CHAPTER FIVE NON-MARINE MOLLUSCS FROM THE ROCKSHELTER SEDIMENTS by Patrick V. Kirch T IERECONSmhUCTION AND INTERu ATON of fonmer environments using non-marine molluscs from archaeological sites and sediments has a long history in some regions, especially the British Isles (Evans 1972) and parts of North America (Bobrowsky 1984). In the Pacific Islands including Hawai'i however, analysis of non-marine molluscan shells from archaeological contexts has been applied only recently, although with significant results (e.g., Kirch 1975; Christensen and Kirch 1981, 1986; Christensen 1983, 1984). The Hawaiian Islands, due in part to their extreme geographical isolation and absence of both predats and competitors, witnessed impressive adaptive radiation in several faunal groups, including terrestrial and arboreal molluscs (Zimmerman 1948). The indigenous and endemic non-marine molluscan fauna of the Hawaiian Islands is estimated to have included about 1,061 endemic species. Many of these species are now extinct, particularly those from lowland habitats. Species tended to be highly localized, both geographically and to particular floral communities, making them good paleoenvironmental indicators. Unfortunately, several factors have hindered the application of landsnail analysis in Hawaiian environmental reconstruction: (1) the unsettled taxonomy of many groups, especially at the species level; (2) the absence of good ecological data for many taxa, especially those now extinct; and (3) the lack of malacologists experienced with the Hawaiian fauna, and willing to undertake the laborious task of sorting and identifying large assemblages of fossil or subfossil snail shells from archaeological contexts. Fortunately, the autior had some years ago received informal training in Pacific Islands malacology from Dr. Y. Kondo of the Bishop Museum, precipitating some early applications of non-marine molluscan analyses in archaeology (Kirch 1972, 1975). Subsequently, collaborative research between the author and Dr. Carl Christensen, formerly of the Bishop Museum Malacology Division, led to the development of standardized methods for analysis of landsnails from Hawaiian and other Pacific Islands archaeological contexts (Christensen 1983, 1984; Christensen and Kirch 1981, 1986; Kirch and Christensen, in press). Thus, when the Anahulu rockshelters were excavated in 1982, the extraction and identification of samples of non-marine molluscs from the shelter sediments was one of several techniques chosen to track the course of environmental change in the valley. METHODS AND MATERIALS Methods used in the analysis of landsnails from the Anahulu sites largely follow those reported by Christensen and Kirch (1986:53-6) and by Christensen (1983:450-51), the main difference being that in the Anahulu case, shells were extracted from the sediment samples by flotation rather than by wet sieving. Samples were analyzed from Sites D6-60 and -36. In both sites, the samples were taken as condnuous columns in 5-cm sampling increments, from cleaned stratigraphic profiles carefully selected to avoid featrs Anahidu or disturbances. (No continuous column samples could be taken from D6-58 because of stratigraphic disturbances.) These were the same columns used for the sediment and archaeobotanical analyses reported in chapters 3, 6, and 7. Indeed, the molluscs themselves were extracted using flotation by M. Allen as part of the initial stage of her archaeobotanical work (see chapter 6). After extraction from the column samples, the landsnail assemblages were identified by the author, with assistance and checking of all identifications by C. Christensen.1 Identifications were made both by reference to the published taxonomic literature, and by comparison to the type and synoptic collections of the Bishop Museum Division of Malacology. Where the condition of the specimens or the unsettled taxonomy of a particular group precluded precise identification, determination was made to the generic level only. The Anahulu specimens were deposited in the Bishop Museum collections, where they may be consulted as voucher specimens. In quantification, only intact shells, or fragments including the apex, were counted so as to avoid double counting of individuals. Results