Vesicular–arbuscular mycorrhizal infection of Quercus rubra seedlings

Authors


Author for correspondence: Roger T. Koide Tel: +1 814 863 0710 Fax: +1 814 863 6139 Email:rxk13@psu.edu

Summary

  •  The influence of vesicular–arbuscular mycorrhizal (VAM) host plants is reported on VAM infection of Quercus rubra seedlings, a predominately ectomycorrhizal species.
  •  In a field experiment, Q. rubra seedlings were planted near Quercus montana and Acer rubrum (a VAM host) stump sprouts and near dead Quercus spp. stumps. In a subsequent glasshouse study, seedlings of Q. rubra were grown with or without VAM inoculum (±V) and with or without Sorghum bicolor (±S), a VAM host.
  •  In the field experiment, seedlings planted near A. rubrum had greater VAM infection than other seedlings; ectomycorrhizal infection was greatest on seedlings near Q. montana. In the glasshouse experiment, VAM infection was increased in both extent (proportion of root length infected in roots of infected seedlings) and frequency (number of seedlings with infection) in the presence of S. bicolor. Seedling mass and N concentration were reduced in the +V treatment; however, these variables were not correlated with extent of VAM infection.
  •  VAM infection of Q. rubra is greatly enhanced in the presence of VAM hosts in both field and glasshouse conditions. No evidence for positive effects of VAM infection on Q. rubra seedlings was found.

Introduction

Quercus rubra L. and other Quercus spp. are usually considered to be ectomycorrhizal. Nonetheless, vesicular–arbuscular mycorrhizal (VAM) infection has been found in both mature and seedling Q. rubra (Henry, 1933; Grand, 1969; Watson et al., 1990), Q. falcata Michaux (Grand, 1969), Q. palustris Münchhausen (Watson et al., 1990), and seedling Q. imbricaria Michaux (Rothwell et al., 1983) suggesting that VAM infection is common within Quercus section Lobatae (classification follows Nixon, 1993). Watson et al. (1990) suggest that VAM infection does not occur in Quercus section Quercus, but roots of only one species within section Quercus, Q. alba L., have been examined for VAM infection (Watson et al., 1990). Kim & Lee (1990) inoculated cuttings of Q. acutissima (in section Quercus) with Glomus sp. spores, but did not check roots for infection and found no effect of inoculation on growth.

No previous studies that we are aware of have examined whether VAM infection influences the growth of Quercus spp. Nonetheless, Watson et al. (1990) suggest that the ability to become infected with both ectomycorrhizal and VAM fungi may permit Q. palustris and Q. rubra to thrive in a diverse range of site conditions. VAM infection of Q. rubra has been found more abundantly in lowland than upland sites (Watson et al., 1990). We are aware of no other studies examining the ecology of VAM infection of Q. rubra.

Lack of fidelity to a single type of mycorrhiza is by no means limited to Quercus. A number of plants, including members of the Salicaceae, Betulaceae, and Myrtaceae, normally become infected with both ectomycorrhizal and VAM fungi (Smith & Read, 1997). VAM infection in the normally ectomycorrhizal Pinaceae has also been described (Vardavakis, 1992; Cázares and Trappe, 1993; Cázares and Smith, 1996; Smith et al., 1998). Levels of VAM and ectomycorrhizal infection may be negatively correlated in plants forming both mycorrhiza types (Chilvers et al., 1987; Lodge, 1989; Lodge & Wentworth, 1990; Watson et al., 1990) with ectomycorrhizal infection apparently capable of displacing VAM infection (Lapeyrie & Chilvers, 1985; Chilvers et al., 1987; Lodge & Wentworth, 1990; Chen et al., 2000). Interactions between ectomycorrhizal and VAM infection appear to vary, however, with species of host plant and habitat (Moyersoen & Fitter, 1999; Smith & Read, 1997). Ectomycorrhizal infection of normally ericoid mycorrhizal Vaccinium spp. has also been described (Stevens et al., 1997).

