to of rstt of Toronto Herbert Otto Frind Esq. WAVES OF THE SEA BY THE SAME AUTHOR. The Panama Canal and its Makers* With a Map and 63 Illustrations from the Author's Photographs. Large Crown Svo, cloth, 5s. " It is a pleasure to have a book on the Panama Canal by an English observer. Mr. Cornish is both qualified and independent. He has produced an admirable description, suited for the general reader, lucid, not overloaded with technicalities, of the right length, yet comprehensive." — Daily Chronicle. " The author of this concise account of the history, the progress, and the possibilities of a gigantic enterprise is to be congratulated on the skilful manner in which he has compressed an enormous amount of information and many valuable opinions and observations into a very small space." —Standard. LONDOX: T. FISHER UNWIN. WAVES OF THE SEA AND OTHER WATER WAVES BY VAUGHAN CORNISH Doctor of Science (Manchester Univ.), Fellow of the Royal Geographical, Geological, and Chemical Societies of London Member of the Japan Society WITH 50 PHOTOGRAPHS TAKEN BY THE AUTHOR T. FISHER UNWIN LONDON: ADELPHI TERRACE LEIPSIC: INSELSTRAS-SE 20 1910 (All rights reserved.} TO SIR CLEMENTS R. MARKHAM, K.C.B., F.R.S. VICE-PRESIDENT AND PAST PRESIDENT OF THE ROYAL GEOGRAPHICAL SOCIETY IN TOKEN OF SINCERE REGARD AND VENERATION I DEDICATE THIS BOOK V. C. PREFACE MOST of us have felt the fascination of a wave. The waves of the sea, which are the prototype of all the phenomena which we now call waves, are perhaps the most fascinating of all. Great as is the beauty of their form, the mystery of their motion is the greater charm. For while they move they live and have a being, which, like our own, is but momentarily associated with the matter of which they are formed. The wave preserves its individuality, its recognisable though not unchang- ing form, its energy, partly active, partly in reserve, whilst its material substance is constantly rejected and renewed. Of all manifestations of the inor- ganic world it is most like a living being. Yet when we watch it to its end we find none of the sad accompaniments of the exhaustion of life. It is most beautiful at the last, as it culminates to its fall and breaks in seething foam. There are two aspects of sea waves which parti- cularly attract our admiration. The first is that 8 PREFACE of a storm at sea, the second is the surf which comes in during calm weather upon a shore facing the open ocean, the breakers booming like minute guns. The first indicates the fury of the wind, but it is the second which we almost instinctively recognise as affording the best index of the greatness of the expanse of water. These things, and others of a like nature, I have watched for many years, and I have set down in this book what I have been able to add to former knowledge. I have written to inform the mind, not to stir the imagination. The appearances are familiar and recurrent, and I do not attempt to recall them by much word-painting, but the precise observation and measurement of waves, and the discovery of the mode of their production, are matters of difficulty, and to these I have given myself. My investigations pn water waves have been prosecuted during the last fifteen years. They have been made in many parts of the world^ which I have visited principally for the study of surface waves of different kinds. The chief results as far ;as they relate to water waves are contained in this volume, which contains also a critical examination of observations made by a number of seamen and others upon the size and speed of ocean waves. PREFACE 9 Most of Part III., relating to tidal bores and other waves in rivers, was published in the Geographical Journal and in Engineering in 1906, but Parts I. and II., which relate to waves of the sea, are new, and were written in 1909-10. Side by side with the observations of water waves, I have during the same period (i.e., since 1895) been investigating the progressive transverse ridges which are produced in sand and snow by the action of water or wind, and I have made observations upon earthquakes and other wave- phenomena which come within the province of physical geography. These I intend to publish later. August, 1910. VAUGHAN CORNISH. CONTENTS PART I ON THE SIZE AND SPEED OF DEEP-SEA WA VES CHAPTER I PAGE Introduction — Waves on ponds — Waves on lakes — Coniston Water — Lake of Geneva — Lake Superior — Waves in semi-enclosed seas — Western Mediterranean — China Sea .... 23 CHAPTER II Observations of waves on the North Atlantic — Conclusion as to height of North Atlantic waves — Waves on the North Pacific — The effective length of fetch of strong winds on the North Atlantic Ocean — The numerical relation between the length of fetch of wind and height of storm-waves — Waves on the South Atlantic and Southern Indian Oceans — Waves on the South Pacific Ocean — The discrepancy between wave-lengths determined respectively by simultaneous and by suc- cessive observation of wave-crests . . . 42 CHAPTER III The wave-length of the swell which reaches the shore after storms — The height of the swell at sea during storms — The co-existence of waves of different lengths — The give and take between air and water in the develop- ment of waves . . . . . -85 11 12 CONTENTS CHAPTER IV PACK The velocity of the wind at sea— The numerical relation between velocity of wind and average height of waves — The relation between the velocity of the wind and the velocity of the waves — The connection between the rate of progress of cyclones and the character of the waves — The effect of squalls and gusts upon waves — On wave-fronts in a veering wind, and on the irregularity of the waves in the region of the Trade winds — The observed profile of waves at sea — The mountainous appearance of waves . . .107 PART II ON THE ACTION OF SEA WAVES TO TRANS- PORT SHINGLE, SAND, AND MUD CHAPTER V On the depth to which wave-agitation extends, and on the transport of fine mud — The action of waves to drive shingle shorewards — The effect of percolation to pro- mote the building-up of beaches — The movement of sand by waves — The undertow . . ... . 141 CHAPTER VI On the actions which determine the 'longshore transport of beach shingle — On the causes which give rise to an arrangement of pebbles according to their sizes along the Chesil Beach and certain other beaches — On the production and maintenance of the Shambles sand-bank near Portland — The formation of patches of shingle upon a sandy beach by the action of breakers . .176 CONTENTS 13 PART III ON STATIONARY AND PROGRESSIVE WAVES IN RIVERS CHAPTER VII PAGE The flood-wave of rivers — The roll- wave in the Tees — In the Ure — In the Nikko torrent — On the advance of a flood over dry ground . . . . .221 CHAPTER VIII Tidal bores — Wave-length of the Severn bore — On the want of concordance between height of tide and height of bore, and on the conditions which deter- mine the starting-point of the bore . . . 230 CHAPTER IX Cinematographing the Severn bore . , . . 254 CHAPTER X Stationary or standing waves — Cross-stream progressive waves : observations in Niagara River . . . 265 CHAPTER XI On streams which flow as a series of roll-waves — Observa- tions on a conduit from Territet to Glion — Observations on the conduit of the Griinnbach (Merlingen,Thunersee) — Observations on the conduit of the Guntenbach (Gunten, Thunersee) . . . , . 294 CHAPTER XII The sounds of running water — The gushing motion of cataracts — Conical structures produced by the break- up of a waterfall — The wave-track of a ship . . 333 LIST OF ILLUSTRATIONS PLATES. WAVES MEETING AND CROSSING IN VERY SHALLOW WATER Frontispiece PAGE A MODERATE GALE IN THE MEDITERRANEAN . . 37 From S.S. Orizaba. Height of camera 22 feet. A STRONG GALE IN THE NORTH ATLANTIC . . -47 From S.S. Ivernia. Height of camera 40 feet. HEAVY SWELL IN THE NORTH ATLANTIC '!" . Ji. $1 WAVES AND SWELL IN THE NORTH ATLANTIC ' .' ' . 57 UNDULATING HORIZON OF A ROUGH SEA . . 65 WEATHER IN THE NORTH ATLANTIC. ' . . . 11$ THE PROFILE OF BREAKING WAVES AT EASTBOURNE . 143 WAVES BREAKING ON A SHINGLE BEACH . . • J51 A WAVE BEFORE AND AFTER BREAKING . . .. 155 WAVE BURSTING AGAINST A VERTICAL WALL . . 159 BREAKER AND BORES ON A FLAT SHORR . . . .163 THE FRINGE OF FOAM LEFT BY THE BORE UPON A FLAT SHORE . . . . . , ' . !67 15 16 LIST OF ILLUSTRATIONS PAGE WAVES APPROACHING ONE ANOTHER IN VERY SHALLOW WATER . • • *73 WAVES SHOWN BY DISTORTED REFLECTION OF MASTS (KINGSTON, JAMAICA) . 177 SHINGLE RIDGES . . • l8l THE ENCROACHMENT OF SHINGLE (SLAUGHDEN, NEAR ALDE- BURGH, SUFFOLK) » . . . .185 STONES ON THE BEACH NEAR CHESILTON AND THE GRA- DATION OF STONES FROM BURTON BRADSTOCK TO CHESILTON .* . . . . . 195 SHORE BETWEEN EAST END OF CHESIL BEACH AND BLACKNOR POINT . . . . -195 SAND BROUGHT BACK AFTER A STORM AND SAND DRIFTED ALONGSHORE . . . . . . 2OI SHINGLE SORTED FROM SAND BY THE WASH OF A HEAVY SWELL (BRANKSOME CHINE, NEAR BOURNEMOUTH) . 213 REMOVING SHINGLE AFTER A HEAVY SWELL, BEFORE THE SAND IS BROUGHT BACK IN CALMER WEATHER . 215 CHEVRON-SHAPED PATCH OF SHINGLE FORMED BY THE WASH OF THE SWELL ON A SANDY SHORE (BOSCOMBE) 217 THE COURSE OF THE RIVER SEVERN FROM GLOUCESTER TO THE SEVERN TUNNEL . . . . 233 LIST OF ILLUSTRATIONS 17 PAGE THE BORE ON THE SHALLOWS OF DENNY ROCK, AND AFTER REACHING DEEPER WATER .... 237 THE NOOSE SANDBANK ..... 243 CINEMATOGRAPHS OF THE BORE APPROACHING STONEBENCH 2$$ CINEMATOGRAPHS OF THE BORE ON REACHING STONEBENCH 259 BORE VIEWED FROM BEHIND AFTER PASSING STONEBENCH 263 STATIONARY WAVES CAUSED BY A WEIR ON THE RIVER AARE, SWITZERLAND ..... 269 BIRD'S-EYE VIEW OF WAVES IN WHIRLPOOL RAPIDS, NIAGARA 275 LEAPING WAVE, WHIRLPOOL RAPIDS, NIAGARA . . 279 WAVES IN WHIRLPOOL RAPIDS, NIAGARA . . y 285 WAVES IN A RAPID ON THE ST. LAWRENCE, TAKEN GOING DOWN-STREAM IN A STEAMER .... 289 ROLL-WAVES IN THE GRUNNBACH CONDUIT, LOOKING UP- STREAM ....*. . 303 ROLL-WAVE LEAPING THE OUTFALL OF THE GRUNNBACH CONDUIT. .> ^ . ; , , . ,,^. M.« 307 ROLL-WAVES IN THE GRUNNBACH CONDUIT, LOOKING DOWN-STREAM . . ,. . ., , » 315 OUTFALL OF THE GUNTENBACH AT THE WAVE-TROUGH, AND (A FEW SECONDS LATER) AT THE WAVE-CREST . 321 2 18 LIST OF ILLUSTRATIONS PAGE THE GUNTENBACH WITH TWO ROLL- WAVES . • 32$ UPPER PART OF FALLS AT NIAGARA, SHOWING UNBROKEN WATER ... • 331 NIAGARA FALLS, SHOWING FLEECY APPEARANCE PRODUCED BY THE BREAKING OF THE FALLING WATER . - 335 THE BREAKING WATER OF NIAGARA FALLS, TAKEN FROM S.S. MAID-OF-THE-MIST . . 339 FALLS OF THE TSCHINGELBACH, BURGLAUENEN , . 343 WATER ROCKETS PROJECTED UPWARDS FROM THE FOOT OF THE HORSESHOE FALLS, NIAGARA . . . 347 WAVE-TRACK OF STEAMER ON THUNERSEE, SHOWING THWART-SHIP AND DIVERGING WAVES . . • 351 WAVE-TRACK OF STEAMER ON CONISTON WATER, SHOWING INTERFERENCE BANDS BETWEEN DIVERGING WAVES OF THE BOW AND STERN SERIES . ' ; *V . 355 WAVE-TRACK OF STEAMER ON CONISTON WATER, SHOWING INTERFERENCE BANDS AND DOUBLE-CRESTED DIVERG- ING WAVES . . . . ... .361 WAVE-TRACK OF STEAMER ON LAKE LEMAN 363 WAVE-TRACK OF STEAMER ON LAKE LEMAN, SHOWING, ON THE LEFT, THE INNER BOUNDARY OF GROUP OF DIVERGING WAVES ,:. . ,, < . . . 365 LIST OF ILLUSTRATIONS 19 FIGURES IN THE TEXT. PAGE COMBINATION OF WAVE AND SWELL . . .98 A CYCLONIC SYSTEM . . . . . .122 VERTICAL AND HORIZONTAL CURRENTS OF WAVES . . 147 THE CHESIL BEACH .... . IQI CHART SHOWING THE SHAMBLES SHOAL . . . 207 DIAGRAM OF VERTICAL AND HORIZONTAL COMPONENTS OF THE TIDE NEAR PORTLAND .... 209 PART I ON THE SIZE AND SPEED OF DEEP- SEA WAVES CHAPTER I Introduction — Waves on ponds — Waves on lakes — Coniston Water — Lake of Geneva — Lake Superior — Waves in semi- enclosed seas — Western Mediterranean — China Sea. Introduction MEASUREMENTS of waves at sea by means of the eye are not susceptible of great accuracy ; but the irregularity of the waves themselves is so con- siderable, especially in their most important condi- tion, which is during storms, that it is more useful to measure many waves somewhat roughly than to obtain (even if it be possible) the precise measurement of a few. The advantage of the mere number of observations does not, however, apply when we pass from rough measurements by careful observers to mere guessing at the dimensions of waves as seen from on board ship . Measurements of the height of waves, for instance, taken in the usual way by finding the height above the ship's water- line from which a neighbouring wave-crest just intercepts the horizon, are believed to be accurate 24 WAVES OF THE SEA to within i foot in 10 when made by a practised observer. This was the estimate of the late Lieutenant Paris, of the French Navy, for his obser- vations, and the late Lord Kelvin informed me that he relied upon his own measurements to the same extent. When, however, an unpractised observer, judging merely by the look of things from the deck of a ship, guesses the height and length of waves, it is possible for him to err much more widely than he would on the land, where he stands on a firm platform, with objects of known size in the neigh- bourhood to afford a scale. The rolling of the ship, in particular, alters the apparent direction of the vertical so as to mislead the judgment as to height. It is difficult to say how widely these guesses may depart from fact, but I do not think it unlikely that waves 20 feet high may, according to the circumstances, be guessed by unpractised or careless observers at anything from 10 to 30 feet. This is a range of error of 100 per cent, as against the i o per cent, of the practised observer. While, therefore, in dealing with the data before me, I have been anxious to obtain numerous records, I have been still more anxious not to include any which might belong to the category of misleading guesses. As far as possible, I have relied upon figures in which the observer has ex- AND OTHER WATER WAVES 25 plained his method of measurement, and detailed the attendant circumstances of wind, weather, and environment. This may have led to the rejection for the purpose of this book of good measurements, but the importance of all measurements or estimates of the size of sea waves being accompanied by a statement of the method of observation and the attendant circumstances cannot be too strongly emphasised. The relation between the wave-length in deep water and the period (or time which elapses between the passage of a fixed point by two succeeding wave-crests) has been calculated mathematically, and verified by observation. I have, therefore, in what follows applied the formula thus obtained, viz. : Wave length = 5^ x square of period, to obtain the wave-length when only the period has been measured. In all cases, however, the reader is informed which observation was made, that of length or period. WAVES ON PONDS As the sea is always heaving with the disturb- ance due to former winds, the commencement of the wave -making action of wind is best observed 26 WAVES OF THE SEA on smaller sheets of water. In a rock pool, 20 feet in diameter, under High Peak, Sidmouth, Devon, I have measured waves of i inch wave-length at the windward end, and 4£ inches at the leeward end, the wind having the force of a gale. Thus if we take the average length of the waves to be 2 J inches, there was in this small pool a series of ninety consecutive waves, the longest of which was i-6oth of the length of the series. On the Round Pond in Kensington Gardens, London, 670 feet in diameter, the water on an absolutely still day has a glassy surface, but any breath of air sufficient to be felt upon the cheek is enough to ruffle its surface and to do away with the mirror -like reflection. It will then be seen that the area of ruffled water has been instan- taneously covered with an almost uniform pattern of little waves about an inch in length from crest to crest. As the minutes pass during which the breeze continues, the height and length of the waves to leeward increases, but those on the windward edge of the ruffled water remain of the original size. Soon the whole pond, except a few feet at the windward end, is covered with waves travelling before the wind whose size increases regularly from the windward to the leeward shore. The maximum size attained at the latter place depends to some AND OTHER WATER WAVES 27 extent upon the strength of the wind, but even in a whole gale (October 6, 1901) the little breakers on the shelving lee shore followed one another at intervals of only i second.1 The wave-length was not measured, but appeared to be about 3 feet. The whole of this effect was soon produced, being a matter of minutes ; the height of the waves quickly diminished when the wind lulled, and could be seen to increase instantaneously when squalls occurred, the height increasing by a considerable fraction during a squall lasting four minutes. If we take the greatest length of waves observed in the Round Pond on the day referred to as 3 feet, and adopt the simple mean of the shortest ( i inch) and the longest (36 inch) wave as the average wave-length for the pond during a gale, we find that there is then a series of 436 consecutive waves, of which the longest is 1-2 2 3rd of the length of the series, or of the length of fetch of the wind, or of the length of run of the waves. On the Serpentine Water in Hyde Park, a much larger pond, the shape of which, however, is not favourable to the development of waves, I measured those produced in a strong SW. breeze, going out 1 The calculated length of a i -second wave in deep water is 5 feet. The velocity of the wind in London was 50 statute miles per hour. 28 WAVES OF THE SEA in a boat not very far from the lee end, where I was sure of a sufficient depth of water for the purpose. Measurements with a wooden rule re- peated during the course of half an hour gave 3 feet wave -length and a height from trough to crest of about 2 inches. In ponds, therefore, we see that when the wind raises waves at all they must be numerous, and that even the longest must be a small fraction of length of the series, and that the fraction decreases as the length of the pond increases. The steepness of the waves formed by wind upon ponds is also found to be small as compared, for instance, with the waves which may sometimes be seen caused by obstructions in a river. Thus the height of the waves in the Serpentine was only i-i8th of their length. WAVES ON LAKES Coniston Water We now proceed to observations on a larger sheet of water, the Lake of Coniston, or Coniston Water, in Lancashire. They were taken near the upper end of the lake, at a distance of 7 statute miles from the lower end. The sheet of water is narrow and nearly straight, the lower end about AND OTHER WATER WAVES 29 S. by W. from the upper. When the wind blew up the lake with the force of half a gale I found that the waves near the upper end succeeded each other at intervals of 2 seconds, corresponding to a wave- length in deep water of 20 feet. At a later date Mr. Hamil, a seaman of experience, and captain of the steam gondola which plies on the lake, sent me the following observations which he made upon the larger waves produced in a whole gale of wind. On September 3, 1902, at 8 a.m., there was a light wind blowing up the lake, i.e., from the soutji. At 10 a.m. it rose to a gale, the wind shifting to a little W. of S. Mr. Hamil timed the waves at Yew dale Beck, near the upper end of the lake, with the following results : Waves passing point of observation. Waves breaking on beach. 