Hydathodal leaf teeth of Chloranthus japonicus (Chloranthaceae) prevent guttation-induced flooding of the mesophyll


Dr Taylor S. Feild. E-mail: feild@tulane.edu


Why the leaves of cold temperate deciduous and moisture-loving angiosperms are so often toothed has long puzzled biologists because the functional consequences of teeth remain poorly understood. Here we provide functional and structural evidence that marginal leaf teeth of Chloranthus japonicus, an understory herb, enable the release of guttation sap during root pressure. When guttation from teeth hydathodes was experimentally blocked, we found that the leaf intercellular airspaces became flooded. Measurements of chlorophyll a fluorescence revealed that internal flooding resulted in an inhibition of photosynthesis, most likely through the formation of a film of water within the leaf that reduced CO2 diffusion. Comparing a developmental series of leaves with and without teeth experimentally covered with wax, we found that teeth did not affect overall leaf stomatal conductance and CO2 uptake. However, maximal and effective light-saturation PSII quantum yields of teeth were found to be lower or equal to the surrounding lamina throughout leaf ontogeny. Collectively, our results suggest hydathodes and their development on teeth apices enable the avoidance of mesophyll flooding by root pressure. We discuss how these new findings bear on the potential physiological interpretations of models that apply leaf marginal traits to infer ancient climates.


Leaf margins of temperate, deciduous and temperate/tropical flowering plants from wet soil and low evaporative-demand environments (such as along riparian corridors, forest understories, and oxbow lakes) are frequently toothed (Wolfe 1971, 1979; Baker-Brosh & Peet 1997; Wilf 1997; Gregory-Wodzicki 2000; Burnham et al. 2001; Kowalski 2002; Kowalski & Dilcher 2003). Bailey & Sinnott (1916) also noted that the proportion of tooth-bearing trees increased in the Northern Hemisphere as one moved north. This led to their landmark study showing a strong correlation between the number of toothed species in a flora and regional mean annual temperature. However, the functional significance of leaf teeth, which could resolve whether the mean annual temperature versus leaf-margin correlation has an underlying physiological basis, is enigmatic (Wolfe 1979; Baker-Brosh & Peet 1997; Wilf 1997; Burnham et al. 2001).

Some have hypothesized that leaf teeth are not directly physiologically significant to plants. For instance, resin secretions from teeth in Populus deter leaf-feeding caterpillars (Curtis & Lersten 1974). Also, spinose marginal leaf teeth can impede insects from chewing along the leaf margin (Ehrlich & Raven 1967; Givnish 1979; Brown & Lawton 1991). Alternatively, marginal teeth in some species act as nectaries, releasing sugars to attract pollinators or insects that protect leaves from herbivores (Curtis & Lersten 1974; Belin-Depoux 1989). Another suggestion is that the evolution of teeth, particularly in thin-leaved deciduous angiosperm taxa, may reflect xylem hydraulic constraints within the leaf vasculature to avoid death of tissues within the leaf that are not well-irrigated by veins (Givnish 1979). This is because, all else being equal, toothed leaves in comparison to entire leaves, will have a lower area of mesophyll tissue between secondary veins near the leaf margin that are less well-supplied with water (Zwieniecki et al. 2002) and therefore more likely to die during drought (Givnish 1979). Nonetheless, several adaptive physiological roles for teeth have been suggested.

Leaf teeth are frequently remarked on as sites of water exudation during guttation in a wide-range and large number of taxa (∼ 400 species; Burgerstein 1904; Haberlandt 1914; Bailey & Sinnott 1916; Stocking 1956; Ivanoff 1963; Curtis & Lersten 1974; Fahn 1979; Lersten & Curtis 1985; Belin-Depoux 1989; Takeda, Wisniewski & Glenn 1991). Guttation occurs under conditions of high soil moisture, combined with high humidity, which act in concert to reduce transpiration (Stocking 1956). These conditions enable root pressure to drive water flow from roots to shoots (Haberlandt 1914; Stocking 1956). In many species, water exits the leaf through specialized stomata on the epidermal surfaces of teeth called ‘water pores’ (Burgerstein 1904; Haberlandt 1914; Bailey & Sinnott 1916; Stocking 1956; Ivanoff 1963; Curtis & Lersten 1974; Fahn 1979; Lersten & Curtis 1985; Belin-Depoux 1989; Takeda et al. 1991). Water pores are typically associated with other cells, including tracheary elements and a network of elongate, thin-walled, and colorless mesophyll cells referred to as epithem (Haberlandt 1914; Fahn 1979; Lersten & Curtis 1985). Collectively, water pores, epithem cells, and tracheary elements constitute a hydathode (Haberlandt 1914; Fahn 1979). Although botanists have long remarked on the association between guttation and hydathodal teeth in a large number of flowering plant species (Burgerstein 1904; Haberlandt 1914; Bailey & Sinnott 1916; Stocking 1956; Ivanoff 1963; Curtis & Lersten 1974; Fahn 1979; Lersten & Curtis 1985; Belin-Depoux 1989; Takeda et al. 1991), the functional significance of the interaction remains poorly understood. One suggestion is that water pores act to prevent flooding of the leaf interior when plants produce root pressure (Haberlandt 1914; Rea 1929).

Others have suggested that marginal teeth increase leaf transpiration and heat convection (Gottschlich & Smith 1982; Canny 1990; Wilson, Canny & McCully 1991; Schuepp 1993; Wilf 1997). By a greater perimeter resulting from the zigzagging of tooth sinuses and apices compared to entire leaves of the same area, toothed leaves have been predicted to exhibit greater evaporative and convective potential, owing to a thinner boundary layer over the lamina margin (Vogel 1970; Gottschlich & Smith 1982; Canny 1990; Wilson et al. 1991; Schuepp 1993). Another reason to suspect teeth as sites of active transpiration is that they appear to be hydraulically well supplied. Major veins, with high hydraulic conductance (Zwieniecki et al. 2002; Sack, Streeter & Holbrook 2004), enter marginal teeth of many species (Curtis & Lersten 1974; Fahn 1979; Lersten & Curtis 1985; Canny 1990; Todzia & Keating 1991; Wilson et al. 1991). An advantage of greater transpiration is a decreased likelihood of leaves overheating under conditions of full sunlight and still air (Schuepp 1993). If greater rates of transpiration are coupled to CO2 uptake, teeth may also be sites of greater photosynthesis relative to the bulk of the leaf (Baker-Brosh & Peet 1997; Wilf 1997). Yet, how teeth affect leaf photosynthesis and water vapour gas exchange during leaf development remains uninvestigated.

