1The flowering of the outstanding archaeological site of La Tolita around 2600 BP is a pertinent question in South American archaeology, because of its wealth and its very fast development (Valdez 1987; De Boer 1996). One of the commonly-accepted answers points to a strategic location that allowed interaction between the Pacific and the Andean (Carchi-Otavalo-Quito) cultural spheres (Fig. 1). The La Tolita site developed near the coast at the mouth of the Cayapas River (Fig. 2), which provided easy communication with the Andean communities. In this paper we show that the geographic setting that eventually favored the cultural flowering of the La Tolita site was inherited from a natural disaster that occurred ~3000 BP (Dumont, Santana et al. 2006), as evidenced by the opening of a new estuary which was shorter and wider than previously, as a result of a drainage deviation by faulting. This area has registered some of the strongest earthquakes of the 20th century (1906, Ms 8.7; 1942, Ms 7.9; 1958, Ms 7.8; 1979, Ms 7.9) (Kelleher 1972; Herd, Youd et al. 1981; Swenson and Beck 1996), as it did in 1868 and 1875 (Scheu 1911), some of them accompanied by tsunamis (Scheu 1911; Herd, Youd et al. 1981; Espinoza 1992) and vertical deformation of the coastal margin (Herd, Youd et al. 1981).
2The hypothesis presented here results of a multidisciplinary analysis of the data provided by three different projects that were developed separately: first a geomorphology study of the La Tolita site (CNRS/UMR 6012 and Pau University/IRSAM); then archeology (Museo del Banco Central -IRD/UMR 208) and more recently geology and tectonic deformations in the Esmeraldas area (IRD/Geoscience Azur-INOCAR cooperation project). The results of these endeavors were finally discussed by the participants of the different projects, in a pluri-disciplinary perspective looking for possible relationships between geomorphic changes and archeological cultural evolution. Finally, a common fieldwork analysis of the lithological sections of the El Indio archeological site brought consistency to the hypothesis formulated in this paper.
Fig. 1. Simplified topographic map of Northwestern Ecuador showing the location of the La Tolita Central Site (star) in relation to the to the Carchi-Otavalo-Quito cultural areas
Figure 2. Simplified map of drainage and structural patterns of the La Tolita and El Placer fan area (see location on Fig. 1)
In green the former drainage pattern of the early Pleistocene El Placer fan. In blue the fan beheading by motion of the Canandé fault zone during middle Pleistocene. In red the southward deviation due to La Boca uplift during the Holocene.
3The early development of the La Tolita culture (Stirling 1963) began on the Pacific coast of northern Ecuador around 3000 BP (Valdez 1987), and quickly spread throughout the interior alluvial plain of mangrove and thick rainforest vegetation up to an elevation of 200m along the western Andean slopes. Sedimentary analysis of the archaeological sequences shows that the main La Tolita settlement -or “La Tolita island site”- was located on a sand spit near the coast at the estuary of the Cayapas River (Tihay 1989). At first, it was probably just another site of the numerous settlements that appeared on the Pacific coast around the Cayapas-Santiago / Mira Delta. During the early occupation (3000-2600 BP) (Valdez 1987; Valdez 2006) numerous small settlements were spread out in an area extending along the coast on the ~3000 BP beach ridge up to a distance of 12 km south of the La Tolita island site (Fig.3). These Early La Tolita culture sites present unity of social organization and cultural values (Valdez 2006). The La Tolita island site and the other sites of the same period in the surrounding area belong to the La Tolita-Tumaco culture sphere of northern Ecuador and South Colombia, that extended over 500 km of the coastal margin. By 2700 BP the site began to grow around a central concentration of several earth mounds of different sizes (tolas or “Tolitas”). From this time onward the archaeological record becomes diversified, with a notable increase in quality, non-utilitarian, goods that denote a high degree of specialization in all the crafts. The traits that define the Tolita material culture become standardized and are all well represented throughout the northern equatorial Pacific coast. Their distinctive style seems to radiate from a central point located in the La Tolita island site. Resident master potters, metallurgists, sculptors (in stone, bone and wood) and very probably weavers (textiles have not survived in the humid acid subsoil) achieved levels of perfection that can rival with any other contemporary art forms from the rest of the world. Although early evidence of metallurgy has been found in the area surrounding the island site, by 3000 BP, the craft achieved its highpoint in La Tolita around 2300 BP (Valdez et al, 2005). Gold and tumbaga (a gold and copper alloy) articles were massively produced as ornaments, fancy vessels, and even everyday utensils such as needles, fishhooks or tweezers. The funerary offerings include body ornaments, laminated assemblages, miniature sculptures, and elegant dishware. In all Tolita representations its art style takes shape with a mixture of figurative and very realistic representations of humans, animals and objects, such as model houses, canoes or complicated geometric forms. A trait of this style is the symbolism that is rendered to all natural beings, evoking the supernatural forces and implying sophisticated ritual practices. Shortly before the turning to the Common Era (2000 BP) the site is probably the most important ceremonial center on the north Pacific equatorial coast, with many satellite centers radiating some 500 km along the seaboard. By this time, La Tolita is a ceremonial center and a regional necropolis, with both elite and commoner burials disposed on an extension of over a square kilometer. The massive amount of archaeological objects that are found today throughout the world, in museums and private collections, were once funerary paraphernalia, deposited to service the deceased in the afterlife.
