n ne rocks

Quaternary
isputable absolute terraces, jence for istropod) iisms. In ine arthro-
Figure 2. Partial view of a huge sambaqui, located beside a sand dune, near the present strandline. It is composed almost exclusively of Anomalocardia brasiliana shells (Jaguaruna, State of Santa Catarina). For scale, compare with the vehicle used during the field surveys.
PaleosAor-elitiesarid Prchistoiy: An Imwfigatior1 of Merlrod

tillstand-!

,Mean sea-leve!
to she (Fairb with 1 sandy highe: imme safety positil well i sea le
Figure 3. Schematic relative sea level fluctuation curve during the last 7000 years for the northern area of Salvador (State of Bahia). From Martin, L., Flexor, J.-M., Vilas-Boas, G.S., Bittencourt, A.C.S.P., and Guimares, M.M.M. (1979a). In: Proc. 1978 Int. Symp. Coastal Evolution in the Quaternary, pp. 264-274. Suguio. K., Fairchild, T.R., Martin, L., and Flexor, J.-M., Eds., S0 Paulo. Used by permission.
of a : establ positi above interp possii were elimii assun paleo site's
present position (about 7000 to 6500 C-14 years ago; Santos transgression), the low-lying parts of the ancient terraces were invaded by the sea, forming extensive lagoonal systems. Relative sea level fluctuation curves for the last 7000 years have been constructed for several sectors of the Brazilian coast, which indicate that:
* The present-day mean sea level was surpassed for the first time about 6500 to 7000

* By about

5100 B.P., the sea level had risen to about 4 to 5 m above today's mean sea level. At about 3800 B.P., the sea level experienced a low stand and was positioned slightly below today's mean sea level. At 3600 B.P., the sea level rose to 2.5 to 3.5 m above today's level. * At 2700 B.P., the sea level again experienced a low stand, dropping slightly below the present-day mean sea level. By about 2500 B.P., a third high stand was reached. At this time, the sea level rose 2 to 2.5 m above today's mean sea level, and since then it has been gradually dropping.

7 (300 to 1. Santa 2.5 b

In summary (Figure 3), three main phases of submergence (7000 to 5100 B.P., 3800 to 3600 B.P., and 2700 to 25001 B.P.) and three main phases of emergence (5100 to 3800 B.P., 3600 to 2700 B.P., and 2500 B.P. to the present) can be recognized along the central Brazilian coast during the Holocene.

stranc only

INFORMATION ABOUT PALEOSHORELINES TO BE DERIVED FROM THE SAMBAQUIS

Geographic Position

In coastal areas subjected to an emergence period following a submergence phase, such as the central coast of Brazil, prehistoric settlement sites located further inland were probably associated with a period of maximum lagoonal extent during a period of maximum sea level. Some central coast sambaquis occur as far as 30 to 35 km inland from the present shoreline. It may be assumed that the location of settlements is decided primarily by proximity

bival and 1

t r i r
Paleoshorelines and the Sambaquis o Brazil f

X1,oOO

area of

. Suguio,

v-lying 'stems. ked for

and Gui-

o 7000

ean sea slightly

to shellfish beds of sufficient richness to provide nourishment over a long period of time (Fairbridge, 1976). Obviously, shallow and quiet water bays, lagoons, and estuarine zones, with muddy surfaces exposed during low tides, would be more favorable locations than sandy areas at the open-ocean margin. In the former case, biological productivity is much higher. It is probable that the prehistoric occupants chose campsites above high-tide level, immediately adjacent to favorable collecting grounds where conditions of comfort and safety were available. If this is true, one may assume a close relationship between the positions of the sambaquis and the paleoshorelines, and the position of some sambaquis well inland can be explained only by a lagoon extension significantly greater, and hence a sea level higher, than at present. While it is relatively easy to establish the geographic relationship between the position of a sambaqui and a nearby ancient lagoon, estuary, or bay, it is much more difficult to establish the vertical relationship between the altitude of the sambaqui's substrate and the position of the sea level during its construction. We can only assume that initially it was above the local high-water spring tide (HWST) level, a very important assumption for interpreting sambaquis whose substrates are now located beneath the present HWST. It is possible that some of the latter sambaquis originated as wastedumps below houses which were built on poles standing in shallow water during high tides. This possibility can be eliminated by a detailed investigation of the sambaqui's substrate and composition. The assumption that sambaquis were built-up close to mollusk-collecting areas allows the nearest paleoshoreline to the site to be cited as the location of shellfishing areas at the time of the site's occupation.

