Homage to Ramon Marga/ef,' or, Why there is such p/easure in studying nature 351 (J. D. Ros & N. Prat, eds.). 1991 . Oec% gia aquatica, 10: 351 -366

THE SUBANTARCTIC FORESTS OF : DISTRIBUTION, STRUCTURE AND PRODUCTION

1 2 2 2 EMILIA GUTIÉRREZ , V. RAMÓN VALLEJ0 , JOAN ROMAÑA & JAUME FONS

1 Departament d'Ecologia. Facultat de Biologia. Universitat de Barcelona. Av. Diagonal, 645. 08028 Barcelona. Spain. 2 Departament de Biologia Vegetal. Facultat de Biologia. Universitat de Barcelona. Av. Diagonal, 645. 08028 Barcelona. Spain.

Received: August 1990.

SUMMARY

Evergreen Nothofagus betuloides and deciduous N. pumilio form the main forest types in Tierra del Fuego. These forests were sampled along two altitudinal gradients to study their structure and dynamics and assess the causes of their distribution. The distribution pattems of the two species of Nothofagus seem to respond to different climates and soils. The dornÍnant soil processes are hydromorphy in the evergreen forest and podzolization in the deciduous one. N. betuloides is an evergreen resilient to short-term environmental fluctuations, due to its ability to retain nutrients. 2 Leaves on the tree may last up to years, with an average density of mg cm, . In contrast, the leaves of 7 2 17 N. pumilio are shed in autumn and reach only 8 mg cm, . In both types of forests the following features can be outlined. Old-growth forest stands develop in the middle and lower slopes. The distribution of diameter sizes of the trees usually shows a pronounced bimodality. Recruitment is discontinuous as shown by the spatial pattem of tree sizes, and regeneration is vegetative in the upper slopes. 2 2 2 2 Leaf area indices range between and m m, for the deciduous forest and from to m m, for 2.3 2.8 2.5 2 4.5 the evergreen forest. It appears that tree standing biomass ranges between and kg e m' in the deciduous 2 6.0 13.6 forest, and between and Kg e m, in the evergreen forest. Production varies from to g e 2 10.5 15.6 2 l 148.0 372.1 m , yr ,1, and from 204.6 to 346.4 g e m , yr, , respecti vely. In relation to boreal forests, biomass can be considered high, but production is medium to low. The relation of wood production to leaf biomass shows that N. betuloides be ha ves similar to coniferous forests and N. pumilio as hardwood forests. There is much accumulation of organic matter on the forest floor, especially coarse wood debris, reflecting the stage of development of these forests and the unfavourable conditions for litter decomposition.

KEY WORDS: Nothofagus, production, forest structure, regeneration, forest soils, Tierra del Fuego.

INTRODUCTION (Mirb.) Oerst. is the main constituent of the evergreen Magellanic forest, whereas N. The forests of Tierra del Fuego extend pumilio (Poepp. and Endl.) Krasser forms along the southernmost foothills of the the deciduous forest. The third species, N. Andes, between 53° 30' and 56° South (Fig. antarctica (Forstf.) Oerst., also deciduous, 1). They are formed by three broad-Ieaved is restricted to the 1east favourable habitats. species of Nothofagus. N. betuloides The forest ecosystems of Tierra del Fuego 352 GUTIÉRREZ, VALLEJO, RO MAÑA & PONS

are poorly known, with the possible and l.7 oC in winter (Fig 2). Thus, the exception of their floral composition. average annual variation is less than 8 oC. Geographical location and climate have Strong winds are the other climatic both contributed to this lack of studies, characteristic of the zone. Climatic even though peculiar environmental subzones are caused by the mountains of conditions make this region a place of the Andes, which in Tierra del Fuego considerable scientific interest. It is the change direction and run from west to east, closest area to the South Pole not covered bordering the westem and southem coasts. by ice. The influence of the Andes, The Andes, as in the rest of the South Quatemary glaciations and the nature of the American continent, are the great structural substrate formed mainly by metamorphic element of the landscape. The distribution and vo1canic rocks, have shaped a unique and type of ecosystems are, to a great landscape. extent, determined by the presence, Tierra del Fuego is a group of islands to arrangement and elevation of the the south of the Magellan Straits, and only mountains, which were lifted in the 1,170 km from Antarctica. The disjunctive Pliocene. The altitude of the Andes in distribution of living Nothofagus attests to Tierra del Fuego is not great, yet sufficient an ancient trans-Antarctic forest belt. to slow down the clouds from the Pacific Nowadays, the proxirnity to this continent ocean. is reflected in the climate, which is much Gradual preclpltation has created a colder than that of the Northem gradient ending in steppes. Consequently, Hernisphere at the same latitude. The the coastal strip occupied by evergreen climate is cold-temperate without forest has a cold-temperate climate with pronounced thermal oscillations during the high humidity (Cfk' c, according to the year. Koppen classification; PISANO, 1981). The Measurements taken at Ushuaia, area occupied by N. pumilio runs between representative of the forest area, show an the steppe and the evergreen forests. The average temperature of 9.4 oC in surnmer climate in this zone is cold-temperate with

TIERRA DEL FUEGO *¡Ushuaia � + -�Wil��_ O. o 10 Km

PACIFIC t::::-..

OCEAN

Cape Horn

FIGURE L . Location map. General distribution of Nothofagus forests on Tierra del Fuego. The striped area corresponds to the deciduous N. pumilio forest and the unstriped coast to the evergreen N. betuloides forest. The samples were taken in the Valle de Andorra (*) and the Costa de Moat (+). NOTHOFAGUS FORESTS OF TIERRA DEL FUEGO 353

steppe degeneration (Dfk'c; PISANO, than 50 cm and N. antarctica usually 1981). Both types of forest are situated appears as an accompanying species. within the c1imatic limits of conifer forests Studies on the forests of Tierra del (Fig. 3) (HAMMOND, 1972). Fuego have been mainly descriptive, The flora and fauna of the different especially biogeographic and taxonomic. terre s tri al ecosystems of Tierra del Fuego Outstanding works are those of PISANO reflect the past and present effects of recent (1975, 1977, 1981) and DIMITRI (1982). glaciations, c1imatic changes, and Fluctuations in the extension of the forests exogenous disturbances. This last factor is based on pollen profiles as well as on other cited by PISANO (1975) as a lack of paleoecological features have been homeostasis. described by AUER (1960), MARKGRAF On the forested mountain slopes (1977) and VUILLEUMIER (1971). The numerous landslides bring abplant relationships for both tree line rarely surpasses 700 m a.s.l. at types of forest (GUTIÉRREZ et at, 1989). these latitudes. There is, however, a lack of quantitive The morphology and degree of studies on most of the ecological aspects of development of the trees are related to the these forests. The results of comparative altitudinal gradient. In the upper part of the studies carried out in the N. pumilio and N. forest and along a strip of more than 40 m betuloides forests are presented in this aboye the timberline, trees take on a study. We have analysed the possible bushlike aspect, known as krummholz. At the tree line the height of the trees is less � GRASSLAND PICAL f'OREST

