Introduction

European canker, caused by Neonectria ditissima (Tul. & Tul.) Samuels & Rossman, is one of the most important diseases of apple in regions with a mild and humid climate (Beresford and Kim 2011). This includes the Lower Elbe region of Northern Germany which comprises some 10,000 ha of fruit trees, about 90% of them being apple (Malus domestica). As a wound pathogen, N. ditissima primarily causes cankers on the bark of twigs and trunks (Saville and Olivieri 2019). In addition, infections at flowering may lead to blossom-end fruit rot in summer (Weber and Dralle 2013), whereas infections of ripening fruit result in a post-harvest rot which emerges during prolonged storage (Xu and Robinson 2010). Canker lesions are perennial. They give rise first to asexual micro- and macroconidia which are dispersed by rainsplash mainly within the tree in which they are formed, and later to sexually produced ascospores which are airborne and can be distributed over long distances (Weber 2014). Both macroconidia and ascospores are highly infectious (Latorre et al. 2002; Weber 2014), whereas microconidia are probably irrelevant at least under commercial fruit production conditions (Wesche and Weber 2023).

Canker lesions thus serve as a source of infective spores both for new cankers and for fruit rots. In all canker-affected growing regions, the control of N. ditissima is achieved by a combination of cultural measures such as canker pruning, and chemical control. In Northern Germany (Weber and Børve 2021) as well as New Zealand (Amponsah et al. 2015), new cankers are initiated predominantly in autumn by infections of fresh scars arising from fruit picking and leaf fall. Protectant fungicides such as captan- or copper-based products are applied specifically against N. ditissima at that time (Walter et al. 2019; Weber and Børve 2021). In order to control fruit infections, fungicide treatments at flowering or before harvest are recommended (Xu and Robinson 2010, Weber 2014), but the choice of compounds is far from settled. In New Zealand, methyl benzimidazole carbamate (MBC) fungicides such as carbendazim have been used for the control of apple canker in the past. In addition to treatment of leaf scars in autumn, applications were made in spring to control fruit rot (Walter et al. 2014). Similar results for the control of fruit rot have been reported from England (Cooke et al. 1993), and for canker control from the Netherlands (de Jong and van der Steeg 2012). However, MBCs have not been registered in most European countries for some time. In addition, there are indications of a partial MBC resistance among northern German (Weber and Palm 2010) and New Zealand isolates (Walter et al. 2014). Among the more modern fungicides, pre-harvest sprays of fludioxonil, fludioxonil + cyprodinil, and boscalid + pyraclostrobin are recommended in England in situations of enhanced storage-rot risk (Saville and Olivieri 2019). In addition, dodine and dithianon + potassium phosphonate have been shown to have an effect against Neonectria fruit rot (Berrie 2016).

Chemical control recommendations for fruit rots continue to differ between fruit-growing regions. This may be due to differences in the national approval of active substances, the use of combination products, and uncertainties about the relevance of different wound types, timing of applications, and inoculum (Weber and Børve 2021). Further, apart from the above-mentioned indications of a reduced susceptibility of N. ditissima to MBC fungicides there appear to be no comprehensive data on the sensitivity of N. ditissima to single-site compounds. For this reason we examined compounds belonging to different chemical classes for their effective concentrations causing 50% inhibition of germ-tube growth (EC50) of a wide range of N. ditissima isolates, and conducted experiments on wound-inoculated detached fruits and on floral infections in the orchard with selected fungicides.

Materials and Methods

Isolates and Preparation of Conidia

A total of 43 N. ditissima isolates (Table 1) were obtained from the north (Lower Elbe region), west (Hesse, North-Rhine Westphalia) and south (Lake Constance) of Germany as well as from South Tyrol. Tree cankers as well as pre- and post-harvest fruit rots were sampled from orchards under Integrated Pest Management (IPM) as well as organically managed or abandoned orchards. Four strains were collected from cankers on Sorbus aucuparia trees that were located several hundred metres away from the nearest site of any fungicide input. All isolates produced micro- and/or macroconidia in agar culture.

