Volume 44 1935 > Volume 44, No. 174 > The chronology of Australian watercraft, by D. S. Davidson, p 69-84
                                                                                             Previous | Next   

- 69
Continued from Vol. 44, No. 1 Page 16.

3. Undercrossed-type. 1 In this method, straight sticks in pairs are crossed and the end of the boom is placed on the fork made by the crossing (fig. 4).

FIG. 4.
Single-outrigger Canoe with “Undercrossed”-method of Attachment. After Roth. Central coast of Queensland.

This type is found, apparently, from the Batavia river on the Carpentaria coast to the region of cape Direction on the east coast of the Cape York peninsula. It is extensively used throughout the contiguous islands of Torres strait and the neighbouring south-eastern coast of New Guinea, thence on to the eastern and northern coasts and to Melanesia, Micronesia, and Polynesia. This method is also employed in the Andaman islands, on the extreme north-western periphery of outrigger use. These appearances are all peripheral to the centralized distribution of the more complex Halmaheran method which seems to be confined to eastern Indonesia. In so far as we are concerned it is important to note that the Australian use of the stick-type is contiguous to its appearance in Torres strait and New Guinea, thus indicating that it is the most recent method to invade Australia, granted that cape York or the immediate vicinity was the point of entry for all three methods.


The direct-lashed-method is associated with the so-called Claremont type of double-outrigger which in 1910 was found between Night island and Claremont point and which is now distributed as far south as port Stewart, Princess Charlotte bay. This type of attachment is common in Polynesia, Indonesia, and Ceylon, but apparently lacking - 70 in New Guinea and the neighbouring Melanesian islands. The nearest appearance to Australia seems to be Nissan in the northern Solomon islands.

The occurrence of this method in Australia is difficult to explain. It is reasonable to suppose that it has not been derived from any of the localities of the present users of this type, hence it is likely that it once was used in New Guinea and has been displaced by the methods now found there. This conclusion is consistent with the supposition that in being a marginal type it is relatively old and that its present distribution has been reached by diffusion from a centralized region. Haddon also believes that it has been derived from New Guinea but considers it to be the most recent type in Australia, a conclusion which is difficult to accept. 2


The undercrossed-method of attachment is found in Australia with the single-outrigger only from Flinders island southward. It is employed in south-eastern, eastern, and northern New Guinea and in the far away Andaman islands. It seems quite obvious that it is an old method and there can be no doubt, as Haddon suggests, that the Australian appearance has been derived from New Guinea.


To summarize the material on the methods of attaching the booms to the float or floats: it seems quite certain that the geographical sequence along the coast of the Cape York peninsula from south to north represents the order in which these methods were introduced into Australia. There can be no doubt that the stick-and the undercrossed-methods came directly from New Guinea by way of cape York but, as we have seen, there is no evidence to indicate that the direct-lashed-type diffused over the same route, although such is probable, unless it is to be supposed that the appearance is the result of a local invention, which is decidedly unlikely, or the direct introduction from overseas to the cape Sidmouth region, a conclusion which seems hardly plausible. In view of the contiguous distribution of outriggers in Australia and New Guinea, the - 71 marginal appearances in other regions of the methods of attachment used in Australia and the probability that Torres Strait has been the only avenue over which these influences have come, it seems reasonable to suppose that the geographical sequence in this case is indicative of the relative ages of the three types in Australia. This conclusion is also supported by the little historical information available, which shows that diffusion has been southward along the coast of Queensland and away from cape York and Torres strait. The Australian evidence also demonstrates that the single-outrigger diffused earlier than the double-type. The finding of this chronology in this particular region, however, should be regarded as a purely local circumstance which has no relation to the possible chronological relations between the two traits in other parts of the world. The historical movements, as accurately as we can judge from the documentary evidence, are indicated in fig. 5.

FIG. 5.

