iron strapping

8/9/2019 Iron Strapping

1/17

The Internutionul Journal of hiiuticd Archueology (1996) 25.3 & 4 . 207-223

The United States mail steamer Humboldt, 1851-53:initial report

Trevor Kenchington and Colin WhitelockGudus Assoriutes, R . R # I . Musyuodoboit Harbour, Nova Scotia, Canada BOJ 2LO

IntroductionIn December 1853, the transatlanticsteamer Humboldt , one of the laof her time at 2181 tons (Fig. l), was runashore on the coast of Nova Scotia. In the1960s, her wreck-site was re-discoveredand became a favourite location for sportdiving. A mile away, near Bear Cove(Fig. 2), divers found two large sections ofcoherent wooden ship structure. One ofthese is notable for its many copper plank-ing bolts (Fig. 3 ) . Since similar bolts hadbeen found on the Humholdts wreck-site,these sections were regarded as parts of thebig steamer.This w reckage lies exposed with n o over-burden or associated artefacts and thusoffers an opportunity for the study ofwooden ship structure without requiringexcavation: research that can be pursuedwithin the logistic constraints of an un-funded, amateur project. In 1989, the sitewas therefore chosen for the survey exer-cise compon ent of a N autical A rchaeologySociety training course. Building on thatexercise, we have begun a detailed study ofthis site, in an attempt to understand thewreckage. T he specific aims are:(1) to determine whether the wreckageat Bear Cove is indeed part of theHumholdt;(2 ) if so, to provide a measure of how farcoherent wooden ship structure canmove from an initial wreck-site;( 3 ) to describe some structural features

adopted by the builders of largewooden steamers1057-2414/96/030207+ 17 $2 5 0010

(4) hereby to show that even relativelymodern and badly degraded wreckstructures merit detailed exam ination.This pap er provides an initial report on theproject, based on fieldwork to the end of1994. A n introduction outlines the histori-cal background to the ship and the site.Historical backgroundThe coming of transatlantic steamAfter 1815, oceanic shipping faced a newimperative. In preceding centuries trans-oceanic communications, though oftenrich and varied, had been highly asym-metric. The Western cultures that con-trolled them had their major governments,principal trading houses and prim ary massof consumers concentrated in Europe. Thisasymmetry was forever altered by the riseof the Un ited S tates into a significant inde-pendent power: a centre of Western com-merce and diplomacy that lay across theAtlantic from the European heartland.The consequences of this change for theshipping industry were delayed by therequirement for convoy sailing in war-time. After 1815 , however, transatlanticshipping was no longer a matter of bulkcargoes and passenger transport alone.Henceforth, communication of infor-mation was as important as movement ofgoods and people.Shipowners, particularly in New York,responded quickly, the Black Ball Line,first of the new packet companies, beingfounded in 1818. Their new challenge wasinitially met by developing the old sailing

0 996 The Nautical Archaeology Society


2/17

NAlJ TICAL ARCHAEOLOGY, 5.3 & 4

Fi


3/17

T. K E N C H I N G T O N & C. WHITELOCK: US STEAMER H U M B O L D T

steamer and they therefore petitioned theirgovernment for support equivalent to thatenjoyed by Cunard. The ensuing compe-tition resulted in the Ocean Steam Navi-gation Company being subsidized to carrythe mails between New York and Bremen,the Washington and Hermann opening thenew service in 1847. Their contract allowedfor an expanded service to Le Havreand, in 1849, an associated company, theNew York and Havre Steam NavigationCompany or Havre Line, was incorpor-ated to take up this opportunity. A consortfor the Wushington and Herrnunn,launched in 1848, was taken over while afourth ship was ordered. She was launchedin 1851 as the Hurnboldt (Ridgely-Nevitt,198 ) .

The shipwrights challengeThese big steamers posed a greater struc-tural problem than had any previous ships.The core of the difficulty lay in their con-densers. Until the 1870s, metallurgy couldnot provide a joint that would withstandcold sea water on one side and steam onthe other. Moreover, the tallow used forlubrication in the cylinder clogged thetubes of surface condensers. Thus, marineengines used jet condensers, in which thewaste steam from the cylinders was con-densed by a spray of sea water; the partialvacuum created by the condensation serv-ing to draw in the spray. These workedwell but they introduced salt into the con-densate, from which the boiler feed waterwas necessarily drawn. That salt becameconcentrated in the boilers and in highpressure (and hence high temperature) sys-tems, it was laid down as a hard scaleon the heated surfaces. There it acted asinsulation. not only cutting efficiency butleading to irregular heating of the boiler,Figure 2 Approaches to Halifax. Nova Scotia, which often ended in failure.Meanwhile, tallow also reached the boilerhowing locations named in the text . The dotsnear Portuguese and Rear coves mark theHunrboldts stranding site and th e location of in th e feed water. At high tempratures,the wreckage described. this broke down into fatty acids which

209


4/17

NAI ' T I CAL ARCHAEOLOGY, 5.3 & 4

Figure 3. View of the southern section of wreckage at Bear Cove, show ing the planking boltsprotruding from the frames. Some bolts bear remnants of planking. N ote the heavily erodedstate of the wreckage. The planking bolts are 22 mm in diameter and provide an indicationof scale. All photographs in this paper were taken through a wide-angle lens and haveunnatural perspective. (Author's photograph.)