VAM infection of predominantly nonVAM host plants may be increased by the presence of predominantly VAM hosts. Smith et al. (1998) found that VAM infection in Pinaceae is most successful in the presence of a predominately VAM host. Normally nonmycotrophic members of the Cruciferae and Chenopodiaceae have been shown to become slightly infected by VAM in the presence of a VAM host (Ocampo et al., 1980), although no arbuscules were observed. It may also be possible for normally host-specific ectomycorrhizal fungi to infect other hosts in the presence of the preferred host plant (Massicotte et al., 1994). In the first portion of the present study we examine naturally occurring VAM infection of Q. rubra seedlings growing in the field. As reported below, we observed high levels of VAM infection, particularly in Q. rubra seedlings planted near established Acer rubrum L., a VAM host plant. This led to the hypothesis that VAM infection of Q. rubra was enhanced in the presence of VAM host plants. We tested this hypothesis by growing Q. rubra seedlings with and without Sorghum bicolor (L.) Moench, a VAM host plant, in the glasshouse. A secondary goal of the glasshouse experiment was to examine the influence of VAM infection on Q. rubra seedling nutrient uptake and growth.

Materials and Methods

Field experiment

The field site was in the Rothrock State Forest (PA, USA), in the Valley and Ridge physiographic province on a north to north-east facing lower slope with an elevation of approximately 420 m (77°52′ W Long, 40°42′ N Lat). Annual precipitation in 1998 and 1999 was 97.7 cm and 98.5 cm, respectively. There was a moderate to severe drought during most of the growing season in 1999. The site had been dominated by a mixture of Quercus spp., including Q. montana Wildenow, Q. alba, Q. coccinea Münchhausen, and Q. rubra (in decreasing order of basal area, Rothrock State Forest timber sale records, sale #5–74 BC3). This stand was partially shelterwood harvested in 1974. In 1990, the site burned in a fire that killed most vegetation but did not burn the lower layers of duff and roots. All remaining trees greater than 5 cm in diameter were removed in a salvage logging operation in 1990. Vegetation on the site is dominated by scattered A. rubrum and Quercus spp. stump sprouts emerging from a dense shrub layer consisting of (in decreasing order of percentage cover) Dennstaedtia punctilobula (Michx.) T. Moore, Vaccinium spp., Comptonia perigrina (L.) J. M. Coult., Kalmia latifolia L., Galussacia bacatta (Wang) K. Koch, grasses, forbs, sedges, Gaultheria procumbens L., Lycopodium sp., and other ferns.

In the yr 1 planting, two treatments were established: Near-Quercus and Near-Acer. For the Near-Quercus treatment, we selected stump sprouts of Q. montana and cleared a 2 × 2 m plot on the south side of the stump at canopy drip line. Stump sprouts were defined as a stump of at least 20 cm diameter with at least one living sprout of at least 2 m in height. Near-Quercus plots were an average of 5.4 m from the nearest A. rubrum stump sprout. Near-Acer plots were similarly located adjacent to sprouted stumps of A. rubrum, but had the additional criterion of being at least 9 m from the nearest ectomycorrhizal tree, stump sprout, or seedling. Plots were sprayed with glyphosate herbicide 2 wk before planting and hand-cleared approximately 2 d before planting. Eleven replicates of each treatment were established, of which 10 from each treatment had surviving seedlings by the end of the first growing season.

Acorns were collected from a single Q. rubra in State College, PA, USA in the autumn of 1997 and stored at −1°C over winter. Acorns were surface sterilized in 10% household bleach for 5 min, planted in Pro-Mix BX potting soil on 27 May 1998 (Premier Brands Inc., Riviere-du-Loup, Quebec, Canada), and kept continually moist until radicle emergence. On 2 and 3 June 1998, germinated acorns with < 2 cm of emerged radicle were planted, four acorns per plot, in a square arrangement at approximately 30 cm spacing and 2 cm depth. Immediately after planting, a 60 × 60 × 30 cm tall, 1.25 cm mesh hardware cloth cage was placed over seedlings to reduce herbivore damage. Following the full expansion of the first pair of leaves on each seedling, the cotyledons were removed to maximize seedling growth responsiveness to environmental conditions and to reduce the risk of seedling mortality during post-germination seed predation. Seedlings were watered on three occasions during early establishment with tap water.

