7 A GEOARCHAEOLOGICAL ANALYSIS OF SEDIMENT SAMPLES FROM THE TO'AGA SITE EXCAVATIONS PATRIcK V. KIRCH, ELIZABETH MANNING, AND JASON TYLER INTRODUCTION A MAJOR THUSr OF OUR research at the To'aga site has been the investigation of the geomor- phological and sedimentary context of the site's extensive buried archaeological deposits. Such an approach is essential both for the underading of the "formation processes" by which the site was created (Schiffer 1987), and for the broader issues of landscape change in relation to human ecology on the Manu'a Islands. An understanding of the site's geomorphology also relates directly to cultural resource management concems for, as we shall argue, there are reasons to believe that te To'aga coastal terrace is now in a stage of active coastal erosion. While the deeply bunred archaeological deposits at To'aga are not immediately theatened by shoreline regression, they could become endangered should this process intensify, particularly through the effects of global warming and sea-level rise (Geo- physics Study Committee 1990). In chapter 4, a model of the morphodynamics of the To'aga coastal terrace was deduced from the general pattem of Holocene sea-level regimes in the southwestem Pacific, and from certain features of Samoan geological history, especially evidence of tectonic subsidence. The stratigraphic results of our transect excavations, presented in detail in chapter 5, substantially confinned this model, while the radiometric chronology presented in chapter 6 provided a temporal calibration of the stages of coastal terrace formation. In this chapter, we focus on another line of evidence for testing the morphodynamic model: that of geoarchaeological analysis of sediment samples obtained from the stratigraphic profiles of the varous transect excava- tion units. Our approach here follows long-established principles of sedimentology (Blatt et al. 1972; Folk 1974; Krumbein and Sloss 1963; Twenhofel 1950, Reineck and Singh 1980) in the context of a geoarchaeological perspective (Stein and Fan-and 1985; Shackley 1975). In particular, our aim is to augment our field stratigraphic descriptions and interpretations through laboratory analysis of sediment samples, using a "life history" approach to sediment interpretation. Any sediment, for example a sand layer in one of our transect units, can be understood in tenns of four "life history" stages. Each sediment has: (1) a source; (2) a tansport history; (3) an environment of deposition; and, (4) post-depositional alterations (Stein 1985:6-7). These stages may be infenred through the application of various analytical techniques including particle-size analysis; grain lithology and mineralogy; point- counting of grain composition; chemical and physical detennination of carbonates, organic matter, and pH; colorimetric analysis; and oher techniques. - 94 The To'aga Site Our geoarchaeological investigation of sediment samples from To'aga was conducted in two phases. In 1987, a series of archaeological and control samples was obtained to check aspects of our statigraphic interpretation. ibese samples were analyzed by Tyler at the University of Washington geoarchaeology laboratory, following procedures outlined by Stein (1985). In 1989, a much more extensive set of samples (both archaeological and contrls) was obtained from our trLansect excava- tions. These 1989 samples were analyzed at the University of Califonia, Berkeley, Archaeological Research Facility by Mmaning, under the direction of Kirch. The analytical methods applied to the 1989 samples differed slightly from those used in 1987, largely because in the latter case we were able to define more precisely the specific research questions to be addiessed by laboratory analysis. For examnple, in 1989 we recognized that time-onsuming pipette determination of the fine-fcion (< 4 phi) was unnecessary to determine the environment of deposition for these predominately coarse-grained, calcareous sediments. lherefore, we concentrated our efforts on mechanical sieving and textual analysis of the pebble- to fine-sand-sized compo- nents, with additional determination of grain lithol- ogy (basalt versus carbonates) in the -1 to 1 phi (?) size ranges. In other aspects, however, such as pH, organic matter, and carbonate determination, our 1987 and 1989 laboratory tecniques were identical. As indicated above, Tyler and Manming were responsible for the laboratory work upon which this chapter is based, while Kirch authored the text and is responsible for the final interpretations and data presentation. METHODS Field Sampling Sediment