Sweet Potato (Ipomoea Batatas) and Banana (Musa Sp.) Microfossils in Deposits from the Kona Field System, Island of Hawaii

Journal of Archaeological Science 36 (2009) 1115–1126

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

Sweet potato (Ipomoea batatas) and banana (Musa sp.) microfossils in deposits from the Kona Field System, Island of Hawaii

Mark Horrocks a,b,*, Robert B. Rechtman c a Microfossil Research Ltd, 31 Mont Le Grand Rd., Mt Eden, Auckland 1024, New Zealand b School of Geography, Geology and Environmental Science, University of Auckland, Private Bag 92-019, Auckland, New Zealand c Rechtman Consulting, LLC, HC1 Box 4149, Keaau, HI 96749, USA article info abstract

Article history: Analysis of plant microfossils (pollen, phytoliths, starch grains and xylem cells) was carried out on 12 soil Received 22 July 2008 samples from a variety of pre-contact archaeological features in South Kona within a portion of what has Received in revised form been termed the Kona Field System, on the Island of Hawaii. The oldest radiocarbon ages of the sampled 12 December 2008 deposits are 1300–1625 AD and 1310–1470 AD. The pollen and phytolith evidence suggests a change Accepted 18 December 2008 from more to less trees and shrubs in the area as a result of human activity. We found phytoliths of banana (Musa sp.) leaves in most and starch grains and xylem cells of tuberous roots of c.f. sweet potato Keywords: (Ipomoea batatas) in all of the samples, suggesting that the sampled features were associated with these Polynesian horticulture Plant microfossils crops and that both crops were cultivated intensively within the study area. Higher concentrations Introduced crops (volumetric) of starch and xylem in samples from older deposits suggest that cultivation was more Ipomoea batatas intensive then. The apparent absence of starch and xylem remains of other tuberous crops archaeo- Musa logically identiﬁed elsewhere in Polynesia suggests that tuberous cropping within the study area was Kona Field System mono-speciﬁc. Hawaii Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction 2005; McCoy and Hartshorn, 2007; Meyer et al., 2007). Many of these studies focused on productive increases in relation to During Polynesian settlement of the Pacific, agricultural prac- increased socio-political complexity. tices combining tropical Asian-Pacific and American crops were Pond fields are restricted mainly to the geologically older adapted to a diversity of island environments. The intensity of this islands, which have well-developed drainage networks. The reached its peak in the Hawaiian Islands (Vitousek et al., 2004). A dryland systems are found almost entirely on the younger islands of total of 26 plant species, mostly food crops, were introduced to the Hawaii and Maui, which have few streams. Their discrete locations islands by early colonists (Whistler, 1991), who arrived between are thought to reflect areas that represent ‘‘sweet spots’’ of rainfall c. 300 and 600 AD (Cordy, 2000). Large areas of irrigated pond fields and soil fertility (Meyer et al., 2007). Most of the dryland systems and of short-fallow, rain-fed dryland field systems were developed, were abandoned within a few decades after significant European reflecting the great environmental heterogeneity of the islands contact in the late 18th century as traditional Hawaiian societies (Rosendahl and Yen, 1971; Clark and Kirch, 1983; Pearsall and were disrupted (e.g., shifts in economic focus, disease-induced Trimble, 1984; Kirch, 1985, 1994; Clark, 1987; Ladefoged, 1993; population decline, demise of the traditional religious system) Rosendahl, 1994; Ladefoged and Graves, 2000, 2005, 2006, in press; (Vitousek et al., 2004). The archaeological remains of the dryland Allen, 2001, 2004; Ladefoged et al., 1996, 2003; Burtchard and systems’ stone field walls, numerous mounds and enclosures, and Tomonari-Tuggle, 2004; Coil, 2004; Kirch et al., 2004, 2005; stone lined trails mark their former extent. The distribution of Vitousek et al., 2004; McCoy, 2005; Mulrooney and Ladefoged, dryland fields was recorded in 1823 by Ellis (1963). Considering the extensive amount of landscape evidence for horticulture in the Hawaiian Islands, there is no large amount of direct evidence in the form of actual remains of cultivated plants. To * Corresponding author. Microfossil Research Ltd, 31 Mont Le Grand Rd., Mt Eden, our knowledge, the direct evidence published to date is limited to Auckland 1024, New Zealand. Tel.: þ64 9 630 5787; fax: þ64 9 373 7434. E-mail address: [email protected] (M. Horrocks). a few sites. Macrofossil fragments of the tuberous root of sweet URL: http://www.microfossilresearch.com potato (Ipomoea batatas) have been identified at Kohala on the

0305-4403/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2008.12.014 1116 M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 island of Hawaii (Rosendahl and Yen, 1971; Clark, 1983b; Ladefoged extending across 100þ jurisdictionally distinct Hawaiian tradi- et al., 2005) and at Kahikinui on Maui (Kirch et al., 2004, 2005). tional land units (ahupua’a). Rainfall and soil age vary considerably Also, pollen of sweet potato (Clark, 1983a) and coconut (Cocos within the Kona region. The median annual rainfall ranges from nucifera)(Ward, 2001), seed coats of candlenut (Aleurites moluc- 500 mm at sea level to 2500 mm at higher altitudes (Juvik and cana) and a charred stalk of Pacific Island cabbage tree (Cordyline Juvik, 1998). The ages of the frequent periodic lava flows, from fruticosa)(Lennstrom, 2001) have been identified at Kealakekua on which the local soils are derived, range from >5 ka to modern times the island of Hawaii. (Lockwood et al., 1988; Moore and Clague, 1991). Archaeological investigations at Ki’ilae and Kauleolı ahupua’a in The Kona Field System is characterised by a variety of archaeo- the southern part of the dryland Kona Field System on the big island logical stone structures, the most prominent of which are inland– of Hawaii (Fig. 1) by one of us (RR) during 2001–2007, present an seaward aligned field walls (kuaiwi)(Allen, 2004). Other features opportunity to look for further evidence of introduced Polynesian include terraces, mounds and enclosures. Based on early ethno- crops in Hawaii, using combined plant microfossil analyses. graphic accounts, Kelly (1983) outlined planting zones of Polyne- Reported here are results of analysis of microfossils in deposits sian introduced crops at Kona based on altitudinal variation in soil, from a variety of suspected agricultural and habitation features at rainfall and temperature. These are: Kula with sweet potato and five archaeological sites. Twelve soil samples were examined for paper mulberry (Broussonetia papyrifera) at 0–150 m altitude, 800– pollen, phytoliths and starch residues, to provide evidence of past 1200 mm rainfall; Kalu’ulu with breadfruit (Artocarpus altilus), environments, vegetation and horticulture. paper mulberry and sweet potato at 150–300 m, 800–1500 mm rainfall; ’Apa’a with taro (Colocasia esculenta), sweet potato, sugar 2. The study area cane (Saccharum officinarum), Pacific Island cabbage tree and banana (Musa sp.) at 300–750 m, 1500–2000 mm rainfall; and The Kona Field System (w140 km2) is situated on central ’Ama’u with banana at 600–900 m, 2000–2500 mm rainfall. Other western Hawaii Island, on the slopes of Hualalai (2100 m) and Polynesian crops introduced to the Hawaiian Islands, namely yams Mauna Loa (4000 m) (Fig. 1). These are currently active volcanoes (Dioscorea spp.), mountain apple (Syzygium malaccense), bottle and were so at the time of European contact. The field system is gourd (Lagenaria siceraria), turmeric (Curcuma longa) and coconut a narrow (w30 5 km) belt with an altitudinal range of 0–2500 m (C. nucifera), are also mentioned.

