Social cancer and the biology of the clonal ant Pristomyrmex punctatus (Hymenoptera: Formicidae)

Abstract

We review some aspects of the biology of the ant Pristomyrmex punctatus, in which the winged queen caste is absent and wingless females reproduce by thelytokous parthenogenesis. The majority of females have two ovarioles, whereas up to 50% of colonies contain large-bodied females which have four ovarioles. We call the former workers and the latter ergatoid queens. Males are rare. Some ergatoid queens have a spermatheca, but no inseminated individual has been found so far. Castes are morphologically defined, and workers engage in asexual reproduction in all colonies regardless of the presence of ergatoid queens. In colonies containing only workers, reproductive division of labor is regulated by age-polyethism: All young workers reproduce and fulfill inside-nest roles, and old workers become sterile and fulfill outside-nest roles. Colonies are founded by fission or budding, and consequently neighboring colonies are often re-lated. Nevertheless, populations are multi-colonial, with strong hostility among neighboring colonies. A genetic analysis revealed that colonies often have multiple genotypes (parthenogenetic lineages), and suggested that the majority of those lineages can produce both workers and ergatoid queens. However, a lineage in a population in central Japan pro-duces only ergatoid queens. We define these queens as cheaters, as they fulfill no other task than oviposition and therefore depend on the work force of other non-cheater lineages. Ergatoid queens in cheater lineages have three distinct ocelli, but those in non-cheater lineages usually have zero to two. As cheaters are likely to be horizontally transmitted, we draw an analogy to transmissible cancers. The coexistence of cheaters and non-cheaters is discussed in the frameworks of multilevel selection in the short term, and local extinction-immigration in the long term. However, many things re-main to be studied, such as the developmental origin of the ergatoid queens, the frequency of sexual reproduction, and how colony identity is maintained.

Full text

Myrmecological News 15 91-99 Vienna, May 2011
Ross H. Crozier Memorial Volume
Social cancer and the biology of the clonal ant Pristomyrmex punctatus (Hymenoptera:
Formicidae)
Kazuki TSUJI & Shigeto DOBATA
Abstract
We review some aspects of the biology of the ant Pristomyrmex punctatus, in which the winged queen caste is absent
and wingless females reproduce by thelytokous parthenogenesis. The majority of females have two ovarioles, whereas
up to 50% of colonies contain large-bodied females which have four ovarioles. We call the former workers and the
latter ergatoid queens. Males are rare. Some ergatoid queens have a spermatheca, but no inseminated individual has been
found so far. Castes are morphologically defined, and workers engage in asexual reproduction in all colonies regardless
of the presence of ergatoid queens. In colonies containing only workers, reproductive division of labor is regulated by
age-polyethism: All young workers reproduce and fulfill inside-nest roles, and old workers become sterile and fulfill
outside-nest roles. Colonies are founded by fission or budding, and consequently neighboring colonies are often re-
lated. Nevertheless, populations are multi-colonial, with strong hostility among neighboring colonies. A genetic analysis
revealed that colonies often have multiple genotypes (parthenogenetic lineages), and suggested that the majority of
those lineages can produce both workers and ergatoid queens. However, a lineage in a population in central Japan pro-
duces only ergatoid queens. We define these queens as cheaters, as they fulfill no other task than oviposition and therefore
depend on the work force of other non-cheater lineages. Ergatoid queens in cheater lineages have three distinct ocelli,
but those in non-cheater lineages usually have zero to two. As cheaters are likely to be horizontally transmitted, we
draw an analogy to transmissible cancers. The coexistence of cheaters and non-cheaters is discussed in the frameworks
of multilevel selection in the short term, and local extinction-immigration in the long term. However, many things re-
main to be studied, such as the developmental origin of the ergatoid queens, the frequency of sexual reproduction, and
how colony identity is maintained.
Key words: Clonal ant, cheating, social cancer, genetic caste determination, parthenogenesis, ergatoid queen, colony
discrimination, review.
Myrmecol. News 15: 91-99 (online 20 April 2011)
ISSN 1994-4136 (print), ISSN 1997-3500 (online)
Received 27 September 2010; revision received 4 February 2011; accepted 8 February 2011
Prof. Dr. Kazuki Tsuji (contact author) & Dr. Shigeto Dobata, Department of Agro-Environmental Sciences, Faculty of
Agriculture, University of the Ryukyus, Nishihara, Okinawa 903-0213, Japan. E-mail: tsujik@agr.u-ryukyu.ac.jp
Introduction
In queenless ant species, the role of morphologically dis-
tinguishable queens as the primary reproductive females is
taken by mated workers called gamergates. Gamergates in
some members of the Amblyoponinae, Ponerinae, and Ec-
tatomminae are functional queens (MONNIN & PEETERS
2008), and colonies show reproductive division of labor be-
tween gamergates and sterile workers. In contrast, although
winged queens are also absent in the myrmicine ant Pristo-
myrmex punctatus (previously called P. pungens) and in
the cerapachyine ant Cerapachys biroi, the reproductive di-
vision of labor is unrelated to mating (TSUJI 1988a, RA-
VARY & JAISSON 2004). Instead, many wingless females
reproduce totally asexually by thelytokous parthenogenesis
(ITOW & al. 1984, TSUJI & YAMAUCHI 1995). These social
systems differ from those recently found in two, also ob-
ligatory parthenogenetic, myrmicine ants, Mycocepurus
smithii and Pyramica membranifera, in which winged
queens produce both sterile workers and new queens com-
pletely parthenogenetically (HIMLER & al. 2009, RABELING
& al. 2009, ITO & al. 2010). In this paper, we review some
aspects of the biology of Pristomyrmex punctatus to cla-
rify its peculiarity among "thelytokous ants". We also dis-
cuss that P. punctatus is a good system to study evolu-
tionary dynamics of cooperation and cheating.
