hawaiian archaeology 4
Volume 4, 1995 Society for Hawaiian Archaeology hawaiian archaeology
Volume 4, 1995
Society for Hawaiian Archaeology
Editor’s Note 3
Sara L. Collins Avifaunal Remains from the Kawailoa Site, O‘ahu Island (BPBM Site 50-OA-D6-62) 4
Frank R. Thomas Excavations at Maunalua Cave, Hawai‘i Kai, O‘ahu 17
Richard Pearson A Brief Report on Test Excavations at the Hawaiian Mission, Honolulu 27
Anne W. H. Garland An Archaeological Predictive Model for the Mission Houses Site in Honolulu and its Value 34
Matthew Spriggs Roadkill Archaeology on Läna‘i: A Historic House Site at Kahemano (State Site 1529) 45
Michael T. Pfeffer Distribution and Design of Pacific Octopus Lures: the Hawaiian Octopus Lure in Regional Context 47 hawaiian archaeology 4
Melinda S. Allen, Editor Production of this volume was made possible with the generous financial assistance of: Robert E. Black Memorial Fund B. P. Bishop Museum Archaeology, Inc. Paul H. Rosendahl, Ph.D., Inc. – Editor‘s Note
This volume marks an important point in the history of Hawaiian Archaeology, as the Society brings to publication the last of its outstanding papers from the late 1980s. Volumes 2, 3, and 4 are witnesses to the commitment of our membership and board to producing a professional and timely journal. I am appreciative of the willingness of Anne Garland, Richard Pearson, Matthew Spriggs, and Frank Thomas to work with me on editorial details, despite the many years between submission and publication.
The papers of Volume 4 reflect three trends in Hawaiian Archaeology. The paper by Collins is an exam- ple of the increasing interest in re-examining previously collected materials, while Thomas synthesizes information from a second previously excavated but poorly known site. Collins takes a fresh look at the avifauna from Kawailoa, O‘ahu, comparing them with finds from other archaeological and paleontolog- ical sites in the region. Notably, this is the first published account of this relatively early occupation site and Collins also summarizes important information on the radiocarbon sequence and non-bird fauna. Likewise, Thomas’ paper makes available stratigraphic, radiocarbon, and artifact data from Maunalua Cave, an important leeward O‘ahu site. Other studies of previously excavated materials are currently underway by University of Hawai‘i and Bishop Museum research teams. Like the contribution of Collins, these efforts underscore both the potential of these older collections to provide new information and the importance of carefully documenting and managing archaeological resources for future studies.
The papers by Pearson, Garland, and Spriggs register an early, and recently accelerated, interest in his- torical archaeology, the period of Hawaiian history after A.D. 1778. The importance historical archaeolo- gy has assumed locally is also indicated by the enthusiastic response to the workshop and symposium organized by Susan Lebo for the 7th Annual Hawaiian Archaeology Conference.
Finally, the paper by Michael Pfeffer, along with recent literature elsewhere, signals renewed interest in traditional material culture studies. Departing from earlier research, Pfeffer examines not only historical aspects of artifact form and distribution related to culture-contact but also functional features of mor- phology which relate to environmental conditions. He offers a regional perspective on the ingenious Pacific octopus lure, and suggests that important design changes in this tool may have occurred in the Hawaiian Islands. M.S.A.
3 Avifaunal Remains from the Kawailoa Site, O‘ahu Island (BPBM Site 50-OA-D6-62)
Sara L. Collins
State of Hawai‘i Historic Preservation Division
A major theme of Pacific prehistory has centered on human agency in the disap- pearance of endemic bird species which once flourished on nearly all of the arch- ipelagos of Oceania. From the smallest and most remote islands, like Tikopia or Henderson (Steadman and Olson 1985; Steadman et al. 1990; Weisler et al. 1991), to the largest island archipelagos of Hawai‘i and New Zealand (Anderson 1989; Olson and James 1984), a growing number of studies have demonstrated the link between human settlement and the extinction or extirpation of avifauna. In the Hawaiian Islands, inferences drawn from research on avifaunal remains from paleontological sites have shaped thinking about the nature of human/bird relationships (Olson and James 1982, 1991). According to these views, the impact of the original Polynesian colonists was immediate and devastating, with the majority of extinctions taking place in the pre-Contact1 era (Olson and James 1984). To a lesser extent, excavations of archaeological sites in Hawai‘i have informed research on avian extinctions by providing additional data on presumed prey species of the Polynesians found in midden and, in some instances, radio- carbon dates of such deposits (James et al. 1987; Schilt 1984). It would seem that avifaunal remains from firmly dated archaeological sites would potentially yield the most significant and detailed information on the scope and pace of bird extinctions and extirpations in Hawai‘i. The goal of this paper is to refine our understanding of human/bird interactions in Hawai‘i by an examination of archaeological data obtained from several pre-Contact habitation sites in the Hawaiian Islands (Fig. 1).
4 5 collins
Materials and Methods
Field Methods and Preliminary Results In 1984, Bishop Museum Anthropology staff con- ducted excavations at site 50-Oa-D6-62, a habita- tion site at Kawailoa on the north shore of O‘ahu Island (Garland n.d., 1984a, 1984b), as shown in Fig- ure 1. Associated with a nearby fishing shrine or ko‘a, Site D6-62 covered an area of approximately 1,152 m2 and had a number of activity areas and unex- pectedly rich and extensive pre-Contact deposits. Features identified at Site D6-62 included a proba- ble cooking area, postmolds, refuse pits, and four human burials (Fig. 2). All excavated deposits were screened through 1/4 in and 1/8 in mesh, with each fraction bagged separately. Figure 1. Map of the Hawaiian Islands, showing location of Kawailoa Site and other archaeological The relatively complex sequence of occupations at localities discussed in text. the Kawailoa Site is represented by the two major cultural layers (called here Layer I and Layer Ib) and the numerous features identified during the excava- tion. The radiocarbon and volcanic glass age deter- minations obtained from fieldwork at Site D6-62 indicate considerable pre-Contact use of the site. As can be seen in Table 1, there appears to be an earlier component to the site, dating to between .. 1280 and 1460, and a later component, which extends from the seventeenth century .. into perhaps the early post-Contact era (circa .. 1800). The vol- canic glass age determinations, as seen in Table 2, corroborate the 14C age determinations, but provide a somewhat more continuous chronology, ranging from .. 1344 to 1674.
Laboratory Methods Laboratory sorting and analysis of midden primarily focused on the feature contents, with the 1/4 in fau- nal bone sampled much more extensively than the 1/8 in fractions, although selected 1/8 in materials were examined. The 1/4 in faunal remains from the southeast quadrants of grid units BB10, CC9, and DD8 were also analyzed. It is estimated that approx- imately 5 to 10 % (by weight) of the total bone mid- den recovered from the Kawailoa Site has been examined to date. The results of bone midden analy-
Figure 2. Base map of south parcel (TMK: 6-1-3:26), sis were originally recorded as weights in order to showing locations of archaeological sites and test compare findings with the results of shell midden excavations (from Athens and Shun 1982). analysis in which the shell is usually weighed. For 6 hawaiian archaeology
Table 1 Radiocarbon Age Determinations from Site 50-Oa-D6-62 Results of Analyses Kawailoa, O‘ahu, Hawaiian Islands
BPBM Provenience 14C Ages in Years Calibrated Range in HRC NO. .. ± 1 Sigmaa Years ..b (2 Sigma) Taxonomic Diversity Table 3 contains a list of the vertebrate faunal taxa HRC 392 T.P. 1/Layer 1 <150 .. <.. 1684 identified from the Kawailoa Site deposits, together HRC 393 T.P. 2/Layer 1 <370 .. <.. 1475 with the common names of the taxa, and Table 4 HRC 811 BB10/ Fe. 1 <170 .. <.. 1678 contains the total NISPs and MNIs for all taxa. In (hearth) general, nearly all vertebrate faunal bones could at least be assigned to zoological class; less than 1% of HRC 812 BB10/ Fe. 4 <180 .. <.. 1675 (postmold) the bone fragments were identified as “Medium Vertebrate,” which could represent bone from a fish, HRC 813 DD7/ Fe. 3 580 ± 80 .. .. 1280–1460 bird, or mammal. Similarly, less than 1% of the (oven) bone midden was probably derived from cartilagi- HRC 814 CC9/ Fe. 1 <150 .. <.. 1684 nous fishes (Sharks and Rays). Bony fish remains (firepit) formed the single largest component of the Kawailoa .. .. HRC 815 CC9/ Fe. 2 150 ± 80 1600–1955 vertebrate faunal midden—83%—while mam- (hearth) malian remains composed about 15% of the total. .. .. HRC 816 CC9/ Fe. 4 600 ± 80 1280–1450 Finally, nearly 2% of the bone was derived from (postmold) avian taxa. Table 5 presents some of these compara- HRC 817 CC9/ Fe. 5 140 ± 90 .. .. 1640–1955 tive statistics. (ash deposit) HRC 818 CC9/ Layer Ib 110 ± 80 .. .. 1670–1955 The NISPs for each zoological class of faunal remains can also be used as measures of diversity. While more HRC 819 DD8/ Fe. 4 210 ± 60 .. .. 1630–1890 (oven) than 6,700 fish bones were identified from these deposits, only about 3% were identified beyond the HRC 820 DD8/ Layer Ia <190 .. <.. 1672 level of class—to family, genus, or species. Approxi- a Dates given are uncalibrated, 13C-adjusted 14C ages (i.e., conventional radiocarbon mately 25% of the avifaunal remains were identified ages). beyond the level of class—to family, genus, or species. b Radiocarbon samples were calibrated with Stuiver and Reimer (1993) CALIB Program Finally, about 31% of the mammalian bone was iden- Version 3.0.2. tified to genus or species. The differences in identifi- cation ratios among the three classes of faunal remains are due to several factors. First, the low NISP of iden- tified bony fish remains is most likely a function of this report, the bone midden was re-analyzed so as incomplete knowledge of fish osteology; the major- to obtain the number of identified specimens per ity of identified fish bones from the Kawailoa Site taxon, or NISP figure for each taxon, and a mini- are mouth parts or distinctive postcranial elements mum number of individuals (MNI), as calculated like the dorsal tang of the acanthurid Naso spp. Sec- from the NISPs (cf. Grayson 1979). In all, approxi- ondly, the relatively high percentage of identifiable mately, 8,100 faunal bones and bone fragments from mammalian remains can be attributed to the fact both the cultural layers and features were examined that a number of otherwise unidentifiable mamma- for this report. In order to simplify reporting for this lian bones were found in features, clearly in asso- brief overview, the results of the vertebrate faunal ciation with obvious dog or pig teeth; as such, these analyses have been combined so as to present all of bones probably represent individual carcasses. the data for the cultural layers and features together. Thirdly, the vast majority (over 90%) of the verte- The author identified specimens by using compara- brate faunal remains are broken and fragmentary, a tive collections in the Departments of Vertebrate state which usually limits the precision of identifica- Zoology and Anthropology at Bishop Museum. Dr. tions, allowing only more general labels of “Pisces” Akira Goto identified the shark specimens. or “Medium Mammal.” 7 collins
Table 2 Volcanic Glass Age Determinations from Site 50-Oa-D6-62 (Uhu or Scaridae), particularly Scarus spp., which Kawailoa, O‘ahu, Hawaiian Islands account for over a third (38%) of the identified fish bone. HHD NO. Provenience Lithic Date in Classification Years .. Margaret Titcomb (1972) notes that the Uhu was a favorite fish of the Hawaiians, prized for its delicious HHD 645 T.P. 1/ Layer I Flake 1451 ± 70 liver as much as for its flesh. Two of the most com- HHD 646 T.P. 1/ Layer I Core 1411 ± 70 mon Scarus species found in Hawaiian waters (S. HHD 952B BB10/ Fe. 6 Diagnostic Flake 1344 ± 60 dubuius and S. perspicillatus) range in length from 36 (postmold) cm to 61 cm, while Calotomus spp. range from 5 cm HHD 953 CC9/ Fe. 4 Diagnostic Fragment 1534 ± 60 to 43 cm in length (Gosline and Brock 1971). Tit- (ash fill) comb (1972) also describes the wrasses (Hïnälea or HHD 954 CC9/ Fe. 4 Diagnostic Flake 1534 ± 80 Labridae) as a popular fish, eaten raw or broiled. HHD 955B CC9/ Layer Ib Diagnostic Flake 1444 ± 80 Thalassoma spp. are usually found in from six to 30 feet of water, and they readily take a hook, making HHD 956 CC28/ Layer Ia Diagnostic Flake 1534 ± 100 them easy prey for anglers; Thalassoma range in size HHD 957 DD7/ Fe. 2 Non-diagnostic 1504 ± 70 from 13 cm to 61 cm (Gosline and Brock 1971). (oven) Flake The Kawailoa midden contains the bones of least HHD 958 DD8/ Layer Ib Diagnostic Flake 1544 ± 40 one other labrid genus—Bodianus spp.—which may HHD 959 DD8/ Fe. 4 Diagnostic Flake 1544 ± 40 be derived from B. bilunulatus or the ‘A‘awa. The HHD 960 DD8/ Fe. 3 Non-diagnostic 1674 ± 60 ‘A‘awa is relatively large, reaching lengths of 61 cm (refuse pit) Fragment or more, and is caught by hook and line in waters up HHD 961 DD7/ Fe. 3 Non-diagnostic 1534 ± 80 to 100 feet deep (Gosline and Brock 1971). Finally, (oven) Fragment the surgeonfish or acanthurid genus Naso accounts for 15%, by NISP, of the identified Kawailoa fish remains. The Naso spp. bone probably comes from the fish called Kala which, Titcomb (1972) points Nonetheless, the numbers of taxa identified are com- out, was also a popular fish among Hawaiians, in parable with other nearby pre-Contact occupation spite of its strong smell. sites on O‘ahu—the rockshelters from the Anahulu River Valley, to the west and south of the Kawailoa Over a fifth (22.5% or 267/1186) of the mammalian Site (Fig. 1). As seen in Table 6 the number of fish, bone found in the Kawailoa midden comes from avian, mammalian taxa found in the vertebrate fau- Domestic Dog and Pig (see Table 4). Of the mam- nal midden from the three Anahulu rockshelters are malian bone identified to genus or species (NISP = comparable to the Kawailoa faunal midden (data 368), dog and pig account for nearly 73% of these taken from Kirch and Collins 1989). remains. In the absence of a final report, the prelim- inary NISP and MNI figures from Table 4 must Fish and Mammalian Taxa serve as an accounting of the number of pigs and The fish taxa identified from the Kawailoa midden dogs represented in the analyzed portions of the are all inshore reef species that could have been Kawailoa midden. It is possible that once the vertical found in waters adjacent to the site. Although, as and horizontal proveniences of the Kawailoa Site are noted above, only 3% of the fish bone could be iden- better understood, the MNIs for these species as seen tified beyond the level of class, 175 specimens repre- in Table 4—17 for the dog and 23 for the pig—may senting at least 129 individuals in seven families of be somewhat high. bony fishes were recorded (see Table 4). In general, fish from three families—Labridae, Scaridae, and Avifaunal Taxa Acanthuridae— account for over 90% of the identi- The bird bones found in the Kawailoa midden rep- fied fish. Judging from the NISPs, the most preva- resent at least seven avian species, three of which are lent fish found in the midden are the Parrotfishes currently endangered or extirpated, one of which was 8 hawaiian archaeology
Table 3 List of Vertebrate Faunal Taxa Identified from Kawailoa Site Midden (50-Oa-D6-62)
Taxon Definition
Medium vertebrate Bone from a vertebrate of unknown class Chondrichthyes Class of Cartilaginous Fishes (Sharks and Rays) Shark Pisces Class of Bony Fishes Sparidae, Monotaxis grandoculis Family Sparidae, Porgy (Mü) Labridae Wrasse Family Labridae, Thalassoma sp(p). Wrasse Family, genus Thalassoma Labridae, Bodianus sp(p). Wrasse Family, genus Bodianus Scaridae, Calotomus sp(p). Parrotfish Family, genus Calotomus Scaridae, Scarus sp(p). Parrotfish Family, genus Scarus Acanthuridae Surgeonfish Family Acanthuridae, Naso sp(p). Surgeonfish Family, genus Naso Balistidae Triggerfish Family Monocanthidae, Pervagor sp(p). Monocanthid, genus Pervagor Diodontidae Spiny Puffer Family Small aves Bird the size of a small songbird Medium aves Bird the size of a chicken Medium Procellariid Mid-sized shearwater or petrela Pterodroma phaeopygia Hawaiian Petrel (Ua‘u) Pterodroma hypoleuca Bonin Petrel Puffinus newelli Hawaiian Shearwater (‘A‘o) Gallus gallus Red Junglefowl (Moa) Small rallid Small rail Myadestes cf. oahensi O‘ahu Thrush (‘Ämaui ) Moho apicalis O‘ahu ‘Ö‘ö Medium mammal Mammal the size of a dog or small pig Rattus sp(p). Rat boneb Rattus exulans Polynesian Rat (‘Iole) Mus musculus House Mouse Canis familiaris Domestic Dog (‘Ilio) Sus scrofa Domestic Pig (Pua‘a)
a The Medium Procellariid bone is most likely derived from one of the three medium-sized procel- lariids identified in the Kawailoa midden: the Hawaiian Shearwater, the Bonin Petrel, or the Hawaiian Petrel. b The rat bone referred to genus Rattus is most likely derived from the Polynesian Rat (R. exulans). 9 collins
Table 4 Total NISPs and MNIs for Vertebrate Faunal Remains at the Kawailoa Site (50-Oa-D6-62) (Layers and Features Combined)
Taxon NISP MNI
Medium vertebrate 20 Chondrichthyes 1 Shark 19 11 Pisces 6593 Sparidae, M. grandoculis 11 Labridae 17 11 Labridae, Thalassoma sp. 18 12 Labridae, Bodianus sp. 7 7 Scaridae, Calotomus sp. 18 14 Scaridae, Scarus sp. 66 51 Acanthuridae 7 6 Acanthuridae, Naso sp. 27 20 Balistidae 1 1 Monocanthidae, Pervagor sp. 2 2 Diodontidae 11 4 Small aves 9 Medium aves 96 Medium Procellariid 20 Pterodroma phaeopygia 22 Pterodroma hypoleuca 33 Puffinus newelli 11 Gallus gallus 65 Small rallid 1 1 Myadestes cf. oahensi 11 Moho apicalis 11 Medium mammal 818 Rattus sp. 6 Rattus exulans 89 23 Mus musculus 64 Canis familiaris 54 17 Sus scrofa 213 23
TOTALS 8134 221 10 hawaiian archaeology
Table 5 Comparative Data on Faunal dence from recently examined paleontological and Identifications from the Kawailoa archaeological sites on O‘ahu and other islands, Site (50-Oa-D6-62) however, it is quite possible that in the past, during the pre-Contact era, local populations of these and Zoological Class NISP % of Total NISP other seabird species existed in, or near, the vicinity Unknown 20 < 1% of the Kawailoa Site (see James et al. 1987; Kirch and Collins 1989; Olson and James 1982). Conse- Chondrichthyes 20 < 1% quently, these local populations may have been of Pisces 6768 83% sufficient size and stability to serve as reliable food Aves 140 2% sources. Of further interest is the fact that eight to Mammalia 1186 15% ten Medium Procellariid elements show evidence of charring; these elements are from cultural Layers I TOTALS 8134 100% and Ib (in Unit DD8) and features (Units BB10/Fe. 6 and DD8/Fe. 4 ). The relatively high percentages of identified procellariid remains among the Kawai- Table 6 Numbers of Identified Vertebrate loa avifauna (18.6% of all bird bone, 74% of all Taxa from Anahulu Rockshelters and Kawailoa avian identifications), plus the presence of charred Site procellariid elements, indicate some form of regular Site Mammals Birds Fish Totals predation.