Fig. 1. Map of Kona Field System showing location of study area and traditional planting zones (Kelly, 1983). M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 1117

Newman (1974), who surveyed the Kona Field System primarily avocado (Persea americana), with a few loulu (Pritchardia spp.) and through aerial photography (Soehren and Newman,1968), seems to citrus trees, and an under-story of Christmas-berry (Shinus ter- have suggested that the fields existed as a cohesive unit. However, ebinthifolius), coffee (Coffea arabica), air plants (Bryophyllum pin- recent research and reinterpretation (see Rechtman et al., 2001) natum), waiawı (Psidium cattleianum), cabbage tree (C. fruticosa), have described the development of collections of widely distrib- liliko’i (Passiflora spp.), kakalaioa (Caesalpinia major) and hoi or uted agricultural fields over time and space into a loosely affiliated bitter yam (Dioscorea bulbifera), along with a variety of introduced socio-political system. In other words, the fields expanded under weeds and grasses. In Kauleolı, below 180 m elevation, the exposed the influence of individuals and small groups as the populations of lava flows support a tangle of koa haole (Leucaena leucocephala), Kona increased. As the Hawaiian socio-political system became opiuma (Pithecellobium dulce) and lantana (Lantana camara), with more centralised, more of the agricultural produce found its way occasional kiawe (Prosopis pallida). (through tribute) into the same coffers, but the fields continued to Agricultural features dominate the archaeological landscape of function independent of one another. Cordy (2000: 257–258) the study area, which is within the Kalu’ulu planting zone (Kelly, described the fields of Kona, albeit within the context of the Kona 1983). Of the 4773 archaeological features recorded, 95% are Field System, as follows: thought to be agriculturally related. These features were segregated into twelve sites (Fig. 2). Six of the sites, comprised of 4312 features Generally, it appears that the Kona field system gradually (95% of the agricultural features) are situated above 180 m eleva- formed, with small clearings in the wetter uplands and some tion. Many of these features may be associated with nineteenth use of the kula, beginning in some ahupua’a ca. A.D. 1000, and in century indigenous agricultural pursuits that were a continuation others as late as the A.D. 1400’s. Then over time – with growing of earlier practices. Archival and oral sources indicate that indige- populations, the chiefly centers, and other factors – the fields nous farming was practiced in Ki’ilae and Kauleolı from the early gradually expanded and intensified. This appears likely to have 1800s to the 1900s and beyond (Rechtman et al., 2001). Five of the taken place at different times in different ahupua’a. By the end of the A.D. 1700’s, the fields of all these lands could be seen by the European visitors as one big complex of near continuous fields. Also these were fields of individual communities with consid- erable variation and differences in extent. The archaeological sites remaining probably number in the thousands. The historically observed and archaeologically documented patterns of cultivation within the agricultural fields of Kona perhaps reflect common cultivation strategies and approaches that have developed in response to centuries of experimentation under the varied geomorphic and climatic conditions of the area. However, the concept of an agricultural system (with respect to defining the agricultural practices over a broad region) suggests that from one end of the region to the other, the agricultural features were temporally, functionally or synergistically interre- lated. Clearly this was not the case; that the products of these agricultural fields may have ended up (through tribute) in the same coffers tells us more about the workings of a socio-political system than it does about an agricultural one. Keeping in mind the diverse nature of the Kona Field System, the most comprehensive study of archaeological features within the Kona Field System published to date has been carried out at Kea- lakekua, one of the field systems’ more productive areas (Fig. 1) (Allen, 2001). This includes the aforementioned archaeo-botanical studies by Lennstrom (2001) and Ward (2001). In a subsequent study looking at risk management strategies as well as productive increases, Allen (2004) provided a fine-grained record of agronomic change from Kealakekua and compared this with patterns from the overall region. She suggested that risk management strategies, such as innovative garden architecture, novel cultivation practices and changes in the scale of agronomic integration were initiated by 1450 AD in even the most productive localities, and a shift to productive maximising strategies occurred after 1650 AD. Recent archaeological investigations in the southern portion of the Kona fields have focused on intensive surface survey, mapping, and excavation of selected features in Ki’ilae and Kauleolı ahupua’a (Rechtman et al., 2001; Loubser and Rechtman, unpublished data). Soils in the study area are characterised as stony and rocky peat, and rough broken land or exposed lava flow (Sato et al., 1973). At the time of the investigations the vegetation pattern within the study area followed the distribution of soil types. The upper portions (above 180 m elevation) of both ahupua’a support a tree canopy of kukui (A. moluccana), mango (Mangifera indica), monkeypod (Samanea saman), guava (Psidium guajava) and Fig. 2. Map of study area showing agricultural sites and sampling locations. 1118 M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 agricultural sites, accounting for only 228 features, exist below to have been cultivated in the prehistoric Pacific. Samples are from 180 m elevation. These lower features appear to be opportunistic herbariums at Auckland War Memorial Museum and Landcare sites taking advantage of localised soil areas in an otherwise soil Research, New Zealand; Forest Research Institute, Malaysia; and poor lava landscape. All of these sites have temporary habitation National Tropical Botanical Garden, Hawaii. Previous reports of features that are centrally located. These sites may represent phytolith morphology were also consulted (e.g. Mindzie et al., 2001; expansion of agricultural practices into marginal areas. An alter- Lentfer and Green, 2004; Ball et al., 2006; Piperno, 2006). native explanation is that these sites closest to the coast represent As well as starch grains, starch residue analysis includes calcium farmed areas proximate to coastal residential habitation. oxalate crystals and xylem cells, all three of which may be found The sampled features (n ¼ 7) reported in this study come from together in tuberous plant material (Torrence and Barton, 2006). five sites and include four enclosures, two platforms and a pave- Starch residues in this study were identified in three forms: ment (Table 1, Fig. 3). These features are dry-stacked rock parenchyma storage cells densely packed with starch grains, indi- constructions used for planting, processing and short-term habi- vidual starch grains and fragments of xylem cells. At least 200 of tation (see Rechtman et al., 2001 for detailed descriptions). these combined forms were counted for each sample. Tablets Prior to this study, standard radiometric age determinations containing a known quantity of exotic Lycopodium spores were were carried out on material associated with several of the samples added to samples to allow absolute counts (volumetric) of starch by Beta Analytic Inc., Miami, United States (Table 1). The results of and associated remains. Starch grain identification was carried out that analysis indicate that one feature (Site 23182, Feature 2) with the aid of a modern reference collection, which includes tissue accounting for five of the analysed samples may date from post- samples from the starch crops and many other plants known to contact times; and the remaining six features and their associated have been cultivated in the prehistoric Pacific. Previous reports of samples date from pre-contact times, possibly as early as the 15th starch grain morphology were also consulted (e.g. Reichert, 1913; century AD. Loy et al., 1992; Horrocks et al., 2007). Starch grains are normally colourless. The position of the hilum (developmental origin of the grain) can be seen under cross-polarised light; grains characteris- 3. Methods tically show birefringence with dark lines intersecting at this point (‘‘Maltese’’ cross). Well-preserved starch grains generally will have Numerous standard bulk archaeological sediment/soil samples retained their optical properties, however as grains degrade and were recovered in the original South Kona excavations (Rechtman lose their structure, the Maltese cross and growth rings progres- et al., 2001; Loubser and Rechtman, unpublished data). We selected sively disappear. Some other types of microfossils may be visible 12 soil samples (arbitrarily numbered 1–12) from this collection, under cross-polarised light. from the seven aforementioned archaeological features, to look for evidence of past environments, vegetation and horticulture (Table 1, Fig. 3). We did this using pollen, phytoliths and starch residue 4. Results analysis. Samples were given priority for analysis based on the existence of associated radiocarbon data and the presence of 4.1. Pollen and spores a stratified deposit within the features. Sampled features were typical of those within the project area and representative of the Pollen and spore preservation in the study samples ranged from types of features found within the greater Kona Field System. good to poor, with most showing signs of corrosion. Samples 7 and 9 Sub-samples (2–4 c.c.) were prepared for pollen (and spore) had insufficient pollen for counting. Samples 1–6 are over- analysis by the standard acetylation method (Moore et al., 1991). At whelmingly dominated (95%) by pollen of Cheno-Am (Chenopo- least 200 pollen grains and spores were counted for each sample diaceae and/or Amaranthaceae, the pollen of which is difficult to (except samples 7 and 9 which had insufficient pollen for counting). differentiate) (Fig. 4). Samples 8 and 10–12 are similarly dominated Slides were scanned for pollen types not recorded in the counts. by monolete fern spores. Small amounts of Myrtaceae, Pandanus Sub-samples (4 c.c.) were prepared for analysis of phytoliths and tectorius and Cichorieae pollen were also found in these four samples. starch residues by combined density separation (Horrocks, 2005). At A spore type tentatively identified as being from a truffle least 200 phytoliths were counted for each sample and slides were (hypogeous Ascomycotina) was found in samples 8,11 and 12 (Fig. 4). scanned for phytolith types not recorded in the counts. Phytolith This putative truffle spore resembles those of the Australasian truffle identification was carried out with the aid of a modern reference genera Labyrinthomyces, Dingleya, and Reddellomyces but is not collection, which includes tissue samples from many plants known identifiable to any known species (Trappe et al., 1992).

