Evolutionary significance of a flat-leaved Pinus in Vietnamese rainforest


Author for correspondence:
Timothy J. Brodribb
Tel: 61 362261707
Fax: 61 36262698
Email: timothyb@utas.edu.au


  • • Pines are generally absent from tropical rainforests. An important exception, Pinus krempfii, is a unique tree that bears flattened needles and competes with evergreen angiosperm trees in southern Vietnam.
  • • Here, the photosynthetic and hydraulic physiology of P. krempfii leaves were examined to determine whether this species departs from the widespread pattern of high-light-demanding photosynthetic physiology displayed in needle-leaved Pinus species.
  • • Maximum photosynthesis and light saturation of photosynthesis, as well as stem and leaf hydraulic efficiencies, were all very low in P. krempfii compared with other Pinus species. These characteristics were consistent with our observations of P. krempfii seedling regeneration under the forest canopy. By possessing shade tolerance coupled with the production of flattened leaves, P. krempfii has converged morphologically and physiologically with many genera of the southern hemisphere conifer family Podocarpaceae. This convergence extends to a key feature of leaf anatomy, the production of tubular sclereids in the leaf for radial transport of water from the vein to the margin.
  • • These observations suggest that few adaptive possibilities are open to conifers when moving into tropical rainforest, meaning that Pinus is forced into direct competition with southern hemisphere conifers for a narrow niche in the equatorial zone.


Pines are enormously successful emblems of natural forest in the northern hemisphere (Pinus has c. 120 species) and invasive exotic crop plants south of the equator. The natural domain of the genus extends from the Arctic Circle (72°N) south to its final outpost in northern Sumatra (2°S). Two features of this distribution are particularly striking: first, the great ecological importance of native Pinus forests in the northern hemisphere; and second, their conspicuous absence (or any other members of the Pinaceae) from native forests south of the equator. Even more curious is the lack of any fossil evidence of the Pinaceae south of the equator. While other major conifer families. Araucariaceae, Podocarpaceae and Cupressaceae show abundant fossil evidence of panglobal distributions in the Mesozoic (Chaney, 1947; Florin, 1963; Mirov, 1967; Stockey, 1994; Enright & Hill, 1995; Hill & Brodribb, 1999), fossils attributable to Pinaceae have never been found outside their current latitudinal confines. This fact is made more surprising by the family's rich fossil record in the northern hemisphere, extending back to the Early Cretaceous (Millar, 1998; Smith & Stockey, 2001).

The absence of Pinus in the southern hemisphere remains an enigmatic pattern. The antiquity of the genus combined with its well-demonstrated ability for both long-distance and rapid dispersal (Davis & Shaw, 2001; Grotkopp et al., 2002) argues against simple geographical containment as preventing invasion from northern to southern land masses. Indeed, Pinus diversity is high in the region of the Tropic of Cancer today (Farjon & Styles, 1997; Nguyen & Thomas, 2004), and there is fossil evidence that pine forests have been unsuccessfully ‘knocking on the door’ of the southern hemisphere for c. 50 million yr (Millar, 1998). Whilst considerable interest has been focussed recently on explaining the biogeography of Pinus in the northern hemisphere (Eckert & Hall, 2006; Willard et al., 2006), little attention has been paid to the inability of Pinus to cross the equator.

An interesting alternative to vicariance as an explanation for the contrasting fortunes of Pinus in the northern and southern hemispheres is that evergreen tropical forests pose an impenetrable ecological barrier to the southward movement of Pinus. Conifers in general have been discussed as ill-equipped to compete with highly productive tropical angiosperms. This is because of their relatively inefficient xylem, low photosynthetic rates, and single-vein leaves that have been suggested to handicap conifer seedling regeneration in the company of angiosperms (Bond, 1989). Of the conifers that do thrive in the tropics, however, most employ a suite of anatomical and structural modifications to produce flattened leaves or short shoots (Hill & Brodribb, 1999). These broad leaves enable shoots to be deployed efficiently with a minimum amount of self-shading, thus facilitating growth under the low and variable light conditions characteristic of equatorial forest (Brodribb & Hill, 1997). By contrast, Pinus (and Pinaceae in general) have been relatively unadventurous in exploring flattened leaf morphologies, with the probable consequence that virtually all species are shade-intolerant and unable to regenerate beneath a forest canopy (see Richardson & Rundel, 1998 for a review). This conservative leaf morphology may underlie the inability of Pinus to penetrate the equatorial zone and into what appears suitable habitat (Grotkopp et al., 2002) in the southern hemisphere.

