SEARCH

SEARCH BY CITATION

Keywords:

  • Mississippi River;
  • inland navigation;
  • steamboat construction;
  • North American river boats

Abstract

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

This paper describes the design and construction of the side-wheel steamer Heroine, a representative of the ‘western river steamboat’ type and the earliest example of its kind to undergo archaeological study. Heroine was built at New Albany, Indiana, in 1832 and sunk on the Red River between Oklahoma and Texas in 1838. The extensive remains of the lower hull show assembly practices in use during the developmental era of Mississippi River steamboats. The wreck also reveals a heretofore-unknown technique for longitudinally strengthening these long, narrow, and very lightly built hulls.

Steam propulsion was introduced to North America's Mississippi River and its many tributaries in 1811 with the building of the side-wheel boat New Orleans at Pittsburgh, Pennsylvania. The shallow, fast-flowing, hazard-filled channels of these rivers presented special challenges for navigation, but the quarter-century that followed New Orleans' first passage saw the creation of a distinctive vessel type known as the western river steamboat. It was characterized by a lightly built, shallow-draft hull outfitted with compact and powerful high-pressure steam engines. Steam propulsion proved vital to the economic and territorial expansion of the United States in the 19th century, but the specifics of vessel design and construction during the first quarter-century of steam power in the Mississippi River watershed are largely unknown. The earliest-known hull plans for a western river steamer, the 393-ton Buckeye State of 1850, show a vessel built long after the developmental period was over (Kane, 2004: 85). The problem was summarized by steamboat historian Louis Hunter:

Unlike the sailing and steam marine, western steamboats gave rise to almost no technical literature. They left behind no collections of hull models, very few prints, and even fewer scale drawings, while their structural evolution was virtually completed when the photographic record began. (1969: 66)

A recent archaeological discovery is expanding our understanding of the formative years. The 160-ton Heroine, sunk in the Red River near Fort Towson in south-eastern Oklahoma, dates to the 1830s and is the earliest wreck of a western river steamboat to be systematically excavated and recorded. Between 2001 and 2008 the Oklahoma Historical Society, the Institute of Nautical Archaeology, and Texas A&M University jointly investigated the site. The excavations, major finds, and history are summarized in IJNA 42.2 (Crisman et al., 2013).

History and archaeological study

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

A medium-sized, general-purpose carrier of people and cargo, the side-wheel Heroine began service in November 1832 and over the next five-and-a-half years annually navigated thousands of miles on the Mississippi, and lower Ohio and Missouri Rivers. Heroine was ascending the Red River near Fort Towson, Oklahoma on 6 May 1838 when the hull was snagged by a submerged log (Crisman et al., 2013: 366, fig. 1). Limited salvage of the cargo and machinery occurred after the sinking, but the hull soon filled with sediment, ending recovery operations. Portions of the upperworks remained visible for five years, until a flood shifted the river channel to the south and capped the wreck with 7.62 m (25 ft) of sand. Between 1843 and 1990 Heroine was preserved in an ideal environment which was consistently wet and protected from air, light, and human activity. Another flood in 1990 shifted the river northward, exposing the wreck for the first time in 147 years.

When study of the wreck began in 2001, about 15.2 m (50 ft) of the stern was partially exposed above the river bottom, from the sternpost forward to the side wheels and the engine-and-machinery supporting cylinder timbers. Test excavations in 2002 showed that the wreck was c.42.7 m (140 ft) in length overall, and hogged at some point after the sinking, with the bow and stern sloping down from midships. The ends of the hull twisted to port: the angle varied, from 6° at the bow, near 0° amidships, and 25° at the stern.

The condition of the hull also varied from stem to stern (Fig. 1). With the exception of the cylinder and bearing timbers and portions of the port-side railing, nothing survived above the level of the main deck. The wreck lies at a slight angle across the river, with the starboard surfaces bearing the brunt of the current and the floating logs it carries. Most of the overhanging deck structure—the guard—was stripped from that side. The orientation and list of the wreck protected the port side from the worst effects of erosion, impacts, and collapse. Heavier construction and deeper burial beneath the river bottom combined to preserve the first 9.14 m (30 ft) of the bow and last 6.1 m (20 ft) of the stern in relatively good shape; both had intact structure up to and including the main deck beams and planking. About 12.19 m (40 ft), of the midships structure was held together by the heavy cylinder-timber assembly at the level of the main deck, although collapse of the sides was evident and some starboard frames and planking were missing. Between the bow and cylinder timbers the hull was in poor shape, with the starboard side and main deck gone above the turn of the bilge, and the port side collapsed outward. The section of hull between the cylinder timbers and stern assembly was missing most of the deck. Hull timbers protected from river-current erosion and periodic exposure were generally in excellent condition, often with original tool marks and coats of paint and tar. The twisting and hogging of the hull left many timbers, particularly on the starboard side, cracked, broken, or partially separated from adjacent structure.

figure

Figure 1. Plan view of wreck. (Drawing K. Crisman)

Download figure to PowerPoint

Diving conditions in the Red River—very low visibility, strong currents, and numerous obstructions on and around the wreck-site—encouraged simplicity in the tools and techniques used to document the hull's construction. Most of the recording was carried out by divers equipped with folding plastic rulers and open-reel measuring tapes. Because we were reasonably certain that the builders of Heroine worked in imperial measures, we used these units for recording (to the nearest quarter-inch). Frame sections, as well as degrees of listing, hogging, and sagging, were measured with electronic levels (goniometers) installed in watertight housings (Cozzi, 1998).

The high costs of conservation and display precluded recovery of the hull, and so every effort was made to record the dimensions of its constituent timbers, techniques of assembly, three-dimensional form, and evidence of its career. The goals of the hull documentation were to create a record of the vessel as found, to generate descriptions and analyses of its design and assembly, and to graphically reconstruct Heroine in the form of lines drawings, construction plans, and scale models. Of particular interest to our study was how Heroine's builders addressed the problem of longitudinal strength. Government enrolment records indicate that, early in their development, western steamboats greatly increased their length-to-breadth proportions and, at the same time, decreased their depth of hold (Works Progress Administration, 1942). The resulting hull form, while well-suited to shallow river navigation, was prone to drooping at its ends (hogging) and sagging in the middle. By the early 1840s fore-and-aft trusses consisting of support posts and wrought-iron rods tightened with turnbuckles were introduced. Known as hog chains (even though there were no actual chains), the trusses became a signature feature of western steamboats built after that time (Hunter, 1969: 94–100; Custer, 1991). Historical records provide few clues about the longitudinal reinforcement of river steamers in the pre-hog-chain era, and Heroine offered an excellent opportunity to examine earlier approaches to the hogging problem.

Construction and materials

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

We know that Heroine was built at New Albany, Indiana, in 1832 (Hall, 1836: 252, 256), but the name of the builder is not yet known. Several individuals constructed steamboats in the Louisville-New Albany area at the time, including Peter Tellon, William Crane, and Messrs Parkinson, French, Cunningham, and Ferguson, and it is likely that one of them assembled the hull (US Treasury Department, 1838: 318–9; Jackson, 1985: 159–60). The hull may have been a speculative venture by the yard, but more likely was built after the original owner, Jeremiah Diller, and the shipwright signed a contract specifying the dimensions, tonnage, materials, special features, timetable for completion, price, and payment schedule (Jackson, 1985: 159–66). The dearth of plans or half models for river steamers of this era leaves the design methods open to question. Half models were widely used in contemporary American coastal yards, and this may have been the preferred means for designing western steamboat hulls as well (Chapelle, 1960: 6–12; Hunter, 1969: 88–9).

There was a clear favourite in the selection of timber for Heroine's assembly: American white oak (Quercus alba). With a few exceptions—pine (genus Pinus) for the deck planking and the tongue-and-groove sheathing on the wheelhouses, and black locust (Robinia pseudoacacia) for the timberheads at the bow and stern—all of the timbers sampled were of white oak.1 This dense, durable hardwood was commonly used by North American shipbuilders, and was widely available in the Ohio River Valley.

The quality of the timbers we examined appeared exceptional, with few knots, blemishes, or rounded log surfaces (waney edges) evident. Frame timbers and deck beams were typically 178 mm (7 in) moulded and 127 mm (5 in) sided, suggesting that they were cut to these dimensions by mills that supplied the yard. Very little naturally curved compass timber was used, and nearly all such pieces were incorporated in the stem and stern assemblies, or in frames at the very ends of the hull. Elsewhere, futtocks at the turn of the bilge were cut to the curve from straight timbers. Lodging, hanging, or dagger knees were not used in the deck assembly.

Heroine was exclusively iron fastened: through-bolts at the stem and stern assemblies, and fastening the keelson to the floors and keel; bolts headed over clench rings at the sister- and bilge-keelsons, clamps, and plank butts; large chisel-pointed square spikes in the frame assemblies; smaller chisel-pointed square spikes for the exterior and deck planking; and square nails for the ceiling, the tongue-and-groove sheathing of the upperworks, and for other interior finishing work. The builders were sparing in their use of fasteners at certain locations: the square-frame floors and futtocks were fastened to adjacent timbers with only two spikes.

The following description of Heroine's construction loosely follows the assembly sequence used by the shipwrights who built the steamboat.

Keel

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

Heroine's upright orientation and largely intact lower hull limited access to the keel, particularly at the bow where the location of the forward end of the keel could only be estimated. However, the top surface (sided dimension) of the keel was measured between all of the frames, and its aftermost end was also uncovered and measured. A split in the after-hull along the starboard side of the keel permitted examination of the top, bottom, and side of the keel.

The keel measured 305 mm (12 in) sided and 102 mm (4 in) moulded, and resembled a thick plank; there were no rabbets or chamfers carved into its sides for the garboards. Its overall length was estimated to be 38.25 m (125 ft, 6 in), and it undoubtedly consists of at least two, and likely as many as three or four timbers scarfed end-to-end. No scarf seams were observed between the frame floors, and so scarf dimensions and types are still unknown.

