much more. Mr. Footner states, judging from an examination of some rusted rails, that * ' Minutes of Proceedings,' Inst. O.E., vol. lxxxiv. NOTES ON RAPID CORROSION, WITH EXAMPLES. 77 the rate of corrosion on the hright surface in contact with the wheel flanges has been approximately four or five times as rapid as upon the rest of the rail. The area of the head or bright surface in contact with the wheel flange is to the rest of the skin of the rail as one is to six approximately. The scale forming on the body of the rail affords a lodgment for dirt, and the two become a natural coating or protection, frequently shaken off, and as frequently renewed, but re- tarding oxidation, while on the other hand the passing wheel keeps the running surface clean and in a condition most favourable for oxidation. He concludes by saying that the relative amount of loss by oxidation depends upon the brightness of the surface, and the absolute amount upon the purity of the atmosphere. It also appeared their abrasion by an unbreaked wheel is small compared with the amount of loss by corrosion when the atmosphere is impure, and that all bright surfaces exposed to moisture similarly suffer, such as railway tires, &c, and lose by corrosion an amount hitherto scarcely suspected. To wash the rails in the Hauenstein Tunnel, the Swiss Central Eailway officials used hot water issuing at a high pressure, for water at a low pressure was found to diminish the adhesion. Slipping of wheels, which was very frequent, consequently now very rarely occurs, and the wear of the rails and tires is found to be much diminished and sand is seldom required. The effect of this intermittent hot water flushing at a high pressure on the metal has not yet been recorded, and it is not easy to foretell, for although the rail may be so cleansed as to free it from active corrosive films, the water may find its way through the pores to the metal, and as a rule, water of the same character has a greater corrosive influence when hot than when cold. Comparative experiments alone can determine the relative rates of corrosion, for although the wear of rails and tires is no doubt reduced very considerably by a free use of water upon them, and the tractive force required lessened, still dampness increases corrosion. 78 CORROSION AND ITS PREVENTION. From a summary of reports and opinions, founded on results, the corrosion of metal sleepers on European railways does not appear to be a source of anxiety. Galvanising, oxidation by steam, a tar or other preservative coating, being seldom considered necessary, except when the sleepers are stacked for renewals on the side of the line, or are to be used in damp tunnels or cuttings, or in a locality where special corrosive influences exist. Then they are either coated with a tar preparation, or paint, as also if they have to be shipped. In northern climates, in a well-drained formation and permanent way under traffic, it is considered corrosion in metal sleepers need not generally be feared as being serious, but if they are stacked and not in use it will be. In fact, experience seems to indicate if a railway is used, corrosion will be inappreciable in northern climates, but if there is little traffic over such sleepers it will soon become important. In tropical climates the corrosion is much increased. Steel in sleepers has been found in some places to rust faster than iron, particularly when it is exposed to moisture. If metallic sleepers work through the ballast to a clay formation, in wet weather, oozing of mud will occur, and the whole road-bed may become mud when wet, and when dry will be dusty, the soil will crack and shrink, and corrosive conditions will be active. The nature of the ballast, among other things, affects the rate of corrosion ; thus, fang-bolts buried in clinker ballast are found to corrode rapidly, and it is the same with any iron that comes in contact with it. It is to be remembered all metallic sleepers, and necessarily their fastenings, which are laid as it were upon the surface of the ballast, and are so packed, are exposed to atmospheric, and therefore corrosive influences more than wooden sleepers, whether their tops are covered with ballast or not. In temperate climates, wrought iron or steel sleepers may not seriously corrode during their serviceable life in the permanent way, but in hot, damp, or tropical countries, thin NOTES ON RAPID CORROSION, WITH EXAMPLES. 79 wrought iron and steel appear to corrode very rapidly as compared with cast iron, although there may be a change in the composition of the latter metal. Where timber is quickly attacked by the white ant, some kind of metallic sleeper, notwithstanding the defects inherent in them, may have to be adopted, unless the wooden sleepers are very carefully and effectively protected from their attacks. The advantages of metallic sleepers, apart altogether from the question of corrosion, have yet to be demonstrated, and, except under peculiar circumstances, their present employment hardly appears to be justified by general results, although here and there they have been approved and used on considerable mileages ; however, much depends on their form, quality of the metal, nature of the ballast, traffic, form of fastenings, and other circumstances. So far as simple corrosion is concerned, from a perusal of reports from all parts of the world, in northern or temperate climates the corrosion, with ordinary precautions, need not be feared when they are in use in the permanent way. The best ballast for them appears to be either sand or gravel, and they should be well covered with it, but sand is not desirable ballast, because it does not afford a rigid bed, it is liable to erosion, and to be blown away, and another objection to it is that i^nsinuates itself into the bearings and machinery, and causes frequent repairs, and a marked deterioration of the rolling stopk and metallic parts with which it may come in contact. Rolled steel plate sleepers, § of an inch in thickness at the centre and £ inch at the edges, on the Indian State Railways, it has been stated, are not found to rust much, except when the ballast is impregnated with saltpetre and known actively corrosive substances. The sleepers are dipped, at a temperature of 300° Fahr., in a boiling solution con- sisting of 3 parts coal tar and 1 part tar oil, which forms a good hard coating, the solution being kept boiling by the passage of the hot plates through it. Mr. H. F. Bamber stated that it is very essential the sleeper should not be so hot at the time of dipping as to cause the solution to give 80 CORROSION AND ITS PREVENTION. off thick suffocating fumes, for the coating then formed flies off like coal dust, and is useless for preserving the metal sleeper from rust. After the execution under his superin- tendence of large quantities of steel sleepers, he found that they should in no case be dipped in tar at such a temperature as to give rise to thick yellow fumes. The lengths of metal sleepers tried in France, when they have been properly coated with coal-tar, it has been found, do not oxidise to any appreciable extent. Irou longitudinal sleepers laid in 1864 on the Brunswick Kail ways, it is said, were found to be but slightly affected by rust 18 years after ; and those on the Bergisch-Markisch line, in badly-drained ballast, after being eight years in use, were found not to be more affected by rust than the rails. Anything from 20 to 50 years has been estimated as the life of a metal sleeper, but the nature of the traffic, quality of the metal, design, thickness, and strength of the sleeper, nature of the ballast, bearing area upon it, climate, absence of special corrosive influences, and care in maintenance, cause any rules to be untrustworthy, except all the conditions are exactly similar. An approved anti-corrosive coating for the metal sleepers, used on the Algerian Bailways, is to temper them with coal-tar. It has been used with success for some 20 years. A disadvantageous feature connected with metallic sleepers is, that the threads of the bolts used for attaching the rails to metal sleepers become rusted, and there is diffi- culty in keeping them tight. However, " after many years' trial, the Netherlands State Bail way decided to keep the screw bolts as a means of attachment between the rail and sleeper." In 1865, the rails were fastened by four bolts 0 *42 inch diameter, to each sleeper. In 1883, for the first time, it was necessary to renew 2000 of these 40,000 bolts, and the remainder were still in the permanent way in 1885. No vegetation should be allowed to grow upon any deposit that may have formed on any metal. When iron or steel arrives at a port in the Tropics it should be at once taken to the site, as it is found oxidation is much more rapid NOTES ON EAPID CORROSION, WITH EXAMPLES. 81 near the coast than inland, unless there are localities in the interior of the country exceptionally exposed to corrosive influences. In tropical climates, iron quickly decays if not thoroughly protected, and, if buried in the earth, the rate of corrosion will vary according as the soil contains active corrosive agents or not. The great heat of the sun, warm damp atmosphere, and increased vegetable life, generate acids which quickly corrode iron. In consequence of the small dimensions, sectional area, exposed position, and numerous connections of the parts of most iron roofs, corrosion will have a serious effect on their strength, and as there are many metallic roofs— such as those over railway stations and goods sheds, riding schools, drill and public halls, &c— of various kinds, careful periodical inspection should be made to ascertain the magnitude of the corrosion, as many are peculiarly subject to corrosive influences. In the ' G-esundheits-Ingenieur,' 1888, particulars are given of some experiments made by the Bonifacias Colliery Co., at Kray, to ascertain the best metal to withstand exposure to the acid waters of the pits. Tests extended for months with wrought iron, steel, and " delta " metal (an alloy of copper, zinc and iron). Test pieces were 0 • 79 by 0*79 by 0-75 of an inch. They were suspended in the water, and were carefully weighed before and after, and the percentages of loss of weight were found to be respectively 45*9 for wrought-iron, 45-45 for steel, and but 1-2 per cent, for "delta" metal. The company therefore adopted "delta" metal for their underground machinery. These experiments can hardly be considered conclusive as regards corrosion, for a simple weight test is insufficient ; and a chemical analysis, before and after testing, is necessary to enable any reliable result to be attained, although the metal may be particularly anti-corrosive. Still, the behaviour of some alloys is very varied and often peculiar ; and, again, it is absolutely neces- sary the tests are conducted so that there are no interfering conditions, such as galvanic action, &c. 82 COKKOSION AND ITS PKEVENTION. To prevent chemical action caused by electrical currents, the wrought-iron bars forming the skeleton of the statue of Liberty on Bedloe's Island, New York Harbour, and to which the copper skin of the statue is fixed, were first painted with shellac, and then enveloped in asbestos fibre, although it has been remarked they may eventually be found to be of little use should electrical currents be set up. Some few years ago, in Paris, bare copper mains were used for the underground electric light wires. The conduits were made of cement concrete, and were considered to be water- tight, but the copper became covered with verdigris, and so reduced in thiokness, that fears were entertained it might become so corroded as to be the cause of overheating with the normal currents. Earthenware troughs were found to be not waterproof, and, by a complicated electrolytic process, explo- sions were considered not unlikely to be brought about. One such occurred, although not a serious one, and gas was proved to have had nothing to do with it. Examination of the conduit showed that carbonate of soda and caustic soda were present, and the conductors were thickly covered with verdigris and copper chloride. The cause of the explosion was considered to be that the explosive gaseous mixture contained oxygen, hydrogen, and chlorine, the latter gas being due to the presence of chloride of sodium in the water which filtered into the earthenware conduit, the roads above having been strewn with salt to melt the snow, and caused electrolysis. The conductors are, generally, now covered with a bituminous compound to prevent electrolytic action, and earthenware troughs have been abandoned.* In the Ferrara district, in North Italy, notwithstanding that the quality of the iron was good, the pipes in some artesian wells, 8^ to 11 inches in diameter, and £ inch in thickness, were badly corroded in six months, a thin outer skin only remaining. On the other hand, the Modanese tube wells, which are generally of small depth, do not reach * For further particulars, see ~U Electricien, 1892. NOTES ON BAPID CORROSION, WITH EXAMPLES. 83 sea level, and their water is from gravelly streams almost en- tirely free from iron. They appear to he quite free from cor- rosion. In the Ferrara wells, it was considered, two causes brought about the destruction of the pipes, one mechanical and the other chemical, viz., the friction of sandy and gravelly particles wearing away the skin, and, the water then being able to set free the iron, slowly dissolves chemically the interior core till it reaches the outer face. Whichever side is the lower, the sand will affect the most by gravitation. Wooden pipes have been, therefore, used in the Galara dis- trict, and, it is considered, long-continued experiment is required to solve the problem created, viz., the manner in which to prevent the rapid wearing away and corrosion of the tube wells. At the Southampton Waterworks, Mr. W. Matthews, M. Inst. C.B.,* has stated, in the apparatus used for softening the water obtained from chalk wells and borings, " All the bright work on the filters and lime cylinders is, so far as possible, nickel-plated, as the great amount of condensation produced by the proximity of tanks, containing water at a low temperature, would otherwise render it almost impossible to keep the fittings free from rust." At some of the soda nitrate works, where, in the water, salt and alkaline matters are present, the boilers have to be cleaned frequently, and at intervals of not more than two months. Raw nitrate of soda is a mineral deposit, which is considered to have been formed by decomposing animal and vegetable matters becoming acted upon by the salts present m sea water. Such soil or substance is actively corrosive. In the case of a cylinder bridge-pier erected on the Bom- bay, Baroda, and Central India Railway, it was thought the rapid corrosion of the bolts was caused by the proximity of the sea, as the water was brackish, and, therefore, the action was much more rapid than in fresh water. * 'Minutes of Proceedings,' Inst. C.E., vol. oviii. g a 84 CORROSION AND ITS PREVENTION. CHAPTER VII. THE CORROSIVE INFLUENCE OF SOILS, VEGETATION, SITUATION, CLIMATE, RAINFALL AND WATER. Corrosion of iron or steel buried in the ground will vary, according to the nature of the soil and its character, the porosity or imperviousness greatly influencing the relative powers of corrosibility. Although an iron column or "base plate is in the soil, it cannot be regarded as altogether free from air or fresh supplies of moisture, and other decaying agents. When water enters the pores of any soil it displaces the air upwards, and the liquid previously present down- wards ; hence, for a certain depth, varying with the com- position of the earth, the metal may be severely tried, as corrosive influences will be frequently renewed, and if such water is derived from manured fields, sewage, or irrigation, or from peaty soils in which humic acid is present, the corrosive effect will be much increased. Those who have made a life study of the question, and have conducted numer- ous experiments, state that when water has reached a distance of even ten feet from the surface, it is doubtful whether it has got entirely beyond all atmospheric influences, and may not again be drawn up within the limits of superficial evaporation. Hence the portions of a metallic structure which are buried in the earth may be subject to severe corrosive influences, depending upon the nature of the soil. The amount of percolation is found to be governed more by the time at which the rain falls, and the manner in which it is distributed, than on its actual quantity. COEEOSIVE INFLUENCE OF SOILS, ETC. 85 The absorbing power of soils has also to be considered, for dry loamy earths imbibe more water than sandy soils, but the percolation may not be so great. The density, looseness, and open nature of the earth have also to be con- sidered, and much affect the absorption, for particles in suspension, smaller than the interstices in the soil, and lighter than the water, pass through or into it till the earth becomes waterlogged, or the water accumulates upon an impervious stratum. Any other transferred matter is re- tained by the soil. Barren soil does not absorb so much urea as that covered by vegetation, but, in the latter case, all the urea in any sewage water is not abstracted, and, therefore, some remains to exert corrosive influences. The effluent of sewage-irrigation water contains some ammonia and chlorine, &c, and these have a decidedly active corrosive influence. The effect upon the paint or composition covering the metal has also to be considered, and this is referred to in the Second Part of this book, relating to Fouling and Corrosion in Submerged Structures and Ships, and their prevention by paint, &c. Earths absorb water, or allow it to percolate. Sand, all loamy soils, gravel, and almost all earths having fine particles not compacted, readily admit water, if they do not absorb it. Gravel on chalk is considered to make a dry foundation, and to conduce to dry air, especially if the localities be at a considerable elevation above the sea, and are fully exposed to the sun. Blue clay, compact gravel, which is sometimes almost of the nature of rough weak concrete, and most rocks consisting of close even grains, and if not fissured, are of an impervious character ; however, clays have moisture in them, and, it may be said, earth, or even rock, upon which metallic engineering structures have to be founded. In deep founda- tions, which may reach to a soil known to be not of a corrosive nature, rain water may percolate through the upper strata, or down fissures, and so corrosive influences may be conveyed to the lower beds. This action has been beautifully- described as follows : — i 86 CORROSION AND ITS PREVENTION. " I see the leaning strata, artful rang'd ; The gaping fissures to receive the rains, The melting snows, and ever-dripping fogs. Strew'd bibulous above I see the sands, The pebbly gravel next, the layers then Of mingled, moulds, of more retentive earths, The gutter'd rocks, and mazy-running clefts ; That, while the stealing moisture they transmit, Eetard its motion, and forbid its waste. Beneath th' incessant weeping of these drains I see the rocky syphons stretch'd immense, The mighty reservoirs, of harden'd chalk, Or stiff compacted clay, capacious form'd. O'erflowing thence, the congregated stores, The crystal treasures of the liquid world, Through the stirr'd sands a bubbling passage burst, And swelling out, around the middle steep, Or from the bottoms of the bosom'd hills, In pure effusion flow. United, thus,' Th' exhaling sun, the vapour-burden'd air, The gelid mountains, that to rain condens'd These vapours, in continual current draw, And send them, o'er the fair divided earth, In bounteous rivers to the deep again, A social commerce hold, and firm support The full adjusted harmony of things." At the point where a porous soil overlies an impervious there will be dampness, even if the water flows away. The word imperviousness is here only used by way of comparison, and not to indicate absolute imperviousness. Earth that has been excavated, deposited, or turned over, is more or less aerated and porous, and therefore is favourable to the admis- sion of dampness, to corrosive influences, and the growth of vegetation. It is well to remember that made or polluted earth, to a few feet below the ground level, contains many organisms and impurities which exist to a very much less extent below any such depth. Experiments have shown that after ten feet or so below the ground, many exist in hardly appreciable number or quantity, and therefore in- creased corrosion may be expected at or near the surface. CORROSIVE INFLUENCE OP SOILS, ETC. 87 Bacteria, however, have been found in wells, and it is thought were introduced into the subsoil during the sinking operations, and that they found a congenial soil for develop- ment. Soils neither equally acquire heat, retain or radiate it, nor so absorb moisture. Differences of polish of a metal, its rough- ness or smoothness, and colour when painted, will also affect the radiation and absorption of heat. Clayey or marshy grounds cause a diminution of temperature, and, in warm humid climates, are unhealthy and actively corrosive. Light, calcareous, and stony soils, have a much healthier effect. The saline nature of a soil causes it to be decidedly corrosive. Structures erected in cultivated soils will be subject to much more active corrosive influences than in barren ground, because the former are comparatively loose and porous, and freely radiate in the night the heat they absorbed in the day ; therefore they become cooler, and hence they condense the vapour of the passing air, and it falls in dew. The effect of the sun is to harden earth, consequent upon the evaporation of moisture, and to bring the particles closer together, and make the mass more compact. If this were not so, heat and drought would penetrate the soil and kill all vegetation. In this respect heat retards corrosion. At what seasons of the year may corroding influences be usually considered to be most active ? In autumn and winter. Why ? Because the capacity of the air for holding water decreases after the summer season is past, and vapour is consequently deposited upon everything with which the air has contact ; hence the necessity of preventing the surface of the metal being exposed to air and water, and the value of even simple grease as a protecting covering to prevent the humidity of the air reaching the metallic surface. For the same reason, in a damp climate corrosion will be much more rapid than in an arid district and one having a drying atmosphere and almost continual sunshine or constantly cold climate ; in fact, it might be said, in comparing districts or countries, that wherever vegetation is least, corrosion will be reduced 88 COEEOSION AND ITS PREVENTION to a minimum — a damp tropical condition, vhich quickly produces vegetable life, whether mould or most, or in higher forms, being the worst. In Great Britain, the rate of evapo- ration, of course, is most during summer, and the amount of humidity in suspension is usually greatest tht last and first two months of the year, but not always so. However, it may be considered there is most moisture, nrt rain, in the atmosphere in the months which generally hive the lower temperatures. Evaporation of moisture in earth being most in the summer months, it is favourable to the production of salts in the earth, which have a corrosive influence. Although water may not percolate a clay soil to any con- siderable depth, cracks and fissures occur in t in very dry weather, down which air and moisture penetrate. A clayey subsoil is also subject to bulging in cold weaiher, so that it cannot be said ironwork bedded in it only mdergoes the corrosive influences that the soil possesses in s>lid mass. In sandy soils, on land reclaimed from the sea, ihe salt water soon loses its deleterious action on plant and vegetable life, but clay soils do so very slowly, and may reqiire two years or more before they can be sown. Hence, clav soils impreg- nated with sea Water will have a more active ils. Alluvial soils often consist of fine mud, and contain minenl constituents and some organic matter, animal and vegetable. Diluvial soils are more usually stony and barren, and consist of large gravel, boulders, stones and clay. The mere memical com- COKROSIVE INFLUENCE OF SOILS, ETC. 89 position of a soil will not indicate its corrosive influence, although necessary to its accurate determination, for the earth, from local causes, may he impregnated with substances and liquids almost unknown in the pure soil ; still, to know the geological formation, with a due regard to any special local corrosive influence, will enable the probable corrosive effect to be judged. Insoluble silicious matter forms in clean sand about 90 per cent, of it, the remaining 10 per cent, being oxide of iron, alumina, carbonate of lime, and various other ingredients, according to the locality. Sandy soils vary greatly ; for instance, the sandy downs along the coast of the North Sea are found to be less contaminated than other soils, and they have very considerable oxidising power. The oxidation or nitrification of water, by percolation through sand, destroys organic matter contained in it. Sand downs decidedly attract atmospheric vapour not precipitated in the form of rain. This is not due to capillarity, i.e., the power which very minute tubes possess of causing a liquid to rise in them above its level, as, after long droughts, at a slight depth on the hill tops, sand is found to be damp, although the lower surfaces are dry to greater depths at such times. A dry sandy soil is warmer than a wet marshy soil subject to great evaporation, and the latter has a corrosive tendency. Peaty soils are chiefly composed of marsh and water plants and moss, the layers having successively dried. Consequent upon the action of humic acid, they become changed, and are actively corrosive. The mists, which are of frequent occurrence over marshes and rivers at night, also have a corrosive effect, and are caused by the air over them being almost always near saturation. The least fall in temperature causes the air to release some of its moisture in the form of dew or mist. Fogs are caused by the con- densation of the vapour of the air,'when near the point of saturation, close to the surface of the earth, whereas rain falls from a considerable height. Soil is sometimes composed of alternations of gravel and sand. The gravel may be consolidated by the admixture of 90 CORROSION AND ITS PREVENTION. carbonate of lime from limestone pebbles. Light loam usually contains a large proportion of micaceous particles. Some arenaceous limestone, on analysis, was found to be composed as follows : clay, 5 • 60 ; fine sand, 9 • 30 ; oxide of iron, 2 • 00 ; carbonate of lime, 80*00 ; carbonate of magnesia, 1 • 90 ; mois- ture and loss, 1*20: total 100. Sulpbur is found in mica slate, limestone, beds of gypsum, sandstone, in alluvium, &c. Soils obtain carbonic acid from the air from which it descends by rain. It is also abundantly present in all decomposed vegetable and animal matters. Limestone, chalk, and all calcareous stones, have carbonic acid in them in a solid state, and the presence of any acid sets it free. The soil of Egypt, some analyses in reports made to the French ministry of agriculture have shown, consists of about 45 per cent, of silica and 53 per cent, of clay, and in general is composed of silica, alumina and oxide of iron, with small quantities of phosphoric acid, and not more than from 4 to 9 per cent, of organic matter ; and a larger proportion of chloride of sodium is found on approaching the Mediterranean, attaining 4 per cent, in the middle of the delta. The Nile mud contains, on an average, at its high and low stages respectively, 55 and 58 per cent, of silica, 20 and 23 per cent, of alumina and oxide of iron, and 15 and 10 per cent, of organic matters, together with small quantities, 0 • 66 to 3 per cent., of phosphoric acid, lime, magnesia, potash, soda, and carbonic acid. The value of chemical analysis of an earth or mud can be judged from the preceding examples in any reliable deter- mination of the presence of corrosive substances in it. The alluvial matter in river water is either deposited on the shores, increasing as the course nears the sea or mouth, or the current is sufficiently strong to carry it away. If not, salt marshes will be created, which is usually the case in estuaries. Such waters and earths have an active corrosive influence on metallic structures, and not infrequently in other situations various corrosive salts are found in them. The soil of forests is generally covered with long damp grass, COKEOSIVE INFLUENCE OF SOILS, ETC. 91 rotting leaves, "brushwood, &c, and there are places full of or holding stagnant water. Low-lying and damp situations also give a maximum of atmospheric pressure. Climate, soil, humidity, light, elevation above the sea, &c, cause plants to grow in some localities and not in others. As regards temperature, it is the degree of heat, and the duration of the summer months, that is the chief agent in limiting the range of vegetable life, and not the mean annual highest or lowest temperature that may prevail. At the sides of a pond, or sheet of water, leaves and debris will be found, unless driven by wind, because the shore attracts the leaves, &c. Vegeta- tion is also, on the margin of a river, more abundant and strong than in an open field, because the porous earth on the bank draws, by capillary attraction, water to the roots of the plants. Consideration of the causes of plant life will enable its probable corrosive effect to be ascertained. The green colour of plants is given by light, the intensity increasing with the brilliancy of the light. Heat and moisture are the other indispensable requirements for the vegetation of plants. One of the chief causes of the remark- able growth of tropical vegetation is the considerable quantity of water contained in a tropical atmosphere in the condition of transparent gas. It is known to be pro- portional to the heat, and to be not affected by what is known as a clear sky. This invisible water is absorbed by the plants, and taken up by their large leaves, and so they luxuriate to such an extent as to be of marvellous growth, whereas in our temperate climate, without rain vegetation would be parched up and perish. This beneficently bestowed property of tropical regions actively produces corrosive and fouling influences. In a valley in a mountainous country corrosion may not equally proceed on both sides of it, for there will be a sunny side and a more or less sunless side, and more moisture necessarily on one than the other. The direction of the valley or gorge will, therefore, govern the rate of corrosion, and the sunny side is almost always the one to be preferred 92 CORROSION AND ITS PREVENTION. for the erection of metallic structures, if preference can be given, simply from the point of view of facility and con- struction and durability. The cbaracter of the vegetation will indicate the warmth of climate. For instance, in the Alps of the Vaiais, on one side may be seen the vine growing, on the other thick ice. The slope of the country, and the aspect it presents to the sun's course, causes this difference, the side on which the vine grows being exposed to the direct rays of the sun. On the southern terraces of the Himalayan slope, the pine-apple, cotton tree, mango, &c, grow, even at an altitude of 5000 feet ; this tropical vegetation then gives place to the plants of a temperate region, and, higher still, to alpine forms of vegetation. The power of the sun in heating the soil varies according to the angle at which the sun's rays strike the ground. When the sun is at an angle of 45° above the horizon, the sun's rays will fall perpendicularly on a hill facing the south having an equal slope. The level land below will receive them at an angle of 45°. If the northern slope of a hill inclines at an angle of 45°, the rays will be parallel to the surface, and their heating effect will be very little. If the northern slope be steeper than 45° it will be in the shade. A S.S.W. or S.W. aspect is the warmest in Great Britain and the Continent, and a N.N.E. or N.E. the coldest, unless special local circumstances cause a modification of the general con- dition ; therefore, corrosive influences will vary, and while one side of a hill may be exposed to them, the other may only be to a much less extent. It is known that some lakes and many rivers present an exceptional climate. Such peculiarities are usually the result of protection by mountains and hills from north and other cold winds, and from their being exposed to the sun. Also, near some rivers, what are called " river clouds " occur at certain seasons, and this is considered due to the temperature being much more constant than further inland, which is quickly cooled by radiation. Vegetation which does not exist inland often thrives on the borders of lakes, and so just around the Lombardy lakes vegetation flourishes that is COKKOSIVE INFLUENCE OF SOILS, ETC. 93 found usually only in much more southern and warmer places. It is the same at Lake Geneva, the north side of the lake, exposed to the sun's rays, and sheltered from north winds, "being more favoured than the south. It was found by experiment that, with the sun at an altitude of 30°, the reflected heat was nothing : and 60 to 70 per cent, of the incident heat between about 4° 30' and 3° 30'. The reflected heat at 16° 30' was about 20 to 30 per cent, of the incident heat. The north and south bank of a river stretching east and west, and sheltered by mountains or hills, will usually show different plants growing on one side to the other, the north side having those which grow in the warmer regions of the earth, for the north side faces the south. These phenomena it is well to remember, as corrosion is not likely to be the same on both sides of a river or mountain under such circumstances. On the north side, having a southern aspect, there would be much more sun and drying of the surface. In connection with the special subjects of this chapter, it is advisable to ascertain the annual quantity of rainfall, and months in which it falls ; whether mists, fogs, dews, &c, are frequent, and the atmosphere usually in a damp, warm con- dition ; the latitude and elevation of the site of a metallic structure above the sea-level; the location of the site as regards the sea; form and nature of the land, whether mountainous, swampy or wooded ; its position with respect to the sun's course, especially in a valley, as on one side the sun may shine, but not on both, and perhaps never on either ; the character of the sea, and waters of rivers, &c, rain, and the nature of the soil ; the prevailing winds, remembering that those countries where the winds are the most variable are the most cloudy, as Great Britain; the rate of growth, and species of vegetation, also of mollusca, and all life that aids corrosion, decay, and fouling ; the effect of these influences on the metal, and on any paint or composi- tion used to cover it or supposed to protect it ; and also, whether, as in screw or other piles or ironwork, the paint 94 CORROSION AND ITS PREVENTION. is likely to be removed or injured in screwing or during erection. The rainfall, or moist character of the air, which is governed, among other causes, by the situation and configura- tion of the land, and its impurity or freedom from corrosive agents, will vary in almost every district on the known surface of the earth. Natural water being impure, because of its contact with the earth and the atmosphere, it is the character and intensity of the impurities in it as regards corrosive influences that have to be here regarded. The quantity varies greatly : thus, at Bergen, the commercial capital of western Norway, the rainfall is excessive, nearly four times that of the east coast of England, and amounts to about 88 inches of rain, the heaviest of any city in Europe, and more than that of many places situated in the tropics. The reason is, clouds are driven forward by south-westerly winds into the Norwegian fiords, until they are arrested by the mountains, and accumulate, then rain is caused. The proximity of mountains attracts vapours and con- denses them, and hence heavy and continuous rain occurs. As a rule there is more rain in islands and on sea-coasts than in inland districts or places far removed from the sea — as the broad plains of Asia, Africa, Siberia, some parts of Eastern Europe, and America — and in mountainous places than in flat countries, and in the tropics than in more temperate zones, with a few exceptions, such as Bergen. In general, the rain- fall becomes less with the increase of the latitude, but mountains or numerous hills will cause a few exceptions, as at Bergen, Keswick, Kendal, &c. The causes of rain and deposition of moisture are so well known that there is no occasion to here refer to them, except so far as is absolutely necessary. The quantity of rainfall or moisture retained must vary according to the form of the land, as witnessed in mountainous countries where the water may at once flow away, or be retained as in marshes or flat localities. The character of the surface soils and subsoils will much affect the corrosive activity of water flowing over COEEOSIVE INFLUENCE OF SOILS, ETC. 95 or through them. In the north and south zones of variable rains, i. e., from between about latitude 20° N. and the Arctic circle, and 20° S. and the Antarctic circle; the zones of periodical rains, say, from between latitude 20° N. and 20° S. ; the zone of almost constant precipitation, situate about 3° to 10° N., extending from South America and reaching to Africa ; and in the rainless and riverless districts in the north of Africa and a few other places ; corrosive influences must necessarily vary, and it is, therefore, indispensable that the characteristics of the country in which a metallic structure has to be erected be duly considered. Dr. Angus Smith has shown that there are innumerable solid particles in the air, consisting of common salt, sulphate of soda, nitrate of ammonia, and sometimes lime-salts and iron, as well as phosphates, iodides, and organic substances, given off from animals and vegetables and living things, and probably a little of nearly everything at times. The per- centage of oxygen in the air varies from all but 21 per cent, to 20 • 40. After rain, the atmosphere contains a larger per- centage of oxygen than it did previously, and there is at the same time a diminution of carbonic acid. Eain brings down the " atmospheric dust," and carbonic acid and other gases, by the mechanical and chemical aotions of the falling water. In rain waters, chlorides, sulphates, and ammonia have been found. Hydrochloric acid in the air is due to the salt found near the sea, and in the neighbourhood of towns, where it is one of the products of coal burning ; but rain water near the sea is not acid, but that taken in towns is so because it contains sulphur compounds, and the latter also become present where decaying animal and vegetable matter are. Sulphuric acid is a measure of decomposition, being a part of the oxidised sewage of the air. The acidity of rain water is also a measure of the impurity of the atmosphere. Dr. Angus Smith found that, taking the inland country parts of England as 0, Glasgow figures 109, Manchester and Liver- pool 83, and London 28. The comparative amounts of combined ammonia are shown by Valentia 1, Glasgow 50, 96 COKKOSION AND ITS PKEVENTION. Liverpool 30, Manchester 36. Hydrochloric acid : Blackpool 100, London 320, Underground Kailway 974. These being comparative figures, as are the following : — Sulphuric acid (anhydrous) : Blackpool 100, London 352, Underground Bail way 1554. Ammonia and albuminoid ammonia : Innellan on Firth of Clyde 100, London 117 and 108, Glasgow 150 and 221, and the Underground Bailway 138 and 271. Single experimental results should, however, not be regarded as absolutely conclusive, but the mean of a number be found before deciding upon the amount of impurity in the air of any particular place. It may be useful to remember these relative quantities when considering the probable corrosive influences to be encountered. The rainfall of manufacturing towns, and of cities and towns, has a much more corrosive effect than that of rain as it falls in the open country or mountainous districts, because of its impregnation with the air ; but comparatively pure rain water may not be so after percolating the earth for a little distance, because of the corrosive effect of manures and substances in it in well-cultivated land. It may be said, generally, the impurities of the air are indicated by an analysis of the rain water of the locality, and that in dense manufacturing districts, or a copper-smelting and chemical- producing town, such as Widnes on the Mersey, acids are present which can hardly be said to exist in more favoured places. Many acids have such a strong affinity for water, that it is very difficult to entirely separate them from it. Water has very great solvent powers, and evaporates at all temperatures to which engineering structures are subject. Bain waters, and dews to a lesser extent, in country districts, in falling to the earth, take up the gases, oxygen, nitrogen and carbonic acid, a little carbonate of ammonia, nitric acid, &c, including any particles which may be floating in the air. In towns and manufacturing districts, in addition, sulphur- ous acid, sulphuretted hydrogen, &c. ; and if rain water runs from a roof or down pipes, the impurities in it will be various. In considering the corrosion caused by rainfall, CORROSIVE INFLUENCE OF SOILS, ETC. 97 it should not be forgotten that in many tropical countries the rain descends in torrents for many hours, and in quantity even as great as about one inch in an hour, and also very violently, so much so as to cause districts to be flooded in a short time. The effect of such rain is to very severely try the paint on a metallic structure, as the constant beating of the rain against its surface will wear away any but the most tenacious and hard coating. An instance of the de- structive effects of tropical rains in damp situations may be mentioned. In the Western Himalayas, Major Browne, R.E., has stated, " Retaining walls of what seemed most compact sandstone have suddenly collapsed, the underground courses have become dissolved into sand." Such walls, from the foundations to two feet above the ground line, are therefore built of granite boulders. This shows the active decaying force that may be at work upon any painted or coated surface of iron or steel, and the