Behav Ecol Sociobiol (1997) 41: 335±341 Ó Springer-Verlag 1997

Douglas J. Emlen

Alternative reproductive tactics and male-dimorphism in the horned

beetle Onthophagus acuminatus (Coleoptera: Scarabaeidae)

Received: 12 October 1996 / Accepted after revision: 8 August 1997

Abstract Adult dung beetles (Onthophagus acuminatus) Key words Alternative reproductive behavior á
exhibit continuous variation in body size resulting from Male dimorphism á Male competition á
di€erential nutritional conditions experienced during Horned beetles á Onthophagus
larval development. Males of this species have a pair of
horns that protrude from the base of the head, and the
lengths of these horns are bimodally distributed in nat- Introduction
ural populations. Males growing larger than a threshold
body size develop long horns, and males that do not Males in many animal species show variation in mor-
achieve this size grow only rudimentary horns or no phology which is associated with di€erences in behavior
horns at all. Previous studies of other horned beetle (e.g., Austad 1984; Dominey 1984; Travis 1994). For
species have shown that horned and hornless males often example, large and small males frequently utilize strik-
have di€erent types of reproductive behavior. Here I ingly di€erent behaviors to encounter and mate with
describe the mating behaviors of the two male morphs of females (Dominey 1980; Rubenstein 1980, 1987; How-
O. acuminatus during encounters with females. Females ard 1984; Gross 1985; Kodric-Brown 1986; Arak 1988;
excavate tunnels beneath dung, where they feed, mate Reynolds et al. 1993). Occasionally, variation in male
and provision eggs. Large, horned males were found to morphology is dimorphic, and two or more distinct male
guard entrances to tunnels containing females. These forms co-occur in populations with intermediate forms
males fought with all other males that attempted to enter scarce or lacking (Shuster 1987; Ryan and Causey 1989;
these tunnels. In contrast, small, hornless males en- Zimmerer and Kallman 1989; Danforth 1991). Species
countered females by sneaking into tunnels guarded by exhibiting dimorphic variation can be especially reveal-
other males. In many instances, this was accomplished ing to the investigator because they generally implicate
by digging new tunnels that intercepted the guarded morphological specializations for alternative behavioral
tunnels below ground. Side-tunneling behavior allowed or ecological situations (e.g., ®ghting and dispersing in
sneaking males to enter tunnels beneath the guarding thrips, Crespi 1988; and beetles, Eberhard 1982; Siva-
male, and mate with females undetected. Both overall Jothy 1987; soft and hard seed diets in ®nches, Smith
body size and relative horn length signi®cantly a€ected 1993; and high and low levels of predation in barnacles,
the outcome of ®ghts over tunnel ownership. These re- Lively 1986a,b; rotifers, Gilbert and Stemberger 1984;
sults suggest that alternative reproductive tactics may and Daphnia, Grant and Bayly 1981; Black and Dodson
favor divergence in male horn morphology, with long 1990; Spitze 1992).
horns favored in males large enough to guard tunnels, A classic example of morphological dimorphism in-
and hornlessness favored in smaller males that adopt the volves the horns of some male beetles (Bates 1863;
``sneaking'' behavioral alternative. Darwin 1871; Wallace 1878; Fabre 1899; Inukai 1924;
Huxley 1931; Beebe 1944; Arrow 1951; Clark 1977;
Eberhard 1982; Cook 1987; Siva-Jothy 1987). Not only
are horned beetles sexually dimorphic (females in most
D.J. Emlen1 species do not have horns), but many species exhibit
Department of Ecology and Evolutionary Biology,
Princeton University, Princeton, NJ 08544-1003, USA dimorphic variation within males (Goldsmith 1987;
Cook 1987, 1990; Eberhard and Gutierrez 1991; Emlen
Present address:
1
Division of Biological Sciences, The University of Montana,
1994a; Rasmussen 1994; Kawano 1995). In these species
Missoula, MT 59812-1002, USA large males possess fully developed horns, while small
Fax: +406 243-4184; e-mail: demlen@selway.umt.edu males have only rudimentary horns, or no horns.
336