Heroine's keel provides us with a clear-cut instance of river navigation requirements outweighing other functional considerations. On an oceangoing wooden ship, particularly a sailing ship, a deep keel provided longitudinal strength and reduced a vessel's tendency to slide to leeward under the pressure of wind on its sails. However, shallow draft was everything for a river steamer, and on this hull the keel protruded no more than 38 mm (1½ in) below the bottom planking. Its contribution to the longitudinal strength of the hull was minimal, and compensation for this deficiency was provided elsewhere in the structure.

Stem assembly

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The bow was in better condition than the rest of the hull due to its heavy construction and deep burial beneath the river bottom (Fig. 2). The interior of the bow and the uppermost 1.83 m (6 ft) of the stem and its cutwater were accessible to recorders, but the lower end of the stem and forward end of the keel were not, and so the dimensions and assembly of the latter elements are conjectural.

figure

Figure 2. Interior profile of bow. (Drawing K. Crisman)

Download figure to PowerPoint

The curve of the apron and the upper stem together indicated that the stem raked forward with a long, easy curve (Table 1). No evidence of a scarf was seen at the upper end of the stem, and it was not clear if this member was fashioned from one timber or two. The heel of the stem was likely notched to lap over the forward end of the keel to protect it when the vessel grounded—a frequent occurrence on river steamers. The after corners of the stem were deeply chamfered to fit the ends of the planking. A two-piece cutwater assembly was bolted to the forward face of the stem head and braced with a pair of horizontal knees—one on each side of the stem; the top of the stem and cutwater supported a short projection of the forward deck that was part of the guard structure. A ‘V’ carved into the starboard side of the stem at the base of the cutwater served as a 1.52 m (5 ft) draft mark, which likely indicated the hull's depth when fully laden; the port side of the stem did not have any draft marks.

Table 1. Stem timber dimensions
 LengthMouldedSided
MetricImperialNoteMetricImperialNoteMetricImperialNote
Stem5.26 m17 ft 3 inestimated229 mm9 inat upper end241 mm9½ inmax.; at head
178 mm7 inon forward face
127 mm5 inon after face
Apron5.18 m17 ft102 mm4 inat heel
229 mm9 inat forwardmost floor305 mm12 inat forwardmost floor
127 mm5 inat head299 mm11¾ inat head
Lower breasthook2.23 m7 ft 4 inbetween arms355 mm1 ft 2 inat the apron203 mm8 in
Upper breasthook2.51 m8 ft 3 inbetween arms432 mm1 ft 5 inat the apron203 mm8 in

The stem was backed by an apron which extended from the forward end of the keel to the underside of the main deck. The sides and top of the apron were obscured in many places by adjacent timbers, but no evidence of a scarf was seen or felt, suggesting that the apron was shaped from a single timber. The forward corners of the apron were chamfered to create planking rabbets with the after corners of the stem.

During a later stage in construction the apron served as the attachment point for many other bow timbers. The doubled floors of the five forward-most square frames fit into notches in the lower (after) end of the apron, the heels of the port and starboard cant frames butted the sides of its upper (forward) end, and the keelson lapped the lower half of the apron and was bolted to the keel and stem. After the cant frames were fitted, two breasthooks of substantial dimensions were bolted to the after face of the apron's upper end and to the adjacent cant frames. Finally, the forward ends of the stringers, bilge-keelsons, and lower clamps terminated at the sides or after face of the apron.

Stern assembly

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The well-preserved stern was more elaborate in design and more solidly built than might be expected in a shallow-draft river vessel (Figs 3 and 4). Stern elements fastened atop the keel included a substantial deadwood timber and stern knee, inner and main posts, and three transom timbers (Table 2). The deadwood was a relatively complex piece, with a sloping top surface on its forward half, a flat top on the middle where it fit the stern knee, and a deeply cut-down after end where it seated the two sternposts. The sternposts were straight lengths of white oak that tapered from bottom to top and angled aft with a rake of 85°.

figure

Figure 3. Interior profile of stern. (Drawing K. Crisman)

Download figure to PowerPoint

figure

Figure 4. Model of Heroine's stern. (Model G. Grieco, photo W. Smith)

Download figure to PowerPoint

Table 2. Stern timber dimensions
 LengthMouldedSided
MetricImperialNoteMetricImperialNoteMetricImperialNote
Deadwood timber4.55 m14 ft 11 in381 mm1 ft 3 in254 mm10 in
Stern knee1.78 m5 ft 10 inlower arm432 mm1 ft 5 inin throat254 mm10 inat top
0.88 m2 ft 10½ inupper arm
Inner sternpost1.96 m6 ft 5 in241 mm9½ inat heel254 mm10 inon forward face
152 mm6 inat head
Main sternpost2.16 m7 ft 1 in343 mm1 ft 1½ inat heel254 mm10 in
178 mm7 inat head
Lower transom0.79 m2 ft 7 inacross forward face216 mm8½ in267 mm10½ inat inner sternpost
Middle transom2.28 m7 ft 6 in216 mm8½ in216 mm8½ inat inner sternpost
Upper transom4.78 m15 ft 8 in229 mm9 in229 mm9 inat sternpost
Standing knees0.64 m2 ft 1 inlower arm349 mm13¾ inheight on standing arm111 mm4⅜ in

Three transom timbers were fitted inside the stern: the lowest was fastened atop the stern knee, the middle to the forward face of the inner post, and the uppermost transom was seated atop the inner post. All tapered toward their outboard ends and all had moulded after faces to better seat the exterior planks as they drew inward and upward to meet the sternpost and underside of the counter. The lowest of the three transoms resembled a triangular wedge with its after corner notched to fit around the inner sternpost. The fashion piece, a two-piece, V-shaped frame that defined the shape of the stern, was spiked to the forward face of this transom. The middle transom timber was notched on its after face to fit around the inner sternpost and fastened to the post by a single bolt.

The upper transom had a shallow notch on its after face to fit the forward face of the main sternpost, and was secured in place by two bolts. The lower corner of its after face was deeply bevelled to seat the ends of the exterior planking. This piece had a number of interesting features. The outboard ends of the transom each supported a standing knee that served as an attachment point for the uppermost side strakes. The top of the transom also had a 25 mm (1 in) deep, fan-shaped notch which was flush with the top of the sternpost; this allowed the tiller to swing freely from side to side without scraping the transom. Finally, the transom's after face had three shallow notches on either side of the sternpost to fit the bases of six aft-angling counter timbers. The finely crafted stern, an attractive feature of the hull, would have been largely invisible under the overhang of the main deck.

Frames

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

Heroine was built with a total of 74 square frames (the 12 forward-most with double floor timbers, and the rest with a single floor), 12 cant frames on either side of the bow, and four pairs of half-frames in the stern (Table 3). The midship frame was not identified, nor has the frame-designation system used by the shipwrights when building the hull been determined. For the purposes of frame description, a division was made at the juncture of the double-floored and single-floored frames near the bow. Thus, the aftermost double-floored frame was assigned the letter A and the forward-most cant frame in the bow was X. The forward-most single-floored frame was numbered 1 and the aftermost half-frame in the stern 66.

Table 3. Frame dimensions
 Moulded averageSided averageOther
MetricImperialNoteMetricImperialMetricImperial
Floor timbers203 mm8 inat keel102 mm4 in
152 mm6 inat head
Limber holes51 x 25 mm2 x 1 in
First futtocks152 mm6 inat heel
102–127 mm4–5 inat head102 mm4 in
Second futtocks152 mm6 inat heel102 mm4 in
102 mm4 inat head

The square frames were on 508 mm (20 in) centres atop the keel—this spacing varied between some frames, but only slightly. Over most of the length of the hull the square frames were consistent in their dimensions and both lightweight and uncomplicated in their assembly (Figs 5 and 6). Frames 1 through to 62 each consisted of five pieces: a single floor timber, two first futtocks at the turn of the bilge, and two second futtocks on the sides. The frames were held together with a minimal number of fastenings: on each side of the hull, four spikes were driven across the longitudinal axis of each frame, two from the forward face of the frame (one through the head of the floor, and the other through the heel of the second futtock) and two from aft (one at the heel, and the other at the head of the first futtock). Thus, only eight spikes were required for the assembly of each frame.

figure

Figure 5. Section at frame 23. (Drawing K. Crisman)

Download figure to PowerPoint

figure

Figure 6. Model of Heroine's hull showing framing and deck assembly. (Model G. Grieco, photograph W. Smith)

Download figure to PowerPoint

All of the floors between the apron at the bow and the deadwood at the stern were straight lengths of timber with a single limber hole centred over the keel. A single 19 mm (¾ in) round, iron through-bolt fastened each floor to the keel, with the bolts offset in a staggered pattern on succeeding frames to avoid the limber holes. The first futtocks were short lengths of sharply curved timber. The wood grain (and splitting) pattern seen on these futtocks suggests that most were cut to the curve from straight-grained timbers rather than fashioned from naturally curved compass timber. This was clearly one instance where the builders took a short cut to avoid the extra expense and effort of finding compass timber to match the curve of the frames. It did constitute a weak point: the sides of the hull between the bow and the cylinder timbers, it will be recalled, were broken off along this hard curve. This framing style with its sharply curved transition from bottom to side is known as ‘rounded knuckle’ construction (Kane, 2004: 103–4).

Not all of the square frames in the hull were built alike. As noted above, the 12 forward-most frames (A-L) each included a second floor timber that extended up to the heels of the first futtocks, an addition clearly intended to reinforce the bow and enable it to withstand repeated groundings and collisions. The first 43 single-floored square frames (1–43) were assembled with their first futtocks forward of the floor timber. The aftermost 19 square frames (44–62) all had their first futtocks attached abaft the floors.