4The causes of the emergence of the La Tolita culture and its main ceremonial center are probably numerous and all very different in nature, but it seems logical to think that the ideological power held by an elite of shaman priests was probably the principal drive behind the social organization of this complex society. By 2700 BP the Tolita center developed a way of life based on the practice of complex rituals, manufacture of related handicrafts and intensive agriculture. The economic base that sustained the center was probably trade in various forms. The nearby Laguna raised fields -artificaial areas of stockpiled earth to prevent cultures from excess dampness- (Fig. 3) produced agricultural high yields that could have been distributed by the Tolita center among the coastal peoples that inhabited the mangroves and tropical forest along the Santiago – Cayapas / Mira Delta (Valdez, 2006). In the same tenure, the production of luxurious and high status goods such as jewelry, fine textiles, symbolically charged pottery objects, etc was definitely suited for long and short distance trade. In this context, the location of the Tolita island ceremonial center was both strategic and tactical. Sea and fluvial navigation was the constant form of communicating with near and distant places. Archeological evidence shows that the center received visitors and merchants from all directions (Valdez, 1987).
5La Tolita was contemporary with other South American complex societies such as the Jama Coaque and Guangala, in present day Ecuador, or the later part of Chavin, Paracas, Nasca and the Vicus Moche, in Peru. Interactions with some of these coeval societies have been documented as far south as Frias, in the north of modern Peru. In this Vicus site, located roughly 1000 km away from La Tolita, some distinctive Tolita metal objects were found in the late 1960’s (Jones 1979). In the hinterland, interactions have been documented as far as Cumbaya in the valley of Quito (Buys and Domínguez 1987; Buys 1988) and in the Carchi-Otavalo region (Vernau and Rivet, 1912); Cesar Vasquez Fuller, personal communication 1985). The interaction, direct or indirect, between the coast and the highlands began surely before 3000 BP, but flourished between 2600 and 1600 BP. The presence of obsidian, a highland commodity coming from the eastern ridge of the Andes, near Quito, has been well documented in the Tolita region by 3000 BP (Burger, Asaro et al. 1994). Tolita ceramic figurines and round well-polished emeralds have also been reported in the Quito area around 1600 BP (Doyon 1988; Buys and Domínguez 1989; Doyon 1995).
6It seems evident that La Tolita’s coastal location, at the mouth of the Santiago Cayapas river basin, benefited from better communication with the hinterland than any other settlement in the region. Exchange developed between the coast and the highlands using the Santiago-Cayapas drainage, that flows straight down slope from the Andes. This is a striking element of the La Tolita culture that contrasts with the development of other sites in the same period. There is evidence of settlements related to La Tolita culture along the Andean tributaries of the Cayapas and Santiago rivers by at least 2600 BP (De Boer 1996). Exchange between the coastal and the highland cultures is evidenced since the Early La Tolita period, but trade seems to have increased notably from 2400 BP onward (Valdez 1987). The outmost expansion and the cultural climax of the La Tolita-Tumaco society was coeval with the fast development of La Tolita island site as a ceremonial center, which reached its zenith around the beginning of the Common Era. The center maintained its prominence until the year 400 AD, in which it was abandoned, but its material culture remained present until the year 800 AD in which all its manifestations ceased in the region. The causes for its collapse and eventual extinction have not yet been truly determined, but it is likely that the interactions that had knitted the cultural sphere began to decline throughout the region. It is possible that the ideological force that was the base of all interaction lost its appeal, probably in view of continuous failures to cope with natural events (i.e. harsh Niño episodes, epidemic outbreaks, etc). The structural collapse of the interrelated parties involved in the interaction sphere never has a single cause, but the evidence inferred from the material culture and the settlement patterns that replaced the Tolita Culture shows a very different general outlook on existence. It is very possible that the physical environment did not change at all, but the dominant worldviews certainly changed and produced a much simpler way of life. By the end of the IXth century AD La Tolita was already a long lost tradition.
