basal song rate variation in male red-winged blackbirds ...dshutler/1991behavecol.pdfmale's...

Basal song rate variation in malered-winged blackbirds:sound and fury signifying nothing?

Male red-winged blackbirds (Agelaius phoeniceus) vary substantially in how much they sing. Wetested whether song rate reflected the male's quality, the quality of his territory, or his frequencyof interaction with potential competitors. Song rate declined with the seasonal decline in nestinitiations, suggesting that song is used to attract females. However, males with higher songrates did not obtain larger harems. This suggests that females do not use male song rate toassess either male quality or territory quality. If song rate signals male quality to male rivals,and if territory owners are of higher quality than floaters, then owners' song rates should behigher than those of floaters. However, when we removed territory owners temporarily, thesong rate of their floater replacements was similar to that of the original owners. Finally, amale's song rate was not affected by how frequently he chased intruders, or by the proximityor number of neighbors he had. In short, the only information that basal song rate of red-winged blackbirds appears to provide to conspecifics is that the territory is occupied. In sim-ulated territory takeover attempts, however, song rate was significantly higher than the basalrate. Thus, more may be revealed by the singer in these situations. [Behav Ecol 1991;2:123-132]

Dave ShutlerPatrick J. WeatherheadDepartment of Biology,Carleton University,Ottawa, OntarioK1S 5B6, Canada

During the breeding season, male birds dis-play to attract mates or repulse rivals or

both. For some species, more vigorous dis-plays reap greater reproductive return (e.g.,Duncan and Bird, 1989; Pruett-Jones andPruett-Jones, 1990; Vehrencampetal., 1989).Song is a common component of passerinedisplay. Both mate attraction and rival repul-sion functions for passerine song have beenvividly illustrated when males have had theirsinging ability surgically muted (McDonald,1989; Peek, 1972; Smith, 1972, 1976, 1979).Muted males have difficulty acquiring andholding territories or mates. When the effectsof the surgical muting are temporary, malesare able to regain territories or mates oncetheir voices return. Song rate varies less innatural populations than in muting experi-ments. However, in natural populations,greater singing vigor has been associated withbetter parental care (Grieg-Smith, 1982),higher male quality (Reid, 1987), and morerapid mate attraction (Alatalo et al., 1990;Reid and Weatherhead, 1990). Despite theease of measuring song rate, it has receivedless attention than more elaborate aspects ofsong (Alatalo et al., 1990). In this paper, weexamine both potential causes and conse-quences of song rate variation in red-wingedblackbirds {Agelaius phoeniceus).

The red-winged blackbird is a commonNorth American passerine that breeds poly-gynously over much of its range (Eckert andWeatherhead, 1987b; Orians and Christman,1968). The species breeds in various habitatsranging from cattail (Typha sp.) marshes wherenest density is high, to upland fields wherenest density is low (Eckert and Weatherhead,1987b). Breeding sites are nonetheless limit-ed; some males defend no territories, somedefend territories that attract no females, andsome defend territories with multiple females(Orians, 1980). These facts, and the fact thatfemales breed exclusively with territory own-ers (Gibbs et al., 1990; Monnett et al., 1984;Searcy, 1979a), lead to intense competitionand selection for territory ownership. The red-winged blackbird territorial system is thuscharacterized by pronounced variation in malesuccess, territory quality, and intrusion rates.We focus on these three variables in testinghypotheses to explain variation in song rate.In essence, we are asking what informationconspecifics can obtain from song rates.

The first hypothesis we consider is the re-source-holding potential (RHP) hypothesis.RHP is usually viewed as the total of an in-dividual's fighting abilities, which often de-pend on size or strength. It has been assumedthat better fighters obtain better or more re-

Received 4 December 1990Revised 14 March 1991Accepted 23 March 1991

1045-2249/91/$2.00© 1991 International Societyfor Behavioral Ecology

Shutler and Weatherhead • Red-winged blackbird song rate 123

sources (Darwin, 1871; Parker, 1974). Thus,this hypothesis predicts that male red-wingedblackbirds with high RHP will control largerharems than individuals with low RHP. Non-territorial "floaters" that have been unable tocapture a territory will have the lowest RHP.If song rate is indicative of a male's RHP, thenmales with higher song rates should have larg-er harems than males with lower song rates.Furthermore, territory "owners" should singat higher rates than their replacements, i.e.,males that were previously floaters. Thus, ifRHP affects song rate, conspecifics could de-termine a territory occupant's quality by mon-itoring his song rate.

