The documentation and explanation of distributions of artefacts are of major concern in prehistoric archaeology. A fundamental component of this concern is the production of distribution maps. In Australia, a strong historical-diffusionist tradition using distribution maps in the interpretation of historical items of Aboriginal material culture flourished in the 1930s (e.g. Davidson 1933, 1934, 1936; McCarthy 1939-40). In contrast, major use of stone artefact distribution maps in Australian archaeology has only occurred in the last two decades following the professional establishment of the discipline in the 1960s and the ensuing accumulation of basic distributional data. During this time, most interest has been focused upon the distribution of backed blades, points, tula adzes and edge-ground axes (e.g. Davidson 1983; Hiscock and Hughes 1980; Morwood and Trezise 1989; Mulvaney 1975; Pearce 1974; Pickering 1990; Smith and Cundy 1985). In a few areas, historical information is available to account for the movement and distribution of these implement types (e.g. points - Davidson 1935; axes - Hiscock 1988a; McBryde 1978, 1984, 1986; tulas - Hiscock 1988b; Roth 1904; see also McCarthy 1977; Mulvaney 1976). However, most of these studies involve trade and little detailed attention has been directed towards identifying other processes (e.g. diffusion of ideas, independent invention) which may account for the occurrence of artefact types in other parts of the continent (Davidson and McCarthy 1957; McCarthy 1977; Pearce 1974; Rowland 1987; Smith and Cundy 1985; Walters 1988).
|
In this paper I examine the occurrence of tula adzes and bifacial points in the coastal and sub-coastal regions of southeast Queensland (Fig. 1). The first section describes known tulas and bifacial points from southeast Queensland, thus expanding upon generally accepted views concerning the distribution and use of these implements across Australia. In the second half of the paper I confront this new distributional information directly, and evaluate it in terms of existing general models of invasion, innovation and diffusion. Following on from these discussions, I develop a hypothesis that the occurrence of these implements outside their main range may have been a result of information flow (diffusion) from 'inland' regions made possible by the establishment of large-scale inter-regional social linkages across southern Queensland during the late Holocene. |
Figure 1. Southeast Queensland showing location of sites and places mentioned in text (1: Tin Can Bay Site 62-63; 2: King's Bore sandblow Site 97; 3: Brooyar Rockshelter; 4: Sandstone Point; 5: Platypus Rockshelter; 6: Mt Crosby Road Site; 7: Gatton Shelter; 8: Broadbeach Aboriginal Burial Ground; A: Bunya Mountains; B: Blackall Range). |
Numerous 19th century ethnographic observations record the use of hafted, chipped stone adzes by arid/semi-arid zone Aboriginal people (e.g. Etheridge 1891:40-2; Liversidge 1894:238; Roth 1904:17, 20; Smyth 1878:379; Spencer and Gillen 1899:583, 594-5, 1904:636-40; Sterling 1896:93-5; Worsnop 1897:114). However, it was only in 1924 that Horne and Aiston (1924:80, 89, 101-3, 106; see also Aiston 1928, 1929) described in detail a specific hafted, woodworking adze called 'koondi tuhla' by the Wonkonguru people of the Lake Eyre district. The adze ('tuhla') was hafted onto the end of a long handle ('koondi') with organic cement (e.g. resins etc.) and periodically resharpened with a hammerstone. This distinctive stone tool was soon recognised archaeologically, and became known as a 'tula' adze (Hale and Tindale 1930:159; see also McCarthy 1976:31; McCarthy et al. 1946:29; Sheridan 1979:10, 17-24).
Tula adzes are retouched flakes usually made from fine-grained siliceous stone (e.g. chert, silcrete, chalcedony). Characteristically, they are semi-discoidal in shape, 15 mm to 60 mm wide, and exhibit wide platforms with dorsal retouch along the distal margin worked back to the bulb of percussion (Hiscock 1988b:63; Hiscock and Veth 1991:333-5; Kamminga 1985:9). The bulb of percussion is usually pronounced giving the adze immense mechanical strength (Sheridan 1979). Thinning (retouching) of the proximal end of the flake may also take place to aid hafting (Hiscock 1988b:65). Following the ethnographic model, archaeological tulas are generally thought to have been hafted onto a handle. The distal end (working edge) was left exposed and successive resharpening (retouching) of this edge gradually reduced the tool distally/proximally to a point where use was impractical (Aiston 1928:127; Horne and Aiston 1924:89; Howchin 1934:41; McCarthy 1976:30-1; McNiven 1991a:493; Mitchell 1949:46; Mulvaney 1975:82, 232). This 'worn-out' (Aiston 1928:127; Horne and Aiston 1924:89) remnant is usually termed a tula 'slug' (McCarthy et al. 1946:27; see also Campbell and Noone 1943a:299, 1943b:381; Noone 1943:275).
