The Bark Canoes and Skin Boats of North America

Chapter Three
FORM AND CONSTRUCTION

Classification of the types of bark canoes built by the Indians is not a simple matter. Perhaps the most practical way is to employ the tribal designation, such as Cree canoe, Micmac canoe, accepting as a criterion the distinctive general appearance of the canoes used by each tribe. It must be emphasized, however, that this method of classification does not indicate the model, or "lines," employed. Both the model and the size of bark canoes were extensively affected by the requirements of use: lake, coastal, or river navigation; smooth, rough, or fast-running water; transportation of a hunter, a family, or cargo; the conditions and length of portages; and the permanence of construction desired. Canoes of various models, sizes, methods of construction, or decoration might be found within the limits of a single tribal classification. Also, within a given area, there might be apparent similarity in model among the canoes of two or three tribal groups. However, a classification based on geographical areas has been found to be impractical, because the movements of tribal groups in search of new hunting grounds tend to make tribal boundaries difficult to define.

Form

The canoes of some tribal groups appear to be hybrids, representing an intermingling of types as a result of some past contact between tribes. Those of other groups are of like model, form, and even appearance, possibly owing to like conditions of employment. The effects of a similarity in use requirements upon inventiveness is seen in the applications for modern patent rights, where two or more applications can cover almost exactly the same device without the slightest evidence of contact between the applicants; there is no logical reason to suppose the same condition cannot apply to primitive peoples, even though their processes of invention might be very slow or relatively rare in occurrence.

The effects of migration of tribes upon their canoe forms can only be studied with respect to those comparatively recent times for which records and observations are available. From the limited information at hand it appears that the Indian, when he moved to an area where use requirements and materials available for building differed from those to which he had been accustomed, was often forced to modify the model, form, size, and construction of his canoe. In some instances this seems to have resulted in the adoption of another tribal form.

The distinctive feature that usually identifies the tribal classification of a bark canoe is the profile of the ends, although sometimes the profile of the gunwale, or sheer, and even of the bottom, is also involved. The bow and stern of many bark canoes were as near alike in profile as the method of construction would permit; nevertheless some types had distinct bow and and stern forms. Among tribes the form of the ends of the canoes varied considerably; some were low and unimpressive, others were high and often graceful.

Obviously practical reasons can be found for certain tribal variations. In some areas, the low ends appear to ensue from the use of the canoe in open water, where the wind resistance of a high end would make paddling laborious. In others the low ends appear to result from the canoe being commonly employed in small streams where overhanging branches would obstruct passage. Portage conditions may likewise have been a factor; low ends would pass through brush more easily than high. Types used where rapids were to be run often had ends higher than the gunwales to prevent the canoe from shipping water over the bow. The high, distinctive ends of the canoes most used in the fur trade, on the other hand, were said to have resulted from the necessity of employing the canoe as a shelter. When the canoe was turned upside down on the ground, with one gunwale and the tops of the high ends supporting it, there was enough headroom under the canoe to permit its use as a shelter without the addition of any temporary structure. The desirability of this characteristic in the fur-trade canoe can be explained by the fact that the crew travelled as many hours as possible each day, and rested for only a very short period, so that rapid erection of shelter lengthened both the periods of travel and of rest.

Yet these practical considerations do not always explain the end-forms found in bark canoes. Canoes with relatively high ends were used in open waters, and similar canoes were portaged extensively. Possibly the Indian's consciousness of tribal distinctions led him to retain some feature, such as height of the end-forms, as a means of tribal recognition, even though practical considerations required its suppression to some degree.

The profile of the gunwales also varied a good deal among tribal types. Most bark canoes, because of the raised end-forms, showed a short, sharp upsweep of the sheer close to the bow and stern. Some showed a marked hump, or upward sweep, amidships which made the sheer profile follow somewhat the form of a cupid's bow. Many types had a straight, or nearly straight, sheer; others had an orthodox sheer, with the lowest part nearly amidships.

The bottom profiles of bark canoes showed varying degrees of curvature. In some the bottom was straight for most of its length, with a slight rise toward the ends. In others the bottom showed a marked curvature over its full length, and in a few the bottom was practically straight between the points at which the stems were formed. Some northwestern types had a slightly hogged bottom, but in these the wooden framework was unusually flexible, so that the bottom became straight, or even a little rockered when the canoe was afloat and manned.