Table 1 Study samples and associated radiocarbon dates (sample numbers were arbitrarily assigned and do not represent a chronological sequence).

Sample no. Site no. Feature no. Layer/level Feature type Beta no. 14C BP Cal. AD (2s)a Intercept 1 23182 2 II/3 Enclosure 180392 100 70 Outside cal. range NA 2 23182 2 II/3 Enclosure 180392 100 70 Outside cal. range NA 3 23182 2 II/2 Enclosure 180392 100 70 Outside cal. range NA 4 23182 2 I/3 Enclosure 180393 210 50 1530–1950 1660 5 23182 2 I/4 Enclosure 180393 210 50 1530–1950 1660 6 23180 101 II/1 Enclosure NA NA NA NA 7 23184 1024 II/4 Enclosure 180381 490 80 1300–1625 1425 8 23184 1024 II/1 Enclosure 180381 490 80 1300–1625 1425 9 23184 1116 II/3 Pavement NA NA NA NA 10 23157 3 II/1 Platform 180383 270 60 1470–1950 1645 154222 420 60 1410–1640 1450 11 23145 C II/2 Platform 180391 510 60 1310–1470 1420 12 23145 C II/2 Platform 180391 510 60 1310–1470 1420

a INTCAL98 Radiocarbon Age Calibration, Stuiver et al., 1998. M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 1119

Fig. 3. Stratigraphic proﬁles of excavation units (*sampled layer). (a) Unit 51, Site 23182, Feature 2. (b) Unit 53, Site 23182, Feature 2. (c) Unit 62, Site 23180, Feature 101. (d) Unit 79, Site 23184, Feature 1024. (e) Unit 83, Site 23184, Feature 1116. (f) Unit 85, Site 23157, Feature 3. (g) Unit 93, Site 23145, Feature C.