samples were systematically taken from all units following th completion of excava- tion, in conjunction with the drawing and description of stratigraphic profiles. In both 1987 and 1989, the exact positions of sampling blocks were recorded on the satigraphic profiles, and these have been indicated on the various section drawings repro- duced in chapter 5. Sample blocks, which usually measured 10 x 15 x 10 cm, were cut into the cleaned section walls with a trwel. Samples were placed directly into heavy plastic bags and sealed for shipment back to the laboratory. As discussed by Stein (1985:7-9, fig. 1), sampling strategies can vary depending upon the questions to be asked in labora- tory analysis. Since we were interested pimarily in depositional events, our samples were taken from within individual strata, avoiding boundares or contacts between layers. Where large clasdcs were present (e.g., > -2 0) which could not be adequately sampled, these were noted in stratigraphic descrip- tions. Subsampling in the laboratory, for various analytic tests, was caried out with the use of a Jones sample splitter. A total of 32 sediment samples (including 3 controls) was obtained and analyzed in the 1987 season. Another 70 samples (including 5 controls) were taken and analyzed from the 1989 excavations. Of the 1989 samples, those from Units 15, 19, and 23 were accorded a full analysis, while samples from other units were analyzed only for color, pH, parficle size, and lithologic composition. Controls Several control samples were obtained in the To'aga area from modem depositional envimrnments, in order to assist in the interpretation of the archaeo- logical sediments. These included samples of active colluvial material from the slope inland of the 1987 excavations, sand from contemporary beach ridges, sand from te beach in front of Transect 7 (fig. 7.1), and a high-energy, sand-and-gravel beach at Fa'ala'aga. Details of tese control samrles are provided in tables 7.1 and 7.2. Particle-Size Analysis A primary aim of laboratory analysis was to interpret the environment of deposition of sediments. This requires a knowledge of the particle-size distribution, since the size ranges and degree of sorting of a sediment reflect the energy levels in that envirnment. For example, low-energy beaches will be characterized by well-sorted (stmngly unimodal) sands in the medium- to very fine-grained size range. Higher energy beaches, on the other hand, display less sorting and a higher frequency of coarse-grained to granule-pebble-sized clastics. Similatly, colluvial sediments deposited by mass wasting can be ex- pected to be very poorly sorted, exhibiting a ful size Geoarchaeological Analysis of Sediment Samples 95 Photomicrograph (lOX) of the 1 phi size fraction of a modem control sample of calcareous beach sand from Transect 7. range from silts and clays up through very large clastics (cobbles to boulders). Unconsolidated sands generally required no pretreatment prior to parficle size analysis. The colluvial samples, however, required pretreatment for the removal of organics, and in some cases, for iron oxides which would otherwise bind individual grains together (Kunze 1965). Organic materials were removed by pretreatment with hydrogen peroxide, H202 (Jackson 1969). A 1:1 solution of sediment and distilled water was prepared, to which H202 was gradually added until frothing and foaming subsided. The beaker was then placed in a water bath at 80 OC, with additional H202 added Table 7.1 Analytic Data for To'aga Control Sediment Samples (1989) Sample % % % Textural % % No. Location Gravel Sand Silt Class pH Calc Basalt 89-131 Fa'ala'aga Beach 2.3 97.7 0 S 9.01 97 3 89-132 Fa'ala'aga Beach 43.0 56.9 <0.1 sG 8.68 90 10 89-133 Fa'ala'aga Beach 97.4 2.5 <0.1 G 8.95 89 11 89-134 T-7, A horizon, 1.7 97.0 1.3 S 8.28 100 0 beach ridge 89-135 T-7, beach sand 0.2 99.8 0 S 8.64 100 0 Figure 7.1 96 The To'aga Site C! - N., Q m m C4 C4 w 04 >4 >, 0>, tn 4n - - ;soo seno - - ^ k _;~ W 4 C ~c o o\ o- in 0x s en r- W) *; *r *6 t- _- ^ t >, t t_ _ lq N o tn sfu00x N r4 o _ t- 1. od o40' e 'r_oo en V- C14 - -o 04 O: 04 00 0 --- '-qo*o4 0 0 44 O ) O>4 O ~o t_ _ < 0%O-t :t t: en CtO en 4)0 W W: Rr)tfL0 * :t -. -4 "t X oo W w W >0 >r O; 1- - In a a (4 >4 > C4 O WO !qQ t \0 t (N cI (NI en - W) C4 E 0 0> > 004 r-4 - - - () er *4 I. 1: s' >.0 r) r) C1 ~o C14"I "-4 - N-4 "t 'tt enON c t 00 t 00 x- t- qt cr m 4 N \0 r- ON t ONi ON t C -4 ~o 47 C C1 CS _o-x ON oN vx r- u N N m a ~o e T--4 m o m r x ur ON N ) 00 n C1 ot W) en r W) xh on xn cr 00 _ 00 (Vs e 1 - W- 00 06~ 60 06 6t:0 60 6~ 6 O0 0 0 0 t 0 0 0 r:6 t~- 0 0r 00 00 00 (N I o,- Q _ot Q w00 V-4 R W _ N _t r QN eon - Wo e %v t- x _ tt oo V W tn en e o " > o _ - C) r- tn V- Os 00 _ o_ om en xy4 e 9 *O o 0*--> -oa ro t _oo - W~~~~~~~~1 0 o o o- m _ 8 _ O _ ~o _ _ tn CZ ON r-4 ON o7Nr o ~o Wo ON 00N 0o en N t o- 0o0 \ t Oo% t e o 'r x _ ox \O00'iO~~~~~~~~l r0 f-400en _ & - 0_ 0 0_ e 00 e I 0-- O\O\N ?o 8 ON N r- 06 li t--: r- W) 00 M U 4 0-4 011, . 