In the present study we examined the functional interactions between teeth with hydathodes and guttation by first addressing whether teeth prevent flooding of the leaf mesophyll during guttation. We also investigated how teeth affect water loss and photosynthesis in a developmental series of leaves that were experimentally manipulated to block water vapour possibly emanating from teeth. Our experiments focused on the toothed, understory herb Chloranthus japonicus (Chloranthaceae; four genera, approx. 75 species, Todzia 1988). We discuss how insights on leaf teeth physiology and structure in C. japonicus may shed light on why teeth appear to have preferentially evolved in angiosperm lineages from temperate moist environments (Bailey & Sinnott 1916; Wolfe 1971, 1979; Baker-Brosh & Peet 1997; Wilf 1997; Gregory-Wodzicki 2000; Burnham et al. 2001; Kowalski 2002; Kowalski & Dilcher 2003).


Plant species and growth conditions

Chloranthus japonicus Siebold is winter-deciduous perennial herb that produces numerous four-leaved shoots from an underground rhizome. Chloranthus japonicus inhabits a variety of wet, shady, disturbed microsites (e.g. steep stream-banks) as well as partially sun-exposed disturbed sites (such as along irrigation canals and collapsed roadside banks) of subtropical evergreen forests of southern China and temperate deciduous forests of China, Korea, Japan, and southern Siberia (Nianhe & Jérémie 1999; Feild et al. 2004). For our experimental work, we raised 10 plants of C. japonicus in an environmentally controlled glasshouse. Day/night temperatures were controlled at 25 ± 3 °C/22 ± 2 °C, respectively, and light intensities averaged 300 µmol m−2 s−1 of photosynthetically active radiation (PAR) by supplemental metal halide lighting. Relative humidity averaged 60% and CO2 concentrations were 380–400 µL L−1 CO2. The potting medium was a mix of 25% Promix (Sun Gro Horticulture Canada Ltd, Seba Beach, AB, Canada), 25% sand, and 50% topsoil. The plants were watered daily and fertilized once a week with a 20 : 20 : 20 NPK mix.

Leaf anatomical/morphological studies and root pressure measurement

Light, scanning (SEM), and transmission electron microscopy (TEM) observations of leaf tooth structure were made on cryogenically fixed/freeze-substituted immature and mature leaves (n = 5–8 leaves from a single plant; Lam et al. 2001; Thien et al. 2003). In addition, histochemical observations of serial sections of resin-embedded material were made using light microscopy. Histochemical features of leaf tooth structure were characterized on whole mounts and free-hand sections of fresh tissue using the following stains: (a) phloroglucinol 1% in 95% ethanol (Ruzin 1999); (b) 0.01% auramine-O in 0.05 m Tris/HCl under UV light for cutin (Heslop-Harrison & Shivanna 1977); (c) 1% alcian blue in 3% acetic acid for acidic polysaccharides, pectins, and mucilages (Jensen 1962; Benes 1968); (d) 0.05% ruthenium red for pectic substances (Gurr 1965); and (e) aniline blue for detection of callose (Martin 1959). Leaf teeth on mature as well as immature leaves (n = 3–5 leaves from a single plant) were cleared in methyl salicylate for observations of tracheary element distribution (Stelley et al. 1984) without Mayer's hemalum. Because leaf teeth may affect gradients of heat and water vapour exchange across the leaf (Vogel 1970; Gottschlich & Smith 1982; Canny 1990; Schuepp 1993), we measured the relative contribution of teeth area to overall leaf area (tooth–area ratio; Huff, Wilf & Azumah 2003). Leaves and centimetre scales were photographed on a white background using a Nikon Coolpix 4500 digital camera at 2048 × 1536 pixels resolution (Nikon, Melville, NY, USA), and an image analysis program (Scion 4.2; Scion Corporation, Frederick, MD, USA) was used to measure surface area. Wavy or curled portions of leaves were pressed flat under glass to minimize image distortion. Leaf teeth areas were manually determined by drawing lines around the hydathodal tips and then bounding the tooth with a straight line from sinus to sinus (Huff et al. 2003). Root pressures were measured from de-topped shoots (cut underwater at 2 cm above the rhizome and attached to a water-filled tube) using a handheld electronic manometer.

