Nothofagus pumilio and N. antarctica: The Most Widely Distributed and Cold-Tolerant Southern Beeches in Patagonia

Abstract

Nothofagus pumilio and N. antarctica are main components of the Andean Patagonian Temperate forests that occur along all the environmental gradients that characterize the Patagonian region. In this chapter, their main botanical features and natural ranges are presented, as well as the state of the art in genetic studies. Phylogeographic patterns based on chloroplast DNA help inferring Pleistocene refugia location and colonization pathways. Both species share haplotypes, and interspecific hybridization occurs naturally. A genetic structure analysis through genomic microsatellites shows latitudinal trends, species admixture, and contact zones. Selective logging evidences an effect on patterns of genetic diversity in N. pumilio. The definition of genetic zones contributes to the conservation and management of the genetic resources of both species. In addition, quantitative genetic variation and phenotypic plasticity are analyzed under nursery and field conditions, in juvenile morphological, phenological, physiological, and growth rhythms traits. Ecotypic variation was detected along longitudinal and altitudinal gradients of N. pumilio. Finally, the domestication and development of low intensity breeding strategies are presented, including a provenance trial network and seed production areas of N. pumilio, and progeny seed orchards of N. antarctica.

1 Main Characteristics, Ecological Features, and Distributional Ranges

Nothofagus pumilio (Poepp. & Endl.) Krasser, known as lenga, and Nothofagus antarctica (G. Forster) Oerster, commonly called ñire, are tree species of the temperate forests of southern South America, occurring in temperate-rainy forest districts, in the subalpine transition area, and in poorly drained sites (Veblen et al. 1996). Their natural distributions in Chile and Argentina mainly correspond to the Cordillera de los Andes, from 36° S to the south of Tierra del Fuego archipelago (55° S). In Chile, N. pumilio also inhabits the Coastal Range (Cordillera de Nahuelbuta) where it grows associated with Araucaria araucana forests above 1400 m asl. In turn, N. antarctica is distributed in the Central Depression of Chile, from Valdivia toward austral latitudes. Throughout their entire distribution in Argentina, lenga and ñire forests co-occur and overlap to a greater or lesser extent. This shared geographical distribution covers approximately 18° of latitude (2200 km of extension) and is the widest among the South American Nothofagus. Recently, the national forest inventory of Argentina (CIEFAP and MAyDS 2016) revealed a total of 1,595,661 ha and 864,148 ha of lenga and ñire forests, respectively. The National Park Administration of Argentina protects a portion of all this forest surface (34% and 15%, respectively), belonging the rest to provincial jurisdictions.

The two species conform an altitudinal ecological gradient along their natural distribution, with N. pumilio dominating at higher elevations and N. antarctica at lower sites. Nothofagus pumilio usually forms large masses of pure stands associated with climax forests (late-successional species) but also in co-dominance with Nothofagus betuloides (in the south) or Araucaria araucana (in the north). Individual trees can be taller than 35 m (Tortorelli 1956) (Fig. 5.1a) and grow in environments with deep, well-drained soils; while reaching the timberline, they tolerate low temperatures and frosts as a shrub. On the other hand, N. antarctica is the species with the widest ecological plasticity and phenotypic variation (Ramírez et al. 1997) among their South American congeners. It occurs in cold humid valleys with heavy clay soils, in peat bogs, and in rocky and xeric sites like the Patagonian steppe, forming monospecific masses of shrubs or small trees (Fig. 5.1b).

Fig. 5.1

Species features. (a) Nothofagus pumilio, (b) Nothofagus antarctica; (1) seeds, (2) male flowers, (3) leaves, (4) arboreal morphotype, (5) view of a forest. (Photos: C. Soliani, M. J. Pastorino, V. Rusch; A2, B2 extracted from Giménez Gowland 2002)

Nothofagus antarctica is considered a pioneer and resprouting species and has a great capacity for clonal reproduction (Premoli and Steinke 2008), while N. pumilio can only reproduce generatively. Clonal reproduction in ñire is suggested as an adaptation to recurrent disturbances, among which fires are the most common in the region.

Both are related species included in the subgenus Nothofagus, one of the four defined within the Nothofagaceae family (Hill and Jordan 1993). Phylogenetic analyses, including morphology, conserved DNA sequences, and fossils, agreed in a common ancestor among Nothofagus betuloides and N. pumilio/N. antarctica sister species (Manos 1997; Sauquet et al. 2012). Recently, it has been suggested that N. pumilio would be the ancestral species within the subgenus (Acosta and Premoli 2010). Extant species included in the clade shared the pollen type, nominated equivalently as “N. fusca Type b” (Manos 1997) or “dombeyi” (Villagrán et al. 1995; Heusser et al. 1999) all of them are cold-tolerant species.

As deciduous species, their dormant buds during winter determine a growing season restricted to spring and summer, whose length varies depending on available resources; particularly drought stress during summer is among the most critical factors. Pollen and seed dispersion is mediated by wind. Most commonly, seeds reach short distances from the mother tree (Rusch 1993). Seed production is not regular over consecutive seasons, being years of greater production associated with greater germination power and viability (masting) (Donoso 1993). Variability in seed production and seed quality was suggested to be linked to stand conditions, e.g., secondary N. antarctica forests have been observed at 51° S to have better seed quality (Soler Esteban et al. 2010), and also differences have been reported between pure and mixed N. pumilio forests (Toro Manríquez et al. 2016). Moreover, in these species, seeds have a stage of reduced viability and do not form persistent seed banks (Cuevas and Arroyo 1999). Seedling recruitment and survival are crucial stages for regeneration, varying in number of established individuals year after year, being tightly linked to the microclimate conditions within natural forests (e.g., Soler et al. 2013; Bahamonde et al. 2018). Regeneration dynamics is the result of large-scale disturbances, which cause the replacement of complete stands (e.g. mass removal on steep slopes), or small-scale disturbances, mainly the gap dynamics, which consist of the natural falling of over-mature trees; both involve a massive recruitment related to canopy opening (Veblen et al. 1996; Donoso 2006).

Temporal overlapping of flower maturity during reproductive stages favors the occurrence of natural interspecific hybridization between these species, which is a widely reported phenomenon among South American Nothofagus (see Chaps. 3 and 4). The natural hybrid between N. pumilio and N. antarctica presents intermediate characteristics between both species (i.e., bark roughness and color, stem straightness, crown form). It has been suggested that N. antarctica acts predominantly as pollen receptor (Acosta and Premoli 2010). The introgression of the chloroplast genome from N. pumilio to N. antarctica, due to repeated interspecific crossings and/or hybrid-parent backcrossing, would have occurred more frequently during unfavorable climatic periods (e.g., glaciations, Palmé et al. 2004; Heuertz et al. 2006) or during forest recovering (e.g., postglacial recolonization), constituting an additional source of variation.