The practical reasons for these bottom forms are not clear. For canoes used in rapid streams or in exposed waters where high winds were to be met many Indians preferred bottoms that were straight. Others in these same conditions preferred them rockered to varying degrees. It is possible that rocker may be desirable in canoes that must be run ashore end-on in surf. Of course, a strongly rockered bottom permits quick turning; this may have been appreciated by some tribal groups. Still other Indians appear to have believed that a canoe with a slightly rockered bottom could be paddled more easily than one having a perfectly straight bottom.

The midsections of bark canoes varied somewhat in form within a single tribal type, because the method of construction did not give absolute control of the sectional shape during the building, but, on the whole, the shape followed tribal custom, being modified only to meet use requirements. Perhaps the two most common midsection shapes were the U-form, with the bottom somewhat flattened, and the dish-shape, having rather straight, flaring sides combined with a narrow, flat, or nearly flat bottom. Some eastern canoes showed marked tumble-home in the topside above the bilge; often they had a wide and rather flat rounded bottom, with a short, hard turn in the bilge. A few eastern canoes, used mainly in open waters along the coast, had bottoms with deadrise—that is, a shallow V-form, the apex of the V being much rounded; the V-bottom, of course, would have aided in steering the canoe in strong winds. One type of canoe with this rising bottom had tumble-home topsides, but another, used under severe conditions, had a midsection that was an almost perfect V, the apex being rounded but with so little curvature in the arms that no bilge could be seen.

Generally speaking, the eastern canoes had a rather well rounded bottom with a high turn of the bilge and some tumble-home above, though they might have a flatter form when built for shallow-water use or for increased carrying capacity. A canoe built for speed, however, might be very round on the bottom, and it might or might not have some tumble-home in the topside. In the West, a flat bottom with flaring topsides predominated; fast canoes there had a very narrow, flat bottom with some flare, the width of the bottom and the amount of flare being increased to give greater capacity on a shallow draft. Some canoes in the Northwest had a skiff-form flat bottom and flaring sides, with the chine rounded off sharply.

The form of the sections near the ends of a canoe are controlled to a great extent by the form of midsection. In canoes having flat bottoms combined with flaring sides this form was usually carried to the ends, where it became a rather sharp V, giving fine lines for speed when the canoe was light, and only moderately increased resistance when it was loaded. Among eastern canoes having tumble-home topsides, the midsection form could be carried to the ends, gradually becoming sharper in canoes having "chin" in the profiles of the ends; in canoes having no chin, the sections necessarily took a pointed oval form close to the ends. A few canoes having flaring sides and chin ends showed a similar change in form. In all, however, the bow and stern showed a tendency toward fullness near the waterline.

Canoes with a strongly U-shaped midsection commonly carried this form to the ends, with increasing sharpness in the round of the U. The U-form predominated in the end-sections of eastern canoes, of course, though a few showed a V-form, as must be expected. The fairing of the end sections into the end profiles appears to have controlled this matter. The outline of the gunwales, in plan view, also influenced the form of the end-sections and of the level lines there. Some canoes, when viewed from above, showed a pinched-in form at the ends, this was caused by the construction of the gunwales or by the projection of the end-profile forms beyond the ends of the true structural gunwale members. Such canoes would have a very strong hollow in the level lines projected through their hull-form below the gunwales, and this could have been accentuated by any strong chin in the bow and stern shapes. On the other hand, many canoes showed no hollow, and the level lines were straight for some distance inboard of the ends, or were slightly convex. Full, convex level lines will appear below the waterline in canoes having a strongly rockered bottom.

It should be noted that the Indians were aware that very sharp-ended canoes usually were fast under paddle; hence they employed this characteristic in any canoe where high speed was desired. However, the degree of sharpness in the gunwales and at the level lines is not always the same at both ends, though the variation is sometimes too slight to be detected without careful measurement; it may at times have been accidental, but in many cases it appears to have been intentional.

