doi:10.1111/j.1420-9101.2006.01165.

Sperm competition in a fish with external fertilization: the

contribution of sperm number, speed and length
J. A. STOLTZ & B. D. NEFF
Department of Biology, University of Western Ontario, London, Ontario, Canada

Keywords:

Abstract

bluegill;
fish;
mechanism;
raffle;
speed;
sperm competition;
sperm number.

The role of sperm number and quality in male competitiveness was

investigated using in vitro fertilization experiments with bluegill (Lepomis
macrochirus). Bluegill males use one of three mating tactics: sneakers, which
streak spawn; satellites, which mimic females; and parentals, which are
territorial. The in vitro experiments mimicked natural spawning by incorporating these males mean proximity to eggs and timing of sperm release. Using a
maximum-likelihood algorithm, raffle equations were fit to paternity data,
which revealed a strong effect of sperm number on male competitiveness.
There was no difference in sperm flagellum length, curvilinear swim speed or
path linearity among the three male mating types, and these traits did not
explain any additional variation in male competitiveness. It was estimated
that, given closer proximity to eggs, satellites need release only 0.34 times as
many sperm as parentals to obtain equal paternity. Despite being farther from
the eggs and releasing sperm about half a second after parentals, sneakers need
only release 0.58 times as many sperm as parentals to obtain equal paternity.
Thus, the increased competitiveness of sneakers sperm must come from a
component of sperm quality other than speed or length.

Introduction
Sperm competition occurs when the sperm from two or
more males compete for the fertilization of a given set of
eggs (Hildemann & Wagner, 1954; Parker, 1970). Sperm
competition is widespread, and occurs in a range of
animal taxa including amphibians, birds, fishes, insects
and mammals (Birkhead & Mller, 1998). Sperm
competition can lead to a wide range of behavioural,
morphological and physiological adaptations that
enhance the success of a males sperm relative to a rival
(Beatty et al., 1969; Parker, 1970; Smith, 1984). It is now
well understood that males typically differ in their
competitiveness during fertilization (Gage et al., 1995;
Arnqvist & Danielsson, 1999). However, the mechanisms
that males use during sperm competition to increase their
Correspondence: B. D. Neff, Department of Biology, University of Western
Ontario, London, Ontario N6A 5B7, Canada.
Tel.: 519 850 2532; fax: 519 661 3935;
e-mail: [email protected]
Present address: J. A. Stoltz, Department of Life Sciences, University of
Toronto at Scarborough, Scarborough, Ontario M1C 1A4, Canada
Both authors contributed equally to this work.

probability of fertilization are just beginning to be

understood (e.g. Birkhead et al., 1999; Evans et al.,
2003; Gage & Morrow, 2003; Garca-Gonzalez &
Simmons, 2005).
In externally fertilizing fishes, sperm competition
ranges from near complete mate monopolization to large
breeding assemblages (Stockley et al., 1997). Taborsky
(1998) postulated that success in sperm competition
depends directly on a males investment in gametes and
semen. Specifically, during spawning, four components
that can influence the probability of fertilization include:
(1) proximity of males to eggs during ejaculation; (2)
timing of sperm release and coordination with female egg
release; (3) number of sperm released and (4) sperm
traits such as swimming speed.
First, the proximity of males to females during sperm
competition is important because males that ejaculate
closer to females should give their sperm a competitive
advantage in a race to fertilize the eggs. Furthermore,
eggs and sperm may quickly diffuse, particularly when
they are released in flowing or turbulent water (Petersen
& Warner, 1998; Petersen et al., 2001). Closer proximity
of a male to a female during egg release has been
demonstrated to increase a males probability of

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

1873

1874

J . A . S T O L T Z A N D B . D . N E FF

fertilization during sperm competition in several species

including brook trout, Salvelinus fontinalis (Blanchfield
et al., 2003), Atlantic cod, Gadus morhua (Hutchings et al.,
1999) and coho salmon, Oncorhynchus kisutch (Schroder,
1973).
Secondly, timing of sperm release is important because
coordination with female egg release can increase the
probability of spermegg collisions and ovarian fluids can
modify the behaviour of sperm influencing the outcome
of sperm competition. If a male releases sperm too early,
then the concentration of sperm could diffuse and
become too dilute (Levitan, 1998; Petersen & Warner,
1998; Petersen et al., 2001) or the sperm could die or
otherwise cease forward motion by the time a female
releases her eggs. If a male releases sperm too late, then
the eggs may be already fertilized. Hoysak & Liley (2001)
concluded that in sockeye salmon (Oncorhynchus nerka) a
brief window exists during which eggs are available to be
fertilized and that the proximity of males to females and
timing of sperm release are critical factors determining
success in sperm competition. In the Arctic charr
(Salvelinus alpinus) the presence of ovarian fluid increases
sperm longevity, swimming speed and linearity of sperm
movement even though ovarian fluid is more viscous
than freshwater (Turner & Montgomerie, 2002). Cations
such as Mg2+ and Na+ in the ovarian fluid may increase
motility and longevity of sperm (e.g. Linhart et al., 2002;
Alavi et al., 2004).
Thirdly, sperm number is important for fertilization
success when sperm competition follows a raffle-like
process. In a raffle, each males probability of fertilization is related to the number of sperm he contributes to
the competition (Parker, 1990). For example, in bluegill
(Lepomis macrochirus), in vitro fertilization trials that
competed ejaculates from pairs of males showed that a
males paternity increased as his relative sperm number
increased (Neff et al., 2003). However, another study on
Atlantic salmon (Salmo salar) found no relationship
between paternity and sperm number (Gage et al., 2004).
Fourthly, sperm traits such as swimming speed can be
important during sperm competition (reviewed by Snook,
2005). Faster swimming sperm may be able to reach the
egg first and thereby increase success during competition.
For example, in Atlantic salmon, males with faster
swimming sperm fertilized a greater proportion of eggs
in two-male competition trials, and relative sperm swimming speed captured about half of the variation in
fertilization success (Gage et al., 2004). Swim speed
would be especially important if sperm also were able to
home in on the egg (e.g. Turner & Montgomerie, 2002).
The purpose of the present study was to examine
mechanisms used during sperm competition by males
that utilize different mating tactics in bluegill. Specifically, we incorporated information on proximity of males
to females and timing of sperm release to assess the
contribution of sperm number, length and speed to
fertilization success.

