Experimental reduction of ultraviolet wavelengths reflected from parasitic eggs affects rejection behaviour in the blackcap Sylvia atricapilla

Interspecific brood parasitism is a reproductive strategy adopted by about one percent of bird species (Payne,1997), and consists of laying eggs in nests of unrelated species,the hosts, which incubate the eggs and feed the parasitic young(Lack, 1968). It has been documented that a successful parasitic act has a generally detrimental effect on the host's fitness because of dramatic reduction of their reproductive fitness (Davies, 2000). Many studies have shown that interaction between avian brood parasites and their hosts represent one of the best examples of coevolution (for a review, see Rothstein, 1990), which has led to a variety of adaptations and counter-adaptations on both sides, and has been described as an evolutionary arms race(Dawkins and Krebs, 1979; Davies and Brooke, 1989). For example, in the case of a host-parasite system in which the female parasite specialises on a single host species, parasitic egg mimicry is one of the outcomes of this struggle (Brooke and Davies, 1988). However, this counter-adaptation of the brood parasite must be very stressful for hosts, since egg discrimination has primarily evolved as a consequence of other selection pressures associated with nesting in dense colonies (Underwood and Sealy, 2002).

A significant effort has been devoted to studying naturally parasitized host populations to explore a role of parasitic egg mimicry(Moksnes et al., 1993; Edvardsen et al., 2001; Moskát and Honza,2002). Similarly, many conclusions have been drawn by an experimental approach testing host responses towards both conspecific eggs(Procházka and Honza,2003; Procházka and Honza, 2004; Lovaszi and Moskát, 2004) and an artificial model egg(Moskát and Fuisz,1999; Honza et al.,2004). Most of these studies have tried to explain why so many different host populations accept non-mimetic eggs. This scenario could be primarily influenced by the fact that humans and birds significantly differ in their colour perception. Indeed, Honza et al.(Honza et al., 2007) in a study of song thrushes Turdus philomelos revealed how questionable the use of egg categories such as mimetic versus non-mimetic might be, since some colours of the parasitic eggs classified by humans as non-mimetic were accepted by the hosts. This discrepancy could be explained by the anatomical differences between bird and human eyes(Chen and Goldsmith, 1986; Harth et al., 1998), because birds are able to detect UV wavelengths in the range of 300–400 nm,invisible to humans (Huth and Burkhardt,1972; Wright,1972). Since the UV visual sensitivity plays an important role in bird recognition ability (Wright,1972; Cherry and Bennett,2001), interest in the role of UV visual signals has been increasing. Moreover, there is evidence about other functional significances of the UV wavelengths in mate choice, foraging and navigation(Bennett and Cuthill,1994).

The evolution of the cues influencing rejection behaviour of hosts has attracted enormous attention from many evolutionary biologists(Moskát and Fuisz,1999; Underwood and Sealy,2006), however, most of these studies did not take into consideration the significance of the UV wavelengths. The role of the UV wavelengths in avian brood parasite–host systems was first investigated by Cherry and Bennett (Cherry and Bennett,2001), who identified the importance of this part of spectrum in the evolution of egg mimicry in the red-chested cuckoo, Cuculus solitarius. Moreover, there are other studies demonstrating that the combination of the UV and the part of the spectrum visible to humans plays an important role in avian brood parasites–host coevolutionary system(Avilés and Møller,2004; Avilés et al.,2006a; Polačiková et al.,2007; Underwood and Sealy,2008).

In our study, we investigated rejection behaviour in blackcaps, Sylvia atricapilla Linnaeus 1758. Although, this species is rarely parasitized at present, there is evidence that blackcaps have been used as hosts by the common cuckoo Cuculus canorus in the past(Honza et al., 2001). Since the rejection behaviour of this species still persists, the blackcaps are considered to be current winners in the coevolutionary struggle with the cuckoo (Honza et al.,2004).

