Fishing in Port Jackson, New South Wales - more than met the eye.

Attenbrow, Val ; Steele, Dominic

Contemporary diaries and the water-colours of artists such as the 'Port Jackson Painter' vividly tell of Aboriginal life when the first Fleet in 1788 settled its cargo of convicts in Australia. Fishing was important around the waters of Port Jackson, whose Aboriginal inhabitants are recorded to have used the techniques of spear-fishing and angling. Were other methods also used? Fish remains from a shell midden provide an opportunity to investigate.

In prehistoric studies historical descriptions are often used to interpret the archaeological record of an area; usually those of the area in which the archaeological site or sites occur are used, but sometimes those of another environmentally similar region (Binford 1983: 25-6; Lampert 1971a: 62-4; 1971b: 116-18; 1988: 43; Meehan & Jones 1988; Poiner 1976: 197; Spaulding 1968:37-9). Historical descriptions are also seen as portraying the most recent 'traditional' phase in the cultural succession of the original inhabitants of a colonized region, i.e. how the people lived before their culture was impacted upon by the colonizers. Some researchers have been critical of the use of ethnographic and ethno-historical sources to extrapolate back into the distant past to interpret archaeological evidence (Meehan & Jones 1988).

This article focuses on fishing methods used in Port Jackson, and assesses the documentary and archaeological evidence for the type of equipment that Aboriginal people used to catch fish. This issue arose as part of a larger project about the role of marine and land resources in the diet and material culture of the Aboriginal people of Port Jackson (Attenbrow 1991).

This article questions the reliability of the First Fleet historical records, but at the same time indicates that both documentary and archaeological sources provide essential evidence and both may be necessary to produce a valid picture of what happened at contact and what changes may have occurred in the past.

Port Jackson historical descriptions

In the earliest historical records for Port Jackson (those of the First Fleet diarists and artists), fishing is the most frequently mentioned subsistence activity of the local Aboriginal people. Only two methods are described and illustrated: spear-fishing and angling. First Fleet documents state that spear-fishing was undertaken by men using multi-pronged spears (often called 'fizz-gigs' or 'gigs') from the rocky shores as well as from bark canoes and in shallow waters. Angling (or line-fishing) was carried out by women who fished from canoes using shell hooks and lines in deep water (Bradley 1786-1793 [1969]: 133; Collins 1798 [1975]: 461; HR.NSW 1893 [1978]: 309; Hunter 1793 [1968: 63]; Tench 1979: 210,285-7 [1793: 96, 193-6]; Worgan 1788 [1978]: 16, 37; see also Lawrence 1968: 143-7, 195-6). The gender division in fishing was not absolute and Tench (1979: 287 [1793: 195]), for example, noted that: 'women sometimes use the gig, and always carry one in each canoe, to strike large fish which may be hooked, and thereby facilitate the capture'. The only reference to men using hooks and lines is Govett's (1837: 7-8) some 50 years after contact; as Bowdler (1976: 253) points out, by that time European fishing tackle was being given to the Aborigines and Govett's description is of an Aboriginal person using European fishing tackle.

The few references to nets in Port Jackson suggest they were not used to catch fish: Stockdale (1789 [1950]: 136-7) describes a 'small net' which 'appears to have been used either as a landing net, or for the purpose of carrying the fish when taken' and 'small hoop nets in which they catch lobsters, and sea crayfish' (see also HR.NSW 1893 [1978]: 132). Tench (1979: 47 [1789: 79]) refers to 'small nets, in which they put the fish they catch'. The dimensions of these 'small' nets are not given. Seine-fishing with large nets and other methods of fishing (e.g. weirs and fish traps), as far as the historical records are concerned, do not appear to have been used in the Sydney region at contact (Lawrence 1968: 144-7). Fish traps of stone or brush are not mentioned at all.

This picture of spear-fishing and angling being the only methods that the Aboriginal people of Port Jackson used to catch fish has continued to be repeated in recent descriptions of Aboriginal life for the Sydney district (Kohen 1993: 23; Kohen & Lampert 1987: 352-4; Lampert 1988; Lampert & Konecny 1989; Lampert & Megaw 1979: 67-8; Turbet 1989: 53-6).

Evidence from other parts of the NSW coast

In other parts of Australia historical and ethnographic accounts describe a wide range of fishing methods: not only spear-fishing and angling, but also tidal weirs and traps, communal drives and a variety of nets and poisons (Bowdler 1976: 249-50; Happ 1977; Lawrence 1968). Archaeological evidence (e.g. stone fish traps and the size of fish represented in excavated faunal assemblages) has also been used to indicate prehistoric fishing methods.

Documentary and archaeological evidence for the NSW north and south coasts indicate that in addition to line- and spear-fishing, the methods described below were also used to catch fish, particularly in tidal and estuarine situations.

Documentary evidence

To the north and south of Port Jackson, the use of poisons as well as brush and stone weirs or traps have been described (Attenbrow 1976; Brayshaw 1986: 77; Campbell 1978: 122-34; Coleman 1982: 5-6, figure 9; Enright 1935: 8-9; Lampert & Sanders 1973: table 2; Lawrence 1968: 146-7, table 7; Robinson 1844: 233-4; Robinson in Mackaness 1941: 336; Vinnicombe 1980: V: 2). South of Port Jackson, the historically described weirs were all made of brush; no stone traps are described for that region.

Fishing nets were used on the NSW north coast (McBryde 1974: 12-13), and in the Hunter Valley, near Newcastle, hand nets were used in shallow water (Brayshaw 1986: 77). However, indisputable references to the pre-contact use of fishing nets by Aboriginal people have not been found for the NSW south coast (Sullivan 1982: 54). A drawing by Brierly (1842-44) shows a bark canoe at Twofold Bay with what appears to be a net, but this dates from the early 1840s when the whaling station was well established and the net could be a European item.

