continue in this tradition by building on existing stud-
ies or moving to new taxa, as well as providing the
communitywide data we currently lack.Acknowledgments We thank the people of Monte-
verde, the Monteverde Cloud Forest Preserve, and
the MCL for supporting ornithological research on
private land. We are grateful to the U.S. National Sci-
ence Foundation and the Organization for Tropical
Studies for financial support. We are indebted to Dou-
glas J. Levey, Bette A. Loiselle, Nalini M. Nadkami, F.
Gary Stiles, Harry M, Tiebout III, and Dan Wenny for
providing many useful suggestions for the manu-
script. Ana Beatriz Azofeifa of the Organization for
Tropical Studies kindly performed bibliographic
searches. The Costa Rican Servicio de Vida Silvestre
has been courteous and efficient in issuing the nec-
essary permits.COOPERATION BETWEEN MALE LONG-TAILED MANAKINS
David B. McDonaldhe courtship displays of Long-tailed Manakins
are elaborate. Their most surprising feature is
that they entail coordinated display by a pair
of males (Fig. 6,1). The courtship song, which sounds
like the word "toledo," is sung in unison and has the
resonant quality of a pair of musical instruments
(Trainer and McDonald 1993; see Trainer, "Roles of
Long-Tailed Manakin Vocalizations," pp. 215-216),
If the singers are successful in attracting a female,
they move to a low dance perch and begin a dual-
male leapfrog dance. One male sidles toward the fe-
male, hovering upward as he nears her. Meanwhile,
his partner has sidled forward, so that the first male
lands behind his partner. The resulting sets of dance
hops may continue for many minutes, interspersed
with dual-male "butterfly flight," in which the part-
ners flutter, using a labored flight within a hemi-
sphere approximately 30 m in diameter, centered on
the dance perch. Eventually, if the female remains,
one male drops out of the display, the other contin-
ues solo butterfly flight, and if the female is recep-
tive, a copulation occurs on the perch (McDonald
1989b). Males form stable partnerships, with a domi-
nant (alpha) and subordinate (beta) male, and a vari-
able number of lower-ranking males (Foster 1977,
1981, McDonald 1989a, 1993a,b). The partnerships
persist for several to many years, but during its course,
only the alpha male mates. Why, then, does the
nonmating beta partner engage in months or years of
vigorous courtship display for no apparent reward?
One possible explanation is that the partners are
close relatives. In that case, the beta male would bene-
fit through the copies of his own genes passed through
the relative he helped, via kin selection (Hamilton
1964). We tested the kin selection hypothesis in Long-
tailed Manakins using a technique called microsatel-
lite DNA (McDonald and Potts 1994). Microsatellites
(Queller et al. 1993) are noncoding regions of DNA
containing short unit repeats of some combination of
the four nucleotides—A, G, C, or T. For example, the
microsatellite might consist of CA dinucleotides re-
peated 10—20 times. Within a population, individu-
als can vary widely in the number of dinucleotide
repeats. One male might inherit a microsatellite with
12 repeats from his father and 14 repeats from his
mother; another individual might have 15 repeats on
one chromosome and 14 on the other. On each side
of the microsatellite is an area of nonrepeating DNA,
whose sequence is invariant across the population.
These "flanking regions" provide the information for
building primers to amplify the microsatellites of
sampled individuals by the polymerase chain reac-
tion (PCR). The amplified product can be visualized
by ethidium bromide staining. The variants then ap-
pear as bands on the gel that can be sized by compari-
son to a known standard. By developing a sufficient
number of highly variable microsatellites, we could
address the question of the degree of relatedness
among partnered males.
We found that the relatedness of 33 pairs of coop-
erating males was no greater than that between any two
males taken at random from the population. The mean
relatedness for the 33 pairs was -0.14 (±0.10), mean-
ing that they were slightly less similar than randomly
chosen males. Further, of the 33 pairs, 17 were nega-
tively related, while 16 were positively related. This
is what one would expect if partnerships developed
without regard to relatedness—some males will pair
by chance with males to whom they are dissimilar
genetically (negative relatedness) and others with
males whom they resemble genetically (positive relat-
edness). The nonmating male thus does not benefit
from the indirect component of inclusive fitness. We
therefore looked for direct benefits to the beta male.204 BirdsT