The dominant conifer family in the southern hemisphere, the Podocarpaceae, contrasts strongly with the Pinaceae in both its pantropical distribution and its diverse range of flattened leaf and shoot morphologies (Bucholtz & Gray, 1948; Brodribb & Hill, 2004). Podocarpaceae use a diversity of sclerified tissues in the mesophyll to transport water radially from the single vascular strand towards the leaf margin (Griffith, 1957). The resultant extravenous hydraulic network appears to overcome the limitations on leaf size imposed by single-vein leaf anatomy (Brodribb et al., 2007). Other gymnosperms (species of Cupressaceae, Taxaceae and Cycas) also show this association of wide leaves and sclereids in the mesophyll (Brodribb et al., 2007). By contrast, very few Pinaceae have comparable sclereids in the leaf (Bucholtz, 1951). Thus, one hypothesis is that the virtual absence of leaf sclereids in Pinus leads to an inability to produce flattened leaves, which in turn limits the success of the genus at low latitudes.

Here we address this biogeographic hypothesis by examining the function and anatomy of the deviant Pinus species, Pinus krempfii (Lecomte). This pine is not only highly significant as one of the southernmost occurrences of Pinus, but also as the only species to possess bilaterally flattened leaves (up to 12 mm wide and 10 cm long; Fig. 1a) (Bucholtz, 1951; Ickert-Bond, 2000). Its occurrence in primary tropical montane rainforest alongside ‘Gondwanan’ podocarp conifers that have crossed the equator and extended into the northern hemisphere adds further intrigue to this remarkable plant. By examining the anatomy, photosynthetic light response of the leaves as well as xylem hydraulic performance of leaves and stems of this species and its conifer associates (including two conventional Pinus species), we investigate the extent to which P. krempfii falls outside the range of needle-leaf Pinus physiology. We also examine the degree of structural and physiological convergence between P. krempfii and the major southern hemisphere conifer family, Podocarpaceae. Of particular interest was the possibility that leaves of P. krempfii might develop a podocarp-like system of lignified tubes used to augment water delivery to the mesophyll in wide, single-vein leaves (Brodribb et al., 2007).

Figure 1.

(a) An emergent tree of Pinus krempfii (c. 40 m) above mixed conifer evergreen forest at 1200 m in Vietnam. (b) Flattened ‘needles’ of a P. krempfii sapling growing in low light in the understorey.

Materials and Methods

Study site and plant material

We conducted our study at Hon Giao station (c. 1500 m above sea level), which is located in Bidoup-Nui Ba National Park, Lam Dong Province, approx. 30 km from Dalat, Vietnam (12°4′60″N, 108°40′0″E). The climate of the Vietnamese Central Highlands is characterized by a dry, cool season from December to early April, and a warm, wet monsoon season from May to November. Mean annual temperature is 18°C and mean annual rainfall is approx. 1800 mm. In this site, three species of Pinus co-occur. Pinus krempfii and P. kesiya were selected for study; the third species, P. dalatensis, could not be studied because we were unable to sample branches in full sun, as they were situated in the canopy (c. 40 m).

Both P. krempfii and P. kesiya grow in close proximity, yet their ecologies are divergent. Pinus kesiya is a stereotypical, tropical, needle-leaved pine species. It forms monotypic stands, and is highly dependent on frequent fire to maintain regeneration (Richardson & Rundel, 1998). Stands of P. kesiya occur on the burned borders of primary rainforest that is codominated by a diversity of broad-leaf angiosperms and conifers. Seedlings of P. kesiya only regenerate in recently burned pastures. By contrast, P. krempfii only occurs in undisturbed primary forest (above 1200 m) and is excluded by fire and logging. In our study site, P. krempfii is locally abundant, with over 200 trees in the population. The species is a canopy emergent (up to 40 m tall) generally occurring on steep slopes at elevations of 1500–2000 m. Emergent trees often attain a great age, with over 800 growth rings recorded from stem bole cores (B. Buckley, pers. comm., 2007). Other conifers, including Fokienia hodginsii and Taxus wallachianna, as well as several podocarps (Dacrydium elatum, Dacrycarpus imbricatus, and Podocarpus neriifolius), are also common in the canopy. Pinus krempfii shows evidence of continuous seedling regeneration in the forest. Seedlings and saplings of P. krempfii were relatively abundant, growing alongside seedlings of the podocarp species under large trees and in treefall gaps.

