How to really help bees: nutrient demand and supply in a changing environment

Local and global changes affect which pollen varieties are available to bees in the environment. Therefore, wild bees cannot always access the optimally balanced diet required for their survival. Our feeding experiment showed that the nutritional quality of the pollen diet eaten by bee larvae is shaped not by pollen diversity but by a specic pollen species composition that results in specic nutrients being scarce or sucient; this species composition inuences bee survivability, development and mass. We proposed that the functioning of bee populations and communities may depend on the oral diversity of the local habitat, which determines whether a nutritionally balanced pollen diet obtained from specic species can be provided to bee larvae. Holistically considering wild bee tness and health and the different characteristics of the food base at both the ecosystem and bee biology levels can provide new, important knowledge for conserving bees and their critical ecological roles. Our preliminary experiment showed that the scarcity of specic nutrients in a larval diet indeed impacted bee development in a sex-dependent manner , as predicted by theoretical calculations based on stoichiometric phenotypes and stoichiometric mismatches between consumers and their food 10 The current study provides the rst detailed insight into this phenomenon that is based on a large pool of specimens utilizing different nutritional treatments. We have shown that the scarcity of specic nutrients (atoms of vital chemical elements) in bee larval food shapes the tness of bee individuals in a sex-dependent manner. Females were strongly negatively affected by the scarcity of Na and P, which resulted in higher mortality rates for females than for males. In fact, the Na- and P-scarce was the only treatment resulting in 100% mortality (only for females); this effect was mitigated when the diet was supplemented with Na and mitigated even more when the diet was supplemented with Na and P.


Introduction
Bees are deeply entrenched in ecological systems and processes, and bee nutritional ecology, health and tness are shaped by ecological interactions [1][2][3] . Although much is understood about the details of the functioning of bees and pollinators in general, connections among the various elements of knowledge about bee nutrition and well-being that would allow for a broad, ecosystem-oriented understanding of wild bee ecology and conservation are still lacking [1][2][3] . The functioning and prosperity of bee populations are in uenced by human activities that cause changes in landscape structure, leading to habitat loss, fragmentation and degradation. As a result, bees may be subjected to de ciencies of food resources, which have been indicated as one of the most important drivers of bee decline [4][5][6] . The term "de ciency of food resources" should be emphasized here; it does not simply mean less food but rather refers to changes in oral species diversity and composition caused by landscape changes due to any agricultural, industrial, construction, or other anthropogenic activity.
The quantity and quality of food available in the environment both strongly impact consumer tness and therefore determine the success of the consumer population 7,8 . Crop monocultures may increase the amount of food available for bees while at the same time decreasing its quality, i.e., the balance of its nutritional composition [9][10][11] . Monocultures can serve as nectar sources for adult bees 12 , but pollen from monocultures may be insu cient to meet the needs of developing bee larvae 10 . This is because nectar, which is rich in sugars, is extremely lacking in other nutrients [13][14][15] ; thus, it is an important food for strongly energetically limited adults. In turn, for growing and developing larvae, pollen provides a source of body-building matter. Nevertheless, mono oral pollen from certain species is not nutritionally balanced enough to allow normal juvenile bee development 10,16 . Moreover, bee sexes differ in their nutritional needs and prefer different plant species as food sources [17][18][19] . However, bee conservation efforts are often based on simplistic assumptions regarding the nutritional ecology of one life stage (usually the adult stage) or sex (usually females). In reality, bee populations consist of individuals in various life stages and of different sexes. We believe that a holistic approach to bee conservation that is based on details of bee biology but also considers bees as part of ecosystems and as being involved in ecological processes and interactions may be important for developing more effective management strategies for maintaining wild bee populations. Therefore, in this study, we connect the basic physiological mechanisms underlying bee nutritional needs with the functioning of bee populations within the context of ecosystems in which the ora offers either a nutritionally balanced or unbalanced food base.
In general, oral diversity has a positive impact on bee populations, providing large amounts of nectar and pollen with diverse botanical origins 6,20−22 . However, the mechanism underlying this positive impact is not understood. Here, we hypothesize that this mechanism is based on the differences between bee nutritional demand and environmental nutritional supply: bees are often unable to achieve a nutritional balance in their diets because only a strictly limited number of plant species provide the required resources in the needed proportion; these species are not available in certain landscapes that have speci c oral compositions 1,10,22,23 . In recent years, a debate has developed over the importance of biological diversity to ecosystem functioning and ecosystem services. The economic value associated with biodiversity is a strong argument for conservation efforts; however, some studies indicate that this value has certain limitations 24 . For instance, only a small minority of bee species provide most of the crop pollination services worldwide 24 . This fact is most likely linked to bee nutritional needs; a more indepth understanding of the nutritional requirements of wild bees is needed, and the relationship between oral species diversity and species composition could become a focal point for future conservation efforts.
