Introduction

Bees obtain almost all nutrients from pollen and nectar. Nectar primarily contains carbohydrates and chiefly consists of sucrose, glucose, fructose and water (Corbet 2003). Other macronutrients are mostly provided by pollen (Roulston and Cane 2000). It is the sugar content of nectar which is usually of primary interest of adults because sugar provides energy that fuels the individual’s activity or is used for provisioning the larvae (Corbet 2003). Pollen is primarily a non-energetic resource which contains proteins, lipids, vitamins and minerals primarily needed for larval growth and for the development of reproductive organs in adults (Plowright and Pendrel 1977; Nicolson 2011). Both, pollen and nectar, are essential for bees and one cannot be completely replaced by the other (Plowright et al. 1993).

The total amount of food ingested during larval development determines the size of an adult and depends upon the larval food supply (Sutcliffe and Plowright 1988). Larvae fed with larger amounts of food were heavier and their survival rate increased (Aupinel et al. 2005). Strong correlation between the weight of the larvae and the amount of ingested pollen (Ribeiro et al. 1993; Pereboom et al. 2003) and protein (Moerman et al. 2016) was determined. Offspring body size generally increases with an increase in the amount of food consumed during the larval stage, but the extent to which the body size is influenced by the type of food consumed is not precisely known (Roulston and Cane 2002), because the consumption of pollen and nectar in the colony is highly correlated (Moerman et al. 2016; Řehoř et al. 2014). The quantity of collected food is relatively easily measured in mass provisioners; therefore, mass provisioners are of frequent use for the quantification of maternal investment (Johnson 1988; Kapheim et al. 2011; Lawson et al. 2016; Radmacher and Strohm 2010). However, most studies do not analyze the food quality and also do not separate pollen from nectar.

Determining the exact amount of consumed pollen and nectar, and their ratio during a development of a single larva is problematic, especially concerning progressive provisioners such as honeybees and bumblebees that are most commonly bred. The larvae of these bees are fed continuously throughout their entire larval development. Nixon and Ribbands (1952) quantified the amount of sugar transferred from a worker honeybee to a larva by trophallaxis using radioactivity; however, they did not identify the overall amount of sugar consumed during development. Surprisingly, the amount of sugar consumed by honeybee larva has never been exactly measured and the estimation is evaluated from an average growth rate, average final weight and a proportion of sugar in food (Nelson et al. 1924; Jay 1963; Rortais et al. 2005; Hrassnigg and Crailsheim 2005; Brodschneider and Crailsheim 2010; Bishop 1961). Unlike in males, the amount of sugar consumed is measured in workers (Rortais et al. 2005). The amount of sugar consumed within a single male bumblebee larva was first established by an application of lanthanides (Řehoř et al. 2014). Řehoř et al. (2014) investigated the sugar consumption of larvae and the sugar distribution in bumblebee colony. The weight of the larvae increased accordingly to the amount of sugar consumed as reported by Řehoř et al. (2014). On average, 128 mg of sugar was consumed by a male larva during its development, thus 2.9 mg of the sugar received by the larva was equal to 1 mg of its dry body weight. The amount of sugar consumed by a worker honeybee larva is estimated to 59.4 mg (Rortais et al. 2005). It is expected that 98.2 mg of sugar is consumed during the development of a male; however, this number is estimated from the honeybee worker data and the weight of an adult male (Rortais et al. 2005).

The exact quantity of protein in pollen consumed and required by a single honeybee and bumblebee larva has not been determined yet. However, 4–6 mg of nitrogen is required to raise a worker larva during feeding experiments (Haydak 1970). Hrassnigg and Crailsheim (2005) calculated that the known quantity of nitrogen corresponds to 25–37.5 mg of protein and about 125–187.5 mg of pollen with 20% protein content. The rearing conditions of one male was also discussed by Hrassnigg and Crailsheim (2005) whose estimation in this case is 65–97.5 mg of protein.

From a qualitative point of view, the chemical composition of bee food has been shown to have major impacts on the colony and on the size of an offspring (Vanderplanck et al. 2014). Protein content is generally used to set the quality of pollen and it is highly variable among the different species of plants ranging from 2.5 to 61% of dry mass (Roulston et al. 2000). Amino acid composition determines the quantity of pollen required for the development of bees more than the actual protein content (Nicolson 2011). The presence or absence of some essential chemical components may have a major influence on the bumblebee colony development. Low quality of pollen (chemical composition, e.g., amino acid and sterol profiles) might negatively impact the colony growth rate (Moerman et al. 2016; Moerman et al. 2017), whereas high pollen quality accelerates the production of eggs and the growth of larvae (Roulston and Cane 2002; Tasei and Aupinel 2008; Moerman et al. 2016). Poor-quality diet with low nutritional content might weaken the health state of the colony. Feeding colonies with the low-quality pollen causes a production of smaller larvae with a decrease in larval and pupal masses, and an increased larval ejection (i.e., mortality) (Roger et al. 2017).

