Rapid nitrogen transfer from ectomycorrhizal pines to adjacent ectomycorrhizal and arbuscular mycorrhizal plants in a California oak woodland

Authors


Author for correspondence: William R. Horwath Tel: +1 530 754 6029 Fax: +1 530 752 1552 Email: wrhorwath@ucdavis.edu

Summary

  • • Nitrogen transfer among plants in a California oak woodland was examined in a pulse-labeling study using 15N. The study was designed to examine N movement among plants that were mycorrhizal with ectomycorrhizas (EM), arbuscular mycorrhizas (AM), or both.
  • • Isotopically enriched N (K15NO3) was applied to gray pine (Pinus sabiniana) foliage (donor) and traced to neighboring gray pine, blue oak (Quercus douglasii), buckbrush (Ceanothus cuneatus) and herbaceous annuals (Cynosurus echinatus, Torilis arvensis and Trifolium hirtum).
  • • After 2 wk, needles of 15N-treated pines and foliage from nearby annuals were similarly enriched, but little 15N had appeared in nontreated (receiver) pine needles, oak leaves or buckbrush foliage. After 4 wk foliar and root samples from pine, oak, buckbrush and annuals were significantly 15N-enriched, regardless of the type of mycorrhizal association.
  • • The rate of transfer during the first and second 2-wk periods was similar, and suggests that 15N could continue to be mobilized over longer times.

Introduction

The role of mycorrhizas in nutrient uptake from soil is well established, but their role in transferring nutrients among plants is less well understood. While many studies demonstrate resource movement between the same or different plant species that share the same mycorrhizal type (arbuscular mycorrhizal, AM or ectomycorrhizal, EM), resource exchange between plants with different mycorrhizal types is not well established. Transfer of carbon, nutrients, and water among plants that are EM or AM has been demonstrated (Brownlee et al., 1983; Finlay & Read, 1986a, 1986b; Newman, 1988; Newman et al., 1992; Simard et al., 2002; He et al., 2003; Leake et al., 2004). Transfers among or between plants of the same mycorrhizal type may occur through mycorrhizal networks that redistribute resources among plants along source–sink gradients (Francis & Read, 1984; Newman, 1988; Bethlenfalvay et al., 1991; Newman et al., 1992; Frey & Schuepp, 1993; Simard et al., 1997; Smith & Read, 1997; Perry, 1998; Lerat et al., 2002; Carey et al., 2004; Leake et al., 2004).

Soil nutrient uptake, particularly N, is usually considered to be unidirectional from soil to roots to shoots (Marschner, 1995; Brady & Weil, 2002). In contrast, C, nutrients and water transfers can be bidirectional between different plant species through mycorrhizal networks, as shown in field, growth-chamber and glasshouse studies using isotopes (Duddridge et al., 1980; Chiariello et al., 1982; Francis & Read, 1984; Newman, 1988; Eissenstat, 1990; Hamel et al., 1992; Newman et al., 1992; Ek et al., 1996; Johansen & Jensen, 1996; Lerat et al., 2002; Simard et al., 2002; He et al., 2003, 2004, 2005; Querejeta et al., 2003; Leake et al., 2004; Simard & Durall, 2004).

Although such transfers are well documented, the quantities transferred are generally low and the ecological implications of sharing or competing for these resources among AM and EM plants are less well known (Newman, 1988; Perry et al., 1998; Robinson & Fitter, 1999; Simard et al., 2002; He et al., 2003; Leake et al., 2004; Simard & Durall, 2004). Pulse-chase experiments lasting a few days may not reflect transfers that would occur over longer time periods. Complex AM and EM networks in blue oak woodlands encourage competition for soil resources (Welker et al., 1991; Cheng & Bledsoe, 2004) and water (Gordon & Rice, 1993). Plant species with AM and EM mycorrhizal types may interact in complicated ways to exchange nutrients and water. Plants sharing the same mycorrhizal species have the potential to share soil resources directly via fungal hyphae. Sharing of nutrients has been reported for C (Francis & Read, 1984; Newman, 1988; Simard et al., 1997; Leake et al., 2004); N (Newman, 1988; Simard et al., 2002; He et al., 2003, 2004, 2005; Simard & Durall, 2004); P (Chiariello et al., 1982; Johansen & Jensen, 1996); and water (Brownlee et al., 1983; Querejeta et al., 2003).