-----t:::::¡;:� HZ § %W7P USHUAIA T = 4.5 "­ � m p= 550 E f'OREST Alt. 14 5 � � 50 100 ---/// 40 80 {.ZZ. 30 60 Nothofagus pumi1 io forest

1II N. betuloides f"orest 20 40

10 20

O O

MEAN ANNUAL PRECIPITATION (cm)

F M A M A S O N D FIGURE 3. The ecoclimatic position of Nothofa gus forests of Tierra del Fuego among the major terrestrial FIGURE 2. Climatic diagram according to the data ecosystems in the ecoclimatic space defined by from Ushuaia, Tierra del Fuego, 54° 48'S and 68° average temperature and precipitation (modified from 19'W. HAMMOND, 1972). 354 GUTIÉRREZ, VALLEJO, ROMAÑA & PONS

causes of the distribution of the two species the higher slopes of the mountains. of Nothofagus, the forrner deciduous and The evergreen forest was sampled in the the latter evergreen. area of Costa de Moat where the Beagle Climatic conditions are considered to be Channel exits towards the Atlantic, at the primary cause of the distribution of around 110 km east of U shuaia. Although these Nothofagus species, but other factors c1imatic data are not available for this site, may also come into play. Soil it is wetter and warmer than the Valle de characteristics may have a relevant role as Andorra. The area receives between 700 well. On the other hand, apart from the and 1000 mm of rainfall annually research carried out by SCHMIDT & (ZAMORA & SANTANA, 1979a). An URZÚA (1982), no other studies have been altitudinal gradient is also evident in this undertaken concerning the structure and forest. production of these forests and hence no A mixed forest is forrned in the lower data are available for the possibilities of parts, with N. betuloides forrning the exploitation and management techniques. canopy. Drimys winteri Forst. and The existence of great extensions of Maytenus magellanica (Lam.) Hook. f. unexploited forest in Tierra del Fuego forrn the understory, both species being implies minimal human disturbance. Given broad-leaved and evergreen. This type of the speed with which these ecosystems are forrnation disappears at an approximate being exploited for timber, crop and altitude of 150 m a.s.l., giving way to livestock all over South America, there is monospecific stands of N. betuloides. an urgent need for a database for a more Isolated populations of N. pumilio also rational management of the environment. appear in well drained zones. The topography is gentler and the mountains are lower than in the Valle de Andorra. The DESCRIPTION OF SAMPLING SITES upper forest limit reaches 350-400 m a.s.l., and is often limited by peat bogs which, in For the study of the two types of forest, this zone, cover the higher parts of the two representative sites were selected on mountains and extend over the middle the Isla Grande de Tierra del Fuego parts, which are interconnected by gentle (Republic of Argentina). The deciduous N. slopes. pumilio forest was sampled in the Valle de Andorra (54° 49'S & 68° 42'W) and the evergreen forest in the Costa de Moat (54° METHODS 49'S & 67° 21 'W (Fig. 1). The Valle de Andorra is 4 km north of The recognition of different phases of Ushuaia. The annual average precipitation forest development is based on the analysis and temperature are 561 mm and 4.5 oC of population structure of dominant species. (HOLDGATE, 1960). The tree stratum is Plots of 26 m x 26 m were delimited represented fundamentally by N. pumilio. wherever possible in both types of forest in The bottom of the valley is occupied by the upper, middle and lower slopes. peat bogs of Sphagnum sp., with the forests Samples were gathered in each plot for the developing on the mountain slopes, from study of both trees and soils. 240 to 500-600 m a.s.l. In the lower and Estimation of biomass and production middle parts of the forest there are large was carried out using dimensional analysis emerging trees and discontinuities in the (WHITTAKER & WOODWELL, 1968; canopy. On the floor there is a large WHITTAKER & MARKS, 1975) from accumulation of wood, and young shoots field measurements of trees. Diameter at appear in the c1earings. Trees are smaller in breast height (DBH), at 1.30 m aboye NOTHOFAGUS FORESTS OF TIERRA DEL FUEGO 355

T ABLE 1. Climatic characteristics of 6 representative sites of the Nothofagus forest area. Due to their location they can be considered to encompass the gradient over which the forests extend. The data from Isla Nueva are the most representative of the zone in which the evergreen forest was sampled. Med: average annual temperature; Mmed: average of the maxima; rnmed: average of the minima; Mabs: absolute maximum registered; mabs: absolute minimum registered. All the temperatures in oC. P, precipitation in mm yr- I . Data from ZAMORA & SANTANA (1979a) except for those from Ushuaia. SITE LAT LONG Med Mmed mmed Mabs mabs P (mm)