The second hypothesis we consider is theterritory-value hypothesis. According to thishypothesis, male song rate reflects the valueof the territory being defended and is inde-pendent of male quality. Radesater and Ja-kobssen (1989) found that, in their populationof male willow warblers (Phylloscopus trochilus),there was no competition for territory vacan-cies, and all replacing males sang at rates in-dicative of the territory's potential for attract-ing females. Hence, males on poor-qualityterritories sang at lower rates than males onhigh-quality territories, regardless of the males'status. If the territory-value hypothesis appliesto red-winged blackbirds, it predicts an asso-ciation between higher song rates and largerharems. However, in contrast to the RHP hy-pothesis, the territory-value hypothesis pre-dicts that the song rate of replacements shouldbe similar to that of owners. If the territory-value hypothesis holds, conspecifics that mon-itored song rate would obtain informationabout the quality of a territory rather thanabout the occupant. For willow warblers, songrate may be an honest signal of food avail-ability on the territory. However, in situationswhere males are able to obtain food from out-side their territories, song rate may no longerhonestly reflect the abundance of food on aterritory.

The final hypothesis we consider is the in-teraction hypothesis. In red-winged black-birds, neighbors and floaters frequently in-trude on territories (Metz and Weatherhead,1991; Peek, 1971). Because singing aids inrepulsing intruders (see above), more fre-quent singing may be used to counter morefrequent intrusions. If, on the other hand,singing functions to repel intruders beforethey actually intrude, then intrusions will beless frequent on territories defended by maleswith greater song rates. This hypothesis pro-poses that the frequency of interactions be-tween males determines song rate. If inter-action frequency is a function of male qualityor territory value (e.g., Arcese, 1987; McNair,1987; Norton et al., 1982), this hypothesis will

not be mutually exclusive of the previous twohypotheses.

We determined harem size and used severalindexes of the frequency of interactions be-tween competing males to test whether songrate revealed information about these factors.We used removal experiments to test whethersong rates of owner and replacement red-winged blackbirds were indicative of RHP,territory value, or both.

MATERIALS AND METHODS

Our study area was in southeastern Ontarionear the Queen's University Biological Station(45°37' N, 76°13' W). The territories in ourpopulation of red-winged blackbirds were ir-regularly spaced along the sides of approxi-mately 30 km of highway. We used most ofthese territories in both years of the study.

All analyses were performed on a micro-computer version of the Statistical Packagefor the Social Sciences (Nie, 1988). All meansreported are ± 1 SD.

Song rate

We observed territory occupants from insidea vehicle parked 20 to 30 m from the territory.We determined song rates (number of songsper minute) while the male occupant was vis-ible and within hearing range. Ninety-fourpercent of our observation periods (N=l 850)were 5 min. We only included observation pe-riods for which the male was visible and withinhearing range a minimum of 10% of the time(N = 1011). We also did not include assess-ments where males gave alarm calls in re-sponse to our presence; this generally did notoccur until females were feeding nestlings. Wevisited each territory repeatedly (mean = 6.5±3 .9 visits per year) to determine the occu-pant's song rate. In most cases, territories werevisited in systematic order (i.e., starting oneterritory beyond the one we had last observedin the previous visit to the field), but we mademore visits to territories where we had per-formed removals. Observations were madebetween 0500 and 1000 h and between 1800and 2030 h from early April to mid-July inboth years.

Breeding activity

We tested whether males with higher songrates attracted larger harems. Harem size wasdefined as the maximum number of simulta-neously "active" nests (having one or moreeggs or nestlings) on a territory. Red-wingedblackbirds have conspicuous nests, and we areconfident that we found all nests on the ter-ritories that we used in this analysis.

124 Behavioral Ecology Vol. 2 No. 2

We removed some males before their fe-males began nesting (see below). To determinewhether this affected the harem size on theterritory when the final occupant was present,we tested if a territory's harem size in yearswhen owners were removed was different fromyears when owners were not. We also per-formed removal experiments in 1988 and de-termined harem sizes on many of the sameterritories used in the present analysis. Thus,we included data from 1988 (and all possiblepairs of years) for this test. Harem size de-creased between years in 8 of 47 cases whereowners were removed, and in 13 of 45 caseswhere no removals were performed in eitheryear (x2 = 1.23, p = .27). Because this indi-cates that our removal experiments had noeffect on a territory's ultimate harem size, intesting the territory-value hypothesis all malestaken from a particular territory were as-signed the ultimate harem size that their ter-ritory supported.