At either end of the hull the frames changed in form and assembly from those in the middle of the vessel; both the deadrise and moulded dimensions of floors increased, while the length of their arms decreased, and the first futtocks expanded in length. These frames appeared to contain more compass timber with grain that matched the curves of the hull. At the bow the floors of frames H-L were notched down over the top of the apron. Forward of frame L, a dozen cant frames were fitted in a radial pattern on each side of the apron, with additional filler pieces between their heads. The cant frames and hawse pieces at the head of the apron were so closely laid as to form a near-solid wall of timber for 1.22 m (4 ft) on either side of the stem. At the stern the single floors of the four aftermost square frames (59–62) were notched down over the deadwood, and had pronounced curves that formed the hollow of the run. The four aftermost frames were composed of three half-frames (63–65) with their first futtocks notched into the side of the deadwood and stern knee, and the aft-canting final frame (66) with futtock heels that butted one another over the centreline of the stern knee. As noted earlier, the fashion piece that formed the corner of the sides and stern was bolted to the forward face of the lowest transom timber.

Keelson

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The keelson was composed of three long white oak pieces that paralleled the keel, starting forward at a point midway up the apron and extending aft to frame 65 and the upper arm of the stern knee (Table 4). The three keelson timbers were secured to one another by long flat scarfs, the first located at the forward end of the cylinder timbers and the second at their after end. The direction of the scarfs indicated that the keelson pieces were laid down from the bow aft. The need for a shallow draft limited the depth of the keel, but the same restriction did not apply to the keelson; it was two-and-a-half times greater in its moulded dimension than the keel, although it was also one-quarter narrower.

Table 4. Keelson dimensions
 LengthMoulded/widthSided/depth
MetricImperialMetricImperialMetricImperial
Keelson (overall)39.01 m128 ft267–279 mm10½–11 in229 mm9 in
Timber 1 (forward)16.59 m54 ft 5 in
Timber 2 (amidships)15.67 m51 ft 5 in
Timber 3 (aft)10.82 m35 ft 6 in
—notches over floor25 mm1 in
Flat scarfs (2)2.03 m6 ft 8 in
Notches for stanchions (average)203–305 mm8–12 in76 mm3 in25 mm1 in (at after end)

The keelson was notched down over the floors, and fastened to the frames and keel by 19 mm (¾ in) diameter iron bolts—one through each floor—offset from its centreline to avoid limber holes and the bolts that fastened the floors in place. The keelson had a series of shallow, sloping notches cut into its top surface to fit deck-beam-supporting stanchions. The majority of these notches (11) were cut into the forward-most keelson timber between the forward hatch and the cylinder timbers; their spacing matches the deck-beam notches on the surviving port-side clamp, indicating that every beam here was reinforced by a stanchion. The reason for the extra posts was obvious: this portion of the main deck supported the weight of the boiler assembly with its many cast- and wrought-iron elements, the brick-lined firebox, and the water within the boilers; a total of c.13.9 metric tonnes, mounted on the relatively flimsy deck structure.2 Abaft the cylinder timbers, the aftermost keelson timber had three notches carved into its top for deck-beam stanchions.

The keelson was greater than the keel in its moulded dimension, but when the overall length of the hull is considered it is still not very large. A long and relatively narrow wooden hull is inherently prone to hogging, and a larger keelson, or multiple keelsons, would surely have strengthened the boat's spine. This modestly sized keelson may be another example of the builder keeping timbers as light as possible to minimize vessel draft. Notching the underside of the keelson to lock it over the floors was a labour-intensive practice, but its contribution to the hull's longitudinal stiffness was evidently considered worth the added effort.

Stringers, bilge-keelsons, and clamps

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The keel and keelson stiffened the hull along its centreline, but additional longitudinal strengthening along the bottom and sides was provided by stringers (also called ‘streaks’), bilge-keelsons, and clamps (Table 5; Kane 2004: 97, 106–7). Three longitudinal members were fitted inside the turn of the bilge on each side of the hull where the bottom curved up to form the sides. The sharp turn of the bilge created a focal point for stresses in the hull; this is also the place where the floor, first futtocks, and second futtocks of each frame met. Each of the longitudinal timbers was placed to reinforce a specific part of every frame. The innermost of the three, the hold stringer, was a thick plank laid over the heels of the first futtocks and secured with two spikes per frame, one each in the floor and first futtock. The forward ends of the port and starboard hold stringers butted over the apron, and each extended back in a long arc to a terminus on frame 61, the penultimate square frame in the hull. Each stringer was made up of several timbers flat-scarfed end-to-end. (Similar flat scarfs were noted on the bilge-keelsons, side stringers, and clamps.)

Table 5. Stringer, bilge-keelson, and clamp dimensions
 Moulded/widthSided/depthNotched over frames
MetricImperialMetricImperialMetricImperial
Hold stringers254 mm10 in51 mm2 in
Side stringers254 mm10 in51 mm2 in
Bilge-keelsons127 mm5 in203 mm8 in38 mm1½ in
Lower clamps330 mm1 ft 1 in70–76 mm2¾–3 in19 mm¾ in
Upper clamps330 mm1 ft 1 in70–76 mm2¾–3 in19 mm¾ in

Bilge-keelsons were fitted over the turn of the bilge on each side, centring on the butts where the floor timber heads met second futtock heels. The importance of the bilge-keelsons was evident in their substantial size and fitting: each was notched down 32 mm (1¼ in) over the doubled frame timbers, and fastened by one to two spikes and one 19 mm (¾ in) bolt per frame. The bolts were inserted into holes drilled through the centre of each first futtock and their heads were clenched over rings, while the spikes were driven into floor heads or second futtock heels. The bilge-keelsons extended aft from the apron, the notching of their undersides commenced immediately abaft frame L and ended at the forward edge of frame 61, and the after ends of the bilge-keelsons terminated over frame 65.

The side stringers, also known as ‘futtock head strakes’, were similar in many respects to the hold stringers (Bates, 1968: 30). They consisted of thick planks laid down over futtock ends—in this case the heads of the first futtocks—and secured to each frame by a pair of spikes, one each through the first and second futtocks. The forward ends of the side stringers were located on either side of the apron, and each terminated abaft frame 58.

Above the waterline, each side of the hull was longitudinally reinforced by a pair of clamps attached one above the other to the insides of the second futtocks. Both clamps were of the same dimensions, resembling broad, thick planks, and both were notched down over and fastened to the second futtock of every frame by one 19 mm (¾ in) bolt clenched over a 32 mm (1¼ in) diameter ring, and by spikes, one through the lower clamp and one, or more commonly two, through the upper clamp. The tapered forward ends of the port and starboard lower clamps butted over the apron between the upper and lower breasthooks, and their after ends terminated at frame 64. The forward ends of the upper clamps butted the outboard ends of the upper breasthook and their after ends terminated at the fashion piece. The upper clamps were notched along their top edges to fit the beams of the main deck. The frame tops were uneven in height, but on average extended 127–140 mm (5–5½ in) above the top of the upper clamp.

Two features of the clamps emphasized their vital role in strengthening the sides of the steamboat above the waterline. The first was the notching of the clamps to lock the frame tops tightly in place. This required much time and labour on the part of the builders to measure then chisel out the 102 mm (4 in) wide by 19 mm (¾ in) deep notches to fit each futtock. The second noteworthy feature was the use of clenched bolts to fasten each clamp to each frame. Spikes alone were not deemed sufficient for this purpose, and the security that the bolts provided was clearly considered worth the extra costs in iron and labour.

Planking

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

Samples of wood taken from around the hull suggest that all planks were of white oak. Planking on the starboard side of Heroine's hull was generally eroded and split, if not missing altogether; planking on the buried port side was typically in very good condition, but the still-tight seams were often hard to locate and obscured by overlying stringers and clamps. Planking thicknesses were obtained in places where hull fractures opened up the seams (Table 6).

Table 6. Planking dimensions
 WidthThickness
MetricImperialMetricImperial
Garboards279 mm11 in64 mm2½ in
Bottom strakes254–330 mm10–13 in51 mm2 in
Bilge and side strakes140–279 mm5½–11 in51 mm2 in
(average)_152–203 mm6–8 in

The innermost strakes, the garboards, were relatively wide and slightly thicker than the 51 mm (2 in) that was standard on all of the other bottom and side planking. Their fit against the side of the keel did not include a rabbet, specially chamfered edges, or other modifications. However, at the bow and stern rabbets were cut into the stem-apron and deadwood-sternpost assemblies to fit the ends of the garboards and other planking strakes. From the garboards to the turn of the bilge the strakes were relatively wide; there were five to seven strakes of this description on each side of the hull. From a point just inboard of the bilge keelson the strakes narrowed, although their thickness remained the same; nine or ten strakes covered each side of the hull. There were no wales (extra-thick planks) near the top of the hull to strengthen the sides above the waterline. The uppermost row of planks around each side, the sheer-strake, was flush with the top of the second futtocks, and, like the upper clamp, was notched to fit the deck beams.

A pair of large iron spikes fastened each plank to each frame. The narrow width of the frame timbers forced the shipwrights to drive the spikes side by side, rather than in a staggered pattern. The narrow frame widths and flexibility of the steamer's hull made ordinary butt joints between the planks a risky, leaky proposition, and the builders instead lapped them with short, vertical, straight scarfs. One such plank join was noted on the first plank below the sheer-strake on the starboard side of frame 52. Elsewhere on the hull, excavators noticed bolts clenched over rings on the inside surfaces of two frames with a single bolt-less frame between them; the bolts appeared to be reinforcing a plank butt located on the intermediate frame.

During the excavation of the rudder, a portion of the previously buried port transom was uncovered, revealing traces of a black surface-coating on the side and counter planks. This was identified as lead-based paint with a covering of pitch. A similar dark-coloured surface-coating was observed on the port side amidships when several of the uppermost planking strakes were uncovered at frame 23.