7The present estuary of the Cayapas and Santiago Rivers, where the La Tolita culture spread and flourished after 3000 BP resulted from a stepped evolution of the drainage pattern of the area triggered by fault motion (Dumont, Santana et al. 2006; Santana, Dumont et al. 2006). We analyzed the drainage pattern on Spot images and aerial photos, looking for river deviation, abandoned valleys, and channel pattern anomalies, using geomorphologic methods of alluvial river analysis synthesized in Schumm et al. (2000), as well as geomorphic evidence of recent fault motion (Santana and Dumont 2002). The first important step of the evolution was the disruption of the El Placer Pleistocene fan, which was beheaded (Dumont, Santana et al. 2006) by the deviation of the Mira River to the north along the active Canandé fault zone (Eguez, Alvarado et al. 2003) (Fig. 2). During the Holocene the drainage of the central part of the El Placer fan was blocked against the rising La Boca uplift, and redirected to the SW toward the Cayapas River (Dumont, Santana et al. 2006; Santana, Dumont et al. 2006).
8Winckell et al. (1997) interpreted the concentration of the drainage in the area of Concepción (southern border of the La Boca Faults and uplift, Fig. 2) as an effect of local subsidence. These authors point out the fact that the Santiago River turns to the NE north of Selva Alegre (Fig. 2) to join the depression area of Concepción, abandoning a valley that led previously the river directly to the estuary. The trend of the new channel suggests a defluviation possibly controlled by NE-SW trending faults, in a style similar to the re-direction of the Mira River, and pre-dates probably the Holocene. This shows that the history of fluvial migration of the area is complex. Downstream from Concepción, the Santiago River together with the other streams of the El Placer fan that are collected along the southern border of the La Boca uplift, join the Cayapas River through a narrow valley that does not fit the importance of the stream (Winckell et al., 1997, unit 109, p. 212). We interpret the local subsidence in the Concepción area and the drainage changes downstream from Concepción as the result of the rise of the La Boca uplift that blocked and re-directed the drainage toward the Cayapas River through a new channel (Fig. 3, segment 1’-2).
Fig 3. Extract of the Spot Image 638-347 of 22/03/94 from the La Tolita area (see location on Fig. 2)
The large red star refers to the El Indio site, and the small stars show the location of the Early La Tolita culture archeological sites. Dotted line shows the central part of the depression of the “Laguna de la Ciudad”, bordered to the south by the Yanayaca fault (Fig. 2). The dashed line shows the limit of an area marked on the image by a fine grain, and interpreted as a flood area. Numbers refer to commentary in the text.
9The drainage re-direction generated an increase of discharge to the Cayapas River downstream from point 2 on Fig. 3. As usually observed in alluvial rivers a significant increase of discharge generate at least a flood, and possibly an avulsion by over bank flow, which is frequently initiated in a meander loop. Either the new channel joins the former channel some kilometers to hundred of kilometers downstream (Smith, Cross et al. 1989; Dumont 1992; Dumont 1996), or conect in any place the regional base level if the avulsion occurs near an estuary or a lake. The excess of discharge that generates the avulsion is generally temporary: when the normal discharge resumes the old channel that has been by-passed by the avulsion is partially or totally abandoned. In the first case the old channel develops a tipical underfit channel pattern that fits the remaining low discharge (Dury 1970). If the increase of discharge is permanent, resulting from a capture or a deviation, there is enough water to maintain active the old channel and the avulsion channel. This is the situation observed here. The geomorphologic analysis of channel characteristics shows that the Cayapas River presents the same meandering pattern (the relation between width and sinuosity) upstream from the confluence with the Santiago River (Fig. 3, segment 1-2), and downstream from the avulsion point along the old river segment (Fig. 3, segment 3-4). According to this observation the lower reach of the Cayapas River is the estuary channel that was active before the drainage deviation and the avulsion occur. It is important to note that we observe only one avulsion, with permanency of the old channel that pre-dates the avulsion. This situation is different from the drainage pattern observed in the areas where flood and avulsion are frequent and present a typical anastomosed drainage pattern (Smith, Cross et al. 1989), or the tectonic basins where an avulsion toward the structural depocenter results in the abandonment of the previous channel (Dumont 1991; Dumont and Garcia 1991). The dual estuary channel observed here (the previous Cayapas River channel and the new one after the avulsion) accounts for the new mean discharge, and do accommodate the temporary discharge variations, as no other avulsion is observed. The area observed on aerial photos and SPOT image with a smooth pattern (Fig. 3), and interpreted as the mark of a flood area, is spatially related to the only identified avulsion (Fig. 5).