Size of the Sambaquis

The sambaquis exhibit a variety of shapes and sizes, ranging from huge, elongated (300 X 60 X 60 m) or ellipsoidal (86 X 40 X 25 m) mounds to small, circular (10 X 1 to 1.5 ni) accumulations. One sambaqui examined by Fairbridge (1976) in the State of Santa Catarina was approximately 20 m high and 100 m in diameter, representing about 2.5 billion shellfish, that is, 100 shellfish per day for a group of 100 people for 500 years. In general, the more inland sambaquis are smaller than those situated near the present strandline. This is to be expected as maximum lagoon extent periods were very short. Thus, only in the outer lagoon zones was there sufficient time for the buildup of huge sambaquis.

low the rose 2 Ipping.

., 3800

.o 3800

central

Faunal Composition

About 50 different species of mollusks have been identified within the sambaquis of the Brazilian coast (Bigarella, 19491, but there are only five dominant species:
;e, such d were iximum present
Atiomalocnrdia brasiliann Gmelin Ostrea brasiliana sp. Ostren arborea Chemnitz Litcina jclmaicetzsis Chemnitz Modiolus brclsiliensis Chemnitz
Except for O. arborea, which lives fixed on aerial roots of mangrove trees, these bivalves live within sandy or clayey-sandy sediments deposited in shallow-water lagoons and bays. Shell remains of other organisms are generally very scarce.

proximity

Paleoshoreliries arid Prehisrory: Ari Irivestigatiori of Method
TABLE I. Variation of 61JC,,, as a Function of Position of the Samples Within a Paleolagoon, Coastal Plain of the State of S0 Paulo -Radiocarbon

Sample Itapo III

Jataituba Vapumava II

Age (B.P.)

5245 t t ft t t 100 3775t 130

613C,,, (o/oo)

Distance from Open Sea (km)

1 24. ? 33

surfacc appror If reoccu level I

- 0.63

-3.26 -0.06 -3.66 -0.12 -6.17 -0.02
Pariqiiera-Au Vapumafiva I A196

Ubatuba A132

imum been t

- 5.69

From Flexor, J.-M., Martin, L., and Suguio, K. (1979). In: Proc. 1978 Int. Symp. Coastal Evolution in the Quatemary, pp. 356-375. Suguio, K., Fairchild, T.R., Martin, L., and Flexor, J.-M. Eds. So Paulo. Used by permission.

during

C-13h2-12 Ratios of the Carbonate Shells
S1TPDB measurements of the carbonate shells show a spectrum of values changing as a function. of the influence of continental environmental conditions during the carbonate formation. Shells from lagoonal organisms show S3C,,B values between those for freshwater (about 13 o/oo) and those for marine organisms (about O o/oo), varying as a function of their geographic position within the lagoon (Flexor et al., 1979). Shells from outer zones (nearer the open-sea edge of a lagoon) show only slightly negative ST,,, values, while those from inner zones are clearly characterized by more negative values (Table I). In the latter cases, carbon derived from the decomposition of land plants has been incorporated by the mollusks within their carbonate shells. Deeper and inner portions of the lagoons are characterized by poor water circulation, promoting the accumulation of large quantities of plant-derived organic material. This is independent of water salinity and explains why the carbonate shells of inner lagoon organisms exhibit S13CpD, values characterized by a strong land influence. Conversely, according to the increase or decrease of the lagoonal area, shells from the same locality will have distinct S13Cp,Bvalues. If such shells are found in stratified sambaquis, their S13CpDB values will be good indicators of lagoon oscillations and thus indirectly of sea level fluctuations (Martin et al., 1985). For example, looking at Figure 4, if the sambaquis located at A, B, and C are of a similar age, they will be expected to show quite different S13Cp,Bvalues for their carbonate shells, with A being quite negative and C close to zero. On the other hand, if B and C show several periods of occupation, the least-negative S13CpDB values will correspond to periods of greatest lagoonal extent and the most negative to periods of lagoon contraction. These data cannot provide information on paleosea-level position with respect to the present level. However, they can indicate falling and rising lagoon levels and the approximate time of maximum and minimum still stands.