Methods
Study species
Male bluegill have a discrete life history polymorphism
called parental and cuckolder (Gross, 1979; Gross &
Charnov, 1980). In lake Opinicon (Ontario, Canada;
4434N, 7619W), parentals delay sexual maturation
until 7 years of age (Gross, 1982). During the spawning
season parentals compete to construct nests in colonies
(Dominey, 1981; Gross, 1982). After nest construction is
completed females arrive at the colony as a group.
Spawning involves a female entering a parentals nest
and through a distinctive motion called a dip, the
female releases a small batch of eggs at the utmost
horizontal position of her dip (Gross, 1982). A female
may dip hundreds or even thousands of times before
leaving the parentals nest. Spawning lasts for a day, after
which the females leave the colony and the parentals
remain to provide sole care to the developing offspring.
Cuckolders mature precociously at 2 years of age and
behave as sneakers (Gross, 1982). Sneakers hide in the
debris at the edge of a parentals nest and streak into the
nest during a female dip, stop momentarily to release
sperm and typically are aggressively chased out by the
parental. Sneakers are successful at entering the nest in
about 17% of intrusion attempts (Gross, 1982), and
when successful they fertilize a mean of 89% of the eggs
(Fu et al., 2001). At around 4 years of age, cuckolders
begin to mimic female morphology and behaviour and
are referred to as satellites (Gross, 1982). A parental will
allow access of a satellite to his nest during spawning and
may even court the satellite, suggesting that the parental
typically cannot distinguish between a satellite and a
female (Dominey, 1980; Gross, 1982; Neff & Gross,
2001). During spawning, satellites typically position
themselves between the female and parental and fertilize
a mean of 67% of the eggs (Fu et al., 2001). Parentals
compete with sneakers in about 7% of female dips and
with satellites in about 3% of female dips; parentals
rarely experience competition simultaneously from both
types of cuckolder males (Neff et al., 2003).
Using underwater cameras, we have previously determined that when parentals and sneakers are in direct
competition for a batch of eggs, a parentals urogenital
pore is a mean of 23 mm (range among nests 8
36 mm) from that of the females, whereas a sneakers
urogenital pore is a mean of 38 mm (range among
nests 2448 mm) from that of the females (Stoltz &
Neff, 2006). Furthermore, sneakers release their sperm
an estimated 0.46 s (range among nests 0.310.97 s)
after parentals. Satellites release their sperm at the same
time as parentals and are a mean of 17 mm (range among
nests 1126 mm) from a females urogenital pore,
whereas when a satellite is present, parentals are a mean
of 64 mm (range among nests 4284 mm) from the
female (Stoltz & Neff, 2006).

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

Sperm competition in bluegill

All of the experiments in the current study were

conducted during the MayJuly breeding seasons of 2004
and 2005 at the Queens University Biological Station on
Lake Opinicon. At the onset of spawning, males and
females were collected from colonies and brought back to
our field laboratory (near the site of collection). An egg or
ejaculate sample was collected by applying gentle pressure on a fishs abdomen. Male types were differentiated
based on behaviour (at the time of collection), size and
coloration (Gross, 1982). Males were anaesthetized with
clove oil immediately prior to ejaculate collection and
only males with free-flowing milt were used. All egg and
ejaculate samples were used within 2 min of collection.
Care was taken to ensure that no water or urine
contaminated or activated the ejaculate sample (which
was verified by viewing sperm samples with a microscope).
Sperm traits
We measured three sperm traits as possible indexes of
quality: curvilinear swim speed, path linearity and
flagellum length. To measure the first two traits, a total
of 5 lL of freshly collected ejaculate was added to 750 lL
high osmolarity extender (20 mM Tris, 2 mM KCl,
200 mM NaCl, pH 9.0). About 5 lL of this mixture was
then placed onto an Improved Neubaur haemocytometer
(Fisher Scientific, Ottawa, ON, Canada) and covered with
a slip. The haemocytometer was then placed under a
Motic Digital Microscope (model DMB1-223; Upstate
Technical Equipment Company, East Syracuse, NY,
USA). The sample was focused at a total magnification
of 100 and then activated by flooding the haemocytometer with lake water (at 20 C). The sample was
recorded for a minimum of 60 s. The video camera
recorded 30 frames per second.
We analysed the videos using the N I H I M A G E J
program (v. 1.34; National Institutes of Health, Bethesda, MD, USA). At both 5 and 10 s post-sperm activation, we analysed 11 consecutive frames (i.e. one-third
of a second). A single primary grid (0.2 0.2 mm) was
selected on the haemocytometer from which 10 sperm
were selected and tracked. The position of the sperm
was recorded in each consecutive frame and these data
were used to calculate the curvilinear speed of the
sperm as well as the path linearity. Curvilinear speed
incorporates the specific path followed by the sperm
and equals the total distance travelled divided by the
time to travel that distance; path linearity varies
between 0 and 1, where a 0 implies the sperm started
and ended at the same location and a 1 implies the
sperm moved in a straight line (see Kime et al., 2001).
The two time intervals were selected to encompass the
possible fertilization window for the eggs in nature and
the speed and path linearity values from the two times
were averaged into a single index of speed for each
male.

