Colonies fed with HFCS had more brood cells during the study than those fed with SS (p=0.048; Table 1; Fig. 3). Also, throughout the assay, T+ colonies had less adult honey bees (p=0.011; Table 1; Fig. 3) and a lower Varroa infestation percentage than T- colonies (p< 0.001; Table 1; Fig. 4). No differences on N. ceranae abundance and nutritional reserves (pollen and honey cells) between HFCS and SS colonies nor T+ and T- were detected (Table 1; Fig. 5). No significant effect of the protein supplementation during the course of the study on any of the response variables was observed (Table 1). Within T+ colonies, the 2016 mite invasion rates tend to be higher in SS colonies (mean=30.63 ± 6.20 mites/week) than HFCS colonies (20.08 ± 4.03 mites/week; p=0.074, Table 1). During 2017, SS colonies from all groups had significantly more mites drop counts than HFCS colonies (278.27 ± 53.39 and 59.16 ± 12.78 mites/week, respectively; p<0.001) (Table 1). By means of acaricide effect, during 2016 T+ colonies had more mites drop counts (mean=24.80 ± 3.53 mites/week) than T- colonies (mean=15.38 ± 2.46 mites/week; p=0.026) (Table 1).

— Short term effects on Varroa destructor. During May 2016 (after acaricide treatment), T+ colonies had less Varroa infestation level (0.91 ± 0.16) than T- colonies (2.51 ± 0.44; p<0.001) (Table 2). At the same moment, P- colonies had more Varroa infestation level (1.93 ± 0.34) than P+ colonies (1.18 ± 0.20; p=0.05; Table 2). In June, the joint effect of acaricide treatment, protein supplementation and kind of syrup was significant (p=0.027). The highest Varroa infestation level was recorded in SS T-/P+ colonies (3.35 ± 1.32; p=0.011) (Table 2; Fig. 4) while the lowest Varroa infestation level was recorded in HFCS T+P- (0.01 ± 0.003; p=0.05).

— Mid-term effects on Varroa destructor. During spring (September 2016), the joint effect of acaricide treatment, protein supplementation and kind of syrup was significant (p=0.016). The highest Varroa infestation level was observed in SS T-/P-colonies (3.25 ± 1.20; 2.58 ± 0.9, respectively) (Table 2; Fig. 4) while the lowest Varroa infestation level was recorded in HFCS T+P- (0.19 ± 0.07; p=0.03).

— Long term effects on Varroa destructor. Differences between T+ and T- colonies deepened in November (1.44 ± 0.26; 5.38 ± 0.94, respectively) (p<0.001; Table 2). Similarly, SS colonies from all groups showed higher Varroa infestation level (3.59 ± 0.66) than HFCS colonies (2.15 ± 0.37; p=0.042; Table 2). However, unlike previous sampling times, no interaction effect was found. During February 2017 (before final acaricide treatment) the differences between T+ and T- colonies became more remarkable (6.04 ± 1.05 and 10.6 ± 1.79, respectively; p=0.02; Table 2).

— Nosema ceranae. N. ceranae abundance was similar between all groups all over the year (repeated measures, Table 1). However, during February 2017 (before the final acaricide treatment) the effect of the interaction between acaricide treatment and protein supplementation on the N. ceranae abundance was significant (p=0.007, Fig. 6). For instance, T-/P- colonies had the highest N. ceranae. abundance (2.02 ± 0.67 Log₁₀ spores/bee) and T+/P+ colonies had the lowest N. ceranae abundance (0.18 ± 0.06 Log₁₀ spores/bee). T+/P- colonies had less N. ceranae abundance (mean=1.39 ± 0.46 Log₁₀ spores/bee) than T-/P+ colonies (1.55 ± 0.5 Log₁₀ spores/bee) (Fig. 6).

Honey production was similar between colonies fed with SS and HFCS (p=0.647), between T+ and T- colonies (p=0.991) and between P+ and P- colonies (p=0.709). Furthermore, no double or triple significant interaction were found.

On the contrary, the number of harvested brood frames during October was higher in the HFCS colonies (2.31 ± 0.31) compared to SS colonies (1.49 ± 0.2) (p=0.022) and was higher in the in T-/P+ (2.96 ± 0.56) compared to T+/P+ colonies (1.12 ± 0.21; p=0.012).