Anahulu D6-36 3 5 7 15 The introduced avian species is the Red Junglefowl Anahulu D6-58 4 6 7 17 (Moa or Gallus gallus) which was brought to the Hawaiian Islands by Polynesian settlers, along with Anahulu D6-60 6 11 7 24 the dog, pig, and Polynesian Rat. The Red Jungle- Kawailoa D6-62 5 7 11 fowl is represented by six elements: two from one provenience within cultural Layer I, and four from introduced to the Hawaiian Islands in pre-Contact features. Despite the small numbers of Moa, we can times by the original Polynesian settlers, and three of say that the species also provided food since three of which are presumed extinct. The Hawaiian Petrel the six identified elements showed evidence of char- (‘Ua‘u or Pterodroma phaeopygia) is represented by ring. These elements came from features in Units two right ulnae, each from a different feature. The CC9 and DD8. It is possible that at least some of the Bonin Petrel (Pterodroma hypoleuca) is represented 96 elements assigned to the general category of by three elements (three ulnar portions), each from a “Medium Aves” are additional Moa remains. different provenience within Layer I. The Newell The Kawailoa avifauna also included bones of at Shearwater (‘A‘o or Puffinus newelli) is represented least three endemic bird species, all of them probably by a single element, a left ulna, from a feature. Of now extinct. One species is represented by a distal the 20 bone or bone fragments assigned to “Medium left humerus, which comes from a small rail (Family Procellariid,” eight were found in various prove- Rallidae), recovered from a feature in Unit CC9. niences within cultural Layers I and Ib, and 12 come Historically, two species of rails were known from from features. All but two of the Medium Procel- the Hawaiian Islands—the Laysan Rail (Porzana lariid bones were wing elements. palmeri) and the Hawaiian Rail (Porzana sandwich- Presumably, the Medium Procellariid remains are ensis)—both of which were flightless and are now derived from one or more of the three shear- extinct (Olson and James 1991). Paleontological and water/petrel species identified in the midden. All archaeological deposits in the Hawaiian Islands have three species are now locally threatened, endangered, yielded the remains of nine or ten additional species or extirpated, with breeding populations restricted to of flightless rails, all of them apparently extinct offshore islets or the northwest Hawaiian Islands (see before the post-Contact era (Olson and James 1991). Pratt et al. 1987; Pyle 1992). Given mounting evi- The rallid postcranial element from the Kawailoa 11 collins
Table 7 Comparative Data on Significant Avian Taxa from Table 8 Comparative Chronologies the Kawailoa, Anahulu, and Kahulu‘u Archaeological Sites of Five Archaeological Sites
Taxon D6-62 D6-36 D6-58 D6-60 7702 SITE Earliest Terminal Occupation Occupation PROCELLARIIDAE .. .. Small Procellariid X X D6-62 (Kawailoa) 1280–1460 1670–1955 .. Medium Procellariid X X X X X D6-36 (Anahulu) 1720–1820 — .. .. Pterodroma phaeopygia XXXX D6-58 (Anahulu) 1220–1520 1460–1480 .. .. Pterodroma hypoleuca XX XX D6-60 (Anahulu) 1390–1480 1630–1955 a .. .. Bulweria bulwerii XXX7702 (Kahulu‘u) 1420–1650 1660–1955 Puffinus pacificus X a Earliest date of .. 670–1180 omitted by Hay et al. Puffinus newelli XXX (1986) HYDROBATIDAE Oceanodroma castro X ARDEIDAE Nyctocorax nyctocorax XX PHASIANIDAE Gallus gallus XXXXX RALLIDAE Small rallid X X LARIDAE Stercorarius pomarinus X STRIGIDAE Asio flammeus X MUSCICAPIDAE Muscicapid X X X cf. Myadestes oahuensi X cf. Myadestes obscurus X MELIPHAGIDAE cf. Moho apicalis X cf. Moho nobilis X cf. Chaetoptila spp. X DREPANIDINAE Drepanid X cf. Psittirostra spp. X cf. Hemignathini, X Loxops spp. 12 hawaiian archaeology
Site is from a bird smaller than the Hawaiian Rail and the Kahulu‘u Habitation Cave (State Site No. and may possibly be derived from what Olson and 50-10-37-7702) on Hawai‘i Island (see Fig. 1). Table James (1982) have called the “small Oahu rail,” and 7 contains summary information on the avifaunal taxa subsequently named Porzana ziegleri (Olson and recorded from middens excavated from the five James 1991). archaeological sites.
The other two endemic bird species found in the The Anahulu rockshelter sites were excavated in Kawailoa avifaunal remains are also each represented 1982, during a Bishop Museum project sponsored by by one element. The right ulna of a muscicapid the National Science Foundation; a complete series (Family of Old World Flycatchers)—referred to the of site reports, with ancillary specialist reports, have O‘ahu Thrush (‘Ämaui or Myadestes oahensi)—was been published elsewhere (Kirch 1989a; Kirch and recovered from a feature in Unit CC9. The ‘Ämaui Sahlins 1992). Dated deposits from the three Ana- is now extinct, although it was apparently present on hulu sites listed above in Table 6 (D6-36, D6-58, O‘ahu as late as the 1820s (Pratt et al. 1987). In and D6-60) contained the remains of extinct, extir- addition to the thrush bone, the left humerus of a pated, and endangered bird species. As such, these melaphagid (Family of Honeyeaters) came from sites provide a valuable source of comparative data for Layer Ib in Unit DD8. This element has been the Kawailoa Site. referred to the extinct O‘ahu ‘Ö‘ö (Moho apicalis) rather than the larger genus Chaetoptila. The latter One radiocarbon age determination was available 14 was known historically only from Hawai‘i Island (as from Site D6-36. The conventional C age was 240 .. the Kioea), although fossil specimens of Chaetoptila ± 50 (Kirch 1989b); the calibrated date range, at .. were found at Barbers Point, O‘ahu (see Olson and the two sigma value, is 1720–1820 (after Stuiver James 1982). and Reimer 1993). The D6-36 midden contained elements attributed to the Bulwer Petrel (‘Ou or Bul- weria bulwerii), the Bonin Petrel, and unidentified medium-sized petrels or shearwaters. A single bone Discussion and Conclusions referable to the Family Muscicapidae (Old World Flycatchers) came from midden deposits at D6-36. Comparative Data from Other Hawaiian There are two endemic muscicapids known from Archaeological Sites O‘ahu from which the D6-36 specimen could have Since the mid-1970s, the understanding of Hawai- been derived. One is the ‘Elepaio (Chasiempis sand- ian avian paleontology has increased in an exponen- wichensis) which inhabits native forests and still is tial fashion, largely due to the results of paleontolog- one of the most commonly seen endemic birds on ical excavations conducted by staff of the Bishop O‘ahu. The other endemic muscicapid is the ‘Ämaui Museum and the National Museum of Natural His- (Myadestes oahensi) which was known historically tory, Smithsonian Institution. Concurrently, the from O‘ahu until the 1820s (see Pratt et al. 1987). vertebrate fauna of archaeological middens have Anahulu Site D6-58 yielded two valid radiocarbon been more thoroughly examined, and in the case of age determinations; the conventional 14C ages were older excavations re-examined, for the presence of 600 ± 110 .. and 280 ± 60 .. (Kirch 1989b). extinct or extirpated avian species. Consequently, The calibrated dates, at two sigma, were .. 1220– there now exists a small but growing number of pre- 1520 and .. 1460–1480 (after Stuiver and Reimer Contact archaeological sites in which extinct avifau- 1993). Like D6-36, Site D6-58 faunal midden con- na have been found in securely dated deposits. In this tained the remains of several procellariid taxa, includ- section, vertebrate faunal data obtained by the author ing the Hawaiian Petrel, the Wedge-Tailed Shear- from midden analyses of several dated Hawaiian water (‘Ua‘u kani or Puffinus pacificus), and the archaeological sites will be compared with the fore- Newell Shearwater. In addition, the indigenous going information from the Kawailoa Site. The sites Black-crowned Night Heron (‘Auku‘u or Nyctocorax to be used include the three rockshelters from nearby nyctocorax) was represented in the D6-58 midden. Anahulu Valley on O‘ahu Island mentioned above, 13 collins
Perhaps the most interesting avifaunal specimen from ing information is taken from the report of this exca- D6-58 was a left tibiotarsus referred to the genus vation prepared by Deborah Hay (Hay et al. 1986). Psittirostra of the Hawaiian Honeycreeper Sub-Family The excavated deposits of the Kahulu‘u Habitation (Drepanidinae). The only member of the genus Psit- Cave yielded extensive amounts of vertebrate faunal tirostra known historically from O‘ahu was the ‘Ö‘ü (P. midden; in addition, many of the recovered artifacts psittacea), now extinct. Olson and James (1982) have were of worked fish, bird, and mammal bone. Sev- also reported the fossil bones from at least one addi- enteen radiocarbon dates and 44 volcanic glass dates tional species of Psittirostra found in paleontological were obtained from the Kahulu‘u Site, with their deposits on O‘ahu; the D6-58 specimen cannot be ranges generally being complementary. As reported positively referred to either of the two species. by Hay et al. (1986), the radiocarbon dates range from the fifteenth century .. to 1950, with one Anahulu Site D6-60 had three radiocarbon age early date of .. 670–1180 which the authors con- determinations, and the uncalibrated conventional sider to be aberrant. The volcanic glass dates provide 14 .. C ages were as follows: 460 ± 70 , 190 ± 60 a similar but tighter chronology, ranging from the .. .. , and 180 ± 80 (Kirch 1989b). The calibrat- fifteenth through the eighteenth centuries .. (Hay .. ed date ranges, at two sigma, are, respectively, et al. 1986). 1390–1480, .. 1650–1890, and .. 1630–1955 (after Stuiver and Reimer 1993). D6-60 midden Like the Kawailoa Site and the Anahulu rockshelters, contained the remains of several petrel species: the the Kahulu‘u Habitation Cave yielded a number of Hawaiian Petrel, the Newell Shearwater, the Bulwer significant avifaunal remains. Numerous petrel bones Petrel, and the Bonin Petrel. In addition, bones from were present, with at least three species represented: the indigenous Black-crowned Night Heron and a the Bonin Petrel, the Bulwer Petrel, and the Har- regular migrant species of gull, the Pomarine Jaeger court Storm Petrel (‘Akë‘akë or Oceanodroma castro) (Stercorarius pomarinus), were found in the D6-60 of the Storm Petrel Family (Hydrobatidae). In addi- deposits. Perhaps the most significant of the identi- tion to these identified petrel species from the mid- fied avifaunal remains from D6-60 are the anterior den remains, at least 39 bone artifacts were attribu- mandible referred to the genus Chaetoptila of the ted to the taxon “Medium Procellariid,” which cor- Honeyeater family and the two humeri from indivi- responds to a petrel or shearwater in the size range of duals in the genus Loxops of the Slenderbilled Honey- the Hawaiian Petrel, Wedge-tailed Shearwater, or creeper Tribe (Drepanidinae, Hemignathini). The Newell Shearwater (Hay et al. 1986). Bones of the only historically known member of the genus Chae- Domestic Chicken were present at Kahulu‘u, as were toptila is the Kioea (C. angustipluma), last reported remains of the Hawaiian Owl (Pueo or Asio flam- from Hawai‘i Island in the 1860s (Pratt et al. 1987). meus). Like the Kawailoa Site, the Kahulu‘u Habita- Recently, specimens referable to the genus Chaetop- tion Cave deposits contained an element from a tila have been discovered in paleontological deposits small flightless rail. It is possible that the specimen is from O‘ahu (Olson and James 1982). The genus derived from the historically known Hawaiian Rail Loxops currently includes the ‘Äkepa (L. coccineus), (Moho or Porzana sandwichensis), now extinct. The now presumed extinct on O‘ahu (Pyle 1992). For- Kahulu‘u rail bone may also be derived from anoth- merly, the ‘Amakihi (Hemignathus virens) belonged er small rail, described by Olson and James (1991) as to the genus Loxops but has now been re-assigned, Porzana sp. Small rallid remains have been found in although James and Olson (1991) dispute this re- at least four west Hawai‘i Island archaeological sites assignment on morphological grounds. Olson and (50-Ha-D6-41, -D7-27, -D8-30, and -D8-33), nine James (1982) reported the presence of fossil speci- archaeological sites in the Pohakuloa Training Area, mens referred to Loxops spp. and Hemignathus spp. and in Site 50-Ha-D24-22 from the Mauna Kea from paleontological sites on O‘ahu. region in Hämäkua (50-Ha-D24-22) (Athens and Kaschko 1989; Collins 1984; Olson and James 1991). The Kahulu‘u Habitation Cave (State Site No. 50- 10-37-7702) was excavated by personnel from Paul At least three passeriform families or subfamilies are H. Rosendahl, Ph.D. Inc. (PHRI), and the follow- represented in remains from the Kahulu‘u midden: 14 hawaiian archaeology
the Muscicapidae, the Meliphagidae, and the Drepa- populations of some size for at least two petrel/shear- nidinae. There is bone from Kahulu‘u which can be water taxa (cf. Olson and James 1982; Collins, In attributed to a muscicapid, but without assignment press). Thus, it is not surprising to find that the pro- to genus or species. Given that there is one specimen cellariiforms are well-represented in archaeological from Kahulu‘u Site which is probably derived from sites. the Hawai‘i Island Thrush (‘Öma‘o or Myadestes obscurus) it is possible that the muscicapid bone may Equally significant is the fact that avifaunal-bearing also be from this species. The meliphagid bone from deposits from all five archaeological sites discussed Kahulu‘u Habitation Cave can be referred to the here seem to date from throughout the pre-Contact Hawai‘i ‘Ö‘o (Moho nobilis), but the drepanid bone era—that is, from earlier through later periods. As can only be referred to sub-family level. Table 8 shows the beginning and terminal dates for each of the sites spans at least 200–300 years, and more in some cases. The significance of these tempo- ral data is that a number of now-extinct or fossil spe- Conclusion cies apparently co-existed with Polynesian popula- tions for at least a thousand years into the human The foregoing discussions of avifaunal specimens era, given an initial colonization of the Hawaiian from the Kawailoa Site, and similar remains from Islands around .. 400 (Kirch 1985). The small other archaeological sites on O‘ahu and Hawai‘i NISPs given here make it difficult to deduce popu- Islands, provide a pattern of human use of bird lation size from the available archaeological data but, species in pre-Contact Hawai‘i. As the data in Table as suggested above, the suite of avifauna in archaeo- 7 make clear, a wide range of avian species was logical sites parallels what has been observed in pale- exploited, apparently both for food and non-food ontological deposits. A basic tenet of Hawaiian avian purposes. The predominance of petrel/shearwater paleontology has been the irrevocable changes species in all of the sites’ deposits probably indicates wrought by the Polynesian settlers whose alteration the relatively greater size and stability of these avian of landscapes and native floras, in particular, are con- populations over time, since the occupation or use of sidered to have had the most devastating effect on each of the five sites seems to span at least 200–300 endemic faunas (Olson and James 1991). If the tem- years. The comparatively smaller numbers of non- poral and faunal data from the Kawailoa, Anahulu, procellariiform taxa (particularly the rallids and pas- and Kahulu‘u sites can be taken at face value, then it seriforms) may indicate either limited local popula- would appear that endemic and indigenous Hawaiian tions of these species or a more remote catchment birds were not driven into extinction or extirpation at area. Another recent study of archaeological middens a precipitous rate but, in fact, persisted well into the from Hawai‘i Island also provides similar figures. Polynesian era. Further study of the Kawailoa depos- Helen James (1989) found that 74% of the avian its—which promise to yield substantially more avi- remains from the Pohakuloa Training Area sites were faunal remains—as well as examination of the avi- procellariid. faunal data already available from other published Although there is no conclusive study of the bio- Hawaiian archaeological site reports will make clear- geography and demography of the Hawaiian avifau- er the pace and scope of avian extinctions in Hawai‘i. na, the data from major paleontological deposits— like Barbers Point, O‘ahu—are similar to what is seen in these archaeological sites in that procellarii- Note forms are by far the most numerous taxa, especially when statistical techniques like rank-order correla- 1. Throughout this paper, the word “Contact” refers tions are applied to the NISP data from faunal analy- to the first documented contact between Europeans ses (Collins, In press). Furthermore, the numerous and Native Hawaiians—that is, the visit of Captain procellariiform remains from Barbers Point sites also James Cook and his crews in .. 1778. include juvenile individuals, which suggests breeding 15 collins
Garland, A. n.d. Draft final report on excavations Acknowledgements at 50-Oa-D6-62 Kawailoa, O‘ahu. MS., on file, Department of Anthropology, B. P. Bishop Museum, I would like to thank Ms. Anne Garland for the Honolulu. opportunity to examine the Kawailoa faunal remains, as well as her willingness to discuss the excavations _____. 1984a. A preliminary report on intensive and their results. I would also like to thank staff of archaeological data recovery at a coastal parcel on the the Bishop Museum for assisting in the production Kawailoa-Waimea boundary in Waialua, O‘ahu. MS., of this paper, particularly in allowing the use of the on file, Department of Anthropology, B. P. Bishop unpublished data and providing copies of the site Museum, Honolulu. maps used. Finally, I would like to thank Dr. M. S. _____. 1984b. Second preliminary report on archae- Allen, also of Bishop Museum, for her thoughtful ological excavations at Site 50-Oa-D6-62, Kawailoa, and attentive editing, and Dr. M. I. Weisler (His- Waialua, O‘ahu Island. MS., on file, Department of toric Preservation Office, Republic of the Marshall Anthropology, B. P. Bishop Museum, Honolulu. Islands) for critical review. All remaining errors of fact or interpretation are my responsibility. Gosline, W. A., and V. E. Brock. 1971. Handbook of Hawaiian fishes. Honolulu: University Press of Hawai‘i. References Grayson, D. K. 1979. On the quantification of ver- tebrate archaeofaunas. In Advances in archaeological Anderson, A. 1989. Prodigious birds, moas and moa- method and theory, vol. 2, ed. M. B. Schiffer, 199– hunting in prehistoric New Zealand. Cambridge: Cam- 237. New York: Academic Press. bridge University Press. Hay, D., A. E. Haun, P. H. Rosendahl, and C. J. Athens, J. S., and M. W. Kaschko. 1989. Prehistoric Severance. 1986. Kahuluu data recovery project. Exca- upland bird hunters: Archaeological inventory sur- vations at Site 50-10-37-7702, Kahuluu habitation vey and testing for the MPRC Project area and the cave. Land of Kahuluu, North Kona, island of Hawai‘i. Bobcat Trail Road, Pohakuloa Training Area, island Kurtistown, HI: Paul H. Rosendahl, Ph.D., Inc., of Hawai‘i. Prepared for U.S. Army Engineer Dis- Report 61-022084. trict, Pacific Ocean Division, Fort Shafter by Inter- national Archaeological Research Institute, Inc., James, H. 1989. Bird bone analysis. In Prehistoric Honolulu. upland bird hunters: Archaeological inventory sur- vey and testing for the MPRC Project area and the Athens, J. S., and K. Shun. 1982. Archaeological test Bobcat Trail Road, Pohakuloa Training Area, island excavations and mapping near Waimea Bay, O‘ahu of Hawai‘i, by J. S. Athens and M. W. Kaschko, (TMK 6-1-3:26). MS., on file, Department of Appendix B, 242–87. Prepared for U.S. Army Engi- Anthropology, B. P. Bishop Museum, Honolulu. neer District, Pacific Ocean Division, Fort Shafter by Collins, S. L. 1984. Analysis of vertebrate faunal International Archaeological Research Institute, remains. In Subsistence and conflict in Kona, Hawai‘i: Inc., Honolulu. An archaeological study of the Kuakini Highway realign- James, H., and S. L. Olson. 1991. Descriptions of ment corridor, ed. R. Schilt, Report 2, 333–54. thirty-two new species of birds from the Hawaiian Departmental Report Series, Report 84-1. Honolulu: Islands: Part II. Passeriformes. Ornithological Mono- Department of Anthropology, B. P. Bishop Museum. graphs no. 46. Washington, D.C.: American Orni- Collins, S. L. In press. Analysis of vertebrate remains thologists’ Union. from Barbers Point, Southwestern O‘ahu. In Archae- James, H., T. W. Stafford, D. W. Steadman, S. L. ology and paleontology at Barbers Point, by B. Davis, Olson, P. S. Martin, A.J.T. Jull, and P. C. McCoy. Technical Report 4. Honolulu: Department of 1987. Radiocarbon dates on bones of extinct birds Anthropology, B. P. Bishop Museum. 16 hawaiian archaeology
from Hawai‘i. Proceedings of the National Academy of _____. 1991. Descriptions of thirty-two new species of Sciences, United States of America 84:2350–54. birds from the Hawaiian Islands: Part I. Non-Passeri- formes. Ornithological Monographs no. 45. Washing- Kirch, P. V. 1985. Feathered gods and fishhooks: An ton, D.C.: American Ornithologists’ Union. introduction to Hawaiian archaeology and prehistory. Honolulu: University of Hawaii Press. Pratt, H. D., P. L. Bruner, and D. G. Berrett. 1987. The birds of Hawaii and the tropical Pacific. Prince- Kirch, P. V.1989a. Introduction. In Prehistoric Hawai- ton: Princeton University Press. ian occupation in the Anahulu Valley, O‘ahu Island: Excavations in three inland rockshelters, ed. P. V. Kirch, Pyle, R. 1992. Checklist of the birds of Hawaii— 1–7. Contributions of the University of California 1992. ‘Elepaio 52(8):53–62. Archaeological Research Facility no. 47. Berkeley: Department of Anthropology, University of Cali- Schilt, R. Subsistence and conflict in Kona, Hawai‘i: fornia. An archaeological study of the Kuakini Highway realign- ment corridor. Departmental Report Series, Report _____. 1989b. The rockshelter excavations. Prehis- 84-1. Honolulu: Department of Anthropology, toric Hawaiian occupation in the Anahulu Valley, B. P. Bishop Museum. O‘ahu Island: Excavations in three inland rockshelters, ed. P. V. Kirch, 9–41. Contributions of the Univer- Steadman, D. W. and S. L. Olson. 1985. Bird sity of California Archaeological Research Facility no. remains from an archaeological site on Henderson 47. Berkeley: Department of Anthropology, Uni- Island, south Pacific: Man-caused extinctions on an versity of California. uninhabited island. Proceedings of the National Aca- demy of Science, United States of America 82:6191–5. Kirch, P. V., and S. L. Collins. 1989. Faunal assem- blages of the Anahulu Rockshelter sites. In Prehis- Steadman, D. W., D. S. Pahlavan, and P. V. Kirch. toric Hawaiian occupation in the Anahulu Valley, 1990. Extinction, biogeography, and human exploi- O‘ahu Island: Excavations in three inland rockshelters, tation of birds on Tikopia and Anuta, Polynesian ed. P. V. Kirch, 61–72. Contributions of the Univer- outliers in the Solomon Islands. Bishop Museum sity of California Archaeological Research Facility Occasional Papers 30:118–53. no. 47. Berkeley: Department of Anthropology, Uni- Stuiver, M., and P. J. Reimer. 1993. Extended 14C versity of California. data base and revised CALIB 3.0 14C age calibration Kirch, P. V., and M. Sahlins. 1992. Anahulu: The program. Radiocarbon 35:215–30. anthropology of history in the Kingdom of Hawai‘i. Titcomb, M. 1972. Native use of fish in Hawaii. Vol. II, The archaeology of history. Chicago: The Uni- Honolulu: University Press of Hawaii. versity of Chicago Press. Weisler, M. I., T. G. Benton, M. de L. Brooke, P. J. Olson, S. L. and H. James. 1982. Prodromus of the Jones, T. Spencer, and G. Wragg. 1991. The Pit- fossil avifauna of the Hawaiian Islands. Smithsonian cairn Islands scientific expedition (1991–1992): Contributions in Zoology no. 365. Washington, First results, future goals. Pacific Science Association D.C.: Smithsonian Institution Press. Information Bulletin 43(3/4):4 –8. _____. 1984. The role of the Polynesians in the extinction of the avifauna of the Hawaiian Islands. In eds. P. S. Martin and R. G. Klein, Quaternary Extinctions: a Prehistoric Revolution, 768–80. Tuscon, AZ: University of Arizona Press. Excavations at Maunalua Cave, Hawai‘i Kai, O‘ahu
Frank R. Thomas
University of Hawai‘i at Mänoa
In 1962, Lloyd Soehren of Bishop Museum and Wilhelm Solheim of University of Hawai‘i discovered a small overhang shelter (Site O-5), situated on Bishop Estate land in the Kaiser Hawai‘i Kai development (Kaluanui Ridge). Site O-5 (State Site 50-80-15-2908) is located near marshy grounds where taro may have been grown, and overlooks Kuapä pond. Preliminary survey of the shelter and adjacent areas revealed thick deposits with abundant shell (exposed by bulldozing activities), as well as artifacts. A decision was made to excavate the site in its entire- ty, as destruction by the housing development appeared imminent. Work at the site was performed between March 1962 and July 1963, in part as a University of Hawai‘i field school under the direction of Wilhelm Solheim. A report was initi- ated by one of the project’s participants, Colin Smart (1965) of the Australian National University, but was never completed. Donn Bayard (1965), then of the University of Hawai‘i, described the artifacts. The site was briefly described dur- ing a recent survey by Beggerly and McNeill (1985). At the suggestion of Matthew Spriggs, I prepared the following summary of the excavations, based on manu- scripts of Smart (1965) and Bayard (1965).
Objectives
Smart’s primary objective in compiling the excavation data was to construct a chronology for southeastern O‘ahu. Earlier excavations at Kuli‘ou‘ou Shelter, O‘ahu (Site O-1), and three other cave sites in the area, provided the only mean- ingful reference points for future work. Further excavations on the dry leeward side
17 18 hawaiian archaeology
recorded in terms of this system. All artifacts and F grid N sample materials are identified by excavation square and layer. Over two-thirds of the shelter deposit was E W E excavated by arbitrary 10 cm levels.
D The most complete and undisturbed stratigraphic S sequence was preserved in the south wall of Square C D.5. Examination of this particular sequence, and comparison with the other well-recorded and/or B photographed sections, provided the basis for the stratigraphic designations used in analysis of the A recovered materials. 0 12345678910
Figure 1. Plan view of shelter floor and excavation Stratigraphy grid.
The following description of stratigraphy is largely of the island were judged to be particularly impor- based on Colin Smart’s 1965 draft, which can be tant for obtaining a more com-plete picture of consulted along with the field notes in the Archaeo- human adaptation in a seemingly marginal environ- logy Lab, Anthropology Department, University of ment. Moreover, because of the site’s depth, investi- Hawai‘i at Mänoa. gators were hoping to identify an early settlement, comparable to the circa eleventh century .. occu- The deposit at Site O-5 was finely stratified. Various pation found at Kuli‘ou‘ou Rockshelter (Emory and mixtures of white ash, brown earth, and black char- Sinoto 1961). coal provided strong color contrasts between many of the layers, while varying textures and degrees of compaction further distinguished a number of them. Methods In some areas, a series of thin, horizontal, and clear- ly defined lenses were preserved. Elsewhere, the stra- tigraphy had been disturbed by later activities, with The shelter was marked off in 1 m grid squares, with pits and depressions marked by a uniform grey fill. each square designated by a letter and number code, beginning at the southwest corner of the site (Figure With the available information, it is possible to infer 1). This resulted in 6 trenches (A through F), cover- at least five periods of occupation. The evidence ing the totality of the shelter floor beneath the over- comes primarily from compact grey-brown layers hang. Altogether, approximately 24 m2 of sediment yielding little or no midden material. The distribu- was excavated within the shelter. No artifacts were tion of these sterile grey-brown layers across the found in Trench A or in the small number of units entire shelter floor (except where there was distur- excavated in Trench F (Bayard 1965). bance), together with their relatively uniform thick- ness, are strong arguments in favor of sterile deposi- The strategy employed in excavating the shelter was tion by aeolian action when the site was uninhabit- initially to remove the sediments in arbitrary 10 cm ed. Aeolian erosion and redeposition in the shelter levels, followed by removal of the sediments by nat- most likely resulted from vegetation clearance in the ural layers. Only squares D.4-7 and E.4-6 were exca- surrounding area (Egler 1947). vated by natural layers, after the stratigraphy had been clearly exposed in section on the south side of Layer designations are given as in the original exca- these units, along Datum Line D (Figure 1). These vation notes, identified alphabetically. Major strati- stratigraphic layers were examined and distinguished graphic units are in some cases divided into one or by a series of letters and numbers (e.g., Layers E.1, more subdivisions, as for example Layer X.1 and E.2, F.1, F.2, etc.). Subsequent excavations were X.2. The matrix descriptions are those provided by 19 thomas
Smart (1965) and in the absence of curated sediment Layer C (circa 50–60 cmbd) Solheim’s excavation samples, no attempt has been made to alter his ter- notes showed a single Layer C, but Smart later dis- minology. Starting at the surface, the stratigraphic tinguished an apparent break in this layer in Square sequence of the rockshelter was as follows. D.5. Layer C is characterized as a “very compact, grey-brown matrix with a marked prismatic struc- Layer X (ca. 0–20 cm below datum) Two or ture” and yielded minimal shell and artifacts. All more “light, dusty grey-brown” layers with little pits, “holes”, and the four burials penetrated this midden and artifacts constitute Layer X. The upper layer or lay immediately above it. portion of Layer X (e.g., X.1) was, in all likelihood, blown into the shelter during recent times, as it was Layer D (circa 50–60 cmbd) Layer D is subdivid- not compact. The subdivision Layer X.2 was based ed into 2 subunits, D.1 and D.2. Layer D.1 consist- on its more compact nature, which could have been ed of an ash matrix with shell and other materials in produced either by a transient occupant or extreme- some quantity. The lower layer, D.2, comprised a ly wet conditions. A single metal fragment (possibly series of silt lenses with abundant shell. A sample a barrel hoop) and a patinated piece of glass were from Layer D.1 yielded a radiocarbon age of recovered from 0–10 cm below the Square D.5 250±100 .. (GAK-354), calibrated (after Stuiver datum. The superficial Layer X.1 sealed in all struc- and Reimer 1993) to .. 1510–1950 at one sigma, tural remains, including the fireplace in Square D.5. with an intercept of 1660 (see discussion below). However, this fireplace appears to have been dug into Layer X.2, and thus may post-date it. Layer E (circa 50–60 cmbd) The upper portion of Layer E (E.1) was a pure white ash with numerous Layer A (circa 20–30 cmbd) This thin (3–5 cm) shells. Toward the back of the shelter, Layer E.1 was and irregular layer of “loose, grey-brown, generally characterized by a series of lenses, including dark uncompacted” matrix yielded concentrated deposits charcoal lenses. The occupation associated with of shell and ash. Like the layers immediately below Layer E.1 appears to have been of fairly long dura- it, Layer A was considerably disturbed by pits (most tion, based on the thick and extensive deposit of commonly burials and firepits) and other smaller shell, other cooking refuse, and the matrix of ash, “holes” (after Smart 1965), possibly postmolds or burnt earth, and charcoal. Layer E.2 is characterized animal burrows. Layer A also seals in some pits by a predominantly sterile “grey-brown dust”. Both which must therefore have pre-dated Layer A. A bur- subunits of Layer E were fairly uniform, averaged ial pit in Square D.7 appeared to penetrate through about 5 cm in thickness, and extended over much of this layer. Three additional burials, which penetrat- the shelter. ed deep into the underlying strata appeared to have been dug from Layer A. Historic artifacts associated Layer F (circa 60–70 cmbd) The upper portion of with Layer A include a nail and another possible bar- Layer F (F.1) consisted of a very thin ash deposit, rel hoop. A sterile deposit of “compacted dust” sep- while the lower Layer F.2 was composed of “com- arates this layer from the underlying Layer B. pacted dust”. The deposition of Layer F.2 was inter- preted by Smart as evidence for a rather lengthy peri- Layer B (circa 30–50 cmbd) Layers B.1 and B.2 od of abandonment. Charcoal collected from Layer were separated by a thin lens of silt and shell. The F.1 yielded a radiocarbon age of 620±150 .. layers were composed of an almost pure, white ash (GAK-302), calibrated (after Stuiver and Reimer with some shell midden. Both ash layers were about 1993) to .. 1280–1440 at one sigma, with inter- 10 cm thick. The deepest pit in the center-rear of the cepts of 1320, 1350, and 1390. shelter cut into these layers and must therefore post- date Layer B. Layer G (circa 60–70 cmbd) This thin layer was comprised mainly of an ash matrix with a few frag- Layers B.2, B.1, A, and X.2 all contained remains ments of shell and only two artifacts, a stone abrad- suggesting that the latest phases of use took place er and a two-piece bone fishhook point. after Euro-American contact. It appears that the final use of the shelter, prior to abandonment, was Layer H (circa 60–70 cmbd) This strata was a for burial. “uniform, wet, brown matrix” of fine-grain size with 20 hawaiian archaeology
Table 1. Site O-5 Radiocarbon Dates the provenience of GAK-355 is incorrect, then the sample result does not date the basal occupation. Square Layer Depth/cmbd Lab No. 14C Age Calibrated Age1 Nevertheless, the other two samples are probably in D.5 D.1 50–60 GAK-354 250±100 .. 1440–1955 an appropriate provenience context. As previously stated, the artifacts (especially the adze type and fish- B.4 F.1 1–50 GAK-302 620±150 .. 1060–1640 hook types) associated with the upper layers are C.4 H.1 60–70 GAK-355 <250 indicative of late developments. With respect to the 1 two sigma Layer F date, it is worth recalling the early date from nearby Kuli‘ou‘ou Shelter (Site O-1).
a considerable number of naturally occurring rock clasts. The lower Layer H.2, rested on the natural Artifacts rock basement of the shelter. The upper Layer H.1 yielded scattered fragments of charcoal and shell. Bayard reported that the area under the shelter over- Because Layer H.1 appeared to represent the deepest hang produced 107 formal (e.g., shaped) artifacts. evidence of occupation, some effort was taken to col- Materials such as sea urchin spine files, unworked lect a sufficient amount of charcoal for dating. Yosi- shark teeth, and unworked stone flakes were not hiko Sinoto and Donn Bayard submitted a relative- counted as artifacts per se. The artifacts, considered ly large charcoal sample that yielded a date of less as a whole, form an assemblage reminiscent of those than 250 .. (GAK-355), and was therefore reject- recovered from the four southeast O‘ahu shelter sites ed. According to Smart, however, the only charcoal described by Emory and Sinoto (1961). Detailed in this layer was small fragments, which turned out descriptions of the Site O-5 materials were limited to to be insufficient for dating. Smart was convinced artifacts of special interest to Bayard. The functional that all the charcoal was collected, and concluded identifications are those provided in Bayard’s (1965) that the dated sample (wherever it came from) prob- manuscript; the author was not able to re-examine ably did not date the Layer H occupation. the artifacts for this paper.
In the discussions below, the prefix EA indicates an Radiocarbon Chronology unknown provenience. The prefix AN indicates an area outside the shelter referred to as the “annex”. Three carbon samples were submitted to the Cutting Tools Gakushuin Laboratory in Tokyo. I have calibrated 14C ages using Stuiver and Reimer’s (1993) CALIB Adze. A small adze, 2.7 cm from bit to shoulder, of program, version 3.02, and two sigma ranges are black fine-grain basalt was identified as quadrangu- provided above. As it is unlikely that these samples lar in cross-section (D.3-4, 6–12 cmbd). were corrected for isotopic fractionation (e.g., Adze chips. Seven adze chips, mostly under 2 cm in 13 12 δ13 C/ C ratios), the program’s optional C esti- length, were recovered. These include Specimens mate for fractionation correction was used. C.2-3, depth unknown; E.6-2, 7 cmbd; E.4-2, 15 The reliability of Gakushuin dates has been ques- cmbd; D.1-3, 20 cmbd; C.5-12 and E.5-2, 20–30 tioned (cf. Kirch 1986:23). The problem at Site O- cmbd; D.6-3, 30–40 cmbd. The two larger adze 5 is compounded by the fact that while Smart flakes, C.2-3 and E.6-2, may have functioned as (1965) rejected the “modern” date for Layer H, knives according to Smart (1965). Bayard (1965) received assurances from the labora- Stone saws. Two basalt saws or files, each with two tory expert that the sample was accurate. The close working edges, were found ( D.3-5, 6–12 cmbd; correspondence between Sample GAK-355 and the D.3-1, 13 cmbd). charcoal collected from Layer D (Sample GAK-354) was presented as additional evidence for assigning Stone knives. Two stone knives were recovered, one the occupation a late prehistoric date. However, if (EA-6, immediately below Layer X.1) of fine-grain 21 thomas
olivine basalt and a second of a black, more coarse- recovered from the site. One pounder (B.6-3) is made grain basalt (D.1-2, 15 cmbd). from a coarse basalt and has a roughly pecked finish. It was located just beneath the surface. The other Volcanic glass knife? One volcanic glass flake may specimen (C.6-2) is made from consolidated sand and have functioned as a knife (E.3-2, 10–20 cm). was found at 30 cmbd. Both specimens are smaller Core. A basalt core (AN-7), with two well-defined than the average given by Buck (1964a) for small flake scars, was found in the terrace in front of the taro pounders. Smart (1965) was inclined to view main overhang of the shelter. these as pestles. However, he also noted that the first specimen (B.6-3) was atypical for a pestle, as the base Abrading Tools flares somewhat and, according to Buck (1964a:28) Coral files. As in the Kuli‘ou‘ou and Hanauma Shel- true pestles are not flared. ters, the most commonly encountered artifacts were Hammerstones. A single waterworn basalt hammer- fine-grain Porites spp. coral files. A total of sixteen stone (EA-8) was recovered from the surface and was specimens were identified (AN-2, AN-3, and D.6-2, partially covered with a calcareous patina. depth unknown; E.4-1, surface; D.1-1, 5 cmbd; D.2- 5, 13 cmbd; B.1-3, 15 cmbd; C.2-2, 10–20 cmbd; Fishing Gear D.2-6, 23 cmbd; EA-10, 28 cmbd; C.5-15, 20–30 Fishhooks. Five fishhook fragments were recovered. cmbd; C.3-1 and C.3-2, 35 cmbd; E.4-6, 45 cmbd; A one-piece, unbarbed, jabbing, bone hook possibly B.8-2, 40–50 cmbd; C.7-4, 50–60 cmbd). originated between 10 and 20 cmbd. The head is a Stone and coral rubbers. Seven rubbing stones were HT4 type (Sinoto 1962:163). The presence of a recovered, three of these were of coral (E.3-1, 20–30 pointed shank with a distinctive protruding knob is cmbd; C.7-7, 40 cmbd; E.4-5, 41 cmbd), one of an important chronological marker, and has yielded pumice (E.4-4, 38 cmbd), one of andesite tuff (E.6- consistent results throughout the island chain (Kirch 1, 7 cmbd), and two of vesicular basalt (C.5-13, 47 1985:207). According to Goto’s (1986:174–6, 182) .. cmbd; B.6-10, 60–70 cmbd). analysis, the HT4 head form appeared around 1400, became dominant after .. 1600, and con- Grindstones. One large grindstone of coarse basalt tinued into the historical period. A comparison with was found (C.7-8, 20-28 cmbd). Three fragments of the 155 one-piece specimens measured by Emory, more carefully made, fine-grain basalt, bifacially Bonk, and Sinoto (1968:14 –15) revealed dimensions smoothed objects were recovered (D.3-2, 13 cmbd; that approximate the average for this type (ranging D.4-1, 20-25 cmbd; E.4-7, 45 cmbd). These may have from 7 to 32 mm, with a mean of 19.8 mm). The been whetstones or pieces of an unfinished mirror. Site O-5 hook has a shank length of 21 mm.
Piercing Tools A pearlshell shank fragment (C.1-1) was also recov- Awls, needles, pickers. Eight bone awls (D.2-4 and ered from the 20–30 cmbd level. It too was of the EA-5, surface; D.2-2, 0–10 cmbd; C.6-4, 20 cmbd; HT4 type. E.3-3, 30–40 cmbd; C.5-8 and C.7-1, 40–50 cmbd; One complete shank of a bone two-piece hook with B.6-9, 50–60 cmbd) and three needles (E.2-2, 10–20 a knobbed base was located in the 15–20 cmbd level cmbd; D.3-3, 20 cmbd; B.1-4, 30 cmbd) were iden- (D.7-1). Another bone shank specimen with a miss- tified. A single bone picker (D.7-7), immediately ing base was found at a depth of 5 cmbd (B.4-1). below Layer X.1, was also recovered. The relative scar- city of this artifact on O‘ahu, as compared to assem- A single two-piece, barbless, bone point with a blages from shelters on Kaua‘i and Hawai‘i Island, was knobbed, unfaced, inner base was found at the 60–70 noted by Emory and Sinoto (1961:39). cmbd level (C.5-11). This two-piece bone point was the only “early” fishhook, on the basis of its strati- Striking Tools graphic position. While there is still an insufficient Food pounders. Two “collar button” variety stone number of two-piece hook assemblages to establish a pounders, as described by Buck (1964a:27–28), were secure chronology beyond the South Point area, 22 hawaiian archaeology
Goto’s (1986) analysis of two-piece hook attach- ished black basalt mirrors representing pieces of at ments in relation to point type showed some promis- least three separate mirrors with diameters approxi- ing results. In comparing the incipient knobbed and mating 8 cm (D.5-1, 0–5 cmbd; B.1-1 and E.2-1, fully knobbed varieties, he concluded that the incip- 0–10 cmbd; D.2-1, 14 cmbd; D.4-1, 20–25 cmbd; ient knobbed type was preferred for small points. In C.7-5, 50–60 cmbd), and an 8 cm length of sennit addition, notches were more common on large cordage (B.6-4, surface) found in association with points than on small points (Goto 1986:187). Never- two pieces of twisted bark or vine (B.6-2, unknown theless, the various strands of information retrieved depth). from the Site O-5 excavations strengthen the assump- tion that knobbed two-piece hooks developed later Two possible ‘ulu maika (gaming stones) were also than the notched varieties. found (AN-5 and EA-1, surface). One specimen may be a blank (AN-5). Additional findings included a A carefully manufactured octopus lure point (E.6- split half of a wooden peg (C.6-3, 8 cmbd) and six 3), with two notches for attachment at the base, was pieces of worked bone (Ea-7, just below X.1; C.4-1; recovered from Layer B. (30–50 cmbd). 20 cmbd; E.4-3, 25 cmbd; C.5-9 and C.5-10, 40–50 cmbd; C.7-3, 50–60 cmbd). Two of the Sinker stones. An octopus lure sinker (B.6-1) of the worked bone specimens (C.7-3 and E.4-3) are perfo- “coffee-bean” variety was located at the 10 cmbd rated and were apparently intended to be fishhooks. level. It was made of coral conglomerate and had a clearly defined central groove. Five smaller sinkers Historical Artifacts were also found. Two specimens (EA-9, just below The historical artifacts consisted of a patinated glass Layer X.1 and C.3-3, at 30–40 cmbd) were proba- fragment (EA-4, just below X.1), two nails (B.6-7, bly used as lure sinkers, as evidenced by their small 25–30 cmbd; C.6-5, 30–40 cmbd), two metal size (less than 5 cm in length) and cross-grooves. Two pieces that were probably barrel hoops (B.1-2, 10 other sinker stones (B.6-8, 55–60 cmbd and C.8-1, cmbd; D.7-5, 30 cmbd), and a twisted length of 36 cmbd) are considerably larger and transversely metallic wire (C.7-6). The metallic wire, found at grooved, suggesting they were used with nets (cf. 50–60 cmbd, was the only metal fragment found Buck 1964b:342). A third specimen (B.8-1, 0–10 below 40 cmbd. cmbd) shows the beginning of transverse grooves.
Netting Gauge. A turtle shell netting gauge (D.6-1), similar to one described by Emory and Sinoto Artifact Summary (1961:60, Fig. 55c), was also recovered. However, its provenience is unknown. The horizontal distribution of the Site O-5 artifacts seemed relatively uniform, although artifacts were Adornment absent from the trench immediately adjoining the Eleven perforated dog canines (Specimens C.6-1, terrace at the extreme front of the shelter (Trench C.7-2, B.7-1, C.2-1, EA-3, AN-4, AN-6, C.8-2, A). Trench C was the most productive, yielding 57 D.7-2, D.7-3, D.7-4) were concentrated in the lev- artifacts, while Trench D produced 26 artifacts (one els above 50 cmbd. Other ornaments included two of doubtful provenience). Square B.6, centrally beads of cone shells (C.2-5 and C.2-7, depth located near the shelter opening, yielded the most unknown), a bead formed from the tip of the inner artifacts (n=10). core of Conus or a related genus (C.4-2, 50+ cmbd), and two shells commonly used in lei (Buck The vertical distribution of artifacts indicates that 1964c:542): one Nerita (C.2-6, depth unknown), most of the specimens came from the upper levels. and an unidentified species (B.6-6, 25–35 cmbd). Only 17 of the 107 catalogued items were found below the 40 cmbd level (corresponding roughly to Miscellaneous the lowest layers of Layer B and upper Layer C in the Miscellaneous items included fragments of a gourd sections excavated by natural stratigraphy). The top vessel (B.6-5, unknown depth), six fragments of pol- 20 cm yielded 43 artifacts, including eight surface finds. 23 thomas
Table 2. Vertical Distribution of Artifacts1 Depth (cmbd): UNK2 Surf 3 0–50 50–60 60–70 Totals Layers: X to B C to E F to H
Tools Cutting Tools basalt adze 1 1 basalt flake 1 6(2) 9 basalt saw 2 2 basalt knife 2 2 basalt core 1 1 volcanic glass flake 1 1 Abrading Tools files, coral 3 1 11 1 16 abraders, stone 1(1) 1 3 abraders, coral 3 3 grindstones 3(1) 4 Piercing Tools awls 2 5 1 8 needles 3 3 pick 1 1 Striking Tools pounder 2 2 hammerstone 1 1 Fishing gear 1-pc bone 1(?) 1 1-pc shell 1 1 2-pc bone shank 1 (1) 2 2-pc bone point 11 octopus lure point (1) 1 coffee-bean sinker 1 1 line sinker 3 3 net sinker 1 1 2 net gauge 1 1 (turtleshell) Adornment Conus bead 2 2 Nerita bead 1 1 muricid(?) bead 1 1 worked Conus core 1 1 dog tooth pendant 2 1 8 11 Miscellaneous vessel, gourd 1 1 mirror, fragment 3(1) 1 5 cordage, sennit 1 1 bark (?) 1 1 ‘ulu maika 22 wooden peg, fragment 1 1 bone blank perforated 1 1 2 unperforated 4 4 Historical Materials nail and spikes 2 2 iron wire (?) 1 1 other metal 2 2 glass mirror frag.? 1 1
1 Numbers enclosed in parentheses indicate items taken from squares excavated by natural strata; those given under the figures for artificial levels at the top of the columns are based on point D.5 as datum, and thus are to be taken as only very approximately coordinate with the levels themselves. 2 UNK=unprovenienced materials of unknown depth 3 SURF=surface finds 24 hawaiian archaeology
In the course of a recent survey, 41 traditional arti- windward valleys of the islands, a general leeward facts were collected from the surface of the excava- expansion may have occurred as early as the eleventh tion backfill at the site. They included 36 volcanic century .. (Davis and Kaschko 1980:1). Within glass and fine-grain basalt fragments. Based on these the general area of southeast O‘ahu, early settlements additional findings, Beggerly and McNeil (1985) esti- may have been in rockshelters, with subsequent per- mated that volcanic glass represented at least 24% of manent housing in the coastal areas. Moreover, the total artifact assemblage. use of such rockshelters might have continued inter- mittently for functionally specific activities (e.g., glass knapping, human burials) (Davis and Kaschko Conclusions 1980:1).
According to Spriggs’ (1988) revision of the Hawai- Smart described the Site O-5 deposit as representing ian sequence, the “Early Expansion Period” (.. five periods of occupancy, separated by intervals dur- 1100–1400) saw the establishment of settlements in ing which the shelter was uninhabited. The earliest favorable leeward locations, such as those adjacent to period remains undated, but the radiocarbon dates natural coastal fishponds and sheltered inlets. A com- indicate both pre- and post-contact uses for the cave. parison between Site O-5 and two other southeast These have been divided into a probable post-con- O‘ahu rockshelters (Kuli‘ou‘ou and Hanauma Shel- tact phase (Layers X through B.2, and the associated ters) indicates a similar pattern of late intensive use, pits), and a probable pre-contact phase (layers below as evidenced by the increase in shellfish consump- .. B.2). Layers C, D, and E date from post 1450 tion. Bird bone, on the other hand, shows the reverse to Euro-American contact, while Layers F, G, and H trend (Emory and Sinoto 1961). Based on the abun- .. date to sometime before 1450, mostly likely the dance of artifacts in the lower layers of Site O-5, fourteenth century. The chronology inferred from (about 2% of total artifacts, excluding those of fishhook types and the quadrangular adze is consis- unknown provenience and surface finds), we may tent with the radiocarbon dates. postulate an initial period of occasional, short-term .. Smart (1965) suggested that the aridity of southeast use of the site. Beginning with Layer E.1 (after O‘ahu may have compelled its inhabitants to rely 1450 and probably post-1600) use of the shelter more heavily upon fishing and shellfish collecting, increased, as shown by a much denser artifact accu- rather than taro cultivation. Although midden was mulation (about 33% of the total assemblage). Even collected and identified, no quantitative analyses were more intensive use is shown later in time, particular- attempted. Smart (1965:9), however, noted dense ly in the early historic period, as represented by shell deposits in Layers B.2, B.1, and A. In a memo roughly the top 30 cm of the deposit (about 65% of addressed to Solheim and Bayard, Smart concluded artifacts). that, We can gain some insights into the subsistence . . . the shelter sites so far examined in this area strategies at Site O-5 by looking at early historical can be regarded as habitation sites in which all accounts of leeward O‘ahu. Fishponds, like the one manner of activities were carried out (i.e. they identified at Kuapä would have provided salt, mul- were equivalent of a dwelling structure). In com- let, and crabs (McAllister 1933:93; Handy and bination, and with certain other evidence, I would Handy 1972:474), and the hillsides would have been suggest that this part of Oahu could have been suitable for limited root-crop and tree-crop cultiva- exploited by occasional visits from inhabitants tion, and for collecting wild forest products. Ethno- from other parts of the Island and was not an area graphically, the area surrounding Site O-5 was capable of supporting a permanent population known for its sweet potatoes (Handy 1940:155). (Smart 1965:4). The seasonal character of Kamilo Nui and Kamilo Iki Streams, which empty into the pond, would While we have accumulated a large corpus of data account for the presence of crops bearing a greater that demonstrates early occupation in the wetter tolerance to dry conditions (Davis 1984:4). While 25 thomas
early historical descriptions of southeast O‘ahu indi- _____. 1964c. Arts and crafts of Hawai‘i: Section XII, cated that Kuapä was an inland arm of the sea, rather Ornaments and personal adornment. B. P. Bishop than a pond, the possibility of repeated, purposeful Museum Special Publication 45. Honolulu. closing and opening of a littoral barrier to develop a fishpond should not be overlooked (Davis 1984:5). Davis, B. D., 1984. A research design for Pahua Heiau at Kamilo-nui in Maunalua, southeastern At the time of Euro-American contact, the popula- Oahu. MS., on file, Anthropology Department, Uni- tion in the ‘ili of Maunalua was quite dense, and versity of Hawai‘i at Mänoa, Honolulu. remained so well into the nineteenth century (Davis 1984:12–13). Possibly, the collapse of the sweet Davis, B. D., and M. W. Kaschko, 1980. Use and potato trade (which depended on the whaling indus- abandonment of habitation caves in the prehistoric try) after circa 1860 had a profound impact on the settlement of southeastern Oahu: a proposed local population (Davis 1984:12–13). With less research design for the 1980 University of Hawai‘i demand for agricultural products, there was presum- archaeological field program. MS., on file, Anthro- ably a decrease in labor requirements, and the people pology Department, University of Hawai‘i at consequently moved elsewhere. Other factors that Mänoa, Honolulu. may have contributed to population decline should Egler, Frank E., 1947. Arid southeast O‘ahu vegeta- also be considered, such as the cumulative effects of tion. Ecological Monographs 17:385–435. historically introduced Old World diseases. Emory, K. P., and Y. H. Sinoto, 1961. Oahu excava- tions. B. P. Bishop Museum Special Publication 49. Acknowledgements Honolulu. Emory, K. P., William J. Bonk, and Y. H. Sinoto. I wish to thank Dr. Matthew Spriggs for helpful com- 1968. Fishhooks. B. P. Bishop Museum Special Pub- ments and criticisms on earlier drafts of this paper. lication 47. Honolulu. Thanks also to Drs. Bert Davis, Patricia Beggerly, and the late Dr. Joyce Bath. The excavated materials from Goto, Akira. 1986. Prehistoric ecology and economy of Site O-5 are currently stored at Bishop Museum, fishing in Hawai‘i: an ethnoarchaeological approach. Department of Anthropology. University Microfilms International, Ann Arbor. Handy, E. S. Craighill. 1940. The Hawaiian planter, Volume I. B. P. Bishop Museum Bulletin 161. Hono- References Cited lulu.
Bayard, Donn T. 1965. Hawai‘i Kai: Artifacts (June, Handy, E. S. and E. G. Handy. 1972. Native planters 1965). MS., on file, Department of Anthropology, in old Hawaii. B. P. Bishop Museum Bulletin 233. University of Hawai‘i, Mänoa, Honolulu. Honolulu.
Beggerly, Patricia P. and J. R. McNeill. 1985. Archae- Kirch, Patrick V. 1985. Feathered gods and fishhooks. ological reconnaissance of the proposed marina zon- Honolulu: University Press of Hawaii. ing project Kaluanui 1, 2, and 3 (Hawai‘i Kai). _____. 1986. Rethinking East Polynesian prehis- Report submitted to Environmental Communica- tory. Journal of the Polynesian Society 95:9–40. tions, Inc. Honolulu. McAllister, J. G. 1933. Archaeology of Oahu. B. P. Buck, Peter H. (Te Rangi Hiroa).1964a. Arts and crafts Bishop Museum Bulletin 104. Honolulu. of Hawai‘i: Section I, Food. B. P. Bishop Museum Spe- cial Publication 45. Honolulu. Sinoto, Y. H. 1962. Chronology of Hawaiian fish- hooks. Journal of the Polynesian Society 71:162–6. _____. 1964b. Arts and crafts of Hawai‘i: Section VII, Fishing. B. P. Bishop Museum Special Publication Smart, Colin. 1965. Memo to Bill Solheim and Donn 45. Honolulu. Bayard-Hawai‘i Kai excavation report, 18 August, 26 hawaiian archaeology
1965. MS., on file, Anthropology Department, Uni- Stuiver, M., and P. J. Reimer. 1993. Extended 14C versity of Hawai‘i at Mänoa, Honolulu. data base and revised CALIB 3.0 14C age calibration program. Radiocarbon 35:215–30. Spriggs, Matthew. 1988. The Hawaiian transforma- tion of ancestral Polynesian society: Conceptualizing chiefly states. In State and society: The emergence and development of social hierarchy and political central- ization, eds. J. Gledhill, B. Bender, and M. T. Lar- sen, 57–73. London: Unwin Hyman. A Brief Report On Test Excavations at the Hawaiian Mission, Honolulu
Richard Pearson
University of British Columbia
During the restoration of nineteenth century buildings at the Hawaiian Mission, immediately south of Kawaiaha‘o Church in Honolulu, limited archaeological excavations were conducted. The project architects, Mr. and Mrs. Lockwood Frost, contacted the University of Hawai‘i Department of Anthropology regarding the potential for archaeological work. Test excavations were carried out over sev- eral short periods in 1968, 1969, and 1970, with the aid of Anthropology gradu- ate students and participants in introductory courses. Washing, numbering, and some repairing of the large number of artifacts was done at the University of Hawai‘i at Mänoa campus. The Hawaiian Mission, occupying its present site since 1820, constitutes a most important resource for the study of the latest period of Hawaiian archaeology. Archaeological study of the Mission is indispensable for understanding the role of the missionaries in the transformation of Hawai‘i.
Three main buildings comprise the Mission Museum at present: the frame Bingham House, built in 1821; the stone Chamberlain House, built in 1831; and between these two buildings, a stone bedroom unit, constructed in 1841. This lat- ter building is referred to as the Depository or Printing House. Changes in the construction details of these buildings have been documented in great detail by the Frosts (Fig. 1) from source material available in the Hawaiian Mission Children’s Society Library (Frost and Frost 1968). Excavations were carried out in two areas: 1) adjacent to the stone bedroom unit; and 2) an area adjacent to the frame Bingham House where an addition had once stood. Additionally, digging activi- ties by the restoration contractors in several historic wells on the property were monitored.
27 28 hawaiian archaeology
gencies. Our first excavation, undertaken in January 1968, was a trench that encircled the stone bedroom unit. At that time, this building was thought to be the Mission Printing House; however, it was later suggested that the Printing House was on the other side of King Street. A trench of units 2 ft x 2 ft (0.61 m x 0.61 m) was excavated, one at a time, next to the building walls; this was done so that some support was available to the walls at all times. Virtually all of this first trench consisted of disturbed soil, turned over either in the many renovations to the building or in the laying of pipes. A manuscript on the arti- facts from this trench was prepared by Michael Seelye and left with the Hawaiian Mission Chil- dren’s Society. Excavation below the level of the foundation, which could have told much about the occupation of the area before the arrival of the mis- sionaries, was prohibited by the architects as they were concerned about wall collapse. However, a number of squared coral blocks could be seen pro- truding from under the foundation and tradition has it that an aboriginal skeleton was discovered at the time of construction.
Frame Bingham House The second increment of excavation was the explo- ration of an addition to the southern side of the frame Bingham House. This addition, which con- sisted of a cellar with a bedroom above it, was built in 1828 (Frost and Frost 1968:CL-21). The dimen- sions are noted as being 15 ft (4.72 m) by 17 ft (5.18 m) and the basement was used for storage. It appears Figure 1. Evolution of Mission buildings (from Frost from old photographs that the addition was removed and Frost 1968:D-1). and the basement filled circa 1880 (Frost and Frost 1968: Plate 42, Map D-6).
In addition, we attempted to follow the outline of One of the most pressing problems faced by the the foundation of a kitchen that was added to the architects was elimination of moisture from the walls south side of the Bingham House in 1825 (Fig. 2). of the old buildings. Wall foundations had absorbed The kitchen was converted to a wagon shed in 1864 ground water, which was in turn absorbed by plaster (Frost and Frost 1968:D-1) and in the early 1900s on the walls, so that even in the second stories, there it, and many other outlying sheds, were removed. were problems with rotting window frames and staining. Given these conditions, the walls needed to At the time of the Bingham House excavation, resto- be supported at all times and archaeological excava- ration was in full progress. Our unit was placed next tion adjacent to them was limited. to the builder’s trenches, which had been opened for capping and water-proofing of the walls (indicated Stone Bedroom Unit by the dotted lines on Fig. 3). Excavation units 5 ft Occurring in the course of the restoration, our exca- x 5 ft (1.52 m x 1.52 m), divided into four quad- vations had to fit a number of rather stringent exi- rants, were placed in the interior of the basement; in 29 pearson
the kitchen area, wide-area shoveling and screening were used, and individual finds were plotted hori- zontally and vertically. The basement deposit was extremely loose and near the floor had large air pock- ets. In the absence of uniform stratigraphy, the exca- vation proceeded in 3 in (7.6 cm) arbitrary levels. We were unable to carry the excavation of each unit all the way down to the stone cellar floor, except in the case of Units 4 and 5 where the floor was reached about 4 ft 4 in (1.3 m) below the present ground sur- face. The other units were excavated to about 4 ft (1.2 m) deep. In general, the upper 2 to 21/2 ft (61 to 76 cm) of the deposit had a great deal of coarse rubble, while the lower portion consisted of fine soil and large pieces of broken furniture, cans, and iron straps.
The west wall, adjacent to the kitchen, was finished with white plaster, while the other three walls showed grouting between the large pieces of cut Figure 2. Unidentified blueprint in the Hawaiian Mission Children’s Society Library, showing Bingham stone. A tremendous quantity of plaster was found House addition with basement and stone bedroom in the fill; unfortunately only a small sample could (Printing House) (from Frost and Frost 1968:D-7). The be saved. In some cases the reverse face showed the Frosts comment that the drawing was probably made imprint of stone walls, presumably that of the base- about 1910 or 1920, from recollection. ment. Several coats of whitewash or greenish-grey paint had been applied to the plaster. Lath imprints were found on a large number of specimens. In many cases, the dimensions of the lath could be clearly determined. Widths of the impressions varied from 7/8 in (2.11 cm) to 15/8 in (4.14 cm), although the most common width seemed to be between 11/4 in (3.18 cm) and 11/2 in (3.8 cm). Space between the lath varied from 1/2 in (1.27 cm) to 3/8 in (0.97 cm). Six waterworn pebbles were also found, all with traces of plaster on them. Five of these were concen- trated in Unit 1, Quadrant 1, at 6 to 12 in (15.2 to 30.5 cm) depth. They may have come from a native style ‘ili‘ili pavement located in some other area of the mission compound. Twenty-four sharp, flat chips of basalt (commonly referred to as “blue- stone”) were found in the excavation. Two of these were embedded in plaster, showing their use as grouting. The flakes may be trimmings from basalt which the Mission received from the quarry near Punahou School.
Figure 3. Plan of the basement of the Bingham Adjacent to the basement was a hard-packed earth House, showing excavation units and adobe walls area (Fig. 3). In this area there was a kitchen, built in of kitchen. 1825, which was later rebuilt as a wagon house in 30 hawaiian archaeology 31 pearson
Figure 4. Bottles from basement and kitchen 1864. The perimeter of this, next to the basement excavation and Wells No. 1, 2, and 3. and the frame house, had been dug through by the a) Medicine bottle, heavy patina, molded design, contractors and the southwest corner was complete- basement, Unit 4, Quadrant 4, 3 to 31/2 ft ly covered by a stand of coconut trees. It was diffi- (91.5 x 106.7 cm) deep. cult to establish a horizontal datum for this area, as b) Dark brown beer bottle, broken bottom, wire recent filling and digging had been extensive. The loop for “lightning type” closure, from kitchen top 2 to 3 in (5 to 7.6 cm) of soil consisted of fill area, Layers I, II (see Fig. 3). from recent trenching and lawn fill, the latter which c) Case gin bottle, square body, dark green glass, had been built up over the past few decades. Under from Well No. 1. this fill was a hard-packed surface, which we cleaned d) Round bottle, brown glass, Well No. 1. off. Artifacts were conspicuously absent on this floor. Along two sides of this floor were the remains e) Beer bottles, Well No. 1. of a wall of dried mud blocks averaging about 13 by f) Beer bottle, Well No. 1. 12 by 6 to 8 in (33 x 30 x 15.2 to 20.3 cm) thick. g) Beer bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in Judging from the width, it looked as if it had been a (1.37 to 1.52 m) depth. low wall which had collapsed. Another possibility is h) Beer bottle, Well No. 2, 5 ft 8 in to 8 ft 6 in that the blocks are the remnants of a complete adobe (1.73 to 2.60 m) depth. floor, the middle portions of which had been i) Beer Bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in removed during construction of a wooden floor. The (1.37 to 1.68 m) depth. Frosts (1968:CL-50) noted an entry for May 8, j) Beer bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in 1850, in the Amos Starr Cook journal, Volume 9, (1.37 to 1.68 m) depth. pp. 120–121, that mentioned removing adobes to make ready for the laying of a floor. k) Beer bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in (1.37 to 1.68 m) depth. Fragments of lead sheeting were found 17 ft (5.2 m) l) Soda bottle, Well No. 2, 4 to 5 ft (1.2 to 1.5 m), from the basement and 3 ft (0.9 cm) from the edge 3 to 4 ft (0.9 to 1.2 m) depth; fragments came of the mud block wall (see Fig. 3). In a small hole, from two different layers in the well. dug through the packed earth floor, a dark brown m) Soda bottle, Well No. 2, 5 ft 8 in to 8 ft 6 in bottle was found. The bottle had been chipped along (1.7 to 2.6 m) depth. the rim of the base, had a “blob top”, and had a wire n) Soda bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in loop that appears to be from a cork closure used in (1.4 to 1.7 m) depth. the 1860s (Munsey 1970:116). The bottom of this o) Medicine bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in bottle has a reddish-black, bare iron pontil mark, that (1.4 to 1.7 m) depth. dates to between 1845 and 1870 (Munsey 1970:48). p) Medicine bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in (1.4 to 1.7 m) depth. Well Excavations
q) Medicine bottle, Well No. 2, 4 ft 6 in to 5 ft 6 in At the time of the excavations three wells, which had (1.4 to 1.7 m) depth. been previously filled in, were cleaned out by the contractors. Although I was unwilling to let the stu- r) Food bottle (?), Well No. 2, 4 ft 6 in to 5 ft 6 in (1.4 to 1.7 m) depth. dents excavate within the wells because of the danger involved, in the case of Well No. 2 it was possible to s) Medicine bottle, Well No. 3. keep a record of the depths from which the materi- t) Beer bottle, Well No. 3. als were recovered, and to have the students screen, sort, and bag the materials. Well No. 1 was located to the east of the basement, Well No. 2 lay to the west of the Chamberlain House, and Well No. 3 was in the garden west of the Bingham House. From the occurrence of “blob-top” bottles in Wells No. 1 and 3, it appears that they were filled in the latter part of 32 hawaiian archaeology
the nineteenth century. From Well No. 1, a case From Well No. 3, the base of a bottle embossed bottle for gin dating to about 1880 (Munsey with the mark of the Wisconsin Glass Company 1970:86), with a square body and dark glass, was (Wis Glass Co 38), which operated from 1882 to recovered. The specimen has no embossing and air 1886 (Toulouse 1971:541), was recovered. Frag- bubbles can be seen inside the glass. Three other ments of a brown-glazed stoneware teapot with the complete bottles were also found (Fig. 4: c, d, e, and motif “Rebekah at the Well,” made by Edwin and f ). Well No. 2, which produced an enormous quan- William Bennett about 1850 in Baltimore, Mary- tity of material, yielded several bottles with “blob land (Schwartz 1969:75), were also found. We made tops” and embossed beer and drug bottles. attempts to compare the excavated stoneware frag- ments with those in the Mission House Museum Of particular interest were three soda bottles. The collection for identification, but found this difficult. first (Fig. 4:a), is an embossed bottle with a Thorough research by someone familiar with nine- Hutchinson stopper and the words Arctic Soda works teenth century European and American utilitarian M. R. Desa, Prop. Elliot (1971:24) mentions that the ceramics would be most profitable. Arctic Soda Water Works was established in 1896 and located at 127 Miller Street near Beretania, In addition to intact bottles, a large sample of late Honolulu. The second specimen (Fig. 4: m) was also nineteenth century artifacts was recovered from the equipped with a Hutchinson stopper and embossed basement, kitchen area, and wells. These include with the name, Hollister and Company Soda Water. iron nails, hinges, buttons of stone, shell, bone, brass Elliot (1971:28,29) reports that H. R. Hollister and wood, stoneware, metal containers, clay bricks, operated the Hollister Soda Works between 1864 furniture fittings, iron rakes, horseshoes, and pins. and 1894. The third soda bottle, of clear but heavily This report has offered only minimal coverage of a weathered glass, bears the words Honolulu, Hawaiian small selection of the entire collection—enough to Islands, at the bottom of the embossing. Another confirm the rough chronological position of the complete bottle with a height of 101/2 in (26.7 cm) deposits. closely resembles a food container illustrated by Munsey (1970:153), who dates it between 1870 and 1890 (Fig. 4: r). Conclusions
A number of fragments with embossed maker’s marks on the bases were also recovered from Well The test excavations at the Mission Site have provid- No. 2. They seem to indicate slightly later dates ed authentication of historical records, information than those mentioned above, and suggest that Well on architectural detail, and abundant evidence of the No. 2 was filled up early in this century. One of material culture of the time. In addition, however, these has the combined AB mark of the Adolphus they show us interesting aspects of life in Hawai‘i in Busch Glass Manufacturing Company, Belleville the late nineteenth century. The construction of Illinois, a type made about 1904 to 1907 (Toulouse New England style basements in Hawai‘i is interest- 1971:26). Below the AB mark, the mark “H14” ing. Since the missionaries expanded them, they appears. Another example, also from the Adolphus must have continued to require them. Later, with Busch Glass Manufacturing Company, is marked the building of the Depository (Frost and Frost with A.B.G.M. Co., in a semi-circle, with “H19” 1968:1–2), which became a store, above-ground below. According to Toulouse (1971:26) this mark storage was favored. With the waning of the Mission was in use from 1886 to 1928. Site as a center, the out buildings were cleared away and the old houses became landmarks. The growth Fragments of a glass chimney from a kerosene lamp of local haole business, and the development of were found at various levels within Well No. 2. In Honolulu as a modern port, can be seen in the bot- addition, pieces of the same bottles distributed at tle record. The pattern of import and distribution of different levels, indicate that the well was filled with glass goods and metal containers, and the entry of accumulated junk all at one time. the inhabitants into commercial activity, are also 33 pearson
documented. In addition, the local production of advances in the historical archaeology of Hawai‘i. All luxury and consumer goods, absent in the first half of the measurements in this paper were originally in of the nineteenth century, can be studied. feet and inches.
Acknowledgments References
Grateful acknowledgment is made to the members Elliot, R. R. 1971. Hawaiian bottles of long ago. Hono- of the Hawaiian Mission Children’s Society for pro- lulu: Hawaiian Service Inc. viding funds for excavation supplies and photo- graphic processing, to Lockwood and Rossie Frost, Frost, L., and R. M. Frost. 1968. A study and report and to many students from the University of Hawai‘i on the old Mission Houses. Prepared for and on file who worked on the project. Particular thanks are with, Hawaiian Mission Children’s Society, Hono- offered to H. Reilly, L. Okouchi, S. Sakai, M. Maruo, lulu. P. Tanaka, and R. Crouch, who aided in the prelim- Munsey, C. 1970. The illustrated guide to collecting inary artifact sorting. Artifacts and photos are bottles. New York: Hawthorn Books Inc. deposited in the Hawaiian Mission Children’s Soci- ety Museum, South King Street, Honolulu. The Schwartz, M. D. 1969. Collector’s guide to antique graphics for this article were prepared by Moira American ceramics. Garden City, NY: Doubleday and Irvine. Company.
I regret that I have not been able to update this Toulouse, J. H. 1971. Bottle makers and their marks. report, prepared 16 years ago, in the light of recent New York: Thomas Nelson and Co. An Archaeological Predictive Model for the Mission Houses Site in Honolulu and its Value
Anne W. H. Garland
University of Hawai‘i at Mänoa
Historical documentation is useful in developing predictive models for archaeolo- gy, especially in the study of historical sites (Deetz 1977). An archaeological research design includes a prediction based upon preliminary findings from histor- ical documentation prior to fieldwork. Usually enough historical documentation is available to aid the archaeologist in developing more informed research designs.
Of course the archaeology often affects the interpretation of the historical docu- mentation. Predictive models for archaeology should, therefore, encourage a mutual exchange of information not only prior to fieldwork, but throughout the research project. Consequently, the choice of a sampling design and field objec- tives should remain flexible when based upon predictions from historical docu- mentation. Either additional documentation or archaeological findings may require reassessment.
This paper is an examination of the primary and secondary documentary sources on the Hawaiian Mission Station’s structures and property in Honolulu, and served as the beginning step of developing a research design for archaeological investigations at the Mission Houses Museum Site (Hawai‘i State Site Number 50-80-14-9991, TMK 2-1-32, Parcel 2). The background research for this predic- tive study dates from May of 1985, and was an effort to realize the potential of the site for culture contact research.
The aim of this study was to use historical records to determine which period fea- tures may potentially still exist archaeologically, and also to identify how they have been disturbed. A second goal was to locate information in the historical record about the various structures and land uses of the property. Eventually, the com- bined historical and archaeological research will incorporate the intrasite data into
34 35 garland
an intersite comparison of early mission stations as a) Will stations be in areas of high population they were interrelated with Native Hawaiian and density, including nucleated settlements of Native non-Hawaiian settlement patterns, especially prior to Hawaiians and their more populous ahupua‘a? 1850. This larger research aim depends upon the quality and quantity of intrasite data generated by the b) Will stations be associated with residential field excavations. regions of Native Hawaiian chiefs and their entourages? Initially, several research questions were proposed for work at the Mission Houses Site in Honolulu. The c) Will stations be in areas of major Euro-Ameri- documentary data collected for the predictive model can habitation, especially port towns? aided the formation of these general hypotheses. These hypotheses are presented as preliminary and Hypothesis 1: The Mission Station in Honolulu will may need revision after a more complete acquain- provide an early nineteenth century context for the tance with the literature, especially additional infor- study of material culture change among the Euro- mation on the organization of the American Board American settlers in Hawai‘i. of Foreign Missions in the Sandwich Islands and how this may have influenced the choice of dwell- a) Will influences from the Native Hawaiian cul- ings or the placement of the mission sites. ture and new environmental resources be more evident during the early settlement period (circa This paper deals with the structures and land use of 1820), before connections with import trading the Hawaiian Mission’s property in Honolulu. A are refined and accessibility to homeland goods is survey of the primary sources, such as diaries, mis- frequent? sion reports and journals, are used. From this data, predictions are made regarding the archaeological b) Will it be possible to determine the retention remains that exist within the present property lines and/or abandonment of local resources over time of the Mission Houses Museum. To begin with, at this site, since there were occupants through- however, a brief discussion about the presence of the out the nineteenth century? mission in Honolulu, as well as how the station’s site was chosen, will provide an historical perspective on Hypothesis 2: The early context of this station will the property itself. provide information about Euro-American influ- ences on Native Hawaiian culture (especially mater- How the American Mission came to send their ial culture) during initial and extended exchanges missionaries to the Sandwich Islands reveals the throughout the nineteenth century. transient nature of Pacific waterways in the early nineteenth century. Some of the young Hawaiian a) Native Hawaiians were domestics for the mis- boys who sailed away with foreign ships landed sionaries at settlement and resided on the sta- in Boston and became attached to the Foreign Mis- tion’s grounds during the nineteenth century. sion Society. After the printing and circulation of Will their structures, assemblages, and activity Henry Obookiah’s memoirs, who was one of these areas reflect Native Hawaiian occupancy and be boys, the American Board of Foreign Missions’ inter- distinct from those of Euro-Americans? est in sending a mission to Obookiah’s homeland b) Will Native Hawaiian style structures on the increased. Obookiah was tutored at Yale and con- property, which were used by the missionaries, verted to Christianity with the eventual desire to predominately contain material culture assem- return to Hawai‘i and spread his new learning, but blages and activity distributions that are non- he died of typhus before this was possible. The Sand- Hawaiian? wich Island Mission became a reality in 1819 when the first members were chosen for it. They did not Hypothesis 3: Initial locations for mission stations, have an invitation from the government of the which were often chosen by the Native Hawaiian Sandwich Islands when they set sail, but were hope- government, will spatially correlate with contempo- ful that they would be accepted upon their arrival. rary settlements. They included in their group three young Hawaiians 36 hawaiian archaeology
according to the king’s wishes. The missionaries requested land in Nu‘uanu Valley, but Boki refused them this site because he had farmers at this loca- tion. After a mutual agreement, a station site was chosen on June 19, 1820, on the property of Kalani- mökü in the ‘ili of Kawaiaha‘o. The site is described in Mrs. Loomis’ diary as, “about three quarters of a mile southeast of Hanarooroo village on the high road or rather footpath to Witete, on an extensive plain, with a view of the open sea in front and lofty mountains and fertile vallies [sic] in the rear” (Forbes 1967–71 v.5:998).
From March when they arrived, until the houses at the site were finished, the missionaries lived in Native Hawaiian style houses offered by some Honolulu traders. On September 16, 1820, the missionaries finally moved into their quarters on the site called Missionary Row.
The exact extent of the original site is not known. However, its overall boundaries were confirmed by the Land Commission on November 18, 1851. By this time, the station consisted of a series of five lots Figure 1. Frame House and Church. Photograph of a in and around the intersection of Mission Lane (now certificate made from a wood-carving done in 1822 Kawaiaha‘o Street) and Church Street (presently (HMCS Library, Visual Collection). King Street). In 1837–38, an uncontiguous lot was awarded on Kïna‘u Street in Kaumakapili for a school (LCA 387, 1851:141); it was northwest of the other lots.
who had been educated in Boston by the Foreign The Mission site of 1820 is within the lot designat- Mission Society. ed “Kawaiahao” in the Land Commission Award book (LCA 387, 1851:137, 510). Missionary Row When they arrived at Kawaihae Bay of Hawai‘i stood to the east of the Mission’s Frame House, the Island, they learned that King Kamehameha’s son, latter being occupied by the Hall family in 1851 Liholiho, was the new ruler. Before visiting him at (Forbes 1992:86, Fig. 9) (Fig. 1 and Fig. 2). The Kailua, they talked with high chief Kalanimökü, probable size of the Mission site in 1820 was 3.29 who was at Kawaihae. He accompanied them on acres (Royal Patent Grant 1600, 1853:5). As the their visit to Liholiho in order to support their mis- mission family grew, the original Kawaiaha‘o station sion. Liholiho decided to allow them to remain for expanded and changed, based on the size of Cham- at least a year, as long as the doctor resided with him berlain and Hall premises from 1851 to 1853 (LCA in Kailua. The mission party divided, leaving the 387, 1851:137, 510; Royal Patent Grant 1600, Thurstons and the doctor in Kailua while the rest 1853:5). The other lots that were awarded in 1851, continued to Honolulu to organize their “main sta- except for the Kawaiaha‘o Church lot, are all on the tion” (Loomis 1951:33). northwest side of Church Street (King Street) across Kalanimökü’s brother, Boki, was the governor of from the original Missionary Row (Fig. 2). The divi- O‘ahu. He refused to allow the missionaries to pro- sion of the station by Church Street was probably cure their own residences, but promised to find a site present from the beginning, since locations of Mis- for the station and have houses built for them, sionary Row by Bingham (in Forbes 1967–71), 37 garland
Chamberlain (ibid.) and Loomis (1951) all mention that it was on a road or footpath from Honolulu to Waikïkï.
The current property held by the Hawaiian Mission Children’s Society (HMCS) is part of the “Kawaia- hao lot” in its location as just described. All the other lots had been sold before the deed exchanges in 1907 and 1909 from Cooke, Atherton, and Chamberlain to the Society (Frost and Frost 1968a). The original 3.29 acres of this lot was reduced to 1.126 acres after two lots (Lot B and C, Fig. 3) at the intersection of Bingham and South Streets were sold and no longer part of the station (R. M. Towill Corporation l968). The “Kawaiahao” lot also was reduced in 1919 when the City and County of Honolulu realigned Mission Lane (Kawaiaha‘o Street) (see Frost and Frost 1968, unnumbered figure, “Reconstruction of the Old Depository”).
Comparison of the original road alignment with the realigned Mission Lane on these maps indicates that the portion of the original station which is now Figure 2. Survey map of the Mission’s claims, ca. under the street was the site of the Depository or 1845, for LCA 387 awarded in 1851 (HMCS Library, storage structure. The Depository was built in 1825 Visual Collection). by the mission and represented a unique architectur- al style, having a “battlement” facade made of stone (Forbes 1992:100). Fortunately, it appears that the northeast side of the Depository may still be intact and within the current property. Consequently, the property lines of the Mission Houses Museum still enclose the main portion of the original site of the mission station. Besides the restored structures now in the yard, which include the Chamberlain House, the Printing Office and the Mission Frame House, there may be quite a few other structures preserved as subsurface remains that were used from the earli- est occupation.
General Architecture and Land Use History
Land use on this lot prior to the mission station is not documented. The original locale of the station is on an open plain and the climate was described by the missionaries as very arid. The gardens they attempted did not survive through the windy, dusty Figure 3. Survey map by J.F. Brown, Surveyor, July 8, summer months. Only after irrigation brought water 1881 (1 in = 50 ft) (HMCS Library, Visual Collection). from Nu‘uanu Valley in 1867, were vegetation and 38 hawaiian archaeology
gardens plentiful. Prior to this, the only crop that spot, but the general locale was populated with a yielded all year was melons, according to Maria nucleated settlement at least by 1810 (Ii 1957). The Loomis’ journal of December 1820 (Forbes 1967–71 map of the port of “Onorourou” drawn by L. I. v.5). Duperrey in August, 1819, shows a number of hous- es in the vicinity of the eventual Mission Station When the first well was dug at the mission station, (Kelly 1978:37). The site was on the footpath from an early soil profile of the Honolulu plain was Honolulu to Waikïkï, which may have been a route recorded and is described in the journal of the brig of long-standing use, along which the local populace “Thaddeus”. Based upon subsequent well excava- preferred to make their residences. John Papa Ii dis- tions connected with the Frosts restoration, this cusses the branching trails at Kawaiaha‘o, which are description is accurate. It describes three feet of rich reconstructed in a map (Ii 1963:93). The station was soil, then three feet of cinders, followed by coral at this intersection. He mentions that “at Kawaiahao composed of seashells and sand “petrified” together a trail passed in front of the stone house of Kaina, (Forbes 1967–71 v.5:958). Further, description of late father of Kikaha” (Ii 1963:92). Whether Native the plain indicates that the shoreline was nearby: Hawaiian domiciles existed on the chosen site prior “Sea surf ran here and it does one mile from this to 1820 is not documented. However, any Native place. The plain, about fifteen miles in length and Hawaiian style structural remains encountered on two in breadth appears to be but a very few feet the property during excavations cannot be assumed above the level of the sea” (Forbes 1967–71 to be prehistoric and their temporal associations v.5:958). (prehistoric, pre-mission or post-mission) need to be confirmed by artifact, features, and stratigraphic The coral was not only subsurface, but is mentioned associations. as extending all over the plain and “frequently lies on top of the ground” (Forbes 1967–71 v.6:1102). The known post-1820 locations of the Station’s structures were researched in an exhaustive survey by Because of these accounts regarding the soil and arid the architectural restoration firm, Frost and Frost, environment, it is probable that land use of this AIA. The survey of exterior and interior information plain before the Mission Station could not be easily came from primary sources available in the HMCS agricultural without irrigation. Support for this inter- Archives. The Frost and Frost (1967–71) report is pretation comes from the account of de Freycinet on an invaluable aid not only for the restoration, but his visit to Honolulu in 1819, before the arrival of also for future archaeology on the property. Their the Mission. He states that near the fort protecting report, in fact, recommends excavations, especially Honolulu Bay, “the waters of a river, the mouth of of the standing houses’ foundations, for the purpos- which is nearby, are detoured and spread out for es of restoration and preservation. Excavations were agricultural needs” (Kelly 1978:41). He also recog- carried out in 1968, 1969, and 1970 by University nized that the large, flat plain, on which the town of of Hawai‘i field schools (Pearson 1969, 1980, this Honolulu was built, had been an ancient submerged volume; Seelye 1969). They concentrated on the reef. “This supposition is confirmed beyond doubt foundations of the Mission Frame House and the by the presence of coralline banks in the soil” (Kelly later Printing Office, plus three wells. Excavations at 1978:48). The distance of the Mission Station from the Mission Frame House consisted of the bisection the mouth of Nu‘uanu Stream, and the aridity of the of the basement fill of the eastern wing and the floor area, suggests that only dryland agriculture was pos- of the second kitchen (see Pearson, this volume, Fig. sible. Consequently, if any evidence of traditional 1), which extended from the eastern rear of the agriculture is found on the station site, it will prob- Frame House. Neither of these structures were visi- ably be subsurface features associated with dryland ble above ground. Today they are demarcated by cultivation. low, whitewashed stonework in the yard.
The presence of Hawaiian structural remains prior The chronological diagram devised by the Frosts’ to the station’s formation on the lot is also possible. (Pearson, this volume, Fig. 1) includes all informa- No direct evidence exists for earlier habitation at this tion about known changes to the main buildings, 39 garland
kitchen wing was added to service the Bingham wing (Forbes 1967–71 v.5:976). The first kitchen of 1825 still stands. The University of Hawai‘i’s archaeologi- cal field school excavated the 1827 kitchen addition, which had been attached to the “Bingham study” (Pearson, this volume)1.
Other cookhouses mentioned in the records include one built by Mr. Clark, the new superintendent, and finished by October of 1830. It was placed “midway Figure 4. “A view of the Mission at Hanaroorah, between our two dwelling houses”, according to Mr. Woahoo” made by Dexter Chamberlain and dated Chamberlain (Forbes 1967–71 v.5:976). This prob- October 11, 1820 (HMCS Library, Visual Collection; ably refers to the Frame House since the Chamber- original in Houghton Library, Harvard University). lain House was not occupied until December of 1831. A cookhouse addition was begun in Novem- ber of 1830 (Frost and Frost 1968a:3–29), during outbuildings and additions. These changes are his- construction of the Chamberlain House cellar. torically documented and two have been confirmed by archaeology: the “Bingham” study and the While the cookhouses were of traditional Hawaiian kitchen addition. The current structures in the yard or dressed coral construction, as well as various com- include the Chamberlain House, the Printing Office, binations, several other buildings were traditional the Mission Frame House, the first kitchen addition, Hawaiian frames structures with thatch coverings. and three wells. Two additional wells have caps. These included a blacksmith shop built in 1824, storehouses, the servants’ quarters, residences for new From the Frosts’ (1968a) study, plus the survey of arrivals of missionaries such as the Richards, the additional records for this paper, it is obvious that Stewarts and Mr. Clark in 1823 (Forbes 1967–71 many more structures existed within the current v.5:965–966), and a traditional Hawaiian house property boundaries. A few locations have been for Mr. Cooke built in 1851 (Forbes 1967–71 inferred from the records, but these have not been v.6:1090). Mr. Clark, who lived across King Street, archaeologically confirmed. These structures include also built a thatched study in 1823. traditional Hawaiian style houses used for early mis- sionary residences (an area called Missionary Row), While adobe was a building material used at the Sta- and the first detached cookhouse and storehouse. tion, some buildings that were described as “adobie, Based upon the engraving in Figure 1, the Frosts stone or unfired brick” in the Mission’s records actu- locate Missionary Row to the east of the Mission ally may have been made of coral blocks (Frost and Frame House. The detached kitchen, according to Frost 1968a). Walls of enclosures and buildings were Figure 4, is on the side of the row with the länai, that made of both coral blocks and adobe (Forbes is the seaward side (Forbes 1967–71 v.5:998). The 1967–71 v.5:982) and they all look similar in draw- länai and doors of residences at Missionary Row also ings of the Mission Station (Forbes 1992:100–101), opened seaward (ibid.). making it difficult to determine from the records what the exact building material was. Coral blocks, A stone cookhouse was built in June 1821, after con- dressed or rough cut, were generally used for build- struction of the Mission Frame House began in ing foundations and walls, which were then covered April. It was ten feet square and the Frosts’ study with thatched roofing. Structures of this kind inclu- places this detached outbuilding on the seaward side ded the Bingham study built in 1828 and two guest of the Frame House (Pearson, this volume, Fig. 1). houses in the front and seaward yards of the Cham- Frosts’ location is conjectural, since the remains are berlain House built between 1832 and 1840 (Forbes subsurface. 1992:100). Exceptions were the Chamberlain The first kitchen addition to the Frame House was House, the Bedroom Wing (the current Printing ready by January of 1825, and in 1827 a second Office), and the Depository, which had coral blocks 40 hawaiian archaeology
for walls and slate shingles for roofs. The preference standing together at any one time (Forbes 1967–71 for stone houses increased through time. In January v.5:965–66). of 1825, Levi Chamberlain reports that “of seven houses built in native fashion, standing in the mis- Enclosures around these structures were also altered sion enclosure at the time of arrival of reinforce- over time. The earliest ones were fences of thin pal- ment, most of which were improved for dwelling ings. “Our fence consists of small sticks, six or eight houses, not one is now safe for a store house—two feet in length, stuck in the ground and lashed by have fallen down, and but one is in any condition to withes to other sticks laid crossways”, according to be inhabited” (Forbes 1967–71 v.5:976). The need Mr. Elisha Loomis in August of 1822 (Forbes to repair or rebuild traditional Hawaiian houses 1967–71 v.6:1104). By April of 1823, 170 fathoms every two to three years encouraged the use of coral of coral wall, seven and one-half feet high, enclosed stones for permanent dwellings (Forbes 1967– the Mission Station with its framed house and four 1971:975). thatched houses (Forbes 1967–71 v.5:964)(Fig. 1). The paling fence was used inside the yard to keep In addition to traditional Hawaiian houses and coral animals away from certain areas. In April of 1827, stone structures, frame buildings with wooden clap- Chamberlain reported that a “fence of pallisades [sic] boards or planks on the sides were also common. A was built around the meeting house by the natives to wagon house built in 1851 for Mr. Cooke was of this keep out goats, cattle and hogs” (Forbes 1967–71 type of construction. The first mission house was an v.6:1109). While the Chamberlain House was under imported frame building sent by sympathetic friends construction (August of 1831), it was enclosed on in Boston. However, its basement and foundations one side with stone and on the other with a mud are of coral stone blocks. Clapboards were added fence (Forbes 1967–71 v.6:1114). By 1836, a high later by reusing the earlier weather boards. coral wall surrounded both the Frame and Cham- berlain houses, with no separating wall. Its purpose, From the above statement by Levi Chamberlain, it is as stated by Mrs. Chamberlain, was “to keep the apparent that a number of structures, in addition to place ‘kapu’ to the natives and the children from the Frame House, existed on the station grounds undue familiarity with them” (Forbes 1967–71 simultaneously. Several of these remained after the v.6:1115). The gate through the wall was primarily Chamberlain House was completed in 1831. How- from King Street, although its position changed ever, even more structures are mentioned in the once in 1848. records for which no provenience or construction type is known. These include at least two bath hous- Several brackish water wells were dug at the station, es, one used next to a well from 1831–39 (Forbes but only six have been relocated. Pearson (this vol- 1967–71 v.6:1113) and a “new bath house” built in ume) monitored the excavation of three wells on the July of 1850 (Forbes 1967–71 v.6:1039). At least current property. Frost and Frost (1968a) reports four privies were dug, in 1830, 1850, l852, and that eight wells are apparent in the records. Only the 1856, presumably with outhouses. More privies are fifth well had spatial references in the yard to aid in likely. A “lanai” was built in September of 1852 to relocating it. It has subsequently been excavated and store lumber, according to Mr. Cooke (Forbes restored (Frost and Frost 1968a; Pearson, this vol- 1967–71 v.6:1092). Another shed or “lanai” for ume). This well was probably dug in 1832, suggest- carts was built in June of 1825 (Forbes 1967–71 ing that many more must exist. Another well on the v.6:1088). Remodeling of the Depository in 1869 Chamberlain house grounds was adjacent to the for the Gulicks included construction of a store- kitchen and supplied the nearby bath house between house, carriage house, horse shed, and bathroom 1832 and 1840 (Forbes 1967–71 v.6:1115). In 1915, (Frost and Frost 1968b). How many of these struc- two wells were visible between the Frame House and tures were contemporaneous is difficult to surmise. the Kaiulani Home for Girls (Forbes 1967–71 The Depository outbuildings of 1869, and an v.6:1137). account of the new families’ residences after their arrival in May of 1823 (7 native houses), are the only Two lime kilns near the Chamberlain House are indications of the maximum number of houses reported by Mr. Chamberlain (in Forbes 1967–71). 41 garland
Frame House (Fig. 7). According to former HMCS librarian Leila Goodell (pers. com., 1985), a Red Cross tent was supposedly staked in the northwest corner of the lot during World War II.
Modern non-structural disturbances of the subsur- face remains are primarily from landscaping: the removal of trees, the installation of a sprinkler sys- tem, and the placement of cement walkways. The realignment of Kawaiaha‘o Street, which has already been discussed, is another alteration that may have destroyed much of the Depository’s foundations. Since the writing of this paper (1985), several new buildings were constructed (in 1987), including an administration building, visitor center, gift shop and a länai; the construction disturbed several subsurface structural remains that probably predate 1850.
Predictive Archaeological Features Figure 5. Frame House and part of Printing House With all the above disturbances noted, it is time to before 1907 restoration (HMCS Library, Visual Collection). suggest where structures and other remains are like- ly to persist and what their archaeological character- istics may be. To determine these predictions prima- They were dug in 1828 and 1830 for the construc- ry documentation, sketches, photographs and maps tion of his new Stone House and the Depository. were utilized. First, much of Missionary Row, or the The location of one is noted on the map in Figure 3. four Native Hawaiian houses built by Boki in 1820, Unfortunately, it appears as if the lot where it was is disturbed. If Figure 1 is accurate, their location located has been lost from the original mission prop- beside the Frame House to the east means that the erty and is now beneath a building in Kawaiahao additions on this side (the Bingham study and the Plaza. The whereabouts of the second kiln is kitchen), with a full basement, eradicated much of unknown. their remnants. Also, the eastern property lines differ from those of 1851, thus reducing the space between In conclusion, the number of structures with sub- the Frame House and the adjacent lot for subsurface surface remains could be quite high. Recent ground investigation. The early cookhouse, servants quar- alterations that might have disturbed these features ters, storehouses and early residences, made in the have been few. Current structures on the grounds traditional Hawaiian custom, may all have subsur- include the library (built in 1950), a dividing wall face vestiges within the present property bounds. If (constructed in 1967), and a garage on the far sea- any evidence of these traditional Hawaiian style hous- ward side of the original lot. Two buildings on the es survive, it will probably be in the form of post- King Street side of the lot are seen in Figure 5, a molds. Reverend Bingham and Levi Chamberlain photograph of the Mission Frame House in the early describe the following pattern of posts for the tradi- 1900s. The store on the left is over part of the east tional Hawaiian style houses at the mission station. side addition, an area excavated by Pearson (this vol- ume) and a University of Hawai‘i field school in the “Round posts, a few inches in diameter, are set in late 1960s. The storefront to the right is not a Mis- the ground about a yard apart, rising from three sion building, but was built within the bounds of the to five feet from the surface. Three poles or posts, original lot. A wagon house was also located over the about three times the length of the side posts, are site of the east addition, as well as near the second set in the ground, one in the centre of the build- kitchen which was built off the rear of the Mission’s ing, and the others at the ends, on which rests the 42 hawaiian archaeology
nether ridge pole supporting the head of the these structures are represented and may have rafters. A door-frame, from three to six feet high, intruded into earlier features. is placed between two end or side posts. The area or ground within, is raised a little with earth, Regarding the enclosures, the high coral wall is no to prevent the influx of water and spread the longer present, but a low one was completed in 1924 grass and mats” (Bingham in Forbes 1967–71 after the HMCS obtained the title. Since the low v.5:944–45). wall borders the property, it is probably on the site of the original wall, except where the road has been “Thatched houses are the most common habita- realigned. Any evidences of the palisade fences will tions in this country they are composed of a cer- be difficult to locate because of the size of the sticks, tain number of posts from 4 to 6 inches in thick- but it is possible to locate activity areas which were ness from 5 to 7 feet high from the ground in enclosed to keep animals out, such as garden plots, which they are firmly fixed at the distance of 4 or or areas used as animal pens (Forbes 1967–71 5 feet apart. The ends formed of posts at the same v.6:1102). distance from each other as those on the sides, of different lengths corresponding to the slope of the roof, the bottom of which enters the ground Conclusions and the top is made fast to the rafters. A grass house is sometimes cover over with a thick coat of The historical evidence outlined in this paper sug- mud and then it is called a hale lapo [sic] or gests that many of the structural remains of the mudhouse” (Chamberlain in Forbes 1967–71 Mission Station still exist within the current proper- v.5:975). ty boundaries. Additionally, they are amenable to Dexter Chamberlain provides a good sketch of archaeological excavations designed to pursue some Missionary Row (Fig. 4). The key with Figure 4 of the hypotheses presented, especially those regard- identifies each house and provides dimensions for ing initial settlement. Success in locating these the posthole patterns of several (Forbes 1967–71). remains will depend upon the effectiveness of the sampling strategy that is chosen for the archaeologi- The structures with coral block foundations should cal investigations. Moreover, the validity of this pre- be readily apparent, unless severely disturbed or dictive model, which is based on historical records, robbed for the reusable blocks. Re-use of coral will be determined by the archaeology. blocks, while not specifically mentioned in the records of the Mission, was a common practice for The value of archaeological excavations on the Mis- the Euro-American period. The cost of the blocks sion Houses’ property is threefold. First, it is one of and the labor to cut them encouraged their reuse. the only non-concreted parcels left in the Kaka‘ako Whether the coral foundations were laid at the sur- Historic District which is relatively undisturbed. face or sunk into a trench would also affect their Second, recovery of provenienced artifacts from a preservation. Buildings with coral block foundations site occupied so early and continuously by Euro- or walls include the Depository, the detached cook- Americans (culturally speaking) will provide an house just south of the Mission Frame house, the important material culture baseline for establishing guest houses in the Chamberlains’ yard, and the when certain imported goods came to Hawai‘i. This adobe study for Bingham, which was southeast of will benefit archaeologists who excavate historical the Frame House. All of these structures are within period sites, as few comparative assemblages are cur- the current property and their hypothesized loca- rently available. Thirdly, the interrelationship among tions appear relatively undisturbed. early settlers and Native Hawaiian populations, as it is portrayed in the portable and non-portable mate- No specific locations are identifiable for many of the rial culture, is of interest to both Native Hawaiians other structures such as privies, wells, bath houses, and the academic community. For example, there sheds, and the kiln. However, given the areal extent appears to be a transformation of the morphology of the current lot, it is highly feasible that many of of Native Hawaiian structures within a European- 43 garland
American context and vice versa. The Missionary observations suggest that the architecture of out- Row houses were built by Native Hawaiians accord- buildings is varied, with much traditional Hawaiian ing to a traditional design, both structurally and spa- construction evident. Stratified refuse deposits were tially. De Freycinet comments in 1819 that “the located in several areas of the yard, which will enable lodging of each fairly well-to-do family consists of at construction of a material culture baseline. Lastly, least three adjoining huts: one is the men’s eating the archaeological potential of the property was con- room, one is for the women, and the third serves as firmed beyond a doubt. In twenty-six square meters a bedroom. Sometimes all three are enclosed in a of excavations, 200 features have been identified. The hedge or palisade” (in Kelly 1978:64). Missionary preliminary report for the 1992 field season contains Row (Fig. 4) exemplifies this arrangement. The a discussion of the survey and testing phases, includ- American missionaries subdivided the interiors ing initial data recovery (Garland 1992b). A second according to their ideas about behavioral space. They preliminary report for the 1993 field season has also had sleeping, living, and meeting places in each been prepared (Garland 1993) and discusses data house, while they set up their dining area, jointly, on recovery. A final report will be included in the doc- the länai (Forbes 1967–71). Excavation of these toral paper to be written for this research project. areas would theoretically produce discrete assem- blages signifying these differences in spatial use. Also, the use of coral for building materials was an Note innovation based upon a new environmental demand. The missionaries appear to have made their 1. It appears that the location of the 1825 kitchen in new surroundings conform to their own traditional the Frosts drawing entitled “Evolution of Mission design. Buildings” (see Pearson, this volume, Fig. 1) is in Consequently, the information presented in this pre- error, based on comments in Levi Chamberlain’s dictive model indicates much potential for archaeo- October 11, 1824 journal entry (in Forbes 1967–71 logical investigations on the property of the Mission v.3:426). Houses Museum. It also suggests that the hypothe- ses presented earlier can be easily investigated with data from this site. And finally, the predictive model References provided assurance about the potential for culture contact research at the Honolulu Mission Station Deetz, J. 1977. In small things forgotten. Garden Site. City, New York: Anchor Press.
The Value of the Predictive Model (1993) Forbes, D. W. 1967–71. Research notes. 12 vols. On file, Hawaiian Mission Children’s Society Library, Three archaeological field schools by the University Honolulu. of Hawai‘i in 1992 and 1993 have provided data from the Mission Houses Site with which to address _____. 1992. Encounters with paradise: Views of the value of the predictive model. The main value of Hawaii and its people, 1778–1941. Honolulu: Hono- the model was in the development of the research lulu Academy of Arts. design. The general and specific locations for struc- tures and activity areas was invaluable in selecting Frost, H., and R. Frost. 1968a. A study and report areas to test, which provided data for the research on the old Mission Houses. Prepared for Hawaiian objectives. Mission Children’s Society, on file, Hawaiian Mis- sion Children’s Society Library, Honolulu. While the final research proposal for these excava- tions included changes in the research objectives, _____. 1968b. The old stone Depository at the nonetheless, culture contact research remains the Mission Houses Museum. Prepared for Hawaiian main focus (Garland 1992a). Data with which to Mission Children’s Society, on file, Hawaiian Mis- address these issues is not yet analyzed, but field sion Children’s Society Library, Honolulu. 44 hawaiian archaeology
Garland, A. 1992a. Proposed: Mission Houses Loomis, A. 1951. Grapes of Canaan: Hawaii 1820. Museum, 1992 University of Hawai‘i archaeological Honolulu: Hawaiian Mission Children’s Society. field school and doctoral fieldwork at Site 50-80-14-9991. MS., submitted to and on file with Pearson, R. 1969. Descriptive archaeology of the Hawai‘i State Historic Preservation Division, Hono- Bingham Wing, Old Frame House, Mission Houses. lulu. Collection of student papers and snapshots of work in progress, on file, Hawaiian Mission Children’s _____. 1992b. A preliminary report about the Society Library, Museum Collection, Honolulu. University of Hawai‘i archaeological field school’s survey and excavation at Site 50-80-14-9991, June _____. 1980. A brief report on test excavations at 9–July 31, 1992. MS. submitted to and on file with the Hawaiian Mission, South King Street, Hono- Hawai‘i State Historic Preservation Division, Hono- lulu. In The ethnic cultures of Japan and surrounding lulu. areas. Shimonosaki, Japan.
_____. 1993. Second preliminary report for archae- _____. 1995. A brief report on test excavations at ology at Mission Houses Museum (TMK 2-1-32; the Hawaiian Mission, Honolulu. Hawaiian Archae- Parcel 2), University of Hawai‘i archaeology field ology 4. schools, spring and summer 1993. MS., submitted R. M. Towill Corporation. 1968. Plan of Hawaiian to and on file with Hawai‘i State Historic Preserva- Mission Children’s Society property as surveyed by tion Division, Honolulu. R.H.L. Au on October 8, 1968. On file, Hawaiian Hawaiian Mission Children’s Society Library, Hono- Mission Children’s Society, Honolulu. lulu. Visual Collection. Rockwood, P. 1957. Honolulu in 1810. Map drawn Ii, J. P. 1963. Fragments of Hawaiian history. Trans- from data complied by D. Barrere, Bishop Museum lated by M. Pukui and edited by D. Barrere. Bishop Press, Honolulu. Museum Special Publication 70. Honolulu: Bishop Royal Patent Grant 1600. Records of the Board to Museum Press. Quiet Land Titles, Book 6:5, Kingdom of Hawaii. Kelly, M., ed. 1978. Hawaii in 1819: A narrative On file, Archives of Hawai‘i, Honolulu. account by Louis Claude de Saulses de Freycinet. Pacific Seelye, M. 1969. Descriptive archaeology of the Anthropological Records No. 26. Honolulu: Hawaiian Mission printing house. MS., prepared for Department of Anthropology, Bishop Museum. and on file with Hawaiian Mission Children’s Soci- Land Commission Award 387. Board of Commis- ety Library, Honolulu. sioners to Quiet Land Titles, Book I:137, 141, and 510, Kingdom of Hawaii. On file, Archives of Hawai‘i, Honolulu. Roadkill Archaeology on Lana‘i: ¯ A Historic House Site at Kahemano (State Site 1529)
Matthew Spriggs
Australian National University
In March 1986, while driving on the jeep road which runs along the east and southeast coasts of Läna‘i from Maunalei (the end of the paved road) to Naha, I noticed glass and ceramic fragments protruding from the sandy roadbed, just south of the mouth of Kahemano Gulch (U.S. Geological Survey 1984, Grid 262 002). This gulch is indicated erroneously as Kapua Gulch on both the 1923 and 1984 U.S. Geological Survey Maps. Kapua gulch is, in fact, the next gulch to the southwest, if Emory’s (1924) informants are to be believed.
The road is graded by a bulldozer at infrequent intervals but the cultural materials are quite localized. The Lanai Government Survey Map of 1878 indicates four houses in this portion of Kaohai Ahupua‘a, amid a coconut grove. The map also erroneously shows a heiau as being in this area.1 During his 1921 survey, Emory (1924:4) noted “several modern sites” along the beach road in the coconut grove, and on the beach itself, “a structure probably a tomb but possibly a ko‘a” (State Site 152).
Although no structural remains were found in 1986, the scatter of historic mate- rials almost certainly marks the site of one of the houses occupied in 1878 but abandoned prior to Emory’s 1921 visit and probably before 1900.2 A selection of ceramics and a bottle base were collected, and have been analyzed by Laura Carter (see Appendix A), supporting an occupation date of approximately 1850 to 1890.3
This short report shows that “keeping your eyes on the road” may lead to the dis- covery of potentially significant archaeological sites, as well as preventing tire dam- age from sharp ceramic and bottle fragments. Artifacts such as these may be the only surficial clues to the presence of buried sites.
45 46 hawaiian archaeology
5. Annular ware (banded), circa 1845–1880 (Don- Notes ham 1985).
1. The 1878 cartographic error caused Emory much 6. Edge-decorated feather edge earthenware, circa wasted time. In September and October of 1921, he 1850 to 1880+ (Donham 1985). searched in vain for the heiau on six different occa- Examples of all of these types of ceramics have been sions before learning from Pia Kauhane, a knowl- recovered at sites on Maui, O‘ahu, and Hawai‘i. edgeable Läna‘i native serving at that time as a policeman in Lähainä, that the map was mistaken and probably referred to the large heiau at Löpä References some two kilometers to the northeast (see Emory 1921-2:34, 36, 41, 40, 56). Donham, T. K. 1985. The Seth Richards and James 2. At Ko‘ele, Emory (1921-2:50) met a Mrs. Maka Mariott house sites: Historic context and material cul- who had lived at Kahemano as a child, presumably ture. Decorah, IA: Luther College Archaeological some time prior to 1900. Research Center.
3. The artifacts are currently stored in the University Emory, K. P. 1921–2. Visit to the island of Lanai, of Hawai‘i at Mänoa Archaeology Laboratory until Hawaiian Islands. Bishop Museum Library, Hono- such time as a museum is established on Läna‘i. lulu. Typescript.
_____. 1924. Village and house sites elsewhere than Acknowledgements at Kaunolu, Lanai. Bishop Museum Library, Hono- lulu. Typescript.
I would like to acknowledge the hospitality of Sol Herskovitz, R. M. 1978. Fort Bowie material culture. and Lynn Kahoohalahala, their family, and other Anthropological Papers of the University of Arizona Läna‘i residents during my various visits to the no. 31. Tucson: University of Arizona Press. island. A tip of the hat to Michael Graves who sug- Jones, O. 1971. Glass bottle push-ups and pontil gested the concept of “roadkill archaeology.” marks. Historical Archaeology 5:62–73.
Lanai Government Survey. 1878. Registered Map Appendix A: Analysis of Historical 1394, by J. F. Brown and M. D. Monsarrat. Materials Noel-Hume, I. 1973. Creamware to pearlware: A Williamsburg perspective. In Ceramics in America, Laura Carter ed. I.M.G. Quimby, 217–55. Charlottesville: Uni- National Park Service versity of Virginia Press. 1. Bottle base, dark green, iron pontil, circa 1845– Ramsay, J. 1939. American potters and pottery. Bos- 1880 ( Jones 1971). ton: Hale, Cushman and Flint. 2. Yellow earthenware sherds, American manufac- U.S. Geological Survey. 1923. Topographic map of ture, circa 1830–1930 (Ramsay 1939). the island of Lanai, Maui County, Hawaii. 3. Buff earthenware sherds, British manufacture, U.S. Geological Survey. 1984. Läna‘i South, Hawai‘i circa 1840–1900 (Herskovitz 1978). (1:25000-scale metric topographic map). Provi- 4. Hand-painted “gaudy Dutch” whiteware and sional edition. sponge-decorated whiteware, circa 1850 to early 20th century (Noel-Hume 1973). Distribution and Design of Pacific Octopus Lures: the Hawaiian Octopus Lure in Regional Context
Michael T. Pfeffer
University of Hawai‘i at Mänoa
Introduction Tracking the historical pathways of human migrations across the Pacific has been a primary goal of Oceanic archaeology. This is accomplished in part by examining material remains in a comparative temporal-spatial framework. Archaeologists have traditionally looked at stylistic variation to build chronologies and infer rela- tionships, but have often neglected functional aspects of artifacts. By separating stylistic and functional dimensions of artifact variability, the origin and distribu- tions of traits may be more readily understood (Dunnell 1978). In this regard, components of marine tool kits have been key in identifying homologous similar- ities within the Pacific region (e.g., Allen 1992; Buck 1930, 1964; Emory et al. 1968). Allen (1992; MS) has recently begun to explore functional aspects of fish- hooks as well. The focus of this paper is a widespread but little investigated fish- ing tool, the octopus lure, which reached its most complex and developed form in the Hawaiian Islands. The emphasis is on functional features of the octopus lure, but distributional patterns within Polynesia, which may reflect ancestry and con- tact, are also explored.
Forms of Octopus Lures The octopus lure is a specialized and complex tool used solely for capturing octopi. Two major forms are known from the Pacific, one which functions primarily as a lure (Fig. 1) and a second which also includes a hooking device (Fig. 2). The non- hooked form is colloquially known as the “imitation rat” lure (hereafter IR). It is widely distributed across the Pacific, being found in both Polynesia and Micro- nesia (Table 1). The hooked form often includes a distinctive coffee-bean-shaped sinker (hereafter CB lure) and is best known from the Hawaiian Islands.
47 48 hawaiian archaeology
Both forms of lure serve to attract octopi from their lairs. In the case of the IR form, the device is solely an attractant with no hooking component. Once in the open, the animal can either be drawn to the sur- face and captured or, with the hooked CB lure, snagged on the ocean bottom. The success of the lure centers around three factors. First, shells of the cowrie (Cypraea), typically used as a “bait,” are a favored food of the octopus. Second, octopi are ter- ritorial and will often attack other octopi. The tassels and leaves often used on these lures imitate octopus tentacles, thus resembling a “competitor.” Third, octopi are inquisitive by nature and attracted to brightly colored materials, such as the rocks, tufts of Figure 1. The imitation-rat (IR) lure (from Samoa) feathers, and flowers found on ethnographic speci- mens (Handy 1923; Kamakau 1976).
The IR lure takes its name from its supposed resem- blance to a drowning rat. Folklore explains that it was an effective lure because of the octopi’s long- standing hatred of the rat (but see below). The IR lure consists of several cowrie shell plates tied over a conical, non-grooved or grooved, stone sinker (Fig. 1). A coconut tree root, or other slender stick, is tied to this component. In some cases, coconut leaflets, feathers, or other tassels are also tied to the lure. An ethnographically documented variant, found in the Society Islands, uses cowrie shell plates tied around a long, slender stick. Unlike most other IR forms, the Society Island lure has no stone weight. This variant continues to be used today, particularly in shallow water environments (Sinoto pers. com., 1993).
CB lures (Fig. 2) consist of a distinctively shaped, grooved stone sinker, a whole cowrie shell or shells, a length of wood, a long bone or wood hook, a length of sennit or other form of line, a bone toggle and, on some varieties, a tassel consisting of strips of ti (Cordyline) or other leaves (Fig. 2). The sinker is tied to one side of a short stick, while the cowrie shell is tied opposite the stone. A sharp bone or wood hook is attached to the end of the stick, with the hook pointing towards the cowrie shell. Eth- nographic specimens from the Bishop Museum (Emory et al. 1968), and ethnographic accounts, indicate that on some specimens, a tassel of ti leaf strips was attached to the end of the lure near the Figure 2. The coffee-bean (CB) sinker lure (from hook. The fisherman’s line is attached to the cowrie Emory et al. 1968) with a bone, coral, or wood toggle. 49 pfeffer
Table 1. Geographic Distribution of Octopus Lures At least six variants of the CB sinker were developed in Hawai‘i (Emory et al. 1968:27; Pfeffer in prep.). Location Octopus Lure Source Hawaiians also developed another sinker form, Marianas IR lure Thompson (1932) referred to as the “bread-loaf” sinker because of its Fiji IR lure Hocart (1929) distinctive shape. Bread-loaf sinkers are not as com- Tonga IR lure Hocart (1929) mon as coffee-bean sinkers and also may have been Niue IR lure Loeb (1926) used in conjunction with parrotfish (Uhu) dip nets Samoa IR lure Buck (1930) (Buck 1964:345). It is interesting that all of the Southern Cooks possible IR lure Allen (1992) ethnographic lures with bread-loaf sinkers use two Societies IR lure/possible CB lure Handy (1932) cowrie shells. It may be that the slimmer bread-loaf Rudiger & Smith (1987) sinker was developed specifically to allow use of a Marquesas IR lure/possible CB lure Handy (1923) shell on both sides of the lure (see Emory et al. Sinoto and Kellum (1965) 1968:28). Of particular note in this respect is the Rolett (1989) following early twentieth century Hawaiian language Aotearoa possible CB lure/hooks Leach (1979) newspaper account (Nalimu n.d.:904–927): Hjarno (1967) Hawai‘i (PN) CB/Breadloaf lures Buck (1964) Squid Fishing—The hook used for this is the Kamakau (1976) kakala or spike. It is bent a little and fastened to Emory, Bonk, a stick. The stick is about the width of two fingers and Sinoto (1968) put together into one. On this stick are fastened Malo (1903) Kahaulelio (1902) two cowry [sic] shells, both tied fast to the stick and back of the place where the cowries are bound to the stick there the lines are tied and on the under side of the cowry and stick is a stone. This stone draws the cowries into the sea where the squids are. Over the cowries hang yellowed ti Hawaiian Coffee-bean Lures leaves, on either side of the shells. Live fishes are Components of the Hawaiian CB Lure. While fastened with the yellowed ti leaves, two live hina- octopus lures are found throughout the Pacific, in lea fishes. The cowries are let down with the live Hawai‘i a distinctive form developed, along with a fish and the fisherman keeps hold of the other end rich record of their use. Survey of the ethnographic of the cord. He remains on the canoe and jerks and literature reveals the importance of the individual pulls the cowries up and down. The canoe moves lure components and the science of octopus capture. about as he fishes for squids with his cowries. The squid is caught by the hook in front of the cow- Malo (1903:19) lists 24 different kinds of stone that rie. When he drew it up, the squid clutched the were used to make octopus lure sinkers. Kamakau canoe if it was a big one a fathom in length. It (1976:68) lists nine stone varieties, including: is killed with a hatchet and knife till the head is komano, pu‘ukuna, maili, polipoli, pupukea, kalapai- cut off. It is a soft but a moving one (Nalimu ki, ‘iole, kauala, and ‘o‘io. Unfortunately, neither Malo n.d.:905–906). nor Kamakau describe these varieties in detail. How- ever, the number of stone types demonstrates the sig- Nalimu’s description is the only account found of nificance placed on the sinker portion of the lure. live fish being used in conjunction with the octopus Nowhere else in Polynesia is there any indication of lure. Hinalea, the common name given to wrasses a comparable diversity of stones being used in octo- generally, could also refer to any one of three specif- pus lure sinkers. The elevated importance of the ic reef taxa: Hinalea hilu (Coris flavovittata), Hinalea sinker stone is also highlighted in one form of lure, ‘akilolo (Coris gaimardi ), or Hinalea ‘akilolo (Macro- the okilo he‘e, where there was no cowrie shell bait pharyngodon geoffroyi ) (Tinker 1978:299–303). and the stone was the primary attractant (Buck 1964; Nalimu’s account is also unique in its mention of see below). multiple cowrie shells. 50 hawaiian archaeology
The particulars of the cowrie “bait” were also impor- numerous sinker stones in underwater, offshore sites tant. Kamakau (1976:67) lists four commonly used at depths of up to a hundred feet. varieties of cowrie: ahi, ‘olupalaha, pauhu, and kupa. The ahi cowrie was artificially reddened by burning Buck (1964) discusses three methods used by Native to make it more attractive to the octopus. Some cow- Hawaiians to catch octopus: 1) the pan-Polynesian rie shells were considered more successful than oth- method of spearing octopus in shallow water; 2) the ers and thus highly prized and carefully curated by kilo, or ‘okilo he‘e method (see below); and 3) lu-he‘e, Hawaiian fishermen. Different types and colors of the use of the cowrie shell lure to catch octopus in cowrie shell were appropriate for different times of deeper waters. Much of Buck’s information comes day and under different lighting conditions. Fisher- from Kamakau (1976). men frequently took several cowrie shells on a fishing In the kilo, or ‘okilo he‘e method, the fisherman had foray, to meet the specific conditions encountered. to be able to see the octopus on the ocean floor. The hook used in Hawaiian CB lures is similar to Kamakau (1976:69) states that this method was used those used for composite bonito lures throughout in “shallow seas from 6 to 10 fathoms deep” (36–60 Polynesia and Micronesia. They differ primarily in ft), nonetheless considerably deeper than practices that the CB hooks are often longer and sometimes elsewhere in Polynesia. Kamakau notes that to use more curved than the bonito hook. Emory et al. this method, the fisherman also had to be able to (1968) in describing excavated assemblages noted recognize the various disguises that the octopus the paucity of bone hooks relative to the number of employs to remain invisible on the bottom: stone sinkers and perforated cowrie shells. They sug- [t]he okilo he‘e method could be used only in gested that many octopus lure hooks were probably shallow seas from six to ten fathoms in depth, and made of wood, a supposition supported by ethno- where the water was clear; it was impossible where graphic specimens in Bishop Museum collections. the sea was dark. The fisherman who would use The toggle, used to hold fast the pulling line to the this method had to learn all the ways of the octo- cowrie shell, was also an integral part of the CB lure. pus before he could become skillful. He was not The toggle strengthens the attachment of this line after the octopus that just “lay by the fireplace” and keeps it from pulling out of, or breaking, the (waiho ka‘e kapuahi ), curled up in a ball—this an perforation in the cowrie shell. This reinforcement is unskilled novice could see. The ways of the octo- necessary because the fishing line is attached, not to pus are countless. It might conceal itself amongst the wooden shaft of the lure, but to the cowrie shell the pebbles, or close up its hole and thrust one itself. long tentacle out; or it might look like the mouth or head of an eel or of a sea urchin, or it might Lure Use-context. A consideration of the use-con- show only its beak; it might look as though it text of Hawaiian CB lures is also instructive with were marching along in a procession, or as respect to their origin and relationship to octopus though it were a blob of excrement. There are lures in Polynesia generally. Notably, the CB lure many other ways of the octopus that were known was distinctive from the IR lure, not only in terms of to the okilo he‘e fisherman (Kamakau 1976:69). design, but also in its application. Like the IR lure, the CB lure could be used in shallow water. Addi- To aid his search for an octopus, the fisherman tionally, the CB lure also could be used in deeper chewed and spat kukui nut (candlenut) on the water, waters where octopi were known to occur but could making an oily surface that improved his ability to not be directly observed. Historical accounts (Kahau- see below. Once the fisherman spotted an octopus, lelio 1902; Kamakau 1976) indicate that use of the he lowered the kilo lure down to its lair. When the octopus lure in Hawai‘i ranged from a depth of a octopus fastened onto the lure the fisherman, sens- few feet to 80–120 fathoms (480–720 ft)—much ing the weight and pull of the animal, jerked the lure deeper than elsewhere in the region. These accounts and hooked his prey. Interestingly, Kamakau (1976) are supported by the author’s own observations of reports that the lure consisted of only a stone sinker, 51 pfeffer
a stick with a hook on the end of it pointing towards octopus. It came up so fast through the water that the stone, a tuft of ti leaves or a few bright flowers, its head stood up straight and its tentacles trailed but with no cowrie shell. A specimen in the Bishop like the branches of a willow, wilou, tree. With Museum’s ethnographic collection conforms com- the fisherman shaking the lure, it was like an ‘ala- pletely with Kamakau’s description. ‘apapa hula, and many he‘e came to embrace the dancer, unaware of the hook underneath. The The third Hawaiian method, lu he‘e (referring to a octopus did not want the cowrie or the stone to up and down shaking motion), used a cowrie shell as eat; papa‘i and ‘ohiku crabs and other small Crus- the main bait for the octopus, and seems to have tacea (mea ‘ano papa‘i) were its food; but the fish- been used only in exceptionally deep waters. Kama- erman enticed it with a sort of hula, and the octo- kau states that this method was used in waters pus was “taken in” (ua puni) (Kamakau 1976:67). 80–120 fathoms deep (480–720 ft). While such depths seem exaggerated, a second early description A choice cowrie was given the name of a grand- transcribed by Mary Kawena Pukui (Kahaulelio parent, a father, a mother, a wife, or of a chief. 1902:15–24), reports an 80 fathom (480 ft) long Mulali was a famous cowrie of ka po‘e kahiko and cord used in the lu he‘e method. Kamakau (1976) so was Hualalahu. Long stories are told of these also notes that the lu he‘e lure was used in shallow, famous cowries and of how, when they were near-shore waters by ali‘i (nobility) in aristocratic merely shown alongside a canoe, the he‘e would sport fishing. just rise up and fill the canoe (Kamakau 1976:69).
Unlike elsewhere in the Pacific, where the success of These accounts point to the importance of octopus the IR lure is attributed to the hatred of the octopus fishing, including its value as sport for the aristocra- for the rat, in Hawai‘i Kamakau writes the following: cy. The use of specific kinds of stones for sinkers, combined (or “mated”) with specific cowrie shells, A fisherman would boast, “I will go after the he‘e demonstrates the care taken in constructing the lure. today—these are the days of rising tides.” Just as The weight portion of the lure was seen as an inte- a woman with lustful eyes (maka leho) entices gral part of the lure’s overall attractive power for many men, so a beautiful leho [cowrie] arouses catching octopi. Well-worked (shaped and polished) the desire of the he‘e, and two or three of them at sinker stones of a variety of materials are often found once might be pierced by the kakala hook, or in Hawaiian archaeological sites, supporting Kama- because they clung fast to the ‘amana, the wood- kau’s and Malo’s statements that certain lithic mate- en stem of the lure. rials were much sought after. A stone had to be used with the cowrie—a hand- Spatial-Temporal Distribution of some one, to enhance the loveliness of the Octopus Lures in Oceania “female,” the cowrie. The handsome stone was the “husband” to the cowrie, and the cowrie was The IR lure is found in various forms throughout “married” to the stone. When the two matched in Micronesia and Polynesia. It occurs in the Mariana beauty, and they swayed in dance in the ocean, Islands, Fiji, Tonga, Samoa, Tahiti, and the Mar- the he‘e came to watch the joyful dance. Those of quesas (Table 1; see also Leach 1979). Worked cow- them who wished to “kiss” (honi) the cowrie, rie shell dorsal caps excavated by Allen (pers. com., leaped to embrace and kiss her because they were 1993) may also indicate its presence in the southern aroused by the dance. When the fisherman saw Cook Islands. The CB lure, in contrast, may have one hug the cowrie, he braced himself and kept been limited to Hawai‘i and the Marquesas. shaking the lure. When the octopus took hold of Archaeological evidence provides temporal depth to the cowrie, the fisherman pulled up the cord these ethnographic records. Non-grooved IR lure swiftly with his right hand, grabbed it with his stones and grooved coffee-bean sinkers have been left hand, and pulled it hard against the side of recovered from archaeological sites in the Marquesas the canoe, which forced the kakala hook into the and Tahiti (Sinoto pers. com., 1993; Suggs 1961). 52 hawaiian archaeology
In the Marquesas, the conical, non-grooved IR vari- device. Traditions say that the octopus attacks the IR ety is restricted to the lower layers at Hane, while lure because long ago the rat defecated (or vomited) grooved, coffee-bean sinkers are found in upper lay- on its head. The following rendition comes from ers at Nukuhiva and elsewhere. Intriguingly, grooved Tutuila, Samoa (Buck 1930:438): conical IR lure stones are also found in the lower lay- ers of the Hane dune site and may be precursors to . . . the unga (hermit crab), the ve‘a (rail) and the the coffee-bean sinker stone used in CB lures. Suggs’ isumu (rat) planned a visit to see the red cliffs, (1961) found both coffee-bean sinker weights and Lenga-a-Taema, near the western end of the whole perforated cowrie shells. Composite hooks island . . . they decided to journey by sea...[but] similar to Hawaiian CB lure hooks were also found, they were struck by a hurricane (afa). As the craft but not identified as such, by Suggs (1961:79–94). foundered, the rail flew away to the land, the crab Informant interviews conducted by Suggs (1961:90) sunk to the bottom where it was quite at home, suggest that the CB lure form was known and used but the unfortunate rat was left swimming for its in the Marquesas until recent times. Suggs (1961:92) life. also identified a component of the Marquesan CB The octopus ( fe‘e) hearing the rat’s wailing, took lure not found on Hawaiian specimens, specifically pity and invited the rat to sit on his head whilst cowrie shell plates with single perforations. He iden- he conveyed him safely ashore. On the journey, tifies these as “flappers” and suggests that they the rat defecated on the head of the fe‘e without increased the effectiveness of the lure. Further work the latter knowing it. When safely ashore, the is needed to determine whether or not hooks and ungrateful rat taunted the octopus by drawing his toggles comparable to the Hawaiian forms were used attention to the insult. The enraged fe‘e could not in the Marquesas. pursue the rat on land but vowed if ever he Archaeological evidence also suggests that the CB lure caught the rat in the water again, he would may have been used in Aotearoa (Leach 1979) and avenge the insult. Tahiti (Sinoto, per.com. 1993). Coffee-bean shaped The hate of the fe‘e has become hereditary and stones have been found in both localities. Nonethe- man has utilized it to his own advantage by shap- less, coffee-bean-like sinker stones in the absence of ing and decorating a lure to represent the rat. whole perforated cowrie shells, diagnostic hooks, or bone toggles preclude a definite identification of the Versions of the legend are found in Fiji, Tonga, Niue, CB lure in these island groups; stone sinkers were in Samoa, and Hawai‘i (P. Nogelmeier pers. com., 1993; some contexts also used for other purposes. Given D. Shideler pers. com., 1993). Additionally, Clark the foregoing, hooked CB lures are only definitively (1976) has reconstructed the legend to Proto-Poly- known from Hawai‘i and the Marquesas. Interest- nesian, further indicating its wide distribution and ingly, in the Hawaiian case, both ethnographic and antiquity. Additional research is needed to establish archaeological evidence indicate that only the CB the occurrence of the legend vis-a-vis use of one or lure was present. both forms of lure. The Hawaiian evidence, howev- er, suggests that functional divergence in lure forms Sinoto and Kellum (1965) hypothesize that the con- was accompanied by loss of the legend—possibly ical IR lure sinker stone pre-dates the CB sinker form, because the morphological similarities no longer suggesting that the CB lure was a later and specifi- held. cally East Polynesian development. This hypothesis is supported by the present research as well, which Evolution of the CB Lure suggests that the hooked CB lure may have been As suggested above, the CB lure allowed exploitation restricted to Hawai‘i and the Marquesas. of both shallow and deep water habitats. In order to exploit these deeper water habitats effectively, sever- Legend of the Rat and the Octopus al design changes were necessary. Specifically, the IR The distribution of IR lures appears to be paralleled lure has no hook, or hook-like projection. As such, it by the folk account for why it is an effective fishing can be dragged behind a canoe with little chance of 53 pfeffer
snagging on the reef. On occasions when it did snag, Environmental Conditions Influencing the shallow waters facilitated easy removal and loss- Octopus Lure Design es were infrequent. What environmental conditions might have selected for the development of a CB lure? Consideration of The CB lure, on the other hand, could be used in the marine environments of those islands where the both shallow and deep water environments. Use in CB lure is known or suggested is instructive. Ethno- deep water environments was effected in part by the graphic accounts indicate that the IR form was used addition of a hook, which allowed the fisherman to primarily in shallow water reef environments, where snag octopi in very deep waters, beyond the range of octopi can be lured to the surface. Given that the human vision. However, the use of a hook required hooked CB lure will not work well if dragged in shal- further design changes: low water, or in areas with large branching corals, it 1) Because the lure now had both a luring and is reasonable to hypothesize that hooked forms devel- hooking function, the flexible coconut root or oped in an island setting where deeper waters were soft stick needed to be replaced by an inflexible available and regularly utilized. In this final section, I one, strong enough to hold the weight of an octo- briefly review the marine conditions of several key pus; East Polynesian island groups and the potential influ- ence of environment on octopus lure design. 2) The cowrie shell bait needed to be presented in such a way that the hook would always be locat- Hawai‘i. In Hawai‘i the CB lure was used in shal- ed below and pointed towards the shell, so as to low and deep waters where it was commonly jigged. snag the octopus; the addition of an inflexible The marine environment surrounding the main stick helped in this regard but another critical Hawaiian Islands is characterized by deep near-shore change was the placement of the hook on the waters. In many areas, the waters can be hundreds of same side as the cowrie shell. feet deep within several hundred yards of the coast. This is especially true around the geologically 3) The foregoing further required that the IR lure younger islands, such as Maui and Hawai‘i, as shell plates be replaced with whole cowries. The marine charts show. addition of a wood, coral, or bone toggle on the lure end of the fishing cord was needed, so that The areas with the greatest potential for marine the cord would not pull through the hole drilled exploitation in the Hawaiian Islands are the near- in the cowrie shell. shore waters between 20 and 200 ft in depth. Deep near-shore marine environments are not conducive 4) Use in deeper waters required the lure to be to spearing, netting, or other forms of shallow water constructed so that it would not catch on the bot- fishing strategies. Bottom fishing, trolling, and tom if jigged (bounced). This was accomplished angling are more effective in these deeper near-shore by putting the cowrie shell bait and the stone areas. Similarly, the hooked CB lure would have weight on opposite sides. If the weight was left on been more productive for catching octopi under the same side as the hook, the lure would fall these conditions. The effectiveness of the CB lure in over, or rest with the weight and hook down both shallow and deeper waters would also have when dragged on the bottom and the hook would been advantageous. snag, break, or be lost. However, with the weight moved opposite the cowrie bait and hook, the The prevalence of deep near-shore marine condi- lure would only touch the bottom on the stone tions in Hawai‘i does not explain the absence of the side. This was further facilitated by attaching the IR lure from Hawaiian artifact assemblages. On the lifting cord directly to the cowrie shell. This islands of Kaua‘i, O‘ahu, Maui, and small parts of forced the lure to consistently fall on the stone Moloka‘i and Läna‘i, there are extensive shallow side when dropped, and also served to present the water reef environments suitable for IR lure use. It lure with the hook pointing towards the cowrie may be that development of the CB lure, which shell bait. works in both shallow and deep water environments 54 hawaiian archaeology
precluded the need for IR lures. However, this does In each case where the CB lure has been definitively not explain the absence of the IR lure from early recorded or potentially recovered, near-shore condi- Hawaiian sites. The question of why the IR lure is tions are typified by deep water environments. The absent from Hawai‘i is intriguing, but remains unan- evidence suggests, preliminarily, that CB lures devel- swered. oped in response to these readily available deep water near-shore conditions and to the constraints The Marquesas. According to Sinoto and Kellum imposed by the relatively limited extent of shallow (1965), the volcanic Marquesas generally lack fring- water fishing grounds. ing reefs. The shoreline is predominantly rugged steep cliffs that continue offshore, resulting in deep near-shore conditions. There are some shallow reef Conclusions areas in these islands (notably near the Hane Dune site), but overall the Marquesas lack extensive shal- low fringing reefs. In the isolated shallow reef areas Based on the foregoing assessment of Polynesian of the Marquesas Islands, the IR lure would be effec- octopus lures, several conclusions are possible. tive, but the CB lure would be more effective in the 1) The non-hooked IR lure is a widely distributed more prevalent deep water areas. Spearing of octopi form, found in several Micronesian islands and in would also be possible in the shallow reef areas. parts of Polynesia, including Fiji, Samoa, Tonga, The Society Islands. The Society Islands have a Niue, Futuna, the Marquesas, Society, and possibly combination of both deep water and shallow reef the southern Cook Islands. It has not been defini- conditions. The reefs here are predominantly barrier tively identified in Hawai‘i, Rapa Nui, or Aotearoa. reefs, which create very shallow offshore areas sur- 2) The hooked CB octopus lure has been conclu- rounding extensive, deep water lagoons. Both types sively identified only in Hawai‘i and the Marquesas of lure might be expected in the Societies on func- Islands. Archaeological finds of CB lure-like compo- tional criteria. In the shallow offshore reef areas, the nents (e.g., coffee-bean sinkers, and possible hooks) IR lure might be the best apparatus, while in the suggest that it may also have been used in Aotearoa deep lagoonal areas and in waters outside the reef’s and Tahiti, but more evidence is needed to establish edge, the CB lure would be more effective. The its presence in these groups. common method of spearing octopi would also be effective in shallow reef areas. 3) Distributional, archaeological, and possibly lin- guistic evidence suggest that the IR form pre-dates Aotearoa (New Zealand). The marine environ- the CB lure. The presence of both the IR and CB ment of Aotearoa is markedly different than condi- forms of octopus lure in the Marquesas indicates a tions found in tropical Polynesia. Due to its south- possible origin here, although a Hawaiian origin ern latitude, Aotearoa’s marine environment lacks with subsequent diffusion to the Marquesas cannot extensive coral growth. The cold waters prohibit the be ruled out. growth of most coral, and taxonomic richness is lower than in tropical waters. The combination of a 4) In Hawai‘i, the CB sinker lure reached its most lack of coral reefs and deep near-shore waters make complex and varied form. Several variants of stone Aotearoa unsuitable for IR lure use. However, these sinkers were developed in Hawai‘i, including six conditions would be favorable for CB lure use, and varieties of coffee-bean sinkers, and two varieties of one might expect its occurrence here. Leach (1979) the bread-loaf sinker (unique to Hawai‘i). notes the presence of octopi in Aotearoan waters and 5) Ethnographic accounts, design features, and the reports possible CB lure hooks and coffee-bean ecological distribution of the CB lure suggest that it sinker stones. Anderson (pers. com., 1993), howev- varied significantly from the IR lure in use. The er, suggests that IR lure sinkers (non-grooved, coni- hooked CB lure was designed to catch octopus in cal sinker stones) may have also been found in New deep waters without snagging the bottom. Zealand sites, but mistakenly identified as “tops.” 55 pfeffer
Future work is needed to document the spatio-tem- Clark, R. 1976. Aspects of Proto-Polynesian syntax. Te poral distribution of octopus lure components Reo Monographs, Linguistic Society of New Zea- throughout the region. Such accumulating evidence land. Auckland. will provide a more accurate and detailed picture octopus lure variability. Distributional information Dunnell, R. C. 1978. Style and function: A funda- will be critical in testing hypotheses of the origins of mental dichotomy. American Antiquity 43:192–202. artifact forms, accounting for variability in terms of Emory, K. P., W. J. Bonk, and Y. Sinoto. 1968. style (homology) and function (analogy). Under- Fishhooks. B. P. Bishop Museum Special Publication standing these dimensions of material culture change 47. Honolulu. will, in turn, allow us to examine processes of change among island populations. Handy, E.S.C. 1923. The native culture in the Mar- quesas. B. P. Bishop Museum Bulletin 9. Honolulu.
Handy, E.S.C. 1932. Houses, boats, and fishing in the Acknowledgements Society Islands. B. P. Bishop Museum Bulletin 90. Honolulu. I would like to acknowledge the contributions made to this paper by thanking: Dr. Terry Hunt for point- Hjarno, J. 1967. Maori fish-hooks in southern New ing me in the right directions, and his comments and Zealand. Otago Museum Trust Board. Records of editorial suggestions; Dr. Melinda Allen for her the Otago Museum, Anthropology No. 3. Dunedin. helpful comments, editorial advice, and information Hocart, A. M. 1929. Lau Islands of Fiji. B. P. Bishop on the southern Cook Islands; Dr. Yosihiko Sinoto Museum Bulletin 62. Honolulu. for information on Tahitian and Marquesan sites; Dr. Atholl Anderson for information on Aotearoa; Kahaulelio, A. D. 1902. Fishing lore, Ka nupepa ku‘o and David Shideler for his assistance in researching ‘koa. MS., on file, B. P. Bishop Museum, Honolulu. the distribution of Polynesian octopus lures. Kamakau, S. M. 1976. The works of the people of old: Na hana a ka po‘e kahiko. B. P. Bishop Museum Special Publication 61. Honolulu. References Leach, H. 1979. The New Zealand octopus lure: Allen, M. S. 1992. Temporal variation in Polynesian Fact or fiction. In Birds of a feather: Osteological and fishing strategies: The southern Cook Islands in archaeological papers from the South Pacific in honour regional perspective. Asian Perspectives 31:183–204. of R. J. Scarlett, ed. A. Anderson, pp. 231–46. New Zealand Archaeological Association Monograph II, Allen, M. S. MS. Style and function in East Polyne- BAR International Series 62, Oxford. sian fishhooks: Understanding the origins and dis- tribution of morphological variability. Loeb, E. M. 1926. History and traditions of Niue. B. P. Bishop Museum Bulletin 32. Honolulu. Buck, P. H. 1930. Samoan material culture. B. P. Bishop Museum Bulletin 75. Honolulu. Malo, D. 1903. Hawaiian antiquities. B. P. Bishop Museum Special Publication 2. Honolulu. _____. 1944. Arts and crafts of the Cook Islands. B. P. Bishop Museum Bulletin 179. Honolulu. Nalimu, Rev. H. B. n.d. Different Kinds of Fishing. Translated from Hawaiian by Mary Pukui. Hawai- _____. 1964. Arts and crafts of Hawaii, Section VII, ian Ethnological Notes, vol. I:904 –24, Bishop Fishing. B. P. Bishop Museum Special Publication Museum Archives, Honolulu. 45. Honolulu. Rolett, B. V. 1989. Hanamai: Changing subsistence Burrows, E. R. 1936. Ethnology of Futuna. B. P. and ecology in the prehistory of Tahuata (Marquesas Bishop Museum Bulletin 138. Honolulu. Islands, French Polynesia). Ph.D. diss., Department of Anthropology, Yale University, New Haven. 56 hawaiian archaeology
Rudiger, J., and B. Smith. 1987. The art of Captain Thompson, L. M. 1932. Archaeology of the Marianas Cook’s voyages. Vol. I, The voyage of the Endeavor Islands. B. P. Bishop Museum Bulletin 100. Hono- 1768–1771. New Haven: Yale University Press. lulu.
Sinoto, Y., and M. Kellum. 1965. Preliminary report Tinker, S. W. 1978. Fishes of Hawaii: A handbook of on excavations in the Marquesas Islands, French the marine fishes of Hawaii and the central Pacific Polynesia. MS., on file, Department of Anthropo- Ocean. Honolulu: Hawaiian Service Inc. logy, B. P. Bishop Museum, Honolulu.
Suggs, R. C. 1961. The Archaeology of Nuku Hiva, Marquesas Islands, French Polynesia. Anthropological Papers of the American Museum of Natural History. Volume 49: part 1, New York. Editorial and subscription notices Notes for contributors Hawaiian Archaeology, founded in 1984, is published Articles on Hawaiian prehistory or archaeology, or by the Society for Hawaiian Archaeology, a registered that contribute to the advance of method and theory tax-exempt organization. as these apply to Hawai‘i, are considered for publication in Hawaiian Archaeology. Send three photocopies of the The Officers of the Board of Directors of the Society text and of illustrations or roughs to the Publications for Hawaiian Archaeology are: Committee of the Society for Hawaiian Archaeology. R. J. Hommon, President Preferred style is as you find it in this volume, and S. L. Collins, Vice-President generally follows The Chicago Manual of Style, 13th or S. A. Lebo, Secretary 14th Edition, for scientific and technical publications L. O. Miller, Treasurer using the short (B) form for bibliographic citations. Authors of articles accepted for publication will be The Publications Committee members are: asked to submit a copy of the text on a diskette format- T. S. Dye, Chair and S. L. Collins. ted for DOS and to provide camera-ready illustrations.
Copyright © 1995 Society for Hawaiian Archaeology. All rights reserved; no part of this publication may be Conventions for radiocarbon dates reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without the prior In citing radiocarbon dates, Hawaiian Archaeology written permission of the Board of Directors, Society for uses the following conventions: .. (before 1950) Hawaiian Archaeology. CRA (conventional radiocarbon age) indicates a radiocarbon age that (i) uses 5568 as the 14C half-life, (ii) was measured against the NBS oxalic acid standard, Address (iii) uses 1950 as the zero date for radiocarbon time, (iv) is normalized for ∂13C, and (v) has not been correct- ed for reservoir effects; ../.. indicates a measurement Hawaiian Archaeology, Society for Hawaiian Archaeology, that has been calibrated to calendar years. P. O. Box 23292, Honolulu, Hawai‘i 96823-3292 is the address for all matters relating to the Society and to Hawaiian Archaeology.
Current subscriptions
Hawaiian Archaeology is published annually and distributed to members of the Society for Hawaiian Archaeology. Current dues are:
Professional—$26 Regular—$22 Associate—$18
Membership applications are available from the Secretary of the Society for Hawaiian Archaeology.