0. ?o oN -, 4-0 4..# ."" .1.4 in 4-0 ?:) ::) .9.4 vo t- t cs oo O O N 11 N O oO O- Co X~ O 00 (In en t N O t O t oO N 0 oO ON O -- t O _ _ O X ON st0 O \ cO F cr o e 0 xO\O x 0t cr m u -- .W .9- 4 . = = , CS c > '-4 -4- - - C _ _ = S: S:4 .4 ._ _= _ o o o_ _ a _ 3>3V4 - V-4 1- & - m -- - > t v4 * .- _ - _ * * *s:. *.= 4- -P *_ "4 .4 0 c.) (A * - 0 z 0 SW a U "a N _- 00 - . Q - 00 0 0 CN 00 vx 0 IS as i g C,' Cu Cu V- c;O t', . v :s co E-4 6 0 ~u O 3f : g 0 .o ?4 . rAZI 00 u co g ci c) i i U C.) D t fi en en Q 'N 1-- 1-.. WI) t- W) en 04 04 N 04 t t 01 1014 V-4 7-4 >4 lw i Geoarchaelogical Analysis of Sedime SWples 97 untl all reaction had ceased. The sanple was then cooled, rinsed witfi distilled water, and centnfuged at 1700 rpm for fifteen minutes. Superatant liquid was thn decaned, and the ctifuging procedure repeated until the soludon was dlear. The metd for removal of inn oxides was as follows: sediment which had lready been pre- trat for removal of organics was placed in a 250 ml centrifuge tube to which about 200 ml of citrte buffer soludon (sodium citrate dihydrate; sodium bicarbonate NaHCO3; sodium chloride NaCI) was added. Samples were then warmed to 80 OC in a water bath within the fume hood. Four grams of sodium dithionite (Na2S204) were slowly added while stiring coantdy for one minute. Following a fifteen-minute digestion period, 10 ml of sated sodium chlonde (NaCI) was added to flocculate te sample, which was then centrifuged at 1700 rpm for fifteen minutes; the supematant liquid was decanted and discarded. The procedure was repeated until samples appeared light gey or white in color, indica- ting the mnoval of mostor all of fe free imn oxides. Pretreated samples were mechanically shaken through a nested set of geological reens of mesh sies -2 to 4 # (plus pan) for fifteen minues. Weights of each subsample were deteimined with a high-precision digital electic balnce. The 1987 samples were futher analyzed by the pipette metod to determine the particle size distribution of the fine frtion; this was not done for the 1989 samples. Following mechanical sieving and weighing, the percentages and cumulative percentages of all* classes were calculated, as was the sand:mud atio (mud is defined as all material > 4 *, i.e., silt plus clay). Tese stastics permit the textual classifica- tion of the sediments according to the system out- lined by Folk (1974:27-30, table 1). This system utlizes a triangular diagram with apices of gravel, sand, and mud, reproduced here as figure 72. The dominant mode is indicated by a capital letter, supplemented by lower case letters for secondary modes. For exanple, a sandy gravel is designated sG, while a slightly gravelly muddy sand would be (g)mS. Figure 7.2 Terminology adopted for the To'aga site sediments, in terms of the relative contribution of gravel, sand, and mud size frac- tions, expresd asa trary plot (Souce: G-dinerand Dackombe 1983:108). 0 Gravel (>2 n11im) 8(1 MUDl)Y a a Z\ tGRAVEL | a (mG) a< 4. GRAVELLY MUD (gM) sIGhTLY SuIIY GRAVLLY G;RAVELLY MUD SANDY Mtl) ((Wg)M)h - / UMUI(M)/ SANDY MUD (sM) Mlud (<0.06 mm) I :9 0 a I GRAVELLY \7 MUDDY \ t f SAND (ginS) SLIGIITL Y GRAVELLY MUI)DY SANI) l(igmS) MUDDY SAND (mS)\ SAND (S\ 1: 1 Sand: mud ratio SLIGHTLY GRAVELLY SAND ((g) S) 9:1 Sand (0.06-2mm) ... i A xx L Irl- v - v - -- Hf IC Iv I A IL Point-Counting (Lithology and Micro-artifacts) The pH (acidity, neutrality, alkalinity) of samples was detennined witfi a Mettler automatic pH meter. A 20 g sample was p d in a 1:1 solution with distfilled water, and standards set to pH 7 and 10. Tlree trials were made for each sample, with the repoited value being the average of these readings The pH values from the To'aga sediments are of intres prmarily for the assessmentof preservadon of organic culural materials, such as bone and shell faunal remains. pH values for the upper colluvial deposits, or of contemporary A soil horizons with humic materials, tended to range between 6.4-8. Deeper sediments, generally calcareous, had more alkalin pH values ranging frxm 8 to 9.5. This alkalinity strngly favors the preservation of bone and shell, reflected in the generally excellent condi- tion of tX faunal assemblage. In