The Brendan Voyage

by Tim Severin

  But this must remain speculation, even for the most imaginative detective hoping to solve the riddle of the Navigatio. A more tantalizing clue is being examined on the coast of Newfoundland. There, overlooking the entrance to Saint Lunaire Bay, archaeologists were investigating some lines incised on a boulder and trying to decipher their meaning even while Brendan was sailing across the Atlantic. The lines on the rock appear to be man-made and to have been cut with a fine pointed tool, possibly of metal. Some of the lines are obscured beneath an encrustation of lichens which covers most of the rock surface. But one criss-cross pattern seems to have been rubbed clean for inspection at some time in the past, and what looks like a fresh lichen growth has begun to grow again on top of it. By chance this fresh lichen includes a variety that grows at a measurable rate, and it has been calculated that even this growth is at least 150 to 200 years old, which is longer than the present habitation of the area. The lichens covering the other lines may be far older, and some observers have been quick to claim that the criss-cross lines are ogham, the early Irish form of writing, particular on stone, favored by the Christian monks. Scholars of ogham are doubtful, and certainly the inscriptions are very cryptic, if not completely undecipherable. Work on them is continuing, including an attempt to date the more heavily encrusted markings and to ascertain if the lines really were incised with metal. But there are other candidates for their authorship: Saint Lunaire lies not far from the presumed Norse settlement at L’Anse au Meadow and the marks could have been cut by a straggler from that camp; and there is the wreck of a cannon-bearing vessel in the waters of Saint Lunaire entrance, right under the lookout where the mysterious rock is set in the hillside. Perhaps a survivor swam ashore and cut the marks with a knife. Whatever the answer, there is plenty of room for speculation and further investigation, particularly now that Brendan has shown that one group of suspects could certainly have reached the scene of the “crime” at the time alleged. Clearly the file on North American’s discovery by Europeans is far from closed.

  APPENDIX III

  BRENDAN

  DESIGN

  Brendan’s design was based on three sources—ethnographic, literary, and archaeological. The ethnographic data were the most fruitful and included the design and construction details of the Irish curraghs of the Dingle pattern as described by James Hornell in “Curraghs of Ireland,” Mariner’s Mirror, January 1938. Hornell’s data was compared with the present-day Dingle curraghs, which range up to 21′4″ in length for the “four-hand” size, and augmented with information from John Goodwin, curragh-builder of Castlegregory, County Kerry. The literary evidence for the historic use of leather boats in western Europe is widely scattered throughout many early texts, and the majority of references have been collected together by G. J. Marcus in his “Factors in Early Celtic Navigation,” Etudes Celtiques, Volume 6, 1952. Leather boats, for instance, are mentioned by Caesar, Pliny, and Solinus. But the most relevant details for Brendan were found in Adomnan’s Life of Columba and, of course, the Navigatio Sancti Brendani Abbatis itself. Early Christian artifacts in the Irish National Museum, Dublin, show how wood, leather, and metal were worked in Ireland, and this information decided the actual construction techniques of Brendan. With particular reference to leather-working techniques, John J. Waterer’s study, “Irish Book Satchels or Budgets” in Medieval Archeology, Volume 12, 1968, was invaluable. Pictures of early boats are rare, but fortunately an outline of an open Irish boat can be seen carved on the vertical shaft of an Irish stone cross, still in situ near Bantry, County Cork. This carving has been dated to the eighth century A.D., and is described by Paul Johnson in Antiquity, Volume 38.

  In designing the boat, Colin Mudie was concerned with keeping the hull weight as low as possible to allow for a heavy load of stores and the intended use of the craft on long open-water passages, beaching, and propulsion by sail and by oar with a crew of only four or five. Also for stability, the sections were made rather firmer than those of some current smaller curraghs. His design was for a boat 36’ overall and with a beam of 8’. The estimated weight was 2,400 pounds for bare hull, plus 1,284 pounds for sails and rowing equipment. With stores, water, crew, and sea-water uptake into the leather, Brendan’s final cruising displacement was close to five tons. Her sail area was 140 square feet on the mainsail and 60 square feet on the foresail. Both sails were made in flax by Arthur Taylor and Son of Maldon, and could be extended by the addition of a bonnet, 3 feet deep on the main and 2 feet deep on the fore, at the foot of the sails. The addition of side panels was found to be effective only in a wind from dead astern. Steering was by a large-bladed paddle on the starboard quarter, though twin steering paddles and a steering sweep were tested in sea trials and found to be ineffective or unnecessary. Oars of different lengths were also tested, and the most suitable length was found to be 12 feet. Oars worked on thole pins and “bulls”—triangular pieces of wood pegged to the oarshaft—and the oar blades were of the traditional curragh pattern, being extremely slim.

