Oecologia (1997) 110:501±507
Ó Springer-Verlag 1997
Eystein Markusson á Ivar Folstad
Reindeer antlers: visual indicators of individual quality?
Received: 8 October 1996 / Accepted: 13 January 1997
Abstract Fluctuating asymmetry (FA) in ornamental characters may re¯ect developmental stability in the translation from genotype to phenotype. Antlers of reindeer show FA, are visually conspicuous ornaments and are important in intraspeci®c assessment. We show that there is a negative relationship between size and asymmetry in visual antler characteristics (i.e., antler length and number of tines) among free-ranging male reindeer in rut. This indicates that individuals that develop large ornaments are better able to buer developmental stress than individuals that develop small ornaments. There is no relationship between asymmetry in antler length and asymmetry in jaw length, suggesting that symmetry in antlers does not re¯ect overall body symmetry. This dierence may be caused by trait-speci®c sensitivity to developmental stress. Such stress may partly be caused by parasites, which show a positive association with asymmetry in antlers, but no relationship to asymmetry in jaws. Our results indicate that antlers may be exposed to directional selection in a visual signaler-receiver system and that information about parasite burden may be obtained from evaluation of asymmetry in antlers developed under exposure to a multitude of environmental stresses. Key words Ornaments á Fluctuating asymmetry á Parasites á Body condition á Reindeer
Introduction Fluctuating asymmetry (FA) is de®ned as small random deviations from perfect bilateral symmetry, where the signed dierences between the two sides are normally distributed with a mean value of zero (Van Valen 1962). Fluctuating asymmetry results from an individual's inE. Markusson (&) á I. Folstad Department of Ecology/Zoology, IBG, University of Tromsù, 9037 Tromsù, Norway fax: +47 776 45600; e-mail:
[email protected]
ability to complete identical development of bilateral symmetrical traits, therefore representing an appealing measure of developmental instability (Palmer and Strobeck 1986; Parsons 1990). Dierent types of environmental stress (e.g., pollution and extreme temperatures) have been found to aect FA (reviewed in Parsons 1990). The sensitivity to developmental stress, and the knowledge of the ideal expression, perfect bilateral symmetry, makes FA a potential indicator of environmental stress experienced by organisms (cf. Leary and Allendorf 1989). In ornamental characters, the relationship between intensity of asymmetry and mean size of a trait is frequently negative, that is, large traits show low asymmetry. In non-ornamental traits showing asymmetry, on the other hand, the usual pattern between size and asymmetry is positive or U-shaped (Mùller and Pomiankowski 1993). The negative relationship between asymmetry and size in ornaments has been suggested to be consistent with the handicap principle (Zahavi 1975; Mùller 1990a; Mùller and HoÈglund 1991). Sexually selected condition-dependent ornaments are often subjected to directional selection for size. Directional selection is likely to result in maximization of trait expression and, assuming maintenance of trait cost, reduced developmental homeostasis of the trait. The latter may ensure that condition-dependent ornaments will be especially sensitive to developmental stress, relative to non-ornamental traits (Mùller and Pomiankowski 1993). If individuals of high quality pay a relatively lower cost for a given ornament size than individuals of low quality, as suggested by the handicap principle (Zahavi 1975), the former are likely to experience a relatively lower degree of developmental stress than the latter. Thus, individuals with large ornaments are likely to develop lower asymmetry in ornaments than individuals with small ornaments (Mùller 1990a). Parasites can be of particular importance in sexual selection (Hamilton and Zuk 1982; Mùller 1990b; Read 1990; Clayton 1991; Folstad and Karter 1992; Zuk 1992). If the full expression of sexually selected orna-
502
ments depends on the health and vigour of the individual, ornaments can signal resistance against parasites that aect host viability. Moreover, continuous co-evolution between host and parasites can in theory maintain variation in heritable resistance against parasites. Given heritability of parasite resistance, individuals choosing mates with elaborate ornaments will acquire superior resistance genes for their ospring, improving ospring ®tness and thereby their own reproductive success (Hamilton and Zuk 1982). Recent studies have found symmetry in ornaments to be related to female choice (e.g., Mùller 1992a; Thornhill 1992; Swaddle and Cuthill 1994) as well as dominance (Malyon and Healy 1994; Mùller et al. 1996). In male barn swallows (Hirundo rustica), asymmetry in length of the tail ornament was positively related to parasite intensity when parasite intensities were manipulated (Mùller 1992b), and females also preferred males with long symmetrical tail ornaments (Mùller 1992a). Additionally, a positive relationship between parasites and asymmetry has been observed in several species (reviewed in Mùller 1996). Thus, the results from both experimental and observational studies suggest that parasitic infections are major contributors to developmental instability causing