Entomological Science (2009) 12, 116–129
ORIGINAL ARTICLE
ens_311
doi:10.1111/j.1479-8298.2009.00311.x
116..129
Segregation of temporal and spatial distribution between kleptoparasites and parasitoids of the eusocial sweat bee, Lasioglossum malachurum (Hymenoptera: Halictidae, Mutillidae) Carlo POLIDORI, Luigimaria BORRUSO, Roberto BOESI and Francesco ANDRIETTI Department of Biology, Section of Zoology and Cytology, University of Milan, Milan, Italy
Abstract Cuckoo bees and velvet ants use different resources of their shared host bees, the former laying eggs on the host pollen stores and the latter on immature stages. We studied the activity patterns of the cuckoo bee Sphecodes monilicornis and the velvet ant Myrmilla capitata at two nesting sites of their host, the social digger bee Lasioglossum malachurum, over a 3 year period. Due to the difference in host exploitation, we expected different temporal patterns of the two natural enemies as well as a positive spatial association with host nest density for both species. At a daily level, S. monilicornis was more abundant between 10.00 and 15.00 h, while M. capitata was most active in the early morning and late afternoon; both species activities were independent from host provisioning activity. The activity of cuckoo bees was in general positively correlated with the density of open host nests (but not with the total number of nests), while that of velvet ants was rarely correlated with this factor. Sphecodes monilicornis was seen both attacking the guard bees and directly entering into the host nests or digging close to nest entrances, while M. capitata only gained access to host nests through digging. We conclude that the temporal and spatial segregation between the two species may be, at least partially, explained both by the different resources exploited and by the different dynamics of host interactions. Key words: host-searching activity, kleptoparasite, Myrmilla, optimal foraging theory, parasitoid, Sphecodes.
INTRODUCTION Major enemies of bees (Apoidea: Anthophila) are other aculeate Hymenoptera. Among this group, two kinds of parasitic strategies are recognized. One is kleptoparasitism – that is, the exploitation of the host’s maternal resources by laying eggs in nests of other species to avoid the costs of brood care – and the other is parasitoidism – namely a life cycle that includes free-living mothers laying eggs in or on the host (O’Neill 2001). At least eight families of Aculeata contain obligate kleptoparasites, while at least 13 families contain parasitoids (Bohart 1970; O’Neill 2001). Cuckoo bees are kleptoparasites.
Correspondence: Carlo Polidori, Dipartimento di Biologia, Sezione di Zoologia e Citologia, Università degli Studi di Milano, via Celoria, Milano, Italy. Email:
[email protected] Received 27 May 2008; accepted 18 November 2008.
© 2009 The Entomological Society of Japan
Every species attacks nests of one or several other bee hosts, often phylogenetically related (Michener 2000) and uses various strategies to get into the host nest (Bogusch et al. 2006). In the genus Sphecodes, for example, entry of parasitic females into the nest can occur both during the absence of the host female and in the presence of the host female. In the latter case, it was observed that host presence does not prevent the kleptoparasite’s entry, and during the contact females usually do not fight (Bogusch et al. 2006). In contrast, velvet ants (Mutillidae) are ectoparasitoids of mature larvae or pre-pupae of other insects, mostly other aculeate Hymenoptera (Brothers et al. 2000). They are characterized, with few exceptions, by a strong sexual dimorphism: males are winged while females are invariably wingless (Deyrup & Manley 1986). Females, probably in response to their apterous condition (higher susceptibility to predators) and due to their host-attacking strategy (higher efficiency in
Distribution of kleptoparasites and parasitoids of a bee
subduing adult host bees or wasps if encountered in host nests), have evolved several adaptations, including a long mobile sting with powerful venom, a strong and slippery cuticle, an ability to run very rapidly and evasively, aposemantic warning coloration patterns, releasing chemical secretions and producing stridulation (Schmidt & Blum 1977). Despite the often great abundance of both groups at aggregations of their hosts, cuckoo bees and velvet ants have been infrequently studied from an ecological and behavioral point of view. In particular, despite behavioral interactions with the hosts and manners of parasitism into nests having been described for some species (cuckoo bees by Sick et