POPULATION DIVERGENCE IN SEXUAL ORNAMENTS: THE WHITE FOREHEAD PATCH OF NORWEGIAN PIED FLYCATCHERS IS SMALL AND UNSEXY

Transcription

1 Evolution, 3(4), 1999, pp POPULATION DIVERGENCE IN SEXUAL ORNAMENTS: THE WHITE FOREHEAD PATCH OF NORWEGIAN PIED FLYCATCHERS IS SMALL AND UNSEXY SVEIN DALE, I TORE SLAGSVOLD,,3 HELENE M, LAMPE,,4 AND GLENNPETER SJETRE, ldepartment ofbiology and Nature Conservation, Agricultural University ofnorway, P.O. Box 14, N143 As, Norway svein.dale@ibn.nlh.no Department of Biology, University of Oslo, P.O. Box 1 Blindern, N31 Oslo, Norway 3 tore.slagsvold@bio.uio.no 4 helene.lampe@bio.uio.no g.p.satre@bio.uio.no Abstract.Models of sexual selection suggest that populations may easily diverge in male secondary sexual characters. Studies of a Spanish population of the pied flycatcher, Ficedula hypoleuca, and a Swedish population of the closely related collared flycatcher, F. albicollis, have indicated that the white forehead patch of males is a sexually selected trait. We studied the white forehead patch of male pied flycatchers (n = 48) in a Norwegian population over seven years. Males with large forehead patches were in general more brightly colored, but patch height was not correlated to body mass, body size, or parasite loads. Conditions during the nestling period did not seem to influence patch height as an adult. Patch height increased slightly from the first to the second year as adults, but then remained relatively constant at higher ages. Patch height was not related to survival. Yeartoyear changes showed that males who increased in patch height also increased in body mass, suggesting that expression of the forehead patch may be partly condition dependent. However, changes in body mass explained only a small proportion of the variance in patch height between males. Thus, patch height would not be a good indicator of male quality. Furthermore, patch size was also not related to male ability to feed nestlings, indicating that females would not obtain direct benefits by choosing males with large patches. However, patch height could be a Fisher trait, but this requires heritability and there was no significant fatherson resemblance in patch height. Comparisons of the males visited by each female during the mate sampling period indicated that chosen males did not have larger forehead patches than rejected males. Experimental manipulation of patch height did not affect male mating success. These results indicate that females do not use patch size as a mate choice cue. Finally, patch height did not predict the outcome of male contests for nestboxes, suggesting that the forehead patch is not an intrasexually selected cue of status. Norwegian pied flycatchers have smaller forehead patches than both Spanish pied flycatchers and Swedish collared flycatchers. We suggest that this pattern may be explained by the lack of sexual selection on the forehead patch in the Norwegian population as compared to the other populations, where the patch is apparently sexually selected. We discuss possible reasons for these population divergences, such as female choice on an alternative secondary sexual character (general plumage color) and speciation among Ficedula flycatchers. Key words.ficedula hypoleuca, mate choice, pied flycatcher, plumage color, population divergence, secondary sexual characters, sexual selection, signals, speciation. Sexual selection on male characters will often lead to the production of large or colorful ornaments (Andersson 1994). However, the exact traits that become targets of sexual selection may vary depending on, for example, environmental conditions and specific female preferences (Endler 19, 199; Lande 1981, 198; Ryan et al. 199; Schluter and Price 1993). Thus, there is substantial variation between species and even within species in the kinds oftraits that have become exaggerated (Kaneshiro and Boake 198; Shutler and Weatherhead 199; Marchetti 1993; Schluter and Price 1993; Young et al. 1994; Endler and Houde 199). Species and population divergences in ornaments may be permanent, but there is also the possibility that ornamental traits may vary continuously over time (Pomiankowski and Iwasa 1998). Prolonged selection may reduce genetic variation (Fisher 193; Gustafsson 198), and in the end all males could be equally, but highly ornamented. Conspecifics may then switch their attention to other male traits, for example, to assess male quality. Sexual selection on specific male characters may therefore proceed in cycles depending on the amount of variability among males and on the information value of the character (Schluter and Price 1993; Iwasa and Pomiankowski 199; Pomiankowski and Iwasa 1998) The Society for the Study of Evolution. All rights reserved. Received March 3, Accepted February 1, In Europe, two closely related species of Ficedula flycatchers have been well studied with regard to sexual selection (Sretre et al. 199). Male collared flycatcher (Ficedula albicollis) have a conspicuous large white forehead patch (mean about 8 mm), and the size of the patch is important in malemale contests and also affects female mate choice (Gustafsson et al. 199; Sheldon et al. 199; Part and Qvarnstrom 199; Qvarnstrom 199, 1998). Males in a Spanish population of the pied flycatcher (Ficedula hypoleuca) also have a conspicuous forehead patch (mean is 4 mm'), and females prefer males with large patches (Potti and Montalvo 1991a). In a Norwegian population of pied flycatchers, the forehead patch is rather small with a mean of about mm? (this study), but the importance of the forehead patch in sexual selection is not known. The three populations also differ in another secondary sexual character. Males in the Norwegian population show a continuous variation in plumage color (Slagsvold and Lifjeld 1988), and males generally become darker with age. Females prefer the darkest males (Sretre et al. 1994), probably because they then get older males, who are better parents than younger males (Seetre et al. 199). However, females in the Spanish population of pied flycatchers apparently do not choose mates on the basis of plumage