10.30 a.m. ... 27 per minute 10.32 a.m. ... ... 24 per minute u.oa.m. ... 23 „ „ n.2a.m. ... ... 22 „ 11.30 a.m. ... 21 „ „ 11.32 a.m. ... ... 20 „ „ 12.0 a.m. ... 21 „ „ 12.2 a.m. ... 20 „ Thus the period appeared to be constant after 1 1.30 a.m. — i.e., an hour and a half after the gale commenced. The length of wave in deep water corresponding to the observed period of 3 seconds 30 WAVES OF THE SEA is 46 feet. The distance from the lower end of the lake being 7 statute miles, or 36,960 feet, the length attained by the waves is i-8o3rd of this dis- tance. If we take the average length of the series of waves as 23 feet (since when they commence their length is very small), the number of waves in series from end to end of the lake was 1,608. The wind conditions were similar to those described for the Round Pond at Kensington and for the Rock Pool at Sidmouth. Thus we see that the length of the storm-waves is increased when the length of the sheet of water is increased, but more slowly. The above, however, are not the greatest waves which can be formed on Coniston Water, although they are probably about as large as are formed in ordinary gales. The following observations sup- plied to me by Mr. Hamil illustrate the way in which waves larger than ordinary are produced. They indicate a fact (which we shall find illus- trated later when dealing with the great waves of high southern latitudes — e.g., between the Cape of Good Hope and Australia) that the wave -raising power of wind is much greater when operating upon water already in waves than upon nearly smooth water. Mr. Hamil finds that the largest waves on AND OTHER WATER WAVES 31 Coniston Water are only formed when, after about three days of steady wind blowing along the length of the lake has produced a steady " run " of waves, it comes on to blow very hard in the same direction. Under such circumstances he recorded, near the upper end, a wave-length of 65 feet and a height of 5 feet. These measurements were made against the side of the steam gondola. He relies upon the length to less than 5 feet either way — i.e., is sure that the waves were more than 60 and less than 70 feet long. The determination of height he found more difficult. The wave-length attained under these somewhat rare conditions was 1-5 69th of the length of the whole series. The length of wave was 13 times the height. Assuming the same steepness of water for ordinary gales as for the above unusual storms, the height of wave corresponding to the wave-length of 46 feet would be 3^ feet, and this, I suppose, is seldom much exceeded on Coniston Water. The Lake of Geneva On Lake Leman, or the Lake of Geneva, Dr. F. A. Forel records during storms wave-periods of 4.7 seconds at Morges and 5.0 at the town of Geneva. The longest run which a wave could have before reaching Morges is 27 statute miles, 32 WAVES OF THE SEA and the calculated wave-length in deep water for a period of 4.7 seconds is 113 feet. The length of run possible at the town of Geneva, which is situated at the lower extremity of the lake, is 43 statute miles, and the wave-length corresponding to a period of 5 seconds is 128 feet. At Morges, therefore, the length of the storm -waves was 1-1261 of the length of the series and at Geneva 1-1774. If we suppose the wind blowing down the lake, and at the Geneva end making waves 1 28 feet long, then if we take as before the average length of waves on the lake to be one -half as long, the number of the waves in the series, from end to end of the lake, is 3,548. The following heights of waves on the lake are recorded by Thos. Stevenson * as having been observed by Buckie at the distances stated from windward shore, viz. : Height of Wave Length of Fetch in feet. in miles. 7-0 31 7-0 38 8.0 38 8.0 40 1 Vide article, " Harbours," " Encyclopaedia Britannica," Qth edition. Of all the bodies of water cited by Stevenson in his table of the increase of height with length of fetch, the Lake of Geneva is the only one which is wholly enclosed. Many of the others are, moreover, shallow or affected by strong currents. AND OTHER WATER WAVES 33 The greatest length (calculated from the observed period) is 15.4 times the greatest height. Lake Superior On Lake Superior waves have been carefully measured from the shore by Colonel D . D . Gaillard, Corps of Engineers, U.S.A. For the size of the waves out in the deep water of the lake, he has, however, had to obtain evidence from the captains of vessels. He writes : l " As the result of inquiries of vessel captains who have navigated Lake Superior for many years, and who have, in some cases, made a special note of the fact that in unusually severe storms the hiorizon could not be seen from the wheelhouse when the ship was in the trough of the sea, on account of adjacent wave -crests, it seems probable that during unusually severe storms upon Lake Superior, which occur only at intervals of several years, waves may be encountered in deep water of a height of from 20 to 25 feet and a length of 275 to 325 feet." He adds that the waves on Lake Superior are larger than occur on any other of this chain of great lakes. 1 " Wave Action in Relation to Engineering Structures " (Washington Government Printing Office, 1904), p. 82. 34 WAVES OF THE SEA In the Duluth Canal at the western end of Lake Superior Colonel Gaillard during 1901 and 1902 frequently measured waves 200 feet in length, occasionally 250, and once 275 feet, in a depth of not more than 27 feet of water. The velocity of such waves is reduced in water of this depth and they close in upon one another, their wave- length diminishing. The measurements taken by Colonel Gaillard in somewhat shallow water there- fore show that the estimates of wave-length made by the captains of vessels were not excessive, and that we may safely say that a length of 300 feet is attained during exceptionally severe storms by the waves of Lake Superior. The possible length of fetch of wind and length of run of the waves at Duluth is 298 statute miles or 259 nautical miles, so that the greatest wave-length is 1-5 248th of the length of the lake, and if the average wave-length of the series be 150 feet, there would be 10,496 successive waves simultaneously between windward and leeward shore. Thus again we find the steady growth of wave-length with the length of the sheet of water, the increase of wave-length taking place, however, more slowly. Taking the height of the waves during severe storms on this lake as 22j feet and their length as 300, we find that the length is 13.3 times the AND OTHER WATER WAVES 35 height, which is nearly the same as that (13) found for an exceptional storm on Coniston Water. WAVES IN SEMI-ENCLOSED SEAS OF CONSIDERABLE DEPTH Waves in the Western Mediterranean The Western Mediterranean is about 1,000 statute miles from east to west, but its area is somewhat broken up by islands. The depths are great and the tidal currents small, so that observa- tions of waves are useful for the determination of dimensions due to the mere action of wind on deep water. On April 7, 1899, I sailed from Marseilles at 4 p.m. on the Orient liner Orizaba, bound for Naples, the weather being fair. Next morning, the 8th, when nearing the Straits of Bonifacio, there was a heavy sea directly following the ship, the wind having risen during the night and having now the force of a moderate gale. The waves were not running in a single series of parallel ridges, but with a good deal of crossing, and the characteristic feature of the scene was the number of bursting billows of more or less pyramidal form produced by the meeting at an acute angle of waves running not quite in the same direction, 36 WAVES OF THE SEA These waves, curling over in a cusp and breaking, flecked the surface of deep blue water with white foam which reflected the bright beams of the sun shining through the spray. I set myself to measure the height of these waves at the times of their greatest elevation. Standing on the bulwark rails of the spar deck, while holding to the steel uprights which supported the promenade deck, I found that my eye was on a level with the crests of the highest waves when 22 feet above the flotation line. The position of the latter I obtained from the com- mander, who informed me that its then position was 3 feet below the Plimsoll mark. The height of the deck above the Plimsoll mark was known from the scale plan of the ship. The ship did not pitch, and her rolling was so slow that it was fairly easy to make the observation when on an even keel. Thus the height from trough to crest of the pyramidal waves, which were the characteristic waves of the day, was ascertained with some approach to accuracy to be 22 feet. I desired to ascertain the length of the waves, which were travelling in the same direction as the ship, but as usual was unable to watch a wave- crest running the whole length of the vessel. The length between two convexities near the ship's side, viewed simultaneously, I judged to be 37 AND OTHER WATER WAVES 39 130 feet, using the length of the promenade deck as standard of measurement. This estimate is quite out of harmony with the successive observation of following waves. The speed of the vessel was 14 knots and the waves travelled past the ship at a very considerable speed. They succeeded one another at an average interval of 17 seconds. This determination by itself does not permit the wave-length to be calculated from a simple formula,1 but it is satisfied by a true period of 8 to 9 seconds with length of 328 to 415 feet. Seven seconds (with a wave-length 250 feet) is definitely too small, and 10 seconds ( with wave length 512 feet ) is definitely too much. It is quite inconsistent with a length pf 130 feet, which corresponds to a speed of only i 5 knots, which is scarcely greater than that of the ship, and it seems likely that this estimate of the wave-length was in fact due to the transverse waves caused by the ship's own motion, which for her then speed of 1 4 knots have a length of 1 1 o feet . It appears, therefore, that the average wave- length was not less than 328 feet (an 8 -seconds wave). The wind was westerly, and the sea -room about 1 It can be calculated, however ; see Monthly Chart, North Atlantic, September, 1909, published by the Meteorological Committee. 40 WAVES OF THE SEA 300 statute or 260 geographical miles. This dis- tance from the windward shores is the same as the maximum possible on Lake Superior, but the area of the Western Mediterranean is many times greater. The height which I observed in the Western Mediterranean in a moderate gale (speed of wind estimated by the captain of the vessel at 30 miles per hour) is the same as that recorded for gales of exceptional severity upon Lake Superior, in which the velocity of the wind would not be less than 53 miles an hour. The height, therefore, attained by the largest waves in very severe storms in the Western Mediterranean must be greater. The storms of the Gulf of Lions situated to the westward of my place of observation are notorious for their severity and for the dangerous sea which rises. Admiral W. H. Smyth, in his book upon the Mediterranean,1 writes that in the worst weather in the Gulf of Lions the waves attain a height which cannot be much less than 30 feet. Perhaps we may interpret this as meaning that they cannot be less than 27 to 28 feet. Waves in the China Sea The China Sea is a body of water lying between the mainland of Asia and the open Pacific, from 1 Quoted by Cialdi, " Moto Ondoso del Mare." AND OTHER WATER WAVES 41 which it is screened by the Philippine Archipelago and other islands. The uninterrupted expanses of water are greater than in the Western Mediterranean. The following observations were made by the late Lieutenant Paris, of the French Navy, whose careful methods of work will be described later. Off Cape Varella, in a violent storm from the north-east which lasted several days, the highest wave was 21.3 feet and the average wave-length 328 feet. The sea -room to windward was about 750 statute or 650 geographical miles. Thus the difference in the sizes of waves in great lakes and in the much larger semi -enclosed seas respectively is less than we should have expected from the observed difference between their size in the smaller and in the larger lakes. We find, however, a great increase when we go from the semi -enclosed seas to the open oceans.1 1 In the Caribbean Sea also, where I have made five voyages, the waves are not quite like those of the open ocean. Between Colon and Kingston they have a clear run of more than 500 nautical miles in the direction of the prevailing NE. wind, which blows strongly. The waves, which I have seen attain a height of above 20 feet during ordinary strong Trade winds, remind one more of those of the Mediterranean than of the Atlantic, and there is an absence of the long swell accompanying the rough sea which is characteristic of the open ocean. CHAPTER II Observations of waves on the North Atlantic — Conclusion as to height of North Atlantic waves— Waves on the North Pacific— The effective length of fetch of strong winds on the North Atlantic Ocean— The numerical relation between the length of fetch of wind and height of storm-waves— Waves on the South Atlantic and Southern Indian Oceans —Waves on the South Pacific Ocean— The discrepancy between wave-lengths determined respectively by simul- taneous and by successive observation of wave-crests. Waves on the North Atlantic Ocean I NOW pass on to the records of storm-waves on the North Atlantic Ocean. One of the best is that of the Rev. William Scoresby,1 which I give almost in his own words. The height of the waves was recorded on the east-bound voyage from America to England on board the S.S. Hlbernla. The construction of the ship afforded several platforms of known elevation above the water-line. On March 5, 1848, the ship 1 British Association Meeting of 1850, Report, published 1851, Part II., pp. 26-31. 