Вы находитесь на странице: 1из 9
Quarter-Inch Screening: Understanding Biases in Recovery of Vertebrate Faunal Remains Brian S, Shaffer American Antiquity, Vol. 57, No. 1 (Jan., 1992), 129-136. Stable URL htp:/ links jstor-org/sic?sici=0002-73 16% 28199201%2957%3A 1% 3C 129%3A QSUBIR%3E2.0,CO%3B2-3 American Antiquity is currently published by Society for American Archaeology ‘Your use of the ISTOR archive indicates your acceptance of JSTOR’s Terms and Conditions of Use, available at hup:/www,jstororglabout/terms.hml. ISTOR’s Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at hutp:/wwww jstor.org/journals/sam. html Each copy of any part of @ JSTOR transmission must contain the same copyright notice that appears on the sereen or printed page of such transmission. STOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact support @jstor.org, hupulwww jstor.org/ Pri Oct 1 18:43:50 2004 QUARTER-INCH SCREENING: UNDERSTANDING BIASES IN RECOVERY OF VERTEBRATE FAUNAL REMAINS Brian S. Shaffer Fine screening of archaeological materials soften too expensive or too time-consuming for large assemblages. Consequently, 4 has become a standard size of mesh used among North American archaeologist. Unfortunately the elects of screening on bone recovery and the biases in interpretations are poorly documented, One-quarte. inch serening biases faunal recovery towards “larger” bone specimens. bt previous studies fail ro document biases in the recovery of spect taxa or elements. To better undersiand these bases, Screening ets were conducted (n 26 modern, comparative specimens. Results of these tests indicate that recovery and lass of specific elements {for each taxon can be predicted. Mammals with lve weights of less than 140 gare almost completely lat by 1 screening. Specimens weighing from 71 to 340 g are poorly represented, while specimens weighing from 340 t0 3.100 g are represented by most elements except foot bones. Taxa greater than 4,300 ware represented by most tlements. Eltamizado fino de materiales arqueoldgicos es a menudo excesiramente costo o consume demasiado tempo ‘enel caso de grandes conjuntos. Como resultado, el uso de 2arandas de 1" se ha converte en préticacorriente ‘ire arquedlogos norteamericanos. Desgractadament, 10s efectos del tamizado con zarandas de W" sobre la reeuperacion de esos y las resltantes sesgas en la interpretacion han sido escasamentedocumentados. pesar de que el uso de este procedimiento distoriona la recuperacion de fauna en favor de huesos de mayor amaho, Drevios estudias no dacumentan sesgos en la vecuperacion de taxones o elementos especfios. Con el propesito ‘de comprender mejor estos sesgas, se realizaron pruebas de ramizado con 26 especimenes mdernas contparativas. ‘Los resultados de estas pruebas indican que es posible predectr la recuperation ypérdida de elementos especficos para cada taxén. Mamiferes con peso en vida menor que 140 se pierden casi por complet al tamizar con I. Especimenes que pesan entre 71 y 340 gse encuentran pobremente reresentades, en tanto que aquellos con peso ‘enive 340 y 3.100 g estan represencados por la mayorta de los elementos, excepto los huesas del pie. Casi todos los elements se encuentran presentesen taxones mayores que 4.500 g Microfauna can often be recovered from archaeological sites through fine-screening procedures (Caste! 1972; Clason and Prummel 1977; DeMarcay and Steele 1986; Dye and Moore 1978; Meighan 1969:418; Payne 1972:53, 1975; Struever 1968:353). However, fine screening can be expensive and time-consuming (Barker 1975:62; DeMarcay and Steele 1986:260; Kobori 1979:229;, Payne 1975:16-17). In shell middens, use of 'j,’ screens can result in a 500 percent increase in screening time relative to use of 8" screens (Meighan 1969:418). Due to the expense in both time ‘and funds, ¥"-mesh screening has become a standard recovery method on many North American archacological sites. Fine screening and flotation are often used only for random samples and for specific archaeological contexts such as features (see examples in Baker etal. 1991:140-141; DeMarcay and Steele 1986:251; and Shatfer 1989:172) ‘Grayson (1984:168-169) noted that the choice of screen size will have a significant impact on. the type of material lost or recovered. To understand taxonomic loss from '" sereening, many sites have been sampled with fine-screen methods (eg., Baker et al. 1991:140-141; Casteel 1972:383- 1387; DeMarcay and