Some eastern canoes having their greatest width, or beam, on the gunwales at midlength had finer level lines aft than forward, apparently to produce trim by the stern when afloat and manned. This made them steer well in rough water. Some northwestern canoes had their greatest beam abaft the midlength, giving them a long, sharp bow; the run was sometimes formed by sweeping up the bottom aft to a shallow stern, as well as by the double-ended form of the canoe. Despite a general similarity in the form of the ends, in some canoes the bow was marked by its greater height, in others, by the manner in which the bark was lapped at the seams, or by the manner of decoration. In a few with ends exactly alike the bow was indicated by the fitting of the thwarts such as, for example, by placing at the forward end a particular style of thwart, intended to hold the torch used in spearing fish at night, or to support a mast and sail.

In examining the lines, or model, of a bark canoe, the limitations imposed upon the builder by the characteristics of bark must be considered. The degree of flexibility, the run of the grain, and the toughness and elasticity of the bark used all influenced the form of canoes. The marked chin in the ends of some canoes, for example, resulted from an effort to offset the tendency of birch bark to split when a row of stitches lay in the same line of grain. The curved chin profile allowed the stitching to cross a number of lines of grain. Sometimes this tendency was avoided by incorporating battens into the coarse stitching; this style of sewing was particularly useful in piecing out birch bark for width in a canoe, where the sewing had to be in line with the grain. The Indians also employed alternating short and long stitching in some form for the same purpose. Spruce bark, as used in canoes in the extreme North and Northwest, could be sewn in much the same manner as birch bark, but with due regard for the longitudinal grain of the spruce bark.

The joining of two pieces of bark by root sewing or lacing, combined with the use of spruce gum to obtain watertightness, formed a seam that could be readily damaged by abrasion from launching the canoe, from pulling it ashore, or from grounding it accidentally. For this reason, seams below the waterline were kept at a minimum and were never placed along the longitudinal centerline of the bottom, where they would have formed a sharp apex to both the V-shaped midsection and to the deadrise bottom form. Likewise, a seam was not used in forming the rocker of the bottom. Though seams had to be used to join the bark at bow and stern, the form of the canoe allowed the seams to be greatly strengthened and protected there.

The restrictions on form imposed by barks such as elm, chestnut, and hickory were very great. These barks, which are not as elastic as birch bark, were sometimes employed in a single large sheet. The sheets were not joined for length; canoes of this material were often formed by crimping, or lap folding, rather than by cutting out gores and then sewing the edges together. The characteristics of these barks can readily be demonstrated with a sheet of paper: such a sheet can be made into a crude canoe-form by bending it lengthwise and joining the ends, but it will be obvious that the midsection takes a very unstable U-form. By forcing the ends inward to give a ram, or chin, effect to bow and stern, a somewhat flatter bottom can be obtained in the midsection. By crimping or folding the paper gore-fashion near each end of the canoe-form at the gunwale edge, some rocker is created in the bottom and the width of the gunwales is increased near the ends, giving more capacity. But without the crimping along the gunwale, when the midsection form is flattened on the bottom, the latter tends to hog. Many of these bark canoes utilized both the rams ends and crimping to obtain a more useful form. However, while a sheet of birch bark could be crimped or gored into a scow-form canoe such as the Asiatic birch-bark canoe, no example of this form from North America is known. On this continent all bark canoes were sharp at both ends, i.e., double-ended, although a number of North American dugouts were scow-(or punt-) shaped.

Birch bark gave much more freedom in the selection of form simply because it could be joined together in small odd-sized sheets to shape a hull, and because it was elastic enough to allow some "moulding" by pressure of the framework employed. Birch bark could be gored, or slashed, and rejoined without resort to folding or crimping; thus it permitted a smooth exterior surface to be achieved. The toughness of the bark was sufficient to allow some sewing in line with the grain, to add to the width of a sheet, if the proper technique were employed (this was also true to a lesser extent of spruce bark).