contast, the relative neutrality of the shallower deposits probably reflects th presence of humic acids. This is evident in te "chalky" nate of shell midden in these upper strata. Organic Matter and Carbonates Ihe p=sence of organic matter and carbonates in sanples was determined by the 'loss-on-ignition" method (Dean 1974; Stein 1984). This is based on the principle that organic materials will begin to ignite at 200 OC and will bum completely when the temperue raches 550 OC. Calcium carbonate (CaCO3) evolves to carbon dioxide gas when heated to 800 OC and is eliminated at 850 OC. Thus te loss-on-ignition procedure involves the controlled burning of samples with precision weighing before ignitions after 500 ?C, and after 1000 OC bums. All samples were processed in a Thermolyne muffle fumace. Color Sediment colors were recorded in situ during the description of profiles, using the Munsell Soil Color charts, and have been reported in chapter 5. In addition, the colors of laboratory sediment samples were wecorded both dry and moist, also using the Munsell system. During the field description of sedimentary units, we observed that the more deeply buiied and older staa tended to have sands with a mixed calcareous-volcanic lithology, designated 'salt-and- pepper' sands (fig. 7.3), whereas higher and more rcent sands were largely or wholly calcareous. These differences were believed to reflect the process of coastal progradation and termce forma- tion, which gradually removed sources of volcanic sediment from the To'aga coastal sediment budget. These sources originally would have been volcanic headlands and talus boulden which when exposed to high-energy wave action would generate volcanic- lithic sand grains. In order to quantify more pre- cisely these field observations, we undertook controlled point-counting of samnples in the labora- tory, following procedures first outined by Galehouse (1971). The -1, 0, and 1 + size classes were selected for point-counting following mechani- cal sieving. Subsamples of these + classes were mounted on petrographic glass slides, which could then be examined under a Nikon stereozoom microscope. An average of 100 grains was counted for each sample, with grains being classified as calcareous (consisting of coral sand, foraminifera, marine shell fragments, etc.) or volcanic lithic (basalt or similar materials). These counts were then calculated as percentages. For Units 15, 19, and 23, we also used the point- counting method to determine the frequencies of the following classes of micro-artifactual or faunal materials: charcoal, marine shell, bone, sea urchin, and terrestial gastrpods. (In order for marine shell to be counted as a micro-artifact constituent, it had to have sharp or fractured edges, as opposed to water- worn or rolled edges, the latter indicating a natural constituent of the sand matrix.) The terrestrial gastro- pods are of particular interest as they consist prima- rily of synanthrpic species; these are the subject of a separate and detailed analysis in chapter 8. RESULTS The 1987 Excavations Thirty-two sediment samples were obtained from the 1987 excavations (including fhree control 98 The To'aga Site pH Geoarchaeological Analysis of Sediment Samples 99 Photomicrograph of a typical "salt-and-pepper" sand with mixed calcareous and volcanic-lithic lithology; from Layer IF of Unit 24 (1 phi size fraction; lOX). samples). Samples were taken from Units 2 and 3 along the initial transect, from Units 5, 6, and 9 in the main trench, and from Unit 10. Analytical data from all of these samples are reported in table 7.2. The sedimentary sequence in the main excava- tion trench is summanzed by samples 13-23 listed in table 7.2. The colluvial units (Layers I and III) are similar in their higher frequencies of silt and clay, contrasting markedly with Layers II and IV which are dominated by sand-sized particles. Similar contrasts are expressed in the percent of calcium carbonates in these strata. The 1989 Transects Transect S Units 15, 16, and 17 were selected for sediment analysis along Transect 5, with a complete set of analyses for Unit 15, and particle-size and lithological analyses only for Units 16 and 17. Complete analyfical data for these units are pre- sented in tables 7.3 to 7.5. The sedimentary sequence in Unit 15 is graphi- cally depicted in figure 7.4, and a more detailed graphic presentation of the grain-size distribution results is provided in figure 7.5. In figure 7.4 several overall trends are apparent. The lower sediments (Layers Ill to IV) are dominated by sands, with significant gravel