al. 1994 and Bogusch et al. 2006; velvet ants by Brothers 1972 and Bayliss & Brothers 1996), quantitative data on activity patterns are still lacking. Very few studies have investigated seasonal and daily patterns of these natural enemies, either at a temporal or at a spatial scale (e.g. Alicata et al. 1974; Bogusch 2003). This information is necessary if we want to know more about host-searching and host-attacking strategies of cuckoo bees and velvet ants. To successfully exploit host brood, cuckoo bees and velvet ants should be synchronous to the breeding period of the host (Wcislo 1987). As Wcislo (1987) pointed out, referring in particular to kleptoparasites, there is a narrow “window of opportunity” for these organisms to complete their egg-laying and development within the nests of their hosts. In addition, optimal foraging theory predicts that animals should behave in such a way as to find, collect and use resources while expending the least amount of time possible in doing so (Kamil et al. 1987). In the case of primitively eusocial, mass-provisioning hosts such as some sweat bees in temperate regions, characterized by discrete foraging phases separated by several days of worker inactivity (Michener 1974), these natural enemies should be constrained by different “windows of opportunity”. In fact, due to the different host resources used by these two groups on the host resources (fresh pollen balls vs mature larvae/prepupae), a number of predictions may arise. Cuckoo bees should be: (i) more active in periods of greatest provisioning activity of the host bee (when they are more likely to find fresh pollen); and (ii) more abundant in areas of high nest density, because in these areas females should have a greater probability of finding a suitable host nest to enter. In contrast, velvet ants should: (i) be more active at the end of an activity phase (when more suitable larvae are present in the nests); (ii) have a daily pattern of activity independent from the foraging pattern of the host (mature larvae and pre-pupae are unlikely to be more abundant at any
Entomological Science (2009) 12, 116–129 © 2009 The Entomological Society of Japan
given hour of the day); and (iii) like kleptoparasites, be more abundant in higher nest density areas. We tested these predictions on the cuckoo bee Sphecodes monilicornis (Kirby) and the velvet ant Myrmilla capitata Lucas, through field observations at two nest aggregations of their shared host, the social digger bee Lasioglossum malachurum (Kirby). We did not plan the study to evaluate the dynamics of activity patterns of these two natural enemies across the whole annual cycle of the host bee (from nest founding to sexual emergence, ~6 months at this locality, Polidori et al. unpubl. data, 2003–2007), but to look for the distribution of these patterns during the “eusocial part” of the nesting cycle (i.e. with workers in the nests).
MATERIALS AND METHODS Synopsis of the biology of studied species Sphecodes monilicornis is widespread in Europe. Females lay eggs in fresh brood cells of the genera Halictus and Lasioglossum (Westrich 1989). Both solitary and social hosts have been recorded (Bogusch et al. 2006). Little is known about the impact of this kleptoparasite on the host population, with recorded rates of parasitism of approximately 20% (Strohm & BordonHauser 2003), although invasion and parasitism of entire nests or colonies have been also observed (Sick et al. 1994). Myrmilla capitata is confined to southern Europe and northern Africa (Invrea 1964); its hosts are groundnesting bees of different families, more often social sweat bees of the genera Lasioglossum and Halictus (Brothers et al. 2000). This species, together with very few others, is atypical in that males are not winged and are quite similar to females. Lasioglossum malachurum is a typical primitively eusocial sweat bee (Paxton et al. 2002). Queens establish their colonies in subterranean nests in spring, and then produce one (in northern Europe) to three (in southern Europe) worker phases and a last phase composed of males and gynes (Knerer 1992; Wyman & Richards 2003), these phases being separated by several days during which no foraging activity takes place. Mated gynes overwinter and found new colonies in the following spring. Colonies include few workers (as few as approximately four per nest on average, Paxton et al. 2002) in the spring phase, but they can become populous in the summer phases (up to 80 workers per nest) (Knerer 1992).
Study area and host nest aggregations The field collection of the data was performed near Alberese, a small town inside the Maremma Regional