4* WAVES OF THE SEA 43 was in latitude 5i°N. and longitude 38° 5o'W<., the wind about WSW., and the ship's course true N. 52° E. By sunset of the previous day the wind was blowing a hard gale, which continued with heavy squalls during the night, so that all sail was taken in except a storm stay-sail forward. The barometer, which had stood at 29.50 inches at 8 p.m. on the previous day, had fallen to 28.30 inches by 10 a.m. on the 5th. On the after- noon of this day Dr. Scoresby took up his post of observation on the saloon deck, which gave an elevation of the eye 23 feet 3 inches above the water-line. He found, however, that every ap- proaching wave intercepted the horizon, so that from this position he could decide little except that the average height, reckoned from trough to crest, was more than 24 feet. He therefore ventured upon the paddle-box, which was about 7 feet higher, giving an eye elevation of 30 feet 3 inches. This level was well maintained during the moments of actual observation, because the whole of the ship's length (220 feet) was clear within the trough of the wave when the next following crest was at its greatest apparent height, and the ship at these moments was on an even keel. From this position quite one half of the waves which overtook and passed the ship were above the level 44 WAVES OF THE SEA of the observer's eye . Sometimes a crest extending in a ridge 100 yards long would be from 2° to 3° above the invisible horizon. This Dr. Scoresby says would give a height from trough to crest of more than 40 feet, but I confess I regard this rather as a guess than a measurement. Sometimes, he says, the crossing of two wave -crests would send up a sharp peak of water to a height which he believed to be 50 or 55 feet, or the crest of a breaking wave would shoot up to a similar height. The average height of the waves during the observations on March 5th was more than 30 feet. On the following day, when, the wind being less violent, the waves had subsided to an average height of 26 feet and were more regular, Dr. Scoresby determined the wave-length in the follow- ing indirect manner. The waves overtook the ship every 16.5 seconds and each wave took 6 seconds to run the whole 220 feet of the ship's length. Then the distance between two succeeding wave- crests as thus observed was : 220 x ^g-5 = 605 feet ; but the ship was not running in exactly the same direction as the waves, and a line from crest to AND OTHER WATER WAVES 45 crest measured at right angles would of course be shorter. Making due allowance for the observed angle, the true wave-length was determined to be 560 feet. So far Dr. Scoresby. I proceed to consider the position of his vessel at sea so as to be able to compare the sea-room with that possible on Lake Superior, which we have stated to be (at Duluth) 259 nautical miles. At noon on March 5th, the day when the waves were highest, the vessel was about 600 geographical miles from the coast of Newfoundland, the direc- tion from which the wind blew, and by 4 o'clock in the afternoon 36 miles farther from this, the nearest shore. At noon on the 6th the distance from shore was about 800 geographical miles. If the highest waves were measured at 4 p.m. on the 5th, the ship had been running twenty hours in a hard gale, and was 180 geographical miles from the point where the full force of the gale first struck her. The cyclone was, presumably, travelling in the same general direction as the ship, but how close to the shore the westerly wind first attained the force of a gale we do not know. This much is certain, that where the waves attained an average height of more than 30 feet, and an occasional maximum estimated at more than 40, the sea -room 46 WAVES OF THE SEA or stretch of water to windward l was 600 geo- graphical miles or more, as compared with the 259 geographical miles which is the maximum amount of sea-room possible upon Lake Superior, where we have accepted 22j feet as the height of waves in exceptionally severe storms. The wave-length of 560 feet 2 was measured on the Atlantic with a sea -room of 800 geographical miles as compared with the 300 feet of Lake Superior with a possible sea -room of 259 geo- graphical miles. The length of the waves on March 6th was 21.6 times as great as the height, but on the previous day they must have been steeper. We will next consider the case of another strong gale in the North Atlantic with sea-room of at least 1,000 instead of 600 geographical miles, in which we shall find that the waves are of about the same height as in the last case. The waves in question were observed by myself on December 7, 1900, when outward bound from Liverpool to Boston by the Cunard S.S. Ivernia. 1 The " room" which concerns a navigator is more often that to leeward, but the expression for space to windward is useful for our purpose. 2 From the building line in Bond Street to the front of the Royal Geographical Society's House, i, Savile Row, is the length of a lo-second wave, 512 Jfeet. 47 AND OTHER WATER WAVES 49 Leaving Queenstown on the morning of the 5th, we met a rather heavy swell on our westerly course, which continued and somewhat increased during the 6th. We drove into a strong gale that night, and the highest waves which I saw during the voyage were on the morning of the 7th. The force of the wind on that day was logged as 9 on Beaufort's scale of 0-12, the number 9 being called " a strong gale." We continued to drive through a gale during the whole of the 7th, 8th, and 9th, the winds varying from S. by W. on the morning of the 7th to WNW., always therefore producing a head sea, at first on the port, afterwards on the starboard. The morning of the 7th (bar. 29.15 inches) was the only time during the three days when the waves were " running true " — i.e., in long parallel ridges exactly at right angles to the wind, of which ridges six or seven were simul- taneously visible when looking upwind from the weather side of the ship. There was no long, flat swell noticeable, neither were there minor waves of such prominence as to distract attention from the principal waves. In the course of eleven voyages across the Atlantic this is the only occasion on which I have seen this " regular sea " during a storm, and the waves were the highest which I have ever observed in a storm. Our noon position 50 WAVES OF THE SEA was N. 50° 56', W. 250 33', so that the distance from the Newfoundland Banks was about 1,000 geographical miles. This was the direction in which we were steaming and apparently the direc- tion from which the bad weather was coming. The wind, however, came from a direction in which there was no land for 2,000 geographical miles. I was able to judge the height of the waves more readily than usual on account of several favourable circum- stances. The ship, though pitching, did not roll, being heeled over to starboard at a moderate angle, which did not vary. Stationing myself amidships, I was subject to neither pitching nor rolling, but merely to a lift and fall as each of the large waves passed beneath us. The two promenade decks afforded platforms which happened to be at just the right altitudes for judging the height of the usual, and of the maximum, waves respectively. The altitude of these decks above the water line of the ship I obtained from the scale section of the ship in charge of the chief engineer. The heeling over of the ship was measured, and its amount allowed for, on the assumption that the height of the deck above the water was reduced, on the lee side, to the full extent possible. The height of the waves is recorded as equal to the height of the eye above the ship's water-line, where they just topped AND OTHER WATER WAVES 53 the horizon, nothing being added for the small amount by which they actually exceeded the height of the eye. Thus, when observing on the lower deck, I found that the waves commonly attained 29 feet, and from the upper deck that they occa- sionally attained 43 feet. My lower platform was about the same altitude as Dr. Scoresby's highest position, and my estimate of the height of the ordinary waves is about the same as his. I had the advantage, however, of a second and higher platform, which brought me on a level with the highest waves. Their height as measured by me from this point of vantage is practically the same as that guessed by Dr. Scoresby from his lower platform. The force of the wind was about the same in both cases, his determination with 600 geographical miles sea-room, mine with 1,000 miles. Next day, December 8, 1900, the wind had shifted to WNW. with force 8 (a " fresh gale ") ; the waves were irregular and not so high, and again on the 9th, with wind SW. of the same strength, there was an irregular sea with no very high waves, though they were magnificent from the tumult of their headlong rush and the white fury of the broken water. The sea was, indeed, covered with spume and veiled in spindrift. A small schooner sighted 54 WAVES OF THE SEA in the afternoon was running before the storm under a minimum of sail, and rolling very heavily. From on board this vessel such waves would tower above the horizon, but none were as high as 30 feet. The position of our ship on the 9th was not far S. and E. of that of Dr. Scoresby's ship when he measured the 56o-feet wave-length. Although the higher platform of my larger vessel gave me an advantage for the measurement of the highest waves, her great length, and the structure of the upper works usual on modern liners, made it very difficult for me, unassisted, to measure wave- length. I have found the same difficulty every time I have been at sea, as I have always voyaged in ships measuring from nearly 500 to 600 feet in length. It has usually happened that when there were large regular waves their course made a con- siderable angle with that of the ship, and even when the waves ran more nearly in our own direc- tion, it was generally . impossible for me to watch the wave -crest during the whole of its passage along the length of the ship. I have, therefore, been generally reduced to judging the wave-length in terms of the known length of the ship from the simultaneous position of two crests. The length of the Ivernia is 600 feet, and the regular waves which we encountered in the strong gale of Decem- AND OTHER WATER WAVES 55 her 7th I judged by the above method to be about 350 feet in length. I have never yet seen storm-waves on the North Atlantic in which the distance between the wave- crests, viewed simultaneously, appeared to me nearly as long as the 6oo-feet ships on which I have voyaged, nor, indeed, any which appeared more than 400 feet in length. This was the ap- parent distance between crests during a heavy swell without wind which I observed from the Red Star S.S. Vaterland in March, 1901, east-bound in N. Lat. 48° 30', W. Long. 21° 40'. The second, and latest, opportunity which I have had of measuring very large waves on the North Atlantic was on board the Atlantic Transport Company's S.S. Minnehaha, east-bound from New York to Southampton, on February 9, 1907, in N. Lat. 48° 54', W. Long. 18° 20'. There was only a moderate breeze from NW., but a huge north-westerly swell came upon us at about 45° abaft the beam. The vessel did not pitch, her slow and stately rolling motion was perfectly rhythmical and regular, and in the absence of strong wind observation was unusually easy. It was evident from the great height of the swell that we were only just outside the storm area, and this conclusion was confirmed by a wireless message received 4 56 WAVES OF THE SEA during the day from the White Star s.S. Cedric to the northward of us reporting that she was in a strong NW. gale. Standing on the lower prome- nade deck, one long ridge after another obscured a considerable arc of the horizon after passing beneath the ship, and continued so to obscure it when at a distance estimated at 400 to 600 feet from us. I allowed 2 feet as a minimum estimate for the excess of height above my eye, based on an observation made a short time before. The observation was simply this, that when sitting in my deck chair I had seen a wave similarly obscure the horizon, and on my rising at once, the horizon remained obscured. The increase in the elevation of eye upon rising was found afterwards to be 2 feet. The roll of the ship at the time of the obscuration of the leeward horizon was in each case to the weather side, so that the deck on the lee side was tilted upwards. The amount of the tilt was measured for several rolls, which were quite regular, and assuming that the full amount ought to be added to the deck height, I obtained 2 feet so to add. I was standing during the observations, and my eye-height is 5 feet 9 inches, or say 5^ feet, so that there is altogether 9^ feet to add to the height of the deck. This, measured with a heavy 57 AND OTHER WATER WAVES 59 rope hung over the lee side, I found to vary with the oscillations of the ship and of the water from 2 9 2- to 332 feetJ giving an average of 31^ feet for smooth water. Two days later, with lighter bunkers, the height above the smooth water pf the Solent was found by the rope to be 32^ feet. Taking, therefore, the height of the deck at the time of observation as 31^ feet, the height of the waves which repeatedly passed us was : Height of deck 31^ feet Added for roll 2 „ Height of eye above deck 5i ?, Height of wave above the eye (at least) ... 2 „ Wave height (at least) 41 feet. Whilst this is 2 feet less than that determined for the highest waves during the strong gale of December 7, 1900, there was much less variation in size from one wave to another, and the average height in this north-westerly swell was no doubt quite as great as that during the strong southerly gale. The sea-room on February 9, 1907, reckoned from the coast of Greenland — the direction of the wind — was about 1,100 geographical miles, and from the Newfoundland Banks to the westward about 1,200. 60 WAVES OF THE SEA Conclusion as to Height of North Atlantic Waves Thus concordant observations indicate that any- where in the North Atlantic with sea-room of from 600 up to certainly 1,000 and perhaps 2,000 miles the height of the large waves during ordinary strong gales is practically constant, being not less than 43 feet.1 With regard to the height which is momentarily attained by peaks of water shooting upwards where waves cross, the late Lord Kelvin informed me that he had measured one 60 feet high, and this measurement confirms the concordant guesses of several officers on North Atlantic liners whom I have consulted on the subject. I have not myself seen anything nearly so high. Accounts not infrequently appear in the news- papers of some great wave encountered by the fast Atlantic liners. These are sometimes reported as 80, 90, and even 100 feet high. This height invariably relates to the altitude above the flotation line of the superstructures which have been deluged with water. This is not, properly speaking, the height of a wave, but merely the height to which a body of water is thrown when a wave breaks on board. This increases with the speed of the 1 See Note on p. 138. AND OTHER WATER WAVES 61 ship, which dips her bows into the rising billow in a head sea. The recorded heights also tend to increase as the ships are built of larger dimen- sions, on account of the fact that the greater height of the navigation bridge and wheel-house allows the attainment of a greater altitude to be recorded with certainty. The wheel-house of the Lusitania, e.g., is 80 feet, or rather more, above the flotation line. Waves on the North Pacific Ocean The North Pacific Ocean has a breadth of open and deep sea about twice as great as that of the North Atlantic. The passage from Victoria, B.C., to Yokohama is about 4,000 geographical miles, as compared with the 3,000 from Liverpool to New York. This route is traversed by a number of liners similar to the medium -sized Atlantic liners. I made this passage once, east-bound from Yokohama to Seattle, in fair weather, when we only encountered a moderate swell similar to that met with in similar weather in the same latitudes on the Atlantic. The great circle course which is followed took us as far north as Lat. 49° 40'. I had opportunities during this long voyage of collecting opinions upon the size of the waves on this route as compared with those of the North Atlantic and of the Southern Ocean from seamen who knew all three. The late 62 WAVES OF THE SEA Dr. Elgar, F.R.S., designer of the Campania, who was also a passenger, gave his views. There was a complete consensus of opinion — First, that the type of storm was the same as on the North Atlantic routes. Second, that the storm-waves on the Pacific route were certainly not higher than those on the Atlantic, and I did not gather that they were any longer. They were said to be as irregular as those of the North Atlantic and not to " run true " as the waves do in the steadier winds of the Southern Ocean, a circumstance which was attributed to the storm being generally of the rotatory character, with a fairly rapid change in the direction of the wind. I conclude from this evidence that the greater size of the ocean does not in this case lead to the development of greater storm -waves than those of the North Atlantic. Thus I have not found in northern latitudes any increase of height of storm -waves beyond a distance of 600 geographical miles from the windward shore. Nevertheless the size of waves observed by Scoresby is not nearly equalled in enclosed seas of 500-700 miles in breadth. I attribute the difference to a smaller size pf cyclonic systems on the semi -enclosed seas.1 1 Cf. A. Buchan, " Meteorology," " Encyclopaedia Britannica," 9th edition. AND OTHER WATER WAVES 63 On the Effective Length of Fetch of Strong Winds on the North Atlantic Ocean The volume of charts illustrating the weather in the North Atlantic from December 18, 1898, to February 15, 1899,* provides detailed and reli- able information as to the effective length of fetch of winds in that ocean. The storms were of unusual strength and persistence, so that the charts give us maximum values. The positions and distances stated below are measured from the charts. At noon, January i, 1899, a west wind of force 7-8 of Beaufort's scale and upwards (i.e., in no place less than a moderate gale) is shown to obtain from N. 49°, W. 40° to N. 49°, W. 60°, a distance of Ij3°° geographical miles or rather more. But, in order that the waves at the lee end should be reinforced by those at the weather end of the strip, time must be allowed for the travel of the waves. Waves of about 8 seconds period would be preva- lent in such winds, and we will consider their move- ment. Their speed is 8X3=24 knots, and even when going as forced waves before the wind they will only travel 576 geographical miles in 24 hours. Now, an examination of the charts for Decem- 1 Charts illustrating weather of the North Atlantic Ocean in the winter of 1898-9, Met. Council, 1901. 