The Scarcity of Specific Nutrients in Wild Bee Larval Food Negatively Influences Certain Life History Traits

Simple Summary Pollen comprises many organic substances (sugars, lipids, proteins, amino acids, vitamins, etc.), all of which are built from elements such as carbon, hydrogen, oxygen, nitrogen, phosphorus, sodium, potassium, zinc, and approximately twenty others. These special nutritional elements compose the cells, tissues, and bodies of all the life forms on our planet and are needed by bee larvae for healthy growth. However, not all plants produce pollen containing these elements in proportions needed specifically by bees, meaning that not all pollens are nutritionally balanced for bees. Moreover, the decrease in plant diversity is thought to be among the main causes of the dwindling numbers of pollinators worldwide. Currently, governments and societies are attempting to combat this pollinator decline by providing nutritionally balanced and diverse food plants to pollinators. Knowing which nutritional elements are crucial for the bee diet and understanding why are prerequisites for tailoring conservation efforts for this group of insects, which are substantially important for human nutrition and ecosystem functioning. Basic information obtained from feeding experiments is important for synergistically understanding how plant diversity within certain species that produce pollens with rich or scarce amounts of certain nutritional elements influences bee health and prosperity. Abstract Bee nutrition studies have focused on food quantity rather than quality, and on details of bee biology rather than on the functioning of bees in ecosystems. Ecological stoichiometry has been proposed for studies on bee nutritional ecology as an ecosystem-oriented approach complementary to traditional approaches. It uses atomic ratios of chemical elements in foods and organisms as metrics to ask ecological questions. However, information is needed on the fitness effects of nutritional mismatches between bee demand and the supply of specific elements in food. We performed the first laboratory feeding experiment on the wild bee Osmia bicornis, investigating the impact of Na, K, and Zn scarcity in larval food on fitness-related life history traits (mortality, cocoon development, and imago body mass). We showed that bee fitness is shaped by chemical element availability in larval food; this effect may be sex-specific, where Na might influence female body mass, while Zn influences male mortality and body mass, and the trade-off between K allocation in cocoons and adults may influence cocoon and body development. These results elucidate the nutritional mechanisms underlying the nutritional ecology, behavioral ecology, and population functioning of bees within the context of nutrient cycling in the food web.

Using a feeding experiment, we determined the fitness-related life history traits (mortality, developmental cocoon stage, and dry mass of the developed adult body) of O. bicornis after exposure to control and nutrient-deficient pollen from the juvenile (three-day larvae) to the adult (imago) stage. In addition, individuals were exposed to pollen supplemented with K, Na, or Zn to confirm or exclude the effects of scarcity of a given element.

Model Organism
Osmia bicornis (O. rufa, Hymenoptera: Megachilidae) wild bees were obtained from a nest trap assembled with ca. 500 empty Phragmites sp. stems (250-300 mm in length; 6-10 mm in diameter) in the form of a case. The nest was located in the vicinity of the Institute of Environmental Sciences, Jagiellonian University, Kraków, Poland (50° 01′ 35″ N; 19° 54′ 05″ E). Female O. bicornis constructed their nests in the cane stems.
The nesting biology of the bee is shown in Figure 1 and was previously described in detail by Filipiak [10]. Usually, female eggs are laid first; therefore, they are located in the rear part of the nest, whereas male eggs can be found near the entrance [8,33]. In early spring, once females started to construct their nests, the stems were checked daily for the presence of closed brood cells. Firstly, stems (N = ca. 250) with ca. 1-3 closed brood cells were collected to obtain female larvae, and when the bees closed the tubes with mud, more stems (N = ca. 250) were collected to obtain male larvae.
All stems were kept at 21 °C and 60% relative humidity (RH) under a 12:12 (L:D)-h photoperiod. Only specimens from the first (females) and last (males) brood cells within each stem were collected for the experiment. Due to the fragility and sensitivity of eggs and possible mechanical damage to the eggs during transfer to experimental containers, 3-day-old larvae were used for the experiment.