Hold, bow and stern compartments

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The interior of the hull mostly served as hold space for cargoes, since passenger and crew accommodations were on the main or upper decks, and the boilers, engine, and paddle-wheel machinery were all mounted on the main deck. The hold was a long (33.35 m (109 ft, 5 in)), narrow (c.5.63 m (18 ft, 6 in)) open space with only 1.6 m (5 ft, 3 in) of clearance between the tops of the floors and the undersides of the deck beams. The central portion of the hold was longitudinally divided by the 13.51-m-long (44 ft, 4 in) support structure for the engine and fly wheels. Bulkheads at either end of the hold created small bow and stern compartments.

Heroine had a snag chamber bulkhead, a watertight transverse partition, installed 4.57 m (15 ft) abaft the head of the stem (Table 7, Figs 2 and 7). Its purpose, as the name suggests, was to prevent flooding of the entire hull if the bow was punctured by a snag or crushed in a collision with another boat. Snags were the leading cause of accidental losses on the western rivers. Snag bulkheads were introduced as early as 1819 and by the late 1820s were fitted on most western steamboats (Marestier, 1957: 57, 60; Hunter, 1969: 80; Hall, 1974: 363–4). A watertight bow compartment seems a sensible precaution, but must not have been very effective, for they fell out of use by the 1840s (Kane, 2004: 90). Heroine's snag bulkhead is, thus far, the only example found or recorded on a western river steamboat wreck.

figure

Figure 7. Section at frame F. (Drawing K. Crisman)

Download figure to PowerPoint

Table 7. Snag chamber bulkhead dimensions
 Moulded/widthSided/depthHeight
MetricImperialMetricImperialNoteMetricImperial
Overallfrom top of frame G to underside of deck1.91 m6 ft 3 in
Support posts (11)89–114 mm3½–4½ in89–114 mm3½–4½ in
Forward horizontal planks (2)102 mm4 in25 mm1 in
Bulkhead planking (7 strakes)203–355 mm8–14 in51 mm2 in

The bulkhead was installed directly over frame G, the double-floored frame against which the heel of the apron butted. Its framework consisted of a line of vertical posts, with one atop the keelson, three on the starboard side, and five on the port side—the two additional port-side posts, located near the centreline, may have been part of a repair made during the boat's career. It was not clear if the heels of the posts were tenonned into the top of frame G, but their heads were clearly notched to fit the after corner of deck beam 6. One additional slanting post was fitted on each side of the bulkhead. The posts were reinforced with two thin horizontal planks on their forward faces, while the after sides were completely covered from the top of frame G to the after face of deck beam 6 with a wall of thick, wide, horizontally laid planks—this was the watertight wall designed to hold back a flood. A rectangular section of planks let in to the port side of the bulkhead appeared to be a repair.

Heroine's watertight bulkhead was useless on 6 May, 1838 when the steamboat ran on a snag that punched through the hull 12.19 m (40 ft) abaft the partition. It may have been of help on another occasion, however. During the excavation of the bow compartment evidence of an earlier collision was found on the starboard side between frames J and K, directly beneath the bilge keelson. A plank strake at this location had impact damage which the crew sealed by fitting a curved wooden plug, covering it with tar, and then driving two wooden wedges between the plug and the underside of the bilge keelson to fix the repair in place.

The small separate compartment in the bow was used by the crew for storing broken or heat-damaged iron castings (destined for recycling), and a few tools and rigging elements. The bilge debris that typically collects in spaces between frames was noticeably absent in the bow, suggesting that someone swept the compartment not long before the sinking.

The after compartment in the hold was created by the addition of a crude, non-watertight bulkhead of narrow, vertical slats nailed to the after side of frame 60 and the forward face of deck beam 44; the bulkhead planks were braced by a horizontal plank that extended across frames 60 and 61, 305 mm (12 in) above the keelson (Table 8, Fig. 3). The enclosed space, also known as the ‘run’, contained the tiller, wheel ropes, and the pair of large, iron-bound single blocks (bolted to the clamp abaft beam 45) that led the wheel ropes forward to the pilot house. When fitted to the rudder head, the tiller extended nearly the full length of the compartment and pivoted from side to side in a wide arc.

Table 8. Dimensions of stern compartment and bulkhead
 LengthWidthHeight/depth
MetricImperialNoteMetricImperialNoteMetricImperialNote
Interior dimensions3.05 m10 ftfrom transom timbers to bulkhead5.08 m16 ft 8 inbreadth between upper clamps at frame 621.98 m6 ft 6 inheight from top of frame 60 to upper side of deck
Bulkhead planking127 mm5 in51 mm2 in

Artefacts located in the run suggested that the crew used it for storage of tools, ship's stores, discarded equipment, and personal belongings (Crisman et al., 2013: 375). The spaces between frames 61 to 63 and the sides of the deadwood contained viscous, oily hydrocarbon, identified as pine tar (Bates, 2003); other debris in the run included wood shavings that resembled waste from a plane or draw knife, and corn cobs. While a contemporary account of the 342-ton western river steamer General Pratte noted that the 19 firemen aboard slept in the run (McDermott, 1968: 173), there was no evidence to suggest that the run of the much smaller Heroine was used as living quarters by crew members.

Main deck structure

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

Heroine's main deck suffered extensive damage and loss since the boat sank in 1838. As noted, the extension of the main deck beyond the hull, the guard, was entirely missing on the starboard side, and many of the deck beams over the forward and after holds were missing, or only survived at the port-side guard. Sections of deck with intact beams and planking were documented at the bow, midships, and stern; the intact beams and beam notches in the port-side upper clamp showed that Heroine had 48 transverse beams in its main deck (consecutively numbered, starting with beam 1 at the bow) (Table 9). Due to the considerable width of the main deck (up to 10.97 m (36 ft)) it was not always possible for the builders to find a single timber of the necessary length; in these cases they lapped two timbers side-by-side and spiked them together to form a single beam. In other places, notably at the bow and beneath the boiler assembly, it appears that beams 6, 8, 9, 10, 12, 17, and 19 were doubled all the way across the hull to strengthen the deck or hatch openings.

Table 9. Deck-beam, guard-strake, waterways, and deck planking dimensions
 Moulded/widthSided/depth
MetricImperialNoteMetricImperial
Deck-beams127–178 mm5–7 in102–152 mm4–6 in
Paddle- and shaft-beams (25, 26, 29, 32, 33)172–178 mm6¾–7 in172–178 mm6¾–7 in
Nosing strakes203 mm8 in51 mm2 in
Half-round moulding102 mm4 in
Guard-cap254 mm10 in44 mm1¾ in
Sheer-cap254 mm10 in51 mm2 in
Main-deck planking89–254 mm3½–10 in51 mm2 in
Guard-deck planking165–216 mm6½–8½ inat stern25 mm1 in
 254–457 mm10–18 inat bow25 mm1 in

Beams were laid down an average of 0.91 m (3 ft) centre-to-centre over the length of the hull, but this varied considerably, from as little as 0.66 m (2 ft 2 in) between beams 47 and 48 in the stern, to as much as 1.32 m (4 ft 4 in) between beams 19 and 20 at the forward end of the cylinder timbers. Clearly, the builders were not obsessed with precise placement of beams, but shifted them to accommodate features such as hatches and cylinder timbers. Each beam was fitted into a shallow notch in the upper clamp and sheer-strake and secured in place by two large spikes, one in the clamp and the other in the sheer-strake (Fig. 6). The junctures of the beams and clamps were not reinforced by any type of knees.

Moulded and sided dimensions of the beams varied within a 51 mm (2 in) range. The two beams immediately forward (26 and 27), and the two immediately aft (33 and 34) of the side-wheel openings in the guard, known as ‘paddle beams’, were slightly larger than average in section. Beams 27 and 33 had separate pieces at either end that lapped inboard of the hull, apparently to increase the length of the paddle-wheel openings by a few inches. Beam 30, directly beneath the main shafts of the paddle wheels, was also slightly larger in dimensions than most other deck beams.

Two pairs of short, diagonally slanting beams were fitted at the bow, with one pair on either side of the stem. The after ends of these beams butted the forward face of beam 1, and the innermost of each pair also lodged against the sides of bitt posts at the bow. These canted beams supported the curving guard structure that overhung the stem. Amidships, the central portions of beams 28–32 were cut away for the flywheels. (These were mounted atop the cylinder timbers and extended down into the hold.) At the after end of the deck, six short, fore-and-aft-oriented beams butted the after face of beam 48 and supported the extension of the deck over the counter and rudder head.

From stem to stern, the outboard ends of the beams were protected by a continuous run of vertical oak planks known as the nosing. These planks were spiked to the ends of the beams, and half-round moulding was spiked down the centre of the strake's outer face. Another run of heavy oaken strakes, the guard-cap, was spiked atop the beam ends and nosing, and also served to define the outboard edges of the main deck. The cap had square mortises cut through it at regular intervals to fit the main-deck rail stanchions.

Each side of the main deck had another run of oak planks, the sheer-cap, which covered the frame heads, clamps, and sheer-strakes along the top of the hull. This was fastened by one or two spikes driven into each deck beam. The strake closely resembled the guard-cap in its dimensions and appearance, and also had a series of square mortises cut through it, in this case to fit the posts, called stationaries, that supported the steamboat's superstructure (Bates, 1968: 23, 30). The individual planks composing the guard- and sheer-caps were flat-scarfed end-to-end, and at the narrowing of the bow both had two short lengths of plank that allowed the strakes to follow the curve. At the after end of the hull, the guard-cap curved around the after edge of the deck, while the sheer-cap terminated at the inside edge of the guard-cap.