4.2. Phytoliths dominated by chloridoid phytoliths (w40–80%) (Fig. 5). Panicoid phytoliths are also present (up to w25%). Samples 7–12 are Phytolith preservation was generally good, with only small dominated by irregular verrucose (up to w45%), spherical spinu- amounts of corroded phytoliths in the counts. Samples 1–6 are lose (up to w75%) and bulliform (up to w25%). Also, phytoliths of 1120 M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126

Fig. 3. (continued).

banana (Musa) leaves were found in small amounts (<5%) in samples 6, 8, 9, 11 and 12 (w60,000–200,000 per c.c). The samples 1, 2, 4, 7, 9, 11 and 12 (Fig. 7). Examples of banana leaf remaining samples had total concentrations of w10,000–30,000 phytoliths from modern reference material are shown in Fig. 8. per c.c. The densely packed cells were by far the most abundant, with individual grains and xylem cells equal to or less than 8000 4.3. Starch and other residues and 4000 per c.c., respectively. Morphological preservation of starch grains ranged from good We identified a single type of starch grain in all samples. These to poor. Most grains showed one or more of the following: soil- were up to 25 mm in diameter, spherical to sub-spherical, often bell- staining (yellow, amber, brown), expansion and distortion of the shaped, with a vacuole (often fissured) at the central hilum, and grain and vacuole, loss of the Maltese cross, pitting and cracking several flattened pressure facets, present as individuals or densely (Fig. 7c,d, and g). These features have been described in detail packed (several dozen grains) in ovate cell membranes. These are elsewhere (Horrocks et al., 2007). Identification of this type of consistent with root of sweet potato (I. batatas)(Figs. 6–8). We also starch grain is based on size and shape of the grain and features of found a single type of xylem cell (vessel) in all samples. These have the hilum. Of the Pacific starch crops, only I. batatas appears to have alternate slit-like pits and are likewise consistent with root of sweet grains with a central vacuole (Loy et al., 1992). However, although potato. The microfossils showed the highest concentrations in starch taxonomy of economic plants is well known and many M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 1121

Fig. 4. Percentage pollen diagram from Kona (þ¼found after count). species have distinctive starch grain morphology (e.g. Reichert, bioturbation or mechanical disturbance, or a combination of these. 1913), the starch taxonomy of non-economic plants is not as well In conclusion, we note that our interpretations based on the known. We therefore cannot unequivocally rule out all local plants microfossils, especially regarding timing of events, are made with as possible sources. Given this and the corroded nature of most of caution. the observed starch grains, our identification is made with caution. The generally corroded nature of the pollen grains and fern Haslam (2006) noted that some types of fungal spores may be spores in the study samples reflects the free-draining local soils, difficult to distinguish from starch grains because they show allowing aerobic activity by soil microorganisms and decomposi- a similar cross under cross-polarised light. However, these are tion of organic material. The abundance of monolete fern spores in generally much smaller than starch grains of the Polynesian starch samples 8 and 10–12 indicates disturbed vegetation but also in crops, with the exception of taro, which does not feature in our large part is the result of the superior preservation of fern spores c.f. study. The xylem cells in our study were mostly fragmented and pollen (Dimbleby, 1985)(Fig. 4). Many of the Hawaiian Islands’ fern showed similar soil-staining, occasional cracking, cross-wise frag- species have this spore type (bean-shaped) (Selling, 1946), a lot of mentation, with wall pits becoming progressively less visible or which are difficult to differentiate, especially if corroded. The great showing distortion, expansion and coalescence (described in detail abundance of Cheno-Am pollen in samples 1–6 probably reflects in Horrocks et al., 2007). their generally younger age (Table 1), with major weed invasions as a result of increased forest clearance and perhaps extended fallow 5. Discussion and conclusions periods (see also phytolith and starch results, below). This disturbance-related taxon is wind-pollinated (as opposed to animal- Theoretical processes of pollen incorporation into soils and pollinated) therefore produces abundant pollen. The indigenous other better-drained substrates were modelled by Dimbleby Chenopodium oahuense is a likely candidate although European (1985). Pollen is deposited on the ground surface and carried introduced species are also possibilities. downwards through the soil by percolating groundwater. The The Myrtaceae pollen in three of the samples may be from percolating profile mainly conforms to the chronological order of a number of indigenous species, with the higher values in samples pollen deposition at the surface with older pollen found increas- 8 and 10 possibly reflecting earlier times with more trees and ingly towards the bottom of the profile. Soils conforming to the shrubs present (Fig. 4). The Polynesian introduced mountain apple Dimbleby (1985) model will most likely be acidic, as this condition (S. malaccense) is also a possible candidate but pollen corrosion reduces the activity of soil invertebrates that may homogenise generally precluded confident differentiation between Myrtaceae profiles. The rate at which these processes operate depends on such taxa. Pollen of P. tectorius, another important economic plant and factors as soil type and drainage, rainfall and vegetation. Preser- possibly indigenous to Hawaii (Whistler, 1991), in samples 8 and 10 vation and movement of phytoliths and starch grains in soils have indicates that it grew and was perhaps cultivated in the area. been discussed more recently (e.g. Therin, 1998; Haslam, 2004; As Hawaii has no indigenous Cichorieae (Lactuceae), pollen of Piperno, 2006). As phytoliths are non-organic, they often show less this disturbance-related taxon in samples 10 and 12 suggests either corrosion than organic microfossils (as in this study). Starch grains Polynesian introduction, probably accidental (although the Sonchus may survive for long periods by their sheer abundance (e.g. 1 g of shoots are edible) or contamination with pollen of historical age Zea mays (maize) starch comprises 1,000,000,000 grains (Swinkels, (Fig. 4). The distribution of Sonchus in Polynesia is unclear and 1985)), decreased surface area exposed to degrading agents due to complicated by early confusion between Sonchus asper and Sonchus clumping, and protection from degrading agents by cell walls. oleraceus. In Hawaii S. oleraceus was naturalised before 1871 while In cultivated soils at archaeological sites such as ours, preser- S. asper is thought to be a more recent introduction (Wagner et al., vation and movement of starch grains and phytoliths may be dis- 1990; Leach, 2005). rupted by burial and mixing by cultivation. Thus, we assume that an Ward’s (2001) pollen study from Kealakekua examined several undetermined degree of mixing of plant remains of different ages test-pit profiles. One of these extended into very early Polynesian or in the South Kona deposits has occurred by percolation, perhaps pre-settlement times, with initial large amounts of grass 1122 M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126

Fig. 5. Percentage phytolith diagram from Kona (þ¼found after count).