Steele 1986:250-264; Kobori 1979:228-229; Payne 1975; Thomas 1969:392- 401; Yates 1987:87). These studies focused on comparing and contrasting differences in recovery ‘between 1" and fine-screen mesh (14" or Yi") OF on the cost efficiency of different methods. While information about specific sites was presented, an overall framework for understanding '4"-mesh biases was not presented. ‘Thomas's (1969:392-401) discussion of recovery of microfauna from Great Basin sites in Nevada provided a preliminary assessment of sieving biases. Thomas examined frequencies of Brian S. Shaffer, Department of Anhropolomy, Texas A&M University, College Staion, TX 77843-4352 American Antiquity, S71), 1992, pp. 129-136. Copyright © 1992 by the Society for American Archacology 00 AMERICAN ANTIOUTTY (Vol. 7, No.1, 1982] recovery of five arbitrarily defined weight classes of mammals from '4", %é", and i” mesh, He documented percentage loss for the two larger sizes of sereen for each of the five weight classes. For example, 1" screening resulted in a 74-100 percent loss of elements among taxa with live weights less than 100 g. Specimens weighing from 100 to 700 g exhibited losses of 39-100 percent. This work clearly demonstrated the value of fine screening in the recovery of microvertebrates and documented that larger-sized mesh screens bias the recovery of faunal material to larger specimens. It did not, however, document the loss of taxon-specifc elements from the 14" screen. To analyze taxonomic loss stemming from use of 4" screens, screening tests were conducted on skeletons of ‘modern mammals, METHODS. This study used 26 disarticulated mammal skeletons representing 25 species (Table 1). Most specimens did not have a recorded live weight from the time of collection. Therefore, live weight ‘and head-and-body-measurement ranges for each taxon were taken from Burt and Grossenheider (1976). Because several species exhibited a wide range of size or weight, the greatest length of the femur of each individual was measured using dial or electronic calipers. Measurements were taken following those described by von den Driesch (1976:84-85) and provided additional relative size information (Table 1), Specimens were selected based on fusion of elements, skeletal completeness, state of disarticu- lation, and lack of bone fragmentation. All but two specimens had fully fused long-bone epiphyses. ‘The two exceptions were a muskrat (Ondatra zibethicus) and a Mexican ground squirrel (Spermo- hilus mexicanus). The choice to use individuals with fully fused long-bone epiphyses simplified tabulating procedures and provided a control for element completeness. Each disarticulated animal skeleton was individually placed in a 4’-wire mesh screen and shaken for 30 seconds. This process was repeated nine times with complete replacement of elements for each taxon for each test. Exceptions to this are the canids. These skeletons were only sereened five times each because there was essentially no change between tests. Afler each test, elements that ppassed through the screen were recorded. Repetition of tests allowed verification of early results ‘and helped control for variation in the aggressiveness of the shaking of the screen from test to test. Recovery of loose teeth was not addressed. Also, due to the necessity for preserving comparative specimen collections, specimens were not altered or damaged in any way. ‘Tests conducted represent situations where every element is complete. This condition, of course, is unlikely to occur in archaeological sites. Moreover, the skeletal elements were placed in the screen ‘without other material such as sediment, plant remains, or rocks. Obviously, other materials in the screen may affect recovery potential, Plant materials may clog sereens and increase recovery of elements, while rocks may break bones during the sereening process and cause additional losses. ‘The sediment in the sereen also may affect recovery of smaller elements in several ways. By turning ‘small long bones on end so that they may pass through the screen, dry matrix may cause a loss of ‘material. Conversely, moist matrix may adhere to specimens, preventing passage through the screen. ‘Another consideration that cannot be controlled for is the role of the field technician operating. the screen. If atentive, the field technician may have the opportunity to remove small bones from the screen before they are