Table 1 Overview of isolates, region of origin, diseased host organ, year of isolation and types of conidia produced in pure culture

All isolates were kept as lyophilized vacuum-sealed conidial preparations (Smith and Onions 1983) and were revived on potato dextrose agar (PDA) to produce fresh conidia. PDA cultures were incubated at room temperature (approximately 20 °C) for a maximum of 3 weeks and irradiated with a daily 10-min burst of UV light (λmax = 365 nm) to stimulate conidial production. All chemicals were supplied by Carl Roth (Karlsruhe, Germany) unless indicated otherwise.

Determination of EC50 Values

Sensitivity tests on agar medium were performed as previously described (Weber and Hahn 2011; Weber et al. 2015). Fungicides were added to the cooling agar medium after autoclaving as stock solutions of commercial preparations to give the concentrations listed in Table 2. Fresh conidia were harvested from PDA plates in sterile deionized water, adjusted to a concentration of 105 ml−1 and then applied as droplets of 15 µl per agar plate. After 12–14 h incubation at 20 °C, 10 representative germ tubes were measured for each fungicide concentration using an Axio Scope A.1 (Carl Zeiss, Germany) fitted with a 10 × objective and an eyepiece graticule to give a 100 × final magnification. All measurements were repeated between two and three times for each isolate and each fungicide concentration. The mean germ-tube length was calculated, and the percentage of germ-tube growth relative to the fungicide-free control was determined. EC50 values were calculated by regressing germ-tube length (percent of control) against log10 fungicide concentration. To examine possible cross-resistances between the succinate dehydrogenase inhibitors (SDHI) boscalid and fluopyram, and between the anilinopyrimidine-type fungicides pyrimethanil and cyprodinil, EC50 values were log10-transformed and subjected to a regression analysis (P < 0.05), followed by the calculation of a Pearson correlation (Amiri et al. 2014).

Table 2 Details of media used in the in vitro germination assay

Where possible, measurements were performed on germ tubes emerging from five-septate macroconidia. However, microconidia were measured for isolates that failed to produce macroconidia in culture. Because macro- and microconidia differ in their germination process and growth kinetics (Wesche and Weber 2023), EC50 values for macro- and microconidia were determined and compared for individual isolates producing both spore types. For this purpose, the Pearson correlation and a regression analysis were calculated to show the correlation between both conidial types.

Preparation of Inoculum for Detached-Fruit Assays and the Field Trial

Fresh conidia from 14-day-old PDA cultures of N. ditissima isolate OVB 13–062 were harvested in sterile deionized water by scraping the colony surface with a sterile glass slide and filtering the spore suspension through sterile cotton wool. Macroconidia were the predominant spore type produced under the experimental conditions, and their concentration was adjusted to give suspensions as required. The suspensions were stored at 2 °C for a maximum of 6 h prior to inoculation. In order to test for spore viability, 100 µl suspension samples were incubated on PDA for 12 h at room temperature. The germination rate was > 98% for all suspensions used in inoculation experiments.

Detached-Fruit Assay

Apple fruits (cultivar ‘Elstar’) were disinfected by swabbing with 70% ethanol (w/v) and wounded at four positions with a sterile nail (2 mm diameter, 2 mm deep). Fruits were then dipped in the test fungicides at concentrations given in Table 3, followed by air-drying in a laminar flow cabinet. After drying, 20 µl conidial suspension of isolate OVB 13–062 was applied to each wound to give 50, 500 or 5000 conidia per wound. The fourth wound served as a control and was treated with 20 µl sterile deionized water. For each variant, 25 inoculated fruits were incubated at room temperature in a randomized block design in five humidity chambers. The radius of each fruit lesion was measured after 21, 28 and 35 days’ incubation as the average of the smallest and largest distance between the edge of the wound and the advancing edge of the rot lesion. In addition, the incidence of infection (%) was determined after 35 days. In order to verify infections by N. ditissima, tissue from infected fruit was incubated on PDA augmented with streptomycin sulphate and penicillin G (each at 200 mg l−1). A total of three separate fruit assays was set up in order to test the effects of different fungicides against untreated control fruit. These assays concerned fludioxonil, trifloxystrobin and cyprodinil; dodine; and boscalid, fluopyram and fludioxonil.