10The dual morphologic pattern of the present estuary of the Cayapas River provides a clear account of the avulsion event. The upper segment, upstream from La Tolita (Fig. 3, segment 3-5) shows a fluvial pattern that cuts sharply through the beach ridge frame of the pre-existing coastal margin, whereas the lower segment, downstream from La Tolita to the estuary mouth (Fig. 3, segment 5-6) presents a normal estuary pattern, with beach ridges curved toward the estuary (Dumont, Santana et al. 2006). The interpolation of 14C dated beach ridges (Tihay 1989; Tihay and Usselmann 1995) near the point of changing pattern of the estuary channel allows an estimation of 2800-3200 BP for the avulsion (Fig. 3, ~3000 BP beach ridge) (Dumont, Santana et al. 2006). The beach ridge accretion pattern presents relatively parallel lines except during the period 4000-3000 BP (Fig. 3). This deviation suggests a faster accretion of beach ridges during the period that predated the avulsion event. The subsidence tendency seems to resume about 500 years after the event, and a situation similar to the present is observed from around 2000 BP (Dumont, Santana et al. 2006). The Early La Tolita archeological sites are presently found 1-2m below the mean upper sea level (Fig 4) (Valdez 2006), suggesting a total subsidence of at least 3-4m since the occupation.
Fig. 4. Field profile section from the El Indio archeological site
1: grey-green shelly sand from the beach ridge; 2: grey sand reworking clay fragments; 3: grey silt-clay sand including early La Tolita ceramics (ELTC); 4: flood level, limy sand with wood fragments, 14C dated (cal. 2 sigma) 3220-2960 (Beta 223218); 5 and 5’: refuse midden, 14C dated (cal. 2 sigma): 2868-2702 BP (Gif 11900) and 2865-2730 BP (Beta 181458). In a 140cm deep dig pit some 5m inland, the same refuse midden layer provided the gold element showed to the right (size 17mmx11mmx0,1mm); 6: Sandy clay including early La Tolita ceramics (ELTC); 7: yellowish sandy clay; 8: silty sand filling a drainage channel; 9: archeological layer including Early La Tolita ceramics (ELTC); 10: ash; 11: Archeological layer including Late La Tolita ceramics (LLTC); 12: soil cover.