Experimental Design
A feeding experiment was designed to determine fitness-related life history traits (mortality, developmental stage of the cocoon, and dry mass of the developed adult body) of solitary bee (O. All stems were kept at 21 • C and 60% relative humidity (RH) under a 12:12 (L:D)-h photoperiod. Only specimens from the first (females) and last (males) brood cells within each stem were collected for the experiment. Due to the fragility and sensitivity of eggs and possible mechanical damage to the eggs during transfer to experimental containers, 3-day-old larvae were used for the experiment.

Experimental Design
A feeding experiment was designed to determine fitness-related life history traits (mortality, developmental stage of the cocoon, and dry mass of the developed adult body) of solitary bee (O. bicornis) larvae fed food (pollen) nutritionally balanced or scarce in specific nutrients (physiologically important chemical elements: K, Na, and Zn). Fitness-related life history traits were chosen for study because: (1) mortality is an evident and relevant trait; (2) cocoons are fitness-enhancing secretions that protect bees for approximately ten months of pre-and overwintering [34,35]; and (3) body mass is positively related to fitness in O. bicornis and other solitary bee females but not males [36,37]. The three studied life history traits are considered separate and competing "sinks" into which organisms allocate resources from the available pool (see, e.g., [38] for more information). For ecological relevance and to make our experimental results relatable to the natural world, we analyzed and discussed the data obtained, focusing on the relevance of the studied traits for bee fitness.
Fifteen replicates (Eppendorf tubes, 2 mL) were prepared per treatment and sex and filled with homogenized pollen of specific nutritional quality. The amount of pollen corresponded to the dry masses of pollen provisions found in nature, i.e., 195 ± 5 mg dm for females and 140 ± 5 mg dm for males. Dry pollen loads were complemented with either demineralized water or salt solutions (to reach concentrations of the studied elements found in Osmia-collected pollen) in an amount reflecting ca. 25% of the dry pollen mass. Before starting the experiment, the Eppendorf tubes were left for 24 h to allow the water and salt solutions (KCl, NaCl, and ZnCl 2 ) to penetrate the pollen loads. The three-day-old larvae were assigned to treatments with one individual per Eppendorf tube. All experimental tubes were kept at 21 • C and 60% RH under a 12:12 (L:D)-h photoperiod for 3 months. The exposure period was long enough to ensure that all larvae had reached the adulthood stage of the life cycle, i.e., the stage where fully developed individuals hibernated in their cocoons [8,35]. At the end of the exposure periods, cocoons were collected to determine the degree of cocoon development (in the case of undeveloped individuals who died as larvae and did not reach maturity, cocoons were not available). Then, the bees were extracted from cocoons, and the mortality rate was assessed. Afterwards, the individuals and cocoons were dried using a vacuum dryer (80 • C, 48 h) to obtain the dry mass.

Pollen Diets
Polyfloral pollen mixtures characterized by differing nutritional qualities expressed using the concentrations of the studied elements were used for the feeding experiment. Pollen mixtures were obtained from either O. bicornis provisions collected manually from brood cells or derived from commercially available polyfloral pollen pellets collected by honeybees (Apis mellifera) in central Europe. The pollen collected by O. bicornis in the field was considered a balanced diet, providing the needed amounts and proportions of nutrients to the bee larvae during development, and was used as a control diet in the experiment (Control-Osmia as described below). For practical reasons, only honeybee-collected pollen pellets could be used as diets depleted of certain nutrients in the experiments; therefore, we used an additional control diet (Control-Apis as described below), i.e., a diet similar to Control-Osmia in nutritional quality but composed of honeybee-collected pollen pellets.
Five packs of honeybee pollen pellets were purchased from different manufacturers and were composed of pollen of various botanical origins. According to the method proposed by Filipiak and colleagues [39], pollen from each pack was divided according to color by the naked eye to obtain pollen pellet pools with specific elemental compositions. Additionally, unsorted pools of pollen pellets were retained from each pack. In total, we obtained 15 different pollen pools from all packs: 5 unsorted and 10 sorted pools.
The concentrations of K, Na, and Zn were determined in all purchased pollen pools (unsorted and sorted) and in O. bicornis provisions. From all pollen pools, we chose the following pools for use in the feeding experiment (their nutritional qualities are given in the "Results" section): (1) control pollen from O. bicornis provisions, i.e., natural larval food, designated Control-Osmia; (2) control honeybee pollen, i.e., the unsorted pollen pellets obtained from one of the packs, which had a chemical composition similar to that of the O. bicornis provisions, designated Control-Apis; (3) three sorted honeybee pollen pellets pools with the lowest concentration of one of the studied elements (Na, K, or Zn), designated Na-deficient, K-deficient, and Zn-deficient; and (4) the same three sorted honeybee pollen pellets pools with the lowest concentration of one of the studied elements (Na, K, or Zn) and supplemented with salt containing the deficient element to reach the same concentration found in Control-Osmia, designated Na + supplemented, K + supplemented, and Zn + supplemented. The pollen pools from each treatment were homogenized manually to obtain a homogenous powder and then freeze dried to obtain the dry mass (dm).