Leaf photosynthetic performance

We selected healthy sun leaves (those exposed to sun for > 50% of the day) from five small trees of each species for measurement of photosynthetic light response. Measurements of photosynthetic capacity were made using chlorophyll (Chl) a fluorescence and leaf gas-exchange.

Fluorescence was measured in individual leaves using a MiniPAM (Walz, Effeltrich, Germany) portable fluorometer. Leaves were illuminated for 15–45 s with an external halogen light until stable steady-state fluorescence (Fs) values were observed (usually c. 15 s of actinic illumination). Once a stable Fs value was observed, the leaf was exposed to a brief saturation pulse (1 s, 4500 µmol m−2 s−1 PPFD) to determine maximal fluorescence yield under an actinic light (inline image). The quantum yield of photosystem II (φPSII) was calculated as inline image (Genty et al., 1989), where Fs is the steady-state fluorescence yield measured under actinic light before application of a saturation pulse. Using φPSII data, we calculated electron transport rates (ETR), a component of leaf photosynthetic capacity (Genty et al., 1989). ETR was calculated as ETR = φPSII × I × α × 0.5, where I is the incident PPFD; α is the leaf absorbance, taken here as 0.84 (Björkman & Demmig, 1987), and the coefficient of 0.5 accounts for the assumption that half of the absorbed light energy is distributed evenly between photosystem II and I (Bilger et al., 1996). The units of ETR are µmol electrons m−2 s−1, although it should be noted that values of ETR may not be precise because of small variations in α among the species we sampled. We believe that differences in α are small, because the leaves of both Pinus species appeared very similar in color, and neither had undersurface anthocyanins or trichomes, which can result in significant differences in whole leaf light absorbance (Lee et al., 1990).

Following determination of maximum ETR in the field, five fully hydrated, sunlit branches were harvested and transported back to the laboratory in sealed plastic bags for determination of the light response of ETR in both species. Following the procedure of Brodribb & Hill (1997), small branches were cut underwater, placed in humid plastic bags and acclimated at an initial light intensity of approx. 1800 µmol m−2 s−1 for 15 min. For each of five replicate leaves, ETR was measured to ensure it fell within the range measured in the field. Following this, actinic light intensities were stepped down gradually to c. 15 µmol m−2 s−1, allowing a period of 5 min for acclimation at each light intensity. By starting with leaves at saturating illumination, we ensured that stomatal conductance was nonlimiting at all light intensities. Leaf temperature remained between 24 and 26°C throughout.

A similar procedure was undertaken using an infrared gas analyzer (Li-Cor 6400; Li-Cor, Lincoln, NB, USA) to determine the response of net CO2 exchange to light intensity in P. krempfii. Five replicate leaves were induced to open stomata inside the cuvette at a [CO2] of 380 ppm, a leaf temperature of 25°C, an initial light intensity of 1800 µmol−1 m−2 s−1 and a vapour pressure gradient of not more than 2 kPa. Net CO2 exchange was logged at this initial light intensity and then at 1200, 1000, 800, 600, 400, 200, 100, and 50 µmol m−2 s−1. A maximum time of 30 s between measurements was allowed, ensuring that leaves did not become dehydrated during measurement. Gas exchange parameters were expressed as a function of projected leaf area, giving maximum CO2 assimilation rates (Amax) in units of µmol CO2 m−2 s−1.

Pinus Krempfii photosynthesis compared with other conifers

Amax in P. krempfii was compared with that in other pines from tropical and temperate regions using published CO2 assimilation data. We compiled a collection of published records of maximum assimilation rates for Pinus needles that were measured under optimal conditions, and where rates were either expressed on a projected area basis, or could be converted to a projected area. For comparison, we also collected published records of photosynthesis in tropical Podocarpaceae.