corroded the engine (Ridgely-Nevitt, 1981;Griffiths, 1993).The salt and tallow could be toleratedin low pressure boilers, which were inter-mittently blown down to empty them ofhypersaline water and re-filled with seawater. Unfortunately, low-pressure steamengines are inherently inefficient at trans-ferring energy from coal to paddle-wheels.So inefficient, indeed, tha t a norm al coasta lsteamer of th e 1830s could not floatenough coal to power herself across theAtlantic.All else being equal, however, th e energyrequired to drive a ship rises with thecross-sectional area of her underwaterbody, whereas her ability to carry coal an dpaying cargo rises with her immersed vol-ume. Thus, large ships could profitablysteam across the ocean. It was exactly thislogic that led Brunel to design the Great2 10

Western at 1230 tons old measure, makingher the largest steamer yet built (Farr,1963; Co rlett, 1975).Iron plates and sections can, in theory,be made to any required dimensions.They can also be fastened together so thatthe joints are almost as stiff and strong asthe pieces joined. Wood, in contrast, isonly available in the sizes to which treesgrow. Moreover, without modern glues,joints between wooden pieces are in-evitably points of weakness and mobil-ity. Thus, large wooden ships werecomposed of thousands of small, flexibly-attached pieces. The larger the ship, therelatively-smaller the individual pieces,the greater the number of joiiits and theweaker the hull. The first part of theshipwrights' problem with the big steamerswas thus that they were necessarily verylarge.


5/17

T K E N C H I N G T O N & C WHITELOCK. US STEAMER H U M B O L D T

Their great weight of coal and enginesalso meant that only a small proportion oftheir displacement could be devoted topaying cargo and passengers, yet the highcosts of fuel and machinery made theseships much more expensive, per ton, tobuild and run than were the sailingpackets. If they were to cover those costswith their reduced relative payloads, thesteamers had to command much higherfreight and passenger rates. Yet their onlyadvantage was the promise of speed andregularity. There w as little to be gained byexpending more power (and hence morecoal and heavier engines) to achieve thespeed. Ra the r, the steame rs were built withfiner lines. For paddlers this was doublynecessary since a bluff bow could cause thebow wave to interfere with the paddles,cutting efficiency and introducing severevibrations (Greenhill, 1993). The highlength-to-beam ratios and lack of buoy-ancy in the fine ends of these ships thenexacerbated the hogging problem thatbuilders of large vessels always struggledwith. Moreover, unlike the ships-of-the-Line and Indiamen, which at least carriedthe weights of their guns an d ballast sprea dalong their lengths, the steamers necessar-il y had their boilers and engines concen-trated amidships, imposing greater stresseson the hull.Lastly, the steamers had to hold theirshape, despite these stresses, in NorthAtlantic winter weather while being drivenhard against a head sea. a situation rarelyencountered by a sailing ship. This was thegreat structural challenge posed to anybuilder of transatlantic steamers. It wasone that they did not always meet; theWushington and Hermann were notablyinade quate (R idgely-Nevitt, 1981). Th eproblem was not fully solved until the mailcontra cts were changed in the mid-1850s topermit iron hulls.From 1838 to 1855, the builders oftransatlantic steamers adopted and devel-oped many features as they sought the

required strength and stiffness in woodenhulls. No comprehensive account of theseis available. However, a m ong the scatteredinformation that has survived, Farr (1963)has noted that the Greut Westerns floorswere filled in solid and then boltedtogether. w ith fo ur rows of 1: (38 mm)[]diameter bolts [each individual bolt being24 feet (7.3 m) long and those in each rowoverlapping by 4 eet (1-2 m) at their ends],while much of her structure was fastenedwith nuts and screwbolts. The Americanships sometimes had their ceiling strakesedge-bolted together, as did the Arugo(2240 tons, 1855), th e Humboldts eventualreplacement. Most prominently, theseships were strengthened with diagonal ironstrapping, in the form of trusses, braces orboth (Ridgely-Nevitt, 1981).The Humboldts career und lossThe Havre Line ordered the Humboldtfrom Westervelt and Mackay of NewYork, who had built the two Bremenships. They had evidently overcome thedeficiencies of those earlier hulls a s the newship drew no criticism. Launched late in1850, she was 282 feet (86 m ) long, 40 feet(12.2m) in beam and of 2181 tons. N odetails of her structure have yet beenfound, except that she was iron strapped.By the time she left New York for hermaiden voyage, in May 185 , she had cost$560,000.Under Captain David Lines, sheimmediately began her regular New York--Cowes-Le Havre-Cowes- New York ser-vice, sailing every eight weeks a nd ma kingsix round-trips each year. The Humboldtand her sister proved reliable, profitableships and enjoyed a special market niche.Much of the mail and passenger traffic onthe transatlantic route ran between Britainand the United States. However, Frenchand Swiss industry was already noted inAmerican markets for producing luxmygoods with the high price-per-unit-volumethat was ideal for transport by steamer.