lost or may pick up bones that they see pass through the screen. In ‘addition, some small bones may be recovered from the matrix before it is screened, although this is unlikely (Casteel 1972:382; Struever 1968:353; Watson 1972:221-223), Thus, tests reported here cannot emulate all possible field circumstances. It is hoped that by providing this idealized data set, researchers can use it as a guide for evaluation of faunal material recovered from specific sites, taking into account the conditions of recovery at each site RESULTS AND DISCUSSION OF TESTS ‘Table 2 documents the results of the '"-screen tests. Table 1 provides taxon reference numbers for Table 2, taxa names, live-weight ranges, head and body lengths, and the femur length for each specimen tested. With few exceptions, body-weight range (Table 1) correlates well with the number REPORTS wa Table 1, Weight and Measurements in Ascending Order by Lowest Weieht. eter ence Weight Length Femur Number Taxon @ (am) (am) 1 Least shrew (Crypitis parva) a 56-64 70 2 Pygmy mouse (aiomys talon) ro S164 93 3. Evening bat (Nyeticeus humeralis) 19 36-38 13 4 House mouse (Mus muscu) 11-22 Bi-86 122 5. Shortail shrew (Biavina brevicauda) 1122 31 6 Deer mouse (Peromyscus maniculatus) 18535 SL 7, Mexican pocket mouse (Liomys irrratus) 3450 wor-127 222 8. Kangaroo rat (Dipodomys oni aon Wor-114 246 9. Eastern mole (Sealopus aquaticus) orto 130 10. Valley pocket gopher (Thomomys boxe) 71-250 10 HI. Pocket gopher (Thomomys sp.) (LA 15049 Grant County, NM) 2 1 24 12 Cotton rat (Sigmodon hispiis) 3-198 wr203 7 13, Plains pocket gopher (Geomys bursarius) 127-354 140-229 290 14, Wood rat (Neovoma albigula) 135.283 190-216 336 15, Thirteen-lined ground squire (Spermophitus tridemcimlineats) 140-252 1446s 29.0 16, Red squirel (Tamlascurus hudsonicus) 198-250 178-203 405 17, Mexican ground squierl Spermophitus mexicans) 198-340 iia 304 Mink (Mustela vison—male) 198340 228266347 Gray squirrel (Seiurus carolinensis) 340-726 200-250 322 Spotted skunk (Spilgate purorius) 363.999 230-30 46.4 Cottontail (Syvilagus audboni) 600-1,200 300-380 65.2 Ringtail eat (Bassariscus asturus) 360-410 638 Muskrat (Ondatra sibehicus) 250-360 43.6 Jackrabbit (Lepus calforncis) 1,300-3,100 430-530 106.2 Red fox (Pulpes flv) 4'300-6,7005,600-6,300 1326 Coyote (Cant latrans) 91000-22,000__8,100-9,400__176.0, of elements recovered (Table 2, Figure 1). For example, the only elements consistently recovered for animals weighing < 140 g were crania, with innominates and sacrae also being recovered for ‘some specimens. The only long bone consistently recovered for a specimen < 140 gis the humerus ‘of the mole (Scalopus aquaticus). The unique, circularly shaped bone cannot pass through '" mesh from any direction. Clearly, the use of 4" screening will miss most elements of mammals listed in this first weight class. Thus, quantitative analysis of this category recovered by %" mesh will not yield an accurate assessment of taxonomic abundance or even presence vs. absence. For example, Driver (1985:5) ‘noted that the only small fauna recovered in this size category (from " mesh) from six sites in the Sierra Blanca Region of New Mexico were Peromyscus (two specimens) and Microtus (two speci- ‘mens). These specimens were represented only by crania and mandibles. Unfortunately, clear distinctions between each of the size categories could not be made. Conse ‘quently, these categories overlap in size. Delineation between the first two categories is based on the recovery of mandibles and scapulae. The largest specimen in the < 140 g weight class is a mole, bbut 1 screening did not consistently recover the mandibles, scapulae, or most long bones. Most clements from animals weighing 71-340 g also will be missed with '" screening. However, these taxa are often large enough for several commonly identifiable elements to be recovered (pelves, scapulae, femora, humeri, skulls, and mandibles). ‘The next weight category is 340-3,100 g. Mammals in this category should be fairly well repre~ sented with ''-screen recovery methods. Even though many identifiable elements are recovered, ‘other elements such as caudal vertebrae, ribs, sternae, patella, sesamoids, podials, metapodials, and phalanges were not recovered consistently. Distinction between this size category and the ‘Table 2. Recovery of Mammalian Elements with 1/4*-Mesh Sereen. ‘Weight Class 35007 & 25 8 0-3 20 7308 3 ide 2 Bs 2 