2 13 SVEIN DALE ET AL. coloration (Potti and Montalvo 1991a). Plumage color among Spanish males is variable, but males do not become darker with age (Potti and Montalvo 1991b), suggesting that females may not get better mates by choosing darker males. There is little color variation between males collared flycatchers (Reskaft and Jarvi 199; Cramp and Perrins 1993), and there have been no reports or suggestions of female choice on plumage color in this species. The pattern of variation in forehead patch size and general plumage color among Ficedula flycatchers in Europe seems to match what would be expected if sexual selection proceeds in cycles with one trait taking over the importance of another. In Spanish and Swedish flycatchers, the forehead patch is sexually selected and has become large, whereas general plumage color is not subject to selection. In Norwegian flycatchers plumage color is sexually selected, whereas the forehead patch is small and may therefore be predicted not to be sexually selected. In the present study, we tested the prediction of a lack of sexual selection on forehead patch size in Norwegian pied flycatchers. First, if the forehead patch is an intersexually selected indicator of male quality, one would expect condition dependence. We therefore analyzed the relationships between patch size and nestling conditions, male characters such as body size and body mass, age, parasite loads, ability to feed nestlings, and survival. We also analyzed changes in patch size in individual males from year to year in relation to changes in other male characters or body condition. Absence of significant relationships in the abovementioned analyses would be taken as evidence that the forehead patch is not an intersexually selected quality indicator. Alternatively, the forehead patch could be a Fisherian trait and then heritability of the trait would be required. Therefore, we compared patch size of fathers and sons to assess heritability. To assess intersexual selection directly, we compared chosen and rejected males sampled by individual females, and we also performed a forehead patch manipulation experiment. The hypothesis of a lack of sexual selection on the forehead patch predicts that no difference between chosen and rejected males should be found. Finally, the forehead patch could be an intrasexually selected badge of status. This was tested with an experiment of male contests for nestboxes, and our hypothesis predicted that male forehead patch size should not determine the outcome of the contests. METHODS The study was carried out during the breeding seasons of near Oslo, southeastern Norway. The study area consisted of three plots: Sinober, Tangen, and Brenna. The Sinober plot consisted of mixed forest, Tangen was mostly coniferous forest, and Brenna was deciduous forest. Males were captured soon after arrival and given individual combinations of color rings. Plumage color of males was scored on a scale from blackandwhite (score 1) to brownish and femalelike (score ; Drost 193). Halfscores were used for intermediate types, and mean values were used in cases of repeat measurements during the same year. Body mass was measured to the nearest.1 g and standardized to 9 h (the regression line for body mass in relation to time of day in o l ll).c E::J Z o o FIG. 1. Frequency distribution of forehead patch heights of male pied flycatchers. For males that had been measured in two or more years, patch height in the first year was used. creased with.4 g/h). Tarsus length was measured with a caliper to the nearest.1 mm with bent toes, and wing length to the nearest. mm using a flattened and straightened wing. Forehead patch height was measured with a caliper to the nearest.1 mm. The forehead patch varies in shape from rectangular to two separate dots (Lundberg and Alatalo 199). Measurements were taken where height was at a maximum. Forehead patch height was measured on all males except for those present in the Tangen plot in 199. A total of 48 different males were measured, and 111 of these were measured in two or more years. The number of males measured in each year was (first measurements, number including repeat measurements in parentheses): 199: 11 (11); 1991: (9); 199: 89 (11); 1993: (11); 1994: (89); 199: (); 199: 9 (). We measured forehead patch height, whereas studies on Spanish pied flycatchers and Swedish collared flycatchers have used patch area (rnm). The relationship between patch height and patch area was determined by measuring 3 skins from the collections at the Zoological Museum in Oslo of male pied flycatchers from Norway. Patch area was estimated as the product of mean patch height and patch width of each male. The regression of patch area on maximum patch height was y = 9.1x.3. Mean patch area (based on the mean of maximum patch height = 3. mm; Fig. 1) of males in our population is therefore estimated at mm. A few males completely lacked a forehead patch. The largest patch measured had a maximum height of. mm, whereas width would rarely be more than about 1 mm, giving a patch area range of about mm. In Spanish pied flycatchers mean patch size is 4 mm (range 18 mm; J. Potti pers. comm.; see also Potti and Montalvo 1991a; Potti and Merino 199). Swedish collared flycatchers have a mean patch size of about 8 mm (range 1 mm; Ellegren et al. 199; Part and Qvarnstrom 199). Male age was determined according to the method described by Karlsson et al. (198), which is based on the color and wear of the outermost greater wing coverts, the wear of wing and tail feathers, and the color of the inside of the upper 3 Forehead patch height (mm) 4 r,..,.1 I

3 POPULATION DIVERGENCE IN SEXUAL TRAITS 13 J: C) 'iii J: J: CO C FIG.. Forehead patch height of male pied flycatchers in relation to minimum age. Bars represent standard errors, and numbers on top are sample sizes. Males are included for every year they were measured. mandible. Previously unringed males were thereby classified as first year (i.e., second calendar year) or second year (third calendar year). We assumed that few males appeared in the study area for the first time as third year birds or older. Natal dispersal of nestlings ringed in our study area is too large to permit checking the accuracy of age determination. However, Lundberg and Alatalo (199) reported that they determined the age correctly for 9% of the males when using the method of Karlsson et al. (198). Most of the males in our study areas probably return if they survive (Slagsvoid and Lifjeld 199); therefore, return rate was used as an index of survival rate. Nestlings were ringed at an age of 13 days, and their body mass and tarsus length were measured at the same time. In some comparisons males from the Brenna plot were excluded because mean forehead patch height and other variables may differ from other plots due to the exclusion of latearriving males in Brenna after all vacant nestboxes were closed before the end ofmigration (see