11The El Indio (also known as Las Balsas) archeological section (Fig. 4) is located south of La Tolita (Fig. 3), north of the main depression of the “Laguna de la Ciudad” (Tihay and Usselmann 1995; Valdez, Gratuze et al. 2005). It is included inside the flood area observed on the aerial photo, near the western border. A field section of ~3m is observed by cleaning a tide channel bank during low tide. The section presents 3 parts (Fig. 4): (A) a massive and homogeneous basement made of beach ridge sand including shell fragments, covered by a thin level of sand from the beach ridge and clay fragments from local origin. The part (B) is complex and corresponds to the fill of a local drainage that was oblique to the present tide channel. The part (B) includes sedimentary and archeological levels from the Early La Tolita Culture (ELTC). The brown level (Fig. 4, n° 3) presents an archeological midden containing ceramics and other cultural debris in a sandy-silt matrix (Valdez 1986; Bouchard and Usselmann 2003; Valdez 2006). An irregular but clear limit separates this lower level from a 50-70 cm thick homogeneous light brown layer of sand, including scattered centimeter-size wood fragments (Fig. 4, n° 4). The granulometry presents few variations across the different levels, all with silt size coming from alluvial deposits or reworked from flood deposits. The flood level is differentiated for its clearer color, homogeneity and the presence of a discontinuous coarser base level. There is no evidence of deposits of marine origin (i.e. from a tsunami) although their occurrence would not be unexpected. The AMS 14C dating of wood fragments scattered in the middle part of the flood deposit yielded a calibrated age of 3230 to 2960 BP. The flood level is covered by two lenses of refuse (Fig. 4, levels n° 5 and 5’) and a layer of brown sandy-silt similar to the lower level (Fig. 4, level n° 6), and includes similar archeological remains. This level of refuse yielded two 14C dates: respectively 2868-2702 and 2865-2730 BP. Both levels of human occupation located below (n° 3) and over (Fig. 4, levels n° 5 and 6) the flood layer, belong to the same cultural occupation. However the refuse level (Fig. 4, level n° 5) that post date the flood event (Fig. 4, level n° 4) includes the oldest occurrence of gold metallurgy of the La Tolita Culture Sphere. An ash layer ends the part (B). The upper part (C) is a monotonous archeological layer (Fig. 4, level n° 11) including late La Tolita ceramics from the Late La Tolita Culture (LLTC). A thick sedimentary cover (Fig. 4, level n° 12) of dark brown soil ends the section.
12The first evidence when comparing geomorphic changes and archeological culture evolution is that the apparition of the La Tolita island site postdates the avulsion that created the present estuary. As this archeological site was settled on an island at the estuary mouth, it is logical to assume that this would only be possible after the estuary existed. However, the most important settlement in La Tolita island appeared by 2778-2545 BP (Valdez 2006), that is about 200 to 300 years after the avulsion (Fig. 5). From the pre-avulsion occupation of the coastal area by ELTC sites to the emergence of the La Tolita island ceremonial center what was the relation between the succession of geomorphic and cultural events?
Fig. 5. Recapitulation of dating elements from the La Tolita area, in relation to the correlation between the geomorphologic changes and the cultural evolution
D and E refer to the location of 14C dates in Santana et al., 2006.
13An important point for the comparison between geomorphic and archeological cultural evolution, and in particular for the interpretation of the El Indio profile section, is the interpretation of the flood deposit as representative of the avulsion that opened the new estuary (Fig. 4). As stated before, there is evidence for only one avulsion episode, and the flood area identified on the remote sensing documents is located around the avulsion channel. The dating of the avulsion is not direct, it is estimated from the ages of the closest beach ridges to the point of change of the morphology of the lower Cayapas-Santiago channel, from river to estuary system. Two beach ridges, distant about 500m, are dated in this area, giving 14C conventional ages of 3251-2854 BP (Fig. 5, D) and 3260-2830 BP (Fig. 5, E) respectively (Santana et al., 2006, fig. 5, D and E). The D point is the closest to the line of changing estuary morphology, but both date fall in the same range (Fig. 5). This is coherent with the age of 3220-2960 BP obtained from wood fragments in the flood level (Fig. 5). In the interpretation of the El Indio profile section we correlate the attribute flood level with the avulsion, on the basis of dating and particular importance of the flood level. This flood level was identified as particular and important before and independently from the analysis of the avulsion geomorphology (Valdez 2006), on the basis of the erosion base level and thick basal granulometry. Other levels of thin sand and silt are observed in the El Indio section and in other profiles attributed to flood deposits, in particular in the upper part of the sections located in the Laguna de la Ciudad, which have been reported to repetitive flood (Valdez, Gratuze et al. 2005; Valdez 2006). They represent repetitive floods, seasonal or pluri-annual, probably associated to the progressive subsidence of the area.
14In the El Indio profile section, the oldest level including ELTC ceramics (fig. 4, level 3) just predates the flood level. This archeological level is just separated from the shelly sand of the beach ridge located below by a thin layer of sand from the beach ridge including reworked clay fragments from the back ridge depression, as frequently observed behind a beach ridge. An erosion contact cut the early ELTC level at the base of the flood level (Fig. 4, levels 3 and 4 respectively). There is no 14C dating from the early ELTC level (Fig. 4, level 3), but the good coherency between the dating of the beach ridge obtained from the extension of the El Indio beach ridge to the south and the flood level leave few possible variations in age (Fig. 5).