Chemical Analysis
To analyze the K, Na, and Zn concentrations, freeze-dried pollen homogenates (five per treatment) were digested on a hotplate in a 4:1 mixture of nitric acid (70%) and hydrogen peroxide (30%). The K, Na, and Zn concentrations were measured using atomic absorption spectrometry (PerkinElmer AAnalyst 200 and PerkinElmer AAnalyst 800) and expressed in ppm dm. To determine the analytical precision, certified reference materials (bush, NCS DC 73349; chicken, NCS ZC 73016; and bovine muscle powder, RM8415) were examined with the samples.

Data Handling and Statistical Analysis
Differences in mortality between the nutritionally deficient and supplemented groups for each element separately, i.e., K-deficient vs. K + supplemented, Na-deficient vs. Na + supplemented, and Zn-deficient vs. Zn + supplemented, as well as between Control-Apis and the other groups were assessed using the Chi-squared test, with Yates's correction for one degree of freedom.
The distribution of body mass was checked for normality with Shapiro-Wilk's W test, and the homogeneity of variances was checked with Levene's test. If the criteria were not met, the data were either log-or square root-transformed, and if these steps failed, a nonparametric (Kruskal-Wallis) test was used. Only individuals who survived until the end of the experiment, i.e., those that had undergone the entire metamorphosis to the imago form, were considered in the analyses of body mass and cocoons. The effects of the treatments on body mass were tested using the Kruskal-Wallis test.
For adult (imago) bees, the degree of cocoon development was assessed by qualitative analysis, and two stages of cocoon development were distinguished. (1) The first stage was an underdeveloped cocoon that covered only part of the bee or not at all, and the cocoon consisted almost exclusively of soft ("wooly") fragments; the cocoon tore easily with bare hands but was impossible to cut using a knife because it was too soft. (2) The second stage was an almost fully or fully developed cocoon that covered the whole bee body and mainly consisted of a hard material; the cocoon was difficult or impossible to tear by bare hands but was possible to cut using a knife because it was sufficiently hard ( Figure 2). The second stage of cocoon development might have the greatest probability of allowing an adult individual to overwinter until the next season and protecting the bee from external factors (e.g., parasites or pathogens). The differences in reaching the second cocoon developmental stage between the deficient and supplemented treatments (each element separately), and between Control-Apis and the other groups were assessed using the Chi-squared test, with Yates's correction for one degree of freedom. All analyses were performed separately for females and males. Biology 2020, 9, x FOR PEER REVIEW 6 of 17 To calculate the percentages of each type of cocoon developed by the bees we considered all 15 bee specimens as 100%, constituting every treatment and control. Therefore, this 100% consisted of the sum of (1) specimens that reached the adult stage and developed to the first cocoon stage, (2) specimens that reached the adult stage and developed to the second cocoon stage and (3) specimens that died before reaching the adult stage (the majority of which did not reach the last larval stage, i.e., spinning larvae that start to produce the cocoon). Organismal death before reaching maturity has obvious negative consequences for fitness, therefore we separated these specimens in our analysis from those that successfully reached maturity and developed cocoons to make our analysis ecologically relevant. In this way, we treated the cocoon developmental stage as an ecologically meaningful trait that influences the fitness of living and mature bees. This analysis was complemented by a simultaneous redundancy analysis (RDA) of the datasets on body mass and cocoon stages performed in Canoco 5 [40], which helped us to determine whether the negative effects of nutrient scarcity on these two life history traits (1) were correlated and (2) differed in strength.