In temperate-breeding birds, breeding op-portunities occur less often as the breedingseason progresses. If males defend territoriesto entice females to breed, a decline in femaleavailability will cause a concomitant decline interritory value. We estimated the seasonal de-cline in territory value for our population bycounting nest initiations (date when first eggwas laid) as the breeding season progressed.We searched territories for nests every 3-7days from the day the first nest was built untilthe last day we monitored song rates. We foundthe great majority of nests at the egg stage orearlier. We estimated the initiation date ofeach nest by backdating, based on an incu-bation period of 11 days and a nestling phaseof 10 days (Nero, 1956). With experience, wecould visually age nestlings to within a day.

Owners and floaters

We performed removal experiments to com-pare the song rates of owners and their floaterreplacements. We assessed a territory occu-pant's song rate before capturing and remov-ing him. Owners were captured and removedusing playback, stuffed male decoys, and traps(Smith, 1972) or mist nets. Owners werehoused in aviaries until we completed removalexperiments on their territory. Owners wereusually replaced within a day. We then mea-sured the song rates of replacements. On someterritories, we removed the replacement andrepeated the procedure. We continued to re-move birds until no replacements occurred oruntil a replacement was too unresponsive tobe captured. The most birds removed fromany one territory was seven. Removals wereperformed until 13 and 25 June in 1989 and1990, respectively.

We termed the first replacement a "shallowfloater." Males that replaced shallow floatersand all subsequent replacements were re-ferred to as "deep floaters." On average, in1989, we obtained song rates of individualowners 4.8 times, shallow floaters 3.4 times,and deep floaters 2.7 times. In 1990, the samevalues were, respectively, 3.6, 3.6, and 5.4times.

We had to be certain that replacements hadpreviously been floaters, since we would oth-erwise have compared song rates of ownersto owners. Prior to removing a territory own-er, we mapped the territory boundaries ofneighbors, and, in some cases, we also bandedneighbors. After removing a territory occu-pant, we ascertained whether the replacementwas banded or whether he also occupied aneighboring site simultaneously. The spacingof many of our territories (see below) alsoreduced the chances that neighbors would re-place the males we removed. If a neighbor didclaim a vacancy, we excluded that male fromour analyses.

Male-male interactions

If songs are used to repulse rivals, song ratesmay be higher where chases occur more fre-quently. While monitoring song rates, we re-corded all instances where owners chasedneighbors or floaters. From these data, wecomputed chases per hour for each territory.

Males from neighboring territories may dis-play to each other without actually becominginvolved in chases. We thus tested whethersong rates were affected by the distance be-tween territories or by the number of neigh-bors. Most of our territories were separatedfrom neighbors by areas of habitat unsuitablefor breeding (i.e., plowed fields or forest). Each"isolated" territory was occupied by one male,and each was centered on patches of cattailapproximately 25 to 250 m2 in size. Someisolated territories were close enough thatneighbors came in contact with each other.However, it was rare that a male was able todefend the perimeter of two isolated territo-ries. "Contiguous" territories were on cattailmarshes that were subdivided into 10 or moreterritories. Contiguous territories ranged insize from 200 to 1000 m2. Vacancies createdby removal experiments on contiguous terri-tories were more likely to be annexed byneighbors than was the case on isolated ter-ritories. As a further index of the frequencyof interactions between males, we counted thenumber of neighbors surrounding each ter-ritory. Here again, we compared the song ratesof males according to the number of neigh-bors they had.

Finally, we tested the effects of simulated


Figure 1Song rates for owners relativeto the hour of the day in1989 (top) and 1990(bottom). Sample sizes aregiven above the vertical linesthat indicate standarddeviations.

o> 2co

oV)

oeoo 0700 oaoo oaooTime of day

1800 1M0

dom sing when they arrive in a territory (Or-ians and Christman, 1968; Reskaft and Roh-wer, 1987). A singing male in a territory iseither an owner or a male attempting to takeover. The aggressive reactions of owners totakeovers is presumably what we exploit withour capture methods. To test whether songrates increased during takeovers, we com-pared an occupant's song rate from 5 minbefore a playback experiment with the songrate during 5 min of playback. Playback ex-periments involved placing a stuffed, male red-winged blackbird in the center of a territory.A cassette player suspended below the decoyplayed territorial song at a rate of 3.5 songsper minute. Songs had been taped from asingle male that occupied a territory on a marshoutside of the study area. Song rates of ter-ritory occupants were monitored for the first5 min after the male detected the decoy (i.e.,once the occupant oriented or moved towardthe source of the playback).

RESULTS

Between years, changes occurred in many ofthe individual males sampled, harem sizes onspecific territories, quality of individual ter-ritories, pairing dates, etc. Because suchchanges make these data partially indepen-dent, we treat each year separately in our tests.

