Version classiqueVersion mobile

Les établissements Asto à l’époque préhispanique

 | 
Danièle Lavallée
, 
Michèle Julien

Etudes complémentaires

Archaeological pollen analysis of sediment samples from asto village sites

James Schoenwetter

Texte intégral

Introduction

1Pollen analysis is a paleobotanical technique which has a number of archaeological applications. Traditionally, pollen analysis is utilized to date Quaternary sediments to time horizons which are identified by characteristic pollen frequencies. If the sediments derive from an archaeological context, or if they can be correlated with an archaeological context in some way, a “pollen date” may be applied to the archaeological record. Another traditional usage is the reconstruction of past vegetational patterns, and paleo-climates. Pollen analysis is also useful in the isolation and recovery of microfossil plant remains extracted by prehistoric men from their local environments.

2Pollen analysis is a statistical technique, since it deals not only with the variety of pollen types found in a deposit but also with the frequencies of each pollen type observed. Because pollen analysis is a statistical technique the analytic procedures used must be consistent with sampling parameters imposed by the problem under investigation. Different problems may be explored by the same statistical approach if the mathematical constraints of that approach are satisfied; but all problems are not amenable to investigation by the same battery of statistical procedures. In some cases specific problems should be approached through creation of uniquely appropriate procedures.

3When confronted with a problem, the pollen analyst must first determine whether pollen exists in the sediment samples which provide qualitative date pertinent to the problem's solution. For example, if the problem is the investigation of the cultivation practices of an ancient society, the pollen analyst must determine if pollen from cultivated plants exists in the sediment, or if frequency variations in the pollen of uncultivated plants exist which indicate human activities attributable to cultivation. The pollen analyst also must establish an analytic framework which may be effectively applied to the problem at hand. An analytic framework pertinent to the establishment of a pollen date need not be effective for determining the character of cultivation practices. All pollen analyses, then, are not alike even if they deal with the same kinds of pollen. As the problem orientations vary, appropriate analytical procedures also vary. Unique analytical formats must be generated for the investigation and solution of unique problems.

4Archaeological pollen analysis (Schoenwetter, 1970:35-36) poses particular difficulties in this regard. Archaeological pollen analysis is both a paleobotanical and an archaeological procedure. Pollen extractive and identification techniques developed by paleobotanists are used, but the problem orientation is a cultural, rather than a strictly botanical one. The framework of analysis must be appropriate to the resolution of the cultural problem. An archaeological pollen analysis cannot be evaluated by reference to botanical and geochronological standards traditionally applied to pollen studies in archaeoly; it must be evaluated in terms of its power in resolving the cultural problem towards which it is directed.

5The archaeological problem investigated by this study was concerned with certain protohistoric village sites occupied during the xvith century, in the Central Highlands of Peru. A number of these village sites are found at higher elevations than contemporary villages located in the same area. Further, some of the sites occur above the elevation at which agriculture is practiced today. The objective of the pollen study was to determine whether the agricultural potential of the zone in which the high elevation villages were situated might not have been greater during the xvith century than it is today. If the agricultural potential was greater, one contributing factor leading to the abandonment of the villages could be the environmental change resulting in essentially modern conditions of agriculcutal potential.

6This study was based upon the assumption that a palynological index of agricultural potential exists. Whatever the climatic, edaphic and biotic variables may be which combine to produce the agricultural potential of a plot of land in this dictrict, it is presumed that the combination in some way affects the pollen types characteristic to that plot, and/or the frequencies of those pollen types. Further, it was assumed that this index of agricultural potential would be revealed by palynological variations which were functions of the elevation of plots in the district. Since the average length of growing season and average temperature decrease with elevation in this area, it was granted that high elevation plots have a lower agricultural potential than low elevation plots. Palynological distinctions between high and low elevation plots would thus serve as an index of agricultural potential.

7In statistical terms, the problem posed by the research objective is one of paired comparisons. Simply stated, we would expect that if the modem environment of the villages is like the ancient environment, the two will be comparable in their production of both kinds and frequencies of pollen grains. If one environment is distinct from the other, either the types of pollen produced today by this environment will be distinct from the types produced in the past, or the pollen frequencies will be distinct, or both the qualitative and quantitative data will be at variance.

8“Environment,” however, is a very generalized term. We recognize both natural and cultural environments; either or both of which may affect the types and frequencies preserved in the pollen record. It is necessary to devise means for discriminating these two types of environments palynologically in order to assess the agricultural potential expressed by the archaeological pollen samples.