desirability of the metal being embedded'in impervious Portland cement concrete, or otherwise effectively protected. It would be of no practical value to give analyses of the water of different seas and rivers, for they are subject to frequent changes and notable differences. The character of the water resulting from rain or floods will generally vary somewhat daily everywhere, and it is advisable to consider the kind of soil, whether argillaceous, calcareous, ferru- ginous, or of a special nature, that the water has flowed over or percolates through. In the sea, rivers and streams, con- sequent upon the ceaseless change of water, and consequent constant fresh supply of corrosive influences, iron or paint are in a perpetual state of siege, and very much more so than if the water was quiescent. Analysis will seldom show the composition of the waters of rivers to be the same for the total solid residue, combustible matter, the ash, sulphuric acid, chlorine and common salt, iron, ammonia, nitrogen, &c. The bed of a stream, in some cases, may be covered with weeds, entangled and dense, and replete with Crustacea, mollusca, and the lower forms of life, and this becomes less H 98 COEEOSION AND ITS PEE VENT ION. till at last the bed appears to be clean, and the water quite clear, although still full of animalcula. Fresh water algae exist in large numbers and great variety in most water, and are very hardy and tenacious of life. They flourish even in the Arctic regions. The species are very numerous ; some seem to live in waters containing sulphur, salt or iron, some in stagnant water, and others in the clearest and purest. Starch and globules of oil have been found to spring from them, and decay and taint the water, also a kind of jelly from diatoms. Especially in many warm climates, rivers are generally in a turbid condition, and the water never free from argillaceous substances, and, if exposed to the sun, the formation of fungus on the surface of the water will be more or less rapid, according to the location and special circumstances of the site of a structure, and will be greatest during the spring and summer season. The composition of water varies greatly, and is affected by the nature of the ground it passes over or through. It has in it organic and inorganic compounds, and the inorganic compounds may be sodium chloride, carbonate, nitrate, sul- phate, sulphate and carbonate of calcium, sulphate and carbonate of magnesium, silicic acid, alumina and oxide of iron, potash, chlorine, ammonia, phosphoric oxide, sulphuric acid, &c. The salts usually contained in hard water are carbonate of lime, chloride of sodium or common salt, sul- phate of lime, sulphate and carbonate of magnesia, and sometimes carbonate and oxide of iron. Spring water con- tains a considerable quantity of carbonic acid, and is, therefore, pleasant to drink. The chlorides, sulphates, and nitrates of calcium and magnesium are easily dissolved and main- tained in solution by water, but the carbonates of these elements can only be maintained in solution by an excess of carbonic acid in the form of bicarbonates. Mechanical impurities in water are removed by filtration, but chemical impurities may not be. A quick method of ascertaining whether there is vegetable and animal matter in water is by adding a little sulphuric acid to it, the water will then become COEROSIVE INFLUENCE OF SOILS, ETC. 99 black. Water charged with carbonic acid disintegrates, by continuous action, limestone, granite and hard rocks; and zinc and galvanised iron, which do not appreciably corrode in a clear and comparatively pure atmosphere, if exposed to the carbonic acid, which is present in the air of some manufacturing towns, and is also in abundance in some water, decay, and the galvanised coating becomes loose or detached, by corrosion of the surface of the iron underneath. How is the carbonic acid in water produced? By the presence of bicarbonate of lime, which is a salt composed of lime and carbonic acid, and is frequently held in solution by hard water, and chiefly causes it to be more agreeable to the taste than soft water ; and hard water is made softer by exposure to the air, because not only do the mineral salts subside which cause it to be hard, but the carbonic acid of the water escapes into the air. A copious supply of water prevents the accumulation of carbonic acid by dissolving it, but it causes the water in which it is dissolved to be slightly acid. An analysis by Dr. Lyon is here given of what may be called an actively-corrosive water, viz., that at the estuarine mouth of a river. The water of the Nerbudda River, in India, taken at a bridge-crossing of the Bombay, Baroda and Central India Bailway, was found to contain per gallon — Grains. Chloride of sod ium 139-40 Chloride of magnesia 8 • 64 Sulphate of zinc 11 -34 Sulphate of magnesium 7 " 84 Carbonate of magnesium 9 • 37 Silica 1-34 Organic matter -56 178-49 Suspended matter, 72 • 91. Specific gravity of clear water, 1-00206. Total solids, dried by evaporation at 280° F., 178-64. h 2 100 CORROSION AND ITS PREVENTION". When water is rich in lime, potash, and sodium chloride, with marked traces of nitrates and phosphoric acid, it is known pollution has taken place. There will be a quick growth of vegetable organisms, weeds and mud will be pro- duced, and then corrosion and fouling may follow. Water containing ammonia, fcecal and excrementitious matter, and that impregnated with salt, is actively corrosive. Sewage water, in addition to other constituents, usually contains nitrogen, sulphuric acid, chlorine, phosphoric acid, potash, soda, lime, ammonia ; the nitrogen being present in organic compounds and as ammonia, but in the effluent chiefly as ammonia and as nitrous and nitric acid, and only in small quantities in organic compounds. Nitric acid is frequently found in impure waters in considerable amount. The quan- tity of suspended matter in, and the character of sewage varies greatly according to the locality. It will also be affected by the waste effluent from breweries, factories, and by water from mines, peat bogs, chemical works, wire-drawing manufactories, where acid is used for pickling the wires, and by any which contains free acids and in which salts have been dissolved, such as chloride of magnesium; also that from dung-heaps and slaughter-houses. Water of this de- scription is decidedly actively corrosive. The character of the drainage waters of towns or manufacturing districts is different on Sundays and holidays to that of the ordinary working days. The first flow of a flood will generally bring down vegetable and organic matter, and then alluvial matter, as the floods, which are frequently helped by gales, increase ; and so corrosive and fouling influences are seldom regular and of the same power. At one time, as in dry weather, the alluvial matter carried may be very small, but it will probably be large in wet weather and at flood time, and particularly so if a flood or gale occurs after a long drought. The character of the sand and mud and matter in suspension in river water can be sufficiently ascertained, but the ma- terial that rolls along the bed and at the sides cannot be so COREOSIVE INFLUENCE OP SOILS, ETC. 101 nearly determined. The dissolved matter, or that in sus- pension, will be dependent upon the condition and nature of the soil it has passed over or percolated through, and the length of time taken to complete either of those operations. The harder silieious rocks will give off but little mineral matter in solution to water passing over them, or even rising in fissures in them. But in the soft rocks, chalk, or calcareous earth, marly and alluvial soil, and such soft earth, it may be found in much increased percentage, either in solution or in suspension. Waters which hold in solution much organic matter quickly putrefy. Ponds, marshes, bog-land, most stagnant waters, &c, are examples of this. The organic matter becomes altered, and is in suspension or dissolution. Anything which causes putrefaction must be considered as an active corrosive agent to the metal and the paint. The effect upon the latter is considered in the second part of this book, on Fouling and Corrosion in Submerged Structures and Ships, and their prevention by paint, &c. Waters flowing over rocky or clean gravel, or sandy beds, are obviously much purer and less corrosive than those which have been contaminated with the surface earth of cultivated land, or deposits that become in a state of decay or putre- faction from the decomposition of vegetable or animal matters. The substances contained in solution in the waters of the Nile do not materially differ from those found in many other rivers, but their relative proportions appear to vary accord- ing to the stage of the river. Again, the amount of silt deposited in reservoirs depends upon the catchment area, rainfall, and the geological character of the soil. For instance, in Algeria, in some reservoirs, it ranges from 0*16 to 1 • 6 cubic yards yearly per acre of catchment. The silt hardens by time, and therefore it is easier to clear, being looser, and to carry off, if done yearly or so. While water flows underground its impurities are held in solution by the presence of carbonic acid, but when the stream reaches the open air its carbonic acid escapes, and these impurities, which are especially carbonate of lime and iron, are precipitated on 102 CORKOSION AND ITS PREVENTION". various substances lying in the course of the stream, and cause petrifaction. A structure exposed to petrifying spring water, or in- crusting springs, i.e., those covering things with a crust of the mineral they hold in solution, obtained in trickling through the ground by dissolving some of the substances they encounter, as in springs from limestone mountains which incrust substances with lime or calcareous matter in a short time, brine springs with salt, sulphur, iron, mineral and medicinal springs, are also numerous, and most of these do not materially change, or have not changed since they were discovered, and may be considered as actively corrosive. The influences of light, air, evaporation, loss of temperature, absorption and escape of carbonic acid, cause them to deposit the mineral matter. Some mineral springs have a high tem- perature. In Italy, Mexico, and some parts of Spain and Portugal, they are numerous. Those in Great Britain are so well known that they need not be specified. There is hardly a country which does not possess several. Chalybeate springs, or those containing iron, are very numerous, and deposit oxide of iron wherever they flow. Saline springs less fre- quently occur. Many of the chalybeate and saline springs are impregnated naturally with carbonic acid gas, which makes them exhilarating. Any metallic structure subject to any such waters requires to be specially protected against corrosive influences, and the effect of the waters on the paint or composition used has to be duly considered, as any coating may soon be deleteriously affected. It may here be well to refer to some experiments made by Professor August Wagner upon the rusting of iron under water at the ordinary temperature. He found that under distilled water, saturated with carbonic acid and air, iron rusts nearly twice as much as under water containing air only. If, instead of distilled water, spring water concentrated by evaporation is used, an essentially different result is ob- tained ; in this case, iron rusts in water containing only air more strongly than in water containing air and carbonic acid. CORROSIVE INFLUENCE OF SOILS, ETC. 103 And it was generally found that iron under water containing air, but free from carbonic acid, with the presence of salts, except alkaline reagents, rusted much more quickly than in pure water. In water containing chloride of barium and chloride of calcium, saturated with air free from carbonic acid, rusting is very energetic, but less striking with air and carbonic acid. Under water containing common salt and chloride of potassium, as well as under water containing ammonia, rusting in the presence of air and carbonic acid proceeds with the greatest energy. Under water containing chloride of magnesium, rusting occurs with air and carbonic acid more strongly than in the presence of air free from car- bonic acid. Additions of oil greatly prevent rust in water containing only air, as well as in water containing air and carbonic acid. Alkaline reagents, as lime and soda, tend to prevent the rusting of iron. Both at the ordinary tempera- ture, and at the boiling point, the presence in water of such dissolved chlorine combinations as chloride of magnesium, of ammonium, sodium, barium, potassium, and lime, is very de- structive to the iron so soon as air is admitted. The addition of alkaline substances, such as lime and soda, to the feed- water of boilers prevents, it was found, rusting of the boiler plate. Experience and experiments show that hot sea water acts more corrosively on boiler plates than hot fresh water, and that the less sea water there is mixed with the fresh, the less the corrosive influence ; also, the less frequently either water is changed or renewed, and that, if the immersed metal is occasionally exposed to the atmosphere, and again immersed, corrosion is accelerated. Mr. Hering, M.I.C.E., with regard to the pollution of streams, draws the following inferences from the best data available in 1888 : " That rivers not to be used for domestic water supplies, but which must remain inoffensive to communities residing a few miles below the outfall, to be fit for all manufacturing purposes, and to sus- tain the life offish, may receive the sewage from 1000 people for at least every 150 or 200 cubic feet minimum flow per 104 CORROSION AND ITS PREVENTION. minute." Any such water, however, will have a corrosive influence. Lakes having no stream flowing through them are to be regarded as stagnant waters, hut the rock on which they generally repose, and aeration of the water by wind, much lessens the effect of no flow, which would otherwise aid the growth of weeds and decomposing influences. Why? Because plants grow in stagnant water, and leaves, &c. become de- composed in it, and insects lay their eggs in the leaves and plants floating on the surface, the result being a gradually increasing swarm of worms and insects. There is usually more suspensory and corroding matter in river than in lake water, chiefly consequent upon the greater number of cities, towns and villages per acre of water area, and the more polluted nature of the supplies, and also because of the increased facilities offered in lakes for the deposition or precipitation of suspensory matter, owing to the comparative stillness of the water. The salt in the water at the mouths of rivers is due to admixture of oceanic brine with the out- flowing fresh water, and at such places corrosion is likely to be severe, for brackish water acts more rapidly on iron than fresh water. Sea water flowing up a river will much increase the percentage of chlorine in it. The salt water so entering, it has been noticed, sinks to the bottom. This is owing to the different specific gravity of fresh and sea water, which, therefore, do not readily commingle, the sea water being separated from the fresh. Fresh water being lighter than sea water, it, as it were, floats and spreads on it until it becomes commingled. A current of saltwater meeting fresh, will, at first, force itself under the fresh water. Also, water heavier than that into which it is discharged, as sewage, displaces the lighter water, and proceeds under it till inter- mingling occurs by deposit of suspended matter, or by other causes. Eemembrance of this may indicate its possible relative corrosive influence, and also where it is likely to be active. The viscosity, or frictional adhesion, of water to the sur- CORROSIVE INFLUENCE OF SOILS, ETC. 105 face of a structure will vary according to the character of the water, and nature of the material used in a structure. Although water, under any probable pressures occurring in practice, is not increased in viscosity by the force it is pressed against a structure, its corrosive effect is greater. Some recent German experiments have shown that a mixture of 25 per cent, of common salt with water increased its viscosity at high pressures and temperatures. The viscosity of con- centrated aqueous solutions of sodium chloride and ammonium chloride are also increased approximately proportional to the pressure, and almost independently of temperature. It is well to bear in mind the influence of pressure and tempera- ture upon the viscosity of fluids in endeavouring to ascertain the probabilities of corrosive effect. In ordinary water, the dominant action upon iron is that of the oxygen of the air dissolved in water, for, although water is decomposed by contact with iron, it is chiefly the oxygen of water which causes the corrosion of iron, all water containing atmospheric air, the percentage varying. Water may act on any surface by percussion, by pressure, by re- action, and by chemical action. The chemical action of gases has to be resisted as effectually as the corrosive or solvent action of water. Mechanical agitation also facilitates solu- tion. Experiments have shown that falling water exerts a pressure due to a column of water of a height equal to the fall. Wind pressure, however, may increase or even diminish it. All dripping water from projecting parts should be kept from a structure. Pressure of water has an effect on the painted surfaces of a structure, for it may find out any weak places, as it does when it presses against scales of iron rust. With a pressure of 1 lb. or so, or that equal to about a depth of 2 feet 4 inches on the square inch, they may readily resist percolation of water, but it may not be so in 20 to 25 feet of water, when the pressure is about 9 lbs. to 11 lbs. per square inch, and water may ooze through. Constant pressure from still water, i.e., without any impactive force, cannot have such an effect on the paint as that of waves, the pressure of 106 CORROSION AND ITS PREVENTION. which on the side of a structure has been recorded as equal to about 50 lbs. per square inch. Some experiments have shown that the corrosive effect of active water in the open sea is some seven times more that of the same sea water when kept still. The different effects of still and moving water upon the rate of corrosion may be shown by examples, namely, that corrosion, except under peculiar circumstances, usually increases externally in ships upwards from a few feet above the keel, the worst places being situated near the wind and water line, and a few feet down, that is, where the plates are alternately wet with the sea water, dry and exposed to the air. Experience and experiments have proved that in water constantly aerated, or changed by currents, or disturbed, such as the surface waters of seas and rivers, which continually supply fresh quantities of active corrosive agents, both iron and steel become corroded much quicker than in still water. That the constant wash of bilge water in ships, when the internal plates are not properly protected, soon corrodes and destroys the heads of rivets, and grooves the plates between the frames on the path the water takes, and that main pipes in which water is constantly in motion become quickly rusty if not covered with an anti-corrosive preservative ; whereas in disused or seldom used branch pipes for firecocks, and houses to which the supply has been discontinued, although full of comparatively still water, and only occasionally subject to a flow, corrosion is comparatively insignificant. Here, also, may be mentioned that corrosion in boilers is usually the quickest in the stays at or about the water line, and in the steam space. These facts show that the more the water is aerated the greater will be the corrosion, other conditions being similar, and that the constant renewal of the water practically re- plenishes the surface with a new and incessantly energetic corrosive fluid. It is important this should be remembered, for it indicates some parts in a metallic structure which are more liable to corrosion than others, and therefore require to be especially protected and examined. Pressure of water, CORROSIVE INFLUENCE OF SOILS, ETC. 107 such as that caused by wave action, will also aid corrosion, and is by no means to be disregarded in submerged or occa- sionally submerged structures, for if scale is formed on the iron underneath the protecting coating, or if a weak place occurs in the covering, pressure will at length cause a per- foration of the paint or scale, and then water will percolate till it fills any unoccupied space, and will become compressed according to the head of water or the force of the waves, the result being that the scale or protecting coating is forced away. The scale, also, not having any elasticity to yield to the straining of a ship or submerged structure from the shock of the waves, cracks and becomes detached, and, even supposing the paint is sufficiently elastic to be unaffected by such sudden and repeated strain, the scale will ultimately cause the paint to crack, and then the pressure of the water will complete the disintegration, and the metal, or what is left of it, if in thin plates, will be bare and entirely unpro- tected, and be subject to influences peculiarly liable to destroy it. Various iron and steel structures subject to tidal immer- sion, both in sea and fresh water, which have come under the observation of the Author during the last quarter of a century, tend to show that in a properly protected and maintained bridge, landing, or promenade pier, the greatest corrosion is upwards from about low-water level, and that from a few feet below any wave action, although there is corrosion, it is much less than that above what may be called the line of still, or nearly still, water. In other words, in iron or steel work constantly submerged, corrosion pro- ceeds more slowly than when it is alternately immersed and exposed to the atmosphere, especially if the water frequently or regularly changes from being fresh to brackish ; the chief exception being in water impregnated with sewage, when the corrosion will be very active, and will continue so until a depth is reached in the ground to which the sewage water has not penetrated. Iron does not appear to corrode in ice, but whether oxida- tion already commenced is arrested is doubtful, although it 108 CORROSION AND ITS PREVENTION. appears to be much less active ; and so long as ice remains in water, and the temperature does not exceed 32° F. or cause wet ice, corrosion is not apparent, unless air is allowed to reach the metal. This is the case with ordinary hard water, such as London water. At about 60° or 70° F., however, corrosion becomes apparent. In an hour or two a little scum of greenish hue can be seen on the surface of the water, and in two or three hours it has a brownish colour at the bottom. After a few weeks, thin rust particles will be formed on the surface, and can generally be washed off with water, but if they are raised lumps, which are usually concentric, the plate will be found to be pitted, and serious corrosion has commenced, and loss of sectional area of the plate. Whether metals suffer from freezing of water in the pores has not been determined ; however, the expansive power of water is very great, and it is well to bear in mind that it expands by heat as well as by cold. It expands from 40° F. at a higher temperature till it becomes sufficiently hot to be converted into steam, or so cold as to become ice, when, by its conversion into solid crystals, which do not fit so closely as the particles of water did, it expands. Ocean currents, or marine rivers that flow from the high to the low beds of the sea, are believed to be promoted by differences in the density of the water, evaporation, rotation of the earth, character of the coast, rivers flowing into the sea, tides, winds, effects of climate, and other causes. Each will have its own corrosive influence, and it may vary con- siderably ; for instance, sea water, by eroding the cliffs or lands in the vicinity of a structure, will probably deposit slime of a similar character upon it. Experiments have shown that light penetrates both fresh and sea water to distances far beyond any upon which an engineering structure is likely to be erected, and that the actinic rays do, to at least 1000 feet in the open sea, and to about half that depth in lake waters, which are open all over their surface to the sun's rays ; so that if light is necessary for the promotion of any fouling, and consequent corrosion, it is present. 109 CHAPTER VIII. GALVANIC ACTION AND CORROSION. The influence of galvanic action in the promotion of corrosion in iron or steel, whether caused by immediate contact or by mediate contact, as sea water, wire, &c, is important. The direction of the current of electricity has been ruled to be, that the current passes from the metal subject to the greatest chemical action to the liquid, thence to the sound metal, and so forward, and therefore the direction of electricity in voltaic arrangements may be known, and the relative designations of positive and negative can be allocated. The decomposing effects of galvanic electricity on water are marked, the oxygen being liberated at the positive end of the battery, the hydrogen at the other; and although energetic decomposi- tion may be slow, consequent upon a bad conductor, such as water, being the conducting power, the latter becomes readily increased by the addition of a little acid or common salt, hence the accelerated decomposing effect of rain water in dense manufacturing districts, and of sea water on iron or steel under such conditions. It is well to bear in mind, in metallic structures, that voltaic action, as evinced by chemical decomposition, may ensue without metallic contact. Electro- lysis, i.e., the decomposition of a compound substance by the action of electricity or galvanism, is now generally accepted as a fact. In considering the effects of galvanic action, and the manner in which it chiefly arises in structures, it is im- practicable to state every condition in which it may occur, for one of the highest practical authorities has shown that " in 110 CORROSION AND ITS PREVENTION. one and the same piece of plate, galvanic action might come into play very decidedly in promoting the corrosion of the metal." Indirect or circuitous connection is sufficient to produce it, 'for instance, if the metals be in some corrosive liquid capable of corroding one of them. The importance of uniformity in the composition and quality of iron or steel has been previously referred to, and it may be taken that uneven corrosion may occur because of electric action gene- rated by oxide being crushed into the metal in the piles during rolling. When this electric condition exists, the effect is necessarily uneven in intensity, and generally variable according to the position and quantity of oxide that was on each pile of the bloom, or on the metal at the time of rolling. There can be no question that the composition and purity of the metal very much influences the decay or corro- sion, but the desired purity must not cause difficulty in working the metal, or reduce its tenacity or other powers. The insertion of a new sheet into a boiler, consequent upon a modification in the form, and not because of excessive or uneven corrosion of a plate, has been known to set up galvanic action so strongly that all the sheets became cor- roded except the new plate. This, by Dr. Cresson's experi- ments, was shown to be because of the different electro- condition of the sheets, the purer iron corroding and protecting that which contained the greatest amount of carbon, the new sheet of iron being electro-negative to the old plates, its position in the electro-chemical scale depending upon the amount of carbon it contained. Neither a black or any deposit should be allowed to remain on a metallic structure, as it will on railway over-bridges be a carbonaceous deposit electro-negative to the iron, causing galvanic action and therefore aiding corrosion, attracting moisture and all other corrosive influences. A similar deposit will occur in other smoky situations, as particles of soot or carbon rest upon an iron or steel surface, sulphurous acid being present, and moisture of the atmosphere; and any such covering will almost always have a corrosive tendency. GALVANIC ACTION AND COEROSION. Ill In order to prevent galvanic action in the same plate or structure, it is advisable to avoid any variation in the electro- chemical properties of the metal. As wrought iron, steel, and cast iron do not possess the same electro-chemical properties, it follows that galvanic action, varying in in- tensity between any two, may be set up by mediate or direct contact. Thus, if steel is joined to wrought iron, the decom- posing effect of galvanic action is in progress, and the iron may suffer ; or steel or wrought iron to cast iron, but especially in the case of steel, as then by experiment the effect is increased, and is stated to be as violent as when copper is made an element in the galvanic circuit in connection with wrought iron. In brief, the direct, indirect, or circuitous contact of all metals of different conducting powers should be avoided as much as possible, or a durable insulator be placed between them, for paint is but a kind of expedient or remedy liable to be rendered useless by decay, and if there happens to be a bare spot it will be quickly corroded, such action will be concentrated on the exposed place, and a bare surface may easily be caused by abrasion, or during the course of erecting a structure. The following electro-chemical scale,* arranged after Berzelius, according to their relative degrees of positive and negative, and negative electro-chemical character, is useful for reference. It commences with those substances possessing the strongest electro-positive properties, and ends with those of the strongest electro-negative properties. Al- though several substances appear in the scale that are seldom employed in engineering structures, they cannot be elimi- nated without disturbing the scale. Some consider sulphur and nitrogen to be less negative than either fluorine or chlorine, and that hydrogen should be towards the end of the positive division, but the general comparative value of the scale is acknowledged. It has, however, been pointed out that the division between gold and hydrogen is in great measure an arbitrary one, and the use of the table is in indicating the general electro-chemical character of the substances. * Vide ' Circle of the Sciences.' J 112 CORROSION AND ITS PREVENTION. Positive End. POTASSIUM. SODIUM. LITHIUM. BARIUM. STRONTIUM. CALCIUM. MAGNESIUM. GLUCINUM. YTTRIUM. ALUMINIUM. ZIRCONIUM. THORIUM. CADMIUM. MANGANESE. ZINC. IRON. NICKEL. COBALT. CERIUM. LEAD. TIN. BISMUTH. URANIUM. COPPER. SILVER. MERCURY. PALLADIUM. RHODIUM, PLATINUM. IRIDIUM. OSMIUM. GOLD. HYDROGEN. SILTCIUM. TITANIUM. TANTALUM. TELLURIUM. ANTIMONY. CARBON. BOKON. TUNGSTEN. MOLYBDENUM. VANADIUM. CHROMIUM. ARSENIC. PHOSPHORUS. SELENIUM. IODINE. BROMINE. CHLORINE. FLUORINE. NITROGEN. SULPHUR. OXYGEN. Negative End. For instance, sulphur and chlorine, two of the most negative of substances, must be viewed as positive in relation to oxygen. In fact, each substance throughout the scale may be viewed as both positive and negative : positive in relation to those below it, and negative in relation to those above it; those of the upper end being strongly positive and feebly negative, and those of the lower end strongly negative and weakly positive. The electric relations of metals, of course, vary if placed in certain solutions or mixtures, none of which are likely to be met with in ordinary engineering struc- tures. The thermic condition has an influence in the electric conductivity of metals, as their order is not the same at all temperatures. At 60° F. the order for voltaic electricity, beginning with those which conduct most freely, is : silver, copper, gold, cadmium, zinc, brass, tin, palladium, iron, steel, lead, platinum, German silver, antimony, mercury, bis- muth, potassium. By these two scales a fair indication of the effects of galvanic action, so far as regards the promotion of corrosion, may be generally gathered. An iron plate placed in sea water will corrode more quickly than a copper one, but if they are in contact the corrosion will be very greatly increased, owing to galvanic action, and the copper will appear to be able to resist corrosion. It will be noticed in the scale that iron is con- siderably higher in the positive end than copper. The accelerated decomposition GALVANIC ACTION AND COKROSION. 113 pf the iron is thus explained. In the case of iron or steel, if the water is able to corrode one of the metals, the oxidising medium is in active operation, and accelerated corrosion, the result of galvanic action, is proceedin g. There- fore an insulator is required to prevent the communication of electricity, but the importance of a complete insulating covering cannot be over- valued, because, if a bare spot should occur in an otherwise coated plate, galvanic action may- become concentrated on that alone, which will then become quickly corroded. Experiments have shown that carbon gives to steel some power of retaining magnetism, and that the permanent magnetism of steel is affected by the amount of carbon which it contains, therefore the strength of the current will vary, but the depth to which magnetism penetrates in iron or steel is a contested point. It exerts an influence on the corrosion of iron or steel. The decomposing effect, or accelerated corrosion, arising from galvanic action may be stated approximately to be according to the juxtaposition of metals, the effect being the greatest when they are in close contact. Every metal being electro-positive to its own oxide, when steel or wrought iron with the oxide scale on it is placed in an oxidising liquid, as salt or sea water, the medium is present to at once promote corrosion j further, it may occur between the metal of a plate and the oxide scale, and this has been proved to have been caused by the oxide formed in the early stages of manu- facture being crushed or pressed into the metal in the rolls, resulting in the metal having impurities in it inducing galvanic action. In fact, a plate with the oxide scale on, and one without, if placed under the condition of electric action, as in salt or sea water, will be an electric battery, and the plate with the oxide coating will produce the necessary decomposing electric condition. Carrying this further, it will be readily understood that even some alloys, composed of metals of varying electro-chemical character, may, in an oxidising medium, as salt or sea water, be in an i 114 COREOSION AND ITS PREVENTION. electrically decomposing condition, inducing internal action with deleterious effect. Usually, alloys are harder than their components. It will be noticed that carbon is placed six- teenth from the negative end, and iron sixteenth from the positive end, and, therefore, carbon is decidedly electro-nega- tive to iron ; consequently accelerated corrosion by galvanic action will be set up by direct or mediate contact. In order to lessen galvanic action as much as possible in engineering structures, the accelerated corrosion so arising should ob- viously be guarded against by providing that, as far as practicable, different metals should not be placed in close union or in close contact, and that the metal is well covered with non-oxidising paint, cement, or an insulating coating, for, by the junction of such metals as steel with wrought iron, cast iron with steel or wrought iron, electrical action is greater than with metals whose electro-chemical properties more nearly approach, the main point being to attain as much similarity as possible. Local galvanism has also to be considered, for the complex composition of steels causes an interchange of the electro-chemical positions, and "it may be inferred that the crystalline formations of their structure nearer the surface are at first more readily acted upon by sea water, until inner and seemingly less vulnerable portions have been reached." When there is a variation in the condition of a mass of iron or steel, one having less affinity for oxygen than the other, contact of the former makes the latter become oxidised more rapidly. The positive metal, being more rapidly oxidised and acted upon than the negative metal, also militates against an equi- librium being maintained, and aids the alternate variation of the electric position. The reliable recent experiments of Mr. T. Andrews, F.E.S., M. Inst. C.E., of the Wortley Iron- works, extending over 300 days, on steels and irons in galvanic action immersed in clear sea water, have shown that a complete interchange of electro-chemical position occurs after very considerable intervals, affecting the rates of corrosion, and that it is doubtful whether a permanent GALVANIC ACTION AND CORROSION. 115 position of rest finally ensues between the metals, and the probabilities are against it, though eventually the galvanic action becomes very small from the accumulation and possible influences of the oxides. He stated in a paper communicated to the Institution of Civil Engineers, that the conditions attending galvanic action between wrought iron, cast metals, and various steels, in. sea water, is important, as the accurate measurement of the electro-chemical positions of the metals when in combination should afford an indica- tion of their relative corrosibility under such circumstances in sea water ; also, in a series of most carefully conducted and reliable experiments,* he "arrived at the following conclusions, namely: that wrought iron and various steels, when exposed singly and separately, without liability to galvanic action other than local, to the action of sea water for long periods, showed a greater corrosion on the part of all the steels than the wrought iron, the advantage in favour of the wrought iron as compared with the steels amounting roughly to 25 per cent, and upwards. It was also noticed that corrosion was increased in the steels in proportion as the percentage of combined carbon was greater. Further, it has been found that the galvanic action between wrought iron and steels induced a largely increased total corrosion in the several metals." Also, " that the corrosion of metals is considerably affected by stress, varying according to the nature and extent of the strain applied. It might have been thought that metals under stress would be more liable to increased corrosion than when in their normal state. His recent experiments, however, indicate the opposite conclusion. That is, when ' strained ' is considered separately from 'unstrained' metal. When, however, the strained metal is in galvanic circuit or combination with the unstrained metal in any solution, an increased total corrosion ensues from the galvanic action which this research has shown to arise consequent upon the difference of potential between the two. The reason is that stress, whether tensile, * ' Minutes of Proceedings,' Inst. C.E., vol. cxviii. i 2 116 CORROSION AND ITS PREVENTION. flexional, torsional, or of any other kind, considerably alters the physical properties of iron and steel. Stress increases the rigidity of both iron and steel and renders the metal harder, also greatly reducing its properties of elongation or ductility. A higher tonnage is required to break a ' strained ' than an ' unstrained ' portion of the same metal. A tensile stress applied to a wrought-iron shaft, producing an elongation of only 2 per cent., increased the tensile resistance of the metal 2-66 per cent. Other investigators have also noticed a similar alteration in the properties of metals referable to stress. From these observations it is manifest that the stresses applied to the metals examined for corrosion, altered their structure, render them harder in nature, i.e., the metals showed a higher tensile strength subsequent to the strain than previously, and hence less liable in the strained con- dition to be acted upon by sea water, or other waters, than in their normal or softer condition. The experiments, however, indicate that an increased total corrosion, in excess of the normal corrosibility of the metal, occurs in a metallic bridge, vessel, boiler, or other structure, from the action of the local galvanic currents which are hereby shown to be induced between ' strained ' and ' unstrained ' portions of even the same piece of iron or steel forging, bar, or plate. Hence a strain occurring in a metallic structure tends, owing to local galvanic action thus set up, to increase any corrosive forces which may be deteriorating the metal of which it is composed." Thus to the fatigue of metals by stress may have to be added that of galvanic action aiding corrosion set up between the strained and unstrained parts of a metallic structure, especially when the strain is locally severe or of an uneven character. Mr. Andrews's further valuable experiments to ascertain whether the normal corrosibility of metals might be consider- ably affected under the influence of strain, are novel and useful. They were made on small round bars. Plate inch in thickness. All metals were perfectly bright. As GALVANIC ACTION AND CORKOSION. 