HH11SK KIRTLANDIA CLEVELAND, OHIO NUMBER 39 •NATURAL HISTORY* KIRTLANDIA David S. Brose, Editor Kirtlandia is an occasional publication of The Cleveland Museum of Natural History and is devoted to scientific papers in the various fields of inquiry within the Museum’s sphere of interest. Published at least twice a year, issues will vary be- tween collections of short papers and single issue-length studies. Kirtlandia is distributed by The Kent State University Press, Kent, Ohio 44242. Copyright © 1983 by The Cleveland Museum of Natural History. Kirtlandia is abstracted in Zoological Record and Biological Abstracts and in- dexed in Bibliography and Index of Geology. ISSN: 0075-6245 KIRTLANDIA THE CLEVELAND MUSEUM OF NATURAL HISTORY Cleveland, Ohio December 1983 Number 39 RECENT EXCAVATIONS AT THE EDWIN HARNESS MOUND, LIBERTY WORKS, ROSS COUNTY, OHIO N’OMI GREBER with James B. Griffin, Tristine Lee Smart, Richard I. Ford, Orrin C. Shane III, Raymond S. Baby, Suzanne M. Langlois, Stephanie J. Belovich, David R. Morse, Kent D. Vickery Abstract Analyses of data recovered during the 1976-1977 excavations at the remnant of the Edwin Harness Mound, a major Hopewell site in Ross County, Ohio, are presented, with some reanalysis of data from earlier (1840-1905) excavations. A large structure has been identified at the base of the mound. Botanical, faunal, and soil analyses indicate that the physical environment in the central Scioto Valley near a.d. 300, when the building was in use, was similar to that found in the area in the late eighteenth century. Detailed studies of ceramics, lithics, and human skeletal remains are presented. The varied activities which took place within and near the major structure, and implications for interpretations of materials from Seip (Pricer) Mound on Paint Creek, are discussed. Kirtlandia No. 39 0075-6245/83/1983-0039 $9.95 Copyright © 1983 by The Cleveland Museum of Natural History • • . ■ ■ ' ■ CONTENTS Figures 5 Tables 5 Acknowledgments 7 1. Introduction 11 2. The Excavations 13 3. The Site 23 4. The Ceramic Complex — James B. Griffin 39 5. Plant Remains — Tristine Lee Smart and Richard /. Ford 54 6. Vertebrate Faunal Remains — Orrm C. Shane HI 59 7. Analysis of Human Skeletal Material Ohio Historical Society Collections — Raymond S. Baby and Suzanne M. Langlois 62 Cleveland Museum of Natural History Collections — Stephanie J. Belovich 62 8. Mollusc Identification and Analysis — David R. Morse 66 9. The Flint Sources — Kent D. Vickery 73 10. Discussion and Conclusions 86 FIGURES 1. 1 General view of excavations, 12 August 1976 1 1 1.2 General view of excavations, 21 July 1977 12 1.3 Base of heavy gravel wall (Feature 1) 12 2.1 General location of site 13 2.2 Edwin Harness excavation grid, 1976 15 2.3 Edwin Harness excavation grid, 1977 16 2.4 Post holes as first found 1976-1977 17 2.5 Sketches of site profiles 18-19 2.6 Remains of major activity floor and other features found at base of mound 20 3. 1 Cuts made by Mills in 1905 through the original floor layers 26 3.2 Floor plan of Harness Big House 28 3.3 Location of post holes with surrounding stains 30 3.4 Examples of post holes with surrounding stains 31 3.5 Pipe fragment found by F. W. Putnam 32 3.6 Sketch profiles of pit features 35 4. 1 Hopewellian and Check-stamped sherds 40 4.2 McGraw Cordmarked, McGraw Plain, and unidentified sherds 41 4.3 Turner Simple Stamped and unidentified simple stamped sherds 46 4.4 McGraw Plain and McGraw Cordmarked sherds 47 9. 1 Areas of outcrop of flint varieties 75 10. 1 Estimated floor plan of Seip Big House 88 10.2 Radiocarbon dates plotted as mean calendrical years 90 TABLES 2. 1 Summary of Features 14 3.1 Non-structural Posts Associated with Edwin Harness Big House 29 3.2 Radiocarbon Dates 34 4.1 Summary Tabulation of Harness 1976-1977 Pottery Collection 39 4.2 Griffin-Morgan Pottery Count of Ohio State Museum Harness Collection 42 4.3 The Occurrence of Turner Simple Stamped B in Ohio, based on Prufer (1968) 43 4.4 The Occurrence of Turner Simple Stamped A in Ohio, based on Prufer (1968) 45 5.1 Charred Posts from the Edwin Harness Mound Floor 55 5.2 Charcoal from the Edwin Harness Mound Features 56 5.3 Carbonized Seeds and Faunal Remains from the Edwin Harness Mound Botanical Samples 58 6. 1 Frequency of Identified and Unidentified Bones by Class 59 6.2 Frequency of Identified Vertebrate Remains from Edwin Harness Mound 60 7.1 Edwin Harness Mound Human Skeletal Material Excavated 1896-1905 63 7.2 Edwin Harness Mound Human Skeletal Material, non-cremated, 1977 64 8.1 Summary of Mollusc Taxa Identified, Edwin Harness Mound 67 8.2 Comparisons of Dimensions Melongenidae — Conch Shells 68 9. 1 Provenience of Debitage from Edwin Harness Mound 77 9.2 Correlation of Flake Type and Technique of Flake Removal, Edwin Harness Mound 78 9.3 Correlation of Debitage Type and Flint/ Chert Variety, Edwin Harness Mound 79 9.4 Correlation of Technique of Flake Removal and Flint/Chert Variety, Edwin Harness Mound 80 9.5 Frequency Distribution of Flint/ Chert Varieties for Cultural Collection and Comparative Local Scioto River Gravel Pebbles 82 10.1 Values of Ranksum F for Seip Mound 1 (Pricer) 89 . ' . ■ . . ; . ■* . ACKNOWLEDGMENTS It was exciting and somewhat unbelievable in Novem- ber 1975 to see clear evidence that sections of a major Hopewell mound were still unexcavated after 1 30 years of intermittent digging. This was my first look at a very level Edwin Harness Mound. Many people shared the effort and work, which have resulted in significant new informa- tion from this classic site. The work was made possible by Robert Harness, who not only gave us permission to ex- cavate on his land, but also helped in many practical mat- ters. These ranged from letting me use the top of his trac- tor as a photography tower in November 1975 to backfilling over many refilled and cemented post holes af- ter the 1977 field season. He and his wife, Marilyn, were most hospitable to their many digging guests. The first season of field work was under the auspices of the Ohio Archaeological Council, whose president, David Brose, and vice-president, Orrin Shane, were instrumen- tal in initiating the support of the council and in encourag- ing me to work at the site. In the field there was a perma- nent staff of three — Wesley Clarke, Michael Hambacher, and myself — and a total of 1 20 volunteers who worked at the site for various lengths of time from June 7 to August 14, 1976. Funding for this season was provided in part by a matching grant to the council from the Ohio American Revolution Bicentennial Advisory Commission. The match for this grant was more than met with the dona- tions received by the council from Ohio corporations and private sources. At the end of this season, when it became evident that there was more detail intact than could pos- sibly be completed within the original time limit, Mr. Harness agreed that 1 could seek funds for a second season. Major funding for this second season, which was under the auspices of the Cleveland Museum of Natural History, came from the National Science Foundation. Field school students worked with a core crew of expe- rienced first season veterans, making a combined total of 43 crew members for the second season. Wesley Clarke was my chief assistant. Franco Ruffini joined the supervi- sory staff with special responsibilities for overseeing stu- dents. Michael Hambacker did what might seem impos- sible by maintaining an excellent, much needed, and often used photo record of our work while still demonstrating his exemplary field techniques. The second season began June 14 and ended August 26, 1977. Both summers, Alva McGraw gave us the use of one of his farmhouses as well as help with innumerable practical problems and invaluable moral support. We, as all ar- chaeologists in Ross County, give him special thanks. The curation and analysis began in the field in 1976. Many hours were spent “after hours” cataloguing, clean- ing, water screening, etc. Jane Busch was the chief field cataloguer in 1976; Shaune Skinner in 1977. The work continued in the winter of 1976-1977 at the museum with Dennis Griffin. In the fall and winter of 1977-1978 Eloise Gadus supervised the laboratory curation and was a general research assistant. Museum volunteers and stu- dents from Cuyahoga Community College, Case Western Reserve University, and Cleveland State University worked on various laboratory connected projects. The re- sults of specialized analyses presented herein were per- formed at the respective laboratories of the various con- sultants noted in the chapter headings. Even with excellent support from public foundations and with private dona- tions of funds and services, choices had to be made con- cerning which samples would be analyzed first, which would wait. Those samples awaiting analysis are stored and ready for new resources, new analytic techniques, and new ideas. The integration of information from earlier excava- tions was an essential part of our work. Martha Otto and Bradley Baker of the Ohio Historical Society and Stephen Williams and Sally Bond of the Peabody Museum, Har- vard University, were, as always, gracious and helpful in allowing access to their collections and documents. The maps and drawings of the field data were prepared for the present report with patience and skill by Mark Schornak. The chapters without author designations were written by myself. Although I have personally gained knowledge from each archaeological excavation in which I have partici- pated and know that each site represents a unique and irreplaceable piece of prehistory or history, Edwin Har- ness is a special place for me. To each and every one who shared the planning, excavation, curation, and analysis I give personal and professional thanks. Without your help we would still only be able to speculate on the existence of the Harness Big House. N’omi Greber THE EDWIN HARNESS MOUND EXCAVATION Field Personnel, 1976 Director: N’omi Greber Field Supervisors: Wesley S. Clarke Michael Hambacher Participating Supervisors David Brose Patricia Essenpreis Marie Freddolino Michael Pratt Jeffrey Reichwein David Stothers Kent Vickery Kelly Adkins Robert Asumc Katie Banzhaf David Barna Sandra Barna Naomi Bauer Tom Berreis Robert Blickensderfer Nancy Bocash Tom Bogus Jonathan Bowen Ken Bowden Helen Bradburn Ken Broberg Jane Busch* Nancy Cameron Ricky Cibulskas Eileen Closs Elizabeth Cole Cindy Cubbison Jane Claire Deward John Doershuk Jim Donald Marion Donald Sally Donald Jim Donald, Jr. Bob Drozda David C. Ebosh Jean A. Ebosh Fran Erwin Dale Fable Brad C. Featherstone Bill Fenton Charles Ford Jon Paul Fry Bob Genheimer Leanna Geriak Mike Glenn Amy Glin * Adjunct Field or Laboratory Assistant General Crew Richard S. Gray Lisa Greber Rebecca Greber* Dennis Griffin* Dan Grossman Joseph T. Hannibal Bonnie Hannon Dan Harmon Crickette Harrell Janet Hart Laura Havasi Chuck Heath Neil Henderson Connie Holden George L. Holler Frank Huntley Frank Johnoff Winifred Kelley Debby Klein Phil Kleinhenz Lois Lambert Peter Lemmerman Timothy Light Janet Lipstreu Lisa Littman Karen Lord Bruce Leland Markley Mike Marmostein Harry Martin Mark McDonald James Trent Metcalf Barbara Mihuta Laura Mihuta Dave Miller Dward A. Moore* Jeannine Moore Dave Morrison Glen W. Nyhart Bradley A. Oen Special Consultants Raymond Cotrill James B. Griffin Don Grimes James Kerr Alva McGraw James Marshall Maynard Munsing Orrin C. Shane III Jose R. Oliver Christine Opfer W. Kevin Pape* A. Lowell Randall James Reckley Frank Reiger Alison Roberts Betty Rock Patricia Rubright Franco Ruffini* Ronald A. Salupo Hattie Sells Ruth Sheard Michael Shore Jon Singer Clement Skehan William Snodgrass Kimberly Jo Solsman Dana G. Staley Patrick Steiner Merry Steward Ruth Streicker Ted Sunderhaus Helen Swanner Christine Tailer* Sue Tituskin* Dale Thomas Bryan Tupper Millie Tupper William D. Ullery John Walters Sally Warrick Carmen Anna Weber Matthew J. Weitendorf A1 Wilson Susan B. Wiseley David B. Woodmansee Nora Wright THE EDWIN HARNESS MOUND EXCAVATION Field Personnel, 1977 Director: N’omi Greber Field Supervisors: Wesley S. Clarke Franco Ruffini Participating Supervisors: Dward Moore Arthur Saxe Mark Bir* Jonathan Bowen Robert Buck* Daniel W. Bull* Lois Cahill* Karen Ciatyk* Hattie S. Clarke** Ann C. Cramer Scott Detroiler* John Doershuk Marjory Forbes-Howard* Lois E. Fultz* Eloise Gadus* Kathy Gladwin* John T. Goerlich* Linda Grand* Richard S. Gray** Rebecca Greber** Dennis Griffin Susan Hammond* Field Photographer: Michael Hambacher Janet Hart Elizabeth Ippolito* Mary Klock* Britt S. Krebs** Janet Lipstreu** Caroline McLeod* Bradley A. Oen** W. Kevin Pape Laurie Patti* Jim Retzler* Alison Roberts** Daniel Simon** Shaune M. Skinner Roy Walsh* William Snodgrass** Christine Tailer Robert E. Thornsberry* Sue Tituskin Bryan Tupper** Sally Warrick** Russell Weisman* * Field School Student ** Volunteer ■ . ■ ' ■ 1 INTRODUCTION The following site report on the 1976-1977 salvage ex- cavations of the remnant of the Edwin Harness Mound, Ross County, Ohio contains types of data which could not have been included in the reports of earlier expedi- tions. Technology has grown rapidly since 1907 when the last report was published. The application of these new analytic techniques was one of the major purposes of our excavations. (Figs. 1.1, 1.2). A second major purpose was to salvage data on the manner in which space had been used by the inhabitants of the site before the mound had been constructed. This data had not been available be- fore, partly because of the checkered series of excavations and the missing field records from major sections of these excavations. These earlier excavations at Harness and other major sites in southern Ohio established that differ- ent cultures were represented among the “mound build- ers.” Edwin Harness was assigned to Hopewell on the ba- sis of artifact similarities (Mills 1 907 : 1 9 1 ). These artifacts do show the high artistic and technical talents of the indi- viduals who made them, and we still restudy them because they are part of the basic data which allow new ideas to be developed and refined. Less exotic, perhaps, but equally important to this data base is the context in which these objects were found. Fortunately some of the gaps in the knowledge of the contexts at Edwin Harness can now be filled in, both with newly found old records (e.g., Murphy 1978) and with new field data (Fig. 1.3). The emphasis of the most recent field work at the site of the Edwin Harness Mound was context; the following chapters report on the data we found and on the special- ized studies of that data which have been completed to date. The design and construction of a major Hopewell civic-ceremonial building, parts of its contents, and the mound that covered it are described. We also now have specific data on environment, subsistence, and chronol- ogy from this classic site. References Mills, William C. 1907 Explorations of the Edwin Harness mound. Ohio Ar- chaeological and Historical Quarterly 16:1 13-193. Murphy, James L. 1 978 William C. Mills’s notes on the Edwin Harness mound excavation of 1903. Ohio Archaeologist 28(3):8— 1 1. Fig. 1.1. General view of excavations, 12 August 1976. The remains of the heavy gravel wall (Feature 1) ringed secondary mound fill and defined the major activity floor. View looking west. Fig. 1.2. General view of excavations, 21 July 1977. View looking south. Fig. 1.3. Base of heavy gravel wall (Feature I), west side. Unit N537.5 E485, 28 July 1976. 2 THE EXCAVATIONS Previous Excavations Putnam began his trench on the north end, inside the heavy cobbles which ringed the mound, and gradually cut down towards the floor. He widened the trench when he reached the first charred areas (Greber 1 979: Fig. 6.4). Moorehead continued south from the end of Putnam’s trench. He did not dig from the surface to the floor but used tunneling techniques to reach the south end of the mound (ibid.). Mills began at both the south and southeast edges of the mound in his first season. He spent 12 days in 1903 digging to and through the floor of the mound to the approximate end of Moorehead’s tunnels. When he re- turned in 1905, he began on the northeast side of the mound east of Putnam's trench and finished on the west side of this trench. The relatively intact main floor extend- ing east-west near the N525 line and the similarly rela- tively extensive remains on the west side of the structure are in areas uncovered, if at all, at the end of each of Mills’s two expeditions. Squier and Davis describe the mound as egg shaped in plan, with the larger end at the north. The height of the mound varied from approximately 20 ft (6 m) in the north Fig. 2. 1. General location of site. Topographic contour intervals, in feet, were taken from U.S.G.S., 7.5 minute series quad, Chillicothe East. The Edwin Harness Mound (lat. 39°15.'4N, long. 82° 52.'6W) was the largest of the 14 mounds associated with the Liberty Earthworks, Ross County, Ohio (Fig. 2.1). Excavations of this mound, as for many of the large clas- sic Hopewell mounds in Ohio, began in the early part of the nineteenth century and have continued intermittently to the present. The results of the first two test shafts in the mound were reported by Squier and Davis in 1848. Sub- sequent digging was done by local schoolboys; by Freder- ick W. Putnam ( 1 885) of the Peabody Museum, Harvard; by Warren K. Moorehead (1897) and William C. Mills (1907, 1903) of the Ohio Historical Society. Mills quotes in some detail from the reports on earlier excavations and from Putnam’s account of the materials found by the schoolboys. One of the pits dug to the floor of the mound by the boys was next to Squier and Davis’s Pit B. This pit, which has been recorded by all the excavators except Moorehead who worked only at the south end of the mound, has provided one of the major reference points for integrating data obtained over a 130 year time span (see Table 2.1:Fea. 24). 14 N’OMI GREBER No. 39 TABLE 2.1 Summary of Features Major mound stratum: 1,41 Local mound stratum: 20, 21, 43 Floor stratum: 3, 3C, 33, 50, 65 Local floor stratum: 3B, 72, possibly 39, 96 Main activity floor South Section Shallow burned area: 82B, 92 Disturbed grave: 70, 82A; immediately west, 71 Middle Section Shallow burned area: 47 Disturbed grave: 83, 84 Heavily burned area: 36 Pit: 19, 30, 79, 89 North Section Shallow burned area: 34, 78; immediately west, 32 Disturbed grave: 27, probably 28, 29; all immediately west Burned log: 26, 35 Probably heavily burned prepared clay basin: 18 Pit: 91 Prepared basin: 62 Depression: 54 Log mold: 48 Pit: 17 Post Hole: 11, 13, 14, 38, 40. 42, 64, 66, 67, 87 Small trench with stake holes: 59, 85, 88, also in 22 Shallow burned area or deposit of burned materials: 6-10, 12, 16, 31,44-47, 49, 51-53, 76 Historic pit in main floor: 24, 25, 77, 93, 94, 95, probably 58, 86, 90 In prior backfill Bone concentration: 2, 4, 5, 23, 61, 74, 75 Log: 15 Below main activity floor Shallow deposit of burned materials On Feature 65: 54A, 97 On Feature 50: 57, 73 Associated with outer areas Burial: 56, 60 Major stratum: 69, 69A Deposit within 69: 68, 80, 81 Deposit within 69A: 55 Pit: 63 Tap Root: 37 to approximately 1 1 ft (3.3 m) at the southern end. Mills reported the maximum height at the north as slightly less (16.75 ft, or 5. 1 m). The recorded decrease is likely due to Putnam’s cut through the crest of the mound and some erosion. The site was backfilled after Mills’s more exhaus- tive excavations to approximately 5 ft (1.5 m) in height. Based on photographs and descriptions of Mills’s work and on our recent excavations, it seems likely that the backdirt was deposited behind the excavators as they worked section by section. The outer boundary of the mound described by all work prior to 1977 was the heavy cobble and bedded stone mantle shown in Squier and Da- vis’s estimated cross section (1848:Fig. 6f). Major por- tions, if not essentially all, of the lower sections of this wall were left intact until heavy power equipment was used in 1975 to level the remnant of the mound and the backfilled excavations to the general ground level. Data from previous work, though varied in their documenta- tion, have been integrated into the already completed re- ports on special aspects of the research goals (Greber 1979; Gadus 1979; Greber, Davis, and DuFresne 1981; Bender, Baerreis, and Steventon 1981) and into the pres- ent site report. Field Methods 1976-1977 During the 1976 season, the remains of the very bottom of the outer stone wall (Feature 1) which had surrounded the mound were defined. Hand excavated trench units were concentrated in the least disturbed sections of the site on the west and in the north in order to obtain undis- turbed prehistoric ecofacts and as much stratigraphic in- formation as possible. These units were taken down through all cultural deposits into the underlying natural soils. Two large areas in the central portion of the site were cleared using power equipment (Fig. 2.2). During the second season, additional hand excavated trench units were placed on the east and in the south, again to give stratigraphic information, particularly as a guide in using power equipment. Backhoe trenches were dug to aid in determining the horizontal extent of the floor strata (Fea- tures 3, 3B, 3C, 33, 39, 50, and 65) and mound strata (Fea- tures 4 1 , 69, and 69 A) found both by the hand excavations and the first backhoe trenches. These trenches were num- bered as they were dug. After the hand excavation trench units were completed, the interior of the site (within Fea- ture 1) was cleared by the backhoe so that the building patterns and also the activities on the remaining floor out- side the building but within the stone wall could be de- fined by final hand clearing (Fig. 2.3). At the request of Mr. Harness the stones of this wall were removed and kept separate from the other backdirt, thus simplifying future farming of the land. Mr. Harness himself very kindly did the backfilling. A 2.5 m unit which contained original mound strata just within this wall was left unex- cavated, and it is hoped that sections of this will remain below the depths reached by farming equipment. At first, all excavated soils were dry screened through !4 in. mesh. As it became apparent that the mound load- ings at the remaining mound levels were culturally sterile and that archaeological redeposited materials were of secondary importance, general dry screening was discon- tinued. When significant deposits of prehistoric materials were found in the archaeological backfill, soils were dry screened. Flotation samples were taken from undisturbed areas of features and post holes as well as systematically from mound loadings in each excavation unit. All the re- maining excavated soils from features and post holes were EDWIN HARNESS 33Ro22 Excavation Grid, 1976 # Fig. 2.2. The Excavation grid was set along the major axis of the mound Grid north is 30° west of magnetic north (1976). The units indicated by dotted lines were excavated to the base of the gravel wall. Grid point N540 E485 is at 624 ft ( 190 m) elevation and at the Ohio Plane System Coordinate (1,893,234; 457,750). 480 485 490 495 500 505 510 515 520 525 530 Fig. 2.3. Trench units excavated by the backhoe are numbered in sequence as dug. After completion of the hand excavated trench units, the area within the gravel wall was excavated by a combination of backhoe and hand clearing (see chap. 2 and Fig. 2.2). Magnetic north as of 1976 indicated. EDWIN HARNESS 33Ro22 1976-1977 Excavations POST HOLES Fig. 2.4. Post holes as first found. Some variations in elevations due to priorexcava- tion (see chaps. 2 and 3). -DISTURBED- $ a a Fig. 2.5. Sketches of site profiles. Vertical scale exaggerated for clarity. The directions indicated on profile sections are grid W and grid N. Magnetic north ( 1976) is indicated on the location key. Section EE' illustrates composite section through Big House; details discussed in chap. 3. EDWIN HARNESS 33Ro22 1976-1977 Excavations Fig. 2.6. Remains of major activity floor and other features found at base of mound. See chap. 3 for explanation of elevation depths represented. Magnetic north as of 1976 is indicated. 1983 EDWIN HARNESS MOUND 21 water screened through fiberglass window screening. The flotation equipment, which was constructed by museum personnel, was basically that described by Patty Jo Wat- son (1976). Screens of 4 mm, 2 mm, and 500 /urn were used. Flotation was done at the Scioto River about 8 km from the site. Water screening was done largely at the field station. One hundred two features were recovered. These fell into several categories, which are tabulated in Table 2.1 and discussed in more detail in later sections. Post holes were tabulated separately from features. Oc- casionally a disturbed post hole was first recorded as a feature and then given a post hole number when it was later excavated and identified ( Fig. 2.4). Once the prehis- toric loadings or historic backfills were removed, the structural post holes were readily identified. The subsur- face depths at which these holes were identified varied. There was a downward slope built into the prehistoric floor from the center towards the outside. However, the major factor affecting the identification depth was the de- gree of historic disturbance (Fig. 2.5). With very few ex- ceptions (e.g.. Fig. 2.6: Feature 87) all major building posts had been set deep enough to be identifiable below extensive disturbances which were mainly due to Mills’s excavations. Smaller, more shallowly set stakes, which may have been placed in sections of the building floor previously excavated by Mills, could not, of course, be found. Since the 1976-1977 excavations were salvaging infor- mation from a severely disturbed site, samples for ecolog- ical data had to be taken from a “what’s left” universe. All samples for soils and pollen analyses were taken only from within an archaeologically undisturbed context. Within this restriction every attempt was made to take samples which represented the horizontal and vertical ex- tent of the site as well as the various types of features (i.e., mound strata, floor areas, pits, etc.). During the 1 976 sea- son, two sets of pollen samples were taken. A set of 12 one-inch core samples were taken from within the undis- turbed mound loadings and directly under Feature 1 (the outer stone wall). Also, four columns which had square cross-sections, 6 cm on each side, were cut through the mound loadings and building floor strata down into the natural subsoils. These columns were wrapped in plaster soaked cloths for shipment to Tinda Shane at the Pollen Laboratory, University of Minnesota. It was hoped that uncontaminated pollen might be found by taking samples from these columns under controlled laboratory condi- tions. No pollen was found in any of the samples ana- lyzed. Broken and fragmentary phytoliths were found in samples from the building floor; however, it could not be determined whether these were of ancient origin within the clays used to construct the floor or whether they were associated with flora blown or carried into the building during the time the floor was in use. During the second season. Dr. Shane came to the site to collect samples which were to receive preliminary process- ing immediately at facilities of Ohio University in Chilli- cothe. She determined that the best chance for finding intact pollen was on the under side of large in situ pieces of charred wood. Unfortunately, once again, no pollen was found. Soil samples for simple comparative chemical analyses were taken from representative mound and floor strata and floor features. Soil samples for thin section analyses were taken from archaeologically undisturbed soils within Feature 1 and within and below Feature 3C in order to study the soil structure. James Kerr of the U.S. Soils Ser- vice did additional field studies of soil structure and pos- sible origins on representative profile walls as well as in the local vicinity of the site. Charcoal samples for possible radiometric dating as- says were collected from all charcoal deposits found in an undisturbed context. In addition, Jeffrey Friedland of Earth Sciences took several burned clay samples from burned areas of the main floor (Feature 3) for archaeo- magnetic dating analysis. The major goals of the excavation were to collect dat- ing and ecological samples and to determine, if possible, the building pattern which was not reconstructible from previous field work. These goals were met. References Bender, Margaret M., David A. Baerreis, and Raymond L. Steventon 198! Further light on carbon isotopes and Hopewell agri- culture. American Antiquity 46(2):346— 353. Gadus, Eloise F. 1979 The Harness copper plate. Ohio Archaeologist 29(3): 27-29. Greber, N’omi 1979 A comparative study of site morphology and burial patterns at Edwin Harness mound and Seip mounds 1 and 2. In Hopewell archaeology: the Chillicothe con- ference., edited by David S. Brose and N’omi Greber, pp. 27-38. Kent State University Press, Kent, Ohio. Greber, N’omi, Richard S. Davis, and Ann DuFresne 1981 The micro component of the Ohio Hopewell lithic technology: bladelets. Annals of the New York Acad- emy of Science 376:489-528. Mills, William C. 1903 Diary. On file. Library Archives, Ohio Historical Center, Columbus, Ohio. 1907 Explorations of the Edwin Harness mound. Ohio Ar- chaeological and Historical Quarterly 16:1 13-193. Moorehead, Warren K. 1897 Report of field work carried out in the Muskingum, Scioto, and Ohio Valley during the season of 1896. Ohio Archaeological and Historical Quarterly 5: 165-274. Putnam, Frederick W. 1885 Explorations of the Harness mounds in the Scioto Valley, Ohio. In Peabody Museum 18th and 19th An- 22 N’OMI GREBER No. 39 nual Reports (1884-1885), bound in Peabody Mu- seum Reports 3(5-6):449-466. Squier, George Ephram, and E. H. Davis 1 848 Ancient monuments of the Mississippi Valley. Smith- sonian Contributions to Knowledge 1. Washington, D.C. Reprinted 1973 with introduction by James B. Griffin as Antiquities of the new world: early explora- tions in archaeology (Vol. 2). A.M.S. Press, New York, for Peabody Museum, Harvard University. Watson, Patty Jo 1976 In pursuit of prehistoric subsistence: a comparative account of some contemporary flotation techniques. Midcontinental Journal of Archaeology 1(1 ):77— 1 00. 3 THE SITE Introduction to the Stratigraphy: Site Profiles The site of the mound was a small knoll situated on a second terrace approximately 3 km east of the Scioto River (Fig. 2.1). The underlying geological strata are sorted glacial outwash (Wisconsin) on which have gener- ally developed Fox-Ockley soils with some associated Warsaw series. These soils are described by the U.S. Soil Survey. Fox Series In the Fox series are well-drained soils that developed on deposits of calcareous gravel and sand of Wisconsin glacial age. These soils are mainly on terraces (glacial outwash plains and valley trains) but locally are also on kames, eskers, and parts of moraines on uplands. Fox silt loams formed in 1 2 to 18 inches of silty material over gravel and sand, whereas the coarser textured Fox soils formed in loamy material over gravel and sand. On terraces the Fox soils occur with the Thackery, Sleeth, Westland, Wea, Warsaw, and Ockley soils. On uplands they occur principally with the Kendallville soils, though in a few places they are close to the Miami, Lorenzo, and Rodman soils. Fox silt loams resemble the Ockley soils but are not so deep to parent material. (Petro et al. 1967:133) Warsaw series The Warsaw series consists of dark-colored, well-drained soils on terraces that developed on stratified, calcareous gravel and sand outwash. These soils occur closely with Wea, Fox, Ockley, and Westland soils. Unlike the Fox soils, which developed under hardwood forest, the Warsaw soils developed under grass and have a darker colored A horizon containing more organic matter than the Fox soils. The Warsaw soils have a less silty upper solum, are shallower to calcareous gravel and sand, and are less acid than the Wea soils. (Petro et al 1967:151) In the following descriptions the Munsell color desig- nations are given in parentheses. The color name used in the text is the common visual color. The major part of the soils and gravels used in con- structing the various mound strata, the floor of the build- ing at the base of the mound, and the various features on that floor both within and without the building itself were the local Fox-Ockley soils and underlying gravels. The only exception was the outer cobble and bedded stone mantle (Feature 1). These stones were brought to the site from Dry Run, the banks of the Scioto to the west, or the hills at the east end of the Scioto Valley, here 6 km east of the site (Fig. 2.1). Within these stones was found a dark (10YR 4/4 dark yellowish brown and 5YR 3/2 dark red- dish brown) soft loam or fine silt which has characteristics associated with the Warsaw series. Jerry M. Bigham of the Department of Agronomy, Ohio State University, analyzed thin sections made from samples that James Kerr of the U.S. Soils Conservation Service had taken from within these soils (see Appendix 3. 1). These soils do not appear to have developed in place under grasses which may have covered the stones since the mound was built. It is possible, though probably unlikely, that they were placed there as fill by the original builders. It is more probable that they filtered down from upper mound lay- ers before overgrowth stabilized the upper strata. We at- tempted, with the assistance of Mr. Kerr, to find possible local sources of such soils. Unfortunately historic and possible prehistoric land use has disturbed the local area too heavily to enable such areas to be found. Pocket prair- ies did exist in this area of the Scioto Valley historically, for example, at Prairie Station 8.5 km north of Liberty. Also, in situ prairie soils were found under a section of the High Banks Earthworks 8 km north of the Edwin Harness site (Shane 1973, personal communication 1982). Thus, although the characteristics of the soils within the mantle are clear and different from the other soils we found in the mound, the exact origin of these dark soils has not been found. Prehistorically the surface of the knoll had been cleared into the B soils horizon with some minor filling required before the desired building level was achieved. Evidence of this filling was found in small pockets of soils with little to no structure at the upper edge of the undisturbed B horizon, as, for example. Sample II in Appendix 3. 1 . It is likely that materials which had been burned in the land clearing were incorporated into the layer which underlay much of the site (Feature 3C). This was a generally thin (5 cm), dark (7.5YR 3/2 dark brown) clay stratum (Sample I, Appendix 3. 1 ). It occurred, as did all of the floor strata, within the area defined by the heavy cobble wall (Feature 1 ). At the outer edges of this area Feature 3C was the only floor stratum present. Sketch profiles of the site are given in Figure 2.5. Prac- tical difficulties prevent the presentation of one-to-one scale profiles. The field data are of course available for study. The original land clearing, construction, and other cultural activities resulted in a complicated stratigraphy of which only disjointed and truncated remnants were left in 1975. Fortunately there was still enough information to allow reasonable reconstructions. The physical character- istics of the strata follow directly; the horizontal extent will be discussed further as the floor maps are discussed. Three profiles along grid N-S lines and three along grid E-W lines are shown. The profile along the E507.5 line (Section DD') gives a general profile of the building floors and remaining mound strata. Since the major axis of the building is east of grid north, no grid line gives a symmet- ric profile of the structure. Section EE' is a composite pro- file based on the floor strata along the E495 line with post 24 N’OMI GREBER No. 39 holes based on data farther to the east; thus this section represents a line parallel to the major axis of the Big House. The number of superimposed floor layers varied from one (Feature 3C) at the outer edges (east end of Section AA', north and south ends of Sections DD' and EE') to five (middle of Section DD'). All the tombs, pits, post holes, and almost all the burned areas were found on the topmost layer. Feature 3. This layer was a heavy gray clay (10YR 5/3 yellowish brown), which, as noted by previous excavators, appears to have been puddled, that is, mixed with water when put down. The major layer beneath this clay (Feature 65) was a very sandy orangish brown clay (7.5YR 5/8 strong brown), at times mottled with gray. Between these two heavy layers, a 2 mm layer of hard, reddish orange sand (7.5YR 5/6 strong brown) was frequently found. Feature 33, which was distinctive in color and texture, was found below Feature 65 on the E507.5 line but not on the E495 line. This stratum was composed of a red, sandy clay (5YR 5/6, 5YR 5/8) mottled with dark gray, pink, yellow, and light brown clays, and relatively evenly scattered charcoal bits, 0.5 to 3 cm in size. There were limited sec- tions within this layer, such as Feature 72 (a 50 cm by 75 cm area in N535 E509), which were of the same texture and mottling as the major extent of the feature but within a brown rather than a pink-red matrix. Feature 50, a soft, moist, dark gray-brown, slightly sandy clay (7.5YR 4/2 strong brown) was found below Feature 33. A construction break in Feature 3 was found in the E495 profile near N543. It appears likely that such a break also existed on the east side along E507.5, although the extensive prior excavations, as illustrated in the deep cut seen in the north end of Section DD', have destroyed the evidence. It is dear that north of this point, although the activity floor level and the construction materials appear to be the same as those of Features 3 and 65, there are differences. The floor is less consistent in total thickness, and there is no consistent ordering of gray clay over orang- ish brown sandy clay. Areas of both types of clay are in- terspersed as can be seen in Section AA'. The area shown at the south end of Section DD' shows another variation in floor construction. Here Feature 39, brown clay (7.5YR 5/6 strong brown) mixed with sands and gravels, lies above Feature 3C. This stratum has a different color and texture from that of Feature 65, which was found directly beneath the main gray floor under the structure. In the N-S profiles, the large posts which were set to the depths of the natural sandy gravel layer can be seen. In Section DD' the remains of the posts defining the East Structure can be seen at the base of the backfill. The loca- tion of Putnam's and part of Mills’s excavations can be seen in Section AA', an E-W profile along the N560 line. This is near the beginning of the wider portion of Put- nam’s trench (Greber 1 979: Fig. 6.4). Here the bottom of his trench was sloping down towards the main activity floor, which it had not yet reached. Mills did, as he wrote in 1907, dig up to the edge of Putnam’s trench and through the mound floor. The remnants of the mound construction itself can be seen in all the sketch profiles. What was an early, if not the first, covering over the parts of the Big House can be seen at the base of Putnam’s trench as well as farther east on the N560 line and at the ends of the N-S profiles. This covering is a brown sandy clay (7. 5YR 4/4 strong brown) with a few pebbles (ca. 12 mm). Loadings are usually dis- tinct and 25-50 cm in size. All of the soils we excavated appeared to contain no cultural debris. In some areas, such as at the north end of Section DD' where these soils have been placed over a cleared soil horizon rather than a prepared clay floor, the boundary between the fill and the in situ soils was less distinct. Several small features are probably a part of the early mound raising. Features 20 and 21, found in unit N560 E495 (Fig. 2.2), appear to be very large loadings or very small mounds of gravels and sands. Feature 43, found in unit N530 E492.5 (Fig. 2.3), was composed of sheet loadings as contrasted with con- tainer type loadings. Alternating sheets of sand, gray sandy clay, and charcoal were found in a very restricted area close to Feature 3. These were found at the edge of a historically disturbed area and could not be traced east to any significant extent. Covering the base of the first fill in most areas is Fea- ture 41, a 10 cm layer of pea gravels. These gravels are identical to those found in the local C horizon. This same type of material has been found as a major cover in many Hopewell contexts, for example, within the Seip Earth- works (Baby and Langlois 1979; Greber and D. Griffin 1982), at Mound City (Brown and Baby 1966), and in a number of other sites where Squier and Davis ( 1 848) de- scribed it as “sand.” A second stratum of mound fill found over Feature 41 is composed of a matrix similar in color to the first fill, but with a higher percentage of gravels so that loadings are more difficult to detect. These two strata are in effect a reversal of the natural B horizon in which the deeper soils contain more gravels. This suggests that the second cover- ing came from a greater depth, perhaps from the same borrow pit. The only artifacts which we found in the soils of this secondary fill were associated with Feature 56, a cremated burial. Feature 1, the bottom edge of a heavy cobble wall, was found over this secondary fill. The base of this wall varied from 1.5 to 2 m from inner to outer surfaces. It was com- posed of waterworn pebbles 10 cm or less in size and larger stones ranging up to 50 cm in length. Included were flint cobbles, sandstone, and limestone, both waterworn and bedded. All the stones in two 50-cm wide sections through the wall at N517 and N528 on the east side were saved as a representative sample. The average depth re- maining of the wall was 10 cm. In general, the smaller 1983 EDWIN HARNESS MOUND 25 cobbles were found on the inner edge with the larger stones on the outside. It is likely that the edge row of large stones extending almost 6 m north from N532 on the west side of the site is the end of the elliptical “ring” which Mills ( 1907, 1903) described on the west Hank of the mound. Over the secondary fill, which, as previously noted, ex- tended under and beyond Feature 1, were soils that ap- parently eroded down from higher parts of the mound. At times the boundary between these soils and the secondary mound fill was difficult to distinguish, possibly because these soils had a similar geological origin. The north end of Section DD' shows an example of this type of stratig- raphy. Auger samples taken on a line continuing north from this profile show that the slough layer ended within 10 m. To summarize, the remnant mound strata found from the southwest corner of the mound, continuing clockwise around the perimeter of the mound towards Backhoe Trench 2, were a first fill covered by a thin layer of pea gravel, followed by a secondary fill capped by a stone wall. The lower fills extended beyond the base of this wall. The secondary fill was covered over time with soil eroding from higher up the sides of the mound. There is a different stratigraphy on the southeast and south edges. In Section FF' Feature 69A is seen directly over the pea gravels of Feature 41. Here these gravels are more spo- radic and appear to have been disturbed, probably by the placement of this upper stratum, which is composed of a distinctive, dark sandy loam ( 10YR 3/2 very dark grayish brown) with a high density of small to medium gravels. We found no artifacts in these soils, nor in the similar soils of Feature 69 (Section BBO- Within both of these ex- tended dark strata were deposits or discrete sections which were darker (2.5YR 2.5/0 black) and more com- pact. These areas (Features 55, 68, 80, 81) contained a high density of rock, much of which was apparently fire cracked, and large pieces of charred wood and charcoal. Feature 55 extended 3 m by 2.6 m and was 55 cm deep. The rocks within the feature were predominantly lime- stone and ranged from pea gravels to 15 cm rocks; the average maximum fracture edge length was 10 cm. Below Feature 55 was a pit extending down into the C horizon. A bundle burial (Feature 60) was found at the base of this pit. No other pits were found associated with any of the other rock concentrations recorded. It is likely that the intrusion of Feature 69A under the rock wall which is seen in Section FF' is the result of pre- historic disturbance. The major archaeological distur- bances on the east side of the site prevent the location of the first fill in some profiles (see Section DD'). Here again the outer dark, humic layer has been placed up to or under the heavy rock wall. A small section of secondary fill ap- pears under the dark layer in Section BB' but not in Sec- tion FF'. The relationship of Features 69 and 69A to each other is not clear from the available evidence. To summarize the mound strata found in 1976-1977, the lowest fill found was composed of soils likely taken from the upper levels of the local B horizon. These soils contained no artifacts and were deposited over the main activity floor and outer fringes surrounding this main area. Over this fill, up to some unknown depth, was a thin layer of pea gravel identical to gravels found in the local C horizon. This thin layer extended outward and usually beyond the lower edges of the first fill. On top of this gravel was a second fill composed of soils apparently taken from lower levels of the local B horizon. These soils contained no artifacts except those associated with a cremated burial found on the east side of the site. On top of this fill was placed a stone wall which was probably intended to retain upper layers of mound fill. This wall was above the irregular outer edges of the prepared knoll surface; however, underlying mound strata extended beyond this wall. On the southern and southeastern edges of the site a dark, humic soil containing gravels but no artifacts was placed up to or upon the mound. There are indications that earlier mound strata were disturbed as this soil was added to the complex. Discrete areas or de- posits within this dark soil contained very dark, compact soils mixed with burned rock, predominately limestone, and burned wood and charcoal. No artifacts were found in these very dark soils. The exact horizontal extent of the one or several areas of these outer dark soil strata is not known. The Horizontal Extent of Floor Strata Mills’s general description of the “clay floor” is in ac- cordance with what we found; viz., “in some places it was only three or four inches in thickness, in other places from ten to twelve” (1907: 1 38). He also describes Feature 3C and the original soils below it: “The original surface of the site was covered with ashes and charcoal. . . . Nowhere were there evidences of any prolonged fire on the original surface, rather only the burning of small limbs as evi- denced by the charcoal remains” (1907:138). The exten- sive areas of Feature 3C which we found did not have large pieces of charcoal, even from the “burning of small limbs.” It may be the lack of large pieces rather than direct evidence of limbs which Mills is describing. The layer, as discussed above, was not in situ. We also did not find any evidence for burning on the natural soil horizon. Includ- ing Mills’s description. Feature 3C, the dark clay stratum, underlay the Big House and the entire activity area; this is generally the same area as defined by the inner edge of the heavy cobble wall. Feature 3C was under or within a me- ter or so of this wall edge around the entire site. The overall extent of the main activity floor, Feature 3, appears to have been from the outer edge of the portico on the north and south, definitely beyond the portico on the west. Feature 3B was a single layer above Feature 3C and abutted the western edge of Feature 3. It was more grav- 26 N’OMI GREBER No. 39 elly but generally similar to Feature 3. It was under the west end of Feature 6 and extended several meters south and appears to have been just a local variant in construc- tion materials. Feature 65 appears to have directly underlain Feature 3 in most areas of the site. At the outer edges there were some minor variations as to which extended farther (e.g., see Fig. 2.5). Feature 39 may have been a part of Feature 65; however, the heavy historic disturbances prevented the establishment of a direct connection. The slight differ- ences in texture may have been accidental or perhaps as- sociated with the end of the portico area. Feature 33 and the underlying Feature 50 appear to have been only in the northeastern area under parts of the Big House and portico ( Fig. 3.1). The horizontal extent of these features does not appear to coincide with any single portion or portions of the Big House. As described above. Feature 33 was the most distinctive floor stratum found. If we assume that the portions found in the same strati- graphic sequence around disturbances caused by Mills’s major excavations are the same stratum, then the west- ernmost piece of Feature 33 was near N546 E500, the northwest corner near N553 E505, and the southeastern edge near N535 E509. Although the exact shape cannot be known, the north and south edges appear generally to be oriented magnetic east-west. In the relatively extensive areas of this feature, which were hand troweled, no post pattern was found. We were rather hoping that an “Ade- na” house would appear. It is possible that a small struc- ture with shallowly set posts (less than 10 cm) could have existed in the southwest heavily disturbed section of the feature. I think it is more likely that the area had been used without a major structure. Perhaps the mixed burned clays and charcoal bits were culturally significant remains from the land clearing and other activities asso- ciated with the first use of the site. Such an origin, as care- fully cleaned up and redeposited debris, would be in keep- ing with the character of various types of features found on the main activity floor. The Building Post Pattern At the base of the Edwin Harness Mound were the re- mains of a large building, which in keeping with known Native American languages we can call “Big House.” This Fig. 3. 1. Cuts made by Mills in 1905 through the original floor layers. The layer (Feature 65), thus exposing the reddish clays of Feature 33. Unit main activity floor (Feature 3) has been cut through, as has the secondary N550 E5 10, level 5, 29 June 1977. 1983 EDWIN HARNESS MOUND 27 is a translation, for example, of the Shawnee m ’sikamekwi (Greber 1979:28) or the Creek tcoko-thlako (Hudson 1976:221). The same word is used in Shawnee for stomp dance ground (Charles Callendar, Case Western Reserve University, personal communication, 1983); thus, the English translation has connotations not of a large resi- dential house but of a special activity area. Also at the base of the mound were the remains of cultural activities which took place about the Big House within an area which, though not symmetric, was apparently well defined. There are several major parts of the Big House (see Figs. 2.4, 2.6). The floor plan outline of each of the largest two parts is a classic Ohio Hopewell nearly rectangular, rounded corner design (Baby 1971; Baby and Langlois 1979; Brown 1979). These two parts, the North Section and the Middle Section, are joined by a rectangular hall. The South Section, which is circular in floor plan, is joined directly to the Middle Section. On the east is a small, again classic in floor plan, structure which is joined to the other elements by a corridor area. A wall extends around the north side of the East Section. The post pat- tern of these basic parts has been abstracted from Figure 2.4 and presented in Figure 3.2. Because of the historic disturbances prior to 1976, the building posts were found at differing depths and with varying degrees of distur- bance from none to total. The latter was rare. The size of each post hole indicated in Figure 2.4 is the size first found in 1976 or 1977. In Figure 3.2 adjustments have been made to an estimated original floor. There is a portico-like area about the entire complex which is demarcated by large posts on all sides. On the south, west, and north along the line of these posts is a shallow, clay-filled trench containing small stake holes. These were likely supports for a screen or narrow wall which would have formed an enclosure about the Big House. There was no evidence of such a shallow trench on the east side of the house because of the extensive removal of the main floor area in the 1903 and 1905 excavations (Mills 1903, 1907). However, considering the east-west asymmetry of the complex, there may not have been such an enclosure on the east side. The two major elements are similar in general architec- tural design. There is a set of 48 inner posts (average post size 24 cm) which forms the structural strength of the building. These posts are arranged in seven rows N-S and E-W. There is some bending of the rows to turn the corners and some space left at the center of each building. The E-W separation of the rows is the same for both, averaging 1.6 m center to center. The N-S separation, however, is greater in the North Section (1.9m compared to 1.6 m). Thus there is a more spacious floor plan in the North Section, but the same number of structural sup- ports. The size and placement of these supports strongly suggest that there had been an upper floor or platform area. There was at least one post from the North Section which was just over 10 ft (3 m) high. This charred post. which was photographed by Mills in 1905, is probably the same one which was sketched by Putnam in his field notes (1884). Within the identified sample from the 1976-1977 excavations, the major construction timbers were young hickory trees (Table 5. 1). The exact length available from a tree which could have fitted into the recorded post holes cannot be calculated because of the many complex envi- ronmental variables involved. However, for similar en- vironments, 58 ft (17 m) heights are recorded for 18 cm diameter trees in Silvics of North America (Fowells 1965:126). The next outward series of posts in both the North and Middle Sections architecturally appear to have supported a facade or, more likely, a roof for the area defined by the 48 inner posts. There is some variation possible in the number of posts contained in this series because of the common boundaries with the hall and the South Section. This series contains posts of different diameters and dif- ferent spacing in the two sections. The M iddle Section has a paired post corner arrangement and at least 2 larger posts in the western line. The majority of these posts aver- age 18 cm in diameter; a few are 25 cm, while the 2 large posts on the west are 40 cm. There are a total of 39 posts in the series plus, possibly, Posts 378 and 387 (Fig. 2.4). For the North Section there are probably 42 posts in this se- ries, plus possibly Posts 42 and 72. The average size of the northern posts is 23 cm. In general, these posts are more widely spaced than those of the Middle Section. There is relatively closer spacing at the corners. The larger post size for this post series in the North Section may be related to the longer spacing between posts. It is possible that in the original building design the same number of posts had been used in this series in both sections. The last series of posts differs on the east and west sides of both structures. Those on the east (13 cm average di- ameter) are evenly spaced and may form a formal facade or decorative line. Those on the west side are also mainly small, but their placement is irregular and scattered. They do not appear to be structural posts. Four short lines of posts form a hallway or corridor between the North and Middle Sections. As is shown in Figure 3.2, the South Section is outlined by posts of three size ranges: 7 cm average diameter, 15 cm, and 24 cm. The central posts of this element may have served as structural supports; however, the contents of at least one of these (Post Hole 216) suggest that they also served other functions. The materials recovered from post holes are described in another section of this chapter. The East Section is outlined by small posts (10 cm). This, as the South Section, is of small scale. It does not appear likely that there had been an upper level in this structure. The outer wall on the north side is composed of posts 18 cm in average diameter. The outer post holes which ring the complex average 27 cm in diameter on the west side and 1 8 cm on the east side. Next to or about these posts on three sides of the structure indicated are estimates of original sizes on the main floor of the building prior 1983 EDWIN HARNESS MOUND 29 was a shallow (8-13 cm), narrow (15-20 cm), clay-filled trench (Features 22, 59, 85, 88) which contained small stake holes ranging in size from 2 to 5 cm (average 3.6 cm). Along the east side, these holes were at varying angles in the clay such that stakes set in the holes would tilt, appar- ently at random, towards the north or south but would remain in the same plane. These stakes likely were part of a screen or fence which was constructed after the main floor had been completed. All the floor graves appear to have been within the area enclosed by the trench and outer posts. Other types of remains of cultural activities — mainly burned areas and deposits, a few post holes, and pits — were found outside this boundary. Other Posts The remaining post holes found fall into three size classes: large (over 25 cm), medium ( 10-20 cm), and small (less than 10 cm). The number of each size associated with the several parts of the Big House are given in Table 3.1. The range in the depth from the top of the main floor (Feature 3) to the bottom of the medium-sized post holes which were found basically intact was 23 to 40 cm. The corresponding depth for the small ones was 7 to 19 cm. Thus there may have been additional small posts in sec- tions where Mills removed both Feature 3 and Feature 65 (total depth ca. 12 cm). The main sections where this had been done were near the center of the site in the vicinity of the hall and the East Section. It is less likely that evidence of medium-sized post holes was completely destroyed in the earlier excavations. The small posts found associated with the South Sec- tion are generally about the perimeter. The four medium and large posts are more centrally located. In the Middle Structure, the largest number (12) of the medium posts found form the last series of posts on the west side of the structure as described above. Approximately half (17) of the small (here 3 cm) posts are located along the east side from the center line of the structure to about 4 m south of the center. The remaining small posts are scattered less densely (Fig. 2.4). TABLE 3.1 Non-structural Posts Associated with Edwin Harness Big House Location Corridor East Diameter North Middle East Central Section South (cm) Section Section Section Hall to Hal! Section <10 39 31 0 0 2 14 10-20 40 25 1 7 5 4 >25 2 6 0 0 0 1 Total 81 62 1 7 7 19 In the North Section the medium posts are also in an irregular line on the west side. Others were found within the last western row of the central structural posts, out- side the northern end, in the northeast corner, and as part of the grouping which forms the central focus of the building. A few posts holes were found outside the portico: Post Hole 2 (20 cm diameter) on Feature 3 near Feature 1 7 and four post holes a short distance north (79,59) or east (156,206) of the main complex (see Fig. 2.6). Contents and Description In general the main structural post holes were straight walled, frequently with a clay lining below floor levels, and also frequently bottoming in the C horizon gravels where drainage is excellent. An exact inventory of the contents of these posts cannot be given due to Mills’s ex- tensive excavations (which were appropriate) and the loss of his field notes (which is unfortunate). Describing the contents, he says, “Very frequently these molds would contain broken animal bones, mussel shells and occa- sionally a piece of mica. We have never been able to find in the great number of molds examined, any implements or ornaments” (1907:138). Any of these objects could have been accidentally swept into the hole by the prehis- toric occupants of the building during the cleaning up of activity debris from nearby floor areas. Such cleaning would be in character with the apparent manner in which shallow deposits of mixed materials had been made on the floor (see following section). The majority of objects which we found within post holes were similar to those found by Mills and appear to have been accidental inclu- sions. The small rocks, pieces of fabric, and perhaps sherds, which we also found, were likely used as wedges. However, there were incidences of apparently deliberate filling of some post holes. These post holes were con- structed in the same manner as were others of similar diameter. Although numbers of posts were burned before the first stage of mound building, there were posts which were re- moved. This procedure has also been found in our current work within the Seip Earthworks. James Brown of Northwestern University has discussed with me similar findings at Mound City, and there are reports of similar activities (for example, at Garden Creek, North Carolina) from Middle Woodland sites outside Ohio (Chapman and Keel 1979). At least one post hole was filled with a collection of materials similar to the deposits on the floor. Post Hole 216 contained a variety of burned woods (Table 5.1), ash, shell, bone, apparently fire cracked rock, and mica pieces, some of which were 8 by 3 cm. The soil surrounding these objects was oily. Another non-structural post which had been removed was Post Hole 178, immediately north of the Big House. A small egg-shaped clay basin had been Fig. 3.3. Location of post holes with surrounding stains. Only those post holes which were sur- rounded by rings of iron stain are indicated by solid circles. Small posts have been relatively enlarged for visibility. See also post descriptions in chap. 3, Fig. 2.4, Table 3. 1 , and Table 5. 1 . 1983 EDWIN HARNESS MOUND 31 constructed over this hole. The basin (Feature 62) had contained a fire which reddened some of the surrounding clays (5YR 4/6 yellowish red). The fire, as in Post Hole 216, had contained several varieties of wood, including the second recorded occurrence of pine (Table 5.1), and six fragments of mammal bones. The post had also been removed from Post Hole 25. This was a 26 cm E-W by 30 cm N-S hole which extended 56 cm below the floor surface and was located along the narrow trench on the west side of the North Section. In this hole a human skull and mandible was found. Parts of the skull were colored red, apparently from red ocher. Holes had been drilled in the ascending rami of the mandi- ble in such a fashion that skull and mandible could have been articulated. The skull has been identified by Ray- mond Baby (personal communication 1977) as that of a young adult male. No ocher, charcoal, or any other cul- tural materials were found within this post hole. One characteristic that divided the post holes into two classes regardless of size was a “rusty” stained ring which surrounded 85 of the post holes (Fig. 3.3). This ring, which averaged 2 cm in thickness, was irregularly shaped and ranged from less than 1 to 25 cm from the edges of the post holes. It was surrounded in most instances by a fine black line. The area between it and the post hole tended to have a “crusty” texture. Putnam’s description of such rings in his field notes coincides with our findings in the central area of the North Section, which was within Put- nam’s trench. This stain was present about the entire post hole; thus, evidence for its existence was not destroyed by previous excavations (see Fig. 3.4). These rings appear to be caused by the movement through the soils of water soluble forms of iron (this leaves an orange stain) which are preceded by the movement of manganese (this forms a thin black line). I have not yet been able to find suitable chemical tests to prove that these rings were caused by deterioration of coloring used on the posts; 1 believe that they were. The patterning of these posts is shown in Fig. 3.3. Of the nine structural posts with such stains in the North Section, one was within the central focus area; six were in the southeast corner; two were near the middle hall. The north three posts on either side of the hall had stains, as did the post in the Middle Section opposite the center line of the hall and the stained post in the North Section. The other stained structural posts in the Middle Section were paired corner posts, two pairs on the west side, one on the east. In the South Section, the stained posts appeared to be symmetrically arranged with respect to the two oak posts (Table 5. 1). This pattern may indi- cate that the entrance to the structure was towards the southwest, and Post Hole 216 would form a line with this entry way. The posts forming the west side of the East Sec- tion and one additional post on the north also had ac- companying rust stains. The area between the East Sec- tion and the middle hall had such posts along the north and the west sides. Fig. 3.4. Examples of post holes with surrounding stains. The rings of iron stain continued the length of the posts as illustrated in the middle sections of the site where Mills removed extensive portions of the floor strata and post holes were found in 1 976 and 1977 under extensive back- fill. Unit N537.5 E497.5, level 3, 75 cm below base of SW stake, PM 34, 37, and 38, 29 July 1976. These posts are on the west side of the Middle Hall. Of the non-structural posts found in the North Struc- ture which had rust stains, six were within the central fo- cus, single posts were at the southwest and northeast corners, and four posts were just outside the north end. In the Middle Section most of the small clustered posts on the east side had stains, as did several within the last south row of the main structural posts. In summary, at least half of these specially marked posts were located near entranceways or near the perime- ters of parts of the Big House. The others appeared to be marking various special areas within the interior of the building. Central Focus of the North Section The center of the North Section had a more compli- cated configuration than that of the Middle Section. This was an area which had been excavated by all previous expeditions except those of Moorehead. The most infor- mation on the original deposits about this area is found in 32 N’OMI GREBER No. 39 Putnam’s field notes. The small (Post Hole 432) and medium-sized (Post Hole 444) posts set on either side of the geometric center of the building were surrounded by rusty stains, as were the other non-structural posts about the main posts 23 and 462 in the next row towards the west. The southern edge of the large (ca. 1.5 by 4.5 m) deposit of burned matting and charcoal described by Putnam was apparently located at the northern end of the area marked by these stained posts. The long axis of the deposit was E-W; thus, it extended relatively symmetri- cally on both sides of the building center line. In the Peabody Catalogue (Number 34982) the deposit is de- scribed as being “1 to 4 inches thick and on burnt clay about 1 foot from the bottom of the mound.” The field notes call the floor on which the tombs were found “hard pan”; this is the same as our Feature 3. The stratigraphy containing the charcoal deposit is given in a label for a sketch as “hard pan” at the base followed by “clay & gravel 6 inch [s/c], sand /i inch, pure clay 4 inches, char- coal layer 2-4 inches, clay, gravel & loam.” Three feet (ca. 1 m) above this deposit was a 1 -ft (30- cm) thick deposit of sands and gravels which extended 3 by 4 ft ( 1 by 1.3 m). A “basin-like cavity” apparently asso- ciated with these sands and gravels contained mica, shell, flint, carved bone, and an unusual human effigy pipe (see Fig. 3.5). The horizontal location of the gravels and cavity are not given; thus, it is not clear whether this cavity was deliberately created or the result of a cave-in of the roof of one of the tombs underneath the gravel deposits. One of the few extended burials found in the mound was located at (or possibly on) the eastern edge of the matting. This is “Skeleton 3” in “Burial Chamber No. 6.” A copper plate had been placed lengthwise on the chest of the individual whose head was towards the southeast. A second ex- tended burial “Skeleton 1,” also with head towards the southeast, was located just north of the northwest corner of the burned deposit. An extensive ash deposit was found southwest of the charcoal deposit in the vicinity of Feature 24 at the southwestern edge of the central space. Another unusual specimen, a carved stone sphere (Willoughby 1916: PI. lOi), was found at the edge of this ash area, which is de- scribed in the field notes as “distinct masses occurring from a few inches to 3 feet above the clay.” In summary, the archaeological evidence about the central space of the North Section suggests much more extensive activity than was found in the center of the Middle Section. The scale of these deposits is also greater than that of any of the other deposits which we found. The major deposit found by the schoolboys was probably part of this central focus. No other extensive deposits have been reported in any of the available records. Burned Areas on the Floor All the burned floor areas found, except Features 18 and 36, were reddened and fire hardened to a depth of 2 Fig. 3.5. Pipe fragment (4.5 cm X 3.5 cm) found by F. W. Putnam over deposit of burned matting immediately north of geometric center of the North Section. Photograph courtesy of Peabody Museum of Ethnology and Archaeology, Harvard University. cm or less. Similarly, the individual charcoal deposits were also thin. Features 18 and 36, which were within the Big House, were different from each other. Feature 18, which was in the southwest corner of the North Section, appears to be the remnant of a heavily fired prepared basin. There was only a strip 40 cm long by 1 cm wide remaining in place, but large chunks (7X 15 cm) of heavily burned clay of the same color and texture were found in the nearby backfill. Feature 36, which was at the center of the Middle Section, contained the deepest evidence of fire harden- ing found. Here in an area 120 cm N-S by 70 cm E-W, hard, red sand (2.5YR 4/6 red, 5YR 5/8 yellowish red) was found through the gray clay floor. The soils were red- dened and hardened to the depths of the underlying C gravels. Thin (1-2 cm) layers of cemented sand and pea gravels and ash (0.5-4 cm) were found on top of the sand. Unfortunately, this, as all fired areas found within the Big House, had been archaeologically disturbed to some de- gree. In a sample taken from the in situ layers of Feature 36, pine wood has been found. This is one of only two examples of this type of wood in the identified floral sam- ples (Table 5.1). Mills in his report describes fires set on top of graves. Such fires may have been the origin of the burned areas on the portico immediately west of the North Section (Fea- ture 78, about Features 28 and 29); and the areas within the South Section (Features 82B and 92). No artifacts were found associated with these or any of the other fired areas within the Big House or the portico area. 1983 EDWIN HARNESS MOUND 33 More complete stratigraphy is known for the fired areas found on the west and north sections of the site. In some of these areas, ash and any burned matter had been cleaned away and only the reddened and fire hardened clay floor was found (Feature 22 and sections of Feature 16). Other discrete areas (Features 6, 7, 8, 9, 10, 49) were fire reddened floor areas covered with a thin (2 cm or less) lense of ash and charcoal, and then a final covering of sand, gravels, or clay. None of these features contained any artifacts. The areas in which some materials were found (Fea- tures 31, 44, 45, 46, 51, 53, and 53A) consisted of thin deposits of mixed charred and sometimes uncharred materials, placed on burned and un burned sections of the floor. As is described in the discussion of stratigraphy, the floor west of the Big House was somewhat different in character from that on the north. North of the building there were more patchy and more mottled sections of floor. Frequently there was only one layer above Feature 3C. This peripheral floor was gray clay, as Feature 3, or orange-brown sandy clay, as Feature 65, which was stra- tigraphically directly beneath Feature 3 under the Big House. Thus the charcoal deposits shown in Figure 2.6 which are not on a gray floor were found on an orange- brown sandy clay floor. The deposits of materials were not the results of in situ burning. For example, in Feature 3 1 a thin ( 1-4 cm) layer of sterile brown sand had been placed over a cleaned red- dened floor area; then a layer (1-2 cm) of thoroughly mixed charcoal, fired clay nodules, and an unidentified mammal bone fragment were placed on top of the sand. In Feature 44, a series of thin layers containing burned beads, cut mica fragments, broken bladelets, and bone fragments were deposited over burned and unburned areas of the peripheral mound floor. The three whole plus several fragments of clay beads and the single fresh water pearl found were likely part of a necklace, which included many canines (see chap. 6). These burnt and broken drilled canines included fox and raccoon. Other faunal material in the deposit included 51 turtle shell fragments; 2 bird bones; 1 bird talon; 3 catfish vertebrae, one of which was modified; 13 mammal bones; and 42 unidenti- fiable bone fragments. All bone was burnt. Twenty-seven bladelets have been identified from 35 pieces. Feature 45 was similar in construction to Feature 44 but contained less material. There were mica fragments; clumps of burnt sand; 28 burnt mammal bone fragments, 3 unburnt; 3 bird bone fragments; 5 unidentifiable bone fragments; several fossil fragments; and one bladelet mid- section (see also Tables 5.2 and 5.3). Feature 46 contained only 5 baked clay nodules and 2 burnt mammal bones mixed in with the charcoal deposit. Feature 51 was smaller in extent than the other deposits and contained a flint chip, 1 1 mammal bones, and 1 left distal deer radius. In Feature 53, a 1-2 cm layer of gray clay had been placed on a fire reddened surface (Feature 53A); another layer of dark clay ( 1 cm thick) had been placed over this gray clay. On top of the dark clay was found 1-2 cm of fire reddened clay covered with scattered charcoal, bladelet fragments, and bone fragments; all these were covered by a mottled gray clay. The 3 bladelet pieces represented 2 different bladelets. Seven mammal bone fragments, 1 burnt mammal bone fragment, 1 unidentifiable bone fragment, and several flint flakes were in the deposit. To summarize, each deposit outside the Big House was carefully prepared; the number of artifacts and volume of deposit were small. The unburned objects found may have been used in the activities immediately associated with the lighting of the fire which burned the remaining materials. No pottery sherds were found in any of these deposits. Feature 54A is an unusual deposit associated with an 80 cm diameter, 19 cm deep depression found in the main floor west of the Middle Section (Fig. 2.6). Feature 3 thinned into a negligible thickness at the center of this depression. Feature 65, which was about 2 cm thick here, followed the line of the depression. A thin layer of dark clay with deposits of charcoal and charred twigs was found between the two main floor layers. At 80-100 cm from the depression center Features 3 and 65 were, as more usual, 5-6 cm thick. The deposit on Feature 65 was offset from the center of the depression (see Fig. 2.6). Soils which were the same in color and in texture as the local B horizon were found extending into the yellow C gravels immediately below the depression. A thin layer ( 1 m diameter) of charcoal was found 19 cm below the center of the depression; however, not the center, but the south edge of this deposit was directly below the center of the depression. A few apparently fire cracked rocks and flint flakes were found within the charcoal on Feature 65; noth- ing was found within the lower deposit. The origin of the depression is not clear. Perhaps it resulted from the set- tling of a relatively large section of fill needed during the original land clearing; perhaps it was intentionally con- structed. There was no open arch above this depression as was usually found over the collapsed tomb roofs (e.g., Putnam 1884). The small trench about the portico area was cut through the depression; the trench also cut along the east edge of the heaviest section of the charcoal de- posit on Feature 65. There were three other deposits of charcoal found below the main activity floor (Feature 3). Feature 57, a deposit of hard ash and charcoal, 1-2 cm thick as usual, extended 40 cm N-S by 60 cm E-W on Feature 50 (below Feature 33) at N548 between E506 and E505. There were scattered bits of charcoal found 10 to 75 cm about this deposit. There was a small (5 X 15 cm) area of orange stain found near one edge, but there was no definite evidence for in situ burning. Similarly, Feature 97 was found on an unburned surface. A thin (less than 2 cm) layer of powdery charcoal lay between Feature 3 and Feature 65 at the edge of the disturbed main floor (N542.5 E495). It extended N-S, and was apparently the western edge of the 34 N’OMI GREBER No. 39 original deposit; no significant extent was found in trac- ing this deposit west. DIC-662 is from this feature (Table 3.2). Feature 73 was a 90 cm N-S by 33 cm E-W layer of charcoal deposited on gray clay east of the North Section and below Feature 33. A dark red burned area was found at the southern edge of the feature. The charcoal was covered with a thin layer of mottled yellow-gray clay which contained many small limestone pebbles. There were also limestone pebbles mixed with the charcoal found in the northern end of the feature. Ex- cept perhaps for the deeper deposits below Feature 54, the deposits of charcoal found under the main activity floor appeared to be similar in character to those found upon the floor; that is, they were carefully constructed, thin, and with few artifacts. Pits Sketches of various pit shapes found are shown in Fig- ure 3.6. Feature 17 was the only pit found which was on the main floor outside the Big House. Features 19, 30, 79, and 89 were within the Middle Structure; Feature 91 was in the North Structure. This known distribution may be close to the original distribution of relatively deep pits since at least the bottom of such features would have been noted in the heavily disturbed areas. If the known distri- bution is the original distribution, then such pits occurred only on the west side of the Big House and the portico. Shallowly cut features, if any, were lost. The apparent major difference between the contents of the pits and the floor deposits discussed above was the presence of pottery in three pits. Probably one cord- marked vessel was represented in Feature 19 and two in Feature 30, while Feature 89 contained cordmarked and Hopewell series sherds, parts of possibly six to nine ves- sels (see pottery analysis in chapter 4). Feature 19 also contained 3 mammal canines; 75 unidentified bone frag- ments (40 burnt, 35 unburnt); 7 shell fragments; and a mica fragment. Feature 30 contained similar items. These were 3 mammal canines; 1 fragment of deer ulna; 96 other burnt bone fragments, 9 unburnt; 13 shell fragments (see shell analysis in chapter 8); 2 mica fragments; 15 flint flakes; 4 fire cracked rocks; and 4 fossils. Feature 89 con- tained 7 burnt canines; 55 burnt unidentified bone frag- ments, 45 unburnt; 6 shell fragments; 10 pieces of worked flint; 2 bladelets; 1 clay bead; 12 fossils; and numerous small mica fragments. Although Features 79 and 91 were heavily disturbed, they contained some charcoal as did the other pits. Also within Feature 91, 12 unidentified bone fragments and 30 identified shell fragments were found. TABLE 3.2 Radiocarbon Dates Radiocarbon Lab. ft Provenience Material Years (B.P.) Comment DIC-661 Fea. 17 Wood Charcoal 1490± 65 DIC-662 Under Fea. 3 Wood Charcoal 2150 ± 155 sample size, 4 g** DIC-663 Fea. 19 Wood Charcoal I620± 65 D1C-664 Fea. 30 Wood Charcoal 1500± 60 D1C-664 Rerun Fea. 30 Unused Benzene from original burn 1600 ± 65 D1C-665 PM. 32 Wood Charcoal 1820 ± 70 D1C-801 PM. 36 Wood Charcoal 1900 + 460 sample too small; indicator date only -500 DIC-802 Fea. 3 1 Wood Charcoal 1630+ 70 DIC-860 Fea. 53A Wood Charcoal 1500+ 50 D1C-II89 Fea. 69 Charcoal [Gleditsia (Sp) 100%]* Modern burnt honey locust root over Fea. 68 DIC-1 187 Fea. 62 Charcoal [Carya (Sp) 55% Quercus (white) 40% Quercus (Sp) 5%]* 1770+ 50 DIC-I 188 Fea. 8 1 Charcoal [Carya (Sp) 55% Fraxinus (Sp) 15% Juglans (Sp) 20% Quercus (white) 10%]* 1 140 + 60 D1C-I 190 Fea. 55 Charcoal [Quercus (red group) 100%]* 1110+50 D1C-1635 Fea. 56 Burnt bone 1200+ 65 ♦Percentages of 20 random pieces identified by University of Michigan Ethnobotanical Lab. ♦♦Using long-term average background count, the date is 1980 ± 1 55 b.p. Clay FEA. 79 Fig. 3.6. Sketch profiles of pit features: Fea. 17, 120 cm N-S X 72 cm E-W, 45 cm deep; Fea. 30, 48 cm N-S X 53 cm E-W, 77 cm deep; Fea. 79, 40 cm X 37.5 cm, 9 cm deep, N-S diagonal 43 cm, E-W diagonal 44 cm; Fea. 89, top 34 cm N-S X 32 cm E-W, base 55 cm N-S X 53 cm E-W, depth 71 cm. (Grid north referenced.) 36 N’OMI GREBER No. 39 Feature 17, which was in situ, showed careful cutting, filling, and covering. Within it were found 4 fragments of human bone, fragments of at least 2 canines, 1 bird bone, 108 mammal bone fragments, 129 unidentified bone fragments, and 2 flint flakes. Large amounts of wood charcoal were found as well as wild food remains (see flora analysis in chap. 5). The rather idiosyncratic nature of the pit profiles and the limited quantities of artifacts found within the pits appeared to reflect single use. They likely contained the remains of one cultural event, which was probably associated with the performance of a ritual or ceremony. Other Features The historic pits noted in Fig. 2.6 are obvious, but of course are not the only spots which were literally dug up prior to 1976. Feature 24 was, as previously described, associated with the work of Squier and Davis ( 1 848) fol- lowed by just about everyone else. The digging style shown in Feature 77 I would identify with “The Boys,” even though Putnam’s notes may indicate another loca- tion for the schoolboys’ second pit. The remaining pits 1 assume were dug about post holes, aboriginal pits, or other possibly deep features. The bone concentrations were usually associated with smears of charcoal in the backdirt. No artifacts can be tied to these. The disturbed graves (see Fig. 2.6) were very disturbed. The location of several outside the Big House but within the area defined by the portico posts does give useful in- formation on the accepted use of these spaces. Within Feature 84 the remains of an extended infant skeleton were found. These remains are discussed in chapter 7 and in Greber (1979). Mica fragments were found in the dirt above the grave, but no artifacts were found associated with the remains in 1977. Two in situ burials were found in the outer areas of the site. Feature 56 was a deposit of cremated human remains centered at N534.4 E512.2 within a secondary mound fill (see Fig. 2.5A Section CCT Associated with the remains were a copper plate (22.6 X 12.7 cm) (see Gadus 1979), a small copper adze (9X 5 X 1.06 cm thick) with some fabric still intact, and a slab of sandstone. This slab was nearly rectangular with maximum dimensions 84.4 X 32.5 cm, weight 54.7 kilos. There was a small amount of pecking at the corners, but the 5 mm depth of the weathering rind suggests that this was a weathered rock before it was placed in the second mound fill (Paul Clifford, Curator of Geology, CMNH, personal communication, 1977). No evidence of a prepared grave was found in the gravelly clay matrix in which the bones were found (see Introduc- tion to the Stratigraphy in this chapter). The second burial found had been placed at the bottom of a pit which was under Feature 69A (see Fig. 2.5B Sec- tion FF'). This was a bundle burial which was accompa- nied by a cut marine shell. David Morse, whose identifica- tion and analysis of the molluscs recovered is given in chapter 8, has separately described this shell. The shell has been modified by the removal of the exterior spikes and trimmed along the edge; probably for use as some type of container. This shell belongs to the species Busy con contrarium (lightning welk) and is native to the western Atlantic coast from North Carolina to Florida. The animal lives primarily in shallow waters. Of all the species in the genus Busvcon , this species is one of the most southerly and restricted species described in the zoological literature. Originally the shell was a whitish buff in color with violet and brown vertical streaks. The shell measures 17.5 cm in length which is average for shells of this species today, but is one of the smallest speci- mens known from the Edwin Harness Mound, (personal communication, 1978) A variety of burned hardwoods as well as a fragment of Zea mays and other seeds were found in flotation samples taken from the earth surrounding the redeposited bones (see Table 5.3). Feature 63 was a small pit centered 2 m south of the edge of the pit containing Feature 60, in the east wall of Backhoe Trench 4. The pit appeared to be generally oval (21 X 35 cm) and 49 cm deep. Charcoal flecks were found in the lower half of the pit within a dark reddish sandy clay matrix (5YR 4/2 yellowish red). No artifacts were found. Feature 37 was a tap root centered at N522.75 E508.7 near the edge of Feature 1. Radiocarbon Assays The results of a series of radiocarbon assays on samples collected during the 1976-1977 excavations are given in Table 3.2. Two of these are not useful DIC-801, an indi- cator date, and DIC-1 189, which is modern. Feature 69 is an extensive deposit of dark earth and rock at the outer edges of the mound with occasional discrete concentra- tions of darker earth, fire cracked rock, and charcoal. Apparently a modern root intruded into this feature di- rectly over but not into one of these darker deposits (Fea- ture 68). DIC-1 190 and 1188 are dates from deposits which are similar to Feature 68 and in similar contexts. The state of the art in radiocarbon work has changed even since 1977, so that samples submitted today can be smaller than the suggested 10-g size with no adjustments necessary in procedures. At the time of the DIC-662 as- say, 4 g of submitted charcoal resulted in a very small prepared sample (0.6869 g). For such a prepared sample size the background count used in calculating the radio- carbon years may be based on the average background count over a month rather than the average for only the 1983 EDWIN HARNESS MOUND 37 two days adjacent to the day on which the sample count is taken. (Irene Stehli, Director, Dicarb Radioisotopes, Inc., personal communication, 1977). Either of the calen- drical dates (200 b.c. ± 155or30B.c.± 155) will be strati- graphically in order with the remaining dates. The sample was taken from a small deposit of burned material be- tween Feature 3 (the central main activity floor) and Fea- ture 65 (sandy clay layer directly below Feature 3) as dis- cussed more fully in the section on the horizontal extent of floor strata. DIC-663 (a.d. 330 ± 65) and DIC-664 (first run a.d. 450 ± 65, second run a.d. 350 ± 65) are from pits within the Middle Section. These pits were cut into and through the main floor (Feature 3). DIC-665 (a.d. 1 30 ± 70) is from a post hole immediately northeast of the Northern Section. DIC-661 (a.d. 460± 65) is from a pit cut through the main floor west of the North Section, while DIC-802 (a.d. 320 ± 70) is from a deposit of burned material on the same floor, again, west of the Middle Section. North of the en- tire building, DIC-860 (a.d. 450 ± 50) comes from a sim- ilar deposit of burned material, and DIC- 1 187 (a.d. 1 80 ± 50) comes from burned hardwoods found in a small basin (Feature 62) constructed over a post hole (see Fig. 2.6 for locations). The date for Feature 56 ( DIC- 1 635, a.d. 750 ± 65) is again stratigraphically in order since this burial was in an outer stratum of the mound construction. The data is very close to the dates from the concentration of black earth, burned rock, and charcoal in the outer strata (DIC-1 188, a.d. 81 0± 60 and DIC-1 190, a.d. 840 ± 50). There can be questions concerning the retention of humic materials in bone use for radiocarbon assays; the standard procedures used for removing these younger contaminates from wood charcoal destroy bone collagen. This destruction would give too young a date. It seems reasonable to as- sume that bone, particularly burned bone, will not ac- cumulate as much humic matter. The bone used for DIC- 1635 was somewhat protected from water by its general location, which was within a gravelly, easily draining soil and beneath an upper strata of heavy stone. In the contin- uing series of dates from Ohio Hopewell sites, other sam- ples of bone and charcoal from a single provenience are being processed. For now, this single date is in proper sequence for site stratigraphy, but it is considered ten- tative. The application of corrections for the Seuss Effect does not significantly change the calendrical years at the time period of this study. For example, using the Arizona cor- rections (Damon, Ferguson, Long, and Wallick 1974) DIC-1635 becomes a.d. 770 ± 83. Other examples using both the Arizona and the Masca correction procedures for dates from Ohio Hopewell sites have been given pre- viously (Greber 1976: Fig. 24). The increase in the stan- dard deviations, which is a result of the correction, may be worth consideration. Comparative materials will be dis- cussed in more detail in the report conclusions. Appendix 3. 1 Thin Section Analysis of Soil Samples From Edwin Harness Mound Jerry M. Bigham The Ohio State University Sample I. (SW end of backhoe trench 7; N W wall. Feature 3C). 1 found no evidence of clay films in this sample and little indica- tion of the structural aggregation which characterizes undis- turbed soil materials. Vertical thin sections revealed several con- tinuous lenses of charcoal (one very prominent), but charcoal fragments were also dispersed throughout the sample. Iron stains often, but not always, paralleled the charcoal lenses. The thin, lenticular nature of these iron stains is quite uncharacteris- tic of soils. Since soil will often redden when heated, perhaps these stains represent materials which have been fired. In addi- tion to these features, I also noted several filled worm casts, but I cannot say if the activity was recent or relict. Sample II. (As for I, but directly below Feature 3C). I saw no evidence of clay films or charcoal in either vertical or horizontal thin sections from this sample. Iron stains, however, were com- mon. In soils, iron oxides often segregate to form mottles and concretions. In contrast, the stains in this sample usually oc- curred as coatings on large grains and/or rock fragments (generally siltstone). In some instances, the stains proved to be completely oxidized fragments of unknown origin. The contacts between the iron stains and the surrounding matrix were gener- ally quite abrupt, suggesting perhaps a mixture of materials. Sample II also contained numerous worm casts and was quite porous (but not as much so as Samples 1 and III). Sample III. (N522.5 E487.5, South Wall, Feature 1). Most of this sample was too loose to impregnate. However, we did man- age to save one unit consisting of a siltstone fragment overlain by unconsolidated material. In thin sections, the siltstone frag- ment appeared to be quite fresh with no evidence of iron stains or weathering rinds. The unconsolidated material was exceed- ingly porous, and I saw no evidence of clay skins, iron stains, or charcoal. In all three samples the sand grains appeared to be fresh, rela- tively angular and unsorted. All three samples also contained more feldspar minerals (I, III, II) than I am accustomed to see- ing in soils. I doubt, however, if there is any significance to this observation. All samples were organic stained to some extent, but I saw little evidence of primary (undecomposed) root tissues. Subsurface layers in soils of this region often contain “clay skins” along root channels and the surfaces of soil aggre- gates due to the dispersion and downward movement of colloi- dal particles in waters percolating through surface layers. Trans- located clay is, therefore, indicative of soil formation. I saw no evidence of translocated clay in any of the samples I examined. Without further knowledge of the excavation site, 1 would con- clude that these samples were taken from disturbed and/or re- cently deposited materials that have not been subjected to soil forming processes over a significant period of time. This state- ment is based on the dark color, absence of clay films, unusual porosity, charcoal content (where present), and lack of natural aggregation (soil structure) in these materials. 38 N’OMI GREBER No. 39 References Baby, Raymond S. 1971 Prehistoric architecture: a study of house types in the Ohio Valley. Ohio Journal of Science 71(4): 1 93— 198. Baby, Raymond S., and Suzanne M. Langlois 1979 Seip Mound State Memorial: nonmortuary aspects of Hopewell. In Hopewell archaeology: the Chillicothe conference , edited by David S. Brose and N’omi Greber, pp. 16-18. Kent State University Press, Kent, Ohio. Brown, James A. 1979 Charnel houses and mortuary crypts: disposal of the dead in the Middle Woodland Period. In Hopewell archaeology: the Chillicothe conference , edited by David S. Brose and N’omi Greber, pp. 21 1-219. Kent State University Press, Kent, Ohio. Brown, James A., and Raymond S. Baby 1966 Mound City revisited. MS on file. Department of Archaeology, Ohio Historical Society, Columbus, Ohio. Chapman, Jefferson, and Bennie C. Keel 1979 Candy Creek-Connestee components in eastern Ten- nessee and western North Carolina and their relation- ship with Adena-Hopewell. In Hopewell archaeology: the Chillicothe conference , edited by David S. Brose and N’omi Greber, pp. 157-161. Kent State University Press, Kent, Ohio. Damon, P. E., C. W. Ferguson, A. Long, and E. I. Wallick 1974 Dendrochronologic calibration of the radiocarbon time scale. American Antiquity 39(2) part 1:350-366. Fowells, H. A. 1965 Silvics of forest trees of the United States. Division of Timber Management Research Forest Service. U S. Department of Agriculture Handbook No. 271. Gadus, Eloise F. 1979 The Harness copper plate. Ohio Archaeologist 29(3): 27-29. Greber, N’omi 1976 Within Ohio Hopewell: analysis of burial patterns from several classic sites. Ph.D. dissertation. Depart- ment of Anthropology, Case Western Reserve Uni- versity. University Microfilms, Ann Arbor, Michigan. 1979 A comparative study of site morphology and burial patterns at Edwin Harness mound and Seip mounds 1 and 2. In Hopewell archaeology: the Chillicothe con- ference, edited by David S. Brose and N’omi Greber, pp. 27-38. Kent State University Press, Kent, Ohio. Greber, N’omi, and Dennis P. Griffin 1982 Comparison of excavations and subsurface remote sensing data from sections of the Seip Earthworks Complex, Ross County, Ohio. Paper presented at the 1982 Annual Meeting of the Southeastern Archaeo- logical Conference, Memphis, Tennessee, October 28-30. Hudson, Charles 1976 The southeastern Indians. University of Tennessee Press, Knoxville, Tennessee. Mills, William C. 1903 Diary. On file. Library Archives, Ohio Historical Center, Columbus, Ohio. 1907 Explorations of the Edwin Harness mound. Ohio Ar- chaeological and Historical Quarterly 16:1 13-193. Petro, James H., William H. Shumate, and Marion F. Tabb 1967 Soil survey, Ross County, Ohio. United States De- partment of Agriculture Soil Conservation Service in Cooperation with Ohio Department of Natural Re- sources. Putnam, Frederick W. 1884 Field notes on excavation of Liberty Group, Ross County, Ohio. On file, Peabody Museum, Harvard University. 1885 Explorations of the Harness mounds in the Scioto Valley, Ohio. In Peabody Museum 18th and 19th An- nual Reports (1884-1885), bound in Peabody Mu- seum Reports 3(5-6):449-466. Shane, Orrin C., Ill 1973 Report on excavations at the High Banks earthwork, Ross County, Ohio. Paper presented at the Annual Meeting of the Ohio Academy of Sciences, Cleveland, Ohio. Squier, George Ephram, and E. H. Davis 1 848 Ancient monuments of the M ississippi Valley. Smith- sonian Contributions to Knowledge 1 . Washington, D.C. Reprinted 1973 with introduction by James B. Griffin as Antiquities of the new world: early explora- tions in archaeology (Vol. 2). A.M.S. Press, New York, for Peabody Museum, Harvard University. Willoughby, Charles C. 1916 The art of the great earthwork builders of Ohio. An- nual Report, Smithsonian Institution, Washington, D.C. 4 THE CERAMIC COMPLEX JAMES B. GRIFFIN In this discussion of the pottery from N’omi Greber’s 1976-1977 excavations of the Edwin H. Harness Mound floor, the specimens have been described and identified according to their location (see Appendix 4. 1). This col- lection has provided information on some of the pottery which was in existence before the completion of the mound and in some instances before different sections of the mound floor had been formed. These pottery frag- ments are not the pottery complex of the Harness people but a very small segment of their pottery production. This mound has had a long history of excavations since at least the mid- 1840s, and the pottery preserved from those ex- cavations in the Peabody Museum of Archaeology and Ethnology, Harvard University, and the Ohio Historical Society probably represent only a small part of the speci- mens in the area excavated. 1 did not find any pottery from Harness in the British Museum Squier and Davis Collection. In addition, the ceramic material from the several segments of the Harness mound area forms a small portion of the ceramic production of the popula- tions who lived at and near the Harness site and partici- pated in the activities centered there. The same is true of all of the other excavated major and minor Hopewell sites in Ohio. Only the Turner site has a pottery sample which can be regarded-as representative, yet that collection also presents many difficulties of interpretation. If the major Ohio Hopewell sites existed over a period of several generations with a population associated with each of some hundreds of people, their year-by-year pottery pro- duction and breakage would reasonably come to a total far in excess of that recovered from the excavations. The McGraw site is one small unit of a village presumably as- sociated with a small Hopewell earthwork, and it pro- duced almost ten thousand sherds. The Turner and Mar- riott sites had 3,806 sherds according to Prufer’s (1968) analysis of the pottery. There are about 100 counted ex- amples from the Edwin Harness floor in the present col- lection, although the total is somewhat greater, for, when possible, sherds belonging to a single vessel were counted as one example. Prufer’s count, from the Harness mound excavations at the Peabody Museum and the Ohio His- torical Society, plus the Russell Brown collection, totals some 1, 1 57 sherds. In the Greber collection of 98 sherds, McGraw Cord- marked is the dominant surface finish and McGraw Plain a distant second (Table 4. 1 ). There are a few examples of the Hopewell Rim and Chillicothe Plain Rocker-stamped. There are a few examples of Turner Simple Stamped A, but none of the micaceous or other sand tempered simple ware, although there were some from the earlier collec- tions. There are a few sherds of Turner Check Stamped, almost certainly from one vessel. The most unusual pot- tery is the vessel represented by limestone tempered Turner Simple Stamped A, sherds which could well have been an import. I have not provided a percentage figure for the identified “types” because they would be even more misleading than usual and can be easily obtained by anyone wishing to do so. A description of each sherd is given by provenience in Appendix 4.1. Photographs are presented in Figures 4. 1-4.4. TABLE 4.1 Summary Tabulation of Harness 1976-1977 Pottery Collection Sherd Count* McGraw Cordmarked Body 56 McGraw Plain Body 7 Rim 3 Chillicothe Plain Rocker Stamped Body 4 Hopewell Rim 2 Turner Simple Stamped A Rim 1 Body 3 Turner Check Stamped Body I Unidentifiable 21 Total 98 * All pottery pieces which can be fitted together into one unit are considered to be one sherd. In 1968, Olaf H. Prufer published an analysis of Ohio Hopewell pottery from a study he had made some years earlier. His typology will be followed in this study in order to enhance comparability. In 1943 Richard G. Morgan, Curator of Archaeology of the Ohio State Museum (now the Ohio Historical Society), and I made a similar study, which I used briefly in a report on Adena pottery (Griffin 1945), but our complete study was not published. Our sherd count is presented in Table 4.2 and is given in Pruf- er’s terminology where possible. We did not use type names in our original study except for the identification of the highly distinctive Hopewell Rim style. Fig. 4. 1 . Hopewellian and check-stamped sherds: a(24E), 6(40A), and d{ 225B), Hopewell Rim; c(2E), e(24C), J[ 103E), g(103E), /?(35B), y'(24H), A:(24G), and /(24B), Chillicothe Plain Rocker Stamped; ;'(103E), Hopewell sherd with incised line; m( 290C), n( 290C), o(290G), p(290C), and g(290C), thin check-stamped sherds. Fig. 4.2. McGraw Cordmarked, McGraw Plain, and unidenti- Cordmarked; e(290C),y(138A, 23 IB) unidentified, g(2B), h( 1 23), fied sherds: a(308C), />(2D,I2H), c(37A),and ^(309A), McGraw /(219V), and /(229C). McGraw Plain. 42 JAMES B. GRIFFIN No. 39 TABLE 4.2 Griffin-Morgan Pottery Count of Ohio State Museum Harness Collection Sherd Count* Sub-total Total McGraw Cordmarked Body 78 Rim 2 80 McGraw Plain Body 18 Rim 7 25 Turner Check Stamped (square) Body (all from one Footed vessel) 23 Turner Check Stamped (diamond) Body 8 Rim 3 34 Turner Simple Stamped A Grit Tempered Body 6 Limestone Tempered Body 7 Limestone Tempered Plain Rims (Probably Turner Simple Stamped A) 3 16 Hopewell Rims 4 4 Chillicothe Plain Rocker Stamped Body 17 Dentate Rocker Stamped Body 8 Banded Dentate Rocker Stamped Body 1 Zoned Dentate Rocker Stamped Body 2 28 Total 187 187 * All pottery pieces which can be fitted together into one unit are considered to be one sherd. Discussion McGraw Cordmarked The McGraw Cordmarked vessels from Greber’s Har- ness floor collection display some variety. On some the cord impressions are close together and on others are widely spaced. Some of the sherds were smoothed while others were not. There is also some variation in thickness, from 4 to 8 mm. There is variation in the amount and character of the grit temper. All of this variation suggests somewhat less attention to the acquisition of raw mate- rials and vessel manufacture than was evident in sherds of the Hopewellian Series. The same can be said regarding the few McGraw Plain sherds in this collection. The McGraw Cordmarked sherds illustrated by Prufer ( 1968: PI. 3) are representative of the sherds at hand. There was a time when some archaeologists did not like to think that cordmarked pottery was a part of the Ohio Hopewell Complex. However, this pottery is a strong component of almost all known Ohio Hopewell sites ex- cept Tremper. While usually regarded as "utility’' ware, which it certainly was, cordmarked pottery is also found in burial association as whole or broken vessels at Turner, Hopewell, and the Martin Mound in Coshocton County in the upper Walhonding valley (Mortine and Randles 1978), where it was the surface finish on a small tetrapod vessel. The size of the reconstructed McGraw Cordmarked vessel 20A, of the McGraw Plain vessel 27 1 0, and of the Turner Simple Stamped 308A suggests that they were probably whole vessels which had been used for some function shortly before they were deposited. What that function (or functions) was is not clear. The same may be said for the larger Hopewell Rims and plain rocker- stamped sherds which seem to be from the same vessel. Hopewell Style Pottery The Hopewell Rims and Chillicothe Plain Rocker- Stamped sherds are the only representatives of Hopewell style pottery in Greber’s collection from the Harness floor. This is well-made pottery, and the slight camber of the rim I regard as indicating the vessels were made fairly early in the life-span of this complex. The reconstructed body fragment (Prufer 1968: PI. 5a) is almost certainly part of the same vessel as my Figure 4. Ic, f-h. The illus- trated Hopewell Rims from the older Harness collections are not as well made and appear to have a more pro- nounced camber (Prufer 1 968: PI. 5b-c). The Hopewell Zoned-dentate Rocker-stamped vertical compound jar from Putnam’s Harness 4 in the Peabody Museum is a unique specimen, and the large rim section from Russell Brown Mound 1, one would think, might well fit into the reconstructed area on the Peabody vessel (cf. Prufer I968:P1. 2 and 46). This would raise the possibility of Put- nam’s Mound 4 being the same mound as Frank Soday’s Mound !. Close comparison, however, by Prufer and others does not support the idea of a single vessel. There are a few other sherds of the Hopewellian Series from Harness and the Brown mounds, but my impression is that Harness does not have the variety of this ceramic complex that is found at Seip, Turner, or Mound City. The McGraw site ceramic collection is also much more varied, and I interpret its Hopewellian Series as later than the available material at Harness, Mound City, and Seip. The Harness examples should be close to the initial ap- pearance of the Hopewellian Series in the Scioto Valley. There is no evidence in southern Ohio for a develop- mental sequence of ceramic decoration which could have developed into the Hopewell Zoned decorated style. While some of the late Adena tablets in Ohio and Ken- tucky do present conventionalized bird designs, the only northern area where zoned stamped decoration appears is in the lower half of the Illinois Valley, where it is found apparently in a time period which precedes the Ohio Hopewell development. I have not seen a vessel or any vessel fragments from Ohio which could be correctly as- 1983 EDWIN HARNESS MOUND 43 signed to an Illinois source. On the other hand, such ves- sels from the southern half of the Illinois Valley were car- ried into northern Illinois, Michigan, and Wisconsin. The most logical area for the generation of the Hopewell Zoned style is in the Illinois Valley even though it is by no means a certainty. Turner Simple Stamped B The known distribution of Turner Simple Stamped ware recorded by Prufer is given in Table 4.3. There are no examples of Turner Simple Stamped B from Greber’s excavation of the Edwin Harness Mound, nor were any identified from the collections made by earlier excavators. There were none from Russell Brown Mounds 1 and 2. There were none reported by Prufer (1968) from Hope- well, Rockhold, Ater, or Marriott 1, which is part of the Turner site. No examples are reported from sites ex- amined in the Scioto Valley survey (Prufer 1967) nor were any mentioned as present from the excavations of the McGraw site (Prufer et al. 1965). At the Turner site, which, of course, has the strongest representation of Turner Simple Stamped B (see Table 4.3 and Prufer 1 968: PI. 34a and c, PI. 40, PI. 44d and g), TABLE 4.3 The Occurrence of Turner Simple Stamped B in Ohio, based on Prufer ( 1968) Site Provenience Body Rim Tetrapod " 'feet " Total Turner Mound 1 , fill 2 2 Mound 3, unit 2 unit 4 12 3 Mound 4, unit 1 40 4 3 Embankment trench 13 2 Cemetery, unit 6 2 Ginther Mound 4 2 Tremper 2 1 Mound Mound 13 1* City General 7 Seip Mound 1, unit 957/ 161 3 1 1 unit 957/ 237 & 238 1 unit 957/260 1 Mound 2, general 5 1 General 10 1 1 Fort Ancient 1 4 Fort Hill 1 Russell Brown Md. 3 3 108 12 14 134 Willoughby had noticed that “some of the clay used in making the smaller and more delicate vessels was tem- pered with sand instead of crushed stone.” He mentioned a tetrapod base and ten other “feet.” He also observed of the complicated stamped sherds from Turner that “sherds showing ornamental paddle marks were extremely rare. Such vessels may have been brought from the southern Appalachian region, or they may possibly have been made by captured women from the South” (Willoughby and Hooton 1922:93). Prufer published a type description for Turner Simple Stamped B (1968), which he included in his Southeastern Series. He felt that type was imported into Ohio from the southeast, primarily because of its sand temper, which is rather rare in the Ohio Valley, and because of the narrow stamp impressions which were similar to Deptford and Mossy Oak Simple Stamped of the Georgia area. He also recognized, as had others, similarities to Paintville Simple Stamped of eastern Kentucky and to Bluff Creek Simple Stamped of northwest Alabama. The connections to the north Georgia area were thought to be particularly strong, and this area was considered as the probable source of imports into Ohio because of the presence in some of the examples of small mica flakes in the paste. Since the publications of Prufer, excavations in south- western North Carolina and in eastern Tennessee have produced ceramic data which serve to alter significantly interpretations of the source for Turner Simple Stamped types. In southwestern North Carolina a number of Mid- dle Woodland sites have been excavated and identified as members of a Connestee phase. One of the sites is Garden Creek Mound No. 2. One of the ceramic types is Connes- tee Simple Stamped with about 300 examples. Keel (1976:1 10) says that “some of the sherds . . . could be classified as Turner Simple Stamped. ’’One lower rim and upper body (Keel 1 976: PI. 16d) has annular punctates in a horizontal row at the base of a smoothed lower rim. In his discussion of trade pottery at Garden Creek, Keel identi- fies the same fine sand or finely ground limestone tem- pered pottery as Turner Simple Stamped (1976:120) and illustrates (PI. 181) two lower rim and upper body sherds with annular punctates in the horizontal row at the base of a smoothed rim which he calls Turner Simple Stamped B. But in the description of Tennessee types he states that the 92 limestone tempered sherds are all regarded as im- ported from Tennessee, although other sources were pos- sible. He particularly identifies the Ice House Bottom site with its Connestee material as similar to the limestone Connestee phase pottery at Garden Creek (1976:1 18). I am puzzled why Chapman and Keel (1979:157) regard sherds identified as Turner Simple Stamped sand tem- pered as probably being from Ohio, when Prufer and at least some other archaeologists have regarded this type as trade pottery from the southeast into Ohio. If the sherds illustrated by Keel ( 1 976: PI. 171) as Turner Simple Stamped B are typical, then there are no known examples at Edwin Harness Mound of this type. In comparing the *whole vessel 44 JAMES B. GRIFFIN No. 39 lower rim punctates at Garden Creek with the Mound City whole vessel from Mound 13, one should note that the former punctates are annular while the latter are verti- cally placed hemiconical punctates. It may be doubted that any simple-stamped sherds at Garden Creek were traded from Ohio. There are, however, Hopewell Cross-hatched rims and plain and dentate rocker-stamped sherds that might well be Ohio imports along with Ohio Flint Ridge blades. It is of special interest to note that Keel identifies the figurine specimen at Garden Creek as having been made from lo- cal paste similar to that of the Connestee pottery. At the Ice House Bottom site in Monroe County, Ten- nessee on the Little Tennessee River there is a ceramic complex which includes simple-stamped pottery that has even more interesting similarities to the Ohio and North Carolina pottery with the same surface finish (Chapman 1973). The Connestee Simple Stamped sherds from this site were so named because of their striking similarity to the Garden Creek and other North Carolina sites with Connestee Series pottery. In contrast to Garden Creek, the Ice House Bottom Collection has both annular as well as angular punctates, and it is the later technique which is found at Mound City, Ohio in Mound 13. The rectangu- lar punctates also appear on what could be Connestee Plain rims; both annular and angular punctates are also found on Connestee Brushed. The Connestee Series pottery at Ice House Bottom constitutes about 20% of the ceramic assemblage, while the Candy Creek Series with limestone temper is, at 71%, the dominant pottery. Chapman uses the term “Bluff Creek” for the simple-stamped pottery with limestone temper following the practice initiated by Haag ( 1 942a) in northwestern Alabama and subsequently followed by others for materials from that area. Its use in eastern Ten- nessee, however, is perhaps a misnomer even though there are obvious resemblances in the material from the two areas. The Ice House Bottom Bluff Creek Simple Stamped pottery does not have punctates similar to those of Con- nestee, and the vessel rims are primarily vertical instead of flaring, which is the dominant form on Connestee Simple Stamped. Over half of the lips of Bluff Creek are notched transversely, which does not appear on Connestee wares at this site or in Ohio. Ice House Bottom perhaps does have a simple-stamped ware which is close to the lime- stone tempered vessel from Harness. The rim and lip treatment at Harness is not, however, either illustrated or described by Chapman or by Gleeson (1970) from earlier excavations. The excavations at the C and O Mounds and village site in Jonathan County, eastern Kentucky, on Levissa Fork of the Big Sandy River recovered a small number of simple-stamped sherds. However, this pottery, described as Paintsville Simple Stamped (Haag 1942b), presents some problems in interpretation for several reasons. In his description of this type Haag says that perhaps all of the sherds could be from one vessel. The large rim (Haag 1 942b:Fig. 17:3) has a high flaring upper rim and annular punctates in a horizontal row at the base of the lower rim. Also an exact provenience within the two excavated mounds and village is not given. Since the flint projectile points from the site range in age from Early Archaic to perhaps Fort Ancient, and the pottery range is from an Adena complex to perhaps Fort Ancient, the attribution of Paintsville Simple Stamped is a bit difficult. There are Montgomery Incised examples and the Adena pottery complex as a whole would seem to be late. The Paintsville Simple Stamped is probably a trade vessel or vessels from eastern Tennessee. This would fit well with the presence of mica at the C and O Mounds and its probable derivation from western North Carolina. There are no specifically Hopewellian artifacts at the C and O Mounds, and if the Adena occupation there were in existence during the life- span of Ohio Hopewell, one would expect to find some indication of such contemporaneity. A look at one final possible southern connection indi- cates that the three sand tempered simple-stamped sherds illustrated in the Tunacunnhee site report from northwest Georgia (Jefferies 1976) cannot be considered close to Turner Simple Stamped B, and the one limestone tem- pered simple-stamped sherd is not close to Turner Simple Stamped A. Turner Simple Stamped A Turner Simple Stamped A sherds at Edwin Harness probably all belong to one vessel (cat. nos. 308A and 237C, plus other limestone tempered examples), except for two grit tempered specimens with a thinner body and less con- spicuous lands and grooves. The sherds of the large vessel are very close in appearance to the sherds illustrated by Prufer ( 1968:P1. 4b and c) from older Harness collections and from Russell Brown Mound 2, which was a part of the Harness Earthwork Complex. For a listing of the oc- currence of Turner Simple Stamped A as stated by Prufer see Table 4.4. As already mentioned, the Ice House Bottom Bluff Creek Simple Stamped illustrated examples do have an appearance similar to Turner Simple Stamped A. This is also true of other Middle Woodland sites in eastern Ten- nessee. At the Pittman-Alder site in Marion County, a short distance southwest of the bridge on which U.S. Highways 41, 64, and 72 cross the Tennessee River, a mi- nority ware (58 examples) of the Middle Woodland com- plex is Bluff Creek (Faulkner and Graham 1965:P1. XXIV). One of these rims has a notched lip (Faulkner and Graham 1965:59). Only two sherds from the site were identified as Benson Simple Stamped, which is the sand tempered variant name for northeastern Alabama (Heim- lich 1952). On the south side of the Tennessee River oppo- site Pittman-Alder is the Lay site, where the continuing excavations in the Nickajack Reservoir by the University 1983 EDWIN HARNESS MOUND 45 TABLE 4.4 The Occurrence of Turner Simple Stamped A in Ohio, based on Prufer ( 1968) Site Provenience Body Rim “feet" bn* Total Rockhold 1 1 Hopewell Mound 2 1 I Mound 17 1 1 General 40 1 1 40 Tremper Mound 3 3 Seip Mound 1 , unit 957/216 1 1 Mound 2 1 1 General 20 2 22 Fort Ancient 12 Turner Mound 3, unit 4 1 1 1 1 Mound 4, unit 2 3 1 2 Mound 6 3 3 Mound 9, unit 2 1 1 Embankment 3 1 2 Harness Edwin Harness Mound 13 8 5 Russell Brown Mound 1 34 2 36 Mound 2 153 4 157 Mound 3 4 4 293 8 4 305 *limestone temper **grit temper of Tennessee uncovered more and better examples, some 919 sherds, of Bluff Creek Simple Stamped (Faulkner and Graham 1 966: PI. XIII, 37-38). They identified this type with late Early Woodland in eastern Tennessee at the time of writing the report, but a later placement in Middle Woodland is probably a better assessment. At the Doughty site in Loudon County, Tennessee, just east of the 1-75 bridge over the Tennessee River, a few sherds of a Bluff Creek Simple Stamped vessel were found (McCollough and Faulkner 1 973: PL 21G). At the nearby Higgs site there were also a few examples (McCollough and Faulkner 1973:89-91). In the Tims Ford Reservoir along the Elk River in Franklin County, Tennessee, at the Mason site, a very minor type is Bluff Creek Simple Stamped (Faulkner 1968:78 and PI. IXG). Other sites in this area with a Middle Woodland limestone tempered complex should also have this type as a minor component. When working with the Norris Basin pottery in the mid 1930s, I identified simple-stamped specimens as “combed” or striated (Griffin 1938). A few such examples were found at Rock Shelter-cave Sites 3 and 12. Examples are illustrated in W. S. Webb’s section of the report (Webb 1938:P1. 13) from Saltpeter Cave (Site 3), and from Wal- lace Cave (Site 12) in my section of the report on Plate 1 52. While limestone tempered, none of the examples are similar to the Harness simple-stamped example. The sites were located in the Clinch-Powell river drainage a short distance north of Norris Dam in Campbell County, Ten- nessee. It is unfortunate that these shelters could not have been excavated in a manner which might have aided a recognition of successive Woodland occupations, for the pottery suggests a time span from Early Woodland cer- tainly well into the Middle Woodland time period. Two of the open sites, the Harris Farm, Site 9, and the Cox Mound, Site 19, also had a few specimens of simple- stamped or brushed grit tempered surfaces (Griffin 1938:305). In the lower Ohio Valley there are two sites which have simple-stamped pottery pertinent to the presence of this ceramic technique in Ohio. The Mann site in Posey County, Indiana, southeast of Mount Vernon, is a large site of some 200 acres with a strong Hopewellian compo- nent(Adams 1949; Kellar 1979, 1973). In the collection of pottery available to Adams, simple-stamped sherds were a minority type, but there are indications from later Indi- ana University work of areas of the site where it occurs in much larger numbers. Adams emphasizes that this type at Mann was sand tempered with small fragments of mica in the paste. He referred to similar inclusions in southeast- ern simple-stamped pottery, and he also noted that there were levels at the Angel Site with micaceous sands identi- fied by soil tests (Adams 1949:59). Kellar’s resume men- tions the Mann site and illustrates a simple-stamped rim with closely spaced lip notches (Kellar 1973:45-46). There Fig. 4.3. Turner Simple Stamped and unidentified simple stamped sherds: a(308A), />(237C), c and 4309A), e( 1 38C), /(308A), gand h( 23 1 A), i(23 1 F), /(237C), and Ar(237E), Turner Simple Stamped A; /and m(2!9A), unusually thin hard simple stamped. Fig. 4.4. McGraw Plain and McGraw Cordmarked sherds: a(2710), McGraw Plain; 6(20A), McGraw Cordmarked. 48 JAMES B. GRIFFIN No. 39 are also complicated stamped sherds from the Mann site. One of these is illustrated by Kellar, and six presumably from there are illustrated by Adams ( 1 949: PI. V). In the Rutherford Mound in Hardin County, Illinois, overlooking the Saline and Ohio river bottoms, M. F. Fowler (1957) excavated a small simple-stamped vessel with tetrapods in association with Burial 6, which was placed with four other burials in the primary mound about 2.5 feet above the mound floor. These burials may or may not be significantly later than the effigy platform pipes, panpipes, copper earspools, and axes placed with burials on the mound floor. While there is a Carbon-14 date of about a. d. 432 ± 1 00, 1 believe it is too late for the Hopewell items and for the tetrapod vessel, for this mate- rial should date close to a.d. 100. While the stamped paddle pottery seems to disappear after the Hopewell occupation in southern Ohio, such is not the case in the lower Wabash Valley. The survey re- port for the Illinois side of the river by H. D. Winters ( 1 963) shows simple-stamped pottery which he has named Embarrass (locally pronounced “Ambraw”) Simple Stamped. Some of the lips are notched. This type is asso- ciated with the FaMotte-Allison complex, which I believe straddles the rather arbitrary dividing line between late Middle Woodland and early Fate Woodland. Check- stamped pottery is also associated with this complex. Check-stamped There are five body sherds with check-stamped impres- sions from Harness. Prufer has called the Ohio Hopewell examples Turner Check Stamped and placed them in his Southeastern Series. Some of the Ohio examples may really be trade vessels from the east Tennessee area. The Harness examples of Greber’s collection, however, are blackened on both surfaces, and their paste characteris- tics are like the paste of the Hopewell Rims and Chilli- cothe Plain Rocker-stamped Harness sherds and are probably the product of Harness potters. I do not believe I have seen very many similar check-stamped sherds from other Ohio Hopewell sites and Prufer (1968) does not il- lustrate any. I have not seen such sherds from southeast- ern collections or in illustrations from that area. The con- cept of check-stamping did, however, almost certainly reach southern Ohio from the southeast, probably the east Tennessee area. I think that the absence of check- stamping on most of the Adena pottery in Kentucky, Ohio, and adjoining areas probably has temporal signifi- cance. One limestone tempered check-stamped sherd was at the late Adena Wright mounds (Haag 1940:8 1) and one grit tempered sherd from Jo9 (Haag 1942b:348). Haag il- lustrates two check-stamped sherds from Wright Mound 6 ( 1 940: PI. 52n and r). Prufer has identified a diamond check-stamped in his Harness site study, a Turner Check Stamped at Russell Brown Mound 1, and two of the same type at Russell Brown 3. None of those are illustrated from the Brown mounds but the Harness example is illustrated in Plate 4. Such sherds also occur in very small frequencies at Rock- hold; at Seip General and Seip Mound 2 General; at Fort Hill; and at Turner Mounds 1, 3, 4, 7, and 9, the Great Embankment, and the cemetery area. The lower rim and body sherd from Turner Mound 4 is illustrated by Prufer ( 1 968: PI. 39a) and has a horizontal incised line separating the plain lower rim from the stamped body. The sherd is placed sideways on the plate instead of vertically. It has stamp size and black outer surface similar to the five spec- imens from the Edwin Harness floor (cat. no. 290C). Prufer does not report check-stamped sherds from Hope- well, Ater, Ginther, Mound City, Tremper, Fort Ancient, or the Marriott Mound at the Turner site. There are cer- tainly more sherds from the Turner excavations than from all the other Ohio Hopewell sites where such sherds are known to occur. This is probably in part because of the much larger amount of pottery excavated and pre- served from the Turner site. At Turner, check-stamped sherds are found on both grit and limestone tempered paste, the latter being most common. One sherd from Seip General is limestone tempered, and the diamond check-stamped sherd from Seip 2 General is sand tem- pered. The few other sherds are grit tempered. The check-stamped presence in Hopewell is even more difficult to pin down in terms of its derivation than was the simple-stamped technique. None of the Ohio exam- ples are very large and very few have rims which might be helpful. In eastern Tennessee, check-stamped is found along with simple-stamped on many of the sites with Middle Woodland components. This is probably the area from which the check-stamped vessels or the concept was moved north into southern Ohio. One large rim and body sherd from the Turner cemetery is identified by Prufer ( 1 968 : PL 45a) as an Untyped Complicated Stamped be- cause it has large squares with a raised central circular area, which is unusual. Another location where such a pattern appears in the southeast is near Savannah, Geor- gia. Quite a few years ago J. C. Caldwell sent me illustra- tions of sherds that he identified as the Oemler complex, which had a limited distribution in the coastal area. Such material is not mentioned or illustrated in the Waring papers, although the Oemler site is identified on maps (Williams 1 968: Figs. 35 and 37). The Oemler style of stamping is regarded by some archaeologists to date about 600 b.c. (DePratter 1979). A connection between the Turner and Oemler examples is not implied, for they are quite different in appearance, and the vessel shape at Turner with its fairly high rim which is angled outward is quite different in shape from any of the Oemler complex. Three sherds with this design are known from the Mann site in southwest Indiana (Kellar 1979: 103). Thediamond variant of the check-stamp is uncommon both in Ohio Hopewell and in the southeast but does occur at the Yearwood site in southwestern Tennessee (Butler 1979: 1983 EDWIN HARNESS MOUND 49 Fig. 20.9). The ones at Harness in the older collections and other Ohio examples are probably local products. Complicated-stamped While there are no complicated-stamped sherds at Harness, they do occur at Seip and Turner particularly. The illustrated examples appear to have their closest con- nections to Early to Middle Swift Creek types. Sites in eastern Tennessee of the Middle Woodland period which have check-, complicated-, and simple-stamping, with ev- idence of trade sherds and other items probably made in Ohio, are the most likely candidates to be on the right temporal level to have furnished those pottery specimens or concepts to Ohio Hopewell. As mentioned, complicated-stamped pottery occurs at the Mann site (Kellar 1979:103). One rim illustrated by Adams (1949: PI. V) has closely spaced lip notches while a second has an undulating appearance. Adams expressed the opinion that these specimens — because of their grit, sand, and clay temper designs and notched lips — more closely resembled Swift Creek pottery in Georgia than the geographically closer Pickwick Complicated Stamped of northern Alabama. This was also my reaction when I saw them in the early 1940s (Griffin 1946:71) and still is. Rocker-stamped The presence of rocker-stamped pottery in the Greber collection or in the earlier collections from the Edwin Harness site is not easily explained. The Hopewell Zoned Plain Rocker-stamped vertical compound vessel, recon- structed from sherds obtained by Putman from his exca- vations, might be regarded as having been deposited as a whole vessel. The plain rocker-stamped sherd illustrated by Mills ( 1907:Fig. 36) is on the right hand side of the four glued sherds illustrated by Prufer ( 1 968: PI. 5a), and these appear to be part of those in the Greber collection. The latter specimens were relatively close together in the northwest part of the southern sub-rectangular structure. It is also possible that Mills’s ( 1 907 : Pig. 37) and Prufer’s ( 1 968: PI. 4b) are from the same vessel as Greber’s check- stamped sherds (cat. no. 290C) from Feature 66, which is also in the southern sub-rectangular structure. Notes from the Griffin-Morgan study state that the check- stamped sherds from the Mills Harness collection seem to be from one footed or tetrapod vessel. Conclusions This brief discussion and survey of the distribution of stamped wares, particularly simple-stamped, indicates the general area of possible southeastern connection for Ohio Hopewell. At present, eastern Tennessee and south- western North Carolina are the more probable loci for the derivation of simple-stamped and check-stamped vessels into the central Ohio Valley or for the manufacturing techniques which produced them. At some of the sites mentioned there are pottery types and other artifacts from Ohio. Some part of this interarea diffusion probably represents the activity of individuals from southern Ohio who participated in the acquisition of mica from North Carolina and in the acquisition of marine shell and other items from the Florida Gulf Coast. A more intensive ef- fort should be made to identify other sites in northeastern Tennessee, along the headwaters of the tributaries of the Tennessee, and in eastern Kentucky and West Virginia, which have occupations of the Middle Woodland peri- ods. Whether this proposed interarea traffic took place along streams or by trails is not definitely known. Proba- bly both were used. The temporal span of both Ohio Hopewell and Middle Woodland sites in the eastern Tennessee area is not too well known, and it is difficult or impossible to be precise in terms of calendar or Carbon- 14 years about when these contacts took place. Given the probable time span of each, of some 400 years or more, the known amount of reasonably identifiable trade goods is not very great, nor can we yet be very specific about which Ohio sites were prime movers of these goods. Such may come with a re- finement of identification techniques of the clays and tempering material. Appendix 4. 1 Description and Sherd Count Harness Mound Pottery — Season 1976 Field Cat. No. 1 Surface Small McGraw Cordmarked specimen with cords closely spaced and smoothed. Thickness 4.5 mm. I 2B, 2D, 2E, 2J, N535 E492.5 Disturbed area 2B Small vertical rim sherd of McGraw Plain with nar- rowed flattened lip 3 mm wide. Slight protrusion on outer upper rim folded or smoothed down 3 mm on that surface. Consequently that area has a thickness of 5 mm as does the lower rim (Fig. 4.2g). 1 Small McGraw Cordmarked body sherds one of which is 6 mm and the other 4 mm thick. 2 2D McGraw Cordmarked body sherds composed of 3 fragments which fit together and vary in thickness from 5 to 12 mm (Fig. 4.2b). I 2E McGraw Plain, probably a rim section, 5 mm thick. 1 An upper shoulder area section of Chillicothe Plain Rocker-stamp which is placed below a shallow hor- izontal line 3 mm wide which delimits the smoothed horizontal rim band from the decorated area on the body. The Rocker-stamp impressions are convex to the right. They are arranged in 2 visible horizontal rows each 1 .9 cm high. Both the inner and outer sur- 50 JAMES B. GRIFFIN No. 39 face have been smoothed. On the outer surface this took place after the design application. Both sur- faces are black, possibly from having been fired in a reducing atmosphere. Lower rim 7 mm and body from 7 to 5.5 mm thick (Fig. 4.1c). 1 2J McGraw Cordmarked body sherd with some large tempering particles 3.5 mm thick extending from the inner to outer surfaces. Thickness 4 to 5 mm. 1 1 1C, N535 E495, Feature 4 Small sherdlets of McGraw Cordmarked. Perhaps from the same vessel. Thickness 4.5 mm. 4 20A, N537.5 E492.5, Feature 13, (in situ) I.S. 21 fragments of the side wall of a McGraw Cord- marked vessel which was in 28 fragments when re- ceived. Thickness 3 to 5 mm (Fig. 4.4b). 1 McGraw Cordmarked sherd with same field number but perhaps a different vessel or near base of vessel listed above. Thickness 7 mm. 1 24B, 24C, 24D, 24E, 24G, 24H, N535 E492.5, Feature 89, I S. 24B Outer wall fragment McGraw Cordmarked. 1 Chillicothe Plain Rocker Stamped carelessly exe- cuted on body fragment with tan outer surface. Probably not from same vessel as 2E. Thickness 5.5 mm (Fig. 4. 11). I 24C McGraw Cordmarked. Thickness 3.5 mm. 1 Chillicothe Plain Rocker Stamped with same execu- tion and size of rocker stamping as 2E but only 4.5 mm thick. Same dark brown to black surfaces but probably not from same vessel. Fine grit temper (Fig. 4.1e). I 24D McGraw Cordmarked 5 mm thick. 1 24E McGraw Cordmarked with coarse grit temper of whitish crushed temper which is not limestone. Very friable and 5.5 mm thick. I Well-made Hopewell rim. The fine incised cross- hatched upper band is 1.6 cm high. The horizontal row of hemiconical punctates below the cross- hatching are 5 to 9 mm long and 4 mm high. The lower rim area is well smoothed. Inner and outer sur- faces are black. There is a slight camber to the rim caused by a shallow channel on the upper inner rim. The lip is rounded and 4 to 5 mm wide while both the upper and lower rim are 7 mm thick. This rim could be from the same vessel as 2E because color, paste, and thickness are similar (Fig. 4. la). 1 24G Chillicothe Plain Rocker Stamped with light tan outer surface. Swing of rocker is 2.3 cm high. Thick- ness 5 mm (Fig. 4. 1 k). 1 24H Chillicothe Plain Rocker Stamped with darker tan outer surface. Swing of rocker is 2.3 cm high. Thick- ness 4 mm. Does not appear to be from same vessel as 24G and neither are from same vessel as 2E (Fig. 4- lj). 1 3 IB, N535 E495 Disturbed McGraw Cordmarked sherdlets 3 mm thick. 1 35B, N535 E492.5, Disturbed Chillicothe Plain Rocker Stamped. Too small to measure the rocker swing. The lines are markedly narrower than those of 2E and it is 5 mm thick (Fig. 4. lh). I 37A, N535 E492.5, PM 16, I.S. McGraw Cordmarked upper body and lower part of rim which may have been smoothed. It is 6 mm thick and has relatively fine grit temper (Fig. 4.2c). 1 40 A, N535 E492.5, PM 14, I.S. 40A Hopewell rim section with same features as 24G and is almost certainly from same vessel. The hemiconi- cal punctates are more closely spaced but otherwise appearance of all visible features is almost identical (Fig. 4. lb). I 44C, N535 E492.5, Feature 19, (pit) I.S. McGraw Cordmarked 4 mm thick. 1 Three sherdlets probably McGraw Cordmarked, one of which is 5 mm thick. These probably from the same vessel. 1 McGraw Cordmarked 5 mm thick. 1 66A, N548.5 E507.5, Feature 23 Two McGraw Cordmarked sherdlets 4.5 mm thick. I McGraw Cordmarked sherds 4.5 mm thick. I 103B, 103C, 103D, N532.5 E495 Redeposited-disturbed 103B Three McGraw Cordmarked sherds 4.5 mm thick probably from same vessel. I McGraw Cordmarked lower rim and upper body 5 mm thick. 1 103C McGraw Cordmarked sherd. 1 103D McGraw Plain sherd 7 mm thick. 1 McGraw Plain sherd 4 mm thick. 1 103E Two Chillicothe Plain Rocker Stamped sherds with 1.8 cm high vertical swing, of very similar color and finish as 24E but are 4.5 to 5 mm thick. They could be from same vessel as 24G but from a different sec- tion of the body (Fig. 4. lf,g). 1 One fine paste Hopewell sherd with horizontal (?) incised line 2 mm wide. Sherd is 4 mm thick (Fig. 4. li). 1 104 Bulldozed area N520 E500 Small McGraw Cordmarked sherds 5 mm thick. Could be from same vessel but impossible to say for certain. 4 107 Bulldozed area McGraw Cordmarked with well-smoothed exterior 5.5 mm thick. I 123 Surface McGraw Cordmarked 7 mm thick. 1 1983 EDWIN HARNESS MOUND 51 McGraw Plain rim with flattened lip 6.5 mm wide and slight outer slope. Whitish grit temper. Rim is 7 mm thick (Fig. 4.2h). I 138A, N522.5 E495 Redeposited-disturbed Small lower rim and upper body sherd 7 mm thick. Is from same vessel as 23 1 B and is glued to that rim (Fig. 4.20. 1 138C, N522.5 E495 Redeposited-disturbed Probably Turner Simple Stamped A, 7 mm thick (Fig. 4.3e). 1 155, N557.5 E495 in mound floor McGraw Cordmarked 5 mm thick. 1 183A, 183B, 183C, 183F,and 183G, N525 E495, Feature 30(pit) Some 24 sherdlets of 183B, six of 183A, seven of 183C, six of 183E, seventeen of 183F, and three of 183G are McGraw Cordmarked. While it is not cer- tain that all of these are from one vessel, the 183A, a 183B, and a 183C fit together, as do a 183C and 183E sherd. If there are at least two vessels, the larger sherds belong to a vessel with 8 mm side walls and another vessel is represented by side walls 4 mm thick. 2 One baked clay fragment of 1 83C does not appear to be from a vessel, or figurine, and is unidentified. 191 A, N532.5 E495, PM 84, I.S. Inner section of McGraw Cordmarked or Plain, probably the former. I 217, Provenience lost This small sherd is either McGraw Plain or Cord- marked but it cannot be identified with any cer- tainty. 1 Harness Mound Pottery — Season 1977 219A, Surface Turner Simple Stamped in Prufer’s( 1968) terminol- ogy. This thin, hard, fine grit tempered sherd is cer- tainly not the same vessel as 308A. Neither have sand or micaceous sand temper. The inclusion of either in a “Southeastern Series” would be a mistake. Thick- ness 4.5 mm (Fig. 4.31 and m). 1 219E, Surface Backhoe Trench 1 backdirt pile Fairly large smoothed over McGraw Cordmarked 6 to 7 mm thick. 1 219U, Surface Two small McGraw Cordmarked sherds glued to- gether. They are 4 mm thick. 1 219V, Surface A McGraw Cordmarked sherd 5.5 mm thick. 1 A McGraw Plain small rim with a narrowed and rounded lip 4 mm wide while the rim is 7.5 mm thick (Fig. 4.2i). 1 22 1G, N530 E507.5, Bulldozed area One sherd 8.5 mm thick called here McGraw Plain. 1 225B, N540 E507.5, Redeposited Small Hopewell rim with reddish tan interior and exterior color. The lip, narrowed and rounded, slopes inward and is 3 mm wide. The incised cross hatching is widely spaced and is 7.5 mm high. The hemiconical punctates are 3 mm long and 3 mm wide. They are spaced 3.5 to 4 mm apart. The upper and lower rim thickness is 4.5 to 5 mm (Fig. 4. Id). I 229C, N550 E507.5, Redeposited Lower rim and upper body sherd of McGraw Plain. It is 8 mm thick and grit tempered (Fig. 4.2j). 1 230N, N522.5 E497.5 Two small McGraw Cordmarked sherds glued to- gether. They are 4 mm thick. 1 231 A, 23 IB, 23 IF, 237C, 237E, N522.5 E490, Disturbed area 231 A Two small Turner Simple Stamped A sherds with limestone temper. They are probably part of 308A. The stamp depressions are 2 to 3 mm wide and the lands are 1 to 2 mm wide (Fig. 4.3g, h). 1 23 1 B A rim sherd of unidentified type with a plain upper body and a horizontally brushed, wiped, or simple stamped outer rim. The lip is narrowed and rounded and is 4.5 mm wide. The rim and body are 8.5 to 9.5 mm thick. It is grit tempered. 1 23 IF Turner Simple Stamped A body sherd with lime- stone temper. Probably same vessel as 308A. Thick- ness 7.5 mm (Fig. 4.3i). 1 237C Rim sherd and upper body of Turner Simple Stamped A. Is very probably same vessel as 308A. The rim has a slight flare or outer slope and is 7 mm thick. The lip slopes to the interior and has shallow depressions caused by thumb (?) impressions while the clay was still soft. This vessel is a real stranger in the Harness ceramic assemblage because of the temper (Fig. 4.3b, j). 1 Turner Simple Stamped A body sherd from same vessel. 1 237E Turner Simple Stamped A body sherd from same vessel (Fig. 4.3k). 1 264E, N530 E485 McGraw Cordmarked 6.5 mm thick. 1 2710, N552 E500, PM 162, I.S. A large McGraw Plain body of 4 glued fragments 5 mm thick (Fig. 4.4a). I 290C, 290G, N525 E502.5, Feature 66 290C Has 8 fragments of a thin checked-stamped vessel. These sherds are thin, 3 mm, grit tempered and are not trade material from the Southeast as far as I can see. Both inner and outer surfaces are black and smoothed (Fig. 4. lm-g). 1 Has 1 lower rim section of McGraw Plain. It is well smoothed on outer and inner surfaces, which are black. The sherd is 9 mm thick (Fig. 4.2e). 1 52 JAMES B. GRIFFIN No. 39 290G Has 4 sherds of McGraw Cordmarked. They do not seem to be from the same vessel. 4 291 A, N530 E502.5 on top of PM 160 A small sherd of McGraw Plain 4 mm thick. 1 295G, N530 E502.5 on top of PM 160 Two sherds of probably same McGraw Cordmarked vessel. 1 Two other unidentified fragments. 2 295L, N530 E502.5, PM 160, I.S. Two sherds of which one has two glued pieces. The glued sherds are from a lower rim section of perhaps McGraw Plain and are 8.5 mm thick. The second is a small fragment of probably McGraw Cordmarked and is 5 mm thick. 2 297G, N535 E509, partially disturbed One McGraw Plain 5 mm thick body sherd. 1 One McGraw Cordmarked 4.5 mm thick. 1 300A, 300B, N525 E509, Redeposited Both of the two sherds are very small. They may be McGraw Plain or Cordmarked and are 3.5 mm thick. 2 308A, N524 E485, PM 270, I.S. Large body sherd of Turner Simple Stamped A glued together from 9 sherds, and one other sherd. This vessel has limestone temper of medium size crushed fragments. There are fragments of this ves- sel from other localities, and the vessel seems to be the only limestone tempered one in this collection. The paddle depressions are about 3 mm wide and the bands are 2 to 1 mm wide. Thickness is 6 to 7 mm (Fig. 4.3a, 0. 1 308C, N524 E485 PM 268 4 body sherds McGraw Cordmarked glued together, reddish to tan in color, may belong to one vessel, but uncertain. Thickness 6 to 9 mm. 1 309A, N524 E492.5 burned area on mound floor, partly dis- turbed. A lower rim (McGraw Plain) and upper body (McGraw Cordmarked) 7 mm thick. Well fired tan in color (Fig. 4. 2d). I Two body sherds of Turner Simple Stamped A from same limestone tempered vessel as 308A, 8 mm thick (Fig. 4.3c, d). 1 504B, N527.5 E495, PM 410, I.S. A McGraw Cordmarked sherd 4.5 mm thick. 1 700B, N522.5 E495, Redeposited McGraw Cordmarked sherd 5 mm thick. 1 References Adams, William R. 1949 Archaeological notes on Posey County, Indiana. In- diana Historical Bureau, Indianapolis. Butler, Brian M. 1979 Hopewellian contacts in southern Middle Tennessee. In Hopewell archaeology: the Chillicothe conference, edited by D. S. Brose and N. Greber, pp. 150-156. Kent State University Press, Kent, Ohio. Chapman, Jefferson 1973 The Ice House Bottom site 40Mr23. Report of Inves- tigations, No. 13, Department of Anthropology, Uni- versity of Tennessee, Knoxville. Chapman, J., and B. C. Keel 1979 Candy Creek-Connestee components in eastern Ten- nessee and western North Carolina and their relation- ship with Adena-Hopewell. In Hopewell archaeology : the Chillicothe conference, edited by D. S. Brose and N. Greber, pp. 157-161. Kent State University Press, Kent, Ohio. DePratter, C. 1979 Ceramics. In The anthropology of St. Catherines Is- land: 2. The Refuge-Deptford mortuary complex, edited by D. H. Thomas and C. S. Larsen. Anthropo- logical Papers of the American Museum of Natural History 56(1): 109- 132, New York. Faulkner, Charles H., and J. B. Graham 1965 Excavations in the Nickajack Reservoir: Season 1. Tennessee Archaeological Society, Miscellaneous Paper, No. 7. 1966 Highway salvage in the Nickajack Reservoir. Tennes- see Department of Highways Contract No. 0402. De- partment of Anthropology, University of Tennessee, Knoxville. Faulkner, Charles H. (editor) 1968 Archaeological investigations in the Tims Ford Res- ervoir, Tennessee, 1966. National Park Service Con- tract 14-10-0131-1631. Department of Anthropology, University of Tennessee, Knoxville. Fowler, Melvin L. 1957 Rutherford Mound, Hardin County, Illinois. Illinois State Museum Scientific Papers 7( 1 ): 2— 43, Spring- field. Gleeson, Paul F. (editor) 1970 Archaeological investigations in the Tellico Reser- voir, Interim Report, 1969. Report of Investigation, No. 9. Department of Anthropology, University of Tennessee, Knoxville. Griffin, James B. 1938 The ceramic remains from the Norris Basin, Tennes- see. In An archaeological survey of the Norris Basin in eastern Tennessee, edited by W. S. Webb. Bureau of American Ethnology, Bulletin 1 18:253-359, Wash- ington. 1 945 The ceramic affiliations of the Ohio Valley Adena cul- ture. In The Adena people, edited by W. S. Webb and C. E. Snow. University of Kentucky, Reports in An- thropology and Archaeology 6:220-246. 1946 Cultural change and continuity in eastern United States archaeology. In Man in northeastern North America, edited by F. Johnson. Papers of the Robert S. Peabody Foundation for Archaeology 3:37-95, Andover. Haag, William G. 1940 A description of the Wright site pottery. In The 1983 EDWIN HARNESS MOUND 53 Wright mounds, sites 6 and 7, Montgomery County, Kentucky, edited by W. S. Webb. University of Ken- tucky Reports in Anthropology and Archaeology 5( 1 ):75— 82, Lexington. 1942a A description and analysis of the Pickwick pottery. In An archaeological survey of Pickwick Basin in the ad- jacent portions of the states of Alabama, Mississippi, and Tennessee, edited by W. S. Webb and D. L. De- Jarnette. Bureau of American Ethnology, Bulletin 129:509-526. 1942b The pottery from the C and O mounds at Paintsville. In The C and O mounds at Paintsville, sites Jo 2 and Jo 9, Johnson County, Kentucky, edited by W. S. Webb. University of Kentucky, Reports in Anthro- pology and Archaeology 5(4):34 1—349, Lexington. Heimlich, Marion D. 1952 Guntersville Basin pottery. Geological Survey of Ala- bama, Museum Paper No. 32, Tuscaloosa. Jefferies, Richard W. 1976 The Tunacunnhee site: evidence of Hopewell interac- tion in northwest Georgia. Anthropological Papers of the University of Georgia , No. 1, Athens. Keel, Bennie C. 1976 Cherokee archaeology: a study of the Appalachian summit. University of Tennessee Press, Knoxville. Kellar, James H. 1973 An introduction to the prehistory of Indiana. Indiana Historical Society, Indianapolis. 1979 The Mann Site and “Hopewell” in the lower Wabash- Ohio valley. In Hopewell archaeology: the Chillicothe conference , edited by D. S. Brose and N. Greber, pp. 100-107. Kent State University Press, Kent, Ohio. McCollough, Major C. R., and Charles H. Faulkner (editors) 1973 Excavation of the Higgs and Doughty sites: 1-75 sal- vage archaeology. Tennessee Archaeological Society, Miscellaneous Paper. No. 12. Mills, William C. 1907 Explorations of the Edwin Harness mound. Ohio Ar- chaeological and Historical Quarterly 16:1 13-193. Mortine, Wayne A., and Doug Randles 1978 The Martin mound: an extension of the Hopewell in- teraction sphere into the Walhonding Valley of east- ern Ohio. Occasional Papers in Muskingum Valley Archaeology 10. Prufer, Olaf H. 1967 The Scioto Valley archaeological survey. In Studies in Ohio archaeology , edited by O. H. Prufer and D. H. McKenzie, pp. 267-328. The Press of Western Re- serve University, Cleveland; rev. ed., Kent State Uni- versity Press, 1975. 1 968 Ohio Hopewell ceramics: an analysis of the extant col- lections. Anthropological Papers No. 33, Museum of Anthropology, University of Michigan, Ann Arbor. Prufer, Olaf H., D. H. McKenzie, O. Pi-Sunyer, H. C. Cutler, R. A. Yarnell, P. W. Parmalee, and D. H. Stansbery 1965 The McGraw site: a study in Hopewellian dynamics. Cleveland Museum of Natural History, Scientific Publications (n.s.), 4(1). Webb, W. S. 1938 An archaeological survey of the Norris Basin in east- ern Tennessee. Bureau of American Ethnology, Bul- letin 1 18, Washington. Williams, Stephen (editor) 1968 The Waring papers: the collected works of Antonio J. Waring, Jr. Harvard University, Peabody Museum of American Archaeology and Ethnology, Papers 58, Cambridge. Willoughby, Charles C., and Earnest A. Hooton 1922 The Turner group of earthworks, Hamilton County, Ohio. Harvard University, Peabody Museum of American Archaeology and Ethnology, Papers 8(3), Cambridge. Winters, Howard D. 1963 An archaeological survey of the Wabash Valley in Il- linois. Illinois State Museum, Report of Investiga- tions 10, Springfield. 5 PLANT REMAINS TRISTINE LEE SMART and RICHARD 1. FORD The Edwin Harness Mound, located in the Scioto River valley near Chillicothe, Ohio, has long been recog- nized as a major Hopewell burial mound. While several excavations were conducted at the site in the past, many questions regarding the mound stratigraphy and sub- mound features, structures, and activity areas remained unanswered. Therefore, when unexcavated portions of this mound were recently threatened by cultivation, sal- vage excavations were conducted under the direction of Dr. N’omi Greber of the Cleveland Museum of Natural History. During the excavations, plant remains were col- lected and submitted to the Ethnobotanical Laboratory of the University of Michigan Museum of Anthropology for identification and interpretation. These samples were returned to the Cleveland Museum of Natural History af- ter analysis for curation and storage. Nature of the Samples Plant remains recovered from the Edwin Harness Mound included primarily charcoal samples and flota- tion samples from post molds and features. In addition, 2 pieces of fabric, 1 seed, and a modern corn cob ( Zea mays) were collected during excavation. Flotation was conducted in the field and at the Cleve- land Museum of Natural History using a SMAP-type flo- tation system (Watson 1976). The light fraction was col- lected in 4 mm, 2 mm, and .5 mm screens. Analytical Techniques The samples of plant material were examined micro- scopically under magnifications ranging from 10 to 30X. Carbonized seeds were separated from the samples and identifications were attempted with the aid of seed manu- als (Martin and Barkley 1961; Delorit 1970) and the com- parative collections of the Ethnobotanical Laboratory. Charcoal was identified by examining a transverse section microscopically with the assistance of wood manuals (Brown 1928; Panshinand deZeeuw 1970) and compara- tive wood samples. A subsample of 20 pieces of charcoal was selected for identification from samples containing large amounts of charcoal. Because the size and shape of a piece of charcoal affects its identifiability, these factors influence the selection of charcoal pieces for identification. Therefore, the charcoal selected cannot be considered an unbiased subsample. However, care was taken to select charcoal pieces of many shapes and sizes, so the charcoal subsamples should be fairly representative of the charcoal from each sample as a whole. Charcoal from Posts The submound deposits from the Edwin Harness Mound included several adjacent structures outlined by post molds. Charcoal was recovered from a number of these post molds (Table 5. 1 ). Hickory ( Carya) was the most common timber used for construction. Out of a total of 33 post samples examined, 25 (76%) contained hickory. All of the identified hickory pieces had annual rings which showed an even, concentric growth pattern. This suggests that these trees were grow- ing in a location that was not stressful at the time the rings were formed. All were saplings or young trees rather than limbs or split sections from larger, more mature trees. In 4 of the post samples, a few pieces of non-hickory charcoal types were found in addition to hickory. These included oak ( Quercus', Post 1 42), the more specific white oak group (Posts 151 and 155), and maple (Acer, Post 146). These may have been from wooden wedges used to tighten the post in place, or alternatively contamination from the general fill or the charred remains of trees which once grew on the site. A woven bast fiber textile was found at the base of Post 1 3 1 and may have served a sim- ilar purpose. Eight post samples did not contain hickory charcoal. The white oak group was present in 6 of these samples ( Posts 85, 122, 132, 1 98, 2 1 3), 1 post was elm ( Ulmus ; Post 138), and another was an unidentifiable diffuse porous wood type (Post 128). The posts from the circular structure at the south end of the mound are interesting. The 2 samples from Post 216 actually contained charcoal of 5 tree types: hickory, chestnut ( Castanea ), honey locust ( Gleditsia ), walnut ( Juglans ), and pine ( Pinas). This suggests that this feature was not a post. Two of the remaining 3 posts that were identified were white oak group, and it might be signifi- cant that they were located opposite each other on the east and west sides of the structure. Charcoal from Features Plant remains were recovered from a number of differ- ent types of features at the Edwin Harness Mound. These included mound loading (Feature 43); a large ring of cob- bles which encircled the other submound deposits (Fea- ture 1 ); a burned area at the center of the middle structure TABLE 5.1 Charred Posts from the Edwin Harness Mound Floor (Percentages based on a total of 20 identified pieces) C. ■3 -C> 3 3 u Possible fence around perimeter Circular structure, south end Middle rectangular structure Northern rectangular structure Eastern structure 116 F6 X 1 19 F4 X 125 F9 X 140 F40 X 151 F76 X 85 192A 142 230M X 198 CCD61 216 CCD64 50 F99 65 127 F31 X 128 F28 131** F27 X F32 X 157 F85 X 230 ¥91 X 122 F8 126 F19 X 132 F48 134 F39 X 136 F26 X 138 F42 141 F65 X 146 F66 X X 152 F75 X 154 F74 X 155 F77 X 162 F84 X 163 F82 X 213 F93 F95 121 F5 X 215 F98 X O "O o 20 30 10 10 c 5, 3 C '£ x x x x 10 X X X X X .§ 3 £ a § X* X x = present *charcoal minute and fragile **woven blast fiber fabric present TABLE 5.2 Charcoal from the Edwin Harness Mound Features (Percentages based on a total of 20 identified pieces) *3 §• §■ © s* 3 o O bo *© 3 bo 3 3 "3 3 3 “3 -3 £ ») 3 3 -Si §■ £ © •Si «3 -3 c. c« C 2 -3 -3 3 -3 3 3 C '* 3 o -3 .2 .3 *3 £ C J2 .3 s. 3 O 3 ■4* 3 O 3 o s *3 © C. 3 £ O ao C3 a> Ob C 3 3 >3 $ < u u Urn U- o —> a. o a o 1 F10A X F23 X 17 F70AA X X X X 33 F25 67 33 F36 100 F41 X F43 100 F44 100 F45 X F57 40 60 F70 50 10 40 CCD19A 5 5 15 75 36 FI A 25 75 43 F29 100 X 45 F53 100 F60 10 30 60 X F64A 30 10 60 F64B 80 20 X F64C 30 5 5 60 F64D 30 30 40 F69 5 50 15 30 X F69B 67 33 F69C 60 40 X 46 F47 20 40 40 F62 45 5 15 35 49 F46* 90 10 F49* 30 70 50 F58 100 52 F56* 100 53 F61 100 F63 70 30 54 F79 X 60 F7 1 A 75 5 5 5 5 5 62 F83 30 10 60 65 F35 100 ** F20 - X x = present ♦directly over feature in loading **N560 E507.5, mound loading 1983 EDWIN HARNESS MOUND 57 on the mound floor (Feature 36); a depression in the main mound floor (Feature 54); strata under the main mound floor (Features 33, 50, and 65); a bundle burial (Feature 60); a large pit located outside of the submound structures (Feature 1 7); a small oval-shaped basin found just north of the submound structures (Feature 62); and various strata located outside of the submound structures. The latter included burned material on clay floors (Features 45, 46, and 53); burned soil (Feature 49); and mica pieces in a dark stained area (Feature 52). Charcoal identifications from botanical samples col- lected from features are presented in Table 5.2. Ten tree types were represented, including such types as beech ( Fagus ) and maple, which are found today in riverbottom and lower slope communities, as well as such trees as chestnut and pine, which are primarily found on the upper slopes (Gordon 1 969). Additional tree types identi- fied included hickory, ash ( Fraxinus ), honey locust, wal- nut, red oak group, and white oak group. As with the post samples, hickory and white oak group were the most common charcoal types in the feature sam- ples. Of the 35 flotation samples and 1 charcoal sample from features, 26 (72%) contained hickory and 20 (56%) contained white oak group charcoal. This suggests that some of the charcoal incorporated in the feature deposits could have come from the remnants of burned hickory and white oak group construction timbers. However, the charcoal in features could also be the remnants of wood used as fuel. Hickory and white oak group wood both burn with a hot, clean flame and may have been selected for use as fuel because of their burning properties. Interestingly, the burned area in the middle structure on the main mound floor (Feature 36) contained 75% pine charcoal. Pine has very different heating properties than oak and hickory, producing a hot fire more quickly than the hardwoods. The bundle burial deposit (Feature 60) contained the greatest diversity of charcoal types includ- ing hickory, chestnut, beech, walnut, white oak group, and red oak group. Seeds from the Edwin Harness Mound Carbonized seeds were recovered from 10 of the feature flotation samples, from 2 of the post samples (1 seed was recovered during excavation), and from a charcoal sam- ple used for C-14 dating (Table 5.3). Corn kernels were the only type of domesticated plant material recovered from the site. Unambiguous kernel fragments were recovered from Feature 45 and Feature 60, the bundle burial. The identification of grains from Feature 65 is more questionable because of their fragmen- tary state. One uncarbonized, 12-row cob fragment of a modern variety was found in Backhoe Trench 4 during excavation. The presence of corn with the bundle burial suggests the use of corn as a mortuary offering and hints at the ritual significance of corn in Hopewell culture. Carbonized seeds from wild plants recovered at the site included goosefoot (Chenop odium), knotweed ( Polygo- num), spurge ( Euphorbia ), and grass (Gramineae). The seeds of the goosefoot and knotweed are difficult to inter- pret. These ruderal plants grow prolifically on disturbed or abandoned sites, and their seeds could be accidental inclusions resulting from unintentional dispersal of seeds from plants growing nearby. Alternatively, these seeds could have been carbonized through the use of dried weeds as kindling. However, both goosefoot and knot- weed seeds are edible. Large quantities of these seeds were recovered from Middle Woodland deposits at the Scovill site in Illinois, indicating that they were used as food in that area (Munson et al. 1971). In most of the samples from the Edwin Harness Mound the low counts of goose- foot and knotweed seeds suggest that they may well have been accidental inclusions in the deposits. However, the moderate number of goosefoot and knotweed seeds along with corn in the bundle burial deposit hint at the eco- nomic use of these seeds. Spurge seeds were also recov- ered from the Feature 60 burial as well as from Feature 17, although these seeds are not known to be edible. The presence of carbonized grass seeds and culms could also have been due to accidental inclusion in the archaeologi- cal record. Alternatively, these plant remains could have come from thatching on the submound structures. Hundreds of an unknown, egg-shaped seed type were recovered from Features 17 and 60, and much smaller numbers of these seeds were found in Features 33 and 65. A sample of 25 of these seeds from Feature 60 had an average size of 1 .7 mm X 1 .2 mm. In addition, other uni- dentified seeds were found in many of the samples. No carbonized nutshell fragments were recovered from the site, even though hickory, oak, and walnut charcoal was present in the post and feature samples. This is quite unlike the reported plant remains from several other Ohio Hopewell sites (Ford 1979:235) and probably reflects the special purpose of the Edwin Harness Mound. The only identified edible tree products from this site were 2 plum pits ( Prunus americana). Uncarbonized, modern weed seeds were also recovered in the archaeological deposits from the mound. These in- cluded not only modern goosefoot, knotweed, and grass seeds, but also the grass genus Digitaria , purslane ( Por- tulaca), campion ( Silene ), and blueberry ( Vaccinium ), which did not occur archaeologically. Other items recovered from the botanical samples were bone and snail shells (Table 5.3). A piece of bone from Feature 60 was identified as possibly human. The identifi- able snail shells from the samples were Hawaiia minus- cula (Amy Shraden Van Devender, Museum of Zoology, Univ. of Michigan, personal communication, 1978). This species prefers a moist, floodplain habitat, but it could be found elsewhere as well (LaRocque 1970:639). 58 TRISTINE LEE SMART and RICHARD I. FORD No. 39 TABLE 5.3 Carbonized Seeds and Faunal Remains from the Edwin Harness Mound Botanical Samples (counts include half to complete seeds) Sample Provenience Number Chenopodium goosefoot Euphorbia spurge Gramineae grass Carbonized Seeds Prunus Polygonum americana knotweed plum Zea mays ( kernel ) corn Unknown (egg shaped) Fauna Unident. Bone Snail Post 141 F65 1 Post 142 230M 1 frag Fea. 17 F70AA 4+ 3 1 8 155+ 34+ x Fea. 33 CCDI9A 5+ F25 1 F36 l+(cf) 1 F70 22+ 7(cf) 3 56 xx Fea. 45 F64C I 3 1 frag F69C 3 culms X Fea. 46 F62 1 X Fea. 52 F56 1 Fea. 60 F71A 4+ 8 16+ 1 frag+* 131 44+ x x Fea. 65 F35 1 frag +(cf) 10 8+ + = additional seed fragment(s) not included in count x = present * one Zea mays kernel fragment plus many small cf Zea mays kernel fragments Conclusion The plant remains from the Edwin Harness Mound were very specialized as one might expect in a mortuary site. Most of the construction wood was hickory. The charcoal from the feature deposits was primarily hickory and white oak group, although other species were present in low frequencies. One sample from a burned area at the center of the middle structure on the mound floor con- tained mostly pine charcoal. The charcoal types recov- ered suggest that the aboreal environment was similar to the contact forest; however, no nuts from this forest were found at the site. Fragments of corn kernels and possibly goosefoot and knotweed seeds found in association with a bundle burial may represent mortuary offerings. Corn kernel fragments were also present in one or possibly two other deposits. Carbonized seeds from wild plants recov- ered from other features were possibly refuse from a meal or were simply the result of accidental dispersal into a hearth or the fire that consumed the structures. References Brown, H. P. 1928 Atlas of the commercial woods of the United States. The Bulletin of the New York State College of For- estry at Syracuse University 1(4). Delorit, R. J. 1970 Illustrated taxonomy manual of weed seeds. Agron- omy Publications, River Falls, Wis. Ford, Richard I. 1979 Gathering and gardening: trends and consequences of Hopewell subsistence strategies. In Hopewell archae- ology: the Chillicothe conference , edited by D. S. Brose and N. Greber, pp. 234-238. Kent State Univer- sity Press, Kent, Ohio. Gordon, Robert B. 1969 The natural vegetation of Ohio in pioneer days. Bul- letin of the Ohio Biological Survey 3(2). LaRocque, Aurble 1970 Pleistocene mollusca of Ohio. State of Ohio Depart- ment of Natural Resources, Division of the Geologi- cal Survey Bulletin 62, Part 4. Martin, Alexander C., and William D. Barkley 1961 Seed identification manual. University of California Press, Berkeley, Calif. Munson, Patrick J., Paul W. Parmalee, and Richard A. Yamell 1971 Subsistence ecology of Scovill, a terminal Middle Woodland village. American Antiquity 36(4):4 1 0—43 1 . Panshin, A. J., and Carl de Zeeuw 1970 Textbook of wood technology (V ol. 1). McGraw-Hill Book Company, New York. Watson, Patty Jo 1976 In pursuit of prehistoric subsistence: a comparative account of some contemporary flotation techniques. Midcontinental Journal of Archaeology 1(1 ):77— 1 00. 6 VERTEBRATE FAUNAL REMAINS ORRIN C. SHANE III During the 1976 and 1977 field seasons nearly 10,000 pieces of bone were recovered from the Edwin Harness burial mound (33Ro22), Ross County, Ohio. Excava- tions were directed by Dr. N’omi Greber of the Cleveland Museum of Natural History as part of a project to salvage information from the site prior to its final destruction by agricultural activity. The Harness Mound is the major earthen mortuary structure associated with the Harness (Liberty) Earth- work Complex, and represents a locus of Hopewellian mortuary-ceremonial activity dating from the first few centuries of the Christian era. The mound is located on the east side of the Scioto River Valley, in an area charac- terized in prehistoric times by a mixture of deciduous forest and open grassland vegetational communities. The site lies with the Carolinian Biotic Province (Dice 1943), at the western margin of the unglaciated Allegheny Plateau. Methods A variety of recovery techniques were employed in the mound excavation, including dry screening through 'A inch hardware cloth, water separation, and flotation. Consequently, the rate of recovery of bone and other cul- tural material was high, yielding large numbers of small, often unidentifiable bone fragments. Therefore, each bone piece was examined and initially placed into one of three categories, either as bone identifiable below the class level, bone identifiable to class only, or bone uniden- tifiable to class. The criterion for identifiability below the class level was the presence of the major portion of an articulatory surface; bone was identified to class on the basis of gross structure. For that portion of the assemblage with bones bearing an articulatory surface, identifications were made after comparison with skeletal reference collections housed in the James Ford Bell Museum of Natural History, Univer- sity of Minnesota, and the Science Museum of Minne- sota. Scientific and common names of mammals follow Hall and Kelson ( 1959), while the names of birds are from the Checklist of North American Birds of the American Ornithologists’ Union ( 1957). The scientific and common names of fishes follow Trautman ( 1957). The minimum number of individuals (MNI) repre- sented by the identifiable bones was determined by simple osteological count of right and left elements. Because of the small size of the sample and the absence of bones in- dicative of the age of individuals, this method of deter- mining MNI is probably satisfactory. Several excavation units and features yielded mammal canine teeth drilled for suspension. While many of these specimens were burned and fragmentary, some teeth were sufficiently complete for identification. A minimum num- ber of drilled canines was obtained from a count of distal, medial, and proximal fragments. Perforation for suspen- sion was near the base of the tooth root, and fracture gen- erally occurred at the point of drilling. Therefore, distal fragments were defined as including the tooth portion from the tip of the crown to the point of fracture at the perforation. Medial fragments lacked the tip of the tooth, while proximal fragments were defined as the tooth por- tion from the base of the root to the point of fracture at the perforation. The minimum number of drilled canine teeth was determined from counts of distal, medial and proximal fragments plus whole canines. Results A total of 9,762 pieces of bone were recovered. Of these, 3,036 pieces, or 31. 10% of the total, were very small frag- ments unidentifiable to class. Of the remaining 6,726 bones, 1 76 pieces were identifiable to the family, genus, or species levels. Table 6. 1 shows the frequency of identified and unidentified bones by class; the frequency of identi- fied vertebrate taxa is presented in Table 6.2. Sixteen complete perforated mammal canine teeth and 225 proximal, medial, and distal fragments represent at least 171 specimens of drilled canines. All but 5 of these objects were burned, and some were completely calcined. Seventeen specimens could be identified to the genus or species level as follows: Raccoon 10 specimens Canis sp. 4 specimens Grey Fox 2 specimens Bobcat I specimen Feature 44, described in the field as a deposit of burned materials and ash, yielded 115 fragments of drilled ca- TABLE 6.1 Frequency of Identified and Unidentified Bones by Class Class Identified % Unidentified % Total % Mammal 133 1.98 6,211 92.34 6,344 94.32 Bird 24 0.36 260 3.87 284 4.21 Reptile 2 0.03 51 0.76 53 0.79 Fish 17 0.25 28 0.42 45 0.67 Totals 176 2.62 6,550 97.38 6,726 100.00 60 ORRIN C. SHANE III No. 39 TABLE 6.2 Frequency of Identified Vertebrate Remains from Edwin Harness Mound Scientific Name Common Name No. of Bones % MNI % Odocoileus virginianus Deer 35 19.89 2 6.25 Procyon lotor Raccoon 11 6.25 6 18.75 Canis sp. Canid 5 2.84 2 6.25 Urocyon cinereoargenteus Grey Fox 1 0.57 1 3.13 Lynx rufus Bobcat 2 1.14 1 3.13 Sylvilagus floridanus Cottontail 1 0.57 1 3.13 * Tamias s that us Chipmunk 28 15.91 1 3.13 Peromyscus maniculatus Deer Mouse 2 1.14 1 3.13 Rodentia Small Rodent 1 0.57 1 3.13 Homo Human 47 26.70 3 9.37 Total Mammal 133 75.57 19 59.37 Meleagris gallopavo Turkey 15 8.52 2 6.25 * Gallus gallus Chicken 5 2.84 2 6.25 Colinus virginianus Bob-white Quail 1 0.57 1 3.13 Buteo sp. Hawk 1 0.57 1 3.13 Anas sp. Duck 1 0.57 1 3.13 Passeriformes Songbird 1 0.57 1 3.13 Total Bird 24 13.64 8 25.00 Colubridae Snake 2 1.14 1 3.13 Total Reptile 2 1.14 1 3.13 Ictalurus sp. Catfish 7 3.98 1 3.13 Micropterus sp. Bass 6 3.41 1 3.13 Catostomidae Suckers 4 2.27 2 6.25 Total Fish 17 9.66 4 12.50 Grand Totals 176 100.00 32 100.00 ♦Intrusive nines representing at least 99 specimens. Also associated with these canines was a fragment of a cut and polished bobcat (Lynx rufus) mandible. This specimen is the ante- rior portion of the right mandible including the alveoli of the canine and the adjacent two premolars. The mandible is cut parallel to the plane of the dentition at the base of the tooth roots, and grinding is evident along the cut sur- faces. A hole was drilled through the mandible, below the first premolar and posterior to the canine. This bobcat mandible fragment and the associated perforated canines may represent a necklace or other similar ornament. Discussion Chipmunk and domestic chicken are both clearly intru- sive to the site. The chipmunk is represented by 28 bones found articulated in situ; apparently the animal died after burrowing into the mound. The occurrence of domestic chicken is interesting, for it is likely that these bones rep- resent meals of such earlier excavation crews as those led by Putnam, Moorehead, and Mills. All of the chicken bones were recovered from previously disturbed portions of the mound, in and under backdirt from prior excava- tions and well below the modern surface. Apart from the intrusive modern domestic chicken, the faunal assemblage is in no way unusual for a Woodland context in the Carolinian Biotic Province. No exotic spe- cies are present, and those animals which are represented are to be expected. White-tailed deer, raccoon, turkey, and fish account for approximately 45% of identifiable bone and almost 50% of the individuals present. These species were also among those most numerous in the envi- ronmental zones around the site. What may be most significant about this assemblage is its very small size. If the sample is truly representative of animal utilization on the Harness Mound floor, then it would appear that animal foods d id not play a particular- ly important role in Hopewellian mortuary-ceremonialism 1983 EDWIN HARNESS MOUND 61 as practiced at Harness. While animal parts may have been used as ornaments, certainly there is no evidence for mass offerings of animals, or for large-scale processing of foods for use on the mound floor. The small size of the assemblage would seem to pre- clude any large habitation at the mound as mortuary ac- tivities were carried out. Furthermore, the paucity of fau- nal remains argues against the mortuary floor being the site of accumulation and exchange of food resources. If anything, animals appear to have been used on the mound floor in much the same manner as at habitation sites, but in far smaller quantity. References American Ornithologists’ Union 1957 Checklist of North American birds. 5th ed. Dice, L. R. 1943 The biotic provinces of North America. University of Michigan Press, Ann Arbor. Hall, E. R., and K. R. Kelson 1959 The mammals of North America. Ronald Press, New York. Trautman, Milton B. 1957 The fishes of Ohio , with illustrated keys. Ohio State University Press, Columbus. 7 ANALYSIS OF HUMAN SKELETAL MATERIAL Ohio Historical Society Collections RAYMOND S. BABY and SUZANNE M. LANGLOIS The fragmentary skeletal materials from the early ex- cavations in the Edwin Harness Mound by Warren K. Moorehead (1897) and William C. Mills (1907), presently in the Ohio Historical Society collections, represent only a small part of the total number of burials recorded in the field. Thus general statements concerning the total popu- lation from which the individuals came cannot be made. There are cremated bone fragments in the collections from the Harness “Rectangular Grave Exhibit.” At least three individuals are represented: an adult male, an adult female, and one immature individual. These bones prob- ably represent a conglomeration of several burials exca- vated and combined for exhibit. The duplication of parts includes condylar fosses, mandible with tooth sockets, right mandibular fragments, and vertebrae. The burning of the remains follows the standard pattern for Hopewell (Baby 1954). The analysis of these bones is included in Table 7.1. The cremated burial excavated in 1977 (Feature 56) contained two thin skull fragments, one anterior inferior parietal, one mastoid process, one mandible, and several long bones. The coronal suture exhibits the beginning of closure. These bones are from an adult female, around 30 years old. Again, the burning of the remains follows the standard pattern for Hopewell. References Baby, Raymond S. 1954 Hopewell cremation practices. Papers in Archaeology No. 1, Ohio Historical Society. Mills, William C. 1907 Explorations of the Edwin Harness mound. Ohio Ar- chaeological and Historical Quarterly 16:1 13-193. Moorehead, Warren K. 1897 Report of field work carried out in the Muskingum, Scioto, and Ohio Valley during the season of 1896. Ohio Archaeological and Historical Quarterly 5: 165-274. Cleveland Museum of Natural History Collections STEPHANIE J. BELOVICH All of the skeletal material included in this analysis was excavated from the Edwin Harness Mound (33Ro22) in 1977 by field crews under the direction of N'omi Greber. The skeletal material analyzed was sparse and fragmen- tary, few bones being complete. As a result, analytic procedures were limited to 1) inventory, 2) age and sex determinations, where possible, and 3) gross macroscopic examination for osteopathology. Anthropometric measurements were not possible due to the incompleteness of the bones. Skeletal age determi- nations were based upon the following criteria: 1) dental eruption (Brothwell 1972) and 2) epiphyseal fusion (Schour and Massler, cited in Brothwell 1972; Bass 1971; and Krogman 1962). Determinations of sex were based upon pelvic examination for major sexing criteria (Bass 1971). The material was also examined for the presence of four categories of pathological skeletal lesions: 1 ) develop- mental, 2) degenerative, 3) infectious, and 4) traumatic. The results of these analyses are summarized in Table 7.2. The three burials recovered during the 1 977 field season can hardly be considered a population, and only insuffi- cient data can be obtained for the other Harness burials; thus, it is impossible to address questions of population dynamics. Nonetheless, some statements can be made. Feature 60 was a charred and partially burned bundle burial. An age determination of 19-22 years was based upon the following observations: 1) absence of epiphyseal fusion of the iliac crest 2) partial epiphyseal fusion of the femur 3) dental development 4) dense, uniform trabecular bone The skeleton was determined to be that of a female be- cause of the presence of a wide sciatic notch and a deep pre- auricular sulcus. No pathological lesions were observed. Feature 75 was a primary inhumation recovered from a disturbed context, which resulted from previous excava- tions. An age determination, other than to classify this individual as an adult, was not possible. The porous na- ture of the trabecular bone and the presence of osteophy- tosis, however, suggest that this adult was past middle age. The major sexing characters of the innominate were missing. A small portion of what appeared to be the be- ginning of a deep pre-auricular sulcus was present. This suggests that the individual was female. Osteophytosis TABLE 7.1 Edwin Harness Mound Human Skeletal Material Excavated 1896-1905 O.H.S. No. Bone Sex Age Comments 7/56 Cranium with facial mask absent Male Approx. 45 years Marked bifrontal flatten- ing, probably mesocranic 7/B4, 14150 Facial mask only Initial development of supraorbital ridges sug- gests male Immature, 4-5 years No evidence of bifrontal flat- tening 7/B2, 13849, & Complete right parietal, part of frontal Immature (child) Not part of7/B4, 14150 13850 7/BI, Intact left ilium Greater sciatic notch sug- Under 13years Elements of innominate un- 13814 gests female united, iliac crest ununited 7/B6, Left tibia Immature Epiphyses ununited 14171 7/B4, 14152 Portion of right frontal Adult probably male, from superior orbital margins — Not related to 7/B4, 14150 13910/ 1391 1 Nearly complete right parietal, portions of right occiput and frontal Male 40 years by suture closure 7/53 Intact mandible Male Approx. 35-40 years Not related to 7/ 56 but could be to 7/ B4, 1 4 1 52 or 1 3910/ 1391 1, slight erosion of right condyle resulting in more wear on right teeth 20070 Intact mandible Male Not related to 7/ 56, could be to 13910/ 11 or 14152, large ( 1 2 x 10 mm) aperture in right ascending ramus 4.5 mm below notch, due to bone tumor; healing begun on exterior surface, draining-type abscess; some trauma to right con- dyle, marked reduction of right ramu compared to left 7/ Parts of both skull and post- cranial skeleton (ca. 19%) represented Acetabulum suggests female Lapsed union on posterior sagittal suture places age above 45 years Bone ranging from com- pletely normal (unburned) to completely incinerated 7/ (exhibit) Male ca. 50 years Vault extremely thick, arthritic lipping of con- dylar fossa and upper lumbar & lower thoracic vertebrae quite extensive 7/ (exhibit) Female 45-47 years Slight arthritic lipping on cervical, none on condylar fossa, sagittal and part of lambdoid suture com- pletely closed 7/ (exhibit) 1 fragment of unburned cervical vertebra Immature 64 STEPHANIE J. BELOVICH No. 39 was the only pathological lesion observed. Degeneration involved lipping and destruction of the vertebral bodies. Finally, copper staining was observed on several of the bones. Table 7.2 details the extent of this observation. Feature 84 was located in a grave cut through the mound floor and into the natural gravels which underlay the mound. Dental eruption and the length of the humer- us were used to establish an age of 6-8 months. Sex, of course, remains undetermined. No pathological lesions were observed. No developmental, infectious, or traumatic lesions were observed for any of the individuals examined in this study. In summary, the three individuals recovered from the Edwin Harness Mound during the summer of 1977 were examined for sex and age determinations and osteopa- thology. The small sample size and its fragmentary nature limited the scope of the present study to inventory and description. TABLE 7.2 Edwin Harness Mound Human Skeletal Material, non-cremated, 1977 Feature # Specimen Comments 60 partially burned bundle burial female, !9-22yrs frontal; missing small portion of squamous and L & R orbits L temporal; missing squamous R temporal; missing squamous and mastoid L zygoma L & R parietal occipital; missing base, R condyle and portion of squamous L & R nasal L & R maxilla; missing frontal process and portion of body L mandible R mandible; missing coranoid process, con- dyle and portion of ascending ramus 148 skull fragments 3 incisors 1 canine 3 premolars 12 molars 2 roots L scapula; missing supe- rior and medial borders, glenoid cavity, coracoid and portions of body, axillary border and spine 5 thoracic spinous processes Feature # Specimen Comments L articular facet of atlas vertebra 17 vertebra fragments 36 rib fragments L & R radius; shaft frag- ments L ulna; shaft fragments R femur L femur; mid-shaft frag- ment, greater trochanter and neck L tibia; proximal and distal fragments R tibia; distal fragment (L/R)? tibia; mid-shaft fragment L & R fibula; distal fragments 84 long bone fragments L & R patella R innominate; articular surface and iliac portion of acetabulum L innominate; small por- tion of ilium and iliac portion of acetabulum 5 innominate fragments L & R calcanous L talus L cuboid fragment L 2nd cuneiform R 2nd cuneiform navicular fragment L 3rd metatarsal fragment L 4th metatarsal fragment R 5th metatarsal fragment foot phalanges: 5 prox- imal, 2 medial, 1 distal hand phalanges: 1 prox- imal 1st metatarsal fragment 4 metatarsal/ metacarpal (?) fragments 3 phalange fragments 1000 plus unidentifiable bone fragments less than 1 cm in size 75 primary inhumation, adult female, osteo- phytosis, copper stain- ing on many of the bones 5 cranial fragments L scapula; acromion and copper staining body fragments R scapula; glenoid and copper staining acromion, 3 body fragments 1983 EDWIN HARNESS MOUND 65 Feature # Specimen Comments Acknowledgments I wish to thank R. P. Mensforth, C. O. Lovejoy, and W. H. Kimbel for their review of the analysis. References Bass, William M. 1971 Human osteology: a laboratory and field manual of the human skeleton. Missouri Archaeological So- ciety, Columbia, Missouri. Brothwell, D. R. 1972 Digging up bones. British Museum (Natural History), London. Krogman, Wilton Marion 1962 The human skeleton in forensic medicine. Charles C. Thomas Publ., Springfield, 111. L & R frontal orbits L & R lateral portions of occipital L & R petrous portions 22 cranial fragments; parietal, temporal, frontal, occipital and sphenoid represented R mandible; condyle and portion of ascending ramus L mandible; body and condyle 1 incus 2 malleus 6 incisors 2 canines 6 deciduous molars 4 permanent molars R neural arch of 1st vertebra 24 neural arches 7 centrums L scapula; portion of axillary border and spine 25 rib fragments L humerus fragment 1 sacral fragment 100 plus unidentifiable bone fragments less than 1 cm in size 35 vertebral fragments L femur; proximal head, portion of shaft R femur; mid-shaft fragment 61 long bone fragments; upper/ lower limbs L patella R patella fragment R innominate; acetabu- lum, portion of ilium L innominate; portion of acetabulum 14 innominate fragments L talus fragment calcaneus fragment navicular fragment L 2nd cuneiform fragment L 3rd cuneiform fragment 19 metatarsal/ metacar- pal (?) fragments 300 plus unidentifiable bone fragments less than 1 cm in size 4 unidentifiable burned bone fragments 5 with copper staining copper staining 5 with copper staining 5 with copper staining 84 Infant, 6-8 mos. 8 MOLLUSC IDENTIFICATION AND ANALYSIS DAVID R. MORSE The archaeological salvage excavations of the Edwin Harness Mound (1976-1977) by the Cleveland Museum of Natural History and the Ohio Archaeological Council recovered 332 molluscan remains (19 species) in various cultural contexts associated with the mound. The shell remains analyzed included local Ohio naiads (51%) and terrestrial gastropods (31%) as well as marine gastropods ( 1 7%) from the Atlantic Coast. As this research project is only the latest phase in the investigation of the Edwin Harness Mound, it is likely that the mollusc material dis- cussed in this report represents only a small sample of the items that were originally associated with the mound structure (over 3,000 shells are part of the Edwin Harness Mound collection at the Ohio Historical Society). Al- though the sample recovered in these recent excavations is small, it is hoped that some interpretational context can be given to the importance and use of mollusc species by Hopewell groups (see Appendix 8.1). As part of the overall project, two major areas of the site were excavated. First, in order to define any remnant sections of the mound structure, the entire floor was ex- posed. Second, once the mound structure had been de- fined, exploratory radial trenches were opened with the aid of power equipment to expose any additional midden areas or structures. Prehistorically used shell material was recovered from both areas. In order to characterize the range of the mollusc species found on this site, an attempt was made to identify the genus and species of all items. At this site, however, most of the material consisted of only small sections of shell, many times lacking the phenotypic characteristics used to distinguish between species. As a result, the specific iden- tification in this report is tentatively offered to facilitate a discussion of dietary, environmental, and other variables. The relative frequency of species (Table 8.1) has been based on a tally of the number of locations from which a species was recovered (maximum number of individuals, MaxNl; Grayson 1973). Although this method has many problems (Grayson 1973), in this case, as the material was widely scattered over the site, it would seem unlikely that fragments are from the same individuals. To add to these problems, a third of the shell material recovered from the 1976-1977 Harness Mound excava- tions were found in a disturbed stratigraphic context. Many of the early investigators of the Ohio Hopewell complex included the Harness Mound in their field inves- tigations. This resulted in much of the interior sections of the mound being devoid of stratigraphically significant material. Most of the mollusc remains discussed in this report were recovered from the exterior edges of the house structures and mound gravel ring (Feature 1). In addition, the mollusc material from the Harness Mound curated at the Ohio Historical Society (OHS) ap- pears to be very selective. Most of this collection consists of exotic marine shell fragments and shell beads. M ollusc species indigenous to Ohio, all naiads, constituted less than 1% of the OHS collection (93% of the material from the 1976-1977 excavations consisted of indigenous spe- cies). It is interesting to note that the relative frequencies of local species that are represented in both collections are similar, suggesting that both samples are part of the same overall statistical population (see Appendix 8.2). Naiads Compared to other sites in southern Ohio, the range of naiad species at the Harness Mound site is somewhat lim- ited (see Table 8.1). A total of three species were identi- fied from the Harness Mound site: Lampsilis ovata (30%), Elliptio dilatatus (45%), Amblema costata (25%). These species are some of the most common naiad (freshwater pelecypods) species found in archaeological sites and in recent collections from central and southern Ohio (Stansbery 1965). Elliptio dilatatus (common filter clam or spike mussel), a very adaptable species, is found in a wide range of riverine habitats, which probably ac- counts for its high frequency at Harness. On the other hand, Amblema costata (common river mussel) and Lampsilis ovata (ovate river mussel) are found in small, slow moving, shallow streams or tributaries of large streams, preferring sandy and gravel bottoms. It is likely that these species somewhat reflect local conditions of the Scioto River near the mound. It is interesting to note that a wider range of naiads were found at the McGraw site (Prufer 1965) and the Morrison Village site (Prufer and Andors 1967), both of which are located near Harness on the Scioto River in Ross County. For example, in the McGraw site report, Stansbery ( 1 965) lists 25 naiad species. It is likely that a wider range of naiad species had existed in the Scioto River during the occupation of Harness than were found in the recent ex- cavations. The differences in the relative numbers of spe- cies between these sites probably reflect some dissimilar- ity in activities (e.g., Brose 1972). If the McGraw site is primarily a habitation site, then one would expect that there would be more evidence of food procurement and tool manufacture activities, which seems to be the case. The low number of naiad shells at Harness might be re- lated to occasional manufacture of a limited range of arti- facts, possibly including shell beads. Even though many have assumed that most shell beads were made from ex- TABLE 8.1 Summary of Mollusc Taxa Identified, Edwin Harness Mound Max NI* Max NI* Taxa 1976-1977 Exc % OHS % Naiads Fragments, species indeterminate 21 43.8 Lampsilis ovata 1 2.1 5 20.8 Lampsilis sp? 5 10.4 3 12.5 Total Lampsilis 6 12.5 8 33.3 Elliptio dilatatus 6 12.5 3 12.5 Elliptio sp? 6 12.5 10 41.7 Total Elliptio 12 25.0 13 54.2 Amblema costata 3 6.3 2 8.3 Amblema sp? 6 12.5 1 4.2 Total Amblema 9 18.8 3 12.6 Total 48 100.1 24 100.1 Terrestrial Gastropods Fragments, species indeterminate 11 28.2 Stenotrema leaii 1 2.6 Mesodon sp? 1 2.6 Anguispira alternata 1 2.6 Anguispira sp? 4 10.3 Total Anguispira 5 12.8 Oxychilus sp? 1 2.6 Zonitoides arboreus 4 10.3 Discus cronkhnitei 2 5.1 Helicodiscus para. 8 20.5 Retinella wheatyi 1 2.6 Vertigo morsei 2 5.1 Pupilla sp? 1 2.6 Cionella lubrica 1 2.6 Hawaiia miniscula 1 2.6 Total 39 100.2 Marine Gastropods Marginella sp? 2 28.6 500** Jaspidella jaspidae 2 28.6 3 Olivella sp? 1 14.3 1400 Oliva sp? 4 Vitrinella sp? 6 Busycon contrarium 1 14.3 2 Busycon spiratum 1 Fasciolariidae 2 unidentifiable shell bead 1 14.3 2000** Total 7 100.1 ♦maximum number of individuals ♦♦estimate 68 DAVID R. MORSE No. 39 otic marine gastropods, it is possible some naiads were used in the manufacture of shell beads. The three naiad species found at Harness are among those few species thick enough for shell bead manufacture. Terrestrial Gastropods At the Harness Mound site, at least ten terrestrial gas- tropod species were recovered from a total sample of 39 individuals (MaxNI). All items were collected as exca- vated; however, most of the smaller sized (below 3 mm) species were recovered from flotation samples of post holes and features (mostly in the genera Zonitoides , Dis- cus, Helicodiscus, Retinella , Vertigo , Pupuilla). All the terrestrial gastropods in the sample are indigenous to Ohio except for Oxychilus (cellar snail). This species was introduced from Europe in the eighteenth century (La- Rocque 1 970) and most likely is related to the excavations early in this century. The single specimen identified was found at the top of Feature 89. The interpretation of gastropods in an archaeological context is often related to paleo-environmental variables, especially in a general description of the vegetation cover of the site. Although it is difficult to relate many of the species discussed here to specific environmental condi- tions per se, some comments seem to be warranted. Most of the species found at Harness, especially in the genera Anguispira (cf. Alternata ) (striped forest snail), Zoni- toides (zonite shell). Discus (common disk shell), and Heli- codiscus (parallel disk shell), are most often found within forest detritus: logs, stumps, and other decaying material. So it seems reasonable to suggest that a certain amount of forest cover existed in the area of the mound. Helicodis- cus is also associated with second growth areas which might have been present near the mound. The absence of a significant number of species which favor either cleared areas, especially in the genus Mesodon , or heavily wooded areas, such as Anguispira kochi, is notable. This is in contrast to other archaeological sites in south-central Ohio (sites along Caesar Creek, Broseetal. 