E ﬀ ect of Beauveria bassiana Fungal Infection on Survival and Feeding Behavior of Pine-Tree Lappet Moth ( Dendrolimus pini L.)

: Research highlights: The pine-tree lappet moth, Dendrolimus pini , can cause serious needle defoliationonpineswithoutbreaksoccurringoverlargegeographicalareas. Under laboratory conditions, the promising potential of the naturally occurring entomopathogenic fungus Beauveria bassiana was tested against D. pini larvae as a biological control method. Background and objectives: The aim of this study was to investigate the most effective concentration and treatment dose of B. bassiana conidial suspension and how it affected the survival and feeding behavior of the pest. Materials and methods: The first experiment applied the fungal suspension directly on the back of selected larvae, and in the second experiment, sporulating cadavers obtained in the first experiment were placed into Petri dishes with healthy individuals. Different doses per larvae [ µ L] and spore suspension concentration [spores / µ L]) were used. The second experiment was designed to investigate the horizontal transmission of fungi by exposing individual caterpillars to a cadaver covered in B. bassiana mycelia. Mortality rates were analyzed by Chi-squared tests using absolute values for total mortality and B. bassiana - attributed mortality. The lethal time and feeding-disruption speed were analyzed with parametric and non-parametric tests with the aim to determine whether statistically significant differences were observed between treatments. Results: Analysis of lethal time revealed that the 20 µ L dose and 7.9 × 10 4 concentration yielded highest mortality, but also the average dieback rate of 9.4 days was significantly faster compared to all other treatments. In order to see whether insects stopped feeding after infection with B. bassiana , larvae weight loss was examined. Results showed that effective treatments induced change in the feeding behavior of infected larvae. The feeding disruption caused by the 20 µ L treatment occurred on average 5.5 days after exposure or 3.9 days before the final dieback of larvae. that application of 20 µL of 7.9 × 104 conidia/µL suspension could yield 100% mortality with 88% of B. bassiana infection, and a Chi-square test showed that both results could be effectively considered as LD100. The 5 µL and 1 µL dose had a significantly lower yield, however both yielded over 50% mortality attributed to B. bassiana infection. Mortality in the control group of the first experiment was at the level of 13% and was attributed to parasitoid infections and bacteriosis. Looking at the differences between total and B. bassiana attributed mortalities among the treatment groups, the average difference was 11%, thus indicating that differences did not diverge substantially from naturally observed mortality. Consequently, the yield of B. bassiana attributed mortalities in the treatment groups can be considered as a result of our treatments. Lethal time analysis revealed an average of 9.4 days until dieback for the 20 µL dose, which was also significantly faster compared to all other treatments, and 2.2 days faster than the control group. The 5 µL dose averaged 9.9 days until dieback and was marginally different from the control group. The 1 µL and 1:100 dilution treatments were not significantly different from the control group. In addition to lethal time, our results showed that effective treatments induced the change in feeding behavior of infected caterpillars. On average, feeding disruption occurred 5.5 days after exposure to 20 µL treatment, which was 3.88 days before the final dieback of caterpillars. Feeding disruption in the control group occurred 9.9 days after the start of the experiment. Comparing these two groups, the disruption in the feeding of caterpillars occurred on average 4.4 days sooner compared to the control group. This is a very interesting result to consider when making forest protection decisions because, not only would treatment disrupt the reproduction and population growth of D. pini through increased mortality, but it would also induce substantially faster cessation of damages inflicted by targeted caterpillars. These findings are in agreement with other studies that reported a decline in food consumption and excretion associated with fungal infections in insects. For example, reduced food intake was observed after treatment with entomopathogenic fungi B. bassiana and Metarhizium anisopliae (Metschn.) Sorokïn (Hypocreales: Clavicipitaceae)


Introduction
Pine-tree lappet moth, Dendrolimus pini L. (Lepidoptera, Lasiocampidae) is considered one of the most important harmful insect species in pine forests in Europe. This insect covers an area from Western Europe to Middle Asia (China and Asian Russia) and was recorded in North Africa, with the highest activity recorded in Central Europe [1]. Currently this forest pest is spreading to the north and south of the continent due to climate change [2]. Considering the fact that the main activity of D. pini was previously more or less concentrated in Poland and Germany [1], outbreaks in Sweden, Scotland, or Croatia are new reports [3][4][5] that could be examples of the increasing pest activity throughout Europe. Flight period of D. pini moths typically lasts from July until middle August, but is variable and temperature dependent [6]. After mating, each female moth lays their eggs (150-300 eggs) on pine