Life cycle and division of labor
The majority of colonies in Pristomyrmex punctatus consist
of only workers that are morphologically defined (TSUJI
1988a, 1995), although some colonies have ergatoid queens
whose definition is described later in detail. First, we de-
scribe the biology of those colonies containing only wor-
kers. In ants and social bees, division of labor among wor-
kers is mostly age dependent, in that young workers ful-
fill inside-nest roles such as nursing, while old workers per-
form outside-nest tasks such as foraging (the temporal castes
or age-polyethism; OSTER &WILSON 1978). In P. puncta-
tus, reproductive division of labor is regulated by this age-
polyethism (TSUJI 1990a): Young workers reproduce and

simultaneously perform other inside-nest tasks, while old
workers cease reproduction and perform outside-nest tasks.
The distribution of the number of mature oocytes that nest-
workers have fits a Poisson distribution (TSUJI 1988a). This
means that all young workers have a similar probability of
egg-laying and the variation of short-term fecundity may
occur just by chance. All workers follow this same beha-
vioral development, and therefore there are neither perma-
nently reproductive nor permanently sterile workers. Fur-
thermore, neither dominance behavior nor policing has been
observed among workers (TSUJI 1988a, 1990a). This si-
tuation has raised semantic arguments of whether this ant
should be regarded as eusocial or not (TSUJI 1990a, 1992,
FUREY 1992). Reproductive division of labor in Cerapachys
biroi is likewise regulated by age-polyethism. Although
C. biroi colonies show cyclic phases as seen in some Eci-
ton army ants (RAVARY & JAISSON 2002, RAVARY & al.
2006), old workers do not reproduce even in the stationary
(non-migratory) phase, in which young workers lay eggs
(RAVARY & JAISSON 2004).
The reproductive biology of Pristomyrmex punctatus
and Cerapachys biroi markedly differ from any other ex-
ample of ants that are known to be able to reproduce by
thelytokous parthenogenesis in that the majority (possibly
all) of females participate in reproduction at least once in
their life. In these other ants, the majority of females are
sterile helpers, with reproductive division of labor regulated
either by the physical caste system (such as in Anoplolepis
gracilipes, see DRESCHER & al. 2007, Cataglyphis spp., see
PEARCY & al. 2004, TIMMERMANS & al. 2008, Pyramica
membranifera, see ITO & al. 2010, Mycocepurus smithii,
see HIMLER & al. 2009, RABELING & al. 2009, Messor ca-
pitatus, see GRASSO & al. 2000, Vollenhovia emeryi, see
OHKAWARA & al. 2006, KOBAYASHI & al. 2008, and Was-
mannia auropunctata, see FOURNIER & al. 2005) or by be-
havioral dominance and / or policing (in Platythyrea punc-
tata, see HARTMANN & al. 2003). Furthermore, Pristomyr-
mex punctatus and Cerapachys biroi contrast with Platy-
thyrea punctata in which either physical caste system or
behavioral interference regulates reproductive division of
labor depending on colonies and populations (HEINZE &
HÖLLDOBLER 1995, SCHILDER & al. 1999, HARTMANN &
al. 2003).
Although age-polyethism is the general pattern that in
turn regulates the reproductive division of labor in Pristo-
myrmex punctatus, individuals show large variation in be-
havioral development: The timing of the switch from inside-
nest tasks to outside-nest tasks occurs two to three weeks
before the worker's death (in lab-reared colonies), yet indi-
vidual life span is highly variable for unknown reasons
(TSUJI 1990a). Consequently, a long-lived worker tends to
have a longer inside-nest phase and therefore more oppor-
tunity for reproduction than a short-lived one. At least two
interpretations of the adaptive significance of this pheno-
menon are possible: either that the long lifespan in some
workers is a manifestation of the individuals' self-interest
(see our later discussion on cheaters), or that this is a means
to maintain the colony's efficient task allocation. The an-
nual reproductive cycle of colonies we studied in central
Honshu, the main island of Japan, shows why such flexi-
bility would be needed: New adults emerge from June to
September (peaking in early August). They withhold repro-
ducing until after overwintering. Overwintered adults begin
oviposition in April; they lay eggs and raise brood until
the end of fall, when they die. The adult emergence period
(June – September) is shorter than the breeding period
(April – September). Therefore, if all individuals followed
the same schedule of age-polyethism, inefficient task allo-
cation would happen at some time of the year; for example,
a shortage of foragers or of egg-layers. Such unbalanced
task allocation lowers colony efficiency under experimen-
tal conditions (TSUJI 1994) and probably also in the field
(TSUJI 1995).
Individual variation might enable a colony to evade such
a problem. TSUJI (1994) found that laboratory colonies of
which all foragers were removed at the outset of the ex-
periment later produced some brood. Indeed, during the ex-
perimental period for 20 days, some young workers which
had never previously foraged started to forage and obtained
food. Although this behavior might have been an artifact
caused by the small size of the lab cage, or more simply
aging (some might have become foragers) could account
for it, it is also possible that some nest-workers could be
flexibly recruited to foraging when a large work force is
needed. So far, division of labor of this species was ob-
served mostly in laboratories with constant conditions.
Behavior, in particular the flexibility of division of labor,
should be studied also under heterogeneous environments
like in the field.
Migration, colony founding, and nestmate discrimination
Despite its unusual biology, Pristomyrmex punctatus forms
a typical multi-colonial population (TSUJI & ITÔ 1986).
The colony size ranges from 4,000 to 320,000 adults (TSUJI
1988a, 1995). Neighboring colonies are strongly hostile to-
ward each other, and lethal fights sometimes occur in the
field (TSUJI 1988b). The absence of alate queens implies
that new colonies must be founded by fission or budding.