1979; Killen A and B sites, Brose et al. 1979; McGraw site, Prufer 1965; and Morrison Village site, Prufer and Andors 1967). Al- though this sample is small, it suggests that the vegetation cover near the mound could best be described as an open forested area or parkland. This range of gastropods at the Harness Mound varies from the published list of molluscan fauna from the Mor- rison Village (Prufer and Andors 1967) and McGraw (Stansbery 1965) sites near the Harness Mound along the Scioto River in Ross County. At both sites Anguispira kochi and Allogona profunda (profound forest snail), primarily a forest adapted species, are important at these sites. So, it would seem that these sites were more heavily wooded than was the area near the Harness Mound. Marine Gastropods Seven marine gastropods were recovered in the 1 976— 1977 excavations at the Harness Mound site. Taxonomi- cally, these molluscs can be grouped into three major fam- ilies: Olividae (olive shells), Marginellidae (marginella shells), Melongenidae (conch shells). It was obvious that these items were part of a much larger sample, as over 3,000 marine gastropods from the Harness Mound are curated at the OHS. In general, all items are native to the southeast Atlantic Coast or Gulf Coast of the United States. Although the items recovered in the present exca- vations seem to be finished artifacts, the earlier collec- tions at the OHS also include a whole range of unmodi- fied, partially finished and worked fragments. The largest item recovered in the excavations was a conch shell or lightning whelk. Busy con contrarium (Fea- ture 60, bundle burial), which had been modified by re- moving the columellae and trimming the edge surface of the aperture. (The term “whelk” as applied to B. contra- rium should not be confused with waved whelks in the family Buccinidae.) In comparison with other shell of B. contrarium , this item is somewhat small but still can be considered a fully developed individual (see Table 8.2). In the family Melongenidae , four species besides B. contra- rium are found along the Atlantic coast in the United States: B. carica( knobbed whelk), B. canliculatum (chan- neled whelk), B. spiratum (fig whelk), B. perversum (per- verse whelk). Of these, B. contrarium has the most south- ern range, which extends from North Carolina to Florida. In the Harness Mound collections at the OHS, two addi- tional B. contrarium , one B. spiratum , and one Fascio- lariidae (tulip shell or horse conch) have been identi- fied tentatively. These species are common to many Hopewell sites in the Ohio Valley. Seip and Mound City have a few molluscs in the genus Crassis (helmet shell) TABLE 8.2 Comparison of Dimensions, Melongenidae — Conch Shells Cat. No. Busycon contrarium Length Width OHS-Collections 1375G 7/34 20.5 cm 12.2 cm OHS-Collections 7/32 27.5 16.8 1976-1977 Excavations 28 IE 17.5 8.5 Mean of OHS Material 24.0 14.5 Overall Mean 21.8 Busycon spiratum 12.5 OHS-Collections 13757 7/34 21.6 cm 14. 1 cm 1983 EDWIN HARNESS MOUND 69 which are absent from Harness. Crassis is native only to southern Florida, which might have been beyond the range of southern contacts for the inhabitants of Harness. The remaining items are within the families Marginel- lidae ( Marginella sp?, marginella shells) and Olividae ( Olivella sp?, dwarf olive shells; Jaspidella jaspidae , jasper dwarf olive shells). The native range for these spe- cies is very similar to Busycon (North Carolina to Flor- ida), which seems to confirm that this region is the overall acquisition area for the Harness Mound shell material. These species, recovered from the 1976-1977 excavations, characterized 97% of the marine shell at the OHS from Harness and seem to be a good sample of the overall collection. Spatial Distributions As mentioned, the molluscan faunal collection from the Edwin Harness Mound is the result of excavations of the remaining mound floor in its entirety and a sampling of selective areas outside of the mound structure. This procedure included the excavation and collection of flota- tion samples from all features (burials, hearth, and activ- ity areas) and all post molds. Although large sections of the mound had been excavated previously, it was noted that some of the mound structure and sections of the mound floor were intact. In addition, the investigators found post hole outlines of a major house structure on the mound floor, part of an early construction phase. As mol- luscs were found within a number of features, an analysis of their distribution was considered important to the overall understanding of the site. Assemblage variability and frequencies of individuals found within a particular feature of significant stratigraphic level were compared to the site as a whole although sample size was not large. A comparison of molluscs within and beyond the limits of the mound found approximately 90% of the same local aquatic and terrestrial species in both areas. That eight times as many individuals were found within the mound area is attributed to the amount of area excavated. No significant differences in relative frequency or species dif- ferences were noted between differing intra mound struc- tures or between stratigraphic features and/or levels. These data suggest that proximal vegetation patterns within the general site area were consistently parkland forest throughout the sequence of structure and mound construction. Fifty percent of the smaller species ( Zoni - toides , Discus , Helicodiscus , Vertigo) were directly asso- ciated with post holes, and an additional 20% of these molluscs were found near the highest density of post holes. As these species commonly feed upon decaying plant material, it seems that some of these posts rotted in situ. Because such posts were evenly distributed in differ- ent structures across the mound floor, it may be that these structures are more or less coeval. At any rate, these data strongly support the position that every structural por- tion of the submound structure underwent some more-or- less extended period of exposure prior to either burning or entombment. Since these samples were small, these tentative hypoth- eses should be evaluated in future research. Discussion The identification and interpretation of mollusc mate- rial from the Edwin Harness Mound excavations and Ohio Historical Center collections has provided some context for the discussion of environmental and cultural variables, in spite of the small sample and extensively dis- turbed areas. Although all of the recommendations were provisional and inductive, this material adds to the gen- eral data base for Hopewell sites. It is hoped that mollusc samples will be collected in the future. Overall, the Harness Mound material seems to exhibit a different range of molluscs than were found on other sites along the Scioto River. Principally, it is suggested that the number of naiads were low because shell was used mainly as raw material for a small number of specific arti- facts (e.g., shell beads) and that indigenous gastropods are related to an open parkland forest environment. Analysis of the spacial distribution seems to indicate that these environmental conditions were present on the site during all construction phases of the mound. Appendix 8. 1 Mollusc Identification Analysis Edwin Harness Mound 1976-1977 Excavations Cat. No. Provenience Taxa 2D N535-E492.5, disturbed 1 frag., gastropod 2F N535-E492.5, disturbed 1 frag., gastropod 3H N547.5-E492.5, dis- turbed but close to original floor 1 frag., gastropod 3J N547.5-E492.5, dis- 1 ventral margin frag.. turbed but close to poss. Lampsilis , original floor burned 13B N547.5-E492.5, dis- turbed but close to original floor 1 frag., gastropod 2 1J N547.5-E502.5, dis- 1 frag., gastropod, poss. turbed near “Boys’ 1880” Anguispira 23A N547.5-E502, disturbed 1 shell bead 24A N535-E495, Fea. 89 1 Oxychilus 24C N535-E495, Fea. 89 1 frag., naiad 1 Vertigo morsei 1 Helicodiscus 70 DAVID R. MORSE 24C N535-E495, Fea. 89 1 dorsal margin frag. (bag #2) w/o beak. Elliptic) dilatalus 24E N535-E495, Fea. 89 1 Cionella lubrica 30 Surface, collected June 1 R dorsal margin frag. 1976 w/o beak, Elliptio dilatatus 32 N540-E492.5, south 1 R dorsal margin frag. balk surface w/o beak, Elliptio dilatatus , burned 37A N535-E492.5, post hole #16 2 frag., naiad 39B N535-E492 5, post hole #15 20 frag., naiad, burned 39B N535-E492.5, post hole 1 frag., naiad, burned (Bag #2) #15 64A N532.5-E492.5, dis- turbed 1 frag., naiad 64C N532.5-E492.5, dis- 1 dorsal margin frag., turbed Marginella 103B N522.5-E492.5, dis- turbed but near other archaeological de- posited material 1 Jaspidella jaspidea 103C N522.5-E492.5, dis- 1 frag., naiad, poss. turbed but near other archaeological de- posited material Elliptio 109 Surface, collected July 1 frag., naiad, poss. 1976 Amblema 1 17B N520-E505, disturbed, Fea. 1 1 frag., naiad, burned 1 18D N535-E495, front 1 R dorsal margin frag. loader cut #2, dis- w/ beak, Amblema turbed costata 128 A N535-E495, poss. post hole #8 10 frag., 1 naiad 130 N535-E500, poss. post hole #10 1 frag., naiad 138C N522.5-E495, floor surface 1 frag., naiad 138C N522.5-E495, floor 4 frag., naiad, poss. surface Elliptio 139B N520-E445, floor surface 1 frag., naiad 168 N522.5-E495, post hole 1 R dorsal margin frag. #80 w/o beak, Elliptio dilatatus 168 N522.5-E495, post hole 1 ventral margin frag. (bag #2) #80 Elliptio 183E N525-E495, Fea. 30 6 frag., naiad, poss. Amblema , burned 183F N525-E495, Fea. 30 1 dorsal section of Marginella 1 frag., gastropod 188A N532.5-E495, post hole 1 dorsal margin frag. #40 w/o beak, poss. Elliptio 217 No provenience 2 frag., naiad, burned 4 frag., gastropod, burned No. 39 1 frag., Amblema costata 219A Surface, collected June 10 frag., naiad, poss. 1967 Amblema, burned 223D N520-E507.5, Fea. 37 1 Anguispira alternata 271 A N552-E500, disturbed 8 frag., naiad 1 ventral margin frag., poss. Lampsilis 27 IQ N552-E500, post hole #179 1 Zonitoides 272C N560-E512.5, backdirt 10 frag., naiad 1 dorsal margin w/o beak, Lampsilis 28 IB N517.5-E502.5, Fea. 60 15 frag., naiad, poss. Elliptio 28 IE N517.5-E502.5, Fea. 60 1 Busycon contrarium 281C N517.5-E502.5, Fea. 60 4 frag., naiad 1 frag., poss. Steno- trema leaii 15 frag., naiad, poss. Anguispira 28 ID N517.5-E502.5, Fea. 60 10 frag., gastropod, poss. Anguispira 290C N535-E502.5, post hole #171, poss. disturbed 1 R dorsal margin frag, w/ beak, Elliptio dilatatus 295B N530-E502.5, post hole #160 1 Olivella 299A N500-E502.5, Fea. 55 6 frag., 2 gastropods, poss. Anguispira 305B Front loader Area #3, post hole #102 1 Helicodiscus 308K N524-E485, post hole #270 3 frag., naiad 3 10C N537.5-E497.5, post hole #252 1 F dorsal margin frag, w/o beak, Elliptio dilatatus 31 1C N522-E495, mound loading directly over Fea. 3 1 L dorsal margin frag, w/ beak, Lampsilis 312C N540-E497.5, post hole #66 30 frag., naiad Flotation Samples F2B N520-E507.5, Fea. 37 7 Zonitoides arboreus 4 Helicodiscus par- allelus 1 Discus cronkhitei 224H N522.5-E485, east of Fea. 1 8 frag., 1 gastropod 228B N547.5-E510, poss. post hole #20 3 frag., 1 gastropod 229 E N550-E507.5, mound 2 frag., naiad, poss. floor Elliptio, burned 229Q N550-E507.5, post hole 1 Helicodiscus par- #141 allelus 230L N522.5-E497.5, post 2 Helicodiscus par- hole #143 allelus 2 Zoniloides 1983 EDWIN HARNESS MOUND 71 2330 240A 240 B 240C 262A 260 D 260P 269C 269C (Bag #2) 269 D 270A F71A F100 F269C CCD68 Cat. No. 13735-7/99 13736 13737 N530-E492.5, Fea. 43 1 frag., naiad 13738-7/99 1 frag., naiad, poss. Elliptio 1 frag., gastropod 13739-7/99 1 frag., naiad, poss. Elliptio N547.5-E507.5, dis- 15frag., poss. Amblema 13740 1 frag., naiad, poss. Elliptio turbed 13741 1 frag., naiad, poss. Elliptio N547.5-E507.5, dis- 9 frag., naiad 13743 1 frag., naiad, poss. Elliptio turbed 1 L dorsal margin frag. 13745 1 whole shell, Elliptio dilatatus w / beak, Amblema 13746-7/99 1 ventral margin frag., poss. Elliptio 1 R dorsal margin frag. 13747-7/99 1 ventral margin frag., poss. Elliptio w/o beak, Lampsilis 14582-7/99 1 whole shell, Lampsilis ovata N547.5-E507.5, dis- 1 Helicodiscus 14583-7/99 1 frag., Lampsilis ovata turbed 14584-7/99 1 whole shell, Amblema costata N550-E509, above 1 L dorsal margin frag. 14585-7/99 1 whole shell, Lampsilis ovata Fea. 33 w/ beak, Elliptio 14586-7/99 1 whole shell, Amblema costata dilatatus 14587-7/99 1 whole shell, Lampsilis ovata N557.5-E500, Fea. 44 1 dorsal margin frag. 14588-7/99 1 whole shell, Lampsilis ovata w/ beak, Amblema 1 ventral margin frag., poss. Amblema costata 14589-7/99 1 ventral margin frag., poss. Lampsilis N557.5-E500, Fea. 44 6 frag., naiad 14590-7/99 1 frag., Elliptio dilatatus backdirt 1 dorsal margin frag, w/ #7 1 frag., naiad, Amblema beak, poss. Amblema 1 frag., naiad, poss. Elliptio Backhoe trench #4, 1 frag., gastropod 1 frag., naiad, poss. Lampsilis post hole #147 outside of Fea. 1 not labeled 1 whole shell, Elliptio dilatatus Backhoe trench #4, 12 frag., naiad Conch post hole #147 1 ventral margin frag., Lampsilis 1 frag., poss. Elliptio 6 frag., naiad 1 375G *-7 / 34 1 Busycon contrarium N517.5-E502.5, -7/32 13757-7/34 1 Busycon contrarium 1 Busycon spiratum backhoe trench #4 13755-7/33 1 Fasciolariidae Fea. 56 1 frag., gastropod 14590-7/99 1 frag., poss. juvenile of Pleuroploca N517.5-E507.5, back- 7 Helicodiscus par- gigantea hoe trench #4, Fea. 60 allels Post hole #147 1 frag., poss. Mesodon Small Marine Gastropods Post hole #147 3 Zonitoides I frag., poss. Pupilla 40 Helicodiscus par- allels 36 Relinella wheatleyi 66 Zonitoides arboreus I I Vertigo morsei 7/36 4 Oliva sp? 1400** Olivella sp? 500** Marginella sp? 6 Vitrinella sp? 3 Jaspidella jaspidea Modified Shell Carbon Samples Backhoe trench #6 3 Retinella sp? 18 Hawaiia minuscula Appendix 8.2 Preliminary Mollusc Identification Analysis Edwin Harness Mound Ohio Historical Society Harness General 20 dorsal sections, Marginella , burnished 1 string — single row, ventral sections, Marginella 1 string — double row, dorsal sections, Marginella 1 string — double row, ventral sections, Marginella , burnished 4 strings — double row, dorsal sections, Marginella 1 string — single row, dorsal sections, Olivella Naiads Taxa 1 frag., naiad, poss. Elliptio 1 frag., naiad, poss. Elliptio 1 frag., naiad, poss. Elliptio 1 frag., naiad, poss. Lampsilis ♦This is the catalogue number written on the specimen. Based on the original catalogue of materials collected by Moorehead, the number likely should be 13756. ♦♦estimate 72 DAVID R. MORSE No. 39 References Abbott, R. Tucker 1 974 American seashells (second ed.). Van Nostrand Rein- hold, New York. Brose, David S. 1 972 The molluscan fauna. In The Schultz site, compiled by J. E. Fitting. University of Michigan Museum of An- thropology Memoirs No. 4, Ann Arbor. Brose, David S., Donald R. Bier, Jr., Judith Astramerki, Fred- rick Chapman, Richard Ford, Robert Mensforth, David Morse, and Paul Storch 1979 Prehistoric occupation of the Killen Electric Generat- ing Station near Wrightsville, Adams County, Ohio. Submitted to the United States Department of the In- terior Interagency Archaeological Services, Atlanta; Contract 5880-7-0070 by the Cleveland Museum of Natural History, Cleveland. Brose, David S., and Nancy M. White with contributions by Stephanie Belovich, Fredrick Chapman, W. B. Clapham, Richard I. Ford, John Lallo, David Morse, and Paul Storch 1979 Archaeological investigations of prehistoric occupa- tion in Caesar Creek Lake: Clinton, Greene, and Warren Counties, Ohio. Submitted to the United States Department of the Interior Interagency Ar- chaeological Services, Atlanta; Contract C-5555 (formerly CX5880-7-01 34) by the Cleveland Museum of Natural History, Cleveland. Cinadr, Thomas J., and David S. Brose with contributions by Stephanie Belovich, Donna Benson, David Morse, Paul Storch, and Nancy M. White 1978 Archaeological excavations in Caesar’s Creek Lake, Ohio: Section 1 1 . The Carr Mill Race Site (33Wa75). Submitted to the United States Department of the In- terior Interagency Archaeological Services, Atlanta; Contract CX-5880-7-0134 by the Cleveland Museum of Natural History, Cleveland. Grayson, Donald 1973 On the methodology of faunal analysis. American An- tiquity 39:432-438. LaRocque, Aurele 1970 Pleistocene mollusca of Ohio. State of Ohio, Depart- ment of Natural Resources Division of Geological Survey Bulletin 62(4). Prufer, Olaf H., D. H. McKenzie, O. Pi-Sunyer, H. C. Cutler, R. A. Yarnell, P. W. Parmalee, D. H. Stansbery 1965 The McGraw site: a study in Hopewellian dynamics. Cleveland Museum of Natural History, Scientific Publications (n.s.), 4(1). Prufer, Olaf H., and Ellen Andors 1 967 The Morrison Village site (33Ro-3): a terminal prehis- toric site in Ross County, Ohio. In Studies in Ohio Archaeology , edited by Olaf H. Prufer and Douglas H. McKenzie, pp. 187-229. Kent State University Press, Kent, Ohio. Stansbery, David H. 1965 The molluscan fauna from the McGraw site. In The McGraw site: a study in Hopewellian dynamics , edited by Olaf H. Prufer, pp. 1 19-124. Scientific Pub- lications of the Cleveland Museum of Natural His- tory, (n.s.), 4(1). 9 THE FLINT SOURCES KENT D. VICKERY Introduction Four collections of chipped stone items were analyzed: 1. artifacts and debitage recovered during the 1976 field season of excavation at the Edwin Harness Mound, 33Ro22, consisting of 127 items; 2. 1,180 bladelets and bladelet cores surface collected by Robert Harness in the vicinity of the Harness Mound; 3. 627 surface collected artifacts and pieces of debitage from Harness’s Site 18; and 4. a similar collection of surface material from Har- ness’s Site 25, consisting of 720 chipped stone items. Chipped stone material from the Harness Mound ex- cavation consisted of 91 pieces of debitage, 23 bladelets, 1 bladelet core, and both unifacial and bifacial tools and weapons. A bipolar core and 1 tool edge rejuvenation flake were treated as artifacts (rather than debitage) for purposes of this analysis. The Site 1 8 collection included 7 projectile points, 4 bifacial artifacts, and 4 bladelets in addition to abundant debitage. Debitage also dominated the Site 25 collection, which included 4 projectile points and 14 bladelets. The remaining surface collection con- sisted of 1,027 bladelets and 153 bladelet cores. For site locations see Greber, Davis, and DuFresne (1981). The present analysis focuses on flint/ chert raw material iden- tifications for the 2,654 items from these four collections and on the chipping techniques which produced the 91 pieces of debitage from the Edwin Harness Mound. This latter analysis focuses on 1) incidence of heat alteration, 2) presence or absence of post-detachment utilization and/ or retouch, 3) debitage type, and 4) technique of de- tachment. Unmodified and broken chert pebbles which were found in post holes and features of the 1976 excava- tions of the Harness Mound were also examined for raw material identifications. These pebbles lacked pronounced bulbar scars and compression ring segments, platform preparation, or other evidence of intentional fracture for artifact production. Methods Flint raw material identifications were made by ex- amining specimens under low magnification using a bi- nocular microscope and comparing them with samples obtained from verified outcrop locations. In addition, cortex characteristics were recorded for the Harness Mound debitage in an effort to determine the general con- texts from which chert was obtained. “Pebble” chert is material dislodged from its naturally occurring matrix, transported by stream or glacial action, and subsequently deposited some distance away from its source area. In the process, such material acquires a hard, smooth surface patination (cortex). Flakes with cortical retention on por- tions of the dorsal face and/or platform, as well as arti- facts, are referred to in this report as “Local Pebble” chert if they cannot be matched to any of the samples from veri- fied outcrop locations. Artifacts and debitage lacking cor- tical remnants and unidentifiable as to flint source are re- ferred to as “Unknown.” Heat alteration was recorded in an effort to detect whether or not thermal treatment was part of the se- quence whereby artifacts were produced from parent raw material. For those specimens exhibiting thermal modifi- cation, macroscopic and microscopic inspection was the basis for distinguishing between heat alteration and heat damage. The former is commonly recognized by adjacent lustrous and non-lustrous scars (Collins and Fenwick 1974; Greber, Davis, and DuFresne 1981:513), a smoky or cloudy appearance, and/or slight discoloration. It is only among these specimens that candidates for proper heat treatment might be present. By contrast, heat dam- aged specimens often showed extensive crazing, potlid depressions, pronounced color changes, scalloped edges or more serious heat-induced breakage, and / or a chalky texture caused by water having been driven out under thermal stress. Utilization and retouch were determined by micro- scopic examination. Retouched artifacts and debitage had had a series of small, contiguous chips intentionally removed from one or more margins, the scars of which were typically uniform in size and shape. Utilized flakes were often recognized by marginally detached chip scars that were less regular in size, shape, and placement and that occasionally resulted in a rather jagged contour. Uti- lized edges commonly assumed one or more of four basic forms: 1) nibbling, consisting of diminutive scars with an ovate configuration that should have resulted from cut- ting or scraping relatively soft materials (Binford 1963: 207); 2) hinge fractures with square or rectangular config- urations and abrupt terminations, probably resulting from contact with more resistant material; 3) attrition, the gradual dulling or wearing away of formerly sharp edges by the sustained removal of tiny chips or pieces of the tool margin(s); and 4) polish, which succeeds attrition as the functional portions of tools undergo progressive modifi- cation with use against relatively soft materials. Polish is recognized as a gloss or sheen under magnification. 74 KENTD. VICKERY No. 39 Debitage: definitions An unmodified piece of flint is reduced to a finished tool through stages. The debris created at each stage is reasonably distinctive, which facilitates a classification of waste material for the sequence of stages. One component of debitage analysis involves the recognition and defini- tion of various debitage types which reflect the reduction sequence. For the Harness Mound collection only, each of several such debitage types is described below in a logical knap- ping order following the off-site selection or importation of raw material for artifact production. Checked pebbles. The first step is determining whether or not the raw material acquired is suitable for knapping. The quality of unmodified pebble chert can be checked by removing one or more flakes to expose the sub-cortical matrix. This activity results in a checked pebble. Com- parable treatment of bedded material, which often exhib- its thinly weathered surfaces from frost action, was not observed in the sample. Primary decortication flakes. The reduction sequence is initiated with the systematic detachment of flakes from a weathered piece of flint, which creates flakes retaining cortex over the entire outer (dorsal) face. The removal of these primary decortication flakes (White 1963:5) achieves partial decortication of the objective raw material and produces a core tool blank. Secondary decortication flakes. Additional flake de- tachment generates secondary decortication flakes (White 1 963:5). These flakes often retain the scars of one or more previously detached primary decortication flakes and some cortex on the dorsal face, but may be wedge-shaped with cortex remaining only on the thickened edge. The result of their removal is a blank in a more advanced stage or an early stage of a preform. Primary flakes. Wilmsen (1970:25) refers to post- decortication flakes associated with the shaping of a blank or preform by reducing its mass as primary flakes and defines them as “those which were struck from a de- corticated core.” Primary flakes retain the scars of pre- viously detached flakes over most or all of the dorsal face, with small amounts of cortex sometimes remaining (commonly in the center). They are characteristically thick and often triangular in section. Their removal pro- duces, or reduces the mass of, a preform. Thinning flakes. Crabtree (1972:94) defines thinning flakes as “flakes removed from a preform by pressure or percussion to thin the piece for artifact manufacture. Thinning flakes are also removed to thin a biface or a uniface. ” Because the process of bifacial reduction is con- tinuous, the distinction between thinning flakes and pri- mary flakes detached in the advanced stages of reducing the mass of blank or a preform is often arbitrary. The former are typically thinner than primary flakes and re- tain scars over the entire dorsal face. Their removal pro- duces an advanced stage preform or a finished bifacial tool or weapon. Not all bifacial artifacts reach an ad- vanced stage of reduction because their suitability as tools for certain kinds of tasks (e.g., heavy cutting, scraping, or chopping) may have been achieved by removing only de- cortication or primary flakes. As the preform approaches the morphology of the fin- ished tool, sharpening flakes may be detached in an effort to strengthen, straighten, and/or sharpen the edges or to shape them by creating indented or projecting contours (Vickery and Lambert 1977). Representing the final stage of the reduction sequence, these tiny chips may not be encountered unless fine mesh screening or flotation tech- niques are employed during site excavation. For the analyzed sample, flake type is used for all flake debitage from the primary decortication stage through the thinning stage, no sharpening flakes having been rec- ognized in the Harness Mound debitage; debitage types include all flake types plus checked pebbles and unclassifi- able core fragments (exclusive of bladelet and bipolar cores). The debitage types used for this analysis pertain to a bifacial reduction sequence. A different set of debitage types may characterize unifacial artifact production (in- cluding bladelets) and such other specialized knapping as the use of bipolar techniques. Another dimension of debitage analysis involves at- tempts to infer various techniques of flake removal. The following categories were used to analyze the Harness Mound flake debitage according to removal technique: 1 ) hard hammer percussion, 2) soft hammer (“billet”) per- cussion, and 3) indirect percussion (“punch”) or pressure. Stone hammers were probably used for hard hammer percussion, deer or elk antler beam segments for soft hammer percussion, and antler tines for pressure or indi- rect percussion. These techniques are recognizd among waste flakes mainly on the basis of shape and platform configurations. Shape is influenced in part by the force used to detach flakes while platform remnants on flakes should correspond in size and shape with the contacted portion of the percussion or the pressure tool used for flake detachment. Hard hammer percussion flakes characteristically have a pronounced bulb of percussion and a relatively large platform remnant. They are usually thick and may either hinge or feather at their terminations. The incidence of hinging is greater than with other percussion techniques. Hard hammer percussion is the predominant technique of flake removal in the initial stages of tool manufacture. As the blank or preform was reduced, a stage would have been reached where continued use of hard hammer likely would have resulted in breakage from overly stress- ful shock. At this point, the shift to a soft hammer for further thinning would be expected. Soft hammer percussion flakes generally have a less pronounced bulb of percussion and a platform remnant that is narrower (as measured from the dorsal face to the 1983 EDWIN HARNESS MOUND 75 ventral face) than those detached by hard hammer per- cussion. Platforms are often of an elongated lenticular shape but may be recurvate from the previous detachment of a flake from the dorsal face at one end of the platform remnant. Soft hammer percussion flakes are characteris- tically thinner than flakes detached by a hard hammer and usually feather at their terminations. No attempt was made to distinguish indirect percussion from pressure as flake detachment techniques. In both cases, a tool such as the tip of an antler tine would have been used to apply force on a point of the core platform rather than an area. Thus, flakes removed by indirect per- cussion and pressure are recognized by the presence of a small circular to oval platform remnant in addition to a bulb of applied force that is only subtly swollen. Such flakes are characteristically thin and have feathered terminations. When correlated, flake type and technique of flake de- tachment determinations provide a basis for inferring the points in the reduction sequence at which shifts occurred from hard to soft hammer percussion, and, if employed, to indirect percussion or pressure. This is one way of de- scribing a lithic waste flake assemblage. More impor- tantly, this method monitors the results of prehistoric cul- tural conditioning with the attendant potential of reveal- ing ethnically diagnostic similarities and differences from one assemblage to the next. Raw Materials Flint raw material identifications were attempted for all four collections. At least 9 varieties are present among the excavated artifacts and debitage from the Harness Mound, while 1 5 were identified in the larger sample that includes Harness’s surface collected material. Figure 9.1 identifies the outcrop locations of these varieties in rela- tion to the Harness site complex. Fig. 9. 1. Areas of outcrop of flint varieties present at Harness. ▲ Edwin Harness Mound (33Ro22); 1-Columbus; 2-Delaware; 3-Prout; 4-Plum Run; 5-Boggs; 6-Flint Ridge; 7-Upper Mercer; 8-Za- leski; 9-Brassfield; IO-Cedarville-Guelph; 1 1— Bisher; 12-Brush Creek; 13-Harrison County; I4-Kanawha Black. 76 KENT D. VICKERY No. 39 Geological documentation of all Ohio varieties is pro- vided by Stout and Schoenlaub (1945); archaeological occurrences of two or more varieties are discussed by Fowke (1894, 1902, 1928), Holmes (1919), Prufer and Baby ( 1963), Converse (1972), and Morton and Carskad- den (1972). Each is summarized here. Flint Ridge. This well-known variety is described by Smith (1885), Moorehead (1892:30-48), Mills (1921), Crawford (1967), and Patterson (1979), among others. It is variegated and fine grained with a vitreous luster. Delaware. Delaware flint is tan to dark grayish brown with sparsely scattered, tiny ostracod inclusions that are white in the non-heat altered state. It is described by Con- verse (1972:37). Brush Creek. Described by Carskadden and Donald- son (1973), Brush Creek flint is tan to light gray to dark brown mottled with diagnostic small orange spots on its cortex. Zaleski. Zaleski would have been the closest of the bedded varieties available to the Harness site inhabitants. This material is brownish black to black and fine grained with a vitreous luster. Kramer (1953) and Morton and Carskadden (1972) discuss its characteristics and prehis- toric utilization. Upper Mercer. Carskadden (1971) and Morton and Carskadden (1972) describe this flint variety, which ranges from light to dark gray and from bluish black to black with white inclusions and thin streaks. The darker material is often fine grained with a vitreous luster. Boggs. Boggs flint is described by Morton and Car- skadden (1972). It is medium-dark gray to black, slightly grainy in texture with a rather dull luster, and may have numerous white fossil (commonly fragmented crinoid stem), quartz, and pyrite inclusions. An orangish, ferru- ginous patination is distinctive. This variety may be un- derrepresented because of the non-availability of com- parative samples when the Harness surface collections were examined. The excavated collection from the Har- ness Mound was reexamined, however, for Boggs chert. Knife River. Clayton. Bickley, and Stone (1970) de- scribe this fine-grained, honey-colored flint, and Braun, Griffin, and Titterington ( 1 982:65—89) summarize mid- western occurrences in Middle Woodland contexts. Knife River flint has a color range from light to rather dark, a diagnostic thin white cortex, and relatively large opaque inclusions in a translucent matrix. That the presence of Knife River represents long-distance acquisition, proba- bly from west-central North Dakota, is noteworthy. Harrison County. Harrison County flint (also known as “Indiana hornstone”) from extreme south-central In- diana also represents long-distance acquisition. Collett (1878:421-423), Shaver et al. (1970:146), Lilly (1937: 101-104), Guernsey (1937), Seeman (1975), and Tomak (1980:110; 1982:37-38) are among those who document the archaeological utilization or geological context of this well-known flint variety. It formed in nodules and typi- cally ranges from dark bluish gray at the center to light gray near the cortex. Harrison County flint is homogene- ous and fine grained, slightly translucent at the edges of thin flakes, and quickly weathers to a uniformly light to medium blue-gray. Cedarville-Guelph. Stout and Schoenlaub ( 1945:20-21) document geological occurrences of this flint variety, and Converse (1972:36-37) mentions its archaeological utili- zation. It is grainy, pinkish tan or light gray, and has nu- merous darker gray spots throughout the matrix which are diagnostic for the variety among those identified in the analyzed sample. Columbus. Stout and Schoenlaub (1945:21-24) and Prufer and Baby (1963:44) discuss this flint variety. It is light and dark brown mottled and rather coarse grained. Prout. Comparative samples of Prout chert upon which identifications were made were collected from out- crop on the Plum Brook NASA base near Sandusky, Ohio. They are fossiliferous with small pyrite inclusions and range from cream to tan and from light to dark gray. Occurrences of chert in the Prout limestone member doc- umented by Stauffer (1909, 1916) and Stumm (1942) in this area probably refer to this variety. James L. Murphy of Ohio State University (letter to K. D. Vickery, April 22, 1982) believes that the chert from Perkins Township in Erie County mentioned by Stout and Schoenlaub (1945: 31) and the “Pipe Creek” chert mentioned by Stothersand Rutter (1978) are also of the Prout variety. Brassfield. Hastings (1969) gives a description and summary of the prehistoric utilization of Brassfield chert. It is white to flesh pink and is abundantly fossiliferous. Bisher. Bisher flint is grainy and homogeneous, tan to light purplish, and occasionally banded. It is described by Hastings (1969). Plum Run. The prehistoric use of Plum Run flint as represented by quarries is discussed by Murphy and Blank (1970); Converse (1972:38) provides a description. Fresh comparative samples are fine grained and light to dark gray and blue mottled, while weathered specimens are tan, orange, red, brown, and green mottled with streaks and patches of white-light gray inclusions. Kanawha Black. The source of Kanawha Black flint is central and southwestern West Virginia. It is bluish black to black, non-vitreous, homogeneous, and very grainy. White (1903:328-332, 1908:487-488), Krebs (1914:255- 266, 643-644), and Price (1921) document geological oc- currences, Olafson (1964, 1972) archaeological occur- rences. Hastings (1971) discusses the use of pebble raw mate- rial, which is ubiquitous in distribution. Results of Analysis and Interpretations Presented below is the distribution of thermally altered 1983 EDWIN HARNESS MOUND 77 specimens according to debitage type in the sample of 91 pieces of debitage from Harness Mound: Debitage Type Heat altered Heat damaged Total Core fragments 2 0 2 Primary decortication flakes 3 2 5 Secondary decortication flakes 2 2 4 Primary flakes 2 2 4 Thinning flakes 0 2 2 Uncertain 1 4 5 Total 10 12 22 Twenty-two specimens (24% of the total debitage) are heat altered or damaged, which exceeds the 20% figure that Collins and Fenwick (1974: 143) represent as “larger than would be expected by chance” in their search for heat treating technologies among the chipped stone assem- blages of 12 archaeological sites in Kentucky. More than half of the heated specimens from the Harness Mound, however, were so damaged that the functional effective- ness of the flakes (or the knapping suitability of the cores from which they were struck) would have been impaired. These specimens may simply represent post-detachment damage unrelated to intentional heat treatment. None of the 10 heat altered pieces of debitage exhibits the fre- quently subtle changes associated with properly heat treated material. Furthermore, heat altered or damaged specimens occur throughout the reduction sequence rather than being clustered at any particular stage. These observations suggest that intentional heat treating of flint and chert was not incorporated into the bifacial reduction sequence. By contrast, the bladelet industry apparently did feature the heat pretreatment of cores (Greber, Davis and DuFresne 1981:513-514). Table 9. 1 gives the proveniences of the Harness Mound debitage. The frequency distributions of debitage and flake types are presented in Tables 9.2 and 9.3. Among the flake debitage, 14 (19.2%) are either utilized or retouched (Table 9.2). One core fragment exhibits a utilized edge (Table 9.3). This indicates a degree of chert resource con- servation on the part of the people who produced the sample. The ratio of retouched to utilized to non-utilized flakes is 1 : 1.8: 14.6. Among the flake debitage, it is interesting to note an increase in the proportion of utilized and retouched flakes from the early to the late stages of the reduction sequence: 1 0% of the primary decortication flakes are uti- lized or retouched; of the secondary decortication flakes, 18% are utilized or retouched, and the percentages of uti- lized and retouched primary flakes and thinning flakes are 50% and 60%, respectively (Table 9.2). If the sample is representative, selection of relatively thin flakes for spe- TABLE 9.1 Provenience of Debitage from Edwin Harness Mound Fragments Checked Pebbles Primary Decort. Flakes Secondary Decort. Flakes Primary Flakes Thinning Flakes Uncertain Total 1 I 1 1 1 1 1 1 4 1 5 15 1 1 1 2 3 1 1 1 1 1 1 1 3 2 14 3 2 1 5 36 5 13 29 1 1 10 5 18 91 5.5 14.3 31.9 12.1 11.0 5.5 19.8 100.1 78 KENT D. VICKERY No. 39 TABLE 9.2 Correlation of Flake Type and Technique of Flake Removal, Edwin Harness Mound Technique of '"^flake Removal Flake Type Hard Hammer Percussion Soft Hammer Percussion Indeterminate Percussion Indirect Percussion Pressure Platform missing Uncertain Total No. % Primary Decort. T 10 1 1 15 2 29 U 1 1 39.7 R 2 2 Secondary Decort. T 4 7 11 U 1 1 15.1 R 1 1 Primary T 1 2 7 10 U 1 2 2 5 13.7 R Thinning T 4 1 5 U 1 1 6.8 R 2 2 Uncertain T 3 1 1 7 6 18 U 1 1 24.7 R Total T 18 7 2 1 37 8 73 100.0 U 3 3 3 9 12.3 R 2 3 5 6.8 % 24.7 9.6 2.7 1.4 50.7 11.0 100.1 T = Totals U = Utilized R = Retouched cific or perhaps multiple tasks is suggested. Primary and thinning flakes should have more acute edge angles and sharper edges than decortication flakes. Hence, if the sample reflects intentional selection of flake debitage for tool use, marginal sharpness seems to have been a more important consideration than strength. This in turn sug- gests a greater association with cutting tasks than with scraping tasks, which parallels the apparent utilization of bladelets in the Edwin Harness Mound and vicinity. Technique of flake removal is correlated with flake type in Table 9.2 and with raw material in Table 9.4. The total number of flakes in the sample is quite small and the often diagnostic platforms of half are missing. Nevertheless, a few observations can be made on the data bearing on dif- ferential tool use in the knapping sequence. Decortication waste flakes are dominated by hard hammer percussion, with a noticeable reduction in the number of hard hammer percussion detached flakes from the primary to the secondary decortication stages. Only one primary flake and no thinning flakes in the sample were detached by this technique. There is a corresponding increase in the use of soft hammer percussion from the stage of primary flake detachment to the thinning stage. Only one soft hammer percussion detached flake is rep- resented in the early stages of decortication. Thus, hard hammers were apparently employed in the initial stages of knapping at least cortical raw material and the shift to a soft hammer occurred when the mass of the blank or pre- form was being reduced by the removal of primary flakes. Indirect percussion or pressure is represented by a single flake of indeterminate type. However, this flake removal technique is often employed in the final stages of the re- duction sequence when sharpening flakes are detached. “Soft” cortex, such as is commonly present on chert nod- ules that have not been transported far from their outcrop areas by natural agencies, was not observed in the debit- age sample. Thus, the presence of cortex on the dorsal faces of flakes and on other debitage is presumed to repre- sent pebble raw material. In the debitage sample, a reduction in the proportion of specimens with remnant pebble cortex is observable from the checked pebble stage (100.0% pebble material) to the primary decortication stage (79.3%) to the secondary de- cortication stage (72.7%) to the primary flake stage (10.0%) to the final thinning stage (0.0%). Seemingly con- tradictory is the lack of a 100% representation of pebble material in the decortication stages. Decortication , how- ever, refers not only to the detachment of pebble cortex but to the removal of portions of very thin patinas result- ing from frost wedging as well. Such frost-cracked mate- rial may have occurred as bedded chert in outcrop loca- tions that served as prehistoric quarries. Otherwise, the 1983 EDWIN HARNESS MOUND 79 TABLE 9.3 Correlation of Debitage Type and Flint/Chert Variety, Edwin Harness Mound Flint / Chert Variety Debitage Type \ Delaware Brush Creek Flint Ridge Cedarville- Guelph Columbus Brassfield Upper Mercer or Boggs Local Pebble Chert Unknown Total No. % Core Fragments T 2 1 1 1 5 U 1 1 5.5 R Checked Pebbles T 3 1 2 1 6 13 U 14.3 R Primary Decort. Flakes T 7 2 1 1 1 14 3 29 U 1 1 31.9 R 1 1 2 Secondary Decort. Flakes T 2 2 3 3 1 11 U 1 1 12.1 R 1 1 Primary Flakes T 2 1 2 1 1 3 10 U 1 1 2 1 5 11.0 R Thinning T 2 2 1 5 U 1 1 5.5 R 1 1 2 Uncertain T 1 2 5 5 5 18 U 1 1 19.8 R Total T 17 10 10 4 3 2 1 30 14 91 100.1 U 1 2 3 1 1 1 1 10 11.0 R 2 2 1 5 5.5 % 18.7 11.0 11.0 4.4 3.3 2.2 1.1 33.0 15.4 100.1 T = Totals U = Utilized R = Retouched debitage assemblage reflects a bifacial artifact industry that was dominated by pebble raw material. This is evi- dent in the distribution of pebble material throughout the knapping sequence and in reduced numbers of flakes from the beginning to the end of the sequence, suggesting that not a great deal of bifacial thinning occurred beyond the decortication stage. Furthermore, no sharpening flakes were recognized in the analyzed sample even though feature contents were consistently subjected to flotation appropriate to their recovery from a soil matrix. Their absence (or scarcity) suggests either that this ad- vanced stage of artifact manufacture occurred at one or more off-site loci or that bifacial tools were considered finished products after the removal of thinning flakes. With this possible exception, the debitage reflects a rather homogeneous assemblage in that the same raw material apparently passed through the entire reduction sequence with neither significant inputs of additional raw material at any one stage in this sequence nor removal of blanks or preforms on their way to becoming finished artifacts. Flint/ chert raw material excavated from the Edwin Harness Mound is correlated with debitage type in Table 9.3 and with technique of flake removal in Table 9.4. Ta- ble 9.5 presents the frequency distribution of identified and unidentifiable flint varieties among the four analyzed collections, including questionable raw material and cor- tex identifications. Items of Flint Ridge flint were exam- ined for the presence or absence of cortex for the Harness Mound sample only. Therefore, Flint Ridge pebble is un- derrepresented in Table 9.5. At least one specimen in Rob- ert Harness’s surface collection was observed with pebble cortex, but none were present in the excavated Harness Mound collection. One striking feature of flint raw material utilization is 80 KENT D. VICKERY No. 39 TABLE 9.4 Correlation of Technique of Flake Removal and Flint/ Chert Variety, Edwin Harness Mound Flint /Chert Variety Technique of Flake Removal Delaware Brush Creek Flint Ridge Cedarville- Guelph Columbus Upper Mercer or Boggs Local Pebble Chert Unknown Total No. % Hard Hammer Percussion T 3 2 1 1 6 5 18 U 2 1 3 24.7 R Soft Hammer Percussion T 1 2 2 1 1 7 U 1 1 1 3 9.6 R 1 1 2 Indeterminate Percussion T 1 1 2 U 2.7 R Indirect Percussion/ Pressure T 1 1 U 1.4 R Platform missing T 7 5 6 1 2 9 7 37 U 1 2 3 50.7 R 1 1 1 3 Uncertain T I 7 8 U 11.0 R Total T 12 9 10 2 3 1 23 13 73 100.1 U 1 2 3 1 1 1 9 12.3 R 2 2 1 5 6.8 % 16.4 12.3 13.7 2.7 4.1 1.4 31.5 17.8 99.9 T = Totals U = Utilized R = Retouched the dominance of Flint Ridge in all but the Harness Mound collection. Flint Ridge flint accounts for at least 93% of Harness’s surface bladelets and bladelet cores. For Sites 18 and 25, 81% and 89% of the items, respectively, are of Flint Ridge flint. This high-grade material also accounts for half of the excavated artifacts from the Harness Mound; all (17 bladelets and 1 bladelet core) are related to the bladelet in- dustry. Only 2 Flint Ridge flakes (that are not necessarily associated with the bladelet industry) were recovered from the Harness Mound. Seven of the remaining 8 flakes are virtually identical in appearance. They may have been struck from the same core, which seems also to have been the parent core for 4 of the 23 bladelets and bladelet frag- ments recovered. At least 6 of these 7 were from the same feature (Feature 30). The single primary decortication flake of Flint Ridge probably represents the acquisition of bedded material without any (or with very little) knapping prior to transport. Two thinning and primary flakes each are not necessarily part of a bifacial reduction sequence; rather, they may have resulted from bladelet core prepa- ration of a suitable preform. Five Flint Ridge flakes, “un- certain” as to flake type, may represent snapped frag- ments from attempted bladelet removal or flake segments detached in an effort to reshape one or more cores. The association of Flint Ridge raw material with blade- let production is also evident, although not quantified, in the Site 1 8 and Site 25 collections. It is possible that the remarkably high proportion of Flint Ridge flint (and/or items related to the bladelet industry) in the surface mate- rial is due to selective collecting. Utilitarian artifacts and associated debitage of non-Flint Ridge materials are present, however, in all three collections. It therefore seems unlikely that selective collecting, had it occurred, significantly inflated the proportional representation of Flint Ridge. The possibility exists that Flint Ridge flint occurred naturally in the vicinity of the site complex, having been transported there by natural agencies. As noted, at least one specimen in the Harness surface collection is indeed 1983 EDWIN HARNESS MOUND 81 pebble material. James L. Murphy (letter to K. D. Vickery, April 22, 1982) cautions that some flint of the Vanport limestone member in Jackson County to the south and east of the Harness site complex is high-grade and resembles material from the famous Flint Ridge de- posits in Licking County and vicinity, to the north and east of Chillicothe. Being located in the ancient Teays val- ley, this more southern variety may have been transported naturally into the area. If the flint being exploited at Har- ness had been carried to the area in this manner, however, one would expect to find at least some low-grade, weath- ered material, or at least a range of variation in quality. Because nearly all of the flint is high-grade, its source area for at least most of the Flint Ridge flint from the Harness site complex was probably the ridge proper or its envi- rons, and it was likely transported by human agency. Of interest in the investigation of flint raw material procurement strategies is the representation of varieties that would have been available locally as redeposited pebble material. This problem was approached in three ways: 1) each archaeological specimen was tabulated ac- cording to whether or not pebble cortex was present; 2) pebbles collected from a gravel bar along the Scioto River close to the Harness Mound were examined for identifia- ble varieties; and 3) non-artifactual chert pebbles from the Harness Mound excavation in broken and unbroken conditions were identified according to variety. Table 9.5 indicates the incidence of pebble chert in the archaeological collections. This form of raw material would have been available locally, and most or all of the pebble specimens were likely collected in the immediate vicinity of the Harness site complex. The present analysis would not have detected the acquisition of pebble raw material from distant locales, if such procurement had in- deed occurred. Debitage and artifacts lacking cortical remnants are ambiguous with respect to context of acqui- sition (in situ or redeposited). The lack of cortex may simply represent decorticated but locally acquired pebble chert rather than “exotic” raw material obtained directly or indirectly from outcrop locations at varying distances from the site complex. No more than 28% of the unidenti- fied flint retains pebble cortex, suggesting at least some decortication at the site(s) of acquisition. The impression remains that much unidentified “Local Pebble” and “Unknown” chert is of local origin, having been dislodged from distant outcrop areas to the north, east, and south and then transported by glaciers and an- cient and modern drainage systems. Such pebble chert presently occurs in gravel bars along rivers and in Wis- consin and Illinoian till and outwash in the site complex vicinity. Included with unidentified flint are 6 flakes with a thin, smooth, white-tan cortex and 10 decorticated flakes of identical raw material from Site 25 that is dark brown, homogeneous, very fine grained and occasionally pitted. A bladelet core in one of Harness’s surface collections may be of the same Hint variety. Although direct compar- isons were not made, this high-grade material seems to be the same as several bladelets and debitage apparently as- sociated with bladelet production excavated from the Hopewell site of Mound City in the Chillicothe vicinity and present among surface collected material in Clermont and Hamilton counties in southwestern Ohio. One speci- men from a workshop near the Turner site in Hamilton County was thin sectioned by Timothy S. Dalbey, who identified it with the Delaware Formation based in part on similarities between its photomicrograph and that from the Kuenzli Quarry at Delaware illustrated by Stout and Schoenlaub (1945: PI. III). The possible link between three “classic” Hopewell sites through the co-occurrence of this distinctive flint variety merits further investigation, particularly in view of its ap- parently exclusive use for bladelet production. If the Del- aware identification is correct, it is not necessarily exotic to the area. It is represented as unidentified because geo- logically documented comparative samples are not cur- rently available. Also from Site 25 are one pebble flake and three flakes without pebble cortex of identical, high-grade but uni- dentified material that is translucent and variegated (with reddish tinges as part of the coloration). These are likely exotic to at least the region from which most of the Har- ness site complex flint/chert was acquired, as is at least one flake in the excavated Harness Mound collection. A sample of 100 chert pebbles was randomly selected from among several hundred collected at a gravel bar near the Harness Mound during the summer of 1976. The re- sults of identifying the flint / chert varieties represented are given in Table 9.5, which shows Delaware present in abundance (46%), followed by Cedarville-Guelph (8%), Bisher (3%), and Columbus, Brassfield, and Upper Mer- cer (l%each). Forty percent were unidentifiable. Identifi- cations of non-artifactual chert pebbles excavated from the Harness Mound are also presented in Table 9.5. Apart from Flint Ridge and unidentifiable varieties, the most abundantly represented raw material is Dela- ware. Delaware dominates the excavated debitage and the non-bladelet artifacts and is prominently repre- sented among the debitage from Sites 1 8 and 25. This sug- gests that it was the most important raw material ex- ploited for tool production unassociated with the bladelet industry, although small amounts were exploited for the latter. Delaware, Columbus, and Cedarville-Guelph all out- crop in areas presently drained by the Scioto River and its tributaries north and northwest of the Harness site com- plex. It is likely that they were dislodged by the river and carried southward where they were then available locally as pebble material in the site vicinity. All three varieties were present in the pebbles collected along the Scioto River, with Delaware accounting for almost half of the sample. The non-cultural pebbles recovered from the TABLE 9.5 Frequency Distribution of Flint /Chert Varieties for Cultural Collection and Comparative Local Scioto River Gravel Pebbles ED WIN HA RNESS MOUND Artifacts Debitage Pebbles Artifacts ROBERT HARNESS Surface Collection Debitage Bladelets sites It 1 8 & Cores Q( & tt25 et TOTA LS FOR CUE TURA L MATERIALS % Total non-flint Subtotals Q* Total Ridge Scioto River Gravels (Pebbles) Flint Ridge 18 10 7 1,103 2 1,151 2,289 2 2,291** Flint Ridge (Pebble) 1 1 1 Delaware 2 3 1 18 2 7 i 31 3 34 7.8 Delaware (Pebble) 4 14 8 1 1* 2* 27 3 30 6.8 46 Brush Creek 2 7 1 1 4 3 15 3 18 4.1 Brush Creek (Pebble) 1 3 2 2 8 8 1.8 Zaleski 1 1 1 1 4 7 1 8 1.8 Zaleski (Pebble) 1 4 1* 5 1 6 1.4 Upper Mercer 2 1 4 1 6 2 8 1.8 Upper Mercer (Pebble) 1 1* 1 1 2 0.5 1 Upper Mercer or Zaleski 1 2 2 5 5 1.1 Upper Mercer or Zaleski (Pebble) 1 1 1 0.2 Boggs 5 5 5 1.1 Boggs (Pebble) 1 1 1 0.2 Upper Mercer or Boggs 1 1 1 0.2 Knife River 2 3 2 3 4 7 1.6 Harrison County 1 2 1 1 2 3 5 1.1 Cedarville-Guelph 1 1 2 2 0.5 Cedarville-Guelph (Pebble) 3 4 7 7 1.6 8 Columbus 1 1 1 0.2 Columbus (Pebble) 2 2 2 0.5 1 Prout (Pebble) 1 1 1 2 1 3 0.7 Brassfield (Pebble) 2 2 2 0.5 1 Bisher 1 1 1 2 1 3 0.7 Bisher (Pebble) 3 Plum Run (Pebble) 1 1 1 0.2 Kanawha Black (Pebble) 1 1 1 0.2 Unknown 4 14 8 36 111 173 173 39.5 Local Pebble 1 30 40 1 2 29 1* 103 1 104 23.7 40 Total 36 91 54 22 1,167 13 1,328 19 2,698 32 2,730 99.8 100 ♦Probably pebble; flint identification definite ♦♦Flint Ridge variety forms 83.9% of total identified cultural objects tQuestionable 1983 EDWIN HARNESS MOUND 83 Harness Mound excavation included both Delaware and Cedarville-Guelph. In the excavated debitage sample, the proportions of Delaware, Cedarville-Guelph, and Co- lumbus that are demonstrably pebble-derived are 82.4%, 75.0%, and 66.7%, respectively. These data indicate that predominantly pebble material was used for bifacial arti- fact manufacture, although the presence of bladelets and bladelet cores of Delaware in the surface collections may represent direct acquisition from outcrop in order to ex- ercise some selectivity regarding quality and perhaps size. Three flint varieties outcrop to the east and south of the Harness site complex, in the path of the Pleistocene Teays River and its tributaries. Brush Creek and Zaleski in Ohio and Kanawha Black in West Virginia were all subject to displacement by the westward and — in the site vicinity and immediately south of it — northward flowing Teays, which undoubtedly carried at least these three varieties to and through the area. The most important of these varie- ties for chipped stone tool manufacture was apparently Brush Creek. Like Delaware, Brush Creek was used for limited bladelet production, as suggested from the occur- rence of one Brush Creek bladelet each in the excavated sample and in Harness’s surface collection. However, it is more abundant in the form of debitage not necessarily associated with bladelet production and ranks second to Delaware among the identified varieties in the total sam- ple. Ranking third is Zaleski, which is the closest source of bedded flint to the Harness site complex. None of these three flint varieties is represented in the Scioto River pebble chert sample, but nearly half of the Zaleski flint in the total sample is demonstrably pebble material and the only specimen of Kanawha Black also retains pebble cortex. The presence of Brush Creek in the mound as unmodi- fied pebbles indicates that it was naturally occurring and locally available. It is represented by one checked pebble in the excavated debitage sample, suggesting that it was exploited in pebble form. Nevertheless, the lowest pro- portion of demonstrably pebble material of all the locally available varieties is Brush Creek. Only 30% of the exca- vated Brush Creek debitage and 26% of this variety in the total sample retains pebble cortex. This may represent the transportation of pebble material to the mound and vicin- ity in partially decorticated conditions, probably in con- junction with at least some direct acquisition of tabular material from the source area (perhaps destined for blade- let production). If such were the case, the procurement strategies for both Brush Creek and Delaware were very similar. Relatively close to the Harness site complex are source areas of Bisher and Brassfield to the west and Upper Mercer to the east. The presence of all of these varieties in the Scioto River pebble chert sample suggests that they were carried into the area naturally. The Harness surface collection contains a bladelet core of pebble Upper Mercer and a bladelet of questionably identified Upper Mercer in addition to two bladelets of either Upper Mercer or Zaleski. Thus, Upper Mercer is yet another flint variety that was used for bladelet production in the Harness Mound vicinity, although apparently in very small quantities. Even though available in the local area in pebble form, bedded material from Upper Mercer out- crops would have been accessible to groups traveling to or from Flint Ridge deposits yet farther north. Also present in small quantities are Bisher and Brass- field. One of the Harness surface collected bladelet cores is probably Bisher, but no bladelets of this variety were recognized. Bisher is noticeably grainier than such high- quality varieties as Harrison County, Knife River, Flint Ridge, and most Upper Mercer and Zaleski. Brassfield has many fossil inclusions that would have deflected shock waves traveling through a core, causing bladelets and flakes to detach unpredictably. These may have been prehistoric considerations in selecting predominantly other varieties for the bladelet industry. With the possible exception of one stemmed knife of either Upper Mercer or Zaleski, neither Upper Mercer nor Bisher is represented among the excavated artifacts and debitage, although both are present in the Scioto River pebble sample. Given their low density in this sam- ple and the small quantity of excavated debitage, how- ever, this lack of representation is probably due to sam- pling error. By the same token, the occurrence of Bisher, Upper Mercer, and Zaleski artifacts unrepresented by debitage in the excavated collection may also be due to sampling problems because the Harness surface material contains debitage of each. Brassfield, Columbus, and Upper Mercer or Boggs are represented by excavated deb- itage and no artifacts, but again the small amounts pres- ent suggest sampling error. Outcropping far to the north and present in the sample in such small quantities that intentional procurement seems remote are Plum Run (in northeastern Ohio) and Prout (in the extreme north-central part of the state). The mechanism or mechanisms by which these varieties were transported to south-central Ohio is unknown. Glacial action is plausible in the case of Prout but cannot account for the presence of Plum Run. Unless misidentified, this item may have been brought to the Harness site complex by visitors carrying indigenous items with them for trad- ing along the way. This may also account for the presence of one Plum Run specimen at nearby Mound City (Vickery 1983). Exotic to Ohio are Harrison County and Knife River flint, both of which are present in the Site 25 collection and, for Harrison County only, among Harness’s surface collected bladelets and bladelet cores. With its source area in extreme south-central Indiana, Harrison County flint was available ca. 200 miles away, while Knife River Hint, if obtained from deposits in west-central North Dakota, was at least 1,100 miles distant from the Harness site complex. Unless these varieties were brought here by vis- 84 KENT D. VICKERY No. 39 iting groups, their occurrence in south-central Ohio re- sulted from direct acquisition or long-distance trade. If the former, they were apparently imported in small amounts (perhaps each representing a single episode of acquisition) and likely specifically for bladelet, cache blade, and/or ceremonial spear point production. Summary of Flint/ Chert Raw Material Utilization In the vicinity of the Harness Mound, Flint Ridge flint was the most heavily relied upon raw material for bladelet production. Though some utilitarian items of Flint Ridge were manufactured — as exemplified by projectile points, a scraper, and unidentifiable biface fragments in the Site 18 and Site 25 surface collections — the main purpose of its importation seems to have been to maintain a rela- tively large supply of high quality raw material on hand for bladelet production. Some of this material may have been available locally, the Teays River perhaps having transported it northward. Nearly all of the Flint Ridge flint observed in the collections, however, was high-grade material, which, in combination with its general lack of pebble cortex, suggests that procurement was predomi- nantly or exclusively from bedded deposits, probably to the north and slightly east of the site complex. Its abun- dance and quality suggest that procurement was both sys- tematic and selective. Supplementing Flint Ridge for bladelet production were Harrison County and Knife River flint, along with small quantities of flint varieties available locally or a rel- atively short distance away (e.g., Delaware, Brush Creek, Upper Mercer). It is likely that some Delaware and Brush Creek were acquired from outcrop. The bladelet industry is represented in the excavated sample from the Harness Mound, but this collection mainly reflects bifacial reduction associated with the man- ufacture of utilitarian artifacts. There is a general correspondence between the debit- age, artifacts, and naturally occurring pebbles from the Harness Mound and the Scioto River pebble chert sam- ples. There is also a fairly even distribution of flint varie- ties throughout the reduction sequence (as represented by the various debitage types). This suggests that locally available pebble chert was exploited for non-bladelet and very limited bladelet manufacture. The territory of systematic flint and chert resource ex- ploitation was probably a linear, north-south one that as- sumed the configuration of an ellipse roughly 150 miles north-south by 80 miles east-west, with the Harness site complex located near the west-central periphery (see Fig. 9. 1). In addition, trade arrangements and/or forays were undertaken to acquire certain high-quality varieties from greater distances away. Visiting groups bringing indige- nous flint with them likely supplemented the raw material acquired in this manner. Acknowledgments I wish to thank James F. Murphy of Ohio State Univer- sity for his critical reading of the manuscript and helpful suggestions. Any errors or omissions remain the respon- sibility of the author. References Binford, Lewis R. 1963 A proposed attribute list for the description and clas- sification of projectile points. In Miscellaneous stud- ies in typology and classification, Anta M. White, Lew- is R. Binford, and Mark L. Papworth, pp. 193-221. University of Michigan, Museum of Anthropology, Anthropological Papers No. 19, Ann Arbor. Braun, David P., James B. Griffin, and Paul F. Titterington 1982 The Snyders mounds and five other mound groups in Calhoun County, Illinois. University of Michigan, Museum of Anthropology, Technical Reports No. 13; Research Reports in Archaeology Contribution 8, Ann Arbor. Carskadden, Jeff 1971 Upper Mercer flint quarries in Muskingum County, Ohio. Ohio Archaeologist 2 1 ( 1 ):3 1 5—3 1 6. Carskadden, Jeff, and Gerald Donaldson 1973 Brush Creek flint quarrying in Perry and Morgan counties, Ohio. Ohio Archaeologist 23( 1):20— 2 1 . Clayton, Lee, W. B. Bickley, Jr., and W. J. Stone 1970 Knife River flint. Plains Anthropologist 15(50): 282-290. Collett, John 1878 Geological report on Harrison and Crawford coun- ties, Indiana, 1878. Eighth, Ninth, and Tenth Annual Reports of the Geological Survey of Indiana, pp. 291-522, Indianapolis. Collins, Michael B., and Jason M. Fenwick 1974 Heat treating of chert: methods of interpretation and their application. Plains Anthropologist 19(64): 34-145. Converse, Robert N. 1972 Flints used by Ohio’s prehistoric Indians. Ohio Ar- chaeologist 22(2): 36-39. Crabtree, Don E. 1972 An introduction to flintworking. Occasional Papers of the Idaho State University Museum No. 28, Poca- tello. Crawford, Betty J. 1967 Famous Flint Ridge flint. Earth Science 20(3): 1 1 8 — 1 19. Fowke, Gerald 1894 Material for aboriginal stone implements. The Ar- chaeologist 2:328-335, Waterloo, Indiana. 1 902 Archaeological history of Ohio. Ohio State Archaeo- logical and Historical Society, Columbus. 1928 Archaeological investigations — II. Bureau of Ameri- can Ethnology, 44th Annual Report , pp. 399-540, Washington, D.C. Greber, N’omi, Richard S. Davis, and Ann S. DuFresne 1981 The micro component of the Ohio Hopewell lithic 1983 EDWIN HARNESS MOUND 85 technology: bladelets. Annals of the New York Acad- emy of Sciences 376:489-528. Guernsey, E. Y. 1937 Certain southern Indiana sources of lithic artifact material. Proceedings of the Indiana Academy of Science 46:47-52, Indianapolis. Hastings, Jerry 1969 Brassfield and Bisher Hints: their use by prehistoric Indians in Adams County, Ohio and adjacent areas. Ohio Archaeologist 19( 1 ): 1 1 . 1971 Nodular chert and Hint: their use by prehistoric Indi- ans in southern Ohio. Ohio Archaeologist 21(1):319. Holmes, William H. 1919 Handbook of aboriginal American antiquities: part I, introductory: the lithic industries. Bureau of Ameri- can Ethnology, Bulletin 60, Washington, D.C. Kramer, Leon 1953 Is there a jet black flint and may it be found in Ohio? Ohio Archaeologist 3(4):29— 3 1 . Krebs, Charles E. 1914 Kanawha County. West Virginia Geological Survey, County Reports , Wheeling. Lilly, Eli 1937 Prehistoric antiquities of Indiana. Indiana Historical Society, Indianapolis. Mills, William C. 1921 Flint Ridge. Ohio Archaeological and Historical Pub- lications 30:91-161, Columbus. Moorehead, Warren K. 1892 Primitive man in Ohio. G. P. Putnam’s Sons, New York. Morton, James, and Jeff Carskadden 1972 Aboriginal flint quarrying activities in the Muskin- gum County area. Ohio Archaeologist 22(2): 15-21. Murphy, James L., and John E. Blank 1970 The Plum Run flint quarries. Ohio Archaeologist 20(2): 198-199, 217. Olafson, Sigfus 1964 West Virginia flints used in Ohio. Ohio Archaeologist 14(2):37— 39. 1972 Some sources of lithic materials in West Virginia. West Virginia Archaeological Society, Newsletter 14(3):3— 9. Patterson, L. W. 1979 Additional data on thermal alteration of Flint Ridge materials. Ohio Archaeologist 29( 1 ):33— 34. Price, W. Armstrong 1921 Chert deposits of West Virginia. In Nicolas County, edited by David B. Reger, pp. 221-240. West Virginia Geological Survey, County Reports , Wheeling. Prufer, Olaf H., and Raymond S. Baby 1963 Paleo- Indians of Ohio. Ohio Historical Society, Co- lumbus. Seeman, Mark F. 1 975 The prehistoric chert quarries and workshops of Har- rison County, Indiana. Indiana Archaeological Bul- letin 1(3):47— 6 1 . Shaver, Robert H., et al. 1970 Compendium of rock-unit stratigraphy in Indiana. Indiana Department of Natural Resources, Geologi- cal Survey, Bulletin 43, Bloomington. Smith, Charles M. 1885 A sketch of Flint Ridge, Licking County, Ohio. Smithsonian Institution, Annual Report for 1884 , pp. 851-873, Washington, D.C. Stauffer, Clinton R. 1909 The Middle Devonian of Ohio. Geological Survey of Ohio, Fourth Series, Bulletin 10, Columbus. 1916 The relationships of the Olentangy shale and asso- ciated Devonian deposits of northern Ohio. Journal of Geology 24(5):476-487. Stothers, David M., and William Rutter 1978 Pipe Creek chert: newly discovered aboriginal quarry source. Ohio Archaeologist 28(3): 13. Stout, Wilbur, and R. A. Schoenlaub 1945 The occurrence of flint in Ohio. Ohio Department of Natural Resources, Division of Geological Survey, Fourth Series, Bulletin 46, Columbus. Stumm, Erwin C. 1942 Fauna and stratigraphic relations of the Prout lime- stone and Plum Brook shale of northern Ohio. Jour- nal of Paleontology 1 6(5):549— 563. Tomak, Curtis H. 1980 Schershel: a Late Archaic occupation in southern In- diana with appended chert descriptions. Central States Archaeological Journal 27(3): 104-1 1 1. 1982 Selected Indiana chert types. Ohio Archaeologist 32( 1 ):36— 39. Vickery, Kent D. 1983 Harness and Mound City: a flint raw material com- parison. Paper presented at the 48th Annual Meeting of the Society for American Archaeology, Pittsburgh, Pennsylvania, April 27-30. Vickery, Kent D., and Lois J. Lambert 1977 Analysis of chert debitage sample from the Nashport mound, 33Mul5, Muskingum County, Ohio. Appen- dix F to Report of the 1975 Excavations at the Nash- port mound (33Mul5), Dillon Lake, Ohio, N’omi Greber, pp. 87-109. National Park Service report, contracts cx-6000-5-0186 and cx-5000-6-0301. White, Anta M. 1963 Analytic description of the chipped-stone industry from Snyders site, Calhoun County, Illinois. In Mis- cellaneous Studies in Typology and Classification, edited by Anta M. White, Lewis R. Binford, and MarkL. Papworth,pp. 1-70. University of Michigan, Museum of Anthropology, Anthropological Papers No. 19, Ann Arbor. White, I. C. 1903 Coal Report, Vol. 2. West Virginia Geological Sur- vey, Reports , Wheeling. 1908 Supplementary Coal Report, Vol. 2(A). West Virginia Geological Survey, Reports , Wheeling. Wilmsen, Edwin N. 1970 Lithic analysis and cultural inference: a Paleo-Indian case. University of Arizona. Anthropological Papers No. 16, University of Arizona Press, Tucson. 10 DISCUSSION AND CONCLUSIONS Environment and Subsistence The botanical, faunal, molluscan, and soils analyses are all consistent. The environment of the Central Scioto when the Harness Big House was used and first covered over was very similar to that found in the area in the eighteenth century, a forest cover with some relatively open areas and pocket prairies. Within 3 km of the site on Gordon’s (1966) map of vegetation at contact are bottom- land hardwoods, oak-sugar maple, oak savannas, and mixed mesophytic forests. Although the debris found at the site reflects specialized and public activities rather than everyday subsistence, it does contain evidence for the use of the several parts of the relatively local environments: scale fish, shell fish, and building materials from the river and the hills; deer, raccoon, turkey and other birds, and small mammals from the woodland and edge areas. The evidence of plant foods comes both from skeletal analysis and botanical samples. Skeletal analysis (Bender, Baerreis, and Steventon 1981) demonstrates that, on the basis of present evidence, corn became a regular part of the diet of humans and of deer after the Middle Woodland time period in the sites tested from Illinois, Ohio, and Wisconsin. Three individuals from Edwin Harness and five from the large Seip Mound were included in the study. The skeletal evidence from Edwin Harness is consistent with the amount of corn recovered. The Zea mays found at Harness was in contexts in which it can easily be interpreted as being a special purpose plant in the same sense as tobacco or other known ceremonially important plants. This is demonstrated within Feature 60 where fragments of Zea mays were found and the ratio of Carbon- 13 to Carbon- 12 in the bones of the individual interred was -22.7%. This value of the ratio is interpreted as indicating no significant amount of corn in the diet (ibid.). Wild fruits and seeds were also identified in the Harness botanical samples analyzed. Of particular interest are the small unidentifiable seeds which composed high percent- ages of some of the notation samples. The unavoidably biased information from Harness does fit the best present estimate of the subsistence pattern of Scioto Hopewell peoples: gathering, hunting, and gardening (Ford 1979). Additional work needs to be done, particularly with respect to the position within the economy of the locally available plants which bear starchy seeds. Relative Intrasite Chronology One of the primary tasks of our excavation was to seek stratigraphic or other physical evidence for rebuilding or multiple building stages of the structure at the base of the mound. All the evidence which we found supports a single stage of both construction and use of the major structure. This is to be distinguished from the possible use through time of the knoll itself and the various building stages of the mound which were placed over the building. There are two somewhat different sequences of con- struction which can be interpreted from the known data. They center on the time lapse between the preparation and use of the sub-main floor area. Feature 33, and its underlying base. Feature 50. No building posts were found that originated on Feature 33; it appears that only activities which would result in few if any subfloor remains took place on the surface of this mixed clay floor. This surface was used, if at all, before the Big House was completed. Large areas of relatively featureless “Floor” which had been separately covered by primary mounds were recorded by Shetrone at either end of Hopewell Mound 25 (Shetrone 1926). Within the Seip Earthworks in Locality 23, which is immediately west of the area excavated by the Ohio Historical Society 1971- 1977 (Baby and Langlois 1979), we have uncovered a plaza-like area which was also relatively featureless and had been separately covered (Greber and D. Griffin 1982). There were undoubtably other such places in central Ohio. Thus the existence and use of such a space is not unusual. Feature 33 underlies the East Section and extends approximately 13 m north and 10 m west of this section. This is an area of particular interest because of the somewhat anomolous character of the burial population associated with the corresponding area in the Seip Big House. At present we cannot identify the nature of the population associated with the East Section at Harness. Hopefully further work will aid in deciding whether the location of Feature 33 with respect to the East Section is of some significance. With respect to possible construction phases of the major structure, we did not find any evidence which would indicate that the surface of the knoll had been cleared in stages. Feature 3C, which underlay the entire structure, was the same everywhere it was found. It is possible that evidence of differences existed originally in the heavily disturbed areas, and that the area under Features 50 and 33 had been cleared first, the remainder later. However, in addition to the physical characteristics of Feature 3C, there is a design feature which makes me tend not to accept two separate clearings. The heavy gravel outer mantle had been placed essentially at the edges of Feature 3C, which would not have been visible when the wall was placed. This and the repetition of the pattern at Seip do appear to indicate there was some basic preconceived design for the total complex. 1983 EDWIN HARNESS MOUND 87 Mills (1907:137) states that the posts were placed and the floor built around them. We did not have the opportunity to see the join of Feature 3, a puddled clay floor, and pristine building posts. However, the character of this main upper floor and the constructions of tomb remains and outer posts which we did find is consistent with Mills’s observation. I do think it would have been possible to have constructed post holes through the softer clays of Features 33 and 50. In sum, the surface of the knoll could have been partially cleared, a special purpose ceremonial area constructed, used, and left. Later, additional clearing and cleaning of the first area, placement of major structural posts (perhaps beginning construction on the east), and laying of the main floor of the complete house followed. Or, what appears more likely, the complete knoll was cleared, and a special area made, perhaps with materials of special significance from previous ceremonies. When the appropriate ceremonies were completed, the major structural posts were placed, the main floor layers put down, and the floor features constructed, all in a rela- tively short span of time. In considering this second as more likely, I am regarding the radiocarbon assay DIC- 662 as a product of old trees and/or chance. The Pattern of the Big House The basic plan of the Big House, three nearly rectan- gular and one circular section, is in contrast with the design of the earthwork which contains three sections that are parts of circles and one section that is square. Out of both designs come the deliberate use of 3, 4, 7, and 48 (4 X 12) construction elements. Within the Big House itself the two larger sections, which are similar in outline, differ in the stylistic implementation of the basic design. The two smaller sections differ in basic outline. Color appears to have been used to distinguish structural posts on perim- eters or entranceways in the three smaller sections and the middle hallways. In the North Section three structural posts in the southeast corner and one immediately west of the geometric center were also marked with red stains. The non-structural posts were distributed differently among the sections. The North and Middle Sections each had an irregular line just west of the building proper. Other small, apparently colored posts were immediately north of the North Section. A cluster of small stained posts was on the east side of the Middle Section, while a cluster of medium and small stained posts was on and immediately west of the center of the North Section. In contrast, at the center of the Middle Section was a long, burning fire. There is also a contrast between the usual small scale (total content) of the deposits found in many places on the floors within and about the Big House and those which were found by earlier excavators about the center of the North Section. Comparison of the placement of these deposits with respect to the central focus posts suggests an east-west division in the design of the center. Thus, there are differences in scale, design, and activity remains within the four sections. These differences indi- cate a likely difference in style if not in actual content of some of the activities which took place in each of the four sections. I have concluded, based on my analysis of burial attributes, that each section of the Big House was the social space of a sub-group within the total society which supported the Big House (Greber 1979). The data given here is separate from that of individual burial attributes. This data, on the design of the complex, is consistent with my additional conclusion that each sub-group represented in the Big House had separate social responsibilities within the whole society. The Pattern at Seip The Harness building design shown in Figure 3.2 was placed over the map of tombs and floor features found at the base of the large Seip Mound (see Greber 1 979: Fig. 1A). It was necessary to rotate the Middle Section 90°. The results of the superposition are given in Figure 10. 1. 1 consider this an excellent fit. This new estimated map suggests a refinement of the original groupings which I had used in studying the social sub-groupings associated with this Seip mound. The section which corresponds to the Harness East Section groups together a small number of individuals who have stood out among the general population associated with Lobe 1 (West Section) (Gre- ber 1976:76). Using the ranksums calculated for the individuals whose tombs were found on the floor of the Seip Big House (Greber 1976:53, 237-253), I calculated new median ranksums using now four separate units. The small group of six individuals in the North Section of the Seip Big House has the highest median ranksum but with a large confidence interval (see Table 10.1). The field notes of the Ohio Historical Society excavations at the large Seip Mound stated very clearly that the primary mound, which covered the largest section (Lobe 1), also covered this small north extension (Shetrone et al. 1925-1928). If this were strictly true, then perhaps this high ranking group has a special relationship with the major group within the Lobe. Such a small group would have a different social character from that of the larger groups associated with the other sections of the Big House. A review of the descriptions of the non-perishable artifacts found with the remains of these individuals suggests that, as a group, these individuals may have been associated with special rituals (Greber 1976:Individuals 37, 39, 44, 45, 46, 51). With the remains of Individual 39 were a miniature copper plate and earspools and four cop- per covered stone buttons. Above the roof of the tomb pieces of a small pottery human head effigy were found. This tomb was unusual in construction. It also had been covered by a primary with two gravel strata. The primary RTH ^ O o "O "O C c c cd g 2 o. - -S £ _ 3- • 3 00 ^ O S 00 S o. E o. o C V £ x \S c C3 E T3 C 3 0J o ■ 2f S o It o c I Harness Middle Structure has been rotated 90°. 1983 EDWIN HARNESS MOUND 89 TABLE 10.1 Values of Ranksum F for Seip Mound 1 (Pricer) (Based on calculations in Greber 1976) House Section No. of Individuals Median Ranksum 97% Confidence Interval Confidence Interval Length Highest Value Lowest Value West (Lobe 1) 41 607 554 to 667 1 13.0 371.0 743.5 North (Lobe 1) 6 555 412 to 685 273.0 41 1.5 685.0 Middle (Lobe 2) 37 656 628 to 689.5 61.5 440.0 750.5 East (Lobe 3) 19 698 643 to 738.5 95.5 297.5 747.5 is described as covering two tombs, but which two is not clear. If it is that which Shetrone and Greenman called Grave 34, then the artifacts found with Individual 37 in this grave complement those of Individual 39. A diminua- tive copper crescent along with four pearl beads, one copper hemisphere, and two copper objects, called “pos- sible effigies of the praying mantis” by Shetrone (Shetrone and Greenman 1931:407), were found. The latter may be metal working tools. The multiple burial (Individuals 44, 45, 46) was covered by a large primary mound with two sand strata, an unusual construction. Associated with Individual 44 were two copper celts; with Individual 45, approximately 20 pounds of galena, pearl beads, burned fabric, a copper celt, and pulley-type stone earspools; with Individual 46, nothing. Deposited with the bones of Individual 5 1 were a small, “fine” ocean shell container, a copper plate, earspools, and 50 fine pearl beads. Within the charcoal platform on which the bones had been deposited, three hollow 1-inch (2.54-cm) diameter copper hemispheres were found. I suggest that the miniature items, the copper hemispheres, the conch shell, and the galena may have been primarily used in rituals. The four sections of the Seip Big House may have, at one time, been reflected in the design of the earthwork walls at Seip. There are some variations in the design of the five major earthwork complexes which I have con- sidered as related: Seip and Baum in the Paint Creek Valley, Frankfort on North Fork, and Works East and Liberty in the main Scioto Valley. The most nearly complete maps for all these come from Squier and Davis ( 1 848: Pis. XX, XXI). All five of these earthworks include complete squares, if we assume the square at Works East was intact before erosion by the lateral movement of the Scioto. All the complexes also contain one large incom- plete circle. The next smaller circle is relatively complete except at Baum, which has an amorphous area joining two partial circles. There is an additional smaller partial circle at Works East and Liberty. An amorphous wall joins the small circle and the square at Seip. There were changes in building design and apparent social groupings at Seip as seen in the structures under the large mound and the smaller conjoined mound (Greber 1979, 1976). Changes appear to have been made in the earthwork walls. Perhaps these two changes are related. The order of the building of the walls and the significance of the design of the parts is currently being studied (Greber and Jargiello 1982). Based on types of variation in designs, the earthworks can be grouped by river valleys. On North Fork the design contains a square and two nearly complete circles; in the Paint Creek Valley each has a square, two circular parts with less geometrically identi- fiable joining walls; and in the Scioto Valley the designs contain a square and three circular elements. Absolute Chronology The radiocarbon dates presented in Table 3.2 are plotted in Figure 10.2 with comparative dates from the Russell Brown Mounds, which are part of the Liberty Earthwork Complex (Seeman and Soday 1980); from the McGraw site, which is on the west bank of the Scioto 8.5 km north of Liberty (Prufer 1968; Prufer et al. 1965); and from the Seip Earthwork Complex (Baby and Langlois 1979; Greber and D. Griffin 1982). Considering first the dates from the base of the Edwin Harness Mound, we can group several of these into reasonable stratigraphic units: the four dates associated with the Big House itself (Features 19, 30, and P. M. 32); those from areas outside the house but on the main, heavy puddled clay floor ( Features 1 7, 3 1 , 62); and the single one from the northern activity area (Feature 53A). Since the radiocarbon years represent a statistical mean, a group average mean can be calculated using the standard deviation associated with each mean as its weighting factor. The weighted average mean for the use of the Big House is 1641 ± 32 radiocarbon years b.p. (a.d. 309 ± 32). The outer areas average to 1619 ± 35 radiocarbon years b.p. (a.d. 331 ± 35); and the single date for the north area is a.d. 450 ± 50. There is of course no justification for computing such averages if the features associated with the radiocarbon assays are not judged to represent a reasonably discrete single cultural time. If all activities on the heavy clay floor are assumed to be culturally contemporaneous, the corresponding weighted mean is 1632 ± 24 radiocarbon years b.p. (a.d. 318 ± 24). It should be noted that the 90 N’OMI GREBER No. 39 Fig. 10.2. Radiocarbon dates plotted as mean calendrical years with one and two stan- dard deviations noted. (See Key to Fig. 10.2 next page.) apparent increase in precision of this latter date is the result of averaging dates judged to represent contempo- raneous cultural events. If this assumption is accepted, the odds are better than 2 to 1 that the actual date represented by this weighted mean is between a.d. 294 and a.d. 342. Additional dates will be added to the sample as dating techniques and resources allow. This further work may indicate whether or not the shift in weighted means for the floor groupings is a coincidence. Based on the present sample size the best estimate for the time interval which encompassed the construction and use of the area at the base of the Edwin Harness Mound is centered around a.d. 300. Unfortunately there are not enough dates to judge the time which intervened between the use of the floor and the final capping of the top of the mound. The dates obtained from the outer strata placed against the lower edges of the mound do appear to indicate that the site, as part of the earthwork complex, was used for many years. Seeman’s analysis of the available radiocarbon dates from the Russell Brown Mounds (Seeman and Soday 1980) places the use of these mounds probably before (Mounds 1 and 2) and after (Mound 3) the use of the main Edwin Harness floor. The sample of dates from these smaller mounds is limited, but the use of the earthwork area over such a time span is quite reasonable. A number of interesting questions are posed both by the dates obtained from the outer features at Edwin Harness and by the type of materials being placed against the mound. The dark, rock laden soils in Features 69 and 69A do not correspond to any other major mound strata found in the above four mounds; however, they do appear to resemble feature fill 1983 EDWIN HARNESS MOUND 91 KEY Figure 10.2 Site Lab. No. Provenience Calendrical Year Harness a D1C-662 Under Feature 3 200 b.c. ± 155 a' DIC-662 Under Feature 3 30 b.c. ± 155 b DIC-665 PM 32 a.d. 130± 70 c DIC-663 Feature 19 a.d. 330 ± 65 d DIC-664 rerun Feature 30 a.d. 350 ± 65 e DIC-664 Feature 30 a.d. 450 ± 60 f DIC-802 Feature 3 1 a.d. 320 ± 70 g DIC-661 Feature 17 a.d. 460 ± 65 h DIC-860 Feature 53A a.d. 450 ± 50 i D1C-1187 Feature 62 a.d. 180 ± 50 J D1C- 1635 Feature 56 A.D. 750 + 65 k DIC-1 188 Feature 81 a.d. 810 ± 60 1 DIC-1 190 Feature 55 a.d. 840 ± 50 Russell Brown m UCLA-244B Mound 1, Burned zone 140 b.c. ± 70 n UCLA-244A Mound 1, Feature 210 a.d. 200 ± 80 o UCLA-245 Mound 2, Feature 73 a.d. 90 ± 90 P UCLA-246C Mound 3, Feature 13 a.d. 430 ± 90 q UCLA-246B Mound 3, Feature 13 a.d. 590 ± 70 r UCLA-246A Mound 3, Feature 13 a.d. 615 ± 70 McGraw s UCLA-685 Excavation unit B-l 230 b.c. ± 80 t UCLA-679A Excavation unit D-l a.d. 140 ± 80 u UCLA-679B Excavation unit D-l a.d. 190 ± 80 V M-1558 Excavation unit D a.d. 230 ± 140 w UCLA-688 Excavation unit B-l a.d. 280 ± 80 x OWU-62 Excavation unit D-2 a.d. 435 ± 166 y UCLA-679C Excavation unit C-l a.d. 440 ± 80 z OWU-61 Excavation unit C-l a.d. 481 ± 65 Seip aa UCLA-292 Mound 1 (Pricer) a.d. 55 ± 100 bb D1C- 1725 Locality 20, Feature 2 a.d. 350 ± 45 cc DIC- 1724 Locality 20, Feature 4 a.d. 470 ± 55 dd DAL-280 House 7, Feature 5 a.d. 90 ± 85 ee DAL-116 Unit D, Midden Layer, Houses 1-3 a.d. 230 ± 80 ff DIC-28I House 4, Feature 7 a.d. 590 ± 105 gg D1C-289 Unit F, External post. House 4 a.d. 350 ± 60 hh DAL-281 House 5, Feature 3 a.d. 590 ± 105 ii DAL-282 House 6, Feature 1 3 a.d. 1055 ± 1 10 from two of the smaller mounds. In Russell Brown Mound 1, one or possibly two out of ten pit features contained dark soil and burned rock, while there were five out of seven pits with such fill in Mound 3. These soils may represent a change in type or in intensity of activities which were carried out in the earthwork area at this later time. Considering the overall archaeological evidence, Prufer has placed the occupation of the McGraw site in the fifth century (Prufer et al. 1965: 106). Accepting his interpreta- tion of the site as a single cultural event, and using all the dates except that on bone, which was an experimental date (Prufer et al. 1965: 104), we can compute a weighted mean for the site of a.d. 318 ± 32. This is well within the range of dates from the Liberty Complex and close to the estimate for the use of the main floor at Edwin Harness. As is shown in Figure 1 0.2, the range of available dates from the Seip Earthworks overlaps that of Liberty. Unfortunately there are no series of dates from the same context within Seip, so relationships among the various parts of the complex are more difficult to evaluate. Two dates were not available when Figure 10.2 was prepared. The first date, a.d. 430 ± 70 (DIC-2471), was obtained from charcoal left from a small fire built on the surface of a plaza-like area in Seip Locality 23 (Greber and D. Griffin 1982) which is immediately west of the Ohio 92 N’OMI GREBER No. 39 Elistorical Society Houses 1-7 (Baby and Langlois 1979). While it was burning, this fire was apparently covered with pea gravels similar to those in Harness Feature 41. These gravels were the first of at least two covers over the entire plaza. A second date, a.d. 330 ± 40 (DIC-2473), has been obtained from the floor of the Seip Big House from charcoal mixed with the cremated bones in Shetrone and Greenman Burial 32. The assay on bone from this same feature is unfortunately unacceptable (a.d. 670 ± 55, D1C-2472). The a.d. 330 mean value (likely range a.d. 290 to a.d. 370) is relatively close to the dates from the floor of the Harness Big House. More dates are needed to corroborate the series. Two dates, a.d. 470± 55 (DIC-1724) and a.d. 350± 45 (DIC-1725) have been obtained from charcoal found in a pit and in a midden-type deposit of burned rock, mica, colored clays, and other materials in Locality 20 which is just west of the earthwork wall extending south from the small circle (Greber and D. Griffin 1982). Sherds which have been identified as Connistee material (Bennie Keel, Chief Consulting Archaeologist, U.S. Dept, of Interior, personal communication 1980; Roy Dickens, University of North Carolina, Chapel Hill; and Jefferson Chapman, University of Tennessee, Knoxsville) were found in the general vicinity of the dated charcoal. In the dates from the Ohio Historical Society units, the assays for Houses 4 and 5 do not follow the stratigraphy as reported by Baby and Langlois (1979). There does appear to be some time range in the use of the seven structures but it is not yet possible to associate specific events in various parts of the earthwork with each other. The time span of use does overlap with that of the Liberty Works; however, more dates, hopefully from a variety of absolute dating techniques, are needed. There is a growing data base of absolute dates which demonstrates, consis- tent with the archaeological evidence itself, that these earthworks are indeed complexes of sites. No single chronological date can be used to represent “Seip” or “Harness”; therefore, discussions of intersite relationships should specify the specific sites within the complex which are pertinent to the points being discussed. Epilogue The Harness Big House was a place full of symbols. We can identify at least some of these symbols; however, we may not be able to interpret their intended meaning. It is very tempting to interpret the data we have found using details from ethnographies of known Eastern North American peoples who are separated by more than a thousand years from Hopewell peoples. In such interpre- tations we are looking for elements which are basic parts of an Eastern Woodlands culture and which could have stability through time; that is, we are looking for a proto- culture as a linguist seeks a proto-language. Yet we wish this exercise to yield something that is not so general that it is uninformative. This search must be done with addi- tional cautions, since, as is well documented, the meanings of a symbol vary not only through time, but within the same time and even within a given culture at a given time. What we can be confident of is the existence of a symbol and a class to which the symbol belongs. The recent salvage work at Edwin Harness has added to our knowledge of Hopewell symbols. The classes we have found include numbers, directions, colors, shapes, opposi- tion or binary contrasts, special trees and plants, and special uses of fire or smoke. These symbols reflect a system of thought and a way of life. One can find examples of the meaning of each of these within historic Eastern Woodlands peoples in their range of activities; gathering, hunting, gardening, curing, giving birth, giving names, marrying, praying, trading, achieving great deeds, and dying. They are part of the oral traditions which explain the origins of the world and of important cultural elements (e.g., Callendar 1978; Swanton 1946). Structural remains which have been found at the base of Hopewell mounds have at times been called “Charnel Houses.” I would consider the use of such a name for the Big House inappropriate. It should not be assumed that all evidence of activities found in such contexts must be interpreted as funeral. Even within modern cultures there are examples of the juxtaposition or superposition of crypts and/or tombs with areas which are used for a variety of civic, ceremonial, and religious purposes, as colonial churches which have burials below the main floor or in adjacent yards. Describing only the funeral aspects can be misleading and then centrality of the Big House within the life of these Hopewell people is lost. We have not yet been fortunate enough to find the house of a family; at least we have not recognized any structure as such. I expect that such a house will lack much of the symbolic detail which can be found in a major civic- ceremonial center because it is the center which is the appropriate place to show and care for the symbolic life of a people. References Baby, Raymond S., and Suzanne M. Langlois 1979 Seip Mound State Memorial: nonmortuary aspects of Hopewell. In Hopewell archaeology: the Chillicothe conference , edited by David S. Brose and N’omi Greber, pp. 16-18. Kent State University Press, Kent, Ohio. Bender, Margaret M., David A. Baerreis, and Raymond L. Steventon 1981 Further light on carbon isotopes and Hopewell agri- culture. American Antiquity 46(2):346-353. Callender, Charles 1978 Shawnee. In Handbook of North American Indians (Vol. 15), Northeast. Smithsonian Institution, Wash- ington, D.C. 1983 EDWIN HARNESS MOUND 93 Ford, Richard I. 1979 Gathering and gardening: trends and consequences of subsistence strategies. In Hopewell archaeology: the Chillicothe conference, edited by David S. Brose and N’omi Greber, pp. 234-238. Kent State University Press, Kent, Ohio. Gordon, Robert B. 1966 Natural vegetation of Ohio at the time of the earliest land surveys. The Ohio Biological Survey, Columbus, Ohio. Greber, N’omi 1976 Within Ohio Hopewell: analysis of burial patterns from several classic sites. Ph.D. dissertation. Depart- ment of Anthropology, Case Western Reserve Univer- sity, University Microfilms, Ann Arbor, Michigan. 1979 A comparative study of site morphology and burial patterns at Edwin Harness mound and Seip mounds 1 and 2. In Hopewell archaeology: the Chillicothe conference , edited by David S. Brose and N’omi Greber, pp. 27-38. Kent State University Press, Kent, Ohio. Greber, N’omi, and Dennis P. Griffin 1982 Comparison of excavations and subsurface remote sensing data from sections of the Seip earthworks complex, Ross County, Ohio. Paper presented at the 1982 Annual Meeting of the Southeastern Archaeo- logical Conference, Memphis, Tennessee, October 28-30. Greber, N’omi, and David Jargiello 1 982 Possible astronomical orientations used in construct- ing some Scioto Hopewell earthwork walls. Paper presented at the Annual Meeting of the Midwest Archaeological Conference, Cleveland, Ohio, Octo- ber 1-3. Mills, William C. 1907 Explorations of the Edwin Harness mound. Ohio Archaeological and Historical Quarterly 16: 1 13-193. Prufer, Olaf 1968 Ohio Hopewell ceramics: an analysis of the extant collections. Anthropological Papers No. 33. Museum of Anthropology, University of Michigan, Ann Arbor. Prufer, Olaf H., D. H. McKenzie, O. Pi-Sunyer, H. C. Cutler, R. A. Yarnell, P. W. Parmalee, and D. H. Stansbery 1965 The McGraw site: a study in Hopewellian dynamics. Cleveland Museum of Natural History, Scientific Publications (n.s.), 4(1). Seeman, Mark F., and Frank Soday 1980 The Russell Brown mounds: three Hopewell mounds in Ross County, Ohio. Midcontinental Journal of Archaeology 5( 1 ) : 7 3— 1 16. Shetrone, Henry C. 1926 Exploration of the Hopewell group of prehistoric earthworks. Ohio Archaeological and Historical Quar- terly 35: 1-227. Shetrone, Henry, Frank Setzler, and Robert Goslin 1925- Field notes of Ohio State Museum Archaeological 1928 Expedition, Seip mound 1. On file. Department of Archaeology, Ohio Historical Center, Columbus, Ohio. Shetrone, Henry Clyde, and Emerson F. Greenman 1931 Explorations of the Seip group of prehistoric earth- works. Ohio Archaeological and Historical Quarterly 40:343-509. Squier, George Ephram, and E. H. Davis 1 848 Ancient monuments of the Mississippi Valley. Smith- sonian Contributions to Knowledge 1. Washington, D.C. Reprinted 1973 with introduction by James B. Griffin as Antiquities of the new world: early explora- tions in archaeology (Vol. 2). A.M.S. Press, New York, for Peabody Museum, Harvard University. Swanton, John R. 1 946 The Indians of the southeastern U nited States. Bureau of American Ethnology, Bulletin 137. ■ CONTENTS No. 39. Recent Excavations at the Edwin Harness Mound, Liberty Works, Ross County, Ohio— N’omi Greber KIRTLANDIA CLEVELAND, OHIO NUMBER 40 •NATURAL HISTORY* KIRTLANDIA David S. Brose, Editor Kirtlandia is an occasional publication of The Cleveland Museum of Natural History and is devoted to scientific papers in the various fields of inquiry within the Museum’s sphere of interest Published at least twice a year, issues will vary between collections of short papers and single issue-length studies Kirtlandia is distributed by The Kent State University Press, Kent, Ohio 44242, Copyright © 1983 by The Cleveland Museum of Natural History, Kirtlandia is abstracted in Zoological Record and Biological Abstracts and indexed in Bibliography and Index of Geology. ISSN: 0075-6245 KIRTLANDIA THE CLEVELAND MUSEUM OF NATURAL HISTORY CLEVELAND. OHIO SEPTEMBER 1983 NUMBER 40 A STUDY OF THE EASTERN BLUEBIRD AT THE HOLDEN ARBORETUM, LAKE COUNTY, OHIO JEAN EAKIN Abstract A study of the nesting of the Eastern Bluebird at the Holden Arboretum in Lake County. Ohio, was conducted from 1965 through 1980. Emphasis was placed on causes of nest failure and methods used tocontrol them The 16-vear volunteer effort contributed to an increase in the Eastern Bluebird population at the Arboretum. Some information on the nesting ofT ree Swallows wasalso gathered during the study. Kirtlandia No. 40 0075-6245/83/1983-0040 $2.00 Copyright c 1983 by The Cleveland Museum of Natural History Fig. 1: Holden Arboretum in August 1980 Introduction In 1965, a volunteer effort was begun to increase the population of the Eastern Bluebird ( Siala sialis) at the Holden Arboretum (hereafter referred to as the Arboretum), in Kirtland and Mentor Townships, Lake County. Ohio. A previous effort had been made at this location to provide additional nesting sites for bluebirds by a local volunteer, Paul Smith, who built and placed 230 bluebird boxes on the property. Only eight pairs of bluebirds, however, nested in this area in 1965. At that time. House Wrens ( Troglo- dytes aedon) occupied 49 boxes and House Sparrows ( Passer domestieus) occupied 134. The boxes, located in 14 areas either in meadows or plantings along hedgerows or woods, were unpainted with side openings and were placed approximately 4 to 5 ft above the ground on metal fence posts. ( A few were placed 6 ft high; one was only 2 ft above the ground.) Project Bluebird was begun in 1965 by Virginia Barrus, a local ornithol- ogist and bird bander, as a volunteer program to study the Arboretum bluebird nesting population in an effort to maintain and, if possible, in- crease it. Prior to the 1 965 nesting season she cleaned and repaired the 230 original boxes. Although there have been several reports of such efforts to increase bluebird populations (Kibler, 1969; Laskey, 1939; Pinkowski, 1979; Musselman, 1935) none covered such a long period of time, 16 years (1965-80), as in this study. Because of a significant number of Tree Swallows ( Iridoprocne Bicolor) nesting in the boxes, they have been included in this study. The Study Area The Holden Arboretum consists of 5 sq mi (approximately 2,800 acres) of rolling land in Lake and Geauga Counties (lat 4 1°37'N; long 81° !9'W). Elevations (excluding the highest point. Little Mountain) range from 750 to 1,200 ft. (The areas included in this study range from 900 to 1,100 ft.) 1 he area is drained by Stebbins Gulch and Pierson Creek, both of which empty into the East Branch of the Chagrin River which runs through the property. The Arboretum is approximately 27 mi northeast of Cleveland and 2 mi south of Lake Erie. As the accompanying map ( Fig. 4) shows, the area is highly irregular in outline due to private ownership of adjoining properties. Although most of the Arboretum’s properties are contiguous, one large area and several small ones are separate. SU3aWflN YEARS Fig. 3 Tree Swallows Nesting Success 1965-1980 - o) r oi w 2 « 2 O) 4) 0)0) □ □ □ □ I I f! ° O 'NX ol o °l 1 ,j7 S 0 1 ^ I o ! \ 1 \ >4fc ^ .L^ v* |0o\ «>n) £ /^AV 1°^ X _CN SHAMAN YEARS 6 JEAN EAKIN No. 40 Fig. 4: Eastern Bluebird nest box areas at Holden Arboretum. Areas active in 1980: T — Thayer Center; CL — Corning Lake; HP — Heath Pond; CA — Chapin Acres; BA Bicknell Acres; N — Norweb; GP — Goose Pond; PM — Paul Martin Farm; P Pines; DMA- Upper Mather Acres; LBF — Lower Baldwin Farm; UBF — Upper Baldwin Farm. Areas discontinued during the study period were: CB — Cooper Building; I — Island in Corning Lake; PA — Park Acres; AC — Around Crabapples; AB— Around Blueberry Pond; AL — Around the Lilacs; LMA — Lower Mather Acres. 1983 EASTERN BLUEBIRD 7 At least 1,700 acres are covered by woodland —coniferous, mixed, and deciduous woods, with the latter predominating. There are 18 lakes and ponds, the largest of which is Corning Lake, approximately 20 acres. The areas included in this study are located either in nurseries or plant- ings with sufficient open space to meet the needs of bluebirds, or in un- developed fields. On the map each area is defined by the point at which field meets roadway, woodland, or. in a few cases, an extension of the field which was unsuitable because of competition from other species or heavy human use. Areas which had boxes at the start of the project but which have been phased out, as well as the present! 1980) placement of boxes, are shown. The following is a detailed account of each study area. Former Areas Cooper Building (CB). The area in the immediate vicinity of the admin- istrative building contained 10 boxes in 1965, all of which were inhabited by House Sparrows. We removed these prior to the 1967 season. Island in Corning Lake (1). This area had 10 boxes in 1965. All boxes were removed after the 1967 season due to inaccessibility for monitoring and occupation by House Sparrows. Park Acres (PA). An area on the southeast corner of Sperry and Kirt land -Chard on Roads contained 10 boxes in 1965 inhabited by House Sparrows. They were removed prior to the 1967 season. Around Blueberry Pond( AB). Most of thisarea iscovered by deciduous woodland and mature coniferous plantings. Seventeen boxes were placed in small open spaces. There was one successful bluebird nesting in 1967. In 1975 and 1976, bluebird attempts were interrupted by predators and wrens. Otherwise there was no bluebird activity. Black-capped Chickadees ( Parus atricapillus) successful in 1967 and 1968, were interrupted by wrens or human beings in subsequent attempts. There are several heavily used trails. Boxes were removed gradually until none remained in 1979. Lower Mather Acres ( LM A). This is a small triangular area with an en- trance on Mitchell's Mills Road. It is surrounded by deciduous or mixed woods and contains an old barn. To the southeast is Stebbins Gulch. This area was heavily populated by House Wrens. Six bluebird attempts were unsuccessful due to interference by wrens. The original nine boxes were gradually reduced to six by 1975. Five of these were removed before the 1976 season. The last box was removed before the 1980 season. Around the Lilacs (AL). This area stretches from the Arboretum re- ception center south along Sperry Road for approximately 1,000 ft to a JEAN EAK.IN No. 40 property line where it is separated by a fence from an undeveloped field. On the north a double row of white pines ( Pinus strobus ) separates it from the parking area. It extends west from the road for approximately 150 ft where it meets deciduous woods and groves of pine and aspen. A small pond and brook separate the lilac collection in the larger northern portion from an open area containing scattered trees and shrubs and continuing to the west where there is another pond. 1 he original 17 boxes were gradually reduced to 14 by 1969. Nine were removed prior to the 1970 season in an effort to reduce Elouse Sparrow and House Wren competition. There were four boxes from 1971 through 1974. In 1979 the last two boxes were removed due to heavy human use of the areas where they were located and continuing House Sparrow activity. During 10 of the first 1 2 years of the study one pair of bluebirds nested or attempted to do so. In 1967 and 1969 there were two pairs. Eight broods fledged, the last one in 1974. Of 14 nestingattempts byTree Swallowsonly five succeeded, the last one in 1973. The Crahapple Collection (AC). This area contained 24 boxes in 1965. In an effort to reduce House Sparrow and wren competition, we reduced the number to eight in 1970. As tree growth gradually eliminated open space, this area failed to attract bluebirds except for an open triangle con- taining three boxes at the extreme northwest bounded on two sides by de- ciduous woods and on one by crabapples (Mains) and hawthornes (Crataegus). By 1979 this remaining section (Thayer Center, T on map) contained only one box. The total AC area almost always attracted one pair of bluebirds, oc- casionally two. However, prior to 1976, when all but three of the boxes were removed, bluebirds were successful in only three years. From 1976 through 1980, with the exception of 1978 when the nesting was interrupted by House Sparrows, one or two broods fledged each year from the re- maining box in section T. Areas of Continuous Use Corning Lake (CL). This area surrounds Corning Lake except on the Sperry Road side (Fig. 5). It is almost entirely surrounded by deciduous woodland or hedgerow and contains collections of evergreens, birch ( Betula), alder (Alnus), and hornbeam ( Ostrya). The open space is broken here and there by maple (Acer) and sassafras (Sassafras alhidum) and there are willows (Salix) near the lake. Both islands are covered with dense de- ciduous growth. Most of the grassy area is mowed. 1983 EASTERN BLUEBIRD 9 Fig. 5. Corning Lake Area. Holden Arboretum. By 1967 the original 29 boxes had been increased to 32. In 1968, we re- moved eight and relocated seven. Since then the number has been reduced to 19. We eliminated five boxes which attracted House Sparrows and/or wrens and relocated two others. In this area the number of bluebird pairs varied from two to four with an average of 3.3, and 56 broods fledged. From 1967 through 1980, with the exception of 1977, at least one pair of Tree Swallows nested. In 1970 and 1974 there were four pairs. During one season there were three, and during three seasons, two pairs. Heath Pond (HP). This small area is almost surrounded by mixed woods. It contains two ponds, two enormous old oaks ( Quercus alba and Quercus rubra ) and the Arboretum’s rhododendron and azalea collec- tions. In 1965 there were nine houses. By 1971 we had removed all but four in an attempt to reduce wrencompetition, which appeared to have stopped bluebird attempts to nest. After 1 976 there were only three boxes. The first bluebird success in this area was in 1971. Since then at least one brood fledged in 8 of 10 years. Bicknell Acres ( BA). This is a strip of land approximately 500 ft wide and 2,200 ft long on the east side of Sperry Road extending north- northeast from Kirtland-Chardon Road to a low wooded area bordering a 10 JEAN EAKIN No. 40 Fig. 6. Bicknell Acres, Flolden Arboretum Fig 7. Chapin Acres, Holden Arboretum 1983 EASTERN BLUEBIRD I I pond (Fig. 6). The magnolia collection at the north end, which contains two boxes, is surrounded by deciduous woodland except where it meets the road. South of the magnolia collection, deciduous woodland nearly meets the road in three places, leaving two fields which slope to the east for 1,700 ft before they meet the woods. There are six boxes in all. There was one pair of bluebirds during two of the first six years. From 1971 through 1980 there have been two with the exception of 1976 when there were three pairs. A pair of Tree Swallows nested in this area during four years, including 1979 and 1980. Chapin Acres(C A). Thisarea is on the west side of Sperry Road ( Fig. 7). At the northeastern end is a pond. Deciduous woodland extends from the road along the far edge of the pond, then southerly, coming within 250 ft of the road at the southern end. The evergreen planting which occupies much of the area continues south to Kirtland-Chardon Road. This southern part of the area had boxes which were removed prior to the 1 976 season because they attracted House Sparrows and wrens. The one bluebird attempt (1974) failed because of wren interference. Since 1976, there have been seven boxes in the northern section of CA. There was one pair of bluebirds in 4 of the first 10 years. From 1975 through 1979 there was one pair. In 1980 there were two pairs. At least one pair of Tree Swallows nested in this area every year since 1966. The Paul Martin Farm (PM). Thisarea is bounded by deciduous or con- iferous woodland and traversed by three hedgerows following old fence lines (Fig. 8). There are three buildings near Kirtland-Chardon Road. On the southern border is a small pond drained by a brook which crosses the property in a north-northeasterly direction. Considerable open space re- mains in the northern field (approximately 375 by 2,000 ft) where the Ar- boretum’s maple collection is located. The southern portion adjacent to Kirtland-Chardon Road is planted with mountain ash ( Sorbus ) and var- ious species of horsechestnut ( Aesculus ), leaving little open space. The remainder of the southern portion consists of an undeveloped and un- mowed field approximately 275 by 1,100 ft. From 1965 through 1975 there were 30 boxes, some of which were grad- ually relocated away from Kirtland-Chardon Road and the buildings in order to avoid House Sparrow occupancy. In 1976, after studying the his- tory of each box, we removed all but 1 2, of which 6 were relocated. In 1980 there were 14 boxes. Pines (P). In 1977 we placed three boxes in the lot on the corner of Kirtland-Chardon and Booth Roads adjacent to the PM area. It is an un- developed field bounded by coniferous hedgerow, deciduous hedgerow, and woodland except where it meets the two roads. A pair of bluebirds 12 JEAN EAKIN No. 40 Fig 8. Paul Martin Farm. Holden Arboretum, nested in this area each year. The number of bluebird pairs in PM gradually increased from one in 1965 to three in 1976. After the addition of pines (P) there were three pairs in the combined areas (four in 1978). Usually there were one or two pairs of Tree Swallows in PM. Goose Pond (GP). Except for the Kirtland-Chardon exposure on the northeast, this area is surrounded by deciduous woodland. It contains five ponds and lakes. There are two dense plantings of mature conifers; one en- tirely white pine; the other, white pine and spruce ( Picea). There are many willows ( Salix). The area available for boxes is divided by a thick hedgerow running north and south. There are six boxes. One pair of bluebirds nested each year of the study. There were also usually one or two pairs of Tree Swallows. Upper Mather Jcra(lJMA). This is an unmowed field of coarse grass and weeds (approximately 400 by 2,000 ft) entirely surrounded bv mixed or deciduous woodland or thick, dense hedgerow. The northeastern half contains a nursery of small conifers. There were 10 boxes from 1965 through 1974. In 1975, due to vandalism and deterioration, only five re- mained. Before the 1976 season, we relocated these in an effort to reduce 1983 EASTERN BLUEBIRD 13 House Wren interference. Another box was installed before the 1980 nest- ing season. With three exceptions (all prior to 1976) there was a pair of bluebirds each year. From 1965 through 1976, presumably due to wren competition, the bluebirds failed in three of nine years. In three additional years, wrens prevented a second nesting by building in the boxes. From 1977 through 1980, one or two broods of bluebirds fledged each year. One pair of Tree Swallows nested in 12 of the last 14 years with only four failures. Lower Baldwin Farm (LBF). Except where it is bordered by the road, this area is almost entirely surrounded by deciduous or mixed woodland. The higher eastern section contains mature nut trees and the younger nut tree collection. The land slopes south and east to the lower field which con- tains a mainly deciduous mixed nursery. Except on one steep grassy slope, a patch of deciduous woods separates the two areas. There were originally 20 boxes in this area. Six of these, south of the ex- tension of Baldwin Road, were relocated in the present area by 197 1 . From 1968 through 1974, while the number of boxes remained almost constant, many were relocated in an effort to lessen competition from wrens and House Sparrows and attract more bluebirds. Prior to 1976, after studying the history of each box, we reduced the number to nine, of which four were in new locations. During three of the first four years of the project, only one pair of blue- birds attempted to nest in this area. Only two broods fledged. From 1970 through 1980 there have been two or three pairs every year and 29 broods have fledged. In 1977 a Tree Swallow nesting was interrupted, presumably by a predator. In 1979 and 1980 two pairs of Tree Swallows were suc- cessful. Post-1965 Areas Upper Baldwin Farm (U BF). This “upper field,” (actually lower in ele- vation than the lower field) is approximately 1,000 by 2,500 ft and is en- tirely surrounded by mature deciduous and mixed woodland. It containsa mixed deciduous nursery of small trees and shrubs. The grass is not mowed. Four boxes were placed in this area in 1968. Three additional boxes were added in 1969. One box was removed in 1975 after House Sparrows had destroyed bluebird nestlings for two successive years. During the 13 years this area always had one or two pairs of bluebirds and 26 broods fledged. Since 1969 there was at least one pair of Tree Swallows every year but one. Six Tree Swallow broods fledged. 14 JEAN EAK.IN No. 40 Norweb (N). This area, added in 1972, is bounded by deciduous woods and by two extensions of its own fields which have not attracted bluebirds. The eastern portion contains a holly planting. The western part is an un- developed, un mowed field. The number of houses varied from five to eight. Three at the south end near buildings were eliminated or moved due to House Sparrow competition. From 1978 through 1980 there were five boxes. During its nine years this area always had one pair of bluebirds. Nine broods fledged. There were usually one or two pairs of Tree Swallows. Five Tree Swallow broods fledged. Methods Monitoring of Boxes. The original plan in this study was to monitor the contents of the bluebird boxes on a weekly basis, recording the contents and removing House Sparrow nests. Sometimes during the first few years the interval between visits to boxes was as long as two weeks. In 1966 there were no visits after July 16. Since 1978 an effort was made to check the boxes twice a week. Beginning in 1975 we tried to check boxes with nest- lings every two days from approximately the fifth to the tenth day. After the twelfth day we used extreme caution in checking. We did not disturb birds which were engaged in house selection or nest building. We disturbed females brooding eggs as little as possible. Box Placement. During the 16 years of the project we removed some boxes and added others. In 1967 we eliminated twoareas(23 boxes) which had attracted only House Sparrows (see Fig. 4, Areas CB and PA). We re- moved II boxes from an island (Area I) in another area for the same rea- son. The number of boxes in two areas near the Arboretum Reception Center was gradually reduced until none remained in 1979. One of these sites ( AB, 17 boxes) was a trail through the woods which lacked sufficient open space to attract bluebirds. The other (AF, 16 boxes)ceased toattract either bluebirds or Tree Swallows after 1975 as it became more heavily used by people and as House Sparrow competition increased. Only one of 28 boxes was left in a remote part of the crabapple area (AC) which was heavily used by people and frequented by wrens and sparrows. Boxes were removed from parts of three areas which attracted only House Sparrows (CA, PM, FBF) and from another area which attracted only wrens (FMA). Roadside boxes, 22 at most in 1969, which were productive for several years, were eliminated gradually due to vandalism, predation and House Sparrow competition. Three new areas with 15 boxes were added, all of which attracted bluebirds (see Fig. 4, UBF, N and P). 1983 EASTERN BLUEBIRD 15 Fig. 9. Bluebird box on post with metal guards, winter 1980. In 1966, we decided that the boxes were placed too close together to meet the territorial needs of bluebirds. Superfluous houses only served to attract wrens and House Sparrows. In an attempt to reduce interference by these two species, we removed some boxes and relocated some, making certain that they were at least 100 ft apart. This was not done during the nesting season. In the winter of 1975 and 1976, after reviewing the history of each box, we reduced the remaining 146 to 100. After this, however, it was still necessary to remove or relocate certain boxes because of new planting or hedgerow growth. In 1980 there were 86 boxes. Since we found that bluebirds, when their nesting was interrupted, often sought a different site in the same area, we provided several boxes as alter- nate choices for pairs that might be disturbed ( Krieg, 1971 : 124). We had enough boxes in every area to accomodate such other species as Tree Swallows and Black-capped Chickadees. 16 JEAN EAKIN No. 40 We gradually replaced the original fenceposts with metal poles and guards to protect the boxes from predators (Fig. 9). Banding of Young. Most of the young were banded (Table 3 and Table 14). The banding was done when the birds were no younger than 8 or older than 12 days. Results An Analysis of Factors Causing Nest Failures House Sparrow. The fJouse Sparrow is an aggressive competitor for bluebird nesting boxes ( Kibler, 1969: 123; Zeleny, 1976: 1 16-1 19). In the 16- year study evidence indicated that House Sparrows destroyed 3 adult bluebirds, at least 33 young and 55 eggs, and interrupted 89 additional at- tempts of the bluebird to nest. There is also evidence that they destroyed 9 adult Tree Swallows, 29 young and 24 eggs, and interrupted 12 additional nest- ings (Table 4). Sometimes we actually witnessed House Sparrows entering boxes and attacking the occupants. In the other cases, we found the House Sparrow nest on top of the bluebird or Tree Swallow nest, often with bro- ken eggs or dead young. Young small enough to be removed were some- times found on the ground under the box. When they were too large, the House Sparrows usually delayed nest building until we removed them. Two of the bluebird adults were buried under House Sparrow nesting ma- terial. The other adult bluebird and the adult Tree Swallows were found in empty boxes early in the nesting season. All of the birds listed as destroyed by House Sparrows were severely pecked on the head. We always found House Sparrow nests when we first checked the boxes in mid-March. Eggs appeared after the middle of April. Nesting activity peaked in May, continued into June and decreased in July. There were some attempts in August. Most bluebird casualties occurred in May and June. We found House Sparrow nests in boxes less than fourfeet from the ground and in a box of smaller than average dimensions. In 1980 House Sparrows built in an experimental box with a screentop designed to be “sparrow-proof.” We routinely removed House Sparrow nests from our boxes. More often than not the House Sparrow returned to build again. In 1972, out of 75 cases of nest removal, 9 did not return; I 1 came back once; 27 returned twice; 5, three times; 3, four times; 2, five times; 4, six times; and 14, seven times. In 1970 a bluebird nest was completed on 22 April. On 27 April there was a House Sparrow nest on top of a dead, severely pecked female bluebird in 1983 EASTERN BLUEBIRD 17 the bluebird nest. (In this same box four bluebird young were pecked to death by House Sparrows in June of the same year.) We removed this box. On 10 April 1976, a dead, severely pecked male bluebird was found in an empty box. House Sparrow nests were removed from this box on 23, 27, and 28 April. On 9 May 1977, I found a House Sparrow nest with three eggs on top of a dead male bluebird which had been pecked severely on the head. This box was removed. We found evidence that suggests that the H ouse Sparrow’s success is not due to the unwillingness of the bluebird to fight or aggressively defend the nesting site or young, but to the fact that the bluebird pairs leave the nest together to seek food and return to the nest almost simultaneously. The eggs and young, therefore, are intermittentlv left undefended (Goldman, 1975:800). In 1965 a dead House Sparrow was found at the foot ofa box with blue- bird young which subsequently fledged. A second brood was destroyed by House Sparrows. On 18 June 1970, 1 founda HouseSparrow pairata bluebird house. The four approximately one-week-old bluebird young were alive but seemed weak. All had been pecked on the head and one had been severely pecked on the back. The bluebird pair returned at intervals, chased the House Sparrows away, and fed the young. The sparrows waited until the blue- birds left, at which time the male sparrow perched on the box and the fe- male entered. I chased them away when the female entered and otherwise waited for the bluebirds to return and defend the box. When the bluebirds seemed to be staying near and defending the box, I left for 20 minutes. When 1 returned the male sparrow was on the box, the female inside. The young had been pecked more severely and were quiet. The male bluebird returned with food and called softly and repeatedly for some time but did not enter the box when he received no response from the young. When the male House Sparrow returned and sat on the box, the male bluebird did not challenge. The female bluebird returned and both she and the male fluttered around the box calling but did not enter They made no attempt to chase the male House Sparrow, which finally flew away. The female bluebird perched on the box until the male House Spar- row returned; then she flew to a nearby stake protesting but not fighting. The male House Sparrow entered the box. During the several hours that 1 observed them the bluebirds fought ag- gressively every time they returned to feed the young until the young be- came too weak to respond to their calls (Hartshorne, 1962:141). At our next visit, three days later, there was a House Sparrow nest on top of the bluebird nest and one dead bluebird young on the ground. 18 JEAN EAK.IN No. 40 In 1975, a House Sparrow was observed several times leaving a box in which a female bluebird was brooding. This nesting was not interrupted. In 1976, a dead House Sparrow was removed from a bluebird nest. The nesting was successful. In 48 cases bluebirds nested successfully in boxes from which House Sparrow nests had been removed ( Kibler, 1969: 123). In 21 cases Tree Swal- lows were successful after the removal of House Sparrow nests (Table 4). House Wren. The House Wren is a formidable competitor for nesting cavities and may destroy eggs and young in boxes which it does not occupy (Zeleny, 1976: 1 20). Since it is a protected native bird we simply included it in our records during the first six years of the project. We attempted to pro- tect the bluebirds by relocating boxes away from hedgerows to discourage wrens (Kibler, 1969: 119). Wrens occupied 49 of 230 boxes in 1965; 95 of 200 in 1970. In six years approximately 900 young fledged. We believe that in this same period wrens destroyed four bluebird young and 40 eggs and interrupted eight other nestings. After 1970 we decided to remove wren nests from boxes which had a bluebird history or were attracting bluebirds. In 1971, wrens occupied or attempted to occupy 92 of 200 boxes; in 1972, 105 of 189. In these two years at least 450 fledged (Tables 6 and 7). In 1973 and 1974, with the approval of the Arboretum naturalist, weex- perimented with nest removal and with removal of all but one or two eggs. The majority of wrens kept returning after nest removal but deserted if only one or two eggs remained. In 1973, 135 young fledged. In 1974, only 35 fledged but the destruction of bluebird eggs increased from an average of 9 to 29. In the 10 years, 1965 through 1974, approximately 1,300 wrens fledged. During this period indications were that wrens destroyed 13 bluebird young and approximately 92 eggs; 5 Tree Swallow young and 14 eggs; and 28 chickadee eggs. When the eggs were pierced and/ or the wren built on top of the nest, I attributed the failure to wrens. Beginning in 1975, we routinely removed wren nests. We occasionally made an exception when bluebirds were nesting nearby. Wrens continued to attempt to occupy about one-fourth of our boxes. Wrens probably destroyed 21 bluebird eggs in five nests in 1975; 17 eggs in four nests in 1976; 8 eggs in two nests in 1977; 1 1 in three in 1978; none in 1979; one in one nest and 2 in another in 1980 (Table 7). The adult bluebird listed as destroyed by House Wrens in Table 7 was a female found in a box in which bluebird nest failure had been followed by two attempts by wrens to nest. Her head was pecked clean of feathers. The wrens subsequently occupied the box. Predators. Predators of several species destroyed 5 adult bluebirds, 55 1983 EASTERN BEUEBIRD 19 young, 1 10 eggs and 13 Tree Swallow young and 32 eggs(Table 8). Some- times it was possible to determine that the destruction was caused by a rac- coon or a red squirrel, while in other cases the evidence indicated a snake. Often we could not determine the identity of the predator. Greasing the metal fence posts was ineffective. In 1975, we began to replace these posts with round metal poles which could be greased, and in 1977 we began to place rectangular metal guards on the poles (as suggested by Zeleny, 1976: 107-109). In 1980 these guards were on all active houses. We have had two cases of predation in boxes so equipped. One box was on a short pole, since replaced. In 1980, one of three eggs disappeared and latertwo banded young disappeared from a box 52 in. above the ground. Human Interference. Human interference caused the loss of 9 bluebird young and 49 eggs, and 14 Tree Swallow eggs. Two chickadee nests con- taining five and six eggs respectively, were removed from boxes in sight of a trail. Three young were found dead at the foot of another box, the door of which was open. We removed these boxes. In 1976 we put Philips screws on boxes in areas open to the public. More recent incidents of human dis- turbance to the boxes in remote areas made it necessary to put such screws on all boxes (Table 9). We attempted to keep human interference at a minimum by making the boxes inconspicuous and by removing ones that proved vulnerable. Unknown Causes of Destruction or Disappearance. We could not al- ways determine the cause of nesting failure. We could not rule out House Sparrows, House Wrens, predators or human beings. Such cases ac- counted for 19 bluebird young, 57 eggs and 25 nest failures. Three Tree Swallow nests were destroyed by unknown causes. Parasitism by Cowbirds. On 8 May, 1972, there were two Brown-headed Cowbird (Molothrus ater) eggs in a bluebird nest with two eggs. On 16 May a third bluebird egg was in the nest. The two cowbird eggs were re- moved. Three bluebirds fledged from this nest. On 15 June, 1971, a young cowbird was found in a Tree Swallow nest with two unhatched eggs (there had been five). In the same year one live cowbird nestling was found in a nest with four dead T ree Swallow young. Wasps. We removed many wasp nests from boxes. In May 1975, after a bluebird nest with three eggs had been deserted, we found an active wasp nest. In 1975 a Tree Swallow nest with five eggs was deserted after an old wasp nest fell on top of it. In 1978 we began to coat the ceiling and upper walls of the boxes with Vaseline. This reduced but did not eliminate the problem. Disappearance of Eggs or Young from Nest. Sometimes it was not pos- sible to determine whether eggs or young had disappeared from a nest con- taining live nestlings. When the nest was infested with fly larvae we as- 20 JEAN EAKIN No. 40 sumed that dead young had been removed by the adults. It is unusual for single eggs to disappear from clutches which are otherwise successful ( Pea kail. 1970:25 1 ). When in doubt as to whether an egg or a nestling had disappeared 1 assumed that a dead nestling had been removed by a parent (Hartshorne, 1962:145-146). Failure of Eggs to Hatch. In 16 years, eight clutches of bluebird eggs failed to hatch. In the case of one second clutch of two eggs the pair was near the nest on the eighteenth day after completion. Four white eggs, warm on the twelfth day, were cold on the eighteenth. One clutch, the third of the season, was found covered with new nesting material five to eight days after laying was completed. The pair did not return. In two cases one egg was laid and then deserted. In 59 otherwise successful clutches of blue- bird eggs, 70 failed to hatch. Two clutches of Tree Swallow eggs failed to hatch. In 15 partially suc- cessful clutches, 38 failed. A paulina sialia. Parasitic fly larvae, A paulina sialia, first became a problem in 1968. After two clutches of five young and nine other young were found dead in infested nests we began to dust the nests ( with or with- out eggs but not with young) with rotenone. There were no further casual- ties in 1968. In 1969 and 1970 the only fatalities from parasites occurred in nests which had not been treated. In 1971 six young disappeared and one was found dead in nests on which rotenone had been used, and in 1972 two clutches of banded young were found dead in infested boxes. One of three was found dead in another nest. Three young fledged from another after fly larvae had been removed by hand. Only one young Tree Swallow fledged from four infested nests in which 18 were dead or missing. Rotenone had been used in all of these nests. In 1973 and 1974, larvae continued to appear in nests treated with rotenone, but 25 young fledged from infested nests af- ter larvae had been removed manually. In 1975, suspecting that some of the parasitic flies might have developed immunity to rotenone, we decided to remove larvae by hand. Examination of nest material spread out on white paper revealed the presence of many larvae too small to be visible in the box. Since it was impossible to remove them without destroying the nest, our solution was to substitute a man- made nest of dry grass, as described by Johnson (1932:29) and Mason (1944:232). From the larvae and pupae which we collected. Dr. Sonja Teraguchi (en- tomologist at the Cleveland Museum of Natural History) hatched adult Apaulina sialia and the hyper-parasite Mormortiella. She advised us to remove infested nests and place them at the foot of the boxes in wire mesh bags which would confine adult Apaulina sialia but permit the smaller Mormoniella to escape. We followed this procedure in most cases. We sub- 1983 EASTERN BLUEBIRD 21 stituted artificial nests of dry grass which the bluebirds never failed to accept. From 1975 through 1979, 60 bluebird and 9 Tree Swallow voungdied in infested nests, most of which probably had not been checked soon or thoroughly enough. During the same period 152 bluebird and 61 Tree Swallow young fledged from nests that had been substituted for infested ones. In 1980, although larvae were present in the nests of 1 8 broods, there were no nest failures. Four bluebird young were found dead or dis- appeared from four nests in which larvae were present. TwoTree Swallows were found dead and one disappeared from two such nests. Apaulina sialia appeared from late May through August. They were present in every area. In some cases they were present in the first nesting but not the second in the same box; in other cases the reverse was true. In 1979, in at least 10 cases larvae were present- even abundant one to three days after the box had been cleaned. They were found afterall or part of the young had Hedged in boxes that had been cleaned on the tenth to twelfth days. During six seasons! 1975-1980) larvae werefound in the nests of 103 of 155 broods in which young survived to the age of one week (see Table 10). Other Parasites. On 10 July 1975, two bluebirds, I or 2 days old, were dead in the nest. Dr. Teraguchi found them covered with microscopic lar- vae. (Of 4 eggs, 2 had disappeared, one at a time.) In 1973 a Tree Swallow nest had six eggs due to hatch on 20 June. On 25 June there were four live young and one dead. On 2 July there were only two young. The nest was infested with tiny Hies which a volunteer re- moved. On 6 July there were a few flies which were removed. The young were banded on 10 July, when they were approximately 20 days old. T wo days later one of the young was seen leaving the nest. Both Hedged. Young Dead in Nest — Cause Unknown. In June of 1976 a volunteerdes- cribed one of four Tree Swallow young, 19 or 20 days old, as“sick. "There had not been and were at no time any visible parasites in the box. Two days later there were three dead young. On 9 June, 1977 I banded four bluebirds approximately eight days old. They were active and seemed healthy. No larvae were present. The adults fed them as soon as I left. Two days later they were dead. I took these young to Dr. Aaron Leash, a local veterinarian, who found them to be well nourished. He dissected one and could find no parasites or abnormalities. In a nearby area on 8 June, two young approximately two days old were found in a nest which had had six eggs. One tiny nestling was dead on the ground at the foot of the box. On the 1 I th, I found the two young ailing, one on its back and one on its side. The adults were in a nearby tree. After consulting the Arboretum naturalist, 1 gave each nestling two drops of vit- 22 JEAN EAKIN No. 40 amins which seemed to revive them. The parents fed them before I left. We administered vitamins again on 13 and 15 June. The young fledged at ap- proximately 17 days. In a nearby area three Tree Swallow young, hatched on 8 June, seemed normal on 1 1 and 13 June but were dead on the 18th. No larvae were found. These three boxes were near ornamental trees and shrubs which had been sprayed. After that date, the Arboretum kept me informed of any ne- cessary spraying, which was always selective. Weather. Kibler states, “That inclement weathercan be a significant fac- tor in bluebird mortality is borne out by the result of my first nesting period in 1968 in which 76% of 51 nestlings died in the nest” (1969:125). On four occasions in the course of our study adverse weather conditions appeared to be the cause of young mortality and/or failure of eggs to hatch. In 1973, from 4 to 6 May, the temperatures of the Chardon, Ohio, weather bureau were recorded as follows: There were six bluebird nests in five of which dead young were discovered or from which young were missing: 1 ) On 6 May, four young, hatched approximately 20 April, were dead. 2) On 7 May, four young, hatched on 1 May, were dead. 3) On 9 May, a nest which had had fouryoung approximately four days old on 3 May, was empty. 4) On 9 May, one of three young, newly hatched on 1 May, was missing. 5) One nest had four eggs on 29 April. On 5 and 1 1 May the incubating female did not leave. On 20 May the female was dead on the nest with two dead young several days old. 6) On 21 April, a female was incubating five eggs. On 29 April there were four young and one egg. Four young were banded on 8 May and Hedged. No fly larvae were found in these boxes. Of seven nests with eggs, one clutch of five and one of two failed to 4 May 5 May 6 May Maximum 48 42 56 Minimum 32 29 30 1983 EASTERN BEUEBIRD 23 hatch. In three clutches, one of five, one of three, and one of four failed to hatch. On 27 April, 1976 daytime temperatures were in the low 30s; there was a ground cover of snow on 3 May, and the temperature was in the low 30s in the daytime of 4 May. During this period there were six nests with young: 1 ) One brood of four hatched approximately 25 April had food in their stomachs on 27 April. They were alive on 4 May, dead on 8 May. 2) One brood of four hatched approximately 2 May were alive on 6 May, dead on 8 May. 3) One brood of five alive and approximately one week old on 5 May were dead on 10 May. 4) Two young alive and approximately three days old on 5 May were dead on 10 May. 5) One nest which contained three eggs on 14 and 28 April contained one egg and one young on 5 May. On 9 May there was only one egg. 6) One nest which had four eggs due to hatch on 3 May had three young and one egg on 8 May. The three young fledged in spite of fly larvae infestation after a new nest was substituted. No fly larvae were found in the first five boxes. Of six clutches of eggs, two failed to hatch. On 8 May, 1977, a cold front moved across Lake Erie with wind gusts in excess of 70 mph. The temperatures (° F) in Chardon for that day and the ensuing four were as follows: Maximum 8 May 56 9 May 65 10 May 46 11 May 55 12 May 66 Minimum 27 3 1 (45° during dav at Arboretum) 28 27 34 During this period there were five nests with young: 1 ) In a box at the highest elevation in the area, only one of five young survived. These young hatched between 3 and 7 May. No fly larvae were found. 24 JEAN EAK.1N No. 40 2) One of a brood of five hatched between 5 and 7 May was alive on 22 May but dead in the box after the others fledged. Fly larvae were present. 3) Two young hatched on 8 May and fledged. 4) One of five young hatched between 2 and 7 May disappeared on or before 8 May. Another disappeared between 8 and 13 May. On 15 May the remaining three were dead. Fly larvae were removed on 13 May. 5) In one box were five eggs due to hatch on 5 May. There were only four young on 8 May. One of these disappeared bv 23 May, when fly larvae were found. The others fledged. There were six nests with eggs: 1 ) In one clutch completed on 3 May, four of five hatched. 2) One clutch of four was completed on 8 May. The female was incu- bating on 15 and 21 May. On 23 May the eggs were gone. 3) One nest contained five eggs on 23 April and 3 May. On 10 May, a Tree Swallow had built on top. We do not know if the eggs failed to hatch or if the Tree Swallow drove the bluebirds away. 4) A clutch of five eggs present on 24 April and 9 May had hatched by 16 May (probably on 9 or 10 May). 5) Of four eggs present on 9 May, one disappeared and three hatched. 6) On 9 May a female was brooding five eggs due to hatch 1 1 May. On 16 May there were two unhatched eggs and one dead nestling. In 1979 the maximum and minimum temperatures (° F) and the precipi- tation ( inches of rain) at Chardon were as follows from 24 through 27 May. Maximum Minimum Precipitation 24 May 73 35 .00 25 May 53 36 2.00 26 May 52 38 .70 27 Mav 5! 35 .73 In five boxes, 24 healthy young free of fly larvae and banded or ready to band were found dead on 26 and 27 May. They were from one to two weeks old on 24 May. Four in another brood which hatched on 6 May and had not developed normally, although free of fly larvae, were also dead on 26 1983 EASTERN BLUEBIRD 25 May. In one clutch of four dead on 26 May, Hv larvae were present. In two cases in which three of five and four of five approximately two weeks old were found dead, fly larvae were present. The three young which fledged from these two nests were the only ones to survive this cold spell. Three nests with eggs were not affected; all hatched. The four Tree Swallow nests which contained eggs were deserted. In every case the pairs made a new start in the same box. (On 26 May there was no activity in the barn where Barn Swallows ( Hirundo rustica) were nesting except for two brooding adults. On 27 and 28 May, seven adults were dead on the floor. The nests containing eggs were deserted. ) White Eggs From 1973 through 1980 we found 44 white or extremely pale blue eggs in 10 clutches, six of which were at least partially successful. Of the 18 banded young, none was recaptured. Two of the females producing white eggs had hatched from blue eggs. Clutch Size Clutch size varied from 3.9 in 1968 to4.66 in 1965 and 1980. Theaverage was 4.22. I used only completed clutches to make this calculation (Table Breeding Season The breeding season in terms of clutch completion extended from the first week in April to mid-August. The 16-year average shows a peak in late April and early May and a lower peak from 9 to 28 June (Table 12). Interval Between Broods In computing the interval between fledging or nesting interruption and the laying of the first egg of the next clutch I have used only pairs which have nested a second or third time in the same house. If 1 did not know the exact dates I used 14 days as the incubation period and 17 days as the age of 26 JEAN EAKIN No. 40 fledging (Clapp, 1974:15-19; Thomas, 1946:156-158). From fledging to laying of first egg in 22 broods of one to three young the average interval was 20 days. The range was from 42 days to 8 days. In 60 broods of four or five the average interval was 24 days. The range was 42 to 9 days. In five nest failures involving eggs the intervals were 65, 54, 32,21, and 2 or 3 days with an average of 25. In 13 nest failures involving young the average interval was 19.4 days with a range of 45 to 8 days. Banding Records Most of our bluebird young were banded (Table 3). We were able to pick up 10 brooding females and record their band numbers. Nine of these had been banded as nestlings. Four were 6, 5, 4, and 3 years old respectively. Two were 2 years old; three only 1 year old. One was found in two different boxes, the second of which was the one from which she had fledged. One, banded as a brooding female, was brooding in the same box the following year. We have a five-year record on our most productive box. The female in this box in 1973 and 1974 had been banded as a nestling in 1969. The fe- male in the box in 1975 had been banded as a nestling in 1972 in another area. In 1976, the female in this box was unbanded. I banded herand was able to recapture the same female in the same box in 1977. On 30 June 1980, I caught a small flock of bluebirds in a net. The adult pair was unbanded. Three juveniles had fledged between 9 and 12 May from a nearby box; two had fledged from another box on 4 June. Tree Swallows Population. In 1966, only one pair of Tree Swallows nested in one of our boxes. Since then there has been a gradual increase. The average number of pairs for 1 5 years ( 1966-1980) is nine. In 1980, there were 13 pairs. They laid 747 eggs, of which 567 hatched. The number of young fledged fluctu- ated from four in 1972 to 5 1 in 1980, with the total at 396 (Table 14 and Fig. 3) and the average clutch size, 5.03 (Table 15). The breeding season computed on the basis of clutch completion ex- tended from 10 May to 22 June. Over 73% of the clutches were completed between 20 May and 8 June (Table 16). In seven cases a Tree Swallow female laid a second clutch in the same 1983 EASTERN BLUEBIRD 27 box after the first was interrupted. One female swallow laid eight eggs be- tween 10 May and 1 June. On 8 June there were eight eggs. On 19 June the box contained one intact egg, a broken egg, and a newly hatched dead young. In one box, four of five eggs hatched after which three of the young disappeared. Following this, two more eggs were laid, but failed to hatch. The remaining young fledged. There were two cases of second broods of Tree Swallows, both in 1980. One brood of six, which hatched on 3 J une, was dead on 1 I June ( last seen alive on 9 June). On 19 June, presumably the same pair of adults was build- ing in a nearby box. On 22 June, there were four eggs; three young fledged. In the other instance, five of seven young fledged after 14 June. On 22 June a new nest with one egg was found on top of the old. Of foureggs laid, one hatched. The bander discovered two dead young from the first clutch un- der the new nest. Tree Swallow broods often include one nestling smaller than the others. Occasionally there are two. These birds often fail to fledge, particularly if fly parasites are present. In our study, excluding cases of total or partial nest failure due to other causes; of two broods of seven all fledged in one, two were found dead in the other; of 20 broods of six all fledged in I 1 , one disappeared from or was found dead in nine; of 38 broods of five all fledged in 23, one disappeared or was found dead in 1 1, and two disappeared or were found dead in four. Banding Records. Tree Swallows were not present in every area each year(Table 2). 1 banded 37 brooding adults in 1975-1980. During the same period, 200 young were banded. From 1966 to 1974, approximately 100 were banded (Table 14). We have had only three recoveries of these banded birds. Two adults banded in 1975 were brooding in different areas in 1976. One adult found dead in a box in 1974 had been banded in a box one mile away in 1973. Although certain boxes appeared to be preferred, we have no evidence that the same swallows return to the same nest sites year after year. Bluebird-Tree Swallow Competition Tree Swallows seldom interfered with the bluebirds. The instances of competition were too rare to be significant. The two species appear to share the same enemies and benefit from the same measures to protect them from House Sparrows, House Wrens, predators, human interference and parasitic fly larvae. 28 JEAN EAKIN No. 40 Productiveness of Boxes — Bluebird and Tree Swallow One box, occupied by bluebirds for 12 consecutive years, produced 45 young; another occupied 10 consecutive years, 54 young. From each of three boxes 26 young fledged. Young fledged from all but one of these boxes in 1980. Of our 12 most successful boxes, 1 1 were still in use in 1980. The following table shows years of bluebird occupancy. Number of Boxes 2 2 4 5 18 19 20 Number of Years 12 1 1 10 9 8 7 4 3 2 Of 82 boxes occupied more than once, 54 were still in use in 1980. The other boxes were removed either because they had ceased to attract bluebirds or because of repeated bluebird nesting failures. Of 86 boxes present in 1980, bluebirds had occupied or attempted to occupy 65; Tree Swallows, 39. Both species had occupied 28. The following table shows occupancy of boxes by Tree Swallows: Number of Boxes 4 I 3 4 18 Number of Years 6 5 4 3 2 1983 EASTERN BLUEBIRD 29 No box has been occupied by Tree Swallows for more than threeconsecu- tive years. Of the 1 2 boxes occupied for three or more years, I I were still in use in 1980. Natural Cavities. On two occasions, we observed bluebirds nesting in natural cavities. In 1966 a pair was seen on 26 April, 5 May, and 9 May en- tering and leaving a hole in a dead stump projecting from the top of an old white oak (Querus alba) in the center of a small bog (Fig. 4, HP). On 17 May, H ouse Wrens had occupied the cavity. During June 1980 a pair nest- ed in a hole in a sassafras approximately 20 ft above the ground ( Fig. 4, CL). Black-capped Chickadees Chickadees attempted 23 nestings from 1967 through 1980. There have been 1 1 broods. Of 108 eggs laid. 58 hatched and 55 young fledged. House Wrens destroyed 28 eggs and interrupted two other nesting attempts. Human interference caused the loss of three young and I I eggs. Bees inter- rupted one nesting (Table 17). Bewick Wrens In July 1969 a pair of Bewick W rens attempted to nest in a box in area CA (Fig. 4). They completed a nest ( more grass than twigs) on the 2 1 st. Ten days later. House Wrens were in the box. Discussion During the 16 years of the project, the minimum numberof nesting pairs increased from 8 in 1965 to 18 in 1980 (Table 3). From 1972 through 1976, the number fluctuated from 18 to 23. In 1972, when there were 19 pairs, 3 were in roadside houses. In 1973, 3 of the 22 pairs were in roadside houses, one in an area which subsequently became unsuitable for bluebird boxes. In 1 974, 3 of the 23 pairs were in roadside boxes, 2 in areas which became un- suitable. In 1976, 2 of the 21 pairs were in roadside houses. After 1976, we removed the remaining roadside houses because of human interference and House Sparrow competition. The average clutch size during the 16 years was4.22. PeakalK 1970:249) found an average of 4. 48 for Ohio based on 63 1 nest cards (97 prior to 1964; 30 JEAN EAK1N No. 40 the remainder from 1964 to 1969). Since he used records from various parts of Ohio and covered a different period, some differences could be expected (Table 1 I ). Following Peakall’s procedure, I have based my computation of seasonal activity upon clutch completion. Table 12 shows the number of clutches completed in each 10-day period of each season. It also shows the percentage of the total number of clutches for 16 years completed during each 10-day period. In Table 13 my figures are compared with those of Peakall Since his study covers a larger part of Ohio and ends in 1969, it is difficult to compare his data with the present study (Peakall, 1970:245- 251). The number of young fledged fluctuated between a low of 31 in 1967 to a high of I 10 in 1980. The 1966 figure is based on incomplete data. In spite of fluctuations, the Hedging trend has been upward (Table 3 and Fig. 2). Nest failures have been reduced by measures taken to deal with specific prob- lems. House Sparrows. Regular removal of Elouse Sparrow nests made more boxes available for nesting bluebirds and Tree Swallows. Both species had successful nestings in boxes from which House Sparrow nests were re- moved (Table 4). We eliminated House Sparrow productivity in our boxes. The number of boxes in which nesting attempts by House Sparrows occurred decreased (Table 5) as we removed or relocated those which at- tracted them, often near buildings (Zeleny, 1976:74). We were not able to keep these aggressive birds from attempting to occupy some of our boxes or to eliminate them as occasional causes of bluebird or Tree Swallow nesting failure (Table 4). House Wrens. The House Wren, a native bird and a natural competitor for bluebird nest sites, posed a more complex problem than the House Sparrow. During the first 10 years of the project, approximately 1,300 wren young fledged from our boxes. After 1975, when we began to remove wren nests routinely, no more young fledged. We believe that wrens will not tolerate competition for available insects within 50 to 100 ft of their nesting cavity. They frequently pierce the eggs in boxes near their own even if they do not subsequently occupy these boxes. To reduce such inter- ference we placed our boxes more than 100 feet apart. Because wrens pre- fer nest sites near shrubbery ( Kibler, 1969: 1 19), we placed our boxes as far into the open as was possible to do and still provide a safe landing for the bluebird young on theirfirst flight (Zeleny, 1976:74). As we removed boxes attractive to wrens, the number which they attempted to occupy decreased (Table 6). We were able to reduce but not eliminate wren interference with nesting bluebirds (Table 7). 1983 EASTERN BLUEBIRD 31 Predation. Failure due to predation seemed to increase with the increase in the bluebird population (Table 8). Regular monitoring of the boxes may have been a factor. From 1977 on, rectangular sheet metal guards folded around the poles seemed to keep most would-be predators out of the boxes. It is probably not possible to eliminate predation by raccoons. Human Interference (Table 9). Part of the losses due to human inter- ference can be described as theft or vandalism. Some were d ue to curiosity and carelessness. We found that boxes fastened with Philips screws were protected from interference by human beings who wished to remove or merely look at their contents. Occasionally an entire box was damaged or removed. We kept this kind of destruction at a minimum by removing vul- nerable houses and by keeping our boxes as inconspicuous as possible. With the exception of one dark green box, the boxes were unpainted. The metal guards were painted dark brown. Apaulina sialia. These larvae were present in two-thirds of our broods of bluebirds and 154 young died in infested nests (Table 10). Rotenone, ef- fective in 1968, 1969, and 1970, lost its effectiveness from 1972 to 1974. Beginning in 1975, we removed larvae manually, which necessitated re- moving the infested nest and substituting one of grass made to simulate the original nest as far as possible. We found larvae as early as the fifth day, when we felt that the young could be handled safely. We found it necessary to change the nest one or two more times at two-day intervals. When this procedure was followed, there were few if any fatalities from these para- sites. The infested nest was placed in a mesh bag at the foot of the box. Adult Apaulina sialia flies could not escape. The smaller Monnoniella which parasitizes the Apaulina sialia pupae could leave or enter the bag. We do not know if larvae pupated or if adult Apaulina sialia emerged with- in these bags. We have no proof of the presence of Monnoniella in the bags. We followed this procedure in an effort to permit natural control of Apaulina sialia by Monnoniella. Removal of the young bluebirds and returning them to a substitute nest did not disrupt the nesting process. One removal could be combined with banding. The fly larvae, if unchecked, caused so many fatalities in our area that a routine nest change on the fifth or sixth day seemed to us to be a wise precaution. Weather. It appeared that in certain years adverse weather conditions in northeast Ohio caused nest failures. This occurred during four years of our study. Fortunately, most bluebird pairs when interrupted go on to a sec- ond or even a third nesting. Eight pairs attempted a second nesting in the same box, and five pairs were successful One pair which lost four of five young raised two more broods in the same box. Often bluebirds seek a dif- 32 JEAN EAK.1N No. 40 ferent box near the one in which they have had a nest failure. We believe that in eight cases of nest failure during cold weather the pairs had second nestings in nearby boxes. One pair apparently moved toa nearby box and, after an interruption by House Sparrows, moved to the original box and raised a brood. One pair probably had two successful nestings in a nearby box. Since we did not dissect any of the young that died during periods of cold weather, we do not know if starvation due to a temporary dearth of avail- able insects was a factor. Of 23 survivors. 14 were less than a week old, 5 approximately one week old, and 4 approximately two weeks old. It may be that birds young enough to be brooded have a better chance of survival. The 82 young that died varied in age from one day to two weeks. Because we cannot eliminate all the causes of bluebird nest failure, we feel that in the future the number of fledglings will continue to fluctuate. It seems clear that humans can help the bluebird population to maintain it- self, if not to increase. If every portion of our study area in use in 1980 were occupied according to its demonstrated capacity, we might expect to have 21 nesting pairs. Bluebirds nest on some properties adjacent to the Arboretum. In two in- stances, brooding females banded as young at the Arboretum were cap- tured on such properties. Other bluebirds nesting in the vicinity have been observed to be banded. It seems reasonable to expect that any population increase will take the form of territorial expansion rather than increased concentration on Arboretum property. Summary During 16 years (1965-1980) a study was made of a volunteer effort to increase the population of the Eastern Bluebird, Sialia sialis, at the Holden Arboretum in Lake County, Ohio. In 1965 there were 230 boxes in 14 areas. Eight boxes were occupied by bluebirds, 49 by House Wrens (Tro- glodytes aedon ), and 134 by House Sparrows ( Passer domesticus). During the study, five areas with 64 boxes were eliminated and three areas with 16 boxes were added. The total number of boxes was gradually reduced to86. In 1980 bluebirds occupied 22; Tree Swallows ( Iridoprocne bicolor) 15; and Black-capped Chickadees ( Parus atricapillus) 2. House Sparrows attempted to occupy 15; House Wrens, 22. The minimum number of nesting bluebird pairs increased from 8 to 18 or 19 (Table 1 ). The number of young which fledged fluctuated from 10 in 1983 EASTERN BLUEBIRD 33 1966 to 1 10 in 1980 (Table 3 and Fig. 2). The following measures were taken to deal with specific causes of blue- bird nest failure: House Sparrows. Removal of boxes that attracted only sparrows re- duced the number of their nesting attempts. Relocation of others dis- couraged but did not stop sparrow competition. Regular removal of spar- row nests eliminated their reproduction in the boxes and made more boxes available for bluebirds. House Wrens. Removal of boxes occupied only by wrens and relocation of others as far into the open as possible reduced the number of attempts by this species to nest in bluebird boxes. From 1975 on, wren nests were re- moved from the boxes. These measures reduced but did not eliminate bluebird nest failures due to wrens. Predators. Rectangular metal guards folded around the posts were found to be effective against most would-be predators. Human Interference. The use of Philips screws in the doors of the boxes prevented nest failures caused by humans. Apaulina Sialia. Larvae of this species were present in two-thirds of the boxes in which bluebird young survived to the age of one week, and 154 young died in infested nests. Rotenone, effective at first, became useless as the flies apparently developed immunity to it. We found that nest failures caused by these parasites could be prevented by removing nests when the young were five days old and substituting nests of grass made to simulate the original. This procedure did not interfere with the nesting process. The following tables show what we believe to be causes of failure of eggs to hatch and of young to fledge in order of numerical significance. Bluebird Eggs Causes of Failure Number of Eggs House Wrens Predators House Sparrows Human Interference Clutch Failed to Hatch or Abandoned Weather Wasps 152 110 55-57 49 36 21 3 34 JEAN EAK1N No. 40 Bluebird Young Causes of Failure Number of Young Apaulina sialia 154 Weather 82 Predators 55 House Sparrows 33-35 House Wrens 13 Human Interference 9 Causes of Failure Predators House Wrens House Sparrows Human Interference Cowbirds Wasps Bluebirds Weather Tree Swallow Eggs Number of Eggs 32 31 24 14 5 5 2 2 Tree Swallow Young Causes of Failure Apaulina sialia House Sparrows Weather Predators House Wrens Cowbirds Number of Young 46 29 21 13 7 4 35 Acknowledgments 1 am indebted to Virginia Barrus who initiated Project Bluebird and to C. W. Eliot Paine who helped to coordinate it during the early years. 1 gratefully acknowledge the cooperation of R. Henry Norweb, Jr., Director of The Holden Arboretum, and the Arboretum staff. Dr. Harold Mahan, Director, and Dr. Sonja Teraguchi, Entomologist, of the Cleveland Museum of Natural History gave valuable assistance in an advisory capacity, and Dr. Mahan edited the manuscript and made use- ful suggestions regarding the use and arrangement of data. Without Elizabeth E. Walker and the Arboretum bluebird volunteers, the study would not have been possible. The photographs were taken by Patti Buchtel of the Arboretum staff and by Elizabeth Walker. o oo — v^fNiCNl ~ oo 04 O' 04 m s© 50 SO in Os 00 C/2 1— C/2 "O o o C/2 00 00 C/2 01} OX} C/2 OX} ft n oo c 3 o 00 c 3 o E o C/2 00 00 O 02 LU lu lu >* LU £ sh- OO o '-H- o o ■S <-*=. o 73 E V- t— !- X J3 1— 00 ftfi 1 'c E 3 z C/2 JD E 3 X e 3 Laid O fl c CJ l— U u, 70 u- o a X z Z CL CL Banded 36 10 20 Total Number of Eggs Laid: 2.004 Total Number of Eggs Hatched: 1.428-30 Total Number of Young Fledged: 1 .015 TABLE 4 Effects of House Sparrows on Eastern Bluebird Box Nesters — — rn — Tt rf m — co — — • C 4 P\J i/~> (N ^ oo n oo - Os C\l o CnJ \0 Tf c*~, ^ C*-> CsJ ‘o^ot^oooso — co sOvOvOvOsOi — r- j — O' Os Os Os 0s Os Os Os rfiTfi^v0h000'O r-r-r-r-r-i^-r^oo OsOsOsOsOsOsOsOs TOTAL 3 33-35 55-57 40 JEAN EAKIN No. 40 TABLE 5 Hpuse Sparrow Occupancy of Eastern Bluebird Boxes Year Total Boxes Number of Boxes House Sparrows Attempted to Occupy 1965 230 134 1966 236 179 1967 228 150 1968 204 1 12 1969 201 lb 1970 200 91 1971 200 100 1972 189 99 1973 178 72 1974 17! 77 1975 146 69 1976 100 65 1977 97 50 1978 89 40 1979 84 26 1980 86 15 1983 EASTERN BLUEBIRD 41 TABLE 6 House Wren Occupancy of Eastern Bluebird Boxes Year Occupied or A t tempted to Occupy Earliest and Latest Dates Found in Boxes Period during which Bluebird Eggs Were Destroyed 1965 49 of 230 No Record 1966 53 of 236 8 May, 16 July May and June 1967 92 of 228 5 May, 18 Aug. May 1968 9! of 204 29 Apr., 23 Aug. May through July 1969 89 of 201 26 Apr., 27 Aug. April through July 1970 95 of 200 29 Apr., 12 Aug. May and June 1971 92 of 200 1 May, 24 Aug. May through July 1972 105 of 189 21 Apr., 17 Aug. June and July 1973 77 of 178 29 Apr., 24 Aug. April through June 1974 93 of 171 29 Apr., 18 Aug. May through July 1975 72 of 146 2 May, 2 Sept. April through August 1976 43 of 100 10 May, 18 Aug. May through July 1977 29 of 97 16 May, 18 July May 1978 21 of 89 16 May, 13 Aug. June 1979 18 of 84 2 May, 13 Aug. None 1980 22 of 86 3 May, 2 Aug. June and July TABLE 7 Effects of House Wrens on Other Eastern Bluebird Box Nesters ^ *3 ^ “Q A « So o "J4 rSV3 £; C Q s h. £ ^ 0 <$ -r- 1 £? 5 ^ Q t$ * b« 5; =2 § o' « o £ r S ^ 2 ^ Q 5j c ^ tb ~Q % -L .a -c, wj ^ 5 „ — — . r*1 .— — (N — (N — \o r-- oo sD \Q vO vO O' C?' O' O' Qs O — CN sO h h h O' O' O' O' n ^ 'O r- r- 0N O' O' 0> r- oo O' o r^- r- r- o© O' <0 O' O' TOTAL 1 adult 13 young TABLE 8 Numerical Summary of Predation on Eastern Bluebirds and Tree Swallows Year Eastern Bluebirds Tree 5 wallows Adults Young Eggs Young Eggs 1965 1966 1967 4 1968 1969 4 1970 4 7 1971 2 10 9 1972 4 17 1973 1 3 9 1974 18 3 1975 1 17 23 5 S3 1976 1 8 16 4 3 1977 4 9 1978 3 4 4 1979 9 1980 2 1 TOTAL 5 55 1 10+ 13 32 TABLE 9 Detrimental Effect of Eluman Interference on Nesting Success Year Bluebird Young Eggs Tree Swallow Young Eggs Chickadee Young Eggs 1969 6 1970 1971 12 1972 3 1973 2 1 1 1974 15 3 1975 5 2 1976 4 1 1977 2 1978 5 1979 2 1980 5 8 TOTAL 9 49 0 14 3 1 1 44 JEAN EAKIN No. 40 TABLE 10 Incidence of Apaulina sialia in Eastern Bluebird Nests Year Number of Broods in Which Young Survived One Week Number Infested With Apaulina sialia 1975 27 23 1976 18 12 1977 30 23 1978 21 9 1979 32 18 1980 27 18 TOTAL 155 103 TABLE 1 1 Computation of Eastern Bluebird Clutch Size Year Completed Clutches Eggs Laid A verage Size 1965 12 56 4.66 1966 1 1 44 4.00 1967 17 68 4.00 1968 19 74 3.90 1969 25 105 4.20 1970 23 92 4.00 1971 34 142 4.17 1972 30 131 4.36 1973 41 164 4.00 1974 41 177 4.31 1975 41 161 3.92 1976 42 176 4 19 1977 34 143 4.20 1978 25 109 4.36 1979 35 160 4.57 1980 30 140 4.66 Average clutch size for the 16 years: 4.22. TABLE 12 Dates of Eastern Bluebird Breeding Season Based on Clutch Completion Os Os "n o 05 Os n ^ oo cn— — r-'iosor^fNvor-fNOo- oo — (N M ■“ (N rn tJ- fN f'' — fN OO < H iOsO^OOO'O^-CNir^, ^ VS ^ h oo av O P sOsOsOvOsOt^t-r^r'-r^r-'t^r'-r-'-r-'G© r- 0’s Qs O' O' O' O' O' O' O' O' O' O' O' O' O' O' 46 JEAN EAKIN No. 40 TABLE 13 Breeding Season of Eastern Bluebird: Percentage of Clutches Completed in Each 10-Day Period Dates Ohio (Peakall, 1970:246) Holden Arboretum 1965-80 1 1 Mar. -20 Mar. 0.20 21 Mar. -30 Mar. 0.17 0.00 3 1 Mar. -9 Apr. 0.33 0.81 10 Apr. -19 Apr. 9.52 5.69 20 Apr. -29 Apr. 18.03 15.85 30 Apr. -9 May 10.52 1 1.79 10 May-19 May 7.35 6.91 20 May-29 May 7.68 7.52 30 May-8 June 11.52 9.55 9 June-18 June 12.02 11.58 19 June-28 June 10.18 1 1.78 29 June-8 July 6.18 7.1 1 9 July-18 July 3.34 4.47 19 July-28 July 2.34 4.26 29 July-7 Aug. 0.33 2.03 8 Aug. -17 Aug. 0.406 TABLE 14 Tree Swallow Nesting Data r\ VO X ro r- O — X OO OO X OO X X O' r^- os X Tf X rt X OO fN X fN OO 'O Os 0 fN x On) fN 00 fN X X X r- x r- X X X " > <*— UJ >" X OX) <-•_ 0 <4— O 0 •0 2: 1— OJ X) E 3 umber [.aid 0) X) S 3 X CJ <3 c OJ O X 0 <3 3C 1— X E 3 C ‘ob "O Ll X 0 c5 X t— dJ X £ 3 X E 3 z z 2: CL 2 0- 2: 2 Number of Young Fledged: 396 48 JEAN EAKIN No. 40 TABLE 15 Computation of Tree Swallow Clutch Size Year Completed Clutches Eggs Laid Average Size 1966 1 5 5.00 1967 4 23 5.75 1968 6 25 4.16 1969 7 33 4.71 1970 9 44 4.99 1971 12 63 5.25 1972 1 1 57 5.09 1973 10 54 5.40 1974 12 57 4.66 1975 15 77 5.13 1976 10 48 4.80 1977 12 67 5.58 1978 8 41 5.12 1979 9 45 5.00 1980 17 82 4.82 Average clutch size for 15 years: 5.03. 1983 EASTERN BLUEBIRD 49 TABLE 16 Tree Swallow Breeding Season Based on Clutch Completion A/av 30- Year May 10-19 May 20-29 June 8 June 9-18 June 19-28 1966 1 1967 4 1968 3 2 1 1969 5 3 1970 3 4 1 1 1971 2 7 2 1 1972 7 2 2 1973 2 6 3 1974 2 8 1 3 1975 2 7 3 1 1976 1 5 3 2 1977 5 6 1 1 1978 2 2 4 2 1979 1 6 2 1980 3 10 1 TOTAL 20 67 39 17 3 Percentages of Clutches Completed in Each 10-day Period 13.81 46.53 27.08 11.8 .69 TABLE 17 Black-capped Chickadee Nesting Success o o be 7? s? & O Cu — — ti- £ o o .be S3 UJ 53 5; — — tj- Tt 3,5 uj1 -3 o rn r- Cl 3 O -c o o a*. QQ (N (N O “ O O > Cl ■g g- cs '-5 oo o c W -3 -O O >. r- oo & o 'O \C L- O' O' O' — rsif^Tfi^soi^ooa^o Os 0s O'' 0s O' 0s 0s O' < H o (- £ - o Q C/3 O u = 3.6; axial telson spine length, »2.6. Remarks Eller ( 1 935) based the species Echinocaris auricula on a well-preserved left valve from the Upper Devonian “Che- mung” Shale at Alfred Station, in southwestern New York state. This species has not previously been photo- graphically illustrated. The West Virginia specimen con- forms closely to Eller’s specimen and his description and sketch. Because the species is readily distinguishable from all others, there can be little doubt of its placement. The general outline of E. auricula is different from that of other members of the genus, particularly because of the pronounced anterior position of the hinge and concomi- tant prolongation of the posterior of the carapace. Addi- tionally, it is one of only two species that possesses three carinae on the carapace. The most closely related species is E. castorensis Copeland, 1 960, reported from the upper Devonian rocks of the Alexo Formation, Alberta, Can- ada. This species, as reconstructed by Copeland (1960: Fig. 1), is also characterized by the development of three carinae. Upon examining the type material we have been able to discern the centroventral ridge and the postero- dorsal ridge on these specimens, but have been able to discern only a suggestion of a very short posterocentral ridge. By contrast, all three ridges are well developed and distinct on E. auricula. The nodes on the carapace of E. auricula are generally similar to those of E. castorensis. The most notable difference in the nodes of these two spe- cies is the presence of a tubercle on the centroventral lobe of E. auricula and the lack of such a feature on the corre- sponding lobe of E. castorensis. The general outline of E. castorensis is also different than that of E. auricula, being less alate and having a more centrally located hinge. Relatively few species of the genus in addition to E. auricula and E. castorensis have a posterocentral ridge. These include: E. pulchra Sturgeon, Hlavin, and Kesling, 1964; E. randalli Beecher, 1902; E. socialis Beecher, 1884; and E. whidbornei Jones and Woodward, 1889. Among these, however, there are also other differences, such as the degree of development of lobes, presence or absence of tubercles, and shape of the hinge that serve to distin- guish these species from E. auricula. The above redescription allows for some intraspecific variation, especially in regard to the development of tub- ercles on the anterodorsal node and the exact nature of the carinae. Since Eller based the species E. auricula on a single specimen, he could not discuss variation and, in fact, little has been done regarding intraspecific variation 26 HANNIBAL AND FELDMANN No. 41 among members of the genus. It is certainly clear that some variation must, indeed, have existed and the devel- opment of minor tubercles and minute elements of orna- mentation on the carapace must be taken as well within the range of individual variation. Discovery and identification of this second specimen of E. auricula lends credence to the validity of the taxon. Ridges, grooves, carinae, and related structures on deli- cate arthropod skeletons are often the result of distortion during preservation. Very often it is difficult or impossi- ble to distinguish between actual ridged structures and artifacts of preservation. Further, different surface struc- tures can often be discerned, depending upon the degree of exfoliation of the exoskeleton. The discovery of a sec- ond specimen conforming very closely to the morphology of the holotype not only reinforces the original descrip- tion but also confirms that the structures described were structures actually present on the living organism. Addi- tionally, the West Virginia specimen provides a view of the abdomen and telson. Unfortunately, both are crushed substantially and very little can be said about the mor- phology of this region except that it seems to be more or less like that seen in better preserved specimens of some other Echinoearis species, notably E. punctata. Acknowledgments John L. Carter, of the Carnegie Museum, Pittsburgh, pro- vided the type specimen and counterpart of Echinoearis auri- cula. Murray J. Copeland, Geological Survey of Canada, Ot- tawa, made arrangements for the loan of type specimens of Echinoearis from Canada. Thomas Rammer, West Virginia University, and Kathleen Farago, Cleveland, aided in an inves- tigation of the West Virginia locality. Murray J. Copeland and W. D. Ian Rolfe reviewed the manuscript. The assistance of these colleagues is greatly appreciated. A portion of this work was supported by NSF grant EAR8312798 to Feldmann. Con- tribution 270, Department of Geology, Kent State University. References Beecher, C. E. 1884. Ceratiocaridae from the Chemung and Waverly groups of Pennsylvania. Second Pennsylvania Geo- logical Survey PPP:!-22. 1902. Revision of the Phyliocarida from the Chemung and Waverly groups of Pennsylvania. Quarterly Journal of the Geological Society of London 58:441-449. Cardwell, D. H., et al. 1968. Geologic map of West Virginia. West Virginia Geological and Economic Survey. 2 sheets. Cooper, G. A., et al. 1942. Correlation of the Devonian sedi- mentary formations of North America. Bulletin of the Geo- logical Society of America 53:1729-1794. Copeland, M. J. 1960. The occurrence of Echinoearis and Spath- iocaris (Phyliocarida) in western Canada. In M. J. Copeland and T. E. Bolton, Canadian fossil Arthropoda, Eurypterida, Phyliocarida and Decapoda. Geological Survey of Canada Bulletin 60:1-1 1. Dennison, J. M. 1971. Petroleum related to Middle and Upper Devonian deltaic facies in central Appalachians. American Association of Petroleum Geologists Bulletin 55:1 179-1 193. Eller, E. R. 1935. New species of Echinoearis from the Upper Devonian, of Alfred Station, New York. Annals of the Car- negie Museum 24:263-274. 1937. Echinoearis crosbyensis , a new species from the Upper Devonian of New York. Annals of the Carnegie Museum 25:257-259. Hennen, R. V., and D. B. Reger. 1914. Preston County. West Virginia Geological Survey. 566 pp. Jones, T. R., and H. Woodward. 1889. Notes on phyllopodi- form crustaceans, referable to the genus Echinoearis , from the Palaeozoic rocks. Geological Magazine, Decade III 1:393-396. Manspeizer, W. 1963. A restudy of the Chautauquan Series of Allegany County, New York. In V. C. Shepps, ed.. Sympo- sium on Middle and Upper Devonian stratigraphy of Pen- nsylvania and adjacent states. Pennsylvania Geological Sur- vey, 4th ser.. Bulletin G39, 4th ser., pp. 259-277. Rickard, L. V. 1975. Correlation of the Silurian and Devonian rocks in New York state. New York State Museum and Science Service. Map and Chart Series No. 24. Rickard, L. V., and D. W. Fisher. 1970. Geologic map of New York. Niagara Sheet. New York State Museum and Science Service. Map and Chart Series No. 15. Sturgeon, M. T., W. J. Hlavin, and R. V. Kesling. 1964. Rare crustaceans from the Upper Devonian Chagrin Shale in northern Ohio. Contributions from the Museum of Paleon- tology of the University of Michigan 19:47-64. Williams, H. S., and E. M. Kindle. 1905. Contributions to De- vonian paleontology. Part 1: Fossil faunas of the Devonian and Mississippian (“Lower Carboniferous”) of Virginia, West Virginia, and Kentucky. United States Geological Survey Bulletin 247. 