below). In cases where males were measured in several years, only the first measurement is used in overall comparisons unless otherwise stated. Patch heights were measured by SD, TS, HML, Trond Amundsen, Andrzej Kruszewicz, and three students. There were significant differences in mean measurements of males between persons, and the means differed with up to.8 mm. However, we have too few repeat measurements to test for differences in height recorded of the same males. We have no reason to believe that differences in measurements between workers introduced systematic errors in the dataset because the identity of the person measuring each male was random with respect to all important male variables as far as we can judge. Furthermore, the consistency in patch height of the same males in different years (Fig. ) when birds were often measured by different persons, suggests that measurement errors are small compared to betweenmale variation. Finally, the large sample sizes in most analyses will also reduce the problem that measurement errors can make it difficult to detect differences between groups of males. Blood samples were taken in 199 and Blood smears Age 8 9 were checked for the presence of the protozoan parasites Trypanosoma and Haemoproteus. Analyses were done using infection status of each male (infected or uninfected). Details of the parasitological analyses are given in Dale et al. (199). Male ability to feed nestlings was studied in 199 and The weight development of the brood was measured during a 4h period of enforced female absence, and males were compared in a pairwise design to control for effects of time of day, date, and weather. Changes in the body mass of the males themselves were also recorded. Further details of the experiments are given in Seetre et al. (199). Female mate choice was studied by releasing colorringed females into study plots where the nestboxes of males were monitored with video cameras. The mate sampling behavior of individual females could then be determined by recording at which nestboxes females appeared (for details, see Dale and Slagsvold 199a). We first present analyses of the influence of the forehead patch on female mate choice by comparing patch height of the chosen male with patch heights of males that were rejected. It has previously been shown that plumage color, nestbox quality, and male mating status influence female mate choice (Dale and Slagsvold 199a). Therefore, we also present analyses where we control for the effect of these choice criteria by comparing the chosen male with only rejected males that were at least as good as the chosen male regarding the three choice criteria. Thus, males that may have been rejected because they were dullcolored, had lowquality nestboxes, or were already mated (if the chosen male was a bachelor) were excluded from the set of males each female could choose from. Such statistical analyses were necessary because it is practically impossible to control for or manipulate male plumage color and male mating status in the field experiments themselves. However, we did control for nestbox quality in the field through the use of only one type of nestbox, except in a minority of the cases (3/1), where we manipulated the diameter of the entrance hole (Dale and Slagsvold 199a). Further details of analyses of female mate sampling behavior and female mate choice are given in Dale and Slagsvold (199a). In 1991 the forehead patch size of the 1 males present in the Brenna plot was manipulated. Males were randomly assigned to one of three treatment groups and subjected to treatment when they were captured early in the season for ringing and biometrical measurements. One group had the forehead patch enlarged by painting with white acrylic paint. Postmanipulation patch height was on average.1 mm (range..3, n = ). Another group had the forehead patch reduced by painting with dark acrylic paint. Postmanipulation patch height was on average 1.1 mm (range 1.1., n = 4). The last group of males was released after handling without changing patch size and served as controls (mean patch height was 4. mm; range 4.., n = 3). There was no difference in premanipulation patch height between groups (Kruskal Wallis test: H =.14, n = 1, P =.34). In the patch manipulation experiment, male mating success was measured as the time elapsing until a female was attracted after males had been restricted to one nestbox each on 3 May. In addition, we recorded the number of females attracted during the remaining part of the breeding season. This was done by daily inspections of the nestboxes of all

4 138 SVEIN DALE ET AL. males in the study plot, and the presence of nest material in the nestbox was used as an indication that a female had been attracted. All females were removed from males soon after settling and released in other areas. Female mate choice among manipulated males was also investigated by comparing chosen and rejected males as revealed by the use of video cameras as explained above. To assess the importance of the forehead patch in intrasexual selection, we induced male contests for nestboxes in aviaries. Each aviary had two compartments with one male defending one nestbox in each part, and a contest was initiated when the wall between the compartments was removed and the two nestboxes were replaced by one in the middle of the aviary (for further details of the experiment, see Seetre and Slagsvold 199). The identity of the winner was determined thereafter on the basis of which male took possession of the single nestbox. Parametric statistical tests were used unless distributions were obviously nonnormal in which case nonparametric tests were used. All tests report twotailed Pvalues. In cases where a lack of significant differences was interpreted as evidence that the forehead patch was not sexually selected, we performed power analyses to assess the reliability ofthe negative results. We report power as the percentage of tests that would reject the null hypothesis at a =. if there was a true relationship of "medium" magnitude (conventional levels given by Cohen 19). Power analyses were performed with the program G*Power (Buchner et al. 199). RESULTS Forehead Patch Height Mean forehead patch height of all males measured during the study was 3. mm (SD = 1.1, n = 48; Fig. 1). There was no significant variation between years (oneway ANO VA, F,48o = 1.13, P =.3). It has been suggested that habitat selection ofpied flycatchers is affected by competitive interactions between males (Lundberg et al. 1981), but we did