components confined to the upper deposits (Layers I and II). This change reflects the shift from a littoral depositional environment to a mass wasting (colluvial) depositional environment. The pH shifts from neutral to somewhat alkaline between Layers IA and IIIA and remains alkaline to the base of the section. Organic matter is greatest in Layers IA and II. The shift from littoral to ternig- enous mass wasting depositional environments is also clearly indicated in the percent of CaCO3 and in the lithological composition of basalt and calcareous grains determined by point-counting of the 1 0 size Figure 7.3 100 The To'aga Site Q zt C4 >- >4 >- "-4 RQI OO 0 Qo sri NN0 C 00 NO00 0'r)0 _- t._. _. - . o- 00 N- 00 0; 00I o 0- oo R 0 0 o t r - w ON C C O C bON ON * Nt Y t r- '%o O- _N -O -4 R CR t ON ON 0000 ON c, w) - N w) r- c7 o XN X0 00000 000 qtC ) ,t "TC C) r- ooCK "-4 d ~0.t-ONt--00 ONt00 0 c- OO (V Wm cf \0 r- < t), _. C4) w c\- C\ C \b C - m m - - -- - 00 ON0 0 00 0000 ON ON ON 000000 ,it )RI~R 00 0000 E 0t I- 0 4- *C c; ;a Cu U -H to z 2 V" .5 qA I' 4-b i cl m u 014 u i. Iw PC 'w Ev c o CX Z cir ad >0 e r- (1 - w od 04 C ~ -C.I Q -1 0 )00 r --- 00 0 0 00 0 ' 00 00 IGIN C O CIA ON -4 e4 V- tn V-- 0 0* ONO NO 00 ON N 4 \0 ? o CN GII r-G Ce a en 0 0 cI O NN V N < m c e _ _ > I -4 0-- 0- -. N-00OGN 0000 00 0 00 00 _-4 00 C0 oo 0 .5 U I 0 S tv Cu 2 0 Cu Cu co *s: Cn "4H e A e "a riA &.. 0.4 I Geoarchaeological Analysis of Sdet Se s 101 Table 7.5 Analytic Data for Unit 17 Sediment Samples 1 + Munsefl Color Sample 9 % % Textural % % No. Layer Gravel Sand Silt Class pH Calc Basalt Dry Moist 89-66 I 49.1 47.8 3.1 sG 7.84 99 1 7.5YR 3/2 5YR 2.5/2 89-67 II 80.7 17.4 1.8 G 8.42 100 0 7.5YR 5/2 7.5YR 5/2 89-68 IIIA 83.5 14.6 1.9 G 8.51 100 0 10YR 5/3 7.5YR 5/2 89-69 IIIB 83.4 15.1 1.5 G 8.47 99 1 7.5YR 5/2 7.5YR 4/2 89-70 IIIC 63.0 35.4 1.6 sG 8.46 100 . 0 10YR 5/3 10YR 5/3 89-71 IVA 1.9 95.9 2.1 S 8.48 99 1 7.5YR 5/8 7.5YR 5/6 89-72 IVB 1.4 94.9 3.6 S 8.57 76 24 7.5YR 6/2 7.5YR 5/2 89-73 V 2.3 97.5 0.2 S 8.54 100 0 10YR 8/3 10YR 8/3 89-74 VI 11.5 88.4 0.04 gS 8.90 100 0 7.5YR 8/4 7.5YR 8/4 89-75 VII 0.4 99.5 0.05 S 9.50 93 7 7.5YR 8/2 7.5YR 8/2 GRAIN SIZE MA-1 pH 8 9 ORGANIC MATTER CARBONATES LITHOLOGY (1+) o 5 10% 0 i I _- 100% TRANSECT 5, UNIT 15 Figure 7.4 Summary diag of grain size, pH, organic matte, carbonates, apd lithology for sdiment samples from Unit 15, Transect 5. I 102 The To'aga Site 2 IA 30 50 50 0 -~~~~~M- 0 III 0 70 80 0~~~~~m 50 70 t X L ~~~mc 50 0 -2 -1 0 I 2 3 4 4 'P' GRAN SIZE t Hstogram plots of grain size distribu dons for sedinent samples from Unit 15, Transect 5. class. PhotomicrograPhs of the 0 + size frcions for sediment samples from Unit 15 are siown in figure 7.6; tese illusae the vadations in lithology between the various cultl and natrl deposits. The detailed grain-size distnbutions shown in figue 7.5 likewise document changes in the environ- ment of deposition. The basal sand, Layer IV, consists of rather poorly sorted sands (dominant mode = coarse sand) with a small granule-sized component, suggestive of a relatively high-enrgy beach environment In Layen HID to lIIA-1 there is a steady shift to well sorted sands dominated by particles in the medium sand range (2 0) with granule-sized particles generally lacking. This shift is presumably correlated with the progradation of the active beach seaward of the Unit 15 locus. The Layer II midden deposit displays a very poorly sorted sediment, reflecting the incorporation of cultual materials (including oven stones and other large clastics) into a medium-sized sand matrix. Layers lB and IA are also poorly sorted, with dominant modes in he pebble-to-cobble size ranges (-2 to -I O). These upper sediments are typical of young colluvium denived from mass wasting of the cliffs and volcanic slopes immediately inland of te site. In sum, the Unit 15 sedimentary sequence reveals the following geomorphological evolution. The intmdal enviromnent of deposidon (at ca. 3,000 Cal B.P., prior to human colonizadon) was a high- energy beach, indicating an active shoreline very close to Unit 15. The source of sediment was primarily the calcareous reef flat, with te minor addition of volcanic lithic grains derived from exposed headlands and/or talus rockfall exposed to wave actdon. During the period of human occupa- tion represented by Layers HID to IIIA-l (ca. 3000- 1600 cal B.P.), the Unit 15 locality shifted to a beach ridge depositional enviromnent, with trasport of calcareous sediment primarily by aeolian processes (e.g., saltation of grains inland). Following the Layer H midden occupation on top of a vegetated beach ridge, the Unit 15 locality began to be covered by terrigenous, poorly sorted, volcanic sediments deriving from the steep talus and colluvial slopes inland. The encraclhnent of xse volcanic sedi- ments probably reflects increasing slope instability due to forest clearance and agricultual modifica- tions. Figure 7.5 Geoarchaeological Analysis of Sediment Samples 103 Figure 7.6 Photomicrographs (lOX) of the 0 phi size fractions of sediment samples trom Unit 15, Transect 5: A, Layer IB; B, Layer II; C, Layer IIIA; D, Layer IIA-1; E, Layer IIIB; F, Layer IIIB; G, Layer IIID; and H, Layer IV. (e.s. = echinoid spine; Lb. = fishbone; s.g. = synanthropic gastropod). 