The Vegetation of Greece During Prehistoric Times: The Palynological Evidence

These diagrams indicate that Greece was covered by a steppe-like vegetation, with grasses, chenopods and mugwort during the last glacial period and after about  8000 B.C. with forests. These forests showed regional and altitudinal variation, being predominantly of oak on the lowlands. They were eventually destroyed by man, a process which began in the south as early as Final Neolithic times. Elsewhere, although man's influence could be detected, the woodland remained virtually intact throughout prehistoric times, even in areas where there is abundant archaeological evidence for settlement.

The Bronze Age appears to have been a time of maximum impact on the vegetation by man. Pollen grains of olive, vine, walnut and chestnut have been found at several sites, indicating that these plants were grown locally.

There are no pollen diagrams from the Aegean islands, and so the vegetation of Thera can only be deduced by comparison with that of the southern mainland.

INTRODUCTION*

*All radiocarbon dates are uncalibrated and are quoted as years b.c. or a.d., to indicate this.

The science of palynology, or pollen analysis, is concerned with reconstructing earlier vegetation and floras from the microscopic pollen grains and spores which are preserved in peat and lake sediments. Such grains are diagnostic of the plant that produced them at either the species, genus or family level and, in suitable deposits, occur in high enough concentrations to allow quantitative analyses of how their frequencies have changed through time. Provided the sediments can be dated by their geological, archaeological or historic associations, or by an independent method such as radiocarbon assay, vegetation change during successive prehistoric periods can be traced.

Many such studies have been carried out in northern Europe since the 1930s, but it was not until 1965 that the first results were available from Greece. This was largely because there are fewer pollen preserving deposits in countries bordering the Mediterranean. I do not want to go into the difficulties associated with pollen analysis in Greece - all research has its problems - but I do want to make it clear that they are such that a palynologist is without the luxury afforded him in wetter and cooler climates, of being able to select the best from several possible deposits, all in the region and covering the period which interests him. In Greece, he is thankful for any data he can obtain, whatever the region, whatever the period. Naturally I would like to be able to talk about the vegetation of Thera, or even the Aegean, on the basis of several pollen diagrams from local deposits securely radiocarbon dated to the Bronze Age, but that is not possible at present.

However, during the last twelve years, a number of pollen diagrams from other parts of Greece have been published, several of them with associated radiocarbon dates, and they do provide information of interest to archaeologists. The major conclusions of these projects concerning the vegetation of Greece during all prehistoric periods, including the Bronze Age, will be presented in this paper.

The first section deals with the nature of the early post-glacial forests and subsequent sections outline region by region the evidence for man's influence upon the vegetation from the Neolithic onwards. Finally, in the light of the evidence presented, the probable nature of Thera's vegetation during prehistoric times will be discussed.

THE EARLY POSTGLACIAL FOREST OF GREECE

The first pollen diagram to be published (van der Hammen 1965) showed that after a long cold period, which was the equivalent of the Würm glaciation of the Alps, Greece was largely forested. The diagram comes from Tenagi Philippon near Philippi in Macedonia (see fig. 1) where, during many tens of thousands of years, over sixty-two metres of peat had formed as a result of long term local tectonic subsidence. Most of the pollen in the top five metres of the sediments, equivalent to the last 10,000 years, belongs to tree genera such as oak (Quercus), pine (Pinus), hazel (Corylus), ash (Fraxinus), elm (Ulmus), birch (Betula), lime (Tilia), hornbeam (Carpinus) and Ostrya, with oak pollen predominating (Wijmstra 1969).

Below five metres, that is, prior to 10,000 years ago, the pollen grains are quite different. There are only minimal amounts of tree pollen but large amounts of the pollen of grasses (Gramineae), chenopods (Chenopodiaceae) and mugwort (Artemisia). This pollen represents an open steppe-like vegetation, and it is clear from the diagram that such vegetation dominated the landscape throughout much of the Würm glaciation.