not find any significant difference in patch height between males in a plot with mostly coniferous forest (Tangen; mean = 3., SD =.88, n = 93) and a plot with mixed forest (Sinober; mean = 3.4, SD = 1., n = 318; ttest: t = 1.3, P =.1). Data from different plots have therefore been pooled in subsequent analyses. AgeCorrelated Changes There was no significant overall change in patch height in relation to age (Fig. ; r s =.4, n =, P =.38). However, more detailed analyses were done using patch height of 111 males that were measured in two or more years. Patch height increased from the first to second year (Fig. 3A; paired ttest: t =.94, df =, P =.48; increase in mean height from 3. mm to 3.88 mm), but there were no further changes at older ages (Fig. 3BD; paired ttests: second to third year:t =.39, df =, P =., third to fourth year: t = 1.1, df =, P =.4; fourth to fifth year: t = 1., df = 1, P =.4). Measurements in two consecutive years were strongly correlated (Fig. 3; first vs. second year: r =.8, n =, P <.1; second vs. third year: r =.1, n = 1, P <.1; third vs. fourth year: r =., n =, P <.1, fourth vs. fifth year: r =.4, n = 11, P =.89). Yeartoyear changes for males in different age interval groups are shown in Table 1. The increase from the first to second year for males measured from the first to third year was significant (paired ttest: t =.8, df = 11, P =.1), all other comparisons were nonsignificant (P >.8). Patch height of males at the age of two years was larger for males that were present in the study area also at the age of one year (mean = 3.88, SD =.9, n = ) than for males that appeared in the study area first at the age of two years (mean = 3., SD =.91, n = 38; ttest: t =.43, df = 93, P =.1). There was a tendency that males had smaller forehead patches in the last year they were present in the study area than the year before; of 8 males, 3 decreased, increased, and three did not change in patch height (X = 4.9, P =.43; excluding 3 males that were only present as first and second year adults, Table 1). However, the decrease was small and not significant in pairwise comparisons (mean patch height in next to last year = 3.8 mm, SD = 1.; mean in last year = 3.1 mm, SD = 1.; paired ttest: t = 1.3, df =, P =.89). Correlates of Patch Height Simple regressions indicated that patch height increased with plumage color, that is, the darkest males had the largest patches (Table ). Among first year males there also seemed to be a correlation between patch height and wing length, but this relation disappeared in multiple regressions, which showed that only plumage color was correlated with patch height (Table ). However, the multiple regressions only explained 8% of the variation in patch height. Return Rate and Parasites There was no evidence that patch height was related to return rate (an index of survival; Table 3). There was no significant difference in patch height among first year males in relation to parasite infection status (Table 4). However, among males at least two years old there were nonsignificant trends that males infected with Haemoproteus had smaller patches and that males infected with Trypanosoma had larger patches than uninfected males (Table 4). Paternal Care Pairwise comparisons of male ability to feed nestlings in the absence of their mate indicated that males with large forehead patches did not provide more parental care than males with small patches. There was no significant relationship between patch size difference in each male pair compared and difference in change in brood mass (r =.1, n = 4, P =.48; the power of the test with a medium effect size [r =.3] was 49%). However, because male age and/ or plumage color affects male ability to provide parental care (Seetre et al. 199), it was necessary to control for these factors. In comparisons where both males were of the same age and plumage color there was still no relation between

5 POPULA TION DIVERGENCE IN SEXUAL TRAITS 139 A C) C) _. 1o. CO 1. 4 o CO 4 U~ U CO"C ~ 3 ~ CO LL LL Forehead patch height (first year) Forehead patch height (third year) B C) C) _ co 4 1o. 4 U _ U CO ;~ 1. CO~ Q.c: o 8 Q. 1. "Co "C::J co U COo _ (/) Q."C 3 Q "C. CO "C:::: CO; ~ LL LL Forehead patch height Forehead patch height (second year) (fourth year) FIG. 3. Forehead patch height of male pied flycatchers measured in one year in relation to their patch height in the previous year. Correlations between: (A) first and second year; (B) second and third year; (C) third and fourth year; and (D) fourth and fifth year. patch size differences and differences in brood mass change (r =.4, n =, P =.; the power of the test with a medium effect size [r =.3] was only 9%, but note that the observed correlation was in the opposite direction of what would have been expected if large patch size indicated more paternal care). Likewise, there was no evidence that males with large patches lost less weight during the period of feeding nestlings

6 14 SVEIN DALE ET AL. TABLE 1. Changes in forehead patch height from year to year for different classes of male pied flycatchers. Figures are mean and SE (in italics), except for classes with only one male, where patch height is given directly. Some males had been present in the study area one or more years before measurements of patch height began, thus the occurrence of age interval groups such as 9 years. (Another six males were measured in two years that were not consecutive.) Age Age (years) interval group (years) (.) 3.9 (.1) alone (i.e., when the female was held in captivity) than males with small patches. There was no relation between patch size differences and differences in male body mass changes in each male pair (r =.1, n = 4, P =.3), even when controlling for male age and plumage color (r =.3, n =, P =.9). Correlates of Changes in Patch Height Changes in patch height from the first to second year were not related to changes in other male characters (Table ). However, yeartoyear changes in patch height at higher ages were related to changes in plumage color and body mass; males that increased in patch height also became darker and heavier (Table ). This suggests that expression of the forehead patch is related to expression of plumage color, and that there is also an element of condition dependence. However, TABLE. Forehead patch height of male pied flycatchers in relation to their plumage color, body mass, tarsus length, and wing length. Age class Simple regression Multiple regression Character 11 P p First year Plumage color Body mass Tarsus length 4.13.Q Wing