The blackcap host is a passerine species, with the ability to perceive short wavelengths including UV (Cuthill et al., 2000). Moreover, our previous study on this species(Polačiková et al.,2007) revealed that the probability of egg rejection significantly increased with decreasing brightness of host eggs at the blunt pole. In the light of this finding, we investigated an influence of reduction of UV wavelengths reflected from parasitic eggs on recognition behaviour in blackcaps. We predicted that experimental manipulation of the UV wavelengths reflected from parasitic eggshell should influence recognition behaviour of this species. We expected that the model eggs with a reduced part of the UV spectrum would be rejected from experimental nests at a different rate than those with unmanipulated UV wavelengths. In our analyses, we investigated the eggshells of both whole parasitic eggs and their blunt poles.

We monitored a population of blackcaps in a suitable habitat daily from mid April to mid June 2006 near Dolní Bojanovice (48°51′N,17°02′E) in the Czech Republic. We searched systematically for nests during the laying period or the first days of incubation. Eggs found during incubation were floated to estimate their approximate laying date according to the methods described by Hays and LeCroy(Hays and LeCroy, 1971). All eggs were marked with small numbers on the middle part of the eggshell using waterproof ink to allow their individual identification. Experimental parasitism was carried out only when at least the fourth host egg was laid. Blackcap clutches were randomly parasitized by replacing an egg with one randomly chosen egg from an original clutch. Even though the modal clutch size in the blackcaps nest varies from four to five eggs, our previous study(Polačiková et al.,2007) on this host species documented that adding or switching host eggs in experimental nests has no effect on host response. The experimental nests were monitored daily for 5 days after manipulation. If the parasitic egg disappeared within this period but the host eggs were left unharmed in the nest, it was considered to have been rejected. Nest desertion was classified as a mode of rejection. If the parasitic egg remained unharmed in the nest, it was considered to have been accepted. Each experimental nest was parasitized only once and each experimental egg was used only in one experimental trial. Prior to each experimental procedure, we measured the reflectance spectra of both experimental and host eggs in each clutch. Two types of model eggs were used: (1) conspecific eggs coated in a UV blocker(UV^–), and (2) conspecific eggs coated in Vaseline™petroleum jelly (UV⁰). The UV blocker consisted of mixture of an octylmethoxysinnamate and benzophynone-3 which reduces a significant part of the UV wavelengths. The spectral reflectance curves of the experimental eggs for both treatments are shown in Fig. 1. The effect of blocker persisted for at least 48 h after coating the experimental eggs, and it had no effect on curvature of the visible part of the spectrum. All experimental eggs were recoated in the test substances daily during examination of the nests. The third experimental group,consisting of conspecific eggs without any coating, was used as a control. This last group was not included in our analyses.

We measured the spectral reflectance of all eggs from the parasitized clutches in the range of 300–700 nm using a USB 2000 spectrophotometer(Ocean Optics, Dunedin, FL, USA). We divided each egg into three regions across its longitudinal axis: the blunt end, the middle and the sharp end. Each region was a third of the length of the egg. Three randomly located measurements on each part of the egg were taken (each covering 1 mm²). A deuterium and halogen light source was used (DT-Mini-GS,Ocean Optics) that was shone on the eggs through a quartz optic fibre(QR400-7-UV/VIS-BX, Ocean Optics), and was reflected from the eggs at an angle of 45° to the surface. Data from the spectrophotometer were loaded into OOIBase 32 software (Ocean Optics). The measurements were relative and referred to a standard white reference (WS-2) and to darkness. Reference and calibration were made before each measurement of each clutch. Total reflectance was obtained for the UV (325–400 nm) interval, which is the invisible part of colour spectrum for humans, and a visible part of a spectrum(400–700nm). We calculated objective variables of the colour, namely brightness (reflectance intensity), chroma (colour purity) and hue [peak wavelengths; (Endler, 1990)]. Brightness was calculated as the sum of the total reflectance values within the 325–700 nm range of wavelengths(R_325–700). Chroma for each part of the spectrum was calculated using the following reflectance ratios: R_325–400/R_325–700 (UV chroma), R_400–475/R_325–700 (blue chroma), R_475–550/R_325–700(green chroma), R_550–625/R_325–700 (yellow chroma) and R_625–700/R_325–700 (red chroma). Hue was estimated by the wavelength of the maximum reflectance(R_max).