From the 38 colonies computed as death at the end of field trial (total losses), 23 were registered between May and September (winter mortality). For total losses, overall comparisons between all groups were non-significant (Chi-square=0.673; p=0.879). Similarly, there were no significant differences in winter mortality when all groups were compared (p=0.725). However, there was a significant effect of the carbohydrate feeding and acaricide treatment on the OI. The highest value was registered in the HFCS/T- colonies (OI=1.16 ± 0.054) and the lowest value was registered in SS/T+ colonies (OI=0.65±0.054; p=0.03).

Feeding the colonies with HFCS during autumn had an effect on population dynamics, particularly on brood cells availability during the year. These results are different from previously reported studies where bees provided with sucrose syrup produce higher numbers of brood cells (Neupane & Thapa, 2005; Sammataro & Weiss, 2013) or no differences were found (Lu et al., 2014). However, methodology is not accurately comparable, as for instance Neupane & Thapa (2005) calculated the number of brood cells including cells having egg or larva or pupa and here only sealed brood cells were counted. During autumn 2016 and 2017 mite invasion was lower in the HFCS fed colonies probably because it was proved to be less attractive to bees than SS in field studies (Sammataro & Weiss, 2013) and laboratory studies (Neupane & Thapa, 2005). Varroa mites horizontal transmission might varied as a function of mite invasion (Fries & Camazine, 2001) driven by several factors including foraging behavior mediated by sucrose responsiveness (Kuszewska et al., 2019). A reduced invasion pressure by means of robbing (Greatti et al., 1992) or drift (Goodwin et al., 2006) at late autumn might be essential to avoid winter losses (Frey & Ronsenkranz, 2014). Thus, providing HFCS as carbohydrate supply might decrease the “robbing” behavior among the colonies and consequently the mite invasion pressure.

High efficacy of the acaricide products applied was expected but opportunely checked. Still, treated colonies showed a reduced colony size during the entire year. The results presented suggested that regardless of the benefits, the same acaricide could shorten the adult honey bee lifespan (Wu et al., 2011) especially oxalic acid application (Martin Hernández et al., 2007; Schneider et al., 2012). Moreover, resistance development (Lodesani and Costa, 2005) and residues of varroacides in bee products (Bogdanov, 2006; Le Conte et al., 2010) should also discourage its prolonged use. This result is different from some previous studies that suggested that exposure to miticides does not affect bee populations at colony level (Berry et al., 2013; Rangel & Tarpy, 2016).

Pollen patties supplementation had neither global effect nor specific effect at any sampling time on colony strength and nutritional reserves. Similar to reported in DeGrandi-Hoffman et al. (2016) nutritional differences not necessarily translate into population sizes differences. Also, Van Dooremalen et al. (2013) demonstrated that abundant pollen cannot compensate for damage caused by Varroa parasitism as mite presence might reduce the benefits of nutritional resources (DeGrandi-Hoffman & Chen, 2015). Another even more feasible explanation is that pollen patties should have been heavier and more frequently administered in order to achieve a significant impact at colony level (Branchiccela et al., 2019). In addition, even when the percentage of crude protein was within the recommended range of concentration (Herbert et al., 1977) it might be lower than optimal amount of protein (van der Steen, 2007) or similar to low protein-content pollen (Basualdo et al., 2013).

Overall, there was no effect of acaricide treatment and nutritional management on N. ceranae abundance and nutritional reserves (pollen and honey cells) probably because they fluctuate seasonally and are strongly influenced by the environment conditions (Alaux et al., 2010; Fries, 2010; Di Pasquale et al., 2016). However, at the end of the study, it was observed that T- colonies and P- colonies showed the highest N. ceranae abundance. T- colonies showed significantly more Varroa than T+ colonies in November and February sampling times what could explain the substantial increase of the N. ceranae spores counts by the end of the study. Interactions between N. ceranae and V. destructor mites (Pacini et al., 2016; Giacobino et al., 2018) and among both pathogens with chemical treatments were previously reported (Little et al., 2016). It seemed that Varroa-controlled colonies and better-nourished honey bees during autumn results in better general sanitary conditions the following honey yield season. This is important since the nutritional status of the colonies is closely related to a negative impact of N. ceranae infection (Branchiccela et al., 2019).