Stem and leaf hydraulics

We sampled sunlit branches from five individuals in the field for determination of stem hydraulic properties. Branches were collected in the early morning (to ensure stems were maximally hydrated), double-sealed in plastic bags, and transferred to a makeshift laboratory at Hon Giao station within 1 h. At the laboratory, stem segments for hydraulic measurements were cut from branches while under water. Stem segments were cut at a standardized length (c. 7 cm) and diameter (3–4 mm). Both ends of the excised segments were then carefully shaved with a double-edged razor blade before attaching them to the measurement system. Hydraulic measurements were made on one stem segment from five individual plants of each species.

A low-pressure steady-state flow meter (Brodribb & Feild, 2000) was used to measure stem hydraulic conductance (KH; kg MPa−1 m s−1). Stems were perfused with a measuring solution of sterile spring water (cationic strength of 15 mm, including monovalent and divalent cations), using a hand pump to pressurize a small (350 ml) stainless steel reservoir. The solution was filtered to 0.2 µm with a syringe filter, and manually degassed by repeatedly placing it under vacuum with a large-volume syringe. Stem segments were measured using a driving pressure of approx. 0.1 MPa. The head pressure as well as the pressure drop at the attachment juncture between the stem segment and a calibrated tube of known conductance were measured with a portable electronic manometer (0.25 MPa range; Sper Instruments, Scotsdale, AZ, USA). Following connection of the stem segment to the flow meter, pressure data were recorded when steady-state pressure readings (< 1% variation in 30 s) were achieved (typically < 5 min). We changed the solution in the tubing system every day to avoid any microbial growth that can clog the flow meter or the stem segment.

We calculated sapwood-area specific conductivity (KS; kg MPa−1 m−1 s−1) by dividing stem KH by the sapwood cross-sectional area (m2). Sapwood cross-sectional area was measured under a dissecting microscope on hand-sectioned portions of stems and the area contributed by the pith was subtracted. Fresh leaves were scanned on a flatbed scanner and the projected leaf area was determined using image analysis software (ImageJ; National Institutes of Health, Bethesda, MD, USA). Leaf specific conductivity (KL; kg MPa−1 m−1 s−1) was calculated by dividing KH by the total leaf area supported (i.e. distally) by that branch.

Leaf hydraulic conductance of 10 leaves (Kleaf) of P. krempfii was measured using the water potential relaxation method (Brodribb & Holbrook, 2006). Pressure–volume curves were constructed for five leaves dehydrated on the laboratory bench and leaf capacitance calculated from the slope of the water potential vs relative water content plot and water content per unit leaf area (Brodribb & Holbrook, 2003). Rehydrations were carried out on sun leaves collected and carefully bagged 24 h before measurement. To ensure that Kleaf was maximal and not affected by cavitation (Brodribb & Holbrook, 2006) or low light depression (Cochard et al., 2007), Kleaf was measured with leaf water potential between −0.5 and −1 MPa at 22°C under a light intensity of 300 µmol−1 m−2 s−1.

Leaf anatomy

Several anatomical traits of leaves were measured to determine if leaf anatomy contributed to the hydraulic function of flattened leaves in P. krempfii. We were specifically interested to know if leaves of P. krempfii fell within a recently described relationship between leaf internal anatomy and leaf hydraulic efficiency (Brodribb et al., 2007). Hence leaves were examined to determine the hydraulic distance between the leaf vein and the stomata (Dm). This parameter is a measure of the distance that must be travelled by water in the transpiration stream from the leaf vein, through the mesophyll to the stomata, and it requires four anatomical parameters to be measured, namely the leaf thickness from vein to stomata, leaf width from (lignified) transfusion tissue to most distal stomatal band, and average mesophyll cell length and width. Assuming water passes apoplastically through the mesophyll, Dm is calculated as the distance through a matrix of tightly packed capsules of known length and radius (cell length and width). Hence paradermal and cross-sections of leaves were prepared from five leaves used in the above measurements, and mean values for each of the four leaf parameters measured with a light microscope and digital camera. One hundred mesophyll cells were measured for length and width, while 10 measurements of leaf width and thickness were made, and mean values for each parameter used to calculate Dm from Eqn 2 (Brodribb et al., 2007) The mean distance traversed along the long axis of cells (Cx) was calculated as the cell length plus the semicircular distance around the tip of the cell (Eqn 1), while the distance around the minor axis of cells (Cy) was calculated from the number of cells traversed multiplied by the distance around half the cylindrical cell perimeter (Eqn 2):