Archaeological evidence

Archaeological evidence from other parts of the NSW coast [ILLUSTRATION FOR FIGURE 1 OMITTED] comes in two forms:

1 stone-built fish traps

2 excavated fish remains.

Several stone-built fish traps have been reported [ILLUSTRATION FOR FIGURE 1 OMITTED]. All of the stone-built tidal fish traps are north of Port Jackson (Campbell 1978: 122-34; Coleman 1982: 5-6, figure 9; Enright 1935: 8-9; NSW National Parks and Wildlife Service Aboriginal Sites Register). The two tidal traps nearest Port Jackson are south of Newcastle. At Norah Head is said to have been a double fish trap, destroyed when a swimming pool was built. The other is in the inter-tidal zone of a small tributary of Broken Bay, where a series of rocks 'appears to have been arranged purposefully' 'and form a loose arrangement of enclosure' [sic] (NSW National Parks and Wildlife Service Aboriginal Sites Register). A stone-built fish trap was also reported in the freshwater reaches of the Nepean River at Castlereagh (McCarthy 1948: figure 1; Sim in Campbell 1978: 122).

Analysis of excavated fish remains from coastal archaeological sites to the north of Port Jackson (Macleay estuary and Barrenjoey peninsula) and on the NSW south coast (Pambula) suggest that in those coastal areas traps and/or nets were used in addition to spear-fishing and angling (Coleman 1980: 73; Sullivan 1982: 240-41; 1984: 12; Wood 1989: 65-70; 1992: 173).

Historical and archaeological evidence from these adjacent areas thus supports the view that tidal traps, and possibly nets, are likely to have been used in Port Jackson (cf. Lawrence (1968: 147).

Port Jackson archaeological evidence

The Port Jackson archaeological fish assemblage used for this study comes from a shell midden in a rock-shelter called Mt Trefle (Attenbrow 1992a; 1992b). The rock-shelter is typical of the Port Jackson area, small to medium-sized measuring 20 m x 5 m in area, with a maximum height of 1.9 m [ILLUSTRATION FOR FIGURE 2 OMITTED]. Shell midden has accumulated within the rock-shelter and extends down the slope in front of the shelter for about 8 m.

The rock-shelter is not far from the harbour mouth on a forested sandstone ridge-side [ILLUSTRATION FOR FIGURE 1 OMITTED] about 30-35 m above sea level and c. 175 m from the shoreline. The local shoreline within a kilometre of the site includes extensive rock platforms, small sandy beaches and bays. A deep-water channel and shallow bay-waters are close by.

The excavated deposits

The excavated trench, 1.75 sq m in area, included deposits both inside and outside the rock-shelter [ILLUSTRATION FOR FIGURE 2 OMITTED]. Within the shelter the midden is shallow (=/[less than]30 cm), and directly overlies sandstone bedrock. The bedrock drops down sharply beneath the dripline, and outside the shelter the deposit is c. 70 cm deep.

The excavated deposits contained shell, stone artefacts, bone and charcoal in varying quantities, The inside deposits consist of relatively well-defined layers of dry, firm to very firm sandy sediments ranging in colour from buff to brown and grey-brown, An ashy hearth area was revealed in Level 2 in Square Ad. The inside deposits are rich in shell but contain relatively little charcoal. The outside deposits are homogeneous, very dark brown (almost black), sandy sediment rich in charcoal throughout the 70 cm depth. The density of shell is much lower than in the inside squares, except for two shallow concentrations of hairy mussel (Trichomya hirsuta).

Stratigraphic layers inside and outside the shelter cannot be correlated because of differences in their depth and nature of their deposits, and the stratigraphic discontinuity beneath the dripline. Data for both areas are thus presented separately (TABLES 2 & 3). Excavation units inside the shelter have been grouped into three analytical levels based on the stratigraphy. Outside the shelter 11 analytical levels were established based on the spits, Excavation methods are described in Attenbrow (1992a), but it is relevant to note here that nested 7-mm and 3-mm meshes were used to sieve the deposits, and that materials retained in the 3-mm sieves were sorted under laboratory conditions.

Radiocarbon dates and length of Aboriginal occupation

Radiocarbon dates (TABLE 1) indicate that occupation of the site began around 1300 years ago and continued until the contact period. Aboriginal people may have continued to use Mt Trefle rock-shelter in the immediate post-contact period as Aboriginal groups continued to live in the South Head area until 1819 at least (Murray 1909). However, artefactual evidence to support this proposal has not been identified in the site.

The faunal assemblage

The fish remains analysed for this study are only part of the faunal assemblage from Mt Trefle which includes shell and crustacea, as well as bone from terrestrial mammals, birds [TABULAR DATA FOR TABLE 1 OMITTED] and reptiles. The bone was not evenly distributed throughout the excavated deposit, and both cultural and non-cultural processes contributed to its vertical and horizontal distribution (Steele 1992: 13-19). Outside the shelter, though there are noticeable 'peaks' in the occurrence of fish bone in Levels 5 and 7 (TABLE 2), an overall decrease in fish bone with depth (particularly evident in Levels 9-11) corresponds with an increase in the proportion of burnt bone. This distribution pattern suggests that more bone was originally present in the lower levels of the deposit and the present pattern is due to preservational factors.

The total bone assemblage comprises 16,224 pieces of whole and fragmented bone with a total weight of 371 g. Much of the faunal assemblage is highly fragmented and approximately 40% by weight is unidentified fragments (10,577, weighing 150 g). Few bones exceed 5 cm in maximum dimension, and most were retrieved from the 3-mm sieve. Even whole bones are small in size.