Among the various measures of the nutritional balance of bee diets, the concept of stoichiometric mismatch and the ratios of atoms of vital chemical elements that form organic molecules seems to be one of the most ecologically relevant; this measure allows bee nutritional physiology and tness to be connected directly with nutrient cycling throughout the whole food web, which is shaped by complicated synergy of ecological interactions 8, 25,26 . The atoms of chemical elements in various proportions build the tissues and bodies of organisms. The most fundamental feature of these elements, which allows them to be used in ecological stoichiometry, is that these speci c atoms cannot be transformed into different atoms by the organism. This feature distinguishes atoms from organic compounds. Within this context, ecological stoichiometry provides a common currency that links the ecology of organisms to life history trade-offs and evolutionary processes that are entrenched in the biogeochemical economy of life 27 . This currency is the ratio of atoms that compose the bodies of organisms and their food 8 . Consumers, especially herbivores, are faced with a high threshold of stoichiometric incompatibility (i.e., stoichiometric mismatch) between the chemical composition of their tissues and their food 8, 25,28 . Within this context, the concepts of the biogeochemical niche and the stoichiometric niche were recently proposed 29,30 . Both concepts share similar ideas and focus on changes in the availability of particular atoms that are needed in speci c proportions to build and maintain organisms as a factor driving the evolution of life and structuring the environment (hereafter, the term "stoichiometric niche" will be used). The stoichiometric niche is de ned as a multivariate niche space occupied by a group of individuals with similar stoichiometries; speci c species occupy speci c niches, and each niche is stoichiometrically explicit [29][30][31] .
To obtain stoichiometrically balanced food, various organisms with speci c optimal elemental phenotypes (also called ionomes) may prefer various food sources that provide nutrients in proportions that re ect their nutritional demand 31 . Feeding on stoichiometrically unbalanced food, which would result in decreased concentrations of some elements and therefore an unbalanced stoichiometric ratio, may negatively affect bee tness 18 . Nutritional niches (irrespective of the speci c de nition applied) may differ between sexes, imposing sex-speci c nutritional constraints on organisms 31,32 . It has been suggested that in the case of wild bees, sexual dimorphism in nutritional demand may be an important factor for and should be considered in bee conservation efforts that focus on providing balanced diets to bees [17][18][19] .
Here, we propose and investigate the mechanisms underlying the link between oral diversity and bee performance. We hypothesize that (1) the nutritional quality of bee larval food is connected to both the species diversity and species composition of the pollen constituting this food but that (2) the factor directly driving the nutritional quality of the food is the species composition and not the species diversity, because (3) the species composition of the pollen provisions, i.e., the presence of pollen from speci c plant species, is associated with the stoichiometric phenotype of the pollen, which re ects its ability to provide atoms of chemical elements in the speci c proportions that meet or do not meet bee nutritional demands. We investigated the abovementioned ideas and hypotheses using a solitary bee model system.
Utilizing this model organism, we were able to perform feeding experiment controlling and measuring independently (1) the pollen diversity and species composition of larval food, (2) the stoichiometric phenotype of this food and (3) the different nutritional needs of bee females and males (Figs. 1 and 2).
Additionally, this study system provided us with a unique opportunity to determine the nutritional needs of female and male bees and to address a general ecosystem ecology problem, that is, the life history of an organism embedded within an ecosystem. We believe that this approach makes the evaluation of oral resource limitations in bees ecologically relevant and may provide precise tools for use in conservation actions. Such precise tools are needed because, as the nutritional ecology of wild bees is still not well understood, management actions to support bee habitats are not effective and often fail [33][34][35] .

Results
Pollen pools -nutritional quality, species composition and species diversity The pollen pools used in the feeding experiment differed in their nutritional quality, as re ected by the concentrations and proportions of the studied elements (Fig. 3). The control-Osmia, Apis1 and Apis2 pollens had similar nutritional quality, -CuZn1 and -CuZn2 were scarce in Cu and Zn, -KP was scarce in K and P, and -NaP was scarce in Na and P (Fig. 3). The exact concentrations of the studied elements are shown in Supplementary Table S2. There was a strong correlation between the elemental composition of the pollen mixtures and the presence of pollen from speci c plant species: scarcity of Cu and Zn was associated with a predominance of Brassica napus pollen, scarcity of K and P was associated with a predominance of Anthriscus sp. pollen, and scarcity of Na and P was associated with a predominance of Aesculus sp. pollen.
The pollen pools differed in both species composition and species diversity. -CuZn1 was the least diverse, and -KP was the most diverse (Tables 1 and 2). The most diverse pollen pool (-KP) also had the lowest dominance value, whereas the -CuZn1 and -CuZn2 pollens, composed of approx. 80-90% of B. napus pollen grains, were the most strongly dominated by a single species (Tables 1 and 2). All the nutritionally balanced pollen pools, i.e., control-Osmia, Apis1, and Apis2, had average species diversity as expressed by the Fisher's alpha and D indices among all studied pools (Table 1) but differed in their species composition ( Table 2). Table 1 Diversity indices characterizing the pollen pools used in the feeding experiment based on the number of bee specimens and the number of taxa of pollen grains composing the pools.  For each pollen pool, the values for the pollen species composing more than 20% of the pool are shown in bold. All species found in the pollen pools and their percentages in each pollen pool are presented in Supplementary Table S1.