Collection of Insects
Larvae of D. pini were collected in August and September 2017 in the area of Nature Park Telašćica, Dugi Otok (43 • 53 20,9724" N, 15 • 10 9714" E) in Croatia. The first collection in August consisted of collecting 300 larvae for the first experiment and another 100 larvae were collected in September for the second experiment. Larvae from the third and fourth larval stages were selected from needles, branches, trunks, and soil, and were transferred in plastic cylinders with a cloth over the opening to the Laboratory for Entomological Analysis at the Croatian Forest Research Institute, where they were incubated overnight under laboratory conditions (L:D = 16:8, T = 23 ± 1 • C, relative humidity = 60 ± 5%).

Fungal Isolation
For preparation of spore suspension, one B. bassiana isolate was obtained from D. pini cadavers, previously collected in 2014 near Skradin, in the vicinity of Šibenik. Fungal cultures were prepared by planting pieces of mycelium on PDA (potato dextrose agar) plates and incubated for a minimum of 14 days at 25 ± 1 • C. Fully sporulated cultures were used for the preparation of suspensions.

Preparation of Spore Suspensions
Spores were harvested by adding 10 mL of Tween 80 and sterile distilled H 2 O solution (0.1%) to each Petri dish with fungal culture, and the conidia from the surface of the agar plates were gently scraped with a sterile triangular cell spreader (15 cm). After that, 1 mL of obtained spore suspension was diluted with sterile distilled H 2 O and Tween solution (0.1%) in 1:10 dilution, and was vortexed for 1 min for homogenization (V-1 Plus Vortex-Mixer). Afterward, spore suspension was filtered through four layers of sterile cheesecloth into 50 mL sterile plastic tubes (Falcon), in order to remove possible residuals of mycelium and agar pieces. Spore concentration was determined with Neubauer hemocytometer at 20× magnification under the phase-contrast microscope (Olympus, model BX53, Olympus, Tokyo, Japan). For the D group in the first experiment, one part of the suspension was separated and adjusted to the desired concentration (1:100 dilution), while the spore suspension for the other experimental groups was not diluted, only differing doses between groups.

Confirmation of Conidial Viability
Conidial viability was assessed by determining the percentage of the germinated conidia in 24 h after spreading 100 µL of conidial suspension on malt extract agar (MEA) medium. To estimate the percentage of germination, 100 conidia in three randomly chosen plate areas were examined (total of 300 conidia per plate). Conidia were considered germinated when the germ tube was longer than the conidial diameter. The dishes were incubated at 25 ± 1 • C under natural day-night regime and there were five replicates (Petri dishes).

Setting up the Experiments
From the larvae collected in the field, the healthy and vital ones were chosen, divided into groups, and separately placed into plastic Petri dishes. Two experiments were conducted. The first experiment consisted of pipetting the suspension directly on the back of the selected larvae ( Table 1). The first experiment was designed to investigate differences in mortality rates, lethal time, and quickness in feeding disruption between different concentrations and doses of spore suspension. This experiment was composed of four treatments: E1T1 (each caterpillar received 1 µL of 7.9 × 104 spores/µL suspension), E1T5 (each caterpillar received 5 µL of 7.9 × 104 spores/µL suspension), E1T20 (each caterpillar received 20 µL of 7.9 × 104 spores/µL suspension), E1T01 (each caterpillar received 1 µL of 7.9 × 102 spores/µL suspension), and E1C as a control group (each caterpillar received 20 µL of 0.1% Tween suspension). The duration of the experiment was 15 days (40 individuals per group, L:D = 16:8 h, T = 23 ± 1 • C, relative humidity = 60 ± 5%). In the second experiment, sporulating cadavers obtained in the first experiment were individually placed into a separate Petri dish with the healthy individuals ( Table 2). The second experiment was designed to investigate horizontal transmission of fungi by exposing individual caterpillars to a cadaver covered in B. bassiana mycelia (E2CAD). In each Petri dish (50 × 9 mm), there was one larva exposed to one cadaver for 30 min. A control group was also set in order to rule out the effect of naturally present infection (E2C). The duration of the experiment was 15 days (50 individuals per group, temperature = 23 • C, photoperiod L:D = 16:8 h).