FA in ornamental characters. Antlers in Cervidae have often been used to illustrate and test ideas concerning ornamental characters (e.g., Andersson 1982). For example, antihelminthic treatment of female reindeer (Rangifer tarandus) fed ad libitum under experimental conditions resulted in lower degree of asymmetry in length of growing antlers compared to controls. Furthermore, asymmetry in length of growing antlers was positively related to parasite burden in the untreated group (Folstad et al. 1996). However, it is still uncertain if asymmetry in full-grown reindeer antlers can reveal any information about the parasite burden of the bearer when antlers have grown under natural conditions, because other environmental (e.g., nutritional and social) stressors may mask the eects of parasitic infections. Therefore, other environmental stresses may be more important than parasites for the developmental stability of antlers grown under natural conditions, and parasite burden may thus not be revealed through the expression of hard antlers in freeranging reindeer. The above considerations led to an observational investigation of antler characteristics, parasite intensities and measures of body condition in sexually active, freeranging male reindeer that had developed their antlers under exposure to a multitude of environmental stressors. Four main questions were raised. First, is there a negative relationship between antler size and asymmetry in antlers, indicating that antlers are ornamental characters exposed to directional selection? Second, is there a relationship between antler symmetry and symmetry in non-ornamental traits indicative of overall body symmetry in reindeer? Third, is size and symmetry of reindeer antlers positively related to traits indicative of body
condition and, last, are antler asymmetry and asymmetry in traits indicative of overall body symmetry positively related to parasite burden in sexually active male reindeer?
Materials and methods The animals Antlers, common to most species of deer (family Cervidae), are bony outgrowths of the skull covered with velvet during the growth phase. Before the mating season the antler ossi®es and the velvet is cast. Both male and female reindeer develop antlers from early spring. Males lose their antlers after rut in late autumn to early winter, whereas females loose their antlers after calving in spring (Skjenneberg and Slagsvold 1968). We examined 59 semi-domestic male reindeer 1.5 years old, slaughtered within 3 h in Guovdageaidnu, Finnmark, northern Norway on 20 September 1993. All animals sampled had entered rut and had antlers without velvet. They originated from one herd in district 34, West Finnmark reindeer management area (Anonymous 1994). As approximately 90% of the 1.5-year-old males in this herd were slaughtered in 1993 (West Finnmark Reindeer Council, personal communication), the sampled individuals were considered representative of their cohort. Morphometric and parasitological data where sampled for all animals, but sample size vary as a consequence of damage to some of the traits examined (e.g., jaws were fractured) during slaughtering.
Morphometric data Body weights (i.e., dressed body weight) were recorded to the nearest 0.1 kg. We weighed kidneys and perirenal fat to the nearest 0.1 g and calculated a kidney fat index using the formula: (mean perirenal fat weight/mean kidney weight) ´ 100 (Finger et al. 1982). There was a positive relationship between dressed body weight and kidney weight (t 0:31; n 39; P 0:005; Kendall rank-order correlation), which is a prerequisite for using the kidney fat index (Adamczewski et al. 1987). The kidney fat index was used as an estimate of relative body fat deposits (Adamczewski et al. 1987). One of us (E.M.) measured jaw length, antler length, antler weight, antler volume and number of tines. Antler length includes the length of all parts of the antler (Fig. 1). The main beam was measured along the inner side from the burr to the outer tip. Tines were measured from the tip to their divergence from the main beam or other tines. All measurements of antler length were done with a tape measure to the nearest 1.0 mm. The number of tines was counted as all points on the main beam exceeding 2.0 cm, excluding the point of the main beam, frontal tines and brow tines (Fig. 1). After conducting length measurements, antlers were sawed o at the burr and weighed to the nearest 0.1 g. Additionally, relative antler weight was calculated using the formula: (antler weight/ dressed body weight) ´ 100. Volume was measured to the nearest 10.0 ml by cutting the antlers into approximately 10-cm pieces and putting them in a known amount of water, then measuring water displacement. The weight loss from cutting the antlers varied from 0.34% to 5.1% of total antler weight, with a mean of 1.4% (SD = 1.04). Length of left and right side of the lower jaw was measured according to von den Driesch (1976), to the nearest 0.1 mm with slide callipers, in order to calculate non-ornamental FA indicative of overall body symmetry. Relative FA (Palmer and Strobeck 1986) was calculated for each measure of antler size and the jaw length, using unsigned values of left-minus-right side of the character divided by mean size of the character. Individuals possessing antlers without tines on the main beam were not included in the analysis of FA in number of tines.