Mortality
Bee mortality differed among treatments, and the differences in mortality were sex dependent (Fig. 4).
The -CuZn1 and -CuZn2 treatments without supplemental nutrients resulted in increased mortality compared with the control in both sexes, and this effect was stronger for males than for females. This effect was mitigated in females when both Cu and Zn were supplied. In males, the effect was mitigated when Zn was added, but adding both Cu and Zn had a stronger mitigating effect. The mitigating effect caused by the addition of both Cu and Zn was stronger for males than for females. The -KP treatment without supplemental nutrients resulted in increased mortality, and this effect was stronger for males than for females. The addition of either K or K and P caused an even stronger increase in female mortality but did not have any effect on males. The -NaP treatment without supplemental nutrients resulted in increased mortality in both sexes, but this effect was stronger for females (100% mortality) than for males. The addition of either Na or Na and P mitigated this effect only for females, and the addition of both Na and P had a stronger mitigating effect than adding only Na.

Cocoon development
Cocoon mass was not correlated with cocoon development, and adult mass was moderately correlated with cocoon mass (Table 3). Spearman's R coe cient, p < 0.05. For both sexes, adult mass was moderately correlated with cocoon mass and slightly correlated with cocoon development, while cocoon mass was not correlated with cocoon development.
The effect of diet on cocoon development differed among treatments, and these differences were sex dependent (Fig. 5). The -CuZn1 and -CuZn2 treatments without supplemental nutrients resulted in underdeveloped cocoons in both sexes. This effect was mitigated in females when both Cu and Zn were added. In the case of males, mitigation was observed when Zn was added, but adding both Cu and Zn had a stronger mitigating effect. The -KP treatment without supplemental nutrients resulted in underdeveloped cocoons in both sexes. Supplementation with either K or K and P mitigated this effect in males. -NaP treatment without supplemental nutrients resulted in underdeveloped cocoons in males, and supplementation, whether with Na or Na and P, had no effect.

Adult mass and cocoon mass
The effect of diet on the masses of cocoons and adults differed between treatments, and these differences were sex dependent (Fig. 6). The -CuZn1 and -CuZn2 treatments without supplemental nutrients resulted in reduced body mass in both sexes, and the effect was stronger for males than for females. At the same time only the female cocoon mass was reduced in these treatments. Supplementation with both Cu and Zn mitigated the negative effect on female body mass but not that on cocoon mass. For males, the negative effect on their body mass was mitigated by supplementation with solely Cu, solely Zn, and both Cu and Zn. The -KP treatment without supplemental nutrients reduced both body mass and cocoon mass in both sexes. Supplementation with both K and P mitigated the negative effect on male cocoon mass but not that on male body mass. The -NaP treatment reduced both body mass and cocoon mass in males, and supplementation, whether with Na or Na and P, had no mitigating effect.

Discussion
The links between oral diversity and bee functioning and between oral species composition and bee functioning are unclear, and a better integration of the approaches and frameworks scattered among various studies is needed 2,22,23,36,37 . Here, we provide a point of view rooted in organismal bee physiology and further extend it over the whole ecosystem, of which the bee population and the oral composition form a part. We suggest that bee tness may be shaped by the availability of vital nutrients at speci c concentrations associated with speci c key plant species. We conclude that more diverse oras provide bees with the opportunity to use their preferred resources; however, the direct mechanism driving the increase in bee tness in diverse environments is the diet balance achieved with the pollen of key plant species.
In our study, the pollen mixtures that were both the most diverse and the least diverse in terms of oral species had limiting effects on bee growth, development and tness. This phenomenon was associated with the presence of speci c plant species that are responsible for dietary imbalances in terms of certain nutrient elements. Diminished concentrations of Cu and Zn were associated with the dominance of Brassica napus; a lack of K and P was associated with a predominance of Anthriscus sp., and other Brassicaceae; and a lack of Na and P was associated with a predominance of Aesculus sp. Interestingly, the two diets (-CuZn1 and -CuZn2) with a high percentage of B. napus pollen were characterized by similar dominance D indices but relatively distinct Fisher's alpha indices, and both diets had similar negative effects on bee survivability and development. However, the addition of Cu and Zn to these diets mitigated their negative effects. Therefore, B. napus pollen, if highly concentrated in the larval diet, increases bee mortality, decreases body mass, and inhibits cocoon development, and this effect is driven by the scarcity of Cu and Zn. Our results are in accordance with those of previous studies 37,38 . Klaus et al. 37 observed diminished reproduction of O. bicornis in mono oral habitats (100% oilseed rape) compared with that in habitats with more complex oral resources (50% wild owers: 50% oilseed rape) in a semi eld experiment with pesticides. Holzschuh et al. 38 in turn found that almost no O. bicornis reproduction occurred in isolated oilseed rape elds that were not adjacent to grasslands, whereas reproduction was prominent in oilseed rape elds that were adjacent to grasslands as well as in grasslands that were adjacent to oilseed rape elds. Our study is the rst to reveal the mechanism behind the widely known phenomenon of the need for diverse oral resources in the vicinity of B. napus monocultures, namely, the de ciency of speci c nutritional elements (Cu and Zn), which in turn can be mitigated by the presence of speci c plant species whose pollen is rich in those elements, e.g., Filipendula sp.