117 regards tensile stress, it appeared the wwstrained metal was "being more rapidly corroded than the strained metal, which latter had in most cases been strained so as to produce an elongation of 20 per cent, between gauge points 3 inches apart. In the case of torsional stress similar results occurred, I viz. the ^restrained metal was being corroded the more. "When the stress was flexional, similar results occurred. In all cases the ^restrained steels were more electro-positive than the strained steels, and, therefore, were more acted upon [by sea water. Mr. Andrews states that " the electric measurements ought perhaps to be regarded as tentative indications establishing a general principle." In the Journal of the Franklin Institute, Prof. Munroe, U.S.N.A., states that two steel chisels found in the U.S.S. ; Triana, were deeply corroded by the action of the salt water, : but the corrosion was confined entirely to the soft metal, the I tempered points not being attacked in the least ; the corro-> sion being deepest at the line of contact between the tempered and untempered metal. Whether the change which takes place in the tempering of steel is a chemical or a physical one it is not attempted to be shown, but it is considered the tempered steel is not so readily acted upon by salt water as reretenipered steel, and it is suggested that when the untempered and tempered steels are brought into contact in the presence of salt water, they constitute an electro- chemical couple, and that this hastens the destruction of the rentempered metal. Among some instances of galvanic action accelerating corrosion, additional to the familiar examples of iron and copper in contact in sea water, and the base of iron railings when fixed with lead run in a hole in stone, instead of by Portland cement, or in glycerine and litharge stirred to a paste, which hardens quickly, and will cement iron in stone ; lor in sulphur, which has been recommended to be preferred to lead for preserving iron bars or staples that have to be fixed in stonework ; may be mentioned that Mr. Hyslop stated at a meeting of the Institution of Engineers and 118 CORROSION AND ITS PREVENTION. Shipbuilders in Scotland, that he had connected the boilers of a steamer with a plate of zinc immersed in the sea, and, by the use of a galvanometer, he was able to observe the effect galvanic action produced at every lurch which the vessel received from the motion of the waves. Mr. Gilchrist, then President, said that galvanic action was set up in vessels having different metals used in the construction of their parts, was a very patent fact ; for if a cast-iron screw were used on board a wooden vessel coated with copper it would, disappear in twelve months, falling to the bottom of the sea. Such vessels have brass instead of cast-iron screws. It has also been recorded that the replacing of the iron screw blades of a steamer by others of manganese-bronze, caused such corrosion of the steel and iron in the stern posts and surrounding parts, during one voyage between London and the Cape of Good Hope, that they had to be entirely renewed, and had now to be carefully protected by zinc. But, perhaps, one of the most striking instances of galvanic action is that related by Mr. John Donaldson many years ago. He then said the practice of his firm with steel plates had been to preserve, as far as possible, the oxide on the plates, with a view to prevent them from getting rusted, the oxide itself being considered to protect the plate covered by it from rust. Favourable accounts were at first received of a boat, the plates being as described, but one day she made so much water as to be in danger of sinking ; the boat was then examined, and it was found that the plates in the bow and in the stern, and others in the boiler and engine room, were pitted with very small holes. Some of the black oxide had been knocked off in the process of working the plates, and the oxide left on had contributed to increase the rusting of other parts. It was the black oxide that had been acting galvanically in oxidising those parts of the plates that were exposed. When the plates of a boat were well painted and kept clean, very little of such action took place. Their practice after that experience was not only to remove the scale from the plates, but to galvanise the whole of the hulls, and since that had been done there had been no trouble as GALVANIC ACTION AND COKROSION. 119 far as oxidation was concerned. Another instance of gal- vanic action, for there was no other way of accounting for it, that came before the author many years ago, was that of a small steel plated vessel which was riveted with iron rivets, the result being that the steel plate around the rivet-heads became sufficiently corroded to cause the heads and a small portion of the iron rivets to project beyond the surface of the corroded steel plates. The grooving of propeller shafts is considered to be generally due to galvanic action between the steel shaft and the sleeves forming the shaft journals in the stern tube. This has been proved to be the case, as the galvanic action has been prevented when the shaft has been covered with india-rubber or with a few turns of marline near the brass sleeves. Professor Lewis has demonstrated by experiment that of two steel wires dipping in sea water, the chemical action on one, which was kept in a current of hot moist air, caused it to be strongly electro-positive to the other. It is considered that similar chemical action would sometimes occur in hot moist steam. It has been found that while wrought-iron pipes taking condensed water from iron steam pipes connected with the same boilers showed no signs of rapid corrosion, those conveying the same water from a copper steam pipe were corroded through, and required to be renewed in a short time. This was probably due to galvanic action kept up by the constant circulation of steam and feed water between the two. Some cast iron coming from the blast furnaces of the South Oural Mountains in Eussia was, on analysis, found to have the following composition : — Per cent. Iron 83-5 Copper 8 • 2 Tin 1-3 Cobalt 0-5 Silicium 1*0 Tungsten 0-1 Carbon 3*0 Manganese 2-4 100-0 120 CORROSION AND ITS PREVENTION. Under the microscope, on the metal being fresh cut, small grains of copper were easily remarked in the mass of the metal.* In such a metal galvanic action is probable, and it is to be regarded as one liable to accelerated corrosion, because the electro-chemical properties of iron, copper, tin are different. Electric railway return currents have been observed to cause corrosion of adjoining water and gas pipes. A paper was read by Professor Jackson f on the subject, in which it appears that experiments were made at Boston, U.S.A., by connecting the reinforcing wire laid between the tracks to the water pipes, but the supplementary wire was destroyed in several places; then the polarity of the generators was reversed, the current being sent out through the rails, and back through the overhead trolley wire, but this current sent through the rails took to the water pipes and the lead cable coverings with disastrous results, following the law of divided currents, and leaving them at many points along the line, causing serious corrosion at these places. On the direction of the current being reversed, so great a current flowed along the water pipes that, at a joint where oakum was used for caulking, it was sufficient to set fire to the oakum. The loss of pressure on the return circuit was from 25 to 100 volts, or from 5 to 20 per cent, of the total pressure. On the water pipes being experimentally connected with the negative polo of the dynamo, a new danger occurred, for the difference of potential between the gas pipes and the water pipes caused a marked electrolytic effect on the former. By connecting the gas and water pipes together in all parts of the city, to arrest the action, fair results were obtained, but the expense to the city and to the company was great, and finally far from satisfactory to either party. In Brooklyn, where there are many electric railways, numerous cases of serious corrosion had occurred, and the * See < Iron,' Sept. 2, 1876. t See 'Journal of the Association of Engineering Societies,' Phila- delphia, 1894. GALVANIC ACTION AND CORROSION. 121 Board of Commissioners of Electric Subways of Brooklyn reported, in 1893, " that all kinds of buried pipes are being eaten away in many places. As an example, an iron service pipe, buried at a depth of 4 feet below the track, had been completely perforated in a month. At Milwaukee, the electrical engineer to the Street Bailway Company reported, in December 1893, that at 200 feet from the power-house, a 6-inch water-main was so badly corroded, after the electric railway had been at work four years, as to render it entirely useless, and when taken out of the ground it was so soft in some places that a cane could easily be poked through it. The corrosion was arrested at Milwaukee by making numerous low-pressure connections between the pipes and the rails, thus keeping the two at the same potential, and by connecting both pipes and rails at the power-station to the negative pole of the generator. As much as 28 per cent, of the total output is now found to be returned by means of the pipes, and this plan has been working satisfactorily for more than a year. At Chicago, Professor Barrett reported, that in some experimental work, in which a current of 0*3 ampere continued for three weeks, it was most destructive to a lead telephone cable, while another, buried in the same soil, which was not subject to the action of the current, was unaffected. In Janesville, Ohio, a 4-inch cast-iron water pipe was completely perforated in two years. At Columbus, Ohio ; Hamilton, Ontario ; Indianapolis ; Philadelphia ; Los Angeles, California; and many other cities, where consider- able electric railway systems are in operation, in every case the corrosion has exhibited the same general features. The iron pipes are usually " pitted " in many places. It is now practically agreed that the reason for the extra- ordinary corrosion is the imperfect character of the return circuit of electric railways. As first constructed, the rails in connection with the surrounding earth were relied upon to carry all the current back to the generator. It was soon found, however, that the current would not confine itself to this path, and that the resistance of the earth was far from 122 CORROSION AND ITS PREVENTION. being as low as was originally supposed. Bending the rails, cross-bending, supplementary wires, and ground plates, were then tried, but have not answered, and the tendency now is to make the return circuit of fully as great conductivity as that of the overhead supply circuit, without relying upon any conductivity from the ground. There is little doubt that with a perfect return system, properly connected to systems of underground pipes, electrolytic disturbances will practically disappear in nearly all cities. Two theories were put forward to account for the corrosion : (1) that it is simply due to chemical action caused by ammonia, saltpetre, leakage from gas mains, &c, found in the earth ; and (2) that it is the result of electrolytic action. Where chemical action alone occurs, water and gas pipes ordinarily last about twenty years or more, but this corrosive action has destroyed new pipes in intervals of from a few weeks to six years, and it has taken place where the electric current left the pipes. This was conclusive proof of elec- trolytic action, and the only question is, whether it occurs, (1) by direct electrolysis of iron, or (2) by the electrolysis of chemical compounds which are held in the water of the soil, causing secondary chemical reactions at the electrodes. A series of laboratory experiments, in which the practical conditions were produced as fully as possible, showed that the latter was the true solution of the problem, the iron pipe acting as the positive plate of a cell ; the water of the soil, with chemical compounds in solution, as the electroly tex, and the rail as the kathode, or negative pole. Analyses of street soils usually show the presence of some soluble salts of ammonia, potash and soda. Experiments were made with such salt solutions, the result being that it was seen only such measures as will prevent electrolytic action of salts in solution in the soil can be relied upon to stop the corrosion of the iron pipes. Professor Jackson concluded by stating the solution of the matter is the proper construction and arrangement of the return currents. Alter- nating currents produce no appreciable electrolysis, and their GALVANIC ACTION AND COREOSION. 123 employment would avoid all difficulty, "but their use for driving street railway motors is not yet an assured success. Connecting the pipes with the rails by means of heavy cables at points where the former are positive to the latter, has hitherto proved the most effective method of prevention. The conductivity of the track circuit must be properly reinforced by feeders, so that an undue drop may not be experienced in the return conductors, and these track feeders should always be insulated and put on the line exactly as are overhead feeders, in order to save them from corrosion. In Milwaukee, with 125 miles of track, and over 200 cars in daily operation, at a cost of about 1600Z., such a connection of pipes and rails has apparently succeeded perfectly. Inves- tigations have shown that when the negative pole of the generator is connected with the trolley, the pipes are positive to the rails over an extended outlying district, and corrosion proceeds over a large area ; while with the reversed arrange- ment, the dangerous area is concentrated about the power- station. The latter method of connection allows the difficulty to be most easily dealt with by making frequent connections of pipes and rails within the affected area. 124 CORKOSION AND ITS PEEVENTION. CHAPTER IX. THE INFLUENCE OP THE SCALE ON CAST IRON, WROUGHT IRON, -AND STEEL, AS REGARDS CORROSION. As there is scale or skin on cast iron, wrought iron and steel when manufactured, the question arises whether it should be removed, or be regarded as a natural preservative. Ex- tensive use seems to show that the natural covering of cast iron as it leaves the moulds should be preserved, although the appearance is not so good j and that cast iron is more durable if the skin is allowed to remain than if the metal is planed or manipulated, as it is generally found to be a film of silicate of the protoxide of iron, and is a protection. As silica is not quickly acted upon by salt water, in all cast-iron work that will be submerged and cannot be periodically inspected, the preservation of the natural skin of cast iron which it obtains when run into a sand mould appears to be advantageous, particularly if it is to be immersed in salt water, and it is a protection which, with care, can be pre- served ; but silica is only an apparently insoluble substance in water, for Mr. Crookes, F.E.S., and Professor Odling found, in experimenting for the Huddersfield Corporation in 1882-4, that silica dissolved in water to an appreciable extent, from to T 7 ^ of a grain being found in a gallon of water. It is important that the silicate scale on cast iron, which is harder than the metal it covers, being fused sand or loam on the molten metal, should be coated with oil or paint soon after casting, and before any oxidation has occurred ; as unless it is protected the metal underneath it will become exposed as the silicious film may fall away. If oxidation of the fused INFLUENCE OF THE SCALE ON CAST IRON, ETC. 125 skin be allowed, unless the rust is carefully removed, paint spread on its surface will in time become detached. When any portion of the scale of cast iron is removed by manipula- tion, turning or planing, vulnerable places for active corrosion are produced. The skin on wrought iron is a chemical combination of iron with oxygen, the proportion of the latter increasing as the iron is exposed to the air. It will fall away sooner or later. With regard to the scale on wrought iron or steel, differences in the process of manufacture and in the nature of the metal cause it to be necessary to remove it in order to effectually prevent corrosion, or paint the metal, and it is decidedly injurious to wrought iron or steel. Iron and steel not being manufactured in the same way, their surfaces are not alike. Steel cast in an iron ingot mould, and cast iron in a sand or loam mould, have dissimilar substances in contact with them during manufacture, and steel which is reheated and passed through a rolling mill has a different kind of scale to wrought iron. Any attempt to preserve the entire skin of wrought iron as it leaves the rolling mill can only end in leaving some portions covered and others bare, cause accelerated and concentrated corrosion in the bare places, and result in ultimate failure, for as iron in passing through the rolls is exposed to the air, oxidation really com- mences, in however small a degree, in passing the rolls, and any such film of oxide is unstable, and may flake and be- come detached, whether it is covered with paint or not. ^ In the case of wrought iron, the scale will fall off at some time or other, depending upon the magnitude of the rusting that took place during the process of manufacture and upon the homogeneousness and general quality of the metal. If a complete and equable covering could be so obtained and maintained it would form a protection, and most probably an air-tight coating. In practice, however, corrosion will continue with any such scale, and the process of oxidation will rather be accelerated than retarded by it. In fact, quite distinct from their chemical composition, the skins COEEOSION AND ITS PEEVENTION. of cast iron, wrought iron, or steel, are dissimilar, because of the difference in the process of manufacture, and the latter requires to be considered, no matter what method of production may be adopted. That galvanic action is set up, when a rolled plate is immersed in sea water, between the black oxide scale and the iron, may be considered as agreed ; but, if a bare plate were immersed in sea water, i.e., one with the oxide scale removed, and one with it remaining, the oxide scale would form a kind of temporary shield. However, the black oxide scale is not permanent, and therefore although, for a short time, it may help to protect the surface, when air permeates to the iron, the scale gradually flakes or falls off with any paint that may have been spread upon it. If the surface of the iron, when so bared, was properly prepared by rubbing or other approved process, it would be preserved, but only when the black oxide scale had been entirely removed, consequently it is far better to remove the scale at first, before painting, than to wait till it falls away with the paint, which would be when a structure is in use, and then, probably, there will be places almost inaccessible to painting and thorough inspection, and the work cannot be done so easily or so effectually. The ordinary cheap market wrought iron has a thicker scale than the higher brands, and may have laminations and perhaps cinder in it, but the cinder in iron is generally regarded as dielectric; however, if it be a carbonaceous deposit it may act deleteriously in another way, namely, by becoming a nucleus in which moisture and acid may be retained and become condensed. In the case of wrought iron, careful consideration confirms the view that it is the best plan to remove any thin scale that may be upon the surface, for it cannot be a preservative ; it in no way adds to the strength, and, moreover, is a cause of accelerated corrosion. This requirement is no mere mechanical refinement, and,' in course of time, may become the universal practice, for metallic structures generally have hardly been introduced INFLUENCE OF THE SCALE ON CAST IKON, ETC. 127 sufficiently long to cause marked attention to be given to corrosion in its various aspects, but every day must increase it; and as iron structures, especially bridges, have to be removed or altered to comply with increasing traffic, ocular demonstration may prove that more attention and special precautions to prevent corrosion may be necessary. With respect to the skin or scale on steel plates, experience shows that the hard black oxide scale on steel is very deleterious in its effects, and more seriously so than in wrought iron. In fact, it approaches that of copper, and on high authority it has been stated, " a tolerably compact coating of oxide is as detrimental to steel exposed with the oxide on as copper." Dr. Siemens said it was of much importance to perfectly clean the surfaces of plates, for if they were carefully cleaned of oxide by dipping in the first instance in an acid solution, it was found corrosion was always much lessened, the reason being that the scale on steel which was produced in rolling, being a magnetic oxide, negative to the steel, had a very deteriorating influence on it ; or if any scale were rolled into steel plates, as occasionally occurred, when the metal was exposed in the presence of such magnetic oxide, corrosion rapidly followed. By removing this skin in steel, this cause of rapid corrosion is prevented. The Admiralty require the magnetic oxide scale to be removed from steel plates, for it was found that unless it was completely cleared, rust formed under the paint when carefully spread on such scale, and also that unless the surfaces of steel plates were carefully cleared of the black oxide produced in the rolls, corrosion was not uniform, but when the scale was removed corrosion was very uniform in salt water. If the surface oxide is not removed, the effect of the oxide in the neighbouring bared metal is as strong and continuous as copper can be. In some locomotive works, the scale of all steel plates is removed either by bathing them in or brushing them over with a solution of sal-ammoniac. It is not only found that this much reduces the tendency to corrosion in the boilers, but that more secure and water-tight joints can then be made. 128 COKEOSION AND ITS PKEVENTTON. This is a decided advantage, and applies equally to girder- work and work made up of plates, for the closer the plates are together, and the tighter the joints and edges, so will active corrosive influences be hindered from coming into action. It has been asserted that if the magnetic oxide scale on a steel plate could be preserved and made to adhere to the solid metal it would be and is a protection against corrosion, but in practice, either from blows or in putting the parts together, patches fall off, even if they do not when untouched, and leave the paint bare, and even do so when the scale has been painted, and so centres of corrosion are formed. For these reasons alone, and apart from the question whether oxide scale may be a protection or not under certain circumstances, the better and safer plan is to remove it by pickling or brushing ; if not, blows, vibration, or attrition may do so and expose a plate. How can the scale or the film of oxide on wrought iron or steel be removed ? By the metal prepared for coating being dipped in a bath of dilute acid, or by a wash of weak acid, and afterwards by careful neutralisation, washing and dry- ing; but this process is not so complete and certain as planing, filing or grinding, the latter two methods being almost confined to test pieces. On account of expense, pickling the plates, and so removing the oxide, is the way in which, in ordinary engineering structures, the thin skin can be removed at trifling cost. Some experiments made by the Dutch Government with respect to the preservation of plate iron in railway bridges gave the following results. Sixteen wrought-iron plates were cleaned by immersion for twenty-four hours in hydrochloric acid, they were neutralised with milk of lime, washed with hot water, and, while warm, dried, and washed with oil, and coated as described. Sixteen were merely cleaned mechanically, by scratching and brush- ing ; four sheets of each kind were then exposed to the weather, and examined after three years. The following INFLUENCE OF THE SCALE ON CAST IRON, ETC. 129 were the results of these experiments : (1) The red lead was perfect on both sides, so that it could not be said whether the chemical cleaning was any use. (2) One kind of iron* oxide red paint gave better results on the chemically-pre- pared plates than on the others ; but another kind of iron- oxide red paint gave not very good results on the plates only scratched and brushed. (3) The coal-tar was considerably worse than the paint, and had even entirely disappeared from the iron sheets which had not been treated chemically, but only been cleaned by brushing. In Holland, especially, great care is taken to remove the scale before painting. Some Prussian experiments were recently made on the effect of pickling and rusting on the strength of iron ; and on steel rails of the State line section ; wrought-iron joists, 9^ inches by 3 \ inches; wrought-iron round bars, 0*8 inch diameter; and steel round bars, 0 • 8 inch diameter. The pickling was done with sulphuric acid, diluted with water in the proportion of 1 to 100. The metal was protected from direct attack by covering one end with melted zinc. The time of immersion was seventeen hours. After pickling, some pieces were immersed in lime water for many hours, so as to remove all traces of acid. The results were considered to show that, as regards chemical composition, the susceptibility of iron to become brittle by pickling and rusting is least in cast iron and silicon steel, and highest in wrought iron, and, according to Badecker, in high-carbon steel. Combined carbon appears to increase the action, and silicon to diminish it. The pieces were treated in the condition as delivered, after exposure for the purpose of rusting ; galvanised and tested at once, galvanised and exposed for a time, pickled in acid and immediately tested, and pickled and kept for a time in a dry place. It was thought, from some results of the experiments, that acid sets up chemical changes in wrought iron which it does not in steel ; however, the steel treated was exceptionally rich in silicon, no doubt adding to its resisting power. These ex- periments are referred to to record them here, and not to K 130 CORROSION AND ITS PREVENTION. consider them as conclusive, nor are they so claimed, although they are well worth attention.* The effect of pickling iron wire in water acidulated with sulphuric acid, in order to remove scale, has been shown by Prof. Hughes and others to be to make it brittle.f It did not affect its tensile strength, so long as the metal was not sen- sibly corroded, but the extension under stress and the capacity of resisting bending strains were notably diminished ; and an action similar to that of weak acid is produced by the atmo- sphere when iron and moderately soft steel are exposed to it in an unprotected condition. Contact of the iron with zinc, which renders the former electro-negative, increases the influence of the acid. Brittleness produced by pickling or rusting is removed by annealing, and also disappears, or is considerably diminished, by allowing the brittle metal to rest for some time in a dry place. Cast iron is not sensibly, or only very slightly, affected by pickling. The pickling was in 1 per cent, sulphuric acid water for twenty-four hours without previous cleaning ; tests after three days. By pickling in 2\ per cent, acid for twenty-three hours after cleaning the surfaces with ether. The mechanical tests were applied immediately. These experiments show the important effect of sulphuric acid on iron and steel. Pickling in acid, then carefully neutralising it, washing and drying, may be considered as the most practical, suf- ficiently trustworthy, effective, and cheap process yet dis- covered of removing the scale. * Vide ' Mitthoilungen aus den Koeniglichen technischen Versuch- sanstalten zu Berlin, 1890, Supplement J. t ' Stahl und Eisen,' vol. vii. 131 CHAPTEE X. THE SERVICEABLE LIFE OF METALLIC STRUCTURES, AND SOME EXAMPLES OF CORROSION. The serviceable life of iron or steel cannot be determined, although, by deduction, it may be approximately estimated. It would be useless and misleading to attempt to establish any rule, the circumstances and conditions being so dissimilar. The corrosive influences to which each metallic structure is subject, and also the other conditions, will affect the decay and durability of the metal. Metallic bridges are of recent origin. The first erected in England was Pritchard's celebrated cast-iron arched-ribbed Coalbrook Dale Bridge, over the river Severn, constructed about 116 years ago, which was followed by Burdon's cast- iron arched-ribbed bridge over the Wear, at Sunderland, some ten years later. Similar bridges were then built by Telford, over the Severn, at Buildwas, Shropshire ; by Eennie, over the Witham, at Boston, Lincolnshire ; and, in 1815-19, Sir John Eennie erected his masterpiece, the Southwark cast- iron ribbed bridge over the Thames, at London. The first boat made of malleable iron plates is generally acknowledged to be that constructed by Mr. Manby in 1820-21, although at Broseley, Shropshire, John Wilkinson, ironmaster, is said to have launched the first iron boat, but the system of con- struction is not clearly stated. Several vessels were subse- quently built of iron plates, by Messrs. Maudslay and Field, about the years 1830-31, and by 1846 about one hundred such British-owned vessels were afloat. Compound bridges, i.e., those constructed of cast-iron girders and wrought-iron k 2 132 COEROSION AND ITS PREVENTION. bars to truss them, received their doom in the failure of a bridge over the river Dee in 1847. With regard to wrought- iron girders : the first employment of wrought iron in the present form of plates and angle-irons was by the Messrs. Fairbairn, about the year 1832, and, with the exception of a few suspension bridges, malleable iron bridges were hardly known in 1837. It was not till Sir Wm. Fairbairn, in 1846, erected for the late Mr. Vignoles, Past-President Inst. C.E., the wrought-iron tubular girder bridge over the Liverpool and Leeds Canal, for the Blackburn and Bolton Kailway, that wrought iron may be said to have commenced to be practi- cally applied on a considerable scale to such important engineering structures as girders. Its practical application in the form of plates riveted to angle or T-irons, &c, may now be said to be about sixty-three years old. We thus have evidence that, with due care, wnsubmerged and properly coated cast iron in arched bridges may be said to be practically not deteriorated from visible corrosion after 116 years, and wrought-iron plate bridges after about fifty years, taking the first very large wrought-iron plate bridge, that of Eobert Stephenson's Britannia Bridge, erected in 1847-50. The application of steel plates to girders, ships and important structures, is quite of recent origin. Cast-iron sheet piles, whether in cofferdams or wharf walls, may be said to have been used since about 1815, and largely between the years 1820 and 1834; and solid, hollow, and screw, cast and wrought-iron piles extensively during about the last fifty years. With the exception of a few small pieces of ironwork embedded in mortar, occasionally found in some ancient building or ruin, or in a wall, or an old gate here and there, very little direct and unquestionable proof exists absolutely testifying that any ironwork used in the form adopted in modern engineering structures, and similarly exposed to corrosive influences, will last 200 or 300 years or more. In order to show how the serviceable life of metal in a form largely used can be prolonged, it is but necessary to mention that, during the construction of the Britannia LIFE OF METALLIC STRUCTURES, ETC. 133 Bridge over the Menai Straits, some rejected plates were left entirely unprotected upon a wooden platform exposed to the wash and spray of the sea. In two years after, they had become so corroded that they could be swept away with a broom. The plates were by no means thin, as they varied from y~s to f of an inch in thickness. The late Mr. W. Baker, when chief engineer of the London and North-Western Railway, found the corrosion which had taken place in the plates of the Britannia Bridge, exposed to the smoke and fumes of the locomotives, all being carefully painted and main- tained, would give no less than 1200 years as the time before the plates would be entirely corroded ; this would be 1 2 2 ° 0 = 600 times more than that of the rejected bared plates left on a stage on the shore as described. Further, with proper precautions Mr. Baker and Mr. Ramsbottom " could not give any practical limit to the endurance of the Britannia and Conway tubes." These facts are here mentioned to show that in very great measure the serviceable life of a metallic structure, so far as corrosion is concerned, can be prolonged to almost any practical limit by constant care of the surfaces, and repainting with an anti-corrosive coating, containing nothing that will produce or induce any deleterious action in or upon the metal. If bridges, roofs, promenade and landing piers, and similar structures, had the care bestowed upon them that iron and steel ships receive, their serviceable life would be much prolonged. It cannot be too much borne in mind that corrosion is very much accelerated by allowing it free action for a time ; in fact, reliable experiments have shown that the corrosion of plates of cast iron, wrought iron and steel, especially the steel, rapidly increases with time ; for it has been found in some experiments that the second year's corrosion compared with the first year's was 50 per cent, greater. This accelerated corrosion can be almost entirely prevented by reasonable care and attention, and at little expense. It is, therefore, well to remember that, before the introduction of 134 COEEOSION AND ITS PEEVENTION. plates of various forms for the floors of bridges, the necessity of renewing the floor planking or timber platform at intervals caused more frequent uncovering of the main and cross girders than is now requisite. Bridges of large span may be said to be constantly watched and maintained, and it is not to them especial reference is here made. The service- able life of iron and steel bridges, roofs, columnar or pile structures, may be much reduced by inattention to their preservation from corrosion, complication of the parts, badly- fitting rivets, rough butt-joints, the main girders being of loosely built-up plates ; also from members of a structure being of such shape, or so, when fitted together, as to be gatherers of moisture and dust, and receptacles for the promotion of decay and decomposition. In bridges, roofs, and unsubmerged structures, the extent of appreciable corrosion chiefly depends upon the attention given to them, the removal of the oxide scale before painting, the complete covering of the metal with an anti-corrosive and insulating coating, preferably possessing the power of hardening under water, the quality of the material, its freedom from galvanic action, and the nature of the atmo- sphere and climate in which the structures are erected. In fact, their serviceable life is governed by the care bestowed upon them in keeping them clean, and entirely and equally coated, which necessitates frequent examination, and renewal of the protection of the surfaces. Then, and then only, may their serviceable life, so far as regards corrosion, be con- sidered as long-continued. In the case of submerged structures, the circumstances are different, for, except by divers, or by the use of compressed-air apparatus in some form or other, they cannot be inspected, nor can they be repainted in water, and the subterranean portion cannot be examined. The duration of the parts of a structure which are either constantly submerged or buried in the earth, as in the case of piles or columns, can only be deductively estimated from the behaviour of similar works subject to like condi- tions and circumstances. An economical method of testing LIFE OF METALLIC STKUCTUEES, ETC. 135 Whether the metal in a pier has corroded to a serious extent below low water, is to suspend, when a structure is erected, a few short lengths of piles of the same dimensions, form, &c, under precisely similar circumstances to those of the permanent columns, taking especial care that no galvanic action is set up. These lengths can be secured to the per- manent work by an insulating medium at the shore and sea ends, and middle of the pier, or where desired or thought advisable, and a few piles can also be screwed or other- wise inserted in the ground. They can then be removed at any time for testing or inspection, and be replaced without interfering with the permanent structure, and will indicate the probable condition of the submerged and buried piles. In a paper read before the American Society of Civil Engineers, by Mr. John D. Van Buren, it was stated, " Sub- merged bolts and other wrought-iron parts are generally badly corroded in less than twenty years ; the corrosion taking place in lines parallel with the fibres. Certain kinds of cast iron could, perhaps, be made to last fifty years; but already (1875) authentic rumours are afloat that our — i.e., American — cast-iron lighthouses are materially injured by corrosion ; fifty years is a generous allowance, and probably greatly exceeds the average life of cast iron in salt water." The cast-iron cannons of the Boyal George, which went down in 1782, and the Boyal Edgar, were found to be quite soft, and like plumbago in some cases. They had been immersed for 62 years and 133 years respectively. The original cast-iron sluice-gates of the Caledonian Canal also corroded seriously, and had to be replaced ; on the other hand, the pier at Milton-on-Thames, erected in 1842-44, designed by Mr. Eedman, was supported upon cast-iron columns 3 feet in diameter and l£ inch in thickness ; and Gravesend Town Pier, designed by Mr. Tierney Clark, erected about the same time, and l£ inch in thickness ; and the Maplin Sands Light- house, also constructed about the same time, of f-inch metal only, appear to be unaffected. Sir J. Brunlees remarked, some sixteen years after the Morecambe Bay Viaduct 136 CORROSION AND ITS PREVENTION. had been erected, that there was no sensible decay in the columns. The whole of the cast-iron work was protected with a preparation of tar and asphalt. The tar was heated in a large tank, and each piece of ironwork was boiled in it for half an hour, as he believed it was the best preparation for coating cast iron. An examination was made by the chief engineer to the Bombay, Baroda and Central India Railway Company, of the condition of the piles of the South Bassein Bridge,* some of which had been fixed 25 years, and exposed during this period to the action of sea water. Specimens were cut from each pile that was considered likely to be corroded, and from an examination of these it was concluded the greatest corrosion in cast-iron piles exists close to low water, and does not extend to any considerable depth below it, a conclusion which also applies to bolts and braces. After 25 years' exposure in a salt-water way the piles were in very good condition, and corrosion has only occurred in places which are easily accessible for repairs and renewals. The Austerlitz Bridge, Paris, erected in 1802-5, which consisted of five cast-iron segmental arches of 106 feet span, was taken down in 1854, and it was found that the cast-iron arches had not given satisfactory results, but were displaced and broken in places. The condition of some of the original cast-iron bridges taken down for the London and South Western widening works over Upper Kennington Lane, South-Lambeth Road, Wandsworth Road, &c, London, was not found to be satisfactory, nor was it expected it would be very reassuring, f It was stated it was caused by defects in the castings, such as air-holes, also that, "taking into con- sideration the condition and design of these ribs, and having carefully examined sixty after being broken up, it suggested that calculations as to their working strength must obviously be misleading, and that it is impossible to assign to them a proper margin of safety." * Vide « Engineering,' February S, 1888. M I t OF MiD 1 utes .. of Pr °ceedings,' Inst. O.E. Paper by Mr. Szlumper, LIFE OF METALLIC STRUCTURES, ETC. 137 At Lowestoft, where cast-iron sheet piling, 1^ inch in thickness, after some forty years' immersion, was examined ; it was found the iron had not become soft, and had only deteriorated slightly at the edges of the piles, where the original skin of the casting had been removed. An examina- tion of the cast iron Fleet sewer, erected by the Metropolitan Eailway Company, London, showed 4hat the cast iron of the original crossing had suffered no deterioration from the constant flow of sewage matter through it; the red lead with which the interior had been cased was still adhering to the sides, although it had been washed away from the invert ; all the joints which had their surfaces truly planed, and were made good with iron cement, were perfect, and the packing was as good as when it had been put in some four or five years before. Professor Kennedy has mentioned that he had to inspect a piece of good wrought iron which had been erected 50 years before. He found that on its being broken across it showed signs of damp having penetrated it in many places, through defects in the welding of the original material ; although it was not corroded on the surface, it being preserved by painting, particles of golden rust were found in the heart of the material. Moisture had found its way into the spaces in the metal where the layers of dirt, during the process of welding, rolling and manufacture, had originally interrupted the continuity of the metal ; hence the more pure and homo- geneous the metal, the less the chance of corrosion. This shows that, even although the surface of iron may appear to be sound and not corroded, it is possible for internal corrosion or electrolysis to be in progress. It would be a valuable guide, when old girders are re- moved, if even a few bare details were supplied, stating the dimensions of the original members of a structure, and they were compared with the thicknesses when the bridge was taken down ; the nature of the traffic; abstract of specification ; date of erection ; how often painted, and with what substance, &c. &c. Then, in time, some valuable data could be tabulated 138 CORROSION AND ITS PREVENTION. showing the most vulnerable parts of an ordinary metallic structure, similarly circumstanced in every way, and to be erected for like purposes of traffic. At present, information is very fragmentary and difficult to obtain. Mr. Ewing Matbeson, M. Inst. C.E., bas mentioned tbat in altering a wrougbt-iron bridge at Stepney, erected 25 years before being taken down, and subject to a heavy and constant railway traffic, the upper boom or box had been riveted and caulked like a boiler, and was perfectly air-tight, and the inside plates were as good as on the day they were first put in ; while some of the parts exposed to the atmosphere, and ineffectively painted, had deep pits bitten out in all directions, materially weakening them. The worst part was where the iron had been brought in contact with wood, the acid of which had so destroyed it that an angle-iron \ inch in thickness was worn down to a knife edge. In another case, an approach to a large terminus, the rivet heads were found to be almost entirely corroded, and the T-iron stiffeners were nearly rusted away. The following detailed description of the condition of the structure, erected about the year 1847, so far as regards corrosion, is extracted from ' The Engineer,' May 26, 1876, in which full particulars of the bowstring bridge, and a drawing were given. " Generally, the ironwork of the main girders was in a very fair state of preservation, but it was much corroded wherever it had been in contact with wood, especially where the latter was decayed, as it was in almost all cases