is 7 16 1s 14 2 7 0 Cranium Ramue Vertebra [AMERICAN ANTIQUTTY Atlas ‘Axis we le des ee tetni ee tert eetedts ee tedii we tette eeredns eOretit bhrediy letestui weedy wererii wen eat eeu ele le ceces [eee Vrimiecprrriyyd Hrimtititie rin Fitetrrirrriens Hrimereiisreens Cervical Thoracic Lumbar Sacrum Caudal Seapil Ciaviele Sternae ‘Manubrium Rib Humerus Una Radius Carpals Metacarpals Pelvis with used (Vol. 57, No.1, 1962) REPORTS imo pao on sss poms Se sour 0 posty wuss “9 uaWoFE saqtoue or Posty PaLONOo9! Sem HHA = apPIN Sos saueyoud reurosa sprouiesog, seis sam snoUP) soyetensy sree inqyorgtt eindit con eertiieriiy leteduuuree sereie riers levrruuuiet Hoeeeneneee we dee titer sreuruiout ee ltieteedd ee crtediiie sleetrvrrini afirtiieriigs arrrererries 8 g 3 +005 aD WEA ‘ponunuoy -Z21g8, 14 ‘AMERICAN ANTIQUITY [Wok 57, No.1, 1992) 200 3 180 160 140 120 100 20 60 202 rt gg BP M2n pie 12-78 9 10 Ul 12 13 4 18 16 17 16 19 20 21 22 25 24 25 26 Taxa (See Teble | or lara description) a Legend — Elements Recovered Figure 1, The numberof elements recovered per taxon with % screening. previous size category is based on the maximum size for the Mexican ground squirrel (Spermophilus ‘mexicanus) and the minimum size for the gray squirrel (Sciurus carolinensis) (see Table 1). The ‘mink (Mustela vison) is included in this category because the individual was a domesticated animal ‘and therefore was probably larger in size than wild individuals. Also the recovery of elements for this individual was greater than for others in the 71-340 g category. This level represents the first size range at which the presacral and sacral vertebrae, and the majority of commonly identified ‘elements are recovered from Ye" mesh screen. In the > 4,500 g category, most elements except sesamoids, carpals, patellae, and middle and distal phalanges were recovered in 70 percent or more of the tests conducted. The 4,500-g level ‘marks the point at which '" screening does not greatly bias the sample toward “larger specimens, since virtually all of the specimens are “larger,” and few small elements are present in animals of this size. lustrated in Figure 1 are the number of elements recovered with i screen in the tests conducted. ‘The sum of the elements that were recovered consistently (Table 2) was plotted for each taxon. Generally, as taxa increased in size, the number of elements recovered increased. Exceptions to this ‘were the small squirrels (Taxa 15-17) and the muskrat (Taxon 23). Examination of Table 1 reveals that the small squirrels and the muskrat are smaller in length than preceding taxa and most also hhad smaller femora, Increases in the number of bones recovered for each size class are due to the recovery of new clement categories for each size class. For the 71-340-g class, additions included rami and long bones. For the 340-3, 100-g class, additional elements recovered included vertebrae and certain long ‘bones. The large increases noted in the red fox and coyote (Taxa 25 and 26) are due primarily to the recovery of podials, metapodials, and phalanges. ‘These results indicate that only taxa with similar recovery rates should be quantitatively compared. ‘That is, regardless of cultural and taphonomic factors that may have influenced assemblage com~ position, quantitative comparison of taxa can be conducted only when recovery methods have not produced biased results. CONCLUSIONS Fine-screen techniques continue to be used on a limited scale at many archaeological sites. Even ‘though '" screening provides the bulk of recovered faunal remains from many sites, screening biases REPORTS. ns largely have been overlooked. Recovery tests conducted on 25 species of mammals using 4? mesh revealed that coarse screening not only biases samples towards larger specimens, but that the bias against smaller specimens can be predicted. Additionally, the lack of recovery in archacological assemblages of those elements consistently found to be recovered by l’-screen mesh in this study should signal the zooarchacologist to question why the material was not recovered from the site. ‘These results provide zooarchaeologists with a useful tool to ad in the interpretation of faunal data. This study also suggests that differential recovery compromises the utility of quantitative com- parisons of different sizes of taxa. Based on the data presented, only taxa with similar recovery rates should be compared unless recovery