Leaf teeth plugging

To examine the influence of marginal teeth on leaf water vapour loss and CO2 uptake, we carefully painted molten paraffin wax (melting point 45 °C, Sigma-Aldrich, St Louis, MO, USA) from tooth sinus-to-sinus on the top (adaxial) and bottom (abaxial) sides of the leaf teeth. We experimented with other substances (i.e. cyanoacrylate and epoxy glues, cooking and vacuum greases, liquid wound covering, petroleum jelly, polyester wax, and silicone) to block transpiration, but each of these methods caused different problems. For example, fumes released during curing of epoxy resin, cyanoacrylate glue, and silicone killed mesophyll cells up to 3 mm away from the apex of each tooth glued. Greases and petroleum jelly did not visually damage the leaf tissue, but we found that these substances were less effective than paraffin at blocking transpiration. Steady-state water loss rates of detached leaves completely coated with vacuum grease, cooking grease, polyester wax, and petroleum jelly placed in the dark were 5.6 × 10−4, 6.5 × 10−4, 1.8 × 10−4 and 8.6 × 10−4 g cm−2 min−1, respectively. Rates of water loss in leaves completely coated in paraffin were 8 × 10−5 g cm−2 min−1 as compared to 5.3 × 10−3 g cm−2 min−1. Water loss rates were measured in the dark and at constant humidity (10% relative humidity) and temperature (22 ± 1.3 °C, measured with a hand-held thermocouple reader) by placing leaves on an electronic balance interfaced to a computer. Paraffin, however, did not block guttation, since water emerging from the teeth pushed the paraffin off. Consequently, we applied a clear, non-toxic, and water-proof adhesive (liquid band-aid; Band-Aid Water Block Plus; Johnson & Johnson, Mississauga, Ontario, Canada) to the adaxial surfaces of leaf teeth. Although the adhesive was not as effective as paraffin in reducing tooth transpiration (minimum water loss rates were 3.6 × 10−4 g cm−2 min−1), it did prevent guttation. When carefully applied, no differences in maximum photosystem II (PSII) quantum yield, as determined by chlorophyll fluorescence (see below), were observed between wax and liquid band-aid-treated teeth compared to adjacent untreated teeth in C. japonicus (data not shown).

Leaf gas-exchange and chlorophyll a fluorescence measurements

An open-flow [infra-red gas analyzer (IRGA) LI-6400; Li-COR, Lincoln, NB, USA] was used to measure water vapour and carbon dioxide fluxes from C. japonicus leaves. We selected five healthy shoots, bearing leaves from 20 to 120 cm2, from five different plants for measurements of net CO2 uptake (A, µmol CO2 m−2 s−1) and stomatal conductance (gs, mol H2O m−2 s−1) of the leaf margin of C. japonicus. Leaves were measured from 0900 to 1130 h to allow for at least 2 h of sunlight exposure to adequately light-induce leaf photosynthesis. Then, we gently sealed a portion of a leaf (only the distal third of the leaf blade with nine to 12 marginal teeth) into the cuvette under conditions of 200 µmol m−2 s−1 PAR, 26 °C, 380 µL L−1 CO2, and 1 kPa VPD (‘prelight curve conditions’). PAR was generated using a mixed red and blue light-emitting diode array. Next, we traced the outline of the cuvette gasket onto the leaf edge to mark the measured portion of the leaf. After steady-state fluxes of CO2 and H2O were established (approximately 5 min), the light intensity was increased to 500 µmol m−2 s−1. Values of A and gs at steady-state were collected for light intensities of 500, 400, 250, 100, 50, and 0 µmol m−2 s−1 PAR. Previous fieldwork on C. japonicus indicated that photosynthesis of fully expanded leaves reached 90% saturation around 400 µmol m−2 s−1 PAR (Feild et al. 2004), so we chose 500 µmol m−2 s−1 PAR as a maximum light intensity to avoid photoinhibiting leaves. Leaves took between 5 and 15 min to equilibrate to a new irradiance, and output signals from both IRGAs were matched to correct for drift before taking data. After completing a light-response curve, we removed the leaf from the cuvette and coated the marginal teeth within the measured leaf area with molten paraffin. The leaf was then returned to the leaf prelight curve environmental conditions in the cuvette until steady-state CO2 and water vapour fluxes were established. Once steady state was reached, the light responses of A and gs were re-measured. We also investigated whether teeth were especially active in water loss by covering an area equal to that contributed by leaf teeth with paraffin located in middle of the leaf blade. With practice, we were able to cover a portion of the bulk leaf (10–15 mm away from the toothed margin) with a roughly square area of paraffin that was within ± 10–15% of the surface area contributed by teeth in a given leaf sample. During these measurements, the teeth were left uncoated. We corrected leaf gas-exchange fluxes for the amount of area lost for gas-exchange resulting by sealing teeth with paraffin using digital photographs of the measured leaf portion next to a scale bar and subtracting the area covered by wax.

For leaves smaller than 20 cm2, which were difficult to fit into the gas-exchange cuvette without injury, we measured tooth photosynthesis using chlorophyll (Chl) a fluorescence emission. The Chl fluorescence parameters were determined with a pulse amplitude-modulated fluorometer (PAM-2100; Walz, Effeltrich, Germany). First, we characterized differences in maximum, dark-adapted PSII photon yield (Fv/Fm) of leaf teeth compared to the bulk lamina in a developmental series of leaves (0.5–120 cm2). We isolated the teeth fluorescence from the rest of the leaf by carefully positioning the measuring beam from the fibre-optic bundle of the PAM-2100 so that only teeth along the leaf margin were exposed to the measuring beam. Fluorescence from the rest of the leaf was blocked with a glass plate covered in non-fluorescent tape. We defined the bulk leaf areoles as areas at least 15 mm away from the leaf margin.

Plants of C. japonicus were taken from the glasshouse in the morning, bagged in plastic, and placed in complete darkness for 3 h to relax light-induced down-regulation of PSII activity (Krause & Weis 1991). Minimal fluorescence emission (Fo) was determined using a non-actinic measuring beam, following exposure to 10 s of far-red illumination (710 nm), to ensure maximal PSII re-oxidation (Feild, Nebal & Ort 1998). A 600-ms saturation pulse was then applied to momentarily close the pool of active PSII reaction centres for determination of Fm. Dark-adapted values for Fm and Fo were used to calculate maximum PSII quantum yield as: Fv/Fm = (Fm − Fo)/Fm (Krause & Weis 1991). We also compared the saturated-light response of effective PSII photon yield (ΔF/Fm′; Genty, Briantais & Baker 1989) of teeth and the bulk of the lamina (separated as above) in a developmental series of leaves. ΔF/Fm′ was calculated as: ΔF/Fm′ = (Fm′ − F)/Fm′, where F is the fluorescence yield of the light-adapted sample at steady state and Fm′ is the maximum light-adapted fluorescence yield when a saturating light pulse is superimposed on the prevailing light (Genty et al. 1989). Leaves were exposed to a saturating light treatment (450 µmol m−2 s−1 PAR) in a humidified chamber with controlled gas concentration [380 µL L−1 CO2, 21% (v/v) oxygen balanced with nitrogen gas] and temperature (24 ± 2 °C) until steady F-values were reached (approximately 2 min).