21 I


6/17

NAUT I CAL ARCHAEOLOGY, 25.3 & 4

The Havre Line captured this trade,while sharing with the Bremen ships thepassenger traffic from continental Europe.The Humboldt did encounter problemsin Janu ary 1852 when, westbound fromCowes, she experienced a series of heavygales. With bulwarks wrecked, paddle-boxes stove in and her rudder damaged,Captain Lines took her into Halifax forcoal and temporary repairs (Ridgely-Nevitt, 1981). This use of Ha lifax as aport of refuge was common for the earlysteamers. The great circle route from theScilly Isles to Nantucket runs close tothe Nova Scotia coast. Halifax was thusthe first port of consequence that thesteamers passed after clearing the Irishcoast westbound and was the one thattheir captains usually turned to when inneed.Late in Octob er 1853, the Humboldt leftNew York on her sixteenth voyage[31.A tLe Havre, she took on a large cargo ofholiday presents an d C hristma s toys sincehers was the last sailing from France thatshould have reached New Y ork in time forthe seasonal shopping spree. She also tookaboard Welsh coal, since Captain Lineswas unable to purchase any French fuel.The American steamers had previouslyhad difficulties with English coal; on theWii.d?itigfons first voyage the heat haddestroyed two sets of furnace grates w ithin24 hours, necessitating a return toSouthampton for replacements (Ridgely-Nevitt, 1981). Faced with W elsh steamcoal. American crews seem to have beenoverly generous with their stoking.The Humboldt left Le Havre on 22ndNovember, reaching Cowes after daylightnext morning. She mad e fast to the Yac htBuoy for a few hours, taking the Englishmails aboard, and then departed for NewYork with 90 passengers and a 450-toncargo. The crossing was rough, with stro ngwinds from south-west to north, the

meather being boisterous with squalls ofrain, hail and snow and some lightning.212

Around noon on 5th December, theengineer reported that the coal was run-ning short and would not last to NewYork. As the ship was then lying abou t 200miles due east of Halifax, Captain Linesdecided to go into that port for a freshsupply. At 0630h next morning, theHumboldt was boarded by a m an namedPowell wh o falsely represented himself as aHalifax pilot. The ship was then somefive miles south-east of Sambro Island,safely in what is now the traffic controlroute.A t tha t point, however, the ship ran in tounusual weather for December: thick fogand calm. The Humboldt neverthelessstood on at half speed and, at 0715h,struck a rock, later supposed to be BellRock. She must have struck hard, as partof the bow was broken off and the shipbegan to fill. Ordering the crew to thepumps, Captain Lines steamed for thesafety of Halifax. As the water rose inthe engine room, he abandoned that planan d turned for the shore. About 0800h, theHumboldt ran aground with her bow in6 m of water a nd her stern in 12 m . Th e fogwas so thick that the land could not beseen, though it was at most 15 m from thebow.When the fog lifted, Captain Linesfound himself a short distance north ofPortuguese Cove. On hearing the news,th e US Consul went down to the wreckwith two small steamers. These took thepassengers, their baggage and the mail,with as much of the cargo as could bereached, and safely delivered them allto Halifax. From there, the CunarderNiagara carried the passengers on toBoston.Meanwhile, the stern of the wreck sankto the saloon deck within three hours ofthe stranding. That night a southerly galeblew in, bringing a heavy sea which grou ndth e Humboldt on the rocks. From themorning of 7th December until the IOth,the weather moderated an d the crew, aided


7/17

T. K E N C H I N G T O N & C . WHITELOCK: US STEAMER H U M B O L D T

by local men, were busy trying to save thecargo. Then, on the 11 h, the wind turnedeasterly producing a rough swell and partof the ships bottom broke up. The nextnight, she split fore-and-aft and on the13th December was declared a total loss.The Royal Engineers made two attemptsto blow up the Humboldts decks toimprove access to the remaining cargo, thesecond blast being successful. Wracked byexplosives and the weather, the Humholdtsoon went to pieces. By early January1854, it was reported that her forward parthad broken up and her afterbody had gonedown in deep water. Where she wasstranded, the 30-m depth contour is lessthan 300 m from the cliff.Little is known of the wreck over thefollowing century. There are records ofsalvage divers working her in 1855, pre-sumably to recover her enginesL4].Other-wise, she was lost to human concern untilthe local advent of sport diving in the1960s. Since then. many divers havesearched the rock gullies just north ofPortuguese Cove and have recovered as-sorted remnants of her cargo. Meanwhile,near Bear Cove other divers found the twolarge sections of coherent ship structurethat form the subjects of the present re-search. Unconfirmed reports speak of ahalf-dozen other similar sections of wreck-age spread between Portuguese and BearCoves, as well as the remains of a paddle-wheel in deep water. There have beenno reports of wooden wreckage southof Portuguese Cove, although the areabetween there and Chebucto Head isintensively dived.Field researchField research has been confined to the twosections of wreckage near Bear Cove. Ageneral examination of the site has beenmade. recording assorted structural details,and a detailed survey of the southern sec-tion has begun. To date, this last hasencompassed the inner sternpost and after