Beetle horns have been shown to function in intra- constructed glass observation chambers. These chambers were
sexual combat over access to females (Palmer 1978; similar to ``ant farms'' (see Klemperer 1981; Hunter et al. 1991;
Emlen 1993 for examples) in that they consisted of two parallel
Eberhard 1979, 1987; Brown and Bartalon 1986; Gold- panes of glass separated by 5 mm and ®lled with soil. Wide boxes
smith 1987; Siva-Jothy 1987; Conner 1988), and the made from clear plexiglass ®t over the tops of these farms, allowing
existence of a ``hornless'' class of smaller males suggests beetles to walk freely on the ``soil'' surface. Pieces of howler
that these individuals may employ an alternative, less monkey dung were placed over the soil, and beetles of known sizes
and sexes introduced. In all cases, beetles tunneled readily into the
aggressive behavioral tactic. In the species examined soil between the panes of glass, and all behavior occurring both
thus far, hornless males do employ reproductive be- above and below the surface could be viewed clearly. Because
haviors that di€er from that of horned males. Hornless tunneling behavior naturally occurs in darkness, all observations
males dispense with courtship and transfer spermato- were conducted using red-®ltered light (beetles cannot detect red
pores more rapidly (Cook 1990), search for females in light; Crowson 1981).
Behavioral observations included ad libitum, focal-animal, and
sub-optimal locations (``satellite'' tactics; Eberhard scan samples (Altmann 1974) until beetles either became dormant
1982; Goldsmith 1987; Siva-Jothy 1987), and sneak or tried to leave the arena (often 2±4 days later). At this time
around ®ghting males (Rasmussen 1994). However, why tunnels were traced onto clear acetate overlays. All beetles were
these behaviors are associated with hornlessness remains wild-collected from the Barro Colorado Island Nature Monument,
Panama (where this study was conducted), and no individuals were
unclear, and the behavioral repertoires of the vast ma- used more than once. Beetle behavior was observed under ap-
jority of horned beetle species have yet to be explored. proximately natural conditions, as well as in four sets of experi-
Onthophagus acuminatus Har. (Coleoptera: Scarab- mental situations.
aeidae) is a horned beetle common in lowland tropical
forests of Central America, where it feeds on dung from
General observations
howler monkeys. Large males of this species possess a
pair of frontal horns, while horns are greatly reduced in To approximate natural conditions, I constructed boxes consisting
small males (Emlen 1994a, 1997). Variation in adult of several adjacent glass chambers (methods described in Emlen
body size in O. acuminatus was found to be predomi- 1993). These boxes were buried in the forest so that the tops of the
chambers were ¯ush with the forest ¯oor. When ®lled with soil,
nantly determined by environmental factors related to boxes provided a smooth, dirt surface which could later be disas-
larval nutrition (Emlen 1994a, 1996). Male horns are sembled into separate glass-walled ``slices''. Boxes were buried
facultatively expressed, and depend on the body size beneath trees where the howler monkeys slept. In the mornings,
attained by an individual: larvae growing larger than a when dung fell to the forest ¯oor, pieces were re-located on to the
genetically-determined threshold size metamorphose surfaces of the boxes. Beetles were allowed to colonize dung and
tunnel undisturbed. Boxes were then returned to the laboratory,
into adults with long horns, and larvae not reaching this and all below ground behaviors were observed under red light.
size metamorphose into adults lacking or with very short Twenty-two of these boxes were monitored to characterize the
horns (Emlen 1994a,1996). natural repertoire of tunneling, feeding and mating behavior. From
The facultative adoption of either a horned or a these observations, two male reproductive tactics were identi®ed,
labeled ``guarding'' and ``sneaking'' (Fig. 1).
hornless morphology by males strongly suggests that
these males may be morphologically specialized for
alternative behaviors or ecological situations. Here I Experiments 1 and 2: reproductive behavior of horned
determine whether horned and hornless male O. acumi- and hornless males
natus di€er in reproductive behavior. I present results To compare the reproductive behavior of horned and hornless
from a series of experiments comparing the tactics em- males, I observed their methods of mate-acquisition both without,
ployed by these two classes of males to encounter and and with competition from a rival male. In the ®rst experiment, one
mate with females. In addition, as a ®rst step towards male and one female were placed in each of 12 observation
identifying the functional signi®cance (if any) of horned chambers. Seven of these males were hornless, ®ve were horned,
and females were selected at random. Beetles were observed for a
and hornless male morphologies, I measure whether minimum of three half-hour intervals (maximum of six intervals),
these males di€er in their ability to perform one of these during which time all behaviors were recorded.
behaviors. Taken together, these results identify an In the second experiment, two males (one horned and one
``alternative'' male mating behavior that may be func- hornless) were placed together in each observation chamber with a
single female, and behaviors of all individuals were monitored as
tionally associated with a hornless beetle morphology, above. This second experiment tested for behavioral di€erences
and suggest that alternative male reproductive tactics arising as a result of direct competition for access to females. Be-
may contribute to the maintenance of dimorphic varia- cause dung always had large numbers of O. acuminatus (Emlen
tion in male horns. 1994b), and because horned and hornless males occurred in ap-
proximately equal frequencies on Barro Colorado Island (Emlen
1994a,b, 1997), this experiment accurately re¯ected natural condi-
tions experienced by males.
Methods
Experiments 3 and 4: tunnel-guarding performance
Observing underground behavior and male morphology