58 pp. Woodward, H. P. 1943. Devonian system of West Virginia. West Virginia Geological Survey [Reports], 15:1-655. LIFE HABITS AND DISTRIBUTION OF RIVERINE LAMP SI LIS RADI A TA LUTEOLA (MOLLUSCA: BIVALVIA) MICHAEL J. S. TEVESZ DAVID W. CORNELIUS Department of Geological Sciences Cleveland State University Cleveland, Ohio 44115 and J. BERTON FISHER Research Center Amoco Production Corporation Tulsa, Oklahoma 74102 Abstract Lampsilis radiata luteola is an abundant, widely distributed, freshwater bivalve whose autecology and habitat preferences are little known. Analysis of populations from an Ohio stream reveal that it is most fre- quently found on compositionally mixed substrata, near shore, where current velocities are low. Although the life position of L. r. luteola is extremely variable, the most common orientation observed was with the shell approximately two-thirds buried in the substratum, the anteropos- terior axis at a 45° angle to the substratum, and the siphonal areas nor- mal to the direction of current flow. While L. r. luteola is an active burrower and crawler, its punctuated and variable mode of locomotion is in contrast to the more constant and predictable life habits of many comparably sized marine bivalves. Biometrical studies show that stream-dwelling L. r. luteola are larger than lake-dwelling forms. Also, the relation of shell length to age is nearly linear for individuals three to eight years old. The paucity of bivalves less than three years old indicates the L. r. luteola experiences sporadic re- cruitment and frequent reproductive failures. Introduction Lampsilis radiata luteola ( Lamarck, 1819) ( Unionidae) is one of the most widely distributed freshwater bivalves in North America. It occurs throughout the Mississippi and Missouri river systems, much of Canada east of the Rocky Mountains, the St. Lawrence drainage, and the Atlantic slope south to South Carolina. In addition, L. r. luteola is extremely widely distributed in aquatic envi- ronments within its geographic range. It is found in streams, ponds, lakes, and other wetland areas on sub- strata ranging from muds to gravel. Moreover, L. r. lu- teola is frequently the most abundant bivalve in these en- vironments (Burch 1973; Clarke 1973) and is a dominant organism in terms of standing crop and biomass. For ex- ample, unionid bivalves have the largest standing crop (82.5 g/m2 wet) and individual biomass (average individ- ual wet weight is about 10.5 g) of any invertebrate group in western Lake Erie, and L. radiata is the most abundant unionid (Wood 1953). It is therefore surprising that very little ecological in- formation has been published concerning this remarkable species. The literature concerning lentic populations mainly focuses on behavioral characteristics and biogenic modification of sediments (e.g., McCall et. al. 1979; McCall and Tevesz 1982). And with the exception of communications by Whittine (1969) and Salmon and Green ( 1 983) there is very little published information on the ecology and distribution of L. radiata in lotic environments. This paper is an attempt to begin building a body of knowledge concerning the autecology and distribution of lotic populations of L. r. luteola. In this paper, we provide new information on the distribution, biometrics, age, substratum preference, and life habits of L. r. luteola. The Vermilion River, Ohio, was chosen as the study area be- cause it is a relatively clean, accessible river, which, like many others, contains large populations of Lampsilis. The Vermilion River is 58.7 mi in length, has an average fall of 7.8 ft/mi, and drains an area of 271.7 sq mi (Ohio Division of Water 1954: Fig. 1). The river flows through mostly rural countryside, where its chief pollutants are silt and fertilizers from farms. The presence of a variety of pollution-intolerant organisms (e.g., stonefly and mayfly nymphs; caddis fly larvae) in the river over much of its length indicates that the river is mostly reasonably clean (Beck 1954; Gaufin and Tarzwell 1952, 1956). Methods Live and freshly dead L. r. luteola and associated Unionidae were collected in summer months during 1 975-82 at thirteen stations along the Vermilion River by bank-combing and wading (Fig. l;Stas. 1-7, 9-14). The general characteristics of the substratum were recorded for each station. In addition, four more stations (Fig. 1; Stas. 8, 15-17) were sampled in detail by wading and Fig. 1. Vermilion River. Scale (inset): ! in. = 5 mi. 1985 LAM PSILIS RADIATA LUTEOLA 29 Key to Figure 1 COLLECTING STATIONS 1 . Location: M ud Lake. At the intersection of U.S. Rt. 250 and U.S. Rt. 224, turn SE and travel 6.2 mi to the first right turn after Crum Rd., then due E 0 I mi; Ashland Co., Oh. 2. From the intersection of U.S. Rt. 250 and Crum Rd., go W on Crum 0.5 mi to St. Rt. 545. Turn right onto 545 heading to an overpass 0.4 mi NE; Ashland Co., Oh. 3. From the intersection of U.S. Rt. 250 and U.S. Rt. 224, take U.S. Rt. 250 SE for 5.0 mi to Clear Creek Rd. in Savannah, Oh. Go W on Clear Creek Rd. 0.8 mi to an overpass; Ashland Co.. Oh. 4. From the intersection of U.S. Rt. 250 and U.S. Rt. 224, take U.S. Rt. 250 SE 3.0 mi to Base Line Rd. Take Base Line Rd. W 0.5 mi to a steel truss span bridge; Ashland Co., Oh. 5. From the intersection of U.S. Rt. 250 and U.S. Rt. 224. take U.S. Rt. 250 SE 2. 1 mi to river overpass; Ashland Co., Oh. 6. From the intersection of U.S. Rt. 250 and U.S. Rt. 224, take U.S. Rt. 224 E 0.1 mi to river overpass; Ashland Co., Oh. 7. From the intersection of U.S. Rt. 250 and U.S. Rt. 224, take U.S. Rt. 250 N W 3.0 mi to Town Line Rd., then E on Town Line Rd. to a curve which parallels the river bed below, then hike 0. 1 mi up river; Ashland Co., Oh. 8. From the intersection of U.S. Rt. 250 and U.S. Rt. 224, take U.S. Rt. 250 NW 2.8 mi to steel truss span bridge, then upstream 0. 1 mi from bridge; Fluron Co., Oh. 9. From the intersection of U.S. Rt. 250 and Fitchville River Rd., take U.S. Rt. 250 E 0. 1 mi to bridge, and from there to 200 m up- stream; Huron County, Ohio. 10. From the intersection of U.S. Rt. 250 and Fitchville River Rd., take Fitchville River Rd. N 1.1 mi to Fayette Rd. Then take Fayette Rd. E 0.6 mi to a steel truss span bridge just W of Palmer Rd; Huron Co., Oh. 1 1. From the intersection of U.S. Rt. 250 and Fitchville Rivei Rd., take Fitchville River Rd. 5.3 mi N to Prospect Rd. Then take Prospect Rd. E 0.5 mi to steel truss span bridge; Huron Co., Oh. 12. From the intersection of U.S. Rt. 250 and Fitchville River Rd., take Fitchville River Rd. 8.2 mi N to Cook Rd. Take Cook Rd. SE 0.4 mi to a steel arch-truss span bridge; Huron Co., Oh. 13. From the intersection of Fitchville River Rd. and Cook Rd.. take Fitchville River Rd. 1.1 mi N to Zenobia Rd., then take Ze- nobia Rd. E 0. 1 mi and turn S (right), fol- lowing this road around a curve to a steel span bridge; Huron Co., Oh. 14. From the intersection of Fitchville River Rd. and Zenobia Rd., take Zenobia Rd. 0.7 mi E to a bridge; Huron Co., Oh. 15. From the intersection of St. Rt. 60 and St. Rt. 116 in Birmingham, Oh., take St. Rt. 1 16 E 1.3 mi to Gore Orphanage Rd. Then take Gore Orphanage Rd. N 1.8 mi to bridge; Lorain Co., Oh. 16. From the intersection of St. Rt. 60 and St. Rt. 1 16, take St. Rt. 1 16 E 1.3 mi to Gore Orphanage Rd.. then take Gore Orphanage Rd. N 2.4 mi to Morse Rd. Follow Morse Rd. to Bank Rd , then take Bank Rd. S 0.9 mi until it ends. From the dead end, hike due SW circa 0.5 mi to river; Lorain Co., Oh. 1 7. From the S end of Bank Rd., hike 0.8 mi NE to river; Lorain Co., Oh. SCUBA for live individuals only, and the following in- formation was recorded where each individual was found: substratum type (estimated as gravel, sand, or mud, or some combination of these); distance from shore; current velocity (measured with a submerged float and stop- watch); inclination of the anteroposterior axis of the bi- valve with respect to the substratum surface (estimated as either 0, > 0< 45°, 45°, > 45° < 90°, or 90°); orientation of posterior with respect to direction of current (measured with protractor); and fraction of shell buried in substra- tum (estimated as> 0. 15, 0.25, 0.33, 0.50, 0.67, or> 0.75). Over 1 30 live L. r. luteola were collected and the follow- ing data were taken in the laboratory: total length (great- est linear dimension parallel the hinge); anterior length (portion of total length anterior of the mid-point of the umbo); height (greatest linear dimension normal to length and in the same plane); and width (greatest linear dimen- sion across both valves normal to length and width). Ab- solute age determinations were made by cutting a valve in two with a rock saw, polishing one of the sections, and counting the annual markings under a microscope. Tax- onomic identifications were provided by Dr. David H. Stansbery, the Zoology Museum, Ohio State University, Columbus, Ohio. Distribution Areal The presence or absence, at 17 sampling stations, of L. r. luteola and associated Unionidae, is represented in Table 1 (see also Fig. 1). The predominant substratum 30 TEVESZ, CORNELIUS, AND FISHER No. 41 TABLE 1 Occurrence of Lampsilis radiata luteola at 17 sampling stations Station No. Predominant Substratum Type Sampled Bivalves Lampsilis radiata luteola (Lamarck, 1819) Lampsilis ventricosa (Barnes, 1823) Anodonta grandis grandis (Say. 1829) Strophitus undulatus undulatus (Say, 1817) Lasmigona costata ( Rafinesque, 1820) Fusconaia flava ( Rafinesque, 1820) Elliptio dilitata ( Rafinesque, 1820) Anodontoides ferussacianus (Lea, 1934) I mud X* 2 mud /sand X* 3 sand /mud X X* 4 mud/sand/gravel X* 5 mud X* 6 gravel X* X 7 sand/mud, with gravel X X X* X X 8 gravel, with pockets of X* X X X sand/ mud 9 bedrock/ gravel X* X X 10 gravel X* X X 11 gravel X* X X X X X 12 gravel X* X 13 sand /gravel X* X X X X 14 bedrock/ gravel X* X X X 15 gravel, with pockets of X* X X sand/mud 16 gravel, with pockets of X* X X sand /mud 17 gravel, with pockets of X* X X X sand /mud ♦Indicates most abundant species at particular site. type found at each station is also listed there. The table and figure demonstrate that Anodonra grandis grandis is the numerically dominant unionid at the sampling sta- tions in Mud Lake and at the source area of the river. But below Station 5 the numerically dominant species at the collecting locales is generally L. r. luteola , although Las- migona costata is the most abundant species at Station 7. Additionally, Table 1 reveals correlations between lacus- trine-conditions/muddy-substrata and the presence of a relatively species-poor fauna dominated by A. g. grandis (one to two species). Typical fluvial conditions and coarser substrata, on the other hand, appear to be asso- ciated with a relatively species-rich fauna (two to six spe- cies) generally dominated by L. r. luteola. Also correlated with this downstream increase in species richness is an increase in substratum heterogeneity. The fact that we have yet to discover Lampsilis living on soupy muds in the Vermilion River is interesting, be- cause it occurs with A. g. grandis on similar bottoms in Lake Erie (McCall et al. 1979; Tevesz and McCall 1979). This difference in distribution may be related to the smaller size and thinner shells of lake-dwelling forms (cf. Clarke 1973; Harman 1970) which would make them less prone to sinking into the substratum. Within- Habitat Table 2 provides more detailed information on the sub- stratum distribution of L. r. luteola and the five unionid species (combined) collected live at Stations 8, 15, 16, and 17. Of three “pure” substratum categories, L. r. luteola was most abundant on sand, compared to mud or gravel, was most frequently found on substrata containing a mix- ture of particle sizes, and was particularly common on a mud/sand mix and a mud/sand/gravel mix. The pattern of substratum preference also obtains for the remainder of the Unionidae observed at these stations. Information on the distance from shore and associated current velocities of all live-collected Unionidae at Sta- tions 8,15,1 6, and 1 7 is presented in Tables 3 and 4. Over 60% of L. r. luteola and other Unionidae are found within 1 m of stream bank as shown in Table 3. Over 75% of the live-collected Unionidae, including L. r. luteola, were found where current velocities were less than 5 cm/sec TABLE 2 Substratum Distribution (%) Mud Sand Gravel M/S G/S M/S/G Lampsilis radiata luteola 0 16 3 32 15 34 n = 134 Other Unionidae1 0 18 15 25 13 29 n = 55 1 Lampsilis ventricosa; Elliptio dititata ; Lasmigona costata; Strophitus undulatus urtdulatus; Anodontoides ferussacianus. 1985 LAM PS I LIS RADIATA LUTEOLA 31 TABLE 3 Distance from Shore Distance from shore (m) Lampsilis radiata luteola (%) n = 136 Other Unionidae (%) n - 52 1 61.8 61.5 2 14.7 21.2 3 8.8 7.7 4 7.4 5.8 5 6.6 3.8 6 0.7 0.0 TABLE 4 Current Velocity Preference Lampsilis radiata Other Current Velocity luteola (%) Unionidae (%) (cm sec) (n - 136) (n = 52) <5 77.9 75.0 >5 but <10 13.2 11.5 >10 but <15 2.2 3.8 >15 but <20 0.0 3.8 >20 but <25 0.0 0.0 >25 but <30 6.6 5.8 (Table 4). Thus, L. r. luteola appears to show definite preferences for certain microhabitats within the river at the sampling sites . It was most frequently found on com- positionally mixed substrata near shore, where current velocities are low. The same statement is true for the other live-collected Unionidae as a group. The within-habitat distribution of all these live-collected Unionidae at the sampling stations is highly similar. Life Position and Locomotion Table 5 presents information on the life position of L. r. luleola collected in situ. While life position of L. r. luteola is extremely variable, the most common orientation re- corded was with the shell approximately two-thirds buried in the substratum, the anteroposterior axis at a 45° angle to the substratum, and the siphonal areas normal to the direction of current flow. The life position and locomotory behavior of L. r. lu- teola from the Vermilion River and adjoining areas of Lake Erie were also observed in laboratory microcosms containing simulated native substrata and regulated to ambient temperature. The burrowing sequence of L. r. luteola observed for this laboratory study from both stream and lake population is similar, and is described by McCall et al. (1979). Individuals from both the stream and lake are active semi-infaunal crawlers. TABLE 5 Life Position of L. r. luteola Orientation of posterior (siphonal) area with respect to current (% of population). 0° — area oriented directly into current 0° 45° 90° 135° 180° n = 1 33 23 14 40 11 13 Orientation of anteroposterior axis with respect to sub- stratum (% of population). 0° >0° < 45° 45° >45° <90° 90° n=136 9 18 29 20 24 Fraction of shell embedded in substratum (% of population). n = 136 < 15 .25 .33 .50 .67 >.75 8 5 18 10 51 8 Burrowing rate indices (BRI’s) (Stanley 1970) for L. r. luteola are presented in Table 6. These data show that the BRI of L. r. luteola is variable, but by comparison to ma- rine bivalves of roughly similar size and shape, Lampsilis, over all, is a slow burrower (cf. Stanley 1 970). In addition, the burrowing rate is dependent on the type of substratum the clams inhabit. The rates on fine substrata are two to four times higher than rates on coarse substrata. This re- lation of BRI to substratum type is otherwise unreported for freshwater bivalves. The burrowing and crawling behavior of L. r. luteola was frequently interrupted by long periods of stasis. Usu- ally there was a lag time of several minutes to several days before animals placed in the laboratory microcosms be- gan locomotory behavior. The time involved to complete a burrowing sequence also took several minutes to several days, depending on the number of pauses, and duration and number of interspersed semi-infaunal crawling epi- sodes. This punctuated and variable mode of behavior is in contrast to the more constant and predictable habits of many shallow-water marine bivalves of comparable size studied by Stanley (1970). Comparable information for other freshwater species is scarce. The meager burrowing ability of Lampsilis, coupled TABLE 6 Burrowing Rate Indices Vmass (g) \ 7 : : : X 100 burrowing time (sec.) / Bivalve 1 Bivalve 2 Bivalve 3 Bivalve 4 (Length - (L - 69.8 (L = 62.4 (L=92.9 Substratum 56.1 mm) mm) mm) mm) mud 0.13; 0.13 coarse sand 0.057 0.049; 0.057 50% sand / 50% 0.031; 0.036; gravel 0.062 33% mud/ 67% 0.13 sand TABLE 7 Lampsilis radiata : Comparative Morphometries bo 7 >, o >> ON A ~ +1 ^ 5 Er t: ^ On ~C) ° o 5 5 Uj V) 5 3 So ^ 2- Co ■ 3: '-J +1 § ■S E ^ +1 ■5 E ^ +1 "? tjJ C Co ^ +1 Co g, Co o o +1 o CM o o © +1 o Co SO © +1 SO o o © +1 SO o o +1 l/~> so o o +1 so Tt OC © © +1 +1 r- — +1 +1 0 i c .2 I > so CM © © +1 7t © © +1 o so © o o +1 OS CM o © © +1 Os o o +1 o 4-1 uo sO O — t/7 4-1 o ro — O 4-1 o o 4-1 o m rn — — U~) 4-1 o tj- o -H o • — co O 4-1 o CM oo — St cd C o o rx x: £ O fe > 3 X> o3 X C/3 < ^ 5 2 o . . (U 7 x 6 CO o3 y o C/3 03 o> *J C sO O ■ — IC'J sd rf — ■Tf r- CM — CM oo o O m Os CM 4-1 4-1 4-1 4-1 4-1 K sd OO — o o o o oo sO V~) »/4 r* 4 o — ■ Os Os Os Os Os 7f OO Os CM 7f 7t CM oo V~) CM m oo sO CM O r- — >r o3 CQ i— O X) .2 ’ C j— u. c3 W c5 UJ cd C/3 W 0) > X (U 33 aj 0) < L) s w o3 j; cd U 03 .C3 s ~s u Sr "O ^ C oo 03 rc y Os i3 U c £ o CD "-5 ^ 5 1L R 5 Os O — 3 w Cr a3 ^ t— •— o3 ^ c ^ c3 > u II 1985 LAM PS I LIS RADI A TA LUTEOLA 33 with its preference for near-shore, shallow-water habitats associated with low-flow velocities, suggests that this bi- valve is not frequently physically removed from the sub- stratum by erosion. Further evidence of this is the absence of appreciable physical wear on almost all live-collected L. r. luteola, compared to the extensively abraded ap- pearance of loose valves collected on the sediment sur- face. L. r. luteola does not selectively reduce the forces exerted on it by currents. In fact, the preferred orientation is normal to flow. This suggests that the organisms are passively oriented by (low much like a flat rock. Biometrics Morphometrical information for L. r. luteola from the Vermilion River and other environments is pre- sented in Table 7. As shown here, L. r. luteola from the Vermilion River are morphometrically similar to other Lampsilis radiala riverine populations and those from ar- tificial impoundments. Further, these riverine/ impound- ment populations tend to have larger average dimensions than those from naturally occurring lakes. These observa- tions support the findings of several studies (cf. Clarke 1973; Harman 1970) who report that L. r. luteola and sev- eral other Unionidae are typically larger in streams than lakes. Additionally, comparative figures on the upstream (Sta. 8) versus downstream (Stas. 15-17) populations of L. r. luteola support the findings of Ortmann (1920) in that they show this species does not follow the “Law of Stream Distribution.” It should be noted, however, that the “Law of Stream Distribution” was mainly derived from study of large rivers in which habitat changes from shallow, rapid-flow environments to much deeper, slower -flow environments. The Vermilion is not such a river. Age Structure The age distribution of live-collected L. r. luteola is shown by the data in Table 8. Age data for all live- collected Unionidae are given in Table 9. The mean age of all live-collected L. r. luteola is 5.55 years, and few indi- viduals less than three years of age were observed. It is possible but doubtful that small individuals were missed during this study, because the length of three-year-old in- dividuals is only slightly less than 9 cm (Fig. 2) which is a size readily collectable by our techniques. Individuals between three and eight years of age show a nearly linear relation of age to mean length, as shown in Figure 2. Younger and older individuals are much smaller than would be predicted by the curve in Fig. 2, which indi- cates that the clams grow most during their second year of life and grow much more slowly after reaching eight years of age. The absence of young individuals has been noted TABLE 8 Age Distribution (%) of Lampsilis radiala luteola Aye (in years) Percent of Individuals (n = 133) 1 1.5 2 1.5 3 9.8 4 15.0 5 25.6 6 21.1 7 11.3 8 12.0 9 1.5 10 0.8 TABLE 9 Mean Age of Live-Collected Individuals Species Mean Aye Number of Individuals L. r. luteola 5.55 129 L. ventricosa 5.87 23 E. dilitata 5.90 1 1 L. costata 5.72 1 1 A. ferussacianus 5.20 5 S. undulatus 4.67 3 for Elliptio complanata in lacustrine environments ( Fisher and Tevesz 1976; Matteson 1948). The observed age structure for L. r. luteola suggests that recruitment is not continuous. This could be due to a number of factors Matteson ( 1948) noted that environ- mental stresses such as low dissolved oxygen and elevated temperatures can result in the death of both newly settled young and gravid females of E. complanata. as well as the expulsion of any maturing glochidia from gravid females. For L. r. luteola. males and females are present in approx- imately equal numbers; there does not appear to be evi- dence for differential mortality. It should also be noted that since L. r. luteola depend upon fish for the dispersal of their parasitic larvae, factors affecting the biology of the host organisms may be important in regulating the age structure of L. r. luteola. Detailed surveys covering a rea- sonably long period of time would be required to resolve this problem. References Beck, W. M., 1954. Studies in stream pollution biology. Quar- terly Journal of the Academy of Sciences 17:21 1-227. Brown, C. J. D., C. Clark, and B. Gleissner. 1938. The size of cer- tain naiades from western Lake Erie in relation to shoal expo- sure. American Midland Naturalist 19:682-701. 34 TEVESZ, CORNELIUS, AND FISHER No. 41 Age (yrs.) Fig. 2. Age/ mean-length relationships for 3- to 8-year-old L. r. luteola. Burch, J. B. 1973. Freshwater unionacean clams of North Amer- ica, biota of freshwater ecosystems. Identification Manual No. 1 1. EPA, U.S. Government Printing Office. 176 pp. Clarke, A. H. 1973. The freshwater molluscs of the Canadian in- terior basin. Malacologia 13:1-509. Cvancara, A. M. 1970. Mussels (Unionidae) of the Red River Valley in North Dakota and Minnesota, U.S. A. Malacologia 10:57-92. 1972. Lake mussel distribution as determined with SCUBA. Ecology 53:154-157. Cvancara, A. M., and P. G. Freeman. 1978. Diversity and dis- tribution of mussels (Bivalvia: Unionacea) in a eutrophic res- ervoir, Lake Ashtabula, North Dakota. Nautilis 92: 1 -9. Fisher, J. B., and M. J. S. Tevesz. 1976. Distribution and pop- ulation density of Ellipio complanata (Mollusca) in Lake Pocotopaug, Connecticut. Veliger 18:332-338. Gaufin, A. R., and C. M. Tarzwell. 1952. Aquatic invertebrates as indicators of stream pollution. Public Health Reports 67:57-64. 1956. Aquatic macro-invertebrate communities as indi- cators of pollution in Lytle Creek. Sewage and Industrial Wastes 28:906-924. Harman, W. N. 1970. New distribution records and ecological notes on central New York Unionacea. American Midland Naturalist 84:46-58. Matteson, M. R. 1948. Life history of Elliptio complanatus (Dillwyn, 1817). American Midland Naturalist 40:690-723. McCall, P. L., M. J. S. Tevesz. 1982. The effects of benthos on physical properties of freshwater sediments. In P. L. McCall and M. J. S Tevesz, Animal-sediment relations, 105-176. New York: Plenum Press. McCall, P. L., M. J. S. Tevesz, and S. F. Schwelgien. 1979. Sediment mixing by Lampsilis radiata siliquoidea (Mollusca) from western Lake Erie. Journal of Great Lakes Research 5:105-111. Ohio Division of Water. 1954 Gazetteer of Ohio streams. Ohio Department of Natural Resources, Division of Water. 1 75 pp. Ortmann, A. E. 1920. Correlation of shape and station in fresh- water mussels (naiades). Proceedings of the American Philo- sophical Society 59:269-312. Salmon, A., and R. H. Green. 1983. Environmental determi- nants of unionid clam distribution in the Middle Thames River, Ontario. Canadian Journal of Zoology 61:832-838. Stanley, S. M. 1970. Relation of shell form to life habits of the Bivalvia (Molluscs). Geological Society of America Memoir 125. 494 pp. Tevesz, M. J. S., and P. L. McCall. 1979. Evolution of substra- tum preference in bivalves (Molluscs). Journal of Paleontol- ogy 53:112-120. Whittine, A. H. 1969. Distribution of unionids in Hinkley Creek, Portage County, Ohio. Compass 46:142-149. Wood, K. G. 1953. Distribution and ecology of certain bottom living invertebrates of the western basin of Lake Erie. Ph D. diss., Ohio State University, Columbus. THE PREHISTORIC OCCUPATION OF THE HALE FARM BATH TOWNSHIP SUMMIT COUNTY OHIO DAVID S. BROSE Head Curator of Archaeology The Cleveland Museum of Natural History Abstract A prehistoric occupation at the Hale Farm site (33Su 1 7) underlay the 1809-27 occupation of a broad tributary floodplain of the Cuyahoga River. During the historical archaeology, four aboriginal pits/hearths and possible houses represented by two single post arcs of 5 m diameter were encountered in 36 m2 excavated over a 1 50 m2 area. These features, and the undisturbed strata in places, yielded a wide range of lithic tools and debitage with over 60 projectile points, intermediate between the styles of the ninth and fourteenth centuries A.D.. At least 1 1 handmade ceramic vessels were represented by over 500 grit-tempered, cord- marked and smooth sherds. These simple ceramic jars had rims with interior cordmarking, and flat lips which were either plain or notched. The ceramics are ancestral to the earliest Whittlesey ceramic types of the thirteenth century. Faunal and floral analyses indicate a late summer campsite with some evidence for maize and squash agriculture. Introduction In 1971 and again in 1980, archaeological investiga- tions were undertaken at the Western Reserve Historical Society’s Jonathan Hale Farm and Village in Bath Town- ship, Summit County, Ohio. The 1971 archaeological excavations, between June 9 and August 14, were supported by a grant from the Na- tional Endowment for the Humanities under my direction (H-4982). Work in the field was supervised by David Frayer and George Miller. Full-time crew members in- cluded Bert Barnard, Bruce Palmer, and Larry Ruben- stein. Part-time crew included A1 Hugley, Meg Conley, Elizabeth Coppedge and Harvey Yates (see Plate I). The limited 1980 excavation was performed by Donna L. Benson as part of her Ph.D. research. As Benson’s dis- Pl. I. 1971 excavations of the Hale Farm site view to east from 1827 house to Oak Hill Road. sertation committee chairman, this work, too, was done under my overall direction. The focus of these archaeological investigations was the historic occupation of the early nineteenth century. It was my aim to study the structural and material differ- ences between the Connecticut Land Company “settlers” and the non-landowning “squatters” whom the settlers frequently encountered occupying the best homesite in their often vast properties. We also hoped to learn some- thing about the rates of the economic effects and the material consequences due to opening the Ohio-Erie Canal through the Western Reserve in the 1820s. The Hale Farm had been occupied for several years by Abra- ham Miller when Jonathan Hale arrived from Connecti- cut in 1810. Hale continued to live in Miller’s cabin until 1826/1827. With the opening of the canal. Hale pros- pered. After solidifying his economic position, he built and occupied the large brick house now on the property, but continued using the cabin as an outbuilding until after 1 840. Excavation of the original cabin and analyses of the artifacts it contained seemed to offer an ideal opportunity to begin the study. The expectation concerning what archaeology could add to the history of the early Western Reserve era largely met with success. Although the final report of the 1971/ 1980 excavations has not yet been published, sev- eral articles have detailed various aspects of that archaeo- logical excavation (Benson 1978, 1980; Brose 1973a; Brose and Benson, 1985; Brose et al. 1981; Miller 1980, 1983). An unexpected result of the investigation was the dis- covery that not even Abraham Miller had been the first inhabitant. In 26 units which were excavated, we encoun- tered the remains of a prehistoric Indian campsite which had been abandoned and buried nearly 500 years before Columbus had set sail. To a degree. Miller and Hale, in their continual historic occupations of the area, had re- moved or disturbed much of this archaeological site which lay beneath their feet. Those storage pits, brick kilns, roads, cellars, and foundations which we had been expecting to find had often been dug through the thin prehistoric stratum. Of course, our digging for historical information was carried deeply to recover all of the prehistoric remains in any area where we worked. Yet our excavations were lim- ited; indeed, we dug into about 20% of the area over which prehistoric remains were found. Nonetheless, enough of the aboriginal camp remained relatively undisturbed that 36 DAVID S. BROSE No. 41 it was possible to carefully reconstruct many aspects of the economy, the technology, and the social lifestyles of these ancient Americans. These I have described in this report, and 1 will attempt to show when and with whom early people of Hale’s Farm were in contact and how these small self-reliant villages stood at a threshold: their ances- tors were the scattered hunting-gathering-gardening bands who had built the mounds and earthworks across Ohio in the first centuries of our era. Their descendants built the agricultural tribal confederations encountered by the European explorers of the seventeenth century. The Environment The prehistoric Hale Farm site (33Su 1 7) is located in the front yard of the present Jonathan Hale Farm on Oak Hill Road, lot I 1, township 3, range 12, in Bath Town- ship, Summit County, Ohio. The site area was estimated from the recovery of prehistoric artifacts and from the thin, discontinuous patches of prehistoric strata and pits. From the north bank of Hale Run the prehistoric site ex- tended about 30 m northward to what had been a smaller east-flowing branch of Hale Run. East of the present 1826/27 Hale House the prehistoric materials appeared concentrated in a roughly 30 to 45 m zone centering about 25 m west of modern Oak Hill Road. A few prehistoric ceramic sherds and one broken chipped stone projectile point were recovered washing out along the north bank of Hale Run about 30 m east of the road but the intervening area had been disturbed by the clay pits and brick burning activities of Hale himself in 1826. Within these 1378 m2, between the branches of Hale Run, the topography is rel- atively level (see Fig. 1), sloping genty to an old terrace east of the road, and further leveling out. The geological deposits in this portion of the Cuyahoga Valley have been actively reworked by glacial ice, by lakes of ponded melt water from retreating glaciers, and for the past 10,000 years the region has been controlled by the Cuyahoga River which flowed first south and then north along sev- eral different channels (Brose et al. 1981; Hall 1980; Rau 1968; White 1953a, 1979; Whiteand Totten 1982; Wittine 1970). The slopes behind the Hale Farm are composed of rather level Mississippian and Pennsylvanian shales and sandstone formations, deposited in near-tropical seas about 200 million years ago. Over the past million years, the area was scoured by glacial ice, and filled by clays, sands, and gravels as the ice melted. The last glacial ice left large gravel ridges called “kames” along its valley mar- gins. Just southeast of the Hale Farm a series of hills were formed when a channel of the lower Cuyahoga flowed south along the western valley wall to join the upper river at Akron about 1 4,000 years ago. With lowered water lev- els in Fake Erie about 3,000 years later, the lower Cuya- hoga flowed north along the east side of the present val- ley. It captured the upper Cuyahoga at Cuyahoga Falls. The old Oak Hill Road channel, filled with clays, silts, and gravels, was covered by a veneer of sandy soils washed from the western valley slopes. Furnace Run and Yellow Creek began to build a level floodplain by sea- sonal flooding, while they, with some even smaller tribu- taries, began to dissect the old glacial deposits into smaller and less steep hills and terrace remnants. By 4000 b.c. Fake Erie and its tributary valleys had reached their modern configuration. By that time also the modern post- Glacial climatic patterns and the developed soil associa- tions had established the mixed hardwoods forest com- munities which were recorded by the first surveyors (Andreas 1980; Gordon 1966, 1969). The Hale Farm site is located at the juncture of several different soil associations. The prehistoric site itself lies upon nearly level areas of Fitchville silt loams. Small areas of Orrville silt loams and moderately eroded Glenford-Geeburg silt loam and Furay loam soils occur on the slightly steeper knolls and terrace formations to the north and south along Oak Hill Road. On the steep slopes to the west of Hale Farm, soils are predominantly those of the Ellsworth-Mahoning association. Along the Hale Run valley east of the Kame Terrace and knolls east of Oak Hill Road are the poorly drained soils of the Conotton-Oshtemo complex. Over most of the level por- tions of the floodplain between Hale Run and Furnace Run, east to the Cuyahoga, the soils are Glenford and Chagrin silt or Chili silt loam, while between Hale Run and Yellow Creek most floodplain soils are either gravelly Chagrin silt loam or the slightly alkaline Chili silt loam ( Ritchie and Steiger 1 974). J’hese soils are of a markedly different agriculture potential. In an unimproved state their yield in bushels of corn per acre would run from over 90 for the Chagrin loam to 70 to 75 for the Fitchville- Orrville silt loam, to less than 50 for the Ellsworth soils (Ritchie and Steiger 1974). Differences in the quality of soils were certainly important to Jonathan Hale, and may also have been significant for the location of the prehis- toric site itself. The typical soil profile of the 2% to 6% Fitchville soils which underlie the Hale Farm site is described as having a 18 cm dark grayish-brown plow zone layer of slightly acid, well-drained silt loam; about an 8 cm layer of light yellowish-brown, strongly acid silt loam; about 20 cm of yellowish-brown and iron-stained acidic, poorly drained, silty clay loam; and a final soil layer of as much as 80 cm of brown, moderately acid silt loam. Below this 150 cm depth lay the sands and gravels of the Pleistocene lake and river deposits. (Ritchie and Steiger 1974:78). In general this soil is seldom flooded and moderately well-drained. It presents little problem for occupation. At the time of its first prehistoric habitation, the Hale Farm site would probably have represented a small, rather open and grassy area with scattered yellow poplar, white oak, and hickory. To the west the slopes would have Fig. 1. Location of the Hale Farm site (33Sul7). 38 DAVID S. BROSE No. 41 been covered with a beech-maple forest with hemlock in the ravines. To the east the lower flat floodplains would have had a mixed forest with elm, ash, yellow poplar, lo- cust, walnut, and basswood, while the old Kame Terrace hills would have been covered with black oak, red oak, beech, and maple. (Braun 1950, 1955; Gordon 1966; Hor- ton 1961; Williams 1936, 1949). Numerous flowers, shrubs, and vines covered much of the forest floor save on the open oak lands. Along the runs and rivers, reeds and rushes grew thickly. Within this wilderness lived a variety of animals includ- ing bears, wolves, and cougars. The largest mammal in these forests was the wapiti or American elk, but it is likely that elk would have been found only scattered widely within the valley during the warmer seasons (see Murie 1951). The most common big game in the site area would have been the whitetail deer, while cottontail rab- bit, squirrel, and raccoon would have been far more nu- merous (ODNR 1975). The Excavations Initially, the discovery of the prehistoric occupation at the Hale Farm was fortuitous. The location and nature of the excavation units had naturally been planned to re- cover information about the nineteenth, rather than the ninth century a.d. Although prehistoric materials were recovered at the base of even the first excavation unit, the investigation of this early occupation was always limited by the historical focus of the archaeological investigations undertaken for the Western Reserve Historical Society. It was further constrained in part by the nature of distur- bance due to the subsequent occupations of the area by Miller and Hale. Thus prehistoric structural data recov- erable were also limited. The upper surfaces of the four aboriginal features encountered along the north side of Hale Run had all been truncated by nineteenth-century activities. Few continuous stratigraphic profiles connect the bases of these shallow storage pits or cooking hearths. Most of the aboriginal materials were recovered from the fill of storage pits, foundations, and cellars built between 1810 and 1826. We also discovered that during the continuous histori- cal occupation of that area from 1826 to as late as 1957, there had been occasional excavations for new fences, for cisterns, for driveways, and to rebuttress the slumping banks for Hale Run. Not only had these more recent ac- tivities disturbed the early historic and prehistoric levels, they had brought to the surface some of the buried mate- rials which were thus reincorporated into the developing topsoil. It was clear enough that a chipped stone projectile point, an 1835 half-dime, and an aluminum wire-drawn nail did not belong to a single occupation. While those artifacts could be assigned to an appropriate period, that was not possible for fragments of charred bone, or for the empty small postholes which we encountered in disturbed areas of the site. Since we cannot know to which period such items pertain, they have been excluded from the fol- lowing discussions of the prehistoric component of the site. All archaeological excavations at the Hale Farm site were based on a 5 ft square grid system (1.52m) oriented north-south, with the front of the 1 826/ 27 Hale House as its western boundary. The actual units chosen for excava- tion were determined by a combination of three types of information. First, various historical sources suggested that the original Miller cabin was located just north of Hale Run and due west of Oak Hill Road. Therefore a series of fifteen excavation units in that area were ran- domly chosen for testing. Secondly, there was a vague leg- end that the 1870 barn which had been located north of the present driveway, was built over the original occupa- tion. In that area ten excavation units for testing were chosen in order to avoid obvious recent disturbance. Fi- nally, a series of ten test units were chosen to investigate otherwise unaccountable depressions, knolls, or topo- graphic irregularities and to “fill in” the coverage of the area west of Oak Hill Road. Excavation began by stripping the recent sod from within each test unit. Below that level the excavations proceeded with shovel and trowel to remove soils by a 10 cm depth or to the point where we noted any changes in soil color or texture which might represent a feature such as the top of a pit or the edge of a wall. At that point a measured drawing of the floor of the test unit was made, photographs and samples were taken, and the new feature was given a field provenience number which was assigned to all material recovered in association. Any concentra- tions of artifacts or biotic remains were treated as separ- ate features. All soils after removal by trowel were screened through 1/4 in or 3/32 in hardware cloth. The decision to expose all features at the level they were first encountered often required excavation of adjacent units and resulted in a rather strange pattern of excavation un- its (Fig. 2). Prehistoric archaeological materials were encountered in 26 of the units excavated (Fig. 3), as well as along the northern bank of Hale Run. Yet only three areas of test units revealed prehistoric archaeological features. To the immediate north of the driveway, below the twentieth- century disturbance resulting from the removal of the 1870 barn, a small shallow pit had been dug into the silty clays. This pit, about 61 cm in diameter and 20 cm deep, contained a broken chipped stone point and a single turtle carapace. Possibly the turtle shell had once been used as a bowl. In parts of three adjacent excavation units just west of Oak Hill Road, another aboriginal pit was located. It had been cut into by one of the posts from the original fence line erected by Hale prior to 1826. Even more shallow than the first pit, it was about 76 cm in diameter, while less than 1 5 cm at its deepest. In it we found a concentration of Fig. 2. Excavation units at the Hale Farm site. Fig. 3. Areas of prehistoric materials at the Hale Farm site. 40 DAVID S. BROSE No. 41 26 prehistoric ceramic fragments from two different vessels. The greatest concentration of prehistoric remains was encountered along the edge of Hale Run between Oak Hill Road and the 1826/27 house. We had excavated two sets of contiguous trenches which crossed at the location of the 1810-27 root cellar below the original cabin. We found a refilled aboriginal storage pit, an aboriginal hearth, eight postholes, and portions of a sheet midden (a scatter of artifacts and charred remains) upon the ground surface where Hale Run had flowed some ten centuries ago. The scale drawings of the north-south trench side- wall stratigraphy, reproduced as Fig. 4, show clearly that these postholes and the associated midden underlay, and thus predate, the early nineteenth-century occupation. However, that nineteenth-century activity had removed large portions of the original site. We were able to expose and record the earlier prehistoric levels only along the floor of our east-west trench (Fig. 5). The western set of five postholes and one pit may repre- sent a small, more or less round structure about 2.4 to 3.5 m in diameter. If the eastern set of three postholes were the remains of another similar structure, then most of it appears to have been destroyed by Miller when he built his cabin. In the middle of the 1971 digging season, there did not appear to be any significance to the distribu- tion of different types of prehistoric material within this area of the site. And yet, we were primarily concerned with the historic remains, and even at that time we had found as much or more prehistoric material redeposited in historic features. It is nonetheless clear that this area, just north of Hale Run, was the core of the prehistoric site occupation area — just as it was to be the core of Miller’s occupation and of Hale himself until 1827. Prehistoric Material Recovered During the excavations at the Hale Farm a variety of prehistoric artifacts and waste was recovered. These came from thin undisturbed aboriginal floor levels, and from the few pits dug for storage or as fire hearths and later filled with trash. We also encountered a larger number of similar prehistoric materials from soils which were dis- turbed and redeposited by Miller or Hale in the early nineteenth century. Altogether there were 153 chipped stone tools, along with over 4,000 waste flakes, many of which were used; ground and polished stone tools and ornaments; several drilled stone and clay beads; over 500 fragments of aboriginal pottery; and nearly 1 70 floral and faunal remains. The detailed analyses of these materials have taken the Fig. 4. Stratigraphic profile of trench at the Hale Farm site. 1985 PREHISTORIC HALE FARM OCCUPATION 41 545 E 255 ^) POST HOLE ABORIGINAL MIDDEN Fig. 5. Floor plan of prehistoric stratum in East West trench at the Hale Farm site. better part of four years by staff members and volunteers of the Cleveland Museum of Natural History, and while these details may seem rather uninteresting and overspe- cialized to the general reader, such studies represent the basic data from which archaeologists have been able to reconstruct much of the economy, technology, and social geography of these early inhabitants of the Hale Farm site. And through the stylistic comparison of artifact de- sign and decoration with similar materials from sites across the Great Lakes/Ohio Valley region, it is possible to assign rather closely a date to this occupation. This is in spite of the fact that the nature of the nineteenth-century activities seem to preclude recovering any amount of charred aboriginal organic material which could be sub- mitted for a direct radiocarbon determination of the date. Floral am! Faunal Remains Altogether, over 2,700 fragments of broken, butchered, or burned animal bone, and 56 pieces of charred plant material were recovered from the 1971 excavation. Most of these were either found in historical or in disturbed contexts. Thus they can have no significance for the un- derstanding of the prehistoric site occupation. From what appeared to be stratigraphically intact prehistoric fea- tures such as postholes or refilled storage/cooking pits, we recovered fragments of only 160 animal bones and only 9 charred plant remains. The animal bone is described in Table I. Although most of the available meat comes from the two elk, it is clear that a wide range of animals was being used for food. The elk, represented by skull fragments and limb bones primarily, may have been killed and butchered away from this location. The young bear must have been taken be- tween spring and late fall. The deer were represented by the jaw of a yearling and by parts of the skull and several ribs and proximal limbs of an adult. The antlers of the adult deer and one elk had been shed, suggesting that they were taken between late winter and summer. Tooth erup- tion and wear on the yearling deer indicate an age of about 15 fnonths (Severinghaus 1949). Most fawns are dropped between May and late June (ODNR 1975). The turtle and the fish remains found, along with the pas- senger pigeon bones, are frequently indications of early spring or late summer hunting. All in all, analysis of the animal bones (viz Cleland 1966) reveals that most of the prey were probably hunted between July and September (see Cleland 1966, 1976; Ziegler 1973). This interpretation is complemented by analysis of the small number of plant remains. These consisted of a single charred half of a hickory nut shell, three wild grape seeds, three charred squash or pumpkin seeds, and two frag- ments of what might have been a single cut cob of ten- or twelve-row maize. While all of these plant products could have been harvested between late August and early No- vember, the hickory nuts could have easily been stored for a long time. The maize cob could have been discarded just after harvest or could also have been dried. However, based upon the historic accounts (Ford 1977); one might expect only the kernels to have been stored. The grapes, normally ripe in late September, could not have been kept long given the damp winters of the valley. The plant and animal remains together suggest that the site was occupied in the late summer. During that season hunting parties brought back two elk, taken from some other part of the valley. All of the plants and most of the animals remains recovered should have been available within the varied environment represented by the Cuya- hoga River itself, its western valley slopes, and the tribu- tary floodplains of Furnace Run and Yellow Creek, bi- sected by Hale Run along which the site was built. 42 DAVID S. BROSE No. 41 TABLE 1 Animal Bone From Prehistoric Context At The Hale Farm Site Species Number of bones present Minimum Number of Individuals Usable Meat for Individuals Total Pounds Usable Meat Per Species Percentage of Usable Meat at Site American Black Bear immature 4 1 75 lb 75 10.5% Whitetail Deer adult 40 2 80 160 22.0% immature 13 I 40 40 5.5% American Elk adult 21 2 200 400 56.0% Racoon 13 3 1.5 4.5 0.5% Beaver 2 2 7 14 2.0% Passenger Pigeon 5 2 0.5 1 — Turkey 6 2 10 20 3.0% Turtle 17 1 0.5 0.5 — Suckers 39 7 1 7 1.0% Total 160 23 — 712. 100.5% Stone Artifacts and Debitage There were several different sources of stone from which Hale Farm tools were made. Most of the unfin- ished chips and cores are debitage representing the knap- ping of the Bois Blanc, Columbus /Delaware, or Onon- daga cherts which occur as nodules of chert in the Silurian and Devonian limestone and dolomite strata forming the Niagara escarpment. The outcrops nearest to the Hale Farm would be those in Monroe County, Michigan, in Essex County, Ontario, or in Erie County, New York. Yet cobbles and pebbles of these types of chert are quite common throughout northeast Ohio as stream gravels washed from the glacial deposits pushed south during the late Pleistocene. A number of finished tools and a surprising amount of the unused debitage recovered at the Hale Farm site came from working tabular Plum Run, Pipe Creek, and Upper Mercer flint which occurs in bands and as lenses inter- bedded with the Mississippian strata of shale and coal in eastern Ohio. A number of these outcrops occur in Stark, Erie, and Coshocton counties, Ohio, at distances from 30 to 75 km from Hale Farm. Many of these outcrops show signs of aboriginal quarrying and several have yielded the waste from the initial stages of knapping (Stout and Schoenlaub 1945). The lithic categories employed in the analysis of the 153 tools recovered at Hale Farm, follow those outlined by Brose ( 1967b, 1978c), with the aim to define functional arti- facts on the basis of selective micro-wear studies and the reconstruction of manufacturing sequences. Table 2 illus- trates the frequencies of formal chipped stone tools recov- ered from the Hale Farm site, and indicates whether tabu- lar flint or nodular cherts were the source. A similar frequency illustration of the debitage, or technological byproducts of the knapping activities are presented on Table 3. By combining both sets of information, it has been possible to consider the alternative technological se- quence which shows major differences from later sites of the general period. Alternative sequences of lithic manu- facture have been recognized at functionally differing site types within this same temporal phase (Brose 1976a, 1978a, 1978c). These sequences involve alternative sources of raw materials, differing initial core preparation TABLE 2 Absolute Frequencies of Formal Lithic Tools at Hale Farm Functional Tool Category Irreg. Blades Bifacial Scrapers Drills Triangular Projectile Points Triangular Knives Flake Gravers Bifacial Knives Wedges Gouges Unifacial Scrapers Subtotal Total Lithic Source TF TF NC TF NC TF NC TF NC TF NC TF NC NC NC TF NC TF NC 7 2 5 1 2 24 43 5 14 0 0 7 19 2 6 3 7 53 100 153 Key TF: Tabular Flint NC: Nodular Cherts 1985 PREHISTORIC HALE FARM OCCUPATION 43 TABLE 3 Lithic Debitage From All Units at Hale Farm Site Reduction Stage Tabular Flint Nodular Cherts Total N % N % N % Pebble cores used 2 0.05 0 0 2 0.05 unused 9 0.22 0 0 9 0.22 Block cores used 0 0 3 0.07 3 0.07 Decortication unused 0 0 2 0.05 2 0.05 Oakes used 1 0.02 99 2.44 100 2.46 unused 282 6.95 400 9.85 682 16.80 Primary shatter used 6 0.15 14 0.35 20 0.50 Secondary flakes unused 283 6.97 286 7.05 569 14.02 with platform used 13 0.32 25 0.62 38 0.94 w/platform unused 576 14.19 875 21.56 1451 35.75 without platform used 1 1 0.27 30 0.74 41 1.01 unused 545 13.43 596 14.68 1141 28.1 1 TOTALS 1728 42.6 2330 57.4 4059 100% SUBTOTAL used 33 0.81 171 4.22 204 5.03 unused 1695 41.76 2159 53.19 3854 94.95 and utilization, differing application of heat treatment, differing reconstructed preform, blank, and finished arti- fact sequences with different functions for morphologi- cally similar artifacts. Some attempt was made to identify, in quasi-ethno- graphic terms, the inferred function of these “finished ar- tifacts” and the quantities of utilized debitage. The degree of detailed investigation needed for much statistical con- fidence in such identification is unfortunately not present in the sample used in this study. Nonetheless it has been possible to reconstruct a model of the lithic reduction strategy used at the Hale Farm (Fig. 6). Hard hammer and soft hammer percussion were the major manufactur- ing technique at all stages, although it was more com- monly employed in the earlier portions. Pressure retouch occurred as the final technique for many of the final tool types, although it occurred earlier for specific artifacts and was altogether absent for some. Thermal alteration, never common in any Whittlesey tradition lithic assem- blage (viz Brose 1973b, 1980, 1985a, 1985b), occurred at several different times in the reduction sequence, and again for some tools seemed to be absent entirely. The products of this lithic strategy, whether unfinished pre- forms, flake debitage, or finished formal artifacts, dis- played variations in their indication of utilization such that less than 20% of the functional stone tools at the Hale Farm consist of utilized debitage and utilized interme- diate reduction sequence forms (Brose 1973b), while in smAll FLAKES LARGE FLAKES TABULAR (QUARRIED) FLINT TABULAR CORE I NODULAR (GLACIAL) CHERT I PEBBLE CORE DECORTICATION FLAKES SMALL BLANK IRREGULAR BLADES LARGE BLANK SMALL FLAKES LARGE DECOR- TICATION FLAKES SMALL FLAKES Fig. 6. Reduction sequence as reconstructed. 44 DAVID S. BROSE No. 41 other coeval sites these categories may represent over 75% of the functional lithic assemblage (e.g., Brose and Scarry 1 976). Many of the intermediate reduction forms also ap- pear to have been used to duplicate the functions of some of the heavy-duty formal tools (e.g., triangular preforms were utilized as knives, large decortication flakes were utilized as gravers). There is considerable redundancy within this system: however inefficient such a system may appear, the possible alternative sequences made it rather stable and very flexible. Known archaeological sites of this early Late Wood- land period show little demographic differences between what are seen as summer multifamily horticultural vil- lages (along alluvial floodplains in mixed mesophytic forests) and spring or fall fishing-fowling sites (along the lakeshore at river mouths) (Brose 1976b, 1978c). Mid- winter occupation sites are known in upland rock shelters or at spring locations. These appear to represent smaller family segments engaged in a rather focused hunting econ- omy. Summer villages such as the Hale Farm site reveal a mixed economy. The site shows a utilization of lithic sources which approximates their relative local availabil- ity. At winter hunting camps, however, glacially derived local chert nodules were selected over quarried sources. These differences in the lithic tool assemblages and in the reduction strategies between summer villages and winter campsites seem relatively minor, although in the winter campsites most cores from tabular quarried flint received thermal alteration and appear to have been used predom- inantly for irregular blade production. The limited ther- mal alteration of nodular glacial cherts appears late in the reduction sequence at many winter campsites and, where present, seem restricted to immediate pre-pressure re- touch of triangular preforms into projectile points and hafted knives. At the Hale Farm site thermal alteration appears to have been applied to about 60% of all quarried flint prior to any knapping and to about 35% of all nodu- lar chert following its reduction to a large rough blank. The relative frequencies of functional tool categories are significantly different from those at winter sites. At this summer occupation there are several varieties of small speical purpose tools, numerous projectile points, and a number of small scrapers. However, at winter occu- pations (viz Brose and Scarry 1976; Lee, Brose, and Weisman 1983; McKenzie et al. 1974) the tool assemblage was characterized by fewer different tool types. There was also a lower frequency of projectile points, and a large number of knives and heavy scrapers. In general, the lithic assemblage at summer villages reflects greater evidence for what Binford (in Whallon 1 974) has termed curatorial activity, and for a much wider range of extractive activi- ties than do winter campsites. Summer sites also reflect an industrial sequence with more carefully chosen raw mate- rials, more evidence for thermal alteration and at earlier and less artifact-specific production stages, and a higher frequency of nonspecific utilized debitage than in winter campsites (Brose 1978c). TABLE 4 Metric Attributes (mm). Unifacial Scrapers, All Units Hale Farm Site ( 33Su 1 7) Catalogue Location Maximum Length Maximum Width Maximum Thickness Shape H73BD 30.65 16.55 5.35 Lunate 9576 HI 16S 29.55 22.85 12.85 Thumb-nail 9618 H170HG 19.45 14.25 5.10 Triangular H78AC 21.45 14.65 4.30 Triangular 9590 H136AK 19.0 14.45 5.40 Rhombridal 9506 H20EE 21.85 10.95 4.95 Ovate 9508 H23DK 13.25 8.30 3.35 Thumb-nail 9561 H 101 AP 16.2 14.65 2.65 Rhombridal 9586 H 1 3 1 A 22.75 13.65 5.75 Ovate 9584 H135KY 20.0 15.65 8.85 Convergent Beyond this there is little in the assemblage of chipped stone tools from Hale Farm of great interest (PI. II). The unifacial scrapers (Table 4), bifacial scrapers (Table 5), triangular preforms and knives (Tables 6 and 7), or drills (Table 8) all seem similar to those from contemporary sites such as the Bass Lake and Columbia Road sites, and to those from components representing the early phases of the Whittlesey tradition after a.d. 1200 (Brose 1973b, 1976b, 1978a, 1980). Assuming that drills were for drilling and scrapers were for scraping, and recognizing that the differing frequencies of different tools probably reflect TABLE 5 Metric Attributes (mm) Bifacial Scrapers from Hale Farm Site (x = Broken) Catalogue Location Maximum Length Maximum Width Maximum Thickness Shape 9523 H44C 25.45 24.35 9.95 Triangular *9535 H65B 18.60 19.65 3.95 Sub-triangular 9549 H86AV 35.60 30.0 6.85 Sub-triangular 9536 H67AQ 28.0 20.25 10.35 Ovate 9500 H12S 26.20 23.10 10.0 Ovate *9605 H134BB 26.20 12.85 4.0 Lunate 9505 H 1 38 Y 20.55x 15.85x 4.45 Triangular PI. II. Chipped stone artifacts from the Late Woodland component at the Hale Farm site. 46 DAVID S. BROSE No. 4! TABLE 6 Attributes (mm) of Triangular Preforms/ Bifacial Knives (x = Broken) Catalogue Location Maximum Length Maximum Width Maximum Thickness Striae, Grinding, Beveling or Edge Damage Present? 9536 H67AH 33.65 29.35 9.85 Yes 9536 H67AQ 28. Ox 20.45x 11.0 None 9618 H170HG 20.35x I4.35x 5.15 None 9618 H170AG 23.20x 30.25x 11.40 None 9530 H58U 28.90 21.25 7.85 None 9538 H69AW 44.75 22.15 7.45 None 9618 H170AC 21.95 23.85 10.10 None 9618 HI70AG 35.35 22.45 11.55 None *9555 H94C 35.95 22.75 7.10 Yes 9656 H214KT 29.0 20.55 8.55 None 9656 H214KR 32.0 29.9 7.75 None 9626 H180D1 23.25 14.55 8.10 None 9578 H120N 28.85x 20.0 9.0 None 9519 H33S 31.55 15.85 6.75 None 9519 H33S 31.55 15.85 6.75 None 9519 H38CK 21.45 22.50 8.65 Yes 9596 H I44DI l!.35x 17.15 7.65 None 9508 H23BV 41.35 30.55 9.35 None 9584 H27A 38.1 15.4 7.85 None 9617 H16BJ 17. Ox 1 1.45x 4.85 None 9574 HI 10AC 13. Ox 9.10 3.60 None 9644 lO.lOx 17.85 5.55 None 9618 H170AF 17.25 25.45x 7.0 None 9623 HI77GR 22.0x 14.45 6.45 None 9547 H82E1 59.10 34.55 12.55 None 9541 H172AF 23.35x 23.20 9.45 None TABL E 7 Attributes (mm) of Triangular Knives (x = Broken) Catalogue Location Maximum Length Maximum Width Maximum Thickness Striae, Grinding, Beveling or Edge Damage Present 9560 H100B 26.75 20.6 4.45 Yes 9561 H 101 AD 25.85 18.85 3.25 Yes 9560 H100A 27.05 16.25x 3.65 None 9560 H100G 30.95 21.65 5.55 Yes 9560 H100D 28.0 19.65 6.75 None 9560 H100E 43.25 28.75 10.0 Yes 9679 H175A 18. Ox 14.0 4.65 None 9545 H78AB 14.65x 15.35 5.90 None 9642 H204CP 20.55x 18.85 7.65 None 9618 H170AD 25.65 23.25 7.55 None 9614 H 166 AH 18.55x 15.35 4.25 None 9647 H204EQ 13.75x 14.85 4.45 Yes 9622 H176HI 13.75x 16.20 5.75 Yes 9530 18.95 13.65x 3.75 None 9565 HI 05 A 35.25x 21.65x 4.55 Yes 9569 H 1091 26.55x 21.45 7.35 Yes 9564 H104H 22. 8x I3.55x 4.0 None 9563 H103FY 14.9x 14.7 3.35 None 9566 H106DI 12.25x 13.4 3.25 Yes differing site-to-site task emphases, there does not seem to be much major difference in the tools themselves during the early Late Woodland in this area except for the pro- jectile points (Table 9). Traditionally, in late prehistoric sites, the most diag- nostic chipped stone artifacts have been the triangular projectile points, and variations in width and base shape have been used to order assemblages in time. The single broadest triangular projectile point does have a concave base, and the single narrowest triangular projectile point has a straight base. However, the statistical analysis of 1985 PREHISTORIC HALE FARM OCCUPATION 47 TABLE 8 Metric Attributes (mm) of Drills From Hale Farm Catalogue Location Point Type Max. Length Max. Width Max. Thickness Shoulder To Base Base Width Min. Haft Width Haft To Base 9499 HI 105 46.5 19.7 8.0 16.1 19.7 6.6 30.4 9532 H61AL 18.55 8.25 3.20 — — — — 9532 H61AM 32.15 10.85 10.10 TABLE 9 Attributes (mm) of Triangular (x = Broken) Projectile Points Catalogue Location Maximum Length Maximum Width Maximum Thickness Base Type Catalogue Maximum Maximum . Maximum 9530 Location Length Width Thickness Base Type H-58 19. lOx 14.0 3.65 Concave 9644 9644 H201 lO.Ox 17.65 5.50 Straight H201GA 14.95x 10.25 3.20 Concave 9679 9643 H 175 17.85 14.15 4.74 Concave H200HE 12.65x 1 1.35 3.15 Straight 9622 *9618 H176HI 13.1 5x 16.25 5.75 Straight HI70C1 30.1 19.9 6.6 Straight 9570 *9612 H-136AR 14. 1 5x 19.25 5.35 Concave HI63AE 26.2 17.7 4.2 Straight 9545 9609 H78AB 14.55x 15.55 6.0 Straight H160BC 21.6 14.4 4.6 Straight 9545 9612 H78AD 30.65 11.45x 5.65 Straight H163AF 1 l.Ox 9.2x 2.1 x X 9536 9617 H67AH 33.45 25.40 9.85 Concave H 169 19.8 15.7 2.8 Straight 9614 9629 H 166 AH 18.65 15.1 5x 3.95 Straight H183FU 24.6 15.7 4.4 Straight 9618 9623 H170AF 1 7. 1 5x 25.35x 6.95 Straight H177GD 12.4x 17.7 4.1 Concave 9574 9629 HI 10AC 12.70 9.20x 3.45 Concave H183MN 19.1 1 !.5x 3.4 Straight 9647 *9635 21.9 19.8 3.7 Straight H204CP 25.35x 14. Ox 5.10 Straight 9629 9618 H187AA 18. 3x 17.9 6.3 Concave H170AC 16.0x 26. lOx 9.65 Straight 9595 9623 H143D 27.3 18.1 5.6 Straight H177GR 1 3.55x 22.35x 7.65 Concave (?) 9586 9647 H131D 21.75 23.75 5.5 Concave H204EO 13.75x 13.65x 4.35 Concave 9589 9651 H135LU 16.25x 13.5 4.75 Straight H208GR 16.85x 15.90 5.95 Straight 9647 9656 H204CR 25.20 12.40x 5.55 Straight H214KP 17.95 12.85x 3.65 Straight 9545 *9656 H78AD 29.25 1 I.OOx 5.45 Straight H214K.U 16.75 12.60 3.85 Straight 9548 9552 H840 24.10 15.10 4.70 Straight H90D 20.20x 16.20 3.40 Concave 9597 9537 H145DG 24.45 13.70 4.55 Straight H65? 19.15 10.35 4.45 Straight 48 DAVID S. BROSE No. 41 Catalogue Location Maximum Length Maximum Width Maximum Thickness Base Type 9509 H24A 12.35.x 17.0 2.75 Concave H20ED 9.80x 22.70 4.35 Concave 9644 H20IGP 1 5.65x I5.35x 4.60 Concave 9644 H20IGA 14.95x 10.25 3.20 Straight 9643 H200HE 1 2.65x 11.35 3.15 Concave *9618 H170C1 30.1 19.9 6.6 Straight *9612 H163AE 26.2 17.7 4.2 Straight *9619 H172AE 36.2 29.8 9.1 Straight 9609 H160BC 21.6 14.4 4.6 Straight 9612 H163AF 1 l.Ox 9.2x 2.7x X 9617 H 169 19.8 15.7 2.8 Concave 9545 H78AC 18.85x I4.45x 3.95 Straight *9592 H145DG 23.75 13.80 4.40 Straight *953 1 H57A 25.75 22.35 3.65 Concave *9535 H65A 26.95 20.75 5.50 Straight 9530 H158EX (9608) 18.10 14.9 5.55 Concave 9512 H30FF 20.55x 12.55 4.35 Straight 9511 H28 1 18.55 14.50 4.75 Concave 9519 H38CZ 17.0 14.75 3.45 Straight 9513 H31AR 25. Ox 21.45 6.75 Concave 9574 HI ION 30.25x 30.10 5.10 Straight 9574 HI I0P 22.20x 22.0 4.80 Straight *9574 HI I0Q 30.85 28.80 4.55 Straight *9574 H110R 28.10 23.65 4.90 Straight 9577 HI 19Z 19.25x 1 5.85x 5.0 X 9656 H214K.Q 15.25 9.85 5.10 Straight 9656 H214KV 27.25 6.65 4.45 Straight 9656 H214KS 22.35 15.0 6.20 Straight length/width ratios show that the complete straight base points (n = 23; x = 1.53; s.d. = 59) and the complete con- cave base points (n = 7; x = 1.20; s.d. = 1.08) both lie within the normal length/ width distribution for all 30 unbroken points (x = 1.45; s.d. = 0.63). That is, it is statistically proper to consider these all as representing the normal variation within a single population. Although virtually identical projectile points were first formally described as Madison points by Scully (1951), the most detailed typo- logical analysis of such projectile points in the Great Lakes area is that by W. Ritchie ( 1 97 1:13—14, PI. XVI; see also Converse 1964; Luedtke 1978). Ritchie had described the Madison point in New York as having replaced the somewhat larger and broader Levanna point with their more concave bases. Levanna points were common from a.d. 900 to a.d. 1300 (Ritchie 1971:31). Morphologically all of these Late Woodland projectile points from the Hale Farm site are transitional but are best seen as an early varient of the New York Madison points rather than as a late variant of Levanna type (cf. Fitting et al. n.d.TO). This suggests the Hale Farm points (PI. Ill) should date to between a.d. 1000 and a.d. 1200. As at most late archaeological sites in northeast Ohio, not all of the chipped stone knives/points were as recent as the site itself (Brose 1973b, 1975, 1976a, 1980). A number of notched Archaic projectile points (some broken) were recovered from Hale Farm (Table 10). Most of these are not very diagnostic in terms of the specific temporal periods they represent, but the resemblance to the types Otter Creek (PI. Ill, bottom row, second) to Normanskill (PI. Ill, bottom row, third) or Lamoka (PL III, bottom row, fourth) suggest they may all have been made between 4000 and 2500 b.c. (Ritchie 1971:29-30, 37-38, 40-4 1 , Pis. XIII, XIV, XVIII, XX, and XXI). Cer- tainly there are numerous Middle/ Late Archaic sites in this portion of the Cuyahoga Valley which were occupied at that time (Brose 1975; Brose etal. 1981). A final class of chipped stone tools recovered from the Hale Farm are represented by nine whole and fragmentary split shale or slate knives (Table 11). All display crudely chipped curved sides with battered and often scratched edges, pre- sumably as the result of heavy cutting use of some sort (Keeley 1974, 1980). Similar split shale knives were noted at the Boston Ledges shelters (Brose and Scarry 1976; Read 1880) which yielded ceramics similar to the Hale Farm site. In addition to the chipped stone materials, the Hale Farm site excavations yielded four drilled limestone and fired clay beads; three fragmentary drilled gorgets, two of slate and one of sandstone; and a single celt (axe or adze) of granodiorite (Table 1 2). Not one of these artifacts is par- ticularly diagnostic; all have a wide distribution in time and space within the aboriginal cultures of the eastern Woodlands. Other than the celt these artifacts were found in somewhat disturbed contexts. The celt was found em- PI. III. Chipped stone projectile points from the Hale Farm site. 50 DAVID S. BROSE No. 41 TABLE 10 Metric Attributes (mm) of Archaic Points, All Units Point Attribute Chart Min. Depth of Catalogue Maximum Maximum Maximum Shoulder Base Haft Haft to Bifurcation Location Length Width Thickness to Base Width Width Base or Indent Grinding or Beveling Present? 9521 H42CW 9547 51.35x 41.95 6.85 7.20 20.55 7.55 6.2 No H82EK 9547 20.0 18.60 8.45 10.5 3.35 3.35 0 No H82EN 17.40x 23.45 8.55 6.10 20.90 13.60 8.0 Shoulder margins and haft edges ground 9500 H21H 13.85x 15.95 3.45 9.55 9.55 3.10 Base and edges ground 9640 H195A 9558 48.25 14.20 9.00 5.5 14.2 9.85 3.85 No H97A 39.90x 38.10 8.70 11.95 30.10 21.95 14.90 2.95 Side notches ground 9578 H120M 34.65x 23.85 7.10 5.20 17.10 14.0 5.65 Side notches ground bedded within the thin aboriginal midden level encoun- tered just northwest of the cabin foundations along Hale Run (see Fig. 5). Aboriginal Ceramics Previous Whittlesey ceramic typology has been both confusingly inclusive, inconsistent, and imprecise. This TABLE 11 Split Slate Knives (in mm) Striae, Grinding, Beveling or Catalogue Location Maximum Length Maximum Width Maximum Thickness Edge Damage Present? 9545 H78A 45.0 16.25 9.25 Yes 9578 H120P 21.65 23.0 6.85 Yes 9578 H120T 31.0 22.75 9.55 None 9622 H176HD 63.7 32.3 10.6 Yes 9615 H167T 40.75 23.20 6.25 None 9525 H98J 47.65 31.0 9.25 None 9511 H26AV 49.55 47.90 9.80 Yes 9509 H214KO 23.40 16.45 5.95 Yes 9615 HI67G 45.9 23.2 8.8 Yes point had earlier been made by Murphy (1971b) who pro- vided the first workable descriptions for many of the Whittlesey ceramic types. I have expanded upon and modified several of the type names proposed by Morgan and Ellis (1943), Fitting (1964) and Murphy (1971a, 1971b), in an effort to look for significant changes in ce- ramics through time and space. While not yet demonstra- ble that the type varieties proposed will reveal everything we always wanted to ask about the late prehistory of northeast Ohio, it is unfortunately true that previous ce- ramic seriations based upon the previous ceramic typol- ogy have not revealed much at all. For example, the mis- named type Fairport Plain was first loosely defined by Fitting ( 1 964: 1 64) who added the statement that most of these ceramics were cordmarked to the lip, although some were plain. Fitting also noted that both plain and notched lips occurred, and, although he reported that there were variations in the temper of Whittlesey ceramics, he did not use temper in his type definitions (1964:162). Based on the ceramics recovered in his 1968 excavation at Fairport Harbor, Murphy (1971a) suggested that there did not appear to be any significance to the distributions of the ca 30% shell-tempered (and mixed shell-and-grit) ceramics at Fairport Harbor in terms of surface finish, or lip mode or decoration, a position supported in his more analytical definition of the type Fairport “Plain” (Murphy 1971 b:299— 300). The term Fairport Plain, with or with- out quotation marks, is still incorrectly inclusive. It is my position here (see Brose 1980) that the distinction between plain or smoothed vessel surfaces and cordmarked vessel surfaces, is likely to carry some degree of archaeological information although whether chronological or socio- geographical is yet unknown. Indeed the use of overly in- clusive and/or borrowed ceramic terms (e.g., Brose 1985 PREHISTORIC HALE FARM OCCUPATION 51 TABLE 12 Miscellaneous Ground and Polished Stone Artifacts Catalogue Location Description Max. L. (mm) Max. W. (mm) Max. T. (mm) Inside Diameter of Drilled Holes (in mm) 9612 HI 63 AC/ 2 portions of 3 hole 37.1 29.6 4.4 3.9 (incomplete) H163AD Slate Gorget 29.1 31.9 4.4 4.6and 5.4(broken) 9656 H214KN Broken 1 hole Gorget of Sandstone 97.65 47.0 12.45 6.2 9647 H-204AA Broken Slate Gorget 51.2 40.4 9.10 _ 9595 H-143A/ Broken clay bead 15.85 7.55 6.45 3.5 H143C Broken clay bead 16.55 7.55 6.65 3.5 9630 H-184CE Complete clay bead 22.9 9.0 8.0 4.1 9630 H-184X Broken limestone bead 12.65 7.9 7.0 3.8 9645 H-202A Granodiorite Celt 110.65 72.95 32.5 — 1973b) is likely to conceal such potential data. The type name Fairport Harbor Plain should be reserved for the plain vessels previously included in the old Fairport Plain of Fitting and the Fairport “Plain” of Murphy. This leaves the type name Fairport Harbor Cordmarked to refer to the cordmarked ceramics previously included by those authors. I have further used the type-variety system as propounded by Phillips (1970) to determine possibly significant varieties within each type, distinguished by the presence (var. Painesville) or absence (var. Willoughby) of the horizontal row of punctates below the vessel lip. The variable lip notching noted by all authors, from Greenman (1935a, 1935b, 1937) to Murphy (1917b), appears to be a freely varying vessel mode in the sense used by Phillips ( 1970( 1):55). As such it deserves descrip- tion, but until its significant distribution in time and/or space is demonstrated, need not be given classificatory status. A resolution for the issue of the ceramic tempering material is less obvious in the late prehistoric ceramics of the Lake Erie region. There seems little justification for cluttering the literature with a plethora of ceramic type and variety names redundant in all save the nature of in- clusive tempering agent when there seems to be a free vari- ation along a continuum from wholly grit-tempered through mixed shell-and-grit-tempered, to wholly shell- tempered ceramics which differ in no other attributes (cf. Brose 1973b; Brose et al. 1976). Indeed, shell tempering seems to appear relatively early in the seriation of Whit- tlesey ceramics (say after about a.d. 1200), at which time its relative frequency may be as high as 20% of some types (Brose 1976b, 1978a, 1980, 1984, 1985a, 1985b). There is not much apparent change in frequencies through time, for the latest dated Whittlesey components also yield about 15% to 20% shell and mixed shell-and-grit temper- ing. Nonetheless, the high relative frequency of shell tem- pering for some types at the Fairport Harbor site itself (Murphy 1971b) or at some structures at the South Park site (Brose 1978b, 1985b) suggests that with closely con- trolled ceramic lots the frequency of shell tempering may represent significant social patterning, if not also being temporally significant (but probably not temporally diagnostic). Using the former Ohio State Museum collections of Greenman (1935a, 1935b, 1937) and Morgan ( 1943), Fit- ting ( 1 964: 1 65ff) developed a seriation from early to late: Fairport Plain to Reeve Opposed-Reeve Horizontal to Tuttle Hill Notched. He concluded that the site sequence was Fairport Harbor, Reeve, South Park, and Tuttle Hill. This sequence, and indeed much of the ceramic seriation, was based on the assumption of single componency for the sites. Murphy clearly recognized the fact that multiple occu- pations had occurred at South Park (1972), at Fairport Harbor (1971a), and at Lyman (1971c). Recent investiga- tions (Belovich and Brose 1982; Brose 1973b, 1975, 1976a, 1976b, 1985b; Brose, Wentzel, et al. 1976; Brose, White, and Ford 1983; Bush 1982; Lee 1982) have confirmed Murphy’s recognition and added data to support the multi-component nature of the Reeve and Tuttle Hill sites also. Lacking the stratigraphic control revealed by subse- quent extensive excavation, Murphy (1972) seriated his South Park ceramics into four typological components. From early to late these were represented by a component characterized by Mixter series and what Murphy called Glen Meyer ceramics (1972:33); a component character- ized by Reeve Horizontal, his newly defined Reeve Fil- leted, and crude Parker Festooned ceramics; a compo- 52 DAVID S. BROSE No. 41 nent characterized by Reeve Opposed and Horizontal and possibly by his recently defined Fairport Filleted ce- ramics; and a final component dominated by Tuttle Hill Notched ceramics. Based upon this study, and upon his earlier typological remarks (1971b), I infer that Murphy felt the Whittlesey ceramic seriation might be dichoto- mized into a Cuyahoga Valley and a somewhat different Fake Shore sequence (but with considerable overlap at times). Murphy’s suggested overall sequence, which seems to have given little significance to the relative frequencies of shell temper (rightly, I believe), began prior to a.d. 1300 with Mixter and Glen Meyer as external ceramic intro- ductions into a local assemblage of Fairport “Plain” and Reeve Horizontal. This was seen as followed by the de- velopment of Fairport Filleted and Reeve Opposed ce- ramic types, possible as a response to the introduction after a.d. 1400 of Parker Festooned from the west and McFate Incised from the east. Reeve Filleted had devel- oped from Fairport Filleted after a.d. 1400, and by a.d. 1600 this had in turn developed into Tuttle Hill Notched, the latest local ceramic type. Murphy recognized that there were morphological varieties in each of these Whit- tlesey ceramic types and suggested that these might have chronological significance. As he stated, “All such defini- tions of formal ceramic types are inevitably subjective and are always subject to future refinement. The same is true of inferred relationships and suggested ages of these types” ( 1 97 1 b;298). Although the decade since his study has indeed pro- vided data for typological refinement, has raised some questions concerning the cultural relationships, and has seen further excavation at stratified sites yielding revi- sions to his suggested ages, the general sequence Murphy proposed in 1971 has been supported. Murphy’s typology was artificially limited at its earlier end, due to the circumstances that no early Whittlesey sites had been identified. The recent excavations at a number of small early Fate Woodland sites in Cuyahoga, Lake, Geauga, and Summit counties ( Belovich and Brose 1982; Brose 1975, 1980; Brose and Pratt 1976; Brose and Scarry 1976; Brose, White, and Ford 1983; Bush 1982; Fienga and Lee 1982; Lee 1982) have shown that the earliest Whittlesey ceramics developed from local Wood- land antecedents in the period between a.d. 1000 and a.d. 1200, during the Riverview phase, as previously defined by lithic seriation (Brose 1978c). The Jonathan Hale Farm site lithic materials were a critical component of that Riverview phase, and the abo- riginal ceramics form an assemblage which by seriation should also fall in the period between a.d. 1000 and a.d. 1 150. The 1971 and 1980 excavations at the Hale Farm recov- ered just over five hundred fragments of aboriginal pot- tery (Tables 13 and 14; Pis. IV and V). The ceramic as- semblage from the Hale Farm represents 1 1 different TABLE 13 Distribution of Surface Finish on Prehistoric Ceramics from the Hale Farm Site (Rimsherds/Body sherds) Exterior Surface Treatment Interior Surface Treatment Cordmarked Plain Brushed or Fabric Impressed Total Cordmarked 1/6 14/446 0/0 15/452 Plain 0/0 3/0 0/0 3/0 Brushed or Fabric Impressed 0/0 0/52 0/0 0/52 Total i/6 17/498 0/0 18/504 vessels. Nine of these, grit-tempered cordmarked jars with no plastic modifications, are typical of the general early Late Woodland period in the Great Lakes Region. Two of the vessels have the notched, thickened lip or folded rim strip or collar which occurs both in some regional early Late Woodland ceramic assemblages at about a.d. 1000, and which, in many varieties, seems to have been consid- ered a characteristic of Whittlesey focus ceramics, usually assigned to this period a.d. 1300 to a.d. 1600. The type Cuyahoga Cordmarked, is represented by five rimsherds, apparently from three different vessels at the Hale Farm site. It was previously defined based on ceram- ics recovered from two rock-shelters located in Boston Ledges along Richie’s Run, about five miles downstream and across the Cuyahoga Valley (Brose and Scarry 1976: 134-136). All Cuyahoga Cordmarked ceramics rep- resent small somewhat outcurved rim, semiconoidal to subglobular jars with massive grit tempering. Flat, slightly thickened lips show a slight exterior bevel and are about 9. 6 mm thick and the rims below the lip are from 8.2 to 12. 6 mm (x = 9.6 mm) in thickness. Vessel exterior TABLE 14 Summary Statistics for Thickness of Prehistoric Ceramics from the Hale Farm Site (in mm) Sample Vessel Portion Size Mean S.D. C. V. A. Lip 18 6.072 1.044 0.172 Neck 17 7.458 1.054 0.141 Shoulder 16 8.425 1.764 0.209 Body 195 10.704 8.995 0.840 B. Body Sherd with brushed or Fabric Impressed Exterior and/or uniform curvature 45 12.691 2.482 0.196 Other Body Sherds 150 9.167 3.556 0.388 PI. IV. Aboriginal ceramics from the Hale Farm site; exterior surfaces. PL V. Aboriginal ceramics from the Hale Farm site; interior surfaces. 1985 PREHISTORIC HALE FARM OCCUPATION 55 surfaces are decorated with vertical to slightly oblique overlapping impressions of a paddle tightly wrapped with a two ply z, z, S cord. Vessel interiors show direct or slightly dragged z, z, S cordwrapped paddle edge or stick in a zone from the lip to a depth of about 40 mm (Brose and Scarry 1976:134-136). From the Hale Farm site one Cuyahoga Cordmarked vessel with no additional decoration clearly corresponds to that type description and is here assigned the name Cuyahoga Cordmarked var. Boston Ledges. While iden- tical in most attributes to the previous type description its unmodified lip is slightly thinner in thickness (x = 7.8 mm), lying at the lower limit of the type distribution. Two additional Cuyahoga Cordmarked vessels from the Hale Farm site are also represented by unmodified lip modes but display the single horizontal row of circular punctua- tions about 25 mm below the lip. These vessels of Cuya- hoga Cordmarked var. Hale also have a somewhat thinner (x = 6.8 mm), as well as a somewhat more out- curved, neck area. In these attributes they differ from the typical Cuyahoga Cordmarked type in the direction of the Fairport Harbor Cordmarked type and seem to represent a typological and temporal intermediate ceramic variety. Fairport Harbor Cordmarked is represented by one vessel each of var. Painesville, plain lip mode, and var. Painesville, notched lip mode; by three vessels of var. Wil- loughby, plain lip mode; and by one var. Willoughby, notched lip mode vessel. For all six of these Fairport Harbor Cordmarked vessels, the ceramic paste is con- torted and no coil breaks are evident. Temper consists of crushed granitic grit with some apparent selection for lighter, acidic angular minerals such as quartz and plagio- clase. Temper particle size ranges from 0.8 mm to 1 .8 mm with a mean of about 1.2 mm. Temper density is relatively high, possibly representing between 20 and 30% by vol- ume. Thickness at the flattened lip ranges from 5. 1 mm to 6.6 mm with a mean thickness of 6.2 mm. Just below the slightly extruded lip, the neck ranges from 6. 1 mm to 8.3 mm in thickness with a mean of 6.8 mm for these vessels with their minimally thinned, minimally outturned rim profiles (Fig. 7). While rim diameters could not be deter- mined from these small sherds, similar ceramics from the Fig. 7. Rim profiles of aboriginal ceramics from the Hale Farm site (interior to right), with recon- struction of small vessel of the type Fairport Harbor Cordmarked, var. Painesville. 56 DAVID S. BROSE No. 41 earlier Fairport Harbor site collections of Morgan, Ellis, and Murphy range from just under 24 cm to over 55 cm. All exterior vessel surfaces have been malleated from base to lip with a rather coarse, loosely cordwrapped paddle. Ceramic cord impressions are predominantly ver- tical in orientation. They appear almost entirely to be rep- resentative of s, s, Z twisted two-ply cords, although a variety of different sized elements are present on the dif- ferent vessels. All vessel interiors are smoothed, and often shallow horizontal finger impressions of the potter can be felt. The three vessels of Fairport Harbor Cordmarked var. Painesville are decorated by a single horizontal row of circular or annular punctates around the upper neck. These punctates range from 3 to 7 mm in diameter, are from 8 to 1 5 mm apart, and from 14 to 40 mm below the lip. Occasionally the punctates were deep enough to create interior bossing and, rarely, they completely pene- trated the inner wall (although the lack of regularity for this attribute on single rimsherds suggests that this was a production mistake rather than a deliberate style). One of the Fairport Harbor Cordmarked var. Painesville vessels has a series of exterior shallow lip edge notches made by the transverse impression of a rounded tool, and one var. Painesville vessel has a plain lip mode. The Fairport Har- bor Cordmarked var. Willoughby vessels, without a row of punctations, are also represented by one vessel with a plain lip mode and with two vessels having the notched lip mode. One grit-tempered vessel of the type Fairport Filletted (Murphy 1 97 1 b:299) is represented by two rimsherds from the Hale Farm site. The folded rim (sometimes called a thickened lip or collar) is weakly developed, ex- tending downward from the exterior lip for about 8.5 mm and being only about 3.8 mm thicker than the flattened lip of 6.5 mm. The thickest portion of the exterior folded rim is transversely notched with narrow, fingernail-like im- pressions. The lip itself has a single shallow longitudinal grove incised along its center. Apart from this unusual lip mode the vessel appears similar to ceramics Murphy re- covered for the Cleveland Museum of Natural History in 1968 which he felt (correctly, I believe) were intermediate between Fairport “Plain” and Fairport Filleted (Murphy 1 97 1 b:299). I have previously proposed two varieties of the type Fairport Filleted, distinguished by the presence (var. Hillside) or absence (var. Fairport) of vertical finger trailing below the rimfold to the shoulder (Brose 1980, 1984). The Fairport Filleted vessel from the Hale Farm could thus be considered var. Fairport with an incised rim mode. I am not comfortable with this pigeonhole, but only the recovery of further examples of such ceramics in significant context would justify the establishment of a new variety. The final aboriginal vessel from the Hale Farm site is represented by one grit-tempered, smoothed or plain sur- face rimsherd with a flat notched lip and a single horizon- tal row of circular punctates. This is similar to many of the plain ceramics with the notched lip mode which had been previously subsumed within the type Fairport Plain (Fit- ting 1964) or Fairport “Plain”(Murphy 1971a, 1971b). To follow the terminological treatment for the cordmarked majority of those ceramics, this plain vessel should be called Fairport Harbor plain, var. Painesville. These aboriginal ceramics can all be considered as a single prehistoric cultural assemblage, assignable to be- tween a.d. 900 and a.d. 1200. They thus represent the ce- ramic complex transitional from the still poorly under- stood early Late Woodland period to the later prehistoric Whittlesey focus, or tradition (Brose 1973b, 1976a, 1976b, 1978a, 1980; Fitting 1964; Greenman 1937; Griffin 1946, 1967; Murphy 1971b). The earlier post-Hopewell societies show widespread ceramic similarity throughout the Midwest. The later Whittlesey tradition represents a geographically localized complex, similar in many ways to those protohistoric Iroquoian and Algonkian groups encountered by the European explorers. The period dur- ing which the prehistoric occupation of the Hale Farm oc- curred is thus of considerable importance in understand- ing the cultural changes associated with the development of tribal agricultural societies in North America. The few archaeological sites known prior to a.d. 750, in northeast Ohio show little evidence for any significant dif- ferences in the size or composition of the groups which occupied them, although these sites clearly seem to be sea- sonally and functionally different. None are clearly agri- cultural villages although corn and squash are present at several. During the past several years a number of early Late Woodland archaeological sites quite similar to the Hale Farm site have been excavated or reported upon. These now permit some understanding of the period between a.d. 800 and a.d. 1300 in northeastern Ohio. At the Columbia Road site atop a steep ridge overlook- ing the Cuyahoga River Valley, test units uncovered four cultural features, two of which were fire pits (Belovich and Brose 1982). Most projectile points were notched but a few were triangular. The plain or cordmarked grit- tempered ceramics had straight or slightly excurvate rims with flat lips, some slightly rounded. Over half of the rims displayed interior cordmarking. Exterior rim decoration consisted of a single nail-punctate sherd. Two radiocar- bon dates were a.d. 1 040 ± 80 (DIC #2605), and a.d. 970 + 60 (DIC #2606). Our analyses suggest the site was a family campsite utilized during the summer. The Bass Lake site (Fienga and Lee 1982) appears to be a rather large plant-collecting and fishing camp located on the north shore of Bass Lake, headwaters of the Cha- grin River. Excavation revealed 26 oblong, basin-shaped pits, and single radiocarbon date (DIC #2457) of a.d. 1200 ± 45 was obtained from charcoal. 1985 PREHISTORIC HALE FARM OCCUPATION 57 The projectile points in situ include a few Corner Notched and Levanna types, as well as numerous points intermediate between Levanna and Madison. The ceram- ics are grit-tempered, with cordmarked exteriors and plain interiors. Rims are everted, or slightly incurving, and may have a folded incipient collar. Vessel bodies are rounded. Decoration, restricted to the exterior neck and rim, to the lower edge of the collar, and to the lips of the vessels, includes incising, punctates, and cordwrapped cord impressions, broadly comparable to Wayne (Halsey 1 968) or to Allegan and Spring Creek wares between a.d. 800 and a.d. 1 350 in Michigan (Brashler 1981); to the a.d. 900 to 1200 Clemson’s Island complex of Pennsylvania; or to the Carpenters Brook phase of New York (Ritchie 1969) . At the Kernisky village site, on the lower Chagrin River, recent excavations (Bush 1982) have revealed what appears to be a floodplain terrace horticultural village with the remains of several structures. The ceramics are cordmarked on the exterior only. They are grit-tempered or grit-and-shell-tempered, with slightly everted rims and flat, slightly rounded lips, most of which are notched. A few ceramic vessels from the earlier Cleveland Museum of Natural History testing in 1971 and 1979 show a folded incipient collar. All of the recovered projectile points could be considered examples of the Madison point type. Although neither the completed analyses nor the absolute dating are yet available, in my judgement the lithic reduc- tion and ceramic tradition seen at the Kernisky village should place the major occupation late in the Fairport phase, dated somewhere between a.d. 11 50 and a.d. 1350 (Brose 1 978c:90— 9 1 ). In 1974, as part of an archaeological survey program for the Cleveland Sewer District, excavations tested and completely exposed a small prehistoric campsite about 25 km to the north of Hale Farm. A remarkably similar as- semblage of stone tools and ceramics was encountered in the refuse-filled hearths and postholes of two circular to oval single post structures on the second terrace of the Cuyahoga River (Brose and Pratt 1976:3-8). While no animal bone was preserved at that site, a charred cob of eight-row maize was recovered. With no organic remains appropriate for direct dating, I speculated that the site was occupied between a.d. 900 and a.d. 1300 based on the styles of its projectile points and ceramics. It also seemed related to an early occupation on the high bluffs to the west where still earlier test excavations had revealed an overlying large late prehistoric village on Tuttle Hill (Greenman 1937). When Tuttle Hill was itself destroyed by the construction of the 1-77 / 1-480 interchange, salvage excavation recovered a single burial associated with grit- tempered, exterior and occasionally interior cordmarked, notched-lip ceramics typical of that early Late Woodland campsite. These were dated to a.d. 1048 ± 100 (CWRU- 14). While no direct dates are available, the artifacts, fea- tures, and possible structures encountered in testing the Drivers site on the lower Tinker’s Creek floodplain, are also nearly identical to those from Hale Farm ( Lee 1 982). Beyond those relatively thoroughly explored sites within the region, similar ceramic and lithic artifacts have been encountered in test excavations. In the river valleys, the Kurtz and the Lee village sites, the Rhodes Farm site, and the Young sites are relatively large agricultural sites located on the second terraces of the Cuyahoga, Chagrin, and Grand rivers. Along the smaller tributaries of these rivers, well into the uplands, the Mohawk Park Shelter, the Pineway Trails site, the Gillie (Bernhardt 1973), the Krill (Prufer, personal communication 1983), the Stow and Little Mountain Shelters, the Cleveland Zoo, and the Doan Brook sites are all very small reoccupied fall and winter hunting camps. It is likely that the sporadic recov- ery of similar artifacts underlying some early nineteenth- century activities on the Cuyahoga River floodplain (cf. Jeff Richner, personal communication 1984; Lee et al. 1983) represent small, even more seasonally limited, sin- gle family activities. These sites were possibly occupied for collecting tubers and greens during the early spring, usually called “the starving moon” by Indians throughout this area. There are also a few contemporary sites of differing sizes such as the Avon Plant site (Brose and Morse 1976), the Cahoon Creek and the Greenhouse sites (Brose, White, and Ford 1983), a lower component of the Reeve site, dated at a.d. 1065 ± 100 (CRWU-13) (Brose 1985a), and the Ashtabula Gulf site reported by Kraus (1942). These sites all sit on sand ridges cut by streams as they enter Lake Erie. They were reoccupied as medium-sized agricultural villages and as seasonal fishing camps. By combining evidence from these sites it is possible to glimpse new economic and geographic patterns which were adopted by the Indians of this early Hale phase of northeast Ohio’s Whittlesey cultural tradition between a.d. 900 and a.d. 1200. This was a period of significant climatic change, beginning with winters which were both colder and drier than normal and with mild summers which were longer and wetter. By the end of this phase, winters had become rather mild and quite wet while, ex- cept for a narrow zone along Lake Erie from just west of Cleveland to Pennsylvania, summers were not only long and warmer, but also drier (cf. Barreis et al. 1976; Brose 1980; Wendland and Bryson 1974). Such summertime changes would have made the social commitment to floodplain corn agriculture easier, and the shorter milder winters with heavier snow cover would have made late winter hunting and early spring fishing and plant collect- ing increasingly reliable. 58 DAVID S. BROSE No. 41 Conclusion It must have been some late summer day nearly a thou- sand years ago that the Hale Farm site, surrounded by forests, was first cleared and occupied. The prehistoric group appears to have included a dozen or so American Indians. Probably there were two or three related fami- lies. Where these people spent that spring and early summer is not known. Possibly they had come from one of those archaeological sites known along the Cuyahoga River terraces in nearby portions of the valley. Those sites had been occupied by somewhat larger groups whose pot- tery and stone tools suggest close relationships with those found at the Hale Farm site. (Brose 1975; Brose et al. 1981) Those sites also yielded some cultivated maize and show evidence of animals taken in the late spring or early summer, including large numbers of fish and migratory waterfowl. One or two semicircular to circular prehistoric struc- tures were built at the Hale Farm site. Both were likely to have had a covering of bark or reed mats tied over a sim- ple ring of saplings placed into the ground and tied to- gether at the top. It is unknown whether any interior brac- ings were used, although some of the historic Indians of the Great Fakes who constructed similar houses did use them (cf. Brose 1970:37-38). While spending that summer at the Hale Farm site, these Indians gathered the ripening wild fruits and nuts and hunted a variety of animals as they waited to harvest their small crops of corn from the sandy floodplain ter- races to the east and north. A wide variety of domestic activities were performed, only some of which have left archaeologically recovered remains. Chipped stone tools were made from cobbles picked up in the nearby Cuya- hoga River and from stone obtained on expeditions to quarries and outcrops which were several day’s travel from the Hale Farm site. The stone tools were used for a range of cutting, scrapping, and piercing work. They were also broken and resharpened, and some were lost. Pottery was also made at the site, from the local clays mixed with fragments of crushed and burned rock. It was a simple shape, hand built in several different sizes to ac- cord with the group’s functional needs for both cooking and for storage. A good deal of it was fragile and several of the vessels were broken. With the coming of cooler fall weather after the har- vest, the natural resources which had sustained the group for several months were less easily obtained. The fires would have been put out and the remains of the last meals, along with the broken and discarded bowls and jars, were thrown into the pits and hearths. The bark or skin mat- tings would be taken from the sapling framework of the houses. Dried nuts and corn and the remaining smoked meat could be packed along with a few personal posses- sions, and each family would leave the valley for their winter hunting camps in the upland. They probably planned to meet again the following summer, but they never returned to the Hale Farm site itself. The sapling framework of the houses decayed in place. Through cen- turies falling leaves and the spring floods from Hale Run filled in and eventually covered completely the aban- doned fire hearths and half-emptied storage pits. The low rise at the forks of Hale Run was reclaimed by the silence and the forests, until these, in turn, were disturbed by Ab- raham Miller’s axe. Acknowledgments Beyond the efforts of the various field crew members, 1 wish to express thanks to the museum volunteers Bettyann Ball, Dan Best, Nathalie Boswell, Robert Burns, Jr., William Fienga, Judy McDonald, Jack Moviel, Susan Russell, Lars Sande, Ronald Skrbin, and Pat Stockwell. Special consideration is also due to Dean Zimmerman who curated and resurrected field notes, artifacts, and field invento- ries from a decade of neglect. Finally I thank Siegfield Buerling for never having given up all hope that I would eventually com- plete this report, among others still due. References Andreas, Barbara K. 1980. The flora of Portage, Stark, Summit and Wayne counties, Ohio. Ph.D. diss.. Department of An- thropology, Kent State University, Kent. Baerreis, David A., Reid Bryson, and Joan Kutzbach. 1976. Climate and culture in the western Great Lakes region. Mid- continental Journal of Archaeology 1(1 ):39— 58. Banks, P. O., and Rodney M . Feldmann, eds. 1 970. Guide to the geology of Northeastern Ohio. Northern Ohio Geological Society, Cleveland. Belovich, Stephanie, and David S. Brose. 1 982. Survey and eval- uative testing within the Cuyahoga Valley N.R.A.: Sites 33Su87 and 33Sul02. Archaeological Research Reports 40:1-86. Cleveland Museum of Natural History, Cleveland. Benson, Donna Lynne. 1978. A reconsideration of the Carolina pattern. Master’s thesis in Anthropology, Case Western Re- serve University, Cleveland. 1980. Taxonomy in historical archaeology. Papers of the Conference on Historical Archaeology, Vol. 1 1. In press. Bernhardt, Jack E. 1973. Gillie Rockshelter: A Late Woodland phase in Summit County, Ohio. Master’s thesis in Anthro- pology, Kent State University, Kent, Oh. Brashler, Janet G. 1981. Early Late Woodland boundaries and interaction: Indian ceramics of southern Lower Michigan. Anthropological Series 3(3): 1-359. Publications of the Mu- seum, Michigan State University, East Lansing. Braun, E. Lucy. 1950. Deciduous forests of eastern North Amer- ica. Philadelphia: Blakeston Co. 1955. The phytogeography of unglaciated eastern United States and its interpretation. Botanical Review 21:297-375. Brose, David S. 1970. Middle Woodland cultural ecology and social organization in the Upper Great Lakes. Case Western Reserve University Papers in Anthropology No. 1 , Cleveland. 1971. The Southeast ceremonial complex in Northeast- ern Ohio. The Ohio Archaeologist 2 1(4): 16- 19. 1985 PREHISTORIC HALE FARM OCCUPATION 59 1973a. Report on the analysis of early pioneer settle- ment in the Connecticut Western Reserve, 1796- 1837: A sta- tistical evaluation of the archaeological and historic data. Report on file with the National Endowment forthe Humani- ties. Washington, D.C. 1973b. A preliminary report on recent excavations at the South Park site, Cuyahoga County. Ohio. Pennsylvania Archaeologist 43( I ):25— 43. 1975. Preliminary results of the archaeological recon- naissance of the Lower Tinkers Creek Valley. The Cuyahoga Valley National Park Federation Board, pp. 1-58. Peninsula. 1976a. Locational analysis in the prehistory of North- eastern Ohio. In Cultural change and continuity: Essays in honor of James B. Griffin, edited by Charles Cleland, pp. 3-18. New York: Academic Press. 1976b. The Whittlesey occupations of Northeast Ohio: A second approximation. In The late prehistory of the Lake Erie drainage basin, edited by David S. Brose, pp. 22-51. Scientific Papers of the Cleveland Museum of Natural His- tory, Cleveland. 1978a. The late prehistory of the Upper Great Lakes. In Handbook of North American Indians. Vol. 15, Northeast, edited by Bruce Trigger, pp. 569-582. Washington, D.C.: Smithsonian Institution. 1978b. Some comments on “Temporal models in East- ern U.S. prehistory” by James B. Stoltman. Current Anthro- pology 19(4): 729-731. 1978c. A model of changing subsistence technology in the Late Woodland of Northeast Ohio. In Lithic Technology and Subsistence, edited by David Davis. Vanderbilt Univer- sity Publications in Anthropology 20:87-1 16. Nashville. 1980. The Late Prehistoric occupations of the South Central Lake Erie area. Paper presented at the 1 4th Annual Meeting of the Ontario Archaeological Society Symposium, London, Ontario. 1 984. A statistically dendritic analysis of Whittlesey ce- ramics with an illustrated type-variety classification. Manu- script on file at the C.M.N.H. Archaeological Archives. 1985a. Recent salvage excavations at two village sites near Cleveland, Ohio: Redating the occupations of the Tuttle Hill and Reeves sites. The Ohio Archaeologist. In press. 1985b. South Park: A stratified Whittlesey site in Cuya- hoga Co., Ohio. Scientific Papers of the Cleveland Museum of Natural History. In press. Brose, David S., S. Belovich, M. Brooslin, R. Burns, J. Hall, H. Haller, C. Pierce, and C. Ubbeohde. 1981. Prehistoric and historic archaeological investigations of the Cuyahoga Valley National Recreation Area, Ohio. Archaeological Research Reports of the Cleveland Museum of Natural History 30:1-524. Brose, David S., and Donna L. Benson. 1985. The Jonathan Hale Farm excavations and a socio-ecological analysis of pi- oneer settlement in the Western Reserve. Manuscript in pre- paration. Brose, David S., and David Morse. 1976. An archaeological survey of the Avon Lake district. Report to the Director of Utilities, Avon Lake, Ohio. Archaeological Research Reports of the Cleveland Museum of Natural History 16:1-23. Brose, David S., and G. Michael Pratt. 1976. Salvage excava- tions at the Walnut Tree site, 33Cu35: Archaeology of an im- pacted prehistoric campsite. Cleveland Regional Sewer Dis- trict Engineer Technical Reports. Brose, David S., and John F. Scarry. 1 976. Boston Ledges shel- ter: Comparative analysis of Early Late Woodland occupa- tions in Summit County, Ohio. Midcontinental Journal of Archaeology 1(2): 1 79-228. Brose, David S., Gregory Wentzel, Helga Bluestone, and Pat Essenpreis. 1976. Conneaut Fort: A prehistoric Whittlesey focus village in Ashtabula County, Ohio. Pennsylvania Ar- chaeologist 46(4):29-77. Brose, David S., Nancy M. White, and Sarah Ford. 1983. Sev- enteen prehistoric sites in Northern Ohio. Manuscript on file at the Cleveland Museum of Natural History, Department of Archaeology, Cleveland. Bush, David R. 1982. The earlier Late Woodland in the Chagrin River — The Kernisky site. Paper presented at the 45th An- nual Meeting of the Midwestern Archaeological Conference. Cleveland. Cleland, Charles E. 1966. Prehistoric animal ecology and eth- nozoology in the Upper Great Lakes. Anthropological Pa- pers No. 26. Museum of Anthropology, University of Michi- gan, Ann Arbor. 1976. A re-examination of the focal-diffuse adaptation model with a note of evolutionary implications. Midcontinen- tal Journal of Archaeology 1( 1 ):59— 76. Converse, Robert. 1964. Ohio stone types. The Ohio Archaeol- ogist 1 3(4):78— 1 20. Fienga, William, and Alfred Lee. 1982. The Bass Lake site, Geauga County, Ohio. Paper presented at the 45th Annual Meeting of the Midwest Archaeological Conference. Cleve- land. Fitting, James E. 1964. Ceramic relationships of four Late Woodland sites in Northern Ohio. Wisconsin Archaeologist 45 (4): 160- 175. 1965. Late Woodland cultures of Southeastern Michi- gan. Anthropological Papers No. 24. Museum of Anthropol- ogy, University of Michigan, Ann Arbor. Fitting, James E., Stephan R. Claggett, and Stephen R. Triech- ler. n.d. Prehistoric projectile points of Michigan; Great Lakes informant. Topics in History Ser. 2, No. 5, pp. 1-11. Michigan Department of State. Michigan History Division, Lansing. Ford, Richard I. 1977. Evolutionary ecology and the evolution of human ecosystems: A case study from the Midwestern U.S. A. In Explanation of prehistoric change, edited by James W. Hill, pp. 153-184. Albuquerque: Univ. of New Mexico Press. Gordon, Robert B. 1966. Vegetation of Ohio at the time of orig- inal survey. Map. Columbus, Ohio. Biological Survey. 1 969. The natural vegetation of Ohio in pioneer days. Bulletin of the Ohio Biological Survey, n.s., 3(2): 1-1 13. Co- lumbus. Greenman, Emerson. 1930. Field notes of survey and excava- tion in Northern Ohio 1929-1930. On file at the Ohio Histori- cal Society, Division of Archaeology. Columbus. 1935a. Excavation of the Reeve Village site. Lake Co., Ohio. The Ohio State Archaeological and Historical Quar- terly 44( 1 ):2— 64. 1935b. Seven prehistoric sites in Northern Ohio. Tlte Ohio State Archaeological and Historical Quarterly 44(2):220— 237. 1937. Two prehistoric villages near Cleveland, Ohio. The Ohio State Archaeological and Historical Quarterly 46(4) :305— 366. 60 DAVID S. BROSE No. 41 1939. The Wolf and Furton sites, Macomb Co., Michi- gan. Occasional Contributions of the Museum of Anthropol- ogy No. 8. University of Michigan, Ann Arbor. Griffin, James B. 1943. The Fort Ancient aspect. Ann Arbor: University of Michigan Press. 1 946. Cultural change and continuity in Eastern United States archaeology. In Man in Northeastern North America, edited by Frederick Johnson. Papers of the Robert S. Pea- body Foundation for Archaeology, 3:37-95. Andover. 1967. Eastern American archaeology: A summary. Science 156(3772): 175- 191. Washington, D.C. 1978. Late prehistory of the Ohio Valley. In Handbook of North American Indians, Vol. 15, Northeast, edited by Bruce Trigger, pp. 547-566. Washington, D.C.: Smithsonian Institution. Hall, John. 1980. Field notes on the Quaternary geology and modern geomorphology, soils, and water resources of North- east Ohio. On file. Department of Geology, CWRU. Halsey, John R. 1968. The Springwells Mound Group of Wayne County, Michigan. In Contributions to Michigan Archaeol- ogy, edited by J. Fitting, pp. 79- 1 59. Anthropological Papers No. 32. Museum of Anthropology, University of Michigan, Ann arbor. Horton, John J. 1961. The Jonathan Hale Farm: Achronicalof the Cuyahoga Valley. The Western Reserve Historical So- ciety, Publication No. 116. Cleveland. Keeley, Lawrence H. 1974. Technique and methodology in mi- crowear studies: A critical review. World Archaeology 5(3):323— 336. London. 1980. Experimental determination of stone tool uses: A microwear analysis. Prehistoric Archaeology and Ecology Series. The University of Chicago Press, Chicago. Kraus, Bertram S. 1942. Archaeological survey of Ashtabula County, Ohio, 1941. Mimeographed manuscript on file at CWRU. Lee, Alfred M. 1982. The Driver’s site, 33Su24; Evaluative test- ing and salvage excavations. Cleveland Museum of Natural History Archaeological Research Reports 41:1 -37. Lee, Alfred M.,and David S. Brose. 1980. Archaic archaeologi- cal complexes in Northeastern Ohio. Paper presented at the 14th Annual Meeting of the Ontario Archaeological Society Symposium. London, Ontario. Lee, Alfred M., David S. Brose, and Russell Weisman. 1983. Archaeological reconnaissance of the Akron Water Pollution Control Facility at Botzum, Northhampton Twp., Summit County, Ohio. Cleveland Museum of Natural History Ar- chaeological Research Reports 43:1-54. Luedtke, Barbara. 1978. Lithic material distributions and inter- action patterns during the Late Woodland period in Michi- gan. Ph.D. diss.. Department of Anthropology, University of Michigan, Ann Arbor. McKenzie, Douglas, Don Bier, Don Ianone, and Chris Pierce. 1974. The Riverview site: An Early Late Woodland village in Cuyahoga County, Ohio. Manuscript on file at Cleveland State University, Department of Anthropology, Cleveland. Miller, Barry B. 1970. The Quaternary period. In Guide to the geology of Northeastern Ohio, edited by Phillip Banks and Rodney Feldmann, pp. 149-164. Northern Ohio Geological Society, Cleveland. Miller, George. 1980. Classification and economic scaling of nineteenth century ceramics. Historical Archaeology 14:1-40. 1983. Ceramic supply in an economically isolated fron- tier community: Portage County of the Ohio [ric] Western Reserve 1800-1825. Historical Archaeology 17:19-32. Morgan, Richard G., and H. Holmes Ellis. 1943. The Fairport Harbor site. The Ohio State Archaeological and Historical Quarterly 52(3): 1-62. Murie, Olhaus J. 1951. The American elk. New York: Stackpole Press. Murphy, James L. 1971a. The Fairport Harbor site (33La5), Lake County, Ohio. Pennsylvania Archaeologist 4 1 (3) :26— 43. 1971b. Whittlesey ceramic types. The Ohio Archaeolo- gist 21(l):298-303. 1971c. The Lyman site (33La2), Lake County, Ohio. Pennsylvania Archaeologist 41(3): 12-25. 1972. A note on the South Park ceramics. The Ohio Archaeologist 22(2):3 1 —35. Ohio Department of Natural Resources, Division of Recreation Planning. 1975. Technical Report for the Cuyahoga River Valley Park Study, 4 vols. Columbus. Phillips, Philip. 1970. Archaeological survey in the Lower Yaz- oo Basin, Mississippi, 1949-1955. Peabody Museum of Ar- chaeology and Ethnology, Harvard University, Papers No. 60. Cambridge. Rau, John L. 1968. The evolution of the Cuyahoga River: Its geomorphology and environmental geology. In Cuyahoga River wastershed symposium, edited by D. Cooke, pp. 9-41. Kent State University, Kent. Read, Matthew C. 1 880. Explorations of a rocky shelter in Bos- ton, Summit County, Ohio. American Antiquarian 2:203-206. Boston. Ritchie, A., and J. R. Steiger. 1974. Soil survey of Summit County, Ohio. USDA Soil Conservation Service in coopera- tion with ODNR, Washington, D.C. Ritchie, William A. 1969. The archaeology of New York Slate. Revised edition. Garden City: Natural History Press. 197 1. A typology and nomenclature for New York pro- jectile points. New York State Museum and Science Service, Bulletin No. 384. Second edition, revised. Albany. Scully, Edward G. 1951. Some Central Mississippi Valley pro- jectile point types. Museum of Anthropology, University of Michigan Technical Report 1:1-15. Ann Arbor. Severinghaus, C. W. 1949. Tooth development and wear as criteria of age in whitetailed deer. Journal of Wildlife Man- agement 13(2): 195-2 16. Shelford, Victor E. 1963. Ecology of North America. Urbana: University of Illinois Press. Stout, Wilber, and R. A. Schoenlaub. 1945. The occurrence of flint in Ohio. Geological Survey of Ohio, Bulletin 46, 4th ser., Columbus. Whallon, Robert C. 1974. Spatial analysis of occupation floors II: Application of nearest-neighbor analysis. American An- tiquity 39:16-34. Wendland, Wayne M., and Reid A. Bryson. 1974. Dating cli- matic episodes of the Holocene. Quaternary Research 4(l):9-24. White, George W. 1953a. Pleistocene geology. (Map of Akron quadrangle) mapping by George White and copying by Jane Forsyth, 1956. Ohio Geological Survey, Columbus. 1953b. Geology and water-bearing characteristics of the unconsolidated deposits of Cuyahoga County. In The wa- ter resources of Cuyahoga County, Ohio, edited by J. D. 1985 PREHISTORIC HALE FARM OCCUPATION 61 Winslow, G. W. White, and E. E. Webber. United States Geo- logical survey. Water Resource Division, Bulletin No. 26. 1960. Classification of Wisconsin glacial deposits in Northeastern Ohio. United States Geological Survey, Bul- letin 1 1 2 1 A: A I —A 1 2. 1 979. Extent of till sheets and ice margins in Northeast- ern Ohio. Geological Note no. 6, State of Ohio Department of Natural Resources, Division of Geological Survey. Co- lumbus. White, George W., with Stanley M. Totten. 1982. Glacial geol- ogy of Northeastern Ohio: With a chapter on Pleistocene beaches and strand lines bordering Lake Erie. State of Ohio, Department of Natural Resources, Division of Geological Survey, Bulletin No. 68. Columbus. Williams, Arthur B. 1936. The composition and dynamics of a beech-maple climax community. Ecological Monographs 7(3). 1949. The native forests of Cuyahoga County, Ohio. Cleveland Museum of Natural History, Science Publication 9(1), Cleveland. Wittine, Arthur H. 1970. A study of Late Pleistocene lake in the Cuyahoga Valley, Summit County, Ohio. Master’s thesis in Geology, Kent State University, Kent. Yarnell, Richard A. 1964. Aboriginal relationship between cul- ture and plant life in the Upper Lakes region. Anthropologi- cal Papers No. 23. Museum of Anthropology, University of Michigan, Ann Arbor. Ziegler, Allan C. 1973. Inference from prehistoric faunal re- mains. Addison-Wesley McCaleb Module in Anthropology, Reading. ' . , . . STYLE GUIDE FOR AUTHORS: KIRTLANDIA Address manuscript correspondence to the editor. Dr. David S. Brose, Department of Archaeology, Cleveland Museum of Natural History, Cleveland, Ohio 44106. Re- prints of this Style Guide for Authors are available on request from the editor, or the publisher, Kent State Uni- versity Press, Kent, Ohio 44242. Manuscripts General Instructions: The manuscript should have a uniform style and be submitted exactly as it is to appear in print. It should consist of the following subdivisions, each prepared as a unit on separate sheets: Title page Abstract Notes Text Tables References Figure legends The manuscript, including references, should be typed double-spaced on bond or heavy-bodied paper 8 Vi" by 1 1 "(22 cm x 28 cm) with a 1 "(2.5 cm) margin on all sides. Number the manuscript pages consecutively beginning with the title page. Submit an original and three copies of text and illustrations. Do not divide words at the end of lines; if they are un- familiar to the printer, they may be incorrectly hyphen- ated. Corrections to the manuscript should be typed or printed legibly in ink. Do not begin sentences with abbreviations. The word Figure is not abbreviated in the text, except when appearing in parentheses: (Fig. 2); (Figs. 4-6). The spelling of nontechnical terms should be that rec- ommended in the current Webster’s Third New Interna- tional Dictionary, Unabridged. Always spell out numbers when they stand as the first word in a sentence; do not follow such numbers with ab- breviations. Numbers indicating time, weight, and mea- surements are to be in Arabic numerals when followed by abbreviations (e.g., 2 mm; 1 sec; 3 mil). Title Page : The title page must contain: Title Author’s name (or names) Insitution from which the paper emanated, with city, state, and zip code Number of text pages plus bibliography, figures, graphs, charts, and tables Abbreviated title (running headline) not to exceed 48 characters and spaces Name, address, and telephone number of the person to whom the proof is to be sent. Abstract: The abstract should consist of 250 words or less. When published, it will precede the introductory sec- tion of the text. The abstract should be written in com- plete sentences and should succinctly state the objectives, the experimental design of the paper, and the principal observations and conclusions; it should be intelligible without reference to the rest of the paper. 7exT:The general style guide for Kirtlandia is the 13th revised and expanded edition of The Chicago Manual of Style (1982). The author-date system of text references and a full (style B) reference list are required. References lists must be double-spaced. Titles of works should be capitalized according to the standard set forth in the Chi- cago Manual sec. 7. 123, and titles of journals, series, and papers or proceedings should be spelled out in full. In the text, references to the literature should be cited by au- thor’s surname followed by year of publication. . . . the study of Kirtland ( 1836) reveals . . . . . . the report of Squier and Davis (1848) describes . . . . . . studies by Darwin (1859) and Wallace (1868) sug- gest . . . . . . the earliest report (Whittlesey 1871) revealed . . . . . . previous discussions (Stout et al. 1943) indi- cate . . . When references are made to more than one paper by the same author or authors published in the same year they should be designated in their order of appearance in the text as, for example. Young (1947a), (Young 1947b), or (Young 1947a, 1947b) and in the reference list as follows: Young, C. C. 1947a. Mammal-like Reptiles from Lufeng, Yunnan, China. Proceedings of the Zoological Society of London 1 17:537-597. 1947b. On Lufengo saurus magnus Young and Additional Finds of Lufengo saurus huenei Young. Pale- ontologica Sinica, n.s., 12:1-53. The list of references must be arranged first alphabetically by author’s or authors’ surname(s) and chronologically for works by any single author or group of authors in the following style: Author’s surname (or names) and initials; year of publication; complete publication data, other than physical facts, as follows: American Ornithologists’ Union. 1957 Checklist of North American Birds. Baltimore, 5th ed. Aronson, J. L., T. J. Schmitt, R. C. Walter, M. Taieb, J. J. Tierelin, D. C. Johanson, C. W. Naeser, and A. E. M. Nairn. 1977. New Geochronologic and Paleomag- netic Data for the Hominid-Bearing Hadar Formation of Ethiopia. Nature 267:323-327. Beck, W. M. 1954. Studies in Stream Pollution Biology. Quarterly Journal of the Florida Academy of Sciences 17:211-227. Gartley, Richard, Jeff Carskadden, and Tim Gregg. 1973. The Philo Site, A Fort Ancient Component in the Cen- tral Muskingum Valley. Ohio Archaeologist 23(4): 15-19. 64 STYLE GUIDE FOR AUTHORS No. 41 Greber, N’omi. 1976. Within Ohio Hopewell. Ph.D. diss.. Department of Anthropology, Case Western Reserve University, Cleveland. Hopwood, A. T. 1926. Fossil mammalia. In The Geology and Paleontology of the Kaiso Bone Beds. Ed. E. J. Wayland. Occasional Papers of the Geological Survey of Uganda, 2: 13-36. Mayr, E. 1969. Principles of Systematic Zoology. McGraw-Hill, New York. Stout, W., and R. A. Schoenlaub. 1945. The Occurrence of Flint in Ohio. Ohio Geological Survey, 4th ser., Bul- letin 46. Uzzell, T. M. 1964. Relations of the Diploid and Triploid Species of the Ambystoma jeffersonianum Complex (Amphibia, Caudata). Copeia 1964:257-300. Notes: Footnotes are not accepted. Endnotes to the text should be as limited as possible. Endnotes must be numbered consecutively. The corresponding reference superscript numbers must be clearly indicated in the text. Additional references to the identical notes are to be numbered with the next following consecutive number. There were subjects for all six experiments.1 These were obtained through a variety of methods.2 Notes to a table should be typed directly beneath the table and numbered, 1, 2, 3, etc. They should not be numbered in sequence with the endnotes to the text. Tables: All tables must be cited in the text. Tabular material should be simple and uncomplicated, with as few vertical and horizontal rules as possible. Table titles should be complete but brief. Information other than that defining the data should be presented in the text. Figures: All figures must be cited in the text. Figure legends are to be numbered consecutively as follows: Fig. 1 . . . , Fig. 2 ... , and should follow the sequence of reference in the text. Whenever possible, figures should be integrated into the text. Reference to relevant text passages can often re- duce the length of legends and avoid redundancy. Metric System: The metric system should be used for all measurements, weights, etc. Temperatures should be expressed in degrees Celsius. Metric abbreviations should be expressed in lowercase without periods. Symbols: When preceded by a digit, the following sym- bols are to be used: % for percent; ° for degree; ' for min- ute; " for second. Abbreviations: Consult the 13th revised and expanded edition of The Chicago Manual of Style for permissible abbreviations. Illustrations: Photographic illustrations are costly. Au- thors should limit plates to those which are needed to adequately present the findings. All illustrations must be original drawings or high-quality photographic prints, submitted in complete and finished form with adequate labeling. Photographic Prints submitted for reproduction must be of adequate contrast and if multiple prints are included in a single figure, they should be of mylar uniform tone. Prints should be on white, nonmatte paper. Line Drawings should be drawn with black ink on medium-weight white paper, mylar or light-weight art- board. To reduce weight and postal charges, photo- graphic prints may be submitted in lieu of original draw- ings. The artwork should be sharp and black to achieve maximum contrast. Use stippling and hatching techniques to achieve tonal quality. Avoid the use of shading (pencil, wash, or air- brush) for tonal effect unless the drawing is to be repro- duced as a halftone. The reverse side of each illustration should indicate: Author’s name; figure number; top side of illustration; reduction requested; “Review copy” on those copies in- tended only for reviewers. Do not fasten illustrations with paper clips, staples, etc., since they will mar the surface of the illustrations. Illustrations should be shipped flat and protected by heavy cardboard. Illustrations cannot be reduced by less than 20% of their submitted size. Submitted line drawings cannot ex- ceed 1 1" x 14" (28 cm x 36 cm). Lettering and labels must be readable after reduction. When reduced, the minimum height of a capital letter should not be less than 2.5 mm for a photomicrograph and 1 mm for a graph or chart. Printed figures may not exceed 9" x 6" (22.9 cm x 15.3 cm). General Illustration Instructions: Original illustrations, and two sets of good-contrast photographic copies should be submitted with the manuscript for review purposes. Photocopies of illustrations made on an office duplicat- ing machine are not acceptable; copies of the manuscript only will be permitted. Figures should be trimmed straight on all sides and “squared.” Figures should be mounted on strong bristol board of about 1 5 points (0.4 mm) thickness with at least a 1" (2.5 cm) margin surrounding the figure or grouping of figures. Illustrations grouped to form a single figure should be of similar density and tone to prevent loss of detail. Labels should be large enough to allow for suitable re- duction and sturdy enough to withstand mailing and handling in the production process. Labeling should be done directly on the drawing or photographic print, never on an overlay. All labeling should be placed at least 1 / 8" (3 mm) in front of the edges of the illustration. To achieve adequate contrast between the label or letter and its back- ground, place white labels over dark backgrounds and black labels over light backgrounds. 1985 STYLE GUIDE FOR AUTHORS 65 Numbering: Figures, including charts and graphs, must be numbered consecutively. Miscellaneous: The editor and publisher reserve the right to return to the author for revision manuscripts and illustrations that are not in proper finished form. Upon acceptance of an article for publication, the au- thor will be asked to sign a transfer agreement, transfer- ring copyright to the publisher. Proofs: A single set of page and illustration proofs will be sent to the author. All corrections should be marked clearly, directly on page proofs. Authors are solely re- sponsible for proofreading their articles. Reprints: Reprints may be purchased at prices quoted on the reprint order form. Reprint orders should be returned with the proofs. It is important to order initially a sufficient quantity of reprints, since the price is substan- tially higher if they are ordered after the paper has been published. Publication in Kirtlandia requires consideration of edi- torial costs. These will be assessed at $32 per page for so- ciety members and $40 per page for nonmembers. Papers in excess of 32 printed pages may be charged full produc- tion costs. A page charge invoice will be sent with page proofs. Unsupported authors may request a waiver of page charges and will be asked to supply statements from their supervisor (department chairman, their advisor in the case of graduate students, and dean) that institutional and grant funds are not available. To help defray abnor- mal costs, authors will be requested to pay $75 per page for photographic illustration pages that exceed 25% of the total number of printed text pages. These payments will commonly be levied but are not a prerequisite to the pub- lication of any article. CONTENTS Benthic Colonization in Fresh Water: A Synthesis Michael J. S. Tevesz 3 An Inventory of the Vertebrates of the Grand River Terraces, Ashtabula County, Ohio Timothy O. Matson 15 An Echinocarid Crustacean, Echinocaris Auricula, from the Late Devonian of West Virginia Joseph T. Hannibal and Rodney M. Feldmann 22 Life Habits and Distribution of Riverine Lampsilis Radiata Luteola (Mollusca: Bivalvia) Michael J. S. Tevesz, David W. Cornelius, and J. Berton Fisher 21 The Prehistoric Occupation of the Hale Farm, Bath Township, Summit County, Ohio David S. Brose 35 Style Guide for Authors: Kirtlandia 63 1 CLEVELAND, OHIO NUMBER 42 ■ KIRTLANDIA •NATURAL HISTORY* PHYSICAL ANTHROPOLOGY Distal Radius and Proximal Femur Fracture Patterns in the Hamann-Todd Skeletal Collection Robert P. Mensforth, Susan A. Surovec, and James R. Cunkle 3 PALEONTOLOGY Echinocaris, A Mid-Paleozoic Crustacean: An Annotated Bibliography Joseph T. Hannibal and Rodney M. Feldmann 25 BOTANY A Bluegrass New To Ohio: Poa Saltuensis Fern. & Wieg. James K. Bissell 53 31 MAY 1987 KIRTLANDIA The Scientific Publication of the Cleveland Museum of Natural History David S. Brose, Editor Brief History and Purpose Kirtlandia is a publication of the Cleveland Museum of Natural History, an orga- nization incorporated in 1922 whose predecessor, the Ark , was founded in Cleveland by Jared Potter Kirtland in 1837. Published since 1926, Kirtlandia is a continuation of the earlier old series, Scientific Publications volumes 1 to 7 (1928-1937), and new series volumes 2 to 4 (1962- 1965), and the Survey of Ohio Fluted Points. Supported by the Kirtlandia Society of the Cleveland Museum of Natural History, Kirtlandia is devoted to the publication of scientific papers in the various fields of inquiry within the Museum’s sphere of interest: Cultural and Physical Anthropology; Archaeology; Botany; Geology; Paleobotany; Invertebrate and Vertebrate Paleontology; System- atic Ecology; and Invertebrate and Vertebrate Zoology. Issues will vary from single monographs to collections of short papers, review articles, and brief research notes. Kirtlandia is distributed by The Kent State University Press, Kent, Ohio 44242. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that the base fee of $00.50 per copy, plus .25 per page is paid directly to CCC, 27 Congress Street, Salem MA 09170. The fee code for users of the Transactional Re- porting Service is 0075-6245/87 $00.50+. 25. Kirtlandia is abstracted in Biological Abstracts and indexed in Bibliography and Index of Geology and Zoological Record. Kirtlandia No. 42 © 1987 by The Cleveland Museum of Natural History KIRTLANDIA THE CLEVELAND MUSEUM OF NATURAL HISTORY Cleveland, Ohio Spring 1987 Number 42 PHYSICAL ANTHROPOLOGY Distal Radius and Proximal Femur Fracture Patterns in the Hamann-Todd Skeletal Collection Robert P. Mensforth, Susan A. Surovec, and James R. Cunkle 3 PALEONTOLOGY Echinocaris, A Mid-Paleozoic Crustacean: An Annotated Bibliography Joseph T. Hannibal and Rodney M. Feldmann 25 BOTANY A Bluegrass New To Ohio: Poa Saltuensis Fern. & Wieg. James K. Bissell 53 ISSN: 0075-6245 KIRTLANDIA EDITOR DAVID S. BROSE Cleveland Museum of Natural History ASSOC I A TE EDITORS MARY BAUM Research Librarian Cleveland Museum of Natural History JAMES K. BISSELL Curator of Botany Cleveland Museum of Natural History BRUCE LATIMER Curator of Physical Anthropology Cleveland Museum of Natural History SONJA TERAGUCHI Manager of Collections Cleveland Museum of Natural History JOSEPH T. HANNIBAL Curator of Invertebrate Paleontology Cleveland Museum of Natural History SIBLEY HOOBLER Editorial Assistant Cleveland Museum of Natural History EDITORIAL A D VISOR Y BOA RD RODNEY FELDMANN Professor of Geology Kent State University MICHAEL C. HANSEN Geologist Ohio Geological Survey RICHARD MEINDL Associate Professor of Anthropology Kent State University G. MICHAEL PRATT Associate Professor of Anthropology Heidelberg University DAVID H. STANSBERY Director, Museum of Zoology Ohio State University FREDERICK H. UTECH Curator of Botany Carnegie Museum of Natural History ED VOSS Curator of the Herbarium University of Michigan ANDREW M. WHITE Professor of Biology John Carroll University DISTAL RADIUS AND PROXIMAL FEMUR FRACTURE PATTERNS IN THE HAMANN-TODD SKELETAL COLLECTION ROBERT P. MENSFORTH Department of Anthropology University of Oklahoma Norman, Oklahoma 73019 and SUSAN A. SUROVEC JAMES R. CUNKLE Department of Anthropology Cleveland State University Cleveland, Ohio 44115 Abstract The study presents a retrospective analysis of distal radius and proximal femur fractures that occurred in 938 Hamann-Todd Collec- tion skeletons. Individuals included in the investigation were retrieved from dissecting room cadavers in Cleveland, Ohio, between the years 1910 and 1938. Demographic analysis showed that mean age at death was 41.9 years for Blacks and 53.8 years for Whites examined in the study. Observations of remodeling status and side of involvement were recorded for all fractures identified. It was found that the age, sex, and race specific fracture patterns, which characterize the early twentieth- century Hamann-Todd sample, strongly correspond to those seen in modern American and European communities. The most striking difference between Hamann-Todd fracture patterns and those seen in modern groups concerns that much greater total frequency of traumatic injuries that occurred in the former group. Although hip fractures appear to be a primary result of age progressive skeletal fragility, it is suggested that the early onset, and high incidence, of distal radius fractures that occur in climacteric Caucasian women may be more directly due to accidents initiated by a greater frequency, intensity, and duration of vasomotor disturbances, which are known to accompany estrogen withdrawal in perimenopausal White females. Introduction Over the past several decades American and European populations have experienced a demographic shift whereby increasing numbers of people survive to old age (Gordon 1984; Johnell et al. 1984). Although continuing advances in medical technology and improvements in health care delivery systems are primarily responsible for such trends, we also have witnessed changes in the major causes of morbidity and mortality in these groups (Kilbourne and Smillie 1969). Thus, it is not surprisingto find that contemporary sociological, epidemiological, and clinical researchers have been devoting greater attention to the health care problems of the aged. One of the major problems that characterizes the geriatric segment of our society is a high incidence of traumatic injuries. For example, Gordon (1984) reports that 225,000 hip fractures occur annually in the United States. Over two-thirds of these occur in elderly women. The financial costs associated with hip fractures in the United States now approximates 3.8 billion dollars a year. However, old-age fractures are costly not only in terms of clinical management, but also in constituting a major risk of mortality. Compared to younger individu- als, it is well recognized that old people who have frac- tures are much more vulnerable to complications as a re- sult of surgery and/or immobilization (Robbins 1974). These include thromboembolism, pneumonia, and death. Thus, it is not remarkable that hip fractures are now the 12th leading cause of death among the elderly in the United States (Gordon 1984). These problems motivated Buhr and Cooke (1959) to conduct one of the first comprehensive surveys of fracture epidemiology in a modern urban industrial community. They examined fractures in individuals from England and Wales whose ages ranged from birth to 80+ years. It was found that traumatic injuries were charac- terized by marked age and sex specific patterns, which could be classified as one of four types (Buhr and Cooke 1959). These are L-Type, J-Type, A-Type, and Compos- ite fracture patterns. For example, an L-Type fracture pattern (e.g., supra- condylar fractures of the distal humerus; Fig. 1, upper left) typically shows a peak incidence at an early age, followed by a subsequent decrease in frequency to levels that are relatively insignificant. In contrast, the J-Type pattern identifies old-age fractures (Fig. 1, upper right). Here, the frequency of traumatic injury is low throughout childhood, and early and middle adulthood. Beyond 60 years of age J-Type fractures (e.g., hip fractures) show an Kirtlandia, No. 42 ® by The Cleveland Museum of Natural History 4 MENSFORTH, SUROVEC, AND CUNKLE No. 42 A-TYPE COMPOSITE (A,J) Fig. 1. Typical age and sex specific fracture patterns observed in modern human groups. Fracture frequencies are variably reported as number of cases per million, annual incidence per 10,000, incidence per 10,000 per age class, or percent of individuals affected per age class. The patterns shown here were adapted from Buhr and Cooke (1959). 1987 HAMANN-TODD SKELETAL COLLECTION 5 age progressive increase, and usually reach peak inci- dence in the oldest decade of life. The A-Type fracture pattern (e.g., phalangeal fractures; Lig. 1, lower left) describes those traumatic injuries that have a low initial frequency, then rise to reach peak incidence, and subse- quently decline to a low frequency once again. Linally, a Composite fracture pattern (e.g., distal radius fractures; Lig. 1, lower right) refers to those traumatic injuries that display two, or more, of the age related patterns described thus far. The different types of fracture pat- terns observed may also be characterized by sex differ- ences in age at onset and peak frequency of occurrence. The two J-Type traumatic injuries of old age that have been examined most thoroughly in contemporary epi- demiological studies are distal radius and proximal femur fractures. The former primarily are represented by fractures of the Colies, Smith, and Barton types (Bacorn and Kurtzke 1953; Older et al. 1965; Koefed 1983). Proximal femur fractures include compression fractures of the femoral head, subcapitular and basilar fractures of the femoral neck, and intertrochanteric fractures (Buhr and Cooke 1959; Bauer 1960; Alffram and Bauer 1962; Knowelden et al. 1964; Little 1973; Lreeman et al. 1974; Johnell et al. 1984; Bengner and Johnell 1985; Solgaard and Petersen 1985). The age and sex specific patterns that characterize distal radius and proximal femur fractures are illustrated in Fig. 2. The shown fracture patterns represent trau- matic injuries that occurred in White males and females from England and Wales (Buhr and Cooks 1959), and Age in Years Fig. 2. Incidence of distal forearm and proximal femur fractures in a contemporary population from England and Wales. Annual fracture rates are represented as number of cases per million. The left vertical margin shows the frequency of distal forearm fractures. The right vertical margin shows the frequency of proximal femur fractures. Adapted from Buhr and Cooke (1959). they are by no means unique. Although annual incidence per 10,000 may vary from one geographic region to another, identical patterns for these two types of trau- matic injury have been observed repeatedly (Bauer 1960; Johnell et al. 1984; Gordon 1984; Solgaard and Petersen 1985). The fundamental age and sex related differences, which discriminate between distal radius and proximal femur fracture patterns, can be described as follows. Dis- tal radius fractures in females show an early age at onset where a high incidence is established in the menopausal years. Thereafter, distal radius fractures in females attain what might be described as a plateau phase where a high frequency is maintained, but does not appear to increase. In contrast, males over 30 years of age have a low frequency of distal radius fractures throughout life and do not exhibit an old age J-Type pattern (Buhr and Cooks 1959; Bauer 1960). Proximal femur fracture rates illustrated in Pig. 2 show that these traumatic injuries are uncommon in males and females under 60 years of age. Thereafter, both sexes exhibit a typical J-Type age progressive increase in the frequency of hip fractures. However, females display fracture rates that are at least two times greater than those of males. Moreover, accelerated fracture rates in males begin to rise approximately ten years later than in females. With regard to etiology, a number of exogenous and endogenous factors are known to promote fracture risk in modern industrial communities (Buhr and Cooke 1959; Alffram and Bauer 1962). The former represent a group of nonbiological environmental circumstances, such as occupational hazards and automobile accidents. Geographic location and seasonality also have been implicated in differential fracture incidence in human groups (Gordon 1984; Solgaard and Petersen 1985; Beng- ner and Johnell 1985). However, among old people it is well recognized that endogenous factors play a dominant role in elevated risk, and increased severity, of traumatic injury (Bauer 1960). These include disease conditions that often result in pathological fractures (e.g., bone cysts, bone cancers, Paget’s disease, etc.) and degenerative consequences of biological aging (Buhr and Cooke 1959). The latter are far more important in fracture risk and include degenera- tive sequelae, such as failing eyesight, impaired motor coordination, muscular weakness and atrophy, and pro- nounced bone loss with age. Thus, as biological age progresses in men and women, the frequency of acciden- tal injuries continues to rise (Alffram and Bauer, 1962; Knowelden et al 1964). Indeed, Solgaard and Petersen (1985) report that the majority of distal radius fractures that occur in women (87%) and men (64 %) are the result of falls from level ground. Studies also have shown that hip fractures in individuals over 70 years of age are most often due to falls. 6 MENSFORTH, SUROVEC, AND CUNKLE No. 42 or movements we would ordinarily regard as trivial (e.g., arising from a seated position; Jaffe 1972; Johnell 1984). Thus, a greater frequency of relatively minor accidents in old people results in a greater incidence of moderate to severe traumatic injuries. A significant body of clinical and epidemiological research has established a strong positive association between age and sex specific fracture patterns and bone loss with age. The results of such studies are summarized elsewhere, and they will not be reviewed in detail here (see Alffram and Bauer 1962; Chalmers 1973; Goldsmith et al. 1973; Wasnich et al. 1985; and references therein). It will suffice to state that a slow normal rate of bone loss begins in males and females during the 4th decade of life (Goldsmith et al. 1973). Elowever, during the menopausal years women experience an accelerated rate of bone loss due to the cessation of ovarian function and estrogen withdrawal (Meema and Meema 1976; Horsman et al. 1977). Therefore, by age 60 adult females often show a significant reduction in skeletal mass, marked osteoporo- sis, and subsequently display an earlier onset age for J-Type fractures compared to males (Chalmers and Weaver 1966; Chalmers 1973). Racial differences in skeletal metabolism have also been demonstrated. Studies show that Black men and women attain higher peak bone mineral density values compared to Whites (Trotter et al. 1960; Goldsmith et al. 1973). Although Blacks lose bone at rates that are comparable to Whites, they enter the older decades of life with a relatively, and absolutely, greater bone mineral content (Goldsmith et al. 1973). Thus, it is not remark- able to find that the incidence of osteoporosis, and old-age fractures, is significantly lower in older Black men and women compared to Whites, and other racial groups as well (Gordon 1984). Given the epidemiological relationships described thus far, the purpose of the present investigation is to examine the frequency with which distal radius and proximal femur fractures occurred in a sample of skeletons that were assembled in Cleveland, Ohio, during the earlier part of the twentieth century. The study is therefore retrospective and replicative in nature. Here, it is specif- ically hypothesized that: (1) White females will display a greater frequency of distal radius fractures compared to all other sex/ race subgroups, (2) age at onset for increased incidence of distal radius fractures in White females will occur during the perimenopausal and menopausal years (i.e. , the period between 40 and 60 years), (3) White females will display a greater frequency of hip fractures compared to all other sex/ race subgroups, (4) age at onset for increased incidence of hip fractures in White females will be 60 years of age, (5) age at onset for increased incidence of hip fractures in White males will be 70 years of age, or approximately 10 years later than in females, (6) a low frequency of hip fractures should charac- terize all sex/ race subgroups under 60 years of age, (7) Black males and females will show a lower frequency of distal radius and proximal femur fractures at all ages compared to Whites, (8) Black males and females will show no significant increase in the frequency of distal radius and proximal femur fractures with age, and (9) Black males will show the lowest frequency of distal radius and proximal femur fractures com- pared to all other sex/ race subgroups. Although age at death, sex, and race are known for the individuals used in this study, a potential problem concerns the fact that adequate medical histories are unavailable for the skeletons examined. This means that the age at which fractures occurred is only known for those traumatic injuries appearing unremodeled at time of death. Therefore, the fracture frequency data to be dis- cussed are best regarded as cumulative in nature. None- theless, sex and race differences in age at onset for in- creased incidence of distal radius and proximal femur fractures should remain unaffected and easy to detect. Therefore, the J-Type fracture patterns that occurred in an early twentieth-century urban industrial American community were investigated in the following manner. Materials The human skeletal remains employed in this study are from the Hamann-Todd Collection. The bulk of this historic collection was assembled between the years 1910 and 1938, by Western Reserve University anatomists. Under the direction of T. Wingate Todd, the skeletons of over 3,400 dissecting room cadavers were retrieved for future scientific research. These skeletal materials are now permanently curated by the Cleveland Museum of Natural History. It is important to realize that many of the individuals included in the Hamann-Todd Collection were transients, indigents, and persons of low socioeco- nomic status. Thus, the sample used in this study is best regarded as a biased cross section of early twentieth- century urban industrial America. At present, the Hamann-Todd Collection contains records that document age and sex for the skeletal remains of 3,157 Black and White Americans. These specimens are partitioned by sex and race in Table 1. It can be seen that sex ratios are markedly skewed in favor of males. The male:female sex ratios for Blacks and 1987 HAMANN-TODD SKELETAL COLLECTION 7 TABLE 1 Black and White Adults in the Hamann-Todd Cadaver Records Race Male n\ Female m Sex Ratio Blacks 901 279 3.23 Whites 1725 252 6.85 Total3 2626 531 4.95b “Among the 3,157 Hamann-Todd Black and White adults (+18 years) listed in cadaver records, a total of 1, 180 (37.4%) Blacks and a total of 1,977 (62.6%) Whites are represented. bLor both races combined males outnumber females by a factor of five. Whites are 3.23 and 6.85, respectively. In addition, 88 adults and an unspecified number of subadults that are listed in cadaver records were returned to relatives for burial, or were cremated, following autopsy. Given these circumstances, the following sampling strategy was employed. At the time of observation a total of 262 Black and 207 White females were available for study. An equivalent number of Black and White males were then age-matched to their respective female samples. This was done to establish balanced adult age distributions and sex ratios. Thus, a total of 938 adult skeletons (Black n = 524; White n= 414) were examined for distal radius and proximal femur fractures. Methods All Hamann-Todd specimens included in the study were macroscopically examined for the presence/ absence of distal radius and proximal femur fractures. Distal radius fractures are defined as those fractures that occurred within 4 cm. of the distal articular surface of the radius. Proximal femur fractures were divided into two classes. The first is intracapsular fractures of the proxi- mal femur (PIC), and the second is extracapsular frac- tures of the proximal femur (PEC). The former are represented by compression fractures of the femoral head and subcapitular fractures of the femoral neck. The latter include basilar fractures of the femoral neck and intertro- chanteric fractures of the proximal femur. No fractures that were situated below the intertrochanteric line were included in the study. Four skeletal sites were examined for each specimen. These were the left and right proximal femora and distal radii, respectively. Adult males who were missing one or more of the sampling sites were replaced by complete age-matched specimens. Only three adult females were missing one observation per individual (i.e., .08% of all observations). For these females the missing skeletal site was assumed not to have been fractured. The identified fractures were then scored with re- spect to symmetry. Unilateral left, unilateral right, and bilateral involvements were recorded. Fractures were also qualitatively assessed as remodeled, or unremod- eled, at time of death. This was done in order to estimate the risk of mortality associated with each fracture type. All fractures that were identified in our initial survey were then re-examined on a second occasion. This proce- dure ensured that bony changes due to degenerative joint disease, and other pathological conditions, would not mistakenly be diagnosed as traumatic injuries (i.e., patho- logical fractures were not included in the analysis). Like- wise, all bones that showed evidence of post-mortem damage were carefully gleaned so that all fracture ‘mim- ics’ were excluded from consideration. Thus, the fracture data to be summarized in the study are reported with a high degree of confidence. The investigation results are presented in graphic and tabular form. Where appropriate, nonparametric proba- bility statistics were used to evaluate differences in fracture symmetry, remodeling status, and age, sex, and race differences in the fracture patterns observed. Results Demographic Composition of the Hamann- Todd Sam- ples The age distributions for Blacks and Whites examined in the study are listed by five-year age intervals in Table 2. For comparison, more recent aggregate census data for the United States in 1 960, is given in Table 3 (Thomlinson 1965). Survivorship curves for these data are illustrated in Fig. 3. It can be seen that the Hamann-Todd Black and White age distributions differ markedly from the modern age profile. The greater survivorship, which characterizes Fig. 3. Survivorship curves for the Hamann-Todd Black and White samples, compared to survivorship reported for more recent aggregate census data for the United States in 1960 (see Thomlinson 1965). MENSFORTH, SUROVEC, AND CUNKLE No. 42 TABLE 2 Age Distributions for the Black and White Male and Female Adult Samples Used in the Fracture Study Age n\ White Sample % h ni Black Sample % h 18-24 5 2.4 100.0 30 11.5 100.0 25-29 14 6.8 97.6 34 13.0 88.5 30-34 10 4.8 90.8 29 11.1 75.5 35-39 24 1 1.6 86.0 50 19.1 64.4 40-44 19 9.2 74.4 29 11.1 45.3 45-49 22 10.6 65.2 21 8.0 34.2 50-54 24 1 1.6 54.6 23 8.8 26.2 55-59 9 4.3 43.0 8 3.1 17.4 60-64 19 9.2 38.7 10 3.8 14.3 65-69 13 6.3 29.5 10 3.8 10.5 70-74 13 6.3 23.2 10 3.8 6.7 75-79 20 9.7 16.9 2 0.8 2.9 +80 15 7.2 7.2 6 2.3 2.1 Total2 207b 262° aThe total number of Hamann-Todd specimens included in the study equals 938. bThe 207 adult white females available for analysis were age-matched with 207 adult white males. The sample of adult whites examined thus equals 414 individuals. "The 262 adult black females available for analysis were age-matched with 262 adult black males. The sample of adult blacks examined thus equals 524 individuals. TABLE 3 Age Distribution for Modern Adults Age3 N % lx 20 558 0.58 100.0 25 611 0.64 99.4 30 732 0.76 98.8 35 1 124 1.17 98.0 40 1719 1.79 96.9 45 2595 2.70 95.1 50 3987 4.15 92.4 55 5616 5.84 88.2 60 7911 8.23 82.4 65 11132 11.58 74.1 70 13851 14.40 62.6 75 14173 14.74 48.2 80 14383 14.96 33.4 +85 17767 18.48 18.5 Total 96159 aThese are aggregate United States census data for 1960 as reported in Thomlinson 1965. the latter group, is undeniably a consequence of techno- logical advances in medicine and improvements in health care delivery systems that have occurred over the past 60 years. Marked differences also characterize the Hamann- Todd Black and White age distributions. Mean ages for the Black and White samples used in the study are 41.9 and 53.8 years, respectively. A Kolmogorov-Smirnov test indicates that the two age distributions are significantly different (x2 = 38.90; p<.001). The lower mean age at death among Blacks may reflect socioeconomic and racial factors that compromised access to health care during the late nineteenth and early twentieth centuries. Here, the circumstance that only a small proportion of the Blacks sampled are over 60 years of age is thus regarded as a potential source of error in our attempt to accurately es- timate the incidence of old age fractures in this group. General Observations Several of the distal radius and proximal femur fractures, which were identified in the Hamann-Todd series, are illustrated in Fig. 4. In general, degree of severity varied widely for the distal radius fractures that were observed. Traumatic involvement ranged from relatively benign hairline fractures of the subchondral bony joint surfaces, to those where severe trauma resulted in marked displacements and angular distortion of the distal radius. Proximal femur fractures were much more obvious. Intracapsular fractures of the proximal femur (PIC) typically displayed a marked shortening of functional femur neck length, combined with distortion of normal spherical contours of the femoral head. Extracapsular proximal femur fractures (PEC) were frequently situated at the base of the femoral neck, and they were often accompanied by the bony fragments of associated intertrochanteric fractures. Among the 938 skeletons that were examined, a total of 77 (8.2%) individuals displayed one or more distal radius fracture. Only 34 (3.6%) specimens exhibited proximal femur fractures. Thus, distal radius fractures were 2.3 times more frequent than hip fractures. A chi square comparison indicates that this difference is highly signifi- cant (x2 = 17.71; p<.001). For proximal femur fractures alone, a total of 22 (2.4%) individuals had extracapsular hip fractures, while only 12 (1.3%) specimens displayed intracapsular hip trauma. The difference in frequency for these two classes of hip fracture is not significant (x2 = 3.00; p>.05). All fractured individuals are partitioned by sex and race in Appendix A. Information summarized there identifies each specimen’s catalogue number, age at death, as well as the specific type, symmetry, and remodel- ing status for those fractures that were observed (Appen- dix A Table 21). Fig. 4. Examples of proximal femur and distal radius fractures which were observed in the Hamann-Todd Collection. A, left proximal femur which manifests a basilar neck fracture combined with an intertrochanteric fracture (CMNH-HT-3032, 89-year-old White female); B, Subcapital fracture of the right femoral neck which subsequently led to ischemic necrosis of the femoral head (CMNH-HT-1024, 84-year-old White female); C, anterior view of a Colie’s fracture affecting the right distal radius (CMNH-HT-14Q2, 81-year-old White female); D , posterior view of Colie’s fracture affecting the left radius; and E posterior view of Colie’s fracture affecting the right radius of the same individual (CMNH-HT-1581, 83-year-old White female). 10 MENSFORTH, SUROVEC, AND CUNKLE No. 42 Fracture Symmetry The frequency with which distal radius and proximal femur fractures were seen on the left, right, or both sides of affected individuals is summarized in Table 4. These data are partitioned by sex and race in Appendix B (Table 22). In general, 8-9% of all fractures were seen to TABLE 4 Distal Radius and Proximal Femur Fracture Symmetry Total No. of Individuals with Left Right Bilateral Fracture Type Fractures m % m_ % m % Distal Radius 77 36 46.8 32 41.6 7 9.1 Proximal Femur 34 16 47.1 15 44.1 3 8.8 occur bilaterally. Nonetheless, it was not possible to determine if bilateral involvements were the result of the same traumatic episode. For several individuals this did not seem to be probable. Patterns for unilateral involve- ment indicate no side preference for distal radius or proximal femur fractures. A chi square comparison confirms that no significant difference in unilateral expression characterizes the fractures which were identi- fied in the study (x2 = 0.02; p>.80). Fracture Remodeling Status The frequencies with which distal radius and proximal femur fractures appeared remodeled, or unremodeled, at time of death is given in Table 5. These data are parti- tioned by sex and race in Appendix C (Table 23). It was found that 88% of distal radius fractures, and 87% of in- tracapsular hip fractures, were healed at time of death. Though varying degrees of functional immobility un- doubtedly accompanied many of these injuries, it is con- cluded that distal radius and intracapsular proximal TABLE 5 Distal Radius and Proximal Femur Fracture Remodeling Status Total No. of Fractures Remodeled Unremodeled Fracture Type Observed m % rii % Distal Radius 84 74 88.1 10 1 1.9 Femur (PIC) 15 13 86.7 2 13.3 Femur (PEC) 22 5 22.7 17 77.3 femur fractures were well tolerated from a clinical perspective. In contrast, only 23% of extracapsular hip fractures were remodeled at time of death. Chi square values listed in Table 6 show that the marked remodeling differential between extracapsular hip fractures, and those described above, is statistically significant. It is concluded that indi- viduals who acquired extracapsular fractures of the prox- imal femur had a very high risk of mortality. That is, approximately eight out of every ten Hamann-Todd spec- imens with femur neck fractures probably died as a direct, or indirect, consequence of the traumatic injury. TABLE 6 Chi Square Values for Comparisons of Distal Radius and Proximal Femur Fracture Remodeling Status Comparison Chi Square Distal Radius vs. 39.24a Femur (PEC) Distal Radius vs. 0.26 Femur (PIC) Femur (PIC) vs 13.14“ Femur (PEC) “Significant at .001 level of probability. Old Age Fracture Patterns in the Hamann- Todd Collection The frequency with which distal radius and proximal femur fractures were observed for Hamann-Todd Black and White males and females is given by decade in Tables 7, 8, 9, and 10. These data are listed by five-year age inter- vals in Appendix D; Tables 24, 25, 26, and 27. The Black and White age and sex specific fracture patterns are illu- strated in Figure 5. The total frequency with which frac- tures occurred is partitioned by sex and race for each frac- ture type in Table 11. Data summarized there are expressed as the total number of individuals affected. Results show that White females experienced a greater frequency of distal radius and proximal femur fractures compared to all other sex/ race subgroups. In descending frequency, the rank order for distal radius fractures is White female (19.3%), White male (7.2%), Black female (6.5%), and Black male (1.9%). The rank order for proximal femur fractures is White female (9.7%), White male (3.4%), Black male ( 1.5%), and Black female ( 1 .2%). Chi square values (Table 1 2) indicate that both Black and White women had a significantly greater frequency of distal radius fractures compared to those of the proximal femur. Similar comparisons for Black and White men were insignificant. 1987 HAMANN-TODD SKELETAL COLLECTION TABLE 7 Age Specific Lracture Lrequencies for White Lemales Listed by Decade TABLE 10 Age Specific Lracture Lrequencies for Black Males Listed by Decade Total Distal Femur Femur Proximal Radius (PIC) (PEC) Femur Age N ri\ % fl2 % m % m % 18-29 19 0 0.0 0 0.0 0 0.0 0 0.0 30-39 34 1 2.9 0 0.0 0 0.0 0 0.0 40-49 41 6 14.6 0 0.0 1 2.4 1 2.4 50-59 33 10 30.0 1 3.0 1 3.0 2 6.1 60-69 32 6 18.8 0 0.0 5 15.6 5 15.6 70-79 33 9 27.3 0 0.0 4 12.1 4 12.1 +80 15 8 53.3 3 20.0 5 33.3 8 53.3 Total 207 40 19.3 4 1.9 16 7.7 20 9.7 TABLE 8 Age Specific Lracture Lrequencies for White Males Listed by Decade Age N Distal Radius rt\ % Femur (PIC) ni % Femur (PEC) W % Total Proximal Femur tlA % 18-29 19 0 0.0 0 0.0 0 0.0 0 0.0 30-39 34 3 0° oo 0 0.0 0 0.0 0 0.0 40-49 41 1 2.4 0 0.0 0 0.0 0 0.0 50-59 33 1 3.0 0 0.0 0 0.0 0 0.0 60-69 32 3 9.4 1 3.1 1 3.1 2 6.3 70-79 33 5 15.2 0 0.0 1 3.0 1 3.0 +80 15 2 13.3 1 6.7 3 20.0 4 26.7 Total 207 15 7.2 2 1.0 5 2.4 7 3.4 TABLE 9 Age Specific Lracture Lrequencies for Black Lemales Listed by Decade Age N Distal Radius rt\ % Femur (PIC) ni % Femur (PEC) m % Total Proximal Femur m % 18-29 64 0 0.0 0 0.0 0 0.0 0 0.0 30-39 79 4 5.1 0 0.0 0 0.0 0 0.0 40-49 50 4 8.0 0 0.0 0 0.0 0 0.0 50-59 31 2 6.5 0 0.0 0 0.0 0 0.0 60-69 20 6 30.0 1 5.0 1 5.0 2 10.0 70-79 12 0 8.3 1 8.3 0 0.0 1 8.3 +80 6 1 16.7 0 0.0 0 0.0 0 0.0 Total 262 17 6.5 2 0.8 1 0.4 3 1.1 Total Distal Femur Femur Proximal Radius (PIC) (PEC) Femur Age N n\ % rt2 % % ri\ % 18-29 64 \ 1.6 1 1.6 0 0.0 1 1.6 30-39 79 0 0.0 1 1.3 0 0.0 1 1.3 40-49 50 1 2.0 1 2.0 0 0.0 1 2.0 50-59 31 1 3.2 0 0.0 0 0.0 0 0.0 60-69 20 2 10.0 1 5.0 0 0.0 1 5.0 70-79 12 0 0.0 0 0.0 0 0.0 0 0.0 +80 6 0 0.0 0 0.0 0 0.0 0 0.0 Total 262 5 1.9 4 1.5 0 0.0 4 1.5 TABLE 1 I Summary of Number and Lrequency of Fractures Observed in the Hamann-Todd Sample White White Black Black Female Male Female Male (N-207) (N-207) (N-262) (N-262) Fracture Type n % n % n % n % Distal Radius 40 19.3 15 7.2 17 6.5 5 1.9 Femur (PIC) 4 1.9 2 1.0 2 0.8 4 1.5 Femur (PEC) 16 7.7 5 2.4 1 0.4 0 0.0 Total Proximal Femur3 20 9.7 7 3.4 3 1.2 4 1.5 Total 53h 25.6 22 10.6 20 7.6 9 3.4 ■“Figures listed for total proximal femur are simply the summed observa- tions for the catergories of femoral ( PIC) and femoral (PEC) fractures. hA total of 7 adult White female individuals displayed combined frac- tures of the distal radius and proximal femur at time of death. Thus a total of 53 White females (60-7) had one or more old-age fractures at time of death. No White males and no Blacks displayed combined distal radius and proximal femur fractures at time of death. TABLE 12 Chi Square Values Comparing the Difference in Total Frequency for Distal Radius versus Proximal Femur Fractures Distal Proximal Radius vs. Femur White Female 7.803 White Male 3.07 Black Female 9.253 Black Male 0.11 “Significant at the .01 level of probability. Proximal Femur Fractures Distal Radius Fractures H-T White H-T Black H-T White H-T Black Fig. 5. Age, sex, and race specific frequency of distal radius and proximal femur fractures that were observed in the Hamann-Todd sample. 1987 HAMANN-TODD SKELETAL COLLECTION 13 Chi square values given in Table 13 show that male/ female differences in the frequency of distal radius fractures are statistically significant for comparisons within each race. These tests also indicate that White males and females had a significantly greater frequency of distal radius fractures compared to Black males and females, respectively. Therefore marked sex and race differences characterize the total frequency of distal radius fractures that were encountered in the study. With TABLE 13 Chi Square Values Comparing Sex and Race Differences in the Total Frequency of Distal Radius Fractures White Female White Male Black Female Black Male White Female — 1 3. !0a 17.84a 40.43“ White Male — 0.01 8.07 Black Female — 6.83b Black Male — “Significant at the .01 level of probability. bSignificant at the .001 level of probability. regard to proximal femur fractures, the chi square values listed in Table 14 show only that White women had a significantly higher frequency of hip fractures compared to all other sex/race subgroups. Given the results pre- sented thus far, the age, sex, and race specific fracture patterns that were found can be described as follows. Distal Radius Fractures in Hamann- Todd Whites White males and females displayed dramatic differ- ences in distal radius fracture patterns (see Fig. 5; Tables 7 and 8). In general, fractures were uncommon in White women under 40 years of age. Thereafter, females showed a marked rise in the frequency of fractures. An early peak occurred in the sixth decade. Slightly lower frequencies were seen in the seventh and eighth decades. Nonetheless, TABLE 14 Chi Square Values Comparing Sex and Race Differences in the Total Frequency of Proximal Femur Fractures White Female White Male Black Female Black Mate White Female — 6.70“ 17.99b 1 5. 76b White Male — 2.77 1.74 Black Female — 0.14 Black Male — “Significant at the .01 level of probability. bSignificant at the .00! level of probability. the frequency of distal radius fractures remained high in females over 40 years of age, and reached peak incidence in the ninth decade. In contrast. White males had a low frequency of distal radius fractures compared to females overall. Men exhibited a minor A-Type fracture pattern during the 20 to 50 year period, and a diminutive J-Type pattern over 60 years of age. The age at onset for increased incidence of distal radius fractures in older men was during the seventh decade. The peak frequency for distal radius fractures in White males occurred in the eighth and ninth decades. Thus, White men and women showed dramatic differ- ences in age at onset, and frequency of occurrence, for distal radius fractures. These findings can be summarized as follows. Females displayed an age at onset for increased incidence of distal radius fractures that was 20 years earlier than age at onset for males. Also, women achieved a fracture incidence in the fifth decade that was not attained in males until the eight decade. Finally, females displayed a ninth decade distal radius fracture incidence that was four times greater than that seen in comparable age-matched males. Distal Radius Fractures in Hamann- Todd Blacks In contrast to Whites, the Blacks examined in the study showed lower frequencies of distal radius fractures, and did not display a typical J-Type pattern (see Fig. 5; Tables 9 and 10). The latter finding may be due to the fact that significantly fewer Blacks survived to the later decades of life. However, age and sex specific fracture patterns were seen among the Blacks who were examined. Although Black women under 60 years of age exhibited a slightly higher frequency of distal radius fractures compared to men, the overall incidence of fractures was low for both sexes during the early and middle years of adulthood. Age at onset for increased frequency of distal radius fractures, as well as age at peak incidence, occurred in the seventh decade for both sexes. The only noteworthy difference between males and females was confined to individuals over 60 years of age. At this time Black women displayed a frequency of distal radius fractures that was 3.5 times greater than that seen in males. Proximal Femur Fractures in Hamann- Todd Whites White males and females displayed a typical J-Type pattern for proximal femur fractures (see Fig. 5; Tables 7 and 8). In general, hip fractures were uncommon in men and women that were under 50 years of age. Age at onset for increased incidence of hip fractures was during the sixth decade in females, and the seventh decade in males. Thereafter, the frequency of hip fractures showed a general trend of age progressive increase in both sexes. White men and women in the ninth decade of life showed the highest frequency of hip fractures. Sex differences in 14 MENSFORTH, SUROVEC, AND CUNKLE No. 42 hip fracture incidence were also observed. For all decades over 50 years, White females exhibited frequencies of proximal femur fractures that were at least two times greater than those seen in males. Thus, hip fractures in the Hamann-Todd White sample showed a 10-year sex dif- ference in age at onset, and sex differences in overall fre- quency of occurrence. Proximal Femur Fractures in Hamann-Todd Blacks In contrast to Whites, the Blacks showed a low overall frequency of hip fractures, and did not display a charac- teristic J-Type pattern (see Fig. 5; Tables 9 and 10). The latter finding is presumed to be, in part, a consequence of poor sampling in the later decades of life, as was the case with distal radius fracture patterns. The age and sex specific patterns for hip fractures that were observed among Blacks are as follows. Hip fractures were uncom- mon in males and females under 60 years of age. Age at onset for increased incidence of proximal femur fractures was in the seventh decade for men and women. With regard to sex differences. Black females over 60 years of age displayed a frequency of hip fracture that was three times greater than that seen in males. Although age at onset appears to be similar, a sex difference in the frequency of hip fractures among older Blacks is thus suggested. Evaluation of Age Specific Fracture Patterns In order to assess the statistical significance of age related differences in the frequency of old-age fractures that were observed, the following procedure was used. Fracture frequency data were compressed into three age groups, which identify early adulthood (18-39 years), middle adulthood (40-59 years), and late adulthood (+60 years). These age intervals also correspond to the pre- menopausal, climacteric, and postmenopausal years in adult women, respectively. Fracture data that were summarized in this manner are listed in Tables 1 5 and 1 6, and illustrated in Fig. 6. Chi square values for sex specific comparisons of age related differences in the frequency of distal radius fractures are given in Table 17. Results show that (a) White women over 40 years of age had a significantly greater frequency of fractures compared to those women under 40 years of age, (b) White men over 60 years of age had a significantly greater frequency of fractures com- pared to men that were 40 to 59 years of age, and (c) both Black men and women over 60 years of age had a significantly greater frequency of fractures compared to those individuals that were 18 to 39 years of age. Thus, significant age-related differences in the frequency of distal radius fractures characterize all sex/race sub- groups that were examined in the study. Chi square values listed in Table 18 compare male/ TABLE 15 White Female and Male Fracture Data Compressed into Age Intervals Age N Distal Radius n\ % Femur (PIC) n2 % Femur (PEC) m % Total Proximal Femur m % White Female 18-39 53 1 1.9 0 0.0 0 0.0 0 0.0 40-59 74 16 21.6 1 1.4 2 2.7 3 4.1 +60 80 23 28.8 3 3.8 14 17.5 17 21.3 Total 207 40 19.3 4 1.9 16 7.7 20 9.7 White Male 18-39 53 3 5.7 0 0.0 0 0.0 0 0.0 40-59 74 2 2.7 0 0.0 0 0.0 0 0.0 +60 80 10 12.5 2 2.5 5 6.3 7 OO oo Total 207 15 7.2 2 1.0 5 2.4 7 3.4 TABLE 16 Black Female and Male Fracture Data Compressed into Age Intervals Age N Distal Radius m % Femur (PIC) m % Femur (PEC) m % Total Proximal Femur «4 % Black Female 18-39 143 4 2.8 0 0.0 0 0.0 0 0.0 40-59 81 6 7.4 0 0.0 0 0.0 0 0.0 +60 38 7 18.4 2 5.3 1 2.6 3 7.9 Total 262 17 6.5 2 0.8 1 0.4 3 1.1 Black Male 18-39 143 1 0.7 2 1.4 0 0.0 2 1.4 40-59 81 2 2.5 1 1.2 0 0.0 1 1.2 +60 38 2 5.3 1 2.6 0 0.0 1 2.6 Total 262 5 1.9 4 1.5 0 0.0 4 1.5 female age related differences in the frequency of distal radius fractures that were observed within each race. Results indicate that (a) White females over 40 years of age had a significantly greater frequency of fractures compared to White males, and (b) age related differences in the frequency of distal radius fractures among Black males and females were statistically insignificant. 1987 HAMANN-TODD SKELETAL COLLECTION 15 A 30- 25' ABC ABC ABC ABC White White Black Black Female Male Female Male B White White Black Black Female Male Female Male TABLE 17 Chi Square Values for Distal Radius Fracture Age Group Comparisons Partitioned by Sex and Race Age Group White White Black Black Comparison Female Male Female Male 18-39 vs. 40-59 9. 1 8a 0.71 2.58 1.22 18-39 vs. +60 14.25b 1.69 1 2.84h 4.25“ 40-59 vs. +60 1.03 5.14“ 3.22 0.78 '“Significant at the .01 level of probability. hSignificant at the .001 level of probability. “Significant at the .05 level of probability. TABLE 18 Chi Square Values for Distal Radius Fracture Age Group Comparisons between the Sexes and Partitioned by Race White Black Age Male vs. Female Mate vs. Female 18-39 0.84 1.53 40-59 1 l.26a 1.52 +60 6.45h 2.44 “Significant at the .001 level of probability. hSignificant at the .05 level of probability. TABLE 19 Chi Square Values for Distal Radius Fracture Age Group Comparisons between the Races and Partitioned by Sex Male Female Age Black vs. White Black vs. White 18-39 4.80a 0.01 40-59 0.01 6.42a +60 0.95 1.45 “Significant at the .05 level of probability. Fig. 6. Frequency of distal radius and proximal femur fractures partitioned by sex and race for the age groups A ( 1 8-39), B (40-59), and C (+60 years). Linally, chi square values given in Table 19 compare male/ male and female/female age related differences in the frequency of distal radius fractures between each race. Results show that (a) White females over 40 years of age had a significantly higher frequency of fractures compared to Black females, and (b) young adult White males had a significantly higher frequency of fractures compared to young adult Black males. No other race differences in the age specific frequency of distal radius fractures were statistically significant. Statistical analyses similar to those presented above were not conducted for proximal femur fractures. This is because hip fractures were infrequent overall, and they were particularly rare in those individuals that were under 60 years of age at time of death. Nonetheless, the age re- lated patterns in Lig. 5 clearly demonstrate that hip frac- tures were of no epidemiological significance in either Blacks or Whites prior to the onset of old age. Discussion With regard to the replicative goals of this study, a number of statements can be made concerning the age. 16 MENSFORTH, SUROVEC, AND CUNKLE No. 42 sex, and race specific fracture patterns that were pre- dicted earlier. Hypotheses 1 , 2, 3, 6, and 7 were confirmed as stated. Hypotheses 4, 5, and 9 were confirmed with minor qualifications. Only hypothesis number 8 was found to be incorrect. For each hypothesis given, the fracture patterns that were observed can be summarized as follows. (1) White females displayed a significantly greater frequency of distal radius fractures compared to all other sex/ race subgroups. (2) Age at onset for increased incidence of distal radius fractures in White women occurred in the 40 to 60 age group. (3) White females displayed a significantly greater frequency of hip fractures compared to all other sex/ race subgroups. (4) Age at onset for increased incidence of hip fractures in White women occurred during the sixth decade of life (i.e., 10 years earlier than predicted. (5) Age at onset for increased incidence of hip fractures in White men occurred during the seventh decade (i.e., 10 years earlier than pre- dicted. Age at onset in males did occur approxi- mately 10 years later than in females. (6) A low frequency of hip fractures characterized all sex/ race subgroups that were under 60 years of age at time of death. (7) Black men and women showed a lower fre- quency of distal radius and proximal femur fractures at all ages compared to White men and women, respectively. The only exception to this pattern was that Black females had a slightly higher incidence of distal radius fractures in the fourth and seventh decades compared to White females. However, these minor differences were insignificant, and not apparent in smoothed comparisons of age grouped data. (8) Contrary to the pattern predicted earlier. Black men and women that were over 60 years of age did display a significantly greater frequency of distal radius fractures compared to Black men and women in the 18 to 39 age group. (9) Black males displayed the lowest frequency of distal radius and proximal femur fractures com- pared to all other sex/ race subgroups. The only exception to this was that Black men showed a slightly greater, but insignificant, total fre- quency of hip fractures compared to Black women. However, data summarized earlier indi- cate that several hip fractures in Black males occurred in individuals that were under 60 years of age at time of death. As predicted, Black women over 60 years showed a higher age specific frequency of hip fractures compared to Black men. Therefore, it is quite clear that the age, sex, and race specific patterns that characterize distal radius and proximal femur fractures among Hamann-Todd Blacks and Whites strongly conform to fracture patterns that have been reported for more recent American and European urban industrial populations. However, one interesting observation concerns the finding that Hamann-Todd Black and White men both displayed an increased frequency of distal radius fractures in individu- als over 60 years of age. Fracture epidemiology studies, which were conducted during the mid 1950’s, report that older males showed no increased incidence of distal radius fractures with age (Buhr and Cooke 1959; Bauer 1960). Nonetheless, more recent surveys document a secular trend whereby men, and women, display marked increases in the incidence of both distal radius and hip fractures in individuals over 60 years of age (Nilsson and Obrant 1978; Zetterberg and Andersson 1982; Frandsen and Kruse 1983; Swanson and Murdoch 1983; Wallace 1983; Zain et al. 1984; Bengner and Johnell 1985; Solgaard and Petersen 1985). Thus, the increased fre- quency of distal radius fractures that was seen in older Hamann-Todd Collection Black and White men should not be regarded as an unexpected, or unusual, finding. The most striking difference found in this investigation concerns the remarkable frequency with which distal radius and proximal femur fractures occurred in Hamann-Todd individuals. In order to emphasize this point, the Hamann-Todd fracture data were converted to total incidence per 10,000. This information is summa- rized in Table 20. Also given is the total incidence per 10,000 for distal radius and hip fractures as computed for data reported for the population in Malmo, Sweden in 1955 (Bauer 1960). Fracture ratios for the distal radius indicate that TABLE 20 Total Incidence of Old Age Fractures Distal Radius Proximal Femur Fractures Fractures Subgroup H-T M-Sb Ratio H-T M-S Ratio White Female 1932 116 16.7 966 93 10.4 White Male 725 22 33.0 338 30 11.3 Black Female 649 116 5.6 115 93 1.2 Black Male 191 22 8.7 153 30 5.1 “Total incidence per 10,000 for the Hamann-Todd study sample of early twentieth century American Blacks and Whites. bTotal incidence per 10,000 for distal radius and proximal femur frac- tures reported for European White males and females in Malmo, Sweden in 1955 (Bauer 1960). 1987 HAMANN-TODD SKELETAL COLLECTION 17 Hamann-Todd total fracture incidence is (a) 17 times greater for White women, (b) 33 times greater for White men, (c) 6 times greater for Black women, and (d) 9 times greater for Black men compared to the modern group. Although less marked, fracture ratios for hip trauma show that Hamann-Todd total fracture incidence is (a) 10 times greater for White females, (b) 1 1 times greater for White males, (c) nearly identical for Black females and (d) 5 times greater for Black males. Given the known biased composition of the Hamann-Todd Collection, and lacking any other fracture data from the earlier part of the 20th century in urban America, it is suggested here that Hamann-Todd individuals represented a cross sec- tion of society at very high risk of traumatic injury. The final issue to be addressed here concerns the etiology of distal radius fractures in Caucasian women. All fracture epidemiology studies that have been con- ducted thusfar (i.e. , those cited herein) are in agreement with the following perspectives. First, the majority of distal radius fractures that occur in all age/sex/race groups are the simple consequence of accidental falls from level ground. Second, the age at onset for a marked increase in the incidence of distal radius fractures in White females is strongly associated with the climateric years of life. Third, studies have shown that bone loss with age is most pronounced in Caucasian women compared to all other age-matched sex and race groups. Fourth, skeletal fragility is regarded as the primary risk factor responsible for the early onset, and high incidence, of distal radius fractures in Caucasian females. Although senile and postmenopausal osteoporosis are clearly implicated in the pathogenesis of hip fractures in men and women that are over 60 years of age, we consider it improbable that bone loss alone could account for the early onset, high frequency, and marked differences that have been observed for distal radius fracture patterns in Caucasian women compared to men. Indeed, do peri- menopausal and menopausal females ‘fall’ more fre- quently than their premenopausal peers? If so, it would be of interest to know whether or not such falls were merely routine accidents, or whether such episodes were initiated by physiological disturbances common to this particular age and sex group. Thus, it is suggested here that physiological conse- quences of estrogen withdrawal other than bone loss may be important factors, which play a role in promoting accidents and elevating fracture risk during the climac- teric years of life in women. The latter is defined as the period encompassing the onset of menstrual irregularity, overt menopause, and several years following the com- plete cessation of menses. That is, the period from 45 to 60 years, with the median age of menopause occurring at approximately 50 years (Frommer 1964; Dennerstein and Burrows 1978). It is important for fracture epidemiologists to realize that over the last decade clinical researchers have devoted greater attention to the psychological, behavioral, bio- chemical, and physiological changes that occur in women undergoing the menopausal transition (McKinley and Jefferys 1974; Dennerstein and Burrows 1978; Casper et al. 1979; Tataryn et al. 1980). The problems encountered by these individuals include an increased incidence of psychosomatic illness, vaginal atrophy osteoporosis and vasomotor disturbances (Voda 1981). However, the latter are by far the most frequent symptoms (e.g., hot flushes, night sweats, etc), which generate discomfort and anxiety. Indeed, studies show that 75% of women undergoing natural menopause experience a variable frequency, intensity, and duration of vasomotor distur- bances during the climacteric period. (McKinley and Jefferys 1974; Voda 1981; Felman et al 1985). One of the most informative studies of the menopausal syndrome in Caucasian women was conducted by Bun- gay and colleagues ( 1 980). They found that (a) symptoms of vasomotor disturbance were strongly associated with the climacteric period, and (b) peak frequency of symp- toms closely corresponded to the median age of meno- pause. The vasomotor disturbances displaying this marked age and sex specific pattern were classified as Type 3a responses (see Bungay et al. 1980). The latter include hot flushes, night sweats, day sweats, dizzy spells, tiredness, palpitations, difficulty in concentration, diffi- culty in making decisions, loss of confidence, forgetful- ness, and feelings of unworthiness. Unfortunately, simi- lar information about the menopausal experience in Black women is unavailable. The age specific pattern for the frequency of vasomo- tor disturbances as seen in modern White climacteric women (Bungay et al. 1980) is illustrated in Fig. 7. Also shown is the age specific frequency with which distal radius fractures were observed in Hamann-Todd White females. The vasomotor disturbance data exhibits a typical A-Type pattern. It can be seen that the age related pattern for distal radius fractures in Hamann-Todd Caucasian women parallels the frequency of vasomotor disturbances for those individuals that ranged in age from 40 to 65 years. Although falls from level ground are recognized as the major circumstance in which distal radius fractures occur, it is perhaps worthy of comment that no studies have ever reported the results of inquires as to why patients fell in the first place. That is, we have no direct knowledge of the extent to which dizzy spells or other vasomotor disturbances may have preceeded, or ini- tiated, accidental falls. Thus, the data illustrated in Fig. 7 prompt us to suggest that vasomotor disturbances may play a significant role in initiating a greater frequency of accidents and, subsequently, an early onset and elevated incidence of distal radius fractures in climacteric White females. 