Adult bees and larvae performed better on higher sugar concentrations (Kaftanoglu et al. 2011). The high sugar content in the diet of honey bee larvae increases the number of queens and intercastes (Kaftanoglu et al. 2011). The glucose and fructose ratio is not crucial for the survival of the developing larvae and they are able to utilize both sugars effectively; however, it affects the live weights of the adult bees (Kaftanoglu et al. 2011).

High metabolic activity and relatively low energy stores have an influence on an occasional possible suffering from an unbalanced nutrient intake if the diet is provided in excess or in deficient amounts (Ruedenauer et al. 2015). During such periods, a colony could be potentially stressed by insufficient food intake (Couvillon and Dornhaus 2010). Bumblebees have an ongoing need for protein at the beginning of their life and throughout their lives, too. Pollen deprivation may tend to prolong the larval stage of all castes and affect the average size of individuals. Pollen-deprived colonies required a longer time to recover and they continued to produce smaller brood under improved conditions (Sutcliffe and Plowright 1988, 1990). Limited access to pollen during larval development directly affects body mass and reproductive quality also in honeybee males (Czekońska et al. 2015). Bumblebee workers and queens are able to perceive the nutritional status of larvae and adjust their behavior accordingly (Smeets and Duchateau 2001). If too little pollen is provided, they uncover the larvae and eject some of them out (Heinrich 1979; Tasei and Aupinel 2008). Experimentally starved larvae were fed significantly earlier and more often than non-starved larvae (Pereboom et al. 2003). Profile detectable from cuticular hydrocarbons differ between larvae fed ad libitum and starved larvae (Boer and Duchateau 2006). This indicates that larvae play an important role in the regulation of the feeding rate and are not passive participants only (Smeets and Duchateau 2001). However, the precise impact of a diet content on an individual growth rate remains poorly understood in progressive provisioners (Moerman et al. 2016). Experiments with mass-provisioning Meliponini suggest that insufficient pollen supplies can be partly replaced by an addition of more sugars (Quezada-Euán et al. 2011).

This paper aims to precisely measure the amount of sugars and protein consumed by a single bumblebee larva. Monitoring the pathway of food starting from sources up to the larvae of progressive provisioning bees requires an appropriate experimental approach and has never been done before, except in our previous research on sugar consumption in bumblebees (Řehoř et al. 2014). We employed lanthanide complexes that are inert and nontoxic towards animals (Řehoř et al. 2014) and humans (Merbach et al. 2013). The lanthanides are elements with very similar chemical properties displaying sensitivity to inductively coupled plasma techniques (ICP-OES, ICP-MS). These are analytical methods that make the precise quantification of elements possible (Řehoř et al. 2014). We have already successfully used lanthanides on bumblebees in complex with DOTA-1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid chelator. In this study, we used different chelator with similar properties (DTPA), due to its low cost. We also investigate the effects of monosaccharide sugars, disaccharide sugars and protein stress on the composition of food which is provided by queen to worker and male bumblebee larvae. The impact on larvae development is evaluated.

Materials and methods

Bumblebee colonies

For our experiments, artificially bred pollen-storer bumblebees Bombus terrestris were used (Agricultural Research, Ltd. Troubsko). Workers and queens of these species feed the larvae through holes made in a wax cover. Bumblebee larvae store feces from all consumed food in midgut until the lumina of the posterior part of intestines retires and the feces leave the digestive tract (Alford 1975). This happens during the final part of larval development which concerns the fourth larval instar around the time of the spinning of cocoon (Michener 2007). Feces are excreted and smeared over the cocoon’s inner surface (Michener 1974).

The hibernation of queens that artificially overwintered in a fridge (from January or February 2015) was terminated on the 7th of July 2015 and the queens were placed in boxes together with a sugar solution (Apiinvert) and pollen pellets. After 4 days, we moved them to separate 17 × 8.5 × 8 cm plastic boxes (PETRA plast Ltd). In total, we prepared five types of colonies. Each colony consisted of a queen and two male cocoons. The purpose of male cocoons was to initiate egg laying in the queen; therefore, after hatching, the males were immediately removed from the colony. Colonies were kept in the dark at 27 ± 2 °C and 60% relative humidity.

Experimental design

The experiment consisted of colonies fed with disaccharide (sucrose; S) or monosaccharides (fructose + glucose 1:1; FG), colonies marked by lanthanides (Z) and control colonies without lanthanides (K). Two control sets with 21 colonies each (SK1–SK21 and FGK1–FGK21) and two sets with 23 colonies each (SZ1–SZ23, FGZ1–FGZ23) were prepared. Additional sets of 23 colonies were prepared for an experiment with pollen stress (H) (FGH1–FGH23). The sugar solutions and pollen paste with chemical tracer (GdDTPA and DyDTPA) were placed in colonies SZ1–SZ23, FGZ1–FGZ23 and FGH1–FGH23. Sugar solutions and pollen paste without marking tracers were administered to control colonies (SK1–SK21 and FGK1–FGK21). FGZ colonies and FGH colonies were supplied with marked monosaccharide sugar solution, SZ colonies with marked disaccharide sugar solution, FGK colonies with pure monosaccharide sugar solution and SK colonies with pure disaccharide sugar solution.