9The research strategy involved five discrete procedures:

  1. Collection of samples from archaeological contexts at sites of different elevation;
  2. Collection of samples of surficial sediment at different elevations from two cultural contexts: (a) villages and (b) cultivated fields;
  3. Collection of samples of surficial sediment from culturally undisturbed botanical environments at different elevations;
  4. Analysis of surficial sediment samples to separate culturally caused pollen variation from pollen variation caused by elevation; and
  5. Analysis of archaeological samples.

Methods

10Surficial sediment samples were collected by the “pinch” sampling procedure described by Dittert and Wendorf (163:9-12). This procedure has been widely utilized for obtaining pollen spectra in modern arid land environments (Hevly, Mehringer and Yokum, 1965; Mehringer, 1967; Schoen-wetter, 1967; Schoenwetter & Doerschlag, 1971) and has been demonstrated to be the procedure least likely to be biased by error due to local pollen over-representation (Hevly, 1964; 1968). Fossil pollen sample collections were from sediments directly associated with cultural features such as houses, hearths, pits and terrace surfaces upon which houses had been built. Such sediments undoubtedly were deposited during the occupation of the village sites and thus must date to the xvith century.

11The method used for extracting pollen from both the surficial and the fossil samples was that described by Mehringer (1967: 136) with slight modification of the HF treatment. This technique concentrates pollen from at large volume of sediment (ca. 125 cc) through gravity separation of the light (polliniferous) sediment fraction. Thirteen surface samples and eleven fossil samples were processed. Nine of the surface and all of the fossil, samples were then submitted to microscopic examination. Six of the nine surface samples observed yielded sufficient pollen for analysis. Only four of the eleven fossil samples yielded sufficient pollen.

12The low proportion of productive samples was not unexpected. Pollen preservation problems are extreme in the sediment types collected, and the amount of pollen per cubic centimeter in such deposits is normally small. Terrestrial deposits are, for this reason, not normally investigated by palynologists (Dimbleby, 1969: 171-2). However, since prehistoric man normally lived upon the dry terrestrial surface rather than upon the wet surfaces which tend to preserve pollen grains in large numbers, archaeological pollen analysis must accept the difficulties inherent in the investigation of terrestrial deposits.

13Observation of the fossil samples was undertaken first. One hundred pollen grains from each sample were identified and the frequency of the various types in this count was recorded (Table 1, lower portion). This tabulation indicated that significant pollen frequency variations occurred within the population of fossil samples. In this case statistically significant difference was assessed at the 95 percent confidence level by reference to a confidence limits table for the binomial distribution. Observation of the surface pollen samples was undertaken next, again involving a one hundred grain count. Larger pollen counts were considered unnecessary at this stage, since it had already been demonstrated that a count of this size was sufficient to indicate variation in the fossil record.

14Variation observed in the surface pollen records (Table 1, upper portion) could be attributed to three causes: (1) statistically significant variation in the proportion of Compositae pollen occurred as a function of elevation: higher elevation pollen records incorporated progressively less Compositae pollen; lower elevation records incorporated progressively more Compositae pollen. (2) Significant variation in the proportion of Chenopodinnae pollen occurred as a function of human disturbance of the sampled environment: cultivated field and village surface pollen records produced statistically significant quantities of Chenopodinnae pollen, while “natural” botanical environments produced little or no such pollen. (3) Significant variation in the pollen category labelled “Other Types” occurred as a function of elevation, but in a pattern distinct from that expressed by Compositae pollen: higher frequencies of “Other Types” pollen were observed at elevations between 3700 and 4250 meters than were observed at elevations above or below these values.

15Two of these patterns of variation (that expressed by Compositae pollen and that expressed by Chenopodinnae pollen) were identical to variation patterns observed in the fossil pollen series. Since variation expressed by Chenopodinnae pollen appears to be a function of human disturbance of the environment, and since such disturbance is expected within the confines of an archaeological site, the fossil Chenopodinnae pollen record is not informative for the purpose of this investigation. Further, because the pollen record is expressed in terms of the relative frequencies of the pollen observed, incorporation of the values for Chenopodinnae pollen in the pollen sum (i.e. the total upon which frequency values are calculated) influences frequencies for all other pollen types in a given spectrum.

16Because of this, both surface and fossil pollen samples were reob-served to generate pollen sums of 100 grains exclusive of Chenopodinnae and cultigen pollen. The pollen spectra so produced (Figure 1) are appropriate for assessing the character of plant environments in the district which are exclusively functions of elevation, unbiased by factors of human disturbance.