The dugout (fig. 6) from the point of view of the historical development of watercraft, is undoubtedly older than the use of outriggers, as already mentioned, for it

FIG. 6.
Dugout Canoe, Drysdale river. From a specimen in the Western Australian Museum, Perth, W.A.
- 72

is the basic trait to which the outrigger-feature has been attached. At the unknown point of origin of the outrigger, therefore, the chronological relationship between the two is obvious. Once the outrigger became a permanent part of the craft, however, the older and younger traits could diffuse together as a unit and could invade new regions conjunctively. In the meantime it seems logical to suppose that the dugout, by itself, by virtue of its greater age, had already diffused into a more or less considerable area. Theoretically then, if the rates of diffusion have been the same, we should expect to find the dugout at all times in areas marginal to the distribution of the use of the outrigger. There are inherent qualities in the two, however, which preclude the possibility that the rates of diffusion would be the same or that the distributions in any particular region should stand in any definite relationship to each other. In the first place the ordinary crude dugout is not a seagoing craft in the usual sense of the term, hence its distribution must be confined to regions which are not separated by great expanses of ocean. For this reason it could not have diffused rapidly except on inland streams or along sheltered coasts. Only in exceptionally good weather or by accident could it cross great distances of water. The outrigger-canoe on the other hand, because of its greater seaworthiness, is not hindered from diffusing over stretches of water which might prohibit the use of a dugout and, therefore, it might easily reach many regions which the dugout could not invade. It is not surprising, as a consequence, to find that the distributions of the two types are not arranged in any consistent order. By the very nature of the two, we should expect to find their distributions irregular and inconsistent.

Now in so far as Australia is concerned we have seen that the dugout with an outrigger entered the continent by way of Torres strait and the Cape York peninsula. In view of the chain of islands between New Guinea, the East Indies and the Asiatic mainland this would seem to be the only logical route by which primitive man could have reached Australia. It is over such a route, therefore, that we should expect the dugout to have diffused at a previous time if it were a part of the equipment of the invading Australians or if it had been introduced at a - 73 relatively later time in a manner similar to that of the outrigger.

The known use of the ordinary dugout in Australia is restricted to the northern coast west of the Sir Edward Pellew islands (fig. 7) and, as we shall see, is the result

FIG. 7.

of Malay influence. Apparently it is not found to the east of the region mentioned, at least no reports of its presence have come from the southern or eastern shores of the gulf of Carpentaria, the Cape York peninsula or the eastern coast of Queensland. There is no evidence, therefore, to support the supposition that the modern appearances could possibly have been derived from New Guinea by way of Torres strait. There are a few reports, however, which can be interpreted to imply the use of dugouts in eastern New South Wales and south-eastern Queensland and, if they are authentic, there is a possibility that these southern appearances may have been derived from New Guinea at an early time and that this type of craft subsequently became abandoned on the coast of Queensland as the more - 74 advanced types of craft, such as the bark-canoes, or the outrigger-canoes, made their appearance. The alternative conclusion would be that they represent an indigenous development in eastern Australia, for it does not appear possible to link these local appearances with the dugout now found on the northern coast of the continent.

The evidence for the supposition that dugouts may have been used in northern New South Wales and the neighbouring region of Queensland is based upon a very few reports which are general in scope, ambiguous in description, or based, in all probability, on hearsay. Angas, for instance, in speaking of eastern New South Wales in general, remarks on the use of the bark-canoe in this region and adds:

“Towards the north the natives have canoes of a more substantial kind, formed out of the trunks of trees, and about twelve or fourteen feet long: they are hollowed-out by fire and shaped with the mogo, or stone hatchet.” 3

Now, since we know that bark-canoes are used as far north as the southern periphery of the outrigger, it is quite possible that this statement refers to the outrigger-canoe in northern Queensland. The report does not indicate any specific locality and it would seem more reasonable to interpret it as referring to a known type north of the use of the bark-canoe than to an unknown appearance towards the north.

In the Bunya mountains near Brisbane, Leichardt 4 reports that “They make little canoes of the stringy-bark tree” but he does not specify whether a dugout or a bark-canoe is meant. Apparently the conclusion rests upon the interpretation of the word “tree.” The stringy-bark tree, however, furnishes the bark for a large proportion of the bark-canoes made in this part of Australia, so it seems more likely that a bark-canoe is implied than a dugout.