Pollen
The concentrations of the studied elements in the control and deficient pollen are presented in Table 1. The potassium concentrations in the Control-Apis, Na-deficient and Zn-deficient pollen loads corresponded to 97-101% of that in Control-Osmia, whereas the K concentration in the Kdeficient treatment was ca. 26% lower than that in both control treatments. Sodium concentrations were ca. 1-16% higher in the Control-Apis, K-deficient, and Zn-deficient treatments and 39% lower in the Na-deficient treatment than in Control-Osmia. The zinc concentration in the Zn-deficient treatment was 39% lower than that in Control-Osmia, while the Zn concentrations in the Control-Apis, K-deficient, and Na-deficient treatments were ca. 2-7% lower than that in Control-Osmia. To calculate the percentages of each type of cocoon developed by the bees we considered all 15 bee specimens as 100%, constituting every treatment and control. Therefore, this 100% consisted of the sum of (1) specimens that reached the adult stage and developed to the first cocoon stage, (2) specimens that reached the adult stage and developed to the second cocoon stage and (3) specimens that died before reaching the adult stage (the majority of which did not reach the last larval stage, i.e., spinning larvae that start to produce the cocoon). Organismal death before reaching maturity has obvious negative consequences for fitness, therefore we separated these specimens in our analysis from those that successfully reached maturity and developed cocoons to make our analysis ecologically relevant. In this way, we treated the cocoon developmental stage as an ecologically meaningful trait that influences the fitness of living and mature bees. This analysis was complemented by a simultaneous redundancy analysis (RDA) of the datasets on body mass and cocoon stages performed in Canoco 5 [40], which helped us to determine whether the negative effects of nutrient scarcity on these two life history traits (1) were correlated and (2) differed in strength.

Pollen
The concentrations of the studied elements in the control and deficient pollen are presented in Table 1. The potassium concentrations in the Control-Apis, Na-deficient and Zn-deficient pollen loads corresponded to 97-101% of that in Control-Osmia, whereas the K concentration in the K-deficient treatment was ca. 26% lower than that in both control treatments. Sodium concentrations were ca. 1-16% higher in the Control-Apis, K-deficient, and Zn-deficient treatments and 39% lower in the Na-deficient treatment than in Control-Osmia. The zinc concentration in the Zn-deficient treatment was 39% lower than that in Control-Osmia, while the Zn concentrations in the Control-Apis, K-deficient, and Na-deficient treatments were ca. 2-7% lower than that in Control-Osmia.

Cocoon Development
The percentages of each type of cocoon developed by bees are presented in Figure 3. In general, 73-74% and 66-73% of female and male bees, respectively, developed almost fully or fully formed cocoons (second stage) when reared on Control-Apis and Control-Osmia pollen.
The comparisons of cocoon status at the second developmental stage revealed that, for females, significantly fewer fully developed cocoons were observed in the K-deficient (χ 2 = 11.25; p = 0.0008), Na-deficient (χ 2 = 14.35; p = 0.0002) and Na + supplemented (χ 2 = 4.80; p = 0.03) treatments than in Control-Apis. Similarly, significantly fewer fully developed male cocoons were observed in the K-deficient (χ 2 = 12.15; p = 0.0005), Na-deficient (χ 2 = 6.80; p = 0.009) and Na + supplemented (χ 2 = 6.80; p = 0.009) treatments than in Control-Apis. In addition, significantly more developed cocoons were observed in the K + supplemented treatment than in the K-deficient treatment for both females (χ 2 = 5.71; p = 0.02) and males (χ 2 = 6.71; p = 0.01). Table 2. Percentages of mortality and cocoons at the second stage of development in O. bicornis female and male bees reared on pollen characterized by different elemental compositions from the 3-day larva to the imago stage. Note that letters and asterisks denote significant differences always within a single sex and between only two treatments (letters: element-deficient vs element + supplemented; asterisks: single treatment vs Control-Apis).  Considering that mortality has the most important and preliminary effect on bee fitness, dead specimens were also included in the graphic to emphasize the overall survival and development patterns for all of the studied individuals. Therefore, all of the percentages were calculated for N = 15 specimens.