117
C. Polidori et al.
Park (Tuscany, Italy: 42°40′5″N, 11°6′23″E). This area is typically Mediterranean, with an average annual temperature of approximately 14–15°C (7.1°C in January, 23.1°C in August); the average yearly rainfall is approximately 690 mm, with a maximum in November– December and a minimum in July–August. The main part of the park is characterized by the Uccellina Mountains, a chain of hills parallel to the coast and covered by the thick Mediterranean maquis. In this area, L. malachurum is commonly found nesting in small (2000) nest aggregations in a variety of locations, such as along tracks in the pine woods, along cultivated fields, or even in small bare soil patches inside the town (Polidori et al., unpubl. data, 2003–2007). Two bee nesting sites were chosen for the study: site A was located in a Quercus wood, while site B was located along an alfalfa field approximately 100 m from the Ombrone River. The two sites were separated by approximately 3 km, and in both L. malachurum nested copiously (>1000 nests at both sites). At site A, the next most abundant ground-nesting hymenopteran (~50 nests) was another sweat bee, Lasioglossum marginatum (Brullè) (whose nests did not intermix with those of L. malachurum), while in site B it was the digger wasp, Cerceris rubida Jurine (~100 nests, intermixed with those of L. malachurum) (Polidori et al. 2006). Apart from Sphecodes bees and Myrmilla velvet ants, which were abundantly found at both sites, several other natural enemies of L. malachurum were observed to be persistent and relatively abundant at the sites (i.e. recorded every year from 2003 to 2007): the scuttle fly Megaselia leucozona Schmidt, at least one species of Bombylius beefly, and the parasitic wasp, Gasteruption rubricans Guerin; in addition, phoretic parasitic larvae of a meloid beetle were sometimes observed on the body of bee workers returning to the nests (Polidori et al. 2005a and unpubl. data, 2003–2007; Boesi et al. 2009).
Host searching activity Temporal and spatial activity patterns of S. monilicornis and M. capitata were assessed through field observations in 2004 and 2005. We obtained data from different periods of the host nesting cycle and from both sites A and B. In 2004, we monitored only site A from 22 May to 2 June; in 2005, site A was monitored in two periods (20–23 May and 21–27 June), and site B was monitored in two periods (24–27 May and 23 July to 8 August); on the whole, 39 days of observation were spent. In both sites, May coincided with the foraging period of the first worker phase, June with a period of non-foraging activity between the spring and summer worker phases, and July–August with the second worker phase foraging period (Polidori et al. unpubl. data, 2003–2007).
118
Because we did not excavate nests of L. malachurum during the study (to avoid affecting data collection on parasite activity), we made some assumptions about the dynamics of the bee nesting cycle: (i) fewer workers were present in the spring phases than in the summer phases; (ii) between activity phases (June), most individuals were at immature stages inside the nests; and (iii) in the last days of an activity period, most workers were still alive in the nests despite their low foraging effort. These assumptions are consistent with observations in several populations of this sweat bee across Europe (Knerer 1992; Wyman & Richards 2003; Mitesser et al. 2006; Weissel et al. 2006). In each day of observation, every hour from 8.00 to 18.00 (solar hours), we observed the cuckoo bee and host activity in eight fixed plots of 1 m ¥ 1 m (except in 2004, when 24 plots of 2.5 m ¥ 2.5 m were used). Plots were chosen in order to obtain a wide range of nest density (per m2), which ranged from zero to seven nests/m2 at site A (2004), from five to 17 nests/m2 at site A (2005) and from eight to 31 nests/m2 at site B. Each plot was observed for 5 min/h of the observation period, during which the following information was obtained: (i) the highest number of S. monilicornis at the same time in the plot; (ii) the highest number of M. capitata counted at the same time in the plot; (iii) the total number of entries of host bee workers with pollen into the nests; and (iv) the total number of open nests of the host bee (i.e. whose entrance was open at the moment of observation, independently by the activity of workers). Sphecodes monilicornis was not the only cuckoo bee present at both sites, but it was by far the most abundant: during 4 years (2003–2006) of collection in the areas, only six out of 134 collected individuals belonged to other species of the genus ( 0.6). No significant trends of activity of S. monilicornis and host resulted in July–August 2005 across the days of observation (cuckoo bee: r = -0.16, n = 12, P = 0.6;
119
C. Polidori et al.
Figures 1–2 Distribution of activities through the days of observation in May 2004 at site A. 1 Sphecodes monilicornis (open dots) and Myrmilla capitata (solid dots) (number of individuals recorded in 5 min/plot); 2 Lasioglossum malachurum (solid dots) (number of entries in the nest with pollen in 5 min/plot) and number of open nests (open dots). Significant regressions were shown (trend lines and equations) in 1 for S. monilicornis and in 2 for L. malachurum (solid line) and for nests (dashed line).
