Ecological intensification seeks to achieve crop yield increases through intensifying complementary or facilitative interactions between plant species or varieties. Different species of arbuscular mycorrhizal fungi (AMF) exhibit niche differentiation and show selectivity towards certain plants, which can further enhance complementarity. It is not clear whether in the presence of one AMF species, where mycelial networks connect crop species, opportunities for complementarity effects may be reduced.

We grew monocultures and mixtures of maize varieties in a greenhouse with one species of AMF, Funneliformis mosseae, during two consecutive years to investigate whether under such conditions the mycorrhizal symbiosis would affect complementarity and overyielding compared to non-mycorrhizal plants.

Variety mixtures showed increased phosphatase activity and mycorrhizal colonization, enhanced P-uptake and overyielding when plants were mycorrhizal. There was no overyielding when plants were non-mycorrhizal. The increase in relative yield total was due to complementarity effects.

Our data show that the magnitude of mycorrhiza-induced overyielding in maize variety mixtures can be similar to that reported for plant species mixtures. Our study implies that appropriate agricultural management that enhances the mycorrhizal fungal contribution to ecosystem services may result in overyielding in yield or P uptake through mixing varieties of one crop species.

Collection of dataset

A calcareous loamy soil was collected from field plots at the Changping Long-Term Fertilizer Station of China Agricultural University in Beijing. The soil contained 17.8 g organic matter kg^-1 soil, 2.9 mg Olsen-P kg^-1 soil, 87.2 mg N kg^-1 soil, 156 mg ammonium acetate-exchangeable K kg^-1 soil, and had a pH (in 0.01 M CaCl₂) of 7.8. The soil was sieved (2 mm) and sterilized by γ -radiation with ⁶⁰Co at 10 kGy.

The following mineral nutrients at the indicated rates (kg^-1 soil) were added uniformly: 200 mg N (as KNO₃), 50 mg Mg (as MgSO₄), 5 mg Zn (as ZnSO₄·7H₂O), and 2 mg Cu (as CuSO₄). To achieve the same soil K level among all treatments, K₂SO₄ was supplied at 25, 25 and 0 mg K kg⁻¹ soil, depending on the P treatments (see below). Three weeks after sowing, another 100 mg of N (as KNO₃) was added to every pot (16 cm in height and 25 cm in diameter). The nutrients were mixed with the soil before pot filling. The soil was then placed in plastic pots (4 kg soil per pot).

The experiment was executed twice, in 2012 and 2013. In 2012, three Chinese maize varieties, which were bred in the last 60 years, were selected based on previous screening experiments for P acquisition strategy from the perspective of root morphological and physiological traits and mycorrhizal responsiveness (Chu et al. 2013): Huangmaya (HMY) with relatively high root length density, Zhongdan2 (ZD2) with relatively high acid phosphatase activity on the root surface, and 197 with relatively high mycorrhizal responsiveness. Moreover, another reason we selected these varieties was that they represented landrace (HMY), hybrid (ZD2), and inbred line (197), respectively. We have also executed earlier studies on genetic variation in mycorrhizal response of maize landraces, inbred lines and hybrids (Wang et al. 2020). In 2013, we tested one additional maize variety, Xianyu 335 (XY335), a newly released and widely grown maize variety, which can take up P from phytate via mycorrhizal hyphae (Wang et al. 2017).

In a field study, Wang et al. (2015) showed that Funneliformis mosseae (F. mosseae) colonized maize roots in different growth stages in the Northern Chinese Plain. Inoculum of the AMF species F. mosseae was obtained from Bank of Glomeromycota of China, Institute of Plant Nutrition and Resources, Beijing Academy of Agriculture and Forestry Research, Beijing. The fungus was propagated in a 5:1 mixture (w/w) of zeolite and river sand with maize for 4 months in a greenhouse. Inoculum consisted of soil containing spores (140 g^-1 soil), mycelium, and fine root fragments. Forty gram of the inoculum was added to pots of the mycorrhizal treatment before pot filling, and 40 g of the sterilized inoculum was added to pots of the non-mycorrhizal treatment. To minimize differences in microbial communities of mycorrhizal and non-mycorrhizal treatments, 10 ml of AMF-free filtrate from the inoculum was added to each pot of the non-mycorrhizal treatment, and 10 ml of deionized water was added to each pot of the mycorrhizal treatment. 20 g non-sterilized soil was submerged into 500 ml deionized water for two hours, and the solution was filtered by two-layer filter paper with 30 µm filter pore size. We checked the filtrate under the stereomicroscope and confirmed there were no AMF spores or hyphae. The filtrate was therefore regarded as AMF-free filtrate.

The diagram of the experimental design is shown in Figure 1. The experiment in 2012 was a three-factor completely randomized block design. The factors included: (i) mycorrhiza (with or without inoculation with F. mosseae); (ii) planting pattern (monocultures and two-variety mixtures); and (iii) P forms (no P, KH₂PO₄ or phytate). Both monocultures and two-variety mixtures contained two seedlings per pot. Each monoculture consisted of two seedlings of the same variety, and each two-variety mixture consisted of one seedling of each maize variety. There were three monoculture treatments and three two-variety mixture treatments of three maize varieties (HMY, ZD2, 197). Phosphorus treatments consisted of a control (no P added), addition of an easily available P source (20 mg P kg^-1 soil as KH₂PO₄), and addition of a sparingly available P source (20 mg P kg^-1 soil as Ca-phytate (Sigma-Aldrich, St. Louis, MO, USA). We tested three P-sources, as our earlier study showed differential abilities of two maize varieties (HMY and XY335) to acquire organic P from phytate (Wang et al. 2017). For comparison, we added treatments without P and with an easily available P source. There were four replicates, giving 144 pots divided over four blocks. The experiment in 2013 was a two-factor completely randomized block design with (i) mycorrhiza (as in 2012); and (ii) planting pattern (monocultures and two-variety mixtures). We used four varieties (HMY, ZD2, 197, XY335), resulting in four monoculture treatments and six two-variety mixtures). Based on the results from 2012, all pots received 20 mg P kg^-1 soil as Ca-phytate. Addition of other nutrients was the same as in 2012. There were four replicates, giving 80 pots divided over four blocks.