1875

To measure flagellum length, 1 lL of the freshly

collected ejaculate was added to 150 lL of fixative (3%
glutaraldehyde in 0.1 M cacodylate buffer, pH 7.0). The
fixed sample was spread across an Improved Neubaur
haemocytometer and covered with a slip. A photograph
was then taken of the sample using the digital microscope. The U T H S C S A I M A G E T O O L (v. 3.0; University of
Texas Health Science Centre, San Antonio, TX, USA) was
used to trace the flagellum of 10 haphazardly selected
sperm and to determine the length of the flagellum. For
each sperm, the flagellum was traced three times and
these measurements were averaged (the three measurements were highly repeatable: r > 0.9 for pairs of data).
Sperm competitiveness
We tested ejaculate competitiveness of sneakers, satellites
and parentals by using an experiment involving sperm
number, distance from eggs during ejaculation and time
difference between male ejaculations (for sneakers only).
For each successive competition trial, we alternated which
male type (cuckolder or parental) was used first to collect
the ejaculate. Various volumes of ejaculate, ranging from
0.5 to 9.6 lL from a sneaker or satellite and 0.5 or 1.0 lL
from a parental, were introduced above 30100 eggs
(freshly collected from a gravid female) from two separate
pipettes. These volumes were selected to provide a range
of sperm numbers for competing males. The pipettes were
positioned to mimic the actual mean distances between
the male and female urogenital pores as determined from
previous video analysis (Stoltz & Neff, 2006).
For the sneakerparental trials the sperm from the
sneaker was introduced shortly after that of the parental
to correspond to the time difference calculated from the
films. To determine our accuracy to simulate the time
difference the experimenter trained using a timing device
that was triggered in a similar manner to the operation of
the pipettes. Based on the timing device, we determined
that we could precisely repeat a time delay of 0.46 s
(mean 0.47 s; range 0.430.53 s; one-sample t19
1.51, P 0.15; 0.46 s was the estimated time delay for
sneakers). For satelliteparental trials the sperm were
introduced simultaneously.
Sperm density was calculated for each male after the in
vitro sperm introduction by taking a sample of 3 lL of the
same milt sample used in the experiment. The 3 lL were
diluted 150-fold with water and a sample of this solution
was placed on an Improved Neubaur haemocytometer.
The sperm were counted in ten 0.2 0.2 mm grids under
a compound light microscope at 400 magnification.
Sperm density was calculated by multiplying the mean
sperm count per grid by the dilution factor and volume of
liquid in each cell.
The fertilized eggs were maintained in 1 L jars with
aeration and at ambient temperature on a natural light
cycle for 7 days. Water changes (with lake water) were
performed daily. At the end of the 7 days, all larvae were

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

1876

J . A . S T O L T Z A N D B . D . N E FF

preserved in 95% ethanol. DNA was later isolated from the

adults and a subset of larvae, and microsatellite DNA
fingerprints were obtained by using multiplex genotyping
(Colbourne et al., 1996; Neff et al., 2000). Up to four primer
sets (locus: RB7, RB20, LMAR10, LMAR14; published in
DeWoody et al., 1998 and Schable et al., 2002) were
simultaneously amplified, and all competing parental
and cuckolder offspring were unambiguously identified
by using exclusion methods (Chakraborty et al., 1988).
Based on the paternity data, we used Neff & Wahls
(2004) maximum-likelihood algorithm to fit the following raffle equation:
Nc
Stc

;
Nc Np Stc rStp

where Nc is the number of offspring sired by the

cuckolder and Np is the number of offspring sired by
the parental, Sc and Sp are the numbers of sperm
provided by the cuckolder and parental, r is a measure
of the loading given to the parentals sperm relative to
the cuckolders sperm and t is a measure of the economy
of scale to sperm number (i.e. a nonlinear effect of sperm
number; see Parker, 1990; Neff & Wahl, 2004). Neff &
Wahls (2004) algorithm determines the most-likely
value of r and t and whether or not these values differ
significantly from the fair-raffle model (r 1; t 1) and
from the sperm independent model, in which case
fertilization success is independent of sperm number
(t 0). Using the best-fit lines derived from eqn 1, we
calculated the coefficient of determinations for the two
competition trails (Zar, 1999, p. 335).
By rearranging eqn 1, we next calculated the relative
sperm number needed to explain equal fertilization
success and the fertilization success observed in nature
for competing males:

1
Sc
pr t

;
2
1p
Sp
where p is the paternity of the sneaker or satellite as
calculated from Nc/(Nc + Np). The value of p was set to
either 50% or 89% for sneakers and either 50% or 67%
for satellites; the latter two numbers in each pair
represents the estimated fertilization success in nature
(see Fu et al., 2001). To determine the sensitivity of the
Sc/Sp ratio for the fertilization success estimates in nature,
we re-sampled the original data set with replacement
1000 times and recalculated r and t and Sc/Sp (Manly,
1997). For the resulting distribution of Sc/Sp values we
assessed the 5th and 95th percentiles.
Statistical analyses
A KruskalWallis nonparametric test was used to compare curvilinear swim speed among the three male types
because this variable deviated from a normal distribution
for one of the male types. Analysis of variance (A N O V A )

was used to compare path linearity or flagellum length

among the three male types because these variables did
not deviate from a normal distribution. A linear regression was used to assess the relationship between flagellum length and either speed or path linearity. A
Spearmans correlation was also used to assess the
relationship between each of the sperm traits and male
competitiveness. This was done by taking the difference
in each trait between competing males in the sperm
number trials [i.e. (sneaker traitparental trait) or (satellite traitparental trait)] and correlating the difference
with the residuals from the raffle equations. For the
correlation analyses, data from all males were combined.
We used Spearmans correlation because, although the
residuals did not significantly differ from zero (P > 0.10
for each) and where symmetrically distributed about a
mean of zero, they did significantly differ from a normal
distribution (ShapiroWilks, P < 0.001). We did not
measure all sperm traits in each field season, so sample
sizes vary among the different analyses. All sperm traits
were performed by a researcher who was blind to the male
type. All statistics were performed using J M P (v. 4.0.4;
SAS, Cary, NC, USA); mean values are reported 1 SE.

Results
Sperm traits
Curvilinear speed was variable among sperm within
ejaculates with a mean coefficient of variation of 35%.
Sperm length was less variable with a mean coefficient of
variation of 8%. However, the 10 sperm measured
should provide precise mean estimates for each male;
for example, the standard error relative to the mean for
speed and length were 11% and 2.5% respectively.
Across all males, the sperm curvilinear speed measurements at 5 and 10 s post-activation were highly correlated (r 0.93, P < 0.001, n 24). The path linearity
index was also correlated at these two times (r 0.70,
P < 0.001, n 24). The means of the 5 and 10 s values
were not significantly different among the three male
types for speed (Fig. 1a; parental: 150 12 lm s)1;
sneaker: 137 13 lm s)1; satellite: 165 11 lm s)1;
v22 3.10, P 0.21) or path linearity (Fig. 1b; parental:
0.75 0.03; sneaker: 0.76 0.03; satellite: 0.78 0.03;
F2,21 0.21, P 0.82). There was also no difference in
the mean flagellum length across the three male types
(Fig. 2; parental: 40.1 0.6 lm; sneaker: 40.4 0.6 lm;
satellite: 40.1 0.6 lm; F2,35 0.10, P 0.91). Flagellum length was not related to either speed (r2 0.0064,
b 0.08, P 0.72, n 24) or path linearity (r2
0.062, b )0.25, P 0.23, n 24).
Sperm competitiveness
We conducted 84 competition trials involving either a
sneaker and parental (n 37) or a satellite and parental

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

Sperm competition in bluegill

44

200

Sperm flagellum length (m)

Sperm curvilinear speed (m s1)

(a) 220

180
160
140
120
100

42

40

38

36

80
Parental

Sneaker

Satellite

Life history

Parental

Sneaker

Satellite

Life history

(b) 1.0

Fig. 2 Flagellum length of sperm from parental, sneaker, and

satellite male bluegill (Lepomis macrochirus). Data are represented by
box plots that summarize the 10th, 25th, 50th, 75th and 90th
percentiles. Dots denote data points that fall outside the 10th90th
percentile range. Sample sizes are 13, 14 and 11 for parental, sneaker
and satellite respectively. There was no statistical difference in
flagellum length among the three male mating tactics (see Results
for statistics).

0.9

Sperm path linearity

1877

0.8

0.7

0.6

0.5
Parental

Sneaker

Satellite

Life history

Fig. 1 Sperm traits for parental, sneaker and satellite male bluegill
(Lepomis macrochirus). Data are (a) curvilinear speed and (b) path
linearity and are represented by box plots that summarize the 10th,
25th, 50th, 75th and 90th percentiles. Dots denote data points that
fall outside the 10th90th percentile range. Sample sizes are 8, 7 and
9 for parental, sneaker and satellite respectively. There was no
statistical difference in either measure among the three male mating
tactics (see Results for statistics). A path linearity value of 0 implies
the sperm started and ended at the same location, whereas a value of
1 implies the sperm moved in a straight line.