A no more satisfactory inference of the use of a dugout in the Blue mountains, New South Wales, near Sydney, has been made by Bennett 5 whose work is not - 75 available to me. Thomas, 6 however, in referring to both Bennettt and Angas says that

“… the absence of detail suggests that both authorities may be relying on hearsay evidence.”

Thomas also states, on the authority of a Mr. Thomas Hardy that a dugout was used on the coast of New South Wales in the region of Richmond. A dugout was recently dredged out of the mud near port Stephens but Enright believes it to have been patterned after the A.A. Company boats or made by Maori sailors who have resided there.

It is most unfortunate that we cannot reach a satisfactory conclusion on this matter. It is too late to secure additional evidence of an ethnological character, and therefore, unless archaeological investigation may unearth the remains of additional dugouts, the question as to the validity of the reports may never be settled.


The modern use of the dugout canoe in northern Australia is definitely known to be the result of Malay influence. These people for at least a century and a half, and for probably an unknown but considerable period of time before that, have visited the coast of Arnhem Land, and undoubtedly other points of the Northern Territory coast, in search of trepang and other commodities for trading in the East Indies. According to Warner 7 the Malays who came to Arnhem Land sailed in double-out-riggers and even taught the art of manufacturing this type of craft to some of the Australians. The latter, however, accustomed to the navigating of sewn bark-canoes, were apparently not impressed by the outrigger-attachments, for they have not adopted them. They were interested in the dugout itself as a substitute for their bark-canoe, and the Malays were quick to realize that an opportunity for a profitable trade had been created. Dugouts were subsequently imported for sale to the natives and this trade flourished for many decades until finally broken up by the Australian government.

- 76

During the period of Malay contacts the natives of Arnhem Land seem to have relied almost entirely upon the invaders for their dugouts for, according to testimony secured by Warner from old natives who were alive at the time the trade was forcibly stopped, once the Malay had departed they were so ignorant of the art of manufacture that they were obliged to return to their bark-canoes. Had it not been for the aborigines of the English Company islands, who possessed the knowledge of dugout construction, this form of craft would have become at least temporarily obsolete in this locality. We cannot assume that it would have become permanently obsolete, however, for it is possible if not probable that the natives further west were building their dugouts at that time and diffusion might well have reintroduced them into Arnhem Land.

In addition to the dugout the Malays were responsible for the introduction of the mast and the pandanus sail, traits which have become integral accessories of both the dugout and sewn bark-canoes in Arnhem Land.

At the present time the dugout is found contiguously distributed from the Pellew islands 8 in the gulf of Carpentaria on the east to the northern shore of the Prince Regent river in the western Kimberley district of Western Australia 9 (see map, fig. 7). Tindale 10 has recorded the evidence of the Malay visits to Groote island and it is possible that they were directly responsible for the appearance of the dugout in the Sir Edward Pellew group. On the other hand, diffusion from the mainland may be the cause of its presence there for Spencer implies the use of only the bark-canoe in this region at the beginning of the twentieth century. 11 At any rate the Pellew islands appear to be the most easterly place from which the dugout has been reported. It is interesting to note, therefore, that in this direction the dugout has not moved very far, although peripheryward, from its point of introduction. To the westward, however, diffusion has been instrumental in causing the spread of the dugout along many hundred miles - 77 of coastal country. According to Stokes, who made a detailed survey of the western and north-western coasts of Australia in the years 1837-1843:

“Upon all this extent of coast, we saw no single instance of the use, or even existence, of any proa, or canoe; and my opinion, strengthened by personal experience, and enforced by the authority of the most recent navigators, is, that the canoe is not used upon the north-west coast. The negative evidence, at least, is strongly in favour of this presumption, for, while we saw the canoe in use at Clarence Strait,—the western boundary of the northern coast,—we saw nothing but the raft to the south of that point. 12

