Statistical Models for Assessing the Influence of Hygienic Behaviour of Worker Bees on the Level of Lysozyme and Total Protein Content in Their Hаemolymph

Iranian Journal of Applied Animal Science

مقاله 20، دوره 10، شماره 2، شهریور 2020، صفحه 365-373   

نویسندگان

S. Bogdanov Lazarov^{* 1}؛ P. Veleva²؛ I. Zhelyazkova¹

¹Department of Animal husbandry–Non Ruminants and other Animals, Faculty of Agriculture, Trakia University, 6000 Stara Zagora, Bulgaria

²Department of Informatics and Mathematics, Faculty of Economics, Trakia University, 6000 Stara Zagora, Bulgaria

چکیده

The hygienic behaviour of 26 bee colonies of apiaries in different regions of Bulgaria was tested. The degree of expression of hygienic behaviour was evaluated by the pin-killing method. A test field of square 5 x 5 cm was stuck onto a section of a honeycomb with a sealed worker brood. The bee colonies are divided into two groups: hygienic (cleaned over 95% of cells in the testing area on the 48^th hour) and nonhygienic (cleaned less than 95% of the cells in the testing area on the 48^th hour). Haemolymph has been taken from bees from each bee colony at 48^th hour and lysozyme levels and total protein content has been defined. Multivariate analysis (ANOVA) was used to determine significant differences between hygienic and nonhygienic colonies. The results obtained show significant differences between both groups (p <0.05) according to percent cleaned cells after killing the brood at 3^rd, 24^th and 48^th hours and insignificant differences between hygienic and nonhygienic colonies (p <0.05) for both parameters-lysozyme and total protein. Bivariate correlation was applied to investigate the impact of hygienic behaviour on lysozyme levels and total protein content. Low negative correlation (r=-0.33) between total protein and hygienic behaviour was established which means that an increase of the activities related to the cleansing of honeycomb cells leads to a decrease in the total protein content. Low correlation (r=0.35) between hygienic behaviour and lysozyme levels was found i.e. higher levels of lysozyme are reported with more pronounced hygiene behaviour. The calculated regression model for lysozyme (P=0.24826) is statistically insignificant but the regression model for total protein is statistically significant (P=0.003153) and can be used to predict the relation between total protein content in the haemolymph and hygienic behaviour of bee colonies.

کلیدواژه‌ها

honeybees؛ hygienic behaviour؛ lysozyme level؛ total protein

اصل مقاله

INTRODUCTION

Honey bees have existed for millions of years and in the process of evolution they have built up mechanisms for survival depending on the changing environmental challenges. A major mechanism that protects individual bee specimens is the bee colony. Tens of thousands of worker bees, from a few hundred to a few thousand drones and queen bee, coexist in it. As a separate biological unit, the bee colony exists due to the unity and harmony between the individual specimens. This synchronization is accomplished through conditional, unconditional reflexes and instincts (Nenchev and Zhelyazkova, 2010; Comman et al. 2012; Haritonov, 2015). One of the important behavioral reactions in honeybees, which is congenital and is passed on to the offspring, is "hygienic behaviour". Hygienic behaviour is a natural mechanism for disease control and is related to the detection and removal of infected and dead larvae and bees outside the hive. In this way, "hygienic" bee colonies clean their nests well ahead of the appearing disease's signs. Rothenbuhler (1964) found that two recessive genes control cell production and the removal of infected pupae. Later, Taber (1982) reported that one of them determines the removal of dead larvae from the beehives and called it "removing." It is marked with the "rr" symbol. The other one determines the uncapping of the cells and is called "uncapping", marked with the "uu" symbol. More recent studies show that hygienic behaviour is a quantitative feature and its expression in bees can be affected by six (Oxley et al. 2010) or seven quantitative loci (Lapidge et al. 2002). Natural immunity in bees largely depends on the content of antimicrobial peptides in their haemolymph (Glupov, 2004). Its protective function includes cellular and humoral barriers (Gillespie et al. 1997). The amount of antimicrobial peptides increases with injury and bacteria entering the haemolymph (Aronstein and Saldivar 2005; Aronstein et al. 2010; Ilyasov et al. 2014). Glinski and Jarosz (2001) found that bacteria swallowed by bees are digested by the released hydrolytic enzymes, predominantly lysozyme (N-acetylmyramylhydrolase), alkaline phosphatase, ribonucleases, and phospholipases. A number of authors believe that lysozyme content of the haemolymph increases the protective functions of the bee organism. According to them, the higher levels are influenced by some factors such as bee age, way of rearing the bee colonies, protein content in food, supplementary feeding of bee colonies with stimulating products, presence of pathogenic microorganisms and parasites, etc. (Zyuman et al. 1988; Gechev, 1995; Nagornaya et al. 1996; Kanchev et al. 1997; Zhelyazkova and Gurgulova, 1997; Zhelyazkova and Gurgulova, 2000; Gurgulova et al. 2001; Darkazanli, 2008; Shumkova, 2016). The analysis of the available literature shows that correlations have been established between the hygienic behaviour and the productivity of bee families (Guler and Hakan, 2013; Mansourizalani et al. 2018), between the hygienic behaviour and the degree of infection with Varroa destructor (Nganso et al. 2017). Lysozyme and total protein in bee haemolymph are considered to be factors of immunity, and hygienic behaviour is a factor for disease resistance. In this regard, it is of interest to develop statistical models to determine the relationship between hygienic behaviour, lysozyme levels, and total protein content in worker bee haemolymph. The objective of the study is to explore the degree of influence of hygienic behaviour of worker bees on the level of lysozyme and the quantity of total protein in their hаemolymph and to develop statistical models that can be used to predict the relation between investigated parameters.