The framework of most bark canoes depended upon the gunwale structure to give longitudinal strength to the hull; for this reason the structure was made sufficiently large in cross-section to be rather stiff, or was formed of more than one member. An inner and outer gunwale construction was employed in many bark canoes. The inner member was the strength member and was sometimes square, or nearly so, in cross-section. In some canoes bark was brought up on the outside of this gunwale member, lapped over the top, and lashed over it; in others the bark was lashed to both inner and outer gunwales. The outer gunwale, a rectangular-sectioned batten bent narrow-edge up, was applied like a guard, outside the bark, and was secured by pegs, by the lashings of the bark cover, or by widely spaced lashings. On top of the large inner gunwale and usually extending outward over the outer gunwale, a thin cap, pegged or lashed in the same manner as the outer gunwale, was sometimes added; this was intended to protect the lashing of the bark to the gunwale rather than to add longitudinal strength.

The corners of the inner gunwale, or of the single gunwale, were all rounded off to prevent them from cutting the sewing and lashings. The bottom outboard corner was sometimes rounded off more than the other, or beveled, in order to form between the outboard face of the gunwale and the bark a slot into which the heads of the ribs could be forced. An alternate method of accomplishing this was to notch or drill holes in the gunwales for the heads of the ribs.

The ends of the gunwales were fashioned in various ways. In some canoes the gunwales were sheered upward at the ends only slightly, the gunwale ends being secured to wide end boards in the stems or extended past them and secured to the stem-pieces. The apparent sheer in the latter might be formed by bending the outer gunwale, or outwale, and the cap (if one existed) to the required curve and then securing the ends to the stem-piece, or to the end boards, as the form of end profile dictated. If either the single gunwale or the outwale or both were sharply sheered, they were split, to a point near the end thwart, into two or four or even more laminations; even the rail cap, which was perhaps half an inch thick, might be split in the same manner to allow a sharp upward sweep at the stems. After being bent, the split members were temporarily wrapped to hold the laminations together. In no bark canoes did the ends of the gunwales curve back on themselves to form a hook just inboard of the bow and stern, despite the numerous pictures that show this feature. The gunwale ends sometimes projected almost perpendicularly upward, slightly above the top of the bow and stern, so that when the canoe was upside down its weight came on these rather than on the sewing of the ends of the craft.

The gunwale ends in some canoes were fastened together by means of one or more lashings, often widely spaced. After being lashed together, a narrow wedge was sometimes driven between the two gunwales from inboard to tighten the lashings. The ends were sometimes beveled on their bearing surfaces so as to make a neat appearance when joined. The various ways in which the gunwale ends at stem and stern were finished can best be described when individual types are under examination. Some canoes had a small piece of bark over the ends of the gunwales but under the outwales that held it in place. Whether these pieces were employed to protect the lashing of the gunwales and adjoining work from the weather, or whether they were the vestigial remains of a decking once used, cannot be determined. In the Canadian Northwest the ends of bark canoes were sometimes decked with bark for a short distance inboard.

The bark was secured to the gunwales by a continuous spiral lashing all along the main gunwale or by separated lashing in series. In the first, the continuous lashing, where it passed through the bark, might show regularly spaced separations to avoid the tops of the ribs. In the second, the lashings were placed clear of the ribs. There were some slight variations in the lashings, but these were of minor importance so far as structural strength is concerned. In all cases, the bark was brought up to or over the top of the gunwale before being secured, so that the holes for the lashing were pierced at some distance from the edge of the bark to prevent it from splitting.

The ends of the thwarts were mortised into the gunwales and also secured by lashings. The number of thwarts varied with the tribal type, the size, and the purpose of the canoe. Usually an odd number, from three to nine, were used, though occasional canoes had two or four thwarts. Very small canoes for hunting might have only two or three thwarts, but most canoes 14 to 20 feet long had five. Canoes intended for portaging usually had one thwart at midlength to aid in lifting the canoe for the carry position. The distance between the thwarts might be determined by structural design, or might be fixed so as to divide the cargo space to allow proper trim. The thwarts might serve as backrests for passengers, but were never used as seats. There was no standard form for the shape of the thwarts, which varied not only to some degree by tribal classification, but even among builders in single tribe. They were usually thickest and widest over the centerline of the canoe, tapering outboard and then spreading again at the gunwales to form a marked shoulder at the mortise. The lashings to the gunwales often passed through two or more holes in this shoulder.

