Journal of Archaeological Science (1996) 23, 175–181

Fish Heads, Fish Heads: An Experiment on Differential Bone Preservation in a Salmonid Fish Patrick M. Lubinski Department of Anthropology, University of Wisconsin, 5240 Social Sciences, 1180 Observatory Drive, Madison, WI 53706, USA (Received 24 March 1994, revised manuscript accepted 7 June 1994) An experiment modelling the effects of cooking and soil pH on salmonid bone survival was completed using cleaned Lake Whitefish bones. The experiment tested for preservation differences between head parts and vertebrae and between raw, boiled, and moderately burned bone under both acidic and alkaline conditions. Elements were submerged in aqueous solution for 24-h periods, removed, dried, weighed, and re-immersed. Resulting weight and element loss curves for acidic conditions suggest that (1) head parts are destroyed more quickly than vertebrae, and (2) element destruction increases with heating intensity. Under alkaline conditions, broadly similar trends were observed, but at reduced rates. ? 1996 Academic Press Limited Keywords: SALMONID ARCHAEOFAUNAS, TAPHONOMY, BONE DIAGENESIS, SKELETAL PART REPRESENTATION, COOKING, PH.

fish skeletal elements as a result of different responses of elements to heating and chemical attrition. There is considerable experimental evidence for differences in preservation potential between skeletal elements (e.g. Binford & Bertram, 1977; Lyman, 1984; Von Endt & Ortner, 1984; Butler and Chatters, 1994). Skeletal parts may be destroyed differentially because they vary considerably in size and structure (shape, density, tissue structure, porosity, etc.). Size and structure determine a bone’s inherent mechanical resistance to physical destruction and the surface area it presents to chemical action. For bony fishes, one might expect that thin, flat cranial bones would be destroyed more quickly than cylindrical vertebrae (Rojo, 1987: 210; Wigen & Stucki, 1988: 106) because the cranial bones appear to have less mechanical strength and more surface area. Additionally, cranial elements are lower in volume density (measured in g cm "3) than vertebrae (Butler & Chatters, 1994). Heating of bone is known to weaken the organic phase (DeNiro et al., 1985; Richter, 1986) which gives bone tissue some of its resilience and chemical stability (Currey, 1990). Heat-induced loss of the organic phase will increase the porosity of the mineral phase (McCutcheon, 1992: 354), which will provide more surface area to chemical action, as well as making the bone more susceptible to mechanical breakage. For these reasons, one might expect that moderately burned elements would be destroyed more quickly than unburned elements under archaeological conditions (Stewart, 1989: 172). However, bones subjected to more intense, high temperature burning may be more

Introduction s with mammal assemblages, measures of the relative frequencies between anatomical parts, such as %MAU (Binford, 1984) may provide useful information in the interpretation of archaeological fish assemblages (e.g. Stewart, 1989; Belcher, 1992; Butler, 1993). However, there are a large number of possible explanations for any observed pattern, because there are a number of processes that might affect skeletal parts differentially. Potential sources of differential representation in fish body parts include methodological bias (in recovery, sampling, identification and aggregation), differential disposal (during butchery, transport, preparation and discard), and differential preservation (before deposition as a result of fish bone structure, processing, ingestion and subaerial weathering; and after deposition as a result of biological, mechanical and chemical attrition). The most important sources of differential skeletal part representation will vary among archaeological contexts; however, the preservation of all fish remains will be strongly affected by the thermal history of the remains and the chemistry of their disposal context. While there has been experimental examination of several other mechanisms of fish bone destruction (e.g. Jones, 1986; Nicholson, 1992; Butler & Chatters, 1994), thermal history and disposal context chemistry have not been the subject of much detailed examination (cf Richter, 1986). The experiments described in this paper were designed to determine if there are any natural tendencies towards differential preservation between