where it touched the iron. The decaying wood seems to have accelerated the oxidation of the iron, which had gone on to a greater extent where the two materials were in contact than can be accounted for by the simple retention of moisture ; parts equally unfavourably situated in this respect being much less affected, as, for instanoe, where the iron was in contact with earth or rested on brickwork unprotected from the wet. These remarks, however, do not apply to portions of the top flanges of the oross girders. To form a bed for the corrugated floor-plates these flanges bad pieces of plank wide LIFE OF METALLIC STKUCTUKES, ETC. 139 enough to extend a little over both edges, bolted tightly down to them, and wherever these planks remained sound they had excluded the wet, and prevented the iron from being much rusted. There was a considerable amount of rust in all the confined spaces open to atmospheric influences, but not so much as in the former case ; whilst inside the arched boom, which is completely closed, there appeared to have been no oxidation whatever, the inner surfaces of the plates remaining as clean as when they left the maker's yard. So entirely had this part escaped, that when the first section, ^ of the entire boom, was lifted out, the threads of some screws which were seen projecting about half an inch through the plate into the closed chamber were still bright, having undergone no change except a very slight browning since they were put in to make the joint nearly thirty years ago. In some of the exposed parts of the main girders the rust had eaten out a number of curious little pits, of a more or less circular outline, varying from ^ inch to 1 inch or more in diameter, and often fully ^ inch deep ; some of these pits were isolated, others were clustered together in considerable numbers ; they were chiefly found in those parts of the main girders which were otherwise least injured by the rust. Some of these pits must have been of long standing, for they were painted over, and the paint had, in some cases, protected them from further action. The lower flanges of the cross girders were made of five thin plates, which were held to- gether by rivets so far apart that the rust had got between the plates and thrust them asunder, so that in many places the part of the flange between two plates was bulged out so as to be much thicker than it originally was. The bulging out of the plates was especially noticeable at the ends of the girders, where, in consequence of the flange having been diminished in width, there was not room to continue the outside rows of rivets. This serious defect in these girders proves the necessity for having flanges which are to be exposed to the weather secured with a sufficient number of rivets to make tight the joints between the plates. The 140 CORROSION AND ITS PREVENTION. wind-bracing was quite ineffective, being only of flat bars 5 inches wide by f inch thick. It had sagged in some places as much as 6 inches, and in one or two places was rusted entirely through. " The condition of this bridge suggests the following observations on some points of practical detail, which may perhaps be advantageously considered in designing struc- tures of iron combined with other materials. Ironwork, which is to be placed in situations exposed to the weather, should not be enclosed by wooden panelling, or otherwise covered with wood, where it can be avoided ; and where it rests on masonry piers or abutments the iron should either be left quite open for the free circulation of air round it, or the masonry should be so built up to it as to exclude the air altogether ; confined spaces open to atmospheric influences should be avoided, not only as being inaccessible fur repainting, but because they are attacked sooner than those parts which, being freely exposed, are soon dried after a wetting, and from which deleterious matter is periodically removed by wind and rain. Very thin plates are open to the objection that they present a number of joints which it is much more difficult to rivet tightly enough to exclude air and moisture, than in the case of thicker and, consequently, stiffer plates ; every joint is thus liable to be speedily attacked by the rust ; and, for the same reason, when corrosion has once set in, they offer less resist- ance than thick plates to the increasing thickness of the layer of rust between them. This is well seen in the example before us, in which the upper flanges of the cross girders are each made of two T V-mch plates, into the joints between which the rust had made but little way, whilst there was none of the bulging out of the plates which might be seen in many places along the lower flanges, although, in consequence of the top flanges having been covered with timber, they were much worse rusted than the bottom ones wherever the timber was decayed. The brightness of the screws inside the arched boom noticed above offers a suf- LIFE OF METALLIC STRUCTURES, ETC. 141 ficient testimony to the value of good work ; in this instance the riveting had been so well done as to render the inside of the boom practically air and water-tight, so that it would probably have remained intact for an indefinite period. It is worth noticing that under the woodwork and in some of the confined spaces there was under the red rust a layer of imperfectly-oxidised iron, sometimes more than ^ inch thick, which was very hard, and could not be removed by scraping, but which, on being struck sharply with a hammer, fell off in thick flakes ; these were quite brittle, and had a rather metallic-looking fracture. By well hammering the iron it could be freed entirely from this scale, leaving a surface somewhat uneven, but quite clean, and with a greenish-black lustre. This suggests a point which should be looked to in repainting rusted ironwork, as of course any paint laid on this scale would speedily be destroyed. The ironwork of this bridge had been painted first with red lead, and after- wards with stone colour, which had protected it fairly well in those parts that were favourably situated, but a com- parison of these portions with those so placed that the paint on them could not be renewed, shows how great is the damage which outdoor ironwork sustains if it is not kept thoroughly painted. The floor not having been made approximately water-tight, the cross girders had suffered greatly in consequence, the drip from the engines having kept them in some places almost constantly wet." Mr. J. W. Wilson, in a paper read before the Society of Engineers, in March 1875, stated that the solid wrought-iron piles at West Brighton Pier, which was opened on October 6, 1866, were originally of a diameter of four or five inches. They had proved unsatisfactory owing to the continuous corrosion, as in several instances after a lapse of time, they were reduced to a diameter of two inches, or even less, and are being gradually withdrawn and replaced by a more durable material, cast iron. Examination of some iron railway bridges by Austrian engineers, consequent upon a State order in 1887, showed that in the case of wrought-iron box 142 CORROSION AND ITS PREVENTION. girders, open at the ends to the free circulation of air, the original colour has still remained after the lapse of 30 years, but it was found locomotive smoke seriously affected the girders under which the trains passed. The girders of the original Tay Bridge, erected some four or five years before, that remained after the disaster of December 28, 1879, which befel a portion of the structure, were found on inspection to be generally in good condition. Mr. St. John Day, in a paper read before the Institution of Engineers and Shipbuilders, Scotland, in February 1880, stated "that the bolts, 1^ inch in diameter, holding the ties to the lugs of the piles of the original Tay Bridge were so corroded that tbey would have had to be replaced every four to six years, and that one of the bolts found was nearly, if not quite, half-way converted into rust." At the new Tay Bridge, the Board of Trade pro- hibited the use or continuance of small cast-iron columns with wrought-iron ties. In a report of Colonel Macomb and Lieut.-Col. Ludlow, Corps of Engineers, U.S.A., on the Improvements of Eivers and Harbours in New Jersey, Pennsylvania and Delaware, 1879-80, the following paragraph occurs with respect to the appearance of some of the screw piles which were lifted and reset at greater lengths, to meet some contingencies of soft stratum, after an exposure of five years to the action of the salt mud and water. " Below the line of the bottom, the outer scale of the wrouglit-iron pile itself was still smooth, hard and bright, with only a few rusty spots here and there. From the level of the bottom up to low water, the surface of the pile — after the mussels, which covered it to the thick- ness of six inches with a closely-adhering coat, were scraped off — was found to be full of small cavities, and of a streaky fibrous appearance. Between low and high-water marks the pile was still smooth; but above high- water mark to the cap, corrosion had taken place, imparting a scaly blister- like appearance. The braces near to and above high- water mark had much corroded; the screw threads are cut per- LIFE OF METALLIC STRUCTURES, ETC. 143 pendicular to the fibre of the iron, and nearly destroyed, making it necessary to cut them off, and weld on pieces for the new threads." Inspection of suspension bridges erected 25 to 30 years ago has shown the imperative necessity of having all parts of a cable accessible for a examination. Several wire-rope bridges have been rendered unsafe for traffic in less than 50 years, and had to be renewed owing to unperceived oxidation in the interior of the wire cables. The difficulty in preventing wire ropes deteriorating from corrosion is that in round wires there are spaces between them, and these must be completely filled with some substance of an elastic but tenacious nature, having great adherence — qualities somewhat contrary. It is advisable not to employ wire ropes in places that cannot be properly inspected. Tor this reason, in some modern wire-rope suspension bridges, links are used in the tunnels and wells, and anchorage abutment chambers, where corrosive influences will be very active owing to the humidity of the air, and moisture. Laminated-wire rope has been introduced, in order to remove the objection to round- wire rope under the circumstances mentioned, and to obtain compactness. Unless the interstices between the wire cables of suspension bridges are completely filled with an anti- corrosive composition, having the requirements stated, and the outer surfaces are properly protected, such cables will have their useful life soon impaired, and ultimately be so weakened as to be worthless. Galvanised-steel wire ropes are, therefore, sometimes used to prevent or lessen corrosion. There is, however, one point in favour of a cable bridge, namely, that the surface exposed to corrosive influences is very small as compared with, what may be called, rigid girders or arches of any design, and it is desirable to reduce the surface to be painted to a minimum ; on the other hand, any corrosion is of more importance, and if tho cables are neglected, and allowed to corrode, the whole stability of the bridge is soon impaired, and may be destroyed. Frequent 144 CORROSION AND ITS PREVENTION. inspection of the junction of cables with the anchorages of any wire-rope suspenders should be made ; also at their junction with the stirrups and any woodwork. The cutting action against one another of the wires forming ropes used for hauling or lifting purposes can be moderated by oiling. The oil should be such that it will permeate the rope. The principal causes of the destruction of wire ropes are the wearing of the outer surface of the outside wires, the rubbing of the wires against one another, the strain on the metal when worked over too small a pulley, and corrosion. It is necessary that the oil contains no acids, or corrosion of the wires will be accelerated. Wire ropes used on inclined planes do not generally wear out by abra- sion, but by crystallisation. As steel-wire ropes, because of their lightness and great strength, are so frequently used, any reduction of their sectional area by corrosion is of much importance. It is especially necessary to keep them in a dry place, at a moderate temperature, and free from vapour or steam. It does not follow because a wire rope is well- greased that it is thereby preserved from corrosion, even if the grease is free from acid and of excellent quality, for if it hardens, rust may form between it and the wire. The mere covering a piece of metal with some substance which soon changes its character will not prevent corrosion. In wire-rope suspension bridges, it is usually the case that those parts which cannot be inspected are most exposed to damp and deteriorating influences. In inspecting several such bridges, with a view to ascertain their condition, it has been noticed when a slimy liquid of a reddish colour appears on the surface, it is almost always an indication that the interior wires have been seriously attacked by rust, and it has been found to be so in several wire-rope bridges. M. Bernadeau, in the 'Annales des Ponts et Chaussees,' 1881, refers to a case in which only 15 out of 180 wires forming each cable were in good condition, the rest being as brittle as glass. The suspension bridges had been erected between 26 and 39 years. Two such bridges fell after 26 and 28 LIFE OF METALLIC STRUCTURES, ETC. 145 years' use. Repairs were effected by removing the rusted wires, and replacing them by new ones, the sound wires i being carefully cleaned and dipped into boiling linseed oil, and i finally the cables were coated with coal-tar. The operation ; was carried out without stopping any but the heavy traffic, only one vehicle being allowed on the bridge at a time. ; Wires have also been considered protected by dipping in asphaltura. It has been stated by an authority that there are some 500 suspension bridges in France. The fact that [ a less weight of wire is required than that of iron links, led to wire cables being very largely adopted instead of bar links, of which latter system the bridge over the Thames, at Hammersmith, is an excellent example. It has been found in France, and it may be also said in America, that the life of wire cables is precarious, for oxidation proceeds I in the interior of the cable unperceived, and there is no proof that the means adopted for their preservation will do so after many years, or even 25 years, although the wires may very likely be preserved, and some engineers avow they are. The failure of a suspension bridge at Angers, many years | ago, showed that it was impossible to keep the hydrate of i lime coating there used in immediate contact with the cables, ithe consequence being that moisture, and other corrosive influences, reached the metal, and the coating of lime was :thus rendered practically useless. In some recent French wire-rope suspension bridges, at the anchorage of the cables, an alloy of 4 parts tin, 5 of lead, and 1 of antimony, has been used for filling between the iwires in the conical hole instead of lead. It is found that [very fusible alloys are too soft to form a solid anchorage ; [and alloys which fuse only at a high temperature weaken the wires, and to obtain a long continuous wire it is necessary [to unite several separate lengths. Instead of binding the iwires together with fine wire, or screwing them into little [hollow cylinders, which latter it is found expose them to ioxidation, a method has been used of filing the extremities !of the wires obliquely, and of soldering them together ; the L 146 CORROSION AND ITS PREVENTION. objection to it being that though the joint is strong, the wire is somewhat weakened by it. In either wire or link suspension bridges, the anchorage and saddles will be found to be the places where corrosion has to be especially guarded against. It is well to remember the larger the superficial area of the wire cables, the more surface there is exposed to oxidation. On September 15, 1886, as a squadron of Uhlans was crossing a suspension bridge over the Ostrawitza, at Mahrisch- Ostrau, the bridge gave way, and many were killed or| injured. It was commenced in 1846, and not finished till June 1851, owing to floods. The fracture occurred in one of the anchor chains. It was found that the material was thoroughly altered in character, the iron being oxidised through, so that it could be crushed in the hand. The! anchor stays in question consisted of 12 links, of which one = was completely corroded ; the others were reduced to about ■ ^th of the original section. The original sectional area! of the chain was 24*4 square inches, but it rusted away to about 4 square inches. The chamber enclosing the portion I of the anchor chains where the fracture occurred was open! to all the surface drainage of the road. A somewhat peculiar! circumstance connected with this failure was that, in July 1885, an official inspection and report had been made, at the urgent request of the municipality, which said " the bridge j has been examined in all its parts, and is in good and safe condition." The ends of the wires are sometimes placed in a tapering or conical metal tube, into which a pin can be screwed so asl to prevent the end of the wire moving upwards. To thel merits of the system of turning back the strands and splicing! them so as to form a thick end, firmly hold the ends and so| distribute the strain by screwing up, no reference is here! made ; but it is pointed out that unless water, air, dirt, andf moisture are prevented from percolating down the strands,! the wire inside the conical piece may soon corrode, and the ' fastenings become ineffective. The Niagara Suspension^ LIFE OF METALLIC STRUCTURES, ETC. 147 Bridge was opened for railway traffic in 1855, and in 1877 Mr. T. C. Clarke, M. Inst. C.E., examined the cable near one of the shoes, and found some of the outer wires corroded through, but the second layers sound. The cause of the corrosion of the outer wires was considered to be that the elongation and contraction of the strands, under moving loads, had loosened the cement from the outside strands, and therefore allowed moisture and air to penetrate. To preserve the wires from rust, they were originally covered with a thick coating of hydraulic cement. Where the hydraulic cement surrounded the wires, they were found to be in a perfectly sound condition, and no signs of oxidation were visible. In the specification of the steel cable wire for the East Eiver Suspension Bridge, drafted by Mr. W. A. Boebling, the following clause referred to galvanising the cables. " 8. The cables of the East Biver Bridge are suspended directly over a salt-water stream, and are, in addition, exposed to the salt air of the neighbouring seashore. Ex- perience has shown that the ordinary means of protection, such as paint, oil or varnish, which would be ample in the interior, are totally inadequate to prevent rusting in localities so near the coast. The only certain safeguard is a coating of zinc, which acts by its absolute air- tightness, as well as by its galvanic action, and is not easily abraded. " The galvanising must be done throughout in a thorough and perfect manner ; each ring will be inspected in this regard by the inspector when he tests the wire. All rings will be rejected which show spots imperfectly covered, or are full of rough lumps, showing a defective stripping. The galvanising must be of uniform thickness, and must not scale off, or show any cracks when the wire is bent. " The attention of the manufacturer is particularly called to the point that he must galvanise at such a temperature, and in such a manner, as not to destroy the temper of the wire. * See ' Scientific American,' 1881. L 2 148 CORROSION AND ITS PREVENTION. The manufacturer must run the whole risk in this respect, because the wire is inspected and tested after it has been galvanised. Samples which have been received and tested show that it is not difficult to reconcile these two operations, and that when proper caution is exercised, and the parties possess sufficient experience, the wire can be properly galvanised without impairing the temper." Mr. F. Collingwood describes * the results of an examina- tion, in August 1883, of the Alleghany Suspension Bridge at Pittsburgh, U.SA., erected and in constant use since 1861. Each inner cable contains 2100 wires laid up in seven strands, and measures 7^ inches in diameter. Each outer 700 wires laid up in two strands, and is 4^ inches in diameter. The serving or wrapping wire on the cables measures 0*098 inch in diameter, and is included, of course, twice in the diameter given for the cables. The interior spaces were filled in solid with hot coal-tar, which had been boiled and treated with quicklime to neutralise any acid it might contain. It was found, in some of the subterranean portions of the cable, which had in addition been covered with canvas, that the tar had partly disappeared, and that the cavity was nearly full of a dirty-greyish liquid. There was also extensive rusting of the wire, so that the seizing wires, 0 • 06 inch in diameter, were in many places rusted through, and the cable wires deeply pitted. A second cable-end was opened with similar results. A general survey of the bridge revealed fine cracks in the paint on the cables, which admitted moisture, &c. Eecommendation was then made that all the cables and other ironwork should be scraped and repainted. The serious damage to the wires was found to extend about 3 feet from the anchorage outward. Beyond this there was a little dry rust, but no pitting, and still further from the anchorage the paint on the interior wires was gummy. The rust appeared to be of two kinds : a red oxide where the wire seemed to have been attacked as if by acid ; the * ' Minutes of Proceedings,' Inst. C.E., vol. lxxvi. LIFE OP METALLIC STRUCTURES, ETC. 149 second form of rust was a hard blackish substance, containing much sulphur, which, when scaled off, left a deep pit. The result of a chemical analysis showed the rust to be a -com- bination of the hydrated peroxide of iron and sulphate of iron. The liquid found among the strands consisted of a weak solution of carbonate and sulphate of ammonia, coloured by tarry matter, and was almost identical with tar water from gasworks. The original cables were first coated with boiled linseed oil, and afterwards with coal-tar. It was remarked that the tar had evidently not been heated long and high enough to drive off all the water, and the salts of ammonia, contained in coal-tar. Professor Wuth, of Pittsburgh, further reported that " the oils contained in the tar first dissolved the coat of linseed oil ; then the sulphuret of ammonia, which is contained in the tar in considerable quantity, acted upon the iron, con- verting it into the sulphuret of iron, which again was converted into the sulphate by the oxygen of the air, which could not have been completely excluded. This alternate action of the sulphuret of ammonia and tar was continued until the sulphuret was entirely exhausted. The oxidation was further carried on by the atmospheric air in the presence of water and carbonate of ammonia." The presence of water was accounted for by air having slowly percolated to and fro as the masonry changed in temperature, and moisture would probably be condensed, and the water slowly collected. The sulphur and ammonia, it was considered, accumulated in this way, as Pittsburgh is a very smoky city. The preserving measures adopted were, first, all the wires were cleansed thoroughly by scraping, as drenching was found to be im- practicable. Wedges were used to force the strands apart. It was found the damage was almost entirely confined to the outer two layers of wires in each strand. The seizing wires had held the strands so close together as to exclude the destructive agent from further penetration. After the wires had been repaired, by jarring with mallets all loose rust and dirt were removed, and the wires were thoroughly 150 COKKOSION AND ITS PKEVENTION. saturated with raw linseed oil. Two days afterwards, a coating of boiled linseed oil was applied, and then the strands were tightened up by the seizing wires. The cable was then saturated with white lead and oil. The cable-end for some two feet back from the clamp, the length being surrounded by canvas, was bathed in melted paraffin until it could hold no more. Changes of temperature had, on examination some months after, had no effect upon the paraffin as a protection. It would appear that, although wire suspension bridges are excellent for sustaining strain, unless corrosion is thoroughly guarded against, the solid link or rod-suspension bridge is to be preferred. It is open to question whether a linseed-oil coating is a durable and effective coating. Anchor blocks of Portland -cement concrete are to be preferred to those of masonry or brickwork, as the former can be made a monolithic mass, but the latter have joints of uncertain and unequal strength, and impervious Portland- cement concrete is a proved preservative against corrosion ; masonry and brickwork not only do not always so act, but may be the cause of the introduction of active corrosive influences, either along the joints, or through the stone or brick. The anchor cables formed of 127 wires, 0*19 inch in diameter, on the right bank of St. Christophe Suspension Bridge,* it was found had been so reduced from the oxidation of the lower portions of the wires, by the oozing of fresh water into the gallery, and the intermittent action of sea water, that they had to be renewed in 1884. Ten per cent, extra thickness of wire was allowed to provide against corrosion at exceptionally exposed places. Experiments on the portions cut off from the old cable showed that the wires erected in 1850 were uninjured, and had lost hardly any of their primitive strength. In the Cincinatti Suspension Bridge, so far as oxidation is * Vide Annates des ' Ponts et Chaussee's,' 1886. LIFE OF METALLIC STKUCTURES, ETC. 151 concerned, all the parts are open to inspection. The wires composing the cables are well protected by varnish, and so closely and compactly compressed that all interstices are filled with linseed oil. This protection is further increased by the outside wrapping, so that it was considered no appre- hension need be entertained as regards corrosion. On taking apart the cables of the suspension aqueduct at Pittsburgh, after seventeen years' exposure, and the effect of dripping water leaking from the trunk, and with little or no care bestowed upon them, the wires inside were found just as free from rust as on the day they were put in. The cables are wrapped from end to end with No. 10 wire, which gives them the appearance of solid cylinders. A well-manufactured cable, it was considered, may be regarded as perfectly free from all danger of rusting; however, the examination of several wire-rope bridges, previously mentioned, hardly con- firms so sanguine an assertion. CORROSION AND ITS PREVENTION. CHAPTER XL NOTES ON THE CORROSION OF METAL EMBEDDED IN CONCRETE, BRICKWORK, OR MASONRY. With regard to metal embedded in concrete, brickwork, or masonry, it is not advisable to assume, when iron or steel is embedded in brickwork, masonry, or concrete, that moisture and air are excluded from it, and that if it be cleaned and tarred or painted before being covered then it cannot corrode ; for vibration and weather influences cause the mortar in the joints to perish, and cracks occur, and this may even be the case with good Portland-cement concrete. When a fissure appears it is a conduit by which moisture, air, and other corrosive influences can reach the metal. By pointing the joints of brickwork or masonry, as required, and closing any cracks or fissures as they occur in Portland-cement concrete, such deleterious action can be lessened. However, if ironwork is free from any corrosion when placed in position, and properly cleaned before it is coated, and is fixed in air, damp-proof, and water-tight concrete, brickwork, or masonry, it is unlikely to appre- ciably corrode, except under peculiar circumstances— such as the material it is embedded in being impregnated with, or containing, some corrosive influence acting in conjunction with galvanic deterioration of the metal. Brick and stone are permeable and porous, and, moreover, retain some moisture, quite apart from consideration of the limiting zone of capillary attraction from the ground. Stone, such as Portland, will absorb water to a considerable amount. For instance, in immersion in distilled water, some experi- THE CORROSION OF METAL IN CONCRETE, ETC. 153 ments made by Mr. Spiller, F.C.S., showed that the "Whit [bed " Portland, used for external purposes, absorbed 92 [grains in 1^ hour, the stone weighing 1421 grains, or 6^ jper cent. "Base bed" Portland, fit for internal decoration only, being so porous, absorbed 126 grains in 1^ hour, the stone weighing 1291 grains, or nearly 10 per cent. TJnder- ; burnt, dusty, soft bricks, soon decay in damp situations. ■ Bricks which are dense, hard, even in texture, and have a vitrified appearance, will resist decay. The weathering j qualities, density and uniformity, should be considered, and that some bricks contain a large percentage of soluble salts. I Porous, open-grained stones or bricks must be capable of ( receiving more moisture than those which are closer grained [or denser. Although they are frequently called solid, their [weight, as compared with the metals, to some extent indicates \ that there are interstices between the grains or atoms which may contain moisture or air, &c. If any material consists of many different ingredients, each has its characteristics, and • therefore there may be causes of decay, disintegration, and [corrosion that would not be present in a less compound material ; there may be mechanical action caused by the different rates of expansion and contraction, as well as the chemical action. Salts are formed in the bricks or stone by decomposition, which implies a degree of decay, and, particularly in brick- work, causes efflorescence, i.e., a white powdery substance to appear, consequent upon the salt losing its water of crystalli- sation. This is believed to be due to the sulphuric acid from ; gas-burners and coal-stoves, in the presence of air and j moisture, acting upon the silicates of lime and magnesia in the brick clay, sulphates of lime and magnesia being formed in the bricks. The bricks becoming wet, the solution evapo- rates, and efflorescence occurs. Some aver it is due to a minute vegetable or fungoid growth. However, no matter how such an effect is caused, it is decidedly corrosive in its tendency, and should be guarded against as much as possible. This j efflorescent and corrosive influence is greatest between any 154 CORROSION AND ITS PREVENTION. portion of a structure liable to be alternately dry and wet, or humid. Therefore, above the ground, and within the range of tides, i.e., where alternately exposed to air and to submer- sion, such influence is greatest. As the cost is so trifling, the ends of all metallic girders or joists that must be embedded, and therefore cannot be inspected, should be completely surrounded by Portland-cement concrete of an impermeable character. For this purpose, as impermeability, and not strength, of the concrete is particularly required, a 3 of fine dry clean sand to 1 of Portland cement, or a 2 to 1 mixture, can be adopted, a poorer concrete not being suitable. The ends of a girder or joist should be coated with approved tar- asphaltum composition, or other acknowledged effectual anti- corrosive paint, before being embedded. Some recent experi- ments by Professor Bauschinger showed that the adherenco between iron and Portland cement is as much as 625 lbs. per square inch. It must, however, vary according to the quality of the Portland cement, and nature of the surface to which it adheres. In some experiments conducted with the view of testing the Monier system of iron wire and rods embedded in concrete, the adhesion between the iron and concrete ap- peared to be perfect, and it was concluded they could be con- sidered a homogeneous mass. Still, much depends upon the dryness and cleanness of the iron, and the adhesiveness and imperviousness of the Portland-cement concrete being uni- form and of very considerable amount. The system of dowelling joints run in Portland cement is much to be preferred to any method of iron cramps for stonework, as the latter not only have an ugly appearance, but soon corrode, become bent, and have a destructive effect upon the stone. When iron has to be placed in contact with any other substance, its porosity, and the capacity the ma- terial possesses to contain water, should be considered, for they will in great measure determine its corrosive influence, disintegration by frost, and the presence of deteriorating agents. Iron embedded in properly-made and mixed water- and THE CORKOSION OF METAL IN CONCRETE, ETC. 155 air-tight Portland-cement concrete, has not yet been shown to rust, and the preservative effects of such concrete may he considered to be established, provided the surface of the metal was clean and dry on the Portland cement coating being applied, and free from corrosion ; and, as the expansion of cement and iron by heat are nearly the same, there is no struggle between the substances to cause cracks, fissures, scaling or disintegration. One of the most conclusive proofs of the value of Portland cement as a preservative against corrosive influences is, that formerly it was the practice, in the case of iron vessels, to coat the inside surface of the plates with a Portland cement fine gravel and sand mixture, and it was found to last as long as the ship, the metal almost invariably being completely preserved from rust. It has now given place to bituminous and other coatings, simply because they are lighter. Other proofs of the value of concrete in preventing corrosion may here be mentioned, for a high authority, Mr. T. Curtis Clarke, has stated, in a report on the Niagara Suspension Bridge, that the painted anchor links of that bridge in 1873, eighteen years after its erection, which were embedded in hydraulic cement, had not rusted, and the paint on them was found to be fresh, and the seven strands forming the southerly pair of cables, so far as they were exposed to view by the removal of the cement sur- rounding them, were in a perfectly sound condition, no signs of oxidation being visible. In another place, that of the cable wires, which, of course, were not embedded in cement, the paint had cracked along the spiral lines of twist, admit- ting water which caused rust. It has been found that water-tight Portland-cement concrete preserves the buried part of a telegraph post from corrosion, and also fulfils the double function of ensuring its stability. Sheet-iron posts not embedded in water-tight Portland-cement concrete soon corrode, and the portion above the ground has to be painted every two or three years. Ironwork which, it is believed, was inserted in the walls of buildings erected 500 or 600 years ago, has been found, when embedded in lime mortar, to 156 CORROSION AND ITS PREVENTION. be in a fair state of preservation. On the other hand, M. Eiffel has stated that some iron rag-bolts, which had been placed in fortifications for 200 years, had become en- larged to from 2 to 2\ times their original diameter by rusting in mortar. Probably this was due to oxidation of the metal having commenced, or to its being damp or dirty when placed in the structure. 157 CHAPTEE XII. THE INFLUENCE OF DESIGN AND WORKMANSHIP WITH REGARD TO CORROSION. The influence of design, so far as regards corrosion, is very considerable, and the serviceable life of a metallic structure may be greatly prolonged by attention being directed to prevent the decay of any of its parts without detriment to the best distribution of the metal to meet the strains. Sim- plicity of construction is one of the chief objects to be desired. However, it may not be easy, or possible, to design a structure so that every part can be inspected and maintained ; still, if any portion is inaccessible, so that it cannot be coated and protected from time to time, corrosion must take place, and increase. But much can be done towards the accessibility of the parts by simplicity in the form of the various members, and by care being exercised that complicated arrangements are avoided as far as practicable, and by all depressions in which water may collect being drained by holes, or filled with a durable waterproof mastic. In bridges, although there may be disadvantages in having as few members or separate parts as possible— because of deflections, bending strains, vibration, greater importance of accurate and perfectly strong joints, inequalities of ma- terial, and faults and flaws, &c, &c. — possible corrosion is undoubtedly lessened in the case of thick, massive sections, as compared with thin and light, which have a larger surface area exposed to corrosive influences as compared with the heavier and thicker members. Wherever bars or pieces cross, or are bolted or riveted 158 CORROSION AND ITS PREVENTION. together, water permeates into the interstices, and deflection and vibration of the parts of a structure, caused tr traffic or shocks, accelerates any corroding action ; therefore it is well to reduce to a minimum the joints and surfaces h contact, the effect of a rolling load being sometimes to rais<, and then to deflect a part of a girder. As a metallic structue consists of a combination of a number of members, none of whose dimensions are usually nearly equal to its length, md few to its height, the joints are of great importance, and, is they are weakened from unequal bearing, and other defecs, a little extra strength will not cause an excess of metal. Corrosion, when accelerated by galvanic action, acts with tie greatest energy in the interior surfaces of joints and such iiaccessible places. Every member should be designed so tlat it may, relatively to the strain it has to bear, be as strcag as any other part, and be free from corrosion ; for surphs strength may cause local weakness, and result in portions of i structure being especially liable to corrosion and deterioraton. Con- tinuous girders, particularly, may be brought tc a serious condition consequent upon looseness of the parts that may not be apparent, and by corrosion ; and there is always more risk in their case from any deterioration of tie section. In metallic structures, in addition to the metd required to meet the strains, the contingencies of manufacture, con- struction, vibration and corrosion, and any otier special circumstances, have to be considered. It may h advisable to design them, even when modifications in thei] theoretic- ally correct form have to be made, so as to allow of inspection and sufficient space for painting and repairs. !f possible, every part should be so designed that it can be examined, repaired, and painted, without danger to life ; am it may be expedient to provide special facilities for thes> purposes, such as examination-platforms, movable ladders, cc. There is no advantage in adopting other than the usua and most suitable forms of iron, if merely with the viev to lessen corrosion, when it can be appreciably prevented ly ordinary care and attention, for they have satisfactorily wihstood the INFLUENCE OF DESIGN AND WORKMANSHIP. 159 test of many years' use. The object in designing ironwork, apart from questions of strain, &c., should be to expose as small and thick a surface as possible to corrosive influences, and allow of easy access to every portion of a structure. Practical experience may often overthrow any purely scientific theory of the best shape or section. The strongest forms are generally the simplest, and the fewer the parts the better, or increased facilities for corrosion will be provided. A section or form difficult to roll or cast is not so strong as one that can be easily made, or without much difficulty. It should be remembered that when sections of iron are wide, as well as deep, they are not easily rolled, nor is a channel-iron of con- siderable width, or a deep wide H -section. In the case of iron of a plastic and inferior nature, such beams are not nearly so difficult to roll. Z-bar sections, deep angle-bulbs, L and T-bulbs, it is said, can be more easily produced in steel than iron, and lengths of 60 to 70 feet have been made. This saves joints, and lessens corrosion. The cellular form of member is not often now used in girder-work in Great Britain, because it cannot be painted internally or repaired. If a box girder, column, or any cellular section, has to be adopted, which cannot be inspected, it is well to line it with Portland cement, or to fill it with comparatively rich Portland- cement concrete, so as to protect the surface from corrosive influences, such as air, moisture and vapour. The form of a piece of iron or steel has, by careful observation, been found to have an effect in preserving it from corrosion, and it has been shown that pointed or square edges, abrupt bends and angles, particularly in submerged, or occasionally sub- merged work, should be avoided, and that round and curved surfaces are less subject to corrosion. Whenever various forms and sections are used in a metallic structure there must be much manipulation in order to make them fit. Im- perfect fitting and joints must then increase, and corrosion ; the strength of a structure be diminished, and be more liable to deterioration. The limit of elasticity is the chief point to ascertain in 160 CORROSION AND ITS PREVENTION. determining the safe strain on metals, for it should not be deduced from the breaking-weight alone. To know the strain which the material will bear without injury to its elastic powers is absolutely necessary to enable a correct design to be made. If every care is taken that corrosion cannot occur, or is at once arrested, and any strain brought upon it is well within the limits of elasticity, the metal should not deteriorate seriously ; but whether age and being in one position do cause deterioration has not been authorita- tively determined ; however, it has been shown that strained and unstrained iron do not equally corrode consequent upon galvanic action. The stretching of a rod, chain, or bolt, when tenfcilely strained, is not uniform unless the mass of the metal is of the same strength. The greatest extension will occur at the weakest part. In compression members it is doubtful if the metal always gives equal resistance, for that portion which is situated at the greater distance from the neutral axis is in a position of increased resistance to strain, as compared with that which is in its immediate proximity ; therefore, unequally-sided sections are not generally subject to the same strain throughout their sectional area. The ties of a bridge should never be cut or nicked, but be left entire. If they have to be joined, corrosion at the joints will be a very serious matter. The best and most careful workmanship cannot make a constructed section of iron equal in strength to the same section of the same quality of metal as it issued from the rolls. In some experiments made by M. Dupuy, with apparatus for measuring strains, it was found that, as the braces or lattices bear only a portion of the shearing strain, and the remainder has to be borne by the flanges, it is important that the flanges should not be greatly reduced at the abut- ments, where the shearing strain is greatest. The strains also produced on the braces of the same panel are not equally distributed between the braces in tension and compression, and the flanges have a tendency to overturn, hence any corrosion of the metal in the flanges at or near the supports INFLUENCE OF DESIGN AND WOKKMANSHIP. 