  MATERIALS RESEARCH

  Perhaps the most fascinating and rewarding phase of the Voyage preparations was the research into the performance of the medieval materials. The aim was to try to determine if the medieval materials had any chance of standing up to a trans-Atlantic voyage. Some of the tests were very simple—for example, samples of leather were hung on frames in sea water to see if they would decompose or pick up barnacles. Other tests were conducted in the laboratory, thanks to the enthusiastic cooperation of the British Leather Manufacturers’ Research Association at Milton Park led by Dr. Robert Sykes, and the associated work by the tannery of W. & J. Richardson of Derby, and by Harold Birkin’s team at the tannery of Joseph Clayton and Son in Chesterfield. There was plenty of room for error, because although leather is a remarkable and rewarding material to work with, it must be handled absolutely correctly or it will be ruined. In preparing the leather for Brendan’s hull, the tanners and scientists had to be concerned with such matters as surface crack, the tightness of the grain in the leather, the lower surface tension of salt water, and a host of other considerations, including the dimensional stability of damp leather. To general satisfaction, the lesson was learned that the medieval materials, if prepared correctly, were exceptionally suited to the task of crossing the Atlantic.

  Schematic diagram of Brendan

  1. Hull leather. The Navigatio states that Saint Brendan’s monks covered the wooden frame of their boat with “oxhides tanned with the bark of oak” and carried similar oxhides on board as spares, and to make two smaller skin boats presumably as tenders to the larger vessel. Dr. Sykes’ researchers tested samples of oak-bark leather supplied by Josiah Croggon and Sons of Grampound, Cornwall, and compared their performance to other types of leather tannages under similar conditions. The properties that were sought were resistance to water; dimensional stability, particularly after immersion in salt water; and tensile strength. The following table, supplied by the B.L.M.R.A., indicates the remarkable stability of the dressed oak-bark leather throughout various testing procedures:

  TABLE I

  In effect, the tests showed that traditional oak-bark leather was very strong even when wet; that this strength did not diminish; and—from earlier data available at the Research Association—that the comparatively open-fiber structure of oak-bark leather rendered it particularly suitable for “stuffing” with a waterproof grease or dressing. Croggon’s of Grampound eventually supplied 57 “butts,” oxhides trimmed of their upper shoulder and leg skin, for the boat. These butts were oiled with cod oil, rolled, and rough-dried.

  2. Dressing. The dressing of the 57 butts with grease was undertaken by Joseph Clayton and Sons under the supervision of Harold Birkin, whose own laboratory had been testing samples of greased leather in conjunction with the B.L.M.R.A. The Navigatio stated that the monks used “fat for preparing hides to cover the boat” and “smeared all the joints of the hides on the outside with fat,” so only fats and greases which were know
n from early medieval times were tested—tallow, beeswax, fish oil, and wool grease, singly and in mixture. The correct dressing was vital to the durability of the leather hull. The problem was not merely to grease the leather well enough so that it kept out the water, but to grease it without damaging the quality of the leather. If too much soft grease was put into the leather fibers, the leather became too limp; too little grease, and the leather was not water-tight. After very many experiments, it was found that the most suitable dressing was raw wool grease. This wool grease was given through the generosity of W. & J. Whitehead (Laisterdyke) Ltd., Top makers and Worsted Spinners of Bradford; and was put into the leather by dipping the oxhides in a bath of wool grease at 50°C. for up to twelve hours and then leaving the hides stacked for weeks, with grease between the layers to ensure complete penetration of the grease into the leather. This produced a 30- to 37-percent uptake of grease. The monks might have achieved the same result much more slowly by hand, applying successive dressings of wool grease to damp oxhides, or by pouring on hot grease and rubbing it in by hand. The success of the collaboration between Joseph Clayton and the B.L.M.R.A. in producing Brendan’s hull leather can be judged from the comparison between the leather first supplied to Brendan (Column a) and the same leather after it had crossed the Atlantic (Column b). In sum, there was no major change in the composition or performance of the leather from the beginning to the end of the voyage.