Females of O. acuminatus excavate vertical tunnels directly beneath To examine whether males in this species di€er in their ability to
dung that are used for feeding, mating, and provisioning eggs perform the ``guarding'' tactic, I measured whether natural varia-
(Emlen 1994b). In order to clearly observe beetles in tunnels I tion in two aspects of male morphology (body size and horn length)
 337

tunnel with the female when the other male exited the arena. Fight
outcomes were compared for larger and smaller males using a chi-
square test.
In a fourth experiment, I measured the e€ects of natural vari-
ation in horn length on male guarding performance. To control for
possible confounding e€ects of variation in body size, I chose 26
pairs of males di€ering in horn length but not in body size, and
compared the number of ®ghts won by the longer- and shorter-
horned contestants. As in experiment 3, the predicted loser (i.e. the
male with the shorter horn length) was placed into tunnels ®rst, and
the second male added less than one minute later. Staged contests
encompassed an extensive range of horn length di€erences: the
di€erence in horn length between competitors ranged from
0.05 mm (less than 5%) to 0.45 mm (over 50%). I therefore mea-
sured the e€ect of horns on male guarding performance by re-
gressing the proportion of ®ghts won by the longer-horned male
against the di€erence in horn length between contestants (propor-
tions were arc-sine transformed prior to analysis; Zar 1984). A
signi®cant regression would indicate that the e€ect of horns on
male guarding performance increases with increasing di€erence in
horn length between opponents. Di€erences in horn length were
measured in both absolute units and as proportions of total length.
The two methods yielded similar results, so only the absolute
measures are presented.

Results
Fig. 1 Biology of Onthophagus acuminatus illustrating the two
reproductive tactics used by males. Horned males ®ght to guard Males of O. acuminatus employed two very di€erent
possession of tunnels containing females (``guarding''), and males able
to successfully guard tunnels mate repeatedly with occupant females. tactics to encounter and mate with females (Fig. 1): they
Hornless males sneak into tunnels by digging horizontal side-tunnels either attempted to monopolize access to a female by
that intercept primary tunnels underground (``sneaking''). In this way guarding the entrance to her tunnel (guarding), or they
hornless males sometimes enter main tunnels beneath the guarding attempted to bypass guarding males (sneaking).
male and mate with the female undetected
Guarding behavior entailed remaining inside a tunnel
with a female, and ®ghting intruding males over pos-
a€ected male guarding ability. To identify an appropriate measure session of the tunnel. Guarding males blocked tunnel
for body size, I used a principal components analysis including ®ve entrances and periodically ``patrolled'' the length of the
measured traits (elytra width, elytra length, prothorax width, head tunnel. Rival males could gain possession of a tunnel
width, body weight) and 368 male beetles. All of these measures only by forcibly evicting the resident male, and both
were found to be strongly positively correlated (Emlen 1994b), and
the ®rst principal component (considered by some authors to be a ®ghts and turnovers were frequent. Fights over tunnel
measure of overall body size; Bookstein 1989; Klingenberg 1996) occupancy entailed repeated butting, wrestling and
explained 90% of the overall variance. The width of the prothorax pushing of opponents, and ®ghts continued until one of
was the trait most strongly correlated with the ®rst principal the contestants left the tunnel.
component (r = 0.981; Emlen 1994b), and was used subsequently
as a linear approximation of overall body size. Horn length was Sneaking involved bypassing the guarding male. The
measured as the linear distance between the base and the tip of the primary method of sneaking into tunnels was to dig side-
horn. Both prothorax width and horn length could be measured to tunnels that intercepted guarded tunnels below ground
the nearest 0.05 mm. (Fig. 1). New tunnels were dug immediately adjacent
In the third experiment, I measured the consequences of vari-
ation in body size on the ability of males to guard tunnels by
(<2 cm) to a guarded tunnel. These tunnels then turned
staging contests between individuals with di€erent body sizes horizontally 1±2 cm below ground, and often inter-
(n = 10 contests). However, body size and horn length are posi- cepted primary tunnels beneath the guarding male (16/
tively correlated in natural populations (Emlen 1994a, 1997). To 24 side-tunnels). In this fashion, sneaking males some-
control for possible confounding e€ects of horn length, I compared times bypassed the guarding male and mated with fe-
guarding abilities only between pairs of males with the same horn
length but di€erent body sizes (i.e., by pairing those males with the males undetected (observed in four instances). Side
largest and smallest body sizes for a given horn length). tunnels were not artifacts from observing beetles in two-
The general observations indicated that ®ghts only occur inside dimensional observation chambers: side-tunnels were
tunnels, and that intruding males needed to maneuver past the always present in tunnel castes formed by injecting sili-
resident beetle before they could push the resident beetle out of the
tunnel from below (62 observed ®ghts). Therefore, all staged con- cone latex into natural tunnels in the forest ¯oor (Emlen
tests were performed in natural tunnels made by females, and to 1994b). In addition, latex castes revealed that individual
provide a conservative estimate of the e€ects of male morphology, side-tunnels often intercepted several di€erent primary
the predicted loser was always placed in the tunnel ®rst. In each tunnels, suggesting that sneaking males may visit mul-
contest, the male with the smaller size was placed in the tunnel ®rst,
and the second contestant added immediately (<1 min) after. All
tiple guarded burrows (Emlen 1994b).
males were wild-caught, and no male was ever used in more than A second method of sneaking into tunnels involved
one contest. Winners were de®ned as the male remaining in the entering guarded tunnels directly. Occasionally, if the
338

guarding male was away from the tunnel (e.g., in the table with Yates' correction for continuity: v2 = 14.22,
dung) or o€ to the side of the tunnel, a hornless male P = 0.000; Fig. 2b).
managed to enter tunnels unchallenged (8/27 hornless Results from experiment 2 were consistent with the
male entrance attempts). Once inside, these males gen- general observations described above. Horned males
erally went directly to the female and mated (observed always guarded tunnels. When horned males were
for 5 of the 8 entrances). evicted from a tunnel, they abandoned the vicinity of
Sneaking behavior was characterized by brief that tunnel, and generally attempted to gain possession
(<10 min) tenure of males inside tunnels. Even when of another tunnel (30/31 evictions). In contrast, when
sneaking males succeeded in entering a tunnel unde- hornless males were evicted from a tunnel, they generally
tected, they exited tunnels immediately after mating. remained nearby before attempting to sneak back into
Most of the time, however, sneaking males were caught guarded tunnels (27/34 evictions). This tendency to leave
by the guarding male and evicted before encountering a tunnel after a ®ght was one of the most characteristic
the female (19/27 entrance attempts at the top of the di€erences in behavior between horned and hornless
burrow; 30/43 total observed entries, also including use males (2 ´ 2 contingency table with Yates' correction for
of side-tunnels). Sneaking males then returned to their continuity: v2 = 35.339, P = 0.001).
original side-tunnels, or made new side-tunnels, and re- With one exception, all males using side-tunneling
mained inactive until several hours later when they tried behavior were hornless (23/24 total observations [in-
to enter the primary tunnel again. cluding general observations and experiments], binomial
probability: P < 0.001), and all nine hornless males
from experiment 2 (males competing for access to fe-
Experiments 1 and 2: reproductive behavior males) used this tactic.
of horned and hornless males Matings generally occurred inside tunnels (55/56
observed matings). Females never rejected a mating at-
When males were placed alone with females (experiment tempt from any male, suggesting that active, precopu-
1), both horned and hornless males remained inside latory mate choice may not be important in this species.
tunnels with females and mated repeatedly (Fig. 2a). Courtship consisted of the male ``drumming'' his fore-
There were no ®ghts in this experiment because there legs over the back and sides of the female, and copula-
were no rival males. However, remaining inside tunnels tion durations were brief (x ‹ SD = 123 ‹ 35 s,
was still considered ``guarding'' because sneaking males n = 21). There was no di€erence in copulation duration
exit tunnels immediately after mating, even when they between horned and hornless males (Mann-Whitney
do not encounter a guarding male. test, U14,7 = 44, P = 0.71), suggesting that females do
When males competed for access to females (experi- not discriminate among horned and hornless males by
ment 2), only the horned males successfully defended the preferentially terminating copulations. Female mating
tunnels. In all nine trials where both a horned and a behavior, and this conspicuous absence of direct female
hornless male competed for a single female, the horned choice, will be discussed more fully in a later paper.
male guarded the tunnel entrance and evicted the In experiment 2, which had equal frequencies of
smaller, hornless male in the process (2 ´ 2 contingency horned and hornless males (one of each per observation
chamber), 75% of the observed matings were by the
horned male (9/12). However, this need not re¯ect the
relative reproductive success of horned and hornless
males under natural conditions, because morph fre-
quencies may vary, and because sneaking males may
sometimes intercept several di€erent primary tunnels
(which was not possible in this experiment).

Experiments 3 and 4: guarding performance

and male morphology

Body size signi®cantly a€ected male performance at

guarding tunnels (experiment 3; Fig. 3). These contests
controlled for possible confounding e€ects of male horn
Fig. 2 Reproductive behaviors used by horned and hornless males length by using pairs of males that di€ered in body size
to gain access to females. a When males were alone with females, all but not horn length. In males with the same horn
males remained inside primary tunnels with females (``guarding''). b lengths, body size signi®cantly a€ected the outcome of
However, when males competed for access to a female, only the ®ghts, with the larger male winning 9 out of 10 staged
horned males were able to successfully guard tunnels. In these
situations the hornless males sneaked into guarded tunnels by
contests (chi-square test: v2 = 6.40, P = 0.011).
digging new tunnels that intercepted the guarded tunnels below Natural variation in male horn length also a€ected
ground guarding ability. Speci®cally, long horns improved male
 339

Onthophagus taurus: Fabre 1899; Moczek 1996;

D.J. Emlen, unpublished work; Phanaeus di€ormis:
Rasmussen 1994; Typhoeus typhoeus: Palmer 1978).
Smaller, hornless males also remained inside tunnels
with females when given the opportunity (i.e., when
there was no competition from rival males; Fig. 2a).
However, whenever males competed for access to fe-
males (the typical situation in natural populations;
Emlen 1994b), hornless males always adopted a non-
aggressive alternative tactic that was never employed by
Fig. 3 E€ect of male body size on guarding performance. Results are
the larger, horned males (Fig. 2b). Hornless males
from 10 staged contests between males with the same horn length but sneaked into guarded tunnels either by digging side-
di€erent body sizes tunnels that intercepted guarded tunnels below ground
(Fig. 1), or by sliding past guarding males at the tunnel
entrance. Sliding past guarding males was similar to
sneaking behavior described for hornless males in other
horned beetle species (e.g., Rasmussen 1994; Moczek
1996; A.P. Moczek and D.J. Emlen, unpublished work),
but this is the ®rst characterization of a sneaking tactic
involving hornless males digging their own tunnels, and
intercepting guarded burrows beneath the soil surface. It
was impossible to estimate the pro®tability of sneaking
into tunnels, as compared with guarding, using the
present methods. Below-ground behaviors could only be
observed inside glass-walled observation chambers.
Fig. 4 E€ect of male horn length on guarding performance. Results
These chambers restricted the directions of side-tunnels,
are from 26 staged contests between same-sized males with di€erent and may have in¯uenced the likelihood of a sneaking
horn lengths male intercepting a guarded tunnel. However, castes
from natural tunnels indicated that side-tunneling males
can intercept multiple guarded tunnels with a single side-
guarding performance when the confounding e€ects of tunnel, suggesting that sneaking males may repeatedly
body size were removed, and when the di€erence in horn visit numerous tunnels. More direct measures are needed
lengths between contestants was large (experiment 4; (e.g., using genetic markers) to adequately assess the
Fig. 4). The probability of a male winning a ®ght was relative fertilization success of guarding and sneaking
signi®cantly and positively a€ected by the di€erence in males.
horn length between contestants (simple regression, Females appeared to mate with hornless males just as
r = 0.696, F = 6.573, P = 0.037). Males with rela- readily as with horned males. No female ever rejected
tively long horns were more likely to win contests over matings with any male, and copulation durations were
tunnel ownership than males with relatively short horns. not di€erent for horned and hornless males. These re-
sults are consistent with a mating system characterized
by intense inter-male competition over access to females,
and where smaller, competitively inferior males adopt a
Discussion non-aggressive behavioral alternative to encounter fe-
males.
Males in natural populations of the beetle O. acuminatus Do guarding and sneaking tactics favor horned and
occur in two forms. Populations on Barro Colorado hornless male morphologies, respectively? As a ®rst step
Island, Panama, contain approximately equal numbers towards addressing this question, this study measured
of males with a pair of fully developed horns, and males the e€ect of natural variation in male horn morphology
with only rudimentary horns or no horns at all (Emlen on male performance at guarding tunnels. In staged
1994a, 1997). Here I show that horned and hornless contests that controlled for confounding e€ects of vari-
males employ two very di€erent behavioral tactics to ation in body size, males with relatively longer horns
encounter and mate with females. Large, horned males won signi®cantly more ®ghts over tunnel ownership than
guard entrances to tunnels containing females (Fig. 1). same-sized males with relatively shorter horns. This
Guarding a tunnel enabled a male to mate repeatedly suggests that for males large enough to guard tunnels,
with the female as she provisioned burrows with dung long horns will be bene®cial.
and oviposited. Guarding frequently involved ®ghting But why should smaller, sneaking males be hornless?
intruding males over tunnel occupancy, and was similar One possibility is that males without horns sneak more
to behavior described for other horned dung beetles e€ectively than males with horns. Horns scrape against
living in burrows (Onthophagus binodis: Cook 1990; tunnel walls as beetles run below ground (Emlen 1994b;
340

Moczek 1996). Sneaking males depend on rapidly en- M.J. West-Eberhard, and several anonymous reviewers.
tering and exiting tunnels for their reproductive success, H.W. Ambrose III and D. Peoples aided with the design and
construction of the observation chambers, and W. Eberhard,
and horns may hinder their sneaking performance. Al- B. Gill, H. Howden and K. Milton patiently taught a novice (me)
though this remains to be tested for O. acuminatus, ex- about the biology of beetles and monkeys. This project bene®ted by
periments conducted on the related species Onthophagus insightful conversations and ideas from H.W. Ambrose III,
taurus demonstrated that for same-sized males, males K.L. Bright, W. Eberhard, K. Harms, E. Leigh, Jr., E. Perry,
A.S. Rand and M.J. West-Eberhard. I thank the Smithsonian
with short horns moved signi®cantly faster inside tun- Tropical Research Institute, and in particular, the sta€ and biolo-
nels than males with longer horns (Moczek 1996; A.P. gists of Barro Colorado Island, for generously providing facilities,
Moczek and D.J. Emlen, unpublished work). laboratory space and housing. Funding was provided by a Smith-
A second possibility is that horns are costly to pro- sonian Institution Predoctoral Fellowship, a National Science
duce. Relatively small males were not successful at Foundation Doctoral Dissertation Improvement Grant (DEB-
9224088) and a Princeton University Graduate Fellowship.
guarding tunnels, and presumably derive little bene®t
from possessing horns. If horns are expensive to pro-
duce, then this might favor males able to facultatively
omit horn growth whenever developmental conditions
preclude the attainment of large body sizes. At least two
References
costs of beetle horns have already been established.
First, production of horns signi®cantly extends the de- Altmann J (1974) Observational study of behavior: sampling
methods. Behaviour 49:227±267
velopment time of O. taurus males, and results in in- Arak A (1988) Callers and satellites in the natterjack toad: evolu-
creased larval mortality from soil-dwelling nematodes tionarily stable decision rules. Anim Behav 36:416±432
(Hunt and Simmons in press). Second, allocation of Arrow GH (1951) Horned beetles. Junk, The Hague
developmental resources to horns in both O. acuminatus Austad SN (1984) A classi®cation of alternative reproductive be-
haviors and methods for ®eld-testing ESS models. Am Zool
and O. taurus results in reduced allocation to other 24:309±319
morphological traits, speci®cally eyes: males with rela- Bates HW (1863) The naturalist on the river Amazon (re-published
tively longer horns developed with signi®cantly smaller 1989). Penguin, Harmondsworth
eyes than males with relatively shorter horns (D.J. Em- Beebe W (1944) The function of secondary sexual characters in two
species of Dynastinae (Coleoptera). Zoologica 29:53±58
len and H.F. Nijhout, unpublished work). Such costs to Black AR, Dodson SI (1990) Demographic costs of Chaoborus-
horn production suggest that sneaking males might induced phenotypic plasticity in Daphnia pulex. Oecologia
bene®t by not developing horns. 83:117±122
One prerequisite for the maintenance of dimorphism Bookstein FL (1989) ``Size and shape'': a comment on semantics.
is that organisms experience a ®tness tradeo€ across Syst Zool 38:173±180
Brown L, Bartalon J (1986) Behavioral correlates of male mor-
environments (Levins 1968; West-Eberhard 1979, 1992; phology in a horned beetle. Am Nat 127:565±570
Stearns 1982; Lively 1986b). If animals encounter sev- Clark JT (1977) Aspects of variation in the stag beetle Luc-
eral discrete environment types, or ecological or behav- anus cervus (L.) (Coleoptera: Lucanidae). Syst Entomol 2:9±
ioral situations, and these di€erent environments favor 16
Conner J (1988) Field measurements of natural and sexual selection
di€erent morphologies, then distinct morphological al- in the fungus beetle, Bolitotherus cornutus. Evolution 42:736±
ternatives can evolve within a single population ± each 749
specialized for one of the di€erent environments. Such Cook D (1987) Sexual selection in dung beetles. I. A multivariate
®tness tradeo€s have been demonstrated for several di- study of morphologcal variation in two species of Onthophagus.
Aust J Zool 35:123±132
morphic species. For example, soft and hard seed diets Cook D (1990) Di€erences in courtship, mating and postcopula-
have favored two divergent bill morphologies within tory behaviour between male morphs of the dung beetle On-
populations of African ®nches (Smith 1993), and high thophagus binodis Thunberg (Coleoptera: Scarabaeidae). Anim
and low levels of predation have favored alternative shell Behav 40:428±436
morphologies in barnacles (Lively 1986a), and spined Crespi BJ (1988) Adaptation, compromise and constraint: the de-
velopment, morphometrics, and behavioral basis of a ®ghter-
and spineless morphologies in rotifers (Gilbert and ¯ier polymorphism in male Hoplothrips karnyi (Insecta: Thy-
Stemberger 1984) and Daphnia (Grant and Bayly 1981; sanoptera). Behav Ecol Sociobiol 23:93±104
Black and Dodson 1990; Spitze 1992). It is possible that Crowson RA (1981) The biology of the Coleoptera. Academic
the alternative reproductive tactics characterized in this Press, London
Danforth BN (1991) The morphology and behavior of dimorphic
study produce a similar situation in O. acuminatus. If males in Perdita portalis (Hymenoptera: Andrenidae). Behav
guarding and sneaking behaviors favor horned and Ecol Sociobiol 29:235±247
hornless male morphologies, respectively, then the re- Darwin C (1871) The descent of man, and selection in relation to
productive behavior of males may have contributed to sex. Murray, London
Dominey WJ (1980) Female mimicry in male bluegill sun®sh ± a
the evolution of male horn length dimorphism in this genetic polymorphism? Nature 284:546±548
species. Dominey WJ (1984) Alternative mating tactics and evolutionarily
stable strategies. Am Zool 24:385±396
Acknowledgements For their help with versions of this manuscript, Eberhard WG (1979) The function of horns in Podischnus agenor
I thank K.L. Bright, W. Eberhard, S. Emlen, B. Gill, P. Grant, (Dynastinae) and other beetles. In: Blum MS, Blum NA (eds)
R. Grant, M. Hughes, H. Horn, H. Howden, C. Martinez del Rio, Sexual selection and reproductive competition in insects.
T. Mendelson, M. Roberts, D. Rubenstein, L.W. Simmons, Academic Press, New York
 341

Eberhard WG (1982) Beetle horn dimorphism: making the best of a Lively CM (1986b) Canalization versus developmental conversion
bad lot. Am Nat 119:420±26 in a spatially variable environment. Am Nat 128:561±572
Eberhard WG (1987) Use of horns in ®ghts by the dimorphic males Moczek AP (1996) Male dimorphism in the scarab beetle On-
of Ageopsis nigricollis (Coleoptera Scarabeidae, Dynastinae). J thophagus taurus Schreiber, 1759 (Scarabaeidae: Onthophagini):
Kans Entomol Soc. 60:504±509 evolution and plasticity in a variable environment. Masters
Eberhard WG, Gutierrez E (1991) Male dimorphisms in beetles dissertation, University of Wurzburg
and earwigs and the question of developmental constraints. Palmer TJ (1978) A horned beetle which ®ghts. Nature 274:583±584
Evolution 45:18±28 Rasmussen JL (1994) The in¯uence of horn and body size on the
Emlen DJ (1993) A new method for observing underground and reproductive behavior of the horned rainbow scarab beetle
soil surface behaviors. Coleopt Bull 47:243±244 Phanaeus di€ormis (Coleoptera: Scarabaeidae). J Insect Behav
Emlen DJ (1994a) Environmental control of horn length dimor- 7:67±82
phism in the beetle Onthophagus acuminatus (Coleoptera: Reynolds JD, Gross MR, Coombs MJ (1993) Environmental
Scarabaeidae). Proc R Soc Lond B 256:131±136 conditions and male morphology determine alternative mat-
Emlen DJ (1994b) Evolution of male horn length dimorphism in ing behaviour in Trinidadian guppies. Anim Behav 45:145±
the dung beetle Onthophagus acuminatus (Coleoptera: Scarab- 152
aeidae). Doctoral dissertation, Princeton University Rubenstein DI (1980) On the evolution of alternative mating
Emlen DJ (1996) Arti®cial selection for horn length-body size al- strategies. In: Stadden JER (ed) Limits to action: the allocation
lometry in the horned beetle Onthophagus acuminatus (Cole- of individual behavior. Academic Press, New York, pp 65±100
optera: Scarabaeidae). Evolution 50:1219±1230 Rubenstein DI (1987) Alternative reproductive tactics in the spider
Emlen DJ (1997) Diet alters male horn allometry in the beetle Meta segmentata. Behav Ecol Sociobiol 20:229±237
Onthophagus acuminatus (Coleoptera: Scarabaeidae). Proc R Ryan MJ, Causey BA (1989) ``Alternative'' mating behavior in
Soc Lond B 264:567±574 the swordtails Xiphophorus nigrensis and Xiphophorus pyg-
Fabre JH (1899) Souvenirs entomologiques (translated excerpts). maeus (Pisces: Poeciliidae). Behav Ecol Sociobiol 24:341±
In: Mattos AT (1922) More beetles. Hodder and Stoughton, 348
London Shuster SM (1987) Alternative reproductive behaviors: three dis-
Gilbert JJ, Stemberger RS (1984) Asplancha-induced polymor- crete male morphs in Paracercis sculpta, an intertidal isopod
phism in the rotifer Keratella slacki. Limnol Oceanogr 26:1309± from the northern gulf of California. J Crustacean Biol 7:318±
1316 327
Goldsmith SK (1987) The mating system and alternative repro- Siva-Jothy MT (1987) Mate securing tactics and the cost of ®ghting
ductive behaviors of Dendrobias mandibularis (Coleoptera: in the Japanese horned beetle Allomyrina dichotoma L. (Scar-