length Second year or older Plumage color Ql Body mass Tarsus length Wing length changes in plumage color and body mass explained only 1% of the changes in forehead patch size. The remaining variation in patch size changes may have been due to measurement errors or other factors. Changes in patch height from one year to another were also related to the change in patch height between preceding years. Males that increased in patch height from year y 1 to y, decreased in patch height from year y to y + 1 (r =., n = 1, P <.1). This relationship may have been due to random measurement errors, a cost of having a large patch, or both. Direct separation of these two explanations is not possible. However, changes in patch height between preceding years were also related to changes in body mass, but not to changes in plumage color, tarsus length, or wing length in the following year (Table ), at least for changes occurring after the second year. Thus, males that increased in patch height from year y 1 to y, decreased in body mass from year y to y + 1, suggesting a cost of having a large forehead patch. In a further analysis we included change in patch height between years y 1 and y as an independent variable in the multiple regression of change in patch height from year y to y + 1. This analysis showed that change in patch height between preceding years had a strong effect ([3 =., P =.), whereas there was a nearly significant relationship with change in plumage color ([3 =., P =.9). Changes in body mass, tarsus length, and wing length had no effects (P >.4 for all). The effect of changes in body mass probably disappeared because of its strong relation with change in patch height in preceding years. Effect of Nestling Conditions Patch height as an adult did not seem to be related to conditions during the nestling period. There was no signifi

7 POPULATION DIVERGENCE IN SEXUAL TRAITS 141 TABLE 3. Forehead patch height (mm) in relation to return rate of male pied flycatchers. Pvalues refer to ttests of the difference in patch height between returning and not returning males. TABLE. Changes in forehead patch height of male pied flycatchers in relation to changes in their plumage color, body mass, tarsus length, and wing length from one year to another. Age Returned Did not return class (year) Mean SE II Mean SE II P Power! % % % % %! Power was calculated for a medium effect size (d =.). cant correlations between patch height and nestling body mass (patch height measured as one year old: r =.14, n = 13, P =.; Fig. 4A; as two years old: r =.14, n = 14, P =.; Fig. 4B; eight males were included in both analyses) or tarsus length at an age of 13 days (patch height measured as one year old: r =.4, n = 1, P =.13; Fig. A; as two years old: r =.3, n = 13, P =.91; Fig. B; eight males were included in both analyses). The power of these tests were not very high due to small sample sizes (calculated for a medium effect size (r =.3): 119%). To increase sample sizes patch height as two years old was used when available and supplemented with patch height at other ages for remaining males. However, nestling conditions still did not seem to affect patch height (nestling body mass: r =., n = 1, P =.; nestling tarsus length: r =.14, n = 19, P =.8; power was 8% in these tests). Heritability There was no positive relationship between patch height of father and son (r =., n =, P =.3), even when the analysis was restricted to cases where patch height of father and son had been measured at the same age (r =.1, n =, P =.38; Fig. ). The power of these tests if there was a medium effect size (r =.3) was 9%. However, note that the observed results went in the opposite direction of what would be expected if patch height was heritable. Female Mate Choice Analyses of the mate choice of females whose mate sampling had been videotaped were made to investigate whether patch height influenced their choices. Without controlling for Simple regression Multiple regression Age class (years) Character II P II ~ p 1 Plumage color Body mass Tarsus length Wing length , 34, 4 Plumage color Body mass Tarsus length Wing length the effect of other mate choice cues, 9 females chose the male with the largest patch, whereas 3 females rejected the male with the largest patch (expected number of females settling with the male with the largest patch with random choice: 4., expected number of rejections: 3.4; X = 1.3, P =.). Patch height of chosen males tended to be greater than the average of rejected males, but the difference was not significant (Wilcoxon signedranks test: Z = 1.8, n = 1, P =.3; Fig. A). However, because patch height was correlated to plumage color, and females prefer males with a dark plumage (Seetre et al. 1994), it was necessary to control for the effect of plumage color (see Methods for how this was done). Then 1 of 33 females chose the male with the largest forehead patch (expected number of females choosing the male with the largest patch with random choice: 1.9, expected number of rejections: 1.91; X =.9, P =.9). Patch height of chosen males was no longer very different from the average of rejected males (Wilcoxon signedranks test: Z =.4, n = 33, P =.9). Furthermore, we controlled for the effect of other mate choice cues (mating status and entrance hole diameter of nestbox; Dale and Slagsvold 199a). In this final analysis, eight of 18 females chose the male with the largest forehead patch (expected number of females choosing the male with the largest patch with random choice:.9, expected number of rejections: 1.; X =., P =.98). Patch height of chosen males was not greater than the average of rejected males (Wilcoxon signedranks test: Z =.8, n = 18, P =.8; Fig. B; power in the case of a medium effect size [f =.] was 1% when a paired ttest was used as an approximation). TABLE 4. Effects of parasites on forehead patch height (mm) of male pied flycatchers. Pvalues refer to rtests of the difference in patch height between infected and noninfected males. TABLE. Changes in plumage color, body mass, tarsus length, and Infected Not infected wing length of male pied flycatchers from year y to y + 1 in relation Age class to changes in their forehead patch height in the previous year (i.e., Parasite Mean SE II Mean SE II P Power! from year y 1 to y). First year Haemoproteus % Year y = second year Year y = third year or older Trypanosoma % Character II P II P Second year or older Plumage color Haemoproteus % Body mass Trypanosoma % Tarsus length.q Wing length ! Power was calculated for a medium effect size (d =.).