  TABLE II

  TABLE III

  TABLE IV

  3. Thongs. W. & J. Richardson of Derby provided the two miles of leather thong which were used to lash Brendan’s frame together, also the thongs employed on board in place of short lengths of rope. Initially five different types of thong were sent to the research laboratories for testing and their values were shown as follows:

  Alum-dressed thong—a system known at least in Roman times—was selected as the most suitable, and the thongs were treated with a tallow and fish-oil dressing before they were sent to Brendan.

  4. Thread. Henry Campbell and Company of Belfast made the 23 miles of flax thread which stitched Brendan’s oxhides together, and also made up the improved flax ropes for her rigging with the help of Belfast Rope Works. At the suggestion of James Henshall, the director concerned, samples of Campbell’s flax thread were tested (in conjunction with Brendan oak-bark leather) by the Shoe and Allied Trades Research Association. Table IV shows how well the flax thread performed in these tests, growing stronger when wetted and stronger again when sandwiched in oak-bark leather. It can also be assumed that the thread became more rot-proof when the tan from the leather migrated into the thread and discouraged bacterial decomposition.

  CONSTRUCTION

  The scantlings in Colin Mudie’s design for Brendan were of the order of 1″×6″ oak for the gunwales, 2″×⅝″ ash for frames, 2½″×⅝″ ash for stringers. The fore-and mainmasts, 12 and 19 feet long respectively, were initially cut as single ash saplings, but these developed splits during trials, and were replaced by single lengths cut from close-grained ash under the personal supervision of Paddy Glennon of Glennon Brothers, County Longford, who also provided all ash used for frames, stringers, and oars. The masts were set in socketed mast-steps of oak which rested directly on the keelson.

  Pat Lake and Michael Murphy of Crosshaven Boatyard faired up the hull frame, which was then fastened with alum thong lashings. Each thong was pre-stretched by hand to reduce its initial elasticity, soaked, and then hand-lashed in place while still wet. The fastened hull frame was then liberally coated with wool grease to preserve the timber. Subsequent inspection showed that the wool grease had penetrated deeply into the wood grain.

  Forty-nine oxhide butts were needed to cover the boat. The hides averaged about one-quarter inch in thickness, and measured on average 45″ × 47″.

  The thread to stitch the hides was “fourteen cord,” i.e., handrolled from 14 single threads. It was then waxed with a combination of wool grease, beeswax, and resin so that it filled the awl holes when drawn through. John o’Connell, harnessmaker, and Eddie Hinton, master saddler from the Walsall saddlery firm of Eldonian Brookes, decided the most suitable method of stitching. This was back-stitch for the hides along the center line of the boat, and double-handed stitching for the remainder of the work. The double-handed stitching had to be done by pairs of stitchers stationed opposite one another on each side of the leather. Stitch length was approximately two to the inch, and double lines of stitches were put in, approximately 1 inch apart. Every joint was a plain overlap of about 2 inches width. At the prow and the leading ridge of the keel, where collision and grounding were anticipated, John o’Connell overlapped four thicknesses of hide for extra strength. And at the stern an extra fillet of leather was applied at the keel to protect against stern beaching.

  The keel-skid of oak was fastened through the leather with ½″ copper rivets. Riveting was a very highly developed technique in early Christian Ireland. The leather skin was not otherwise fastened directly to the hull frame but tensioned over it, pulled over the upper gunwale, and lashed down to the lower gunwale with hair-on round leather belting. Harold Birkin suggested this unusual use for hair-on belting, which was supplied by the Lion Leather Company.