Group E2CAD E2C
Treatment Healthy larvae + cadaver Only healthy larvae After treatment, larvae were checked every day for mortality and appearance of fungi mycelium on cadavers. Each larva was fed daily with fresh Aleppo pine needles and all excrement was removed from the container. In the first experiment, each larva was also weighed every day (analytical scale Acculab ATILON ATL-224-I, 220 × 0.0001 g).
During both experiments, each dead larva was placed into the moisture chamber to stimulate fungal sporulation outside of the cadaver. For confirmation of the cause of death, re-isolation of fungi on the new PDA medium was carried out from the individuals on which the white mycelium appeared, and every new culture obtained was microscopically examined. Dead individuals with either no or poorly developed mycelium were dissected to examine the interior of the body for the presence of fungi.
Experiments were designed to last for 15 days, which is considered as the threshold for acute mortality [25]. Mortality rates were determined by counting the individuals that died within the 15 days of the experiment and are represented in absolute and relative value. The mortality of individuals with symptoms of infection by B. bassiana were attributed to the entomopathogenic fungus.

Statistical Analyses
Mortality rates were represented in absolute and relative value for both total and B. bassiana attributed mortality. Abbot's formula was used in order to evaluate the treatment results with correction for mortality observed in the control group [26]. Lethal time was determined by recording consecutive day of dieback for each individual caterpillar after the start of the experiment. Likewise, feeding disruption speed was determined by recording consecutive day after the start of the experiment at which individual caterpillars stopped feeding. For purpose of feeding disruption analysis, the chrysalis forming day and day 15 were used respectively for individuals that did not die during the experiment, all with the aim to keep the whole set of observations and maintain robustness of analyses by including both infection induced and naturally induced disruption in feeding. Mortality rates were analyzed by Chi-squared tests using absolute values for both the total and B. bassiana attributed mortality in order to determine if a significant difference existed between observed mortalities and lethal dose (LD100: dose required to kill 100% of D. pini larvae). Lethal time and feeding disruption speed were analyzed by parametric (ANOVA) and non-parametric (Kruskal-Wallis) tests with the aim to determine whether statistically significant differences were observed between treatments. Tukey and Steel-Dwass-Critchlow-Fligner multiple pairwise comparison where performed based on parametric and non-parametric tests, respectively, in order to distinguish the differences between treatments.
All analyses were performed using MS Office Excel (Microsoft) and XLSTAT (Addinsoft) statistical tools add-on.

Mortality Rates
According to overall mortality, E1T20 treatment yielded 100% total mortality and 88% B. bassiana attributed mortality, while the lowest concentration and doses yielded lower mortalities (Table 3). There was 13% mortality observed in the control group, although none of the dead caterpillars in this group showed symptoms of B. bassiana infection. The Chi-square test showed that E1T20 treatment was the only one not significantly different from LD100 for both the total and B. bassiana attributed mortality. Treatment E1T5 was not significantly different in terms of total mortality, however, significant difference from LD100 could not be rejected for mortality attributed to B. bassiana infection. Two other treatments and control were significantly different from LD100 looking at both the total and B. bassiana attributed mortality (Table 3). Dose response curves depict a segment of the hyperbolic curve in I quadrant with a steady positive slope, showing positive correlation between mortality and dose increase ( Figure 1). Total mortality reached the maximum at the highest dose used in the experiment while the B. bassiana curve did not reach the maximum as it did not account for whole observed mortality. Neither curve reached a plateau. High coefficients of determination indicate high consistency in mortality increase as a result of dose increase.

Lethal Time
The lethal time results showed lowest values for 20 μL treatment, followed by 5 and 1 μL. Highest value was observed for the control group and 1:100 diluted spore solution treatment. Differences between treatment groups were significant, as determined by ANOVA (Table 4). However, the Shapiro-Wilk's test showed that residuals did not follow normal distribution (W = 0.961; p-value (Two-tailed) = 0.009), indicating that condition of validity was not met and therefore a non-parametric test was also performed on the same set of data. A Tukey multiple pairwise comparison of the differences between the categories with a confidence interval of 95% grouped treatments in two groups: Group A-treatment with 1:100 diluted spore solution and Group B-1 µ L, 5 µ L, and 20 µ L dose treatment, while the control group could not be discriminated between groups (Figure 2).