64 WAVES OF THE SEA her 3 ist and for January 2nd shows that on neither day was there so long a strip of water simultaneously subject to west wind of force 7-8 as on January ist. Therefore the apparent length of fetch on that day never became effective. But we see on comparing the chart of December 3 ist with that of January ist that a strip of 550 geo- graphical miles was continuously subject to west wind of force 7-8 for 24 hours, and towards the end of this period this was the effective length of fetch of the wind for 8 -second waves, and for any swifter waves which the force of the wind may have been capable of producing. At noon on January 10, 1899, winds of Beaufort force 7-8 with direction a little W. of N. prevailed from N. 47°, W. 46° to N. 48°, W. 32°, a distance of 600 geographical miles. On the previous day (January 9th) at noon the wind was blowing in the same direction between the same two positions, with a force of 9-12, so that for 24 hours there were winds of constant direction, with a force vary- ing from a moderate gale to a hurricane, simul- taneously and continuously affecting the water over a stretch of 600 geographical miles. Towards the end of the period, therefore, there was an effective length of fetch of wind of 600 geographical miles for all waves of rather more than 8 seconds period. . AND OTHER WATER WAVES 67 Throughout the whole of the nine weeks of ex- ceptionally stormy weather which is covered by these charts I cannot be sure of any greater effec- tive length of fetch of wind than 600 geographical, or 700 statute, miles. This is the wave -making1 length of wind-fetch for the North Atlantic which we have to compare with the 259 geographical or 289 statute miles of Lake Superior. The wave- making effect of a strip of wind of 600 geographical (700 statute) miles long on the mid-Atlantic is, however, increased by the swell which is always entering at the weather end. This result agrees with the conclusion already arrived at that Scoresby observed, at 600 geographical miles from the wind- ward shore, waves of the greatest height producible in the Atlantic by the then force of wind. The Numerical Relation between Length of Fetch of Wind and Height of Storm-waves Thomas Stevenson's empirical formula ( height of wave in feet = 1.5 X square root of length of fetch in geographical miles) was shown by him to apply to distances of rather more than 100 geographical miles. Colonel Gaillard observed waves 23 feet high in the Duluth Canal with a length of fetch of 259 nautical miles, the height calculated from Stevenson's formula being 24. i feet. The same 68 WAVES OF THE SEA formula gives 36.75 feet for 600 geographical miles from the windward shore, which is in fair agreement with Scoresby's observations on March 5, 1848 (see ante, p. 42). Wares on the South Atlantic and Southern Indian Oceans Sailors agree that storm -waves are seen in their fullest and most typical development in those Southern latitudes where the ocean uninterruptedly encircles the globe. They are observed from ships which go round the Cape of Good Hope, and by those which go round Cape Horn. I proceed to the observations made by the late Lieutenant Paris,1 of the French Navy, in 1867, when proceeding by the corvette Dupleix to the China station via the Cape of Good Hope. The Dupleix was a sailing vessel with auxiliary steam. She ran before a prolonged westerly gale during the last days of October, 1867, having passed the Agulhas Bank, but being still west of the Island of St. Paul. The vessel seems, therefore, to have been somewhere between 40° and 8o°E. Long, and about 40° S. Lat. After this storm the vessel proceeded northwards under steam through the calms of Capricorn, 1 See Revue Maritime et Coloniale, vol. xxxi. (1870), pp. m- 127, " Observations sur 1'Etat de la Mer." AND OTHER WATER WAVES 69 accompanied by a swell from the south-west. Lieutenant Paris made regular observations of waves during the whole of this voyage, and when cruising in the seas of China and Japan, but his opportunities of measuring large ocean waves occurred only in the southern ocean, and principally during the storm above referred to. For obtain- ing data as to the size of full-grown storm -waves from his valuable paper, we are, in fact, almost restricted to a single day's observations, but the value of this day's work is greatly enhanced by his daily practice in observing. His method, as he explicitly states, is that of Dr. Scoresby, the wave-lengths being usually obtained by noting the time occupied by the wave in running the length of the ship combined with the interval elapsing be- tween the arrival of the waves. The height he calculates as equal to that of the eye when the wave just obscures the horizon ; but his observa- tions in this respect are superior to Dr. Scoresby's, for, by nimbly running up or down the shrouds, he got on a level with each succeeding wave, and was not reduced to Scoresby's expedient of guessing the height above him of the largest waves. His results are as follows. On October 25th, 1867, between the Cape of Good Hope and the Islands of St. Paul and Amsterdam, during a strong gale 70 WAVES OF THE SEA from the north-west with violent snow squalls, he measured at different times of the day 30 waves which averaged 29.5 feet in height, and of these 6 succeeded one another in a procession, all the members of which were of equal height, viz., 37.7 feet (11.5 metres). Later in the day he saw waves which he says were " certainly higher " than these, but he was not at the time so placed as to be able to measure them. Their height, we may safely conclude, was not less than 40 feet and probably a little more. Several sailors on board who had been much at sea had never seen waves so high. They occurred on what appears to have been the fourth day of the gale, but the narrative is not quite clear upon this point. It will be noticed that both in average and extreme height the results are practically the same as were obtained during strong gales in the North Atlantic by Scoresby on March 5, 1848, and by myself on December 7, 1900. The waves were, however, much longer in the Southern Ocean, the greatest average on any day being 771 feet, with not a few of 900, and several waves surpassing 1,312 feet (400 metres) in length. Scoresby, measuring in the same indirect way, found the Atlantic waves 560 feet long. Although at the outset of our discussion of the AND OTHER WATER WAVES 71 relation between the size of waves and the size of the basin in which they were formed, I used either observed lengths, or lengths calculated from the period, according to which were available, I find that in observations from shipboard in the open ocean there is in this matter a troublesome discrepancy. Most of the " wave-lengths " given in published records such as those of Scoresby and Paris were not really observed as an apparent distance separating two ridges of water viewed simultaneously, but are deduced by calculation from observed speeds. The speed is obtained in the way described by Dr. Scoresby, viz., the observa- tion of the time taken by a wave -crest to run the length of the ship, combined with the interval be- tween the arrival of two wave -crests, the speed of the ship being known. When the ship is broad- side on to the waves their apparent period should enable one to calculate at once the wave-length. I found, however, when observing at sea, that the length so calculated was much greater than the apparent distance between successive convexities of the water's surface. I have also found that many officers on the Atlantic liners disclaim ever having witnessed in the North Atlantic waves of the great lengths which are measured indirectly by system- atic observers from the observed speeds of, and 72 WAVES OF THE SEA the interval of time between, successive waves. The discrepancy I suppose to be in some way con- nected with the fact that in a storm there are always waves of different length and speed run- ning in the same direction. The waves made by the ship, which are stationary with respect to the ship herself, also I think produce a confusing effect on the direct method of estimation by in- creasing the height of the progressive waves at certain definite positions. I do not think it desir- able to set aside either mode of observation, and in what follows I shall record both. The reader will in consequence find that, while the records of heights provide consistent numbers which are comparatively easy to interpret, the records of length are more confused. Lieutenant Paris remarks of the great series of six waves above described that, as they passed, they left the ship, which was 230 feet long, in a valley "a cable's length" (600 feet) across. The following is a case of a great wave-length recorded by direct observation in the same part of the ocean, viz., on the route from the Cape of Good Hope to Australia. Major Leonard Darwin, who communicated the facts to me, was on this voyage in a vessel 400 feet long, when they fell in with a gale of such unusual severity that the AND OTHER WATER WAVES 73 Captain went twelve hours out of his course in order to partly avoid its fury. Major Darwin made a special effort to judge the distance separating crest from crest of the great rollers, and came to the conclusion that it was three times as great as the ship's length— /.£., 1,200 feet. Thus we see that if we compare the wave-lengths, calcu- lated from the velocity of the waves, as observed between the Cape of Good Hope and Australia with those obtained by the same process in the North Atlantic, the Southern Ocean has much the longer waves, and, again, if we compare the eye- estimated length in the Southern Ocean with that in the North Atlantic, the Southern Ocean has also much the longer waves. The following observations by Captain Hugh F. David, of the White Star S.S. Corinthic,1 were made during August, 1907, somewhere between S. Lat. 45° 30', E. Long. 61° and S. Lat. 46° 45', E. Long. 98° 25'. This is between the Island of St. Paul and Kerguelen Island, and about 600 geographical miles farther south than the position where Lieutenant Paris made the observations above described. The direction of the wind was westerly, and its greatest force was logged as 9, 1 Letter to Captain Campbell Hepworth, R.N.R., of the Meteorological Office, and log of the Corinthic, 74 WAVES OF THE SEA which on the Beaufort scale of o- 1 2, is equivalent to a wind-velocity of 44 statute or 38.2 geo- graphical miles per hour, and is termed " a strong gale." An interesting photograph was taken at the after -end of the promenade deck, with camera 26 feet above the sea level, showing a neighbouring wave following the ship and eclipsing the distant horizon. Captain David writes : " With regard to the lengths and heights of the waves at the time of the photograph, I did take quite a few observations, more particularly of the heights, which ranged from 38 to 45 feet, though I have a vivid recollection of one which I think was quite 50 feet in height. Standing on bridge at 50 feet above sea level, the crest of this wave appeared level with the eyes of jigger rigging just before ship's stern commenced to rise. The height of rigging from the horizontal line drawn from bridge rail to jiggermast taken from scale slightly exceeds 50 feet. This was, of course, an exceptional wave, and I felt quite glad when it passed without doing any damage. Referring to their length from crest to crest, I would not be quite so definite, though an average would be about 600 to 750 feet during the worst part of the storm and indeed afterwards." (The length of the Corinthic is 500 feet.) AND OTHER WATER WAVES 75 In the absence of any statement to the contrary it may be safely concluded that the wave-length was judged by eye, not determined by indirect measurements as in the cases of Paris and Scoresby, eye estimation of the distance between two crests simultaneously observed being the almost universal practice of officers on the bridge. Thus this length for Southern (Indian) Ocean of 600- 750 feet should be compared with my eye estimate of 350-400 feet for similar weather in the North Atlantic, and not with Scoresby's indirect measure- ment there of a length of 560 feet. Captain Percy Howe, who has voyaged on the same route, informs me that between the Cape of Good Hope and Adelaide he was in 1907 subject to a gale of 21 days' duration, from July i 5th to August 5th, much the most prolonged which he has ever experienced. During the most violent parts of the storm the waves which passed the ship obscured his horizon when the ship was on an even keel. His eye-height on the bridge of the Owe sir y Grange was 45 feet, so that the waves exceeded this height. The ship's length was 480 feet, and she was wholly within the trough of the waves, the length of which he estimates at 750 feet. These estimates, which are almost the same as Captain David's, are for a part of the same series of exceptional storms. 76 WAVES OF THE SEA Waves on the South Pacific Ocean Leaving now that part of the Southern Ocean which is on the routes from the Cape of Good 'Hope to Australia and New Zealand, we will examine records from the Cape Horn route from New Zealand to Europe, which touches higher latitudes . The following observations for height were given me by Mr. G. T. Ogilvie. He came home (east- bound) by this route in 1880 on a i,3oo-ton ship, with a length of 230 feet. Whenever it blew hard he used to get into the mizzen rigging and sight on to the horizon from various heights. Near Cape Horn, and therefore more than 53° S. of the Equator, in a full gale blowing from about SW., he estimated the highest wave as 2 feet above his plane of observation, which, on the Captain's estimate, was then 40 feet above the water-line, making a total wave -height of 42 feet. Waves 30 feet high were comparatively common. He thought he saw one or two rollers not less than 45 feet in height from trough to crest, and possibly 48 feet, but attaches little weight to the figures for the altitude of a wave apparently so far above the line of sight. The following observations by the Hon. Ralph AND OTHER WATER WAVES 77 Abercromby 1 relate to waves observed in the South Pacific Ocean on a voyage between New Zealand and Cape Horn on S.S. Tongariro in 1885. On July 1 6th the ship was in S. Lat. 55°, W. Long. 105°, in a hard gale from SW. The waves were the largest seen on the voyage. For the measure- ment of their height he used a 4^-inch aneroid barometer with a very open scale divided to .01 inch. He found (on the assumption that a difference in .001 inch in the aneroid reading was equivalent to a change of i foot in level) that in passing from trough to crest the greatest lift experienced by the aneroid in the cabin was 40 feet. This was a solitary instance. The next greatest was 30 feet. Now, on a previous day, he had found by measurement with a piece of string that, when the wave -crest passed the cabin, the porthole was 6 feet nearer the water than it was at the trough of the wave. Assuming the same difference to hold during the day of heavier sea, he adds 6 feet to the lift of 40 feet in order to obtain the total height of the greatest wave, which he therefore considers to have been 46 feet. He reckons the liability to error of the aneroid read- ing at 2 to 2.5 feet, and that of the measurement 1 Phil. Mag., April, 1888 (vol. xxv., 5th series), " Observa- tions on the Height, Length, and Velocity of Ocean Waves." 