However, so far no one has directly observed a colony be-
ing founded in the field. In laboratories, small colonies,
with less than 100 individuals, always show very low per-
formance and often fail to produce any adult offspring
(K. Tsuji & S. Dobata, unpubl.). This suggests that colony
founding should occur with a large group of individuals,
i.e., possibly at least hundreds. Colonies do not construct
an elaborate underground nest, but instead use naturally pre-
existing cavities, such as the space under a stone. They
frequently move nest sites, and the average residence time
is around two to three weeks (TSUJI 1988c). Colonies are
usually monodomous, but become polydomous under some
conditions (TSUJI 1988c). Polydomous colonies are more
frequently found in summer, when they actively forage,
than in spring and fall, when foraging is less active (TSUJI
1988c). Colony fission is inferred to occur by chance as a
consequence of the frequent migrations, which can also
occasionally give rise to polydomy (TSUJI 1988c).
Neighboring colonies of fission-founding ants are likely
to be related. An isolation-by-distance pattern revealed by
genetic markers empirically supported this prediction in
Pristomyrmex punctatus (see DOBATA & al. 2011). This
situation raises the question of nestmate discrimination:
How can these ants discriminate colony membership so
strictly (TSUJI & ITÔ 1986, TSUJI 1988b) despite the gene-
tic similarity? They may use environmentally derived chem-
icals as colony-specific labels (TSUJI 1990b). Indeed, arti-
ficial mixing of different colonies and thus of the putative
92

93
Tab. 1: A list of names previously used to describe phenotypic variants of Pristomyrmex punctatus females. * + is present,
– is absent, ? means information insufficient or non-available (GOTOH & al. 2011). ** TERANISHI (1929) wrote that the
ergatoid queens he defined can be also called pseudogynes. *** In non-cheater lineages, ocelli (if any) are always vestigial
and can be observed only by using strong light that passes through the ant's head (DOBATA & al. 2011). Therefore, it is not
surprising that many previous authors did not recognize them.
Phenotype References and used terminology
Number of
ovarioles Number
of ocelli Sperma-
theca*
Description of its occurrence
This paper,
DOBATA &
al. (2011)
WANG (2003), ITOW &
al. (1984), TERA-NISHI
(1929)**
TSUJI (1988a,
1995), SASAKI
& TSUJI (2003)
DOBATA &
al. (2009)
2 0 – More than 50% of colonies consist
of only this female category. worker worker small (or nor-
mal) worker S-type
0 – (?) (no name is provided)
1-2 ?
Found in 5 - 50% of colonies in all
populations. Its frequency varies
over years. Its intracolonial propor-
tion is 0 - 30%. (no name is provided) ***
4
3 + Very rare. So far, collection is re-
peatable only in 2 populations. Its
intracolonial proportion is 0 - 50%.
ergatoid
queen
ergatoid queen
large worker L-type
labels indicated that ants can absorb labels from outside
(TSUJI 1990b). Regardless of whether the colony label is
genetic or environmental, one can infer that neighboring
colonies tend to share more similar labels with each other
than with distant colonies. Therefore, one could predict a
positive correlation between geographical distance and in-
tercolonial hostility, like the "dear enemy" phenomenon
(FISHER 1954) that is known also in ants (e.g., HEINZE &
al. 1996). However, SANADA-MORIMURA & al. (2003) ob-
served the opposite in P. punctatus: Ants strongly attacked
their neighbors but ignored ants from distant nests! The
proximate behavioral explanation is that individual ants be-
come more aggressive to a specific foreign colony because
of frequent contact with members of that colony (SANADA-
MORIMURA & al. 2003): Outside the nest, ants should more
frequently meet neighbors than members of a distant nest
under natural conditions. Such a flexible response can be
adaptive, because neighbors should directly compete for re-
sources such as food and nest sites, and thus pose a threat,
but a worker from a distant nest is less likely to recruit
colony members en masse and is therefore less of a threat.
These phenomena might be more or less general in group-
foraging ants (see also BROWN & GORDON 1997). These
mechanisms assume that ants can discriminate nestmates
from non-nestmates, but do not attack those from distant
nests. However, colony-discrimination mechanisms in P.
punctatus still largely remain to be studied.
NISHIDE & al. (2007) discussed the possible occurrence
of colony fusion in a Pristomyrmex punctatus field popu-
lation, because of the frequently observed intracolonial var-
iation in microsatellite loci. However, we wonder about its
generality for two reasons.
First, there is a frequent misunderstanding here: Asex-
ual reproduction itself does not necessarily lead to a gen-
etically homogenous colony or population unless experi-
encing a bottleneck in their specific lifecycle or in their en-
tire population dynamics. High intracolonial genetic vari-
ation may simply reflect the long-lasting large effective
population size (Ne) in each colony, because neutral theory
predicts that a (nearly-)neutral mutant allele takes 4Ne gen-
erations to be fixed in a diploid population, and the muta-
tion rates of microsatellite loci in Pristomyrmex punctatus
were estimated as ca. 10-5 per generation (DOBATA & al.
2011). In the colony's lifecycle of P. punctatus, there is
seemingly no genetic bottleneck, such as the founding stage
of a single queen, similar to the single zygotic stage of an
embryo that grows into a multicellular organism. Therefore
"somatic" mutations can build up within the colony. Note
that this situation remarkably differs from that of another
parthenogenetic species, Platythyrea punctata, in which be-
havioral interaction among workers in queenless colonies
leads to monopolization of the colony's reproduction by a
single parthenogenetic worker (SCHILDER & al. 1999, HART-
MANN & al. 2003). Interestingly, for Platythyrea punctata,
KELLNER & al. (2010) discussed that the loss of intraco-
lonial genetic diversity is recovered by colony fusion.
Second, instead, high genetic differentiation among col-
onies within a local population is maintained in Pristomyr-
mex punctatus, contradicting the frequent occurrence of
colony fusion. As measured by intracolonial relatedness
(r), genetic differentiation among colonies was r = 0.4325
± 0.0749 (DOBATA & al. 2009). If colony fusions or inter-
colonial exchanges of individuals were frequent, such high
genetic differentiation among colonies would not be ex-
pected. Alternatively, colony fusion could occur by chance
when the colonies shared the same environmentally derived
cues, but such a "hybrid" colony would soon be eliminated
by colony-level selection. In fact, NISHIDE & al. (2007)
experimentally revealed the low performance of genetical-
ly polymorphic colonies, shown as slower nest relocation
and difficulty in aggregating, as is seen in chimeric slime
molds (FOSTER & al. 2002), but this is unlike examples of-
ten discussed in other social insects in which high genetic
diversity leads to high colony performance (e.g., MATTILA
& SEELEY 2007, WIERNASZ & al. 2008). Many aspects of
the maintenance of colony identity and the colony found-
ing process remain to be studied.