Since our previous study(Polačiková et al.,2007) revealed that mainly characteristics of the blunt egg pole are important for recognition behaviour in blackcaps, we calculated all colour variables reflected from both the whole egg surface and the blunt egg pole separately.

The effect of both colour variables characterizing the parasitic egg appearance and contrast between host and parasitic eggs (both calculated for the whole egg surface and blunt egg pole separately, before and after experimental manipulation) on host's recognition ability was examined by a binomial logistic regression with backward stepwise elimination, which included experimental treatments (UV^– and UV⁰) as categorical variables and host response as a dependent variable (1=rejection,2=acceptance). We started analysis by including all variables in the model,dropping non-significant steps until the last step (step 28), which was considered to be the most significant model.

Prior to the experimental manipulation of eggshell reflectance, there was no difference between the host eggs and the parasitic eggs in the colour characteristics of the whole eggshell surface (t-test; Table 1) or of the blunt egg poles (t-test; Table 2). However, after the experimental manipulation, the brightness in UV part of the spectrum, of the eggshells coated in the UV block(UV^–), significantly decreased for both the whole surface(uncoated, 5759.61±1385.94 vs coated 3665.05±1118.78;mean ± s.d.; t-test=6.11, N_1,2=27; P<0.001) and their blunt poles (3602.46±1777.11 vs2313.45±1159.89; t-test=3.16; N_1,2=27; P=0.003). The eggs coated in the Vaseline remained absolutely unchanged (for the whole eggs: 5644.56±2283.84 vs5644.56±2283.84; t-test=0.00; P=1.00; for the blunt poles: 3303.59±1301.64 vs 3303.59±1301.64; t-test=0.00; P=1.00).

Fig. 2 shows rejection rate of the experimental eggs in all three experimental groups. There were no differences in rejection rate between UV⁰ and the control unmanipulated group (χ²-test=0.78, P=0.37) showing no effect of the grease surface on the bird's response. However, significantly more of the experimental eggs coated in the UV blocker (UV^–)were rejected than either the model eggs coated in the Vaseline(UV⁰: χ²=4.04, P=0.02). Similarly, a logistic regression revealed that the only two variables has a significant effect on the host response (rejection=1; model significanceχ ²=15.7, P<0.001), the brightness of whole conspecific eggs (Wald χ²=7.06, d.f.=1, P=0.008) and the UV brightness at blunt egg poles of experimental eggs (Waldχ ²=6.13, d.f.=1, P=0.013). The likelihood of alien egg rejection increased with increasing values of the brightness of the whole surface of experimental eggs [B=0.00±0.00 (± s.e.m.)],but with decreasing UV brightness at the blunt poles of experimental eggs[B=–0.01±0.0 (± s.e.m.)].

Parent birds are usually able to discriminate a foreign egg in two ways:either using recognition of their own eggs, i.e. true recognition, or recognition of a discordant egg in their own clutch(Rothstein, 1982a). However,other particular cues facilitating recognition of a foreign egg play an important role in the host recognition processes. For example, Rothstein(Rothstein, 1982b) in a study of the American robin (Turdus migratorius) found that smaller parasitic eggs of brown-headed cowbirds, Molothrus ater, elicited a quicker host rejections, probably because of a combination of visual and tactile stimuli, since undersized parasitic eggs are possibly easier to reject. However, Underwood and Sealy(Underwood and Sealy, 2006)argued that warbling vireo (Vireo gilvus) hosts usually reject brown-headed cowbird's eggs only based on eggshell spot pattern, which indicates that heavy spottiness of the parasitic eggshells is primarily responsible for egg recognition and this host species does not need a stimulus summation (Underwood and Sealy,2006). In the cuckoo-host system, there are only a few studies documenting which parameters of the eggshell appearance affect the recognition of the parasitic eggs (Stokke et al.,1999; Procházka and Honza, 2003; Procházka and Honza, 2004). Moreover, it is not just the colour characteristics of the whole eggshell that are important in host recognition behaviour, but also particular eggshell parts(Polačiková et al.,2007).