A number of flaked adzes have been surface collected and excavated from the Cooloola-Gympie region (McNiven 1988, 1990a) (Fig. 1). Four of these implements conform to descriptions of tulas. They are described as follows:
|
1. Tula adze (TCB62-63/IF3)
The tula was surface collected from tidal mudflats adjacent to Sites 62-63 along the eastern periphery of Tin Can Bay, Cooloola (Fig. 1). It is apparent that the implement had recently eroded from a stratified shell midden located a few metres away in eucalypt forest. Given that a basal radiocarbon date of 807 CAL BP was obtained from the site (see McNiven 1991b for details), it is probable that the tula dates to within this period. The adze is made from grey silcrete and exhibits edge rounding (use-wear) along most of its working edge (Fig. 2a). 2. Tula adze (Brooyar IF1) The adze was collected from the surface of Brooyar Rockshelter (Fig. 1). Given that the cultural deposit of the site has a near basal radiocarbon date of 2762 CAL BP (see McNiven 1988 for details), it is apparent that the adze was discarded after this time. The semi-circular-shaped adze is made from green chert and similarly exhibits edge rounding (use-wear) along the entire length of the convex working edge (Fig. 2b). 3. Tula adze (Brooyar IF6) This adze was also found on the surface of Brooyar Rockshelter and has a similar antiquity of less than ca. 2800 CAL BP. It is made from very dark green chert with the ventral surface exhibiting a series of step terminated shallow flake scars removed from the distal margin (i.e. working edge) (Fig. 2c). These flake scars are consistent with impact damage resulting from use (Kamminga 1982). More diagnostic edge rounding (use-wear) is located along the length of the convex working edge. 4. Tula slug (Brooyar IF7) This tula slug also comes from the surface of Brooyar Rockshelter with an antiquity similar to the other two described adzes. It is made from very dark green chert and exhibits edge rounding (use-wear) along the entire length of the working edge and use-polish across most of the ventral surface (Fig. 2d). Small particles of black residue located across the proximal end of the implement appear to be the remains of hafting resin. |
Figure 2. Tula adzes from Cooloola and Brooyar Rockshelter (A: TCB62-63/IF3; B: Brooyar IF1; C: Brooyar IF6; D: Brooyar IF7). |
|
Artefact No |
Raw material |
Platform thickness (mm) |
Platform width (mm) |
Length
(mm) |
Width
(mm) |
Weight
(g) |
|
TCB62-63/IF1 |
silcrete |
5 |
21 |
22 |
35 |
10.9 |
|
Brooyar IF1 |
chert |
2 |
9 |
14 |
19 |
2.5 |
|
Brooyar IF6 |
chert |
11 |
19 |
8 |
29 |
2.2 |
|
Brooyar IF7 |
chert |
7 |
21 |
10 |
22 |
5.3 |
Metrical variables recorded for each tula are presented in Table 1. While the four implements vary in size displaying little standardisation, they remain within the documented size range mentioned above. A larger sample of tulas is needed before further reliable inferences concerning tula morphology can be made.
It is generally accepted that tula adzes functioned as hafted woodworking tools (Flood 1983:191-2; Kamminga 1982:76; Sheridan 1979). The identification of major edge rounding and polish on the Cooloola and Brooyar tulas and the possible hafting resin on one of these implements supports strongly the hafted woodworking adze function associated with the inland tulas. In this connection, it is worth noting that Mathew (1910:120) mentions the existence of stone 'chisels' (vis-a-vis stone 'spokeshaves') in the study area. I suggest that Mathew may have been referring to hafted adzes (including tulas) similar to those documented from more arid regions. However, the occurrence of tulas on the sub-tropical east coast indicates that these implements were not exclusively an arid zone adaptation (cf. Flood 1983:192; Kamminga 1982:76; Mulvaney 1975:235; Sheridan 1979; White and O'Connell 1982:131). In fact, the Tin Can Bay tula was found in seawater less than 5 km from sub-tropical rainforest. Further insights into the function of the coastal tulas will depend upon the results of continuing use-wear and residue studies.