A

175 0305-4403/96/020175+07 $12.00/0

? 1996 Academic Press Limited

176 P. M. Lubinski Table 1. Summary of experiments

Beaker Experiment 1: A B C D E F G H Experiment 2: A B C D E F

Solution

Processing

Acid Acid Acid Acid Base Base Base Base

Boiled Boiled Burned Burned Boiled Boiled Burned Burned

Acid Acid Acid Base Base Base

Raw Boiled Burned Raw Boiled Burned

27 27 27 27 27 27 27 27

Bones used

Initial weight* (g)

Head, Head, Head, Head, Head, Head, Head, Head,

3·881, 2·818, 1·436, 2·909, 3·580, 1·986, 2·199, 1·657,

10 10 10 10 10 10

51 51 51 51 51 51 51 51

trunk trunk trunk trunk trunk trunk trunk trunk

Vertebrae Vertebrae Vertebrae Vertebrae Vertebrae Vertebrae

6·956 4·730 2·241 3·178 6·320 3·297 3·424 2·545

0·703 0·877 0·656 1·096 0·706 0·703

Fish specimen no.

Fork length† (cm)

3 4 5 7 2 8 9 11

51 46 38 47 50 44 47 39

6 6 6 10 10 10

41 41 41 45 45 45

*After processing. †Fork length is a measure of fish size from the tip of the snout to the fork of the tail.

stable, because the overall shrinkage and crystal reorganization which take place at high temperatures (Shipman et al., 1984) lower a bone’s surface area. The stability of intensely burned bone has been demonstrated in several recent experimental studies (Knight, 1985; Linse & Burton, 1990: 7). The chemical environment of deposition is widely known to affect skeletal element preservation, principally through the pH of the soil solution (e.g. Gordon & Buikstra, 1981; White & Hannus, 1983). Since conventional wisdom holds that bone is destroyed in acidic environments, the following experiments were designed to test for differential bone destruction under acidic conditions. Additionally, the study models alkaline conditions because recent experiments have shown that fish bone is destroyed in alkaline burial environments as well (Linse & Burton, 1990; Linse, 1992).

Table 2). For Experiment 2, vertebral columns from two fish were removed and split into three sections each to compare preservation under raw versus boiled versus burned heating states. For each segment, 10 vertebrae (five with spines and five without spines) were selected. All bones used in the experiments were cleaned with a toothbrush and dental pick to remove any adhering connective tissue. Skeletal elements were prepared in three different heating states; raw, boiled, and burned. Raw specimens (vertebrae only) were soaked in cool water, then separated and cleaned. Boiled specimens were boiled in tap water for 1 h and then cleaned. Burned specimens were boiled for 1 h (to facilitate cleaning), cleaned, dried, then heated in a muffle furnace at 250)C for 2 h. This oven temperature and duration were chosen to remove a portion of the Table 2. Elements selected for Experiment 1

Methods The study utilized Lake Whitefish (Coregonus clupeaformis), an important prehistoric and modern commercial fish species in the Great Lakes (Heidenreich, 1978: 382, Smith, 1979; Cleland, 1982). This species is a member of the family Salmonidae, which also includes chars, ciscos, salmons, and trouts. Ten filleted whitefish carcasses, ranging from 38 to 51 cm in fork length (tip of snout to fork of tail), were obtained from a local fish market. These were used for two experiments, which are summarized in Table 1. Experiment 1 utilized head and trunk elements from eight fish carcasses to compare preservation of head versus trunk elements and boiled versus burned elements. From each specimen, 27 of the most easily recognized and robust head elements and 51 of the vertebrae were selected to represent the head and trunk regions respectively (see

Part Head*

Region Orbital Otic Basicranial Oromandibular Hyoid

Pectoral Trunk†

Vertebral Caudal

Element and number Frontal (2) Posttemporal (2) Parasphenoid (2) Maxilla (2), dentary (2), articular (2), quadrate (2) Hyomandibular (2), urohyal (1), ceratohyal (2), opercular (2), preopercular (2), subopercular (2), interopercular (2) Cleithrum (1) Total head=27 Thoracic vertebra (31), precaudal vertebra (6) Caudal vertebra (14) Total trunk=51

*Head terms (except pectoral) after Wheeler and Jones (1989: Table 7.1). †Trunk terms after Cannon (1987: Table 2).