after 1 provic

sandy after

Dating of the Sambaqui

The usefulness of dates from sambaquis depends on the position of the dated sample within the mound. If a sample from the basal portion has been dated, it should correspond to the beginning of the sites occupation and thus provide a minimum date for the establishment of the neighboring shoreline. On the other hand, if the sample comes from the

Figure coast 2

surface of the sambaqui and thus dates the end of occupation, it should indicate the approximate date at which shellfish were no longer available in the vicinity. If central portions of inland sambaquis are dated, particularly periods of desertion and reoccupation, the departures may be tied to periods of lagoonal desiccation, or low-sealevel periods, and the reoccupations to periods of resurgent high sea level.

Types of Substrates

Sambaquis are located on four distinctive types of substrates (Figure 5).
Type I: Pleistocene marine terraces. These terraces are found inland from the maximum extent of Holocene lagoons. Therefore, the sambaquis found on them should have been built during periods of highest Holocene sea level. Type II: Holocene marine terraces. These terraces were deposited as beach ridges during the regressive phases that followed the sea level maxima, particularly the 5 100B. P. Santos transgression. Sambaquis on these terraces have necessarily been constructed after high Holocene-sea-level stands, and in some cases their positions on the terraces can provide additional data on sea levels. Type III: Ancient lagoonal deposits. These lagoon deposits are located in front of sandy marine terraces. Hence, the sambaquis found on them must have been constructed after a period of high sea level, probably during the ensuing lowering of the sea level. One
nging as arbonate )r freshfunction er zones s, while 1. In the rporated ions are itities of why the a strong al area, 'ound in ons and

Marimum lagoonal extent:

sambaquis in A , Band C. Intermediate lag. extent= sambquis in Band C. Present lagoonal extent: sombaquis only i n C.
are of a trbonate I and C pond to raction. present nate time
n A = sambaquis of only one age In B = samboquis of several ages

Oceanword

d sample mes pond he estabfrom the
Figure 4. Age variations of the sambaquis as a function of changes in lagoonal extension along the Brazilian coast during the Holocene.
Paleoshor-elines arid Prehistory: Ari Iiiwsrigario~zof Method

1 " 9Landward (-

Oceanward$ I
Figure 5. Distinctive types of substrates of the sambaquis along the Brazilian coast. (Modified from Fairbridge, R.W. (1976) Science. 191: 353-359.)
must be cautious in interpreting their elevation because the lagoonal organic clay is subject to compaction under the weight of the shell midden, thus diminishing the height of the substrate.

Type IV: Crystalline rocky hills. These hills are located at the margins or centers of ancient lagoons or bays. As they are never covered by high sea stands, they may have been occupied at any time and are thus not sensitive time markers. However, those situated more inland may have been settled when the lagoon reached its greatest extent.
COMPARISON OF GEOLOGICAL AND ARCHAEOLOGICAL DATA ON SEA LEVEL
Canania-Iguape Region (State of So Paulo)
The Canania-Iguape sedimentary plain has the shape of a large crescent and covers an area of about 2500 km*. Its outer portion is presently drained by a lagoonal system and tidal estuaries (Figure 6). A large part of the plain is occupied by remnants of more or less dissected Pleistocene terraces and by dried-out ancient lagoons, at the far inland margins of which several sambaquis have been fcund. Shell debris and wood fragments from littoral marine or lagoonal deposits have been dated, and this information, when associated with the nature of the sediments, indicates that (Figure 7A): About 6600 to 6500 B.P., the relative sea level reached approximately its present level and was rising.

Figui regio:

mat i on

Jr ,: : ;

Legend
Holocene paleo-lagoons Pleistocene marine terraces Lagoonal zone considered variation for t h e 6
.;!.. Holocene marine terraces :,.