Lethal Time
The lethal time results showed lowest values for 20 µL treatment, followed by 5 and 1 µL. Highest value was observed for the control group and 1:100 diluted spore solution treatment. Differences between treatment groups were significant, as determined by ANOVA (Table 4). However, the Shapiro-Wilk's test showed that residuals did not follow normal distribution (W = 0.961; p-value (Two-tailed) = 0.009), indicating that condition of validity was not met and therefore a non-parametric test was also performed on the same set of data. A Tukey multiple pairwise comparison of the differences between the categories with a confidence interval of 95% grouped treatments in two groups: Group A-treatment with 1:100 diluted spore solution and Group B-1 µL, 5 µL, and 20 µL dose treatment, while the control group could not be discriminated between groups (Figure 2).
The non-parametric test also showed significant difference in lethal time between treatments in experiment 1, and multiple pairwise comparison using Steel-Dwass-Critchlow-Fligner procedure resulted in three groups. According to this analysis, the 20 µL treatment was clearly designated to Group A, the 5 µL and 1 µL treatments could not be discriminated between Groups A and B, the control could not be discriminated between Groups B and C, while the treatment with 1:100 diluted solution was designated to Group C. The differences between treatments are clearly visible in the Box-Whiskers histograms (Figure 3). The non-parametric test also showed significant difference in lethal time between treatments in experiment 1, and multiple pairwise comparison using Steel-Dwass-Critchlow-Fligner procedure resulted in three groups. According to this analysis, the 20 µ L treatment was clearly designated to Group A, the 5 µ L and 1 µ L treatments could not be discriminated between Groups A and B, the control could not be discriminated between Groups B and C, while the treatment with 1:100 diluted solution was designated to Group C. The differences between treatments are clearly visible in the Box-Whiskers histograms (Figure 3).  The non-parametric test also showed significant difference in lethal time between treatments in experiment 1, and multiple pairwise comparison using Steel-Dwass-Critchlow-Fligner procedure resulted in three groups. According to this analysis, the 20 µ L treatment was clearly designated to Group A, the 5 µ L and 1 µ L treatments could not be discriminated between Groups A and B, the control could not be discriminated between Groups B and C, while the treatment with 1:100 diluted solution was designated to Group C. The differences between treatments are clearly visible in the Box-Whiskers histograms ( Figure 3).  Twenty µL treatment showed the lowest mean and dissipation of data, followed by 5 µL and 1 µL, while the 1:100 dilution showed highest value, even higher than the control group. Cross represents the arithmetic mean, central line of the box represents the median, upper line of the box represents the third quartile, lower line of the box represents the first quartile, outreaching lines represent standard deviations, dots represent outlayers.

Feeding Disruption Speed
The feeding disruption speed results also showed lowest values for the 20 µL treatment, followed by 5 and 1 µL. The highest value was observed for the control group and 1:100 diluted spore solution treatment. Differences between treatment groups were significant, as determined by ANOVA. However, the Shapiro-Wilk's test showed that residuals did not follow a normal distribution (W = 0.977; p-value (Two-tailed) = 0.002), indicating that the condition of validity was not met and therefore a non-parametric test was also performed on the same set of data.
A Tukey multiple pairwise-comparison of the differences between the categories with a confidence interval of 95% grouped treatments in three groups; the treatment with 1:100 diluted spore solution was designated to Group A, the control could not be discriminated between groups A and B, the 1 µL and 5 µL could not be discriminated between Groups B and C; and the 20 µL treatment was designated to Group C (Figure 4).
The feeding disruption speed results also showed lowest values for the 20 μL treatment, followed by 5 and 1 µ L. The highest value was observed for the control group and 1:100 diluted spore solution treatment. Differences between treatment groups were significant, as determined by ANOVA. However, the Shapiro-Wilk's test showed that residuals did not follow a normal distribution (W = 0.977; p-value (Two-tailed) = 0.002), indicating that the condition of validity was not met and therefore a non-parametric test was also performed on the same set of data.
A Tukey multiple pairwise-comparison of the differences between the categories with a confidence interval of 95% grouped treatments in three groups; the treatment with 1:100 diluted spore solution was designated to Group A, the control could not be discriminated between groups A and B, the 1 µ L and 5 µ L could not be discriminated between Groups B and C; and the 20 µ L treatment was designated to Group C (Figure 4). The non-parametric test also showed a significant difference in lethal time between treatments in experiment 1, and multiple pairwise comparison using the Steel-Dwass-Critchlow-Fligner procedure resulted in three groups. According to this analysis, the 20 µ L treatment was clearly designated to Group A, the 5 µ L and 1 µ L treatments could not be discriminated between Groups A and B, the control could not be discriminated between Groups B and C, while the treatment with 1:100 diluted solution was clearly designated into Group C. The differences between treatments are clearly visible in the Box-Whiskers histograms ( Figure 5). The non-parametric test also showed a significant difference in lethal time between treatments in experiment 1, and multiple pairwise comparison using the Steel-Dwass-Critchlow-Fligner procedure resulted in three groups. According to this analysis, the 20 µL treatment was clearly designated to Group A, the 5 µL and 1 µL treatments could not be discriminated between Groups A and B, the control could not be discriminated between Groups B and C, while the treatment with 1:100 diluted solution was clearly designated into Group C. The differences between treatments are clearly visible in the Box-Whiskers histograms ( Figure 5).