 

MATERIALS AND METHODS

Testing bee colonies for the degree of expression of hygienic behaviour

Testing about the degree of expression of the hygienic behaviour of bee colonies and the analysis of the haemolymph obtained from them were conducted during the active beekeeping season from May to September. Bee colonies with bees from the Local Bulgarian bee (Apis mellifera) were used for studying. Twenty-six bee colonies from apiaries in different regions of Bulgaria shown at Figure 1 were tested for expression of hygienic behaviour. Haemolymph for analysis was obtained from the tested bee colonies. Testing for the degree of expression of hygienic behaviour was carried out in bee colonies equal by strength, quantity of sealed worker brood and food stock. The modified method of Gurgulova et al. (2003), similar to the method by Petrov (1997) has been applied. Honeycombs with the largest area of sealed worker brood have been selected. The test field on the combs has been marked by using a square template measuring 5 × 5 cm (100 worker bee cells), which is placed in a section of the comb with continuous worker brood. The brood within the borders of the template was killed by using a thin entomological needle by piercing the caps of the sealed cells without breaking them. Empty cells within the test area have been reported in advance and excluded from the test field. The cells uncapped and cleaned by the bees were counted on the 3^rd, 24^th and 48^th hours of the piercing (Darkazanli, 2008; Gramacho and Gonçalves, 2009).

 

Obtaining haemolymph and establishing the lysozyme levels and total protein content in it

From each bee colony young worker bees were taken (200-250 pieces), from which haemolymph (cumulative samples) was obtained. The haemolymph was extracted by a Pasteur pipette at the border between the II^nd and III^rd abdominal tergite. The lysozyme content in the haemolymph was determined by the method of Motavkina et al. (1979), modified by Kostov and Bonovska (1983). Live culture of Micrococcus lisodeicticus was used in the assay, which is lysed specifically by the lysozyme. To determine the total protein quantity in the haemolymph samples the Biuret method was used.

 

Figure 1 Geographical location of tested apiaries in the Republic of Bulgaria

 

 

Laboratory assays to determine the lysozyme and total protein content in the studied haemolymph samples were carried out at the National Reference Laboratory "Bee Health" at the National Diagnostic Research Veterinary Medical Institute – Sofia, Bulgaria.

 

Statistical processing of the data

The data obtained were processed variationally and statistically with software STATISTICA 12, Copyright ^© Stat Soft Inc. 1984-2014 (StatSoft, 2014). To determine the presence of significant differences between bee colonies with hygienic or nonehygienic behaviour Multivariate ANOVA with Hotelling and Wilks Tests for the 3^rd, 24^th and 48^th hours was applied. The same statistical analysis was used to define significant differences between hygienic and nonehygienic groups based on lysozyme levels and total protein content measured at 48 h. Correlation analysis was used to determine the strength and direction of the relation between the levels of lysozyme in the haemolymph and the hygienic behaviour of worker bees and between the total protein content in the haemolymph and the hygienic behaviour of bees as well. To assess the degree of influence of hygienic behaviour of worker bees on the level of lysozyme and the quantity of total protein in the hаemolymph, regression models have been developed.