161 should be especially prevented, corrosive influences being fre- quently active about the ends and portions of girder-flanges near piers or abutments. Some flaws and defects of workmanship are likely to occur in built-up metallic structures, and when to these corrosion is added, a serious reduction of strength inevitably follows, which, sooner or later, will affect the stability of the whole structure. Apart from the question whether the same sec- tional area of metal, if in one piece of iron, will bear without fatigue more strain, or the same strain, with less deteriorat- ing effect, than if it were in two or more pieces, it is a protection against corrosion to have all parts of considerable sectional area for the reasons previously stated in this chapter. No metallic structure, when the sun shines on one side of it and not on the other, and the undersides are always in the shade or exposed to cool winds, is at rest, for there will be strain from this cause, and it will tend to loosen the parts, because of the unequal expansion and contraction, and one side or part may be more susceptible to corrosive influences than another. Thus, the top and bottom members of a girder may be unequally exposed to the sun's rays, and also one side receive almost all the sunshine, and so any tie-rods and bracings may work loose. Alterations of weather have not such an effect, as they are generally felt throughout a structure. For these reasons, although a plate girder of considerable span affords a cheap screen or parapet, it is doubtful if it is as durable, even when carefully maintained, as an open-web girder, for there will be a tendency for moisture to be attracted from the colder towards the hotter side, and the paint — which in order to increase radiation, and diminish the effect caused by alterations of temperature, should be white, or nearly so — is more severely tried on one side than the other, according as moisture or dryness may deleteriously affect it ; therefore, corrosion is likely to be unequal. If one side of a bridge or structure is shaded, the temperature of its surface may be less than that of the air, 31 162 CORROSION AND ITS PREVENTION. and if the latter is in a saturated condition, a deposition of moisture will take place, and any corrosion be accelerated. Those portions of a structure liable to be shaded, or in a dark and partly enclosed position, are especially likely to be corroded. It is advisable, in designing, to avoid troughs and channels in which one or more surfaces are shaded. Although the less surface exposed to the air per unit of sectional area the better, that which in the nature of things, when properly painted, must be exposed, it is well to expose to the air as much as possible to counteract dampness ; and also to avoid all forms and sections in which water can collect, or dust, earth, or vegetation, come in contact with the coated iron. Open-web girders, such as the lattice, Warren, the various trusses, &c, have one advantage so far as regards preserva- tion from corrosion, namely, that the sun and air reach them more uniformly, it being well to expose the metal to the air, unless it can be hermetically kept from it. In designing lattice, truss, or any form of girder bridge, if considerations of strength, rigidity, riveting, and bearing as a strut do not prevent it, the thicker a bar, and the less the surface exposed to the air and weather, the better ; and should light bracing be necessary, if it can be done so as to effectually brace the parts, without too great concentration of the strain on the principal members, it is advisable, considered from the point of view of corrosion, to reduce the number of bracing pieces to a minimum, so as to have as few joints as possible, and thicker bars ; and also in lattice girders, or any in which bars cross, to have as few intersecting members as possible, because moisture and dust, &c, penetrate and accumulate at the points of crossing, and the vibratory action caused by a rolling load helps to loosen the joints and aids corrosion by mutual pressure of the parts. All plates, angle, T, or other pieces of iron, except where especially shown or required by the drawings and specifica- tion for the purposes of camber or otherwise, should, by preference, be perfectly straight or regularly curved, so that INFLUENCE OF DESIGN AND WOKKMANSHIP. 163 no depressions are formed for the accumulation of dirt, dust, moisture, damp air, smoke, earth, vegetation, or gases, &o.', which will quickly cause corrosion to actively proceed ; and every angle, bend, and corner, should he so arranged that it can be inspected, cleaned and painted, and care should be taken in the design to prevent water finding its way down any joint, or between plates, or any contiguous members of a structure. As pressing iron into any reasonable shape is now so often done while it is hot, and without the injurious effect of hammering, there may be cogent reasons for using bends, curved, and other forms. However, there is no occasion in adopting them to do otherwise than avoid depressions or receptacles in which corrosive influences can accumulate. Apart from consideration of strain, care should be taken that all flat bars are perfectly straight, and that they are not slack, or moisture and dirt are certain to collect, and corrosion be accelerated. Packing-pieces should be avoided as much as possible, as they are not always shaped or planed so that they fit tightly, and they should be of the same metal m every respect as that with which they are to be in contact, and not always of cast iron. One of the most important provisions against corrosion is that all plates of a built-up metallic structure should fit so tightly, and be in such contact, as to exclude all air, smoke, vapour, moisture, water and dirt ; therefore, a closely-riveted structure is likely to be less affeoted by corrosion than one with more widely-pitched rivets, always provided that the edges of the plates are flat, and not turned up or distorted. To prevent air, smoke, vapour and moisture, from penetrating between the plates of a girder or structure, it has been suggested that preservative material might be forced be- tween the plates by a powerful syringe. It is certain, if any preservative could be made to permanently fill such vacuities, it would be more efficacious than any covering of the edges of the plates. Close riveting, if not weakening the effective area of the plates, tends to exclude corrosive influences, pro- si 2 164 CORROSION AND ITS PREVENTION. vided it does not damage the edges of the plates, or cause them to turn up or "bend. Thick plates are usually stronger per unit of area than thin plates. If it be necessary to use a thin plate in building up the flange of a girder, it should be placed at the bottom, or it will be liable to buckle. Although it may be advisable to have the thickest plate in the top of any flange, as it is considered the strain is greatest there, plates of equal thick- ness are, on the whole, generally preferred. Experience has proved that which was to be expected from the less exposure : namely, that plates do not corrode so quickly if they are united in a considerable mass, or if a thick one forms the bottom or top flange of a girder, provided the joints and riveting are water and air-tight. Sir B. Baker's reliable experiments have shown that it is best to make the thickness of the plates of flanges of uniform size, and that narrowness of the bottom flange may cause a girder to fail before the real strength of the metal is brought into action. If desired, the width of the top flange, when the load is on the bottom flange, can be made narrower, and the thickness increased, but the bottom flange should have a reasonable width ; and rigidity of the flanges, and web, which is generally con- sidered as being subject to diagonal compression due to the shearing strain, should be assured. In so designing a flange, attention can be given to the prevention of corrosion by having no thin plates. The strength of a wrought-iron plate is generally not uniform. In the middle, where the puddle-bars forming the pile are most drawn out, it will probably be greatest, and decrease gradually towards the edges. The designed width of a plate-girder bridge varies in different countries. In this, wide flanges are adopted to ensure stability, and to obviate the necessity of overhead bracing. In America narrow flanges, compared with ours, are used— say, from 0 • 75 to 0 • 50 of the width here adopted. Of course, with a narrow flange, thicker plates are required to make up the necessary sectional area of the metal, but in any than short-span plate girders for light INFLUENCE OF DESIGN AND WORKMANSHIP. 165 railways, the area of metal required causes a f-inch or more plate to be used ; consequently, any apparent advantage the narrow-flange system possesses, on the ground of thicker plates and less possible corrosion, because of the less surface exposed to corrosive influences, does not then accrue. It is an advantage if all the plates are of the same pattern, and to use one thick plate in preference to two, and none less than f inch in thickness, and to build up any section with the least number of plates that due consideration of strain and manufacture may require. Although T 3 ^-inch plates are used in plate-built structures, it is better to have them not less than ^ inch where only thin plates are required to give the necessary strength. It may be generally accepted that in built-up plate flanges it is almost impossible to make them fit so closely as to prevent the ingress of air, smoke and vapour; and even if this can be done, when such a girder is erected, vibration, strain, deflection and set, &c, will prevent them being an absolutely air and water-tight mass. The lower flanges and parts of girders are particularly liable to become quickly corroded, for water, dust, and vege- table growth, ballast, earth, and the drip of locomotives, are all, more or less, there concentrated. Around the angle- irons, junction of bars, cross-girder bearings, all joints and connections, rivet-heads and holes, and any part of a bridge in which water or dust can accumulate, the point of attach- ment of J_-iron and floor-plates to girders, where any timber is in contact with the metal, as in the wooden platforms of bridges, corrosion is to be expected, because of the presence of tannic and gallic acids ; and the wood will also suffer if placed on bare iron, especially in tropical climates : in fact, all places that can be receptacles for, or are exposed to, the influences mentioned are particularly liable to corrosion. If rust has loosened the attachment of rivets, angle, T, channel, or other pieces of iron, to a flange-plate, there is no uniformity of strain, and each part in turn may be tried severely instead of all acting-together. The late Mr. Cowper, M. Council, Inst. C.E., said with respect to built-up girder- 166 CORROSION AND ITS PREVENTION. work, " He would suggest a mode of protection should be adopted, by forcing varnish or boiled oil, or whatever might be the best material, between the plates by a powerful syringe ; the oil might not go all the way in, but it would go as far as the air would, and two or three injections might entirely stop the entrance of air. In that way he believed many bridges of laminated plates, which were now in a state of deterioration, might be preserved from further deterioration between the plates. Thirty years was too short a life for a bridge, and some improved method of protection ought to be adopted." It may be well to remember, in designing, that rough and uneven, and polished surfaces, of the same metal, will not equally corrode, because it is not easy to make water adhere to a polished surface — as, for instance, to a needle or a razor — until it becomes dirty, and it then finds a place to rest, as no indentations or cavities are present to contain any rain- water or moisture, and the area exposed to corrosive influences is less than that of a rough surface ; for it is an axiom, which cannot be too much borne in mind in designing metallic structures, that the greater the extent of surface exposed to atmospheric influences, tho more powerful will be the action of air and water to corrode it. In designing, what additional thickness should be allowed for corrosion? It principally depends upon the active power of the agents of corrosion to which any structure will be subject. These must necessarily vary according to the location, material employed — whether steel, wrought or cast j ron — a nd all the other influences that promote corrosion. The amount of the corrosion of cast iron in a century, when placed in clear sea-water, has been variously estimated — by deductive reasoning from experience, observation, and the results of tests during a few years — at from x yth to T %ths of an inch, according to the quantity of graphite or carburet of iron in it ; white iron being considered to corrode the slowest. The corrosion of wrought iron, under the same conditions, was estimated at from T \jth to T %ths of an inch. However, INFLUENCE OF DESIGN AND WOKKMANSHIP. 167 such general estimates are unreliable, for the varying quality of the iron or steel, local conditions, and the intensity or comparative absence of corrosive influences, render any such deductions almost worthless, for in some places iron has been very seriously corroded in a very short time, reference to which has been made in Chapter VI. ; on the other hand, in some structures, it hardly appears to have commenced. To follow any attempted formula, derived by generalising from a few examples, and declaring a fixed rate of corrosion, is to be subject to a false guide. Each case and structure must be separately considered, and every circumstance be duly noted ; and that the action of corrosion generally becomes intensified by age. Perhaps the time will come when it will be the custom to add to the sectional area of the metal required to meet the calculated strain a certain percentage to allow for corrosion, and this has been done in a few cases ; for, wrought iron and steel, when very thin, are much more liable to waste away, from corrosion and the effects of vibration, than fail from the strains brought upon them. Mr. Shelford, M. Council, Inst. C.E., in the designs of the iron bridges of the Hull and Barnsley Railway, carefully considered the question of durability as against first cost cheapness, and made the sections thicker according to the exposed situation of the bridge, so as to reduce the cost of maintenance. One-sixteenth of an inch in thickness is, in some American specifications, allowed when one face only of the metal is accessible. In all metallic structures to be erected in the tropics, ample allowance for corrosion and deterioration should be made in the thicknesses, and all plates and parts should be of the same pattern, so that they can be easily and accurately fitted. It is desirable that any metallic bridge should be put together at the makers, if possible, as then there is less trouble, and no risk on erection at the site, and parts can be made to fit better, and the possibilities of corrosion are reduced. In the case of underhung cross girders, which rely upon bolts and rivets to hold them to the main girders, and to keep the 168 CORROSION AND ITS PREVENTION. floor of a bridge in position, special provision should be made to protect the connections from, corrosion, and to so arrange them that they can be easily inspected and renewed. As the upper portions of a bridge are the most exposed to carbonic, sulphuric, and sulphurous acids, generated by the fuel, steam, and heated gases of locomotives— chlorine and ammonia also probably being present either from the air or water of locomotives — it would seem that it would be advisable to have as little ironwork as possible above the rail level, and therefore to have the load on the top of girders. However, special circumstances, such as headway, must govern each design, and the load on the top of a girder system is not much used in this country, as parapets or screens are then required in many situations ; but, considered merely from the point of view of the prevention of corrosion, the system would appear to be preferable, provided the bridge floor is properly designed and maintained, so as to prevent percolation of water to the main and cross girders. The Russian Ministry of Roads, a few years ago, published some rules respecting the employment of steel in bridges, and one is that iron and steel may be used in the same structure, but in each member of a group of similar parts the same material must be adopted. Thus, the top and bottom booms of a girder, the diagonals and verticals, cross and longitudinal roadway bearers, form such a group. This is a provision against corrosion by galvanic action, but it is always preferable to employ only one metal, and every care should be taken to ensure that it is homogeneous and of the same composition. Sir G. B. Bruce, in his address, in 1877, as President of the Institution of Civil Engineers, expressed the opinion that fiat iron girders for ordinary road, river, and accommodation bridges, not for bridges of large span, where a brick or stone arch is admissible, are greatly inferior both as regards appearance and durability, and, he asked, " Where would the works of past centuries have been now had they been of iron ? " As a metallic viaduct becomes older, the cost of main- INFLUENCE OF DESIGN AND WOEKMANSHIP. 169 tenance will increase, as compared with, one constructed of stone or brick, and the charges for maintenance of an iron or steel bridge may make it advisable to construct it of stone, brick, or Portland-cement concrete. The extra strength of steel enables a lighter section to be used ; there is, therefore, a greater area per unit of section exposed to atmospheric and other corrosive influences, hence an even increased necessity for protection against corrosion. In the new Tay Yiaduct, to provide against the effects of possible corrosion, the outer wrought-iron casings of the cylinders of the piers only extend to low-water level, which is about fifteen feet below high-water mark. Inside them is a ring of brickwork surrounding the concrete hearting, of sufficient thickness to encase it if the wrought-iron skin, which was adopted for convenience of construction, should perish. Great care should be taken that any holding-down bolts are kept free from corrosion, as the stability of a pier or structure, particularly when it is severely strained, may almost depend upon their resistance. In order to enable their lower ends to be inspected, a small subway beneath their anchorage can be constructed, of sufficient size to enable a man to get along it and do repairs, the dimensions increasing according to the length of the tunnel; 2 feet 6 inches to 3 feet may be sufficient width ; the height being not less, for a very short length, than 2 feet 6 inches, and in- creasing, as may be thought desirable, according to the length of the subway. If it can be made reasonably larger, so much the better. In the Transactions of the American Society of Civil Engineers, vol. xvii., it is stated by Mr. Buck, that when the old rollers of the Niagara Suspension Bridge were in- spected, during the reconstruction of the towers, the old saddles, together with the main cables lying over them, were raised from the old rollers by means of hydraulic jacks. They were found to be embedded in cement and iron rust, the spaces between their ends and the ribs of the bed-plate 170 CORROSION AND ITS PREVENTION. being filled with this cement, which had become so firm that a chisel-bar was required to extricate the rollers. Thus each roller was lying in a trough, and any water finding its way into it could only escape by evaporation. The old rollers appear to have been placed side by side without a roller- frame, with a play of a J inch between their ends and the fixed ribs of the bed-plates. The new rollers were formed with trunnions, running in holes in the side-bars of a roller- frame, whose depth is greater than the diameter of the new rollers, and overlaps the steel-bearing plates above and below the rollers, the edges of the frame being planed to come into neat contact with the edges of the steel plates and the shoulders of the rollers ; the latter are therefore protected from water, and as far as possible from dust and corrosive influences. This shows that, in consequence of corrosion, the rollers upon which the ends of a girder may rest require to be carefully protected from rust or they may cease to act as rollers, and, in addition, cause corrosive agents to accumulate, and be active centres of decay. The bottoms of petroleum-storage tanks, when placed on earthy soil, especially that containing much moisture, soon decay, as they cannot be inspected and repaired ; and all tanks should be raised sufficiently above the ground so that the bottoms can be inspected, repaired, and painted when necessary. In the case of plates riveted together to hold oil, the rivets must have a less pitch than that necessary to hold water. This is an advantage, as the plates are likely to be held together tighter, and corrosive influences excluded. The system of Messrs. Moreland of encasing the columns, girders, and cross-bearers of a floor in Portland-cement concrete, apart from the claim that it is fireproof, is to be commended — provided the inside of the columns are filled with Portland-cement concrete after being made perfectly clean and empty — as being an excellent method of preserving the ironwork, the construction being a continuous mass of concrete. It may, perhaps, be found even more effectual in preventing corrosion than in resisting a fierce fire. INFLUENCE OP DESIGN AND WORKMANSHIP. 171 In the case of iron or steel dock-gates and caissons, vibration caused by blows of vessels produces leakage when they are constructed on the box principle, but if the water is at the same level on both sides, such shocks are very much reduced in intensity and effect. In order to lessen corrosion, all parts should be open for inspection, repair, and painting, as they are generally exposed to severe corrosive influences ; and therefore, to lessen corrosion, single skin-gates are to be preferred. It is generally difficult to keep dock-caissons and gates water-tight, as they are usually subject to considerable shaking and jarring, and therefore to somewhat rough treat- ment. In order to simplify the connection of the skin-plates, if of varying thickness when designed to resist the strains, it is desirable to consider whether they had better not be made of one thickness throughout. This can be done by arranging the position of the ribs so as to obtain the required strength. In making water-tight joints between iron and timber, it is well to remember that a small face-area of timber is easier to make water-tight than a large one. After having made the timber-end to fit the ironwork, it can be coated with Stockholm tar, or other known preservative, lines of tarred spun-yarn can be placed on the timber-face, and it can be fixed tightly to the metal by screwing up the bolts. Portland-cement grout can also be used, or tarred felt. The heads of the connecting bolts can be sunk below the surface of the timber, see Figs. 1 and 2, p. 172, and be covered with neat Portland cement, and the nuts, if possible, should be placed inside. The holes are sometimes also plugged with greenheart, run in with marine glue. If a soft substance is used for washers, such as indiarubber, on tightening-up they may turn up or be torn. The ends of the bolts sometimes have cast-iron washers, with indiarubber packings under the nuts, so as to cause them to be water-tight. If iron washers are slightly cup-shaped, so as to receive a rubber washer and hold them in place, they are much less likely to be torn, bent, or turned up at the edges. Hard rubber gene- rally resists chemical action much better than the soft 172 CORROSION AND ITS PREVENTION. material. Indiarubber containing mineral matter becomes hard and brittle if stored for several years. Joints are sometimes made by placing slips of wood, 1^ inch or so in thickness, between the metals, and by caulking. As the wood will be more or less always in a damp state, it cannot be said to be an anti-corrosive joint. To obtain thorough caulking in timber, about 3^ inches of planking is desirable. A water-tight junction between two built-up parts of an iron-plate caisson has been made by a strip of india- rubber | inch in diameter, laid in a groove at the top of one portion. The iron lock-gates of the canal at Havre are galvanised, and the rivet-heads painted with zinc- white. The steel lock-gates of Limerick floating-dock were repaired Fig.i. HORIZONTAL SECTION . SHOWING BOLT HEAD ■ % • by removing all scale and rust from the inside plates, pure Portland-cement wash being run into all the seams and joints of the cells, the utmost care being taken that it penetrated everywhere. By attention being directed to the preceding points, and by care in construction, painting, and maintenance, corrosion in dock-gates and caissons can be much lessened, and perhaps prevented. With regard to workmanship, it has an important effect on durability and the magnitude of the corrosion. Care should be taken that all plates and bars, angle, T, channel, and other members, are in close contact over the entire meeting-surface, and that there are no gaping joints ; therefore, the ends of plates, and all exposed edges and abutting surfaces should be INFLUENCE OF DESIGN AND WORKMANSHIP. 173 perfectly flat, and be planed, at least, in structures of any importance, and in all, if possible. Planing the edges of plates, so that they are in even and perfect contact, is far preferable to mere caulking, for the caulking-tool tends to separate the plates. Any plates, angle -irons, &c, forming or bearing on the flanges, should be carefully butted, squared, and close-jointed, so that they are in complete contact, especially if they act as struts, and do not depend upon the rivets to take the thrust ; and all web-stiffeners should have tight bearings against the flanges and web. In lieu of a better joint, where two uneven cast-iron surfaces, having no permanent load to sustain — as the cylinders of bridge piers — have to be joined, in order to obtain a com- paratively even bearing, and to prevent corrosion or water penetrating, and so introducing corrosive influences, it may suffice if they have tarred hemp wrapped on hoop iron around the surfaces. Rivet-holes should be truly concentric in girder-work. They are usually made about ^ inch larger than the rivet for rivets of the most usual sizes — viz., say between •§ and 1 inch in diameter — so as to admit the rivet when hot, and therefore somewhat expanded, to allow for the scale on the rivet, and want of exact correspondence in the holes, &c. All rivet-holes should be drilled, as punching the plates weakens them by straining the fibre of the metal, and they should be made so that the pieces joined come into close contact. The rivets must completely fill the holes, and have full heads, holding the plates evenly, firmly, and tightly together, and leaving no vacuities or interstices which may cause weeping of water. Unless riveting is very carefully done the rivet-holes are likely to be receptacles from which may spring active corrosive influences. The rivets sometimes have a less pitch in the top flange than in the bottom so as to prevent buckling. They also cause the plates to be so close that moisture, air, smoke, dirt, decaying matter, and other corrosive influences, cannot get between the plates. In some cases it may be advisable to make the pitch of the rivets less 174 CORROSION AND ITS PREVENTION. than usual ; for instance, a f -inch rivet, with a pitch as little as 2 inches to 2| inches, in order to make the plates or pieces joined more air and water-tight. Corrosion in girders, as in toilers, is generally active along the joints, seams, and lines of rivets, unless they are well protected and the riveting properly effected. The difference between hand and machine- riveting, which causes the latter to be so much more reliable, is that the effect of the blows in hand-riveting is chiefly confined to the particles nearest the hammer, while machine- riveting has the great advantage that it communicates the force to the interior metal of the rivet — which should be wholly heated and be hot when riveted — and so it completely fills the hole ; but care should be taken to avoid any bulging or splitting of the metal, or corrosion will be accelerated. Especially when the rivet-holes in the plates are badly made, it has been shown by experiment machine-riveting will cause a rivet to completely fill a hole, whereas when it is done by hand, it only partly does so, and vacuities occur, and the heads may be misshapen and misplaced. Moisture, in time, will reach such voids, and then corrosion of the rivet and plates will quickly proceed, and the whole structure will lose stiffness and rapidly deteriorate, consequent upon cor- rosion, looseness of the parts, and the mechanical wedge- action of the badly-fitting rivets. Machine-riveting, from its uniform pressure and certainty of action, is therefore to be preferred to hand-riveting. Experiments have shown that the outside fibres of rivets appear to be more strained than those near the centre. It is doubtful whether a number of plates are generally made sufficiently tight at the edges to permanently exclude air and vapour. If they are not, they become conduits to convey moisture, air, smoke, and other corrosive influences, to the mass. In built-up girders, every effort should be made to keep the metal from air and moisture by accurately-fitting work, and therefore it is absolutely necessary the riveting should also be excellent. 175 CHAPTER XIII. CORROSION IN PILES AND COLUMNS. With regard to the corrosion of metallic columns, pillars, and piles, "which are so largely nsed, their preservation from corrosion and decay is of much importance. As their thick- ness, when they are cast iron, is usually from ^ to 1^ inch, and generally does not exceed 1 inch, any decrease in the strength of the metal is not to be lightly regarded, for, when a column or pile is erected, the interior cannot be seen, and in any case comparatively little can he done to preserve it internally, or arrest any oxidation of the inside surface. Their outer surfaces can usually he inspected and protected, but attention is especially required to be directed to the inner surface, so that, although it is hidden when the column is placed in position, it is known that at least every reasonable precaution has been taken to prevent corrosion and decom- position. For cast-iron submerged piles, the cylindrical form has, among other advantages, that of having no angles or abrupt bends ; accord in the order and direction of crystallisation ; greater homogeneousness of the metal, and therefore more immunity from imperfections ; and the important one that whatever the direction of a blow the same amount of surface is exposed, therefore any corrosive or abrading effect is likely to be uniform. There is, however, a disadvantage in ordinary hollow columns, because, although they may be periodically painted and maintained exteriorly, their interior surfaces cannot be inspected, and therefore their freedom from corro- sion must be to a certain extent a matter of conjecture or 176 COKEOSION AND ITS PEEVENTION. deduction from the behaviour of similarly exposed columns under the same general conditions. This being the case — although the circular is undoubtedly the strongest form for the amount of metal — in order that all the surfaces of the column may be inspected, and also, in special cases, the cruciform section may have to be preferred, the latter fre- quently being the handier for attaching or conforming to the internal fittings of a building. By preference there should be no inside obstruction, and never more than is absolutely necessary in columns, pillars, or piles, unless they are filled with impermeable Portland-cement concrete or protective filling, in order that saw-dust, wood-chips, dirt, debris, and corrosive agents may not prevail. It is advisable to avoid columns that consist of parts joined together by bolts. If compound piles lave to be adopted, it is better to use those which are firmly riveted together with rivets that completely fill the drilled holes, and whose flanges and rivets can be inspected on the outside ; always endeavouring to have as few pieces as pcssible, and to avoid all complicated sections, for they are not only difficult to preserve from corrosion, but their strength depends upon their perfect fitting and joint action, which may or may not be the case, and cannot be declared as certain, the probabilities being that, except where great care is exercised, some parts are not much strained till others have been unduly so. As the age of such a structure increases, the parts are sure to set unequally, and become not so tight as when they were erected, then corrosion and strain will quickly complete the deterioration, and will demand serious attention. By requiring the riveting to be most carefully executed, so that the plates are tightly gripped, and by not having too wide a pitch of the rivets, corrosion of the plates forming a wrought-iron column will be reduced, and the strength increased. Many columns of closed cross-section possess the disadvantage that all their surfaces after erection cannot easily be examined or painted. American engineers especially, in many cases, have therefore adopted angle, COEROSION IN PILES AND COLUMNS. 177 channel, and trough-shaped sections, braced with light lattice bars, for the struts of bridges. Inequalities in the thickness of the metal in cast-iron columns not unfrequently exist, although they can, by care, be cast of uniform thickness ; however, the thinner part will usually be the harder. Mr. Hodgkinson found that the external portion of a casting, whether hollow or solid, was always harder than that near the centre, the hardness increasing with the thinness, and the difference of strength was greatest in the small castings. Corrosion is, therefore, likely to be unequal, the internal surface being more easily affected by corrosive influences. Removing the skin of a casting reduces its compressive strength. - An equal H- shaped pillar or stanchion, either cast or formed of two channel irons, or one of cruciform section, can be painted easier than a circular one, but the strength is considerably less. An angle or T-iron light strut is not only stronger, but more easily painted than a channel or cruciform strut, and has a greater thickness, area for area, and is, therefore, not so quickly affected by corrosion. Columns are now cast with greater precision than formerly, and uniform thickness is to be attained if the best methods of casting are adopted ; still, it is prudent to consider that a deviation of 25 per cent, in the thickness of thin columns is not improbable, and that unequal bearing and loading greatly reduce the strength, therefore all joints should be flat and smooth, exclude moisture and air, and be firmly fixed, in order that the strain may be in the line of the axis ; and in estimating the strength, an allowance should be made for continued corrosion and decomposition of the ma- terial, because the interior cannot be painted, or even in- spected, after erection. White cast-iron piles are sometimes preferred, as they are believed to be less likely to corrode than other kinds, but tough grey metal of specified good quality is generally used. Sufficient time has hardly elapsed, since the introduction of iron piles, for any decided opinion to be given, although inferences may be logically deduced. Th& U 178 CORROSION AND ITS PREVENTION. forms of cast-iron columns generally used are the circular, the +> the \-\, and sometimes the U. Columns of cruciform section are occasionally adopted when there is a heavy strain upon them, because their thickness can be visibly and re- liably ascertained ; and, as all the sides are exposed, it may be said that such a form is easier to'preserve from corrosive effects than a circular pillar, the internal surface of which is hidden, but the even distribution of the load is not so easily attained. As it has been found that box girders closely riveted and caulked like a boiler, and thus made water-tight, and properly painted before erection, are most effectually preserved from corrosion, it would appear that wrought-iron columns are to be preferred to cast-iron for positions in which a heavy strain has to be borne, and with the view to obtain resistance to corrosion. Wrought-iron or steel columns or supports are quickly replacing, in modern railway structures, cast-iron columns, owing to the accidents that have occurred through flaws or local weakness in castings. The effective protection of the plates from corrosion will, however, require attention. It is sometimes assumed, because a column is placed upon a column with an intervening cover to it, that it is air-tight and vapour proof, or almost so, and cannot corrode. This is a sanguine assumption. The corrosive influence of any water, moisture, sawdust, wood-chips or debris, that during construction may accumulate inside the column, has to be considered. In every case the interior should be thoroughly cleared and cleaned, and all slime, dirt, and debris be removed, but care should be taken that the silicious skin of cast iron is not disturbed. The interior of piles is seldom sufficiently large to be painted by hand, and the bottom of a column has, therefore, been closed, and paint poured in and allowed to remain for some time ; it is then opened and the surplus paint permitted to run out. This process requires more paint, but less labour is necessary. A method of coating to be preferred is that of Dr. Angus Smith's black enamel for protecting water-pipes, &c, or a boiling coal-tar CORROSION IN PILES AND COLUMNS. 179 and asphaltum preparation, applied hot as soon as possible after casting, the castings being dipped in a bath of it for not less than thirty minutes. The interior surface should always be coated with an anti-corrosive composition or paint. All the cast-iron piles of the Morecambe Bay Viaduct are protected by a preparation of tar and asphalt, the tar being heated in a large tank, and each piece of ironwork boiled in the mixture for half an hour. It has been very effective. A solution of coal-tar, lime, and sand, poured in at the top, was used to fill the piles of the Clevedon Pier. At Woolwich Arsenal, some wrought-iron cylinders, forming a pier, were coated before immersion with a solution of coal-tar, naphtha, and resin, no rusting being allowed to commence, as then nothing would arrest corrosion except scraping and dipping the iron in an acid bath, neutralising by lime-water, and afterwards properly cleaning it as has been described. Frequently the base of a column is simply bolted to the sole- plate at the ground line, a position in which it is exposed to dust, moisture, air, and, may be, water. It would be preferable to make the joint a little distance below the ground line and bed and encircle it in Portland-cement concrete, after being properly coated with tar-asphaltum paint, not only to lessen corrosion, but to stiffen the joint, the concrete having an asphalt floor upon it, or to have the joint above the ground line. If the joint is properly made, and all corrosive influences excluded from the column, convenience and strain have then only to be considered, and the positions of the base and joints are of less consequence. Columns sometimes rest on sheet lead, or even sheet copper, which may cause serious galvanic action. A fine Portland- cement grout, or tarred felt covered with such grout, and allowed to soak into it as much as possible, would be as effectual in obtaining an even bed, and act as an anti- corrosive covering, instead of as one which is likely to be a corrosive stratum. The base of a column, whether held in position by a plate, bedded upon concrete or stone, or placed on sheet lead laid in a recess in a stone, is especially liable to n 2 180 CORROSION AND ITS PREVENTION. corrosion ■unless protected. When the base-plate upon which a column rests is held by holding-down or anchor bolts passing through it, some means should always be adopted to preserve them, for should the base become unstable, or cor- rosion be unequal — and it is seldom uniform, from various causes — the strength will be much reduced. The capital is often required to support the base of a column, and also a beam upon which are fixed the floor joists. Corrosion may then occur, not only from accumulation of dirt, but also from the timber resting upon the metal, the acids in the wood having been proved to act injuriously upon iron. A properly- composed tar-asphalt, or Portland-cement mortar cushion, or a layer of tarred felt, soaked in Portland-cement'grout, upon the metal, so that the timber does not touch it, would tend to neutralise this action, if it did not prevent it. Mr. Mortimer Evans, at the Craigmore Pier, near Rothesay, adopted the system of casing the iron piles with pipes of fire-clay secured with Portland cement. The iron is considered to be the strength of the structure, and the Portland cement the pre- servative anti-corrosive element, for it is thought the pier will be as indestructible as any that can be made. On certain coasts, such as the Brazilian, iron piles frequently become covered with oysters to a considerable thickness, and other shell and marine life. If the covering is even and continuous, and sufficiently thick to exclude air and moisture, and no issue from the marine animals dissolves or corrodes the metal, it is a protection, but if corrosion has commenced before they seal the surface, it will continue, notwithstanding such, a covering, and it cannot be a reliable protection. The screws of cast-iron piles should be of exceptionally good metal, and great care be exercised in casting, so as to prevent flaws, air-holes, and other defects, and every means should be employed to make them homogeneous. When this is done, as they are not so easy to bend, or liable to rust away, as wrought-iron screws, being thicker, they are largely used with success. In a cylinder-bridge pier,* as the hearting bears * Tide ' Cylinder Bridge Piers, and the "Well System of Foundations, published by Messrs. Spon, Strand, London. CORKOSION IN PILES AND COLUMNS. 