We could not simultaneously test the rela-tionship between all of our variables and songrate because the resulting design had samplesizes ranging from 1 to 15 in individual cells,and this weakens the multiway analysis of vari-ance (Tabachnick and Fidell, 1989). Instead,we ran separate tests on the importance ofeach variable. In many cases, we subdividedour data for additional tests. This approachrequired that we divide our experimentwiseerror rate by the number of tests. Thus, whenwe tested for a relationship between song rate

and a variable in eight nonindependent tests,we could only reject the null hypothesis of nodifferences between groups when p < .006.

Song rate

For various reasons, we had more samplingperiods for some days than for others. Toweight each day equally and to remove within-day variation, we computed an average songrate for each day. Based on owners only, dailyaverage song rate declined as the breedingseason progressed (for 1989, r = —.34, p =.005, N = 66 days; for 1990, r = - .48 , p <.001, N= 60 days). Our removal experimentswere done in a way that required us to obtainmore owner song rates early in the season andmore floater replacement song rates late inthe season. In comparing song rates, we con-trolled for potential seasonal biases by com-puting a regression of owner song rate againstJulian date. We then adjusted song rates ofboth owners and replacements to reflect themiddle of the season (15 May or Julian date135) by using the slope (m) from the ownerregression as follows:

adjusted rate = actual rate — m(day — 135).

After adjusting for seasonal effects, wefound that song rate still varied among ownersfrom hour to hour (Figure 1; ANOVA for1989,7^ =4.94, p< .001; for 1990, F= 3.34,p = .007). To avoid any potential biases fromtime of day effects, we readjusted song ratesof both owners and replacements as follows: •

readjusted rate =mean rate for hour

mean rate for all hours

x adjusted rate.

Finally, we determined whether a bird's in-creased energy requirements in cold weatherresulted in a concomitant decrease in songrate (e.g., Reid and Weatherhead, 1990). Wefound that, after having controlled for seasonand time of day, song rate was not affectedby temperature (for 1989, r = -.06,p = .26,N = 123 hourly temperature averages; for1990,r=.02,/>=.41,./V= 105). Except whereindicated, we only used song rates adjustedfor season and time of day.

The average song rate of individual malesstill varied after these adjustments (Figure 2).Some males consistently sang at low rates (e.g.,one owner had a mean rate of 1.3 songs perminute, range 0.1-3.9, N= 11), whereas oth-ers consistently sang at high rates (e.g., an-other owner had a mean rate of 3.6 songs perminute, range 1.8-5.5, N = 9). We testedwhether male quality, territory quality, or the


frequency of male-male interactions couldexplain the remaining variation.

Breeding activity

We used average song rates for males that hadbeen recorded more than once. We first testedwhether males with higher song rates attract-ed larger harems. Harem sizes ranged from 0to 5 in 1989 (N =* 76 territories, mean 1.3 ±0.9 females) and from 0 to 4 in 1990 (N -63, mean 1.6 ± 1.0). In neither year was songrate correlated with harem size (Figure 3), Theresults were similar when we used owners only(Figure 3). Because most females arrived a fewweeks before nesting began, we repeated theanalyses to determine whether females usedmale song rate as a cue for choosing nest sites.To make this determination, we consideredsong rate before the date the population's firstegg was laid. Again, we found no relationshipbetween male song rate and ultimate haremsize in either year, whether we tested using allmales or owners only (Figure 3).

Early nesters have greater reproductive suc-cess in many bird species (reviewed in Priceet al., 1988). Thus, we tested whether femaleslaid eggs earlier on territories where males(owners only) sang at higher rates. We foundno support for the prediction; the date thefirst egg was laid on a territory was not relatedto owner song rate in 1989 (r = — .12, p —.16, iV= 67 males) or 1990 (r - .ll,/> = .20,N = 56). The relationship was also not signif-icant using only song rates from the periodbefore egg laying (for 1989, r = .02, p = .45,N = 47; for 1990, r = - . 13 , p = .23, N =38).

A male red-winged blackbird in a low-den-sity population can be more certain of his pa-ternity in the nests on his territory (Gibbs etal., 1990). We tested whether the relationshipbetween breeding activity and song rate wasmore reliable among males from territories inlow-density habitats that had no neighbors.Again, all results were nonsignificant (data notshown). Thus, extrapair fertilizations wereprobably not confounding the relationshipbetween nesting activity and male song rate.

Nesting in both years peaked between 15and 20 May (Figure 4). Song rate (withoutmaking adjustments) was negatively correlat-ed with cumulative nest initiations in 1989(Figure 4; r = - .52, p < .001) and 1990 (r =— .38, p = .004). This suggests that males sangat a rate that was related to the likelihood thatthey would attract an additional female.