The effective catchment area of this pollen is only a small part of Macedonia, but palynologists have found similar evidence from other parts of Greece. For example Bottema (1974) has recorded this change from an open steppe-like vegetation to oak forest for the regions around Lake Khimaditis and Lake Ioannina and Edessa in north-west Greece (see fig. 1). At Edessa it is dated to 8695 ± 100 b.c. and at Ioannina to 8240 ± 90 b.c., dates consistent with those for the beginning of the Holocene or post-glacial in north-west Europe. Bottema has also demonstrated the existence of forest around Giannitsa and Kastoria during Holocene times, and Athanasiadis (1975) for the regions around Litochoro Pierias and Pertouli Trikalon, in northern Greece. For southern Greece, results from Kopais, east of Levadia, indicate the same replacement of steppe-like vegetation by oak forests at the end of the Würm (Turner & Greig 1975).

In the first part of the Holocene, oak appears to have been the major tree species in all the areas studied, except at high altitudes, i.e. above about 1400 m., in the north, where coniferous trees, pine and sometimes fir (Abies) were growing.

At lowland northern sites, and at Kopais in the south these early oak woods were open in nature with a grassy ground flora. There were low growing shrubs like junipers (Juniperus), and small trees such as Pistacia terebinthus, as well as a variety of herbs including salad burnet (Sanguisorba minor). Above about 800 m in the north of the country, however, the trees were growing more densely and mixed with elm, lime and hazel.

Later in the Holocene, dated at one site to about 6150 b.c., the open oak wood became more dense all over Greece. North of the Pindus range and in Macedonia elm, lime and hazel began growing at lower altitudes although in the south of the country these closed oak woods appear to have been almost pure oak. At Kopais there are surprisingly few pollen grains of other tree species and one can only conclude that the thermophilous trees such as elm, lime, ash and hazel were not as abundant as they were in the north.

Then, from about 4550 b.c. onwards, in the north, the pollen of Carpinus orientalis/Ostrya type became more abundant and this has been interpreted by Bottema as indicating the spread of Carpinus orientalis. Later still, beech (Fagus) spread in the north, but this and associated changes in the vegetation will be discussed later. No corresponding spread of Carpinus orientalis was recorded at Kopais. The development of these forests during the early Holocene was a natural process. Wijmstra and Bottema both attribute the changes in forest composition they have described, to changes in climate during the first half of the Holocene and I am of the same opinion for the Kopais region. Bottema's diagrams are the most helpful in this connection for they cover a range of altitudes within a region. He suggests that there was initially an increase in temperature which allowed trees to replace the steppe-like vegetation of the glacial period and that there was then an increase in precipitation which allowed the various tree zones to move to lower altitudes. Later still a decrease in precipitation led to a raising of the tree lines.

But from about 3250 b.c. onwards, many of the changes in vegetation that have been recorded have undoubtedly been caused by man and it is these that will now be considered. It is convenient to do so on a regional basis.

NORTH-WEST GREECE

Bottema's diagrams from Edessa, Khimaditis, Kastoria, Giannitsa and Ioannina all show that the areas concerned were well wooded throughout the last half of the Holocene. Nevertheless from the beginning of his pollen assemblage zone Z, which has been dated to 2585 ± 40 b.c. at Ioannina and 2045 ± 60 b.c. at Khimaditis, there are indications that man was beginning to have a noticeable affect on the vegetation in the region north of the Pindus mountains. At first sub-zone Z1, it was very small, a slight opening up of the canopy, perhaps a little agriculture. The pollen diagrams show only very low values for certain weeds, plantain (Plantago) species and thistles (Centaurea solstitialis type). Subsequently. however, in sub-zone Z2, the beginning of which has been dated to 1330 ± 55 b.c. at Edessa, pollen of olive (Olea), chestnut (Castanea), walnut (Juglans) as and also plane (Platanus) occurs in low frequencies and there are increased values for that of the weeds of cultivation. Bottema has also found pollen of vines (Vitis) and pollen of Cannabis /Humulus type (the two are indistinguishable).

All this points towards settled agriculture in the region with chestnuts, walnuts, grapes and olives being harvested.

Grass pollen of the cereal-type has also been recorded, but as wild cereals are native in this part of the world, and since cereal-type pollen has also been found in sediments of early Holocene age, long before there can have been any question of agriculture, it is not possible to argue from the pollen that cereals were being cultivated at the time.