Morphological castes and cheaters
In Pristomyrmex punctatus, the majority of adult females
have two ovarioles (one in each ovary). In some colonies,
females that have four ovarioles (two in each ovary) are
also found. Different authors have used various terminolo-
gies for these morphological variants in P. punctatus (Tab. 1),

Fig. 1: The body-size distributions of workers and ergatoid
queens in Pristomyrmex punctatus that were re-drawn using
the data of DOBATA & al. (2009) studied in Kihoku popu-
lation. (a) Colonies containing only workers (N[indv.] = 90;
N[colony] = 9). (b) Colonies containing both workers and
ergatoid queens with zero to two ocelli (N[indv.] = 228;
N[colony] = 14). (c) Colonies containing cheaters (ergatoid
queens with three distinct ocelli) (N[indv] = 198; N[colony]
= 10). Workers, ergatoid queens with zero to two ocelli and
cheaters are shown in white, grey and black bars, respec-
tively.
causing confusion. Focusing on the discontinuity in the
ovariole number and judging from morphology of wor-
kers and queens in related species, DOBATA & al. (2011)
proposed calling the females with two ovarioles "workers"
and those with four ovarioles "ergatoid queens" (permanent-
ly wingless queens as defined in MOLET & al. 2009). The
se of the term "ergatoid queen" is justified also by the fact
that a portion of ergatoid queens, at least those with three
ocelli (see later), have a spermatheca (ITOW & al. 1984,
GOTOH & al. 2011). These morphological conditions in
"ergatoid queens" of P. punctatus resemble ergatoid queens
of other ants including some Pristomyrmex species (WANG
2003). However, we acknowledge that ergatoid queens in
P. punctatus are exceptional, because they reproduce asex-
ually (DOBATA & al. 2009, 2011), whereas most ergatoid
queens known in other ants reproduce sexually (MOLET &
al. 2009). Like in queenless colonies of other ants, P. punc-
tatus workers keep reproducing asexually even in the pres-
ence of ergatoid queens (TSUJI 1988a, 1995). This is also
an exceptional characteristic in P. punctatus.
u
The body size of ergatoid queens in Pristomyrmex punc-
tatus is on average larger than that of workers, but with an
overlapping distribution (Fig. 1b). Most ergatoid queens
have zero to two ocelli, but some have three and are dis-
tinctly larger. Those with three ocelli are usually very rare,
but are common in some populations. So far, the collec-
tion of ergatoid queens with three ocelli is repeatable only
in two populations, i.e., Kihoku on Honshu island (DOBATA
& al. 2009, 2011) and Takamatsu (Nomoto) on Shikoku is-
land (HASEGAWA & al. 2011). Colonies containing many
ergatoid queens with three ocelli have a bimodal distribu-
tion of body size (Fig. 1c; ITOW & al. 1984, SASAKI &
TSUJI 2003, DOBATA & al. 2009), forming a striking con-
trast to the unimodal distribution of the majority of col-
onies (Fig. 1a, b). Ergatoid queens without ocelli can be
found in most populations: depending on population and
year, in up to 50% of colonies within a population, and up
to 30% of individuals (TSUJI 1988a, 1995, DOBATA & al.
2011). The average proportion of ergatoid queens without
ocelli in the entire population is usually 0.01 - 5% (TSUJI
1988a, 1995). However, a thorough dissection might detect
ergatoid queens without ocelli at low proportions in most
colonies. Colony size is not correlated with the proportion
of ergatoid queens (TSUJI 1995). Much more rarely than
those without ocelli, some ergatoid queens have one or two
ocelli that seem more or less vestigial. The body-size dis-
tribution of these queens is in between those with three
ocelli and those with none (DOBATA & al. 2009) (Fig. 2).
We do not use the ocellus as the key character discrimi-
nating castes in P. punctatus, because we wonder about its
function. In insects, ocelli function as light meters and are
involved in stabilization reflexes during flight (KRAPP 2009);
accordingly, they are often lost in non-flying insects such
as workers of many ants. So ocelli are likely to be non-
functional in ergatoid queens of P. punctatus.
Workers have no spermatheca, but ergatoid queens with
three ocelli have a spermatheca with a seemingly normal ac-
cessory gland (GOTOH & al. 2011). Much less is known
about the spermathecal condition in ergatoid queens with
zero to two ocelli, though all ergatoid queens without ocelli
examined so far (N > 200) apparently had no spermatheca
(TSUJI 1988a, TSUJI 1995, SASAKI & TSUJI 2003, DOBATA
& al. 2009).
Irrespective of the number of ocelli, ergatoid queens
share the same behavioral characteristics: They lay more
eggs than workers but rarely perform other tasks such as
nursing (SASAKI & TSUJI 2003). Ergatoid queens were nev-
er observed to forage in the laboratory (SASAKI & TSUJI
2003), although they were occasionally collected on trails
in the field (TSUJI 1988a). This suggests that ergatoid queens
94

Fig. 2: The relationship between the body size (head width) of ergatoid queens and their number of ocelli. Data are from
DOBATA & al. (2009) from the Kihoku population. Measures of cheaters (filled triangle) and non-cheaters (open circles)
were combined. Correlation was positive and statistically significant (r = 0.760, p < 0.0001).