18L the weight, the bolts and joints are not so important as in pile structures, the especial object of the cylinder-rings being to enable the pier to be erected in their interior, and to afford a protection to the hearting until it is set, and then to shield its surface from weather and currents. The joints should, however, be caulked, or so made that water cannot permeate, and the flanges and bolts should be inside the cylinder — the hearting will then cover them. The joints are generally the weakest parts of any metallic structure, and they are particularly so in piles, pillars, and columns, hence they should be reduced to a minimum, and careful provision against corrosion should be made. . Imper- fections of workmanship, inequalities of the material and bearing, may cause leaky joints in piles, then corrosion will be internally active. Metal to metal joints are considered to be preferable, but very perfect fitting is required to make an air and water-tight joint. White and red-lead joints may aid corrosion, and they cannot be regarded as perfect. As a provision against corrosion, apart from the other advantages, caulking the joints with iron cement, in order to make them water-tight, is to be commended; but, if it can be effected, a metal to metal water-tight joint is to be preferred, but this requires planing, fitting, or machine-faced ends, &c. Among the means which can be employed to obtain tight joints may be mentioned, all the bearing-places to be truly faced in a lathe or planed, so that the abutting ends are perpendicular to the axis of a pile, the bolt-holes drilled, and the bolts turned. A close fit between the bolts and holes is of the utmost importance to prevent corrosion. In designing columns or pillars, care should be taken that the joints are not so placed as to be specially liable to receive corrosive influences, as at the ground level, and where joists and planking rest against them ; for washing the planking may cause trickling or permeation of water down them, and into the joints and connections. As Mr. Hodgkinson observed, in making his celebrated experiments, that "pillars with rounded ends generally broke at the middle only ; pillars with flat ends usually broke in three places, at the middle, and near each 182 CORROSION AND ITS PREVENTION. end ; " these places, subject to other conditions, are those frequently strutted and tied. Any corrosion at them is likely to produce decided weakness, and, therefore, it is advisable to have no joints at or near them. The ordinary exposed projecting flanges, fastenings, and joints of piles for bridge piers, and landing and promenade piers, appear in not a few cases to have been designed some- what regardless of corrosive influences. Experience with such structures has demonstrated that the joint-bolts quickly corrode ; and it is well to remember, apart from other causes, that this accelerated corrosion is assisted by the different electro-chemical properties of the materials used, the columns frequently being of cast iron and the bolts of wrought iron or steel. Even if iron cement or caulking is used to fill the joints or openings, it will generally not remain long perfectly air-tight and damp-proof. If the only connection between the different lengths forming a pile, or between a pile and a screw, are the bolts, when they fail in the latter case, the column alone must support the whole load, for the screw is then no longer attached to it. None of these joints or con- nections can be considered satisfactory, so far as regards durability and resistance to corrosive influences. If the diameter of a cylinder or column will permit, joint flanges and bolts should be internal, and external flanges and bolts, should only be used if absolutely necessary, but some other COKROSION IN PILES AND COLUMNS. 183 more preferable arrangement can generally be adopted. In designing joints of columns, piles, &c, anything which tends to prevent moisture or any corrosive influence penetrating to the interior, is to be preferred; thus, the joint, Fig. 3, is better than the joint, Fig. 4, as in the latter any water or air has a direct course into the column. Figs. 5 and 6 show a simple joint for piles, sometimes adopted to avoid the use of bolts. It is much stronger laterally than that shown in Figs. 3 and 4. The casting is also stronger, as practically there are no angles, and as there are no bolt-holes, the skin of the cast iron is not disturbed. If the junction be made water-tight, it is to be regarded as a good one for the preven- tion of corrosion. In order to prevent loose joints and places for the accumu- lation of corrosive influences, and any mechanical action that may assist to wear away the ends of joints, and so cause openings and depressions, it is well if any bracing to piles or columns can act as struts as well as ties, unless it is certain no compressive strain can come upon a tie. The latter should not be cut or jointed. Bracing should have combined action, so as to prevent the strains being concen- trated upon one point. The bracing of metallic structures is intimately associated with that of joints, and is necessary to ensure rigidity, and to maintain the required strength, and any deterioration from corrosion cannot but be regarded as serious. Pier bracing exposed to wave action, if of light section, is of very little use for structures in the sea, for it soon becomes bent and loose by the force of the waves, or the constant vibration. For instance, at Westward Ho Pier, 5 inches by \ an inch T-iron braces were found to be bent by the force of the waves 2 inches out of a straight line in a length of 12 to 13 feet. Thin flat bars should be avoided in bracing piles and columns, as they are much more liable to contraction and expansion, owing to changes of temperature. A much more solid and rigid section should be adopted, as not only being better able to resist strain, but also corrosion, as there will be less exposed surface. Double-headed rails, 184 CORROSION AND ITS PREVENTION. 56 lbs. per yard, are used at the St. Leonards new promenade pier as struts, and there are no cast-iron lugs on the piles, the bracing being attached to the columns by straps passing round them between two collars, cast on to prevent them slipping up or down. The tie-rods, If inch in diameter, being bolted to the rails, are secured with patent lock nuts. This arrangement allows of inspection, and any corrosion can be seen and remedied. The vibration caused by passing trains on a high, thin column, is considerable, and will have a tendency to make the joints loose. Strong bracing is required to lessen this, and the consequent corrosion. The height of the piers in situations where there is ample headway, can be lessened by causing the load to be on the top of the girder instead of on cross- girders resting on the bottom flange, the height of the piers being lessened by about the depth of the girders. Bolts passing through any external flanges are especially liable to be corroded, and as the corrosion has been found to be very considerable — even as much as two-thirds of their strength has been destroyed in a short time— it is advisable to do without them, if possible, by the adoption of some other arrangement, such as a slot and stud joint, in which any pressure on the pile is directly transmitted to the thickened end of the flange, or after the manner of the boltless joint just previously described. Sleeve-bracing can also be applied instead of bracing bars joined by bolts or other connections to external flanges or lugs, which latter are weak and unre- liable, and especially exposed to corrosive influences, the difference between the safe tensional and compressive strain of cast iron also being disadvantageous. It has been sug- gested that the bolts should be made of steel in order to prevent the severe corrosion, but as corrosion of wrought iron and steel is about the same, it is difficult to see that any advantage would accrue, and, moreover, the different electro- chemical properties of the metals may set up galvanic action and accelerate decomposition, instead of diminishing it. On the Bombay, Baroda and Central India Railway, some of the COKROSION IN PILES AND COLUMNS. 185 external bolts, 1\ inch in diameter, all of which were above the water level, were found to he so corroded, after only four years' exposure, as to he useless. Internal flanges were used below the water level. "When the lateral strength of a structure depends upon the secure bolting of the flanges of the piles, which, in this instance, were 30 inches in diameter and 1 inch in thickness, the importance of their always being unimpaired is evident. Plates and flanges can be evenly and equally painted, but it is difficult to protect bolts over all their surfaces, and dipping in approved tar-asphaltum com- position, when- they are heated, is probably the best way to preserve them from corrosion. It is the want of equal protection which produces local corrosion, and causes them to be vulnerable, or sound in one place, and corroded in another. It has been proved, by experience, that when brass bolts have been used to fasten the planking or wooden sheathing of ships to the steel or iron plates of a vessel, no deterioration has resulted, whereas iron bolts in the same situation have soon become corroded. If lugs on the piles and flanges must be used for holding the tie-bolts, the holes in them should be drilled, so as to ensure that they are truly cylindrical and give a direct and equal bearing. In cooling it is sometimes found that the joint flanges break away from the lugs, which may conse- quently be only attached to the column, thereby failing to support and strut the flanges. Lugs and such projections are, in addition to being a very weak arrangement, decided aids to corrosion, and it is much better if the ties pass through a column by some water-tight device or around it by means of straps or rings passing round collars. Cast-iron lugs on columns are particularly liable to failure from a sudden strain or shock, for in a casting it is usually found that any considerable projection will have impurities in it, and that at the point where the projection or lug joins the column, the metal will be spongy and porous ; corrosion will, therefore, be especially severe wherever there are lugs and bolts, &c. The edges or outer sides of the lugs are sometimes 186 CORROSION AND ITS PREVENTION. unduly subject to stress if the pins or "bolts have not sufficient bearing area, do not fit properly, and are unequally strained ; then corrosion will complete the failure of the joint or the bracing. There can be no question that passing bracing tie- bolts through the columns, as at the Huelva Pier, designed by Sir G. B. Bruce, is a far better system than placing them through eyes in lugs cast on a column, for the lugs may be wrenched off. Care, however, should be taken that no air or moisture is thereby admitted into the interior of the column through the holes made in it for the tie-bolts, or corrosion will occur. This method, or that of passing the bracing round a pile or column between two collars, is an excellent arrangement, as a reliable union with the piles is ob- tained, the parts are rigid, and local corrosion much lessened. There is reason to believe that bolts of different diameters have not the same strength per unit of section. Mr. Brunei found, in testing some iron bolts, all made of the same iron, that a 1^-inch bolt broke at 23 tons per square inch, a 1-inch at 25 tons per square inch, and a § and |— inch at 27 and 32 tons per square inch respectively. Hence, corrosion in bolts of the larger sections may be of more relative importance than in those of smaller dimensions. Usually, when made from the same metal, square or round bars are stronger than plates by at least 3 tons for every square inch of section. Tie-bolts are sometimes used either to tie a fender pile to a wall, or to bind together a pile and a wall. The nuts and washers of such bolts should be designed so as to specially protect them from corrosion at the front and the back of the wall, for the tie-rods may be in a good state of preservation, being surrounded by the material of the wall, and their nuts, washers, heads, &c, be much corroded. If the heads are below the surface, and the recess be completely filled with Portland cement, they will be preserved from corrosive influences. See sketch of bolt, Figs. 1 and 2, in Chapter XII. In certain sea water, iron bolts are not so durable as wooden treenails, if made of exceptionally durable timber, as the blackwood of Australia, teak, &c. CORROSION IN PILES AND COLUMNS. 187 In all metallic hollow pile structures, but especially those for bridge piers, promenade or landing piers, and similarly submerged or occasionally submerged works, in order to pro- tect them from corrosion, it has to be determined whether the interior of the piles Bhall be filled with some preservative substance, be left open, or the surface be coated. With regard to filling the interior of a hollow pile, the chief object is to prevent water accumulating in the column, and fresh supplies of air, especially moist air or steam, and all corrosive agents having contact with the metallic surface. Metallic cylinders or columns that are not filled with hearting, should be so constructed that they can be thoroughly examined and painted ; this, of course, can only be done with those of considerable size, and ventilation by means of man-holes and air spaces should also be provided, and means adopted by which any moisture confined in a filled-in iron cylinder or column will be led away, and not allowed to accumulate. Every care should be taken, in filling columns with Portland- cement concrete, that the unequal contraction of cast iron and concrete by cold is provided against, also that a non- swelling concrete is used, or internal strains will occur, and may cause a column to crack. Weeping holes can be pro- vided in order to allow of the escape of water accumulated in the pile, or proceeding from an excess of moisture in the hearting, so that any water, after the concrete has set, can flow away. Cracking and bursting of a pile during frosty weather will then not occur. Apart from the expansion of the internal water during freezing, the contraction of the iron during cold weather upon a firm and compact hearting may be injurious, for if concrete filling is used, and it is solidly rammed, the contraction of the iron ring will compress it, or a column will crack or burst, but it is an excellent pre- servative against corrosion. Columns have been filled with concrete rammed tightly to protect them from rust, and to avoid painting, but unless provision is made in piles of the usual dimensions employed in piers for the contraction of the iron ring, although the column may not crack or burst, 188 CORROSION AND ITS PREVENTION. some strain must be brought upon the iron from this cause. Wooden staves, wedges, and planking have been fixed between the iron and the concrete in cylinder-bridge piers, but tarred felt is to be preferred to relieve the metal from this strain. In piles of small diameter, it is unimportant. Hollow piles are sometimes filled with clean dry sand. Before any filling is deposited in a pile, its interior should be thoroughly cleared of all dirt and debris. If clean dry sand is available, the interior of the piles may be filled with it, as it has been found that water was so excluded, but, of course, impervious Portland-cement concrete is to be pre- ferred. If concrete be used, no gravel should be mixed with it, but clean sharp sand; the Portland cement should be non-swelling, quick drying, and only just enough water be employed in mixing to enable it to properly combine with the cement, so that no moisture remains in the column. The concrete should also be rich in Portland cement, so as to prevent any water permeating the mass that may from any cause penetrate into the column, for a weak cement concrete is porous, there being capillary orifices in it, and gases can also pass through it, but Portland-cement concrete is an excellent protection against corrosion if properly proportioned and applied.* To prevent corrosion in the interior of metallic piles or columns in which the hearting does not Support the load, strength is not required, but the concrete should dry and set quickly and equably, be thoroughly homogeneous and impermeable, and be able to resist any corrosive influences to which it may be subject. All piles, columns, and pillars used in a building, or of too small dimensions to admit of inspection, can, after being properly cleaned and coated, be sealed at the top and bottom to prevent the admission of air or moisture. The hollow cast-iron piles of the Morecambe Bay Viaduct, erected in 1853, were boiled in a preparation of coal-tar and asphalt for half an hour, and the interior was simply filled with * For full information with regard to concrete, see ' Notes on Concrete and Works in Concrete,' second edition, published by Messrs. ' Spcn Strand, London. ' CORROSION IN PILES AND COLUMNS. 189 clean sand. Mr. Brunlees, the engineer of the viaduct, found that water was thus excluded, which is necessary, or the piles would prohably burst in frosty weather. In screwing hollow cast-iron screw piles having open ends, not only to lessen the chance of breakage, but also to retard corrosion, all earth in the pile should be removed by a shell- auger as it proceeds, but not in advance of the bottom. Such a pile, as it has open ends, should be filled with water-tight Portland-cement concrete, and not sand, so as to exclude the atmosphere and damp. In most soils it is preferable to have a gimlet-pointed screw pile if the diameter is not too large, as then the interior is less liable to corrosive influences. Examination has shown that in iron pile structures in ordinary sea or brackish water, and, in a lesser degree, in tidal river water, such as the Thames, corrosion appears to be greater in cast-iron piles at or about low- water mark than above it, and does not apparently energetically extend much below that level. 190 CORROSION AND ITS PREVENTION. CHAPTER XIV. NOTES ON CORROSION IN BRIDGE FLOORS, ROADWAYS, AND PLATFORMS. With regard to bridge floors, roadways, and platforms, and corrosion: strength, durability, lightness, and incombustibility are the chief points to be considered. In order to lessen the dynamic effect of a rolling load upon a bridge, elastic support between the rail and the girder should be provided, for if a longitudinal beam merely rests upon the upper flange of a girder, and the rail upon such timber, the dynamic effect is increased, and is considerable. The tendency in wooden platforms for railway bridges will probably be to increase their thickness under the rails. Their renewal will there- fore be more expensive, and inspection of the girders, cross- girders, and floors not so easy ; and it will be the more important to provide against the effects of corrosion. Flooring plates are sometimes placed on the bottom flanges of the cross -girders, and the permanent way is the same as that on the formation of cuttings and embankments, and is unconnected with the platform ; the ballast being purposely of extra thickness, even as much as 18 inches, so as to counteract the momentum of a train being suddenly brought to rest, and the transmission of such a thrust direct to the girders, piers, or abutments. A rigid floor, such as Portland-cement concrete on a metal trough, cannot be an elastic medium and therefore any special provision against corrosion may require to be modified, in order to construct the floor so as to neutralise or lessen sucb strain. A water- tight floor to an ordinarily-constructed bridge is not easily CORROSION IN BRIDGE FLOORS, ETC. 191 obtained. In many cases zinc protectors have been fixed upon the under side of girders over footpaths, to keep the pavements and footpaths dry, as it is practically impossible to prevent water percolating through ballast and joints of timber. In many floors, even if ample provision is made for leading away the surface water, it will find the course of least resistance, and this may not be towards a drain ; conse- quently, in designing a floor, this should be considered, or corrosion may be considerable, and unequal in intensity. Corrugated steel and wrought-iron floor plates can be made nearly water-tight, but provision against corrosion is re- quired. It is not easy, however, to make the whole of a bridge floor constructed of built-up troughing permanently water-tight, unless it is covered with asphalt, or other similar coating of an anti-corrosive and impervious nature, for all metallic floor plates are subject to contraction and expansion, consequent upon changes of temperature, and therefore the joints are likely in time to admit moisture. The angle or T-irons to which the floor plates are at- tached should have sharp, clearly-defined angles, so that the plate edges can very closely rest against them, and not have round angles, as then they cannot be made to fit so well, and each will form a ready receptacle for active corrosive influences. It is necessary to exercise considerable care in order to make the edges of the floor plates perfectly flush, and to tightly fit the T-iron bearers, and make them rest evenly. If this is not done, rapid corrosion of the rivets, T-irons, and ends of the floor plates will almost certainly follow. Everything should fit perfectly, and the whole floor bo as one mass. Bends and angles, unless effectively and permanently covered, form receptacles for water and sub- stances inducing corrosion. Many railway and road bridges have floor plates of corrugated wrought iron, and consist of a series of troughs. If any moisture or dirt accumulates in the depressions of the bends, actively corrosive conditions will be produced, therefore their surface should be well protected before erection, and, if practicable, a non-corrosive, 192 CORROSION AND ITS PREVENTION. impermeable material be laid upon them. If water be held by a trough, the side plates, unless completely protected by an an ti- corrosive coating, may quickly corrode, and the rust would probably gradually fall off, sink to the bottom, and form a skin or deposit of a brownish hue. Flooring plates are sometimes joined together as shown in Fig. 7, which cannot be said to be a good arrangement to prevent corrosion, as there are two top and two lower joints, the rivets can work loose, and water may permeate between the plates presumed to be always in close contact, and then corrosion will occur, and unequally. Fig. 8 is a preferable joint, for there can be but little lodgment of water on the joint plate, and there is no joint at the bottom of the trough, and, if the rivets should fail from corrosion, the trough will probably not buckle seriously; but if any of the rivets in Fig. 7 become corroded, and the plates act as single pieces, the strength is impaired and may be seriously so. In any CORROSION IN BEIDGE FLOORS, ETC. 193 case, Fig. 8 is a much better arrangement than Fig. 7, and is more easily examined. In preference to trough plates — which, unless completely filled with Portland-cement concrete or other water-tight ma- terial, are liable to hold water, or cause moisture to reach the plates — in Fig. 9 an arrange- ment, which appeared in 'The < ' • • - " + to the top of the floor plate can be but a few inches, and the ballast be laid upon it, and be as little as 6 inches in depth, so that to inspect the plate, or to repaint it with tar-asphaltum, or other anti-corrosive com- position, the ballast has but to be removed, and the sleeper supported. Any corrosion can thus be prevented. In the case of trough plates a greater depth of ballast must be used, perhaps twice as much, and deadweight is, therefore, added to a bridge for about half the extra depth, because there is no ballast on the upper portion of the trough. The junctions of floor plates or troughs with the main girders are generally likely to leak after a time, and should be carefully inspected periodically. Obviously, flooring plates which have the fewest joints are the least subject to local corrosion. Any local accumulation of water, caused by corrugated plates, stays, or deposit of any kind, should be avoided. Portland-cement concrete and asphalt can readily be made to be water-tight, but macadam, metalling of any kind, stone and wood paving, unless kept in excellent repair, and placed upon a water-tight Portland-cement concrete bed, pro- perly proportioned, made, and deposited, are not water-tight, and cannot be considered as affording protection from cor- rosive influences, drainage, or rain water, or the dripping from locomotives, &c. When porous material has to be used for the roadway, a good plan to prevent corrosion is to cover the metal with a layer of asphalt or impervious Portland- Kail way Engineer,' is shown for places where the headway is of importance, as the distance from the under side of the girders Fig. 9. o 194 CORROSION AND ITS PREVENTION. cement concrete, or an approved bituminous concrete. Bitu- minous concrete is sometimes used for the first or bottom layer, and tben Portland-cement concrete, and asphalt at top. Such a roadway covering should be water-tight if it be sufficiently elastic, and properly made. Before any asphalt or other coating is placed upon the floor plates, the whole should be scraped, swept, and made quite clean, dry, and free from all decaying or corrosive matter. This is best done in convenient lengths, so that, after cleaning, the floor plates and bearers are not exposed to the atmosphere or dirt longer than is necessary. It can be so arranged that, almost immediately after being thoroughly cleaned, the plates are coated. The asphalt can be dressed to the proper falls, and very small short pipes, or openings, covered with grating flaps, to carry off rain and surface water, can be inserted. If it is desirable to prevent leakage from a wooden platform of a bridge, in order to lessen any corrosion of the girders beneath, the flooring can be tongued, and the joints caulked, so as to make it like the deck of a ship. The floors of street bridges, owing to dung and other corrosive matter, which is not present on the floors of a railway bridge — other conditions being of a similar character — cause more active corrosion in the iron or steel main or cross-girders, or other beams upon which they may rest. Perhaps the worst floor for corrosion is planking upon timber beams, as, in addition to any acid that the wood possesses, there is the almost constant more or less damp state of such a wooden covering, for wood when perfectly dried is extremely hygroscopic. There is also the leakage of the dirty impure water and snow from the roadway to the bearers and girders. The chief redeem- ing feature would appear to be the necessity of a timber floor of a street bridge having to be frequently renewed, and consequently the girders have to be occasionally inspected. A wooden block floor on concrete is an improvement on the old planking on wooden bearers. The trough system, the metal being well coated with asphalt in every direction, is, however, much to be preferred, as the filling can be CORROSION IN BEIDGE FLOORS, ETC. 195 light, dry, and consist of damp-absorbing material, such as ashes, care being taken to obtain sufficient rigidity for the upper surface of the road, and then the roadway can be made of any desired material, and any part repaired or inspected without structural alterations, as in timber floors. In any iron or steel structure liable to be covered or be- spattered with street mud, it is advisable to know the nature of the mud to form an idea of its probable corrosive effect. Dr. Letheby (see ' The Chemical News '), in a series of experi- ments conducted under his direction in the City of London, found that the composition of mud from the stone-paved streets of the city, compared with fresh horse-dung and farmyard dung dried at 300° F., was : — Constituents. Fresh horse- dung. Farmyard dung. * Maximum organic. Minimum organic. Average. Organic matter Mineral matter 82-7 17-3 69-9 30-1 *58-2 41-8 *20-5 79-5 47-2 52-8 100-0 100-0 100-0 100-0 100 s 0 The largest amount of mineral matter is always found in the mud in wet weather, when the abrasion of the stone and iron is the greatest. It then may amount to 79 per cent, of the weight of the dry mud ; whereas in dry weather it does not exceed 42 per cent. Taking the average of all weathers, the amount of horse-dung to abraded matters is 57 per cent. In the case of wood pavement, the amount of organic matter in the dried mud was larger than in the case of the stone pavement. It was 60 per cent. The amount of moisture in the street mud varies to a considerable extent according to the state of the weather, but it is rarely less than 35*3 per cent, of the weight of the mud in the driest weather, the average in ordinary weather being 48-5 per o 2 196 CORKOSION AND ITS PREVENTION. cent., and in wet weather it ranges from 70 to 90 per cent. The mud was always so finely comminuted that it floated freely away in a stream of water, and the inference is that it would not subside to any great extent in a sewer with a moderate flow of water. This information affords a good criterion of the amount of ammonia and other corrosive agents to he met with in street surface water, also that which percolates into the soil and passes over any metal work before it flows into the sewers. Chloride of calcium has been used in Eouen for watering the principal streets, and it was found that it impregnates the soil with hygrometric matter, and maintains its humidity for a week, the economy effected being 30 per cent. It is to be remembered that any such solution produces dampness and accelerates corrosion, apart from any special effect of the chemical introduced. Sea water, or water mixed with salt, also acts similarly, but not so powerfully. Where damp sea air reaches the under side, and rain the upper portion of iron or steel bridge flooring, corrosive influences will be active. The drainage of the platforms of bridges is of importance, and the protection of the floors from the drip of locomotives, and the removal of any dust or earth that may accumulate. All ballast on bridge floors should be carefully selected, be dry, and entirely free from earth, decomposing, or friable matter, and be capable of absorbing water, but only to quickly evaporate it ; small shells, and clean washed gravel and sand, being better for this purpose than broken stone, as there is dust with the latter, unless it is carefully washed, and it frequently is not so able to withstand weather influences as the other materials mentioned. Ashes are naturally hygroscopic, and, owing to the calcining of the calcareous matter, which usually is found in fuel, they readily assimilate water by chemical action. The portion of carbon which remains unburnt acts as an oxidising and antiseptic agent in the destruction of the putrefying organic matter occurring in damp earth or marshes, and, being reduced to a fine CORROSION IN BRIDGE FLOORS, ETC. 197 powder, the ashes form a compact mass, which prevents the nitration of water, and the formation of pools. This was found to be the case in using ashes to reclaim deadly malarial swamps in some parts of the South Austrian Railway, and it is equally effective in any similar situation. By crushing, coke has been reduced in bulk by about 40 per cent., and this may be taken as a rough approximation of the proportion of pores to solid matter. With regard to asphalt, it is carbonate of lime naturally impregnated with bitumen in very variable proportions, but which for road-making should be limited to between 7 and 12 per cent. Exposed to the atmosphere, asphalt gradually assumes a grey, almost a white tint, caused by the bitumen evaporating from the surface and leaving a film of limestone. Gas tar is "decidedly the worst form of bitumen for paving purposes, as it passes from the dry to the liquid sate, and vice versa, according to the season. It is also very brittle. Bituminous coatings are frequently used so as to afford a solid and water-tight layer. Mineral tar, powdered lime, and fine baked sand have been used, but they deteriorate in time. However, by fresh coating with about equal parts of tar and powdered lime, laid on hot, as is sometimes done in reservoirs, they become solid and water-tight, but a neat Portland-cement coating is of a more permanent character, except where con- siderable elasticity is required in the coating. When as- phalt is specified for covering road or railway bridge floor plates, it is important to know that asphalt is being used, and not a mixture of ground limestone, ground slate, and Trinidad bitumen, or ground chalk, fire clay and gas tar, which is dull and black, has a tarry smell, and a hard metallic sound when struck against iron in cold weather. These are sometimes passed off as asphalt. The former is not so good as real asphalt, and is nearly as dear, and the latter becomes soft in summer and cracks in winter, and therefore admits corrosive influences. No preparation of gas tar is asphalt, nor is bitumen, shale grease, pitch from suets, or Stockholm tar. It is, therefore, always advisable to know 198 COEROSION AND ITS PREVENTION. that the asphalt comes from well-known mines, or that the quality is vouched for by experts. If the layer of asphalt is of a total thickness of § inch, it is generally sufficiently thick, and it is usually laid on in two layers. A kind of sawdust asphalt, made with coal tar and sawdust, is sometimes used to cover floor plates, instead of concrete, in order to make a very considerable saving in the weight of the covering. If spurious mixtures called asphalts are used, corrosive in- fluences must be much increased, and the composition may cease to be an antiseptic and anti-corrosive material. 199 CHAPTER XV. CORROSION IN WATER-PIPES, SEWERS, ETC. With regard to the corrosion of cast-iron water-pipes, it varies considerably, according to the locality, composition of the water, &c. In some cases, the pipes appear to be almost free from corrosion ; for instance, the cast-iron mains of the Seville Waterworks, laid in 1884, as far as can be ascertained, were free from corrosion and deposit ten years after. Their state, however, in many other cases has become most serious. The pressure of the water in the mains has been used on the Appold system to propel or drive pipe-scraping machines, having steel scrapers, through some water-mains. To show that corrosion does occur to a considerable extent in water- pipes, it has been recorded that the motion of the scraper has been arrested for about an hour until the scrapings accumulated in front of it were washed away. At Grenoble, the pipes became covered with rust carbuncles, which M. Lory, of Grenoble, found consisted mainly of hydrated oxide of iron, and contained from 5 to 10 per cent, of organic matter. It is believed that, apart from the general chemical composition of the water, which obviously will cause different effects as it varies, water containing organic matter in solution quickly attacks iron pipes. That some kinds of water do have a destructive action on iron is shown by examples. The magnitude of the corrosion depends in great measure upon the quality of the iron, see Chapter III. The external corrosion, should the earth be at all saturated with foul water or gases, but not otherwise, is likely to be greater than the internal corrosion ; however, if the water is fit for 200 CORROSION AND ITS PREVENTION. drinking, and does not contain much air in combination, internal corrosion will not be severe, provided the iron is of good quality. If bad, weak, heterogeneous, and carelessly made metal is used for pipes, it will soon corrode, and usually in tubercular form. In uncoated pipes, corrosion commences and proceeds somewhat uniformly if the metal is of good and homogeneous quality, and its degree of action gradually increases. It usually presents a rough uneven surface, although it will prevail over it, if the pipes are uncoated. If protected by composition, the surface will not appreciably corrode until the coating becomes exhausted by time, or has weak places in it forming active centres of cor- rosion, upon which bunches of rust are formed. Therefore, the efficiency of an anti-corrosive coating much depends upon its being thoroughly incorporated before being applied, and evenly spread. With respect to the deposits which form upon, adhere to, and have been removed from water-pipes, some substances are stones, lead, pieces of pipes, spades, crow-bars, peaty matter, gravel, calcareous earth and slime, pieces of wood and mussels ; and these from pipes varying from 6 inches to 24 inches in diameter, and in one case the gain in delivery of the main pipe after scraping was not less than 25 per cent., and averaged 45 per cent., and in special cases very much more, even as high as 250 per cent. Eusty pipes also contain algae and other vegetable and animal growth. Earthy salts contained in spring water also form a deposit upon the pipes. The nature of the deposit will vary according to the charac- ter of the water ; in some cases, carbonate of lime may be deposited, in others, hydrous oxide of iron, &c, &c. Mr. Mansergh, M. Council, Inst. C.E., gives the result of an analysis of iron rust and deposit, made by Dr. Harker, the medical officer, raked out from the 8-inch main pipes of the Lancaster Waterworks, laid in 1852, and scraped in 1878, as the reduced discharge demanded such clearing * : — * See • Minutes of Proceedings,' Inst. C.E., vol. Ixviii. CORROSION IN WATER-PIPES, SEWERS, ETC. 201 Chemical Examination. — Water, 64; organic matter and water of composition, consisting chiefly of brown insoluble vegetable matter, common to all soils, 23 ; mineral matter after incineration, 13 (consisting of iron protoxide 7, silica in minute crystals, silver sand 2, alumina 2, lime and other soils and loss 2, total 13) : total 100. The larger flakes of scale removed measured f inch in thickness. The material was of a light spongy brown appearance. The mean result of two examinations is given, as by evaporation the sub- stance changes in density from day to day. The scrapings were fine soil matters deposited under high water-pressure, and iron rust from the pipes. In considering the probable corrosion, it is well to remem- ber that water always has a tendency to follow along a pipe embedded in the earth, especially when its mouth is entirely submerged, and there is a considerable head of water on it, hence corrosion may be expected. Although pipes may be efficiently laid and coated, and in dry soil, leaks at the joints and through defects in the castings may make the ground in a constantly damp condition, and therefore one favouring corrosion. It is always advisable that each length of pipe should be either properly tested, to discover any concealed flaws, with a pressure in excess of that it will have to sus- tain, either at the foundry or in the trench, and all defective lengths rejected. Cast-iron pipes are usually preferred to wrought-iron for durability, and if they are coated with Dr. Angus Smith's composition soon after casting, their durability is increased, but they have rougher internal surfaces, which increase friction and the formation of deposits, and may have hidden flaws. Wrought-iron pipes are sometimes adopted simply to save the cost of carriage ; for instance, in order to save the enormous expense of transport, at Kimberley Waterworks, the pipes are of wrought iron, \ inch in thickness, instead of being ordinary cast-iron pipes. In a dry climate, where the rainfall does not permeate more than about two feet or so, if wrought-iron pipes are laid in such soil as clean sand, the 202 CORROSION AND ITS PREVENTION. opinion has been expressed that any objection on account of corrosion does not hold. Such may be the case, and can only be proved by time, but if the sand is impregnated with saline or other actively-corrosive substances, and the pipes are not coated by an approved system, it cannot be shown that serious corrosion will not result. Their lightness as com- pared with cast-iron pipes, and absence from shrinkage- strains in castings — which cause weak places, flaws, air-holes, and variations in thickness — are decided advantages, as also their ability to resist sudden shocks, such as those caused by the shutting of valves or excess of water, but their less thick- ness is not in their favour, although, on this point it can even be said the cast iron may change in character, and, further, if flaws or air-holes exist, each will be an active corrosive nucleus. Wrought-iron pipes are not often used in Scotland for subterranean petroleum conduits. In many Scotch oil- works, cast-iron pipes are adopted, but, on the other hand, cast-iron for the man-hole mountings of petroleum sheet-iron tanks has been condemned as liable to crack in frosty weather, and also by heat, and it is considered should not be used. Experienced opinion has declared that a £ inch in thick- ness steel pipe is safer than a 1-inch cast-iron pipe under similar conditions, and whereas cast iron will break on undue loads being applied, steel will only flatten or bend. A large steel water-pipe, five feet in diameter, has been recently laid in Paris. The thickness of the mild rolled steel is 0'235 inch, and is made slightly thicker than the stress limit, 3 • 8 tons per square inch, required, in order to allow for oxidation. The rivets are of mild steel, and are allowed to stand out if the heads are round, but the tubes may be lap-jointed. So far as regards strength, a steel pipe is to be preferred ; however, consequent upon the thinner section of metal, corrosion may be more serious. In some experiments made on galvanised service-pipes at Boston, it was found by Messrs. Nicholls and Eussell that the insoluble precipitate formed by the action of water on CORROSION IN WATER-PIPES, SEWERS, ETC. 203 zinc, when analysed, consisted of zinc, water, and carbonic acid, forming a zinc hydrocarbonate. Tests were also made of a pipe coated with a composition of lead, tin, and anti- mony ; and of a brass pipe. Lead and tin in the former, and zinc and copper in the latter, were found in solution. Carbonic and nitric acids are most injurious to lead. Carbonic acid generally exists in the atmosphere, and it readily attacks lead in the presence of moisture. Lead pipes are liable to be attacked by carbonic acid in a soil, if not protected by being laid in wooden troughs and surrounded with pitch, or otherwise protected. Wrought iron is attacked by any ammonia or sulphur compounds left in gas. Lead is quickly altered by contact with damp plaster. Sulphate of lead is thus formed, which is a most energetic oxidising agent. Saltpetre, which may be found in the steining of wells and damp cellars, is also destructive to lead pipes in pumps. In damp air, a bright or freshly-cut surface of lead soon becomes coated with a thin scale of grey oxide, which ad- heres closely to the metal and prevents further oxidation. Lime-water, lime-putty, or lime-mortar attacks lead, and forms a pale yellow deposit of oxide of lead in a short time, if the air is not excluded, the presence of air and moisture being necessary for oxidation. If the mortar is decidedly alkaline, the effect will be greater. This action seldom happens in water-service pipes, as the necessary conditions do not then often exist, but it may do so if lead be used for packing in engineering structures, and lime-mortar or limy compounds, or water are present. A piece of pipe, which for six years had formed the supply pipe to a fountain basin, was analysed at the chemical labora- tory at Berlin. It contained 99*05 parts of oxide of lead, the residue consisting of carbonic acid, with traces of lime and silica, which latter, it was considered, may have been due to small quantities of cement adhering to the outer surface of the pipe. The pipe was much corroded where embedded in cement, the corrosion being most marked at the end of the pipe next the basin, and the effect diminishing as the 204 CORROSION AND ITS PREVENTION. pipe receded from the water. The pipe was coated with a chocolate-coloured layer of oxide of lead of the hardness of glass, which adhered strongly to the metal. Dry Portland and Koman cement were found by Dr. Peschke to have no action upon lead ; the presence of water appears to be neces- sary to effect corrosion, although no explanation, it was con- sidered, could then be given. It has been found in Germany, in taking up some lead pipes, that they had become brittle and porous, and some of the lead had been converted into a basic carbonate white lead, and this had only occurred where the lead pipe had been in contact with mortar or cement. By experiment, Dr. Eossel found that lead buried in most earth that contains chlorides, saltpetre, and sal-ammoniac, lost weight, but to a much less degree than in mortar ; and that the sulphates, like plaster of Paris and Glauber salts, had no action upon lead — however, this does not quite agree with the experience of some others — neither had the carbonates, like chalk, soda and potash, nor the silicates, sand and clay. He considered lead pipes should never be brought in contact with any kind of mortar or cement. Mr. Eeichardt, of Jena, examined, a few years ago, the leaden pipes which had been in use for upwards of 300 years for the supply of water from springs to the town of Andernach, and found them to be coated internally with a layer, about ^ millimetre in thick- ness, of phosphate and chloride of lead, with a little free oxide of lead. Very small quantities of lime and magnesia were present in this coating, and the metal of the pipes, after being in use for this long period, was perfectly good in quality. Various experiments were made, and showed that it was important the lead pipes should always be full of water, so as to exclude the air and prevent oxidation of the lead, and not alternately emptied and filled, as, if oxidation occurred, the metal readily became soluble in water. To prevent or lessen corrosion on the outer surface of water or gas-mains, perhaps there is no more effectual plan than surrounding them with Portland-cement concrete, properly proportioned, and mixed so as to be impervious, COKROSION IN WATER-PIPES, SEWERS, ETC. 205 continuous, uniform, and an unvarying protection.