(n 47). For parentals, sperm density had a mean

of 4.3 106 sperm lL)1 (range 0.9 10614.7 106
sperm lL)1), for sneakers, sperm density had a mean
of 10.5 106 sperm lL)1 (range 5.0 10622.7
106 sperm lL)1), and for satellites, sperm density had a
mean of 10.7 106 sperm lL)1 (range 0.8 106
23.7 106 sperm lL)1).
In total, we determined the genotypes of 1632
offspring with a mean of 19 offspring per trial
(range 736 offspring per trial). For the sneaker
parental trials, the maximum-likelihood algorithm
revealed that the sperm loading r was 0.58 and the
economy of scale t was 1.00 (Fig. 3a). This model was
nonlinear and did not conform to the fair-raffle because
the r parameter was significantly different from 1

(P < 0.001). The equation of the line captured 75% of

the variation (i.e. r2 0.75). For the satelliteparental
trials, the algorithm revealed that the sperm loading r
was 0.35 and the economy of scale t was 0.95 (Fig. 3b).
This model was similarly nonlinear and did not conform
to the fair-raffle because the r parameter was significantly
different from 1 (P < 0.001). The equation of the line
captured 51% of the variation (i.e. r2 0.51). For both
sets of trials the t parameter was not significantly
different than 1 (P > 0.48 for both), but t was significantly different than 0 (P < 0.001 for both) and therefore
we were able to reject the sperm independent model.
When we removed the effect of sperm number by
using the residual paternity from the raffle equations, we
found no relationship between males competitiveness
and the difference in males sperm flagellum length
(rS )0.20, P 0.49, n 14), curvilinear speed (rS
)0.11, P 0.75, n 11), or path linearity (rS )0.34,
P 0.31, n 11). Sample sizes are smaller for these
latter analyses because we did not always have sperm
trait measurements for both males in a competition trial.
Using eqn 2 and setting p to 50%, we determined that
despite the disadvantaged position, sneakers need to
release only 0.58 times the number of sperm that
parentals release to obtain equal fertilization success.
For satellites, we determined that they need to release
only 0.34 times the number of sperm that parentals
release to obtain equal fertilization success. By instead
setting p to 89% for sneakers and 67% for satellites (the
estimated fertilization success in nature; Fu et al., 2001),
we determined that sneakers release an estimated 4.7
(5th and 95th percentiles 2.38.3) times the number

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

1878

J . A . S T O L T Z A N D B . D . N E FF

(a) 1.0

Sneaker paternity

0.8

0.6

0.4

0.2

0.0
0

10

15

20

25

30

35

Relative sperm number (sneaker / parental)

(b) 1.0

Satellite paternity

0.8

0.6

0.4

0.2

0.0
0

10

15

20

25

30

Relative sperm number (satellite / parental)

Fig. 3 The relationship between sperm number and male competitiveness in in vitro fertilization trials involving parental, sneaker and
satellite male bluegill (Lepomis macrochirus). Competition trials
involved (a) sneakers vs. parentals (n 37) where the equation of
the line is: sneaker paternity S1:00
=S1:00
0:58S1:00
c
c
p  or (b) satellites vs. parentals (n 47) where the equation of the line is: satellite
paternity S0:95
=S0:95
0:35S0:95
c
c
p . The lines were fit using the
raffle model of Neff & Wahl (2004).

of sperm relative to parentals and satellites release an

estimated 0.70 (5th and 95th percentiles 0.391.1)
times the number of sperm relative to parentals.

Discussion
Many fishes with external fertilization are characterized
by sperm competition (Petersen & Warner, 1998). Within
these species, sperm traits such as swimming speed,
number of sperm released, timing of sperm release and
proximity of males to eggs all can influence a males
fertilization success (Hoysak & Liley, 2001; Blanchfield
et al., 2003; Neff et al., 2003; Gage et al., 2004). In this
study, we examined the role of sperm number and
quality in male competitiveness in the bluegill (L. macrochirus) using in vitro fertilization experiments that

mimicked natural spawning by incorporating proximity

and timing of sperm release.
We found that sperm competition in bluegill does not
conform to a fair raffle, but instead represent a loaded
raffle whereby one males sperm has a competitive
advantage over another males sperm (Parker, 1990).
Based on eqn 1, an r parameter <1 indicates that
parental sperm are competitively inferior (see Neff &
Wahl, 2004). For sneakers, even though they are in a
disfavoured role based on proximity and timing of sperm
release, our competition trials revealed that their sperm
still have an advantage over parental sperm (r 0.58). It
is unlikely that this competitive advantage relates to
differences in sperm flagellum length or swim speed
because we found no difference in these traits between
any of the three male types (also see Leach &
Montgomerie, 2000; Burness et al., 2005), and there
was no relationship between the difference in these traits
between two competing males and their competitiveness.
However, caution is warranted for the latter conclusion
because our sample sizes were small for those particular
analyses (n 11 or 14) and the large variation in sperm
numbers used in the competition trials may have reduced
statistical power to detect independent effects of the
sperm traits. Two studies also report no difference in
sperm longevity between sneakers and parentals (Leach
& Montgomerie, 2000; Burness et al., 2005), although we
did not assess longevity here.
Alternatively, it is possible that ATP concentrations
contribute to the difference in competitiveness because
sperm from sneakers have higher ATP concentration
than those from parentals (Burness et al., 2004, 2005).
For satellites, the model revealed that their sperm
similarly have an advantage over parental sperm (r
0.35), but this could relate simply to the proximity
advantage of satellites (Stoltz & Neff, 2006). Interestingly,
we have previously shown that when equal volumes of
milt are mixed from pairs of parentals and either sneakers
or satellites and placed in direct contact with eggs,
parental sperm appear more competitive than cuckolder
sperm (Neff et al., 2003; also see Schulte-Hostedde &
Burness, 2005). In conjunction with the data presented
here, which incorporate proximity of males to eggs, the
results suggest that the race to the egg is paramount in
understanding males competitiveness; although some
care is needed when interpreting the data of Neff et al.
(2003) because they did not have actual sperm counts for
their males.
There are differences in the mechanisms that sneakers
and satellite use to out-compete rival parentals during
sperm competition. Sneakers intrude on a spawning
female and parental by streaking in from the periphery of
the parentals nest. Previously, we have shown that
sneakers gain access to eggs about half a second after
parentals and end up 1.6 times farther from the eggs than
parentals during ejaculation (Stoltz & Neff, 2006).
Nevertheless, sneakers are able to fertilize an average of