In California oak woodlands, plants are highly dependent on mycorrhizae for growth and survival. A major species in these woodlands is blue oak (Quercus douglasii Hook. & Arn.) that forms predominantly EM associations, but AM symbionts have also been observed (Egerton-Warburton & Allen, 2001; Douhan et al., 2005). Other major species in these woodlands include gray pine (Pinus sabiniana Douglas) that forms EM associations, and buckbrush [Ceanothus cuneatus (Hook.) Nutt., Rhamnaceae] and annual species that form AMs (Rose & Youngberg, 1980; Cheng & Bledsoe, 2002a; Douhan et al., 2005). In this oak woodland community with its mixed mycorrhizal associations, we measured 15N movement from gray pine saplings to nearby saplings of gray pine and blue oak, and to buckbrush shrubs as well as to adjacent annuals: hedgehog dogtail (Cynosurus echinatus L.), hedge parsley (Torilis arvensis (Hudson) Link), and rose clover (Trifolium hirtum All.). Our goal was to determine whether mycorrhizal type (EM or AM) affects N transfer between plant species growing in a Mediterranean oak savanna ecosystem. We designed a field study to examine three questions: (1) Does N move from an EM host to nearby EM plants? (2) Does N move from an EM host to nearby AM plants? and (3) Does N transfer occur preferentially from EM donors to EM receivers?

Materials and Methods

Site description

The University of California Sierra Foothill Research and Extension Center (SFREC) at 39°17′ N, 121°17′ W, elevation 520–560 m, 30 km east of Marysville, CA, USA, served as the experimental site. The area has a Mediterranean climate characterized by hot, dry summers and mild, rainy winters (http://danrrec.ucdavis.edu/sierra_foothill/resources.html). Annual precipitation averages 710 mm, occurring mostly as rainfall from October to May. Mean annual temperature is 16°C; mean minimum temperature 4.5°C (November–March); mean maximum temperature 32°C (July–August). Soils are developed from basic metavolcanic bedrock (greenstone) and are classified as fine, mixed, active thermic Mollic Haploxeralfs. Dominant overstory tree species include gray pine, blue oak, and evergreen interior live oak (Quercus wislizenii A. DC.), with oaks occurring at a density of 100–500 trees ha−1. Common chaparral shrubs include buckbrush and white leaf manzanita (Arctostaphylos viscida C. Parry, Ericaceae). Exotic annual plants include hedgehog dogtail (Cynosurus echinatus, Poaceae), hedge parsley (Torilis arvensis, Apiaceae), and rose clover (Trifolium hirtum, Fabaceae).

Plant selection and experimental design

Using existing gray pine and blue oak saplings (approx. 2.5 m tall) and buckbrush shrubs (approx. 1.5 m tall) we selected combinations of pine–pine, pine–oak, and pine–buckbrush that were growing approx. 1 m apart as our experimental units. Each pair of woody species was replicated four times. In these pairings of woody species we treated gray pine saplings (donors) with 15N, and observed N transfer to the adjacent members of the pair (receivers) and to grasses and forbs growing between the woody species. The grass species was hedgehog dogtail, forb species were hedge parsley and rose clover.

15N labeling of donor pines

In April 2002 we added 1.6 ml K15NO3 (0.334 mole or 0.5% N, 99.30 at %15N, Cambridge Isotope Laboratory) to a 2.0-ml microcentrifuge tube and inserted two fascicles into each tube. Isotope was supplied to five separate branches within the canopy to facilitate 15N distribution within the plant. A total of 8.0 ml 15N solution provided 40 mg 15N to each donor. Tubes were attached vertically to branches and sealed using laboratory labeling tape (Fisher Scientific, Hampton, NH, USA) to reduce evaporation and avoid spillage. By 2 wk after application all the 15N solution had been imbibed and the tubes were removed.