18 MENSFORTH, SUROVEC, AND CUNKLE No. 42 White Female E o a H E 3 O a o c a o CO - Fig. 7. Frequency of distal radius fractures in Hamann-Todd White females, compared to the age specific frequency of Type 3a menopausal symptoms reported for modern White females (see Bungay et al., 1980). Type 3a responses are a class of symptoms due to vasomotor disturbances that accompany estrogen withdrawal in women during the climacteric years of life. Based on the age and sex specific physiological rela- tionships described thus far, it is further suggested that distal radius fractures in adult White women may be characterized by an overlapping A/ J-Composite fracture profile. The A and J-Type components that correspond to the hypothetical fracture pattern are illustrated in Fig. 8A. The additive distal radius fracture incidence which would result from such a pattern is shown in Fig. 8B. It is hypothesized that the A-Type distal radius fracture component should bracket the climacteric period. Dur- ing this time, vasomotor disturbances would be regarded as the dominant, but not sole, factor in fracture etiology. Therefore, the age at onset, peak incidence, and age at decline for distal radius fractures should parallel chang- ing frequencies of vasomotor disturbances. The second, J-Type distal radius fracture component would be regarded as a primary consequence of age progressive bone loss and other degenerative sequelae discussed earlier. Thus, as vasomotor disturbances de- crease in frequency at the end of the climacteric period, it would be expected that skeletal fragility will play a dominant role in the increased incidence of distal radius fractures, as is the circumstance with hip trauma among older men and women alike. The striking similarity between the hypothetical com- posite White female distal radius fracture pattern, and the pattern that was observed for Hamann-Todd White females, has not escaped our attention. However, several problems which limit our basis of inference deserve comment here. The hypothetical composite fracture pattern would receive support from our findings if the A White Female B White Female Fig. 8. Hypothetical age specific distal radius fracture pattern for Caucasian women. The proposed A-Type pattern during the climac- teric years, and the J-Type pattern for postmenopausal females are illustrated separately in Fig. 8A. The cumulative frequency of distal radius fractures which would be observed for the A/ J-Composite pattern is shown in Fig. 8B. majority of distal radius fractures that were seen in Hamann-Todd individuals had occurred with five years of age at death. This circumstance is considered to be unlikely. Most of these traumatic injuries were, to some 1987 HAMANN-TODD SKELETAL COLLECTION 19 degree, remodeled at time of death and true age at occurrence remains unknown. Therefore, the age specific frequencies that were reported for Hamann-Todd distal radius fractures are best regarded as cumulative data. Furthermore, the high 6th decade incidence of distal radius fractures that was seen in Hamann-Todd White females, relative to 7th and 8th decade values, can just as readily be interpreted as a consequence of sampling errors. More recent epidemiological studies do not permit evaluation of the hypothetical White female composite distal radius fracture pattern for other reasons. The most important among these is the effective clinical manage- ment of bone loss. Over the past two decades, prophylac- tic and therapeutic use of hormonal and nutritional supplements has become increasingly widespread as a means to slow bone loss in aging Caucasian women (Gordon 1961; Nicholas and Wilson 1963; Davis et al. 1966; Lafferty et al. 1969; Riggs etal. 1976; Gallagher and Riggs 1978). Numerous clinical studies document that accelerated rates of bone loss can be retarded if estrogen replacement therapy is begun within the first few years of natural, or surgical, menopause (Meema et al. 1975; Meema and Meema 1976; Aitken et al. 1976; Nordinet al. 1976; Horsman et al. 1977). With regard to fracture patterns, it is important to realize that widespread use of estrogen replacement therapy in climacteric women also has the effect of reducing the frequency, intensity, and duration of vaso- motor disturbances that would otherwise occur in these individuals (Bungay et al. 1980; Hammar et al. 1984). Thus, it is reasonable to expect that the hypothesized A-Type climacteric distal radius fracture component in modern White females will be reduced in magnitude and shifted upward in age. The A- and J-Type fracture patterns would then show significant overlap. The result- ant modern fracture profile might then give the appear- ance of an early onset J-Type pattern that attains a plateau phase, or simply an age progressive J-Type pattern. The most recent distal radius fracture patterns, which have been reported for Caucasian women, are in accord with the expectations posited above (see Solgaard and Petersen 1985; and Bengner and Johnell 1985). The retrospective analysis presented here, and more contemporary research described above, do not allow us to assess the validity of the hypothetical A/ J-Composite distal radius fracture pattern as posited for Caucasian women. Nonetheless, it is interesting that Hamann-Todd White females display an age specific distal radius frac- ture pattern similar to the one we would predict to occur in a population that did not experience the benefits of estrogen replacement therapy and recommended nutri- tional supplementation. Therefore, resolution of the vaso- motor disturbance hypothesis, as well as information about fracture patterns and other aspects of aging in modem Black men and women, deserve attention in fu- ture gerontological research. Summary and Conclusion The study presents a retrospective analysis of distal radius and proximal femur fractures that occurred in 938 Hamann-Todd Collection skeletons. The Black and White individuals included in the investigation were retrieved from dissecting room cadavers in Cleveland, Ohio, between the years 1910 and 1938. Thus, the sample represents a biased cross section of an early twentieth cen- tury American urban industrial society. Demographic analysis demonstrated that the mean age at death for Blacks, 41.9 years, was significantly lower than the mean age at death for Whites, 53.8 years, for those individuals examined. Therefore, the small propor- tion of Blacks over 60 years of age that were represented was considered a potential source of sampling error, with respect to our ability to accurately estimate the frequency of old age fractures in this group. Fracture symmetry data indicate that 8-9% of all distal radius and hip fractures occurred bilaterally. However, it was not possible to determine if bilateral involvements were the result of one or more traumatic injuries. It is probable that several of these occurred at different times during the life of affected individuals. Unilateral involve- ments showed no side preference for any of the fracture types that were examined. With regard to fracture repair status, it was found that 88% of distal radius fractures, and 87% of mtracapsular hip fractures, displayed moderate to marked bone re- modeling at time of death. It was concluded that these fractures were well tolerated from a clinical perspective. In contrast, only 23% of extracapsular hip fractures were remodeled at time of death. It was therefore concluded that Hamann-Todd individuals with these traumatic injuries were at a significantly high risk of mortality. The age, sex, and race specific fracture patterns, which characterize the Hamann-Todd sample, were found to correspond strongly to those seen in modern European and American communities. Therefore, the replicative goals of the study were confirmed with few exceptions. Results are summarized as follows. (1) White women displayed a significantly greater frequency of distal radius fractures compared to all other sex/race subgroups. (2) Age at onset for increased incidence of distal radius fractures in White females occurred dur- ing the 40 to 60 year period. (3) Caucasian women displayed a significantly greater frequency of hip fractures compared to all other sex/ race subgroups. (4) A low frequency of hip fractures characterized all sex/ race subgroups that were under 60 years of age at time of death. (5) Age at onset for increased incidence of hip fractures in White females occurred during the 20 MENSFORTH, SUROVEC, AND CUNKLE No. 42 6th decade, and peak frequency of hip trauma was seen in the ninth decade. (6) Age at onset for increased incidence of hip fractures in White males occurred during the seventh decade, 10 years later than females, and peak frequency of hip trauma was observed in the ninth decade. (7) Black men and women showed a low frequency of distal radius and proximal femur fractures at all ages compared to White men and women, respectively. (8) Black men and women that were over 60 years of age showed a significantly greater frequency of distal radius fractures compared to those Blacks in the 18 to 39 year age group. (9) Black males exhibited the lowest frequency of distal radius and proximal femur fractures com- pared to all other sex/ race subgroups. The only exception to this was that Black females had a slightly lower total number of hip fractures compared to males. The most dramatic difference between fracture pat- terns reported in modern groups and those which were observed for the Hamann-Todd sample concerns the much greater total frequency with which distal radius and proximal femur fractures occurred in the latter group. Distal radius fracture ratios indicate that the total frequency of traumatic injury for Hamann-Todd was (a) 17 times greater in White women, (b) 33 times greater in White men, (c) 6 times greater in Black women, and (d) 9 times greater in Black men, compared to a modern group. Similarly, hip fracture ratios show that the total frequency of these traumatic injuries for Hamann-Todd was (a) 10 times greater in White females, (b) 1 1 times greater in White males, (c) nearly identical in Black females, and (d) 5 times greater in Black males, compared to a modern group. These findings support the conclu- sion that individuals in the Hamann-Todd Collection represent a cross section of early twentieth century urban industrial society that was at very high risk of traumatic injury. Although hip fracture patterns appear to be a primary consequence of age progressive skeletal fragility, it is suggested that bone loss alone may not explain the early onset and high incidence of distal radius fractures that are known to characterize middle-aged Caucasian women. Alternatively, it is posited that vasomotor disturbances, which accompany estrogen withdrawal in climacteric women, may play a more important role in initiating a greater frequency of accidental falls. The latter would result in elevated distal radius fracture rates in climacteric White females, prior to significant reduction of skeletal mass and biomechanical strength. Finally, it is further suggested that adult Caucasian women in pre-estrogen replacement therapy societies may be characterized by a A/J-Composite distal radius fracture pattern. Here, vasomotor disturbances would dominate fracture risk in the climacteric years, and skeletal fragility would dominate fracture risk in individ- uals over 60 years of age. With respect to Black men and women, we regard the lack of comparative information about fracture patterns, aging effects in general, and the menopausal syndrome in Black women in particular, to constitute gerontological issues that deserve attention in future research. Acknowledgments The CSU-OAFS (Cleveland State University— Old Age Fracture Study) research group would like to express our gratitude to the faculty and staff of the Cleveland Museum of Natural History and Cleveland State University department of anthropology for their assistance in and support of this investigation. The authors would specifically like to thank Bruce Latimer, Dr. Byron Hoffman, Lyman Jellema, Dave Brose, Stephanie Belovich, Lori Linden, Jari Cardinal, Toni Hutton, Luba Gudz, Doris Webster, and the wonderful volunteer staff at the CMNH. We would also like to thank Leon Soule, John Lallo, Laura Martin, John Blank, and Paul Aspelin at CSU for their kindness and encouragement. Finally, we thank Jim Ohman, Ted Coombs, C. Owen Lovejoy, and Richard Meindl who tolerated endless bad puns about break- throughs in our research. The senior author would also like to thank the junior authors whose interest, motivation, and pro- ductivity were enormously appreciated. Also, I hope that the junior authors did not find this project to have been too much of a traumatic experience. This research was funded in full by Pass- the-Hat, Inc. References Aitken, J. M., D. M. Hart, R. Lindsay, and J. B. Anderson. 1976. Prevention of bone loss following oophorectomy in premonopausal women. Israel Journal of Medical Sciences 12:607-614. Alffram, P. A., and G. C. H. Bauer. 1962. Epidemiology of fractures of the forearm. A biomechanical investigation of bone strength. Journal of Bone and Joint Surgery 44A: 105-1 14. Bacorn, R. W., and J. F. Kurtzke. 1953. Colles’fracture: a study of two thousand cases from the New York State workmen’s compensation board. Journal of Bone and Joint Surgery 35A: 643-658. Bauer, G. C. H. 1960. Epidemiology of fractures in aged persons. Clinical Orthopedics 17:219-225. Bengner, U., and O. Johnell. 1985. Increasing incidence of forearm fractures. A comparison of epidemiologic patterns 25 years apart. Acta Orthopedica Scandinavica 56:158-160. Buhr, A. J., and A. M. Cooke. 1959. Fracture Patterns. Lancet (1):53 1-536. Bungay, G. T., M. P. Vessey, and C. K. McPherson. 1980. Study of symptoms in middle life with special reference to the menopause. British Medical Journal 28 1 : 1 8 1 - 1 83. 1987 HAMANN-TODD SKELETAL COLLECTION 21 Casper, R. F., S. S. C. Yen, and M. M. Wilkes. 1979. Menopausal flushes: a neuroendocrine link with pulsatile luteinizing hormone secretion. Science 205:823-825. Chalmers, J. 1973. Distribution of osteoporotic changes in the aging skeleton. Clinics in Endocrinology and Metabolism 2:203-219. Chalmers, J., and J. K. Weaver. 1966. Cancellous bone: its strength and changes with aging and an evaluation of some methods for measuring its mineral content. Journal of Bone and Joint Surgery 48A:299-308. Davis, M. E., N. M. Strandjord, and L. H. Lanzl. 1966. Estrogens and the aging process. The direction, prevention, and retardation of osteoporosis. Journal of the American Medical Association 196:219-224. Dennerstein, L., and G. D. Burrows. 1978. A review of studies of the psychological symptoms found at the menopause. Maturitas 1:55-64. Feldman, B. M., A. Voda, and E. Gromseth. 1985. The prevalence of hot flash and associated variables among perimenopausal women. Research in Nursing and Health 8:261-268. Frandsen, P. A.,andT. Kruse. 1983. Flip fractures in the county of Funen, Denmark. Acta Orthopedica Scandinavica 54:68 1 — 686. Freeman, M. A. R., R. C. Todd, and C. J. Pirie. 1974. The role of fatigue in the pathogenesis of senile femoral neck fractures. Journal of Bone and Joint Surgery 56B:698~702. Frommer, D. J. 1964. Changing age of the menopause. British Medical Journal 2:349-35 1 . Gallagher, J. C., and B. L. Riggs. 1978. Current concepts in nutrition. (Nutrition and bone disease). New England Jour- nal of Medicine 298: 193-195. Goldsmith, N. F., J. O. Picetti, and C. Garcia. 1973. Bone mineral in the radius and vertebral osteoporosis in an insured population. Journal of Bone and Joint Surgery 55A:1276- 1293. Gordon, G. S. 1961. Osteoporosis: diagnosis and treatment. Texas State Journal of Medicine 57:740-747. Gordon, G. S. 1984. Prevention of bone loss and fractures in women. Maturitas 6:225-242. Hammar, M., G. Berg, L. Fahraeus, and U. Larsson-Cohn. 1984. Climacteric symptoms in an unselected sample of Swedish women. Maturitas 6:345-350. Horsman, A., B. E. C. Nordin, J. C. Gallagher, P. A. Kirby, R. M. Milner, and M. Simpson. 1977. Observations on sequen- tial changes in bone mass in postmenopausal women: a controlled trial of estrogen and calcium therapy. Calcified Tissue Research 22:217-224. Jaffe, H. L. 1972. Metabolic, Degenerative, and Inflammatory Diseases of Bones and Joints. Philadelphia: Lea & Febiger. Johnell, O., B. Nilsson, K. Obrant, and I. Sernbo. 1984. Age and sex patterns of hip fracture — changes in 30 years. Acta Orthopedica Scandinavica 55:290-292. Kilbourne, E. D.,and W. G. Smillie. 1969. Human Ecology and Public Health. Fourth edition. London: MacMillan Com- pany. Knowelden, J., A. J. Buhr, and O. Dunbar. 1964. Incidence of fractures in persons over 35 years of age. British Journal of Preventive Social Medicine 1 8: 1 30- 141 . Koefed, H. 1983. Comminuted displaced Colles’fractures. Acta Orthopedica Scandinavica 54:307-31 1. Lafferty, F. W., G. E. Spencer, and O. H. Pearson. 1969. Effects of androgens, estrogens, and high calcium intake on bone formation and resorption in osteoporosis. American Journal of Medicine 36:514-528. Little, K. 1973. Bone Behavior. New York and London: Academic Press. McKinlay, S. M., and M. Jefferys. 1974. The menopausal syndrome. British Journal of Preventive Social Medicine 28:108-1 15. Meema, S., M. L. Bunker, and H. E. Meema. 1975. Preventa- tive effect of estrogen on postmenopausal bone loss. Archives of Internal Medicine 135:1436-1440. Meema, S., and H. E. Meema. 1976. Menopausal bone loss and estrogen replacement. Israel Journal of Medical Sciences 12:601-606. Nicholas, .1. A., and P. D. Wilson. 1963. Osteoporosis of the aged spine. Clinical Orthopedics 26:19-33. Nilsson, B. E., and K. J. Obrant. 1 978. Secular tendencies of the incidence of fracture of the upper end of the femur. Acta Orthopedica Scandinavica 49:389-391. Nordin, B. E. C., A. Horsman, R. Brook, and D. A. Williams. 1976. The relationship between estrogen status and bone loss in postmenopausal women. Clinical Endocrinology 5, sup- plement, 353-361. Older, T. M., E. V. Stabler, and W. H. Cassebaum. 1965. Colles fracture: evaluation and selection of therapy. Journal of Trauma 5:469-476. Riggs, B. L., J. Jowsey, P. J. Kelley, D. L. Hoffman, and C. D. Arnaud. 1976. Effects of oral therapy with calcium and vitamin D in primary oseoporosis. Journal of Clinical Endocrinology & Metabolism 42: 1 139-1 144. Robbins, S. L. 1974. Pathologic Basis of Disease. Philadelphia: W. B. Saunders Company. Solgaard, S., and V. S. Petersen. 1985. Epidemiology of distal radius fractures. Acta Orthopedica Scandinavica 56:391- 393. Swanson, A. J. G., and G. Murdoch. 1983. Fractured neck of the femur. Pattern of incidence and implications. Acta Orthopedica Scandinavica 54:348-355. Tataryn, 1. V., P. Lomax, .1. G. Bajorek, W. Chesarek, D. R. Meldrum, and H. L. Judd. 1980. Postmenopausal hot flushes: a disorder of thermoregulation. Maturitas 2:101- 107. Thomlinson, R. 1965. Population Dynamics. Causes and Consequences of World Demographic Change. New York: Random House. Trotter, M., G. E. Broman, and R. R. Peterson. 1960. Densities of bones of White and Negro skeletons. Journal of Bone and Joint Surgery 42A:50~58. Voda, A. M. 198! Climacteric hot flash. Maturitas 3:73-90. Wallace, W. A. 1983. The increasing incidence of fractures of the proximal femur: an orthopdaedic epidemic. Lancet (i): 1413-1414. Wasnich, R. D., P. D. Ross, L. K. Heilbrun, and J. M. Vogel. 1985. Prediction of postmenopausal fracture risk with use of bone mineral measurements. American Journal of Obstetrics and Gynecology 153:745-751. Zain Elabdien, B. S., S. Olerud, G. Karlstrom, and B. Smedby. 1984. Incidence of hip fracture in Uppsala County 1965-1980. Acta Orthopedica Scandinavica 55:284-289. Zetterberg, C., and G. B. J. Andersson. 1982. Fractures of the 22 MENSFORTH, SUROVEC, AND CUNKLE No. 42 proximal end of the femur in Goteborg, Sweden, 1940-1979. Acta Orthopedica Scandinavica 53:419-426. Appendix A CSU-OAFS (Cleveland State University — Old Age Fracture Study) List of All Fractured Individuals Code: a) H-T No., Hamann-Todd Catalogue Number. b) Age, Cadaver Record Stated Age at Death. c) Distal Radius, all fractures that were located within 4 cm. of the distal articular surface of the radius. d) Femur (PIC), Intracapsular Proximal Femur Frac- tures. e) Femur (PEC), Extracapsular Proximal Femur Frac- tures. f) R, Right Side Affected. g) L, Left Side Affected. h) r, all fractures which were remodeled, or evinced only minimal and initial remodeling, at time of death. CSU-OAF STUDY Adult White Female Fracture Specimens H-T No. Age Distal Radius Femur Head Femur Neck 2478 35 R-r/L-r 0228 40 R-r 0415 40 R-r 0783 42 R-u 1119 45 R-r 2437 45 R-r 2414 47 R-r/L-r 0355 49 L-u 0229 50 L-r 0411 50 L-u 0742 50 R-r 1149 50 R-r 3123 51 L-r 1846 53 R-r 3140 53 L-r 0785 54 L-r 3118 54 L-r 3352 54 L-u 2949 58 L-u/ R-u 3327 58 R-r 0022 60 R-r/L-r L-u 0541 60 R-u 1426 60 L-r 3132 61 R-r R-r 3164 62 L-r 0234 65 R-u 1200 68 L-U 2307 68 R-r R-r 1413 72 R-r H-T No. Age Distal Femur Radius Head Femur Neck 1811 73 L-r 3151 74 R-u 1451 75 L-r 1750 75 R-r 2188 75 R-r 2021 76 R-u 3183 77 L-r 3278 77 L-u 1191 78 L-r 1779 78 L-r 3359 78 L-r L-u 0927 80 L-r 1680 80 R-r 1753 80 R-r 1754 80 L-u 1402 81 L-r/ R-r R-r 1505 81 L-r L-r 2007 81 R-r 1581 83 L-r/ R-r 1024 84 R-r/L-r 1433 84 R-r 2121 85 L-r 3032 89 R-r L-u 1708 93 R-u Adult White Male Fracture Specimens 2689 38 L-r 2580 40 R-r 1765 40 L-r 2618 48 L-u 2198 58 R-r 1064 60 L-r 1464 60 L-u 0963 65 R-u 2626 65 R-r 3147 69 L-r 0759 73 L-r 3341 75 L-u 1989 75 L-u 1 166 76 L-r 2801 78 R-r 3081 78 R-r 1021 81 L-r 1307 83 R-u 2635 84 R-r 1663 85 R-r 3214 89 L-u 3025 93 R-u Adult Black Female Fracture Specimens 0918 35 R-u 2311 35 R-r 2612 39 R-r 2942 39 L-r 1702 40 R-r 3131 43 L-r 1987 HAMANN-TODD SKELETAL COLLECTION 23 H-T Distal Femur Femur Appendix B No. Age Radius Head Neck TABLE 22 1022 46 R-r Distal Radius and Proximal Femur Fracture 3182 48 L-r Symmetry Data 2660 5 1 R-r Individuals 2269 58 R-u/ L-u with 0839 60 R-r Fractures Left Right Bilateral 0773 60 R-u N n\ % n2 % W % 1551 60 L-r 1912 60 L-r DISTAL RADIUS 3174 60 L-r White Female 40 16 40.0 19 47.5 5 12.5 2773 64 L-r White Male 15 10 66.7 5 33.3 0 0.0 0751 65 R-r Black Female 17 9 52.9 7 41.2 1 5.9 2039 65 L-r Black Male 5 1 20.0 3 60.0 1 20.0 1367 72 L-r Total 77 36 46.8 32 41.6 7 9.1 0967 87 R-r FEMUR (PIC) Adult Black Male Lracture Specimens White Female 4 0 0.0 2 50.0 2 50.0 White Male 2 0 0.0 2 100.0 0 0.0 0598 25 R-r Black Female 2 2 100.0 0 0.0 0 0.0 2368 26 L-r/R- r Black Male 4 2 50.0 1 25.0 1 25.0 1245 31 L-r 129! 42 L-r Total 12 4 33.3 5 41.7 3 25.0 1338 42 R-r FEMUR (PEC) 2339 52 L-r White Female 16 10 62.5 6 37.5 0 0.0 1452 60 R-r White Male 5 2 40.0 3 60.0 0 0.0 1 /33 60 R-r Black Female 1 0 0.0 I 100.0 0 0.0 1528 67 R-r/L-r Black Male 0 0 0.0 0 0.0 0 0.0 Total 22 12 54.5 10 45.5 0 0.0 TABLE 21 Summary of Lracture Observations Appendix C White White Black Black TABLE 23 Female Male Female Male Total Distal Radius and Proximal Femur Fracture Remodeling Status Distal Radius Lx 45 15 18 6 84 Total Remodeled 41 12 15 6 74 Fractures Remodeled Unremodeled Unremodeled 4 3 3 0 10 Observed n\ % n2 % Unilateral Left 16 10 9 1 36 Unilateral Right 19 5 7 3 34 DISTAL RADIUS Bilateral 5 0 1 1 7 White Female 45 41 91.1 4 8.9 White Male 15 12 80.0 3 20.0 Lemur (PIC) Lx 6 2 2 5 15 Black Female 18 15 83.3 3 16.7 Remodeled 4 2 2 5 13 Black Male 6 6 100.0 0 0.0 Unremodeled 2 0 0 0 2 Total 84 74 88.1 10 1 1.9 Unilateral Left 0 0 2 2 4 Unilateral Right 2 2 0 1 5 FEMUR (PIC) Bilateral 2 0 0 1 3 White Female 6 4 66.7 2 33.3 Lemur (PEC) Lx 16 5 1 0 22 White Male 2 2 100.0 0 0.0 Remodeled 5 0 0 0 5 Black Female 2 2 100.0 0 0.0 Unremodeled 11 5 1 0 17 Black Male 5 5 100.0 2 13.3 Unilateral Left 10 2 0 0 12 Total 15 13 86.7 2 13.3 Unilateral Right 6 3 1 0 10 Bilateral 0 0 0 0 0 FEMUR (PEC) White Female 16 5 313 11 68.8 TNILxa 53 22 20 9 104 White Male 5 0 0.0 5 100.0 TNLxb 67 22 21 1 1 121 Black Female 1 0 0.0 1 100.0 Black Male 0 0 22.7 17 77.3 “Total no. of individuals that displayed one or more OAF fractures. Total no. of OAF fractures observed in each subsample. Total 22 5 22.7 17 77.3 24 MENSFORTH, SUROVEC, AND CUNKLE No. 42 Appendix D TABLE 24 TABLE 26 White Female Age Specific Fracture Frequencies Black Female Age Specific Fracture Frequencies Age Group N Distal Radius rt\ % Femur (PIC) n2 % Femur (PEC) m % Total Proximal Femur ru % Age Group N Distal Radius n\ % Femur (PIC) 12 % Femur (PEC) rt 3 % Total Proximal Femur ru % 18-24 5 0 0.0 0 0.0 0 0.0 0 0.0 18-24 30 0 0.0 0 0.0 0 0.0 0 0.0 25-29 14 0 0.0 0 0.0 0 0.0 0 0.0 25-29 34 0 0.0 0 0.0 0 0.0 0 0.0 30-34 10 0 0.0 0 0.0 0 0.0 0 0.0 30-34 29 0 0.0 0 0.0 0 0.0 0 0.0 35-39 24 1 4.2 0 0.0 0 0.0 0 0.0 35-39 50 4 8.0 0 0.0 0 0.0 0 0.0 40-44 19 3 15.8 0 0.0 0 0.0 0 0.0 40-44 29 2 6.9 0 0.0 0 0.0 0 0.0 45-49 22 3 13.6 0 0.0 1 4.5 1 4.5 45-49 21 2 9.5 0 0.0 0 0.0 0 0.0 50-54 24 9 37.5 0 0.0 1 4.2 1 4.2 50-54 23 1 4.4 0 0.0 0 0.0 0 0.0 55-59 9 1 11.1 1 11.1 0 0.0 1 11.1 55-59 8 1 12.5 0 0.0 0 0.0 0 0.0 60—64 19 5 26.3 0 0.0 2 10.5 2 10.5 60-64 10 4 40.0 1 10.0 1 10.0 2 20.0 65-69 13 1 7.7 0 0.0 3 23.1 3 23.1 65-69 10 2 20.0 0 0.0 0 0.0 0 0.0 70-74 13 3 23.1 0 0.0 0 0.0 0 0.0 70-74 10 0 0.0 1 10.0 0 0.0 1 10.0 75-79 20 6 30.0 0 0.0 4 20.0 4 20.0 75-79 2 0 0.0 0 0.0 0 0.0 0 0.0 +80 15 8 53.3 3 20.0 5 33.3 8 53.3 +80 6 1 16.7 0 0.0 0 0.0 0 0.0 Total 207 40 19.3 4 1.9 16 7.7 20 9.7 Total 262 17 6.5 2 0.8 1 0.4 3 LI TABLE 25 White Male Age Specific Fracture Frequencies TABLE 27 Black Male Age Specific Fracture Frequencies Age Group N Distal Radius ru % Femur (PIC) ni % Femur (PEC) rt) % Total Proximal Femur ru % Age Group N Distal Radius rt\ % Femur (PIC) rt2 % Femur (PEC) rt 3 % Total Proximal Femur ru % 18-24 5 0 0.0 0 0.0 0 0.0 0 0.0 18-24 30 0 0.0 0 0.0 0 0.0 0 0.0 25-29 14 0 0.0 0 0.0 0 0.0 0 0.0 25-29 34 1 2.9 1 2.9 0 0.0 1 2.9 30-34 10 0 0.0 0 0.0 0 0.0 0 0.0 30-34 29 0 0.0 1 3.4 0 0.0 1 3.4 35-39 24 1 4.2 0 0.0 0 0.0 0 0.0 35-39 50 0 0.0 0 0.0 0 0.0 0 0.0 40-44 19 2 10.5 0 0.0 0 0.0 0 0.0 40-44 29 1 3.4 1 3.4 0 0.0 1 3.4 45-49 22 1 4.5 0 0.0 0 0.0 0 0.0 45-49 21 0 0.0 0 0.0 0 0.0 0 0.0 50-54 24 0 0.0 0 0.0 0 0.0 0 0.0 50-54 23 1 4.3 0 0.0 0 0.0 0 0.0 55-59 9 1 11.1 0 0.0 0 0.0 0 0.0 55-59 8 0 0.0 0 0.0 0 0.0 0 0.0 60-64 19 2 10.5 0 0.0 0 0.0 0 0.0 60-64 10 1 10.0 1 10.0 0 0.0 1 10.0 65-69 13 1 7.7 1 7.7 1 7.7 2 15.4 65-69 10 1 10.0 0 0.0 0 0.0 0 0.0 70-74 13 1 7.7 0 0.0 0 0.0 0 0.0 70-74 10 0 0.0 0 0.0 0 0.0 0 0.0 75-79 20 4 20.0 0 0.0 1 5.0 1 5.0 75-79 2 0 0.0 0 0.0 0 0.0 0 0.0 +80 15 2 13.3 1 6.7 3 20.0 4 26.7 +80 6 0 0.0 0 0.0 0 0.0 0 0.0 Total 207 15 7.2 2 1.0 5 2.4 7 3.4 Total 262 5 1.9 4 1.5 0 0.0 4 1.5 ECHINOCARIS , A MID-PALEOZOIC CRUSTACEAN: AN ANNOTATED BIBLIOGRAPHY JOSEPH T. HANNIBAL Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 and RODNEY M. FELDMANN Department of Geology Kent State University Kent, Ohio 44242 Abstract Echinocaris is a genus of malacostracan crustacean that has been re- ported in rocks ranging in age from Early Devonian through Early M is- sissippian in North America, Europe, Asia, and New Zealand. More than 200 citations in primary and secondary literature have focused on one or more aspects of the taxon including its morphology, paleoecol- ogy, and phylogenetic position. This annotated bibliography represents a compilation of citations to works on the taxon and summarizes impor- tant observations in each reference. Introduction Purpose The genus Echinocaris embraces a small group of mal- acostracan arthropods known exclusively from the fossil record of the mid-Paleozoic. Twenty-seven species have been formally assigned to the genus and more than 20 citations have associated specimens with Echinocaris, which were so poorly preserved as to preclude precise identification. Of those identified and described, proba- bly more than 1 5 are valid species. The remainder are either synonymous with valid taxa or are assignable to different organisms. The described species range in age from Early Devonian through Early Mississippian. Al- though the preponderance of taxa have been described from marine rocks in the central and eastern United States, some have been identified in Canada, Great Brit- ain, the Soviet Union, Burma, and New Zealand. Thus, the echinocaridids are a relatively small group of organisms that one might assume would occupy a rela- tively insignificant position in the study of earth history. On the contrary, Echinocaris — or taxa that at one time were referred to the genus — and trace fossils presumed to be due to the activity of Echinocaris have been cited in more than 200 technical and nontechnical publications. Kirtlandia, No. 42 © by The Cleveland Museum of Natural History Echinocaris has often been illustrated (Fig. 1), and has been selected by authors as either the sole example, or one of a few examples, of Paleozoic crustaceans. Reference to the genus is made, not only in the primary literature, but also in a number of textbooks of paleontology and histor- ical geology published in the last 100 years. The prominence of Echinocaris in the literature proba- bly stems, in part, from the influential role played by clas- sical, early American paleontologists, such as James Hall, Robert Parr Whitfield, and John Clarke. Their work not only formed the basis for description of the rich and var- ied Devonian faunas of northeastern United States but also helped to set the standard of excellence for subse- quent paleontological work in America. These authors described in detail, and considered the implications of, Echinocaris in the record of Devonian rocks. With the exception of the trilobites and the eurypterids, Echinocaris was probably the most common large ar- thropod to be collected from Paleozoic rocks. As such, the taxon was recognized early as a potential ancestor to subsequent malacostracans and thus, served as a splendid illustration of the rootstock of this group. It was also rec- ognized, in the nineteenth century, that Echinocaris bore resemblance to the living leptostracans, including Nebalia and its allies. The presumed central position of Echino- caris in the phylogeny of the Malacostraca and its mor- phological similarity to living phyllocarids, coupled with its large size and distinctive morphology, rendered the genus a useful and lasting example of early arthropod. Because of the extraordinary prominence of Echino- caris in the literature and because the genus has been cited as the possible precursor to subsequent crustacean groups it is appropriate to focus attention on the knowledge of the group accumulated since the first specimens of Cera- tiocaris\-Echinocaris]punctata were described by Hal! in 1 863. The purposes of this work, therefore, are to present Fig. 1. Echinocaris punctata (Hall, 1863). A. Reproduction of the original illustration by Beecher (1884, Plate I, fig. 1 3), which is the most widely reproduced, and probably the best known, illustration of a specimen of Echinocaris. Beecher described the specimen as, “a nearly entire individual showing the form and relation of the parts, and the number of naked abdominal segments.” B. Photograph of the specimen of E. punctata , New York State Museum 13342/4 (old number, NYSM4401), from which the drawing was made. Bar scales = I cm. as complete a bibliography of the genus Echinocaris as possible and to provide annotations summarizing the sig- nificant systematic, morphologic, stratigraphic, and eco- logic observations in these works. Additionally, trace fos- sils ascribed to the work of Echinocaris spp. and taxa previously thought to be included within Echinocaris (particularly I Dunsopterus wrightianus and Eleuthero- caris wrightiana ) are treated. Finally, because many of the references cited treat a broad spectrum of Paleozoic ar- thropods, the bibliography will serve as a general entry into the literature of early malacostracans. Annotation procedure Every attempt has been made to secure as complete a bibliography on Echinocaris as possible. Standard biblio- graphic sources, such as the Bibliography and Index of Geology and its precursors, including the Bibliography of North American Geology and the Bibliography and In- dex of Geology Exclusive of North America , as well as the Zoological Record and Van Straelen and Schmitz’s Fos- silium Catalogus (1934) on phyllocarids, served as the basic resources. Additionally, references within articles on Paleozoic arthropods were searched for citations of the genus. Examination of stratigraphical and general pa- leontological articles relevant to the units known to con- tain echinocaridids provided additional references. A number of references were also drawn from W. D. E Rolfe’s manuscript of an updated manuscript Fossilium Catalogus on phyllocarids. Each of the citations has been examined to provide de- tailed page references to significant information relative to Echinocaris. Although these page citations refer to nearly all mentions of the genus, some passing notices have not been cited in situations where more inclusive or specific record is made in the same work. We have at- tempted to identify citations in textbooks and nontechni- cal literature, but we have not attempted to confirm ci- 1987 ECHINOCA RIS: AN ANNOTATED BIBLIOGRAPHY 27 tations in every edition of every textbook. Unpublished theses, with the exception of two which contain impor- tant, relevant remarks, are not included. In all cases, the original materials, or photocopies of the original mate- rials, have been examined so that errors of citation in sec- ondary literature have been identified and eliminated. Original spellings have been preserved and, where incor- rect, corrected spellings are indicated in square brackets. When the trivial portion of species names have been de- rived from the names of individual people, and their ini- tial letter was capitalized in the original article or book, they are not capitalized here (nor is this considered a misspelling herein). Authors of species, when not sup- plied by authors cited, are added. Two kinds of documenting information are given within the annotations. References to previous works and acknowledgments of illustrations cited by the author are presented in the text or are set off by parentheses. Those that were not cited by the original author, but are inserted by us, are set off by square brackets. Additionally, com- ments intended to clarify points of misinterpretation in the original text, to correct misspellings, or to guide the user to other, relevant references are placed in square brackets. Specimen numbers and depositories, when not provided previously in the literature, if known to us, are also provided in square brackets. To avoid misunderstanding, we have attempted to uti- lize standard terminology with reference to citations. Thus, when the term, “after” is used, it means that il- lustrative material has been taken directly from a pre- vious author, without modification. In the event that il- lustrations have been altered, but an original source can be identified, we have used the terminology, “modified from.” Cross references are included where a paper could be construed as having been written by an author other than the first author listed. Also, cross references are provided where an author’s name may have varied spellings in English. Acknowledgments W. D. 1. Rolfe, The Royal Museum of Scotland, Edin- burgh, generously allowed Hannibal access to his exten- sive files on Echinocaris , including his manuscript copy of an updated Fossilium Catalogus on phyllocarids, at an early stage of this work. Natalie Sidel, Cleveland Public Library, provided translations of materials in Russian. Rolfe, Loren Babcock, then at Kent State University, D. E. Butler, then with the British Geological Survey, and Andrew K. Rindsberg, Golden, Colorado, provided us with copies of important articles. Roy E. Plotnick, the University of Illinois, Chicago, provided information on ? Dunsopterus wrightianus. Kathleen M. Farago, Lake- wood Public Library, provided invaluable assistance with proofreading. Rolfe and Murray J. Copeland, Geological Survey of Canada, provided helpful reviews of this paper. This work was supported by NSF Grant EAR 83 1 2798 to Feldmann. A portion of this study was supported by a Cleveland Museum of Natural History Staff Enhance- ment Award, funded by Mr. and Mrs. Willard Hirsh, to Hannibal. Contribution 312, Department of Geology, Kent State University, Kent, Ohio 44242. Bibliography Allan, R. S. 1935. The fauna of the Reefton beds (Devo- nian), New Zealand; with notes on Lower Devonian animal communities in relation to the base of the Devonian System. New Zealand Department of Scien- tific and Industrial Research, Geological Survey Branch, Palaeontological Bulletin 14:1-72. Reported, based on an elongate, tapering, cercopod (p. 30; PI . Ill, fig. 8), Echinocaris sp. indet. from the Fig. 2. Illustrations of an arthropod spine (New Zealand Geological Survey AR1 131), originally identified by Allan (1935, p. 30) as Echino- caris sp. indet. The specimen does not, however, seem to be referable to Echinocaris. A. Counterpart of the sole specimen. B. Specimen illus- trated by Allan ( 1935, Plate 3, fig. 8). C. Enlargement of a portion of the specimen illustrated in B showing articulated spinelets arranged along the convex margin of the spine. Bar scales = 1 cm. 28 HANNIBAL AND FELDMANN No. 42 Reefton beds of New Zealand. [The specimen. No. AR 1131 in the collection of the New Zealand Geo- logical Survey, is from the Bolitho Mudstone For- mation, Middle Siegenian to Lower Emsian (Fig. 2). It is five or more sided and bears spinelets that are less than one mm long on at least two of these sides. It may be a phyllocarid telson-spine, but it is unlike that of any Echinocaris species.] Baird, G. C. 1978. Pebbly phosphorites in shale: a key to recognition of a widespread submarine discontinuity in the Middle Devonian of New York. Journal of Sedi- mentary Petrology 48:545-555. Reported (p. 548) the uppermost portion of the Kashong Shale Member of the Moscow Formation in western New York state to be, “characterized by the brachiopods Tropidoleptus carinatus and Lin- gula sp., ramose and fenestrate bryozoans, the trilo- bites Greenops boothi and Dipleura dekayi , and the phyllocarid Echinocaris sp.” 1979. Sedimentary relationships of Portland Point and associated Middle Devonian rocks in central and western New York. New York State Museum Bul- letin 433:1-24. Noted (p. 12) Tropidoleptus, Pleurodictyum [sic], Modiomorpha, Pseudoaviculopecten, Orthonata [.sic], Grammysia and Echinocaris as common taxa in the Middle Devonian Kashong Member of the Moscow Formation in central and western New York. Barrois, C. 1891. Memoire sur la faune du gres armori- cain. Annales de la Societe Geologique du Nord 19:134-237. Compared Trigonocarys lebescontei n. gen. and sp. with some other phyllocarids, including (p. 224) Equisitides [-1 Duns opt erus wrightianus ] and (p. 225) Echinocarvs [= Echinocaris-, Rolfe, 1969, noted that Echinocarys was a nomen vanum], Barron, L. S., and F. R. Ettensohn. 1980. A bibliography of the paleontology and paleoecology of the Devonian- Mississippian black-shale sequence in North America. U.S. Department of Energy, Morgantown Energy Technology Center, DOE/METC/5202-13, 86 p. Listed several publications dealing with, or mention- ing, Echinocaris . 1981. Paleoecology of the Devonian- Mississip- pian black-shale sequence in eastern Kentucky with an atlas of some common fossils. U.S. Department of Energy, Morgantown Energy Technology Center, DOE/ ET/ 12040-151, 75 p. Noted (p. 21) that, “fossils such as Echinocaris and Spathiocaris are commonly reported from the black shales” and that, “if they were indeed crustaceans, they probably led a nektonic life in the upper part of the water column of the black-shale sea and were as- sociated with flotage as are modern arthropods in the Sargasso Sea. ” [ Echinocaris is certainly a crusta- cean. However, unlike Spathiocaris, it is not com- monly found in classic black shales, such as the Cleveland Shale; see also Hlavin, 1976.] Bate, R. H., J. S. H. Collins, J. E. Robinson, and W. D. I. Rolfe. 1967. Arthropoda: Crustacea. In Die Fossil Record , p. 535-563. Geological Society of London. Noted (p. 555) the range of the order Archaeostraca Claus to be Tremadocian through Carnian. Bather. See British Museum (Natural History). Beecher, C. E. 1884. Ceratiocaridae from the Chemung and Waverly groups of Pennsylvania. Second Geologi- cal Survey of Pennsylvania, Report of Progress, PPPT-22. Reviewed (p. 1-3) literature on Paleozoic “phyllo- pods,” discussed (p. 3-4) “the optic spot” of these crustaceans, and listed (p. 5) geologic ranges for Echinocaris punctata (Hall), E. [ -"'Ceratiocaris ”] longicauda Hall [see Hall, 1863], E. sublevis Whit- field, E. pustulosa Whitfield, E. multinodosa Whit- field, and E. socialis Beecher. Redescribed E. punc- tata (p. 6-10; PI. I, figs. 13-16) [Fig. 1 herein] from the Hamilton Group of New York and described E. socialis n. sp. (p. 10-13; PI. I figs. 1-12) from the shales at the base of the Chemung Group at Warren, Pennsylvania. Also described mandibles (p. 9-10; PI. 2, figs. 9-11) associated with specimens of E. punctata , but belonging to “a species otherwise un- known.” Presented (p. 4) a labeled, diagrammatic il- lustration of Echinocaris. [Museum numbers for E. socialis, PI. I, figs. 5 and 6, E. punctata , PI. I, figs. 13-16, and for the mandibles, PI. II, figs. 9-11, are given in Clarke and Ruedemann, 1903. These au- thors listed the mandibles as E. punctata .] 1900. Restoration of Stylonurus lacoanus, a giant arthropod from the Upper Devonian of the United States. American Journal of Science, Fourth Series, 10:145-150. Also, Geological Magazine, New Series, 7(XI):48 1-485. Discussed (p. 148) Stylonurus (?) ( Echinocaris ?) Dunsopterus] wrightianus (Dawson), suggesting that the type specimen represents two proximal joints of one of the large crawling feet of a form re- lated to Stylonurus. 1902. Revision of the Phyllocarida from the Chemung and Waverly groups of Pennsylvania. Quar- terly Journal of the Geological Society of London 58:441-449. Described Echinocaris randallii n. sp. (p. 443; PI. XVIII, fig. 8) and E. clarkii n. sp. (p. 443-444; PI. XVIII, fig- 9) from the Waverly Group, Lower Car- boniferous, near Warren, Pennsylvania. Both forms were compared to E. socialis Beecher. E. clarkii was also compared to Pephricaris horripilata Clarke. 1987 ECHINOCARIS: AN ANNOTATED BIBLIOGRAPHY 29 Also presented additional information on the mor- phology of E. socialis (p. 441-442; PI. XVII; PI. XVIII, figs. 1 -7) based on specimens from the “phyl- locarid-beds” in the Upper Devonian Chemung Group at Warren, Pennsylvania. [Beecher’s primary type, and figured, specimens are in the Peabody Mu- seum of Natural History, Yale University.] Bernard, F. 1895. Elements de Paleontologie. Librairie J.-B. Bailliere et Fils, Paris, 1 168 p. Diagnosed (p. 328) Echinocaris and provided (fig. 159A) a labelled, diagrammatic illustration [modi- fied from Beecher, not Beecker, 1884] of E. punctata (Hall). Bigsby, .1. J. 1878. Thesaurus Devonico-Carboniferus: The Flora and Fauna of the Devonian and Carbon- iferous Periods. John van Voorst, London, 447 p. Listed (p. 26) Ceratiocaris armata Hall [= Echino- caris punctata (Hall)] and Ceratiocaris punctata Hall [= Echinocaris punctata (Hall)] from New York State, after Hall, 1863. Bolton, T. E. 1966. Catalogue of Type Invertebrate Fos- sils of the Geological Survey of Canada. Vol. III. Geo- logical Survey of Canada, Department of Mines and Technical Surveys, Ottawa, 203 p. Listed (p. 115) type and figured specimens of Echi- nocaris beecheri Copeland, E. castorensis Copeland, E. consanguina Eller, and E. sp. figured and de- scribed by Copeland in Copeland and Bolton, 1960. [See Copeland, 1960a.] Boule, M., and J. Piveteau. 1935. Fossiles: Elements de Paleontologie . Masson et cie, Paris, 899 p. Mentioned (p. 175) Echinocaris as a Devonian malacostracan. British Museum (Natural History). 1907. A Guide to the Fossil Invertebrate Animals in the Department of Geology and Palaeontology in the British Museum ( Natural History ) . . . The Trustees of the Museum, London, 182 p. Noted (p. 95) that the shield [carapace] of Echinocaris is bivalved. “This book has been written by Dr. Francis Arthur Bather . . .” [from the Preface], 1911 .A Guide to the Fossil Invertebrate Animals in the Department of Geology and Palaeontology in the British Museum ( Natural History) . . . Second Edition. The Trustees of the Museum, London, 183 p. Made the same comments (p. 95) as in the entry above. “The First Edition of this Guide . . . was written by Dr. Francis Arthur Bather . . . who has also . . . revised the book for this, the second, Edi- tion” [from the Preface], Brooks, H. K. 1957. Chelicerata, Trilobitomorphia, Crus- tacea (exclusive of Ostracoda) and Myriapoda. In Treatise on Marine Ecology and Paleoecology , Vol- ume 2, Paleoecology, edited by H. S. Ladd, p. 895-929. Geological Society of America Memoir 67. Noted (p. 898) the burrows on the carapace of a spec- imen of Echinocaris punctata (Hall) shown in Clarke, 1919 [= Clarke, 1921], Brown, R. W. 1956. Composition of Scientific Words. Revised edition. Smithsonian Institution Press, Wash- ington, D.C., 882 p. Included (p. 235) Echinocaris punctata (Hall) as an example of an animal named using the Latin caris. Buehler, E. J., and I. H. Tesmer. 1963. Geology of Erie County, New York. Buffalo Society of Natural Sciences Bulletin 21 (3): 1-1 18. Listed (p. 51) Echinocaris punctata (Hall) and E. sp. as occurring in the Ledyard Shale Member and (p. 56) Echinocaris punctata (Hall) and “mandibles of Phvllocaris" [.vzc] as occurring in the Middle Devo- nian Wanakah Shale Member of the Ludlowville Formation in Erie County, New York. Other taxa, including additional phyllocarids, occurring in these units are also listed. Butler, D. E. 1980. North Devon Athenaeum — Barnsta- ple, B. Figured Devonian fossils in the collections. The Geological Curator 2(9- 1 0): 588— 592. Listed (p. 59 1 ) specimens of Echinocaris whidbornei Jones and Woodward and E. sloliensis Partridge, figured by Partridge (1902), as being in the collec- tions of the North Devon Athenaeum. Supplied specimen numbers. 1981. Marine faunas from concealed Devonian rocks of southern England and their reflection of the Frasnian transgression. Geological Magazine 118: 679-697. Reported Echinocaris from Frasnian or early Fa- mennian rocks from the Little Chishill borehole (p. 689), Essex, England and the Steeple Aston borehole (p. 691), Oxfordshire, England. Carll, J. F. 1883. Geological Report on Warren County and the neighboring oil regions with additional oil well records. Second Pennsylvania Geological Survey Re- port: 1-439. Included (p. 304-307) a description of a stratigraph- ic section of the rocks in the Warren area by F. A. Randall. Reported (p. 306) crustaceans ( Ceratio - caris!) from division R at Tanner’s Hill near the brewery in western Warren. Also reported (p. 304) that fossils given to the state museum by Randall were keyed to this section. [The brewery mentioned was probably that of Adolph Saltsman, once located in King’s Hollow. The rocks noted as containing crustaceans are in the Conewango Formation (=Chadakoin & Venango formations). The crusta- ceans probably included one or more of the species of Echinocaris described in Beecher, 1884 and/or 1902, although it is not possible to reconcile Ran- 30 HANNIBAL AND FELDMANN No. 42 dall’s section with the exact horizons given by Beecher.] Case, G. R. 1982. A Pictorial Guide to Fossils. Van Nos- trand Reinhold, New York, 515 p. Suggested (p. 130-132) Pseudodontichthys whitei Skeels to be a junior synonym of Echinocaris. [This is probably incorrect; see Rolfe and Denison, 1966.] Illustrated (fig. 15-3) a specimen of P. whitei from the Silica Shale from Milan, Michigan and (fig. 15-6) Echinocaris socialis Beecher from the Upper Moravian beds of the Tully Limestone, Ithaca Group, Moravia, New York. [The latter specimen is not E. socialis. It is, however, comparable to E. punctata (Hall).] Caster, K. E. 1930. Higher fossil faunas of the Upper Al- legheny. Bulletins of American Paleontology 15(58): 1-332. Listed and illustrated Echinocaris socialis Beecher (p. 97; PI. 55, figs. 4, 6-8 after [modified from] Beecher, 1902) from the “Chemung Group” at Warren, Pennsylvania and E. clarkii Beecher (p. 97; PI. 54, fig. 5 after [modified from] Beecher, 1902) and E. randalli Beecher [=E. randallii] (p. 98; PI. 55, fig. 5, after [modified from] Beecher, 1902) from the Waverly Group at Warren. 1934. The stratigraphy and paleontology of northwestern Pennsylvania. Part I: Stratigraphy. Bul- letins of American Paleontology 2 1 (7 1 ): 1 - 1 85. Listed (p. 75) the following occurrences of Echino- caris in the “Bradfordian fauna” of Ohio, Pennsylva- nia, and New York; E. socialis Beecher in the Upper Chadakoin (Ellicott) Member, E. socialist and E. clarkiC. Beecher in the Amity Shale Member, and E. randalli Beecher [-E. randallii ] in the Oswayo — lower Riceville Member and the Cussewago series. Chadwick, G. H. 1935. Faunal differentiation in the Upper Devonian. Bulletin of the Geological Society of America 46:305-342. Listed Eleutherocaris whitfieldi (Clarke) and Cera- tiocarisf!) \=“ Ceratiocaris"] heecheri as a diagnostic species (p. 3 15) and Echinocaris punctata (Hall) as a nondiagnostic and last-appearing species of the Na- ples Group (p. 317). Listed Echinocaris condylepis Hall and Clarke as a diagnostic species of the Can- adaway Group (p. 325), Echinocaris socialis Whit- field as a diagnostic species of the Conewango Group (p. 330), Echinocaris clarkii Beecher and E. randalli Beecher [=iT. randallii ] as diagnostic of the Cus- sewago-Knapp (p. 366), and Stylonurus (?)[-r!Dun- sopterus ] wrightianus (Dawson) diagnostic of the Chemung Group (p. 320). Chamberlin, T. C., and R. D. Salisbury. 1905 (a publica- tion date of 1907 is indicated on the title page). Geol- ogy. Vol. II, Earth History. Genesis — Paleozoic. Sec- ond Edition, Revised. Henry Holt and Company, New York, 692 p. Noted that, in contrast to trilobites, phyllocarids and cirripeds were amply represented in the Hamilton fauna (p. 473), illustrating (fig. 210a) Echinocaris punctata (Hall) [modified from Beecher, 1884], 1909. A College Text-book of Geology. Henry Holt and Company, New York, 978 p. Illustrated (fig. 419d) [modified from Beecher, 1884] Echinocaris punctata (Hall) on a figure showing “representative Hamilton fossils,” and referred to the species as, “a crustacean more highly organized than the trilobites.” Chernyshev. See Tschernyshev Chhibber, H. L. 1934. The Geology of Burma. McMillan, London, 538 p. Listed (p. 168) Echinocaris asiatica Reed as one of the most important fossils found in the Devonian Wetwin shales of Burma. Chlup&c, I. 1963. Phyllocarid crustaceans from the Silu- rian and Devonian of Czechoslovakia. Palaeontology 6( 1 ):97— 1 18. Corrected (p. 98, 1 13) the family name Echinocari- dae Clarke (in Zittel, 1900) to Echinocarididae. Clarke, J. M. 1885a. A brief outline of the geological suc- cession in Ontario Co., N. Y., to accompany a map. Report of the [New York] State Geologist for the year 1884-. 9-22. Listed (p. 20) Echinocaris [= Eleutherocaris ] whit- fieldi Clarke as occurring in the “Portage” group in Ontario county. 1 885b. On the higher Devonian faunas of Ontar- io County, New York. U. S. Geological Survey Bulletin 16:1-86. Commented (p. 44-45) on Echinocaris and described Echinocaris [= Eleutherocaris] whitfieldi n. sp. (p. 45, PI. II, figs. 3, 4) from the Devonian Naples beds at Hatch Hill, Naples, New York. [Clarke (1902) later erected the genus Eleutherocaris for this species.] Also described Ceratiocaris [-“Ceratiocaris"] bee- cheri n. sp. (p. 44, PI. II, fig. 1) [see Clarke, 1892], Noted (p. 66) that he had not found any specimens of Equisetides wrightiana Dawson [=? Dunsopterus wrightianus (Dawson)]. 1891. The fauna with Goniatites intumescens , Beyrich, in western New York. The American Geolo- gist 8:86-105. Listed (p. 93) Echinocaris [= Eleutherocaris ] whit- fieldi Clarke and Ed beecheri Clarke [nomen nu- dum] as occurring in the G. intumescens fauna of the Naples beds. Noted, in introducing the list, that, “many interesting forms are undescribed and their affinities can only be indicated.” [Echinocarisl beecheri Clarke, 1891, not Copeland, 1960, has never been illustrated or described. Although it is probable that Clarke was referring to an undescribed form, it is possible that he could have been referring to “ Ceratiocaris ” beecheri Clarke, 1885.] 1987 ECHINOCARIS. AN ANNOTATED BIBLIOGRAPHY 31 1892. List of the original and illustrated speci- mens in the palaeontological collections, Part I Crustacea. 45th Annual Report of the New York State Museum : 373-437 and 1 1th Annual Report of the New York State Geologist'. 57-121. Listed type and figured specimens (including repro- ductions) of crustaceans in the collections of the New York State Museum. Included, under the main head- ing Echinocaris (p. 427-430), specimens of E. punc- tata ( Hall), mandibles of Phyllocarida ( E. punctata ), (?) Ceratiocaris [ Ceratiocaris ”] longicauda Hall, Echinocaris [-Eleutherocaris] whitfieldi Clarke, (?) Ceratiocaris [-” Ceratiocaris”] beecheri Clarke, E. condylepis Hall and Clarke, E. socialis Beecher, E. multinodosa Whitfield, E. pustulosa Whitfield, and E. sub/aevis [ -sublevis ] Whitfield. [It is unclear why C. beecheri was listed under Echinocaris.] Also in- cluded, (p. 426) Stylonurusl ( Echinocaris ?) [=? Dun- sopterus] wrightianus (Dawson). 1898a. The stratigraphic and faunal relations of the Oneonta sandstones and shales, the Ithaca and Por- tage groups in central New York. New York State Mu- seum, Forty-ninth Annual Report of the Regents , 1895, 2:27-81. Listed (p. 53) Echinocaris [-Eleutherocaris] whit- fieldi Clarke, and E. (?) beecheri Clarke [nomen nu- dum; see Clarke, 1891] in a list of species of the nor- mal Portage (Naples) fauna in Ontario and Living- ston counties. New York. 1898b. Notes on some crustaceans from the Chemung Group of New York. I. A singularly orna- mented phyllocarid genus, Pephricaris. New York State Museum, Forty-ninth Annual Report of the Re- gents, \S95,2:73\-733. Compared (p. 731-733) Pephricaris horripilata n. sp., from the Chemung Sandstone at Alfred, New York, with Echinocaris , and concluded that the new species was closely related to Echinocaris. 1902. Notes on Paleozoic crustaceans, 3, some Devonic Phyllocarida from New York. 54th Report of the New York State Museum , 1 (Appendix 3):97— 103. Noted (p. 98) that, although no phyllocarids were common in the Hamilton fauna, Echinocaris punc- tata (Hall) is more common than other species and found (p. 99) some Ithaca rocks at Laurens and near Noblesville crowded with E. punctata and Rhino- caris columbina Clarke. Also noted that these forms “in every recorded instance” are found in true marine faunas. Discussed (p. 103) the history of the name Eleutherocaris , noting that he had introduced the name in “Eastman-Zittel’s” Text-book of Paleontol- ogy for Ceratiocaris [= Eleutherocaris ] whitfieldi Clarke [undoubtedly Clarke was referring to the taxon then known as Echinocaris whitfieldi Clarke; seeZittel, 1900] and briefly compared Eleutherocaris to Echinocaris (p. 103). 1904. Naples fauna in western New York. New York State Museum Memoir 6:199-454. Listed (p. 352) Ceratiocaris ( Echinocaris ?) as occur- ring in the Kellwasser Limestone of the Eifel area [Germany], Included Eleutherocaris whitfieldi Clarke and Stylonurusl [=? Dunsopterus] wrightia- nus (Dawson) in a list of taxa (p. 358) found in the Cashaqua shale in the Naples section. Listed (p. 360) S.l wrightianus and E. whitfieldi as occurring, rarely, in the Naples subprovince, and Echinocaris ? longicauda Hall [-E. punctata (Hall)] as occurring in the Styliola or prenuncial fauna of the Naples sub- province. Noted (p. 373) Whitfield’s description of Echinocaris multinodosa, E. pustulosa , and E. sub- laevis [-E. sublevis] from the Huron [actually the Chagrin] Shale and that Echinocaris could be found in the Intumescens zone of New York. 1905. Ithaca fauna of central New York. New York State Museum Bulletin , 82:53-70. Listed (p. 61 ) Echinocaris punctata (Hall) as having been collected from the Ithaca beds at Norwich (p. 55) and Burdick (p. 58) and from the Sherburne Sandstone at Noblesville and Laurens (p. 59). 1921. Organic dependence and disease; their origin and significance. New York State Museum Bul- letin, 221-222:1-113. Also, Yale University Press, New Haven. Illustrated (fig. 86) with a specimen of Echinocaris punctata (Hall) from the Hamilton Group, consist- ing of a single left valve, with marks of Clionolithes borings. [The cover gives a date of publication of 1919 in addition to 1921, thus this article is some- times cited as being published in 1919. The specimen of E. punctata is Princeton University 89238, from the Moscow Shale, Smyrna, New York.] Clarke, J. M. See also Zittel, K. A. Clarke, J. M.,and D. D. Luther. 1904. Stratigraphic and paleontologic map of Canandaigua and Naples quad- rangles. New York State Museum Bulletin , 63, Paleon- tology, 7: 1-76. Listed (p. 59) Echinocaris ? [=“ Ceratiocaris”] longi- cauda Hall as occurring in the Genundewa Limestone of the Genesee [Group] and (p. 61) Eleutherocaris whitfieldi Clarke and Stylonurusl [=? Dunsopterus] wrightianus (Dawson) as occurring in the Naples fauna of the Cashaqua Shale of the Portage [Sonyea Group], Clarke, J. M., and R. Ruedemann. 1903. Catalogue of type specimens of Paleozoic fossils in New York State Museum. New York State Museum Bulletin , 65, Pa- leontology, 8:1-847. Listed museum numbers, rock units, localities, and other data for type specimens (including hypotypes and plastotypes) of Echinocaris condylepis Hall and Clarke (p. 700-701), E. multinodosa Whitfield (p. 701), E. punctata (Hall) (p. 701-703), E. pustulosa 32 HANNIBAL AND FELDMANN No. 42 Whitfield (p. 704), E. socialis Beecher (p. 704), and E. sublaevis Whitfield [ =E . sublev is] (p. 704), as well as Eleutherocaris whitfieldi Clarke (p. 705), Ceratio- caris [=“ Ceratiocaris”] longicauda Hall (p. 668) and Stylonurus (?) [=? Dunsopterus] wrightianus (Daw- son) (p. 761). 1912. The Eurypterida of New York. New York State Museum Memoir 14:1-439. Discussed (p. 311-312), provided a synonymy for (p. 311) and illustrated (text fig. 68) the type and sole specimen of Stylonurus (?) [=? Dunsopterus] wright- ianus (Dawson) (-Echinocaris wrightianus), and concurred with Beecher, 1900, that the specimen rep- resented the crawling legs of a form related to Stylo- nurus. [See also Dawson, 1881b, and Waterston, 1968.] Claypole, E. W. 1903. The Devonian Era in the Ohio basin. Part II, Devonian palaeontology of the Appa- lachian Gulf. American Geologist 32:240-250. Listed (p. 249), after Newberry, 1873, Ceratiocaris ? and Ceratiocaris sp. [not Ceratiocaris , but at least in part Echinocaris ; see Newberry, 1873] as occurring in the Erie [^Chagrin] Shale of Ohio. Cleevely, R. J. 1983. World Palaeontological Collections. British Museum (Natural History) and Mansell Pub- lishing, London, 365 p. Noted (p. 52) that C. E. Beecher donated specimens of Echinocaris [including a metatype of E. socialis], and Tropidocaris to the British Museum (Natural History) and (p. 225) that the North Devon Athe- naeum has material of E. M. Partridge, including Echinocaris whidbornei. [See also Butler, 1980.] Cleland, H. F. 1903. A study of the fauna of the Hamilton Formation of the Cayuga Lake section in central New York. U.S. Geological Survey Bulletin 206:1-112. Reported (p. 81) specimens of Echinocaris punctata (Hall) in the Upper and Lower Hamilton in the Cayuga Lake section. Listed (p. 104) E. punctata as occurring, very rarely, in the Michelinia zone, first Cvpricar della- Athyris zone, and second Cypricar- della-Athyris zone of the Lower Hamilton, and in the Cystodictya zone of the Upper Hamilton of the section. 1911. The fossils and stratigraphy of the Middle Devonic of Wisconsin. Wisconsin Geological and Nat- ural History Survey Bulletin, Scientific Series No. 6, 21:1-222. Described (p. 145-146), after Hall [and Clarke], and illustrated (PI. XLIV, fig. 3) a specimen of Echino- caris punctata (Hall) from Middle Devonian rocks of the Milwaukee cement quarry, Berthelet, Wiscon- sin. Also described (p. 146) and illustrated (PI. XLIV, fig. 4) a specimen consisting of an abdomen and telson that, “differs from those of New York E. punctata . . . and may be a new . . . species.” [Both specimens are now in the United States Na- tional Museum of Natural History, USNM 78216. See also Monroe and Teller, 1899.] Coogan, A. H., L. E. Babcock, J. T. Hannibal, D. W. Martin, K. S. Taylor, and D. C. Wehn. 1986. Late Devonian and Early Mississippian strata at Stebbins Gulch, Geauga County, and Quarry Rock, Cuyahoga County, Ohio. Field Trip Guidebook, Field Trip Number 1, Geological Society of America Northcen- tral Section Meeting, Kent, Ohio, 16 p. Noted (p. 4) that the Chagrin Shale, at some locali- ties particularly to the east of Stebbins Gulch, con- tains a fauna with Echinocaris, brachiopods, gas- tropods, bivalves, nautiloids, and other taxa. Copeland, M. J. 1960a. The occurrence of Echinocaris and Spathiocaris (Phyllocarida) in western Canada. In Canadian fossil Arthropoda, Eurypterida, Phyllocar- ida and Decapoda, edited by M. J. Copeland and T. E. Bolton. Geological Survey of Canada Bulletin 60: 1 - 1 1 . Described Echinocaris castorensis n. sp. (p. 4, fig. 1, PI. I, nos. 1-5) from the Upper Devonian Alexo Formation, “Beaver Ridge,” Alberta, comparing it to E. auricula Eller; E. consanguina Eller (p. 5-6, PI. I, nos. 6, 6a) from the same unit and locality, com- paring it to E. condylepis Hall and Clarke; E. sp. (tel- son) (p. 6, PI. 1, no. 7) from the Upper Devonian Duvernay Formation, Alberta; and E. beecheri n. sp. [not E. beecheri of Clarke (nomen nudum); see Clarke, 1891] (p. 6-7, fig. 2, PI. I, no. 8) from the Mississippian Banff? Formation, Alberta, compar- ing it to E. whidbornei Jones and Woodward and E. randalli Beecher [ -E . randallii]. Also discussed the stratigraphic position of these species (p. 2) and listed (p. 3) previously described species of Echino- caris, including “E.” [=? Dunsopterus] wrightiana (Dawson) (= Stylonurus ? wrightianus ) from North America. 1960b. New occurrences of Ceratiocaris and Pty- chocaris (Phyllocarida) from the Canadian Arctic. In Canadian fossil Arthropoda, Eurypterida, Phyllocar- ida and Decapoda, edited by M. J. Copeland and T. E. Bolton. Geological Survey of Canada Bulletin 60: 49-54. Diagnosed (p. 51) the Echinocaridae (sic) and pre- sented a key to the genera in the family: Echinocaris, Silesicaris, Eleutherocaris and Ptychocaris. Listed (p. 52) Echinocaris sp. McLaren (then a manuscript name) in the synonymy for Ptychocaris novaki n. sp. [See McLaren 1963.] Copeland, M. J., and T. E. Bolton. 1985. Fossils of Ontar- io. Part 3: The eurypterids and phyllocarids. Royal On- tario Museum Life Sciences Miscellaneous Publica- tions: 1-48. Noted (p. 39) and illustrated (fig. 25) the occurrence of Echinocaris sp. [£. cf. E. punctata], collected from the Middle Devonian, Arkona Formation, Hamilton Group, at Hungry Hollow, Ontario. Also mentioned 1987 ECHINOCARIS'. AN ANNOTATED BIBLIOGRAPHY 33 (p. 39) that Echinocaris pustulosus ( Hall) is “the typ- ical Middle Devonian representative of the genus in New York State.” [E. pustulosa Whitfield is a Late Devonian form; E. punctata (Hall) is the typical Middle Devonian form.] Dacque, E. 1921. Vergleichende biologische Formen- kunde der fossilen niederen Tiere. Gebruder Born- traeger, Berlin, 111 p. (Reprinted, 1980, by Arno Press, New York) Discussed (p. 289) Paleozoic phyllocarids, illustrat- ing (fig. 113) Echinocaris [ punctata (Hall)], from [after] Zittel, 1915 [which is, in turn, modified from Beecher, 1884], [The figure is reversed.] Dana, J. D. 1895. Manual of Geology , Lourth Edition. American Book Company, New York, 1087 p. Mentioned (p. 614-615) that there are a number of species of Echinocaris in Portage rocks and noted the occurrence of Echinocaris [= Eleutherocaris ] whitfieldi Clarke and E.l beecheri Clarke [nomen nudum; see Clarke, 1891] in the Portage Naples beds (p. 620-621, after Clarke 1891, 1892) and the occur- rence of E. socialis Beecher in the Chemung beds of New York and Pennsylvania (p. 621). Illustrated E. punctata (Hall) (fig. 920), [modified from] Beecher [1884], Dawson, J. W. 1881a. On new Erian (Devonian) plants (abstract). Canadian Naturalist , New Series, 9:475-476. Briefly described (p. 476) a new fern, Equisetites wrightianus n. sp., from New York. [See also Daw- son, 1881b.] 1881b. Notes on new Erian (Devonian) plants. Geological Society of London Quarterly Journal 37: 299-308. Described (p. 301, PI. XII, fig. 10; PI. Xll I, fig. 20) a new species of fern, Equisetides wrightiana [ =? Dun- sopterus wrightianus (Dawson)], from the Portage Group at Italy, New York. [Several names have been applied to this enigmatic species. Dawson himself re- ferred the species to both Equisetites and Equisitides in 1881. This taxon was subsequently considered re- ferable to Echinocaris , later, to the Eurypterida, and, most recently, questionably to the genus Dun- sopterus : see Waterston, 1968. We have not been able to locate the holotype of this enigmatic fossil, but casts of the original, and sole, specimen of the taxon are in at least two museums, the British Mu- seum (Natural History) and the New York State Mu- seum. A copy of correspondence between J. Hall and H. Woodward on the specimen is also deposited in the British Museum. One eurypterid specialist, R. Plotnick (personal communication) doubts that this specimen really is a portion of an eurypterid. Based upon the asymmetry of the specimen, as seen in the illustrations, and our examination of the material at the British Museum, it could even be a trace fossil.] 1882. The fossil plants of the Erian (Devonian) and upper Silurian formations of Canada. Part 2. Geo- logical Survey of Canada Report ; 93-142. [Report 429 in Ferrier’s catalogue.] Noted (p. 126) Wright’s discovery of Equisetites wrightiana Dawson [ -1 Dunsopterus wrightianus (Dawson); see Dawson, 1881b.] Derby, A. G. 1906. A bibliography of Ohio geology. Part 1. A subject index of the publication of the Geological Survey of Ohio. . . . Geological Survey of Ohio, Fourth Series, Bulletin 6:1-233. Indexed (p. 179) descriptions of Echinocaris in Whit- field, 1893. Duluk, C. E. 1964. Fossil fauna of the Silica Formation. Earth Science 17(6):250— 255. Reprinted in Fossils of the Mid-continent of North America , 1965, p. 37-42. Earth Science Publishing Co., Downers Grove, Illi- nois. Reported Echinocaris (p. 40) and “jaws” variously identified as fish and mandibles of crustaceans, “par- ticularly Echinocaris" (p. 40-41, fig. 15b), from Middle Devonian rocks exposed in the Medusa Port- land Cement Quarries in Sylvania, Ohio [see Dun- kle, 1965; Rolfe, 1966; and Stumm and Chilman, 1969, for further discussion of the identification of the mandibles], Dunkle, D. H. 1965. The presumed holocephalan fish Pseudodontichthys whitei Skeels. Scientific Publica- tions of the Cleveland Museum of Natural History, New Series, 4(2): 1-10. Discussed (p. 3-5) and illustrated (figs. 1 and 2; PI. 1 ) mouth parts, previously identified as representing a holocephalan fish, from the Silica Shale of Ohio. Observed that these mouth parts closely resembled those occurring with Echinocaris punctata as noted by Beecher ( 1884) and Hall and Clarke (1888). [See also Rolfe and Denison, 1966.] Dzik, J. 1980. Isolated mandibles of Early Palaeozoic phyllocarid Crustacea. Neues Jahrbuch fur Geologie und Palaontologie, Monatshefte 1 980( 2):87— 1 06. Observed (p. 97) that the mandibles of Montecarisl sp. from the Devonian of Poland had a gnathal lobe like that of Echinocaris , but with different tooth morphology. Also concluded (p. 101-102) that, “ar- chaeostracans were feeding on relatively firm food, which did not need cutting into pieces, but tritura- tion [sic],” consistent with a postulated habitat dose to the sea bottom. Eastman, C. R. 1913. See Zittel 1900. Easton, W. H. 1960. Invertebrate Paleontology. Harper and Brothers, New York, 701 p. Noted (p. 563) Echinocaris to be a characteristic Pa- leozoic genus of phyllocarid. Illustrated (fig. 1 3. 1 2.8) E. socialis Beecher, after [modified from] Hall and Clarke, 1888. 34 HANNIBAL AND FELDMANN No. 42 Edmonds, E. A., B. J. Williams, and R. T. Taylor. 1979. Geology of Bideford and Lundy Island. Geological Survey of Great Britain. Memoir for 1 :50 000 geologi- cal sheet 292, New Series, with sheets 275, 276, 291 and part of sheet 308 , 143 p. Reported (p. 20) Echinocaris sloliensis Partridge and E. whidbornei Jones and Woodward from the shaly sequence of the coastal sections of the Upper Devo- nian Baggy Sandstones (after Goldring, 1971) and (p. 21) E. sp. and E. whidbornei (PI. 2, fig. 10) from the inland sections of the Baggy Sandstones in a quarry near Croyde Hoe, Georgeham, England. Eller, E. R. 1935. New species of Echinocaris from the Upper Devonian, of Alfred Station, New York. Annals of the Carnegie Museum 24:263-274. Described Echinocaris consanguina n. sp., (p. 268- 269, PI. Ill, figs. 1-4), E. turgidan. sp., (p. 269-270, PI. Ill, figs. 5, 6) and E. auricula n. sp., (p. 271, PI. Ill, fig. 7) from the Upper Devonian [Famennian?] Alfred Shale [Canadaway Group] exposed at Alfred Station, New York. The species are compared with each other and with E. punctata (HaU), E. condylepis Hall and Clarke, E. socialis Beecher (also incorrectly spelled E. sociales) and E. sublaevis Whitfield [-E. sublevis], Eller also discussed (p. 264-267) several aspects of the morphology of Echinocaris finding no evidence of an eye on the “eye node” and doubting the existence of a “nuchal furrow” separating the ce- phalic and thoracic regions. 1937. Echinocaris crosbyensis , a new species from the Upper Devonian of New York. Annals of the Carnegie Museum 25:257-259. Described (p. 257-258) and illustrated (fig. 1) Echi- nocaris crosbyensis n. sp. from near Crosby Creek, New York, “possibly from the same horizon” as spe- cies previously described (Eller, 1935) from the Alfred Shale, and compared it with E. whidbornei Jones and Woodward, E. condylepis Hall and Clarke, E. socialis Beecher, and E. turgida Eller. Etheridge, R., H. Woodward, and T. R. Jones. 1884. Re- port of the committee . . . on the fossil Phyllopoda of the Palaeozoic rocks. Report of the fifty-third meeting of the British Association for the Advancement of Science held at Southport in September, 7553:215-223. Listed (p. 217) Echinocaris as a leperditioid phyl- lopod [sic] with four spiny exposed abdominal seg- ments and three telson styles or caudal spines. 1886. Third Report of the committee ... on the fossil Phyllopoda of the Palaeozoic rocks. Report of the fifty-fifth meeting of the British Association for the Advancement of Science held at Aberdeen in Sep- tember, 7555:326-361. Abstracted and commented on (p. 358-361) refer- ences to species of Echinocaris , and taxa at one time referred to that genus, in Hall, 1863; Whitfield, 1 880; Dawson, 1881; Beecher, 1884; and Jones and Wood- ward, 1884. 1889. Sixth report of the committee . . . on the fossil Phyllopoda of the Palaeozoic rocks. Report of the fifty-eighth meeting of the British Association for the Advancement of Science held at Bath in Sep- tember, 7555:173-181. Abstracted and commented on references to Echino- caris, and taxa at one time referred to that genus, in Clarke, 1885 (p. 175) and Hall and Clarke, 1888 (p. 180). Noted (p. 180) that Hall preferred the name E. punctata (Hall) over E. armata (Hall). 1891a. Seventh report of the committee ... on the fossil Phyllopoda of the Palaeozoic rocks. Report of the sixtieth meeting of the British Association for the Advancement of Science held at Leeds in Sep- tember, 7590:63-68. [A reprint of this article indicates that this paper was read at the Newcastle-upon-Tyne meeting.] Briefly commented (p. 63-64) on a newly discovered specimen of Echinocaris found in a quarry near Sloly, England. [This specimen was the basis for the description of E. whidbornei Jones and Woodward, see Jones and Woodward, 1889.] 1891b. Eighth report of the committee ... on the fossil Phyllopoda of the Palaeozoic rocks. Report of the sixtieth meeting of the British Association for the Advancement of Science held at Leeds in September, 7590:424-428. Noted (p. 426) the presence of Echinocaris in Devon and in the Devonian strata of North America. Feldmann, R. M., R. M. Boswell, and T. W. Kammer. 1986. Tropidocaris salsiusculus, a new rhinocaridid (Crustacea: Phyllocarida) from the Upper Devonian Hampshire Formation of West Virginia. Journal of Pa- leontology 60(2):379— 383. Noted (p. 379) references to Echinocaris sp. (Wil- liams and Kindle, 1905) and Echinocaris auricula Eller (Hannibal and Feldmann, 1985) as theonly rec- ords of the genus in West Virginia. Feldmann, R. M., and J. T. Hannibal. 1985a. Finger- printing fossils. Earth Science 38(1): 14-15. Commented on the biological relationships of Echi- nocaris, as presented by Rolfe, 1981. Illustrated a specimen of E. punctata (HaU) (p. 15), after Hall and Clarke, 1888, and included a map (p. 15) showing some locations in North America and England where echinocaridids [specifically Echinocaris ] have been found. 1985b. Paleobiogeography of Echinocaris. Geo- logical Society of America Abstracts with Programs 17(5):287. Reviewed the paleobiogeography of Echinocaris and reported a new West Virginia locality [see Hannibal 1987 ECHINOCARIS: AN ANNOTATED BIBLIOGRAPHY 35 and Feldmann, 1985] for Echinocaris. Questioned certain finds of echinocaridid “cercopods”[see Feld- mann, Hannibal and Babcock, 1986and Allan, 1935, herein] and found the genus “to be a good index in Middle and Late Devonian fine clastic rocks depos- ited in tropical and subtropical marine habitats lack- ing in shelly faunal elements.” Feldmann, R. M., 3. T. Hannibal, and L. E. Babcock. 1986. Fossil worms from the Devonian of North Amer- ica (. Sphenothallus ) and Burma (“Vermes”) previously identified as phyllocarid arthropods. Journal of Pa- leontology 60(2):341-346. Illustrated (fig. 3), and redescribed as “Vermes” (p. 343-345), fossils previously identified as “cerco- pods”of Echinocaris asiatica Reed [see Reed, 1908] from the Wetwin shales of Burma, [if. asiatica is in- correctly spelled E. asiaticus in the abstract of this paper.] Feldmann, R. M., and S. McKenzie. 1981. Echinocaris multispinosis, a new echinocarid (Phyllocarida) from the Chagrin Formation (Late Devonian) of Ohio. Journal of Paleontology 55(2):383— 388. Described Echinocaris multispinosis n. sp. (p. 385-386; text fig. 2) based on a single specimen from the Chagrin Shale (Famennian) at Indian Point, Lake County, Ohio. Noted (p. 386) that the type lo- cality of the new species, “may well be the type local- ity of the species described by Whitfield ( 1880).”[See Whitfield, 1880.] Found that arthropods, predomi- nantly Echinocaris , comprised 50% of the fauna found in a sample of concretions at this locality. Provided a key to the identification of arthropods, including E. pulcra Sturgeon, Hlavin and Kesling [=£. pulchra ], E. ohioensis Sturgeon, Hlavin and Kesling, E. sublaevis Whitfield [-E. sublevis ], and E. multinodosa Whitfield from the Chagrin and sug- gested that E. sublaevis and E. ohioensis are quite similar. Also discussed the paleoecology of Echino- caris and provided (text fig. 1 , modified from Stur- geon, Hlavin and Kesling, 1964) an illustration of a generalized Echinocaris. Feldmann, R. M., R. G. Osgood, Jr., E. J. Szmuc, and D. W. Meinke. 1978. Chagrinichnites brooksi , a new trace fossil of arthropod origin. Journal of Paleontol- ogy 52(2):287— 294. Noted (p. 292) that Echinocaris multinodosa Whit- field, E. sublevis Whitfield, E. pulchra Sturgeon, Hlavin and Kesling, and E. ohioensis Sturgeon, Hlavin, and Kesling have been described from the Famennian age Chagrin Shale in northeastern Ohio but concluded (p. 292-293) that these, and other, ar- chaeostracans had morphology that would make it difficult to ascribe the traces from the Chagrin de- scribed as Chagrinichnites brooksi n. sp. to them. Fenton, C. L., and M. A. Fenton. 1958. The Fossil Book. Doubleday, Garden City, New York, 482 p. Illustrated (p. 235) Echinocaris punctata (Hall) [modified from Beecher, 1884], Fisher, D. W. 1951. Marcasite fauna in the Ludlowville Formation of western New York. Journal of Paleon- tology 25:365-37 1 . Listed (Table 1) Echinocaris (?) sp. as occurring, rarely, in the marcasite horizon of the Ledyard Shale in western New York. Frey, R. W., H. A. Curran, and S. G. Pemberton. 1984. Tracemaking activities of crabs and their environmen- tal significance: the ichnogenus Psilonichnus. Journal of Paleontology 58(2):333— 350. Listed (p. 333) Chagrinichnites as one of several ich- nogenera of trace fossils of crustacean origin. Gekker, R. F. 1941. Deposits, fauna and flora of the main Devonian field. In Fauna of the Main Devonian Field, /, edited by R. F. Gekker, p. 1 7-84. USSR Academy of Sciences, Palaeontological Institute, Moscow. Noted (p. 51) the occurrence of Echinocaris in the main Devonian field. Listed, in a faunal distribution list (p. 65), Echinocaris tudrensis Tschernyshev as occurring in the Bilova Series (Famennian) and E. sp. as occurring in the Pskov beds (Frasnian). [In Russian, with English summary. See also Tscherny- shev, 1941.] 1983. Tafonomicheskiye i ekologicheskiye oso- bennosti fauny i flory glavnogo Devonskogo polya [Taphonomic and ecological characteristics of the fauna and flora of the main Devonian Field], Aka- demia Nauk SSSR. Trudy Paleontologicheskoga Insti- tuta 190:1-144. Summarized (p. 75, after Tschernyshev, 1941) in- formation on Echinocaris from the main Devonian field of the Russian Platform, contrasting the pau- city of specimens with the greater number of speci- mens of the genus found in the Devonian sediments of the Urals. Mentioned E. tudrensis Tschernyshev. [In Russian.] Goldring, R. 1971. Shallow-water sedimentation as illus- trated in the Upper Devonian Baggy Beds. Memoirs of the Geological Society of London , 5:1-80. Reported (p. 9) that Echinocaris sp. has been found in the Diplocraterion yoyo facies near Croyde Hoe, England and (p. 32) that E. sloliensis Coomaraswany [sic] [the author of the species is actually Partridge] and E. whidbornei Jones and Woodward have been associated with lingulids (p. 31) in the Lingula facies of the Baggy beds (Famennian). Speculated on the environment in which Echinocaris lived, noting that Echinocaris “seems to be very much a facies fossil” (p. 32), and mentioned Partridge’s description of a species of Echinocaris (p. 2). 1978. Devonian. In The Ecology of Fossils .edited 36 HANNIBAL AND FELDMANN No. 42 by W. S. McKerrow, p. 125-145. MIT Press, Cam- bridge. Illustrated (fig. 36) Echinocaris [with a somewhat disproportionately small carapace] in a reconstruc- tion of the sedimentary environment in which por- tions of the Upper Devonian Baggy Formation of North Devon, England, were deposited. Suggested (p. 140) Echinocaris to be associated with Lingula in the bay and coastal lagoon facies. One of the echino- caridids is shown (fig. 36) inside the burrow Diplo- craterion. Suggested that Echinocaris was probably epifaunal, based on the morphology of the carapace. Goldring, R., and F. Langenstrassen. 1979. Open shelf and near-shore clastic facies in the Devonian. Special Papers in Palaeontology 23:81-97. Reported (p. 89) phosphatized carapaces of Echino- caris to be common in the Lingula facies of the Baggy Sandstones (Famennian) of North Devon, England. Goldring, R., I. P. Tunbridge, A. Whittaker, et al. 1978. North Devon. In A Field Guide to Selected Areas of the Devonian of Southwest England , edited by C. T. Scrutton, p. 8-27. Palaeontological Association, Lon- don. Specimens of Echinocaris , as phosphatised internal molds, were reported (p. 23) to occur with Lingula and Diplocraterion in a sequence of interbedded shales and sandstones within the Famennian age Baggy Sandstones at Path Cove, near Baggy Point, England. Goldring, W. 1929. Handbook of Paleontology for Be- ginners and Amateurs. Part 1. The Fossils. New York State Museum Handbook 9: 1-356. Noted (p. 203) Echinocaris on exhibit in the New York State Museum, Albany, and illustrated (fig. 56B) E. punctata (Hall), after [modified from] Beecher, 1884 [the illustration is reversed], Grabau, A. W. 1921. A Textbook of Geology. Part II: Historical Geology. D. C. Heath and Co., Boston, 976 p. In a discussion of the Devonian (p. 424) noted that, “the fresh water (river and estuarine) deposits are characterized by the remains of crustaceans (fig. 1336) and eurypterids.” Illustrated (Fig. 1336b) Echinocaris punctata (Hall) [modified from Beecher, 1884] from the Hamilton. [The illustration is reversed.] Grabau, A. W.,and H. W. Shimer. 1910. North American Index Fossils: Invertebrates . Vol. II. A. G. Seiler and Company, New York, 909 p. Diagnosed Echinocaris (p. 376), noting that it had a free rostrum and no posterolateral spines [no ros- trum is known; most of the values of E. clarkei[-E. clarkii ] Beecher, including their posterolateral por- tions, are surrounded by spinose ornamentation] and Eleutherocaris (p. 380). Briefly described: E. punctata (Hall) (p. 376, fig. 1678) after [modified from] Beecher, [1884] [the figure is reversed]; E. so- cialis Beecher (p. 376-377, fig. 1679, after [modified from] Beecher, [1902]); E. sub/aevis Whitfield [=£. sublevis ] (p. 377, fig. 1680a&b) [modified from Whitfield, 1890]; E. pustulosa Whitfield (p. 378, fig. 1680c) (modified from Whitfield, 1890]; and E. mul - tinodosa Whitfield (p. 378, fig. 1 680d) modified from Whitfield, 1890J. Diagnosed Eleutherocaris and briefly described Eleutherocaris whitfieldi Clarke (p. 380). These species were also assigned (p. 686) to faunal provinces. Grasso, T. X. 1981. Stratigraphy, paleontology and pa- leoecology of the Upper Hamilton Group (Middle Devonian), in the Genesee Valley, Livingston County. In Guidebook to Field Trips, Annual Spring Meeting — National Association of Geology Teachers — Eastern Section, June 5-7, 1981, edited by R. M. Liebe, p. B1-B36. Reported the presence of Echinocaris in the Ka- shong Shale Member of the Moscow Formation ex- posed at Little Beards Creek (Stop #4, p. B— 36) near Leicester, New York. Brachiopods, bivalves, bry- ozoans, crinoids, trilobites and Echinocaris are listed (p. B— 8) as occurring in the Kashong fauna. The Tropidoleptus — bivalve community, in which Echi- nocaris occurs, is described (p. B— 1 6) as inhabiting, “soft substrates in shallow water of low energy, rem- iniscent of the Ledyard but containing higher levels of dissolved oxygen.” Also illustrated (fig. B— 1 3) Echinocaris punctata [after Beecher, 1884], Gtirich, G. 1929. Silesicaris von Leipe und die Phylloka- riden uberhaupt. Mitteilungen aus dem mineralogisch- geologischen Staatsinstitut in Hamburg 1 1:21-90. Described (p. 76-78) Echinocaris, commenting on the following Devonian species: E. whidbornei Jones and Woodward, E. punctata (Hall), E. socialis Beecher, E. sublaevis Whitfield [=£. sublevis ], E. condylepis Hall and Clarke, E. pustulosa Whitfield and E. multinodosa Whitfield. Distinguished Echi- nocaris from Ceratiocaris and Aristozoe. Noted (p. 82) that Pephricaris horripilata may belong to the ge- nus Echinocaris. Found (p. 82) Ceratiocaris^ Cera- tiocaris"\ longicauda Hall to be problematical. Re- described (p. 86) and commented on the Lower Carboniferous E. clarke i \-E. clarkii] Beecher, find- ing the serrated margin of its carapace to be reminis- cent of Pephricaris. Found (p. 86) the Lower Carbo- niferous E. randallii Beecher to be little distinguished from E. socialis. Also redescribed (p. 81-82) Eleu- therocaris whitfieldi Clarke, distinguishing the genus Eleutherocaris from Echinocaris . Illustrated E. punc- tata (Text plate 7, fig. 4) [modified from Hall and Clarke, 1888], E. socialis (Text plate 8, fig. 1) [modi- fied from Beecher, 1902], E. sublaevis [=E. sublevis] (Text plate 8, fig. 2) [modified from Hall and Clarke, 1987 ECHINOCARIS'. AN ANNOTATED BIBLIOGRAPHY 37 1888?], E. condylepis (Text plate 8, fig. 3) [modified from Hall and Clarke, 1 888], E. clarkei (Text plate 9, fig. 7) [modified from Beecher, 1902] and Eleuthero- caris whitfieldi (Text plate 8, fig. 10) [modified from Hall and Clarke, 1888], Hall, J. 1863. Contributions to palaeontology, no. 6, on the occurrence of crustacean remains of the genera Cer- atiocaris and Dithyrocaris, with a notice of some new species from the Hamilton Group and Genesee Slate. 16th Report of the New York State Cabinet of Natural History , Appendix D:71-75. Described Ceratiocaris armatus n. sp. (p. 72-73; PI. I, figs. 1-3) [-Echinocaris punctata ( Hall)] based on abdomen and telson material from the shales of the Hamilton Group, in Ontario County, New York and Ceratiocaris! punctatus n. sp. (p. 74; PI. I, fig. 8) [-E. punctata (Hall)] based on the left half of a carapace from the shales of the Hamilton Group on the east shore of Cayuga Lake. [Although E. armatus should have had priority over its synonym C. [-E.]punctata, most subsequent authors, with a few exceptions, in- cluding Jones and Woodward, 1884, have used the name E. punctata to refer to this species. As most authors, we refer to this species as Echinocaris punc- tata.] Also described (p. 73, PI. I, fig. 4-7) Ceratio- caris longicaudus n. sp. [-"Ceratiocaris" longicauda . This taxon has been referred to as Echinocaris longi- cauda in Beecher, 1884, and Miller, 1889, and listed under Echinocaris in Clarke, 1892. Hall and Clarke (1882) and Clarke (1892) noted that the specimens illustrated as fig. 4-6 were other than arthropods.] 1876. Illustrations of Devonian fossils: Gaste- ropoda, Pteropoda, Cephalopoda, Crustacea and cor- als of the Upper Helderberg, Hamilton, and Chemung groups. New York Geological Survey, Palaeontology. 39 pis. Illustrated [after Hall, 1863] an abdomen and telson referred to as Ceratiocaris armatus Hall \= Echi- nocaris punctata (Hall)] (PI. XXIII, figs. 4-5) and a carapace referred to as Ceratiocaris (Aristozoe) punctatus [=£. punctata ] (PI. XXIII, fig. 7). Noted (note accompanying PI. XXIII) that, “as this sheet is going to press” a specimen was found with a cara- pace like that of C. punctatus and an abdomen like that of C. armatus. 1884. [Note accompanying PI. XV] Thirty-fifth Annual Report on the New York State Museum of Natural History 35: PI. XV. Remarked on Equisetides wrightiana Dawson [-IDunsopterus wrightianus ], referring to it as a form not unlike Stylonurus. [See also Wright, 1884, and Dawson, 1881b. ] Hall, J., and J. M. Clarke. 1888. Trilobites and other Crustacea of the Oriskany, Upper Helderberg, Hamil- ton, Portage, Chemung and Catskill groups. New York State Geological Survey, Palaeontology 7 : i— lxiv, 1-236. Provided (p. liii) a synonymy for the genus Echino- caris. Diagnosed (p. liv) Echinocaris and compared the genus (p. liv— lv) to Aristozoe and Ptychocaris. Included (p. liv) a sketch of an Echinocaris[E. punc- tata, modified from Beecher, 1884], Described E. condylepis n. sp. from greenish shales of the Che- mung Group, Belmont (p. 173-174; PI. XXIX, figs. 14-17). Provided synonymies and described: Echi- nocaris punctata (Hall) from the Hamilton Group of New York State (including mandibles of Phyllocar- ida found associated with E. punctata ) (p. 166-171, PI. XXVII, fig. 10; PI. XXVIII, figs. 1-7 [figs. 6 and 7 modified from Hall, 1 863, figs. 3 and 5 modified from Beecher, 1884], PI. XXIX, figs. 1-8 [figs. 1-2 modi- fied from Beecher, 1 884] ); Echinocaris [= Eleuthero- caris] whitfieldi Clarke from the lower beds of the Portage (“Naples beds” of Clarke), Hatch Hill, Na- ples, New York (p. 172-173; PI. XXIX, figs. 20-21) [note the difference between fig. 21 and fig. 4 in Clarke, 1885; see also Rolfe, 1969, fig. 142,4b]; E. socialis Beecher from the base of the Chemung Group as exposed at Warren, Pennsylvania (p. 1 74- 1 76; PI. XXX, figs. 1-12) [figs. 1 -6, 8-9, 11-12 modified from Beecher, 1884]; E. sublaevis Whitfield [= E. sublevis ] from the Erie Shales [= Chagrin Shale] at LeRoy, Ohio (p. 176-178; PI. XXIX, figs. 11-13); E. pustulosa Whitfield from the Erie Shales at LeRoy (p. 178-179; PI. XXIX, figs. 9-10); and E. multino- dosa Whitfield from the Erie Shales at LeRoy (p. 180-181; PI. XXIX, figs. 18-19). Noted (p. 1 74) that E. condylepis is “almost a miniature of E. punctata." Provided (p. 160) a synonymy for and described (p. 160-162; PI. XXVII, figs. 7-9), as a possible stylo- nurid, the sole specimen of Stylonurus(!) (Echino- caris!) [=? Dunsopterus] wrightianus (Dawson). Also provided synonymies and described Ceratiocaris [-"Ceratiocaris"] longicauda Hall (p. 163-4, PI. XXXI, fig. 1 ), noting that some specimens originally referred to the taxa were, in fact, specimens of Coleo- lus aciculum Hall. [Many of the illustrations in this volume are similar to illustrations that have ap- peared previously in the literature, but were likely redrawn from original specimens or casts.] Hannibal, J. T., and R. M. Feldmann. 1981. Arthropod trace fossils, interpreted as echinocarid escape bur- rows, from the Chagrin Formation (Late Devonian) of Ohio. Geological Society of America Abstracts with Programs 13(7):467. [See Hannibal and Feldmann, 1983.] 1983. Arthropod trace fossils, interpreted as echinocarid escape burrows, from the Chagrin Shale (Late Devonian) of Ohio. Journal of Paleontology 57(4):705— 716. 38 HANNIBAL AND FELDMANN No. 42 Discussed (p. 713-714) the possible relationships of Echinocaris to the trace fossil Chagrinichnites os- goodi n. sp., concluding (p. 713) that, “it is most probable that the trace maker . . . was one or more species of Echinocaris." Illustrated (fig. 7) a general- ized Echinocaris associated with such a trace and (fig. 8) an Echinocaris with an exploded block dia- gram of the trace. [The trace had been described pre- viously by Lesquereux, 1891.] 1985. A phyllocarid crustacean, Echinocaris au- ricula, from the Late Devonian of West Virginia. Kirt- landia 41:22-26. Redescribed (p. 23-25) Echinocaris auricula Eller based on the holotype and an additional specimen from Upper Devonian Chemung rocks in Preston County, West Virginia. Discussed variation in the species and compared E. auricula to E. castorensis Copeland and other species. Illustrated the holotype (fig. 2A), its counterpart (fig. 2B), and the West Vir- ginia specimen (fig. 2C). 1986. Late Devonian Crustacea, Palaeopalae- mon newberryi and Concavicaris sp., from the Chagrin Shale of northeastern Ohio and Alfred Shale of west- ern New York state. Geological Society of America Abstracts with Programs 18(4):291. Noted that the Chagrin Shale of Ohio and the Alfred Shale of New York contained a fauna of arthropods dominated by Echinocaris. Hecker. See Gekker. Herries, R. S. 1896. Long excursion of West Somerset and North Devon. Proceedings of the Geologist's As- sociation 14( 10):433— 440. Reported (p. 440) the collecting of part and counter- part of a “fine specimen” of Echinocaris whidbornei Jones and Woodward “with an impression of an- other on the same slab” from the Baggy beds in Sloly Quarry, Georgeham, England. [See Jones and Wood- ward, 1899, p. 394 for a brief description of this specimen.] Hlavin, W. J. 1976. Biostratigraphy of the Late Devonian black shales on the cratonal margin of the Appalachian Geosyncline. Unpublished Ph.D. diss., Boston Univer- sity, 194 p. Listed (Table 3) the occurrence of “indet. phyllo- carid”and cf. Echinocaris sp. in the upper five feet of the Cleveland Shale on Townes Creek, Lorain Co., Ohio. [The specimens studied by Hlavin and depos- ited in the Cleveland Museum of Natural History, however, do not appear to include any Echinocaris .] Hoover, K. V. 1960. Devonian-Mississippian shale se- quence in Ohio. Ohio Department of Natural Re- sources, Geological Survey Information Circular Num- ber 27:1-154. Listed the occurrence (Appendix, p. 144) of Echino- caris multinodosa Whitfield, E. pustulosa Whitfield and E. sublevis Whitfield in the Upper Devonian Chagrin Shale of Ohio. Jones, T. R. 1883. II — Palaeozoic phyllopoda; as re- ported on to the British Association, Southport, 1883, Section C. Geology. Geological Magazine , Decade 2, 10:461-463. Listed (p. 463) Echinocaris as being leperditioid and as having four spiny exposed abdominal segments and three telson elements. 1898a. The fossil Phyllopoda of the Palaeozoic rocks. Thirteenth report of the committee consisting of Prof. T. Wiltshire (Chairman), Dr. H. Woodward, and Prof. T. Rupert Jones (Secretary). Report of the sixty- seventh meeting of the British Association for the Ad- vancement of Science held at Toronto in August, 1897: 343-346. Noted (p. 345-346) Whidborne’s (1896b) redescrip- tion of Echinocaris whidbornei Jones and Wood- ward and reported (p. 346) the discovery of two addi- tional specimens of Echinocaris \= E. whidbornei , see Jones and Woodward, 1899] from the Sloly Mudstone. 1898b. Ill — The fossil Phyllopoda of the Pa- laeozoic rocks. Geological Magazine , New Series, De- cade 4, 5:41-45. Essentially a reprint of Jones, 1898a. The portion on Echinocaris whidbornei Jones and Woodward (p. 44-45) is identical. 1 900. Fossil Phyllopoda of the Palaeozoic rocks — fifteenth report of the committee . . . Report of the sixty-ninth meeting of the British Association for the Advancement of Science held at Dover in September, 1899 68:403-405. Noted (p. 405) that one of the specimens of Echino- caris whidbornei Jones and Woodward discussed in Jones, 1898a and 1898b, was figured and redescribed in Jones and Woodward, 1899. Jones, T. R. See also Etheridge, Woodward, and Jones. Jones, T. R., and H. Woodward. 1884. Notes on phyllo- podiform crustaceans, referable to the genus Echinoca- ris, from the Palaeozoic rocks. Geological Magazine , New Series, Decade 3, 1:393-396, PL 13. Gave a history of the specimens referred to Ceratio- caris armatus Hall [= Echinocaris punctata (Hall)] (p. 393, PI. XII, fig. 2, after Hall [1863]) and its syn- onyms, noting that the trivial name armatus deserves priority [but see Hall, 1 863]. Reviewed (p. 394) Whit- field, 1880, and illustrated, for the first time, E. sub- laevis Whitfield [=£. sublevis] (PI. XIII, figs. 3-5) and E. pustulosa Whitfield (PL XIII, fig. 6) from a yet unpublished plate supplied by Whitfield. Recog- nized (p. 394) only three “well-determined” species of 1987 ECHINOCA RIS: AN ANNOTATED BIBLIOGRAPHY 39 Echinocaris : E. armata (Hall), E. sublaevis Whit- field, and E. pustulosa Whitfield. Redescribed (p. 395-396) the sole specimen of Equisetides wrighti- ana Dawson (PI. XIII, fig. la-b; A and B on p. 395) as representing a portion of the abdomen of a large Echinocaris, E. wright iana (Dawson) [=? Dunsop- terus wrightianus]. [See Dawson, 1881b.] 1888. A Monograph of the British Palaeozoic Phyllopoda (Phyllocarida, Packard). Part 1. Ceratio- caridae. The Palaeontographical Society, London, 72 p. Noted (p. 3), in a table of genera of fossil Phyllocar- ida, that Echinocaris was leperditioid, with spinous segments, and had a pod-like, oculate, bivalved carapace. 1889. I. On some new Devonian fossils. Geo- logical Magazine, New Series, Decade 3, 6:384-388. Described (p. 385-386) and illustrated (fig. 1) Echi- nocaris whidbornei n. sp., from the “leaden-blue shales of the Lingula squamiformis beds in a quarry near Sloly, close to the three-milestone on the Barn- staple and Ilfracombe road.” The authors distin- guished the specimen from E. punctata ( Hall) (illus- trated as fig. 2, after [modified from] Beecher [ 1884]) and other previously described species on the basis of the presence of two parallel ridges on the carapace of the species and other characters. [The holotype, H 419, is deposited in the Sedgwick Museum, Cam- bridge University. See also Whidborne, 1896b.] 1899. II. — Contributions to fossil Crustacea. Geological Magazine, New Series, Decade 4, 6:388- 395. Briefly described (p. 393-394, PI. XV, fig. 6) speci- mens of Echinocaris whidbornei Jones and Wood- ward from the Sloly Quarry [near Barnstaple], [See also Partridge, 1902.] Jux, U. 1960. Montecaris lehmanni, a new crustacean from the Rhenish Devonian and the problem of its sys- tematic position. Journal of Paleontology 34(6): 1 1 29— 1152. Noted (p. 1 147) the fossil assemblage containing Echinocaris socialis Beecher, described in Hall and Clarke, 1888, in a discussion of phyllocarid living habits. Discussed (p. 1 147-1 148) various aspects of the carapace of E. crosbyensis Eller, suggesting that the “eye notches” of Echinocaris and other phyllo- carids were “the original sites of eyes.” Also dis- cussed (p. 1148) the mandibles of phyllocarids, in- cluding those of Dilhyrocaris [that is, Echinocaris ]. Illustrated E. crosbyensis (Text fig. 4B) after [modi- fied from] Eller, 1 937 [not 1 935] and a specimen of E. punctata (Hall) (Text fig. 9F) after [modified from] Hall and Clarke, 1888. Kesling, R. V., and R. B. Chilman. 1975. Strata and mega- fossils of the Middle Devonian Silica Formation. Uni- versity of Michigan Museum of Paleontology Papers on Paleontology, Number 8:1-408. Gave (p. 155) a short synonymy for Echinocaris punctata (Hall) and briefly described the species as found in the Middle Devonian Silica Formation. Noted that the species occurs at the Martin-Marietta Quarry (in Washtenaw County, Michigan), as well as Silica, Ohio. Illustrated specimens of E. punctata ( PI. 73, fig. 1 ; PI. 93, fig. 5; PI. 1 29, figs. 6-9) [some of which were previously illustrated by Stumm and Chilman, 1969] and provided a diagram of the spe- cies (p. 155, after Stumm and Chilman, 1969). Kjellesvig-Waering, E. N. 1961 . Eurypterids of the Devo- nian Holland Quarry Shale of Ohio. Fieldiana, Geol- ogy 1 4( 5 ): 79— 98. Listed (p. 8 1) Stylonurus (?) [- 1 Dunsopterus ] wrigh- tianus (Dawson) in a list of eurypterids from the Devonian of North America. [See Dawson 1881b.] 1966. A revision of the families and genera of the Stylonuracea (Eurypterida). Fieldiana, Geology 14(9): 169-197. Recognized the species Stylonurus ? [=? Dunsop- terus ] wrightianus (Dawson), as questionably be- longing to the genus Stylonurus (p. 178). [See Daw- son 1881b.] Krestovnikov, V. N. 1960. Nadotryad Phyllocarida Pack- ard, 1879. In Osnovv Paleontologii, 8, Trilobitoobv- raznye i rakoobraznye, edited by N. E. Chernysheva, p. 425-429. Akademia Nauk USSR, Moscow. Diagnosed the Echinocaridae [szc] and Echinocariss (p. 426). Illustrated E. punctata (Hall) (fig. 1247) [modified from] Hall and Clarke, 1888 [which is, in turn, after Beecher, 1884], Mentioned (p. 427) Eleu- therocaris Clarke as a member of the family. [In Russian.] 1961. Novye rakoobraznye fillokaridy Paleozoya Russkoi platformy, Urala, Timana i Donbassa. Akademia Nauk SSSR, Trudy Geological Institute, Vpy, 52:1-67. Reviewed the occurrences of, and summarized strati- graphic information on, Echinocaridae [sz'c], includ- ing E. archae [=£■. arschae ] Tschernyshev, E. uralen- sis Tschernyshev, E. (?) brevicarinata Tschernyshev, E. tudrensis Tschernyshev, and various E. spp. Tschernyshev, from the Upper Devonian deposits of the Ural Mountains and the northwestern region of the Russian Platform (p. 10, 33, Table 5). Noted the similarity of Echinocaris sp. from the bank of the Arsha to E. randalii Beecher [-E. randallii], Pro- vided maps (figs. 1, 3) showing the distribution of Echinocaris in the Soviet Union. Also provided a chart (folded, between p. 8 and 9) showing the distri- bution of Echinocaris (worldwide) as Middle Devo- nian through Lower Carboniferous. [In Russian.] 40 HANNIBAL AND FELDMANN No. 42 La Touche, T. H. D. 1913. Geology of the northern Shan States. Memoirs of the Geological Survey of India 39(2): 1-379. Included Echinocaris asiatica Reed in a list of fauna from the Wetwin shales [in what is now Burma] (Ta- ble 9) and identified, as its nearest ally, E. punctata (Hall) from the Hamilton Group of North America. Lesley, J. P. 1889-1890. A dictionary of the fossils of Pennsylvania and neighboring states. Second Pennsyl- vania Geological Survey Report P4: 1-1283. Illustrated (p. 214) Echinocaris punctata (Hall) [modified from] Zittel after [modified from] Beecher [ 1 884] [the illustration is reversed] (and upside down as noted by J . Hall in the Errata, p. xxi) and Echino- caris [-Eleutherocaris] whitfieldi Clarke [modified from] Clarke, 1885. Echinocaris socialis is listed (p. 214) with the note “see Appendix.” [We could not, however, locate a reference to this taxon in any appendix.] Lesquereux, L. 1891. Remarks on some fossil remains considered as peculiar kinds of marine plants. Proceed- ings of the U.S. National Museum 13:5-12. Described (p. 9-1 1, PI . 1 , figs. 4-9), as a fossil plant, Physophycos bilobatus n. sp. from rocks of the Por- tage Group [Chagrin Shale] along Lake Erie, near Cleveland. [Later authors have overlooked this arti- cle and the name has been unused subsequently. Specimens have been redescribed as trace fossils under the generic name Chagrinichnites. A request to suppress the name P. bilobatus has been submit- ted to the International Commission on Zoological Nomenclature. Hannibal and Feldmann, 1983, in- terpreted one of these forms as having been pro- duced by Echinocaris .] McLaren, D. J. 1955. Devonian formations in the Alberta Rocky Mountains between Bow and Athabasca rivers. Geological Survey of Canada Bulletin 35:1-59. Reported (p. 47) well preserved carapaces of Echino- caris sp., along with Aulopora, Schuchertella, Cyr- tospirifer and other taxa, from a limestone unit in Member B of the Upper Devonian Alexo Formation at Beaver Ridge, Alberta. [The Echinocaris material is described in Copeland, 1960a.] 1963. Southwestern Ellesmere Island between Goose Fiord and Bjorne Peninsula. In Geology of the North-Central Part of the Arctic Archipelago, North- west Territories (Operation Franklin), by Y. O. Fortier et al. Geological Survey of Canada Mermoir 320: 310-338. Reported (p. 321) Echinocaris sp. [= Ptychocaris novaki Copeland] from a stromatoporoid-coral bio- herm within the limestone and shale (lower) member of the Blue Fiord Formation (Middle Devonian) on the south side of Eids Fiord, Ellesmere Island. [Copeland (1960a) described the “ Echinocaris " material reported here as Ptychocaris novaki n. sp.] Miller, S. A. 1889. North American Geology and Pa- laeontology for the Use of Amateurs, Students, and Scientists. Western Methodist Book Concern, Cincin- nati, 664 p. and Second Appendix, p. 719-793. Diagnosed (p. 545) Echinocaris and listed (p. 545-546, 787) articles describing E. condylepis Hall, E. longicauda Hall ( Ceratiocaris [=“Ceratiocaris”\ longicauda ), E. multinodosa Whitfield, E. punctata (Hall), E. pustulosa Whitfield, E. socialis Beecher, E. sublaevis (also, spelled correctly as E. sublevis) Whitfield, E. [= Eleutherocaris] whitfieldi Clarke, and E. wrightiana (Dawson)[-I Dunsopterus wright- ianus (Dawson)]. Illustrated (fig. 1002) E. punctata (Hall) [modified from Beecher, 1884; the figure is reversed], Monroe, C. E., and E. E. Teller. 1899. The fauna of the Devonian formation at Milwaukee, Wisconsin. Jour- nal of Geology 7:272-283. Listed (p. 279) Echinocaris (= Ceraliocaris ) [s/c] sp. as rare in the Devonian rocks in the lower 2 1 feet at the cement quarry on the Milwaukee River, imme- diately north of Milwaukee. [See Cleland, 191 1, fora description of Echinocaris from this locality.] Moore, R. C., and L. McCormick. 1969. General features of Crustacea. In Treatise on Invertebrate Paleontol- ogy, edited by R. C. Moore, Part R, Arthropoda 4(1), p. R57-R 120. Geological Society of America and Uni- versity of Kansas Press, Lawrence. Included (p. R 1 1 3) the Echinocaridinae and the Echinocarididae in an outline of classification of crustaceans. Moore, R. C., C. G. Lalicker, and A. G. Fischer. 1952. Invertebrate Fossils. McGraw Hill, New York, 766 p. Illustrated (fig. 14-17C) Echinocaris socialis Beecher [modified from Hall and Clarke, 1888] from Upper Devonian, Chemung, sediments of New York. Moret, L. 1940. Manuel de Paleontologie Animate. Mas- son et cie, Paris, 675 p. Considered (p. 294) Echinocaris as a Devonian rep- resentative of the Phyllocarida which typically in- habited shallow marine habitats. The phyllocarids were considered intermediate between entomostra- cans and malacostracans. 1966. Manuel de Paleontologie Animate. Fifth Edition. Masson et cie, Paris, 781 p. Mentioned (p. 329) Echinocaris as a Devonian rep- resentative of the phyllocarids. The phyllocarids were considered intermediate between entomostra- cans and malacostracans. Morris, S. F. 1980. Catalogue of the type and figured specimens of fossil Crustacea (excl. Ostracoda), Che- licerata, Myriapoda and Pycnogonida in the British Museum (Natural History). British Museum (Natural History), London, 53 p. Listed (p. 6) specimen 1.3945, Echinocaris whidbor- nei Jones and Woodward, figured by Jones and 1987 ECHINOCA RIS: AN ANNOTATED BIBLIOGRAPHY 41 Woodward, 1899, p. 393, PI. 15, fig. 6, and its coun- terpart, In. 38080, as being deposited in the British Museum (Natural History). Muller, A. H. 1963. Lehrbuch der Palaozoo/ogie. Band II, Invertebraten. Teil 3, Arthropoda 2 — Stomochorda. Gustav Fischer Verlag, Jena, 698 p. Diagnosed (p. 92) Echinocaris and illustrated (fig. 1 14) E. socialis Beecher from Roger 1953 [which, in turn, was modified from Beecher, 1902], Muller [Mueller], K.. J. 1967. Devonian of Malaya and Burma. In International Symposium on the Devonian System , edited by D. H. Oswald, Vol. 1:565-568. Al- berta Society of Petroleum Geologists, Calgary. Noted (p. 567), after Pascoe (1959), that the presence of Echinocaris [E. asiatica Reed] in the Devonian Wetwin fauna of Burma “seems to indicate the prox- imity of estuarine conditions.” [ Echinocaris , how- ever, does not indicate such conditions.] Murphy, J. L. 1972. Echinocaris punctata (Had) from the Hamilton Group, Thedford, Ontario. Ohio Journal of Science 72(3): 155- 157. Described and illustrated (figs. 1, 2) two specimens of Echinocaris punctata (Hall), USNM 170561 and 170562, from the Middle Devonian Hamilton Group in, and near, Thedford, Ontario. Greenops hoothi (Green), Ponderodictya punctulifera (Hall), Tor- noceras uniangulare (Conrad), “ Chonetes" lepida Hall, Styliolina fissurella (Hall) and Tasmanites sp. are listed as being associated with the Echinocaris. Newberry, J. S. 1873. Devonian System. Report of the Geological Survey of Ohio Vol. I, Geology and Paleon- tology. Part I, Geology: 140-167. Noted (p. 166-167) that concretions occurring in the Erie [=Chagrin] Shale, “in the bed of Paine’s Creek, in the north part of Leroy township,” Ohio, con- tained two crustaceans, one “probably a species of Ceratiocaris ,” the other an allied but new genus. [The crustaceans, Echinocaris and Palaeopa/aemon , were described by Whitfield in 1880.] Nicholson, H. A., and R. Lydekker. 1889. A Manual of Palaeontology for the Use of Students with a General Introduction on the Principles of Palaeontology. Third edition. Vol. 1. William Blackwood and Sons, Edin- burgh and London, 885 p. Noted (p. 5 14) that, “the curious genus Echinocaris ,” occurs in the Devonian of North America and briefly described its characters including, “an ovoid folded carapace.” O'Connell, M. 1916. The habitat of the Eurypterida. Bul- letin of the Buffalo Society of Natural Sciences XI(3): 1-277. Concluded (p. 23) that the single specimen of Stylo- nurus (?) wrightiana (Dawson) [=? Dunsopterus wrightianus ( Dawson)] was part of a jointed append- age. Provided a synonymy (p. 50) including Echi- nocaris wrightiana Jones and Woodward = Stylonu- rus (?) wrightianus (Dawson). [See Dawson 1881b.] Oehlert, D.-P. 1889. Sur le Devonien des environs d ’Angers. Bulletin Societe Geologique de France , Se- ries 3, 17:742-791. Reviewed (p. 770-771) the observations of Hall, Whitfield (1880), and Jones and Woodward (1884) that Echinocaris, Tropidocaris , and Elymocaris could confidently be assigned to the phyllocarids and were comparable to the living Nebalia and that Aris- tozoe, Callizoe, and Orozoe were probably also phyl- locarids, not leperditioid [ostracodes]. Olsson, A. 1912. New and interesting fossils from the Devonian of New York. Bulletins of American Paleon- tology 5(23): 1 -7. Reported (p. 7; PI. 7, figs. 2-4) two specimens [one of which, fig. 2, is now deposited in the collection of the Paleontological Research Institute, Ithaca, num- bered 28297] of Echinocaris punctata (Hall) col- lected from the Ithaca Shale beds of the Portage Formation in the McGraw, or University, Quarry on the Cornell University campus. These specimens were associated with Spirifer mesastrialis ( Hall) and Cryptonella eudora Hall. Olsson noted, “its appear- ance in higher beds [Frasnian], associated with re- current Hamilton species, and as having changed but little in the time interval.” Osgood, R. G., Jr. 1976. Trace fossils of Chagrin Forma- tion (Upper Devonian, Northeast Ohio). AAPG Bul- letin 60(4):704. Mentioned, “enigmatic resting and dwelling traces of what may be a phyllocarid or eocarid crustacean.” [The likely phyllocarid would have been Echino- caris. These trace fossils were later assigned to the genus Chagrinichnites\see Feldmann et al., 1978 and Hannibal and Feldmann, 1983.] Packard, A. S. Jr. 1882. The Palaeozoic allies of Nebalia. American Naturalist 16:945-953. Noted (p. 953) that, “ Echinocaris punctatus (Hall) [ -E. . punctata ] must have been nearly a foot in length, while the Echinocarides [s/o] . . . described recently by Mr. R. P. Whitfield were considerably smaller.” Illustrated an abdomen of E. punctatus (Hall) [=E. punctata ] (fig. 12), [modified] from Hall [1863], a carapace of E. multinodosus Whitfield [=E. multinodosa ] (fig. 10), [modified from? Whitfield’s unpublished plate], and an abdomen of E. sublevis Whitfield (fig. 1 1) [modified from? Whitfield’s un- published plate], 1883. A monograph of the phyllopod Crustacea of North America, with remarks on the Order Phyllo- carida. U.S. Geological and Geographical Survey of the Territories, Twelfth Annual Report , Part I: 295-592. Diagnosed (p. 450) Echinocaris , listed (p. 450-451) 42 HANNIBAL AND FELDMANN No. 42 E. sublevis Whitfield (fig. 7 IB, after [modified from?] Whitfield[’s unpublished plate]), E. punctatus (Hall) [ -E . punctata] (fig. 70, [modified] from Hall [1863]), E. armatus (Hall) = E. punctatus, E. pustu- losus Whitfield [ -E . pustulosa] and E. multinodosus Whitfield [-E. multinodosa ] (fig. 71 A, after [modi- fied from?] Whitfield['s unpublished plate]), and provided (p. 457) a bibliography of fossil species of phyllocarids. [The figures are the same as those in Packard, 1882.] Partridge, E. M. 1902. Echinocaris whidbornei (Jones & Woodward) and Echinocaris sloliensis, n. sp. Geologi- cal Magazine , Decade 4, 9:307-308, PI. 17. Described E. sloliensis n. sp. (p. 307-308, PI. XVII, figs. 8, 9) and the abdomen of Echinocaris whidbor- nei Jones and Woodward (p. 307, PI. XVII, fig. 7) based on specimens from the Marwood beds of Slo- ley Quarry, near Barnstaple, England. Compared (p. 308) E. sloliensis to E. socialis Beecher. [The deposi- tory of these specimens is recorded in Butler, 1980.] Pascoe, E. H. 1959. A Manual of the Geology of India and Burma. Vol. 2. Government of India Press, Cal- cutta, 2:485-1343. Listed (p. 688, after Reed, 1908) Echinocaris asiatica Reed as occurring in the Upper Devonian Wetwin fauna of Burma and noted that, “ Echinocaris is characteristic of brackish if not fresh water,” indicat- ing, “the proximity of estuarine conditions.” [The reference to salinity is incorrect.] Prosser, C. S. 1 898. The classification and distribution of the Hamilton and Chemung series of central and east- ern New York, Part 1. New York State Museum, Forty-ninth Annual Report of the Regents (New York State Geologist Annual Report 15), 1895, 2:87-222. Reported (p. 183) Echinocaris cf. [£.] punctata (Hall) as occurring, very rarely, at a shaly arenaceous horizon, “probably in the lower part of the Ithaca formation,” near Stetsonville, New York. Read, M. C. 1873. Geology of Lake County. Report of the Geological Survey of Ohio , Vol. I, Geology and Pa- leontology, Part I, Geology:5 10-519. Noted (p. 519) the occurrence of a new crustacean in nodules found in the Erie [Chagrin] Shale in Paine’s Creek, LeRoy Township, Ohio [This crustacean was probably Echinocaris ; see Whitfield, 1880.] Reed, F. R. C. 1908. The Devonian faunas of the northern Shan States. Palaeontologia Indica , New Series, 2(5): 1-183. Described (p. 179-180) and illustrated (PI. XX, fig. 21) Echinocaris asiatica n. sp. from the Devonian Wetwin shales near Wetwin, in what is now Burma. The description is based on a portion of a single left valve of a carapace. Reed also described (p. 180) and illustrated (PI. XX, figs. 22-25) “elongate styliform bodies” that he determined could be “regarded as the cercopods of a species of Echinocaris ,” probably E. asiatica. [These “cercopods” were redescribed by Feldmann, Hannibal and Babcock, 1986 as “Vermes.” Reed’s figured specimens are now depos- ited in the Geological Survey of India, Calcutta, with the type numbers 9335-9339; the carapace bears the number 9335.] Reimann, I. G. 1942. Hamilton phyllocarids in western New York. Buffalo Society of Natural Sciences, Bul- letin 1 7(3):48— 5 1 . Reported (p. 48) the occurrence of Echinocaris punc- tata (Hall) in the Ludlowville Shale at Wanakah and in the Ledyard Shale at Alden and Cazenovia Creek. Described (p. 49) as Echinocaris sp. nov.? a speci- men, differing from E. punctata by the shape and spinosity of the border of its carapace, from the Led- yard Shale at Alden. Reported E. sp., a fragmentary specimen, from the Ledyard Shale at Alden (p. 49) and two mandibles of undetermined phyllocarids from the Wanakah Shale (p. 48). Also noted (p. 48) that Echinocaris occurs in the Ludlowville along Lake Erie. Other phyllocarids were also reported. Roger, J. 1953. Sous-classe des malacostraces (Malacos- traca Latreille 1806). In Trade de Paleontologie , edited by J. Piveteau, Tome 3. Onychophores, Arthropodes, Echinodermes, Stomocordes, p. 309-378. Masson et cie, Paris. Diagnosed the family Echinocaridae [sic] and the genera Echinocaris and Eleutherocaris (p. 312) and illustrated a specimen of Echinocaris socialis Beecher from the Devonian of Pennsylvania (PI. I, fig- 5) [modified from Beecher, 1902], Rolfe, W. D. I. 1962a. The cuticle of some middle Silurian ceratiocaridid Crustacea from Scotland. Palaeontol- ogy 5(2):30— 5 1 . Noted (p. 48) Clarke’s (1921) figuring of borings, “Clionolithes ,” on Echinocaris . 1962b. Grosser morphology of the Scottish Silu- rian phyllocarid crustacean, Ceratiocaris papilio Salter in Murchison. Journal of Paleontology 36(5): 9 1 2-932. Noted (p. 917) that some Echinocaris species have nodes along the hinge line that differ from the tuber- cles on the valves of the carapace, referring to such nodes on E. punctata (Hall) and E. socialis ( Beecher). Also (p. 925) quoted remarks of Beecher (1884) on the mandible of E. punctata. 1962c. A new phyllocarid crustacean from the Upper Devonian of Ohio. Breviora 151:1-7. Noted (p. 4) that the dorsal head of the telson style of Ohiocaris wycoffi Rolfe is similar to that of Echi- nocaris sublevis Whitfield and (p. 5) that the smooth last abdominal segment of O. wycoffi differs from every species of Echinocaris. Also noted (p. 6) that the Harvard Museum of Comparative Zoology spec- imen of O. wycoffi is part of a collection containing E. multinodosa Whitfield and E. sublevis. 1987 ECHINOCA RIS. AN ANNOTATED BIBLIOGRAPHY 43 1969. Phyllocarida. In Treatise on Invertebrate Paleontology, edited by R. C. Moore, Part R, Arthrop- oda 4( 1 ), p. R296-R33 1 . Geological Society of Amer- ica and University of Kansas Press, Lawrence. Diagnosed, and corrected, when necessary, the names of (p. R3 1 7- R3 1 8) the Echinocarididae, the Echino- caridinae, Echinocaris and related taxa. Commented on various aspects of phyllocarid biology and ecol- ogy, including (p. R305) the mandibular palps of Echinocaris and (p. R303) Beecher’s correlation of carapace lobes to cephalic appendages, and (p. R308) the lack of evidence for archaeostracans being fluviatile or “continental.” Noted (p. R298) that prismatic structures are found in the integument of Echinocaris and provided a chart (fig. 122) of mor- phological features observed in various phyllocarid genera, including Echinocaris. Illustrated E. socialis Beecher (fig. 141,1), E. punctata (Hall) (fig. 142,3a&b) [after Hall, 1888, which, in turn, were modified from Hall, 1863], and E. randalliil Beecher (fig. 128). Diagnosed (p. R319) Eleutherocaris and illustrated (fig. 142,4a&b) rubber molds of the holo- type of Echinocaris (= Eleutherocaris ) whitfieldi (Clarke). 1981. Phyllocarida and the origin of the Mala- costraca. Geobios 14:17-27. Suggested (p. 18) homologies of the carapace grooves of Echinocaris with those of Montecaris and deca- pods, illustrating the carapace of Echinocaris multi- nodosa Whitfield (Fig. 1 A, C, drawn from Sturgeon et al., 1964). Reported (p. 21) and illustrated (fig. 3; PI. IB) apertures for limb insertion on abdominal somites 3 and 4 of Echinocaris punctata (Hall) from Kashong Glen, New York. Suggested (p. 21) that the large size, and large mandibles, of archaeostracans indicate that they were not filter feeders, but were probably macrophagous. Rolfe, W. D. I., and R. H. Denison. 1966. The supposed fish Pseudodontiehthys Skeels, 1962, is the phyllocarid crustacean Dithyrocaris. Journal of Paleontology 40(0:214-215. Concluded that mandibles erroneously identified as belonging to a chondrichthyan by Skeels (1962) “are identical with the ‘Mandibles of Phyllocarida’ fig- ured by Hall and Clarke . . . although it is still un- certain whether these belong to Dithyrocaris (= Me- sothyra) or to unusually large Echinocaris .” [See also Dunkle, 1965.] Rolfe, W. D. I., and V. A. Edwards. 1979. Devonian Ar- thropoda (Trilobita and Ostracoda excluded). Special Papers in Palaeontology 23:325-329. Discussed the stratigraphic value of phyllocarids, providing (text fig. 2) a chart showing the geologic age of species of Echinocaris, Eleutherocaris , and other Devonian phyllocarids, as well as their inferred phylogenetic links. The range of Echinocaris in the Devonian was given as earliest Givetian to late Fa- mennian. Also noted (p. 328) that revision of the Russian species of Echinocaris is needed. Rollins, H. B., N. Eldredge, and R. M. Linsley. 1972. Pa- leontological problems of the Hamilton Group (Mid- dle Devonian). In Field Trip Guidebook (44th Annual Meeting, Hamilton) New York State Geological Asso- ciation, edited by J. McLelland, p. F-l-F-28. Listed (p. F — 7) Echinocaris sp. from the Solsville Member, Marcellus Formation, at the Peterbo- rough South Quarry, near Morrisville, New York, and E. punctata ( Hall) from the Ludlowville Forma- tion at the “Pierceville” (=Bradley Brook) Quarry, Pierceville, New York (p. F— 18) and from the Mos- cow Formation at the Deep Spring Quarry, Leb- anon, New York (p. F-26). Sartenaer, P. 1969. Late Upper Devonian (Famennian) rhynchonellid brachiopods from Western Canada. Geological Survey of Canada Bulletin, 169:1-269. Lists (p. 4) Echinocaris sp. and other taxa as occur- ing in the Eoparaphorhynchus zone (Lower Fa- mennian) of western Canada. [The Echinocaris material was described in Copeland, 1960a as E. cas- torensis Copeland and E. consanquina Eller.]. Sass, D. B., and R. A. Condrate. 1985. Destruction of a Late Devonian ophiuroid assemblage: a victim of changing ecology at the Catskill delta front. In The Catskill Delta , edited by D. L. Woodrow and W. D. Sevon. Geological Society of America Special Paper 201:237-246. Noted (p. 238) that the earliest occupants of the en- vironment represented by the lower member of the Upper Devonian Alfred Shale, exposed in the vicin- ity of Alfred Station, N. Y., may have been malacos- tracans, for example Echinocaris, and infaunal bur- rowers. These were succeeded by a more diverse fauna. Also illustrated phyllocarids [including Echi- nocaris?\ in a reconstruction (fig. 6) of the sea bot- tom during “Alfred Station time.” Schram, F. R. 1986. Crustacea. Oxford University Press, New York, 606 p. Referred, after Krestovnikov, 1961, and Chlupaiuc, to echinocaridines as, “benthic reef-dwelling forms” (p. 328) and illustrated (fig. 26-4A) Echinocaris so- cialis Beecher [incorrectly identified as E. punctata (Hall)] (after Rolfe, 1969). Concurred (p. 329) with Rolfe (1969) that large mandibles possessed by ar- chaeostracans may not confirm a carnivorous habit but that, alternatively, they may have been scav- engers and detritus feeders. Schuchert, C. 1943. Stratigraphy of the Eastern and Cen- tral United States. Wiley, New York, 1013 p. Included (p. 105) Echinocaris punctata (Hall) in a list of taxa (after Clarke, 1905) from the Ithaca fauna (Upper Devonian, Naples Stage) of New York. Re- 44 HANNIBAL AND FELDMANN No. 42 ported (p. 106) Echinocaris [=Eleutherocaris] whit- fieldi Clarke, and, after Clarke, 1904, Eleutherocaris whitfieldi , from the Naples fauna of New York. Also reported (p. 112, after Caster, 1934) Echinocaris from the Devonian Conneaut beds of Pennsylvania. Scott, W. B. 1907. An Introduction to Geology. Second Edition. Macmillan, New York, 816 p. Illustrated (PI. IX, fig. 19) Echinocaris punctatus (Hall) [=£’. punctata ] on a plate illustrating Devo- nian fossils. [The illustration is a combination of figs. 1 and 2, PI. 28, in Hall, 1888, with a restored abdomen added.] Sepkoski, J. J., Jr. 1982. A compendium of fossil marine families. Milwaukee Public Museum Contributions in Biology and Geology , Number 51:1-125. Noted (p. 60) the range of the Echinocarididae as be- ing Siegenian through Tournaisian. Shimer, H. W., and R. R. Shrock. 1944. Index Fossils of North America. Technology Press, Massachusetts In- stitute of Technology, Cambridge, 837 p. Gave a diagnosis (p. 657) of Echinocaris and illus- trated (PI. 279) E. socialis Beecher (figs. 32, 39) and E. sublevis Whitfield [spelled incorrectly as E. sub- laevis in the figure caption] (figs. 37, 38, 41) after Hall and Clarke, 1888, and E. punctata (Hall) (fig. 42) after Hall and Clarke, 1888, which, in turn, is modified from Beecher, 1884. Skeels, M. A. 1962. Two new fishes from the Middle Devonian Silica Formation, Lucas County, Ohio. Journal of Paleontology 36(5): 1039- 1046. Described (p. 1043-1045, text figs. 2A-D, 3), as a chondrichthyan, Pseudodontichthvs whitei n. gen. and sp. [This fossil was later identified as represent- ing the mandible of a phyllocarid. Some authors have considered it to be Echinocaris. See Dunkle, 1965 and Rolfe and Denison, 1966.] Smith, B. 1935. Geology and mineral resources of the Skaneateles quadrangle. New York State Museum Bul- letin 300:1-120. Included Echinocaris punctata (WaM) in a list (p. 106) of conspicuous and characteristic Paleozoic fossils of the Skaneateles quadrangle having been noted in the Otisco Member of the Ludlowville Formation in the platform which underlies the Staghorn Point submember. Spinar, Z. 1960. Zaklady Paleontologie Bezobratlych. Ceskoslovenske akademie ved, Prague, 834 p. Mentioned (p. 598) Echinocaris as a Paleozoic phyl- locarid and illustrated (fig. X-85A) Echinocaris so- cialis Beecher after [modified from] Roger, 1953 [which, in turn, was modified from Beecher, 1902], Steinmann, G., and Doderlein, L. 1890. Elemente der Palaontologie. Wilhelm Engelmann, Leipzig, 848 p. Described (p. 501) Echinocaris and provided a dia- gramatic illustration (fig. 596) [modified from Beecher, 1884] of E. punctata (Hall). Straelen, Van. See Van Straelen. Stukel, D. J. II. 1986. Ichnology and environmental anal- ysis of the Chagrin Shale (Famennian), Northeast Ohio. Geological Society of America Abstracts with Programs 18(4):326. Commented on the occurrence of Zoophycos and Chagrinichnites within the same horizon of, and throughout, the Chagrin Shale of northeastern Ohio. Suggested, based on the occurrence of Lingulichnus and Chagrinichnites that the Chagrin represented a shallow water environment, deepening to the west. Stromer von Reichenbach, E. 1909. Lehrbuch der Pa- laozoologie. I. Teil. Wirbellose Tiere. B. G. Teubner, Leipzig, 342 p. Briefly discussed (p. 286) the Archaeostraca, noting that some forms probably had a double-valved shell and a spiniform telson. Illustrated (fig. 359) Echino- caris socialis Beecher, after Beecher, 1902, as an ex- ample. Stumm, E. C., and R. B. Chilman. 1969. Phyllocarid crus- taceans from the Middle Devonian Silica Shale of northwestern Ohio and southeastern Michigan. Con- tributions from the Museum of Paleontology, Univer- sity of Michigan 23(3):53— 7 1 . Provided (p. 65-66) a short synonymy and a diagno- sis for Echinocaris. Described E. punctata ( Hall) (p. 66, PI. 1, figs. 1-15; PI. 2, figs. 1 -5) from the Middle Devonian Silica Shale of Ohio, and reported the spe- cies as being abundant in the same beds as Rhino- caris. Included a description (p. 66) of the mandibles of E. punctata , with illustrations of a specimen from the Windom Shale of New York ( PI. 2, fig. 6) as well as the Silica Shale. Also described (p. 54, 60, 63, PI. 7, figs. 10-15) mandibles, belonging either to Dithyro- caris or Hebertocaris , some of which have been con- sidered by other authors (e.g. Dunkle, 1965) to be- long to, or possibly belong to, Echinocaris. Sturgeon, M. T., W. J. Hlavin, and R. V. Kesling. 1964. Rare crustaceans from the Upper Devonian Chagrin Shale in northern Ohio. Contributions from the Mu- seum of Paleontology, University of Michigan 19:47-64. Reported, described or remarked on, and illustrated, Echinocaris multinodosa Whitfield (p. 49-51, PI. I, figs. 1 -5; PL II, figs. 1-4; PL V, fig. 10), Echinocaris sp. cf. E. multinodosa (p. 51, PL V, figs. 7, 8), E. sub- levis Whitfield (p. 52, PL V, figs. 1, 2), E. pulchra n. sp. (p. 52-53, PL III, figs. 1 -5), E. ohioensis n. sp. (p. 53-55, PL IV, figs. 1-4), and E. sp. (p. 55-56, PL IV, fig. 5; PL V, fig. 3). All of the taxa reported were found in concretions from exposures in the Mill Creek area, Ashtabula County, Ohio. Echinocaris multinodosa was also reported from Euclid Creek, Cuyahoga County, Ohio. A composite diagram (fig. 1 ) of Echinocaris was also presented and some other fossils found in concretions in the Mill Creek area 1987 ECHINOCA RIS: AN ANNOTATED BIBLIOGRAPHY 45 were also described. [Weidner and Feldmann, 1985, redescribed some of these taxa, synonymizing E. pulchra with E. sublevis .] Szmuc, E. J. 1970. The Devonian System. In Guide to the Geology of Northeastern Ohio , edited by P. O. Banks and R. M. Feldmann, p. 9-21. Northern Ohio Geolog- ical Society. Noted (p. 13) that the fossil assemblages of the Cha- grin Shale are, “remarkable in that, except for iso- lated occurrences of arthropod remains (Echinoca- ris), they consist almost entirely of brachiopods . . . and mollusks.” Tchernychev. See Tschernyshev Tesmer, I. H. 1975. Geology of Cattaraugus County, New York. Buffalo Society of Natural Sciences Bulletin 27:1-105. Illustrated (PI. 5, fig. A) a specimen of Echinocaris, identified as E. cf. [if.] socialis Beecher, from undif- ferentiated Canadaway in Freedom Township, Cat- taraugus County, New York. The right valve of the carapace and much of the abdomen can be seen in Tesmer’s photograph. Tschernyshev, B. I. 1928. Phyllocaridae from the Devo- nian of the Urals. Annuaire de la Societe Paleontolo- gique de Russie 7:132-135. Described and illustrated (p. 132-133; PI. IX, figs. 1-3) Echinocaris arschae n. sp. and (p. 133-134, PI. IX, figs. 4, 5) Echinocaris n. sp. from the Devonian, “reef limestone overlying the Stringocephalus Bur- tini beds on the Arsha River,” in the South Urals. Briefly compared E. n. sp. to E. randallii Beecher and noted that the specimens referred to as E. n. sp. may be young individuals of E. arshae. [In Russian, with an English summary.] 1933. Arthropods from the Urals and other re- gions of the U.S.S.R. Paleontology and Stratigraphy Magazine , Leningrad 1:15-24. Described Echinocaris tudrensis n. sp. (p. 19, 23, PI. I, figs. 5-5a) from the Devonian rocks on the right bank of the M. Tudr River, at Bilovo Village. Com- pared the new species with E. condylepis Hall and Clarke, E. punctata (Hall), and E. socialis Beecher. [In Russian, with an English summary. See also Tschernyshev, 1941.] 1938. Some Phyllocarida of the Urals and the north-western district. Materials of the Central Geo- logical and Prospecting Institute , Leningrad ( Gen. Se- ries) 3:71-79. Discussed the occurrences of Echinocaris in the Urals and northwestern district, postulating that these phyllocarids lived in a district isolated from other phyllocarids and that their fossils were al- lochthonous. Discussed the genus Echinocaris (p. 73), mentioning the American species E. multino- dosa&nd E. clarke\-E. clarkii ]. Noted (p. 73-74, 78) the occurrence of E. arschae Tschernyshev (spelled, incorrectly, E. aschae on p. 73). Described £. uralen- sis n. sp. (p. 74, 78, PI. I, figs. 4-6) from Sukhoy Spring, E. sp. 1 (p. 74-75, 78, PI. I, fig. 7) from oppo- site Terebum Village along the Syas River, E. sp. II (p. 75, 78, PI. I, fig. 10) from Bogoslovskaya Dacha, E. sp. Ill (p. 75, 78, PI. I, fig. 1 1) from the Ura-tubin southern slope, and E. (?) brevicarinata n. sp. (p. 75, 79, PI. I, figs, 8-9) from the Bogoslovskaya Dacha and the Sungurduk River. [In Russian, with English summary. Rolfe and Edwards (1979) noted that E. uralensis and other Russian species were too poorly illustrated for identification, even to the generic level.] 1941. Phyllocarida of the Main Devonian Field. In Fauna of the Main Devonian Field /, edited by R. F. Gekker, p. 3 1 3-3 17. USSR Academy of Sciences, Pa- laeontological Institute, Moscow. Described the genus Echinocaris (p. 315) and de- scribed E. tudr ensis Tschernyshev, 1933 (p. 315-316, 317, PI. II, fig. lOa&b) from the Bilovo Series (of the Upper Variegated series, Famennian) near Bilova and E. sp. (p. 316-317, PI. II, fig. 1 1) from the Pskov beds (Frasnian) opposite Terebuny. Noted (p. 317) that phyllocarid remains are extremely scarce in the Main Devonian Field of the Russian Platform. [In Russian, with English summary. See also Gekker, 1941.] Twenhofel, W. H., and R. R. Shrock. 1935. Invertebrate Paleontology. McGraw-Hill, New York, 51 1 p. Noted (p. 444) that Echinocaris is a typical genus of fossil Nebaliacea and illustrated (fig. 163D), after Beecher, 1902, Echinocaris socialis Beecher. Van Straelen, V., and G. Schmitz. 1934. Crustacea Phyl- locarida (=Archaeostraca), pt. 64:1-246. In Eossi- lium Catalogus 1, Animalia , edited by W. Quenstedt. W. Junk, Berlin. Listed (p. 88-94, 95, 179-180, 182-183) most arti- cles published through 1930 that discussed, men- tioned, and/or illustrated Echinocaris and species and specimens that had been referred to the genus, including Eleutherocaris ( Echinocaris ) whitfieldi Clarke, “ Ceratiocaris ” longicauda Hall and “ Echi- nocaris" ( StylonurusI ) [='1 Dunsopterus] wrightia- nus (Dawson). Also corrected many names. [Many, but not all, of the corrections are used by us in this bibliography.] Articles are listed by taxon and syn- onyms are given. Vogdes, A. W. 1889. A catalogue of North American Pa- laeozoic Crustacea confined to the non-trilobite genera and species. New York Academy of Science Annals 5:1-38. Provided (p. 16) a diagnosis of Echinocaris after Hall and Clarke, 1888, and listed (p. 16- 18) articles refer- ring to the following species of Echinocaris : E. con- 46 HANNIBAL AND FELDMANN No. 42 dylepis Hall and Clarke, E. multinodosa Whitfield, E. punctatus Hall (also spelled E. punctata ), E. pus- tulosa Whitfield, E. socialis Beecher, E. sublevis Whitfield (also spelled E. sublaevis ), and Echinoca- ris \= Eleutherocaris] whitfieldi Clarke. Also listed (p. 35) articles referring to Stylonurus [ -1 Dunsopterus ] wright ianus (Dawson). The names used are those of the original authors. Entries for some forms referred to Echinocaris were also included under other ge- neric names, for instance, Ceratiocaris. Illustrated (p. 16) Echinocaris \punctata (Hall), modified from Hall and Clarke, 1888, which, in turn, was modified from Beecher, 1884], 1890. A bibliography of Paleozoic Crustacea from 1698 to 1889. U. S. Geological Survey Bulletin 63:1-177. Diagnosed (p. 16) Echinocaris. Listed (p. 16-18) ar- ticles referring to Echinocaris condylepis Hall and Clarke, E. multinodosa Whitfield, E. (= Ceratiocaris ? = Ceratiocaris - Ceratiocaris [Aristozoe]) punctatus (Hall) ( -E . armatus) [=£. punctata ], E. pustulosa Whitfield, E. socialis Beecher, E. sublevis Whitfield ( -E . sublaevis ), and E. whitfieldi Clarke. Also listed articles (p. 19, 35) referring to Stylonurus (-Equise- tides = Echinocaris = ? Echinocaris) wrightianus (Dawson). 1893. A classed and annotated bibliography of the Palaeozoic Crustacea, 1698-1892, to which is added a catalogue of North American species. Occa- sional Papers of the California Academy of Sciences 4:1-412. In an annotated list of works arranged alphabetically by author, noted species referred to Echinocaris (at one time or another) as being included in Beecher, 1884 (p. 18), Clarke, 1885 (p. 37), Hall, 1863 (p. 84), Hall, 1876 (p. 87), Hall and Clarke, 1888 (p. 88-89), Jones and Woodward, 1884 (p. 110), Etheridge, Woodward and Jones, 1885 [=1886] (p. 112) and 1889 [=1891] (p. 119), Packard, 1883 [= 1882] (p. 171), Whitfield, 1880 (p. 235), and Woodward and Jones [=Jones and Woodward], 1884, (p. 250). Also listed, in a separate list arranged alphabetically by taxon, articles mentioning E. condylepsis [ -E . condylepis ] Hall and Clarke, E. multinodosa Whitfield, E. ( Cer- atiocaris) punctatus ( Hall) [-E. punctata ], E. pustu- losa Whitfield, E. socialis Beecher, E. sublevis Whit- field (also spelled E. sublaevis), E. [-Eleutherocaris] whitfieldi Clarke, E. wrightiana (Dawson) (p. 383-384), Equisetides [-1 Dunsopterus] wrightiana (p. 385) and Stylonurus ( Equisetides ) [=? Dunsopte- rus] wrightianus (Dawson) (p. 407). Synonyms were supplied. 1917. Palaeozoic Crustacea. The publications and notes on the genera and species during the past twenty years, 1895-1917. Transactions, San Diego So- ciety of Natural History 3( 1 ): 1 — 1 4 1. Listed species referred to Echinocaris mentioned in various articles as follows: Echinocaris punctata (Hall) (p. 24) in Cleland, 1903; E. whidbornei Jones and Woodward (p. 46) in Jones and Woodward, 1899; E. socialis Beecher (p. 66) in Partridge, 1902 [not 1912]; E. sublevis Whitfield, E. pustulosa Whit- field and E. multinodosa Whitfield (p. 135) in Whit- field, 1893; and Stylonurus [=1 Dunsopterus] wright- ianus (Dawson) in Clarke and Ruedemann, 1912. 1925. Palaeozoic Crustacea. Part I, A bibliog- raphy of Palaeozoic Crustacea, supplementing the au- thor’s previous papers on the same subject. Transac- tions of the San Diego Society of Natural History 4:5-88.' Abstracted (p. 48) Reed’s 1909 description of Echi- nocaris asiatica Reed and noted (p. 88) the inclusion of Echinocaris in the 5th edition of Zittel’s Grund- zuge der Palaeontologie. Waterston, C. D. 1 968. Further observations on the Scot- tish Carboniferous eurypterids. Transactions of the Royal Society of Edinburgh 68:1-20. Noted (p. 12) the similarity of the massive prosomal appendages associated with Dunsopterus stevensoni (R. Etheridge, Jr.) to “the controversial specimen . . . still known as Stylonurus (?) wrightianus (Dawson)” and, on this basis, placed (p. 18) the ge- nus Dunsopterus in the Stylonuracea. Placed (Table 1) S’. (?) wrightianus questionably in Dunsopterus as Dunsopterus (?) wrightianus (Dawson) and that ge- nus tentatively in the Stylonuridae. [See also Daw- son, 1881b.] Weidner, W. E.,and R. M. Feldmann. 1983. Paleoecolog- ical interpretation of echinocarid arthropod assem- blages in the Late Devonian Chagrin Shale, northeast Ohio. Geological Society of America Abstracts with Programs 15(4):248-249. [See Weidner and Feldmann, 1985.] 1985. Paleoecological interpretation of echino- carid arthropod assemblages in the Late Devonian (Famennian) Chagrin Shale, northeastern Ohio. Journal of Paleontology 59(4):986— 1 004. Described five echinocarid-bearing concretionary horizons from two exposures of the Chagrin Shale (Famennian) in Lake and Ashtabula counties, Ohio. Remarked on the genus Echinocaris (p. 989) and on E. multinodosa Whitfield (p. 989, figs. 3. 5-3.9), E. ohioensis Sturgeon, Hlavinand Kesling(p. 989-990, figs. 4.9-4.11), E. sublevis Whitfield (p. 990-992, figs. 3. 1-3.4) and E. sp. (p. 992-995, figs. 4. 1-4.3). Considered E. pulchra Sturgeon, Hlavin and Kes- ling, 1964, to be a synonym of E. sublevis Whitfield, 1880, providing a synonymy of E. sublevis. Noted the occurrence of a presumed echinocarid trace, Chagrinichnites osgoodi (p. 995-996, figs. 5. 1-5.2) from along the Chagrin River. Divided the upper 1987 EC H IN OCA RIS\ AN ANNOTATED BIBLIOGRAPHY 47 portion of the Chagrin Shale into two ichnofacies, the Lingulichnus ichnofacies and the Chagrinich- nites ichnofacies. Discussed the sedimentological setting of Echinocaris in the Chagrin Shale and characterized the crustacean (p. 1002-1003) as an epifaunal detritus feeding arthropod, “well adapted to the turbulent, periodically nutrient-poor . . . offshore environment, owing to their mobility and success in utilizing a broad range of food.” Whidborne, G. F. 1896a. A preliminary synopsis of the fauna of the Pickwell Down, Baggy, and Pilton beds. Proceedings of the Geologist’s Association 14(9): 371-377. Listed (p. 371) Echinocaris whidbornei Jones and Woodward as occurring in the Marwood, Baggy and Sloly beds. [See Whidborne, 1896b.] 1896b. A monograph of the Devonian fauna of the south of England. Vol. 3, Part 1, The fauna of the Marwood and Pilton beds of North Devon and Somerset. London, Palaeontographical Society, 1 1 2 p. Redescribed (p. 6-7) Echinocaris whidbornei Jones and Woodward, 1889. Cited two specimens of the species, the type specimen (PI. I, fig. 3) from the Marwood series, near Sloly and another, fragmen- tary, specimen from Pilton (in the Porter Collec- tion). Whitfield, R. P. 1880. Notice of new forms of fossil crus- taceans from the Upper Devonian rocks of Ohio, with descriptions of new genera and species. American Jour- nal of Science , Third Series, 19:33-42. Proposed (p. 33-34) the name Echinocaris for Cera- tiocaris punctatus Hall, 1863, and three newly de- scribed species from Ohio. Provided a diagnosis of the genus (p. 34) and compared it to Ceratiocaris and other genera then considered members of the Cerati- ocaridae (p. 34-36). Described E. sublevis n. sp. (p. 36-37), designating it (p. 34) as the type species of the genus, E. pustulosa n. sp. (p. 38), and E. multinodosa n. sp. (p. 38-39) from calcareous concretions [the concretions are actually phosphatic; there is little calcareous material in them] in the Erie Shale [=Chagrin Shale, Famennian] at Leroy, Ohio. Also described the decapod Palaeopalaemon newberryi n. gen. and sp. (p. 39-42), found associated with Echi- nocaris. [Although this work was published without illustrations, plates were evidently supplied with the author’s special edition (fide Vogdes, 1893, p. 383) and other authors copied some of the illustrations on these plates. Feldmann and McKenzie ( 198 1, p. 386) suggested that Indian Point, Lake County, Ohio, may be Whitfield’s type locality. However Indian Point is actually in Perry Township, whereas New- berry ( 1873) indicated that the specimens were found in northern Leroy township. The type locality is probably somewhere along Paine’s Creek upstream of Indian Point, perhaps at Hell Hollow (Hell Hole), which is just to the east of the present town of Leroy Center. Both Indian Point and Hell Hollow are Lake County metropolitan parks. Permits are necessary to collect fossils in them.] 1890. Contributions to invertebrate paleontol- ogy [of Ohio], New York Academy of Science Annals 5:505-620. Described Echinocaris. Reported and described (p. 560-573) various fossils, including E. sublevis Whit- field (p. 565-567, PI. XII, figs. 12-14), E. pustulosa Whitfield (p. 567, PI. XII, fig. 15), and E. multino- dosa Whitfield (p. 568, PI. XII, fig. 16), from the Erie Shales [=Chagrin Shale, Famennian] at Leroy, Ohio. [Pages 562-572 of this article are essentially a reprint of Whitfield, 1880, which was published without il- lustrations. The specimens of Echinocaris described and/ or figured in this article are now in the Ameri- can Museum of Natural History, Columbia Univer- sity Collection, numbered 12281 and 12282 (E. sub- levis), 5512G (E. pustulosa ), and 551 1G, 12278 and 12280 (E. multinodosa ).] 1892. Discovery of a second example of the mac- rouran decapod crustacean, Palaeopalaemon newber- ryi. The American Geologist 9:237-238. Described a specimen of Palaeopalaemon newberryi that had been sent to Whitfield “under the name Echinocaris sp.” 1 893. Contributions to the paleontology of Ohio. Report of the Geological Survey of Ohio 7(2):407— 494. The sections dealing with or mentioning Echinoca- ris, “Fossils from the Erie Shales” (p. 452-462) and “Species from the Huron Shales” (p. 462-464) are reprints of p. 560-572 and p. 573-575 of Whitfield, 1890. The plate, PI. VIII, is a reproduction of PI. XII in Whitfield, 1890; both plate numbers are indicated at the top of the plate. 1 899. List of fossils, types and figured specimens, used in the palaeontological work of R. P. Whitfield, showing where they are probably to be found at the present time. New York Academy of Sciences Annals 12:139-186. Listed (p. 182) the type specimens of Echinocaris multinodosa Whitfield, E. pustulosa Whitfield, and E. sublaevis Whitfield [ =E . sublevis], as published in Whitfield, 1890, as being located at Columbia Col- lege. [These specimens are now at the American Mu- seum of Natural History; see Whitfield, 1890.] Willard, B. 1932. Devonian faunas in Pennsylvania. Penn- sylvania Geological Survey, Bulletin, Fourth series, G4: 1-43. Reported Echinocaris sp. in a faunule 3.8 miles east of Wellsboro, Tioga County (p. 28), E. cf. E. sublae- vis [ -E . cf. E. sublevis ] Whitfield in a faunule found in the Hollenback region, Bradford County (p. 31, 34), and E. cf. [£.] multinodosa Whitfield in a faunule found in Monroe County (p. 37, 40), Pennsylvania. 48 HANNIBAL AND FELDMANN No. 42 The age of the first two faunules is given as Che- mung; the age of the third as Ithaca. The first and third faunules are composed of marine elements, the second, found in redbeds and associated strata, is composed of marine and nonmarine elements, in- cluding the fish Bothriolepis. 1935. Hamilton Group of central Pennsylvania. Geological Society of America Bulletin 46: 195-224. Listed ( p. 2 1 1 ) Echinocaris punctata ( Hall) as occur- ring very rarely and E. sp. nov. as occurring rarely in the Ludlowville faunal facies of the Mahantango Formation of the Hamilton Group in Pennsylvania. Willard, B., F. M. Swartz, and A. B. Cleaves. 1939. The Devonian of Pennsylvania. Pennsylvania Geological Survey, Bulletin , Fourth Series, G 19: 1-481. Reported (p. 177) Echinocarisl from the Moscow beds (Hamilton) north of Auburn, Pennsylvania. Listed E. punctata (Hall) (p. 190, PI. 31, fig. 26) as occurring, very rarely, and E. sp. nov. as occurring, rarely, in the Ludlowville portion of the Mahan- tango Formation (Hamilton). Listed (p. 213) Echi- nocaris (?) sp. nov. as occurring in the Trimmers Rock fauna (Portage Group) of south-central Schuylkill County. Also, illustrated E. multinodosa Whitfield (PI. 31, fig. 25) [probably -E. cf. E. multi- nodosa of Willard, 1932] from the Trimmers Rock Sandstone in Monroe County and E. sublaevis Whitfield [=£■. sublevis , probably E. cf. E. sublaevis in Willard, 1932] from the Canadaway in Bradford County. [The specimens are too poorly illustrated to judge the accuracy of the identifications of the echi- nocaridids, even to the generic level.] Williams, H. S., and E. M. Kindle. 1905. Contributions to Devonian paleontology, 1903. Part 1. Fossil faunas of the Devonian and Mississippian (“Lower Carbonifer- ous”) of Virginia, West Virginia, and Kentucky. United States Geological Survey, Bulletin 244, Part 1 :9-58. Reported (p. 37; chart facing p. 55) the rare occur- rence of Echinocaris sp. in a Devonian (Chemung) faunule near White Sulphur Springs in southeastern West Virginia. [The whereabouts of the document- ing specimen(s) is unknown.] Woods, H. 1909. Palaeontology: Invertebrate. Fourth edition. The University Press, Cambridge, England, 388 p. Mentioned (p. 323) that Echinocaris is found in the Devonian and (p. 322) that the Leptostraca (Phyllo- carida) “are all marine, and live mainly in shallow water or at moderate depths.” Woodward, H. P. 1943. Devonian System of West Vir- ginia. West Virginia Geological Survey [ Report] 15:1- 655. Listed (p. 366) Echinocaris punctata (Hall) as occur- ring in Hamilton rocks elsewhere than West Vir- ginia. Wright, B. H. 1884. Notes on the geology of Yates County, N.Y. Thirty-fifth Annual Report on the New York State Museum of Natural History 35:195-206. Described (p. 196-197), after Dawson, 1881b, and illustrated (PI. XV, figs. 1-3, figs. 1 and 3 after [mod- ified from] Dawson, 1881b) Equisetides wrightiana Dawson \=r!Dunsopterus wrightianus (Dawson)]. [See also Hall, 1884.] Zell, P. D. 1985. Paleoecology and stratigraphy of the Middle Devonian Moscow Formation in the Che- nango Valley, New York. Unpublished M. S. thesis, University of Pittsburgh, 128 p. Found phyllocarids, particularly Rhinocaris colum- bina Clarke and Echinocaris punctata (Hall) to be rare but persistent elements in the Moscow fauna of the Chenango Valley (p. 55) with the latter, “more common in the coarse siltstones in the upper, shal- lower, portions of cycles.” Reported E. punctata as occurring in a Devonochonetes- Mucrospirifer com- munity (p. 44, Table 6) and a Tropidoleptus commu- nity (p. 50, Table 8). Listed E. punctata (p. 83, 1 18, 122) from six localities in the Chenango Valley. Commented on the life habits of E. punctata (p. 55-56), finding it to be, “an epifaunal scavenger/ predator on silty substrates with currents of moder- ate energy.” Reported (p. 56) the epizoans IConcho- trema (Appendix D, PI. 2.4) and Orbiculoidea doria (Appendix D, PI. 1.3) associated with the carapace of E. punctata. Zittel, K. A. 1880-1885. Handbuch der Palaeontologie. Part I. Palaeozoologie. Vol. II. Mollusca und Ar- thropoda. R. Oldenbourg, Munich, 893 p. Described (p. 658) Echinocaris and illustrated (fig. 846) E. punctata (Hall), Beecher, 1884. [The illustra- tion is reversed. This same picture is reproduced in the following Zittel entries.] 1887. Traite de Paleontologie. Tome II. Paleo- zoologie. Partie I. Mollusca et Arthropoda. Octave Doin Editeur, Paris, 897 p. Described (p. 655) Echinocaris and illustrated (fig. 863) E. punctata ( Hall), [after an earlier Zittel which, in turn, was] after [modified from] Beecher [1884], [The illustration is reversed.] Echinocaris is included in a table (p. 658) listing the geologic distribution of phyllocarids. 1900. Text-book of Palaeontology .Vol. 1, edited by C. R. Eastman. Macmillan and Co., New York, 706 p. J. M. Clarke, in this work, named and diagnosed the Echinocaridae [j/c] (p. 655) [see Chlupdc, 1963], and diagnosed Echinocaris (p. 655-6) and p. 656) Eleutherocaris [the latter for the first time, but without an illustration or a mention of any species be- longing to it; see Clarke, 1902.] Illustrated (fig. 1373) 1987 ECHINOCARIS : AN ANNOTATED BIBLIOGRAPHY 49 Echinocaris punctata (Hall), [after an earlier Zittel which, in turn, was] after [modified from] Beecher, 1884. [The illustration is reversed.] Also illustrated “gastric teeth” [mandibles] of E. punctata (fig. 1369a&b) [modified from Hall and Clarke, 1888, which, in turn, were modified from Beecher, 1884], 1903. Grundziige der Palaontologie (Palaozool- ogie). Second edition. I. Abteilung: Invertebrata. R. Oldenbourg, Munich, 558 p. Briefly described (p. 515) Echinocaris and illustrated (fig. 1319) E. punctata (Hall), [after an earlier Zittel which, in turn, was] after [modified from] Beecher, 1884. [The illustration is reversed.] 1913. Text-book of Paleontology, V ol. 1, Second Edition, edited by C. R. Eastman. Macmillan and Co., London, 839 p. J. M. Clarke, in this work, diagnosed (p. 751) the Echinocaridae [szc], Echinocaris , and (p. 752) Eleuth- erocaris. Illustrated (fig. 1455) E. punctata (Hall), [after an earlier Zittel which, in turn, was] after [modified from] Beecher, [1884], [The illustration is reversed.] Also illustrated “gastric teeth” [mandi- bles] of Echinocaris punctata (Figure 1451) [modi- fied from Hall and Clarke, 1888], [At least one re- print edition exists with a later date of publication; that one is dated 1937.] 1915. Grundziige der Palaontologie (Palaozool- ogie). Neubearbeitet von Ferdinand Broili. 1. Abtei- lung: Invertebrata. R. Oldenbourg, Munich, 694 p. Diagnosed the Echinocaridae [sz'r] and briefly de- scribed Echinocaris ( p. 626). Illustrated (fig. 1368) E. punctata (Hall), [after an earlier Zittel which, in turn, was] after [modified from] Beecher, [1884], [The illustration is reversed.] 1924. Grundziige der Palaontologie (Palaozool- ogie). Neubearbeitet von Ferdinand Broili. Vol. 1, In- vertebrata. Sixth and revised edition. R. Oldenbourg, Munich, 733 p. Diagnosed the Echinocaridae [szc] and Echinocaris (p. 660) and illustrated (fig. 1 377) E. punctata ( Hall), [after an earlier Zittel which, in turn, was] after [modified from] Beecher [1884], [The illustration is reversed.] Mentioned the presence (p. 659) of “eye tubercles” on Echinocaris. Index to Generic and Species Taxa Cited The index lists citations, by author and date, for all ref- erences to generic and species taxa of fossil arthropods (excluding trilobites) referred to, or cited as being illus- trated, in annotations [but not our bracketed comments] in the bibliography. However, no citations to the genus Echinocaris itself are listed, as such a list would be too long to be useful. In addition to the references to taxa assigned, at one time or another, to Echinocaris , the index also lists cit- ations to other arthropod taxa made in the original litera- ture and mentioned in the annotations. The references to these arthropods are not exhaustive but do, in many cases, provide an entry into the literature on these forms. The combinations and spellings presented in the index are those of the original author and do not reflect the most recent combinations or the correct spelling of the taxon, necessarily. Current combinations, correctly spelled, are indicated in the index by boldface type. Aristozoe: Giirich, 1929; Hall and Clarke, 1888; Oehlert, 1889. Callizoe: Oehlert, 1889. Ceratiocaris: Giirich, 1929; Newberry, 1873; Vodges, 1889; Whitfield, 1880. Ceratiocaris?: Carll, 1883: Claypole, 1903. Ceratiocaris sp.: Claypole, 1903. Ceratiocaris (Echinocaris?): Clarke, 1904. Ceratiocaris armata Hall: Bigsby, 1878. Ceratiocaris armatus Hall: Hall, 1863; Hall, 1876; Jones, 1884; Jones and Woodward, 1884; Packard, 1883. Ceratiocaris beecheri Clarke: Clarke, 1885b; Clarke, 1892. Ceratiocaris (?) beecheri Clarke: Chadwick, 1935; Clarke, 1892. Ceratiocaris longicauda Hall: Giirich, 1924. ? Ceratiocaris longicauda Hall: Clarke, 1892; Hall and Clarke, 1888. “Ceratiocaris” longicauda Hall: Van Straelen and Schmitz, 1934. Ceratiocaris longicaudus Hall: Hall, 1863. Ceratiocaris punctata Hall: Bigsby, 1878. Ceratiocaris punctatus Hall: Vogdes, 1889; Whitfield, 1880. Ceratiocaris ? punctatus Hall: Hall, 1863. Ceratiocaris (Aristozoe) punctatus Hall: Hall, 1876. Ceratiocaris whitfieldi Clarke: Clarke, 1902. Chagrinichnites: Frey, Curran, and Pemberton, 1984; Stukel, 1986; Weidner and Feldmann, 1983. Chagrinichnites brooksi Feldmann et al.: Feldmann et al„ 1978. Chagrinichnites osgoodi Hannibal and Feldmann: Han- nibal and Feldmann, 1983; Weidner and Feldmann, 1983. Dithyrocaris: Stumm and Chilman, 1969. Dithyrocaris (= Mesothvra): Rolfe and Denison, 1966. Dunsopterus: Waterston, 1968. Dunsopterus stevensoni (R. Etheridge, Jr.): Waterston, 1968. Dunsopterus (?) wrightiamus (Dawson): Waterston, 1968. Echinocaris (- Ceratiocaris [sic]) sp.: Monroe and Teller, 1899. Echinocaris sp.: Allan, 1935; Baird, 1978; Beecher, 1884; 50 HANNIBAL AND FELDMANN No. 42 Bolton, 1966; Buehler and Tesmer, 1963; Copeland, 1960a; Copeland and Bolton, 1985; Edmonds, Wil- liams, and Taylor, 1979; Feldmann, Boswell, and Rammer, 1986; Gekker, 1941; Goldring, 1971; Hlavin, 1976; Krestovnikov, 1961; McLaren, 1955; McLaren, 1963; Reimann, 1942; Rollins, Eldredge, and Linsley, 1972; Sartenaer, 1969; Sturgeon, Hlavin, and Kesling, 1964; Tschernyshev, 1928; Tschernyshev, 1938; Whit- field, 1892; Willard, 1932; Willard, Swartz, and Cleaves, 1939; Williams and Kindle, 1905. Echinocaris? sp.: Fisher, 1951; Willard, Swartz, and Cleaves, 1939. Echinocaris archae Tschernyshev: Krestovnikov, 1961. Echinocaris armata (Hall): Etheridge, Woodward, and Jones, 1889; Jones and Woodward, 1884. Echinocaris armatus (Hall): Packard, 1883. Echinocaris arschae Tschernyshev: Tschernyshev, 1928; Tschernyshev, 1938. Echinocaris asiatica Reed: Chhibber, 1934; Feldmann, Hannibal, and Babcock, 1986; La Touche, 1913; Pas- coe, 1959; Reed, 1908; Vogdes, 1925. Echinocaris asiaticus Reed: Feldmann, Hannibal, and Babcock, 1986. Echinocaris auricula Eller: Copeland, 1960a; Eller, 1935; Feldmann, Boswell, and Kammer, 1986; Hannibal and Feldmann, 1985. Echinocaris? beecheri Clarke (nomen nudum): Clarke, 1891; Clarke, 1898a; Dana, 1895. Echinocaris beecheri Copeland: Bolton, 1966; Copeland, 1960a. Echinocaris? brevicarinata Tschernyshev: Krestovnikov, 1961; Tschernyshev, 1938. Echinocaris castorensis Copeland: Bolton, 1966; Cope- land, 1960a; Hannibal and Feldmann, 1985. Echinocaris clarkei Beecher: Grabau and Shimer, 1910; Gurich, 1929. Echinocaris clarkii Beecher: Beecher, 1902; Caster, 1930; Chadwick, 1935. Echinocaris clarkii? Beecher: Castor, 1934. Echinocaris condylepis Hall and Clarke: Chadwick, 1 935; Clarke, 1892; Clarke and Ruedemann, 1903; Cope- land, 1960a; Eller, 1935; Eller, 1937; Gurich, 1929; Hall and Clarke, 1888; Miller, 1889; Tschernyshev, 1933; Vogdes, 1889; Vogdes, 1890. Echinocaris condylepsis Hall and Clarke: Vogdes, 1890; Vogdes, 1893. Echinocaris consanguina Eller: Bolton, 1966; Copeland, 1960a; Eller, 1935. Echinocaris crosbyensis Eller: Eller, 1937; Jux, 1960. Echinocaris longicauda (Hall): Beecher, 1884; Miller, 1889. Echinocaris ? longicauda Hall: Clarke, 1904; Clarke and Luther, 1904. Echinocaris longicauda Hall (Ceratiocaris longicauda)-. Miller, 1889. Echinocaris multinodosa Whitfield: Beecher, 1884; Clarke, 1892; Clarke, 1904; Clarke and Ruedemann, 1903; Feldmannand McKenzie, 1981; Feldmann et al., 1978; Grabau and Shimer, 1910; Gurich, 1929; Hall and Clarke, 1888; Hoover, 1960; Miller, 1889; Rolfe, 1962c; Rolfe, 1981; Sturgeon, Hlavin, and Kesling, 1964; Tschernyshev, 1938; Vogdes, 1889; Vogdes, 1890; Vogdes, 1893; Vogdes, 1917; Weidner and Feldmann, 1983; Whitfeld, 1880; Whitfield, 1890; Whitfield, 1899; Willard, Swartz and Cleaves, 1939. Echinocaris cf. E. multinodosa Whitfield: Willard, 1932. Echinocaris sp. cf. multinodosa Whitfield: Sturgeon, Hlavin and Kesling, 1964. Echinocaris multinodosus Whitfield: Packard, 1882; Pack- ard, 1883. Echinocaris multispinosis Feldmann and McKenzie, 1981: Feldmann and McKenzie, 1981. Echinocaris ohioensis Sturgeon, Hlavin, and Kesling: Feldmannand McKenzie, 1981; Feldmann etal., 1978; Sturgeon, Hlavin, and Kesling, 1964; Weidner and Feldmann, 1983. Echinocaris pulchra Sturgeon, Hlavin, and Kesling: Feldmannand McKenzie, 1981; Feldmann et al., 1978; Sturgeon, Hlavin and Kesling, 1964; Weidner and Feldmann, 1985. Echinocaris pulcra Sturgeon, Hlavin, and Kesling: Feld- mann and McKenzie, 1981. Echinocaris punctata (Hall): Beecher, 1884; Bernard, 1895; Bigsby, 1878; Brooks, 1957; Brown, 1956; Buehler and Tesmer, 1963; Chadwick, 1935; Chamberlin and Salisbury, 1905; Chamberlin and Salisbury, 1909; Clarke, 1892; Clarke, 1902; Clarke, 1905; Clarke, 1921; Clarke and Ruedemann, 1903; Cleland, 1903; Cleland, 1911; Copeland and Bolton, 1985; Dacque, 1921; Dana, 1895; Dunkle, 1965; Eller, 1935; Etheridge, Woodward, and Jones, 1889; Feldmann and Hannibal, 1985a; Fenton and Fenton, 1958; Goldring, 1929; Gra- bau, 1921; Grabau and Shimer, 1910; Grasso, 1981; Gurich, 1929; Hall and Clarke, 1888; Jones and Woodward, 1889; Jux, 1960; Kesling and Chilman, 1975; Krestovnikov, 1960; La Touche, 1913; Lesley, 1889-1890; Miller, 1889; Murphy, 1972; Olsson, 1912; Reimann, 1942; Rolfe, 1962b; Rolfe, 1969; Rolfe, 1981; Rollins, Eldredge, and Linsley, 1972; Schram, 1986; Schuchert, 1943; Shimer and Shrock, 1944; Smith, 1935; Steinmann and Doderlein, 1890; Stumm and Chilman, 1969; Tschernyshev, 1933; Vogdes, 1889; Vogdes, 1917; Willard, 1935; Willard, Swartz, and Cleaves, 1939; Woodward, 1943; Zell, 1985; Zittel, 1880-1885; Zittel, 1887; Zittel, 1900; Zittel, 1903; Zit- tel, 1913; Zittel, 1915; Zittel, 1924. Echinocaris cf. punctata (Hall): Prosser, 1898. Echinocaris punctatus (Hall): Packard, 1882; Packard, 1883; Scott, 1907; Vogdes, 1889. Echinocaris ( Ceratiocaris ) punctatus (Hall): Vogdes, 1893. Echinocaris (= Ceratiocaris ? = Ceratiocaris - Ceratiocaris 1987 ECHINOCA RIS: AN ANNOTATED BIBLIOGRAPHY 51 ( Aristozoe )) punctatus (Hall) (= armatus ): Vogdes, 1890. Echinocaris pustulosa Whitfield: Beecher, 1884; Clarke, 1892; Clarke, 1904; Clarke and Ruedemann, 1903; Grabau and Shimer, 1910; Gtirich, 1929; Hall and Clarke, 1888; Hoover, 1960; Jones and Woodward, 1884; Miller, 1889; Vogdes, 1889; Vogdes, 1890; Vogdes, 1893; Vogdes, 1917; Whitfield, 1880; Whit- field, 1890; Whitfield, 1899. Echinocaris pustulosus (Hall): Copeland and Bolton, 1985. Echinocaris pustulosus Whitfield: Packard, 1883. Echinocaris randalii Beecher: Tschernyshev, 1928; Kres- tovnikov, 1961 . Echinocaris randalii Beecher: Caster, 1930; Caster, 1934; Chadwick, 1935; Copeland, 1960a. Echinocaris randallii Beecher: Beecher, 1902; Gurich, 1929. Echinocaris randalliil Beecher: Rolfe, 1969. Echinocaris sloliensis Coomarasway: Goldring, 1971. Echinocaris sloliensis Partridge: Butler, 1980; Edmonds, Williams and Taylor, 1979; Partridge, 1902. Echinocaris sloliensis Partridge: Butler, 1980; Edmonds, Williams and Taylor, 1979; Partridge, 1902. Echinocaris sociales Beecher: Eller, 1935. Echinocaris socialis Beecher: Beecher, 1884; Beecher, 1902; Case, 1982; Caster, 1930; Caster, 1934; Chad- wick, 1935; Clarke, 1892; Clarke and Ruedemann, 1903; Dana, 1895; Easton, 1960; Eller, 1935; Eller, 1937, Grabau and Shimer, 1910; Gurich, 1929; Hall and Clarke, 1888; Jux, 1960; Lesley, 1889-1890; Miller, 1889; Moore, Lalicker, and Fischer, 1952; Muller, 1963; Partridge, 1902; Roger, 1953; Rolfe 1962b; Rolfe, 1969; Schram, 1986; Shimer and Shrock, 1944; Spinar, 1960; Tschernyshev, 1933; Twenhofel and Shrock, 1935; Vogdes, 1889; Vogdes, 1890; Vogdes, 1893; Vogdes, 1917. Echinocaris socialis? Beecher: Castor, 1934. Echinocaris cf. socialis Beecher: Tesmer, 1975. Echinocaris sublaevis Whitfield: Clarke, 1892; Clarke, 1904; Clarke and Ruedemann, 1903; Eller, 1935; Feld- mann and McKenzie, 1981; Grabau and Shimer, 1910; Gurich, 1929; Hall and Clarke, 1888; Jones and Woodward, 1884; Miller, 1889; Shimer and Shrock, 1944; Vogdes, 1889; Vogdes, 1893; Whitfield, 1899; Willard, Swartz, and Cleaves, 1939. Echinocaris cf. E. sublaev is Whitfield: Willard, 1932. Echinocaris sublevis Whitfield: Beecher 1884; Feldmann et al., 1978; Gurich, 1929; Hoover, 1960; Miller, 1889; Packard, 1882; Packard, 1883; Rolfe, 1962c; Sturgeon, Hlavin and Kesling, 1964; Vogdes, 1889; Vogdes, 1890; Vogdes, 1893; Vogdes, 1917; Weidner and Feldmann, 1983; Whitfield, 1880; Whitfield, 1890. Echinocaris tudrensis Tschernyshev: Gekker, 1941; Gek- ker, 1983; Krestovnikov, 1961; Tschernyshev, 1933; Tschernyshev, 1941. Echinocaris turgida Eller: Eller, 1935; Eller, 1937. Echinocaris uralensis Tschernyshev: Krestovnikov, 1961; Tschernyshev, 1938. Echinocaris whidbornei Jones and Woodward: Butler, 1980; Cleevely 1983; Copeland, 1960a; Edmonds, Wil- liams, and Taylor, 1979; Eller, 1937; Goldring, 1971; Gurich, 1929; Herries, 1896; Jones, 1898a; Jones, 1898b; Jones, 1900; Jones and Woodward, 1889; Jones and Woodward, 1899; Morris, 1980; Partridge, 1902; Vogdes, 1917; Whidborne, 1896a; Whidborne, 1896b. Echinocaris whitfieldi Clarke: Clarke, 1885a; Clarke, 1885b; Clarke, 1891; Clarke, 1892; Clarke, 1898a; Clarke, 1904; Dana, 1895; Hall and Clarke, 1888; Les- ley, 1889-1890; Miller, 1889; Schuchert, 1943; Vogdes, 1889; Vogdes, 1890; Vogdes, 1893. Echinocaris (= Eleutherocaris) whitfieldi Clarke: Rolfe, 1969. Echinocaris wrightana (Dawson): Miller, 1889. Echinocaris wrightiana (Dawson): Jones and Woodward, 1884; O’Connell, 1916; Vogdes, 1893. “ Echinocaris ” wrightiana ( Dawson ){- Stylonurus ? wrigh- tianus ): Copeland, 1960a. "Echinocaris" ( Stylonurus ?) wrightianus (Dawson): Van Straelen and Schmitz, 1934. Echinocarys: Barrois, 1891. Eleutherocaris: Clarke, 1902; Copeland, 1960b; Grabau and Shimer, 1910; Gurich, 1929; Krestovnikov, 1960; Roger, 1953; Rolfe, 1969; Rolfe and Edwards, 1979; Zittel, 1900; Zittel, 1913. Eleutherocaris whitfieldi Clarke: Chadwick, 1935; Clarke, 1904; Clarke and Luther, 1904; Clarke and Ruedemann, 1903; Dana, 1895; Grabau and Shimer, 1910; Gurich, 1929; Schuchert, 1943. Eleutherocaris ( Echinocaris ) whitfieldi Clarke: Van Straelen and Schmitz, 1934. Elymocaris: Oehlert, 1889. Equisetides wrightiana Dawson: Clarke, 1885b; Dawson, 1881b; Hall, 1884; Jones and Woodward, 1884; Vogdes, 1893; Wright, 1884. Equisetites wrightiana Dawson: Clarke, 1885; Dawson, 1881a; Dawson, 1882. Equisetites wrightianus Dawson: Dawson, 1881a. Hebertocaris: Stumm and Chilman, 1969. Montecaris: Rolfe, 1981. Montecaris? sp.: Dzik, 1980. Ohiocaris wycoffi Rolfe: Rolfe, 1962c. Orozoe: Oehlert, 1889. Palaeopalaemon newberryi Whitfield: Whitfield, 1880; Whitfield, 1892. Pephricaris: Gurich, 1929. Pephricaris horripilata Clarke: Beecher, 1902; Clarke, 1898b; Gurich, 1929. Physophycos bilobatus Lesquereux: Lesquereux, 1891. 52 HANNIBAL AND FELDMANN No. 42 Pseudodontichthys whitei Skeels: Case, 1982; Skeels, 1962. Ptychocaris: Copeland, 1960b; Hall and Clarke, 1888. Ptychocaris novaki Copeland: Copeland, 1960b. Rhinocaris: Stumm and Chilman, 1969. Rhinocaris columbina Clarke, 1888: Clarke, 1902; Zell, 1985. Silesicaris: Copeland, 1960b. Stylonurus: Beecher, 1900; Clarke and Ruedemann, 1912; Hall, 1884. Stylonurus (?) wrightiana (Dawson): O’Connell, 1916. Stylonurus wrightianus (Dawson): Chadwick, 1935; Clarke, 1904; Vogdes, 1917. Stylonurus ? wrightianus (Dawson): Chadwick, 1935; Clarke, 1904; Clarke and Luther, 1904; Clarke and Ruedemann, 1903; Kjellesvig-Waering, 1961; Kjellesvig- Waering, 1966; Waterston, 1968. Stylonurus'? wrightianus (Dawson) (-Echinocaris wright- ianus)-. Clarke and Ruedemann, 1912. Stylonurus (?) ( Echinocaris ?) wrightianus (Dawson): Beecher, 1900; Clarke, 1892; Hall and Clarke, 1888. Stylonurus (= Equise tides - ? Echinocaris) wrightianus (Dawson): Vogdes, 1890. Stylonurus (Equisitides) wrightianus (Dawson): Vogdes, 1893. Stylonurus (= Equisetides - Echinocaris - ? Echinocaris) wrightianus (Dawson): Vogdes, 1890. Trigonocarys lebescontei Barrois: Barrois, 1891. Tropidocaris: Cleevely, 1983; Oehlert, 1889. A BLUEGRASS NEW TO OHIO: POA SALTUENSIS FERN. & WIEG. JAMES K. BISSELL Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Pasture bluegrass ( Poa saltuensis Fern. & Wieg.), a northern species which ranges from western Ontario and Minnesota to eastern Quebec, southward through north- ern and western New England to northern Maryland and West Virginia (Fernald 1950), has been discovered re- cently in northeastern Ohio in Chardon Township, Geauga County. P. saltuensis is a species of open woods, thickets, and recent clearings (Fernald 1950). The Ohio specimen (Bissell 1984: 23) was collected on 23 May 1984 just below a bluff rim of an eroding, south-facing clayey slump along the west valley wall of Big Creek just north of the corporate limits of Chardon. A duplicate of the mu- seum specimen has been annotated by A. A. Reznicek at the University of Michigan, Ann Arbor. A diligent search for additional occurrences of P. sal- tuensis was made on similar slumping valley walls of the Chagrin River to the west of Chardon and slumping val- ley walls of the Grand River to the east of Chardon during 1985 and 1986. However, no additional Ohio records for P. saltuensis were discovered. P. saltuensis is fairly common in adjacent Pennsylva- nia, occurring across the northern portion of the state ex- cept the extreme northwestern counties, and extending south through the state on the Appalachian Mountains (Wherry et al. 1979). This newly discovered Ohio locality of P. saltuensis is about 90 miles west of its westernmost occurrence in Pennsylvania (Wherry et al. 1979). Within Michigan P. saltuensis is confined to the Upper Peninsula and central and northern Lower Peninsula with a disjunct county occurrence in Allegan County on Lake Michigan in the southern lower Peninsula (Voss 1972). P. saltuensis is uncommon at the single Ohio locality, growing in partial shade on the upper slump face just below a forested bluff rim. Several breaks occur on the forest along the bluff rim. The most dominant tree along the bluff rim is white oak (Quercus alba L.), with a few scattered beech (Fagus grandifolia Ehrh.). Shrubs asso- ciated with P. saltuensis include buffaloberry (shepherdia candensis L.), leatherwood ( Dirca palustris L.), witchha- zel ( Hamamelis virginiana L.) and maple-leaf viburnum ( Viburnum acerifolium L.). The most common herbace- ous associates of P. saltuensis at the site are longstalked sedge ( Carex pedunculata Willd.) and wreath goldenrod (Solidago caesia L.). P. saltuensis is similar to another species of bluegrass (Poa languida Hitchc.) (Gleason and Cronquist 1963) which does occur in northeastern Ohio (Braun 1967). The length of the ligules in P. saltuensis runs between .6 and 1.5 mm (3 mm), while ligule length in P. languida ranges between (2. 1 mm) 2.4 and 4 mm (Voss 1972). The margins of the lemmas in both species are glabrous (Voss 1972). P. saltuensis has acute or acuminate, distinctly 5-nerved lemmas, while P. languida has obtuse, obscurely nerved lemmas (Gleason and Cronquist 1963). Another species of bluegrass (Poa cuspidata ), common throughout the eastern half of Ohio (Braun 1967), grows in habitats sim- ilar to the P. saltuensis habitat in Chardon. Many of the northeastern Ohio P. cuspidata specimens in the CLM herbarium contain the label comment: “in partial shade on eroding upper slopes just below bluff rim.” The keels of the lemmas of P. cuspidata are pubescent in contrast to the glabrous keels of P. languida and P. saltuense (Fer- nald 1950). References Braun, E. L. 1967. The Monocotyledonaea (of Ohio): Cat-tails to orchids. With Gramineae by Clara G. Weishaupt. Colum- bus: Ohio State University Press. Fernald, M. L. 1950. Grays manual of botany. 8th ed. New York: American Book Company. Gleason, H. A., and A. Cronquist. 1963. Manual of vascular plants of northeastern United States and adjacent Canada. New York: D. Van Nostrand. Voss, E. G. 1972. Michigan flora part E Gymnosperms and Monocots. Cranbrook Institute of Science Bulletin 55. Bloomfield Hills, Mich. Wherry, E. T., J. M. Fogg, Jr., and H. A. Wahl. 1979. Atlas of the flora of Pennsylvania. Philadelphia: The Morris Arbore- tum. Univ. Penn. Kirtlandia, No. 42 © by The Cleveland Museum of Natural History Membership and Subscription Membership in Kirtlandia Society is open to any member of the Cleveland Mu- seum of Natural History. Current individual membership is $30.00 ($10.00 for stu- dents); participating members, $100.00; institutional, $30.00; sustaining, $250.00. Applications and fees should be sent to the Treasurer, Kirtlandia Society, Cleve- land Museum of Natural History, Cleveland, Ohio 44106. Membership includes subscription to Kirtlandia. There is a non-U. S. postal sur- charge of $5.00. All subscriptions outside North America will be sent by air. Pay- ment must be made in U.S. dollars. Payments not drawn on a U.S. bank will be subject to bank collection charges. Change of Address '. Please send to Museum six weeks prior to move; enclose present mailing label with change of address. Claims for missing issues: Claims cannot be honored beyond four months after mailing date. Duplicate copies cannot be sent to replace issues not delivered be- cause of failure to notify Museum of change of address. Back issues: Individual issues of Kirtlandia are available from the Museum Pub- lications Office. A complete list is available upon request. CLEVELAND, OHIO NUMBER 43 KIRTLANDIA •NATURAL HISTORY* PALEONTOLOGY A New Species of Sauropod Dinosaur, Haplocanthosaurus delfsi sp. nov., from the Upper Jurassic Morrison Fm. of Colorado John S. McIntosh and Michael E. Williams 3 Isolated Tetrapod Remains from the Carboniferous of West Virginia Stephen J. Godfrey 27 JULY 1988 KIRTLANDIA The Scientific Publication of the Cleveland Museum of Natural History David S. Brose, Editor Joseph T. Hannibal, Assistant Editor Brief History and Purpose Kirtlandia is a publication of the Cleveland Museum of Natural History, an organization , incorporated in 1922 whose predecessor, the Ark, was founded in Cleveland by Jared Potter Kirtland in 1837. Kirtlandia is a continuation of the earlier series. Scientific Publications volumes 1 to 7 (1928-1937), and new series volumes 2 to 4 (1962-1965), and the Survey of Ohio Fluted Points. Supported by the Kirtlandia Society of the Cleveland Museum of Natural History, Kirtlandia is devoted to the publication of scientific papers in the various fields of inquiry within the Museum’s sphere of interest: Cultural and Physical Anthropology; Archaeology; Botany; Geology; Paleobotany; Invertebrate and Vertebrate Paleontology; Systematic Ecology; and Invertebrate and Vertebrate Zoology. Issues will vary from single monographs to collections of short papers, review articles, and brief research notes. Kirtlandia is distributed by The Kent State University Press, Kent, Ohio 44242. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that the base fee of $00.50 per copy, plus .25 per page is paid directly to CCC, 27 Congress Street, Salem, MA 09170. The fee code for users of the Transactional Reporting Service is 0075-6245/88 $00.50+.25. Kirtlandia is abstracted in Biological Abstracts and indexed in Bibliography and Index of Geology and Zoological Record. Kirtlandia No. 43 © 1988 by The Cleveland Museum of Natural History KIRTLANDIA THE CLEVELAND MUSEUM OF NATURAL HISTORY Cleveland, Ohio July 1988 Number 43 PALEONTOLOGY A New Species of Sauropod Dinosaur, Haplocanthosaurus delfsi sp. nov., from the Upper Jurassic Morrison Fm. of Colorado John S. McIntosh and Michael E. Williams Isolated Tetrapod Remains from the Carboniferous of West Virginia Stephen J. Godfrey $M\lTHS6ivJ^ AUG 2 6 1988 3'fiRARIES 3 27 ISSN: 0075-6245 KIRTLANDIA EDITOR DAVID S. BROSE Cleveland Museum of Natural History ASSOCIATE EDITORS MARY BAUM Research Librarian Cleveland Museum of Natural History JAMES K. BISSELL Curator of Botany Cleveland Museum of Natural History BRUCE LATIMER Curator of Physical Anthropology Cleveland Museum of Natural History SONJA TERAGUCHI Manager of Collections Cleveland Museum of Natural History JOSEPH T. HANNIBAL Curator of Invertebrate Paleontology Cleveland Museum of Natural History MARTIN ROSENBERG Editorial Assistant Case Western Reserve University EDITORIAL ADVISORY BOARD RODNEY FELDMANN Professor of Geology Kent State University MICHAEL C. HANSEN Geologist Ohio Geological Survey RICHARD MEINDL Associate Professor of Anthropology Kent State University G. MICHAEL PRATT Associate Professor of Anthropology Heidelberg University DAVID H. STANSBERY Director, Museum of Zoology Ohio State University FREDERICK H. UTECH Curator of Botany Carnegie Museum of Natural History ED VOSS Curator of the Herbarium University of Michigan ANDREW M. WHITE Professor of Biology John Carroll University KIRTLANDIA THE CLEVELAND MUSEUM OF NATURAL HISTORY Cleveland, Ohio July 1988 Number 43:3-26 A NEW SPECIES OF SAUROPOD DINOSAUR, HAPLOCANTHOSAURUS DELFSI SR NOV., FROM THE UPPER JURASSIC MORRISON FM. OF COLORADO john s. McIntosh Wesleyan University Middletown, Connecticut 06457 and MICHAEL E. WILLIAMS Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Abstract A skeleton mounted in the Cleveland Museum of Natural History is assigned to the rare genus of sauropod dinosaur, Haplocanthosaurus , as a new species, H. delfsi. The seventy-foot-long skeleton is the largest known specimen of Haplocanthosaurus and the only mounted one. It was collected between the years 1954 and 1957 by a museum party from the clays of the lower part of the Upper Jurassic Morrison Formation on Oil Creek in Fremont County, Colorado, north of Canon City. 4 mcintosh and williams No. 43 Abstract A skeleton mounted in the Cleveland Museum of Natural History is assigned to the rare genus of sauropod dinosaur, Haplocanthosaurus , as a new species, H. delfsi. The seventy-foot-long skeleton is the largest known specimen of Haplocanthosaurus and the only mounted one. It was collected between the years 1954 and 1957 by a museum party from the clays of the lower part of the Upper Jurassic Morrison Formation on Oil Creek in Fremont County, Colorado, north of Canon City. Introduction In 1901 the newly appointed curator of vertebrate pale- ontology at the Carnegie Museum in Pittsburgh, John Bell Hatcher, decided to reopen the Marsh-Felch quarry in Garden Park, Colorado, about ten miles north of Canon City. This quarry had produced the type specimens of a number of species of dinosaurs including Diplodocus lon- gus, Allosaurus fragilis, Ceratosaurus nasicornis, Stego- saurus stenops , and “Morosaurus” agilis. W. B. Utterback was sent to the field and during the summer’s operations, collected two medium-sized, partial skeletons of a new genus of sauropod dinosaur. The more complete of these, CM 572, was subsequently described in a brief paper by Hatcher (1903a) as Haplocanthus priscus. Four months later he altered the name to Haplocanthosaurus (1903b: 100) because Haplocanthus was “essentially preoc- cupied” by a genus of fish named by Agassiz (1844). In November 1903, Hatcher published a monograph on Hap- locanthosaurus in which the second skeleton was also described as a new species, H. utterbacki (CM 879). Since the publication of Hatcher’s memoir, very little has been added to our knowledge of this animal. In a review of the sauropods, von Huene (1929) modified several of Hatcher’s conclusions and suggested that the number of dorsal verte- brae should be reduced from 14 to 12. In a paper redescrib- ing “ Morosaurus ” agilis Marsh, 1889, Gilmore (1907) conjectured that its type specimen, a cranial fragment and cervicals 1 to 3, might belong to Haplocanthosaurus, and this remains a distinct possibility as discussed below. Finally, in 1981, McIntosh referred some limb bones from the type locality to this form as probable. No new material belonging to this genus was reported until summer of 1954, when a field party from the Cleve- land Museum of Natural History began excavating a large sauropod skeleton on the east bank of Oil Creek (Four Mile Creek, Nine Mile Creek) less than a mile south of the historic Marsh-Felch Quarry 1 (Fig.l). History of the CMNH Quarry (The “Sawropod Lode”) Early in the summer of 1954, William E. Scheele, then director of the CMNH, dispatched a small field party to several western states with the expressed aim of finding a mountable dinosaur for exhibit. The search initially cen- tered around Vernal, Utah, where the crew was “shown the (★), and the nearby Marsh-Felch (■) and the Cope-Lucas Quarries (•). Cooper Mtr. Quadrangle, Colorado. ropes” by Leroy “Pop” Kay, long-time curator at the Carnegie Museum. Kay also provided leads to several promising sites. Although several specimens were located, none proved workable due to problems of obtaining permis- sion or the difficulty of the required excavation. The field party then split in two, one group headed by director Scheele travelling northward into Wyoming and the second moving eastward into Colorado. While camped at the Colorado National Monument near Grand Junction, a member of the Cleveland crew overheard a conversation between students in a geological field party who had found and collected a partial bone in a stream exposure near Canon City. Dr. Carl Sanderson, a geologist at Louisiana State University and leader of that University’s summer field camp, kindly provided a map of the site. The bone fragmnent was given to the CMNH crew, and later proved to be the posterior end of the fourth cervical vertebra. The first day’s excavation demonstrated the presence of a considerable amount of bone, and it soon became apparent that the major portion of a skeleton was present. The quarry site (Figs. 1, 2A) is located just off the road, on the east bank of Four Mile Creek in the NW 1/4, NW 1/4, Sec. 34, T. 17S, R.70W, Fremont County, Colorado (Cooper Mt. Quad.). Since this was during a major uranium boom, it was deemed necessary to protect the site from potential pros- pecting damage by filing a claim with the Fremont County recorder. A letter from the County Clerk and Recorder’s office, dated October 21, 1954, states “The name of this lode is Sawropod, and the reception number is 289758.” Kirtlandia, No. 43, July 1988 © by the Cleveland Museum of Natural History 1988 NEW SPECIES OF SAUROPOD DINOSAUR 5 Fig. 2A The CMNH Quarry from the road leading north from Canon City. Early summer, 1954. The bones were recovered from the lower part of the Morrison Formation, in a light gray clay layer, bound both above and below by massive sandstone ledges (Figs. 2B, 2C, 3). The upper sand was five- to eight-feet thick and was overlain by some eight feet of sand and gravel. Quarrying operations during the first field season consisted of tunnel- ing under the overlying sandstone which was shored up with vertical timbers. By the end of the season, the practical limits of this technique had been reached and the remaining excavation, in 1955 and 1957, (Fig. 4) was accomplished by removing the overburden with a bulldozer and blasting away the sandstone ledge. Each of the three years’ excavations was led by Edwin Delfs, then an undergraduate biology major at Yale Univer- sity. Other field crew members were high school and college students Wesley Williams, William West, and Richard Jones (1954), Wesley Williams and Joseph Hurley (1955), and Ralph Wrisley (1957). A local rancher, Joe Rhode, of Garden Park, Canon City, did the bulldozer work and provided invaluable advice on construction matters and in handling the large and often very heavy blocks. Members of the “permanent” field crew were supplemented from time to time by a number of Cleveland Museum staff members and volunteers, most notably Mary Flahive, Elizabeth Olmstead, David Roberts, William Scheele, Dan Snow, and Ellen Walters. The skeleton was lying on its left side and was largely articulated. As so often happens with sauropod skeletons, the neck was drawn sharply backward and the skull had snapped off and was not found. The anterior dorsal verte- brae and all but the first four cervicals had been eroded away by the stream. The posterior two-thirds of the tail was also missing. The ribs and girdle bones of the under (left) side were in place, but only the ilium remains of the right side. The only limb bones preserved were the left femur and the heads of two bones restored as the left tibia and fibula. Other vertebrate remains found with the skeleton include a number of turtle fragments, an isolated theropod tooth, and 6 McIntosh and williams No. 43 Fig. 2B. Installing timbers, early summer, 1954. Left to right, Wesley Williams, Richard Jones, William West. the skull and partial skeleton of a new goniopholid croco- dile, Eutretauranosuchus delfsi (Mook 1967). The members of the field party are to be congratulated for their skill in exhuming the huge but often very fragile bones, particularly the vertebrae. Some of the packages were opened in the winter of 1954-55 (Delfs 1961) and the skeleton was determined to be Haplocanthosaurus. Further preparation and restoration of the missing parts prior to mounting the skeleton were accomplished at the American Museum of Natural History in New York under the super- vision of veteran preparator George Whitaker. The speci- men was first displayed in Kirtland Hall at the Cleveland Museum of Natural History in 1961, with the body resting on the ground. It was remounted in the upright pose shown in Figure 5 two years later. Although brief notices have appeared in the popular press (Anonymous 1959: Delfs 1961; Pearl 1975; Piel 1963; Anonymous 1966; Williams 1982), the skeleton has not been described. Abbreviations AMNH American Museum of Natural History CM Carnegie Museum of Natural History CMNH Cleveland Museum of Natural History OUM Oxford University Museum USNM National Museum of Natural History YPM Yale Peabody Museum Systematic Paleontology Class Reptilia Subclass Archosauria Order Saurischia Seeley Suborder Sauropoda Marsh Family Cetiosauridae Seeley Genus Haplocanthosaurus (Hatcher 1903b) Diagnosis — Cervical vertebrae of only moderate length, pleurocoels prominent but simple, neural spines of poste- rior cervicals and anterior dorsals not divided. Dorsal 1988 NEW SPECIES OF SAUROPOD DINOSAUR 7 Fig. 2C. Quarry site, late summer, 1954, showing the removal of the first of 3 large plaster jackets, the 2500 lb. “Iceberg." Ed Delfs (by jacket), Joe Rhode (wearing hat) and unidentified truck driver. centra relatively small, all containing prominent pleuro- coels; dorsal arches high with diapophyses extending up- ward at 45 degrees as well as outward, spines short and broad. Sacrum consisting of the usual dorso-sacral, three primary sacrals and a caudo-sacral with centra coosified, as are the sacral ribs, to form a yoke; small pleurocentral cavities in at least some of the centra; spines relatively low with a tendency toward coalescence of all, but particularly numbers one to three. Caudal centra amphicoelous, short, and without pleurocentral cavities; chevron facets very prominent and give the underside of the centrum a sculp- tured appearance; caudal spines slender and curved back- ward in the anterior region; they are of moderate height anteriorly and low further back. Distal end of scapula thin and broadly splayed; proximal plate relatively smaller than in most sauropods. Sternal plates large and subquadrangu- lar. Proximal part of ischium relatively small, shaft straight, distal end slighty broadened but not thickened. Femur neither overly slender nor stout, the latero-medial diameter of the shaft significantly exceeding the antero-posterior one as in Brachiosaurus. Other possible significant generic characters are discussed below in connection with speci- mens probably, but not certainly, belonging to Haplocan- thosaurus. Haplocanthosaurus priscus (Hatcher 1903b) Haplocanthus priscus Hatcher 1903a Haplocanthosaurus priscus (Hatcher 1903b) Haplocanthosaurus utterbacki (Hatcher 1903c) Holotype - CM 572 (Hatcher 1903a) Horizon and locality — Upper Jurassic Morrison Fm., Marsh-Felch Quarry, Garden Park, Colorado Amended specific diagnosis — Medium sized Haplocantho- saurus with comparatively slender femur and pelvic girdle. Distal ends of ischia narrowed, rotated inward and fused to their opposite in the midline. Type species Haplocanthus priscus Hatcher 1903a Haplocanthosaurus delfsi sp. nov. Holotype - CMNH 10380 Horizon and locality — Upper Jurassic Morrison Fm. East bank of Four Mile Creek NW 1/4, NW 1/4, Sec. 34, T. 17S., R.70W. , Fremont County, Colorado (Cooper Mt. Quad.). 8 McIntosh and williams No. 43 Fig. 3. Dick Jones exposing a section of rib, early summer 1954. Material — CMNH 10380 cervicals 1-4, nine posterior dorsals with ribs of the left side, five sacrals, caudals 1-14, several chevrons, shaft and distal end of left scapula, fragmentary coracoid?, right sternal plate, proximal end of left radius, proximal end of left ulna, both ilia, left pubis, left ischium, left femur. Diagnosis — Very large Haplocanthosaurus with most measurements 35-50% greater than that of the fully adult holotype of H. priscus (CM 572). Girdle bones and femur more robust than in H. priscus. Pubis in particular, much heavier distally. Distal end of ischium broader, not rotated inward and not fused to its mate. V-shaped, anterolaterally projecting laminae present on neural spines of the middle dorsals, and greater developement of median laminae on the posterior dorsal spines than in H. priscus. Designation of the specimen As mentioned in the introduction, Hatcher changed the generic name from Haplocanthus to Haplocanthosaurus because Haplocanthus was “essentially preoccupied” by a genus of fish named by Agassiz in 1844. Agassiz used the spelling Haplocanthus. A similar situation exists with the stegosaur originally named Kentrosaurus by Hennig (1915) but altered to Kentrurosaurus (Hennig 1916) because Lambe had used Centrosaurus for a genus of ceratopsian. Following Romer (1966), Galton (1982) has recently re- turned to the use of Kentrosaurus , arguing the “two generic names cannot be considered homonyms even if it is only one letter which is different (International Code of Zoological Nomenclature 1961 Article 56a).” This discussion has been challenged by Anderson (1982) who argued that Hennig made the correct decision based on the rules in 1915. In the present case, Haplocanthosaurus is clearly the valid form, because the senior synonym Haplocanthus must be consid- ered a forgotton name (nomen oblitum) due to lack of use (ICZN, 1961 art. 23b). The most recent edition of the Code (ICZN, 1985), however, does not use the term nomen oblitum and requires that the current usage be maintained while the matter is referred to the commission for a ruling (art. 23b). Skull and mandible No part of the skull, mandible, hyoid bones, proatlas, or any portion of the dentition has been found with any of the three partial skeletons of Haplocanthosaurus. The skull on the mounted skeleton is modelled. However, see comments concerning Morosaurus agilis below. Vertebrae Cervicals. The number of cervical vertebrae in Haplo- canthosaurus is not known. In CM 572 only the last two were preserved. In CM 879 there were seven complete cervicals, an additional centrum, an arch, and a fragmen- tary arch. The atlas and axis are not represented. From this material Hatcher surmised that the total number of cervical vertebrae was fifteen, the same as in Diplodocus, the only sauropod in which the number was known with certainty at the time. Since then, a number of sauropod genera have been found to have fewer than fifteen cervicals, and indeed, von Huene (1929) revised Hatcher’s figure down to thir- teen. The Cleveland skeleton does not help resolve this question, but by providing the atlas and axis (Fig. 6), which are missing in the Pittsburgh material, it does add signifi- cantly to our knowledge of the animal. Although most of the neural arch of the atlas is missing, what remains shows that it was firmly coalesced to the intercentrum as in all other adult sauropods. The odontoid is firmly united with the anterior end of the axis and extends straight forward, tapering and ending in a blunted point. The intercentrum is relatively longer than in Apatosaurus , but as in the atlas of the latter, well-developed articular facets for a single headed cervical rib occur on the posterior part of the lateral face. The ribs mounted on the atlas are both plaster, however, the cervical rib attached to the left side of the axis is real, and is clearly single-headed, lacking a dorsal tubercular process. The capitular end is consider- ably expanded and conforms well to the articular surface of the atlas. The distal end of the rib is rounded and restored in plaster. It seems likely that it was displaced from the atlas and that the axis bore a double-headed rib as is typical of sauropods. 10 ft NEW SPECIES OE SAUROPOD DINOSAUR Fig. 4. Quarry map redrawn from field notes. The extent of the three summers’ activities are indicated by dashed lines. Jacket no. 2, containing portions of rib and apparently the neural spine of the dorsal vertebrae in PK 13, was lost in an early flash flood. (Not to scale.) 