The small size of the fragmented bone is likely to have resulted from the nature of the occupation and/or site maintenance activities in the area excavated. Large and obtrusive refuse may have been either discarded elsewhere or removed from the area excavated. However, the smallness of the fragments and the fact they appear to have been reduced to a stable size is commensurate with assemblages commonly found in areas which have been the focus of activity (Schiffer 1987: 129). The area excavated was in the optimum habitable area of the shelter which would have received maximum treadage and use. Scavengers appear to have contrifuted little to the accumulation, fragmentation and spatial configuration of the sample (Steele 1992: 14-15).

TABLE 2. Mt Trefle. Distribution of fish bone according to level.

 proportion
 density of total bone
 weight (g/kg of assemblage
level (g) deposit) % (weight)

outside

1 7.82 0.09 21.3
2 14.12 0.16 30.8
3 7.97 0.21 29.0
4 8.23 0.21 22.1
5 13.09 0.31 42.5
6 2.34 0.06 15.2
7 14.05 0.30 33.1
8 5.83 0.09 26.5
9 1.58 0.02 26.9
10 0.67 0.02 43.5
11 0.02 [less than]0.01 6.1

total 75.72

inside

1 41.07 1.78 67.1
2 7.27 0.32 34.5
3 8.60 1.02 37.9

total 56.94

site total 132.66 35.8

inside: Aa, Ab, Ad, Ba, Bb (surf, 1 east, 2-4)

outside: Bb (1 west, 5-8), Bc, Bd, Ca, Cb

The fish assemblage

The identified fish species are listed in TABLE 3.

Quantification and determination of abundance

Quantifying and determining the abundance of faunal remains in a site, whether for the total assemblage or individual species, is often difficult (Colley 1990: 217-18). Often a single [TABULAR DATA FOR TABLE 3 OMITTED] measure will not provide a figure suitable for comparing the abundance of one taxa or species with another. In this article, weight and number of identified specimens (NISP) are used to measure the relative frequencies of taxa and of skeletal elements within each taxa. These two measures were considered appropriate primarily because much of the assemblage was highly fragmented (cf. Wood 1992: 154).

The fish remains consist of 5254 whole and fragmented bones weighing 133 g(1) (TABLE 2). They comprise 36% by weight, 32.5% by count and 92% by NISP of the bone assemblage, and are a significant component of the identified faunal remains (approximately 60% by weight).

The identified fish remains consist of 304 skeletal elements (almost 6% of the fish component by number) which were identified to taxa or species level wherever possible. The wide number of species identified (TABLE 3) suggests that fish played an important part in the economy of the site's inhabitants.

Remains belonging to the family Sparidae (snapper, bream and tarwhine) dominate the identified fish component (76% by NISP; TABLE 3), with snapper (52% NISP) being the most frequently identified. Snapper is possibly over-represented because of the large number of loose molars which were identified in comparison to other skeletal elements. Bones from leatherjacket and wrasse represent the bulk of the remaining identified specimens. Leather-jacket was identified exclusively upon dorsal spine fragments. The remaining fish species, which include flathead, whiting, groper and mulloway, are represented by only a small number of specimens. Other fish species which inhabit Port Jackson (e.g. garfish, mullet) may be absent due to the fragility of their skeletons and differential preservation factors rather than the fact they were not eaten.

Differential survival, or the small size of the identified component in some levels, may account for the difference in particular fish categories between excavation contexts inside and outside the shelter (TABLE 3).

[TABULAR DATA FOR TABLE 4 OMITTED]

Ascertaining fishing methods

Our approach in investigating fishing methods used in Port Jackson is based on the view that certain fishing methods and strategies result in the capture of fish of particular species and/or particular sizes (TABLE 4). Leach (1979), Balme (1983: 24), Colley & Jones (1987) and Colley (1990: 218-22) previously used this approach to infer prehistoric fishing methods or the use of particular types of fishing gear from excavated fish bone assemblages. Inferences are based on the estimated size and age range (derived from selected skeletal elements in a reference collection - see below) and the known habitat of the fish (Balme 1983; Bowdler 1970; 1976; Coleman 1980; Colley 1983; Coutts 1975: 270 in Colley 1990; Kefous 1977 in Coleman 1980; Sullivan 1984: 240-41; Wheeler & Jones in Shackley 1981: 185-6; Wood 1989; 1992: 167-8).

Making inferences upon such data is not straightforward. Coleman (1980), Colley (1987; 1990: 222) and Owen & Merrick (1994a) question such methods. Coleman and Colley are critical because 'different fishing methods often produce a similar catch composition, and depending on when and where it is used, the same fishing gear may produce a different catch.'

Owen & Merrick (1994a: 15), in contradiction to the anticipated catch characteristics set out in TABLE 4, believe that nets and traps would be non-selective in terms of species captured, spearing would be most selective of all methods, and angling semi-selective. However, their discussion on size selectivity is limited to the use of fish hooks.

We acknowledge that problems exist with this approach, but consider our conclusions are valid since we have not attempted to be too specific in our inferences.

Age structure and habitats

The age structure of fish populations in different habitats depends upon factors such as the growth rate and breeding patterns of individual species as well as the degree of predation (including fishing by humans). The following information on fish species in the Mt Trefle assemblage illustrates this point.

Snapper, Pagrus auratus, change their behaviour and preferred habitat according to age (Roughley 1961: 77-8). Young (sexually immature) snapper (referred to as 'cockneys' and 'red bream') tend to school in age classes in shallower estuarine waters and bays where they feed and grow before moving out into deeper offshore waters. Sexually mature snapper are essentially bottom-dwelling fish and generally solitary; they occur in greatest numbers over rocky reefs and on shell or gravel bottoms. At sexual maturity snapper are approximately 230 mm standard length (State Pollution Commission 1981: 108).