Differential Bone Preservation 177 –2

–3

log H2(PO4)– or H(PO4)2–

–4

–5

–6

–7

–8

–9

3

4

5

6 pH

7

8

9

Figure 1. Solubility of hydroxyapatite (after Lindsay 1979: figure 12.10).

organic phase (DeNiro et al., 1985: 5) without causing significant splitting, shrinkage or crystal reorganization (Shipman et al., 1984). Additionally, this temperature was chosen to mimic the core temperature of a small open campfire, such as the experimental fire monitored by Shokler (1991). Two solvents were used in the experiments, one an acidic solution with a pH of 4 and the other an alkaline solution with a pH of 10. Aqueous test solutions were prepared using mixtures and pH values given by Robinson (1976: 133–134). The solutions used were 0·05  potassium acid pthalate with a pH of 4·008, and a buffered solution of borax (sodium tetraborate) and sodium hydroxide with a pH of 10·0. These strong pH values were chosen not to mimic archaeological conditions, but rather to produce rapid results in the laboratory. The operating assumption is that increased acidity and alkalinity over a short, experimental time span can model gross patterns of preservation bias that would be obtained over archaeological time spans under more moderate conditions. These pH values also were chosen to bracket neutral pH (7) and the pH at which bone mineral is the most stable. Hydroxyapatite, the primary mineral of bone, is the most stable (least soluble) under slightly alkaline soil conditions at about pH 7·9 (Figure 1). By choosing solution pH values to either side of the values at which bone is probably the most stable, it was expected that the experimental solutions might serve as general models for acidic and alkaline conditions.

Bone units were placed along with test solution in beakers sealed with Parafilm (to prevent evaporation) and immersed in solution for twelve 24-h periods. For Experiment 1, head and trunk units from each fish were placed into the same beaker with 150 ml of solution. In Experiment 2, each group of 10 vertebrae was placed into a separate beaker with 50 ml of solution. Test solutions were checked with indicator paper daily to detect any pH fluctuations. Beakers were rinsed and filled with fresh solution on days 4 and 9 and on any day when the pH changed by a factor of 1·0. In this way, the experiment models an open system with a continuous ion supply rather than a closed system in which the bone might come into equilibrium with the beaker solution. After each 24-h immersion period, the bones were removed, dried, counted, weighed to the nearest mg, and re-immersed. Experiment 1 bones were oven dried at 80)C for 1 h while Experiment 2 bones were air dried in order not to cook the raw bone. Total drying time from bone removal to re-immersion (including oven drying, counting, and weighing) averaged 3·8 h. The raw bone was inadvertently placed in the drying oven for 10 min on day 1 and should therefore be considered ‘‘minimally roasted’’ rather than raw. Bone decay was measured in terms of both element loss and weight loss per day of immersion. Element loss is directly relevant to zooarchaeological research because bone counts are the basic analytic units used by most faunal analysts (Grayson, 1984). While weight loss may be directly relevant to researchers who employ weights as a basic analytic unit (e.g. Wing & Brown, 1979), it is used here primarily as a proxy measure of preservation potential. Weights were obtained using all fragments identifiable to element each day. Unidentifiable fragments were not included. Due to this practice, bone loss in the experiment can result from chemical weathering or from fracture brought about by wet/dry cycles and handling, which I believe approximate common conditions of preservation and archaeological recovery. Cumulative weight and element loss curves were used to discern any patterns in decay magnitude on a given day or in decay rates over a number of days. All data points are included on the weight loss versus time graphs, except those representing absolute bone gain (negative % bone loss) due to bone moisture. These points were rejected because (1) they are theoretically impossible and (2) the random error is strongly nonhomogeneous, so that the ‘‘true’’ relationship must be a regression line or curve very near or above the collected data points. There is much more random error to overweighing than to underweighing, since bone wetness may increase weight dramatically (in one case, a 0·225 g gain, which represents 29·1% of total weight), while instrumental error for the balance is only &0·0005 g (0·06% of total weight). Clearly, the rejection of negative bone loss values does not remove

178 P. M. Lubinski 60

80

Cumulative weight loss (%)

Cumulative weight loss (%)

100

60

40

20

0

0

2

4

6 Time (days)

8

10

12

50 40 30 20 10 0

0

2

4

6 Time (days)

8

10

12

Figure 2. Experiment 1 acidic solution weight loss. (/) boiled; (-) burned; (——) head; (– –) trunk.