Owners and floaters

In 1989, we used 20 territories for removalexperiments from which we obtained song rate

IOO-I

TJ 60-C

O

<D

E3 20

0.S 1.5 2.5 3.5 4.5 5.5

Songs per min (adjusted)

data for both owners and floater replace-ments. We also had two territories where theowner disappeared without our intervention.In 1990, the equivalent values were 27 re-moval experiments and two territories wherethe owners disappeared naturally. In total, weobtained song rates for 142 owners (includingowners from territories where no removal ex-periments were performed), 47 shallow float-ers, and 42 deep floaters.

To avoid pseudoreplication, we averagedsong rates for individuals we had observedmore than once. Comparisons of song ratesof owners and floater replacements revealedno significant differences (Figure 5; ANOVAfor 1989, F = 0.13, p = .88; for 1990, F =

Entire season

1989 a-

Figure 2Variation in average songrates* fer individual males (TV= 249) after cprrecting forseasonal and i Meffects.

Figure 3Song rate of males sampledon a territory combined overthe whole season and beforeegg laying only. Song rates ofdifferent males from the sameterritories are treated asseparate values. Blank barsare data for all males; solidbars are data for owners only.No significant relationshipsbetween harem size and songrate were detected (eightseparate ANOVAs, all p values

Before egg-laying

25 17

1Harem size

2+ 0 1 2-f

Harem size


1989 1990

Figure 4Daily average song rates(owners only) versus nestinitiations for the wholepopulation (N = 125 nests for1989, N= 141 for 1990).Intervals are 5 days.

n nllfL „

Figure 5Song rates of owners, shallowfloaters, and deep floaters.See text for explanation ofdesignations.

0.15, p = .86). We then compared the songrates of owners and shallow floater replace-ments on a territory-by-territory basis. In the2 years, 21 shallow replacements had songrates greater than the owners they replaced,and 18 had song rates less than the ownersthey replaced (paired t test, t = —1.33, p =.19). Similarly, 11 deep replacements sang athigher rates than owners, and 8 sang at lowerrates (t = -1.04, p = .31). Thus, in neitherof our tests did song rate betray any differencein the RHP of owners and floaters.

Male floater red-winged blackbirds com-pete most intensely shortly after vacancies ap-pear (Peek, 1971). We tested whether songrate declined with the duration of a floaterreplacement's tenure. In both years, floatersong rate did decline the longer the floaterhad been on the territory (for 1989, r— —.12,p = .08, N = 146 individual song rates; for1990, r = - .27 , p < .001, N = 179). This

1989

1990

cE ji_s>Q. 2

W

to

4

n

§

2325

64 24 17

Ownars Shallowfloatcra

suggests that, if anything, song rate is nega-tively correlated with male experience. It alsosuggests that, if contests for territory own-ership are most intense shortly after vacanciesappear (Peek, 1971), song rate should be re-lated to the frequency of male-male interac-tions.

Male-male interactions

We first tested whether song rates were higherduring chases. We did not include data fromterritories adjoining forested areas becausewe could not reliably monitor chases in thesesituations. Chases were observed in 60 of 534observation periods in 1989, and in 91 of 453observation periods in 1990. Song rates dur-ing observation periods in which chases oc-curred were not significantly different fromobservation periods in which they did not oc-cur (Table 1). These findings were the samewhen we considered owners only (Table 1).We then compared song rates on territorieswhere we observed at least one chase duringthe breeding season with song rates on ter-ritories where chases were never observed. Tocontrol for differences in the number of ob-servation periods on different territories, weaveraged song rate for each territory. Wefound no difference in song rate on territorieswhere chases occurred versus where they didnot occur, whether we considered all males orowners only (Table 2). Thus, we have no evi-dence that the frequency of chases affects songrate, or vice versa.

We then examined the effects of proximityto neighbors. We observed 58 isolated and 27contiguous territories in 1989 and 49 isolatedand 21 contiguous territories in 1990. Weagain used mean song rate observations forindividual males for this test. In neither yeardid song rate differ between isolated and con-tiguous habitats, whether we included all malesor just owners (Table 3).

As an additional index of the potential forinteraction with other males, we counted thenumber of territories that shared boundarieswith each territory and tested whether thisaffected song rate. In neither year was songrate affected by the number of neighbors amale had (Figure 6; ANOVA for 1989, F =1.68, p = .19; for 1990, F = 1.28, p = .28).This result was unchanged if owners alonewere considered.