Table 3 Fungicides used in inoculation experiments

For statistical analysis, the mean lesion radius (mm) was log10 transformed and the disease incidence (%) was arcsin transformed. A two-factorial analysis of variance (ANOVA) (P < 0.05) with block design was then conducted to test for differences between fungicide variants and conidial concentrations. A multiple comparison of means was based on the Tukey test. Homogeneity of variance was previously tested for all analyses using the Levene test. All statistical analyses were performed using the SPSS 27.0 software (IBM, Armonk, NY, USA). In addition, the degree of efficacy of each fungicide on the basis of disease incidence was calculated according to Abbott (1925).

Field Trial for Blossom-End Rot

The experimental plot comprising 5‑year-old slender spindle trees (cv. ‘Nicoter’ on rootstock M9) was located on the premises of the Esteburg Fruit Research and Advisory Centre (Jork, Northern Germany; 53.51 °N, 9.75 °E). The orchard was treated according to IPM guidelines, except for the time of flowering, when no fungicides were applied. The trial was set up as a completely randomized block design with four replicates per variant, each replicate comprising two neighbouring trees. All experimental trees showed adequate flowering and were free from visible cankers at the time of inoculation. Fludioxonil, trifloxystrobin and cyprodinil were applied as commercial fungicides at the recommended rates (Table 3). Each experimental tree was sprayed to surface wetness with about 300 ml fungicide preparation. The fungicides were applied on 30 April 2022 at 7–8 a.m. before bee flight. At this time the flowers on 1‑year-old wood were in full bloom (BBCH 65), whereas those on older wood were at incipient petal fall (BBCH 67), thus showing the most susceptible developmental stage to floral infections by N. ditissima (Holthusen and Weber 2021). At 2–4 p.m. on the same day, experimental trees were inoculated twice with 250 ml conidial suspension (105 macroconidia ml−1) at an interval of 60 min. A fungicide-free control was also inoculated with the conidial suspension. To determine the natural infestation incidence in the plant, additional experimental trees were sprayed with water only. All experimental trees were repeatedly moistened with water after inoculation until no further drying occurred beyond 10 p.m. The weather was cloudy at the time of inoculation, and the temperature was 14 °C in the early afternoon, dropping to 11 °C in late evening.

Fruits showing clear blossom-end rot symptoms were removed from the trees and counted. To confirm N. ditissima as the pathogen, the characteristic separation of decayed and healthy tissue (Weber 2014) was examined in the lab, and tissue samples from representative fruits were incubated on PDA augmented with antibiotics. Similarly, tissue samples from all fruits with ambiguous disease symptoms were incubated on PDA. The incidence of infection was determined for each tree.

The SPSS 27.0 software was used for statistical analysis. A two-factorial (block, treatment) ANOVA analysis (P < 0.05) was conducted, based on generalized linear models with binomial distribution considering logistic regression and Pearson correlation. The dependent variable was the number of infected fruits per tree and the experimental variable was the total number of fruits. The multiple mean comparison was based on the Sidak correction (P < 0.05).

Results

Fungicide Sensitivity in Vitro

All N. ditissima isolates were highly sensitive to fludioxonil, with EC50 values ranging from 0.009 to 0.046 ppm (mean 0.019 ppm; Table 4), and also to trifloxystrobin (0.018–0.426 ppm; mean 0.150 ppm). In contrast, all EC50 values were above 1 ppm for cyprodinil (5.04–83.66 ppm; mean 25.98 ppm), thiophanate-methyl (2.05–15.90 ppm; mean 5.69 ppm), fluopyram (1.35–140.05 ppm; mean 13.98 ppm), and also dodine (1.38–6.33 ppm; mean 3.03 ppm) with one exception (0.61 ppm). There were no obvious differences between isolates from different growing regions or between isolates from IPM, organic or abandoned orchards. The responses to fluopyram were highly heterogeneous, varying by a factor of > 100 between the most and least sensitive; for trifloxystrobin the range was nearly 25-fold and for cyprodinil nearly 17-fold. Differences of < 10-fold between the lowest and the highest EC50 values were recorded for dodine, fludioxonil and thiophanate-methyl.