 

RESULTS AND DISCUSSION

Table 1 presents the results from the conducted Multivariate ANOVA tests on bee colonies for the degree of expression of hygienic behaviour. From the total of 26 bee colonies included in the study 14 of them (53.85%) have been determined as colonies with a high degree of hygienic behaviour (hygienic) and 12 of them (46.15%) as colonies with a low degree of hygienic behaviour (nonhygienic), respectively. As can be seen from Table 1 bee colonies from the hygienic group had uncapped and cleaned on the 3^rd hour an average of 2.20 ± 0.32% of the cells with killed brood within the template area. In the 3^rd–24^th hour range these bee colonies had cleaned over 90% (91.24%) of the total number of uncapped and cleaned cells established detected on the 48^th hour (98.83±0.23%). The analysis of the results concerning the percentage of uncapped and cleaned cells from the group of non-hygienic bee colonies shows that on the 3^rd hour they had cleaned an average of 0.63 ± 0.33%, which is 1.57% lower than the hygienic group at that time. On the 24^th and 48^th hour, bee colonies from this group had uncapped and cleaned an average of 69.04 ± 1.99% and 86.28 ± 1.09%, which is, respectively, 23.33% and 12.55% less than the hygienic group. In the period from the 3^rd to the 24^th hour, bees from colonies with a low degree of hygienic behaviour had cleaned less than 70% (68.41%) of cells, which compared to the number recorded on the 48^th hours is 79.29%. It was found that for the indicated period (3^rd-24^th hour) the nonhygienic bee colonies had cleaned 11.95% fewer cells compared to the group of hygienic colonies. Based on the results obtained, it can be concluded that the hygienic bee colonies recognize at an earlier stage the killed brood.

 

Table 1 Results from testing bee colonies for the degree of expression of hygienic behaviour^{1, 2, 3, 4}

 

¹ Multivariate tests of significance for levels of expression of hygienic behaviour (ANOVA).

² Sigma-restricted parameterization.

³ Effective hypothesis decomposition.

⁴ n= 26.

* (P<0.05).

SE: standard error and SD: standard deviation.

 

According to Spivak and Downey (1998), the presence of a dead brood provokes the hygienic behaviour of worker bees. Bee colonies with a high degree of expression of hygienic behaviour quickly discard the dead larvae and pupae from the beehive well before signs of disease appear, which contributes to the resistance of bees to pathogenic microorganisms and ectoparasites (Nemkova, 2004). A similar opinion is shared by other authors (Palacio et al. 2010). Table 1 shows the significant differences between bee colonies with different levels of hygienic behaviour. Based on the Hotelling and Wilks tests, a statistically significant difference (P≤0.05) was found between the groups of hygienic and non-hygienic bee colonies at the 3^rd, 24^th and 48^th hours, separately. Table 2 shows the results of lysozyme levels in the haemolymph of worker bees with different degrees of expression of hygienic behaviour. A higher level of lysozyme in the worker bees’ haemolymph has been found in the group of nonhygienic colonies (13.71±1.76 μg/mL). In this group, a wide variation in terms of the minimum and maximum value of the studied trait has been reported, 3.13 µg/ml and 38.28 µg/mL, respectively. The mean value of lysozyme content in the haemolymph of bees in the group with a high degree of expression of the hygienic group is 10.33 ± 0.93 μg/mL, which is 3.38 μg/mL less than the value of this indicator in the nonhygienic bee colonies. As can be seen from Figure 2 the differences between hygienic and nonehygienic group for lyzozyme levels in bees’ haemolymph based on Hotelling and Wilks tests are statistically insignificant (P=0.11613<0.05). A probable cause for the higher lysozyme levels in the nonhygienic bee colonies is due to the fact that they detect the presence of the dead brood at a later stage. This, in turn, leads to the activation of humoral protective immunity factors, one of which is lysozyme. According to some authors, this enzyme is not the only factor of immunity, but it is important in initiating the synthesis of peptides and low molecular proteins with bactericidal activity (Nagornaya et al. 1996; Wilson-Rich et al. 2008; Gätschenberger et al. 2013). The obtained values of the studied parameters lysozyme level in the haemolymph and degree of hygienic behaviour of worker bees reflected on Figure 3 have distribution close to the normal distribution with mild left asymmetry at P < 0.05, i.e. the test for normality of distribution is statistically reliable. This is a prerequisite for the results of the correlation and regression analyses of the relation between the lysozyme level and the hygienic behaviour of worker bees to be considered common and applicable to all bee colonies from the studied population. The correlation analysis performed (Figure 3) has established the strength and direction of the relationship between lysozyme levels in the haemolymph and the hygienic behaviour of worker bees. The results show a weak relation between the studied parameters (r=0.351) i.e. higher levels of lysozyme are reported with more pronounced hygiene behaviour. The results from correlation analysis are evidenced by the regression analysis presented in Table 3 and Figure 4. The resulting regression model is statistically insignificant (P>0.05). The value of the determination coefficient R²= 0.01662784 (Table 3) in turn indicates that only 1.7% of the variations in the lysozyme levels studied are accounted for by the degree of hygienic behaviour of the bees. The insignificance of the model is a reason to continue the assay with a larger data set or to further study the nature of the dependence between the variables. So far, no similar study has been conducted in Bulgaria on the influence of worker bees’ hygiene behaviour on lysozyme levels and total protein content in their haemolymph. In the studied literature the authors hadn't found a similar analysis on these parameters. Table 4 presents the results of total protein content in the haemolymph of worker bees in colonies with high (hygienic) and low (nonhygienic) degree of expression of hygienic behaviour. In the group of nonhygienic colonies, the value of this indicator is higher (22.63±2.97 g/L), which is 3.58 g/L higher than that found in the group of hygienic bee colonies (19.05±1.95g/L).