X = (ν/Cx)[(Cx – Cy) + πCy/2](Eqn 1)
Y = (t/Cy) (πCy/2)(Eqn 2)

(X, horizontal apoplastic path length from the vein to stomata; ν, mean longest distance from veins to stomata; Y, vertical apoplastic path length from the vein to stomata; t, mean leaf thickness from vein to stomata).

Finally the total distance through the mesophyll (Dm) was taken as the hypotenuse of a right-angled triangle produced by these x and y dimensions (Eqn 3).

Dm = √(X2 + Y2)(Eqn 3)

Frozen sections of leaves were examined to determine the structure of lignified mesophyll cells bridging the leaf. Cross-sections cut parallel to the midvein were examined using an epifluorescence miscroscope (Axioskop, Carl Zeiss, Oberkochen, Germany) to identify lignified cells in the mesophyll.


Photosynthesis in leaves of P. krempfii became light-saturated at a mean photosynthetic photon flux density (PPFDsat) of 498 µmol m−2 s−1 (Fig. 2a). This was about one-third of the PPFDsat (1530 µmol m−2 s−1) found in leaves of the associated P. kesiya. Light-saturated ETR was correspondingly low in P. krempfii: 61 µmol m−2 s−1 compared with 184 µmol−1 m−2 s−1 in the needle-leaved P. kesiya (Fig. 2b). We found that both the hydraulic conductivity of the sapwood (KSP) and the leaf area specific stem conductivity (KL) were significantly lower (P < 0.05, two-tailed t-test) in P. krempfii than in P. kesiya. The very large difference in KL between these species (approx. 300%) was of the same order as the differences in light-saturated ETR. Mean maximum net CO2 assimilation (Amax) in leaves of P. krempfii was very low (4.1 ± 0.1 µmol m−2 s−1) and fell outside the range of published data from pines native to tropical and temperate climates (8.0–16.9 µmol−1 m−2 s−1; Fig. 3). This low photosynthetic rate was, however, within the range of published values for tropical Podocarpaceae species, which exhibited a smaller range of photosynthetic capacity than the pines (3.5–6.5 µmol m−2 s−1; Fig. 3).

Figure 2.

(a) Light saturation characteristics of photosynthetic electron transport rates (ETR) in sun leaves from flat-leaved Pinus krempfii (open circles) and its needle-leaved neighbor P. kesiya (closed circles) (n = 5; error bars show ± 1 SD). The shaded region indicates the range of saturating light intensities from 10 rainforest Podocarpaceae (data from Brodribb & Hill, 1997). (b) Leaf (hatched) and sapwood-area (grey) normalized stem hydraulic characteristics of P. krempfii and P. kesiya sunlit branches (n = 5; error bars show ± SD).

Figure 3.

Mean maximum net CO2 assimilation rates for Pinus krempfii (grey bar) compared with 15 needle-leaved Pinus species (black bars) and 10 flat-leaved Podocarpaceae (white bars). Data from: 1, Turnbull et al. (1998); 2, Jach & Ceulemans (2000); 3, Zhang & Marshall (1995); 4, calculated from electron transport rates (ETR); 5, Day et al. (1991); 6, Will et al. (2001); 7, Tan & Hogan (1995); 8, Peters et al. (2003); 9, Lusk et al. (2003); 10, Gower et al. (1993); 11, Brodribb et al. (2007); 12, Koskela et al. (1999); 13, Jose et al. (2003); 14, Dick et al. (1991); 15, Warren & Adams (2000).