Finally, we considered the reaction of own-ers to simulated takeovers. When we placedstuffed decoys on territories and played re-cordings of territorial song, the song rates ofterritory occupants increased by 20% (unad-justed song rates, data pooled for 1989 and1990; before playback, mean songs per min= 2.96 ± 1.7; during playback, mean = 3.53


± 1.6, paired t test, t = -3 .31, p < .001).Together with the other results for male-maleinteractions, these data indicate that male songrate does not change in response to passiveinteractions, but increases dramatically whenan interaction becomes more threatening.

DISCUSSION

We found no evidence that song rate was anindex of RHP. Owners with large harems sangat rates similar to males with small harems,and owners sang at rates similar to floaterreplacements. Several other studies have alsofailed to find evidence that measures of RHPrelate to reproductive success for adult malered-winged blackbirds (Beletsky and Orians,1989; Eckert and Weatherhead, 1987a,b,d;Shutler and Weatherhead, 1991). The dif-ferences between successful and unsuccessfulterritory claimants in this system remain elu-sive, and chance may play an important role(Eckert and Weatherhead, 1987b).

There was little evidence that song rate wasrelated to territory value. Males did sing lessas the seasonal reproductive value of the ter-ritory declined, as has been shown for severalother avian species (reviewed in Welty, 1982).Because floater replacements sang at ratessimilar to owners, all males could have ad-justed their song rate to reflect the value ofthe territory they had claimed. However, thecrucial evidence needed to support the ter-ritory-value hypothesis was a relationship be-tween a male's harem size (which was our mea-sure of territory value) and his song rate.Because we did not find this relationship, wecannot conclude that the relative value of amale red-winged blackbird's territory influ-ences his song rate.

The absence of a relationship between songrate and territory value in our population con-trasts with evidence from food supplementa-tion experiments. With additional food, malesof many species, including red-winged black-birds, increase their song rate (e.g., Gottlan-

1989

1990

s

E *

s , 2

I '0

5

!» 3Q.

CO 2o>cO 1

2117 22

2116

14 10

0 1 2 3+

Number of neighbors

der, 1987; Searcy, 1979b), suggesting that songrate is an honest signal of territory quality.Females may use this signal because red-wingedblackbird territories with additional food at-tract larger harems (Ewald and Rohwer, 1982),and females on these territories breed earlier(Wimberger, 1988).

However, food supplementation may causegreater variation in song rate than occurs innatural situations, which may be a partial ex-planation for our negative results. Yasukawaet al. (1987) also found no relationship be-tween male display intensity and harem sizein experiments where food was not supple-mented. An additional factor that would ex-plain why we found no relationship betweensong rate and harem size is that males in ourpopulation frequently forage off the territory,which would decrease any correlation be-tween food abundance on the territory andsong rate (Ydenberg, 1984). Thus, our resultsand those of Yasukawa et al. (1987) suggestthat food availability in natural situations doesnot vary enough to affect song rate in a waythat affects female settlement patterns.

Figure 6Owner song rates relative tothe number of neighborsassociated with the territory.

Table 1Effect of male-male chases on song rates

Song rate for

1989All malesOwners only


Observation periodswith chases

2.5 ± 1.6(474)2.4+ 1.6(318)

2.6 ± 1.5 (362)2.6 ± 1.4(186)

Observation periodswithout chases

2.6 ± 1.3(60)2.6 ± 1.5(39)

2.4 ± 1.4(91)2.4 ± 1.2(44)

t

-0.64-0.76

0.940.85

P

.53

.45

.35

.40

The data were compared using / tests. Means + SD of song rates are indicated. The numbers of observation periodsare given in parentheses.


Table 2Effects on song rates of defending territories that were susceptible to chases

Song rate for



Territories wherechases were seen

2.4 ± 1.2 (73)2.3 + 1.3(48)

2.6 ± 1.1 (59)2.6 ± 1.1 (34)

Territories wherechases were not seen

2.6 ± 0.8 (40)2.6 ± 0.8 (26)

2.6 ± 0.8 (52)2.8 ± 0.8 (29)

t

-1.10-1.02

-0.11-0.67

P

.28

.24

.91

.50

Data were compared using t tests. Means ± SD of song rates are indicated. The numbers of territories associated witheach category are given in parentheses.