15Valdez (Valdez, Gratuze et al. 2005; Valdez 2006) points out the presence of several ELTC sites along the 3000 BP beach ridge, as well as further inland in the area of the Laguna de la Ciudad (Fig. 3). This supposes that these areas represented at that time relatively safe morphology regarding the risk of flood from rivers and destruction by waves for the sites located near the beach ridge. This is a situation drastically different from the one observed now. The beach ridge in the El Indio section in now near the low sea level registering a subsidence of at least 3-4m, and the Laguna de la Ciudad is (as its name implies) an extensive wetland with 2m of flood deposits covering the last archaeological level (Valdez, Gratuze et al. 2005; Valdez 2006). Both of these locations show a clear indication of subsidence. The emergence of the La Tolita culture and its permanency raises the question of possible specific geomorphologic conditions that benefited the area and made the regional subsidence less effective for a period of some hundred years at least. The geological context suggests three hypothesis. The first one considers that higher sediment transport to the coast generated more active beach ridge accretion during a relatively large period, at least from the 4000 BP beach ridge. The second hypotesis considers a variation of the sea level, bsically a light lowering, since the mid Holocene (about 6500 BP) post glacial transgression. The third hypotesis contemplates a temporary vertical uplift of the coastal margin due to stress accumulation in the frame of a seismo-tectonic cycle. The hypothesis (1) lacks arguments, as climate and sedimentary transport conditions to the coast do not seem to have changed drastically since the mid Holocene. Also a faster beach ridge accretion can be obtained by a lesser subsidence as stressed in hypothesis (3). Regarding the hypothesis (2) there is some indices to set the onset of the post glacial transgression at a higher sea level than the present one (~2m) as suggested by Tihay (1989), in agreement with the global post glacial sea level curve in the equatorial zone (Pirazzoli 1991). The hypothesis (3) lacks of precise control but is coherent with observations of light uplift and subsidence in relation with the seismic cycle. According to this interpretation stress accumulation raises the coastal margin before a tectonic event, and subsidence is observed as the stress is released. A combination of the processes hypothesized in (2) and (3) may give account of particularly favorable geomorphologic conditions in the area, allowing a relatively safe occupation of the coastal margin.
16The effect of tsunami is not stressed in the interpretation, principally because we did not find clear evidence of such an event in the lithologic record, or in the El Indio profile section. Tsunami waves are related from the 1906 (Scheu 1911) and 1979 (Herd, Youd et al. 1981) earthquakes in the area of Tumaco, in relation to focal mechanism located offshore, between the trench and the coast. The effect of a tsunami on a specific place of a coast depends principally from the type and position of the offshore source, and the bathymetry of the near shore area off the coast. Two remarks can be made relatively to the present case: A) the coast in front of the Esmeraldas-Tumaco area presents two caps, one of them the Galera Cap associated with an offshore promontory, that drastically reduces the effect tsunami waves coming from southwest of Esmeraldas (Mansour Ioualalen, personal communication, September 2009); and B) the context of the drainage re-organization related to the La Tolita site suggests fault motion -and possibly source- located inland or near the coast (the rise of the La Boca uplift) that should not be able to produce a tsunami wave.
17As far as the correlation between the flood level observed in the El Indio profile section and the avulsion is accepted, the El Indio section shows that the ELTC occupation predates the flood, and resumes later. The stratigraphic record is continuous, but as the section is located in a local topographic low used at least episodically by local drainage, some layers are reduced or may lack due to erosion. An erosive contact is observed at the bottom of the refuse midden (Fig. 4, level 5) that overlay directly the flood level (Fig. 4, level 3). The 14C ages obtained from the midden suggest that the erosion contact between the flood and the refuse deposit represents a gap of at least 90 years (Fig. 5). At regional scale there is no evidence of a gap of human occupation or a cultural change inside the ELTC period (Valdez, Gratuze et al. 2005; Valdez 2006). Thus the time gap observed between the flood level and the midden may represent the irregular conditions of sedimentation in the low drainage line where refuse deposit was accumulated.
