- 257

THE netting and basketry techniques employed by the Australians are characterized by simplicity and, relatively speaking, by a lack of variation. Compared with neighbouring Melanesia, where both netting and basketry techniques are found in a much greater variety, and with Malaysia, where basketry is exceptionally well developed, these facts are significant, for they serve to illustrate another instance in which Australia, in its peripheral location, is marked by general backwardness.

Netting techniques may be classified into two main groups, (1) Knotted Netting and (2) Knotless Netting. Both are employed in the manufacture of carrying or dilly-bags and in the construction of game- and fish-nets. Occasionally they may be found in skull-caps or hair-nets as the impressions in discarded clay widows'-caps reveal. In the larger nets used for kangaroo or emu the knotted variety of net seems to be always used. This is as one might expect, however, for since these nets may be over 100 feet long 2 and as much as 6 feet wide, it seems obvious that a secured meshwork is imperative for practical uses.

(1) Knotted netting, as made in Australia, is identical with that of modern Europeans. Its antiquity in Europe goes back to the Neolithic age and its wide distribution in almost all parts of the world indicates that its use may have been known at an even earlier time. It is common to New Guinea and Melanesia, as well as to other Pacific areas. In Australia its distribution is the most widespread of any netting technique or basketry type. The known - 258 occurrences are shown in figure 1. 3 There can be little doubt but that if our information were more extensive this type would be found throughout most of the eastern and northern portions of the continent. In south-western Australia, however, there are no indications that it ever has been used in the region roughly south and west of a

1. M—Normanton, Mornington Is. (dugong net); A—Bags: Barrow Point, Bathurst Head, Flinders Is., Port Stewart, Cape Bedford; S—Staaten River, Princess Charlotte Bay (fish net), Burketown (fish net), Normanton (fish net), Cairns (fish net), Atherton (large kangaroo net), Port Douglass (net cap), Wellesley Is.; B—Cardwell, Mornington Is., Normanton, Mitchell River., Roth, W. E., North Queensland Ethnography, Brisbane, Bull. 7, p. 28, Mitchell to Flinders River and across peninsula to east coast, Princess Charlotte Bay, Palmer River., 2. S—Whitsunday Is., Fitzroy River, Great Keppel Is., Rockhampton (fish net)., Roth, op cit., Bull. 7, p. 28, Gladstone to Broadsound; Bull. 1, pl. 13 and p. 14, Keppel Is., Rockhampton., 3. S—Bags: Bellinger River, Richmond River, Roth, Bull. 7, p. 29, Brisbane, Morton Bay., Enright, W. J., Mankind, vol. 1, no. 6, Sydney, 1932.

- 259

4. M—Woewurong, Gippsland, Lake Tyers (fish net); A—Nets: Point McLeay, Darling River; S—Bags: Lachlan and Darling Rivers, Louth and Moorarar Holding (widow's caps)., Cawthorne, W. A., Rough Notes on The Manners and Customs of the Natives [1844], Adelaide, 1926., Woods, J. D., The Native Tribes of South Australia, p. 41, Narrinyerri; p. 193, Encounter Bay, Brough Smyth, Aborigines of Victoria, I, p. 390, Yarra., 5. M—Arunta, South Arunta (wallaby net), Worgaia (fish net); A—Bags: Coopers Cr., West Queensland; A—Nets: Kopperamanna, Coopers Cr., Frew River, Lake Eye, Georgina Dist., Birdsville, Warburton River; S—Cloncurry (fish net), Glenormiston (fish net), Herbert Downs (net cap); Head bands: Boulia, Georgina, North Gregory Dist.; B—Coopers Cr., Horn and Aiston, p. 62, Wonkonguru (2′-6′ x 30′-90′)., Brough Smyth, I, fig. 225, Burdekin River., 6. M—Anula (fish net), Gnanji (fish net), Anula (bag); A—Bags: Darwin, Alligator Rivers, Melville Is., Port Essington; A—Nets: McArthur River, Newcastle Waters, Borroloola; S—Alligator River (bags and nets); Bags: Port Essington., 7. P—Beagle Bay., 8. M—Ngurla, Mindaroo; P—Roebourne., 9. New Guinea, South Collingswood Bay, Avid River (Papuan Gulf), Bartel Bay, Goodenough Bay, D'Entrecasteau Is., Fly River.

line drawn from North-west Cape to Spencer Gulf. It is also lacking in Tasmania. Such a distribution would seem to point to New Guinea as its place of derivation in so far as Australia is concerned.

(2) Knotless netting, a half-hitching technique, commonly called “coiling without a foundation,” is found in Australia in three varieties, (A) The Simple Loop, (B) The Loop and Twist, and (C) The Double Loop or Hourglass pattern.

A. Simple Loop. The simple loop, as the term implies, is the most simple of the three in appearance as well as in the procedure of its construction. It consists of a continuous series of loose half-hitches taken at definite intervals on a basal string for the first row, the subsequent rows being added in the same fashion by a half-hitch in each pendant loop of the preceding row (see fig. 2). This

FIG. 2.

Simple Loop.

- 260

1. S—Rockhampton, Whitsunday Is.; A—Batavia River., Roth, Bull. 1, p. 13, Morehead River, Musgrave River, Rockhampton; Bull. 7, p. 28, Whitsunday Is., Gladstone to Broadsound, Rockhampton to Yeppoon., 2. S—Clarence River, Richmond River, Bellinger River., Roth, Bull. 7, p. 29, Morton Bay and Brisbane., 3. M—Victoria, Lake Tyres (hand net for fish bait), Woewurong., Brough Smyth, 1, pp. 343-344, Yarra and Goulburn tribes (already obsolete in his time), p. 381, Lake Tyers (fish net)., 4. M—Arunta, Lake Frome; S—Ooldea; Widow's caps: Morarra, Louth (N.S.W.); A—Pituri bags: Coopers Cr.; Bags: Coopers Cr, Lake Eyre, Warburton River., Spencer, B., Wanderings in Wild Australia, 2 vols., London, 1928, vol. 1, p. 223, South Arunta., Eylmann, L. K., Die Eingeborenen der Kolonie Sud Australien, 1908, p. 383, Urabunna., 5. S—Camooweal, Boulia; Net caps: Boulia, Cloncurry, Georgina River, Herbert Downs; B—Kilkummin (headdress net)., Roth, Bull. 1, p. 13, Boulia; Bull. 7, p. 28, Boulia, Upper Georgina, Camooweal (head net or hair cap)., 6. M—McArthur River; S—Alligator River, Roper River; A—Point Charles, Alligator River, Daly Waters, Darwin, Port Essington, Daly River, Port Keats, Liverpool River, Melville Is., Newcastle Waters, McArthur River, Roper River, Katherine River, Tennant Cr., Borroloola., Basedow, H., Anthropological Notes on the Western Coastal Tribes of the Northern Territory of South Australia, Trans. Royal Society of South Australia, Adelaide, 1907, p. 24, north-west coast tribes; Davidson, Katherine-Victoria Rivers district, 7. S—Strickland River, Goodenough Bay and Hinterland. Graebner, see footnote 4, p. 27, Geelvink Bay, Gilbert Is.

- 261

technique, as is shown in fig. 3 is widely distributed in Queensland, New South Wales, Victoria, South Australia, Central Australia, and North Australia. It is not found in Tasmania 4 nor, on the basis of present information, in Western Australia. It would seem to be a possibility that it might be forthcoming in the Wyndham district of Western Australia, in view of its prominent use in the contiguous region of North Australia. For the greater part of Western Australia, however, all forms of knotless netting appear to be absent. Museum collections from this region are consistent in their lack of specimens of this nature, and the ethnographic literature which is available either makes no mention of this technique or denies its presence.