Imago Body Mass
The effect of treatment on body mass was significant for both females (p = 0.00002) and males (p ≤ 0.0001). In total, 78 females were included in the analysis, for which significantly lower body masses Considering that mortality has the most important and preliminary effect on bee fitness, dead specimens were also included in the graphic to emphasize the overall survival and development patterns for all of the studied individuals. Therefore, all of the percentages were calculated for N = 15 specimens.

Imago Body Mass
The effect of treatment on body mass was significant for both females (p = 0.00002) and males (p ≤ 0.0001). In total, 78 females were included in the analysis, for which significantly lower body masses were observed for individuals in the K-deficiency, K + supplemented, and Na-deficiency treatments compared with Control-Osmia, and no differences were observed between the other treatments ( Figure 4). For males, 75 individuals were included in the analysis. The body masses of male individuals in both the K-deficient and K + supplemented treatments were lower than those in the Control-Osmia and Zn + supplemented treatments. Moreover, significantly higher body masses were observed for males exposed to Zn + supplemented pollen than for males exposed to Zn-deficient pollen.
Biology 2020, 9, x FOR PEER REVIEW 10 of 17 were observed for individuals in the K-deficiency, K + supplemented, and Na-deficiency treatments compared with Control-Osmia, and no differences were observed between the other treatments ( Figure 4). For males, 75 individuals were included in the analysis. The body masses of male individuals in both the K-deficient and K + supplemented treatments were lower than those in the Control-Osmia and Zn + supplemented treatments. Moreover, significantly higher body masses were observed for males exposed to Zn + supplemented pollen than for males exposed to Zn-deficient pollen.

Simultaneous Redundancy Analysis (Body Mass Plus Cocoon Stage)
The RDA of the imago body mass and cocoon stage ( Figure 5) suggested that the negative effects of nutrient scarcity in the diet on these two traits were not correlated. For females and males, the first two axes explained 36.82% and 45.36% of the total variance, respectively. Relationships between adult mass/cocoon stage and the experimental diets are denoted by vectors. For both sexes, a vector symbolizing the cocoon stage positioned between the axes was situated perpendicular to the vector symbolizing adult mass, with the number of well-developed cocoons in a treatment increasing from the right-lower to the left-upper corner of the graphs and the adult mass increasing contrarily from the right-upper to left-lower corner of the graphs. For both sexes, the vector symbolizing adult mass was larger than the vector symbolizing the cocoon stage, suggesting a stronger effect of nutrient scarcity on mass than on cocoon development. Similar to previous analyses, the strongest negative effect of nutrient scarcity on cocoon development was observed for K in both sexes; additionally, the RDA suggested a similar effect for Na in females. For males, a positive effect of Zn supplementation on body mass was revealed, similar to previous analyses.

Simultaneous Redundancy Analysis (Body Mass Plus Cocoon Stage)
The RDA of the imago body mass and cocoon stage ( Figure 5) suggested that the negative effects of nutrient scarcity in the diet on these two traits were not correlated. For females and males, the first two axes explained 36.82% and 45.36% of the total variance, respectively. Relationships between adult mass/cocoon stage and the experimental diets are denoted by vectors. For both sexes, a vector symbolizing the cocoon stage positioned between the axes was situated perpendicular to the vector symbolizing adult mass, with the number of well-developed cocoons in a treatment increasing from the right-lower to the left-upper corner of the graphs and the adult mass increasing contrarily from the right-upper to left-lower corner of the graphs. For both sexes, the vector symbolizing adult mass was larger than the vector symbolizing the cocoon stage, suggesting a stronger effect of nutrient scarcity on mass than on cocoon development. Similar to previous analyses, the strongest negative effect of nutrient scarcity on cocoon development was observed for K in both sexes; additionally, the RDA suggested a similar effect for Na in females. For males, a positive effect of Zn supplementation on body mass was revealed, similar to previous analyses.