Mulloway, Sciaena antarctica, are found in rivers up to and beyond the tidal limit; when schooling they congregate near estuary mouths and pass beyond into the surf zone of coastal beaches (Roughley 1961: 70). Yellowtail kingfish, Seriola lalandi, observed in both offshore waters and estuaries, are frequently caught about rocky headlands and reefs, in the surf zone, and in estuaries and sheltered bays (Roughley 1961: 54). Eastern blue groper, Achoerodus viridus, generally restricted to ocean foreshores, are commonly caught around rocky coastal reefs and headlands (Roughley 1961: 102; Thomson 1978: 117).

Numerous species of leatherjacket (Monacanthidae), wrasse (Labridae) and cod (Serranidae) occur along the NSW central coast and it was not possible to identify these remains more specifically. Leatherjackets, wrasse and cod inhabit both offshore and estuarine waters in a variety of habitats. Tarwhine, Rhabdosargus sarba, are also found in both estuarine and offshore waters, but large tarwhine seldom enter the estuaries (Thomson 1978: 96).

Of the remaining fish species identified within the Mt Trefle assemblage, the majority are largely estuarine in habitat (such as yellow fin bream, Acanthopagrus australis) and/or prefer shallow water conditions (e.g. dusky flathead, Platycephalus fuscus, and sand whiting, Sillago ciliata). Yellowfin bream, with slow growth rates, are sexually mature when about 210-230 mm long (4 years old). They are found in coastal and estuarine waters but can inhabit estuaries throughout their life cycle; as bottom feeders, they move about the creeks and rivers in scattered shoals (Roughley 1961: 823). Sand whiting and dusky flathead are both adapted to life on the bottom and inhabit shallow waters over sand-flats (Roughley 1961: 46, 132, 136-7). Sand whiting inhabit sand-flats and sandy channels near estuary mouths and periodically the surf zone. Spawning appears to occur near estuary mouths and the surf zone during summer (Roughley 1961: 46). Dusky flathead move in the summer months from the creeks and upstream flats towards the estuary mouth to spawn (Roughley 1961: 136).

Estimating fish size and age ranges for the Mt Trefle assemblage

Analytical methods used

Individual fish are usually sized by comparing selected bone measurements from archaeological specimens with measurements from reference collection specimens (Casteel 1976; Owen 1984; Owen & Merrick 1994b; Shackley 1981: 182-4). In our study, the size and age range of the fish-in the faunal assemblage were calculated using comparative data from previous studies.

Few detailed studies of comparative collections have examined the relationship between bone size and fish length for species commonly identified from Australian coastal archaeological sites. Relevant studies have focused on snapper (Bowdler 1970; Owen 1984; Owen & Merrick 1994b; Wood 1989; 1992; see details below), though some data are available for bream and flathead (Bowdler 1970; Coleman 1980).

Owen (1984: 18, table 10; Owen & Merrick (1994b: 5)) recorded 22 standard measurements on a suite of snapper (Pagrus auratus) bones using a reference collection of 42 individuals graded for an even size-range. Statistical methods tested which bones were the best predictor of fish size (length) and produced the standard estimate of error in each prediction. Owen (1984: 24) demonstrated that each of the bone measurements she used has a high correlation with fish length.

Coleman (1980) provides comparative data for bream supra-occipitals and flathead dentaries, but no bream supra-occipitals were identified in the Mt Trefle assemblage, and the few flathead dentaries are too fragmentary to use for estimating fish length with accuracy. A number of diagnostic skeletal elements (particularly otoliths) from other fish species are complete enough to allow measurement, but comparative data for these species were not available (Serranidae, Labridae, Platycephalus fuscus, Pseudolabrus tetricus, Rhabdosargus sarba, Sciaena antarctica, Sillago ciliata, Sillago sp.).

Our analysis and discussion therefore concentrate on the dominant portion of the Mt Trefle identified sample: snapper and, to a lesser extent, bream.

[TABULAR DATA FOR TABLE 5 OMITTED]

Snapper (Pagrus auratus)

The three most commonly occurring skeletal elements were measured: dentaries, pre-maxillae and supra-occipitals. The level of integrity of the skeletal elements included in the analysis (complete or slight damage), the measurement used (breadth, height, length, process length), as well as the measurements recorded for each skeletal element, and the derived estimates of standard fish length for each are provided in TABLE 5.

Size estimations calculated for snapper specimens are based on Owen (1984) and Owen & Merrick (1994b). Wherever possible, two measurements were recorded for each specimen: for dentary, length (L) and height (H); for pre-maxillae, process length (PL) and length (L) ; for supra-occipital, process length (PL) and breadth (B). This procedure enables two independent estimates of standard fish length (SL) to be gained from the one bone, and minimizes the loss of information through damage. It was possible to calculate two estimates of fish length for only a few specimens. Where only a single calculation for a specimen with minor damage was made, care was taken to ensure the specimen had no missing bone portions. Measurements were recorded using digital callipers in millimetres to two decimal places, according to the methods, locations and orientations illustrated by Owen (1984: 18, plates 2-6, figure 3; Owen & Merrick 1994b: figure 1). Estimates of standard fish length were calculated using the formula Y = a + bx in conjunction with the data provided by Owen from the reference collection, where Y is standard fish length, a is the constant, b the exponent and x the bone measurement (Owen 1984; Owen & Merrick 1994b: 5).