104 The To'aga Site Transect 7 Sediment analyses were perfonned on samples from Units 18 and 19 along Transect 7. Tables 7.6 and 7.7 summanze the detailed analytical results. A graphical summaiy of the Unit 19 sedimen- tary sequence is shown in figure 7.7, and the detailed grain-size distnbutions are depicted in figure 7.8. The overall sequence minors ta described above for Unit 15, with such features as the shift from neutral to alkaline pH, the rapid decease in organic matter, and the relative contributions of basalt and carbonate sediment grains. Layers VI to II are all reladvely well-sorted sands dominated by a medium- sand (2 0) mode, with the exception of Layer V, in which some larger (-1 and -2 0) clastics represent the incorporation of cultural materials into the sand matrix. The Layer IA colluvium is typically poorly sorted. Transect 9 Units 21, 22, and 23 were selected for sediment analysis along Transect 9. Tables 7.8 to 7.10 present Table 7.6 Analytic Data for Unit 18 Sediment Samples 1 + Munseil Color Sample % % % Textural % % No. Layer Gravel Sand Silt Class pH Calc Basalt Dry Moist 89-86 IA 43.8 50.1 6.1 msG 7.42 37 63 5YR 3/3 5YR 3/3 89-87 IB 27.3 64.9 7.8 gmS 8.46 95 5 IOYR 3/4 10YR 3/4 89-88 II 1.1 95.6 3.3 (m)S 8.63 99 1 IOYR 5/4 5YR 4/6 GRAIN SIZE 0 50 100% IATG IB m IA IVB V VI No Data Sand I pH 7 9 ORGANIC MATTER CARBONATES 0 50 100% I LITHOLOGY (1+) o so 100% a 4 i1~ Calcareous TRANSECT 7, UNIT 19 Summary diagram of grain size, pH, organic matter, carbonates, and lithology for sediment samples from Unit 19, Transect 7. Figure 7.7 4 0 4 0 4 0 4 p I 4 Geoarchaeological Analysis of Sediment Swpks 105 00 t00 00 lx IY od o 00tt00 " - -4 -- Q oo : r00 >00I 0 1- V- ( i x00 000 ) 00 eir-: -4 -4 --_ ', 00 000 0 Ck (2 N It C4 t- N 6 t r; - - t I'- - - cr4 C-- CN _ > cr, vo m m tl 1 t r-:00 00 00 00 00 00 00 to trxocraxo c' 6 6 6 6 6 c c en Cq4-4 - - as 0| ON as >e q |t r-4 -4 -t 0 6 0 6- S S< > S1 S . m > > > _0 04 ON o _- N en WI m o x ON ON ON ON O ON oE oao b oo o o co F* 0 ce - C "a *so %,wI a u A a 2 0 'a "A 40 Ia I I U #1 &o 0 04 a Oa Qr o tn[- Q 4> I Q r- Os en d 4- r-00 N% o t) m 00 0000 0X 0% 00 00 CO) C CT C en in 0 0 " _s 0oe N 1-- P0% -m m - --~ 0 0 o00 0 00 00 00 2 m CU 0 I- Cu aN *a: co t.o 0 "a u gO a ~u 13 t 0 Eu 106 The To'aga Site the detailed analytical results. A graphic swumary of the Unit 23 deposidonal sequence is shown in figure 7.9, and the detailed grain-size distributions are illuste in figure 7.1 1. otomicrographs of the 0 0 size fraction for sediment samples from Unit 23 are shown in figure 7.10. As seen in figure 7.9, the overall pattem of .2 E2 U .~ 30 so 0 60 70 I IVA 0 60 . l ~~~~~IVB 0- 0 soV o- ~~~~~~VI -2 ' -1 1 ' 2 ' 3 ' 4 ' PN' GRAIN SIZE (#) Figure 7.8 Hisgram plots of grain size distribu- tions for sediment samples from Unit 19, Transect 7. change is much like that described above for Transects 5 and 7. Of particular note is the prsence of basalt grains in the basal sands (Layers IV and IIIC), reflecting the 'salt-and-pepper' lithology observed in the field. The detailed grn-sze histograms (figure 7.1 1) indicate that Layer IV is a coarse-to-medium-grained sand with a 'tail' extend- ing into the larger clastic size ranges, suggestive of a higher-energy beach depositional environment In Layers MIB and IIA, the larger dastics are cultual materials (such as oven stones, marine shells, and large sea urhin spines) which were added to a coarse-to-medium-grained sand matrix. Layer HA, intepreted in the field as a paleosol horizon, is less well sorted than the underlying sands and begins to incorporate fine-grained terigenous sediments. The Layers IC to IA colluvial deposits are all very poody sorted, with ineasing quantities of terrgenous (basalt) materals. Figue 7.12 presents a graphical summary of grain-size distribution and lithology for Units 21, 22, and 23. These results clearly reflect the temporal shift from (1) high-energy