Discussion
The comparative experimental approach presented in this study provides evidence that deficiencies in specific elements in larval food impose constraints on certain life history traits and on the fitness of wild bee Osmia bicornis.
An important but understudied component of bee nutritional ecology is the relationship between K and Na concentrations in food [27]. In our study, K deficiency had similar effects on both sexes as follows: reduced survivability, reduced body mass, and underdevelopment of cocoons. K supplementation improved survivability and increased the proportion of well-developed cocoons, but had no effect on body mass. This effect is in line with that suggested by a theoretical study demonstrating that the trade-off for K may occur between allocation to the adult bee body and allocation to its cocoon [9]. Such a phenomenon was observed in our study, where the allocation of K to cocoons resulted in a smaller body size.
In the current study, Na scarcity strongly reduced survivability for both sexes; however, Na supplementation had a slight positive effect on female fitness, which manifested as increases in adult body mass. Among female larvae fed Na-deficient and Na + supplemented pollen, only three and five individuals, respectively, survived, whereas for male larvae, four individuals survived on both Na-deficient and Na + supplemented pollen. The facts that only a small number of specimens survived and even fewer of them developed cocoons suggest that something other than sodium might have affected the bees. A possible explanation is the scarcity of other colimiting nutrients (apart from Na) or the presence of poisonous substances; for example, bees may be negatively affected if their food consists of a large proportion of pollen having unfavorable chemical properties [41]. Additionally, the digestibility of pollen from specific species might affect bee fitness [42]. The pollen pools used in our experiment were hand-sorted based on color, therefore species composition was not assessed for the pollen pools, and no nutrients other than K, Na, and Zn were analyzed in the pools. Thus, we cannot conclude with 100% certainty that Na scarcity was the sole driver of such low survival and poor development in bees fed Na-deficient pollen. Nevertheless, a slight but significant effect of Na supplementation on female body mass was observed irrespective of any factors that might have affected the outcome of our study. Importantly, body size positively influences the fitness of females but not males [36,37]; therefore, the observed effect has ecological relevance.
The levels of potassium and sodium are essential for homeostasis in living cells. Both elements are maintained in gradients that are involved in the maintenance of transmembrane electrochemical