of the identification and quantification of non-marine molluscs from the D6-60 and -36 column samples are presented in tables 5.1 and 5.2. No adjustments in specimen counts have been made for differential sample sizes (due to varying volumes and weights of sediment from each 5-cm level), but a standardized shells per milliliter of sediment is provided as an index to the concentration of shells in any sample, and the volumes and weights of floated sedinent are also indicated. The environmental interpretation of these data will be presented following a systematic review of the taxa present in the rockshelter samples. SYSTEMATIC REVIEW Family HELICINIDAE Pleuropoma sp. A few specimens of an unidentified species of Pleuropoma were present in the D6-60 column, mostly at lower depths (especially 35-65 cm). This snail may be a variety of P. laciniosa, the most common of only two species known from the Ko'olau Mountain region of O'ahu Island (Neal 1934). Neal reports that these operculates are consistently terrestrial, living "in damp places in woods on dead leaves on the ground, some on damp rocks" (1934:4). Family HYDROBLIDAE Tryonia sp.? Four specimens of what may be a Tryonia sp. were found in the D6-60 column. Hydrobiids are aquatic snails, which could have derived from the nearby Anahulu StreanL Unfortunately, the Hawaiian representatives of the family are poorly known, and sparsely represented in the Bishop Museum collections. Thus, identification could not be taken to a more precise level. Family ACHAT1NELLIDAE The Achatinellidae is a relict family of pulmonate molluscs confimed in its distribution to the Pacific Basin, with Hawai'i being the center of greatest specific- level diversity (Cooke and Kondo 1960). Several genera are abundantly represented in the Anahulu samples. Laenllidea sp. Specimens of Lamellidea were abundant in all stratigraphic levels of both rockshelters. At least some of these shells may be of the species L. gracilis, which occurs commonly throughout the islands up to 300 m elevation. Cooke and Kondo note that Lamellidea is both terrestrial and arboreal (1960:182). L. gracilis is "one of the few native land snails commonly found in association with non-native plant communities" (Christensen and Kirch 1986:57). Elasmias sp. Examples of this genus were very rare, only two occuring at the base of the D6-60 column, and one in the D6-36 column. Cooke and Kondo remark that "Elasmias have been taken alive in the Hawiian Islands only at altitudes of several hundred to 4,000 or 5,000 feet, most frequently along the edges of native forests" (1960:221). They appear to be associated with wet forests, and are not common in lowland fossil deposits. Tornatellides spp. Specimens of Tornatellides spp. were more abundant in both columns than those of any other taxon. More than one species is present, but identifying members of this genus is exceedingly difficult, particularly with imure or broken shells. There are 48 endemic species in the Hawaiian Islands (Cooke and Kondo 1960:247). Christensen and Kirch observe that "little is known of the ecological preferences of species 11 take this opportunity to thank Dr. Christensen for his substantial assistance in this work, and for his pleasant collaboration over several years in the subject of archaeo-malacology. 74 Non-Marine Molluscs of Tornatellides," but that they appear to have "a rather broad tolerance" for chging ecological conditions (1986:57). Pacoflcella cf. baldwini A few specimens of P. baldwini are present in both columns, especially at the base of the Site D6-36 column. The species is distributed throughout the larger Hawaiian Islands (Cooke and Kondo 1960:167), but litde is recorded of its ecological preferences. Christensen (1983:435) notes that it may be tolerant of ecological disturbance. Auriculella sp. Two specimens of an unidentified species of Auriculella were recovered from the basal portion of the D6-60 column. The genus is endemic to the Hawaiian Islands, where it is arboreal, and usually confined to native wet forests. Achatinella spp. At least three species of Achatinella are represented, primarily from the basal levels of the D6-60 column, although there are also three individuals from the D6-36 column. One of