The climatic changes that occurred during the Quaternary modified the areas covered by forests in Patagonia and have strongly impacted in the population dynamics, possibly restricting gene flow, isolating them geographically, or even hindering their chances of regeneration. This regional disturbance left a trace on the distribution of genetic variation. At a narrower level, other processes could have shaped the population structure, such as the mating system or, more recently, anthropogenic disturbances. The genetic variability of these species, shaped along their evolutionary history, should be preserved in order to ensure their adaptability and, finally, their persistence.

2 Phylogeography: Inspecting Nothofagus Evolutionary History Through Chloroplast DNA

Climatic changes during the Quaternary imposed a great selection pressure along the distributional range of forest species. After ice advance, the remaining patches of forests withstood adverse climatic conditions. In Patagonia, glaciations occurred from the Late Miocene to the Pleistocene, being the Great Patagonian Glaciation (GPG ; about 1 million years before present [M years BP]) the maximum expansion of ice in extra-Andean Patagonia (Flint and Fidalgo 1969; Rabassa et al. 2005). The Last Glacial Maximum (LGM) occurred about 18,000–20,000 years BP (Porter 1981), when the ice covered the Patagonian plains beyond the mountain range (Rabassa and Clapperton 1990; Glasser et al. 2008). However, several ice-free areas remained (Markgraf et al. 1995) and constituted refugia for vegetation, and some became the center of expansion of the biota after the ice retreated. Multiple evidence suggests a latitudinal trend in the type of glaciations during the LGM: valley-type glaciations characterized the north, whereas continuous ice layers covered the southern region (Glasser et al. 2008). A transitional zone at mid-latitudes (42° S–44° S) was established. The predominance of westerlies determined intermediate climatic conditions at these latitudes by the early Holocene (10,000–8500 years BP) (Markgraf et al. 2003), i.e., lack of seasonality and a drier and warmer environment (Manzini et al. 2008). Reinforcing this, three different zones based on a palynological reconstruction representing paleo-climates (Markgraf et al. 1996) were identified (north of 43° S, between 43° and 51° S, and south of 51° S).

The distribution of chloroplast genetic variation helps to thoroughly comprehend how forests were modeled after glaciation. Considering its slow evolution rate as well as the uniparental (maternal) inheritance, chloroplast DNA is a useful tool tracing the effective dispersal of seeds. Polymorphisms from cpDNA non-coding regions were screened in 40 Argentinean populations of N. pumilio and N. antarctica (Soliani et al. 2012; Table 5.1) and used to define haplotypes, coming from the following restricted regions: trnD-trnT/HinfI, trnC-trnD/taqI, and atpH-atpI/HinfI. Point mutations in the restriction site and indels (insertion/deletions) allowed identifying 9 and 13 haplotypes in N. pumilio and N. antarctica, respectively (based on different combinations of length variants). Eight haplotypes were shared among the species, with N. antarctica being the most variable with five unique haplotypes.

Table 5.1 Geographic location of coupled Nothofagus pumilio and Nothofagus antarctica sampled populations ordered from north to south

A genetic diversity trend along latitude, decreasing in both species from north to south, as well as a significant phylogeographic structure between the two main groups of populations and haplotypes (north and south of 42° S), evidenced regional footprints of glaciations (Fig. 5.2). The hypothesis of multiple glacial refugia is supported by these results (Premoli et al. 2000; Marchelli and Gallo 2006; Sérsic et al. 2011). In addition, a meeting area where migration routes could have encounter was proposed (ca. 42–43° S), which agrees with stratigraphic and palynological evidence. A geographical segregation of genetic lineages was identified (Mathiasen and Premoli 2010; Soliani et al. 2012), like in other widely distributed species of the region, such as Austrocedrus chilensis (Pastorino and Gallo 2002) and Pilgerodendron uviferum (Premoli et al. 2002). Alternatively, the great divergence between haplogroups was interpreted as isolated forest patches due to the settlement of pre-Quaternary depression areas (paleobasins) (Premoli et al. 2012).

Fig. 5.2

Geographic location of sampled populations. Haplotype diversity is expressed by different colors in each species and with different tones for each population (darker colors represent higher diversity levels). Population codes have correspondence with Table 5.1. From Soliani et al. (2012)

Both species presented similar levels of average within-population gene diversity (h_s), total genetic diversity (h_t), and gene differentiation based on frequency (G_ST) and ordered alleles (N_ST) (Table 5.2). Cryptic refugia might be inferred from population allelic richness (A_R), a parameter independent from population size (Widmer and Lexer 2001) and with a significant value in conservation decisions (Petit et al. 1998). Northern populations (40° S) (4, IV) harbor the highest diversity in both species; at mid-latitudes (42–43° S), three populations of lenga (a, 12, 14) and two of ñire (XIII, XIV) showed a higher allelic richness. In southern Patagonia (54° S), one lenga population (XIX) was the most diverse. Trends of genetic diversity clearly follow the geographical latitude.

Table 5.2 Average values of within-population gene diversity (h_s), total genetic diversity (h_t), and gene differentiation in all populations for unordered alleles (G_st) and for ordered alleles (N_ST), for the analyzed populations of each species

Haplotype sharing (cpDNA or mtDNA) among closely related species that hybridize naturally and occur in sympatry is very common (Rieseberg and Soltis 1991), signaling population variation. Introgression in this Nothofagus species complex (IG = 0.90; Soliani et al. 2012) could be occurring due to interspecific gene flow and backcrossing offspring. A similar geographical pattern for haplotype distribution in both species supports the idea of recent or at least postglacial hybridization. Ecological features also support hybridization in our species complex: flowering phenology and pollen release overlap (González et al. 2006), although other pre- and even post-zygotic barriers can also occur. The predominance of westerlies in the region and the altitudinal ecological gradient formed by both species might favor N. antarctica acting as the mother tree. However, the lack of evidence supporting backcrosses of hybrids toward N. antarctica as the parental species and the scarce genetic information about F1 and F2 generations do not allow to thoroughly conclude about the directionality of the hybridization. Moreover, since individuals from both species coexist in some places, hybridization and directionality of the crosses (hybrids to parental species) could vary according to the relative abundance of the taxa (Lepais et al. 2009) or particular conditions of the site where they co-occur (e.g., Heuertz et al. 2006).