The ribs, or frames, of most canoes were very closely spaced and were wide, flat, and thin. They ran in a single length from gunwale to gunwale. In canoes having V-sections near the ends, the ribs were often so sharply bent as to be fractured slightly. Across the bottom they were wide but above the bilge they tapered in width toward the end, which was either a rounded point or a beveled or rounded chisel-edge. The ribs were forced under the gunwales so that the heads fitted into the bevel, or into notches or holes at the underside and outboard edge of the gunwale, between it and the bark cover. By canting the rib to bring its ends into the proper position and then forcing it nearly perpendicular, the builder brought enough pressure on the bark cover to mold it to the required form. Bulging of the bark at each frame was prevented by a thin plank sheathing. The ribs in many Eastern canoes were spaced so that on the bottom they were separated only by a space equal to the width of a rib.

Each piece of sheathing, better described as a "splint" than as "planking," was commonly of irregular form. The edges were often beveled to a marked thinness. While some builders laid the sheathing edge-to-edge in the bark cover, others overlapped the edges. Nearly all builders feathered the butts and overlapped them slightly. The sheathing was held in position by a number of light temporary ribs while the permanent frames, or ribs, were being installed. It is to be noted that the sheathing was neither lashed nor pegged; it remained fixed in place only through the pressure of the bent ribs and the restraint of the bark skin.

The exact method of fitting the sheathing varied somewhat from area to area, but not in every instance from tribe to tribe. The bottom sheathing used by some eastern Indians was in two lengths. The individual pieces were tapered toward the stems and the edges butted closely together. The sides were in three lengths, but otherwise similarly fitted. The butts lapped very slightly. In a second method, used to the westward, the sheathing was laid edge-to-edge in two lengths, with the butts slightly lapped. The center members of the bottom, usually five, were parallel-sided, but the outboard ends of those at the turn of the bilges were beveled, or snied, off. The members further outboard were in one length, with both ends snied off. The bottom thus appeared as an elongated diamond-form. The topside sheathing was fitted as in the first instance.

A variation in the second style used three lengths in the centerline sheathing. In still another variation a centerline piece was laid in two lengths without taper, the next outboard piece was then cut in the shape of a broad-based triangle, and the rest were laid in two lengths, with the sides parallel to the sides of the triangular strake and with their ends snied off against the centerline pieces. In a fourth style short pieces, roughly elongate-oval in shape, were overlapped on all sides and laid irregularly so that when in place they appeared "thrown in." With this style, the midship section was laid first and secured by a temporary rib, then the next toward the ends, with the butts shoved under the ends of the middle section. The next series was similarly laid so that the top member of each butt-lap faced toward the ends of the hull and was under a rib. The ends were not cut square across, but were either blunt-pointed or rounded. Five lengths of sheathing were often used, and the widths of the individual pieces of sheathing were rarely the same, so the seams were not lined up and presented an irregular appearance in the finished canoe. The sheathing was thin enough to allow it to take the curve of the bilge easily.

If the sheathing was lapped, the overlap was always slight. In some old canoes a small space was left between the edges of the sheathing, particularly in the topsides. In some northwestern bark canoes there was no sheathing; these used a batten system somewhat like that in the Eskimo kayak, except that in the bark canoes the battens were not lashed to the ribs, being held in place only by pressure. These kayak-like bark canoes had a bottom framework formed with chine members; some had a rigid bottom frame of this type, while others had bottom frames secured only by rib pressure. The purpose of the sheathing, it should be noted, was to protect the bark cover from abrasion from the inside, to prevent the ribs from bulging the bark, and to back up the bark so as to resist impacts; but in no case, even when battens were employed, as in the Northwest, did the sheathing add to the longitudinal strength of the bark canoe. The principle of the stressed rib and clamped sheathing, which is the most marked characteristic in the construction of the North American Indian bark canoe, is fundamentally different from that used in the construction of the Eskimos' skin craft.

A wide variety of framing methods are exhibited in the construction of the ends, or stems, of bark canoes. In the temporary types of the East, the bark was trimmed to a straight, slightly "ram" form and secured by sewing over two battens, one outboard on each side. Birch-bark canoes of the East usually had an inside stem-piece bent by the lamination method to the desired profile, the heel being left unsplit; as usual, the laminations were spirally wrapped, often with basswood-bark thongs. The stem-piece was then placed between the bark of the sides, and the bark and wood were lashed together with an over-and-over stitch. Sometimes variations of the short-and-long form of stitch were used here, and some builders also placed a halved-root batten over the ends of the bark before lashing to form a stem-band as protection to the seam. In some canoes the end lashing passed through holes drilled in the stem-pieces, often with the turns alternating in some regular manner through and around the stem-piece.

