Marine Biology (2005) 148: 189–196 DOI 10.1007/s00227-005-0056-8
R ES E AR C H A RT I C L E
Christophe Pe´labon Æ A˚sa A. Borg Æ Jens Bjelvenmark Iain Barber Æ Elisabet Forsgren Æ Trond Amundsen
Do microsporidian parasites affect courtship in two-spotted gobies?
Received: 31 March 2004 / Accepted: 20 May 2005 / Published online: 13 August 2005 Springer-Verlag 2005
Abstract Courtship and body condition of male twospotted gobies (Gobiusculus flavescens) harbouring naturally acquired microsporidian infections were compared with those of microsporidian-free males under standardized conditions in the lab. Although parasite infection had no apparent effect on individual condition, it significantly affected male courtship intensity. This effect, however, was affected by the intensity of the female courtship. Our study is one of the first to demonstrate a sub-lethal behavioural effect of a microsporidian parasite that could negatively affect the reproductive success of infected individuals. Our results also suggest that secondary sexual traits like courtship may be more sensitive to moderately detrimental effects of parasite infection than classical condition indices.
Introduction Parasite infections can dramatically affect the population dynamics of their hosts by reducing host recruitment or survival (Minchella and Scott 1991; Poulin 1999). However, parasites may also negatively affect the mating success of their host either by castration mechanisms (Hurd 2001) or by the alteration of sexual signals such as ornaments or behavioural displays (Kennedy et al. 1987; Communicated by M. Ku¨hl, Helsingør C. Pe´labon (&) Æ A˚sa A. Borg Æ E. Forsgren Æ T. Amundsen Department of Biology, Norwegian University of Science and Technology, 7491 Trondheim, Norway E-mail:
[email protected] J. Bjelvenmark Department of Marine Ecology, Go¨teborg University, Kristineberg Marine Research Station, Fiskeba¨ckskil, Sweden I. Barber Institute of Biological Sciences, University of Wales, Aberystwyth, UK
reviewed in Barber et al. 2000). By utilising nutrients normally used by their hosts for maintenance or by deteriorating specific host tissues, parasite infections can negatively affect condition-dependent secondary sexual traits (Hamilton and Zuk 1982; Folstad and Karter 1992; Houde and Torio 1992; Andersson 1994; Thompson et al. 1997) including courtship display (e.g. Borgia et al. 2004). Parasite infections may have complex effects on courtship display. Because selection should favour individuals that identify and select against parasitized mating partners, conspicuous infections can negatively affect attractiveness and mating success even in the absence of any dramatic debilitating effect on host condition (Milinski and Bakker 1990; Dugatkin et al. 1994; Rosenqvist and Johansson 1995; Barber et al. 2000). By incorporating individual differences in attractiveness into theoretical models of condition-dependent sexual selection, Reynolds (1993) suggested that morphologically less attractive individuals might increase their attractiveness by courting at a higher rate, thus rendering courtship dishonest with respect to individual condition. Therefore, an increase in courtship rate by parasitized males may be expected in the case of conspicuous parasite infections with low virulence. An increase in courtship rate by parasitized individuals may also be expected if parasite infections reduce host survival. Normally, investment in a reproductive attempt is made at a sub-maximum level due to the trade-off between current and future reproduction (Williams 1966). Individuals infected with parasites that reduce longevity, and hence the prospect of future reproduction, may be selected to invest more heavily in the current reproduction, a phenomenon referred to as ‘‘terminal investment’’ (Forbes 1993; Polak and Starmer 1998; Candolin 2000). Because parasites might affect the amount of energy available for mate sampling, or interfere with sensory performance, infections can also affect mating preferences of parasitized individuals, hence reducing their choosiness (Poulin 1994; Poulin and Vickery 1996; Lopez 1999; but see Zuk et al. 1998). Also, any decrease in
190
the attractiveness of parasitized individuals and the potential loss of mating opportunities associated with it could affect the costs of being choosy. Therefore, parasitized individuals are expected to be less choosy than parasite-free ones (Poulin 1994). In this study, we tested whether visually conspicuous microsporidian (Phylum Microsporida) infections in male two-spotted gobies (Gobiusculus flavescens) affected their courtship behaviour. We used a study in which parasite-free males were allowed to choose between large and small females as a control (Pe´labon et al. 2003), and tested, in simultaneously performed trials, whether males infected by microsporidian parasites behave differently in terms of courtship intensity or preference. We also tested whether parasite infection was associated with poorer condition. Microsporidians are obligate intracellular parasites that infect a wide range of invertebrate and vertebrate hosts including fish (Dunn and Smith 2001; Lom and Nilsen 2003). In fish, microsporidians can infect several organs such as liver, skin, gills, or muscles (Lom and Dykova´ 1992; Lom and Nilsen 2003). After being swallowed or inhaled by a new host, spores reach the intestine where they penetrate host cells and subsequently migrate to reach the final tissue site (Bush et al. 2001; Dunn and Smith 2001). The parasite transforms the infected cell to form a cyst-like structure, the xenoma, in which uninucleate spores are produced (Dunn and Smith 2001). In the two-spotted goby, microsporidian parasites infect the musculature of the fish, inducing white patches highly visible through the skin (Fig. 1). The taxonomy and some pathological effects of microsporidian parasites were extensively studied in other host species (Figueras et al. 1992; Matthews et al. 2001; reviewed in Lom and Nilsen 2003). However, to our knowledge, no studies have so far tested the effect of microsporidians on fish mating behaviour.