3 Genetic Structure at Nuclear Markers Across Species Ranges

3.1 Latitudinal Trends, Species Admixture, and the Identification of a Contact Zone

In addition to the phylogeographical analysis presented in the previous section, the same 20 pairs of populations of sympatric interspecific natural forests were studied with seven microsatellite loci in N. pumilio and six in N. antarctica (Soliani et al. 2010) (five shared by both species). Besides, one sample from a single N. antarctica site (42° S, Cholila (00) Table 5.1, Fig. 5.2) that showed admixture of maternal lineages (Pastorino et al. 2009; Soliani et al. 2012) was also analyzed. All genetic diversity parameters evaluated were significantly higher in N. antarctica than in N. pumilio (Wilcoxon paired test: N_e = 3.55 vs. N_e = 3.16, p = 0.022; A_R = 4.80 vs. A_R = 4.45, p = 0.015; H_E = 0.656 vs. H_E = 0.598, p = 0.029; H_O = 0.520 vs. H_O = 0.450, p = 0.017). The hierarchical AMOVA showed that the proportion of genetic variance partitioned between the two species was 15% for nSSRs (φ_RT = 0.15, p = 0.001), much greater than the percentage for cpDNA variation (φ_RT = 0.012, p = 1). Standardized genetic differentiation between the species was also higher for nSSRs (G’_ST = 0.335) with respect to cpDNA (G’_ST = 0.061) data.

A latitudinal trend was revealed in both species, which is probably related to the impact of past glaciations. Higher allelic richness and gene diversity values were found in N. pumilio populations located around 37°–42° S (populations 2, 3, 4, 6, and 8), whereas at about 40°–43° S for N. antarctica (V, 00, VIII, and IX). A decrease in A_R and H_E toward southern populations (southward 42° S) was more evident in N. antarctica than in N. pumilio, in agreement with previous results based on cpDNA data (Soliani et al. 2012). Genetic differentiation was significant in both taxa and slightly higher in N. pumilio (F_ST = 0.094, p = 0.001) than in N. antarctica (F_ST = 0.083, p = 0.001). The standardized differentiation (G’_ST) was higher but similar in both species (G’_ST = 0.296 and G’_ST = 0.303, respectively).

A clear separation between both species was found with a Bayesian clustering analysis (STRUCTURE; Pritchard et al. 2000) of the 41 populations (ΔK = 2 indicated the optimal number of groups) (Fig. 5.3). Some populations presented a high level of admixture, suggesting hybridization and even introgression through backcrosses, e.g., populations 8 (42°59′ S, 71°11′ W) and 9 (43°04′ S, 71°35′ W) in N. pumilio and X (43°51′ S, 71°33′ W), IX (43°04′ S, 71°34′ W), XII (43°50′ S, 70°45′ W), and XIII (44°51′ S, 71°38′ W) in N. antarctica. Evidence of interspecific gene flow was also observed when cluster partitioning increased (e.g., at K = 4 and K = 5).

Fig. 5.3

Inferred genetic structure at the between-species level (K = 2 to K = 5, STRUCTURE, Pritchard et al. 2000) (a) and past demography scenarios tested with DIYABC (b). Populations are ordered from north to south and are coded according to Table 5.1 and Fig. 5.2 (except populations 10 and X that were added for the genotyping with microsatellite markers). Representation of clusters is indicated by the same colors in each species; figure was split in two for better comprehension and legibility, from Soliani et al. (2015)

Through a coalescent model, the putative origin of population genetic variation of each species at intermediate latitudes was inferred (approximate Bayesian computation (ABC) in DIYABC v1.0.4.39; Cornuet et al. 2008, 2010). Species divergence was estimated around 302,500 years BP. Then, a recent species admixture could have occurred ~18,950 years BP (50 years/generation time) across individuals from the 8 (42° S) and 9 (43° S) populations (ABC1, scenario 1; Fig. 5.3). Although hybridization could have determined N. pumilio population variation, it seems not to be the most probable explanation of the N. antarctica sympatric populations. In this species, divergence from an ancestral population to around 26,700 BP could have been the origin of the current variation (ABC2, scenario 4; Fig. 5.3). The settlement of a hybrid zone at intermediate latitudes (42° ~ 43° S, 8–9 and VIII–IX populations) could have been facilitated by niche overlapping and a flowering lag along altitude (Rusch 1993), an aspect that fosters hybridization in contact areas. Accordingly, the high proportion of admixed individuals, mainly detected around 42–44° S, suggests a tight correspondence between the frequency and the geographic location of hybridization.

Colonization from multiple refugia, as reported before, implied population bottlenecks, founding events, and admixture of genetic lineages. The convergence of isolated, independently evolving, intra-specific lineages during colonization might increase genetic diversity (Comps et al. 2001; Alberto et al. 2008; Durand et al. 2009) due to their admixture (e.g., Vendramin et al. 1998; Petit et al. 2003). The secondary contact zones that are therefore established constitute genetic reservoirs relevant for their conservation and might enrich sources of material in ex situ breeding programs (Petit et al. 2003; Grivet et al. 2008).

The identification of a contact zone where the northern lineages mixed with the southern lineages was inferred for N. pumilio and N. antarctica (Soliani et al. 2015). These results are in agreement with previously reported evidence of contact zones for Patagonian taxa, e.g., in fishes (Zemlak et al. 2008), forest trees (Pastorino et al. 2009), and herbs (Cosacov et al. 2010; Sérsic et al. 2011). The high allelic richness, low level of inbreeding, and a lack of evidence of genetic bottlenecks in populations around the contact zone add support to the meeting of colonization routes from northern and southern refugia (Soliani et al. 2015). In addition, the contact zone might be the result of immigrants from local or nearby refugia that remained in unglaciated areas at central latitudes. The described patterns of genetic variation were taken into consideration for the definition of preliminary operational genetic management units of N. pumilio and N. antarctica in Argentina (see next section). The identification of at least one population with high level of genetic diversity within each unit could be considered as base material for future breeding programs.