10 McIntosh and williams No. 43 Fig. 5. Haplocanthosaurus delfsi, mounted skeleton on display in Kirtland Hall, Cleveland Museum of Natural History. The axis (Fig. 6) is complete except for the lower front part of the arch, which would have included the articulation with the neurapophysis of the atlas. A moderate depression occupies a large part of the lateral face of the centrum rather than a true pleurocoel. The anterior part of this depression is deepened into a small round cavity that penetrates the medium septum. Below and slightly anterior to this opening is a well developed, dorso-ventrally elongate, articular facet, the parapophysis. The transverse process arises on the rear of the lower part of the arch, however, the distal portions of both processes are restored. As in other sauropod axes, the arch and spine sweep back and upward from front to rear and the postzygapophyses lie directly beneath the high point of the spine and directly above the rearmost part of the centrum. In cervical three the lateral depression occupies about half of the side of the centrum. The vertebra is strongly opisthocoelous and the prezygapophyses extend beyond the anterior ball of the centrum. The diapophyses lie further forward than in the axis, and both the neural spine and the postzygapophyses are higher. As the distal end of the cervical rib is not preserved, it cannot be ascertained whether or not it extends beyond the back end of the centrum. The fourth cervical is considerably larger than the third and is also strongly opisthocoelous. The edges of the lateral depression are sharper but do not yet define a true pleuro- coel. The zygapophyses resemble those of cervical three, but the postzygapophyses are placed higher. The undivided spine has assumed a triangular shape. Again nothing can be said concerning the extent of the cervical rib. The fragment identified by Hatcher as a postzygapophy- sis of cervical three in CM 879 is too fragmentary to be of any comparative value. Unfortunately the centrum of cer- vical 4? of CM 879 has been damaged since Hatcher’s day and only the rear half remains. Its lateral depression is similar to that of CMNH 10380. The arch is much dis- torted, but differs in no important way from that of the Cleveland specimen, so while Hatcher’s identification of it as the fourth may be correct this cannot be verified with certainty. A more significant comparison can be made between the cervicals of CMNH 10380 and those of a specimen found directly beneath a left femur associated with USNM 4275, discussed below as probably belonging to this genus. The specimen USNM 5384, consists of the braincase, proatlas, atlas, axis, and cervical 3. It was described briefly by Marsh (1889) as Morosaurus agilis sp. nov., and in detail 1988 NEW SPECIES OF SAUROPOD DINOSAUR 11 Fig. 6. First four cervical vertebrae of Haplocanthosaurus delfsi, CMNH 10380. Most of the arch of the atlas and the cervical ribs are restored. Left lateral view. by Gilmore (1907), who suggested that it might belong to Haplocanthosaurus since it came from the same quarry as the holotype. A direct comparison of this specimen with the Cleveland skeleton appears to bear out Gilmore’s conjec- ture. The Washington specimen has been crushed as are most of the specimens from the Marsh-Felch quarry, but when one takes into account 1) Gilmore’s observation that, “the spinous process of cervical 3 has been crushed forward somewhat from its normal position” and that 2) “the transverse process, postzygapophysial lamina and postzygapophyses are wanting,” the two specimens appear quite similar. This is not only apparent in the general shapes of the first three cervicals but more specifically in the simple characters of their pleurocentral depressions. These are not crossed by laminae nor puctuated with subsidiary cavities as in many of the more advanced sauropods, and their margins lack the sharp lips characteristic of a true pleurocoel. It will not be our intent here to discuss the skull fragment; suffice it to say that it differs significantly from that of Camarasaurus ( Morosaurus ) and the other Morrison sau- ropods. The presence of complete right and left halves of the proatlas in position in USNM 5384 is of great interest, since this element is rarely reported in the sauropods. Although it seems likely that Morosaurus agilis is syn- onymous with Haplocanthosaurus priscus, we refrain from invoking the law of priority at this time, because of the distorted and incomplete nature of USNM 5384, and because the first three cervicals are not among the most diagnostic elements of the sauropod skeleton. Dorsal Vertebrae. Nine dorsal vertebrae are preserved in CMNH 10380. The first six of these appear to have been in articulation, but the seventh had been displaced upward and rotated onto its side (see additional comments concerning the numerical sequence, below). The eighth dorsal was displaced further back and lay near the ilium. The ninth had been carried still further back and was found beyond the fourteenth (last preserved) caudal vertebra. As it turned out, collecting ended in 1954 with the seventh dorsal in the series (Fig. 4). Collecting resumed in 1955 with the eighth dorsal and the sacrum, but the widely displaced ninth dorsal was not recovered until 1957. Field notes indicate that the latter was displaced at least twenty feet from the articulated series. Thus, the question arises as to whether additional dorsals between the seventh and ninth may have also been displaced and lost. For a number of reasons, we believe that this is not the case, and that the series which is mounted as the nine dorsals anterior to the sacrum is correctly restored. Unlike most of the other Morrison sauropods, in Haplocan- thosaurus the variation from one posterior dorsal to another is minor. Using the two Pittsburgh skeletons, Hatcher determined the number of dorsal vertebrae in Haplocanthosaurus as fourteen. In CM 572 there were three articulated vertebral segments, the first of which he took to be the last two cervicals and first dorsal. The second segment consisting of nine dorsals, he took to be the last nine, while the third segment represented the sacrum and tail. In CM 879 the presacral vertebrae lay in approximate order, but only a series of six posterior dorsals were actually articulated. He 12 McIntosh and williams No. 43 TABLE 1 Measurements of Cervical Vertebrae, CMNH 10380 No. Centrum or Intercentrwn Height Spread length anterior posterior overall pre- post- Diapo- breadth height breadth height height zygap. zygap. physes 1 65 66 65 63 85 e200 — 2 192 70 98 80 97 277 194 113 3 237 106 102 55 72 282 186 203 172 4 300 123 123 72 71 305 197 232 197 All measurements here and elsewhere in this paper are in millimeters, e - estimated took thirteen of these to be dorsals two through fourteen; dorsal one represented in CM 572 being absent. Some years later von Huene (1929) used Hatcher’s figures to reinterpret these assignments and concluded that Hatcher’s dorsal one was really the penultimate cervical and that his dorsal two was actually the last cervical, thus reducing the dorsal count to twelve. Briefly stated, the reasons for the change were the occurrence of the parapophysis below the pleurocentral cavity and the fact that Camarasaurus had recently been shown to have twelve dorsals. The other two sauropod genera with known dorsal counts at the time of von Huene’s paper were Diplodocus and Apatosaurus both of which had ten. Subsequently the Chinese genus Euhelopus was shown to possess fourteen, and the prosauropod Plateosaurus had fifteen, including the last one which is taken into the sauropod sacrum as a dorso-sacral. Thus it would not be surprising if a relatively primitive sauropod like Haplocan- thosaurus did indeed possess fourteen dorsals. Theoretical arguments aside, the empirical evidence also favors Hatch- er’s interpretation. As students of the sauropods know well, the transition from cervicals to dorsals in this group is quite gradual, and the determination of where the change occurs is based on the ribs. The transition from the last cervical rib to the first thoracic rib is abrupt. In Diplodocus the parapophysis drops significantly on the arch of the fifth dorsal from the position seen in subsequent dorsals. In the fourth and third it is lower on the centrum in front of the pleurocentral cavity, and in dorsals one and two it lies beneath this cavity. Dorsal one resembles the last cervical more closely than it does a typical dorsal, but it bears a true thoracic rib. There is another difference: in mature sauro- pods the cervical ribs are invariably coalesced to their vertebrae whereas the thoracic ribs are not. In CM 879 the neural arches of most of the presacrals and sacrals are not fused to their centra, and the cervical ribs are not fused to their vertebrae, thus indicating an immature individual. In CM 572, however, all the arches are firmly fused to their centra, no trace of the line of fusion remaining. Likewise the cervical ribs on both sides of the two anteriormost vertebrae of segment one are firmly fused to their vertebrae, while the ribs of the third are free (and have been lost). Furthermore, the parapophyses lie beneath the pleurocen- tral cavity at the base of the centrum, which is to be expected in the first dorsal. It would appear to us that the only possibility for reducing the number of dorsals from fourteen would be if Hatcher’s dorsal two of CM 879 corresponds to his dorsal one of CM 572. This would reduce the number to thirteen, and might conceivably be attributed to individual variation, although not even von Huene suggested this possibility. While admitting that future discoveries might bear out this possibility, we believe that based on current evidence, it is more prudent to stick with Hatcher’s original determination of fourteen dorsals. What then are the locations in the series of the CMNH 10380 dorsals? It seems reasonably certain that the first six and probably the first seven occurred in serial order. The first package taken out, PK 1, was a 750 lb. block said to contain “assorted bones’’ (Fig. 5). A later, typed version of the field notes lists PK 1 as containing “vertebra in scrappy condition.’’ Package no. 2, which was immediately behind PK 1, contained the neural arch of another dorsal, labelled PK 13 on the quarry diagram. Unfortunately, PK 2, plastic jacket and all, was lost in a flood early in the excavation (Fig. 7). In the mounted skeleton, two dorsals were placed anterior to the one missing the neural arch (Fig. 8), suggesting that the first two were in PK 1 . Evidence tending to confirm this is the considerable difference in the anterposterior expanse of the neural spines of the first two vertebrae. This is in contrast with the remaining dorsals, whose neural spines are essentially uniform in this respect, as are the rearmost dorsals of both CM 879 and CM 572. The two displaced dorsals have been mounted as the two between the sacrum and the former group and they seem to fit well in this series (Fig. 9). A hyposphene-hypantrum articulation is present in the dorsals of both Carnegie specimens from the sixth to the last. These articulations are likewise present in the dorsals of CMNH 10380, save for the first and third where they are restored in plaster. The hyposphene on the first dorsal may be partially real, but the presence of a hypantrum on the second demonstrates its presence nonetheless. It would seem, therefore, that the series is correctly restored, and represents dorsals six through fourteen (dorsals four through twelve in the mount). One apparent difference, however, is the position of the capitular articulation on the sixth dorsal, which in CMNH 10380 is at the same level as in succeeding vertebrae, but 1988 NEW SPECIES OF SAUROPOD DINOSAUR 13 Fig. 7. One of several flash floods that plagued the excavation. Joseph Hurley is seen standing atop a large plaster jacket. Summer 1955. distinctly lower and greatly enlarged on the sixth dorsal of both Carnegie Museum specimens. Detailed descriptions of the dorsals are unnecessary since Hatcher’s suffice, but several points should be noted con- cerning the new specimen. First, all the neural arches are firmly fused to their respective centra. The centra are relatively small, as in CM 572 and CM 879, with large, sharpley defined pleurocentral cavities high up on the centra. The anterior and middle centra are strongly opistho- coelous, the posterior ones less so (Fig. 8). Indeed, the anterior ball of the last three are only slightly convex to nearly flat. The centra of dorsals 13 and 14 are noticeably shorter than the others, but number 13 has clearly received some antero-posterior flattening (crushing) so this feature may be exaggerated. The neural arches and spines are, for the most part, in accord with the Carnegie Museum specimens. The arch is high, the spine short and broad and the diapophyses are directed strongly upward as well as outward. The arches of dorsals 6 and 7, the first bones collected and those just in front of the segment of the column eroded away by the stream, have required some restoration (e.g. the left diapophysis of dorsal 6 has been restored in plaster), but they are for the most part intact. As noted above, the arch and spine of dorsal eight were lost in a flood and have been completely restored. The remaining six dorsals are essentially complete, and with the exception of the minor crushing already noted in dorsal 13, they are little distorted. Finally, J. F. Bonaparte (personal commu- nication) has pointed out some differences in the develop- ment of the laminae of the dorsal arches; in particular, the presence of V-shaped antero-laterally projecting laminae on the neural spines of the middle dorsals and a greater 14 mcintosh and williams No. 43 Fig. 8. Dorsal vertebrae of CMNH 10380 (seen from right side) in the process of being mounted at the AMNH in New York. A cervical rib was later added to the second (modeled) vertebrae in the series, leaving 12 dorsal vertebrae in the mount. The last 9 dorsals are completely or in part real and are herein regarded as the posteriormost 9 dorsals of the animal (i.e. 6-14). See text for additional comments. Photo courtesy of the American Museum of Natural History. development of the median laminae on the posterior dorsal spines of CMNH 10380. He believes these characters may indicate a new genus “related to Haplocanthosaurus but different. We believe these differences indicate at most, a separate species. Sacrum. As far as can be determined the sacrum is in almost complete agreement with that of CM 572. The dorso-sacral and caudo-sacral centra are totally fused to those of the three primary sacrals. Furthermore, the arches of all five vertebrae are firmly fused to one another and to the sacral ribs. As in CM 572 the short to moderately long spines of sacrals one through three are firmly united throughout, and those of sacrals four and five are united to midheight. In CM 572 spines four and five are free, but this is probably an individual character because in the even younger CM 879 all five spines are united. As mounted it is not possible to determine the existence of pleurocentral cavities. As in CM 572, the spine of the dorso-sacral is strongly inclined backward and that of the caudosacral strongly inclined foreward. Caudal Vertebrae. The first fourteen caudal vertebrae were found articulated with the sacrum. Other than appear- ing slightly heavier than those of CM 572 and CM 879 (a probable age character), they resemble the latter closely (Fig. 10). The centra are short, amphicoelous, lack side cavities, and bear prominent chevron facets, particularly on the posterior end of the centrum. In the first two vertebrae the diapophyses (more correctly caudal ribs) bear a hint of the wing-like processes seen on the anterior caudals of the diplodocids (in contrast to those of Camarasaurus and Brachiosaurus). Further back they are simple and extend horizontally, diminishing in size from number one to fourteen (the last preserved). The prominence of the diapo- physes on the fourteenth caudal suggests that they were 1988 NEW SPECIES OF SAUROPOD DINOSAUR 15 TABLE 2 Measurements of the Dorsal Vertebrae No. length anterior breadth height Centrum posterior breadth height overall height pre- zygap. Height post- zygap. 6 e255 158 172 193 238 795 — — 1 e245 172 193 190 236 875 465 — 8 230 182 215 213 248 — — — 9 e255 191 233 214 242 840 — — 10 e248 211 263 237 274 913 — — 11 e243 245 263 271 288 935 — — 12 201 319 254 370 313 915 — — 13 144 334 298 349 337 1040 — — 14 170 338 278 349 293 1010 — 590 Note: These measurements were taken with some difficulty from the mounted skeleton. Minor errors may be present. present on several succeeding vertebrae. This is in contrast to the situation in CM 572, where the diapophyses have all but disappeared on caudal thirteen. The first caudal bears a typical spine which is directed first upward and then curves around toward the back, resembling that of CM 879 closely. In CM 572, caudal one was displaced and badly crushed. When restored by the artist Sydney Prentiss, this curvature was not indicated. Behind caudal number one the spines are relatively slender, straight, and directed upward and back- ward. Fig. 9. Last 6 dorsal vertebrae of CMNH 10380 as seen from the left side. Photo courtesy of the American Museum of Natural History. Ribs. In the mounted skeleton the left ribs attached to dorsals six through thirteen are at least partially bone. The other ribs have been restored in plaster. Field diagrams show that the left ribs nine, ten, and eleven were articulated TABLE 3 Measurements of the Sacrum of CMNH 10380 Length of five centra 1090 Length of centrum of Sacral 1 202 Anterior breadth of same 301 Length of centrum of Sacral 2 234 Height of Sacral 1 1020 Height of Sacral 2 1010 Height of Sacral 3 1000 Height of Sacral 4 910 Height of Sacral 5 854 Distance between distal ends of ilia externally 640 with their respective vertebrae and that twelve was little displaced. These ribs are all characteristic of sauropod dinosaurs and do not exhibit any unusual features. Chevrons. Several incomplete anterior chevrons were found with the tail, but were displaced above it. They appear to be relatively long and straight with the bifurcate heads relatively short. The heads resemble those of Cama- rasaurus, in having no bridge of bone above the haemal canal, and are in contrast to the situation in Diplodocus and Apatosaurus , where a transverse “crus” exists. Sternum One of the most important discoveries with the Cleveland specimen was that of a complete sternal plate (Fig. 11), previously unknown in Haplocanthosaurus. Considerable controversy has existed in the literature concerning the positioning of these sternal plates. It was partly resolved —with the discovery in Camarasaurus USNM 13786 and Alamosaurus USNM 15560— that the heavy pointed end of the plate was directed forward (Gilmore 1946). This being the case the element preserved with CMNH 10380 would be the right sternal plate. It is roughly quadrangular with the length greatly exceeding the breadth. It is very gently convex downward. The lateral and medial margins are parallel and very gently curved to bulge outward from the 16 mcintosh and williams No. 43 TABLE 4 Measurements of Caudal Vertebrae of CMNH 10380 No. length anterior breadth Centrum posterior height breadth height overall height pre- zygap. Height post- zygap. Spread Diapo- physes 1 176 346 370 270 370 865 — — — 2 155 294 355 226 330 770 — — — 3 154 266 340 231 258 728 515 460 — 4 157 289 300 231 280 680 446 430 377 5 165 257 273 220 269 620 424 390 389 6 166 257 272 191 271 578 423 386 392 7 155 217 251 203 244 547 372 372 379 8 155 216 232 210 237 530 369 356 362 9 163 313 239 213 217 491 336 349 *306 10 163 227 204 194 201 454 348 328 — 11 162 207 192 205 193 415 305 278 278 12 163 192 182 188 187 430 279 310 223 13 174 187 175 197 164 359 279 271 e247 14 167 178 147 175 161 323 261 257 231 * = distorted midline. The medial margin is smooth, but the lateral one is irregularly rugose for the attachment of the cartilaginous sternal ribs. The posterior margin also has indentations and trends slightly forward from the outside in. The thickened anterior margin is least straight of all. It curves forward to form a blunted point and then recedes to the median margin. The plate is relatively much longer than in any other sauropod except for Alamosaurus and Opisthocoelicaudia. If and how the sternal plate articulates with the coracoid in sauropods has not been determined. The sternal bone of CMNH 10380 more nearly resembles that of Camarasaurus lentus USNM 13786 than any other, but sauropod sternal plates show great variation within the same genus, so this similarity may be fortuitous. This bone does not resemble the sternal of Camarasaurus grandis YPM 1901. After many attempts to place the sternal plates reasonably, the AMNH preparator George Whitaker was unable to settle on an arrangement that he regarded as satisfactory, and in the end, they were not placed on the mounted skeleton. TABLE 5 Measurements of the Sternum of CMNH 10380 Length of sternal plate 755 Breadth of same 358 Pectoral girdle Scapula. The shaft and distal end of the left scapula are preserved in CMNH 10380. This is to be compared to the scapula of CM 879 and another left scapula-coracoid, CM 33995. The latter was miscatalogued CM 94, the number of the cotype of Diplodocus carnegii. It does not belong to that animal but rather to Haplocanthosaurus. Both scapulae of CM 94 are preserved. All the bones found in Quarry D, Sheep Creek, Wyoming (from which it derives) are repre- sented on an excellent quarry map, and all are accounted for. In addition, the color and preservation of CM 33995 is distinctive of the Garden Park quarry and quite unlike those of bones from the Sheep Creek quarries. Finally, the quarry map of the Garden Park quarry (Hatcher 1903c) shows a complete scapula-coracoid among the bones of CM 572. It is shaded to show that it belonged to a different genus, but it is not present elsewhere in the collections in Pittsburgh. It appears to us that this bone was somehow misplaced and very likely belongs to CM 33995. If so, Hatcher’s failure to refer it to the H. priscus skeleton is puzzling. Perhaps he considered it a bit too large. The incomplete Cleveland specimen shows the same widely flared distal end exhibited by CM 879 (Fig. 12). The general appearance of the bone as restored would appear a bit more massive than the latter, but this may be a matter of restoration or possibly of age. The proximal (anterior) plate is restored in plaster to resemble the “normal” sauropod scapula. Haplocanthosaurus , like Cetiosaurus, differs from the norm in having a much less developed anterior plate. Coracoid? An incomplete girdle bone found near the sternal plate may be an incomplete coracoid. If so, it is of little diagnostic value as none of the characteristic features —the coracoid foramen, the thickened glenoid surface, etc.— is preserved. It is also possible that the element is the posterior portion of the left sternal plate. One of the borders does have a scalloped outline similar to the lateral margin of the right sternal plate. On the other hand, it would appear to be slightly wider and perhaps a little thicker than the latter. Fore Limb Radius and Ulna. All that remains of the fore limb are the proximal ends of the left ulna and radius. They were 1988 NEW SPECIES OF SAUROPOD DINOSAUR 17 Fig. 10. Caudal vertebrae of CMNH 10380 seen from the right side. The transition from the first 14 (real) caudals to the modeled ones behind is marked by a dimunition in size of the vertical support rods beneath each vertebra. mistaken for parts of the left tibia and fibula and were incorporated into these restored elements in the mounted skeleton. These pieces are the only fore limb bones known to belong to Haplocanthosaurus , but they have lost much of their value because of their fragmentary condition. The proximal end of the radius differs from that of most other 18 McIntosh and williams No. 43 Fig. 11. Right sternal plate of Haplocanthosaurus delfsi. A) Ventral View B) Dorsal View. sauropods in that its greatest diameter (345mm) greatly exceeds its perpendicular one. Viewed from the front, the medially-directed pointed process is prominent. Little can be said about the imperfectly preserved head of the ulna except that the anterior trough, which cradles the end of the radius, is shallow. As in other sauropods, the proximal end is V-shaped when viewed from above, but the corners of both legs of the V are incomplete, particularly that of the medial branch. This apparently led the restorers to mistake it for the cnemial crest of the tibia. Pelvis The bones of the pelvis and the femur of CMNH 10380 show more differences from those of the type of H. priscus, CM 572, than do any other of the preserved parts. Ilia. Both ilia are present and the left one is virtually complete (Fig. 13). It is more robust than the illium of CM 572. In front of the pubic peduncle, the anterior tip of the blade is deflected outward in a manner reminiscent of the Upper Cretaceous titanosaurids, where this feature is greatly accentuated. No such deflection is observed in either the left or right ilium of CM 572. Both of the latter have been crushed but in different directions, and this may account for the apparent difference between the ilia of the two animals. The crushing is most severe in the left bone of CM 572 causing the lobe in front of the pubic peduncle to appear much longer than it was. Pubis. The left pubis is nearly complete, and while resembling that of CM 572 it appears somewhat more robust, particularly at the distal extremity. A swelling on the upper portion of the anterior margin represents the attach- ment for the ambiens muscle, but there is no indication of the hook-like process that develops here in the diplodocids. As preserved the pubic foramen was open posteriorly, leaving an incomplete border for articulation with the ischium. This is almost surely the result of incomplete preservation rather than immaturity. In all known sauropod genera (certainly in all adult specimens), the foramen is ringed with bone. It has been restored to indicate the closed condition. The shaft of the pubis is stocky and the distal end is moderately expanded to meet its opposite member in the midline, although the two bones are not coalesced. TABLE 6 Measurements of the Pectoral Girdle CMNH 10380 CM 879 CM 33995 Length of scapula el 290 800 1063 Breadth, proximal -- 396 440 Breadth of shaft — 137 154 Breadth, distal end 497 372 410 Length of coracoid ? — 295 320 Breadth, greatest — 350 360 Ischium. The left ischium is complete and, except for the distal end, it resembles that of CM 572. The pubic articular margin is arced forward, more so than in other sauropod genera. The process for articulation with the ilium is typically long. The shaft is slender but only a little twisted. It broadens slightly at the distal end while not becoming any thicker. In CM 572 the broadening is less and the shaft is more twisted so that the two ischia meet at the distal end edge to edge as shown in Figure 14. In fact, in CM 572, the shafts of the two ischia are firmly coosified at their distal ends and for a considerable distance forward. An identical arrangement is found in USNM 4275 from the same quarry as the type and clearly referrable to Haplocanthosaurus. As both the ischia in CMNH 10380 and CM 572 were found articulated there can be no question that, indeed, each did belong to its assigned skeleton. Careful examination of the distal end of the left ischium of the Cleveland specimen shows that it was not coosified with that of the missing right one. Various misalignments of both the pubes and ischia of this specimen cause the distal ends of the ischia to meet in a sharp “V” rather than the shallow trough described by Hatcher (see additional comments in the final section describing the mounted skeleton). As in Camarasaurus and Brachiosaurus , none of the three specimens shows any thickening of the distal end in the perpendicular direction so typical of the diplodocids, particularly Apatosaurus (Fig. 15). There is in the collection of the National Museum of Natural History an articulated sauropod hind limb of the left side with part of the foot and the associated right femur and both ischia, USNM 4275. Only the ischia bear the cata- logue number USNM 4275, but the other elements clearly 1988 NEW SPECIES OF SAUROPOD DINOSAUR 19 Fig. 12. Left scapula-coracoid of CMNH 10380. Only the shaft and upper end of the scapula are bone. belonged to the same individual and will henceforth be referred to by the same number. The specimen was found in the East End of the Marsh-Felch quarry whence came the types of Haplocanthosaurus priscus and Diplodocus lon- gus. Marsh arbitrarily assigned the ischia to the D. longus and figured them as such (Marsh 1896). Recognizing that these ischia clearly had nothing to do with Diplodocus , Gilmore (1907) referred them, “to some large species of the Morosauridae,” and refigured them as Morosaurus. They are coalesced for half their length as in CM 572 and resemble it very closely. Although bearing some resem- blance to those of Camarasaurus ( Morosaurus ) these bones certainly belong to Haplocanthosaurus. Hind Limb The hind limb in CMNH 10380 is represented by the left femur only (Fig. 16). Femur. At first glance, the femur of CMNH 10380 would appear to show marked differences from that of the holotype of H. priscus, CM 572. Aside from its smaller size (see Table 8), the Pittsburgh specimen appears to be more slender and to have a more circular shaft, although it has suffered from latero-medial crushing; whereas the Cleve- land specimen has, if anything, undergone some antero- posterior flattening. TABLE 7 Measurements of the Pelvis CMNH 10380 CM 572 USNM 4275 Length of ilium 1315 827 — Height of pubic peduncle 955 512 — Breadth of acetabulum 370 — — Length of pubis 1100 693 — Breadth, proximal end 504 — — Breadth, distal end 453 — — Length of ischium, distal end to pubic 1000 790 800 articulation Length of ischium, distal end to iliac 995 — — articulation Least breadth of shaft 144 — — Breadth, distal end 236 85 — Length of articular surface with pubis e440 — — The femur of CMNH 10380 is nearly complete although an area on the medial side of the lower half of the shaft has been restored in plaster. It is a straight heavy bone expanded at both ends, in which the latero-medial diameter consider- ably exceeds the antero-posterior one. The cross-section of the shaft is thus a flat oval, most closely resembling Brachiosaurus in this respect. The fourth trochanter is 20 McIntosh and williams No. 43 Fig. 13. Left illium and sacrum of CMNH 10380. located at mid-length on the postero-medial border of the shaft. The slight swelling on the upper part of the lateral- border of the shaft, which Galton (1981) has interpreted as the final vestige of the lesser trochanter, is very weak, unlike that of Brachiosaurus. The proximal end is broad with the head directed nearly at right angles to the shaft, and not rising much above the remnant of the greater trochanter. The horizontal projection of the head into the acetabulum is greater than in most sauropods, but that may be due to a slight deformation of the Cleveland specimen and is not observed in CM 572. The distal end is moderately expanded with the tibial condyle exceeding the fibular one in extent. The comparative measurements given below may be some- what misleading due to the crushing of CM 572 noted above. The hind limb associated with the ischia USNM 4275 has peculiarities that preclude its reference to any of the other five Morrison sauropods: Diplodocus , Barosaurus , Apato- saurus, Camarasaurus , and Brachiosaurus. The femora are in general agreement with that of CMNH 10380 although the comparative measurements indicate a slightly more robust form. All four limb bones have been crushed; the left femur antero-posteriorly to flatten it, the right one latero- medially; although they are identical in length and were lying across one another. The left tibia and fibula were in position at the lower end of the femur, and the ischia were also in position. The worth of USNM 4275, important in itself in providing another specimen of this rare animal, is further enhanced by providing information about the lower segment of the leg described below. Tibia, Fibula, and Pes. No trace of the lower segment of the hind limb or foot have been preserved with CMNH 10380. As noted previously the proximal ends of the left radius and ulna have been incorporated into the restored left tibia and fibula of the mounted skeleton. In USNM 4275 the left tibia, fibula, and astragalus are complete but somewhat crushed. Their most notable feature is their massiveness. The short stocky tibia is expanded at both ends, but the cnemial crest is only moderately devel- oped. The distal end is greatly expanded, more so than in any other Morrison sauropod except Brachiosaurus. The latero-medial diameter of the shaft greatly exceeds the antero-posterior one, but this may have been exaggerated by crushing. The astragalus is in place at the end of the tibia, but the matrix between the bones has not been removed. As presently visible, the astragalus presents no noteworthy features. The left fibula is also stocky. Its proximal end is unexpanded, the antero-posterior diameter exceeding the 1988 NEW SPECIES OF SAUROPOD DINOSAUR 21 Fig. 14. A) Pelvis of the holotype of Haplocanthosaurus priscus (CM 572) as seen from behind. Note the fusion of the distal ends of the ischia in the midline. (From Hatcher, 1903c). Fig. 14. B) Left ishium of Haplocanthosaurus delfsi CMNH 10380, Length = 1000 mm. latero-medial one as usual. The lower end is slightly expanded, the two breadths being subequal. The tibial articular scar on the medial face at the proximal end is Fig. 15. Lateral Views of Left Ischia. A) Cetiosaurus, B) Haplocantho- saurus priscus, C) H. delfsi, D) Titanosaurus , E) Brachiosaurus , F) Camarasaurus , G) Diplodocus, H) Dicraeosaurus , I) Apatosaurus. Fig. 16. Left femur of CMNH 10380, posterior view, with Gil Stucker, Martin Cassidy and George Whitaker. Courtesy of American Museum of Natural History. prominent. It terminates below at a bulge on the anterior margin of the bone. In addition to the stockiness of the tibia and fibula the most striking feature of the USNM 4275 limb 22 mcintosh and williams No. 43 is the small tibio-femoral length ratio, .52. Among the other five North American Jurassic sauropods Brachiosaurus is closest to this with .59, but the others are over .60. It may be added that an almost identical right tibia, fibula, and astragalus (CM 2043) were collected in the Marsh-Felch Quarry at the same time, but some little distance to the east of CM 572 and CM 879. These have been assigned to Haplocanthosaurus by McIntosh (1981). The tibia has a length of ,627m., which is too short for CM 572, although it may belong to CM 879. TABLE 8 Measurements of the Hind Limb CMNH 10380 CM 572 USNM 4275 left right Length of femur 1745 1275 1110 1105 Breadth, proximal end 555 353 350 370 Breadth of shaft — 207 200 195 Breadth, distal end 545 309 315 245 Thickness, tibial condyle 360 — 215 275 Thickness, fibular condyle 245 — 185 255 Least circumference of shaft 755 518 491 502 Ratio, circumference: length .43 .41 .44 .45 Length of tibia 580 Breadth, proximal end 240 Breadth of shaft 130 Breadth, distal end 256 Least circumference of shaft 321 Ratio, circumference: length .55 Length of fibula 610 Breadth, proximal end — Breadth of shaft 79 Breadth, distal end 142 Least circumference of shaft 218 Ratio, circumference: length .35 Ratio of lengths, tibia: femur .52 Ratio of lengths, fibula: femur .55 Note: caution should be observed in using these measure- ments as most of the bones have suffered distortions. The lengths are probably little affected however. The Species of Haplocanthosaurus Hatcher referred his two partial skeletons to separate species based solely on the state of coosification of the sacral spines. H. priscus CM 572 is about 5% larger than H. utterbacki CM 879 and as noted previously has all its vertebral arches firmly coosified to their respective centra, even the suture line being obliterated. In H. utterbacki almost all the arches are separate from their respective centra; the scapula and coracoid are likewise not coalesced. These characters indicate that CM 879 is a younger animal than CM 572. However, in CM 572 the spines of the first three sacrals only (i.e. , the dorso-sacral and primary sacrals one and two) are coosified for their entire lengths, those of the last two sacrals (primary sacral three plus the caudo- sacral) remaining free. In the younger animal, on the other hand, the spines of sacrals one to four are fused from top to bottom and even that of number five is fused to the spine in front of it at the top and bottom. Mook (1917) studied the problem of speciation in sauropods. In his terminology the fusion problem is one of “acceleration” in CM 879. In some unpublished notes he accepted Hatcher’s criterion for separating the species and noted further a second supposed difference, namely, the lack of the hyposphene in dorsals eight through eleven of CM 572 and the presence in CM 879. In their present state of preservation and restoration, dorsals eight through eleven of CM 572 do indeed appear to lack a hyposphene and have been so drawn by Sydney Prentiss, but one must remember that six through fourteen of CM 572 were found tightly articulated, and the process of disarticulation, particularly of the hyposphene- hypantrum, was most difficult. After examining these specimens, we are convinced that these vertebrae originally possessed a normal hyposphene. Hatcher himself stated, “All the dorsals in this region, nos. 6-14 exhibit the hyposphene-hypantrum articulation” (Hatcher 1903c). The separation of the species must therefore stand or fall on Hatcher’s original criterion regarding the union of the sacral spines. Riggs (1903) made a detailed study of the sacra of a number of individuals of Apatosaurus (Bronto- saurus) and Morosaurus (= Camarasaurus) in an attempt to show that the use of the number of centra fused together in the sacrum, whether three, four, or five, was an age character rather than a generic one as employed by Marsh and others. He also discussed the union of the spines. It is now generally accepted that he made his point at least at the generic level. A great deal more study is needed before distinctive criteria can be stated for the separation of species in the dinosaurs in general and the Sauropoda in particular. This must await the full preparation and study of large population samples from quarries like those at Dinosaur National Monument, the Cleveland-Lloyd Quarry, and Como Bluff Quarry 13. There are very few such quarries, and the problems relating to preparation and study are vast. Clearly it will be some time before the necessary criteria are available. As to the problem at hand, we believe that the variations observed in the coalescence of the various ele- ments of the sacrum of different individuals showing no other characters worthy of specific differentiation are inad- equate grounds for separating H. priscus and H. utterbacki. The fact that the two skeletons were found only a few feet from one another in the same quarry at the same level in the same stratum and are so similar in all other characters adds weight to this conclusion. The question remains as to whether CMNH 10380 belongs to H. priscus in light of the apparent differences in the pelvis and femur discussed above. There is also the question of the great disparity in size between the two adult animals CM 572 and CMNH 10380. There are a number of bones in the National Museum from the younger, Marsh- 1988 NEW SPECIES OF SAUROPOD DINOSAUR 23 Felch Quarry at Garden Park, Colorado, which probably belong to different individuals of Haplocanthosaurus (evi- dence to be presented elsewhere). All of these indicate a species relatively small by sauropod standards, none of them approaching in any way the size of the gigantic Cleveland skeleton. As pointed out by Mook (1917), size may indeed be significant in determining species. There are also the special characters of the laminae in the dorsal arches noted earlier. We conclude that these differences, together with the overall differences in size and robustness as well as the unfused ischia, indicate that CMNH 10380, from an older horizon, represents a distinct species, which we hereby designate H. delfsi after Dr. Edwin Delfs. Relationships of Haplocanthosaurus to other Sauropods Noting that the division of the Sauropoda by Marsh (1896) into six families would probably have to be reduced, Hatcher ( 1903c) indicated that he accepted at least the three families Atlantosauridae, Diplodocidae, and Morosauridae, and that it was to the latter family that Haplocanthosaurus belonged. He then compared that genus with three British sauropods and concluded that its relationship was closest to Cetiosaurus. The following year Riggs (1904) erected a new family for his recently discovered Brachiosaurus and assigned Haplocanthosaurus to the Brachiosauridae. The family characters he chose were 1) fore limbs longer than hind; 2) vertebral spines simple throughout; and 3) number of dorsals more than ten. Of course evidence regarding 1) was not available for Haplocanthosaurus. Some years later in reviewing the sauropods, von Huene (1929) referred both Haplocanthosaurus and Brachiosaurus to the Cetiosau- ridae, which had been established by Lydekker in 1888. Later classification schemes often place Cetiosaurus and Haplocanthosaurus together in one subfamily, the Cetio- saurinae, and Brachiosaurus in a second, the Brachiosau- rinae, of a single family sometimes called Cetiosauridae, sometimes Brachiosauridae. We agree that Haplocantho- saurus should be grouped with Cetiosaurus in the Cetio- sauridae, but believe Brachiosaurus and its allies have advanced sufficiently to be grouped in a separate family, the Brachiosauridae. To compare Haplocanthosaurus and Cetiosaurus di- rectly, we note that the skull is not known in either, at least not the complete one. The cranial fragment USNM 5384 which likely belongs to Haplocanthosaurus cannot be compared to Cetiosaurus because the corresponding frag- ment referred by von Huene (1906, 1932) to Cetiosaurus OUM 13596 does not belong to a sauropod. No teeth are known in Haplocanthosaurus , but the fragmentary tooth described by Phillips (1871) found with the Cetiosaurus oxoniensis skeleton is of the broad spatulate type. The heart-shaped teeth named by Owen (1840-45) Cardiodon rugulosus were considered to belong to Cetiosaurus by Marsh (1896)— a judgment augmented by the discovery of very similar teeth in the Middle Jurassic Argentine cetio- saurid Amygdalodon (Cabrera 1947). Similarities between Haplocanthosaurus and Cetiosaurus suggest that the former will also be found to have teeth of this sort. Comparison of the cervicals of Haplocanthosaurus with those of C. oxo- niensis is difficult because of the incomplete condition of the latter, but the partial skeleton of Cetiosaurus from Rutland, England, in the Leicester Museum, reported by Jones (1970) seems to agree closely with Haplocanthosaur- us in the simplicity of the lateral cavities and the undivided neural spines. Further preparation and a detailed study of the Leicester specimen is being pursued by John Martin and when completed will allow a more critical comparison. The dorsal vertebrae of the two genera are also similar, the one complete dorsal of C. oxoniensis exhibiting the same high arch, short spine, and having diapophyses directed outward and upward at 45° (Fig. 17 A, B). The sacrum is unknown Fig. 17. Anterior Views of Antero-median Dorsal Vertebrae. A) Cetio- saurus, B) Haplocanthosaurus, C) Brachiosaurus, D) Titanosaurus , E) Dicraeosaurus, F) Camarasaurus , G) Apatosaurus , H) Diplodocus. in Cetiosaurus, but the caudals are very similar to those of Haplocanthosaurus , one complete centrum from Bucking- hamshire (OUM 13876) showing the same greatly enlarged chevron facets. The scapula of Cetiosaurus has a reduced proximal plate and a broadly expanded distal end; however, the expansion is not as extreme as in that of its American counterpart. Both genera have relatively broad pubes and ischia with distal ends not greatly expanded (Fig. 15 A, B). The ischia of Cetiosaurus from Gloucestershire, England in the Stroud Museum (Reynolds 1937) compare very favor- ably with those of CM 572, but less so with that of CMNH 10380 where the distal end is more broadly expanded. The femora of both animals are moderately broad, show little evidence of a lesser trochanter, and the head of each rises little above the great trochanter. Finally, the tibio- femoralratio is only .6 in Cetiosaurus, which is signifi- cantly greater than the .52 of the referred specimen of Haplocanthosaurus. Other genera of the Cetiosauridae include Rhoetosaurus Longman, Amygdalodon Cabrera, Patagosaurus Bonaparte, and Shunosaurus Dong, Zhou, 24 mcintosh and williams No. 43 and Zhang. Preliminary notices relating to the latter two genera (Bonaparte 1979 and Dong, Zhou, and Zhang 1983) suggest more primitive forms but a full comparison must await more detailed descriptions of the vertebrae in those forms. Comparison of Haplocanthosaurus with other Morrison genera shows that it agrees with Brachiosaurus in the undivided presacral spines (Fig. 17 B, C), but differs in having shorter cervicals with less complicated pleurocentral cavities, dorsals with smaller centra, shorter spines in the anterior region, and differently directed diapophyses. The ilium is lower in Haplocanthosaurus. With the diplodocids Diplodocus , Apatosaurus , and Barosaurus the contrast is very great. These taxa have presacrals with deeply cleft spines in the shoulder region (Fig. 17 G, H), fewer dorsals with lower arches, much higher spines, particularly in the rear, and diapophyses directed horizontally. Their sacra have much higher spines with the dorso-sacral spine not coalesced with primary sacral 1. Their caudal vertebrae contain pleurocentral cavities anteriorly, have far less prom- inent chevron facets, and these vertebrae are much more numerous and more elongate in the median and posterior regions, resulting in a vastly longer tail. The two heads of the anterior chevrons bridge across in the diplodocids but not in Haplocanthosaurus, and in the former the median and posterior chevrons exhibit the peculiar fore and aft expansion which reaches its extreme in Diplodocus ; those in Haplocanthosaurus are simple throughout. The scapulae are quite different in the diplodocids and the distal end of the ischium is expanded greatly both in breadth and thick- ness (Fig. 15 G, H, I); finally the tibio-femoral ratio is well over .60. The comparison with Camarasaurus is somewhat closer but here again the divided presacral spines in the latter (Fig. 17 F) are in sharp contrast. Its cervicals have more complicated pleurocentral cavities than those of Haplocan- thosaurus; it has two fewer dorsals, 12 instead of 14, its dorsal centra are larger; the arches are somewhat lower and the diapophyses are directed outward; however the neural spines of the posterior dorsals are similar. The caudals in the two genera resemble one another, but the chevron facets are much less developed in Camarasaurus , and the poste- rior caudals are noticeably shorter in Haplocanthosaurus than in the latter. The chevrons themselves are quite similar, with no transverse bridge above the haemal canal and no diplodocoid fore and aft expansion. Both genera have scapulae with expanded distal ends, and both have ischia with distal ends little expanded (Fig. 15) and meeting one another edge to edge. The tibio-femoral ratio in Camara- suarus is a bit larger than .60; in Haplocanthosaurus it is apparently smaller. The Mounted Skeleton As restored, the skeleton is fully seventy feet long, a large sauropod (Williams 1982). All preserved elements of CMNH 10380 have been incorporated into the mounted skeleton except the sternal plate, the supposed coracoid, and a few chevrons (Figs. 18, 5). The missing portions have been modelled or cast from specimens in the American Museum of Natural History in New York. The skull and mandible have been modelled in generalized sauropod fashion. The neck has been restored to contain fourteen cervicals of which nos. one through four are bone whereas the other ten are modelled from the Pittsburgh skeletons. Twelve dorsals have been assigned to the thorax, the first three modelled from Pittsburgh specimens and the last nine real. Left thoracic ribs four through eleven as mounted are, at least in part, real. The others have been restored in Fig. 18. Diagram of the mounted skeleton. The shaded parts are bone. plaster. The sacrum and first fourteen caudals are real. These are followed by several modelled vertebrae and the tail is completed with casts of an articulated series of thiry-nine caudals of Camarasaurus AMNH 825. The first two chevrons are real, all others restored. The left scapula, excepting the proximal plate, is bone. The right one is modelled after it and both coracoids are restored. The fore limbs and feet have been completely restored from speci- mens in the American Museum, based largely on Apato- saurus. The complete left pelvis and right ilium as well as the sacrum are real, the right pubis and ischium having been modelled from their counterparts. The left femur is real and the right one modelled from it. The tibiae, fibulae and pedes have been modelled, the (real) heads of the left ulna and radius having been incorrectly incorporated into the left tibia and fibula. For the most part the pose of this skeleton follows the pattern of traditional sauropod mounts. The placing of the scapula was the subject of considerable debate among sauropod scholars until Gilmore’s (1925) paper on Cama- rasaurus CM 11338, where the scapula was found articu- lated and in position for the first time. In the present mount the scapula is placed somewhat higher on the rib cage, and its orientation more horizontal than in CM 11338. Unlike many sauropod mounts the restored fore and hind feet are constructed with the properly reduced carpus and tarsus and the single claw on digit I of the manus. The only possible criticism with the manus is that the five metacarpals are mounted side by side instead of in circular fashion where the first and fifth almost meet (Gilmore 1936). In addition, 1988 NEW SPECIES OF SAUROPOD DINOSAUR 25 Gilmore (1932) has shown that in Diplodocus, and probably the other sauropods as well, the tail proceeds straight out from the sacrum for a considerable distance before it begins to descend. As mounted in this specimen, it begins to descend almost at once and continues to do so sharply. One final criticism concerns the misorientation of the pubes and the ischia. As mounted, the medial borders of the pubes are directed caudally and do not meet ventrally in the midline. This causes them to appear much broader in side view than would normally be the case and also causes the ischia to meet in a sharp “V” rather than a shallow, nearly horizon- tal trough. In addition, the acetabular borders of the pubes are not in line with those of the ischia, resulting in an open acetabulum which appears much larger than it was in life. This skeleton is one of only two sauropod skeletons to be mounted in the United States in the last twenty-five years, the other being the Diplodocus in the Houston Museum of Natural Science. Acknowledgments We wish to thank Dr. Wann Langston and Dr. Philp Currie for having carefully read the manuscript and for numerous, valuable suggestions, and Dr. J. F. Bonaparte for valuable observations concerning the laminae of the dorsal vertebrae. We also wish to thank Bruce Frumker, staff photographer of the Cleveland Museum of Natural History and his assistant, Toni Hutton, for their fine photographic work. References Agassiz, L. 1844. Monographien des poissons fossiles du Vieux Gres Rouge ou systeme Devonien (Old Red Sandstone) des lies brittaniques et de Russie. Neuchatel xxvi + 171 pp. Anderson, H. J. 1982. Kentrosaurus Oder Kentrurosaurus. Eine bemerkung zur Nomenklatur. Geologica et Palaeontologica 15:147. Anonymous. 1959. Dinosaur remains main guest at museum dinner. Science News Letter, Sept. 21, 1959, 76:341. Anonymous. 1966. After 120 million years-on its feet again. Explorer 8:2, 11-13. Bonaparte, J. F. 1979. Dinosaurs: a Jurassic assemblage from Patagonia. Science 205:1377-1379. Cabrera, A. 1947. Un sauropodo nuevo del Jurasico de Patagonia. Notas Museo de La Plata 12 Paleontology 95, 1-17. Delfs, E. 1961. Operation dinosaur. Explorer 3:1,22-24. Dong, Z. M., S. W. Zhou, and Y. H. Zhang. 1983. The dinosaurian remains from Sichuan basin, China. Palaeontolo- gia Sinica Whole No. 162, C No. 23:136 pp. 44 pis. Galton, P. M. 1981. A juvenile stegosaurian dinosaur, “Astrodon” pusiilus from the upper Jurassic of Portugal. Journal of Verte- brate Paleontology 1:3-4, 245-256. Galton, P. M. 1982. The postcranial anatomy of the stegosaurarian dinosaur Kentrosaurus from the Upper Jurassic of Tanzania, East Africa. Geologica et Palaeontologica 15:139-160, 6 pis. Gilmore, C. W. 1907. The type of the Jurassic reptile Morosaurus agilis with a note on Camptosaurus. Proceedings of the United States National Museum XXXII: 151-165, pis. XII-XIII. Gilmore, C. W. 1925. A nearly complete articulated skeleton of Camarasaurus. Memoirs of Carnegie Museum X:347-384, pis. xii-xvii. Gilmore, C. W. 1932. On a newly mounted skeleton of Diplodocus in the United States National Museum. Proceedings of the United States National Museum 81:1-21, 6 pis. Gilmore, C. W. 1936. Osteology of Apatosaurus. Memoirs of the Carnegie Museum XI, 4:300 pp. 35 pis. Gilmore, C. W. 1946. Reptilian fauna of the North Horn Forma- tion of Central Utah. United States Geology Sur\>ey Professional Papers 210-C:29-53, pis. 3-14. Hatcher, J. B. 1903a. A new sauropod dinosaur from the Jurassic of Colorado. Proceedings of the Biological Society of Washing- ton XVI: 1-2. Hatcher, J. B. 1903b. A New Name for the dinosaur Haplocanthus Hatcher. Proceedings of the Biological Society of Washington XVI: 100. Hatcher, J. B. 1903c. Osteology of Haplocanthosaurus, with description of a new species. Memoirs of the Carnegie Museum II, 1:1-72, 5 pis. Hennig, E. 1915. Kentrosaurus aethiopicus, der Stegosauridae des Tendaguru. Sitzungsberichte der Berliner Gesellschaft naturforschender Freunde. Berlin. 6:219-247. Hennig, E. 1916. Zweite Mitteilung iiber den Stegosauriden des Tendaguru. Sitzungsberichte der Berliner Gesellschaft naturfor- schender Freunde. Berlin. 6:175-182. Huene, F. von. 1906. Uber das Hinterhaupt von Megalosaurus aus Stonesfield, Neues Jahrbuch f ir Mineralogie, Geologie and Palaeontologie 1:1-12. Huene, F. von. 1929. Los Saurisquios y Ornitisquoios del Creta- ceo Argentino. Anales del Museo de La Plata 2a, 3: 196 pp. 44 pis. Huene, F. von. 1932. Die fossile Reptil-Ordnung Saurischia. Monographien zur Geologie and Palaeontologie 1,4:361 pp. 56 pis. Jones, M. D. 1970. Cetiosaurus oxoniensis, Phillips, J.. Middle Jurassic sauropod from Rutland, England. Transactions of the Leicester Literary and Philosophical Society 64:144-160. Lydekker, Richard. 1888. Catalogue of the fossil Reptila and Amphibia in the British Museum (Natural History). Pt. 1:309 pp. McIntosh, J. S. 1981. Annotated Catalogue of the Dinosaurs (Reptilia, Archosauria) in the Collections of the Carnegie Museum of Natural History No. 18:65 pp. Marsh, O. C. 1889. Notice of new American Dinosauria. The American Journal of Science (3) XXXVIL33 1-336. Marsh, O. C. 1896. The dinosaurs of North America. Sixteenth Annual Report of the United States Geological Survey 1895, pp. 133-414, pis. II-XXXV. Mook, C. C. 1917. Criteria for the determination of species in the Sauropoda, with description of a new species of Apatosaurus. Bulletin of the American Museum of Natural History XXXVIIL355-360. Mook, C. C. 1967. Preliminary description of a new goniopholid crocodilian. Kirtlandia. 2:1-10. Owen, R. 1840-1845. Odontography, London. Pearl, R. M. 1975. The dinosaur menagerie. Earth Science. 28:6, 299-301. Phillips, J. 1871. The Geology of Oxford. Oxford, Clarendon Press. Piel, G. 1963. Last word from the dinosaur. Explorer 5:6, 14-19. 26 McINTOSH AND WILLIAMS No. 43 Reynolds, S. H. 1936. On a collection of reptilian bones from the Oolit of Stow-on-the-Wold, Gloucestershire. The Geological Magazine 76:193-214. Riggs, E. S. 1903. Structure and relationships of opisthocoelian dinosaurs. Part I. Apatosaurus. Field Columbian Museum Publications 82, Geological Series II, 4:165-196, pis. xlv-liii. Riggs, E. S. 1904. Structure and relationships of opisthocoelian dinosaurs. Part II. The Brachiosauridae. Field Columbian Museum Publications 94, Geological Series II, 6:229-247, pis. lxxi-lxxv. Romer, A. S. 1966. Vertebrate Paleontology. University of Chi- cago Press: Chicago and London. Williams, M. E. 1982. Paleontological displays of the Cleveland Musuem of Natural History. Fossils Quarterly 1:20-26. KIRTLANDIA THE CLEVELAND MUSEUM OF NATURAL HISTORY Cleveland, Ohio July 1988 Number 43:27-36 ISOLATED TETRAPOD REMAINS FROM THE CARBONIFEROUS OF WEST VIRGINIA STEPHEN J. GODFREY Redpath Museum, McGill University 859 Sherbrooke St. West, Montreal, Quebec, Canada H3A 2K6 Abstract In addition to Greererpeton (Amphibia; Temnospondyli) and Protero- gyrinus (Amphibia; Anthracosauria), Carboniferous tetrapod fossils from Greer, West Virginia, include fragmentary remains attributed to Crassigy- rinus, Eoherpeton, and at least one other taxon, a large temnospondyl amphibian. The description of additional tetrapods from Greer demonstrates that in terms of its taxonomic diversity it resembles other Euramerican Carbonif- erous localities which span the Visean-Namurian boundary. 28 GODFREY No. 43 Abstract In addition to Greererpeton (Amphibia; Temnospondyli) and Protero- gyrinus (Amphibia; Anthracosauria), Carboniferous tetrapod fossils from Greer, West Virginia, include fragmentary remains attributed to Crassigy- rinus, Eoherpeton, and at least one other taxon, a large temnospondy! amphibian. The description of additional tetrapods from Greer demonstrates that in terms of its taxonomic diversity it resembles other Euramerican Carbonif- erous localities which span the Visean-Namurian boundary. Introduction The first vertebrate remains recovered from Carbonifer- ous deposits within a limestone quarry operated by the Greer Limestone Company near Greer, West Virginia, were discovered in 1948 by an amateur, L. R. Collins, (Hotton 1970). The quarry lies just north of State Route 7, about 10.5 km southeast of Morgantown, Monongalia County (Universal Transverse Mercator Grid, Zone 17 NJ 991809). More detailed descriptions of the locality are provided by Hotton (1970) and Romer (1969, 1970). In July of 1968, J. J. Burke and W. E. Moran discovered, in a slab which had fallen from the quarry wall, the incomplete remains of a new temnospondyl amphibian, Greererpeton burkemorani Romer, 1969. By 1969, this quarry was recognized as yielding some of the oldest non-ichthyostegalian tetrapods. In 1970, A. S. Romer and N. Hotton III published descriptions of the new anthracosaurian amphibians from Greer, Proterogyrinus scheelei and Mauchchunkia bassa respectively. Because no consistently unique features dis- tinguish the latter from the former. Holmes (1984) con- cluded that Machchunkia was the junior synonym of Pro- terogyrinus. Between 1969 and 1973, field parties from the Cleveland Museum of Natural History under the direction of D. H. Dunkle recovered a large number of vertebrate skeletons from a restricted bed within the quarry (Fig. 1). The main bone-bearing layer, exposed for some 15m along the vertical face of the quarry, contained a nearly solid mass of fish and amphibian remains. Greererpeton, represented by at least 60 individuals, was by far the most common vertebrate recovered. Several virtually complete specimens were pre- served literally head-to-tail (Godfrey 1986). This deposit also produced articulated skeletons of the lungfish Tranodis castrensis Thomson and articulated skeletons of the anthra- cosaur Proterogyrinus scheelei Romer. In addition to these spectacular finds, the Cleveland Museum of Natural History (CMNH) and the Museum of Comparative Zoology (MCZ) recovered numerous spines of the acanthodian Gyracanthus, one shark spine (M. E. Williams, CMNH, personal communication), disarticu- lated remains of palaeoniscoid fishes, isolated elements of a large rhizodont crossopterygian, cf. Strepsodus, and iso- lated postcranial remains of several heretofore undescribed tetrapods. This paper deals with only the undescribed tetrapod remains. Vertebrate remains at Greer were recovered from the Bickett Shale (Bluefield Formation), which forms the lower subdivision of the Mauch Chunk Group (Busanus 1974). The Bickett Shale is lithologically the most variable unit within the Bluefield Formation, exhibiting rapid lateral changes in lithology. This nonmarine unit consists primarily of red and green mudstones but medium-grained sandstones —which have been interpreted as fluvial point bar and overbank deposits— are also present (Busanus 1976). The tetrapod remains (which include fragments of Greererpeton and Proterogyrinus ) recovered by the Museum of Compar- itive Zoology were apparently found about lm below the main vertebrate bearing bed quarried by the Cleveland Museum (C. Scaff, MCZ, personal communication). It is not known whether the fragmentary tetrapod remains col- lected by the Cleveland Museum were removed from the main layer or from the ‘bone bed’ below it. Fig. 1. Quarrying operations by members of the Cleveland Museum of Natural History field party in a section of the Greer Limestone Company quarry, Monogalia County, West Virginia. Fossiliferous rock within the Bickett Shale is being removed from the West wall of the quarry. Circa 1970. Kirtlandia, No. 43, July 1988 © by the Cleveland Museum of Natural History 1988 TETRAPOD REMAINS FROM GREER, W.VA. 29 Because Visean and Namurian tetrapods are exceedingly rare, a description of this material is important for it substantiates a more diverse assemblage than previously recognized at Greer. Recent descriptions of Scottish Car- boniferous tetrapods by Panchen (1985) and Smithson (1985a, 1985b) permit a more precise taxonomic assign- ment for some of the tetrapod remains from Greer. Com- paring taxonomic diversity indicates that Euramerican Lower Carboniferous and basal Upper Carboniferous tetra- pod assemblages are similar. I have opted to refrain from erecting formal taxonomic names because to do so on such fragmentary material would be premature. Although it seems unlikely, the isolated remains could represent elements from five different species. Description Vertebrae CMNH 11239 (Figs. 2-5) consists of two large rhachit- omous vertebrae. Both vertebrae conform to the typical rhachitomous pattern, being composed of a relatively mas- sive intercentrum and paired pleurocentra that bear large facets to support the neural arch. These vertebrae are clearly distinct from the essentially embolomerous verte- brae of Proterogyrinus, and their size and proportions preclude the possibility that they pertain to Greererpeton. The three intercentra preserved in CMNH 11239 are massive and remarkably similar to those of Eryops (Moulton 1974) (Fig. 2, A-D). Ventromedially, a flat longitudinal ridge is only weakly developed and may represent the zone of fusion between paired anlagen. In lateral view, the ventral margin of the intercentrum is gently concave. The median ridge is flanked by two shallow concavities formed by periosteal bone that is deflected along the anterior and posterior margins of the intercentrum (Fig. 2, A and D). The periosteal bone is perforated by minute pits. The concave ventrolateral surfaces vanish dorsally as the periosteal bone gathers to form a small semicircular facet that articulates with the capitulum of the rib (Fig. 2A). The parapophyses project little beyond the margin of the bone but can be seen in anterior or posterior view. A narrow strip of finished bone curves around each parapophysis to termi- nate on the apex of the wedge. On the two anterior-most intercentra, the facet for the capitulum of the rib is very small and only weakly developed (Fig. 2D). The anterior, posterior, and dorsal (notochordal) surfaces are rough and unfinished. Because the intercentrum is very thick mediolaterally, the notochordal notch is relatively small. In typical rhachitomous fashion, the massive intercentra from adjacent segments abut ventromedially (Fig. 5 A). Fig. 2. CMNH 11239, elements of the centrum. A-C, a thoracic intercentrum in left lateral, anterior, and ventral views respectively; D, a cervical intercentrum in left lateral view; E-F, a pleurocentrum half in posterior and anterior views respectively. Scale bar equals 1 cm. In life, the pleurocentra of CMNH 11239 were situated above the notochord and the main ossified portion of the intercentrum (Figs. 2 E-F and 5 A-B). In lateral view, each pleurocentrum half is an elongated oval that bears a protu- berant semicircular articular facet on its anterodorsal mar- gin that received the pedicle of the neural arch. This facet of unfinished bone forms an angle of approximately 60 ° with a plane tangential to the dorsolateral curvature of the external surface of the pleurocentrum. The external surface below and around the posterior margin of the facet is finished with smooth bone. Ventrally, the pleurocentrum ends in a blunt point. In anterior or posterior view the pleurocentrum is gently curved to surround the noto- chordal space. The internal surface is unfinished. Opposing antimeres of one pleurocentrum probably abutted above the notochord and immediately below the spinal cord (Fig. 5 B). 30 GODFREY No. 43 Fig. 3. CMNH 11239, A-C, specimen drawings of a ‘thoracic’ neural arch in left lateral, anterior, and dorsal views respectively. Scale bar equals 1 cm. The two neural arches preserved in CMNH 11239 differ in the structure of the neural spine and pedicle (Figs. 3 and 4). The distal end of one of the neural spines is greatly expanded laterally and markedly rugose (Fig. 4), whereas the other is much more slender and laterally compressed (Fig. 3). Based on the structure of the neural arches in Greererpeton (Godfrey 1986), laterally expanded neural spines and poorly developed diapophyses occur in the cervical series, whereas narrow, laterally compressed spines and longer transverse processes characterize more posterior vertebrae. Because both spines are unfinished dorsally, each was probably capped by cartilage in life. One of the more remarkable features of both arches is the retention of a large supraneural space (Figs. 3 B and 5 B). The space remains open for half the height of the spine as measured from the base of the postzygapophyses. A supra- neural space (canal) is primitive for tetrapods and is thought to have accommodated a supraneural ligament (Smithson 1985a). In anterior view, the neural spine is roughly the shape of an inverted V. On the posteromedial surfaces of each spine, immediately above the postzygapophysis, is a conspicuous groove that passes dorsally to terminate where the arch halves meet above the supraneural space. Both pre- and postzygapophyses are roughly ovoid in outline and inclined very little from the horizontal. Imme- diately below each prezygapophysis is a deep pit which may have led to a foramen (Fig. 3 B). This feature is not known in any other early tetrapod. Both arches possess large, broadly oval to rectangular facets on their ventromedial surfaces to receive the articular facets of the pleurocentrum. Immediately above the lateral margin of the pedicle is a narrow strip of periosteal bone. Above this strip the transverse process is poorly developed. In the cervical neural arch, the diapophyses are almost non-existent and probably did not hold the tuberculum of a rib in life (Fig. 4 A and B). On the other neural arch, protuberant flange-like diapophyses are developed. Judging from the relatively small size of these narrow, anteroven- trally directed facets, the tuberculum must have been small. Fig. 4. CMNH 1 1239, A and C, specimen drawings of a cervical neural arch in left lateral and dorsal views respectively; B and D, restorations of the arch in lateral and dorsal views respectively. Scale bar equals 1 cm. To summarize: although the vertebrae of CMNH 11239 resemble superficially those of Eryops, they differ from that genus in the possession of the following features: 1) a large supraneural space; 2) a relatively short neural spine; 3) widely separated contralateral pre- and postzygapophyses; and 4) a deep pit or foramen below each prezygapophysis. These vertebrae are probably derived from a large rhachi- tomous temnospondyl amphibian. 1988 TETRAPOD REMAINS FROM GREER, W.VA. 31 Fig. 5. Restoration of vertebrae based on CMNH 11239; A, left lateral view; and B, posterior view. Scale bar equals 1 cm. Ribs CMNH 1 1230 consists of a consecutive series of 10 or 11 large but incompletely preserved ribs lying on the visceral surface of an articulated series of ventral scales (Fig. 6). Although no rib heads are preserved, the morphology of the gastralia indicates that the right side of the body is repre- sented and that the narrow end of the slab in Figure 6 is the anterior end. The ribs lack characteristic features of cervi- cal, sacral, or caudal series and thus probably represent a segment from the midtrunk region. Although all the ribs are broken or variably crushed, they retain much of their original curvature. Because the radius of curvature of the eighth rib is just over 7 cm, the trunk diameter at that point was probably 15 cm or more if rib curvature followed that of the body. All the ribs are ornamented with longitudinal striations as in Crassigyrinus (Panchen 1985). The rib shafts lack an uncinate process or flange found in some temnospondyls. The large size of the ribs and the morphol- ogy of the gastralia indicates that they do not pertain to Proterogyrinus. Although some of the scales are scattered, articulated series show that the original pattern was en chevron (Fig. 7 A). The scales are shaped much like an elongate tear drop. The anterolateral margin of each scale is more conspicu- ously convex than the trailing posteromedial edge. Their ventral surface is smooth and convex whereas their visceral surface is concave. A rounded ridge (Fig. 7 B) runs the length of the scale along the posterodorsal (internal) sur- face. In ventral view, the medial margin of each scale underlaps the trailing edge of its medial neighbor. The scales in CMNH 11230 resemble those of Crassigyrinus. Pending further discoveries, CMNH 11230 is tentatively assigned to the genus Crassigyrinus. Scapulocoracoid MCZ 8950 (Fig. 8) is a single incompletely preserved right scapulocoracoid. The thin coracoid plate appears to be virtually complete, whereas the margin of the scapular blade anterodorsal to the supraglenoid buttress is poorly represented. The position of foramina and its overall pro- portions preclude it from belonging to either Greererpeton or Proterogyrinus (Fig. 8, D-F). The glenoid is fairly well developed and forms a concave horizontal strap of unfinished bone that shows very little 32 GODFREY No. 43 Fig. 6. CMNH 11230, 10 or 11 incomplete ribs and ventral scales from the right side of the body in dorsal view. Two ribs, 8 and 9 in the series (counting from the narrow end of the slab), were fractured in life and show a swollen mass of ankylosed bone distally. Scale bar equals 1 cm. torque along its surface. Except for the anterior one-third of the glenoid, which is directed postero-ventrolaterally, the fossa is directed laterally. Above the anterodorsal margin of the glenoid on the posterodorsal-facing margin of the ascending scapular blade, a small boss is developed that is visible in lateral view (Fig. 8 A and B). Extending posteriorly from the base of this small tuberosity, along the surface immediately medial to the glenoid, is a small ridge which ends on the posterior margin of the bone. A second, more prominent ridge lies medial and parallel to this ridge and continues from the posterior margin of the bone anterodorsally onto the supraglenoid buttress, where it loses prominence. In medial view, the supraglenoid buttress thins rapidly anterior to the infraglenoid buttress. The ascending scapular moiety is not preserved. The supraglenoid buttress is not perforated by the supraglenoid foramen, but a C-shaped notch (Fig. 8 A), immediately anterior to the supraglenoid buttress and level with the glenoid articulation, may repre- sent the remnants of this foramen. The foramen passes through the 3 mm thick bone in a ventrolateral direction. Alternatively, this foramen might represent the supracora- coid foramen, and the supraglenoid foramen may have been lost. A second large foramen is visible laterally, just Fig. 7. Restoration of gastralia in ventral view based on CMNH 11230 (the narrow end of each scale points medially and the anterior end of the body lies towards the top right-hand corner of the page); and B, one scale enlarged, in dorsal view. Scale bar equals 1 cm. 1988 TETRAPOD REMAINS FROM GREER, W.VA. 33 beneath the anterior margin of the glenoid. It passes through the scapulocoracoid in a dorsomedial direction to open on the medial surface at the base of the supraglenoid buttress. The homology of this foramen is difficult to determine but, as in Eoherpeton (Smithson 1985a), it could represent the common opening of the supracoracoid and glenoid foramen. The unfinished bone forming the thin ventromedial margin of the coracoid plate was probably finished by cartilage in life. Both medial and lateral surfaces of the plate are perforated by numerous ‘nutritive’ foramina. The antero-ventromedial portion of the scapulocoracoid is per- forated by another fairly large foramen which passes vertically through the bone. Its location is almost identical to the anterior coracoid foramen seen in Protorothyris archeri (Clark and Carroll 1973, Figs. 5 A and 6 F). Humeri MCZ 8951 (Fig. 9) is an incomplete right humerus that possesses several features that are not present in Greerer- peton or Proterogyrinus. On the dorsal surface of the humerus immediately above the proximal articulation is a conspicuous boss or tubercle. Immediately behind the tubercle is a concavity which may have marked the insertion of m. subcoracoscapularis. The processus latissimus dorsi is damaged in MCZ 8951. The ectepicondyle ridge begins at the base of the processus latissimus dorsi and broadens as it curves gently towards the anterodistal corner of the bone. Distally, the dorsal margin of the condyle was damaged and, therefore, the exact height of the ridge is not known. A small posteriorly directed flange of bone is present on the postaxial surface of the humerus (Fig. 9 A and C). Aside from Eoherpeton (Smithson 1985a), this flange has not been described on the humerus of other early tetrapods; consequently, its function remains unclear. Between the aforementioned flange and the processus latissimus dorsi is a shallow longitudinal trough which deepens distally and marks the postaxial base of the ectepicondyle. The deltopectoral crest in MCZ 8951 is remarkably large. Anteroventrally, the crest exhibits a shallow pit, above which lies a large tubercle. The pit and tubercle mark the major insertion of m. pectoralis and m. deltoideus respectively. A postaxially directed ridge extends from the deltoid tubercle across the dorsal surface of the humerus to merge with the anterior end of the ectepicondyle ridge at the base of the processus latissimus dorsi (Fig. 9 A). As in some early amniotes (Romer 1956), this ridge presumably marks the insertion of m. scapulohumeralis anterior. The proximoventral surface of MCZ 8951 displays two prominent areas of muscle attachment. A pronounced preaxial tubercle marks the insertion of m. supracoracoid- eus and, more posteriorly, a broad concavity marks the insertion of m. coracobrachialis. The area of insertion of these and other muscles on MCZ 8951 is based on the description of other more completely known tetrapods Fig. 8. MCZ 8950, A-C, specimen drawings of an incomplete right scapulocoracoid in lateral, medial, and posterior views respectively. Partially restored right scapulocoracoid of D, Greererpeton burlcemorani; E, Proterogyrinus scheelei; and F, MCZ 8950, all drawn to the same scale. P. scheelei after Holmes (1980). Scale bars equal 1 cm. Abbreviations: anterior coracoid foramen, f ant cor; supracoracoid foramen, f scor; supraglenoid buttress, sgl bt; supraglenoid foramen, f sgl. (Romer 1956; Holmes 1980; Smithson 1985a). The supinator flange and entepicondyle are incomplete in MCZ 8951. The preserved section of the entepicondyle differs from that seen in Eoherpeton in that it extends distally as a smooth, broadly rounded flange beyond the trochlea (ulnar articulation), much as it does in Eryops (Miner 1925). The distal and postaxial margins of the entepicondyle are finished with periosteal bone that is rounded from dorsal to ventral surfaces. The entepicondyle foramen passes from the proximal anterodorsal surface of the entepicondyle in a ventrodistal direction. In ventral view, the foramen exits below a distinct ridge that sweeps 34 GODFREY No. 43 Fig. 9. MCZ 8951, A-D, specimen drawings of an incomplete right humerus in dorsal, preaxial, ventral, and proximal views respectively. Scale bar equals 1 cm. Abbreviations: deltopectoral crest, delt cr; ectepicondyle, ect; entepicondyle foramen, f ent; m. coracobrachialis, cbl; m. scapulohumeralis anterior, scha; m. subcoracoscapularis, sbcsc; m. supracoracoideus, spec; processus latissimus dorsi, pr lat d; supinator flange, sup. postaxially along the proximal margin of the entepicondyle. As the entepicondyle is incomplete, the extent of this ridge is unknown. A shallow groove extends in an anterodistal direction across the ventral surface of the entepicondyle from the entepicondyle foramen to vanish below the ulnar articulation (Fig. 9 C). Radial and ulnar condyles are poorly defined. In MCZ 8951, the long axis of the proximal articulation forms an angle of 45-50 ° with the plane of the entepicondyle (Fig. 9 D). Although somewhat smaller, MCZ 8951 bears a striking similarity to the humerus of Eoherpeton as described by Smithson (1985a). Minor differences between the two may be attributed to size. Among Lower Carboniferous tetra- pods, only Eoherpeton approaches MCZ 8951 in the large number of conspicuous areas for muscle attachment. MCZ 8951 is provisionally assigned to this genus. Fig. 10. MCZ 8952, A-E, specimen drawings of a left humerus in preaxial, dorsal, slightly ventral of the postaxial surface, ventral, and proximal views respectively. Scale bar equals 1 cm. Abbreviations as in Fig. 9. Although the second undescribed humerus from Greer (MCZ 8952, Fig. 10) is of the tetrahedral type common to most early tetrapods, it is unlike other humeri from this locality in that it is a more robust element with a higher degree of torsion and is without most of the tuberosities or ridges marking the origin or insertion of limb musculature. 1988 TETRAPOD REMAINS FROM GREER, W.VA. 35 If the plane of the entepicondyle is positioned horizon- tally, the unfinished convex strap of bone forming the proximal articulation is nearly vertical and remarkably compressed anteroposteriorly (Fig. 10 E). The articular surface is poorly defined. A short distance behind the postaxial margin of the proximal articulation is a small tuberosity, probably marking the insertion of m. subcora- coscapularis. This tuberosity forms a ridge that continues posterodistally to form the anterodorsal margin of the entepicondyle. On the preaxial side of this tuberosity, a shallow groove passes distally and becomes progressively more pronounced as it undercuts the ridge that forms the ectepicondyle. The posterodorsally projecting ectepi- condyle is unfinished. The deltopectoral crest in MCZ 8952 is poorly devel- oped, and the insertions of m. pectoralis and m. deltoideus are not defined. A small ridge passes postaxially from the posteroventral corner of the deltopectoral crest onto the entepicondyle above the entepicondyle foramen. Anterior to this ridge the proximoventral surface of the bone in MCZ 8952 shows two V-shaped concavities that are separated by a second low, rounded ridge. The anterior-most concavity is larger and lies immediately postaxial to the deltopectoral crest. The second concavity, with the top of its V-shaped margin directed proximally, lies adjacent to the small tuberosity that may represent the insertion of m. subcora- coscapularis. The deltopectoral crest and the supinator process are continuous, forming a large preaxial flange. The dorsal preaxial surface of the supinator flange bears a bulbous tubercle that may have held the origin of m. supinator. The dorsal surface of the entepicondyle bears a broad but low ridge that extends proximodistally. This ridge increases substantially the thickness of the entepicondyle distally and presumably accommodated the radial, ulnar, or both radial and ulnar condyles. MCZ 8952 displays a remarkably high degree of longi- tudinal torsion. The long axis of the proximal articulation forms an angle of about 64° with the plane of the entepi- condyle. Discussion The description of several additional tetrapods from Greer indicates that this assemblage is more diverse than previously recognized. In terms of its taxonomic comple- ment, the Greer tetrapod fauna resembles two localities from the Midland Valley of Scotland and three North American localities (Table 1). The faunal similarities be- tween Greer, Gilmerton (Smithson 1985b), Keokuk County (McKay et al. 1987), Cowdenbeath (Smithson 1985b, 1986), Point Edward (Carroll et al. 1972), and Hinton (Smithson 1982 and personal communication) are not all that surprising. All Visean and Namurian tetrapods are derived from a narrow equatorial strip along the southern region of ‘pre-drift’ Euramerica (Milner et al. 1986). These sites extend from present-day Iowa in the west to the Federal Republic of Germany in the east and demonstrate that known localities are remarkably restricted latitudinally. Their faunas were dominated by large, primarily aquatic tetrapods. TABLE 1 Distribution of Euramerican Amphibians across the Visean-Namurian Boundary STAGE < X LOCALITY eS Crassigyrinus Loxomma Spathicephalus Eoherpeton Doragnathus Proterogyrinus Pholiderpeton Undescribed Anthracosaur Pholidogaster Greererpeton Large Temnospondyl Adelogyrinidae c .2 Hinton, West Virginia *9 *? *? *9 * p £ Point Edward, Nova Scotia * * *9 p Z Cowdenbeath, Fife Region * * * * * * G 03 Greer, West Virginia *9 *9 * * * - CO Keokuk County, Iowa * *? > Gilmerton, Lothian Region * * * * The ventral surface of the entepicondyle is conspicuously concave, with its deepest point marked by the opening of the entepicondyle foramen. The foramen passes from the ven- tral suface in a posterodorsal and slightly medial direction and opens near the proximal margin of the entepicondyle. Acknowledgments I am indebted to Michael E. Williams (Cleveland Museum of Natural History) and Farish A. Jenkins and Charles R. Schaff (Museum of Comparitive Zoology, Harvard University), for the 36 GODFREY No. 43 loan of specimens that formed the basis of this paper. Collection and initial preparation of the material from the Cleveland Museum were supported by the U.S. National Science Foundation under grant G.B.— 35474. Many comments from R. W. Hook (Austin, Texas), J. R. Bolt (Field Museum of Natural History), anonymous reviewers, T. R. Smithson (University of Newcastle-upon-Tyne), R. L. Carroll (Redpath Museum, McGill University), and D. W. Dilkes (Uni- versity of Toronto) improved the contents of this paper. The assistance of D. Baird (Princeton University) and the late J. Moulton in the early stages of this work is also gratefully acknowledged. This research was supported by the Natural Sciences and Engineering Research Council of Canada, and the National Science Foundation of the United States of America. References Busanus, J. W. 1974. Paleontology and paleoecology of the Mauch Chunk Group in Northwestern West Virginia. Unpub- lished M. S. thesis. Bowling Green State University, Bowling Green Ohio. Busanus, J. W. 1976. Faunal distribution within the Greenbrier- Mauch Chunk transition (Chesterian: Elviran) along the Chest- nut Ridge Anticline. Geological Society of America, Abstracts with Programs 8(2): 143. Carroll, R. L., E. S. Belt, D. L. Dineley, D. Baird, and D. C. McGregor. 1972. Vertebrate Paleontology of Eastern Canada. Excursion A59, 24th International Geological Congress. Clark, J., and R. L. Carroll. 1973. Romeriid reptiles from the Lower Permian. Bulletin of the Museum of Comparative Zool- ogy 144:353-407. Godfrey, S. J. 1986. The skeletal anatomy of Greererpeton burkemorani Romer 1969, an upper Mississippian temno- spondyl amphibian. Unpublished Ph.D. dissertation, McGill University, Montreal. Holmes, R. 1980. Proterogyrinus scheelei and the early evolution of the labyrinthodont pectoral limb. In The Terrestrial Environ- ment and the Origin of Land Vertebrates. Edited by A. L. Panchen. Systematics Association Special, 15:351-376. Lon- don, Academic Press. Holmes, R. 1984. The Carboniferous amphibian Proterogyrinus scheelei Romer, and the early evolution of tetrapods. Philosoph- ical Transactions of the Royal Society of London, (B) 306:431- 527. Hotton, N. 1970. Mauchchunkia bassa, gen. et. sp. nov. , an anthracosaur (Amphibia; Labyrinthodontia) from the Upper Mississippian. Kirtlandia 12:1-38. McKay, R. M., B. J. Witzke, M. P. McAdams, and J. R. Bolt. 1987. Early tetrapods and deposition of the Upper St. Louis Formation (Mississippian; Visean) in Keokuk County, Iowa. No. 129575. Geological Society of America, Abstracts with Programs 19(4):233. Milner, A. R., T. R. Smithson, A. C. Milner, M. I. Coates, and W. D. I. Rolfe. 1986. The search for early tetrapods. Modem Geology 10:1-28. Miner, R. W. 1925. The pectoral limb of Eryops and other primitive tetrapods. American Museum of Natural History, Bulletin 51:145-312. Moulton, J. M. 1974. A description of the vertebral column of Eryops based on the notes and drawings of A. S. Romer. Breviora 428:1-44. Panchen, A. L. 1985. On the amphibian Crassigyrinus scoticus Watson from the Carboniferous of Scotland. Philosophical Transactions of the Royal Society of London, (B) 309:505-568. Romer, A. S. 1956. Osteology of the Reptiles. The University of Chicago Press, Chicago and London. Romer, A. S. 1969. A temnospondylous labyrinthodont from the Lower Carboniferous. Kirtlandia 6:1-20. Romer, A. S. 1970. A new anthracosaurian labyrinthodont, Proterogyrinus scheelei, from the Lower Casrboniferous. Kirt- landia 10: 1-16. Smithson, T. R. 1982. The cranial morphology of Greererpeton burkemorani Romer (Amphibia: Temnospondyli) . Zoological Journal of the Linnean Society 76:29-90. Smithson, T. R. 1985a. The morphology and relationships of the Carboniferous amphibian Eoherpeton watsoni Panchen. Zoo- logical Journal of the Linnean Society 85:317-410. Smithson, T. R. 1985b. Scottish Carboniferous amphibian local- ities. Scottish Journal of Geology 21:123-142. Smithson, T. R. 1986. A new anthracosaur amphibian from the Carboniferous of Scotland. Palaeontology 29:603-628. Membership and Subscription Membership in Kirtlandia Society is open to any member of the Cleveland Museum of Natural History. Current individual membership is $30.00 ($10.00 for students); participating members, $100.00; institutional, $30.00; sustaining, $250.00. Applications and fees should be sent to the Treasurer, Kirtlandia Society, Cleveland Museum of Natural History, Cleveland, Ohio 44106. Membership includes subscription to Kirtlandia. There is a non-U. S. postal surcharge of $5.00. All subscriptions outside North America will be sent by air. Payment must be made in U.S. dollars. Payments not drawn on a U.S. baank will be subject to bank collection charges. Change of Address: Please send to Museum six weeks prior to move; enclose present mailing label with change of address. Claims for missing issues: Claims cannot be honored beyond four months after mailing date. Duplicate copies cannot be sent to replace issues not delivered because of failure to notify Museum of change of address. Back issues: Individuaal issues of Kirtlandia are available from the Museum Publications Office. A complete list is available upon request. •NATURAL HISTORY* Anthropology A Method for Making Three-dimensional Reproductions of Bones and Fossils 3 Jenny A. Smith and Bruce M. Latimer Archaeology The Squaw Rockshelter (33CU34): A Stratified Archaic Deposit in Cuyahoga County 17 David S. Brose Skeletal Remains from Squaw Rockshelter 55 Fred Prior Knotted Cordage from Squaw Rockshelter (33CU34), Aurora Run, Cuyahoga County 59 R. L. Andrews and J. M. Adovasio MARCH 1989 KIRTLANDIA The Scientific Publication of The Cleveland Museum of Natural History David S. Brose, Editor Joseph T. Hannibal, Assistant Editor Brief History and Purpose Kirtlandia , a publication of The Cleveland Museum of Natural History, is named in honor of Jared Potter Kirtland, a noted nineteenth-century naturalist who lived in the Cleveland, Ohio area. It began publication in 1967 and is a continuation of the earlier series Scientific Publications volumes I to 10 (1928-1950), and new series volumes 1 to 4 (1962-1965). Supported by the Kirtlandia Society of The Cleveland Museum of Natural History, Kirtlandia is devoted to the publication of scientific papers in the various fields of inquiry within the Museum’s sphere of interest: Cultural and Physical Anthropology; Archaeology; Botany; Geology; Paleobotany; Invertebrate and Vertebrate Paleontology; Systematic Ecology; and Invertebrate and Vertebrate Zoology. Issues will vary from single monographs to collections of short papers, review articles, and brief research notes. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted for libraries and other users registered with the Copyright Clearence Center (CCC) Transactional Reporting Service, provided that the base fee of 50 cents per copy, plus 25 cents per page is paid directly to CCC, 27 Congress Street, Salem, MA 09179. The fee code for users of the Transactional Reporting Service is 0075-6245/89 $00.50+. 25. Kirtlandia is abstracted in Biological Abstracts and indexed in Bibliography and Index of Geology and Zoological Record. Kirtlandia No. 44 © 1989 by The Cleveland Museum of Natural History Cleveland, Ohio KIRTLANDIA The Cleveland Museum of Natural History March 1989 Number 44 Anthropology A Method for Making Three-dimensional Reproductions of Bones and Fossils 3 Jenny A. Smith and Bruce M. Latimer Archaeology The Squaw Rockshelter (33CU34): A Stratified Archaic Deposit in Cuyahoga County 17 Day id S. Brose Skeletal Remains from Squaw Rockshelter Fred Prior 55 Knotted Cordage from Squaw Rockshelter (33CU34), Aurora Run, Cuyahoga County 59 R. L. Andrew’s and J. M. Adovasio KIRTLANDIA Editor Dr. David S. Brose Assistant Editor Joseph T. Hannibal The Cleveland Museum of Natural History Associate Editors James K. Bissell, Curator of Botany The Cleveland Museum of Natural History Bruce Latimer, Curator of Physical Anthropology The Cleveland Museum of Natural History Martin Rosenberg, Editorial Assistant Case Western Reserve University Sonja Teraguchi, Manager of Collections The Cleveland Museum of Natural History Editorial Advisory Board Rodney Feldmann, Professor of Geology Kent State University Michael C. Hansen, Geologist Ohio Geological Survey Richard Meindl, Associate Professor of Anthropology Kent State University G. Michael Pratt, Associate Professor of Anthropology Heidelberg University David H. Stansbery, Director, Museum of Zoology Ohio State University Frederick H. Utech, Curator of Botany Carnegie Museum of Natural History Ed Voss, Curator of the Herbarium University of Michigan Andrew M. White, Professor of Biology John Carroll University KIRTLANDIA The Cleveland Museum of Natural History March 1989 Number 44:3-16 A Method for Making Three-dimensional Reproductions of Bones and Fossils Jenny A. Smith and Bruce M. Latimer Department of Physical Anthropology The Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Abstract The production and dissemination of high quality casts is an integral part of the paleontological and anatomical sciences. However, detailed descriptions of the techniques for making high resolution, three dimensional reproductions are not readily available. The casting laboratory at the Cleveland Museum of Natural History has developed a technique for making detailed, dimensionally stable reproductions of fossil and skeletal materials. Experimentation with molding compounds has shown Dow Corning Silastic, RTV E, to be an excellent material for our purpose, in spite of its extreme sensitivity to certain organotin compounds. Proper preparation of the original specimens prior to molding is therefore crucial in the molding procedure. The original specimen is placed in a clay base and is covered with a number of thin coats of deaired silastic. A plaster jacket is formed over the flexible rubber layer to prevent angular distortion. The half mold is inverted, the clay removed, a thin coat of parting compound painted over the exposed silastic and the process is repeated. Casts are made in urethane plastic or dental stone. Hollow casts are made by partially filling and rotating the mold. These easily learned techniques allow the production of three dimensional casts of exceptional quality. Introduction The casting facilities of the Department of Physical Anthropology at the Cleveland Museum of Natural History were developed in 1975. At that time, the primary function of the casting laboratory was to duplicate the three-to-four- million-year-old hominid fossils that Donald Johanson had recently discovered in the Hadar region of Ethiopia. The techniques developed to mold and cast these fossils are now being used to replicate select specimens from the Hamann-Todd Osteological Collection, a large assemblage of recent human and nonhuman primate skeletal materials which is curated by the Museum’s Department of Physical Anthropology. We believe that these procedures produce accurate, stable reproductions of unequaled quality in both 4 Smith and Latimer No. 44 plaster and plastic. They are, moreover, easily learned and are readily adaptable to many situations, both field and laboratory. This paper specifically describes the molding and casting techniques developed and presently in use in the Laboratory of Physical Anthropology of the Cleveland Museum of Natural History (Smith and Latimer, 1985; 1986). It is intended for those with some experience in molding and casting and is not meant to be a review of, or introduction to casting techniques. Several articles, however, have detailed some of the many other molding and casting techniques that are currently in use (Burke, 1983; Heaton, 1980; Jensen, 1961; Madsen, 1974; O’Donnell and Hanley, 1983; Parsons, 1973; Quinn, 1940; Reser, 1981; Rigby and Clark, 1965; Schrimper, 1973; Siveter, 1982; Waters and Savage, 1971). Also, because molding and casting is a "hands-on” activity we encourage beginners to practice with less valuable objects. As many of the materials used in casting procedures are potentially hazardous (e.g. organic solvents, epoxies, urethanes), instructions should be followed carefully. Protective clothing including gloves (we recommend latex disposable) should be worn. Adequate ventilation, such as a fume hood, is a necessity during the mixing of the urethane plastics, Dynacast and MasterCast, and the application of trichloroethane, Glyptal, Krylon and silicone sprays. It is also advisable to leave the newly unmolded plastic casts under the hood for a 24-hour curing period. Our presentation is divided into three sequential sections. The first section describes specimen preparation before molding, the selection of the molding material and the seven basic steps of the molding process. The second section describes casting techniques for the production of plaster and plastic casts. The final section details finishing methods, repair of defects and painting techniques. The rotating casting machine used to make hollow casts is described in Appendix 2. Molding Molding is the most time consuming part in the entire procedure and should be done very carefully as mistakes are very difficult to remedy later. The selection of an appropriate molding material is crucial, as is the proper preparation of the original specimen prior to molding. The following description and illustrations will, for demonstrative purposes, focus on a simple two-part mold. However, the same techniques can be applied with slight modification to create a more complex, multiple-piece mold. Specimen Preparation The primary objective in mold making is to reproduce the greatest amount of detail without placing the original specimen at risk. In order to maximize mold detail and simultaneously protect the original specimen, preparation techniques must be followed carefully. A good mold will reproduce microscopic detail. It is therefore very important that the surface of the original specimen be cleaned so that it is free of any rubber from previous moldings, unnecessary adhesives, clay, pieces of matrix, plaster, etc. At this point, because the physical properties of the molding rubber could be affected by prior specimen preparation, it is necessary to discuss the choice of molding material. Our experience with various molding compounds has shown that Dow Corning Silastic RTV E (Appendix 1) molding rubber is superior in its ability to retain intricate detail over the duration of the mold life. A major drawback, however, is its extreme sensitivity to certain compounds. Contamination by sulfonates, amines and neoprene can severely inhibit this rubber’s vulcanization. It is therefore necessary to review the original specimens “history” as to how and when was it originally prepared and what types of materials (especially adhesives) were used. It is especially important to find out if the specimen had been molded previously. This latter issue is pertinent because many molding compounds are incompatible with one another and the residue from a previous molding may result in inhibition of the silastic rubber. In light of this, we strongly encourage that detailed and easily accessible records are kept regarding all preparation, molding and casting procedures. Finally, some fossil deposits naturally contain elements that will inhibit molding rubber. It has been our experience that fossils taken from a coal or shale matrix are likely to result in some degree of molding rubber inhibition due to the high sulfur content in these materials. In any case the important point is to test all specimens for chemical contamination prior to molding. This axiom holds for modern bone specimens as well as fossils; all originals must be checked for chemical contamination. Molding rubber inhibition has occurred if the rubber is gummy or uncured after the 24-hour curing period has elapsed. Compatibility should be tested by placing a drop of the mixed, uncured rubber on the surface to be reproduced. An individual drop of the silastic should be placed on each area exhibiting different surface structure or coloration. Again, we strongly recommend testing any object that is suspected as a potential problem; uncured, semi-liquid silastic is very difficult to remove. In the event of contamination, other silastics, for example Silicone H (see Dow Coming Product Guide for a list of Silastics), which are not as easily inhibited, may be used in place of Silastic E. Although this publication recommends the use of Silastic RTV E, the procedures can be easily modified for use with other Silastic rubbers and molding materials. Chemically contaminated areas can often be cleaned with a weak organic solvent such as acetone and, if necessary, treated with a barrier coat, such as Glyptal lacquer cement (Appendix 1 ) dissolved in acetone. After 1989 Bone and Fossil Reproductions 5 Fig lire 1. The frontal and inferior view of an adult male gibbon cranium that has been prepared for molding. Note that the nasal aperture and all foramina and canals are blocked with clay. coating the specimen or specific areas of the specimen with thinned Glyptal, the surface should be washed with acetone to reduce surface glaze. It is important to wait several hours to allow any remaining solvent to volatilize, a step which can be hastened by placing the specimen under a heat lamp. Because acetone will dissolve most commonly used Figure 2. Coronal section through skull showing reference plane used in Figures 5,7,9,11 ,13,15 and 16. adhesives, care must also be taken to ensure that any prior reconstructions or repairs are not damaged. Again, it is necessary to check that the adhesives are completely dry as any partially dried glue or softened adhesive will result in contamination of the molding rubber. Water when present on the original specimens can also result in inhibition. Having once tested the specimen to ensure its compatibility with the molding rubber, it is then necessary to “stabilize” fragile areas so they will not be damaged in the subsequent procedures. The glyptal acetone mixture described above can also be used to stabilize or strengthen fragile surfaces that might otherwise be damaged during the molding process. Painting especially delicate surfaces with petroleum jelly dissolved in 1-1-1 Trichloroethane (CH3CCI3) will prevent adhesion of the molding rubber and in so doing will facilitate removal of the original from the mold. This latter step reduces mold detail, but is often necessary with extremely fragile specimens. If the petroleum jelly solution is used, it must be carefully removed from the original specimen after the molding procedure is completed. After checking for molding rubber compatibility and stabilizing any especially fragile areas, the next step is to block small openings to prevent the penetration of the liquid molding rubber. This includes openings such as exposed medullary canals, foramina, areas between teeth or deep cracks in the surface of the specimens. These should be blocked to within a few millimeters of the external surface with clay or beeswax. We use Plastalina clay, a wax-type plastic clay. 6 Smith and Latimer No. 44 Figure 3. Step 1 — Lateral view of the skull partially embedded in clay block (shown as transparent) to demonstrate position of flash line. Note that the flash line follows natural contour lines and ridges, such as the edge of the tooth row, the zygomatics and the nuchal ridge along the back of the skull. but any clay which is compatible with the silastic rubber can be used. Again, it is important to check that the clay is compatible with the molding rubber. Many kinds of oil- base modeling clay inhibit RTV curing. We have found when blocking small foramina, beeswax due to its tackiness, works better than clay. In addition, because it is sticky and semi-transparent beeswax can also be used to create a thin “window” of material to block small openings. The resulting membrane of casting material formed in the completed cast can be easily removed with a needle or sharpened dental tool. This technique is especially useful when molding jaws and teeth Figure 4. Step 1 — Skull partially embedded in clay block to establish part line Note elevated tabs which act as registration guides. Note also that all elevations and depressions on the face of the mold are angular with planar surfaces. where spaces between individual teeth must be maintained. Again, if there is any question regarding compatibility, any clays and/or waxes should be tested prior to use. Deep cracks or fissures in the original makes “demolding” of the original as well as subsequent casts difficult. If this is judged to be the case, cracks and fissures should be blocked or filled to within a millimeter or two of the surface with wax or clay. Reconstructions of this type can strengthen an especially fragile original, but these must be sharply delimited so that there is no confusion on the finished cast as to the reconstructed and original areas. We believe that permanent fossil reconstructions should, when possible, be done on highly detailed plaster casts and not on actual specimens. However, if for some reason it is necessary to reconstruct an original specimen, the material chosen must be dimensionally stable as well as reversible and easily removed. Epoxy putty is a material which, when soft, can be configured to a particular shape and then allowed to dry away from the original specimen. Once hardened, the putty can be sanded and shaped and then glued in position with a reversible adhesive. Plaster or clay should not be utilized for fossil reconstructions as these materials are very difficult to remove from the surface of the bone. In addition, clay-like materials shrink as they age which can result in serious damage to the original. Making the Mold The molding procedure is discussed and illustrated in seven sequential steps to make it easier to follow the instructions. Please note that much of the important detail is contained in the figure captions. The specimen used in this example was a male gibbon ( Hylobates ) skull (Figure 1). Although the methods discussed here are for a two piece 1989 Bone and Fossil Reproductions 7 mold, the same basic procedures are followed for a multiple-piece mold. However, because of the extra part- lines and registration problems in complex molds, they should always be made as simple as possible. Specimens which are substantially larger or smaller than the gibbon skull will require some modification of the procedures. Step 1 The gibbon skull is partially embedded in clay to establish the mold “part” line. The positioning of the “part” or “flash” line around the specimen is critical and important anatomical features must not be obscured. Furthermore, deep undercuts and especially fragile areas must be taken into account when designing the mold to ensure that the original and plaster casts can be removed from the completed mold without damage. In the sample specimen the flash line is positioned to follow the natural contours of the skull and to produce a two-part mold, one- part approximately two-thirds of the total volume of the skull (Figures 3 and 4). This is an important factor when pouring hollow casts, when the volume of casting material used is approximately two-thirds of the total volume of the specimen (see Casting in Plastic, this article). Molds that are primarily for plaster casts are more easily handled when the two parts are of equal size (see Casting in Plaster , this article) but we have found that problems with registration occur if one part of the mold is less than one- third of the volume of the specimen. Thus the flash-line in the Hylobates skull was designed to follow natural lines and ridges on the skull, along the edge of the tooth row and zygomatics and the nuchal ridge. The clay is formed into a rectangular block which is deep enough to accept approximately half of the specimen. We find that the easiest way to do this is to roll the clay into approximately lcm-thick slabs. We have tried melting the clay and pouring it into shallow pans to make the clay slabs but this procedure proved less than satisfactory for numerous reasons and we now simply roll the clay mechanically. Rolling can be accomplished with a baker’s rolling pin, a process made easier by first working the clay under a heat lamp. In addition, a mechanical wringer such as those used on old fashioned washing machines also works to roll the clay slabs. If possible, it is important to keep the clay block in a square or rectangular shape with vertical walls as this will assist mold registration in later stages. Also, although the block must be of adequate size to accept the original, it should not be much more than 3cm to 5cm larger on any side. This facilitates handling of the completed mold (small molds are easier to manipulate) and also allows the use of a dental irrigation syringe to inject plaster into the closed mold (see below). After making sure that the clay is deep enough, it is placed on a base made of particle board, plywood or plexiglass. We recommend spraying the base with silicone, or putting wax paper or a paper towel between the clay and the base, to prevent the clay from sticking. The outline of the specimen is then inscribed on the clay surface and a hole is cut large enough for the specimen to be inserted with the predetermined part-line in the same plane as the clay surface. Small amounts of clay are added as needed and pushed against the specimen along the part-line, leaving a clean, sharp contact continuous with the plane of the clay. A small double bladed, stainless steel spatula with the tips of both the rounded and tapered ends bent to a 45=° angle is an excellent tool for this purpose. In cases where the natural topography of the specimen requires a depression or elevation of the clay, the edges should be sharply delineated and perpendicular. As much of the clay surface as practical should be smooth and horizontal in order to minimize potential shear strains and Smith and Latimer No. 44 Figure 6. Step 2 — First thin coat of silastic molding rubber spread over the original specimen. Care must be taken to ensure that air bubbles are not trapped between this primary coat and the surface of the original specimen. consequent deformation between the two halves of the mold when it is closed. During the formation of the part-line any unneeded clay should be removed to ensure a precise, continuous junction between the clay and the specimen. A series of elevated tabs, and/or regular depressions, are placed at regular intervals around the specimen to act as registration guides. These tabs should be angular with flat, beveled edges. Lastly, the edges of the clay slab should be cut to make a square or rectangle with sharp vertical sides. This will also assist in proper registration of the two halves of the closed mold. Step 2 This is the most important step during the molding procedure. Great care must be taken to ensure that no air bubbles are trapped at the boundary of the object’s surface and the first coat of molding rubber. The subsequent layers of the rubber serve only to strengthen the primary coat so that the finished mold can only provide as much detail as the first layer of molding rubber. The entire surface of the original specimen must be coated in a thin layer of molding rubber with no air bubbles at the rubber-specimen interface. The silastic rubber and curing agent should be thoroughly mixed immediately before using in the ratio recommended by the manufacturer. After mixing, the rubber should be deaerated to reduce air bubbles that were introduced during mixing. The mixed molding rubber is placed under vacuum (approximately 25mm pressure) until it completely expands and recedes to its original level. Experience will dictate the amount which can be vacuumed at one time. Caution should be exercised to guard against the expanding rubber overflowing its receptacle and being drawn into the hose connecting the vacuum pump and bell jar. We find that large disposable plastic or paper cups are satisfactory for vacuuming the rubber. Avoid Styrofoam cups for this purpose. After the molding rubber has been vacuumed a disposable wooden spatula works well as a tool for spreading the liquid silastic over the original. However, touching the surface of the specimen should be avoided and a stream of compressed air should be used to direct the liquid molding rubber over the surface and especially into small cracks, crevices or undercuts. In the case of small specimens or difficult to reach areas a small, clean paint brush combined with compressed air can be used. Often under field conditions neither a vacuum pump nor an air compressor is available. In this situation, less vigorous stirring and mixing of the molding rubber is advisable to reduce the introduction of air into the mixture. The air which is inevitably introduced into the molding CLAY REGISTRATION TAB S I LASTIC Figure 7. Step 2 — Coronal section. 1989 Bone and Fossil Reproductions 9 Figure 8. Step 3 — Second coat of silastic molding rubber applied and clay retaining walls added. Areas on photograph have been shaded to highlight detail. rubber during the mixing process will coalesce into bubbles which gradually rise to the surface of the liquid rubber. It is therefore necessary to check the liquid rubber which has been placed over the specimen approximately every 20 minutes (for the first hour) in order to remove these bubbles. The number of bubbles can also be reduced by mixing the silastic in a paper cup and allowing the mixture to stand for about ten minutes; a period which allows the air bubbles to rise to the top. A small hole can then be cut in the bottom of the cup and the silastic poured over the specimen through this hole. An additional aid is to use a stream of compressed air which can be created by attaching a hose to the valve of a spare tire which has been previously overinflated. Any bubbles rising to the surface of the silastic can be removed by directing the air across them. Although this may appear to be additional work, special care must be taken to avoid entrapment of bubbles at the part-line or the rubber- specimen interface. Step 3 A second thin coat of silastic is applied after the initial coat has vulcanized (approximately 24 hours) and retaining walls of clay are built around the mold. We do not recommend placing retaining walls around the mold any earlier than this stage. The reason is to permit easy access to all areas of the specimen during the first, crucial coat. Again, the mold and the retaining walls should be angular to assist registration. Step 4 Additional coats of RTV E are applied until approximately 3mm of the material evenly covers the specimen and surrounding clay base. It is important to wait the recommended amount of time between coats as improper curing might otherwise ensue. Importantly, thicker and more layers of silastic are not necessarily better because the thicker the silastic, the less flexible the mold and the more difficult the unmolding process. To avoid creating a rigid, solid block of silastic completely covering the specimen, ground cotton or cabosil can be added to decrease the fluidity of the silastic so that it can be applied to strongly contoured surfaces and/or deep indentations. In addition, surgical gauze can be placed over the external surface of the mold to serve as a strengthening agent. -CLAY WALL SILASTIC Figure 9. Step 3 — Coronal section. 10 Smith and Latimer No. 44 Figure 10. Step 4 — Final coat of silastic molding rubber applied to this half of the mold. Note that gauze has been applied to the specimen at this stage to increase strength of the rubber. Note also that the gauze and rubber follow the natural contours of the original specimen. Figure 12. Step 5 — Completed half of mold. Plaster jacket partially cut away to demonstrate the shape and position of plaster inserts. Note also the addition of silicone rubber registration tabs. A set of small, precut registration tabs of hardened rubber are affixed to the surface of the final coat of uncured silastic to firmly secure the mold to the plaster jacket. These registration tabs should have flat, beveled edges to ensure easy removal of the plaster jacket which is described in the following step. Step 5 A plaster jacket (“mother mold") minimizes distortion in the flexible silastic rubber mold. In the absence of a plaster jacket, any pressure on the highly flexible rubber mold can potentially result in a distortion of the original contours and subsequent error in the cast. The external surface of the rubber mold must be smooth to prevent adherence of the plaster jacket and any roughened area should be covered with a coat of vaseline. In addition, a liberal coating of talcum powder over the silastic will also help the removal of the plaster jacket. Before pouring the plaster for the external jacket, all undercuts and insets (e.g. orbits) are filled with individual GAUZE SILASTIC TAB CLAY WALL Figure 11. Step 4 — Coronal section. 1989 Bone and Fossil Reproductions 11 plaster plugs (Figures 12 and 13). These plugs are liberally covered with vaseline to prevent their bonding to the jacket. Gypsum Hydrocal A- 11 plaster (Appendix 1) is a good, jacketing material. A single layer of fast setting plaster bandage (Appendix 1) applied initially, will strengthen the jacket. The addition of sissal fiber to the liquid plaster can also be used to strengthen and contour the plaster jacket. Fiberglass also makes a rigid yet lightweight material for jacketing. Step 6 The clay walls and base are removed and the half finished mold is inverted. Any clay adhering to the specimen and the surrounding silastic is carefully removed. Check especially for traces near the surface of the part line by gently pulling the silastic away from the specimen in the area of the part line. Having done this, it is necessary to make sure that the silastic edge returns exactly to the established part line and does not pucker against the edge of the specimen. A thin coat of release agent (petroleum jelly dissolved in trichloroethane, CH3CCL3) is then painted over the exposed silastic, care being taken not to wet the original specimen. The petroleum jelly trichloroethane mixture will create the part line by ensuring that the silastic layers will not bond to each other. Care should be taken to coat the entire exposed surface of the silastic or else adherence will occur between the two sides. Figure 14. Step 6 — The mold has been inverted and clay removed. The release agent has been applied to the silastic rubber surface and the specimen is now ready for the next step. Step 7 Steps 2-5 are repeated for the exposed half of the skull. During the casting procedure a mold is normally taped closed, but we have found that it is advisable to incorporate a threaded bolt system into the design of the larger, heavier molds. During the first step of the molding SKULL CLAY BASE CLAY WALL PLASTER / PLUS Figure 13. Step 5 — Coronal section. Plaster jacket shown in its entirety. 12 Smith and Latimer No. 44 procedure, four unthreaded metal rods (equal in diameter to the bolt) should be placed vertically in the clay base, one in each corner. The rods should stay in position throughout Steps 1 - 7 to ensure that the holes for the bolts through the rubber and jacket are properly aligned. The plaster jacket can be reinforced with a piece of hollow, threaded steel lamp pipe placed over the rods during Step 5. Four threaded bolts closed with wing nuts will give excellent registration for larger molds. To unmold the original specimen, the external jacket is first removed. The two rubber halves are then carefully separated as far as the specimen, care being taken not to twist or damage the original. The two halves of the mold are then cautiously peeled away from the specimen taking special care in the areas of foramina and undercuts. Upon removal of the original specimen, the molding process is nearly complete. The final step is cleaning and restoring the original specimen. Wear and tear of the mold surface occurs from the moment the original specimen is removed. Occasionally, small pieces of the molding rubber caught in cracks and crevices may tear away during the unmolding process, emphasizing the importance of good preparation before molding. Torn pieces of the mold can be reattached by gluing with Silicon RTV 700 (Appendix 1 ). These repairs are temporary and have to be repeated each time a cast is poured. The first cast is the most accurate and can be saved as a "record”. Then, if and when a second mold is required, the record cast can be used in place of the original. In fragile and rare fossil specimens it is especially critical to use a “record cast” for remolding in order to keep the original as pristine as possible. We use hardened dental stone (see below) for our record casts because of its dimensional stability. This material, although very hard, is brittle and easily broken. Record casts, therefore, should always be placed in a protected area and not handled. Also, because of the natural fragility of the dental stone casts we usually pour two record casts. A final note concerning record casts is our recommendation not to clean the flash lines off the cast. By leaving the flash, if and when it becomes necessary to remold the specimen, the technician can easily recognize and duplicate the original part-line. Casting Accurate casts can be poured in a variety of materials, including plastics, epoxy compounds and plasters. These materials vary in terms of physical properties, working times and cast detail. Therefore, the casting materials should be chosen on the basis of the number of casts needed and the ultimate use of casts. Because these various materials differentially effect mold life, thought should also be given as to the total number of “pulls” needed. We recommend the use of the following materials because they produce dimensionally stable, highly detailed casts while minimizing mold damage. Casting in Plaster Coecal (Appendix 1) is a dimensionally stable, chip resistant dental stone that gives excellent detail. We have found that mold life using Coecal is excellent; 25-35 “pulls” as opposed to 15-20 “pulls” for plastic casts. The trade off is, of course, the natural fragility of the dental stone relative to the more durable plastics. Prior to the introduction of the liquid dental stone into the mold, a mold dressing (Appendix 1) is painted on the mold's internal surface to reduce surface tension and prevent the 1989 Bone and Fossil. Reproductions 13 Figure 16. Coronal section of completed mold. See text for details. entrapment of air bubbles. Coecal dental stone is mixed with liquid Gypsum Hardener (Appendix 1) at a ratio of 100:34 by weight, in a Whip-Mix machine under vacuum. Gypsum Hardener is used in place of water because it reduces the pore spaces and increases the compressive strength of the dental stone. However, one cautionary note must be added regarding the use of Gypsum Hardener instead of water. The liquid hardener significantly hastens the “setting time” and as a consequence reduces the handling time. Prior to actually filling the mold halves, small indentations and concavities should first be injected using a disposable dental irrigation syringe (Appendix 1). The two parts of the mold are then filled with liquid plaster over a vibrating plate to dislodge trapped air bubbles adhering to the surface of the mold. Just before closing, a little extra plaster should be added to both halves of the mold, which are then held side-by-side one in each palm and “slapped” shut. When first practicing this procedure it is advisable to wear a lab coat. The mold should immediately be taped (using fiber strapping tape; Appendix 1 ) or bolted closed to prevent the loss of the liquid plaster. Importantly, if using a threaded bolt system, the wing nuts should never be secured beyond “finger tightness”. Any further pressure will distort the mold and lead to shearing along the plane of the part-line. In addition, in especially large molds the small amount of expansion that occurs when the plaster hardens can actually crack the mold jacket if it is tightened too securely. To prevent a bubble from coalescing on the surface of the cast, the mold can be rotated by hand for two to three minutes while the plaster hardens. If deemed necessary, extra plaster can be added after the mold is taped shut by gently inserting the nozzle of a syringe between the mold halves and injecting liquid plaster. This latter step is often required if large amounts of liquid plaster are spilled during the closing of the mold halves. Dental stone casts can be solid or hollow. For hollow casts the ratio of dental stone to Gypsum Hardener is increased to 100:36 by weight to increase the handling time and the mold is rotated in the casting machine (Appendix 2). Casting in Plastic Frequent use of epoxies should be avoided as epoxies chemically contaminate silicone rubber reducing mold life. Manufacturer specifications recommend the use of polyesters and polyurethanes as casting materials with RTV E. We have found that a 30:50 mixture (by weight) of two polyurethanes (Dynacast and Master Cast 750; Appendix 1) gives excellent detail and minimizes mold damage (after 15 “hollow” castings most molds show no appreciable damage). However, Master Cast and other polyurethanes are extremely sensitive to moisture and should be mixed in plastic containers with metal or plastic utensils. A thin coat of silicone sprayed over the mold surface will facilitate unmolding and can also help protect the mold rubber from contamination. Dynacast and Master Cast are 2-component thermosetting urethane plastics. Precautions therefore must be taken to prevent the formation of the small air bubbles in 14 Smith and Latimer No. 44 the finished cast, which are the by-product of the heat produced by the chemical reaction in the plastics. “Shelling” the mold surface with a thin coat of liquid plastic using a disposable brush will reduce volume and heat and ensures an even flow of plastic over the entire mold surface. In difficult to reach areas, (e.g. teeth, styloid processes, pterygoid plates) plastic should be injected using a plastic disposable dental irrigation syringe. After approximately an hour, when the first coat of plastic has hardened, the mold should be filled 2/3 full of the liquid plastic, closed, and taped or bolted shut. Again, as with the plaster, extra plastic can be injected into the closed mold using a disposable syringe. The silastic is flexible enough to insert syringe tip gently between the two halves of the mold. Because this mixture of urethanes thickens fairly rapidly (has a working life of less than ten minutes) handling time is an important factor. The filled mold should be wrapped to prevent spillage and fastened onto the rotating machine (Appendix 2) while the plastic casting material is still liquid. Although no longer liquid after 20 minutes the plastic remains malleable and the mold should be rotated for approximately an hour. Hollow casts can be poured in Dynacast alone. Dynacast, however, has a pot life of only five minutes (at 25°°C) and it takes an experienced caster to work under these conditions. Most casts can be unmolded four hours after closing the mold. However, the urethanes often remain somewhat malleable in small casts and should harden for at least six hours before unmolding. We generally leave urethane casts in a well ventilated area (under a fume hood) for 24 hours after unmolding for the completion of the curing process. Finishing Techniques When the casts are fully hardened, the flash along the “part” line is trimmed with a scalpel. A well-registered mold produces casts that have an all but invisible part line. This is especially important when painting, as the part-line takes pigment differently from the surrounding casting material. Small defects or bubbles are filled with plaster, urethane or epoxy putty. Positive bubbles caused by defects in the mold surface can be flicked off with a needle point. Painting can duplicate either the original color(s) or, in a study cast, enhance surface detail. Record casts should always remain unpainted as pigment will obscure surface detail during the remolding process. Painting Plaster Casts Dry artist’s pigment dusted over the cast’s external surface and sealed with an acrylic agent (Appendix 1) highlights detail by filling cracks and fissures. For study casts, a wash of Grumbacher raw umber pigment mixed in acetone will intensify surface detail. Lightly rinsing with water will ensure an even distribution of color over the surface of the cast. A mixture of dry pigments, acrylic paints and magic markers can also be used to duplicate the original colors. Painting Plastic Casts Plastic casts are more difficult to color, owing to the material’s natural impermeability and repellency to liquids. However, liquid and/or dry pigments can be applied to the mold’s internal surface prior to adding the plastic. Upon hardening, the plastic will bond to the pigments, resulting in a permanent finish. This method, although time consuming, gives excellent results. Plastic casts can also be colored by painting the external surface after unmolding. A satisfactory paint is obtained by dissolving acrylic pigments in a liquid matte medium to which an Acrylic Flow Improver (Appendix 1) has been added (2 parts to 20). An acrylic flow improver slows the setting time which is important if an even color is required. Reconstructed areas can be delineated with acrylic or enamel paints. Special urethane pigments are available and can be added while mixing. Difficulties arise, however, in obtaining identical shading if casts or portions of the cast are poured at different times. Conclusion These molding and casting techniques are the result of experimentation over the past 12 years in the Laboratory of Physical Anthropology. The methods described are easily learned and enable the production of economical, dimensionally stable, high resolution casts, that are suitable for display, research and educational purposes. A Appendix 1 Materials The materials listed below are those currently being used in the casting laboratory at the Cleveland Museum of Natural History. This list includes materials that we have chosen through extensive trial and error. Molding Materials and Equipment • Silastic RTV E molding rubber: Dow Corning, Midland, Michigan. • Silicon RTV 700: General Electric Silicone Products, Waterford, New York. • Glyptal Lacquer Cement 1726: General Electric, Schenectady, New York. • Plastalina Leisure Clay #10422: Leisure Craft Company, Compton, California. • CH3CCI3 Trichloroethane: Fisher Scientific Company, Springfield, New Jersey. • Kling rolled surgical gauze: Johnson and Johnson, medical supply store. • Fast setting plaster bandage: Johnson and Johnson, medical supply store. 1989 Bone and Fossil Reproductions 15 5" AUXILLARY DRIVE PULLEY 28 RPM Figure 17. Schematic diagram of the rotating casting machine used in the casting laboratory at The Cleveland Museum of Natural History. See Appendix 2 for details. 5" AUXILLARY DRIVE PULLEY 14" STATIONARY PULLEY 6" MAIN DRIVE PULLEY Figure 18. Diagram of the rotating casting machine, shown in previous figure. See Appendix 2 for details. 16 Smith and Latimer No. 44 • Hydrocal Plaster A- 1 1 : United States Gypsum Company, Chicago, Illinois. Casting Materials and Equipment • Dental Stone (Coecal plaster): Coe Laboratories Inc., Chicago, Illinois. • Plaster Gypsum Hardener: Whip Mix Corporation, Louisville, Kentucky. • Permaflex Concentrate Mold Dressing: Permaflex Mold Company, Columbus, Ohio. • Whip Mix Casting Vacuum Mixer, Model B: Whip Mix Corporation, Louisville, Kentucky. • Dental Irrigation Syringe, Type #8881-411012 Curved Tip: Monoject; Sherwood Medical, St. Louis, Missouri. • Dynacast: Kindt Collins Co., Cleveland, Ohio. • Master Cast 750: Kindt Collins Co., Cleveland, Ohio. • Pure Silicon Spray, Cat. #3302.7: Brookstone Company, Peterborough, New Hampshire. • Tape, Scotch Brand Tape No. 898. • Dry Pigment: Grumbacher Artist Quality Dry Color, art supply store. • Acrylic Agent #1303 Crystal Clear Acrylic Spray Coating: Krylon, art supply store. • Liquid Acrylic Matte Medium: Liquitex, art supply store. • Acrylic Flow Improver: Windsor and Newton, art supply store. Appendix 2 Casting Machine The rotating casting machine permits the production of "hollow” plastic and plaster casts by rotating the molds as the liquid casting material hardens. The machine rotates the mold slowly in three dimensions, allowing the casting material to flow equally on all internal mold surfaces. It is important that the machine not develop sufficient centrifugal force to overcome gravity and that the auxiliary and main shafts not turn at identical RPM values. In our machine we use different sized pulleys producing a speed of 10 RPM for the main shaft and 28 RPM for the auxiliary shaft. The platform upon which the mold is secured is recessed and adjustable so that the center of mass can be concentrated near the rotational axes. Acknowledgements We gratefully acknowledge Donald Johanson, Tim White and William Kimbel for providing the impetus for the casting program; Bill McIntosh for most of the original ideas; and Rob Black and Cindy Luchetti for continual improvements. Special considerations are due to John Aicher who designed and built the rotating casting machine and Bruce Frumker and Toni Hutton for their photographic assistance. We would like to thank Luba Gudz and Kip Dieringer for their excellent illustrations and Anthony Mitri for his diagrammatic representation of the casting machine. We would also like to thank Laurie Linden for her patience for typing various versions of this manuscript. Lastly, we would like to sincerely thank the numerous volunteers whose dedication, hard work and creativity have contributed greatly to the success of the casting program. Development of these casting techniques were in part, accomplished through support provided by the Institute of Museum Services special project supporting grant (IP- 4001 1-84). References Burke, Ann C., Michael Anderson, Alison Weld, and Eugene S. Gaffney. 1983. The Reconstruction and Casting of a Large Extinct Turtle, Meiolania. Curator 26:5-26. Heaton, Malcolm J. 1980. New Advances in Latex Casting Techniques. Curator 23:95-100. Jensen, James A. 1961. A New Casting Medium for Use in Flexible and Rigid Molds. Curator 4:76-89. Madsen, James H. 1974. Derakane Vinyl Ester Resins — An Alternative to Plaster-of-Paris. Curator 17(l):64-75. O'Donnell, Robert, and John H. Hanley. 1983. The “Gourmet” Staining of Plaster Casts. Curator 26(4):307-3 13. Parsons, Kenneth C. 1973. Precision Casting: A New Method in Museum Technology. American Journal of Physical Anthropology 38:789-802. Quinn, James H. 1940. Rubber Molds and Plaster Casts in the Paleontological Laboratory. Technique Series, Field Museum of Natural History, 6. Reser, Peter K. 1981. Precision Casting of Small Fossils: An Update. Curator 24(3 ): 157-1 80. Rigby. J. Keith, and David L. Clark. 1965. Casting and Molding, pp. 389-413 in Handbook of Paleontological Techniques. edited by Bernhard Kummel and David Raup. San Francisco and London: Freeman. Schrimper, George D. 1973. Hollow Casting of Fossil Skulls in Epoxy Plastic. Curator l6(4):286-305. Siveter, David J. 1982. Casts Illustrating Fine Ornament of a Silurian Ostracod. The British Micropalaeontological Society. Fossil and Recent Ostracods 6: 1 05- 1 22. Smith, Jenny, and Bruce Latimer. 1985. The Joy of Casting. The Explorer 27(4): 4-5, 22-24. Smith, J. A. and B. Latimer. 1986. A Method for Making High Quality Three-dimensional Reproductions of Vertebrate Fossils. American Journal of Physical Anthropology 69(2):265-266. Waters, Barbara T., and Donald E. Savage. 1971. Making Duplicates of Small Vertebrate Fossils for Teaching and for Research Collections. Curator 14(2): 1 23- 1 32. KIRTLANDIA The Cleveland Museum of Natural History The Squaw Rockshelter (33CU34): A Stratified Archaic Deposit in Cuyahoga County The Squaw Rockshelter site (33CU34), located along a tertiary tributary to Lake Erie on the edge of the Glaciated Allegheny plateau in northeast Ohio, contained stratified sealed deposits. A small remnant dated 9240±160 B.P., yielded fragments of a human skeleton, unifacial and bifacial scrapers, unstemmed lanceolate points, and a corner-removed indented-base point. Additional lanceolate and bifurcate-base points have eroded from this deposit. This lower cultural level, relatively continuous across the rear of the shelter, represents a drip-line depression washed by water. Within a thin zone overlying superimposed rockfall, a hearth dated to 5500±85 B.P. is associated with a fragment of charred cordage. This level yielded expanded stemmed points, bifacial knives and drills. The occupation surface could not be traced across disturbed areas of the shelter. The early Archaic lithic assemblage included both Plano projectile point styles of the Great Lakes and corner-notched types of the southern Appalachians, while the late Middle Archaic projectile points represent a local blending of styles typical in the riverine midwest and the Mid-Atlantic states. Skeletal remains recovered from the Early Archaic level of the Squaw Rockshelter were identified as belonging to a young female Amerindian. Analyses of the limited fragments and dentition suggest a balanced diet of moderate coarseness. No pathologies or trauma were noted. Two teeth from other individuals were also present. March 1989 Number 44:17-53 David S. Brose Department of Archaeology The Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Abstract Introduction Squaw Rockshelter (33CU34) is a small and partially disturbed site on the southwest bank of Aurora Branch, a second order tributary of the Chagrin River. The site is located in the South Chagrin Reservation of the Cleveland Metropolitan park system, Bentleyville Township, Cuyahoga County, Ohio (Figure 1 ). The discovery of the site occurred in 1 974. 1 Having followed the riverbank 500m upstream from the 1884 carving “Squaw Rock," Robert M. Brose noticed a lanceolate biface of Upper Mercer flint exposed in a silty colluvial deposit veneering the riverbank. Within minutes a bifurcate-base projectile point, also of Upper Mercer 18 Brose No. 44 Figure 1. The location of northeast Ohio sites mentioned in text: 1 — Cooper Hollow: 2 — Burrill Hill Orchard; 3 — Ziegler; 4 — Regis; 5 — Hogue's Spring; 6 — House and Lukens; 7 — Squaw Rockshelter; 8 — McKiben; 9 — Holdson District flint, was recovered from a higher gully in the same sediments by Thomas A. Brose. By the end of that afternoon a small lanceolate point of mottled glacial chert and a second bifurcate-base point of Upper Mercer flint had been collected. The condition of three of the four artifacts suggested a primary depositional context. Permission to investigate was obtained from Mr. Harold Growth II, Cleveland Metroparks Director. That autumn the colluvial fan was traced to a presumed terrace behind several large sandstone blocks, 3.5m higher than the river. Excavation quickly revealed a partially collapsed rockshelter, with deposits protected from the river by massive blocks of sandstone detached from the overhanging cliff 2m south. By the end of the first day only a single unit, 1.0m x 1.25m, had been excavated to a depth of 1.40m. The entire deposit was estimated to include a zone less than 2.25m from front to back, extending along Aurora Branch for 15m. This zone Table 1 Archaeological Excavations at 33CU34 Excavation Unit Surface Area Maximum Depth Actual Volume Excavated 1 1.5m x 1.5m 145cm 2.4m-1 2 1.25m x 1.0m 137cm 1 ,4m3 3 1 ,50m x 2.0m 128cm 2.8m1 4 1 ,50m x 1.75m 95cm 1 .8m3 5 2.0m x 1.75m 153cm 2.5m1 TX 1.50m x 0.50m 85cm 0.6m1 1989 Squaw Rockshelter 19 Figure 2a. View from old quarry on north bank, of large roof fall blocks protecting the Squaw Rockshelter site. contained large blocks of sandstone rising like icebergs through the unconsolidated deposits (Figure 2a, b). The largest area between such obstacles was only 2.5m x 1 ,5m. The Cleveland Metroparks insisted that no excavation be left open; that personnel be limited; and that there be no publicity until work was completed. Due both to previous commitments and current academic concerns, I agreed to personally undertake the fieldwork and analyses, and to curate recovered materials at The Cleveland Museum of Natural History. This arrangement was accepted by Harold Mahan, Director of the Museum in December 1974. Through the 1975 summer 13m2 in four units and a single exploratory trench were excavated into the unconsolidated deposits at the Squaw Rockshelter site (Table I, Figure 3). It is estimated that the 11.5m3 excavated represent 85% of the site preserved from erosion. This is a small and somewhat biased sample of what may have been 40m3 of sediment. Nevertheless, these deposits offered evidence to support a different interpretation of the Archaic of northeast Ohio than existed prior to 1975. Cultural and Historical Background Early descriptions of Ohio antiquities concentrated upon earthworks and the artifacts they contained (Brose Figure 2b. View to west in Squaw Rockshelter, figures on surface of Unit 3 before excavation. 1973). Stone tools that appeared more primitive, and badly decayed human bone from river banks, wells, and mines were compared to materials recovered from purported Pleistocene deposits in other states. Most such reports were ignored, since human occupation of the New World was not considered to be of great antiquity (cf. Fowke 1902), and those sites subjected to detailed scrutiny turned out to be of dubious antiquity (Holmes 1919). The association of artifacts with extinct fauna not only demonstrated respectable antiquity for western Amerindians, but revealed projectile points with analogues throughout North America. Shetrone (1936) noted the frequency of such points in Ohio, but lack of any stratigraphy limited temporal interpretation. Subsequent studies (Tuck 1978) have revealed eastern Paleolndian sites with a variety of projectile point forms. Later Paleolndian lanceolate point types documented from the plains were also recovered in the Great Lakes-Ohio Valley, and geochronological correlations suggested considerable age for these types in the midwest (Mayer-Oakes 1955; Mason 1958). There seemed good reason to regard the east as having been occupied as early as the west (Mason 1962). Highland rockshelters and deep stratified sites in floodplains yielded fluted points as well as partially fluted. 20 Brose No. 44 Figure 3. Excavations at Squaw Rockshelter site. unfluted, and notched projectile points showing a transition in lithic technology from the Paleolndian period into the early Middle Archaic period (Coe 1964; DeJarnette et al. 1969; Griffin 1974; Chapman 1977). These sites have been taken to indicate a wide-spread, and synchronous series of changing projectile point styles with little or no overlap of differing types at any component or at any given time over nearly 1500 air miles from Maine to Mississippi (Broyles 197 1 ; Dincauze 1975; Bense et al. 1983). It was a situation unparalled by the Early Archaic in the Midwest. From the Dakotas into western Ohio, Paleolndian points precede development of the prairies. With the Hypsithermal, both prairies and gallery forest zones yield assemblages of stemmed and unstemmed lanceolate and partially fluted projectile points. By 9500 B.P., as the prairie fingered eastward, the grasslands yield assemblages in which lanceolate projectile points predominate, while forested areas yield assemblages in which corner-notched and side-notched projectile points predominate. Short-duration components at ecotones from 9500 B.P. to 6500 B.P. lithic assemblages contain a variety of stemmed and unstemmed lanceolate projectile points as well as a variety of expanded stemmed and/or corner- removed, and side-notched and corner-notched projectile point types. While the stemmed and unstemmed lanceolate projectile points from the Great Lakes suggested relationships to the Plano complex further west (cf. Mason 1981:114-126), they were also compared to Late Archaic points, from 5000 B.P. to 3500 B.P. further west, or to similar projectile points in the St. Lawrence or Atlantic drainages of Pennsylvania and New York, dated to the end of the Middle Archaic (cf. Dragoo 1959). Those who argued a late date for the stemmed and lanceolate points in the Great Lakes claimed that the accompanying corner- removed or notched points related to Normanskill points. 1989 Squaw Rockshelter 21 Brewerton points. Otter Creek points or even Snook Kill points, dated between 6500 B.P. and 3500 B.P. in the east. Those who viewed the Great Lakes lanceolates as early related the corner-removed or notched points to Thebes, Hardin barbed, Graham Cave side-notched, Kirk corner- notched, and St. Albans side-notched types, dated between 9500 B.P. and 6500 B.P. in river drainages to the south and west (Brose 1975). The absence of components that yielded only a single clearly identified point type suggested to some that there were no Early or Middle Archaic occupants of the region at all. Assuming a Paleolndian focal adaptation to tundra- edge hunting. Fitting (1968, 1970) argued for a post- Paleolndian depopulation of the Great Lakes, suggesting the 9000 B.P. rise from low levels created a lakeshore unfavorable to human occupation. Those few possible large Early and Middle Archaic sites, he believed, were now drowned, having been located along waterways, as in the southeast, and only smaller interior hunting camps still existed. This ignored controversy about the reality of any such pattern in the south. Mason however, argued that the 4500 years between Paleolndian and Archaic economy and technology represent ...gradual transition from one dominant culture type to another.. .in the Great Lakes, enough is known to encourage the view that the perception of two partly coetaneous cultural traditions is accurate. (1981:114-115) Indeed, Ellis and Deller ( 1986:56-57) speculated that in southern Ontario and lower Michigan Early Archaic Kirk and Plano types would co-occur by 8900 B.P, while Middle Archaic side-notched forms should resemble Godar, Brewerton-like Thebes, and Otter Creek by 4500 B.P. (cf. Lovis and Robertson 1985). Candidates for the 9500-6500 B.P. period in the Great Lakes, thus included both lanceolate projectile points, and the corner-removed, stemmed, bifurcated base, and notched projectile point clusters dated between 10,000 B.P. and 6500 B.P. to the west, and dated between 9500 B.P. and 3500 B.P. in areas east of Ohio. Ohio remained an enigma. In the upper Ohio Valley Mayer-Oakes (1955) had assigned stemmed and lanceolate projectile points to an early Archaic horizon, while Dragoo assigned a similar lithic assemblage (which included corner-removed/side- notched points) to a “proto-Laurentian culture” with at least one radiocarbon date as late as 53 1 0± 1 80 B.P. (Dragoo 1959:238-239). Nonetheless, the lowest levels of the Rohr shelter, with Steubenville lanceolate points, yielded a Dalton point while in the Allegheny River valley lanceolate-free Brewerton components date to 5800 B.P. (Caulkin and Miller 1977) and to 6090 B.P. (George and Davis 1986). As in the upper Ohio Valley, the lanceolate points of northern Ohio have been called Early Archaic, Late Archaic, or both. Few components reported dated between 9500 B.P. and 3500 B.P. The first serious attempt to address the Paleolndian and early Archaic occupation in northern Ohio, was initiated by Olaf Prufer ( i 96 1 ). Based upon survey by Prufer and his associates between 1958 and 1963, a study of the distribution of various types across Ohio was published (Prufer and Baby 1963). Given the nature of the sampling, chronological typology was based upon distant analogues. Not all of these have received radiometric support, and the relationships between the survey sample, sampling methods and sampling framework have been a source of contention (Seeman and Prufer 1982; Lepper 1983). Beyond stimulating research, these studies illustrated the variability and standardized the terminology for early Ohio projectile points. Although recognizing their differing temporal positions, the Plano points in northwestern Ohio and the fluted points of southern Ohio were attributed to contemporary immigrations from the upper Great Lakes and the Appalachians. After discussing specific sites Prufer and Baby found ...no evidence permitting the linkage between the Paleolndian assemblages of Ohio and established local Archaic complexes. (1963; emphasis added) The problem was that in 1963 there were no established local Archaic complexes other than mortuary aspects of the Late Archaic. Although Mayer-Oakes suggested continuity, Prufer and Baby agreed with Dragoo that there were no Archaic predecessors in any Ohio late Paleolndian site assemblage. They accepted Dragoo’s chronology, and his implication that no lanceolate projectile point in eastern Ohio was early. Geistweit argued that Kirk and LeCroy points had been recovered in all regions of Ohio save the northeast (1970:1-32, figs. 9-16). She too saw no evidence for Paleolndian Early Archaic continuity, and offered no explanation for a proposed rapid introduction of Early Archaic as a style horizon or as an economic pattern (1970:162-164). Claiming dense Late Archaic occupation of those same regions (1970:45-88) Geistweit suggested that the Middle Archaic in Ohio was represented by a “...continuation of Early Archaic tool types and way of life...” (1970:44, 164) The occupational hiatus was between 5000 B.C. and 3000 B.C. for Geistweit, who found little local precedent for regional Late Archaic variants, especially along the Ohio River valley. Blank (1970) synthesized data from collections in northern and west central Ohio with excavations in east central Ohio. Despite the paleoecological data with which he sought to support his model, every surface collection was selective, and other than at quarries all sites showed a 22 Brose No. 44 mixture of lanceolate, stemmed and notched point types. No site yielded floral or faunal remains, and there was confusion of single site activity with economic adaptation. Placing the late Paleolndian and Early Archaic periods between 8000 and 5000 B.C., Blank suggested that lanceolate points characterized northwestern Ohio groups hunting elk and moose in the till and lake plain swamps. The notched points of southeastern Ohio were used by economically diffuse hunting and gathering groups (1970:363). While coeval. Blank saw little cultural inter- action, suggesting seasonally exclusive occupations of sites where both appeared. He concluded that the Kirk-like complex of southeast Ohio was initiated by and continued to receive direct influences from the Mid-South (cf. 1970:366). The resultant Middle Archaic cultures moved into northwestern Ohio following its abandonment by inflexible Plano peoples. For Blank, although Early-Late Archaic continuity anywhere in Ohio was problematic, in southeast Ohio there was no occupational hiatus. Rather ...this apparent gap in cultural chronology results. ..from our inability to recognize other than a limited number of the cultural elements. (1970:355-6) Accepting Blank's thesis. Fitting had argued for a 4,500 year abandonment of northern Ohio until the appearance of a Late Archaic adaptation between 3000 B.C. and 2000 B.C. with the establishment of modern forests. Neither archaeological nor paleoecological data offer support for such a model. Fitting had assumed that the settlement types to be found in northern Ohio fit his Michigan model. Yet, as early as 1966 (cf. Brose and Essenpreis 1973) I argued that the drop of western Lake Erie between 12,000 and 9500 B.P. so altered the relationships of shoreline-interior resource availability that no Great Lakes analogy would apply. Prufer and Long reviewing the northeastern Ohio Archaic, now suggest that, ...it would be wiser to study.. .the local Archaic as a fluid continuum rather than along the usual tripartite lines of fixed Early, Middle and Late units. This is so because, throughout there seem to have been no variations in life-style patterns from the beginning to the end of the Archaic, at least through the Laurentian tradition. (1986:50) Certainly that approach seems required for sites such as McKibben, House or Lukens Hill, intermittently occupied from 7000 B.C. to 3000 B.C., with all diagnostic artifacts in plow-disturbed soils (Prufer and Long 1986:6-10, Table 1 1 ). In the absence of intra-site spatial analyses, functional morphological and use-wear studies, or floral or faunal recovery, it is difficult to see how Prufer and Long could escape the impressions that the Archaic artifact inventory was rather static (1986:25) and that the nature of activities carried out at Archaic sites in differing topographic locations could not have been very different (1986:26). However, their tables showing relative frequencies of chipped stone scrapers, bifaces and debitage; ground stone tools, notched and stemmed points; indications of chert heat pre-treatment; and breakage patterns all suggest significant technological differences between these sites. Of course, the extent to which differences may be attributed to site function, as opposed to period(s) of occupation, could not be evaluated from these mixed components. Nonetheless, Prufer and Long saw no discontinuity in Archaic occupation but rather a low population density with a conservative cultural tradition. The sealed and/or stratified sites which might have verified their conjecture remained unpublished, and their report has spurred me to make some known. Environmental Setting Regional Paleoecology South of Lake Erie a zone of proglacial beach formations narrows from Michigan disappearing at Niagara. These deposits lap flat till plains in the west, while in the east they abut the Allegheny Plateau. With the 12,500 B.P. drop of Lake Erie the islands which define the lake’s western basin were dry, with a conjoined Thames- Maumee-Sandusky-Cuyahoga River cutting to a shrunken basin off Erie, Pennsylvania (Coakley and Lewis 1985). Tributaries, below former base levels, must have been quite dynamic. The lake effect, which currently buffers seasonal differences between shoreline and interior, could not have existed in the Early and early Middle Archaic. At 9500 B.P. uplift began refilling the western basins, creating biotically rich habitats like those of the 19th century Black Swamp. During much of the Archaic period vegetation represented local mosaics on uplands (now islands). Altered groundwater and effective climate resulted in concentric floristic zones with extensive interfluvial grassland/oak openings with gallery forests. Through time along the lower course of the major rivers there was succession from beech-maple; to mixed-maple; mixed mesophytic; and finally, mixed oak-hickory and elm-ash as lake levels rose. South of this first zone, old lake beds extend around the old Maumee estuary, thence northeastward with decreasing width to merge with beach formations and disappear west of the Cuyahoga. The zone is marked by rolling topography with numerous kettle lakes and bogs. Most rivers and streams are immature with stable channels and limited biotic diversity, although there must have been changes with the short-term lake fluctuations between 9500 and 6500 B.P. 1989 Squaw Rockshelter 23 The Glaciated Allegheny Plateau begins south of the lake along the East fork of the Rocky River. The escarpment gradually approaches the modern shoreline, abutting the Lake at Dunkirk, New York. On the plateau (increasingly so to the south) streams slow where they encounter more resistant strata, creating variable valley morphology along short segments, and major differences in tributary frequency and catchments between adjacent systems. Due to altitude, seasonal precipitation differs significantly from other zones. Although exposure creates edaphic communities, flora generally is beech-maple forest on the rolling interfluves, with elm-ash or hemlock valley facies, and mixed tulip-oak- chestnut or oak-hickory-butternut facies on ridges (Williams 1940; Gordon 1969). Local studies (Potter 1947; Ogden 1966, 1967; Shane 1975) and regional syntheses (Webb, Cushing and Wright 1983; Davis 1983; Hollaway and Bryant 1985) suggest two Holocene periods of major floral change in northeast Ohio. Between 10,000 and 9500 B.P. there was a rapid shift from the short-lived Hudsonian pine/oak forests to the mixed Canadian hardwood/deciduous forests of the Great Lakes; and between 6500 and 4500 B.P. a number of typically 24 Brose No. 44 Carolinian species (such as chestnut, walnut, and rhododendron) appeared in the region. The palynological evidence for an early appearance of modern forests, and for the lack of any Hypsithermal prairie development in northeast Ohio, is fully corroborated by faunal evidence indicating a turnover in the period between 11,300 and 9300 B.P. to communities which are essentially modern, with little significant change due to Hypsithermal wanning (Semken 1983:193ff). Holocene Geomorphology Squaw Rockshelter lies at the southern edge of the Defiance moraine (Goldthwait, White and Forsyth 1967). It sits on the west bank of Aurora Branch, a tributary which joins the Chagrin River 2.5km below the Falls, and 3km above the junction with Griswold Creek (Figure 4). The upper Chagrin above the Falls, and the ower Chagrin below Griswold Creek, flow through large valleys carved into Paleozoic rock, and now filled with fluvio-glacial and lacustrine deposits (Winslow, White, and Webber 1953:41- 43). Through these two segments which originally drained south (but not in the intervening section) the river now flows southwest, then north. Earlier Pleistocene drainage (Rau 1969:10-14) was probably controlled by the Ancient Chagrin River whose buried valley, at an elevation between 400' and 500' above mean sea level (AMSL), flowed southward up Aurora Branch (Winslow, White and Webber 1953:41-43, PI. 3). During the Wisconsinan stadial, ice overrode the region at least four times. The Woodfordian advance was marked by the Defiance moraine, about 19,000 B.P. This moraine crosses the lower Chagrin valley 2km west of the Falls, just north of the junction with Aurora Branch. The Defiance moraine was subsequently overlain by a veneer of Hiram outwash and till (cf. White 1982). With the melting of Hiram ice, after 14,000 B.P. (White and Totten 1982:48; 1988:66-68), the upper Chagrin River valley drained southwest to the upper Cuyahoga (Wittine 1970). Its channel lay 1500m to the east of present Aurora Branch. Aurora Branch, above the riffles at Solon Road, now flows north-northwest through a buried valley which, during the Altonian interstadial, formed a portion of a stream system which included Griswold Creek and flowed southward into the Upper Cuyahoga, thence south across the present Akron divide to the Tuscarawas River. That portion of Aurora Branch below Solon Road, and the section of the Chagrin River between the Falls and Griswold Creek, were abandoned during much of the late Pleistocene. Reoccupation of the present Aurora Branch valley, with reversed (north flowing) stream entrenchment and erosion of the Hiram Till which filled it, began when headward erosion of the north flowing lower Chagrin River downcut through the Defiance moraine north of the site and captured the Upper Chagrin River (Feldmann, Coogan and Heimlich 1977:192-193). The hydrostatic event which triggered this erosion was the drop to lowest Lake Erie levels, around 12,500 B.P. (Lewis 1969; Forsyth 1973; Coakley and Lewis 1985). Aurora Branch appears to have downcut to a resistant high of Sharon Conglomerate, and then shifted 1.5km west to reoccupy its present channel during the subsequent low-water period (Feldmann, Coogan and Heimlich 1977:107-142; John Hall, personal communication, 1983-1985). With a rapid rise in the level of Lake Erie to 495' AMSL, fluvial erosion decelerated by 10,000 B.P., and ended between 7500 and 6000 B.P, with Lake Erie at a stillstand between 538' and 554' AMSL. There appears to have been virtually no downcutting between 5000 and 4300 B.P. when lake levels rose rapidly to ca. 590' AMSL (Coakley and Lewis 1985:198-200, 208-210). After 3900 B.P. there was again a drop to a level controlled by the Niagara Sill at 555' AMSL. With crustal upwarping after 3000 B.P, there was a slow rise to the present level of 572.5' AMSL (Calkin 1970:Table 2). At present the bed of Aurora Branch, where it joins the Chagrin River 2km downstream from the Squaw Rockshelter, is composed of nearly 200' of alluvial deposits with surface elevation of 830' AMSL. Some 20m upstream from Squaw Rockshelter the floor of Aurora Branch is exposed on the upper surface of Bedford Shale (Prosser 1912:529) at an elevation of 854' Figure 5. Berea Sandstone roof at Squaw Rockshelter 1989 Squaw Rockshelter 25 Figure 6. Stratigraphic profile across Aurora Branch at the Squaw Rockshelter site. AMSL. Thus, retrenchment of the 3km section of Aurora Branch in which the site is located, was rather rapid. It began after ca. 12,000 B.P. when Lake Erie stood below 440' AMSL. The surface of the old buried southflowing stream in upper Aurora Branch crosses Orange and Solon townships at an elevation just under 500' AMSL (Winslow, White and Webber 1953:42) and thus could not have been tributary to a stream whose surface stood at 500' AMSL, the lowest floor for sediments within the valley of the present Aurora Branch-Chagrin River junction (Rau 1969:figure 2, II; 1, 12, 31-32). That suggests that headward downcutting in the relatively recently captured northflowing Aurora Branch-Chagrin drainage ended before 10,000 B.R when a relatively rapid rise brought lake Erie to a stand at 523' AMSL (cf. Lewis 1969:276; Coakley and Lewis 1985). While lateral erosion of this segment of Aurora Branch undoubtedly occurred after 10,000 B.P. it must have been at a slower rate than during previous millenia. Indeed, the lower Aurora Branch valley may have been partially re-filled between 4300 B.P. and 3900 B.P. as base level seems to have been higher than present (ibid.). Bedrock Lithology Squaw Rockshelter formed beneath an overhang of Berea Sandstone, a thick, variably bedded Mississippian sandstone, lying unconformably upon thinly bedded late Devonian shale of the Bedford Formation (Newberry 1870; Bownocker 1915; Prosser 1912). The Berea Sandstone is overlain by thin siltstones and shales of the Sunbury and Aurora submembers of the Orangeville Member of the Mississippian Cuyahoga Formation (Coogan et al. 1986). At the site the section forms one of the classic exposures which led to considering Berea Sandstone as a massive deltaic channel fill in the underlying Bedford mudflat shales (Prosser 1912:530-534; Pepper et al. 1954:204, pi 5; cf. Coogan et al. 1986; however, see Lewis 1988 for an alternative interpretation) (Figure 5). About 200m southwest of the shelter. Deer Lick Run flows north from a small kettle lake at 943' AMSL, cascading into Aurora Branch 40m downstream of the Shelter. The Orangeville Member here is an 8' to 10' (2.5- 3m) thick stratum of thinly bedded, soft argillaceous and grayish-black bituminous shales, interbedded with the layers of blue fine-grained sandstone and siltstones of the Aurora submember. These lie upon Berea Sandstone. The Berea Sandstone at the site consists of three layers with a total thickness of 42' to 64' ( 12.7 to 19m). The upper 18' (5.5m) are flat, thinly bedded, ripple marked sandstone composed of medium to medium fine sand grains. Below these lie two or three distinguishable strata of cross-bedded sandstones, with a total thickness of 22' (6.7m). These are composed of medium to coarse sand grains. Below these is a massive stratum of festoon cross- bedded sandstone (Coogan et al. 1986:7) composed of 26 Brose No. 44 medium to medium-fine sand grains, and having a thickness of 23' (7m). This stratum occupies a “channel” deeply cut into the bluish sandy shale of the Bedford Formation (Figure 6). Across Aurora Branch, Berea Sandstone was quarried by the Independent Stone Company beginning somewhere about 1900 (Bownocker 1915). Prior to that the valley was about 35m wide at the top (cf. Newberry 1870). Stratigraphy and Sedimentology Analyses Sedimentological studies were carried out during and after the excavation of Squaw Rockshelter (Appendix). Because Holocene processes operated upon formations with facies displaying considerable variation in short distances, results are less than ideal for understanding deposition during the occupation of the site (cf. Farrand 1985). Under similar circumstances (whatever the actual antiquity of the site) the quartz sand grains at Meadowcroft Rockshelter reflect the primary Paleozoic depositional processes, not the secondary Holocene processes of archaeological interest. Only with great caution ( contra Donohue 1976) could shelter evolution be reconstructed from studies of sediment derived with little transport from variable sandstones such as the Berea Sandstone at Squaw Rockshelter or the Connellsville Sandstone at Meadowcroft. At Squaw Rockshelter profile inspection of in situ sediments and subsequent analyses distinguished five strata (Appendix). Stratum 1 This lowest unconsolidated post-incision fill of the shelter represents water-sorted Hiram till, deposited along the undercut wall of Aurora Branch. Samples A1 and B1 both consisted of poorly sorted and poorly consolidated lenses of dark yellowish brown (10YR 4/6) to greenish gray (5GY 5/1) fine sandy loam and silt loam with sub- angular flat lying cobbles of greenish-gray to dark bluish- gray shale, and small angular to sub-angular cobbles of reddish-yellow to dark grayish-brown stratified sandstone. These samples lay conformably upon and graded into the underlying sediment (Ao), at depths below datum of 140cm and 135cm, respectively. There is a relative decrease in very fine silt to clay-sized particles to the rear of the deposit, suggesting some low energy groundwater movement along the rear wall. No bedding differences could be noted across what seems to be this same level in those excavation units where it could be identified (XU 1 , 5, 3, 5, and the excavation at the north end of the deposit), although there were fewer cobbles and finer sediments in higher levels. Stratum 2 These deposits, encountered at 110cm and 125cm below datum consisted of a very compact lens of dark grayish brown (10YR4/2) to dark strong brown (4.5YR 4/6) medium to coarse sandy loam flecked with granules of charcoal, and lying conformably upon the underlying silt loam and inclusive cobbles. This lens, which ranged from 10cm in thickness at the front of excavation Unit 3, to 20cm in thickness at the back of Excavation Unit 5, was virtually free of inclusive cobbles. As identified lithologically it occurred in all areas of the site except for the eastern halves of Excavation Units 2 and 4 and the excavation where recent erosion appears to have removed it. Within Excavation Units 1, 3, and 5, where continuous stratigraphic exposures could be observed, the upper 5 to 10cm of the deposit contained small areas with thin laminae of yellowish brown (10YR 5/4) or light grayish brown ( 10YR 2/4) fine silts and clays. These laminae were more common to the south. Horizontally lying flint tools and chipping debris were encountered throughout this lens. Within the lowest and thickest portion of this deposit running about 160cm along the southern wall of Excavation Units 1 and 5, 115cm below the datum, human skeletal remains were encountered within a thin zone (Feature 2) 30cm wide. This stratum appears to represent at least the basal portions of an anthropic epipedon. The relative removal of silts and clays, the horizontal placement of larger (cultural) particles, the few finer laminar facies within this deposit, which parallel the overall westward dip toward the rear wall and strike to the north, all suggest that this level represents some portion of a surface in part deposited by gravity, minimally altered by prehistoric occupation, and then significantly reworked by low energy water transport. In Excavation Unit 4, at a depth below site datum of 120cm, a concentrated lens of oak charcoal associated with the midsection and tip of a lanceolate point, yielded a radiocarbon date of 9480±160 B.R (DIC-586). This stratum is called Cultural Level II. Stratum 3 This is an unconsolidated zone of numerous large angular cobbles and boulders of cross-bedded Berea Sandstone within a matrix of fine to very fine sands. Encountered in all excavation units between 95cm and 110cm below datum, this stratum was about 45cm thick at the front of the shelter and about 65cm deep along the rear wall and lay conformably upon Stratum 2. Stratum 3 appears to represent in place chemical weathering of material derived by mechanical processes from the massive Berea Sandstone “channel fill” which forms the present back shelter wall and which must have formed the shelter overhang during the deposition of Stratum 2. The massive blocks of cross-bedded Berea Sandstone which presently lie between these unconsolidated deposits and the Aurora Branch, seem to lie directly upon Stratum 2. Where recent rodent disturbance has not confused subsequent depositional sequences (in XU 1 and 3), Stratum 3 appears to have been deposited between these blocks and the rear wall of the shelter. 1989 Squaw Rockshelter 27 Stratum 4 This deposit lies conformably on the surface of Stratum 3. It is a 15cm to 25cm deep zone of yellowish brown (10YR 5/4) medium to coarse loamy sands including relatively few pebbles or small cobbles. Large cobbles of massive unbedded Berea Sandstone occurred throughout the stratum, however, and while portions of stratum 4 were encountered at depths between 55cm and 70cm below datum in every excavation unit except the northern test trench, it was not possible to trace this deposit continuously for distances greater than 80cm in any direction. Indeed the correlation of Stratum 4 in excavation units 1, 5, and 3 and in Excavation Units 2 and 4 is a conjecture largely based upon similar granulometric analysis and relative stratigraphy. Although no vertical sorting or pedogenic structures of any sort could be identified in Stratum 4, there was some clear front-to- back sorting with coarser sands at the front. Several very shallow cultural features were encountered within this stratum, and their origins varied in surface elevation by as much as 20cm. The most unambiguous of these. Feature 3, encountered at a depth of 70cm below datum in XU 2, consisted of a shallow oblate depressed area of fire reddened (2.5YR 3/5) silty clays about 25cm by 20cm in surface area and about 5cm thick. This ‘hearth’, located about 45cm from the rear shelter wall, was about 10cm deeper in the center than at the periphery. The base of the center rested on a block of sandstone buried in Stratum 3. Within this depressed central area a matrix of dark yellowish brown (10YR 5/4) to dark grayish brown (10 YR 4/2) silty loam contained small flakes of chert, charcoal identified as maple, and a small charred fragment of walnut hull. The wood charcoal fragments yielded a radiocarbon date of 5500±85 B.R (DIC-321). Between this feature and the rear wall, at slightly shallower depths, were two discontinuous horizontal areas of more compact and darker ( 10YR 3/2) silty sands, each about 10cm thick. These zones yielded several lithic tools and flakes and a single fragment of charred cordage. A second possible hearth. Feature 1, was encountered at the northern edge of Excavation Unit 1 between 75cm and 85cm below datum. No fire-reddened subsoil was noted, but a lens of sandy ash (pH 6.0) filled the central portion of this circular clayey lens, some 30cm across. Between the northern periphery of Feature 1 and the block of fallen sandstone fronting the river, were loose sandy silts riddled with recent rodent burrows. The feature contained no organic material, although an expanding stemmed projectile point was recovered from the compact silts at the southeastern edge of the feature. Several other formal lithic tools and as well as lithic debitage were recovered from this stratum in excavation units 2, 3. and 5, but no other cultural features and no other areas of apparent concentration were found. Stratum 4 is called Cultural Level 1. Stratum 5 This stratum rests unconformably upon the surface of Stratum 4 with a clear contact. Many of the upper surfaces of the large cobbles and blocks of Stratum 4 (and several of those which obtrude from Stratum 3 through Stratum 4) are incorporated into the lower portions of this 40cm to 60cm thick zone of dark grayish brown (10YR 4/2) to dark yellowish brown (10YR 3/2) sandy loam and sandy clay loam. Large cobbles of Berea Sandstone and pebbles and small cobbles of siltstone from the overlying Orangeville member are entirely incorporated in this matrix. At the time of excavation the surface of this stratum in Excavation Unit 2, 3, and 5 was capped by a 20cm to 35cm thick zone of decomposing organic material (leaf litter, dead rodents, cigarette filters) and some amount of very recent inorganic trash. Humic acid staining from this Ao horizon had penetrated to variable depths within the upper 20cm of Stratum 5, but no soil development had occurred (cf. U.S.D.A. 1977). During the accumulation of at least the lower portion of Stratum 5 sediments apparently weathered from some portions of the Orangeville Member and perhaps even some weathered from Hiram Till were being introduced to the front of the shelter by a higher energy transport medium. This situation seems to have ceased during the period in which the upper portions of the stratum were deposited. While unconfirmed by stratigraphic excavation, it is possible that the large foreset-bedded Berea Sandstone block presently forming the southeast wall of Excavation Unit 4 fell onto lower Stratum 5 during this depositional episode. Unfortunately, the area along that face of the Excavation Unit 4 block and the entire stratigraphic sequence above Stratum 3 were represented by rodent disturbed deposits (Sample X). All cultural materials encountered within Stratum 5, whether prehistoric or recent, were recovered from such rodent-disturbed areas in Excavation Units 1, 3, and 4. The presence of mid-20th century material in these areas suggests that their absence on the surface of Stratum 5 is due to Metropark cleaning efforts since the 1960’s. Summary Most non-cultural sediment within the site is attributed to the in-place deposition of sands and silts chemically and mechanically weathered from bedrock. There was sheet runoff in portions of the shelter, and with the exception of a short period around 5500 B.R the net effect of water transport after 10,000 B.R appears to have been erosive. Sedimentological analysis and the vertical and horizontal stratigraphy within and downstream from the shelter suggest three relatively distinct episodes of human occupation of variable intensity and importance. Squaw Rockshelter appears to have been relatively deep when occupied about 9200 B.R The shelter was more exposed when occupied around 5500 B.R Finally, the r r 28 Brose No. 44 v ' N .kH *. Figure 7. Excavation of Unit 5 showing limits of space. barely protected surface of shelter deposits has been used intermittently during past decades. Evidence for anthropic episol formation appears only in discontinuous portions of Cultural Level II, and at the top of Stratum 5 postdating aboriginal use. Archaeological Recovery Methods Rockfall made shovelling impossible within the confines of excavation units (Figure 7). All sediments were removed by hand and trowel. When a distinct stratum surface was encountered the contact was cleaned across the floor of the excavation unit. Large fallen blocks often made this procedure difficult to follow. After their non-cultural origins had been confirmed, there was no screening of strata 5, 3, or 1. Although all feature matrix and fill was screened through 3/32" mesh, fine screening of cultural “floors” was not systematic. No flotation samples per se were collected at the site, but flotation was undertaken on each soil sample collected. These methods yielded fragmentary and complete mollusk shells and minute fragments of wood charcoal. No pollen was preserved in any of the sediments inspected, and no attempt to identify phytoliths was made. Overall, fine screening and flotation yielded little independent data on the proximal environment of Squaw Rockshelter when Cultural Level II (9480± 1 60 B.P.) and Cultural Level I (5500 + 85 B.P.) sediments accumulated. Mollusks Analysis of recovered mollusks was performed by Brose in 1984. A total of 36 identifiable gastropods and 61 Table 2 Identifiable Molluscs Recovered from Squaw Rockshelter Cultural Level I Cultural Level II Aquatic Gastropods N (% class) N (% class) Valvata tricarinata (Say) 2 (15.4) • Stagnicola sp. 6(46) 1 (50) S. reflexa (Binney) 3(23) • Helisoma sp. 2(15.4) 1 (50) Subtotal 13 (99.8) 2(100) Terrestrial Gastropods Stenotrema fraternum 4 (17.4) 3 (33.3) Triodopsis alholahris 8 (34.8) 1 (H.l) Zonitoides arboreus 6 (26.1) 1 (ll.D Discus cronkhitie 5 (21.7) 4 (44.4) Subtotal 23(100) 9(99.9) TOTAL 36 11 1989 Squaw Rockshelter 29 Table 3 Charcoal Fragments Recovered from Squaw Rockshelter Cultural Level I Cultural Level SI Genus N (% ID) N (% ID) Pinus 5 (9.6) 10(23.3) Tsuga 16(30.8) 8 (18.6) Acer 10(19.2) 8 (18.6) Quergus 5 (9.6) 17 (39.5) Fagus 8 (15.4) 0(*) Ulnuis 5 (9.6) 0(0 Jugulans 3* (5.8) 0(0 Total Identified 52(100) 43 (99) Unidentifiable 21 31 TOTAL 73 74 * 2 fragments of charcoal and 1 charred nut shell fragment unidentifiable fragments was recovered from Cultural Level I, while 1 1 identifiable gastropods and 27 unidentifiable fragments were recovered from Cultural Level II sediments (Table 2). No bivalves were recovered, suggesting that all mollusks present were due to non-cultural factors. Even standardized for the relative volumes of sediment analyzed from Cultural Level II (.94m3) and Cultural Level I (1.75m3) it is clear that there was a somewhat greater frequency of gastropods of all types in Cultural Level I than in Level II, and there was a significantly greater frequency of aquatic gastropods in Level II. There are few environmental differences among the recovered aquatic gastropods. All are extant and have inhabited the region since the mid-Pleistocene, living on emergent vegetation in a wide range of energy environments (LaRoque 1968:367-506). The relative paucity of Cultural Level I aquatic gastropods is consistent with sediment analyses suggesting little flooding of the shelter during its early history. Differences in overall frequencies of the terrestrial gastropods from the two cultural levels reflect both the relative loss of shelter overhang during the millennia between depositional events, and the greater duration of Level I sediment accumulation. Differences suggested by the habitat preferences of the 47 terrestrial gastropods must be tempered by small sample sizes, the unsystematic recovery strategy, and by the assumption that for over 4,000 years the shelter represented a varying taphonomic catchment. Ignoring these caveats, the shift from an assemblage of Stenotrema fra termini and Discus cronkhitei to an assemblage dominated by Triodopsis albolabris and Zontoides arboreus reflects the change from cool damp hardwood/conifer Canadian forests to forests in which beech-maple and mixed Carolinian mesophytic hardwoods predominated. However, all of these species are extant locally and have been present since pre- Wisconsinan time (LaRoque 1968:570-680). Charcoal Fragments of charcoal and charred nut hull (Table 3) recovered from Squaw Rockshelter were identified in 1976 by Dr. F. DiMarinis, Cleveland State University Biology Department. None of the uncharred seeds or nut hulls recovered from potentially rodent disturbed soils (x) were included in these analyses. Assuming these specimens have been humanly introduced to the shelter, each assemblage represents a mixture of local availability and cultural selectivity. The absence of walnut and beech in level II is predictable, since neither were present in northeast Ohio at 9500 B.P. (Webb et al. 1983:154-157; Davis 1983:173). Elm, however, should have been even more common than at present and its absence from Level II samples may be due to the fact that deadfall of prefered species was more easily obtained. Hemlock is presently a common species in the Chagrin River valley ravines (Williams 1940:19-23) and apparently has been for millennia. Other significant differences between the assemblages, from one dominated by oak and pine at 9500 B.P. to one of beech and maple at 5500 B.P. may reflect general environmental availability rather than shifts in cultural preference. The fragment of charred walnut hull from Level I is a poor seasonal indicator, walnuts in the hull being storable for several seasons. Certainly, plants yielding fruits widely utilized throughout the prehistoric record existed in the proximity of the shelter, so that the lack of other charred seeds could be considered at least seasonally indicative had the recovery methods been more thorough. 30 Brose No. 44 i METRIC li , , : 2| , ,1,3; 1 4i 5 6 ! o ' | j m ' : j ; | j n s j ! ill! ;|]j jm| 1 1 [ j i ! Figure 8. Fragment of charred cordage from Cultural level I. Charred Cordage A single piece of knotted cordage was recovered during the excavation of Excavation Unit 4. Within the sandy Stratum 4, at a depth of 70cm, Feature 3 consisted of a shallow fire reddened silty clay hearth 25cm in diameter and 5cm thick. Between this hearth which contained flakes of chert and charcoal dated to 5500±85 B.P., and the rear shelter wall 45cm south, slightly shallower compact dark silty sands 10cm thick yielded several lithic tools and flakes and a single fragment of charred cordage. The cordage, analyzed by Drs. Andrews and Adavasio (this volume), consists of a single fragment of spun and twisted plant fibers knotted at both ends. It may have been a knotted clothing fringe, a bundle of construction material, or a part of a snare or trap (Figure 8). Human Skeletal Material From the lowest portion of Stratum 2, running for 160cm along the southern wall of Excavation Units 1 and 5 at a depth of 115cm, fragmentary remains representing at least two individuals were recovered within a thin zone (Feature 2) 30cm wide. The stratum is a dark brown sandy loam flecked with granules of charcoal dating to 9200+150 B.P. Feature 2 within this stratum was from 10 to 20cm thick at the back of Excavation Units 1, 3, and 5, and contained small areas with thin laminae of yellowish or light grayish-brown fine silts and clays. These laminae parallel the rear wall of the shelter and represent a surface reworked by low energy water transport, perhaps seasonally. Based upon the analyses (Prior, this volume) the majority of the material is from a young adult female, additional individuals being represented by three isolated teeth not in association. Unfortunately, while adding to a limited data base (Protsch 1977), the skeletal material does not allow much comparision of Early Archaic regional demography or morphology possible at other Table 4 Combined Squaw Rockshelter Debitage Categories Cultural Level II Used Unused Total Block Cores Upper Mercer • 2(1) 2(1) Plum Run • 2(2) 2(2) Glacial • • • Pebble Cores Glacial e 1 1 Decortication Flakes Upper Mercer • • • Plum Run 9 • • Glacial 4(2) 9(3) 13(5) Block Shatter Upper Mercer 9 4(3) 4(3) Plum Run 1 (1) 5 (4) 6(5) Glacial 2 3(2) 5(2) Thinning/Retouch Flakes Upper Mercer 5(3) 3 (3) 8(6) Plum Run 1 (1) 8(8) 9(9) Glacial 2(1) 4(1) 6(2) Subtotals Upper Mercer 5 (3) 9(7) 14(10) Plum Run 2(2) 15 (14) 17(16) Glacial 8(3) 17(6) 25(9) TOTAL 15 (8) 41 (27) 56 (35) 1989 Squaw Rockshelter 31 sites ( e.g . Charles and Buikstra 1983; Redder 1985; Dickel and Doran 1988). Chipped Stone During the excavation of the shelter 168 pieces of debitage were recovered. Application of several distributional analyses (cf. Brose and Scarry 1977), adjusted for volume of sediment from each excavation unit, does not indicate statistically significant clustering. No doubt, more rigorous analyses (Vance 1987:59) would suggest otherwise, but the waterworked surfaces of Cultural Level II, the longer but unknown erosional processes during the formation of Cultural Level I, and the large size of most debitage recovered, argue that concentrations would be lag deposits, geological rather than cultural in origin. Those processes urge caution in any attempt to reconstruct not only spatial but technological aspects of lithic reduction during either period of occupation. The distribution for each cultural level (Tables 4 and 5) is presented in terms of debitage categories as stages in the manufacturing trajectory (Brose 1970:97-106, 1978; Prufer and Long 1986:42-44). Analyses of debitage weight and surface area reveal that debris from all stages of biface reduction and use are present in both occupations (Stahle and Dunn 1984:4-9). In that sense, the Early Archaic lithic procurement and reduction is similar to the Kirk occupation at the Calloway Island site (Chapman 1979:36-62). Identification of lithic sources was primarily visual, although petrographic analyses confirm some attributions. The nearest outcrops of the black Upper Mercer flint lie 55km east, along the Portage/Trumbull County line (Prufer and Long 1986:18). The mottled orange to greyish-blue Plum Run flint is from a small quarry 70km south in Stark County (Murphy and Blank 1970: 198). Glacial flints are mostly brownish-grey, dark grey, and mottled blue-grey Palaeozoic cherts of the Niagaran formations in Ontario (Parkins 1977), present in local stream gravels throughout glaciated portions of Ohio (Brose et al. 1981). Heat treatment was identified by the presence of texture and color changes, as well as by “pot-lid” fractures underlying primary flaking scars. Utilization of debitage edges for expedient tools was determined by low powered opaque microscopic observation, following methods described by Frison (1968), Brose (1970, 1975, 1978), Tringham et al. (1974) and Keeley (1980). Although no significant changes in knapping chipped stone can be demonstrated within Squaw Rockshelter, procurement and reduction differed between levels. There is a small, but statistically significant increase Table 5 Combined Squaw Rockshelter Debitage Categories Cultural Level I (Number showing heat-treatment) Used Unused Total Block Cores Upper Mercer • 3(1) 3(1) Plum Run 2(1) 3(2) 5(3) Glacial • • • Pebble Cores Glacial • 4 4 Decortication Flakes Glacial 6(2) 8(4) 14(6) Block Shatter Upper Mercer 1 (1) 7(5) 8(6) Plum Run 5(2) 12(8) 17(10) Glacial 8(2) 11 (6) 19(8) Thinning/Retouch Flakes Upper Mercer 4(4) 7(5) 11 (9) Plum Run 1 (1) 8(7) 9(8) Glacial 7(1) 15(5) 22(6) Subtotals Upper Mercer 5 (5) 17(11) 22(16) Plum Run 8(4) 23(17) 31 (21) Glacial 21 (5) 38 (15) 59 (20) TOTAL 34(14) 78 (43) 1 12 (57) 32 Brose No. 44 Figure 9. Projectile points from Cultural Level II. See Table 6 for sizes. in the frequency of the more local of the three sources of raw material used from the Early Archaic (9300 B.P.) to the late Middle Archaic (3300 B.P.). The Plum Run quarry was initially favored over the closer Upper Mercer outcrops, possibly because the former source displays thicker bedding. Through time, however, there was a greater utilization of the local glacial cherts for the production of expedient tools. This localization of procurement is related to the highly significant decrease in heat treatment in utilized and unutilized debitage. In the earlier period almost all Plum Run flint was heated, while in the later period only one third was. A rough index, comparing initial reduction stage decortication flakes and block shatter, to final reduction stage thinning flakes, suggests little change over four millennia. However, it is unlikely that all areas of a rockshelter would have been equally attractive for performing different aspects of flint knapping. Expedient tool use was more likely to have taken place in the well-lighted shelter opening, while curated tool loss was more likely to have occured in the darker rear areas of the shelter. Most frontal portions of Level II had been lost to erosion before Level I formed so that if similar portions of the shelter were used for rather similar activities differing segments of the production sequence could be recovered from sequent levels in a single unit, even if overall reduction strategies were similar. Gramley (1980) discussed site function in terms of a curatorial index, comparing ratios of bifaces and formal tools to the ratios of expedient tools and utilized flakes. Fitting (1967) had used a similar index to infer group sexual composition. Clearly, the interaction of site function and demography is responsible for the fact that at Squaw Rockshelter 40% of lithic artifacts in the Early Archaic Level II were curated, while only 20% were in late Middle Archaic Level I. Beyond the utilized debitage, 28 formally produced tools were recovered from the shelter. Not every provenience from the site is unimpeachable, and even though there is stratigraphic separation between cultural levels, neither demonstrate the internal integrity needed for satisfactory social interpretations. Nonetheless, these examples of local style and technology can be compared to coeval assemblages in other regions to illuminate the Early and Middle Archaic occupations in northeast Ohio. Indeed, the bifacial projectile points from the Early Archaic level are evidence for the interaction of cultural patterns hitherto isolated in Ohio, with exceptions so infrequent their very existence had been questioned. Table 6 Projectile points from Squaw Rockshelter (33CTJ34) Cuyahoga Co., Ohio, Level II (in millimeters) 1989 Squaw Rockshelter 33 c Ol E E o U 3 H > cS Cu Oh o o cd X) OX) C o -CD JD S 'Oh o 2 X 04 )— i > xr x - o 04 > X 04 o X o ©X) C 2 £ 0 04 X X On 73 04 1 3 O .2? db ^ £ £ > S cd 04 c X 3 C/D Cd 8> E X 3 73 C cd 11 o — Otj 3 3 E -O ^ •g 3 a ) cp S ON cd Cu C/D Cd cd cp Xi UM 04 X OD rn C ip) Of) 04 ON 'S 3 O S3 c Q ’E D 5-h c cd 04 04 S_ D -x X Id D 04 c Z3 , . Lh g s £ £ -2 c X 04 X E C/D cd 04 C o X c cd C/D & on 04 X > o ^ 3 X Of) O E '% 6 E c O 03 CQ O CQ 3 o c/d cd 04 x ■3 E « -r 2; 60 J3 3 ^ _ 04 4-h XJ — (1) On 2 S3 2! 