(Poaceae) and sedge (Cyperaceae) pollen followed by an increase in large amounts of these in samples 1–6 suggest a reduction in trees monolete fern spores, little Cheno-Am pollen and increased vege- and shrubs. tation disturbance. However, in the Lalamilo Field System study, In a phytolith study of soil proﬁles within the Lalamilo Fields, Bennett (1983) found high proportions of Chen-Am, grass and Pearsall and Trimble (1984) linked disturbance to indigenous Asteraceae (Compositae) pollen. vegetation and subsequent cropping/fallowing. This was evidenced Phytoliths provide further evidence of the past vegetation by a decline of large cell grass phytoliths (which include bulliform) within the study area. Spherical spinulose phytoliths occur in palms and coincident increase in short cell phytoliths (which include (Arecaceae) and bromeliads (Bromeliaceae) (Piperno, 2006). In our chloridoid and panicoid). With generally larger amounts of bulli- case these phytoliths are most likely from Pritchardia spp. palm, the form phytoliths in older samples and higher chlorodoid and pani- Hawaiian Islands’ only indigenous representative of this family. coid phytoliths in younger samples, our results show a similar trend However, the introduced coconut may also be a contributor. The (Fig. 5). islands have no indigenous bromeliads. The abundance of palm Phytoliths of banana leaves are diagnostic (Mindzie et al., 2001; phytoliths in the older samples (7–12), coincident with large Lentfer and Green, 2004; Ball et al., 2006). All bananas in Remote amounts of irregular verrucose phytoliths (probably mainly from Oceania are human introductions. They include distinctive culti- dicotyledonous trees) supports the pollen evidence of more trees vars, belonging to the sections Musa (formerly Eumusa) and Calli- and shrubs present in earlier times (Fig. 5). The grass phytolith musa (formerly Australimusa) of the genus Musa that were evidence also supports this; chloridoid, festucoid, panicoid and prehistoric introductions (Simmonds, 1959; Whistler, 1991; Ken- bulliform phytoliths are exclusively from grasses (possibly nedy, 2008). The presence of these in most of the samples (58%) including sugar cane). As grasses are typically shade-intolerant, the suggests that the sampled archaeological features were associated

Fig. 6. Concentration starch diagram from Kona. M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 1123

Fig. 7. Plant microfossils from Kona (a, b mounted in Caedax; c–h mounted in glycerol jelly; 100,400 or 600; scale bars: 20 mm). (a, b) Musa leaf phytoliths showing rectangular/squarish base with protuberances all along the sides of the base and truncated cone (c.f. Fig. 8a, b). Of the three examples shown here, two are viewed looking down into the cone (a, right panel; b) and the other (a, left panel) looking at the cone side-on. (c) Low magniﬁcation view of soil-stained dense material within ovate cell membranes (four) (c.f. parenchyma starch storage cells of Ipomoea batatas, Fig. 8c). (d) High magniﬁcation view of fragment of ovate cell showing densely packed starch grains (membrane has been lost) (c.f. parenchyma storage cells of I. batatas, Fig. 8c). Individual soil-stained starch grains are more easily seen in cells that are disintegrating than in intact cells because the grains are tightly packed in the latter. (e, f) Faceted starch grains with central vacuole (c.f. starch grains of I. batatas, Fig. 8d, e). (f) Shows the grains in (e) viewed under cross-polarised light, showing central Maltese cross. (g) Soil-stained starch grain with central vacuole (c.f. starch grains of I. batatas, Fig. 8d). Although facets are not clearly visible, two straight edges of these are. (h) Xylem vessel element with alternate slit-like pits (c.f. xylem vessel element of I. batatas, Fig. 8f). 1124 M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126