these appears to be A. decora. Two other species can be distinguished from each other on the basis of their dextral and sinistral coiling, but could not be further identified with the material at hand. The polymorphic genus is endemic to the island of O'ahu (Cooke and Kondo 1960), where it has undergone extensive radiation (Welch 1938). All species of the genus are exlusively arboreal, and generally do not tolerate distubed conditions or exotic vegetation. Consequently, the genus is today restricted to the higher elevation forests of the Ko'olau and Wai'anae Mountains, and its conservation status is in question. Family AMASTRIDAE The family Amastridae is a relict group of pulmonates endemic to the Hawaiian Islands (Pilsbry and Cooke 1914-1916). Amastra spp. Two species of the large-shelled genus Amastra are present in the D6-60 column only, from the basal levels. One of these is from the subgenus Cyclamastra, while the other is an unidentified dextral species of Amastra (s.s.). "For the greater part, these were shells of the plains and low elevations, down to near sea level. With the passing of the low forests these Amastras have become extinct, with few exceptions" (Pilsbry and Cooke 1914-1916:20). Leptachatina spp. The Leptachatina are somewhat smatler members of the Amastridae, represented in low frequencies in both sites. As most Leptachatina species are extinct, little is known of their ecology, although they evidently were associated with lowland forests. Family PUPILLIDAE Lyropupa spp. Several species of both dextral and sinistral Lyropupa are represented in the collections, although the sinistral species is present only in Site D6-60. In the D6-60 column, Lyropupa is present throughout the sequence. The genus is terrestrial, associated with leaf litter or grass. Christensen and Kirch note that "dextral species of Lyropupa are generally restricted to arid low- elevation sites" (1986:58). The sinistral species, in contrast, "inhabit the humid forest zone where they may be found on mossy stones, fallen twigs, and dead leaves" (Pilsbry and Cooke 1918-1920:227). Nesopupa spp. Representatives of Nesopupa are found throughout the sediment columns of both sites, and include the species N. newcombi and N. dispersa, as well as at least one additional unidentified species. Both N. newcombi and N. dispersa are distributed throughout the main Hawaiian Islands. Both species appear to be associated with native dry forest (Christensen 1983:456-7). Pronesopupa sp. Only two individuals of an unidentified species of Pronesopupa were recovered, from the basal levels of the D6-60 column. Gastrocopta servilis G. servilis appears in the upper levels of both sites. It is a Neotropical species, first reported from the Hawaiian Islands in 1892 (Ancey 1892). The snail has been widely distributed throughout the islands of the Pacific region as a result of commerce (Christensen and Kirch 1986:59). Family ENDODONTIDAE Cookeconcha sp. This is an endemic Hawaiian genus for which the species-level taxonomy remains unsettled (Solem 1982:275-76), thus precluding identification of the specimens occuring tiroughout the D6-60 column. Eco- logically, Cookeconcha species "live on dead stumps and logs, and under the bark of dead trees, but also under fallen leaves" (Pilsbry and Vanatta 1906:783). 75 Anahsu - 0 eq 4 e ( - e + + + - oo 00 '0 IVa - C4 e in N q om e V- W- - - 't - V-4 C-I + -4 e v co veq C- e4 q v-4e4 - eq + C,4 - t- e4 eq 0 qe W- (f m + - - . + + N eq ~ V-_ - N V- *4 c eq -- t- t 0% - - c o IV e q t eq ( N cn 0 0 -4 0 W- - - - %0 4 a B. R -a b^ , %1 . s s . N~~~4 76 0 0 co oo 0 en 6 en CD 0 en II 0 0o 0 0 0 0A o- %n o- m o- . o-- %A 40 o 0 z z 0 z ;E Po Vz 42% N qq Non-Marine Molluscs -4 }s- -_1 _~ - -. V- -" eqi "- 77 00 0 Iz Ci4 Ci %O 0 ci 0%n I.C 0 N~ %OOci4 'Ir~ - r- % ~ C O rior 00 C 0%0 O %n tn v-%'s W- -- cm V- 0 _n O en 4 - 0 02 rA 02 rA 0~~~~~~~0 0 7 0 I ,~~~~~~~~r -o -O 0 W- 0% 0 00 en O t ON co in O _ _nv- - O 4 O in- V- '-4 00 l co 0 co 00 t'r) 0 0 ~o WI 0 0A %A %A U o '14- 0 0 CD) '14- 0 '-4- '-4- 0 000 '4,0 'n O4T4 0 0% r- W-) .o *_ 0 %D N V.-I 78Anaiu 4 t--- v'4. V- -4 00 C1 00 * - -o 'cr J. W-4 - + 00 + mr NO t N00 V-q V- - 0cn e14 %O 00O(D - -4 -04 0 o eN in W- 1- 't 00 00 o o - o- 00 C - _e4 - + - eq 00 - - + e - - I >- eI q -- 0C ii il % 3 P'iI* t E tji 3 3 | 4) I 'a In 0 C6 H 0 t 0 0 .