A negative effect was also imposed on bee growth, development and tness by the two most diverse diets, which had either no dominant pollen species (treatment -KP) or moderately dominant pollen species (treatment -NaP). This negative effect was most likely caused by the speci c species composition of these diets, which resulted in a scarcity of vital nutrients (either K and P or Na and P); however, additional causal factors cannot be excluded. The supplementation of -KP with K or both K and P had no effect on females. In males, the supplementation of the -KP pollen pool had a positive effect on cocoon development and mass, indicating that to some extent, the negative effects were associated with K and P de ciency. Regarding -NaP, the positive effect on the studied parameters after supplementation was more pronounced than that in the supplemented -KP diets, but the survival, cocoon development and masses still did not reach the levels of the control individuals. In general, the relatively high diversity indices calculated for -KP and -NaP do not imply an adequate diet, and even supplementation did not mitigate the negative effects of these diets; certain other factors may have been related to the observed results, e.g., colimiting nutrients (apart from K and P) or the presence of poisonous substances 39 . These additional effects might also be related to the speci c species composition of the pollen, with -KP having a 25% concentration of Anthriscus pollen and -NaP having a 65% concentration of Aesculus pollen.
In contrast, the pollen diet collected in nature by O. bicornis (Control-Osmia) and the two Apis diets (Apis1 and Apis2), which did not have any negative effects on the bees, had moderate pollen species diversity and dominance. What differentiated these diets from the limiting diets was the speci c pollen species composition, which resulted in a stoichiometric phenotype that was nutritionally balanced for O. bicornis. Therefore, we infer that, considering the pollen pool available directly to bee larvae, the species composition of the larval pollen diet is more important to an adequate nutritional balance than the diversity of the larval pollen diet. However, we emphasize that the pollen pool that is directly available to bee larvae is not identical to the pollen pool provided by particular ora in the environment to adult bees that collect pollen for their larvae. Therefore, below, we extend our point of view to the whole ecosystem. Population (a) thrives and prospers because the bees are able to collect pollen from their preferred species and thereby compose a nutritionally balanced diet that allows for proper larval development. Therefore, in the next generation, the number of individuals increases, and the bees are appropriately sized and healthy. In the case of population (b), the poor ora offers mainly stoichiometrically unbalanced pollen, resulting in stoichiometric mismatches 25,26 for the bees. The bees experience high mortality, generate underdeveloped cocoons that further increase their mortality and have small body sizes. Since the negative effects of dietary imbalances affect females to a greater degree than males, as shown in the current study, the next generation is dominated by males. Moreover, smaller bees can y only shorter distances and can carry less pollen to their progeny than larger bees, which further negatively affects future generations. Overall, population (b) is in decline in this scenario. Therefore, even though all the pollen species occurring in ora (b) also occurred in ora (a), the additional species that occurred only in ora (a) allowed bee population (a) to thrive and prosper. This positive effect on bee populations happens more often in diverse oras than in poor oras simply by chance -the more plant species are available, the greater the chance of nding pollen that allows a nutritionally balanced diet. Therefore, the oral composition may shape bee populations by controlling the nutritional supply available to bees. The occurrence of key plant species that provide the correct dietary stoichiometric balance for bees may be a factor in shaping bee populations. Access to these key plant species is essential for bee growth and development regardless of whether the pollen from these plants is gathered intentionally or not. Bee populations are in uenced by the nutritional balance of the bee larval diet, and this balance depends on the oral composition of the bee habitat (the stochiometric niche). High oral diversity may be necessary to maintain populations of pollen eaters by providing key plant species that allow for dietary nutrient balancing; single-species crop plantations, even if they are rich in nectar and pollen, might limit bee development. Consequently, changes in local oral communities may shape bee colonies, populations and communities. Therefore, not only the quantity but also the quality of food sources for bees should be considered in intervention strategies aimed at improving the nutritional base for bees.