3.2 Impact of Selective Logging on Patterns of Genetic Diversity: A Case Study in Nothofagus pumilio

Logging is one of the human activities that has impacted on the natural evolution of the forest, by altering the genetic diversity and structure of main species like trees (Rajendra et al. 2014). In particular, the removing of trees and the impoverishment of a forest could lead to within-population changes in genetic variation and diversity, which is the key to adaptation (El-Kassaby et al. 2003; Finkeldey and Ziehe 2004). Signs of impact could be a decrease in allelic richness or modifications in heterozygote proportions, a reduction in allele frequencies, or loss of variants between the adult cohort and its regeneration (Cornuet and Liukart 1996; Rajora et al. 2000). Then, erosive forces (i.e., genetic drift, selection) might affect the remnant population. Logging could also affect the spatial genetic structure, i.e., the amount and distribution of genetic variation between and within local populations and individuals of a species, with consequences in regeneration recruitment.

In Patagonian natural forests, selective extraction in a high grading management system was implemented over many decades (Bava and Rechene 2004; Bava et al. 2006; González et al. 2006). The removal of best-featured individual trees (stem straightness and best sanitary conditions) was the most frequent technique employed, which may be expected to induce changes in allelic richness or modifications in the spatial distribution of alleles. Because of its excellent wood properties (high quality, long-time durability; González et al. 2006), N. pumilio has historically been one of the most exploited native species in Patagonia, threatening its populations.

A comparison between logged and non-logged stands of N. pumilio was made through the estimation of genetic diversity of adults and their regeneration by means of nuclear microsatellites (Soliani et al. 2016). Three pairs of stands from areas around 42° S, which were reported as strongly intervened and degraded after a high grading logging (Bava et al. 2006), were sampled. A slight decrease in allelic richness (A_R) and a lower number of rare alleles (frequencies ≤5%) were observed in adults from the logged stand with respect to adults from the non-logged-stand in one population (Lago Guacho (11), Table 5.3). However, an unexpected result with the opposite trend was found in the other two studied populations (Huemules (6) and Lago Engaño (10)), with more alleles in the logged stand. A tendency toward a decrease in number or frequency of alleles was detected in old growth and remnant adults of managed forests in all populations. These results could be a sign of the impact of logging on the adult cohorts. The regeneration cohort from the logged sites had greater allelic richness than non-intervened forests, which could not be attributed unequivocally to logging but to other factors not contemplated by the study.

Table 5.3 Location and genetic characterization of sampling sites of Nothofagus pumilio representing stands with selective extraction of individuals (LOGG) and natural forest (NONL)

Even though a genetically impoverished population is expected after logging under a high grading system, no signs of recent bottlenecks were observed (Bottleneck software, Cornuet and Liukart 1996). The proportion of genetic variance partitioned between the two treatments (i.e., logged vs non-logged) was low but significant in only one population located at 42° 50′ S (5%, F_RT = 0.048; p = 0.001). Spatial autocorrelation and heterogeneity tests between management treatments were non-significant in the populations, making inviable to assess the true impact of management on spatial variation. The effects of management varied widely according to the type of treatment, having multiple effects (positive, negative, or neutral) on the genetic diversity and the mating system. Additional factors might be playing a role in the final determination of genetic variation, i.e., intensity of management, time elapsed since the last intervention, recent practices, and presence of livestock.

4 Genetic Zones: On How Molecular Tools Can Contribute to the Conservation and Management of Forest Resources

In widely distributed species conformed by hundreds of natural populations, as many tree species, it is unfeasible the management from a genetic perspective of each population separately. The definition of operational genetic management units (OGMU) overcomes the limitation of making decisions at the level of single populations. Knowledge on the genetic pool of a species gained by sampling Mendelian populations can contribute to the application of specific management actions to groups of populations. Those management decisions are commonly related to the conservation and use of genetic resources and involve actions such as the planning of reforestation or restoration programs. Properly designed strategies will focus on the preservation of the local provenance in order to avoid maladaptation and genetic contamination. In this regard, genetic zones’ (GZs) delineation is the first step toward OGMUs’ definition. GZs are known as genetically homogeneous regions (Bucci and Vendramin 2000) within which genetic material can be moved with minimum risk of altering the genetic constitution of the local and nearby populations (McKay et al. 2005). At the same time, GZs would represent discernible genetic pools that are desired to be conserved because of its distinctive genetic attributes. To accomplish this purpose, both a genetic inventory and a representation of the natural distributional range (mapped geographic area) of the species are crucial requirements. In order to classify the genetic information, a hierarchical clustering of sampled populations is useful to group populations sharing the same genetic background. Among the available methods, Bayesian clustering represents the most accurate and reliable.

Genetic analyses using molecular markers provide information based on the neutral evolution of the populations (i.e., not affected by selection) for identifying GZs. This type of markers could offer information about demographic history of a population and allow identifying the genetic structure of a set of populations modeled by historical processes. Still, adaptive traits could be evaluated by quantitative genetic studies and, combined with neutral marker analyses, might provide a complete assessment of the genetic resources. Thus, GZs constitute a first step toward the definition of provenance regions (PRs) , which ensure not only the conservation of evolutionary significant variants related to its life history (Crandall et al. 2000; de Guia and Saitoh 2007) but also the ecological viability and local adaptation of populations. The final goal of OGMUs’ definition is to conserve relevant ecological entities that represent short- and long-term genetic processes (Fraser and Bernatchez 2001).

Population genetic diversity of a species is shaped in terms of evolutionary time as the product of the interplay between enhancing forces (i.e., gene flow, generation of new mutations, hybridization) and erosive forces (i.e., genetic drift, selection). Highly diverse populations should be prioritized for conservation, since they probably expose a better response to environmental changes that might risk their persistence (e.g., extreme climatic events, biological invasions). Unique variants (i.e., infrequent genotypes, private alleles, or haplotypes) or geographically restricted variants are also relevant to face new and unpredictable environments. It is crucial to understand the distribution of the genetic diversity along the natural species range since it reveals the degree of interconnection among them. The variation level of each population might be related to past evolutionary footprints, the mating system, or current genetic processes. Whatever the cause, its identification could assist managers when collection of propagation material is needed.

Based on molecular markers, standardize allelic richness SAR (Marchelli et al. 2017) was obtained for Nothofagus pumilio and N. antarctica populations, and the more diverse were identified (1–5, 8, 12, 14, 16, and 19 and II, IV, IX, XIII, XIV, XV, and XX, respectively (Table 5.1); Soliani et al. 2017). Considering the emergent patterns of genetic diversity and structure, but also the prevailing topography and the geographic isolation of the populations, preliminary genetic zones were proposed for both species. Several GZs include large areas of the natural distribution of the target species, demanding the screening of more and new populations to refine the divisions. In N. pumilio case, a re-delineation of GZs has being carried out by adding 14 new populations to the previous set and thus covering almost completely the main distribution range of this species (Mattera et al., in press). As a result of this, 18 GZs were defined in the species within four main regions along Patagonia (36–42° S, 42–44° S, 44–51° S, and 55° S corresponding to populations included in the island of Tierra del Fuego). In N. antarctica, nine GZs were proposed (Fig. 5.4) based on the genetic data from the 21 populations presented in Table 5.1. From north to south, they are North, Tromen, Central, Chubut, Río Grande, LGM East, LGM West, South, and Tierra del Fuego.