Figure 4. Sketch-plan of site and surroundings.The southern section of wreckage is labelled S,the northern N.and areas, a boulder beachand the subtidal bedrock slope are indicateddiagrammatically.

deadwood. Some loose fragments havebeen raised for detailed study but thewreck has otherwise been left undisturbed.Site layoutThe two sections of wreckage lie nearly20 m below low tide level, at the base of abedrock slope extending from the shoreabout 175 m away (Fig. 4). In the vicinityof the wreckage, the bedrock dips below athin sediment cover composed of sand,gravel and small stones. Sport diversreport that mobile sand sometimes coversthe entire site, leaving only the uppermosttips of the planking bolts visible. In the lastfive years, however, all but the most low-lying parts of the wreckage have been fullyexposed.Of the two sections, the northern one isabout 21 m in length by 7 m in breadth andis oriented approximately northlsouth. Itlies ceiling upwards, as shown by thepresence of hanging knees on its uppersurface. From its lack of longitudinalcurvature, the presence of the knees and amarked upward curve at its western edge,this section appears to be the side of theship, extending roughly from the curve ofthe bilge nearly to the main deck, andseems to be from somewhere about themid-point of the ships length. In contrast,the southern section, which lies about 14 maway, is oriented approximately eastlwest

21 3


8/17

N A I ric A L ARCHAEOLOGY, 25.3 & 4n-I \- -e

CentimetresTnches

Figure 5. Scale drawings of some fastenings found on the Bear Cove site. a: ends of a copper plankingbolt found loose on the site, with the associated rove shown in section; b: rove from the planking boltshown in (a); c: end of a broken, copper-alloy spike of the type used to fasten plank butts (foundin s i t i c in loose fragment of frame timber: The curvature is original and apparently produced as the spikewas driven); d: section of a hardwood treenail remnant found in the same fragment of timber as thespike; e: end of a copper-alloy screwbolt seen in a fragment of frame on the site. Details of the screwthread reconstructed from observations made underwater.

and lies outboard up, as shown by thein-situ survival of some copper sheathing.I t is about 12m long by 8 m wide andtakes the form of a large number of framesheld together by underlying ceiling (Fig. 3).Very little outer planking has survived.Near the western end of this section, theframes are strongly concave towards theirsouthern ends. Inspection of contemporarylines-plans of American transatlanticsteamers shows that such curvature onlyoccurred right aft and only above thewaterline (Ridgely-Nevitt, 1981). Thus,thi> section is from the aftermost part ofthe ship and from the port side.Fustenings and wood typesA typical planking bolt (22 mm in diam-eter, perhaps originally 718 in) was foundloose on the site, with a rove still in place(Figs 5a,b). Both bolt and rove are appar-ently copper. The former is about 800 mmin surviving length, sufficient to extendthrough outer planking, frame and ceiling.One end had been shaped into a slightly214

expanded and rounded head, while theother may have been distorted by post-wrecking damage but appears to showsome spreading as though it had beenhammered into place.One small fragment of frame and a pieceof ceiling, each badly eroded and foundloose on the seabed near the southernsection, were raised for wood analysis.The frame was of some species of oak(Quercus spp.) while the ceiling fragmentwas pine; probably Pinus rigida, one of thespecies known in shipbuilding as southernhard pine. Underwater macroscopic in-spection of the rest of the wreckage sug-gested that all of the principal pieces wereof one or other of these wood species, theceiling being consistently pine, the timbersoak and the outer planking perhaps amixture of the two.

The two fragments raised each bore rem-nants of a number of fastenings. The framehad a planking bolt, a spike (Fig. 5c) of thekind that, elsewhere on the wreckage, wasevidently used to fasten the butt ends of


9/17

T . K E NCH I NG T ON & C . WHITELOCK: US STEAMER H U M B O L D T

Tuhk 1. Some scantlings measured on the siteItem Moulded SidedFrames 9.5-1 2 in/240-300 mm Over 9.5 id 240 m mTraces of planksCeiling in holdCeiling near Berth DeckInner post at load waterline About R d 2 0 0 m m8 in1200 mmft deadwood at load waterlineKeelward edge of' aft dcadwood .- Over 12 id 30 0 mm

About 6 id150mm9-1 1 in/23@280 m m4 id100 mmAbout 18 id4 60 mm~ ~.

planks, a hole for a fore-and-aft bolt(somewhat over I in or 25 mm in diameter)that doubtless once attached adjacenttimbers, three similar transverse holes(0.75 in o r 19 mm in diam eter) and twotreenails (Fig. 5d). The ceiling fragmentbore two treenails. of the same form asthose in the frame timb er. an d two holes ofthe right size for planking bolts. All fourof these ran transversely.Several other loose fragments of thewreckage have been examined in situ. O none of them. a broken piece of frametimber revealed a copper-alloy screwbolt(Fig. 5e) which reached to the end of thehole that had been bored for it . The headof this screwbolt has not been seen.Tlir ufirr drrid\t.ood Lilt(/ i i i n i v .vtcrripostDetailed survey of the wreckage com-menced at the western end of the southernsection. the structures there being theinner sternpost and the after deadwood(Figs 6 & 7). The port faces of thesetimbers lie approximately horizontal butdip as they extend both fhrwards and awayfrom the keel.Th e sternpo st itself is missing. Th e after-most surviving material comprises a row of(probably brass) bolts. 1 in (35 mm) indiameter. arranged on approxim ately 18 in(460 m m ) centres where they pass throughthe inner post. These bolts originally pen-etrated the sternpost. inner post and dead-wood. though there is no evidence of their