Tulas have also been reported from other sites in southeast Queensland. They include the Mt Crosby Road Site (Norma Richardson, Queensland Museum, pers. comm. 1989), Gatton Shelter (Morwood 1986), Platypus Rockshelter (Hiscock and Hall 1988) and Broadbeach Aboriginal Burial Ground (Haglund 1976) (Fig. 1). While the antiquity of the Mt Crosby Road tulas is unknown, available data suggest that tulas have a maximum antiquity of 2700 BP at Platypus Rockshelter (Hall and Hiscock 1988:48; Hiscock and Hall 1988:84), 1300 BP at Broadbeach Aboriginal Burial Ground (Haglund 1976) and 1000 BP at Gatton Shelter (Morwood 1986:109).
Current distribution maps of tula adzes do not include the east coast of Australia (Flood 1983:187; Kamminga 1985:19). The most easterly extent of these implements is generally accepted to be the Central Highlands of Queensland located some 400 km inland from the east coast (Hiscock 1988b:68; Morwood 1981, 1984a, 1984b). The recovery of tulas from at least six sites in southeast Queensland demonstrates that the presence of these artefacts on the eastern seaboard is not 'fortuitous' (Mulvaney 1975:235).
Australian bifacial points consist of 'leaf-shaped' retouched flakes exhibiting bifacial secondary working along both lateral margins and often around the platform or proximal end of the flake. They tend to be triangular in section, approximately three times longer than wide, with a mean length of 40 mm (range = 25 mm to 70 mm) (see Flood 1983:186-8; McCarthy 1976:42-4; McCarthy et al. 1946:38-40; Mitchell 1949:63-4; White and O'Connell 1982:106).
Six bifacial points have been recovered from Site 97, Cooloola (McNiven 1985, 1990b, in press). The site is located within King's Bore sandblow, a huge (ca. 700 m x 300 m) mobile parabolic sand-dune situated approximately 300 m inland from Teewah Beach midway between Noosa Heads and Double Island Point (Fig. 1). The points were collected from an approximately 10,000 m2² stone artefact surface scatter located across the dune floor. All of these artefacts (including the points) appear to have been deflated from their original stratigraphic context by massive wind erosion. Recent radiocarbon age determinations obtained from remnants of this original stratigraphic context (cultural layer) within the dune apex yielded an antiquity of between 2332 CAL BP and 3833 CAL BP (see McNiven, in press for details). A brief description of the points follows.
|
Collection unit |
Raw material |
Weight (g) |
Length (mm) |
Width (mm) |
|
Circle 9/1 |
chert |
1.12 |
19 |
11 |
|
IF 8 |
andesite |
37.42 |
72 |
28 |
|
IF 9 |
silcrete |
21.06 |
51 |
27 |
|
IF 10 |
andesite |
14.62 |
48 |
24 |
|
IF 11 |
silcrete |
18.10 |
57 |
28 |
|
IF 13 |
silcrete |
28.17 |
54 |
28 |
Recorded dimensions of the points are presented in Table 2. Length was measured along the central axis of the point while width was maximum width perpendicular to the central axis. The points display reasonable standardisation, with most 50 mm to 60 mm in length and 25 mm to 30 mm in width. The size range of these points is in accord with the size range for previously recorded points in Australia (White and O'Connell 1982:106).
Most Australian bifacial points are believed to have functioned as hafted spearheads (Flood 1983:186; McCarthy 1976:40; Schrire 1982:247; White and O'Connell 1982:123; see also Kamminga 1978:335, 1985). While the specific function(s) of the Cooloola points is unknown, it is noted that the broken tips on five of the six Cooloola points are consistent with the projectile point hypothesis (Akerman 1978; Bergman and Newcomer 1983; Epstein 1963; Kamminga 1978:338; Witthoft 1968). Future use-wear and residue analyses may shed more light on the function of these implements.
Platypus Rockshelter is the only other documented site in southeast Queensland containing points. A single specimen was recovered from Stratigraphic Unit 1 which formed some time after ca. 2400 BP (Hall and Hiscock 1988:47-8; Hiscock and Hall 1988:84).
Bifacial points were previously unknown from the east coast of Australia (Flood 1983:189). Although unifacial points have been recorded from the Central Highlands of Queensland (Morwood 1981), bifacial points are generally accepted as being restricted to the central northern part of tropical Australia (Mulvaney 1975:217; Smith and Cundy 1985:36; White and O'Connell 1982:116). This places the Cooloola points some 400 km from the nearest documented point producers in the Central Highlands (Morwood 1981) and over 1000 km from the nearest documented bifacial point producers in northwest Queensland (Hiscock 1988c).