Fig. 5.4

Nothofagus antarctica genetic zones (GZ) proposed for Argentinean populations (Soliani et al. 2017). Colored dots represent highly diverse populations identified based on chloroplast and nuclear markers

5 Adaptive Genetic Variation of N. pumilio: Assessment of Juvenile Traits in Common Garden Experiments

Genetic diversity is a population intrinsic property, the basis on which the evolutionary force of natural selection operates. The geographic variation of quantitative traits displayed by individuals in situ can be associated with their distribution in environmental gradients and inferred as a result of adaptation processes, thus providing a first approach to the analysis of the genetic diversity of a species. A generalized interpretation of the genotype-environment relationship will make it possible to discern whether the type of variation is clinal, i.e., gradually changing characters, or ecotypical, that is, variation occurring in discrete changes of the character (it does not correspond to a gradual change of the environmental conditions). However, the establishment of common garden experiments is needed to distinguish between environmental and genetic effects. To accomplish this, breeding the bulk progeny of several natural populations in a common location is a further step on the way to unravel the genetic patterns of a species. Yet, a next step can be advanced if we keep the identification of the parental relationships of those progenies assayed. The analysis of variation of quantitative traits in progeny trials allows us to estimate the genetic variation within and between populations.

Out of the two species considered in this chapter, N. pumilio is the one with more breeding potential, due to its better productivity and forestry shape. Additionally, the interest on cultivating these species is higher in N. pumilio due to the fact that it lacks the ability to resprout from the stumps, thus making its cultivation for restoration purposes more necessary than in N. antarctica. Consequently, there are more advances in quantitative genetics of N. pumilio, which will be presented below following the three levels of analysis: (1) in situ geographic variation, (2) among populations’ variation in common garden trials, and (3) genetic variation by means of progeny trials.

5.1 In Situ Geographic Variation

Variation in seed traits across environmental gradients could evidence adaptive processes. Aiming to analyze this possibility, Mondino (2014) collected seeds in 14 natural populations of N. pumilio in the Province of Chubut (a small portion of the wide Argentine distribution of the species), representing a latitudinal range of 2 degrees and a precipitation range from 400 to 1000 mm of annual average. The mean weight of 100 seeds was 1.36 g, and differences among populations were shown by means of an ANOVA for weight, width, length, and length/width ratio of the seeds (80.6% of the total variance was explained by population in the weight trait). However, it was not possible to recognize a consistent pattern associated with the two environmental characters considered (different interactions were shown between precipitation and latitude factors depending on the trait).

Based on a new seed collection in six populations, now representing three altitudinal levels in two sites (Mondino 2014), interaction between altitude and sites was shown for 100-seed weight (the highest population produced the heaviest seeds in one site, while the opposite was verified in the other). A different picture was found for seed-shape traits, where those of the highest altitude were the narrowest in both sites.

5.2 Variation Among Populations in Common Garden Trials

Annual growth synchronization with climate usually reflects a compromise between tolerance to cold and optimization of growth. In trees of temperate climates, the initiation and cessation of primary growth determine the duration of the stem elongation period and indicate the transition between resistant and frost-vulnerable stages. Therefore, the synchronization is critical for both optimal biomass production and fitness. Because of its wide latitudinal range, N. pumilio is expected to show deep phenological differences among populations.

Torres et al. (2017) carried out a population variation study concerning the bud burst process, including the entire latitudinal range of the species. They installed a high-density provenance trial in Trevelin Forest Station of INTA (43° 7′ 17″S, 71° 33′ 41″W, 390 m asl) representing 12 Argentine populations (4 blocks, 24 seedlings per block, N = 1152). Budburst phenology was registered in all seedlings every 3 days during 3 months at the beginning of their second growing season, according to five phenophases as described in Chap. 3 for Nothofagus alpina. The population factor had a significant effect on the day of the year to reach phenophase 3 (completely open buds) and explained 21% of the total variance. However, a geographically defined pattern was not found.

Similarly, Mondino (2014) studied the effect of the main environmental gradients on the phenology of the growth process. First, he installed a greenhouse provenance trial in Trevelin Forest Station with 27 potted seedlings (in a three-block design) from each of 11 Argentine populations from the Province of Chubut, representing a two degrees’ latitudinal range and a 600 mm precipitation range. At the beginning of the second growing season, the height of each seedling was measured every 10 days in order to build individual growth curves fitting a Boltzmann sigmoidal model by regression. Growth initiation (t₁₀: time to reach 10% of the season growth) and cessation (t₉₀: time to reach 90% of the season growth) were estimated in days since the first changes in the budburst phenology (bud swelling). Interaction between latitude and precipitation factors, as well as significance of the precipitation level, was not detected. On the other hand, the latitude factor was significant for t₉₀ and for the duration of the growing season (Dur = t₉₀ – t₁₀) (northern populations ceased growing later and presented a longer growing period).

In a new study, Mondino et al. (2019) collected seeds from six populations from three altitudinal levels (200 m of difference between each other) in two sites of similar latitude and precipitation regime and installed a greenhouse provenance trial similar to the previous one. Again, the height of each seedling was measured every 10 days during the second growing season, and individual growth curves were regressed. The average final height of the entire trial was 29.2 cm. Interaction between altitude and site and differences between sites were not significant for any of the traits considered. On the contrary, significant differences among altitudinal levels were shown for several variables (t₁₀ and Dur among them), and in all of them, the significance was due to the difference between the highest-level plants and the other two undifferentiated. Thus, in both sites, seedlings from the uppermost populations initiated growth later, had a shorter period of growth, and displayed a steeper growth curve than those from the other two altitudinal levels.