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

Sperm competition in bluegill

89% of the eggs (Fu et al., 2001). Here, we used the in

vitro fertilization experiment to predict that sneakers on
average must release about 4.7 times more sperm than
parentals in order to fertilize 89% of the eggs. Conversely, satellites temporarily position themselves
between the female and parental during spawning and
end up closer to the female at only an average of 26% of
the distance that parentals are to females when eggs are
released (Stoltz & Neff, 2006). Based on the in vitro
fertilization experiment, we predict that satellites release
only about 70% as many sperm as parentals to fertilize
the 67% of the eggs that Fu et al. (2001) calculated. These
calculations of relative sperm number are only estimates
as our analysis utilized the mean proximity of males to
females and did not incorporate the variance in these
estimates. This approach could introduce error into our
sperm number estimates if there is a nonlinear relationship between proximity and sperm competitiveness.
Furthermore, we were unable to incorporate potential
differences in mechanisms of ejaculation, such as differences in musculature, and the effect of water turbulence
from male and female movements during spawning.
Nevertheless, our analysis suggests that sneakers outcompete parentals in part by releasing more sperm,
whereas satellites out-compete parentals in part by
obtaining a superior position during spawning. To confirm the former, however, requires measuring the numbers of sperm released in nature.
How is it possible that sneakers are able to release
nearly five times more sperm than parentals? Leach &
Montgomerie (2000) showed that the total volume of
milt in a stripped ejaculate of a sneaker had a mean of
11.9 lL, whereas that for parentals was over five times
greater at 63.4 lL. First, we found that the density of
sperm in ejaculates of sneakers is greater at about 2.4
times the density of the ejaculates of parentals (
10.5 106/4.3 106). Leach & Montgomerie (2000)
estimated the relative density of sperm in ejaculates of
sneakers and parentals to be 1.4 (16.5 106/
11.5 106). Secondly, because sneakers overall access
only about 7% of spawning events (i.e. dips) and
outnumber parentals 3 : 1 (Neff et al., 2003), individual
sneakers access only 2.3% (7/3) the number of
spawning events that individual parentals access. Based
on these spawning numbers and the sperm density
estimates of Leach & Montgomerie (2000) (the lower
estimate of relative sperm number), we can estimate the
relative numbers of sperm per spawning event (dip)
available for sneakers and parentals. Sneakers have a
total
of
196 106 sperm
(11.9 lL 16.5 106
sperm lL)1), whereas parentals have a total of
729 106 sperm (63.4 lL 11.5 106 sperm lL)1).
Thus, parentals have 3.7 times the number of sperm that
sneakers have. However, because individual parentals
participate in 43 times (1/0.023) the number of
spawning events that sneakers do, individual sneakers
actually have about 11 times (43/3.7) the number of

1879

sperm available per spawning event (dip). This latter

number suggests that sneakers could easily release our
estimated 4.7 times more sperm than parentals during
individual intrusions. These calculations, however, do
not incorporate potential regeneration rates of sperm
during the single day of spawning. We are currently
investigating testes histology and sperm regeneration
rates in bluegill.
The importance of sperm traits for male competitiveness appears to differ among externally fertilizing fishes.
In Atlantic salmon, using in vitro fertilization experiments, Gage et al. (2004) showed that sperm number had
no effect on fertilization success of pairs of competing
males, whereas relative sperm swim speed captured a
large portion of the variance in competitiveness. In
bluegill, our data suggest the opposite relationship; i.e.
number is more important than speed. It is possible that
the difference in these two studies is attributable to
variation in ecology between the species. Salmon typically spawn in flowing rivers, whereas the bluegill we
studied spawn in relatively calm waters of the lake.
Indeed, the experiment of Gage et al. (2004) was
conducted in a funnel stream that mimicked the flowing
water of natural rivers and our experiment was conducted in still water that mimicked conditions on the lake
bottom. Flowing water may favour faster swimming
sperm, whereas still water may select for more sperm.
Differences in the effect of ovarian fluid on sperm traits
(e.g. Turner & Montgomerie, 2002) also may differ
among species and deserves attention. To our knowledge
the Atlantic salmon study by Gage et al. (2004) is the
only other study of an externally fertilizing fish that uses
controlled experiments to investigate the roles of sperm
quality and number on males competitiveness (for an
example in an internal fertilizing fish, see Evans et al.,
2003 and for examples from other taxa see Birkhead
et al., 1999; Gage & Morrow, 2003; Garca-Gonzalez &
Simmons, 2005).
In conclusion, we found no statistical difference in
flagellum length, curvilinear speed, or path linearity of
sperm from male bluegill that utilize three different
mating tactics. Competition trials revealed that sperm
number strongly influences fertilization success, and
independent of sperm number, sperm from sneakers
and satellites have a competitive advantage over those
from parentals. Although satellites are closer to the eggs
than parentals, sneakers are more distant than parentals
and therefore some other aspect of sperm quality must
contribute to the increased competitiveness of sperm
from sneakers.