One seedling from each plot was destructively harvested 3 October 1998 to measure mycorrhizal infection. Seedlings were removed from the soil with care to minimize damage to root systems; the extremely stony soil made extraction of intact soil cores impossible. Because of damage to root systems during extraction from soil, all data on mycorrhizal infection are on a percentage basis, rather than absolute values, and root mass is not presented. Seedlings were harvested on a cool, overcast day and placed in moist toweling. After returning to the lab, seedlings were soaked in water for 24 h to loosen remaining soil, then gently washed under running water. Intact root systems were stored under refrigeration for up to 5 d before measuring mycorrhizal infection. For a separate experiment (unpublished) an additional seedling was harvested from four plots from each treatment, but at least one seedling was left in every plot.

Seedlings remaining from the yr 1 plantings were left in place for a second year. Plots were weeded in the spring of 1999, and cages replaced with 30 cm diameter, 120 cm tall, open-topped tubes of 2.5 cm mesh chicken-wire to reduce deer browse on larger seedlings. Seedlings were harvested on 4 October 1999.

The yr 2 planting was similar to yr 1, with one added treatment. Acorns were also planted in proximity to Quercus spp. stumps that had not resprouted, hereafter referred to as the ‘Near-Dead-Quercus’treatment. Dead Quercus stumps were selected with similar criteria used for the Near-Acer treatment: a minimum of 20 cm diameter and at least 9 m from the nearest ectomycorrhizal tree, stump sprout or seedling. All stumps were either Q. montana or Q. alba. Both Near-Quercus and Near-Dead-Quercus seedlings in the yr 2 planting averaged 4.7 m from the nearest A. rubrum stump sprout. Acorns were collected from a single tree near the study site in 1998, stored as in yr 1, and placed in peat moss to germinate on 21 May 1999; planting took place on 27 and 28 May 1999. Cages were somewhat smaller in footprint; but taller than in yr 1 (30 × 30 × 60 cm tall) with three acorns planted in an equilateral triangle at about 15 cm distance. Seedlings were watered with tap water on two occasions during early establishment. All surviving seedlings were harvested 3–4 October 1999. Harvesting procedures were as in yr 1.

A subsample of fine roots from each seedling was examined for vesicular–arbuscular mycorrhizal (VAM) infection. Roots were cleared for 30 min in 10% KOH at 121°C, rinsed in water, acidified for 5 min in 5% HCl, and stained for at least 24 h in 45% H2O, 50% glycerol, 5% acetic acid, 0.01% trypan blue. Stained roots were examined as in Koide & Mooney (1987), with a root recorded as VAM if both vesicles and hyphae were present within a section of root. Ectomycorrhizal infection was also quantified in stained samples, with all samples where a mantle was present recorded as ectomycorrhizal.

Stem and leaf tissues were digested and analysed for N and P. Dried (70°C) and ground tissue (c. 0.1 g samples) was digested in 5.0 ml H2SO4 and 5.0 ml 30% H2O2 for 60 min at 400°C. Samples that retained colour after 30 min of digestion received an additional 3.0 ml H2O2 and a total of 90 min at 400°C. After digestion, N was quantified with the Nessler method (Jensen, 1962), and P was quantified with the molybdo-phosphate method (Wantanabe & Olsen, 1965).

Glasshouse experiment

The glasshouse experiment was a 2 × 2 factorial design, with VAM inoculum (±V) and sorghum (Sorghum bicolor, ±S). Each treatment combination was replicated in 10 pots, each containing one tree seedling. Pots (3.8 l ‘Tall Ones’, Steuwe and Sons, Corvallis, OR, USA) were filled with a 1 : 3 mix of soil and sand. Soil was collected from an agricultural field, dried, autoclaved for 60 min at 121°C, and permitted to rest for several weeks before planting to reduce possible phytotoxic effects of soil heating (Rovira & Bowen, 1966). The VAM pots (+V) received 200 ml of Glomus intraradices pot culture mixed into soil in the upper 1/2 of each pot. Pot culture contained soil, sand, root fragments, and VAM spores (c. 225 spores ml−1) in which mycorrhizal sorghum had been previously grown. NonVAM pots (−V) received 100 ml of water that had been mixed with pot culture soil and passed through a 63-µm soil sieve (Koide & Li, 1989). All pots were established on 8 January 2000.

Nine sorghum seeds were planted in each +S pot on 8 January 2000. Pots in which fewer than four sorghum germinated 14 d after initial planting were planted with an additional 4–6 sorghum seeds. Pots with more than four sorghum were thinned to four plants per pot on 3 February 2000, so that all pots had four sorghum plants. Nonsorghum treatment pots (−S) were established simultaneously and treated identically to +S pots, except for the planting of sorghum. Pots were watered once per week with 200 ml of 300 p.p.m. N as NH4NO3 and twice per week with water.