do not follow the age-polyethism and old individuals may
continue to lay eggs. This behavioral propensity of ergatoid
queens in Pristomyrmex punctatus is normal among queens
of species in which the division of labor between repro-
ductive queens and sterile workers enhances colony per-
formance, but, as TSUJI (1995) discussed, ergatoid queens
in P. punctatus could be harmful to the colony (see also
SASAKI & TSUJI 2003) for the following reasons. Empirical
evidence indicates that having only workers is enough for
P. punctatus colonies to prosper (MIZUTANI 1980, TSUJI
1990a, 1994). More importantly, it is known that colonies
with only workers are the most productive, and the average
lifetime reproductive success of nestmates is estimated to
decrease as the proportion of ergatoid queens increases in
field colonies (TSUJI 1995). The contextual analysis ap-
plied to a field population detected no fitness peak (no sta-
bilizing selection) at a small proportion of ergatoid queens
(TSUJI 1995: tab. 4), suggesting that the idea that a low
but non-zero ergatoid queen proportion in a colony en-
hances the colony's fitness through reproductive division of
labor is not supported. Although we need more data across
populations and years on the relationship between these
individual and colony phenotypes and their reproductive
outputs, their harmful effect to the colony is evident when
ergatoid queens are genetically distinct cheaters as de-
scribed below.
In Kihoku, central Japan, we found parthenogenetic
lineages that give rise only to ergatoid queens (DOBATA &
al. 2009). We defined these lineages as cheaters, because
they can obtain a large reproductive output at the cost of
fitness of other nestmate lineages that produce workers.
Non-cheaters, or cooperators, are defined as lineages that
can become workers. We use the plural term "lineages", be-
cause as mentioned previously, in Pristomyrmex punctatus,
colonies are often genetically heterogeneous, particularly
the colonies with cheaters always contain more than one
genotype. This is owing to their social-parasitic character-
istics, i.e., cheaters can exist only with workers that be-
long to a different lineage. In other words, cheaters de-
pend on the work force of other lineages. Cheating in P.
punctatus is therefore a genotypic concept, and is a speci-
fic case of genetic caste determination. A genetic analysis
suggests that some non-cheater lineages produce only wor-
kers and other non-cheater lineages can produce both wor-
kers and ergatoid queens (DOBATA & al. 2011). The latter
cases are most likely an account of phenotypic plasticity af-
fected by larval nutrition as seen in the caste differentia-
tion of most ants (DOBATA & al. 2011). Given the castes are
mostly environmentally determined in those non-cheater
lineages, we, however, do not rule out the possibility that
some non-cheater lineages have a propensity to produce er-
gatoid queens. This is because such weak cheating is re-
ported in some patrilines of a polyandrous Acromyrmex ant
(HUGHES & BOOMSMA 2008), but is yet to be seen in P.
punctatus. Interestingly, almost all adults of the cheater
lineage (98.8%) examined so far have three ocelli. In con-
trast, ergatoid queens in non-cheater lineages have no or
more or less vestigial ocelli that can be usually recognized
only by using strong light that passes through the ant's
head. Moreover, most of them have zero to two ocelli, and
only 11.2% had three ocelli that are also more or less de-
generated (DOBATA & al. 2011). Therefore, in practice, one
can label ergatoid queens with three distinct ocelli as chea-
ters.
The monophyly of cheaters was suggested by analyses
using microsatellite nuclear DNA and mitochondrial haplo-
types in our study population in Kihoku (DOBATA & al.
2009, 2011). The existence of another cheater lineage of
independent origin is inferred from mitochondrial DNA
data in a population in Takamatsu, western Japan (HASE-
GAWA & al. 2011). However, as the rare occurrence of sex-
ual reproduction in this species has been suggested (see
later), further studies incorporating nuclear genomic data
are needed. Cheaters of the identical multiloci genotype are
found in many colonies. This occurrence, as also suggested
by a more rigorous population genetic analysis (DOBATA
& al. 2011), implies the horizontal transmission of chea-
ters. In other words, cheaters migrate to other intact colo-
nies. Nothing is known about this process, however. Such
horizontal transmission of cheaters despite the presence
of strict nestmate discrimination is not surprising, because
many social-parasitic species can evade the host's discri-
mination behavior. A parallel phenomenon is known in the
Cape honey bee (Apis mellifera capensis), in which wor-
kers can reproduce by thelytokous parthenogenesis, as does
Pristomyrmex punctatus. Some parthenogenetic worker lin-
95

eages of the Cape honey bee have become social-parasitic
and migrated among captive colonies of the neighboring
subspecies A. m. scutellata, causing mass extinction of the
host colonies (NEUMANN & MORITZ 2002, DIETEMANN &
al. 2007). One can draw an analogy also to the transmissible
cancer cells that infect and kill the host, as found in some
mammals (e.g., MURGIA & al. 2006, PEARSE & SWIFT
2006). For example, a facial cancer transmitted among Tas-
manian devils (Sarcophilus harrisii) through biting now
poses a serious threat to the persistence of the species. By
this analogy, the parasitic Cape honey bee workers and the
P. punctatus cheaters are called social cancers (OLDROYD
2002, DOBATA & al. 2009; but see KORB & HEINZE 2004
for the case of Cape honeybee).
A question arises about the cause of the difference in
the long-term consequences between these two social can-
cers; one leads to mass extinction within a decade or so (the
Cape honeybee) the other persists much longer (Pristomyr-
mex punctatus). DOBATA & al. (2011) discussed that differ-
ence in migration ability of social cancers may account
for the differential consequences. In theory, it is well es-
tablished that limited migration of parasites can contribute
to the persistence of host-parasite systems by creating the
local extinction-recolonization process. The parasitic Cape
honey bee workers have wings and thus can migrate fur-
ther than the cheaters in P. punctatus that have no wings,
which might hinder the former to persist long. Future theo-
retical and empirical studies should test this idea.
Similar to Pristomyrmex punctatus, Cerapachys biroi
also shows a variation in body size and ovariole number
among wingless females (RAVARY & JAISSON 2002, 2004),
therefore some of these can be called ergatoid queens. More
recently, LECOUTEY & al. (2010) revealed that colony
demography and nutrition affects the production of erga-
toid queens in C. biroi. However, if cheaters also exist in
C. biroi or not is yet to be studied.
Parthenogenesis and rare sexual reproduction
ITOW & al. (1984) studied chromosomes in the ovaries of
pupae of Pristomyrmex punctatus and revealed that the first
meiosis occurs normally in developing oocytes. Micro-
satellite marker loci are often heterozygous (DOBATA & al.