Punta Arenas 53°08'S 70053'S 6.2 10.8 2.0 26.4 -1 1.8 438.9 San Isidro 53°47' S 70059'W 5.9 9.0 3.1 22.0 -6.0 876.5 Ushuaia 54°48'S 68°19'W 5.4 9.6 1.2 29.5 -21.0 550.5 Puerto Navarino 54°57'S 68°20'W 6.4 8.9 3.3 23.2 -8.2 447.8 Puerto Williams 54°56'S 67°38'W 5. 9.4 1.6 26.4 -10.0 553.4 ° � Isla Nueva 55° 1O'S 66 36'W 5.� 9.3 2.8 23.8 -6.2 738.1 ground, tree height, and bark thickness the dominance of deciduous over evergreen were measured in aH trees within the plots. appear evident (Table 1). This is also the W ood cores were extracted from the trunks explanation given from field observations. at 1.30 m covering aH the tree sizes in McQUEEN (1976), HUECK (1978) and order to assess age structure, and to PISANO (1977) attribute a strong ca1culate increases in biomass and sapwood resistance to low temperatures to N. area. Production was estimated from the pumilio. This explanation is supported by wood increments of the last ten annual the presence of species with laurel-like rings. Branch and leaf biomass were leaves, sueh as D. winteri and M. ca1culated through aHometric relationships magellanica, in the evergreen forest. These from 6 trees of N. pumilio and 7 of N. thermophile species are distributed along betuloides, which were feHed, measured the coastline and at low altitude, in areas of and weighed. relative c1imatic mildness where the AH the trees in each plot were mapped evergreen forest is fo und. In the same way, in adjoining squares of 2 x 2 m, forming a evergreen flowering have a lattice of 169 squares. Spatial structure was thermophile distribution compared with the analysed with the hierarchical squares deciduous flowering plants (W ALTER, technique (GREIG-SMITH, 1964) and the 1976), this character being interpreted as distribution pattem was expressed as adaptation to extremely low temperatures. Morisita's index (WILLIAMSON, 1975; The aboye explanation of the distribution ELLIOT, 1983). of Notho fagus species is not entirely Total nitrogen and total carbon in leaves satisfactory. The existence of two distinct were measured with a CarIo Erba forest types is not related to a strong Autoanalyzer. Water-soluble phosphorus in environmental gradient (Table 1). the forest floor and mineral soil was Furthermore, both species are often found determined using HUMPHREYS & growing side by side. Yet, the transition PRITCHETT' s (1972) procedure. A from evergreen to deciduous forest takes detailed explanation of the complete place in just a few kilometers. The limits of sampling methodology can be found in each species could be both c1imatic and A ROMA Ñ et al. (1989). edaphic (PISANO, 1977). N. betuloides survives on soils with rather poor drainage, whereas N. pumilio is much less tolerant to THE DISTRIBUTION OF waterlogging. The soils of the evergreen NOTHOFAGUS SPECIES forest are predominantly hydromorphous, while the processes of podzolization are In a first approximation, the roles of relevant in the deciduous forest (ROMAÑA continentality and of winter harshness In et al. , 1989). The question thus arises as to 356 GUTIÉRREZ, VALLEJO, ROMAÑA & PONS

2 TABLE 11. Specific leaf weight (mg cm- ) for the two TABLE III. Percentage of carbon (C) and nitrogen (N) Nothofagus species. MOl, M02 and M03, sample in leaves (one year leaves, * and two year leaves, +). sites from lowest to highest altitude in the N. MOl, M02 and M03, sample sites from lowest to betuloides forest on the Costa de Moat. V Al, V A2 highest altitude in the N. betuloides forest on the Costa and V A3, the same for N. pumilio in the Valle de de Moat. VA l, VA2 and VA3, the same for N. Andorra. Date of sampling January 1989. pumilio in the Valle de Andorra. LR, Lake Roca (near Ushuaia). species N. betuloides N. pumilio %C %N C/N Species

YEAR MOl M02 M03 VAl VA2 VA3 VA l 44.77 2.43 18.42 N. pumilio SITE VA2 45.22 2.10 21.53 N. pumilio 1 (1988) 14.35 10.89 12.55 8.28 8.61 9.14 VA3 45.45 2.35 19.34 N. pumilio 2 (1987) 19.24 17.71 16.95 MOl 49.50 1.67 29.64 N. betuloides (*) 3 (1986) 19.30 19.10 17.37 48.17 0.85 56.06 D. winteri 4 (1985) ?16.10 19.33 18.15 M02 49.50 1.67 29.64 N. betuloides (*) 5 (1984) 18.54 18.59 49.58 l.l8 42.02 N. betuloides (+) 6 (1983) 18.91 19.20 M03 49.80 1.32 37.73 N. betuloides (+) 7 (1982) 20.04 19.6 1 LR 47. 12 2.17 21.71 N. antarctica the relative advantages of the deciduous or The differences in nutrient content the evergreen habit. between the two species are high. N. The evergreen condition is advantageous pumilio has large accumulations of nitro gen for the absorption of nutrients when growth in its lea ves (Table I11). Thus, the C/N ratio is limited by low soil nutrient reaches values around 20. In contrast, the concentrations. The evergreen character of nitrogen content in N. betuloides is low. Its N. betuloides gives it a certain C/N ratio is almost double, similar to the independence from environmental values found in Quercus ilex. In the soil, fluctuations owing to a greater retention of the C/N ratio is 40 in the Valle de Andorra, nutrients in the tree. Furthermore, the and 80 in the Costa de Moat, values much longer lifespan of the leaves, up to seven higher than those found in the leaves. The years (Table I1), allows greater large accumulation of wood debris on the photosynthetic production per nutrient unit ground and the strong hydromorphism in in the leaves, increasing the efficiency of the Costa de Moat, which retards the their use. This may represent an advantage decomposition of organic matter, seem to both in habitats with low nutrient contents cause these high concentrations of carbon and those in which nutrient recycling is on the ground. difficult. Sc1erophyIlous leaves of N. In swamped boreal environments the betuloides (RAGONESE, 1981) are sc1erophyllous and evergreen habit of adaptative in this kind of habitat. Average various species is associated with low leaf density, an index of sc1erophylly, is nutrient contents in the plant tissues 2 slightly higher than 8 mg cm- in N. (LARSEN, 1982). This has been interpreted 2 pumilio and 17 mg cm- in N. betuloides as an adaptation to restrictions imposed by (Table I1). For comparison, Fagus sylvatica a lack of 02 in soil water, low temperatures 2 has values between 2.2 and 7 mg cm- in in the radicular environment, absence of the Montseny mountains, NE Spain mycorrhizae, and the existence of free (TERRADAS, 1984), lower than those of toxins. Mycorrhizae are indeed very the No thofagus species. The index of abundant in the N. pumilio forest, and sc1erophylly ca1culated for N. betuloides scarce under N. betuloides. The formation (Table II) is similiar to that of the of mycorrhizae has been described for other Mediterranean sc1erophyIles, whose values Nothofagus species (MEYER, 1966). 2 range between 10 and 25 mg cm- (SALA, However, they have never been reported in 1986). waterlogged soils. NOTHOFAGUS FORESTS OF TIERRA DEL FUEGO 357