Photosynthetic responses to guttation-blockage

When guttating teeth were coated with liquid band aid, the intercellular airspaces of the leaf flooded (see below). We examined how guttation back-up impacted leaf physiology by measuring Fv/Fm and photosynthetic induction of ΔF/Fm′ of teeth with flooded airspaces as compared to teeth and areoles near the middle of the lamina with gas-filled intercellular airspaces on three leaves (90–120 cm2) from three different plants. At the beginning of an experiment, we coated three randomly selected teeth from the leaf apex with non-fluorescent liquid-band aid and left several adjacent teeth untreated. After the polymer dried (< 2 min), the whole plant was watered to saturation and placed in a humidified Plexiglas chamber. Time courses of Fv/Fm were based on averages from three areoles measured on a single leaf from the following leaf regions: flooded teeth ‘plugged’ with glue; unglued, freely guttating (and thus free of liquid water); the middle of the lamina. The rest of the leaf was protected from stray light during applications of saturation pulses using small aluminium leaf clips (Walz, Effeltrich, Germany). After 1 h of guttation, we determined the induction of ΔF/Fm′ to saturating light (i.e. 450 µmol photons m−2 s−1 PAR) from the same areas of the leaf. Saturating pulses were applied every min for calculation of ΔF/Fm′. To examine how leaves of C. japonicus responded to prolonged internal flooding, we followed changes in Fv/Fm from the three sample locations (as above) for 48 h.

Leaf wettability

Hydrophobicity of the leaf surfaces was determined by measuring the contact angle of a 0.5-mm3 droplet of distilled water pipetted onto C. japonicus leaves (Adam 1963; Brewer & Smith 1997). The contact angle was taken as the angle (θ) of a line tangent to a droplet through the point of contact between the droplet and a horizontal leaf surface. The θ values were measured for the abaxial/adaxial surfaces of the teeth and the abaxial/adaxial surfaces of the lamina from digital photographs (at 63×) using a horizontally mounted stereomicroscope. Before measurement, margins of leaves were affixed to glass slides by double-sided adhesive tape to ensure an even surface.

We also measured the amount of water retained on a leaf with and without teeth following immersion. A leaf was weighed on a balance, and then it was dipped vertically into a bath of distilled water. After 5 s, the leaf was pulled out and allowed to drip vertically for 30 s before weighing. A 30-s period was selected to allow leaves sufficient time to drip before weighing. Then, we carefully removed the teeth from the same leaf, cutting the teeth off sinus-to-sinus with a double-edged razor blade (the apical drip-tip tooth was left intact). The toothless leaf was carefully dried with paper towels, and then re-dipped in water so the measurement sequence could be repeated.


Leaf morphology, teeth structure, and guttation patterns

As leaves expanded, tooth area increased, reaching 2 cm2 in fully expanded leaves (Fig. 1a). The contribution of teeth to total leaf area (leaf teeth ratio) was greatest in the smallest leaves measured (0.8 cm2). Teeth area contribution decreased from approximately 13% to less than 2% in fully expanded (approximately 120 cm2) leaves (Fig. 1b). Once fully expanded, leaves of C. japonicus began to senesce, as evidenced by chlorosis along the leaf margin. Chloranthus japonicus leaves bore on average 48 teeth per leaf (SD, 7; n = 22), which protruded 2–5 mm from the leaf margin (average = 4.2, SD 1.5; n = 45; Fig. 2a).

Figure 1.

Morphometrics of expanding Chloranthus japonicus leaves. (a) Developmental changes in tooth area (cm2) in relation to leaf area (cm2) (b) Changes in tooth area ratio (tooth area leaf−1 area) in relation to total leaf area (cm2). Note that the relative contribution of teeth to total leaf area decreases with leaf size. Each point represents a single leaf.

Figure 2.

Shoot morphology and leaf tooth structure/development in Chloranthus japonicus. (a) Flooded leaf margin, note that the airspaces of the areoles near the leaf margin are water-filled (indicated by a dark green colour) whereas ones closer to the middle of the leaf remain gas-filled. Flooding was induced by blocking tooth hydathodes with liquid band aid. (b) A tooth cleared with methyl salicylate. Xylem enters the tooth with one vein entering the tooth head-on and two lateral veins. An arrow points to dark-staining substances in the tooth (c) SEM micrograph of the adaxial surface of a mature leaf tooth. Note the water pores located at the apex of the tooth. (d) SEM micrograph of the tooth abaxial surface. Note the absence of stomata at the tooth apex, and that stomata are only developed in areoles away from the tooth. Each micrograph represents a representative image based on three to five sampled leaves. Scale bars: (a), 8 mm; (b), 1 mm; (c) and (d), 250 µm.

Tooth tips of methyl-salicylate cleared leaves stained darkly (Fig. 2b). Tip coloration may result from tannins/essential oil bodies, which have been previously reported in Chloranthus leaves (Todzia 1988). With the exception of tracheary elements, leaf teeth were histochemically negative for phloroglucinol staining indicating an absence of lignin/schlerencyhma (observation not shown). Tooth tips of C. japonicus contained a yellowish, opaque, glandular region composed of numerous sinuously walled cells that are comprised of a primary wall rich in pectins based on positive staining for alcian blue and ruthenium red (observation not shown). This region is referred to as epithem (sensuHaberlandt 1914; Fig. 2b).