A detailed study on an inert nature of lanthanide complexes is presented in our previous study (Řehoř et al. 2015). For this reason, we carried out a basic test of compatibility of results with control samples only.

Preparation of lanthanide complexes GdDTPA and DyDTPA

DTPA ligand (diethylene-triamine-pentaacetic acid) (5.895 g, 15 mmol, Sigma-Aldrich) was dissolved in distilled water (10 ml) in weighed glass flask. The pH of the mixture was adjusted to 5 with NaOH (5M) solution. GdCl3·6H2O (5.573 g, 15 mmol, Sigma-Aldrich) or DyCl3·6H2O (5.655 g, 15 mM, Sigma-Aldrich) was dissolved in distilled water (10 ml) in glass vial and slowly dripped to a flask with dissolved ligand. The pH of the mixture was very slowly adjusted to 8 (pH must be adjusted very carefully to avoid precipitation of lanthanide hydroxide) and the mixture was stirred for 12 h at room temperature. The pH was checked and readjusted to final 8 again. The solution was filled with distilled water to a final weight of 50.00 g to yield 300 mM solution.

Preparation of sugar solutions

Four types of sugar solutions were prepared once at the beginning of the experiment. FGZ sugar solution was prepared twice, once for FGZ colonies and once for FGH colonies.

A mix of anhydrous d-glucose (600 g) and d-fructose (600 g) (PENTA s.r.o.) or only sucrose (1200 g) (PENTA s.r.o.) was dissolved in hot water (792 g) for the preparation of FGZ and SZ sugar solutions. The solutions were stirred with a glass stick until all sugar was dissolved. Evaporated water from the solution was refilled after dissolving the sugar. Subsequently, 8 g of GdDTPA solution (~ 0.3 M) was added to each solution and thoroughly mixed. The resulting concentration of Gd in the feeding sugar mixture after the measurement of samples on ICP-OES was 186.97 mg/kg. Completed sugar solutions were poured into jars. Caps were taped with a parafilm and placed in a fridge.

Sugar solutions for the control colonies were prepared analogously and were marked with an equivalent amount of deionized water added instead of GdDTPA solution.

Preparation of the pollen paste

Approximately 3 kg of honeybee pollen pellets (a mixture of pollen from a local beekeeper) was ground in a mortar until the structure of pollen pellets was the same as the structure of loose pollen and thoroughly mixed.

In total, 17 g of DyDTPA solution (5.1 mmol) was dissolved in 595 g of deionized water and homogenized with 1800 g of prepared loose pollen. Mixing in a mortar was made 36 times independently in smaller volumes (50 g of pollen) and resulting pollen paste was continuously stored in a bowl located in the fridge. Homogenization was made in a mortar until the pollen paste was created (which took approximately 10 min). The total volume of the pollen paste was carefully mixed in the bowl at the end of the process again. Pollen paste was inserted in plastic vials, and caps were taped with a parafilm and stored in a freezer. Reference samples were taken from different pollen paste locations and then stored in plastic vials in a freezer as well. The resulting concentration of Dy in the feeding pollen paste after the measurement of samples on ICP-OES was 340.9 mg/kg.

The pollen paste utilized for the control colonies was prepared analogously with an equivalent amount of deionized water instead of the DyDTPA solution. The procedure of mixing loose pollen with water in a mortar was carried out 20 times. The total amount of loose pollen was 1000 g and total amount of deionized water was 340 g.

Nectar and pollen feeders

FGZ, SZ, FGK, and SK colonies were supplied with sugar solutions and pollen paste according to their needs. FGH colonies were reared with an unlimited access to the sugar solution and pollen paste until the seventh day after the first egg cell was established by the queen. Subsequently, FGH colonies had an unlimited access to the sugar solution and limited access to the pollen paste. The pollen paste was removed for 6 h at the same time every day. The duration of period without pollen paste was determined according to experience and published information on frequency of feeding of larvae (seconds to 3 h at maximum according to larval stage; Ribeiro et al. 1999) and ability of offspring number reduction under insufficient supplying of colony by food (ejection of eggs and larvae from cells; Heinrich 1979). The consumption of both nutrients in all colonies was monitored during the course of the experiment. Sugar was resupplied after a major amount of the previous dosage/batch was consumed, whereas the pollen paste was delivered fresh every other day. The sugar solution was supplied in 10-ml plastic feeder (Tachovske SPV Ltd) and the pollen paste in small plastic caps.