Fairbridge,

Figure 6. Schematic geologic map of the Canania-Iguape coastal plain (State of So Paulo).

is subject ht of the

A-Information furnished by the geological samples
maximumYminimun modmum :*---- ---*! ----, !< ---- ---
:enters of ave been.led more
6 -Information furnished by the sambaquis
C-Information furnished by 6'3Cp0B
id covers stem and re or less margins ave been indicates
6 '3cpDB variations as a function
of time within three homogeneous zones O f the paleo-lagoon

D-Curve of synthesis m

esent level

ages B.Px1,OOO years

Figure 7 Reconstruction of relative sea level fluctuation curve for the past 7000 years in the Canania-Iguape. region (State of So Paulo).
Palcos/~orelir~es Preliistory: An Iiivesrigatioii of Metlrod aird

"

About 5400 B.P., the relative sea level was about 1.5 m above its present level. About 4400 B.P., the relative sea level was about 2 m highqrJhan its present level, but was falling. Hence, between 5400 and 4400 B.P., the relative sea level passed through a maximum height which, according to morphological data, must have been 3.5 to 4 m above the present one. Between 4400 and 3700 B.P., the relative sea level passed through a minimum. Between 3800 and 3700 B.P., the relative sea level rose very rapidly, reaching a second maximum of about 3 m above its present level. After the second maximum, the relative sea level returned more or less steadily to its present level. Unfortunately, we have insufficient data for detailing this return. Dating of sambaquis in this area confirms the conclusions on relative sea level reached on the basis of geological evidence. As indicated above, between 5400 and 4400 B.P., the relative sea level was higher than at present, attaining a maximum of 3.5 to 4 m above its current level. The maximum lagoonal extent which would accompany this high stand should result in the establishment of inland sambaquis (Tables II and III). The location of these sambaquis only makes sense if the sea level is higher than at present (Figure 7B) and the ages of these shell middens lie between 5200 and 4800 B.P. These dates narrow the estimate of the time for maximum lagoonal extent, which we now estimate at about 5100 B.P. Moreover, the location of later sites suggests that by 4100 B.P., the sea level was no more than 0.5 m above the present level. This confirms the conclusion reached on the basis of geological evidence that the sea level reached a minimum between 4100 and 3800 B.P. Table IV shows that the oldest of the late-Holocene sambaquis (Estaleiro, 80 B.P.) is situated on a Pleistocene terrace, which should indicate a high lagoon level. This hypothesis seems to be confirmed by the 613C,,, value of its shell carbonate. This value is much less negative than is that of the shells of Sambaqujnho, which was built in the same area when the sea level could have been 0.5 m above sea level at the most. This

>O m

< +OS m < +OS m < +1Sm

Indefinite

< + l m
Note: H.T. =Holocene Terrace and L= lagoonal.

- 3.17

-5.39 -4.67
1.63 3.26 1.48 1.06 -3.88 -0.76 - 1.34

- 1.28

Aplicada, ur-Yvette,
confirms the geological conclusion that between 3800 and 3700 B.P., the relative sea level reached 3 m maximum. The sambaqui Boguacu I (3080 k 55 B.P.) was constructed when the sea level could not have been more than 2.5 m above the present level, indicating that by 3100 B.P. the maximum of 3 m was passed. Finally, the information from the sambaquis Guarapari and Sambaquinho indicates that the relative sea level at about 2300 and 1500 B.P. was no more than 1 and 0.5 m above present level respectively. This provides data on the rate of sea level fall in the latest period not available from geological data. As discussed above, 613C,,B values can be used as an index to marine vs. continental influences and as indirect evidence of relative sea level change (Figure 7C). Three groups of sambaquis in the Canania-Iguape coastal plain can be used to assess the value of 83cp,B values in estimating sea level change (Figure 4: A, B, and C). The sambaquis in Group A were inhabited between 5200 and 4800 B.P., those in Group B between 5200 and 4800 B.P., and those in Group C between 3800 and 1500 B.P. Group A consists of seven sambaquis on the Cananeia coastal plain (Figure 4). The first six have ages ranging from 5200 to 4200 B.P. during the first Holocene transgression (Table V). It would be valuable to have information on a sambaqui constructed about 4100 B.P., when the sea level could not have been higher than 0.5 m above its present level. We could then compare the 8*3Cp,Bvalues of carbonate shells formed in the same geo-
Paleosliorelir~s Prehistory: An Iiivestigation of hfethod arid
graphic position between 5200 and 4100 B.P. encompassing the first Holocene regressive phase. Lacking this, we have considered a sambaqui dated to abput 1000 B.P., when the sea level was quite similar to that at 4100 B.P., assuming that the S13CpDB values will also be comparable. When plotted as a function of time (curve 1 in Figure 7C), the S13C,DB values show a maximum between 5200 and 5100 B.P., decrease rapidly until about4800 B.P., and then decrease slowly until 4200 B.P. Thereafter, the interpolated data suggest another rapid decrease to 4100 B.P. These 613Cp,Bdifferences can be interpreted both in terms of changing lagoonal extent and, assuming constant local tide range, in terms of changing relative sea level as follows:

Maximum lagoonal extent, and sea level maximum, occurred between 5200 and 5100 B.P. The maximum was followed by a rapid decrease in lagoonal extent, indicating a quick sea level lowering between 5100 and 4800 B.P. Lagoonal extent and, hence, relative sea level were stationary between 4800 and 4200 B.P. Lagoonal extent reduced rapidly after 4200 B.P., suggesting an equally rapid sea level lowering. Group B consists of three sambaquis from an area of the inland portion of the Iguape coastal plain (Figure 4, area B). The data derived from these sambaquis (Table VI) allows
the esta 510 are; are rapi rap. and and infc a f acc of i
TABLE V. Variation of 6'3CpDB a as Function of Age Within the Same Portion of a Paleolagoon, Coastal Plain of the State of So Paulo, Brazil
Sample Itapo III Guaxixi Juruvahva I Juruvahva III Juruvahva II Itapo II

Vamiranga

5245 & 5010i -C i 140 4215-C & 70

(o/oo)

- 0.63 - 0.48
-0.78 - 1.34 - 2.07 - 2.21 -3.81
are; of ' this are, tair allc reg difj the 8B
From Flexor, J.-M., Martin, L., and Suguio, K. (1979). In: Proc. 1978 Int. Symp. Coastal Evolution in the Quaternary, pp. 365-375. Suguio, K., Fairchild, T.R., Martin, L., and Flexor, J.-WI., Eds. S0 Paulo. Used by permission.
TABLE VI. 613CpD, Values of Carbonate Shells from Three Sambaquis Constructed in the Same Inner Zone of the Iguape Coastal Plain
Sambaqui Jataituba Pariqiiera-Au Momuna Radiocarbon Age. (B.P.)

+- 115

1% val da1

-3.26 -3.88 -5.81

egressive when the will also
TABLE VIL 6'3C,,, Values of Carbonate Shells from Four Sambaquis of the Canania Coastal Plain
Sambaqui Fosfasa II Estaleiro Guarapari Sambaquinho Radiocarbon Age (B.P.)
3790 +- +-45 1500a 120 613C,,, (o/oo) -2.63 0.67 -2.24 -2.61

:s show a

and then her rapid terms of changing

and 5100

5 a quick

md 4200 sea level

Iguape

:) allows

the creation of a curve of variation of 613CpDB which shows that the maximum was well established between 5200 and 5100 B.P. and that the fluctuation was very rapid between 5100 and 4800 B.P. (curve 2 in Figure 7C). Group C consists of four sambaquis located on the Canania coastal plain (Figure 4 , area C) which date from 3800 to 1500 B.P. (Table VII). The 613CpDB variation curve for area C is given in Figure 7C (curve 3). Between 3800 and 3500 B.P., values increase very rapidly and then decrease. Put in terms of relative sea level, it appears that there was a rapid rise between 3800 and 3500 B.P., followed by a drop. The results of 613CpDB analyses provide additional confirmation of both the geological and archaeological conclusions adumbrate above. Thus, radiocarbon dating of shell debris and fossil wood fragments from both sedimentary deposits and sambaquis, as well as information on the 6'3CpDB carbonate shells from sambaquis, have allowed us to establish of a fluctuating relative sea level curve for the Canania-Iguape region (Figure 7D). The accuracy of this curve is supported by the agreement between the three independent lines of information used to create it.