18However, the importance of the El Indio profile section is less due to the relative discontinuity of the ELTC deposits -the constant occupancy of the area is assessed at regional scale (Valdez, Gratuze et al. 2005; Valdez 2006), - than to the occurrence of the oldest manufactured gold artifact of the ELTC (found in the midden) that postdate the flood level (Fig. 5). This is considered as an important change in the evolution of the ELTC (Valdez, Gratuze et al. 2005). On the basis of chemical analysis Valdez et al. (2005) stress that the gold nuggets came from a placer washing area located at the Santiago river. The geomorphic evolution shows that there was no connection between the Santiago River and the La Tolita area before the avulsion event. That does not mean that gold nuggets from the Santiago River were unknown before, but it is easier to understand the occurrence of gold artifacts in the La Tolita area after the geomorphic reorganization. An other point underlined by Valdez et al. (2005) is the technique used to work the early gold artifact; it was hammered and assembled to form a thin sheet that was subsequently cut, shaped and perforated to be assembled as an ornament. These techniques are simple but need some ability and knowledge. The time gap of at least 90 years between the flood and the first appearance of gold artifacts is too long to establish a close relation of cause/effect between the two events. On one hand, we refer here to the El Indio lithologic section showing an erosion gap, thus leaving the possibility of a shorter gap not documented in this section. On the other hand, the geomorphic reorganization has obviously modified the sedimentary transport and deposition of the area, redirecting the drainage from the gold bearing area of the Santiago River toward a new single estuary channel. It necessarily needed some time to be discovered as a new opportunity to find the raw material that could be worked here or in some other place. It must be remembered that the wet tropical forest conditions make the discovery of archeological sites a hard task. What seems clear is that the new channel gave access to gold deposits, and that could have been crucial to the development of a goldsmith specialization in the Tolita center. As the gold placers began to be exploited the Tolita site gained fame as a unique production center that was visited by a growing population which had contact with different areas, that included the sierra. In turn, the availability of the new resource allowed the center to gain control of the extraction, the transformation and the eventual trade of the gold artifacts (Valdez, Gratuze et al. 2005).
19According to the available dates, the ceremonial center in the La Tolita island (2778-2545 BP) begins to grow soon after metallurgy is practiced in the Santiago- Cayapas drainage area, as suggested by the presence of early gold artifacts found in the El Indio site (2868-2702 BP and 2865-2730 BP) (Fig 4). If we examine the regional context and the chronology of these events, it seems that the inital settlement on the La Tolita island was determined by economic and communication criterion than by a random choice. In any case, the strategic location, both coastal and riverine, and the close access to the gold bearing area on the Santiago river certainly seem to have played a major role in de development of the site as a regional ceremonial center. The new access to placer area, through the avulsion channel and the re-directed drainage system, made the gold deposits more evident and perhaps even easier to wash. Thus the economic impulse given by the strategic location and the goldsmith craft specialization, made La Tolita island a very suitable place for cultural interactions involving ideological power management.
20The La Tolita area in northern Ecuador and southern Colombia presents exceptional scenarios of both geomorphologic and cultural changes, apparently structured together in time and space. Whether this is a fortuitous relation, or not, can not be clearly established, but what seems clear is that the geomorphologic reorganization of the estuary did favor the cultural events that followed.
21The geomorphologic reorganization could have influenced two of the main elements of the La Tolita cultural preponderancy: the occurrence of a specialization in gold metallurgy, and the communication with the deep hinterland (and eventually up to the Andean sierra). The first point to be stressed is that the geomorphologic reorganization stands -probably abruptly- on the scene on which the La Tolita Culture developed. How did this happen? A deterministic view may consider that the early La Tolita people, that were present in the area before the geomorphologic reorganization, took immediate advantage of the new opportunity for communications with the interior and its newly found resources. This seems to be conform to the general picture of the geomorphologic-cultural relationships, in which two elements favor this interpretation: (1) the fact that coastal - sierra relationship did exist before the geomorphologic changes, but were never predominant or brought forth a major cultural change. This was certainly emphasized when the new channel access opened a shorter and more direct connection to places and new resources. (2) The chronologic succession of events that shows the apparition of the La Tolita island ceremonial center soon after the avulsion and opening of the new channel.
22The chronology of events shows that a major cultural event marked a change in the activities of the region after the geomorphologic reorganization. This event was an increase of gold extraction, metallurgy and craft specialization in the area. This new activity favored the economy and triggered exchanges with both coastal and sierra regions, where gold metallurgy was poorly known. Finally, it is only about 300-400 years after the geomorphologic reorganization that the people reorganized the activities in the La Tolita island site, the center was the only point on the coast that intersected the fluvial communication lines imposed by the introduction of intensive gold extraction and symbolic trade in the economy of the area.