Mortality Rates
Treatment by exposure to cadavers yielded 85% total mortality and 67% B. bassiana attributed mortality, while there was 30% total and 5% fungus attributed mortality observed in the control  Figure 5. Box plot histograms of feeding disruption speeds observed for different treatments in experiment 1. The 20 µL treatment showed the lowest mean and dissipation of data, followed by 5 µL and 1 µL, while the 1:100 dilution showed the highest value, even higher than the control group. Cross represents the arithmetic mean, the central line of the box represents the median, the upper line of the box represents the third quartile, the lower line of the box represents the first quartile, outreaching lines represent standard deviations, dots represent outlayers.

Mortality Rates
Treatment by exposure to cadavers yielded 85% total mortality and 67% B. bassiana attributed mortality, while there was 30% total and 5% fungus attributed mortality observed in the control group and thus Abbott's formula was applied to calculate the corrected relative mortality for treated group (Table 5). Table 5. Total and B. bassiana attributed absolute and relative mortality rates for treatments in experiment set two and corresponding results of the Chi-square tests. Values that are not significantly different from LD100 are highlighted in bold. Abbott's formula was used to calculate relative mortality for thee treated group (E2CAD) due to mortality observed in the control group (E2C). The Chi-square test showed that cadaver exposure treatment (E2CAD) total mortality rates were not significantly different from LD100, while mortality attributed to fungus infection was significantly different from LD100 even though it reached 67%. Mortality rates in the control group were significantly different from LD100 looking at both the total and B. bassiana attributed mortality ( Figure 6).

Discussion
This study was performed with the aim to investigate the potential of B. bassiana isolates for control of D. pini. We compared the susceptibility of D. pini larvae to different doses and concentrations of spore suspensions and revealed that application of 20 µ L of 7.9 × 104 conidia/µ L suspension could yield 100% mortality with 88% of B. bassiana infection, and a Chi-square test showed that both results could be effectively considered as LD100. The 5 µ L and 1 µ L dose had a significantly lower yield, however both yielded over 50% mortality attributed to B. bassiana infection. Mortality in the control group of the first experiment was at the level of 13% and was attributed to parasitoid infections and bacteriosis. Looking at the differences between total and B. bassiana attributed mortalities among the treatment groups, the average difference was 11%, thus indicating that differences did not diverge substantially from naturally observed mortality. Consequently, the yield of B. bassiana attributed mortalities in the treatment groups can be considered as a result of our