The stem-pieces were generally very light, and in some canoes the head was notched and sharply bent down and inboard, so that it could be secured to the ends of the gunwales. Some tribal types had no inner stem-piece, and the stem profiles were strengthened merely by the use of two split-root or halved-sapling battens, one on each side, outside the bark and under the sewing.

Birch-bark canoes to the westward used battens under the end lashing as well as rather complicated inside stem-pieces. In some parts of the West and Northwest, the ends were formed of boards set up on edge fore-and-aft, the bark being lashed through all, with the boards projecting slightly outboard of the ends of the bark cover to form a cutwater.

To support the inside stem-piece, some form of headboard was usually fitted near each end after the sheathing was in place. These were shaped to the cross-section of the canoe so as to form bulkheads. In some canoes, these miniature bulkheads stood vertical, but in others they were curved somewhat to follow the general curve of the end-profile, and this caused them to be shaped more like a batten than a bulkhead. Bent headboards were sometimes stepped so as to rake outboard. Sometimes the form of the headboard permitted the gunwale members to be lashed to it, and often there was a notch for the main gunwale on each side.

The headboards were sometimes stepped on the unsplit heel of the stem-piece; a notch was made in the bottom of the headboard to allow this. In two types of canoe in which there was no inner stem-piece, the headboards were stepped on short keel pieces, or "frogs," fore-and-aft on the bottom and extending slightly forward of the end of the sheathing to reinforce the forefoot. The purpose of the headboard was to strengthen the stem-piece, and in many cases it was an integral member of the end structure itself and helped to maintain its form. The headboard usually served to support the gunwale ends in some manner, it stretched the bark smooth near the stems, and it secured the ends of the sheathing where support from a rib would have been most difficult to obtain. Many canoes had the space between the headboard and the stem-piece stuffed with shavings, moss, or other dry material to help mold the bark to form beyond the sheathing in the ends. Some tribal groups decorated the headboards.

In a few canoes, the stem-piece was additionally supported by a short, horizontal member stepped in the forward face of the headboard and projecting forward to bear on the after side of the stem-piece. The latter was sometimes bent back onto itself above this member to form a loop around the top of the end-profile, and the gunwale ends or a part of the gunwale structure were secured to it. This complicated bending of the stem-piece, in conjunction with use of a headboard and a brace member, served to stiffen the end structure sufficiently to meet the requirements of service.

The use of a bark cover over the gunwale ends has already been mentioned. In some eastern canoes, this was placed under the cap and outwale pieces and extended below the latter in a shallow flap on which the owner's mark or other decoration might appear; the flap was in fact a kind of name board. Such flaps do not appear on the partly decked bark canoes of the Northwest.

This general description of the structure of the bark canoes is sufficient to permit the explanation of the actual construction of a bark canoe to be more readily understood, and it also serves to illustrate the close connection between the method of construction and the formation of the lines, or model, of bark canoes. From the description, too, it can be seen that while the shape of a bark canoe was partially planned during the construction the control of every part of the model could not be maintained with the same degree of precision as in the building of an Eskimo skin boat or an Indian dugout.

Construction

One aspect of canoe construction, the Indian method of making measurements, was briefly mentioned (p. 8) under a discussion of the origin of the measurement known in French Canada as the brasse. This was the distance from finger-tip to finger-tip of the arms outstretched; in the fur trade in English times it was known as the fathom and it appears to have been about 64 inches, or less than the nautical fathom of 6 feet. Other measurements used were the greatest width of the ball of the thumb, which is very close to an English inch, and the width of the four fingers, each finger-breadth being close to three-fourths of an English inch. The length of the forearm, usually from the knuckles of the clenched hand to the elbow, was also employed by some Indians, as a convenient measurement.

Measurements in these units might be memorized and used in building, but many Indians used measuring sticks, and these served as "foot-rules." They were sometimes squared and were painted as well as notched.