In less than a century, therefore, the dugout has diffused westward for a distance of over six hundred miles. In 1917 its boundary was the Prince Regent river, according to Love, 13 and it seems that it has not passed that point at the present time, although I have been informed by Mr. Laves 14 that the natives of Sunday island, King sound, about 100 miles to the south, know of its use. It will be interesting to learn whether this southern and western trend of diffusion will bring this type of craft into the King sound region within the next few years. The change in watercraft types which has taken place on the northern coast in the last century is also indicated by other writers. In 1818 King 15 saw a bark-canoe at Knocker bay, port Essington, but by 1834 Campbell 16 found there only dugouts “like those of the Malays.” He was under the impression that they had been left by the latter or that they had been stolen from them by the Australians. It must not be assumed that the dugout was introduced into Australia between 1818 and 1834, for, as already stated, it is quite certain that the Malays have visited these shores for centuries. It may be that this new type of craft had been adopted at a much earlier date by the natives of eastern Arnhem Land, although Flinders makes no mention of its use by natives in 1803. The acceptance of the dugout in the Port Essington district, however, was apparently assuming an intense form in the period - 78 mentioned, for King seems to have been the last one to have seen a bark-canoe in this locality, subsequent visitors such as Macgillivray 17 (1852) and Foelsche 18 (1881) finding only the dugout form. The dugout seems to be becoming more and more popular on Melville island where possibly it may eventually displace the bark-canoe. On Bathhurst island, however, the bark-canoe is still the prevalent form although the dugout is making inroads there also. 19

East of port Essington, the dugout is almost the universal type of craft. As early as 1818 King 20 saw it in use at Goulburn island. It is found in the Wessel islands and among all the peoples and islands on the east coast of Arnhem Land as far south as the Sir Edward Pellew islands.

Although several descriptions of the method of making the dugout-canoes are available, they are all more or less brief with the exception of Tindale's detailed report for Groote island. 21

The history of the dugout in Australia may be considered a good example of how a foreign trait may be diffused once it has been introduced into a new area. It is unfortunate that we do not have a greater number of records of the limits of the dugout taken at different time intervals at various places along the northern coast, for it would be interesting to know whether diffusion has been gradual and constant in certain areas, or during certain periods of time, and hasty in other places, or at different times. The most important value of the early reports, however, is their information that other types of craft, which have now disappeared from use in many localities, were formerly common in the regions now monopolized by the dugout. With such data we have a check on the conclusions which we may see fit to adopt from theoretical points of view. Mention has been made in passing of the former presence of both the bark-canoe and raft on the northern coast at the time the dugout was becoming popular in the Clarence strait region. Unless the dugout canoe - 79 has been instrumental in completely annihilating them within its present distribution, we should expect to find them scattered in non-contiguous regions within the distribution of the dugout, as well as, perhaps, in areas which are peripheral to the present use of the latter. In other words, knowing the historical sequence of watercraft in this area, we have a means of testing the validity of the theory that the relative distributions should indicate the chronology of the traits involved. For our first consideration let us turn to the bark-canoe. Bark-canoes in Australia can be classified into three main types: (1) the simple bark-canoe, (2) the tied bark-canoe, and (3) the sewn bark-canoe.


The simple bark-canoe, without sewing or stitching of any kind, is illustrated in fig. 8. It is found only in

FIG. 8.
Simple Bark-canoe. Victoria and Murray river.

western Victoria, south-eastern South Australia, and the Murray-Darling basin of New South Wales. Indications of its use appear at Avoca, Darling river and Riverina region, Goolwa, Murray river, southern coast, interior of New South Wales, Encounter bay, lake Alexandrina and Yass 22 (see fig. 9). As can readily be seen it is most primitive, but nevertheless an ingenuous type of watercraft. A large sheet of bark is stripped off a tree selected for its shape so that the natural curves determine the contours of the canoe. The ends, when necessary, are filled with clay or mud, and the same material is usually used for a hearth in the bottom of the craft. No sewing or stitching of any description is found in this type, nor are stretchers,

- 80
FIG. 9.

ties, ribs, or reinforced gunwales, the natural shape of the bark selected being relied upon for the maintenance of the shape of the canoe. Even at the present time hundreds of trees which have furnished bark for these canoes are to be seen along the rivers of the region indicated.