Methods Study species The two spotted goby is a small (40–50 mm long) fish inhabiting shallow waters along rocky shores of the Fig. 1 Male two-spotted goby infected with microsporidian parasites
Baltic and European Atlantic coasts (Miller 1986). During a single breeding season (May to August in the studied population), males defend nest sites in empty bivalve shells or crevices in the algal vegetation (Gordon 1983; Forsgren et al. 2004) to which they attract females with a series of courtship displays (Amundsen and Forsgren 2001; Pe´labon et al. 2003; Forsgren et al. 2004). After spawning, egg care (i.e. fanning, cleaning and nest defence) is performed exclusively by the male (Skolbekken and Utne-Palm 2001; Bjelvenmark and Forsgren 2003). Male two-spotted gobies are slightly larger than females, and display highly developed dorsal fins with iridescent blue markings and iridescent blue spots along the body sides. Females develop colourful yellow–orange bellies during breeding (Amundsen and Forsgren 2001). Males display only a weak preference for large females (Pe´labon et al. 2003), but a clear preference for colourful females, at least late in the season (Amundsen and Forsgren 2001). Small males appear less discriminating than large ones (Amundsen and Forsgren 2003). Study design The study was conducted at Kristineberg Marine Research Station on the west coast of Sweden (5815¢N, 1127¢E) from 20 June until 15 July 2001. Fish were hand-netted by a snorkeler in waters adjacent to the research station. Before being used in trials, fish were kept in 75-l single-sex aquaria. We offered males a choice between two stimulus females of different size, controlling for female colour and stage of egg maturation (i.e. roundness, see below). Trials were conducted in threecompartment aquaria (60 l), with the male placed centrally and one female placed on each side, randomly in regard to their length. Test fish were separated by transparent Plexiglas partitions perforated to allow water flow (see Pe´labon et al. 2003 for a detailed description of the study protocol). Because courtship is generally interactive, male courtship is likely to be distorted if females are not allowed to respond sexually. Therefore, we allowed the two sexes visual access to each other (instead of using one-way mirrors). Although this set-up hampers our ability to infer causality in some of
191
the relationships between male courtship, female courtship and the parasite load of the males, it allows us to test whether microsporidian infection affects male courtship. We have previously reported the results of trials testing the preference of parasite-free males (i.e. males free of visible microsporidian infection) in relation to female size (Pe´labon et al. 2003). Here, we report the results from trials conducted simultaneously, in which the responding males were naturally infected with microsporidian parasites, and we compare the results from the two sets of trials. The trials lasted 20 min during which we continuously recorded three different courtship behaviours for the male: Fin display, Shiver and Lead. In short, courtship is initiated by a male swimming up to the female conspicuously erecting his fins (fin display). Courtship may then be escalated by the male swimming in parallel with the female, shivering his body (shiver), and he may finally swim towards his nest in a particular undulating manner (lead) (Amundsen and Forsgren 2001; Pe´labon et al. 2003). Females may perform a courtship display in which they bend the abdomen so as to display the belly maximally to the male (Amundsen and Forsgren 2001). Similar female courtship behaviour was observed in other fish species including gobies and was described as a ‘‘sigmoid-display’’ (Swensson 1997; Takahashi 2000). This sigmoid display typically occurs after initial solicitations by the male when the courtship intensity has already escalated. We continuously recorded the occurrence of female sigmoid display during the 20 min of observation. The male courtship intensity was calculated as the total number of displays performed by the male to both females during the 20-min trial. The time spent by the male within response zones £ 5 cm from the female