Discussion
The comparative experimental approach presented in this study provides evidence that deficiencies in specific elements in larval food impose constraints on certain life history traits and on the fitness of wild bee Osmia bicornis.
An important but understudied component of bee nutritional ecology is the relationship between K and Na concentrations in food [27]. In our study, K deficiency had similar effects on both sexes as follows: reduced survivability, reduced body mass, and underdevelopment of cocoons. K supplementation improved survivability and increased the proportion of well-developed cocoons, but had no effect on body mass. This effect is in line with that suggested by a theoretical study demonstrating that the trade-off for K may occur between allocation to the adult bee body and allocation to its cocoon [9]. Such a phenomenon was observed in our study, where the allocation of K to cocoons resulted in a smaller body size.
In the current study, Na scarcity strongly reduced survivability for both sexes; however, Na supplementation had a slight positive effect on female fitness, which manifested as increases in adult body mass. Among female larvae fed Na-deficient and Na + supplemented pollen, only three and five individuals, respectively, survived, whereas for male larvae, four individuals survived on both Na-deficient and Na + supplemented pollen. The facts that only a small number of specimens survived and even fewer of them developed cocoons suggest that something other than sodium might have affected the bees. A possible explanation is the scarcity of other colimiting nutrients (apart from Na) or the presence of poisonous substances; for example, bees may be negatively affected if their food consists of a large proportion of pollen having unfavorable chemical properties [41]. Additionally, the digestibility of pollen from specific species might affect bee fitness [42]. The pollen pools used in our experiment were hand-sorted based on color, therefore species composition was not assessed for the pollen pools, and no nutrients other than K, Na, and Zn were analyzed in the pools. Thus, we cannot conclude with 100% certainty that Na scarcity was the sole driver of such low survival and poor development in bees fed Na-deficient pollen. Nevertheless, a slight but significant effect of Na supplementation on female body mass was observed irrespective of any factors that might have affected the outcome of our study. Importantly, body size positively influences the fitness of females but not males [36,37]; therefore, the observed effect has ecological relevance.
The levels of potassium and sodium are essential for homeostasis in living cells. Both elements are maintained in gradients that are involved in the maintenance of transmembrane electrochemical potential differences, which are essential for cell signaling and secondary transport [29]. Conversely, disruption of potassium and sodium cation gradients can result in paralysis or death. Regarding potassium, its homeostasis in insects is associated with adaptation to extreme cold and heat [30]. For example, studies on adults of the true bug Pyrrhocoris apterus and the beetle Alphitobius diaperinus revealed that during a seven-day exposure at low temperatures (−5 • C and 4 • C, respectively), a gradual increase in potassium cations was observed within the hemolymph of both species, whereas for the other studied elements, i.e., magnesium and sodium, almost no changes were observed, indicating the importance of potassium homeostasis in response to cold stress [43]. Overall, it is not surprising that potassium had such a strong influence on the survival and development of cocoons in our experiment. However, because plant tissues in general contain high levels of K and low levels of Na, the K:Na ratio in herbivores' foods strongly influences their fitness and must therefore be adequately balanced [27]. For example, acute bee paralysis may be caused by an excessively high K:Na ratio in their food [27,44]. Although, in diverse floras, K is not expected to have a limiting effect on herbivorous insects, including wild bees; nonetheless, low levels of potassium can be found in the pollens of several plant species [9]. Therefore, in the case of monocultures or habitats with low species richness, such a phenomenon might occur. For instance, based on data available from the literature, stoichiometric mismatches were calculated for different pollen species, showing that Silybum marianum, Olea europaea, and Lavandula sp. produce stoichiometrically unbalanced pollen for O. bicornis bees in terms of the potassium content [28]. Importantly, these plants are usually grown in large agricultural areas.
Regarding sodium, its gradient maintains the secondary transport system, which mediates the transport of other ions, substrates (e.g., glucose), and neurotransmitters across the plasma membrane [29]. Most importantly, Na is one of the most limiting elements for herbivores [27], including bees [39,45], and strong preferences of different bee species for sodium have been shown [27,[45][46][47]. The sodium concentration in pollen depends on the species and is the most variable among all the elemental concentrations studied, differing fivefold between the pollens of species with the minimum and maximum concentrations [9]. Therefore, considering the availability of Na for developing bees, the occurrence of plant species producing Na-rich pollen in bee habitats may be important for both females and males, potentially influencing the growth of the entire bee population.
