Estudio especializado N° 1. Pollenanalysis of the telarmachay rockshelter (PERU)
p. 365-373
Texte intégral
INTRODUCTION
1The Telarmachay rockshelter is situated on the edge of the high plain of Junin (Perú)’, at an altitude of 4 420 m, and at a distance of 20 km from the large Junin Lake (approx. 11°11’ lat. S 75°52’ long. E). The area is in the upper part of the « alpine zone », in the Puna grassland formation. The rockshelter was excavated by a team guided by D. Lavallée and M. Julien of the Centre de Recherches Archéologiques (C. N. R. S.), between 1975 and 1980. The samples were sent to us by D. Lavallée, who also informed about the archaeological sequence and 14C dates from the rockshelter. The samples received consisted in the first place of a series of 9 samples from one sector (A12), in the back of the shelter, not far from the rockwall, and representing the entire excavated sequence: layers I (top) to VII (bottom, resting on the rocky basement). In the second place there are a number of samples from the layers IV to VII incl. from different sectors of the excavation, and some samples from puna soil outside the rockshelter.
2All the samples, collected by D. Lavallée, were prepared in 1981 in the Laboratoire de Palynologie du Centre de Recherches Archéologique du C. N. R. S. (Valbonne) by P. Guyomarc’h, under the direction of M. Girard1. 28 samples were treated chemically in the normal way, mostly completed with separation in a heavy liquid; they were treated according to the Schulze-method (nitric acid and potassium chlorate). Slides and residues were sent to Amsterdam for analysis (M. Girard, written communication). All the samples from the A12 series contained sufficient pollengrains. From the other 13 rockshelter samples 7 contained sufficient pollen, while from the puna soil only the deepest (50 cm) sample out of a total of 4 contained pollen. The total of 17 pollen-containing samples was analysed palynologically. Of the A12 series a pollendiagram is given (Fig. 1), the other samples are given as sepárate pollen spectra placed in stratigraphic order (Fig. 2). For the situation of the sectors where the samples were taken, we refer to the map in Fig. 3.
3As to the 7 layers differentiated by the excavators, the following data are of importance (see Lavallée & Julien, 1979 and 1983):
Layers II and III: « formative » period, ca. 3.800-2000 BP
Layers IV and V: « preceramic » period, ca. 6.800-3800 BP
Layers VI and VII: « archaic » period, before ca. 6.800 BP.
4Domestication of Camelidae was found to have taken place approximately between 6000 and 5800 B. P. In the upper part of layer V domesticated Camelidae are definitely present.
The main pollendiagram: description and interpretation
5The spectra of the pollendiagram of the A12 sector (Fig. 1) are based on a pollensum (sum of calculation) that contabas all pollendiagrams; spores (Lycopodium, Monolete spores, Isoetes, Fungi) are excluded from the sum.
6Three major (provisional) pollenzones may be distinguished, two of them subdivided in two subzones. A short description follows.
Zone III | – Gramineae very hight percentage |
Zone II | – Gramineae and Compositae lower than in III |
Zone I | – Compositae very high percentage |
7Comparison of the boundaries of these zones with the « archaeological » layers and the 14C dates, allows the following approximate correlation and dating.

8Hence, the boundary of pollenzones I a/b might have age of c. 7.000 years B. P., the boundary I/II c. 6.500 B. P., II a/b c. 5.000 B. P. and II/III c. 3.000 B. P. Although there are clear differences in pollen content between the zones and the approx. ages are known, a sure interpretation in terms of vegetation and climate is difficult because of lack of pollendiagrams from lake-deposits of the area and because of lack of detailed knowledge of the vegetation in the area.
9One of the principie early studies on the vegetation of the Peruvian Andes is that of WEBERBAUER (1911). Recently extensive vegetation studies in the area were carried out by Dr. P. Gutte, Leipzig. Although his data were not yet fully published (but see GUTTE & GUTTE, 1976), Dr. Gutte was so kind as to inform us of some of his results, and express some ideas on the possible interpretation of the pollendiagram.
10The presence of Isoetes in Ib and of relatively abundant Fungi spores in I and IIa seem to indicate generally more humid circumstances than during zones IIb and III. This agrees with the opinion of DOLLFUS, 1965 (who remarks that there were more lakes in the Peruvian high Andes at that time) and others (see the review of published data in MARKGRAF & BRADBURY, 1982).
11The presence of the Chenopodiaceae-Amaranthus-type pollen in zones II and III, but principally in IIb and III, most probably indicates anthropogenic influences (cultivation).