2009). These results suggest that parthenogenesis in P.
punctatus is a central-fusion-type automixis.
Although reproduction of this species is considered to-
tally asexual, males are occasionally found. Males are ha-
ploid (n = 12; females have 2n = 24) and have seemingly
normal sperm (ITOW & al. 1984). As already mentioned,
workers have no spermatheca, but ergatoid queens with
three ocelli have a seemingly functional spermatheca (GO-
TOH & al. 2011). Therefore, ergatoid queens might occa-
sionally reproduce sexually, although we have found no
inseminated individuals despite having dissected more than
a thousand ergatoid queens with three ocelli (T. Sasaki &
K. Tsuji, unpubl.). In the Kihoku population, there is some
mismatching between the mitochondrial DNA-based phylo-
geny and the nuclear microsatellite-based phylogeny, sug-
gesting sexual reproduction in the past leading to genetic
introgression among lineages (DOBATA & al. 2011). How-
ever, we stress that the dominant mode of reproduction is
thelytokous parthenogenesis, which is indicated by several
population-genetic indices; for example, the observed re-
dundancy of the same multilocus genotypes due to distinct
sexual reproductive events has at most a probability of
0.8%, which indicates that this redundancy is derived main-
ly from thelytokous parthenogenesis (DOBATA & al. 2011).
Multilevel selection for the short-term dynamics of
cheater and non-cheater populations
Although we previously mentioned that the long-term co-
existence of cheaters and non-cheaters in Pristomyrmex
punctatus can be described as a local extinction and im-
migration dynamics, another approach might be possible
to understand their short-term dynamics. Ross Crozier was
the first to point out that multilevel or group selection can
be a useful framework to understand the maintenance of
cooperative societies in this ant (see pages 99 - 100 in
Discussion by ITOW & al. 1984 that was written by RC).
TSUJI (1995) also discussed how the coexistence of non-
cheaters and cheaters can be explained by multilevel selec-
tion. Individual selection favors cheaters, whereas colony-
level selection favors non-cheaters. A balance of these two
forces operating in opposition may lead to coexistence at
least in the short term. Note, however, that group selec-
tion and kin selection are not alternative mechanisms but
rather different descriptive models for the same phenome-
non of evolutionary dynamics (QUELLER 1992). In fact,
multilevel selection in P. punctatus can be transformed to
kin selection. Cheaters obtain a higher individual fitness
than non-cheaters (minus c) but harm the fitness of nest-
mates (minus b). Therefore they can increase in the (meta-)
population when relatedness is low: That is, a cheater is no
more likely than by chance to interact with cheaters in the
background metapopulation. In contrast, cheaters can de-
crease when relatedness is high: That is, a cheater is more
likely than by chance to interact with cheaters in the meta-
population, leading to mutual exploitation. We are conduct-
ing long-term research to monitor the microevolutionary
process of cheater-non-cheater dynamics in the field and
to test both group selection and kin selection models (see
TSUJI 1995 for the earliest study).
Intraspecific parasitism or parasitic species?
People often wonder whether cheaters and non-cheaters are
different species, in which case prey-predator or host-para-
site ecological dynamics could provide a more appropriate
description of the phenomenon than group or kin selection.
We consider this view to be wrong for three reasons: (1)
The biological species concept is not applicable to parthe-
nogenetic organisms, and thus we have to resort to other
criteria such as phylogenetic distance. (2) Although chea-
ters and non-cheaters form separate clades in the genetic
analysis of the Kihoku population (DOBATA & al. 2009,
2011), both are nested in a deeper clade of the morpholo-
gical species Pristomyrmex punctatus among other popula-
tions of Japan. Thus, cheaters and non-cheaters in Kihoku
are phylogenetically more closely related to each other than
to P. punctatus found in Okinawa. If cheaters and non-
cheaters were regarded as separate species, many cryptic
species should be discernable within the morphological spe-
cies currently called P. punctatus. This would not be a
practical decision. (3) A model of the biological dynamics
in P. punctatus that includes two parties – one parasitic, the
other non-parasitic – might be called a metapopulation host-
parasite ecological model if the two parties are regarded as
distinct species. Or it might be called a multilevel selec-
96

tion or family-structured kin-selection model if they are
regarded as the same species. However, mathematically,
both models are in principle the same with regard to pre-
dicting the relative abundance of the two parties when one
is exploiting the other. Based on these arguments, we con-
clude that the criticism claiming that the two parties might
be different species is semantic. Regardless of the taxono-
mic status of cheaters and non-cheaters, we believe that this
system provides an ideal opportunity to test models of eco-
logical and evolutionary dynamics.
The origin of cheaters
The body-size difference between workers and ergatoid
queens (with three ocelli) in Pristomyrmex punctatus cor-
responds almost exactly to that of congeneric species with
alate queens (DOBATA & TSUJI 2009). Furthermore, rudi-
mentary wings can be seen under the cuticle of the thorax
of these ergatoid queens. This evidence suggests that erga-
toid queens in P. punctatus are real queens with arrested
development.
However, we have to stress that there is an exceptional
characteristic of ergatoid queens in Pristomyrmex punc-
tatus and in Cerapachys biroi, in addition to their asexu-
ality. Ergatoid queens in these two species reproduce to-
gether with many workers, forming a striking contrast to
other cases in which reproductive division of labor between
reproductive ergatoid queens and sterile workers is conspi-
cuous.
The other congeneric species are all known to have
queen (either winged or ergatoid) and worker castes (WANG
2003). Furthermore, males are found in many of those spe-
cies (WANG 2003), suggesting their sexual reproduction.