beach depositional environments in the basal levels of Units 23 and 21 to (2) beach ridge depositional environments with human occupations, to (3) the encroachment of terigenous colluvial deposits in the vicinity of Unit 23. A micro-artifact constituent analysis of sediment samples from Unit 23 is presented in table 7.1 1, with frequencies for five categores of culural material indicated by grain size (-1 to 10 size classes). Charcoal was represented only in the upper colluvial deposit Layer IB, presumably reflecting human- induced buming up-slope, associated with agricul- tral activities. Other cultural materials, mainly sea urchin and bone, were confined to Layers IIA to IUB. Transect 17 Sediment analyses were performed for samples from Units 24 and 25 along Transect 17. Analytical results are summarzed in tables 7.12 and 7.13. In figure 7.13, the results of grain-size distribu- tion and lithology are presented for both Units 24 and 25. A shift from high to lower energy deposi- tional environments is indicated by the grain-size distribution, while a decreasing contribution of volcanic sand grains is reflected in the 10 lithology. Geoarchaeological Analysis of Sediment Sampks 107 ORGANIC MATTER 5 10 15 CARBONATES LITHOLOGY (10) TRANSECT 9, UNIT 23 Figure 7.9 Summary diapm of grain size, pH, organic matter, carbonates, and lithology for sediment sanples from Unit 23, Transect 9. Table 7.9 Analytic Data for Unit 22 Sediment Samples 1 ? Sample % % % Textural % % No. Layer Gravel Sand Silt Class pH Calc Basalt 89-111 I 15.5 79.8 4.7 gS 8.36 100 0 89-112 II 31.5 67.0 1.5 sG 8.64 100 0 89-113 III 0.1 99.6 0.2 S 8.84 100 0 This indices the effects of coastal progradation, and the gradual eliminadon of volcanic boulders and headlands as a source of sand sediment. The detailed grain-size distribution for Unit 24 clearly reflects the changing depositional environ- ments at Transect 17 (figure 7.14). Layers IF hough IC are indicative of very high-energy beaches with substantial components in the granule- pebble-cobble size modes (O to -2 O). Layer ID, in particular, probably represents a major stonn event, possibly associated with a tropical cyclone. Layer IB reflects a rapid shift to a lower energy beach, dominated by coarse-to-medium-grained sands, while the upper Layer IA reflects the development of a stable beach rdge depositional environment There was no significant encroachment of colluvium over Transect 17. It is clear that for most of its late Holocene GRAIN SIZE pH 108 The To'aga Site *9 . . . 94 * .- > t F .- t_ - ln t tn W WI WI WI 0%sr4n te m 0 '8It C J C14 %o 0o00 0^ 0 %O0 N~ E- 00 00 00 en m 00 00 0 00 00 V) o) n _4V 4C o 000000Oa 0a " x x o *0> o - -D""0 N -ItV e rt 0m o %t 0 C e cl 4fiw 14 6 V-4 t V- t- 4 v- en 6< m C; 6 m < m u > 0 %4 0 %4 0 4 % " 4 "04 " 4 " 4 " 4 "-4 oa a a a a a a a ao 000000000 000000000 * sn" 5 C,, 2 co ri ;~a e4 * a a) 0 3- U *= I' *;. hia ;, a ha "4^ 0+ Q ha 0 'U I a a - I a a I I I I I a I Ia I I I I I I a a a - * a I a I a a cl | s. ~oc V- I a I I 0 6 "-4 0 6 " 4 I I I I a I I I I I I I I a I I I a I I I I a (NI I I a I I I I I a 6< m c; ;5 f<< m u >, a- 4 a- 4 a a a a a 0- 0-~-~ 0x 00 00 00 0% C0 x0 8 oo "-4 0 "-4 00 0 o4 00 0 "-4 00 a an 0 x4 00 0 Iid 0 r5o a U) la q *. i U) as c; *sA H e | "4 I DI as 00 lid 8 *E oo 'I 8 -N * I 1 C'4 Cfi I I I I I I I . IL - N I I I I I I -- I I I I I I en i I I I I . Geoarchaeological Analysis of Sediment Samples 109 i i i I i i I Figure 7.10 Photomicrographs (lOX) of the 0 phi size fractions of sediment samples from Unit 23, Transect 9: A, Layer IB; B, Layer IC; C, Layer IIA; D, Layer IIB; E, Layer IIIA; F, Layer IIIB; G, Layer IIIC; and H, Layer IV. I 1.. I. I I. I 110 The To'aga Site It 0 Q en C4 0 9114 - -- r-0000 00 V--4 -4 -4 I- cn CZ - *N t \0 0-00 00 C00 00 (jA 00 00 lqt'-(14 IOCN00 QON \ -C1 00 00 C1400 \OC1 000C 'r) \or' eNcNw 0000 00 00 CO o 00 00 00 U) 112 ;._ C * cnCA t- C 0 *& co *a co &W 0 ~u 4) Ig .U de u i2 co N r PC rj, 'U 'Ut 04) Nx Q '.4 1-- m W) Cd C'i >-, Cd W) >4 t- W) '-4 c o o '0e en _B _t >~~~~~~~~~~~~~~~~~ en en o~~ ~~~~~~~~~~~~~~~~~= , I- I---- * - 00 it C4r:400 en 0 00 0 od C1 Q C) b- C> so - - r o o 0000000en00000 CF CD O N Ch N IT m ? r- 0 od 0x 0x 0x 0o 0o s N x00 N N b0t 0 0CC 00 o00 CI40WIr) - 0 - - bi : ^ .< m U ;~ a e O.4 -40- -4e - oo~ 00 N X0 00 V-- - c 0 N 00 00 X0 00 C0 00 Li 0 0 2 I" 0 'U v ~a N Cu v exi Ono Cu co eu CX *s: V1) N) Ul Li 1-4 '0 N U) Q 'U 4) NW Geoarchaeological Analysis of Sedcm Sanples 111 40 '_~ ' I I~~I 40 30 _ IA 0 I0 IC o- I . HIA 0 40 lA 0 m 50 _r_ 0 50 50 ~~~~L o~~~~ rm ~~~IV 0 history, Transect 17 has been an area of high-energy beach deposition, which conelates with the lack of significant cal deposits. The stable coastal ten-ace probably did not extend into this Fa'alapaga area unil the last 1-1.5 kyr B.P. SUMMARY AND CONCLUSIONS Th results of detailed sdiment analyses frm the To'aga excavations provide a set of independen data with which to t th morphodynamic model of coasal change developed in chapter 4. These rsts ar highly consistent between tasects and confirm that dtere has been a significant shift in source, mode, and environment of deposition at To'aga which conmlates with shoreine prg ion and formation of the coasal terrace. Pidor to, and at the time of initial human colonization of Ofu, the shoreline along the southem pail of the island was much closer to the volcanic cliffs, in the vicinity of the present surficial contact between the talus slope and coastal terrace. These basal deposits are consis- tently reflective of high-eergy beacrs, and the admixture of basalt with calcareous grains indicates ta volcanic headlands along with the coral reef provided sources of sediment A shift to low-energy beach ridge depositional environments by 1900 cal B.P. reflects coastal progradation, as predicted by our morphodynamic model. Subsequently, inrased deposition of young, pootly sorted, temgenous colluvial sedimets onto the calcareous coastal terrace suggests inreased human disuibance (burning and agricultural activity) on the interor volcanic slopes after about 1900 cal B.P. Further implications of ftese results are explored in grater detail in chapter 15. -2 '-I 'PA OIs 2 3' GRAIN SIZE (4) Figure 7.11 Hisg plots of grain size distribu- tions for sediment samples from Unit 23, Transect 19. . 112 The To'aga Site UNIT 22 I MLI GRAIN SIZE (%) IA IB IIA fB 'IC 0 I 50 100 Calcareous - SEAWARD 0 50 100 I II Calcareous m "Salt and Pepper" Sands "Salt and Pepper" Sands GRAIN LITHOLOGY I (%) Figwe 7.12 Summary diagrams of grain size and lithology for Units 21, 22, and 23, Transect 9. UNIT 23 UNIT 21 0 100 100 IA IB HA fIB RC LA IB IC I1A LB mA mc wV INLAND IA IB IC IIA fiB MfA MB mc IV 46 GeoarchaeologicalAnalysis of Sediment Swmp1es 113 UNIT 24 GRAIN SIZE Lll 0 THOLOGY (1) 50 100% IA IB IC HIA LIf) IIB UNIT 25 GAIN SIZE LITHOLOGY (14) 0 50 100% 0 50 100% INLAND -' * SEAWARD Figure 7.13 Summary diagrams of grain size and lithology for Units 24 and 25, Transect 17. IA IB IC ID is EF 114 The To'aga Site Jill~' siH t~~ Q '5| E 3 A 40 I 70 - O- ~ ~~ r 40 IC 40 50 so L: IE 0 so IF 0 -2 -1-I F 0 1 ' 2 GRAIN SIZE (#) 3 4N Figure 7.14 Hisgraam plots of grain size distribu- tions for sediment samples from Unit 24, Transect 17. REFERENCES CITED Blatt, H., G. Middleton, and R Murray 1972. Origin of Sedimentary Rocks. New Jersey: Prentice Hall. Dackombe, R V., and V. Garliner,1983. Geomor- phological Field Manual. London: Allen and Unwin. Dean, W. E., Jr. 1974. Detennination of carbonate and organic matter in calcareous sedimentary rocks by loss-on-ignition: Comparison with other methods. Journal of Sedimentary Petrol- ogy 44:242-48. Folk, R. L. 1974. Petrology of Sedimentary Rocks. Austin: Hemnphll Publishing Co. Galehouse, J. S. 1971. Point counting. IN R. Carver, ed., Procedures in Sedimentary Petrology, pp. 385-407. New York: Wiley and Sons. Geophysics Study Committee 1990. Sea Level Change. Studies in Geophysics. Washington D.C.: National Academy Press. Jadcson, M. L. 1969. Soil Chemical Analysis: An Advanced Course. 2nd ed., Department of Soil Science, University of Wisconsin, Madison. Krumbein, W. C., and L. L. Sloss 1963. Stratgraphy and Sedirnentation. San Frandsco: Freeman. Kunze, G. W. 1965. Pretreatment for mineralogical analysis. IN C. A. Black, et al., eds., Methods of SoilAnalysis, pp. 569-77. Madison: American Society of Agronomists. Reineck, H. E., and I. B. Singh 1980. Depositional Sedimentary Environments. New Yodk: Springer-Verlag. Schiffer, M. B. 1987. Formation Processes of the Archaeological Record. Albuquerque: Univer- sity of New Mexico Press. Shackley, M. L. 1975. ArchaeologicalSediments. New York: Wlley and Sons. Stein, J. K. 1984. Organic matter and carbonates in archaeological sites. Journal of Field Archaeol- ogy 11:239-46. 1985. Interpreting sediments in cultural settings. IN J. K. Stein and W. R. Farrand, eds., Archaeo- logical Sediments in Context, pp. 5-20. Orono: Center for the Study of Early Man, University of Maine. Stein, J. K., and W. R. Farrand 1985. Archaeological Sediments in Context. Orono: Center for the Study of Eai?y Man, University of Maine. Twenhofel, W. H. 1950. Principles of Sedimenta- tion. New York: McGraw- Hil.