12Besides the interpretation of these two characteristics of the pollendiagram, it is as yet difficult to interprete with certainty the other salient features. The main problem is the interpretation of the very high percentage of Cruciferae-pollen (Draba-type) in zone II (especially Ilb), correlated with low Compositae percentages (Gramineae percentages are lower before and higher afterwards). At first sight this seems to mean that vegetation of the subnival zone, with abundantly Draba (to day probably occuring some 100 to 200m above the site), prevailed. This would mean a some what lower temperature during zone II than during zones I and III; this seems to be in contradiction to the world-wide trend of a warmer climate before and a cooler after 3.000 B. P. It is however, imaginable that a drier climate would also have a lowering effect on the lower limit of the subnival zone. Another explanation could be that low open bush of Escallonia and Buddleja (with species that release only very little or no pollen in the air) spread in the area, and that the Draba-type curve represents other Cruciferae, that could have been abundant in such open scrub of the « subalpine » zone (like Eremodraba, Sisymbrium and several others2). This would mean a rise of some 500m of the subalpine scrub-puna grassland boundary, and hence a considerable rise in temperature. However, if this were true, Polylepis and Eugenia that produce a reasonable amount of pollen could then also be expected (ELLENBERG, 1979), but not a single pollengrain was found. Another possibility is that a number of features is caused by anthropogenic influences. As domestication of Camelidae started in zone II, grazing should have had (besides cultivation of plants) an ever increasing influence on the natural vegetation. There are many indications that the abundance of trees and scrubs was originally much greater in the (wetter) puna, and was greatly reduced during several millennia of human influence (ELLENBERG, 1979). In the Southern part of South America, there are clear indications that the highest temperatures of the Holocene occur in its early part, while the later Holocene is somewhat cooler. According to the available data in the Northern tropical Andes, warm and dry conditions culminate around 7.500 B. P. Between approx. 6.000 and 5.000 B. P. there is a cooling of the climate and an increase of moisture (lake levels high). Consequently there is, in Bolivia and N. Argentina, an increase in aridity, while it became warmer in Columbia and Venezuela (MARKGRAF & BRADBURY, 1982; VAN DER HAMMEN, 1974). If we accept this, and also take into account the human influence, the following rather logical interpretation is possible.
13During zone I a Puna grassland rich in shrubs and herbs occured (Compositae, Malvaceae (like Nototriche), Caryophyllaceae, ferns and Isoetes). Cactaceae were possibly abundantly represented by the cushion plants Tephrocactus floccosus and/or T. lagopus, today occuring between 4.000 and 4.500 m. The climate was relatively warm and humid (Isoetes; Fungi), especially shortly after 7.000 B. P. (zone Ib). During subzone IIa this puna vegetation is becoming poorer in shrubs and herbs, while there is a remarkable extension of Draba (type). This might be explained by a lowering of the subnival vegetations because of a cooler climate (approx. 6.500 – 5.000 B. P.). In this period domestication of Camelidae starts, and perhaps cultivation (Chenopodiaceae-Amaranthus type); jthis may equally have contributed to an increase in grasses and decrease in shrubs and (other) herbs.
14During subzone IIb (c. 5.000 – 3.000 B. P.) the Draba-type reaches its maximum, Fungi decrease sharply. Apparently there is a subnival type of vegetation: the relatively low lower limit of the subnival vegetation may now have been caused by a notably drier climate (with greater diurnal temperature extremes). There might also be an effect of increased grazing, but by the time there is no way of knowing this for certain. The scarceness of (other) herbs in this subzone however, might best be explained by anthropogenic influence.
15During Zone III (after c. 3.000 B. P.) there is a considerable increase of Gramineae (grasses) and certain Compositae; the Draba-type almost disappears. This might have been caused by a rise of the lower limit of the subnival zone because of an increase of humidity. An expected increase of Fungí however does not take place, possibly because of a general drying up of the surface of the soil of the rockshelter due to continuous inhabitation. Human influence (grazing, burning) on the vegetation may now have become very important.
Other pollen spectra
16A number of samples from different depths and belonging to the archaeological layers V, VI and VII were analysed, for comparison with the main pollendiagram of sector A12. A number of these samples did not contain sufficient pollen (5 out of 12), and of the remaining 7, seperate spectra were drawn (Fig.2).
17The spectra corresponding to archaeological zone V correspond very well to those of pollenzone Ha, with lower percentage of the Draba-type in the lower part and a high percentage in the upper part; Fungí are especially high in the sample from the lower part. There is some variation in the Gramineae values, and percentages both somewhat higher and lower than in the diagram occur.