8 14 SVEIN DALE ET AL. A A 4 4 o 3 3 E E E E.c C).c C).c 11.c c o s: o lu Q. lu Q. "C lu "C lu.c Jo...c B B c9 Jo U LL o '"...&..."...'' OL...'..."" ""_I Nestling body mass (g) FIG. 4. Relation between nestling body mass and forehead patch height of male pied flycatchers as first year birds (A) and as second year birds (B). In the patch height manipulation experiment, males with enlarged patches did not attract more females than control males and males with reduced patches (KruskalWallis test: H =.11, n = 1, P =.9; Table ), and males with enlarged patches did not attract females sooner (KruskalWallis test: H =.8, n = 1, P =.9; Table ). Unfortunately, only one female whose mate sampling behavior had been filmed could choose between males that differed in patch treatment. She chose a control male and rejected a male with enlarged forehead patch. Nestling tarsus length (mm) FIG.. Relation between nestling tarsus length and forehead patch height of male pied flycatchers as first year birds (A) and as second year birds (B). Male Contests Winners of contests between males over ownership ofnestboxes did not have larger forehead patches than losers (Wilcoxon signedranks test: Z =., n = 33, P =.). In these experiments there was also variation in male plumage color, and it was found that brown males won contests more often than black males (Seetre and Slagsvold 199). If contests between brown and black males were excluded, leaving contests between intermediately colored and black males

9 EE J: C) ') J: C 4 J:O ou) mo 3 Q. " m CI) J: CI)... LL. a a o 3 POPULATION DIVERGENCE IN SEXUAL TRAITS EE,c )._ ca E,c,cC u _U) ca O c.,c U '_ cao... s: o LL A ocp 8 3 Forehead patch height (mm) of father FIG.. Relation between forehead patch height offather and son in pied flycatchers. 4 o o 3 4,...T"'"""or.."T"'"""T"'"""T"'"""" (where color did not predict the outcome of the contests), there was still no difference in forehead patch size between winners and losers (Wilcoxon signedranks test: Z =.1, n = 1, P =.1; power in the case of a medium effect size [f =.] was 1% when a paired ttest was used as an approximation). 4 3 DISCUSSION In the present study we did not find evidence that the white forehead patch of males in a Norwegian population of pied flycatchers is sexually selected. First, the results indicated that patch size was not heritable or had a low heritability. Second, there was no direct evidence that females preferred males with large forehead patches. We tested for preferences by analysing the mate sampling records of individual females released in study areas where the nestboxes of all males were monitored with video cameras. Therefore, we were able to compare the male chosen by a female with the males she had rejected. This provides one of the most powerful methods for determining whether females have preferences for specific male traits, and the method has revealed female preferences for other male traits in pied flycatchers (Dale and Slagsvold 199a). In addition, the patch size manipulation experiment did not provide any evidence for female choice on patch size. However, there was some evidence relating to yeartoyear changes in patch size that suggested that expression of the trait was partly condition dependent and thus could be used by females in mate choice. Males that increased in patch size from one year to another also increased in body mass, which could indicate that only the most healthy males were able to increase in patch size. Furthermore, males that increased in patch size from one year to another decreased in body mass o o 3 Mean forehead patch height (mm) of rejected males FIG.. Forehead patch height of males chosen as mates by female pied flycatchers in relation to mean patch height of rejected males as revealed by videotapes of female mate sampling behavior. No control for other mate choice cues (A) and control for the effect of male plumage color, male mating status, and nest site quality on female choice (B). to the following year, suggesting a cost of having a large patch. Finally, males that were present in the breeding area as firstyear birds had a larger patch as secondyear birds than males that were first seen in the breeding areas as secondyear birds. These three findings suggest that a large forehead patch may signal good condition of the bearer. However, in most cases changes in patch size from yeartoyear were small, and, although significant, the relationships with body mass were weak. Females would therefore not gain very pre 4

10 144 SVEIN DALE ET AL. TABLE. Effect of manipulation of forehead patch height of male pied flycatchers on the total number of females attracted and the number of days elapsing until the first female was attracted. Data refer to a starting date of 3 May, when all males were restricted to one nestbox each. Male group Enlarged patch (..3 mm) Male 1 Male Male 3 Male 4 Male Reduced patch (1.1. mm) Male Male Male 8 Male 9 Control (4.. mm) Male 1 Male 11 Male 1 Premanipulation patch height (mm) No. of females attracted Days until first female cise information on male quality by using the size of the forehead patch as a mate choice cue. In addition, analyses of the effect of nestling conditions, male survival, blood parasites, and male parental care did not support the hypothesis that a large forehead patch is an indicator of male quality. We did not find evidence that the patch was influenced by intrasexual selection either. Intrasexual selection in the pied flycatcher is mostly linked to competition for nestboxes, and forehead patch size did not determine the outcome of contests for nestboxes. This parallels previous experiments that have shown that darkcolored males are not more likely to win a contest than dullcolored males (Slagsvold and Lifjeld 1988; Lundberg and Alatalo 199; Huhta and Alatalo 1993; Seetre and Slagsvold 199). In general, the outcome of contests for nestboxes will usually be determined by an ownerintruder