Acknowledgments
Elizabeth Clare, Angela Dang, Tim Hain, Joy Marcus,
Margaret Orlowski and Nick Sweet provided field and
laboratory assistance. Tim Hain, Trevor Pitcher and four
anonymous reviewers provided helpful comments on the

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

1880

J . A . S T O L T Z A N D B . D . N E FF

manuscript. Parts of the work were conducted at the

Queens University Biological Station where Frank
Phelan and Floyd Connors provided logistical assistance.
The work was supported by an Ontario Graduate Scholarship to JAS, and the Natural Science and Engineering
Research Council of Canada and a Premieres Research
Excellence Award (grants to BDN).

References
Alavi, S.M.H., Cosson, J., Karami, M., Amiri, B.M. & Akhoundzadeh, M.A. 2004. Spermatozoa motility in the Persian
sturgeon, Acipenser persicus: effects of pH, dilution rate, ions
and osmolality. Reproduction 128: 819828.
Arnqvist, G. & Danielsson, I. 1999. Copulatory behavior, genital
morphology, and male fertilization success in water striders.
Evolution 53: 147156.
Beatty, R.A., Bennett, G.H., Hall, J.G., Hancock, J.L. & Stewart,
D.L. 1969. An experiment with heterospermic insemination in
cattle. J. Reprod. Fert. 19: 491502.
Birkhead, T.R. & Mller, A.P. 1998. Sperm Competition and Sexual
Selection. Academic Press, CA.
Birkhead, T.R., Martnez, J.G., Burke, T. & Froman, D.P. 1999.
Sperm mobility determines the outcome of sperm competition
in the domestic fowl. Proc. Roy. Soc. Lond. (B) 266: 17591764.
Blanchfield, P.J., Ridgway, M.J. & Wilson, C.C. 2003. Breeding
success of male brook trout (Salvelinus fontinalis) in the wild.
Mol. Ecol. 12: 24172428.
Burness, G., Casselman, S.J., Schulte-Hostedde, A.I., Moyes,
C.D. & Montgomerie, R. 2004. Sperm swimming speed and
energetics vary with sperm competition risk in bluegill
(Lepomis macrochirus). Behav. Ecol. Sociobiol. 56: 6570.
Burness, G., Moyes, C.D. & Montgomerie, R. 2005. Motility, ATP
levels and metabolic enzyme activity of sperm from bluegill
(Lepomis macrochirus). Comp. Biochem. Physiol. 140: 1117.
Chakraborty, R., Meagher, T.R. & Smouse, P.E. 1988. Parentage
analysis with genetic markers in natural populations. I. The
expected proportion of offspring with unambiguous paternity.
Genetics 118: 527536.
Colbourne, J.K, Neff, B.D., Wright, J.M. & Gross, M.R. 1996.
DNA fingerprinting of bluegill sunfish (Lepomis macrochirus)
using (GT)n microsatellites and its potential for assessment of
mating success. Can. J. Fish. Aquat. Sci. 53: 342349.
DeWoody, J.A., Fletcher, D.E., Fletcher, E.E., Wilkins, S.D.,
Nelson, W.S. & Avise, J.C. 1998. Molecular genetic dissection
of spawning, parentage, and reproductive tactics in a population of redbreast sunfish, Lepomis auritus. Evolution 52: 1802
1810.
Dominey, W.J. 1980. Female mimicry in male bluegill sunfish
a genetic polymorphism? Nature 284: 546548.
Dominey, W.J. 1981. Maintenance of female mimicry as a
reproductive strategy in bluegill sunfish (Lepomis macrochirus).
Environ. Biol. Fish. 6: 5964.
Evans, J.P, Zane, L., Francescato, S. & Pilastro, A. 2003.
Directional postcopulatory sexual selection revealed by artificial insemination. Nature 421: 360363.
Fu, P., Neff, B.D. & Gross, M.R. 2001. Tactic-specific success in
sperm competition. Proc. Roy. Soc. Lond. (B) 268: 11051112.
Gage, M.J.G. & Morrow, E.H. 2003. Experimental evidence for
the evolution of numerous, tiny sperm via sperm competition.
Curr. Biol. 13: 754757.