The course of an experiment and its termination

Queen right colonies were set and experiments were initiated on the 11th of July 2015. The initiation of cell formation by queens was supported by insertion of two male cocoons glued on a piece of carton. All of the colonies were observed regularly every day, and their development and progress were recorded (construction and abolition of brood cells, cocoon development, offspring emergence). Cocoons of the first brood generation were removed after the pupation of the last larva; the queens then have been used for an initiation of a second round of brood, i.e., “second brood generation”. Colony keeping was successfully terminated together with the queen after the pupation of the last larva from the “second brood generation.” In some cases, the colony keeping had to be terminated after the pupation of the last larva from the first brood generation, because queens died or did not lay eggs after the cocoons from the first brood generation had been removed from the colony. On the 21st of September 2015, all remaining colonies were terminated. If the queen died before the pupation of larvae from the first brood generation, the experimental colony was terminated and the results from an unsuccessful colony were not analyzed. The cocoons of the successfully terminated colonies were carefully handled, separated and reared independently in small plastic breathable containers until the emergence of the bumblebees from cocoons. Separation of cocoons from one cluster of cocoons was complicated and some of them were partly compressed, causing that some individuals did not emerge.

Body mass analysis of the bumblebee brood

Bumblebee offspring from all of the colonies were stored in a freezer until the body weight was determined. Individuals that failed to emerge or died in a cocoon were removed from the cocoon. Emerged individuals, unemerged dead individuals and farate imagoes were used for further analyses; larvae and pupae were discarded. The individuals were dried in an oven at 56 °C for 7–10 days according to a protocol from Řehoř et al. (2014).

Preparation of cocoons for Gd and Dy content analyses

Each bumblebee cocoon was placed into a glass vial; 7 ml of distilled water and 0.30 ml of Triton X-100 (Sigma–Aldrich) were added. The filled vial was heated until the mixture started to boil. The cocoon was disrupted using metal spatula and the mixture was thoroughly mixed and boiled for 10 min. 0.31 g of HNO3 (65%, Suprapur grade, Merck) was added to a new weighed polypropylene vial. Subsequently, the cold cocoon extract was taken out using a syringe (10 ml) with needle (0.80 × 50 mm, P-LAB a.s.) and filtered through a GF syringe filter (pore 1.0 µm, Chromservis s.r.o.) into the polypropylene vial. The empty glass vial from the sample was carefully washed with distilled water (~ 2 ml) and the residue was added to the rest of the sample through the same filter. Distilled water was then added to the solution to achieve final weight of 10.00 g where the final concentration of HNO3 is 2%. Such prepared samples were subjected to the ICP-OES analysis.

ICP-OES technique

The ICP-OES (inductively coupled plasma optical emission spectroscopy) measurements were performed using the SPECTRO ARCOS optical emission spectrometer (SPECTRO Analytical Instruments, Kleve, Germany) with a radial plasma observation. The spectrometer was equipped with Modified Lichte nebulizer and cyclonic spray chamber, and operated under following conditions: plasma power 1500 W, coolant gas flow 13.5 l/min, auxiliary gas flow 0.80 l/min, nebulizer flow 0.75 l/min, and sample aspiration rate 2 ml/min. Dy and Gd contents in the samples were measured simultaneously.

For calibration, commercially available single-element standard gadolinium, dysprosium and yttrium (nitrate) solutions (Analytika) were used. In all cases, 2.00 mg/L Y was used as an internal standard. Individual samples were also spiked with a final Yttrium concentration 2 mg/l. The Y II 324.228-nm and Y II 371.030-nm lines were used for an internal standardization. All calibration measurements were performed in 2.0% (w/w) HNO3 solution (prepared from 67% HNO3 solution, suprapur grade, Merck) and 0.5% Triton X-100 (laboratory grade, Sigma) for partial matrix matching. All calibrations and sample solutions were prepared or controlled using an analytical scale for weighing.

Analysis of protein and sugar compositions in pollen paste

The dry weight of the sample was determined gravimetrically by drying the sample (0.3 g) at 105 °C for 180 min in two parallel experiments.

The Kjeldahl’s method was used for the determination of the total crude protein content. The total nitrogen content was determined. Content of the crude protein was determined by a factor of 5.60 (Rabie et al. 1983). The determination was done in parallel with two input weightings (0.5 and 1.3 g) used for the mineralization of the sample.

Sugars in pollen pastes were analyzed in multiple steps. A sample of pollen paste (0.1 g) was dispersed in 100 ml of ultra-pure water (18.2 MΩ.cm) and sonicated for 10 min at the K10 device (Kraintek s.r.o., Slovakia). Subsequently, the sample was homogenized for 5 min at 9500 rpm using dispergator DI 25 (IKA, Germany). The samples were further diluted eight times, filtered through a 0.45-µm nylon filter and used for an analysis. High-performance anion-exchange chromatography with a pulse amperometric detection (HPAEC/PAD) DX-600 (Dionex, USA) was used for the analysis, a CarboPac PA1 column (2 × 250 mm) with a pre-column guard (Dionex, USA), mobile phase: the first 30 min isocratic in 16 mM NaOH (own analysis), then 10 min column regeneration with 200 mM NaOH and subsequent re-equilibration in 16 mM NaOH for 15 min before further injecting, mobile phase flow rate 0.25 ml/min, temperature 25 °C, and detection: pulsed amperometric with gold electrode, sample volume injection 10 µl. The method was modified based on Dionex Application Sheet No. 46 (Anon 1997).