 

Table 2 Amount of lysozyme in the haemolymph of worker bees from bee colonies with different level of expression of hygienic behaviour

 

* (P<0.05).

SE: standard error and SD: standard deviation.

 

Figure 2 Multivariate ANOVA analysis for significant differences between hygienic and nonehygienic groups based on lysozyme levels and total protein content

 

 

Figure 3 Histogram of distribution and correlation between lysozyme levels in haemolymph and hygienic behaviour of worker bees

 

 

Table 3 Regression model showing the dependence of lysozyme levels in haemolymph from the hygienic behaviour of worker bees¹

 

¹ Regression summary for dependent variable: lysozyme, mg/mL; R= 0.12894898, R²= 0.01662784; Adjusted R²= 0.00433569; F (1; 80)= 1.3527; P

* (P<0.05).

 

Figure 4 Regression model showing the dependence of lysozyme levels in haemolymph and hygienic behaviour of worker bees

 

 

Table 4 Amount of total protein in the haemolymph of worker bees from bee colonies with different level of expression of hygienic behaviour

 

* (P<0.05).

SE: standard error and SD: standard deviation.

 

Figure 5 Histogram of distribution and correlation between a total protein in haemolymph and hygienic behaviour of worker bees

 

 

Larger variation in minimum - maximum limits are reported in colonies with a low level of expression of hygienic behaviour (min=3.18 to max=71.04 g/L), compared to variations in colonies with a high level of expression (min=1.20 to max=43.78). The reported differences between hygienic and nonhygienic group for total protein content calculated by Hotelling and Wilks tests (Figure 2) are statistically insignificant (P<0.05). Concerning the total protein content in the haemolymph and the degree of hygienic behaviour of the worker bees (Figure 5), the results show that the data are close to the normal distribution with a small asymmetry on the left, i.e. the distribution normality test is statistically reliable at significance level P < 0.05. This again gives reason to assume that the results of the correlation and regression analyses for the indicator total protein in the hemolymph could be applied to the entire studied population of bee colonies. The results of the conducted correlation analysis (Figure 5) show that there is a weak negative relation (r=-0.33) between total protein content and hygienic behaviour was established which means that an increase of the activities related to the cleansing of honeycomb cells leads to a decrease in the total protein content. As can be seen from Figure 3 and Figure 5 the are a lot of outliers for both data sets (lysozyme and total protein quantity). A possible reason for this is the larger variations in the data. This could be avoided in subsequent experiments by working with more samples or further investigating the nature of the relationship between the variables. The parameters of the calculated linear regression model are presented in Table 5 and Figure 6 graphically shows the predicted values of the studied parameter total protein content and the overall view of the regression equation. As can be seen from the Table, the model is statistically significant (P<0.05).

 

Table 5 Regression model showing the dependence of total protein content on worker bee haemolymph from hygienic behaviour¹

 

¹ Regression summary for dependent variable: total protein, g/L; R= 0.33019552; R²= 0.10902908; Adjusted R²= 0.09730578; F (1; 76)= 9.3002; P < 0.00315 and Standard error (SE): 14.477.

* (P<0.05).