Table 4 EC50 values [ppm] of all fungicides tested for all isolates

Selected isolates were also examined for sensitivity to boscalid and pyrimethanil. A regression analysis revealed no significant relationship between the EC50 values of the same isolates for the anilinopyrimidine fungicides cyprodinil and pyrimethanil (P = 0.779; r = −0.102). For boscalid, the EC50 values were > 500 ppm in all isolates tested, and no significant relationship between the SDHI fungicides fluopyram and boscalid was recorded (P = 0.653; r = 0.099). In contrast, the EC50 values for germ-tube growth from macro- versus microconidia of eight isolates capable of producing both spore types showed a very high correlation for all active ingredients (P < 0.001 in all cases; r = 0.642 to 0.961; not shown).

Detached-Fruit Assay

In all three fruit assays the number of macroconidia used for inoculation (50, 500 or 5000 per wound) showed a direct and significant (P < 0.05) correlation with the size of the resulting fruit rot lesion. This had an effect on the activity of the fungicides tested. In the first fruit assay, no significant differences between fungicides and the untreated inoculated control were observed at the lowest conidial inoculum dose (50 conidia per wound). At the two higher conidial concentrations, fludioxonil showed a significant (P < 0.05) reduction in lesion size per wound, whereas cyprodinil and trifloxystrobin did not (Fig. 1). Sporodochia of N. ditissima were observed on the fruit surface of four of the 25 control fruits after 35 days (Fig. 2a), but they did not form in fruit lesions on any of the fungicide-treated apples (Fig. 2b, c and d). Fruits treated with fludioxonil showed an externally visible brown-rot lesion only in rare cases. However, upon sectioning of the fruits at the end of the experiment browning around the inoculation site was frequently observed, especially at the highest inoculum dose. This infection-related symptom differed from the superficial phenolic browning caused by the mechanical wounding process (Fig. 2e) in extending into the adjacent fruit flesh tissue which had a corky texture, indicating a stalled growth of the pathogen into the fruit (Fig. 2f). Similar observations of cryptic infections not apparent from the outside were made for trifloxystrobin and cyprodinil in wounds inoculated with 500 and 50 conidia, respectively. When tissue samples from such fruits were incubated on PDA augmented with antibiotics, N. ditissima was isolated from 10 of 25 wounds with cryptic symptoms that had been treated with fludioxonil and subsequently given the highest dose of 5000 conidia. Similarly, the fungus was isolated from eight such cryptic infections at 500 conidia for trifloxystrobin, and two infections at 50 conidia for cyprodinil.

Fig. 1
figure 1

Lesion radius (mean ± SEM) on apple fruit 35 days after treatment with cyprodinil, trifloxystrobin or fludioxonil followed by wound inoculation with three different macroconidium concentrations of Neonectria ditissima, in comparison to a wound-inoculated untreated control. Different letters indicate significant differences between test variants (Tukey, P < 0.05)

Fig. 2
figure 2

Representative fruits wound-inoculated with 50 (bottom right), 500 (bottom left) or 5000 (top) macroconidia of Neonectria ditissima and incubated for 35 days in a damp-chamber at room temperature. a Fungicide-free control. b Cyprodinil treatment. c Trifloxystrobin treatment. d Fludioxonil treatment. e Cross-section through a non-inoculated control wound showing phenolic browning. f Cross-section through a retarded infection of 5000 conidia on a fludioxonil-treated fruit

In the second assay, dodine significantly reduced lesion growth compared to the untreated control (P < 0.05) at spore concentrations of 5000 or 500 conidia per wound (Fig. 3). As for the other fungicides, cryptic infections were occasionally found and were included in the infection incidence. Because of this, the efficacy of dodine against artificial fruit infections by N. ditissima was 55% at 50 conidia per wound but declined to 0% for the highest dose.