-4 0 t 0 _ e z 0v 0 ino S e r_^ m N _14 ,A 1- W) o- 8 1- s; V- 0o z w 0 U, *A 0 q 0 CA u rA z Iu 'a U 78 Non-Marine Molluscs Endodonta sp. Seven specimens of an unidentified Endodonta were recovered from the basal levels of the D6-60 column. As Solem (1982:275-77) notes, the specific-level taxonomy of this group remain unsettled, with perhaps only 13% of the collected species having been described and named. The probable ecological niche of this endodontid is, as for other members of the family, "litter-and-leaf-mould forest dwellers" (Solem 1976:100). Family SUCCINEIDAE Succinea sp. Specimens of Succinea are found in both sites, especially in the lower half of the D6-60 column. Succinea are generally terrestrial, and are common in the lowland, more arid regions of the Hawaiian Islands. The Anahulu specimens may be S. caduca, a widely distributed species in the archipelago, associated with dry forest conditions, although a defimitive determination could not be made. Family HELICARIONIDAE Philonesa sp. A few specimens of Philonesia were present, from the base and top of the D6-60 column. The genus inhabits a wide range of habitats, and lacking a specific identification, no definitive ecological conditions can be inferred. Family ZONITIDAE Hawaiia sp. A few specimes of Hawaiia are present in the upper levels of the D6-60 column. As Christensen (1983:461) remarks, most authorities regard the Hawaiian representatives of Hawaiia as conspecific with the North American species H. minuscula, and an adventive in the Hawaiian Islands. The snail was first reported from Hawai'i in 1854. The restriction of this species to the upper levels of the D6-60 column would appear to reinforce this interpretation of the species as an historic introduction to the islands. Family FERRUSSACIIDAE Ceciliodes aperta Only one specimen of this species is present, from the base of the D6-60 column. The species is an historic introduction to the Hawaiian islands, and its presence deep in the D6-60 site can be explained by its burrowing habit. Family SUBUTLLDAE Lamellaxis clavulinum A single specimen of this exotic, widely distributed snail appears in the uppermost level of the D6-60 column. Lamellaxis gracilis A single specimen of L. gracilis was recovered from D6-60, but the species occurs throughout most of the D6-36 column. This adventive species is widely associated with indigenous cultivation systems in the Pacific, and is now known to have become established on many Pacific Islands in the prehistoric period (Christensen and Kirch 1981, 1986; Hunt 1981; Christensen 1984). Indeed, its presence in a landsnail assemblage is a good index of the presence of human activity and ecological disturbance. Prosopeas achadnacewn A few specimens of this historically-introduced species are present in the upper levels of both sites. Subulina octona Two specimens of this adventive, historically- introduced species are present from the top of the D6-36 column. Family BRADYBAENIDAE Bradybaena similari Two specimens of this large, adventive, historically-introduced species were recovered from the upper part of the D6-60 colulmn. PALEOENVIRONMENTAL INTERPRETATION The column from Unit D20 of Kuolulo rockshelter spans a period from the mid-fifteenth cenury up to the present, thus providing an opportunity to trace changing ecological conditions in the immediate vicinity of the site. The non-marine molluscan asemblages from this column have been plotted as a frequency diagram in figure 5.1, based on the data presented in table 5.1. (It should be noted that the two deepest samples (75-85 cm) consisted of only 11 and 13 specimens, and that sample size effects may have skewed the representation of particular taxa in these basal levels.) Before attempdng to interpret this column, we must ask whether the pattems evident in figure 5.1 may simply reflect differing sample sizes (see also discussion of sample size effects in chapter 6). In fact, there is lide correlation between number of 79 Ansaiu OSAJIUOAPV pOonpoJIUI- Agg.?lFol.I. s!xel,liw3 e1sau@Ol4d eeugoonS eBuopopu3 e4ouoo0a0oo - - 6l ~~~~~~~~ [o 1 J - [~~~ E~~~~~1 - - - edndo.ON (leiisiuis) edndojAi (Iejixep) edndojAi - - - [ C,O X 0 0 II * m eu!le44e3ldol e.asewv eI11ulIeI4oV iellelnoijnV S opIigoleujojl seiw.egs eepillgews, ewodoaneld luewipos Jo IHI/.I!,us IE0 _ KLo I - _ _ I 0 0 at 0 c el *o 0) to U cm 't (0 C \ I._ .V" CA, 0; as 0 .i 19 ;D El 0 0 0 ra a) a) 4-i 0z 80 Non-Marine Molluscs identified specimens (NISP) and number of identified taxa (NTAXA) in the D6-60 column (Pearson's r=0.317, r2=10%). Thus, changing patterns of taxonomic richness in the column are clearly meaningful in an ecological sense. On the other hand, the relative frequencies or abundances of particular taxa do appear to reflect sample size effects. Lamellidea and Tornatellides frequencies correlate significantly with NISP (Spearmen's rho values of 0.88 and 0.85, p=O.001, respectively). Thus, we must be very cautious in relying upon frequency data in interpreting the D6-60 landsnail sequences, and primary weight will be given to pattems of taxonomic richness (i.e., to nominal, or presence/absence, data). The D20 column can be readily divided into three major phases on the basis of the non-marine molluscan fauna, each zone representing a somewhat different local ecology. Zone 1 consists of the lowest six samples, from 55 cm to the base of the column, thus spanning the lowest 10-15 cm of the cultural deposit along with the underlying pre-occupation deposit. This zone is marked by the most diverse array of taxa in the entire column. Of particular note are the presence of Auriculella, Achatinella, Amastra, Elasmias, Endodonta, and Philonesia, taxa that do not occur in any samples higher in the column. In addition, the zone includes Lamellidea, Pleuropoma in some units, Tornatellides, Leptachatina, Lyropupa of both dextral and sinistral forms, Nesopupa, Cookeconcha, and Succinea. The presence of the arboreal Auriculella and Achatinella, of the large, extinct Amastra, and of such taxa as Pleuropoma, Elasnmias, Endodonta, and Philonesia that prefer wetter microenvironments can all be taken to indicate a dense (if not fully closed) canopy, forested environment at this time period. The overall picture indicated by Zone 1 is of an undisturbed mesic native forest. Zone 2 spans the 8 to 9 sample units from 55 cm up to 10/15 cm below surface, that is, the major portion of the prehistoric cultural deposit. Auriculella, Achatinella, Amastra, Elasmias, Endodonta, and Philonesia are absent from this zone. This zone is dominated by Tornatellides and Lamellidea, with lesser frequencies of Pleuropoma, Leptachatina, Lyropupa, Nesopupa, Cookeconcha, and Succinea. The upper part of the zone is marked by increasing frequency of Lamellidea and dextral Lyropupa, with decreases in Pleuropoma, Tornatellides, and Succinea. Zone 2 contrasts markedly with Zone 1, and indicates a period of significant local environmental change. In particular, the absence of arboreal snails suggests the removal or opening up of the former tree canopy, and the extirpation of taxa requiring wet microenvironments suggests a concomitant trend toward drier conditions. These changes were almost certainly the result of human interference with the local vegetation, probably due to clearing of forest, for shifting cultivations, in the vicinity of the site. It is noteworthy that the anthropophilic snail Lamellaxis gracilis, intimately associated with Polynesian gardening activity, appears in the 55-60 cm level of the column. Zone 3 consists of the upper two or three samples, from 0-10/15 cm below surface. This zone is marked by the presence, in high frequencies, of the historically adventive taxa Gastrocopta, Hawaiia, and subulinids, by a marked decrease in the native Tornatellides and dextral Lyropupa, and by the total disappearance of Succinea, Cookeconcha, sinistral Lyropupa, and Pleuropoma. This zone thus reflects the major environmental changes associated with the historic-period introduction of exotic vegetation, as well as the deforestation of the lowlands due to the introduction of grazing animals. The column from Ke'eke'e Iki rockshelter spans a shorter time period than that from D6-60, and thus correlates with Zones 2 and 3 of the latter column. The paleoenvironmental picture from D6-36 closely matches that summarized above for D6-60. In the D6-36 column, the lower 