Nothofagus pumilio builds the treeline of the Subantarctic forests in Patagonia. It is possible to find different morphotypes coexisting at the maximum altitude reached by the species in a 50 m altitudinal range strip: arboreal, shrubby, and crawling (Fig. 5.5). In order to analyze whether these deep phenotypic differences are genetically determined or are the expression of the phenotypic plasticity according to the reaction norm of architectural and growing traits, two different studies were carried out. In a first case study (Mondino 2014), seeds from 20 trees corresponding to each of the three morphotypes were collected in two nearby sites at 42° 50′ S. The altitudinal ranges of both sites were around 1450 m asl and around 1550 m asl, respectively, that is, an intermediate altitude between the low-altitude good shaped forest and the high-altitude Krummholtz (Fig. 5.5). Seedlings were produced and a greenhouse trial was installed with potted plants arranged in a three-block design (9 seedlings per block; N = 162) to test morphotype and site differences by means of ANOVA. Plant growth rhythm was characterized by measuring the height of each seedling every 10 days during the second vegetative period and regressing 162 individual growth curves from these data (Boltzmann model). Then, several variables were derived: initial (H₀) height, time to growth initiation (t₁₀), cessation (t₉₀) and to reach 50% of growth (t₅₀), total length of the growing period (Dur), and the maximum growth rate through the slope of the curves (S) (Fig. 5.5). Plant architecture at the end of the second season was characterized by measuring collar diameter (d), total height (H), H/d ratio, number of buds in the main stem (NBu), apical meristem necrosis (AN, binary trait), forking (F, binary trait), first- and second-order branch length (BL1, BL2), number of branches of the first and second order (NB1, NB2), apical dominance (H/BL1), branchiness (H/NB1), internode length (H/NBu), and stem decumbent habit (Dh, binary trait). ANOVA showed differences between morphotypes only in the following traits: H, S, d, H/NBu, and H/d. This was surprising, particularly because the most vigorous seedlings corresponded to the crawling morphotype (Table 5.4). Thus, the different growth habits found in the nature among plants close to each other at that intermediate altitude seem to be a plastic response in most of the juvenile traits surveyed, especially in those that were expected to present a difference among the three morphotypes. Perhaps this evaluation has been too early to find a genetic cue.

Fig. 5.5

Nothofagus pumilio morphotypes coexisting at the maximum altitude reached by the species in a 50 m altitudinal range strip: arboreal, shrubby, and crawling. Bottom right box: effect of altitude of origin for the variables: (A) mean time (t₅₀), (B) form of the growth curve (S), (C) time of onset of growth (t₁₀), and (D) duration of the growth period (Dur). Different letters indicate significant differences with a P < 0.05. Altitudinal floors: low, white; medium, light gray; high, dark gray. (Photos: Víctor A. Mondino)

Table 5.4 Average values and variation coefficients of analyzed architectural variables in three different morphotypes of Nothofagus pumilio growing in a mountain slope: arboreal, shrubby, and crawling (see Fig. 5.5) (from Mondino 2014)

In a second case study (Soliani and Aparicio, 2020), seedlings were produced with seeds collected from two stands of contrasting altitude located at 41° 15′ S: (L) low altitude, arboreal morphotype (1200 m asl), and (H) high altitude, shrubby morphotype (1560 m asl). A greenhouse common garden trial was established with ~100 potted seedlings from each stand. Once seedlings had entered their first dormant period (June 15), their height (h₀) from the collar to the top of the most distal bud of the main axis (A1) and their diameter (d₀) at the base of the most distal bud were recorded. In the second year growing period, the cumulative length of the main axis was measured, six times along 172 days, finishing when the plants had reached asymptotic growth (December 12). With those data, individual growth curves were fitted using a sigmoidal equation (Boltzmann model), and traits were calculated as before t₁₀, t₉₀, Dur, and the maximum growth rate grmax (mm.day⁻¹). Budburst phenology of the main axis terminal bud (BB) was registered at the second and third growing seasons (considering four phenophases). Day of the year until phenophase 3 (at least one leaf is unfolded and spreading) was registered. Architectural traits were recorded during the second growing period: dominance (dom₁) of the main axis and delayed (dlb) and/or immediate (ilb) lateral branching of second-order branches (A2). A whole-plant size and form characterization was performed when plants were 6 years old, by obtaining the total height (h6), number of co-dominant axes (nAx) (which was also used to calculate the dominance of the main axis: (dom6)), length of the largest co-dominant axis (lAx), total number of secondary branches (nA2br), and length of the largest branch (lBr) carried by the largest co-dominant axis. With these measures, two continuous traits were constructed: branchiness (Br = nA2br/(lAx)/10) and slenderness (Sl = lAx/(lBr/nAx)). Together with h6, Br and Sl describe plant architecture within a continuum from shrubby, multi-stemmed to slender, single-stem growth forms.

The majority of the variables of phenology, growth rhythm, and architecture traits showed significantly different means between stands H and L. Although the individual age-age correlations for height (h₀, h₆) and dominance (dom₁, dom₆) of the main axis were not significant, the plants from stand H were on average consistently shorter and had lower dominance of the main axis. Besides, they had a later budburst phenology and growth initiation, i.e., an average temporal lag of 6 days, similar to the findings of Premoli et al. (2007), and were more branched and less slender than plants from stand L.

From the whole-plant analysis , three archetypes representing the juvenile growth form were retained: archetype 1 represents plants that were small and typically formed by several (2 to 4) co-dominant axes, relatively dense in lateral branches. On the other hand, archetype 3 reflects large single-stemmed, slender plants, with low branchiness. Archetype 2 represents highly branched plants, which, although in general had one main stem, was short (Fig. 5.6). The frequency of plants resembling each archetype (according to their nearest Euclidean distance) was not independent of the provenance stand. Within the group of plants from stand H, 74.5% were phenotypically closer to archetype 1, 15.5% to archetype 3, and 10% to archetype 2. In stand L, the proportions of plants resembling archetypes 1 and 3 were similar (48.7% and 50%), while those closer to archetype 2 were only 1.3%.

Fig. 5.6

Archetypal growth habits in 6-year-old plants of Nothofagus pumilio from the low and high extremes of an altitudinal gradient, according to three size and architecture traits: height (h6), branchiness (Br), and slenderness (Sl). From Soliani and Aparicio (2020)

The juvenile growth habit of N. pumilio differed between the lower and higher extremes of a 360 m elevation gradient. This should be attributed to genetic determination and not solely to plastic responses to varying environmental constraints imposed by altitude. The significant temporal lag of ca. 6 days in budburst phenology between plants of high and low altitudes marked a shortening of the growing season that is perhaps the most relevant constraint for plants at high altitudes (Meloche and Diggle 2003; Cox 2005). Plants from stand H were shorter, more densely branched (immediate and delayed), and had a lower proportion of individuals whose main axes were dominant. At year 6, they presented a higher mean branchiness, and the majority (74.5%) of them clustered near archetype 1, which represents small multi-stemmed, densely branched phenotypes (Fig. 5.6), resembling the dominant architectural forms observed at high altitude. Instead, plants from stand L did not display one clearly prevailing growth habit. Thus, at high altitudes, N. pumilio seems to experience natural selection favoring late flushing and fast shoot extension to avoid frost damage. With increasing altitude, trees could be maximizing carbon gain by a more efficient photosynthetic performance (e.g., Premoli and Brewer 2007; Molina-Montenegro et al. 2012) and the efficient allocation of non-structural carbohydrates (e.g., Fajardo et al. 2013).