ends on the forw ard face of the deadw ood.which instead has traces of the heads ofsome iron bolts. The surviving ends abaftthe inner post have been wrenched andbent.The inner post is badly eroded at itslower end and its heel may be missingentirely. It bears two irregular rows offastenings, the forward one comprisingplanking bolts (one with a rove in place)and the after one spikes (in each casesimilar to those shown in Figure 5). Mostof the spikes are broken off flush with thesurviving timber but those that are nothave flattened heads. These two rows offastenings presumably held the hood endsof the ou ter planking. T here is no sign of arabbet in the inner post to receive thoseends, suggesting that it was formedbetween this timber and a sternpost ofmuch greater siding.The inner post also bears a number oftreenails (of similar dimensions to that inFig. 5d) though their faster erosion relativeto the oak of the post has m eant that theyhave only survived near the limit of sandcoverage. The original function of thesetreenails is unclear since the planks wouldseem to have been fully fastened by thebolts and spikes.The deadwood is composed of fivepieces: the aftermo st may once have been aknee between the inner sternpost andtimbers lying on the keel and is n o w verybadly eroded. Interestingly, the next three

215


10/17

NAUT IC 'AL ARCHAEOLOGY. 25.3 & 4

k ' i qwc 6 . Plan of the surviving deadw ood and attached structures, projccted o nto the plane o f the ship'scentreline. (Scale bars are in metres.) For clarity. the sym bols fo r bolts, spikes, treenails and fasteningholes a re drawn twice the scale size of the corresponding objects . The arrangement of nails on th ecopper sheathing is shown schematically. A11 other features arc shown to scale. as s u r w y e d . Irontii\tcnings are labelled with asterisks, all other metallic fastenings bcing coppcr o r copper alloys. Thercmnants of three cant frames are shaded, as are tw o coaks. A third coak. one treenail and someplanking holts arc hidden under the coppcred planking.

picccs lay more nearly vertical than hori-/ o i i t a l . in contrast to n orma l British ship-building practice. The sixth piece forms aforuard projection. The after face of thisdcadwood wits rayed against the innerstcrnpost while its forward face wasformcd a s a fair curve, tapering to a point.There is no evidence of the keelson which,in normal practice, would have overlainpart o f the forw ard face of the deadwood.Thc presence of rounded bolt heads ont h a t [orward face precludcs the possibility-7 1 h

tha t a lost deadwood wac formerlyattached there.Thc port lateral face of the deadwoodappears to be an entirely flat surface,though the degree of erosion that hasoccurred would conceal any bearding linetha t might once have been present. Therewere certainly no housings cut into thisface for the heels of cant frames.Th e joints between the second a n d thirdand between the third and fourth dead-wood pieces are close-fitting and straight.


11/17

T K E N C H I N G T O N & C W H I T E L O C K : US STEAMER H U M B O L D T

Figure 7 . View of the deadwoo d a nd adjacent structures on the southern section of wreckage.looking towards the form er position of the sternpost heel. Th e forward face of the deadwoodi s in the immediate foreground, Note the rem nant of coppered planking a t lef t centre. theheavily-eroded cant timbers extending across the deadwood in the foreground, and the innerpost and bolts in the background. The latter are 35 mm in diameter. Othe r dimensions maybe found fr om comparison with Figure 6 . (Authors photograph.)

They are locked by transverse coaks, about6 in (150 mm) square in section. The fiveparts are bolted together and to the postand other timbers. As noted above, not allof these bolts were through bolts. Threeiron ones can, however, be traced extend-ing through the foremost piece of the dead-wood towards where the keel once was.The concreted remnant of one of thesesuggests that i t was originally about 2in(50mm) in diameter. There are hints ofother iron bolts o n the keelward edge ofthe deadwood.Overlying the deadwood are the rem-nants of a few cant frames, which have yetto be examined in detail, and traces of asmall area of outer planking. The latterbears copper sheathing which accords inall observable details with standard 19th-century practice, as described by Campbell

(1974). The edges of the copper sheets werenailed on approximately 45 mm centres,while th e faces of the sheets were nailedwith the typical rhomboidal pattern. Thenails were hammered flush but were notindented to produce the quilted appear-ance mentioned in some contemporaryaccounts. This sheathing was laid overwhat appears to have been tarred felt.Where it is less eroded, the inner stern-post has about 12 planking bolts and asmany spikes in a 2-m length. If there wasoriginally a pair of each fastening thehood-end of every plank, this suggests anaverage plank width of about 330 mm or13 in.Further forward, the lateral face of thedeadwood bears surprisingly few remnantsof fastenings; certainly not enough for evenone bolt per plank/frame intersection.