22 c 3 3 g .2? pu c/) -a 3 03 c 3 H 03 r3 U « *" « oo c E § p£ cd 04 -1 E 53 | CQ . X 04 0) -j cd « 2 =j 2 o c CL) O x a> c s 3 »- 2 13 “•■3 P> CD •? E 03 60 3 C — o £ 3 00 t/3 T3 £ LU 5 . ©£) H _Q4 "04 > 2 04 (L) X X 04 04 X J- 3 O r~| ©X) ©Xj X C 'S 2 C r- 04 2 S , — 1 04 04 n c Of) Qh 4_J c E 4_> o a O- 'S 04 X X 'S U o D ’> c D C/D C/D C/D 04 4— * OXJ 04 04 C/D O c o i— Of) C/D E- c cd CJ cd 04 cd ee C 04 Os _< PI GO X X X CP IP) X c X WD oo p o e 04 # ON » pi X X X P; P CP CP X pi ON pi pi CP £ OX) -o d> a X 03 CO X3 .g d> £ -X C / 3 C > oij o >A c > o d> > d> X *5 Dh -a d> OX) E c c o > C > o l-H C d> d> > be X OJ CO .2* « X) 3 O bO OX) rZ .E D- c n. ■ != 03 C C 3 Q£ E 3 aj ex 3 2 Q. _G 3 -J D. ~ a. 3 E- c . o c/3 -r* S..2 T3 d) C/3 X X C X 03 , td ^ j: d> x x . 3 D O O 04) X jr1 c/3 OX) Uh C 03 O O C/3 03 d) X 03 OX) T» X — 1 OX) d> c OX) O s— -—h o3 03 ^ X • o ^ 3 ■s s > d) d) *~ X a. . a> ^ > d) X *c d> Of) 5 X c D D g " •— C/3 D z D O X C/3 d> OX) C/3 & X O i— C/3 O z .12 £ “O Q W < Z o x o o 00 aj 0/ I in a o \D ‘t: OX) “ e SS 5 rj < z < z < z 3 sc 3|- CN < z < z 3 ~ i r. "O 3 •— so £ o [V < z c X ON 00 o 04 — or Os O'’ 04 X Of) c QJ X x CO or 04 04 of oT Os ir~t sc © Z z g ^5 U CJ < < CQ < CQ X O' 0 0 O Tf ^r •o •o 0 O 0 0 QO 00 00 00 00 (bifacial) convex scraper. Minimal retouch on distal ventral surface. Transverse striations on all ventral edges. Heavy platform crushing. Glacial chert. Figure 13e. 42 Brose No. 44 that each areally extensive sealed horizon shows a waxing and waning of differing projectile point hafting categories, but that every unmixed sealed Koster component yielded from three to seven named corner-notched, side-notched, stemmed and lanceolate point types. The variable haft styles showed definite clusters so that any one of the activity foci at any occupation level might yield a predominance of one haft category (cf. Brown and Vierra 1983). Similar situations seem to have existed in Missouri, (O'Brien and Warren 1983:82-93, fig. 5.13), and in central Illinois (Lewis 1983:102-105), and also at Modoc Shelter in the Mississippi Valley (Styles, Ahler, and Fowler 1983). Montet-White (1973) documented the statistical significance of this phenomenon, but failed to explicate potential causes. The faunal assemblages indicate similar diffuse hunting/gathering economies with small game predominant. From Late Paleo through Middle Archaic there is unambiguous evidence for increasing reliance on seeds and nuts (Brown 1983). In the east, a sub-continental synchrony of change in Early Archaic projectile points types seems anchored by a suite of radiometric determinations. It has been argued (Brose and Lee 1980) that hafting morphology within any eastern point “type" shows as much variability as do the sets of different point types from coeval sites to the west. Even allowing Coe (1964:123) to order the undated site profiles as best he can, his classic excavations do not exactly show one point type for one level.2 This is clearly seen in his summary figure (1964:121). From 8400-7500 B.P. Palmer and Kirk corner-notched, Kirk side-notched and stemmed, and Stanly stemmed co-occur. From 7500- 6500 B.P. Kirk stemmed and corner-notched, Stanly stemmed. Morrow Mountain I, and Guilford Lanceolate points exist at virtually every level. Excavation of the St. Albans site along the Kanawha River in West Virginia exposed two areas of Early Archaic and stratified early Middle Archaic occupation (Broyles 1971:19- 20, 47-48). Type descriptions and accompanying illustrations (1971:53-76) suggest morphological and metric continua from large Kirk corner-notched, to Amos corner-notched, to Charleston corner-notched, to small Kirk corner-notched, to large Kirk corner-notched; and from Kanawha stemmed, to Kirk stemmed, to MacCorkle stemmed, to St. Albans side-notched variety B, to St. Albans side-notched variety A, to LeCroy Bifurcate base, to Kanawha stemmed. Indeed, Broyles herself (1971:71) noted this intergradation. Brose and Lee (1980) suggested that using the type descriptions, only Broyles’ Kessel side- notched projectile points (1971:60-61) might be resorted into the same type. The statement that, “Each type of projectile point was confined to one or two zones...” (Broyles 1971:24), even ignoring presumably displaced Kirk and St. Albans points at the site (ibid.), reflects a situation where each “type” is arbitrarily segregated.3 The Little Tennessee river terraces revealed deeply buried sites, stratified from late Paleolndian through Middle Archaic. At Icehouse Bottom (Chapman 1973, 1976); Rose Island (J. Chapman 1975, 1977); Patrick (Chapman 1977) Harrison Branch, Calloway Island, and Thirty Acre island (Chapman 1977), Bacon Farm (Chapman 1978), and at Howard Farm, with the largest Early Archaic exposures (Chapman 1978) there are consistent series of what Chapman (1977:52-55) called projectile point/knife Type Clusters. There is a Kirk corner- notched cluster dated between 9500 B.P. and 8500 B.P. (although Chapman prefers 9400 B.P. to 9100 B.P.) a Bifurcate Point Cluster between 9400 B.P. and 7800 B.P.; a Stanly (Kanawha?) Cluster between 8000 B.P. and 7000 B.P. and a Morrow Mountain phase from perhaps as early as 7000 B.P. (Chapman 1977; 16 1 - 167). Within each discrete stratigraphic level, at each site where the “type cluster” exists, a variety of hafting morphologies co-occur.4 Thus, even in the prototypic sealed stratified floodplain sites of the Appalachian southeast. Late Paleolndian stemmed and unstemmed lanceolate points, and Early Archaic side and corner-notched projectile points co-occur. Further, in the levels running into the Middle Archaic, the named “type clusters” represent as much variability in hafting as would define distinct projectile point types in the midwest. It seems clear that, not only is there some importance to the excavated area of relevant levels (Koster = 625m2; Napoleon Hollow = 435m2; St. Albans = 240m2; Icehouse Bottom = 170m2; Neville = 54ni2; Hardaway = 42m2; Gaston = 28m2) as I had suggested (Brose and Lee 1980), but that differing schools of typology might be responsible for the apparent difference between Archaic horizons in the east and the midwest. If Early Archaic assemblages from the riverine midwest and the southeast indicate the potential variability in single components between 9400 B.P. and 4500 B.P., then stylistic or functional lithic variability is also expected in the ecologically complex region of northeast Ohio. Local Lithic Assemblages While diagnostic projectile points have been recovered from loci throughout northeast Ohio, which appear to be relatively “pure” Plano, “pure” Appalachian Archaic, or “pure” Laurentian, given the environmental complexity, there is little reason to expect that all Archaic sites would display similar variability. The McConnell workshop at the Nellie Chert quarry in the Glaciated Plateau, with a preponderance of lanceolate points is one example of a site of limited function, although even there other types were manufactured from specific preforms (Prufer 1963:13, 31 fn.2). The recovery of only Matanzas/ Brewerton eared-notched points from the small Merkle 2 site, on a buried Cuyahoga River terrace (Brose 1975c: 14, 20) related to a channel shift long after 8540±70 B.P. (Miller 1983), appears to be a site with similarly limited lithic style. 1989 Squaw Rockshelter 43 Based upon recent paleoecological and archaeological data, Brose and Lee (1980) argued that while some Archaic sites in northern Ohio may have been drowned, points recovered from the intermoraine region suggest that rather large sites now lie buried along major rivers, or were located in the adjacent uplands (Brose 1975c, 1976a, Brose et a! . 1981). Those sites, not functional equivalents of smaller occupations, should yield a considerable range in projectile point styles often considered representative of discrete Late Paleolndian through early Middle Archaic occupation. 1 do not intend to suggest that multi-component sites with millennial lacunae do not exist: they surely do, and the Welti site (Fitting 1963), the Hospital and Academy sites (Brose 1 975b:24-38), the Norman ‘P’ site (Brose 1976b) and perhaps the McKibben site (Prufer and Sofsky 1965) appear to be of that type. But, just as small “pure” sites have been reported, larger sites occupied between 9000 B.P. and 4500 B.P. exist and, like Squaw Rockshelter, display considerable lithic variability. Rather than an arbitrary segregation into Paleolndian and Archaic occupations the assemblage from these sites may represent a single component. Among reported sites of this type are the Hogue’s Spring site in Cuyahoga County, Ohio (Brose 1975c: 15- lb), and the Holdson District sites 2,3, and 4 in Ashtabula County, Ohio (Brose 1977:83- 91). The recovery of Kirk, Dalton-like, and St. Albans side-notched ‘A’/LeCroy Bifurcate Base projectile points, along with ovate bifacial knives in salvage of the Regis site, on a plateau above a tributary of the Cuyahoga (Brose 1975c: 17,48) in Summit County, Ohio, may also be noted. Several undescribed sites known through test excavations also illustrate the lithic variability of this period. The Cooper Hollow site on a bluff of the deeply incised Vermilion River in northwest Lorain County was identified by a band of unprincipled amateurs. I tested the area in the fall of 1972 and extensive excavations were conducted by Lee through the summer of 1974. The excavations revealed a sub-plowzone paleosol containing discrete areas of firecracked rock. Within a 180m2 zone there was a large focal cleared zone and a large focal roasting pit, as well as several shallow pits. The overall site configuration is rather like that at Holcomb Beach (Fitting, DeVisscher, and Wahla 1966). The site yielded twelve complete finely made lanceolate bifaces which in hafting morphology grade from narrow based Cumberland to stemmed Scottsbluff lanceolate bifaces. They duplicate points from the Fisher site in Ontario (inspection of materials at the R.O.M. courtesy of Dr. Peter Storck, 1988). Twenty tip and midsection fragments also reflect this continuum. Lateral edge grinding is common but not ubiquitous. From sealed features and paleosol Lee recovered a corner-notched/stemmed point, and the lower lateral portion of the base of a Stringtown Spured-Stem or Hardin Barbed point. Excavation also yielded “Micro” drill tips, rough flake tools, cobble tools and choppers, and over 1100 fragments of debitage (15% used). Flotation of the roasting pit produced fragments of nut shell and fragmented calcined bone representing a large cervid. It also contained charcoal yielding an unacceptable 12,100±250 B. P. date for the Paleolndian-Early Archaic occupation. The Zigler Farm site, on the lake plain in southeastern Lorain county, sits on a knoll at the edge of a bog draining the West Branch of the Black River. Testing by N’omi Greber in the summer of 1973 recovered a single Cumberland point, two edge ground and four unground Scottsbluff points and five McCorkle-like points from undisturbed paleosols in ground undulations. The Burrill Orchard site, sits on the point where French Creek joins the Black River in Lorain County. Excavation by Brose in the summer of 1971 revealed discontinuous areas of sealed paleosol and two pits. Among minute fragments of charred bone, the pits yielded five points grading from Hardin Barbed to Scottsbluff, and three points intermediate between Kanawha Side-notched and St. Albans Type B. Edge grinding occurs on most points. The presence of nearly lm of overburden, and wishes of the sometime owners to preserve a peach orchard, prevented exposure of continuous site areas. Subsequent excavations 40m south along the same plateau (Brose 1978b) revealed an arc of post holes associated with a charcoal and ash-filled pit dated 7 1 20± 1 50 B.P. (DIC- 734) containing one Hardin Barbed point, and one point which Chapman (c.f. 1 977:30-35, 163-168) calls Stanley/Morrow Mountain II. With the clarity of hindsight, we can argue that not all of the “sites” documented by Prufer and Baby were multi-component. The data used to construct their Paleo Complex were abstracted, post hoc, from assemblages which contained bifurcate-based, expanding stemmed, and notched points in numerous varieties and frequencies. At the Sawmill site in Erie County, whose late Paleolndian complex had been separated from a “later” component on the basis of 1963 professional typologies. Smith (1967) illustrated unambiguous Hardin Barbed, Charleston Corner-notched (some approaching Palmer in morphology) McCorkle Stemmed, St. Albans Side-Notched ‘A’ (or Kirk) and Thebes points as well as the types defined as late Paleolndian. Like Squaw Rockshelter, these sites show no Archaic occupational hiatus across northeast Ohio. They reveal local assemblages which show stylistic and technological continuities despite the frequent morphological assignment of specific artifacts to discontinuous distant prototypes. Speculations To what cultural processes may we attribute the distinct Early Archaic lithic variability and the Middle 44 Brose No. 44 Archaic stylistic homogenization at Squaw Rockshelter? Answers must rest on assumptions concerning the number and composition of occupants; the season or seasons during which they used the shelter; and the economic activities that drew them to this portion of Aurora Branch. Despite evidence from Michigan to Maine for a Clovis settlement system exploiting migrating caribou, Payne (1987:2) admitted that no clear evidence for any Paleolndian settlement pattern has yet emerged in northwest Ohio. The same must be admitted for all of the state, and for the entire Archaic as well. No stratigraphic evidence recovered from either the Early or late Middle Archaic levels permits a decision as to whether Squaw Rockshelter was occupied at a single time or was intermittently reoccupied. Although artifacts in both levels were distributed through 10 to 30cm, in sediments of similar texture microbiotic processes and soil plasticity can vertically displace artifacts by 40cm or more (Hoffman 1986:167). In eastern Pennsylvania, at Shawnee-Minisink (McMillan 1986:264), and at Sheep Rock Shelter (Michels 1968; Michels and Smith 1967) the Early Archaic is taken to represent transitory, limited purpose occupation by small band segments at the limit of their large range. At Sheep Rock the Middle Archaic occupation was seen as multi- purpose reoccupations for longer segments of the year by several families occupying a more demographically circumscribed territory. The latter interpretation appears to have been derived principally from analysis of regional projectile point relationships, not from any spatial data recovered at the site. So too, Michlovic (1976:14-15) offered ethnographic analogs suggesting that the observed blending of Archaic projectile point stylistic modes reported in western Pennsylvania may have been the result of ephemeral interaction by short term, limited work parties of mixed ethnic or social composition. Anderson and Hansen (1988) have proposed a social and economic geography model for the Early Archaic of the southeast coastal plain. Their model involves up to 10 bands of 50- 150 people occupying the territory of a single river valley, congregating where resources are annually abundant to exchange information and genes (cf. Brose 1979). Matching the ethnoarchaeological data of Wiessner ( 1983), Anderson and Hansen see assemblages of hafted lithic bifaces displaying the greatest variability at the sites which represent the aggregation of several small bands. Curated tools and possibly the range of raw materials may also have higher frequencies at such sites. Somewhat in opposition to these conjectures, Behm (1985) has suggested an absence of distinct social territories in the Early Archaic of the Upper Mississippi Valley based upon analyses of spatial patterning of stylistic variation in projectile points. While there would seem little potential for Squaw Rockshelter to have been inhabited by even one small band, several similar (yet untested) shelters in the environmentally diverse kame and kettle topography for 2km downstream along Aurora Branch could have been occupied contemporaneously. It actually remains unclear whether the Squaw Rockshelter ever housed a family. There is every reason to believe that most, if not all, of the stone tools recovered in either level could have been used (and many, if not most, of the tools could have been produced) by women (Gero 1988). Perhaps the increasing use of locally available and unheated chert to produce more expedient tools represents a shift to more female occupants from Early to late Middle Archaic period use of the site. If the demography of the Squaw Rockshelter occupants is uncertain, the season in which occupation was probable may be approximated. Deller (1976:5-6) suggested that Ontario Paleolndian and Early Archaic sites had southern exposures to protect from the north wind and to provide light. The latter would only be true in the winter, however, and early Amerindians must have lived somewhere during other times of the year. Hall and Klipple (1988) argue that most occupied rockshelters on the Cumberland plateau had a southern exposure, suggesting occupation primarily in the winter and early spring for protection from northern storms. Size of protected area, proximity to varied resource catchments and avoidance of flooding seem the factors which conditioned choices among shelters. Similar patterns have been observed in southern (Bush 1987:121-129) and northern Kentucky (C. Wesley Cowan, personal communication 1988), and to a lesser extent in eastern Ohio (Olaf Prufer, personal communication 1988), but most of those occupations postdate the Archaic. If the northern exposure of Squaw Rockshelter argues against a winter occupation, the total absence of faunal remains, bivalves or any edible floral remains save a single burned fragment of walnut hull in Level I, may argue against occupation in the summer or fall. Storck (1982:23-25) suggests that fishing had more significance for late Paleolndian site location than is usually considered, but it is unlikely many fish could have been procured this far upstream. In the late spring, with most resources in short supply, subsistence efforts may have concentrated at lakeshore marshes or upland fens and bogs where emergent vegetation and tubers could be obtained with little effort. While such resources were available nearby as early as the occupation of Level II, they could never have been capable of supporting social aggregates of any size, and the limited area available within Squaw Rockshelter may have been more than adequate for any population the catchment could sustain. Like most such ethnoarchaeological models, the relevant environmental and cultural factors can be demonstrated only for late prehistoric or for subsequent periods. Their applicablity to any portion of the Archaic is conjectural at best, and present data are far too thin to support arguments concerning whatever cultural changes 1989 Squaw Rockshelter 45 Figure 14. Location of soil samples from the Squaw Rockshelter site. may have occured over the first two-thirds of that period. Nonetheless, I believe that the technical and stylistic data to be derived from such Archaic lithic assemblages as were present at Squaw Rockshelter can no longer be seen as mere indicators of gross chronological position and general inter-group relationships. For a vast stretch of time they constitute the only evidence we are likely to have from which to recover details of how critical functional and economic vectors may be tempered by variable social structure. To be sure, the quantity and the quality of the data are less than we might wish. But our potential for discovering more Archaic sites remains obscure, and our potential for less archaic interpretations is clear. A Appendix Sample Lithology During the excavation at Squaw Rock Shelter a series of 1.50 liter soil samples were collected from the eastern corner of Unit 3. A series of column-like samples were also taken from the western edge of floors of Unit 5 (Figure 14). The frequency of fragments of sandstone embedded within the finer matrix of sands and silts precluded strict column sampling (cf. Farrand 1985). Sampling was also biased because irregular cobbles of sandstone roof fall, which far exceeded 64cm in every dimension, could hardly be included. Nor were the very fine silt and clay sized particles separated by hydrometry. The remaining samples, from pebble (-5 Phi) to fine silt (+7 Phi) were air dried, quartered and mechanically shaken through a nested column of Wentworth grade sieves. The relative frequency of each grade class by weight is presented as granulometric histograms (Figure 15) in which column A represents the excavation Unit 3 sample while Column B represents the sample from Excavation Unit 5. Small samples were also studied for grain lithology and wear (cf. Brose 1970:27-9), with the assistance of the late Dr. John Hall, Case Western Reserve University Department of Geology. Sample AO A moderately well sorted deposit of yellowish brown ( 10YR4/2) to strong brown (7.5YR5/6) fine sandy silt loam lying unconformably upon the eroded surface of the Bedford Shale at a depth of 155cm below survey datum, itself about 10cm higher than the surface in XU1. Included within the coarse fraction of this sample were several small rounded pebbles of black shale and several of quartzite. Coarse sands included about 60% quartz, and about 40% 46 Brose No. 44 Granulometric Histograms Squaw Rockshelter 33CU34 Figure 15. Granulometric histogram of soil samples from the Squaw Rockshelter site. 1989 Squaw Rocks h e l t e r 47 feldspars along with a few carbonate grains. Virtually no heavy minerals were identified in the sample. Hiram Till is always characterized as a silty clayey till with very few pebbles, predominantly of black shale, and rarely containing cobbles or boulders. The grain size distribution for local deposits is variously reported as sands: 15-20%; silts: 45-48%; clays: 33-40% (Shepps 1953; Winslow and White 1954; White 1982:25; White 1984:17; Szabo and Ryan 1981: 242). Samples A I , B1 The small coarse sand sample analyzed revealed about 60% rounded grains, about 50% of which were feldspars, although a few grains of chert and tourmaline were present. The presence of the numerous large cobbles has shifted the relative histograms (Figure 15) toward the negative Phi units, but the matrix in which they are embedded appears to represent detritus chemically weathered from both the Bedford Shale and Berea Sandstone. The larger fragments are presumably detached by mechanical weathering processes such as freezing. Samples A2, B2 Qualitative geochemical analyses indicate a rather high organic content (5% loss on ignition) and a neutral pH. The sedimentary sources for this deposit appear to have been chemically weathered predominantly from the Berea Sandstone. There were, within the medium to coarse sand sample, only a few grains of chlorite, garnet and zircon and most of the quartz grains were sub-angular (cf. Pepper et al. 1954:91-95). The A2 and B2 samples, when compared to the relative granulometric distribution of the local Berea “channel fill,” show strong positive Phi skewness (Krumbein and Pettijohn 1938:229-254), especially laminae from within portions of Sample B2 along the rear shelter wall. Samples A3, B3 Relatively similar predominantly dark yellowish brown (10 YR 4/4 to 10YR 3/6) and light yellowish brown (10YR 6/4) fine sands although small indistinct mottles of dark reddish brown (2.5YR 3.5/5) and strong brown (7.5YR 4/6) medium sands occur throughout the stratum. Quartz grains within these mottled areas tend to show considerably more rounding than do those in lighter colored areas of the deposit. No geochemical differences between mottled and unmottled areas was noted. Nor were there any significant differences in the low frequencies of heavy minerals (zircon, tourmaline, chlorite, and apatite), or in the rare presence of well-rounded pebbles of dark bluish gray (5B4/1) shale. Although a few coarse sand- sized grains of dolomite or ankerite were present in Sample A3, the entire deposit appears relatively acidic (pH 4.5). Neither organic nor cultural material was encountered. Samples A4, B4 Analyses suggest some degree of water transport of sediments fallen from the roof and back wall with lower energy levels (or less frequent episodes) in the front shelter area. Those sedimentological processes resulting in the formation of Stratum 4 appear to have persisted throughout the period (or periods) of human occupation. It seems likely that persistent seepage across the fore-sloping zone was responsible for the removal of relatively finer sediments from the rear of the shelter. The lack of contiguous floor areas may be in part due to such a phenomenon. While the exact depositional nature of the stratum remains somewhat ambiguous, introduction of coarser sediments by overbank flooding or by wash in from areas upstream or upslope would seem unlikely. The rocks of the Orangeville Member and Bedford Formation exposed in those areas are both consistently finer than the Bedford Sandstone or the Stratum 4 sediments, and they display a far greater incidence of rounded grains and a very different suite of heavy minerals (Prosser 1912:519-30; Pepper et al. 1954:42-45,91-95). Samples A5, B5 and A6, B6 Within Stratum 5 there was a consistent difference in the grain size distribution skew from lower levels (Samples A5, B5) to upper levels (Samples A6, B6). Lower levels had a far greater frequency of silts, pebbles and small cobbles than did the upper levels. Rear wall areas of this deposit also exhibited a higher relative frequency of fine silts than riverward areas of the same depth. Although no bedding or lamellar structures could be identified within Stratum 5, there were significant differences between the front and the rear in the small pebbles included within the samples analyzed. Numerous sub-rounded shale and quartzite pebbles were encountered in Samples A5 and A6 while only a few well-rounded shale pebbles occurred in Samples B5 or B6. Within the coarse sand fractions, A5 and A6 yielded dolomite, quartzite and zircon while B5 and B6 yielded tourmaline and chert only. Although all of these minerals occur in the Berea Sandstone (Pepper et al. 1954:92-94) it seems clear from their spatial distributions and from inclusive clasts, that the depositional processes responsible for Stratum 5 were rather more complex than those responsible for the grossly similar Stratum 3. Historical Reconstruction The following scenario for the evolution of the Squaw Rockshelter is based upon regional geomorphology, analyses of deposits in the shelter itself, and a tentative sequence for collapse of large portions of the roof overhang5’. 1 . Following the 12,500 B.P. drop of Lake Erie and the subsequent Chagrin River capture of upper Aurora Branch, lower Aurora Branch began retrenchment of late Pleistocene deposits in a pre-Wisconsinan valley. 2. Bedrock floor of Aurora Branch was exposed by fluvial erosion and Stratum I, resorted Hiram 48 Brose No. 44 Till, deposited on Bedford Shale floor by 10,000 B.P. 3. Upper Stratum 1 minimally reworked by ground water flow. Large block of Berea Sandstone east of the Test Trench Excavation is detached from the edge of the shelter roof. 4. Prehistoric occupation in Squaw Rockshelter 9200 B.P Stratum 2 deposits formed. 5. Reworking and sheet erosion of Stratum 2 surface. Fluvial removal of northern edge of deposit. 6. Large blocks of Berea Sandstone detached from overhang fall between Aurora Branch and Excavation Units 1, 5, 3, and 2. 7. Accumulation of Stratum 3 primarily by in-place weathering of overhang. Much roof fall, little chemical weathering or water sorting. 8. Large block of Berea Sandstone from overhang falls into position south of Excavation Unit 3, west of Excavation Units 2 and 4. 9. Prehistoric occupation in shelter while chemically and mechanically weathered Stratum 4 sediments accumulate through gravity and low energy ground water. 5500 B.P 10. Deposition of lower Stratum 5 from roof fall and reworked upstream deposits. Large fractured block of Berea Sandstone roof falls between test trench and Excavation Unit 1. Large block of Berea Sandstone falls south of Excavation Unit 4. 11. Deposition of upper Stratum 5, sediments from chemically weathered shelter wall. 12. Modern use of surface of Stratum 5, A.D. 1870- 1970. Acknowledgements Few sections of this report could have been written without the advice and enthusiasm of the late Dr. John Hall, former research associate at The Cleveland Museum of Natural History, and professor of geology. Case Western Reserve University. Beyond John’s advice, thanks are due to many collegues who shared with me their knowledge of the Archaic Period in the mid-continent. Among those deserving signal recognition are James Brown, Jefferson Chapman, William Fox, Ian Kenyon, William Lovis, Olaf Prufer and Peter Storck. If I have not always accepted their interpretations, I have never forgotten their generosity. Notes 1 It is possible that these disturbed and redeposited sediments had previously been noted as an archaeological site. A 1953 catalogue card of the Western Reserve Historical Society (53.471) lists an earlier donation by Mr. George E. Tow[N]er of the "Seventeen Indian relics picked up near Squaw Rock at Chagrin Falls.” Though shown as being in box 305 on shelf 15, those "relics” could not be located with certainty nearly 25 years later. One artifact box with no provenience designation, found on what may have been shelf 15 within that storage room, contained two complete and three broken bifacial blades and a single St. Albans side-notched point (Variety A [Broyles 1971]). All were made of Upper Mercer flint. 2 Coe’s excavations at the Hardaway site (basal level IV) yielded only Hardaway points. Level III yielded Hardaway blades and Hardaway side-notched. Palmer Corner-notched, Kirk Corner- notched and stemmed, and even a few Stanly and Morrow Mountain stemmed points. Level II contained everything from Stanly through protohistoric triangular Caraway points (Coe 1974:56-83). At Doerschuk zone XI yielded Stanly stemmed points with hafting morphology ranging nearly from Bifurcate base to side-notched (Coe 1964:36). Zone X yielded Morrow Mountain I and II stemmed points only, while Zone V, VI, and VII each yielded a variety of stemmed and unstemmed lanceolate points (Coe 1964:14-50). At the Gaston site each level from 24" to 68" below surface showed mixed assemblages of unstemmed lanceolate blades and stemmed and side-notched points (Coe 1964:84-91, 107-112). 3 Every sealed level at St. Albans site shows variation in the projectile point/kmfe hafting: The earliest levels yielded both Dalton/Hardaway and Palmer corner-notched. All Charleston components (Levels 32-36 at 9800 + 500 B.P.) included unstemmed lanceolate: (Palmer and Charleston) corner-notched; and (Kessell) side-notched points and knives. All Kirk components (Levels 14-20 dated between 8930±160 B.P. and 8800±32Q B.P.) included (Kirk and/or MacCorkle) corner- notched; (Kirk) stemmed; and (St. Albans) side-notched points as well as unstemmed lanceolate knives. St. Albans components (levels 11 and 12, dated from 8830+700 B.P. to 8820±500 B.P.) included Kirk and MacCorkle corner-notched; (St. Albans) side- notched; and (Kirk or tentative MacCorkle) stemmed. LeCroy components (Levels 6 and 8 dated at 8250±100 B.P.) included (LeCroy Bifurcated based) corner-removed or corner-notched; (St. Albans B) side-notched; and (Kanawha) stemmed projectile points. And, finally, all of the Kanawha components (Level 4 and 6 dated to 8250+100 B.P. and 8160+100 B.P.) contained (LeCroy Bifurcated base) corner-removed or corner-notched; and (Kanawha) stemmed projectile points and unstemmed lanceolate and side-notched knives and blades. 4 For example, the stratum X Kirk Complex at Howard Farm included (Kirk) corner-notched; (category 6, 13/14, and 17) stemmed; and (category 7) expanded stemmed/side-notched point/knives (Chapman 1978:16-21-32) while at Icehouse Bottom the 170m2 exposures of the equivalent “Upper Kirk" component (Strata I-L) included (large and small Kirk) corner-notched; Hardaway-Dalton; (St. Albans and Pseudo) side-notched; and a variety of (LeCroy Bifurcated) stemmed point/knives (Chapman 1977:26-57). The same variability is seen in late Middle Archaic components. Morrow Mountain levels VII and VIII at Howard Farm yielded (Kirk, Morrow Mt. Indeterminate, and Categories 6 and 17) stemmed; (category 7) expanding stemmed/side-notched; and (Kirk and Eva II) corner-notched projectile point/knives (Chapman 1978:14-16, 21-30). 5 Just below Bridge Street in Chagrin Falls, the walls of the Chagrin River gorge are composed of laminar-bedded Berea Sandstone. The existing property surveys and historical records from 1834 (Johnson 1879) indicate that the topography of this area has remained relatively unchanged for over a century and a half. The Berea outcrop, where unaffected by recent protective backcutting, is a light buff to steel blue grey in color. 1989 Squaw Rockshelter 49 characteristic of what Bownocker (1915:114) called “recent exposure”. In the South Chagrin Metropark, the bas-relief Squaw-and- Snake figure was carved into a fallen block of Berea Sandstone in 1884. The foreset beds of the block are now vertical, showing discontinuous and moderate degrees and depths of banded ferruginous staining on only a few of the higher relief areas. A comparison with the Chagrin Falls gorge suggests that the Squaw Rock block had been in place for over 150 years when the carving was made. Some estimate of the relative sequence, if not (he period, at which large blocks of roof fall took place has been obtained by a comparison of the degree of iron staining seen in their cross-bedding and, the staining seen the cross-bedding exposed in still intact sections of the shelter wall. References Adovasio, James, and Ronald Carlisle. 1982. Meadowcroft : Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage. Pittsburgh: Department of Anthropology, University of Pittsburgh. Ahler, Stanley. 1971. Projectile Point Form and Function at Rodgers Shelter, Missouri. Missouri Archaeological Society Research Series 8. Ahler, Stanley A., and Bruce McMillan. 1976. Material Culture at Rodgers Shelter: a reflection of past human activities, pp. 163- 199 in Prehistoric man and his environments, edited by R. Bruce McMillan. New York: Academic Press. Anderson, Duane C. 1980. Stone Tool Assemblages from the Cherokee Site. pp. 197-238 in The Cherokee Excavations: Holocene Ecology and Human Adaptations in Northwestern Iowa, edited by D. C. Anderson and H. A. Semken, Jr. New York: Academic Press. Anderson, David G., and Glen Flansen. 1988. Early Archaic Settlement in the Southeastern United States: A Case Study from the Savannah River Valley. American Antiquity 53(2): 262-286. Anderson, David G., and Joseph Shuldenrein. 1985. Prehistoric Human Ecology Along the Upper Savannah River: Excavations at the Rucker's Bottom, Abbeville and Bullard Site Groups. Russell Papers. Atlanta: Archaeological Services, National Park Service. Behm, Jeffery. 1985. Stylistic Analysis of Hardin Barbed Points: Cultural Interaction Within the Early Archaic. Paper Presented at the 47th Annual Meeting of the Midwest Archaeological Conference. Bense, Judith A. (editor). 1987. Final Report: The Midden Mound Project. University of West Florida, Office of Cultural and Archaeological Research Report of Investigations 6. Pensacola. Blank, John E. 1970. The Ohio Archaic: A Study in Culture History. Unpublished Ph.D. dissertation. University of Massachusetts, Amherst. Bownocker, J. A. 1915. Building Stones of Ohio. Ohio Department of Natural Resources, Geological Survey of Ohio, Fourth Series, Bulletin 18. Columbus. Brose, David S. 1970. The Archaeology of Summer Island: Changing Settlement Systems in Northeastern Lake Michigan. Museum of Anthropology, University of Michigan. Papers 41. 1972. The Northeast, pp. 11 6- 139 in Development of North American Archaeology, edited by J.E. Fitting. New York: Doubleday. 1975a. Functional Analysis of Stone Tools: A Cautionary Note on the Role of Animal Fats. American Antiquity 40:86-94. 1975b. The Hospital and Academy Sites, Archaic Occupations in Monroe Co., Michigan. Aboriginal Research Bulletin 5(3):24-38. 1975c. Preliminary Results of a Recconnaissance of the Area of Lower Tinker's Creek Valley. Peninsula: Cuyahoga Valley National Park Board. 1976a. Locational Analysis in the Prehistory of Northeast Ohio, pp. 3-18 in Cultural Change and Continuity: Papers in Honor of James B Griffin, edited by C. Cleland. New York: Academic Press. 1976b. Botanical Archaeology at the Norman P Site: A Method for the Identification of Recently Disturbed Stratigraphy at an Early Alluvial Floodplain Occupation. Journal of Field Archaeology 2(4):293-307. 1977. Archaeological Reconnaissance and Sub-surface Investigations of the Proposed U.S. Steel Co. Greenfield Steel Plant. Ashtabula Co. Ohio, and Erie Co., Pennsylvania . Cambridge: Arthur D. Little, Inc. 1978a. Lithic Technology and Cultural Ecology in Northeastern Ohio. pp. 87-115 in Lithic Technology and Prehistoric Subsistence, edited by D. Davis. Vanderbilt University Publications in Anthropology 20. Nashville. 1978b. Archaeological Investigations at the Burr ill Farm, Lorain County, Ohio, Prehistoric and Historic Evidence. The Cleveland Museum of Natural History. Report to the Lorain County Metropolitan Park District, Lorain, Ohio. 1979. A Speculative Model of Prehistoric Exchange in the Eastern United States, pp.3-8 in Hopewell Archaeology: The Chillicothe Conference, edited by D. Brose and N. Greber. Kent: Kent State University Press. Brose, David S., and Stephanie J. Belovich. 1984. Archaeological Investigations at the Greenwood Village Site, 33Su92 Cuyahoga Valley N.R.A. Museum of Natural History, Archaeological Research Report 47( 1 ): 1-194. Cleveland. Brose, David S., S. Belovich, M. Brooslin. R. Bums, J. Hall, H. Haller, C. Pierce, and C. Ubbeohde. 1981. Prehistoric and Historic Archaeological Investigations of the Cuyahoga Valley National Recreation Area, Ohio. Museum of Natural History, Archaeological Research Report 30:1-599. Cleveland. Brose, David S., and Patricia Essenpreis. 1973. A Report on the Preliminary Archaeological Survey of Monroe Co., Michigan. The Michigan Archaeologist 19( 1 &2): 1-182. Brose, David S., and Alfred M. Lee. 1980. A Reinterpretation of the Late Paleoindian to Early Middle Archaic Period in Northern Ohio. Presented at the Ontario Archaeological Society Symposium on the Prehistory of Lake Erie. London, Ontario. Brose, David S., and John F. Scarry. 1977. Boston Ledges: Spatial Analyses of a Late Woodland Rockshelter in Summit Co., Ohio. Mid-Continental Journal of Archaeology 1(2): 179-228. Brown, James A. 1983. Summary, pp. 5-11 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Brown, James A. 1986. Food for Thought: Where has Subsistence Analyses Gotten Us? pp. 315-330 in Foraging, Collecting and Harvesting: Archaic Period Subsistence and Settlement in the Eastern Woodlands, edited by Sarah W. Neusius. Center For Archaeological Investigations. Occ. Paper 6. Carbondale: SIU. 50 Brose No. 44 Brown, James, and Charles Cleland. 1969. The Late Glacial-Early Post-Glacial Faunal Resources in Midwest Biomes Newly Opened to Human Adaptation. University of Illinois, College of Agriculture, Special Publications 14. Brown. James A. and Robert K. Vierra 1983. What Happened to the Middle Archaic: Introduction to an Ecological Approach to Roster Site Archaeology, pp. 165-322 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Broyles, B.J. 1971. Second Preliminary Report: the St. Albans Site, Kanawha County, West Virginia. Report of Archaeological Investigations No. 3. West Virginia Geological and Economic Survey, Morgantown. Bush, David R. 1987. Prehistoric Lifeways in the Cumberland Plateau. Unpublished Ph.D. Dissertation in Anthropology, Case Western Reserve University, Cleveland. Calkin, Parker E. 1970. Strandlines of Glacial Great Lakes in Northwestern New York. Ohio Journal of Science 70(2)78-96. Calkin, P. E. and K. E. Miller. 1977. Late Quaternary Environment and Man in Western New York. Annals of the New York Academy of Science 288:296-315. Chapman, Carl H. 1975. The Archaeology of Missouri, Volume 1 . Columbia: University of Missouri Press. Chapman, Jefferson. 1975. The Rose Island Site and Bifurcate Point Tradition. Report of Investigations 14, Dept, of Anthropology, University of Tennessee, Knoxville. 1976. The Archaic Period in the Lower Little Tennessee River Valley: The Radiocarbon Dates. Tennessee Anthropologist ( 1 ): 1 - 1 2. 1977. Archaic Period Research in the Lower Little Tennessee River Valley. Report of Investigations No. 18, Dept, of Anthropology, University of Tennessee, Knoxville, Tennessee. 1980. The Early and Middle Archaic Periods: A Perspective from Eastern Tennessee, pp. 123- 132 in Proceedings of the Conference on Northeastern Archaeology, edited by J. A. Moore. University of Massachusetts, Department of Anthropology, Research Report 1 9. Amherst. Charles, Douglas K., and Jane E. Buikstra. 1983. Archaic Mortuary Sites in the Central Mississippi Drainage: Distribution, Structure, and Behavioral Implications, pp. 117- 146 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Coakley, J. R, and C. F. M. Lewis. 1985. Postglacial Lake Levels in the Erie Basin, pp. 195-212 in Quaternary Evolution of the Great Lakes, edited by P. F. Karrow and P. E. Calkin. Geological Association of Canada Special Paper 30. Coe, Joffre C. 1964. The Formative Cultures of the Carolina Piedmont. Transactions of the American Philosophical Society. New Series 54(5). Collins, Michael B., Editor. 1979. Excavations at Four Archaic Sites in the Lower Ohio Valley, Jefferson County, Kentucky. Occasional Papers in Anthropology Number 1 ( I & 1 1 ) . Lexington: Department of Anthropology, University of Kentucky. Coogan, Alan H., Loren E. Babcock, Joseph T. Hannibal, Doyle W. Martin, Karen S. Taylor, and David C. Wehn, 1986, Late Devonian and Early Mississippian Strata at Stebbins Gulch [,] Geauga County, and Quarry Rock, Cuyahoga County, Ohio. Field Trip Guide Book Field Trip Number One. Geological Society of America, Northcentral Section Meeting, edited by Alan Coogan. Kent: Kent State University Department of Geology, and Geological Society of America. Cook, Thomas G. 1976. Koster: An Artifact Analysis of Two Archaic Phases in Westcentral Illinois. Prehistoric Records 1 , Koster Research Report 3. Evanston: Northwestern University. 1980. Artifactual [sic] Studies, pp. 222-455 in Archaeological Salvage Excavations At Patoka Lake Indiana, edited by Cheryl A. Munson. Bloomington: Glenn Black Laboratory of Archaeology, Indiana University. Cunningham, Roger. 1973. Paleo Hunters Along the Ohio River. Archaeology of Eastern North America 1(1 ): 120-126. Davis, Margaret Bryan. 1983. Holocene Vegetation History of the Eastern Unites States, pp. 166-181 in Late Quaternary Environments of the United States: Volume 2: The Holocene. edited by H. E. Wright Jr. Minneapolis: The University of Minnesota Press. Deller, D. Brian. 1976. Paleolndian Locations on Late Pleistocene Shorelines, Middlesex County, Ontario. Ontario Archaeology 26:3-19. Dincauze, Dena F. 1976. The Neville Site: 8000 Years at Amoskeag Manchester, New Hampshire. Cambridge: Peabody Museum Monographs 4. Donahue, Jack. 1976. Geological Framework of Meadowcroft Rockshelter. Paper prepared for Meadowcroft Rockshelter Symposium, 41st Annual Meeting of the Society for American Archaeology. St. Louis. Doran, Glen H., David N. Dickel, William E. Ballinger Jr., O. Frank Agee, Philip J. Laipis, and William W. Hauswirth. 1986. Anatomical, Cellular and Molecular Analysis of 8000 year old Human Brain Tissue from the Windover Archeological Site, Nature 323(6091 ):803-806. Dragoo, Don W. 1959. Archaic Hunters of the Upper Ohio Valley. Carnegie Museum Annals, 35:139-246. 1976 Some Aspects of Eastern North American Prehistory: A Review 1975. American Antiquity 4( 1 ):3-27. Ellis, Chris. 1984. Paleolndian Lithic Technological Structure and Organization in the Lower Great Lakes: A First Approximation. Unpublished Ph.D. Thesis, Department of Archaeology, Simon Fraser University, Burnaby, B.C. 1987. Stanley/Neville Points. KEWA: Newsletter of the London Chapter, Ontario Archaeological Society 87-9:21. Ellis, Chris J., and D. Brian Deller. 1982. Hi-Lo Materials for Southwestern Ontario. Ontario Archaeology 38:3-22. 1986. Post-Glacial Lake Nippising Waterworn Assemblages from the Southeastern Huron Basin Area. Ontario Archaeology 45:27-38. 1988. Some Distinctive Paleolndian tool types from the Lower Great Lakes. Mid-Continental Journal of Archaeology 13(2): 111-158. Evans, June 1988. Paleolndian to Early Archaic Transition at the Shawnee-Minisink Site. pp. 221-259 in Shawnee-Minisink: A Stratified Paleolndian-Archaic Site in the Upper Delaware Valley of Pennsylvania, edited by C. W. McNett Jr. New York: Academic Press. Farrand, William R. 1985. Rockshelter and Cave Sediments. pp.21-40 in Archaeological Sediments in Context, edited by Julie K. Stein and William R. Farrand, Orono: Center for the Study of Early Man. Feldmann, Rodney, Alan Coogan, and Richard Heimlich. 1977. Southern Great Lakes: Field Guide. K/H Geology Field Guide Series, edited by J. Harbaugh. Dubuque: Kendale Hunt. Fitting, James E. 1963a. The Welti Site: A Multi-component site in southeastern Michigan. The Michigan Archaeologist 9(3 ):204-2 14. 1989 Squaw Rockshelter 51 1963b. The Hi-Lo Site: A Late Paleolndian Site in Western Michigan. Wisconsin Archaeologist 44( 2 ): 87-96. 1964. Bifurcate-stemmed Projectile Points in the Eastern United States. American Antiquity , 30( 1 ): 92-94. 1967. The camp of the careful Indianian Upper Great Lakes chipping station. Papers of the Michigan Academy of Science , Arts and Letters 52:237-242. 1968. Environmental Potential and the Post Glacial Readaptation in Eastern North America. American Antiquity 33(4)441-445. 1970. The Archaeology of Michigan. Natural History Press. New York. Ford, John P. 1987. Glacial and Surficial Geology of Cuyahoga County, Ohio. State of Ohio, Department of Natural Resources, Division of Geological Survey, Report of Investigations 1 34. Fowke, Gerard. 1902. Archaeological History of Ohio. Columbus: Ohio Archaeological and Historical Society. Fox, William A. 1980. The Kirk Cluster in Southwestern Ontario. Presented at the Ontario Archaeological Society Annual Symposium on the Prehistory of Lake Erie. London, Ontario. Frison, George. 1968. A Functional Analysis of Certain Chipped Stone Tools. American Antiquity 33: 149-1 55. Funk, Robert 1978. Post-Pleistocene adaptations, pp. 16-27 in Handbook of North American Indians. Volume 15. Northeast. edited by Bruce G. Trigger. Washington: Smithsonian Institution. George, Richard L. and Harry Bassinger. 1975. The Wadding Rockshelter, 36AR2 1 . Pennsylvania Archaeologist 45(4): 1-21. George, Richard L., and Christine E. Davis. 1986. A Dated Brewerton Component in Armstrong County, Pennsylvania. Pennsylvania Archaeologist 56( 1-2): 1 2-20. Gero. Joan. 1988. Women's Roles in Stone Tool Production. Paper presented at the Women and Production in Prehistory Conference. Columbia, S.C. Geisthweit, Barbara. 1970. Ohio Archaic. Unpublished Masters thesis in Anthropology. The Ohio State University, Columbus. Goldthwait, Richard, G.W. White, and J.L. Forsyth. 1961. Glacial Map of Ohio. Miscellaneous Geological Investigations. Map 1-316. U.S. Geological Survey, Washington, D.C. Gordon, Robert B. 1966. Map — Natural Vegetation of Ohio at the Time of the Earliest Land Surveys. Ohio Biological Survey, Columbus. 1969. Vegetation of Ohio at the Time of Original Survey. Ohio Biological Survey, Columbus. Gramly, Richard M. 1980. Raw Material Source Areas and “Curated" Tool Assemblages. American Antiquity > 45:823-833. Griffin, John W. 1974. Investigations in Russell Cave. U.S. Department of the Interior, National Park Service Publications in Archaeology 13. Washington, D.C. Hall, Charles L., and Walter E. Klipple. 1988. A Polythetic- Satisficer Approach to Prehistoric Natural Shelter Selection in Middle Tennessee. Mid-Continental Journal of Archaeology 13(2)159-186. Hassen, Harold, and Kenneth Farnsworth. 1987. The Bullseye Site, ll-GE-127: A Floodplain Archaeology Mortuary Site in the Co. Illinois River Valley. Illinois State Museum Report 42. Hoffman, Jack L. 1986. Vertical Movement of Artifacts in Alluvial and Stratified Deposits. Current Anthropology 27(2): 163-171. Holloway, Vaughn M., and Richard G. Bryant. 1985. Late Quaternary Pollen Records and Vegetational History of the Great Lakes Region: United States and Canada, pp. 205-245 in Pollen Records of Late Quaternary North American Sediments, edited by R. G. Bryant and V. M. Holloway. Dallas: The American Association of Stratigraphic Palynologists Foundation. Holmes, William H. 1919. Handbook of Aboriginal American Antiquities: Part I. Introductory. The Lithic Industries. Smithsonian Institution, Bureau of American Ethnology, Bulletin 60. Jeffries, Richard W., and B. Mark Lynch. 1983. Dimensions of Middle Archaic Cultural Adaptation at the Black Earth Site, Saline County, Illinois, pp. 299-322 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Justice, Noel D. 1987. Stone Age Spear and Arrow Points of the Midcontinental and Eastern United States: A Modern Survey and Reference. Bloomington: Indiana University Press. Keeley, Laurence H. 1980. Experimental Determination of Stone Tool Uses, A Microwear Analysis. Chicago: The University of Chicago Press. Kenyon, Ian. 1980. The Satchell Complex in Ontario: A View from the AuSauble Valley. Ontario Archaeology 34: 28-41. Krinsley, D. H. and J. Donahue. 1976. Application of Sand Grain Sufacetextural Analysis to Archaeology. Paper prepared for Meadowcroft Rockshelter: Interim Statement Symposium. 41st Annual Meeting of the Society for American Archaeology. St. Louis. LaRoque, Auriel. 1967. Pleistocene Mollusca of Ohio. State of Ohio, Department of Natural Resources, Division of Geological Survey Bulletin 62(2): 1 1 3-356. Lewis, C.F.M. 1970. Late Quaternary History of Lake Levels in the Huron and Erie Basins. Proceedings of the 12th Conference on Great Lakes Research , 1969:250-270. Ann Arbor: International Association of Great Lakes Research. Lewis, T. L. 1988. Late Devonian and Early Mississippian Distal Basin-Margin Sedimentation of Northern Ohio. The Ohio Journal of Science 88( 1 ):23-39. Lewis R. Barry. 1983. Archaic Adaptations to the Illinois Prairie: The Salt Creek Region, pp. 99-116 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Logan, Wilfred D. 1952. Graham Cave: An Archaic Site in Montgomery County, Missouri. Missouri Archaeological Society Memoir 2. Lovis, William A., and James A. Robertson. 1985. Rethinking the Archaic Chronology of the Saginaw Valley Michigan. Paper presented at the 12th Annual Symposium of the Ontario Archaeological Society, London. Luchterhand, Kubet. 1970. Early Archaic Projectile Points and Hunting Patterns in the Lower Illinois Valley. Illinois State Museum, Report of Investigations 19. Marshall, Sydne. 1985. Paleolndian Artifact Form and Function at Shawnee-Minisink. pp. 165-209 in Shawnee-Minisink: A Stratified Paleolndian- Archaic Site in the Upper Delaware Valley of Pennsylvania, edited by C. W. McNett Jr. New York: Academic Press. Mason. Ronald J. 1963. Two Late Paleolndian Complexes in Wisconsin. Wisconsin Archaeologist 44(4): 199-21 1 . 52 Brose No. 44 1958. Late Pleistocene Geochronology and the Paleo-Indian Penetration into the Lower Michigan Peninsula. University of Michigan, Museum of Anthropology, Anthropological Papers No. 11. 1962. The Paleolndian Tradition in Eastern North America. Current Anthropology 3(3): 227-78. 1981. Great Lakes Archaeology. New York: Academic Press. Mayer-Oakes, William J. 1955a. Prehistory of the Upper Ohio Valley: An Introductory Archaeological Study. Annals of Carnegie Museum , 34. 1955b. The Globe Hill Shell Heap (Site 46Hk34-l), Hancock County, West Virginia. West Virginia Archaeological Society, Publications Series 3. McKenzie, Douglas H. 1967. The Archaic of the Lower Scioto Valley, Ohio. Pennsylvania Archaeologist 37( 1 -2): 33-5 1 . McMillan. Barbara. 1985. A Technological Analysis of the Early Archaic, pp.26 1-320 in Shawnee-Minisink: A Stratified P aleolndian- Archaic Site in the Upper Delaware Valley of Pennsylvania, edited by C. W. McNett Jr. New York: Academic Press. McNett, Charles W. Jr. 1985. Artifact Morphology and Chronology at the Shawnee-Minisink Site. pp. 83-122 in Shawnee-Minisink: A Stratified Paleolndian- Archaic Site in the Upper Delaware Valley of Pennsylvania, edited by C. W. McNett Jr. New York: Academic Press. Michels, Joseph W. 1968. Settlement Pattern and Demography at Sheep Rock Shelter: Their Role in Culture Contact. Southwest Journal of Anthropology 24:66-82. Michels, Joseph W., and Ira F. Smith (editors). 1967. Archaeological Investigations of Sheep Rock Shelter (2 vols.). Department of Anthropology. Pennsylvania State University. University Park. Micholovic, Michael G. 1976. Social Interaction and Point Types m the Eastern U.S. Pennsylvania Archaeologist 46( 1-2): 13- 16. Miller, Barry B. 1983. Late Quaternary Fluvial History of the Lower Cuyahoga River in Northeastern Ohio, Journal of Great Lakes Research 9 ( 1 ):97- 1 05. Montet-White, Anita. 1973. Significance of Variability in Archaic Point Assemblages. Plains Anthropologist 19(3): 14-24. Munson, Cheryl A., and Thomas G. Cook. 1980. The French Lick Phase: A Dimensional Description, pp. 721-741 in Archaeological Salvage Excavations At Patoka Lake Indiana. edited by Cheryl A. Munson. Bloomington: Glenn Black Laboratory of Archaeology, Indiana University. Murphy, James, and John E. Blank. 1970. The Plum Run Quarries, The Ohio Archaeologist 20: 198-2 1 7. Newberry, J. S. 1873. Report on the Geology of Cuyahoga County. Report on the Geological Survey of Ohio 1 ( 1 ): 1 70- 200. Columbus: Nevins and Myers. O'Brien, Michael J., and Robert E. Warren. 1983. An Archaic Projectile Point Sequence from the Southern Prairie Peninsula: The Pigeon Roost Creek Site. pp. 71-98 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Ogden, J.G., III. 1966. Forest history of Ohio. I. Radiocarbon dates and pollen stratigraphy of Silver Lake, Logan County, Ohio. Ohio Journal of Science 66(4 ):387-40 1 . Parkins, W. 1977. Onondaga Chert: Geological and Palynological Studies as Applied to Archaeology. Unpublished MSc thesis. Department of Geological Sciences, Brock University, St. Catharines, Ontario. Payne, James H. 1982. The Western Basin Paleolndian and Early Archaic Sequences. Unpublished honors thesis in Anthropology. University of Toledo. 1987. Some Notes on Paleolndian Cultures in Northwestern Ohio. Paper presented at the December Meeting of the Ohio Archaeological Council, Cincinnati. Pepper. James F., Wallace DeWitt, Jr., and David F. Demarest. 1954. Geology of the Bedford Shale and Berea Sandstone in the Appalachian Basin. U.S. Geological Survey Professional Paper 259:1-1 14. U.S. G.P.O. Washington, D.C. Potter, Loren D. 1946. Post-Glacial Forest Sequence of North- Central Ohio. Unpublished M.A. thesis in Botany, Oberlin College, Oberlin. Prosser, Charles S. 1912. The Devonian and Mississippian Formations of Northeastern Ohio. Ohio Geological Survey. Fourth Series, Bulletin 15:323-534. Protsch, Reiner R. R. 1978. Catalog of Fossil Hominids of North America. New York: Gustav Fischer. Prufer, Olaf H. 1961. Survey of Ohio Fluted Points (Cleveland, The Cleveland Museum of Natural History), No. 4; No. 5; No. 6. 1963. The McConnell Site: A Late Paleolndian Workshop in Coshocton County, Ohio. Scientific Publications of The Cleveland Museum of Natural History , (New Series) 2(1). Cleveland. Prufer, Olaf H., and Raymond S. Baby. 1963. Paleolndians of Ohio. Columbus: Ohio Historical Society. Prufer, Olaf H., and Dana A. Long. 1986. The Archaic of Northeastern Ohio. Kent State Research Papers in Archaeology 6:1-89. Kent: Kent State University Press. Prufer, Olaf H., and Charles Sofsky. 1965. The McKibben Site (33TR-57), Trumbull Co., Ohio: A Contribution to the Late Paleolndian and Archaic Phases of Ohio. Michigan Archaeologist 11(1 ):9-40. Quimby, George I. 1960. Indian Life in the Upper Great Lakes. Chicago: University of Chicago Press. Redder, Albert J. 1985. Horn Shelter Number 2: The South End: A Preliminary Report. Centred Texas Archaeologist 10:37-65. Rippeteau, Bruce E. 1973. The Principles and Theory of Radiocarbon Chronology and a Radiocarbon Chronology for Lithic Variation in the Northeastern U.S. ...3000 to 500 B.C. Unpublished Ph.D. dissertation in Anthropology. Case Western Reserve University, Cleveland. Ritchie, William A. 1965. The Archaic in New York. New York State Archaeological Association Bulletin 52:2-12. 1971. A Typology and Nomenclature for New York Projectile Points. Bulletin Number 384. New York State Museum and Science Service , Albany. Ritzenthaler, Robert. 1975. A Guide to Wisconsin Indian Projectile Point Types. Popular Science Series II . Milwaukee Public Museum. Rule, Pamala, and June Evans. 1985. The Relationship of Morphological Variations to Halting Techniques among Paleolndian Endscrapers at the Shawnee-Minisink Site. pp.211-220 in Shawnee-Minisink: A Stratified Paleolndian- Archaic Site in the Upper Delaware Valley of Pennsylvania. edited by C. W. McNett Jr. New York: Academic Press. Sears, Paul. 1931. Pollen Analysis of Mud Lake Bog in Ohio. Ecology 1 2 (4):75 1 -76 1 . Seeman, Mark F., and Olaf H. Prufer. 1982. An Updated Distribution of Ohio Fluted Points. Mid-Continental Journal of A n haeology 7(2): 1 55-1 67. 1989 Squaw Rockshelter 53 Semenov, S. A. 1964. Prehistoric Technology. Translated by M. W. Thompson. London: Cory, Adams and MacKay. Semken, Holmes A. Jr. 1983. Holocene Mammalian Biogeography and Climate Change in the Eastern and Central United States, pp. 182-207 in Late Quaternary Environments of the United States: Volume 2; The Holocene, edited by H. E. Wright Jr. Minneapolis: The University of Minnesota Press. Shane, Linda. 1975. Palynology and Radiocarbon Chronology of Battaglia Bog, Portage County, Ohio. Ohio Journal of Science 75 (2):96- 1 02. Shepps, Victor C. 1953. Correlation of the Tills of Northeastern Ohio by Size Analysis. Journal of Sedimentary Petrography 23:34-48. Shetrone, Henry C. 1936. The Folsom Phenomena as Seen from Ohio. Ohio State Archaeological and Historical Quarterly 55:240-256. Smith, Arthur G. 1967. The Sawmill Site, Erie County, Ohio. Ohio Archaeologist 10: 84-97. Snow, Dean. 1977. Archaeology of the Lake George Region, pp. 431-438 in Amerindians and Their Paleoenvironments in N.E. North America, edited by Walter S. Newman and Bert Salwen. Annals of the New York Academy of Science. 288. Stafford, C. Russell. 1985. The Campbell Hollow Archaic Occupations: A Study of Intrasite Spatial Structure in the Lower Illinois Valley. Kampsville Archaeological Center. Research Series 4. Kampsville: Center for American Archaeology. Stahle, David W., and James E. Dunn. 1984. An Experimental Analysis of the Size Distribution of Waste Flakes from Biface Reduction. Arkansas Archaeological Survey, Technical Paper 2. Stewart, Andrew. 1984. The Zander Site: Paleolndian occupation of the southern Holland marsh region of Ontario. Ontario Archaeology 41 :45-79. Storck. Peter L. 1972. An unusual late Paleolndian projectile point from Grey County, southern Ontario. Ontario Archaeology 18:37-45. 1978. The Coates Creek Site: A possible late Paleo-Indian-Early Archaic site in Simcoe County, Ontario. Ontario Archaeology 30:25-46. 1979. A Report on the Banting and Hussey Sites: Two Paleolndian campsites in Simcoe County, southern Ontario. Archaeological Survey of Canada Survey Paper 93 Ottawa: National Museums of Canada. 1982. Palaeo-Indian Settlement Patterns Associated with the Strandline of Glacial Lake Algonquin in Southcentral Ontario. Canadian Journal of Archaeology 6: 1-31. 1984. Research into the Paleolndian occupations of Ontario: A review. Ontario Archaeology 41:3-28. Styles, Bonnie W., Steven R. Ahler, and Melvin L. Fowler. 1983. Modoc rock Shelter Revisited, pp. 261-298 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Szabo, John P, and Dale E. Ryan. 1981. Quaternary Stratigraphy of the Lower Mud Brook Basin, Northhampton Township, Summit County, Ohio. The Ohio J ow ned of Science 81:239- 246. Tippitt, V. Ann, and William H. Marquardt. 1984. The Gregg Shoals and Clyde Gulley Sites: Archaeological and Geological Investigations at Two Piedmont Sites on the Savannah River. Russell Papers, Archaeological Services, National Park Service. Atlanta. Tringham, R., G. Cooper, G. Odell, B. Voytek, and A. Whitman. 1974. Experimentation in the Formation of Edge Damage: An Approach to Lithic Analysis. Journal of Field Archaeology 1:171-196. Tuck, James. A. 1974. Early Archaic Horizons in Eastern North America. Archaeology of Eastern North America 2( 1 ):72-80. Vance, Elizabeth D. 1987. Microdebitage and Archaeological Activity Analysis. Archaeology 40 (4):58-60. Webb, T. Ill, E. J. Cushing, and H. E. Wright, Jr. 1983. Holocene Changes in the Vegetation of the Midwest, pp. 142-165 in Late-quaternary Environments of the United States: Volume 2; The Holocene, edited by H. E. Wright, Jr. Minneapolis: The University of Minnesota Press. White, George, and Stanley Totten. 1982. Glacial Geology of Northeastern Ohio. State of Ohio, Department of Natural Resources, Division of Geological Survey Bulletin 68. 1984. Glacial Geology of Summit County, Ohio. State of Ohio, Department of Natural Resources, Division of Geological Survey, Report of Investigations 123. White, Nancy M., Chun Ho Lee, and Judith A. Bense. 1983. Archaeological Investigations in the Upper Tombigbee Valley, Mississippi: Phase II. University of West Florida, Office of Cultural and Archaeology Research Report of Investigations 4. Pensacola. Wiant, Michael D., Edwin R. Hajic, and Thomas R. Styles. 1983. Napoleon Hollow and Koster Site Stratigraphy: Implications for Holocene Landscape Evolution and Studies of Archaic Period Settlement Patterns in the Lower Illinois River Valley, pp. 147-164 in Archaic Hunters and Gatherers in the American Midwest, edited by James H. Phillips and James A. Brown. New York: Academic Press. Wiessner, Polly. 1983. Style and Social Information in Kalahari San Projectile Points. American Antiquity 48:253-276. Williams, Arthur B. 1940. The Native Forests of Cuyahoga County, Ohio. Scientific Publications of The Cleveland Museum of Natural History 2. Winslow, John, George W. White, and Earl E. Webber. 1953. The Water Resources of Cuyahoga County, Ohio. Ohio Department of Natural Resources, Division of Water Bulletin 26:1-123. Witline, Arthur H. 1970. A Study of a Late Pleistocene lake in the Cuyahoga River Valley, Summit Co., Ohio. Unpublished M.S. thesis in Geology, Kent State University, Kent. Wright, J. V. 1978. The Implications of Probable Early and Middle Archaic Projectile Points from Southern Ontario. Canadian Journal of Archaeology 2:59-77. KIRTLANDIA The Cleveland Museum of Natural History March 1989 hi umber 44:55-58 Skeletal Remains from Squaw Rockshelter Fred Prior Research Engineering Division Information Systems, Siemens International Wheaton, Illinois 60191 Abstract Skeletal remains recovered from the Early Archaic level of the Squaw Rockshelter were identified as belonging to a young female Amerindian. Analyses of the limited frag- ments and dentition suggest a balanced diet of moderate coarseness. No pathologies or trauma were noted. Teeth from at least two other individuals were present. Introduction This report on the skeletal material from Squaw Rockshelter includes an inventory and a brief anatomical description of pathologies and anomalies where evident. It also provides estimates of the sex and age at death of the single most complete specimen. The materials are curated in the Physical Anthropology collections of the Cleveland Museum of Natural History as CMNH 8080. As noted by Brose (this volume), the circumstances of recovery suggest surface abandonment of the body in the rear of the shelter, with natural internment occuring only after some period of loss due to erosion. With the exception of a relatively com- plete mandible and five small cranial fragments all remains are post-cranial. Where it can be determined, the majority of this material is from the right side of a single individual. Inventory Post-cranial Remains A right proximal humerus with the head well preserved was present. The shaft is broken slightly inferior to the deltoid tuberosity. The distal portion was not recovered. All epiphyses of the head are fused, and there is no observable pathology. Diameter of head: vertical — 38.4mm horizontal — 36.0mm A complete left clavicle with slight damage to the acromion is present. Both epiphyses are fused. Maximum length — 138.4mm Circumference — 32.4mm Robustness index — 23.4 Fragments of one cervical and three thoracic vertebrae are present. All are marked by a lack of robustness, and exhibit no gross abnormalities. The cervical (possibly the fourth) vertebrae is complete with the exception of the anterior and posterior tubercles on both sides and the lateral borders of both transverse foramina. The three thoracic vertebrae are extremely 56 Prior No. 44 fragmentary. The most complete (possibly the second or third) consists of the posterior half of the centrum and the left transverse process with the spinous and right transverse processes missing. The remaining two thoracic vertebrae are represented by a spinous process and limina with both superior articular processes, and a small body fragment with left superior articular process respectively. A fragment of the right scapula containing the intact glenoid fossa is present. The scapula has been fractured (post mortem) along a line from a point in the scapular notch just medial to the base of the coracoid to a point on the axillary boundary about 5cm inferior to infraglenoid. The bases of both the coracoid process and the spine are present, but neither the coracoid nor the acromion processes remain. There is no observable lipping of the glenoid margin. There is also a second, smaller fragment of the right scapula composed of a section of the spine medial to that of the first fragment. The two sections do not articulate. Length of glenoid — 33.6mm Width of glenoid — 24.3mm Width of axial border (3cm below infraglenoid) — 6.5 A badly preserved right distal extremity of the femur with no articulated shaft, was recovered. The patellar and tibial surfaces are largely intact as is the lateral condyle. Only a small section of the medial condyle remains. The right tibia is represented by two fragments: a 92mm section of shaft of the proximal tibia beginning immediately distal to the popliteal line, and a section of the proximal extremity composed of the articular surface of the lateral condyle and the fibular articular facet. A small section (44mm in length) of the right distal ulna (shaft) is present. A complete right first metacarpal and a right second with the proximal end damaged. Four phalanges are present. These include a proximal, middle and terminal from the right hand and a terminal from the fifth digit of the foot. Five small rib fragments lacking any distinguishing landmarks are the only remains of the rib cage. Miscellaneous fragments include eighteen small, indeterminant, irregular pieces of compact tissue and four irregular chunks of predominantly cancelous tissue are present. All are presumed to be long bone fragments. Cranial Remains The cranial remains consist of five small fragments and a relatively complete mandible. The body of the mandible was broken (post mortem) at the socket of the right canine, but has subsequently been restored. The left ramus is missing, the body having been fractured through the midline of the socket of M3. The right ramus is present but broken superior to the mandibular foramen. Both the coronoid and condylar processes are thus absent. The first and second molars on both sides were lost ante-mortem and the sockets had been resorbed, thus giving the mandible a low profile. There is some evidence of periodontal infection of the alveolar process in the region of the anterior dentition. The chin is relatively square, but the mandible as a whole lacks robustness. Symphysial height - 30.5mm The parietal is represented by a large rectangular section of the right parietal. No sutures are evident. The meningial grooves are not pronounced. The maxilla is represented by a small section of the right alveolar process containing the socket for M3 and the distal wall of the socket of M2 with the channels for the lingual and disto-buccal roots. Sections of both the right and left greater sphenoid wing are present. The fragment of the left greater wing is roughly triangular in shape having three surfaces: the posterior, the superior or cerebral, and the orbital. The inferior section is not preserved, thus the sphenoidal spine and the sphenoidal foramina (ovale, rotundum, spinosum) are absent. The fragment of the right greater wing is smaller and is predominantly the cerebral surface. Only a small triangular section of the posterior surface and a small section of the spheno-squamosal suture are present. Dental Remains All mandibular dentition can be unambiguously associated with the remains of the single individual discussed above. In addition there are 14 isolated teeth, seven of which were found in association with these remains. The remaining seven were recovered from other units. Due to an accident in the CWRU offices, all dentition was placed together making it impossible to determine which were associated. Triplication of mandibular second and third molars (three left M2 and three right M3) indicates the minimum number of individuals is three. Table 1 lists the materials present and the state of attrition or loss. With the exception of several isolated teeth, the general state of wear is relatively low. In the mandibular dentition M, and M2 are missing on both sides and the alveolar processes have been resorbed. There is a small cavity on the mesial surface of the right P2 and the tooth is rotated slightly distally and lingually. No other caries are evident. Slight shoveling of the three incisors is evident. A protostylid appears on the right M3. The isolated teeth are in a very poor state of preservation. All are cracked or split, some to the extent of having the entire pulp cavity eroded. Age and Sex Due to the paucity of material in general, the absence of complete longbones, and the lack of reliable indicators (e.g. pelvic or cranial material), determinations of age at death and sex are uncertain. All available indicators were utilized. Table 2 lists the metric analyses employed for the determination of the sex of the specimen. All listed measures indicated that the individual was 1989 Skeletal Remains 57 Table 1 Dental Attrition and Loss Mandible Misc. Dentition Maxillary central I Right • Left • Right 0 Left 0 Maxillary lateral I • • 3* • Maxillary canine • • 6 0 Maxillary premolar 1 • • • 9 Maxillary premolar 2 • • 0 0 Maxillary molar 1 • • 6 * Maxillary molar 2 • • 6 5 Maxillary molar 3 • • 0 5 Mandibular central I 4 1 0 0 Mandibular lateral I 4 4 * 0 Mandibular canine 1 4 0 0 Mandibular premolar 1 5 4 0 0 Mandibular premolar 2 5 5 0 5* Mandibular molar 1 2 2 0 0 Mandibular molar 2 2 2 0 6* (6*)# Mandibular molar 3 3 1 6 (6 *)# 5 I — postmortem loss, socket present; 2 — antemortem loss, socket resorbed; 3 — little; 4 — dentine visible; 5 — cusps gone; 6 — pulp exposed; • — not available; * — poorly preserved; # — two specimens Table 2 Determination of Sex Characteristic Measurement Indication Reference Humeral Head Diameter Female Krogman 1962:144 vertical 38.4mm horizontal 36.0mm Clavicle Length 138.4mm Female Krogman 1962:148 Scapular Dimensions length of glenoid 33.6mm width of glenoid 24.3mm width of axillary border 3cm below infraglenoid 6.5mm Mandibular Incisor and Canine Z=7.2423 Female Ditch and Rose 1974 female. This finding is consistent with the general lack of robustness which characterizes the remains. The only contradictory evidence is the squareness of the chin which appears somewhat masculine. It must be emphasized, however, that all available indicators of sex are not only of relatively low reliability, but also are all measures of sexual dimorphism. A tentative age at death of 25 years for this individual has been based on the following criteria. Both maxillary and mandibular third molars had erupted. The right M3 was in occlusion, but is only slightly worn (enamel polished) implying an age of 17-25 years (Brothwell 1965:69). The epiphysis of the head of the humerus was completely fused indicating an age in excess of 24 years (Bass 1971:1 16); as were both epiphyses of the clavicle indicating 25-28 years (Krogman 1962:32). Small sections of the spheno-frontal and spheno-squamosal sutures are present, but disarticulated, perhaps indicating a relatively low degree of suture closure. No lipping of the glenoid of the scapula is observed (lipping normally commencing at age 30-35, Krogman 1963:55) nor are arthritic changes of the vertebrae evident. 58 Prior No. 44 Discussion The fragmentary nature of this material makes comparison with material from sites of similar antiquity practically impossible. The few measurements possible are of relatively little comparative interest. The relatively moderate tooth wear stands in distinction to the finding of extreme wear reported by Angel (1966) for the Tranquility Site skeletal material. It is unknown whether this discrepancy is due to a dietary difference or the relatively young age of the Squaw Rockshelter individual. Perhaps the ante-mortem loss of M[ and M2 is indicative of the effects of a similarly coarse diet. The skeleton is in general lightly built and rather petite. There are no observable pathologies except those associated with the mandible and mandibular dentition. There is indication of neither trauma nor cause of death. The predominance of remains from the right side would indicate that the individual was buried lying on her right side. A Acknowledgements The author thanks Professor Joseph Katich, CWRU School of Dentistry, for independent classification of dental remains. References Angel, Lawrence. 1966. Early Skeletons from Tranquility, California. Smithsonian Contributions to Anthropology 2(1). Bainbridge, D., and S. Genovese. 1956. A study of the sex differences in the scapula. Journal of the Royal Anthropological Society 86(2): 1 09- 1 34. Bass, William. 1971. Human Osteology: A Laboratory and Field Manual. Columbia: Missouri Archaeological Society. Brothwell, D. 1965. Digging Up Bones. London: The British Museum. Ditch, L., and J. Rose. 1974. A Multivariate Dental Sexing Technique. American Journal of Physical Anthropology 37:61-64. Krogman, Wilton. 1962. The Human Skeleton in Forensic Medicine. Springfield: C. C. Thomas. KIRTLANDIA The Cleveland Museum of Natural History March 1989 Number 44:59-62 Knotted Cordage from Squaw Rockshelter (33CU34), Aurora Run, Cuyahoga County, Ohio R.L. Andrews and J.M. Adovasio Department of Anthropology U niversity of Pittsburgh Pittsburgh, Pennsylvania 15260 Abstract A charred fragment of cordage from the late Middle Archaic level of Squaw Rockshelter represents a Z- spun, S-twist, 2-ply 4-yarn construction of bast fiber. A complex series of overhand knots cannot be attributed to specific function. With other specimens, this fragment documents an aboriginal textile industry of great antiquity in the Eastern Woodlands. Introduction A single piece of knotted cordage was recovered during the excavations at Squaw Rockshelter (33CU34). In the present context, cordage refers to a class of elongate fiber constructions which herein includes only knotted cordage. Methods The single artifact thus distinguished was classified according to the ply (1, 2, or more); the direction of initial spin (S or Z); and the direction of final twist (S or Z). Also, the specimen was scrutinized for the number and types of knots present. Knots were identified and described according to the terminology contained in Shaw (1972). All pertinent measurements were made with a Helios needlenose sliding dial caliper, and angles were measured with a protractor. Measurements taken follow the specifications of Emery (1966). The single piece of cordage recovered from the site is ascribable to one basic construction type which is described below. Description Type 1: Four-ply, Z-spun, S-twist Cordage Technique and comments This construction essentially consists of four Z-spun single elements with a final S-twist (Figure la, b). The specimen has been knotted at both ends in the following fashion. Initially, two unspun, single-ply elements were 60 Andrews and Adovasio No. 44 Figure 1. Reconstruction of Squaw Rockshelter Cordage: a — twisted yarns and piys; b — knotted element ; c — reconstruction groups and knots knotted near their passive ends with an overhand knot. These elements were laid next to another pair of unspun elements, and the free ends were then manipulated into another overhand knot. Of the eight emergent ends, six were further manipulated while two remained inactive. These elements extend from the knot in groups of two and four. At this point, a slight spin was imparted to those six elements (the term spin denotes only initial twisting and not spinning as that term is used in describing yarn preparation). The bundle of four elements was immediately divided into two groups of two each. The resultant pattern of free ends is 2-2-2, and these units are labeled exterior group, interior subgroup, and exterior subgroup, respectively. The interior subgroup was looped around the exterior group and subgroup as shown (Figure 1 c) and was finally combined with the exterior group of fibers. The exterior subgroup of elements was looped around and behind the other four elements, and its two plys were knotted into an overhand knot, then ended by wrapping one ply around the other and tucking under its loose end. The remaining four actively spun elements were again subdivided into two groups and S twisted. Each group functioned essentially as one element for ca. 2cm after which the subgroups were united again. The entire construction was then terminated by whipping. The final whipping knot is the functional equivalent of an overhand noose and is the fourth in a series of overhand knots. 1989 Knotted Cordage 61 The specimen is thoroughly charred and highly friable. The function of this construction is unknown. Raw material All elements of this specimen are constructed with an unknown genus and species of plant material. Fred H. Utech (1980, pers. com.) suggested that this raw material consisted of grass or sedge leaves. Richard I. Ford (1987, pers. com.) however, suggested to Brose that the material was some sort of bast fiber. More specifically, James K. Bissell and David S. Brose (1987, pers. com.) suggest the specimen was made of cedar {Juniperus spp.) or hemlock ( Tsuga spp.) root. Pollen diagrams in the region of the site indicate that although cedar and spruce were moderately abundant at ca. 10,000 B.P., they had vanished by 6500 B.P. whereas hemlock appeared ca. 7500 B.P. and is still a ravine dominant (Brose, this volume). The method of preparation of the raw material is unknown. Length — 6.40cm Range in diameter of individual Z-spun elements — 1.75-2. 70mm Mean diameter of individual Z-spun elements — 2.23mm Range in diameter of two-ply units — 2.70-4. 65mm Mean diameter of two-ply units — 3.68mm Range in diameter of finished four-ply cordage — 6.95-7. 05mm Mean diameter of finished four-ply cordage — 7.00mm Number of twists per cm — 1 Angle of twist: 28° Provenience and Chronology This single specimen of cordage is ascribable to a level in excavation unit 2 of the shelter which has been radiocarbon dated at ca. 5300 B.P. Internal Correlations Despite the paucity of the data, a number of observations can be made on the cordage construction from Squaw Rockshelter. First, the single specimen is de facto proof that a relatively sophisticated cordage industry existed in this part of the Midwest in the mid-fourth millennium B.C. Second, the specimen is technically well- made and exhibits all of the “earmarks” of a complex perishables industry. Specifically, the diameter of the individual elements is remarkably consistent. The knots, though simple, are combined in a complicated fashion, and the overall impression of the piece suggests that its makers were more than casually familiar with the manipulation of string and rope. The function of this particular specimen is unknown, but it could represent one of a variety of items, such as knotted fringe from a skirt or loin covering, a bundle of twined basketry construction material, a part of a snare or trap, or, perhaps, a “doodle” fashioned for no specific purpose. Whatever its function, we stress that this item was manufactured by an individual who was no stranger to the production of high-quality cordage. External Correlations The specimen under discussion represents at once the oldest well-dated piece of cordage recovered in the state of Ohio as well as one of the older pieces of cordage ever recovered east of the Mississippi River. Far older cordage materials are known from western North America where they are ascribable to at least the 1 1th millennium B.C. (see Adovasio 1974, 1977). Other early specimens are known from portions of Mesoamerica (Taylor 1966; MacNeish et al. 1967) and South America (Adovasio and Lynch 1973). Actual specimens of cordage or, indeed, any other kind of perishables from the eastern United States are relatively uncommon. Though it has been widely assumed that the perishable industries of Archaic populations east of the Mississippi were as old and as complex as those in the arid West, actual proof is somewhat difficult to acquire because of factors of preservation. This is the principal value of the piece from Squaw Rockshelter as well as the scattered perishable remains from localities with Paleolndian and/or Archaic period components, such as Meadowcroft Rockshelter, Pennsylvania (Stile 1982); Petit Anse Island, Louisiana (Wilson 1888; Figure C VI 1 ); Graham Cave, Missouri (Logan, 1952); Icehouse Bottom, Tennessee (Chapman and Adovasio 1977); the Harts Falls site, Maine (James L. Petersen 1984, pers. com.); Russell Cave, Alabama (Griffin 1974); the Long Branch site, Alabama (Webb and DeJarnette 1942); Salts Cave, Kentucky (King 1974); the Ozark Bluff Shelters, Arkansas (Schotz 1975); the Picton site, Ontario, Canada (Ritchie 1949); the Riverside site, Michigan (King 1968); and the Windover site, Florida. Cordage remains per se are limited in distribution to Icehouse Bottom (Chapman and Adovasio 1977; 622), Salts Cave (King 1974), the Picton site (Ritchie 1949), and the Ozark Bluff Shelters (Scholtz 1975). However, multiple-ply warp and/or weft elements produced via cordage- making techniques are identifiable in twined basketry and/or fabric from a variety of Archaic period sites including Graham Cave (Logan 1952: Plate XXlc), Icehouse Bottom (Chapman and Adovasio 1977), Salts Cave (King 1974), the Riverside site (King 1968), and Windover. While both final S- and/or Z-twist cordage and cordage warps/wefts are represented in the aforementioned assemblages, the diversity of both geographic and chronological distributions precludes any statistically significant correlation of twist types. Taken as a unit, these widely distributed remains indicate that populations east of 62 Andrews and Adovasio No. 44 the Mississippi did possess a complex and technologically sophisticated cordage and basketry industry of an antiquity comparable to those in the Desert West. It is also interesting to note that although the production of cordage and basketry from grass stems is ethnographically well documented in the Great Lakes region (Jones 1936; Douglas 1939), the use of “soft,” bast fibers is widespread among both prehistoric and ethnographic populations (Driver and Massey 1957; Spencer and Jennings et al. 1977; Holmes 1896; Mason 1904; Swanton 1946). Unfortunately, the condition of the Squaw Shelter specimen precludes both an exact identification of raw material and also the “method(s)” of its preparation. Presumably, the recovery of more materials such as the piece described here will greatly augment the very dim picture we presently possess of the production of perishables in portions of eastern North America. A Acknowledgements Special thanks are extended to Dr. Fred H. Utech, Botany Department, Carnegie Museum of Natural History, Pittsburgh; Dr. Richard I. Ford, Ethnobotanical Laboratory, Museum of Anthropology, University of Michigan; and James K. Bissell, Department of Botany, Cleveland Museum of Natural History, for suggestions concerning the identification of plant materials reported in this study. References Adovasio, J. M. 1974. Prehistoric North American Basketry. Nevada Slate Museum Anthropological Papers 16:100-145. Carson City. 1977. Basketry Technology. Chicago: Aldine Publishing Company. Adovasio, J. M., and T. F. Lynch. 1973. Prehistoric Textiles and Cordage from Guitarrero Cave, Peru. American Antiquity 38 ( l):84-90. Chapman. J., and J. M. Adovasio. 1977. Textile and Basketry Impressions from Icehouse Bottom, Tennessee. American Antiquity 42(4):620-625. Douglas, F. H. 1939. Indian Basketry East of the Rockies. Denver Art Museum Leaflet 87. Driver, H. E., and W. C. Massey. 1957. Comparative Studies of North American Indians. Transactions of the American Philosophical Society. New Series 47(2). Emery, I. 1966. The Primary Structure of Fabrics, an Illustrated Classification. Washington: The Textile Museum. Griffin, J. W. 1974. Investigations in Russell Cave. U.S. Department of the Interior, National Park Service Publications in Archaeology 13. Washington D.C. Holmes, William H. 1896. Prehistoric Textile Art of Eastern United States. Thirteenth Annual Report of the Bureau of American Ethnology, 1891-1892. Washington, D. C. Jones, V, H. 1936. Some Chippewa and Ottawa Uses of Sweet Grass. Papers of the Michigan Academy of Science, Arts, and Letters 2 1 . King, M. E. 1968. Textile Fragments from the Riverside Site, Menomine, Michigan. Verhandlungen des XXXVIII internationalen Amerikanistendongresses I: 1 17-123. 1974. The Salts Cave Textiles: A Preliminary Account, pp. 31- 40 in Archaeology of the Mammoth Cave Area, edited by P. J. Watson. New York: Academic Press. Logan, W. D. 1952. Graham Cave: An Archaic Site in Montgomery County, Missouri. Missouri Archaeological Society Memoir 2. MacNeish, R. S., A. Nelkin-Terner, and 1. W. Johnson. 1967. The Prehistory of the Tehuacan Valley. Volume 11. Austin: University of Texas Press. Mason, O. T. 1904. Aboriginal American Basketry: Studies in a Textile Art without Machinery. Report of the U.S. National Museum for 1902: 171-548. Ritchie, W. A. 1949. An Archaeological Survey of the Trent Waterway in Ontario, Canada and Its Significance for New York State Prehistory. Researches and Transactions of New York State Archaeological Association 12(1). Albany. Scholtz, S. C. 1975. Prehistoric Plies: A Structural and Comparative Analysis of Cordage, Netting, Basketry, and Fabric from Ozark Bluff Shelters. Arkansas Archaeological Survey Research Series 9. Shaw, G. R. 1972. Knots. New York: Collier Books. Spencer, R. F., and J. D. Jennings, et al. 1977. The Native Americans. New York: Harper and Row. Stile, T. E. 1982. Perishable Artifacts from Meadowcroft Rockshelter, Washington County, Southwestern Pennsylvania, pp. 130-141 in Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage, edited by R. C. Carlisle and J. M. Adovasio. Department of Anthropology, University of Pittsburgh. Swanton, J. R. 1946. The Indians of the Southeastern United States. Bureau of American Ethnology Bulletin 137. Washington, D.C. Taylor, W. W. 1966. Archaic Cultures Adjacent to the Northeastern Frontiers of Mesoamerica. pp. 59-94 in Handbook of Middle American Indians, Volume 4: Archaeological Frontiers and External Connections, edited by G. F. Ekholm and G. R. Willey, pp. 59-94. University of Texas Press, Austion and London. Webb, W. S., and D. L. DeJarnette. 1942. An Archaeological Survey of Pickwick Basin in the Adjacent Portions of the States of Alabama, Mississippi, and Tennessee. Bulletin of the Bureau of American Ethnology 1 29. Washington, D.C. Wilson, T. 1888. Ancient Indian Matting - From Petit Anse Island, Louisiana. Report of the U.S. National Museum for 1888: 673-676. M0( IkT o ■NATURAL HISTORY- Published by The Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Design and Production by Melody Oakes KIRTLANDIA Invertebrate Paleontology Stratigraphic Distribution of Brachiopods and Bivalves in the Upper Devonian (Famennian) Chagrin Shale in the Cuyahoga River Valley , Northeast Ohio 7 Barbara A. Schwimmer and Rodney M. Feldmann Zoology Estimation of Numbers for a Riverine Necturus Population Before and After TFM Lampricide Exposure 33 Timothy O. Matson •NATURAL HISTORY* JUNE, 1990 KIRTLANDIA The Scientific Publication of The Cleveland Museum of Natural History David S. Brose and Joseph T. Hannibal. Editors Brief History and Purpose: Kirtlandia, a publication of The Cleveland Museum of Natural History, is named m honor of Jared Potter Kirtland, a noted nineteenth-century naturalist who lived in the Cleveland, Ohio area. It began publication in 1967 and is a continuation of the earlier series Scientific Publications volumes 1 to 10 (1928-1950), and new series volumes 1 to 4 (1962-1965). Supported by the Kirtlandia Society of The Cleveland Museum of Natural History, Kirtlandia is devoted to the publication of scientific papers in the various fields of inquiry within the Museum’s sphere of interest: Cultural and Physical Anthropology; Archaeology; Botany; Geology; Paleobotany; Invertebrate and Vertebrate Paleontology; Systematic Ecology; and Invertebrate and Vertebrate Zoology. Issues will vary from single monographs to collections of short papers, review articles, and brief research notes. Kirtlandia is abstracted in Biological Abstracts and indexed in Bibliography and Inde x of Geology and Zoological Record. Kirtlandia No. 45 ISSN 0075-6245 © 1990 by The Cleveland Museum of Natural History Cleveland, Ohio KIRTLANDIA The Cleveland Museum of Natural History June 1990 Number 45 Invertebrate Paleontology Stratigraphic Distribution of Brachiopods and Bivalves in the Upper Devonian (Famennian) Chagrin Shale in the Cuyahoga River Valley, Northeast Ohio 7 Barbara A. Schwimmer and Rodney M. Feldmann Zoology Estimation of Numbers for a Riverine Necturus Population Before and After TFM Lampricide Exposure 33 Timothy O. Matson KIRTLANDIA Editors David S. Brose Joseph T. Hannibal The Cleveland Museum of Natural History Associate Editors James K. Bissell, Curator of Botany The Cleveland Museum of Natural History Bruce Latimer, Curator of Physical Anthropology The Cleveland Museum of Natural History Martin Rosenberg, Editorial Assistant Case Western Reserve University Sonja Teraguchi, Manager of Collections The Cleveland Museum of Natural History Editorial Advisory Board Rodney Feldmann, Professor of Geology Kent State University Michael C. Hansen, Geologist Ohio Geological Survey Richard Meindl, Associate Professor of Anthropology Kent State University G. Michael Pratt, Associate Professor of Anthropology Heidelberg University David H. Stansbery, Director, Museum of Zoology Ohio State University Frederick H. Utech, Curator of Botany Carnegie Museum of Natural History Ed Voss, Curator of the Herbarium University of Michigan Andrew M. White, Professor of Biology John Carroll University KIRTLANDIA The Cleveland Museum of Natural History June 1990 Number 45 :7 -31 Stratigraphic Distribution of Brachiopods and Bivalves in the Upper Devonian (Famennian) Chagrin Shale in the Cuyahoga River Valley, Northeast Ohio Barbara A. Schwimmer * and Rodney M. Feldmans Department of Geology Kent State University Kent, Ohio 44242 Abstract The Upper Devonian Chagrin Shale, exposed in the Cuyahoga River Valley in northeast Ohio, has a relatively rich, brachiopod-dominated fauna. Two stratigraphic sections were sampled to determine the distribution of brachiopods in these shale and siltstone beds. Nineteen brachiopod and four bivalve taxa have been described from the unit in this study. Conodont faunas, and to a lesser extent, brachiopod ranges, provide evidence of a late Famennian age for the Chagrin Shale The Chagrin sediments were deposited on a relatively shallow shelf, below normal wave base. Sedimentation rates were slow, energy was low, and the substrate was fluid. Turbidity was at least moderate, as evidenced by dominant colonization by brachiopods instead of bivalves. Episodic storms carried sediments westward from the prograding Catskill Delta complex, leading to the periodic burial of brachiopod communities. The low degree of fragmentation and lack of abrasion of specimens suggest rapid burial rates. Preservation was influenced by the processes of dissolution, pyritization, and phosphatization. * Present address : The Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 8 SCHWIMMER AND F ELD MANN No. 45 K M i j FIGURE 1. Upper Devonian shale outcrop pattern in northeastern Ohio, showing the locations of the two study sites in the Cuyahoga River Valley. Introduction The purpose of this study is to describe brachiopod and bivalve faunas and their stratigraphic distribution in the Chagrin Shale from Chippewa Creek at Brecksville Reservation on the western side of the Cuyahoga River Valley, and from Brandywine Creek at Brandywine Falls, on the eastern bank (Figure 1). The brachiopod fauna supports a late Famennian age for the Chagrin Shale. The Chagrin Shale is exposed across 150 kilometers of northeast Ohio, along the southern shore of Lake Erie, and along many of the streams and their tributaries which flow into the lake (Figure 1). The Chagrin attains a maximum thickness of 400 m at the Ohio-Pennsylvania border, and thins to an estimated 35 meters in the Cleveland area according to Szmuc et al. (1976). Pepper et al. (1954) described the Chagrin as a thick wedge of interbedded siltstones and shales, bounded above by the Cleveland Shale, and below by the Fluron Shale, and pinching out between these two black shales in a westward direction. The Cuyahoga River flows through the western part of the Upper Devonian shale outcrop area, where its valley is floored by the Chagrin Shale. From Cuyahoga Falls, Summit County, the river travels northwest across Summit and Cuyahoga Counties in a pre-glacial valley, characterized by steep hills and cliffs for many miles. The Cuyahoga River tributaries often cut valleys or gorges through the hills, in which many of the best Chagrin Shale exposures are seen. Various aspects of the Chagrin Shale have been studied in recent years: the petrography (Broadhead and Potter, 1980); facies relationships (Potter et al., 1980); sedimentation (Potter et al., 1980); trace fossils (Feldmann et al., 1978; Hannibal and Feldmann, 1983; Stukel, 1987); and arthropods (Weidner and Feldmann, 1983). Previous reports (Prosser, 1912; Cushing et al., 1931; Hannibal and Feldmann, 1983; and Weidner and Feldmann, 1983) cited the sparsely fossiliferous nature of the Chagrin Shale. While faunal elements are restricted, recent studies, listed below, concluded that the Chagrin fauna is more diverse and abundant than previously supposed. Stukel (1987) described fifteen ichnotaxa from the unit; Weidner and Feldmann (1983) reported several species of arthropods found in the Chagrin; and Feldmann et al. (1986) identified fossil worms from the unit. The most abundant megafaunal elements appear to be brachiopods, with bivalves as a minor constituent of the total fauna. Localities Locations of the two sections studied are shown in Figure 1. Chippewa Creek, located on the western side of the Cuyahoga River Valley, flows through Brecksville Reservation, one of the Cleveland Metroparks. The Brecksville section is located in the west-central 1/9 of the Northfield, Ohio 7.5 min. topographic map (1963, photorevised 1979). The section begins on the southern side of the creek, at the base of the Bridle Path, and extends upstream, approximately 900 m, to a steep cliff on the north side of the creek. The upper part of the measured section is approximately 170 m north of the Trailside Interpretive Center on Chippewa Creek Road. Brandywine Creek, part of the Cuyahoga Valley National Recreation Area, lies on the eastern side of the Cuyahoga River Valley. It is located in the southeast 1/9 of the Northfield, Ohio 7.5 min. topographic map (1963, photorevised 1979). From the intersection of Brandywine Road and 1-271 on the map, the measured section is 580 m west on Stanford Road, and 50 m north into the park area. The Chagrin crops out on the banks of Brandywine Creek for almost a kilometer, ending west of a waterfall, approximately one-half km upstream. Lithologies The Chagrin Shale consists of interbedded greenish- gray to bluish-gray shales, and discontinuous siltstones. The siltstones range in thickness from one to tens of centimeters, generally increasing in number and thickness eastward, toward the Ohio-Pennsylvania border (Stukel, 1987). At Brecksville Reservation, the Chagrin Shale exposure measures approximately 6 m from stream level, at the base of the Bridle Path, to the contact with the overlying Cleveland Shale. Here the Chagrin consists of 1990 Brachiopods and Bivalves in the Chagrin Shale 9 generally fissile, greenish-gray shales, current-rippled, cross-laminated siltstones, and plane-parallel laminated siltstones (Stukel, 1987). Near the contact with the Cleveland Shale, the Chagrin at Brecksville exhibits several thin, horizontal, discontinuous pyrite beds. Pyrite is abundant in other forms near the contact, and is manifested in concretions, as nodules, as coatings on fossil molds, and as vertical burrow fillings. Brachiopods at Brecksville occur throughout the shales and at the bases and tops of many of the siltstone beds. Fossils are generally not found within the siltstones themselves. In contrast, the Chagrin Shale at Brandywine Falls measures about 10 m in thickness, and consists of bluish- gray shales with interbedded, discontinuous siltstones. Siltstones exhibit current-rippled, cross-laminated bedding, plane-parallel bedding, and very gentle hummocky cross- stratification. At Brandywine the types of pyrite found at Brecksville are typically absent. Pyrite occasionally occurs as coatings on brachiopod molds. Brachiopods at Brandywine are distributed throughout the shales, and at the bases and tops of some siltstone beds. Again, fossils are rare within the siltstones. In general, the units within the Chagrin at the Brecksville Reservation and Brandywine Falls sites cannot be correlated on the basis of lithology. The discontinuous nature of the siltstone beds, and the absence of horizontal pyrite layers at Brandywine Falls preclude using these parameters for correlation. There is no distinctive bedding feature of the shales or siltstones that permits a bed to bed correlation between sections. Some patterns of faunal distribution only provide an approximate correlation between the two sections. Therefore the two sections must be correlated using the Chagrin Shale-Cleveland Shale contact as a datum, even though it is erosional in nature. Any differences in elevation between the two sites, which are approximately 9 km apart, might be explained in terms of regional dip and/or differential rates of erosion at the sites. The effect of regional dip is minimized because the two sections are in close proximity. Methods Two sections in the Cuyahoga River Valley were sampled to determine the stratigraphic distribution of brachiopods and bivalves (Figures 2 and 3). Each section was trenched and sampled on a bed by bed basis using a pocket knife to remove individual shale and siltstone units. The trench at Brecksville Reservation exposed approximately 6 m of section, while the trench at Brandywine Falls exposed approximately 10 m of section. The trenched surface area averaged 20 cm by 30 cm. At each level where brachiopods and bivalves were discovered, collections were made, fossil orientation noted, and distance from the datum was measured. No attempt was made to determine numerical abundance at each level. Rather, generic identification and stratigraphic position were the primary goals. Paleontology Preservation The Chagrin Shale megafauna in the Cuyahoga Valley consists almost exclusively of brachiopods, but includes a number of bivalves as well. Original shell material and internal structures of brachiopods are often preserved, notably of the spiriferids, orthids, and inarticulates. Densities of fossils in the Chagrin vary; often a bedding surface will contain clusters of one or two species, or randomly spaced taxa, and other surfaces may be barren. The degree of valve articulation is moderate, and the degree of fragmentation is low. Many specimens occur in life position, while others occur in more hydrodynamically stable positions. Some fossils have original shell material preserved so that delicate ornamentation is visible. Others are well- preserved molds of the interior, exhibiting muscle scars and other structures. However, most of the brachiopods are characterized by poorly preserved molds of the interior. Juvenile and adult stages were observed in most species. Evidence of predation is lacking, and epibionts are rare. In some forms, shell material is absent, and molds of the exterior are rare. Sometimes critical characters, such as cardinalia and interareas, are not available for identification. While some of this may be due to selective positioning after death, or to some disarticulation and fragmentation in the normal course of events between storm activities, diagenesis should also be considered. Rapid burial may be responsible for a variety of chemical reactions, including skeletal dissolution, pyritization, and phosphatization of some faunal elements. In the Chagrin brachiopods, skeletal material is often dissolved, leaving only molds or casts. Burial in organic rich muds often accelerates dissolution (Brett and Baird, 1986), probably due to the acidic nature of the sediments. Pyrite assumes a variety of forms in the Chagrin Shale, including concretions, nodules, burrow infillings, and thin coatings on fossil molds. Since most bottom waters contain some dissolved oxygen, Berner (1984) indicated that pyrite usually precipitates in anoxic environments below the sediment-water interface. During periods of rapid burial, organic material is exposed to interstitial waters rich in reactive iron compounds. Together, these are subject to sulfate reduction by anaerobic bacteria, and pyrite is formed. In some of the Cuyahoga Valley brachiopods, pyrite occurs as thin coatings or as crystals encrusting fossil molds. From the sporadic occurrence of pyrite associated with fossils, it can be concluded that anaerobic microenvironments were locally present in the sediments. FIGURES 2 (BreCKSVILLE) AND 3 (Brandywine Falls). Distribution patterns of taxa at Chippewa Creek, Brecksville Reservation, and at Brandywine Creek. Brandywine Falls, respectively. Black intervals represent siltstones, and white intervals represent shales. Twenty— cm sample intervals are indicated directly to the right of the stratigraphic column. 10 SCHWIMMER AND FELDMANN No. 45 interval Acanthatia ? Ambocoelia Athyris - Composita Aulacella Cent ror hynchus Cyrtospirifer "Leiorhynchus" Lingula Orbiculoidea productellid gen indet Retichonetes Schellwienella Sphenospira Spinospiriter Toryniferella Trigonoglossa ? Leptodesma interval Acanthatia ? Ambocoelia Athyris- Composita Aulacella Centrorhynchus Cyrtospirifer "Leiorhynchus" Lingula Orbiculoidea productellid gen. indet Retichonetes Schellwienella Sphenospir a Spinospiriter Toryniferella Trigonoglossa ? Leptodesma Pterinopecten ? BRECKSVILLE BRANDYWINE FALLS 1990 Brachiopods and Bivalves in the Chagrin Shale 11 It must be noted that several thin (<1 cm), horizontal pyrite layers were found near the erosional Chagrin Shale- Cleveland Shale contact at Brecksville Reservation. The topmost of these is the Skinner’s Run pyrite bed, also found elsewhere in the Chagrin outcrop area, at the contact with the overlying black Cleveland Shale. Pyritized material was exhumed and reworked during periods of erosion on normally anaerobic sea floors (Zagger and Banks, 1989). The Skinner’s Run pyrite bed may, in turn, be analogous to the Leicester Pyrite Member, which is associated with a regional unconformity {Baird and Brett, 1986). Phosphatic concretions are sometimes found in the Cuyahoga Valley, but can be found in much greater numbers at numerous sites in the central and eastern sections of the Chagrin outcrop area (Schwimmer et al., 1987). Manheim et al. (1975) listed conditions necessary for formation of modern marine phosphorite. These included: sediments with high TOC (total organic carbon) associated with dysaerobic waters, low rates of sedimentation and a low volume of calcium carbonate in sediments. Slansky (1986) noted that certain modern phosphates result from the reworking and concentration of existing phosphatic bioclasts, such as inarticulate brachiopod shells, arthropod remains, or bone fragments. Apatite precipitation may be triggered at one or more “nucleation sites,” which initially become coated with humic acids. Phosphate precipitation is catalyzed by the acids, and not directly by the phosphatic surface of the organism. Once the process is initiated, precipitation continues, due to the lateral diffusion of ions from interstitial waters. Wetzel (1983, p. 264) stated that, if sediments are anoxic, phosphates may be released from sediments, diffused into surface waters and, over time, would be lost. However, if an oxidized zone is present near the sediment-water interface, then phosphate could be reduced by ferrous iron and precipitated as ferric phosphate and by absorption onto ferric hydroxide and calcium carbonate (Wetzel, 1983, p. 261-3). Phosphate formation is favored by rapid burial of organic material, followed by long periods of little or no sedimentation (Brett and Baird, 1986). Wet chemical tests and analysis by energy-dispersed x-rays (EDX) have determined that concretions found in the Chagrin are phosphatic. The surrounding silts and shales tested negative for phosphorus. Generally, phosphatic material such as lingulid brachiopod shells, arthropod remains, and fish fragments acted as nucleation sites for phosphate formation in Chagrin concretions. EDX analyses also showed that non-phosphatic material in the concretions, such as pelmatozoan fragments, and the matrix itself, are enriched in phosphate (Joseph Hannibal, pers. comm.). In summary, the Chagrin brachiopods show varying qualities of preservation. Some display shell material and delicate internal structures. Others are less well preserved, and are oriented so that vital identifying characters are either absent or obscured. Often molds are the only evidence of ancient life, and these cannot always be removed from the surrounding matrix to expose undersurfaces. The unbroken, articulated appearance of many of the specimens, as well as their orientation, distribution, and density, is evidence for quiet water conditions, with periodic burial of life assemblages by storm sediments. Chemical alteration includes shell dissolution, pyrite formation in local anaerobic microenvironments, and phosphatization during times of low sedimentation, after episodes of rapid burial. Summary oftaxa The Cuyahoga Valley Chagrin fauna consists predominantly of brachiopods, with fewer bivalves. The description of taxa is designed not as a formal systematic description of the fauna, but as a guide for identification. The most important diagnostic features are presented along with remarks about preservation, distribution, and comparisons with similar taxa. The vertical distribution of brachiopods and bivalves at Brecksville Reservation and at Brandywine Falls are shown in Figures 2 and 3, respectively. The black intervals in each stratigraphic section represent silts, while the white intervals represent shales. For ease in charting distributions, each stratigraphic section was arbitrarily divided into consecutively numbered 20-cm intervals, shown to the right of each column. Distribution patterns discussed for each taxon will refer to these figures. All figured specimens are deposited in The Cleveland Museum of Natural History (CMNH). The remainder of the material from this study is deposited in the paleontology collection in the Department of Geology at Kent State University (KSU). Six orders of brachiopods have been identified in the Upper Devonian Chagrin Shale of the Cuyahoga Valley. The Inarticulata are represented by Lingulida and Acrotretida, while the Articulata are represented by Orthida, Strophomenida, Rhynchonellida, and Spiriferida. Four taxa of bivalves, representing three orders, are also presented. Systematic paleontology Phylum Brachiopoda Dumeril, 1806 Class Inarticulata Huxley, 1869 Order Lingulida Waagen, 1885 Superfamily Lingulacea Menke, 1828 Family Lingulidae Menke, 1828 Genus Lingula Bruguiere, 1797 Lingula eriensis Girty, 1939 Figures 4.5a, b, 4.6 12 SCHWIMMER AND FELDMANN No. 45 FIGURE 4. Inarticulata. 1-2, Orbiculoidea sp. 1, pedicle interior, Mill Creek, KSU 4899; 2, mold of brachial valve interior, Brecksville Reservation, interval 6, CMNH 8401 ; 3a, b. Lingula arcta, partially exfoliated pedicle valve in concretion, Brandywine Falls, interval 59, CMNH 8402; 4, Trigonoglossa sp., partial pedicle valve, Brecksville Reservation, interval 17, CMNH 8403; 5-6, Lingula eriensis. 5a,b, partially exfoliated specimen in concretion, Brandywine Falls, interval 67, CMNH 8404; 6, specimen found in living position, valves splayed, Brecksville Reservation, interval 20, CMNH 8405. Bar scale = 1 cm. Description of material Shell medium size, ovate, anterior end rounded, posterior end slightly pointed. Length approximately 1.2- 1.5 times width. Valves generally flat, slightly biconvex in umbonal regions. Lateral margins rounded. Surface ornamentation consists of irregularly spaced growth lines, and very faint radial striations. Exfoliated specimens show alternating chitinous and phosphatic shell layers. Chitinous shell layers marked by closely spaced, regular, concentric lines of a different character from the external growth lines. Pedicle valve interior with cordate visceral area highlighted by curved, transverse striations. Central muscle scars ovate. In some specimens, pedicle groove is seen at posterior end. Brachial valve interior displays median ridge which extends almost entire length of shell. Muscle scars not visible. Remarks Lingulids are rare and occur either as sets of splayed valves in the shales (Figure 4.6), or in concretions in the siltstones. Although original shell material is present, often it is only partially preserved, and internal structures cannot be seen. Lingulids are far more common in the eastern part of the Chagrin outcrop area, where sediments are coarser. The burrowing lifestyle of lingulids is more suitable to coarser sediments than to fine grained muds (Thayer and Steele-Petrovi'c, 1975). The low bulk density of lingulids would cause them to float on a fluid substrate, unable to assume their normal feeding position. Lingulids were most abundant in the lower half of both the Brecksville Reservation and Brandywine Falls sections (Figures 2 and 3). Girty (1939) reported L. eriensis and L. arcta Girty from the Chagrin Shale at Brecksville Reservation. However, the two species of Lingula found at these study sites cannot always be distinguished in the field, therefore distribution patterns outlined in Figures 2 and 3 reflect generic and not species patterns. Lingula arcta Girty, 1939 Figures 4.3a, b Description of material Small size, elongate oval outline, length greater than width. Lateral margins parallel. Surface ornament consists of very fine, evenly spaced, concentric growth lines. Pedicle valve umbo elevated posteriorly, and slightly pointed. Brachial valve and valve interiors unknown. Remarks One good specimen of a partially exfoliated pedicle valve was recovered from an in situ concretion at the Brandywine Falls section (Figure 4.3). Girty (1939) identified this species from the Chagrin Shale at Brecksville Reservation, and noted that it often occurred with L. eriensis. He stated that the two species were distinct, based upon shape and surface ornamentation. Growth lines on L. arcta are very fine and closely spaced. 1990 Brachiopods and Bivalves in the Chagrin Shale 13 FIGURE 5. Aulacella sp. !, interior, pedicle valve, Brecksville Reservation, interval 28, CMNH 8406; 2, interior mold of brachial valve, Brecksville Reservation, interval 27, CMNH 8407; 3a, b, posterior view, and brachial valve exterior, Brecksville Reservation, float, CMNH 8408 ; 4, brachial valve interior, Brecksville Reservation, interval 26, CMNH 8409; 5, pedicle valve interior, Brecksville Reservation, float, CMNH 8410; 6, interior mold of pedicle valve, Brecksville Reservation, interval 28, CMNH 8411 ; 7, pedicle valve exterior, Brecksville Reservation, float, CMNH 8412. Bar scale = 1 cm. while those on L. eriensis are coarser, farther apart, and less regular. Girty (1939) also reported shell fragments of what could be L. limatnla from the Chagrin Shale at Brecksville. The latter species is characterized by a markedly papillose shell surface. Neither of the lingulids collected from the Chagrin Shale in this study exhibit this character. Genus Trigonoglossa Dunbar and Condra, 1932 Trigonoglossa sp. Figure 4.4 Description of material Shell medium size, triangular outline, gently convex pedicle valve, brachial valve unknown. Posterior end pointed. Length greater than width; greatest width in anterior part of specimen. The axial region of the specimen appears to be inflated, forming a gentle fold. Valve surface marked by strong, evenly spaced, concentric lines prominently elevated above surface of the valve, separated by flat interspaces. Very faint radial striations. Valve interiors unknown. Remarks One partial, somewhat crushed specimen was recovered from Brecksville Reservation. The prominent growth lines and triangular shape distinguish it from species of Lingula. Order Acrotretida Kuhn, 1949 Suborder Acrotretidina Kuhn, 1949 Superfamily Discinacea Gray, 1 840 Family Discinidae Gray, 1840 Subfamily Orbiculoideinae Schuchert, 1929 Genus Orbiculoidea d’Orbigny, 1847 Orbiculoidea sp. Figures 4.1, 4.2 Description of material Small to large size, circular outline, convex brachial valve, concave pedicle valve. Brachial valve apex pointed, posteriorly eccentric; diameter of figured specimen (Figure 4.2) approximately 3.2 cm. Ornament includes strong, elevated, evenly spaced, concentric growth lines. Smooth, wide depressions separate growth lines. Very faint radial striations are present, especially in the apical area, dying out anteriorly. Posterior margin rounded, not truncated. Muscle scars and other internal structures not observed. Remarks Three occurrences of Orbiculoidea were noted only in the section at Brecksville (Figure 2). One partial brachial valve, found atop a silt bed, was complete enough for identification. In contrast, Weidner (1983) found many well preserved orbiculoids in concretions in other parts of the Chagrin outcrop area. The pedicle valve of one such specimen from Mill Creek is illustrated (Figure 4. 1 ) for comparison. The 14 SCHWIMMER AND F ELD MANN No. 45 FIGURE 6. Schell wienella sp. 1, interior mold of pedicle valve, with dental plate impressions, Brandywine Falls, interval 57, CMNH 8413; 2, posterior mold of interarea, deltidium and dental plates, Brandywine Falls, float, CMNH 8414; 3a. b. interior mold of pedicle valve, brachial valve, Brandywine Falls, float, CMNH 8415; 4, interior mold of pedicle valve and interarea, Brandywine Falls, float, CMNH 8416; 5, interior mold of pedicle valve, Brandywine Falls, interval 58, CMNH 8417; 6, interior mold of brachial valve, cardinal process, Brecksville Reservation, interval 24, CMNH 8418. Bar scale = 1 cm. size, shape, and ornamentation appear to closely resemble the Brecksville specimen (Figure 4.2). Pedicle foramen is narrow, linear; 8 mm length; posteriorly located with respect to the apex. Internal structures unknown. Class Articulata Fluxley, 1 869 Order Orthida Schuchert and Cooper, 1932 Suborder Orthidina Schuchert and Cooper, 1932 Superfamily Rhipidomellacea Alichova, 1960 Family Onniellidae Opik, 1933 Genus Aulacella Schuchert and Cooper, 1931 Aulacella sp. Figures 5. 1-5.7 Description of material Small subcircular shell; valves unequally biconvex; width greater than length; maximum width at mid-valve; hinge narrow, short, approximately half width of shell; delthyrium open; valve surfaces finely costellate; costellae extend from beak to anterior margin, and increase by intercalation; pedicle valve fold and brachial valve sulcus strong posteriorly, becoming weak anteriorly; anterior margin rectimarginate; valves inflated in areas of umbones, flattening laterally and anteriorly. Concentric ornamentation present near anterior margin. Pedicle valve interior with large, flabellate diductor muscle scars, which almost enclose small, spatulate adductor scars; a forked myophragm is seen anteriorly, separating diductor scars. Brachial valve interior with prominent, trilobed, cardinal process, and diverging brachiophore bases beneath elongate socket ridges; posterior flabellate, and anterior subquadrate, adductor muscle scars located on either side of dorsal fold. Remarks Aulacella can be distinguished from Cariniferella by the presence in the latter of a convexo-concave profile, a sharp fold, and a deep, narrow, sulcus (Schuchert and Cooper, 1932). Although Prosser (1912) reported species of Cariniferella in the Chagrin, none was found in this study. Aulacella is ubiquitous in the lower half of the Chagrin at Brecksville, decreasing in numbers and in occurrence in the upper half. In contrast, Aulacella is less abundant and is found in only seven intervals in the lower third of the section at Brandywine Falls, and in only one instance in the upper third (Figures 2 and 3). Aulacella is the only punctate taxon of Chagrin brachiopods. This feature may have enabled it to respire at reduced capacity, even when valves were shut during episodes of turbulence, which would normally clog the lophophore (Thayer, 1986; Shumway, 1982). Order Strophomenida Opik, 1934 Suborder Strophomenidina Opik, 1934 Superfamily Davidsoniacea W. King, 1850 Family Meekellidae Stehli, 1954 Subfamily Meekellinae Stehli, 1954 Genus Schellwienella I. Thomas, 1910 SCHELLWIENELLA Sp. Figures 6. 1-6.6 1990 Brachiopods and Bivalves in the Chagrin Shale 15 FIGURE 7. Retichonetes sp. la,b, partially exfoliated brachial valve with four hinge spines on either side of pedicle valve umbo, and enlarged view of pedicle valve interarea, Brecksville Reservation, float, CMNH 8419; 2, interior mold of pedicle valve, with costellae and growth lines creating reticulated appearance, Brandywine Falls, interval 59, CMNH 8420; 3, interior mold of brachial valve, median septum, Brandywine Falls, interval 73, CMNH 8421 ; 4, partially exfoliated, pyritized brachial valve and pedicle valve spines, Brandywine Falls, interval 62, CMNH 8422; 5, interior mold of pedicle valve, flabellate diductor scars, divergent dental plates, Brecksville Reservation, interval 19, CMNH 8423; 6, interior mold of brachial valve, Brecksville Reservation, interval 23, CMNH 8424; 7, mold of brachial valve, median septum, alveolus, and divergent lateral septa, Brandywine Falls, interval 73, CMNH 8425 ; 8, mold of brachial valve interior, pedicle valve interarea, Brecksville Reservation, interval 23, CMNH 8426. Bar scale - 1 mm. Description of material Small to medium size, semicircular outline, slightly resupinate profile; width greater than length. Valves generally flat, slightly convex in umbonal regions; hinge width slightly less than greatest shell width, which occurs at mid-length. Interarea straight, relatively high; triangular delthyrium covered by convex pseudodeltidium. Pedicle and brachial valves finely costellate; costellae increase by intercalation, and appear to curve gently as they approach posterior lateral margins; concentric ornamentation sparse. Fold and sulcus either weak or absent; anterior margin rectimarginate. Pedicle valve interior exhibits short dental plates diverging at wide angles; diductor muscle scars faint, large, flabellate. Brachial valve interior with small, bilobed, cardinal process, thin, short median ridge, bisecting slender, spatulate, adductor muscle scar. Remarks Schellwienella is distinguished from Schuchertella by the presence of short, diverging, dental plates. These plates are lacking in the latter genus. Most of the Chagrin specimens of Schellwienella are incomplete; generally the anterior and lateral margins are broken off, and original shell material is absent. In one specimen, however, (Figure 6.3a), the original shell material comprising the dental plates is present. A spirolophous lophophore was probable in all stocks of the superfamily Davidsoniacea, although impressions or calcareous supports are rare (Muir- Wood and Williams, 1965). Schellwienella is moderately abundant in the lower half of the Brecksville section, but occurs in only one interval in the upper half. At Brandywine Falls, this genus is more abundant, occurring in the lower two-thirds of the section, but with only one occurrence in the upper third of the site (Figures 2 and 3). Suborder Chonetidina Muir- Wood, 1955 Superfamily Chonetacea Bronn, 1862 Family Chonetidae Bronn, 1862 Subfamily Devonochonetinae Muir- Wood, 1962 Genus Retichonetes Muir- Wood, 1962 Retichonetes sp. Figures 7. 1-7.8 Description of material Small size, semicircular outline; gently biconvex; width greater than length; hinge line straight, and equal to greatest width of shell; interareas moderately high; minimum of four spines project at high angles (45-60 degrees) from pedicle valve interarea, adjacent to umbo; delthyrium covered by pseudodeltidium. Valve exteriors 16 SCHWIMMER AND FELDMANN No. 45 FIGURE 8. Productellids. 1-3, Productellid, gen. indet. 1, interior mold of pedicle valve, sulcus, anterior rugae, Brandywine Falls, interval 57, CMNH 8427; 2a, b, interior mold of pedicle valve, with beak and divergent “septa” , Brecksville Reservation, interval 24, CMNH 8428; 3, interior mold of pedicle valve, spine ridges, Brandywine Falls, float, CMNH 8429; 4-14, Acanthatia sp. 4, interior mold of brachial valve, Brecksville Reservation, interval 24, CMNH 8430; 5, interior mold of brachial valve, endospine impressions, Brandywine Falls, interval 58, CMNH 8431 ; 6, pedicle valve exterior, elongate hinge spines (arrow), prostrate body spines, borehole, Brecksville Reservation, interval 4, CMNH 8432 ; 7, mold of pedicle valve exterior, hinge and lateral spines, Brandywine Falls, float, CMNH 8433; 8, interior mold of brachial valve, cardinal process, breviseptum, Brandywine Falls, float, CMNH 8434; 9, Brachial valve exterior, spine impressions, Brecksville Reservation, interval 28, CMNH 8435; 10, exterior mold of pedicle valve, small ears, spine impressions, Brandywine Falls, interval 62, CMNH 8436; 11, interior mold of brachial valve, breviseptum, Brandywine Falls, float, CMNH 8437; 12, interior mold of brachial valve, bilobed cardinal process, endospine impressions, Brandywine Falls, float, CMNH 8438; 13, interior mold of brachial valve, alveolus; interior mold of pedicle valve interarea, pseudodeltidium, hinge spines, lateral spines (arrow), Brecksville Reservation, interval 27, CMNH 8439; 14, interior mold of pedicle valve, hinge spines, prostrate spines, Brecksville Reservation, float, CMNH 8440. Bar scale = 1 cm. 1990 Brachiopods and Bivalves in the Chagrin Shale 17 capillate; capillae increase by intercalation; fold and sulcus absent; anterior margin rounded. Pedicle valve interior with short median septum, originating at posterior margin, and extending approximately one-fourth shell length; dental plates present, diverging at wide angle; diductor muscle scars large, flabellate. Brachial valve interior with small, lobate cardinal process; muscle scars small, slightly flabellate. Remarks Prosser (1912) reported and described Chonetes minutus Prosser, 1912 from the Chagrin Shale in Ashtabula County, and Chonetes scitulus Hall, 1857 from the Chagrin Shale at Tinker’s Creek in Cuyahoga County. Retichonetes is differentiated from Chonetes by the reticulate appearance of the valves, and by the presence of an anterior median septum and two short lateral septa in the brachial valve, as opposed to three or more fine, long, diverging septa in the brachial valve of Chonetes (Muir- Wood, 1962). Retichonetes occurs throughout both of the measured sections (Figures 2 and 3). Suborder Productidina Waagen, 1883 Superfamily Productacea Gray, 1840 Family Productellidae Schuchert, 1929 Subfamily Productellinae Schuchert, 1929 Productellid gen. indet. Figures 8. 1-8.3 Description of material Several medium size pedicle valves are not assignable to a genus or to a species. They retain the convex, subtrigonal, productellid shape, but an absence of brachial valves, interareas, and cardinalia preclude more specific identification. The inflated pedicle valve appears to be rugose anteriorly, and to be ornamented with evenly spaced spine ridges. The beak extends over the hinge area, and curves inward. Width is greater than length, with the greatest width at midlength. A mold of the exterior of the pedicle valve confirms the presence of ears. Although one specimen (Figure 8.1) has a prominent sulcus, a feature common in leioproductids, the lack of a median ridge ornamented with a spine row makes assignment to this group questionable. Another specimen (Figure 8.2b) has two converging “septa” originating at the umbo and extending down the center of the valve to midlength. Specimens of this taxon appear to be present, in small numbers, throughout both Brecksville and Brandywine Falls (Figures 2 and 3). Subfamily Chonopectinae Muir- Wood and Cooper, 1960 Genus Acanthatia Muir- Wood and Cooper, 1960 Acanthatia sp. Figures 8.4-8.14 Description of material Small to medium size, semicircular outline. Width greater than length. Hinge length slightly less that greatest width, which occurs at midlength. Pedicle valve moderately convex, brachial valve concave. Pedicle valve ears small, if present, and not differentiated from visceral disc. Pedicle valve interarea longer and higher than brachial valve interarea, and characterized by pseudodeltidium (Figure 8.13). Pedicle valve ornamented by evenly-spaced, concentrically arranged prostrate spines over surface, and long balancing or attachment spines at hinge. Pedicle valve slightly rugose, especially anteriorly. Brachial valve with depressions corresponding to pedicle valve spines (Figure 8.12). Pedicle valve interior often marked by very fine punctation, representing endospines. Brachial valve interior with bilobed cardinal process, and very thin breviseptum. Alveolus not well preserved. Interior surface marked by numerous, small endospines. Muscle scars not seen. Remarks Acanthatia is characterized by its semicircular outline, lack of costation, lack of spine ridges and distinct interarea in each valve. The finely punctate interior surfaces are evidence of endospines, which represent projections of the taleolae (pseudopunctae) into the body cavity. This genus occurs throughout both sections in this study. Prosser (1912) reported Productella hirsuta Hall, 1867 from the Chagrin Shale at Tinker’s Creek, Cuyahoga County. Productella is characterized by suberect or recumbent spines which arise from scattered pustules (Muir- Wood and Cooper, 1960). The specimens in this study show no evidence of surface pustules. Further, spines are arranged in rows, and are not scattered. Order Rhynchonellida Kuhn, 1949 Superfamily Rhynchonellacea Gray, 1848 Family Trigonirhynchiidae McLaren (in Schmidt and McLaren), 1965 Genus Centrorhynchus Sartenaer, 1970 Centrorhynchus sp. Figures 9.1-9.12 Description of material Small size, biconvex, rounded outline; length approximately equal to width; greatest height at midvalve. Ornament of simple, sharply angular costae, which arise in umbonal regions, and terminate at anterior margins. Costae internally thickened in axial regions, leaving a more rounded impression on interior molds. General costal formula 3/4; 1-1/0-0; 7+/8+ (Figure 9.1). Pedicle valve umbo inflated, beak slightly incurved over interarea. Pedicle valve sinus wide, beginning about midvalve, and 18 SCHWIMMER AND FELDMANN No. 45 FIGURE 9. Centrorhynchus sp. la-e, ventral, dorsal, lateral, posterior, and anterior views, Brecksville Reservation, interval 28, CMNH 8441 ; 2a, b, interior mold of brachial valve and latex mold of same specimen, Brandywine Falls, interval 39, CMNH 8442 ; 3, posterior view, Brecksville Reservation, interval 28, CMNH 8443; 4a-d, posterior, dorsal, lateral and ventral views, Brecksville Reservation, interval 28, CMNH 8444 ; 5, posterior view, pedicle foramen, Brecksville Reservation, float, CMNH 8445; 6, crenulate anterior margin, Brecksville Reservation, interval 28 ,CMNH 8446; 7a, b, pedicle and brachial valve interiors, Brecksville Reservation, interval 5, CMNH 8447; 8, brachial valve interior, Brandywine Falls, float, CMNH 8448; 9, mold of pedicle valve, Brandywine Falls, float, CMNH 8449; 10, mold of pedicle valve, Brandywine Falls, float, CMNH 8450; 11, interior mold of brachial valve, Brandywine Falls, float, CMNH 8451 ; 12, interior molds of pedicle and brachial valves, Brandywine Falls, float, CMNH 8452. Bar scale = 1 cm. extending to anterior margin. Tongue well-defined. Anterior margin crenulate (Figure 9.6). Pedicle valve interior with short dental plates. Brachial valve interior with covered, U-shaped septalium, strongly arched anteriorly (Figure 9.2). Septum stout, extending approximately one half shell length. Remarks Angular costae, a distinctive general costal formula, a robust septalium and cover, generally rounded outline, inflated pedicle valve, and suberect pedicle valve beak are diagnostic characters of the genus Centrorhynchus. Sartenaer (1970) pointed out that this Famennian genus appears in New York and Pennsylvania. Identification of the Chagrin Shale specimens would be reinforced if specimens in better condition could be recovered and serially sectioned. Distribution is shown in Figures 2 and 3. This rynchonellid is the most numerous and persistent taxon found in the study area. Family Camarotoechiidae Schuchert, 1929 Subfamily Camarotoechiinae Schuchert, 1929 Genus Leiorhynchus Hall, 1860 “Leiorhynchus" sp. Figures 10.1-10.9 1990 Brachiopods and Bivalves in the Chagrin Shale 19 FIGURE 10. “Leiorhynchus” sp. 1, interior mold of brachial valve, Brecksville Reservation, interval 5, CMNH 8453; 2a, b, interior of pedicle valve umbo and interior mold of pedicle valve and brachial valve, Brecksville Reservation, interval 5, CMNH 8454 ; 3, interior, brachial valve umbo, Brandywine Falls, float, CMNH 8455; 4. posterior view, Brecksville Reservation, interval 5, CMNH 8456; 5, antero-lateral mold of pedicle valve, Brecksville Reservation, float, CMNH 8457; 6, mold of posterior, Brandywine Falls, float, CMNH 8458; 7, interior mold of brachial valve, septum and muscle scars, Brecksville Reservation, interval 2, CMNH 8459; 8a, b, internal mold of brachial valve and latex mold of same specimen, Brecksville Reservation, interval 5, CMNH 8460; 9, interior mold of brachial valve, Brandywine Falls, float, CMNH 8461 Bar scale = / cm. Description of material Medium size, biconvex, globular; outline circular. Pedicle valve beak strongly incurved and touching brachial valve. Pedicle foramen, if present, very small, and rounded. Pedicle sinus and brachial fold are weak and wide, originating from about midlength of the valve. The tongue is not visible. Surface ornament of a few coarse median costae which originate in umbonal regions and extend to anterior commissure. Brachial valve fold with at least five costae, pedicle valve sulcus with fewer. Costae on lateral slopes weak or absent. Pedicle valve interior with median septum extending one third shell length. Shell appears thickened in posterior region. Pedicle valve interior with short, divergent dental plates and robust teeth. Brachial valve interior with amphora-shaped septalium, with deep umbonal cavities on either side. Septum extends at least one half shell length. Remarks Prosser (1912) reported three species of Liorhynchus (= Leiorhynchus ) from the Chagrin Shale of Lake County. Although Leiorhynchus is not considered a Famennian genus (Sartenaer 1967), preservation in the Chagrin specimens is too poor and features are too indistinct for a more concrete diagnosis. No complete, articulated specimens were found, and the general costal formula could not be determined. This taxon is fairly abundant at Brecksville Reservation, occurring throughout the section (Figure 2). On the other hand, at Brandywine Falls, “ Leiorhynchus ” was recovered from only four intervals (Figure 3). It is possible that this taxon appeared more frequently, but was unrecognizable due to conditions of preservation. Order Spiriferida Waagen, 1883 Suborder Athyrididina Boucot, Johnson, and Staton, 1964 Superfamily Athyridacea M’Coy, 1844 Family Athyrididae M'Coy, 1844 Subfamily Athyridinae M’Coy, 1844 Genus Athyris M'Coy, 1844 Athyris sp. Figures 11.1-5, 1 1 .7 Description of material Small to medium size, subcircular, biconvex, inflated in umbonal areas, flattened on lateral slopes and around anterior margin. Hinge length less than greatest width, which occurs at midvalve. Pedicle beak extends over hinge, and incurves slightly; foramen rounded, moderately large. Pedicle valve weakly sulcate, from about midvalve to anterior margin; ornament of concentric growth lines, some lamellose. Pedicle interior with divergent dental plates extending approximately one third shell length; spiralium impressed upon interior mold of one specimen (Figure 1 1.7). 20 SCHWIMMER AND FELDMANN No. 45 FIGURE 11. Athyris-Composita. 1-5,7, Athyris sp. 1, pedicle valve exterior, dental plate impressions, Brecksville Reservation, interval 18, CMNH 8462 ; 2, interior mold of brachial valve, median septum and elongate muscle scar, Brandywine Falls, interval 59, CMNH 8463; 3, pedicle valve interior, dental plates, Brandywine Falls, float, CMNH 8464 ; 4, mold of posterior umbones, Brandywine Falls, float, CMNH 8465; 5, mold of brachial valve interior, median septum and elongate muscle scar, Brandywine Fads, float, CMNH 8466; 7, interior mold of spiralium, Brandywine Falls, float , CMNH 8468; 6,8-11, Composita, sp. 6, brachial valve exterior, Brecksville Reservation, interval 29, CMNH 8467; 8, pedicle valve interior, dental plates, chordate muscle scar, Brecksville Reservation, interval 28, CMNH 8469; 9a-c, dorsal, ventral, lateral interior molds. Brandywine Falls, float, CMNH 8470; 10, pedicle valve exterior, dental plates, Brecksville Reservation, interval 29, CMNH 8471; 11, interior mold of pedicle valve, chordate muscle scar, Brandywine Falls, float, CMNH 8472. Bar scale = 1 cm. Brachial valve interior with median septum extending half of shell length, bounded by thin, elongate muscle scars. Tooth sockets elongate. Remarks Athyris and Composita are members of the same subfamily, and externally, resemble one another. The former genus is more flattened laterally, and less convex overall than the latter. Differences between the two genera can be found internally, especially in the length and angle of divergence of the dental plates. Unfortunately, in the field, it is rare to be able to distinguish between the two genera, when weathered molds or fragments are the only material available. For this reason, Athyris and Composita are considered together when plotting occurrence at the two study sites (Figures 2 and 3). Both appear to be fairly abundant throughout the section at Brecksville Reservation. At Brandywine Falls, both are found in the lower half of the unit, and occur again sporadically in the upper third. Genus Composita Brown, 1849 Composita sp. Figures 11.6, 11.8-11.11 Description of material Small to medium size, ovate, biconvex, length greater than width; maximum width at midvalve. Hinge line short. Ornament of concentric growth lines, becoming somewhat lamellose anteriorly; faint radial striations originate at beak and extend to margins. Pedicle valve with poorly defined fold; anterior margin appears uniplicate. Pedicle beak extends over hinge line; foramen rounded, open. Pedicle valve interior strong teeth supported by divergent dental plates, which extend about one fourth sheil length, recurving slightly at the anterior ends. Muscle scars cordate (Figure 11.11). Brachial valve interior with median septum originating at beak and extending one half shell length, bounded by thin, elongate muscle scars (Figure 1 1 ,9a). Remarks Composita closely resembles Athyris, as pointed out in remarks about the latter taxon. 1990 Brachiopods and Bivalves in the Chagrin Shale 21 FIGURE 12. Spinospirifer and Ambocoelia. 1-4, Spinospirifer. 1, pedicle valve mold, Brandywine Falls, interval 73, CMNH 8488; 2, interior mold of pedicle valve with punctae, Brandywine Falls, interval 59, CMNH 8473 ; 3, interior mold of pedicle valve, punctae, Brecksville Reservation, interval 3, CMNH 8474; 4, pedicle valve mold, Brandywine Falls, interval 43, CMNH 8475; 5-9, Ambocoelia. 5, latex mold of pedicle valve, Brecksville Reservation, interval 19, CMNH 8476 ; 6, interior mold of brachial valve, Brandywine Falls, interval 68, CMNH 8477; 7, interior mold of pedicle valve, Brandywine Falls, interval 58, CMNH 8478; 8, interior mold of brachial valve, Brecksville Reservation, interval 19, CMNH 8479; 9, interior mold of pedicle valve, and brachial valve interior, Brecksville Reservation, interval 5, CMNH 8480. Bar scale = 1 cm. Suborder Spiriferidina Waagen, 1883 Superfamily Cyrtiacea Fredericks, (1919) 1924 Family Ambocoeliidae George, 1931 Genus Ambocoelia Hall, 1860 Ambocoelia sp. Figures 12.5-12.9 Description of material Small size, pedicle valve strongly convex, brachial valve weakly convex. Trigonal to subtrigonal outline; length greater than width; hinge line less than greatest width, which occurs at midvalve. Interareas not visible. Surface ornament of concentric growth lamellae, especially evident in the anterior part of the shell, as well as concentrically arranged spine rows. Pedicle valve with deep, narrow sulcus originating at beak and extending to rectimarginate anterior commissure. Posterior two-thirds of pedicle valve, and umbonal region, inflated; anterior third of valve, and lateral flanks, flattened. Brachial valve circular, with shallow sulcus extending from umbonal region to anterior margin; slightly inflated in posterior half of umbonal region; anterior half of valve, and flanks, flattened. Brachial valve interior with short divergent crura which are attached to socket ridges by hinge plates. Cardinal process small, triangular. Cardinal extremities rounded. Remarks A distinctive quadrate muscle pattern, described by Hall (1860), is characteristic of Ambocoelia, as well as a deeply triangular pedicle interarea. Very few specimens of this taxon were collected in the study area; unfortunately, neither of these features was exposed in the specimens collected, nor was the brachial valve muscle pattern. The taxon is common in the lower half of the section at Brecksville, but is sparse in the upper third. Distribution at Brandywine Falls is limited to the lower and upper thirds of the section (Figures 2 and 3). Superfamily Spiriferacea King, 1846 Family Mucrospiriferidae Pitrat, 1965 Genus Spinospirifer Martynova, 1961 Spinospirifer sp. Figures 12.1-12.4 Description of material Transverse, rhomboidal shells. Ornament of simple plications; shell spinose. Pedicle valve with sulcus defined by one larger plication on either side, and at least three smaller plications on the sulcus. Pedicle valve interior with short, divergent dental plates. Brachial valve, interareas, and other internal structures unknown. Remarks A few specimens of this taxon were found to be spinose upon microscopic examination. Material was scarce, and consisted mostly of fragments, making identification difficult. It is possible that this taxon may be 22 SCHWIMMER AND FELDMANN No. 45 FIGURE 13. Cyrtospirifer leboeufensis. 1, partially exfoliated pedicle valve, Brandywine Falls, float, CMNH 8481; 2, interior mold of pedicle valve, Brecksville Reservation, interval 11 , CMNH 8482; 3a,b, interior mold ofspiralium and counterpart, Brandywine Falls, interval 45, CMNH 8483; 4, posterior view, pedicle valve interarea, upper half of delthyrium covered, Brandywine Falls, float, CMNH 8484; 5, posterior view, interarea, Brecksville Reservation, interval 24, CMNH 8485; 6, pedicle valve interior, convergent dental plates, flahellate muscle scar, Brandywine Falls, float, CMNH 8486; 7, pedicle valve interior, Brandywine Falls, float, CMNH 8487. Bar scale - 1 cm. FIGURE 14. Cyrtospirifer spicatus. 1, mold of pedicle valve exterior, elongate mucrons, Brandywine Falls, float, CMNH 8489; 2, partial mold of pedicle valve interior, divergent dented plates and spatulate muscle scar, Brandywine Falls, float, CMNH 8490; 3, pedicle valve interior, dental plates diverging at 45 degree angle, Brecksville Reservation, interval 28, CMNH 8491 ; 4, interior mold of pedicle valve, divergent dental plates, spatulate muscle scar, Brandywine Falls, float, CMNH 8492; 5, interior mold of pedicle valve, spiralium impression (arrow), Brandywine Falls, float, CMNH 8493; 6, posterior view of narrow interarea, Brandywine Falls, float, CMNH 8494. Bar scale = 1 cm. 1990 Brachiopods and Bivalves in the Chagrin Shale 23 FIGURE 15. Sphenospira alta. I. interior mold of pedicle valve interaea with horizontal and vertical striations, delthyrium, stegidium, Brecksville Reservation, interval 21 , CMNH 8495; 2, brachial valve exterior, Brecksville Reservation, interval 8, CMNH 8496; 3, interior mold of pedicle valve interarea, stegidium, Brecksville Reservation, interval 8, CMNH 8497; 4, interior mold of pedicle valve interarea, stegidium. Brecksville Reservation, interval 28, CMNH 8498. Bar scale = I cm. more abundant in the study area than reported, but due to the paucity of well-preserved material from the unit, may not have been identifiable. Partial molds and fragments are particularly difficult to identify. Family Cyrtospriferidae Termier and Termier, 1949 Genus Cyrtospirifer Nalivkin (in Fredericks, 1924) Cyrtospirifer leboeufensis Greiner, 1957 Figures 13.1-13.7 Description of material Medium to large size; generally transverse; subpentagonal profile; unevenly biconvex, with pedicle valve slightly more inflated. Hinge line straight, equal to or slightly less than greatest width of shell; interareas moderately high, vertically striated; brachial interarea anacline, pedicle interarea apsacline; large, open, triangular delthyrium. Valve exteriors ornamented with low, rounded, costae which originate at beak and extend to anterior margin; lateral costae simple. Costae on fold and sulcus increase by bifurcation. External lamellae common on anterior half of shell. Pedicle valve exterior with greatest convexity in area of umbo; beak projects over hingeline; long, V-shaped sulcus is moderately deep on the posterior half of the valve, becoming shallow and even obsolete for the anterior half. Anterior margin gently uniplicate. Brachial valve exterior with a sharply defined fold originating at the beak and terminating at the anterior margin. Pedicle valve interior with pair of dental plates extending from the beak, initially diverging at a wide angle then recurving around the anterior end of the spatulate diductor muscle scar, which bears radial striations on the anterior two-thirds of its length. The adductor muscle scars are seen on a slender median groove, located between two low ridges on the diductor scar, which extends two-thirds the length of the shell. Spiral brachidia impressed upon the internal mold of the lateral mantle cavity in some specimens. The brachial valve interior not observed. Remarks Cyrtospirifer is particularly abundant and well preserved in the Chagrin Shale, and often original shell material can be recovered. The preservation of delicate dental plates suggests moderate to rapid burial, before disarticulation and breakage could occur. Greiner (1957) noted in C. leboeufensis that the pedicle valve dental plates join, forming a low ridge around the anterior portion of the diductor muscle scar. This ridge is not prominent in the Chagrin specimens. He further stated that this species is distinguished by its size, shape, and distinct, elongated muscle scar, which is partly enclosed by the dental plates (Figure 13.7). Prosser (1912) reported Cyrtospirifer from Brecksville Reservation and from Brandywine Falls. 24 SCHWIMMER AND FELDMANN No. 45 FIGURE 16. Toryniferella sp. la,b, interior mold of pedicle valve, median septum and dental plates, with enlargement of double-barreled spines, Brandywine Falls, interval 69, CMNH 8499; 2, posterior view of pedicle valve beak and interarea, Brandywine Falls, interval 57, CMNH 8500; 3, interior mold of pedicle valve. Brandywine Falls, float, CMNH 8501; 4, interior mold of pedicle valve, Brecksville Reservation, interval 19, CMNH 8502; 5, interior mold of pedicle valve. Brandywine Falls, interval 69, CMNH 8503; 6, interior mold of brachial valve, Brandywine Falls, float, CMNH 8504 ; 7, interior mold of brachial valve, weak fold. Brandywine Falls, interval 43, CMNH 8405. Bar scale = 1 cm. Cyrtospirifer spicatus Greiner, 1957 Figures 14.1-14.6 Description of material Medium to large shell, extremely transverse; biconvex; hinge length increased by long, attenuated mucrons. Fold and sulcus fairly well-defined from beak to anterior margin. Interareas, costae, and other exterior features similar to C. leboeufensis. Pedicle valve interior distinguished by short dental plates, which extend one-third the length of the shell and diverge at a 45 degree angle. Spatulate muscle scar not enclosed by dental plates. Remarks Distinguishing characters for this species include faint growth lines as micro-ornament, the extremely transverse profile with exaggerated mucrons, the short, diverging dental plates, and the shape of the diductor muscle scar (Greiner, 1957). Cyrtospirifer leboeufensis and C. spicatus are for the most part indistinguishable in the field at either Brecksville or Brandywine Falls. Often, only fragments or partial specimens are recovered, and identifications to the species level are not possible. Therefore, in the brachiopod distribution data (Figures 2 and 3), the species are combined and charted together, at the generic level. Cyrtospirifer exhibits almost continuous occurrence in both sections studied. A few intervals in the upper third of the Brandywine Falls section lack species of this genus. Genus Sphenospira Cooper, 1954 Sphenospira alta (Hall, 1867) Figures 15.1-15.4 Description of material Large size, triangular outline; valves unevenly biconvex; pedicle valve pyramidal, brachial valve shallow; greatest width at hinge; pedicle valve interarea finely striated with vertical and faint horizontal lines; posterior two-thirds of delthyrium covered by delthyrial plate, while stegidium covers anterior one-third, growing in increments as pedicle atrophies (Cooper, 1954). Lateral slopes, fold and sulcus are costate; costae originate at beak and terminate at anterior margin, increasing by bifurcation on lateral slopes. Pedicle and brachial valve interiors not observed. Remarks Prosser (1912) reported this species from Chippewa Creek at Brecksville Reservation. This species was also collected and described from Chippewa Creek by Cooper in 1954. Specimens in the National Museum of Natural History, Washington exhibit divergent dental plates in the interior of the pedicle valve, as well as radiating muscle scars between the plates. Museum specimens also show the cardinal process, teeth, and sockets. Sphenospira alta occurs, in low abundance, in the lower half of the section at Brecksville. Only two occurrences in the upper half were 1990 Brachiopods and Bivalves in hie Chagrin Shale 25 recorded. Two specimens were found in situ in the lower half of the Brandywine Falls area, although several very large specimens were recovered from float in the creek. Occurrences are charted in Figures 2 and 3. Superfamily Reticulariacea Waagen, 1883 Family Elythidae Fredericks, 1924 Genus Toryniferella Weyer, 1967 Toryniferella sp. Figures 16.1-16.7 Description of material Medium to large size (width greater than 25 mm in adults); transversely elliptical outline; biconvex; rounded margins; width greater than length; hinge length less than greatest width, which occurs at mid-valve; low interareas; open delthyrium. Valve exteriors ornamented by prominent concentric growth lines, each bearing numerous, regularly spaced, fine, double-barreled spines (Figure 16.1). Pedicle valve exterior with high umbonal region, and small, pointed, well-defined beak, which incurves over interarea; sulcus shallow or obsolete. Brachial valve exterior with low fold, set apart by a shallow groove on either side; umbo elevated, broad. Pedicle valve interior displays long, median septum, originating at umbo, extending approximately two-thirds of valve length. Median septum flanked by initially diverging dental plates, which extend one-half valve length, and then recurve slightly toward septum. Brachial valve interior with a thin, median ridge which begins approximately one-fourth the shell length from the posterior margin, and extends one- half of valve length. Two thin, short, parallel crura extend from hinge, tenninating before median ridge begins. Crural plates form single plate across inner surface of crura, but do not attach to the valve floor. Remarks Toryniferella is easily distinguished from Reticularia praematura (Hall, 1866), which was reported in the Chagrin by Prosser (1912), by the presence of biramous, or double- barreled spines in the former genus, and uniramous spines in the latter genus. Toryniferella can be differentiated from Torynifer based on brachial valve morphology. Torynifer possesses a median septum in the brachial valve, which supports the cardinalia (Carter, 1988). The Chagrin specimens have no brachial median septum. Toryniferella is distinguished from Kitakamithyris, another elythid, by the presence in the former of crural plates which are not separated or divided, and which do not connect to the floor of the shell (Weyer, 1967). Specimens of Toryniferella in the Chagrin Shale are generally incomplete. Original shell material is almost absent, and dorsal (brachial) beaks are often broken off. At this time, identification below the genus level is not possible. Toryniferella is a relatively recently named genus (Weyer, 1967), and has not been reported widely in the literature. It is possible that reclassification of some specimens of Torynifer spp. and Reticularia praematura (Hall, 1866), the latter species having been reported from the Chagrin by Prosser (1912), may result in further additions of members to the genus. Toryniferella occurs in abundance in the lower half of the section at Brandywine Creek, and again in the upper third of the same section. In contrast, at Brecksville, this genus is sparse, occurring in four intervals in the lower half of the section, and in five intervals in the upper third (Figures 2 and 3). Phylum Mollusca Cuvier, 1797 Remarks The Chagrin megafauna in the Cuyahoga River Valley is almost exclusively restricted to brachiopods. The only other faunal elements, with the exception of two unidentifable, pyritized, epibionts, are four taxa of bivalves. Their distribution is plotted in Figures 2 and 3. Class Bivalvia Linne, 1758 (Buonanni, 1681) Subclass Pteriomorphia Beurlen, 1944 Order Pterioida Newell, 1965 Suborder Pteriina Newell, 1965 Superfamily Pteriacea Gray, 1847 (1820) Family Pterineidae Miller, 1877 Genus Leptodesma Hall, 1883 Leptodesma sp. Figures 17.6, 17.8-17.10 Description of material Small to medium size, biconvex. Hinge line straight from anterior extremity to posterior wing. Posterior wing terminates as a short spine. Surface ornament of fine to coarse, irregularly spaced concentric lines. Dentition, hinge, ligament and internal structures unknown. No shell material is preserved. Remarks Very small specimens of Leptodesma are often found concentrated on bedding planes in the Chagrin Shale. These could represent spatfalls, or juvenile specimens, which were buried before growth could proceed. This taxon is found at both locations, mainly in the lower halves of each section (Figures 2 and 3). Superfamily Pectinacea Rafinesque, 1815 Family Pterinopectinidae Newell, 1938 Genus Pterinopecten Hall, 1883 Pterinopecten? sp. Figures 17.1-17.4 26 SCHWIMMER AND FELDMANN No. 45 FIGURE 17. Bivalvia. 1-4, Pterinopecten? sp. 1, interior mold, Brandywine Falls, float, CMNH 8506; 2, interior mold, Brandywine Falls, interval 58, CMNH 8507; 3, mold of exterior, with reticulate pattern formed by costa and growth lines, Brandywine Falls, interval 58, CMNH 8508; 4, exterior mold, subequal auricles, Brandywine Falls, interval 74, CMNH 8509; 5, Sanguinolites sp., interior mold of right valve, growth lines increase by intercalation and bifurcation (arrow), Brandywine Falls, float, CMNH 8510; 7, Spathella? sp., interior mold, right valve, Brandywine Falls, float, CMNH 8512; 6, 8-10, Leptodesma sp. 6, interior mold of left valve, with posterior spine, Brecksville Reservation, interval 25, CMNH 8511 ; 8, partial interior mold of left valve, Brandywine Falls, interval 62, CMNH 8513; 9, interior mold of disarticulated specimen, Brandywine Falls, interval 58, CMNH 8514; 10, molds of juvenile specimens, Brandywine Falls, float. CMNH 8515. Bar scale = 1 cm. Description of material Small to medium size, generally flattened. Pectinid shell shape with subequal anterior and posterior ears. Hinge line straight, extending length of ears. Surface ornament of prominent radial ribs, crossed by concentric growth lines, resulting in a reticulated appearance. Shell material, dentition, hinge features, ligament and musculature unknown. Remarks Specimens were all collected from the Brandywine Falls location, either as float or in situ (Figure 3). No specimen is complete; information was obtained from partial molds. Order Modiomorphoida Newell, 1969 Superfamily Modiomorphacea S.A. Miller, 1877 Family Modiomorphidae S.A. Miller, 1877 Genus Spathella Hall, 1885 Spathella? sp. Figure 17.7 Description of material One medium size specimen was collected as float from a slab at Brandywine Falls. Right valve ovate, ornamented by evenly spaced, concentric growth lines. Beak prosogyrous. One shard of shell material remains on the specimen. Lunule and escutcheon, ligament, hinge, and musculature unknown. Subclass Anomalodesmata Dali, 1889 Order Pholadomyoida Newell, 1965 Superfamily Pholadomyacea Gray, 1847 Family Grammysiidae S.A. Miller, 1877 Genus Sanguinolites M'Coy, 1844 Sanguinolites sp. Figure 17.5 1990 Brachiopods and Bivalves in the Chagrin Shale 27 Description of material One specimen was recovered from float at Brandywine Falls. Specimen large, elongate, with small umbones located in anterior third of shell. Beak prosogyrous. Ornament of very coarse comarginal rugae, which increase by bifurcation and intercalation (Figure 17.5). Brachiopod distribution The total number of genera and their distribution throughout both stratigraphic sections shows that brachiopods are fairly abundant in the lower one half to two thirds of each section (Figure 18). This is followed by a sharp drop in the number of genera to two at Brecksville and to zero at Brandywine. An especially silty interval is present at this position at Brandywine, and could account for the decrease in taxa, and for their inability to recolonize during the time of silt deposition. Silt was transported into the area by periodic storms (Weidner and Feldmann, 1983), and tended to obliterate brachiopod assemblages with rapid burial. There is also a possibility that turbidites were the dominant mechanism of rapid sedimentation. Lewis (1988, p. 24) evaluated different studies, and favors a storm mechanism and deposition of silts in the form of tempestites. After intervals of low abundance, abundance increases in the upper third of each section. Diversity is never as great as in the lower parts of both sections (Figure 18). Diversity declines near the contact with the Cleveland Shale, concommittant with a change to anoxic conditions. Brachiopods such as Cyrtospirifer and Centrorhynchus are well distributed throughout both sections (Figures 2 and 3). This supports a conclusion that these two taxa were eurytopic, and represented pioneer species. Cyrtospirifer possessed a complex, spirolophous lophophore, which enhanced food and oxygen gathering capabilities. Centrorhynchus also had a spirolophous lophophore, but its resurgence after periodic disasters might possibly be due to an initial, epiplanktonic mode of existence, analogous to that suggested by Ager (1962; 1965). Some genera, such as Sphenospira and Aulacella, are abundant at one site and sparse at the other, while the faunal distributions of other genera, such as Retichonetes , Acanthatia, and Schellwienella, appear similar in both sections. The composition and vertical distribution of the Chagrin fauna reflects colonization by opportunistic, pioneering species. Levinton (1970) distinguished between equilibrium species populations and opportunistic species populations. The former are resource-limited, or near the carrying capacity of the environment, which is stable and less likely to fluctuate over time. As a result, population levels tend to remain constant. Opportunistic species, on the other hand, are not resource limited, and increase rapidly in numbers. Under unstable environmental conditions, these species are considered physiological generalists; they tend to increase in numbers when space, temperature, salinity and other environmental factors become favorable. Levinton (1970) offered several criteria for identifying opportunistic species. First, there is random orientation of specimens and lack of size sorting in individual beds, although dominant taxa tend to be grouped by size. In the Chagrin, this is especially true of the cyrtospiriferids and the rhynchonellids, which are numerically dominant and ubiquitous at the two sections. Second, distribution over a limited area is characteristic, with adjacent, nonfossiliferous horizons. Within the study area, some layers are fossiliferous, while adjacent layers and areas may be fossil poor. The pattern of NUMBER OF GENERA 0 6 12 0 6 12 BRANDYWINE Figure 18. Vertical distribution of taxa in the Cuyahoga Valley. The number of taxa within each 20 cm interval is plotted on a scale of zero to twelve . Both sections exhibit abundant taxa in the lower halves, a sharp drop in abundance in the upper one half to one third, and an increase in abundance in the upper portion. Taxa decrease again near the Chagrin-Cleveland boundary. 28 SCHWIMMER AND FELDMANN No. 45 distribution appears random. On a regional scale, very few Chagrin Shale outcrops in northeast Ohio contain the abundant brachiopod fauna found in the Cuyahoga River Valley. Adverse environmental conditions or lack of colonizing stock may account for this condition, although differential preservation must also be considered. Third, stable faunal assemblages may be invaded by one dominant opportunistic species, leading to overwhelming domination of the assemblages (85-100%). This does not appear to be the case with the Chagrin fauna. Stable faunal assemblages never seemed to develop. After periods of rapid silt deposition, many of the species which were able to recolonize appear to be opportunists, with special morphologies for coping with stressed environments. New taxa did not appear in the section over time, possibly because the initial colonists had not sufficiently stabilized the substrate, and because the physical environment remained changeable. Most Chagrin brachiopods represent pioneer species, never rising much past an entry level for community succession. The same faunal elements continue to reestablish themselves after storm events. Very few taxa show distribution patterns which are anomalous in that they do not reflect pioneering characteristics throughout both sections (Figures 2 and 3). Aulacella is reported only once from the upper two thirds of the Brandywine Falls section, although it appears throughout the Brecksville section. Sphenospira is present in the lower half of Brecksville, and in two intervals in the upper half of that section, and in two intervals in the lower half of Brandywine Falls section. Schellwienella occurs in the lower halves of both sections, and once in the upper parts of both sections. The lingulids show the same general distribution pattern as Schellwienella. Orhiculoidea is found only at Brecksville and is not considered in this analysis. The other Chagrin brachiopods remain fairly constant in their distributions throughout the sections. Pioneer assemblages do not appear to have been replaced by more stable aggregations. Storm events were too frequent to allow for an orderly progression of faunal elements and community succession over time. Alexander (1977) studied opportunistic brachiopods in Idaho and Utah. Many developed compressed body plans for suspension on a fluid substrate, which also served to increase oxygen diffusing mantle surfaces under oxygen deficient conditions. According to Alexander (1977), stressed communities are characterized by one dominant type of opportunist which can adapt to adverse conditions; by an absence of rugose and colonial corals, sponges, and bryozoans, which thrive in clear, oxygenated, less turbid waters; and by a variety of vacant niches, as evidenced by a lack of mobile faunal elements. The Chagrin environment was probably stressed, and faunal diversity is low, consisting of sessile organisms which can float in, or on, an unstable, muddy substrate, and which can feed and breathe under low oxygen, turbid conditions (Schwimmer, 1988). In this respect, the Chagrin brachiopods are opportunists, which exploited stressed habitats, and managed to establish and maintain pioneer communities. The brachiopods collected in the Cuyahoga Valley Chagrin sequence provide the best evidence for a paleoenvironmental interpretation of the Chagrin Shale, due to their almost exclusive occurrence in the sediments at both locales. Their morphological adaptations and inferred life styles enhance our understanding of conditions at the time of deposition. Further, their distribution patterns at the two sites may aid in correlation of the two sections. It is anticipated that detailed collection at sites in the central and western portions of the Chagrin outcrop belt would yield similar brachiopod faunas and distributions. Age of the Chagrin Shale On the evidence of its conodont and brachiopod faunas, the Chagrin Shale is Late Devonian (Famennian) in age. Schopf and Schwieterling (1970) and Murphy (1973) used the presence of the alga Foerstia, obtained from locales in Ashtabula County, to correlate the Chagrin with other Upper Devonian units in northwestern Pennsylvania and southwestern New York. Murphy (1973) reported Foerstia near the base of the Ellicott Shale Member of the Chadakoin Formation in Pennsylvania and New York. This formation is included in the Conneaut Group of the upper Cassadagan Stage. Schopf (in Feldmann et al., 1978) commented that the Foerstia zone probably was within the Upper Cassadagan Stage. Feldmann et al. (1978) speculated that the Chagrin could extend into the Bradfordian Stage (Fammenian). Brachiopods found in the Chagrin Shale support a Famennian age. Dutro (1981) reported that Cyrtospirifer spicatus Greiner and Sphenospira alta (Hall) occur in the Venango and Cattaraugus formations, assigned to the Famennian Conewangoan Stage of Cooper et al. (1942), equivalent to the Bradfordian Stage of Rickard (1975). The most definitive age data for the Chagrin Formation come from a recent conodont analysis, reported by Zagger and Banks (1988), from the Skinner's Run pyrite bed, mentioned earlier. Anita Harris (U.S. Geological Survey, 1988, pers. comm.) identified specimens of Polygnathus exerplexus Sandberg and Zeigler, 1979 and Bispathodus aculeatus (Branson and Mehl, 1934) from the Zagger and Banks (1988) collection. These indicate correlation with the Middle expansa subzone (Sandberg and Zeigler, 1984) of late Famennian age. Polygnathus exerplexus is restricted to the Lower to Middle expansa Zone, and Bispathodus aculeatus, although with a greater stratigraphic range, first occurs in the Middle expansa subzone (Harris, 1988, pers. comm.). 1990 Brachiopods and Bivalves in the Chagrin Shale 29 Conclusions Nineteen taxa of brachiopods and four taxa of bivalves are described here from the Upper Devonian Chagrin Shale of the Cuyahoga River Valley in northeast Ohio. The Chagrin sediments were deposited on a relatively shallow marine shelf, below normal wave base. Episodic storms are thought to have carried sediments westward from the prograding Catskill Delta complex, leading to the periodic burial of brachiopod assemblages. Brachiopods are generally preserved as molds, with occasional shell material. Some molds are coated with pyrite. Pyrite also occurs in sediments as nodules or burrow infillings. Lingulid brachiopods are preserved in phosphatic concretions. The infrequent fragmentation and abrasion of specimens suggests rapid burial rates. The composition and distribution patterns of brachiopod assemblages implies population by pioneering species, which were able to survive under adverse conditions. Stressed conditions precluded colonization by stenotopic faunal elements. Brachiopod distribution patterns at the two sites suggest a rough correspondence in assemblage composition. Conodont evidence provides an Upper Famennian age for the Chagrin, as does the occurrence of the brachiopod Sphenospira alta. Acknowledgements We wish to acknowledge the Cleveland Metroparks and The National Park Service for granting collecting permits. Joseph Hannibal, The Cleveland Museum of Natural History, made the Museum collections available to us during this project, and made the EDX analysis by Scott Pluim, Tulsa, available. Lisa Stillings, formerly of Kent State University, performed phosphorus lab tests. We are appreciative of the counsel provided by Dr. J. Thomas Dutro, Jr. of the United States Geological Survey, and Dr. John L. Carter of the Carnegie Museum. Dr. Jed Day, Illinois State University, and an anonymous reviewer reviewed this manuscript and suggested improvements. Contribution 398, Department of Geology, Kent State University, Kent, Ohio 44242. References Ager, D. V. 1962. The Occurrence of Pedunculate Brachiopods in Soft Sediments. Geological Magazine 99:184-186. Ager, D. V. 1965. The Adaptation of Mesozoic Brachiopods to Different Environments. Palaeogeography, Palaeoclimatology, Palaeoecology 1:143-172. Alexander, R. R. 1977. Growth, Morphology and Ecology of Paleozoic and Mesozoic Opportunistic Species of Brachiopods from Idaho-Utah. Journal of Paleontology 51:11 33-1 149. Alichova, T. N. 1960. Otryad Orthida: Mshanki, Brakhiopody [T. G. Sarycheva, ed.], p. 183-197. In Osnovy Paleontologii , edited by Y. Orlov. Baird, G. C., and C. E. Brett. 1986. Erosion on an Anaerobic Seafloor: Significance of Reworked Pyrite Deposits from the Devonian of New York State. Palaeogeography, Palaeoclimatology, Palaeoecology 57:157-193. Berner, R. A. 1984. Sedimentary Pyrite Formation: An Update. Geochimica et Cosmochimica Acta 48:605-615. Beurlen, K. 1944. Beitrage zur Stammesgeschichte der Muscheln. Bayerische Akademie der Wissenschaften , Sitzungsberichte, No. 1-2:133-145. Branson, E. B., and M. G. Mehl, 1934. Conodonts from the Grassy Creek Shale of Missouri. In Conodont Studies No. 3: Missouri University Studies 8(3): 1 7 1 -259. Brett, C. E., and G. C. Baird. 1986. Comparative Taphonomy: A Key to Paleoenvironmental Interpretation Based on Fossil Preservation. Palaios 1:207-227. Broadhead, R.F., and PE. Potter. 1980. Petrology of the Devonian Gas-Bearing Shale Along Lake Erie Helps Explain Gas Shows. United States Department of Energy, Morgantown Energy Technological Center, Morgantown, West Virginia. 49 p. Bronn, H.G. 1862. Die Klassen und Ordnungen der Wiechtiere (Malacozoa) 3(1): 1-518. Brown, T. 1849. Illustrations of the Fossil Conchology of Great Britain and Ireland with Descriptions and Localities of all Species. London, 273 p. Bruiguiere, J.G. 1797. Encyclopedic Methodique . Tableau Ency dope clique et Methodique des Trois Regnes de la Nature. Vers, Coquilles, Mollusques et Polypiers, Vol. 2. pis. 96-134. Carter, J.L. 1988. Early Mississippian Brachiopods from the Glen Park Formation of Illinois and Missouri. Bulletin of the Carnegie Museum of Natural History No. 27, 82 p. Cooper, G.A. 1954. Unusual Devonian Brachiopods. Journal of Paleontology 28:325-332. Cooper, G.A., C. Butts, K.E. Caster, G.H. Chadwick, W. Goldring, E.M. Kindle, E. Kirk, C.W. Merriam, F.M. Swartz, PS. Warren, A.S. Warthin, and B. Willard. 1942. Correlation of the Devonian Sedimentary Formation(s) of North America. Geological Society of America Bulletin 53:1729-1794. Cooper, G.A., and J.T. Dutro, Jr. 1982. Devonian Brachiopods of New Mexico. Bulletins of American Paleontology Vol. 82 and 83, 215 p. Cushing, H.P., F. Leverett, and F.R. Van Horn. 1931. Geology and Mineral Resources of the Cleveland District, Ohio. U. S. Geological Survey Bulletin 818, 138 p. Cuvier, G. 1797. Tableau Elementaire de T Histoire Naturelle des Animaux. Paris, 1-710. Dali, W. H. 1889. Scientific Results of Explorations U.S. Fish Comm. Albatross. No. VII. Preliminary report on the collection of Mollusca and Brachiopoda Obtained in 1887- 1888: U.S. National Museum Proceedings 12:217-362. Dumeril, A.M.C. 1806. Zoologie Analytique on Metlwde Naturelle de Classification des Animaux. Allais (Paris), 344 p. Dunbar, C.O., and G.E. Condra. 1932. Brachiopoda of the Pennsylvanian System of Nebraska. Nebraska Geological Survey Bulletin, 2nd Series, 51:1 -377. Dutro, J. T., Jr. 1981. Devonian Brachiopod Biostratigraphy of New York State, p. 67-82. In Devonian Biostratigraphy of New York State, edited by W. A. Oliver and G. Klapper. 30 SCHWIMMER AND FELDMANN No. 45 Feldmann, R. M., R. G. Osgood, Jr., E.J. Szmuc, and D.W. Meinke. 1978. Chagrinichnites brooksi, a New Trace Fossil of Arthropod Origin. Journal of Paleontology 52:287-294. Feldmann, R.M., J.T. Hannibal, and L.E. Babcock. 1986. Fossil Worms from the Devonian of North America (Sphenothallus) and Burma (“Vermes”) Previously Identified as Phyllocarid Crustaceans. Journal of Paleontology 60:34 1 -346. Fredericks, G.N. 1924. O Verkhe-Kammennougol’nykh Spiriferidakh Urala. Geologicheskiy Komitet Izvestia 38:295-324. George, T.N. 1931. Ambocoelia Hall and Certain Other British Spiriferidae. Geological Society of London, Quarterly Journal 87:30-61 . Girty, G. H. 1939. Some Linguloid Shells from the Late Devonian and Early Carboniferous Rocks of Pennsylvania and Ohio. U.S. Geological Survey Professional Paper 193-C, p. 47-67. Gray, J.E. 1840. Synopsis of the Contents of the British Museum. 42nd ed., London, 370 p. Gray, J.E. 1847. A List of the Genera of Recent Mollusca, Their Synonyms and Types. Zoological Society of London, Proceedings 15:129-219. Gray, J.E. 1848. On the Arrangement of the Brachiopoda. Annals and Magazine of Natural History, Series 2, 2:435-440. Greiner, H. 1957. Spirifer disjunctus : Its Evolution and Paleoecology in the Catskill Delta. Yale University, Peabody Museum of Natural Histoiy Bulletin 11:1-75. Hall, J. 1860. Observations of Brachiopoda. New York State Cabinet Natural History, Annual Report 13:65-75. Hall, J. 1867. Descriptions and Figures of the Fossil Brachiopoda of the Upper Helderberg, Hamilton, Portage, and Chemung Groups. New York Geological Survey, Palaeontology 4( 1 ): 1 -428 p. Hall, J. 1883. Lamellibranchiata, Plates and Explanations. Natural History New York, Paleontology 5(1 ): 1- 20, 79 pi. Hall, J. 1885. Lamellibranchiata II, Descriptions and Figures of the Dimyaria of the upper Helderberg, Hamilton, Portage and Chemung Groups. New York Geological Survey, Paleontology 5( 1 ):269-562. Hannibal, J.T., and R.M. Feldmann. 1983. Arthropod Trace Fossils, Interpreted as Echinocarid Escape Burrows from the Chagrin Shale (Late Devonian) of Ohio. Journal of Paleontology 57:705-716. Huxley, T.H. 1869. An Introduction to the Classification of Animals. London, 147 p. King, W. 1846. Remarks on Certain Genera Belonging to the Class Paliobranchiata. Annals and Magazine of Natural Histoiy 18:26-42, 83-94. King, W. 1850. A Monograph of the Permian Fossils of England. London Palaeontographical Society, 258 p. Kuhn. O. 1949. Lehrbuch der Paldozoologie. Stuttgart, 326 p. Levinton, J.S. 1970. The Paleoecological Significance of Opportunistic Species. Lethaia 3:69-78. Lewis, T.L. 1988. Late Devonian and Early Mississippian Distal Basin-Marginal Sedimentation of Northern Ohio. The Ohio Journal of Science 88(l):23-39. Linne, C. 1758. Systema Naturae, 10th Edition, Tomus 1, Holmiae, Laurentii Salvii: 1-824. Manheim, F., G.T. Rowe, and D. Jipa. 1975. Marine Phosphorite Formation off Peru. Journal of Sedimentary Petrography 45:243-251. Martynova, M.V. 1961. Stratigrafiia i Brakhiopody Famenskogo Iarusa Zapadnoi Chasti Tsentral’nogo Kazakhstana. Materialy po Geologii Tsentral' nogo Kazakhstana 2:1-210. M’Coy, F. 1844. A Synopsis of the Characters of the Carboniferous Limestone Fossils of Ireland. Dublin, 207 p. Menke, C.T. 1828. Synopsis Methodica Molluscorum Generum Omnium et Specierum Earum Quae in Museo Menkeano Adservantur. Pyrmonti, 91 p. Miller, S.A. 1877. The American Paleozoic Fossils. Published by the Author, Cincinnati, 245 p. Muir- Wood, H.M. 1955. A History of the Classification of the Phylum Brachiopoda. British Museum of Natural History, London, 24 p. Muir- Wood, H.M. 1962. On the Morphology and Classification of the Brachiopod Suborder Chonetoidea. British Museum of Natural History, London, 132 p. Muir-Wood, H.M., and G.A. Cooper. 1960. Morphology, Classification and Life Habits of the Productoidea (Brachiopods). Geological Society of America Memoir 81, 447 p. Muir-Wood, H.M., and A. Williams. 1965. Strophomenida, p. H36 1-370. In Treatise on Invertebrate Paleontology, Part H, Brachiopoda. edited by R. C. Moore. The Geological Society of America and University of Kansas Press. Murphy, J.L. 1973. Protosalvinia (Foerstia) Zone in the Upper Devonian Sequence of Northeastern Ohio, Northwestern Pennsylvania, and Western New York. Geological Society of America Bulletin 84:3405-3410. Newell, N.D. 1938. Late Paleozoic Pelecypods: Pectinacea. Kansas Geological Survey 10:1-123. Newell, N.D. 1965. Classification of Bivalvia. American Museum Novitates, No. 2206, 25 p. Newell, N.D. 1969. Order Modiomorphoida, p. N393. In Treatise on Invertebrate Paleontology, Part N, Mollusca 6, Vol. 1 , Bivalvia, edited by R. C. Moore. Geological Society of America and University of Kansas Press. Opik, A. 1933. Uber einige Dalmanellacea aus Estland. University of Tartu (Dorpat). Acta et Commentationes, Series A, 25( 1 ): 1-25. Opik, A. 1934. Ober Klitamboniten. University of Tartu (Dorpat), Acta et Commentationes, Series A, 26(3): 1 - 239. Orbigny, A. d’. 1847. Considerations Zoologique et Geologiques sur les Brachiopodes ou Palliobranches. Academie des Sciences Paris, Comptes Rendus 25:193-195, 266-269. Pepper, J.F., W. De Witt, Jr., and D.F. Demarest. 1954. Geology of the Bedford Shale and the Berea Sandstone in the Appalachian Basin. United States Geological Survey Professional Paper 259, 109 p. Pitrat, C.W. 1965. Family Mucrospiriferidae, p. H686. in Treatise on Invertebrate Paleontology, Part H, Brachiopoda, edited by R. C. Moore. Geological Society of America and University of Kansas Press. Potter, P.E., J.B. Maynard, and W.A. Pryor. 1980. Final Report of Special Geological, Geochemical and Petrological Studies of 1990 Brachiopods and Bivalves in the Chagrin Shale 31 the Devonian Shales in the Appalachian Basin. Prepared for United States Department of Energy Eastern Gas Shales Project. University of Cincinnati, 86 p. Prosser, C.S. 1912. The Devonian and Mississippian Formations of Northeastern Ohio. Geological Survey of Ohio, Bulletin 15, 574 p. Rafinesque-Schmaltz, C.S. 1815. Analyse de la Nature on Tableau de I'Univers et des Corps Organises, etc. Palermo, 224 p. Rickard, L.V. 1975. Correlation of the Devonian Rocks in New York. New York State Museum and Science Service Map and Chart Series 24, pi. 3. Sandberg, C.A., and W. Zeigler. 1984. Palmatolepsis- Based Revision of Upper Part of Standard Late Devonian Conodont Zonation, p. 179-189. In D.L. Clark ( ed . ) , Conodont Biofacies and Provincialism. Geological Society of America Special Paper 196. Sartenaer, P. 1961. Late Upper Devonian (Famennian) Rhynchonellid Brachiopods. Institut Royal des Sciences Nature lies de Belgique Bulletin 37:1-10. Sartenaer, P. 1967. Correlation, p. 1043-1060. In International Symposium on the Devonian System, edited by D. H. Oswald. Alberta Society of Petroleum Geologists, Calgary. Sartenaer, P. 1969. Late Upper Devonian Rhynchonellid Brachiopods from Western Canada. Geological Survey of Canada, Bulletin 169, 269 p. Sartenaer, P. 1970. Nouveaux Genres Rhynchonellides (Brachiopodes) du Paleozoique. Institut Royal des Sciences Naturelles de Belgique Bulletin 46: 1 -32. Schmidt, H., and D.J. McLaren. 1965. Paleozoic Rhynchonellacea, p. H552-H597. In Treatise on Invertebrate Paleontology, Part H, Brachiopoda, edited by R. C. Moore. Geological Society of America and University of Kansas Press, Lawrence. Schopf, J.M., and J.F. Swieterling. 1970. The Foerstia Zone of the Ohio and Chattanooga Shales. U.S. Geological Survey Bulletin 1294-H:1-15. Schuchert, C. 1929. Brachiopoda, p. 1-140. In Schuchert and C.M. Levene. Brachiopoda. Fossilium Catalogus I: Animalia, Pars 42, Junk, ( Berlin). Schuchert, C., and G.A. Cooper. 1931. Synopsis of the Brachiopod Genera of the Suborders Orthoidea and Pentameroidea, with Notes on the Telotremata. American Journal of Science, Series 5, 22:241-251. Schuchert, C., and G.A. Cooper. 1932. Brachiopod Genera of the Suborders Orthoidea and Pentameroidea. Yale University Peabody Museum of Natural History, Memoir 4, No. 1 , 270 p. Schwimmer, B.A. 1988. Brachiopoda from the Chagrin Shale (Late Devonian) of Ohio. Geological Society of America, Program with Abstracts 20(5):387. Schwimmer, B.A., J.T. Hannibal, D.J. Stukel II, and R.M. Feldmann. 1987. The Paleontology and Depositional Environment of the Chagrin Shale (Famennian) in Northeastern Ohio. Second International Symposium on the Devonian System, Program and Abstracts, Calgary, Alberta, p. 204. Shumway, S.E. 1982. Oxygen Consumption in Brachiopods, and the Possible Role of Punctae. Journal of Experimental Marine Biology and Ecology 58:207-220. Slansky, M. 1986. Geology of Sedimentary Phosphates. Elsevier, Amsterdam, 210 p. Stehli, F.G. 1954. Lower Leonardian Brachiopoda of the Sierra Diablo. American Museum of Natural History Bulletin 105:257-358. Stukel, D.J. II. 1987. Ichnology and Paleoenvironmental Analysis of the Late Devonian (Fammenian) Chagrin Shale of Northeast Ohio. Kent State University, Masters Thesis (unpubl.), 95 p. Szmuc, E.J., R.G. Osgood, Jr. and D.W. Meinke. 1976. Lingulichnites , a New Trace Fossil Genus for Lingulid Brachiopod Burrows. Lethaia 9: 163-167. Termier, H., and G. Termier. 1949. Essai sur I'Evolution des Spiriferides. Maroc Service Geologique du Division des Mines et de la Geologie, Notes et Memoires 74:85- 1 1 2. Thayer, C.W. 1986. Respiration and the Function of Brachiopod Punctae. Lethaia 19:23-31. Thayer, C.W., and H.M. Steele-Petrovic. 1975. Burrowing of the Lingulid Brachiopod Glottidia pyramidata: Its Ecologic and Paleoecologic Significance. Lethaia 8:209-221 . Thomas, I. 1910. British Carboniferous Orthotetinae. Great Britain Geological Survey Memoir 1 ( 2 ):83- 1 34. Waagen, W.H. 1882-1885. Salt Range Fossils, Part 4, (2) Brachiopoda. Palaeontologia lndica, (Memoirs of the Geological Survey of India), Series 13, 1:329-770 (1882), 2:391-546 (1883), 5:729-770 (1885). Weidner, W.E. 1983. Paleoecological Interpretation of Echinocarid Arthropod Assemblages in the Late Devonian (Famennian) Chagrin Shale, Northeast Ohio. Kent State University, Masters Thesis (unpubl.), 66 p. Weidner, W.E., and R.M. Feldmann. 1983. Paleoecological Interpretation of Echinocarid Arthropod Assemblages in the Late Devonian (Fammenian) Chagrin Shale, Northeastern Ohio. Journal of Paleontology 59:986-1004. Wetzel, R.G. 1983. Limnology. Saunders College Publishing, Philadelphia, 753 p. Weyer, D. 1967. Kitakamithyris Minato, 1951 (Brachiopoda, Spiriferida) aus dem Etroeungt (Oberdevon) und Tornai (Unterkarbon) des Rheinischen Schiefergebirges. Geologie [Berlin] 16:433-451. Zagger, G., and P. Banks. 1989. Age and Origin of the Skinner’s Run Pyrite Bed, Cuyahoga County, Ohio. The Ohio Journal of Science 89(2):9. KIRTLANDIA The Cleveland Museum of Natural History June 1990 Number 45 :33-38 Estimation of Numbers for a Riverine Necturus Population Before and After TFM Lampricide Exposure TIMOTHY O. MATSON The Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Abstract The purpose of this research was to assess the impact of the lampricide, 3-trifluormethyl- 4-nitrophenol, upon the mudpuppy, Necturus maculosus, in the Grand River, Lake County, Ohio. Pre-treatment and entire season population estimates for 1987 and 1988 were made using Schnabel estimation. Population estimates ranged from 556-1,118 per km in 1987 but declined to 280-397 per km in 1988. Comparison of pre-treatment or entire season estimates between years indicates a minimum 29% decrease in population size. Individuals within the range of total lengths examined appear to exhibit equal sensitivity to TFM. Intensive seining and manual turning of rock slabs were effective methods for capturing Necturus: trapping and electroshocking proved ineffective. Introduction The lampricide 3-trifluoromethyl-4-nitrophenol (TFM) has been used in control of sea lamprey in the Great Lakes drainage since 1958 (National Research Council of Canada, 1985). The chemical has been found toxic to a number of vertebrate taxa, and the degree of toxicity is dependent upon application concentration, temperature, pH and hardness of the water. Relatively little field research has been published concerning effects of TFM on amphibians even though field treatment summaries frequently refer to mortality of amphibians. Gilderhus and Johnson (1980) note mortality in anuran tadpoles in 16% of treatment summaries; mortality in Necturus was noted in 32% of observations from Lake Superior tributaries, 36% 34 Matson No. 45 from Lake Michigan tributaries, and of those observations 18% referenced high mortality. In October 1986, TFM was applied to Conneaut Creek located in Ashtabula County, Ohio and Crawford County, Pennsylvania. Mortality in Necturus was high, but no population data were available either before or after the TFM application. The Grand River in Lake and Ashtabula Counties of Ohio was scheduled to receive TFM treatment during spring of 1987. To assess the impact of the lampricide upon the population of Necturus maculosus , pre-treatment and post-treatment population estimates were made and then compared. For additional comparative purposes, the study was continued during 1989, and statistical comparison of population estimates was made between years. Study Area and Methods Field work extended from March through July 1987 and from April through July 1988. High water, shortage of time, and difficulty in capturing animals limited the number of sites where the population was studied to one. The 600 m study site was located in the Grand River, Madison Twp., Lake County, Ohio (Thompson, Ohio 7.5 min. topographic map, 1960, photorevised 1970). Two meanders, one riffle, three gravel bars, and several pools were major channel features. Pools present were of three types: isolated, flood-plain pools (IFPP); rapid-bounded pools (RBP); and sequential, incised channel-margin pools (SICMP). Water depth ranged from 10 cm to approximately 2 m during the summer months but was subject to dramatic fluctuations from low water under drought conditions during summer, 1988, to spring flood stage in early March or April. Substrate in the river channel consisted of sand, gravel, cobbles, large glacial erratics, and slabs of siltstone from the Chagrin Shale, some of which exceeded two meters in length. Pools had substrates of sand, silt and large boulders or siltstone slabs, whereas others collected detritus and silt to a depth of 10-15 cm. Four methods were used to capture Necturus during 1987: intensive seining with 3.7 m and 7.6 m bag seines, manual overturning of rock slabs capturing animals in a 1 .8 m seine, trapping with baited minnow-like traps, and electroshocking using both back-pack and small pull- behind units. Intense seining was productive from late March to early May and consisted of seining SICMP and RBP repeatedly with assistants churning the substrate before the net. This produced chocolate-colored water with much floating and suspended organic material. Overturning slabs was productive from mid-May through July when DATE OF CAPTURE FIGURE 1. Histogram of the number of Necturus maculosus captured on different dates during 1987 and 1988. Exposure to TFM occurred on 27 April 1987 and is indicated by the arrow. 