Figure 3. Experiment 1 basic solution weight loss. (/) boiled; (-) burned; (——) head; (– –) trunk.

the effect of bone moisture. There is no way to discern the effect of moisture on any particular weight value. However, the ‘‘dip’’ patterns seen on several graphs at day 4 indicate excess bone moisture on those days relative to the surrounding days. This reflects inadequate drying, so drying time was doubled and oven temperature was increased to 90)C beginning on day 5.

quantify the rates of destruction, linear regressions were calculated on the values for days 5–12 weight loss for each fish specimen (Table 3). Based on the regression slopes, heads are destroyed 1·5–2·5 times as fast as vertebrae among specimens with the same heating state. Also, comparing like skeletal parts, burned units are destroyed 1·4–2·5 times faster than their corresponding boiled units. Under alkaline conditions, weight loss curves show a similar pattern of burned head part destruction as under acidic conditions (Figure 3). Burned head units (the two upper curves) are destroyed the most quickly, but all boiled parts and burned vertebrae appear to decay at about the same reduced rate. Days 5–12 linear regressions (Table 3) show burned head elements are destroyed 1·9–4·2 times faster than burned trunk elements or boiled elements. The patterns obtained from element loss, under both acidic and alkaline conditions, are broadly similar to

Results The results (in terms of weight loss) of Experiment 1 bones under acidic conditions are shown in Figure 2. The graph shows that burned head units clearly decay the fastest and boiled trunk units decay the most slowly. Burned trunk elements and boiled head elements lose weight at about the same rate. Burned heads have the largest weight loss on any given day and have the fastest decay rate based on the curve slope. To Table 3. Linear regressions for Experiment 1 Specimen no.

Part

(Treatment)

r2

Cumulative weight loss (days 5–12, fixed zero intercept):* Acidic solution: 3 Trunk (Boiled) 0·97 4 Trunk (Boiled) 0·79 3 Head (Boiled) 0·89 4 Head (Boiled) 0·94 5 Trunk (Burned) 0·95 7 Trunk (Burned) 0·95 5 Head (Burned) 0·05 7 Head (Burned) 0·93 Cumulative element loss (days 0–12, fixed zero intercept): Acidic solution: 5 Trunk (Burned) 0·71 5 Head (Burned) 0·93

Slope

Specimen no.

Part

(Treatment)

r2

Slope

1·40 2·09 3·13 3·45 3·54 2·92 7·91 6·37

Basic solution: 2 8 2 8 9 11 9 11

Trunk Trunk Head Head Trunk Trunk Head Head

(Boiled) (Boiled) (Boiled) (Boiled) (Burned) (Burned) (Burned) (Burned)

0·69 0·46 0·68 0·51 0·94 0·95 0·79 0·93

1·18 1·10 1·09 1·39 1·02 1·19 2·68 4·31

0·53 3·92

*Days 1 to 4 omitted from weight loss regressions because of inadequate drying in that period. All correlations are significant at the 0·05 level except no. 5 head weight loss (its slope was not used for rate comparisons). The non-linearity of this curve is presumed due to moisture-related weighing error.

Differential Bone Preservation 179

increasing intensity of heating from minimally roasted to boiled to burned conditions. Under alkaline conditions, the burned vertebrae were destroyed the most quickly, but the minimally roasted and boiled vertebrae decay at about the same rate. In terms of element loss, the burned vertebrae in acidic solution (the only bone unit to show any element loss) lost three of the 10 vertebrae by day 12.