Average quantities of sugars, nitrogen, protein and dry matter found in a sample of pollen paste are listed in Table 1. The measurements were carried out in two parallel determinations. Average values are given in percentages with a standard deviation.

Table 1 Sugar, nitrogen and protein percentage contents in pollen paste and in dry matter of pollen paste
Full size table

Calculation of values from lanthanide concentrations

The amount of sugar (S1), pollen paste (PP), dry pollen paste (PD) and water (W) consumed by the larvae were calculated from the results of ICP-OES analysis:

$$\begin{aligned} {\text{S1}}\;{\text{(mg)}} & ={\text{1}}/{\text{concentration}}\,{\text{of}}\,{\text{Gd}}\,{\text{in}}\,{\text{sugar}}\,{\text{solution}}\,{\text{(mg/kg)}} \\ & \quad \times \,{\text{proportion}}\,{\text{of}}\,{\text{sugar}}\,{\text{in}}\,{\text{solution}}\, \\ & \quad \times \,{\text{measured}}\,{\text{Gd}}\,{\text{concentration}}\,{\text{in}}\,{\text{sample}}\,{\text{(mg/kg),}} \\ {\text{PP}}\,{\text{(mg)}} & ={\text{1}}/{\text{concentration}}\,{\text{of}}\,{\text{Dy}}\,{\text{in}}\,{\text{pollen}}\,{\text{paste}}\,\left( {{\text{mg}}/{\text{kg}}} \right) \\ & \quad \times \,{\text{measured}}\,{\text{Dy}}\,{\text{concentration}}\,{\text{in}}\,{\text{sample}}\,\left( {{\text{mg}}/{\text{kg}}} \right), \\ {\text{PD}}\,{\text{(mg)}} & ={\text{1}}/{\text{concentration}}\,{\text{of}}\,{\text{Dy}}\,{\text{in}}\,{\text{pollen}}\,{\text{paste}}\,\left( {{\text{mg}}/{\text{kg}}} \right) \\ & \quad \times \,{\text{proportion}}\,{\text{of}}\,{\text{dry}}\,{\text{matter}}\,{\text{in}}\,{\text{pollen}}\,{\text{paste}} \\ & \quad \times \,{\text{measured}}\,{\text{Dy}}\,{\text{concentration}}\,{\text{in}}\,{\text{sample}}\,\left( {{\text{mg}}/{\text{kg}}} \right), \\ {\text{W}}\,{\text{(mg)}} & =[{\text{1}}/{\text{concentration}}\,{\text{of}}\,{\text{Gd}}\,{\text{in}}\,{\text{sugar}}\,{\text{solution}}\,\left( {{\text{mg}}/{\text{kg}}} \right) \\ & \quad \times \,{\text{proportion}}\,{\text{of}}\,{\text{water}}\,{\text{in}}\,{\text{solution}} \\ & \quad \times \,{\text{measured}}\,{\text{Gd}}\,{\text{concentration}}\,{\text{in}}\,{\text{sample}}\,\left( {{\text{mg}}/{\text{kg}}} \right)] \\ & \quad +\,[{\text{1}}/{\text{concentration}}\,{\text{of}}\,{\text{Dy}}\,{\text{in}}\,{\text{pollen}}\,{\text{paste}}\,\left( {{\text{mg}}/{\text{kg}}} \right) \\ & \quad \times \,{\text{proportion}}\,{\text{of}}\,{\text{water}}\,{\text{in}}\,{\text{pollen paste}} \\ & \quad \times \,{\text{measured}}\,{\text{Dy}}\,{\text{concentration}}\,{\text{in}}\,{\text{sample}}\,\left( {{\text{mg}}/{\text{kg}}} \right)]. \\ \end{aligned}$$

Subsequently, we calculated the total sugar consumption (ST), protein (PR), and total dry food (TF) from data of pollen paste analysis:

$$\begin{aligned} {\text{ST}}\,{\text{(mg)}} & ={\text{S1}}\,+\,[{\text{proportion}}\,{\text{of}}\,{\text{sugar}}\,{\text{in}}\,{\text{pollen}}\,{\text{paste}} \\ & \quad \times \,{\text{consumed}}\,{\text{pollen}}\,{\text{paste}}\,{\text{(mg)}}], \\ {\text{PR}}\,{\text{(mg)}} & =[{\text{proportion}}\,{\text{of}}\,{\text{protein}}\,{\text{in}}\,{\text{pollen}}\,{\text{paste}} \\ & \quad \times \,{\text{consumed}}\,{\text{pollen}}\,{\text{paste}}\,{\text{(mg)}}], \\ {\text{TF}}\,{\text{(mg)}} & ={\text{S1}}+{\text{PD.}} \\ \end{aligned}$$

We used only the amount of sugar and protein values for most statistical analyses, because other values are strongly correlated. All values are used for basic statistics.