Acorns were obtained from the same lot as those used in yr 2 of the field experiment. Acorns were surface sterilized as above and planted into a coarse vermiculite seedbed on 3 February 2000. Germinated acorns with radicles of 1–15 mm in length were transplanted into pots on 11 February 2000.

No Q. rubra mortality occurred during the study. Seedlings were harvested on 14 April 2000. All seedlings had completed two flushes of growth and had set bud at time of harvest. Roots and soil were removed from pots, soil was gently shaken free from roots, and roots were washed by immersion in water. Post-harvest measurements of growth and nutrient status were made as in the field experiment. VAM infection was quantified as in the field experiments but with an additional 30 min clearing period in hot 10% KOH following autoclaving to improve clearing.

For a bioassay of mycorrhizal inoculation potential, smaller pots (600 mL) of soil from the +V−S treatment were established concurrent with the main experiment. These pots were treated exactly as the main experiment, except that only 100 mlof nutrient solution was given weekly in order to compensate for smaller pot size. Four maize seeds (Zea mays cv. Bodacious) were planted in each pot on 11 February 2000. Maize emerged by 16 February 2000 and was harvested on 7 March 2000 for quantification of VAM infection. Maize roots were stained as above, except with 15 min clearing in 10% KOH. Sorghum roots removed from pots during thinning were also stained to detect VAM infection.

Sorghum shoots were clipped just above soil in order to reduce competitive effects of sorghum on seedling growth on 25 February 2000. Shoots were then re-clipped every 7–14 days. After 13 March 2000 sorghum failed to regrow after clipping in some pots.

Statistics

The field experiment was analysed using either t-tests (yr 1 planting) or analyses of variance (ANOVAs) with protected least significant differences (PLSD) for means separation (yr 2 planting), provided that the assumptions of the tests were met. Data were checked for normality and homoscedasticity using Shapiro-Wilks and Levenne’s tests, respectively. Where needed, Box-Cox transformations were used. Data that could not be successfully transformed were analysed using nonparametric statistics (Mann–Whitney U for two sample tests and Kruskal–Wallis with Nemenyi’s test for means separation of three sample tests; Zar, 1999). Data and standard errors are presented untransformed in graphs and tables for ease of interpretation.

Results from the glasshouse study were analysed as a 2 × 2 factorial fixed model analysis of variance (ANOVA). All data were examined for normality and homogeneity of variances and transformed with Box-Cox transformations as needed. In the glasshouse study, the proportion of root length infected had highly heterogeneous variances, as all −V+S seedlings and nine of 10 −V−S seedlings had 0% infection. The frequency of infection (number of seedlings with infection) was therefore analysed with a χ2 statistic. Proportion of root length infected was compared between infected seedlings in the +V−S and +V+S treatments using a t-test for unequal variances.

Statistics were performed in SAS version 7, except for Box-Cox transformations which were determined with Minitab version 12.

Results

Field experiment

The extent of VAM infection was consistently higher in Near-Acer seedlings than either Near-Quercus or Near-Dead-Quercus seedlings (Fig. 1). The extent of VAM infection of Near-Quercus and Near-Dead-Quercus seedlings was not significantly different. The extent of ectomycorrhizal infection, in contrast, was significantly greater in Near-Quercus seedlings than in either Near-Acer or Near-Dead-Quercus seedlings.

Figure 1.

The extent of ectomycorrhizal (ECM) and vesicular–arbuscular mycorrhizal (VAM) infection for yr 1 planting at 16 (a) and 64 wks (b), and for yr 2 planting at 16 wks (c). For all graphs, Near-Acer, black bars; Near-Quercus, white bars; and Near-Dead-Quercus, crosshatched bars (yr 2 only). Standard errors are shown with bars. Bars with different letters within each experiment and age are significantly different (P < 0.05).

Where present, VAM infection consisted of vesicles and hyphae. Arbuscules were never observed; coils were observed very rarely. VAM and ectomycorrhizal fungi were never observed simultaneously infecting the same portion of root. Beaded root tips (Grand, 1969; Mineo and Majumdar, 1996) were only rarely observed, and were not infected by either VAM or ectomycorrhizal fungi.