Cumulative element loss (%)

50

40

30

20

Discussion and Conclusions 10

0

2

4

6 Time (days)

8

10

12

Figure 4. Experiment 1 acidic solution element loss for fish no. 5 (burned). (——) head; (– –) trunk.

Cumulative weight loss (%)

80

60

40

20

0

0

2

4

6 Time (days)

8

10

12

Figure 5. Experiment 2 acidic and basic solution vertebrae weight loss. (*) raw; (/) boiled; (-) burned; (——) acidic solution; (– –) basic solution.

those obtained from weight loss. By the conclusion of the experiment, burned units had lost up to 44% of their elements, but boiled units had not lost any elements. Additionally, all burned fish lost at least 18% of their head elements, but only one of the four lost any trunk elements. Figure 4 provides an example of element loss for the only specimen that lost vertebrae (one of the burned fish under acidic conditions). Days 0–12 linear regression slopes (Table 3) suggest the head loses elements at 7·4 times the rate of the vertebral column. Weight loss curves for Experiment 2 vertebrae are shown in Figure 5. As one might expect, the elements in acidic solution collectively show greater weight loss than those in alkaline solution. In each solution group, burned bones show the greatest loss. Under acidic conditions, bone destruction is more rapid with

Figure 6, depicting the total weight loss at the conclusion of Experiment 1 under each test condition, can serve as a summary of the results of the experiments. The general patterns in this graph are representative of all experimental results. Before discussing the overall patterns of bone destruction and implications of the study, several limitations should be noted. First, the experiments obviously are not directly analogous to archaeological conditions, because the pH and wet/dry cycles in the study are more extreme than at most archaeological sites. For this reason, the specific values obtained in the experiment and given in Figure 6 are not directly applicable to archaeological conditions and should not be used as ‘‘correction factors’’ for archaeological assemblages. However, if one assumes that fish bone reacts in a similar manner from moderate to intense acidity, alkalinity and wet/dry cycles, then the directions of preservation bias observed in this experiment should be appropriate analogues for Lake Whitefish bone decay under archaeological conditions. Second, while the results of these experiments might be applicable to species other than Lake Whitefish, it is important to note that fish bones vary markedly between taxa (Gregory, 1933; Cannon, 1987; Wheeler & Jones, 1989; Colley, 1990). Significant variations in size, shape, and composition will result in different head versus trunk element destruction patterns and may result in different burned versus unburned patterns. While one might expect that all salmonid fishes would share the general whitefish pattern, salmon (Oncorhynchus spp.) may, for example, exhibit a less-marked difference in destruction between skeletal parts because they have more robust mouth parts than the Lake Whitefish used here. For non-salmonid fishes, the head versus trunk element destruction pattern may be very different than Lake Whitefish, because whitefish and other salmonids have highly cartilaginous skulls compared to other fishes (Gregory, 1933). Experimental results differ between alkaline and acidic conditions. In acidic solution, the effect of heating state on a given skeletal part is increasing destruction from minimally roasted to boiled to moderately burned conditions. These results imply that there will be a bias against the recovery of moderatelyheated fish bones in archaeological sites. (Note that, based on other experiments (Knight, 1985; Linse &

180 P. M. Lubinski

Head-acid

Trunk-acid

Head-base

Trunk-base

0

20

40 Weight loss (%)

60

80

Figure 6. Experiment 1 total weight loss at conclusion of experiment (day 12). ( ) burned; ( ) boiled.