1990 Estimation of Numbers of Necturus 35 1987 1988 DATE OF CAPTURE FIGURE 2. Number of Necturus maculosus within size cohorts captured on different dates during 1987 and 1988. Exposure to TFM occurred on 27 April 1987 and is indicated by the arrow. water depth was reduced. Trapping and electroshocking were unsuccessful for capturing Necturus. Trapping produced only 1 capture in 126 trap nights, and electroshocking was totally ineffective. Seining and slab turning were the only methods used in 1988. All animals captured were marked by toe-clipping. Individuals marked in the pre-treatment phase during 1987 were marked by removing the outermost toe on the left hind foot; individuals captured during the post-treatment phase of the study were marked in similar fashion on the right. All animals were toe clipped on the left front foot in 1988. The Schnabel method of population censusing and estimation was used to estimate population size before and after treatment of the river with TFM. Chapman’s Poisson table was used to calculate 95% confidence interval estimates for <50 recaptures. The standard normal approximation to the Poisson was used to estimate 95% confidence intervals for >50 recaptures (Seber, 1973). Population estimates were compared using the normal approximation z-statistic (Seber, 1973). Null hypotheses tested were: 1) There was no difference in population size between pre-treatment and post-treatment estimates during 1987; 2) There was no difference in population size between intervals corresponding to the pre-treatment period between years; 3) There was no difference in population size for entire data sets between years. Total length (TL) of each animal was measured to the nearest millimeter (mm) in the field to examine differential mortality associated with size and to compare catchability using various methods. The site was searched for dead Necturus during and following TFM exposure on 27 and 28 April. Results Intensive seining yielded the greatest number of captures and detected most small larvae and juveniles. Seining accounted for 181 of 212 (85.4%) animals examined in the field in 1987 and for 244 of 290 (84.1%) captures in 1988. Overturning rock slabs resulted in capture of 77 animals (15.3%). All but four animals captured using this technique had minimum TL of 150 mm. Few animals were obtained from 21 to 29 March. Number of captures increased in early April, peaked during mid-month, and then decreased to fluctuate about mid- March levels through July (Figure 1). Captures of various size classes of Necturus were not evenly distributed temporally (Figure 2). No juveniles or small larvae were obtained before 29 March due at least in part to technique and high water, and only 51 animals with TL<150 mm were captured after 1 May. Most animals captured between 29 March and 6 May were <150 mm TL and accounted for the elevated number of captures during this period (Figs. 1, 2). TFM treatment was initiated upstream of the site on 26 April and reached the site early morning on 27 April. Repeated searching of the site on the 27th and 28th of April for dead Necturus yielded six individuals of which three were marked. One individual was an adult in excess of 175 mm TL, while the remaining five were under 75 mm TL. One hundred eighty-one (181) Necturus were captured during the pre-treatment phase of the study in 1987, and 31 individuals were obtained during the post-treatment phase. Population estimates and 95% confidence interval estimates are presented in Table 1. The comparison of pre-treatment with post-treatment population estimates for 1987 was invalid because only two Necturus were recaptured during the post-treatment phase having only the post-treatment 36 Matson No. 45 Table 1. Data obtained during two years of study of the Grand River Necturus Population. Comparisons of population estimates between years for both the pre-treatment and the entire season were significantly different (P<0.001). Number Number Number Population 95% examined Marked Recaptured Estimate(N) C.I.E. Pre-treatment 1987 178 153 25 805/km 513, 1188/km 1988 205 120 85 262/km 212, 323/km Entire Season 1987 209 171 * 35 803/km 556, 1118/km 1988 290 165 125 333/km 280, 397/km * 3 marked animals found dead after TFM treatment mark. Comparison of pre-treatment population estimates between years was statistically significant (PcO.OOl), as was the entire season estimate (PcO.OOl) between years. Discussion Cagle (1954) found baited nets, dip nets, seines and trotlines to be ineffective in capturing Necturus maculosus louisianensis in Louisiana. Setlines produced satisfactory catches of adults, and intensive dip netting and seining produced larvae. Shoop (1965) reported traps constructed of 6 mm wire mesh unsuccessful in capturing Necturus maculosus louisianensis and N. beyeri. However, unusually high success was achieved using baited setlines. Although no direct indication of success was stated, dip netting and seining were used by Shoop (1965) for capturing Necturus in Louisiana. Results reported here show that intensive seining produced the greatest number of captures and almost all animals <75 mm TL. Neither setlines nor trotlines were used in this study as both techniques would yield a bias toward large juveniles and adults and would result in a serious underestimate of population size and in increased mortality because Necturus tends to swallow baited hooks. Shoop (1965) used electroshocking with some success to capture Necturus in excess of 140 mm snout-vent length in Big Creek, Louisiana. Eleven of 50 individuals were obtained by electroshocking; the remaining individuals being taken on baited set lines. No indication of time, effort required or success rate was stated. Fitch (1959) was successful in collecting 28 Necturus with the aid of an electroshocker in April at Sugarloaf Lake, Michigan. My results contrast with those in Louisiana and Michigan in that electroshocking was found totally ineffective. Employment of the technique under different water conditions in the Grand River or where animals are not concealed beneath slabs may produce some catches, but it seems improbable that the method is useful for capturing many individuals as is required for population estimates. This assertion is supported by the observation that after using an electroshocker in a SICMP with no Necturus apparent, the pool was intensively seined and 13 individuals, small larvae, juveniles and adults, were captured. The dramatic increase in number of Necturus caught from late March through the end of April is difficult to explain. Bishop ( 1926) mentioned movements or migrations in streams and lakes, and vernal movements from lakes into streams were noted by Pope (1962). Movements into lake shallows during April and May were reported by Fitch (1959) and by Gibbons and Nelson (1968); adult Necturus were reported to retreat to deeper, colder waters in summer, while larvae remained in shallower, warmer waters (Harris, 1958). Movement into shallows does not appear to be the factor involved at this site in the Grand River. On the contrary, nearly all Necturus were obtained during March, April, and early May in SICMP where water depth was greater than in the main river channel. Field technique improved over time during March, 1987, and poor technique may partially explain the low number of captures early in the study. High water in early spring hampered activities in the river in both years and contributed to the low number of captures. However, the same seining techniques which produced high catches in April were less successful from late April through July in the same pools. Furthermore, the capture pattern in 1988 is nearly identical to 1990 Estimation of Numbers of Necturus 37 that of 1987. Small larvae and juvenile classes that were common in SICMP collections from late March through late April or early May were absent thereafter. Movement from pools into the channel may explain these observations. Consequently, for population size estimates or age-size population studies this procedure produces a biased estimate against large juveniles and adults; whereas slab turning produces a bias against larvae and juvenile classes. Inspection of the data from 1987 with respect to the size and number of individuals captured over time suggested that Necturus <150 mm TL may have suffered higher mortality than larger animals (Figure 2). This hypothesis was formulated on the observation that after I May 1987 few animals <150 mm TL were captured. If this hypothesis was correct, then proportionately fewer animals within this size cohort would have been expected to appear in the 1988 capture data. On the contrary, more Necturus within this cohort were captured in 1988; apparently differential mortality did not occur. The total number of Necturus examined in 1988 was higher than that in 1987 (Table 1); improved capture techniques and reduction of water flow and depth resulting from drought are probable causes for higher capture in 1988. My data do not agree with those derived from caged Necturus experiments in the Grand River (Daugherty and Klar, 1988) where observed mortality was restricted to individuals with TL <50 mm. The experimental procedure is not explicitly described by Daugherty and Klar, and it is not stated if individuals of different size classes were placed into a single cage or were placed by similarity of size into separate cages. Furthermore, it is unclear how mortality was observed, i.e. if the animals were found dead or if some were injured or disappeared, perhaps due to cannibalism. They suggest overcrowding in cages as possible cause for observed mortality. This is unlikely for the day or two the animals were confined unless cages were very small and crowding was intense resulting in physical interaction or injury (personal observation). It is possible that mortality observed in caged experiments did reflect increased susceptibility to TFM by small Necturus and that small individuals in pools where much of my data were obtained were able to escape some of the TFM effects by burrowing into the substrate. Both pre-treatment and entire season comparisons between years indicate large reduction in population size from 1987 to 1988 following exposure to TFM within the treatment range of 2.1 mg/1 to 6.6 mg/1 (U.S. Fish and Wildlife chemical treatment summary). The 29% reduction in population size hinges in part upon accuracy of population estimates and was based upon the percentage decrease between the lowest 95% confidence interval estimate for the entire year of 1987 (556) and the upper estimate for 1988 (397). Schnabel estimates may reflect only the general order of magnitude of population size (Schnabel, 1938) and are dependent upon validity of certain assumptions (Seber, 1973; Caughley, 1977). Recruitment from reproduction was not a factor since no hatchlings, which are readily identifiable, were included in the estimates. Immigration and emigration may be factors that effect the population estimates; I am currently investigating these aspects of Necturus biology. Assuming immigration and emigration rates are equal and remain similar between years, they likely have negligible impact on the comparison of population size between years. The assumptions of random sampling and equal catchabi 1 i ty of marked and unmarked individuals probably are valid. Toe clipping temporarily marks Necturus ; however, clipped animals were readily identifiable as marked animals within the few months following marking. Therefore, loss of marks was not considered to effect estimates. Consequently, based upon two years of study, population size and degree of population reduction can only be approximated. Because mortality of Necturus during the TFM treatment was apparent (personal observation; U.S. Fish and Wildlife chemical treatment collection summary, 1987), notable decrease in population size can, at least in part, be attributed to the lampricide. Acknowledgements Partial funding for the research was received from the Ohio Department of Natural Resources, Division of Wildlife and from the Kirtlandia Society of The Cleveland Museum of Natural History. I thank R.A. Pfingsten, D. Waller and an anonymous reviewer for reviewing the manuscript and making helpful suggestions for its improvement. I am indebted to Nick Ashbaugh, Katrine Bosley, and to the numerous volunteers who generously gave their time in the field. References Bishop, S.C. 1926. Notes on the Habits and Development of the Mudpuppy Necturus maculosus (Rafinesque). New York State Museum Bulletin No. 268: 1-60. Cagle, F.R. 1954. Observations on the Life History of the Salamander Necturus maculosus louisianensis . Copeia 1954:257-260. Caughley, G. 1977. Analysis of Vertebrate Populations John Wiley and Sons. Daugherty, W.E. and G.T. Klar. 1988. Sea Lamprey Management in the United States in 1987. U.S. Fish and Wildlife annual report to Great Lakes Fishery Commission. Fitch, K.L. 1959. Observations on the Nesting Habits of the Mudpuppy, Necturus maculosus Rafinesque. Copeia 1959:339-340. Gibbons, J.W., and S. Nelson, Jr. 1968. Observations on the 38 Matson No. 45 Mudpuppy, Necturus maculosus, in a Michigan Lake. American Midland Naturalist 80:562-564. Gilderhus, P.A., and B.G. Johnson. 1980. Effects on Sea Lamprey (Petromyzon marinas ) Control in the Great Lakes on Aquatic Plants, Invertebrates and Amphibians. Canadian Journal of Fisheries and Aquatic Science 37:1985-1905. Harris, J.P. 1959. The Natural History of Necturus: I. Habitats and Habits. Field and Laboratory 27: 1 1-20. National Research Council of Canada. 1985. TFM and Bayer 73 Lampricides in the Aquatic Environment . Publication No. NRCC 22488. Pope, C.H. 1964. Amphibians and Reptiles of the Chicago Area. Chicago Natural History Museum Press. Schnabel, Z.E. 1938. The Estimation of the Total Fish Population of a Lake. American Mathematical Monthly 45:348-352. Seber, G.A.F. 1973. The Estimation of Animal Abundance and Related Parameters. Hafner Press. Shoop, C.R. 1965. Aspects of reproduction in Louisiana Necturus Populations. American Midland Naturalist 74:357-367. United States Fish and Wildlife Service. 1987. Chemical Treatment Collection Summary. Grand River. Lake County, Ohio. United States Fish and Wildlife Service. 1987. Chemical Treatment Summary. Grand River. Lake County. Ohio. •NATURAL HISTORY* Published by The Cleveland Museum of Natural History Wade Oval, University Circle Cleveland, Ohio 44106 Production by Wendy Donkin NUMBER 46 ff i/ELAND, OHIO KIRTLANDIA Anthropology P ale o epidemiology of For otic Hyperostosis in the Libben and Bt-5 Skeletal Populations 1 Robert P. Mensforth Paleontology North American Late Devonian Cephalopod Aptychi 49 Calvin J. Frye and Rodney M. Feldmann •NATURAL HISTORY* KIRTLANDIA The Scientific Publication of The Cleveland Museum of Natural History David S. Brose and Joseph T. Hannibal, Editors Brief History and Purpose Kirtlandia, a publication of The Cleveland Museum of Natural History, is named in honor of Jared Potter Kirtland, a noted nineteenth-century naturalist who lived in the Cleveland, Ohio area. It began publication in 1967 and is a continuation of the earlier series Scientific Publications volumes 1 to 10 (1928-1950), and new series volumes 1 to 4 (1962-1965). Supported by the Kirtlandia Society of The Cleveland Museum of Natural History, Kirtlandia is devoted to the publication of scientific papers in the various fields of inquiry within the Museum’s sphere of interest: Cultural and Physical Anthropology; Archaeology; Botany; Geology; Paleobotany; Invertebrate and Vertebrate Paleontology; Systematic Ecology; and Invertebrate and Vertebrate Zoology. Issues will vary from single monographs to collections of short papers, review articles, and brief research notes. Kirtlandia is abstracted in Biological Abstracts and indexed in Bibliography and Index of Geology and Zoological Record. Associate Editors James K. Bissell, The Cleveland Museum of Natural History Bruce Latimer, The Cleveland Museum of Natural History Martin Rosenberg, Case Western Reserve University Sonja Teraguchi, The Cleveland Museum of Natural History Editorial Advisory Board Rodney Feldmann, Kent State University Michael C. Hansen, Ohio Geological Survey Richard Meindl, Kent State University G. Michael Pratt, Heidelberg University David H. Stansbery, Ohio State University Frederick H. Utech, Carnegie Museum of Natural History Ed Voss, University of Michigan Andrew M. White, John Carroll University Kirtlandia No. 46 ISSN 0075-6245 © 1991 by The Cleveland Museum of Natural History Cleveland, Ohio Copies of Kirtlandia, and many issues of the Scientific Publications series of The Cleveland Museum of Natural History, are available for sale. Write to: Library, The Cleveland Museum of Natural History, 1 Wade Oval Drive, University Circle, Cleveland, Ohio 44106-1767 for a current price list. KIRTLANDIA The Cleveland Museum of Natural History August 1991 Number 46: 1 -47 Paleoepidemiology of Porotic Hyperostosis IN THE LlBBEN AND BT-5 Skeletal Populations Robert P. Mensforth Department of Anthropology Cleveland State University Cleveland, Ohio 44115 Abstract The frequencies with which porotic hyperostosis occurred in the Libben Late Woodland (n = 580) and Bt- 5 Late Archaic (n = 247) skeletal populations were examined. Aside from temporospatial and cultural affiliations, the skeletal samples representing these two prehistoric band level societies principally differ with respect to level of subadult mortality. That is, juvenile death rates were substantially greater for the Libben group. The goals of the study were to (1) identify those factors which played a paramount role in the etiology of the skeletal lesion in the two groups and (2) evaluate the extent to which porotic hyperostosis serves as a useful bioassay of health status in earlier human groups. The epidemiological patterns that were observed support the conclusion that Libben and Bt-5 porotic hyperostosis was a fundamental consequence of iron deficiency anemia. In addition, the only significant difference in the frequency of porotic hyperostosis that was observed between the two groups was confined to subadults. Here, Libben children displayed a significantly greater frequency of unremodeled lesions and total lesions compared to those at Bt-5. Evidence is presented which suggests that dietary inadequacy and parasitism played a minor role in the etiology of porotic hyperostosis in the two groups. Alternatively, local environmental circumstances associated with habitation and exploitation of the Black Swamp may have played a fundamental role in elevating infectious disease loads resulting in a greater prevalence of iron deficiency anemia in Libben infants and children compared to those at Bt-5. Factors which contribute to a low level of non-specificity, and age related differences in biological and demographic sensitivity for porotic hyperostosis are discussed. ? Mensforth No. 46 Introduction Diseases which become manifest in the human skeleton have piqued the curiosity of physical anthropologists for a considerable period of time. These include both (a) specific and (b) non-specific skeletal and dental lesions. A specific skeletal lesion is one that can be attributed to a single disease entity (eg., multiple myeloma), whereas a non- specific lesion is one that can be induced by a wide variety of disease states and which must therefore be regarded as multicausal . The etiology of a specific skeletal lesion can usually be established with a high level of confidence, and often provide us with information about the health status of a particular individual. However, such abnormalities are infrequent in small human groups. Because of this, specific skeletal lesions provide us witli very limited knowledge about the major forces of morbidity and mortality which operated in earlier human groups. The paleoepidemi- ological utility of specific skeletal lesions is therefore markedly restricted. Alternatively, non-specific skeletal lesions, though multicausal in nature, are sufficiently frequent to yield epidemiological information on a populational basis. Thus, it is not remarkable that paleoepidemiologists now employ a variety of non-specific indicators of disease and nutritional stress in order to evaluate the potential significance of pathological responses that occurred in earlier human groups. Such responses are presumed to reflect demographic adaptations to diverse and often changing environmental, sociocultural, political, and socioeconomic circumstances. The stress indicators which have been used most frequently include: (1) enamel hypoplasias, (2) enamel hypocalcifications, (3) radiographic evidence of long bone growth perturbations as reflected in Harris lines, (4) patterns of long bone diaphyseal lengths attained at each age, (5) cortical bone remodeling dynamics, and (6) skeletal lesions such as porotic hyperostosis and periosteal reactions (Goodman et al ., 1984; Cohen and Armelagos, 1984). It is important to realize, however, that non-specific stress indicators are not of equal paleoepidemiological utility. This is due to the fact that the underlying genetic and environmental factors which interact to determine the phenotypic expression of skeletal and dental stress indicators varies considerably. Thus, each stress indicator is characterized by its own tissue-specific set of benefits and liabilities as a potential bioassay. Skeletal biologists who employ non-specific indicators of disease and nutritional stress to evaluate population fitness in earlier human groups must therefore address two cardinal issues. The first concerns etiology of the stress indicator in question. Thus, in order to identify probable cause of a non-specific lesion on a populational basis, differential diagnoses must be used to evaluate a series of tenable explanatory hypotheses. Those hypotheses which best explain the age and sex specific epidemiological patterns that are observed must then be ascribed greater valence as a source of inference. The second issue concerns the extent to which a particular population pathology serves as a useful bioassay of disease and/or nutritional stress in earlier human groups. This problem can be approached by examining the demographic sensitivity of our paleoepidemiological tool. Therefore, if a skeletal response is to be regarded as a useful indicator of disease stress, it would be predicted that differences in the age and sex specific frequency distributions of the lesion will correspond in direction, though not necessarily in magnitude, to differences in age and sex-specific mortality rates that are observed for two or more groups. For those population studies where the relationship between lesion frequencies and mortality rates are found to be discordant, or totally non-existent, then the epidemiological utility of the bioassay must seriously be questioned. In other circumstances the relationship between lesion frequencies and mortality rates may be weak, but nonetheless apparent. Here, we must carefully consider the extent to which potential interpretive value of a lesion outweighs the manifest limitations that characterize such a lesion. Furthermore, in those circumstances where the demographic sensitivity of a particular bioassay is found to be compromised, the paleoepidemiologist must critically evaluate the potential roles that other factors may have played in generating the patterns of lesion distribution that were observed. The latter include various physiological and developmental phenomena, methodological and sampling problems, cultural practices, etc. Descriptive research in paleopathology has thus far contributed greatly to our knowledge about the distribution, and nature, of diseases that have influenced the course of human evolution in time and space. The primary value of such studies is clearly recognized. However, the view posited here is that paleoepidemiology should be regarded as a subdiscipline of paleodemography, and should not be considered as a parallel, or independent, line of inquiry. Moreover, studies in paleopathology which are not conducted in a demographic context must also be regarded as devoid of evolutionary significance. Thus, it is suggested here that non-specific stress indicators of maximum value will be those which possess three fundamental properties described as follows: 1) The stress indicator should exhibit a high degree of biological sensitivity to one or more environmental perturbations. 2) The stress indicator should be characterized by a high level of demographic sensitivity such that population patterns in the frequency of occurrence 1991 PA LEOEP1DEMIOLOG Y OF POROTIC HYPEROSTOSIS 3 C FIGURE 1. Libben children, la, Libben child (KSU-08027, 7 years) that displays extensive unremodeled parotic hyperostosis affecting the superior orbital plates, lb, Libben child (KSU-04035, 12-18 months) which exhibits a pronounced unremodeled porotic hyperostotic lesion affecting the right parietal bone fragment. Note the enlarged pore channels enveloped by a well defined microporous cribriform mesh, lc. Shown here are the complimentary photographic and radiographic images of a right parietal bone cross-section. This Libben child (KSU-01258, 3 years) widening of diploic spaces, irregular trabeculation, and the hair-on-end striations that are characteristic of erythroid bone marrow hyperplasia. of a lesion (i.e., morbidity) can clearly be related to patterns of age and sex specific mortality. 3) The stress indicator should be characterized by a low level of non-specificity. That is, we should be able to infer probable cause of the stress indicator, on a populational rather than individual basis, with a relatively high level of confidence. Given these perspectives, the purposes of this monograph are to (1) examine in detail the factors which promote the development of erythroid marrow hyperlasia associated anemias in human groups, (2) establish the demographic significance of such anemias in human groups, (3) develop hypothetical model of iron deficiency anemia related porotic hyperostosis that would be expected to occur in earlier human groups, (4) conduct a case study which explores the extent to which this model provides insight with regard to the etiology of porotic hyperostosis as seen in two prehistoric band level societies, and (5) evaluate the demographic sensitivity and paleoepidemi- ological utility of porotic hyperostosis as a bioassay of population fitness in earlier human groups. Porotic Hyperostosis in Prehistory Porotic hyperostosis is a descriptive term for cranial lesions that display a coral, cribriform, or sieve-like porosity (Angel, 1966; Cule and Evans, 1968; El-Najjar, 1975). Common referents for the skeletal lesion that have been reported in the earlier literature include: cribra orbitalia (Welcker, 1888), symmetrical osteoporosis (Hrdlicka, 1914; Williams, 1929), osteoporosis of the cranium (Muller, 1935), cribra cranii (Williams, 1929; Henschen, 1961), and spongy hyperostosis (Putshcar, 1966). The lesions are often symmetrical in disposition, exhibit variable degrees of osseous tissue hypertrophy, and frequently display peripheral hypervascular channels (Hengen, 1971). These skeletal changes are usually visible upon close macroscopic examination (see Figures la-c). Radiographic and histological studies of porotic hyperostosis show that the lesions are characterized by irregular trabeculation, widening of diploic spaces, thinning of the endo-and-ectocranial tables (primarily the outer), and a radial pattern of bone spiculation known as hair-on-end striations which are oriented perpendicular to the endocranial table of an affected bone (Moseley, 1965; 4 Mensforth No. 46 Greenfield, 1975; El-Najjar and Robertson, 1976). Porotic hyperostotic skeletal changes can be observed most frequently affecting the anterior portion of the superior orbital plates (Hengen, 1971). Lesions affecting the pericranial surfaces of the frontal, parietal, and occipital bones are more commonly seen in subadults (Carlson et at ., 1974). Endocranial skeletal changes tend to be uncommon relative to those which are described above (Henschen, 1961; Hengen, 1971). Severe manifestations of the lesion are accompanied by excessive osseous tissue hypertrophy, obliteration of the outer table, marked resorption and thinning of the inner table, and involvement of temporal, sphenoid, and facial bones (Moseley, 1965). Skeletal changes that are identical to those described as porotic hyperostosis have been identified in a number of pathological conditions (see Table 1). Radiographic evidence of porotic hyperostosis has been observed most frequently as a consequence of iron deficiency anemia and the chronic hemolytic anemias (Sheldon, 1936; Caffey, 1937; Eng, 1958, Sax, 1963; Baker, 1964; Powell et al. , 1965; Aksoy et al., 1966; Lanzkowski, 1968; Agarwal et al., 1970; Shahidi and Diamond, 1960; Britton et al., 1960; Burko et al., 1961; Moseley, 1974; and Williams et al., 1975). The lesions are less frequently seen in association with cyanotic congenital heart disease (Nice, 1964), hereditary spherocytosis (Trucco and Brown, 1967), polycythemia vera (Dykstra and Halberstma, 1940), and pyruvate kinase deficiency (Becker et al., 1971). For all of the conditions listed above the skeletal changes that affect the cranium are the direct result of erythroid bone marrow hyperplasia which occurs in response to an underlying anemic stimulus (Moseley, 1974). Thus, porotic hyperostosis is best considered a non-specific consequence of bone marrow proliferation. The earliest descriptive reports that identified porotic hyperostosis in prehistoric human groups were published in the late 19th and early 20th centuries (Welcker, 1888; Wood-Jones, 1910; Hrdlicka, 1914; Moore, 1929; Hooton, 1930; and Muller, 1935). Although several hypotheses were posited to explain the skeletal lesion, the pathogenesis of porotic hyperostosis remained vague and enigmatic prior to the mid 1960‘s (Henschen, 1961; Nathan and Haas, 1966; Angel, 1964; and Moseley, 1965). Historical reviews of the earlier accounts are available in the more recent literature (see El-Najjar and Robertson, 1976; El-Najjar et al., 1976). In general, the earlier studies demonstrated that porotic hyperostosis occurred in skeletal groups of widespread temporospatial distribution, and that the lesions were somewhat more common in populations that had lived in the equatorial regions of the Old and New Worlds. Currently, two hypotheses are regarded as primary explanations for the etiology of porotic hyperostosis that has been observed in earlier human groups. One considers TABLE 1. List of the Major Clinical Conditions Where Radiographic Evidence of Skeletal Changes in the Cranium Have Been Reported as a Primary Consequence of Erythroid Bone Marrow Hyperplasia I. Congenital Hemolytic Anemias: A. Thalassemias 1 . Thalassemia Major (i.e., Mediterranean Disease, Cooley’s Anemia, Erythroblastic Anemia) 2. Thalassemia Intermedia - Severe Heterozygous 3. Thalassemia Minor - Mild Heterozygous B. Sickle Cell Disease 1. Sickle Cell Anemia (Hemoglobin S- Homozygous) 2. Hemoglobin C - Homozygous 3. Hemoglobin E - Homozygous 4. Hemoglobin S-C 5. Hemoglobin S - Thalassemia 6. Other Infrequent Abnormal Hemoglobins C. Hereditary Nonspherocytic Hemolytic Anemias 1. Glucose-6-Phosphate Dehydrogenase Deficiency 2. Pyruvate Kinase Deficiency 3. Other Rare Enzyme Deficiencies D. Hereditary Spherocytosis (i.e.. Spherocytic Anemia and Congenital Hemolytic Jaundice E. Hereditary Elliptocytosis (rare) II. Iron Deficiency Anemia III. Cyanotic Congenital Heart Disease (rare) IV. Polycythemia Vera in Childhood (rare) Adapted from Moseley ( 1965), The paleopathologic riddle of “symmetrical osteoporosis.” American J. of Roentgenology 95:135-142. the lesion to be a fundamental response to some variant of hemolytic anemia that occurred as a consequence of endemic falciparum malaria (Angel, 1966; Zaino, 1964; Ascenzi and Salistreri, 1977). The alternative hypothesis suggests that porotic hyperostosis is a direct result of iron deficiency anemia (Moseley, 1965, Moseley 1966; Hengen, 1971). Historically, both views have been offered to explain the occurrence of porotic hyperostosis in Old and New World skeletal populations alike. The geographic distribution of the lesion in some areas of the Old World due not conflict with the hemolytic anemia hypothesis (Angel, 1967). However, a substantial 1991 Paleoepidemiolog y of P orotic Hyperostosis 5 body of clinical, experimental, and epidemiological research supports the view that iron deficiency anemia played a substantial, if not the major, role in the patho- genesis of porotic hyperostosis in all earlier human groups (Mensforth et al. 1978, and references therein). The major proponents of these divergent perspectives, and evidence offered to support such inferences, will thus be considered. Angel (1967) has favored the view that porotic hyperostosis that is seen in several Old World circum- Mediterranean skeletal groups represents a bony response to the hemolytic anemias of thalassemia and/or sickle cell anemia. The lesion is interpreted to be direct skeletal evidence of balanced polymorphic adaptations which arose in response to the selection pressures of falciparum malaria in pre-modem Old World agricultural communities (Angel, 1966). Evidence given in support of this hypothesis includes geographic distribution and frequency of the lesion in circum-Mediterranean skeletal groups, and overall patterns of cranial and post-cranial skeletal involvement. For example, the greater incidence of slight and healed lesions that were observed in these groups are considered to be those skeletal changes which would be expected to occur in thalassemic heterozygotes. Angel (1967) suggested that the low incidence of severe porotic hyperostosis was due to an early and intense selection against individuals that were homozygous for the condition. Thus, the marked skeletal changes that are often observed in modern thalassemic homozygotes are attributed to modern medical intervention which has allowed affected individuals to survive for longer periods of time. As a consequence, these individuals are assumed to sustain more pronounced degrees of osseous tissue hypertrophy than would have been the case for pre-modem thalassemic homozygotes. In addition, cortical thinning of the long bones and ribs are cited as post-cranial evidence of erythroid marrow hyperplasia which occurred in response to the hemolytic anemia of thalassemia in these earlier human groups (Angel, 1966; Angel, 1967). However, some more recent workers concerned with the origin of porotic hyperostosis in Old World circum- Mediterranean archeological populations consider the skeletal evidence for thalassemia to be highly suspect (Ascenzi and Salistreri, 1977). As an alternative, it is suggested that the lesion may be the result of compensatory erythroblastic anemia which occurs in association with malarial infection alone (Germana and Ascenzi, 1980). The etiology of porotic hyperostosis in New World skeletal groups has likewise been attributed to the hemolytic anemias of thalassemia or sickle cell anemia (Wakefield et al., 1937; Zaino, 1964; Zaino, 1967; Zaino, 1968; Zaino and Zaino, 1969). Evidence in support of this view is weak, and is primarily limited to the observation that a greater incidence, and degree of severity, of the lesion characterizes several New World skeletal groups (for discussion see Moseley, 1965). Angel (1967) considered iron deficiency anemia, primarily as a result of hookworm infestation, to be a more probable cause of porotic hyperostosis in New World populations. Paleo- epidemiological research discussed thus far has focused on a general relationship whereby some form of parasitism (i.e., falciparum malaria or hookworm infestation) has been regarded as the principal factor involved in the etiology of porotic hyperostosis that occurred in Old and New World skeletal populations alike. The major alternative hypothesis is that the majority of porotic hyperostosis that has been observed in human skeletal groups occurred as a result of iron deficiency anemia in direct response to elevated levels of infectious disease and nutritional stress (Moseley, 1965; Moseley, 1966; Hengen, 1971; Carlson et al., 1974; El-Najjar et al., 1975; El-Najjar and Robertson, 1976; El-Najjar et al., 1976; Lallo et al., 1977; Mensforth et al., 1978; Schutte, 1979; Cassidy, 1980; Von Endt and Ortner, 1982; Palkovich, 1985, 1987; Stuart-Macadam, 1985, 1987a, 1987b, 1989; Walker, 1986; Hodges, 1987; Fairgrieve, 1990). Considerable evidence supports this inference. First, contemporary epidemiological studies have demonstrated that iron deficiency anemia is the most common nutritional deficiency that affects human groups on a world-wide basis (Witts, 1966; World Health Organization, 1968; Robbins, 1974; Baker, 1978; Betts and Weidenbenner, 1986). Porotic hyperostosis likewise has a very widespread distribution in Old and New World skeletal populations (Moseley, 1965; Hengen, 1971; El-Najjar et al., 1976; Lallo et al., 1977). Second, the incidence of iron deficiency anemia in modem groups, and the frequency of porotic hyperostosis in skeletal groups, is greater for those populations that subsist (or subsisted) on dietary staples that are low in bioavailable iron (Carlson et al., 1975; El-Najjar and Robertson, 1976; Lallo et al., 1977; Martinez-Torres and Layrisse, 1974; Witts, 1966; Jeliffe and Blackman, 1962; Scrimshaw and Young, 1976; Grantham-McGregor et al., 1974; Burks et al, 1976; and Ashworth et al., 1973). Third, no evidence exists to support the view that any of the hemolytic anemias which arose in response to falciparum malaria in the Old World were operative as selective factors in the pre-Columbian New World (El-Najjar, 1976 and references therein). Nonetheless, hookworm infection most likely contributed to a greater prevalence and incidence of iron deficiency anemia in certain Old and New World human groups (Hengen, 1971). Benefits and Liabilities of Porotic Hyperostosis Studies Paleoepidemiological analyses that have employed samples of sufficient size to investigate porotic hyperostosis in earlier human groups have demonstrated a 6 Mensforth No. 46 marked concordance with the iron deficiency anemia hypothesis (El-Najjar, 1976; Lallo et al., 1977; Mensforth et al. , 1978; Walker, 1986; Stuart-Macadam, 1989). For example, El-Najjar (1976) examined the frequency of occurrence for porotic hyperostosis in 4,146 individuals from fourteen New World skeletal groups. This macrosample was further subdivided into maize- dependent (n= 1,722) and non-maize-dependent (n= 2,424) groups. The rank order frequency of occurrence for porotic hyperostosis in the maize-dependent sample was subadults (31.7%), adult females (25.3%), and adult males (21.6%). Likewise, the rank order frequency of occurrence for porotic hyperostosis in the non-maize-dependent sample was subadults (8.6%), adult females (5.2%), and adult males (2.0%). Thus, two patterns emerge that are in accord with the iron deficiency anemia hypothesis. First, both subsamples exhibit a rank order frequency of porotic hyperostosis which identifies those age/sex groups at greatest risk of acquiring iron deficiency as a result of intrinsic physiological factors alone. Second, the markedly higher incidence of porotic hyperostosis that characterizes the maize-dependent skeletal groups is a pattern which would be expected to occur as a consequence of diets that are low in bioavailable iron. However, there are theoretical biases and method- ological problems which limit the inferential utility of studies such as the one described above. The first problem concerns the relationship between the frequency of porotic hyperostosis and levels of morbidity and mortality that characterize our comparisons of prehistoric human groups. We may find that one skeletal groups exhibits a frequency of porotic hyperostosis that is statistically significantly greater than another skeletal group. However, differences in lesion frequency alone do not allow us to assess the demographic impact that such disease loads may have had on earlier human groups. That is, if morbidity is high and mortality is low, than many individuals who survive periods of elevated risk would be expected to exhibit lesions that had undergone various stages of healing or remodeling. Alternatively, certain high risk age/sex groups may display elevated frequencies of active lesions at time of death. Thus, before we accept porotic hyperostosis as a useful bioassay of disease and nutritional stress we must first determine whether or not differences in the frequency of the skeletal lesion correlate with differences in mortality rates that are observed. If a positive correlation exists, we must then evaluate the strength of such a relationship. If the relationship is strong, we can then ascribe demographic sensitivity, and evolutionary significance, to the non- specific stress indicator. The second problem that characterizes many paleoepidemiological inquiries is an over emphasis on dietary hypotheses. Thus, it is routinely assumed that the nutritional status of prehistoric horticultural and agricultural groups will be impaired by the ingestion of domesticated cultigens. It is well recognized that maize, and other Old and New World cereal grains, are low in bioavailable iron and other essential nutrients (Scrimshaw and Young, 1976). However, we often overlook floral and faunal evidence which indicate that many prehistoric horticultural and agricultural groups supplemented their diets with substantial amounts of nutrient rich plant and animal foods that were acquired via foraging, hunting, and/or trade. Furthermore, it is very probable that few prehistoric groups practiced the intense levels of monocropping that occur in many contemporary under- developed societies where levels of sanitation and hygiene are poor, population density is high, and the incidence of protein deficiency and protein-calorie malnutrition is exceptionally high (Scrimshaw and Young, 1976; Beisel, 1982; Herbert, 1985; Herbert, 1987). Perhaps equally important as diet are the demographic changes that occurred in human groups which made widespread use of domesticated cultigens. Overwhelming empirical evidence has shown that increases in local and regional population densities, combined with the hygiene and sanitation problems of sedentary village life, results in a significant rise in the frequency, duration, and severity of infectious diseases in agricultural communities (Scrimshaw and Young, 1976; Herbert, 1985, and references therein). Moreover, several researchers have documented that half of all nutritional crises are precipitated by infectious disease episodes alone (Gordon et al., 1963; Scrimshaw and Suskind, 1959; Maynard and Hammes 1970; Martorell, 1980). Thus, the extent to which infectious diseases may have contributed to the increased incidence of porotic hyperostosis that occurred in certain prehistoric human groups has not received due consideration in light of its demonstrated importance (Lallo et al., 1977; Mensforth et al., 1978). The primary methodological problems which limit the utility of many porotic hyperostosis studies are ( 1 ) inadequate demographic documentation for the samples being investigated, (2) routine use of broad age intervals for reporting skeletal lesion frequencies, and (3) lack of information about lesion activity status (i.e., active versus healed lesions) at time of death for affected individuals. First, it is important to realize that many human skeletal collections that were assembled during the earlier part of the 20th century are demographically unsound as a result of biased sampling techniques. These include selective recovery of (a) cranial versus post-cranial skeletal remains, and (b) adults versus subadults. Furthermore, the growing interest in human skeletal diseases that fascinated physical anthropologists at the turn of the century may have introduced a further bias 1991 Paleoepidemiology oe Porotic Hyperostosis 7 whereby pathological specimens were recovered in greater numbers relative to non-pathological individuals. The extent to which the latter bias may affect many extant human skeletal samples is difficult to assess in retrospect. Nonetheless, it is quite clear that the demographic composition of many skeletal collections is markedly unbalanced. Thus, the quality of age and sex information that has been reported for such materials remains highly suspect (Weiss, 1972; Weiss, 1973; Ruff, 1981). Second, paleoepidemiologists have often reported skeletal lesion frequencies for very broad, and non- standardized, age groups (i.e., subadults, adult males, and adult females). The use of broad age intervals not only masks useful age-specific patterns in the frequency of occurrence of skeletal lesions, but also masks the effects of census errors. These affects are usually most pronounced for subadults where infant underenumeration can markedly skew the skeletal lesion frequencies that are observed in such groups. Thus, erroneous conclusions about differential health status may arise when skeletal lesion frequencies for demographically balanced skeletal samples are compared to the skewed lesion frequencies of selectively biased skeletal samples. Moreover, the width in years that is used to define the subadult age interval varies from one paleoepidemiological report to the next. Therefore, lesion frequencies that are given for subadults grouped in 0-10 year, 0-15 year, and 0-20 year intervals must be adjusted before meaningful comparisons of health status can be made. However, the information required to carry out such adjustments is all too often lacking in the original reports. Third, most paleoepidemiological studies make no attempt to discriminate between skeletal lesions that are active versus those which are inactive at time of death. This is peculiar given the fact that unremodeled lesions provide us with our only reasonable estimates of the age- specific frequency of occurrence, and levels of mortality, which may have occurred as a direct, or indirect, result of the disease process which produced the skeletal lesion. The total lesion frequencies most often reported are combined measures of unremodeled and remodeled lesions that were observed in a group. These may provide us with a general index of overall morbidity that accompanied a particular stress indicator. However, combined measures which include remodeled lesions may also introduce a significant amount of noise. That is, our ability to estimate the real total frequency with which a skeletal lesion occurred in an earlier human group will be distorted by age progressive bone remodeling. The latter may be further confounded by a sex differential in bone remodeling rates as is the case with Harris lines (see Garn, 1968). In general, these effects will be most pronounced in adults where the (a) frequency of unremodeled lesions that are observed is likely to be lower than the frequency of total lesions that are observed. and (b) the frequency of total lesions that are observed is likely to be lower than the frequency of total lesions that occurred in the group. Significance of the Iron Deficiency Anemia Hypothesis Contemporary epidemiological surveys have shown that iron deficiency anemia is so common in human groups throughout the world that the prevalence of iron deficiency is now regarded as one of the best indices of the nutritional health status of a population (Witts, 1966; World Health Organization, 1968; Kilpatrick, 1970; Baker, 1978; Wen- guang et al., 1986). The biological significance of iron resides in the fact that it is required to sustain important physiological processes. These include hemoglobin synthesis, tissue respiration, enzyme activity, and oxidative-reduction reactions (Macdougall et al., 1975). Iron also plays a paramount role in the maintenance of host immunological competence (Prasad, 1979), and the maintenance of normal epithelial tissue structure and function (Naiman et al., 1969). Furthermore, studies have shown that an individual’s ability to maintain adequate body tissue iron stores and dietary absorption of the nutrient are directly related to health status (World Health Organization, 1968 and references therein). Therefore, the intrinsic physiological and extrinsic environmental factors that are involved in the etiology of iron deficiency anemia will be considered. Intrinsic Physiological Risk Factors that Elevate Risk of Iron Deficiency Anemia Clinical studies have shown that iron deficiency anemia occurs most frequently in those age and sex groups of a population where physiological demands for the nutrient are greatest (Heath and Patek, 1937; Hallberg et al, 1970; Robbins, 1974; Olivares et al., 1986). Thus, four high risk groups can be identified in any human group. These can be further subdivided into two major and two minor high risk groups. The two groups at greatest risk of acquiring iron deficiency anemia are (1) infants and young children that range in age from 6 months to three years, and (2) adult females in their child-bearing years (Finch, 1968a, and 1968b). The two minor high risk groups are (1) adolescent males and females, and (2) men and women over 60 years of age (Scott et al., 1970). Birthweight and rate of somatic growth are now recognized as the two most important constitutional factors which elevate the risk of iron deficiency anemia in infants and young children (Josephs, 1956; Jacobs and Wormwood, 1982; Laskari, 1984; Betts and Weidenbenner, 1986). Infants are born with iron concentrations that are proportional to body weight, and iron stores established at birth are the most important source of the nutrient for the first six months of life (Mackay, 1931; Mackay, 1933). Mensforth No. 46 Indeed, in otherwise healthy full-term infants a rapid rate of somatic growth alone often results in significant tissue iron depletion by six months of age (Josephs, 1956). Thereafter, exogenous sources of iron are required in order to meet the infant’s physiological demands for the nutrient. Thus, it is not remarkable that the combined effects of (1) rapid somatic growth, (2) frequent infections, and (3) weaning diets of poor nutritional quality often give rise to a high frequency of iron deficiency anemia in infants and young children (Josephs, 1936; Josephs, 1953; Betke, 1970; Smith, 1972; Scrimshaw and Young, 1976). In modern well-nourished societies pediatricians often consider iron deficiency anemia in infants to be a transitional state in which the child is more anemic than ill (Sturgeon, 1956; Betke, 1970; Smith, 1972; Thomas et ai, 1977). In developing countries, however, the condition is often much more prevalent, severe, and contributes significantly to elevated levels of subadult morbidity and mortality (Akel et al. , 1963; Manchandra et ai, 1969; Maynard and Hammes, 1970; Ashworth, 1973; Grantham- McGregor et ai, 1974; Burks et ai, 1976). With regard to adult females, pregnancy is the major physiological circumstance that elevates risk of acquiring iron deficiency anemia (Witts, 1966; Yusufji et ai , 1973). Population studies have reported a world-wide prevalence of nutritional anemias in pregnancy that range between fifteen and ninety-nine percent (Witts, 1966; World Health Organization, 1968; Finch et ai, 1968; Scott et a!., 1970). Hunter’s ( 1960) survey of nutritional anemias in pregnancy further showed that ninety-eight percent were the direct consequence of iron deficiency, with folate and B12 deficiencies occupying a minor role in overall etiology. The important intrinsic factors which interact to increase the risk of iron deficiency anemia during pregnancy include ( 1 ) low tissue iron stores at the onset of pregnancy due to the cumulative effects of regular menstrual iron losses, (2) the physiologic hydremia of pregnancy where expanding plasma volume has the net effect of decreasing maternal hemoglobin concentration, packed red cell volume, and red cell count, (3) maternal excretion of iron in urine and sweat, (4) fetal and placental iron requirements, and (5) blood loss at parturition (Finch, 1968a; Finch, 1968b; Pritchard and Scott, 1970). More specifically, the third trimester fetal iron demands are so great that maternal tissue iron stores rapidly become depleted (Yusufji et ai, 1973). At this time maternal iron balance becomes highly dependent upon dietary absorption which may increase three to four times above normal in order to meet excessive maternal and fetal iron requirements (Apte and Iyengar, 1970). These observations are concordant with the finding that iron deficiency anemia in pregnant women is most prevalent during the third trimester (Yusufji et ai, 1973). In addition, the period of lactation following parturition is accompanied by additional iron requirements that amount to 0.5- 1.0 mg. iron per day above the normal 1.5 mg. of iron per day that is normally absorbed from the diet (Finch, 1968a; Finch, 1968b). Among the two minor population age/sex groups that are at elevated risk of acquiring iron deficiency anemia, the principal risk factor in adolescents is accelerated rate of somatic growth (Saddi and Schapira, 1970; Kenney, 1985; Liebman, 1985). The increased tissue iron requirements that accompany the adolescent growth spurt can rapidly deplete body iron stores and establish a latent iron deficient state. In addition, the onset of menarche and subsequent menstrual blood losses contribute to a sex differential whereby adolescent females experience a higher incidence of iron deficiency anemia relative to males (Rybo, 1970; Saddi and Schapira, 1970). Finally, individuals of both sexes that are over sixty years of age are at elevated risk of developing iron deficiency anemia due to age progressive degeneration in the absorptive capacity and efficiency of the gastrointestinal tract (Finch, 1968a; Finch, 1968b). Extrinsic Factors Which Elevate Risk of Iron Deficiency Anemia Diet. — The absorption of dietary iron is dependent upon the biochemical properties of food items that are ingested, and the organism’s physiological controls which are designed to maintain a conservative equilibrium (Conrad, 1970; Davis, 1970). In humans, iron is absorbed primarily in the duodenum, the most alkaline portion of the gastrointestinal tract (Davis, 1970). In order to meet nutritional requirements, iron must be exposed to the mucosa of the small intestine in a soluble state, in reasonable quantities, and for a sufficient length of time (Conrad, 1970). Under normal circumstances the body exerts rigid control over iron homeostasis through conservation, reutilization, and by regulating the processes by which iron losses are replenished (Conrad, 1970). For example, healthy infants absorb approximately ten percent of available dietary iron (Heinrich, 1970). In contrast, iron deficient infants have the capacity to absorb two to three times the normal amount as a means of compensating for iron loss and tissue iron depletion (Andelman and Sered, 1966; Saddi and Schapira, 1970). Studies concerned with the bioavailability of dietary iron have demonstrated that the iron content of food varies tremendously (Wretlind, 1970; Martinez-Torres and Layrisse, 1974; Morck et ai, 1981; Palazzari et ai, 1986). Even among similar food items the iron content may vary depending on how the food is prepared and, for vegetable food items, where it is grown (Bressani, 1958; Cook and Monsen, 1976). For example, heme iron is more readily absorbed than ferrous iron, and the latter is better absorbed than ferric iron (Callender et ai, 1957; Hallberg and 1991 Paleoepidemiology of P orotic Hyperostosis 9 Sovell, 1967; Turnbull et al, 1967). Also, meat products generally contain more iron that is in a readily absorbable form (Layrisse et al, 1968; Layrisse et al, 1969; Martinez- Torres and Layrisse, 1971). The bioavailability of iron is also dramatically influenced by other dietary constituents such as chelating agents (Hwang and Brown, 1965; Kuhn et al, 1968; Davis, 1970). These compounds can either promote or inhibit iron absorption. Ascorbic acid is a chelating agent which promotes iron absorption by producing a water soluble iron complex (Moore et al, 1940). Several sugars and amino acids also facilitate iron absorption by decreasing the precipitation and polymerization of dietary iron (Charley et al, 1963; Pollack et al, 1964; Martinez-Torres and Layrisse, 1970). In contrast, compounds such as phytates, phosphonates, carbonates, and oxylates strongly inhibit the absorption of dietary iron by effectively binding iron into insoluble macromolecules (Hegsted et al, 1949; Sharpe et al, 1950; Foy et al, 1959; Hussain and Patwardhan, 1959; Conrad, 1970). Qualitatively superior and inferior dietary regimens have been reported in association with iron deficiency anemia in children (Davidson et al, 1935; Josephs, 1956; Dawson and Desforges, 1958; Woodruff, 1958). In general, a consistent relationship between anemia and artificial, or prolonged, milk feeding has been demonstrated in modem groups (Mackay, 1931; Fullerton, 1937; Smith, 1972). Epidemiological studies have also confirmed that prolonged breast feeding and weaning diets of maize or com gruels often occur in association with a high incidence of iron deficiency anemia in infants and young children (Ashworth, 1973; Jelliffe and Blackman, 1962; Grantham- McGregor et al, 1974). These findings are attributed, in part, to the high phosphorous content in milk and com, as well as the high concentration of phytic acid in com, which inhibits the absorption of dietary iron (Lanzkowski and McKenzie, 1959; Martinez-Torres and Layrisse, 1974). Another factor to be considered with regard to nutritional status, in infants and children in particular, is the extent to which acute and chronic gastrointestinal infections lead to the malabsorption of dietary iron and other essential nutrients (Conrad, 1970; Carpenter and Sack, 1981; Gryboski and Walker, 1983; Santos, 1986). Diarrheal episodes promote malabsorption by increasing intestinal motility (Gordon et al, 1963; Fagundes-Netto, 1984). Therefore, both the quantity of iron and the amount of time that it is made available to the absorptive surfaces of the intestinal tract are decreased. In addition, diarrheal episodes are accompanied by dehydration, electrolyte imbalance, negative nitrogen balance, loss of appetite, and the substitution of solid foods by starchy gruels of lower nutritional quality (Gordon et al, 1963; Scrimshaw, 1964). Moreover, tissue iron depletion alone appears to be an important factor which promotes malabsorption syndrome in iron deficient individuals (Naiman, 1969). Severe degenerative epithelial tissue changes (i.e., glossitis, stomatitis, and koilonychia) are not commonly observed in iron deficient adults unless the condition persists for an indefinite period of time (Halsted et al, 1965; Yusufji et al, 1973). However, iron deficient infants and children are reported to experience a relatively high incidence of gastrointestinal dysfunction that results in malabsorption syndrome (Hawksley et al, 1934; Wilson et al, 1962; Halsted et al., 1965; Guha et al, 1968; Naiman et al, 1969). Tissue iron depletion alone is a major factor responsible for defects in epithelial tissue structure and function that occurs in iron deficient children (Naiman et al, 1969). The most important cytological defects include: atrophy of the gastric and duodenal mucosa, diminished synthesis and secretion of gastric acid, and reduced gastrointestinal enzyme activity (i.e., mucosal disacharidase and cytochrome oxidase activity) (Guha et al, 1968; Halsted et al, 1965; Mahoney and Hendricks, 1975). Atrophy of the gastric mucosa is usually the most frequent and pronounced tissue change that occurs in iron deficient children (Davidson and Markson, 1955; Badenoch et al, 1957; Rawson and Rosenthal, 1960). The functional consequence of atrophic gastritis is a reduction in gastroferrin synthesis and gastric acid secretion (Stewart, 1937; Shearman et al, 1966; Ghosh et al, 1972; Smith, 1972). The latter abnormality has important implications for those individuals who already exist in an overt state of iron deficiency. Under normal circumstances gastric acidity promotes the absorption of dietary iron and calcium by preventing the formation of insoluble macromolecules (Cook et al., 1964; Mahoney and Hendricks, 1975). The altered gastric pH which occurs more commonly in iron deficient children exacerbates the circumstance by further reducing the amount of dietary iron that is available to the host in a soluble form which can readily be absorbed in the duodenum. Gastric mucosal atrophy also effects erythropoiesis by interfering with the synthesis and secretion of intrinsic factor (Guyton, 1976). This compound is produced by parietal cells of the gastric mucosa and plays a significant role in facilitating the absorption, and preventing the digestion, of vitamin B j 2 in the intestinal tract. Vitamin B 1 2 is required for red blood cel! maturation. In the absence of intrinsic factor vitamin B j 2 absorption is impaired. The process of hemoglobin synthesis is not affected, but a macrocytic anemia is a common consequence (Guyton, 1976). Though less pronounced, atrophy of the duodenal mucosa and reduced enzyme activity of the small intestine promotes the malabsorption of fats, some carbohydrates, and contributes to occult blood loss in iron deficient subjects (Wilson et al, 1962; Naiman et al. , 1969). Furthermore, iron deficiency anemia in children is regarded 10 Mensforth No. 46 as one of several protein-losing enteropathies (Lahey, 1962). Pathological effects include malabsorption of dietary amino acids and leakage of plasma proteins in the GI tract. Thus, cytological damage to the GI tract which occurs in iron deficient children has a negative impact on nutritional status in general, and further reduces the bioavailability of dietary iron at a time when physiological demands for the nutrient are great. Infectious disease: impaired immune response and nutritional immunity. — The synergistic relationships between iron deficiency anemia and infectious disease has received considerable attention in recent years. Clinical and experimental studies have repeatedly demonstrated that iron deficient humans and laboratory animals experience a greater incidence and severity of infectious disease episodes compared to normal healthy subjects (Shaw and Robertson, 1964; Werkman et al., 1964; Kilpatrick, 1970; Baggs and Miller, 1975). In humans, this relationship is particularly marked for children under five years of age (Andelman and Sered, 1966; Arbeter et al. , 1971; Joynson et al., 1972; Chandra, 1973; Chandra and Saraya, 1975; Macdougall et al., 1975; Scrimshaw and Young, 1976; Krantman et al., 1982; Chandra, 1985; Walter et al, 1986). Indeed, the high incidence of respiratory and gastrointestinal infections that occur in iron deficient infants and children significantly contribute to growth retardation and elevated levels of subadult morbidity and mortality in many contemporary under-developed societies (Witts, 1966; Gordon et al., 1967; Jose and Welch, 1970; Maynard and Hammes, 1970; Krantman et al, 1982; Forman et al, 1984; McMurray, 1984; Hercberg et al., 1986). Thus, the interactions of anemia and infection have important demographic consequences for populations that are subjected to intensified levels of disease and nutritional stress. One of the most significant findings in recent years concerns the fact that impaired immune response in an early manifestation of tissue iron depletion (Chandra, 1973, 1985; Macdougall et al, 1975; Prasad, 1979). The principal implication is that host resistance to infectious disease is compromised prior to the onset of overt iron deficiency anemia. Studies have shown that both (1) cell mediated immunity and (2) bactericidal capacity of leukocytes are suppressed in iron deficient subjects (Arbeter et al., 1971; Macdougall et al, 1975; Srikantia et al, 1976; Weinberg, 1977; Root and Cohen, 1981). Though similar defects in immune response occur in protein-calorie malnutrition, the site of the biochemical defect in iron deficiency differs and does not appear to be directly related to protein deprivation (Macdougall et al., 1975; Purtillo and Connor, 1975; Bhaskaram and Reddy, 1975; Bhaskaram et al., 1977; Prasad, 1979; Hoffman- Goetz and Kluger, 1979; Chandra and Au, 1980; Gross and Newberne, 1980). Suppressed leukocyte function in iron deficient subjects is a direct consequence of tissue iron depletion (Chandra, 1973, 1985). One proposed defect involves myelo-peroxidase (MPO) synthesis (Baggs and Miller, 1975; Yetkin et al, 1979; Parry et al, 1981). MPO is an enzyme that is dependent on heme iron for its activity. It plays a substantial role in the metabolic pathway that mediates the phagocytic capacity of polymorphonuclear (PMN) leukocytes. Simply stated, certain phagocytic cells of the mammalian reticuloendothial system (i.e., eosinophils, neutrophils, and macrophages) require MPO to synthesize hydrogen peroxidase. The latter enzyme is involved in the intracellular processes of lysosomal pathogen killing (Baggs and Miller, 1975). The reduction in leukocyte MPO concentrations, and the subsequent decrease in leukocyte bactericidal activity, appears to be a major factor contributing to the high incidence and greater severity of infections that have been observed in iron deficient children (Arbeter et al, 1971; Macdougall et al., 1975). Recent research has also demonstrated that reduced quantities of the iron-containing enzyme ribonucleotide reductase, and other cytochrome heme enzymes, impair DNA and protein synthesis in iron-deficient subjects (Joynson et al., 1972; Jacobs and Joynson, 1974; Beisel, 1982). The resultant effects include suppressed (a) lymphocyte production, (b) cytotoxic T lymphocyte activity, and (c) neutrophil ferritin and lactoferrin production (Prasad, 1979; Root and Cohen, 1981; Beisel, 1982; Kuvibidila et al., 1981, 1983a, 1983b; McMurray, 1984). Thus, impaired immune response is a direct, and early, consequence of tissue iron depletion that compro- mises host resistance to infectious disease. In addition, the competitive relationship between host and microbe iron requirements, and the subsequent host response to infectious episodes, can substantially reduce the amount of iron that would normally be available to the host for the maintenance of physiological processes. It is now well recognized that iron is a prime nutrilite that is required by bacterial and viral pathogens in order to survive and multiply in mammalian host tissues (Brendstrup, 1950; Weinberg, 1966; Weinberg, 1974). Moreover, many pathogens have evolved the capacity to synthesize and secrete siderophores (Weinberg, 1966; Garibaldi, 1972; Weinberg, 1977; Kochan, 1977b). These are iron-binding compounds which enable the microbe to compete effectively with host iron-binding proteins and host tissues for essential iron (Weinberg, 1977 and references therein). Studies have demonstrated that both the (1) rate of growth and (2) virulence of bacterial pathogens are directly related to the amount of free iron that is available to the microbes (Martin et al, 1963; Shade, 1963; Kaye and Flook, 1963; Brubaker et al, 1965; 1991 Paleoepidemiology of Porotic Hyperostosis Weinberg, 1966; Bullen and Rogers, 1968; Polk and Miles, 1971; Kochan, 1977a, 1977b, 1978; Kluger and Rothenburg, 1979; Hoffman-Goetz el al. , 1981; Hoshishima et al, 1985). Under ordinary circumstances, iron-binding proteins play a critical role in regulating the amount of free iron that is available to pathogens (Weinberg, 1974). Serum proteins such as transferrin, and large concentrations of lactofeirin contained in human breast milk, exert a strong bacteriostatic effect on microbial growth (Bullen and Rogers, 1968; Bullen et al. , 1968; McFarlane et al., 1970; Fletcher, 1971; Bullen et al., 1972; Hanson and Winberg, 1972; Purtillo and Connor, 1975; McFarlane, 1976; Faulk, 1976). This bacteriostatic effect is greatly diminished, or completely abolished, in individuals with severe protein deficiency where cause of death is most often due to overwhelming infections (McFarlane and Hamid, 1973). Iron-binding proteins are therefore considered to be an important non-specific factor in host resistance to infectious disease. The hypoferremia which is induced by host response to local and systemic infections also plays a key role in iron economy (Grieger and Kluger, 1978). It has long been recognized that infectious disease episodes are accompanied by a dramatic reduction in serum iron concentrations (Cartwright et al. , 1946; Greenberg et al, 1947; Kuhns et al. , 1950). This effect is mediated by plasma transferrins which bind free iron and store it in the reticuloendothelial system (i.e. , liver, spleen, and bone marrow) for the duration of the infection (Cartwright et al., 1946; Vannotti, 1957; Kochan, 1978). The physiological response whereby the host induces a hypoferric state in an attempt to deprive the pathogen of necessary iron has been termed nutritional immunity by Weinberg (1974). Therefore, iron which would normally be mobilized for hemoglobin synthesis is sequestered from both host and pathogen for the duration of the infectious episode. Mild infections of short duration usually have no significant effect on hemoglobin levels in adults (Kuhns et al, 1950; Chandra, 1985). If the infection persists a normochromic normocytic, or slightly hypochromic microcytic, anemia can develop. This is commonly referred to as the anemia of infection (Cartwright et al, 1946). However, among infants and children both mild and severe, acute and chronic, infections can result in a marked decrease in hemoglobin levels (Mackay, 1933; Davidson and Fullerton, 1938; Manchandra et al, 1969). In effect, erythropoiesis is suppressed and hemoglobin synthesis is markedly inhibited (Cartwright et al. , 1946). Furthermore, the regain of normal erythropoietic activity following recovery from an infectious episode can be delayed for several weeks to several months in young children (Fullerton, 1937). Thus, respiratory and gastrointestinal tract infections that occur during periods of rapid growth can interfere with the bioavailability of iron and precipitate, or exacerbate, an iron deficient state. Iron deficiency anemia may ensue even though the amount of iron available in the diet is more than adequate. Some clinical researchers have challenged the concept of nutritional immunity and have questioned the adaptive significance of hypoferremia due to infectious episodes (Srikantia et al, 1976; Prasad, 1979). These researchers emphasize the fact that impaired immune response is an early consequence of tissue iron depletion. It is argued that a hypoferric state would further compromise immunological competence of the host. However, these investigators have failed to consider that during infectious episodes iron, particularly heme iron, is diverted to those tissues of the immune system, the reticuloendothelial system (RES), which require it most in order to maintain, or enhance, cell- mediated immunity, phagocytic activity, and the bactericidal capacity of PMNs. Studies have shown that infections inhibit RES erythropoietic activity and hemoglobin synthesis (Kuhns et al, 1950 and references therein). At the same time, however, infections stimulate RES myeloid tissues (Cartwright et al, 1946; Smith, 1972). The implication is as follows. Although heme iron may not be available for hemoglobin synthesis, it may well be used to promote protein synthesis, lymphocyte production and cytotoxic competence, and leukocyte phagocytic activity and bactericidal capacity (Vannotti, 1957; Kochan, 1978). Viewed in this perspective, host response to infectious disease (i.e., hypoferremia) and its subsequent effects on the differential bioavailability of iron appear to be complimentary physiological adaptations. Cultural practices and parasitism. — In addition to constitutional factors, diet, and microbial infection, a number of cultural and environmental variables are also known to play a role in the etiology of iron deficiency anemia (Shah and Seshadri, 1985). For example, Gordon and associates (1963) have shown that (1) culturally prescribed weaning practices and (2) age and sex specific food restrictions and/or taboos, may contribute to patterns of chronic malnutrition in certain human groups. Likewise, an important environmental variable in tropical and subtropical regions of the world is parasitic hookworm infestation (Roche and Perez-Gimenez, 1959; Bradfield et al, 1968; Venkatachalam, 1968). This promotes intestinal blood loss which can lead to chronic iron deficiency anemia in some individuals. With respect to trends in morbidity and mortality, it is worthy of comment that adults and subadults generally tolerate uncomplicated iron deficiency anemia rather well. In children the anemic condition may further operate as a 12 Mensforth No. 46 homeostatic mechanism that functions to balance iron metabolism during developmental periods of fluctuating supply and demand. Such a physiological homeostat has a strong selective value for rapidly growing organisms that must constantly balance iron economy with respect to (1) nutritional requirements and bioavailability on the one hand, and (2) host nutritional immunity in response to infectious diseases on the other. Arbeter and associates (1971) have expressed a similar view with respect to the iron deficiency anemia that accompanies protein deficiency. These workers suggest that the anemia may be an adaptation to the lowered metabolism of protein- deprived tissues. The epidemiological relationships discussed thus far can be summarized as follows: 1) Although prevalence may vary, iron deficiency anemia occurs in all human groups, and is the single most common nutritional disorder that affects humans on a world-wide basis. 2) Intrinsic physiological risk factors play a major role in predisposing certain age/sex groups to iron deficiency anemia. The age/sex groups at greatest risk are infants and young children six months to three years of age, and adult females in their peak period of fertility. Those age/sex groups at lesser risk are adolescents, females in particular, and elderly individuals. 3) Infants and young children represent those individuals at greatest risk of elevated morbidity and mortality as a consequence of physiological dysfunctions accrued in association with iron deficiency anemia. These include impaired immune response and malabsorption syndrome which may occur as a result of marked tissue iron depletion alone. 4) The prevalence of iron deficiency anemia is sensitive to differences in extrinsic environmental factors that vary in human groups. Thus, a high incidence of iron deficiency anemia has repeatedly been observed in societies that (a) subsist on diets low in bioavailable iron, and (b) experience elevated levels of infectious diseases. Thus, it is clear that iron deficiency anemia has a measurable impact on the health status and demographic characteristics of extant human populations. Similarly, if porotic hyperostosis in earlier human groups occurred as a primary consequence of iron deficiency anemia, it would be reasonable to infer that strong demographic correlates exist which argue in favor of the skeletal lesion’s utility as a bioassay of disease and nutritional stress. Hypothetical Model of Skeletal Lesion Differential Sensitivity Paleoepidemiologists are clearly aware that hematological data on the living provide the most accurate, and sensitive, measure of the prevalence, morbidity, and mortality associated with iron deficiency anemia in extant human groups. Skeletal lesions, by comparison, must always be regarded inferior with respect to their ability to assay such conditions in extinct human societies. The extent to which a skeletal lesion conforms to predicted patterns of a particular disease age/sex distribution will thus strongly influence the degree to which such lesions provide information useful for generating inferences about the health status of earlier human groups. With regard to the property of sensitivity. Figures 2a-c illustrate the lesion frequency distributions which would be expected for a hypothetical skeletal response that exhibited high sensitivity, low sensitivity, differential sensitivity based on age and/or sex, and no sensitivity to a hypothetical disease and/or nutritional stress identified here as Disease X. For this simple model let us assume that (a) no sex differences characterize Disease X, (b) three discrete age groups are at risk, and (c) those at greatest risk are subadults, those at moderate risk are adolescents, and those characterized by a lower risk are middle aged adults. For this model the magnitude of Disease X is irrelevant. Hence, frequency can represent the number of individuals affected for a designated radix, or frequency can simply refer to the percent of individuals affected at each age where age is in years. Here, we are only concerned with the extent to which our Active skeletal lesion mimics the epidemi- ological pattern of the hypothetical disorder. Relative to the epidemiological distribution of Disease X , as measured by modern clinical techniques, Figure 2a identifies skeletal lesion frequencies that display high sensitivity (dot screen), and low sensitivity (black screen), to the disorder. Nonetheless, in each case the pattern of lesion distribution corresponds well with that of Disease X , and correctly identify those age groups at risk and the differential magnitude of such risk. Figure 2b illustrates the circumstance where a skeletal lesion displays differential age-related sensitivity. Here, the Active skeletal response (dot screen) provides a poor assay of the frequency with which Disease X affected subadults and middle aged adults. Nonetheless, the skeletal lesion provides a useful index of the extent to which Disease X affected adolescents in the population. Finally, Figure 2c illustrates the circumstances where the skeletal lesion itself may be either frequent (dot screen) or infrequent (black screen), but in each case the pattern of lesion distribution bears no relationship to the underlying epidemiological profile of Disease X. With respect to various non-specific skeletal and dental indicators of disease and nutritional stress that are currently 1991 Paleoepidemiology of P orotic Hyperostosis 13 AGE in use, it is clear that many factors may synergize to generate one, or more, of the patterns illustrated above, compromise their sensitivity as a bioassay, and restrict their epidemiological utility. Primary among these are ( 1 ) tissue- FlGURE 2. Hypothetical models of differential skeletal lesion responsiveness to a hypothetical clinical disorder (i.e., disease or nutritional stress). Shown in 2a is the pattern we would expect to observe for a skeletal lesion that displayed high fidelity where age groups at risk are clearly identified. The dot screen distribution portrays a lesion of high sensitivity to the underlying disorder, whereas the black-screen distribution indicates a low sensitivity (i.e., infrequent) lesion pattern. Model 2b illustrates a skeletal lesion ( dot-screen ) that exhibits differential age-related sensitivity to clinical Disease X. Here, the skeletal lesion is sensitive to, and provides a useful bioassay of, the disease as it affects adolescents. However, the lesion pattern is relatively insensitive to the frequency with which Disease X affects subadults and middle aged adults. Model 2c illustrates skeletal lesions which are both common (dot-screen) and uncommon (black-screen). However, neither pattern of skeletal lesion distribution corresponds with the epidemiological profile of Disease X. Skeletal lesions such as these (i.e., those which are insensitive to the underlying clinical disorder) must be regarded as poor, or useless, bioassays of health status in earlier human groups. specific rates of growth, development, maturation, and aging, (2) age and sex related differences in bone remodeling rates, and (3) limited ability to distinguish between active versus inactive lesion status at time of death. Lesions which are active at time of death provide our best measures of potential age-specific mortality associated with a particular disorder, whereas combined measures of active and healed lesion frequencies provide us with our best approximation of overall morbidity for the disorder in earlier human groups. However, differential age and sex related bone-specific remodeling rates will distort indices of morbidity derived from skeletal lesions. Thus, the magnitude of such distortion, and the ways in which paleoepidemiologists deal with such phenomena, must be addressed for each particular skeletal lesion that is used to assay disease response in earlier human groups. Hypothetical Model of Iron Deficiency-Related Porotic Hyperostosis If it is assumed that iron deficiency anemia was the primary factor involved in the etiology of porotic hyperostosis in a prehistoric human group, a hypothetical model of the expected age and sex specific frequency distributions for unremodeled and remodeled skeletal lesions can be posited. Figure 3a identifies the age and sex specific frequency of occurrence that we would expect to see for porotic hyperostotic lesions that were unremodeled (i.e., active) at time of death. Again, depending upon the sensitivity of the skeletal lesion to the underlying disorder, the overall magnitudes (i.e., frequencies) of the skeletal lesion would be expected to vary from one population to the next due to differential interaction of intrinsic and extrinsic risk factors. Nonetheless, it would be predicted that the ages at onset, peak incidence, and remission would conform to the iron deficiency anemia hypothesis and bracket those POROTIC HYPEROSTOSIS 14 Mensforth No. 46 (%) SNOIS31 Q313aOIAI3dNn 30 A0N3nO3d3 (%) A0N3H03H3 NOIS33 Q3"l3QOI6l3d 1991 Paleoepidemiology oe P orotic Hyperostosis 15 age/sex groups at greatest risk of acquiring the nutritional disorder in each group. These are: A, infants and children that are six months to three years of age; B, adolescents; C, adult females during their peak child-bearing years; and D, post-reproductive adult males and females. Remodeled lesion frequency distributions shown in Figure 3b should differ from unremodeled lesions in two fundamental ways. First, the ages at onset, peak incidence, and remission for remodeled lesions should display a lag phase and translate to the right of each high risk component in the age distribution. Second, remodeled lesion frequencies should display continuous decay as a result of age progressive bone remodeling. The cumulative age and sex specific frequency distributions for remodeled porotic hyperostosis that we would expect to observe in a prehistoric skeletal series are shown in Figure 3c. This is an additive representation of information summarized in Figure 3b. Here, it is recognized that several factors will confound our ability to identify remodeled porotic hyperostotic lesions that occurred in association with each high risk age/sex group. The most important among these concerns neurocranial bone remodeling rates. As was discussed earlier, porotic hyperostosis is almost exclusively confined to bony elements of the neurocranium. Growth studies have shown that the human neurocranium achieves 95 percent of adult size by approximately seven years of age (Malina, 1975). Thereafter, neurocranial bone remodeling rates diminish considerably to reach a low level throughout life. It is thus reasonable to assume that (a) many lesions of slight expression that occurred in the early years of life will be completely remodeled away by adulthood, and (b) a substantial proportion of remodeled lesions that are seen in adults will represent stress episodes that occurred in late infancy, childhood, adolescence, or any combination of these. This results in a circumstance where remodeled lesions will accumulate in young adults, even though these lesions are continuously undergoing reduction in frequency due to a slow rate of age progressive bone remodeling. The remodeled lesion frequency distributions illustrated in Figure 3c therefore represent a more realistic expectation of the cumulative remodeled lesion distributions that would be observed for porotic hyperostosis in earlier human skeletal groups. With regard to the demographic focus of the hypothetical porotic hyperostosis model, the two most important age/sex groups of interest to the paleoepidemiologist are infants and young children and adult females in their child-bearing years. Elevated levels of subadult mortality that occur as a result of disease and nutritional stress will directly influence both the (a) number of individuals that survive to adulthood, and (b) mean fertility rate that adults must achieve in order to replace the population in succeeding generations. In addition, the latter requirement influences the degree to which nutritional stress may affect the reproductive performance of adult females. It is thus reasonable to presume that such additional compromises may further elevate the mean fertility requirements of the group. Background for the Libben and Bt-5 Skeletal Populations Libben is a multi-component Late Woodland cemetery site located in the Black Swamp on the banks of the Portage River in Ottawa County, Ohio. The site was excavated in 1967-1968 and yielded the human skeletal remains of 1,327 individuals that ranged in age from 4 months in utero to 50+ years (Lovejoy et al. , 1977).1 Radiocarbon dates ranging from A.D. 850 to A.D. 1250 indicate that the Libben site was formed over a 400 year period, with primary use of the cemetery concentrated in a 200 year span from A.D. 900 to A.D. 1100. Based on ceramic and lithic analyses the Libben site has been assigned to the Younge phase of the Western Basin Tradition. In contrast to earlier reports (Prufer and Shane, 1976; Lovejoy et al ., 1977), no evidence presently exists to support the conclusion that Libben was a semi-permanent village occupied on a year-round basis. It is now clear that the site represents one of a small number of pan-regional cemeteries that was inhabited seasonally, and discontin- uously, for several generations (D. Brose, personal communication). The aboriginal peoples that created the Libben site most likely consisted of a small number of culturally affiliated bands which inhabited the northwestern and southwestern shores of Lake Erie, and that placed heavy reliance on local dietary resources (Keenlyside, 1978; Lennox, 1982; Krakker, 1983, 1984; Ferris and Mayer, 1990). These included an abundance of freshwater fish, small mammals, migratory birds, acorn and hickory nuts, several species of berries (Harrison, 1978). In addition, a small quantity of maize was recovered from the site indicating that these peoples also indulged in rudimentary maize horticulture, a finding concordant with observations reported for several other Western Basin Younge phase population groups (see Bowen, 1990; and references therein). Nonetheless, it is quite clear maize horticulture was not an important element in the subsistence economy of the Libben people at this time. Given the number of burials recovered (n=1327), the annual crude death rate observed (CDR=.050), and a two hundred year period of primary use, it is estimated that the Libben cemetery was produced by a small regional population of approximately 130 to 150 individuals distributed among several seasonally mobile bands. The Carlston Annis Bt-5 shell midden is a sister site of Indian Knoll. It represents a Late Archaic habitation and cemetery site situated on the banks of the Green River in 16 Mensforth No. 46 TABLE 2. Comparison ofLibben and Bt-5 Age-Specific Mortality Rates Libben Bt-5 Age Libben:Bt-5 Group dx >'x dx dx l'x dx q ratio 0 226 1289 .175 76 354 .215 0.81 1 102 1063 .096 19 278 .068 1.41 3 68 961 .071 10 259 .039 1.82 5 117 893 .131 17 249 .068 1.93 10 94 776 .121 14 232 .060 2.02 15 92 682 .135 27 218 .110 1.23 20 63 590 .107 35 191 .183 0.58 25 78 527 .148 34 156 .218 0.68 30 1 15 449 .256 31 122 .254 1.01 35 154 334 .461 22 91 .242 1.90 40 97 180 .539 20 69 .290 1.86 45 50 83 .602 18 49 .367 1.64 +50 33 33 1 .000 31 31 1 .000 1.00 Total 1.289 354 Symbols: c/v, the absolute number of individuals dead at agev; / 'v, the absolute number of survivors to aget; and qx, the probability of dying in the succeeding age class for those individuals that survive to agev. the Western Coalfield region of Kentucky (Webb, 1950). The site was excavated in the late 1930s and early 1940s and yielded the human skeletal remains of approximately 390 individuals ranging in age from 7 months in utero to 70+ years (Webb. 1950; Mensforth, 1990). 2 Radiocarbon dates ranging from B.C. 5350 to 2515 indicate that the site was formed over a 2500 year period (Winters, 1974). However, recent studies have shown that the site was occupied most intensively over a 1500 year period from approximately 3500 to 4000 y.b.p. (Marquardt and Watson, 1983). Floral and faunal analyses indicate that Bt-5 was a late summer and fall occupation characteristic of a seasonally mobile semisedentary group. The Bt-5 hunter-gatherers likewise placed heavy reliance on local dietary resources. These included mussels, deer, turkey, waterfowl, and an abundance of hickory nuts and acorns (Marquardt, 1972; Winters, 1974; Marquardt and Watson, 1983). Although seven small fragments of squash were recovered at Bt-5, it is very unlikely that this domesticated cultigen contributed to the subsistence economy of these peoples. Alternatively, it has been suggested that people of the Late Archaic Green River Culture may have made occasional use of squash gourds as containers, thus implying a utilitarian function (Marquardt and Watson, 1983). Given estimates of regional site density, length of occupation, and ethnographic analogy, it has been suggested that individual groups ranged in size from 30 to 50 individuals per given time (May, 1969). Paleodemographic reconstructions and composite life table analyses with archetype fertility data are currently available for Libben and Bt-5 (Lovejoy et ah, 1977; Mensforth, 1990). Age-specific mortality rates for the two skeletal populations and mortality ratios which compare the two groups are given in Table 2. It can be seen that infants and children are well represented in each group. Thus, census error due to infant underenumeration is not problematic. The similarities which characterize the Libben and Bt-5 demographic profiles are as follows. Both skeletal groups exhibit (1) type II survivorship curves, (2) high infant mortality, (3) low mortality in the adolescent years, (3) early onset of elevated mortality rates in young adults, and (4) a sex differential in adult survivorship where males over 35 years of age experience higher death rates relative to females. However, the latter trends that were observed in each group are statistically insignificant (Lovejoy et al., 1977; Mensforth, 1990). The manner in which Libben and Bt-5 demographic parameters differ are as follows. Ratios that compare Libben and Bt-5 age-specific mortality rates are given in Table 2. Except for infants in the first year of life and 3rd decade adults, it can be seen that Libben mortality rates exceed those of Bt-5 at all ages. A comparison of 1991 Paleoepidemiology of Porotic Hyperostosis 17 survivorship at age 15 (1 f 5: Libben, 52.9; Bt-5, 61.6) shows that 8.7 percent fewer Libben individuals survived to adulthood relative to Bt-5. This difference in subadult survivorship is statistically significant .05; total lesions %2=16.77, pc. 001). With regard to degree of involvement, it was found that the majority of Libben and Bt-5 porotic hyperostotic lesions were the result of a minimal osseous tissue response to the stimulus of erythroid marrow hyperplasia. Among subadults, lesions of slight expression occurred in 85.1 percent (114/134) of Libben and 87.0 percent (20/23) of Bt-5 individuals. Lesions of moderate expression were observed in only 14.2 percent (19/134) of Libben and 13.0 percent (3/23) of Bt-5 subadults. The majority of moderate tissue responses occurred in specimens that ranged in age from six months to three years of age at time of death in each skeletal series. Only one Libben child displayed severe porotic hyperostosis (see Figure la). No Bt-5 subadults were affected to such a degree. No adults in either skeletal series displayed severe bony changes in response to erythroid marrow hyperplasia. Lesions of slight and moderate expression occurred in 95.7 percent (66/69) and 4.3 percent (3/69) of Libben adults with porotic hyperostosis, respectively. All lesions that were encountered in Bt-5 adults were slight involvements. In addition, all lesions that were identified in Libben and Bt-5 adults were confined to the superior orbital plates. 1991 Paleoepidemiology of P orotic Hyperostosis 21 TABLE 4. Comparison of the Total Frequency With Which Porotic Hyperostosis Was Observed in the Libben and Bt-5 Skeletal Groups Unremodeled Remodeled Total Lesions Lesions Lesions Skeletal Group ni n2 (%) n ? (%) n4 (%) Bt-5 247 25 10.1 26 10.5 51 20.6 Libben* 580 1 16 20.0 87 15.0 203 35.0 Total 827 141 17.0 113 13.7 254 30.7 Symbols: nr total number of specimens that could be examined for the presence/absence of porotic hyperostosis; n2, number of specimens that were scored positive for unremodeled lesions; n number of specimens that were scored positive for remodeled lesions; n4, total number of specimens that were scored positive for remodeled or unremodeled lesions. * Chi square values indicate that the Libben group experienced a statistically significantly greater frequency greater frequency of unremodeled and total porotic hyperostotic lesions compared to the Bt-5 group (unremodeled lesions, %2=1 1 .95; remodeled lesions, ^2=2.94;total lesions, yf= 1 6.77). Age and Sex Specific Frequency Distributions for Unremodeled Porotic Hyperostosis Data enumerating the age and sex specific frequency of occurrence for porotic hyperostosis in the Libben and Bt-5 skeletal groups are given in Tables 5, 6, 7, and 8. Chi square values that compare the total frequency with which remodeled and unremodeled lesions were observed in subadults and adults partitioned by sex in the two skeletal series are listed in Tables 9 and 10, respectively. The general similarities that were observed for each groups are as follows: (1) subadults displayed the highest frequency of unremodeled and total lesions at time of death, (2) adult females showed the second highest frequency of unremodeled and total lesions, and (3) adult males exhibited the lowest frequency of unremodeled and total lesions. The only exception to the pattern described above was that Bt-5 adult females displayed a slightly greater frequency of total lesions than subadults in this group. The Libben rank order frequency for total lesions observed is subadults (43.2%), adult females (35.3%), and adult males ( 16.1%). The comparable information for Bt-5 is adult females (25.0%), subadults (23.0%), and adult males (12.7%). The Libben rank order frequency for lesions that were unremodeled at time of death is subadults (30.3%), adult females (14.3%), and adult males (2.2%). The comparable Bt-5 rank order is subadults (14%), adult females (8.8%), and adult males (2.8%). Chi square values for within group comparisons (see TABLE 5. Age-Specific Frequency of Occurrence for Porotic Hyperostosis in Libben Subadults (0-15 years) Age Group ni Unremodeled Lesions Remodeled Lesions Total Lesions Extra-orbital Lesions n2 (%) n3 (%) n4 (%) n5 (%) 0.0 - 0.5 61 0 0.0 0 0.0 0 0.0 0 0.0 0.5 - 1.0 38 10 26.3 0 0.0 10 26.3 7 18.4 1.0 - 3.0 71 45 63.4 6 8.4 51 71.8 18 25.4 3.0 - 5.0 26 8 30.8 7 26.9 15 57.7 2 7.7 5.0 -10.0 54 14 25.9 17 31.5 31 57.4 4 7.4 10.0 -15.0 60 17 28.3 10 16.7 27 45.0 2 3.3 Total 310 94 30.3 40 12.9 134 43.2 33 10.7 Symbols: n,, total number of specimens that could be examined for the presence/absence of porotic hyperostosis per age group; n2, number of specimens that were scored positive for unremodeled lesions per age group; ip, number of specimens that were scored positive for remodeled lesions per age group; n4, number of specimens that exhibited remodeled and/or unremodeled lesions per age group; ny number of specimens that manifested extra-orbital lesions per age group. 22 Mensforth No. 46 TABLE 6. Age-Specific Frequency of Occurrence for Porotic Hyperostosis in Libben Adults Female Male Age Group "l Unremodeled Lesions Remodeled Lesions Total Lesions »/ Unremodeled Lesions Remodeled Lesions Total Lesions n2 (%) n3 (%) >U (%> n2 (%) n3 (%) n4 m 15-19 18 6 33.3 5 27.8 1 1 61.1 16 i 6.3 4 25.0 5 31.3 20-24 17 5 29.4 3 17.7 8 47.1 17 0 0.0 3 17.7 3 17.7 25-29 17 1 5.9 4 23.5 5 29.4 17 i 5.9 4 23.5 5 29.4 30-34 15 1 6.7 3 20.0 4 26.7 31 0 0.0 4 12.9 4 12.9 35 -39 17 1 5.9 4 23.5 5 29.4 30 1 3.3 2 6.7 3 10.0 40-44 26 2 7.7 8 30.8 10 38.5 11 0 0.0 i 1.9 1 1.9 45 -49 13 2 15.4 1 7.7 3 23.1 9 0 0.0 i 11.1 1 11.1 +50 10 i 10.0 0 0.0 1 10.0 6 0 0.0 0 0.0 0 0.0 Total 133 19 14.3 28 21.0 47 35.3 137 3 2.2 19 13.9 22 16.1 Symbols: n /, total number of specimens that could be examined for the presence/absence of porotic hyperostosis per age group; n2, number of specimens that were scored positive for unremodeled lesions per age group; nj, number of specimens that were scored positive for remodeled lesions per age group; n4, total number of specimens that were scored positive for remodeled or unremodeled lesions per age group. TABLE 7. Age-Specific Frequency of Occurrence for Porotic Hyperostosis in Bt-5 Subadults (0-15 years) Age Group "l Unremodeled Lesions Remodeled Lesions Total Lesions Extra-orbital Lesions n2 (%) n3 (%) n4 (%) n5 (%) 0.0 - 0.5 38 0 0.0 0 0.0 0 0.0 0 0.0 0.5 - 1.0 17 3 17.7 0 0.0 3 17.7 2 12.5 1.0 - 3.0 14 7 50.0 1 7.1 8 57.1 0 0.0 3.0 - 5.0 10 1 10.0 3 30.0 4 40.0 0 0.0 5.0 -10.0 9 1 11.1 3 33.3 4 44.4 1 11.1 10.0 -15.0 12 2 16.7 2 16.7 4 33.4 0 0.0 Total 100 14 14.0 9 9.0 23 3.0 3 3.0 Symbols: n]t total number of specimens that could be examined for the presence/absence of porotic hyperostosis per age group; n2, number of specimens that were scored positive for unremodeled lesions per age group; n ^ number of specimens that were scored positive for remodeled lesions per age group; n4, number of specimens that exhibited remodeled and/or unremodeled lesions per age group; number of specimens that manifested extra-orbital lesions per age group. TABLE 8. Age-Specific Frequency of Occurrence for Porotic Hyperostosis in Bt-5 Adults Female Male Age Group "i Unremodeled Lesions Remodeled Lesions Total Lesions «/ Unremodeled Lesions Remodeled Lesions Total Lesions n2 (%) n r (%) n4 (%) n2 (%) (%) n4 (%) 15-19 9 i 11.1 3 33.3 4 44.4 10 0 0.0 0 0.0 0 0.0 20-24 8 2 25.0 1 12.5 3 37.5 13 i 7.7 1 7.7 2 15.4 25-29 10 2 20.0 1 10.0 3 30.0 12 0 0.0 1 8.3 1 8.3 30-34 9 0 0.0 1 11.1 1 11.1 11 0 0.0 3 27.3 3 27.3 35-39 7 1 14.3 1 14.3 2 28.6 7 0 0.0 1 14.3 1 14.3 40-44 7 0 0.0 0 0.0 0 0.0 8 0 0.0 0 0.0 0 0.0 45-49 8 1 12.5 0 0.0 1 12.5 5 1 20.0 0 0.0 1 20.0 +50 18 2 11.1 3 16.7 5 27.8 5 0 0.0 1 20.0 1 20.0 Total 76 9 11.8 10 13.2 19 25.0 71 2 2.8 7 9.9 9 12.7 Symbols: /? /, total number of specimens that could be examined for the presence/absence of porotic hyperostosis per age group; n2, number of specimens that were scored positive for unremodeled lesions per age group; nj, number of specimens that were scored positive for remodeled lesions per age group; n4, total number of specimens that were scored positive for remodeled or unremodeled lesions per age group. 1991 Paleoepidemiology of P orotic Hyperostosis 23 TABLE 9. Chi Square Values for Within-Group Comparisons of the Frequency of Occurrence for Porotic Hyperostosis in the Libhen and Bt-5 Skeletal Samples Libben Bt-5 Age and Sex Group Comparisons Unremodeled Lesions Total Lesions Unremodeled Lesions Toted Lesions Subadults vs. Adults 44.35° 19.81° 2.78 0.57 Subadults vs. Adult Males 44.26° 30.87° 6.12* 4.56* Subadults vs. Adult Females 12.60° 2.40 0.18 0.10 Adult Males vs. Adult Females 13.19° 13.19° 4.32* 3.62 a: significant at the .001 level of probability. b: significant at the .05 level of probability. Table 9) show that subadults in each group experienced a significantly greater frequency of unremodeled and total lesions compared to adult males, and adult females displayed a significantly greater frequency of unremodeled lesions compared to adult males. In addition, Libben subadults had a significantly greater incidence of unremodeled lesions compared to adult females. Therefore, both skeletal groups are characterized by marked age and sex differences in the frequency of occurrence of porotic hyperostosis. The only exception to this was that Bt-5 subadults and adult females showed no significant differences in the frequency of unremodeled or total lesions that were observed. Chi square values for between group comparisons (see Table 10) show that the only significant difference between Libben and Bt-5 was confined to subadults. Here, Libben subadults displayed a significantly greater frequency of unremodeled and total lesions compared to Bt-5 subadults. The age and sex specific frequencies with which unremodeled porotic hyperostosis occurred in the Libben and Bt-5 groups are illustrated in Figures 5a and 5b. These frequency distributions identify those age/sex groups where erythroid marrow hyperplasia was sufficiently active to invoke a skeletal response at, or near, time of death. It was previously shown that Libben and Bt-5 adults are characterized by a low incidence of unremodeled porotic hyperostosis. Therefore, unremodeled lesion frequency data for adults over 20 years of age are plotted by decade here in order to smooth effects of small subsample sizes combined with low overall incidence. Comparisons of Libben and Bt-5 age-related porotic hyperostosis frequency distributions for subadults are illustrated more clearly in Figure 6. These patterns show that the lesion bears a strong relationship with developmental age. The following general relationships characterize subadults in both skeletal series. No lesions were seen in individuals that were under six months of age at time of death. A low incidence of unremodeled porotic hyperostosis first occurred in the six to twelve month period. During the one to three year period the frequency of unremodeled lesions increased dramatically to reach peak incidence in each group. Subsequently, the frequency of active lesions decreased throughout the childhood years. During the adolescent period the frequency of unremodeled porotic hyperostosis showed a slight increase once again. Figures 6a-c show that the age at onset, peak incidence, and age at remission for unremodeled porotic hyperostosis are virtually identical in Libben and Bt-5 subadults. Chi square values listed in Table 1 1 indicate that Libben subadult age-related patterns for unremodeled porotic hyperostosis are statistically significant. The absence of lesions prior to six months of age, peak incidence in the one to three year period, and subsequent decline are all well defined. Chi square values for the Bt-5 subadults show a similar, though less pronounced, trend. This is primarily TABLE 10. Chi Square Values for Between-Group Comparisons of the Frequency of Occurrence for Porotic Hyperostosis in the Libben and Bt-5 Skeletal Samples Chi Square Values Age Group Unremodeled Lesions Total Lesions Subadults 10.38* 13.09" Adult Females 0.25 2.39 Adult Males 0.08 0.51 Total Adults 0.06 2.26 a: significant at the b: significant at the .001 level of probability. .01 level of probability. 24 Mensforth No. 46 Libben A Bt-5 B F IGURE 5. Age and sex specific frequencies of occurrence for unremodeled porotic hyperostosis that were observed in the Libben (5a) and Bt-5 (5b) skeletal samples. Individuals over 15 years of age are partitioned by sex. a AGE IN YEARS b AGE IN YEARS c F IGURE 6. Histograms and polygon frequencies that compare the age-specific frequencies of occurrence for unremodeled (6a-c) and remodeled (6a-b) porotic hyperostosis in Libben and Bt-5 subadults ranging in age from birth to 15 years at time of death. 1991 Paleoepidemiology of P orotic Hyperostosis 25 TABLE 1 1. Chi Square Values for Within-Group Comparisons of the Age-Specific Frequency of Occurrence for Unremodeled P orotic Hyperostosis in Libben and Bt-5 Subadults Age Group Age Group Skeletal Group 1-3 3-5 5-10 10-15 0.5- 1.0 Libben 13.60" 0.35 0.26 0.05 Bt-5 3.68 0.29 0.19 0.00 b (jj o Libben 6.51h 23.57" 16.02" Bt-5 4.20b 3.65 3.17 b L/i o Libben 1.22 0.21 Bt-5 0.01 0.21 5.0-10.0 Libben 0.69 Bt-5 0.13 a: significant at the .001 level of probability. b: significant at the .05 level of probability. due to the fact that a smaller number of Bt-5 subadults were available for comparison in each age class. Further reference to Figures 6a-c shows that Libben subadults consistently displayed a greater frequency of unremodeled and total lesions per age class compared to Bt-5. Chi square values given in Table 12 indicate that none of the subadult age class comparisons between the two groups are statistically significant. However, this is not surprising given that unremodeled lesions in both groups displayed a marked conformity in age specific distribution, and a much smaller number of Bt-5 subadults were available for comparisons. Among adult females in each group it can be seen that the highest frequency of unremodeled porotic hyperostosis occurred in individuals that were 15 to 30 years of age at time of death. Thereafter, adult females displayed a low frequency of unremodeled lesions. In contrast, adult males in each group showed a low frequency of unremodeled lesions at all ages. Therefore, the patterns that were observed for porotic hyperostosis indicate that ( 1 ) the ages at onset, peak incidence, and remission for unremodeled lesions are markedly similar in the two band level societies, and (2) the frequency distributions that characterize unremodeled lesions identify those age/sex groups at highest risk of acquiring iron deficiency anemia. Age/Sex Frequency Distributions for Remodeled Porotic Hyperostosis Among those subadults that displayed porotic hyperostosis in the two skeletal series (see Tables 5 and 7), remodeled lesions were observed in 29.9 percent (40/134) of Libben and 39. 1 percent (9/23) of Bt-5 individuals. Thus, the majority of subadults with porotic hyperostosis in the two skeletal series had active lesions at time of death. In contrast, the vast majority of adults with porotic hyperostosis had remodeled lesions at time of death. In Libben adults with porotic hyperostosis, remodeled lesions were seen in 86.4 percent ( 19/22) of affected males and 59.6 percent (28/47) of affected females (see Table 6). Similarly, in Bt-5 adults with porotic hyperostosis, remodeled lesions occurred in 77.8 percent (7/9) of affected males and 52.6 percent (10/19) of affected females (see Table 8). TABLE 12. Chi Square Values That Compare the Age- Specific and Total Frequency of Occurrence for Porotic Hyperostosis Among Libben and Bt-5 Subadults Chi Square Values Age Unremodeled Remodeled Total Group Lesions Lesions Lesions 0.5 - 1.0 0.49 — 0.49 1.0 -3.0 0.88 0.03 1.19 3.0 - 5.0 1.66 0.03 0.91 5.0 -10.0 0.93 0.01 0.53 10.0 -15.0 0.70 0.00 0.56 Total \038b 1.09 13.09" a: significant at the .00 1 level of probability. b: significant at the .01 level of probability. 26 Mensforth No. 46 TABLE 13. Porotic Hyperostosis Frequency Data and Chi Square Values That Compare Adult Females 12-30 Years of Age Versus Those +30 Years of Age Lesion Frequency Unremodeled Total Chi Square Values Adult Female Samples Age Group n i ll 2 Lesions <%) Lesions (%) Unremodeled Lesions Remodeled Lesions Libben 15-30 52 12 23.1 24 46.2 5.39b 4.37* +30 81 7 8.6 23 28.4 Bt-5 15-30 27 5 18.5 10 37.0 1.79 3.24 +30 49 4 8.2 9 18.4 Combined 15-30 79 17 21.5 34 43.0 7 22 a 7.22" +30 130 1 1 8.5 32 24.6 Symbols: /, total number of adult females that could be scored for the presence/absence of porotic hyperostosis per age group; ni number of adult females that were scored positive for unremodeled lesions per age group; nj total number of individuals that were scored positive for remodeled or unremodeled lesions per age group; a, Chi square values significant at the .01 level of probability; b, Chi square values significant at the .05 level of probability. The similarities which characterize the age-specific frequency of remodeled lesions that occurred in the two groups are as follows. A low incidence of remodeled lesions first appeared in the one to three year period. The frequency of remodeled lesions gradually increased to reach peak incidence in the five to ten year period in each group. A low frequency of remodeled lesions was once again seen in adolescents. Remodeled lesions in adults were most commonly seen in individuals that ranged in age from 1 5 to 35 years in each group. The only exception to this was that Libben females over 35 years of age showed a remodeled lesion frequency which was 14.8 percent higher compared to young adult females. However, the latter difference is statistically insignificant %2= 1 .04; p>.05). Age Related Porotic Hyperostosis Frequency Distribution in Adult Females Archetype fertility data reported by Weiss (1973) show that age specific fertility rates in contemporary primitive societies are highest in adult females that are 15 to 30 years of age. As a corollary, it would be predicted that peak periods of reproductive activity would predispose young women to a greater incidence of iron deficiency anemia. The latter would give rise to variable levels of erythroid marrow hyperplasia and result in a greater frequency of unremodeled porotic hyperostosis in these individuals. Therefore, Libben and Bt-5 adult female age- related porotic hyperostosis frequency distributions were assessed in the following manner. Information summarized in Table 13 compares the frequencies with which unremodeled porotic hyperostosis, and total lesions, occurred in Libben and Bt-5 adult females that were 15 to 30 years of age versus those that were +30 years of age at time of death. For the 15 to 30 year age group results show that 23.1 percent of Libben females and 18.5 percent of Bt-5 females had unremodeled lesions at time of death. Libben and Bt-5 adult females that were over 30 years of age displayed unremodeled lesion frequencies of 8.6 percent and 8.2 percent, respectively. Thus, young adult females in both groups exhibited a greater frequency of unremodeled lesions compared to their older peers (Libben, 14.5% higher; Bt-5, 10.3% higher). However, chi square values listed in Table 11 indicate that this difference is only significant for Libben adult females. Comparisons of total lesion frequencies that were observed for Libben and Bt-5 adult female age categories yielded similar results. However, these data are inappropriate for the following reason. Remodeled lesions that occurred in association with childhood and adolescent stress episodes are likely to accumulate in young adults where a substantial degree of age- progressive bone remodeling has not yet occurred. In contrast, it would be expected that many lesions of youth would be completely remodeled away in older individuals. Therefore (1) age-progressive bone remodeling alone can give rise to age related differences in the frequency of occurrence for porotic hyperostosis in adults, and (2) combined measures of remodeled and unremodeled porotic hyperostosis that disregard lesion activity status at time of death will result in inflated young adult lesion frequencies and deflated older adult lesion frequencies. 1991 Paleoepidemiology of Porotic Hyperostosis 27 Age/Sex Frequency Distribution for the Libben and Bt-5 Porotic Hyperostosis Macrosample: a Test of the Iron Deficiency Anemia Hypothesis Paleoepidemiological analyses of skeletal pathologies have generally emphasized the comparative approach. Here, the primary goals are to identify and interpret the ways in which two or more skeletal series differ. Nonetheless, the only way to evaluate the extent to which one or more factors may have played a common role in the etiology of a stress indicator is to examine the similarities in lesion frequency patterns that characterize two or more skeletal groups. This can be accomplished, in part, by the use of more refined age categories (eg., 5 year age intervals in adults). However, when this procedure is employed with skeletal samples that are derived from small anthropological populations an additional problem is introduced. Thus, the combined effects of small subsample sizes and low lesion frequencies that characterize various age/sex categories will give rise to unstable, or spurious, lesion frequencies that are of limited inferential utility. The problems described above can be overcome by the use of macrosamples. Here, the age and sex specific skeletal lesion frequencies that are observed for two or more groups are combined. Although between group differences in skeletal lesion frequency will be masked, the macrosample summary data will more clearly identify the age and sex specific lesion patterns that are common to all human groups. The latter permit the skeletal biologist to infer probable cause of a disorder with greater confidence. Data which summarize the age and sex specific frequency of occurrence for porotic hyperostosis in the combined Libben and Bt-5 macrosample are given in Tables 14 and 15. TABLE 14. Age-Specific Frequency of Occurrence for Porotic Hyperostosis in the Combined Sample of Libben and Bt-5 Subadults (0-15 years) Age Group ni Unremodeled Lesions Remodeled Lesions Total Lesions Extra-orbital Lesions n2 (%) n3 (%) n4 (%) n5 (%) 0.0 - 0.5 99 0 0.0 0 0.0 0 0.0 0 0.0 0.5 - 1.0 55 13 23.6 0 0.0 13 23.6 9 16.4 1.0 - 3.0 85 52 61.2 7 8.2 59 69.4 18 21.2 3.0 - 5.0 36 9 25.0 10 27.8 19 52.8 2 5.6 5.0 -10.0 63 15 23.8 20 31.8 35 55.6 5 7.9 10.0 -15.0 72 19 26.4 12 16.7 31 43.1 1 2.8 Total 410 108 26.3 49 12.0 157 38.3 36 8.8 Symbols: total number of specimens that could be examined for the presence/absence of porotic hyperostosis per age group; /?„ number of specimens that were scored positive for unremodeled lesions per age group; n3, number of specimens that were scored positive for remodeled lesions per age group; n4, total number of specimens that were scored positive for remodeled and/or unremodeled lesions per age group; number of specimens that manifested extra-orbital lesions per age group. TABLE 15. Age-Specific Frequency of Occurrence Porotic Hyperostosis Among the Combined Libben and Bt-5 Adult Samples Female Male Age Group Unremodeled Lesions Remodeled Lesions Total Lesions "i Unremodeled Lesions Remodeled Lesions Total Lesions n2 (To) n3 (To) >U (%) n2 (To) n3 (To) »4 (To) 15-19 27 i 25.9 8 29.6 15 55.6 26 i 3.9 4 15.4 5 19.2 20-24 25 i 28.0 4 16.0 11 44.0 30 i 3.3 4 13.3 5 16.6 25-29 27 3 11.1 5 18.5 8 29.6 29 i 3.5 5 17.2 6 20.7 30-34 24 1 4.2 4 16.7 5 20.8 42 0 0.0 7 16.7 7 16.7 35 -39 24 2 8.3 5 20.8 7 29.2 37 1 2.7 3 8.1 4 10.8 40-44 33 2 6.1 8 24.2 10 30.3 19 0 0.0 1 5.3 1 5.3 45-49 21 3 14.3 1 4.8 4 19.1 14 1 7.1 1 7.1 2 7.3 +50 28 3 10.7 3 10.7 6 21.4 1 1 0 0.0 1 9.1 i 9.1 Total 209 28 13.4 38 18.2 66 31.6 209 5 2.4 26 12.1 31 14.9 Symbols: nt, total number of specimens that could be examined for the presence/absence of porotic hyperostosis per age group; n2, number of specimens that that were scored positive for unremodeled lesions per age group; n3, number of specimens that were scored positive for remodeled lesions per age group; n4, total number of specimens that were scored positive for remodeled or unremodeled lesions per age group 28 Mensforth No. 46 B FIGURE 7. Libben and Bt-5 combined macrosample age specific frequencies of occurrence for (7a) unremodeled and (7b) remodeled porotic hyperostosis. The age and sex specific frequency distributions for unremodeled and remodeled lesions in the macrosample are illustrated in Figures 7a and 7b. respectively. The ages at onset, peak incidence, and remission for unremodeled lesions that characterize the macrosample are similar to those that were described earlier for each group. Thus, peak frequencies for unremodeled lesions occurred in young children one to three years of age, and adult females 15 to 25 years of age at time of death. In addition, adolescents and adults over 45 years of age showed slight, but insignificant, increases in the frequency of unremodeled lesions. However, adult males showed a low incidence of active lesions at all ages. Thus, age and sex related macrosample patterns for unremodeled lesions identify those age/sex groups of a population at highest risk of acquiring iron deficiency. Moreover, these patterns exhibit a marked concordance with the hypothetical model of unremodeled porotic hyperostosis lesion frequencies that would be expected as a result of iron deficiency alone (see Figure 3a). The ages at onset, peak incidence, and remission for remodeled lesion frequencies in the Libben and Bt-5 macrosample also conform to the pattern that would be expected to occur in association with iron deficiency anemia. Figure 7b clearly shows that remodeled lesion frequencies display a lag phase which translates to the right of each high risk age/sex group. These patterns are concordant with those established in the hypothetical model illustrated in Figure 3c. Macrosample data given in Table 13 also show that young adult females 15 to 30 years of age have a frequency of unremodeled porotic hyperostosis which is 13 percent higher than adult females over 30 years of age at death. A chi square comparison indicates that this difference is statistically significant (%2=7.22; p<. 01). Thus, the age-related frequency distribution for unremodeled porotic hyperostosis in adult females identifies those individuals at greatest risk of acquiring iron deficiency anemia during the peak years of reproductive activity. Results that have been presented thus far can be summarized as follows: 1) The vast majority of porotic hyperostotic lesions that were observed in the Libben and Bt-5 groups consisted of slight bony involvements that were primarily restricted to the superior orbital plates. Extra-orbital lesions, and lesions of moderate expression, were most commonly seen in subadults that were six months to three years of age at time of death in each group. 2) Patterns in the ages at onset, peak incidence, and remission for unremodeled porotic hyperostosis were similar in the two groups such that; a) the highest frequency of unremodeled lesions occurred in subadults, b) adult females showed the second highest frequency of unremodeled lesions at time of death, and 1991 Paleoepidemiology of Porotic Hyperostosis 29 c) adult males showed the lowest incidence of unremodeled lesions. 3) Within group comparisons indicated that; a) subadults in each group experienced a significantly greater frequency of unremodeled lesions compared to adult males, and Libben subadults also displayed a significantly greater frequency of unremodeled lesions compared to adult females, b) adult females in each group exhibited unremodeled lesion frequencies that were sig- nificantly greater compared to adult males, and c) young adult females that were 15 to 30 years of age showed a greater frequency of unremodeled lesions compared to females over 30 years of age in each group. However, this difference was only significant for the Libben adult female comparison. 4) Unremodeled and remodeled porotic hyperostosis frequency data for the Libben and Bt-5 macrosample indicate a marked concordance with the hypothetical iron deficiency anemia model. Thus, age and sex specific patterns in the frequency of occurrence for porotic hyperostosis clearly identify those age/sex groups at greatest risk of acquiring iron deficiency anemia in human groups. 5) Between group comparisons showed that: a) unremodeled lesion frequencies that were observed in Libben and Bt-5 adult males, and adult females, were similar in magnitude, and b) the only significant difference in the frequency of porotic hyperostosis that characterized the two groups was confined to subadults. Here, Libben children displayed a significantly greater frequency of unremodeled and total lesions compared to Bt-5 subadults. Discussion Level of Non-Specificity and Differential Diagnosis for Porotic Hyperostosis With regard to level of non-specificity, it must be emphasized that disease prevalence differentials enhance our ability to infer the probable cause of porotic hyperostosis with a high degree of confidence on a populational basis. For example, due to their rare occurrence in small anthropological populations, many of the conditions which promote erythroid marrow hyperplasia (see Table 1) could have played little or no singular, or collective, role in the etiology of the lesion (eg., hereditary spherocytosis, congenital nonspherocytic anemias, etc). Iron deficiency anemia and the congenital hemolytic anemias are the only disorders that are sufficiently prevalent to account for the frequency with which porotic hyperostosis occurred in earlier human groups. In addition, the congenital hemolytic anemias can only be invoked to explain frequency of the lesion in certain geographically restricted populations in the Old World. Regardless, the view posited here is that porotic hyperostosis, as a bioassay of population fitness, is characterized by a very low level of non- specificity relative to most other non-specific indicators of disease and nutritional stress. The various pathological skeletal changes that occur in response to the anemias, and which aid in the differential diagnosis thereof, have been thoroughly investigated by Moseley (1965, 1966, 1974). His observations, combined with the age and sex related patterns reported here, served as the basis for evaluating the pathogenesis of Libben and Bt-5 porotic hyperostosis. The skeletal remains of all Libben and Bt-5 individuals were therefore surveyed for pathological changes known to occur in iron deficiency anemia and the congenital hemolytic anemias (i.e., sickle cell anemia and thalassemia) (Moseley, 1975). For reasons discussed above, other conditions were excluded from consideration. Skeletal changes in the skull that are produced by erythroid marrow hyperplasia were described earlier. Those that are observed most frequently affecting the long bones include (1) widening of the marrow spaces, (2) cortical thinning, and (3) coarsening of trabecular patterns (Angel, 1967; Moseley, 1974). However, these skeletal changes occur in response to the hemolytic anemias, iron deficiency anemia, and are also known to accompany a wide variety of other disease and nutritional disorders (Jaffe, 1972; Aegerter and Kirkpatrick, 1975; Greenfield, 1975). These pathological changes are therefore highly non-specific and of little diagnostic utility. Here, it should be emphasized that skeletal changes other than those produced by marrow hyperplasia are more important for diagnosis of the hemolytic anemias. For example, vertebral step deformity and hand-foot syndrome are the most reliable skeletal criteria for the identification of sickle cell anemia (Moseley, 1974). Also, children with sickle cell anemia often exhibit bone infarctions, osteomyelitis, and periostitis (Jaffe, 1972). However, it is well known that infants and young children are at greater risk of acquiring bone infections compared to older individuals (Robbins, 1974). Furthermore, periosteal reactions can be induced by a wide variety of disease agents (Greenfield, 1975). Therefore, the latter types of skeletal lesions cannot be regarded as diagnostic for the hemolytic anemias. In other circumstances, the extent of bone marrow proliferation is a more important indicator as to the nature of the underlying anemic stimulus. Studies concerned with the biodynamics of marrow response to anemia have shown that bone marrow proliferation is dependent on the amount 30 Mensforth No. 46 of iron available to the tissue (Hillman and Henderson, 1969; Hillman, 1970). For example, the level of erythroid marrow hyperplasia that occurs in iron deficiency anemia is usually self-limiting and generally approximates only 2 to 3 times the normal rate (Finch, 1970). In severe chronic iron deficiency anemia bone marrow proliferation may on occasion reach levels that are 4 to 6 times normal (Giblett et al. , 1950; Hillman and Henderson, 1969). In contrast, individuals suffering from hemolytic anemias frequently exhibit a marrow response that is 5 to 10 times above normal which then results in excessive hypertrophic bony changes in the skull (Finch, 1970; Moseley, 1974). Moseley (1974) has remarked that the skeletal changes due to marrow hyperplasia which occur in thalassemic homozygotes are more pronounced than those found in any other condition. Facial bones are often involved to the extent that malocclusion and a rodent fades deformity become manifest. The most reliable post-cranial skeletal change that has been observed in thalassemic homozygotes is bulbous expansion of the ribs. This primarily affects the posterior portions of the ribs and represents an exaggerated response to subperiosteal bone marrow proliferation (Moseley, 1974). Thus, the excessive levels of marrow proliferation which occur in the hemolytic anemias are determined, in part, by the increased amount of bioavailable iron that is retrieved from lysed red blood cells (Smith, 1972). None of the skeletal changes that are considered pathognomonic for the hemolytic anemias were observed in any Libben or Bt-5 individuals. The overwhelming majority of porotic hyperostotic lesions that were seen in the two groups involved only slight degrees of osseous tissue hypertrophy. These findings suggest that most Libben and Bt-5 individuals with porotic hyperostosis experienced a limited level of marrow hyperplasia which is similar to that reported in acute iron deficiency anemia. This is in contrast to the pronounced skeletal changes that would be expected to occur in greater frequency as a result of severe chronic iron deficiency anemia or the hemolytic anemias. It is concluded here that (1) the marked age and sex specific frequency distributions for the lesion, (2) the absence of skeletal changes that are considered pathognomonic for the hemolytic anemias, and (3) the low levels of osseous tissue response that were observed, support the inference that the majority of porotic hyperostosis in the Libben and Bt-5 skeletal groups was the result of acute iron deficiency anemia. Age and Sex Related Demographic Sensitivity of Porotic Hyperostosis Mortality data reported earlier show that no significant differences characterize between group comparisons of Libben and Bt-5 adult survivorship. Similarly, no significant differences were observed for between group comparisons of Libben and Bt-5 porotic hyperostosis frequency in adults. These results held true for both sex-combined and sex- specific intergroup comparisons of mortality and lesion frequency data. Therefore, the demographic and paleoepidemiological patterns that were observed for adults in the two skeletal groups display a marked concordance. Even though adult survivorship distributions were not substantially different for comparisons within or between each group, the average Libben:Bt-5 adult q.v ratio of 1.24 shows a slight trend favoring higher age-specific mortality rates in Libben adults.5 Similarly, the Libben:Bt-5 adult sex- combined porotic hyperostosis ratios for the frequency of unremodeled and total lesions are 1.09 and 1.35, respectively. Therefore, the minor between group differences in the frequency of porotic hyperostosis correspond in direction, though not necessarily in magnitude, to the minor difference in mortality rates that characterize Libben and Bt-5 adults. Within group comparisons of adult male versus female porotic hyperostosis showed that Libben and Bt-5 females both had a significantly greater frequency of unremodeled lesions at time of death. The Libben adult femaleimale porotic hyperostosis frequency ratios for unremodeled lesions are 6.50 and 4.21, respectively. Comparable sex ratios for total lesions at Libben and Bt-5 are 2.19 and 1.97, respectively. The sex-related lesion frequency differences described above would prompt some skeletal biologists to suggest that elevated pregnancy and lactation stress would result in higher levels of morbidity and mortality in adult females. However, it is unwise to base such conclusions on skeletal lesion frequency data alone. It was previously stated that no significant differences characterize patterns of adult male and female survivorship in each group. Indeed, Libben and Bt-5 adult females displayed a trend where age-specific survivorship either equaled, or exceeded, that of males in each series. Given that factors involved in the risks of acquiring iron deficiency anemia differ markedly for adult males and females, the finding that Libben and Bt-5 females displayed a significantly greater frequency of unremodeled lesions is not unexpected. However, it is clear that these sex- related differences in the frequency of porotic hyperostosis are demographically insensitive. Therefore, while sex- related differences in the frequency of the skeletal lesion may be of value for purposes of differential diagnosis, they do not appear to provide useful information about Libben and Bt-5 adult sex-specific mortality experience. Information reported earlier shows that the only significant differences between Libben and Bt-5 porotic hyperostosis frequencies and demographic parameters are confined to subadults. That is, (1) Libben children displayed a significantly greater frequency of unremodeled and total porotic hyperostotic lesions at time of death, and (2) significantly fewer Libben individuals survived to age 1991 Paleoepidemiology of P orotic Hyperostosis 31 15 compared to Bt-5 subadults. These results indicate the frequency of porotic hyperostosis in subadults and sur- vivorship at age 15 are inversely related in the two groups. With regard to demographic sensitivity, the additional findings are of interest. The Libben:Bt-5 subadult porotic hyperostosis ratios for the frequency of unremodeled and total lesions are 2.18 and 1.88, respectively. The average Libben:Bt-5 mortality ratio for subadults 0 to 15 years of age is 1 .60. Differences in the frequency of Libben and Bt- 5 subadult porotic hyperostosis therefore correspond in direction and magnitude to differences in subadult mortality that was observed for the two groups. Indeed, ratios that compare Libben and Bt-5 subadult lesion frequencies exceed the subadult mortality ratio and exhibit demographic hypersensitivity for individuals in this age category. For inferential purposes, the relationships described above were formalized so that subadult porotic hyperostosis lesion frequencies could be used to predict survivorship at age 15 ( 1 5 ) in earlier human groups. The algorithms are presented as follows: la. y = y = x\ = lb. y v = *2 = .05337(.Y!) + 69.072 1 1 5 frequency of unremodeled porotic hyperostosis that is observed in subadults (0-15 years) -0.4307(x2) + 71.506 •15 frequency of total porotic hyperostotic lesions that are observed in subadults (0-15 years). Given that only two skeletal samples were used to generate the equations listed above, it is clear that the relationships posited must be regarded as hypotheses. The extent to which the models accurately reflect the relationships between porotic hyperostosis and survivorship at age 15 in earlier human groups awaits more rigorous tests that employ a substantially greater number, and temporospatial diversity, of skeletal groups. Future analyses are required to determine (a) the precision with which lesion frequencies can accurately estimate demographic parameters of earlier human groups, and (b) whether a linear or non-linear model provides the best measure of a relationship as such. Alternatively, it may be found that porotic hyperostosis lesion frequencies provide no useful information for demographic inference whatsoever. These are matters that must be addressed in future research. Developmental Factors Which Influence Age-Related Differences in the Frequency of Occurrence and Morphological Expression of Porotic Hyperostosis Results presented here, and elsewhere, document that children under ten years of age generally display greater frequencies and degrees of severity for porotic hyperostosis compared to adults in earlier human groups (Moseley, 1965; El-Najjar et al ., 1976; Lallo et al. , 1977). The intrinsic physiological and extrinsic environmental factors which promote risk of acquiring iron deficiency anemia in subadults, particularly those that are six months to three years of age, have been discussed. There are, in addition, developmental factors that differentially influence the extent to which subadults and adults are likely to manifest skeletal changes in response to erythroid marrow hyperplasia. For example, the liver and spleen play a dominant role in red blood cell formation during much of the fetal period (Smith, 1972). However, shortly prior to birth these tissues quiesce. Thereafter, bone marrow cavities become the principal sites of hematopoiesis throughout life (Sodeman and Sodeman, 1974). In the infant virtually all bone marrow cavities are actively involved in red blood cell formation. This results in a circumstance where, until four years of age, a delicate balance exists between the limited bone marrow space that is available for hematopoiesis, and the bone marrow space that is required to meet the rapidly growing child’s increased red blood cell requirements (Smith, 1972; Sodeman and Sodeman, 1974). By four years of age the skeleton has achieved sufficient volumetric growth to provide bone marrow space in excess of hematopoietic needs. Then, during the second decade of life the bone marrow cavities in the appendicular skeleton undergo a slow transition where active red marrow is replaced by inactive yellow marrow. By twenty years of age active red marrow in the appendicular skeleton is confined to the proximal regions of the long bones. In contrast, bones of the axial skeleton (i.e., skull, vertebrae, sternum, and ribs) remain active in red blood cell formation throughout life (Robbins, 1974; Hardesty and Weatherall, 1982). Therefore, skeletal changes that are due to erythroid marrow hyperplasia are most likely to occur in anemic children under four years of age. This is a fundamental consequence of (a) the competitive relationship between limited supply and increased demands for active marrow space, combined with (b) the numerous constitutional and environmental factors that promote the risk of developing iron deficiency anemia at this age. Infants and young children can thus be regarded as hypersensitive to skeletal changes in the anemias. For adults and subadults alike, it is reasonable to presume that the first tissues to respond to an anemic stimulus will be those that are already active in red blood cell formation (i.e., bones of the axial skeleton) (Smith, 1972; Sodeman and Sodeman, 1974). In contrast to young children, however, the adult appendicular skeleton provides a substantial reserve of yellow marrow that is capable of reverting to active red marrow in response to a hypoxic stimulus (Sodeman and Sodeman, 1974). Therefore, the 32 Mensforth No. 46 extent to which skeletal changes will occur in the adult cranium as a result of erythroid marrow hyperplasia depends on ( 1 ) the duration and severity of the underlying anemic condition, and (2) the rate at which appendicular bone marrow sites are recruited for compensatory red blood cell formation. Given that substantial hematopoietic reserves are readily available to the adult, it is reasonable to suggest that skeletal changes due to erythroid marrow hyperplasia will be uncommon, or extremely limited, in these individuals. In accord with this relationship are results reported by Stuart-Macadam (1985) and results obtained herein (i.e., that Libben and Bt-5 adults displayed low frequencies of unremodeled porotic hyperostosis overall, and that such lesions were minimal bony responses to marrow proliferation). Thus, adults are best regarded as hyposensitive to skeletal changes in the anemias. It is well recognized that porotic hyperostosis of the superior orbital plates (i.e., cribra orbitalia) is by far the most common expression of the disorder. In this study orbital lesions were observed in 97.2 percent (247/254) of all Libben and Bt-5 individuals that had either remodeled or unremodeled porotic hyperostosis. Among the small sample of individuals that had one or more extra-orbital lesions in the two groups 80.6 percent (29/36) also displayed orbital lesions at time of death. The specific reason as to why the superior orbital plates are so responsive to erythroid marrow hyperplasia has eluded functional explanation. Nonetheless, this skeletal site exhibits a marked sensitivity to bone marrow proliferation and must be considered as a superior index of the skeletal disorder. Furthermore, porotic hyperostosis affecting the superior orbital plates cannot be detected by standard clinical x-ray apparatus unless the skeletal changes are as pronounced as those shown in Figure I . Therefore, workers who (1) restrict their definition of porotic hyperostosis to include only those skeletal lesions that are seen affecting the frontal, parietal, or occipital bones proper, and (2) attempt to identify porotic hyperostosis in the living by radiographic techniques, will grossly underestimate the frequency with which the pathological bony response occurred. The way in which erythroid marrow hyperplasia gives rise to the morbid appearance of porotic hyperostosis as seen in the skeleton is relatively easy to comprehend. Laboratory studies have shown that tissue hypoxia initially stimulates a small number of hematopoietic stem cells to enter into an intense phase of proliferation (Till and McCulloch, 1961; Trentin, 1971). This results in a circumstance where clonal colonies of undifferentiated stem cells quickly become established in the bone marrow (Sodeman and Sodeman, 1974). Then, after a few days, these differentiate into discrete colonies or nests of hematopoietically active bone marrow cells. Hyperplastic activity of discrete bone marrow cell colonies is the best explanation for the manner in which porotic hyperostotic lesions acquired their porous appearance (see Figure lb). The degree to which osseous tissue hypertrophy accompanied such lesions was probably dependent on ( 1 ) the level of hyperplasia that was attained by discrete bone marrow cell colonies, and (2) the extent to which additional marrow space was required to accom- modate cells that were actively involved in hematopoiesis. Factors Involved in the Etiology of Iron Deficiency Anemia at Libben and Bt-5 Malabsorption syndrome as a consequence of intestinal cestodiasis is a factor which may have contributed to iron deficiency anemia in the Libben and Bt-5 groups. Parasitic tapeworm infestation is usually contracted by ingestion of raw, or under-cooked, meat or fish, and occurs in geographically widespread human populations (Robbins, 1974). Libben and Bt-5 faunal analyses described earlier indicate that both groups relied heavily on seasonally available freshwater marine resources. This dependence was particularly marked for the Libben group where an abundance of freshwater fish were exploited. Thus, it is conceivable that fish tapeworm infestation (i.e., Diphyllobothrium latum), and subsequent intestinal malabsorption of dietary nutrients, affected some individuals in both skeletal populations. However, epidemiological surveys in contemporary aboriginal societies have shown that microbial respiratory and gastrointestinal tract infections play a much more substantial role in precipitating, and exacerbating, nutritional crises, particularly in infants and young children (Maynard and Hammes, 1970; Jose and Welch, 1970). Although parasite loads reach higher levels in anemic and malnourished children, they generally play no major role in initiating nutritional deficiency syndromes during the early years of life (Jose and Welch, 1970). Therefore, it seems unlikely that intestinal cestodiasis was important in the etiology of iron deficiency anemia and nutritional stress in Libben and Bt-5 individuals. Constitutional factors such as prematurity and low birthweight also probably played no substantial role in the pathogenesis of Libben and Bt-5 porotic hyperostosis in subadults for the following reasons. First, the iron deficiency anemia of prematurity would be expected to produce skeletal changes that could be seen in individuals that were three to six months of age at time of death. However, no lesions were observed in Libben or Bt-5 infants prior to six months of age. Second, premature infants in contemporary under-developed societies rarely survive the neonatal period due to the numerous physiological handicaps that these individuals experience 1991 Paleoepidemiology of P orotic Hyperostosis 33 at birth (Levine and Gordon, 1942; Maynard and Hammes, 1970; Korones, 1976). Given that premature infants only constitute 6-12 percent of live births in human groups (Maynard and Hammes, 1970 and references therein), combined with their high risk of mortality, it is highly improbable that a sufficient number of premature infants would have survived to account for the frequencies with which porotic hyperostosis occurred in Libben and Bt-5 infants. Similarly, diet does not appear to be a major factor in the etiology of Libben and Bt-5 porotic hyperostosis. Floral and faunal analyses indicate that resources available to the two groups provided an adequate supply of bioavailable iron. Furthermore, these resources were low in chelating agents that are known to inhibit the absorption of dietary iron. It remains likely, however, that dietary factors were to some extent involved in promoting the onset of iron deficiency anemia that occurred during the weaning period in each group. Many human societies routinely incorporate high carbohydrate gruels in the weaning diet for reasons principally related to ease of preparation, mastication, and assimilation (Gordon et ai, 1963; Scrimshaw and Young, 1976; Farb and Armelagos, 1980). These are usually manufactured from wild or domesticated cereal grains or other indigenous vegetable resources. In addition, the introduction of solid food items in the diet exposes the weanling to novel pathogens. The latter play a substantial role in precipitating nutritional crises and growth retardation due to recurrent episodes of acute and chronic gastrointestinal tract infections (Gordon et ai, 1963). Still, there is little evidence to suggest that chronic iron deficiency anemia due to dietary inadequacy was an important epidemiological factor in the development of Libben and Bt-5 porotic hyperostosis. The dramatic decrease in the frequency of unremodeled porotic hyperostosis that was seen in subadults over three years of age, and low levels of osseous tissue response that were observed in each group, are concordant with this view. These findings are in contrast to those reported for several maize dependent New World skeletal groups where the incidence of porotic hyperostosis remained high throughout the childhood years, and lesions displayed more severe degrees of bony involvement (Moseley, 1965; El-Najjar et al., 1976; Lallo et al., 1977). In order to more clearly evaluate the extent to which diet may have played a role in the etiology of Libben and Bt-5 subadults, lesion frequencies for these individuals were compared to those that occurred in subadults from two maize dependent prehistoric groups. The latter represent subsamples of the skeletal materials that were employed by El-Najjar (1976) where the relationship between diet and the frequency of occurrence for porotic hyperostosis was investigated in several New World prehistoric groups. The Anasazi subadults (n= 54) were recovered from Canyon de Chelly in northeast Arizona and are affiliated with the Basketmaker II-III cultural period that extended from 400 to 700 A.D. (El-Najjar et al., 1976). 6 The Peruvian subadults (n = 70) represent a poorly documented skeletal series of prehistoric South American Indians that were recovered in the earlier part of the 20th century (El- Najjar, 1976). 7 For these individuals age at death was estimated by assessments of dental maturation and eruption status in accordance with published standards (Krogman and Iscan, 1986). Subadults that were represented in the Anasazi and Peruvian skeletal samples ranged in age from birth to ten years. Although both skeletal series have been classified as maize dependent, the Anasazi group were primarily horticulturalists that supplemented dietary intake by foraging (El-Najjar et al., 1976). In contrast, the Peruvian group practiced intensive agriculture where subsistence was highly dependent upon the use of domesticated cultigens (El-Najjar, 1976). For the Anasazi and Peruvian subadult samples combined, this worker observed a total frequency of porotic hyperostosis that was 75.8 percent (94/124). Similarly, frequency data reported by El-Najjar (1976) yielded a figure of 74.8 percent (80/107) for subadults in these two groups. Inter-observer error was approximately one percent. Therefore, the methods used by El-Najjar and associates (1976) to identify porotic hyperostosis were found to be highly replicable. Data enumerating the age-specific frequencies for unremodeled, remodeled, and extra-orbital porotic hyperostosis that was observed in Anasazi and Peruvian subadults (0-10 years) is given in Table 16. Summary data which compare the total frequency for porotic hyperostosis in Bt-5, Libben, Anasazi, and Peruvian subadults that ranged in age from birth to ten years are listed in Table 17. The age-specific frequency with which unremodeled porotic hyperostosis occurred in these groups is illustrated in Figure 8. Results show that the age at onset, peak incidence, and age at remission for unremodeled lesions are markedly similar in each group. The low-to-high rank order frequency for unremodeled, total, and extraorbital lesions that were observed is Bt-5, Libben, Anasazi, and Peruvian (Table 17). Chi square values that compare the total frequencies with which porotic hyperostosis occurred in each group are listed in Table 18. Results show that (1) the frequency of total lesions and extra-orbital lesions were significantly different in each group, and (2) the maize dependent subadults had a significantly higher frequency of total lesions and extra-orbital lesions compared to the non-maize dependent Libben and Bt-5 subadults. However, there is strong evidence to suggest that the two maize dependent subadult samples are demo-graphically 34 Mensforth No. 46 TABLE 16. Age-Specific Frequency of Occurrence for Porotic Hyperostosis in Prehistoric Anasazi and Coastal Peruvian Subadults (0-10 years) Anasazi Coastal Peruvian Age Group «/ Unremodeled Lesions Remodeled Lesions Total Lesions "i Unremodeled Lesions Remodeled Lesions Total Lesions n2 (%) (%) »4 (%) n2 (%) (%) >U (%) 0.0 - 0.5 1 0 0.0 0 0.0 0 0.0 2 0 0.0 0 0.0 0 0.0 0.5 - 1.0 4 2 50.0 0 0.0 2 50.0 4 2 50.0 0 0.0 2 50.0 1.0 - 3.0 18 13 72.2 1 5.6 14 77.8 24 19 79.2 4 16.7 23 95.8 3.0 - 5.0 13 5 38.5 2 15.4 7 53.8 13 4 30.8 5 38.5 9 69.2 5.0 -10.0 18 3 16.7 10 55.6 13 72.2 27 8 29.6 16 59.3 24 88.9 Total 54 23 42.6 13 24.1 36 66.7 70 33 47.1 25 35.7 58 82.9 Symbols: tip total number of specimens that could be examned for the presence/absence of porotic hyperostosis per age group; n2i number of specimens that were scored positive for unremodeled lesions per age group; n2 number of specimens that were scored positive for remodeled lesions per age group; n4 total number of specimens that were scored positive for remodeled or unremodeled lesions per age group. TABLE 17. Comparison of the Total Frequency With Which Porotic Hyperostosis Occurred Among Subadults (0-10 years) From Four Culturally and Temporospatially Diverse Skeletal Groups Unremodeled Remodeled Total Extra-orbital Lesions Lesions Lesions Lesions Skeletal Group 1l n2 (%) (%) n4 (%) »5 (%) Bt-5 88 12 13.6 7 8.0 19 21.6 3 3.4 Libben 250 77 30.8 30 12.0 107 42.8 31 12.4 Anasazi 54 23 42.6 13 24.1 36 66.7 18 33.3 Peruvian 70 33 47.1 25 35.7 58 82.9 46 65.7 Total 462 145 31.4 75 16.2 220 21.2 98 21.2 Symbols: n]t total number of specimens that could be examined for the presence/absence of porotic hyperostosis; n-,, number of specimens that were scored positive for unremodeled lesions; n3, number of specimens that were scored positive for remodeled lesions; n4, total number of specimens that were scored positive for remodeled and/or unremodeled lesions; n5 total number of specimens that exhibited extra-orbital lesions. unsound as a result of selective recovery practices. For example, even though the state of preservation is nothing less than superb for these two groups, each skeletal collection primarily consists of isolated crania alone. That is, post-cranial materials were saved for only 22.2 percent (12/54) of Anasazi subadults and 14.3 percent (10/70) of Peruvian subadults. In addition, reference to Table 16 shows that Anasazi and Peruvian infants (0-1 year) are grossly under-represented in each group. When compared to Libben and Bt-5, where infants with a low frequency of porotic hyperostosis are present in substantial numbers, this will have the net effect of inflating lesion frequencies in the Anasazi and Peruvian samples when all individuals from birth to ten years of age are compared. Therefore, the comparisons presented earlier are biased and require adjustment before valid conclusions can be reached. Because infants are markedly under-represented in the maize dependent samples, these individuals were excluded from further consideration. Adjusted summary data given in Table 19 compares the total frequency with which porotic hyperostosis was observed in Bt-5, Libben, Anasazi, and Peruvian subadults that ranged in age from one to ten years. Chi square values that compare these adjusted lesion frequencies are listed in Table 20. Most between group differences that were reported earlier are still apparent for the adjusted data. However, unremodeled and total lesion frequencies in Libben and Anasazi subadults are now quite similar. Thus, the extent to which dietary differences may have resulted in a greater frequency of porotic hyperostosis in Anasazi versus Libben subadults (i.e., the pattern that was observed for comparisons of unadjusted lesion frequencies) is now obscure. Only extraorbital lesion frequencies remain significantly different for all four subadult samples that were examined. 1991 Paleoepidemiology of P orotic Hyperostosis 35 FIGURE 8. Age-specific frequencies of unremodeled porotic hyperostosis that were observed among subadults ( 0-10 years ) from four culturally diverse skeletal sample. Alternatively, it is suggested that a fundamental difference in levels of infectious disease load may have been more important in generating the subadult porotic hyperostosis frequency distributions that were observed among Libben subadults. That is, in addition to the demographic patterns reported here for the Libben and Bt-5, comparative analyses of tibia long bone growth perform- ance, and frequency data for periosteal reactions, that were observed in subadults from these two groups, support the inference that Libben infants and children experienced higher levels of morbidity and mortality in response to elevated levels of infectious disease. Results published elsewhere showed that Libben children six months to two years of age experienced a major period of slowed growth relative to Bt-5 (Mensforth, 1985). Growth suppression continued until four years of age when Libben children accrued a maximum 9.94 percent decrement in tibia growth relative to Bt-5. Furthermore, Libben subadults, displayed a frequencies of periosteal reactions that were significantly greater compared to Bt-5 children (%2=7.29; p<.01). In particular, for those ranging in age from birth to three years in particular, 51% of Libben and 27% of Bt-5 infants and young children displayed periosteal reactions (%2=12.24; p<.001). Among individuals affected in this age group, 79% of Libben and 65% of Bt-5 young children displayed unremodeled, or active, lesions at time of death. These skeletal lesions were interpreted to represent common microbial infections that had become established in bony tissues as a result of hematogenous seeding (Mensforth, 1986, 1985; Mensforth et al., 1978). TABLE 18. Chi Square Values That Compare the Frequency of Occurrence for Porotic Hyperostosis in Subadults (0-10 years) From Four Culturally and Temporospatially Diverse Skeletal Groups Skeletal Group Skeletal Group Lesion Status Libben Anasazi Peruvian Bt-5 Unremodeled Lesions 9Mb 15.1 1° 21.49" Extra-orbital Lesions 5.82c 23.78" 70.74" Total Lesions 15.52" 28.65" 58.57" Libben Unremodeled Lesions 2.80 6.47c Extra-orbital Lesions 14.39" 85.07" Total Lesions 10.15/j 35.13" Anasazi Unremodeled Lesions 0.25 Extra-orbital Lesions 12.80" Total Lesions 4.36" a: significant at the .001 level of probability b: significant at the .01 level of probability. c: significant at the .05 level of probability. 36 Mensforth No. 46 TABLE 19. Comparison of the Total Frequency With Which Porotic Hyperostosis Occurred Among Subadults (1-10 years) From Four Culturally and Temporospatially Diverse Skeletal Groups Skeletal Group nl Unremodeled Lesions Remodeled Lesions Total Lesions Extra-orbital Lesions n2 (%) n3 (%) n4 (%) n5 (%) Bt-5 50 9 18.0 7 14.0 16 32.0 1 2.0 Libben 151 67 44.4 30 19.9 97 64.2 24 15.9 Anasazi 49 21 42.9 13 26.5 34 69.4 16 32.7 Peruvian 64 31 48.4 25 39.1 56 87.5 44 68.8 Total 314 128 40.8 75 23.9 203 64.6 85 27.1 Symbols: nt, total number of specimens that could be examined for the presence/absence of porotic hyperostosis; n2, number of specimens that were scored positive for unremodeled lesions; n ,, number of specimens that were scored positive for remodeled lesions; n4, total number of specimens that were scored positive for remodeled and/or unremodeled lesions; n5 total number of specimens that exhibited extra-orbital lesions. TABLE 20. Chi Square Values That Compare the Frequency of Occurrence for Porotic Hyperostosis in Subadults ( 1-10 years) From Four Culturally and Temporospatially Diverse Skeletal Groups Skeletal Group Skeletal Group Lesion Status Libben Anasazi Peruvian Bt-5 Unremodeled Lesions 11.11° 1.24b 1 1 .42° Extra-orbital Lesions 6.66b 16.35° 52.35° Total Lesions 15.86° 13.84° 37.16° Libben Unremodeled Lesions 0.03 0.30 Extra-orbital Lesions 6.49c 58.07° Total Lesions 0.43 11.85° Anasazi Unremodeled Lesions 0.35 Extra-orbital Lesions 5.62c Total Lesions 14.52° a: significant at the .001 level of probability b: significant at the .01 level of probability. c: significant at the .05 level of probability. The extent to which habitation of the Black Swamp elevated risk of acquiring infectious diseases among the Libben people remains enigmatic. Prior to being drained, in recent historic times the relatively unchanged Black Swamp was dreaded by pioneers. However, the Black Swamp acquired its reputation because it was a formidable barrier to travel and settlement (Kaatz, 1955). Thus, at least in the recent historic past, the Black Swamp has no reputation for being a source of profound human misery related to disease. Nonetheless, a more specific scenario could be posited where we might expect aboriginal inhabitants of the rich Black Swamp to have experienced high and prolonged seasonal exposures to mosquito, and other, insect bites. Among infants and children, these circumstance alone could result in cumulative blood losses that could precipitate, or exacerbate, a latent iron deficient state. These, in turn, could easily elevate risk of microbial invasion. The latter would occur either by direct introduction of pathogens at the time when a blood meal is being secured, or indirectly as a consequence of skin irritation and inflammation. In these 1991 Paleoepidemiology of P orotic Hyperostosis 37 ways, pathogens which normally colonize the skin and maintain a symbiotic relationship, such as Staphylococcus aureus, experience greater opportunities for capillary invasion and hematogenous seeding of the host. Alternatively, food preparation techniques, or the selective lack thereof, may have promoted a greater frequency, intensity, and duration of gastrointestinal tract infections in Libben infants and children. While neither of the hypotheses presented above, either alone or in combination, can be accepted or refuted, skeletal evidence presently available favors the former relationship. That is, the majority of periosteal reactions observed in Libben subadults (1) occurred in individuals under three years of age (51% affected), (2) the highest frequency characterized infants that died in the first year of life (57% affected), (3) patterns of skeletal involvement clearly indicate that the majority of affected individuals suffered from a systemic blood-borne infectious disease, and (4) the infectious disease episodes suffered by affected individuals persisted for periods of time sufficient to elicit a frank bony response to the underlying stress episode (Mensforth et al, 1978; Mensforth, 1986). Although specific agents involved in the etiology of Libben and Bt-5 porotic hyperostosis and periosteal reactions remain speculative, the strong synergistic relationships between iron deficiency anemia and infectious disease discussed earlier cannot be ignored (i.e., impaired immune response, hypoferremia, and malabsorption syndrome). Comparative population studies have repeatedly demonstrated that infections, primarily those of the respiratory and gastrointestinal tracts, have a negative influence on hemoglobin status and play a fundamental role in precipitating nutritional crises such as iron deficiency anemia (Maynard and Hammes, 1970; Jose and Welch, 1970; Kaplan et al, 1973; Burks et al., 1976). These effects are most pronounced in children under three years of age where infectious disease and nutritional deficiency commonly results in the pneumonia-weanling diarrhea syndrome (Gordon et al, 1963). The synergistic consequences of anemia and infection in contemporary underdeveloped societies are elevated morbidity, mortality, and early onset of growth retardation (Maynard and Hammes, 1970; Jose and Welch, 1970; Burks et al., 1976). The extent to which increased population densities and degrees of sedentism may have resulted in a greater prevalence of microbial and parasitic infectious diseases, nutritional deficiencies, and growth retardation in the Anasazi and Peruvian skeletal groups cannot be assessed. This is due to the fact that these samples lack suitable demographic documentation and are further characterized by infant under-representation and poor recovery of post- cranial skeletal materials. However, Kent (1987) has recently reported a case study where it is cogently argued that higher population densities, greater sedentism, and elevated levels of infectious disease may have played a more substantial role than maize diet in the etiology of porotic hyperostosis in prehistoric Anasazi populations. For purposes of demographic inference, algorithms la and lb were modified so that Libben and Bt-5 porotic hyperostosis frequencies adjusted for the one to ten year subadult age period could be used to predict survivorship at age 15 (115) for the Anasazi and Peruvian groups. These equations are as follows: 2a. y = y - X\ - 2b. y = y = X2 ~ -0.3295(xi) + 67.532 ll5 unremodeled porotic hyperostosis frequency that is observed in subadults (1-10 years) 0.2702(a2) + 70.246 •l5 total porotic hyperostosis frequency that is observed in subadults (1-10 years) These algorithms predict average values for survivorship at age 15 in the Anasazi and Peruvian groups which should approximate 52.5 + 1.3 and 49.1 + 3.5, respectively. The degree to which porotic hyperostosis lesion frequencies accurately reflect these demographic projections will await further investigations of these two groups that employ sound demographic data. Finally, reference to Tables 19 and 20 show that extraorbital porotic hyperostotic lesion frequencies were significantly different for comparisons of all four subadult skeletal samples. Indeed, it was found that extra-orbital lesions accumulated at an exponential rate relative to increments in total lesion frequencies that were observed in band level foragers and maize dependent subadults. This relationship can be expressed as follows: y-e ax - 1 y = extra-orbital lesion frequency predicted .V = total porotic hyperostosis frequency observed e = 2.7182818 a = .04721 The low-to-high rank order frequencies for extra-orbital lesions are Bt-5 (2.0%), Libben (15.9%), Anasazi (32.7%), and Peruvian (68.8%). These figures suggest that maize utilization by the Anasazi and Peruvian groups, combined with higher levels of population density and degree of sedentism, resulted in a greater duration and severity of iron deficiency anemia compared to the Libben and Bt-5 skeletal populations. Moreover, it is most probable that endemic hookworm disease played a substantial role in the etiology of chronic iron deficiency anemia seen in Peruvian infants and children. 38 Mensforth No. 46 With regard to porotic hyperostosis, the view posited here is that extra-orbital lesion frequencies may provide a superior skeletal index of the extent to which chronic iron deficiency anemia was prevalent in earlier human groups. As such, a useful index as such can be calculated by dividing the frequency of extra-orbital lesions observed by the frequency of total lesions observed. For the subadult skeletal samples described above these ratios are as follows: Bt-5, 0.06; Libben, 0.25; Anasazi, 0.47; Peruvian, 0.79. The benefit of this measure is its ability to detect the presence of chronic or protracted episodes of erythroid marrow hyperplasia regardless of whether total lesion frequencies are low, moderate, or high in skeletal samples under investigation. Summary and Conclusions The age and sex related frequencies with which porotic hyperostosis occurred in the Libben Late Woodland (»=580) and Bt-5 Late Archaic (n= 247) skeletal populations were examined. The goals of the study were to (1) identify those factors which were important in the etiology of the skeletal lesions that were observed in each group, and (2) evaluate the extent to which porotic hyperostosis serves as a useful bioassay of disease and nutritional stress in prehistoric human populations. It was suggested that stress indicators of greatest paleoepidemiological utility will be those that display high levels of biological and demographic sensitivity combined with a low level of nonspecificity. Qualitative assessments of lesion activity status (i.e., remodeled or unremodeled) and patterns of skeletal involvement (orbital versus extra-orbital; cranial versus post-cranial, and slight, moderate, and severe degrees of bony response) were recorded for all Libben and Bt-5 individuals that displayed porotic hyperostosis at time of death. Unremodeled lesions were used to establish the age- specific frequency distributions for the lesions and identify those age/sex groups at greatest risk of acquiring the disorder. Patterns of skeletal involvement aided in differential diagnoses and served as a skeletal index of the degree to which an underlying hypoxic stimulus resulted in a greater duration and severity of the stress episode. Results showed that the majority of Libben and Bt-5 porotic hyperostosis consisted of slight bony involvements that were restricted to the superior orbital plates of affected individuals. More pronounced degrees of osseous tissue response were somewhat more common in individuals that were six months to three years of age at time of death. No individuals in either group displayed skeletal changes that are considered diagnostic for hemolytic anemias such as sickle cell anemia or thalassemia. Within group comparisons for broad age/sex categories showed that (a) subadults had a significantly greater frequency of unremodeled lesions compared to adults, (b) adult females had a significantly higher frequency of unremodeled lesions compared to adult males, and (c) young adult females had a greater incidence of unremodeled lesions compared to females over 30 years of age. However, the latter difference was only significant for the Libben group. Patterns in the ages at onset, peak incidence, and remission for unremodeled and remodeled porotic hyperostosis were similar in the two groups. These data differed in no substantial way from patterns predicted by the hypothetical iron deficiency anemia related porotic hyperostosis model that was presented. Indeed, Libben and Bt-5 age and sex specific frequency distributions for unremodeled porotic hyperostosis clearly identified those age/sex groups at greatest risk of acquiring iron deficiency anemia as a result of intrinsic physiological and extrinsic environmental risk factors. Sex-specific between group comparisons showed that Libben and Bt-5 adult males, and adult females, displayed similar frequencies of porotic hyperostosis. The only substantial difference that was observed between the two groups was confined to subadults where Libben children displayed a significantly greater frequency of unremodeled and total lesions compared to Bt-5 children. Compared to most other conditions that promote erythroid marrow hyperplasia, only iron deficiency anemia and/or the congenital hemolytic anemias (i.e., sickle cell anemia or thalassemia) are sufficiently prevalent to account for the frequencies with which porotic hyperostosis has been observed in earlier human groups. Therefore, it was concluded that porotic hyperostosis is characterized by a very low level of non-specificity relative to most other skeletal indicators of disease and nutritional stress. Comparisons of Libben and Bt-5 average mortality ratios and porotic hyperostosis frequency ratios for subadults, and adults (sex-combined), were directionally concordant. Thus, porotic hyperostosis exhibits demographic sensitivity as a bioassay of health conditions for the two groups. The only exception to this was that sex ratios for adult lesions frequencies did not conform to the patterns of adult male and female survivorship that characterized each group. That is, mortality rates were approximately equal for the sexes in each group whereas adult female: male lesion frequency ratios were skewed. Therefore, it was concluded that porotic hyperostosis is demographically insensitive as an indicator of differential mortality experience in adult males and females. Infants and young children from six months to four years of age were found to be hypersensitive to skeletal changes in the anemias. This appears to be a primary consequence of the age-specific competitive relationship between limited supply and increased demands for active marrow space, combined with numerous constitutional and environmental factors that elevate risk of developing iron deficiency 1991 Paleoepidemiology oe P orotic Hyperostosis 39 anemia in very young individuals. In contrast, adults are regarded as hyposensitive to skeletal changes in the anemias due to the fact that substantial hematopoietic bone marrow reserves are available to these individuals. Thus, porotic hyperostosis exhibits a biological sensitivity that bears a strong relationship with developmental age. The epidemiological patterns that were observed for Libben and Bt-5 porotic hyperostosis support the interpretation that iron deficiency anemia was the primary, if not sole, cause of the skeletal lesion in these two groups. With regard to etiology, however, the following circumstances must be emphasized. Floral and faunal analyses indicate that both band level societies exploited a diverse abundance of local dietary resources. Moreover, no subadults or adults in either group exhibit skeletal evidence to suggest that chronic malnutrition owing to dietary inadequacy was a major factor influencing health status. Therefore, diet probably played a very minor role in the etiology of porotic hyperostosis at Libben and Bt-5. Similarly, evidence from contemporary epidemiological surveys suggests that parasitism played no significant role in precipitating nutritional crises in these earlier human groups. Aside from cultural affiliations and temporospatial distributions, the principal difference between the Libben and Bt-5 skeletal samples concerns the fact that levels of subadult mortality were substantially greater. Flere, it is suggested that local environmmental circumstances associated with the habitation and exploitation of the Black Swamp may have played a fundamental role in elevating infectious disease loads which resulted in a greater prevalence of iron deficiency anemia in Libben children. The strong synergistic relationships between iron deficiency anemia and infectious disease are known to have their greatest impact on patterns of infant and child morbidity, mortality, and growth performance. Moreover, these relationships are concordant with the findings that Libben subadults experienced greater frequencies of porotic hyperostosis, periosteal reactions, and an early onset of long bone growth retardation compared to Bt-5 children. Given that subadults exhibit a marked biological and demographic sensitivity to erythroid marrow hyperplasia (1) algorithms were presented whereby subadult porotic hyperostosis frequencies can be used to predict survivorship at age 15 in earlier human groups, and (2) porotic hyperostoss frequencies for Libben and Bt-5 subadults were compared to those observed in two prehistoric maize dependent skeletal samples. When the latter comparisons were restricted to the one to ten year age interval, in order to adjust for infant under-representation in the Anasazi and Peruvian samples, it was found that Libben and Anasazi subadults had similar frequencies of unremodeled and total lesions. These findings suggest that similarities in degree of sedentism and population density may be responsible for the similarity in lesion frequencies that were observed. Alternatively, it is suggested that increased sedentism, population density, and high levels of endemic hookworm infestation contributed to the elevated frequency of porotic hyperostosis that characterizes the sample of Peruvian subadults. In contrast, the frequency with which extra-orbital porotic hyperostosis occurred was significantly different in all four subadult samples that were examined. It was found that extra-orbital lesions accumulate at an exponential rate relative to total lesions, and may therefore provide a superior skeletal index of the extent to which chronic iron deficiency anemia occurred in earlier human groups. The greater prevalence of extra-orbital lesions in the maize dependent subadult samples may reflect the extent to which dietary inadequacy promoted a greater duration and severity of iron deficiency anemia in these groups. Finally, it is interesting that few investigators concerned with the pathogenesis of porotic hyperostosis in Old World skeletal groups have seriously entertained the iron deficiency anemia hypothesis (for exceptions see Hengen, 1971 and Carlson et al., 1974). This is particularly true for researchers who have been concerned with the etiology of the lesion in circum-mediterranean skeletal groups (Angel, 1967; Ascenzi and Salistreri, 1977; and Germana and Ascenzi, 1980). Given the fact that iron deficiency anemia is prevalent throughout the world (Witts, 1966), particularly in agricultural communities (World Health Organization, 1968), it is clear that the iron deficiency and hemolytic anemia hypotheses cannot reasonably be considered as mutually exclusive explanations for the occurrence of porotic hyperostosis in Old World circum-mediterranean skeletal populations. It is highly improbable that those human groups subjected to the selection pressures of falciparum malaria would be uniquely exempt from basic human, and mammalian, pathophysiological responses to disease and nutritional stress. Likewise, the iron deficiency and hemolytic anemia hypotheses cannot reasonably be invoked to explain the occurrence of the skeletal lesion on a discrete continental basis as has been suggested (Angel, 1967). Thus, many studies that have investigated the pathogenesis of porotic hyperostosis in Old World prehistoric groups are clearly in need of substantial revision. Acknowledgements I would like to thank George Amielagos, Richard Meindl, and several anonymous reviewers who provided valuable comments and criticisms on earlier versions of this manuscript. I would also like to thank John Blank, members of the Cleveland State University Instructional Media Services, and members of the University of Oklahoma Electron Microscopy Laboratory who assisted in the preparation of photographs and illustrations used in this study. 40 Mensforth No. 46 Notes 1. The Libben site skeletal materials are permanently curated by the Department of Sociology and Anthropology at Kent State University, Kent, Ohio. 2. The Carlston Annis Bt-5 site skeletal materials are permanently curated by the Department of Anthropology at the University of Kentucky, Lexington. 3. The criterion that was used to identify lesions of minimal expression conform to those employed by El-Najjar et al. (1976) where one or more clusters of pores which extended 5 millimeters in diameter or greater were considered as indicating presence of the lesion. Lesions of moderate expression typically involved a greater surface area of the affected bone, displayed a well-developed cribriform mesh, and usually exhibited clearly discernable osseous tissue hypertrophy. Lesions of severe expression were interpreted to represent skeletal changes that were associated with extensive osseous tissue hypertrophy, marked expansion of diploic spaces, and commonly involved the frontal, parietal, and occipital bones in combination with the superior orbital plates. 4. In the present study porotic hyperostotic lesions were simply classified as orbital or extra-orbital with respect to cranial bones affected. Other workers have suggested lesion typologies based on physical appearance of the lesion or the anatomical disposition of the lesion in the skull. For example, Nathan and Haas (1966) described orbital porotic hyperostosis as being porotic, cribrotic, or trabeculated in appearance. Carlson and associates (1974) have alternatively suggested that lesions be classified as cribra orbitalia, spongy hyperostosis, and osteoporotic pitting. These descriptive terms were not used in the present study because they primarily reflect degree of osseous tissue response and no further pathophysiological significance can be attributed to the use of such jargon. Thus, the terminology employed here corresponds with that currently advocated by several researchers who have directed their long term efforts toward developing a bettern comprehension of the nature and significance of porotic hyperostosis in earlier human groups (G. J. Armelagos, personal communication). 5. Age-specific Libben:Bt-5 qx ratios are listed in Table 2. The average Libben:Bt-5 qx ratios for adults and subadults were estimated by calculating the mean for all qx ratios that were observed for each age group. The values for adults (+15 years) and subadults (0-15 years) are 1.24 and 1.60, respectively. 6. The Anasazi skeletal materials are permanently curated by the American Museum of Natural History, New York. 7. The Peruvian skeletal materials are permanently curated by the Chicago Field Museum, Illinois. References Aegerter, E. and J.A. Kirkpatrick. 1975. Orthopedic Diseases: Physiology, Pathology, and Radiology. (4th edition.) W.B. Saunders Co., Philadelphia. Agarwal, K., N. Dhar, M.N. Shah, and O.P. Bhardway. 1970. Roentgenologic changes in iron deficiency anemia. American Journal of Roentgenology, Radiation Therapy, and Nuclear Medicine 110:635-637. Akel, R.N., J.D. Frankish, C.P. Powles, K.R. Tyler, J. M. Watt, H. J. Weston, and I. A.M. Prior. 1963. Anemia in Maori and European infants and children on admission to hospital: a cooperative survey from six New Zealand hospitals. New Zealand Medical Journal 62:29-33. Andelman, M.B., and B.R. Sered. 1966. Utilization of dietary iron by term infants. American Journal of Diseases of Children 111:45-55. Aksoy, M., N. Camli, and S. Erdem. 1966. Roentgenographic bone changes in chronic iron deficiency anemia. Blood 27:677-686. Angel, J.L. 1964. Osteoporosis: thalassemia? American Journal of Physical Anthropology 22:369-374. Angel, J.L. 1966. Porotic hyperostosis, anemias, malarias, and marshes in the prehistoric eastern Mediterranean. Science 153:760-763. Angel, J.L. 1967. Porotic hyperostosis or osteoporosis symmetrica, pp. 378-389. In: Diseases in Antiquity. Edited by Sandison and Brothwell. Thomas Publishing Co., Springfield, Illinois. Apte, S.V., and L. Iyengar. 1970. Absorption of dietary iron in pregnancy. American Journal of Clinical Nutrition 23:13-71. Arbeter, A., L. Echeverri, D. Franco, D. Munson, H. Velez, and J. J. Vitale. 1971. Nutrition and Infection. Federation Proceedings 30:1421-1428. Ashworth, A., P.F. Milner, J. C. Waterlow, and W. B. Walker. 1973. Absorption of iron from maize (Zea mays L.) and soya beans (Glycine hispida Max.) in Jamaican infants. British Journal of Nutrition 29:269-278. Ascenzi, A., and P. Salistreri. 1977. Porotic hyperostosis and the problem of the origin of thalassemia in Italy. Journal of Human Evolution 6:595-604. Badenoch, J., J.R. Evans, and W.C.D. Richards. 1957. The stomach in hypochromic anemia. British Journal of Hematology 3:175-185. Baggs, R.B., and S.A. Miller. 1973. Nutritional iron deficiency as a determinant of host resistance in the rat. Journal of Nutrition 103:1554-1560. Baker, D.H. 1964. Roentgen manifestations of Cooley’s anemia. Annals of the New York Academy of Sciences 1 19:642-661 . Baker, S.J. 1978. Nutritional anaemia — a major controllable public health problem. Bulletin of the World Health Organization 55:659-675. Becker, M.H., N.B. Genieser, S. Piomelli, D. Dove, and R.D. Mendoza. 1971. Roentgenographic manifestations of pyruvate kinase deficiency hemolytic anemia. American Journal of Roentgenology, Radiation Therapy, and Nuclear Medicine 113:491-498. 1991 Paleoepidemiology of P orotic Hyperostosis 41 Beisel, W.R. 1982. Single nutrients and immunity. American Journal of Clinical Nutrition 35: 417-468. Betke, K. 1970. Iron deficiency in children, pp. 519-528. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Betts, N.M., and A. Weidenbenner. 1986. Dietary intakes, iron status and growth status of Southeastern Asian refugee children. Nutrition Research 6:509-515. Bhaskaram, P., J.S. Prasad, and K.A. Krishnamanchari. 1977. Anemia and immune response. Lancet 1 : 1000. Bhaskaram, P. and V. Reddy. 1975. Cell-mediated immunity in iron and vitamin-deficient children. British Medical Journal (Clinical Research) 3:522. Bowen, J.E. 1990. The late prehistory of northwestern Ohio. Ph.D. Dissertation, Ohio State University. Bradfield, R.B., V.V. Jensen, L. Gonzales, and C. Garrayar. 1968. Effect of low-level iron and vitamin supplementation on a tropical anemia. American Journal of Clinical Nutrition 21:57-67. Brendstrup, P. 1950. Intravenous iron loadings of patients with infections. Acta Medica Scandinavica 239:201-209. Bressani, R„ R. Paz, and N.S. Scrimshaw. 1958. Chemical changes in corn during the preparation of tortillas. Journal of Agriculture and Food Chemistry’ 6:770-774. Britton, H.A., J.P. Canby, and C.M. Kohler. 1960. Iron deficiency anemia producing evidence of marrow hyperplasia in the calvarium. Pediatrics 25:621-627. Brubaker, R.R., E.D. Beesley, and M.J. Surgalla. 1965. Pasteurella pestis: role of pesticin 1 and iron in experimental plague. Science 149:422-424. Bullen, J.J., H.J. Rogers, and L. Leigh, 1972. Iron-binding proteins in milk and resistance to Escherichia coli in infants. British Medical Journal 1:69-75. Bullen, J.J., and H.J. Rogers. 1968. Effect of haemoglobin on experimental infections with Pasteurella septica and Escherichia coli. Nature 217:86. Bullen, J.J., L.C. Leigh, and H.J. Rogers. 1968. The effect of iron compounds on the virulence of Escherichia coli for Guinea- pigs. Immunology 15:581-588. Burko, H., H.Z. Mellins, and J. Watson. 1961 . Skull changes in iron deficiency anemia simulating congenital hemolytic anemia. American Journal of Roentgenology 86:447-452. Burks, J.M., M.A. Siimes, W.C. Mentzer, and PR. Dallman. 1976. Iron deficiency in an Eskimo village. The value of serum ferritin in assessing iron nutrition before and after a three month period of iron supplementation. Journal of Pediatrics 88:224-228. Caffey, J. 1939. The skeletal changes in the chronic hemolytic anemias (erythroblastic anemia, sickle cell anemia, and chronic hemolytic icterus). American Journal of Roentgenology 37:293-324. Callender, S.T., B.J. Mallett, and M.D. Smith. 1957. Absorption of hemoglobin iron. British Journal of Hematology 3: 1 86- 1 92. Carlson, D.S., G.J. Armelagos, and D.P. Van Gerven. 1974. Factors influencing the etiology of cribra orbitalia in prehistoric Nubia. Journal of Human Evolution 3:405-410. Carpenter, C. 1987. Infectious diarrheal syndromes, pp. 502-506. In: Harrison's Principles of Internal Medicine. (11th edition.) Edited by E. Braunwald et al. McGraw Hill, New York. Cartwright, G.E., M.A. Lauritsen, P.J. Jones, I.M. Merril, and M.M. Wintrobe. 1946. The anemia of infection. I. Hypoferremia, hypercupremia, and alterations in porphyrin metabolism in patients. Journal of Clinical Investigation 25:65-80. Cassidy, C.M. 1980. Nutrition and health in agriculturalists and hunter-gatherers: a case study of two prehistoric populations, pp. 117-145. In: Nutritional Anthropology: Contemporary Approaches to Diet and Culture. Edited by R. F. Kandal and G. H. Pelto. Redgrave, Pleasantville, New York. Chandra, R.K. 1973. Reduced bactericidal capacity of polymorphs in iron deficiency. Archives of Diseases of Childhood 48:864-866. Chandra, R.K. 1985. Trace element regulation of immunity and infection. Journal of American College of Nutrition 4:5-16. Chandra, R.K. and B. Au. 1980. Single nutrient deficiency and cell- mediated immune responses. III. Vitamin A. Nutritional Research 1:181-1 85. Chandra, R.K. and A.K. Saraya. 1975. Impaired immuno- competence associated with iron deficiency anemia. Journal of Pediatrics 86:899-902. Charley, P.J., B. Sarkar, C.F. Stitt, and P. Saltman 1963. Chelation of iron by sugars. Biochemica and Biophysica, Acta 69:3 1 3-32 1 . Cohen, M.N. and G.J. Armelagos. 1984. Paleopathology at the Origins of Agriculture. Academic Press, New York. Conrad, M.E. 1970. Factors affecting iron absorption, pp. 87-120. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Cook, J.D., and E.R. Monsen. 1976. Food iron absoiption in human subjects. III. Comparison of the effect of animal proteins on nonheme iron absorption. American Journal of Clinical Nutrition 29:859-867. Cook, J.D., G.M. Brown, and L.S. Valberg. 1964. The effect of achylia gastrica on iron absorption. Journal of Clinical Investigation 43: 1 1 85- 1 1 9 1 . Cule, J., and I.L. Evans. 1968. Porotic hyperostosis and the Gelligaer Skull. Journal of Clinical Pathology 2 1 :753-758. Davidson, L.S.P., H.W. and Fullerton. 1938. Chronic nutritional hypochromic anemia. Edinburgh Medical Journal 45:1-23. Davidson, L.S.P., H.W. Fullerton, and R.M. Campbell. 1935. Nutritional iron deficiency anemia with special reference to prevalence and age and sex incidence. British Journal of Medicine (August):389 1-3894. Davidson, W.M.B., and J.L. Markson. 1955. The gastric mucosa in iron deficiency anemia. Lancet (2):639-643. Davis, PS. 1970. The effect of iron-binding proteins on iron absorption, pp. 129- 134. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Dawson, J.P, and J.F. Desforges. 1958. Dietary and storage factors in iron deficiency anemia of infancy. American Journal of Diseases of Children 96: 169-178. 42 Mensforth No. 46 Dykstra. O.H., and T. Halberstma. 1940. Polycythemia vera in childhood (report of a case with changes in the skull). American Journal of Diseases of Children 60:907-916. El-Najjar, M.Y. 1976. Maize, malaria, and the anemias in the pre- Columbian New World. Yearbook of Physical Anthropology 20:329-337. El-Najjar, M.Y., B. Lozoff, and D.J. Ryan. 1975. The paleo- epidemiology of porotic hyperostosis in the American Southwest: radiological and ecological considerations. American Journal of Roentgenology, Radiation Therapy, and Nuclear Medicine 125:91 8-924. El-Najjar, M.Y., and A.L. Robertson. 1976. Spongy bones in prehistoric America. Science 193:141-143. El-Najjar, M.Y., D.J. Ryan, C.G. Turner, and B. Lozoff. 1976. The etiology of porotic hyperostosis among the prehistoric and historic Anasazi Indians of Southwestern United States. American Journal of Physical Anthropology 44:477-488. Eng, L.L. 1958. Chronic iron deficiency anemia with bone changes resembling Cooley’s anemia. Acta Haematologica 19:263-268. Fagundes-Netto, U. 1984. Ultrastructural study of alterations in the small intestinal epithelium of children with acute diarrhea. Journal of Pediatric Gastroenterology Nutrition 3:5 10-5 15. Fairgrieve, S.I. (1990) Size of the sella turcica and its relation to iron deficiency anemia: a prehistoric example. American Journal of Physical Anthropology 83:111-121. Farb, P., and G.J. Armelagos. 1980. Consuming Passions: the Anthropology of Eating. Houghton Mifflin Co., Boston. Faulk, W.P. 1976. The immunological system in health and malnutrition. Proceedings of the Nutrition Society 35:253-261. Ferris, N. and R.G. Mayer 1 990. Rescue Excavations of the Younge Phase Cherry Lane (AaHp-21) site, Leamington, Essex County, Ontario. Report prepared by Mayer, Poulton & Associates, Inc. 134 Commissioners Rd., London, Ontario, NGJ 1X8 Finch, C.A. 1968a. Want amid plenty. American Journal of Clinical Nutrition 21:1138. Finch, C.A. 1968b. Iron deficiency in the United States. Journal of the American Medical Association 203:407-412. Finch, C.A. 1970. Diagnostic value of different methods to detect iron deficiency, pp. 409-422. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Fletcher, J. 1971. The effect of iron and transferrin on the killing of Escherichia coli in fresh serum. Immunology 20:493-500. Forman, M.R., B.I. Graubard, H.J. Hoffman. R. Beren, E.E. Harley, and P. Bennett. 1984. The Pima infant feeding study: breast- feeding and respiratory infections during the first year of life. International Journal of Epidemiology 13:447-453. Foy, H., A. Kondi, and W.H. Austin. 1959. The effect of dietary phytate on fecal absorption of radioactive ferric chloride. Nature 183:691-692. Fullerton, H.W. 1937. The iron deficiency anemia of late infancy. Archives of Diseases of Childhood 12:91-1 10. Garibaldi, J.A. 1972. Influence of temperature on the biosynthesis of iron transport compounds by Salmonella typhirium. Journal of Bacteriology 1 10:262-265. Germana, F., and A. Ascenzi. 1980. Porotic hyperostosis in Sardinia. Journal of Human Evolution 9:75-78. Ghosh, S., S. Daga, D. Kasthuri, R.C. Misra, and H.K. Chuttania. 1972. Gastrointestinal function in iron deficiency states in children. American Journal of Diseases of Children 123:14-17. Goodman, A., D. Martin, and G.J. Armelagos. 1984. Indicators of stress from bones and teeth, pp. 13-49. In: Paleopathology at the Origins of Agriculture. Edited by M. Cohen and G.J. Armelagos. Academic Press, New York. Gordon, J.E.. I. D. Chitkara, and J. B. Wyon. 1963. Weanling Diarrhea. American Journal of Medical Science 245:345-377. Gordon, J.E., J.B. Wyon, and W. Ascoli. 1967. The second year death rate in less developed countries. American Journal of Medical Science 254: 1 2 1 - 1 44. Grantham-McGregor, S.M., P. Desai, and P.F. Milner. 1974. Hematological levels in Jamaican infants. Archives of Diseases of Childhood 49:525-530. Greenberg, G.R., H. Ashenbrucker, M. Lauritsen, M.M. Wintrobe. 1947. The anemia of infection. IV. The lack of relationship between the diversion of iron from the plasma and the origin of the anemia. Journal of Clinical Investigation 26: 114-1 20. Greenfield, G.B. 1975. Radiology of Bone Diseases. (2nd edition), J. B. Lippincott Co., Philadelphia. Grieger, T.A., and M.J. Kluger. 1978. Fever and survival: the role of serum iron. Journal of Physiology (London) 279:187-196. Gross, R.L. and P.M. Newberne. 1980. Role of nutrition in immunologic function. Physiology Reviews 60:188-302. Gryboski, J., and W.A. Walker. 1983. Gastrointestinal Problems in the Infant. (2nd edition.) W. B. Saunders Co., Philadelphia. Guha, D.K., B.N.S. Walia, B.N. Tandon, and O.P. Ghai. 1968. Small bowel changes in iron deficiency anemia of childhood. Archives of Diseases of Childhood 43:239-244. Guyton, A.C. 1981. Medical Physiology. W.B. Saunders Co., Philadelphia. Hallberg, L., H. G. Harwerth, and A. Vanotti. 1970. Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Academic Press, New York. Hallberg, L., and L. Sovell. 1967. Absorption of hemoglobin iron in man. Acta Medica Scandinavica 181 :335-354. Halsted, J.A., A.S. Prasad, and M. Nadimi. 1965. Gastrointestinal function in iron deficiency anemia. Archives of Internal Medicine 16:253-256. Hanson, L.A., and J. Winberg. 1972. Breast milk defense against infection in the newborn. Archives of Diseases of Childhood 47:454-460. Hardesty, R.M., and D.J. Weatherall. 1982. Blood and Its Disorders. Blackwell Scientific Publications, Oxford. Harrison, M.L. 1978. Taphonomy of the Libben site, Ottawa County, Ohio. Thesis, Kent State University, Kent, Ohio. Hawksley, J.C., R. Lightwood, and U.M. Bailey. 1934. Iron deficiency anemia in children: its association with gastrointestinal disease, achlorhydria and hemorrhage. Archives of Diseases of Childhood 9:359-372. Heath, C.W., and A.J. Patek, 1937. The anemia of iron deficiency. Medicine (Baltimore) 16:267-350. 1991 Paleoepidemiology of P orotic Hyperostosis 43 Hegsted, D.M., C.A. Finch, and T.D. Kinney. 1949. The influence of diet on iron absorption. II. The inter-relation of iron and phosphorous. Journal of Experimental Medicine 90: 147- 1 56. Heinrich, H.C. 1970. Intestinal iron absorption in man — methods of measurement, dose relationship, diagnostic and therapeutic applications, pp. 213-296. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Hengen, O.P. 1971. Cribra orbitalia: pathogenesis and probable etiology. Homo 22:57-75. Henschen, F. 1961. Cribra cranii, a skull condition said to be of racial or geographical nature. Pathologica et Microbiologia (Basel) 24:724-729. Herbert, J.R. 1985. Effects of water quality and water quantity on nutritional status: findings from a South Indian Community. Bulletin of the World Health Organization 63:143-153. Herbert, V. 1987. Recommended dietary intakes (RDI) of iron in humans. American Journal of Clinical Nutrition 45:679-686. Hercberg, S., M. Chauliac, M. Devanlay, P. Galan, J.L. Pureur, Y Soutre, A. Houdegbe, Y. Agboton, I. Zohoun, A. Masse- Raimbault, and H. Dupin. 1986. Evaluation of the iron status of a rural population in South Benin. Nutrition Research 6:627-634. Hillman, R.S. 1970. Factors affecting hemoglobin regeneration, pp. 531-536. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Hillman, R.S., and P. A. Henderson. 1969. Control of marrow production by level of iron supply. Journal of Clinical Investigation 48:454-460. Hodges, D.C. 1987. Health and agricultural intensification in the prehistoric valley of Oaxaca, Mexico. American Journal of Physical Anthropology 73:323-332. Hoffman-Goetz, L., and M.J. Kluger. 1979. Protein deprivation: its effects on fever and plasma iron during bacterial infection in rabbits. Journal of Physiology ( London) 295:419-430. Hoffman-Goetz, L., D. McFarlane, B. Bistrian, and G.L. Blackburn. 1981. Febrile and plasma iron responses in rabbits injected with endogenous pyrogen from malnourished patients. American Journal of Clinical Nutrition 34: 1 109-11 16. Hooton, E.A. 1930. Indians of Pecos Pueblo. Yale University Press, New Haven, Connecticut. Hoshishima, K., M. Ito, and S. Hyodo. 1985. The trace dose of metal(s) and the immunological reaction in mice. Nutritional Research (Supplement l):740-744. Howell, N. 1982. Village composition implied by a paleodemographic life table. American Journal of Physical Anthropology 59:263-269. Hunter, C.A. 1960. Iron deficiency anemia in pregnancy. Surgery, Gynecology, and Obstetrics 110:210-214. Hrdlicka, A. 1914. Anthropological work in Peru in 1913, with notes on the pathology of ancient Peruvians. Smithsonian Miscellaneous Collections 1914, LXI, (18), 1-69. Hussain, R., and V.N. Patwardhan. 1959. The influence of phytate on the absoiption of iron. Indian Journal of Medical Research 47:676-682. Hwang, Y.F., and E.B. Brown. 1965. Effect of desferrioxamine in iron absorption. Lancet 1 : 1 35- 1 37. Jacobs, A., and M. Wormwood. 1982. Iron metabolism, iron deficiency, and overload, pp. 149-197. In: Blood and Its Disorders. Edited by R.M. Hardesty and D.J. Weatherall. Blackwell Scientific Publications, Oxford. Jacobs, A. and D.H.M. Joynson. 1974. Lymphocyte function and iron-deficiency anemia. Lancet (2):844. Jaffe, H.L. 1972. Metabolic, Degenerative, and Inflammatory Diseases of Bones and Joints. Lea & Febiger, Philadelphia. Jelliffe, D.B., and V. Blackman. 1962. Bahima disease. Possible “milk anemia” in late childhood. Journal of Pediatrics 61:774-779. Jose, D.G., and J.S. Welch. 1970. Growth retardation, anemia, and infection with malabsorption and infestation of the bowel. Medical Journal of Australia 1 :349-356. Josephs, H. 1956. Hypochromic microcytic anemia of infancy: iron depletion as a factor. Pediatrics 18:959-978. Josephs, H. 1953. Iron metabolism and the hypochromic anemia of infancy. Medicine (Baltimore) 32:125-213. Josephs, H. 1936. Anemia of infancy and early childhood. Medicine (Baltimore) 15:307-451. Joynson, D.H., A. Jacobs, D.M. Walker, and A.E. Dolby. 1972. Defect of cell-mediated immunity in patients with iron- deficiency anemia. Lancet 2:1058-1059. Kaatz, M.R. 1955. The Black Swamp: a study in historical geography. Annals of the Association of American Geographers XLV(l):l-35. Kaplan, G.J., J.K. Fleshman, T.R. Bender, C. Baum, and PS. Clark. 1973. Long-term effects of otitis media. A ten-year cohort study of Alaskan Eskimo children. Pediatrics 52:577-585. Kaye, D., and E.W. Hooke. 1963. The influence of hemolysis of blood loss on susceptibility to infection. Journal of Immunology 91:65-75. Keenlyside, D.L. 1978. Late prehistory of Point Pelee, Ontario and environs. Archaeological Survey of Canada, Paper No. 80, National Museums of Canada, Ontario. Kenney, M.A. 1985. Factors related to iron nutrition of adolescent females. Nutrition Research 5:157-166. Kent, S. 1987. The influence of sedentism and aggregation on porotic hyperostosis and anaemia: a case study. Man 21:605-636. Kilpatrick, G.S. 1970. Prevalence of anemia in the United Kingdom, pp. 441-446. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Kluger, M.J., and B.A. Rothenburg. 1979. Fever and reduced iron: an immune response to bacterial infection. Science 203:374-376. Kochan, I. 1978. The role of iron in bacterial infections, with special considerations of host-tubercle bacillus interaction. Current Topics in Microbiology and Immunology 60: 1 -30. Kochan, I. 1977a. The role of serum iron in the regulation of nutritional immunity. Advances in Chemistry Series 162:55-77. 44 Mensforth No. 46 Kochan, I. 1977b. Role of siderophores in nutritional immunity and bacterial parasitism, pp. 251-288. In: Microorganisms and Minerals. Edited by E. D. Weinberg. Dekker Press, New York. Korones, S.B. 1976. High-risk Newborn Infants. (2nd edition), C.V. Mosby Co., St. Louis, Missouri. Krakker, J.J. 1984. Investigations of a small late prehistoric site in St. Clair County, Michigan. The Michigan Archaeologist 30:86-123. Krakker, J.J. 1983. Changing sociocultural systems during the late prehistoric period in southeast Michigan. Ph.D. Dissertation, University of Michigan. Krantman, H.J., S.R. Young, B. J. Ank, C.M. O’Donnell, G.S. Rachelefsky, and E.R. Stiehm. 1982. Immune function in pure iron deficiency anemia. American Journal of Diseases of Children 136:840-844. Krogman, W.M., M.Y. and Iscan. 1986. The Human Skeleton in Forensic Medicine. (2nd edition.) C. C. Thomas, publisher, Springfield, Illinois. Kuhn, I.N., M. Layrisse, M. Roche, C. Martinez, and R.B. Walker. 1968. Observations on the mechanisms of iron absorption. American Journal of Clinical Nutrition 21:11 84- 1 1 88. Kuhns, W.J., C.J. Gubler, G.E. Cartwright, and M.M. Wintrobe. 1950. The anemia of infection. XIV. Response to massive doses of intravenously administered saccharated oxide of iron. Journal of Clinical Investigation 29: 1 505- 1513. Kuvibidila, S.R., B.S. Baliga, and R.M. Suskind. 1983a. The effect of iron-deficiency anemia on cytolytic activity of mice spleen and peritoneal cells against allogenic tumor cells. American Journal of Clinical Nutrition 38:238-244. Kuvibidila, S.R., B.S, Baliga, and R.M. Suskind. 1983b. Impairment of blastogenic response of splenic lymphocytes from iron- deficient mice: in vivo repletion. American Journal of Clinical Nutrition 37:15-25. Kuvibidila, S.R., B.S. Baliga, and R.M. Suskind. 1981. Effects of iron-deficiency anemia on delayed cutaneous hypersensitivity in mice. American Journal of Clinical Nutrition 34:2635-2640. Lahey, M.E. 1962. Protein-losing enteropathy. Pediatric Clinics of North America 9:689-702. Lallo, J.W., G.J. Armelagos, and R.P Mensforth. 1977. The role of diet, disease, and physiology in the origin of porotic hyperostosis. Human Biology 49:47 1-483. Lanzkowski, P. 1968. Radiological features of iron deficiency anemia. American Journal of Diseases of Children 1 16:16-29. Lanzkowski, P„ and D. McKenzie. 1959. Iron deficiency in Cape Coloured and African children in Cape Town. South African Medical Journal 33:22-24. Laskari, A. 1984. Hematologic Manifestations of Childhood Diseases. Thieme-Stratton, New York. Layrisse, M., J.D. Cook, C. Martines-Torres, M. Roche, I.N. Kuhn, R.B. Walker, and C.A. Finch. 1969. Food iron absorption: a comparison of vegetable and animal foods. Blood 33:430-443. Layrisse, M., C. Martinez-Torres, and M. Roche. 1968. Effect of interaction of various foods on iron absorption. American Journal of Clinical Nutrition 2 1 : 1 1 75- 1 1 83. Liebman, M 1985. Iron and folate status of an adolescent female population. Nutrition Research 5:621-625. Lennox, P.A. 1982. The Bruner-Colasanti site: an early late woodland component, Essex County, Ontario. Archaeological Survey of Canada, Paper No. 110, National Museums of Canada, Ottawa. Levine, S.Z., and H. H. Gordon. 1942. Physiological handicaps of the premature infant. I. Their pathogenesis. II. Clinical applications. American Journal of Diseases of Children 64:274-296, 297-312. Lovejoy, C.O., R.S. Meindl, R. P. Mensforth, and T. J. Barton. 1985. Multifactorial determination of skeletal age at death: a method and blind tests of its accuracy. American Journal of Physical Anthropology 68:1-14. Lovejoy, C.O., R.S. Meindl, T.R. Przybeck, T.J. Barton, K.G. Heiple, and D. Kotting. 1977. Paleodemography of the Libben site, Ottawa County, Ohio. Science 198:291-293. Macdougall, L.G., R. Anderson, G.M. McNab, and J. Katz. 1975. The immune response in iron deficient children: impaired cellular defense mechanisms with altered humoral components. Journal of Pediatrics 86:883-843. Mackay, H.M.M. 1933. Factors causing variation in hemoglobin level with age in the first year of life. Archives of Diseases of Childhood 8:251-263. Mackay, H.M.M. 1931. Nutritional anemia in infancy. With special reference to iron deficiency. Medical Research Council Special Reports, Series No. 157:1-125. Mahoney, A.W., and D.G. Hendricks. 1975. Utilization of dietary calcium by iron-deficient rats. Nutrition and Metabolism 18:6-15. Malina, R.M. 1975. Growth and Development: the First Twenty Years in Man. Burgess Publishing Co., Minneapolis, Minnesota. Manchandra, S.S., H. Lai, and S. Khanna. 1969. Iron stores in health and disease. Bone marrow studies in 1134 children in Punjab, India. Archives of Diseases of Childhood 44:580-584. Marquardt, W.H.. and P.J. Watson. 1983. The shell mound archaic of Western Kentucky, pp. 323-339. In Archaic Hunters and Gatherers in the American Midwest. Edited by Phillips and Brown. Academic Press, New York. Marquardt, W.H. 1972. Recent investigations in a Western Kentucky shellmound. Paper presented at the Annual Meetings of the Society for American Archaeology, Miami Beach, Florida. Martin, C.M., J.H. Jandl, and M. Finland. 1963. Enhancement of acute bacterial infections in rats and mice by iron and their inhibition by human transferrin. Journal of Infectious Diseases 112:158-163. Martinez-Torres, C., and M. Layrisse. 1974. Interest for the study of dietary absorption and iron fortification. World Review of Nutrition and Dietetics 19:51-70. Martinez-Torres, C., and M. Layrisse. 1971. Iron absorption from veal muscle. American Journal of Clinical Nutrition 24:53 1-539. Martinez-Torres, C., and M. Layrisse. 1970. Effect of amino acids on iron absorption from a staple vegetable food. Blood 35:669-682. Martorell, R. 1980. Interrelationships between diet, infectious disease, and nutritional status, pp. 81-106. In: Social and Biological Predictors of Nutritional Status. Physical Growth, and Neurological Development. Edited by L. S. Green and F. E. Johnston. Academic Press, New York. 1991 Paleoepidemiology of Porotic Hyperostosis 45 May, J.A. 1979. Shell midden formation processes: a methodological perspective. Paper presented at the 36th Annual Meeting of the Southeastern Archaeological Conference, Atlanta, Georgia, November, 1979. Maynard, J.E., and L.M. Hammes. 1970. A study of growth, morbidity and mortality among Eskimo infants in Western Alaska. Bulletin of the World Health Organization 42:6 1 3-622. McFarlane, H. 1976. Malnutrition and impaired immune response to infection. Proceedings of the Nutrition Society 35:263-282. McFarlane, H., and J. Hamid, 1973. Cell-mediated immune response in malnutrition. Clinical Experimental Immunology 13:153-164. McFarlane, H., S. Reddy, K.J. Adcock, H. Adeshina, A.R. Cooke, and J. Akene. 1970. Immunity, transferrin, and survival in Kwashiorkor. British Medical Journal (October):268-270. McMurray, D.N. 1984. Cell-mediated immunity in nutritional deficiency. Progress in Food and Nutrition Science 8: 193-228. Mensforth, R.P 1985. Relative tibia long bone growth in the Fibben and Bt-5 prehistoric skeletal populations. American Journal of Physical Anthropology 68:87-106. Mensforth, R.P. 1986. The pathogenesis of periosteal reactions in earlier human groups: diagnostic, epidemiological, and demographic considerations. Paper presented at the 55th Annual Meeting of the American Association of Physical Anthropologists, Albuquerque, New Mexico, April 1986. American Journal of Physical Anthropology 69:239-240 Mensforth, R.P. 1990. Paleodemography of the Carlston Annis (Bt- 5) Fate Archaic skeletal population. American Journal of Physical Anthropology 82:81-99. Mensforth, R.P, C.O. Lovejoy, J.W. Fallo, and G.J. Armelagos. 1978. The role of constitutional factors, diet, and infectious disease in the etiology of porotic hyperostosis and periosteal reactions in prehistoric infants and children. Medical Anthropology 2: 1-59. Moore, S. 1929. The bone changes in sickle cell anemia with a note on similar changes observed in skulls of Ancient Mayan Indians. Missouri Medicine 26:561-564. Moore, C.V., H.Q. Bierman, V. Minnich, and W.R. Arrowsmith. 1940. Studies on iron absorption with special reference to the relationship of iron metabolism to ascorbic acid. American Association for the Advancement of Science 1 3:34-37. Morck, T.A., S.R. Lynch, B.S. Skikne, and J.D. Cook. 1981. Iron availability from infant food supplements. American Journal of Clinical Nutrition 34:2630-2634. Moseley, J.E. 1974. Skeletal changes in the anemias. Seminars in Roentgenology 9: 1 69- 1 84. Moseley, J.E. 1966. Radiographic studies in hematologic bone diseases: implications for paleopathology, pp. 121-130. In: Human Paleopathology. Edited by S, Jarcho. Yale University Press, New Haven, Connecticut. Moseley, J.E. 1965. The paleopathologic riddle of “symmetrical osteoporosis.” American Journal of Roentgenology 95 : 1 35- 1 42. Muller, H. 1935. Osteoporosis of the cranium in Javanese. American Journal of Physical Anthropology 20:493. Naiman, J.L. , F.A. Oski, L. K. Diamond, G.F. Vawter, and H. Swachman. 1969. The gastrointestinal effects of iron- deficiency anemia. Pediatrics 33:83-99. Nathan, H., and N. Haas. 1966. Cribra orbitalia, a bone condition of the orbit of unknown nature. Israel Journal of Medical Science 2:171-191. Nice, C.M., M.L. Daves, and G.H. Wood. 1964. Changes in bone associated with cyanotic congenital cardiac disease. American Heart Journal 68:25-3 1 . Olivares, M., F. Pizzaro, P. Chadud, and A. Stekel. 1986. Age- related changes in laboratory measures of iron nutrition in childhood. Nutrition Research 6:1045-1049. Palazzari, G.L., K. Matyniak, and L.O. Schulz. 1986. Iron status independent of socioeconomic variables in school-age children. Nutrition Research 6:1 131-1 137. Palkovich, A.M. 1987. Endemic disease patterns in paleopathology: porotic hyperostosis. American Journal of Physical Anthropology 74:527-537. Palkovich, A.M. 1985. Agriculture, marginal environments, and nutritional stress in the prehistoric southwest, pp. 425-438. In: Paleopathology at the Origins of Agriculture. Edited by M. Cohen and G.J. Armelagos, Academic Press, New York. Parry, M.F., R.K. Root, J.A. Metcalf, K.K. Delaney, L.S. Kaplow, and W.J. Richard. 1981. Myeloperoxidase deficiency: prevalence and clinical significance. Annals of Internal Medicine 95:293-301. Polk, H.C., and A. A. Miles. 1971. Enhancement of bacteria! infection by ferric iron: kinetics, mechanisms, and surgical significance. Surgery 70:7 1 -77 . Pollack, S.. R.M. Kaufman, and W.H. Crosby. 1964. Iron absorption: effects of sugars and reducing agents. Blood 24:577-581. Powell, J.W., H.S. Weens, and N.K. Wegner. 1965. The skull roentgenogram in iron deficiency anemia and in secondary polycythemia. American Journal of Roentgenology, Radiation Therapy, and Nuclear Medicine 95: 143-147. Prasad, J.S. 1979. Leukocyte function in iron-deficiency anemia. American Journal of Clinical Nutrition 32:550-552. Pritchard, J.A., and D.E. Scott. 1970. Iron demands during pregnancy, pp. 173-212. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Prufer, O.H. and O.C. Shane, III. 1976. The Portage-Sandusky- Vermilion River Region in Ohio, pp. 283-304 in The Late Prehistory of the Lake Erie Drainage Basin: A 1972 Symposium Revised. (Edited by D.S. Brose) Scientific Papers of The Cleveland Museum of Natural History, Cleveland, Ohio. Purtillo, D.T., and D.H. Connor. 1975. Fatal infections in protein- calorie malnourished children with thymolymphatic atrophy. Archives of Diseases of Childhood 50: 149- 1 53. Putschar, W.G.J. 1966. Problems in the paleopathology of bone, pp. 57-65. In: Human Paleopathology. Edited by S. A. Jarcho. Yale University Press, New Haven, Connecticut. Rawson, A., and F.Q. Rosenthal. 1960. The mucosa of the stomach and small intestine in iron deficiency. Lancet ( 1 ):730-731 . Robbins, S.L. 1974. Pathologic Basis of Disease. W.B. Saunders Company. Philadelphia. 46 Mensforth No. 46 Roche, M., and M.E. Perez-Gimenez. 1959. Intestinal loss and reabsorption of iron in hookworm infection. Journal of Laboratory and Clinical Medicine 54:49-52. Root. P.K., and M.S. Cohen. 1981. The microbial mechanism of human neutrophils and eosinophils. Review of Infectious Disease 3:565-598. Ruff, C.B. 1981. A reassessment of demographic estimates for Pecos Pueblo. American Journal of Physical Anthropology 54:147-151. Rybo, G. 1970. Menstrual loss of iron, pp. 163-172. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Saddi, R., and G. Schapira. 1970. Iron requirements during growth, pp. 183-198. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Santos, J.I. 1986. Nutritional implications and physiologic response to pediatric diarrhea. Pediatric Infectious Disease 5 (supplement): 1 52- 1 54. Sax, B. 1963. Roentgen manifestations of iron deficiency anemia in the skull of infants and children simulating those seen in Cooley’s and sickle-cell hemolytic anemia. Germantown Hospital, Philadelphia Journal 4:72-75. Schade, A.L. 1963. Significance of serum iron for the growth, biological characteristics, and metabolism of Staphylococcus aureus. Biochemische Zeitschrift 338:140-148. Schutte, D. 1979. Iron deficiency anemia as the cause of porotic hyperostosis among the PreColumbian Maya. Thesis, California State University, Fullerton. Scott, D.E., J.A. Pritchard, A.E. Saltin, and J.M. Humphreys. 1970. Iron deficiency during pregnancy, pp. 491-504. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Scrimshaw, N.S. 1964. Ecological factors in nutritional disease. American Journal of Clinical Nutrition 1-: 1 12-122. Scrimshaw, N.S., and R.M. Suskind. 1959. Interactions of nutrition and infection. American Journal of Medical Science 237:367-403. Scrimshaw, N.S., and V.R. Young. 1976. The requirements of human nutrition. Scientific American 235 (September):51-64. Shah, A.G., and S. Seshadri. 1985. Anemia in relation to dietary iron deficiency, iron availability and parasitic infestation. Nutrition Research 5:341-347. Shahidi, N.T., and L.K. Diamond. 1960. Skull changes in infants with chronic iron deficiency anemia. New England Journal of Medicine 262:137-139. Sharpe, L.M., W.C. Peacock, R. Cooke, and R.S. Harris. 1950. The effect of phytate and other food factors on iron absorption. Journal of Nutrition 41 :433-446. Shaw, R., and W.O. Robertson. 1964. Anemia among hospitalized infants. Ohio State Medical Journal 60:45-47. Shearman, D.J. C., I.W. Delamore, and D.L. Gardner. 1966. Gastric function and structure in iron deficiency. Lancet ( 1 ):845-848. Sheldon, W. 1936. Anemia with bone changes in the skull. Royal Society of Medicine, London Proceedings 29:743. Smith, C.H. 1972. Blood Diseases of Infancy and Childhood. (3rd edition). C.V. Mosby Company, St. Louis, Missouri. Sodeman, W.A., and W.A. Sodeman. 1974. Pathologic Physiology: Mechanisms of Disease. W. B. Saunders Company, Philadelphia. Srikantia, S.G., J.S. Prasad, C. Bhaskaram, and K. Krishnamachari. 1976. Anemia and immune response. Lancet ( 1 ): 1 307-1 309. Stewart, A. 1937. Gastric acidity in infants and young children under normal and pathological conditions, with special reference to nutritional anemia. British Journal of Children 's Diseases 34: 1 -27. Stuart-Macadam, P. 1985. Porotic hyperostosis: representative of a childhood condition. American Journal of Physical Anthropology 66:391-398. Stuart-Macadam, P. 1987a. A radiographic study of porotic hyperostosis. American Journal of Physical Anthropology 74:511-520. Stuart-Macadam, P. 1987b. Porotic hyperostosis: new evidence to support the anemia theory. American Journal of Physical Anthropology 74:521-526. Stuart-Macadam, P. 1989. Porotic hyperostosis: relationship between orbital and vault lesions. American Journal of Physical Anthropology 80: 1 87-193. Sturgeon, P. 1956. Iron metabolism: a review with special consideration of iron requirements during normal infancy. American Academy of Pediatrics 18:267-300. Thomas, W.J., H.M. Koenig, A.L. Lightsey, and R. Green. 1977. Free erythrocyte protoporphyrin: hemoglobin ratios, serum ferritin and transferrin saturation levels during treatment of infants with iron deficiency anemia. Blood 49:455-462. Till, J.E., and E.A. McCulloch. 1961. A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiation Research 14:213. Trentin, J.J. 1971. Determination of bone marrow stem cell differentiation by stromal hemopoietic inductive microenvironment. American Journal of Pathology 65:621. Trucco, J.I., and A.K. Brown. 1967. Neonatal manifestation of hereditary spherocytosis. American Journal of Diseases of Childhood 1 13:263-270. Turnbull, A.F. Cleton, C.A. Finch, L. Thompson, and J. Martin. 1967. Iron Absorption. IV. The absorption of hemoglobin iron. Journal of Clinical Investigation 41:1897-1907. Vannotti, A. 1957. The role of the reticuloendothelial system in iron metabolism, pp. 172-186. In: Physiology of the Reticuloendothelial System. Blackwell Scientific Publications, Oxford. Venkatachalam, P.S. 1968. Iron metabolism and iron deficiency in India. American Journal of Clinical Nutrition 21:1 156-1 161. Von Endt, D.W., and D.J. Ortner. 1982. Amino acid analysis of bone from a possible case of prehistoric iron deficiency anemia from the American southwest. American Journal of Physical Anthropology 59:377-385. 1991 Paleoepidemiology of P orotic Hyperostosis 47 Wakefield, E.G., S.C. Dellinger, and J.D. Camp. 1937. A study of the osseous remains of the “mound builders” of Eastern Arkansas. Journal of American Medical Science 193:489-495. Walker, P.L. 1986. Porotic hyperostosis in a marine-dependent California Indian population. American Journal of Physical Anthropology 69:345-354. Walter, T., S. Arrendondo, M. Arevalo, and A. Stekel. 1986. Effect of iron therapy on phagocytosis and bactericidal activity in neutrophils of iron-deficient infants. American Journal of Clinical Nutrition 44:877-882. Webb, W.S. 1950. The Carlston Annis mound site 5 Butler County, Kentucky. University of Kentucky Reports in Anthropology, volume 7, (number 4), Lexington, Kentucky. Weinberg, E.D. 1966. Roles of metallic ions in host-parasite interactions. Bacteriological Reviews 30: 1 36- 151. Weinberg, E.D. 1977. Microorganisms and Minerals. Edited by E.D. Weinberg. Marcel Dekker, Inc., New York. Weinberg, E. D. 1974. Iron and susceptibility to infectious disease. Science 184:952-956. Weiss, K.M. 1972. On the systematic bias in skeletal sexing. American Journal of Physical Anthropology 37:239-250. Weiss, K.M. 1973. Demographic models for anthropology. Memoirs of the Society for American Archaeology (number 27), American Antiquity 38, (number 2), part II. Welcker, H. 1888. Ci'ibra orbitalia. Archive fur Anthropologie 17:1-18. Wen-guang, W.C. Xue-cun, and L. Dong-sheng. 1986. Hematologic response to iron and ascorbic acid administration in preschool children with anemia. Nutrition Research 6:241-248. Werkman, S.L., L. Shifman, and T. Skelly. 1964. Psychosocial correlates of iron deficiency anemia in early childhood. Psychosomatic Medicine 26: 1 25- 1 34. Williams, A.O., S.B. Lagundoye, and C. L. Johnson. 1975. Lamellation of the diploe in the skulls of patients with sickle cell anemia. Archives of Diseases of Childhood 60:948-952. Williams, H.U. 1929. Human paleopathology: with some original observations on symmetrical osteoporosis of the skull. Archives of Pathology 7:839-902. Wilson, J.F., D.C. Heiner, and M.E. Lahey. 1962. Studies on iron metabolism. I. Evidence of gastrointestinal dysfunction in infants with iron deficiency anemia: a preliminary report. Journal of Pediatrics 60:787-800. Winters, R. H. 1974. Introduction to the new edition of Indian Knoll, by W.S. Webb, University of Tennessee Press, Knoxville. Witts, L.J. 1966. Anemia as a world health problem. International Society of Hematology, 11th Congress of the University of Sidney, Plenary Sessions, 85-102. Wood-Jones, F. 1910. General pathology (including diseases of the teeth). In: Report on Human Remains. Edited by G. E. Smith and F. Wood-Jones. Archaeological Survey of Nubia report for 1907-1908, volume 2, Cairo. Woodruff, C.W. 1958. Multiple causes of iron deficiency anemia in infants. Journal of the American Medical Association 167:715-720. Woodruff, C.W., and E.P. Bridgeforth. 1953. Relationship between the hemogram of the infant and that of the mother during pregnancy. Pediatrics 12:681. World Health Organization. 1968. Nutritional anemias: report of a WHO scientific group. WHO Technical Report Series No. 405. Wretlind, A. 1970. Food iron supply, pp. 39-70. In: Iron Deficiency: Pathogenesis, Clinical Aspects, Therapy. Edited by Hallberg, Harwerth, and Vanotti. Academic Press, New York. Yetkin, S., S. Altay, G. Ciliv, and Y. Lalely. 1979. Myelo- peroxidase activity and bacterial function of PMN in iron deficiency. Acta Haematologica 61:10-14. Yusufji, D., V.I. Mathan, and S.J. Baker. 1973. Iron, folate, and vitamin B12 nutrition in pregnancy: a study of 1000 women from Southern India. World Health Organization Bulletin 48:15-22. Zaino, D.E., and E.C. Zaino. 1969. Cribra orbitalia in the aborigines of Hawaii and Australia. American Journal of Physical Anthropology 42:91-94. Zaino, E.C. 1964. Paleontologic thalassemia. Annals of the New York Academy of Sciences 1 1 9:402-4 1 2. Zaino, E.C. 1967. Symmetrical osteoporosis, a sign of severe anemia in prehistoric Pueblo Indians of the Southwest, pp. 40-77. Miscellaneous Papers in Paleopathology /, Museum of Northern Arizona, Flagstaff. Zaino, E.C. 1968. Elemental bone iron in the Anasazi Indian. American Journal of Physical Anthropology 29:433-436. KIRTLANDIA The Cleveland Museum of Natural History August 1991 Number 46:49-7 1 North American Late Devonian Cephalopod Aptychi Calvin J. Frye and Rodney M. Feldmann Department of Geology Kent State University Kent, Ohio 44242 Abstract Enigmatic, Hat body fossils have been collected from at least 9 localities in dark shales of Late Devonian age in northeastern Ohio. The fossils have been found mainly in the Cleveland Shale, a black shale interpreted to represent an anoxic environment, and more rarely in the Chagrin Shale, which was deposited in a dysaerobic environment. The benthic fauna of these shales is sparse and restricted. These Ohio fossils are comparable to similar structures found within the Woodford Shale of Oklahoma, a formation of equivalent age and depositional environment as the Chagrin and Cleveland shales. The enigmatic fossil remains comprise at least seven authentic species referrable to the genus Sidetes Giebel, 1849. Five of these species are found in Ohio. Spathiocaris tenuicosta Cooper, 1932 is morphologically indistinguishable from Sidetes chagrinensis (Ruedemann, 1916), and is, therefore, placed in synonymy. Similarly, Spathiocaris striatula Cooper, 1932 is the junior subjective synonym of Sidetes lata (Ruedemann, 1916) and Spathiocaris woodfordi Cooper, 1932 and Spathiocaris plicifera Cooper, 1932 are junior subjective synonyms of Sidetes newberryi (Whitfield, 1882). All of the fossils are extremely thin and, typically, flat structures marked with fine, nearly concentric, corrugations or folds and range from 0.7 cm to 8 cm in length. At various times they have been considered to be brachiopods, barnacle plates, cephalopod aptychi, or the phyllopod crustaceans Spathiocaris or Aptychopsis. Scanning electron microscopy reveals no ultrastructure within the fossils. Electron-dispersive X-ray spectroscopy indicates they contain neither calcium, strontium, nor phosphorus. Brachiopods and arthropods from the same units do contain phosphorous. Their general morphology and ornamentation is also unlike that of brachiopods or arthropods, permitting their assignment to the Cephalopoda. They appear to be the aptychi of ammonites, structures which probably served as the animal’s lower jaw. Their probable preservation as carbon films remnant of degraded organic material is consistent with what is known of cephalopod aptychi. The reconstruction of two specimens that had been cracked and Battened during compaction shows the original form of the structures to have been broadly curved and scooplike. This is consistent with reconstructions of Mesozoic ammonite jaws, and strengthens the assignment of these fossils with the cephalopods. 50 Frye and Feldmann No. 46 Introduction In recent years there have been widely scattered reports of fossils believed to be cephalopod aptychi from the Paleozoic of North America (Brady, 1955, but see Yochelson, 1971; Closs, et al., 1964; Saunders and Spinosa, 1974; Saunders and Richardson, 1979; Thompson, et al., 1980; Yochelson, 1983; Kues, 1983; Mapes, 1987; Landman and Davis, 1988; Harper, 1989). Although several of these reports claim to be one of only a handful of such findings, many more specimens are available for study. The collections of The Cleveland Museum of Natural History contain over a hundred such specimens from the Late Devonian Chagrin and Cleveland shales. These fossils are preserved as flat, glossy black, carbonaceous structures in the dark shales. They are marked with fine concentric ridges which parallel the margin or are truncated by it along the periphery. The ridges resemble those sometimes seen in brachiopods or bivalves, and the general outline could be suggestive of some early crustaceans. Similar fossils also have been identified as fish scales, barnacle plates, or perhaps gastropod opercula. In the earliest description of these Devonian fossils in North America they were identified as a form of phyllocarid crustacean, Spathiocaris (Clarke, 1882). Woodward (1885), however, acknowledged that some “phyllocarids” could, in fact, be goniatite aptychi. Later, Clarke (1902) expressed doubts, admitting that they could be cephalopod aptychi or brachiopod fragments instead. This study was begun in an attempt to determine their affinities. It was expected that the application of new methods might yield additional insight into the subject. Examination of the microstructure of these structures using the scanning electron microscope might reveal details of their formation and growth which would permit a more definitive identification. Determination of the chemical composition of these structures by x-ray spectroscopy might also confirm their affinities. Cephalopod shell material is aragonitic, whereas their mandibles are calcitic (Lowenstam, et al., 1984). Aragonite frequently contains strontium as a significant trace element. Inarticulate brachiopod shells and arthropod carapaces are composed primarily of calcium phosphate, not calcium carbonate. New chemical data would not necessarily be definitive, however, for bivalves are predominately aragonitic. The principal purposes of this study are to examine and describe the “spathiocarids” of the Cleveland and Chagrin shales, investigate their relationships with similar Devonian taxa, and attempt to provide solid identification of their nature, if possible. This latter goal was not fully realized, but two of the most likely alternatives have been eliminated. It is probable that these fossils should be referred to the Cephalopoda. Morphological Terminology The terms describing the various forms of these fossil structures are complicated and somewhat confused. Many of the terms, originally defined as morphological features, have subsequently been adopted as taxonomic names. Further, a number of names of taxa have since been considered as morphological terms. It seems clearest to use the simple set of terms proposed by Moore and Sylvester- Bradley (1957a) in the Treatise on Invertebrate Paleontology. Aptychus (plural; aptychi) is considered to be a general name for this group of fossil cephalopod fragments, although it has a restricted meaning as the group of bivalved forms. Anaptychus is used to refer to the univalved structures found alone or in association with the pair of aptychi sensu stricto in younger rocks. For general discussion, these terms are useful: "aptychus [broad sense | — All types of calcareous or corneous structures presumed to serve as opercula [or mandibles] of ammonoid conchs. diaptychus [= aptychus sensu stricto] — Aptychus composed of two discrete valves, anaptychus — Univalved type of aptychus.” (Terms modified from Moore and Sylvester-Bradley, 1957a.) With the multiplicity of interpretations, published descriptions of aptychi and similar specimens are hard to reconcile. The life orientation of these fossil fragments is necessarily different if they are arthropod carapaces (Clarke, 1882; Ruedemann, 1916), brachiopods (Clarke, 1902), cephalopod mandibles (Lehmann, 1970), or opercula (Trauth, 1927;Turek, 1978). The orientation of these structures would be the same whether they were interpreted as crustacean carapaces or cephalopod mandibles. In each case, the rostrum or apex (center of the concentric ornamentation) is anterior; the hinged region is medial. The morphological terms used here (Figure 1) are based on those defined by Clarke (1962) for the description of coleoid cephalopod mandibles. The anterior angle is here defined as the angle of the anterior margin of the flattened fossil; convex if the rostrum is emergent, concave if it is reentrant. Should it prove correct that aptychopsid plates were nautiloid opercula (Holland, et al., 1978; Turek, 1978) whereas aptychi proper served as ammonite mandibles (Lehmann, 1970), then new terminology would have to be created for the former. Not only would the function be different, but the two structures, otherwise similar in appearance, would have had opposite orientations in the living animals. Previous Work In 1847, C.G. Giebel reported the finding of some enigmatic fossil molds in the Cretaceous sandstones around 1991 Late Devonian Cephalopod Aptychi 51 Anterior Posterior 2 Interior FIGURE 1. Stylized diagrams of an early cephalopod lower mandible, showing reference attitude and morphological terms. Terminology after Clarke (1962). 1, Ventral view, extended and flattened. 2, Exterior lateral ( oblique ) view. Salzburg. He described the genus Sidetes Giebel, 1849, two years later, concluding that these structures were the aptychi of Sepia Linnaeus, 1758 (Giebel, 1849). The specimens he observed were semicircular, and marked with fine concentric lines. Later, John M. Clarke (1882) described several odd fossils from Naples, New York, found in Givetian and Frasnian age shales of the Hamilton and Genesee formations. He collected thirty specimens over the course of several years, but remained doubtful as to their biological affinities. The fossils ranged in size from a few millimeters in breadth to as much as 90 mm, were flattened elliptical bodies marked with concentric lines or ridges, and bore a wedge-shaped cleft. He concluded they were not brachiopods as they were too large and did not display any trace of a corresponding ventral valve. Clarke's descriptions were based on what he thought were incomplete specimens, consisting of isolated carapaces. He noted their general similarity to the Silurian arthropod genus Discinocaris Woodford 1856, and suggested these new specimens belonged properly with the phyllopod crustaceans, and that the discovery of a complete specimen would come with time. He assigned them to two crustacean genera, Spathiocaris Clarke 1882, and Lisgocaris Clarke 1882. At the same time, Whitfield (1882) found similar fossils, which he referred to the crustacean genus Plumulites Barrande 1872. These specimens were all recovered from the Cleveland Shale in Erie County, Ohio. Over the course of several decades of work with the British Museum, Henry Woodward had occasion to describe many small fossils of similar type. He proposed (Woodward, 1865; and see Woodward, 1885a) that some of these “ink- flecks” were chiton plates, while others were isolated barnacle plates, for which he proposed the genus Turrilepas Woodward, 1865. He subsequently (Woodward, 1882) referred similar specimens from the Devonian of Biidesheim, Germany to the new phyllopod genera Cardiocaris Woodward, 1882, and Pholadocaris Woodward, 1882, and a specimen from the Silurian of Wales to the phyllopod genus Aptychopsis Barrande, 1872. Later, he (Woodward, 1885) agreed that some such fossils may be cephalopod aptychi, but felt that others were certainly phyllopod carapaces. Ruedemann’s discovery, in 1901, of a very large brachiopod in the “Hudson river shales” (sic] of New York prompted Clarke to summarize 25 years of collecting Spathiocaris and similar fossils (Clarke, 1902). He observed that none had yet been discovered with abdominal fragments, and a specimen of Spathiocaris had been found in the body chamber of the Devonian goniatite Manticoceras intumescens in Germany (Kayser, 1882). He had earlier illustrated a similar occurrence of a “phyllocarid” (Dipterocaris Clarke, 1883) within a goniatite from the Naples (Portage) shales of New York (Hall, 1888, pi. 35). His conclusions at this time were that Spathiocaris and similar forms were probably cephalopod aptychi, but that Discinocaris was perhaps a brachiopod. He gave this interpretation for the latter as its occurrence in the Silurian preceeded the appearance of the ammonites. At about the same time, other dark shales were being studied. Girty’s monograph (1909) on the Caney Shale (Devonian-Mississippian) of Oklahoma included a new genus, ldiotheca, which Girty hesitantly described as a conulariid. He briefly stated other possibilities, including its interpretation as a cephalopod aptychus. He ruled out the possibility of its being an inarticulate brachiopod. Later, Ruedemann (1916) described four new species of Spathiocaris from New York and northeastern Ohio. He suggested several reasons to consider these fossils as belonging to the Cephalopoda, observing that the method of growth seen in Spathiocaris and related genera is similar to that seen in aptychi and not like that of brachiopods or arthropods. He supposed that the horny anaptychus would logically preceed calcareous diaptychi 52 Frye and Feldmann No. 46 in the evolution of these structures, so the presumed lack of diaptychi from the Paleozoic does not mean that spathiocarids could not have been anaptychi. Finally, he suggested that similar structures would, “also have existed in the Ordovician and Silurian cephalopods and thus account for those earlier anaptychi considered as Discinocarina [sic]” (Ruedemann, 1916, p. 102). Six new species of Spathiocaris were described by C.L. Cooper (1932) from the Woodford Formation of Oklahoma, including the redescription of Girty’s Idiotheca specimen as the new species Spathiocaris woodfordi. The Woodford is an interbedded black shale/chert unit of Late Devonian to Early Mississippian age. Many of Cooper’s descriptions are similar to species from the east. He described the fauna strictly as crustacean. Ruedemann (1934) expanded upon the idea that Spathiocaris was a cephalopod aptychus, citing as evidence Matern’s (1931) discovery of Spathiocaris koeneni Clarke, 1884 within the body chamber of Crickites holzapfeli Wedekind, 1913, a European Devonian goniatite. Ruedemann said it was unlikely that this represents the preservation of a phyllocarid preying upon a goniatite. The suggestion was made by Matern, and echoed by Ruedemann, that the anaptychi were separated from the conchs as the cephalopods decomposed while still buoyed by gases in the conch. In his description of the New Albany Shale of Indiana, Campbell (1946) mentioned several thin beds within this Devonian black shale as “ Spathiocaris beds.” He suggested these horizons, where these fossils were locally abundant, might be useful in stratigraphic correlation. Unfortunately, such occurrences are too rare to be helpful (Lineback. 1970; Hasenmueller and Leininger, 1987). Spathiocaris has also been identified in drill cores from western Canada (Copeland and Boulton, 1960), along with a phyllopod (crustacean) telson. This last has been removed from association with Spathiocaris and redescribed as Montecaris (Pratt, 1987). Materials Specimens in this study were collected by many different individuals from 1925 to 1989. Most were collected by RA. Bungart or F. Thompson incidental to collection of Cleveland Shale fish material for The Cleveland Museum of Natural History (CMNH). All specimens studied were borrowed from The Cleveland Museum of Natural History, except for three specimens of Aptychopsis Barrande, 1872 which were kindly loaned by the Palaeontological Institute of Lund, Sweden (LO), Whitfield’s type specimens provided by the American Museum of Natural History (AMNH), and Cooper’s type specimens, which were borrowed from the National Museum of Natural History (USNM). All the fossils studied from Ohio were preserved as carbonized films, flattened and compacted into the shale. Rarely was there a good interface between the specimen and the matrix, so preparation was held to a minimum. In a few cases folded specimens were separated from the rock along their outer surfaces. Most were prepared for photographic illustration by coating with finely particulate ammonium chloride. A few fragments were removed and coated with a thin film of gold for examination with the scanning electron microscope, but uncoated specimens were also examined by this technique with good results, probably due to their carbon content. The USNM specimens, from the Woodford Shale of Oklahoma, are preserved as three-dimensional ellipsoidal packages of thin, sheetlike fossil material within phosphatic concretions from the shale matrix. Stratigraphy and Localities The fossils in this study have come from two units within the Late Devonian of northeastern Ohio; the dark colored Chagrin Shale and the overlying black Cleveland Shale (Fanrmenian). These shales are exposed along the southern shore of Lake Erie for 150 kilometers and extensively along many of the streams draining into the lake. The Devonian shale outcrop belt in northeast Ohio is illustrated in Figure 2, along with the known sites from which “spathiocarids” have been collected. These shale units represent prograding distal deposition of fine-grained sediments from the Catskill Delta to the east during a time of marine transgression in the Appalachian Basin (Lewis, 1988). Prosser first used the name “Chagrin Shale” to describe the unit of interbedded gray shales and siltstones which Newberry called the “Erie Shale,” as the latter term was preoccupied (Prosser, 1912, pp. 14-15). The unit is a wedge- shaped body thickening eastward into western Pennsylvania, where it is correlative with the Riceville Shale. It is underlain by the Huron Shale, another black shale. The Chagrin thins westward and pinches out east of the Huron River (Lewis, 1988). The Chagrin Shale consists primarily of greenish-gray or bluish-gray clayey shales interbedded with discontinuous siltstones. The shales frequently are bioturbated and often contain other traces of benthic life (Barron and Ettensohn, 1981; Hannibal and Feldmann, 1983; Schwimmer, 1988; Schwimmer and Feldmann, 1990). They represent gradual deposition within a dysaerobic low- energy environment (Barron and Ettensohn, 1981). The siltstones probably represent episodic storm events, washing coarser deltaic sediments westward in the basin (Hannibal and Feldmann, 1983). The Cleveland Shale, named by Newberry in 1870, thins both eastward and westward from its maximum thickness west of Cleveland (Lewis, 1988). It consists primarily of black, laminated, fissile shales containing more organic matter and quartz and less clay (illite) than the gray shales of the Chagrin (Broadhead, et al., 1982). The lack of an 1991 Late Devonian Cephalopod Aptychi 53 FIGURE 2. Outcrop map of Devonian shales in northeastern Ohio. Numbered localities refer to sites where “ spathiocarids ” have been collected. active benthos and the enrichment in sulfides and organics indicates that the Cleveland Shale was deposited in deeper, anoxic conditions (Barron and Ettensohn, 1981). While it lies over the Chagrin Shale, there is a general east-west transition between the two units as the gray Chagrin grades westward into the black Cleveland Shale (Prosser, 1912; Szmuc, 1970a). The Cleveland Shale is overlain in turn by the Bedford Formation. The contact is sharp but conformable, marked by a thin bed of pyritized brachiopods and vertebrate fragments (Szmuc, 1970b). Localities The known localities at which these fossils have been found are listed here, in order from west to east. The numbers correspond to those in Figure 2. 1) Chance Creek — An easterly tributary of the Vermilion River in Forain County. Kipton 7.5 Minute Quadrangle, Brownhelm Township, T6N, R19W, 41°21'40''N, 82° 18 '00 "W. Exposures occur on Chance Creek, 400 m south of the intersection of Vermilion Road and Gifford Road. The Cleveland Shale in this region is approximately 15 m thick (Fewis, 1988). Specimens CMNH 3744, 3746, and 3747 were collected by William J. Hlavin from a zone 0.5 to 1.5 m below the Cleveland-Bedford formational contact. “The base of the invertebrate zone is characterized by a thin bone bed which contains water-worn, disarticulated elements and teeth of fossil fish” (Hlavin, 1976). 2) Cahoon Cliffs — Cliffs along the Fake Erie shore in Bay Village, Ohio. North Olmstead 7.5 Minute Quadrangle, T7N, R15W, 41°29'15"N, 81°55'30"W. The 10 m cliff east of the mouth of Cahoon Creek is an excellent exposure of the Chagrin Shale (Prosser, 1912). CMNH 6620 was collected as float along the beach, 100 meters east of the creek mouth. 3) Little Cedar Point — A bluff at the confluence of the East and West branches of the Rocky River. North Olmstead 7.5 Minute Quadrangle, T6N, R15W, 41°24'40"N, 81°53’20"W. The Cleveland Shale in this area forms steep, high cliffs along the river. The thickness of the unit appears to be 30 m, with the upper third somewhat more resistant than the rest (Prosser, 1912). Numerous concretions, cone-in-cone structures, and pyrite nodules have been found in the Cleveland Shale in this area. Forty-four specimens were located in the collections of The Cleveland Museum of Natural History in association with labels which read, “100 yds. E of ford at base of Fittle Cedar Point, 6/24/51” and “5/29/51, in a landslide.” Although it is not clear which specimens, if any, are rightfully associated with these labels, the locality has been productive of both vertebrate and invertebrate material. 4) Abram’s Creek — A southerly tributary of the Rocky River. Fakewood 7.5 Minute Quadrangle, T6N, R14W, 41 °25 '05 "N, 8 1 °52’00"W. CMNH 8312 was collected where Abram’s Creek meets Rocky River, not far downstream from Fittle Cedar Point. 5) Big Creek Localities — Along Big Creek from Brookside Park to the Big Creek Metropark and beyond, the upper 30 m of the Chagrin and at least 15 m of the Cleveland are exposed. The Cleveland appears at the top of the cliffs on the south side of the creek just above Brookside Park; about three km upstream it reaches the stream bed south of West Park Cemetery, near the western edge of the Cleveland South Quadrangle (Prosser, 1912). Focalities in the upper reaches of Big Creek include four along the northwest branch of the creek, and a region west of Finndale and north of Memphis Road. The fossils were collected as float and occasionally in situ in the creek bed. Fakewood 7.5 Minute Quadrangle: 5a) 30 m east of W. 140th St., T7N, R14W, 41°26'15"N, 81°47'30''W. CMNH 8317. 5b) First bend below W. 130th St., T7N, R13W, 41°26'15”N, 81°46'45"W. CMNH 8304. 5c) Above W. 128th St., T7N, R13W, 41°27'00"N, 81°46'45"W. CMNH 8303, CMNH 8315, CMNH 8316, CMNH 8318. 5d) At W. I 17th St. T7N, R13W, 41°27T5"N, 8 1 °46 '00 " W. CMNH 3745. 5e) Region of the Metropark north of Memphis Road, T7N, R13W, 4 1 °23 ’45 "N to 41°26'30"N, 54 Frye and Feldmann No. 46 8 1 °45 '15 "W to 8 1 °45 '30 "W. CMNH 7942, CMNH 8159, CMNH 8306, CMNH 8309, CMNH 8310, CMNH 831 1, CMNH 8314. 6) Brookside Park — Located along Big Creek between the Big Creek Metropark and the creek’s terminus at the Cuyahoga River. Trace fossils have been collected as float along the base of outcrops south of Big Creek in this area (Hannibal and Feldmann, 1983). Cleveland South 7.5 Minute Quadrangle, T7N, R13W, 41°26'30''N to 41°27'30"N, 8 1 °42'15 ”W to 81°44'00"W. CMNH 8305, CMNH 8307, CMNH 8313. 7) Skinner's Run — On the border of Brooklyn Heights, Parma, and Seven Hills, Ohio. The Chagrin and the Cleveland are exposed along this tributary of the Cuyahoga River. At their contact is the Skinner’s Run pyrite bed (Hlavin, 1976), a pyritized lag deposit enriched in fossils. Specimens were collected from the lower portion of the Cleveland Shale, somewhat above the pyrite bed. South central */9th of the Cleveland South 7.5 Minute Quadrangle, T6N, R12W, 41°24'30"N, 81°40'30"W. 8) Euclid Creek - — East of Cleveland, in Euclid. The Chagrin Shale is exposed along Euclid Creek’s west branch in the Euclid Creek Metropark. East Cleveland 7.5 Minute Quadrangle, T8N, R11W, 41°33'00''N, 81°31'45"W. CMNH 6576 was collected in situ north of a small gully extending from the park road, 1.8 km (1.1 mi) south along the park road from its intersection with Highland Road. 9) Mill Creek — Camp Stigwandish, Lake County, Ohio. The Chagrin Shale is exposed in the cliffs along this tributary of the Grand River. Thompson 7.5 Minute Quadrangle, T11N, R6W, 41°44'15"N, 81°02'00”W. CMNH 7948, collected as float from a steep cliff on the west side of the stream, 87 m upstream from the Doty Road bridge. Systematic Paleontology General Aptychi present a particular problem to systematists. They are distinctive enough to be useful in stratigraphy, particularly where they are locally abundant enough to constitute “aptychus beds” (Campbell, 1946; Trauth, 1930). It is useful to be able to distinguish the forms by name, and indeed, many names were applied to these fossils by earlier workers before their nature was understood. However, the variability resulting from preservation and compaction has led to the erection of more species than was perhaps warranted (Turek, 1978). As cephalopod conchs were discovered with associated aptychi, taxonomic conflicts arose; aptychus names often had priority over those well-known for the conch. Article 23 of the Code of Zoological Nomenclature (International Commission on Zoological Nomenclature, 1985) would, in strict interpretation, have the earlier established name become that of the taxon, although section (b) of that article indicates that if this were to cause instability or confusion, an author can refer the case to the Commission for an individual ruling. In either case, one of the two names would have to be suppressed. There is a complex heirarchy of ammonite taxonomy, based entirely on characters exhibited by the conchs. Aptychi do not possess sufficient morphological variation to permit diagnosis at the generic or specific levels. Thus in some cases, single “species” of aptychi have been found to belong to two or more genera of ammonites as distinguished by conchs. A simple solution to this problem was proposed by Arkell (1954; 1957a). He favored the suppression of all names based solely on the aptychus of an ammonite. This proposal does assure the stability of ammonite nomen- clature, at the expense of abandoning names useful for identifying aptychi as discrete entities separate from the remainder of the animal, as is often the case. A more sweeping suggestion was made by Moore and Sylvester-Bradley (1957b) that a separate, parallel system of nomenclature be established for “parataxa;” names based on aptychi, individual conodonts, and isolated holothurian elements. In particular these names would compete with whole-animal names for the purposes of homonymy but not for priority. This proposal was fully supported by Arkell (1957b) as an extension of his original intent. This parataxon proposal has provoked much debate, such that nearly thirty years later the question of parataxa has again been “put aside... for further in-depth study and future consideration.” (International Commission on Zoological Nomenclature, 1985, p. xii). Until this question is resolved, the assignment of specimens to specific taxa is necessarily a cautious endeavour. Arkell (1957a) suggests using Trauth’s (1927, 1928, 1930, 1931, 1935, 1936) system of nomenclature as form-genera only, while others simply refer to aptychi or anaptychi in general terms if association with specific cephalopods cannot be proved (Lehmann, 1971, 1981; Harper, 1989). Trauth’s genera, however, are in many cases junior synonyms of older taxa. In this work, the taxa described for Devonian specimens will be considered appropriate, reserving “anaptychus” as a morphological term only. The “genus” Anaptychus Oppel, 1856, is an erroneous citation, as Oppel used the word merely as a morphological term describing the aptychus seen in Ammonites planorbis (citation of Oppel, 1856, in Moore and Sylvester-Bradley, 1957b). Anaptychus Stimpson, 1860 (Crustacea), and Anaptychus Schlumberger, 1868 (Cephalopoda), are junior subjective synonyms of Sidetes Giebel, 1847. 1991 Late Devonian Cephalopod Aptychi 55 Class Cephalopoda Leach, 1817 Genus Sidetes Giebel, 1847 Anaptychus Schlumberger, 1868; non anaptychus Oppel, 1856, morphological term; non Anaptychus Stimpson, 1860, Crustacea. Pholadocaris Woodward, 1882 Ca nlioca ris W c )o D wa r [ ) , 1882 ?Ellipsocaris Woodward, 1882 Lisgocaris Clarke, 1882 Spathiocaris Clarke, 1882 Idiotheca Girty, 1 909 Palanaptychus Trauth, 1927 Neoanaptychus Nagao, 1931 Type species Sidetes stricitus Giebel, 1 849. Diagnosis Semielliptic carbonaceous structure, weakly convex. Ornamentation of fine concentric lines about a medial anterior (by definition) rostrum, parallel to posterolateral margin, intersecting anterior margin at nearly right angles. Description Structure semielliptic, length 22 mm, weakly convex. Broad, width twice length. Anterior margin straight. Posterior margin smoothly curved. Ornamentation concentric with posterior margin, perpendicular with anterior margin, finely spaced at about 15/cm. Composition carbonaceous, probably conchiolinous, with no evidence of calcareous component. Locality of type species Unknown, “from hard sandstone banks near Salzburg” (Giebel, 1849). Age is Late Cretaceous (Senonian). Type Location unknown. Remarks Five species of Sidetes, described below, are recognised from the Ohio Shale. All are preserved as carbonaceous films, compressed and flattened to varying degrees. None appears to be accompanied by a calcitic or aragonitic component. All have a roughly semielliptic outline, and bear concentric ornamentation which terminates at the anterior margin in a manner unlike that of brachiopods or bivalves. The specimens described by Cooper (1932) have been reexamined, as they were collected from the time-equivalent Woodford Shale of Oklahoma, a unit similar in character and depositional setting to the Cleveland Shale (Cardott and Lambert, 1985). The Woodford specimens are preserved in a different manner than the Ohio forms, however. They do not exhibit the extreme flattening common with the Ohio specimens, but are to varying degrees three-dimensional, with a significant mass of matrix material preserved within the interior of the structure, almost as a “steinkern.” In general, they seem to have been preserved within calcareous or phosphatic concretions, a common alternative mode of preservation for [Upper-] Paleozoic aptychi (Mapes, 1987). All are similar to the Ohio specimens, with the addition of one species, S. gouldi, not recognised in the Ohio material. Clarke’s (1882) type species of Spathiocaris was Spathiocaris emersoni , originally described from New York. His original specimens have not been discovered. We were able to examine two specimens from Virginia (Butts, 1942), deposited at the National Museum of Natural History. In his discussion of Spathiocaris lata , Ruedemann described transverse frontal grooves extending halfway to the anterolateral angles. These grooves caused him to, “recall those of the aptychus of some ammonites” (Ruedemann, 1916, p. 95). Such grooves are not seen in any specimen in this study. It is possible that they may represent in some manner a reflexed portion of the anterior margin analogous to the short inner lamella seen in anaptychus-type ammonite jaws from the Mesozoic (Lehmann, 1979; Kanie, 1982; Tanabe, 1983). Such structures may well have been obliterated in highly compressed material as is common in the Ohio Shale. Exterior molds, of course, would not reveal the inner lamellae. Finally, the Woodford Shale (Cooper, 1932) material has not been prepared to reveal the interiors of the convexly folded specimens, so such structures may well be preserved within the matrix which remains. The species are differentiated on the basis of the ratio of breadth to length, the angle of the anterior margin, the general outline of the structure in its extended, flattened form, and, to a lesser degree, the nature of ornamentation. To this end, the following key is provided as an aid in species identification: KEY TO DEVONIAN APTYCHI REFERRABLE TO SIDETES la) Width greater than length 2 lb) Width equal to or less than length 3 2a) Width twice length; ornamentation finely spaced, 1 6/cm 5. newberryi 2b) Width about 46 length; ornamentation coarse, about 8/cm S. gouldi 3a) Width half to 4/5 length 4 3b) Width about equal to length 5 4a) Anterior margin acutely convex; outline elliptical S. ulrichi 4b) Anterior margin straight or broadly concave; outline triangular S. chagrinensis 4c) Anterior margin acutely concave, notched S. emersoni 56 Frye and Feldmann No. 46 FIGURE 3. Sidetes chagrinensis ( Ruedemann, 1916). 1, Neotype , CMNH 3745, from Big Creek in Cleveland. 2, CMNH 3746, from Chance Creek in Lorain Co. 3, USNM 112031 , from the Arbuckle Mountains of Oklahoma. (Spathiocaris tenuicosta Cooper, 1932). Scale is one centimeter. 5a) Outline sub-pentagonal, lateral margins nearly parallel, meeting anterior margin at distinct angle 5. lutheri 5b) Outline semielliptical, greatest width near anterior margin, narrowing posteriorly Anterior margin broadly concave S. lata Sidetes chagrinensis (Ruedemann, 1916) Figures 3.1 - 3.3 Spathiocaris chagrinensis Ruedemann, 1916, p. 95. Spathiocaris tenuicosta Cooper, 1932, p. 350. Diagnosis Structure elongate semielliptical, length greater than width. Posterior margin strongly rounded and narrow; greatest width anterior near rostrum. Lateral margins but slightly curved, extending obliquely forward; anterior margin straight or broadly concave. Concentric ridges closely arranged, not very prominent. Fine longitudinal lines along hood radiate from rostrum to posterior tip. Description Fossil elongate semielliptical, appearing as a rounded isosceles triangle with base at anterior margin. Length about one quarter greater than width, ranging from 25 to 38 mm. Posterior margin strongly rounded and narrow, lateral margins slightly sinuously curved, extending obliquely forward to anterior margin, which is straight or slightly obtuse and concave. Concentric ridges closely arranged (14- 28/cm), not very prominent. Fine longitudinal lines along hood radiate from rostrum to posterior tip, diverging slightly. Type Ruedemann’s holotype was collected from the Chagrin Shale at Chippewa Creek in Brecksville, Ohio. It was in the Western Reserve University collection, parts of which have been transferred to The Cleveland Museum of Natural History. This specimen has not been located, however. For this reason, we designate CMNH 3745 as the neotype to serve in place of the missing holotype. Material Examined in this study were two specimens collected in June or July of 1965 by William Hlavin from the Cleveland Shale. CMNH 3745 is from Big Creek near W. 117th St. in Cleveland, and CMNH 3746 is from Chance Creek, Lorain County. Also studied were two specimens from the Woodford Shale assignable to this species, namely USNM 1 12031, Cooper’s (1932) holotype of S. tenuicosta , and one of the paratypes of S. gouldi , USNM 112035, both from the Arbuckle Mountains of Oklahoma. Remarks Sidetes newberryi, S. lutheri , and S. lata are each broader than S. chagrinensis , with width/length ratios approximately 1991 Late Devonian Cephalopod Aptychi 57 one or greater. Sidetes ulrichi and S. emersoni , while also narrow, differ significantly in that the anterior margins are strongly curved rather than straight. Few specimens from the Chagrin Shale are well-preserved, and none of those now available represent this species. Even though the specimens in this study are from a different unit than the primary types, in all other respects they appear to conform to Ruedemann’s description and illustrations of the species. Cooper’s (1932) species, Spathiocaris tenuicosta , differs from S. chagrinensis only in the manner of its preservation. It is folded along the median into a groove and ridge rather than a simple crease. In all other respects, it resembles the other specimens described here, and is properly assigned to S. chagrinensis. Sidetes emersoni (Clarke, 1882) Figures 4. 1 , 4.2 Spathiocaris emersoni Clarke, 1882, p. 477. Diagnosis Structure semielliptical, length greater than width. Posterior margin rounded to subtriangular. Narrow, greatest width at anterior wing angle. Anterior margin deeply concave, notched. Concentric ridges closely arranged, well-marked. Description Fossils are semielliptical to subtriangular, width 2/3 to 4/5 length. Greatest width measured between wing angles, anterior of rostrum. Length ranges from 1 8 to 40 mm, width from 12 to 32 mm. Posterolateral margins straight to broadly curved. Ornamentation fine, about 20/cm, concentric with posterolateral margin, bending toward median at anterior margin. Anterior angle deeply concave, angle near 120°. Types Clarke’s type specimens were from the Portage shales in Naples, Ontario Co., New York. They have apparently been lost. Two specimens collected by Butts (1942) and now in the National Museum of Natural History, USNM 97992-a and -b, are designated as “hypotypes.” Material The specimens studied were the hypotypes, from Millboro, 1.6 km south of Shawver Mill, Virginia. Remarks Sidetes emersoni is narrower than S. newberryi or S. gouldi , while the deep anterior angle on S. emersoni differentiates it from S. ulrichi , S. chagrinensis , and S. lata. In outline, S. emersoni does not possess the sharp angle 2 FIGURE 4. Sidetes emersoni (Clarke. 1882). Hypotypes from Millboro, Virginia. 1, USNM 97992-a. 2, USNM 97992-b. Arrow points to fragment of opposite wing presetted on elevated portion of the matrix. Scale is one centimeter. where posterior and lateral margins meet, as does S. lutlteri , being instead nearly triangular. It is interesting that none of the Ohio nor Oklahoma specimens can be assigned to this, the type species for the genus Spathiocaris. Only the one specimen of S. lutheri may possess an anterior margin as concave as S. emersoni. All other specimens studied have anterior margins which are less concave or even convex. Sidetes gouldi (Cooper, 1 932) Figures 5.1, 5.2 Spathiocaris gouldi Cooper, 1932, p. 349. Diagnosis Structure large, semielliptical, broad; width about 1.5 times length. Anterior angle straight or broadly concave. Ornamentation concentric with posterolateral margin, terminating anteriorly with inward bend toward median; spacing coarse, about 8/cm. Description Fossils are large, length about 40 mm; outline semielliptical; broad, width about 65 mm (1.5 times length). Anterior angle appears nearly straight. Ornamentation consists of ridges concentric with posterolateral margin, spaced about 8/cm. Anterior portions of ridges curve inward toward median. 58 Frye and Feldmann No. 46 Figure 5. Sidetes gouldi f Cooper , 1932), from the Arbuckle Mountains. 1, USNM 112023, holotype, from Sycamore Creek, OK. 2, USNM 112034. Scale is one centimeter. Types Cooper’s (1932) type material includes the holotype, USNM 112023, and a paratype, USNM 112034. USNM 112035, another paratype, is smaller and is ornamented with much finer ridges than the other two specimens. It is also much narrower in outline than the others, and properly should be referred to S. chagrinensis. The two type specimens definitely referrable to this species are from the Arbuckle Mountains in Oklahoma; the holotype was collected from Sycamore Creek. Material Specimens studied were USNM 112023 and 112034. None of the Ohio specimens of appropriate breadth exhibit ornamentation so coarse as to permit their assignment to this species. Remarks Sidetes gouldi and S. newberryi are the broadest species studied, all others being much narrower. Sidetes gouldi is slightly less broad than S. newberryi , which approaches a width/length ratio of 2 to 1. Sidetes gouldi is further distinguished from S. newberryi in that the ornamentation is half as finely spaced as that of the latter. The two specimens have been removed from their surrounding concretions, and the anterior margins are poorly preserved, making determination of the anterior angle difficult. There is no indication, however, that it was significantly concave, but rather it appears to be nearly straight. While the spacing of the ornamentation near the rostrum frequently is finer than elsewhere on specimens of Sidetes, even this finer region is more coarsely ornamented than the ridges of S. newberryi , allowing easy distinction of the two species. Sidetes lata (Ruedemann, 1916) Figures 6.1 - 6.7 ?Cardiocaris lata Woodward, 1882, p. 388. Spathiocaris lata Ruedemann, 1916, p. 94. Spathiocaris striatula Cooper, 1932, p. 351. Diagnosis Structure semielliptical, length about equal to greatest width, which is near anterior margin. Anterior margin broadly concave. Ornamentation fine, concentric with posterolateral margin. Description Fossil semielliptical, posterolateral margin nearly circular or slightly flattened posteriorly. Greatest width, near anterior margin, approximately equal to length. Anterior angle broadly concave. Ornamentation fine, about 24/cm, concentric with posterolateral margin, bending toward median at anterior margin. 1991 Late Devonian Ceehaeopod Aptychi 59 FIGURE 6. Sidetes lata (Ruedemann, 1916). 1, CMNH 83 1 3a. from Big Creek at Brookside Park. Cleveland. 2, CMNH 8307a. also from Big Creek. 3, USNM 112038. syntype of Spathiocaris striatula Cooper. 1932. 4, L1SNM 112032. S. striatula syntype. 5. USNM 112028 6, USNM 112029 7, USNM 112041. Scale is one centimeter. 60 Frye and Feldmann No. 46 FIGURE 7. Sidetes lutheri ( Clarke , 1882). 1, CMNH 8302. Scale is one centimeter. 2, USNM 264093, the holotype. Scale is one millimeter Type Ruedemann’s holotype was collected in 1895 from the “Chemung beds” near Avoca, Steuben Co., New York. It is now in the New York State Museum, NYSM 9860. Material Specimens studied which can be assigned to this species include CMNH 8307a and CMNH 8313a from Big Creek at Brookside Park. Woodford Shale specimens which can be assigned to this species are the type specimens of S. striatula , USNM 1 12038 and 112032, and three specimens formerly identified as S. williamsi , USNM 112037, 1 12029, and 112041. These last specimens do not conform with Ruedemann’s description (1916) of Spathiocaris williamsi , being symmetrical and more narrow. Remarks Some specimens show superimposed, concentric undulations which are flatter and not as pronounced as those seen in S. ulrichi. Sidetes lutheri is distinctly different in outline from S. lata , while the other taxa are either decidedly narrower or significantly broader. Ruedemann described short carbonaceous lines radiating from a semicircular area at the anterior angle. These may be artifacts of preservation, as they have not appeared in any other specimen described from the North American Paleozoic. Spathiocaris striatula was distinguished from the other Woodford Shale species by striations “radiating from the apex to the lateral and posterior margins of the shell” (Cooper, 1932). This feature is variously affected by preservation. Ruedemann (1916) noted imperfectly preserved striations in several species. Such striations are not seen in any of the Ohio specimens, but are present in several of the Oklahoma fossils, which are generally better preserved. The proportions of S. striatula and the ornamentation are identical with Sidetes lata , to which it should be referred. Sidetes lutheri (Clarke, 1 882) Figures 7.1, 7.2 Lisgocaris lutheri Clarke, 1882, p. 478. Pholadocaris lutheri Ruedemann, 1916, p. 94 Diagnosis Outline sub-pentagonal, lateral edges parallel, meeting posterior margins at sharp angles. 1991 Late Devonian Cephalopod Aptychi 61 FIGURE 8. Sidetes newberryi ( Whitfield , 1882). Type specimens, from Erie County, Ohio. The remaining specimen in the type series is too degraded to reveal detail, and is not illustrated here. 1, AMNH CU 7452G. 2, AMNH CU 6686G. 3, AMNH CU 551 3G. 4, AMNH CU 6063G. Scale is one centimeter. Description Fossil large, length greater than 47 mm, width 43 mm. Outline sub-pentagonal; lateral margins parallel and meeting posterior margins at a sharp angle. Surface ornamented with concentric ridges parallel to margins, spaced about 16/cm. Ridges turn inward toward rostrum near anterior margin, which is not preserved. Type The holotype, USNM 264093, is in the collection of the U.S. National Museum. It was collected from near the base of the Hamilton Formation in Mile’s Gully, Hopewell, New York. Material In addition to the holotype, one Ohio specimen was examined. CMNH 8302 was collected in 1938 by P.A. Bungart from an unidentifiable locality near Linndale in the Berea quadrangle map. No stratigraphic information was recorded. Remarks The unique outline and approximately equal length and width serve to distinguish S. lutheri from all other taxa in this study. Clarke's (1882) description of Lisgocaris lutheri was based on a single, very small specimen measuring only two by three millimeters. The distinctive configuration of the margins and concentric ornamentation, however, is maintained in the much larger CMNH specimen. Clarke described the species in reverse orientation to that given above, with an “abdominal” cleft beginning centrally and widening to the “posterior” margin. This anterior region is not preserved in the CMNH specimen, so the size and shape of such a cleft cannot be determined. There is evidence that the lateral margins extended anterior of the rostrum a slight distance, but whether the anterior angle was obtuse or acute remains unknown. Sidetes newberryi (Whitfield, 1882) Figures 8. 1-8.4, 9. 1-9.7 Plumulites newberryi Whitfield, 1882, p. 217. Turrilepas newberryi Hall, 1888, p. 219-220. Idiotheca rugosa Girty, 1909, p. 40. Spathiocaris cushingi Ruedemann, 1916, p. 96. Spathiocaris woodfordi Cooper, 1932, p. 351. Spathiocaris plicifera Cooper, 1932, p. 350. Diagnosis Structure large, semielliptical, broad; width about twice length. Anterior angle straight or broadly concave. Ornamentation concentric with posterolateral margin, terminating anteriorly with inward bend toward the rostrum; spacing fine, about 16/cm. Description Structure large, length from 10 to 43 mm, semielliptical in outline, folded upon itself along the median line, forming a curved hinge. Broad, width 1.5-2 times the length. Anterior angle straight or very obtusely concave; anterior wings extend beyond hinge about one-fifth of total length. Entire structure extremely thin and flattened; in some specimens lateral margin is interrupted by one or more fissures, perpendicular to margin. Ornamentation of ridges concentric with posterolateral margin, spaced about 16/cm. Ridges continuous, terminating at anterior margin; anteriormost portion bent inward toward rostrum. Types Sidetes newberryi is the widest species herein studied. Only S. gottldi approaches it in width, but has much coarser ornamentation. All other species are rather narrower. Ruedemann (1916) described the new species Spathiocaris 62 Frye and Feldmann No. 46 Figure 9. Sidetes newberryi ( Whitfield , 1882). Specimens from the Cleveland and Woodford shales. 1, CMNH 8311, and 2, CMNH 8318, from Big Creek, arrows showing fissures extending inward from posterior lateral margin. 3, CMNH 8320. 4, CMNH 8327a, with gypsum encrustation. 5, USNM 112030, holotype of Spathiocaris plicifera Cooper, 1932. 6, USNM 112040, and 7, USNM 112033, types of Spathiocaris woodfordi Cooper, 1932. Scale is one centimeter. 1991 Late Devonian Cephalopod Aptychi 63 Figure 10. Sidetes ulrichi (Cooper, 1932). 1, USNM 112036, the holotype. 2, CMNH 3744, from Chance Creek. 3, CMNH 8305b, also from Big Creek. 4, CMNH 8328a. 5, CMNH 8328b. Scale is one centimeter. cushingi based upon two specimens of Turrilepas newberryi collected from the Cleveland Shale. These specimens were part of the Western Reserve University collection, parts of which ar now at The Cleveland Museum of Natural History. These two specimens have not been located, however. Whitfield’s specimens of Plumulites ( Turrilepas ) newberryi are at the American Museum of Natural History (AMNH), and include the syntypes AMNH CU 7452G, AMNH CU 6685G, AMNH CU 6063G, AMNH CU 6686G, and AMNH CU 5513G. They were collected from the Cleveland Shale near Sheffield and Birmingham, in Erie County, Ohio. These agree in all details with Ruedemann’s description and the specimens studied here, and do not exhibit the median carina nor multiple imbricate plates which might place them with the cirripedes. This species belongs with the aptychi. 64 Frye and Feldmann No. 46 FIGURE 11. Scanning electron micrographs of the cross-sections of the unmineralized wing area of cephalopod jaws. No internal structure is discernable in either specimen. 1, Sidetes sp. CMNH 8317. a presumed Devonian anaptychus. 2, Nautilus pompilius, a modern nautiloid. Scales are in microns as indicated. The Devonian specimen has been considerably compressed. Material Material studied includes Whitfield’s five specimens of Turrilepas, CMNH 8311 and CMNH 8318, from Big Creek, collected as float material, and CMNH 8312, from Abram’s Creek. CMNH 8320 and CMNH 8327a, also studied, have no reliable collection data recorded. Ruedemann’s types were collected by Professor H.P. Cushing from the Cleveland Shale along Cahoon Creek, not far from Cahoon Cliffs. One specimen, CMNH 8668, has been found within the old Western Reserve University collection, but it does not appear to be one of Ruedemann’s type specimens. Woodford Shale material assigned to this species are Cooper’s (1932) types of S. woodfordi , USNM 112033 and 112040, and the holotype of S. plicifera , USNM 112030. Remarks Ruedemann’s description of Spathiocaris cushingi included mention of a second set of concentric lines centered upon the wing angles of the valve. This set “is but faintly shown’’ in the second, larger specimen he described, and may be invisible depending on preservation. Its absence should not rule out the assignment of the CMNH specimens to this species. Girty (1909, p. 40) illustrated a new genus and species of Pteropoda, Idiotheca ritgosa , from the “Woodford chert” [sic] at the base of the Caney Shale (Devonian-Mississippian) in Oklahoma. He was uncertain of its affinities, granting that it “may possibly be an aptychus... which occur so abundantly at a little higher horizon...” (Girty, 1909). Cooper (1932) redescribed the same specimen, USNM 112044, along with another. USNM 112033, as the new phyllocarid species Spathiocaris woodfordi. His Spathiocaris plicifera , holotype USNM 112030, differs from the rest only in preservation, bearing secondary corrugated folds. These Woodford Shale specimens are indistinguishable from Sidetes newberryi , and belong within this taxon. Ruedemann (1916) also described another species from the Cleveland Shale along Mill Creek in Newburg, Ohio (Cleveland), Spathiocaris williamsi, which may represent another folded specimen with a hinge more strongly curved than S. newberryi. He described S. williamsi as being asymmetrical and having the apex or rostrum displaced to one side of the “median line.” This asymmetry is suspect. If we consider the lateral margin of his specimens nearest the rostrum to be the folded hinge line of a compressed specimen, the half which remains visible strongly resembles S. newberryi. Possible differentiating features might be coarser ornamentation (about 10/cm) and a superimposed concentric furrowing with a spacing of about 3-4 mm. Insofar as Ruedemann’s types have not been located for study, it seems prudent to consider the two taxa as separate. Sidetes ulrichi (Cooper, 1932). Figures 10.1 - 10.5 Spathiocaris ulrichi Cooper, 1932, p. 352. Diagnosis Structure sub-elliptical, narrower at rostrum. Anterior margin acutely convex. Broadly convex, highest point 1991 Late Devonian Cephalopod Aptychi 65 0.0 2.0 4.0 6.0 8.0 10.0 Energy [KeV] 2 Concauicaris sp., Cleveland Shale 0.0 2.0 4.0 6.0 8.0 10.0 Energy [KeV] 3 CMNH 8317, Cleveland Shale 0.0 2.0 4.0 6.0 8.0 10.0 Energy [KeV] FIGURE 12. X-ray emission spectra revealing relative composition of some Cleveland Shale specimens. 1, a lingulid brachiopod. 2, Concavicaris, a crustacean, CMNH 3740. 3, CMNH 8317, a presumed Devonian anaptychus. posterior of rostrum. Superimposed upon fine, concentric ornamentation are broad undulations parallel to ornamentation. 0.0 2.0 4.0 6.0 8.0 10.0 Energy [KeV] 0.0 2.0 4.0 6.0 8.0 10.0 Energy [KeV] FIGURE 13. X-ray emission spectra revealing relative composition of 1, an unidentified cephalopod from the Cleveland Shale, CMNH 8705 2, the wing or collar region of the jaw of modern Nautilus. Description Outline nearly elliptical, narrower at rostrum, length 10 to 60 mm, width about two-thirds of length. Surface corrugated in broad concentric undulations subparallel with posterior margin and intersecting lateral margins, spaced about 3 mm apart in a 30 mm specimen. Ornamentation similarly oriented, much more finely spaced, 30-60/cm. Type Cooper’s holotype of Spathiocaris ulrichi was loaned to him by Dr. George H. Girty of the U.S. Geological Survey. It was collected by E.O. Ulrich from the Woodford Formation (Late Devonian) near Dougherty, Oklahoma. It is now at the U.S. National Museum, USNM 112036. The type is 60 mm long and 44 mm wide, somewhat larger than those from the Ohio Shale. 66 Frye and Feldmann No. 46 Material Specimens studied include the holotype; CMNH 8303 and CMNH 8305b, from Big Creek; CMNH 3744 from Chance Creek; and CMNH 8328a and CMNH 8328c, from an unknown locality. All the Ohio specimens are from the Cleveland Shale. Remarks Sidetes ulrichi is the single species examined which bears a convex anterior margin. All others have straight or indented margins anteriorly. The dimensions of the specimens studied range from a width-to-length ratio of 0.6 to 0.8; all but the most compacted and flattened exhibit the corrugations superimposed upon the finer concentric ridges. The flattest are assigned to this species primarily on outline and fineness of ornamentation. Microstructure and Composition Several of the better preserved specimens, interpreted to be aptychi, were prepared for examination with the scanning electron microscope. The calcareous portions of cephalopod aptychi have a distinctive internal microstructure (Lehmann, 1981). This structure, if identifiable in the fossil specimens, would confirm their identification. The microstructure of authentic arthropod cuticle, and brachiopod and bivalve shells was examined for comparison with the study specimens. Fish scales were eliminated from consideration as those in The Cleveland Museum of Natural History collection from the Late Devonian have morphologies distinctly different from these aptychi (M. Williams, CMNH, personal communication). An International Scientific Instruments Model SX-40A SEM was used, with an attached Princeton Gamma Tech System 4 Plus energy dispersive x-ray spectrometer. It is routine practice in electron microscopy to coat the surface of the specimen with a conductive material to drain the accumulation of electric charge built up by electron bombardment. Several of the first specimens examined were coated with gold to a thickness of approximately 500 angstroms, using ISI’s P-Sl diode sputter coater. This procedure interfered with the use of the x-ray spectrometer, however. The K-alpha emission line for phosphorous has an energy of 2.014 KeV, while gold has an M-alpha emission line at 2. 1 23 KeV, too close to the phosphorous line to be resolved (Goldstein, et al., 1981). A commonly used alternative is carbon coating, since the emission spectrum of carbon is entirely absorbed by the beryllium window in the detector apparatus. In the absence of a carbon coater, uncoated specimens were examined. This was successful, perhaps due to the high (4.60 ± 1 .09 wt.%) average organic carbon content of the Cleveland Shale (Broadhead, et al., 1982), and the carbonaceous nature of the specimens themselves. Examination of several aptychus specimens revealed no discernable structure remaining within the thin carbonaceous film (Figure 11.1). All thicker regions examined were indistinguishable from the shale matrix and appeared to be molds. It seems that this method of investigation is of little value with material reduced to a carbonaceous film under anaerobic or dysaerobic preservational regimes. The instrument was used to search for elements that might allow identification of the material composition of the aptychi. Qualitative energy-dispersive X-ray (EDX) spectra were obtained from aptychi and similar appearing fossils and parts from living organisms. The X-ray energy range from 0 to 20 thousand electron volts was scanned, allowing for the detection of nearly all the elements. Only those with atomic number less than berylium were undetectable, as the detector apparatus blocks X-rays from these elements. From the spectra obtained, the relative concentrations of calcium, strontium, and potassium within the samples were used to test for the calcitic, aragonitic, or phosphatic nature of the fossils. Other elements discovered were identified and indicated in the figures. The spectra were compared to test the usefulness of the method in resolving this question. Inarticulate brachiopods and certain arthropods have long been considered the most likely alternative taxa to which aptychus-like structures might be assigned (Clarke, 1902). Therefore, specimens of a lingulid brachiopod (CMNH uncataloged) and the crustacean Concavicaris (CMNH 3740) from the Cleveland Shale were prepared for SEM and EDX examination. Analyses of these specimens and a representative anaptychus, CMNH 8317, are shown in Figure 12. While phosphorous is clearly present in both the phosphatic brachiopod and the crustacean, it is just as clearly absent from the anaptychus. Also conspicuously absent from the anaptychus is calcium or strontium, the latter being a common marker impurity used to identify aragonite (Crick, et al., 1987). It may be impossible to positively identify originally calcareous material from these units, however. Cephalopods tested from the Cleveland Shale, for example the one shown in Figure 13.1, were found to be significantly replaced with pyrite, which can be expected in the metal-rich, anaerobic conditions of deposition (Baird and Brett, 1986). Strontium in particular, present originally in trace amounts, may be undetectable in these altered specimens. The lack of pyrite replacement, common to many shelly fossils in the Cleveland, may suggest that the anaptychi were entirely chitinous in nature, with no mineralized portions. Modem Nautilus mandibles contain phosphorous, but in trace quantities. While the exposed, oral portions of the mandibles are mineralized, the muscle insertion areas are often only lightly calcified if at all, consisting largely of a chitin/protein complex, only thinly coated with aragonite. 1991 Late Devonian Cephalopod Aptychi 67 FIGURE 14. Aptychopsis Barrande, 1872. 1, LO 5270 and 2, LO 5268, from a Silurian shale quarry’ in southern Sweden (Stridsberg, 1984). Specimens from the Department of Historical Geology and Palaeontology, University of Lund, Sweden. 3, Reconstruction of opercular position of aptychopsid plates in an ortlwconic nautiloid, after Turek, 1978. This aragonitic layer contains small deposits of brushite, a phosphatic mineral (Lowenstam, et al., 1984; Lowenstam and Weiner, 1989). Total phosphorous content of this posterior region is on the order of 0.30%. This posterior region of the Nautilus mandible also reveals little internal microstructure beyond subtle layering, even with SEM examination (Figure 11.2). What structure is seen seems to be an artifact of breakage. The thinness of the jaw specimen allowed for electron beam penetration through it and into the aluminum mounting stub when in the SEM. The X-ray analysis in Figure 13.2 reveals a strong aluminum peak for this reason. The other elements present in the analysis are common trace elements in sea water, and appear to have been incorporated into the structure in significant amounts. Perhaps diagenetic alteration resulted in the depletion of chlorine and enrichment in iron seen in the Cleveland Shale aptychi. The phosphate minerals were shown by Lowenstam, et al. (1984) to be limited to the carbonate layers, which are not present in the Devonian material. Thus, the x-ray analysis presents evidence denying an arthropod or brachiopod affinity for these fossils. While there are other possible origins for carbonaceous fossil fragments, two of the most likely alternatives based upon the morphology of the fossils are eliminated from consideration. The most parsimonious interpretation is that these are indeed cephalopod aptychi, as suspected by Clarke ( 1902), Girty (1909), and particularly Ruedemann (1916). Aspects of Functional Morphology Historically, when cephalopods were first found with aptychi in place, the approximate match between their outline and the aperture of the conch suggested they served as opercula (Woodward, 1885b; Clarke, 1902; Trauth, 1927). This correspondence is by no means exact, however (Lehmann, 1972, p. 42). Other functions postulated for these structures were as covers for the nidamentary glands or as cartilaginous plates for funnel muscle attachment. Ruedemann (1916, p. 102) suggested the latter, “would naturally also have existed in the Ordovician and Silurian cephalopods...” in attempting to explain the Discinocarida. This prescient speculation was proven sound by the discovery of specimens of Aptychopsis Barrande, 1872 in situ in the apertures of orthoconic nautiloids from the Silurian of central Bohemia (Turek, 1978) and southern 68 Frye and Feldmann No. 46 FIGURE 15. Hypothesized original shapes of aptychi. 1, Reconstruction of a Cleveland Shale anaptychus. 2, Reconstruction of the mandibles of Psiloceras, Hyatt 1867 ( from Lehmann 1971.) Sweden (Holland, et al., 1978). Aptychopsid plates form a neat, flat circular structure almost precisely fitting the aperture of the nautiloid (Stridsberg, 1984). The three plates involved would be difficult to fold into a concave structure (Figure 14.1 ). It would seem that these are indeed opercula. The structures known as aptychi in Mesozoic ammonites seem to have served a different function. Again, discoveries of aptychi in situ in cephalopod body chambers provided evidence of their function. These were found to be curved structures, sometimes found associated with an element resembling the upper jaw of Nautilus (Lehmann, 1971, 1972, 1978, 1981; Tanabe, 1983; Tanabe and Fukuda, 1987). It has become accepted by many that these aptychi are the lower jaws of ammonites (Lehmann, 1981; Morton, 1981; Mapes, 1987) . One difficulty with this interpretation is the sheer size of the structure, relative to both the upper jaw components and the diameter of the body chamber. In many cases, the aptychus approximates the cross-sectional area of the body chamber. Lehmann’s work, cited above, involved serial sectioning of cephalopods, revealing the three-dimensional relationship of the body chamber and its contents, and seems to be the most accurate means of determining the nature of these structures, at least for the Mesozoic ammonites studied. There are no rigid calcareous plates found with the Devonian fossils, so their preservation as flattened bodies does not necessarily reflect their original morphology. Two specimens of Sidetes newberryi, CMNH 8311 and CMNH 8318, exhibit short cracks radiating inward from the posterolateral margin (Figures 9.1, 9.2). These cracks are precisely what would be expected in a convex structure which has been flattened. If this is the origin of the cracks, then closing the gaps should approximate the original form of the structure. As these specimens were folded in half along the medial line, photographs of them were prepared in both normal and reversed orientations. These were then photocopied and enlarged. The resulting reproductions were then cut out and the mirror images were attached along the midline. This caused the paper models to become slightly convex. The cracks along the margin were then cut away and the edges rejoined. This caused the models to assume a broad, scooped shape, similar to that seen in Lehmann's (1970) reconstruction of the jaws of Psiloceras from the Jurassic of Germany (Figure 15). Many of the specimens preserved as flattened bodies are folded along the median. In some cases they can be separated from the matrix and studied from both sides, or often the remains of the upper half are preserved along the margins of the extant lower portion (see the arrow in Figure 4.2, S. emersoni). These reconstructions of ammonite jaws (Lehmann, 1970; 1975), suggest to some a scooping, shoveling application rather than a true biting action seen in modem Nautilus (Lehmann, 1972), and in other Mesozoic forms (Nixon, 1988). This conclusion is based on the large size of the lower jaws (particularly anaptychi), both relative to the size of the upper jaw and to the total size of the conch. The lack of calcified rostra, or conchorhynchs, to serve as biting surfaces further serves as a basis of interpretation. Calcified beaks have been reported from the Permian (Closs, 1967) and perhaps from the Mississippian (Landman and Davis, 1988) , but no older specimens have been found. Where associated upper jaws are of comparable scale, however, it would seem that, even lacking calcified beaks, the jaws could be capable of strong biting action similar to modem coleoids, which lack such calcified surfaces (Tanabe, et al.. 1991 Late Devonian Cephalopod Aptychi 69 1980). The shape of the lower jaw is similar for both applications. Still, the curved shape revealed in the Ohio Shale specimens by these methods, if genuine, fits known jaw structures far better than hypothesized opercula or Aptychopsis. We must conclude that these forms of Sidetes from the Late Devonian represent cephalopod jaws which have been preserved separately from the animal’s conch. Summary and Conclusion Approximately 120 specimens of Hat, carbonaceous body fossils have been collected from the Cleveland and Chagrin shales of northeastern Ohio. The fossils have been found predominantly in the black Cleveland Shale, which probably represents an anoxic basinal environment (Broadhead, et ah, 1982). A few have been collected from the underlying Chagrin Shale, a gray-green unit deposited in dysaerobic conditions (Barron and Ettensohn, 1981; Schwimmer, 1988, Schwimmer and Feldmann, 1990). In addition, specimens housed at the National Museum of Natural History which were collected by Cooper in 1932 have been reevaluated. These fossils have been identified as representing seven species of Sidetes Giebel, 1847, and interpreted to be cephalopod jaw elements. This identification is based upon the general outline of the specimens, and particularly the pattern of ornamentation they exhibit. This concentric pattern of fine ridges is unlike that seen in arthropods, gastropods, bivalves, or brachiopods. Some of the specimens are preserved unfolded and flattened, while others are folded in half along the median line. Further evidence for this interpretation is furnished by energy dispersive X-ray spectrometry. These fossils show no trace of having once contained phosphorous within the structures studied. It is unlikely that alteration after burial would have so completely removed the element, as both arthropods and inarticulate brachiopods from the same unit have remained phosphatic. On this basis, we can conclude that the Sidetes specimens in this study are neither arthropods nor brachiopods. Two specimens showed signs of compaction damage. When restored to their presumed original shape, they resemble reconstructions of Mesozoic ammonite jaws. It appears that these fossils served the same function in some Devonian cephalopods. By contrast, the Silurian Aptychopsis may have been a nautiloid operculum rather than a jaw element. Acknowledgements Ellis Yochelson, of the Department of Paleobiology, U.S. National Museum of Natural History, Washington, D.C., helped launch this study and offered helpful comments on the morphology of aptychi. His review of a previous version of this manuscript was quite helpful. While our conclusions differ in some particulars, his comments were most constructive. Roy Mapes, Ohio University, also helped to considerably improve the manuscript. The Palaeontological Institute of the University of Lund loaned the Aptychopsis specimens for comparison. Jann Thompson of the National Museum of Natural History located one of Clarke’s holotypes and those from Cooper's Oklahoma work, and was of much assistance in our examination of them. Melvin Hinkley provided information and a loan of Whitfield’s type specimens from the American Museum of Natural History, New York. Mary Baum and Wendy Weitzner, of The Cleveland Museum of Natural History, helped uncover the more obscure references, and Shya Chitaley and JoAnn Cobum, CMNH, searched their collections for a few stray aptychi. Joe Hannibal, CMNH, offered tireless assistance and made possible the loan of the CMNH specimens studied. Thanks are also due to Richard E. Carroll of Michigan State University for collecting one of the specimens in situ and donating it to The Cleveland Museum of Natural History for our study and to Donald F. Palmer and Alan H. Coogan who read an earlier draft of the manuscript. Contribution 455. Department of Geology, Kent State University, Kent, Ohio 44242. References Arkell. W.J. 1954. Proposed declaration that a generic or specific name based solely upon the "aptychus" of an ammonite be excluded from availability under the Regies. Bulletin of Zoologic Nomenclature 9(9):266-269. Arkell. W.J. 1957a. Aptychi, pp. L437-L441. In: Treatise on Invertebrate Paleontology , Volume L, Mollusca 4, Cephalopoda, Ammonoidea. Edited by R.C. Moore. Geological Society of America and University of Kansas Press, Lawrence. Arkell, W.J. 1957b. The Moore/Syvester-Bradley "parataxa" proposals. Bulletin of Zoologic Nomenclature 15:78. Baird, G.C., and C.E. Brett. 1986. Erosion on an anaerobic seafloor: significance of reworked pyrite deposits from the Devonian of New York State Palaeogeography, Palaeoclimatology, Palaeoecology 57:157-193. Barron, L.S., and F.R. Ettensohn. 1981. Paleoecology of the Devonian-Mississippian Black-shale Sequence in Eastern Kentucky With an Atlas of Some Common Fossils. U.S. Department of Energy, Morgantown Energy Technology Center, DOE/METC/ 12040- 151, 75 pp. Brady, L.F. 1955. Possible nautiloid mandibles from the Permian of Arizona. Journal of Paleontology 29: 1 02- 1 04. Broadhead. R.F., R.C. Kepferle, and PE. Potter. 1982. Stratigraphic and sedimentologic controls of gas in shale — example from the Upper Devonian of Northern Ohio. American Association of Petroleum Geologists Bulletin 66:10-27. Butts, C. 1942 (1941). Geology of the Appalachian Valley in Virginia, Pt. 2, Fossil plates and descriptions. Bulletin 52, Virginia Geologic Survey. 27 1 pp. 70 Frye and Feldmann No. 46 Campbell, G. 1946. New Albany Shale. Geological Society of America Bulletin 57:829-890. Clarke, J.M. 1882. New phyllopod crustaceans from the Devonian of Western New York. American Journal of Science , 3rd Series 23:477-478. Clarke, J. 1902. A new genus of Paleozoic brachiopods, Eunoa , with some considerations therefrom on the organic bodies known as Discinocaris, Spathiocaris, and Cardiocaris. Bulletin of the New York State Museum 52:606-615. Clarke, M.R. 1962. The identification of cephalopod “beaks” and the relationship between beak size and total body weight. Bulletin of the British Museum (Natural History ), Zoology 8:419-480. Closs, D., M. Gordon, Jr., and E.L. Yochelson. 1964. Cornaptychi from the Permian of Utah. Journal of Paleontology 38:899- 903. Cooper, C.L. 1932. A crustacean fauna from the Woodford Formation of Oklahoma. Journal of Paleontology 6:346-352. Copeland, M.J., and T.E. Bolton. I960. Canadian fossil Arthropoda, Eurypterida, Phyllocarida, and Decapoda. Geological Survey of Canada Bulletin 60:35-38. Crick, R.E., K.O. Mann, and J.A. Chamberlain, Jr. 1987. Skeletal chemistry of Nautilus and its taxonomic significance. Biochemical Systematic :s and Ecology 15:461-474. Giebel, C.G. 1849. Briefliche Mittheilungen an Herrn Beyrich. Zeitschrift der Deutschen Geologischen Gesellschaft 1 ( 2):99- 1 00. Girty, G.H. 1909. The Fauna of the Caney Shale of Oklahoma. Bulletin of the United States Geological Survey 377 . 106 pp. Goldstein, .1.1. , D.E. Newbury, P. Echlin, D.C. Joy, C. Fiori, and E. Lifshin. 1981. Scanning Electron Microscopy and X-ray Microanalysis. Plenum Press, New York, 673 pp. Hannibal, J.T., and R.M. Feldmann. 1983. Arthropod trace fossils, interpreted as echinocarid escape burrows, from the Chagrin Shale (Late Devonian) of Ohio. Journal of Paleontology 57(4):705-7 16. Harper, J.A. 1989. Occurrence of a cephalopod aptychus in the Ames Limestone (Virgilian) of western Pennsylvania. Journal of Paleontology 63( 1 ): 1 1 9- 1 2 1 . Hasenmueller, N.R., and R.K. Leininger. 1987. Oil-Shale Prospects for the New Albany Shale in Indiana. Special Report of the Indiana Geological Survey, Dept, of Natural Resources, Bloomington, 40:31 pp. Hlavin, W.J. 1976. Biostratigraphy of the Late Devonian Black Shales on the Cratonal Margin of the Applachian Geosyncline. Unpublished Ph.D. Dissertation, Boston University. 194 pp. Holland, B., S. Stridsberg, and J. Bergstrom. 1978. Confirmation of the reconstruction of Aptychopsis. Lethaia 1 1 : 144. Internationa] Commission on Zoological Nomenclature. 1985. International Code of Zoological Nomenclature , Third Edition. LIniversity of California Press, Berkeley. 338 pp. Kanie, Y. 1982. Cretaceous tetragonitid ammonite jaws: A comparison with modern Nautilus jaws. Transactions and Proceedings of the Palaeontological Society of Japan. N.S. 125:239-258. pis. 39, 40. Kayser, E. 1882. Protokoll der Decembersitzung der Deutschen Geologischen Gesellschaft, Vortrag iiber Anaptychen. Zeitschrift der Deutschen Geologischen Gesellschaft 34:818-819. Kues, B.S. 1983. Cephalopod aptychi from Los Moyos Limestone, Madera Group (M. Pennsylvanian), near Albuquerque, New Mexico. New Mexico Geology 5:78-80. Landman, N.H., and R.A. Davis. 1988. Jaw and crop preserved in an orthoconic nautiloid cephalopod from the Bear Gulch Limestone (Mississippian, Montana). Memoir of the New Mexico Bureau of Mines and Mineral Resources, 44:103-107. Lehmann, U. 1970. Lias Anaptychen als Kieferelemente (Ammonoidea). Paldontologische Zeitschrift 44:25-31. Lehmann, U. 1971. New aspects in ammonite biology. Proceedings of the North American Paleontological Convention 1:1251-1269. Lehmann, U. 1972. Aptychen als Kieferelemente der Ammoniten. Paldontologische Zeitschrift 46:34-48. Lehmann, U. 1975. Uber Nahrung und Ernahrungsweise von Ammoniten. Paldontologische Zeitschrift 49: 187-195. Lehmann, U. 1978. Uber den Kieferapparat von Ammoniten der Gattung Parkinsonia. Mitteilungen aus dem Geologisch- Paldontologischen Institut der Universitat Hamburg 48:79- 84, pi. 9. Lehmann, U. 1979. The jaws and radula of the Jurassic ammonite Dactylioceras. Palaeontology 22( 1 ):265-27 1 , pi. 27. Lehmann, U. 1981. Ammonite jaw apparatus and soft parts. In: The Ammonoidea. Edited by M.R. House and J.R. Senior. Systematics Association Special Volume. Academic Press, New York 18:275-287. Lewis, T.L. 1988. Late Devonian and Early Mississippian Distal Basin-Margin Sedimentation of Northern Ohio. Ohio Journal of Science 88(1 ):23-39. Lineback, J.A. 1970. Stratigraphy of the New Albany Shale in Indiana. Bulletin of the Indiana Geological Survey, Dept, of Natural Resources, Bloomington, 44:67 pp. Lowenstam, H.A., W. Traub, and S. Weiner. 1984. Nautilus hard parts: a study of the mineral and organic constituents. Paleobiology 10(2):268-279. Mapes, R.H. 1987. Upper Paleozoic cephalopod mandibles: Frequency of occurrence, modes of preservation, and paleoecological implications. Journal of Paleontology 61:521-538." Matern, H. 1931. Oberdevonische Anaptychen in situ und iiber die Erhaltung von Chitin-Substanzen. Senckenbergiana 13:160-167. Moore, R.C., and PC. Sylvester-Bradley. 1957a. Taxonomy and nomenclature of aptychi, pp. L465-L471. In: Treatise on Invertebrate Paleontology , Volume L, Mollusca 4, Cephalopoda, Ammonoidea. Edited by R.C. Moore. Geological Society of America and University of Kansas Press, Lawrence. Moore, R.C., and P.C. Sylvester-Bradley. 1957b. Second supplemental application: Application for a ruling by the International Commission directing that the classification and nomenclature of ammonoid aptychi (Class Cephalopoda) be in terms of “parataxa.” Bulletin of Zoological Nomenclature 15:35-70. 1991 Late Devonian Cephalopod Aptychi 71 Morton, N. 1981 Aptychi: the myth of the ammonite operculum. Lethaia 14:57-61. Nixon, M. 1988. The buccal mass of fossil and recent Cephalopoda, pp. 103-121. In: The Mollusca , Volume 12: Paleontology and Neontology of Cephalopods. Edited by K.M. Wilber. Academic Press, San Diego. Pratt, B.R. 1987. The phyllocarid crustacean Montecaris from the Devonian of Western Canada. Canadian Journal of Earth Sciences 24:1267-1269. Prosser, C.S. 1912. The Devonian and Mississippian formations of Northeastern Ohio. Bulletin of the Ohio Geological Survey, Fourth Series, Columbus, Ohio. 15:574 pp. Ruedemann, R. 1916. Account of some new or little-known species of fossils, mostly from the Paleozoic of New York. Bulletin, New York State Museum, 1 89:7- 1 1 2. Ruedemann, R. 1934. Paleozoic Plankton of North America. Geological Society of America Memoir, 2:57-6 1 . Saunders, W.B., and E.S. Richardson, Jr. 1979. Middle Pennsylvannian (Desmoinesean) cephalopoda of the Mazon Creek fauna. Northeastern Illinois, pp. 333-359. In: M.H. Nitecki (ed. ). Mazon Creek Fossils. Academic Press, New York. Saunders, W.B., and C. Spinosa. 1974. Unusual fossil cephalopod jaws from Nevada ( abst .). Abstracts with Programs, Cordilleran Section Annual Meeting, Geological Society of America. 6(3):248. Schwimmer, B.A. 1988. Stratigraphic Distribution of Brachiopods and Pelecypods in the Upper Devonian (Famennian) Chagrin Shale in the Cuyahoga River Valley, Northeast Ohio. Unpublished Masters Thesis, Kent State University. 134 pp. Schwimmer, B.A., and R. M. Feldmann. 1990. Stratigraphic Distribution of Brachiopods and Bivalves in the Upper Devonian (Famennian) Chagrin Shale in the Cuyahoga River Valley, Northeast Ohio. Kirtlandia 45:7-31. Stridsberg, S. 1984. Aptychopsid plates: Jaw elements or protective operculum? Lethaia 17:93-98. Szmuc, E.J. 1970a. The Devonian System, pp. 9-21. In: Guide to the Geology of Northeastern Ohio. Edited by PO. Banks and R.M. Feldmann. Northern Ohio Geological Society. Szmuc. E.J. 1970b. The Mississippian System, pp. 23-67. In: Guide to the Geology of Northeastern Ohio. Edited by PO. Banks and R.M. Feldmann. Northern Ohio Geological Society. Tanabe, K. 1983. The jaw apparatuses of Cretaceous desmoceratid ammonites. Palaeontology 26(3):677-686. Tanabe, K., H. Hirano, and Y. Kanie. 1980. The jaw apparatus of Scalarites mihoensis, a Late Cretaceous ammonite, pp. 1 59- 165, pi. 20. Professor Saburo Kanno Memorial Volume. Tanabe, K., and Y. Fukuda. 1987. The jaw apparatus of the Cretaceous ammonite Reesidites. Lethaia 20:41-48. Thompson, E.H., E.L. Yochelson, and R.H. Flower. 1980. Aptychi from the Pennsylvanian of West Texas. Journal of Paleontology 54:903-909. Trauth, F. 1927. Aptychenstudien I: Uber die Aptychen im Allgemeinen. Annalen des Naturhistorische Museums in Wien 41:171-259, 1 pi. Trauth, F. 1928. Aptychenstudien II: die Aptychen der Oberkreide. Annalen des Naturhistorische Museums in Wien 42:121-193, pis. 2-4. Trauth. F. 1930. Aptychenstudien 1 1 1 - V : Nachtrag zu den “Aptychen im Allgemeinen;” Nachtrag zu den “Aptychen der Oberkreide;” Die Aptychen des Dogger. Annalen des Naturhistorischen Museums in Wien 44:329-41 1, pis. 3-5. Trauth, F. 1931. Aptychenstudien VI- VII: Zweiter Nachtrag zu den “Aptychen im Allgemeinen;” die Aptychen des Malm und Unterkreide. Annalen des Naturhistorischen Museums in Wien 45: 17-136, I pi. Trauth, F. 1935. Die Aptychen des Palaozoikums. Jahrbuch Preussische Geologische Landesanstalt 55( 1 934 ):44-83, pi. 1-12. Trauth, F. 1936. Aptychenstudien VIII: die Laevillamellaptychi des Oberjura und der Unterkreide. Annalen des Naturhistorischen Museums in Wien 47: 127- 145, pi. 3. Turek, V. 1978. Biological and stratigraphical significance of the Silurian nautiloid Aptychopsis. Lethaia 1 1 : 127- 1 38. Whitfield, R.P 1882. Species from the Huron shales. Annals of the New York Academy of Sciences 2:217. Woodward, H. 1865. On the discovery of a new genus of Cirripedia in the Wenlock Limestone and Shale of Dudley. Quarterly Journal of the Geological Society of London, 21(8 0:486-489, pi. 14. Woodward, H. 1882. On a series of phyllopod crustacean shields from the Upper Devonian of the Eifel. The Geological Magazine, new series. Decade 2, 9:385-390, pi. 9. Woodward, H. 1885a. On a new species of Helminthochiton from the Upper Bala (Silurian) of Girvan, Ayrshire. The Geological Magazine, new series. Decade 3, 2:352-355. Woodward, H. 1885b. On some Paleozoic phyllopod-shields, and on Nebalia and its allies. The Geological Magazine, new series. Decade 3, 2:345-352, pi. 9. Yochelson, E.L. 1971. The Permian nautiloid mandible Rhynchoteuthis kaibabensis reinterpreted as a polyplacophoran. Journal of Paleontology 45:130-133. Yochelson, E.L. 1983. A Devonian aptychus (Cephalopoda) from Alabama. Journal of Paleontology 57: 1 24- 1 27. ■NATURAL HISTORY' Published by The Cleveland Museum of Natural History 1 Wade Oval Drive, University Circle Cleveland, Ohio 44106-1767 Production by Mary Szelagowski and Wendy Ryan 'ELAND, OHIO NUMBER 47 KIRTLANDIA Archaeology Late Woodland Fortifications in Northern Ohio: The Greenwood Village Site 3 Stephanie J. Belovich and David S. Brose •NATURAL HISTORY* KIRTLANDIA The Scientific Publication of The Cleveland Museum of Natural History David S. Brose and Joseph T. Hannibal, Editors Brief History and Purpose Kirtlandia, a publication of The Cleveland Museum of Natural History, is named in honor of Jared Potter Kirtland, a noted nineteenth-century naturalist who lived in the Cleveland, Ohio area. It began publication in 1967 and is a continuation of the earlier series Scientific Publications volumes 1 to 10 (1928-1950), and new series volumes I to 4 (1962-1965). Supported by the Kirtlandia Society of The Cleveland Museum of Natural History, Kirtlandia is devoted to the publication of scientific papers in the various fields of inquiry within the Museum’s sphere of interest: Cultural and Physical Anthropology; Archaeology; Botany; Geology; Paleobotany; Invertebrate and Vertebrate Paleontology; Systematic Ecology; and Invertebrate and Vertebrate Zoology. Issues will vary from single monographs to collections of short papers, review articles, and brief research notes. Kirtlandia is abstracted in Biological Abstracts and indexed in Bibliography and Index of Geology and Zoological Record. Associate Editors James K. Bissell, The Cleveland Museum of Natural History Bruce Latimer, The Cleveland Museum of Natural History Martin Rosenberg, Case Western Reserve University Sonja Teraguehi, The Cleveland Museum of Natural History Editorial Advisory Board Rodney Feldmann, Kent State University Michael C. Hansen, Ohio Geological Survey Richard Meindl. Kent State University G. Michael Pratt, Heidelberg University David H. Stansbery, Ohio State University Frederick H. Utech, Carnegie Museum of Natural History Ed Voss, University of Michigan Andrew M. White, John Carroll University Kirtlandia No. 47 ISSN 0075-6245 © 1992 by The Cleveland Museum of Natural History Cleveland, Ohio Copies of Kirtlandia, and many issues of the Scientific Publications series of The Cleveland Museum of Natural History, are available for sale. Write to: Library, The Cleveland Museum of Natural History, 1 Wade Oval Drive, Cleveland, Ohio 44106-1767 for a current price list. KIRTLANDIA The Cleveland Museum of Natural History September 1992 Number 47:3-23 Late Woodland Fortifications in Northern Ohio: The Greenwood Village Site Stephanie J. Belovich and David S. Brose Department of Archaeology The Cleveland Museum of Natural History l Wade Oval Drive, University Circle Cleveland, Ohio 44106-1767 Abstract Since the late nineteenth century all of the archaeological sites located on steep bluffs overlooking the major rivers of northern Ohio have been considered fortified villages. The belief that such sites were defensive was given support by the results of some excavations which also demonstrated a Late Woodland temporal placement. Recent excavation at one such site, the Greenwood Village site, demonstrated that its still visible earthworks are not defensive but used construction techniques similar to those at Southern Ohio Middle Woodland ceremonial earthworks. Carbon- 14 and thermoluminescence dates place the earthworks’ construction and ceremonial use of the Greenwood Village site plateau between A.D. 460 and A.D. 1040. Critical review of several similar hilltop enclosures shows many are neither defensive works, nor are they all from late prehistoric periods. Introduction Since initial documentation, the presence of prehistoric villages and accompanying fortifications have come to be the hallmark of the Whittlesey Tradition in northern Ohio. Recorded by Colonel Charles Whittlesey in the nineteenth century, these sites, high atop steeply sided bluffs overlooking major river valleys and their tributary streams, are characterized by the presence of earthen walls and ditches. During the years since Whittlesey’s survey many of these sites have been destroyed through urban expansion, vandalism or both. Excavation on some Whittlesey Tradition sites (Greenman, 1935a, 1935b, 1937; Morgan and Ellis, 1943; Murphy, 1971a) demonstrated similarities in topographic location and ceramic assemblage, but failed to recover data sufficient to discuss spatial organization, subsistence patterning, or resource procurement and scheduling. With little evidence these sites were assumed to represent large fortified villages whose inhabitants focused on maize agriculture. It was not until 1969, when Brose conducted extensive excavations at the Whittlesey Tradition South Park site (33Cu8), that a detailed picture of the tradition emerged. Three stratigraphically distinct occupations dating from A.D. 1000 to A.D. 1640 were identified at South Park (Brose, 1973). Evidence for increasing reliance upon maize-bean-squash agriculture and 4 Belov ich and Brose No. 47 eventual year-round village occupation was found. Information on domestic structures, site patterning, ditch and earthen fortifications and seasonal scheduling was also obtained (Brose, 1992). In essence Brose confirmed the validity of the characterization of the Whittlesey Tradition. Before Brose's detailed contribution to understanding the Whittlesey Tradition, the archaeological community had begun to take for granted the defining characteristics of that tradition. Essentially, all of the sites located by Colonel Whittlesey, and for that matter any northern Ohio site on a steeply sided bluff with ditches, embankments or other earthworks which could be considered fortifications, were uncritically assigned to the Whittlesey Tradition. The issue we address in this paper then, is whether such assignments are valid. In so doing we will report on recent excavations con- ducted by Belovich and Brose (1983) and Belovich (T985a, 1985b) at the Greenwood Village site (33Su92: Whittlesey Fort No. 5). We also will discuss the development of Middle Woodland earthworks into Late Woodland fortifications. The Environment The Greenwood Village site is situated on the western edge of the glaciated Appalachian Plateau Province. This area was covered by ice during several Pleistocene stages, most recently during the Wisconsinan (Belovich, 1985b; White and Totten, 1982). After the formation and draining of a series of pro-glacial lakes, the middle-lower Cuyahoga River flowed north into Lake Erie. The subsequent formation of numerous tributary streams caused the eastern and southern highlands along the river valley to become extensively dissected by small seasonal streams (Belovich, 1985b; Brose et al., 1981; Williams, 1949). Archaeological sites located within the Cuyahoga River Valley are often encountered on the summits of these steeply sided plateaux. Severe erosion has undoubtedly reduced site sizes (Belovich, 1985b). The soils of the uplands are derived from underlying sandstones and shales, or are of glaciofluvia! origin (Williams, 1949; U.S. Department of Agriculture, 1971). The narrow floodplain of the Cuyahoga River was built by successive layering of flood-water sediments. These soils are highly fertile, and the floodplain has always been considered prime agricultural land (Brose et ah, 1981, p. 9). Since 8,000 B.C. the Cuyahoga River Valley has been covered by a nearly continuous deciduous forest canopy. Within this mixed mesophytic forest, several specific floral associations can be identified (Braun, 1950; Gordon, 1969; Williams, 1949). Along the uplands farthest from the river, soils are dry and oak-hickory (and at one time, chestnut) associations predominated. On the slopes, lower elevations and moister areas, beech-maple associations are found. Elm-ash forests inhabit the bottomlands, while com- munities of hemlock and pines thrive along the cool, damp f.ortr/rtnt o/ //tf tt/frirnf rt/rf/ ttwr/'A o/ /Ar ft//* ft// />,/ OZf/ff .. " S s ;, >ttn t. f /nr ftt, D // r////f/&tf \‘r,/ //// ,/ + o \r«/ • + /// ///// -' n'f /■•/ ,V".i -f D /Jrrr/sf f//t \ + , f { Sorf/t/it/t/ V □ V. ,iv '/■ + < 0 V X»/f. □ , Y"/Z +• /ffffU/f f tifff/t □ /httr/stfZf '± y»& Yf/ ft ' O if « vy < 0 ' yoy Yf/f/A \ I/tini/i/on 0M Pttrtoffr d timihgtu. d ■ , firm l/r/w/a f's////r//r/// a// ,v . - - /'V//V// monarf. «• f f ! ///• r r // / forts CO f't/r/tr sc O o //v v r /r > > r t/r/f/r s S rt/ tr ,)//// //' S/tr ///tf/ FIGURE 1. Map of Whittlesey forts located in northeast Ohio. Taken from Ancient Earth Forts of the Cuyahoga Valley, Ohio 1871. ravines of small streams (Williams, 1949). Detailed reconstructions of past faunal communities within the region are presently impossible. Early ethnohistoric accounts provide only general descriptions with little specific information. The soils of the uplands, the rich alluvial bottoms, and the temperate climate would have facilitated even technologically limited horticulture, while the mosaic forest patterns would have yielded a large variety of floral and faunal resources more than adequate to meet the subsistence needs of the aboriginal populations. 1992 Late Woodland Fortifications in Northern Ohio 5 Historical Background: 1847-1982 The Greenwood Village site (33Su92) was recorded in 1847 by Colonel Charles Whittlesey, then Ohio's Surveyor General (Whittlesey, 1850, 1867, 1871). This site was the fifth of eleven “forts” Whittlesey located and reported along both banks of the Cuyahoga River (Figure 1 ). All but one of these “forts” sit atop steeply sided and relatively isolated plateaux. Each site also has at least one earthen embankment and ditch crossing the single, usually narrow, level access to the site. At Greenwood Village (Fort No. 5) and Fort No.’s 1, 3 and 11, more than one ditch/ embankment were noted. Gateways and/or mounds were also recorded. Noting the topographic location, Whittlesey not only referred to these sites as “forts” but described each with reference to its potential defense against assault. The persistence of opinion that these sites were, in fact, defensive has been primarily based on their topographic location and single, narrow level access. Previous excavations at several of these sites (Tuttle Hill [Fort 3]: Greenman, 1937; and South Park: Brose, 1973, 1992) did reveal fortified late prehistoric villages, giving apparent confirmation to this idea. Greenwood Village (Fort No. 5) with its location and extensive earthworks was thought to be another late prehistoric fortified village. First mentioned in Descriptions of Ancient Works in Ohio (1850) Whittlesey called Greenwood Village a “fortification,” and described three wells, two mounds, one gateway, and five ditch/embankment lines. The relevant portions of the text read as follows: The engineers who selected the site of this fortification, understood very well the art of turning natural advantages to good account. Why they did not embrace in their plan the whole of the level space within the crest of the bluff, is not easily explained, unless we presume that their numbers were few, and not sufficient to defend the whole. On all sides, the gullies are from eighty to one hundred and ten feet deep, worn, by running water, into the blue and yellow hard pan that here forms the bluffs of the valley of the Cuyahoga River. The earth is as steep as it will stand; and, in fact, is subject to slides, that lie in terraces, resembling platforms, made by art. Before the ground was cultivated, the ditches are said by Milton Arthur, Esq., the owner of the land, to have been so deep that a man standing in them could not look over the wall. ... At the north end of the ditch of the inner wall, at the neck, there was a narrow space left as a passage into the work, but none in the outer wall. There are low mounds at m,m. The approach is along a sharp ridge called a “hog’s back,” merely broad FIGURE 2. 1850 Whittlesey map of Fort No. 5: the Greenwood Village site (33Si<92). Taken from Descriptions of Ancient Works in Ohio 1850. enough for a single road track, for the distance of thirty rods, and the sides are as steep as any part of the bluffs adjacent . . . It is not very evident why a few rods of ground were cut off by lines at the south-west angle, nor why part of the ditch was made on the inside on the north and west. It is very remarkable that, while all the works in northern Ohio are of a military character, there are no evidences of attacks by a foe, or of the destruction or overthrow of any of them . . . (Whittlesey, 1850, p. 17-18) 6 Belovich and Brose No. 47 FIGURE 3. 1871 Whittlesey map of Fort No. 5: the Greenwood Village site (33Su92). Taken from Ancient Earth Forts of the Cuyahoga Valley, Ohio 1871. The accompanying published map (Figure 2) was evidently drawn by Whittlesey himself. It is inaccurate in some aspects: (1) The narrow “hog’s back” does not run due east as shown, but is oriented east by northeast. 2) The long axis of the plateau lies in a northwesterly direction. 3) The Cuyahoga River and Ohio Erie Canal lie more to the northwest of the site). However, it is the most accurate of all the maps he published of the site (Belovich 1985b). Unfortunately the inaccuracies of the 1850 map reappear in Whittlesey’s later maps. Nevertheless, this map indicates that Whittlesey observed two mounds along the northern edge of the site. It is also clear that four of the embankments had exterior ditches; one was without a ditch (no doubt due to its placement along the bluff edge), and an interior ditch, associated with an embankment which cuts the plateau in half indicated to Whittlesey (1867) “. . . a state of siege.” Whittlesey carefully depicted the “. . . narrow space left as a passage into the work . . .” (Whittlesey, 1850, p. 17-18) as a small western extension perpendicular to the innermost eastern embankment which crossed the narrow hogback leading to the site. One gateway was also present. Subsequent maps, published by Whittlesey in 1867 and 1871 (Figure 3), were accompanied by essentially similar descriptions. However the maps themselves had changed (Belovich, 1985b). On the 1867 map both mounds grew somewhat larger while the earthwork extension became smaller. By 1871 the western earthwork extension had disappeared completely from the map. In addition, not only had the Ohio Erie Canal and the Cuyahoga River changed orientation, but the north arrow had joined them. The shape of the plateau, and the distances between the embankments also changed. The wells vanished but pits appeared, and the mounds not only moved but increased from two to three. These revisions seem so extensive that it is clear that Whittlesey not only never revisited the site, but he even failed to revisit his own notes apparently drawing the later maps from memory, if he drew them at all. In his 1871 publication Whittlesey brought together all the data he collected during his survey on aboriginal occupation sites in the Cuyahoga Valley. Due to the ditches and embankments at many of these sites, Whittlesey called them “forts” and assigned to them a numerical order. At this time the Greenwood Village site was called “Fort No. 5” (Belovich, 1985b, p. 17). Information for the site (apparently gleaned from Whittlesey’s reports) was eventually placed in the official state files. At this time the site was given two names. It was called “Arthur Fort,” after the landowner mentioned by Whittlesey in the 1850 and 1867 publications. The second name, “Whittlesey Fort No. 5,” was taken from Whittlesey’s 1871 publication. Finally the site was assigned the number 33Sul0. Not until 121 years later, in 1971, was the Greenwood Village site (Fort No. 5) revisited. It was investigated as 1992 Late Woodland Fortifications in Northern Ohio 7 (33Su92), Columbia Road Village (33Su87), and Stanford Knoll (33Su99) sites. part of a systematic environmentally stratified survey of northeast Ohio (Brose, 1976a, 1976b). One of the sampling quadrants chosen for testing comprised three plateaux situated along the western bluffs of the Cuyahoga River, on properties then owned by the Greenwood Village Development Corporation (Brose, 1976a; Belovich and Brose, 1983; Belovich, 1985a). The plateaux were labeled according to cardinal directions and it was atop Greenwood Village West that testing was performed (Brose, 1976a; Belovich and Brose, 1983). Three 5 x 5 ft squares were excavated but only one yielded cultural material. This unit, number three, was located “about 2 m south of the northern edge of the narrow plateau neck; to the north of the path” (Belovich and Brose, 1983, p. 15). Cultural material consisted of burned and unburned bone belonging to deer, bear, rodents, birds and fish, and fragments of shell. Worked, polished and incised bone was recovered as were several sandstone abraders, bifacial and unifacial tools, and a Madison projectile point fragment. Shell and grit FIGURE 5. 1983/1984 Map of the Greenwood Village site (33Su92) with test unit locations indicated. tempered ceramics included the types Fairport Harbor Cordmarked, Fairport Filleted, South Park Notched, and Tuttle Hill Notched. A few Wellsberg Simple-Stamped sherds were also recovered. These materials suggested that the site was a small agricultural village occupied from fall through spring, ca. A.D. 1300-A.D. 1400 (Brose, 1976a), relatively late in the Whittlesey Tradition (Brose, 1973). It wasn't until 1979, when Brose and Belovich were performing a survey of the newly formed Cuyahoga Valley National Recreation Area for the National Park Service, that there was occasion to compare the 1971 field notes with the Whittlesey maps. At that time it was concluded that the plateau Brose had labeled Greenwood Village West was the same site Whittlesey had called Fort No. 5 (Brose et al., 1981). Field crews were sent to test the plateau. At that time dense vegetation prevented observation of any earthworks. Of the 10 shovel tests excavated, only three yielded artifacts, none of which were diagnostic. Nevertheless based upon 1971 data and field observations in 1979, it was determined that large portions of the site remained intact (Brose et al., 1981). The site was registered with the state and given a late prehistoric Whittlesey temporal placement, with a suggested date of A.D. 1250- 1400. In 1982, interests again focused on the Greenwood Village site. Further archaeological research had revealed that the site had two different USNM numbers and two Belovich and Brose No. 47 different names. Eventually it was decided that the Greenwood Village site (33Su92), would become the official designation for the prehistoric village first discovered by Col. Charles Whittlesey in 1847 (c.f. Belovich. 1985b, p. 29-30). At this time it was realized that the site, unique in its existence, preservation, and the presence of visible ditches and earthen embankments, was the only Whittlesey Fort remaining along this portion of the Cuyahoga River Valley. Cognizant of the fact that 1979 testing failed to recover diagnostic artifacts and that the site was now suffering from erosion, the National Park Service permitted two years of phased test excavations at the site. The Greenwood Village Site The Greenwood Village site (33Su92), in Northfield Township, extends over an entire plateau overlooking the east side of the Cuyahoga River Valley, in Summit County, Ohio (Figures 4 and 5). On all sides of the site there is evidence of erosion. It is likely that the plateau, as well as the site, was much larger when it was occupied by prehistoric groups. Fimited reconnaissance in the spring of 1983 (Belovich and Brose, 1983; Belovich, 1985b) revealed that the only clearly identifiable ditch/embankment line was the inner- most one along the southern point of the plateau (Figure 2 and Figure 6). At the western extremity of this embankment another embankment joined it to form an “L” extending northwestward along the west side of the plateau. While the short portion of the “L” appeared to be present at the time of reconnaissance (Figure 6), its northwestward extent was not visible. All other ditch/embankment lines identified by Whittlesey had disappeared due to erosion or farming (Belovich, 1985b). While no mounds were observed, several undulations across the eastern narrow hogback were considered to be the possible ditch/embankment lines recorded by Whittlesey (Figure 2). These lines, as well as the interior and exterior of the earthwork, were tested (c.f. Belovich, 1985b). Portions of the “highway” reported by Whittlesey to run through the site and down the western slope to the Canal were present, though the path had varied and was much narrowed (Belovich, 1985b, p. 33). It has been established that this path was once an 1838 roadway (Brose et al., 1981). This path was later used as a jogging and motorcycle trail. It has now been converted by the National Park Service into a pedestrian nature trail. To test the site function and chronology suggested by Brose (1976a) and the accuracy of the reports by Whittlesey (1850, 1867, 1871), four areas of the site were identified for intensive investigation: 1 ) the eastern, narrow neck entrance to the site; 2) the area “inside the lines (where] the ground was much richer than without them” (Whittlesey, 1871, p. 13); 3) the inner ditch/embankment FIGURE 6. Southern ditch/embankment line, view east, March, 1992. (Note arrow indicating day pack on embankment against tree, and tree stump in ditch ). line whose placement resulted in “. . . reducing the fortified area to about one-half the space . . .” (Whittlesey, 1867, p. 38); and 4) the combined areas outside the enclosure, where Whittlesey noted that the ground was not as “rich” as within the enclosure (Whittlesey, 1871, p. 13), and the mounds Whittlesey indicated along the northeastern end of the plateau (Belovich, 1985b). The Features Fifteen test units were excavated during the 1983 field season (Figure 5). After the first season it became clear that the Greenwood Village assemblage being recovered from these sub-surface excavations was not related to the Whittlesey period, but rather to earlier assemblages (Belovich and Brose, 1982; Brose, 1983; Brose and Scarry, 1976). In light of this, Belovich (1985b, p. 4-5) hypothesized that not all hilltop enclosures were defensive works, nor did they all date to the late prehistoric Whittlesey period. The nine test units excavated during the 1984 field season (Figure 5) were placed to obtain additional data on the construction, function and age of the earthen ditches and embankments, and to recover samples suitable for chronometric dating techniques to clearly place all structural and non-structural features in time (Belovich, 1985a, 1985b). Excavations outside the western-most earthwork yielded little cultural material. Soil development was minimal with glaciolacustrine gravels at or near the surface (Belovich, 1985b). This confirmed Whittlesey’s observation that the area outside the enclosure was not as “rich” as that within. Inside the enclosure, the excavation of 18 test units (40m2) revealed 15 cultural features. Most were shallow, basin- shaped, rock-filled, fire pits (Belovich, 1985a, 1985b). Feature 6, located within the enclosure, was a unique pottery deposit containing fragments of a single, Fairport Harbor Cordmarked var. Willoughby vessel (Figure 7). 1992 Late Woodland Fortifications in Northern Ohio 9 \ ■ M FIGURE 7. Reconstructed Fairport Harbor Cordmarked var. Willoughby pottery vessel from Feature 6. a, reconstructed vessel; b, cross section of vessel rim showing large grit-tempering. FIGURE 8. Feature 10, east profile. Feature 6 was encountered 36 cm below the surface and extended to a depth of 58 cm. First identified when excavators observed pottery sherds standing on end, the majority of recovered sherds were piled atop one another in a manner consistent with the interpretation that the vessel collapsed upon itself. The placement of the sherds in this tight cluster suggested that some had been thrown or swept onto the main collapsed pile. Slightly over 200 pottery sherds were recovered from Feature 6 but not all of them could be fitted onto the reconstructed vessel. It is likely that before the vessel was buried some of its constituent sherds may have been scattered. Feature 6 represents a vessel possibly broken while sitting in a shallow pit or depression on the ground’s surface. Five pieces of lithic debitage, one worked piece of shale, some minute bone flecks and a small charcoal sample were also recovered from Feature 6. Feature 10, a meter east of Feature 6, was about 2 m in diameter and nearly 1 m deep. Areas of dense charcoal concentration and a discontinuous ring of burned orange- red soil outlined the feature. The innermost areas were composed of dark yellowish-brown silts and hard-packed clayey silts (Belovich, 1985b, p. 55). Six layers were identified within this great fire pit which yielded lithics, burned soils, dense charcoal deposits, and nearly 300 kg (657 lb) of firecracked rock (Figure 8). Samples of pottery 10 YR 4/4 SANDY-SILT WITH PEBBLES B- 10 YR 3/3 H- 10 YR 2/2 C- 10 YR 3/2 SANDY-SILT WITH FOR D- 10 YR 2/1 FOR AND CHARCOAL F- 5 YR 3/4 SILTY- CLAY WITH PEBBLES 0- 10 YR 5/6 SANDY- SILT HUMUS M ROCK Q GLACIAL TILL 0 TL SAMPLE A C“ 14 SAMPLE POST MOLD 20cm N 10 Belovich and Brose No. 47 for thermoluminesence dating were recovered from the third layer, and 17 carbon 14 samples mostly from the fourth and fifth layers were collected from feature fill (Belovich, 1985b). Feature 10 was devoid of any plant food remains and yielded only exceedingly fragmentary pieces of bone. Similar features were recorded at the Bugai site (Halsey, 1976) and also at the early Late Woodland, Lichliter site near Dayton, Ohio (Allman, 1957). Within the enclosure no evidence for long-term occupation was found (Belovich. 1985a). There were no domestic structures, storage pits, midden or village debris encountered. Even in features, floral and faunal remains were rare or absent. This is in marked contrast to the faunal evidence gathered by Brose in 1971 from areas further east. Excavations begun in 1983 at one of the undulations crossing the eastern, narrow hogback entrance revealed it to be the innermost of the two lines of ditch and embankments Whittlesey recorded in that area (Figures 2 and 5). The aboriginal ditch had been excavated 30 cm into the glacial gravels. These excavated gravels were then mixed with sandy-silts to create a 15-20 cm-thick bed upon which was placed a 20-25 cm-thick layer of shale, sandstone, and igneous rock which varied in size from small pebbles to large cobbles and flagstones (Belovich, 1985b). This prepared rock pavement (Feature 3), encountered some 75 cm below the surface, formed a revetment or foundation for the earthen embankment above. Cobbles and flagstones were noted along the eastern portion of the embankment, some at increasing depths within the ditch. This suggests that they were used to face the eastern, outer slope of the embankment and retard erosion. A refilled post hole measuring 12.5 cm in diameter was observed within Feature 3 (Belovich, 1985a). This post hole may have been part of some superstructure constructed in front of the earthen embankment to prevent erosion of the embankment into the adjacent ditch. No clear evidence of a palisade was observed. The prepared rock pavement appeared to have been less than 2.5 m wide, roughly corresponding to the width of the base of the innermost embankment. If not for a truly serendipitous event this may have marked the end of our earthwork investigations. One excavated test unit, located on the earthen embankment to avoid the roots of several large trees, failed to find the western end of the prepared rock pavement. Two other 1 m x 1 m units excavated adjacent to the west wall of that unit also showed the pavement continuing. At that point it was decided that excavation would proceed in a “westerly direction until we encountered the western termination of Feature 3 or the Cuyahoga River; whichever came first” (Belovich, 1985b, p. 66). Feature 3 ended at a distance of 9.1 m, nearly 30 ft (Figure 9). This prepared rock pavement was constructed in the same manner as that found beneath the embankment. It lay upon a sandy-silt gravel matrix, and maintained a thickness of 20-25 cm until it tapered rather dramatically at its western margin (c.f. Belovich, 1985b, Figures 12, 13 and Plate 6). Like the embankment pavement, this extension was constructed of cobbles and tabular stones laid flat. The stones were from 5-30 cm long and 8-20 cm thick (Belovich, 1985b). Most of the rocks were shale or sandstone, no doubt gathered from the stream beds below. Careful reexamination of Whittlesey’s 1850 map and notes indicated that at the innermost wall “. . . there was a narrow space left as a passage into the work, but none in the outer wall.” This short western extension to the embankment (Figure 2) was undoubtedly the western extension Belovich 1992 Late Woodland Fortifications in Northern Ohio 11 HUMUS SILTS ROCK PAVEMENT SILTY-SAND AND GRAVEL MATRIX GLACSO-LACUSTRINE GRAVELS NORTH FIGURE 10. Hypothesized reconstruction of Feature 3 ditch and embankment. had uncovered. Its graphic representation on the 1850 map, however, contrasts sharply with its length as documented by Belovich. At the time Whittlesey recorded the site, a dirt road leading to the Cuyahoga River crossed the ditch/embankment line at this spot. This road, as well as farming on the plateau, had probably reduced or covered much of the original western extension which was only revealed by these excavations. An hypothesized reconstruction of this extension and prepared rock foundation under the embankment, with a rock facing found within the ditch to retard erosion (Belovich, 1985a, 1985b), is presented in Figure 10. Based on these data it was concluded that . . the ditch and embankment constructions [were] non-defensive earthworks, and the western extension represent[ed] a pedestrian ramp . . .” leading into the main area on the plateau (Belovich, 1985b, p. 180). The few post molds that were identified at the site were located within the enclosure. All of them have been assigned to the Woodland occupation. Three post molds were located within Feature 10, the great fire pit, and one adjacent to it. Three additional post molds were identified about a meter to the southeast of Feature 10. None of these post molds yielded cultural material. Stratigraphically and horizontally they all are associated with Feature 10 and may have functioned as a windbreak and/or a drying rack for Feature 10 (Belovich, 1985). One post mold was identified within eastern profile of Feature 3, the prepared rock pavement. This was the only post mold identified in association with Feature 3 and it may have been part of some type of superstructure constructed in front of the earthen embankment to afford some protection against erosion of the embankment in to the adjacent ditch (Belovich, 1985, p. 42). Five other post molds were located in the central part of the plateau near Features 15 and 16. Based upon limited ceramic evidence these two fire pits and their associated post molds have been assigned to the early Late Woodland occupation at the site (Belovich, 1985). No cultural material was recovered from the post molds, but it is thought that they may have served as windbreaks, spits, or drying racks for these fire pits. The Artifacts The ceramics from the 1983 and 1984 excavation of Greenwood Village site are predominantly coarse grit- tempered wares (97%) although limestone tempering is present (Belovich, 1985b, Table 7). Vessels are simple sub- conoidal in shape, with weak shoulders, straight necks, and flat to slightly rounded lips (Figure 7 and Figure ll:a-b). The average thickness for rim sherds was 0.74 cm (measured 2 cm below the lip), while the average thickness for body sherds was 0.84 cm (Belovich, 1985b). The ceramics are cordmarked or fabric-marked (Figure ll:i). Cordmarking is medium to coarse and usually vertical although other orientations are known (Figure 1 1 :c-d and h). Decoration is rare and when present consists of two to four thin, weakly incised, discontinuous lines over the cordmarking at the neck (Figure 1 l:e-g) (Belovich, 1985b). Similar ceramics found in early Late Woodland sites in Ohio, Indiana, Michigan, New York, western Pennsylvania, and Kentucky have been variously called Fairport Harbor 12 Belovich and Brose No. 47 1992 Late Woodland Fortifications in Northern Ohio 13 Cordmarked (Brose, 1983, 1985), Cuyahoga Cordmarked (Brose and Scarry, 1976), Jack’s Reef Corded (Ritchie, 1965), Wayne Cordmarked (Fitting, 1964), Mixter Cordmarked (Shane, 1967), Watson and Mahoning Cordmarked (Mayer-Oakes, 1955), Newtown Cordmarked (Oehler, 1950; McMichael, 1984), or Peters Cordmarked (Prufer, 1967). While displaying similarities to ceramics from southern Ohio and western Pennsylvania, Greenwood Village ceramics may be more closely related to other types in northeast Ohio (Belovich, 1985b). Some grit-tempered ceramics (including the reconstructed vessel from Feature 6) appear to be examples of the undecorated grit-tempered type Faiiport Harbor Cordmarked var. Willoughby (Brose, 1983, 1985). This type is characterized by an unmodified, sub-conoidal vessel form with flat, slightly everted rims. Surface treatment consists of cordmarking to the lip. Fairport Harbor Cordmarked may date to as early as A.D. 900 (Brose, 1985). Greenwood Village ceramics also possess traits ascribed to the type Cuyahoga Cordmarked (Brose and Scarry, 1976; Brose, 1985). This type has globular to semi-conoidal vessels with straight to moderately outcurved rims, and “massive grit-tempering” (Brose, 1985, p. 52; Brose and Scarry, 1976). Like the type Fairport Harbor Cordmarked, Cuyahoga Cordmarked vessels are usually vertically cordmarked to the lip. Cuyahoga Cordmarked is distinguished from Fairport Harbor Cordmarked by the presence of interior cordmarking on the former (Brose and Scarry, 1976; Brose, 1985). The type Cuyahoga Cordmarked has been dated to between A.D. 900 and A.D. 1200 although it may be an early Late Woodland type which might occur to as early as A.D. 600 (Brose, 1985; Brose and Scarry, 1976, p. 185). Greenwood Village rims display weakly incised horizontal lines (Figure ll:e-g), which can easily be seen as antecedent to the types McFate Incised (Mayer-Oakes, 1955; Murphy, 1971b; Johnson, 1976) and Reeve Horizontal (Belovich, 1985b; Fitting, 1964; Murphy, 1971b; Johnson, 1976). Some Greenwood Village sherds were similar to some ceramics recovered from the Mixter site (Shane, 1967a; Prufer and Shane, 1976), the Lyman site (Muiphy, 1971c), the Fairport Harbor site (Morgan and Ellis, 1943; Muiphy, 1971a), and to the East Wall site, the Reeves site, and some ceramics from the earliest levels at the South Park site (Belovich, 1985b; Brose, 1973, 1985, 1992). All of these FIGURE 11. Selected grit-tempered rim and body sherds from the Greenwood Village site (33Su92). a, straight cordmarked rim sherd ; b, two slightly everted cordmarked rim sherds fitted to one body sherd; c-d, horizontal cordmarked interior and vertical cordmarked exterior of a single body sherd ; e-g,five cordmarked rim sherds and one body sherd with two to four horizontal incisions; h, overlapping cordmarked body sherd; i, four fabric- marked body sherds. sites date between A.D. 800 and A.D. 1200. Comparisons can also be made to earlier time periods and other geographic areas. Watson and Mahoning Cordmarked ceramics from the Upper Ohio Valley (Mayer- Oakes, 1955; Maslowski, 1973; Hemmings, 1984) have granite or limestone tempering and straight necks with flattened or slightly rounded lips. They have vertical exterior cordmarking to the lip, horizontal interior cordmarking, and even fabric-marking (Mayer-Oakes, 1955, p. 191-195). Greenwood Village ceramic assemblage displays all of these characteristics (Belovich, 1985b). Watson and Mahoning ceramics are considered to be Middle Woodland to early Late Woodland utilitarian wares. The Greenwood Village ceramics also show similarities to the cordmarked grit-tempered wares assignable to the Newtown Focus of southwestern Ohio, southeastern Indiana and portions of Kentucky (Oehler, 1950, 1992; Seeman, 1980; McMichael, 1984). Newtown Cordmarked ceramics have vertical to slightly flaring rims with vertically oriented cordmarked exteriors. Some of these sherds display horizontally cordmarked interiors (Oehler, 1950; McMichael, 1984). Like the variety of Peters Cordmarked identified at Chesser Cave in southeastern Ohio (Prufer, 1967, p. 11- 12), Greenwood ceramics have straight necks and flattened lips. They are grit and limestone tempered, cordmarked, and, as Prufer stated, exhibit “shoddily and weakly incised thin lines applied to the neck" (Prufer, 1967, p. 12). Prufer noted the similarity between Peters Phase ceramics and ceramics from sites grouped together in the Cole Complex (Baby and Potter, 1965). Potter (1968, p. 62) suggested a placement of A.D. 800 to A.D. 1300 for the Cole Complex but, as Belovich noted, “Just as Prufer related Peters to Cole, so Potter related Cole to the intrusive Mound Culture of south and central Ohio, suggesting both were the same” (Belovich, 1985b. p. 165- 166). Using chronology and artifact styles Halsey disagrees and suggested . . the Cole Complex post- dates the Mills Phase [his name for Intrusive Mound], although it could possibly be partially contemporaneous” (Halsey, 1976, p. 525). Halsey then fixed the Mills Phase somewhere between A.D. 500 and A.D. 1100 (Halsey, 1976, p. 441-446, p. 519-526). Seven samples for radiometric analysis, four thermoluminescence samples and three carbon- 14 samples (Table 1 ) were sent to three different laboratories; DICARB Radioisotope, Beta Analytic, and Alpha Analytic, by two different submitters; The National Park Service and The Cleveland Museum of Natural History (Belovich, 1985b). Five of these samples were collected from layers within Feature 10, the great fire pit; one from near Feature 10; and one sample from the prepared rock pavement (Feature 3). Unfortunately, for technical reasons, the TL sample 14 Belovich and Brose No. 47 TABLE 1. Radiocarbon and Thermoluminescence Dates for Greenwood Village site (33Su92). Provenience Material Date In Years B.P. Corrected* Date In Years A.D. Corrected* Range Date In Years A.D. Laboratory A0 19-53 TU #10 and 20 Fea. #10, Sec. B 77cm below surface sample # 1 pottery grit-tempered cordmarked body sherd 11801 180 A.D. 770 A.D. 590- A.D. 950 ALPHA- 1484 A0 19-54 TU #19 and 20 Fea. #10, Sec. B 82cm below surface sample #1 pottery grit-tempered cordmarked body sherd 12001210 A.D. 750 A.D. 540- A.D. 960* ALPHA-1485 A0 19- 16 TU #19 and 20 Fea. #10, Sec. C 82cm below surface sample #1 charcoal 12601 80 A.D. 750 A.D. 650- A.D. 850* BETA- 107 15 A0 19-22 TU # 1 0 and 20 Fea. #10, Sec. D 100cm below surface sample #5 charcoal 1500160 A.D. 525* A.D. 460- A.D. 590 BETA- 107 16 A0 19-24 TU #19 and 20 Fea. #10, Sec. D 101cm below surface sample #7 charcoal 1020170 A.D. 985* A.D. 930- A.D. 1040* DICARB-3072 A0 10-23 TU #10, N 1/2 west side of unit. Level 3: 26-36cm limestone-tempered rim sherd (once thought to be shell-tempered) 11101 170 A.D. 840 A.D. 670- A.D. 1010 ALPHA- 1486 A016-36 TU #16, Fea. #3 Level 9: 1 10- 130cm 1 1cm N, 167cm W, 122cm below surface cordmarked grit-tempered body sherd 9.29cm x 636cm 1 . 1cm No date obtainable, no fine grain, fraction present ALPHA- 1487 (ALPHA-1487) from the pavement was undatable. Nevertheless, the cordmarked grit-tempered ceramics from the pavement, identical to ceramics dated elsewhere on the site, lend confidence to assigning the earthworks’ construction to the early Late Woodland Period. Based upon all the dates run, the major period of occupation and earthwork construction at the Greenwood Village site occurred some time between A.D. 460 and A.D. 1040, with the most likely period being between A.D. 600 and A.D. 800. With the dates from the nearby Stanford Knoll (Lee, 1986) and Columbia Road Village (Belovich and Brose, 1982) sites, we have identified the temporal placement of coarse, grit- and limestone- tempered, cordmarked ceramics within the Cuyahoga River Valley. All radiometric information for these three northeastern Ohio sites is displayed in graphic form in Figure 12. The lithic assemblage from Greenwood Village also contains elements common to early Late Woodland sites: shale discs and knives, and a projectile point cluster variously called Chesser Notched, Lowe Flared Base, or what Lee (1986) called Anthony Side Notched (Figures 1992 Late Woodland Fortifications in Northern Ohio 15 GREENWOOD STANFORD KNOLL VILLAGE 33Su99 33Su92 COLUMBIA ROAD VILLAGE 33Su87 A. D. 1 250 1 000 750 500 250 A.D. 1 250 500 750 1 000 B. C. 1 250 □ □ i i FIGURE 12. Radiocarbon dates for the Greenwood Village (33Su92), Columbia Road Village (33Su87), and Stanford Knoll (33Su99) sites. 13 and 14) (Belovich, 1985b). The shale discs and knives, important identifying characteristics of early Late Woodland lithic assemblages, have been referred to as slate discs by Halsey (1976) and others (Allman 1957, 1961). This is clearly incorrect since shale, not slate, occurs in Ohio and slate is almost non-existent in lower Michigan as well. Small, crudely flaked points, with diamond or humpbacked cross sections generally referred to as “fishspears” (Prufer, 1967; Converse, 1984) were also common (Figure 13). Pitted stones, a net sinker, hammerstones, a small shale hoe, as well as other groundstone tool fragments were recovered. Finally a modest amount of firecracked rock totalling 395.47 kg (871.86 lb) was collected and analyzed. The lithic assemblage needs little more comment than reiterating that its constituent “fishspears,” Chesser/Anthony Side Notched/Lowe Flared projectile points and shale knives and discs consistently occur at early Late Woodland sites across Ohio (Schatz, 1957; Allman, 1957, 1961; Converse, 1963, 1984; Prufer, 1965, 1967, 1981; Halsey, 1976; Oplinger, 1981; Ormerod, 1983). An Archaic component at the site is represented by a LeCroy bifurcate and a crudely stemmed projectile point (Belovich, 1985). Both were surface finds. No features could be assigned to the Archaic period. Comparative Discussion The 1983-1984 Greenwood Village site excavations failed to yield any evidence for long-term village occupation. Almost all features excavated were smali firepits and there were no domestic structures, storage pits or deep midden zones of any kind (Belovich, 1983, 1985a, 1985b). Even such large, deep, fire pits, as Feature 10 are essentially devoid of plant food or faunal remains. Like the other small firepits on the site. Feature 10 represents a singular utilization event (Belovich, 1985b). The Greenwood Village site shares traits in common with other early Late Woodland sites. The Bugai site from Saginaw County, Michigan, belongs to the Wayne Mortuary Complex. The site is dated between A.D. 500 and A.D. 1100 (Halsey, 1976, p. 445, 473). At the Bugai site a basin-shaped fire pit five feet in diameter and about a foot deep (Halsey, 1976, p. 475, 480-481) yielded fired ocher, charcoal, and two chipped discs of what Halsey (1976, p. 506) called slate. Directly below this fire pit were two bundle burials, one associated with another “slate" disc. Additional “slate" discs were recovered from other features at the Bugai site. While even a pottery concentration was excavated at Bugai (Halsey, 1976, p. 483), there were no earthworks. The Lichliter site is a late Middle Woodland or early Late Woodland site near Dayton. Ohio (Allman, 1957). 16 Belovich and Brose No. 47 1992 Late Woodland Fortifications in Northern Ohio 17 Four houses were located and the one completely excavated circular house measured 48 ft in diameter. At its north end was a nearly square fire pit about 4 ft on a side and almost 2 ft deep. Allman observed heavy con- centrations of ash and charcoal, noting that “. . . the sides were quite deeply burnt . . .” (Allman, 1957, p. 60). Faunal remains were absent but Allman did recover many stones, . . some over 6 inches in size . . .” and one pottery sherd. Allman concluded . . that this might have been a ceremonial fireplace . . .” (Allman, 1957, p. 60). This feature is similar to Feature 10 at Greenwood Village. Allman also recorded that over 55 whole and broken “slate” discs were recovered from the Lichliter site. From his published photographs, the projectile points from the site are similar to Hopewell, Chesser, and Anthony Side Notched points (Allman, 1957, p. 62). Rock pavements and foundations like those at Greenwood Village are not unique. The Pollock Works, first recorded by Squier and Davis, were recently investigated by Riordan (1982). The site is composed of an earthen wall, cut by three gateways, stretching across a plateau situated 30 ft above Massies Creek in Greene County, Ohio (Riordan, 1982). Outside and some distance from these earthen constructions, are three semicircular earthen walls separated by three more gateways. Riordan’s excavations uncovered three limestone rock pavements at each gateway which he believes functioned to protect the earthworks from erosion and may also have served as walkways to and from the plateau’s interior. Initial radiometric dates placed the Pollock Works between 230 and 400 B.C. suggesting a late Adena affiliation (Riordan, 1982, p. 15-16). Additional radiocarbon dates obtained from wood charcoal recovered from on and beneath the Pollock earthworks clearly indicate that the site’s major period of construction was during the first and second centuries A.D. (Riordan, 1986). Riordan considers the earlier dates aberrant (Riordan, 1986). The Fort Ancient site also has Middle Woodland earthen embankments and rock pavements (Essenpreis and Moseley, 1984). As at Greenwood Village, embankments outline the plateau, and ditches are found inside the embankment. Essenpreis and Moseley demonstrated that stone was used extensively; as facing for the outer embankment slopes, short walls to retard erosion down the FIGURE 13. Selected projectile points and scrapers from the Greenwood Village site (33Su92). a-c, Anthony Side Notched projectile points; d, hafted scraper made from an Anthony Side Notched projectile point; e, corner-notched projectile point similar to Lowe Flared Base type; f, thick, reworked projectile point similar to Lowe Flared Base or Chesser Notched types; g-o, projectile points with diamond-shaped or humpbacked cross sections similar to "fishspears”; p-q, Chesser Notched projectile points; r, hafted end scraper; s, thumbnail scraper. plateau slopes, and as cores or foundations for some of the embankments. Rock pavements, noted as early as 1 940 by Morgan, seemed to have served as roadways between various structures within the enclosure (Morgan, 1946, 1970; Morgan and Thomas, 1950; Essenpreis and Moseley 1984) . The numerous gateways, absence of structures preventing travel through the gateways, and the place- ment of ditches inside the embankments all suggest a non-defensive function. Miami Fort, Fort Glenford, Fort Hill, Windsor Fort, Turner, Spruce Hill, Big Creek and Indian Fort Mountain have all yielded evidence of earthen embankments and rock pavements (Fischer, 1974, p. 87). Fort Hill ( 3 3 C u 1 ) was first surveyed by Charles Whittlesey prior to 1888. The site is located on a plateau 200 ft above the confluence of the East and West Branches of the Rocky River. At the eastern tip of this plateau Whittlesey identified three parallel 15 ft-wide walls with 1 1 ft-wide ditches along their western faces (Whittlesey, 1888). At the time of his survey the maximum distance from the top of an embankment to base of a ditch was 4 ft (Whittlesey, 1888). Eight days of shovel testing in late 1985 failed to recover any evidence for a late prehistoric Whittlesey period village occupation (Lee and Belovich, 1985) . There was no midden deposit of any kind, and shovel testing across the entire plateau recovered only two flint flakes (one utilized). Test excavations of the most visible of the three embankments (all have been severely disturbed by erosion, plowing or dirt bike traffic) indicated that the feature was clearly of cultural origin. Artifacts recovered from the embankment consisted of one flint flake and one quartz flake. Despite the absence of diagnostic ceramics, the similarities of Fort Hill to Greenwood Village (steeply sided plateau [twice the elevation of Greenwood Village], complete with a series of ditches and embankments, and exhibiting no midden deposit and very few artifacts) led Belovich to consider this as evidence for an early Late Woodland placement for Fort Hill (Belovich, 1985b). Windsor Fort (33Ab3) is located on a peninsular plateau 90 ft above the west bank of Phelps Creek in Ashtabula County. The western, landward side of the plateau is crossed by two parallel ditches and embankments, about 150 ft long. The ditches lie adjacent to the west side of each embankment. The deepest, outermost ditch gives its associated embankment an apparent 7 ft height; however, when viewed from the plateau interior this wall appears only about 3 ft high. Test excavations conducted at Windsor Fort by Lee, as part of an archaeological reconnaissance of Ashtabula County, failed to uncover any evidence of deep midden deposits or village debris. Artifacts recovered were limited to small amounts of lithic debitage and three cordmarked grit-tempered body sherds. These limited data suggested to Lee that Windsor Fort was 18 Belovich and Brose No. 47 FIGURE 14. Selected shale knives and tools from the Greenwood Village site (33Su92). a, lanceolate shale knife; b-c, unifacial shale scrapers; d-e, shale discs; f, shale “hoe. ” similar to Greenwood Village and more likely to date to the early Late Woodland period rather than the late prehistoric Whittlesey period (Lee, 1987). Sites with artifact assemblages similar to Greenwood Village have also been noted in other regions of Ohio. The Water Plant site (33Fr 155) in central Ohio, sits atop a high bluff overlooking Big Walnut Creek, a tributary of the Scioto River (Dancey, 1988). Dancey identified eleven discrete debris clusters which he interpreted as distinct household units (Dancey, 1988, p. 223) although no structural evidence for houses was observed. The artifact assemblage from this site is characterized by Chesser 1992 Late Woodland Fortifications in Northern Ohio 19 Notched projectile points and cordmarked grit-tempered pottery. As at Greenwood Village, ceramic decoration on Water Plant site sherds is rare. When present it consists of poorly incised lines nearly parallel to the rim (Dancey, 1988). Dancey states that the ceramics resemble the type Newtown Cordmarked (c.f. McMichael, 1984). The lithics, ceramics, and radiocarbon dates place this small village site firmly within the early Late Woodland period. Another site in northeast Ohio, the Columbia Road Village site (33Su87), is situated atop a high ridge along the west side of the Cuyahoga River Valley (Belovich and Brose, 1982). The plateau is surrounded on three sides by very steep ravines with small intermittent streams. The fourth (south) side trails into a narrow neck which descends to another deep ravine. While no earthworks were encountered at this site, four features were excavated. Feature 2 was a random deposit of broken sherds. Feature 1 was rock-lined fire pit with a double post mold (Feature 3) adjacent to it. Feature 4 was partially excavated and appeared to be a fire pit. The large number of artifacts recovered from these excavations included projectile points, point fragments, tools, chert and shale flakes, and lithic debitage. The points and point fragments are morphologically similar to Chesser Notched points reported from the Ohio Valley (Mayer-Oakes, 1955, p. 83b; Prufer, 1967, p. 54, fig. 5, n-z). Some of these points also resemble Lamoka, Jack’s Reef Corner Notched, and Lowe Flared point types (Ritchie, 1971; Reidhead and Limp, 1974). Grindingstones, hammerstones, a pipe bowl and shale gorget fragment complete the stone artifact inventory (Belovich and Brose, 1982). All of the sherds recovered from the Columbia Road Village site were grit-tempered and some were cordmarked. Rims were cordmarked, with the cordmarking usually oriented vertically. In some instances cordmarking occurred on the inside of the rim. Most rims were straight sided or slightly excurvate while lips were usually flat. Decoration was rare; only one everted rim with finger-nail punctates was collected (Belovich and Brose, 1982). Pottery from the Columbia Road Village site resembles the limestone-tempered Watson Cordmarked ceramics from the Upper Ohio Valley (Mayer-Oakes, 1955, p. 193-195) and the limestone-tempered cordmarked sherds from pre- Fort Ancient contexts at the Haag Site (ca. A.D. 700 - ca. A.D. 950) in southeastern Indiana (Reidhead and Limp, 1974, p. 9). Similar grit-tempered ceramics were found at the Gillie Rock Shelter near Twinsburg, Ohio, dated between A.D. 200 and A.D. 700 (Bernhardt, 1973). While Columbia Road Village site ceramics appear similar to Peters Cordmarked from Chesser Cave (Prufer and McKenzie, 1966, p. 60), Prufer stated that the ceramics from 33Su87 are somewhat thicker, are better fired, and had finer fabrics used in their manufacture than those he had found at Chesser (personal communication, 1985). By far the strongest resemblances are with the grit-tempered cordmarked and plain ceramics from nearby Hale Farm (Brose, 1985), and the Boston Ledges Rock Shelters A and B (Brose and Scarry, 1976). Though absolute dates are not available for either of those sites, their assemblages suggest a chronological placement between A.D. 600 and A.D. 1000. Charcoal samples from Feature 1 at the Columbia Road Village site (Figure 12) date between A.D. 780 and A.D. 1005 (A.D. 930, 1090 + 85 B.P., DICARB - 2605 and A.D. 850, 1160 + 60/-50 B.P., DICARB - 2606) (c.f. for a discussion of pertinent issues surrounding DICARB dates Belovich, 1985b; Belovich and Brose, 1982). These dates plus the regional comparison of the site assemblage clearly indicate an early Late Woodland chronological placement (Figure 12). The multicomponent Stanford Knoll site (33Su99), located east of the Cuyahoga River had Early Woodland (950 B.C.), Middle Woodland (A.D. 170-300) and early Late Woodland (A.D. 600) manifestations (Lee, 1986). While no structures were encountered, cordmarked, grit- tempered ceramics dating to the early Late Woodland period were recovered from Feature 6, a 53 cm diameter, 26 cm-deep pit. These well-fired ceramics were from 8.4- 14.0 mm thick. Shallow cord impressions were “fairly widely spaced (average 1.6 mm)” (Lee, 1986, p. 40). They are virtually indistinguishable from ceramics recovered from the Greenwood Village site and could be lost within its ceramic assemblage. A thermoluminescence (Figure 12) date obtained for one of these sherds was A.D. 620 ( 1330 + 150, B.P. ALPHA-2621) (Lee, 1986). A carbon sample (Figure 12) from Feature 11 placed a rim sherd, virtually indistinguishable from ceramics from the Greenwood Village site, as early as A.D. 235 (1780 + 60 B.P, BETA- 15012) (Lee, 1986). Two Racoon Side Notched projectile points recovered from Feature 6 suggest a very early Late Woodland date (Ritchie, 1965, p. 228, p. 258-260). Belovich (1985b) clearly noted that early Late Woodland phases, complexes, and cultures were poorly defined both in terms of their characteristics and chronology. In an attempt to bring order to these data for southern Ohio, Seeman (1980), noting the similarity between sites originally assigned to the Cole Complex, Newtown Focus, and Peters Phase, suggested they be grouped together. He then gave them the descriptive, yet cumbersome name of Central Ohio Valley early Late Woodland and suggested a time span of A.D. 500 to A.D. 800 (Seeman, 1980, p. 17). Seeman also suggested that such Central Ohio Valley early Late Woodland cultures would extend from southwestern Ohio (Peters Cordmarked), northeastward to Pittsburgh, Pennsylvania (Watson Cordmarked) (Seeman, 1980, p. 16-17). In 20 Belovich and Brose No. 47 Seeman’s scheme, the distinct Intrusive Mound Culture spans the period A.D. 800 to A.D. 1000 (Seeman, 1980, p. 16-18). As Belovich noted, under this scheme the Cole Complex also remains, “. . . but it is now represented only by those sites not already subsumed (with Newtown and Peters) under the nomen Central Ohio Valley early late (sic) Woodland. Due to their late temporal placement (post A.D. 900) and certain shared characters (guilloche designs) these Cole Complex sites are considered “to represent a late Late Woodland complex at least partially contemporaneous with Fort Ancient components to the south” (Seeman 1980:18). Finally, Seeman considers the Chesser Phase to be so poorly defined that not much more than its latest temporal placement (A.D. 1200) is suggested.” (Belovich, 1985b, p. 166) It bears repeating, that the Greenwood Village ceramics share traits with Mixter Cordmarked, Cuyahoga Cordmarked, and Fairport Flarbor Cordmarked (var. Willoughby) ceramics which may date from as early as A.D. 600. Few Middle Woodland sites from which a local antecedent for the Greenwood Village ceramics can be identified are known for northeast Ohio. Belovich (1985b) suggested that the early Late Woodland Greenwood Village ceramics will in fact be found to develop out of indigenous Middle Woodland styles such as those types seen at Stanford Knoll (Lee, 1986) and at other late Hopewellian sites such as Everett Knoll (Brose, 1974), the North Benton Mound (McGrath, 1945) and the Huntington Road Village site (Evangilista and Dodd, personal communication). Conclusions It now seems clear that the late prehistoric Whittlesey materials Brose reported in 1973 (Belovich and Brose, 1983) cannot be associated with the construction of any earthwork reported by Whittlesey at Greenwood Village. Nor do they represent any significant occupation within the enclosure. Similar late Whittlesey materials were recovered in 1985 from contexts associated with ephemeral occupations outside the enclosure and approximately 107 m further east along the narrow hogback entrance to the site, where there is some geophysical evidence for an embankment not recorded by Whittlesey (c.f. Belovich, 1985b). These investigations place the occupations responsible for the earthwork constructions at a pre-Whittlesey date between A.D. 600 and A.D. 800 (Belovich, 1985a, 1985b; Belovich and Brose, 1983). The absence of structures, deep midden deposits, or storage pits with significant faunal or floral remains, and the numerous single-use fire pits, suggest limited domestic activities which seem functionally, structurally, and even chronologically more closely related to preceding Middle Woodland phenomena than to the subsequent late prehistoric Whittlesey Tradition. Indeed, there is unambiguous evidence that burial mound construction continued into the 14th century A.D. (Belovich, 1986). The Greenwood Village earthworks and mounds further suggest that much of the cultural resources and energy expended at the site was focused on the construction activities themselves. Greenwood Village can therefore be seen as a continuation of traditional mound and earthwork construction first seen in the Early Woodland period. Construction techniques were little changed, despite the apparent social and organizational changes presumed to have occurred by the Late Woodland period. The apparent absence of exotic artifacts, mound burials, or cremations or other internments at Greenwood Village may very well be attributed to the erosion of the mounds once located along the northwestern edge of the site, and the poor preservation of bone (Belovich, 1985b). Their absence at other early Late Woodland sites may also be an artifact of preservation or it may be due to the collapse of the elaborate social ceremonialism of the Middle Woodland period. Additional research in northern Ohio may clarify the nature of this transition from horticultural Middle Woodland, to more agricultural early Late Woodland villages, to the economic dependence on cultigens characteristic of the late prehistoric Whittlesey Tradition. While further excavations may be needed before any definitive statements can be made regarding social function at such sites as Greenwood Village or Fort Ancient, these investigations have clearly supported Belovich’s (1985, p. 191) hypothesis that not all hilltop enclosures are defensive works, nor can it be assumed that all of Whittlesey’s “forts” date to the late prehistoric Whittlesey period . Acknowledgements We wish to acknowledge the help and contributions of Hiram College, Cuyahoga Community College, The Cleveland Museum of Natural History and especially the William Bingham Foundation. Without their financial support this research would not have been possible. Thanks must also go to F. Calabrese and Mark Lynott, of the Midwest Archaeological Center and to Lewis Albert, Superintendent, the Cuyahoga Valley National Recreation Area for granting permission to test the site. We also wish to mention the immeasurable contributions of Catherine Hovey, Sean Coughlin, Alfred M. Lee, David McNickle, Robert R Mensforth, Laura Pompignano, Judy Placko, and Mark Ohlberg in the areas of organization, fieldwork, photography, drawing and data analysis. Wendy Wasman and Dan Flocke lent us assistance, as did a skilled crew of 1992 Late Woodland Fortifications in Northern Ohio 21 field and laboratory volunteers who shared in this project. The hard work, enthusiasm, and good humor from all of these folk were important contributions to the success of this project. The authors would also like to thank the reviewers for their helpful suggestions. Finally, the senior author extends a very special thanks to the late Patricia S. Essenpreis for providing the impetus to write this paper by requesting the inclusion of my Masters research in her and Robert Riordan's Woodland Enclosures Symposium at the 1986 meetings of the Society for American Archaeology. Thanks Pat: Miss you. References Allman, John C. 1957. A new Late Woodland culture for Ohio. The Lichliter Village site near Dayton. Ohio Archaeologist 7(2):59-68. Allman, John C. 1961. Late Woodland stone discs from Preble County, Ohio. Ohio Archaeologist 11(4): 136. Baby, Raymond S., and Martha A. Potter. 1965. The Cole Complex: A Preliminary Analysis of the Late Woodland Ceramics in Ohio and Their Relationship to the Ohio Hopewell Phase. Ohio Historical Society, Papers in Archaeology 2. Columbus. Belovich, Stephanie J. 1985a. Evaluative testing at the Greenwood Village Site (33Su92): 1984 season. Archaeological Research Report 56. The Cleveland Museum of Natural History, Cleveland. Belovich, Stephanie J. 1985b. The Greenwood Village site (33Su92): an early Late Woodland Site along the Cuyahoga River. Unpublished M.A. thesis, Kent State University, Kent. Belovich, Stephanie J. 1986. The Staas site (33Cu224): 1985 excavations. Archaeological Research Report 66. The Cleveland Museum of Natural History, Cleveland. Belovich, Stephanie J., and David S. Brose. 1982. Survey and evaluative testing within the Cuyahoga Valley National Recreation Area: Sites 33Su87 and 33Sul02. Archaeological Research Report 40. The Cleveland Museum of Natural History, Cleveland. Belovich. Stephanie J., and David S. Brose. 1983. Evaluative testing at the Greenwood Village site (33Su92). Archaeological Research Report 47. The Cleveland Museum of Natural History, Cleveland. Bernhardt, Jack E. 1973. Gillie Rockshelter: A Late Woodland phase in Summit County, Ohio. Unpublished M.A. thesis, Kent State University, Kent. Braun, E. Lucy. 1950. Deciduous Forests of Eastern North America. Blakiston Co., Philadelphia. Brose, David S. 1973. A preliminary analysis of recent excavations at the South Park site, Cuyahoga County, Ohio. Pennsylvania Archaeologist 43( 1 ):25-42. Brose, David S. 1974. The Everett Knoll: a late Hopewellian site in northeastern Ohio. Ohio Journal of Science 74(1): 36-46. Brose, David S. 1976a. An initial summary of the Late Prehistoric period in northeastern Ohio. In The Late Prehistory of the Lake Erie Drainage Basin: A 1972 Symposium Revised, edited by D. Brose, The Cleveland Museum of Natural History, Cleveland. Brose, David S. 1976b. Locational analysis in the prehistory of northeast Ohio. In Cultural Change and Continuity: Essays in Honor of James Bennett Griffin, edited by C. Cleland, p. 3-18. Academic Press, New York. Brose, David S. 1983. A statistically dendritic analysis of Whittlesey ceramics with an illustrated type-variety classification. Manuscript on file in The Cleveland Museum of Natural History Archaeological Archives. Brose, David S. 1985. The prehistoric occupation of the Hale Farm, Bath Township, Summit County, Ohio. Kirtlandia 41:35-61. Brose, David S. 1992. The South Park Site and the Whittlesey Tradition of Northeast Ohio. The Cleveland Museum of Natural History, Cleveland. Brose, David S., S. Belovich, M. Brooslin, R. Burns, Jr., J. Hall. H. Haller, C. Pierce, and C. Ubbelohde. 1981. Archaeological investigations in the Cuyahoga Valley National Recreation Area. Archaeological Research Report 30. The Cleveland Museum of Natural History, Cleveland. Brose, David S., and John F. Scarry. 1976. The Boston Ledges Shelter: comparative spatial analyses of early Late Woodland occupations in Summit County, Ohio. Midcontinental Journal of Archaeology 1 (2): 1 79-228. Converse, Robert N. 1963. Ohio flint types. Ohio Archaeologist 13(4):77- 1 2 1 . Converse, Robert N. 1973. Ohio Flint Types. (Revised) Archaeological Society of Ohio. Converse, Robert N. 1984. Fishspear points. Ohio Archaeologist 34(4):5. Dancy, William S. 1988. The community plan of an early Late Woodland village in the middle Scioto River valley. Midcontinental Journal of Archaeology 1 3(2):223-258. Essenpreis, Patricia S., and Michael E. Moseley. 1984. Fort Ancient: citadel or coliseum? Field Museum of Natural History Bulletin 55(6):5-10, 20-26. Fischer, Fred William. 1974. Early and Middle Woodland settlement, subsistence, and population in the central Ohio valley. Unpublished Ph.D. dissertation, Washington University, St Louis. Fitting, James E. 1964. Ceramic relationships of four Late Woodland sites in northern Ohio. Wisconsin Archaeologist 45(4): 160-175. Gordon, Robert B. 1969. The natural vegetation of Ohio in pioneer days. Bulletin of the Ohio Biological Survey 3(2): 1 - 1 13. Columbus. Greenman, Emerson F. 1935a. Excavation of the Reeve Village site. Lake County, Ohio. Ohio State Archaeological and Historical Quarterly 44( 1 ):2-64. Greenman, Emerson F. 1935b. Seven prehistoric sites in northern Ohio. The Ohio State Archaeological and Historical Quarterly 44(2):220-237. 22 Belovich and Brose No. 47 Greenman, Emerson F. 1937. Two prehistoric villages near Cleveland, Ohio The Ohio State Archaeological and Historical Quarterly 46(4):305-366. Halsey, John Robert. 1976. The Bussinger site: A multicomponent site in the Saginaw Valley of Michigan with a review of early Late Woodland mortuary complexes in the northeastern woodlands. Unpublished Ph.D. dissertation. University of North Carolina, Chapel Hill. Hemmings, E. Thomas. 1984. The Fairchance Site: Middle Woodland settlement and subsistence in the upper Ohio Valley. West Virginia Archaeologist 36( 1 ):3-5 1 . Johnson, William C. 1976. The Late Woodland period in northwestern Pennsylvania: a preliminary survey and analysis for the symposium on the Late Woodland period in the Lake Erie drainage basin. In The Late Prehistory of the Lake Erie Drainage Basin: A 1972 Symposium Revised , edited by D. Brose, p. 48-75. The Cleveland Museum of Natural History, Cleveland. Lee, Alfred M. 1982. Archaeological Survey of Trumbull County. Prepared by the Trumbull County Planning Commission and The Cleveland Museum of Natural History, Cleveland, and Warren, Ohio. Lee, Alfred M. 1986. Excavations at the Stanford Knoll site, Cuyahoga Valley National Recreation Area. Archaeological Research Report 65. The Cleveland Museum of Natural History, Cleveland. Lee, Alfred M. 1986. Unpublished type descriptions for Anthony Side Notched projectile points in author’s possession. Lee, Alfred M. 1987. Archaeological reconnaissance in Ashtabula County, Ohio. Archaeological Research Report 78. The Cleveland Museum of Natural History, Cleveland. Lee, Alfred M., and Stephanie J. Belovich. 1985. Field notes and forms from test excavations in 1985 at Fort Hill ( 3 3Cu 1 ) on file at The Cleveland Museum of Natural History, Cleveland. MacGrath. Willis H. 1945. The North Benton Mound: a Hopewell site in Ohio. American Antiquity 11(1 1:40-46. Maslowski, R. F. 1973. An analysis of cordmarked Watson Ware. Pennsylvania Archaeologist 43(2): 1-12. Mayer-Oakes, William J. 1955. Prehistory of the upper Ohio valley. Annals of Carnegie Museum 34. Anthropological Series No. 2. McMichael, Edward V. 1984. Appendix A: Type descriptions for Newtown series ceramics. Posthumously published In The Pyles Site (15MS28). a Newtown Village in Mason County Kentucky, edited by J. Railey, p. 132-135. William S. Webb Archaeological Society, Occasional Paper 1. Lexington. (Publication of ms circulated in the 1960s). Morgan, Richard G. 1946. Fort Ancient. Ohio State Archaeological and Historical Society, Columbus. Morgan, Richard G. 1970. Fort Ancient. The Ohio Historical Society, Columbus. Morgan, Richard G., and H. Holmes Ellis. 1943. The Fairport Harbor site. The Ohio State Archaeological and Historical Quarterly. 52(l):3-64. Morgan, Richard G., and Edward S. Thomas. 1950. Fort Hill. The Ohio State Archaeological and Historical Society, Columbus. Murphy, James L. 1971a. The Fairport Harbor site (33La5) Lake County, Ohio. Pennsylvania Archaeologist 41 (3):26-43. Murphy, James L. 1971b. Whittlesey ceramic types. Ohio Archaeologist 2 1 ( 1 ):298-303. Murphy, James L. 1971c. The Lyman site (33La2), Lake County, Ohio. Pennsylvania Archaeologist 41(3): 12-25. Oehler, Charles. 1950. Turpin Indians. Cincinnati Museum of Natural History Popular Publication Series 1. (Revised and reprinted for The Journal of the Cincinnati Museum of Natural History 23[2] 1973). Oplinger, Jon. 1981. Wise Rockshelter: A multicomponent site in Jackson County, Ohio. Kent State Research Papers in Archaeology 2. Kent State University Press, Kent. Ormerod, Dana E. 1983. White Rocks: A Woodland rockshelter in Monroe County, Ohio. Kent State Research Papers in Archaeology 4. Kent State University Press, Kent. Potter, Martha A. 1968. Ohio’s Prehistoric Peoples. The Ohio Historical Society, Columbus. Potter, Martha A., and Edward S. Thomas. 1970. Fort Hill. The Ohio Historical Society, Columbus. Prufer, Olaf H. 1965. The McGraw Site: A Study in Hopewellian Dynamics. Scientific Publications of The Cleveland Museum of Natural History 4(1). Prufer, Olaf H. 1967. Chesser Cave: a Late Woodland phase in southeastern Ohio. In Studies in Ohio Archaeology, edited by O. Prufer and D. McKenzie, p. 1-62. The Press of Western Reserve University, Cleveland. Prufer, Olaf H. 1981. Raven Rocks: A specialized Late Woodland rockshelter occupation in Belmont County, Ohio. Kent State Research Papers in Archaeology 1. Kent State University Press, Kent. Prufer, Olaf H., and Douglas H. McKenzie. 1966. Peters Cave: two Woodland occupations in Ross County, Ohio. Ohio Journal of Science 66(3):233-253. Prufer, Olaf H., and Orrin C. Shane, III. 1976. The Portage- Sandusky-Vermillion River region in Ohio. In The Late Prehistory of the Lake Erie Drainage Basin: A 1972 Symposium Revised, edited by D. Brose, p. 283-304. The Cleveland Museum of Natural History, Cleveland. Riordan, Robert V. 1982. The Pollock Works: report on the 1981 field season. Wright State University Laboratory of Anthropology, Reports in Anthropology 6. Reidhead, Van A., and William F. Limp. 1974. The Haag site ( 1 2D 19): A preliminary report. Indiana Archaeological Bulletin 1(1 ):4- 19. Ritchie, William A. 1965. The Archaeology of New York State. New York: Natural History Press. Ritchie, William A. 1971. A Typology and Nomenclature for New York Projectile Points. New York State Museum and Science Service, Bulletin No. 384. Second edition, revised. Albany. Schatz, John W. 1957. Late Woodland projectile types from central Ohio. Ohio Archaeologist 7(4): 1 34- 1 35. 1992 Late Woodland Fortifications in Northern Ohio 23 Seeman, Mark F. 1980. A taxonomic review of southern Ohio Late Woodland. Paper presented at the 25th Annual Midwestern Archaeological Conference, Chicago, Illinois. Seeman, Mark F. (editor). 1992. Cultural Variability in Context: Woodland Settlements of the Mid-Ohio Valley. MCJA Special Paper No. 7. The Kent State University Press, Kent. Shane, Orrin C., III. 1967. The Mixter Site: a multicomponent hunting station in Erie County, Ohio. In Studies in Ohio Archaeology , edited by O. Prufer and D. McKenzie, p. 121- 186. The Press of Western Reserve University, Cleveland. U. S. Department of Agriculture. 1971. Soils, the Seventh Approximation (with supplements). Washington, D.C. Department of Agriculture, Soil Survey Service. White, George W„ with Stanley M. Totten. 1982. Glacial geology of northeastern Ohio, with a chapter on Pleistocene beaches and strandlines bordering Lake Erie. State of Ohio, Department of Natural Resources, Division of Geological Survey, Bulletin 68. Columbus. Whittlesey, Col. Charles. 1850. Descriptions of ancient works in Ohio. Smithsonian Contributions to Knowledge, Vol. II, Article VII, pp. 17-18, Plate VII, No. A. Whittlesey, Col. Charles. 1867. Early History of Cleveland, Ohio. Fairbanks, Benedict and Company. Whittlesey, Col. Charles. 1871. Ancient earth forts of the Cuyahoga Valley, Ohio. Western Reserve and Northern Ohio Historical Society Tracts: Number 5, Fairbanks, Benedict and Company. Whittlesey, Col. Charles. 1888. Ancient earthworks - northern Ohio. Western Reserve and Northern Ohio Historical Society, Tract Number 41, Vol. 11:34-35. Williams, Arthur B. 1949. The native forests of Cuyahoga County, Ohio. Scientific Publications of The Cleveland Museum of Natural History 9( 1 ): 1 -90. •NATURAL HISTORY* Published by The Cleveland Museum of Natural History 1 Wade Oval Drive Cleveland, Ohio 44106-1767 Production by Melody Oakes HECKMAN |X| BINDERY INC. p| OCT 98 Bound 'To -Pleas^1 fN*gR