8 sampling units correlate with Zone 2, the period of prehistoric human interference in the mid-valley ecosystem. Thus, these samples lack such wet forest taxa as Auriculella, Amastra, Cookeconcha, Endodonta, or Philonesia, as well as the sinistral Lyropupa and the Pleuropoma. Also, the Polynesian anthropophilic snail Lamellaxis gracilis is present in most of these sample units, a good marker of gardening activity in the vicinity of the site. The upper two sample units (0-10 cm) correlate with Zone 3, with high frequencies of the historic adventives Gastrocopta, Prosopeas, Subulina, and Bradybaena. In sum, the non-marine molluscs contained within the rockshelter sediments of Sites D6-60 and D6-36 present a consistent picture of local environmental changes in the middle Anahulu Valley, from about the mid-fifteenth century up into the historic period. At the time Kuolulo Shelter was first occupied, the site was situated within a relatively undisturbed, closed or dense canopy native forest. This forest was quickly opened up through human clearance, presumably for shifting cultivation, with the more open conditions maintained throughout the prehistoric occupation period. There is some suggestion of increasingly open or arid conditions in the later part of the prehistoric period. A final phase of marked ecological change occured with the historic introduction of exotic flora and fauna. 81 82 REFERENCES CITED Ancey, C.F. 1892. Etudes sur la faune malacologique des Iles Sandwich. Memoirs de la Soci6tJ Zoologiques FranCaise 5:708-22. Bobrowsky, P.T. 1984. The hisaty and science of gastropods in archaeology. American Antiquity 49:77-93. Christensen, C.C. 1983. Analysis of land snails. In J. Clark and P. Kirch, eds., Archaeological Investigations of the Mudlane-Waimea-Kawaihae Road Corridor, Island of Hawaii, pp. 449-71. Anthropology Dept. Report 83-1. B.P. Bishop Museum. Honolulu. . 1984. Analysis of nonmarine mollusks. In A.R. Schilt, Subsistence and Conflict in Kona, Hawaii, pp. 355-76. Anthropology Dept. Report 84-1. B.P. Bishop Museum. Honolulu. Christensen, C.C., and P.V. Kirch. 1981. Non-marine molluscs from archaeological sites on Tikopia, Solomon Islands. Pac#flc Science 35:75-88. __. 1986. Nonmarine mollusks and ecological change at Barbers Point, O'ahu, Hawai'i. B.P. Bishop Museum Occasional Papers 26:52-80. Cooke, CM., Jr., and Y. Kondo. 1960. Revision of Tornatellinidae and Achatinellidae. B.P. Bishop Museum Bulletin 221. Honolulu. Evans, J.G. 1972. Land Snails in Archaeology. Seminar Press. London. Hunt, Tl. 1981. New evidence for early horticulture in Fiji. Journal of the Polynesian Soceity 90:259-69. Kirch, P.V. 1972. Subfossil non-marine gastropods from Molokal, Hawaiian Islands. The Nautilus 86(1):23. . 1975. Excavations at Sites A1-3 and A14: early settlement and ecology in Halawa Valley. In P.V. Kirch and M. Kelly, eds., Prehistory and Ecology in a Windward Hawaian Valley, pp. 17-70. Pacific Antrpological Records No. 24. Depart of Anthropology, Bishop MuseunL Honolulu. Kirch, P.V., and C.C. Christensen. in press. Extinct achaInellid lasnails from Easter Island: biogeographic, ecological, and archaeological implications. Burke Museum Contributions in Anthropology and Natural History. Neal, M.C. 1934. Hawaiian Helicinidae. BP. Bishop Museum Bulletin 125. Honolulu. Pilsbry, H.A., and C.M. Cooke, Jr. 1915-16. Manual of Conchology, Vol. 23. Appendix to Amastridae, Tornatellinidae. Academy of Naural Scienes. Philadelphia. . 1918-20. Manual of Conchology, Vol. 25, Pupillidae. Academy of Natual Sciences. Philadelphia. Pilsbry, H.A., and Vanatta. 1906. New Hawaiian species of Endodonta and Opeas. Proceedings of the Academy of Natural Sciencs, Philadelphia 1905:783-86. Solem, A. 1976. Endodontid Land Snailsfrom Pacific Islands. Part I: Family Endodontidae. Field Museum of Natural History. Chicago. . 1982. Endodontid Land Snailsfrom Pacific Islands. Part!!: Families Punctidae and Charopidae, Zoogeography. Field Museum of Natural History. Chicago. Welch, D'A.A. 1938. Distribution and Variation of Achatinella mustelina Mighels in the Waianae Mountains, Oahu. B.P. Bishop Museum Bulletin 152. Honolulu. Zimmerman, E.C. 1948. Insects of Hawaii, Vol. 1, Introduction. University of Hawaii Press. Honolulu.