21 7


12/17

NAIJTICAL ARCHAEOLOGY, 25.3 & 4

There are traces of a number of iron boltsand spikes (which may once have fastenedthe heels of some frames t o the deadw ood),a few planking bolts and a scatter oftreenails and empty holes. Some of thelatter may have contained planking bolts.The apparent paucity of plank fasteningsseems to have resulted from the ship-wrights, when working abreast of theforward part of the deadw ood, using shortbolts that ended in the cant timbers, out-board of the deadwood itself. As shownby their surviving remnants, these cantsprojected mo re than 500 mm abreast of thedeadwood at i ts forward end; an ampledistance for a planking bolt in a blind-ending hole. It is still surprising th at thereare not more plank fastenings in the first50cm forward of the inner sternpost , anarea where the planking must have lainvery close to the deadwood if not actuallytouching it.The considerable lateral distance be-theen the face of the deadwood and theplanking which once lay abreast of itclearly shows that the keel rabbet cannothave lain near th e surviving keelward edgeofthe wreckage. Rath er, if the hull taperedin this area to the degree that contempo-rary lines plans suggest was normal, therabbe t, an d hence also the keel itself, musthave lain about 1 m below the structurethat h as survived.Other features of the southern sectionThe remainder of the southern wreckage5ection has only been examined super-ficiafly. Tow ards the forwa rd (eastern) end,the prominent squ are fram es are of typical19th-century paired construction, with theadjacent timbers held together by fore-and-aft iron bolts abo ut one inch (25 mm)in diameter. Near the deadwood, in con-traht. the timbers are close-set and stronglycanted.Between the frames and the underlyingceiling there are traces of diagonal ironstrapping. This was composed of double218

diagonals, trusses that rose towards thestern and braces that dipped towards thekeel as they headed aft. The straps were5 in by 1 in (125 by 25 mm ) in section(Fig. 8). The trusses were let into theframes while the ceiling was scored to fitover the braces. The aftermost trace of abrace extends almost to the forward endof the deadwood, while a truss can befollowed to a point near the load waterlineand about 3.5m forward of the innersternpost.Northern sectionThe northern section of wreckage has yetto be examined closely. It has a completeceiling from the turn of the bilge to th eberth deck beams. However, the fourstrakes immediately below those beams(collectively 0.95 m wide) were only 4 in(10 0m m ) thick in contrast to the 9 in(230 mm) elsewhere. There is no sign of athicker clamp or beam shelf. Above theberth deck , there is n o sign of any ceiling atall.

Below the load waterline, the ceiling isfastened, at least in part, by copper plank-ing bolts which pass through from theoutside of the hull. In most cases, thesebolts end in heads formed over roves. Ifplankin g bolts were used higher in the ship ,they have not survived and may not havebeen copper.The wooden hanging knees of the berthdeck had their heels on the step in theceiling and lay under, not alongside. theirbeams. A significant part of one kneeremains, along with traces of others atintervals of 1.75 to 2.18 m.Edge bolts are not visible in a finishedhull while iron fastenings have survivedless well than the pine ceiling on thiswreck and thus cannot project beyondthe wood. On the northern section, how-ever, the corrosion products of a fewsuch bolts have preserved the surround-ing wood from erosion, confirming thatthe ceiling was edge-bolted. The exact


13/17


14/17

N A U TI C A L ARCHAEOLOGY, 25.3 & 4

straps be placed outside the frames, wherethey remained until the last large woodenships were built. Thu s, the wrecks double -diagonal strapping between frames andceiling suggest that she was built in themid-19th century and, most probably, inthe early 1850s.Equally, while planking bolts (otherthan butt bolts) have rarely been men-tioned in textbooks, Lloyds introducedthe abbreviation C.T. in the 1841 editionof their Register to indicate tha t a ship hadcopper bolts substituted for treenails, adesignation that probably refers to thearrangement seen on the wreckage. It isnot certain that this was then new but thelimited number of ships designated C.T.in 1841 suggests that such bolts were rarein the 1830s. The screwbolt observed o nthe wreckage is equally suggestive ofmid-I 9th cen tury, o r later, construction.This wreckage shows othe r features typi-cal of large American mid 19th-centuryships, the Western Ocean Packets as muchas the big steamers. These include edgebolting of the ceiling an d hard pine on oakconstruction. The only parallel, known tothe authors, of the sub-vertical arrange-ment of the deadw ood pieces seen at BearCove is a marginal illustration in an 1826US Navy contract for construction of aSloop of War (reproduced in Chapelle,1949). Thus, the two sections of wreckageappear to come from a large to very largeAmerican ship built around 1850.The Nova Scotia Museum is currentlypreparing a database of recorded wrecksaround the coastline of the Province.While no such listing can be complete,large ships lost in the 19th century in theapproaches t o Halifax are unlikely to havebeen missed. The only ship in the data-base lost in the vicinity of Bear Cove andfitting the characteristics of the wreckageis the I lumboldt . Thus, while this identifi-cation cannot be absolutely confirmed,there is little remaining doubt that it iscorrect.20