Maize seeds were surface-sterilized in 10% (v/v) H₂O₂ for 10 min and rinsed in deionized water. In the monoculture treatments, four seeds were sown in each pot which were thinned to two seedlings after emergence. In the mixed treatment, two seeds of each variety were sown in each pot which were thinned to one seedling per variety after emergence. Tap water was supplied daily and the pots were weighed twice per week to adjust soil moisture content to 18% (w/w); differences in plant weight between treatments were ignored. The glasshouse temperature range was 23–31 ˚C. The natural light was given without supplementary light.

Plants were harvested about 6.5 weeks after sowing in both years (47 days, sowing date 12^th July in 2012; 45 days, sowing date 7^th July in 2013). At harvest, roots were carefully removed from the soil and washed under running water. Roots were cut into 1-cm segments and thoroughly mixed. A 0.5-g sub-sample from each root sample was cleared with 10% (w/v) KOH at 90 °C for 2 h and stained with Trypan blue. Mycorrhizal colonization was assessed by the method of Trouvelot et al. (1986) , where 30 root segments per root sample were observed under the microscope at 200x magnification. The shoots were oven-dried at 70 °C for 3 days, and dry weights recorded. Shoot P concentration was determined by the vanado-molybdate method (Murphy & Riley 1962).

Acid phosphatase (phosphomonoesterase) activity of rhizosphere soil was assayed in 2013, according to the method by Neumann (2006): 0.5 mL soil solution was transferred into 2 mL Eppendorf reaction vials, to which 0.4 mL of 200 mM acetate buffer (pH 5.2) and 0.1 mL of 150 mM substrate [p-nitrophenylphosphate (pNPP); Sigma St. Louis, MO, USA] were added. The mixture was incubated for 30 min at 30 °C, after which the reaction was terminated by addition of 0.5 mL of 0.5M NaOH and centrifugation for 10 min at 12000 × g which also removed soil particles. Controls (one control for each soil sample to correct for background signal caused by humic substances) consisted of the same materials as soil samples and the substrate was added only after incubation. The concentration of pNPP in the supernatant was measured spectrophotometrically at 405 nm (Joner & Johansen 2000).

Processing of the dataset

We assessed both relative overyielding, through the calculation of RYT, relative yield total, and absolute overyielding, calculated according to Loreau & Hector (2001). The relative yield total (RYT_biom or RYT_P) was calculated as:

                         (1)

where Y1 and Y2 are the biomass or P content of maize variety 1 and 2 in mixture, and M1, M2 are the biomass or P content of maize variety 1, 2 in monoculture (Gliessman 1985). Overyielding occurs if RYT is significantly larger than one. Overyielding was also assessed for AMF by calculating RYT for fractional root colonization. Finally RYT was calculated for acid phosphatase activity.

We also calculated absolute overyielding, the difference between observed and expected yield and partitioned it in its two components, selection effect and complementarity effect according to Loreau & Hector (2001), who provided the ecological theory underlying this partitioning. We did this partitioning for every mixture (15 mixtures):

∆Y=Y_O-Y_E= NΔRY M+Ncov(∆RY, M)                 (2)

where Y_O and Y_E refer to observed and expected yield, respectively; NΔRY M  measures the complementarity effect, and Ncov(∆RY, M) measures the selection effect. We calculated ΔY, NΔRY M  and Ncov(ΔRY, M) for the average (with 95% confidence intervals) of the experiments in 2012 and 2013.

The relative mycorrhizal biomass-responsiveness and mycorrhizal P-responsiveness (MBR; MPR) was calculated according to:

MBR or MPR=Mc-NMcM*100             (3)

where Mc and NMc refer to the shoot biomass or shoot P content of mycorrhizal and non-mycorrhizal plants, respectively (Plenchette et al. 1983) . We calculated MBR/MPR for every pair in the same block regardless of monoculture or mixture.

We tested for significant overyielding for every maize variety combination in every year by calculating, per block, the expected yield or P uptake of the mixture (expected refers to the means of shoot biomass or P content of two mono-cultured maize varieties) and compared that with the actual yield of the mixture. We subsequently tested, through a t-test with n = 4, whether the difference between observed and expected yield or P uptake was significant for the two pot experiments. We evaluated the role of mycorrhiza in overyielding in two ways. First, we tested, through Fisher’s exact test, whether there was an association between mycorrhiza and significant overyielding across two pot experiments (based on n = 15). Then we tested whether the average RYT for all 15 maize mixtures for either biomass yield or P uptake differed significantly between the non-mycorrhizal and mycorrhizal treatment.

For mycorrhizal colonization, hyphal length density and acid phosphatase activity, we calculated the RYT as described above. We applied t-tests to test for significant differences (based on the average of four replicates) between predicted and observed values.

Data were tested for the requirements on homogeneity of variance (Levene’s test), and in case of non-compliance were subjected to logarithmic or angular transformation. Statistical analyses were performed with SPSS 20.0 (IBM Corp., Armonk, NY, USA). We used the 5% level of probability to judge whether effects were statistically significant. Correlation between variables was tested using Pearson’s correlation coefficient (P < 0.05).