1 With respect to phosphorus, the TABLE IV. Water soluble phosphorus (Kg P ha- ) in the organic layer, H, and in the first 15 cm of mineral differences between one type of soil and soil. MOl, M02 and M03, sample sites from lowest another are also pronounced (Table IV). to highest altitude in the N. betuloides forest on the The data for soluble phosphorus in the Costa de Moat. VA l, VA2 and VA3, the same for N. pumilio in the Valle de Andorra. organic horizons of the deciduous forest are 1 higher than average annual requirements of Available P (kg P ha- ) forest ecosystems in different environments H 0- 15 (COLE & RAPP, 1981). Values for the SITE MEAN STO MEAN STO evergreen forest range from average to VAl 14.33 6.79 0.5 0.08 moderately low. Sc1erophylly is present in VA2 22.03 3.42 0.2 0.08 sorne Australian species in spite of high VA3 11.68 4.36 0.4 0.22 rainfall; the anomalous presence of this trait MOl 2.30 1.27 0.4 0.22 M02 4.13 3.02 0.1 0.01 has been attributed to low soil fertility, M03 12.10 2.55 because the sc1erophylly index decreases after the addition of nitrogen and the distribution area of N. betuloides phosphorus (BEADLE, 1966). Likewise, in roughly coincides with the rainier coastal well aerated soils at mid-latitudes, areas. In addition, these southernmost areas sc1erophylly has been related to low seem to have been more affected by availabilty of nutrients, especial1y glaciation, producing gentle slopes which phosphorus (LOVELESS, 1961). facilitate water accumulation. The decomposition of organic matter and the release of nutrients in waterlogged soils is slower than that expected under cold STAND STRUCTURE c1imates with low temperatures and minimum thermal fluctuation. Waterlogging The distribution of size frequencies, of the soil appears to be related to the either of DBH or of the biomass of distribution of the two species in the Valle individuals, is frequently used to de Andorra and in the Costa de Moat. The characterize the structure of a forest stand, characteristics of hydromorphic soil would together with the ranges and values of the promote physiological strategies for other variables. Although without cornmon nutrient conservation in the trees, together agreement regarding the size interval used with a greater efficiency in nutrient use. by different authors, we have defined The maintenance of the leaves throughout twelve size c1asses of DBH (FORD, 1975; the year, as well as their greater longevity MOHLER et al., 1978; Figs. 4 and 5). and strong retranslocation of nutrients, can These distributions reflect the differences of be interpreted as adaptative in this context. the forest stands according to altitude. The Such a strategy is possible under c1imatic biggest trees grow in the lower and medium conditions allowing the persistence of slopes, where better edaphic conditions and leaves in winter. Slight differences in the a milder c1imate prevail. Other parameters absolute minimum temperatures seem to be reflect these differences among stands. Yet sufficient for the establishment of such the density of trees is quite high in the conditions, as shown by the c1imatic data in upper slopes. Thus, the basal area increases Table 1. with elevation as a consequence of the At the regional level, however, the greater packing together of individuals. hypothesis of hydromorphy is still Dendrochronological values at 1.30 m questionable, as it is difficult to imagine tend to underestimate tree age. Hardwood that a parameter of this type could have of sorne cores was rotten so the age was such a pronounced geographical estimated dividing the tree radius by the component. One element to consider is that annual mean increment. To avoid over- or 358 GUTIÉRREZ, VALLEJO, RO MAÑA & PONS

underestimation of tree ages, cores rotten to forests (cf. SCHMIDT & URZÚA, 1982). more than half of tree radius were rejected. Most of the observed size distributions are In this manner, our results indicate that the bimodal (Figs. 4 and 5). The regenerating biggest trees are not always the oldest (Fig. c1ass, formed by trees with les s than 7 cm 6). Tree ages show a high variation with of DBH, is scarce or lacking in many respect to size in both forests. This may be stands except the one in the lower part of a consequence of the great heterogeneity the Valle de Andorra (Fig. 4) and in the of habitats. In the evergreen forest, a high upper slope of Moat (Fig. 5). But hardly variability is also found in the ages reached any of these stands can be considered within a narrow category of DBH, uneven-aged. The bimodal distributions especially abo ve 45 cm. In contrast with N. correspond more to the fi nal stage of pumilio, betuloides N. reaches older ages � 1400 VA3 . s.. 1 = 3 with smaller diameters. This may be related v cm o- d = ·1008 to the oligotrophic condition of the soil in U) t- [:j p:; 1000 which the latter species grows (Table IV). �

Nevertheless, diffe rences exist among the o :z: stands according to the site characteristics. 600 In the lower slopes of the evergreen forest, c1asses of DBH less than 25 cm do not exist and there is no c1ear relation between 200 age and DBH. In the remaining stands the o 7 35 63 17 relationship between size and age is more 21 49 DBH SIZE CLASSES (cm) evident and could be statistically described al VA2 :r: 1400 by allometric and/or exponential functions. s.. 1 = 5 cm v = The structure of a forest is dynamic. It o- d 1000 1000 &lw changes over time and depends on the p:; ecological characteristics of both the � o component species and the whole habitat. :z: 600 The dependence on the characteristics of the species is manifest in the degree of tolerance to shade. This trait determines to 200 a large extent the way in which 0+-���-r��=T�r-.-. O 7 77 regeneration is produced and is reflected in 21 35 49 63 DBH SIZE CLASSES (cm) size and age structure of the stand. The � 1400 VAl s.. more shade-tolerant species tend to form v 1 = 7 o- cm d = 799 multiple-cohort stands and tree recruitment U) [:j 1000 is continuous throughout time. On the other p:; �

hand, when the trees vary little in size and o have originated in a re latively short interval :z: 600 of time, the structure of the stand is even-aged, corresponding to single-cohort stands. This type of stand can be formed by 200 shade-tolerant or shade-intolerant species, 35 49 63 17 although even-aged stands of intolerant DBH SIZE CLASSES (cm)

I species are more common. According to the HGURE 4. Distribution of trees acconling to thlo! DHH literature, Nothofagus species are sizes in the lower (VAl) , medium (VA2) and upper (V A3) slopes of the deciduous pumilio forest in the shade-intolerant (HUECK, 1978). Thus, the N Valle de Andorra. Notice that the interval (1) defining expected distribution of DBH's would size classes and the density of the stand (d, number of correspond to that of even-aged stand trees per hectare) varies among sample plots. NOTHOFAGUS FORESTS OF TIERRA DEL FUEGO 359

self-thinning and even-aged stands of the plot of the higher slope in the Valle de mature foresto In these stages of Andorra. development, distributions are often indeed Analysis of the spatial distribution of all bimodal when twelve size categories are the trees in the plots gives c1ues to the used (FORD, 1975; MOHLER et at, 1978). regeneration pattem. Two examples of Bimodality is a consequence of the aging of spatial distribution are shown in figure 7 the stand in which the largest size for the lower part of the Valle de Andorra, categories in trunk diameter and tree height VAl, and the higher part of the Costa de form a distinct group of trees. When age is Moat, M03. In the first case, regeneration considered (Fig. 6), most of the stands is produced in areas free of large cannot be c1assified as even-aged, except individuals. Therefore, the bimodality observed in DBH size distributions (Figs. 4 al :x: 1400 M03 and 5) may be due to the recruitment in lo. 1 = 4 cm v Po d = 2692 bursts of new individuals when the canopy ti) rz¡ rz¡ 1000 has disappeared. Although it is not known ¡t; E-< up to what point the Nothofagus species are ci z 600 shade-intolerant, it is very significant that in the low part of the Costa de Moat where the N. betuloides forest has D. winteri and 200 M. magellanica in the understory, there are 49 63 77 91 105 DBH SIZE CLASSES (cm) Nothofagus betuloides �7al 00 lo. M02 al • � 1400 ,;6 00 1 = 5 cm lo. v d = 1361 500 • Po <�

�rz¡ 1000 400 • ¡t; E-< o • o 300 z o • o A A 600 200 • 100 o • 200 04---�--r---��--�---r---r---r- o 63 17 91 10 20 30 40 50 60 70 80 21 35 49 105 DBH SIZE DBH SIZE CLASSES (cm) (cm)

� 1400 MOl lo. v = cm Po N.b. 1 7 Nothofagus pum ilio = 00 c:::J d 251 -;;;lo. 7 rz¡� 1000 al ¡t; 1 = 2.5 cm E-< D.w. ,;600 Cl d = 3595

1 = 2.5 cm zci M.m. �< 500 600 - d = 281 • 400

200 300 • • 200 . .. . 21 35 49 17 63 91 105 A A DBH SIZE CLASSES (cm) A .... 100 O AA FIGURE 5. Distribution of trees according to the DBH lJOO sizes in the lower (MOl), medium (M02) and upper O+-�AT---.---r--'--�---r---r---r­ O (M03) slopes of the evergreen N. betuloides forest on 10 20 30 40 50 60 70 80 DBH SIZE (cm) the Costa de Moat. Notice that the interval (1) defining size elasses and the density of the stand (d, number of FIGURE 6. Relation between age and size of trees in trees per hectare) varies among sample plots. Data are 6 sampled plots. Data from the lower (triangles), also given for Drimys winteri (D.w.) and Maytenus middle (filled cireles) and upper (open cireles) slopes magellanica (M.m.). of each type of forest. 360 GUTIÉRREZ, VALLEJO, ROMAÑA & PONS

neither saplings nor poles (Fig. 5, MOl). which becomes random as the area If the recruitment of new individuals is increases (Morisita index values around produced in bursts and these Nothofagus one). However, when aH the trees are species do not regenerate beneath their included, the distribution is aggregate up to 2 canopy, the regenerating class should areas of 64 and 80 m in the low and exhibit a spatial pattem in accordance with medium parts of both forests, and 16 and 2 environmental rninor disturbances. Spatial 36 m in the higher parts. These cluster analysis was carried out twice, once for the dimensions indicate that the recruitment of large trees, with a DBH aboye 30 cm, and new individuals is produced in a then including all the trees. The results are discontinuous form and in a space defined displayed in figure 8. The large trees by other physical attributes. The result is a feature a regular distribution in small areas mosaic of patches of sizes and ages, (Morisita index values of less than one), sometimes overlapping and sometimes distinct (cf. PEET & CHRISTENSEN, VA l 1987). The development of these patches M - • . • may be subpopulations or different groups w . of trees, but other new ones also appear as · • . the distributions of diameters (Figs 4 and 5) · .o • • and ages (Fig. 6) seem to indicate. On the · other hand, the spatial distribution of trees • is very different between the upper and G . . • ·í lower slopes of both forests (Fig. 7). The o . .'- O I . distribution of large trees in the upper parts • ! is ag regate, the clusters being of 16 and e· . DBH � , • • 32 m . When aH the trees are included the I I 01 I .0�=�� 2 O :_20-30 size of the patches is 4 and 8 m , these . e-30-40 • .-,40-50 being included in the former clusters. The l . o - • . >50 smaHer clusters present in the higher parts 1 2 3 4 5 6 7 8 9 10 11 12 13 of the forest correspond to a different type of regeneration, i.e. vegetative. . M03 , M • _ el .e ._ Regeneration is accomplished by sprouting +--=0¡+!, _;.-+- -+--+--_I,! .-�I ! •-----'-- 1 d.+e-_ • +- . -'.J I !, • -f- , ! • I ¡ . at the base of the trunks of living trees. l • : ,�• ' , . ¡ � ' : : 1 ¡ •• This type of regeneration renders a different . • • J •••• • i e. l· i pattem from those of the lower and middle •• I .�: ..¡ •. .: .• parts of the forest. . . ¡ H • . ,; • Another very important structural t . G ' °0 : .. :e I parameter is the leaf area index, LAI. The ·· e 1 · I I leaf area index is fairly similar in aH the 1 e : . �I . . � sampled plots; values are between 2.3 and . :1 . 2 2 o • 1;:-:1 • :.: 4.5 m m- , being higher in the evergreen e �+-4-·4-· �-'-�-�I T: -+-+�-.-+-� forest (Fig. 9 and Table V). The range of . . . . . ! I • values for the leaf area index are coherent o . A o: . • . .e • : • • with those expected according to latitude, 1 2 3 4 5 6 7 l 8 l 9 10 11 12 13 climate, and the structure and dynamics of forest stand s subjected to frequent treefalls. FIGURE 7. Spatial distribution of the trees in the Fagus sylvatica lower slope of the Valle de Andorra (V Al) and in the features values of 3.2 in upper slope of the Costa de Moat (M03). The leaf area index at similar latitudes in the distribution at middle elevations is very similar to that N orthem Hernisphere owing to the harsher of the lower slopes. DBH in cm, grid in m. White climate (De ANGELIS et al., 1981). circles represent dead individuals. NOTHOFAGUS FORESTS OF TIERRA DEL FUEGO 361

The relation between leaf area and only small-diameter vessels (RIVERA, sapwood area is interesting due to their 1987, 1988). In principIe, this implies a mutual functional implication. These smaller LAI per unit of sapwood with No thofn.f!.us species are diffuse-porous with respect to ring-porous trees such as oaks. The values found for N. pumilio are 0. 12 'O VA1 2 2 H and 0. 13 m cm- in the stands situated in 2 the medium and lower slopes and 0.07 m 1 2 cm- in the upper slopes. Results show that 4 a greater sapwood surface is needed in the stand of the subalpine zone, i.e. the area of evaporation per unit of sapwood area is lower. These variations may reflect aspects 3 related to the permeability of the wood, the increased viscosity of water, a possible adaptation to more dessicant conditions 2 imposed by wind and a higher demand of carbohydrates. The values are similar to those fo und for N. solandri in montane and subalpine stands, respectively (BENECKE .....-_ ...... -- .... -- A--- & NORDMA YER, 1982, in W ARING & ,..- -- / SCHLESINGER, 1985). The ratio of leaf / / / to sapwood area ca1culated for the stands of betuloides forest, gives values of 0.32 N.2 2 2 3 4 5 6 7 m cm- for the lower slope c10se to the sea 2 2 AREA UNITS (L) and 0. 11 and 0.09 m cm- for the medium 'O and upper slopes, respectively. The M03 H requirements of the sapwood area appear to be reduced in the rnilder c1imate, particularly in the lower coastal location. 3 q\ \ These and other previously mentioned \ \ differences among stands at different \ \ elevations may be due to either genetic \ 2 \ differentiation (WARING & SCHLESINGER, 1985) or environmentally Q , , , dependent different gene expressions. ,