Deck planks survived at the bow, the stern, and between the port-side cylinder timber and sheer-cap. All the planks between the sheer-caps ran parallel with the centreline of the hull and their ends were not nibbed into the caps. Individual planks within strakes were butt joined and fastened with two spikes per beam—squared at the butts and in a staggered pattern elsewhere. At the bow a total of 30 strakes were still attached to the beams inside the sheer-caps, with an estimated four to five missing on the port side. Plank widths varied considerably, from 102 to 254 mm (4 to 10 in). Planks at the bow were sawn from pine, with the exception of four wide centreline strakes forward of the hatch, which were of oak. The capstan (missing from the wreck) was located here, and the oak strakes provided a stronger surface for mounting this device. Also noted at the bow were three small plank repairs on the port side of the fore hatch. Amidships, a total of 13 deck strakes were preserved on the port side of the cylinder timbers between deck beams 27 to 33, and at the stern, 15 deck strakes were attached to the port side of beams 45–48.

Intact guard-planks were found at two port-side locations: between the stem and beam 6 in the bow and from the after end of the paddle-wheel opening to the stern (beams 34 to 48). Guard-planks were thinner and slightly wider than deck planks inboard of the sheer-cap; the guard-planks at the bow were exceptionally wide (up to 457 mm (18 in)). Wider planks tend to yield leakier seams, but this was not a problem on the guard-planks since they were outboard of the hull.

Hatches and companionways

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

Five deck openings were recorded during the study of the hull: a pair of companionways, a small scuttle, and the forward hatch at the bow, and a single companionway located on the port side of the stern (Table 10). Another opening in the deck, through the port-side guard at the stern, will be discussed below under the heading ‘Deck features’.

Table 10. Companionway and hatch dimensions
 NoteLengthWidthHeight
MetricImperialMetricImperialNoteMetricImperial
Bow companionwaysOD813 mm2 ft 8 in724 mm2 ft 4½ in
ID686 mm2 ft 3 in597 mm1 ft 11½ in
coamings64 mm2½ in51 mm2 in
Bow scuttle457 mm1 ft 6 in381 mm1 ft 3 in
Bow hatchOD1.93 m6 ft 4 in1.89 m6 ft 2½ in
ID1.78 m5 ft 10 in1.74 m5 ft 8½ in
side coamings76 mm3 inlower piece159 mm6¼ in
upper piece127 mm5 in
forward and after coamings (estimated)76 mm3 in76 mm3 in
Stern companionwayOD (est.)670 mm2 ft 3½ in711 mm2 ft 4 in
ID559 mm1 ft 10 in610 mm2 ft
side coamings51 mm2 in25 mm1 in
after coaming piece76 mm3 in 25 mm1 in

The companionways in the bow, located between beams 3 and 4, provided access to the forward compartment (Fig. 2). Each was nearly square and framed by a coaming that consisted of four pieces, half-lapped at the corners, and nailed down to the deck planking. The port companionway was missing one of its side pieces. Two squared notches were cut into inside edges of each side piece to fit a cover or grating. A wooden cleat nailed between frames K and L, between the bilge keelson and the side stringer, provided a step for crew members entering and exiting the compartment via the port companionway. It is not clear why Heroine had two openings here instead of one, but the feature did provide more light and ventilation into the compartment. This space was used as crew living quarters on some western steamboats, but there was no evidence to suggest this was true on Heroine (Ingraham, 1839: 196; Jackson, 1985: 161).

The small square scuttle on the starboard side of the bow was located between deck beams 6 and 7, immediately abaft the watertight bulkhead. The opening was too small for people or cargo, and was most likely used for stowing anchor cables or mooring lines in the hold. No evidence of coamings was noted around the opening, but it was probably covered when not in use.

The forward cargo hatch was located between deck beams 7 and 9, and was almost (but not quite) square. The sides consisted of fore-and-aft headers between beams 7 and 9, each of which was notched in the centre to support the inboard ends of beam 8; the side coamings were fastened directly atop the headers. The forward and after edges of the hatch were made up of single-piece coamings (missing when the bow was uncovered in 2005); these were half-lapped over the side coamings and spiked to the deck planks and beams 7 and 9. The hatch's side pieces were notched inside to fit a cover or grating, but curiously, the notches on the two sides of the opening did not correspond in size or location: the starboard coaming had five small notches spaced over its length, while the port-side timber had one long notch near the middle and two shorter notches further aft.

The fifth recorded opening in the main deck, a companionway that allowed entry into the stern compartment, was located between beams 46 and 47. The opening was offset to the port side of the deck, since the tiller would have blocked a centreline opening when the rudder was in its neutral position. The sharply curving half-frames below the companionway made for poor footing, a problem that someone addressed by nailing a crude plank step across frames 64 and 65. The coaming around the stern companionway was rudimentary, consisting of thin wooden slats nailed down over the planking. A large wrought-iron hasp fastened to the starboard side of the opening with a staple indicated that it once had a cover. This provision for locking the cover suggests that theft of the contents of the compartment, perhaps by the deck passengers who typically occupied the after half of the main deck, was considered a potential problem (Ingraham, 1839: 196; Hunter, 1969: 422–30). Plans of the 1850 side-wheel steamer Buckeye State show a similar companionway on the after deck (Tredgold, 1851: Pl. 1).

There was likely at least one more hatch through the main deck for loading cargoes into the after hold. If the pattern of doubling the intermediate beam seen on the forward hatch (beam 8) was repeated aft, then it is possible that the after hatch was located somewhere between beams 36 and 42, since beams 37, 39, and 41 were doubled. According to a local resident who explored the wreck soon after it was first uncovered in 1990, another companionway existed directly abaft the cylinder timbers (Ricky Martin, 2006: pers. com.). There may have been additional openings in the deck, but their size and location are currently unknown.

Cylinder and bearing timbers

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The builders of Heroine's hull and the mechanics who installed its propulsion system had to combine a lightly built, flexible wooden hull with a heavy and relatively complex arrangement of reciprocating and rotating cast-iron machinery. Their principal concerns were to keep everything properly aligned and supported. Two approaches were taken to achieve this aim. First, the machinery was designed with loose tolerances between moving parts and with bearings that could be adjusted to re-align shafts and couplings. Second, every effort was made to limit flexibility in the machinery's wooden support structure. The central part of the hull where the engine, crank, flywheels, and side wheels were mounted was made as rigid as possible by the installation of heavy timbers that were cross-braced and held in compression by wrought-iron reinforcing rods.

The principal support structure was an assembly of parallel oak timbers located on the centreline of the hull amidships (Table 11 and Fig. 8). Known as the cylinder timbers, these cradled the single-piston engine at their forward end, served as the alleyway for the crank arm (called the pitman) in the centre, and at their after extremity supported the shafts for the side wheels, each of which had, attached to its inboard end, a one-ton flywheel composed of a cast-iron flange and a rim supported by timber arms (Hodge, 1840: pl. XXXI). This was a lot to ask of any wooden structure, and the shipwrights went to considerable effort to ensure that what they built was up to the job. They began by laying down a pair of sister-keelsons over frames 18 to 43, each located 229 mm (9 in) outboard of the central keelson. The two timbers were the same sided dimensions as the main keelson, though of lesser moulded dimensions; like the keelson, they were notched down over the floors about 38 mm (1½ in) and through-bolted in place, with one bolt per frame, laid down in a staggered pattern with the bolts clenched over rings.

figure

Figure 8. Interior profile of cylinder timbers and supports. (Drawing K. Crisman)

Download figure to PowerPoint

Table 11. Dimensions of cylinder and bearing timbers
 NoteLengthMoulded/widthSided/depth
MetricImperialMetricImperialMetricImperial
Sister-keelsons12.9 m42 ft 4 in152 mm6 in229 mm9 in
Stanchions (7 per sister-keelson)large (1, 3, 5, 7)1.37 m4 ft 6 in127 mm5 in203 mm8 in
small (2, 4, 6)89 mm3½ in89 mm3½ in
Diagonal braces (paired, 6 per sister-keelson)3.66 m12 ft152 mm6 in64 mm2½ in
Lower cylinder timbers13.46 m44 ft 2 in241 mm9½ in229 mm9 in
Middle cylinder timbers13.15 m43 ft 2 in (est.)305 mm12 in229 mm9 in
Upper cylinder timbers12.57 m41 ft 3 in (est.)305 mm12 in229 mm9 in
Main shaft outboard bearing timberslower5.44 m17 ft 10 in267 mm10½ in216 mm8½ in
middle3.68 m12 ft 1 in184 mm7¼ in216 mm8½ in
upper3.4 m11 ft 2 in343 mm1 ft 1½ in216 mm8½ in
Paddle shaft inboard bearing timberslower1.8 m5 ft 11 in419 mm1 ft 4½ in216 mm8½ in
middle1.8 m5 ft 11 in343 mm1 ft 1½ in178 mm7 in
upper1.8 m5 ft 11 in44 mm1¾ in178 mm7 in
Paddle shaft outboard bearing timberslower6.76 m22 ft 2 in432 mm1 ft 5 in178 mm7 in
stack of 4-bottom1.36 m4 ft 5½ in127 mm5 in1526 in
stack of 4-top three1.36 m4 ft 5½ in102 mm4 in152 mm6 in

The sister-keelsons provided the foundation for the stack of cylinder timbers located at deck level directly above them (Fig. 5). Three types of supports or braces extended between the hull bottom and main deck. The first of these was a row of seven stanchions evenly spaced atop each sister-keelson. The stanchion heads fit into mortises cut into the bottom-most cylinder timbers and their tenonned heels fit into sloping grooves in the sister-keelsons. The posts were of two sizes: stanchions 1, 3, 5, and 7 (from forward) were more substantial in section than the three intermediate stanchions (2, 4, and 6). The larger stanchions had angled notches in their fore and after surfaces to fit the ends of the three pairs of diagonal braces between each sister-keelson and its overlying cylinder timbers. The braces crossed on either side of the intermediate stanchions and were through-bolted to one another.

During a slightly later stage of construction when more cylinder timbers were fitted at deck level, seven wrought-iron rods (six of them hinged in the centre by a pair of linking eyes) were inserted into vertical holes in each sister-keelson and the stack of cylinder timbers above. Iron plates fit around the lower ends of the rods where they extended beneath the sister-keelsons. When square nuts on the ends of the rods were tightened, they compressed the stanchions and diagonal braces, ensuring that the engine and flywheel mountings stayed as rigid as possible.