 

Figure 6 Regression model showing the dependence of total protein content in the haemolymph from hygienic behaviour of worker bees

 

 

The value of the determination coefficient R²= 0.10902908 (Table 5), in turn, shows that around 11% of the variations in the studied total protein content in the haemolymph can be explained by the degree of hygienic behaviour of the bees. This percentage is much higher compared to the percentage of variations in lysozyme levels that can be explained by the bees’ hygienic behaviour. As this is the first study in Bulgaria concerning the relationship between the hygienic behaviour of bees and lysozyme levels and total protein content of their haemolymph, it is not possible to answer unequivocally whether the percentage is sufficient for the relevant analysis.

 

CONCLUSION

The mean values obtained for both analyzed parameters (lysozyme and total protein) in the haemolymph of worker bees from non-hygienic bee colonies were higher (13.71±1.76 μg/mL and 22.63±2.97g/L) compared to those of hygienic bee colonies (10.33±0.93 μg/mL and 19.05±1.95 g/L). The differences found between the mean values of the stated parameters for hygienic and nonhygienic bee colonies are statistically insignificant at significance level (P≤0.05). The correlation analysis shows a weak relationship (r=0.35) between the lysozyme levels in the haemolymph and the hygienic behaviour of worker bees and a weak negative relationship (r=-0.33) between the total protein content and the hygienic behaviour, respectively. Based on the correlation analysis, it can be concluded that there is a tendency to increase the levels of lysozyme and decrease the total protein content in the worker bees' hemolymph with increase the degree of hygienic behaviour. The developed regression model showing the degree of dependence of lysozyme levels from hygienic behaviour of bees is statistically insignificant but the regression model for the dependence of total protein is statistically significant and can be used to predict the relation between total protein content in the haemolymph and hygienic behaviour of bee colonies. In summary, it can be concluded that the insignificancy of the model for predicting the lysozyme levels in the haemolymph confirms the need to continue the analysis with a greater number of observations, as well as further study of the nature of the dependencies among the variables.

مراجع

Aronstein K. and Saldivar E. (2005). Characterization of a honeybee toll related receptor gene Am18w and its potential involvement in antimicrobial immune defense. Apidologie. 36, 3-14.

Aronstein K.A., Murray K.D. and Saldivar E. (2010). Transcriptional responses in honey bee larvae infected with chalkbrood fungus. BMC Genom. 11, 391-403.

Comman R.S., Tarpy D.R., Chen Y., Jeffreys L. and Lopez D. (2012). Pathogen webs in collapsing honey bee colonies, PLoS One. 8, e43562.

Darkazanli S. (2008). Study on the hygiene behavior of the honey bee and its relation with some factor diseases. Ph D. Thesis. University of Forestry, Sofia, Bulgaria.

Gatschenberger H., Azzami K., Tautz J. and Beier H. (2013). Antibacterial immune competence of honey bees (Apis mellifera) is adapted to different life stages and environmental risks. PLoS One. 8, e66415.

Gechev I. (1995). On some morphological, physiological and biochemical changes in bees infested with Varroa jacobsoni. Ph D. Thesis. University of Forestry, Sofia, Bulgaria.

Gillespie J.P., Kanost M.R. and Trenczek T. (1997). Biological mediators of insect immunity. Annu. Rev. Entomol. 42, 611-643.

Glinski Z. and Jarosz J. (2001). Infection and immunity in the honey bee Apis mellifera. Apiacta. 36(1), 12-24.

Glupov V.V. (2004). Basic Mechanisms of the Insect Constitutional Immunity. Dissertation paper DSc of biology, Novosibirsk, Russia.

Gramacho K.P. and Gonçalves L.S. (2009). Sequential hygienic behavior in Carniolan honey bees (Apis mellifera carnica). Genet. Mol. Res. 8(2), 655-663.

Guler A. and Hakan T. (2013). Relationship between dead pupa removal and season and productivity of honey bee (Apis mellifera, Hymenoptera: Apidae) colonie. Turkish J. Vet. Anim. Sci. 37, 462-467.

Gurgulova K., Jeliazkova I. and Stoilov N. (2003). A study on the hygienic behavior of bee workers Apis mellifera. J. Anim. Sci. 40(1), 127-129.

Gurgulova K., Valchovski R., Jeliazkova I. and Nenchev P. (2001). Effect of some stimulating substances on natural immunity of honey bee Apis mellifera. J. Anim. Sci. 36(2), 113-115.

Haritonov N. (2015). Selecting honey bees for resistance to Varroa jacobsoni. Apidologie. 30(2), 183-196.