The bark-canoe with purse-like, gathered ends, secured by wrapping and tying with bark-strands or cord, is another type of primitive craft found in the south-eastern part of Australia (fig. 10). It occupies a contiguous distribution from the northern coast of New South Wales to the

FIG. 10.
Tied Bark-canoe. East coast of south-east Australia.
- 81

Gippsland lakes area of eastern Victoria 23 (see fig. 9). There is no information to indicate how much overlapping there is between this distribution and that of the ordinary bark-canoe but it seems likely that the border peoples may have been acquainted with the use of both.

It seems obvious that this type of craft is a decided improvement over the simple bark-canoe. As the result of the sides being forcibly held up by artificial means at the bow and stern, a much greater depth can be attained which, in turn, contributes to an increased seaworthiness. Other features in construction which are of an advanced nature include the use of stretchers to maintain the spread of the bark at a minimum beam, the application of ties to prevent any widening of the beam, the insertion of ribs to strengthen the hull, and the reinforcing of the gunwales by lacing a band of rushes along the edge of the bark hull. Not all of these features are to be found in all the canoes of this type, and sufficient information is not available to indicate the relative use of each. It is probable, however, that the reinforcing of the gunwales was not extensively practised in Victoria, if at all; indeed, many of the Victorian canoes of this type do not show the use of ties or ribs. These features assume more and more importance as the northern coast of Australia is approached. Whereas they are inconsistently used and but crudely fashioned to the primitive canoes in the south, in the northern areas they are quite institutionalized and appear in a greater degree of refinement in association with the most advanced types of bark-canoe. Such a progression should not be unexpected, for probability favours the theory that these elements in construction have diffused from north to south.

In spite of the advanced traits associated with the tied bark-canoe it is still a very primitive type of craft. A good description of the manufacture of this type is given by Howitt. 24

- 82
FIG. 11.
Sewn Bark-canoe. Melville island type. After Basedow.

The sewn bark-canoe (fig. 11) has been reported from many non-contiguous areas between the western coast of the Northern Territory and the southern coast of Queensland. For convenience these appearances may be arranged into three groups: (1) the Northern Territory, (2) the gulf of Carpentaria coast of Queensland, and (3) the east coast of Queensland.


The earliest report of the sewn bark-canoe for this region comes from the celebrated explorer, Captain Matthew Flinders, who found this type of craft at the Sir Edward Pellew islands at the time he discovered them. 25 He described it as being clinker-built, with gunwales of mangrove poles lashed to the bark hulls, obliquely arranged wooden struts and a series of ties to maintain the spread of the bark, and short wooden wedges placed in the bow and stern for the same purpose. On the floor were flat pieces of sandstone which served as a hearth. At Blue Mud bay, on the east coast of Arnhem Land, Flinders saw a somewhat similar canoe made of two pieces of bark which were sewn together lengthwise so that the seam ran along one side of the canoe. The ends were sewn and caulked with gum. Five ties of vine are mentioned as are also gunwales made of poles lashed to the bark. This craft was capable of carrying six people.

- 83

Most of the canoes of this region, however, have a hull made of one piece of bark to which small additions may be applied. It has been mentioned before that King, in 1818, found a one-piece bark-canoe at Knocker bay, port Essington. 26 Short pieces of bark were placed crosswise in the bottom, probably to help maintain the spread of the craft, but possibly just as a floor-protection for the bottom of the canoe. The gunwales were of poles as in the canoes already mentioned. The craft was 18 feet long and could hold eight people. A somewhat similar canoe from Darwin is in the Australian Museum, Sydney. It has three ribs, bark-fibre rope-ties, pole-gunwales but no crossbeams. Examples in the South Australian Museum, Adelaide, include one with fifteen ribs and one without the polegunwales. They are from the coast of the Northern Territory.

At about the beginning of the twentieth century Spencer 27 visited the Macarthur and Roper rivers on the east coast of Arnhem Land and reported the sewn bark-canoe as a characteristic trait of all the northern coast between the Sir Edward Pellew islands, where it was still in use, and Melville island. Although Spencer himself did not visit the entire coastline, it is quite likely that his statement is true if accepted in a general sense. The canoe which he saw at Borraloola was composed of three pieces of bark sewn together lengthwise for the hull, to which small pieces of bark were added at each end to raise the bow and stern. Poles were lashed to the gunwales and a series of nine ties was employed to retain the shape of the craft. Three sticks were arranged at each tie to help support the shape, one running from side to side just under the tie, the other two were braced against the upper stick, one against each end, and crossed to the floor of the canoe where they met and held in place pieces of bark which formed a floor and served after the fashion of ribs.