Therefore, asexual reproduction by workers instead of a
winged queen in Pristomyrmex punctatus is likely a derived
condition. Interestingly, P. rigidus, which is presumably
the species most closely related to P. punctatus, exhibits
a conventional social system with winged queens (WANG
2003). From the above circumstantial evidence we set out
the hypothesis that cheaters in P. punctatus arose from a
mutation that reuses the hidden developmental trajectory
to a winged queen. In other words, although the ancestor of
P. punctatus suppressed the queen trajectory, the genes en-
coding the developmental pathway are retained, and there-
fore mutations in regulatory regions can restore the path-
way (except for full development of wings). Among such
putative mutations, the ones with strong effects gave rise to
cheaters in the Kuhoku population and the ones with weak
effects lead to lineages producing both workers and erga-
toid queens. Alternatively, the following scenario is also
possible. Wing morphology of ant queens is likely an evo-
lutionary labile trait (HEINZE & TSUJI 1995, MOLET & al.
2009, PERFILIEVA 2010). Indeed, a few species of Pristo-
myrmex, such as P. africanus and P. wheeleri, have both
winged queens and ergatoid queens (WANG 2003). These
intraspecific variation indicates winglessness of queens was
evolutionary attainable also in Pristomyrmex. The ability
of thelytokous parthenogenesis might have evolved in an
ancestral species, where ergatoid queens reproduced sexual-
ly as in some other species of Pristomyrmex (WANG 2003).
Then, parthenogenetic reproduction by ergatoid queens to-
gether with young workers prevailed. Later, in some popu-
lations, parthenogenetic queens became selected for, where-
as in other populations ergatoid queens were selected against.
A comparative study across populations and across related
species, with detailed descriptions of morphology and di-
vision of labor together with phylogenetic data, is needed
to test these hypotheses.
Concluding remarks
Although the above scenarios on the origin of the unusual
reproductive biology in Pristomyrmex punctatus are largely
hypothetical, the ideas provide an interesting model system
for future studies in developmental biology and in evolu-
tionary biology from genomics to ecology.
Acknowledgements
We thank T. Sasaki and C. Peeters for discussion. We also
thank two anonymous referees for their comments. This
work is supported by a JSPS Research Fellowship for
Young Scientists (S.D.), KAKENHI (No.18370012 and
21247006 to K.T) and Mitsui and Co., Ltd. Environment
Fund (08R-B047).
References
BROWN, M.J.F. & GORDON, D.M. 1997: Individual specialisation
and encounters between harvester ant colonies. – Behaviour
134: 849-866.
DIETEMANN, V., NEUMANN, P., HARTEL, S., PIRK, C.W.W. &
CREWE, R.M. 2007: Pheromonal dominance and the selection
of a socially parasitic honeybee worker lineage (Apis mellifera
capensis ESCH.). – Journal of Evolutionary Biology 20: 997-1007.
DOBATA, S., SASAKI, T., MORI, H., HASEGAWA, E., SHIMADA, M.
& TSUJI, K. 2009: Cheater genotypes in the parthenogenetic ant
Pristomyrmex punctatus. – Proceedings of the Royal Society
B-Biological Sciences 276: 567-574.
DOBATA, S., SASAKI, T., MORI, H., HASEGAWA, E., SHIMADA, M.
& TSUJI, K. 2011: Persistence of the single lineage of trans-
missible "social cancer" in an asexual ant. – Molecular Ecol-
ogy 20: 441-455.
DOBATA, S. & TSUJI, K. 2009: A cheater lineage in a social in-
sect: implications for the evolution of cooperation in the wild.
– Communicative & Integrative Biology: 2: 67-70.
DRESCHER, J., BLÜTHGEN, N. & FELDHAAR, H. 2007: Population
structure and intraspecific aggression in the invasive ant spe-
cies Anoplolepis gracilipes in Malaysian Borneo. – Molecular
Ecology 16: 1453-1465.
FISHER, J.B. 1954: Evolution and bird sociality. In: HUXLEY, J.,
HARDY, A.C. & FORD, E.B. (Eds.): Evolution as a process. –
Allen and Unwin, London, UK, pp. 71-83.
FOSTER, K.R., FORTUNATO, A., STRASSMANN, J.E. & QUELLER, D.
C. 2002: The costs and benefits of being a chimera. – Proceed-
ings of the Royal Society B-Biological Sciences 269: 2357-2362.
FOURNIER, D., ESTOUP, A., ORIVEL, J., FOUCAUD, J., JOURDAN, H.,
LE BRETON, J. & KELLER, L. 2005: Clonal reproduction by
males and females in the little fire ant. – Nature 435: 1230-1234.
FUREY, R.E. 1992: Division of labour can be morphological and/
or temporal: a reply to Tsuji. – Animal Behaviour 44: 571.
GOTOH, A., BILLEN, J., TSUJI, K., SASAKI, T. & ITO, F. 2011: His-
tological study of the spermatheca in three thelytokous par-
thenogenetic ant species, Pristomyrmex punctatus, Pyramica
membranifera and Monomorium triviale (Hymenoptera: Formi-
cidae). – Acta Zoologica, doi: 10.1111/j.1463-6395.2010.00498.x.
GRASSO, D.A., WENSELEER, T., MORI, A., LE MOLI, F. & BILLEN,
J. 2000: Thelytokous worker reproduction and lack of Wol-
bachia infection in the harvesting ant Messor capitatus. –
Ethology Ecology & Evolution12: 309-314.
97

HARTMANN, A., WANTIA, J., TORRES, J.A. & HEINZE, J. 2003:
Worker policing without genetic conflicts in a clonal ant. – Pro-
ceedings of the National Academy of Sciences of the United
States of America 100: 12836-12840.
HASEGAWA, E., KOBAYASHI, K., YAGI, N. & TSUJI, K. 2011: Com-
plete mitochondrial genomes of normal and cheater morphs
in the parthenogenetic ant Pristomyrmex punctatus (Hymeno-
ptera: Formicidae). – Myrmecological News 15: 85-90.