Statistical analyses

All of the statistical analyses were carried out using the R program (R Core Team 2014). Data were tested using linear mixed models. We used the nlme package for fitting models and the package car for obtaining the significance of the results. Nest identity was used as a random factor.

We did not use data concerning males for the analyses except the basic summarization of consumed pollen/protein and nectar/sugar due to the small overall number.

Results

Influence of lanthanide marking and sugar type on individual body weight

Weight of the bumblebee feeding with lanthanide-labeled diets was not affected in any individual regardless of the sex (workers, males) (Table 2; Fig. 1a, c). We also found a non-significant effect of a sugar type on body weight of bumblebees (Table 2; Fig. 1b, d) with which the colonies were fed (disaccharide or monosaccharides). Similarly, there was no significant effect of an interaction between these parameters on the body weight of individuals (Table 2).

Table 2 The effects of lanthanide marking, type of consumed sugar and interaction between both effects on the body dry weight of the individuals
Full size table
Fig. 1
figure 1

Non-significant effect of lanthanide marking (a, c) and sugar type (b, d) on weight of males (c, d) and workers (a, b). Results of statistical analyses are shown in Table 2

Full size image

Influence of sugar type to worker features

The type of sugar did not influence any feature of the food consumed by a worker larva. We have not found any significant differences in the sugar consumption from the feeder (F = 1.6381, p = 0.2133), pollen paste consumption from the feeder (χ = 0.7528, p = 0.3946), amount of protein contained in pollen paste (χ = 0.7528, p = 0.3946) or the total amount of consumed sugar (sum of sugar from feeder and sugar from pollen paste) (χ = 1.7556, p = 0.1982) according to the type of consumed sugar by worker larvae (disaccharide or monosaccharide). Also, the amount of sugar was not influenced by the amount of protein consumed per milligram of dry weight (F = 0.2587, p = 0.6159) and the amount of sugar consumed per milligram of body weight (F = 0.711, p = 0.4078) was not influenced by the type of consumed sugar.

Consumption of protein (pollen) and sugar

Males fed ad libitum during their development ingested 0.52 mg of protein and 4.43 mg of sugar, while workers fed ad libitum during their development ingested 0.54 mg of protein and 4.26 mg of sugar per 1 mg of dry body weight on average. Food-stressed worker larvae ingested 0.51 mg of protein and 4.65 mg of sugar per 1 mg of adult body weight on average. Final weights and exact amounts of consumed sugar, pollen and protein for each group (males, workers fed ad libitum, food-stressed workers) are listed in Table 3.

Table 3 Dry body mass and consumption of food components
Full size table

Relationship between the characteristics of consumption and body weight

The amount of consumed sugar and pollen was strongly correlated (F = 851.1920, p ≤ 0.0001, Fig. 2c). The resulting weight of the workers significantly grows with the amount of sugar ingested during their development (F = 16.037, p ≤ 0.0001, Fig. 2b) and protein (F = 485.090, p ≤ 0.0001, Fig. 2a); we found a significant influence of an interaction between the consumed amount of protein and the amount of consumed sugar on the weight of a worker (F = 10.114, p = 0.0017). The influence of consumed sugar on body weight was more steep and pronounced in workers which consumed a smaller amount of protein than in workers which consumed a large amount of protein during their development.

Fig. 2
figure 2

Relationship between protein (a) and sugar (b) with final body mass of workers and correlation of ingested protein and ingested sugar (c) (gray spots = sucrose diet, black spots = fructose and glucose diet)

Full size image

We found a significant influence of the total amount of consumed protein on the amount of protein consumed per 1 mg of body weight (F = 25.7747, p ≤ 0.0001), and the total amount of consumed sugar on the amount of consumed sugar per 1 mg of body weight (F = 6.4758, p = 0.0117). Worker larvae which consumed larger amount of protein or sugar need a larger amount of these nutrients for 1 mg of adult body weight. There was also a significant relationship between the total adult body mass and the amount of protein consumed per 1 mg of body weight (F = 16.763, p = 1e−04), and between the body weight and the amount of sugar consumed per 1 mg of body mass (F = 51.1531, p ≤ 0.0001). Heavier workers consumed a smaller amount of protein and sugar per 1 mg than lighter workers. These findings are not in contradiction. One finding suggests higher food effectivity of individuals which consumed less food, but second finding suggests higher food effectivity of heavier individuals. Both relationships with the amount of consumed food per milligram of body weight were weaker than the relationship between the total amount of consumed protein and sugar.