Near-Quercus seedlings had greater concentrations and content of both N and P than Near-Acer seedlings in both plantings. In the yr 2 planting, Near-Dead-Quercus seedlings had N and P concentrations similar to Near-Acer seedlings, but were intermediate between the other two treatments in N and P content (Fig. 2). There were no significant differences among treatments in seedling dry mass at 16 wk after planting in either planting, but by 64 wk after planting Near-Quercus seedling mass was significantly greater than Near-Acer seedling mass (Fig. 2). Basal diameter was marginally larger in Near-Quercus seedlings than Near-Acer seedlings at 16 wk in the yr 1 planting, and significantly larger at 64 wk (Fig. 2). There were no significant differences in basal diameter in the yr 2 planting.

Figure 2.

Shoot N and P content and concentration, dry mass, and basal diameter of Near-Acer (black bars), Near-Quercus (white bars), and Near-Dead-Quercus (crosshatched bars) seedlings for field experiment by year of planting and age of seedlings. Bars with different letters within each experiment and age are significantly different (P < 0.05).

Glasshouse experiment

Infection by VAM fungi occurred in all +V+S seedlings, in four of 10 +V−S seedlings, in one of 10 −V−S seedlings, and in no −V+S seedlings (Table 1, χ2 = 25.9, P < 0.0001). The pairwise difference in presence of infection between +V+S (all seedlings infected) and +V−S (4 of 10 seedlings infected) was also significant (χ2 = 8.57, P = 0.0034). Proportion of root length infected was higher in the +V+S seedlings than in +V−S seedlings, even where only infected seedlings are considered (P = 0.0036, Fig. 3). As in the field experiment, arbuscules were never observed in seedling roots.

Table 1.  Presence of vesicular–arbuscular mycorrhizal (VAM) infection in glasshouse experiment seedlings. Overall 3 df chi-square P < 0.0001. Rows with different letters are significantly different in 1 df Chi-square tests at P < 0.05
TreatmentNumber of seedlings
InfectedNot infected
   −V−S 1 9a,b
   −V+S 010a
+V−S 4 6b
+V+S10 0c
Figure 3.

Extent of vesicular–arbuscular mycorrhizal (VAM) infection of seedlings in glasshouse experiment (only seedlings with infection present). The number of seedlings with infection present is noted above bars. Error bars are ±1 standard error. The difference between +V+S and +V−S infection extent is significant at P = 0.0036. Since only one −V−S seedling was infected, extent of infection can not be statistically compared with other treatments.

Maize bioassay roots were extensively infected in soil from the +V−S treatment. Excessive clearing time in KOH prevented exact quantification. Sorghum roots removed during thinning were extensively infected in the +V treatment and uninfected in the −V treatment.

Adding VAM inoculum reduced shoot N concentration and N content, and marginally reduced P content (Table 2). Seedlings grown with sorghum had reduced shoot contents of N and P, but no change in concentration of either nutrient. No interaction between treatments in nutrient concentration or content occurred.

Table 2.  Shoot dry mass, basal diameter, and nutrient content for glasshouse experiment seedlings1
 N concentration (%)P concentration (%)N content (mg)P content (mg)Shoot dry mass (g)Basal diam (mm)
  • 1

    Denotes means and standard errors.

VAM
 −V2.01 (0.05)0.032 (0.002)65.3 (3.3)1.07 (0.08)   3.29 (0.18)3.28 (0.06)
+V1.82 (0.05)0.031 (0.003)52.3 (3.6)0.90 (0.12)   2.87 (0.18)3.06 (0.10)
P value0.01700.5504 0.00280.0832   0.04570.0531
Sorghum
  −S1.89 (0.06)0.033 (0.002)67.6 (3.1)1.19 (0.12)   3.58 (0.15)3.31 (0.08)
+S1.94 (0.05)0.031 (0.002)50.0 (3.3)0.77 (0.06)  2.58 (0.16)3.03 (0.08)
P value0.55920.3065 0.00010.0005< 0.00010.0138
Interaction
 −V−S1.99 (0.09)0.033 (0.001)73.4 (3.6)1.25 (0.10)   3.74 (0.21)3.34 (0.09)
 −V+S2.02 (0.06)0.032 (0.004)57.2 (4.4)0.89 (0.11)   2.85 (0.23)3.21 (0.09)
+V−S1.80 (0.06)0.032 (0.005)61.9 (4.4)1.14 (0.22)   3.43 (0.20)3.28 (0.14)
+V+S1.85 (0.08)0.030 (0.003)42.7 (3.8)0.66 (0.04)   2.30 (0.18)2.84 (0.11)
P value0.89860.5656 0.71420.7974   0.55720.1667