� ...-- _ ...... _- ...... ; BIOMASS, PRODUCTION AND SOIL ORGANIC MATTER

2 3 4 5 6 7 The range of aboveground biomass and

AREA UNITS (L) production values is shown in Table V. The FIGURE 8. Representation of the values of the evergreen forest features a greater biomass Morisita index (Id) according to area. Each area unit but a lower production than the deciduous 2 2 measures (2 * L) m , L being the unit of length. Open forest. This difference translates into a dots are data points significant to 95 % according to the test. The val ues around are tho e that take a lower biomass turnover rate for the former. F 1 random distribution, those over 1 appear in aggregates, The Nothofagus forests of Tierra del Fuego and those below I are regular. The broken line have c1imatic conditions corresponding to corresponds to trees with DBH > 30 cm, the continuous line to all trees in the plot. the zone of boreal conifer forests (Fig. 3), 362 GUTIÉRREZ, VALLEJO, ROMAÑA & PONS

but show higher biomass values than the TABLE V. Structural characteristics, including biomass and production, of the Nothofagus forests latter (Table V). This is not observed in the sampled in Tierra del Fuego. pumilio stands of the upper slope, with a N. 2 -I production of 148.0 g e m- yr , a figure Deciduous forest Evergreen forest that lies in the low range of observed Age 91 - 247 339 - 628 I values. High values of biomass are Trees ha- 799 - 4008 25 1 - 2692 2 2 translated into a much lower productivity LAI (m m- ) 2.3 - 2.8 2.5 - 4.5 2 I and a higher turnover rate in relation to the Basal area (m ha- ) 50.5 - 66.9 62.6 - 107.7 Biomass 6.0 - 13.6 10.5 - 15.6 boreal forests. These data agree with the 2 (kg e m- ) results and conc1usions that refer to other Production 148.0 - 372.1 204.6 - 346.4 2 structural parameters. ( g e m- -1 ) 2 1 Relative to the incident radiation, the Leaves (gY e m- i) 85.3 - 100.6 90.2 - 143.5 evergreen forest has a greater average PIB (yr- I) 0.022 - 0.03 1 0.018 - 0.027 B/P (yr) 46 - 32 56 - 37 production efficiency than the deciduous one. Values range from 0.32 to 0.54%, and Boreal Forests Temperate Forests

Biomass 2.5 - 16.6 2.5 - 25.7 "i E -2 500 E LJ1 (kg e m ) LJ1 C\J '" Production 166.7 - 833.4 250.0 - 1041.8 �450 '" E 2 I e, (g e m- yr- ) en M O 400 O � � � � :E N 350 O �o:l from 0. 17 to 0.44%, respectively. In the 300 � � relevant calculations involved we � 250 t1l considered the average incident radiation in > 200 each area (ZAMORA & SANTANA, <2 150 1979a). The average estimated incident f-o 100 < �>- O radiation for the deciduous forest is 85 Kcal < :E ..J 50O t1l 2 U el cm- yr- and that of the evergreen forest is H z < 2 1 < f-o 65 Kcal \cm- yr- . Although these data are <:> 50 en g§ 100 8 not referred to the sampling stations, they g provide an adequate quantification. The u: 150 E E efficiency of these forests appears to be LJ1 O 300 LJ1 en LJ1 low, being more similar to that of en � 250 M N hardwood than coniferous forests (WEBB oc{ oc{ :;¡:> �o:l 200 > > et al�, 1983), in spite of having c1imatic el 150 , §O similarities with the latter. But if we o:: � <:> t1l 100 consider relations such as the production of > 8 50 � <2 O t1l wood relative to leaf biomass (Fig. 10), the el '" lIf.' "' ,, betuloides " . r- ill,,' N. forest has a lower wood en �..J ::¡ < 50 � 1 > production per unit of photosynthetic 1,1: � '"1 :'i 100 '" surface than N. pumilio. Furthermore, given u L- :1:,lo betuloides � 150 that N. has a greater LAI its <:> 200 ¡;]l growth efficiency, expressed in g of wood g§ ..J 2 H 250 , , , per m of leaves per year, is lower, being O O O eS O en O O O bO

maintammg high leaf area indices are chemical conditions, affect, in tum, greater in cold c1imates. Besides the effect biological activity. Thus, the rate of of c1imate, in Tierra del Fuego the decomposition integrates the biological following factors should be considered: potential of the soil and determines both the windfall disturbance, fungal disease, tumover and availability of nutrients for landslides and snow-related dynamism of plants, mainly N and P. The decomposition soil, all of which prevent higher LAI and rate can be considered as a soil fertility leaf biomass values. If indeed, forests that indexo do not reach and maintain an LAI of 6 do In the topographic sequence of the Valle not reach a steady state (WARING & de Andorra, the organic matter, and SCHLESINGER, 1985), the forests of especially the wood debris in the organic Tierra del Fuego may be an example of horizons, decreases from the lower to the forests that do not reach a successional upper slopes, in accordance with biomass steady state. (Fig. 9). These results indicate that the In the general context of production in differences in organic matler content are terre s tri al ecosystems, the accumulation of explained by the quantity of litter falI. biomass in the forests of Tierra del Fuego Furthermore, the harsher c1imate hi\� th� appears to be limited by low temperatures. upper slopes does not appear to delay the Temperature is indeed one of the main decomposition of organic matter. Since the factors limiting production although it is estimated contribution from leaves is not adequately described by means of simple statistical relations (WARING & C\J I SCHLESINGER, 1985). This dependence E on temperature is also conspicuous at a tI:J CI:l 2400 local level along the altitudinal gradient in CI:l the Valle de Andorra. The limitations on �H 1Il2000 potential production in the upper slope of ¡... < the deciduous forest can be related to a ¡.¡ ...:1 harsher c1imate. On the other hand, soil 1600 conditions may have a greater importance in the evergreen forest (Table IV, Fig. 9). 1 200 Thus, the altitudinal gradient is blurred by the heterogeneity of soil hydromorphism. o 800 0 In agreement with the structural and 0