Abaxial and adaxial surfaces of teeth differed in stomatal development. Two types of stomata occur on the adaxial surface of a leaf tooth. One group of stomata (9–15 in number) appeared first on the adaxial tooth tip surface on leaves as small as 3 mm. These stomata had guard cells slightly sunken below the epidermis, and a smooth surrounding cuticle (Fig. 2c). A second group of stomata (15–20) developed later (i.e. on leaves 5–7 mm long) at the base of the tooth adaxial surface. These stomata were flush or slightly higher than the epidermal plane and were embedded in striated cuticle (Fig. 2c). Stomata were absent from the abaxial tooth tip and developed away from the base of the tooth beginning in leaves approximately 15 mm long (Fig. 2d). Abaxial stomata were morphologically similar to adaxial stomata located at the tooth base (Fig. 2d).

Compared to leaf mesophyll cells, cells of the epithem were smaller, more elongate and tightly packed, which resulted in a near absence of intercellular airspaces (Figs 3a & b). Tooth epithem cells also exhibited abundant connections to xylem tracheary elements that occurred along the elongate sinuous walls of the epithem (Fig. 3c). The epithem contained numerous starch-containing chloroplasts and mitochondria (Fig. 3c) and extensive endomembrane system (observation not shown). Based on a lack of cellular aniline-blue staining, no phloem parenchyma cells and sieve-tubes were observed in the tips of C. japonicus teeth (observation not shown). Intercellular air spaces in the epithem occurred only in the substomatal chambers at the tooth tip (Fig. 3d). At the tips of hydathodes, epithem cells rimming the stomatal chambers were surrounded by calcium-containing crystals (Fig. 3d). Consistent with this observation, white chalky encrustations were observed on the tips of C. japonicus teeth that had recently guttated (data not shown).

Figure 3.

Light and TEM micrographs of Chloranthus japonicus leaf teeth anatomy. (a) For comparison, a cross-section through the leaf mesophyll, illustrating an abundance of intercellular airspaces (denoted by *). (b) Longitudinal section through the tooth epithem tissue (labelled E); visible are abundant epithem cells, which are the elongate sinuously walled cells, and occasional tracheary elements (labelled T), which are the dark, mutibarred cells threading through epithem tissue. Note the absence of intercellular airspaces. (c) Epithem cell with numerous organelles juxtaposed with tracheary elements. (d) Cross-section of a water pore chamber located at the leaf adaxial surface. The white spheres with dark grey halos (denoted by arrows) are calcium crystals in the apoplast and vacuoles of cells rimming the chamber. Each micrograph represents a representative image based on three to five sampled leaves. Scale bars: (a) and (b), 100 µm; (c), 10 µm; (d), 15 µm.

After 20 min exposure to 100% relative humidity, all leaves of C. japonicus, from expanding (≥ 3 mm long) to fully expanded (approximately 120 cm2), became studded with droplets of water on the adaxial surface of each tooth tip. Viewing guttation under a stereomicroscope (at 100×) revealed that water emerged specifically from adaxial stomata at the tooth tip (Fig. 2c). This observation indicates that these stomata function as ‘water pores’ (Haberlandt 1914). Guttation also occurred from the adaxial tips of stem stipules and bracts on inflorescences of C. japonicus in all developmental stages (from 3 to 50 mm long). Shoots detached underwater and placed in the dark or under weak light (approximately 15 µmol photons m−2 s−1 PAR), however, did not exhibit marginal or inflorescence guttation. Measurements of pressures built up from de-topped shoots of C. japonicus ranged from 0.02 to 0.065 MPa (mean = 0.045 MPa ± 0.012 SD; n = 7 plants). Guttation in shoots, detached underwater, could be restored by applying approximately 0.05 MPa of pressure (observation not shown). Eventually, droplets forming along the leaf margin rolled off of the tip of each tooth and did not coalesce to form a water film over the top or bottom surfaces of the leaf.

Leaf hydathodal teeth and root pressure-induced mesophyll flooding

Teeth on leaves of intact plants placed under high humidity and coated with liquid-band aid polymer flooded, as noted by the appearance of dark green coloration (Fig. 2a). In contrast, the intercellular air spaces of guttating leaves remained gas-filled since the leaf undersurface exhibited a matte light-green colour, which indicates light scattering by internal air–water interfaces (Fig. 2a; Smith et al. 1997). Patterns of mesophyll flooding during guttation blockage were not uniform across leaves of C. japonicus. Filling of the intercellular air spaces occurred only in the leaf periphery, in areoles bounded by tertiary veins (Fig. 2a).

Exposure to saturating PAR (450 µmol m−2 s−1) for 15 min led to a 50% and 40% reduction in PSII quantum yields (ΔF/Fm′) from areoles of flooded teeth compared to areoles of non-flooded teeth from the middle of the leaf, respectively (Fig. 4). We found that flooded leaves detached from guttating plants required approximately 10 min to clear out under 40% relative humidity (RH) and saturating PAR – conditions that C. japonicus only naturally encounters during brief sunflecks. However under more typical understory conditions of low PAR and high humidity (5 µmol m−2 s−1 PAR, RH > 90%, and no wind), we found that the intercellular airspaces of areoles required 4 h to become gas-filled.

Figure 4.

Responses of photosynthetic activity to flooding of the mesophyll induced by root pressure in C. japonicus from different portions of the leaf. Fluorescence parameters were measured under the conditions described in the Materials and Methods section. Symbols are ▴, flooded leaf teeth areoles; ▵ and ○ are tooth areoles and areoles from the bulk of the leaf with non-flooded intercellular airspaces, respectively. Each point represents an average of three measurements taken from three different leaves (n = 3), and error bars denote the standard deviation. The bar above (a) denotes times when measurements were made in darkness (filled) and under saturating illumination (open, 450 µmol m−2 s−1 PAR).

Leaf wettability

The leaf adaxial surface on both the tooth and lamina was more hydrophobic than the abaxial surface, based on greater contact angles [tooth abaxial mean = 54 ± 2°, lamina abaxial mean = 56 ± 3° compared with the tooth adaxial mean = 116 ± 3°, lamina adaxial mean = 116 ± 3°; n = 10 leaves from five different plants for each mean). We also observed that C. japonicus leaves with teeth retained less water when dipped into a water bath compared to the same leaves with their teeth experimentally removed (Fig. 5). As leaves increased in surface area, the amount of water retained on the leaf increased. The relative difference in amount of water retained on leaves between toothed and untoothed leaves also decreased slightly as leaves became larger, from 46% in 20 cm2 leaves to 38% in approximately 120 cm2 leaves.