* In special cases, where severe corrosion is expected, and iron pipes must be used, they can be surrounded with Portland- cement concrete, after receiving a coating of tar-asphaltum composition, and on the inside be covered with a 1 of Portland cement and 1 of washed sand coating, about f inch in thick- ness, and so manipulated as to present a smooth, glassy, and water-tight surface. The cast-iron circular culverts, 8 feet in diameter, in the Langton Dock, Liverpool, Mr. Lysterf has lined with a |-inch coat of Portland cement mixed with fine gravel, kept in place by dove-tail ribs cast at near intervals on the inner surface of the pipes (see Fig. 10). Fig. 10. Periodical examination has shown the condition of the pipes continued to be perfect. This system can be readily applied to pipes of large diameter, and could, by an arrange- ment of an annular movable board, to hold the concrete until set, be adapted for any reasonably small diameter. The conduit of the Paris new water supply consists of iron cir- cular pipes about 5 feet 8 inches in diameter, having an outside lining of masonry 8 inches in thickness, and an inside coating of f -inch cement over the lower two-thirds. Where the pipes are under pressure, they are lined inside all round. It has been found, in repairing and replacing water-mains embedded in concrete and used for the Vienna water supply, that the oxidised iron entered into an intimate combination with the concrete. If the under side of an iron or steel water-tank, for some reason, cannot be made to be accessible for inspection, it * See 'Notes on Concrete and Works in Concrete,' published by- Messrs. Spon, Strand, London. f See ' Minutes of Proceedings,' Inst. C.E., vol. c. 206 CORROSION AND ITS PREVENTION. should be closely and firmly bedded in impermeable Portland- cement concrete. An allowance for corrosion should always be made in determining the thickness of the plates. At Bordeaux, an iron service reservoir, in two basins, each 55 feet in width, 130 feet in length, and 6 feet 7 inches in depth, with semicircular ends, was constructed on columns resting upon bed stones on concrete, placed 13 feet apart longitudinally, and 18 feet 6 inches transversely, because of the very bad foundation, and the necessity of placing the bottom level of the reservoir a few feet above the ground. In order to provide against corrosion, the bottom is every- where made accessible for painting. Tie rods and stay bolts are used in cast-iron water-tanks, and, as they generally corrode quickly at the points of attachment to the tank plates, the strains on the tank be- come very severe, and are local. Cast iron is not a proper material to employ, for such a tank may explode, but wrought iron or mild steel plates should be used, as they would bend before giving way. There need be no fear of corrosion in wrought-iron or steel tanks if they are properly coated with tar-asphaltum or other really anti-corrosive composition. ZJraprotected cast-iron plates in the beds of some waterworks have been found to last but a short time, because of the action of the water upon them. The methods of making joints for metallic pipes are very numerous. In order to prevent corrosion, special care is ne- cessary to secure a tight joint. The steam-pipes of the New York Steam Distribution Company have faced joints, and are rendered steam-tight by the introduction of a gasket of thin corrugated copper. The gaskets are painted with thin red lead, thus excluding moisture. The corrugated copper gaskets were so made that the outer diameter cleared the insides of the flange bolts, and were dropped into the pipe joint from the top after putting in place two or three of the lower bolts to centre them. This much facilitated the work, and no difficulty from leakage has ever occurred at such joints, notwithstanding the very considerable erosive action CORROSION IN WATER-PIPES, SEWERS, ETC. 207 of steam jets. As an additional precaution, the gaskets were painted with thin red lead, as previously mentioned. Eed lead putty joints were first used, but they took a considerable time to complete, and portions of putty were squeezed into the pipe. A recently-adopted method of making a joint for pipes is by carefully cleaning and painting the joints, before the pipes are laid, with thin lead paint, no lead being used. Tape and red lead, and millboard and white zinc have also been employed, and special proprietary compositions of various kinds. In the case of sewer-pipes, cement, tarred gasket, and other water-tight joints are used. Joints of pipes are also sometimes well caulked with tarred gasket, the remainder of the joint being run with molten lead 2 inches deep, and properly caulked afterwards. Sewer joints are also made good with canvas and red lead. At Croydon, only lead was used for making the joints in some pipes, the old lead and gasket joints being abandoned, and the sockets made very much shorter and smaller than formerly. It is to be remembered that galvanic action may occur when iron or steel is joined by lead, and corrosion be thus accelerated. It is advisable to externally and internally inspect all metal pipes before they are coated with any preservative composition, because defects of casting are hidden by paint or tar. Dr. Angus Smith's composition appears to be univer- sally approved when properly applied. Water-pipes should be treated by being completely dipped in a bath of the varnish, so as to coat the entire surfaces inside and out, as directed by him : i.e., the pipes are to be cleaned, and at a temperature sufficient to expel as much as possible the air from the skin of the pipe, 600° P. being sometimes adopted, and to dry the varnish properly, without burning or de- teriorating it, but to produce a smooth, even and regular coating. Common gas"tar will not be effectual, as it does not combine with the iron or harden, but washes off, and the tar-asphaltum process must be used. Seven to nine years may elapse before there is any appreciable deteriorating effect upon a water-pipe coated with asphalt. If tar only is used, 208 CORROSION AND ITS PREVENTION. the pipes should be dry before being laid ; if they are wet, not only will the anti- corrosive coating be more or less a failure, but any water will be polluted, and if they are gas- pipes the moist tar will have a deleterious ^effect on the illuminating power of the gas. If a pipe be coated with the most certain preservative against corrosion, and its surface is afterwards damaged or disturbed, a vulnerable place will be caused, and corrosion will follow. By the coating being thoroughly hard before the pipes are laid, the risk of damage is much reduced. Before being connected, all iron pipes should be thoroughly coated externally and internally with Dr. Angus Smith's composition or coal-tar pitch, applied hot by immersion, or with some equivalent satisfactory substance, and the large mains can receive an internal coating of Portland cement, as previously described. The American Waterworks Association, Philadelphia, 1891, in the report of the committee on specifications for cast-iron water-pipes, recommended, " That pipes are to be thoroughly cleaned without the use of acid, heated to 300° F., and plunged into the coal-pitch varnish. When removed, the coating to fume freely, and set hard within an hour." Pipes not coated before being laid are generally soon encrusted with rust, and deposits, such as carbonate of lime, &c, depending upon the matter in suspension, the character of the water, and other circumstances. It has been recom- mended that all coatings be put on immediately after the pipes are cast, and that the pipes should be tested with oil, and not by water pressure. The great point in cast-iron pipes is not to disturb the skin, and to preserve it at once by coating it. If the metal be chipped or exposed by any other means under the skin, corrosion will occur, and may work its way under the coating or varnish of the adjacent surfaces. After about four years' use, it was found that the inte- rior of the cast-iron pipe of the Fleet Sewer crossing London Metropolitan Railway, which was coated with red lead, had suffered no deterioration from the constant flow of sewage matter through it. The paint still adhered to the sides, CORROSION IN WATER-PIPES, SEWERS, ETC. 209 although it was washed away from the invert. The steel submerged sewer, f inch in thickness, that is inserted below the bed of the Shirley Gut, which is tidal, Boston, U.S.A.,* is lined with three rings of brickwork in Portland cement, mixed in the proportion of 1 of cement to 2 of coarse, clean, sharp sand, in order to provide against corrosion. The dia- meter inside the brickwork is 6 feet 4 inches. At Yarmouth, the sewerage syphon-pipes are made of sheet iron, galvanised so as to resist corrosion. All metallic pipes for conveying sewage should have a thick, hard, and tenacious coating, applied while the pipe is hot, if neither a Portland-cement concrete, nor brickwork in cement lining, is applied, and they are not galvanised ; and only metal made to a properly- drawn specification should be used. * See ' Engineering,' March 16, 1894. 210 CORROSION AND ITS PREVENTION. CHAPTER XVI. CORROSION BY CONTACT WITH WOOD, AND IN GIRDER-SEATS AND BEDS. In works of construction, metals and woods have very often to be placed in contact, or used in combination, therefore any decay of the timber will affect the iron or steel. It is advisable to consider the effect timber may have upon the metal, and some means of preserving the wood from decom- position. " Thro' trunks of trees fermenting sap proceeds, To feed and tinge the living boughs it feeds." It is generally agreed that the decay of timber is due to the sap, which, originally liquid, becomes gradually solid, and then ferments under the influence of the atmosphere, the process being greatly assisted by heat and damp. The air and moisture that are present in the cells of the wood re- quire to be replaced by an approved antiseptic solution, creosote being generally considered the best. The chief points are, therefore, to extract all the sap, make the wood perfectly dry, and then impregnate it with an approved antiseptic, to prevent r fermentation ; or, as a makeshift, if this cannot be done, by coating the timber after the sap is extracted, and the wood thoroughly dried. In engineering structures, it is seldom air and moisture can be excluded, hence the decay of wood will proceed and promote corrosion. Wood treated with an antiseptic which has no deteriorating influence on iron will have a decidedly less corrosive effect on iron or steel than ordinary timber, CORROSION BY CONTACT WITH WOOD, ETC. 211 even if it be dried before being used. The decomposition and fermentation of the sap, when wood is exposed to the action of warm, damp air, will continue unless the sap is extracted or the wood properly creosoted or treated, and any fungoid growth on the timber will have an active corrosive influence on metals. There are two conditions especially favourable to decomposition in wood : the confine- ment of water in it by paint, or a covering producing " wet rot ; " and " dry rot," in which the wood is exposed to warm air. When the deleterious effect of air on wood is extreme, dry rot, which occurs only in timber that is dead, takes place, the residue being brown or black ; and when water, wet rot, which may have commenced before the tree was felled, and the residue of which is of whitish hue. The latter is quicker in effect than dry rot. If the fibre of wood is bruised, it aids the dry-rot fungus. If timber is placed under water, and is not subject to any action of air, wood does not decay for a very long time, perhaps centuries, if it was sound and free from decaying agents when submerged, although its strength as regards strain may be decreased. By air and moisture wood becomes superficially changed, becoming of a darker colour, the process of oxidation and combustion is slow, simultaneous action of air and water being necessary for rapidity of change. Timber telegraph-posts decay most at the ground line, because of the incessant changes of moisture and tempera- ture. Some pine foundation-piles in clay, constantly wet, at Albany, N.Y., were found to be not materially decayed after fifty years, but, wherever the moisture was drawn off, the wood did not last more than twelve years. Wood that is free from large and loose knots, wind shakes, sap, brash wood, worm-holes, is close-grained and even fibred, and which possesses strength, stiffness, hardness, and durability, is the less likely to become decayed, or enable depressed or damp spots to be formed in it. A chemical analysis, however, is necessary in order to ascertain whether it contains anything that will help or cause corrosion of any p 2 212 CORROSION AND ITS PREVENTION. metal in contact with it. Wood absorbs water, and, there- fore, a girder placed upon it will always be in contact with a substance that may cause corrosion. The absorptive power varies very considerably. Some experiments have shown freshly-cut wood, felled about March, to contain by weight water to the extent of £ to \ of its total weight. Ash con- tained about 28 per cent., birch 31, English oak 35, fir 37, pine and beech 40, elm 44*5, red deal 45, larch 48, Lom- bardy poplar 48, white poplar 50 • 6, and black poplar 52 per cent. A large quantity evaporates, but on the hygroscopic state of the atmosphere and that of the wood approaching equality, drying takes place very slowly, and is never complete under such conditions. Air-dried timber usually contains 15 to 20 per cent, of water.* Timber imported from warm climates should be examined as soon as possible after it is unloaded, in order that it may be known whether it contains any water possessing a de- composing or putrefactive power ; because, if so, although not apparently so, it is on shipment in a state of decay, and the damp, warm air of the hold of a ship will help its develop- ment, hence the value of thorough drying and expulsion of air, and impregnation with creosote according to the most approved manner. Although tannic acid in wood acts deleteriously on iron and steel, it is considered to have a preservative effect on wood ; for instance, the Quebracho Colorado wood of North Argentina, which is of blood-red colour, is exceptionally durable and proof against insects, and this is attributable to the large proportion of tannin it contains, said to be from 15 to 20 per cent, of its weight, which is about 78 lbs. a cubic foot. Uncoated or bare iron quickly decays wood in contact with it. In an able article in ' Engineering,' August 20, 1875, the destructive action of iron on wood is so clearly stated that it is desirable to make the following quotation * See ' Bulletin Mensuel de la Societe des anciens Eleves des Ecolea nationales d'Arts et Metiers,' vol. x. COREOSION BY CONTACT WITH WOOD, ETC. 213 from it. " In a damp medium, and exposed to the air, iron rusts and forms sesquioxide of iron. When in contact with organic matters, the sesquioxide yields its oxygen to the hydrogen of the wood, and passes into the condition of pro- toxide. But this combination rapidly absorbs oxygen from the air, and produces a new oxidation in the wood. The sesquioxide, under these conditions, is really a reservoir of oxygen, which, on one hand, discharges itself at the expense of the wood, and, on the other, fills itself at the expense of the air, and the wood is gradually consumed. Under such conditions, sulphate of copper is more destructive than use- ful, because, as soon as it is in contact with the iron, it becomes transferred into a salt of iron, which is of itself very destructive." The tannin and the wood-gum or resin in wood, some recent experiments have shown, are not attacked by the dry- rot fungus, but the coniferin and cellulose are absorbed. In Alsace, it is customary to specify that only raft timber, i.e., that floated down the river, shall be employed, as it has been found timber thus immersed is no longer liable to the attack of dry rot. The reason is considered to be, that the water slowly dissolves out the albumen and salts, and thus deprives the fungus of the nutriment needful for its develop- ment. Experiments, by burying fresh sawdust in damp earth, confirm this, for it rots away in a few years, but sawdust, which has been soaked for some time in water and has been thereby deprived of soluble matters, will remain in the ground, under similar circumstances, apparently unchanged, and only slightly tinged on the exterior with earthy matters dissolved from the soil. It should be borne in mind that if timber is painted when it is sappy, or before it has been properly seasoned or dried, fermentation and decay will be more active. The sap of the wood contains easily decomposed substances, such as albumen, gelatine, gum, &c, and when warm, damp air, is present to aid fermentation, the woody fibres soon become impaired. To winter-felled, seasoned, sound, and air-dried timber, paint 214 CORROSION AND ITS PREVENTION. is a protection, but to sappy wood it is not equally so, nor to timber not thoroughly dry, but effective air-drying has been known to require four to six years, and therefore an artificial drying process has to be employed. Paint prevents moisture penetrating to act as a reagent to decompose the albumen of the wood. Planed and worked surfaces are in some well-directed works merely oiled two or three times, which prevents cracking and drawing, and are not painted when used indoors. As a rule, it will be found that the baulks forming the main support of a temporary bridge or timber platform decay the quickest. Anything which causes condensation of moisture decays and corrodes the wood and the iron. Wounds in the surface of timber, caused by fastenings, nails, spikes, &c, accelerate decay. Iron in rusting has the effect of decaying vegetable fibre with which it may be in contact, and this deleterious action, being additional to that of water and air, corrosion of the iron and decay of the wood result. Driving spikes and nails, &c, and fixing bolts, removes or impairs, to some extent, any but the very hardest and most persistent protecting covering on the iron. The end of a spike or nail, or driven fastening, will almost always be bare, also the screwed portion of a bolt. Covering the heads and around them, of spikes, bolts, nails, &c, after driving or fixing, lessens corrosion, especially if done so effectually as to prevent the percolation of air, water, and dirt. This can be done by Portland cement being pressed upon the head, if it is sunk a little distance below the surface, so as to form a receptacle for it, see Chapter XII. Bare iron bolts in contact with wood rust rapidly, and, if with oak or wood having much acid in it, very quickly. The comparatively considerable power of resistance of larch to decay by mois- ture, the toughness of the fibres, its incombustible and incor- ruptible nature, and its soundness, whether exposed to the atmosphere or when buried in the earth, in comparison with many other woods, are most valuable properties. Its uses are various. In France it has been employed for water-pipes ; COEEOSION BY CONTACT WITH WOOD, ETC. 215 in Switzerland for vine-props. Its other usual applications need not be mentioned. Its durability, strength, and general resisting powers cause it to be a valuable wood to use in contact with iron or steel, in order to lessen corrosion in places where timber and iron must be in contact, rest one on the other, or be used in combination. Larch sleepers with the bark on, it is found, deteriorate more rapidly than when it is removed, which is believed to be due to the absorption of moisture in the bark. Wood having holes in it will have a zone of rottenness round the holes, which quickly spreads. If the fibre of wood is bruised, it aids the dry-rot fungus, consequently there should be as few iron or steel bolts in wood as possible, for timber united by bolts, whether under the surface of the earth or above it, is liable to active corrosive influences. It is found in wooden ships that treenails that have been squeezed in driving, become rotten quickly. They should be properly creosoted to counteract this. As holes made by the Teredo navalis in timber piles cause water and corrosive influences to penetrate to any iron bolts in them, it is well to remember that the Teredo navalis does not eat below the ground line, nor do they like sewer or gaswork water. The Teredo navalis perforates wood by mechanical means. Three circumstances are known to favour its ravages — a moderate rainfall, increased saltness of the water, and an increase of temperature. In some experiments made by the Academy of Sciences of Amsterdam, it was found that all exterior applications failed. A coat of mail, consisting of nails, was not only costly, but was not successful, for the covering of iron and rust did not arrest the ravages of the teredo. Creosoting, carefully effected, and caused to penetrate the whole of the timber, and not merely superficially, with oil of good quality, by thorough impregnation, was alone successful of all impregnations tried. Sheets of iron, copper, or zinc were effectual so long as they remained intact and undamaged. Neither the Teredo navalis nor the Limnoria terebrans will attack wood properly and wholly impregnated with good 216 COEKOSION AND ITS PKEVENTION. creosote oil. Piles have been protected in special cases, from a little below tbe ground line, by clay pipes filled in tightly with Portland-cement concrete. From many reli- able experiments in northern seas it appears, if timber be thoroughly dried till every particle of moisture is evaporated before creosoting, and as much creosote be absorbed by the pores of the wood as they can be made to contain, that, when treated in the most approved manner, it is an effectual preservative against the teredo. The creosoting process of preserving the timbers of the bridges, and also the piles of the bridges crossing the numerous bays and inlets, against the ravages of the teredo was adopted on the New Orleans and Mobile Eailroad some twenty years ago, as decay of the untreated timber took place very rapidly in the long warm seasons of that latitude. It was most successful, and it was found that there was no danger of fire from the impregnated timber as compared with untreated wood, but the reverse. To show the varying character and increased tendency to cause fouling and decay, and the greater virulence of the attack of marine worms in warmer climates than Great Britain, it may be mentioned that the recent (1885) report of the Committee of the American Society of Civil Engineers mentioned that, although from 10 to 12 lbs. of creosote per cubic foot of timber is sufficient as a protection in English harbours, for the higher temperatures of the sea water on the French coast, and in the southern harbours in the United States, about 19 lbs. per cubic foot are required to secure immunity from attack. A paper read before the Society of Engineers by Mr. John Blackbourn, C.E., Public Works Department, Melbourne, contains much information on ' The action of marine worms and the remedies applied in the harbour of San Francisco, California.' * The chief object of processes for preserving timber from decay is to completely seal the pores of the wood, i.e., the minute holes through which the sap circulates, with the * See ' The Engineer,' October 23, 1874. CORROSION BY CONTACT WITH WOOD, ETC 217 preservative matter injected into or formed within the pores themselves. It is advisable, whenever timber, preserved by antiseptic treatment, has to be in contact with iron or steel, to consider whether such treatment of the timber may not accelerate corrosion. Most probably it will not, but it can hardly be taken for granted, for treatment by tar process introduces tar acids, and experiments have shown that tar acids absorb moisture before finally evaporating, and it is sometimes specified that as much as 10 per cent, of tar acids is required, in others about 5 per cent., but it may be much less, even as low as 1 per cent. ; however, creosote obtained from coal tar contains tar acids to an appreciable extent. In some tests, lasting several years, by Mr. C. Coisne, for the Belgian Government, the creosote oil which had no tar acids in it gave the best results. Some were tested with 15, 8, and 7 per cent, of tar acids. Experiments have shown that wood absorbs moisture when treated with oils containing tar acids, which do not, therefore, conduce to arrest the corrosion of iron or steel. In the case of timber, metallic salts cannot be as effectual as such known antiseptics as those of the nature of carbolic acid and tar, which coagulate the albumen of the wood. Oil of tar, called creosote, is one of the products obtained from the distillation of coal tar, by which its ammoniacal liquor and other impurities are expelled, and possesses powerful antiseptic properties, hence the superiority of the creosote- oil process of preserving timber. Some of the other pro- cesses are Burnetising, chloride of zinc being used ; Kyan- ising, corrosive sublimate being employed ; Boucherising, treatment by sulphate of copper. They retard decay for a time, and then usually become redissolved and washed out by rain. Among recent processes introduced are those of impregnating the wood, previously dried, with boracic acid. Soaking timber in naphthalin : this is applied to either green wood or seasoned timber. Also that of Col. Haskin, U.S.A., by treatment of timber in its green state. It thus saves the trouble, time, and expense of drying the wood ; but 218 CORROSION AND ITS PREVENTION. whether it will be effective as a preservative, time will determine, and whether wood so treated has greater strength or not as compared with ordinary green timber, which has considerably less strength than seasoned wood. The process is simple, for after the wood in its green state is placed in an air-tight chamber and subject to an air pressure of from 150 to 200 lbs. to the square inch, according to the character of the wood, the air being first dried and heated by passing through pipes over a stove, the ultimate temperature, it is stated, varies from 200° to 450° F. All fluid matter is said to be thus caused to be insoluble and coagulated in the pores, and so prevents decomposition; however, nearly 40,000 tests of timber made at the labora- tory of the Washington University at St. Louis, under the direction of the forestry division of the Department of Agriculture, have shown that seasoned timber is about twice as strong as green timber, but well-seasoned timber loses its strength with the absorption of moisture. These indicate that green timber should not be used, and, when dried, that moisture should be excluded from it ; however, if, by treat- ment, green timber can be made as strong and durable as thoroughly sound and seasoned timber, much will be gained. Impregnating timber with creosote, after thorough dry- ing of the wood and expulsion of sap, air, and moisture, appears to be the most generally approved method of pre- serving timber. Damp or cold air is found to be injurious to the effective impregnation with creosote. An examination of best Memel telegraph posts, treated by the zinc chloride process, fixed in different earths, Mr. Preece has stated, showed that after thirteen years' use, 40 per cent, of those in sand had decayed, about 33 per cent, in clay, and about 28 per cent, in chalk. This, of course, does not show that the effect would have been the same in the case of iron or steel posts, but it does show that some soils either have a greater decaying or corrosive effect than others, or permit such corrosive and decaying influences to penetrate them. When creosote was used, only a very few piles were CORROSION BY CONTACT WITH WOOD, ETC. 219 at all decayed, about ^ per cent., and those at the top. Comparative tests were most pronounced in favour of creo- soting the piles, and Mr. Preece stated that, after thirty- years' experience, he had never seen a properly creosoted pole showing the slightest sign of decay. Testimony from almost all parts of the world bears witness that creosoting timber, when properly done with heavy creosote oil of attested quality, and not light oils of a volatile nature — the heavy oils and not the tar acids being generally considered the desirable preserving agent — is a reliable preservative of wood for engineering structures, always provided the wood was not cut or adzed as in sleepers and piles after creosoting, and the untreated or partly treated timber reached and exposed. In order to comply with this requirement, which fre- quently cannot be done in using railway sleepers, piles, bridge-floor timbers, and wharf work, the preservation in- ternally and externally has been obtained by carbolising the interior and creosoting the exterior. On the Great Western Eailway, Mr. W. L. Owen stated that in bridges and viaducts, uncreosoted yellow pine lasts 12 years; creosoted yellow pine, 20 years; creosoted yellow pine in the permanent way, 30 years ; timber prepared in any other way in the permanent way, 12 to 15 years ; Baltic sleepers, creosoted, 8 to 10 years ; ditto, not treated, 5 years. The preservative power of creosoting sleepers being fully acknowledged, on more than one main line railway, where the traffic is exceptionally heavy, it has been a question to decide whether the sleeper will not be worn out by the traffic before it has had time to decay. However, any decay of the wood loosens the fastenings, and, mechanical action being thereby accelerated, a sleeper's life is reduced when not creosoted, but the extent of the traffic influences the result. To preserve with tar it is necessary to apply it hot, and it should be heated to boiling point, but not be over-boiled, or the essential oils, which are volatile, will be lost, and they 220 COEEOSION AND ITS PKEVENTION. help the tar to penetrate the wood. Wood tar is considered by many as superior to coal tar for timber, because it pene- trates the wood more easily, and contains a large amount of antiseptic substances. If wood that will be in contact with iron cannot be creosoted, it can generally be coated with pitch oil and coal tar. When iron bolts or rods pass through a pole or post, it is advisable to char the end that will be buried, and give the wood above it two coats of Stockholm tar, put on hot, or Stockholm and coal tar mixed, or the wood will soon decay and cause corrosion of the iron- work ; or it can be blocked off the ground, scorched with burning shavings, and well tarred. But if nothing can be done to preserve the wood, decay may be lessened by simply placing the wood in the ground in the opposite direction to that in which it grew : the reason is considered to be that the capillary tubes in the tree are so adjusted as to oppose the rising of moisture when the wood is inverted. American experience with cribwork dams shows that a wooden dam should not be left hollow, but be filled with gravel, as it tends to preserve the timber, and therefore any ironwork connected with it, and it is found stone will not prevent decay of the wood. A committee of the American Eoadmakers' Association recently investigated the subject of the preservation of timber, and reported that " The theory that decay is due to atmospheric germs, finds support in the fact that timber constantly under water does not decay. That which is constantly dry decays slowly, due, probably, to the coagulation of the albumen. Timber thoroughly seasoned by heat decays less rapidly than that treated by any other mechanical process. This is due to the coagulation of the albumen. Timber subject to being alternately dry and moist decays most rapidly, owing, doubtless, to the softening and drying of the albuminous substances of the timber, thus rendering it more certain of attack by atmospheric germs." With regard to girder-seats, it will frequently be found in girders of small span resting on stone or timber, that corrosion commences about the girder beds. It is always CORROSION BY CONTACT WITH WOOD, ETC. 221 well in all small and other girders to have a hearing plate riveted to the under side of the "bottom flange, so as to preserve the entire sectional area of the bottom flange. A layer or two of tarred felt of the best quality properly placed between any bearing or wall plate is usually effectual in preventing serious corrosion of the girder-bed end, provided no depres- sion is caused in which water can lodge, or moisture, vapours, gases, &c, accumulate. Simply because of such depression, the felt may be very successful in preventing corrosion at one end, and yet not be so at the other. The remedy is to so arrange the girder beds that they can be entirely inspected, and, therefore, to slightly raise them above the level of a pier or abutment. This permits of repairs being more easily executed, and even the ends of girders being sufficiently raised for inspection without interfering with the traffic, or altering the masonry, brickwork, or concrete upon which they may rest. As an additional precaution, the felt can be soaked and bedded in Portland-cement grout. If timber has to be used for a girder bed, it should be chiefly selected for durability and absence of any deleterious effect on the metal, and especially if it be durable when alternately wet and dry. Wooden wall plates, unless well creosoted, will change in their corrosive effect on metal rest- ing upon them according to the weather, and apart from any acids or other corrosive substances present in them. The ends of girders also sometimes rest upon a pair of planed cast-iron rubbing plates, the lower of which is bolted to the masonry, asphalt roofing-felt in several layers being used upon which to bed them. An arrangement sometimes adopted is that the bearing plates are riveted to the under side of the girders, and rest directly upon the abutments, to which the girders are secured by holding-down bolts passing through oval holes, and allowing for a little adjustment or expansion of the girders, the holes being covered by large wrought-iron washers to prevent the access of dirt and other corrosive influences — this arrangement being for spans between 30 and 75 feet. In those up to about 20 feet they are simply 222 CORROSION AND ITS PREVENTION. placed upon the wall plate. The ends of girders are not unfrequently bedded upon sheet lead, so as to cover any inequalities and cause a perfect bearing. A non-conducting substance should be interposed between the lead and the iron, or galvanic action may be caused.* A modern specification thus refers to a girder seat on lead : " The girder seats shall be of cast iron, and of the quality specified for cast-iron work. The upper and under bearing surfaces must be planed perfectly true and parallel, and between the wrought-iron girder seats and the masonry must be a sheet or stratum of 12-lb. lead, covered with a sheet of vulcanite, to prevent contact between the iron and the lead." In designing bearing plates, bed plates, rollers, holding-down bolts, frames, &c, to avoid or lessen corrosion, reference to Chapter XII., 4 The Influence of Design and Workmanship with regard to Corrosion,' may be advantageous. In Part II. of this book, following this page, will be found practical information relating to the fouling, &c, of ships, pile, bridge, promenade and landing piers, and similar works of construction, anti-corrosive, and anti-corrosive and anti- fouling paints and compositions, products of great importance in the preservation and protection from corrosion, fouling, and decay of metallic structures, whether submerged or unsubmerged. * See Chapter VIII., 1 Galvanic Action and Corrosion.' 223 PART II. THE PEEVENTION OF FOULING AND C0KE0SI0N IN SUBMEEGED STEUCTUEES AND SHIPS. CHAPTER L FOULING BY SCUM, MUD, AND MARINE PLANTS. In the First Part of this volume the subject is specially- regarded with reference to engineering and other structures on land. Here, more particularly, attention is directed to the corrosion and fouling of fixed, floating, and moving structures in fresh water and in the sea, and to compositions and paints used to coat their surfaces so as to shield them from corrosive influences, and to prevent fouling by marine life of every kind. It is well to define what is generally understood by the word fouling. It may be said to mean the adherence of scum, mud, moss, grass, seaweed, subaquatic plants, mollusca, coral, barnacles, and marine life. The elements necessary for its propagation are everywhere present, and have been thus magnificently described by Milton, London's greatest poet, in 4 Paradise Lost ' : — " And God said, * Let the waters generate Eeptile with spawn abundant,' . . . 4 Be fruitful, multiply, and in the seas, And lakes, and running streams, the waters fill : ' Forthwith the sounds and seas, each creek and bay, 224 PKEVENTION OF FOULING AND COKEOSION. With fry innumerable swarm, and shoals Of fish, that with their fins and shining scales Glide under the green wave, in sculls that oft Bank the mid sea : part single, or with mate, Graze the \ sea- weed their pasture, and through groves Of coral stray, or sporting with quick glance Show to the sun their wav'd coats dropt with gold ; Or in their pearly shells at ease attend Moist nutriment, or under rocks their food Iu jointed armour watch : on smooth the seal And bended dolphins play : part, huge of bulk, Wallowing unwieldy, enormous in their gait, Tempest the ocean : there Leviathan, Hugest of living creatures, on the deep Stretch'd like a promontory, sleeps or swims, And seems a moving land, and at his gills Draws in, and at his trunk spouts out a sea. Meanwhile the tepid caves, and fens, and shores, Their brood as numerous hatch from the egg that soon Bursting with kindly rupture forth disclosed Their callow young." . It is necessary to consider the nature of the substances and life which become attached to a submerged structure, whether fixed or floating, and the means.employed to prevent or diminish their adherence. The formation of scum or mud can hardly be prevented, for it may be figuratively described as the dust of the sea, and, therefore, as practically impossible to avoid as dust on land, which, sooner or later, will become attached to or upon everything that is exposed. However, it may be much lessened by the hardness, polish, freedom from propensity to form attachments, and non-viscidity of the exposed exterior surface ; and also by its possessing the power of poisoning or paralysing marine life, so that infusoria, or vegetable organisms in suspension, coming in contact with the surface, are not held because of its viscidity, arrested by friction, or attracted ; but repelled. Scum usually originates from the organic or inorganic suspensory matter in water, and which is not in solution, becoming deposited. It may be vegetable slime, have the FOULING BY SCUM, MUD, AND MARINE PLANTS. 