Fig. 8. Modern reference samples ((a, b mounted in Caedax; c–f mounted in glycerol jelly; 400 or 600; scale bars: 20 mm). (a, b) Musa (AAB group) leaf phytoliths. For detailed descriptions see Mindzie et al. (2001) and Ball et al. (2006). Musa leaf phytoliths consist of a base (mostly rectangular or squarish) with protuberances all along the sides of the base and a raised truncated cone or sub-cylinder (crater). Right panel of (b) shows three phytoliths in a chain, as they form in the leaf. Of the five examples shown here, four are viewed looking down into the crater (b) and the other (a) looking at the crater side-on. (c) Parenchyma storage cell of Ipomoea batatas root, with densely packed, immature starch grains. Storage cells are typically ovate, up to w115 mm diam., with thin walls, <0.5 mm thick. (d) Typical starch grains of I. batatas root. Grains are spherical to sub-spherical, often bell- shaped, up to w26 mm diam., with a vacuole at the central hilum (seen here as a dot) and up to six flattened pressure facets. Reichert (1913) described the hilum as ‘‘eccentrically placed. in or slightly to one side of the median line’’. However, as this appears more or less central, to avoid confusion with the several Pacific starch taxa that have elongated grains with very eccentric hila (i.e. well off the median line, e.g. four of the five Remote Oceanic Dioscorea spp.), we describe the Ipomoea hilum as central, as did Loy et al. (1992). Many of the grains shown here are more mature than those in (c), hence are larger. (e) Starch grains in left and upper right panels of (d) viewed under cross-polarised light, showing central Maltese crosses. (f) Xylem vessel element of I. batatas root. Vessel elements are up to c. 90 mm across with walls up to 5 mm thick, and alternate slit-like pits up to 20 mm across. with this crop, and that banana was cultivated intensively within starch grains and xylem cells found in the cultural deposits with the study area. equivocal remains of another cultivated crop (i.e. Musa, see above) The presence of c.f. sweet potato microfossils in all of the from almost certain agricultural features are most likely sweet sampled features suggests that this introduced crop was also potato. The very high concentrations of starch storage cells c.f. cultivated intensively within the study area (Fig. 6). The very high individual starch grains in all samples are probably due to protec- concentrations of these in some of the older samples suggest more tion from soil microorganisms by dense packing within an intensive cultivation in earlier times, with a later reduction possibly enclosing cell membrane. as a result of longer fallow periods. Identifying Ipomoea starch in Given the phytolith evidence of banana, we consider the the Hawaiian Islands (and elsewhere) is complicated by the pres- possibility of banana starch within the samples. Starch grains of ence of several indigenous species of Ipomoea (e.g. Ipomoea indica). banana fruits (fan-shaped) are readily differentiated from those of However, these do not have tuberous roots like sweet potato, so the sweet potato (Reichert, 1913; Fullagar et al., 2006). As the fruit M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126 1125 ripens the starch loses its structure. We would not expect starch in in zones determined by rainfall and moisture. Near the dry seacoast aerial storage organs such as banana fruits to have the same potatoes were grown in quantity. Up to 1,000 feet grew small potential for archaeological preservation that tubers and root crops bananas which rarely fruited. from 1,000 to 3,000 feet, they have because the latter are presumably mostly already buried at prospered increasingly. From approximately 1,000 to 2,000 feet, time of death. This would reduce their rate of decomposition c.f. breadfruit flourished.’’ organic material such as bananas that had fallen onto the ground surface. Although the banana phytoliths are also from aerial tissue Acknowledgements (leaves), because they are not organic they have a better chance than organic plant material of preservation and eventual incorpo- We thank both Ki’ilae Estates LLC for supporting the project and ration in the soil if deposited on the ground surface. Relatively the archaeologists who conducted the field investigations (partic- limited preservation of starch grains in aerial storage organs also ularly Matthew R. Clark and J. David Nelson). We also thank Amy applies to the other main starch fruit crop of the early Hawaiian Ketner for help in preparing the manuscript. We extend our sincere Islanders, breadfruit. This has starch grains similar in shape to some mahalo to the native families of the study area. of sweet potato but distinctly smaller and lacking the distinctive Ipomoea vacuole (Fig. 8d). Apparent absence of other starch taxa in the study samples may References be significant. The other major tuberous crops introduced by early Polynesians to consider in a Hawaiian context include three species Allen, M.S. (Ed.), 2001. Gardens of Lono: Archaeological Investigations at the Amy B. Greenwell Ethnobotanical Garden, Kealakekua, Hawaii. Bishop Museum Press, of yam (Dioscorea alata, D. bulbifera and Dioscorea pentaphylla), two Honolulu. aroids (Araceae): taro (C. esculenta) and elephant ear taro (Alocasia Allen, M.S., 2004. Bet-hedging strategies, agricultural change, and unpredictable macrorrhiza), turmeric (C. longa), ginger (Zingiber zerumbet) and environments: historical development of dryland agriculture in Kona, Hawaii. J. Anthropol. Archaeol. 23, 196–224. arrowroot (Tacca leontopelaloides)(Whistler, 1991). Starch grains of Ball, T., Vrydaghs, L., Van Den Hauwe, I., Manwaring, J., De Langhe, E., 2006. the three Dioscorea species (elongated, ovate and fan-shaped, Differentiating Musa phytoliths: wild and edible Musa acuminata and Musa respectively) are readily differentiated from the smaller, spherical balbisiana. J. Archaeol. Sci. 33, 1228–1236. to sub-spherical, faceted grains of sweet potato (Loy et al., 1992; Bennett, T.M., 1983. Analysis of pollen and fern spores. In: Clark, J.T., Kirch, P.V. (Eds.), Archaeological Excavations of the Mudland-Waimea-Kawaihae Road Fullagar et al., 2006) and both taxa have been identified in Corridor, Island of Hawaii: an Interdisciplinary Study of an Environmental archaeological deposits elsewhere in Polynesia (e.g. Horrocks and Transect. Bernice Pauahi Bishop Museum, Honolulu, pp. 498–513. Report 83-1. Lawlor, 2006; Horrocks and Wozniak, 2008; Horrocks et al., 2008). Burtchard, G.C., Tomonari-Tuggle, M.J., 2004. Agriculture on leeward Hawaii Island: the Waimea agricultural system reconsidered. Hawaii. Archaeol. 