A Malecite Indian, interviewed in 1925, had three such sticks for canoe building. One, for the length of the gunwale frame, was half the total length required; it was notched to show the distance at which the ends of the gunwales were lashed and also the position of the thwarts. Such a stick would be about 7 feet long for a 16-foot canoe, 8 feet for an 18-foot canoe. The second stick was notched to show half the length of each of the thwarts. The third stick had notches showing the height of the gunwale at each thwart and at the end, four notches in all for the half-length of the canoe. This stick measured from the surface of the building bed, not from a regular base line.

The method of measuring canoes appears to have been fairly well standardized, at least in historical times. As stated earlier, length was commonly taken over the gunwales only, and did not include the end profiles, which might extend up to a foot or slightly more beyond the gunwale ends, bow and stern. However, in certain old records the overall length is given, and in various areas other methods of measurement existed. Where a building frame was used, the given length of the canoe was the length of this frame; usually this approximated the length of the gunwales. The width of a canoe was measured by the Indian from inside to inside of the main gunwale members. The extreme beam might be only 2 or 3 inches greater than the inside measurement of the gunwales, but if the sides bulged out, the beam might actually be 6 or more inches greater. The depth was usually measured from the inside of the ribs to the top of the gunwale but in building it was measured from the surface of the building bed to the bottom of the main gunwales, as noted above in the description of the measuring sticks.

Thus it will be seen that the Indian measurements constituted a statement of dimensions primarily useful to the builder, for their main purpose was to fix the proportions rather than establish the actual length, width, and depth. Today we state the length of a canoe in terms of extreme overall measurement; the Indian was inclined to state the length in building terms, giving dimensions applicable to the woodwork only, just as the old-time shipbuilder gave the keel length of a vessel instead of the overall length on deck.

The building site was carefully selected. The space in which the canoe was to be set up had to be smooth, free of stones and roots or anything that might damage the bark, and the soil had to be such that stakes driven into it would stand firmly. A shady place was preferred, as the bark would not dry there as fast as in sunlight. Since the construction of a canoe required both time and the aid of the whole Indian family, the site had to be close to a suitable place for camping, where food and water could be obtained. It is not surprising, therefore, to find canoe building sites that apparently had been used by generations of Indians.

The preparation of the building bed was controlled by the intended form of the canoe to be built. If the bottom of the canoe was to be rockered, the cleared ground was brought to a flat surface for the length required for setting up the canoe. If the rocker was to be great, the middle of the bed would be slightly depressed. If the bottom was to be straight fore-and-aft, or very nearly so, the bed was crowned from 1½ to 2 inches higher in the middle than at the ends, so that the canoe was first set up with a hogged bottom. Very large canoes such as were used in the fur trade required as much as 4 inches crown in the building bed. Other dimensions being equal, the amount of crown was usually somewhat greater in canoes having bulging sides than in ones having more upright or flaring sides. Canoe factories such as were operated in certain fur-trading posts sometimes had a plank building bed suitably crowned and drilled for setting the stakes.

Two methods of setting up the canoe were used. In most of the eastern area, the gunwales were put together and used to establish the plan outline of the canoe on the building bed. But a building frame was used for constructing the various narrow-bottom canoes having flaring sides, and for some other tribal forms. The frame, made in the same general form as the gunwales when assembled, but less wide and sometimes much shorter, could be taken apart easily, allowing it to be removed after the canoe was built; hence it could be used to build as many canoes as desired to the same dimensions as the first, and was retained by the builder as a tool, or pattern, for future use.

The method of construction in which gunwales only were used in setting up the canoe will be explained first in order to show the general technique of construction. Use of the building frame will then be described. Important deviations from these methods will be described in later chapters under the individual tribal types in which they occur.

The Malecite canoe, a straight-bottomed craft about 19 feet long and 36 inches beam, is used as the example, hence the method of building to be described is that generally employed in the East, where variations in construction mainly involve the use or omission of structural elements.