Discussions
This study was performed with the aim to investigate the potential of B. bassiana isolates for control of D. pini. We compared the susceptibility of D. pini larvae to different doses and concentrations of spore suspensions and revealed that application of 20 µL of 7.9 × 104 conidia/µL suspension could yield 100% mortality with 88% of B. bassiana infection, and a Chi-square test showed that both results could be effectively considered as LD100. The 5 µL and 1 µL dose had a significantly lower yield, however both yielded over 50% mortality attributed to B. bassiana infection. Mortality in the control group of the first experiment was at the level of 13% and was attributed to parasitoid infections and bacteriosis. Looking at the differences between total and B. bassiana attributed mortalities among the treatment groups, the average difference was 11%, thus indicating that differences did not diverge substantially from naturally observed mortality. Consequently, the yield of B. bassiana attributed mortalities in the treatment groups can be considered as a result of our treatments.
Lethal time analysis revealed an average of 9.4 days until dieback for the 20 µL dose, which was also significantly faster compared to all other treatments, and 2.2 days faster than the control group. The 5 µL dose averaged 9.9 days until dieback and was marginally different from the control group. The 1 µL and 1:100 dilution treatments were not significantly different from the control group.
In addition to lethal time, our results showed that effective treatments induced the change in feeding behavior of infected caterpillars. On average, feeding disruption occurred 5.5 days after exposure to 20 µL treatment, which was 3.88 days before the final dieback of caterpillars. Feeding disruption in the control group occurred 9.9 days after the start of the experiment. Comparing these two groups, the disruption in the feeding of caterpillars occurred on average 4.4 days sooner compared to the control group. This is a very interesting result to consider when making forest protection decisions because, not only would treatment disrupt the reproduction and population growth of D. pini through increased mortality, but it would also induce substantially faster cessation of damages inflicted by targeted caterpillars. These findings are in agreement with other studies that reported a decline in food consumption and excretion associated with fungal infections in insects. For example, reduced food intake was observed after treatment with entomopathogenic fungi B. bassiana and Metarhizium anisopliae (Metschn.) Sorokïn (Hypocreales: Clavicipitaceae) in maize stem borer Chilo partellus Swinhoe (Lepidoptera: Crambidae) larvae [27] and in malaria vector mosquitoes [28,29]. Significant reduction in feeding was also demonstrated in green peach aphid Myzus persicae Sulzer (Homoptera: Aphididae) mycosed with entomopathogenic fungus Lecanicillium longisporum (Petch) Zare & W. Gams (Hypocreales: Cordycipitaceae) [30] as well as in desert locust Schistocerca gregaria Forskål (Orthoptera: Acrididae) [4] and variegated grasshopper Zonocerus variegatus L. (Orthoptera: Pyrgomorphidae) [31], both infected with Metarhizium flavoviride Gams and Rozsypal (Hypocreales: Clavicipitaceae).
Looking back at the mortality rates, the relative increase in mortality as a result of dose increase was 2.73% per µL for 5 µL dose and 2.95% per µL for 20 µL dose, all compared to the 1 µL dose. This finding showed that increase in dose to 20 µL yielded higher relative increase in mortality than the 5 µL dose, a result that further bolsters the conclusion that an increase in dose to 20 µL is justified in order to reach significantly higher mortality.
Experiment two resulted in high mortality rate among caterpillars exposed to cadavers covered in mycelia of B. bassiana, demonstrating the ability of horizontal transmission by the fungus. Mortality among caterpillars exposed to cadavers was rather high at 85%, with 67% attributed to fungus infection, while there was 30% total mortality in the control group, 5% of which was attributed to fungus infection. These were 80% and 65%, respectively, when Abbott's correction was applied due to mortality observed in the control group.
Whether this was an artifact of contamination or a naturally present infection remains unclear. However, the rate observed in the treatment group was significantly greater, thus rendering the results indisputably meaningful. For example, it was demonstrated that horizontal transmission of the B. bassiana infection to healthy Diaphorina citri Kuwayama (Hemiptera: Liviidae) adults [32] and spruce bark beetle Ips typographus L. (Coleoptera: Curculionidae) [33] yielded high mortality rates, indicating that microbial control can be subsequently promoted through conidial cycling (i.e., by generating new infection cycles). In Drummond and Groden [34], infection from B. bassiana sporulating cadavers resulted in significant reduction of a Colorado potato beetle Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae) population in years following treatment. Horizontal transmission enables self-sustainment and spread of the biocontrol agent, a trait of major importance in light of modern approaches in integrated forest protection.

Conclusions
In conclusion, biological methods of controlling D. pini by using entomopathogenic B. bassiana as a natural enemy seems to represent an environmentally acceptable solution that could replace, or at least reduce, the use of chemical insecticides. These laboratory studies should be extended to the field in order to determine the impact of tested concentration and doses of B. bassiana suspensions on D. pini in natural conditions.