Harvest

Root and foliar samples were collected from all species before treatment. After 2 wk treatment, foliar samples were collected from all species. After 4 wk, foliar and root samples were collected from all species. Annual species were collected midway between the donor and receiver pair in all plots. Foliar samples from woody species were collected from different canopy locations. In order to identify the correct source, roots of woody species were traced back to the plant stem. Then fine roots were harvested from laterals of known woody species. All samples were placed on ice, transported to the laboratory, and stored at 2°C until processed. All roots samples were separated from soil with distilled water; a subset was preserved in 50% ethanol–water solution for determination of mycorrhizal colonization (Brundrett et al., 1996).

Mycorrhizal colonization

Mycorrhizal colonization was expressed as percentage root length colonized. Root colonization by AM fungi was determined by examining Trypan blue-stained fine roots with a compound microscope (Nikon Eclipse E600, Nikon Inc., Melville, NY, USA). The extent of infection was quantified using the gridline intersect method (Brundrett et al., 1996). The gridline intersect method was also used for EM roots. In this case, unstained roots were examined with a dissecting microscope.

Nitrogen analyses

Foliage and root samples were oven-dried (60°C) and milled into fine powder using a Model 6 Wig-L-bug Amalgamator (Crescent Dental MFG Co, Lyons, IL, USA). Powdered samples were analyzed for total N and 15N on a Geo 20/20 Isotope Ratio Mass Spectrometer (PDZ Europa Ltd, Crewe, UK) at the UC Davis Stable Isotope Facility. Ammonium sulfate (Fisher Scientific International Inc., Hampton, NH, USA) was used as a standard for the isotopic analyses. The 15N signature for (NH4)2SO4 was 0.366831 ± 0.000039 (15N, at %) or 1.45 ± 0.10 (δ15N, ) (means ± SE, n = 86). We calculated δ15N (), percentage 15N derived from source (NDFS) as follows (Shearer & Kohl, 1993):

δ15N ()  =  [(Rsample/Rstandard)  −  1]  ×  1000(Eqn 1)

where R is the ratio of 15N/14N (at %) of the sample and standard.

15NDFS  =  (at %15Nsample in week  ×   −  at %15Nsample in week 0)/(at %15Nlabeling source)(Eqn 2)

Root δ15N (y, ) at week 2 (X) was calculated using the following relationship:

y  =  Y0  +  (X  −  X0)(Y4  −  Y0)/(X4  −  X0)(Eqn 3)

where Y4 and Y0 are the root δ15N () at week 4 (X4) and week 0 (X0), respectively.

Statistical analyses

When necessary to achieve a homogeneous distribution of the data, mycorrhizal colonization and N concentration data were transformed by taking the arcsine of the original data. Transformed values were analyzed by anova procedures according to Sokal & Rohlf (1995). Differences in treatment means were compared using Tukey's honestly significant difference with P ≤ 0.05 (Sokal & Rohlf, 1995).

Results

Mycorrhizal colonization

All root samples were colonized to varying extents by either AM or EM mycorrhizal fungi. Colonization of gray pine by EM fungi averaged 64%; oak roots averaged 53% (Table 1). Blue oak mycorrhizas were predominantly EM, but AM infection levels up to 15% were observed. Buckbrush roots had 18% AM infection compared with 25 and 27% for hedgehog dogtail and hedge parsley (Table 1). The highest AM infection occurred in rose clover at 36%. Buckbrush and rose clover roots were also nodulated.

Table 1.  Mycorrhizal colonization (percentage root length colonized) and nitrogen concentration in leaves and roots
Plants15N statusColonization percentageFoliar N percentageRoot N percentage
AMEMWeek 0Week 4Week 0Week 4
  • **

    Donor leaves of gray pine were labeled with 15NO3 for 4 wk. Values are means ± SE (n = 4–12). For percentage colonization, values in columns followed by different letters (x–z) between plant species are significantly different (P ≤ 0.05, Tukey's honestly significant difference). For percentage N, values in columns followed by different letters (a–d) are significantly different (P ≤ 0.05). For percentage N there were no significant changes between weeks 0 and 4, except for hedgehog dogtail and hedge parsley samples ().