The demand for resources for growth and development, as re ected in organismal stoichiometry, is usually studied by comparing different species, but research has also started to focus on individual variations in the chemical compositions of bodies; much of this variation is expected to be attributable to sex differences 19,31,32 . This is expected because processes involved in life history evolution and population dynamics are likely to differentially affect females and males, thus imposing sex-speci c nutritional limitations 27,32 . By considering such within-species variance, evaluations of resource limitations in a given species can increase their ecological relevance. In our previous study, we presented the idea that both sexes of O. bicornis have different stoichiometric niches 19 . It has also been hypothesized that O. bicornis females collect pollen species in proportions that re ect the sex-speci c nutritional needs of their daughters and sons 10,42 . Sex-speci c differences in stoichiometric phenotypes that can be re ected in stoichiometric niches have also been detected in other invertebrates, including amphipods and spiders 31,43 . Our preliminary experiment showed that the scarcity of speci c nutrients in a larval diet indeed impacted bee development in a sex-dependent manner 18 , as predicted by theoretical calculations based on stoichiometric phenotypes and stoichiometric mismatches between consumers and their food 10 . The current study provides the rst detailed insight into this phenomenon that is based on a large pool of specimens utilizing different nutritional treatments. We have shown that the scarcity of speci c nutrients (atoms of vital chemical elements) in bee larval food shapes the tness of bee individuals in a sex-dependent manner. Females were strongly negatively affected by the scarcity of Na and P, which resulted in higher mortality rates for females than for males. In fact, the Na-and P-scarce treatment was the only treatment resulting in 100% mortality (only for females); this effect was mitigated when the diet was supplemented with Na and mitigated even more when the diet was supplemented with Na and P. Interestingly, apart from the obvious functions related to maintaining transmembrane electrochemical potential differences in living cells 44 , Na plays important roles in regulating the assimilation of N and especially P from food, in phosphate homeostasis and in phosphate sensing at the cellular and organismal levels [45][46][47] . This function may be more important for P-limited females that need to develop their ovaries and to produce eggs than for males that do not develop N-and P-demanding reproductive apparatuses 19 . In contrast, males had higher mortality than females when fed on pollen scarce in K; while K scarcity resulted in decreased masses of bodies and cocoons of both sexes, supplementation with K had a positive effect only on male cocoon masses. These results are in line with those of a previous study showing that Na and K are assimilated and allocated to speci c functions differently in both sexes of O. bicornis 19 . Males experienced higher mortality than females when feeding on pollen that was scarce in Cu and Zn. The scarcity of Cu and Zn also resulted in lower female cocoon mass but had no effect on male cocoon mass. Both sexes also developed lower body masses when fed on pollen that was low in both Cu and Zn, but supplementation with Zn mitigated this effect for males (supplementation with both Cu and Zn had a positive effect on both sexes). These differences are again in line with our study considering the elemental budget, assimilation and allocation of elements by O.
bicornis and are discussed therein 19 . Together, these detailed results show that in terms of the dietary nutrient balance, individual tness is regulated differently and is linked to a sex-speci c optimal proportion of nutrients for the larval food of bees. This is important information that should be considered when designing regulations, laws and actions aimed at wild bee conservation. Existing regulations, laws and actions are based mostly on data obtained for females, especially adult females, and related to their energetic needs rather than to detailed nutritional biology data.
Vaudo et al. 9 stated that bee population declines are linked to nutritional shortages and that possible host plant species vary in their nutritional quality; thus, knowledge of bee nutrition should be applied to the selection of oral resources during habitat restoration. The authors also observed that little is known about the nutritional requirements of bees, which remains true. A better understanding of the nutritional ecology of wild bees may be one of the most critical focus areas in bee ecology 25,48 . The attractiveness of wild ower mixtures for wild bees has been shown to depend on several key plant species 49 , but future studies that link bee nutrition to bee life history traits and tness in the context of oral preferences and oral habitat composition are needed to elucidate the dependency of bees on these factors. Nutrient collapse in plant tissues has been reported in recent studies; speci cally, elevated concentrations of atmospheric CO 2 reduce the concentrations of important nutrients in plant tissues 50 , including the nutrients in the pollen utilized by bees 51 . In this context, understanding the demands of growing bees for a nutritionally balanced diet is even more important; apart from contributing to the development of ecology and evolution studies, the current project will also impact conservation biology.