The dissimilar results thrown by the two last essays may be due to the different ages of evaluation, what was evidenced by the lack of age-to-age correlation in the second and longer trial. Alternatively, differences could be related to the sampling sites. In the first case, the contrasting morphotypes were vegetating close to each other, and consequently gene flow likely exists among them, thus restricting differentiation and consequently adaptation. On the contrary, in the second study, both morphotypes are separated by a relatively large distance, and gene flow is likely more limited than in the first case, and therefore adaptation is more probable.

Based on the previous results, altitudinal and latitudinal zonation should be considered in the definition of management units and/or in the delineation of provenance regions in N. pumilio, although not in a clinal way, since ecotypic variation seems to prevail. The seed sources for restoration programs, as well as the seed orchards for low-intensity breeding, should avoid the admixture of genetic materials from stands markedly separated in altitude or latitude. Observed gaps in bud sprouting and growth rhythm development between provenances in both gradients evidenced adaptation to local conditions. Particular site conditions should not be dismissed in active restoration. On the other hand, the amount of annual precipitation does not seem to make a difference among populations.

5.3 Genetic Variation by Means of Progeny Trials

A preliminary study of intra-population genetic variation (Mondino 2014) was conducted by means of a greenhouse progeny test that included the progeny of 68 open-pollinated mother trees corresponding to four natural populations of the Province of Chubut (N = 897). At the end of the first growing season, several architectural traits were measured in each potted seedling. The dispersion of data was large due to the low sampling level; however, the difference among populations and the significance of the family factor could be shown for some traits. Differences among populations were proved to be significant for height, slenderness index (height/diameter ratio), and branchiness index (height/number of branches ratio), for which a differentiation among populations of Q_ST = 15%, Q_ST = 16%, and Q_ST = 17% was estimated, respectively. It must be highlighted that the differentiation was due to only one population that was different from the other three: San Martin, which is in fact a marginal population of the steppe, completely isolated from the forest continuum (43° 49′ 47′ S, 70° 45′ 33′ W; 1350 m asl), and subjected to stressful precipitation conditions (annual average of 300 mm). The heritability estimated for the traits whose family factor resulted significant were mostly moderate, ranging from 0.15 to 0.57 (Table 5.5).

Table 5.5 Additive genetic variance (σ_A), variation coefficient (CV_A%), and heritability (h²) in five variables measured in Nothofagus pumilio seedlings

6 Domestication and Low-Intensity Breeding Strategies

6.1 First Steps in N. antarctica

Although regeneration by sprouts (even after fire) is common in the species, viable seeds are produced regularly. In primary old growth stands of Tierra del Fuego (Soler Esteban et al. 2010), a one-season production of 1.85 million of viable seeds per hectare was estimated, which in fact represent a very low proportion of the 10.28 millions of seeds produced. These values surely vary among years and stands, but anyway, seeds are commonly available. This opens the possibility of seedling production and consequently active restoration (plantation) in the management of markedly degraded ñirantales (i.e., natural forests of the species), ensuring and accelerating the times of forest regeneration. Likewise, planting of N. antarctica for productive purposes is also possible, mainly in silvopastoral systems related to cattle raising (Hansen et al. 2004), in the forest-steppe ecotone. Seedlings suitable for planting can be obtained in one season by means of ferti-irrigation under greenhouse conditions, as in the other Nothofagus of Argentina (Schinelli Casares 2012) (Fig. 5.7).

Fig. 5.7

Greenhouse seedling production of Nothofagus. (a) 1-month-old plantlets of Nothofagus antarctica, (b) 8-month-old plants of Nothofagus pumilio, with an outstanding seedling in foreground. (Photos: Mario J. Pastorino)

Nothofagus antarctica is a species with a great adaptability to dry and waterlogged environments and to different types of soils. Since 2003, 16 plots of ñire plantations have been established at different site conditions in the mountain ranges of the Province of Chubut (~42–44° S). Both survival and growth rate significantly changed depending on site conditions (Luis E. Tejera, personal communication). In humid sites, an average height of 3.5 m has been reached at 5 years of age, with tallest trees presenting 4.5 m of total height. Similar to other Nothofagus, the species usually presents apical death probably related to stressful conditions (i.e., peaks of high temperatures or drought), which causes stem damage in juvenile individuals. Among the main limitations for its establishment are damage by rodents such as tuco-tucos (Ctenomys spp.) , European hares (Lepus europaeus), and rabbits (Oryctolagus cuniculus) (Contreras 1973; Vincon 2010) (Fig. 5.8). In addition, the attack of wood-boring insects (Lautarus concinnus, Calydon submetallicum, Calydon globithorax, Phymantoderus bizonatus, Callisphyris semicaligatus) on the trunks of living trees (Rizzuto 2003) is highly widespread, which greatly affects its timber (Fig. 5.8).

Fig. 5.8

(a) A specimen of tuco-tuco (Ctenomys spp.) and (b) the damage it causes in N. antarctica roots, (c) damage caused by wood-boring insects. (Photos: Luis E. Tejera)

Breeding activities are barely initial for N. antarctica. The main subject related to the productive cultivation of the species is its poor forestry shape, with many stems, often tortuous, and with forking and coarse branches. However, there is much variation in growth form, and trees growing straight upright can be found in any forest, which gives the chance to think about selection. In order to test the genetic control of the forestry shape (Schinelli Casares et al. 2016), 39 trees were selected in a natural stand according to two categories: monopodial and straight trees (named as “plus”) and common phenotypes (named as “general,” including multi-stem, tortuous, or forked trees). Seeds were collected, seedlings were produced, and a greenhouse progeny trial was established with 60 seedlings per open-pollinated family randomly ordered in three repetitions (N = 2340). At the end of the first growing season, the total height (H) of each seedling was measured, and they were categorized in three shape types (S): straight, crooked, or decumbent. Differences between the progeny of plus and general mother trees were shown for the proportion of S (Fig. 5.9), but not for H. Plus trees had a progeny with a greater proportion of straight seedlings. On the contrary, when testing the family factor, it was not significant for the proportion of S but for H, and in fact a high heritability was estimated (h² = 82%). We still do not know the evolution of these traits at older ages, but it seems wise to select mother trees in the natural forest according to their shape. On the other hand, the selection according to the growth in height would also be quite convenient. However, this could hardly be done in the natural forest, since it is ignored if the higher relative height of a tree is due to a better growth or an older age.