The following sub-zone Z3 is characterised by lower values for the pollen of weeds of cultivation and by increased values for chestnut pollen. Bottema suggests that this represents a period with herding, but no agriculture, and that chestnut was perhaps spreading rapidly in the abandoned clearings.

Renewed human pressure on the woods then occurred (sub-zone Z4) with the re-introduction of agriculture. This has been dated by extrapolation to about  a.d. 270 at Edessa and about a.d. 850 at Khimaditis.

South of the Pindus mountains at Ioannina there are fewer indications of man having affected the natural vegetation during zone Z, although archaeological investigations have demonstrated his presence during both the Neolithic and Bronze Ages.

Athanasiadis' diagrams from Litochoro and Pertouli, also in this part of Greece, have detailed sequences of the vegetation showing man's influence.

At Pertouli, north of the Pindus, there is evidence for cereal cultivation during the last part of the early and the middle Bronze Age, up to about 1600 b.c. A further period of increased cereal cultivation began at about a.d. 900 and it was only then that chestnut spread. After about a.d. 1400 there was an intensification of activity with a new cereal maximum and high values for weeds of cultivation and, for the first time, the tree pollen drops to as little as 16 %, indicating that large areas of woodland were being cleared. The trees recovered quickly, however, and the corresponding decline in weed pollen is maintained to the top of the diagram.

At Litochoro, the sequence is somewhat different. No serious cultural influence was detected until about 900 b.c., when walnut and chestnut spread and when cereal and vine pollen was also recorded. During Roman times and up until the mid β-Byzantium period, about 200 b.c. to a.d. 900, there was a decline in farming activity, followed, as at Pertouli, from about a.d. 900 onwards, by an intensification, although tree pollen values did not decline so significantly indicating that less woodland was being cleared than at Pertouli.

It is clear from this short summary that there has been a certain amount of variation in the vegetational history of north-west Greece. If there are any similarities, it is that there were two periods when human influence was fairly intense. The first was some time within the Bronze Age, although exactly when seems to vary from place to place. The second was towards modern times, at several sites encompassing the last thousand years.

MACEDONIA

Towards the top of Wijmstra's pollen diagram from Tenaghi Philippon there is a slight decrease in tree pollen and the appearance of beech pollen, but none of the other features found by Bottema and Athanasiadis for north-west Greece. As it seemed possible that the most recent Holocene was missing from the top of the diagram, Mr Greig and I decided to take some further cores for analysis from the uppermost sediments, in an attempt to supplement the data for the area.

Our two diagrams indicated that oak forest was dominating the area from the level radiocarbon dated to 3081 ± 180 b.c. right through until at least the time of Christ (extrapolated date), although a slight but steady decrease occurred during the period. There were two shorter periods during these three millenia when olive pollen was recorded at both sites.

The first of these was Bronze Age. It has a radiocarbon date of 1338 ± 100 b.c. at site I and at site III falls between levels dated to 1790 ± 60 and 917 ± 60 b.c. The second can only be dated indirectly. It starts just above and so post-dates the level radiocarbon dated to 917 ± 60 b.c. But like Wijmstra we failed to find an increase in walnut or chestnut or plane, or indeed significant increases in the pollen of weeds of cultivation, although plantain and Centaurea type pollen grains were recorded in low frequencies. We were left with the impression that there had been a densely wooded environment throughout the three millenia before Christ with only the smallest amount of disturbance by man, presumably in the immediate vicinity of his settlements.

The final clearance of the woods has been recorded in a pollen diagram from a small lake near Gravouna to the east of Philippi, but as it was not radiocarbon dated, we can only postulate that it occurred at the beginning of the Turkish occupation which followed the fall of Constantinople.

THE PELOPONNESE

Wright (1972) has studied sediments in two coastal lagoons on the western shores of the Peloponnese, one near Kaiafa, the other at Osmanaga, just north of Pylos.

The diagram from Osmanaga starts at a level radiocarbon dated to 1990 ± 270 b.c.  At that time there were some pine forests, probably the aleppo pine, Pinus halepensis, in the area. They had disappeared by 1270 ± 120 b.c. and Wright attributes this to the need for timber, fuel and agricultural land during the Middle Bronze Age. Above the level dated to 1270 ± 120 b.c. oak pollen amounts to some 20% of the total, indicating a modest amount of oak wood, and olive pollen reaches equally high values, forming two distinct peaks with lower values between.

The high values can only represent extensive olive cultivation.

The middle of the first period of olive cultivation has a radiocarbon date 870 ± 100 b.c. and Wright suggests that the period encompasses the Greek Dark Ages. It is worth noting that he found a similar peak in olive pollen at Voulkaria on the west coast of mainland Greece near Vonitsa, where it was dated to the same period, namely 1090 ± 120 b.c.

The second peak is not so pronounced, nor as securely dated, owing to the very slow accumulation of sediment, but it appears to have occurred some time in the first millenium a.d. Following this second peak the diagram shows a gradual decrease in oak pollen, indicating the progressively greater land clearance towards modern times.

The pollen diagram from the peat marsh at Kaiafa covers only the period from a.d. 300 ± 100 onwards. It gives evidence for a decrease in oak and an increase in pine woods and also for the cultivation of olives throughout.

SOUTHERN MAINLAND GREECE

The pollen diagram from Lake Kopais (Turner and Greig 1975) shows a large decrease in the amount of oak forest at a level radio-carbon dated to 3255 ± 120 b.c. This is associated with the inwash of clay to the lake, the result of soil erosion on the surrounding hills, presumably caused by a combination of activities by prehistoric man. The felling of trees and the cultivation of crops would have rendered the soils more susceptible to erosion than they had been under a stable forest cover.

Lake Kopais has a very large pollen catchment area and the diagram therefore gives a regional, rather than a local, picture. Herbaceous pollen is often under-represented in such diagrams, because it does not travel as far as tree pollen, but, even so, there is some evidence of how the cleared land was subsequently used.

Walnut and plane pollen were not recorded, and only one grain of cf. chestnut. But olive pollen was, at values of up to 10 % in some levels. Also one grain of vine and some of weeds of cultivation, including the cornflower (Centaurea cyanus).

There is no evidence for regeneration of forest at a later date. The land remained largely cleared and farmed.

CONCLUSIONS

Such varied data cannot readily be summarised, and as they constitute a mere handful of jigsaw pieces in a very large puzzle, it would be unwise, at this stage, to attempt too many generalisations. I would, however, like to make three points.

1. The Bronze Age appears to have been a time of maximum human impact on the vegetation. Extensive forest clearance, which had begun as early as Final Neolithic times in the Kopais area, was maintained, and in northern Greece, north-west of the Pindus, there was more human activity within the forests than either before or immediately after.
2. There are some northern areas, such as around Ioannina and Philippi, where, in spite of archaeological evidence for settlement, there is little sign that people were having any impact on the natural vegetation.
3. A difference therefore does seem to be emerging from this limited data between northern and southern Greece. The vegetation of the north, although influenced by man, was basically woodland throughout prehistoric times, whereas in the south extensive tracts of land were being farmed from the Bronze Age onwards.

THE VEGETATION OF THERA

This brings us to the question of the vegetational history of the Aegean islands and whether it resembled that of southern or northern Greece. Certainly, today, Santorini lies in the same climatic region of Greece as Athens and Lake Kopais with a mean annual-temperature of 17° to 18° C and a mean annual rainfall of 350 to 450 mm. This climate differs from that of both the Peloponnese and northern Greece (Walter and Lieth 1967). And so on climatic grounds one would expect the vegetation of Santorini to be, and to have been, similar to that of the Kopais region. And if this is so, then it means that the islands were originally covered with forest, at first open oak woods with juniper and Pistacia and then, still using Kopais as a guide, with more dense oak woods containing few other tree species.

This argument, of course, only applies to the period before prehistoric peoples began to modify the vegetation. At that point, in the absence of local pollen evidence, the palynologist has to turn to the archaeologist and ask - from what you know of the early settlement of the islands in relation to the mainland, what do you think the vegetation was like?

--------------------------------------

---------------------------------------