The cylinder timbers consisted of two parallel assemblies, each composed of three large oak timbers. The lower timber was fitted under the deck and its top surface was notched to lock around the lower halves of deck beams 20 to 34. The middle timber was atop the deck and notched on its underside to fit around the top halves of the same beams. Thus, the lower and middle timbers effectively clamped around the beams, holding this part of the deck structure very tightly in place. The uppermost cylinder timber had the same moulded and sided dimensions as the middle timber, although it was slightly lower at its forward end; three 102 mm (4 in) square blocks of wood were set into notches cut along the seams between the upper and middle timbers to prevent them from loosening under the forces generated by the moving piston, crank, and flywheels. The three cylinder timbers had exceptionally close-fitting seams and they were heavily bolted to one another. The top of the upper timber yielded evidence of its function: the forward end was notched to fit the cylinder head, and further aft a pair of cast-iron bearings, also known as pillow blocks, supported the main shafts that extended outboard of the flywheels.

Heroine's builders attached three additional wooden structures to each side of the main deck to support the paddle-wheel elements (now missing on the starboard side). The first of these, the main shaft outboard bearing timbers, was located inboard of the sheer-cap and held the pillow block for the outboard end of the main shaft (Fig. 9). This bearing assembly contained three timbers. The lowest was notched down over deck beams 27–33 and bolted to the paddle beams at its ends. The middle and upper pieces were considerably shorter in length and tapered toward the top. The upper timber was notched on its top surface to fit the pillow block. Two wrought-iron rods, each threaded at either end for nuts, spanned the deck between the cylinder and bearing timbers to limit flexing between the two shaft bearings. Three wooden battens nailed to the inside face of the port bearing timbers secured treads and risers for a set of stairs that allowed passage over the main shaft (the shaft's centreline was 0.61 m (2 ft) above the main deck) (Hunter, 1969: 145).

figure

Figure 9. Model of Heroine wreck showing bearing support timbers. (Model G. Grieco, photograph W. Smith)

Download figure to PowerPoint

Two other bearing-timber assemblies on the guards supported the inboard and outboard ends of the paddle-wheel shaft. The inboard bearing timbers consisted of a stack of three timbers located 254 mm (10 in) outboard of the sheer-cap and bolted atop the truncated ends of deck beams 29 to 31. The lower timber (notched down over the three beam ends) and the middle timber were heavy balks, while the upper was a plank that raised the pillow block to the correct height. In addition to being bolted to the beam ends, this bearing support was also attached laterally to the main shaft outboard bearing timbers by four long bolts with threaded ends as well as two blocks of wood that served as spacers and also as axis points for the lever that disengaged the cast-iron coupling connecting the main and paddle shafts.

The outboard paddle shaft bearing timbers were attached to the top of the guard-cap at the outer edge of the guard. The lowest timber in the assembly straddled the paddle-wheel opening between the two paddle beams forward (26–27) and the two aft (33–34); it was both long and of considerable moulded dimension, since it had to support the great weight of the paddle wheel and not sag over the lifetime of the steamboat. A squared stack of four short bearing timbers was bolted atop the centre of the lowest piece; together these elevated the outboard pillow block to the correct height for the paddle shaft.

Deck and superstructure features

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

Three-quarters of the main deck was missing or severely damaged, but excavators were able to identify several deck features or traces of now-missing elements (Table 12). Evidence of the railing that extended around the outside of the main deck survived in three locations on the port-side guard-cap: at the bow, forward of the sidehouse (the enclosed space on the guard immediately forward of the paddle-wheelhouse), and from the after side of the wheelhouse to the corner of the guard at the stern. The stanchions that supported the rail fit through mortises cut in the guard-cap and were spiked to the sides of deck beams. A cap rail was fastened atop each stanchion by a pair of spikes; there were no side rails. The stanchion at deck beam 37 supported two sections of rail connected by an iron hinge; the forward section was folded back, indicating there was an entryway for cargo and passengers directly abaft the wheelhouse. Further forward, at deck beam 22, excavators found evidence of a rail stanchion half the thickness of the others, suggesting that it was fastened to the forward corner of the sidehouse.

Table 12. Dimensions of deck features
  LengthMoulded/widthSided/depthHeight above sheer-capDistance between posts
MetricImperialMetricImperialMetricImperialMetricImperialMetricImperial
Rail stanchions489 mm1 ft 7¼ in102 mm4 in102 mm4 in
Rail489 mm1 ft 7¼ in152 mm6 in51 mm2 in
Timber headsat bow127 mm5 in152 mm6 in330 mm1 ft 1 in483 mm1 ft 7 in
at stern114 mm4½ in140 mm5½ in305 mm12 in355 mm1 ft 2 in
Bitts152 mm6 in152 mm6 in254 mm10 in

With its top surface only 0.43 m (17 in) above the main deck, the rail was very low, about knee-height for an average-sized adult. While better than no rail at all, it offered limited protection against falling overboard. However, the low-profile rail was practical for everyday operations, for lines could easily be led over the side, and small boats were accessible from any location along the guard, not just at rail openings. Paintings and prints of contemporary steamboats compiled for comparison with Heroine mostly show conventional waist-high railings, but a few, notably a watercolour of the 160-ton steamboat Ouishita (1833), show a knee-high rail (Levinge, 1836; Crisman et al. 2013: 376, fig. 13).

The main deck was fitted with four pairs of short, upright black locust posts called timberheads, used to tie off lines and hawsers to the shore, to anchors, or to other vessels (Lehman, 1972: 446) (Figs 2 and 3). The timberheads fit in mortises cut through the sheer-cap and their lower ends were fastened to frame tops and clamps. Two pairs were located on either side of the bow between deck beams 1 and 3. Their upper corners were chamfered, giving them a slightly rounded appearance on top. The other two pairs of timberheads were located at the stern, straddling beam 47 at the port and starboard sheer-caps. Their tops were squared rather than rounded. The corners of the timberheads were rounded by rope wear.

A pair of small, upright posts, the bitts, was located on the foredeck at deck beam 1. Their heels were fastened to the top of the lower breasthook, and they extended up through mortises cut in the deck planks. The original height of the bitts is unknown, since a small fire on the foredeck burned both posts down to the planking (this probably happened between 1838 and 1843). Bitts are traditionally used for securing anchor cables, but the principal purpose of these two posts was clearly to hold the heel of a short bowsprit (bowsprits were fitted on many western steamers until the mid 1830s), or the heel of a vertical sighting pole used by the pilot to take bearings on the shoreline.

Contemporary steamboat prints and paintings often show a capstan on the bow (Hunt and Gallagher, 1984: 116). No capstan was found on Heroine, but evidence of a mount was discovered (Fig. 2). A small, square opening—presumably for the heavy retaining bolt—was cut through the centre of the foredeck, beneath which two heavy carlings extended between beams 4 and 5. A square oak block with a square hole through the centre, sized to fit between the carlings, was discovered loose in the bow compartment. Finally, the centreline of the deck between the forward hatch and stem was covered by four oak planks, the only oak decking found inboard of the sheer-cap. A small square of wood nailed over the deck opening indicated that the capstan was removed prior to Heroine's loss (salvors had no reason to cover the hole).

Evidence of three structures extending above the main deck was found on the after portion of the port guard and stern deck. The first was indicated by a pair of square impressions in the guard planking and a nailed-down batten slightly outboard of the sheer-cap at deck beams 36 and 37; together these showed that a rectangular assembly extended out from the superstructure. An 1833 watercolour of the steamboat Delphine shows a similar feature at this place on the starboard side of the deck, possibly an enclosed staircase that provided access to cabins on the upper deck (Hunt and Gallagher, 1984: 116).

The second feature, a square opening in the guard planking between beams 45 and 46 near the stern, had a corner-notched plank resembling a door sill nailed to the deck on its after side. The deck planks around the opening had deteriorated, and a repair plank was nailed over the worst of the erosion. This feature was identified as the opening for a head (toilet). The main deck plan of the steamboat Buckeye State (1850) has a square structure in the same location labeled ‘crew head’ (Tredgold, 1851: Pl. 3).

The third feature consisted of three wooden battens nailed to the after deck inboard of the crew head and forward of the run's companionway. Contemporary illustrations show that the after end of the superstructure was in this vicinity, and it thus seems likely that the battens were used to anchor the lower edge of the deck-level sheathing planks.

Very little of Heroine survived above the main deck, but clues to the location and assembly of the superstructure remained. The stubs of 102 mm (4 in) square support posts protruded from mortises cut in the port-side sheer-cap between the fore hatch and after timberheads; the heels of the posts were spiked between frame heads. Spaced an average of 2.43 m (8 ft) apart, these stationaries provided the framework for the superstructure. Square mortises were also found atop the port-side guard structure, evidence of framing posts for the wheelhouse and sidehouse. Remains of 25 mm (1 in) thick tongue-and-groove pine boards were still attached to the outside of the outboard paddle shaft bearing timber; it is likely that similar sheathing once covered most of the superstructure (Jackson, 1985: 161–162). Together, the remnants of the posts and their mortises provided a ‘footprint’ of Heroine's lightly built upperworks.