18The pollenspectra from archaeological layer IV corresponds rather well with the zone I (b?) of the diagram. Fungí spores are relatively frequent, but Isoetes is lacking. The spectrum from layer VII is similar to that of pollenzone Ia; Isoetes is lacking and Fungí are higher.
19The conclusion is that the spectra from layers V, VI and VII from other sectors of the excavation area in the rockshelter confirm the general sequence of the pollendiagram.
20From a sequence of 4 samples (5 – 50 cm depth) from a puna-soil outside the rockshelter, only the one at 50 cm contained sufficient pollengrains. In the corresponding spectrum (Fig. 2) it is noteworthy that regional « background » pollen like that of Alnus, Podocarpus, Hedyosmum and Ephedra are represented with a much higher percentage than in the rockshelter. The percentage of the Amaranthus-type resembles that of the youngest spectra of the diagram. The low background pollen in the rockshelter sediment may be caused by the more wind protected situation.
DISCUSSION AND CONCLUSIONS
21The pollendiagram of the Telarmachay rockshelter shows a clear vertical differentiation and some 5 subdivisions may be distinguished (pollenzones la and b, IIa and b, and III). The diagram covers at least 9.000 years (the Holocene). Pollenspectra from other sectors of the rockshelter and from different archaeological layers confirm the sequence.
22Without pollendiagrams of lake sediments from the area and a more detailed knowledge of the vegetation types and vegetation belts, a sure interpretation of the pollendiagram in terms of vegetation and climate is difficult.
23The present altitude of the site is 4.420 m a.s.l. The present main climatic vegetational belts are the following (WEBERBAUER, 1911; GUTTE, pers. inform):
nival zone: above 5.200 m
Subnival zone: 4.600/4.700 m – 5.200 (communities with incomplete vegetation cover)
« Alpine » zone: 3.900 m – 4.600/4.700 m (tussock grasslands)
« Subalpine » zone: 3.000 m – 3.900 m (evergreen scrub and grass « steppe »).
24The vegetational succession reflected in the pollendiagram, starts (zone I) with grassland rich in Compositae, Malvaceae (prob. Nototriche, Acaulimalva), Caryophyllaceae (prob. Arenaria, Cerastium, Pycnophyllum), Gentianaceae (prob. Gentiana/Gentianella) and Cactaceae (prob. Tephrocactus floccosus and/or lagopus). The compositae might partly represent shrubs (shrubby species often produce much more flowers and therefore much more pollen than herbaceous species), but this is, of course, not certain. The relatively high percentage of Fungi and the relatively high percentage of Isoetes spores is noteworthy. The interpretation seems to be that a puna vegetation rich in herbs (and possibly some low shrubs) dominated near the rockshelter and that the local environment was relatively humid.
25During zone II the Draba-type increases (IIa) and reaches a maximum (IIb). During IIa a number of herbs decreases or almost disappears (Malvaceae, Caryophyllaceae, Cactaceae, ferns); the Compositae equally decrease strongly. At the end of lia the Fungi decrease sharply. The most logical and simple interpretation could be a fall of temperature, bringing a few hundreds of meters down the subnival zone, with abundant occurence of Draba species. This would logically lead to a more humid climate, that indeed might be deduced from the relatively high Fungi values (and possibly the Cyperaceae) during IIa. A cooler climate and increased moisture is indeed known to have occured in the northern Andes between approximately 6.000 and 5.000 B. P., the approx. age of zone IIa.
26Zone IIb, with its high Draba values and low Fungi, seems to be cold and dry: the most logical interpretation seems to be that the vegetation of the lower subnival zone dominated on the spot. An increased aridity has been reprted for Bolivia and N. Argentina for this period, and a warmer climate in Colombia and Venezuela (see above). Accepting a much drier climate for this period (appr. 5.000 – 3.000 B. P.) could explain the scarceness of herbs: it may moreover have led to much greater diurnal temperature fluctuations (more frequent high frost; the average temperatures may have remained the same or even have been higher), that caused the low position of the lower boundary of the subnival zone.
27It can moreover not be excluded, that anthropogenic influence (domesticated Camelidae) had some additional influence.
28At the beginning of zone III there is a steep fall of the Draba-type, and Compositae and some herbs show an increase. Gramineae (grasses) are at the same time very frequent.
Fig. 1 – Pollendiagram of the sequence in sector A12 of the Telatmachay rockshelter (for locality in excavation, see Fig. 3).

Fig. 2 Pollenspectra from different sectors and layers of the Telarmachay rockshelter.

Fig. 3 – Plan of the excavation (each sector is 1m x 1m) with pollen sample numbers.