compartments was estimated by scan-sampling male positions every 20 s. The positions of both females (in or out of the zones £ 5 cm from the male compartment) were recorded in the same way. A switching rate (between females) was calculated as the number of times the male was observed in different response zones between two consecutive scan-samplings during the 20 min of the trial. This variable may be indicative of the sampling effort of the male but underestimates the true switching rate because males may have sometimes switched but returned to the previous response zone during the 20 s sampling intervals. We measured the total length of all test fish (to the nearest 0.5 mm) using a grid, and recorded their wet body mass (to the nearest 0.01 g) using a digital balance. There was no significant length difference between the two categories of males (Table 1) and no significant difference in the time the fish were held in captivity before the start of the trial (parasite-free males: median=2 days (range 1–7); parasitized: 2 days (0–7), Mann–Whitney test, Z= 0.29, P=0.77). The average length of stimulus females did not significantly differ between the two sets of trials nor did the length difference between the two females within trials (Table 1). Females become more round during the maturation of the eggs. We used the residuals of the regression of body mass over total length (on a log-scale) as an index of roundness to compare and match small and large stimulus females for the maturation stage of their eggs. There was no significant difference in the roundness between small and large females within or between the two series of trials (ANCOVA: length · size category · trial F1, 164=0.08, P=0.77, size category · trial F1, 164=2.83, P=0.09, length · size category F1, P=0.88, length · trial F1, 164=1.59, 164=0.10, P=0.64). Furthermore, the belly of the stimulus females
Table 1 Descriptive statistics (mean ± SE) of two-spotted gobies used to test for effects of microsporidian infection on male courtship and mate preference (n=number of trials). Comparisons between the two sets of trials with parasite-free or parasitized males are provided in the last column Parasite-free males (n=56) Male Total length (mm) Mass (g) Condition factor HSI
47.8±0.40 0.81±0.02 0.0053±0.007
47.6±0.60 0.81±0.03 0.0094±0.010
2.07±0.17 Small
Female Total length (mm) Mass (g) Rednessb Length difference (mm)
Parasitized males (n=30)
1.70±0.12 Large
Small
43.0±0.20 47.9±0.20 0.68±0.01 0.89±0.01 0.58±0.002 4.87±0.15
a Results of the ANCOVA with body mass as response variable, body length as covariate and parasite status as factor
Test
F1,85=0.094; P=0.76 a parasite: F1,82=0.34, P=0.56 parasite · length: F1,82=0.27, P=0.61 t=1.77; df=31; P=0.09.
Large
42.6±0.30 47.7±0.37 0.66±0.02 0.88±0.02 0.58±0.004 5.10±0.19 b
Test F1,84=0.18; P=0.67 See text F1,84=0.56; P=0.46 F1,84=0.64; P=0.42
See Pe´labon et al. (2003) for details of colour quantification
192
presented to parasite-free and parasitized males did not significantly differ in colour (Table 1). Parasite infection and condition indices Individuals infected with microsporidian parasites are readily identified by the white, subcutaneous patches, often several millimetre in diameter, where muscle fibres are degenerated (Fig. 1). In the studied population, both males and females were infected by microsporidians with a prevalence of ca. 25% (June–July 2001; C. Pe´labon, I. Barber and T. Amundsen, unpublished data). We counted the number of distinct white muscle patches visible through the skin and used this as an index of infection level. In this study, males referred to as parasitized showed more than five separate and distinct patches (range 6–60, median=10), while parasite-free males did not have any distinct patches. All females used as stimulus fish were uninfected. An analysis of the parasite fauna of the studied two-spotted goby population (210 females and 56 males) revealed infections by six other common macroparasites, four of which were of significant prevalence in males (C. Pe´labon, I. Barber and T. Amundsen, unpublished data). However, none of these macro-parasite infections were correlated with the number of microsporidian-induced muscle patches. Therefore, differences observed in the present study between males with and without microsporidian infection do not reflect correlated infections by any of these other macroparasites. We estimated an index of male body condition (condition factor, CF) as the residuals from the regression of body mass over total length (on a log-scale). For a sub-sample of males dissected for parasite analysis (n=31), we weighed the liver (wet weight, to the nearest 0.001 g) and calculated the hepatosomatic index (HSI: 100 · liver mass/body mass) as a second condition index. Both indices are expected to reflect the energy reserves. The former non-invasive method roughly reflects the total energy reserves, while the second reflects the glycogen reserves and may be more sensitive to shortterm variations in condition (Chellappa et al. 1995). Notably, there was no significant correlation between these two indices (r=0.05, P=0.77).