503 the non-zero samples of that parasite (cf. Karter 1993). When only presence of a parasite group could be established a score of 1 was given. The scoring for all parasite groups in each individual was then summed and the sum constitute the index value. The parasitological terms ``intensity'', ``mean intensity'' and ``density'' are used according to Margolis et al. (1982). Statistical analysis
Fig. 1 Antler nomenclature. All parts on the antler were included in antler length (i.e., length of the main beam from the burr to outer tip, length of all tines, length of frontal tine and length of brow tine including tines on frontal and brow tine)
Statistical analyses were done using Statsoft Statistica, Macintosh version 3.0, (Statsoft 1993). Associations were tested using Kendall rank-order correlations. Tests are two-tailed. A result was considered statistically signi®cant if P 0:05. The term abbreviations t, n and P refer to the value of Kendall's coecient of rank correlation, sample size and signi®cance level for the speci®c test, respectively. We did not adjust for multiple comparisons when testing for a priori predictions. According to Rothman (1990) this leads to fewer errors of interpretation when the data under evaluation are not random numbers, but actual observations on nature. In order to test for measurement errors, we measured antlers and jaws from 5 random selected individuals twice and conducted a two-way mixed model ANOVA (see Palmer 1994), with individual (random) and side (®xed) as independent variables and length as dependent variable. The analysis showed that measurement error is small relative to dierences between individuals in both antler length and jaw length (F4;10 340:71; P < 0:0001 and F4;10 188:71; P < 0:0001, respectively).
Parasitological data
Results
Adult abomasal nematode intensities were estimated from subsampling male and female nematodes in the abomasum (Bye 1987). Propagules from the following parasite groups were recorded in faeces: the nematodes Elaphostrongylus rangiferi, Dictocaulus spp., Nematodirus sp., Trichiuris sp. and Skrjabinema sp.; the cestode Moniezia sp. and Coccidia parasites. Density of E. rangiferi and Dictocaulus spp. larvae were estimated using the extraction method described by Halvorsen and Wissler (1983). Density of nematode eggs (Nematodirus sp., Trichiuris sp. and Skrjabinema sp.), cestode eggs (Moniezia sp.) and coccidia oocysts was estimated using the modi®ed McMaster ¯otation method (Whitlock 1948). Parasite larvae found in the ¯otation method were used to establish presence, not density, of a parasite group. A parasite index was constructed to quantify an individual's overall burden of the investigated parasites. Infections of adult abomasal nematodes were given ordinal scores from 1 to 4, where each number represents the corresponding quartile of the original distribution of parasite intensities. Parasitic eggs and larvae recorded in faeces were given a score of 0, 1 or 2 based on the faecal density of each parasite group (Karter 1993). Zero scores were given to individuals with faecal samples without transmission stages; a score of 1 was given to individuals with ``low'' density of transmission stages and 2 to individuals with ``high'' density of transmission stages. The cut-o between ``low'' and ``high'' scores for each parasite was determined by the median faecal count for all
Descriptive statistics of morphometric and parasitological data are given in Tables 1 and 2, respectively. Mean dressed body weight was 23.5 kg (SD 2:8; n 59) and mean kidney fat index was 15.4 (SD 8:2; n 39). None of the distributions of left-minus-right (L ÿ R) values diered signi®cantly from normality with a mean of zero (Kolmogorov-Smirno one-sample test). However, several distributions departed from normality with regard to skew and kurtosis. These statistics were sensitive to one extreme individual, but after removing the outlier, none of the distributions for L ÿ R diered signi®cantly from normality with a mean of zero for either skew or kurtosis (Table 1). Body weight was positively correlated with the kidney fat index (t 0:24; n 39; P 0:03), but was not correlated with the parasite index (t ÿ0:09; n 45; P 0:39). Furthermore, the parasite index was neither correlated with the kidney fat index (t ÿ0:04; n 35; P 0:76), nor with jaw length (t 0:03; n 38; P 0:81).