78 WAVES OF THE SEA of water-level by the string at not less than 2 feet either way. Thus, his determination of a single wave -height of 46 feet may really be due to a wave of anything from 41 J feet to 50^ feet in height, and therefore, as is the case with so many other measurements in storms at sea, can only be taken actually to establish a wave -height of a little more than 40 feet. Dr. G. Schott, using a sensi- tive aneroid with microscopic reading, recorded a maximum wave -height of 39.4 feet in the South (Atlantic) Ocean. On the Discrepancy between Wave-lengths deter- mined respectively by Simultaneous and by Successive Observation of Wave-crests On March 15, 1903, I was on the S.S. Hitachi, of the Nippon Yusen Kaisha, in N. Lat. 28° 26', E. Long. 125° 53', bound for Kobe, from Hong- kong. The position is in the East China Sea, which is here only partially screened from the open Pacific by the Loo-choo Islands. We encountered in the afternoon a heavy northerly swell, which met the ship at an angle estimated by eye at 45°. One crest was at the stern when the next following was at the bow, and, knowing the length of the ship, the true distance from crest to crest, reckoned at right angles to the course of the waves, was AND OTHER WATER WAVES 79 known to be 280 feet. But the interval between the arrival of the waves was 8f seconds, and, as the vessel was travelling in a direction somewhat opposed to the direction of the waves, the true wave-period must have been somewhat greater. An 8 J -seconds wave has a length in deep water of 370 feet and a 9-seconds wave of 41 5 feet. Assuming the determination of the period to be fairly accurate, as the observation is an easy one, and the application of the mathematical formula to be valid, as there is every reason to suppose, we must seek some reason why the apparent wave- length should be quite 90 feet less than the true. Two possibilities suggest themselves. The first is that minor sea-waves are noticed in the simul- taneous observations of wave -crests, but are passed over by the eye when watching the progress of the more rapidly moving crests of the principal sea -waves. The second possibility is that the shortening of apparent wave-length was due to the increased height of the sea -waves in the vicinity of the bow and the stern of the vessel. A large vessel pro- ceeding at a fair speed — in the above case a vessel of about 6,400 tons going at 12 knots — produces a short, steep wave, several feet high, at the bow, and a similar one at the stern. In smooth water 80 WAVES OF THE SEA these are stationary relatively to the ship, and have a constant height. In a slight head sea, however, the stationary bow-wave is replaced by an intermittent wave, rising up, when the sea-wave meets the bow, and subsiding at the trough of the sea-waves. In a heavy head sea the ship's own bow-wave becomes inconspicuous, but, nevertheless, contributes as much as before to the total height of the sea-waves as they reach the bow and the stern. Now, looking aft, the steeper ship-wave will maintain the apparent crest of the flatter sea -wave for some distance after the crest of the true sea-wave has passed. Conversely, looking forward, the steep bow-wave will cause an anticipation of the true crest of the advancing, flatter sea -wave. Consequently the above measure- ment of wave-length by simultaneous position of crests along the ship's side was too small by twice the distance through which the steeper ship-wave shifted the position of the combined wave-crest.1 Where the sea waves are large and regular and long, and the ship small and slow, this error is least. With a moderate sea and a large and swift ship the error is greatest. Officers of the Atlantic liners generally estimate 1 The presence of ship's waves need not affect the deter- mination of wave-length from observed velocity, for the interval between the arrival of a sea-wave at st^rn and bow is the same as if there were no ship-wave. AND OTHER WATER WAVES 81 wave-length by the position of crests with relation to the bow and stern of the vessel. Thus, when sailing by the Allan liner Tunisian, in 1901, the chief officer informed me that in a storm in the Atlantic the ship generally " took three waves/' which means that when one crest is at the bow and another at the stern there would be a third crest between. In other words, there would be two sea -waves to the ship's length. The Tunisian is 520 feet long, which would make the apparent wave-length 260 feet. In my endeavours to get at the truth about the dimensions of sea -waves, I have done three things : first, made measurements myself ; second, examined the records of measurements made by others ; third, consulted officers on all ships by which I have travelled as to what their experi- ence leads them to suppose the height and length of waves to be. Now, the officers of the Merchant Service have far more experience of weather in the open oceans than most naval officers, for ships of war keep mostly to the vicinity of land. A fortiori their experience is far greater than that of landsmen such as myself. On the other hand, those to whom I have spoken on the subject have not actually made measurements. The result of my three lines of inquiry is as follows : As regards the height of waves, the general 82 WAVES OF THE SEA opinion of the officers of the Merchant Service accords with remarkable closeness with the measurements which have been made by Dr. Scoresby, Lieutenant Paris, Captain David, and Mr. Ogilvie. My own measurements confirm both, as far as the heights of waves in the North Atlantic are concerned. " About 40 feet " for the fairly frequent larger waves in an ordinary North Atlantic storm is the general verdict of the officers on the liners, and they are generally prepared to concede a few feet more for waves in exceptional storms, especially in the region of westerly winds in the Southern Ocean. From the records discussed in detail above I find that the larger waves in ordinary North Atlantic storms attain 43 feet and in ex- ceptional storms both here and in the South Indian Ocean attain, and perhaps surpass, 45 feet. The possibility of an occasional peak of water shooting up to a height of 60 feet before breaking is some- times admitted, but those whom I have consulted generally feel that there is little to be gained by guessing at the figures applicable to such circum- stances when they have had no reliable standard of measurement. In the matter of wave-length in the North Atlantic, however, I find the general opinion pf officers on the liners to be that 600 feet would be an enormous wave-length, and, if intended as an AND OTHER WATER WAVES 83 average, and not merely the distance between a single pair of crests in a confused sea, would not be met with there in ordinary storms. Such a sea, in which a large ship of 500 feet long running directly before the wind is left, time after time, within the trough of the wave, they have only witnessed in the Southern Ocean, par- ticularly in the part east of the Cape of Good Hope, where the sea is more regular and is prob- ably longer than that near Cape Horn, though perhaps not higher. Thus, it is only in the Southern Ocean, par- ticularly in the eastern parts, where the waves are not only large but regular, that the officers' estimate of wave-length agrees with the measurements from speed and periodic time. I have not yet traversed the Southern Ocean, but in my efforts to judge wave-length from on board ship in the Irish Channel, the Mediterranean, East China Sea, Caribbean, North Pacific, and North Atlantic I almost despaired of getting any results worth recording on account of the dis- crepancies above described. It seems as if the measurements from velocity and period were a nearer representation of the natural state of the sea than those gained by officers from their ex- perience on the bridge. Yet the latter must not be lightly dismissed, and more attention should I 84 WAVES OF THE SEA be given to unravelling this part of the subject. My explanation of the systematic effect of the ship's waves to shorten the apparent wave-length I believe to be an important part of the whole explanation, but it may not be the sole cause. It is possible that the observed elevation of the wave-crest above the ship's flotation line is some- times increased on account of the presence of a ship -wave tending to make the recorded heights of waves too great, especially when observed from large, fast ships. On the other hand, a large vessel, among waves shorter than herself, neither rises to the crests nor sinks to the troughs, so that in observations such as my own on the Minnehaha and the Ivernia the recorded height of the wave-crest above the still- water-line of the ship is probably less than the height above the trough. Thus, these two possible sources of error tend to neutralise each other, and as the heights recorded by eye on large ships agree with those recorded on smaller ones, and both are in accord- ance with aneroid determinations, as far as these have been carried, we may regard them as probably free from any large systematic error such as that which the ship-wave, and perhaps other superposi- tion, introduces in one of the methods of measuring wave-lengths. CHAPTER III The wave-length of the swell which reaches the shore after storms — The height of the swell at sea during storms — The co-existence of waves of different lengths — The give and take between air and water in the develop- ment of waves. The Wave-length of the Swell which reaches the Shore after Storms WHEN the waves produced upon the deep sea run into water of which the depth is less than i-4th the wave-length, theory shows that their speed is reduced 6.7 per cent., and when the depth is no more than i-8th of the wave-length their speed is only 4-ioths of what it was in deep water.1 This reduction of speed goes on as the wave approaches the shelving shore. Finally, in water of which the depth is only a small fraction of the original wave-length, the wave -speed is the same in all cases, no matter what was the original speed and wave-length in deep water. This progressive 1 Vide Sir G. G. Stokes' " Admiralty Manual," 1886. 85 I 86 WAVES OF THE SEA change of speed causes the wave -crests to close up, so that the space separating them gets less and less as they approach the shore, but the interval of time between the arrival there of successive crests is unchanged. Consequently, by timing the arrival of a number of breakers, or the passage of a fixed point by a number of waves just before breaking, we know at once the period of these waves, not only as we see them in shallow water, but as they were in water so deep that the wave dis- turbance did not reach nearly to the bottom. From the period we can calculate the speed in deep water by the mathematical theory of waves, using the formula : Period (in seconds) = speed of wave in feet per second ~ 5^ (nearly), or, more roughly : Period (in seconds) = speed of wave in knots per hour -r- 3 ; and we can calculate the wave-length for deep water thus : Square of period = length of wave -f- $J. The breakers which arrive somewhat irregularly during storms do so at intervals which, as far as I have noticed them, do not differ much from the intervals observed on board ship in the deep sea during storms ; but after storms the period of the AND OTHER WATER WAVES 87 breakers is sometimes much greater. Now, this means that their wave-length in deep water was much greater than that observed during storms at sea, and that the speed in deep water of these subsequent waves was also much greater than that of the waves then observed. But when the wind no longer acts upon the water, there is no agent to accelerate the waves. Theoretically, they should travel by gravity at the same speed as that which the action of the wind has induced, or, if there be any change of speed while in deep water, it could only be some very gradual diminution. A real increase in the speed of transmission may be pro- nounced with confidence to be an impossibility in the absence of any new source of energy. The only possible explanation, therefore, is that the waves of longer period which come in upon the shore after storms are present, but escape observation, during storms at sea. Not only so, but they have attracted comparatively little atten- tion from on board ship even after storms. The following observations of the swell, follow- ing a westerly storm, were made by myself at Branksome Chine, on the Dorsetshire coast, between Bournemouth and Poole. The storm had been violent and long at Bournemouth, and the weather in the Atlantic exceptionally stormy. I 88 WAVES OF THE SEA On the morning of December 29, 1898, the wind, now light, having drawn more to the north, the sky was brighter and the weather pleasant, as it generally is here when the wind blows from the north-west after the passing of a storm from the Atlantic . At 1 1 a .m . an unusually heavy surf came rolling in upon the sandy beach, the waves maintained unbroken almost to the shore by the action of the off-shore breeze, which, at Branksome Chine, always favours the production of a well- formed breaker. At ii a.m. came four large breakers at the following intervals, viz. : 22, 1 8, 20, average 20 seconds. They were followed by an almost smooth sea ; and then, after a short time, came a second series of four large breakers at the following intervals, viz. : 16, 22, 19, average 19 seconds. At 11.25 a .m . a fine series of seven breakers* arrived at the following intervals, viz. : 21, 17, 22, 23, 23, 20, average 21 seconds. I did not notice any minor waves between the members of the series. At 1 1.35 a.m. a set of six breakers came at the following intervals, viz. : 16, 19, 25, 16, 24, average 20 seconds. AND OTHER WATER WAVES 89 And at 11.50 a.m. a set of seven breakers at the following intervals, viz. : 17, 21, 16, 22, 23, 18, average 19.5 seconds. The general average of the periods of the five sets of waves is 19.9 seconds. Later in the day large breakers followed one another in continuous succession without the occur- rence of any smooth water, and for three-quarters of an hour I noted, watch in hand, a succession of 139 consecutive breakers, occurring at the following intervals, viz. (reading from left to right): 20, 20, IQ, 21, 19, 19, 17, 12, 20, 20, 19, 22, 21, 21, 20, 22, 15, 18, 19, 17, 17, 20, 22, 2O, 18, 12, 15, 19, 16, 18, 18, 22, 21, 19, 20, 18, J5(?), i5(?)i 16, 19, 20, 18, 20, 25, 16, 21, 25, 10, 12, 22, 21, 18, 21, 19, 21, 21, 23, l8, 14, 15, 15, 19, 19, 14, 21, 18, 19, 14, 15, 21, 18, 15, 22, I9, 17, I9, I3, 20, 23, 24, 15, I4, 17, 21, 19, 20, 28(?), 12, 24, 19, 20, 19, 19, 21, 18, 16, 17, 22, 26, 20, 16, 21, 22, 23, 25, 17, 15, 16, 23, 21, 20, 15, 15, 18, 19, 26, 19, 21, 21, 18, 21, 19, 19, 18, 23, 18, 19, 22, 20, 20, 15, 23, 20, l8, 2lJ, 2l£, 21, 20. The average interval was 19 seconds, and this determination is accurate to a degree not attained by observations from moving ships. The recorded intervals between the successive breakers probably vary more than the real intervals, because if the I 90 WAVES OF THE SEA observer is too late in his determination of the time of breaking for any wave, he not only makes that interval too long, but the next too short. Thus we see that we had here a fairly uniform series of waves whose average period, 19 seconds, shows that their velocity, when in deep water, was 57 knots, or 66.5 statute miles, per hour, and their average wave-length 1,850 feet.1 It was not easy to make exact comparison between the height of these breakers and those which I have seen at Branksome Chine during storms, for the latter break farther out and in deeper water. However, the impression conveyed to the eye is that there is no very great difference between the heights of the breakers above the water in front of them in the two cases. Branksome Chine is sheltered from the west by the promontory called Purbeck Island, so that the waves of westerly gales and the swell which follows are both reduced by change of direction in rounding the promontory. If, however, we recall to mind the numerous references in nautical writings to the great breakers which come in during calm weather upon oceanic islands and other shores directly ex- posed to the ocean, we shall recognise that their 1 This is about equal to the distance from Park Lane to Devonshire House, measured along Piccadilly. AND OTHER WATER WAVES 91 height must be of the same order of magnitude as that of the breaking waves of storms. On December 29, 1898, in N. 47°, W. 19°, waves 45-52 feet high were reported from S.S. St. Simon. In absence of details I have not used these in the above records of height of waves, but I take them here to calculate a maximum length of storm -wave in the Atlantic during the exceptionally wild weather prevailing at the time of my observation at Branksome. We have already seen that when the length of waves is determined in the manner which gives the highest values (viz., by recording speed and period of arrival from on board ship), tnaximum waves of 40 feet are associated in the North Atlantic with lengths not more than 600 feet, or fifteen times the height . The corresponding length for a height of 45 feet is 675 feet, and for 52 feet, 780 feet. The latter is a figure prob- ably considerably in excess of the real average wave-length during the storm, for which 600 to 700 feet would be a more likely estimate. Never- theless, if we took the excessive value of 780 feet for the average length of the storm -waves, this would be less than one-half the average length of the swell observed at a distance. On February i, 1899, I recorded at the same place a swell of even longer period. The day 6 92 WAVES OF THE SEA was calm, after some days of light winds from north and east. A heavy swell began to come in upon the shore at about n a.m., and between 3 and 4 p.m. I timed the arrival of twelve succes- sive breakers, between which the eleven intervals were as follows : 23, 23, 22, 22, 22, 23, 21, 24, 24, 22, 22, average 22-5 seconds. This is the longest period which I have ever observed for a group of waves. It was noted at the time as a good observation — i.e., the determina- tion of the times was rendered easy by the regu- larity in the form and progress of the waves. Multiplying the period by three, we obtain with sufficient accuracy the speed in deep water, ex- pressed in knots, which is therefore 67.5. This multiplied by 7-6 gives the speed in statute miles per hour, viz., 78.75. The corresponding wave-length in deep water is 2,594 feet. Although observed in the English Channel, it is an accepted view that such a swell is due to an Atlantic storm. The period is almost twice as great as that (11.7 seconds) recorded by Dr. Scoresby in the storm of March 5-6, 1848, and the speed therefore is like- wise nearly twice as great. In a later chapter an account will be given of the probable origin of this set of breakers and of the distances they had travelled ( see p . 