Santos-Itanham region (State of S0 Paulo)
The Santos-Itanham coastal plain has the same characteristics as the Canania-Iguape area. A lagoonal system originated during each high-sea-level period and sambaquis, some of which have been dated, were established at the margins of paleolagoons. The data from this area will be summarized rapidly as it confirms the results from the Canania-Iguape area. A large number of radiocarbon dates of shell debris and fossil wood fragments contained in lagoonal and shallow-marine deposits, as well as of vermetid incrustations, have allowed us to delineate a rather detailed curve of relative sea level fluctuations for this region (Figure SA). This curve is quite similar to that for Canania-Iguape, although some differences in amplitude are apparent. Dated sambaquis are much less numerous here than in the Canania-Iguape area, but they are important for corroborating the information from dated geological samples (Figure 8B; Table VIII). The sambaquis Rio Preto, Arara, and Mundo Novo are situated in an inner portion of the Itanham paleolagoon. Hence, they should have been inhabited during maximum lagoonal extent. However, if this hypothesis is correct, the conditions during maximum lagoonal extent were different in Itanham than in the Canania-Iguape region. The 613Cp,B values of the shells from these sambaquis are much more negative than those of the samples dating the high-sea-level period of Canania-Iguape. The 6'3Cp,B values of the sambaquis from Itanaham are greater than that of the sambaqui Momuna (Table II), which is the
Paleoshoreliries and Prchistoty: An Iiivesrigatiori of Method
A - Information furnished by the geological samples 8

ages B.P'x1,OOOyean

6-Information furnished by the sambaquis

C - Curve o f synthesis

Figure 8. Reconstruction of relative sea level fluctuation curve for the past 7000 years in the Santos-ltanham region (State of So Paulo).
TABLE VIII. General Characteristics of the Dated Sambaquis of the Santos-Itanham Coastal Plain

Sambaqui Piaaguera

Rio Preto
Radiocarbon Age (B.P.) f 145 3865295
Laboratory Number I4491 Bah 331 Bah 296 Bah 446 Bah 328 Gif 1194 Bah 382 Gif 9185

Nature of

Substrate C.B. P.T. C.B. C.B. P.T. H.T. L
Position of Sea Level >Om

>O m >O m

6WC,,,, (o/oo)
Arara Mundo Novo A229 Mar Casado Maratu
< $3.5 m < $3.5 m <O m

>Om

- 6.07 -5.29 -8.91 - 2.98
who sea 1 infoi leve shor shel' to tl. only the loca the cont

Nore: 1 = Isolopes, US.; Bah = Laboratrio de Fisica Nuclear Aplicada, Universidade Federal da Bahia, Brazil; Gif = Laboratoire de Faibles Radioactivits, Gif-sur-Yvette, France.
most negative of the Canania-Iguape sambaquis from this 4600-B.P. period. This suggests that the continental influence in the inner portion of the Itanham paleolagoon was very strong, perhaps because the maximum had passed and the paleolagoon was beginning to dry out. The sambaquis A229 and M r Casado (Table VIII), located on a Holocene terrace, a obviously were constructed after the transgression maximum, when the sea level was probably less than their 3.5 and 3-m elevations above the present level. This is supported by the moderate 613C,,, value of sambaqui A229.

Biga 6 Biga

Paleoshorelines and the Sambaqicis of Brazil
The sambaqui Maratu, which has, unfortunately, been completely destroyed, has furnished the most interesting information (Figure 8B). The base of this sambaqui was located below the present sea level. This fact cannot be attributed to substrate subsidence due to the sambaquis weight because the sambaquis layers were not deformed in any way. If the sambaqui had sunk under its own weight, the center would be expected to have lowered farther than the margins and the layers would be deformed. Laming-Empraire (1968) submitted two samples from this sambaqui to be dated at the beginning of radiocarbon investigations. The results, B.P. (Gif-15) and B.P. (Gif-16), were initially accepted because the dates did coincide with a period when relative sea level was definitely lower than today. However, archaeological remains within the sambaqui of Maratu suggested much later dates. Redating of Laming-Emperaires samples produced ages of 39252145.B.P. (Bah-382) and 3865-195 B.P. (Gif-9185), which are more in keeping with the artifactual evidence. They also confirm the hypothesis adduced from the Canania-Iguape region that relative sea level about 3800 B.P. was lower than today.