In the eastern part of the continent, however, it seems permissible to assume a more general use of netting than that indicated on the map. New South Wales and southern Queensland are very poorly known, and it is now too late to secure any additional information from these regions. The lack of positive appearances on the map, therefore, should not be considered as indicative of an absence of the simple loop in this area. The extensive and almost contiguous distribution of the known appearances in surrounding regions would seem to constitute sufficient grounds for this belief. The simple loop is found in a number of places in New Guinea, and there can be no doubt that the Australian and Papuan appearances are historically related.

B. The Loop and Twist. Slightly more complicated than the simple loop is the loop and twist, which is made by taking a turn in the ascending strand with its continuation, the descending strand. There are two varieties of this technique, the single-twist and the double-twist. The difference lies in the order in which the ascending strand is first taken over the pendant loop of the preceding row. Assuming that the progression is moving from left to right the single-twist is made by passing the ascending strand - 262 behind and over the pendant loop and then by taking a turn in this ascending strand with the descending strand (fig. 4). In the double-twist method the ascending strand

FIG. 4.

Loop and Single Twist.

is first placed over the pendant loop from the front and the strand brought downward as in a half-hitch. It is then given a full turn about the ascending strand (fig. 5).

FIG. 5.

Loop and Double Twist.

The relationship of the two seems obvious. The loop and double-twist, however, does not seem to be extensively used in Australia, and apparently is not found where the single-twist is not also present. Its greatest use seems to be in the Boulia district of western Queensland, and with this exception the distribution shown in fig. 6 refers to the loop and single-twist. This technique, as the map demonstrates, appears to be found exclusively in eastern Australia. It

- 263

Unlocated., Brough, Smyth, 1, p. 344, figs. 155-156, possibly Victoria., 1. S—Batavia River; B—Cape York, Roth, Bull. 7, pp. 27-28, Batavia, Pennefather and Embley Rivers., 2. S—Normanton, Staaten River; B—Normanton, Mitchell River., Roth, Bull. 7, p. 13, Laura, Maytown, Highbury, Musgrave, Coen, Gilbert River, Delta, Normanton; pp. 27-28, Mitchell to Flinders Rivers and across peninsula to junction of Palmer and Mitchell Rivers and north of Gilbert River, Princess Charlotte Bay., 3. B—Glenormiston (headdress net)., Roth, Bull. 7, p. 27, Boulia; ibid, loop and double twist., 4. A—Kopperamanna, Birdsville, Warburton River (net)., 5. S—Euston., 6. Peron, Novelle Hollande, plate; North Holland (locality not specified)., Phillips, p 179, cited by Graebner 1913, p. 26, Edge Partington, Album, III, p. 139, South-east, Interior., 7. S—Bellinger River., 8. S—Fly River, Papuan Gulf, Strictland River.

is unknown in Tasmania and Western Australia and also seems to be lacking throughout most of Victoria, South Australia, Central Australia, and North Australia. Similarly to the other techniques considered we have no information - 264 for this type from a large part of southern Queensland and northern New South Wales. The peculiarities of distribution, however, would seem to imply again that if our information were more complete we should undoubtedly find it in a large part of this region. This technique, like the simple loop and the knotted methods, is also found in New Guinea and a historical connection cannot be doubted.

C. The Double Loop or Hourglass technique. The most complicated form of knotless netting found in Australia is the so-called double loop or hourglass technique, illustrated in fig. 7. This method seems to be a direct outgrowth of

FiG. 7.

Hourglass or Figure Eight type of Knotless Netting.

the loop and single-twist technique, for the latter is the basis of its construction. The loop and single-twist pattern is first made and the end of the strand then carried up and around the first descending strand, just below its supporting strand, before it is carried to the next pendant loop. After it is taken over the next pendant loop it is brought back under itself, just below its intersection with the original descending strand, and then carried onward again as described. This technique, as shown in fig. 8,

- 265

1. A—Port Stewart, Flinders Is., Cape Melville, Bathurst Head, Cape Bedford, Princess Charlotte Bay; M—Cape York Peninsula, Princess Charlotte Bay; S—Bags: Cape Bedford, Batavia River, Bloomfield River, Palmer River, Port Douglass, Cooktown; Nets: Batavia River, Palmer River, Bloomfield River, Cooktown; B—Cape Bedford, Cape York, Coen River, Musgrave River, Annan River, Glenormiston, Moreton, Mareeba, Butcher's Hill, McIvon River, Laura River, Batavia River, Albatross Bay., Roth, Bull. 7, pp. 27-28, Mitchell River to Flinders River and across peninsula to east coast, Gladstone to Rockhampton; Bags: Normanton, Delta, Gilbert River, Cooktown, Cape Bedford, Cape Melville, Morehead, Musgrave, Middle Palmer River, Bloomfield; Fish nets: Laura, Maytown, Palmerville, Morehead River, Musgrave River, Pennefather, Embley and Batavia Rivers., 2. B—Sangar Tribe (Eilander River); S—Astrolabe Bay, Huon G., Finschafen, Collingswood Bay, Cloudy Bay, Redscar Bay, Ope River., Williamson, footnote 17, p. 204, calls this technique “Mafulu Network,” and says that it “is never adopted on the coast or Mekea plains (all nets of this description found there having come down from the mountains).”, Graebner, p. 27, North Coast: Geelvink Bay to north-east New Guinea; South Coast: Papuan Gulf to Massim Dist. (compare Williamson). Also north coast of New Britain, Baining tribe

- 266

seems to be confined entirely to the Cape York Peninsula and the north central coast of Queensland. It is also extensively used in New Guinea and historical relationship between these two appearances again seems obvious.

Now there can be no doubt but that the three forms of knotless netting in Australia are genetically related, and that the order of their development has been from simple to complex in the order of (1) Simple Loop, (2) Loop and Twist, and (3) Hourglass-pattern. We have seen that from the technological point of view the hourglass-pattern is derived directly from a loop and twist base. The loop and twist method is obviously but a slight alteration of the simple loop, and the relationship of the three, therefore, seems plain. This sequence, however, is also found geographically in Australia. The simple loop is found throughout the areas of the loop and twist and the hourglass-patterns, as well as in areas peripheral to them, North Australia, Central Australia, South Australia, and Victoria. The loop and twist technique is present throughout the distribution of the hourglass-pattern and is also peripheral to it in southern Queensland and in New South Wales. Geographically, therefore, we find the same sequence as we pass from south to north as was determined on a technological basis.

There can be no doubt, at least in so far as Australia is concerned, that New Guinea is the home of these techniques. The question of the ultimate point of origin of knotless netting is another problem, and one which does not concern us at present. From the Australian point of view it is sufficient to trace the types found on that continent to the nearest locality from which they could have been derived. The reasons are numerous for believing that New Guinea, rather than Australia, was the place from which diffusions took place to result in the described distributions.

In the first place it is a commonly-accepted belief that the influence of the Australians upon other ethnic areas has been practically nil, and, as a result, we should hardly expect that a series of netting techniques would be the exceptions to the rule. On the other hand it is well known - 267 that New Guinea has influenced Australia, particularly the Cape York peninsula-region, in a number of ways. The derivation of the netting-techniques from New Guinea would be consistent, therefore, with the movement of a number of traits into Australia.

Secondly, it is important to note that New Guinea exhibits not only the three techniques found in Australia but also a number of additional variations of knotless netting, many of which are much more complex than those found in Australia, and for the construction of which the latter are fundamental requirements. These varieties have been illustrated and discussed by Graebner. 5

Thirdly, the Australian and New Guinean appearances are but a part of a much more extensive and involved diffusion of at least the simple loop technique from west to east in New Guinea, as Graebner has indicated. It seems quite possible, from a theoretical point of view, that the Pacific appearance of knotless netting may have had its beginning pre-historically in what is now Malaysia 6 or south-eastern Asia.

Fourthly, the geographical distribution of the three techniques in Australia is consistent with—indeed, even indicative of—diffusions from New Guinea, for not only they, but also appearances in the Melanesian Islands, definitely point to New Guinea as a centre from which they have radiated. The New Guinean appearances, on the other hand, do not appear to be arranged in a distribution which would indicate a peripheral condition. The distribution of all the varieties of knotless netting on this great island are very poorly known, but the non-contiguous appearances of the more fundamental techniques and the apparently more or less contiguous and localized distributions of the more complex ones would seem definitely to indicate a much longer history in knotless netting than do the simple but contiguous distributions in Australia.

- 268

For these reasons the conclusion that the three forms of knotless netting in Australia have diffused from New Guinea would seem to be evident without question. Furthermore, it seems to be likewise apparent that the three probably diffused into Australia at different times, the intervals separating the independent diffusions being, of course, impossible to determine. Apparently the simple loop spread across Torres Strait at a time before even the simple mutations now found in New Guinea had been invented. Later, when the loop and twist technique was perfected, it, too, was carried into Australia and started on a southward diffusion. Finally the hourglass-pattern was developed and, in turn, carried to the Cape York Peninsula. In view of the absence of other New Guinean forms in Australia, it would seem to be inferable that they had not been present in the Torres Strait region of New Guinea at the time when any of the three diffusions took place. We cannot conclude that they were also lacking in other parts of New Guinea, however, where a number of them seem to be centered, although it would seem likely, in view of their comparatively localized distributions, that they may be fairly recent developments.

Now the establishing of a chronology of knotless netting types in Australia on a geographical basis has been justified primarily because we have been treating with traits which appear to be genetically related. The danger of applying such a method to traits which are not shown to be historically related is well known and need not be elucidated here. It is with caution, therefore, that I introduce into discussion the possible chronology between the knotless netting techniques and that of knotted netting.

Knotted netting, as we have seen, is the most wide-spread technique in bag- and net-making in Australia. In addition to being found throughout the regions where the three knotless techniques are found, respectively, it is also present in regions peripheral to them, the northern part of Western Australia and possibly western Central Australia. In view of the similarity of the relationship of this distribution to those of the knotless forms which, we concluded, - 269 demonstrate a chronology, the question arises whether we should permit ourselves to extend this chronology to include the last step in a regular and obvious geographical progression (see fig. 9). Considering the function which all

Illustration

these techniques serve, that of net- and bag-making, there would seem to be good grounds for inferring such a sequence. On the other hand a logical objection would be that the knotted and knotless techniques are not obviously nor necessarily related and that, therefore, the former could very well be a recent invader of Australia which has enjoyed a very rapid diffusion, a diffusion which has carried it across and beyond the distribution of its knotless predecessors. There is a possibility, however, that the knotted and knotless forms may not be fundamentally unrelated, - 270 and the evidence to support this contention is to be found in a comparison of the techniques of their construction.

Knotted netting, like the various forms of knotless netting, is based upon the prinicple of constructing a body, row after row, with one continuous strand (fig. 10). The

FIG. 10.

Knotted Netting.

process is similar in each case, in that each row of meshes forms a series of loops which serve as the points of attachment for the meshes of the succeeding row. The difference between the two types lies in the manner of construction of the meshes, for the knotless varieties are based upon the principle of taking a half-hitch in each loop, whereas the knotted form is based upon the tying of a becket-knot in each loop. The first results in a mesh-work with considerable give, whereas the second, in so far as each mesh is concerned, is secure and definitely arranged to prevent slipping. The technique followed, however, is strikingly similar in both, the main difference being that in the knotless type the strand is brought back, after being carried over the pendant loop, and passed over or around itself, whereas in the knotted type it is carried upward and taken around the sides of the pendant-loop before being brought down to cross itself. The similarity of the two methods can be seen by comparing figs. 2 and 10.

On the basis of this comparison we thus see that there would seem to be a good probability that knotless and knotted netting may be historically related. Their techniques of manufacture are similar; they are used for the same purposes; and they are made from the same materials. - 271 We have no way of telling, however, the historical order of their development, if such actually happened, nor the locality where it could have taken place, for it would seem to be possible to derive either one from the other. At first glance the knotted type is certainly more complicated than the simple loop and might, therefore, seem more likely to be an outgrowth of the knotless type than the basis from which the latter developed. On the other hand it seems just as obvious that the simple loop may represent a simplification of the more involved knotted technique. From a technological point of view there seems to be no answer to the question as to which may have been the original form, for both, at best, are simple, and neither type can be shown to be a necessary foundation for the development of the other. The world-wide distribution of knotted netting, and its Neolithic appearance in Europe, however, would seem to imply the greater antiquity.

Now the argument as to the priority in historical development is concerned only with the hypothetical place of origin of these forms of netting, and is not necessarily involved in the question of their historical sequence in Australia, for it is quite possible, if not very probable, that the later form arrived in Australia before the historically earlier one diffused there. We are not concerned at the moment, therefore, with the problem of the priority of development but only with the question of the priority of sequence of these techniques in Australia. As we have already seen, the facts of geographical distribution suggest that the knotted form preceded the knotless forms, and this conclusion is not denied by technological considerations. The only alternative explanation would be that knotted netting arrived in Australia later than knotless netting and diffused more rapidly until it not only had occupied the distribution of the knotless forms but also had spread beyond them, to be taken up by peoples who supposedly were resisting the diffusion of the knotless forms or at least who had had an opportunity to acquire the latter but had not done so. Such a conclusion would be neither plausible nor consistent with what we know of the processes of culture diffusion. It cannot be supposed, until we have evidence to the contrary, that the aborigines of southern and western Australia were resisting the diffusion of any - 272 kind of netting. All of them, in so far as we know, seem to be acquainted with the art of making string from human hair, fur, or fibres, for use in belts, ligatures, head-bands, ceremonial objects, or string-figures, and with this necessary equipment for netting it seems logical to suppose that these natives would welcome the knowledge of netting techniques, once they were exposed to their influences. The fact that they do not have them can be best explained only in historical terms, that the concept of netting has not yet diffused to them and that they would as readily take over one form of netting as another. On the basis of this reasoning the geographical progression of knotted netting, simple loop, loop and twist, hourglass-pattern, is further supported as the temporal sequence in which these techniques arrived in Australia, and subsequently diffused to their present distributions.

We cannot dispose of questions of the origin of knotless netting, however, without consideration of coiled basketry and the part which it may have played in such a development. As already stated, knotless netting usually has been regarded as a coiling technique without a foundation and, for this reason, a study of the former must involve an examination of the latter. Before attempting to draw final conclusions on the history of knotless netting in Australia, therefore, we must devote attention to the coiling-techniques found there.

Australian baskets, as a rule, are made in two principal techniques, coiling and twining. Plaiting is also known but appears to be unimportant. Roth reports its presence in the Cape York Peninsula at Cape Bedford and the Morehead, Musgrave, Embley, Morton, and Upper Batavia rivers. 7 In the National Museum of Victoria there are two plaited rush-baskets said to be from Victoria. Plaited mats and baskets are also reported from Torres Strait. With such scanty information, however, it is impossible to consider this technique in its relationship to plaiting in other areas. Were it not for the Victorian specimens the other appearances would appear to have been the results of a recent - 273 influence from Torres Strait. The Victorian examples, however, if related to those from Cape York Peninsula, introduce a problem which cannot be profitably discussed until more information is forthcoming.

The distribution of twined basketry is shown in fig. 11. The modern use of this technique extends throughout northern Australia as far west as King's Sound, but in south-eastern Australia and in Tasmania it is now extinct.

There seems to be a lack of consistency in the reports of the early observers in Tasmania on the subject of basketry, but such authorities as Roth 8 and Noetling, 9 who have summarized the available information, agree that, in all likelihood, the twining-technique was the only one known. At least there are neither descriptions nor actual bona fide Tasmanian baskets which illustrate any other technique. The Tasmanian baskets in the museum at Hobart are characterised by their pliability and open work. It is also possible that a closely woven basket was made, for mention was made of baskets used as water-carriers. However, since no specimens of this type are known, it is possible that there has been a confusion in identity with skin-bags which were used for that purpose. According to Noetling, the Tasmanians wove their baskets from top to bottom, in contrast to what is believed to be the common practice in Australia of starting at the bottom and ending at the top or mouth. Just what significance this difference may have historically is difficult to say on the basis of present information, for we have very little knowledge concerning technology in Australia.

The Australian twined baskets may be classified in two ways, rigid or non-rigid, closely woven or openly woven. In either case it is difficult to draw a line of distinction, for the technique, apparently, is similar in all cases, rigidity or pliability being determined primarily by the type of materials used. Materials are also important in determining the open or closed types of baskets, for the fine grass-like

- 274

1. B—Dunk Island, Torres Strait (western islands), Daintree River, Ducie River; M—Cardwell, Cairns, Burdekin River, Russell River (eel trap); S—Bag: Normanton; Baskets: Staaten River, Palmer River; Eel traps: Atherton, Russell River. Port Denison; Cairns type basket: Dunk Is., Atherton, Russell River, Cardwell, Herberton (suggestive of Tasmanian type); Baskets: Charters Towers, Cook-town, Cairns, Morton, Bloomfield River. Cape Melville, Morehead River, Starcke River, Pennefather River, Dunk Is.; A—Port Stewart, Flinders Is., Cape York Peninsula, Cardwell, Batavia River., Brough Smyth, I, p. 346, fig. 16, Burdekin River, Roth, Bull. 7, p. 28, Gulf Coast, Mitchell River to Staaten River (pliable or sieve bags), Princess Charlotte Bay (baskets), Palmer River, Endeavour River, Cape Bedford, Bloomfield River, Cairns, Cardwell and Atherton, Gladstone to Broadsound; Bull. I, p. 14, Cape Bedford and various North Queensland (dilly bags), Batavia and Pennefather River (mats), Atherton, Cairns, etc. (fish and wallaby traps)., Report of the Cambridge Expedition to Torres Strait, vol 4, p. 74, and pl. 17, 5 and 6, Western Islands., 2. M—New South Wales (location not given)., 3. M—Pyalong (location ?)., 4. Peron, pl. 13, no. 4 (apparently rigid with open work); Launceston and Hobart: flexible and open., 5. A—Coopers Creek (said to be from Queensland)., 6. S—Darwin, Alligator River (also mats), Mary River, Port Essington, Merkinalal Creek; A—Alligator River, Melville Is., Bathurst Is., Groote Is., Daly River, Newcastle Waters, Port Essington, Borroloola, Daly Waters, Darwin, McArthur River, Roper River, Pine Creek, Katherine River., Basedow, 1907, p. 40, north-west coastal tribes, B—Daly River, Darwin, Roper River, Boolman, Wilton River., 7. M—East Kimberley, Fitzroy River; P—Derby and King Leopold Range; S—Hinterland of King's Sound.

- 275

strands can be woven very closely, whereas the stout strands by their nature dictate a relatively open weave.

Although the twining-technique seems to predominate along the northern coast of Australia, sporadic appearances are reported from South Australia, Victoria, and New South Wales. The South Australian example comes from Cooper's Creek and is represented by a single specimen in the South Australian Museum. It is definitely said to have come originally from Queensland. Presumably south-western Queensland is implied. The other south-eastern appearances are based upon two specimens in the National Museum of Victoria. One, said to come from New South Wales, is not localized; the other is reported to be from Pyalong, Victoria. The former is like the baskets common to Queensland and the northern coast. The latter is definitely of the Tasmanian type.

This is not the only appearance of Tasmanian type baskets on the continent, however, for a number of specimens from the Cardwell-Cairns region of Queensland have this same openly spaced twining, are made from similar soft pliable material and are difficult to distinguish from the Tasmanian examples. In reality, since these specimens from Tasmania, Pyalong, and Queensland are collapsible, they should be called basket-bags.

In view of the scanty and unsatisfactory information from south-eastern Australia we cannot be certain of any implications that might be drawn. In view of the fact that southern Queensland and northern New South Wales are so poorly known, it may be that these few specimens represent what was formerly a more important basketry-technique in this general region. On the other hand, the information is very indefinite and it must be realized that too much reliance must not be placed upon it. It would seem, however, that there must be some relationship between the Tasmanian appearances and the Australian twined baskets and in this respect the authenticity of the specimens from Victoria and New South Wales seems more plausible. It cannot be doubted but that twined basketry is a relatively old trait in this part of the world. Its appearance in Tasmania would seem to be sufficient basis for this conclusion.

- 276

FIG. 12A.

Typical Coiled basket, south-east Australia.

FIG. 12B.

Detail of Bundle-coil of basket 12A.

The distribution of the coiling-technique is shown in fig. 13. As can be readily seen it is confined to eastern Australia, appearing in Queensland, New South Wales, Victoria, and eastern South Australia. It has not been reported from the western part of the continent nor from Tasmania. It is to be noted that no information is available from the greater part of New South Wales and southern Queensland. Presumably, however, the distribution would be more extensive if those regions were known.

- 277

It is strange to note that coiling and twining are seldomly found together in the same area at the same time (compare figs. 11 and 13). In view of the widespread distributions of each in eastern Australia a stratification in which coiling has replaced twining seems to be impliable. It is important to note, however, that twining has not given way in the Cape York Peninsula, where coiling is not reported, and through which coiling must have passed in its southward diffusion.

DILLY AND PITURI BAGS: 1. Roth, Bull. 7, p. 28, Boulia, U. Georgina, Camooweal; see also Roth, Bull. 1, pl. 15, no. 4., ROUND BASKETS: 2. S—Barcoo River., ROUND BASKETS AND BAGS: 2. Roth, Bull. 7, p. 14, Gladstone to Broadsound, Rockhampton., S—Keppel Is, Fraser Is., Stradbroke Is., Holly Hill; B—Bribie Is., Comet Waters.,

- 278

ROUND BASKETS: 4. Roth, Bull. 7, p. 14, Morton Bay., S—Clarence River, Richmond River, Copmanhurst., ROUND AND FLAT BASKETS AND MATS: 5. M—Lower Murray River, Lake Conday, Yarra tribe, Western Victoria, Mortlake., Brough Smyth, I, p. 345, Southern Australia (?), Gippsland., A—Narrinyerri, Port Victor, Victoria, Lake Condah (fish trap); S—Warnambool., Howitt, A. W., The Native Tribes of South-east Australia, London, 1904, p. 719, Kurnai (similar to Narrinyerri)., Woods, p. 43, Narrinyerri., Taplin, G., The Narrinyerri, Adelaide, 1878, mentions “basketry” (and mats in some cases), p. 18, Moroura Tr. and junction of Murray and Darling Rivers; p. 30, Moorundie tribe, Manuum to Overland Corner and 12 miles back on each side; p 40, Narrinyerri; p 58, Tatiara and south-east tribes; p. 62, Wallarooh (Yorke's Peninsula); p. 64, Flinders Range and Crystal Brook; p. 65, Mt. Remarkable tribe; p. 96, Port Lincoln to Fowlers Bay; p. 102, Port Lincoln; p. 103, Absent at Fowlers Bay, and Gawler Range; p. 88, Absent at Mt. Remarkable., ROUND BASKETS: 6. B—Coopers Creek (Queensland type)., 7. Cambridge Ex. to Torres Straits, vol. IV, p. 82, Mer. Davidson, Thursday Is. (place of origin unknown). Flat basket., 8 Astrolabe Bay: Coiled baskets. Coiled bags are also found in New Caledonia.

The coiling-technique is applied to three types of objects in Australia—mats, baskets, and dilly-bags. The mats seem to be confined exclusively to Victoria and adjacent portion of South Australia and probably southern New South Wales, in which regions baskets are also typical. Baskets are also found in north-eastern New South Wales and in southern and south-eastern Queensland (see figs. 12, 15, 16 and 17). Dilly-bags made by a coiling-technique are found in eastern Queensland, for the most part in areas where coiled baskets are not reported, and in the Boulia district of western Queensland where, according to Roth, no baskets of any description are present. 10 A similar technique is found in New Caledonia where it is employed in the manufacture of sling-stone bags. 11

Coiled baskets are also reported from Torres Strait. 12 At Thursday Island the writer secured one of unusual pattern quite unlike those of southern Australia (fig. 14). The locality of manufacture of this specimen is unknown, but it is said that it might have come from the Australian - 279 mainland. Coiled baskets are also found in New Guinea, and are extensively distributed in the Melanesian Islands where somewhat different techniques from those used in Australia are encountered.

We thus find that in Australia the typical coiled baskets, in the sense of the term generally understood, are found exclusively southward from the central part of Queensland, whereas north of this region the technique is, so far as we know, applied only to dilly bags, except at Torres Strait where what seems to be a local type of basket is made. It is thus seen that throughout Australia the technique of manufacture seems to be fundamentally the same. There are two differences in details of construction, however, which appear to be of importance because of similarities found in Melanesia. One is the type of foundation or coil employed; the other is the number of coils circumscribed by each turn of the binding strand.

Multi-strand Coil. In many regions of the world where coiled basketry is found the coil consists of a number of pliable strands bunched together (fig. 12b). This type of coil is continuous only in that the numerous individual strands which compose it overlap and are bound together by a circumscribing binding strand. The use of coils of this type gives comparative thickness as well as semi-rigidity to the walls of the basket and thus ensures a consistent shape. This type of coil is associated with the baskets and mats of south-eastern Australia and is not found in the north where the bags are manufactured. It is also found in New Guinea at Astrolabe Bay and in the Carolines. 13

Single-strand Coil. A different type of coil is made from a single continuous strand or by a number of single strands fastened together or overlapped at their ends into a continuous strand. If material of a pliable nature, such as string, is used, rigidity cannot be secured, and the product will be a bag rather than a basket. On the other hand, a rigid or semi-rigid basket can be made if the coil is of a stout material not readily flexible.

- 280

The single-strand type of coil is found in Australia in a pliable material, string, in the coiled dilly-bags of Queensland (fig. 19), and in a semi-pliable material, pandanus-strips, if we accept as Australian the specimen from Torres Strait shown in fig. 14. It is found in Melanesia

FIG. 14A., FIG. 14B.

Coiled flat basket, collected at Thursday Island., Detail of basket in 14A.

- 281

in the form of a liana (figs. 19 and 20), and, according to Graebner, is the only type of coil used in that region. Coiled baskets in this technique apparently are not found in New Guinea, but the Melanesian system of coiling which is based on a single-strand coil is found on the north coast as well as in the Fly River region of the southern coast, associated with woven armour and breast-shields. It is important to note, therefore, that although different materials are used in the different regions, there is a suggestion of a contiguous distribution in the use of a single-strand coil from Melanesia to the southern distribution of bags in Queensland.

The multi-strand type of coil, on the other hand, as we have seen, occupies a non-contiguous distribution, appearing at present only in south-eastern Australia, New Guinea, and the Carolines. On the basis of this distribution of the two types of coil, the conclusion is indicated that in this part of the world, at least, the multi-strand coil is the older, and that the single-strand, as it is now utilized in Australia and Melanesia, is a more recent development which has diffused into a large part of the area formerly occupied by the older multi-strand coil. This does not necessarily imply that the multi-strand type of coil was at one time widely distributed in Melanesia, but it would seem to definitely indicate that this type of coil was formerly used throughout the northern part of eastern Australia.

One Coil at a Time. In regard to the number of coils circumscribed by the working-strand, in Australian specimens it is to be noticed that, with a few exceptions, the coils are taken one at a time in both baskets and mats in Victoria and South Australia and in the baskets and bags of New South Wales, Queensland, and Torres Strait. The exceptions to this rule are well illustrated in the examples shown in figs. 15, 16, and 17. In the first instance we have an unusual basket from Framlingham, Victoria, made out of two mat-like plaques, now in the Australian Museum, Sydney. The coiling-technique in this specimen follows, for the most part, the southern pattern by circumscribing with the working-strand only one coil at a time.

- 282

FIG. 15.

Coiled basket from Framlingham, Victoria. (Specimen in The Australian Museum, Sydney.)

In a number of instances, however, and apparently more as a decorative feature than for a utilitarian purpose, two coils are included in a single turn of the binder. These instances are so few, however, that the general character of the basket is not altered. For the second type of exception we find in a basket from Fraser Island, Queensland, also in the Australian Museum, that, except for the coils near the rim, the entire surface has been fabricated on the plan of including two coils in each turn of the binder

- 283

FIG. 16.

Coiled basket from Fraser Island, Queensland. (Specimen in The Australian Museum, Sydney.)

(fig. 16). A close inspection, however, reveals the fact that beneath this binding the ordinary coiling method of taking one coil at a time is found. Such a result is as one might expect, however, for a multi-strand coil requires a circumscribing binder for each coil separately in order to tie and hold together the numerous loose strands of the coil. The circumscribing of more than one coil in this type of basket or mat, without the corresponding binding together of the strands of each coil separately, undoubtedly would weaken the product so that the result would not be so well suited to practical uses. A somewhat similar example, in which three coils are bound together by half-hitches over the half-hitches taken around the individual coils, was secured by the writer at Swan Reach, South Australia. This basket, which is illustrated in fig. 17, seems

- 284

FIG. 17.

Coiled basket collected at Swan Reach, South Australia.

- 285

to be of unusual shape for this region. The technique employed, however, serves to illustrate the point made above, that a multi-strand coil more or less dictates a circumscribing of each individual coil by the binder. This may be done in a number of ways, as the baskets with bundle-coils in various parts of the world reveal. The half-hitching method is only one means, although a very efficient one, of achieving such a result.

A similar method of circumscribing each coil individually by half-hitches is also found in the bags of Queensland (fig. 18) and in the pandanus strip-basket

FIG. 18.

Queensland Coiled Bags.

from Torres Strait. It is thus apparent that this method of coiling has a contiguous distribution from south-eastern Australia to Torres Strait. At Astrolabe Bay in New Guinea the same technique is, as already mentioned, present in coiled baskets, and a similar appearance has been reported from the Carolines. The sling-stone bags from New Caledonia are also made in the same technique as the Queensland bags.

Two Coils at a Time. In contrast to the Australian, Astrolabe Bay, Caroline Island, and New Caledonian method of taking one coil at a time in the half-hitches of the binding-strand, Melanesian coiled baskets are characterized, according to Graebner, by the exclusive fashion of including - 286 two adjacent coils in each turn of the binding-element. Such a trait was noticed in a few examples from southeastern Australia, but in all instances, except one, it was seen that such a wrapping was supplementary to the usual practice of taking each coil individually. In the one exception we saw that the fundamental half-hitching of the separate coils was omitted in only a small portion of the basket and that the wrapping of two coils at a time undoubtedly served only as a decorative device. These Australian appearances must be regarded as only incidental in importance, for they seem to be no more than embellishments which have been added to the fundamental method of circumscribing one coil at a time. In being innovations they must be later in origin than the latter, but they would seem to have no historical connection with the coiling-method now found in Melanesia.

The modern Melanesian method of coiling is quite different. Instead of the binding being based upon half-hitching, which, as we have seen, characterizes the Australian mats, baskets and bags, the Astrolabe Bay and Caroline Bay baskets and the New Caledonian sling-stone bags, we find a process which can best be described as spiralling. This technique is present in two varieties. In one (fig. 19) the binder is wrapped spirally around each

FIG. 19.

Melanesian Spiral Coiling (after Graebner).

- 287

couple of adjacent coils in such a manner that the binding strands do not cross each other. In the other (fig. 20) the same spiral movement is employed, but the binding-strands cross each other at each point where they pass over the foundation-coils. There can be no doubt, as Graebner 14 suggests, that they are fundamentally the same technique. Both are widely distributed in Melanesia, being found, at least as far east as the New Hebrides, and in some regions are applied to various objects, such as shields and armour,

FIG. 20.

Melanesian Spiral Coiling (after Graebner).

in addition to, or other than, baskets. Since their individual distributions are concerned with problems local to Melanesia they need not be given in detail here. 15 It is important for us to note, however, that the first variety is reported from the Angriffe Harbour and Fly River regions of New Guinea, where it is associated with woven armour, and that the second is found along the north coast of New Guinea, including the Sepik River district, in breast-shields and other objects. 16

- 288

Now to consider the relative distributions of the two fundamental methods of coiling we find that the first, in which the binder is taken in half-hitches around each coil individually, has a much more widespread distribution than the second, in which the binder is carried spirally around each couple of adjacent coils. The appearances of the first are not only non-contiguous to each other, since they occur in Australia, the Astrolabe Bay region of New Guinea, the Carolines, and New Caledonia, but are also peripheral to the second on three sides of its distribution, Australia, the Carolines, and New Caledonia. It would seem, therefore, that there are good grounds for suspecting that the various appearances of half-hitch coiling are related and that spiral coiling, which occupies a relatively continuous and centralized distribution, represents a more recent development than the former. To my mind such a conclusion would be indicated even if we had only the Australian and New Guinean evidence to go by, for since the method found in Australia undoubtedly diffused from New Guinea, where it is still found, it would be obvious that the second had replaced it in the intervening regions of New Guinea. Such an explanation is indicated by the concentration of coiling in the eastern third of the continent on the direct corridor from New Guinea, by the finding of baskets of the same type in south-eastern Australia and Astrolabe Bay, and by the general principle that many culture traits have diffused to Australia from New Guinea.

The conclusion that the half-hitching method of coiling is older in these regions than spiral coiling is implied by Graebner, for he definitely assigns the former to what the German school calls the Old Australian Kulturkreis. He denies, however, the possibility of genetic relationship between the two methods by placing the latter in the Two Class Kulturkreis as an original trait. Graebner considered it strange that both cultures should possess such a similar trait as coiled basketry, but goes on to state that he regards this similarity as only apparent, and proceeds to point out that the baskets of south-eastern Australia have grass-bundle coils whereas the Melanesian baskets have single-liana coils. Not being satisfied to conclude a lack of historical relationship of the basis of the different materials used, however, he proceeds to contrast the Australian method - 289 of taking one coil at a time in the binder, with the Melanesian technique of including two coils in each spiral of the binder. Graebner explains the difficulty presented by the single-strand coil of the Queensland coiled bags by attributing them to a local combination of coiled basketry and knotless netting, and since he regarded the latter as belonging, in all probability, to the Old Australian Kulturkreis no further explanation seemed necessary. On the basis of such differences is materials and techniques and the peculiarities of distribution, Graebner assumes his case proven, that coiled basketry was an original possession of both the Old Australian Culture and the Two Class Culture.

Now such conclusions may or may not seem warranted to those who accept the Kulturkreis point of view. If the validity of this method of approach is not accepted, however, and the facts are considered in their own right and without prejudice toward any world scheme of culture-development and diffusion, somewhat different conclusions suggest themselves.

We have already seen that the relative distributions of the two types of coil and of the two methods of coiling imply that the bundle-coil and half-hitching are respectively earlier in this general region than the single-strand coil and spiralling. Such a chronology is not inconsistent with Graebner's construction, for he regards what is called the Old Australian Kulturkreis as earlier than what is known as the Two Class Kulturkreis. The distributions also suggest, however, that there may be a genetic relationship between the bundle- and single-strand types of coil, as well as between the half-hitching and spiralling methods of coiling. The possibilities of such relationships have been denied by Graebner on two bases of consideration; first, the differences in materials and in techniques in the two regions; second, the peculiarities of distribution, as interpreted in accordance with Kulturkreis hypothesis. It is necessary, therefore, that we consider the technological and distributional factors in order to ascertain whether they appear to support or deny the conclusions drawn from them.

In the first place it is not at all obvious that the half-hitching and spiralling methods of coiling are as fundamentally different as Graebner would have us believe. Both - 290 are applied in a circumscribing motion to the foundation, the main difference being that one is brought back under itself, whereas the other continues without such a crossing. We have already seen that some kind of a circumscribing binder for each coil is more or less a requirement for the bundle type of coil. For that reason areas in which bundle coils are the only type known would seem to constitute rather effective barriers against the introduction of two-coil spiralling.

A spiral binding, furthermore, is not as serviceable as half-hitching in bags, as a moment's reflection will indicate, for with a soft pliable material such as string, a hitch of some kind is necessary to prevent a “pull” from disorganizing the entire structure of the bag. Without the half-hitches each string in the Queensland coiled bags could act in the manner of a shirr-string, but with half-hitches taken about each coil it would be quite difficult to disarrange the structure of the bag by accidentally catching any part of a coil or binding-strand.

The restriction of the spiralling method to Melanesia and the lack of it in Australia, therefore, although probably due primarily to historical reasons, may also be partly explainable in terms of the types of coil present. In order that spiralling might diffuse into Australia it would be first advantageous to change from string-coils and bundle-coils to single-coils of some stout material. Such a beginning is possibly indicated in the pandanus-strip coils at Torres Strait. It is impossible at the present time, however, to tell whether this appearance represents the result of a diffused influence from the Melanesians, or the local application of a new coil-material to an old technique.

In those regions where a stout single-coil was in use, it would have been relatively easy to have changed to spiralling from half-hitching. In the baskets of Melanesia lianas are used for both coils and binders. Since a liana is a relatively stiff strand, not readily pliable, it would seem obvious that a sharp turn in it, such as a half-hitch, would be not only difficult to make but also impractical in a basket where hundreds of hitches are required. This is not the only reason, however, why half-hitching might be apt to change in these baskets; for since these coils - 291 are stout single-strands, they do not require a tying, which is a primary function of half-hitching in the multi-strand grass-coils, nor do they need hitches to prevent shirring, as in the coiled bags.

The two techniques of coiling, therefore, appear to show a reasonable correlation in distribution with the kinds of materials used. It would not be correct to say that interchange between them could not take place, but it does seem reasonable to suppose that such a combination would require some modification in technique or some alteration in materials.

As the result of such a correlation it would seem that materials may have played a role of much greater significance in the history of coiling-techniques than Graebner has been willing to concede. From the technological point of view, therefore, there appears to be no basis for a belief that the half-hitching and spiralling-techniques are fundamentally different, or opposed in principle to each other, or that the latter could not have been a direct outgrowth from the former, as the result of a change in material. In so far as the relationship between bundle-coils and single-strand-coils is concerned it seems obvious that a change from the one to the other readily could have been made.

From the point of view of distribution there likewise would seem to be no basis for objecting to the conclusions that spiralling may have been derived from half-hitching and that single-strand-coils may be no more than substitutions for bundle-coils. From the distributional angle such conclusions appear to my mind as much more probable than the explanations offered by Graebner. This does not necessarily imply, however, that the changes involved must have taken place in Melanesia or New Guinea, nor that the complete distribution of spiralling, as now found in Melanesia, was once occupied by half-hitching. Allowance must be made for the diffusion of the former into many localities where the latter has never been known, as the result of relatively recent influences. The lack of stratification of the two traits in certain regions, or the lack of their contemporaneity in others, therefore, must not be interpreted in terms of relationship or the lack of relationship. Similarly, the absence of bundle-coils throughout most - 292 of Melanesia should not be regarded as evidence that there is no relationship between them and single-strand-coils.

In the broader aspects of these problems it must not be forgotten that Melanesia, like Australia, is peripherally located in respect to the East Indies and south-eastern Asia. The traits not indigenous in origin, found in either of these marginal regions, therefore, must have come from the West. There can be little doubt that spiralling is not a locally invented Melanesian trait but one the origin of which should be sought in the East Indies or on the Asiatic mainland. Whether it diffused into Melanesia or was brought by a migrating people cannot be determined on the basis of present information. It also seems quite certain that half-hitching came originally from the West, at a somewhat earlier time. Since we do not know the place of origin of either of these two traits we cannot gainfully discuss their ultimate relationship, except in terms of probabilities. It does seem permissible to say, however, that there seems to be no technological reasons which deny the possibility that spiralling could have been derived from half-hitching, or which refutes the conclusion that single-strand-coils may represent a direct transition from bundle-coils. Since all these traits seem to have traversed the same route into the Pacific they appear to go back to a common area of departure. The possibility that they may be genetically related, therefore, would seem to carry more weight than the assumption that they have no relationship whatever. When we also take into consideration that the products are not unalike in purpose and use, that the techniques of construction are quite similar, that a change from one to the other involves no difficulties, and that their distributions in the Pacific imply a relationship, the probability of their relationship would seem to be greatly strengthened.

In respect to the coiled bags of Queensland there would seem to be at least four possible explanations of their origin. They may represent (1) a local combination of Australian coiled basketry and knotless netting, as Graebner suggests, (2) an independent application of string to the process of coiling, (3) the results of a diffusion from New Guinea - 293 of the idea of a single-strand-coil, which the Australians locally applied to string, or (4) the results of an actual diffusion of coiled bags from New Guinea.

1. If the Queensland coiled bags represent the local combination of coiled basketry and knotless netting there are a number of factors and conclusions which would seem to be inconsistent with the facts, as well as with the Kulturkreis hypothesis. In the first place if both belong to the Old Australian Kulturkreis, as Graebner states to be the probability, then both must have diffused from New Guinea. Such a conclusion is indicated by the facts, as we have seen and is implied by Graebner. Graebner, however, regards the Australian type of coiled basketry as the direct ancestor of knotless netting, a conclusion which would appear to be a possibility, since both are based upon a half-hitch technique, but in spite of the fact that the first is built upon a coil or foundation, whereas the second is manufactured without such a base. However, if both diffused into Australia it follows that the place where knotless netting was derived from coiled basketry could not have been Australia.

Now if knotless netting has been derived from coiled basketry the change could have first taken place, seemingly, only through the substitution of string for the materials used in the baskets. The first step in such a transition could have been the substitution of string for (a) the binding-element, (b) the grass-bundle-coil, or (c) for both binder and coil. If the substitution were made for the binder only, coiled baskets would still be the result, and progress toward knotless netting would have hardly begun. In the second instance, if the changes were made only in the coil, a result similar to the coiled bag would be attained. By such a process, however, a single-strand-coil would have been derived from a multi-strand-coil and such might have lead directly to the use of the liana-coil such as is now found in Melanesia. In the third case, if string were substituted for both coil and binder, a coiled-string bag would result. Coiled bags, therefore, instead of being the product of the combination of coiled basketry and knotless netting might equally as well represent a transitional step between them, if they are genetically related, for the with- - 294 drawal of the string-coil is all that would be necessary to change coiled bags into a product identical to simple loop-bags.

2. Coiled bags may represent the independent application of the coiling-process to string. In this respect it is important to note that the same half hitch-binding of individual coils is found in both the Australian baskets and bags. Coiled bags, therefore, could have been derived easily from coiled baskets by a simple substitution of materials, and independent of influence from any other product.

3. A similar result could have been attained by the diffusion from New Guinea of the idea of using a single-strand-coil which the Queenslanders applied to string. We have already noticed that the distribution of single-strand-coils appears to be relatively contiguous from Melanesia to southern Queensland. Different materials are employed in the various regions, but it is possible that they may represent local applications in reaction to the same historical influences.

4. Australian coiled bags may represent the results of an actual diffusion from New Guinea. Such a happening would be consistent with the diffusions of the various netting and basketry techniques which we have considered. Support for this hypothesis is also given by the relative distributions in Australia of coiled baskets and coiled bags, for, if coiled bags have been derived from coiled baskets at some place outside of Australia, the chronology is indicated by the peripheral location of coiled baskets in Australia. We may feel certain that coiled basketry diffused from New Guinea, and the inference from the distributions is that coiled bags may have done likewise.

We must also take into consideration the coiled bags of New Caledonia; for, unless we are to suppose convergent or parallel development in Queensland and New Caledonia, it follows that these two appearances are historically related. If, as Graebner concludes, coiled bags were developed in Queensland, then the New Caledonian bags must represent a diffusion from Queensland, a conclusion which seems hardly plausible. On the other hand a diffusion from New Caledonia to Queensland would appear to be also untenable. - 295 Sight must not be lost of the possibility, however, that both may have diffused from a common centre. In such a case New Guinea would appear as the most likely place of derivation in so far as Queensland is concerned. Whether coiled bags originated in New Guinea is another matter, but it would not be inconsistent with what is known of Melanesian history to suppose that this great island might also have been the ultimate centre from which the New Caledonian bags were derived. In this respect it is important to note that the Queensland and New Caledonian coiled bags are both peripheral to the common types of liana-baskets of Melanesia. A similar marginal condition is also manifest in them, in the half-hitching method of coiling in comparison with the spiralling method, and in the practice of binding one coil at a time, compared with the spiralling of two coils in each turn of the binder.

In view of all these possibilities it would thus seem to be apparent that coiled bags may not represent the combination of coiled baskets and knotless netting, as Graebner supposes. When all factors are taken into consideration, it would seem to be just as likely, if there is any relationship between coiled bags and knotless netting, that the former may have been genetically earlier.

Disregarding the possibility of genetic relationship between coiled bags and knotless netting, however, it is interesting to note that the present northern boundary of the distribution of coiled bags ends approximately where the southern boundary of the hourglass-technique begins. There would seem to be no rational explanation why these two traits should be mutually exclusive in their distributions, for coiled bags and the techniques of the simple loop and the loop and twist do occur together. Be that as it may, the distributions would seem to imply that coiled bags, if derived by diffusion from New Guinea, preceded the hourglass-technique. If the coiled bags are indigenous to Queensland, which seems hardly probable, it also appears that their development took place before the hourglass-technique invaded Australia, for the latter seems to be diffusing toward or into the area occupied by the former.

- 296

There would seem to be little doubt but that coiled bags are directly related to coiled baskets. The relative distributions and the similarities in techniques appear to be sufficient grounds for this conclusion. We cannot be so certain, however, that there is a direct relationship between knotless netting and coiled baskets or coiled bags.

In order to derive knotless netting from coiled baskets, it is necessary to advance a number of assumptions which lead to difficulties. In the first place, we must postulate the substitution of string for vegetable strands. As we have seen, however, the mere substitution of string would not involve a change in the process of manufacture and, as a result, we should have, at most, a coiled bag rather than a basket. It is easy enough to point out that a knotless simple-loop bag would be obtained if the string coils of such a bag were withdrawn. This is obvious and, on first thought, such an explanation of the origin of knotless netting would seem to be simple and reasonable. A moment's reflection, however, will lead one to see that what has been discovered is not knotless netting but the fact that coiled bags are still serviceable even though their coils have been withdrawn. Such products, although identical with knotless netting bags, should not be confused with the latter, for the processes of construction are not the same. In reality they should be termed “coiled bags with the foundation withdrawn.”

In spite of the fact that the processes of coiling, as found in the Australian baskets and bags, and in knotless netting generally, are similar in that both are based upon half-hitching, it would seem that we have no right, solely because of this similarity, to assume them to be historically related. Surely a half-hitch is sufficiently simple in principle to have been applied independently to different uses by various peoples. In this respect it is important to note that the half-hitches in knotless netting are taken directly about the pendant loops of the preceding row, whereas in coiled basketry they are taken about the coil. Incidentally, in the Australian type of basket, the continuation of the half-hitch is carried over the binding half-hitch of the preceding coil, with the result that each half-hitch is actually, but - 297 fortuitously, found around this loop if the coil element is withdrawn. Such a difference in application cannot be considered as proof that the two processes are not related, but it does indicate that they could be the results of independent developments. At any rate it would appear that we have no right to assume that the process of knotless netting would introduce itself merely in consequence of the observation that coiled bags could be made into an altered but still serviceable product by the removal of the coil-strands. For these reasons it would seem to be no more obvious that knotless netting has been derived from coiled work by the withdrawal of the coils than that coiled work has been derived from knotless netting by the insertion of the coils.

The possibility that there may be no direct relationship between coiled basketry and knotless netting is further indicated when one takes into consideration the great distribution throughout the world of each of them and their many varieties, but does not find coiled bags, in so far as the writer has been able to determine, except in Queensland and New Caledonia. If the half-hitch type of coiled basketry were the basis of knotless netting it would seem that some evidences of coiled bags should be forthcoming in other regions. However, since there is always the chance that some other people might apply the technique of coiling to string, the finding of coiled bags in other areas would not necessarily prove anything. Furthermore, there is always the possibility that the motor-habits acquired in the development of one technique may so influence the application of another technique, which is similar but unrelated, that the original differences in process between the two may become concealed.

It thus seems to be quite evident that the relationship between coiled basketry and knotless netting is neither obvious nor provable. Nor is it possible to state definitely that they are not related. On the other hand, there is the alternate possibility that knotless netting may have been derived directly from knotted netting. As we have already seen in the first part of this paper there appears to be a much greater similarity between knotless netting and knotted netting in process of construction, as well as in - 298 materials and uses, than between knotless netting and coiled basketry. On the present basis of knowledge it is impossible to state which of knotless or knotted netting, if related, is the parent trait, for either one, apparently, could have been derived from the other. On the basis of the geographical distribution of the two, however, knotted netting, in its almost universal appearance, would seem to be possibly the older, for knotless netting, although also very widespread, is found in a much more limited area.

By way of conclusion it is possible to indicate a tentative chronology for the traits discussed, in so far as Australia is concerned.

In view of the occurrence of twined basketry in Tasmania there can be but little doubt that the twining technique is a very old one in the Pacific. Its concentration in eastern and northern Australia and its absence in the central and south-western parts of the continent are indeed puzzling. Were it not for the Tasmanian appearance such a distribution would not indicate a very ancient history in this part of the world. To assume that it is of recent derivation, however, would also imply the assumption that the Tasmanians were in recent contact with the Victorian tribes. Such a conclusion, it seems unnecessary to add, would not be consistent with the important differences in culture between Victoria and Tasmania, which, in themselves, imply a relatively long isolation of the latter. Of the basket- and netting - techniques, therefore, twining appears to be the oldest in this region.

Both coiling and netting apparently have relatively less antiquity in Australia, but since we have been unable to find any genetic relationship between these two groups of traits it is impossible to infer their chronological order of appearance on that continent. Of the traits associated with coiling, however, we have seen that coiled baskets with a bundle-coil and half-hitching appear quite definitely to be older than coiled bags and the single-strand string coil. This chronology is indicated regardless of whether coiled bags originated in Australia or New Guinea. We also saw that coiled bags appear to be older in Australia than the - 299 hourglass-technique of knotless netting and that historically there may be some relationship between the string-strand of Queensland and the liana-coils of Melanesia.

Spiralling is not directly involved in Australian chronology, but since it plays such an important part in Melanesia, it is important to recall that it appears to be later in that region than half-hitching.

In respect to the netting-techniques there would seem to be no doubt that their chronological order in Australia has been, first, knotted netting; and second, knotless netting, the latter in the sequence of simple loop, loop and twist, and hourglass-techniques.

These chronologies apparently indicate the maximum antiquity, relatively speaking, for each of these traits in Australia. Twining must have been known before the Tasmanians became isolated from Australia and coiling and netting must have arrived subsequently. The maximum antiquity of each coiling and netting trait, in turn, can be relatively established as the minimum amount involved in the antiquity of the preceding trait.

It would seem to be possible, however, to infer a minimum relative antiquity of the three forms of knotless netting, the bundle-coil, and half-hitch coiling (and of coiled bags, if of New Guinean derivation) on the basis of the break of their distributions at Torres Strait. Since all of these traits are found in Australia, on the one hand, and New Guinea, on the other (except coiled bags), but are lacking in the intervening regions under Melanesian influence, Torres Strait, 17 or the coastal country of south-eastern New Guinea, 18 it would seem plausible to regard their minimum antiquity in Australia as the maximum length of time of the Melanesian occupation of these regions. Eventually, it is to be hoped that it will be possible to determine these relative time-limits more accurately from sources of knowledge unknown at present.