Ilyasov R.A., Gayfullina L.R., Saltaykova E.S., Poskryakov A.V. and Nikolenko A.G. (2014). The role of the antimicrobial defensis peptide in the immunity of the bee colony. Pchelovodstvo. 1, 20-28.

Kanchev K., Gechev I., Parvanov P. and Bonovska M. (1997). Amount of total protein, protein fractions and concentration of lysozyme in the haemolymph of newly hatched worker bees Apis mellifera infested with Varroa jacobsoni. Vet. Med. Sci. 29, 12-20.

Kostov G. and Bonovska М. (1983). Methods for determining the lysozyme activity of serum from farm animals and poultry, Vet. Collect. 11, 30-32.

Lapidge K., Oldroyd B. and Spivak M. (2002). Seven suggestive quantitative trait loci influence hygienic behavior of honey bees. Naturwissenschaften. 89, 565-568.

Mansourizalani A., Tahmasbi G.H., Emam Jomeh kashani N., Amin Afshar M. and Ghazi Khani Shad A. (2018). Study on the hygenic and grooming behaviours of Iranain honeybees colonies (Apis mellifera meda) in the third and fourth generation of breeding plan for resistance to Varroa destructor. J. Entomol. Res. 10(1), 65-76.

Motavkina N.S., Kovalev B.M. and Sharonov A.S. (1979). Micromethod for quantitative determination of lysozyme. Lab. Works. 12, 722-723.

Nagornaya I.M., Bodnarchuk L.I. and Levchenko I.A. (1996). Lyzozyme in the salivary gland secretion of bees. Pchelovodstvo. 3, 19-21.

Nemkova S.N. (2004). Activeness of the hygiene behavious of bees Apis mellifera (Hymenoptera: Apoidae). Bull. Kharkiv Entomol. Soc. 12(1), 191-194.

Nenchev P. and Zhelyazkova I. (2010). Beekeeping. Academic Publishing House of Thracian University, Stara Zagora, Bulgaria.

Nganso B.T., Fombong A.T., Yusuf A.A., Pirk C.W., Stuhl C. and Torto B. (2017). Hygienic and grooming behaviors in African and European honeybees—New damage categories in Varroa destructor. PLoS One. 12, e0179329.

Oxley P.R., Spivak M. and Oldroyd B.P. (2010). Six quantitative trait loci influence task thresholds for hygienic behaviour in honeybees (Apis mellifera). Mol. Ecol. 19, 1452-1461.

Palacio M.A., Rodriguez E., Goncalves L., Bedascarrasbure E. and Spivak M. (2010). Hygienic behaviors of honey bees in response to brood experimentally pin-killed or infected with Ascosphaera apis. Apidologie. 41(6), 602-612.

Petrov P. (1997). Morpho-etological characteristics of honey bee in the Strandzha region. J. Anim. Sci. 7, 137-140.

Rothenbuhler W.C. (1964). Behaviour genetics of nest cleaning in honey bees. I. Responses of four inbred lines to diseased-killed brood. Anim. Behav. 5, 578-583.

Shumkova R. (2016). Comparative study on the effect of stimulating products on the biological development of bee colonies, Ph D. Thesis. Kostinbrod Univ., Kostinbrod, Bulgaria.

Spivak M. and Downey D.L. (1998). Field assays for hygienic behaviour in honey bees (Hymenoptera: Apidae). J. Econ. Entomol. 91(1), 64-70.

StatSoft, Inc. (2014). STATISTICA (data analysis software system), version 12. Available at: www.statsoft.com

Taber S. (1982). Bee behaviour. American Bee J. 6, 422-425.

Wilson-Rich N., Dres D.T. and Starks P.T. (2008). The ontogeny of immunity: Development of innate immune strength in the honey bee ( Apis mellifera ). J. Insect Physiol. 54, 1392-1399.

Zhelyazkova I. and Gurgulova K. (1997). Study of lysozyme level in honey-bee workers`chemolymph under different conditions of rearing and feeding. J. Anim. Sci. 1, 278-280.

Zhelyazkova I. and Gurgulova K. (2000). Influence of feeding with pollen substitute on the protein and lysozyme level in the hemolymph of worker bees (Apis mellifera). Pp. 39-44 in Proc. Natl. Conf. Achiev. Field of Agrar. Social Sci., Zagora, Morocco.

Zyuman B. (1988). Factors of natural immunity. Pchelovodstvo. 11, 21-23.