Methods
We investigated the in uence of various larval diets on bee tness. The diets were composed of pollen mixtures differing in species diversity, species composition, and nutritional quality, as re ected in the concentrations and proportions of studied elements (the P, Na, K, Zn and Cu; the diets were either nutritionally balanced or low in speci c elements). To allow for the clear interpretation of the obtained results, every diet that was low in speci c elements was provided followed by a diet composed of the same pollen mixture that was arti cially supplemented with a salt of the lacking element(s). Our experimental setup, described below, allowed us to (1) study the direct effects of the nutritional quality of the diet on bees, (2) study the indirect effect of plant diversity on bees, and, most importantly, (3) determine whether these effects are related to each other. Such an approach allowed us to clearly indicate which factor was responsible for the obtained results. Moreover, because we studied different features of the food base, the setup allowed us to interpret the results in the context of natural conditions; this increased the ecological relevance of our evaluation of the effects of resource limitation on bees. We  (Fig. 1). This system allowed for easy sex determination, since female progeny are usually located in the rear part of the stem, while males can be found near the entrance ( Fig. 1; 18 ). The stems were checked daily for the presence of larval brood cells, and 500 stems containing 1-3 larval cells with female eggs were rst collected to obtain female specimens to be used in the feeding experiment. New empty cane stems were then mounted to allow the adult bees to continue laying eggs, and the stems were again checked on a daily basis. Every completed stem, i.e., those that were lled with eggs and closed with mud, was collected to obtain male specimens to be used in the feeding experiment. All stems were kept at 21°C, 60% relative humidity, and a 12:12 light:dark photoperiod to allow the eggs to hatch. Three-day-old larvae were used for the experiment due to the fragility and sensitivity of eggs and the possibility of mechanical damage to the eggs during the transfer to experimental containers 18 . To eliminate possible genetic biases, only one female and one male specimen from each stem was collected to be used in the feeding experiment.

Experimental design
With a feeding experiment, we studied whether and how tness-related life history traits (the mortality, adult body mass, cocoon mass and cocoon development) of wild solitary bees (O. bicornis) depended on the species diversity and nutritional composition of the pollen provided as larval food. We studied the abovementioned tness-related life history traits since (1) mortality is the most relevant trait to study; (2) cocoons are secretions made for a speci c purpose to enhance bee tness (providing protection during prewintering and overwintering, i.e., during the rst ten months of the adult part of the life cycle) 52,53 ; and (3) the mass of adults is correlated with the tness of O. bicornis solitary bee females (this effect is not observed in males 54,55 ).
Twenty-ve replicates (Eppendorf tubes, 2 ml) were prepared per treatment and sex. The tubes were lled with speci c pollen diets (see below). The amount of pollen corresponded to the dry mass of pollen provisions typically found in nature, i.e., 195 ± 5 mg dm for females and 140 ± 5 mg dm for males. The dry pollen loads were complemented with either demineralized water (in the case of experimental controls and the pollen mixtures that were low in speci c nutrients) or salt solutions (in the case of the pollen mixtures that were low in speci c nutrients to obtain treatments with the same pollen species composition but different nutritional quality) in an amount that was ca. 25% of the dry pollen mass.
Pollen collected by Osmia in nature is considered to be nutritionally balanced for larvae; thus, the treatments supplemented with salt solutions had concentrations of elements that were similar to those of pollen mixtures collected by Osmia females and provided to their larval progeny. Before starting the experiment, the Eppendorf tubes were left for 24 h to allow the water and salt solutions to penetrate the pollen loads. 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 gone through the life cycle to reach adulthood, i.e., to the stage where fully developed individuals hibernated in their cocoons 53,56 . At the end of exposure, cocoons and undeveloped individuals were collected to determine the degree of cocoon development. Then, live bees were extracted from cocoons, and the mortality rate was assessed. Afterwards, the individuals (i.e., the adult bodies and their cocoons) were frozen and dried using a vacuum drier (80°C, 48 h) to obtain their dry mass. This procedure was previously used in preliminary experiment 18 .

Pollen diets
We established seventeen larval pollen diets differing in their elemental composition and species composition. For every diet that was low in speci c nutrients, we also prepared the same diet supplemented with the lacking nutrients. The nutritional quality of the diets was measured in terms of the balanced/unbalanced proportions of vital nutritional elements (P, Na, K, Zn and Cu), as in previous studies 10,18 . Balanced nutritional quality was de ned as the proportion of the measured elements found in pollen collected by O. bicornis in nature (Control-Osmia pollen diet). Unbalanced nutritional quality was de ned as a scarcity (decreased concentration) of one or more elements in relation to the level of those elements in the Control-Osmia pollen diet. Pollen collection and analyses (Fig. 2 -stages 1 and 2) To prepare the diets for the different experimental treatments, we rst purchased ve packs of honey bee poly oral pollen pellets from different manufacturers (Fig. 2 -stage 1). We divided the pollen from each pack according to color by the naked eye to obtain pollen pellet pools that had speci c species compositions and nutritional qualities. The control-Osmia pollen was collected from Osmia nests. Every pollen pool was mixed manually to obtain a homogenous powder and then freeze-dried to obtain its dry mass (dm) without changing the pollen nutritional properties, as freeze-drying allows vital molecules to be preserved. Then, the pools were analyzed stoichiometrically, i.e., The proportions of various elements were measured and compared in order to select the pollen pools to be used in the feeding experiments ( Fig. 2 -stage 2a). The chemical analyses considered ve elements and their ratios: Zn, Cu, Na, K and P.