We also found little support for the hy-pothesis that the frequency of male-male in-teractions was related to song rate. If greaterdisplay rates signal better quality resources,floaters may preferentially intrude on terri-tories where display rates are high becauseusurpation of these territories will producegreater payoffs for the victor (Tamm, 1985).On the other hand, males that display moremay be more formidable foes and worth avoid-ing. Unless these factors balance each other,a relationship between intrusion rate and dis-play rate would result. We found no such re-lationship, indicating that song rate conveyedno useful information to floaters. Song ratewas also not affected by the likelihood of in-teraction with neighbors. However, intensesimulated male—male interactions did causean increase in song rate. During simulatedintrusion, territorial males infected with bloodparasites are less aggressive than unparasi-tized males (Weatherhead, 1990). Althoughsong rate may be important in this latter con-text, we cannot evaluate how important in-creased song rate would be to preventing ac-tual takeovers.

We have thus far ignored the possibility thatother aspects of red-winged blackbird songare better correlated with the variables wemeasured. The only other aspect of red-wingedblackbird song that has received any attentionis repertoire size. Repertoire size has been

shown to be correlated with male age, breed-ing experience, and harem size (Yasukawa etal., 1980; but see Smith and Reid, 1979). Rep-ertoires may be more important in femalechoice than in male-male interactions be-cause males present their repertoire morerapidly in the presence of females than in thepresence of males (Searcy and Yasukawa, 1990;Smith and Reid, 1979). However, repertoiresare more effective at repelling male intrudersthan is repetition of a single song (Yasukawa,1981). Whether repertoires serve a female at-traction or a rival repulsion function, thesedata are consistent with an RHP function forrepertoire size. However, we cannot separateout the effects of territory quality because wedo not know whether males use larger rep-ertoires to signal the quality of their territoriesindependent of their age or experience.Nonetheless, it appears that red-winged black-bird repertoires convey more information thansong rate in most situations.

An inconsistency between the repertoiredata and our data is Yasukawa's (1981) findingthat males with larger repertoires [and hencemore experience (Yasukawa et al., 1980)] singat faster rates (but see Smith and Reid, 1979).Our data suggest that song rate was unrelatedto experience, and thus faster song rates wereprobably not associated with larger reper-toires. However, our song rate data for in-experienced males were obtained from males

Table 3The effects of neighbor proximity on song rates

Song rate for Isolated Contiguous



2.5 ± 1.1(85)2.3 ± 1.1 (53)

2.7 ± 1.0(88)2.8 ± 1.0(46)

2.5 ± 1.1 (33)2.6 ± 1.2(25)

2.4 ± 0.9 (24)2.4 ± 0.8 (18)

-0.250.93

1.041.53

.80

.36

.30

.13

Data were compared using I tests. Means ± SD of song rates are indicated. The numbers of males sampled are givenin parentheses.


that had just taken over territories, whereasYasukawa's (1981) inexperienced males hadbeen in his study area for 1 year or more.Because we found that recent replacementselevate their song rates for a brief period dur-ing territory establishment (see also Peek,1971), we have overestimated their long-termsong rates. The important issue, however, iswhether song rate is an honest signal. Becauserecent replacements sang at the same rate asmore established owners, this argues againstthe possibility that higher song rate is an hon-est signal of male quality. In fact, just theopposite is suggested, because floater replace-ment song rate declined with time of tenure.

In general, despite much variation, the bas-al song rate of red-winged blackbirds appearsto indicate only that a territory is occupied.We conclude that prospecting females thatmonitored a male's song rate would gain littleinformation about the quality of the male orthe quality of his territory. Similarly, compet-ing males would gain little information frommonitoring other males' song rates. Otherstudies have also found that red-winged black-bird display rates are weak predictors of nestdefense or provisioning rate (Eckert andWeatherhead, 1987c; Yasukawa et al., 1987).If there is information to be gained from dis-plays, it probably comes either during terri-tory takeovers (Weatherhead, 1990) or fromsources other than song rate. Alternatively,the variation among males in song rate is in-significant in most situations (Rohwer, 1982)and may be worth evaluating only when foodresources limit males' abilities to display (Ra-desater and Jakobsson, 1989; Reid andWeatherhead, 1990).

Drew Hoysak made important contributions to both fieldwork and data analyses. Kit Muma provided valuable fieldassistance. The Natural Sciences and Engineering Re-search Council of Canada has provided both of us withfinancial support. We have also benefited from grantsfrom the American Museum of Natural History's FrankM. Chapman Memorial Fund. Statistical advice was pro-vided by Sanping Chen of the Carleton University Statis-tical Consulting Centre. The comments of Val Nolan, Jr.,Bill Searcy, Donald Kramer, and an anonymous reviewerimproved the quality of the manuscript.

REFERENCES

Alatalo RW, Glynn C, Lundberg A, 1990. Singing rateand female attraction in the pied flycatcher: an exper-iment. Anim Behav 39:601-603.

Arcese P, 1987. Age, intrusion pressure and defenceagainst floaters by territorial male song sparrows. AnimBehav 35:773-784.