asymmetry (Maynard Smith and Parker 19; Dale and Slagsvold 199). We also observed that there was no difference in patch size between males in mixed versus coniferous forest, even though it has previously been suggested that habitat settlement is the result of competitive interactions with dominant (larger) males occupying the best habitats (forest with deciduous trees; Lundberg et al. 1981). Taken together, these findings make it unlikely that intrasexual selection in our population affects forehead patch size. In conclusion, the forehead patch of Norwegian pied flycatchers does not seem to be sexually selected. Even though some of our tests had a low statistical power, other tests were based on large sample sizes. Some tests also showed trends in the opposite direction of what would be expected if patch size was sexually selected. Given the large number of independent tests made that encompassed almost all comparisons that are relevant in studies of sexual selection, we feel confident that the general absence of significant results can be taken as evidence that sexual selection is not operating 8 8 on the forehead patch in our population at present. This contrasts with the studies of Spanish pied flycatchers and Swedish collared flycatchers, where the forehead patch is larger and subjected to sexual selection. The present study together with the previous ones on Spanish and Swedish flycatchers suggest that the intensity of sexual selection may have affected the size of a secondary sexual ornament, the forehead patch, differentially between populations. The reason why the forehead patch is not subject to sexual selection in the Norwegain population may be that general plumage color is more important in this population than in the Spanish and Swedish populations (Seetre et al. 1994). One possible scenario is that female choice has operated on plumage color in the latter populations before, but that betweenmale variation is now limited and the trait has lost its value for females. If one sexually selected trait drops in information value, another trait may take over as the main cue for female choice (Schluter and Price 1993). It will be a matter of speculation what the ultimate cause of these populational differences are, but character displacement through reinforcement of speciesspecific mate choice in areas of sympatry of the pied and the collared flycatcher may explain why variation in plumage color in the collared flycatcher has been depleted (Seetre et al. 199). In the Norwegian population there is variation in male coloration perhaps due to gene flow from areas of sympatry with the collared flycatcher where pied flycatchers are brown (Lundberg and Alatalo 199). Because dark male plumage and possibly also female preference for dark males may be the ancestral state in the Ficedula flycatcher group in Europe (Seetre et al. 199), selection against brown males will be strong and may override selection on forehead patch size. The Spanish population of pied flycatchers is far from areas of sympatry with collared flycatchers, and selection may be at an equilibrum with respect to plumage color. Selection on another 'trait such as forehead patch size may thereby have been facilitated. This idea is supported by analyses of mtdna, which suggest an absence of gene flow between populations in Spain and central Europe (i.e., the area of sympatry with the collared flycatcher), but that gene flow occurs between Norwegian and central European populations (G.P. Seetre and T. Moum, unpubl. data). Other possible causes of population divergence in sexual selection could be environmental differences or random events (see references in Introduction). Why do male pied flycatchers in Norway have a white forehead patch at all if it does not convey an advantage in sexual selection? Increased risk of predation might be expected to favor cryptically patterned males (Slagsvold et al. 199) and thus reduce patch size to a minimum. However, males may need to have a minimum amount of contrast in their plumage to be easily detected by females or male competitors (Dale and Slagsvold 199, 199b), which may explain why very few males lacked the forehead patch completely. Furthermore, the forehead patch may be necessary in the courtship of cavity nesting flycatchers where demonstration of the nestbox is important (Curio 199). During courtship, the male often tries to attract the female to the nest hole sitting in the opening with only the head showing from the outside (Curio 199). Interestingly, another European cavity nester, the redstart Phoenicurus phoenicurus, also

11 POPULATION DIVERGENCE IN SEXUAL TRAITS 14 has a white forehead patch. Phylogenetic analyses should be done to test the idea that a male frontal patch is associated with cavity nesting. In conclusion, there does not seem to be any evidence that sexual selection favors males with larger white forehead patches in this Norwegian population of pied flycatchers. This may explain why the patch is relatively small compared to that of Spanish pied flycatchers and Swedish collared flycatchers, where the patch is subject to sexual selection. The Ficedula flycatchers may therefore provide an example of divergence in secondary sexual characters between closely related populations (Endler 19; Lande 1981, 198; Seetre et al. 199). ACKNOWLEDGMENTS We are grateful to T. Amundsen, G. Bergendal, T. Fossnes, M. Gunther, R. Karlsen, A. Kruszewicz, H. Rinden, and H. Viljugrein for help in the field; to A. Kruszewicz for parasitological analyses; and to J. Potti and A. Qvarnstrom for comments on the manuscript. Financial support was provided by the Norwegian Research Council and the Nansen Foundation. LITERATURE CITED Andersson, M Sexual selection. Princeton Univ. Press, Princeton, NJ. Buchner, A., F. Faul, and E. Erdfelder G*Power: a priori, posthoc, and compromise power analyses for the Macintosh. Verso.1.1. University of Trier, Trier, Germany. Cohen, J. 19. Statistical power analysis for the behavioral sciences. Academic Press, New York. Cramp, S., and C. M. Perrins, eds The birds of the western Palearctic. Vol. VII. Oxford Univ. Press, Oxford. Curio, E Verhaltensstudien am Trauerschnapper, Z. Tierpsychol. Beih. 3: Dale, S., and T. Slagsvold Random settlement of female pied flycatchers, Ficedula hypoleuca: significance of male territory size. Anim. Behav. 