Table 1 Descriptive data for FA in 1.5-year-old male reindeer. Descriptive statistics for antler length, weight and volume are presented with one outlier removed. Individuals without tines on the main beam were not included in the analysis of number of tines Character
Antler length (cm) Antler weight (gram) Antler volume (10 ml) Number of tines Jaw length (mm)
n
53 50 50 48 44
(L + R)/2
L)R
Mean SE
Mean SE
Skew SE
78.1 163.3 11.5 1.8 223.2
1.96 )1.51 )0.12 )0.08 )0.17
0.0002 )0.25 0.33 )0.22 0.10
2.25 7.04 0.53 0.12 1.11
jL ) Rj
1.43 2.89 0.24 0.15 0.10
jL ) Rj/(L + R)/2
Kurtosis SE Mean SE
0.33 0.32 0.34 0.07 0.34 0.36 0.34 0.34 0.36 )0.42
0.65 0.66 0.66 0.67 0.70
8.05 16.21 1.28 0.75 0.54
0.94 1.75 0.16 0.10 0.06
Mean SE 0.11 0.11 0.11 0.58 0.002
0.013 0.011 0.013 0.10 0.0002
504 Table 2 Parasite prevalence (percentage of infected individuals) and mean intensity (mean number of parasites in infected individuals) with standard deviation (SD) of the parasite groups found in 1.5-year-old male reindeer. Abomasal nematodes are estimated from subsampling adults and are given as the estimated total number of adults in the abomasum. All other parasite intensities are given as density of parasitic propagules per gram faeces. Sample size is 45 reindeer for abomasal nematodes and 50 for all other parasite groups investigated. For two individuals presence, but not intensity, of Elaphostrongylus rangiferi could be established, and these individuals are not included in the calculation of mean intensity and SD of E. rangiferi Parasite group
Prevalence (%)
Mean intensity SD
Abomasal nematodes Trichiuris sp. Coccidia spp. Nematodirus sp. Dictocaulus spp. Elaphostrongylus rangiferi Moniezia spp. Skrjabinema sp.
100 86 60 22 12 10 10 6
4941.33 4.47 3.91 60.54 10.17 7.0 48.07 11.78
2395.61 6.53 6.47 85.23 10.53 10.39 60.24 19.53
There was a negative relationship between number of tines and asymmetry in number of tines (t ÿ0:3; n 48; P 0:0002) and antler length was also negatively correlated with asymmetry in antler length (t ÿ0:19; n 54; P 0:05). However, there was no signi®cant relationship between antler weight and asymmetry in antler weight (t ÿ0:006; n 51; P 0:95) or between antler volume and asymmetry in antler volume (t ÿ0:007; n 51; P 0:95). There was also no signi®cant relationship between jaw length and asymmetry in length of jaws (t ÿ0:17; n 44; P 0:1). Moreover, asymmetry in jaw length was not signi®cantly correlated with either asymmetry in antler length (t 0:08; n 44; P 0:43), asymmetry in antler weight (t 0:16; n 43; P 0:14), asymmetry in antler volume (t 0:14; n 43; P 0:18), or asymmetry in number of tines (t 0:15; n 40; P 0:18). Both variance and mean asymmetry in antler length were signi®cantly larger than the corresponding measures in jaws [F2;43 50:0; P < 0:01, F-max test, (Sokal and Rohlf 1973), U 104:0; P < 0:0001; n 54; Mann-Whitney U-test respectively]. Body weight was positively correlated with antler length, antler weight and antler volume, but not with Table 3 Kendall rank order correlations between measures of antler size, relative asymmetry (FA) in antlers, body weight and body fat deposits (kidney fat index)
Character
Antler length Antler weight Antler volume Number of tines FA in antler length FA in antler weight FA in antler volume FA in number of tines
Table 4 Kendall rank order correlations between relative asymmetry (FA) in antlers and jaws and the parasite index in 1.5-yearold male reindeer. The parasite index is computed by scoring each host's estimated intensity of adult abomasal nematodes and faecal density of parasite propagules Character
FA FA FA FA FA
in in in in in
Parasite index
antler length antler weight antler volume number of tines jaw length
t
n
P
0.26 0.22 0.27 0.08 0.11
40 39 39 36 38
0.018 0.05 0.014 n.s. n.s.
number of tines (Table 3). Additionally, relative antler weight was positively correlated with body weight (t 0:22; n 50; P 0:02). No measure of antler size was signi®cantly correlated with the kidney fat index (Table 3). The parasite index did not correlate with antler length (t 0:001; n 40; P 0:99), antler weight (t 0:00; n 39; P 0:99), antler volume (t ÿ0:007; n 39; P 0:95) or number of tines (t 0:08; n 40; P 0:45). No measures of asymmetry in antlers were signi®cantly correlated with body weight or kidney fat index (Table 3). Additionally, asymmetry in jaw length was not correlated with body weight (t ÿ0:15; n 44; P 0:16) or the kidney fat index (t ÿ0:02; n 28; P 0:91). The parasite index was positively correlated to asymmetry in antler length, antler weight and antler volume, but not to asymmetry in number of tines or asymmetry in jaw length (Table 4).
Discussion The documented negative correlation between size and asymmetry in both antler length and number of tines among 1.5-year-old male reindeer correspond with results reported from several species of birds (Mùller 1990a; Manning and Hartley 1991; Mùller and HoÈglund 1991), primates (Manning and Chamberlain 1993) and has also previously been found in Cervidae (Smith et al. 1982; however see Solberg and Sñther 1993). Intraspeci®c competition favouring individuals with large antlers
Body weight
Kidney fat index
t
n
P
t
n
P
0.24 0.44 0.42