L. malachurum: r = -0.29, n = 12, P = 0.19), but M. capitata in this period linearly increased its activity across the 12 days of observation (r = 0.59, n = 12, P = 0.041). The two natural enemies did not differ in frequency (values per date per hour) in May 2004 at site A (Student’s t-test for paired data: t = 1.21, d.f. = 59, P = 0.22). In 2005, M. capitata was more abundant than S. monilicornis at site A (May: Student’s t-test for paired data: t = -2.51, d.f. = 39, P = 0.016; in June only mutillids were recorded); in 2005, at site B, this pattern was only significant in the summer months of July and August (May: Student’s t-test for paired data: t = 0.45, d.f. = 39, P = 0.65; July–August: Student’s t-test for paired data: t = -5.01, d.f. = 34, P < 0.001). At both sites in 2005, host foraging activity was weak or absent in summer (site A: May = 2.02 ⫾ 0.3 trips/day, June = 0; site B: May = 3.04 ⫾ 0.5, July– August = 0.52 ⫾ 0.22 trips/day), although a number of open nests was recorded in these periods. At site A, the number of open nests decreased significantly (May = 18.07 ⫾ 0.35, June = 13.6 ⫾ 0.3; Student’s t-test for unpaired data: t = 9.69, d.f. = 80, P < 0.001), while at site B it increased weakly significantly (May = 18.42 ⫾ 1.17, June = 22.8 ⫾ 1.7; Student’s t-test for unpaired data: t = -2.17, d.f. = 73, P = 0.03).
120
In 2004 at site A, the daily activity of cuckoo bees (values per 5 min plot observation) approximately followed a quadratic curve across the day (r = 0.91, n = 10, P < 0.001) (Fig. 3), and the same pattern was recorded in 2005 in all periods and sites (Figs 4,5,7) (site A: May, r = 0.51, n = 10, P = 0.04; site B: May, r = 0.80, n = 10, P < 0.001; July–August, r = 0.79, n = 10, P < 0.001). In contrast, M. capitata daily activity approximately followed an inverse quadratic curve across the day in all periods and sites except in May at site A (both years) (Figs 3–7) (site A: May (2004), r = 0.39, n = 10, P = 0.09; May (2005), r = 0.2, n = 10, P = 0.12; June, r = 0.91, n = 10, P < 0.001; site B: May, r = 0.85, n = 10, P < 0.001; July–August, r = 0.83, n = 10, P = 0.041). Generally differing from both cuckoo bees and velvet ants, L. malachurum activity followed a negative linear trend across the day in all periods and sites except at site A in May 2004, when the distribution fitted a quadratic regression (May 2004, r = 0.95, n = 10, P < 0.001; site A: May, r = 0.89, n = 10, P < 0.001; site B: May, r = 0.92, n = 10, P < 0.001; July–August, r = 0.73, n = 10, P = 0.015) (Figs 8–11). Consequently, no linear correlations resulted between the activity of natural enemies and that of the host bee except for May 2004 for S. monilicornis (May 2004, r = 0.69, n = 10, P = 0.005; site A: May, S. monilicornis, r = 0.29, n = 10, P = 0.41; M. capitata, r = -0.18, n = 10, P = 0.6; site B, May, S. monilicornis, r = 0.50, n = 10, P = 0.14; M. capitata, r = 0.33, n = 10, P = 0.35; July–August, S. monilicornis, r = 0.035, n = 10, P = 0.92; M. capitata, r = 0.42, n = 10, P = 0.22). The number of open nests was not significantly related to time of day except for site A in 2004, where a negative linear correlation was detected (r = -0.90, n = 10, P < 0.001); no correlations were found between this factor and the cuckoo bee and velvet ant (and even the host bee) activities (P > 0.1 for all periods and sites). Thus, with the 2004 exception, nests closed well after the end of foraging activity of the host bee.
Spatial patterns of activity Activity of S. monilicornis (values per 5 min plot observation) was correlated linearly with the number of open nests in the plot only in 2004 and at site B, both in spring and summer (2004, r = 0.64, n = 34, P < 0.001; site A: May, r = -0.35, n = 16, P = 0.17; site B: May, r = 0.44, n = 30, P = 0.012; July–August, r = 0.46, n = 17, P = 0.045) (Figs 12,14,16), while at site A the relationship was quadratic (r = 0.59, n = 16, P < 0.01) (Fig. 13). The activity of M. capitata was correlated linearly with the number of nests only in June at site A (2004, r = 0.18, n = 34, P = 0.29; site A: May, r = -0.36,
Entomological Science (2009) 12, 116–129 © 2009 The Entomological Society of Japan
Distribution of kleptoparasites and parasitoids of a bee
Figures 3–11 Daily distribution of activities of Sphecodes monilicornis (open dots) and Myrmilla capitata (solid dots) (number of individuals recorded in 5 min/plot) (3–7) and of the host bee (number of entries in the nest with pollen in 5 min/plot) (8–11). 3 May 2004; 4 May 2005 site A; 5 May 2005 site B; 6 June 2005; 7 July–August 2005; 8 May 2004; 9 May 2005 site A; 10 May 2005 site B; 11 July–August 2005. Significant regressions were shown (trend lines and equations; S. monilicornis: full line, M. capitata: dashed line).
