Effects of Past and Present Fire on the Vegetation of the French Mediterranean Region1 Louis Trabaud2

Since 1960 wildfires have burned an average of 36,000 hectares a year of natural vegetation in southern France. Some years the area burned greatly exceeds this average (e.g. 52,800 hectares in 1976). The most important trends for the last two decades are a steady increase in number of fires, but a relative constancy in the total area burned (Trabaud 1980). The most numerous and most harmful wildfires occur in the Mediterranean zone, an area which stretches from the eastern Pyrenees to the southern Alps, including Corsica. This area is characterized by its geographical position near the Mediterranean sea and by its climate with hot dry summer and mild rainy winters. The long dry summer season, with warm temperatures and low relative humidity, is particularly favourable to fire. Besides, the mediterranean vegetation type is highly flammable and combustible due to the preponderance of species with a high content of resins or essential oils. In addition, the rugged topography of the area increases the ease of fire spread. The presence of millions of people (both residents and tourists) in this region can also contribute to create an "explosive" fire situation under certain conditions. It is apparent that fire is an important past and present ecological force which has played and continues to play a significant role in fashioning, positioning, and developing numerous mediterranean plant communities. Used for a long time by man to reclaim virgin lands or to maintain already burned plots, fire associated with other human activities has had a significant impact on shaping the current landscape of the French Mediterranean region.

Abstract: Fire has shaped the French Mediterranean vegetation for a long time. Man used and abused fire for reclaiming agricultural lands. Today, fire is mainly lit by carelessness. Most plants of the mediterranean vegetation have acquired strategies to withstand and survive fire. After wildfire, vegetation turns back quickly towards its initial stage. By using prescribed burnings according to different frequencies and seasons the structure of a Quercus coccifera garrigue was changed, but not the floristic composition. FIRE HISTORY ELEMENTS Fire during Prehistory Man had probably begun to frequent the French Mediterranean region half a million years ago. At that time he had little impact on the native vegetation as he was primarily fed by hunting animals and gathering fruits and edible roots. As soon as man possessed fire (about 300,000 years B.C.; Perles 1975, 1977) he had many reasons for burning. These have remained essentially the same until the present time. In fact, man has probably voluntarily used fire for hundreds of thousands of years. Without knowing how to create fire, man carefully preserved natural fire whenever possible. Hough (1926) said that the oldest known torch dates from about 100,000 years ago in France. It was not until later (40,000 to 50,000 years B.C.) that prehistoric man learned to create and control fire. First the hunter and the herdsman must have noted that the animals they fed upon were attracted by burned zones, where vegetation appeared greener and softer. Hence, it appears that man has set fire, accidentally through lack of care, but also intentionally, for a great number of purposes. It seems that fire has been used to increase food choice, to facilitate the gathering of edible plants, and to attract and hunt game. Testimonies are available to show that Paleolithic man used this technique in Greece (Liacos 1974), Israel (Naveh 1974), Spain and France (Perles 1977). With the domestication of animals, which took place in the French Mediterranean region about 7,000 years B.C. (Bailloud 1975), man needed rangelands. He burned the forest to be able to see more clearly and protect his domestic animals against their natural predators.

2-Departement d'Ecologie Générale - Centre d'Etudes Phytosociologiques et Ecologiques B.P. 5051 - 34033 - Montpellier Cedex - FRANCE

During the Neolithic (about 5,000 years B.C.), men of this area were primarily field growers and sedentary herdsmen (Pannoux and Pannoux 1957, Dugrand 1964). They occupied fields after partial burning of the forest, supplementing their resources by picking fruit. Beaulieu (1969) interpreted evidence of fire in the pollen diagram of the Font-Salesse peat-bog (Monts de 1'Espinouse) as being from this period, and concluded that it must be ascribed to human intervention. Those farming peoples, or their immediate predecessors, could have engaged in nomadic agriculture, a sort of shifting cultivation such as is still used in some

450

Gen. Tech. Rep. PSW-58. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, U.S. Department of Agriculture; 1982.

1-Presented at the Symposium on Dynamics and Management of Mediterranean-type Ecosystems, June 22-26, 1981, San Diego, California

areas of Africa and Asia where fire is set before or after tree-felling, followed by reclamation and cultivation. Cultivation shifted as soon as production decreased as a result of the soil becoming poorer. The village might also be moved to remain near the arable lands.

Fire has also been used as a technique for comtry around Montpellier was regularly burned to promote proliferation of the kermes oak (Quercus coccifera L.) cochineal bettle (Kermococcus vermilio) in order to obtain from it a red dye for dyeing wool (Delmas 1958).

The successors of the paleolithic hunter-gatherer, the mediterranean shepherds, used and still today are using fire to open the dense thickets of maquis and garrigue to increase the area of rangelands for their herds (goats, sheep and cattle). Thus, Niederlender and others (1966) ascribe the modification of the vegetation of calcareous regions of southern France to man, who set fires repeatedly during the Neolithic.

Fire has also been effectively used as a war weapon for destroying crops. The country to be subdued was ravaged and the inhabitants famished and demoralized by setting afire forests and fields. In 970 A.D., the people, around Nice set fire to the forest to hinder the ambushes of the Saracens (Kuhnholtz-Lordat 1938). In the Middle Ages, the "Grandes Compagnies" and bands of ruffians pillaged and set fire to fields and woods (Fourquin 1975). During the 16th century, great areas of forest were burned as the armies of the constable of France Charles de Bourbon, of Charles V, and of Emmanuel de Savoie passed through. The forests of Mont des Maures were burned during the battles between the armies of Charles V and those of Francois I (Seigue 1972). It was the same during the rebellion of the Camisards in the Cevennes (Le Roy Ladurie 1976).

Fire during historical times During the historical period (which runs from the first Greek factories, 750 years B.C., up to the beginning of the 20th century), that country today occupied by the garrigue underwent a series of changes. Periods of population increase produced new waves of land reclamation, and simultaneous increase in the number and size of herds. There followed both an expansion of cultivated lands at the expense of rangelands and of rangelands at the expense of forests. On the other hand, during periodic disasters, due to wars, invasions, epidemic diseases and economic crises, cultivated lands were abandoned, returning to post-cultural or pre-forested stages (Dugrand 1964). Most fires ignited by man were set for agricultural or pastoral purposes. Due both to poor yields and the need to feed growing numbers of people, man had to increasingly reclaim new cultivated lands. Fire was associated with the "essartage" or "ecobuage". In the first case, trees were first felled, then burned and the land was sown. In the case of "ecobuage" the technique was more complex. The ground was "fleeced" with adapted tools and according to precise methods; afterwards, the clods were dried up, then burned. The ashes were used as fertilizer. These agricultural practices were in use throughout the French Mediterranean region at least from the Middle Ages through the end of the 19th century (Ribbe 1865, Kuhnholtz-Lordat 1938, 1958, Sigaut 1975). Ecobuage was mainly employed in Languedoc and Roussillon where, during the 18th century, authorities intervened to limit cultivation on burned lands. It was between 1820 and 1830 that the largest areas were burned by ecobuage in southern France (Sigaut 1975). Essartage was still used in 1870 in crystalline Provence (Les Maures, 1'Esterel) and in the Montagne Noire. In Provence the forests of maritime pine (Pinus pinaster Soland) were burned. The burned pines were then cut and sold at Marseille and the cleared lands cultivated (Kuhnholtz-Lordat 1958). In Corsica that technique was still in use in 1930 to create arable fields (Sigaut 1975).

In later periods (towards the middle of the 19th century, Sigaut 1975), with the emergence of new technologies, fire has been less and less used to reclaim lands. Only shepherds continued to set fire, primarily through tradition to open rangelands on abandoned fields that had become increasingly invaded by shrubby vegetation. Fire was required to promote the growth of new shoots of palatable species. As Le Roy Ladurie (1976) states, burning off pasturelands in Languedoc was the only way used during the 17th century to improve feed for animals. Animal husbandry was regulated by "transhumance" (travelling into the mountains to graze in summer and coming back down to the plains in winter ; transhumance was documented 1,000 years B.C., Bailloud 1975). Shepherds burned every four or five years either in February or March, before the growth of herbaceous species, or at the beginning of autumn before the flocks come down from transhumance.

Fire today While the causes of burning have shifted over the years the French Mediterranean region has continued to be fired. It is difficult to know the importance of the lightning fires in the past because we have no precise records, but today in southern France only 1 percent of the wildfires are caused by lightning. At present, flocks of sheep, although less numerous, continue to graze the garrigue rangelands. But the origin of fire is no longer the exclusive liability of shepherds. The town dwellers, eager for calm and clean air and wishing to "come back to nature", are building homes in increasing numbers in the country. Large numbers of people who now travel through southern France are tourists or

451

holiday-makers. They create a new danger. As a consequence of current ways of life (homes in the woods, barbecues, hikers) fire risks are increasing. Besides, people living or travelling through these zones which are susceptible to fire are often ignorant of the chance they take or the imprudence they create for other people. Today the areas burned by wildfires change according to the vagaries of meteorological conditions. In the area called "the red belt" in Provence and Corsica (Southeastern France) each piece of natural vegetation would be burned, on an average, once every 25 years (Seigue 1972). This fact clearly shows the problem of vegetation fire in southern France, but also the importance of fire as an ecological force shaping the landscape and the plant communities. This emphasizes the necessity to obtain a scientific knowledge of the effects of fire.

ECOLOGICAL EFFECTS OF FOREST FIRES In France, to date, only a few authors have studied the role of fire as a factor in succession (Braun-Blanquet 1936, Laurent 1937, KuhnholtzLordat 1938, 1958 , Kornas 1958, and Barry 1960). Generally these authors describe stages corresponding to vegetation types ; they give some information about succession following fire, and compare the different stages to each other only in a synchronic way. Precise diachronic observations are very rare on the actual succession of plant communities after fire. There are also few available data about the ecological consequences of fire.

Effects of fire on mediterranean plants Plants react to fire in very different ways. They possess numerous fire-related adaptations. Nobody has studied in a precise manner the problem of the relative resistance of species to fire, probably because of the numerous factors which are acting. For any given species there is a range of fire-resistance possibilities which vary according to fire intensity. These possibilities may vary with growing season and maturity. For example, observations show that winter or spring fires do not harm the subsequent development of sprouts of Quercus ilex L. (holly oak) or Q. coccifera L. (kermes scrub oak ; Trabaud 1970, 1974, 1980), while fires in summer or autumn, which are more intense, decrease sprouting ability. The plants of sclerophyllous communities in the mediterranean region are adapted to survive after the passage of fires, particularly light fires, where the plant cover quickly builds up again (Le Houerou 1974, Naveh 1974, Trabaud 1974, 1980). Among the fire adaptations observed are plants with little or no bark, but which can survive owing to subterranean organs (bulbs, rhizomes, and tubers), geophytes such as Asphodelus cerasifer Gay (Trabaud 1974) and Ophrys sp.p (Naveh 1974,

452

1975), and grasses such as Brachypodium ramosum (L.) R. et S. (Trabaud 1974). The latter species shows a noteworthy indifference towards burning dates and fire frequency. Some species possess strong vegetative subterranean systems, like Quercus coccifera (Trabaud 1974, 1980). Other kinds of vegetation shaped by fire include tall maquis dominated by Q. ilex, and Ericaceae including Arbustus unedo L., Erica arborea and E. scoparia L. All these species produce vigorous stump sprouts; consequently they dominate the burned forests or maquis (Le Houerou 1974, Naveh 1974, 1975). Some trees, such as Quercus suber L. (cork oak), have relatively thick bark permitting them to resist fire. According to Le Houerou (1974) cork oak forests would be a stage in fire succession of the Q. faginea Lamk. forests in North Africa, Portugal and Spain, where the latter oak has been eliminated by burning. In France, cork oak would be only a vicariant of holly oak on warmer and moister siliceous soils. When fire is neither very frequent nor very intense, open forests of Q. suber maintain themselves. However, despite the role of bark thickness, many species with thin bark, such as Q. ilex, withstand fire very well. This species sprouts abundantly from stumps, but when fires are not too intense it can produce sprouts from epicormic buds. Some species do not withstand fire directly but, owing to their ability to disseminate numerous seeds, they can survive fire and also colonize large areas. Such species include Cistus sp.p. (Trabaud 1970). Le Houérou (1974) quotes 12 species of mediterranean Cistus considered as typical active pyrophytes propagating only by seed. Another example is that of coniferous species, which are among the most flammable ecosystems in the region. Pinus halepensis Mill. (Aleppo pine) and P. pinaster Soland. (maritime pine) constitute huge forests in the Mediterranean region. They can regenerate only by seed, disseminated by cones which often burst during fires (Le Houérou 1974, Trabaud 1970, 1980). Generally after fire the pine forests are colonized by a large number of pine seedlings which produce a new pine forest, with no change in the floristic composition. Since seeds of these two Pines germinate on mineral soil in sunny places, where competition is weakest and where mineral nutrients are abundant, the pine forests may perpetuate themselves indefinitely.

Ecological effects of fire on plant communities of Southern France After fire the vegetation of the garrigue returns quickly towards its initial state (Trabaud 1974, 1977, 1980, Trabaud and Lepart 1980). Most often the species which are present 12 years after fire are the first ones to appear immediately following fire and become more and more numerous through the years. To study this, 47 observation plots were established in areas burned by wildfires in Bas-Languedoc. These were located in 8 types of plant communities representative of the area. One year after fire 70 percent of the study plots

the period of observation (Trabaud and Lepart 1980). The fugacity of any species is measured by the number of observations in which it is missing. The fugacity index can be considered as a measure of floristic stability: if the index is high, the community has not reached a stable state, if it equals zero the community is floristically stable. Fugacity is low immediately after fire and reaches its maximal value during the first and second year, when floristic richness is also at its maximum. Therefore the richness of the intermediate stages results from shortlived species which are progressively eliminated during vegetation recovery. Most often these species are therophytes which do not belong to the communities.

Figure 1 -- Floristic richness of dense Quercus ilex coppices. possessed more than 75 percent of the species which are present 10 or 12 years later. Two years after fire, this percentage was over 80 percent; and in five years it reached 100 percent. The reversion towards a metastable state, at least for the time considered, was quickly accomplished. This study of the development of floristic richness following fire shows that the different communities follow a highly general model (fig. 1). During the first months after fire there were relatively few species. Floristic richness reaches a maximum between the first and the third year after fire, and finally tends to stabilize after the fifth year. The generally higher number of species during the first three years can be attributed to the fire-induced opening of vegetation cover, the disappearance of litter, and to the richness in nutrients of the upper soil. These conditions favour the establishment of exogenous species which will later disappear as plant cover closes up. The index of species fugacity is similar to the model of floristic richness (fig. 2). A "fugacious" species is defined as one that does not survive

Actually, after fire, there is no real succession in the sense of substitution of species or communities, but only a progressive return towards a stage similar to the one before fire. Most often the species of the "terminal" community (defined by the latest dated observation) appear long before the alien species. Floristic composition of present communities is changed very little by fire. The return towards a metastable state occurs quickly and the floristic diversity of the landscape does not seem to be affected. At least for the length of time of this study and for the plant communities in Bas-Languedoc there is no convergence towards similar vegetation types. Each community keeps its own characteristics. This result is also confirmed by the structural development of the communities. The passage of fire has not only an effect upon floristic composition but it modifies the plant arrangement and phytomass. After fire and through succeeding years, stand strata get more and more complex and vegetation tends to grow up from the lower layer to the upper layers with a progressive multiplicity of strata similar to those of origin (fig. 3): the higher strata appear later in succession. These trends were quantified by recording the number of hits on plants at various layers by a needle dropped along a transect line. During the 12 years of observation only Quercus ilex coppices reach the fifth layer (from 2 to 4 m). Pinus halepensis stands, though belonging to forest communities generally do not grow beyond the third layer (1 m). This difference is due to the kind of survival strategy used by the dominant species of each community to regenerate after fire. Q. ilex sprouts vigorously from stumps and rapidly grows reaching heights 'of 2 m in 70 months, while P. halepensis can only regenerate by seed and thus shows a slower growth following fire (80 cm in 80 months).

Changes in Q. coccifera garrigue with different prescribed fire regimes

Figure 2 -- Development of the fugacity index: dense Quercus ilex coppices

The study of the development of post-wildfire vegetation showed that plant communities quickly recover, and that after about 10 years they are not much different from those which existed before the fire. To analyze more precisely and to better understand the effects of fire on plants and vegetation, it is necessary to set up experimentation 453

with prescribed fires. Only an experimentation on vegetation stands identified and studied before burning will give reliable results which can be compared with observations from plots burned by wildfires. Such experimentation was set up in a Quercus coccifera garrigue 10 km north of Montpellier. The objective was to analyze: 1) the development of vegetation with regards to the initial pre-fire state to determine the fire resistance of species; 2) the impact on vegetation of different frequencies of repeated burns (a burn every six years, a burn every three years, and a burn every two years); 3) the effects on vegetation of the time (season) of burning, to determine how seasonal conditions, which affect the phenological behavior of species, relate to fire effects. The presentation of the experimental techniques and partial results were published previously by Trabaud (1974, 1977, 1980). Thus only a brief summary of effects on floristic composition and changes in vegetation structure will be discussed in the present paper. The combination of burning frequency and burning season gives six treatments: 6 P - vegetation burned every six years in spring 3 P - vegetation burned every three years in spring 2 P - vegetation burned every two years in spring 6 A - vegetation burned every six years in autumn 3 A - vegetation burned every three years in autumn 2 A - vegetation burned every two years in autumn The unburnt vegetation was considered as a control (T). All of these treatments have five replicates. As for the study of the development of vegetation structure after wildfire, the importance of vegetation is given by the number of hits on plants of defined layers by a needle. The floristic composition of the Quercus coccifera garrigue is on the whole very stable. After burning it is identical to that which existed at the beginning of the experiment. The dominant and characteristic species are present at all times. The advent or the disappearance of some of them produces only a small floristic change. This relative stability is due to the fact that most of species present in the plots before burning regenerate predominantly by vegetative means (sprouts), while the "invading" species appear only the first year after burning as individuals coming from seeds. They are rapidly eliminated in later years. The influence of season of burning is much more important than burning frequency; with autumn burnings many alien short-lived species appear increasing

454

Figure 3 - Development through time of the number of hits according to the height of strata of the communities burned by wildfires. the floristic richness for a short time whereas other species of the original community disappear. Frequently repeated fires alter the structure of the vegetation. Immediately after each burning there was a decrease in the number of hits (disappearence of above ground vegetation), then a progressive increase, following autumn burns this increase is slower. When burning is only every six years, the vegetation of the Quercus coccifera garrigue reaches a level similar to that of the unburnt garrigue within five year after prescribed fire. On the other hand, burnings regularly set every two years produce a decrease in the total number of hits. This repetition of burns leads to a disappearance of the upper layers (above 50 cm) but it favours an increase in the amount of vegetation in the lower layers (< 25 cm). The decrease in the number of hits is mainly due to the woody plants which have fewer long sprouts bearing fewer leaves. But the number of hits from herbaceous plants is not diminished by successive fires; rather it is increased and particularly so with autumn burnings. The influence of fire on the increment of herbaceous plant phytomass has been recognized for a long time (Ahlgren and Ahlgren 1960, Daubenmire 1968, Vogl 1974). In the unburnt number of hits on herbaceous plants (fig. 4). This is in the spring. In

vegetation, the proportion of the woody plants as compared with does not vary through the years also true for vegetation burned the vegetation regularly burned

Figure 4 -- Changes in the ratio of hits of woody plants to the total vegetation number of hits.

every two years in autumn, this proportion goes down from about 4/5 in 1969, before prescribed burns, to 1/3 in 1975 and again in 1978, or lesser. These results are confirmed by a brief study of above-ground phytomass as a function of the different treatments. In may 1979, for each treatment seven 1 m2 samples were collected from each different plot. Then vegetation was hand-sorted, separating woody from herbaceous plants, dried, and finally weighed. Total above ground phytomass of the unburnt vegetation weighed, on an average, 30 t. -1 -1 -1 (table 1), or a production of 1 t.ha yr ha of dry matter since the last wildfire. This value is similar to that reported already by Long and others (1967) twelve years earlier. Burning frequency has an effect on the amount of vegetation produced. Four years after a second burn (1979; 6P, 6A) total phytomass only represent one third of that of the unburnt vegetation. But herbaceous plants do constitute a greater proportion than in the unburnt vegetation (respectively 12.9 percent for treatment 6P; 18.7 percent for 6A, and 3.2 percent for T). For the other treatments, comparison becomes more difficult because of differences of elapsed times from the latest burns and the sampling date.

As regards the effect of season of burning, the total mass of vegetation burned in spring is always greater than that burned in autumn. In spring-burned vegetation the above-ground mass of woody plants is always greater than that of vegetation burned in autumn. On the contrary, the mass of herbaceous plants of autumn-burned vegetation is higher than that of spring-burned vegetation. About two years after the fifth burn the herbaceous phytomass of the vegetation burned every two years in autumn (2 A) is greater than that of the unburnt vegetation (T).

CONCLUSIONS The results presented here show that the development of vegetation after fire follows the "initial floristic composition" model described by Egler (1954): all the pre-fire species are present immediately after fire, even if later on, the relative abundance of individuals changes. Thus there is no succession of floristic relays, or different communities succeeding on the same sites as is characteristic of secondary succession. The presence of invading therophytes which often characterize young communities (Bournerias 1959, 455

Table 1 -- Mean above-ground phytomass (g per m2 of dry matter) of the Quercus coccifera garrigue according to the different prescribed fire regimes (May 1979). (Extreme values are presented in parentheses). Fire Regimes Phytomass (g m-2) Woody plants

Herbaceous plants

Total plants

Age of community

T 2.944.9 (4000.0 1450.0) 97.0

6 P 983.4 (1124.0 610.0)

6 A

3 P

754.8 1060.0 175.0)

3 A

135.0 (218.8 68.6)

2 P

19.1 (38.0 8.2)

440.5 (733.9 226.6)

2 A 157.5 (336.0 28.9)

145.5

173.7

23.1

24.0

33.0

117.5

(285.0 1.5)

(278.4 24.3)

(441.5 27.9)

(70.0 0.7)

(78.0 5.7)

(78.1 8.0)

(223.7 30.7)

3041.9

1128.9

(4170.0 1475.5)

(1317.3 888.4)

(1160.4 422.3)

(235.5 93.6)

28 years after latest wildfire

48 months after 2nd prescribed fire

44 months after 2nd prescribed fire

12 months after 4th prescribed fire

928.6

Bazzaz 1968) is practically non-existent except for communities where herbaceous plants are dominant. In fact, mediterranean vegetation does present a recovery by "direct" endogenous process, i.e. the species which existed before fire reoccupy the burned ground, as opposed to an "indirect" or exogenous recovery characterized by a succession of stages as Bournerias (1959), Drury and Nisbet (1973) and Guillerm (1978) have described for old fields. The plants which persist are those which appear immediately after fire and which existed before fire. This is due to the strategies used by the plants to regenerate: mainly by vegetative survival, or by residual seeds buried in the soil or coming from nearby unburnt communities.

158.1

43.1 (105.6 13.9) 8 months after 4th prescribed fire

473.5

275.0

(749.2 234.6)

(417.5 162.9)

24 months 20 months after 5th after 5th prescribed prescribed fire fire

This evolutionary impact has been manifested by positive and negative feedback responses that enable direct fire tolerance or permit its avoidance, followed by vegetative and reproductive regeneration. Acknowledgments: I particularly thank Dr. Lepart who performed the indices of fugacity and discussed thoroughly the French manuscript. I thank also Dr. G. Long for fruitful discussions and comments of the French manuscript. I am grateful to the technicians for their work during the burnings. I greatly appreciated the help from Dr. D. Parsons for reading and correcting the English text.

LITERATURE CITED If fires are lit more frequently than the normal cycle (i.e. about every five or six years at least for such communities as the Quercus coccifera garrigues of Bas-Languedoc), the tendency towards a return to the initial stage is slowed and phytomass gradually decreases. This reduction is due to a decrease in the phytomass of woody plants. Their stems which are located above the ground are constantly burned and do not have enough time to sprout again before the next fire. However, in spite of the burning frequencies, there is little change in the flora of the Quercus coccifera garrigue. The impact of fire only leads to a change in vegetation structure as the upper layers of woody plants disappear and the amount of herbaceous species increases. Thus, the use of prescribed burning would be a possibility to transform these shrubby rangelands as was sometimes done by shepherds many years ago. The plants of the French Mediterranean region are well adapted to withstand fire. The present vegetation of Bas-Languedoc results from many years of evolution during which plants have acquired mechanisms to overcome the effects of fire as well as climatic (especially summer drought) factors.

456

Ahlgren, I.F.; Ahlgren, C.E. Ecological effects of forest fires. Bot. Rev. 1960;26:483-533. Bailloud, D. Avant l'histoire. in Histoire de la France rurale, Ed. Seuil, Paris ; 1975;1:119-189. Barry, J.P. Contribution à l'étude de la végétation de la région de Nimes. Année Biologique, 1960; 36:311-550. Bazzaz, F.A. Succession on abandoned fields in the Shawnee Hills, Southern, Illinois. Ecology, 1968; 49:924-936 Beaulieu, J.L. de. Analyse pollinique dans les Monts de l'Espinouse (Hérault). Pollen et Spores. 1969; 11:83-96 Bournerias, M. Le peuplement végétal des espaces nus (Essais expérimentaux sur la génèse de divers groupements pionniers). Mem. Bull. Soc. Bot. Fr. 1959 ;300 p. Braun-Blanquet, J. La forêt d'Yeuse Languedocienne (Quercion ilicis) Monographie phytosociologique. Mem. Soc. Etud. Sci. Nat. Nimes. 1936 ;5:147 p. Daubenmire, R. Ecology of fire in grasslands. Advances in Ecol. Res., 5, Academic Press, New York 1968 ;209-226.

Delmas, M. La diversité de la vie animale dans la garrigue. Ann. Soc. Hort. Hist. Nat. Hérault, n° spécial la garrigue, 1958; 59-70. Drury, W.H., Nisbet, I.C.T. Succession. J. Arnold Arbor. Harv. Univ., 1973; 54:331-368. Dugrand, R. La garrigue montpellieraine. P.U.F., Paris, 1964; 292 pp. Egler, F.E. Vegetation science concepts. I. Initial floristic composition, a factor in old field vegetation development. Vegetatio, 1964; 4:414-417. Fourquin, G. Le premier Moyen Age. Le temps de la croissance. in Histoire de la France rurale, Ed. Seuil, Paris, 1975;1: 297-550. Guillerm, J.L. Sur les états de transition dans les phytocénoses post-culturales. Thèse Etat, Univ. Sci. Techn. Lang., Montpellier, 1978; 127 pp. Hough, W. Fire as an agent in human culture. U.S. Natl. Mus. Bull., 1926; 139, 270 p. Kornas, J. Succession regressive de la végétation de garrigue sur calcaires compacts dans la montagne de la Gardiole près de Montpellier. Acta. Soc. Bot. Poloniae, 1958; 27:563-596. Kuhnholtz-tordat, G. La terre incendiée. Essai d'agronomie comparée. Ed. La Maison Carrée, Himes, 1938; 361 pp. Kuhnholtz-Lordat, G. L'écran vert. Mem. Mus. Natl. Hist. Nat. 1958; B, 9:276 pp. Laurent, L. A propos des incendies de forêts en Basse Provence. Le Chêne, 1937; 44:139-148. Le Houerou, H.N. Fire and vegetation in the mediterranean basin. Proc. Annu. Tall. Timbers Fire Ecol. Conf., 1974; 13:23/-277. Le Roy Ladurie, E. De la crise ultime à la vraie croissance. in Histoire de la France rurale, Ed. Seuil. Paris 1976; 2:359-611. Liacos, L. Present studies and history of burning in Greece. Proc. Annu. Tall. Timbers Fire Ecol. Conf., 1974; 13:65-95. Long, G., Fa, F., Thiault, M., Trabaud, L. Essais de détermination expérimentale de la productivité d’une garrigue de Quercus coccifera, CEPECNRS, 1967. Doc. n° 39, 28 p. Naveh, Z. The ecology of fire in Israel. Proc. Annu. Tall. Timbers Fire Ecol. Conf., 1974; 13:131-170. Naveh, Z. The evolutionary significance of fire in the Mediterranean region. Vegetatio, 1975; 29:199-208.

Niederlender, A., Lacam, R., Arnal, J. Le gisement néolithique de Roucadour (Themines, Lot) Suppl. Gallia Prehistoria, 1966; 3, 226 pp. Pannoux, P., Pannoux, C. Monographie préhistorique de la région des Matelles (Hérault). Compte-rendu 15° session Congrès Préhist. Fr., 1957; 781-814. Perles, C. L'homme préhistorique et le feu. La Recherche, 1975; 60:829-839. Perles, C. Préhistoire du feu. Masson, Paris, 1977; 180 pp. Ribbe, C. de. Des incendies de forêts dans la région des Maures et de l'Estérel. Rev. Agr. Forest. Provence, 1865; 201-213. Seigue, A. Les incendies de la forêt méditerranéenne. Historique, Essai prospectif. Bull. Tech. Inform., 1972 ; 268:415-423. Sigaut, F. L'agriculture et le feu. Mouton et Cie Paris, 197 ; 320 pp. Trabaud, L. Quelques valeurs et observations sur la phytodynamique des surfaces incendiées dans le Bas-Languedoc. Natur. Monspel., 1970; 21:231-242 Trabaud, L. Experimental study of the effects of prescribed burning on a Quercus coccifera L. garrigue. Proc. Annu. Tall. Timbers Fire Ecol. Conf., 1974; 13:97-129. Trabaud, L. Comparison between the effects of prescribed fires and wildfires on the global quantitative development of the kermes scruboak (Quercus coccifera L.) garrigues. Proc. Symp. Environm. Consequences of Fire and Fuel Manage. in Medit. Ecosystems. USDA Forest. Serv. Gen. Techn. Rep. WO-3, 1977; 271-282. Trabaud, L. Impact biologique et écologique des feux de végétation sur l'organisation, la structure et l'évolution de la végétation des zones de garrigues du Bas-Languedoc. Thèse Etat. Univ. Sci. Tech. Lang. Montpellier, 1980; 291 p. Trabaud, L., Lepart, J., Diversity and stability in garrigue ecosystems after fire. Vegetatio, 1980;43:49-57. Vogl, R.J. Effects of fire on grasslands. in Fire and ecosystems. Academic Press. New York, 1974; 139-194.

457