39: Female contests for nest sites and mates in the pied flycatcher Ficedula hypoleuca. Ethology 99: a. Mate choice on multiple cues, decision rules and sampling strategies in female pied flycatchers. Behaviour 133: b. Plumage coloration and conspicuousness in birds: experiments with the pied flycatcher. Auk 113:8498. Dale, S., A. Kruszewicz, and T. Slagsvold Effects of blood parasites on sexual and natural selection in the pied flycatcher. J. Zool. 38: Drost, R Uber das Brutkleid mannlicher Trauenfliegenfanger, Muscicapa hypoleuca. Vogelzug : Ellegren, H., L. Gustafsson, and B. C. Sheldon Sex ratio adjustment in relation to paternal attractiveness in a wild bird population. Proc. Natl. Acad. Sci. USA 93: Endler, J. A. 19. Geographic variation, speciation, and clines. Princeton Univ. Press, Princeton, NJ Signals, signal conditions, and the direction of evolution. Am. Nat. 139:S1S13. Endler, J. A., and A. E. Houde Geographic variation in female preferences for male traits in Poecilia reticulata. Evolution 49:448. Fisher, R. A The genetical theory of natural selection. Clarendon Press, Oxford. Gustafsson, L Lifetime reproductive success and heritability: empirical support for Fisher's fundamental theorem. Am. Nat. 18:14. Gustafsson, L., A. Qvarnstrorn, and B. C. Sheldon Tradeoffs between lifehistory traits and a secondary sexual character in male collared flycatchers. Nature 3: Huhta, E., and R. V. Alatalo Plumage colour and malemale interactions in the pied flycatcher. Anim. Behav. 4:1118. Iwasa, Y., and A. Pomiankowski Continual change in mate preferences. Nature 3:44. Kaneshiro, K. Y., and C. R. B. Boake Sexual selection and speciation: issues raised by Hawaiian Drosophila. Trends Ecol. Evol. :1. Karlsson, L., K. Persson, and G. Wallinder Ageing and sexing in pied flycatchers, Ficedula hypoleuca. Vih. Fagelv, 4: Lande, R Models of speciation by sexual selection on polygenic traits. Proc. Natl. Acad. Sci. USA 8: Rapid origin of sexual isolation and character divergence in a cline. Evolution 3:133. Lundberg, A., and R. V. Alatalo The pied flycatcher. T. and A. D. Poyser, London. Lundberg, A., R. V. Alatalo, A. Carlson, and S. Ulfstrand Biometry, habitat distribution and breeding success in the pied flycatcher Ficedula hypoleuca. Ornis Scand. 1:89. Marchetti, K Dark habitats and bright birds illustrate the role of the environment in species divergence. Nature 3: Maynard Smith, J., and G. A. Parker. 19. The logic ofasymmetric contests. Anim. Behav. 4: 191. Part, T., and A. Qvarnstrom Badge size in collared flycatchers predicts outcome of male competition over territories. Anim. Behav. 4: Pomiankowski, A., and Y. Iwasa Runaway ornament diversity caused by Fisherian sexual selection. Proc. Natl. Acad. Sci. USA 9 :1111. Potti, J., and S. Merino Decreased levels of blood trypanosome infection correlate with female expression of a male secondary sexual trait: implications for sexual selection. Proc. R. Soc. Lond. B BioI. Sci. 3: Potti, J., and S. Montalvo. 1991a. Male arrival and female mate choice in pied flycatchers Ficedula hypoleuca in central Spain. Ornis Scand. : b. Male colour variation in spanish pied flycatchers Ficedula hypoleuca. Ibis 133:9399. Qvarnstrom, A Experimentally increased badge size increases male competition and reduces male parental care in the collared flycatcher. Proc. R. Soc. Lond. B BioI. Sci. 4: Sexual selection in the collared flycatcher (Ficedula albicollis): a life history perspective. Ph.D. diss., Uppsala University, Uppsala, Sweden. Reskaft, E., and T. Jarvi Interspecific competition and the evolution of plumagecolour variation in three closely related old world flycatchers Ficedula spp. J. Zool. 8:13. Ryan, M. J., J. H. Fox., W. Wilczynski, and A. R. Rand Sexual selection for sensory exploitation in the frog Physalaemus pustulosus. Nature 343:. Szetre, G.P., and T. Slagsvold The significance of female mimicry in male contests. Am. Nat. 14: Ssetre, G.P., S. Dale, and T. Slagsvold Female pied flycatchers prefer brightly coloured males. Anim. Behav. 48: Sretre, G.P., T. Fossnes, and T. Slagsvold Food provisioning in the pied flycatcher: do females gain direct benefits from choosing bright coloured males? J. Anim. Ecol. 4:13. Seetre, G.P., T. Mourn, S. Bures, M. Kral, M. Adamjan, and J. Moreno A sexually selected character displacement reinforces premating isolation. Nature 38:899. Schluter, D., and T. Price Honesty, perception and population divergence in sexually selected traits. Proc. R. Soc. Lond. B BioI. Sci. 3:111. Sheldon, B. c., J. Merila, A. Qvarnstrorn, L. Gustafsson, and H. Ellegren Paternal genetic contribution to offspring condition predicted by size of male secondary sexual character. Proc. R. Soc. Lond. B BioI. Sci. 4:93. Shutler, D., and P. J. Weatherhead Targets of sexual selec

12 14 SVEIN DALE ET AL. tion: song and plumage of wood warblers. Evolution 44: Slagsvold, T., and J. T. Lifjeld Plumage colour and sexual selection in the pied flycatcher Ficedula hypoleuca. Anim. Behay. 3: Return rates of male pied flycatchers: an experimental study manipulating breeding success. Pp in J. Blondel, A. Gosler, J. D. Lebreton, and R. McCleery, eds. Population biology of passerine birds. SpringerVerlag, Berlin. Slagsvold, T., S. Dale, and A. Kruszewicz Predation favours cryptic coloration in breeding male pied flycatchers. Anim. Behav.: Young, J. R., J. W. Hupp, J. W. Bradbury, and C. E. Braun Phenotypic divergence of secondary sexual traits among sage grouse, Centrocercus urophasianus, populations. Anim. Behav. 4: Corresponding Editor: P. Gowaty