Gage, M.J.G., Stockley, P. & Parker, G.A. 1995. Effects of

alternative male mating strategies on characteristics of sperm
production in the Atlantic salmon (Salmo salar): Theoretical
and empirical investigations. Phil. Trans. Roy. Soc. Lond. (B) 350:
391399.
Gage, M.J.G., Macfarlane, C.P., Yeates, S., Ward, R.G., Searle, J.B.
& Parker, G.A. 2004. Spermatozoa traits and sperm competition
in Atlantic salmon: relative sperm velocity is the primary
determinant of fertilization success. Curr. Biol. 14: 4447.
Garca-Gonzalez, F. & Simmons, L.W. 2005. Sperm viability
matters in insect sperm competition. Curr. Biol. 15: 271275.
Gross, M.R. 1979. Cuckoldry in sunfishes (Lepomis: Centrarchidae). Can. J. Zool. 57: 15071509.
Gross, M.R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Z.
Tierpsychol. 60: 126.
Gross, M.R. & Charnov, E.L. 1980. Alternative male life histories
in bluegill sunfish. Proc. Natl. Acad. Sci. USA 77: 69376940.
Hildemann, W.H. & Wagner, E.D. 1954. Intraspecific sperm
competition in Lebistes reticulatus. Am. Nat. 88: 8791.
Hoysak, D.J. & Liley, N.R. 2001. Fertilization dynamics in
sockeye salmon and a comparison of sperm from alternative
male phenotypes. J. Fish Biol. 58: 12861300.
Hutchings, J.A., Bishop, T.D. & McGregor-Shaw, C.R. 1999.
Spawning behaviour of Atlantic cod, Gadus morhua: evidence
of mate competition and mate choice in a broadcast spawner.
Can. J. Fish. Aquat. Sci. 56: 97104.
Kime, D.E., Van Look, K.J.W., McAllister, B.G., Huyskens, G.,
Rurangwa, E. & Ollevier, F. 2001. Computer-assisted sperm
analysis (CASA) as a tool for monitoring sperm quality in fish.
Comp. Biochem. Physiol. Part C 130: 425433.
Leach, B. & Montgomerie, R. 2000. Sperm characteristics
associated with different male reproductive tactics in bluegills
(Lepomis macrochirus). Behav. Ecol. Sociobiol. 49: 3137.
Levitan, D.R. 1998. Sperm limitation, gamete competition, and
sexual selection in external fertilizers. In: Sperm Competition
and Sexual Selection (T. R. Birkhead & A. P. Mller, eds), pp.
175217. Academic Press, CA.
Linhart, O., Cosson, J., Mims, S.D., Shelton, W.L. & Rodina, M.
2002. Effects of ions on the motility of fresh and demembranated paddlefish (Polyodon spathula) spermatozoa. Reproduction 124: 713719.
Manly, B.J.F. 1997. Randomization, Bootstrap and Monte Carlo
Methods in Biology, 2nd edn. Chapman and Hall, London.
Neff, B.D. & Gross, M.R. 2001. Dynamic adjustment of parental
care in response to perceived paternity. Proc. Roy. Soc. Lond. (B)
268: 15591565.
Neff, B.D. & Wahl, L.M. 2004. Mechanisms of sperm competition: testing the fair raffle. Evolution 58: 18461851.
Neff, B.D., Fu, P. & Gross, M.R. 2000. Microsatellite multiplexing
in fish. Trans. Am. Fish. Soc. 129: 584593.
Neff, B.D., Fu, P. & Gross, M.R. 2003. Sperm investment and
alternative mating tactics in bluegill sunfish (Lepomis macrochirus). Behav. Ecol. 14: 634641.
Parker, G.A. 1970. Sperm competition and its evolutionary
consequences in the insects. Biol. Revol. Chamb. Phil. Soc. 45:
525567.
Parker, G.A. 1990. Sperm competition games: raffles and roles.
Proc. Roy. Soc. Lond. (B) 242: 120126.
Petersen, C.W. & Warner, R.R. 1998. Sperm competition in
fishes. In: Sperm Competition and Sexual Selection (T. R. Birkhead
& A. P. Mller, eds), pp. 435463. Academic Press, CA.

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

Sperm competition in bluegill

Petersen, C.W., Warner, R.R., Shapiro, D.Y. & Marconato, A.

2001. Components of fertilization success in the bluehead
wrasse, Thalassoma bifasciatum. Behav. Ecol. 2: 237245.
Schable, N.A., Fischer, R.U. & Glenn, T.C. 2002. Tetranucleotide
microsatellite DNA loci from the dollar sunfish (Lepomis
marginatus). Mol. Ecol. Not. 2: 509511.
Schroder, S.L. 1973. Effects of density on the spawning success of
chum salmon in an artificial spawning channel. Master thesis,
University of Washington, Seattle.
Schulte-Hostedde, A.I. & Burness, G. 2005. Fertilization
dynamics of sperm from different male mating tactics in
bluegill (Lepomis macrochirus). Can. J. Zool. 83: 16381642.
Smith, R.L. 1984. Sperm Competition and the Evolution of Animal
Mating Systems. Academic Press, London.
Snook, R.R. 2005. Sperm in competition: not playing by the
numbers. Trend. Ecol. Evol. 20: 4653.

1881

Stockley, P., Gage, M., Parker, G. & Mller, A. 1997. Sperm

competition in fishes: The evolution of testis size and ejaculate
characteristics. Am. Nat. 149: 933954.
Stoltz, J.A. & Neff, B.D. 2006. Male size and mating tactic
influence proximity to females during sperm competition in
bluegill sunfish. Behav. Ecol. Sociobiol., 59: 811818.
Taborsky, M. 1998. Sperm competition in fish: Bourgeois males
and parasitic spawning. Trend. Ecol. Evol. 13: 222227.
Turner, E. & Montgomerie, R. 2002. Ovarian fluid enhances
sperm movement in Arctic charr. J. Fish Biol. 60: 15701579.
Zar, J.H. 1999. Biostatistical Analysis, 4th edn. Prentice Hall, NJ.
Received 17 December 2005; revised 27 April 2006; accepted 9 May 2006

2006 THE AUTHORS 19 (2006) 18731881

JOURNAL COMPILATION 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY