P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
C 2002)
Journal of Insect Behavior, Vol. 15, No. 6, November 2002 (°
Physical Restraint or Stimulation? The Function(s)
of the Modified Front Legs of Male Archisepsis
diversiformis (Diptera, Sepsidae)
William G. Eberhard1
Accepted August 9, 2002; revised September 20, 2002
Four hypotheses that could explain the elaborate species-specific morphology
of the clasping organs on the front legs of male Archisepsis diversiformis flies
were tested: direct male–male combat, mechanical fit, male–female conflict
of interests, and male stimulation of the female. Experimental modification
of the shape of the male clasper and of the female wing base where the male
clasped the female both strongly reduced the chances that a mount would result
in copulation. This reduction was not predicted by the male–male combat
hypothesis but was predicted by the others. Males in the field did nave to
fight other males to remain mounted on females, as expected by the male–
male combat hypothesis. Reduced male copulatory success was not due to
inferior male ability to grasp and hold onto the female’s wings, as predicted
by the mechanical fit and male–female conflict hypotheses but to a reduction
in the likelihood that the female would allow intromission, as predicted by the
stimulation hypothesis. By a process of elimination, and in combination with
data from a previous morphological study, the data support the hypothesis
that the species-specific aspects of grasping organs in these flies function to
stimulate females. Further behavioral data will be needed to test alternative
possibilities.
KEY WORDS: sexual selection; sepsid flies; lock-and-key; male–female conflict of interests.
1Smithsonian Tropical Research Institute, and Escuela de Biologı́a, Universidad de Costa Rica,
Ciudad Universitaria, Costa Rica.
2To whom correspondence should be addressed at Biologı́a, Universidad de Costa Rica, Ciudad
Universitaria, Costa Rica. Fax: +506 228 0001. e-mail: archisepsis@biologı́a.ucr.ac.cr.
831
C 2002 Plenum Publishing Corporation
0892-7553/02/1100-0831/0 °
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
832
16:10
Style file version Feb 08, 2000
Eberhard
INTRODUCTION
Nongenitalic male body parts that are modified to contact females during courtship or copulation often evolve relatively rapidly and divergently
compared with other body parts (Eberhard, 1985). Such contact structures
are often structurally complex and useful for distinguishing closely related
species because of their species-specific forms in many animal groups, including nematodes, spiders, schizomids, mites, copepods, isopods, eubranchiopods, decapods, millipedes, collembolans, and many insects (summary
by Eberhard, 1985). The portions of the male’s body that are specialized
for contact vary widely and include legs, cephalothorax, chelicerae, telson,
hysterosome, dorsal setae, antennae, periopods, mandibles, maxillary palps,
abdominal sterna, wings, forceps, and frontal horns. This general trend to
rapid divergent evolution resembles the even more widespread and general
trend for male genitalia to diverge rapidly and acquire elaborate morphology
(Eberhard, 1985).
Just as in genitalia, one common function of nongenitalic male contact
structures is to grasp the female. The elaborate forms of nongenitalic male
grasping structures often seem overly complex and diverse for the seemingly
simple mechanical demands of holding onto the female (e.g., Fig. 1), however, suggesting that they may also have additional functions. Several of the
hypotheses that were originally proposed to explain the evolution of male
genitalia could also explain the evolutionary tendency for nongenitalic male
grasping structures to evolve rapidly and divergently.
1. Direct male–male combat. Male grasping organs may enable males
to resist attempts by other males to displace them from females after
they have mounted (Darwin, 1871). Diversity in grasping structures
could result from adaptations to resist different types of displacement
attempts.
2. Mechanical fit. A greater degree of physical fit between male and
female structures could enable the male to hold onto the female
more effectively and, thus, avoid being displaced. Two types of selection could favor especially good fits. (A) The species-specific male
structures may fit only conspecific females, and be mechanically incapable of grasping females of other species, and the differences in
female morphology responsible for these fits may evolve to prevent
grasping by heterospecific males (Fraser, 1943; Freitag, 1974; Toro
and de la Hoz, 1976; see Shapiro and Porter [1989] for similar arguments regarding genitalic structures). (B) Females may discriminate
among conspecific males on the basis of mechanical fit (Eberhard,
1985). Diversity in male structures could thus result from selection on
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Function of Clasping Organs of Male Flies
Fig. 1. Front femur of males of four species of Archisepsis flies, illustrating the
species-specific shapes of the portions that clasp the female wing.
833
P1: ZCK
Journal of Insect Behavior [joib]
834
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
females to avoid heterospecific pairings (A) or from sexual selection
on males (B).
3. Male–female conflict. Male grasping organs may evolve as weapons
in coevolutionary races with females to obtain greater control over
critical reproductive processes. Diversity in male grasping structures
would result from the need to overcome different female defenses
that have evolved to reduce her susceptibility to being grasped. The
advantage to the female of avoiding being grasped would be to avoid
possible costs of being seized, such as reduced chances to feed or lay
eggs or increased danger from predation (Arnquist and Rowe, 1995;
Holland and Rice, 1998; see Lloyd [1979] and Alexander et al., [1997]
for similar arguments regarding genitalic structures).
4. Stimulation. Species-specific male organs may stimulate the female
in ways that induce the female to favor the male with cooperative responses. Selection could favor male stimulatory abilities in two contexts: (A) discrimination of conspecific versus heterospecific males
and (B) discrimination of conspecific males with designs better able to
elicit cooperative female responses (Robertson and Paterson, 1982;
Eberhard, 1985; Battin, 1993a,b). Diversity in male structures could
thus result from either selection against heterospecific pairing (A) or
sexual selection on males (B).
The present study of the functional significance of the specialized front
legs of the sepsid fly Archisepsis diversiformis (Ozerov) tests these hypotheses. As is common in many sepsids (Parker, 1972; Ward, 1983; Schulz, 1999;
Eberhard, 2000; Blankenhorn et al., 2000), males of A. diversiformis wait
for females at oviposition and feeding sites (dung and carrion), mount them
before and during oviposition, and copulate when oviposition is finished.
Copulation lasts approximately 20–25 min (Eberhard and Huber, 1998).
Typically there are many more males than females at a given site (the mean
for 297 flies at eight dung pats in Costa Rica was 2.9 males for each female).
Male sepsids generally use their specialized front legs to clasp the bases
of the female’s wings (Šulc, 1928; Hennig, 1949; Parker, 1972; Pont, 1979;
Eberhard, 2001a). The male’s front femur and tibia are often more or less
elaborately sculpted and provided with spines and setae on their ventral surfaces (Figs. 1 and 2A) and generally have species-specific shapes (e.g., the
male’s front femur and tibia often differ among species in the same genus
[Duda, 1925, 1926; Hennig, 1949; Pont, 1979]). There are stress receptors
(campaniform sensilla) in the veins of the female wing near the area where
the male’s leg clasps it (Eberhard, 2001a). Males mount females prior to
copulation, and the female often resists a mounted male by shaking energetically from side to side, kicking, running, and bending her abdomen
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Function of Clasping Organs of Male Flies
Fig. 2. (A) Front leg of a male A diversiformis just beginning to release its grasp
on the wing of a female, showing how the specialized ventral surface of the femur
meshes with the dorsal surface of the base of the wing; the curved ventral surface
of the anterior–ventral thumb fits against the curved dorsal surface of the stem vein.
(B) Male femur with a droplet of glue on the modified ventral surface of the femur.
Scale bar, 200 µm.
835
P1: ZCK
Journal of Insect Behavior [joib]
836
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
ventrally (Parker, 1972; Ward, 1983; Ward et al., 1992; Allen and Simmons,
1996; Blanckenhorn et al., 2000; Eberhard, 2001a, 2000). Mounted males
court females (Eberhard, 2001b), and intromission requires active female
collaboration in which she raises her abdomen and lifts its tip (proctiger)
to expose the opening of her vulva (Eberhard, 2002). Allen and Simmons
(1996) showed that males which were found copulating had more symmetric
foretibiae than those found riding females prior to copulation. They argued,
in possible reference to mechanical fit, that this difference was due to an
inferior clasping ability of less symmetric front legs. The possibility that different stimuli from more asymmetric males induced differences in female
cooperation and resistance behavior was not taken into account, however, in
their discussion, despite the fact that they found that the intensity of female
resistance behavior varied and thus might have influenced male success.
Apparently there has been only one explicit attempt to explain why the
front legs of male sepsids are so elaborate. Šulc (1928) argued that the surfaces of the femur and the tibia fit tightly against each other just behind the
rear margin of the female wing, forming a clamp which snaps shut. A morphological study of six species (including A. diversiformis), in which pairs
were frozen while the male clamped the female’s wings (Eberhard, 2001a),
showed that this idea is incorrect: in all species the specialized areas of both
the femur and the tibia pressed on the female’s wing, not against each other
as proposed by Šulc. The elaborate forms of the male femur and tibia meshed
very precisely with the contours of the female’s wing veins. Contrary to the
predictions of the mechanical fit and male–female conflict hypotheses, however, female morphology proved to be relatively constant interspecifically.
Most of the differences among the designs of males of different species seem
to represent alternative ways of grasping the same generalized female morphology. Female sense organs are present in the wing near where the male
grasps her. The stimulation hypothesis was thus favored by these results
(Eberhard, 2001a), but mostly by a process of elimination of the others.
The present study tests the four hypotheses in A. diversiformis Ozerov,
using field observations, and direct manipulations of the shape of the male’s
femoral clamp (Fig. 2) and of the female’s wing base. All of the hypotheses
except direct male–male combat predict that experimental alteration of the
form of the femoral clamp should reduce male mating success when male–
female pairs are isolated from other males. A lack of reduction in mating
success in this context would thus constitute strong evidence favoring the
direct male–male combat hypothesis over the other three. In addition, the
mechanical fit and male–female conflict of interests hypotheses predict that
this reduction will be due to a reduction in the male’s ability to remain
mounted. In contrast, the stimulation hypothesis predicts that the reduced
mating success will be due to reduced female acceptance behavior but that
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
Function of Clasping Organs of Male Flies
16:10
Style file version Feb 08, 2000
837
this reduced acceptance will not necessarily be associated with a reduction
in the male’s ability to stay mounted. A lack of reduction in the modified
male’s ability to stay mounted would thus constitute evidence against the
mechanical fit and conflict of interests hypotheses but would be compatible
with the stimulation hypothesis.
METHODS
Field observations were made around fresh cow dung near San Antonio
de Escazu (elevation, 1350 m), San Jośe Province, Costa Rica. Experimental
modifications of males and females were performed in captivity in Panama
and Costa Rica under similar conditions. All experimental flies were virgin
A. diversiformis that had been raised in captivity on previously frozen cow
dung and separated by sex within 18 h of emergence as adults. They had
emerged as adults at least 3 days before they were used in experiments (the
flies are not sexually mature until 1–2 days of age). All behavioral observations were made in the afternoon, and flies were prepared for observation
at least 2 h previously that same morning. Flies in Panama were raised from
females collected near the dung of howler monkeys (Alouatta palliata) on
Barro Colorado Island, Panama (elevation, 20 m); flies in Costa Rica were
raised from females collected at cow dung near San Antonio de Escazu.
Both populations are classified as A. diversiformis, but some details of male
behavior during copulation differ (W. Eberhard, in preparation). They also
behaved differently in the experiments (below), so their results were analyzed separately.
