Since 1990, the Tongass National Forest has supported an innovative program of interdisciplinary research involving archaeologists, geologists, spelunkers, and others to identify caves and rockshelters on the Prince of Wales Archipelago. This effort was prompted by passage in 1988 of the federal Cave Resource Protection Act which recognized that the scientific potential of many of these unique places had already been inadvertently damaged. These caves and rockshelters contain rare and well-preserved records of glacial and climatic history, vertebrate paleontology, paleoecology, sea level history, and human history. The acronym, SEACAVES, stands for Southeast Alaska Caves Project.

University of Oregon scholars have participated in some of this research since 1994. Following up on our geochemical study of Suemez Island obsidian (Erlandson et al. 1992), in 1994, Forest Service archaeologist John Autrey found a rockshelter in the forest behind the beach perched atop a talus cone from which obsidian was eroding. In 1995, with the assistance of Forest Service archaeologist Jane Smith, we tested this rockshelter, but found no evidence of human occupation. We did find occupational evidence closer to the modern shoreline, under one rock overhang and in two nearby seacaves. Even these sites, dating to the last 2000 years, contained little in the way of obsidian use. Material recovered from these sites is currently being analyzed.

At a huge seacave known as Wolf's Lair on Baker Island, a raft of drift logs has been dated between 3400 and 4400 years old (Dixon et al. 1997). Although the site does not appear to have been occupied extensively, some unusual wooden artifacts were found amidst the ancient flotsam in 1994. These include a 5000 year old cedar implement. The implement is L-shaped and its working end is studded with rockfish spines. Whether it was used as a fish club, meat tenderizer, or fiber comb is unknown. No exact analogs have been identified in ethnographic collections or in the rare archaeological collections of wood artifacts. Also found in Wolf's Lair were two decorated wood planks which once were lashed together. The plank edges have small rectangular holes, two of which contain cedar fasteners that originally joined the two planks together. The planks are AMS-dated between 1200 and 1350 years old. The presence of red ochre and the fragile cordage fragments strongly suggest that these boards were intentionally brought into the seacave, not washed in. After comparing these boards with ethnographic artifacts, they don't look like they are remains of a bentwood box or a wooden shield. They may be "recycled" canoe parts. Our analyses of these artifacts appeared in the 2000 Canadian Journal of Archaeology 24:107-128.

 Two Engraved Planks from Wolf's Lair  

 Kit'n'Kaboodle Cave, located on Dall Island, is a complex solution cave formed by the dissolution of limestone bedrock by acidic groundwater. It has at least four entrances, three of which contain evidence of human occupation. A small rockshelter adjacent to the cave has also seen pre-contact use. A large resurgence flows through the lower levels of the cave, and a red ochre outcrop is found below the site. Limited testing and dating were conducted in 1992 and 1993. In 1996, we tested the site with Forest Service archaeologist Terry Fifield to attempt to answer some questions raised by earlier investigations. At the main entrance, we reached bottom at 110 cm, dated to ca. 3000 years ago. At another entrance, surface shell samples dated to 5300 years ago. We have yet to reach the bottom of the cultural deposit in the rockshelter, but we did recover faunal remains dating to 5100 years ago. In 1998, Erlandson returned to the site with Jane Smith, and found an obsidian microblade core at this part of the site at a 60 cm depth. Microblade tools have been thought to be early Holocene "index fossils" in southeast Alaska until recent studies (Bowers et al. 1996; Moss et al. 1996) have shown that microblade technology continued into the middle Holocene in this region. The deposit here extended to at least 110 cm beneath the surface, but because we did not reach the base of the cultural deposit, we will undoubtedly return to Dall Island. Identification of the faunal remains collected in 1996 is nearly complete, and additional analyses are underway.

Biologist Doug Larson & Moss inspecting surface features, 49-CRG-188, photo by T. Fifield

In 1997, Moss received a small National Science Foundation grant to investigate the 49-CRG-188, a rockshelter site located on Noyes Island. Based on the age of raised beaches located nearby, we had hoped that this site might also contain early Holocene evidence of use of the outer islands. Erlandson and University of Oregon Ph.D. students Mark Tveskov and Robert Losey also took part in this expedition. A cultural deposit nearly 3 m deep was excavated, all of which was deposited within the last 2000 years. Although we didn't find evidence of early occupation, the rockshelter contained a number of distinctive artifacts and remarkably well-preserved faunal and botanical assemblages. Faunal analyses conducted to date indicate that the site residents were expert sea lion, seal, and deer hunters, and fisherfolk who feasted on large salmon, halibut, Pacific cod, lingcod, and rockfish. The marine bird assemblage is also diverse.

For the 49-CRG-188 work, collaborating researchers include Julie Stein (University of Washington) who has studied the geochemistry of the stratigraphic levels within the site. Paleobotanist Dana Lepofsky and her student, Natasha Lyons, (Simon Fraser University) have been studying the plant remains. It is very unusual to recover uncharred plant remains in a Northwest Coast shell midden, but the rockshelter protects the deposits from the region's heavy rainfall and the site's geochemistry favors plant preservation in at least some stratigraphic levels. Identified plants include tree species well known to the region (yellow and red cedar, fir, spruce, lodgepole pine, hemlock, alder, willow, crab apple) but also Douglas fir, a non-local tree. Lepofsky and Lyons found that elderberry seeds were abundant, suggesting summer occupation. This is the same season when conditions are favorable for fishing and sea mammal hunting in the outside waters near the site.

Although people camped in the rockshelter only during certain seasons, at other times, animals used the site. In the Southeast Alaskan caves and rockshelters, we have seen surface evidence of deer that appear to have died naturally, and of shell and bone brought into sites by land otters, minks, and eagles. Black bears and wolves are known to hibernate or den in caves and shelters, and various animals may transport the remains of their prey into these sites. These bones and shells can then be mixed with cultural material left behind by people. With the help of Alaska Fish and Game biologist Doug Larson, Moss identified and mapped surface remains and features that are likely the result of animal activity. The next step is to attempt to identify signatures of non-human activity on the bones themselves, and see if any of the subsurface faunal remains bear such traces. Distinguishing between the natural and cultural origin of the faunal remains at the site will help other coastal archaeologists approach similar problems. Much of this research is in the analytical phase, while additional fieldwork is called for at several sites.

Jon Erlandson & Debby Head examine artifacts photo by T. Fifield

Ulna Awl from 49-CRG-188 photo by T. Fifield

References Cited

Bowers, Peter M., Catherine M. Williams, Robert C. Betts, Owen K. Mason, Russell T. Gould, and Madonna L. Moss 1996 The North Point Site: Archaeological Investigations of a Prehistoric Wet Site at Port Houghton, Alaska. Prepared for USDA Forest Service, Sitka, AK, and Parametrix, Inc, Kirkland, WA, by Northern Land Use Research, Fairbanks, AK.

Dixon, E.J., T.H. Heaton, T.E. Fifield, T.D. Hamilton, D.E. Putnam, and F. Grady 1997 Late Quaternary Regional Geoarchaeology of Southeast Alaska Karst: a Progress Report. Geoarchaeology 12(6):689-712.

Erlandson, Jon M., Madonna L. Moss, and Richard Hughes 1992 Archaeological Distribution and Trace Element Geochemistry of Volcanic Glass from Obsidian Cove, Suemez Island, Southeast Alaska. Canadian Journal of Archaeology 16:89-95.

Moss, Madonna L., Jon M. Erlandson, R. Scott Byram, and Richard E. Hughes 1996 The Irish Creek Site: Evidence for a Mid-Holocene Microblade Component on the Northern Northwest Coast. Canadian Journal of Archaeology 20:75-92.