Root colonizing Trichoderma fungi can stimulate plant immunity, but net effects are strain × cultivar-specific and changing ambient conditions further contribute to variable outcomes. Here, we used four Trichoderma spp. to inoculate seeds of four common bean (Phaseolus vulgaris) cultivars and explored in three different experimental setups the effects on fungal anthracnose after leaf inoculation with Colletotrichum lindemuthianum. Plants growing in pots with field soil under greenhouse conditions exhibited the highest and those in the open field the lowest overall levels of disease. Among 48 Trichoderma strain × bean cultivar × setup combinations, Trichoderma-inoculation enhanced disease in six and decreased disease in ten cases, but with the exception of T. asperellum B6-inoculated Negro San Luis beans, the strain × cultivar-specific effects on anthracnose severity differed among the setups, and anthracnose severity did not predict seed yield in the open field. In the case of Flor de Mayo beans, Trichoderma even reduced yield in anthracnose-free field plots, although this effect was counterbalanced in anthracnose-infected plots. We consider our work as a case study that calls for stronger emphasis on field experiments in the early phases of screenings of Trichoderma inoculants as plant biostimulants.

The seeds of the four common bean cultivars Flor de Junio Marcela (FJM), Flor de Mayo Anita (FMA), Pinto Villa (PV) and Negro San Luis (NSL) were inoculated each of with the Trichoderma strains  T. longibrachiatum MK1, T. asperellum B6, T. atroviride P1 and T. harzianum T2 (or no Trichoderma as control) and cultivated in three different experimental setups: (A) in pots with a sterile, commercial substrate in a greenhouse, (B) in pots with non-sterile field soil in the greenhouse and (C) in open field plots.

The field experiment was performed in the experimental field of CINVESTAV Unidad Irapuato (State of Guanajuato, 1.800 m above sea level: 20°43′13″ N; 101°19′43″ W) during the spring-summer season of 2019 (from March to August). We established four subplots, two subplots assigned to inoculation with Colletotrichum, the other two subplots as anthracnose-free controls. On 26 and 27 March, n = 8 seeds per strain × cultivar combination and subplot were sown directly into the soil in a completely randomized spatial distribution.

In all three setups, plantlets with 3–5 trifoliate leaves were leaf-challenged with Colletotrichum lindemuthianum. The severity of anthracnose disease was quantified 20 days after the challenge with C. lindemuthianum, collecting one randomly selected leaf per plant which was scanned using a printer equipped with scanning function (Brother DCP-1602) to subsequently determine the total leaf area and the diseased leaf area using the image analysis software Image J and to calculate the percentage of affected area for each leaf.

In case of the field—grown plants, in addition to quantifying the diseased area, the area removed or visibly damaged by herbivores was quantified to calculate herbivore damage as percentage of the total area. Subsequently, plants were allowed to finish their growing cycle. At the end of the reproductive phase, i.e., when plants started to dry naturally, we first counted the number of plants that had survived until reproduction. Plants were harvested individually (either all plants or in case of survival of more than five plants per subplot, a maximum of five randomly selected plants per strain ×x cultivar combination and subplot), the pods were removed and opened manually and the seeds dried in an oven at 50 °C for two days, to determine seed dry weight per plant.