Consumption of pollen and sugar in food-stressed workers

We did not find a significant effect of food stress on any feature of worker ingestion when we tested these features alone. There was no effect of food stress on the amount of consumed sugar from feeder (F = 0.0936, p = 0.7625), the amount of consumed pollen from feeder (F = 2.5444, p = 0.1249), total amount of consumed protein (F = 2.5444, p = 0.1249), total amount of consumed sugar (F = 0.0018, p = 0.9666), amount of consumed protein per 1 mg of body weight (F = 2.1085, p = 0.1606) and the amount of consumed sugar per 1 mg of body weight (F = 1.1657, p = 0.292).

However, we acquired a different pattern, when we tested more factors with the effect of food stress in model. The effect of interaction of the total amount of consumed protein and interaction between total amount of consumed protein and effect of food stress are significant, when both tested factors influenced the total amount of consumed sugar in one analysis (Table 4a). When we tested factors influencing the total amount of ingested protein, all main factors (food stress and the amount of ingested sugar) and also their interaction were significant (Table 4b). Significant interaction means that there is a different relationship between the amount of ingested pollen and sugar in food-stressed and well-fed workers. The amount of ingested sugar grew steeper with the amount of consumed protein in food-stressed workers than in well-fed workers (Fig. 3c).

Table 4 Factors influencing total amount of consumed sugar and protein
Full size table
Fig. 3
figure 3

Influence of ingested protein (a) and sugar (b) on adult body mass of food-stressed workers and relationship of ingested protein and ingested sugar (c) (gray spots = ad libitum-fed worker larvae, black spots = food-stressed worker larvae)

Full size image

Body weight was not influenced by food stress (Table 5). On the other hand, body weight was influenced by the amount of consumed pollen and also sugar (Table 5; Fig. 3a, b). Moreover, we found a significant interaction between food stress and the amount of consumed sugar to adult body weight; the growth of the body weight was less steep with the amount of consumed sugar in food-stressed workers than in well-fed workers.

Table 5 The effects of food stress and ingested protein and sugar on body mass of workers (protein.total = amount of protein, sugar.total = amount of sugar, starvation = effect of food stress)
Full size table

Discussion

Food tracking by lanthanide mark

The total amount of consumed food and components of food by an animal is often difficult to quantify. Only the indirect methods are usually used (Kapheim et al. 2011; Lawson et al. 2016; Radmacher and Strohm 2010). Inventing a direct method which precisely quantifies total amounts of consumed food can be very useful for the quantification of food efficiency (Waldbauer 1968), influence of a focal organism to an environment, or the costs of parental care. In aculeate Hymenoptera, the main focus of research is influenced by the maternal manipulation contributing to the evolution of eusociality (Alexander 1974; Brand and Chapuisat 2012; Kapheim et al. 2011). Relevant results concerning the consumed food by an offspring can be obtained by weighting and composition analysis of provisioned mass in mass-provisioning species (Kapheim et al. 2011; Lawson et al. 2016). However, in progressive provisioners it is impossible to measure the amount of food or food components easily. Chemical marker can, therefore, enable the quantification of the total amount of ingested food. These features are important for the marker used in our study: regular application in food source, possibility of the exact measuring, accumulation of all amount of marker at one place (usually in a body of an organism) and non-toxicity. Some studies used radioactive markers for measuring the amount of consumed food or determining food components (Ashford et al. 2000; Nixon and Ribbands 1952); however, radioactivity is harmful for a research object as well as for researchers and offers only one type of signal.

Complexes of lanthanides are good candidates for useful markers used in quantifying the amount of ingested food in Hymenoptera. We were able to evenly dissolve lanthanide complexes in two types of sugar solutions and also in protein paste. Accurate measuring is possible on ICP-MS or ICP-OES. Lanthanide complexes are completely excreted from the body to cocoon and the measurement is possible thanks to the delayed excretion in Hymenoptera. Non-toxicity of lanthanide complexes was confirmed in previous study (Řehoř et al. 2014), which used the DOTA chelator to mark the sugar solution in bumblebee microcolonies. In this study, we used lanthanide complexes with DTPA chelator which is extremely similar to the DOTA in chelating properties, however, can be purchased at much lower cost. The DTPA substance or lanthanide complex with DTPA can also be freely purchased. Like DOTA lanthanide complex, diets labeled with DTPA lanthanide complex did not significantly influence the weight of individuals regardless of the sex (males, workers) and we did not notice any differences in survival and the number of brood of bumblebees. Our study confirmed the non-toxicity of DTPA lanthanide complex for insects according to our methodological approach.

Total amount of ingested protein and sugar

Using our methodology, we were able to precisely quantify the amount of sugar and pollen, and hence the protein required for the development of males and workers in case of ad libitum feeding. We found out that males fed ad libitum during their development consumed 4.4 mg sugar per 1 mg of their adult body weight which is much higher than the previously published 2.9 mg per 1 mg of the body weight (Řehoř et al. 2014). Most of this difference is made by the measurement of natural sugar included in the pollen paste in current research. Sugar content already in pollen pellets brought by workers is often ignored and is not counted in the food composition of corbiculate bees including honeybees (Rortais et al. 2005). Thus, this is the first study that measures all sugar in the food of larvae of progressively provisioning bees. The amount of sugar in ad libitum-fed workers was similar to that measured in males. We also found out that workers fed ad libitum during their development ingested 0.54 mg of crude protein per 1 mg of the body weight, similarly to larger males (0.52 mg).

