Global Ecology and Biogeography, (Global Ecol. Biogeogr.) (2008) 17, 164–174 Blackwell Publishing Ltd

RESEARCH PAPER

Do leaf margins of the temperate forest flora of southern South America reflect a warmer past? Marcelo A. Aizen1* and Cecilia Ezcurra1,2

1

Laboratorio Ecotono, Centro Regional Bariloche, Universidad Nacional del Comahue, Quintral 1250, 8400 San Carlos de Bariloche, Río Negro, Argentina, 2Departamento de Botánica, Centro Regional Bariloche, Universidad Nacional del Comahue, Quintral 1250, 8400 San Carlos de Bariloche, Río Negro, Argentina

ABSTRACT

Aim Our aims were: (1) to characterize the linear relationship between the proportion of woody dicotyledonous species with entire-margined leaves (E) and mean annual temperature (MAT) from a southern temperate flora that still harbours many lineages that originated under warmer climates; (2) to compare this relationship with those developed from floras of different regions of the world; and (3) to contrast temperature predictions based on leaf margins of the native southern flora versus the naturalized alien flora, mostly of boreal origin. Location The temperate forest of southern South America (TFSA). Methods At each 1° latitudinal band, we estimated E based on species latitudinal ranges and MAT from both an isotherm map and a global temperature grid. We also calculated E from five local floras located between 40 and 43° S, and from the naturalized alien flora of Nahuel Huapi National Park in southern Argentina. Results We found a close relationship between E and MAT for the TFSA. Equations developed from floras of the Northern Hemisphere overestimated extant temperatures of this biome by 6–10 °C at both geographical and local spatial scales. On the other hand, MAT predictions from leaf margins of the alien flora were similar to the actual MAT. A published regression between E and MAT from tropical South America was remarkably similar to the one we estimated from the TFSA. This tropical equation predicted accurately the temperatures observed for this temperate biome based on leaf margins of the native flora.

*Correspondence: M. A. Aizen, Laboratorio Ecotono, Centro Regional Bariloche, Universidad Nacional del Comahue, Quintral 1250, 8400 San Carlos de Bariloche, Río Negro, Argentina. E-mail: [email protected]

Main conclusions Despite massive plant extinction due to environmental cooling and biogeographical isolation during the Tertiary, leaf-margin analysis reveals that the flora of the TFSA still reflects its original development under the warmer conditions of western Gondwana and its past connections with low-latitude forest floras of tropical South America. Keywords Alien flora, biogeographical history, Gondwana, leaf-margin analysis, native flora, Neotropics, South America, temperate forest, temperature prediction.

INTRODUCTION Leaf morphology and function are greatly influenced by climate. Thus, different leaf traits can be used as biological proxies to infer environmental conditions in present and past terrestrial environments. In particular, the proportion of woody dicotyledonous species with entire-margined leaves in a given local flora (E) closely reflects mean annual temperature (MAT) from that locality. This association is usually so strong that empirical

164

simple linear regressions, such as MAT = 30.6E + 1.14 and other similar equations (Wolfe, 1979; Wing & Greenwood, 1993; Wilf, 1997; Kowalski, 2002), are used by palaeoecologists to estimate ancient temperatures from fossil leaf assemblages (Wolfe, 1979, 1995; Wilf, 1997; Wiemann et al., 1998; Graham et al., 2001). This technique, known as leaf-margin analysis, could also be used as a comparative tool to assess the influence of biogeographical history and phylogenetic inertia on present leaf– climate relationships (e.g. Kowalski, 2002; Greenwood et al.,

DOI: 10.1111/j.1466-8238.2007.00350.x © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd www.blackwellpublishing.com/geb

Leaf margins in the temperate forest flora of South America 2004). Although the physiological basis for the relationship between E and MAT is still poorly understood, it has been proposed, among other hypotheses, that serrate or toothed leaf margins allow enhanced rates of photosynthesis under low temperatures by increasing evapotranspiration when water is not limiting (Baker-Brosh & Peet, 1997; Wilf, 1997; Royer & Wilf, 2006). Several linear relationships between leaf margin and temperature have been estimated from different leaf data sets, including both temperate and tropical forest communities (reviewed in Kowalski, 2002). However, their application in regions or sites not considered in their original calculation may be informative about the influence of historical events on present ecological patterns and processes (Greenwood et al., 2004). In particular, these regression equations should overestimate present temperatures in regions of the world where extant temperate floras still reflect a strong tropical heritage, either due to less exposure to extreme cold over evolutionary time or to the absence of an important source of cold-adapted taxa, thus exhibiting a high proportion of taxa with entire-margined leaves. Here we conducted a leaf-margin analysis (LMA) of the woody dicotyledonous flora of the humid-temperate forest of southern South America (hereafter TFSA). This biome reveals its warmer past not only through congeneric relationships with extant tropical taxa (Arroyo et al., 1996), but also through different physiognomic and ecological traits, such as a high incidence of epiphytism and an unexpectedly high incidence of plant–animal mutualisms (Armesto & Rozzi, 1989; Aizen & Ezcurra, 1998; Aizen et al., 2002). However, the comparison of present temperatures within the TFSA with those predicted using regressions developed from other regions of the world may provide a more definitive assessment of this view. The TFSA stretches as a narrow belt, 100–250 km wide, along the western rim of the continent between 35 and 55° S (Fig. 1). Nowadays, this biome is surrounded by the Pacific Ocean to the west and south and a series of arid and semi-arid ecosystems to the east and north (Cabrera & Willink, 1973; Grau, 1995). However, during much of the Palaeocene and early Eocene most of what is today southern South America was covered by a rich tropical-like forest continuous with other forest masses at lower and higher latitudes (Romero, 1993; Arroyo et al., 1996; Markgraf et al., 1996; Hinojosa & Villagrán, 1997; Villagrán & Hinojosa, 1997), even though a large arid zone seems to have existed towards the west of South America at mid-latitudes (Ziegler et al., 2003; Wilf et al., 2005). The separation and drifting away of South America from Antarctica about 30 Ma (Hinojosa & Villagrán, 1997; Poulin et al., 2002; Ortiz-Jaureguizar & Cladera, 2006), and the rain shadow created by the major uplift of the southern Andes since approximately 15 Ma (Solbrig, 1978; Simpson, 1983; Villagrán & Hinojosa, 1997; Ortiz-Jaureguizar & Cladera, 2006; Meudt & Simpson, 2006), left the forest of southern South America biogeographically isolated. In addition, the establishment of the Antarctic circumpolar current and the development of Antarctic glaciations during the last million years resulted in the cooling of southern South America, which initiated a strong south–north temperature gradient (Simpson,

Figure 1 Map showing the extent of the temperate forest of southern South America (modified from Veblen et al., 1996, and Ezcurra & Brion, 2005), and the location of the five local floras analysed (Ch, Chiloé; Py, Puyehue; PB, Puerto Blest; Ll, Llao-llao; Ba, Bariloche). The positions of the north and south Patagonian ice fields are also shown. High-Andean tundra-like vegetation is too patchy to depict at the scale of this map.

1983; Hinojosa & Villagrán, 1997; Ortiz-Jaureguizar & Cladera, 2006). Thus, the highly endemic flora of this region, which developed during much of the Cenozoic under a warmer climate than today’s, may be considered a biogeographical relict of a more widespread and ancient forest biome with strong Gondwanan and Neotropical past connections (Arroyo et al., 1996; Villagrán & Hinojosa, 1997). Biogeographical isolation and particularly cool conditions caused the extinction of many plant taxa with tropical affinities from the Oligocene to middle Miocene onwards (Menéndez, 1972; Romero, 1993; Markgraf et al., 1996; Villagrán & Hinojosa, 1997; Wilf et al., 2003, 2005; Zamaloa et al., 2006). The establishment of the cold corridor of the Andes, however, allowed the

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

165

M. A. Aizen and C. Ezcurra southwards migration of several taxa from the Holarctic region from late Miocene onwards (Raven & Axelrod, 1974; Simpson, 1983). Transoceanic dispersal from the west through the Pacific also appears to have introduced species since those times (Wardle et al., 2001). More recently, during the 19th and 20th centuries, European colonists also brought a large number of cold-adapted alien boreal species that have flourished in the TFSA, particularly in disturbed areas (Brion et al., 1988a; Ezcurra & Brion, 2005). The comparison of temperature predictions based on leaf margin proportions between this naturalized woody flora, which has not had enough time to evolve novel characters in its new environment, and the extant native forest flora, provides an interesting test of the influence of different biogeographical origins on these climatic estimates. Here we asked the following questions: (1) What are the temperatures predicted for southern South America by various different linear regressions between the proportion of woody dicotyledonous species with entire-margined leaves E and mean annual temperature MAT? (2) Does E increase with increasing MAT within the TFSA? (3) Does the relationship between E and MAT from the TFSA resemble equivalent relationships estimated from other floras of the world? (4) Do MAT predictions based on the naturalized alien woody flora introduced during the last five centuries, since European colonization, differ from those based on the native flora? By providing answers to questions 1–4, we approach a more fundamental and general question: how much does past biogeography influence present ecological features of a biome? MATERIALS AND METHODS Within the TFSA, we calculated E and estimated MAT along a latitudinal range of 15°, from 40 to 55° S. As the MAT versus leaf margin relationship is sensitive to the inclusion of dry sites with little rainfall during the growing season (Wilf, 1997), we excluded the woody dicotyledonous species occurring mostly in more deciduous forest types north of 40° S, a region characterized by a more mediterranean type of climate with frequent summer droughts (Grau, 1995). Leaf physiognomy and its relationship to present and past climate of the coastal Cordillera forests of this region has been recently analysed by Hinojosa et al. (2006). The woody dicotyledonous flora of the TFSA south of 40° S includes a total of 167 native species. We constructed a data base (see Appendix S1 in Supplementary Material) that incorporated the following information for each species: growth form (tree, shrub, epiphyte, vine or hemiparasite), leaf margin (entire or toothed) and southernmost latitudinal limit of distribution within southern South America. The list of species was compiled from Correa (1969–99), Muñoz-Schick (1980), Moore (1983), Hoffmann (1991), Donoso-Zegers (1994), Donoso-Zegers & Ramírez-García (1994), Henriquez et al. (1995), Landrum (1999) and Ezcurra & Brion (2005). Information on the latitudinal distribution of each species was obtained from literature as indicated for each species in Appendix S1. Here the southernmost limit of a species represents not only a measure of its southern extension, but also a good estimate of its total latitudinal

166

extent. With three exceptions (Appendix S1), all the woody species of the TFSA have their northern limit near the northern limit of this biome north of 40° S, mostly between 37 and 40° S (Hoffmann, 1991; Arroyo et al., 1996). We also compiled species lists for the local floras of five different localities between 41 and 43° S, along the longitudinal extent of the TFSA (Fig. 1, Appendix S1). The dominant forest type present at each locality was, from west to east, coastal Valdivian rainforest at Chiloé [42°13′ S, 73°50′ W, 200 – 600 m above sea level (a.s.l.); species information from Armesto & Rozzi, 1989], Andean Valdivian rainforest at Puyehue (40°40′ S, 72°09′ W, 450–750 m; Muñoz-Schick, 1980) and Puerto Blest (41°02′ S, 71°49′ W, 750–1000 m; Brion et al., 1988b), mesic Nothofagus forest at Llao-llao (41°04′ S, 71°36′ W, 750–1000 m; Brion et al., 1996), and a transitional Nothofagus–Austrocedrus forest at Bariloche (41°07′ S, 71°13′ W, 750–1000 m; Naumann, 1987). Chiloé and Puyehue are located in Chile, whereas the other three localities are in Argentina, within Nahuel Huapi National Park (see Aizen & Ezcurra, 1998, for more details). Additionally, a list of woody alien dicotyledonous species naturalized in the forests of Nahuel Huapi National Park and adjacent areas (see Appendix S2 in Supplementary Material) was obtained from Rapoport & Brion (1991) and Ezcurra & Brion (2005). Following the methodology traditionally used by palaeobotanists (Wolfe, 1979; Wilf, 1997; Kowalski, 2002), we classified each species based on leaf-margin morphology in one of two categories, entire versus toothed. From species latitudinal ranges, we calculated the proportion of woody species with entire-margined leaves (E) present in each 1° latitudinal interval, from 40 to 55° S. Although to our knowledge, this macroecological approach had not been previously used in the context of leaf-margin analysis, it allowed us to count on samples with ≥ 20 species to estimate E with precision at all latitudes, particularly for the more species-poor forest regions of the far south (see Wilf, 1997; Kowalski, 2002). The inclusion of the few high-altitude, cold-adapted species (e.g. toothed Nothofagus pumilio) in this type of latitude-interval based analysis, would only make any test of temperature overestimation more conservative. We also calculated E for each of the five local floras and for the alien flora of Nahuel Huapi National Park. Observed mean annual temperature (MAT) was calculated for each 1° latitudinal band along the extent of the TFSA through two different methods, because of inaccuracies associated with climatic estimates at large spatial scales. First, we used isotherm maps for southern South America depicting mean annual temperature isolines with a 5 °C resolution (Hoffman, 1975). Along lines separated by 1° latitude between 40 to 55° S, we estimated MAT at points 50 km apart within the limits of the TFSA by linear interpolation between consecutive isotherms. We then averaged MAT over the data points included in each latitudinal line. With the aid of satellite images, we excluded from this calculation points occurring at high-Andean tundra vegetation or ice fields (Clark, 1992). This MAT estimation involved the averaging of four to nine temperature data points per 1° latitudinal line. Second, we used a 360 × 720 global Cartesian, orthonormal

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

Leaf margins in the temperate forest flora of South America geodetic (lat./long.) grid based on digital raster data depicting monthly values of temperature for each 30′ × 30′ cell (Leemans & Cramer, 1992). We averaged monthly values for each cell with > 50% of its area within the limits of the TFSA, and chose the cell with the highest MAT within each 1° band as the representative mean temperature of the forest environment at that latitude. In this way, we avoided the influence of low temperatures associated with the treeless environments of the high Andes and ice fields in the MAT calculation. Estimations based on these two methods were contrasted, for calibration purposes, with actual MATs for 12 cities and localities within the TFSA region (Dirección Meteorológica de Chile, 2002; Servicio Meteorológico Nacional de Argentina, 2004). This calibration showed that our temperature estimations corresponded to an altitude of 0 –200 m, where most forest species present at any latitudinal band occur (e.g. Correa, 1969–99, Muñoz-Schick, 1980, Moore, 1983). In addition, MAT predictions based on E from the local floras were compared with actual MATs from nearby localities at similar or lower altitudes. MAT estimates from the Chiloé flora sample were compared with actual MATs from Ancud (41°52′ S, 73°49′ W, 3 m a.s.l.; time series 1961–70) and Guafo Island (43°34′ S, 74°49′ W, 7 m; 1910–50), the Puyehue flora sample with MAT from Osorno (40°34′ S, 73°9′ W, 86 m; 1990 –2002) and MATs from the three samples of Nahuel Huapi National Park with MATs from Bariloche (41°09′ S, 71°09′ W, 780 m; 1914–90) and Mascardi Lake (41°9′ S, 71°17′ W, 800 m; 1969 –85). All these localities are < 50 km from their respective flora sites, except for Guafo Island, which is about 150 km from the Chiloé sampling site. Data analysis We compared temperatures based on isotherms and on cells of the global temperature grid (hereafter isotherm MAT and cell MAT, respectively) along the latitudinal extent of the TFSA with those predicted using six different published simple linear equations of E versus MAT (Table 1). Equations 1– 4 are based on

Northern Hemisphere sites (East Asia and North America), although eqn 2 includes some Neotropical sites in the Southern Hemisphere. Equation 5 is exclusively based on South American data, which includes sites distributed over a wide geographical range within the continent and a variety of low-latitude Neotropical forest floras, but no temperate flora. A second South American equation restricted to Bolivia (Gregory-Wodzicki, 2000) is not included in our analysis because of peculiarities with the leaf data set already discussed by Kowalski (2002). Another equation (eqn 6) has been recently published for Australia (Greenwood et al., 2004), and provides an interesting comparison between southern lands of former Gondwana. We conducted a leaf-margin analysis from our latitudinal data set studying the relationship between E and each of our two different MAT estimators through simple linear regression. We also included latitude as a second independent variable in a multiple linear regression analysis to test whether north–south variation in both MAT and E could account for the relationship between these two variables. Spatial autocorrelation could also represent another confounding factor of a geographicallystructured data set like ours. To address this possibility, we used the Durbin–Watson test (SAS, 1999) to analyse whether residuals from either our simple or multiple regression equations were autocorrelated. In none of the regressions did we detect any significant first- or higher-order autocorrelation (results not shown). Thus, for the purpose of our analysis, our latitudinal data can be considered as a group of 16 independent observations. The linear regressions between E and MAT estimated for the TFSA were compared with those obtained for other regions of the world (Table 1). Assuming that the intercepts and slopes of the linear eqns 1–6 represent true parameter values, we tested whether our sampling statistics differed from those parameters (Sokal & Rohlf, 1995). To increase the power of these tests, we weighed each observation by the number of species used to estimate E at each latitude. Because of inflated type-I errors associated with multiple comparisons, statistical significance was established at α = 0.05/6 = 0.0082 (Rice, 1989).

Table 1 Simple linear regression models used to predict mean annual temperature (MAT) from the proportion of woody species with entire margined leaves (E). For each equation, information is provided on the number of sites surveyed (n), coefficient of determination (r 2) and standard error (SE) of the fitted regression, geographical region where samples were obtained, approximate latitudinal extension and literature source. The linear regressions estimated in this paper for the TFSA are provided in Table 2a. Equation

n

r2

SE

Region

Latitude

Source

(1) MAT = 30.6E + 1.14

34

0.98

0.8

Eastern Asia

20° N–60° N

(2) MAT = 28.6E + 2.24

9

0.94

2.0

20° S–40° N

(3) MAT = 29.1E – 0.266

106

0.76

3.4

30° N–50° N

Wolfe (1993), Wilf (1997)

(4) MAT = 24.4E + 3.25

74

0.84

2.1

30° N–50° N

Wolfe (1993), Wilf (1997)

(5) MAT = 38.5E – 10.24

30

0.47

3.4

(6) MAT = 27.0E – 2.12

74

0.63

2.2

Western Hemisphere (from Pennsylvania to Bolivia) North America (CLAMP data base) North America (CLAMP data base with 32 cold sites removed) Tropical South America (from Colombia to Bolivia and Brazil) Australia

Wolfe (1979), Wing & Greenwood (1993) Wilf (1997)

9° N–21° S 12° S–40° S

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

Kowalski (2002) Greenwood et al. (2004)

167

M. A. Aizen and C. Ezcurra

Figure 2 Number of woody species with entire and toothed leaf margins present at 1° latitudinal bands along the north–south extent of the temperate forest of South America.

We used the proportion of species with entire-margined leaves from the five local floras and from the naturalized alien flora of Nahuel Huapi National Park to predict MAT at a more local scale using two contrasting equations (Table 1): eqn 1, a widely used regression in the context of LMA and representative of the equations from the temperate Northern Hemisphere (Wolfe,1979; Wing & Greenwood, 1993; Wilf, 1997), and eqn 5, the regression equation from sites in tropical South America (Kowalski, 2002). Standard errors of predicted MATs were estimated following Wilf (1997) as b [E (1 − E)/s], where b is the slope of the E versus MAT regression, and s is the number of plant species in the flora sampled. RESULTS The proportion of species with entire-margined leaves in the dicotyledonous woody flora of the TFSA is 0.517 (n = 167). Although we found a clear decrease in species richness with latitude for species with either entire or toothed leaf margins, the shape of this latitudinal decline differed between these two groups (Fig. 2). Thus, E varied throughout the latitudinal extent of the TFSA, from values > 0.50 recorded at latitudes < 42° S to values as low as 0.37 at about 46 and 50° S (Fig. 3). The latitudes with the lowest E values coincided with the occurrence of the northern and southern Patagonian ice fields (Figs 1 & 3). Isotherm- and cell-based temperature estimates were similar and practically indistinguishable when averaged over the latitudinal extent of the TFSA (mean ± 1 SE = 6.1 ± 0.54 versus 6.5 ± 0.69 °C, respectively; paired t-test, t15 = 1.21, P = 0.24). Also, latitudinal trends in these two estimators of observed temperature were relatively accurate in describing average temperatures of nearby cities and localities. Comparisons of MATs for localities and either isotherm or cell estimates at similar latitudes differed typically by < 1 °C (Fig. 3). Leaf margin analysis using eqns 1– 6 produced different MAT predictions. Considering latitude as a blocking factor, an 

168

Figure 3 Proportion of woody species with entire-margined leaves (E) and mean annual temperature (MAT) along the latitudinal extent of the temperate forest of South America. In the lower panel, the dashes indicate the approximate latitudinal location and extent of the northern and southern Patagonian ice fields. In the upper panel, MAT estimates based on isotherm maps and cells from a global temperature grid are shown with the temperature predictions made by eqns 1–6 (Table 1). From north to south the crosses depict actual MATs for the following cities and localities: Valdivia (39°47′ S, 73°13′ W, 10 m a.s.l.), Osorno (40°34′ S, 73°9′ W, 86 m), Bariloche (41°10′ S, 71°30′ W, 825 m), Puerto Montt (41°28′ S, 72°56′ W, 3 m), Ancud (41°52′ S, 73°49′ W, 3 m), El Bolsón (41°58′ S, 71°31′ W, 310 m), Esquel (42°50′ S, 71°10′ W, 755 m), Puerto Aisen (45°24′ S, 72°42′ W, 11 m), Coihaique (45°34′ S, 72°4′ W, 302 m), Balmaceda (45°55′ S, 71°41′ W, 583 m), Punta Arenas (53°09′ S, 70°48′ W, 10 m), Ushuaia (54°48′ S, 68°17′ W, 7 m).

comparing temperatures predicted by these six equations and estimated through the isotherm and cell methods showed highly significant differences (F7,127 = 330.9, P < 0.0001). An a posteriori pairwise Ryan’s Q test grouped these eight MAT predictions and estimates into four different clusters (Fig. 3). The first cluster included MATs predicted by eqns 1–2 and 4. Temperature predictions based on these equations overestimated extant temperatures (as estimated through the isotherm and cell methods) by 7–8 °C on average, although at high latitudes (> 50° S) differences were greater (about 10 °C). The second cluster included MAT predictions made by eqn 3, which includes data from cold sites in North America (see Wilf, 1997, for more details). This equation overestimates observed temperatures by about 6 °C. Closer temperature predictions were provided by the Australian eqn 6 (third cluster), which were, on average, only 3 °C higher than the observed temperatures. Finally, predictions

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

Leaf margins in the temperate forest flora of South America Table 2 Simple and multiple linear regressions of mean annual temperature (MAT) within the temperate forest of South America as a function of (a) the proportion of woody species with entire margined leaves (E), and (b) latitude and E. MAT was estimated from isotherm maps and from cells of a global temperature grid (see Materials and Methods). Sample size is 16 in all cases.

Isotherm MAT Estimate (SE) (a) Simple Intercept Leaf margin (E) (b) Multiple Intercept Latitude Leaf margin (E)

Cell MAT t

P

Estimate (SE)

t

P

–10.09 (2.63) –3.85 0.0018 38.25 (6.16) 6.21 < 0.0001 (r2 = 0.733, SE = 1.15)

–11.37 (4.49) –2.53 0.0241 42.36 (10.56) 4.01 0.0013 (r2 = 0.535, SE = 1.97)

2.51 (4.60) 0.55 0.59 – 0.18 (0.059) –3.07 0.0090 28.63 (5.79) 4.94 0.0003 (r2 = 0.854, SE = 0.90)

10.26 (7.88) 1.30 0.22 –0.31 (0.10) –3.07 0.0090 25.86 (9.92) 2.61 0.0217 (r2 = 0.730, SE = 1.56)

Table 3 F values testing whether the intercepts and slopes of the simple regressions reported in Table 2a are proper estimators of those from eqns 1–6 (Table 1). In all cases, numerator and denominator degree of freedoms are 1 and 14, respectively. F values significant at α = 0.05/6 = 0.0082 are shown in bold. Isotherm MAT

Figure 4 Simple linear regressions of mean annual temperature (MAT) within the temperate forest of South America (TFSA) as a function of the proportion of woody species with entire-margined leaves (E). MAT was estimated from isotherm maps and from cells of a global temperature grid (see Table 2a for fitted equations). Equations 1–6 (Table 1) are depicted for comparison. Equation 5, the equation from tropical South America, overlaps the one derived from isotherm maps of the TFSA.

made by eqn 5, the equation from tropical South America, clustered together with the estimates of observed MAT based on isotherms and the temperature grid. This equation predicts a MAT, averaged over the latitudinal extent of the TFSA, of 6.1 ± 0.46 °C, the same overall mean temperature value estimated from isotherms (see above). The proportion of woody species with entire-margined leaves increased with MAT along the latitudinal extent of the TFSA. Because of a strong relationship between these two variables, simple regression equations could be used to predict MAT (MAT = 38.25E – 10.9, r 2 = 0.733, for isotherm MAT, and MAT = 42.36E – 11.37, r 2 = 0.535, for cell MAT, n = 16 in both cases; Table 2a, Fig. 4). These two regressions share a similar slope (F1,28 = 0.11, P = 0.74) and intercept (F1,28 = 0.06, P = 0.81). The relationship between E and MAT persisted after accounting for the significant effect of latitude on temperature using a multiple regression analysis (Table 2b). In no case could we reject the hypothesis that the slopes of the simple regressions estimated from the flora and actual temperatures of the TFSA were statistically different from those of eqns 1– 6 in Table 1 (Table 3, Fig. 4).

Eqn 1 Eqn 2 Eqn 3 Eqn 4 Eqn 5 Eqn 6

Cell MAT

Intercept

Slope

Intercept

Slope

27.94 33.78 21.29 39.63 0.02 13.89

2.80 4.43 3.99 9.05 0.00 5.99

16.03 18.72 12.90 21.37 0.45 9.29

4.09 5.23 4.93 8.07 0.95 6.24

Despite homogeneous slopes, and in agreement with the comparisons between extant temperatures and those predicted through LMA (Fig. 3), intercepts of the regressions estimated for the TFSA were improper estimators of those of eqns 1–6 with one exception. This was the equation from tropical South America (eqn 5), which was practically identical to the regression equation we estimated using isotherm data (Fig. 4). Temperature overestimation of the TFSA by regressions from other continents was not an artefact of the macroecological scale we used, but an intrinsic characteristic of the forest flora of this temperate region. MAT predictions using eqn 1 and E values estimated from the five local floras (0.500, n = 60 for Chiloé; 0.484, n = 62 for Puyehue; 0.426, n = 47 for Puerto Blest; 0.466, n = 58 for Llao-llao; and 0.410, n = 39 for Bariloche) were 6–7 °C higher than the actual temperatures recorded for those localities. For the Chiloé and Puyehue flora samples, differences could even be 1–2 °C higher if differences in altitude with their respective meteorological stations were taken into account (i.e. considering an altitudinal lapse rate of 0.6 °C per 100 m). On the other hand, using eqn 1 the naturalized woody alien flora of Nahuel Huapi (n = 34) predicted a mean temperature that differed by 1–2 °C from the actual mean temperature for this region, with a sampling error interval that largely overlaps the range of recorded MATs (Fig. 5a). Equation 5, the tropical South American

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

169

M. A. Aizen and C. Ezcurra

Figure 5 Predictions of mean annual temperature ± 1 SE based on the proportion of species with entire-margined leaves calculated from five local floras in the northern region of the TFSA and using (a) eqn 1 and (b) eqn 5 (Table 1) (open symbols and black error bars). MAT predictions (means ± 1 SE) based on the naturalized woody alien flora of Nahuel Huapi National Park are also shown (black squares and black error bars). Mean annual temperature (MAT) predictions at the Chiloé site were compared with actual MATs from Ancud (left) and Guafo Island (right), at Puyehue with MAT from Osorno, and MATs predicted for the three localities at Nahuel Huapi National Park and the alien flora of this National Park with MATs from Bariloche (left) and Mascardi Lake (right). Crosses depict the means and the grey bars the range of observed MAT values. Geographical coordinates for these cities and localities as well as the time series over which temperature was recorded are provided in Materials and Methods.

equation, produced more accurate predictions of observed MATs based on leaf margins of the native flora, whereas it clearly underestimated MAT for the Nahuel Huapi region when considering leaf margins of the alien flora (Fig. 5b). DISCUSSION Leaf-margin analysis (LMA) of the TFSA flora, using regression equations from the Northern Hemisphere, overestimated present-day temperatures by about 6 –8 °C and up to 10 °C in southern South America. This overestimation occurred at both

170

regional and local spatial scales within this region. Only an equation from tropical South America (Kowalski, 2002) succeeded in making accurate predictions of extant temperatures at these more temperate latitudes. This tropical equation was strikingly similar to the relationships between the proportion of species with entire-margined leaves (E) and mean annual temperature (MAT) we reported for the TFSA. In addition, temperature predictions made by an Australian linear equation (Greenwood et al., 2004) were also more accurate than those made by its Northern Hemisphere counterparts. Finally, at a more local scale, a comparative LMA applied to the alien versus native forest flora of Nahuel Huapi National Park resulted in divergent temperature predictions, clearly reflecting their different biogeographical origins. These findings suggest that leaf characters, and particularly leaf margins, in the currently isolated flora of the TFSA still reflect ancient connections to other southern floras with warmer and more equable climates. During the Palaeocene and early Eocene (about 50 Ma), plant lineages with tropical and austral (Gondwanan) affinities grew at mid and high latitudes in South America (Romero, 1993; Arroyo et al., 1996; Villagrán & Hinojosa, 1997; Wilf et al., 2005; Ortiz-Jaureguizar & Cladera, 2006). Indeed, one of the most diverse palaeofloras of Eocene age has been reported from Patagonia (Wilf et al., 2003, 2005). This palaeoflora was rich in lineages now extinct in southern South America (Menéndez, 1972; Wilf et al., 2005; Zamaloa et al., 2006), but which still survive in mountain and lowland forests at lower latitudes in tropical America and Australasia (Mabberley, 1993; Markgraf et al., 1996; Wilf et al., 2005; Zamaloa et al., 2006). Plant extinctions observed in the fossil record can be attributed mainly to the general cooling and isolation of forests of southern South America during the mid-Eocene onwards (Ortiz-Jaureguizar & Cladera, 2006), as has also been found for New Zealand (Lee et al., 2001). This temporal pattern of decreasing taxonomic diversity in relation to cooling may also be inferred from the present strong spatial north–south gradient in woody plant species diversity along the latitudinal extent of TFSA (Fig. 2; Arroyo et al., 1996; Veblen et al., 1996; Villagrán & Hinojosa, 1997). Also, the relative decrease of species with entire-margined leaves with decreasing temperature that we found for the TFSA suggests that plant traits that evolved under a warmer climate might have increased susceptibility to extinction. Therefore, species extinctions appear to have occurred in a flora that developed mostly under more tropical-like conditions (Romero, 1993; Hinojosa & Villagrán, 1997; Wilf et al., 2005; Hinojosa et al., 2006; Ortiz-Jaureguizar & Cladera, 2006), which unlike floras from the Northern Hemisphere had relatively few representatives of forest lineages that evolved under a truly cold climate. Hence, because of this biogeographical background, the present flora of the TFSA seems to have proportionally more extant lineages originated under warmer and more equable environments than its northern counterparts, an origin revealed by other physiognomic and ecological traits besides leaf margins (Armesto & Rozzi, 1989; Aizen & Ezcurra, 1998; Aizen et al., 2002). This heritage is also reflected in the leaf physiognomy of the modern, more deciduous, forest flora of the Chilean Coastal

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

Leaf margins in the temperate forest flora of South America range just north of our study region. Using a multivariate approach and comparisons with pre-Pleistocene palaeofloras, Hinojosa et al. (2006) concluded that the forest patches found along the Pacific coast of South America between 33° S and 40° S are relicts of a more ancient tropical-like forest flora that occupied this region from the Palaeogene to early Neogene (65– 10 Ma). They also suggest that this relictual flora maintains many original foliar physiognomies due to longstanding phylogenetic inertia (Hinojosa et al., 2006). Here, using a different methodological approach, we expand this view by proposing that most of the flora of the temperate forest region that extends all the way south to Tierra del Fuego can be considered reflective of a warmer past. Most of the few genera of the native woody flora from the TFSA that originated under truly cold conditions are not endemic. Despite a high incidence of endemism at the generic level (Aizen & Ezcurra, 1998), the majority of these cold-adapted floristic elements are, geologically speaking, new arrivals in southern South America that migrated from the Northern Hemisphere after the establishment of the cold Andean corridor about 10 Ma (Simpson, 1983). These northern elements are a minor component of the present flora (about 10% of the extant woody genera; Aizen & Ezcurra, 1998), however. Interestingly, E estimated from the 26 native woody species from genera of putative boreal origin (including Baccharis, Berberis, Empetrum, Ribes and Rubus; Arroyo et al., 1996; Aizen & Ezcurra, 1998) is 0.308 (Appendix S1), remarkably similar to the value of 0.294 estimated from the naturalized alien woody flora of Nahuel Huapi we report here. This striking similarity can be attributed to a common biogeographical origin rather than taxonomic resemblance, as these two groups of species share just one genus in common (Rubus). In contrast, E calculated from the other 271 native species, mostly belonging to Austral (Gondwanan), Neotropical, and endemic lineages (Aizen & Ezcurra, 1998), was 0.550, a value more typical of floras from warmer climates (Jordan, 1997; Kowalski, 2002; Greenwood et al., 2004). The near absence in the TFSA of endemic lineages evolved under, and probably adapted to, extreme cold conditions could be attributed to the fact that these conditions were never so extensive in southern South America, perhaps due to the influence of a more equable climate associated with the low continentality and strong oceanic influence that prevail in this region to present times (Arroyo et al., 1996). Some relevant conceptual and practical implications derive from our findings. In both hemispheres E reflects MAT, but whereas Northern and Southern Hemisphere regressions share a similar slope, they differ in their intercepts (Tables 1–3, Fig. 4). The relative invariance of the slope could reflect the universal influence of factors linking temperature with leaf margins in mesic environments (Wilf, 1997). These factors (e.g. evapotranspiration rates) should represent general constraints on leaf function reflecting the influence of climate on leaf metabolism. In our study, the sensitivity of leaf margins to temperature was so pronounced that changes of E along the latitudinal extent of the TFSA even reflected the location of the two Patagonian ice fields (Fig. 3). On the other hand, differences in the intercept might relate to the different biogeographical templates on which

this functional relation was imposed. That is, in a comparative sense this latter parameter might be indicative of the average temperature settings under which different forest floras, and their constituent plant lineages, originated. Thus, the intercept of the leaf-margin/temperature relationship could provide a quantitative estimate of the composite influence of different historical effects (sensu Herrera, 1992). For instance, the much lower value of this parameter for the LMA regressions we report for the TFSA (Table 2a) could relate to the warmer conditions that prevailed during most of the Palaeogene, under which many of the ancestors of the present-day TFSA species evolved. Further, these differences in the value of the intercept imply that the application of LMA regressions for palaeotemperature estimations should be done with extreme caution using regionally specific models (see also Stranks & England, 1997; Greenwood et al., 2004). A key question that remains to be answered is whether ancestral characters (e.g. entire leaf margins) that still persist today in the extant flora of the TFSA can be considered adaptive or not in the present-day climate. One point of view is that these conserved characters are non-adaptive under present cooler conditions because of their warmer origin. This view might explain, for instance, why there is such a large number of alien plant species of cold boreal origin brought by European settlers that have been successful in colonizing this southern biome (Dimitri, 1972; Moore, 1983). Today about 25% of the extant flora of Nahuel Huapi National Park and adjacent areas is non-native (Ezcurra & Brion, 2005). An alternative view is that the adaptive nature of these traits persisted over time, becoming well-suited to presentday environmental conditions (‘exaptations’ sensu Gould & Vrba, 1982). After all, most of the extant taxa of this southern flora represent the survivors of ancient lineages that were able to pass through different past climatic sieves and endure many environmental transitions, including recent climatic cycles associated with the Pleistocene glaciation events (Markgraf, 1991; Markgraf et al., 1995, 1996; McGlone, 1996). Under this view, the more cold-adapted alien, boreal taxa, with their low E, could be less adapted to the cool but milder and more equable conditions of southern South America. A comparative assay of leaf performance (e.g. photosynthesis rates) between native versus alien species under different temperatures could prove valuable to discriminate between these views. The LMA we conducted here, plus ancillary experimental information, could help us not only to understand how the history of the TFSA flora shaped modern ecological traits but also how this flora might respond to future global warming scenarios. ACKNOWLEDGEMENTS We thank Cecilia Nuñez, Thomas Kitzberger, Vera Markgraf, Adriana Ruggiero, Diego Vazquez, Tom Veblen, Peter Wardle and Peter Wilf for discussion and useful comments on a previous version of this manuscript. Constructive suggestions by the handling editor, Matt McGlone, and three referees are also acknowledged. This work was partially funded by grants from the National Research Council of Argentina (PIP 5066) and Universidad Nacional del Comahue (B126/04).

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

171

M. A. Aizen and C. Ezcurra REFERENCES Aizen, M. & Ezcurra, C. (1998) High incidence of plant-animal mutualisms in the temperate forest of southern South America, biogeographical origin and present ecological significance. Ecología Austral, 8, 217–236. Aizen, M.A., Vázquez, D.P. & Smith-Ramírez, C. (2002) Historia natural de los mutualismos planta-animal del Bosque Templado de Sudamérica Austral. Revista Chilena de Historia Natural, 75, 79 –97. Armesto, J.J. & Rozzi, R. (1989) Seed dispersal syndromes in the rain forest of Chiloé: evidence for the importance of biotic dispersal in a temperate rain forest. Journal of Biogeography, 16, 219 –226. Arroyo, M.T.K., Cavieres, L., Peñaloza, A., Riveros, M. & Faggi, A.M. (1996) Relaciones fitogeográficas y patrones regionales de riqueza de especies en la flora del bosque lluvioso templado de Sudamérica. Ecología de los bosques nativos de Chile (ed. by J.J. Armesto, C. Villagrán and M.T.K. Arroyo), pp. 71–99. Editorial Universitaria, Santiago de Chile. Baker-Brosh, K.F. & Peet, R.K. (1997) The ecological significance of lobed and toothed leaves in temperate forest trees. Ecology, 78, 1250 –1255. Brion, C., Calvelo S. & Lorenzo, L. (1996) Plan de manejo del parque Municipal Llao Llao, Anexo 4, Flora. Universidad Nacional del Comahue, Centro Regional Universitario Bariloche, San Carlos de Bariloche. Brion, C., Grigera, D., Puntieri, J. & Rapoport, E.H. (1988a) Plantas exóticas en bosques de Nothofagus, comparaciones preliminares entre el norte de la Patagonia y Tierra del Fuego. Monografías de la Academia Nacional de Ciencias Exactas, Físicas y Naturales (Buenos Aires), 4, 37– 47. Brion, C., Puntieri, J., Grigera, D. & Calvelo, S. (1988b) Flora de Puerto Blest y sus alrededores. Universidad Nacional del Comahue, Centro Regional Universitario Bariloche, San Carlos de Bariloche. Cabrera, A.L. & Willink, A. (1973) Biogeografía de América Latina, Biology Series, Monograph 13. General Secretariat of the Organization of American States, Washington, DC. Clark, K. (1992) FNOC Global elevation, terrain, and surface characteristics. Global ecosystems database, version 2.0. NOAA National Geophysical Data Center, Boulder, CO (available at http://www.ngdc.noaa.gov/seg/ecosys/ecosys.shtml). Correa, M.N. (ed.) (1969–99) Flora Patagónica 1–7. Colección Científica INTA, Buenos Aires. Dimitri, M.J. (1972) Consideraciones generales. La Región de los bosques Andino-Patagónicos, sinopsis general (ed. by M.J. Dimitri), pp. 15–26. Colección Científica INTA, Buenos Aires. Dirección Meteorológica de Chile (2002) Estadísticas de medio ambiente 1993–2002. Instituto Nacional de Estadísticas, Chile (http://www.ine.cl/canales/chile_estadistico/ estadisticas_medio_ambiente/medio_ambiente.php). Donoso-Zegers, C. (1994) Árboles nativos de Chile, guía de reconocimiento. Marisa Cúneo Ediciones, Valdivia. Donoso-Zegers, C. & Ramírez-García, C. (1994) Arbustos nativos de Chile, guía de reconocimiento. Marisa Cúneo Ediciones, Valdivia.

172

Ezcurra, C. & Brion, C. (2005) Plantas del Nahuel Huapi, catálogo de la flora vascular del Parque Nacional Nahuel Huapi, Argentina. Universidad Nacional del Comahue y Red Latinoamericana de Botánica, San Carlos de Bariloche. Gould, S.J. & Vrba, E.S. (1982) Exaptation – a missing term in the science of form. Paleobiology, 8, 4 –15. Graham, A., Gregory-Wodzicki, K.M. & Wright, K.L. (2001) Studies in Neotropical paleobotany. XV. A Mio-Pliocene palynoflora from the eastern Cordillera, Bolivia: implications for the uplift history of the Central Andes. American Journal of Botany, 88, 1545–1557. Grau, J. (1995) Aspectos geográficos de la flora de Chile. Flora de Chile (ed. by C. Marticorena and R. Rodríguez), pp. 63– 83. Universidad de Concepción, Concepción. Greenwood, D., Wilf, P., Wing, S.L. & Christophel, D.C. (2004) Paleotemperature estimation using leaf-margin analysis: is Australia different? Palaios, 19, 129–142. Gregory-Wodzicki, K. (2000) Relationships between leaf morphology and climate, Bolivia: implications for estimating palaeoclimate from fossil floras. Palaeobiology, 26, 668 – 688. Henriquez, J.M., Pisano, E. & Marticorena, C. (1995) Catálogo de la flora vascular de Magallanes (XII° Región), Chile. Anales del Instituto de la Patagonia, Serie Ciencias Naturales, 23, 5–30. Herrera, C.M. (1992) Historical effects and sorting processes as explanations for contemporary ecological patterns: character syndromes in Mediterranean woody plants. The American Naturalist, 140, 421–446. Hinojosa, L.F. & Villagrán, C. (1997) Historia de los bosques del sur de Sudamérica, I: Antecedentes paleobotánicos, geológicos y climáticos del Terciario del cono sur de América. Revista Chilena de Historia Natural, 70, 225–239. Hinojosa, L.F., Armesto, J.J. & Villagrán, C. (2006) Are Chilean coastal forests pre-Pleistocene relics? Evidence from foliar physiognomy, palaeoclimate, and phytogeography. Journal of Biogeography, 33, 331–341. Hoffman, J.A.J. (1975) Atlas climático de América del Sur: mapas de temperatura y precipitaciones medias. WMO, UNESCO, Geneva. Hoffmann, A.E. (1991) Flora silvestre de Chile, zona araucana. Fundación Claudio Gay, Santiago de Chile. Jordan, G.J. (1997) Uncertainty in palaeoclimatic reconstructions based on leaf physionomy. Australian Journal of Botany, 45, 527–547. Kowalski, E.A. (2002) Mean annual temperature estimation based on leaf morphology: a test from tropical South America. Palaeogeography, Palaeoclimatology, Palaeoecology, 188, 141– 165. Landrum, L.R. (1999) Revision of Berberis (Berberidaceae) in Chile and adjacent southern Argentina. Annals of the Missouri Botanical Garden, 86, 793–834. Lee, D.E., Lee, W.G. & Mortimer, N. (2001) Where and why have all the flowers gone? Depletion and turnover in the New Zealand Cenozoic angiosperm flora in relation to paleogeography and climate. Australian Journal of Botany, 49, 341–356. Leemans, R. & Cramer, W.P. (1992) IIASA Database for mean monthly values of temperature, precipitation, and cloudiness on a global terrestrial grid. Global ecosystems database, version

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

Leaf margins in the temperate forest flora of South America 2.0. NOAA National Geophysical Data Center, Boulder, CO (available at http://www.ngdc.noaa.gov/seg/ecosys/ecosys.shtml). Mabberley, D.J. (1993) The plant book. Cambridge University Press, Cambridge. Markgraf, V. (1991) Late Pleistocene environmental and climatic evolution in southern South America. Bamberger Geographische Schriften, 11, 271–281. Markgraf, V., McGlone, M.S. & Hope, G.S. (1995) Neogene paleoenvironmental and paleoclimatic change in southern temperate ecosystems – a southern perspective. Trends in Ecology & Evolution, 10, 143 –147. Markgraf, V., Romero, E. & Villagrán, C. (1996) History and paleoecology of South American Nothofagus forests. The ecology and biogeography of Nothofagus forests (ed. by T.T. Veblen, R.S. Hill and J. Read), pp. 354 –386. Yale University Press, New Haven, CT. McGlone, M.S. (1996) When history matters: scale, time, climate and tree diversity. Global Ecology and Biogeography Letters, 5, 309 –314. Menéndez, C.A. (1972) Paleofloras de la Patagonia. La región de los bosques Andino-Patagónicos, sinopsis general (ed. by M.J. Dimitri), pp. 129–165. Colección Científica INTA, Buenos Aires. Meudt, H.M. & Simpson, B.B. (2006) The biogeography of subalpine, austral Ourisia (Plantaginaceae) based on molecular phylogenetic evidence: South American origin and dispersal to New Zealand and Tasmania. Biological Journal of the Linnean Society, 87, 479 –513. Moore, D.M. (1983) Flora of Tierra del Fuego. Missouri Botanical Garden, St Louis, MO. Muñoz-Schick, M. (1980) Flora del Parque Nacional Puyehue. Editorial Universitaria, Santiago de Chile. Naumann, C.M. (1987) El gradiente de vegetación desde la montaña hasta la meseta en las cercanías de San Carlos de Bariloche. Licenciatura thesis, Universidad Nacional del Comahue, Argentina. Ortiz-Jaureguizar, E. & Cladera, G.A. (2006) Paleoenvironmental evolution of southern South America during the Cenozoic. Journal of Arid Environments, 66, 498 –523. Poulin, E., Palma, A.T. & Féral, J.P. (2002) Evolutionary versus ecological success in Antarctic benthic invertebrates. Trends in Ecology & Evolution, 17, 218 –222. Rapoport, E.H. & Brion, C. (1991) Malezas exóticas y plantas escapadas de cultivo del noroeste patagónico: segunda aproximación. Ediciones Imaginaria, San Carlos de Bariloche. Raven, P.H. & Axelrod, D.I. (1974) Angiosperm biogeography and past continental movements. Annals of the Missouri Botanical Garden, 61, 539 –673. Rice, W.R. (1989) Analyzing tables of statistical tests. Evolution, 43, 223 –225. Romero, E.J. (1993) South American paleofloras. Biological relationships between Africa and South America (ed. by P. Goldblatt), pp. 62–85. Yale University Press, New Haven, CT. Royer, D.L. & Wilf, P. (2006) Why do toothed leaves correlate with cold climates? Gas exchange at leaf margins provides new insights into a classic paleotemperature proxy. International Journal of Plant Sciences, 167, 11–18.

SAS (1999) SAS/ETS user’s guide, version 8. SAS Institute, Cary, NC. Servicio Meteorológico Nacional de Argentina (2004) Base de datos ORACLE. Servicio Meteorológico Nacional, Buenos Aires (available at http://www.meteofa.mil.ar/?mod=acerca&id=5). Simpson, B.B. (1983) A historical phytogeography of the high Andean flora. Revista Chilena de Historia Natural, 56, 109 –122. Sokal, R. & Rohlf, F.J. (1995) Biometry. Freeman, New York. Solbrig, O.T. (1978) The origin and floristic affinities of the South American temperate desert and semidesert regions. Evolution of desert biota (ed. by D.D. Goodall), pp. 7– 49. University of Texas Press, Austin, TX. Stranks, L. & England, P. (1997) The use of a resemblance function in the measurement of climatic parameters from the physiognomy of woody dicotyledons. Paleogeography, Paleoclimatology, Paleoecology, 131, 15–28. Veblen, T.T., Donoso, C., Kitzberger, T. & Rebertus, A.J. (1996) Ecology of southern Chilean and Argentinean Nothofagus forests. The ecology and biogeography of Nothofagus forests (ed. by T.T. Veblen, R.S. Hill and J. Read), pp. 293–353. Yale University Press, New Haven, CT. Villagrán, C. & Hinojosa, L.F. (1997) Historia de los bosques del sur de Sudamérica, II: Análisis fitogeográfico. Revista Chilena Historia Natural, 70, 241–267. Wardle, P., Ezcurra, C., Ramírez, C. & Wagstaff, S. (2001) Comparison of the flora and vegetation of the southern Andes and New Zealand. New Zealand Journal of Botany, 39, 69–108. Wiemann, M.C., Manchester, S.R., Dilcher, D.L., Hinojosa, L.F. & Wheeler, E.A. (1998) Estimation of temperature and precipitation from morphological characters of dicotyledonous leaves. American Journal of Botany, 85, 1796–1802. Wilf, P. (1997) When are leaves good thermometers? A new case for leaf margin analysis. Paleobiology, 23, 373–390. Wilf, P., Cúneo, N.R., Johnson, K.R., Hicks, J.F., Wing, S.L. & Obradovich, J.D. (2003) High plant diversity in Eocene South America: evidence from Patagonia. Science, 300, 122–125. Wilf, P., Cúneo, N.R., Smith, M.E., Singer, B.S. & Gandolfo, M.A. (2005) Eocene plant diversity at Laguna del Hunco and Río Pichileufú, Patagonia Argentina. The American Naturalist, 165, 634–650. Wing, S.L. & Greenwood, D.R. (1993) Fossils and fossil climate: the case for equable continental interiors in the Eocene. Philosophical Transactions of the Royal Society B: Biological Sciences, 341, 243–252 Wolfe, J.A. (1979) Temperature parameters of humid to mesic forests of eastern Asia and relation to forests of other regions of the northern hemisphere and Australia. United States Geological Survey Professional Paper, 1106, 1–37. Wolfe, J.A. (1993) A method of obtaining climatic parameters from leaf assemblages. US Geological Survey Bulletin, 2040, 1–71. Wolfe, J.A. (1995) Paleoclimatic estimates from Tertiary leaf assemblages. Annual Review of Earth and Planetary Sciences, 23, 119–142. Zamaloa, M.C., Gandolfo, M.A., González, C.C., Romero, E.J., Cúneo, N.R. & Wilf, P. (2006) Casuarinaceae from the Eocene of Patagonia, Argentina. International Journal of Plant Sciences, 167, 1279–1289.

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd

173

M. A. Aizen and C. Ezcurra Ziegler, A.M., Eshell, G., McAllister Rees, P.M., Rothfus, T.A., Rowley, D.B. & Sunderlin, D. (2003) Tracing the tropics across land and sea: Permian to present. Lethaia, 36, 227–254.

BIOSKETCHES Marcelo A. Aizen is a scientific researcher of the National Research Council of Argentina (CONICET) and a lecturer in ecology at the Universidad Nacional del Comahue. His research interests are in the areas of plant ecology and plant–animal interactions. Part of his research focuses on understanding the role of past historical events in moulding present ecological patterns and processes. Cecilia Ezcurra is a scientific researcher of the National Research Council of Argentina (CONICET). Her research interests are in the areas of plant taxonomy, evolution and biogeography. She is particularly interested in the different origins of plant taxa that characterize the extant flora of southern South America. Editor: Matt McGlone

174

SUPPLEMENTARY MATERIAL The following supplementary material is available for this article: Appendix S1 List of 167 woody dicotyledonous species present in the temperate forest of South America with information on leaf margins, growth form, latitudinal extent and their presence/ absence in the five local floras described in Materials and Methods. Appendix S2 List of 34 exotic woody dicotyledonous species naturalized in Nahuel Huapi National Park, Argentina, with information on leaf margins. This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/j.14668238.2007.00350.x (This link will take you to the article abstract). Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

© 2007 The Authors Global Ecology and Biogeography, 17, 164–174, Journal compilation © 2007 Blackwell Publishing Ltd