Acaricide treatment and nutritional supplementation impact on short term Varroa loads as infestation level was lower in T+ and P+ colonies during May 2016 and in HFCS T+/P- colonies during June 2016. High efficacy of the acaricide is revealed as T+ colonies begin winter season with significantly less Varroa mites. This is important since lowest winter losses were associated to varrocides use comparing among several Varroa control methods (Haber et al., 2019). Additionally, carbohydrate feeding might enhance this positive effect by reducing mite invasion during early autumn as it was observed in HFCS fed colonies. During autumn, mite invasion and reproduction are key topic for beekeepers, especially those who treated their colonies earlier in the season (Frey & Rosenkranz, 2014). A short-term effect of pollen patties was difficult to recognize in the observed pathogen levels, as P+ colonies had less Varroa mites in May but had higher infestation levels during June. Different from expected, pollen supplementation did not enhance chemical treatment effectiveness as T+/P- had the lowest mites. Benefits of pollen diet might be limited by Varroa infestation (DeGrandi-Hoffman & Chen, 2015). Nevertheless, as the weight and frequency of supplementation should be revised, further studies should be conducted in order to elucidate this effect.

Similar to what was observed before winter (June 2016) infestation level was lower in HFCS T+/P- colonies after winter season (September 2016). This is a key moment in the beekeeping cycle as the colonies start to grow and benefit from spring forage, but also because Varroa reproduction starts to speed up as bee brood became available (DeGrandi-Hoffman & Chen, 2015). Beginning honey yield season with a low infestation level almost certainly guarantee a healthy condition in the colonies up until the end of the season avoiding additional intervention (Lodesani et al., 2019).

During November 2016 (late spring) Varroa infestation was lower in HFCS fed colonies compared to SS fed colonies. However, as exponential growth occurs in both adult bees and mites (DeGrandi-Hoffman & Chen, 2015) this slightly (but significantly) difference disappeared during summer. On the contrary, T+ colonies went through the honey yield season with a significantly lower number of mites in adult bees (from November 2016 to February 2017). In late summer (February 2017) differences in Varroa infestation level were only explained by the treatment application during autumn 2016. The nutritional management seems to have a short or mid-term impact on the colony dynamics and pathogen loads (Branchiccela et al., 2019) whereas the acaricide treatment effect was observed throughout the year, resulting in a long-term positive effect.

Although Varroa infestation levels were remarkable different in November 2016 (late spring) and February 2017 (late summer), honey production was similar between T+ and T- colonies and between SS and HFCS colonies. This was unexpected but consistent with the fact that colony size was similar between all groups from September on and that previous results suggested that colonies fed with Sucrose or HFCS showed similar honey production among these diets (Severson & Erickson, 1984). As discussed previously here, the HFCS colonies showed significantly more brood during the entire study, so the amount of harvested brood during October was higher in this group than in SS colonies. After that, brood cells were similar between groups. Similarly, more frames fully covered with worker brood were gathered from T- colonies consistent with the reported side effect of chemical treatment on brood (Berry et al., 2013; Zhu et al., 2014; Terpin et al., 2019) and with the fact that treated colonies showed reduced bee population during the entire study. Differences registered in the overwintering index between HFCS/T- and SS/T+ colonies were probably a consequence of those HFCS colonies had more brood cells and T- colonies had more bee population during the entire year.

Two possible hypotheses were raised given that, surprisingly, total losses and winter mortality were similar between all groups. The first hypothesis is that concerning treated and non-treated colonies, the time frame of the study was not enough to observe the negative effect of the presence of Varroa and so it is necessary to study the colonies for longer periods. Varroa infestation impact may be undercover as colonies were headed by young queens with a probable higher resilience (Straub et al., 2015). In addition, is quite difficult to differentiate the negative effect of Varroa presence from the use of acaricide side effects (Boncristiani et al., 2012). Regarding nutritional management, recently a supplemental forage study in overwintering colonies showed that colony losses were not necessarily linked to pollen or nectar scarcity (Carroll et al., 2018). There are multiple drivers associated with colony losses including the acaricide treatment timing (Beyer et al., 2018) as well as the improvement of the best management practices (El Agrebi et al., 2021) and queen replacement (Gray et al., 2020). The second and more likely hypothesis is that all colonies were exposed to a climatic stress factor that could had veiled the nutritional and treatment effect as autumn 2016 was unusual rainy. Previously it has been observed that environmental background conditioned the impact of the management practices on the colony losses (Molineri et al., 2018). Honey bee colonies are frequently exposed to multiple stressors, so trying isolating the relevant variables is challenging (Boncristiani et al., 2012).