9, 50–73. Turmeric starch grains are also fan-shaped (Reichert, 1913), as are Clark, J.T., 1983a. Archaeological investigations of agricultural sites in the Waimea those of ginger. Taro starch grains (at 3 mm in diameter relatively area. In: Clark, J.T., Kirch, P.V. (Eds.), Archaeological Excavations of the Mudland- tiny) are likewise readily differentiated from sweet potato and Waimea-Kawaihae Road Corridor, Island of Hawaii: an Interdisciplinary Study of an Environmental Transect. Bernice Pauahi Bishop Museum, Honolulu, pp. although starch grains of elephant ear taro may be similar in size 293–314. Report 83-1. and shape to sweet potato, like breadfruit they lack the distinctive Clark, J.T., 1983b. Radiocarbon chronology: report 9. In: Clark, J.T., Kirch, P.V. (Eds.), central vacuole (Loy et al., 1992)(Fig. 8d). Again, arrowroot has Archaeological Excavations of the Mudland-Waimea-Kawaihae Road Corridor, Island of Hawaii: an Interdisciplinary Study of an Environmental Transect. similar-sized and -shaped grains to sweet potato but lacks the Bernice Pauahi Bishop Museum, Honolulu, pp. 317–324. Report 83-1. vacuole. In a similar Easter Island study, abundant starch grains of Clark, J.T., 1987. Waimea-Kawaihae, a Lee-ward Hawaii Settlement System. Ph.D. D. alata and small amounts of sweet potato and taro starch were diss., University of Illinois, Urbana-Champaign. Clark, J.T., Kirch, P.V. (Eds.), 1983. Archaeological Excavations of the Mudland- identified in dryland basaltic deposits (Horrocks and Wozniak, Waimea-Kawaihae Road Corridor, Island of Hawaii: an Interdisciplinary Study 2008). Thus, if yams and aroids had been intensively cultivated of an Environmental Transect. Bernice Pauahi Bishop Museum, Honolulu Report within the present study area, we would have expected to find their 83-1. microfossils. Coil, J.H., 2004. ‘‘The Beauty that was’’: Archaeological Investigations of Ancient Hawaiian Agriculture and Environmental Change in Kahikinui, Maui. Ph.D. diss., Regarding the xylem evidence, we use this to support the starch University of California, Berkeley. grain evidence; other plants have xylem cells similar to sweet Cordy, R., 2000. Exalted Sits the Chief. The Ancient History of Hawai’i Island. Mutual potato (but not yams or aroids; see Horrocks and Barber, 2005; Publishing, Honolulu. Dimbleby, G., 1985. The Palynology of Archaeological Sites. Academic Press, London. Horrocks and Lawlor, 2006). However, in our case this was the only Ellis, W., 1963. Journal of William Ellis, Narrative of a Tour of Hawaii, or Owhyhee. type of xylem found. Lack of pollen of the starch crops we identified Advertiser Publishing, Honolulu. is not unexpected; fossil sweet potato (c.f. Clark, 1983a) and banana Fullagar, R., Field, J., Denham, T., Lentfer, C., 2006. Early and mid Holocene tool-use and processing of taro (Colocasia esculenta) and yam (Dioscorea sp.) and other pollen finds are extremely rare due to low pollen production. plants at Kuk Swamp in the highlands of Papua New Guinea. J. Archaeol. Sci. 33, The results of the current study demonstrate the potential of 595–614. using combined plant microfossil analyses of archaeological Handy, E.S.C., Handy, E.G., Pukui, M., 1991. Native planters in Old Hawaii: Their Life, Lore and Environment. In: Bishop Museum Bulletin, revised ed., vol. 223. Bishop deposits from the dryland field systems of Hawaii to identify crop Museum Press, Honolulu. type and range. The ubiquity and exclusivity of sweet potato and Haslam, M., 2004. The decomposition of starch grains in soils: implications for banana in our samples suggest a focus on these crops to the archaeological residue analyses. J. Archaeol. Sci. 31, 1715–1734. Haslam, M., 2006. Potential misidentification of in situ archaeological tool-residues: exclusion of other common, especially tuberous, Polynesian culti- starch and conidia. J. Archaeol. Sci. 33, 114–121. vars. These data seem to support the model of crop specific Horrocks, M., 2005. A combined procedure for recovering phytoliths and starch resource zones identified in the ethno-historic record. The current residues from soils, sedimentary deposits and similar materials. J. Archaeol. Sci. study area lies within the Kalu’ulu zone (Kelly, 1983), which was 32, 1169–1175. Horrocks, M., Barber, I., 2005. Microfossils of introduced starch cultigens from an reportedly dominated by the cultivation of sweet potato. The other early wetland ditch in New Zealand. Archaeol. Oceania 40, 106–114. crop that was reported as common in this zone was breadfruit; Horrocks, M., Campbell, M., Gumbley, W., 2007. A short note on starch and xylem of however no evidence of this cultigen was identified in the study Ipomoea batatas (sweet potato) in archaeological deposits from northern New Zealand. J. Archaeol. Sci. 34, 1441–1448. samples. While it is possible in our case that breadfruit was invis- Horrocks, M., Lawlor, I., 2006. Microfossil analysis of soil from Polynesian stone ible in the archaeological record, it is also possible that local fields in South Auckland, New Zealand. J. Archaeol. Sci. 33, 200–217. geological and climatological conditions were more favourable for Horrocks, M., Smith, I.W.G., Nichol, S.L., Wallace, R., 2008. Sediment, soil and plant microfossil analysis of Maori gardens at Anaura Bay, eastern North Island, New the cultivation of banana. As Handy et al. (1991: 525) pointed out, Zealand: comparison with descriptions made in 1769 by Captain Cook’s expe- ‘‘In the time of intensive native cultivation, South Kona was planted dition. J. Archaeol. Sci. 35, 2446–2464. 1126 M. Horrocks, R.B. Rechtman / Journal of Archaeological Science 36 (2009) 1115–1126