The hydraulic conductance of leaves of P. krempfii was low (mean Kleaf of 2.9 ± 0.25 mmol m−2 s−1 MPa−1) compared with other needle-leaved Pinus (Fig. 4). A low Kleaf in this species was expected because leaf flattening produced a very long Dm (mean of 2106 ± 213 µm) from the midvein to the stomata (Brodribb et al., 2007). However, the Kleaf value measured for P. krempfii was considerably higher than the low value predicted on the basis of its broad single-vein leaves and consequently large Dm (Fig. 4). Based on the published data set correlating Dm with Kleaf across the range of terrestrial plants (Brodribb et al., 2007), a Kleaf of 0.87 mmol m−2 s−1 MPa−1 would be predicted for P. krempfii. Higher-than-expected Kleaf values in P. krempfii appeared to be associated with lignified tissues in the leaf which bridged the gap between the midvein and the stomata (Fig. 5a–c). Examination of the anatomy of fully expanded P. krempfii leaves confirmed the presence of large lignified cells elongated perpendicular to the midvein (Fig. 5) in a manner similar to that seen in the flattened leaves of many Podocarpaceae species. However, the density and interconnection of these large cylindrical cells (sclereids) were not as extensive as those seen in broad-leaved Podocarpaceae species (Fig. 5d–f). Cross-section of the leaf lamina revealed lignified cells to be long (mean 194 ± 51 µm), unbranched tubular cells with thick walls and a large internal diameter (mean 41 ± 8 µm). Unlike similar tissue in Podocarpus leaves, the degree of lignification appeared to be variable, and chloroplasts were present in approximately half of the lignified cells.

Figure 4.

Leaf hydraulic conductivity in three needle-leaved Pinus species (P. palustris, P. echinata, P. strobus– open triangles; data from Brodribb et al., 2007) was related to mesophyll hydraulic length in the same fashion as 25 other land plants (closed circles; data from Brodribb et al., 2007). Pinus krempfii (closed triangle) exhibited a mean Kleaf lower than three needle-leaved Pinus species; however, this value exceeded predictions based upon leaf anatomy (i.e. the regression line shown). In this characteristic, P. krempfii leaves were similar to those of eight flattened Podocarpaceae leaves (grey squares; data from Brodribb et al., 2007).

Figure 5.

Anatomical details from leaves of Pinus krempfii (a–c) and Podocarpus neriifolius (d–f), showing lignified tubes extending from the midvein towards the leaf margin. Paradermal (a, d) and transverse (b, e) sections show the density and interconnection of lignified cells (stained blue in toluidine blue), while longitudinal sections (c, f) show the diameter and wall thickness of lignified cells (blue fluorescence from lignin UV-autofluorescence).


Our examination of the structure and function of the flattened needles of P. krempfii highlight this remarkable tropical pine as a species of great significance in the biogeography and evolution of the genus. Even though it is phylogenetically well nested within Pinus (Wang et al., 2000), we found that P. krempfii shows photosynthetic, hydraulic and anatomical characteristics more akin to a southern hemisphere podocarp than a pine tree.

Our findings demonstrate that the flattened leaves of P. krempfii are adapted to function optimally under low-light conditions. Low light saturation intensities for ETR and low maximum rates of photosynthesis correspond closely with the physiological pattern observed in the major southern hemisphere conifer family, the Podocarpaceae. Among podocarps, it has been shown that broad-leaved species require lower PPFD for light saturation, and exhibit lower maximum ETR than narrow-leaved species (Brodribb & Hill, 1997). This was clearly the case when comparing P. krempfii with its neighbor P. kesiya, a typical needle-leaved pine that occurs only on highly disturbed (high light) forest edges. Pinus kesiya leaves required three times the light intensity to saturate ETR and produced a maximum ETR three times higher than P. krempfii leaves. When compared with data taken across the phylogenetic breadth of Pinus, light saturation of photosynthesis at a PPFD < 500 µmol m−2 s−1 and Amax of only 4.1 µmol CO2 m−2 s−1 in P. krempfii stands out as very unusual for Pinus needles. Pinus needles typically saturate at light intensities above 1000 µmol m−2 s−1 (Teskey et al., 1994; Richardson & Rundel, 1998) and yield net photosynthetic rates between two and four times higher than that observed in P. krempfii (Fig. 3).

