Seedlings of I. deltoidea preferentially occur under shady conditions. Extensive sampling at two sites in western Amazonia showed that more than 91% of I. deltoidea seedlings were found in understory conditions. The total number of seedlings in the Northeastern Peru transects was 660, 94% of these seedlings were located at understory conditions (canopy scope <5). The negative correlation between the number of I. deltoidea seedlings and canopy openness in the 102 transects (Spearman r = −0.117, P<0.05) was estimated using n = 280 5×5 m subplots with at least one I. deltoidea seedling. Statistical significance was assessed as a one-tailed test and correcting for spatial autocorrelation using Dutilleul's approach for computing the geographically effective degrees of freedom  = 227 [30]. The total number of seedlings in the Southeastern Peru plots was 518 seedlings (less than 25 cm), 91% seedlings were located in dense understory (∼55–120±15 µmol m⁻² s⁻¹).

We found that D. mutila is beneficial to I. deltoidea under understory conditions, but strongly reduces I. deltoidea's capacity for recruitment in high-light forest gaps. Additionally, the foliar necrotic spot symptoms produced by D. mutila appeared more frequently in seedlings and juveniles that grew in gaps or diffuse open canopy conditions: Plants with visible symptoms caused by D. mutila received significantly higher illumination, 408.3±17.3 µmol m⁻² s⁻¹ than plants with no visible symptoms receiving lower illumination, 208.2±6.1 µmol m⁻² s⁻¹, (mean ± SE, t test, n = 808, P<0.0001). Disease development was faster and more lethal in seedlings with two leaves or less when exposed to higher light conditions (F _1,22 = 55.4, P = 0.0001, r² = 0.73) (Fig. 1A). These results are consistent with the landscape recruitment patterns of I. deltoidea mentioned above. Ontogenic or age-related resistance may be responsible for differences in disease expression between seedlings in different stages of development [31]. An additional experiment showed that pathogenicity of D. mutila increased with light availability. We inoculated 22 healthy 6-month old I. deltoidea seedlings (no foliar spots or insect marks) with D. mutila, following inoculation procedures from previous studies [25]. Foliar spots produced by D. mutila had a higher growth rate and mortality was greater and faster at higher light availability (F_1,22 = 93.26, P = 0.0001, r² = 0.816) (Fig. 1B).

Using transplant experiments we demonstrated that increased light availability switched the endosymbiotic phase of the fungus to its pathogenic phase. In a greenhouse experiment diametric growth rate of foliar spots produced by D. mutila was higher in full sun conditions (mean 19.5±2.5 SE cm/day) than in reduced light (10.0±2.5 cm/day) and full shaded conditions (0.52±2.5 cm/day; analysis of variance (ANOVA), F_3,30 = 12.62, P = 0.0001; Fig. 2). Diplodia mutila-induced seedling mortality in plants exposed to full sun was 80% after ten days. Seedlings under shaded conditions had 10% mortality and seedlings in the greenhouse had 40% mortality.

Laboratory assays showed that fungal growth (measured as diameter of mycelial colonies or as density of mycelium comprising colony) was greater when a 12-hr alternating light-dark cycle was provided than when periods of light were restricted to 3 hours. On Water Agar medium (WA) the average growth rate per day of the colony mycelium for five days was higher under a 12-hour light cycle (mean 0.52±0.03 SE cm/day) than under a 3-hour light cycle (0.38±0.03 cm/day, t test, n = 12, P<0.004). On Potato Dextrose Agar medium (PDA) the average growth rate per day of the colony mycelium was faster and also higher under the longer light period (mean 1.25±0.01 SE cm/day), compared to 1.11±0.11 cm/day for the 3-hour photoperiod (t test, n = 12, P<0.018) and the mycelium was notably denser with more aerial mycelium (Fig. 3A). We recorded greater melanization of mycelium in colonies exposed to the longer light period. This was especially evident in colonies grown on PDA: Colonies grown in PDA under the 12-hour light cycle had significantly faster growth of the central melanized area (mean 0.71±0.05 SE cm/day) than colonies exposed to the 3-hour light treatment (0.5±0.05 cm/day, t test, n = 12, P<0.022; Fig. 3B, 3D). The absorption spectrum of methanolic extracts of D.mutila cultures showed a characteristic peak in the UV region of 200–260 nm [32]. HPLC chromatograms of methanolic extracts of light and dark grown culture extracts showed that light grown cultures have at least twice the amount of melanin than dark grown cultures. Melanization of mycelium has been linked to enhanced virulence in numerous plant and animal pathogenic fungi [33]. Similar significant results were obtained for colonies growing in WA medium (Fig. 3B).