(infected or not) as factor and the female courtship (the total number of sigmoid displays performed by the two stimulus females) as a covariate. Both continuous variables (male and female courtship) were log-transformed [log(X+1)] to approach normal distribution. Note that, although the linear model of the ANCOVA suggests a known direction of causality (female courtship stimulating the males rather than vice versa), the correlational nature of the data implies that it cannot be told from the present study if the females stimulate the males, the males stimulate the females, or both sexes stimulate each other mutually. Residuals and cook distances of the model were inspected and revealed no outlier or observation with especially high leverage effect. All analyses were performed in S-Plus 6.2 (Venables and Ripley 2002).
Results Parasite infection and male condition We found no significant correlation between microsporidian infection and male condition factor (CF), while the hepatosomatic index (HIS) tended to be lower, though not significantly, in parasitized males (Table 1). Furthermore, we found no effect of male condition factor or hepatosomatic index on any of the courtship displays (Spearman correlations with CF: 0.03 < q < 0.08, all P>0.48; with HSI: 0.230.25). Male mate preference Parasite-free males showed, at most, only a weak preference for large females (Pe´labon et al. 2003). Results for the parasitized males were very similar, with no significant preference for large females (Table 2). However, because of the weak preference among parasite-free males, the potential for detecting any reduction in choosiness among parasitized males, as suggested by Poulin (1994), is very limited. Therefore, preference patterns will not be further analysed or discussed. Parasite infection and courtship
Statistical analysis We conducted a total of 86 trials, 56 with parasite-free males and 30 with parasitized males as respondents. Trials in which the male did not perform any courtship display were discarded from the analysis (n=2 for parasite-free males and n=5 for parasitized males). Because, in an interactive setting, male courtship can be affected by female courtship in addition to male parasite status, the relationships between these variables were analysed using an analysis of covariance (ANCOVA) on the male courtship intensity with the male parasite status
Parasitized males showed on average a 30% decrease in courtship rate (mean ± SE total number of displays per 20 min trial; parasite-free males: 53.93 ± 7.17; parasitized males: 39.32 ± 11.32). The three courtship displays adopted by males were all negatively affected by parasite status, and very similarly so (Fig. 2). Male and female courtship rates were strongly correlated (Fig. 3), as might be expected in a set-up where the sexes are allowed to interact. When we accounted for this correlation, we found a significant interaction between the effect of the parasite infection on the male courtship and
193 Table 2 Mating preference for larger females among male two-spotted gobies infected with microsporidian parasites and comparison with similar data for parasite-free males (Pe´labon et al. 2003). Proportions were tested versus an expectation of 0.50 (50%), after transformation [Arcsin(P)1/2]. The mean difference corresponds to the mean proportion of displays directed at large females by parasite-free males minus the same proportion for parasitized males (parasite-free males: n=54, parasitized males: n=25) Proportion with/or to the larger female
Time in response zone Fin display Shiver Lead
Comparison with parasite-free males (Mann– Whitney U)
Mean ± SE
t (df)
P
Mean difference
Z (n)
P
0.50±0.05 0.58±0.07 0.49±0.11 0.66±0.12
0.99 (24) 1.09 (24) 0.14 (12) 1.48 (10)
0.33 0.28 0.89 0.16
0.07 0.00 0.08 0.11
1.31 (84) 0.16 (77) 0.62 (47) 1.15 (38)
0.19 0.87 0.53 0.25
Fig. 2 Number of courtship displays (mean + SE) performed by male two-spotted gobies that were uninfected (open bars, n=54) or infected (grey bars, n=25) by microsporidian parasites during male mate choice trials of 20-min duration
the intensity of the female display (Fig. 3). The difference in the courtship rate between parasitized and parasite-free males was stronger when female courtship intensity was low. We further tested whether heavily parasitized males had a lower courtship rate than the more lightly infected. We found no significant negative relationship among parasitized males between infection intensity (number of visible white patches) and courtship rate (Spearman correlation: q=0.16, n=25, P=0.44). Parasitized and parasite-free males spent a similar proportion of their time in the response zones near the stimulus females (mean ± SE parasite-free males=0.54±0.03; parasitized male=0.50±0.05; t=0.79, df=77, P=0.42). Parasitized males, however, showed a lower switching rate, moving from one response zone to the other less often than the parasite-free males (mean ± SE switching rate: parasite-free males=16.7±0.7; parasitized males=14.3±0.9, t=2.01, df=77, P=0.04).
Fig. 3 Relationship between female and male courtship rate of two-spotted gobies, for males parasitized (open dots, dash line) or not parasitized (black dots, black line) by microsporidians. Numbers are total numbers of displays during 20 min. Both male and female courtship rates are log-transformed [ln(x+1)]. Coefficients ±SE and associated P-value: Parasite infection= 0.42±0.15; P