Bream (Acanthopagrus australis)

Data for a single sexually mature and relatively large black bream reference specimen with a gutted weight of 450 g and a fork length of 320 mm are provided by Bowdler (1970). Length of the teeth-bearing surfaces of the pre-maxilia and dentary measured 24 mm and 25 mm respectively. Length estimates based on a single reference specimen are statistically unsatisfactory (Owen 1984; Owen & Merrick 1994b). However, comparison of Bowdler's values, the information on the size of bream at sexual maturity, and measurements recorded for two bream dentary in the Mt Trefle assemblage (TABLE 5) allows some assessment of the general size of the fish from which the Mt Trefle dentaries derive.

Results

Estimated fish lengths for the Mt Trefle snapper range from 66 mm to 550 mm. Two distinct size groupings are apparent: 66 mm-289 mm (27 estimates); 413 mm-550 mm (6 estimates) [ILLUSTRATION FOR FIGURE 3 OMITTED]. Most length estimates in the first group (24 out of 27) are under 230 mm and may be considered sexually immature fish. In addition, many of the fragmentary snapper elements not suitable for metrical analysis could be seen to come from small individuals.

Mt Trefle bream elements appear, in general, to derive from individuals less than 320 mm in length, and to be mostly small, sexually immature individuals (i.e. [less than]230 mm). They closely parallel the smaller size range (150 mm-250 mm) of bream encountered by Coleman (1980).

The diminutive size of the sparid jaw elements not identified beyond a family level complement the evidence provided by the estimated size range for snapper and bream: the Mt Trefle fish assemblage is dominated by a high frequency of small juvenile fish of the family Sparidae.

The small and sexually immature individuals in Mt Trefle normally inhabit, and hence are more likely to be caught in, shallow water estuarine habitats. The remainder of the assemblage supports the conclusion that these fish came largely from estuarine environments.

Wrasse and groper, a minor component of the assemblage, as well as adult snapper, perhaps reflect the proximity of the shelter to the harbour mouth or rocky reefs such as nearby Bottle and Glass Point.

Comparative studies

Wood's (1989: 65-70, 1992: 167-8) analysis showed that the Angophora Reserve snapper remains were dominated by the bones of smaller ([less than]280 mm in length) juvenile fish. In comparison, Owen's (1984; Owen & Merrick 1994a) re-analysis of the Bass Point and Currarong fish remains indicate a more restricted and larger size range than those from either Mt Trefle or Angophora Reserve. Snapper from Bass Point Upper Midden range in size between 161 mm and 551 mm, while those from the pre-fish hook levels range between 180 mm and 556 mm. Analysis of variance shows no significant size difference between snapper in the Upper Midden and those in pre-fish-hook levels (Owen 1984: 44). The mean size of snapper in the Currarong shelters varies between 335 mm and 412 mm in Shelter 1 and 278 and 499 mm in Shelter 2, with most being between 300 mm and 400 mm long (Owen 1984: 46, 60, 84).

The pattern at Currarong and Bass Point suggest that the majority of fish were adults. However, judging the proportion of adult to juvenile individuals in Bass Point and Currarong in relation to the proportions in Mt Trefle and Angophora Reserve is made difficult by the different mesh size in the sieves used at each site. At Mt Trefle and Angophora Reserve 3-mm mesh was used, but at Bass Point 3/16-inch (4.8-mm) mesh (Bowdler 1970: 34, 78). The mesh used at Currarong, not reported, was probably the same as at Burrill Lake - 3/16-inch (4.8 mm) (Lampert 1971: 7).

Inferred fishing methods

The size of fish sought, as well as the behaviour and preferred habitat of each species, is likely to have played a significant role in the fishing methods used by Mt Trefle's inhabitants, especially for fish such as snapper whose behaviour and preferred habitat change with age.

Lampert (1988: 43) considers that the predominant species in some of the NSW central and south coast sites can be correlated with fishing methods and gender divisions: '... women caught most snapper from canoes with their long lines, while men took mainly bream using their spears either from canoes or from the shore'.

Radiocarbon dates for the deposits (TABLE 1) show that the site was occupied during the period when fish-hooks were in use along the NSW south coast (i.e. the last 800 years, Sullivan 1987: 98). Neither fish hooks nor stone files were retrieved during the excavations, but a number of Turbo torquata shell fragments which may derive from fish-hook manufacture and two bone unipoints were found in Levels 1, 2 and 3 (Attenbrow 1992: 14; Fullagar & Szpak 1992: 5; specimens 142, 143, 145-152). We believe spears and line-fishing were used principally to catch the minor component of adult and/or deeper water fish in the Mt Trefle assemblage.

According to TABLE 4, the Mt Trefle fish remains do not match the anticipated catch characteristics for spear- or line-fishing with their size selection towards larger fish. The Mt Trefle assemblage has a wide range of species and is characterized by numerous small sexually immature individuals (especially sparids), which inhabit shallow sheltered estuarine waters. The assemblage strongly suggests tidal traps or some form of net were used in addition to spears and line-fishing. It is probable that fish were also collected from tidal rock pools (Dyall 1982: 58; Vinnicombe 1980: V: 3).

Discussion and conclusions

The foregoing discussions relate to fish remains from one site, Mt Trefle, on the shores of Port Jackson. Previously excavated Port Jackson middens either contained very little fish bone or were excavated at a time when such evidence was not systematically retrieved. The Mt Trefle fish remains are thus the first assemblage available for Port Jackson on which size estimates can be based.

Conclusions based on the analysis of the Mt Trefle assemblage raise doubts about the reliability of the First Fleet records to provide a full description of the fishing methods used in Port Jackson prior to and/or at contact.