M ov e me n t of the wreckage froin th estranding siteT he Humboldts stranding site is, however,well known and lies one nautical milesouth o f Bear Cove. Th at is a considerabledistance for large, coherent pieces of shipstructure to move after a ship breaks up.The common belief has been that thewreckage floated to Bear Cove and sankthere. This is unlikely since, except whentrapped air escapes, buoyant wreckagecannot lose its buoyancy quickly. If thesehull sections could ever have floated, withall their weight of bolts and stfaps, i t isimprobable that they would sihk beforegoing ashore. For two sections to havesunk in the same place is extremelyunlikely.We suggest instead that this wreckagemoved north by a process analogous togeological bed-load transport, but on agigantic scale. The general water circu-lation in Halifax Harbour is that typical ofestuaries, flowing in (north) on the bottomand out on the surface. In the approachesto the Ha rbo ur, the surface flow is supple-mented by the coastal current which, nearBear Cove, is deflected southwards by theland. Together, these form a net southerlysurface flow of about 0.05 m/s (Lawrence,1989; Fad er & Petrie, 1991). Th e net n ear-bo ttom flow is less certain. A t the depth ofthe wreckage, it is probably slower than0.05 m/s and may be directed northward.Superimposed upon these net flows aresemi-diurnal tidal streams and short-period wind-induced movements. A milenorth of Bear Cove, the streams have beenmeasured at up to 0 ,05 m/ s (Fader &Petrie, 1991). The average winter wind-fields should not induce surface watervelocities of more than 0.1 m / s in this area(Law rence, 1989), except when the shore-line channels south-easterly winds north-wards. Hence, even combining mean windand maximum tide effects with the netflow, short-period northward watervelocities between Portuguese and Bear


15/17

T KENCHINGTON & C WHIT LLOCK US STEAMER H U M B O L D T

coves should not exceed 0.5m/s at thesurface. Near-bottom flows should beslower and too weak to move large piecesof wreckage.There appear, however, to be strongerephemeral north-going bottom flows in thisarea. Seaward of the wreck-site, at around30m depth, the sandy bottom is formedinto three-dimensional megaripples--indicative of turbulent northward bottomcurrents. There are also unusual gravelcircles, possibly formed by spinning vor-tices created during intense storms. Some-what shallower. at about the depth ofthe site, there are large ripples in graveldeposits formed by oscillatory wave action(Fader & Petrie, 1991; Fader et ul., 1994).While i t cannot be certain, it seems likelythat the strong flows that created thesefeatures are caused by occasional hurricaneevents. The shape of the land protects thisarea from south-westerly and southerlywinds but, when a hurricane blows fromthe south-east. i t sends heavy waves ontothis shore. Such a storm would interactwith the land to drive a short-lived north-ward flow up the channel towards Halifax.The authors therefore, suggest that thewreckage currently at Bear Cove movedduring a hurricane event, probably notlong after the Humboldt broke up. With itswood still retaining some buoyancy, thewreck structure would have rested lightlyon the bottom. Even at its 15 to 20mdepth, the extreme wave action, evidentlysufficient to form ripples in gravel, wouldhave moved the wreckage to and fro--thusovercoming friction. A slow wind-inducednorthward drift, such as the 0.5 m/s flowthat rippled sand at 30 m. would then besufficient to displace the wreck sectionsand carry them from Portuguese to BearCoves in about an hourPerhaps only very robust wooden shipstructure, with iron strapping and close-setplanking bolts, could have held togetherwhile being moved like this. Nevertheless,unless this wreckage is wrongly identified,

its present location demonstrates that sub-stantial pieces of coherent ship structurecan be found as much as a mile from thesite at which a wreck broke up.Structure of the hullMost structural features seen on this wreck-age are among those that might have beenexpected on a very large, American-builtsteamer of c. 1850. These include thewood species used, the heavy scantlings, thediagonal iron strapping, the copper plank-ing bolts, the edge-bolted ceiling, the use ofwooden hanging knees under (rather thanbeside) the deck beams and the arrangementof the coppering. The absence of housingscut into the deadwood to receive the heels ofthe aftermost frames, while it conflicts withsome recent paper reconstructions, mightalso have been expected.Deadwood coaks are rarely mentionedin contemporary accounts but their use isnot surprising in such a ship. The reduc-tion in ceiling thickness below the berthdeck and its apparent absence betweenthat deck and the main deck are moreremarkable.The wreckages strangest feature, how-ever, is the form of the after deadwood.Not only is this made of pieces arrangedsub-vertically but it has no provision forthe keelson either to scarf onto its foremostend or to sweep up its forward face, thealternatives usually presented in textbooks.The only obvious explanation for theobservations reported here is that thesurviving deadwood was placed on top ofthe keelson, or a timber that extended thelatter, which itself ran aft to the sternpostwith filling pieces between it and the keel(Fig. 9)[61. Although rarely noted inpublished accounts, such an arrange-ment would not be unique. The Jhelum(428 tons; Liverpool-built in 1849), forexample, had a keelson extending to hersternpost, with below, filling pieces,which Stammers and Kearon (I9921termed a deadwood and a simple

22 1


16/17

N , 4 C I I CAL ARCHAEOLOGY, 25.3 & 4

Figure Y. Diagrammatic reconstruction of thecentreline structure of the Humboldt near the afterdeadwood, showing the hypothetical keelsonextension and filling pieces abaft the last squareframe. The floors of square frames are shown insection and shaded.

standard knee linking keelson to sternpost.Considering their different sizes, it is notsurprising that the Humboldt had a multi-part deadwood in place of the Jhelumsknee. Other similar examples are knownamong the mid to late 19th-century wrecksin the Great Lakes, though their detailshave yet to be published (Peter Engelbert,pers. comm.).

