The outcomes of pathogen infection can be sensitive to temperature, interactions with other pathogen species, and host age. Yet few studies have experimentally tested how warming alters infection outcomes for multiple, potentially interacting, pathogen species across host ages. In this study, we conducted a factorial experiment to test how plant age (7-week vs. 13-week plants) and temperature (21 °C vs. 29 °C) influence infection outcomes of two foliar fungal pathogens with contrasting feeding strategies—Rhizoctonia solani (a necrotroph) and Colletotrichum cereale (a hemibiotroph)—in the grass species tall fescue (Lolium arundinaceum). Contrary to expectations of co-infection with pathogens of opposing life-history strategies leading to increased disease symptoms, co-infection had relatively minor effects across disease metrics. Instead, infection outcomes were driven by host age, pathogen identity and temperature. In plants inoculated with R. solani, higher temperature reduced lesion severity, and independently, severity was less in older plants. C. cereale lesion development showed a strong age x temperature interaction, with older plants being more resistant to disease in cooler conditions but not under warming. In plants co-inoculated with both pathogens, elevated temperature reduced disease severity, and this effect was stronger in older plants. These findings demonstrate that environmental conditions and host age can both interact and outweigh within-host pathogen interactions, highlighting the importance of incorporating host demographic structure in predicting disease responses to climate warming.  

Data analysis: All analyses were performed using RStudio (version 4.4.2, Pile of Leaves) and each model included random intercepts for treatment blocks nested within replicates to account for variation between blocks. Plant survival was assessed using a generalized linear mixed model (glmer) with a binomial error distribution and logit link (Bates, 2014). Survival to the end of the study (number of plants surviving out of the total inoculated) was modeled as a function of the interaction between temperature, age group, and inoculum. Significance of effects was assessed using Type II Wald chi-square tests, implemented with the Anova() function from the car package (Fox, 2019). Only surviving plants were considered in further analyses. 

The primary analysis of disease used two metrics -- lesion development and disease severity. Lesion development  (whether the plant had developed lesions matching the inoculated pathogen(s) by the final day of observation) was modeled using the same model structure and significance testing as the survival model, with lesion presence at Day 20 as a function of the interactive effects of temperature, age group, and inoculum. For R. solani, infection was nearly universal across inoculation treatments, precluding tests of inoculum effects; analyses therefore focused on the effects of temperature and age group. 

Disease burden was calculated at each time point by assessing Percent Leaf Damage (hereafter, PLD). PLD was determined by visually assessing the proportion of the inoculated leaf covered in lesions--for example, an inoculated leaf with 50% lesion coverage would receive a PLD score of 0.5. Inoculated leaves which died over the course of the survey period were said to have 100% damage, with a PLD score of 1. The severity of the disease burden experienced by plants over the course of the experiment (disease severity) was quantified by calculating the Area Under the Disease Progression Stairs (AUDPS) using the agricolae package. AUDPS was defined as the integrated total of PLD over time. To separate disease severity from disease presence/absence, AUDPS was calculated only for plants which developed lesions matching their respective inoculation treatment (R. solani: n = 230, C. cereale: n=235, coinfection: n=215). Due to overlap between R. solani and C. cereale infection symptoms (chlorosis, necrosis, and senescence), disease severity measurements for co-inoculated plants were not differentiated by lesion type. Because the distribution of AUDPS measurements was strongly right-skewed, values were log-transformed prior to statistical analysis. Within each inoculum grouping (R. solani, C. cereale, or co-inoculation), log-transformed AUDPS was modeled as a function of the interactive effects of temperature and age group with a random effect of replicate that was allowed to vary by treatment block. 

Differences in dry biomass across temperature, age group, and inoculation treatments were modeled using linear mixed-effects models, with log-transformed dry biomass as a function of the interaction between temperature, age group, and inoculum. This model included random intercepts for the treatment blocks nested within replicates to allow for variation between the treatment blocks within each temporal replicate.