The gunwales are the first members to be formed. In the Malecite canoe these are the inner gunwales, as the canoe will have outwales and caps. The gunwales are split from white cedar to produce battens that will square 1½ inches when shaped. The gunwales are tapered each way from midlength, where they are 1½ inches square, to a point 3 inches short of the ends, where they are ¾ by 1 to 1¼ inches. The edges of the gunwales are all rounded, and the outboard bottom edge is beveled almost ½ inch, at 45° to the bottom of the member. The last 3 inches at each end is formed like half a blunt arrowhead, as shown in the sketch of the member on page 31. The gunwales will be bent, side to side, on the flat as far as the ends are concerned, so the blunt arrowhead is formed on one of the wide faces of the ends as shown. The arrowhead form allows a neat joint when the gunwale ends are brought together, pegged athwartships, and then wrapped with a root lashing. In forming and finishing the gunwales, a good deal of care is required to get them to bend alike, so that the centerline of the finished frame will be straight and true.

To take the ends of the middle thwart, a mortise ¼ by 2 inches is cut in each gunwale member athwartships at exactly midlength, the length of the mortise being with the run of the gunwale. In it, the middle thwart, 33 inches long, is fitted. Made of a ⅞-inch by 3-inch piece of hard maple, the thwart tapers slightly in thickness each way from its center to within 5 inches of the shoulders, which are 30 inches apart. The thickness at a point 5 inches from the shoulder is ¾ inch; from there the taper is quick to the shoulder, which is ⁵⁄₁₆ inch thick, with a drop to ¼ inch in the tenon. The width, 3 inches at the center, decreases in a graceful curve to within 5 inches of the shoulder, where it is 2 inches, then increases to about 3 inches at the shoulder. The width of the tenon is, of course, 2 inches, to fit the mortise hole in the gunwale. The edges of the outer 5 inches of the thwart are rounded off or beveled a good deal; inboard they are only slightly rounded.

The thwart is carefully fitted to the gunwale members and the ends are pegged. Some builders wedged the ends of this thwart from outside the gunwales, the wedge standing vertical in the thwart so that the gunwale would not split; however, it is not certain that wedging was used in prehistoric times, although it is seen in some existing old canoes. The pegs used in this canoe are driven from above, into holes bored through the gunwale and the tenon of the thwart to lock all firmly together. Three holes are then bored in the broad shoulders of the thwart about 1½ inches inboard of gunwale for the root lashing that is also used.

The ends of the gunwale members are now brought together, and to avoid an unfair curve appearing at the thwart in place, short pieces of split plank or of sapling, notched to hold them in place, are inserted between the gunwale members as temporary thwarts at points about 5 feet on each side of the middle thwart. After the ends are brought together and the final fitting is carried out, a peg is driven athwartships the ends and a single-part root lashing is carefully wrapped around the assembly.

Some canoe builders omitted the blunted half-arrowhead form at the gunwale end. Instead, the inside faces were tapered to allow the two parts to bear on one another for some distance. The gunwales were then pinched together and lashed with one or more wrappings. Finally, a thin wedge was sometimes driven from inboard between the two gunwale ends to tighten the wrappings. The wedges were usually so carefully fitted as to be difficult to identify. It is probable that this wedged gunwale ending represents the prehistoric form, and the blunted half-arrowhead ending is a result of the use of steel tools.

After the ends of the gunwales have been securely fastened together, the first pair of permanent thwarts is fitted. These are located 36 inches, center to center, on each side of the middle thwart, a distance that determines the centers of the mortises in each gunwale member. Each thwart, made from a ¾-inch by 3-inch piece, tapers smoothly in thickness from the ¾-inch center to the ⁵⁄₁₆-inch shoulder. The tenon is of the same dimensions as that of the middle thwart, the width takes the same form as that of the middle thwart, and the edges are similarly beveled and rounded. The distance between the shoulders, taken along the centerline, is 22½ inches, and the centerline length of the thwart 25½ inches. However, the shoulders and ends of the tenons must be bevelled to follow the curve of the gunwales hence the extreme length of the thwart is actually very close to 26 inches. The worker determines the bevel of the shoulders by fitting the thwart to the run of the gunwales, the temporary thwarts being shifted so that the distance between the gunwales equals that set by the measuring stick. These two thwarts having been fitted, the tenons are pegged as before, but in the shoulders only one lashing hole is bored instead of the three employed in the middle thwart.