1 1 8 ) . AND OTHER WATER WAVES 93 On September 16, 1900, in fine weather at Branksome Chine, I heard the boom of surf, and, looking from my window, timed a set of well- formed though not large breakers at the following intervals, viz. : 18, 17, 19, 23, 19, 21, average 19-5 seconds. The following observation of a swell from the Atlantic was made on the north coast of Ireland, near the Giant's Causeway, in the autumn of 1870, by the late Sir G. G. Stokes.1 One morning a grand surf came rolling in. There had been, some days before, a long succession of heavy gales in the North Atlantic. The period determined from different sets of six or eight waves was 1 7 seconds . The average difference between the mean periods of the different sets of waves was only about i-5th of a second. The differences between the periods of individual waves is not recorded, but would, of course, be much greater. Somewhat later the period sank to 16 seconds, in the latish afternoon to 14, and next day to 13. The surf was highest for the 1 7 -second period. During several other summers, when Sir G. G. Stokes spent a month or two on that coast, he never saw anything so striking. 1 Discussion in Section A, British Association, Dover, 1899, on a paper by the present author. I 94 WAVES OF THE SEA Some interesting deductions can be made from the above record. The mean between the period at commencement (17 seconds) and that next day (13 seconds) is 15 seconds; and in 24 hours 5,760 waves of this period would have discharged themselves upon the shore. The length of a i 5 -second wave is 1,153 feet, so that 5,760 such waves in series would occupy a space of 1,090 geographical miles. The length of the waves at the front of the group was 1,481 feet, and of those at the rear 866 feet. If we suppose the waves travelling freely after the storm, the rate of progress of the group, if reckoned by that of a 15 -second wave, would be, not 45, but 22| knots,1 so that the advance per 24 hours would be 540 geographical miles, and the interval between the storms and the arrival of the swell was " some days." The character of the group of waves at the moment when the storm ceased must have been 1 Vide Nature, vol. xvi., 1877, p. 343, for Osborne Reynolds on the relation between group-velocity and wave-velocity in deep water. In deep water a group of trochoidal waves travelling freely under the action of gravity advances at half the speed of the individual waves. If we follow the motion of the first wave of the group, we shall find that it dies out, and the wave behind it has now taken the lead. If, on the other hand, we watch the last wave of the group, we shall soon find that another one has appeared behind it, and the sum total of these AND OTHER WATER WAVES 95 very different from that at the end of their long journey. In the first place, the waves which would have been noticed and measured from ship-board would have comprised none of the length of even the shortest seen at the Giant's Causeway. They would have been the waves from 400 to 600 feet long, with periods of less than 1 1 seconds, and they could not have arrived at the Giant's Causeway until later. But when they did arrive — as arrive they must— it is evident that they gave rise to no remarkable breakers, for the surf was decreasing two effects may be represented by the following scheme, in which we take the front wave and place it at the back : — 1234567 7123456 6712345 5671234 4567123 3456712 2345671 1234567. In each succeeding horizontal row the individual waves have moved forward two places, but the group has advanced only one. The late J. Scott Russell's early recognition of two wave velocities appears to have been forgotten. He says, in his Report on Waves, B. A. meeting at York, 1844 : " I have found that the motion of propagation of the whole group is different from the apparent motion of wave transmission along the surface ; that in the group whose velocity of oscillation is as observed, 3-57 feet per second, each wave having a seeming velocity of 3*57, the whole group moves forward in the direction of transmission with a much slower velocity." 96 WAVES OF THE SEA as the period decreased, and observations were dis- continued when the period was 13 seconds. Theory shows that when deep-sea waves are left to them- selves, the time which they take to flatten out so as to become invisible is proportional to the square of the wave-length.1 Hence the longer waves are more permanent. Again, waves of all lengths are reduced to the same small speed by the time they break upon the shore ; so that waves originally long and swift are shut up, or telescoped, to a much greater extent than the slower ones, and they undergo in the process a correspondingly greater increase in height. On both accounts, therefore, the height of the breakers given by the swifter waves is increased out of all proportion to their original height during the storm. TJ- Is as w CD w £ ta o W ^* S II AND OTHER WATER WAVES 99 be represented by replacing the theoretically infinite number of wave-lengths by two, which I call the storm -wave, or the dominant wave, and the swell, the former 30 feet high and the latter 20. The lengths arbitrarily chosen are 600 feet for the former and 1,150 for the latter. It will be noticed that the combined wave in the third line represents fairly well the appearance of a toler- ably regular sea, and it would do so still better if the curves had been drawn in the form of a trochoid instead of a curve of sines, which makes the trough similar to the crest. Measuring from left to right, the vertical distance from each crest to the trough next on the right is : 22*50 feet 37-50 „ 1875 „ 40-00 „ 27-50 n Average 28-30 feet. The average, if carried on through the whole gamut of the combined wave, would be 30 — i.e., the same as the height of the dominant or storm- wave. Thus, in the above example we have a range of observed wave -height from rather less than 20 feet up to 40 feet. This result, correspond- ing to the heights observed by me in the storm of 100 WAVES OF THE SEA December 7, 1900, is due to a swell of two-thirds the height of the storm -wave and nearly double its length, the hypothetical height of the swell being 20 feet. This, however, is not an observation but a case chosen for examination. On June 10, 1885, Abercromby, using an aneroid as already described, observed individual wave-heights of 26, 21, 23.5, and 26 feet (average 24), but obtained an absolute difference of level between lowest trough and highest crest (not, however, one of those recorded above) of 35 feet. Let x = height of storm-waves, „ y = height of swell, Then x + y = 35 feet ; and, as we saw from the diagram referred to above — x = 24 (the average apparent height), therefore— y = 35 -24= ii feet. This recorded case, therefore, is consistent with an amplitude of swell nearly one half as great as that of the storm -wave which dominates the eye. Had I not drawn for myself combined curves to test the effect produced by combining together un- dulations of a certain length with others nearly twice as long, I should not have supposed it possible to obscure a longer undulation of such considerable amplitude. AND OTHER WATER WAVES 101 The diagrams, which I published in Knowledge in 1 90 1, and afterwards in the Geographical Journal, May, 1904, show what is geometrically and numerically possible in the way of such obscuration. The observations made on the north coast of Ireland by Sir G. G. Stokes (as well as my own observations) show how necessary it is to allow for the presence with the storm -waves of longer waves of considerable amplitude. Abercromby's single observation supplies a confirmation of the apparently somewhat extreme case shown in the figure given on p. 98. It is, however, much to be desired that further observations similar to those of Abercromby should be obtained, for it must be confessed that it is a narrow foundation upon which to build. On the Co-existence of Waves of Different Lengths In studying deep-sea waves one is often called upon to decide whether to fix the attention upon the actual surface of irregular or complex form and deal with it as the concrete wave, or to think of this irregular wave as composed of, or resolved into, a number of simple harmonic waves each of regular form and of a different speed. As a matter of mathematics all irregular waves can be so re- 102 WAVES OF THE SEA solved, and the behaviour of the roughened sea, when the wind drops, exactly performs this act of analysis — i.e., of resolution into components. On account of this physical circumstance it is impos- sible to obtain a thoroughly satisfactory under- standing of deep-sea waves as long as we restrict our mental view by the limitation of our eyesight, which generally shows us only one set of waves of somewhat irregular form running in any particular direction. Other sets of waves crossing these may be perceived by the eye, but for the present I am only concerned with those which run in one direction.1 The following is a generalised description of what I actually see when the wind blows upon water. 1 A curious condition occurs when two deep-sea swells meet one another from exactly opposite directions. I observed such a case from R.M.S.P. Atrato on June 27, 1910. We were bound for Barbados from Southampton, and had passed a few hours before St. Michael's, in the Azores. All the way out from the English Channel we had been accompanied by a north-westerly swell. From the Azores to Barbados the sea and swell were from a south-easterly and easterly direction. On the day in question the south-easterly swell met the north- westerly, both being of only moderate height. The appearance was that, again and again, a great round-topped billow formed, which did not travel, but (a furrow appearing along its summit) quickly became double-crested, the two crests then travelling away in opposite directions. AND OTHER WATER WAVES 103 First comes the simultaneous creation of a uniform pattern of minute waves all over the sur- face, then the growth of the waves to leeward, so that after a time there is a certain size of wave corresponding to the distance from the windward shore, which size is not afterwards exceeded. The growth of these larger waves at any place is accom- panied by the failure and partial obliteration of the shorter waves which were there before, and this is due to the circumstance that the growth in height of the shorter waves is hampered by the vortex or eddy of the air caused by the larger series of waves. At each position there is finally a definite length of wave which is the dominant wave for that locality for the actual force of wind —i.e., the class of wave which so dominates the eye that any shorter wave there appears as a mere ripple upon its surface and any longer wave is only to be detected by the presence of a sort of heaving motion which runs through the whole system of the dominant waves. On small sheets of water, or near the windward shore at sea, this swell is insig- nificant, but as the length of fetch of the wind increases it becomes an important part of the whole disturbance. This fact is most easily understood if we consider what happens when the wind ceases and the waves are left to themselves. Travel over 104 WAVES OF THE SEA a considerable distance analyses the originally com- plex, irregular waves into series of simple, regular waves of graduated length. The longer and swifter are in front, the shorter and slower are in the rear. But this is not all, for the shorter com- ponents flatten out very quickly as they travel, whereas the longer components preserve their height with but little diminution for long distances. Consider now the effect of this upon the surface- water at a place far distant from the windward shore. The short -length (and therefore short- period) waves will reach this place so flattened that they will produce no appreciable effect, and may, therefore, be regarded as not reaching it at all, and the water will heave with a long-period undulation, the surface exhibiting therefore only long, swift waves. Now, this gravitational travel, with its accompanying analysis of the wave -com- ponents, must go on in just the same way when the wind is blowing as it does after storms. There- fore at a considerable distance from the windward shore the state attained by the sea during a storm does not depend only upon what the wind does there, but also upon the transmission by gravity, independently of the wind, of the longer -period components of the irregular waves which the storm has created to windward. The greater the length AND OTHER WATER WAVES 105 of fetch the greater is the distance from which the surface-water draws the reinforcement of its long-period heaving, and the greater, therefore, is that part of the wave disturbance which is of greater wave-length than the dominant wave. What, now, precisely is this dominant or storm- wave, and how is it evolved? The answer is not difficult if we think of the mode of motion of the wind as it blows over the wave -water. There must be a continual give and take between the wind and the water, such that the air above tends to go into a regular series of travelling vortices or eddies, with long -extended horizontal axes, rolling along in the hollows between the crests of a regular series of travelling water-ridges. Above this series of travelling eddies the air must flow in undulating lines, the amplitude of the undulations diminishing with the height above the water -surf ace, so that at a considerable altitude the air flows in straight lines. When the sea has attained to an approxi- mately steady wave condition under the action of the wind, there is superposed upon it a train of wind-eddies (and above them, aerial undulations), which are of regularly increasing size for a long distance from the windward shore. At each suc- cessive position as we recede from the windward shore there is a characteristic, and successively 106 WAVES OF THE SEA larger, size of dominant or storm-wave, and this is the wave of length identical with that of the air-eddy. These air-eddies, as has been already said, hamper the development of shorter waves whose full growth would require the existence of shorter air -eddies ; but they have little effect upon the longer and flatter swell, which possesses great energy, stirs the water to considerable depths, and is continually reinforced by gravitational transmis- sion from great distances.1 Thus it seems that in the attempt to extract precise and even numerical results for the visual observation of waves at sea, we may neglect that part of the wave disturbance which is of shorter wave-length than the domi- nant or storm-waves ("the waves" of common parlance), but that we must not ignore the swell, which is of greater wave -length.2 1 I suspect the existence of another mechanism contributing to the same result. Referring once more to the swell running by gravity — suppose this to have attained a regular gradation of wave-length. Then each section will presumably be opaque to and absorb vibrations of its own period and transmit or be transparent to vibrations of greater period. 2 The theory of the action of wind to increase the height of waves already running before it is that the horizontal velocity of the air being greatest at the crest, the downward pressure of the atmosphere is least there. Conversely at the trough, where horizontal velocity is least, downward pressure is greatest. Hence the trough is pushed farther down and the crest is sucked up. CHAPTER IV The velocity of the wind at sea — The numerical relation between velocity of wind and average height of waves — The relation between the velocity of the wind and the velocity of the waves — The connection between the rate of progress of cyclones and the character of the waves — The effect of squalls and gusts upon waves — On wave-fronts in a veering wind, and on the irregularity of the waves in the region of the Trade winds — The observed profile of waves at sea — The mountainous appearance of waves. The Velocity of the Wind at Sea IN order to arrive at an opinion upon the rela- tion between the velocity of the wind and the size and speed of the waves of the sea, we have to rely mainly upon the conventional numbers entered by the navigating officer in the ship's log, by which he expresses the force of the wind as he judges it to be. The watch on the bridge being of four hours' duration, the number generally represents the average force of the wind throughout this time. The connection between these conventional num- bers and the velocity of the wind has been deter- mined by experiments with anemometers. In the r 107 108 WAVES OF THE SEA case of the ordinary cup -anemometer the number of revolutions recorded during " strong breeze," " moderate gale," &c., had been compared with those obtained by whirling the instrument on a turn-table, by which means a counter air-current of known velocity is produced. The earlier experi- ments indicated that the number of revolutions of the standard cup -anemometer must be multiplied by three to give the velocity of the wind in statute miles per hour. More complete experiments, however, have shown that this " reduction factor " was much too high, and the factor, or multiplier, now adopted by meteorologists is 2.2. Hence the velocities of wind found in records of some years back are greatly in excess of the values now adopted. Wherever the actual logged number ex- pressing the sailor's estimate of the force of the wind can be obtained, it is, however, easy to calcu- late anew the velocity of the wind in statute miles per hour, and this has been done in the present book. Thus the velocities of wind quoted by me as observed by Lieutenant Paris are not those stated by him in metres per second, but those recalculated from his logged numbers. The follow- ing table is taken from a paper l by Mr. R. H. 1 Q.y.R.M.S., Jan., 1897, vol. xxii., No. 101, pp. 24-55, discussion on pp. 56-61. AND OTHER WATER WAVES 109 Curtis on " An Attempt to Determine the Velocity Equivalents of Wind Forces Estimated by Beau- fort's Scale" The third column gives what is usually called the velocity of the wind, by which is meant its average velocity. But the velocity of the wind varies very rapidly. This is particularly noticeable when the wind is strong. Every gale is gusty and affected by squalls, and columns 4 and 5 sho.w the average range of wind velocity, corresponding to each of the numbers or degrees of the Beaufort scale. Description. Beaufort's Number. Average Velocity of Wind in Statute Miles per Hour or "the Velocity" of the Wind. Average Maximum Velocity of the Wind. Average Minimum Velocity of the Wind. Calm O 2 O Light air I 4 5 3 Light breeze 2 7 9 5 Gentle breeze ... 