biases are taken into account. For different-sized taxa, only elements with comparable recovery rates should be used for quantitative comparison. Acknowledgments. Barry W. Baker, John E. Dockal, Joseph F. Powell, Julia L. Sanchez, Harry J. Shafer, 1D. Gentry Steele, and four anonymous American Antguity reviewers provided valuable commenis on this ‘manuscript. Teresita Majewski is thanked for her editorial assistance. Bonnie C. Yates provided aditional ‘consultation, Specimens used inthis study were borrowed from the following collections (1) the Comparative ‘Zocarchacological Research Collection, Department of Anthropology, Texas A&M University; (2) the Depart- ‘ment of Applied Sciences Zooarchacoiogy Laboratory fuunal collection, University of North Texas; (3) the University of Texas, Department of Geology, Vertebrate Paleontology Collection; and (4) from my personal collection. Juan José Aguirre translated the abstract into Spanish, REFERENCES CITED Baker, B. W., B.S. Shaffer, K. D_Sobolik, and D. G. Steele 1991 Faunai Analysis, Par: Analysis ofthe Vertcbrate Faunal Remains. In Alabonson Road: Early Ceramic Period Adaptation to the Inland Coastal Prairie Zone, Harris County, Southeast Texas, edited by H. B. Envor and D. L. Carlson, pp. 139-161. Repors of Investigations No, 8. Archeological Rescarch Laboratory, ‘Texas A&M University, College Station, Barker, G. 1973. To Sieve or Not to Sieve. Antiguity 49:61-63. Burg, W, H, and R. P. Grossenheider 1916 4 Field Guide to the Mammals of Ameria North af Mexico. Houghton Millin, Boston Castel, R, W. 1972 Some Biases in the Recovery of Archaeological Faunal Remains. Proceedings of the Pehistorie Society 38:328-38, (Clason, A... and W. Prummel 1977 Collecting, Sieving and Archaeozoological Research. Jownal of Archaeological Science 4:171-175, DeMareay, G.B, and D. G. Steele 1986 The Vaive of Fine Screening On Inland Based Hunter-Gatherer Habitation Sites. In Archaeological Investigations at 41 LK 201, Choke Canyon Reservoir, Southern Texas, by C. L. Highley, pp. 250-264 ‘Choke Canyon Series No. 1. Center for Archaeological Research, University of Texas, San Antonio Driver, J.C. 1985 -Zooarchacology of Six Prehistori Sites im the Sierra Blanca Region, New Mexico, Research Reports ‘in Archaeology Contribution No. 12, Museum of Anthropology, University of Michigan, Ann Arbor Dye, D-H, and K. H. Moore 1978. Recovery Systems for Subsistence Data: Water Screening and Water Floatation. Tennessee Anthro: pologist 3:39-68. Grayson, D. K 1984." Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas. Academie Press, Or- Tando, Kobon, LS. 1979" Dierentia Bone Recovery Experiment. In Ezra's Retreat: Rocksheltr/Cave Site inthe North Central Great Basin, by J.C. Bard, C. 1 Busby, and L. S, Kobori, pp. 228-229. Center for Archaeological Research, University of California, Davis, Meighan, C: W. 1969 "Molluscs as Food Remains in Archaeological Sites. In Science in Archaeology, 2nd ed, edited by D. Bothwell and E, Higgs, pp. 415-422, Thames and Hudson, London Payne. 1972 Partial Recovery and Sample Bias: The Results of Some Sieving Experiments. In Papers in Economie Prehistory, edived by E. 8. Higgs, pp. 49-64. Cambridge University Press, London. 1975. Partial Recovery and Sample Bias. In Archaeozdologcal Studies, edited by A. T. Clason, pp. 7-17 “American Elsevier, New York. 16 AMERICAN ANTIQUTY (Vol. 7, No.1, 1082] Shaler, B.S. 1989 Laie Prehistoric Faunal Subsistence on the South Texas Plains: Analysis of the Vertebrate Faunal ‘Remains from 41VT66, Vietoria County. Plains Anthropologist 34:171-178 Struever, 1968.” Fiotation Techniques for the Recovery of Small-scale Archaeological Remains, American Antiquity 33353362, ‘Thomas, D. H. 1969 ” Great Basin Hunting Patterns: A Quantitative Method for Treating Faunal Rem: tiguity 34:392-401 von den Drisch, A 1976. A Guide to the Measurement of Animal Bones from Archaeological Sites, Bulletin No. 1. Peabody ‘Museum of Archacology and Ethnology, Harvard University, Cambridge ‘Watson, W. P.N. 1972’ Fragmentation Analysis of Animal Bone Samples From Archaeological Sites. Archaeometry 14:221~ 228, Yates, B.C. 1987 "Appendix C: Faunal Data from 41CO141, In Test Excavations at 41COI41 Ray Roberts Reservoir, Cooke County, Texas, pp. 133-146. Inttute of Applied Sciences, North Texas State University, Deaton. 3. American An Received January 8, 1990; accepted August 14, 1991

Вам также может понравиться