The rudder

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The rudder was recovered in 2003 and transported to Texas A&M University's Conservation Research Laboratory for recording, disassembly, and conservation (Fig. 10). It was a broad, substantial assembly with a core composed of seven vertical timbers: the rudder post, five central pieces (four quite thick and the fifth resembling a plank), and a narrow back piece that projected diagonally upward to the same height as the post (Table 13). The timber adjacent to the post had a knee-like head that reinforced the attachment of the blade to the post. The post was parallel-sided in profile, but the remaining pieces increased in width toward the base. The entire rudder tapered in thickness from top to bottom.

figure

Figure 10. The rudder of Heroine. (Drawing K. Crisman, Grace Tsai)

Download figure to PowerPoint

Table 13. Rudder dimensions
 NoteHeight/lengthWidth/diameterNoteThicknessNoteSpike holes
MetricImperialMetricImperialMetricImperial
Shoe to rudder-head2.59 m8 ft 6½ in
Blade——1.21 m4 ftat top292 mm11½ inat rudder-head collar
——1.44 m4 ft 9 inat base121 mm4¾ inat shoe
Shoe89 mm3½ in
Side planks (port)——

229 mm;

448 mm;

445 mm

9 in

1 ft 5⅝ in

1 ft 5½ in

top to bottom32 mm1¼ in
Side reinforcing strapsport1.14 m3 ft 9 in16 mm⅝ in
starboard0.85 m2 ft 9½ in16 mm⅝ in
Tiller mortise178 mm7 in127 mm5 in
Upper pintlestraps0.91 m3 ft48–76 mm1⅞–3 intaper from pintle16 mm⅝ in

5 (port)

6 (starboard)

pintle133 mm5¼ in44 mm1¾ in
Lower pintlestraps1.07 m3 ft 6½ in44–64 mm1¾–2½ intapered16 mm⅝ in7 per side
pintle133 mm5¼ in44 mm1¾ in
Upper gudgeon straps355 mm1 ft 2 in51 mm2 in25 mm1 in2 per side
Lower gudgeon straps635 mm2 ft 1 in51 mm2 in25 mm1 in3 per side

The rudder's seven core pieces were held together by two fore-and-aft through-bolts, by a shoe nailed to the bottom of the blade, and by three oak planks fastened to each side of the blade with many small iron nails. The ends of the side planks were bevelled, as were the upper edges of the topmost planks; the lower edges of the bottom-most planks fit flush with the top of the rudder shoe. Near the top of the blade the core rudder timbers were further secured to one another by a pair of wrought-iron straps, each with eight spike holes.

The force required to pivot the rudder against the flow of water must have been considerable, for the head of the rudder was squared around the tiller mortise and heavily reinforced to prevent splitting. Two wrought-iron collars circled the post, one above and one below the mortise. The D-shaped upper collar was fitted with its flat side forward and secured by many small shims or wedges. The squaring of the post at the tiller mortise made it impossible to slide the lower collar into place, so it was made of two semi-circular pieces held together forward and aft by door-hinge-like arrangements with three knuckles and a pin. This collar was also wedged in position by shims. The sides of the tiller mortise were reinforced by two pairs of vertical iron straps through-bolted to one another above and below the mortise.

The rudder retained evidence of maintenance, damage and repairs. The top of its port side had patches of grayish-white paint, a coating that was presumably applied to the entire rudder. Small nails around the head of the post likely held a waterproof cover or boot that kept water from splashing into the hull through the tiller opening. Five longitudinal splits were noted on the starboard side of the post, above the upper pintle; these were filled by nails and small spikes to keep the cracks from spreading. The starboard surfaces of the rudder blade were also damaged beneath the upper pintle. When the rudder was disassembled for conservation its fore-and-aft bolts were found to be broken in the centre of the blade, and damage was noted at the bottom of the central core pieces. This suggests the rudder experienced a violent impact, perhaps when the vessel ran aground or collided with a snag. The plank-like piece in the centre of the core timbers, the shoe at the bottom, the side planks, and the two iron reinforcing straps appear to have been added to repair the blade.

The purpose of the high back piece on the rudder is something of a mystery. Its eroded head retained evidence of a notch or mortise, suggesting it was attached to something at one time. In his Sketch of the Civil Engineering of North America, David Stevenson illustrated a similar rudder projection, noting that on American steamboats it was used as an attachment point for wheel ropes, permitting boats to dispense with a tiller (1838: 132). The back piece on Heroine's rudder seems too weak for this purpose.

The tiller (found in the run) was squared at its after end to fit into the rudder-post mortise, but was otherwise rounded and tapered toward its forward end. Curiously, the tiller was assembled from two pieces that fit together in a long tongue-and-groove joint secured by wooden pegs. This feature seems unnecessarily complicated, since it would have been easier, faster, and probably stronger to fashion the tiller from one piece, but this may be an intentional weak point in the steering system, designed to give way and prevent damage to the rudder head and the tiller ropes and blocks if the rudder received a hard blow or was unshipped from the gudgeons.

Western river steamboat development

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The creation of the western river steamboat was an empirical, trial-and-error process, and the shipwrights who assembled them seemingly had little time or inclination to prepare detailed plans and written specifications of their work. If such technical material was generated, little of it has survived. The historical record tells us that the deep-draft hulls and underpowered engines of the first steamers proved unsuited for conditions on the Mississippi and Ohio Rivers, and were replaced by lightly built, shallow hulls propelled by powerful—though dangerous—high-pressure engines. This much was known, but a clearer picture of the rapidly evolving designs, assemblies, and operations of early boats has been lacking. The wreck of Heroine provides a new benchmark for understanding the first quarter-century of western river steamboats.

What has Heroine told us? It is clear that when Heroine was built in 1832 many salient features of ‘classic’, later 19th-century western river steamboats were in the process of development and adoption. The heavy construction said to be typical of early river steamboat hulls had metamorphosed into a structure composed of fewer, lighter timbers. Heroine's keel was little more than a thick plank that projected slightly beyond adjacent bottom planking. The frames were assembled from a minimal number of timbers, the floors and futtocks were of modest moulded and sided dimensions and lightly fastened with a few spikes, and the frames were widely spaced along the keel.

The deck assembly was likewise lightweight, and consisted of small beams attached to the upper clamp and sheer plank with only two spikes. The omission of hanging, lodging, or dagger knees from the deck structure was a notable departure from wooden shipbuilding practices on oceans and lakes (Robinson, 1999: 169–73). Knees or other forms of bracing at the juncture of the beams and clamps were considered essential reinforcement for vessels subjected to high, ship-twisting waves on open waters. Such waves are not found on the western rivers, but steamboats did frequently run aground, which presumably had much the same effect on their hulls. While knees improve the strength and durability of a vessel, they add extra weight, and this apparently ruled out their use in Heroine; the practice of omitting knees is also evident in the hulls of later steamboats (Kane, 2004: 107). Heroine's main deck assembly was not very rugged, and would have fared poorly in a collision with another boat, a wharf, or a river bank.

The ends of the hull differed: the dimensions of the principal members were larger, the timbers were closely spaced, and both the form of the hull and its assembly were more complex. In addition to its substantial stem and apron, the bow had doubled floor timbers in the first 12 square frames, closely spaced cant frames, and two large breasthooks to reinforce the stem and cant frames. The builders obviously wanted this part of the hull to withstand repeated groundings and collisions. The bow also had a feature common to western river steamboats of the 1820s and 1830s, a snag chamber bulkhead intended to contain flooding if the bow was damaged. The stern was equally notable for its carefully shaped and closely fitted deadwood timber, stern knee, inner and outer posts, and three transom timbers.

One goal of the Heroine investigation was to discover the means by which the boat's long, narrow wooden hull was longitudinally reinforced, since it pre-dated the introduction of hog chain trusses in the late 1830s and early 1840s. Heroine revealed a heretofore-unknown practice in river steamboat construction: major longitudinal timbers—the keelson, sister-keelsons, bilge-keelsons, and upper and lower clamps—were notched down over every frame. Fitting the keelson over floors was a common shipbuilding practice for heavily built open-water vessels, but has not been previously seen on other archaeologically studied western river steamboat wrecks, all of which post-date Heroine by 20 or more years (Kane, 2004: 34–43, 93–107). The joggling of the principal longitudinal timbers was a labour-intensive, and costly, task, requiring the measuring and cutting of 550–600 notches. It is not surprising that builders abandoned this practice and embraced hog chains when the new trussing system was introduced a few years later. While Heroine did not have hog chains, it did have at least one cross chain, a transverse wrought-iron truss located forward of the wheelhouses which supported the weight of the guards and side wheels.

Steamboat builders of the mid and later 19th century found a second way to strengthen hulls longitudinally, by constructing a non-watertight bulkhead between the keelson and deck that extended the full length of a vessel. Large steamers often had two additional bulkheads fitted, with one on either side of the keelson. Bulkheads were typically composed of stanchions and planks, sometimes with diagonally laid braces. The archaeological record confirms their ubiquity, for wrecks of later-era river steamers consistently show evidence of these longitudinal stiffeners, but they were absent from Heroine (Kane, 2004: 108–10).

Heroine's cylinder and bearing support timbers indicated the challenges of installing a powerful reciprocating-engine propulsion system on a lightly built, flexible hull. The wooden longitudinal elements composing this structure were all heavily notched, bolted, and braced, and all were held in compression by iron reinforcing rods with threaded ends. The result was, in effect, a torsion-resistant structure resembling a box girder built within and atop the hull. This was complemented by two features in the machinery: loose tolerances in the connections between moving parts, and by provisions for easily adjusting the bearings to re-align main and paddle shafts.

Samples taken from timbers throughout the hull showed that white oak was the preferred species for nearly all elements (the exceptions were pine deck planking and tongue-and-groove sheathing, and black locust for timberheads at the bow and stern). The quality of wood available to builders was good, but certain pieces were shaped in a manner that weakened the overall structure. The practice of cutting frame timbers across the grain to achieve a hard bilge curve at the first futtocks was a particularly noteworthy example of a hull-weakening construction shortcut. This practice saved material and labour expenses, but doubtless resulted in boats sometimes not surviving relatively minor groundings or collisions. Frames at the bow and stern, on the other hand, appear to have incorporated more compass timber in their construction. Iron fasteners were used exclusively in Heroine's assembly, sparingly in the case of the spiked frames, and more lavishly at the clench-bolted sister-keelsons, bilge-keelsons, clamps, and plank butts.