Fig. 3
figure 3

Lesion radius (mean ± SEM) on apple fruit 35 days after treatment with dodine followed by wound inoculation with three different macroconidium concentrations of Neonectria ditissima, in comparison to a wound-inoculated untreated control. Different letters indicate significant differences between test variants (Tukey, P < 0.05)

In the third assay, which compared the SDHI fungicides boscalid and fluopyram with fludioxonil, the latter again showed a superior effect. Fluopyram also significantly reduced lesion growth at 5000 and 500 conidia, whereas boscalid did not (Fig. 4). At the lowest inoculum dose, all three fungicides had a significant effect, which was strongest for fludioxonil and weakest for boscalid.

Fig. 4
figure 4

Lesion radius (mean ± SEM) on apple fruit 35 days after treatment with boscalid, fluopyram or fludioxonil followed by wound inoculation with three different macroconidium concentrations of Neonectria ditissima, in comparison to a wound-inoculated untreated control. Different letters indicate significant differences between test variants (Tukey, P < 0.05)

A calculation of the efficacies of all fungicides in terms of infection incidence at the three different conidium concentrations used for inoculation illustrated the superior activity of fludioxonil as compared to all other compounds (Fig. 5). Further, it became apparent that the efficacies of all fungicides markedly declined with increasing spore concentrations.

Fig. 5
figure 5

Efficacy of six fungicides in detached-fruit assays as a function of macroconidial wound-inoculum of Neonectria ditissima. Stalled as well as spreading infections were included in this analysis

Field Trial for Blossom-End Rot

Symptoms of blossom-end rot became visible over a long time span from mid-June to mid-August 2022. Incipient and fully developed lesions were often seen on different fruits on the same tree (Fig. 6). Such different developmental stages were observed in all artificially inoculated variants of the trial. Fruits with symptoms were sampled repeatedly between 21 July and 19 September 2022. By far the highest disease incidence was recorded on trees that had been inoculated and had been sprayed either with water or with cyprodinil at flowering. Strong differences in the incidence of blossom-end rot were found between the various fungicide treatments. Only fludioxonil gave a significant (P < 0.05) reduction as compared to the inoculated water-sprayed control (Fig. 7). The blossom-end rot control efficacies were 0% for cyprodinil, 37.6% for trifloxystrobin, and 54.9% for fludioxonil. Where trees had been sprayed with water instead of macroconidia, the natural disease incidence was very low (0.2%) as compared to the inoculated control (9.0%).

Fig. 6
figure 6

Blossom-end rot symptoms on cv. ‘Nicoter’ in the cyprodinil treatment of the field trial on 21 July 2022

Fig. 7
figure 7

Incidence of blossom-end rot on cv. ‘Nicoter’ treated at flowering with various fungicides, followed by inoculation with macroconidia of Neonectria ditissima (n = 4; 1208–1296 fruit per variant). Different letters indicate significant differences between treatments (Sidak, P < 0.05)

Discussion

In this study, in vitro sensitivities to several single-site fungicides were determined for a wide range of N. ditissima isolates obtained from IPM, organic and abandoned orchards in several localities in Germany and South Tyrol. All isolates showed the lowest EC50 values and thus the greatest sensitivity to fludioxonil, followed by trifloxystrobin. This picture was confirmed by field inoculation assays, in which fludioxonil and, to a lesser extent, trifloxystrobin provided a control of N. ditissima floral infections at dosage rates according to current pome fruit registration in Germany. Sensitivity to most of the other fungicides, especially SDHI and anilinopyrimidine products, was very low in vitro, and this corresponded to a reduced or negligible effect against wound infections in detached-fruit assays. These results are in line with previous trials which revealed no significant effect of pyrimethanil, fluopyram and thiophanate-methyl against artificial floral infections leading to blossom-end (RWS Weber and HHF Holthusen, unpublished). Unfortunately, fludioxonil was not tested in these early trials.