We compared the nutritional quality of the pollen types as re ected in their stoichiometric phenotypes (i.e., the proportions of elements; see 10,25 ). Control-Osmia pollen, Apis-collected pollen pellets that were nutritionally similar to control-Osmia pollen (two pollen pools, namely: Apis1 and Apis2) and Apiscollected pollen pellets lacking vital elements (four pollen pools that were scarce in Zn, Cu, Na, K and P, namely: -CuZn1, -CuZn2, -KP, and -NaP) were chosen to be used in the feeding experiment. These pollen pools were analyzed botanically, i.e., their species composition and diversity were investigated (Fig. 2 -stage 2). The exact species composition (the percentage of every noted taxon) of each distinct pollen pool was estimated by counting the pollen grains under a microscope using two samples, each of 3.5 g dry mass (d.m.), from every homogenized pollen pool.
Pollen pool selection and treatments used in feeding experiment (Fig. 2 -stages 3 and 4) The following treatments were established (Fig. 2 -stage 3): (i) Control-Osmia, the natural larval food collected from O. bicornis nests; (ii) Apis1 and Apis2, two unsorted honeybee pollen pellets that were similar to control-Osmia in terms of their nutritional quality but differed in their species composition and diversity; and (iii) four diets that were low in speci c nutrient(s), namely: -CuZn1, -CuZn2, -KP, and -NaP. Additionally, each diet described in (iii) was accompanied by a diet containing the same pollen pool but supplemented with salt of the scarce element; the concentration of the salt was adjusted to re ect the concentration of the corresponding element in Control-Osmia. The pollen pools obtained allowed us to investigate the effects of a scarcity of ve nutrient elements, Cu, Zn, Na, K and P, in pollen mixtures of various botanical origins. Following the results of a preliminary experiment 18 , we used ZnCl 2 and CuCl 2 to supplement Zn and Cu in the bee diets, respectively, and used KCl, KH 2 PO 4 , NaCl, and NaH 2 PO 4 to supplement Na, K and P. All the treatments used in the feeding experiment are shown in Fig. 2 -stage 4.

Chemical analyses
The Zn, Cu, Na, and K concentrations were determined using atomic absorption spectrometry (Perkin-Elmer AAnalyst 200 and Perkin-Elmer AAnalyst 800), and the P content was determined by colorimetry (MLE FIA). Homogenized and dried samples were used to prepare a liquid solution (digested on a hotplate in a 4:1 mixture of nitric acid (70%) and hydrogen peroxide (30%)) that allowed us to perform the analyses. To determine the analytical precision, certi ed reference materials (bush, NCS DC 73349; chicken, NCS ZC 73016; and bovine muscle powder, RM8415) were tested with the samples.
Statistical analyses (Fig .2 -stage 5) We used a series of analyses to explore the relationships between the species composition, diversity and nutritional quality of pollen and its in uences on female and male bee tness.
For the pollen pools used in the feeding experiment, we calculated Fisher's alpha as the index of diversity and Dominance D (i.e. 1 minus the Simpson index; ranging from 0 (all taxa are equally present) to 1 (one taxon dominates the community completely)) 57-59 as the index of dominance, calculated using PAST 4.05 60,61 . Fisher's alpha was used because it is suitable for a data set characterized by a high number of rare species, which was the case in the present study (see Supplementary Table S1). We investigated differences in nutritional quality (i.e., the stochiometric phenotype) between pollen pools and correlated the nutritional quality (i.e., the stochiometric phenotype) of the pollen pools with their species composition using redundancy analysis (RDA). RDA was performed only on the pollen species that contributed the most to each pollen pool (i.e., the concentration of pollen grains in a speci c pollen pool had to be higher than 20%) (CANOCO 5 62 ).
To consider bee tness, we rst analyzed the entire dataset for mortality (Fig. 2 -stage 5). To maintain ecological relevance, we assessed the differences in mortality using the following procedure (chi-square test, p < 0.05, Statistica 13): to study the effects of pollen species composition/diversity on bees via nutritional scarcity we rst compared, for each sex separately, the results for control-Osmia (1) with those for the other pollen pools that were not supplemented with nutrients, i.e., control-Osmia vs. (Apis1, Apis2, -ZnCu1 + H2O, -ZnCu + H2O, -KP + H2O, and -Na + H2O) and (2) with those for the other pollen pools that were fully supplemented with nutrients, i.e., control-Osmia vs. (-CuZn1 + CuZn, -CuZn2 + CuZn, -KP + KP, and -NaP + NaP). Second, to investigate the power of the effect of nutritional scarcity, we compared, for each sex separately, every nutrient(s)-scarce treatment with the corresponding nutrient(s)-supplemented treatments. Third, to assess the sexual differences in the impact of the scarcity of speci c nutrients on tness, we compared the results for females from each treatment to the results for males from the same treatment.
In the second step of considering bee tness, we analyzed the remaining life history traits (cocoon development, cocoon mass, and adult body mass) measured in the specimens that survived the feeding experiments. Before performing detailed analyses, we checked whether and to what degree the measured traits were correlated. To that end, we calculated the Spearman's R coe cient for the whole dataset for all traits for both sexes separately (adult mass vs cocoon mass, adult mass vs cocoon development and cocoon mass vs cocoon development; Statistica 13). Since the measured life history traits were not strongly correlated, in the following steps, we investigated how each trait was exclusively in uenced by the different diets used in the feeding experiment. We assessed cocoon development with Fisher's exact probability test (p < 0.05, VassarStats 63 ). The degree of cocoon development was assessed by qualitative analysis. Four stages of cocoon development were distinguished: (1) a fully developed cocoon that covered the whole bee body and consisted of a hard material; these cocoons were impossible to tear with bare hands but could be cut with a knife because they were su ciently hard; (2) an almost-developed cocoon that covered the whole bee body but consisted of a soft material; these cocoons were impossible to cut with a knife because they were too soft, but they could be torn by hand; (3) an underdeveloped cocoon that covered only part of the bee body and was soft; and (4) a very underdeveloped cocoon that did not cover any part of the bee body and consisted only of "woolly" matter. See supplementary Fig. S1 for details.
We performed statistical comparisons between the same groups as in the case of mortality but considered only groups having at least ve surviving specimens to avoid arti cial statistical results. For the data on the masses of the adult bees and their cocoons, we performed ANOVA (p < 0.05; Statistica 13).

Declarations Data availability
Source data are provided with this paper. All relevant data in this study are available from the corresponding author upon request.
Code availability Not applicable. Figure 1 Nesting biology of solitary Osmia bees.   Mortality of growing and developing bees fed on various diets. Nutritional scarcity and supplementation affected bee mortality in a sex-dependent manner. Alive vs Dead = number of specimens that either survived to the adult stage or died before maturation; N=25 per treatment and sex; chi-square test, p<0.05.

Figures
For all diets, the measured effects were compared to maintain the ecological relevance of the study: Control-Osmia was compared with every pollen pool that was not supplemented with nutrients as well as with the same pollen pool supplemented with all scarce nutrients for each sex separately; every nonsupplemented pollen pool was compared with its supplemented counterpart for each sex separately; and every treatment was compared between sexes. Statistically signi cant differences are indicated with red asterisks, and NS indicates no signi cant difference. Effect of diet on cocoon development. Zn, Cu and K scarcity negatively affected cocoon development in both sexes, but this effect was stronger on males than on females. For males, a similar negative effect of Na scarcity was also observed. Signi cant effects were determined according to Fisher's exact probability test, p<0.05 For all diets, the measured effects were compared to maintain the ecological relevance of the study: Control-Osmia was compared with every pollen pool that was not supplemented with nutrients as well as with the same pollen pool supplemented with all scarce nutrients for each sex separately; every nonsupplemented pollen pool was compared with its supplemented counterpart for each sex separately; and every treatment was compared between sexes. Statistically signi cant differences are indicated with red asterisks, minus signs indicate that a statistical comparison was not possible (too few replicates), and NS indicates no signi cant difference. The degree of cocoon development was assessed by qualitative analysis, and four stages of cocoon development were distinguished: (1) a fully developed cocoon that covered the whole bee body and consisted of a hard material; these cocoons were impossible to tear with bare hands but could be cut with a knife because they were su ciently hard; (2) an almost-developed cocoon that covered the whole bee body but consisted of a soft material; these cocoons were impossible to cut with a knife because they were too soft, but they could be torn by hand; (3) an underdeveloped cocoon that covered only part of the bee body and was soft; and (4) a very underdeveloped cocoon that did not cover any part of the bee body and consisted only of "woolly" matter. The effect of diet on adult mass and cocoon mass. ANOVA, p<0.05, calculated separately for each sex. A negative effect of Cu and Zn scarcity on body mass was observed for both sexes and was stronger in males than in females. In females, cocoon mass was also negatively affected by Cu and Zn scarcity, but Cu and Zn supplementation did not reverse the effect. K scarcity resulted in lower cocoon masses in both sexes, and in males, the effect was mitigated when the pollen was supplemented with K and P. For all diets, the measured effects were compared to maintain the ecological relevance of the study: Control-Osmia was compared with every pollen pool that was not supplemented with nutrients as well as with the same pollen pool supplemented with all scarce nutrients for each sex separately; every nonsupplemented pollen pool was compared with its supplemented counterpart for each sex separately; and every treatment was compared between sexes. Statistically signi cant differences are indicated with red asterisks, minus signs indicate that a statistical comparison was not possible (too few replicates), and NS indicates no signi cant difference.

Figure 7
Impact of pollen species available in the environment on the functioning of bee populations. (a) -diverse ora and (b) -poor ora. Stoichiometric mismatches experienced by bees feeding on unbalanced pollen