Archaeological documentation of Heroine yielded a myriad of other early steamboat construction and layout details, including the existence of separate compartments at the bow and stern, the size and placement of the forward hatches and bow and stern companionways, the locations of stairs and the crew head, and the surprisingly low height of the railing around the main deck. The stumps of the superstructure-supporting stationaries showed the wide spacing and modest size of these elements, and hint at lightweight, flimsy construction in the upper deck and cabins.

Frame sections recorded over the length of the wreck and documentation of the bow and stern assemblies enabled reconstruction of Heroine's lines (Fig. 11). These show a vessel with an overall length of 41.65 m (136 ft, 8 in) between perpendiculars, a moulded breadth of 6.2 m (20 ft, 4 in), a maximum breadth across the main deck of 10.97 m (36 ft), and a depth of hold of 1.98 m (6 ft, 6 in) from the top of the keel to the main deck (the contemporary standard for measuring depth of hold on river steamboats) (US Treasury Department, 1838: 312–15). Over most of its length the floors were nearly flat, the bilges had a hard curve, and the sides were near-vertical, giving the hull a boxy section amidships. Heroine was hardly a tub, however. The length-to-beam ratio was quite high at 6.72:1, and this, combined with the long, raking entrance and short but finely moulded run, made an efficient form that likely passed cleanly and swiftly through the water. The 1.52 m (5 ft) draft mark on the stem tells us about how much water the boat was expected to draw when laden (the freeboard at these times would have been about 0.61 m (2 ft)). According to contemporary sources the tonnage of this hull ranged between 146 and 160 tons (Collins, 1836: 62; Hall, 1836: 252, 256; Lyford, 1837: 464); the discrepancy in the listings may reflect the fact that steamboats over 150 tons were required to pay their agents a fee for each voyage that was twice as much as that for boats under 150 tons (Keemle, 1836: 34).

figure

Figure 11. The reconstructed lines of Heroine. (Drawing K. Crisman)

Download figure to PowerPoint

Was Heroine typical of its day? Limited historical evidence suggests that in terms of tonnage and dimensions, it fell in the average-sized range. An 1832 inventory of 183 river steamboats and their tonnages (of the 220 reportedly operating this year) shows the majority had around the same tonnage (Otis, 1832: 127–30; Chevalier, 1961: 209). An 1838 US Treasury Department report provides dimensions for 113 of the estimated 400 steamboats operating on the western rivers at this time. Their average length was between 42.7 and 45.7 m (140 and 150 ft), just slightly larger than Heroine (US Treasury Department, 1838; Morrison, 1958: 219). As noted earlier, other types of comparison are difficult: there are no other archaeologically studied examples of contemporary river steamboats, nor have any lines, construction plans, or detailed specifications been found.

Successful navigation of the Mississippi River system by shallow-draft, high-pressure steamers conferred great benefits on the inhabitants of the mid continent region, but it came at a price. The lightweight construction seen on Heroine was vital to its operational flexibility and ultimately to the boat's profitability, but it was not durable. The grind of daily operations quickly wore out both hulls and machinery; the working life of a western river steamboat rarely exceeded five years. Many failed to last even that long, for fires, ice packs, collisions with other vessels, groundings, explosions, and especially snaggings took a steady annual toll of both old and new boats working on the rivers (Hunter, 1969: 100–3, 112, 271–304; Brockmann, 2002). Viewed in those terms, the fact that Heroine lasted five-and-a-half years of hard service before being fatally snagged might be viewed as a testament to both its relative quality of construction and its luck in escaping the other perils of the rivers.

Conclusions

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

North America's cultural and environmental landscapes experienced profound changes over the course of the 19th century, processes greatly hastened by the introduction of steam propulsion technology. Nowhere was this more evident than on the great rivers of the interior, the Mississippi and its tributaries. Here a distinctive style of vessel, the western river steamboat, facilitated immigration, growth of agriculture and industry, and expansion of trade. While these boats have come to symbolize a region and an era, the dynamic early years of their evolution was poorly documented; we know what these boats accomplished, but relatively little about their design, construction, and appearance. The wreck of Heroine (1832–1838), the oldest archaeologically investigated example of its type, has yielded the first detailed, keel-to-main-deck look at a western river steamboat from the developmental era.

Notes
  1. 1

    All samples of wood from Heroine's hull were identified by Dr Roy Whitmore, Professor Emeritus, Forestry Department, University of Vermont.

  2. 2

    The weight estimate calculations were prepared by Glenn Grieco from his CAD reconstruction of Heroine's boiler system, based on the remains of the steamboat's boiler system and on the arrangement seen in P. Hodge's The Steam Engine, Plate XXXIII. Materials included the cast- and wrought-iron in the boilers and their mountings (35.7 cubic ft @ 491 pounds per cubic ft = 17,528.7 lbs); brick for the lining of the fireboxes (18.65 cubic ft @ 120 pounds per cubic ft = 2234.4 lbs [the standard weight for red brick, Heroine's fire brick may be heavier but the difference is probably negligible]); and the weight of the water which filled the boilers about half way when the boilers were used (174.44 cubic ft @ 62.4 pounds per cubic ft = 10885.056 lbs). This totals 30,648 lbs (15.32 tons, or 13,898 kg, or 13.9 metric tonnes).

Acknowledgements

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References

The study of Heroine was, and continues to be, supported by the Oklahoma Historical Society, Texas A&M University (TAMU), the Institute of Nautical Archaeology, and the Center for Maritime Archaeology and Conservation at TAMU. Analysis and reconstruction of the hull would not have been possible without the painstaking recording work carried out by project crewmembers over seven field seasons. The author also extends many thanks the University of Maine's Darling Marine Center for providing an ideal location for completing the preparation of this article.

References

  1. Top of page
  2. Abstract
  3. History and archaeological study
  4. Construction and materials
  5. Keel
  6. Stem assembly
  7. Stern assembly
  8. Frames
  9. Keelson
  10. Stringers, bilge-keelsons, and clamps
  11. Planking
  12. Hold, bow and stern compartments
  13. Main deck structure
  14. Hatches and companionways
  15. Cylinder and bearing timbers
  16. Deck and superstructure features
  17. The rudder
  18. Western river steamboat development
  19. Conclusions
  20. Acknowledgements
  21. References
  • Bates, A., 1968, The Western Rivers Steamboat Cyclopoedium. Leonia, NJ.
  • Bates, T., 2003, Analysis of Tar-like Material from the Red River Wreck, December 16, 2003. Manuscript report, Exxon-Mobil Petroleum Engineering Laboratory, Houston, TX.
  • Brockmann, J., 2002, Exploding Steamboats, Senate Debates, and Technical Reports: The Convergence of Technology, Politics, and Rhetoric in the Steamboat Bill of 1838. Amityville, NY.
  • Chapelle, H., 1960, The National Watercraft Collection. Washington, D.C.
  • Chevalier, M., 1961, Society, Manners, and Politics in the United States. Ithaca, NY.
  • Collins, G., 1836, The Louisville Directory for the Year 1836. Louisville, KY.
  • Cozzi, J., 1998, The Goniometer: an improved device for recording submerged shipwreck timbers. IJNA, 27.1, 6480.
  • Crisman, K, Lees, W., and Davis, J., 2013, The Western River Steamboat Heroine, 1832–1838, Oklahoma, USA: excavations, summary of finds, and history. IJNA, 42.2, 365381.
  • Custer, J., 1991, ‘Hogchains: Their Whys and Wherefores,’. The Egregious Steamboat Journal, 15 (July–August), 1015.
  • Hall, B., 1974, Travels in North America in the Years 1827 and 1828. New York.
  • Hall, J., 1836, Statistics of the West, at the Close of the Year 1836. Cincinnati.
  • Hodge, P., 1840, The Steam Engine, Its Origin and Gradual Improvement. New York.
  • Hunt, D and Gallagher, M. (eds), 1984, Karl Bodmer's America. Lincoln, NE.
  • Hunter, L., 1969, Steamboats on the Western Rivers: An Economic and Techno logical History. New York.
  • Ingraham, J. H., 1839, Dots and Lines No. IV; or Sketches of Scenes and Incidents in the West. The Ladies' Companion, A Monthly Magazine, Embracing Every Department of Literature, XI, New York.
  • Jackson, D., 1985, Voyages of the Steamboat Yellowstone. New York.
  • Kane, A., 2004, The Western River Steamboat. College Station, TX.
  • Keemle, C., 1836, The St. Louis Directory for the Years 1836–7. Saint Louis, MO.
  • Lehman, C. F., 1972, A Riverman's Lexicon. Florissant, MO.
  • Levinge, R. G. A., 1836, The Paddle Steamer Ouishita on the Red River, Louisiana Territory. Watercolor painting 1968.271, Amon Carter Museum, Fort Worth, TX.
  • Lyford, W. G., 1837, The Western Address Directory. Baltimore.
  • Marestier, J. B., 1957, (translation of 1824 edition). Memoir on Steamboats of the United States of America. Mystic, CT.
  • McDermott, J. F., 1968, Before Mark Twain: A Sampler of Old, Old Times on the Mississippi. Carbondale, IL.
  • Morrison, J. H., 1958, History of American Steam Navigation. New York.
  • Otis, R. W., 1832, The Louisville Directory for the Year 1832. Louisville, KY.
  • Robinson, D., 1999, Indiana: The History and Archaeology of an Early Great Lake Propeller. M.A. Thesis in Anthropology, Texas A&M University, College Station, TX.
  • Stevenson, D., 1838, Sketch of the Civil Engineering of North America. London.
  • Tredgold, T., 1851, Machinery Used in Steam Navigation, Vol. II, Part I. London.
  • US Treasury Department, 1838, Steam Engines: Letter from the Secretary of the Treasury, Document 21, December 13, 1838. Washington, D.C.
  • Works Progress Administration, 1942, Ship Registers and Enrollments of New Orleans, Louisiana. Baton Rouge, LA.