In general, the hydraulic efficiency of wood and leaves has been shown to reflect the intrinsic leaf photosynthetic capacity (Brodribb & Feild, 2000; Hubbard et al., 2001; Brodribb et al., 2005). Hence the low hydraulic conductivity of P. krempfii wood and leaves (Fig. 2b) further supports the conclusion that the physiology of this species is geared towards producing low maximum photosynthetic rates under low light intensity. In this way, the hydraulics of P. krempfii align better with the rainforest genera of southern hemisphere Podocarpaceae than with other species of Pinus (Brodribb & Feild, 2000). Another similarity between the wood of P. krempfii and rainforest Podocarpaceae was high specific density (> 0.6 g cm−3, data not shown) (Pitterman et al., 2006).

The leaves of P. krempfii also showed striking convergence with the flattened leaves of rainforest Podocarpaceae. Arrays of elongated sclereids in the mesophyll tissue of P. krempfii were highly reminiscent of those found in most broad-leaved podocarp species. These sclereids appear to be critical for facilitating radial water transport from the vein toward the leaf margin in broad, single-vein leaves (Brodribb et al., 2007). In P. krempfii the configuration of these sclereids may be less efficient than that seen in many Podocarpaceae. Genera such as Acmopyle, Falcatifolium, Podocarpus and Sundacarpus all produce interconnected sclereids that also connect to the midrib via lignified transfusion tissue. By contrast, the sclereids in P. krempfii are weakly interconnected and attach to the endodermis rather than to lignified transfusion cells, probably creating a less efficient hydraulic pathway than in podocarp leaves (Fig. 5). This anatomy may underlie the observation of lower Kleaf in P. krempfii than in podocarp leaves of similar sizes (Fig. 4).

These observations provide the first evidence that Pinus has the physiological (and anatomical) capabilities to invade equatorial evergreen forest. The question remains as to whether the evolution of broad leaves and shade tolerance is a recent development, representing a new southward invasion of the genus, or alternatively the remnant of an ancient and unsuccessful invasion. Molecular phylogenetic results are more in line with a recent origin of shoot widening and shade tolerance in Pinus. Fossil-calibrated molecular clock studies indicate that the diversification of the subgenera within Pinus is relatively recent, with divergence of the lineage containing extant P. krempfii occurring c. 14–27 million yr ago (Willard et al., 2006). While this split is rather recent when viewed relative to the ancient timeframe of conifer evolution (Florin, 1963; Mirov, 1967; Millar, 1998), it also suggests that P. krempfii has produced only limited success in spearheading a Pinus invasion southward towards the equator. Pinus krempfii today exhibits a range of < 2000 km2 in central Vietnam, contrasting with the pattern observed in the competing species of Podocarpaceae, which are abundant and extremely widespread throughout Southeast Asia (e.g. Dacrycarpus imbricatus, Podocarpus neriifolius, Dacrydium elatum and Nageia fleuryi). It is plausible that the success of these podocarp competitors may, in fact, limit the southward migration of Pinus. The potential of podocarp species to displace P. krempfii is evidenced by the higher number of podocarp seedlings and apparently strong recruitment underneath many stands of P. krempfii in Vietnam (T. J. Brodribb, pers. obs.).

Historically, pines have not fared well under stable warm and wet climates. Their fossil record suggests that throughout the Cenozoic, pines were only abundant under conditions of climatic upheaval (the onset of glacial cycles) or volcanism (Millar, 1998). Hence it is possible that P. krempfii represents the first step in the genus towards a successful invasion into low-latitude rainforest. By contrast, the Podocarpaceae appear to have been associated with stable, humid climates since the earliest Cenozoic (Hill & Brodribb, 1999) and, as such, retain a strong foothold in tropical rainforest today. Understanding the ecology and interaction between ‘northern’ (Pinaceae) and ‘southern’ (Podocarpaceae) conifers in countries such as Vietnam promises to provide us with new insights into the recent and future movements of these important plant groups.


The authors thank Ton that Minh, Do van Ngoc, Le van Huong, Doan Doan Ai, Nguyen van Hung, Le Canh Nam, Vo Duan, and Nguyen Anh Tuan (Bidoup Nui Ba National Park Service) for logistical support and expertise in Vietnam. This work was funded by an ARC discovery grant DP0559226 (to TB) and an NSF grant IOB-0714156 (to TSF).