N. ditissima is known as a relatively crude pathogen which requires numerous germinating macroconidia or ascospores in order to establish a necrotrophic disease lesion (Xu et al. 1998; Walter et al. 2016). Low spore numbers or suboptimal conditions may lead to stalled or latent infections in which limited canker or fruit rot lesions arise or no trace of colonisation is visible, respectively (Swinburne 1971). These may or may not develop aggressive infections at a later stage. In our study we observed that the relationship between spore numbers and lesion size was influenced by several of the fungicides tested. Thus, at 500 conidia per wound the weak fungicides cyprodinil and boscalid restricted fruit rot lesions to a size comparable to 50 conidia without fungicide treatment, whereas fluopyram achieved a similar effect at 5000 conidia. Furthermore, fludioxonil and, to a lesser extent, other fungicides confined infections to the immediate vicinity of the wound so that these remained invisible from the outside. This symptom resembled infections caused by the relatively non-aggressive microconidia with their slow-growing germ tubes (Wesche and Weber 2023). In N. ditissima more so than in many other pathogenic fungi, therefore, fungicide efficacy is dependent on the infection pressure. This may go some way towards explaining the immense variations between different fruit-growing regions in assessing the suitability of different fungicides (Weber 2014; Weber and Børve 2021).

To the best of our knowledge there are no comparable previous reports of fungicide EC50 data of N. ditissima. Weber and Palm (2010) and Walter et al. (2014) conducted studies with MBCs, but these were based on mycelial plug tests, making a comparison with our data difficult. Nonetheless, our EC50 values for all isolates exceeded 2 ppm and were thus in a similar range as reported by Weber and Palm (2010). This contrasts with fully sensitive isolates of other fungal pathogens such as Botrytis cinerea (0.05–0.17 ppm; Weber and Hahn 2011) or Venturia inaequalis (0.07–0.11 ppm; RWS Weber and R Busch, unpublished) determined with the same test system as used here. In fact, the sensitivity of all N. ditissima isolates in our study and in previous studies was so low that it resembled partial MBC resistance in other pathogens such as Neofabraea spp. (Weber and Palm 2010). Since all isolates of N. ditissima fell within the same EC50 range to thiophanate-methyl, a naturally low sensitivity appears a more likely explanation than an acquired mutation-based fungicide resistance. It remains curious why this group of fungicides has been widely recommended for N. ditissima canker and fruit rot control in other growing areas in the past. Previous field trials in Northern Germany did not indicate any effect of thiophanate-methyl against infections in autumn leading to tree canker (Palm 2009). Meanwhile, MBC-type fungicides have been phased out at least in Europe, so that this puzzle may remain unsolved.

Dodine is a major current apple and pear scab fungicide in Northern Germany and elsewhere, used during heavy ascospore infections between bud break and the beginning of flowering (Palm and Kruse 2010). A protective side effect against N. ditissima has been assumed for this period. High efficacies against Neonectria fruit rot have also been reported from Great Britain (Cooke et al. 1993; Berrie 2016). For the apple scab fungus, V. inaequalis, resistant strains possessing EC50 values > 1 ppm were readily distinguished from sensitive ones with EC50 < 0.2 ppm (Köller et al. 1999; Beresford et al. 2013; RWS Weber and R Busch, unpublished). For N. ditissima we determined EC50 values at concentrations exceeding 1 ppm in all isolates but one. Nonetheless, in the fruit assay dodine significantly reduced lesion growth with a moderate efficacy resembling fluopyram, and this effect was strongly influenced by the inoculum dose.