Colonies grown on PDA under high light conditions showed higher hydrogen peroxide production than those grown under low light conditions as evidenced by staining using the DAB/peroxidase H₂O₂ assay (Fig. 4). Hydrogen peroxide has been shown to be a signal molecule for plants that triggers hypersensitivity cell and tissue death [34].

We found that stem borers (order Coleoptera) are the main mortality cause of I. deltoidea seedlings. After 50 days, 9% of the marked plants had died. Stem borers were the cause of 73% of the total palm mortality after 50 days, while D. mutila (in its pathogenic phase) caused just 2% of the mortality. After 150 days, stem borers induced 43% of mortality, while D. mutila (pathogenic phase) was responsible for a mere 0.1% of the mortality. We did not find any relationship between light availability and endophyte presence.

The proportion of plants affected by stem borers within the first 2.5 m near the parental tree was significantly higher than proportions in the other 4 annuli far away from the parental tree, in the first census: After ∼7 days, stem-borer infection was 8%±0.01% (mean ± SE) in the inner annulus, and 6%±0.01%, 4%±0.01%, 3%±0.01%, and 3%±0.01% in the outer annuli (going outwards) (One-way ANOVA, F_4,50 = 2.65, P>0.045*). After ∼50 days, the same numbers were 10%±0.01%, 4%±0.01%, 1%±0.01%, 2%±0.01%, and 0.7%±0.01% (One-way ANOVA, F_4,50 = 4.28, P>0.0051*).

We found evidence that D. mutila benefits its host plants by enhancing resistance to herbivory by some insects. Field surveys in the ten surveyed plots, showed that insect herbivory (stem borers) decreased with increasing incidence of D. mutila infection. Plots with few D. mutila-infested I. deltoidea plants had higher incidence of stem-borer mortality, whereas plots with higher incidence of plants colonized by D. mutila had lower rates of stem-borer- induced mortality (ANOVA, F_{1, 10} = 18.49, P = 0.0026, r² = 0.69). Additional feeding experiments employing I. deltoidea fruits and PDA media (Potato Dextrose Agar) colonized by D. mutila showed that adults of the beetle Coccotrypes sp., and two unidentified species of larvae of the order Coleoptera avoided consumption of fruits and PDA colonized by D. mutila. Beetles consistently preferred PDA and avoided D. mutila infested PDA, mean 4.8 beetles ±0.14 SE versus 1.4 beetles±0.14 (Repeated measures ANOVA, F_{4, 456} = 160.13, P = 0.0001**). Similar results were also obtained when the experiment was performed using Coccotrypes sp. adults and I. deltoidea fruits instead of PDA.

In Northeastern Peru [cf. 39] we arbitrarily placed 102 transects (5×500 m, divided in 5×5 m subunits) located in mature primary tropical rain forest within 300 km of Iquitos, Peru (excluding transects located in secondary forests, white sand soils, steep topographical conditions and human disturbed forests). Sites in Southeastern Peru were located at Cocha Cashu, (CCBS) [40] and Los Amigos, (LABS) [41]. Ten plots were established in May 2007, in primary floodplain forest, with similar floristic composition and topographic characteristics. Five of plots were located at CCBS and five at LABS. Nine plots measured 900 m² and one plot at CCBS measured 2.25 ha. In each plot all I. deltoidea plants were tagged with numbered plastic tags and mapped in an X - Y coordinate system. The total number of plants located in the ten plots was 1068: 63 fruiting adults, 518 seedlings and 487 were considered juveniles-adults (non-fruiting). We measured height of the tallest photosynthetic leaf (cm) and number of leaves and diameter of foliar spots caused by D. mutila (cm) for all seedlings. The amount of disease was calculated by dividing the diameter of the foliar spot by the diameter of the affected leaf and expressed as ‘% disease’. Disease development over a period of 150 days was calculated by subtracting the % disease in the initial estimate from the final estimate.

Plants damaged and killed by epicotyl borers, such as caterpillars, beetle larvae or crickets, were considered as “damaged by stem borers”. Plants located in the Southeastern Peru plots were monitored for presence/absence of D. mutila and stem borers, three times after initial establishment (7, 50 and 150 days). In each plot, the minimum distances from all seedlings to the nearest I. deltoidea fruiting plant were computed using the coordinates of the labeled plant under consideration and the coordinates of the nearest fruiting tree within the plot. We surveyed seedlings in 5 concentric 2.5 m annuli centered on a focal fruiting tree. The number of seedlings affected by stem borers and D. mutila was tallied for each 2.5 m annulus and then divided by the total number of plants located in the selected annulus to yield proportions. One-way ANOVA was used to compare diseases and mortality proportions among plots for each distance annulus (Tukey's HSD used to contrast means).