Male front leg morphology was experimentally modified without anesthesia. Each experimental (modified) male was aspirated into a plastic sack,
where he was immobilized by pressing the walls of the sac gently against a
flat surface. One wing was then seized with a fine forceps, and the fly was
removed and grasped by another forceps whose tips were covered with soft
foam rubber. This forceps was held shut by slipping a ring over the tips, and
the immobilized fly was placed under a dissecting microscope, positioned so
that the portion of his body to which glue was to be applied was exposed
to view. Modified male flies received a drop of water-soluble, polyvinyl glue
(Resistol; H. B. Fuller Co., Costa Rica) (which was nearly transparent when
dry). The glue was applied with the tip of a fine pin to the ventral surface of
the distal portion of each front femur (Fig. 2B). To apply the drop and allow
it to dry undisturbed, it was necessary to hold the leg partly extended by
seizing the tarsus with another forceps. The tips of this forceps were covered
by a thin layer of foam rubber and a piece of paper towel to avoid damage
to the male’s tarsi. The male’s leg was extended for approximately 30–60 s
P1: ZCK
Journal of Insect Behavior [joib]
838
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
until the glue had dried. Each control male was immobilized as above and
received a drop of glue on the dorsal surface of his thorax.
Experimental females were also immobilized as above and received
glue on the dorsal surface of the basal portion of each wing. The drop of
glue was applied to the anterior portion of the wing at the level of the anal
lobe and then gently spread toward the base of the wing with the tip of the
needle, taking care not to contact the wing’s articulation with the thorax.
Each control female received a drop of glue on the dorsum of her thorax.
After the glue had dried, each fly was immediately introduced into the
mating chamber (a 5-cm-diameter clear plastic petri dish with a droplet of
honey on one wall) and left there undisturbed for at least 2 h before being
tested. The entire process from removing a fly from the culture to introducing
it into the mating chamber usually lasted between 5 and 10 min. Flies were
assigned to control and experimental categories in strict alternation as they
climbed the sides of a plastic sack over the culture. The petri dishes were
labeled with nonconsecutive numbers and mixed together so that any given
male’s or the female’s treatment category was not known during behavioral
observations.
A mating trial began when a fly of the opposite sex was removed from
the culture container and introduced into the mating chamber. These flies
were also assigned to be paired with control and experimental flies in strict
alternation as they climbed the sides of a plastic sack over the culture. Each
trial lasted 1 h, unless the male had not yet attempted to mount, in which
case observations continued until he did attempt to mount or another hour
had elapsed. Observations were made without magnification, so that it was
impossible to see the glue droplets on the legs or the thorax. Not all behavioral details were noted in all pairs, so some sample sizes varied. Several
trials were run simultaneously, with new pairs being formed as the trials of
other pairs ended. Petri dishes were often moved with relation to each other
during this process, further guaranteeing that data were taken blindly with
respect to male and female treatment. A mount was judged to be successful
if the male copulated, and unsuccessful if he dismounted without copulating.
Durations of successful mounts refer to the time elapsed between initiation
of the mount and initiation of copulation. Males often mounted the female
more than once, but because subsequent mounts are usually much shorter
and did not occur in some pairs, only data from first mounts were analyzed.
Each individual was used only once.
Males sometimes lost glue droplets from their legs before, during, or
after mating trials, so each male was checked for the presence of glue soon
after the trial ended. Because the moment when the glue was lost was not
known, data were discarded if the glue was missing from both legs. Data on
durations were highly skewed, so medians as well as means (each followed
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
Function of Clasping Organs of Male Flies
16:10
Style file version Feb 08, 2000
839
by ±1 standard deviation) are reported. Unless noted otherwise, all tests of
significance employed nonparametric Mann–Whitney U-tests.
RESULTS
The behavior of pairs with control males or control females differed
between Panama and Costa Rica, despite the similarity of the conditions
under which flies were raised and observed. In Panama, 82.4% of 74 control
pairs copulated (combining data from Tables I and II), while only 34.1%
of 44 copulated in Costa Rica (χ 2 = 25.0, df = 1, P ¿ 0.001). Several
patterns indicate that this difference was probably due largely to differences
in female receptivity. Those females which did copulate did so with fewer
mounts in Panama than in Costa Rica (Table I). In 59 of 61 control pairs that
copulated in Panama, copulation occurred on the first mount (Tables I and
II); the corresponding value in Costa Rica was 7 of 15 (P ¿ 0.001, Fisher
exact test). The duration of successful mounts (beginning of mount until
intromission) was also shorter in Panama (Table I). In contrast, variables that
were presumably controlled largely if not exclusively by the male differed in
some respects but not others between the two sites. There were differences
in the distribution of giving up times by males (e.g., duration of first mount
for unsuccessful modified males in Table I), but neither the amount of time
elapsed before the male’s first mounting attempt nor the number of mounts
in pairs that did not copulate differed between the two sites (P = 0.59 and
0.62, respectively, for totals for control males). Because of these differences,
apparently due largely to differences in female receptivity, data from the two
sites are analyzed separately below.
As predicted by three of the four hypotheses, the addition of glue to
the male’s front femur resulted in a sharp reduction in the probability that
he would succeed in copulating. Of those males that mounted the female in
Panama, only 19.2% of 52 modified males copulated (including those with
glue on either one or both femora at the end of the experiment), compared
with 87.5% of 48 control males (χ 2 = 43.1, df = 1, P ¿ 0.001). Corresponding values in Costa Rica showed the same pattern: 0% of 34 modified males
compared with 34.1% of 44 control males (χ 2 = 14.2, df = 1, P < 0.001).
In addition, males in Panama that had glue on only one leg at the end of
the experiment had a greater likelihood of copulating than did males that
had glue on both legs (33% of 18 vs. 3% of 32; P = 0.006, Fisher exact
test) (two males were discarded from this analysis because only one leg was
checked for glue). Males in Costa Rica did not lose glue from their legs and,
thus, could not be analyzed in this respect. Glue on the dorsal surface of
the female’s wings also reduced the likelihood that copulation would occur
21 b1
9.5
53.5
29.5 b2
68 b1
258 c1
—
9 b2 c1
49.7 ± 68.1
101.5 ± 204.5
61.6 ± 54.6
75.7 ± 115.3
174 ± 316
690 ± 961
—
28.4 ± 63.0
Median
—
4.0 ± 3.4
2.8 ± 2.6
3.1 ± 2.8
2.4 ± 2.1
6.8 ± 4.2
1.2 ± 0.8
3.7 ± 2.7
—
4 c3
2 c1
2
1 c2
6 c3
1 c1 c2
3.5
Median
Number of mounts
Mean
0
34
15
29
10
42
42
6
N
—
14.3 ± 21.6
14.3 ± 21.6
3.8 ± 3.8
20.8 ± 25.8
15.2 ± 22.6
10.1 ± 6.6
13.7 ± 17.3
13.1 ± 16.0
14.1 ± 16.8
27.7 ± 29.7
15.8 ± 19.0
Mean
—
7
7
2.5
12.5
5
10
6
6
8.5
19
9
Median
0
34
34
14
29
43
8
41
49
42
6
48
N
Time elapsed before first mount (min)
16:10
840
U tests (an exception was the number of mounts; they could not be compared between successful and unsuccessful males, because successful
males had shorter lengths of time with a virgin female). Only values followed by the same italic letter and number differ significantly (b, P < 0.01;
c, P < 0.001).
November 8, 2002
0
34
15
29
10
42
42
6
N
pp699-joir-457250-08
a Means are given for illustrative purposes only, as the data were highly skewed. Medians in the same column were compared with Mann–Whitney
Panama
Control males
Successful
Unsuccessful
Total
Modified males
Successful
Unsuccessful
Total
Costa Rica
Control males
Successful
Unsuccessful
Total
Modified males
Successful
Unsuccessful
Total
Mean
Journal of Insect Behavior [joib]
Duration of first mount (s)
Table I. Durations and Numbers of Mounts by Males with Glue on Their Front Femora (Modified Males) and on the Thorax (Control Males)a
P1: ZCK
Style file version Feb 08, 2000
Eberhard
19
2
77
11
23.5 ± 12.7
2.9 ± 3.2
77.0 ± 94.8
117 ± 320
2
10
13
7
N
1.0
4.2 ± 3.3
1.0
3.7 ± 2.7
1
4
1
3.5
Median
Number of mounts
Mean
2
10
19
7
N
6.5 ± 6.4
30.8 ± 39.9
26.8 ± 37.4
23.4 ± 40.9
43.9 ± 38.0
30.6 ± 40.2
Mean
6.5
12.5
11
8
28.5
12
Median
2
10
12
13
7
20
N
Time elapsed before first mount (min)
November 8, 2002
U tests; no differences were significant.
pp699-joir-457250-08
a Means are given for illustrative purposes only, as the data were highly skewed. Medians in the same column were compared with Mann–Whitney
Control females
Successful
Unsuccessful
Total
Modified females
Successful
Unsuccessful
Total
Median
Duration of first mount (s)
Mean
Journal of Insect Behavior [joib]
Table II. Durations and Numbers of Mounts of Females in Panama with Glue on Their Wing Bases (Modified Females) and with Glue on the
Thorax (Control Females)a
P1: ZCK
16:10
Function of Clasping Organs of Male Flies
Style file version Feb 08, 2000
841
P1: ZCK
Journal of Insect Behavior [joib]
842
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
(16.7% of 12 modified females compared with 73.0% of 26 control females)
in Panama (χ 2 = 10.6, df = 1, P = 0.0012).
The reductions in copulatory success were not clearly due to modified
males being less able to remain mounted on females, as predicted by the
mechanical fit and male–female conflict hypotheses. In Costa Rica mount
durations of unsuccessful modified males were shorter than those of unsuccessful control males as predicted, but this was not the case in other
within-site comparisons for which sufficient data were available. The durations of both successful and unsuccessful mounts by modified males in
Panama were not significantly different from the control values (P = 0.22
and 0.30, respectively; P = 0.26 when they were combined), and the trend
was the opposite of that predicted. Taking as a reference point the median
duration of a mount by a control male that ended in success in Panama
(21 s before intromission), 26 of 42 modified males in Panama that failed
to copulate nevertheless stayed mounted for longer than 21 s. In addition,
durations of unsuccessful mounts by males in Panama that had glue on both
front legs were, if anything, longer (mean = 72.1 ± 119.7 s, median = 29 s;
N = 32) than those of males with glue on only one front leg (mean = 46.5 ±
79.1 s, median = 10 s; N = 18; P = 0.66). Similarly, the durations of unsuccessful mounts on modified females were also longer (mean = 117 ± 320 s,
median = 11 s; N = 10) than those on control females (mean = 2.9 ± 3.2 s,
median = 2 s; N = 7).
In fact, longer mounts were not positively correlated with mating success, as expected. The most extreme deviation occurred in control males
in Costa Rica, in which 22 of 29 failed mounts lasted longer than the median duration of 14 successful mounts. Data from modified females also
showed a trend in the opposite direction from that predicted by the hypotheses: unsuccessful mounts were slightly longer than successful mounts)
(P = 0.06).
A second trend also argues against the mechanical fit and male–female
conflict hypotheses. Mounts by modified males seldom ended with events
indicating that the male had been forcibly shaken off by the female. Dismounts by modified males that occurred during shaking movements by the
female (“forced dismounts”) were rare (5.8% of 121 mounts checked in
Panama for this detail) and were not less common in control males (8.9%
of 45; χ 2 = 0.49, df = 1, P > 0.4). The small difference that existed was the
opposite of that predicted by the mechanical fit and male–female conflict
hypotheses (modified males were forcibly shaken off slightly less often). In
the large majority of cases, the male climbed off during a period when the
female was not shaking. Video recordings confirmed that dismounts during
periods of shaking by the female are also unusual in Costa Rica (Baena and
Eberhard, in preparation).
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
Function of Clasping Organs of Male Flies
16:10
Style file version Feb 08, 2000
843
There was no indication that the reduced copulatory success of modified
males was due to their being less motivated to mount females. The time
elapsed before the male’s first mounting attempt did not differ significantly
between modified males and control males (Table I; P = 0.65 and 0.65 for
totals in Panama and Costa Rica). The frequency with which a male failed
to attempt to mount a female at least once during the observation period
was no greater in modified males (7 cases among 90 males) than in control
males (8 cases among 99 males). Similarly, neither the duration of the first
mount nor the delay to mount differed significantly when males paired with
modified and control females were compared (Table I; P = 0.65 and 0.64,
respectively).
The ability of modified males to remain mounted was not due to a lack
of female resistance. In fact, females in Panama were more likely to perform
especially energetic rejection movements (strong shaking, often combined
with leg pushes and ventral flexion of the abdomen) when they were mounted
by modified males (90.7% of 43 mounts by modified males checked for this
detail, 44.8% of 29 mounts by control males; χ 2 = 16.6, df = 1, P ¿ 0.001).
Observations of males mounted on females in the field as the female
moved over the surface of fresh cow dung showed that males did not have to
battle other males to remain mounted. In more than 25 encounters between a
mounted male and a solitary male, the males never struggled with each other.
The most aggressive interactions involved only a single brief strike by the
solitary male (possibly an aborted mounting attempt) that was immediately
followed by his walking away.
DISCUSSION
The consistent failure of solitary males to attempt to forcefully dislodge
mounted males they encountered in the field argues against the importance
of direct male–male combat as an explanation of the male clasping organ.
Violent male–male battles over females do occur in some sepsids (Zerbe
[1993] on Sepsis punctum; Eberhard [2001a] on A. pleuralis) but are rare or
have never been seen in several others (Eberhard [2001a] on A. ecalcarata, A.
discolor, and the closely related Palaeosepsis pusio). An additional indication
that male–male combat has not been important in the evolution of front leg
clasping organs in A. diversiformis is that it does not predict the reduction
in copulatory success that was observed in isolated pairs when the male’s
front femur or the female’s wings were modified. This hypothesis thus seems
improbable and is omitted from the following discussion.
The prediction by the remaining three hypotheses, that experimental modification of the male’s clasping organ or the female wing would
P1: ZCK
Journal of Insect Behavior [joib]
844
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
reduce the male’s copulation success, was confirmed. But the prediction
of the mechanical fit and male–female conflict hypotheses that this decrease
in copulations would be due to a reduction in the male’s ability to hold onto
the female was not confirmed. Males generally dismounted spontaneously,
rather than being forcibly dislodged by the female; forced dismounts were
not more common in modified males than control males, despite the fact that
females resisted modified males more actively; mounts by modified males
were not consistently shorter; mounts by males with both front legs modified were not shorter than those of males with only one leg modified; and
mounts of females with modified wings were not consistently shorter. The
expectation that males which held on longer would be more likely to obtain
copulations was also contradicted. Longer mounts were also not associated
with copulation success in Sepsis cynipsea (Blankenhorn et al., 2000).
The significance of the short durations of mounts by modified males in
Costa Rica remains uncertain. While their overall brevity fits the expectations of the mechanical fit and male–female conflict hypotheses, the distribution of mount durations, most of which were extremely brief, does not fit
the expected gradual elimination of males which were less able to remain
mounted. Control males from the same site that failed to copulate tended
to remain mounted for relatively long periods.
Reduced female cooperation after the male had mounted, with respect
to both his remaining mounted and his achieving intromission, probably
contributed to the lower copulation success of both modified males and males
mounted on modified females, as predicted by the stimulation hypothesis.
In Panama female resistance to modified males was more vigorous than to
control males. Active female cooperation (lifting the abdomen and flexing
the proctiger dorsally to expose the vulva) is needed for intromission to
occur in this and related species (Eberhard, 2002).
It might be that other stimuli from the male rather than those from his
front legs were responsible for reductions in female acceptance. For example,
the glue on the male’s femur or on the female’s wing may have altered the
stimuli the male received when he grasped the female’s wings and reduced
his tendency to perform courtship behavior or attempts to establish genitalic
contact, causing him to thus fail to elicit the female cooperation that leads to
copulation (Eberhard, 2002). Subtle differences in the male’s behavior after
he mounted the female, rather than in the morphology of his front femora,
might have been responsible for the increased female rejection behavior.
The possibility that manipulations of male display traits result in subtle alterations of male behavior of this sort has been ignored in many experimental
studies of female choice (Andersson, 1982; Basolo, 1990a,b; Thornhill and
Sauer, 1991; Møller, 1994). But ignoring this possiblity seems unjustified,
as some experimental modifications of male morphology do indeed cause
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
Function of Clasping Organs of Male Flies
November 8, 2002
16:10
Style file version Feb 08, 2000
845
changes in male behavior (Briceño and Eberhard, 2002). Further behavioral
observations will thus be needed to test the possibility of such changes in
A. diversiformis. Similarly, further behavioral observations of mixed-species
aggregations in the field will be needed to discriminate possible selection for
species isolation versus sexual selection.
Because the differences in the durations of mounts in this study were
generally the result of male decisions to dismount, rather than of the
male being physically forced off, the related possibility that differences were
due to male choice must also be considered. Male choice seems unlikely
to be important in this species, at least in nature, because A. diversiformis
sex ratios at mating sites are highly male-biased, and the majority of male
mounting attempts in the field do not even result in the male clamping the
female’s wings (Eberhard, 2001a). Interspecific male choice on the basis of
mechanical stimuli from the female wing also seems improbable due to the
long durations of male mounts when female wings were modified and to the
high degree of similarity between the wings of females of different species
in this genus (Eberhard, 2001a).
With respect to the stimulation hypothesis, the sharp reduction in male
copulatory success documented here is important, because if the data had
failed to show a reduction, they would have justified the rejection of this
hypothesis. In one sense, the experimental test was conservative, in that
only virgin females were used. Females need to mate at least once to reproduce, so virgins may be less selective than nonvirgins (Jackson, 1981;
Jennions and Petrie, 1997). If sperm precedence in A. diversiformis is similar
to the almost-complete last male precedence in another sepsid, S. punctum
(Schulz, 1999), then female A. diversiformis may be able virtually to erase
the effects of early matings and may, thus, be even more disposed to accept
copulation when they are virgins. Despite these reasons to expect a lack of
female selectivity, females rejected many males in the experiments described
here.
In another sense, the test of the stimulation hypothesis was crude. The
experimental modifications of male femur form were relatively large compared with the differences among closely related species (compare Figs. 1
and 2B). Similarly, experimental modification of the female wing altered
both the form of the female wing and probably also the responses from the
stress receptors in the area clasped by the male. Females of another species,
A. armata, were apparently able to distinguish, however, mounts by males of
A. diversiformis (Eberhard, 2001a), despite the fact that the male front leg
morphology of this species is especially similar to that of A. diversus (Fig. 1).
Virgin female A. armata vigorously resisted mounts by male A. diversiformis
and failed to copulate with them but then quickly copulated when they were
mounted by male A. armata (Eberhard, 2001a).
P1: ZCK
Journal of Insect Behavior [joib]
846
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
The three hypotheses can be evaluated more extensively by combining
the results reported here with independent data from a detailed comparative
study of male and female morphology in six species of sepsid (Eberhard,
2001a). The morphological data also contradicted predictions of both the
mechanical fit and the male–female conflict hypotheses for five of the six
species, because female morphology was uniform in all species except one,
instead of being species-specific as predicted by these hypotheses (the possibility of species-specific female resistance behavior was not tested, however). In addition, female wings also generally lacked structures predicted
by the mechanical fit and male–female conflict hypothesis that could impede
clamping attempts by either heterospecific or conspecific males. The morphological data confirmed one prediction of the stimulation hypothesis: in
all six species there are female sense organs in the wing near the site gripped
by the male’s leg. The present study and the morphological study are thus in
accord in providing data that are difficult to reconcile with the mechanical fit
and male–female conflict hypotheses and that conform to some predictions
of the stimulation hypothesis. They are also in accord, however, in needing
further observations of behavioral details before final conclusions can be
drawn.
One striking result of the morphological study was that there were very
detailed and precise fits between the male and the female structures in different species: female structures were relatively invariable, and males of
different species grasped the relative uniform female morphology in different ways. The present study suggests the somewhat surprising conclusion
that such precise fits are not necessary to allow a male to maintain his grasp
on the female. Even males with gross modifications of their clamping structures (Fig. 1) and males that grasped modified female wings were able to
stay mounted for long periods, despite active female resistance.
The evidence against the male–female conflict hypothesis is particularly significant in these flies, because females often gave clear behavioral
indications of resistance to male attempts to copulate. As in other sepsids
(Parker, 1972; Ward, 1983; Ward et al., 1992; Allen and Simmons, 1996), they
often resist males in nature (Eberhard, 2001a). Nevertheless, male–female
conflict is not reflected in either the designs or the functional attributes
of the male’s front legs and the female’s wings. Possibly female resistance
behavior such as shaking functions to screen males on the basis of the resulting stimulation of her wings or to communicate to the male her unwillingness to copulate (or both) (see Linley and Adams [1974], Linley and Hinds
[1975], Linley and Mooks [1975], Arnqvist [1997], Crean and Gilburn [1998],
Rodriguez [1998], and Blankenhorn et al. [2000] for studies of this and other
possible functions of female resistance behavior in insects). Blankenhorn
et al. (2000) concluded that female shaking behavior in another sepsid,
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
Function of Clasping Organs of Male Flies
November 8, 2002
16:10
Style file version Feb 08, 2000
847
S. cynipsea, did not appear to function in assessing male ability or inclination to stay mounted; they did not evaluate the possibility that females
were assessing male stimulation of their wings.
There are three other experimental studies of species-specific nongenitalic structures specialized to grasp females during or immediately preceding
copulation whose results can be compared directly with those of this study.
Experimental disabling of the clasping notal organ of male Panorpa vulgaris
scorpionflies by covering it with an adhesive resulted in changes that favored
the male–female conflict hypothesis and argued against the stimulation hypothesis (Thornhill and Sauer, 1991). Panorpa females apparently determine
the duration of copulation, and modified males, which could not grasp the female’s wing, copulated for shorter times and transferred smaller amounts of
sperm (Thornhill and Sauer, 1991). Thornhill and Sauer rejected an alternative explanation based on the stimulation hypothesis, because when females
were hungry (deprived of food for the 7–8 days of their adult life) and males
were well fed and able to feed the female abundantly with saliva masses,
the effect of modifying the male clasping organ was eliminated. Copulations
with modified males were as long as those with control males. They argued
that this trend, which was predicted by the male–female conflict hypothesis, constitutes evidence against the stimulation hypothesis because they
supposed that female stimulation criteria for copulation duration are likely
to be consistent under all conditions. This supposition is weak, however,
in light of the substantial intraspecific variation in mating criteria in many
other species (Jennions and Petrie, 1997). Their rejection of a stimulatory
function was also weak because they did not consider the possibility that
stimulation of the female by the notal organ is important in other contexts.
For example, it might affect crucial female reproductive processes such as
oviposition and remating (Eberhard, 1996). In sum, the rejection was not
completely convincing.
The results of two other studies were similar to those of this study, in that
they favored the stimulation hypothesis over the mechanical fit and male–
female conflict hypotheses. In two species of Ischnura damselflies, modification of the species-specific form of the male’s abdominal clasping organ
did not reduce the male’s ability to grasp the female, but a female grasped
by a modified male was much less likely to bend her abdomen forward to
copulate (Kreiger and Kreiger-Loibl, 1958; Loibl, 1958). Removal of the
elaborate, species-specific portion of the clasping organ on the male’s antenna in the fairy shrimp Eubranchipus serratus also reduced the likelihood
of copulation but did not reduce the male’s ability to grasp females (Belk,
1984). All three studies have the shortcoming that they did not include tests
for possible changes in male behavior resulting from modification of their
clasping organs.
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
November 8, 2002
16:10
848
Style file version Feb 08, 2000
Eberhard
ACKNOWLEDGMENTS
I thank A. Ozerov for kindly identifying flies, H. Herz for help with
German translations, Angel Aguirre for help obtaining references, and
Göran Arnqvist for criticizing an early draft. The manuscript also benefited
substantially from the comments of three anonymous reviewers. As have
many others before me, I profited from Stan Rand’s patient and thoughtful
advice. Part of this work was carried out on Barro Colorado Island, and I
thank the staff there for making it so easy and pleasant to work there. Financial support was provided by the Vicerrectorı́a de Investigación of the
Universidad de Costa Rica and the Smithsonian Tropical Research Institute.
REFERENCES
Alexander, R. D., Marshall, D., and Cooley, J. (1997). Evolutionary perspectives on insect
mating. In Choe, J., and Crespi, B. (eds.), The Evolution of Mating Systems in Insects and
Arachnids, Cambridge University Press, Cambridge, pp. 4–31.
Allen, G. R., and Simmons, L. W. (1996). Coercive mating, fluctuating asymmetry and male
mating success in the dung fly, Sepsis cynipsea. Anim. Behav. 52: 737–741.
Andersson, M. (1982). Female choice selects for extreme tail length in a widowbird. Nature 299:
818–820.
Arnqvist, G. (1997). The evolution of water strider mating systems: Causes and consequences
of sexual conflicts. In Choe, J., and Crespie, B. (eds.), The Evolution of Mating Systems in
Insects and Arachnids, Cambridge University Press, Cambridge.
Arnqvist, G., and Rowe, L. (1995). Sexual conflict and arms races between the sexes: A morphological adaptation for control of mating in a female insect. Proc. Roy. Soc. Lond. B
261: 123–127.
Basolo, A. (1990a). Female preference for male sword length in the green swordtail, Xiphophorus helleri (Pisces: Poeciliidae). Anim. Behav. 40: 332–338.
Basolo, A. (1990b). Female preference predates the evolution of the sword in the swordtail fish.
Science 250: 808–810.
Battin, T. J. (1993a). The odonate mating system, communication and sexual selection. Boll.
Zool. 60: 353–360.
Battin, T. J. (1993b). Revision of the puella group of the genus Coenagrion Kirby, 1890 (Odonata,
Zygoptera), with emphasis on morphologies contributing to reproductive isolation. Hydrobiologia 262: 13–29.
Belk, D. (1984) Antennal appendages and reproductive success in the Anostraca. J. Crust.Biol.
4: 66–71.
Blankenhorn, W., Mühlhäuser, C., Morrf, C., Reusch, T., and Reuter, M. (2000). Female choice,
female reluctance to mate and sexual selection on body size in the dung fly Sepsis cynipsea.
Ethology 106: 1–17.
Briceño, R. D., and Eberhard, W. G. (2002). Courtship in the medfly Ceratitis capitata includes
probable tactile stimulation with the male’s aristae. Entomol. Exp. Appl. (in press).
Crean, C. S., and Gilburn, A. S. (1998). Sexual selection as a side-effect of sexual conflict in the
seaweed fly, Coelopa ursina (Diptera: Coelopidae). Anim. Behav. 56: 1405–1410.
Darwin, C. (1871). The Descent of Man and Selection in Relation to Sex, Modern Library, New
York.
Duda, O. (1925). Monographie der Sepsiden (Dipt.) I. Ann. Naturhist. Mus. Wien, 39: 1–153.
Duda, O. (1926). Monographie der Sepsiden (Dipt.) II. Ann. Naturhist. Mus. Wien 40: 1–110.
P1: ZCK
Journal of Insect Behavior [joib]
pp699-joir-457250-08
Function of Clasping Organs of Male Flies
November 8, 2002
16:10
Style file version Feb 08, 2000
849
Eberhard, W. G. (1985). Sexual Selection and Animal Genitalia, Harvard University Press,
Cambridge, MA.
Eberhard, W. G. (1996). Female Control: Sexual Selection by Cryptic Female Choice, Princeton
University Press, Princeton, NJ.
Eberhard, W. G. (2000). Behavior and reproductive status of Microsepsis armillata flies away
from oviposition sites. Ann. Entomol. Soc. Am. 93: 966–971.
Eberhard, W. G. (2001a). The functional morphology of species-specific clasping structures on
the front legs of male sepsid flies. Zool. J. Linn. Soc. 133: 335–368.
Eberhard, W. G. (2001b). Courtship and multi-stage transfer of material to the female’s wings
during copulation in Microsepsis armillata (Diptera: Sepsidae). J. Kans. Entomol. Soc. 74:
23–31.
Eberhard, W. G. (2002). Female resistance or screening? Male force vs. female collaboration in
intromission in sepsid flies. Rev. Biol. Trop. 50: 485–505.
Eberhard, W. G., and Huber, B. A. (1998). Copulation and sperm transfer in Archisepsis flies
(Diptera: Sepsidae) and the evolution of their intromittent genitalia. Stud. Dipt. 5: 217–248.
Fraser, F. C. (1943). The function and comparative anatomy of the oreillets in the Odonata.
Proc. Roy. Entomol. Soc. Lond. A 18: 50–56.
Freitag, R. (1974). Selection for nongenitalic mating structure in female tiger beetles of the
genus Cicindella (Coleoptera: Cicindelidae). Can. Entomol. 106: 561–568.
Hennig, W. (1949). Sepsidae. In Linder, E. (ed.) Die Fliegen der Palaearktischen Region, D.
Schweizerbart’sche Verlagsbuchhandlung (Erwin Nagele), Stuttgart.
Holland, B., and Rice, W. R. (1998). Chase-away selection: Antagonistic seduction vs. resistance.
Evolution 52: 1–7.
Jackson, R. R. (1981). Relationship between reproductive security and intersexual selection in
a jumping spider, Phidippus johnsoni (Araneae: Salticidae). Evolution 35: 601–604.
Jennions, M. D., and Petrie, M. (1997). Variations in mate choice and mating preferences: A
review of causes and consequences. Biol. Rev. 72: 283–327.
Kreiger, F., and Kreiger-Loibl, E. (1958). Beiträge zum Verhalten von Ischnura elegans und
Ischnura pumilio (Odonata). Z. Tierpsychol. 15: 82–93.
Linley, J. R., and Adams, G. M. (1974). Sexual receptivity in Culicoides melleus (Diptera: Ceratopogonidae). Trans. Roy. Entomol. Soc. Lond. 126: 279–303.
Linley, J. R., and Hinds, M. J. (1975). Quantity of the male ejaculate influenced by female
unreceptivity in the fly Culicoides melleus. J. Insect Physiol. 21: 281–285.
Linley, J. R., and Mook, M. S. (1975). Behavioural interaction between sexually experienced Culicoides melleus (Coquillett) (Diptera Ceratopogonidae). Behaviour 54: 97–
110.
Lloyd, J. E. (1979). Mating behavior and natural selection. Fla. Entomol. 62: 17–23.
Loibl, E. (1958). Zur Ethologie un Biologie der deutschen Lestiden (Odonata). Z. Tierpsychol.
15: 54–81.
Møller, A. P. (1994). Sexual Selection and the Barn Swallow, Oxford University Press, New
York.
Parker, G. (1972). Reproductive behaviour of Sepsis cynipsea (L.) (Diptera: Sepsidae). I. Preliminary analysis of the reproductive strategy and its associated behaviour patterns. Behaviour
41: 172–206.
Pont, A. (1979). Sepsidae. Diptera Cyclorrhapha Acalyptrata. Handbook for the Identification
of British Insects, Vol. 10, 5(c), pp. 1–35.
Robertson, G. G., and Paterson, H. E. H. (1982). Mate recognition and mechanical isolation in
Enallagma damselflies (Odonata: Coenagrionidae). Evolution 36: 243–250.
Rodriguez, R. L. (1998). Possible female choice during copulation in Ozophora baranowskii
(Heteroptera: Lygaeidae): Female behavior, multiple copulations, and sperm transfer. J.
Insect Behav. 11: 725–741.
Shapiro, A. M., and Porter, A. H. (1989). The lock-and-key hypothesis: Evolutionary and biosystematic interpretation of insect genitalia. Annu. Rev. Entomol. 34: 231–245.
Schulz, K. (1999). The Evolution of Mating Systems in Black Scavenger Flies (Diptera: Sepsidae),
Ph.D. thesis, University of Arizona, Tucson.
P1: ZCK
Journal of Insect Behavior [joib]
850
pp699-joir-457250-08
November 8, 2002
16:10
Style file version Feb 08, 2000
Eberhard
Šulc, K. (1928). Biologicky vyznam vyzbroje prednich samcich nohou u sepsid (Muscidae). Biol.
Spisy Vysoke Skoly Zverolekarske Brno. 7: 181–194.
Thornhill, R., and Sauer, K. P. (1991). The notal organ of the scorpionfly (Panorpa vulgaris):
An adaptation to coerce mating duration. Behav. Ecol. 2: 156–164.
Toro, H., and de la Hoz, E. (1976). Factores mecánicos en la aislación reproductive de Apoidea
(Hymenoptera). Rev. Soc. Entomol. Arg. 35: 193–202.
Ward, P. I. (1983). The effects of size on the mating behaviour of the dung fly Sepsis cynipsea.
Behav. Ecol. Sociobiol. 13: 75–80.
Ward, P. I., Hemmi, H., and Roösli, T. (1992). Sexual conflict in the dung fly Sepsis cynipsea.
Funct. Ecol. 6: 649–653.
Zerbe, F. (1993). Innerartliche Grössenvariabilität und Paarungsverhalten bei Sepsis punctum (Fabricius, 1794) [Diptera, Sepsidae], Diplomarbeit, Julius-Maximilians-Universität
Würzburg, Würzburg.