17It should perhaps be pointed out that the pollen analytic principle subtending such an adjustment of the pollen sum is not an unusual one nor is it unique to this study. It is championed by Faegri and Iverson (1964:83) as critical to pollen analysis, and it is normally employed in European pollen studies through exclusion of aquatic and/or bog plant pollen types as an aid to forest type reconstruction. Adjustment of the pollen sum is not a device to confuse or deceive the unwary. It is a procedure which takes advantage of the quantitative character of pollen data to provide a firmer quantitative basis for interpretation.

Results

18Table I gives results obtained in the pollen analyses based on the 100-grain total count. Table II gives results obtained in the pollen analyses based on the 100-grain pollen sum exclusive of Chenopodinnae and Zea pollen. Figure I illustrates Table II data in a fashion which emphasizes patterned sample variation meaningful to the archaeological problem of agricultural potential.

19The identification of some of the pollen observed is ambiguous (exemplified by cf. Leguminosae), but the greater number of grains could be identified to a useful taxonomic level. Since the number of analyzable samples is small, more exact taxonomic identification would not lead to more precise paleoecological reconstruction and was considered unnecessary. Some of the pollen observed (Pinaceae, cf. Ulmus, cf. Quercus) probably represents long distance transport or sample contamination. The relative frequency of such pollen, however, is too low to affect the conclusions of this study.

20The distinction in total Compositae pollen frequency for surface samples collected at the extreme high and extreme low elevational ranges (4450 and 3550 M) is dramatically illustrated in Figure I. These frequencies are statistically distinct at the 95% level of confidence according to the test utilized in this analysis. The Compositae pollen frequencies of the 3450 and 4450 M samples are, in fact, significantly different from those of any other surface sample analyzed. However, the Compositae pollen frequencies of the other samples are not significantly different statistically from each other, only so may be considered representatives of the same Compositae pollen population.

21Fossil Compositae pollen sample frequencies also relate to three segregate Compositae pollen frequency populations. The sample from Chunlamarka pertains to the same population as the 4450 M surface sample but is unlike any other sample. The samples from Waraqolaqay and Shanki pertain to the same population as the 3700 - 4250 M surface samples but are unlike any other sample. The sample from Laiwe is drawn from the same population as the 3450 M surface sample but is unlike any other sample.

22The segregate Compositae frequency populations appear to be paly-nological reflections of ecological variation due to elevational position. With the surface samples, at least, there is an inverse correlation between Compositae pollen frequency and altitude. Too few samples exist of each known botanical environment to demonstrate that the type of flora native to an elevational position is responsible for the Compositae pollen frequency characteristic of that altitude. But the frequency of Compositae plants and the frequency of Compositae species does decrease with altitude in this district, and it may be confidently suggested that the surface pollen records are largely controlled by this factor. In any case, the segregate Compositae pollen frequency populations appear to be a reliable index to conditions of growing season and average temperature. As such they may serve as an index of agricultural potential. The three Compositae pollen frequency populations of the surface samples, then, may be considered indicative of relatively low, relatively moderate and relatively higher sub-districts of agricultural potential occurring within this one thousand meter elevational range in the district today.

23Four surface samples are available which represent the Compositae pollen frequency indicative of relatively moderate agricultural potential. Three of the four contain significantly more pollen of the AP + Other Types category. One of these three also contains significantly more Ephedra pollen than occurs in any other sample. The three samples which are thus distinguished ostensibly reflect some ecological condition existing today between 3900 and 4250 M elevation. The exact nature of this ecological condition remains unknown, and could only be isolated by further observation and sampling. But since this elevational range represents the agricultural frontier zone today, these pollen frequencies would serve to identify a zone of limited agricultural potential.

24None of the fossil pollen records provide Ephedra or AP + Other Types pollen frequencies comparable to those of the surface samples from the 3900-4250 M elevation. Therefore, none of the fossil pollen spectra pertain to this particular statistical population.

Interpretations

25The basic assumption of paleoecological research (Odum, 1959:96) is that the present is the key to the past. Ecological variations which are expressed by specific biological phenomena today (e.g. the Compositae pollen sum frequency) are assumed to have been expressed by similar biological phenomena in the past. A comparison of modern and fossil pollen records, then, allows interpretation of the fossil record as equivalent to, or distinct from, the ecological variations expressed by the modern record.

26It seems clear, then, that the Compositae pollen frequency of the surface samples can be recognized as an index of agricultural potential within the 3450-4450 M elevational range. Applying this index to the fossil record, it would appear that the sample from Chuntamarka reflects a condition of low agricultural potential, more or less equivalent to that found today at 4450 M elevation. Chuntamarka was more probably a site where economic patterns were dependent on pastoralism rather than one where an agricultural economy was emphasized.