Archaeological evidence and historical documents both provide information about the past. Each set of information is formed in a different way and may portray different aspects of human behaviour; neither has primacy over the other (Knapp 1992: 2). Neither can be taken at face value and each needs interpreting and assessing for biases so that valid descriptions of a past culture may be portrayed. In many instances, historical documents are accepted as portraying a more complete picture than the archaeological evidence. In this case, however, the archaeological evidence questions the reliability of the historical descriptions, suggesting they are incomplete about a particular aspect of Aboriginal life.

It is probable that some small fish were caught with spears and fish-hooks even though the intent may have been to catch large mature individuals. Still, the dominance of small, juvenile individuals in the Mt Trefle assemblage suggests a wider range of fishing methods was used in Port Jackson than described in the historic records, i.e. that tidal traps or nets were used as well as spears and hooks and lines.

If large fishing nets (e.g. drum or fixed gill nets, or seine nets) were used in Port Jackson, surely the British settlers would have seen them lying in the Aboriginal campsites or being carried about, even if they had not seen them being used. It is possible that large nets were not used in Port Jackson because of its generally deep waters and the relatively small number and size of the areas of shallow water which occur. It is also possible that large nets had been used in the past, but their use had ceased shortly before contact. However, our preferred explanations are that:

* the small nets referred to in the historical accounts, e.g. scoop nets, were used for catching fish in shallow waters as well as landing and carrying them; and/or

* tidal traps, weirs and rock pools were used. If tidal traps and weirs were used they may not have survived for very long after traditional fishing activities ceased if many (or all?) were built of perishable organic materials such as wood, branches and/or brush. If built of stone, they may not have survived recent human activities or the rough seas of southeastern Australia. Yet the question remains, why was the use or presence of such traps or weirs not reported in the historical accounts of Port Jackson?

It may be that the use of traps and weirs was not observed by the early settlers, as they were used infrequently and/or they were built in more closed and less visible estuarine and bay settings, e.g. on small beaches or at the tidal mouths of small streams running into the estuary where a dense tree canopy often exists. Lack of visibility has also been put forward for the scarce historical references to some other activities, for example, collecting plant foods and hunting land animals. Many women's activities (except fishing in canoes on the harbour) may have had very low visibility as much of the plant and shellfish gathering would have taken place away from the area of (British) settlement (Bowdler 1976: 253), and hunting land animals may simply have been less visible in timbered landscapes than fishing on open estuarine waters (Lampert 1971a: 63; Attenbrow 1988: 47; McDonald 1992: 135). Collecting fish caught in tidal traps or rock pools may have been embedded within women's shellfish gathering activities. However, men may have also collected fish from pools when they were spear-fishing from rock platforms.

Another factor that may affect 'visibility' or what is written is the perception of the observer or recorder as to what is normal or unusual, or expected. The First Fleet diarists were not trained ethnographers and, except for descriptions made during Cook's earlier visit (Beaglehole 1955), were not familiar with the indigenous culture of the land they were colonizing. Although Governor Phillip was asked to ensure descriptions of the Aboriginal people and their customs were reported to the English administrators, the observations were usually made during the course of the officers' normal activities. Records by such writers are therefore likely to be biased or incomplete because of such things as the times and places of the observations, the preconceived ideas of the observer or recorder, the fact that the indigenous people only showed certain aspects of their culture to the colonizers or did not show themselves at certain times or in certain situations.

However, assessing the First Fleet records for their biases and omissions and the reasons for them is beyond the scope of this present article.

Acknowledgements. We gratefully acknowledge funding provided by the Australian Institute of Aboriginal and Torres Strait Islander Studies, Canberra. We also wish to thank members of the Australian Museum's Ichthyology Department, particularly John Paxton, Mark McGrouther and Tom Trnski, for their help and advice in identifications and access to their collections. We wish to thank Sarah Colley, Margrit Koettig and Robin Torrence for their comments on drafts of this article, as well as the unnamed 'Port Jackson referee'. We also wish to thank the many volunteers who assisted in the fieldwork and analysis, particularly Frank Sinn and Cheryl Szpak. Advice and help received at all levels was very much appreciated.

1 The low weight of the bone assemblage should not be taken as an indication that the sample size was too small to make valid statements. The number of bones and fragments (5254) is a better indication of the size of the assemblage. The weight of the bone reflects the fact that fish bones are extremely light and indicates that, indeed, the assemblage contains the bones of a large number of very small fish.

References

ATTENBROW, V.J. 1976. Aboriginal subsistence economy on the far south coast of New South Wales. Australia. Unpublished BA (Hons) thesis, Department of Anthropology, University of Sydney.

1988. Research into the Aboriginal occupation of the Hunter's Hill Municipality. Unpublished report to the Hunter's Hill Municipal Council, Sydney.

1991. Port Jackson archaeological project: a study of the prehistory of the Port Jackson catchment, New South Wales. Stage I - site recording and site assessment, Australian Aboriginal Studies 1991(2): 40-55.

1992a. Port Jackson Archaeological Project - Stage II: Report on work carried out between January 1990 and 30 June 1992. Unpublished report to Australian Institute of Aboriginal and Torres Strait Islander Studies, Canberra.

1992b. Shell bed or shell midden, Australian Archaeology 34: 3-21.

BALME, J. 1983, Prehistoric fishing in the lower Darling, western New South Wales, in Grigson & Clutton-Brock 1983: 19-32.

BEAGLEHOLE, J.C. (ED.) 1955. The journals of Captain James Cook - the voyage of the Endeavour 1768-1771. London: Cambridge University Press.

BINFORD, L.R. 1983. In pursuit of the past. New York: Thames & Hudson.