Burton, 1990), intensely burned bone may not follow this pattern.) This may be significant if burning is the common fate of fish bone at an archaeological site. It may also be significant for faunal analysts who use burning to distinguish cultural versus natural bone accumulations or to infer cooking technique. Under alkaline conditions, the effect of heating state is far less clear. There is a strong bias against burned head elements, but vertebrae (in any heating state) and unburned heads appear to have the same preservation potential in alkaline archaeological sites. In light of these results, it would seem wise for faunal analysts to consider burned and unburned fish elements separately because of their different natural potential for destruction at both acidic and alkaline archaeological sites. As for differential preservation between skeletal parts, Experiment 1 indicates that, for a given heating state under acidic conditions, head elements are destroyed more quickly than vertebrae. These results imply that there will be a bias against the recovery of head elements relative to vertebrae in acidic archaeological sites, which has important implications for zooarchaeological studies. The bias should be taken into account when, for example, employing fish skeletal part representation measures (such as %MAU) to infer butchery patterns or bone transport. Under alkaline conditions, the pattern is again far less simple. Head element are destroyed more quickly than vertebrae amongst burned specimens, but there is no strong preservation bias amongst boiled specimens. The fact that these experiments indicate different patterns of destruction under acidic and alkaline conditions has an additional implication for archaeofaunal comparisons. When comparing assemblages from very different sediment pH values, it would be prudent to limit comparisons to like body parts and heating states (for example unburned vertebrae to unburned

vertebrae) in order to avoid attributing simple differences in preservation to culturally-meaningful patterns. These experiments involved only one species under a few pH and heating states. Further experiments might broaden the focus and consider other fish taxa and conditions. I, nonetheless, hope that these experiments have provided useful information for a more rigorous interpretation of archaeological fish remains.

Acknowledgements Earlier versions of this paper were presented at the 58th Annual Meeting of the Society for American Archaeology and written for a seminar given by J. M. Kenoyer at the University of Wisconsin-Madison. The author is indebted to Goeden’s Fish Restaurant for providing the fish used in the experiment, and to J. H. Burton of the Laboratory of Archaeological Chemistry for allowing access to facilities. J. H. Burton, V. L. Butler, C. P. Lipo, C. J. O’Brien, M. J. Schoeninger, and an anonymous reviewer provided helpful comments. I am particularly appreciative of the many suggestions and thoughtful comments given by W. R. Belcher and M. E. Madsen.

References Belcher, W. R. (1992). Fish resources in an early urban context at Harappa. In (R. H. Meadow, Ed.) Harappan Excavations 1986– 1990: A Multidisciplinary Approach to Third Millennium Urbanism. Monographs in World Archaeology No. 3. Madison, Wisconsin: Prehistory Press, pp. 107–120. Binford, L. R. (1984). Faunal Remains from Klases River Mouth. Orlando: Academic Press. Binford, L. R. & Bertram, J. B. (1977). Bone frequencies—and attritional processes. In (L. R. Binford, Ed.) For Theory Building in Archaeology. New York: Academic Press, pp. 77–153. Butler, V. L. (1993). Natural versus cultural salmonid remains: origin of The Dalles Roadcut bones, Columbia River, Oregon, U.S.A. Journal of Archaeological Science 20, 1–24.