Workers consumed 121.4 mg of pollen paste on average which is 72.4 mg of dry pollen paste and 17.3 mg of protein. It is comparably lower than the amount of pollen and protein consumed by honeybee workers (Hrassnigg and Crailsheim 2005). Our values can be slightly underestimated because we used gentle rinsing methods that produce around 9% of lanthanide loss during processing (Řehoř et al. 2014), and the data on honeybees can be overestimated, because the conversion coefficient from nitrogen to protein was not presented (Hrassnigg and Crailsheim 2005).

There was a strong positive correlation between the body weight, total amount of protein ingested, and the total amount of sugar ingested in all our experiments. Therefore, larger workers ate more proteins and also more sugar than smaller workers during their development. These results are consistent with Aupinel et al. (2005) findings for honey bee, where larvae fed with larger amounts of food were heavier. Similarly, Ribeiro et al. (1993) and Pereboom et al. (2003) found a strong correlation between the weight of an individual and the amount of pollen and protein (Moerman et al. 2016) ingested. Also, a high correlation between the consumption of pollen and nectar in the colony was found in previous studies (Moerman et al. 2016; Řehoř et al. 2014).

The amount of consumed protein during development per 1 mg of adult body weight was positively correlated with the total amount of consumed protein during development. Similar pattern was found for the amount of consumed sugar. Therefore, we can suggest that worker larvae that consumed larger amount of food use this food less effective. On the other hand, the negative association between the total amount of consumed protein during development per 1 mg of adult body weight and adult body weight of workers was found. And again the same pattern was found for sugar. Therefore, heavier workers spend their nutrients more efficiently than lighter workers, but individuals that are same in weight can differ in efficiency in case of different amount of food consumed.

Effect of sugar type

We compared the body weight, the total amount of ingested sugar and the total amount of ingested pollen in colonies which were fed with two types of substrates—sucrose and the mixture of glucose and fructose in ratio 1:1. We did not find a significant difference between diets consisting of different sugar types in any measured feature. Also, the relationship between the body mass, the total amount of ingested pollen and the total amount of ingested sugar was the same for both types of sugar. Therefore, we can conclude that monosaccharides (glucose and fructose) and disaccharides (sucrose) are a fully equivalent food source for bumblebees.

Our findings are in contradiction with an assumption that disaccharide-fed larvae might need more energy to split the disaccharide double bond which could be reflected in a different weight of individuals. This pattern was found in honeybee. Honeybee brood were able to utilize both sugars fructose and glucose effectively; however, the ratio of both sugars determines the live weights of adult bees (Kaftanoglu et al. 2011). However, the effect of sugar types could be obscured by high variability of individual body mass in our results.

Effect of food stress

The abundance of food sources influences the size of an offspring in aculeate Hymenoptera. In case of low abundance of food sources, mother produces smaller offspring (Rosenheim et al. 1996; Kim 1999; Alcock et al. 2005; Peterson and Roitberg 2006) or invests more in cheaper sex (Kim 1999; Peterson and Roitberg 2016; Rosenheim et al. 1996; Strohm and Linsenmair 1997).

Bees use two main sources of nutrients (pollen and nectar) and one of them can be limited. Then, some larvae in the colony may be exposed to food stress. Workers or queen can respond to food stress by increasing the feeding of larvae by the food source that is currently available. Likewise, the pollen of poor quality may lead to a higher need of sugar supplies to the colony (Vanderplanck et al. 2014).

We did not find an effect of food stress on the total body weight of bumblebee workers. We also did not find a direct effect of food stress on the total amount of consumed pollen or sugar, and the amount of consumed pollen and sugar per 1 mg of body weight. This finding is quite surprising. A high variance in adult worker sizes in combination with relatively low pollen stress can be the explanation. On the other hand, we show that some food source optimization mechanism can exist on the colony level.

We found out, that the body weight, the total ingestion of protein and the total ingestion of sugar were highly correlated in well-fed and also in food-stressed workers. However, the relationship between these parameters differs in food-stressed and well-fed colonies. Small workers consumed during development a similar amount of pollen and nectar in well-fed colonies and also in starved colonies. On the other hand, large workers consumed a lower amount of protein and higher amount of sugar during development in food-stressed colonies in comparison with well-fed colonies. Also, the relationship between the total amount of ingested protein and ingested sugar was different in well-fed and food-stressed worker larvae. The ingestion of sugar grew steeper with the ingestion of protein in food-stressed colonies than in well-fed colonies. Although we did not prove an effect of food stress on average values of worker weights or the total amount of food consumed, we showed that there was a different pattern in ingestion of protein and sugar between well-fed and food-stressed colonies.