Females and males differed in their responses to Zn levels. Supplementation with Zn had the strongest effect on males, with lower mortality rates and higher body masses being observed upon exposure to Zn + supplemented pollen in comparison with Zn-deficient pollen. Although the percentage of Zn in the body is estimated to not exceed 0.02%, Zn is the most important trace element for the proper functioning of various tissues, organs and systems [31]. For instance, Beanland and colleagues [48] showed that the proportion of Zn in relation to two other minerals (Fe, B) in soybean (Glycine max) affected the development of three herbivorous insects (Pseudoplusia includens, Epilachna varivestis, and Anticarsia gemmatalis). A study on Zn supplementation in the sucrose diet of the honeybee A. mellifera ligustica revealed that 30 mg Zn kg -1 in food was sufficient to maintain the antioxidative (Cu/Zn-SOD activity) status of bees and to increase the survival of worker bees in comparison to those of bees exposed to lower (0-15 mg Zn kg -1 ) and higher (>45 mg Zn kg -1 ) Zn levels [49].
Interestingly, it has been shown for various bees that adults of different sexes use different plant species as food resources [50]. In O. bicornis, larval diets composed of pollen gathered by a female for her daughters and sons differed in nutritional quality, and this difference reflected sex-specific nutritional optima [28]. Moreover, female O. bicornis bees have a higher demand for Zn than males [28]. In general, in our study, zinc was the diet element to which females were the least sensitive (in terms of mortality, cocoon development and body mass). The explanation for these results might be associated with the function of zinc in female bee reproduction. For example, Cane [51] showed that after emergence from the cocoon, adult female Osmia californica bees require access to pollen to mature their oocytes and reproduce. Wasielewski and colleagues [52] observed that the first oocytes and ovaries of O. rufa (bicornis) bees developed gradually during wintering, and the authors linked the development to the vitellogenin content, whereas Lee and colleagues [53] found that after diapause, the length of the ovary and first oocytes as well as the number of oocytes were correlated with the vitellogenin secretion level in Osmia cornifrons. Vitellogenic proteins are female-specific egg-yolk precursors transferred to oocytes, where they provide nourishment for embryos [54]. Interestingly, the vitellogenin content was found to be closely related to Zn levels in female honeybees, because this protein acts as a Zn carrier [55,56]. Thus, we hypothesize that bee mothers provide both female and male eggs with pollen that contains a sufficient Zn level for development and functioning, but at later stages (i.e., after emergence from the nest), females can replenish zinc levels for continued functioning, e.g., vitellogenin and egg production, by eating pollen [51]. Importantly, similar supplementation strategies for other nutritional elements are impossible because adequate amounts and ratios of these elements are needed during larval growth and pupation. Therefore, Zn deficiency might exert constraints on developing males, manifesting as reduced survivability and body mass, whereas in females, Zn scarcity during the larval stage might negatively affect the reproductive system (not studied in our experiment). The reproductive system might be further rebuilt by adult females to ameliorate this negative effect.
Various bees, even those feeding on a variety of plant species, show preferences for particular plant species as food sources, especially considering pollen food for larvae [3,57,58]. Moreover, these preferences may be driven by specific nutritional needs reflected in the chemistries of the gathered food [13,[59][60][61]. However, the biochemical metrics commonly used in bee nutrition studies, although ideal when focusing on bee biology, seem to be insufficient when considering the bee as part of the ecosystem and biogeochemical cycle-an organism involved in nutrient cycling. Therefore, adopting approaches complementary to traditional approaches, such as the biochemistry-oriented view, and focusing on nutrient flow through ecosystems allow for a better understanding of interactions between pollinators and other food web components (e.g., soil-plant-pollinator interactions) [62][63][64]. According to Paseka and colleagues [26], the frequency of element colimitation in terrestrial ecosystems suggests that stoichiometric effects on plant productivity may, in turn, affect pollen production and thus pollinators, although no studies on the relationships between elemental ratios in the environment and pollen production have been performed.

Conclusions
Bee conservation efforts are often based on simplistic assumptions, considering the nutritional ecology of only one life stage (usually adults) or sex (usually females). However, bee populations consist of individuals of various life stages and different sexes. Effective management strategies for maintaining populations of wild bees may be achieved only by obtaining and understanding the relationships between the complex nutritional demands of the whole bee population and the nutritional supply of pollen produced by different plants, including sex and life-stage differences in bee nutritional needs. Within this context, the current study provides the first insight into the effects of specific-atom scarcity in larval food on the life history traits and fitness of bees, thereby revealing the nutritional mechanisms underlying the nutritional ecology, behavioral ecology and population functioning of bees within an ecosystem context.
In this study, we confirmed earlier theoretical predictions, showing the following:

1.
O. bicornis life history traits and fitness are shaped by the availability of atoms of specific chemical elements in larval food.

2.
Some of these traits might be shaped by the availability of specific elements in a sex-specific manner: Na might influence female body mass, whereas Zn might influence the mortality and body mass of males.

3.
A trade-off between the K allocation to cocoons and the adult body may exist and might influence the development of cocoons and the body mass of adult bees.