BOWDLER, S. 1970. Bass Point: The excavation of a south east Australian shell midden showing cultural and economic change. Unpublished BA (Hons) thesis, Department of Anthropology, University of Sydney.

1976. Hook, line and dilly bag: an interpretation of an Australian coastal shell midden, Mankind 10(4): 248-58.

(Ed.). 1982. Coastal archaeology in eastern Australia. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National University.

BRADLEY, W. 1786-1792 [1969]. A voyage to New South Wales, the journal of Lieutenant William Bradley, RN of HMS Sirius 1786-1792. Sydney: Ure Smith.

BRAYSHAW, H. 1986. Aborigines of the Hunter Valley. Scone: Scone & Upper Hunter Historical Society. Bicentennial Publication 4.

BRIERLY, O.W, 1842-44. Journal of a visit to Twofold Bay, Maneroo and districts beyond the Snowy River. Unpublished manuscript A535 in Mitchell Library, Sydney.

CAMPBELL, V.M. 1978. Two fish traps located on the mid-north coast of NSW, in I. McBryde (ed.), Records of times past: 122-34. Canberra: Australian Institute of Aboriginal Studies.

CASTEEL, R.W. 1976. Fish remains in archaeology and palaeoenvironmental studies. London: Academic Press.

COLEMAN, J. 1980. Fish bones for fun and profit, in I. Johnson (ed.), Holier than thou: 61-75. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National University.

1982. A new look at the north coast: fish traps and 'villages', in Bowdler 1982: 1-10.

COLLEY, J. 1983. Interpreting prehistoric fishing strategies: an Orkney case study, in Grigson & Clutton-Brock 1983: 157-71.

COLLEY, S.M. 1987. Fishing for facts: can we reconstruct fishing methods from archaeological evidence? Australian Archaeology 24: 16-26.

1990. The analysis and interpretation of archaeological fish remains, Archaeological Method and Theory 2: 207-53.

COLLEY, S.M. & R. JONES. 1987. New fish data from Rocky Cape, north west Tasmania, Archaeology in Oceania 22: 67-71.

COLLINS, D. 1798 [1975], An account of the English colony in New South Wales I. Republished. Sydney: Reed in association with Royal Australian Historical Society.

COUTTS, P.J. 1975. Marine fishing in archaeological perspective: techniques for determining fishing strategies, in R.W. Casteel & G.I. Quimby (ed.), Maritime adaptations of the Pacific: 265-306. The Hague: Mouton.

DYALL, L.K. 1982. Aboriginal fishing stations on the Newcastle coastline, New South Wales, in Bowdler 1982: 52-62.

ENRIGHT, W.J. 1935. An Aboriginal fish trap, Mankind 1: 8-9.

FULLAGAR, R. & C. SZPAK. 1992. Port Jackson Archaeological Project: use-wear and residue analysis of selected artefacts, Appendix 2 in Attenbrow 1992a.

GOVETT, W.R. 1837. Sketches of New South Wales XV. The natives: their method of fishing etc., Saturday Magazine X (290): 7-8.

GRIGSON, C. & J. CLUTTON-BROCK (ED.). 1983. Animals and archaeology 2: Shell middens, fishes and birds: Oxford: British Archaeological Reports. International series S183.

HAPP, G. 1977. To catch a fish ...! the aquatic factor in Aboriginal fishing technology. BA(Hons) thesis, Department of Anthropology, University of Sydney.

HISTORICAL RECORDS OF NSW. 1893 [1978]. Volume I, Part 2 - 1788-1792. Sydney: Government Printer.

HUNTER, J. 1793 [1968]. An historical journal of the transactions at Port Jackson and Norfolk Island, with further accounts by Governor Phillip. London: John Bach. [Facsimile, Adelaide: Libraries Board of South Australia.]

HUTCHINS, B. & R. SWAINSTON. 1986. Sea fishes of southern Australia. Perth: Swainston.

KEFOUS, K. 1977. We have fish with ears and wonder if it is valuable. Unpublished BA (Hons) thesis, Department of Prehistory, Australian National University, Canberra.

KNAPP, A.B. 1992. Archaeology and Annales: time, space, and change, in A.B. Knapp (ed.), Archaeology, Annales, and ethnohistory: 1-21. Cambridge University Press.

KOHEN, J. 1993. The Dharug and their neighbours. Sydney: Darug Link in association in Blacktown and District Historical Society.

KOHEN, J.L. & R.J. LAMPERT. 1987. Hunters and fishers of the Sydney region, in D.J. Mulvaney & I.P. White (ed.), Australians to 1788: 343-65. Sydney: Fairfax, Syme & Weldon Associates.

LAMPERT, R.J. 1971a. Burrill Lake and Currarong. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National University, Terra Australis 1.

1971b. Coastal Aborigines of southeastern Australia, in D.J. Mulvaney & J. Golson (ed.), Aboriginal man and environment in Australia: 114-32. Canberra: Australian National University Press.

1988. Aboriginal life around Port Jackson, 1788-92, in B. Smith & A. Wheeler (ed.l, The art of the First Fleet and other early Australian drawings: 19-69. Melbourne: Oxford University Press in association with Australian Academy of the Humanities' and British Museum (Natural History).

LAMPERT, R.J. & T.A. KONECNY. 1989. Aboriginal spears of Port Jackson type discovered - a bicentennial sequel, Antiquity 63: 137-41.

LAMPERT, R.J. & V. MEGAW. 1979. Life around Sydney, in P. Stanbury (ed.), 10,000 years of Sydney life: 65-71. Sydney: The Macleay Museum, University of Sydney.

LAMPERT, R.J. & F. SANDERS. 1973. Plants and men on the Beecroft Peninsula, N.S.W., Mankind 9(2): 96-108.