3 10 J3 7 Moderate breeze 4 14 18 10 Fresh breeze ... 5 *9 25 J4 Strong breeze ... 6 25 33 18 Moderate gale ... 7 31 41 22 Fresh gale 8 37 47 27 Strong gale 9 44 58 31 Whole gale 10 53 73 36 Storm ii 64 83 45 Hurricane 12 77 ? ? 110 WAVES OF THE SEA The Numerical Relation between Velocity of Wind and Average Height of Waves The heights of waves discussed in detail in pre- ceding chapters are those produced when the wind has had opportunity to develop them fully, and most attention was given to the size of the maximum waves then produced. Circumstances frequently prevent the waves from attaining the full size which the velocity of the wind is capable of producing1, of which fact examples were noted in my voyage on the Ivernia (p. 53). The numerical relation between the velocities of wind and average height of wave, obtained from the records of daily obser- vations on long cruises, depends in part on cases where the velocity of wind has no physical relation to the height of the waves. Such averages, therefore, blur the truths, which the writer, as a student of physical geography, desires to eluci- date. There is, however, a practical point of view from which these averages may be useful — that, viz., of the naval architect. As the ships which he designs may have to sail on any seas and to encounter all weathers, it is sometimes desirable to eliminate local conditions. The tables given in my paper in the Geogra- phical Journal, May, 1904, show how closely con- AND OTHER WATER WAVES 111 cordant are the results from the data of Desbois, of Paris, and of Antoine, when recalculated by the modern reduction -factor of wind velocity. On an average the velocity of the wind in statute miles per hour was found to be 2.05 times the height of the wave in feet. Thus the average height of the waves in a whole gale, Beaufort's force 9, wind velocity 44 s.m.p.h., is : 44 -r- 2-05 = 21-5 feet. As already explained, this average would be exceeded when the wind had full opportunity to do its work. The Relation between the Velocity of the Wind and the Velocity of the Waves The greatest average length of storm -waves on any one day recorded in the preceding pages is that observed by Lieutenant Paris, viz., 771 feet, which corresponds to a wave velocity of 43 statute miles per hour. The average velocity of the wind, as re-calculated from the conventional number in Paris 's table, was 46 statute miles per hour. The gale had lasted (or the ship, running before the wind, had been in it ) for four days, or, say, 100 hours. During the first day of the gale the wave-length was only 371 feet, corresponding to 112 WAVES OF THE SEA a speed of 30 statute miles per hour. Thus, during the first day the air blew over the travelling wave- crests at 16 statute miles per hour, and during the last day at 3 miles per hour. During the storm in the Southern Indian Ocean (described on p. 73) Captain David, of the S.S. Corinthic, estimated the average length of the waves at about 675 feet, which corresponds to a wave velocity of 40 statute miles per hour. The wind was logged as 9 on Beaufort's scale, so that its average velocity must be taken as 44 statute miles per hour. The air, therefore, blew over the ridges of the travelling waves at an average speed of 4 statute miles per hour. During the storm in the North Atlantic (de- scribed on p. 42) the wind was recorded by Scoresby on March 5th as a " hard gale," which I take to be the same as a " strong gale," number 9 on Beaufort's scale, corresponding to an average velocity of 44 statute miles per hour. Now, the average length of the waves measured by Scoresby on March 6th was 560 feet, corresponding to an average velocity of 38 statute miles per hour. The ship, which was running before the wind, had then been exposed to the gale for about 40 hours, but the force of the gale had by this time some- what abated. When the gale was at its height AND OTHER WATER WAVES 113 the wind was therefore 6 s.m.p.h. swifter than the waves at their swiftest. In the Mediterranean and other semi -enclosed seas, even where the water is deep, the length and speed of the waves are much less. Conse- quently the effective velocity of the wind, which maintains the eddy on the lee of the travelling ridges, is much greater. In all the above cases, which are typical, the velocity of the storm -wave is a few miles per hour less than the velocity of the wind as averaged over a period of from 4 to 100 hours. The period of the swells which break upon our shores after storms shows that they travel when in deep water at much greater speeds than do the highest waves of storms. Thus the calculated speed of the unbroken series of 139 waves which I observed at Branksome Chine on December 29, 1898, was 66.5 statute miles per hour, and that of twelve successive waves observed on February i , 1899, was ?8 -5 statute miles per hour. Other observations recorded at the same locality, of which particulars have already been given, show that velocities of between 68 and 78 statute miles per hour are normal, though not frequent, for break- ing swells coming to our shores from the west after storms in the Atlantic. 114 WAVES OF THE SEA Many considerations crowd in upon the mind when we endeavour to reason upon the physical connection between the swells of this speed #,nd the pressure of the wind during the storm. How far, for instance, can we regard the longest of these subsequent swells as having had an independent existence during the storm? Eor the present, at all events, I shall set aside such refinements, and simply consider the numerical speed-relation of the swiftest observed swells to the swiftest observed winds in the same part of the world, and see where this will lead us. The storms in the North Atlantic during De- cember, 1898, and January and February, 1899, were of such exceptional violence that the Meteoro- logical Council made them the subject of a special inquiry, which has beene mbodied in a valuable report,1 commenced by Lieut. C. W. Baillie, R.N., and completed by Commander Campbell Hep- worth, R.N.R. The charts show that between December 25th and 29th very strong westerly winds prevailed between the Newfoundland Banks and the entrance to the English Channel. In the notes to these charts it is recorded that the velocity of the wind at Alnwick Castle, Northumberland, 1 Charts illustrating the weather of the North Atlantic Ocean in the winter of 1898-9 (Meteorological Council, 1901). Hwi^ * v \ v'S-'-rr-rrX^'^ ski LOCAL NOON 301" JANUARY, 1899. WEATHER IN THE NORTH ATLANTIC. (From a report of the Meteorological Council, 1901.) AND OTHER WATER WAVES 117 attained 77 statute miles per hour at 10 p.m. on the 27th, and 7 i statute miles per hour at 2 a.m. on the 28th. Beaufort's 11-12 (i.e., 64-77 statute miles per hour) was recorded at sea in N. 49°, W. 35°, on 28th, and N. 52°, W. 19°, on 29th. Thus it is proved that the winds that blew had at times an average velocity 10 statute miles per hour greater than that of the swell observed at Brank- some Chine on December 29th. I am not able, however, to derive the observed swell from the hurricane arrows shown on the charts for noon December 28th and 29th, for they are too far off for the swell, if travelling with the group velocity of 34 statute miles per hour, to reach Branksome Chine at the observed time. A feature of the chart for the 29th which is worth noticing is the existence of W. and SW. winds of Beaufort's 9-10 (44-53 statute miles per hour) near the entrance to the English Channel. In the moderate depths of the Channel the speed of the swell must have been so reduced I that the observed velocity of wind on the morning of the 29th was sufficient to exert pressure upon the swells. This may account for their reaching Branksome Chine with a height unusually great for this locality, and much greater 1 The maximum speed of a wave in 20 fathoms is about 43 statute miles per hour. 118 WAVES OF THE SEA than the height of the swells observed on February ist. The charts throw a clearer light upon the swell observed at Branksome Chine on February ist. This had a period of 22.5 seconds, and therefore a speed in deep water of 78.5 statute miles per hour. A group of these swells travelling freely for a long distance under the action of gravity advances at 39.25 statute miles per hour. Now, the charts for the days January 29th to Febru- ary ist show Beaufort's 11-12 (64-77 statute miles per hour) only on January 3Oth, and the positions of the arrows showing force 9-10 on 3 ist and ist are too distant to affect the water at Branksome Chine on ist. The cyclonic system of Janu- ary 3oth half-way between North America and England is a very deep isolated depression, with wind of force 11-12 in N. 45°, W. 41°, the direc- tion of this wind being a little S. of W. This wind, and that a little in advance of the black arrow, would send a swell almost directly to the entrance of the English Channel. The distance of the straight run from N. 45°, W. 41°, to Cape La Hogue (which I shall take as the distance to Branksome Chine) is 28°, 1,680 geographical miles, 1,960 statute miles. The difference of time between local noon January 3oth in W. 41° and AND OTHER WATER WAVES 119 3 p.m. February ist at Branksome Chine (about W. 2°) is about 48 hours. The westerly swell at N. 45°, W. 41°, if starting at noon on January 3Oth, would reach Branksome Chine at the required time if it travelled at the speed of 1,960 -~ 48 = 40.8 statute miles per hour. Now the observed group- velocity of the swells recorded at Branksome Chine on February ist was 39.25 statute miles per hour. There is, therefore, strong ground for regarding that swell as being the result of the storm which occurred two days before at a distance of nearly 2,000 statute miles. From Mr. F. J. Brodie's paper » on " The Prevalence of Gales on the Coasts of the British Isles during the Years 1871-1900 " we learn that in the great storm of December 22, 1894, the wind had a velocity of not less than 53 statute miles per hour for 14 hours (as observed, of course, at a fixed station), 64 statute miles per hour for 9 hours, and 76 statute miles per hour for 2 hours. On January 12, 1899, 53 statute miles per hour was maintained for 6 hours and 70-76 for i hour. The velocities attained in gusts are recorded in another paper.2 In January, 1899, a rate of 90 statute miles per hour was recorded at Southport * Q.y.R.M.S., 1902. 2 Symonds, Met. Mag., May, 1900. 120 WAVES OF THE SEA in one gust and between 80 and 90 in several others. The highest recorded velocity in any gust was that registered by a Dines' pressure tube anemometer at the Rousdon Observatory, South Devon, viz., 101 statute miles per hour. The results may also be expressed in this way, viz., that, at a fixed station, a wind velocity of above 70 miles per hour has been maintained for as long as 2 hours, but velocities of from 80 to 100 s.m.p.h. have only been maintained for minutes or seconds. The greatest period of any short group of swells which I have observed is 22.5 seconds, with a speed, therefore, of 78.5 statute miles per hour. M. Bertin ' says that 24 seconds (speed 84 statute miles per hour) is certainly beyond all observed periods in European waters. The greatest speed which I have observed in a group of many swells was 66.5 s.m.p.h. on December 29, 1898. The greatest recorded average velocities of wind in that weather were 77 and 71 s.m.p.h. on 27th and 28th of the same month. The greatest speed which I have recorded for a group of a few swells is 78.5 s.m.p.h. on February i, 1899. The greatest speed of wind 1 " Experimental Study of Waves," Inst. of Naval Architects, 1873. AND OTHER WATER WAVES 121 in gusts during that weather was frequently 80 s.m.p.h., ranging up to 90 s.m.p.h. in one case. This was in January. As far as the evidence goes in this difficult part of our subject, the indication is that the maximum average velocity of wind maintained for i hour is a few s.m.p.h. greater than the average maxi- mum velocity of any long group of swells originating therefrom and forming breakers upon the shores of England. It is conceivable that shorter groups of greater speed may owe their origin to gusts, but during lulls these would be running against a current of air. The maximum recorded wind velocity in gusts in 1898-99 was n| m.p.h. greater than that of the swiftest of the short groups of swells. On the Connection between the Rate of Progress of Cyclones and the Character of the Waves Strong winds in the North Atlantic are developed in that part of an area of low atmospheric pressure where the barometric gradient is steep. These atmospheric depressions are very frequently of the form and nature known as cyclones, of which the general character is that shown in the accompany- ing diagram. 122 WAVES OF THE SEA The oval indicates the area covered by the depression, the long arrow the direction of advance, the short arrows the direction of wind in certain parts of the cyclone. In the position marked A the direction of the wind is contrary to the direction of advance of the cyclone. Hence in this quadrant the cyclone is continually re- ceding from the waves which its wind creates, and A CYCLONIC SYSTEM. along the line of advance of this part we should not expect that there would be any great develop- ment of waves, even if the winds there were strong, which they are usually not. Obviously, for a given velocity of wind, the position most favourable to the development of waves is that where the direction of the wind coincides with the direction of advance of the cyclone. This is where the direction of AND OTHER WATER WAVES 123 the wind is shown by the arrow B. Here, moreover, at about 4-ioths the distance from centre to edge, the strongest winds are usually developed. The line of advance of B will, there- fore, be the line along which the greatest wave- development will occur. There is not any constant relation between the rate of advance of a cyclone and the velocity of the winds locally developed within its area. Considering only the critical posi- tion B in the cyclone, the rate of advance of the cyclone is, from our present point of view, simply the rate of advance of the locus of the force which is there creating the waves ; we have to do, in fact, with waves created by a travelling disturbance . Let us consider groups of simple harmonic waves of different lengths, and therefore different speeds, to be already formed and to be travelling together (and therefore superimposed upon one another) as forced waves pressed upon by the wind. This wind, however, in a progressive cyclone is a " travelling disturbance," and while it will, to some extent, increase all the waves beneath it which it can press upon at all, the waves which move slower than the travelling disturbance are being left behind all the time. On the other hand, all waves whose velocity is greater than that of the travelling dis- turbance run ahead of it, and are no longer subject . 