Pine + pine (n = 4 pairs)
Gray pineDonor 065 ± 11x1.71 ± 0.13c1.86 ± 0.16c0.45 ± 0.05c0.44 ± 0.01c
Gray pineReceiver 064 ± 9x1.69 ± 0.10c1.82 ± 0.13c0.41 ± 0.03c0.45 ± 0.05c
Pine + oak (n = 4 pairs)
Gray pineDonor 066 ± 7x1.71 ± 0.15c1.83 ± 0.22c0.45 ± 0.05c0.41 ± 0.05c
OakReceiver15 ± 3z53 ± 11y2.26 ± 0.18b2.27 ± 0.25b0.86 ± 0.03b0.84 ± 0.08b
Pine + buckbrush (n = 4 pairs)
Gray pineDonor 063 ± 9x1.70 ± 0.11c1.84 ± 0.20c0.45 ± 0.05c0.42 ± 0.10c
BuckbrushReceiver18 ± 3z 02.24 ± 0.13b2.24 ± 0.03b0.94 ± 0.03b0.89 ± 0.04b
Understory annuals (n = 12, pooled from all plots)
Hedgehog dogtailReceiver25 ± 5y 02.17 ± 0.29b1.54 ± 0.20d, **0.94 ± 0.16b0.94 ± 0.09b
Hedge parsleyReceiver27 ± 4y 02.08 ± 0.32b1.58 ± 0.17d, **0.91 ± 0.07b0.91 ± 0.07b
Rose cloverReceiver36 ± 6z 02.93 ± 0.34a2.66 ± 0.23a1.26 ± 0.03a1.26 ± 0.13a

Foliar and root N concentration

Total foliar and root N concentrations were essentially constant over the experimental period for all samples except foliage of annuals (Table 1). Because N concentrations were not significantly different among the donor–receiver pairs, data were averaged. Nitrogen levels decreased significantly with senescence in hedgehog dogtail and hedge parsley. In all species, foliar N was significantly greater than root N. Leaf N in blue oak and buckbrush averaged approx. 2.25% with little change during the 4-wk experiment. Root N levels were not significantly different between buckbrush (approx. 0.9% N) and oaks (approx. 0.9% N). Gray pine had significantly lower foliar and root N than blue oak and buckbrush, with needles averaging 1.7% and roots 0.5%. Rose clover, a nitrogen-fixer, had an average foliar N concentration of 2.8%, while the roots averaged 1.3% N. Hedgehog dogtail and hedge parsley foliage averaged 2.1% N initially and decreased to 1.6% 4 wk later. Root N levels averaged 0.9% and did not change over the 4-wk experimental period.

15NDFS (percentage 15N-nitrogen derived from the labeled 15N source)

The percentage of 15N-nitrogen derived from source (NDFS) increased over time, but was low in both donor and receiver plants as the amount of 15N was added at tracer levels. After 2 wk, NDFS ranged from 0.001 to 0.01% for all plants (Table 2), and showed that little N had been translocated from donors to woody species. Interestingly, annuals had higher 15N levels at week 2. By week 4, foliar NDFS of annuals was significantly greater than in samples collected before treatment (week 0). All annual species showed accumulation and translocation of 15N. After 4 wk, all 15N-treated gray pines had NDFS values approx. 0.16%, while roots of these donors averaged approx. 0.074%. NDFS of oak and buckbrush leaves ranged from 0.001 to 0.015%, and the leaf samples from annuals had nearly twice that found in woody plants (0.022–0.028%). Rose clover had the highest NDFS. Similarly, NDFS in root samples of annual species was greater than that in pine donor roots, as well as in pine receivers, oak and buckbrush.

Table 2.  Percentage 15N-nitrogen derived from source (15NDFS) in leaves and roots
Plants15N statusFoliar 15NDFS percentageRoot 15NDFS percentage
Week 2Week 4Week 4
  1. Donor leaves of gray pine were labeled with 15NO3 for 4 wk. Values are means ± SE (n = 4–12). Values in rows (x, y) and in columns (plant species a–c; donor saplings vs receiver saplings, α, β) followed by different letters are significantly different (P ≤ 0.05, Tukey's honestly significant difference).