Figure 5.

Retention of water on leaf surfaces of Chloranthus japonicus leaves with and with teeth experimentally removed. Open triangles (▵) represent with teeth intact and open circles (○) are the same leaves following severing of the leaf teeth. Each point in represents five leaves from five plants (n = 5), and error bars are standard deviations about the mean.

Leaf tooth photosynthesis and stomatal conductance

No differences in light-saturated leaf CO2 uptake (A) and stomatal conductances (gs) to water vapour were observed among leaves 20–120 cm2 in area with and without their teeth coated with paraffin in comparison with leaves with a tooth surface area equivalent of paraffin placed on the middle of the leaf (Fig. 6a & b). There were also no differences observed in the initial slope of the light response of A among the control and two experimental treatments (data not shown). Maximum (Fv/Fm) and effective (ΔF/Fm′) PSII photon yields under saturating light (450 µmol m−2 s−1 PAR) were generally higher for the middle part of the lamina compared to the leaf margin, although the overall ontogenetic patterns in photosynthetic performance were similar between different areas of the leaf (Fig. 7a & b). Early in leaf expansion (> 1 cm2), we observed the largest percentage differences in photosynthetic performance, with Fv/Fm approximately 8% and ΔF/Fm′ approximately 30% lower in leaf teeth in comparison to the middle of the leaf. Differences in PSII photon yields between teeth and the surrounding mesophyll decreased as leaves expanded, such that leaves approximately 20 cm2 and approximately 30 cm2 in size exhibited similar values ΔF/Fm′ and Fv/Fm. When leaves began to senesce (> 110 cm2), photosynthetic properties diverged again, with teeth Fv/Fm and ΔF/Fm′ values 3 and 15% lower than for the middle of the leaf, respectively (Fig. 7a & b). Under glasshouse conditions, Fv/Fm and ΔF/Fm′ patterns during leaf expansion were not dependent on the time of year that a shoot cohort developed; the first shoot cohort expanding in May exhibited similar photosynthetic performance as later cohorts emerging in July and August (data not shown).

Figure 6.

Effects of different wax-treatments on light-saturated net CO2 uptake rate (a) and stomatal conductance to water vapour (b). For (a) and (b), open circles (○) are untreated leaves, open triangles (▵) leaves with their teeth coated with wax, and open squares (□), leaf samples with an area of wax painted on the bulk of the leaf equivalent to the surface area contributed by teeth for the sample. Error bars denote standard deviations about the mean. Leaves were measured under controlled temperature and CO2 concentration as described in the Methods section. Values are averages of five leaves from five different plants (n = 5), and error bars denote the standard deviation about the mean.

Figure 7.

Maximum (Fv/Fm) and effective (ΔF/Fm′) PSII photon yields measured under saturating illumination (450 µmol m−2 s−1 PAR) for teeth (▵) compared to the middle of the lamina (○) of Chloranthus japonicus leaves. Fluorescence parameters were determined as described in the Materials and Methods section. Values are averages of five leaves from five different plants (n = 5), and error bars denote the standard deviation about the mean.


Our results indicate that hydathodal teeth of Chloranthus japonicus prevent flooding of the leaf interior during guttation. When guttation was experimentally blocked, we found that the leaf intercellular air spaces became infiltrated with water as root pressure irrigated the venation system and water spilled over into the surrounding mesophyll (Fig. 2a). Hydathodes of C. japonicus appeared to be functionally ‘passive’ (sensuHaberlandt 1914), depending upon root pressure for guttation (Stocking 1956; Sperry 1983). In support of this conclusion, shoots detached underwater, thus lacking root–shoot continuity, and placed in darkness or under low light (approximately 15 µmol photons m−2 s−1 PAR) did not guttate. Guttation, however, was restored by applying a pressure of 0.05 MPa, which corresponds to measured root pressures of C. japonicus. Interestingly, root pressure induced-flooding of the leaf mesophyll was not uniform over the leaf surface. Filling of the intercellular air spaces, as evidenced by a colour change from matte green to dark translucent green, occurred only in the peripheral areoles of the leaf that were bounded by tertiary veins (Fig. 2a). Studies on patterns of water infiltration in leaves under positive pressure have generally demonstrated that flooding occurs homogenously from base to tip (Nardini, Tyree, & Salleo 2001), indicating an equitable distribution of water (Roth-Nebelsick et al. 2001; Zwieniecki et al. 2002). However in C. japonicus, water flow occurs preferentially out toward the margin under low positive pressure. Future studies are needed to understand how veins located near the leaf periphery are more leaky under pressure on the marginal side and how this anatomy may affect the distribution of water within leaves under tension.

Our experiments revealed that when leaf teeth were experimentally plugged, congestion of the intercellular air spaces with water impedes photosynthesis. Effective PSII photon yields (ΔF/Fm′) were 40% lower in flooded teeth compared to freely guttating, non-flooded teeth (Fig. 4). However, maximum PSII photon yields (Fv/Fm), and thus the capacity for photosynthetic electron transport, were not different in flooded leaves before and after a brief exposure to saturating light that simulated a sunfleck (Fig. 4). Because CO2 diffuses through water 10 000 times more slowly than in air (Denny 1993), we suggest that as guttation backs-up the formation of a film of water within the intercellular air spaces constricts CO2 diffusion and therefore accounts for the decrease in photosynthesis observed in flooded leaves. However, ΔF/Fm′ of flooded leaf portions did not collapse completely, suggesting that photorespiration and mitochondrial CO2 reduction remain significant during intercellular inundation.