LAWRENCE, R.J. 1968. Aboriginal habitat and economy, Canberra: Department of Geography, Australian National University. Occasional Paper 6.

LEACH, B.F. 1979. Fish and crayfish from the Washpool midden site, New Zealand: their use in determining season of occupation and prehistoric fishing methods, Journal of Archaeological Science 6: 109-26.

MACKANESS, G. 1941. George Augustus Robinson's journey into southeastern Australia, 1844, Journal of the Royal Australian Historical Society 27(5): 318-49.

MCBRYDE, I. 1974. Aboriginal prehistory in New England. Sydney: Sydney University Press.

MCCARTHY, F.D. 1948. The Lapstone Creek excavation: two culture periods revealed in eastern New South Wales, Records of the Australian Museum 22: 1-34.

MCDONALD, J. 1992. The archaeology of the Angophora Reserve rock-shelter. Sydney: NSW National Parks and Wildlife Service. Environmental Heritage Monograph Series 1.

MEEHAN, B. & R. JONES (ed.). 1988. Archaeology with ethnography: an Australian perspective. Canberra: Department of Prohistory, Research School of Pacific Studies, The Australian National University.

MURRAY, A. 1909. Letter dated 18 February 1909 from Aubrey Murray, Compiling Branch, Lands Office, to the Australian Museum. Australian Museum Archives, Series 9 correspondence, M7/1909.

OWEN, J.F. 1984. Bones to scale: the interpretation of fish remains from New South Wales coastal middens. Unpublished BA (Hons) thesis, Department of Anthropology, University of Sydney.

OWEN, J.F. & J.R. MERRICK. 1994a. Analysis of coastal middens in southeastern Australia: selectivity of angling and other fishing techniques related to Holocene deposits, Journal of Archaeological Science 21:11-16.

1994b. Analysis of coastal middens in southeastern Australia: sizing of fish remains in Holecane deposits, Journal of Archaeological Science 21: 3-10.

POINER, G. 1976. The process of the year among Aborigines of the central and south coast of New South Wales, Archaeology and Physical Anthropology in Oceania XI(3): 186-206.

ROBINSON, G.A. 1844. Papers of George Augusts Robinson. Journal commencing Wednesday 28 August at Parnbuller and terminating 15 September 1844 Biggah Country. Ms A7035 in Mitchell Library, Sydney.

ROUGHLEY, T.C. 1961. Fish and fisheries of Australia. Sydney: Angus & Robertson.

SCHIFFER, M.B. 1987. Formation processes of the archaeological record. Albuquerque (NM): University of New Mexico Press.

SHACKLEY, M. 1981. Environmental archaeology. London: Allen & Unwin.

SPAULDING, A.C. 1968. Explanation in archaeology, in S.R. Binford & L.R. Binford (ed.), New perspectives in archaeology: 33-9. Chicago (IL): Aldine.

STATE POLLUTION CONTROL COMMISSION. 1981. The ecology of fish in Botany Bay - biology of commercially and recreationally valuable species, Sydney.

STEELE, D.S. 1992. Port Jackson Archaeological Project: a report on the faunal (bone) component from six archaeological assemblages, Appendix 1 in Attenbrow 1992a.

STOCKDALE, J. 1789 [1950]. The voyage of Governor Phillip to Botany Bay, with an account of the establishment of the colonies of Port Jackson and Norfolk Island. Adelaide: Australiana Society.

STUIVER, M. & P.J. REIMER, 1993. Extended 14C database and revised CALIB radiocarbon calibration program, Radiocarbon 35: 215-30.

SULLIVAN, H. 1982. Aboriginal use of the forest environment, Appendix 1 in D. Byrne, The five forests: an archaeological and anthropological investigation. Report to NSW National Parks and Wildlife Service, Sydney.

SULLIVAN, M.E. 1982. Aboriginal shell middens in the coastal landscape of New South Wales. Unpublished PhD thesis, Department of Prehistory and Anthropology, Australian National University, Canberra.

1984. A shell midden excavation at Pambula Lake on the far south coast of New South Wales, Archaeology in Oceania 19: 1-15.

1987. The recent prehistoric exploitation of edible mussel in Aboriginal shell middens in southern New South Wales, Archaeology in Oceania 22(3): 97-106.

TENCH, W. 1979. Sydney's first four years, being a reprint era narrative of the expedition to Botany Bay (1789), and a complete account of the settlement at Port Jackson (1793). Sydney: Library of Australian History.

THOMSON, J.M. 1978. A field guide to the common sea and estuary fishes of non-tropical Australia. Sydney: Collins.

TURBET, P. 1989. The Aborigines of the Sydney district before 1788. Sydney: Kangaroo Press.

VINNICOMBE, P. 1980. Prediction and predilection. A study of Aboriginal sites in the Gosford-Wyong region. Unpublished report to the NSW National Parks and Wild-life Service, Sydney.

WHEELER, A. & A. JONES 1976. Fish remains, in A. Rogerson (ed.), Excavations on Fullers Hill, Great Yarmouth: 212-25. East Anglian Archaeological Report 1, Norfolk Archaeological Unit.

WOOD, V. 1989. Angophora Reserve rockshelter: a faunal analysis. Unpublished B. Litt thesis, Department of Prohistory and Anthropology, Australian National University, Canberra.

WOOD, V.K. 1992. Excerpt from Angophora Reserve rockshelter: a faunal analysis, Appendix I in McDonald 1992.

WORGAN, G.B. 1788 [1978]. Journal era First Fleet surgeon. Sydney: Library Council of New South Wales in association with Library of Australian History, Sydney. William Dixson Foundation Publication 16.