Entomological Science (2009) 12, 116–129 © 2009 The Entomological Society of Japan
121
C. Polidori et al.
Figures 12–16 Relationships between Sphecodes monilicornis (open dots) and Myrmilla capitata (solid dots) activities (number of individuals recorded in 5 min/plot) and the number of open host nests (in 5 min/plot). 12 May 2004; 13 May 2005 site A; 14 May 2005 site B; 15 June 2005; 16 July–August 2005. Significant regressions were shown (trend lines and equations; S. monilicornis: full line, M. capitata: dashed line).
n = 16, P = 0.16; June, rSpearman = 0.74, n = 9, P = 0.022; site B: May, r = 0.004, n = 30, P = 0.98; July–August, r = 0.41, n = 17, P = 0.09), while in May at site A the relationship was quadratic (r = 0.74, n = 16, P < 0.01) (Figs 12–16). When the total number of nests in the plots rather than the number of open nests is considered as the dependent variable, S. monilicornis activity was not correlated with this factor except in 2004 (2004, r = 0.56, n = 24, P = 0.005; May at site A: rSpearman = -0.21, n = 8, P = 0.56; May at site B: rSpearman = 0.57, n = 8, P = 0.12; July–August at site B: rSpearman = 0.04, n = 8, P = 0.89); the
122
same results were obtained for M. capitata activity (2004, r = 0.62, n = 24, P < 0.001; May at site A: rSpearman = 0.06, n = 8, P = 0.87; June at site A: rSpearman = -0.06, n = 8, P = 0.85; May at site B: rSpearman = 0.42, n = 8, P = 0.26; July–August at site B: rSpearman = -0.35, n = 8, P = 0.34). The importance of the number of open nests over other factors (number of host foraging trips, hour of the day, total number of nests and activity of the other natural enemy) in determining the presence/absence of the cuckoo bee and the velvet ant, was demonstrated by a series of binary logistic regressions (Table 1). In fact, the number of open nests was significant for both species
Entomological Science (2009) 12, 116–129 © 2009 The Entomological Society of Japan
Entomological Science (2009) 12, 116–129 © 2009 The Entomological Society of Japan
Presence of Myrmilla capitata
Presence of Sphecodes monilicornis
Dependent binary variable
Number of host foraging trips Estimate = 0.098 c2 = 0.12 P = 0.72 Estimate = 0.057 c2 = 0.006 P = 0.93 Estimate = 0.44 c2 = 3.13 P = 0.077 Estimate = -0.23 c2 = 2.12 P = 0.19 Estimate = -0.52 c2 = 1.53 P = 0.21 Estimate = -0.47 c2 = 4.67 P = 0.03
Goodness of fit of the model R2 = 0.085 R2 (McFadden) = 0.105 P < 0.0001 R2 = 0.002 R2 (McFadden) = 0.009 P = 0.72 R2 = 0.095 R2 (McFadden) = 0.107 P < 0.0001 R2 = 0.012 R2 (McFadden) = 0.020 P < 0.001 R2 = 0.003 R2 (McFadden) = 0.003 P = 0.63 R2 = 0.031 R2 (McFadden) = 0.03 P < 0.0001
Estimate = -0.31 c2 = 1.88 P = 0.32 Estimate = -0.08 c2 = 0.05 P = 0.81 Estimate = -0.06 c2 = 0.04 P = 0.68
c2 = 1.34 P = 0.12
Estimate = 0.13 c2 = 0.2 P = 0.68 Estimate = 0.07 c2 = 0.01 P = 0.9 Estimate = 0.22
Hour of the day
Estimate = 0.21 c2 = 16.3 P < 0.0001 Estimate = 0.04 c2 = 0.071 P = 0.8 Estimate = 0.43 c2 = 20 P < 0.0001
c2 = 18.7 P < 0.0001
Estimate = 0.48 c2 = 101.4 P < 0.0001 Estimate = 0.43 c2 = 1.07 P = 0.3 Estimate = 0.73
Total number of nests
Estimate = 0.02 c2 = 0.003 P = 0.95 Estimate = -0.02 c2 = 0.0004 P = 0.98 Estimate = -1 c2 = 3.98 P = 0.04
c2 = 3.13 P = 0.077
Estimate = -0.03 c2 = 0.006 P = 0.93 Estimate = -0.06 c2 = 0.004 P = 0.95 Estimate = -0.76
Other natural enemy activity
The dependent binary variable was the presence or absence of the species in each plot/day/h, while the explanatory variables were the four in the last four columns. The variable “other natural enemy activity” is the highest number of individuals (recorded in each plot/day/hour) of S. monilicornis (if M. capitata is the binary dependent variable) or of M. capitata (if S. monilicornis is the binary dependent variable). The probability of the significant models and factors are highlighted in bold.