Pine + pine (n = 4 pairs)
PineDonor0.010 ± 0.001y, a, α0.157 ± 0.033x, a, α0.073 ± 0.013c, α
PineReceiver0.001 ± 0.002y, d, β0.011 ± 0.002x, c, β0.039 ± 0.006d, β
Pine + oak (n = 4 pairs)
PineDonor0.010 ± 0.005y, a, α0.156 ± 0.067x, a, α0.074 ± 0.021c, α
Blue oakReceiver0.001 ± 0.001y, d, β0.015 ± 0.003x, c, β0.019 ± 0.005e, β
Pine + buckbrush (n = 4 pairs)
PineDonor0.010 ± 0.001y, ab, α0.164 ± 0.051x, a, α0.076 ± 0.025c, α
BuckbrushReceiver0.001 ± 0.003y, d, β0.014 ± 0.003x, c, β0.030 ± 0.002d, β
Understory annuals (n = 12, pooled from all plots)
Hedgehog dogtailReceiver0.007 ± 0.006y, c0.022 ± 0.007x, b0.078 ± 0.027bc
Hedge parsleyReceiver0.009 ± 0.006y, b0.026 ± 0.009x, b0.086 ± 0.024b
Rose cloverReceiver0.007 ± 0.001y, c0.028 ± 0.006x, b0.100 ± 0.032a

15N atom percentage of labeled plants after week 4

As expected, 15N-treated gray pines foliar samples had the highest 15N at % excess (16 × 10−3%) after 4 wk (Fig. 1), while all receiver species had lower foliar 15N enrichment than the donors (1.1–1.4 × 10−3% in perennials; 2.2–2.8 × 10−3% in annuals). As with NDFS, root samples from annual species had the highest 15N at % excess (7.5–9.9 × 10−3%) compared with donor roots and perennial receiver roots (3–7 × 10−3%). Blue oak root samples had 15N at % excess values that were statistically similar to gray pine donors. Among roots of annual species there were no significant differences in 15N at % excess. In addition, no significant difference in 15N at % excess was found between receiver gray pine and buckbrush roots.

Figure 1.

Distribution of 15N as 15N at % excess in leaves and roots of donor gray pine and receiver gray pine, blue oak, buckbrush, hedgehog dogtail (grass), hedge parsley (forb) and rose clover after 4 wk labeling. Data are means ± SE (n = 4–12). Different letters above bars (x–z for leaves; a–d for roots) designate significant differences (P = 0.05).

Translocation: δ15N following labeling

Before treatment with 15N (time 0), δ15N natural abundance values differed among plant species (Fig. 2). Natural abundance values for foliage samples from non-N-fixing plants ranged from 2.22 to 2.75, while values for N-fixing plants (buckbrush, red clover) were lower, −0.13 to −1.32. Natural abundance δ15N values in gray pine and blue oak roots were the highest among the plants studied at 10.4 and 14.1, respectively. The δ15N values for hedgehog dogtail and hedge parsley roots were 4.24 and 4.39, respectively. The lowest δ15N was found in N2-fixing buckbrush and rose clover at 0.59 and 2.24, respectively.

Figure 2.

δ15N values () of leaves and roots in donor and receiver plants. Data are means ± SE (n = 4–12). Different letters designate significant differences (P = 0.05) between weeks for each plant species (x–z); between plant species for each week (a–e); and between donor and receiver gray pine for each week (α, β). Root δ15N values () at week 2 were calculated according to Eqn 3.

Following 15N applications, pine needles were enriched after 2 wk and were highly enriched by week 4 (46, Fig. 1). Significant enrichment in the foliage of perennial receivers occurred between weeks 2 and 4. By week 4, gray pine and blue oak samples had δ15N values of 6, while buckbrush samples reached 4. Except for rose clover, which had a strongly negative δ15N at week 0, samples from the other annuals had leaf δ15N values of 15–20 2 wk post-15N application. These δ15N values were significantly greater than those found in either donor or receiver perennials. In week 4 samples, foliar δ15N values of annuals remained higher than in perennial receivers and approached values found in 15N-treated pines.

Except for buckbrush, root samples had δ15N values that were similar among the species (approx. 15). By week 4, δ15N values were higher in annual root samples than in woody root samples, and were similar to values found in donor roots. The rate of increase in annual roots was similar to that of donor roots. Because the initial levels were negative, buckbrush shoot and root δ15N values were consistently lower than those of other receiver saplings and annuals. However, the rate of δ15N increase in buckbrush was similar to that of other receiver saplings.