Both the addition of VAM inoculum and planting of sorghum significantly reduced shoot dry mass of seedlings (Table 2). Basal diameter was significantly reduced in the +S treatment, and marginally reduced in the +V treatment (Table 2). There was no interaction between treatments for either growth variable.

Although the introduction of VAM inoculum decreased shoot dry mass and N concentration and content of seedlings, there was no correlation between proportion of root length infected and these variables within the +V+S treatment (P > 0.1 for each).

Discussion

VAM infection

Despite normally being considered ectomycorrhizal, we observed high levels of VAM infection of Q. rubra grown in the field. This is consistent with the VAM infection levels of up to 72.6% of root tips in Q. rubra reported by Watson et al. (1990).

VAM infection of Near-Quercus and Near-Dead-Quercus seedlings was similar, suggesting that VAM infection is not inhibited by the presence of Q. montana stump sprouts. This also suggests that the lower levels of VAM infection observed in Near-Quercus seedlings were not due to antagonistic influences of ectomycorrhizal fungi on VAM fungi, as might be suggested by Lodge & Wentworth (1990) or Chen et al. (2000). In part, the lack of evidence for antagonistic interactions between VAM and ectomycorrhizal infection may reflect the relatively young age of seedlings, as ectomycorrhizal infection might cause reductions in VAM over time (Chilvers et al., 1987).

In the glasshouse experiment, VAM infection of Q. rubra was more frequent and present in a higher proportion of root length in +V+S pots than in +V−S pots. Both −S pots and +S pots were established simultaneously, 34 d before planting of Q. rubra. One possible explanation for low VAM infection in +V−S pots is that VAM infection potential dropped during the 34-d period without a VAM host present. Maize bioassay seedlings, however, were extensively colonized by VAM when planted in +V−S soil, suggesting that VAM infection potential remained high. Biomass of VAM fungi might, however, have been greater in the +V+S pots than the +V−S pots; this is one mechanism through which sorghum may increase VAM infection of seedlings.

The data suggest therefore that VAM infection of Q. rubra is enhanced by the presence of predominately VAM host plants. This pattern was suggested by field results, and we were able to replicate it under controlled conditions in the glasshouse. These results are consistent with other studies that have found that the presence of VAM hosts can increase the VAM infection of normally ectomycorrhizal Pinaceae (Smith et al., 1998), and nonmycotrophic members of the Cruciferae and Chenopodiaceae (Ocampo et al., 1980), as well as other VAM plants (Eissenstat & Newman, 1990). Whether infection was directly from hyphae supported by VAM hosts, or from spores stimulated to germinate by the presence of VAM host root exudates is unknown (Smith et al., 1998).

Growth and nutrient status

In the field experiment Near-Quercus seedlings had higher nutrient concentration and contents than the other treatments, most likely as a result of greater ectomycorrhizal infection (Dickie, 2000). Had VAM infection been beneficial to Q. rubra seedlings, Near-Acer seedlings would be expected to have greater nutrient uptake and growth than Near-Dead-Quercus seedlings (which were as low in ectomycorrhizal infection as Near-Acer seedlings, and as low in VAM infection as Near-Quercus seedlings). This was not the case, suggesting that VAM infection of seedlings in the field had little or no benefit to seedling nutrient acquisition or growth. This result is in contrast with the finding of Smith et al. (1998) that VAM infection may increase tissue P content of Pseudotsuga menziesii.