Taken together, these various featuresshow that there is something to be learntfroin surviving ship structure, even when

the site is relatively modern and thematerial badly eroded. If such wreckage isexamined closely and subjected to the com-parative analysis routinely applied to otherartefacts, it can supply a precise date andsome indication of the geographic origin ofa ship, while also adding to knowledge ofthe evolution of nautical technology. Suchanalysis will become progressively moreinformative as detailed descriptions ofadditional wrecks are published.

AcknowledgementsThis project was designed to obviatethe need for external funding or support.We do, however, thank our families fortolerating the time spent at Bear Cove, aswe do the several divers who have assistedthere. We owe special debts to PeterEngelbert, Eric Lawson and DavidRoberts for information o n 19th-centuryship construction, to The Parker Gallery(London) for the photograph reproducedas Figure 1 and to Gordon Fader (BedfordInstitute of Oceanography) for advice onthe sedimentology of Halifax Harbour.

NotesAll tonnages are old measure.In this paper, original scantlings are quoted in imperial units with metric equivalents. All otherdimensions are in metric units.The only official account of the Humboldts loss is the vessels Protest, signed before the U S Consulo n 16 December 1853 (US National Archives RG84, Vol. 145, Class 22, pp. 166-173). Thisdocument exists in two somewhat contradictory versions, the earlier unsigned and struck out. Theaccount presented here of the ships final voyage relies on the signed version, although supplemen-tary details have been added: from the unsigned one; from several newspapers, The Novascotian (12 ,19 and 26 December 1853,4 January 1854), the Halifax Daily Sun (7, 13, 14 and 20 December 1853),the Acudiun Reporter (10 and 31 December 1853), and the Mew York Times (7, 8, 9, 10 December1853); and from the log of the revenue schooner During (Public Archives of Nova Scotia file RGI.Vol. 427). All dates follow the civil calendar.Public Archives of Nova Scotia file MGlOO, Vol. 225, 12b.Other such survey reports doubtless exist but no comprehensive study of them is available.I t is, of course, possible that the keelson-extension and filling pieces were regarded by theshipwrights as parts of a much larger deadwood.

ReferencesCampbell, G., 1974, China Tea Clippers. London.Chapelle, H . I.. 1949, The History of the American Sailing Navy: the Ships and their Development.Chapelle, H. I. , 1967, Th e Search fo r Speed Under Sail 1700--1855.New York.22 2

New York.


17/17

T. K E N C H I N G T O N & C WHITELOCK: S STEAMER HU M BO L D T

Corlett, E., 1975, The Iron Ship: the History und Significance qf Brunels Great Britain. Bristol.Fader, G. B. J . , Miller. R . 0. & Pecore, S. S ., 1994, Sample control, anchor marks, anthropogenicfeatures and lacustrine sediments of Halifax Habour. Geologicul Surve y of Cunoda Open File Report2958.Fader, G. B. J. & Petrie, B.. 1991. Halifax Harbour: how the currents affect sediment distribution. In:Science Review, 1988 trnd 1989. Bedford Institute of Oceanography. Dartmouth.Farr, G. , 1963, The Steamship G r m t Western: the first Atlantic liner. Locul History Pumphlets, BristolBrunch of [he Historicd Associution, 8.Greenhill, B.. 1993. Steam before the screw. In: Gardiner, R. 8c Greenhill, B. (Eds), The Advent ofSteam: the Merchant Steumship hejbre 1900. London.Griffiths. D., 1993, Marine engineering development in the nineteenth century. In: Gardiner, R. &Greenhill, B. (Eds) . The Advent of Steam : T he Merchant Steumship before lWj. London.Knowles, J. . 1822. The Elements und Practice of Nuvul Architecture, etc. Third edition. London.Lawrence, D. J.. 1989. Physical oceanography and modelling in Halifax Harbour: a review.In: Nicholls. H . B. (Ed.). Investigations of marine environmental quality in Halifax Harbour.Can.Tech. Rep.Fish. yi ta t . Sci. 1693.Lloyds Register. 1858, L l o ~ t l sRcgister of British und Foreign Skipping from I s / July, 1858, to t lw30th June, I K W . London.MacGregor, I).R., 1984. Mm+umt Sailing Ships 1815--1850:Suprcwacy of Sail. London.Ridgely-Nevitt, C., 1981, American Ste am shi p on the Atlantic. Newark, New Jersey.Roff, W. J . , 1993, Early steamships in eastern waters. In : Gardiner, R. & Greenhill, B. (Eds), The Advent

Stammers. M. K . & Kearon. J . , 1992, The Jhelum: A Victoriun Merchunt Ship. Stroud.of Steam: The Merchunt Stramship before 1900. London.

223

iron strapping

Documents