27The sites of Waraqolaqay and Shanki appear to reflect conditions of moderate agricultural potential. Further, as the frequency of AP + Other Types pollen is low, these sites would have been situated in zones characterized by conditions more favorable to agriculture than exist today between 3900 and 4250 M elevation. While the site of Laiwe is located today at an elevation characterized by moderate conditions of agricultural potential, it appears that it enjoyed conditions of higher agricultural potential in the past. Waraqolaqay, Shanki and Laiwe thus appear to represent sites where an agricultural economy would have been practical and productive during the period of occupation.

28If agricultural potential as expressed by Compositae pollen frequency, is actually a function of growing season and temperature, the fossil pollen record indicates that longer growing seasons were present at somewhat higher elevations in this district during the xvith century. It seems likely, then, that the agricultural frontier existing at that time occurred at a somewhat higher average elevation, and was constrained within narrower elevational limits, than is the case today.

29I estimate that the agricultural frontier during that period of the xvith century represented by the fossil pollen spectra was located normally between 4200 and 4300 M. If this is true, a substantially greater amount of agriculturally useful land would have been available to the occupants of the district in the past than is available to the present day inhabitants. Finally, it would appear that lands located in the agricultural zone lying between 3700 and 4200 M had a greater agricultural potential in the xvith century than exists at these elevations today. The agricultural potential of lands lying between 3700 and 4200 M 300 years ago seems to have been on the order of the agricultural potential of lands which today occur at elevations between 3400 and 3700 M.

Table I. 100-grain pollen counts of the analyzable samples.

Table II. Pollen observed in analyses providing a 100-grain pollen sun wich excludes Chenopodinnae and Zea.

Bibliographie

Bibliography

Dimbleby, G. W.

1969 Pollen Analysis. In Brothwell, D. and E. Higgs (eds.) Science in Archaeology. Thames and Hudson, London.

Dittert, A.E. and F. Wendorf

1963 Procedural Manual for Archaeological Field Research Projects-of the Museum of New Mexico. Museum of New Mexico Papers in Anthropology N° 12.

Faegri, Knut and J. Iverson

1964 Textbook of Pollen Analysis. Hafner Publ. Co. New York.

Hevly, Richard H.

1964 Pollen Analysis of Quaternary Archaeological and Lacustrine Sediments From the Colorado Plateau. Unpubl. Ph. D. Diss. Botany, University of Arizona, Tucson.

1968 Studies of the Modern Pollen Rain in Northern Arizona. Jour. Arizona Acad. Sci. 5: 116-127.

Hevly, R.H., P.J. Mehringer, Jr. and H. Yokum.

1965 Modern Pollen Rain in the Sonorem Desert. Jour. Ariz. Acad. Sci. 3-:-123-135.

Mehringer, Peter J., Jr.

1967 Pollen Analysis of the Tule Springs Area, Nevada. In H.M. Wormington and D. Ellis (eds.), Pleistocene Studies in Southern: Nevada. Nevada State Museum Anth. Papers N° 13.

Odum, Eugene P.

1959 Fundamentals of Ecology. Saunders Co. Philadelphia.

Schoenwetter, James

1967 Pollen Survey of the Chuska Valley. In Harris, A. H., J. Schoenwetter and A.H. Warren. An Archaeological Survey of the Chuska Valley and the Chaco Plateau, New Mexico Part 1: Natural Science Studies. Museum of New Mexico Research Records N° 4.

1970 Anchaeological Pollen Studies of the Colorado Plateau. American Antiquity 35: 35-48.

Schoenwetter, James and L. Doerschlag

1971 Surficial Pollen Records from Central Arizona, 1: Sonoran Desert Scrub. Jour, of Ariz. Acad. Sci. 6:216-221.

Table des illustrations

URL http://books.openedition.org/ifea/docannexe/image/1360/img-1.jpg
Fichier image/jpeg, 173k
Légende Table I. 100-grain pollen counts of the analyzable samples.
URL http://books.openedition.org/ifea/docannexe/image/1360/img-2.jpg
Fichier image/jpeg, 502k
Légende Table II. Pollen observed in analyses providing a 100-grain pollen sun wich excludes Chenopodinnae and Zea.
URL http://books.openedition.org/ifea/docannexe/image/1360/img-3.jpg
Fichier image/jpeg, 499k

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.

Acheter

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search