Beletsky LD, Orians GH, 1989. Territoriality among red-winged blackbirds: III. Testing hypotheses of territorialdominance. Behav Ecol Sociobiol 24:333-339.

Darwin C, 1871. The descent of man in relation to sex.New York: Random House.

Duncan JR, Bird DM, 1989. The influence of relatedness

and display effort on the mate choice of captive femaleAmerican kestrels. Anim Behav 37:112—117.

Eckert CG, Weatherhead PJ, 1987a. Competition for ter-ritories in red-winged blackbirds: is resource-holdingpotential realized? Behav Ecol Sociobiol 20:369-375.

Eckert CG, Weatherhead PJ, 1987b. Ideal dominancedistributions: a test using red-winged blackbirds (Age-laius phoeniceus). Behav Ecol Sociobiol 20:43-52.

Eckert CG, Weatherhead PJ, 1987c. Male characteristics,parental quality and the study of mate choice in thered-winged blackbird (Agelaius phoeniceus). Behav EcolSociobiol 20:35-42.

Eckert CG, Weatherhead PJ, 1987d. Owners, floaters andcompetitive asymmetries among territorial red-wingedblackbirds. Anim Behav 35:1317-1323.

Ewald PW, Rohwer S, 1982. Effects of supplemental feed-ing on timing of breeding, clutch size and polygyny inred-winged blackbirds, Agelaius phoeniceus. J Anim Ecol51:421-450.

Gibbs HL, Weatherhead PJ, Boag PT, White BN, TabakLM, Hoysak DJ, 1990. Realized reproductive successof polygynous red-winged blackbirds revealed by hy-pervariable DNA markers. Science 250:1394-1397.

Gottlander K, 1987. Variation in the song rate of themale pied flycatcher (Ficedula hypoleuca): causes andconsequences. Anim Behav 35:1037-1043.

Grieg-Smith PW, 1982. Song-rates and parental care byindividual male stonechats (Saxicola torquata). Anim Be-hav 30:245-252.

McDonald MV, 1989. Function of song in Scott's seasidesparrow, Ammodramus maritimus peninsulae. Anim Be-hav 38:468-485.

McNairJN, 1987. The effect of variability on the optimalsize of a feeding territory. Am Zool 27:249-258.

Metz KJ, Weatherhead PJ, 1991. Color bands functionas secondary sexual traits in male red-winged black-birds. Behav Ecol Sociobiol 28:23-28.

Monnett C, Rotterman LM, Worlein C, Halupka K, 1984.Copulation patterns of red-winged blackbirds (Agelaiusphoeniceus). Am Nat 124:757-764.

Nero RW, 1956. A behavior study of the red-wingedblackbird: I. Mating and nesting activities. Wilson Bull68:5-37.

Nie NH, 1988. SPSS statistical package for the socialsciences. New York: McGraw-Hill.

Norton ME, Arcese P, Ewald PW, 1982. Effect of intru-sion pressure in territory size in black-chinned hum-mingbirds (Archilochus alexandri). Auk 99:761-764.

Orians GH, 1980. Some adaptations of marsh-nestingblackbirds. Princeton, New Jersey: Princeton Univer-sity Press.

Orians GH, Christman GM, 1968. A comparative studyof the behavior of red-winged, tricolored, and yellow-headed blackbirds. Univ Calif Publ Zool 84:1-81.

Parker GA, 1974. Assessment strategy and the evolutionof fighting behaviour. J Theor Biol 47:223-243.

Peek FW, 1971. Seasonal change in the breeding behaviorof the male red-winged blackbird. Wilson Bull 83:383-395.

Peek FW, 1972. An experimental study of the territorialfunction of vocal and visual display in the male red-winged blackbird (Agelaiusphoeniceus). Anim Behav 20:112-118.

Price T, Kirkpatrick M, Arnold SJ, 1988. Directional se-lection and the evolution of breeding date in birds.Science 240:798-799.

Pruett-Jones SG, Pruett-Jones MA, 1990. Sexual selectionthrough female choice in Lawes' parotia, a lek-matingbird of paradise. Evolution 44:486-501.

Radesater T, Jakobssen S, 1989. Song rate correlationsof replacement willow warblers, Phylloscopus trochilus.Ornis Scand 20:71-73.

Reid ML, 1987. Costliness and reliability in the singingvigour of Ipswich sparrows. Anim Behav 35:1735-1743.

Reid ML, Weatherhead PJ, 1990. Mate choice criteria of


basal song rate variation in male red-winged blackbirds ...dshutler/1991behavecol.pdfmale's...

Documents