0 .. dynamic characteristics, the amount of 0 * o * . organic material accumulated in the forest 400 . . ..:...... soil of Tierra del Fuego is high (Fig. 9). * • . The c1imatic conditions, low temperatures O+-----.------r----�------�­ and an average annual fluctuation that does o 200 400 600 800 not exceed 8 oC, favor a slow WOOD INCREMENT (g m-2yr-1 ) decomposition. The accumulation of organic matter on the soil results from the o Needle-leaved stands * N. betuloides • Broad-l eaved stands * N. pumi lio balance between litter inputs and the rate of decomposition, the former being dependent FIGURE 10. Wood production in relation to leaf upon forest production. The rate of biomass of the sampled Nothofagus forests in Tierra decomposition is determined by the del Fuego, together with published data for other c1imatic conditions, the nature of the litter, forests (O'NEILL & De ANGELIS, 1981). Notice the similarity of the evergreen betuloides with and the characteristics of the soil itself. The N. coniferous forests and of deciduous N. pumilio with soil characteristics, e.g., waterlogging and hardwood forests. 364 GUTIÉRREZ, VALLEJO, ROMAÑA & PONS

similar in the three plots (Fig. 8, Table IV), sampled. These values indicate a relatively the diffe rences in litter accumulation must rapid incorporation of litter remains into the be accounted for by the wood fraction. H horizon. The mean residen ce time of The accumulation of organic matter in leaves in the LF horizon, that is 1/k, ranges the sampled stands of the evergreen forest, between three and five years. However, the in contrast to the Valle de Andorra, values of k, calculated for the leaf fraction, increases from the lower to the upper parts. may be underestimates, since wood debris This difference is conditioned by the role of less than 2 cm in diameter were inc1uded. excess of water in the evergreen forest, These woody remains were not inc1uded in where it causes a great spatial variability in the estimates of litter fallo The mean the accumulation of organic matter. residence time for all litter fall inc1uding Decomposition takes place temporarily leaves and coarse wood debris ranges from under anaerobic conditions, which are quite 15 to 60 years in both Nothofagus forests. restrictive to mineralization. The H soil horizon in the stands on the lower slopes of both forests has a high mineral content, CONCLUSIONS probably due to more intense biological activity. A lot of worms were observed The forest landscape of Tierra del Fuego during sampling. has a mosaic pattern at different spatial High amounts of litter and wood debris scales. Within a given forest type, patches may control the long-term productivity of of younger trees are inc1uded in the biggest forest commumtles (CHARLEY & patches of old forest. This structure may be RICHARDS, 1974). The accumulation of caused by strong disturbances like organic matter in the H horizon is very high landslides and windthrown trees in the in the forests studied: up to around 200 M slopes. The gaps produced after such events 1 ha- in Valle de Andorra and 135 Mg ha­� are colonized by Nothofagus seedlings, in Costa de Moat (Fig. 9). These values are while there is little tree regeneration under similar to those found in the boreal the old-growth forest canopy. The result is coniferous forests of Vancouver, 50° N of multicohort stands. There is also a c1ear latitude (VALLEJO, 1987). In contrast, altitudinal gradient in forest structure, the typical Mediterranean forests, such as those extreme conditions of the upper slopes of evergreen Quercus ilex in Montseny, NE preventing the development of old-growth 1 Spain, reach only 10 to 50 Mg ha- forests. (HERETER, 1990). Owing to the limititations imposed by The H horizon has plenty of tree roots in c1imate, the Nothofagus forests of Tierra both Nothofagus forests. However, in Moat del Fuego have low productivity. These most of the roots usually do not reach the unexploited forests have a temperature­ mineral horizons because of waterlogging. limited accumulation of biomass and AIso, the H horizon has the highest production. Yet there is a huge concentrations of nutrients in the soil accumulation of organic matter on the (Table IV). Therefore, the H horizon plays forest floor with an important contribution a key role in the maintenance of soil of decaying wood. The large amounts of fertility. Rational forest management accumulated litter reflect the slowness of practices should take this into account. the decomposition process and the stage of The rate of decomposition (k; OLSON, the forest. 1963), calculated for the LF horizon and In adverse soil conditions, such as taking into account the estimated waterlogging, the evergreen N. betuloides is contribution of fal len leaves from Table V, the dominant forest species. However, it ranges between 0.20 and 0.38 in the stands has a lower productivity than N. pumilio, NOTHOFAGUS FORESTS OF TIERRA DEL FUEGO 365

even though it is distributed in the rainier ACKNOWLEDGEMENTS and milder areas of Tierra del Fuego. Broad-Ieaved evergreen trees are We are grateful to the Servei exceptional in extreme latitudes whereas d'Espectroscópia de la Universitat de they are common in very different habitats, Barcelona, and the Centre de Recerca such as under Mediterranean type Ecológica i Aplicacions Forestals, CREAF, conditions. This study shows that the for the technical assistance provided. This evergreen and sc1erophyllous characteristics study was funded by the Instituto de of N. betuloides are related to shortage of Cooperación Iberoamericana, ICI, nutrients in waterlogged soils. This is a Comisión Nacional V Centenario. Thanks physiological limitation similar to those are given to Robin Rycroft and Cristian attributed to typical Mediterranean RuÍz for their accurate translation and ecosystems in much drier conditions. interesting comments.