Horrocks, M., Wozniak, J.A., 2008. Plant microfossil analysis reveals disturbed forest Meyer, M., Ladefoged, T.N., Vitousek, P.M., 2007. Soil phosphorus and agricultural and a mixed-crop, dryland production system at Te Niu, Easter Island. development in the leeward Kohala Field System, Island of Hawaii. Pac. Sci. 61, J. Archaeol. Sci. 35, 126–142. 347–353. Juvik, S.P., Juvik, J.O., 1998. Atlas of Hawaii. University of Hawaii, Honolulu. Mindzie, C.M., Doutrelepont, H., Vrydaghs, L., Swennen, R.L., Swennen, R.J., Kelly, M., 1983. Na Mala o Kona: Gardens of Kona, A History of Land Use in Kona, Beeckman, H., de Langhe, E., de Maret, P., 2001. First archaeological evidence of Hawaii. Department of Anthropology, B.P. Bishop Museum, Honolulu. banana cultivation in central Africa during the third millennium before present. Kennedy, J., 2008. Pacific bananas: complex origins, multiple dispersals? Asian Veg. Hist. Archaeobot. 10, 1–6. Persp. 47, 75–94. Moore, R.B., Clague, D.A., 1991. Geologic Map of Hualalai. In: Miscellaneous Investi- Kirch, P.V., 1985. Feathered Gods and Fishhooks. An Introduction to Hawaiian gations Series. US Department of the Interior. US Geological Survey. Map I-2225. Archaeology and Prehistory. University of Hawaii Press, Honolulu. Moore, P.D., Webb, J.A., Collinson, M.E., 1991. Pollen Analysis, second ed. Kirch, P.V., 1994. The Wet and the Dry: Irrigation and Agricultural Intensification in Mulrooney, M.A., Ladefoged, T.N., 2005. Hawaiian Heiau and agricultural production Polynesia. University of Chicago Press, Chicago. in the Kohala dryland field system. J. Polynesian Soc. 114, 45–67. Kirch, P.V., Coil, J.H., Hartshorn, A.S., Jeraq, M., Vitousek, P.M., Chadwick, O.A., 2005. Newman, T.S., 1974. Kona Field System. Hawaii Register of Historic Places Nomi- Intensive dryland agriculture on the leeward slopes of Maui, Hawai’i: archae- nation Form. Site 10-37-6601. ological, archaeobotanical, and geochemical perspectives. World Archaeol. 37, Pearsall, D.M., Trimble, M.K., 1984. Identifying past agricultural activity through soil 240–258. phytolith analysis: a case study from the Hawaiian Islands. J. Archaeol. Sci. 11, Kirch, P.V., Hartshorn, A.S., Chadwick, O.A., Vitousek, P.M., Sherrod, D.R., Coil, J., 119–133. Holm, L., Sharp, W.D., 2004. Environment, agriculture, and settlement patterns Piperno, D.R., 2006. Phytoliths: a Comprehensive Guide for Archaeologists and in a marginal Polynesian landscape. Proc. Natl. Acad. Sci. U.S.A. 101, 9936–9941. Paleoecologists. Altamira Press, Lanham. Ladefoged, T.N., 1993. Hawaiian dryland agricultural intensification and the Pacific Rechtman, R., Maly K., Dougherty D., Clark M., Maly O., 2001. Archeological economy. Pac. Stud. 16, 119–131. Inventory Survey of the Ki’ilae Estates Development Area (TMK:3-8-5-05:19, Ladefoged, T.N., Graves, M.W., 2000. Evolutionary theory and the historical devel- 22, 26, 27), Ki’ilae and Kauleolı Ahupua’a, South Kona District, Island of Hawai’i. opment of dry-land agriculture in North Kohala, Hawaii. Am. Antiq. 65, Reichert, E.T., 1913. The Differentiation and Specificity of Starches in Relation to 423–448. Genera, Species, etc. Carnegie Institution of Washington, Washington D.C. Ladefoged, T.N., Graves, M.W., 2005. Modelling the human ecodynamics of Kohala, Rosendahl, P.H., 1994. Aboriginal Hawaiian structural remains and settlement Hawaii. In: Stevenson, C.M., Ramirez Aliaga, J.M., Morin, F.J., Barbacci, N. (Eds.), patterns in the upland agricultural zone at Lapakahi, Island of Hawaii. Hawaii. The Rena Papers: VI International Conference on Rapa Nui and the Pacific. Archaeol. 3, 14–70. Easter Island Foundation, Los Osos, pp. 155–160. Rosendahl, P., Yen, D., 1971. Fossil sweet potato remains from Hawaii. J. Polyn. Soc. Ladefoged, T.N., Graves, M.W., 2006. The formation of Hawaiian territories. In: 80, 379–385. Lilley, I. (Ed.), Archaeology of Oceania. Blackwell Press, London, pp. 259–283. Sato, H., Ikeda, W., Paeth, R., Smythe, R., Takehiro Jr., M., 1973. Soil Survey of the Ladefoged, T.N., Graves, M.W. Modeling agricultural development and demography Island of Hawaii, State of Hawaii. U.S. Department of Agriculture, Soil Conser- in Kohala, Hawaii. In: Kirch, P.V., Rallu, J.-L., (Eds.), Long-term Demographic vation Service and University of Hawaii Agricultural Experiment Station. Evolution in the Pacific Islands. University of Hawaii Press, Honolulu, in press. Government Printing Office, Washington, D.C. Ladefoged, T.N., Graves, M.W., Coil, J., 2005. The introduction of sweet potato in Selling, O.H., 1946. Studies in Hawaiian Pollen Statistics. Part I. The Spores of the Polynesia: early remains in Hawaii. J. Polynesian Soc. 114, 359–373. Hawaiian Pteridophytes. Bernice P. Bishop Museum, Honolulu. Ladefoged, T.N., Graves, M.W., Jennings, R.P., 1996. Dryland agricultural expansion Simmonds, N.W., 1959. Bananas. Longmans, London. and intensification in Kohala, Hawaii Island. Antiquity 70, 861–880. Soehren, L., Newman, T.S., 1968. Archaeology of Kealakekua Bay. Department of Ladefoged, T.N., Graves, M.W., McCoy, M.D., 2003. Archaeological evidence for Anthropology, B.P. Bishop Museum, Honolulu, and University of Hawaii-Manoa. agricultural development in Kohala, Island of Hawaii. J. Archaeol. Sci. 30, Stuiver, M., Reamer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., 923–940. McCormac, G., van der Plicht, J., Spurk, M., 1998. INTCAL98 Radiocarbon Age Leach, H.M., 2005. Gardens with weeds? Pre-European Maori gardens and inad- Calibration. Radiocarbon 40, 1041–1083. vertent introductions. New Zeal. J. Bot. 43, 271–284. Swinkels, J.J.M., 1985. Sources of starch, its chemistry and physics. In: van Lennstrom, H.A., 2001. Macrobotanical materials. In: Allen, M.S. (Ed.), Gardens Beynum, G.N.A., Roels, J.A. (Eds.), Starch Conversion Technology. Marcel Dekker of Lono: Archaeological Investigations at the Amy B. Greenwell Ethnobo- Inc., New York, pp. 15–46. tanical Garden, Kealakekua, Hawaii. Bishop Museum Press, Honolulu, pp. Therin, M., 1998. The movement of starch grains in sediments. In: Fullagar, R. (Ed.), 119–125. A Closer Look: Recent Australian Studies of Stone Tools. Vol. 6. Archaeological Lentfer, C., Green, R.C., 2004. Phytoliths and the evidence for human cultivation at Computing Laboratory, University of Sydney, Sydney, pp. 62–72. the Lapita Reber-Rakival site on Watom Island, Papua New Guinea. In: Torrence, R., Barton, H. (Eds.), 2006. Ancient Starch Research. Left Coast Press, Attenbrow, V., Fullager, R. (Eds.), A Pacific Odyssey: Archaeology and Anthro- Walnut Creek, CA. pology in the Western Pacific. Australian Museum, Sydney, pp. 75–87. Papers in Trappe, J.M., Castellano, M.A., Malajczuk, N., 1992. Australasian truffle-like fungi. II. Honour of Jim Specht. Records of the Australian Museum Supplement 29. Labyrinthomyces, Dingleya and Reddellomyces gen nov. (Ascomycotina). Aust. Lockwood, J.P., Lipman, P.W., Petersen, L.D., Warshauer, F.R., 1988. Generalized Ages Syst. Bot. 5, 597–611. of Surface Lava Flows of Mauna Loa Volcano, Hawaii. In: Miscellaneous Vitousek, P.M., Ladefoged, T.N., Kirch, P.V., Hartshorn, A.S., Graves, M.W., Investigations Series. US Department of the Interior. US Geological Survey. Map Hotchkiss, S.C., Tuljapurkar, S., Chadwick, O.A., 2004. Soils, agriculture, and I-1908. society in precontact Hawaii. Science 304, 1665–1669. Loy, T.H., Spriggs, M., Wickler, S., 1992. Direct evidence for human use of plants Wagner, W.L., Herbst, D.R., Sohmer, S.H., 1990. Manual of the Flowering Plants of 28,000 years ago: starch residues on stone artefacts from the northern Solomon Hawai’i. In: Bernice P. Bishop Museum Special Publication, vol. 83. Islands. Antiquity 66, 898–912. Ward, J.V., 2001. Pollen analysis. In: Allen, M.S. (Ed.), Gardens of Lono: Archaeo- McCoy, M., 2005. The development of the Kalaupapa Field System, Molokai Island, logical Investigations at the Amy B. Greenwell Ethnobotanical Garden, Keala- Hawaii. J. Polynesian Soc. 114, 339–358. kekua, Hawaii. Bishop Museum Press, Honolulu, pp. 127–135. McCoy, M.D., Hartshorn, A.S., 2007. Wind erosion and intensive prehistoric agri- Whistler, W.A., 1991. Polynesian plant introductions. In: Cox, P.A., Banack, S.A. (Eds.), culture: a case study from the Kalaupapa Field System, Hawaii. Geoarchaeology Islands, Plants and Polynesians: an Introduction to Polynesian Ethnobotany. 22, 511–532. Dioscorides Press, Portland, pp. 41–65.