Once leaves of C. japonicus flooded, returning to a gas-filled intercellular air space system was protracted. Under high evaporative conditions [i.e. low ambient humidity (approximately 40%) and saturating PAR; conditions that understorey plants like C. japonicus would only encounter during brief sunflecks (> 2 min long); Chazdon & Pearcy (1996), flooded leaves required approximately 10 min to return to an open air space system. However under more typical understorey conditions of low PAR and high humidity (5 µmol m−2 s−1 PAR, RH > 90%), the intercellular air spaces contained liquid water for nearly 4 h. In contrast, leaves with non-manipulated teeth continually purged guttation sap to the leaf exterior and did not become congested. Collectively, these observations support earlier suggestions that hydathodes act as release valves for guttation sap, thus avoiding injection of the intercellular air spaces with water during root pressure (Haberlandt 1914; Rea 1929).

Based on our anatomical observations, the hydathodal tips of C. japonicus teeth appeared similar to those of other angiosperms in possessing epithem tissue with abundant starch-containing chloroplasts and mitochondria as well as vascular tissue that contains xylem but not phloem (Haberlandt 1914; Curtis & Lersten 1974; Fahn 1979; Lersten & Curtis 1985; Belin-Depoux 1989; Takeda et al. 1991). These observations suggests that epithem cells are metabolically active. In support, epithem appears to function as a crossroads for diverse metabolic/developmental processes within leaves. For example, epithem-containing teeth of Arabidopsis leaves are often illuminated in in situ hybridization studies of metabolic and developmental gene expression, including hormone (i.e. auxin, cytokinin), flavonoid, nitrogen, potassium, and phosphorous metabolism (Largarde et al. 1996; Aloni et al. 2003; Burkle et al. 2003; Pilot et al. 2004; Wang et al. 2004). Furthermore, numerous substances such as carbohydrates, amino acids, ions (calcium, potassium, boron), and proteins (peroxidases) have been identified in guttation sap (Curtis & Lersten 1974; Belin-Depoux 1989; Magwa, Lindner, & Brand 1993; Fukui, Fukui, & Alvarez 1999; Kerstetter, Zepp, & Carreira 1998; Grunwald et al. 2003). Epithem is likely to be the source for many of these substances, owing to a high density of metabolic organelles. In addition, epithem cells are well poised to secrete substances into the guttation sap, since sap is sieved through the epithem before exiting the leaf (Wilson et al. 1991). Our anatomical observations revealed that epithem cells of C. japonicus contain calcium crystals in the apoplast and vacuoles. Calcium crystals, complexed to an unknown anion, occurred only in epithem cells that line the subwater pore chamber (Fig. 3d). Similar appearing crystals under TEM have been identified as calcium-dominated from the leaves and stems of several species (Mazen, Zhang, & Franceschi 2004).

Since guttation occurs from entire-margined leaves in some species – through hydathodes along the leaf blade margin (particularly in monocots which are not known to produce teeth, Ivanoff 1963; Fahn 1979) or through the epidermis directly (Meidner 1977; Wagner, Wang, & Shepard 2004) – why are hydathodes so frequently developed on the tips of marginal teeth (Burgerstein 1904; Haberlandt 1914; Bailey & Sinnott 1916; Stocking 1956; Curtis & Lersten 1974; Fahn 1979; Lersten & Curtis 1985; Donnelly, Skelton & Nelles 1987; Belin-Depoux 1989; Takeda et al. 1991)? Our results suggest that one functional consequence of developing hydathodes on marginal teeth is that guttation droplets are more easily shed. We found that toothed leaves of C. japonicus retained less water compared to the same leaves with their teeth severed sinus-to-sinus, thus in effect entire-margined (Fig. 5). Although vertical immersion of leaves is not the same wetting process as that occurring during guttation where droplets bead along the margin, these experiments demonstrate that water droplets fall off more readily from a serrated margin compared to a smooth one, due to decreased surface area for adhesion (Lightbody 1985). Thus, marginal teeth of C. japonicus appear to function analogously to leaf drip tips by increasing leaf water runoff (Lightbody 1985; Ivey & DeSilva 2001).

Interestingly, if guttation droplets stay on leaves too long, they can be detrimental. For instance, resorption of guttation droplets once transpiration resumes can injure leaves, producing marginal necrosis – a disease called ‘tipburn’ (Curtis 1943; Ivanoff 1963). The causes of tipburn are unclear, but high concentrations of excreted salts/antimicrobial proteins present in guttation sap may injury mesophyll cells. Furthermore, spores of pathogenic fungi and bacteria imbibed with guttation droplets resting on the leaf margin may cause infection (Curtis 1943; Carlton, Braun, & Gleason 1998). Another cost of prolonged retention of guttation droplets is that they may coalesce to form a continuous film of water over the leaf. External water films can favour pathogen invasion, epiphyll colonization, and leaching of nutrients by over-hydrating the cuticle (Curtis 1943). An external water film covering stomata-bearing areas may also reduce photosynthesis by increasing CO2 diffusion resistance (Brewer & Smith 1997; Feild et al. 1998). However, an effect of water films on leaf CO2 uptake in C. japonicus is sidestepped by spatially separating water pores and stomata–water pores occur on the adaxial tooth tip surface whereas stomata occur on the leaf abaxial side and not the abaxial tooth tip surface (Fig. 2c & d). Contact angle measurements also demonstrate that the adaxial surfaces of C. japonicus leaves are more hydrophobic than the leaf undersurface, which further increases the shedding of water. Species with entire-margin guttation either occur in more evaporative environments or, if under shade, exhibit other means for shedding guttation sap (e.g. waxy/curvy margins; Meidner 1977; Wagner et al. 2004; T.S. Feild, unpublished observations, 2004).