FINAL CONSIDERATIONS

Sambaquis cannot be used directly to reconstruct, in space and time, the positions of former sea levels as it is not possible to directly establish a vertical relationship between the base of a sambaqui and the sea level. The only thing of which we are more or less certain is that the base of the sambaqui was above the local HWST level at the beginning of its occupation. If we postulate that the paleoinhabitants established their campsites near the mollusk-harvesting place, it is easy to establish the geographical relationship between a sambaqui and an ancient lagoonal, estuarine, or shallow-marine zone. Sambaquis located further inland were probably associated with a period of maximum lagoonal extent, correlated with a sea level maximum. However, it is necessary to ascertain that coastal progradation was unrelated to intense continental sedimentation, but, rather, to a relative sea level change. If one has a series of dates from a number of sambaquis, they can be used to confirm or adjust the dates derived from geological materials. Thus, for example, sambaquis whose substrates are located beneath the present high-tide level suggest a period of lower sea level. Values of 613C,,, for carbonate shells from dated sambaquis give us additional information about the high- or low-sea-level positions, as well as about the trend of sea level change. Archaeologically, we: have confirmed the close association of sambaquis with past shorelines. This supports our hypothesis that the ancient inhabitants did not carry their shellfish dinners long distances, but, rather, established their settlements in close proximity to their molluscan resources. In studies of sambaquis in other zones, it will be possible not only to reconstruct the paleoenvironment of the site from its location relative to features of the paleoshoreline, but also to hypothesize a range of possible dates from a sambaqui located in its situation based on our sea level-fluctuation curves. Excavation and dating of the sambaqui will further test and refine the curves, so that the two areas of research continue to support and amplify one another.

suggests vas very nning to terrace, :vel was upported

REFERENCES

Bigarella, J.J. (1949) Nota prvia sobre a composio dos sambaquis do Paran e Santa Catarina. Bol. Arq. Biol. Tecnol. 4: 95-106. Bigarella, J.J. (1951a) Contribuio ao estudo dos sambaquis no Estado do Paran. I. Regies adjacentes s baas de Paranagu e Antonina. Bol. Arq. Biol. Tecnol. 516: 231-292.
Paleoshorelirtes and Prehisrorq: An Inivsiigariori of hdethod
Bigarella, J.J. (1951b) Contribui0 ao estudo dos sambaquis no Estado do Paran. II. Regies adjacentes baia de Guaratuba. Bol. Arg. Biol. Tecnol. 5/6: 293-314.Bittencourt, A.C.S.P., Martin, L., Vilas-Boas, G.S., and Flexor, J.-M. (1979) Quatemary marine formations of the coast of the State of Bahia, Brazil. In: Proc. 1978 Int. Symp. Coastal Evolution in the Quatemary, pp. 232-253. Suguio, K., Fairchild, T.R., Martin, L., and Flexor, J.-M., Eds. S0 Paulo. Empraire, J. and Laming, A. (1956) Les sambaquis de la cte mridionale du Brsil. Campagne de Fouilles (1954-1956). J. Soc. Am. Mus. de IHomrne. 45: 5-163. Fairbridge, R. W. (1976) Shellfish-eating preceramic indians in coastal Brazil. Science. 191: 353359. Flexor, J.-M., Martin, L., and Suguio, K. (1979) Utilisation du rapport isotopique 13C/12C comme indicateur doscillations lagunaires. In: Proc. 1978 Int. Symp. Coastal Evolution in the Quatemary, pp. 356-375. Suguio, K., Fairchild, T.R., Martin, L., and Flexor, J.-M., Eds. So Paulo. Hurt, W.R. (1974) The interrelationships between the natural environment and four sambaquis, coast of Santa Catarina, Brazil. Occas. Pap. Monogr. 1, 23 pp. Indiana University Museum, Bloomington. Laming-Empraire, A. (1968) Missions archologiques franaises au Chili et au Brsil. J. Soc. Am. 57: 77-99. Lfgren, A. (1893) Os Sarnbaqzris de S0 Paulo. Comiss. Geogr. Geol. Estado S0 Paulo, Bol. 9, So Paulo. Martin, L. and Suguio, K. (1975) The state of So Paulo coastal marine Quatemary geology: the ancient Strandlines. An. Acad. Bras. Cihic 47: 249-263. Martin, L. and Suguio, K. (1976) Les variations du niveau de la mer au Quatemaire Rcent dans le sud de ]tat de So Paulo (Brsil): utilisation de sambaquis (kjokkenmodings) dans la dtermination des anciennes lignes de rivage holocnes. Actes 42nd Congr. Int. Am. 9A: 73-83. Martin, L. and Suguio, K. (1978) Excursion route along the coastline between the town of Canania (State of S0 Paulo) and Guaratiba outlet (State of Rio de Janeiro). In: 1978 Int. S y i p. Coasfal Evolufioit in the Quatertiary Spec. Publ. 2, S0 Paulo. Martin, L., Flexor, J.-M., Vilas-Boas, G.S., Bittencourt, A.C.S.P., and Guimares, M.M.M. (1979a) Courbe de variation du niveau rlatif de la mer au cours des 7000 dernires annes sur un secteur homogne du littoral brsilien (nord de Salvador, Bahia). In: Proc. 1978 Int. Symp. Coastal Evolution in the Quaternary, pp. 264-274. Suguio, K., Fairchild, T.R., Martin, L., and Flexor, J.-M., Eds. S0 Paulo. Martin, L., Suguio, K., and Flexor, J.-M. (1979b) Le Quatemaire marin du littoral brsilien entre Canania (SP) et Barra de Guaratiba (RJ). In: Proc. 1978 Int. Symp. Coastal Evolution in the Quatemary, pp. 296-331. Suguio, K., Fairchild, T.R., Martin, L., and Flexor, J.-M., Eds. So Paulo. Martin, L., Suguio, K., Flexor, J.-M., Bittencourt, A.C.S.P., and Vilas-Boas, G.S. (1980) Le Quatemaire marin brsilien (littoral pauliste, sud-fluminense et bahianais). Cah. O.R.S.T.O.M. Sr. Gol. 9(1): 95-124. Martin, L., Suguio, K., and Flexor, J.-M., (198111982) Utilisation des amas coquilliers artificiels dans les rconstructions des anciennes lignes de rivage: exemples brsiliens. Cali. O.R.S.T.O.M. Sr. Gool. 12(2): 135-146. Martin, L., Flexor, J.-M., Blitzkow, D., and Suguio, K. (1985) Geoid change indications along the Brazilian coast during the last 7000 years. In: Proc. 5th Int. Coral Reef Congr., Vol. 3, 85-90. Tahiti. Martin, L., Suguio, K., Flexor, J.-M., Dominguez, J.M.L., and Bittencourt, A.C.S.P. (1987) Quatemary evolution of the central part of the Brazilian coast: the role of relative sea level variations and the shoreline drift. In: UNESCO Rep. Mar. Sci. 43: 97-145. Suguio, K. and Martin, L. (1976a) Presena de tubos fsseis de Callianassa nas formaes quaternrias do litoral do Estado de S0 Paulo e sua utilizao na reconstruo paleoambiental. Bol. IG, Instituto de Geocincias, USP, vol. 7, p. 17-26.

Sugui an Sugui

SS sugui ali E(

Sugui A

lit 2 f Villu
Paleoshorelitres and the Sambaqiiis o Brazil f
ry marine Evolution -M., Eds.
npagne de 191: 353C comme ln in the Eds. So uis, coast Imington.

Soc. Am.

Suguio, K. and Martin, L. (l976b) Brazilian coastline Quaternary formations: the states of S0 Paulo and Bahia littoral zone evolutive schemes. An. Acad. Bras. Cinc. 48: 325-334. Suguio, K. and Martin, L. (1978) Quaternary marine formations of the states of S0 Paulo and southern Rio de Janeiro. In: 1978 Int, Symp. Coastal Evolution in the Quaternary. Spec. Publ. 1, S0 Paulo. Suguio, K., Martin, L., and Flexor, J.-M. (1980) Sea level fluctuations during the past 6000 years along the coast of the State of So Paulo. In: Earth Rheology, Isostasy and Eustasy, pp. 471-486. Ed. Mmer N.-A., John Wiley & Sons, New York. Suguio, K., Martin, L.,Bittencourt, A.C.S.P., Dominguez, J.M.L., Flexor, J.M., and de Azevedo, A.E.G. (1985) Flutuaes do nvel relativo do mar durante o Quatemrio Superior ao longo do litoral brasileiro e suas implicaes na sedimentao costeira. Rev. Bras. Geocinc. 15(4): 273286. Villwock, J.A., Tomazelli, L.J., Loss, E.L., Dehnhardt, E.A., Hom Filho, N.O., Bachi, EA., and Dehnhardt, B.A. (1986) Geology of the Rio Grande do Sul coastal province. Quat. South Am. Antarct. Penin. 4: 79-97.

Bol. 9,

)logy: the it dans le la dter183. Canania

. Coastal

. (1979a)