  PERFORMANCE

  1. Rowing. Five weeks after landfall at Musgrave Harbor, Brendan was in a very different setting to the wild Newfoundland coast. In sight of the skyscrapers of Boston on the lower Charles River she was rowed by ten volunteer oarsmen from the Union Boat Club. The purpose of the experiment was to try to find out just how many oarsmen, albeit inexperienced in rowing leather boats, were needed to row Brendan effectively under different states. Table V illustrates what we had also learned from the voyage itself: namely that Brendan was severely undermanned with a crew of only four. But it also shows that a crew of ten oarsmen could make headway in all but very adverse conditions; and that under more favorable conditions a rowing crew of six or eight men could edge a cruising curragh upwind. It should be stressed, however, that the tests, our own ocean experience, and the medieval texts indicate that long-distance voyages in Irish leather boats would have been made under sail. oars would normally have been used in emergencies, when coasting, or when in sight of a landfall.

  TABLE V

  ROWING TESTS

  Average of up- and downstream runs over 510-meter course, lower Charles River, in westerly wind Force 3 with a current of c. 1.3 knots.

  NOTE: A coasting load was calculated as the stores and equipment needed for a curragh making day passages within sight of land. A full ocean load was calculated as the weight on board of water, food, and gear needed for a 3-week ocean passage for 10 men.

  2. Sail. The lack of suitable instruments made it difficult to measure accurately Brendan’s performance under sail. obviously it was impractical to fit a “through hull” log impellor in an oxhide, and the only feasible instrument was a trailing log. A Walker Knotmaster was used, and the manufacturers warned that their instrument was ineffective at speeds below two knots. However, on a cruising basis, the following observations were made:

  a. The maximum distance achieved in a 24-hour period was 115 miles. The minimum day’s run was, of course, nil in a flat calm, and on bad days Brendan was actually set back on her course by adverse winds. The average day’s run under sail was 40 miles, and a cruising speed of 2 to 3 knots was considered satisfactory. This required a wind F 3–4 where the boat was fully laden. In ideal conditions of a following wind of F 5–6 the log would register sustained periods of 5 to 7 knots. The maximum reading on the log scale was 12 knots and this speed was achieved comparatively frequently in heavy weather and high seas, though in Mykines Sound in the Faroes a prolonged burst of 12+ knots was achieved due to the funneling effect of the surrounding cliffs.

  b. Leeway had to be judged by eye, taking the angle between the wake and the trailing safety line. Against a headwind Brendan would point about 50° to 60° off the wind, but leeway of 30° meant that she made an effective 90° to th
e wind. At F 6 and above, it became dangerous in most sea states to continue to reach across the wind, and it was safer to run downwind. The use of leeboards may have been anachronistic, as there is no evidence for or against the use of leeboards in the early Middle Ages. But we did not use leeboards above F 5 to 6 because they tended to scoop too much water into the boat, and at lesser wind strengths they were not as effective as lee-side rowing would have been with a larger crew. Several leeboard positions were tried, and a position 16 feet from the bow was found to be most effective. The addition of a second leeboard, aft by the helmsman, was found to be detrimental as it actually increased leeway. In sum, the use of a leeboard reduced leeway by up to 10° when sailing, but did not enable Brendan to point higher. Lee-side rowing by a larger crew would have been more effective.

  3. Sea keeping. The most impressive aspect of Brendan’s performance was her sea-keeping ability even in severe weather. She successfully negotiated prolonged periods of heavy seas and strong winds. The normal tactic was to reduce sail by one reef in the mainsail at F 5–6, and at F 6+ to lower and stow the mainsail, and then turn downwind under headsail alone, which was later reefed. It was the designer’s opinion that Brendan was undercanvased, but this was a deliberate safety factor. In heavy seas both main warps were streamed in loops from the stern and played a vital part in keeping the boat straight to the sea. oil was also spread from a half-gallon canvas bag and was a help. However it was necessary, and often difficult, to stay directly downwind of the oil slick. Extra oil bags on a spreader bar would have made this easier. Fish oil is considered to be better than mineral oil for this task, and whale oil, which we used, the best. A sea anchor was streamed from the bow and proved effective in moderate seas, but it was considered too dangerous to try using the sea anchor in very heavy seas. It was feared that the boat would be held up into the force of the waves, and the protective tarpaulins would be split by the weight of water.