At Bathurst island and Melville island, Basedow 28 found similar canoes in use, the only differences being a - 84 fishtail-like stern, a concave bow and the custom of sewing together the tops of the side walls for a slight distance from the bow and the stern (see fig. 11).


For the gulf of Carpentaria coast of Queensland, Roth 29 reports that the sewn bark-canoe prevailed about the beginning of this century in the region between the Batavia and Archer rivers. Some of the canoes used here lacked pole-gunwales and ribs according to Roth's description, and a specimen in the Australian Museum, Sydney, but others apparently incorporated these features, as is shown in one of Roth's plates. The former example is reminiscent of canoes of Gippsland, except that they are sewn at the extremities and not wrapped and tied. Two stick-stretchers and two or three ties are usually employed. The latter are held taut by the use of two forked sticks which cross and whose bases are braced against pieces of bark on the floor of the canoe. This practice is somewhat similar to that noticed in some of the Northern Territory craft. There are minor details of construction which distinguish the Gulf coast canoes from those on the east coast of Queensland and those of Arnhem Land.


On the east coast of Queensland, the single-piece type of sewn bark-canoe is reported by Roth 30 for the area from the Johnstone river to just south of Cardwell, where King found them as early as 1819 31 (fig. 12). In general they are similar to those already described, but differ, of course, in details. Gunwales are made of poles or withies

FIG. 12.
Sewn Bark-canoes propelled by small pieces of bark. Gould island, 1819. After King.

(To be continued.)

1   Dixon, pp. 94-96; Haddon, 1920, p. 126.
2   Haddon, 1913, p. 631.
3   Angas, 1847, 2, p. 230.
4   In Lang, p. 375.
5   Bennett, 1, p. 115, cited by Thomas, 1905, pp. 66, 72.
6   Thomas, 1905, p. 72.
7   Warner, pp. 482-483; see also Stokes, 1, p. 394; Curr, 1, p. 273; Spencer and Gillen, 1904, p. 630; Basedow, 1907, p. 53; Spencer, 1928, 2, p. 569; Tindale, 1926, pp. 130-132.
8   Tindale, verbal information.
9   Love, pp. 32-33.
10   Tindale, 1926, pp. 130-132, and verbal information.
11   Spencer, 1928, 2, pp. 569-570.
12   Stokes, 1, pp. 89-90.
13   Love, pp. 32-33.
14   Laves, correspondence.
15   King, 1, p. 90.
16   Campbell, p. 170, quoted by Basedow, 1913, p. 305.
17   Macgillivray, 1, pp. 146-147.
18   Foelsche, p. 12.
19   Basedow, 1913, pp. 303-305; Spencer, 1914, pp. 397-400.
20   King, 1, p. 67.
21   56Tindale, 1926, pp. 103-112.
22   South Australia Museum, Victoria Museum, Brough Smyth, 1, pp. 408-410; Woods, pp. 41, 193; Angas, 1846, plate 2, no. 14; plate 9, no. 18; Newland, p. 5; Flanagan, p. 58; Mitchell, 1, p. 223; Thomas, 1905, p. 57, cites other sources.
23   Peron, plate 23; Brough Smyth, pp. 408, 410, 416, 417; Howitt, p. 424; King, 1, p. 175; Tench, p. 81; Thomas, 1905, p. 57, cites other sources.
24   Howitt, A. W., in Brough Smyth, 1, pp. 408-410.
25   Flinders, 2, p. 171.
26   Flinders, 2, p. 198.
27   Spencer, 1928. 2, pp. 569-570.
28   Basedow, 1913, pp. 303-306.
29   Roth, W. E., 1908. 1910, p. 6.
30   Roth, W. E., 1908, plate 1, fig. 2, 1910, p. 6.
31   King, 1, p. 198.