The nutrient conditions of the glasshouse study were designed to maximize any possible growth increases due to VAM infection by keeping available P levels low. Despite this, there was no evidence of positive effects of VAM infection on Q. rubra. Indeed, adding VAM inoculum actually decreased shoot dry mass, and N concentration and content. There was no interaction between treatments for any of these variables, suggesting this was not due to increased competition with sorghum in the presence of VAM fungi. There also was no correlation between the proportion of root length infected and reductions in any of these variables, suggesting that detrimental effects of adding VAM inoculum may not have been due to VAM infection per se. Pot-culture inoculation is known to potentially cause growth reductions unrelated to mycorrhizal infection of seedlings (Koide & Li, 1989). In the glasshouse study, spore washings were added to −V pots, however, this technique may be an imperfect control for the presence of other microorganisms. Additionally, +S pots received 200 ml of VAM pot culture grown in a 1 : 1 ratio of sand and soil. Other pots had entirely 3 : 1 sand : soil mixture, resulting in a 3.6% difference in percentage of soil between treatments. This represents a small, but not properly controlled, difference in growing conditions.

Seedlings were grown for 71 d in the glasshouse study, which may limit the generality of nutrient uptake and growth results. Substantial acorn nutrient reserves of Q. rubra may limit responsiveness of seedlings to environmental conditions (Hanson, 1986; Newton & Pigott, 1991). Further, a lag-time in growth response to mycorrhizal infection can occur even in typically VAM hosts (Smith & Read, 1997). The results of this study suggest that VAM infection has little benefit for nutrient uptake and growth of Q. rubra seedlings in the first season of growth. Further work is needed to determine if VAM infection has different influences on older trees.

The reduction in growth and nutrient status of Q. rubra seedlings in +S treatment was probably the result of competition with sorghum for P, as N and water were abundantly supplied.

Absence of arbuscules

In our observations of Q. rubra seedling roots in both these and other studies (approximately 300 seedlings in total) we never observed arbuscules, although we have observed abundant VAM hyphae and vesicles and, rarely, intracellular coils. It is possible that arbuscular structures are present, but are destroyed in the rather protracted KOH clearing procedure necessary for oak roots. In prior descriptions of VAM infection of Quercus, arbuscules were observed and photographed by Grand (1969) in Q. rubra and Q. falcata.Watson et al. (1990) report the presence but not frequency of arbuscules in both Q. rubra and Q. palustris.

It is also possible that arbuscules were never present. Rothwell et al. (1983) found vesicles and hyphae of Glomus sp. infection of Q. imbricaria , but did not observe arbuscules or hyphal coils, and only rarely observed intracellular vesicles. Similarly, Ocampo et al. (1980) observed VAM infection of normally nonmycotrophic plants, but found only hyphae and vesicles. The apparent absence of arbuscules might explain the lack of seedling nutrient or growth responses to VAM infection.

Conclusions

The most important finding of this study is that both frequency and extent of VAM infection of Q. rubra seedlings are greatly enhanced by the presence of predominately VAM hosts under both field and glasshouse conditions. This study was not designed primarily to test the influence of VAM infection on Q. rubra nutrition or growth. Nevertheless, the data suggest that VAM infection does not increase either nutrient uptake or growth of Q. rubra seedlings early in development. Until shown otherwise it probably is not justified to assume that VAM infection has any beneficial effects on Q. rubra.

Given the extent of VAM infection observed in this study and others (Watson et al., 1990), more research on the effects of VAM infection on Quercus spp. nutrition and growth is warranted. Since the presence of a predominately VAM host appears to be necessary for extensive VAM infection of Q. rubra, future studies of the influence of VAM infection on Q. rubra will need to develop techniques to avoid confounding competitive and mycorrhizal effects. The difficulty is that the addition of VAM inoculum may make VAM host plants more competitive with Q. rubra seedlings – making the cause of any growth reductions equivocal.

Cázares & Trappe (1993) suggest that studies of the mycorrhizal ecology of Pinaceae are incomplete if VAM infection is not quantified. The same may be true for studies of mycorrhizas of Quercus.

Acknowledgements

P. Kelly, R. Rennig, and R. Weiss gave vital field assistance. D. Eissenstat, L. McCormick, K. Steiner, and two anonymous reviewers provided valuable criticisms and insight. IAD was funded by the Life Sciences Consortium and School of Forest Resources of the Pennsylvania State University, USA. A grant from the College of Agricultural Sciences of the Pennsylvania State University facilitated participation by ACF. Additional support was given by the A. W. Mellon Foundation. The field experiment took place on the Rothrock State Forest, PA, USA as study #SFRA 9708.

Ancillary

Advertisement