124 WAVES OF THE SEA to the reinforcing action of the wind. The waves whose speed is identical with that of the travelling disturbance will be continually subject to the re- inforcing action of the wind during the whole life- time of the cyclone. I think, therefore, that the wave-length of the principal, or dominant, or storm -wave generated in Atlantic storms should depend, not only upon the velocity of the wind, but also upon the rate of advance of the cyclone. It is, I believe, a matter of common remark that in some storms a short, steep sea is soon formed and in others a longer sea. The general rate of advance of Atlantic storms towards our shores has been investigated,1 and the figures are worth examining in connection with our subject. Storms advancing from points between WSW. and WNW. to ENE. and ESE. travel at an average speed of 28.9 statute miles per hour. Of the 264 storms examined, only 60 travelled at more than 35 statute miles per hour, and of these only 10 travelled at more than 52^ statute miles per hour. It appears, therefore, taking the average of a large number of storms coming from the Atlantic, that waves travelling about 30 statute miles per 1 Q.J.RM.S., 1902, loc. oil AND OTHER WATER WAVES 125 hour1 (length 371 feet) should enjoy special opportunities for development in depressions where the wind exceeds that velocity. We see, therefore, that waves of 68-78 statute miles per hour, which are the greatest speeds I have recorded for swells after Atlantic storms, would in almost all cases outrun them, even when the depression advances along a straight path. The reason that such waves are not developed to greater heights is, therefore, not only that the wind is, during most of the time, not strong enough, but also that the cyclones advance too slowly. The above method of theoretical treatment is of only occasional application on account of the fact that there are generally several neighbouring cyclonic systems on the North Atlantic and that the actual winds are a compromise between them. Fortunately, the charts for December 27, 28, and 29, 1898, enable us to examine an instance of this commoner condition. A series of cyclonic systems were following one another across the North Atlantic from S. of W. to N. of E., and strong westerly winds prevailed during the whole of these three days between N. 50°, W. 40°, and 1 This is a wave of a little more than 8 seconds period. The 8-seconds wave has a speed of about 28 statute miles per hour and a length of 328 feet. 8 126 WAVES OF THE SEA the entrance to the English Channel. The unusual force of Beaufort's 11-12 was at local noon on December 28th recorded in the neighbourhood of N. 50°, W. 37°, and not elsewhere, and on local noon December 2 9th in the neighbourhood of N. 49°, W. 19°, and not elsewhere. As the storm does not appear to have abated meanwhile, we may infer that the locus of the hurricane wind progressed continuously along this path of 840 statute miles in the course of these 23 hours. This is at the rate of 36.5 statute miles per hour. The average wind velocity for 11-12 is 70.5 statute miles per hour. Waves travelling at nearly this speed would be running on ahead all the time. The waves which would be all the time from noon Decem- ber 28th to December 29th subject to the maxi- mum force of wind would be those with a speed of 36.5 statute miles per hour, and, therefore, with a length of about 558 feet. As they would be subject to an effective wind velocity of 70.5 — 36.5 = 34 statute miles per hour, there would be a great deal of spraying from their crests. Two hours at a fixed station is the greatest time recorded above for wind with an average velocity of more than 70 statute miles per hour. If the locus of that wind force were advancing at 36.5 AND OTHER WATER WAVES 127 statute miles per hour, as in the above case, the length of fetch, or stretch of water, at any one time subject to this wind would be 73 statute miles, in which there would be a train of only 200 of the 1,918 foot -waves. It may in this connection be noted that in a gust lasting i minute, and in which the locus pf application is advancing at the same rate (36.5 statute miles per hour), the length of fetch would be only about half a mile — i.e., less than two pf the above wave-lengths. The absence of breakers with speeds nearly as great as the maximum wind speed of gusts may, therefore, be properly attri- buted as much to their insufficient length of fetch as to insufficient time of action. The charts for February 2 and 3, 1899, in the publication of the Meteorological Council to which I have so often referred, afford an excellent example of a long -continued hurricane -force of wind which was stationary in position for 24 hours. After examining these two maps, I made the following note, viz : " Maps for Eebruary 2nd and 3rd show hurricane -force (centre about N. 45°, W. 45°), with no appreciable length of fetch indicated, occupying almost the same considerable frontage for 24 hours. In this area must have been a 128 WAVES OF THE SEA very wild sea with the waves spraying very much, from which must have emerged a very long swell." After making this memorandum I referred to the notes on page 6 of the publication, and found : " On February 2nd . . . the Quernmore, in N. 41°, W. 46°, reported that during the storm the atmosphere was so heavily charged with spoondrift as to appear as a heavy snowstorm." My reason for diagnosing an exceptional amount of spraying was that the locus of the hurricane - force did not advance, so that the wind would not have been able to develop the longer waves. The Effect of Squalls and Gusts upon Waves On December 22, 1906, I was on board the Leyland S.S. Jamaican, bound for Puerto Colombia from Liverpool, in N. Lat. 38° 21', W. Long. 35° 43. There was a heavy sea and a moderate gale. At 4 p.m. a violent squall of wind, with rain, occurred, lasting about 4 minutes, which was accompanied by very big waves, and succeeded by comparatively calm water. I guessed the largest waves during the squall to be 7 feet higher than those which preceded or followed. On the following day, December 23rd, we were in N. Lat. 35° 47^', W. Long. 39° 4$', with a AND OTHER WATER WAVES 129 strong breeze verging on a moderate gale — i.e., a wind velocity of about 28 statute miles per hour. At 3 p.m. there occurred a squall lasting 3 minutes, which converted a moderate to a large sea. I esti- mated that the height of the waves increased ,at least 2 feet per minute, finally attaining rather more than 20 feet. Not only did the waves increase in height, but the crests lengthened out transversely, so that the waves ran in longer and more regular ridges. The squall came from the starboard and abaft. Two minutes after it had passed us we were among waves no larger than before, but I could see a group of about four great ridges travelling away with the squall. The length of the waves, as judged by their appearance in relation to the ship's length, was never more than 200 feet. As has been already explained, this estimate is generally less than the measure- ments obtained by following the waves in their course . In another squall, occurring at 5.2 p.m. and lasting 5 minutes, the height of the waves was perceptibly increased. Next day, December 24th, we were in N. Lat. 33° 3$', W. Long. 43° 58'. The princi- pal " waves " were mounds of water, produced by the crossing of two sets of waves. At 3.30 p.m. 130 WAVES OF THE SEA the wind began to fall, and at 4 p.m. the sea was slight, except for the occasional formation of a larger mound of water. At 4.55 p.m. a black being very dense and apparently rainy. The part, head. It reached from the horizon on one side of us to the horizon on the other side, the two ends being very dense and apparently rainy. The part of the cloud directly overhead was thin, and dis- charged only a few small drops of rain. The passage overhead of this central part was accom- panied by only a slight additional breath of wind, but iwith it came a big swell, comprising 1 2 or more large waves, and the ship, previously steady, began to roll heavily. In 5 minutes the cloud had passed, and in another i o minutes the sea had quite returned to its former state. These observations relate to squalls occurring towards the end of stormy weather, coming upon a fairly large swell running in the same general direction as the wind. They show that (according to my guesses of height) a sudden increase of wind can restore the height of waves at the rate of i or 2 feet per minute. They illustrate the fact that, when the system of travelling ridges is already formed, the wind, falling into eddies between them, has greater power to raise the crests and depress the troughs, a power of swift and immediate action AND OTHER WATER WAVES 131 which contrasts strongly with the slight power of the wind when blowing upon smooth water. My observations from the S.S. Jamaican relate, as I have said, to the action of squalls at the tail end of a storm. Their action during the height of a storm is often to lower the waves. Thus, during my voyage from Liverpool to Boston on the Cunard S.S. Ivernia, December, 1900, of which some par- ticulars have already been given, I often saw the effect (which others have frequently described) of showers of spray from every wave -crest during each gust. The steepness and the slowness of the waves were such that the cohesion of the water at the crests could not resist any increase of the upward suction and the horizontal pressure there. Hence in the gusts the water was torn in showers of spray from the crests, falling finally, one must presume, mainly in the sheltered troughs, thus tending to flatten out the sea. On March 1 1, 1901, on the Red Star S.S. Vater- land, east-bound from New York to Southampton, I saw the much greater lowering effect of a wind meet- ing the waves. We were in N. Lat. 44° $6'* W. Long. 36° 54', with a rough sea from the star- board, the waves occasionally rising above the horizon of our promenade deck. At about 4 p.m. 132 WAVES OF THE SEA the wind quite suddenly chopped round and blew from the opposite direction, meeting the waves in their advance. The ridges of the waves, as usual, were not uniform, but had, on the contrary, an undulating outline. The wind caught the higher portions, and bit huge pieces out of them which momentarily formed a milk-white cloud, which in its turn was quickly dissipated in spray. On Wave -fronts in a Veering Wind, and on the Irregularity of the Waves in the Region of the Trade Winds Wind is never really steady. Not only is it always more or less gusty, but it is always veer-* ing — i.e., changing its direction. Apart altogether from the progressive variation in the general direc- tion of the wind which is characteristic of a cyclonic system, there is a rapid veering about a mean position, even in the Trades. The amount of this veering is sufficient to exercise an important effect upon the character of the waves and the appearance they present. It results in the formation of waves running simultaneously in slightly different directions, and thus, even in the regions of the Trade winds, the open sea does not present a series of parallel ridges, AND OTHER WATER WAVES 133 each one of uniform height, with a lateral extension many times greater than the distance from crest to crest. In 1907 I made the following observations on the want of regularity of waves created by the Trade winds. On January 8th, en route from Colon to Kingston, Jamaica, in a strong NE. Trade wind, I estimated the height of the principal waves at 15 feet. Their apparent length was about 200 feet. I noted at the time that this strong breeze had continued, to my knowledge, for eight days, and that for some days past there had been no increase either in the size of the waves or in their regularity, indicating, therefore, not only that the maximum size is soon attained (as M. Bertin l has pointed out), but also that there is no continuous approximation towards regularity. The captain of the vessel (S.S. Jamaican) stated that on such a day as this the wind would not vary more than one point of the compass in the course of a four- hour watch, and this is reckoned a very steady wind. But one point of the compass is 1 1 J degrees of arc, which is a very appreciable angle. If one watches the arrow of a sensitive weathercock in a strong breeze, in our own country, it will be seen 1 Memoir on the " Experimental Study of Waves," Inst. of Naval Architects, 1873. 134 WAVES OF THE SEA that it never remains stationary for more than one or two seconds, but is continually shifting through a large angle. Another cause can also be divined for the unlevel tops of the wave -ridges and their small lateral ex- tension. This cause is a greater force of wind to right or left of the ship's course. Suppose, in the first place, that the ship is in a calm, and that to the starboard and in front there is a head wind blowing. Then a swell will spread out laterally from that area, meeting the ship obliquely on the starboard quarter. Secondly, if a lighter head wind be blow- ing where the ship is, the wind-waves will be crossed obliquely by this swell from the starboard, even though the wind there be blowing in the same direction as that where the ship is. The velocity of the wind being less where the ship is, it will not be able to regularise the obliquely-running, swifter swell. The wind -formed waves which run in the most regular ridges (i.e., of the greatest lateral exten- sion as compared to their wave length) are of two orders of magnitude. First, the small waves of a few feet in length and not more than one foot high; second, the large ocean waves. The regu- larity of the first is not affected by long-period veering of the wind, for they die out completely in AND OTHER WATER WAVES 135 the interval and fresh sets are formed. The regu- larity of the second is not visibly affected by short- period veering, being too massive. Waves of intermediate size are affected by both kinds of veering and several crossing sets are formed. The Observed Profile of Waves at Sea Both during a wind and in the case of a swell travelling calmly in deep water, the trochoid repre- sents the apparent form of the wave much more nearly than the curve of sines, for it is easy to observe that the convex part of the wave has a steeper curve than the concave part, the crest being narrower than the trough. But, except in the case of a somewhat flat swell, the greater steepness of the front face of the wave is obvious to the eye, and most markedly so when the wind is blowing. Now, Sir G. G. Stokes found * that waves of permanent type in deep water, what- ever be the order of approximation to which the calculation is pushed, must be symmetrical with respect to vertical planes passing through their ridges. I conclude, therefore, that steep waves at sea are never waves of permanent type. Again, 1 " Math, and Phys. Papers," vol. i., p. 193, from Trans. Camb. Phil. Soc., vol. viii., p. 441. 136 WAVES OF THE SEA both Stokes and von Helmholtz showed that for a single series of waves of permanent type the condition under which greatest steepness could be attained was that the speed of wind and wave should be equal. Now, according to observation, the steepest wind-waves in deep water are formed where the speed remains small relatively to that of the wind— e.g., in lakes and small seas. Also in the Trades, where the wind is perpetual, there is no sign or symptom of an approach to the " highest wave " of Helmholtz and Stokes. I conclude, therefore, that observation indicates that under natural con- ditions of wind there is no tendency to progressive approximation towards the state of one set of waves of permanent type. In this connection, I note that Professor Horace Lamb, who has investigated water-waves as a mathematician, writes l that " the possible form of waves of permanent type ... is very interesting mathematically . . . but no reason has been given, so far as I know, why free water-waves should tend to assume a form consistent with per- manence." 1 Presidential Address, Section A, British Association, Cam- bridge Meeting, 1904. AND OTHER WATER WAVES 137 On the Mountainous Appearance of Waves The phrase " waves mountains high " has long given offence to literary landsmen, who quote measurements to show that waves are not even as high as hills. I have never met a seaman who either supposed or pretended that waves were as high as mountains, but there are conditions when waves look like mountains, or, at any rate, like large hills. This happens when the majority of the crests rise well above the line of sight, especially if the atmo- sphere be rather thick, so that minute detail is obliterated. Four or five ridges, with the inter- vening three or four troughs, then fill all the space between the eye and the horizon. Being mounted on a deck, there is a feeling or impression that the horizon is at the distance which it would have on land with such an eye -elevation. This would mean a mile or more from ridge to ridge, which is ten times the actual distance ; and the apparent height is consequently increased in the same ratio, making a wave of 40 feet look as high as a hill of 400. I have seen, and recorded,1 a case where absence 1 Geographical Journal, Jan., 1900, "On Desert Sand-dunes Bordering the Nile Delta." 138 WAVES OF THE SEA of surface detail produced a somewhat similar illu- sion of great size among desert sand-dunes. Par- ticularly when under a low sun they produced as great an impression of size as mountains thousands of feet high — i.e., of ten times their real height. NOTE TO PAGE 60. In storms of exceptional duration as well as of more than ordinary severity, such as are not encountered every year, the waves are somewhat higher. Thus in February, 1910, the R.M.S.P. Oruba, between Southampton and Barbados, en- countered waves which, from the account given me by Captain C. P. Langmaid, appear to have attained 45 feet. In March, 1904, Captain J. G. K. Cheret, on another of the R.M.S.P. steamers, encountered between Southampton and the Azores exceptional weather, of which he has given me an account. When he was on the bridge the waves frequently obscured the horizon at times when the ship was on an even keel, from which it appears that they surpassed 45 and may have attained 50 feet. PART II ON THE ACTION OF SEA WAVES TO TRANSPORT SHINGLE, SAND, AND MUD CHAPTER V On the depth to which wave-agitation extends, and on the transport of fine mud — The action of waves to drive shingle shorewards — The effect of percolation to promote the building-up of beaches — The movement of sand by waves — The undertow. On the Depth to which W