Discussion

The goal of pulse labeling is to follow the path of 15N without significantly altering the N status of the target species. In our study the total N (Table 1) and NDSF (Table 2) data indicate that, in this experiment, this goal was accomplished. Except for the decrease in total N accompanying senescence of annuals, total N was constant throughout the 4-wk experiment in both roots and foliage of all woody species. Foliar N levels in the species sampled were low but not deficient. In Mediterranean climates mycorrhizal associations are known to be important for nutrient and water relations of native species; the plants in our experiment were symbiotic with both EM and AM fungi (Table 1). This difference in mycorrhizal status allowed us to compare N movement between and among plants with these two mycorrhizal types. We found that annuals, oak saplings and buckbrush formed AM fungal associations, and that the oaks and pines formed EM associations. These findings confirm literature reports that oaks are predominantly EM but can form AM, especially when young (Rothwell et al., 1983; Egerton-Warburton & Allen, 2001; Valentine et al., 2002). Cheng & Bledsoe (2002a, 2002b) reported that EM colonization of blue oak roots at a nearby site fluctuated seasonally from a low of 0–10% in the dry season to a high of 80% in the wet season. We conducted our study in the spring (cool and wet) when roots and mycorrhizae are more active. In a preliminary study we determined that 15N natural abundance values in pines, forbs and grasses changed little over a 4-wk period in spring (data not shown). In this study we did not identify the fungal species on roots. However, in other studies at our field site the most abundant EM genera associated with blue oak were Cenococcum, Tuber, Tomentella and Sebacina (Cheng & Bledsoe, 2002a; Douhan et al., 2004; M. Smith, personal communication). The most abundant AM genus was Glomus (Douhan et al., 2005).

15N movement from EM donors to EM and AM receivers

Both EM and AM plants in close proximity to labeled gray pine acquired 15N during the 4-wk experimental period. Annual plants growing near the labeled pines became enriched rapidly and contained significant 15N concentrations 2 wk following 15N application. Roots and needles of gray pine saplings within 1 m of the donor pines also became enriched. The increased 15N content of receiver pine roots and shoots demonstrates that pine transferred N to other pines. However, this N was not transferred preferentially to other pines. 15N at % excess was greater in receiver oak saplings than in receiver pine and buckbrush sapling. The NDFS was lower in oaks, showing that the amount of N transferred comprised a smaller quantity of the blue oak total N pool compared with gray pine and buckbrush. In contrast, δ15N values were greater in gray pine and blue oak receivers than in buckbrush saplings because of different background δ15N values. Annuals had greater 15NDFS, 15N at % excess and δ15N values than perennial receivers, showing that the total N pool of the annuals was more affected by translocation of N from the donor pine. The larger NDFS in annuals may be related to a smaller total N pool in the annuals, such that a small amount of 15N could significantly alter the NDFS.

Nitrogen transfer between plants can take place by direct or indirect routes (McNeill & Wood, 1990). Direct N transfer occurs through mycorrhizal connections among plants (Newman, 1988; Frey & Schuepp, 1993; Simard et al., 2002; He et al., 2003, 2004, 2005; Leake et al., 2004; Simard & Durall, 2004). However, amounts of N transferred were low and may not be ecologically relevant (Newman, 1988; Simard et al., 2002; He et al., 2003; Leake et al., 2004; Simard & Durall, 2004). Direct links between EM and AM are unlikely: AM and EM fungi are taxonomically distinct and cannot fuse their hyphae and exchange/transfer materials.

Nitrogen transfer from EM pines to AM receivers shows that direct fungal connections are not necessary for N transfer among plants in this oak woodland community. Other indirect mechanisms must be operative. Nitrogen transfer from donor roots could be explained by at least six processes: (1) turnover of mycorrhizal hyphae (Staddon et al., 2003); (2) N movement from mycorrhizal mycelia to soil (Johnson et al., 2002); (3) N foraging in the rhizosphere by receiver hyphae (Read & Perez-Moreno, 2003; Cairney, 2005); (4) root exudation (Jones et al., 2004); (5) faunal grazing (Klironomos & Hart, 2001; Perez-Moreno & Read, 2001; Johnson et al., 2005); and (6) recapture of N-containing materials from rhizodeposition (Stern, 1993; Dubach & Russelle, 1994; Chalk, 1996; Paynel et al., 2001; Walker et al., 2003) which, in turn, is regulated by interactions between plants and mycorrhizas (Jones et al., 2004). Our data document a rapid transfer among AM and EM plants, but do not allow us to determine which processes are responsible for the 15N transfers we observed.

The transfer of foliar 15N from the donor to its own roots and to roots of receivers and annuals occurred rapidly (Fig. 2). The steep increase in δ15N and the linear response between the second and the fourth week suggests that N transfer from donor roots continued at similar rates during the 4-wk sampling period. This ongoing N transfer may be a significant pathway for N exchange among plants. The roots and leaves of the annual plants had greater 15N derived from source (NDFS) and were more enriched (15N at % excess and δ15N values) than perennial receivers, irrespective of the mycorrhizal type of the receivers. Apparently annuals, with their extensive fibrous root systems, especially in upper soil layers, were a strong sink for N (Jackson et al., 1988; Cheng & Bledsoe, 2002a). Root systems of perennial species (blue oak, gray pine and buckbrush) are unevenly distributed and are deeper with less branching than annuals (Millikin & Bledsoe, 1999). The difference in root distribution may affect competition for soil resources such as water and N (Tilman, 1989; Welker et al., 1991; Gordon & Rice, 1993; Momen et al., 1994; Koukoura & Menke, 1995; Cheng & Bledsoe, 2004).

We measured 15N transfer against a concentration gradient – from pines with lower percentage N to grasses and forbs with higher percentage N. However, we did not measure two-way or net transfer, only one-way transfer. Recently, two-way or bidirectional N transfer has been documented from barley to pea (Johansen & Jensen, 1996) and from Eucalyptus to Casuarina (He et al., 2004, 2005). If two-way N transfer is a universal phenomenon, N transfer against a concentration gradient, for example from pine to oak, may occur.

Foliar labeling and N movement

Under controlled conditions, leaf 15N feeding is an efficient technique for labeling above- and below-ground plant biomass (Schmidt & Scrimgeour, 2001; Yasmin et al., 2006). Although labeling foliage using 15N to observe N movement below ground has not been reported extensively for field conditions (Horwath et al., 1992; Schmidt & Scrimgeour, 2001), this technique worked well at our oak woodland site. Movement of N in trees and other perennials has been thought of as a primarily unidirectional flow from roots to above-ground components during leaf expansion, to structural components during senescence, and, in early spring to fine roots from large roots as roots elongate (Krammer & Kozlowski, 1979). However, our results show that N movement can be a two-way dynamic process, and can occur from needles to roots in donors and from roots to needles in receivers, even after leaf expansion and early root growth. Our study examined the net transfer of 15N among AM and EM plants. We recognize that N transfer/turnover may have occurred several times during the 4-wk period. If that is the case, then we believe that a small pool of actively cycling N is responsible for N transfers among plants.

In summary, our results suggest that mycorrhizas play an important role in transferring N between plants, including release and recapture of N from the rhizosphere. Nitrogen transfer among plants in California oak woodlands expands our view of below-ground complexity associated with N cycling. Nitrogen translocation among perennial and annual plant species is more complex than expected. Nitrogen transfers occurred beyond the boundaries of a common mycorrhizal linkage among plants of the same species. Therefore indirect pathways, in addition to direct transfers, must be invoked to describe the complex N movement observed in the oak woodland ecosystem.

Acknowledgements

For valuable comments on the manuscript we thank four anonymous reviewers and Dr Francis Martin, the New Phytologist Section Editor. We thank the University of California Sierra Foothill Research and Extension Center for field assistance. This research was supported by NSF Biocomplexity grant DEB-9981711 to Drs C.S. Bledsoe, R.J. Zasoski, W.R. Horwath and D.M. Rizzo. Dr He is grateful to Yunnan Normal University and Education Department of Yunnan Province, China for permission to study overseas.

Ancillary