In Northeastern Peru light availability was measured using the canopy scope methodology [42]. In Southeastern Peru light availability was estimated above the tallest photosynthetic leaf of each I. deltoidea seedling, using the average value of light intensity over the leaf with a light meter (Environmental Concepts Plant Light Intensity Meters, LIM2500, USA). The average value was obtained from three measurements over each plant at 6 am, 12 pm and 5 pm for three consecutive days.

We transplanted 30 I. deltoidea seedlings from one plot at Cocha Cashu where adults, juveniles, seeds and fruits were colonized by D. mutila. Ten seedlings were transplanted to shade conditions, ∼55±15 µmol m⁻² s⁻¹, ten to a reduced light environment in a greenhouse, ∼491±34 µmol m⁻² s⁻¹, and ten to full sun exposure, ∼1058±23 µmol m⁻² s⁻¹. All seedlings had two leaves and did not have any visible disease symptoms produced by D. mutila or any other foliar spot. Light availability was measured three times a day (6 am, 12 pm and 5 pm) for a period of ten days and all disease symptoms and insect damage were recorded and measured daily. The average daily temperature in the understory and full sun conditions was 23±3°C and 26±5°C in the greenhouse.

On December 2007, ∼370 Coccotrypes sp. beetles and larvae were extracted from more than 100 fruits and seeds of I. deltoidea. In a Petri plate (60×15 mm, Fisher Scientific Co. Canada) we placed two 1-cm² of PDA (Potato Dextrose Agar) and two 1-cm² of PDA infested with D. mutila, covered with squares of non-acidic paper to simulate dark conditions found inside seeds and fruits. PDA was replaced everyday for the duration of the experiment to avoid contamination of non-infested PDA by D. mutila. We set up 12 repetitions following this procedure. Six to ten beetles were released in the each Petri plate and monitored daily for 8 to 12 days.

To assess the effect of light on the fungus, laboratory observations were made on mycelial growth in Water Agar (WA) and Potato Dextrose Agar (PDA). Two photoperiod treatments were employed for five days with six D. mutila samples per treatment. The first treatment consisted of 12-hour cycle of darkness and 12-hour cycle of white light (fluorescent, 100±10 µmol m⁻² s⁻¹), while the second consisted of 21-hour cycle of darkness and 3-hour cycle of light for five days, (6 repetitions per treatment, constant temperature for all treatments was 24°C).

Methanolic extracts of culture discs from D. mutila grown at room temperature and 30°C in dark and light conditions were analyzed by HPLC [34]. Fungal cultures were extracted in methanol 80% at day 10 of the experiment. Samples were concentrated using a rotovapour (model 11, Buchi, Flawil, Switzerland) at 45°C and dried in speed vacuum (Savant PSD 2010 Speed Vac Concentrator, Thermo). Dried samples were dissolved in 1 ml of methanol 80% with add of vortex and sonication, and filtered using 0.45 µm Nylon (CostarR) filters held in a 2.0 ml polypropylene tube and centrifuged at 5000 rpm for 45 sec before analysis. Sample analysis was conducted with HPLC (Waters, Milford, MA) using a C 18 Luna column (4.6×150 mm; particle size 5 µm; Phenomenex, Torrance, CA). Samples (20 µl) were injected into the column, and gradient elution was used for fractional separation with two solvents at a flow of 1 mL/minute, Solvent A consisted of 10% Methanol in H₂O adjusted pH 3.5 with formic acid0. Solvent B consisted of 20% water (pH 3.5), 20% methanol and 60% acetronitrile. The gradient consisted of 0 minute at 100% A, followed by 5 minutes at 85% A, 10 minutes at 80% A, 20 minutes at 75% A, 25 minutes at 73% A27 minutes at 60% A, 30 minutes at 60% A, 35 minutes at 50% A, 40 minutes at 10% A, 45 minutes at 0% A, and 50 minutes at 100% A and holding at 100% A for a final 55 minutes. Five minutes of equilibration at 100% A was performed before and after each injection.

Fungal cultures were grown on PDA medium and incubated in two different temperatures (25°C and 30°C) and light conditions (dark and light). Accumulation of hydrogen peroxide was evaluated after 5 days of growth following method described by Munkres (1990) [43]. In brief, hydrogen peroxide was indicated by the accumulation of a red precipitate after being flooded with 5 ml solution of 100 mM potassium phosphate buffer, pH 6.9, 2.5 mM diaminobenzidine tetrachloride and 5 purpurogallin units/ml of horseradish peroxidase (Type VI, Sigma Chemical Company, St. Louis, MO). Plates were incubated at 30°C in dark for one hour before the solution was discarded. Afterward, plates were inverted and incubated at 30°C in the dark for 24 h and then photographed.