Despite recent euphemisms, analyses of stone artefact distributions are generally reduced to a discussion of the relative merits of invasion, innovation and diffusion. The invasion hypothesis (Clark 1966) or more generally the theory of successive populations (Adams 1968) posits that major changes in regional artefact assemblages result from the physical replacement of one group/culture by a different group/culture (see Rouse 1986; Trigger 1968). The explanation is generally restricted to areas exhibiting major and abrupt changes in society (e.g. material culture, religion, burial practices, social organisation etc.) and skeletal (genetic) make-up (e.g. the pre-Neanderthal hypothesis - Howell 1957; Howells 1976; Kennedy 1975; Valladas et al. 1987; Wolpoff 1981:289-96).
The problem of isolating population succession becomes more clouded and difficult to identify archaeologically when only partial migration of peoples occurs (Trigger 1968:41-5). For example, it is possible for a small group of people to become incorporated within another group and introduce new cultural traits to the adopter group. Such group amalgamations can take place for a variety of social, political and economic reasons (e.g. exogamous marriage arrangements) and can be incorporated within a more encompassing model of inter-societal/inter-regional linkage (see below).
Identifying whether the appearance of a new material culture item/trait in a region results from either internal processes and developments (independent invention/innovation) or through external stimulation and information flow (diffusion) has its difficulties. The problem of differentiating between these two processes is of fundamental concern not only to archaeology, but to anthropology in general. Even last century, Tylor (1884:547) referred to 'this great problem' in anthropology, claiming its resolution would bring 'the study of civilization into its full development as a science'.
Unfortunately, the basis of many earlier independent invention-diffusion arguments is more emotional than factual, reflecting the personal biases of individual researchers. An example of this phenomenon is the high frequency of diffusionist explanations during the first half of this century for the external origin of many historic Aboriginal cultural traits (e.g. Elliot-Smith 1930:135-9), 1933:5, 8; Rivers 1926:158-66; Thorpe 1924). It is no coincidence that these explanations were most prolific at a time when Aboriginal cultures were seen as 'lowly' (Elliot-Smith 1930; Rivers 1926) and uninventive (cf. British 'invasion neurosis' - Clark 1966:172; see also Kuklick 1991).
Independently invented material culture traits are internally generated by a cultural system and generally termed innovations (cf. Layton 1989). Barnett (1953:7) defines an innovation as:
Innovations can be generated by 'goal-oriented' or 'incidental random' activities (Barnett 1953:109). In the case of subsistence items and production tools, goal-directed innovations can be seen as adaptive processes, responding to changes in the functional requirements of these items initiated by changes in subsistence activities, social organisation and the like (Schiffer and Skibo 1988:598-9). Despite these basic observations, the archaeological identification of independent invention remains poorly developed. This results from the lack of theoretical insight into the process of innovation and the nature of cultural antecedents (Hester and Grady 1982:91; Leeuw and Torrence 1989; Rogers 1983:135), and our subsequent inability to identify these phenomena in the archaeological record (White and O'Connell 1982:122; see also Dickson 1975; Pearce 1974, 1975). At present, the archaeological identification of independent invention/ innovation is generally applied only to contexts where external influences can be ruled out.
Diffusion can be defined as:
Although Rogers' (1983) definition has wide acceptance, it should be noted that diffusion is not a process per se, but the result of a complex set of interactive events (cf. Hodder 1986:11). In this connection, mechanisms of diffusion concern social interactions and linkages that provide the agent(s) that have the means to effect the transfer (diffusion) of phenomenon from one group to another (Morrill et al. 1988:9). One mechanism of information diffusion is through idea (e.g. language and/or kinesics) communication/exchange. In such cases, transfer of information on a particular artefact form/trait occurs from one group to another. An alternative, though by no means mutually exclusive mechanism for information transfer is through object exchange (Renfrew 1977; Renfrew et al. 1968). In a very general sense, exchange can be defined as 'the spatial distribution of materials from hand to hand and from social group to social group' (Earle 1982:2). Whether or not exchange recipients also acquire technological information on the manufacture of an exchange item, and/or attempt to replicate the exchange item (e.g. Wardaman bifacial point replication - Davidson 1935:170-2; see also Akerman and Bindon 1983; Ride 1958:175) and/or perceive the item in similar or differing ways is another matter (cf. stimulus diffusion - Kroeber 1940; Trigger 1968:28; see also Davis 1983:66-7).
A major criterion for isolating object exchange in a region is the manufacture of items from exotic raw materials (e.g. Allen and Duerden 1982; Cummins 1974; Dixon 1976; Egloff 1978; Hammond et al. 1977; Jack 1976; Lilley 1988; McBryde 1984; Pires-Ferreira 1976; Reeves and Ward 1976; Weigand et al. 1977; see also Bahn 1982; Heizer 1968). Whether such exchanges reflect kin-based gift reciprocity, redistribution or specialised marketing systems is difficult to determine and has stimulated much debate (Allen 1985; Hodder 1982: 146-52; Pires-Ferreira and Flannery 1976; see also Polanyi 1957; Sahlins 1972). While these higher order debates continue, it is clear that the more fundamental issue of differentiating indirect procurement (exchange) from direct procurement remains with us, plagued by problems or equifinality (Meltzer 1989).
I suggest that from an archaeological perspective, and valid hypothesis concerning information flow (diffusion) as an explanation for the occurrence of a material culture trait in a region should accommodate at least four conditions. First, the trait must be a new addition (intrusion) into a region, be of local manufacture, and exhibit no evidence of an in situ developmental history. Second, an external origin for the trait must be found. In this connection, the closer the potential source area to the study (adopter) region, the more plausible the diffusion hypothesis. Third, the diffused trait must exist in the potential source area immediately prior to its uptake in the study (adopter) region. Such chronological continuity should manifest both a chronosequence and a positive relationship between the distance the trait has diffused from its source and the timing of trait adoption. Whether or not traits are adopted gradually through spatially contiguous groups (cf. 'contagious wave' or 'neighbourhood effect' diffusion models - Ammerman and Cavalli-Sforza 1971; Clark 1968:435-8; Hudson 1969:45-9; Pedersen 1970:204-7; see also Gregory 1985; Hagerstrand 1967) and/or leapfrog across a landscape due to differential exposure and/or acceptance of traits resulting from a variety of social, cultural and/or political factors (e.g. east coast Australian fishhooks - Rowland 1981:67-8, 1982:114-4, 1983:135; Walters 1988) is a matter for empirical determination. Fourth, some form of social linkage between the diffusion source and potential trait adopter should be identified to provide a mechanism and processual baseline for inferences concerning the transfer of information (see White and O'Connell 1982:121).
The southeast Queensland tulas and points could occur as a result of either population succession, independent invention or information flow (diffusion). At present, no evidence is available to suggest that any form of invasion or population succession occurred in southeast Queensland during the Holocene prior to the appearance of Europeans (e.g. Haglund 1976). Similarly, the manufacture of all tulas and points from local raw materials rules out the possibility of object exchange from distant inland areas. While it is possible that independent invention/innovation may account for the occurrence of these implements, I suggest that available evidence is more supportive of idea communication/exchange. This hypothesis is developed below in terms of the four archaeological conditions for identifying information flow (diffusion) detailed above.
As mentioned above, unfortunately our ability to differentiate between trait intrusion via information flow and trait innovation in archaeological contexts is negligible. In the case of tulas and points, few data are available concerning either the nature of technological antecedents or the degree of chronological resolution required before meaningful inferences can be drawn. as a result, the question of local innovation in terms of the southeast Queensland tulas and points remains unanswered.
Both tulas and points are well documented for southern central Queensland (Morwood 1981). Although located some 300 km to 400 km inland from southeast Queensland, these 'inland' implements provide a plausible source area for information flow.
The only dated tulas and points from southern central Queensland come from the Central Highlands. Results from a series of excavations indicate that tulas have an antiquity of ca. 100 BP to ca. 4000 BP while points are more restricted in time from ca. 2800 BP to ca. 4300 BP (Morwood 1981, 1984a). As noted above, such chronologies match remarkably well those obtained for southeast Queensland. For example, tulas date to within the last 3000 years, while the bulk of bifacial points (e.g. King's Bore Sandblow Site 97) were dated somewhere between ca. 2300 BP and 3800 BP. Given that the chronology for tula and point use in the Central Highlands is slightly older than that recorded for southeast Queensland, it is clear that the former region represents a plausible source area for information transfer to the latter region.
McCarthy 1940:249) states 'the frequent gathering together of groups for ceremonies and other purposes provide a simple mechanism whereby ideas diffused throughout Australia'.