The EC50 values for fluopyram varied widely between 1.35 and 140 ppm. The baseline for sensitive isolates of other fungal pathogens in conidial germination assays is generally much lower, e.g. at 0.02–0.20 ppm in B. cinerea (Weber et al. 2015), and 0.013–0.094 ppm in V. inaequalis (RWS Weber and R Busch, unpublished). Cross-resistance to other SDHI fungicides in Botrytis spp. is partial in the sense that it is conveyed by some but not all of the many point mutations that are associated with SDHI resistance (Amiri et al. 2014). Therefore, EC50 values for boscalid were additionally determined for 23 selected N. ditissima isolates that had low, intermediate or high fluopyram EC50 values. These were all above 500 ppm, indicating a lack of correlation between sensitivities to boscalid and fluopyram. This has also been reported previously for Botrytis spp. (Amiri et al. 2014). In addition, the lower sensitivity of N. ditissima to boscalid as compared to fluopyram ties in with reports on other fungal pathogens (Ishii et al. 2011; Amiri et al. 2014). The extraordinarily high boscalid EC50 values were consistent with the poor efficacies on artificially infected fruits. Further studies involving molecular biological analysis of target gene sequences are required to elucidate the basis of low SDHI sensitivity of N. ditissima.

In this study, EC50 values for cyprodinil were in the range of 5.04–83.66 ppm and thus appreciably higher than those known for other fungi. In general, resistance tests with anilinopyrimidines based on germ-tube growth are more difficult to evaluate than tests for other compounds because of a very gradual decline of germ-tube growth with increasing concentrations (Weber and Hahn 2011). In the taxonomically related fungus Cylindrocarpon destructans, EC50 values > 5 ppm were determined for cyprodinil (Rego et al. 2006). The low effect against N. ditissima was confirmed both by artificial fruit infections and by flower infections in the field. Similarly high EC50 values were also found for pyrimethanil in this study. Based on these results, anilinopyrimidines cannot be recommended for control of N. ditissima.

In the spore germination test, very low trifloxystrobin concentrations (EC50 approx. 0.01 ppm) are sufficient to inhibit sensitive isolates of B. cinerea (Weber and Hahn 2011) and V. inaequalis (RWS Weber and R Busch, unpublished), provided that the alternative oxidase inhibitor salicyl hydroxamic acid (SHAM) is added to the medium. Mutations at G143A of the cytochrome oxidase b gene give rise to a total loss of activity (Grasso et al. 2006), even at high concentrations of 50 ppm or more (Weber and Hahn 2011). Such resistance is extremely widespread in Northern European populations of V. inaequalis (RWS Weber and R Busch, unpublished) and B. cinerea (Nielsen et al. 2021). In the current work, EC50 values of all N. ditissima strains were below 0.5 ppm trifloxystrobin, indicating a sensitivity, albeit reduced, to the fungicide. Nevertheless, with a maximum of 40% the efficacy in the fruit assay and in the field trial was very limited, precluding the effective use of trifloxystrobin and related compounds against N. ditissima fruit rots in IPM.

Of all fungicides tested, fludioxonil was effective at the lowest inhibitory concentrations and produced the highest efficacy in our experiments. The EC50 values for all 43 N. ditissima strains were below 0.05 ppm, which is comparable to the baseline of Botrytis spp. (Weber and Hahn 2011) and V. inaequalis (Chatzidimopoulos et al. 2022). In general, the active ingredient fludioxonil is viewed as being effective against most diseases and possessing a reduced risk of resistance development (FRAC resistance group 12; Anon 2022). Similarly, fludioxonil was the most promising fungicide against C. destructans (Rego et al. 2006). However, the present results on N. ditissima indicated a declining efficacy with increasing inoculum concentrations. On the basis of disease incidence, the efficacy of fludioxonil was only 50% at the maximum inoculum dose of 5000 conidia, which was mainly due to the inclusion of stalled infections in the evaluation. It is well known that N. ditissima causes latent infections on woody tissues if a high fungicide input delays symptom expression but does not kill the infection outright (McCracken et al. 2003; Saville and Olivieri 2019). It is also known that spore inoculum dose has a critical impact on canker incidence (Xu et al. 1998; Walter et al. 2016), as it does on the development of fruit rot (this study). Accordingly, the high intrinsic efficacy of fludioxonil against N. ditissima fruit rot should be supported by canker pruning before flowering and in late summer if it is to be used for the control of blossom-end rot or storage rot, respectively. This applies even more to other compounds: the weaker the fungicide, the more important the reduction of inoculum.