Differential Bone Preservation 181 Butler, V. L. & Chatters, J. C. (1994). The role of bone density in structuring prehistoric salmon bone assemblages. Journal of Archaeological Science 21, 413–424. Cannon, D. Y. (1987). Marine Fish Osteology: A Manual for Archaeologists. Publication No. 18. Department of Archaeology, Simon Fraser University, Burnaby, British Columbia. Cleland, C. E. (1982). The inland shore fishery of the northern Great Lakes: its development and importance in prehistory. American Antiquity 47, 761–784. Cleland, C. E. (1989). Comments on ‘‘A reconsideration of aboriginal fishing strategies in the northern Great Lakes region’’ by Susan R. Martin. American antiquity 54, 605–609. Colley, S. M. (1990). The analysis and interpretation of archaeological fish remains. In (M. B. Shiffer, Ed.) Archaeological Method and Theory, Vol. 2. Tucson: University of Arizona Press, pp. 207– 253. Currey, J. D. (1990). Biomechanics of mineralized skeletons. In (J. G. Carter, Ed.) Skeletal Biomineralization: Patterns, Processes and Evolutionary Trends, Vol. 1. New York: Van Nostrand Reinhold, pp. 11–25. DeNiro, M. J., Schoeninger, M. J. & Hastorf, C. A. (1985). Effect of heating on the stable carbon and nitrogen isotope ratios of bone collagen. Journal of Archaeological Sciences 12, 1–7. Gordon, C. C. & Buikstra, J. E. (1981). Soil pH, bone preservation, and sampling bias at mortuary sites. American Antiquity 46, 566–571. Grayson, D. K. (1984). Quantitative Zooarchaeology: Topics in the analysis of Archaeological Faunas. Orlando: Academic Press. Gregory, W. K. (1933). Fish skulls: A study of the evolution of natural mechanisms. Transactions of the American Philosophical Society (New Series) 23, 75–481. Heidenreich, C. E. (1978). Huron. In (W. C. Sturtevant, Ed.) Handbook of North American Indians, Vol. 15: Northeast. Washington D.C.: Smithsonian Institution, pp. 368–388. Jones, A. K. G. (1986). Fish bone survival in the digestive systems of the pig, dog and man: Some experiments. In (D. C. Brinkhuizen & A. T. Clason, Eds.) Fish and Archaeology: Studies in Osteometry, Taphonomy, Seasonality and Fishing Methods. Oxford: BAR International Series No. 294, pp. 53–61. Knight, J. A. (1985). Differential preservation of calcined bone at the Hirundo site, Alton, Maine. MS Thesis, Quaternary Studies, University of Maine, Orono. Lindsay, W. L. (1979). Chemical Equilibria in Soils. New York: Wiley. Linse, A. R. (1992). Is bone safe in a shell midden? In (J. K. Stein, Ed.) Deciphering a Shell Midden. San Diego: Academic Press, pp. 327–345.

Linse, A. R. & Burton, J. H. (1990). Bone solubility and preservation in alkaline depositional conditions. Paper presented at the 55th Annual Meeting of the Society for American Archaeology, Las Vegas. Lyman, R. L. (1984). Bone density and differential survivorship of fossil classes. Journal of Anthropological Archaeology 3, 259–299. McCutcheon, P. T. (1992). Burned archaeological bone. In (J. K. Stein, Ed.) Deciphering A Shell Midden. San Diego: Academic Press, pp. 347–370. Nicholson, R. A. (1992). An investigation into the effects on fish bone of passage through the human gut: some experiments and comparisons with archaeological material. Circaea 10, 38–50. Richter, J. (1986). Experimental study of heat induced morphological change in fish bone collagen. Journal of Archaeological Science 13, 477–481. Rojo, A. (1987). Excavated fish vertebrae as predictors in bioarchaeological research. North American Archaeologist 8, 209–225. Shipman, P., Foster, G. & Schoeninger, M. J. (1984). Burnt bones and teeth: An experimental study of color, morphology, crystal structure and shrinkage. Journal of Archaeological Science 11, 307–325. Shokler, J. E. (1991). Understanding Wisconsin aboriginal pyrotechnology through the experimental heat treatment of Galena chert. Unpublished manuscript in the author’s possession. Smith, P. W. (1979). The Fishes of Illinois. University of Illinois, Urbana. Stewart, K. M. (1989). Fishing Sites of North and East Africa in the Late Pleistocene and Holocene: Environmental Change and Human Adaptation. Oxford: BAR International Series No. 521. Von Endt, D. W. & Ortner, D. J. (1984). Experimental effects of bone size and temperature on bone diagenesis. Journal of Archaeological Science 11, 247–253. Wheeler, A. & Jones, A. K. G. (1989). Fishes. Cambridge Manuals in Archaeology. Cambridge. White, E. M. & Hannus, L. A. (1983). Chemical weathering of bone in archaeological sites. American Antiquity 48, 316–322. Wigen, R. J. & Stucki, B. R. (1988). Taphonomy and stratigraphy in the interpretation of economic patterns at Hoko River Rockshelter. In (B. L. Isaac, Ed.), Prehistoric Economies of the Pacific Northwest Coast. Research in Economic Anthropology Supplement No. 3. Greenwich, Connecticut: JAI Press, pp. 87– 146. Wing, E. S. & Brown, A. B. (1979). Paleonutrition: Method and Theory in Prehistoric Foodways. New York: Academic Press.