225 appearance of slaked lime or pulverised coral, be mud, or eroded matter from the shore or land, cliffs or rocks of the adjacent sea coast, and be assisted, to some extent, by some of the salts in sea water being partly precipitated. Slime may also be dust in a damp state, and frequently consists of a film of microscopic or other plant life. In fresh and brackish water there is usually little fouling, other than slime and grass, although corrosion is severe in brackish waters. It is known that if it were not for the winds, cur- rents, and tides, even the purity of the sea would not be maintained, as its saltness is insufficient to prevent the growth of fungi and animalcula. Its constant motion pre- vents this, the spores and germs being tossed about until they are destroyed and eaten by other inhabitants of the ocean, who devour every kind of organic matter which is in the sea. A familiar example of the effect of but few currents in the sea on the formation and growth of seaweed is the Sargasso Sea, the immense tract of weed in the mid North Atlantic Ocean. It is a characteristic of stagnant water, whether fresh or salt, that it is generally full of life, and, if fresh water, it contains insects, &c, which lay their eggs in the leaves and plants floating on the surface ; but as the sea is so vast, that which is appreciable in stagnant fresh water is not so in the sea, because of dispersion, except in certain places. Infusoria, the wonderful and extremely minute enclosed granular gela- tinous creatures made known by the aid of the microscope, inhabit stagnant water, whether fresh or salt, in which plants are growing, or in which an abundance of decayed vegetable or animal matter is contained. Their food consists partly of vegetable and animal decomposing matter, and they also prey upon each other. They occur in immense quantities in the water of the ocean, and contribute much to the nourishment of animals of a higher order, particularly in the ocean in high latitudes, where vegetable life ceases to be represented, but animal life is still in abundance. Infusoria are there Q 226 PREVENTION OF FOULING AND CORROSION. found in inconceivable numbers, and form a principal food of the fishes inhabiting those regions. Some of the infusoria are able to withstand very considerable changes of tempera- ture without losing their vitality, even from as low as 8° F. to as high as 260°. When scum becomes attached to a painted surface it may dissolve the paint. This may be called the worst case. A favourable condition may be considered to be that in which the scum and mud are very thin, and easily washed off, leaving the surface of the paint but slightly soiled. A thin scum or layer of mud, if uniform, complete, and not deleteri- ously affecting the paint underneath it, may act as a pro- tective covering; although any slight usefulness that may so accrue will probably be more than balanced by the at- traction afforded by its formation— for, when suitable soil is produced for any marine life it will soon discover it and begin to grow thereon. On a round surface a film or covering of scum or mud is more quickly deposited, or gathers sooner than upon pointed or square edges, chiefly because it offers less opposition to the action of the waves; and therefore friction, abrasion, agitation of the water, and concussive action are reduced. Principally from these causes, the for- ward portions of a ship, except when a vessel remains fast in harbour, are usually cleaner than those nearer amidships and aft— especially so in the case of wall-sided vessels. If a ship's bottom has a semicircular form, which offers the least friction for a given displacement, it will usually have more scum or covering of mud on it than one moulded on triangular lines, which, however, afford the most stability ; all the surfaces being immersed in the same waters. Examination by Prof. Tyndall, of nineteen bottles of sea water, filled under his direction between Gibraltar and Spit- head, on a voyage of H.M.S. Urgent, indicated, as stated by him at the Eoyal Institution, that the general tendency of the examination was to show that the yellowish water of coasts and harbours held in suspension a large quantity of particles; that the particles in the green water were less FOULING BY SCUM, MUD, AND MARINE PLANTS. 227 abundant and in finer division ; and that the blue water of the deep ocean was comparatively free. It is considered probable that the colour of the sea in some places is due to the presence of animalcules or plant life, intermediate be- tween the animal and vegetable kingdoms, for the appearance of the sea varies. It seems white, yellow, of reddish tinge, green, deep bluish green, purple, and even black, the different colours being due to local causes, and the grey and blue tints more especially to sunshine and passing clouds. The constant roughness or calmness of the sea has a con- siderable effect upon the magnitude and character of the fouling. Seas possessing powerful and constant currents, especially when carrying sand in suspension and at consider- able velocity— as must be the case, or it would sink — have a scouring effect, and, therefore, there will be less fouling than in tranquil waters of a similar nature ; but this may be partly counteracted by the waters of the current being warmer than the surrounding ocean, and by their containing more marine life, as witness the warmer oceanic currents and drifts. Such anti-fouling currents, as they may be called, are, however, generally found in the warmer latitudes, such as those within the Northern Warm, Tropical, and Southern Warm Zones, or between about 35° N. to about 50° S., and sometimes have a large quantity of seaweed floating on their surfaces, which also somewhat neutralises their anti-fouling action as regards a ship's bottom; for these floating seaweed meadows contain countless numbers of marine animals, ready to seek a more stable and benignant surface than that afforded them by floating on seaweed, tossed, and rent about far from the shore. On the Gu^f Stream, driftwood has been known to float from the Guff of Mexico to the western shores of Europe, and may form depositories for Crustacea and molluscs, taking them out of the usual range of latitudes in which they are found. A somewhat similar occurrence may be noticed where mounds and hills are formed of the ballast discharged from ships engaged in foreign trade, as at South Shields, for exotic q 2 228 PREVENTION OF FOULING AND CORROSION. plants may be observed growing on them which are seldom to be seen in this country, except where specially culti- vated. It is important to prevent fresh supplies of water, unex- hausted of its combined air, reaching metal, therefore even a film or covering of mud tends to preserve it, and forms a shield to defend the surface from externally active corrosive influences. However, although there is no doubt such a film of mud acts as a preservative coating, it cannot be relied upon for protection, because it is seldom uniform. Careful observation shows that form has a considerable influence in preserving submerged metalwork from fouling and corrosion, for generally on immersed rounded surfaces matter is de- posited, but on pointed or square edges such a coating is either prevented, much lessened, or removed by the action of the water, with the result that from the constant renewal and aeration of the water, a certain amount of air in the water becomes absorbed by the metal, which causes oxida- tion. . Sharp and abrupt edges are also more exposed to the force of the waves or current of water than rounded surfaces, and any protective coating is more likely to be washed off and eroded ; hence the importance of equal protection, or the result will be that while the condition of the greater portion of the metallic surface may be excellent, other parts, because of inequality in the protective covering, may be blistered, pitted, have bunches of rust, or be very much corroded ; and as the real strength of a structure is that of its weakest part, the mud coating is an unreliable preservative, although, for a time, it may be a perfect shield. This inequality in the thickness of the muddy film, by indicating those places which are least covered, other conditions being similar, will more particularly show where corrosion will probably be the greatest, but the surface will be less foul. No vegetation should be allowed to grow upon any deposit that may have formed upon the metal. The shells and means of attach- ment of marine animals also protect the surface to some extent, but unequally. In order to counteract any such FOULING BY MUD, SCUM, AND MAKINE PLANTS. 229 reduction of protection, and the increased abrasion at or about all pointed edges, their additional protection by means of an extra preservative coat, whether paint or composition of any kind, would appear to be advisable. There is one point to be considered with reference to any protection that may result from the formation of scum on the surface of a submerged structure. It is, will such a film of mud continue to be a preservative, or will it, by attracting marine life of any kind, become so changed or destroyed as to be converted into an active deteriorating influence ? It may, under certain conditions ; for, when marine plants grow thereon, the scum undergoes an action that may be described as one of combustion, resulting in perforation, however minute, and when that occurs the film loses much of any value it formerly possessed as a shield of protection. The different stages of this scum and mud covering from that of a painted surface being simply covered with white scum only, to the very foul condition of being covered with moss and grass, very thick and long, may be tabulated from the following photographic examples : — (1) Covered with white scum only. (2) Moss commenced to grow. (3) Moss about § inch in length. 230 PKEVENTION OF FOULING AND COKKOSION. (4) Moss and grass about 4 inches in length. (5) Moss and grass, very thick and long, more than 4 inches in length. (6) Moss and grass growth on chains after about eight months' im- mersion. The mnd and scum film is usually from T \th to ^th of an inch in thickness ; then moss appears, and inayjbecome of any thickness, to about 5 inches or more ; and grass and kelp, w deterioration of the paint, which it is impossible to prevenfit, commencing almost from the time a ship is afloat, or a strucc- ture submerged. This leads to the question : How long wilill a composition retain its protective qualities ? The life of thae paint is influenced, among other causes previously mentionedd, by the foulness or purity, and the temperature of the watenr, and consequently by the location of the ports and seas inn which a ship trades ; to some extent the speed of the vessel 1 ; whether almost constantly steaming, or in port for a conn- siderable time ; and if a ship passes through drifts or currenttts having sand in suspension travelling at a considerable velocity, which have a scouring effect on the paint ; or in i a nearly stagnant sea, such as the Sargasso Sea, or great sea oof weed, thus described by Humboldt : " Those evergreen masseies of Fucus natans, one of the most widely distributed of the sociaal sea plants ; driven gently to and fro by mild and warm breezees, are the habitation of a countless number of small marinne animals " ; or in comparatively cold waters, say N. of 50° NN. The protective life of a good composition, properly applieied upon an ordinary steamship, trading from, say, London Mo well-known ports in the Eastern Ocean, would probably bbe from six months to one year, and, in extreme cases, fifteen t to eighteen months ; and in those N. of 50° N., should not bbe ANTI-FOULING PAINTS, COMPOSITIONS, AND FLUIDS. 341 less than about one year ; which means that a ship's bottom would require to be scraped, cleaned, and fresh painted twice, or, say, once a year — more often the former than the latter — except when a vessel trades in northern waters. Many steamships are, however, with great advantage, docked and re-painted much more frequently. For vessels constantly in latitudes N. of 50° N., the anti-fouling compositions abso- lutely necessary in more southern waters, and possessing qualities enabling them, till exhausted by time, to repel animal or vegetable growth of every description, are not required in such strength ; although their anti-corrosive qualities are equally necessary. Therefore a composition, which may be most successful in northern and clean waters, may by no means be effectual in more southern and genial latitudes, say about S. of 50° N. It is also the same in the case of submerged structures. If a ship is docked frequently, it is no reason for using a paint which is not decidedly anti-corrosive and anti- fouling, for if it does not possess such properties, the plates will become rusty, consequently to use any composition solely because it is cheap and can be easily removed on re-docking, is illogical, for protection from corrosion must be afforded, and this, in the case of ships and submerged structures, cannot be properly attained without due preservation from fouling ; and it should never be forgotten that when corrosion has commenced, it will proceed unless all rust is removed, and the bare and solid metal coated with a really anti-corrosive and anti-fouling paint that will maintain its protective qualities in sea water. It is always much to be preferred if a paint for a submerged structure, or a ship, is anti-corrosive and anti-fouling, so that when from any cause the second 'coat becomes impaired or destroyed, the first is equally anti-fouling, or sufficiently so to prevent fouling as well as corrosion. Eespecting the numerous compounds which have been used with the idea of producing an indestructible and impregnable composition, whether for ships, submerged, or occasionally 342 PREVENTION OF FOULING AND CORROSION. wetted surfaces, it can be said that some of the most offensiwe substances have been employed, and thick liquors possessing so strong an alcoholic odour as to force the remark from a painter that " it was a crying shame to waste good liquor iin paint," the desire apparently being that so long as a compoo- sition bad an overpoweringly disagreeable or strong odomr, it would cause marine animals to make tracks in anothter direction. The delicate sense of smell possessed by the loweer animals generally has been referred to in Chapter II. Thejre is no doubt a composition, so long as it has an odour dis- agreeable to the marine life which causes fouling, has an antti- fouling effect, and therefore the property is not to be despiseed. Among the substances which have been employed maay be mentioned, white and red lead, white zinc, spirits NOTES ON PAINT SPECIFICATIONS. 365 corrosive influences of various kinds, some of little and others of special power, that ordinary provision against corrosion and fouling may be adequate or inadequate accord- ing to the special circumstances. A comprehensive perusal of the chapters of this book will enable a specification to be composed respecting the preparation of the surface to be coated, the character of the paint to be used, the method of application, &c. &c, that will suffice to secure, as far as practicable, any metallic structure from corrosive influences and fouling, which is of vital importance, for the commerce of the world and the lives and property of its inhabitants could not now be sustained or defended as they are without the existence and preservation of iron and steel ships and metallic structures. Read the First Part of this volume, on ' The Safety and Preservation of Iron and Steel Structures,' for much infor- mation relating to corrosion and its prevention in bridges, ironwork in buildings, promenade and landing piers, rails, water pipes, and all metallic structures, whether submerged, buried in the earth, unsubmerged, partly, or occasionally i submerged. 366 PREVENTION OF FOULING AND CORROSION. CHAPTEE XIII. SCAMPING TRICKS AND PAINTING. The following narrative, founded on fact, shows the manmeBr in which scamping tricks have "been carried out in paintiingjj. Two old sub-contractors met, and the following is whatt passed between them. " Well, how are you ? " "All right; I've been painting, and I'll tell you how ]I got ' extras ' out of it. The specification said, ' the iron-i- work to have three coats of good oil paint.' It did not e^vern say the metal was to be cleaned before the paint was spreadl, but we did it, more because we were afraid the stuff we ]putt on would never hold on long if we did not clean the plates a a bit, and that the paint might begin to crack and fall oflff before we were clear of the job altogether." " You mean, left it and gone to another ? " " Precisely. So we did rather more than the specificationn mentioned, although, perhaps, some of the other generail clauses would have made it somewhat awkward for us. A&s for ' good oil paint,' have you ever seen anything else ac- cording to the sellers ? " " Why, never ; of course, never. Everything is goodd that any one wants to sell, and I don't know it is not a liibebl to say it is not the best. Well, there happened to be a Sialee on at a builder's near the works, and I bought the thimgsjs there, white lead, oil, turps, and driers, and they went cheap : ; but I think the price was worth the goods. The white headd was streaky and hardly white, the oil was called ' best oil// and that was enough for me, for the price suited ; the tur- pentine was equal to it ; and what the driers and colouriingg SCAMPING TRICKS AND PAINTING. 367 matter were made of, any' one that wants had better find out. I mixed the things as I liked. My lad did it. It was a funny kind of paint, but we used plenty of oil to make it flow nice and easy, besides, we had to make it stick to the metal, but when the oil is gone, it will be like whitewash, anyhow it made it stick and look shiny. Don't ask me whether it was anti-corrosive. It suited me well enough, and ran in at about Id. a, lb." " Oh gracious ! It must have been choice." " It was to me, for it suited. You know, I believe a lot of people never think it possible for rusting to go on under the paint, but, when rust forms, it will go on, and increase till entirely removed, and really anti-corrosive paint is put on the dry, bare, pure metal surface, and not on the scale of wrought iron or steel ; besides, the paint itself may cause the metal to rust. I knew my man though, so just put it on a bit thick, which I could afford to do, and made it shine lovely." " Fill your glass. Joe, bring the lads in, for I'm going to tune up. Here is an old friend, and now I'll strike up, and this is my latest. It is entitled : — 'The Festive Painter, his Little Game.' " Three cheers for oil and lead I say, They give me smoke and beer each day. We buy it cheap to lay on thick, And add some oil if 't will not stick, Yes, rather. " Don't ask of what the lead is made, Or how much we, for it, have paid, Or if there's acid in the oil. You should not ask this son of toil. No, rather. " The driers, turps, and all the rest, Of course are just the very best. We mix them till we cannot stand, And all the skin comes off our hand. Yes, rather. 368 PREVENTION OF FOULING AND COKROSION. " We dab it on as it will go, And thin it out to make it flow. Yes ! on it goes, and on it must, For we don't care for dirt or rust. No, rather. " It's nought to us if rust works through, In pits or flakes in year or two, Or paint falls off by night and day, That is if we are miles away. Yes, rather. " We care not how the paint is made, So long as we are nicely paid, And it has got a lovely gloss, For that is sure to please the boss. Why, rather. Yes, rather. Good nighhtt." For descriptions of scamping tricks in other branchhies of engineering, see 'Scamping Tricks and Odd Knowleddjge occasionally practised upon Public Works. Chronicled frcwm the confessions of some old practitioners.' Published 1 tby Messrs. E. & F. N. Spon, 125 Strand, London. 369 INDEX TO PAETS I. AND II Accessibility of parts, 10 Acids in oil, 274-276, 288 Air and corrosion, 28, 29, 87, 88, 94-96 Algse, formation of, 234, 235 — in water pipes, 234, 235 Aluminium, 56-58 Analysis and practical use, 5 — instructions respecting, 3-5 — of paint, 280, 292 — of rust, 28, 201 — value of, in determining the corrosion, 5 Animalcula and fouling, 237 Anti-corrosive paints, 277-282, 285, 334, 335 requirements, 278, 280, 288- 290 Anti-fouling agents, 338, 340, 342- 346 — currents, 227 — influence of mud or slime cover- ing, 228, 229, 231, 232 — paints, 285, 320-358 deterioration, 230, 231 essentials of, 232, 236, 320- 324, 337, 338, 343 summary of requirements, 278, 279, 323, 337, 338, 343 Appearance of corrosion, 15 Asbestos paint, 348 Asphalte, 193, 194, 197, 198 Asphaltum paints, 305-311, 319,364 Atmosphere and corrosion, 29, 74, 75, 87, 88 Bacteria, 235 Balanus or Sea Acorn, 244, 245 Barnacles, fouling by, 243, 244 Base of piles and columns, 179, 180 Beds of girders, 220-222 Bituminous paints, 303-306, 308, 319 Blistering of paint, 265-267 Boiled oil, 275 Boilers, external corrosion, 312, 313 Bolts, 142, 169, 171, 185, 186, 357 Bracing, 183-186 — corrosion and fouling of, 356 Brackish water, fouling and corro- sion by, 225 Brickwork, metal embedded in, 152- 156 — pipes lined with, 209 Bridges, deterioration, 8-10 — floors of, 141, 190-198 — inspection, 7, 8 — maintenance, 2, 3, 5, 8-10, 168, 169 — rapid corrosion, 67, 68, 168 — rollers, 169, 170, — serviceable life, 131-151 Buildings, maintenance, 2 — rapid corrosion, 67, 68, 73, 74 Buoys, fouling, 355 Caissons, 171, 172 — fouling, 358 Calm seas, 227 Carbonaceous deposit on iron, 19 2 B 370 COEEOSION AND FOULING. Casing columns with pipes, 180 Cast iron, corrosion of, 43-51, 124, 125 relative corrosibility of, 59- 66 scale, 124, 125 water pipes, 201 Causes of corrosion, 11, 14-22 Cellulose, 349 Cheap paints, 285, 286, 292, 294, 361 Choosing a paint, 279, 280, 287, 292-294, 297, 301, 315, 319-324, 334, 335, 341, 359, 360 Cleaning ships' bottoms, in dry dock, 263 without docking, 262 Climate and corrosion, 15, 23, 69, 70, 91, 92 Coal, quality of, and corrosion, 19, 20 Coats of paints, 291, 292, 350-352 Colour of paint, 287, 355 — of rust, 13, 15 — of sea-water and fouling, 226, 227 Columns, 175-189 Compound columns, 176, 177 Compressed air, painting by, 263 Concrete, metal embedded in, 152- 156, 170 — pipes coated with, 205 — — embedded in, 204, 205 Contact of wood with iron, 210-220 Copper sheathing, 320, 339, 345 Coppering ships' bottoms, 349 Coral, fouling by, 237, 242, 247 Corrosion, analysis, 3-5 — appearance, 252 — ascertaining magnitude, 3, 5, 6 — by contact with wood, 210-222 — causes of, 11, 14-22, 252- 254 — commencement, 1-3, 12, 230, 231 — dangerous, when specially, 9 — dissimilar, 3, 4 — exceptional, 3, 6, 107, 116-123 — in girder seats and beds, 210- 222 — points of greatest, 253-255, 351 — probable magnitude of, 234, 247, 249, 341, 351 — progressive action, 12, 15, 16 Cracking of paint, 265-267, 296 Creosote, 321 Creosoting timber, 211, 215-219 Crustacea, fouling by, 234, 237-242 Currents, anti-fouling, effect of. 225, 227 Dampness, and oils, gums and varnishes, 270, 276 — and paint, 281 Dangerous corrosion, 9 — paints, 316-319 Deposits in water-pipes, 20'0 Design, 157-172 — of bridge floors, 190-193 — of piles and columns, 17 5-189 Deterioration, bridges, 8-10 — oils, 270, 271, 276 — paints, 231, 232, 236, 352 Disinfectants, 25 Dissimilar corrosion, 3 Dissolving paint, 261, 262 Dock gates, 171, 172 fouling, 358 Drainage of bridge floors, 196, 197 Dredging paint with sand, 297 Driers used in paint, 287, 295, 31 1 Drying a ships' bottom, 264 — of paints, 286, 311 Dynamic effect, 190 Electric action, 35, 36, 40, 60, 61, 109-123 Electricity and preservative coats, 312 Electro-plating ship's bottoms, 349 Enamelling iron, 315 Essential qualities of an anti-corro- sive paint, 288-290 Estimating the corrosion, 4-7, 131- 151 Euphorbium paints, 297, 346-348 Examples of corrosion, 131-151, 155, 156 — of fouling, 325-336 Experimental paints, 333-336 Explosive paints, 316 Failure, causes of, in paint, 260, 261, 264 Flaking of paint, 265-267, 295, 296 INDEX. 371 Form, 157-159, 162-164, 175-178 — influence of, 226, 228 — of piles and columns, 175-178 Foul waters, 251 Fouling, brackish water, 225 — coral, 237, 242 — definition, 223 — examples of, 229, 230, 325-336 — fresh water, 225 — greatest, 356 — magnitude of, 234, 235, 247, 249 — mollusca, 180, 237-246 — mussels, 243 — oysters, 180, 243 — saltness of the sea, 247-249 — seasons of the year, 246 — speed of ships, 255, 256, 320 — temperature of the sea, 250, 251 — universality of, 223 — unusual, 227, 246 Friction and speed of ships, 255, 256 Frost and painting, 264 Galvanic action, 109-123, 168 Galvanised sheets and pipes, 66, 202, 209 Galvanising, 314 Girder-beds, 220-222 Grass, formation of, 229, 230, 232 Gravel and wood, 220 Gums and paints, 271, 294 and dampness, 276 Hardening of paints, 286 Hardness, relative, of metals, 42 Holding-down bolts, 169, 180 Holes in timber, 214, 215 Homogeneousness, 287 — of iron and corrosion, 30-42 — of steel and corrosion, 52-56 Hydraulic mains (see Water pipes) corrosion, 314 Ice, 107, 108 — and painting, 264 Infusoria, 225, 226, 232, 235, 237 Ingredients used in anti-fouling paints, 342, 343 Inspection, bridges, 5-10 Interior of piles and columns, 178, 187-189 Joints, 154, 160, 165, 171-173, 181 — of piles, 181-183 — of pipes, 206, 207 Kelp, formation of, 230 Kyanising timber, 217 Landing stages, fouling of, 356 Lead paints, 269, 280-285, 294, 295, 334, 335 Lead pipes, 203-204 Life of anti-fouling paint, 340 — serviceable, of metallic structures, 131-151 Light, 108 Lighthouses on iron piles, fouling, 358 Lightships, fouling, 355 Lime, sulphate of, paint, 285 Limewash for walls, 307 Linseed oil, 295, 311 Localities and corrosion, 17, 19, 20, 69, 70, 91-93 Lugs on piles, 185, 186 Magnitude of the corrosion, 3-6, 131-151 Maintenance, bridges, 2, 3, 5, 8-10, 168, 169 — buildings, 2 — of a painted surface, 277, 279, 280 Making paint, 286 Marine plants, formation, 232-236 fouling by, 232-234 Masonry, metal embedded in, 152- 156 — pipes embedded in, 205 Metallic bridge-floors, 190-194 — paints, 279, 294, 334, 335 — sleepers, 68, 69, 78-80 — structures, maintenance, 2, 3 Metalling of bridge floors, 191-194 Methylated spirit and paint, 318 Mixing paint, 286 Moist paints, 311 Moisture on metal on painting, 260, 264 266 279 Mollusca, fouling by, 142, 237-247 Moss, formation of, 229-231 2 B 2 372 CORROSION AND FOULING. Moulds and moulding sand, 37, 38, 49, 50 Mud, formation of, 224-226, 228, 229, 237 — protection by, 228-232 Mussels, fouling by, 142, 243 Naphtha and paint, 317, 318 — dangerous nature, 318 Naphthalin, soaking timber, 217 Number of coats of paint, 291, 292, 350-352 Oil, danger with, 318, 319 — metals and, 270, 273, 279, 284 — in paint, 268-276, 294-296, 306, 811, 318, 334, 335 — use in paint, 269, 270, 283, 284 — wood and, 270 Oiling sandstones and bricks, 307, 308 — surface of metals, 259, 300 Ornamental ironwork, painting, 287 Oysters, fouling by, 180, 243 Paint, anti-fouling, 285, 320-358 — causes of failure, 260, 261, 264 — cheap, 285, 286, 292, 294, 361 — choosing a, 279, 280, 287, 292- 294, 297, 301, 315, 319-324, 334, 335, 341, 359, 360 — consistency of, 288 — deterioration, 231, 232, 236, 352 — explosive, 316 — making and mixing, 286 — tropics, 301, 304 — (See also Anti-corrosive and Anti-fouling) Painting, compressed air, 263 — damp surface, 260, 264, 267 — method of, 263 — object of, 288-290, 320, 321 — scamping tricks, 366-368 — smooth and rough surface, 267 — time of, 264 — weather, 264 — workmen, 260 Petrifaction, 241 Petroleum spirit and paint, 318, 319 Phosphorescence of the sesa and fouling, 242 Pickling iron, 128-130 Piers, 133, 141, 142, 169, 183,;, 184 — fouling, 355-357 Piles, 142, 175-189 Pins v. Rivets, 41 Pipe-joints, 206, 207 Pipes, 199-209 — coated with concrete, 204, , 205 — embedded in concrete, 204 1 — lined with brickwork, 209 > Pitch, 306 Platforms of bridges, 141, 1900-198 Poisons in anti-fouling paintt, 236, 302, 320-322, 340, 344, 346 ! Polished metal, 42, 166 Pontoons, fouling of, 355 Porosity of iron and steel, 38, , 39 Precautions before painting^, 257, 260-262 Preparation of the surface 1 before painting, 257-267, 291, 300 ) Pressure of water, 105-107 Prevention of corrosion, 3 Probable corrosion, 8-5 Process of corrosion, 13-21, 344, 102- 105, 137, 149, 213 Progressive action, 12, 15, 16 Properties of anti-corrosive paint, 288-290 Protection, equal, 228, 229, 2331 — marine animals, 228, 229 — mud and slime, 228, 2299, 231, 232 — permanent, 229-232 Protective coatings, special,, 315, 316 Purity of iron and steel, 30-333 Quality of ingredients in paint, 282 — of iron and corrosion, 30, 331, 38- 43, 113, 137 Quantity of paint required,, 3;52- 354 Quick-drying paints, 311 Rails, corrosion of, 21-23, 75—77' Rainfall, 94-97 Rapid corrosion, 14, 15, 23, 411, 8J7- 83, 99, 100, 105, 106, 119, 2CC6 INDEX. 373 Relative corrosibility of steel, wrought iron and cast iron, 59-66 Removal of the scale on metals, 128 -130 Resins and paints, 271, 272, 291, 311 Riveting, 41, 163, 173, 174, 176 Roadway of bridges, 190-198 Rollers of bridges, 169, 170 Roofs, 134 Rough seas, 227 Rust, analysis of 28, 148, 149, 201 — no protection, 12, 15, 16 Sand-Ctjerents, 255, 340 Sand-dredging paint, 297 Saturation, 28 Scale on metals, 49, 50, 124-130, 141 painting on, 257, 258 Scamping tricks and painting, 366- 368 Screws of piles, 180, 189 Scum, formation of, 224-226, 228- 231, 235, 237, 249, 323 — protection by, 229-231 Sea, maintenance of its purity, 225, 227 — saltness, 247-249 — temperature, 250, 251 Seats of girders, 220-222 Sea water, colour, 226, 227 composition of, 248, 249 Sea-weed, formation of, 232 fouling by, 233, 234 Selection of metal, 59-66 Serviceable life of metallic struc- tures, 131-151 Sewage, 69, 84 Sewer and drainage pipes, 208, 209 Sheathing, 320, 339, 345 Shipowner's requirements in paint, 323, 324, 352 Situation and corrosion, 15-20, 69, 70 91-93 Sleepers, metallic, 68, 69, 78-80 Slime, formation of, 224-226, 228- 231, 235, 237, 249, 323 — protection by, 228-232 Smell of paint, and fouling, 238 Soils, 84-91 Specifications, 8, 359-365 Speed of ships and fouling, 255, 256 Stagnant water, 225, 227 Steel, corrosion, 52-58 — preparation of surface, 258-260 — relative corrosibility of, 59-66 — scale, 127-129 — sewer-pipe, 209 — water-pipes, 202 Stone and wood, 220 Storage tanks, 170, 205, 206 Strained and unstrained metal, 115- 117, 160 Street mud, analysis of, 195, 196 Structures, maintenance of metallic, 2, 3, 5, 8, 9 Substances in anti-fouling paints, 341-344 Suction power in paint, 266, 279, 301, 322, 324 Superheated steam coatings, 312 Surface, influence of, 226, 232, 322 — of metal, preparation of, 257-267, 291, 300, 324 — painting, rough or smooth, 2G7, 322 Suspension bridges, 143-151 cables, paint for, 293, 310, 311, 314 Tanks, 170, 205, 206 Tar, 149, 270, 308 — and tar-asphaltum paints, 298- 300, 303-309, 319, 364 Tarring, 149, 219, 220 Temperature, 91-93 — of the sea and fouling, 250, 251 Thickness, additional, to allow for corrosion, 166, 167 — columns, 177 Tidal action and corrosion, 71, 72 Timber floors of bridges, 194 — preservation when in contact with iron, 210-220 Time of painting, 264 Tropics, paiuts for the, 301, 304 Tunnels, 24-28, 72-75 Turpentine, influence of, 295, 296 Unequal corrosion, 3, 4, 30, 31, 36, 37, 39, 46, 47, 49 Uniformity in metal, 30, 31, 33-37, 40, 41, 49 374 CORROSION AND FOULING. Uniformity of coat, 281 Unusual fouling, 227, 246 Varnish and paint, 268, 272, 273 — paints, 302-311, 322 Vegetable slime, formation of, 224 Vegetation, 18, 80, 87, 88, 91, 236 — fouling by. 228-231, 236, 237 Vibration and corrosion, 20, 21 Viscosity of water, 104, 105 Wash for walls, 307 Water, 97-108 — brackisb, 225, 249 — in wood, 212 — pressure, 105-107 — quickly fouling, 251 — sea, 225, 247-251 — stagnant, 104, 225 — still and moving, 105, 106 Water-pipes, cast and wrought iron, 201 coating, 207, 208 corrosion in, 199-208 Water-pipes, examination of, 207/ rapid rusting, 82, 83 steel, 202 Waterproofing sandstones amd bricks, 307-308 Weather and painting, 264, 265 Weight of paint, 352, 354 Welding, 34, 35 Wires, 143-151 Wood, in contact with iron or steeel, 210-220 Workmanship, 40, 41, 58, 137, 1661, 163, 173, 174, 181 Workmen, painting, 260 Wounded timber, 214, 215 Wrought iron, corrosion, 13-18, 221, 34, 44, 55, 125 preparation of the surfacce, 257-260 relative corrosibility, 55, 559- 66 scale, 124-130 water pipes, 201 Zinc-paints, 269, 281, 283, 2887, 295, 334, ; 335 LONDON : PRINTED Blf WILLIAM CLOWES AND SONS, LIMITED, STAMFORD STREET AND CHARING CROSS. ADVEKTISEMENTS. 7 HOLZAPFEL'S Antigorrosive & Antifouling Compositions FOR IRON AND STEEL SHIPS' BOTTOMS. Holzapfel's Compositions Company, Ltd. jyiARITIJVlE BUILDINGS, NEWCASTLE-ON-TYNE. LONDON, LIVERPOOL, GLASGOW, CARDIFF, WEST HARTLEPOOL, COPENHAGEN, HAMBURG, BREMERHAVEN, GENOA AND NEW YORK. Agencies and Stocks at all Principal Ports at home and abroad. To follow Index.] 2 B 4 8 ADVERTISEMENTS. WELLINGTON " "ADAMANTINE" fc- — n COMPOSITION FOR SHIPS' BOTTOMS. Is not acted upon by Salt or Frash Water, or Air. 'J 1 HIS Special Composition has been manufactured with parti- cular regard to all its constituents being subjected to no chemical change. Its adhesive power is so strong that it will not rub off readily by abrasion, or blister or flake off by exposure to very high or sudden changes of temperature, making its dura- bility as a coating unparalleled. m m It will withstand chemical vapours by which other paints are injured, rendering rust on iron or steel a practical impossibility under any circumstances. For iron, steel or wood, ADAMANTINE COMPOSITION will be found an invaluable preserver. DIRECTIONS FOR STORAGE AND USE. i, T £e "Adamantine" Iron Coating being a quick-drying liquid should be kept well corked in a dry place, and where artificial light is not used. Contents should be well shaken before use and fre- quently stirred, and only sufficient quantity taken from vessel as would be used m an hour, as it will thicken by exposure to the air Hoy ships bottoms and all wood, iron and steel work, it is onlv necessary that surface to be coated should be perfectly dry and free from all loose extraneous matter. The thin coating will be found best for first surfacing, and the thick for second; but this can be left entirely to the discretion of the user. MANUFACTURERS— ~ S. BOWLEY & SON, WELLINGTON WORKS, BATTERSEA BRIDGE, XjOnsriDonsr, s.w. ADVERTISEMENTS. 9 « ROSBONITE. Wilkinson, Heywood & Clark s HIS Preparation effectively protects Iron from Rust or injury from Sulphurous Fumes, or Vapours, or Steam. Is a Perfect Coating for Water Tanks, Double Bottoms of Ships, and all submerged Ironwork, being absolutely Watertight and Imper- vious. Is Elastic and very Adhesive. Easily applied with a suitable brush, and leaves a perfectly dry surface. ROSBONITE requires no thinnings for its application, contains no ingredient liable to decay, and when once effectively applied requires no renewal. Full information as to price of BOSBONITE, BOSBONITE PBIMING LIQUID and BB USEES to he obtained only from the Manufacturers — Wilkinson, Heywood & Clark, 7 CALEDONIAN ROAD, LONDON, N, Used by the British Admiralty. SPECIAL COATING FOR IRON. Patented and Registered by them. 10 ADVERTISEMENTS. JUST PUBLISHED. In Demy 8vo, cloth, 600 pages, and 1420 Illustrations, 6«. SPONS' MECHANICS' OWN BOOK. A MANUAL FOR HANDICRAFTSMEN AND AMATEURS. Contents. Mechanical^Drawing — Casting and Founding in Iron, Brass, Bronze, and other Alloys — Forging and Finishing Iron — Sheetmetal Working — Soldering, Brazing, and Burning — Carpentry and Joinery, embracing descriptions of some 400 Woods, over 200 Illustrations of Tools and their uses, Explanations (with Diagrams) of 116 joints and hinges, and Details of Construction of Workshop appliances, rough furniture, Garden and Yard Erections, and House Building — Cabinet Making and Veneering — Carving and Fretcutting — Upholstery — Painting, Graining, and Marbling — Staining Furniture, Woods, Floors, and Fittings — Gilding, dead and bright, on various grounds — Polishing Marble, Metals, and Wood — Varnishing — Mechanical Movements illustrating contrivances for transmitting motion — Turning in Wood and Metals— Masonry, embracing Stonework, Brickwork, Terracotta, and Concrete — Booting with Thatch, Tiles, Slates, Felt, Zinc, &c. — Glazing with and without putty, and lead glazing — Plastering and Whitewashing — Paper- hanging — Gas-fitting — Bell-hanging, ordinary and electric Systems — Lighting — Warming — Ventilating — Roads, Pavements, and Bridges — Hedges, Ditches, and Drains — Water Supply and Sanitation — Hints on House Construction suited to new countries. E. & F. N. SPON, 125 Strand, London. New York: SPON & CHAMBERLAIN , 12 Cortlandt Street. Crown 8vo, doth, 2s. 6d. SCAMPING TRICKS AND ODD KNOWLEDGE OCCASIONALLY PRACTISED UPON PUBLIC WORKS. CHRONICLED FROM THE CONFESSIONS OF SOME OLD PRACTITIONERS. By JOHN NEWMAN, Assoc. M. Inst. C.E., F.I. Inst. REVIEWS OF THE PRESS. Engineering News (New York). "This readable and interesting book is arranged as a conversation between two old sub-contractors, in the course of which they deliver themselves of numerous yarns relating to methods practised on various kinds of works, to deceive the engineers and obtain the much-desired ' extras,' thus indicating some of the points to be especially looked after in superintending the construction of works. A still more interesting and valuable feature of the book, however, is that it is full of practical hints and notes upon different methods of carrying out different kinds of work under varying circumstances, giving also advice as to the merits of the different methods." The British Architect. "We take the following stoiy from a series of amusing narratives of ' Scamping Tricks and Odd Knowledge occasionally practised upon Public Works.' " ( ii ) Industries. " This book is out of the run of ordinary professional works, inasrmuch as it is intended, not so much for the purpose of showing how puiblic works are to be carried out, as to point out some of the tricks which; are practised by those who do not wish to carry them out properly, anid to name some methods, founded on practical experience, adopted by 'sub- contractors and others to cheaply and quickly execute work. "The young engineer or inspector will find many things in the b:>ook which will at least cause him to pay attention to special points in, the different departments of civil engineering construction. Such mattters as piles, which are chiefly hidden from view, seem to require caireful inspection, and in fact all work which is covered up when the structure is completed." Indian Engineering. "This is an entertaining little book. It abounds with stories of giross cheating. The ingenuity displayed in hiding the results of some of the frauds may be useful in setting young engineers on their guard agaiinst the over-plausible." The Engineering Review. "The book should be read by all who are in charge of works of this kind." The Engineer and Iron Trades Advertiser (Scotland). " The somewhat uncommon title of this book will in itself prowe a ready attraction to the ordinary student of current literature. The 1 title page alone is characterised by a curious vein of humour. The autthor has, however, a serious and a most important object in view. " There is a peculiar charm in it not usually found in works wlhere technical details require to be recorded. The many ' dodges ' indullged in by these ideal contractors will come as 'eye-openers' to those ur.nac- quainted with the subject. We have no hesitation in saying that the volume before us is likely to serve a good purpose, and it is deserwing of a wide circulation." E. & F. N. SPON, 125 STRAND, LONDON. ( Hi ) NOTES ON OONOEETE AND WORKS IN CONCEETE. By JOHN NEWMAN, Assoc. M. Inst. C.E., F.I. Inst. Second Edition, Revised and much Enlarged. REVIEWS OF THE PRESS. FIRST EDITION. (Similar notices have appeared with respect to the Second Edition.) Engineering. " An epitome of the best practice, which may be relied upon not to mislead. " The successful construction of works in concrete is a difficult matter to explain in books. "All the points which open the way to bad work are carefully pointed out." Iron. "As numerous examples are cited of the use of concrete in public works, and details supplied, the book will greatly assist engineers engaged upon such works.'''' The Builder. "A very practical little book, carefully compiled, and one which all writers of specifications for concrete work would do well to peruse. " The book contains reliable information for all engaged upon public works. "A perusal of Mr. Newman's valuable little handbook will point out the importance of a more careful investigation of the subject than is usually supposed to be necessary." american press. Building. " To accomplish so much in so limited a space, the subject-matter has been confined to chapters. " We take pleasure in saying that this is the most admirable and com- plete handbook on concretes for engineers of which we have knowledge." E. & F. N. SPON, 125 STRAND, LONDON. 2 c ( iv ) EARTHWORK SLIPS AND SUBSIDENCES UPON PUBLIC WORKS. By JOHN NEWMAN, Assoc. M. Inst. C.E., F.I. Inst. REVIEWS OF THE PRESS. Engineering News (New York). "The book is of a practical character, giving the reasons for :slips in various materials, and the methods of preventing them, or of making repairs and preventing further slips after they have once occurred. The subject is treated comprehensively, and contains many notes of practical value, the result of twenty-five years' experience." The Builder. " We gladly welcome Mr Newman's book on slips in earthworks as an important contribution to a right comprehension of such matters. " There is much in this book that will certainly guard desigmers of engineering works against probable, if not against possible, sllips in earthworks. ^ "The capital cost of a work and the cost of its maintenance may both be very sensibly reduced by attention to all the points alluded to bv the author. J "We are glad to see that the 'author enters at some length imto the subject of the due provision of drainage at the backs of retaining walls a matter so often neglected or overlooked, and carries this subject t