29It seems therefore that bunch grass puna established near the rockshelter, not very rich in herbs. This might mean a rise of the lower limit of the subnival zone. The climate may have become wetter and/or warmer. As there is a general trend in other places of the N. Andes to a cooler and wetter climate, it may be that a wetter climate caused a return to the conditions before zone IIb. It is most probable, however, that the vegetation is now strongly changed because of intensive grazing, leading to the predominance of bunch grasses and the drying of the soil in the rockshelter, making a sure climatic interpretation difficult.
30This for the time seems to be the most logical interpretation, if we take into account paleoclimatological data from other places in the Northern Andes. It will be clear, however, that this interpretation is partly tentative, that the interpretation of a warmer zone II cannot entirely be ruled out, and that only the analysis of sequences of lake sediments from the area can give a more definite answer to the questions of the causes of the marked changes of vegetation reflected in the rockshelter sediments.
Bibliographie
REFERENCES
DOLLFUS, O. (1965): Les Andes centrales du Pérou et leurs piémonts. Trav. Inst. Franc. Etudes Andines 10: 1-104.
ELLENBERG. H. (1979): Man’s influence on tropical mountain ecosystems in South America. Journ. Ecol. 67: 401-416.
GIRARD, M. & J. RENAULT-MISKOVSKY (1969): Nouvelles techniques de préparation en Palynologie appliquée à trois sédiments du Quaternaire final de l’Abri Comille (Istres, B.-du-Rh.). Bull. Assoc. Franq. Etude Quatern., n° 4, p. 275-284.
GUTTE, P. & G. GUTTE (1976): Vegetationskundlich-floristische Studien in Peru. Wissenschaftliche Zeitschr. der Karl-Marx-Universität Leipzig 25: 319-346.
LAVALLEE, D. & M. JULIEN (1979) Les sites ancients: l’abri de Telarmachay. In: Mission archéologique française à Junin, Rapport Annuel 1978, Unité de recherches archéologiques N° 25 & Instituí frangais d’études andines: 24-46.
LAVALLE, D. & M. JULIEN (1983): Programme Junin. In: Rapport d’Activité 1980-1983. Unité de recherches archéologiques N° 25, Préhistoire des régions Andines. C. N. R. S., Centre de recherches Archéologiques: 15-31.
MARKGRAF, V. & J. P. BRADBURY (1982): Holocene climatic history of South America. In: J. Mangerud, H. J. B. Birks & K. D. Jäger (eds.), Chronostratigraphic subdivision of the Holocene. Striae, Vol. 16: 40-45. Uppsala.
SITTLER, Cl. (1955): Méthodes et techniques physicochimiques de préparation des sédi-ments en vue de leur analyse pollinique. Revne Inst. Franq. Pétrole et Ann. Combustibles liquides, Vol. X, n° 2, Fév. 1955: 103-114.
VAN DER HAMMEN, T. (1974): The Pleistocene changes in vegetation and climate in tropical South America. Journ. Biogeography, 1: 3-26.
WEBERBAUER, A. (1911): Die Pflanzenwelt der peruanischen Anden. Die Vegetation der Erde 12 (Leipzig) (reprint 1976, Gantner Verlag, Vaduz), 359 p.
WRIGHT, H. E. (1980): Environmental history of the Junin Plain and the nearby mountains. In: Rick, J. W., Prehistoric hunters of the high Andes. Academic Press: 253-256.
Notes de bas de page
Auteurs
Hugo de Vries Laboratorium, Universiteit van Amsterdam, Pays-Bas.
Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
Sociedad y gobierno episcopal
Las visitas del obispo Manuel de Mollinedo y Angulo (Cuzco, 1674-1694)
Pedro Guibovich Pérez et Luis Eduardo Wuffarden
2008
Indigenismo y nación
Los retos a la representación de la subalternidad aymara y quechua en el Boletín Titikaka (1926-1930)
Ulises Juan Zevallos Aguilar
2002
Los Andes y el reto del espacio mundo
Homenaje a Olivier Dollfus
Jean-Paul Deler et Évelyne Mesclier (éd.)
2004
Memorias en conflicto
Aspectos de la violencia política contemporánea
Raynald Belay, Jorge Bracamonte, Carlos Iván Degregori et al. (éd.)
2004
De los Andes hasta Pará
Ecuador - Perú - Amazonas
Marcel Monnier Edgardo Rivera Martínez (trad.)
2005
Del trono a la guillotina
El impacto de la Revolución Francesa en el Perú (1789-1808)
Claudia Rosas Lauro
2006
Ladrones de sombra
El universo religioso de los pastores del Ausangate (Andes surperuanos)
Xavier Ricard Lanata
2007