Last century, certain Aboriginal groups in southeast Queensland hosted large Bunya nut harvests and associated social, political and ceremonial events for Aboriginal people from as far away as northern New South Wales and southern central Queensland (Sullivan 1977). The main focuses of these gatherings were the Bunya Mountains and the Blackall Range, located in the hinterland region of Cooloola (Fig. 1). During such gatherings, it is well documented that exchange of material goods (e.g. hunting items, possum skin rugs, necklaces etc.) and information (e.g. kinship systems, ceremonies etc.) took place (Calley 1959:132; Petrie 1904:19, 56). As Petrie (1904:56-7) states:
The most westerly participants in what might be called the Bunya interaction sphere appear to have been the Mandandanji people from southern central Queensland (Tindale 1974:125). Consequently, their presence at the Bunya gatherings would have provided an excellent opportunity for the dissemination of information concerning 'inland' cultural traits to the coast (Sullivan 1977:41) and vice versa. I speculate therefore, that inter-regional linkages of a similar scale and area may have provided a mechanism for the diffusion of information concerning tulas and points from southern central to southeast Queensland (Fig. 4).
A major implication of this diffusion hypothesis is that inter-regional social linkages between southern central and southeast Queensland have an antiquity of at least ca. 3000 years. Other researchers have speculated, however, based largely on excavations at a large coastal gathering site at Sandstone Point, that large-scale gatherings in southeast Queensland have a much more recent (BP) antiquity (Nolan 1986:98; Walters 1987:102, 1989:221; see also Morwood 1987:346). While I acknowledge that large-scale linkages of the form recorded last century (e.g. Bunya gatherings, Sandstone Point gatherings) may date to the recent prehistoric past, I believe the appearance of tulas and points in southern central and southeast Queensland some 3000 to 4000 years ago following comparable ca. 20,000 year antiquities for human occupation (Mulvaney and Joyce 1965; Neal and Stock 1986) supports the view that large-scale inter-regional linkages may have been operating across southern Queensland for a much greater time. What specific form these hypothesised social linkages across southern Queensland may have taken prior to 1000 BP is a matter for future research.
From the available data, I believe that information flow (diffusion) is a plausible hypothesis to account for the occurrence of tulas and points in southeast Queensland (cf. Davidson and McCarthy 1957:397, 405-6). Further theoretical and substantive research is required, however, to explore and expand upon criteria and data used to support this hypothesis. In particular, emphasis should be placed upon documenting the character, distribution and antiquity of both tula and point production and use from southeast Queensland through o southern central Queensland.
The suggestion that information flow through social linkages may have influenced the distribution of tulas and points in southeast Queensland is consistent with aspects of arguments concerning late Holocene intensification in Australia (Lourandos 1983, 1985). While lack of testability and falsifiability plague the intensification debate, available evidence is consistent with the hypothesis that the spread of many tool types (e.g. adzes, axes, points, backed blades, etc.) across many parts of Australia in the mid- to late-Holocene was primarily a result of a restructuring of inter-regional social networks (e.g. McBryde 1978; Morwood 1984b, 1987; White and O'Connell 1982:133; see also David and Cole 1990; Jones 1977). In all cases, however, it is important to realise that these suggested material expressions of inter-regional linkages may not necessarily correspond to the commencement of social interaction. They simply represent a particular manifestation of that hypothesized interaction that has a tangible archaeological signature. Therefore, the challenge of future research is to develop new ways of identifying past interactive systems, and to critically examine under what socio-cultural and chrono-spatial circumstances associated material traits/items appear and indeed disappear from the archaeological record (Schortman and Urban 1987).
This paper expands present knowledge on the distribution of tula adzes and bifacial points across Australia. It is hoped that the information flow (diffusion) hypothesis will stimulate further research into an important, though poorly understood area in Australia prehistory; namely the processes responsible for the development and distribution of implement types across the continent (Smith and Cundy 1985). In particular, it is hoped that more objective and rigorous investigations of the relative roles of innovation (independent invention) and information communication/exchange (diffusion) are initiated (Dortch 1981). As Leslie White (1957:218) prophetically stated more than 30 years ago:
Bruno David, Jay Hall, Peter Hiscock, Ian Lilley, Harry Lourandos, Lynette Russell, Leonn Satterthwait and Ian Walters made helpful criticisms of earlier drafts of this paper. Advice on tula adzes was kindly provided by Fred McCarthy. All errors in fact or interpretation are my responsibility.