Since Cherry and Bennett (Cherry and Bennett, 2001) revealed the importance of the part of the spectrum that is invisible to humans in host recognition of parasitic eggs, many researchers have been investigating the significance of this trait(Soler et al., 2003; Avilés and Møller,2004; Cherry et al.,2007) (see also Moskat et al.,2008). However, to our knowledge, there is only one study,conducted by Avilés et al.(Avilés et al., 2006a),which experimentally manipulated UV reflectance to reveal the importance of this part of the spectrum in host recognition. This study investigated recognition behaviour of magpies (Pica pica) towards great spotted cuckoo (Clamator glandarius) eggs, however, their research did not reveal an effect of artificial reduction of the UV reflectance on host perception. In contrast to this result, our study revealed a positive effect of lowering UV reflectance at the blunt pole of experimental eggs on egg rejection in blackcaps. To our knowledge, this is the first experimental evidence showing that a reduction of UV wavelengths reflected from the parasitic eggshells may significantly influence host recognition behaviour.

There are several possible explanations of why our results differ from those of the Avilés et al.(Avilés et al., 2006a)study. First, corvids and small passerines may vary in their UV signal perception (Avilés et al.,2006a). Second, the logistic regression in our study revealed that it is not just UV perception that is important in recognition behaviour of blackcaps but also other parameters including host eggs characteristics. Another possibility for the differences could be the chemicals used. Our UV-blocking chemical covered a larger range of UV wavelengths than that used by Avilés et al. (Avilés et al., 2006a) and the experimental eggs coated in this UV block were darker (for comparison see Fig. 1 in both studies). Moreover, Avilés et al.(Avilés et al., 2006b)discovered that nest light properties may influence host discrimination towards parasitic eggs as the UV wavelengths reflected from the egg surface could be chosen to provide detectable cues for hosts in poorly lit environments. However, this does not apply in our case, as blackcaps usually breed in open nests. Finally, another reason for the different results could be different recognition abilities of the two tested hosts (magpies and blackcaps). Since blackcaps exhibit good recognition abilities resulting in a high rejection rate (Honza et al.,2004; Polačikováet al., 2007), their egg recognition should favour a greater variety of different shell cues in comparison with species that exhibit lower recognition abilities. Variation in egg recognition by different hosts has been explained by different duration of co-evolution between the brood parasite and its host (Brooke and Davies,1988; Soler and Møller,1990). Although magpies live in sympatry with cuckoo, this species belongs to a naive host group with lower recognition abilities, that are not as sophisticated as those of blackcaps.

Our results show that not only UV wavelengths but also total brightness of experimental eggs within the 325–700 nm wavelength range significantly influenced host rejection behaviour. Our previous results on blackcaps showed that birds reject every egg that is even a little bit lighter than the host eggs and accept all eggs that are as dark as or much darker than their own host clutch. This may be connected with the spot concentration of the blackcap's eggshell (Makatsch,1976), which might theoretically serve as a fingerprint of individual females (Kilner,2006). Alternatively, if synthesis and allocation of pigments into the eggshell is costly (Moreno and Osorno,2003), the darkness of the blunt end of the egg may be linked to the quality and age of the bird (Moreno et al., 2006; Siefferman et al.,2006), which can be further reflected in the host rejection response (Lotem et al.,1995).

Our findings supported the significance of ultraviolet reflectance in the egg recognition processes of hosts. However, further studies are needed to disentangle the role of functional mechanism of the UV wavelengths in egg recognition for better understanding of this co-evolutionary brood-parasitism system.

The study was supported by grants no. 524/05/H536, IAA600930605 from Grant agency of the Academy of Sciences of the Czech Republic and by grant LC06073 from Ministry of Education, Youth and Sport. We thank P. Procházka and M. Pozgayová for their help with fieldwork. P. Procházka also made valuable comments. Experiments were conducted in accordance with current laws of the Czech Republic and the Academy of Sciences Animal Care Protocol.