Although leaf teeth have been suggested to increase leaf transpiration and photosynthesis (Canny 1990; Wilson et al. 1991; Wilf 1997), our gas-exchange measurements indicated that teeth of C. japonicus did not function as ‘cool sites’ for intense transpirational water loss or CO2 assimilation (Fig. 6a & b). Stomatal conductances and rates of net CO2 uptake in leaves coated with paraffin to block tooth gas-exchange were indistinguishable from leaves with their teeth uncovered or from leaves with freely transpiring teeth but with a paraffin covering equal to the tooth area placed in the middle of leaf. However, gas-exchange data cannot rule out a ‘tooth effect’ on leaf water loss/photosynthesis in C. japonicus. For instance, our procedure, which involved frequent disturbance of the leaf (i.e. the cuvette was attached, then removed, wax applied, and the cuvette re-attached) could have induce physiological transients in stomatal opening thus obscuring the effect of teeth on leaf water loss and carbon uptake. This explanation may account for the high variability among measured leaves (Fig. 6a & b). Furthermore, when leaves are manipulated with wax, gradients of water flow within the leaf vasculature are disturbed. Thus, stomatal opening in portions of leaves that are not covered by paraffin may increase as water previously destined to areas of the leaf shutdown by paraffin become available. Nonetheless even if cuvette gas-exchange approaches reveal tooth effects, because these methods reduce the leaf boundary layer it may be difficult to apply these results to natural conditions (Gottschlich & Smith 1982; Schuepp 1993). However, independent measurements of tooth photosynthetic activity, based on Chl fluorescence emission that does not disrupt the leaf boundary layer, demonstrated that teeth of C. japonicus have lower photosynthesis. Fv/Fm and ΔF/Fm′ yields were found to be lower or equal to the surrounding mesophyll during leaf development (Fig. 7a & b).

Our results provide a new context for discussions over the physiological mechanisms underlying correlations between leaf margin morphology and regional mean annual temperature (Wolfe 1971, 1979; Wilf 1997; Gregory-Wodzicki 2000; Burnham et al. 2001; Kowalski 2002; Kowalski & Dilcher 2003). Based on this correlation, paleobotanists have used the margin morphology of a fossil flora as a ‘paleothermometer’ of the pervading climate before deposition (Wolfe 1971, 1979; Wilf 1997; Gregory-Wodzicki 2000; Burnham et al. 2001; Kowalski 2002; Kowalski & Dilcher 2003; Greenwood et al. 2004). Until now, adaptive hypotheses for tooth evaluation have focused on the view that teeth increase leaf physiological performance, acting as microsites of greater photosynthesis and transpiration (Baker-Brosh & Peet 1997; Wilf 1997; Burnham et al. 2001). Motivating this idea, Baker-Brosh & Peet (1997) found that teeth in some sun-adapted temperate deciduous angiosperms provided an early, but brief pulse of photosynthates during the expansion of spring leaves. Based on snapshot observations of 14CO2 uptake, tooth mesophyll cells of immature Acer, Carya, Liquidambar, and Ulmus leaves (approximately 15 mm long) became photosynthetically active before the rest of the leaf. However like C. japonicus, understorey deciduous toothed shrubs (Rubus and Viburnum species) consistently lacked greater photosynthesis in teeth relative to the rest of the leaf during expansion (Baker-Brosh & Peet 1997). To link these observations to the colder mean annual temperature/toothed margin correlation, early tooth photosynthesis was suggested to allow leaves to become carbon sources earlier in the year – a potential advantage in seasonally cold-temperate climates with short growing seasons (Baker-Brosh & Peet 1997). But how significant photosynthate generated by teeth is to spring leaf out relative to other carbon sources, such as stored sugars in the stem xylem, is unclear. Indeed, tooth photosynthetic activity may be a by-product of early hydathode maturation for leaf developmental patterning since hydathodes appear to be important sources for free-auxin production during vascular differentiation (Aloni et al. 2003). Alternatively, tooth photosynthesis may provide carbon for the local production of auxin and other substances regulating leaf development.

Perhaps a more relevant hypothesis as to why leaves of mesic temperate deciduous angiosperms are frequently toothed is that teeth favour opportunistic leaf expansion, not as sources of carbon, but in enabling root pressure to drive leaf expansion since an inhibition of subsequent photosynthesis by a flooded intercellular air space system can be avoided. Root pressure may increase leaf expansion directly by maintaining leaf turgor a physiological set point (Van Volkenburg 1999). Or, root pressure may favour leaf expansion indirectly by refilling winter-accumulated freeze–thaw embolized vessels in the stem xylem or emboli accumulating following springtime frosts in newly expanding shoots (Sperry et al. 1994). Importantly, favourable conditions (wet soils with cool, moist evening air; Stocking 1956) for guttation pervade spring climates of the Northern Hemisphere temperate zone. Extending this idea, we can also begin to understand why riparian plants exhibit a biased distribution of toothed leaves (Wolfe 1971; Burnham et al. 2001; Kowalski & Dilcher 2003). Root pressure, favoured by frequently wet soils and locally elevated humidity along watercourses, may be important in the expansion of new shoots literally in the wake of frequent traumas that remove parts of plants. Although a pervasive role of leaf teeth in purging guttation sap (as in Chloranthus) is suggested by the widespread correlation between leaf teeth and hydathodes, much more experimental work is needed. Future approaches comparing diverse lineages that represent different ecological guilds (e.g. understorey shrubs and herbs versus early and late-successional trees) from temperate deciduous forests compared with lowland tropical rainforests (where toothed taxa are generally rare, but see Burnham et al. 2001) will be useful in testing potential functional linkages among root pressure, hydathodes, and leaf expansion.


We thank Bruce Hall and Andrew Petrie for excellent assistance in maintaining Chloranthus plants used in this study. We also thank H.S. Feild for help in transporting plants from Alabama to Toronto. This manuscript was also improved by helpful comments from Nan Crystal Arens, Tim Brodribb, Robin Burnham, Kevin Boyce, and Dana Royer. This research was supported by NSERC Discovery Grants to Taylor Feild and Tammy Sage as well as a NSERC USRA fellowship Christine Czerniak.