Cerambycidae). Behav Ecol Sociobiol 20:111±115 abaeidae). J Ethol 5:165±172
Grant JWG, Bayly (1981) Predator induction of crests in morphs of Smith TB (1993) Disruptive selection and the genetic basis of bill
the Daphnia carinata King complex. Limnol Oceanogr 26:201± size polymorphism in the African ®nch Pyrenestes. Nature
218 363:616±620
Gross MR (1985) Disruptive selection for alternative life histories Spitze K (1992) Predator-mediated plasticity of prey life history
in salmon. Nature 313:47±48 and morphology: Chaoborus americanus predation on Daphnia
Howard RD (1984) Alternative mating behaviors of young male pulex. Am Nat 139:229±247
bullfrogs. Am Zool 24:397±406 Stearns SC (1982) The role of development in the evolution of life
Hunt J, Simmons LW (in press) Patterns of ¯uctuating asymmetry histories. In: Bonner JT (ed) Evolution and development.
in beetle horns: an experimental examination of the honest Dahlem Konferenzen. Springer-Verlag, Berlin, pp 237±258
signalling hypothesis. Behav Ecol Sociobiol Travis J (1994) Size-dependent behavioral variation and its genetic
Hunter JS, Fincher GT, Lancaster JL (1991) Observations on the control within and among populations. In: Boake CRB (ed)
life history of Onthophagus medorensis. Southw Entomol Quantitative genetic studies of behavioral evolution. University
16:205±213 of Chicago Press, Chicago, pp 165±187
Huxley JS (1931) Relative growth of mandibles in stag beetles. Linn Wallace AR (1878) Tropical nature and other essays. Macmillan,
J Zool 37:675±703 London
Inukai T (1924) Statistical studies on the variation of stag beetles. West-Eberhard MJ (1979) Sexual selection, social competition and
Trans Sapporo Nat Hist Soc 9:77±91 evolution. Proc Am Philos Soc 123:222±234
Kawano K (1995) Horn and wing allometry and male dimorphism West-Eberhard MJ (1992) Behavior and evolution. In: Grant PR,
in giant rhinoceros beetles (Coleoptera: Scarabaeidae) of trop- Horn HS (eds) Molds, molecules and metazoa: growing points
ical Asia and America. Ann Entomol Soc Am 88:92±99 for evolutionary biology. Princeton University Press, Princeton,
Klemperer HG (1981) Nest construction and larval behaviour of pp 57±75
Bubas bison (L.) and Bubas bubalus (Ol.) (Coleoptera: Scarab- Zar JH (1984) Biostatistical analysis, 2nd edn. Prentice-Hall, En-
aeidae). Ecol Entomol 6:23±33 gelwood Cli€s, New Jersey
Klingenberg CP (1996) Multivariate allometry. In: Marcus LF (ed) Zimmerer EJ, Kallman KD (1989) Genetic basis for alternative
Advances in morphometrics. Plenum, New York, pp 23±49 reproductive tactics in the pygmy swordtail, Xiphophorus
Kodric-Brown A (1986) Satellites and sneakers: opportunistic male nigrensis. Evolution 43:1298±1307
breeding tactics in pup®sh (Cyprinodon pecosensis). Behav Ecol
Sociobiol 19:425±432
Levins R (1968) Evolution in changing environments. Princeton
University Press, Princeton Communicated by L.W. Simmons
Lively CM (1986a) Predator-induced shell dimorphism in the acorn
barnacle Chthamalus anisopoma. Evolution 40:232±242