Fig. 5.9

Contrasting Nothofagus antarctica individual phenotypes photographed in 2019 in the progeny seed orchard installed in 2011 in Trevelin. (Photos: Luis E. Tejera). Central box: proportion of seedlings of the defined shape categories for each of the two types of progenitors selected as “plus” and “general” trees

After these preliminary results, a low-intensity breeding strategy was planned by INTA for N. antarctica, based on the establishment of progeny seed orchards with seedlings obtained from seeds of good forestry shape individuals selected in the natural forest. Thus, 40 trees with monopodial and straight stems were selected in two natural populations of the Province of Chubut (Trevelin and Lake Rosario). In 2009, open-pollinated seeds were collected from those trees, expecting a greater proportion of good shape seedlings in the nursery production, which was conducted by INTA. Finally, in 2011, two seed orchards were established with 2-year-old seedlings. The first one was settled in Trevelin Forest Station of INTA (43° 05′ 29″S, 71° 32′ 21″W, 450 m asl; Fig. 5.9), with 1750 seedlings and without irrigation, and the second in a nearby private property (43° 7′ 6″S, 71° 27′ 30″W, 360 m asl), with 400 seedlings and furrow irrigation. In order to perform a preliminary evaluation, the height of the saplings was measured in 2015 in both orchards, together with a characterization of their shape by means of three binary traits: mono-/sympodial growth, straight/tortuous stem, and presence/absence of forking. The mean height of each orchard was 113.5 ± 49 cm and 196 ± 48 cm, respectively. No differences could be shown between the populations in both orchards for any of the shape traits, nor for the height, and the family factor was significant only for the height in the irrigated orchard and in fact with a h² = 0.24. It may be that the environmental heterogeneity is too large in the non-irrigated orchard to detect differences so soon.

6.2 First Steps in N. pumilio

Seed production has been estimated in old growth natural forests of Tierra del Fuego (Martínez Pastur et al. 2008) and does not seem to be a limiting factor for seedling production (42.5 million of viable seeds per hectare in one season). However, in the experience of INTA in central and north Argentine Patagonia, seed provision represents a true bottleneck. Seed production fluctuates greatly from year to year in the natural forest, and the proportion of empty or nonviable seeds is rather large. It may take several years until an abundant collection of good quality seeds is achieved. Furthermore, the seeds of N. pumilio are almost recalcitrant, and, therefore, their viability decreases considerably in only 1 year, even when stocked at low temperature. Thus, the strategy of stocking seeds during mast years, as recommended for N. alpina and N. obliqua, does not apply in this case, at least till a specific stocking protocol can be adjusted for this species. With viable seeds, very good seedlings (more than 35 cm tall and 0.5 mm of collar diameter) can be produced in 8 months by means of ferti-irrigation under greenhouse conditions (Schinelli Casares 2012; Fig. 5.7).

Although the species has been planted profusely on a very low scale, even in Great Britain (Mason et al. 2018), Denmark, and Norway (Sondergaard 1997), forest plantations with N. pumilio (as with N. antarctica) are scarce. Just in the last decade, the species has been planted on a larger scale in the frame of active restoration programs related to extensive forest fires that occurred in Chile and Argentina. Good examples of these efforts are the restoration programs performed in Torres del Paine National Park (Chile) and in the Argentine provinces of Río Negro, Chubut, Santa Cruz, and Tierra del Fuego, where hundreds of thousands of lenga seedlings have been planted (Pastorino et al. 2018; Guzmán et al. 2019; Mattenet et al. 2019; Mestre et al. 2019; Paredes et al. 2019; Salinas et al. 2019). However, the vast majority of those seedlings were not produced in nurseries but collected from the regeneration banks of the natural forests. In any case, these experiences have boosted interest in the breeding of the species at both sides of the Andes Cordillera.

In Chile, a trial was installed in 2000 in the Forest Reserve of Coyhaique (Ipinza and Gutierrez 2015) with the open-pollinated progeny of 111 mother trees sampled in the natural forest of three provenance regions (N = 2466). Plus and not selected trees were included in the sample. The average survival of the trial was 58.5% at 11 years of age, with 2.14 m of mean height and 21.8 mm of mean collar diameter. The heritability of these two last traits was h² = 0.18 and h² = 0.27, respectively. It is remarkable that among the 20 better mothers, there was not any plus tree, and among the 50 better trees, there was not any progeny of the plus trees. This result leads to discard the strategy of selecting trees from the natural forest. Instead, the mass selection from planted trees, where the method of comparison with neighbor individuals is applied (Ledig 1974), or better yet, the genetic selection in progeny tests seems recommendable.

In Argentina, the first steps toward the genetic improvement of the species have been taken recently. A small provenance trial network has started to be established in 2017. Seeds were collected from 12 natural stands of the whole distribution area of N. pumilio in Argentina, from south of Tierra del Fuego to north of the Province of Neuquén. Seedlings were produced and four provenance trials were installed in (1) Río Turbio (Punta Gruesa Forest Reserve, 51° 32′ 44.7″S, 72° 07′ 22.2″W; 535 m asl; N = 675; first year survival = 92%), (2) Trevelin Forest Station of INTA (43° 05′ 29″S, 71° 32′ 21″W; 450 m asl; N = 900; first year survival = 97%), (3) Bariloche alto (private property, 41°13′38.86″S, 71°14′32.85″W; 1060 m asl; N = 480; second year survival = 51%), and (4) Bariloche Experimental Station of INTA (41° 7′ 21.17″ S, 71°14′ 56.95″ W; 795 m asl; N = 270; second year survival = 80%).

Finally, aiming to count with seeds of known origin for seedling production, in 2017, a production seed area was registered with the National Seed Institute (INASE). It is a natural pure stand of 24 ha located in the Trevelin Forest Station of INTA (43° 3′ 49″S, 71° 34′ 33″W, 1100 m asl) with very good accessibility, seed productivity, and seed quality.