REFERENCES

AUER, V. 1960. The quatemary history of Ecología, composlclOn e importancia económica. Fuego-. Proc. Roy. SOCo (B), 152: GTZ. Berlin. 507-515. HUMPHREYS, F.R & PRITCHETT, W.L. 1972. A BEADLE, N.C.W. 1966. Soil phosphate and its role in soil testing for predicting phosphorus status of molding segments of the Australian flora and Pinus elliottii plantations. Plant and soil, 37: vegetation, with special reference to xeromorphy 479-489. and sclerophylly. Ecology, 47: 992-1 007. LARSEN, J.A 1982. Ecology of the northern lowland CHARLEY, J.L. & RlCHAROS, B.N. 1974. Energy bogs and conifer fo rests. Academic Press. New and nutrient dynamics of forest floors. Proc. 7th. York. Triennial Con! IFA . Calouncha, 1: 25-34. LOVELESS, AR 1961. A nutritional interpretation of COLE, D.W. & RAPP, M. 1981. Cycles élémentaires sclerophylly based on differences in the chemical dans les ecosystemes forestiers. In: Dynamic composition of sclerophyllous and mesophytic properties of fo rests ecosystems. (D.E. Reichle, leaves. Ann. Bot., 25: 179-184. ed.): 34 1 -409. Cambridge Univ. Press. Cambridge. MARKGRAF, V. 1977. New data on the late- and De ANGELIS, D.L., GARDNER, R.H. & SHUGART postglacial vegetational history of "La Misión", Jr., H.H. 1981. Productivity of forests ecosystems Tierra del Fuego, Argentina. Proc. IV Inter. Pa/yn. studied during IBP: the woodlands data set. In: Con! Lucknow. Dynamic properties of fo rests ecosystems. (D.E. McQUEEN, AR 1976. The ecology of Nothofagus Reichle, ed.): 567-672. Cambridge Univ. Press. and associated vegetation in South America. Cambridge. Tuatara, 22: 36-38. DIMITRI, M.J. 1982. La región de los bosques MEYER, F.H. 1966. Mycorrhiza and other plant andino-patagónicos. INT A. Buenos Aires. symbiosis. In: Symbiosis (S.M. Henry, ed.): ELLIOTT, J.M. 1983. Some methods fo r the statistical 171-255. Academic Press. London. analysis of benthic brates. Wilson. Kendal. MOHLER, C.L., MARKS, P.L. & SPRUGEL, D.G. FORO, E.D. 1975. Competition and stand structure in 1978. Stand structure and allometry of trees during sorne even-aged plant monocultures. 1. Eco/., 63: self-thinning of pure stands. 1. Eco!. , 66: 599-614. 311-333. OLSON, J.S. 1963. Energy storage and the balance of GREIG-SMITH, P. 1964. Quantitative plant ecology. producers and decomposers in ecological systems. 2nd ed. Butterworths. London. Ecology, 44: 322-33 1. GUTIÉRREZ, E., FONS, J., ROMAÑÁ, J. & O'NEILL, RV. & De ANGELIS, D.L. 1981. VALLEJO, RV. 1989. Soil-vegetation Comparative productivity and biomass relations of relationships in the Nothofagus forests of Tierra del forest ecosystems. In: Dynamics properties of Fuego. Proc. XVIPacific Inter. Congress. Chile fo rest ecosystems. (D.E. Reichle, ed.): 411-450. HAMMOND, A.L. 1972. Ecosystem analysis: biome Cambridge Univ. Press. Cambridge. approach to environmental science. Science, 175: PEET, R.K. & CHRISTENSEN, N.L. 1987. C i i d d 46-48 . ompet t on an tree eath. BioScience, 37: HERETER, A 1990. Els sois fo restals del Massís del 586-595. Montseny. Ph. D. Thesis, Universitat de Barcelona. PISANO, E.V. 1975. Características de la biota HOLDGATE, M.W. 1960. Vegetation and soils in the magallánica derivadas de factores especiales. Ans. South Chilean Islands. J. Ecol., 49: 559-580. Inst. Pat., 6: 123-1 37. HUECK, K. 1978. Los bosques de Sudamérica. PISANO, E.V. 1977. Fitogeografía de 366 GUTIÉRREZ, VALLEJO, RO MAÑA & PONS

Fuego-Patagonia chilena. 1. Comunidades vegetales V ALLEJO, V.R. 1987. Ana1isi i valora ció de sols entre las latitudes 52° y 56° S. Ans. Inst. Pat., 8: fo restals. Report CIRIT. Barcelona. 121-250. VUILLEUMIER, B.S. 1971. Pleistocene changes in PISANO, E.V. 1981. Bosquejo fitogeográfico de the fauna and flora of South America. Science, Fuego-Patagonia. Ans. Inst. Pat., 12: 159-171. 173: 771 -780. RAGONESE, A.M. 1981. Anatomía foliar de las WALTER, H. 1976. Vegetació i dimes del món. especies sudamericanas de Nothofagus. Uniersitat de Barcelona. Barcelona. Darwiniana, 23: 587-603. W ARING, R.H. & SCHLESINGER, W.H. 1985. RIVERA, S.M. 1987. Aspectos silológicos vinculados Forest ecosystems. Concepts and management. a la dendrocronología de Nothofagus pumilio Academic Press. Orlando. (Poepp. et Endl.) Krasser. 1. Area de vasos. WEBB, W.L., LAUENROTH, W.K., SZAREK, S.R. Parodiana, 5: 9-16. & KINERSON, R.S. 1983. Primary production and RIVERA, S.M. 1988. Revisión xilológica del género abiotic control s in forests, grasslands and desert Nothofagus Bl. (Fagaceae) para la Argentina. Mon. ecosystems in the United States. Ecology, 64: Acad. Nac. Cien. Exac. Fea. & Nat., 4: 73-84. 134-151. ROMAÑA, J., FONS, J., VALLEJO, R.V. & WHITTAKER, R.M. & WOODWELL, G. IM. 1968. GUTIÉRREZ, E. 1989. Condiciones edáficas en Dimension and production relations of trees and relación con la vegetación fo restal en la Isla shrubs in the Brookhaven forest, New York. J. Grande de Tierra del Fuego. Final Report. ICI, Eco 1. , 56: 1-25. Madrid. WHITTAKER, R.M. & MARKS, P.L. 1975. Methods SALA, A. 1986. Evolució de l'estructura vertical de of assessing terrestrial productivity. In: Primary la garriga després del fo c: índex fo liar i gradients productivity of the biosphere. (H. Leith & R.M. morfologics. Master Thesis. Universitat de Whittaker, eds.): 55-118. Springer. Berlin. Barcelona. WILLIAMSON, G.B. 1975. Pattern and seral SANTANA, A. 1984. Variación de las precipitaciones composition in an old-growth beech-maple forest. de 97 años en Punta Arenas como índice de Ecology, 56: 727-73 1. posibles cambios climáticos. Ans. Inst. Pat., 15: ZAMORA, E. & SANTANA, A. 1979a. 51-60. Características de la costa occidental de la Ú SCHMIDT, H. & URZ A, A. 1982. Transformación y Patagonia entre las latitudes 46° 40' Y 56° 30' S. manejo de los bosques de lenga en Magallanes. Ans. Inst. Pat., 10: 109-144. Ciencias Agrícolas, 11. Universidad de Chile. ZAMORA, E. & SANTANA, A. 1979b. Oscilaciones TERRADAS, J. (ed.) 1984. Introducció a l'ecologia y tendencias térmicas en Punta Arenas entre 1888 del fa ig al Montseny. Diputació de Barcelona, y 1979. Ans. Inst. Pat., 10: 145-154. Barcelona.