2005 – B (n = 682)
2005 – A (n = 652)
2004 – A (n = 1464)
2005 – B (n = 682)
2005 – A (n = 652)
2004 – A (n = 1464)
Period/site
Table 1 Results of the binary logistic regressions performed on the data collected through the “patrolling of the area” method
Distribution of kleptoparasites and parasitoids of a bee
123
C. Polidori et al.
in both sites and years, with the exception of 2005 at site A (where the binary logistic model did not explain the presence/absence of the species using any variables) (Table 1). In addition, only in 2005 at site B, the presence of M. capitata was explained by two other negatively (and more weakly) associated factors (host bee foraging activity and cuckoo bee activity) (Table 1). Finally, in 2004 in 18 out of 1464 and in 2005 in 12 out of 1334 (both sites pooled) 5 min observations S. monilicornis and M. capitata were recorded at the same time, and these events occurred more often in areas of lower density of open nests (2004: r = -0.43, n = 34, P = 0.009; 2005: r = -0.41, n = 30, P = 0.022).
Temperature and activity In spring (May), tests for linear associations were possible only for site A in 2004, because in 2005 we performed 4 days of observations at each site. In 2004 at site A, mean S. monilicornis activity per day was marginally and negatively associated with mean air temperature (r = -0.55, n = 12, P = 0.06), and not correlated with maximum air temperature (r = -0.49, n = 12, P = 0.1), while mean M. capitata activity per day was only marginally and negatively correlated with maximum air temperature (mean: r = -0.24, n = 12, P = 0.45; maximum: r = -0.54, n = 12, P = 0.06). In June 2005 at site A, mean M. capitata activity per day was correlated only marginally and negatively with maximum air temperature (mean: rSpearman = -0.61, n = 7, P = 0.13; maximum: rSpearman = -0.78, n = 7, P = 0.05). In July–August 2005 at site B, mean S. monilicornis activity per day was positively associated with mean and maximum air temperature (mean: r = 0.75, n = 12, P = 0.002; maximum: r = 0.67, n = 12, P = 0.015), while mean M. capitata activity per day was negatively correlated with mean air temperature (mean: r = -0.70, n = 12, P = 0.008; maximum: r = -0.33, n = 12, P = 0.28). Total period of activity (i.e. the number of hours in which the species was recorded) of M. capitata per day was negatively correlated with mean temperature in both spring and summer (May 2004: r = -0.65, n = 12, P = 0.02; June 2005: rSpearman = -0.73, n = 7, P = 0.06; July–August 2005: r = -0.61, n = 12, P = 0.03). On the contrary, cuckoo bee’s total activity period was not correlated to this factor (May 2004: r = -0.29, n = 12, P = 0.35; July–August 2005: r = 0.34, n = 12, P = 0.27).
Behavior close to the host nests During the 2004 “nest observations”, S. monilicornis was seen successfully entering nests a total of 40 times, that is, only 1.1 times on average per day/nest, staying inside from less than 1 up to 28 min (7 min 28 s on
124
average). Although we did not observe longer residence in the nests during the observations of bee colonies, during patrols of nesting site B (2004–2005) we recorded a total of five cases of much longer activity (>1 h) of S. monilicornis, which entered the nest and killed several host workers one by one (extracting them from the nest) (Figs 17,18). Up to 19 dead workers were discovered close to a nest entrance, with the cuckoo bee still involved in this operation. On another 260 occasions (7.4 on average per day/nest) the cuckoo bee approached the nest entrance, either landing directly on it (34% of cases) or walking to it (66% of cases), but was rejected by the guard at the entrance. The guard was never seen to exit as a defensive response to cuckoo bee detection, and no fights occurred in these cases. However, on 23 occasions a worker bee returning to the nest with pollen intercepted S. monilicornis that was very close (