HEINZE, J., FOITZIK, S., HIPPERT, A. & HÖLLDOBLER, B. 1996:
Apparent dear-enemy phenomenon and environment-based re-
cognition cues in the ant Leptothorax nylanderi. – Ethology
102: 510-522.
HEINZE, J. & HÖLLDOBLER, B. 1995: Thelytokous parthenogen-
esis and dominance hierarchies in the ponerine ant, Platythyrea
punctata. – Naturwissenschaften 82: 40-41.
HEINZE, J. & TSUJI, K. 1995: Ant reproductive strategies. – Re-
searches on Population Ecology 37: 135-149.
HIMLER, A.G., CALDERA, E.J., BAER, B.C., FERNANDEZ-MARIN,
H. & MUELLER, U.G. 2009: No sex in fungus-farming ants or
their crops. – Proceedings of the Royal Society B-Biological
Sciences 276: 2611-2616.
HUGHES, W.O.H. & BOOMSMA, J.J. 2008: Genetic royal cheats in
leaf-cutting ant societies. – Proceedings of the National Acade-
my of Sciences of the United States of America 105: 5150-5153.
ITO, F., TOUYAMA, Y., GOTOH, A., KITAHIRO, S. & BILLEN, J.
2010: Thelytokous parthenogenesis by queens in the dacetine
ant Pyramica membranifera (Hymenoptera: Formicidae). –
Naturwissenschaften 97: 725-728.
ITOW, T., KOBAYASHI, K., KUBOTA, M., OGATA, K., IMAI, H.T. &
CROZIER, R.H. 1984: The reproductive cycle of the queenless
ant Pristomyrmex pungens. – Insectes Sociaux 31: 87-102.
KELLNER, K., BARTH, B. & HEINZE, J. 2010: Colony fusion causes
within-colony variation in a parthenogenetic ant. – Behavioral
Ecology and Sociobiology 64: 737-746.
KOBAYASHI, K., HASEGAWA, E. & OHKAWARA, K. 2008: Clonal
reproduction by males of the ant Vollenhovia emeryi (WHEE-
LER). – Entomological Science 11: 167-172.
KORB, J. & HEINZE, J. 2004: Multilevel selection and social evo-
lution of insect societies. – Naturwissenschaften 91: 291-304.
KRAPP, H.G. 2009: Ocelli. – Current Biology 19: R435-R437.
LECOUTEY, E., CHÂLINE, N. & JAISSON, P. 2010: Clonal ant so-
cieties exhibit fertility-dependent shifts in caste ratios. – Be-
havioral Ecology, doi: 10.1093/beheco/arq182.
MATTILA, H.R. & SEELEY, T.D. 2007: Genetic diversity in honey
bee colonies enhances productivity and fitness. – Science 317:
362-364.
MIZUTANI, A. 1980: Preliminary report on worker oviposition in
the ant Pristomyrmex pungens MAYR. – Kontyû 48: 327-332.
MOLET, M., FISHER, B., ITO, F. & PEETERS, C. 2009: Shift from
independent to dependent colony foundation and evolution of
"multi-purpose" ergatoid queens in Mystrium ants (subfamily
Amblyoponinae). – Biological Journal of the Linnean Society
98: 198-207.
MONNIN, T. & PEETERS, C. 2008: How many gamergates is an ant
queen worth? – Naturwissenschaften 95: 109-116.
MURGIA, C., PRITCHARD, J.K., KIM, S., FASSATI, A. & WEISS,
R.A. 2006: Clonal origin and evolution of a transmissible can-
cer. – Cell 126: 477-487.
NEUMANN, P. & MORITZ, R.F.A. 2002: The Cape honeybee phe-
nomenon: the sympatric evolution of a social parasite in real
time? – Behavioral Ecology and Sociobiology 52: 271-281.
NISHIDE, Y., SATOH, T., HIRAOKA, T., OBARA, Y. & IWABUCHI,
K. 2007: Clonal structure affects the assembling behavior in
the Japanese queenless ant Pristomyrmex punctatus. – Natur-
wissenschaften 94: 865-869.
OHKAWARA, K., NAKAYUMA, M., SATO, A., TRINDL, A. & HEINZE,
J. 2006: Clonal reproduction and genetic caste differences in
a queen-polymorphic ant, Vollenhovia emeryi. – Biology Let-
ters 2: 359-363.
OLDROYD, B.P. 2002: The Cape honeybee: an example of a
social cancer. – Trends in Ecology & Evolution 17: 249-251.
OSTER, G.F. & WILSON, E.O. 1978: Castes and ecology in social
insects. – Princeton University Press, Princeton, NJ, 352 pp.
PEARCY, M., ARON, S., DOUMS, C. & KELLER, L. 2004: Condi-
tional use of sex and parthenogenesis for worker and queen
production in ants. – Science 306: 1780-1783.
PEARSE, A. M. & SWIFT, K. 2006: Transmission of devil facial-
tumour disease. – Nature 439: 549.
PERFILIEVA, K.S. 2010: Trends in evolution of ant wing vena-
tion (Hymenoptera, Formicidae). – Entomological Review 90:
857-870.
QUELLER, D.C. 1992: Quantitative genetics, inclusive fitness, and
group selection. – The American Naturalist 139: 540-558.
RABELING, C., LINO-NETO, J., CAPPELLARI, S.C., DOS-SANTOS,
I.A., MUELLER, U.G. & BACCI, M.J. 2009: Thelytokous par-
thenogenesis in the fungus-gardening ant Mycocepurus smithii
(Hymenoptera: Formicidae). – Public Library of Science ONE
4: e6781.
RAVARY, F., JAHYNY, B. & JAISSON, P. 2006: Brood stimulation
controls the phasic reproductive cycle of the parthenogenetic
ant Cerapachys biroi. – Insectes Sociaux 53: 20-26.
RAVARY, F. & JAISSON, P. 2002: The reproductive cycle of thely-
tokous colonies of Cerapachys biroi FOREL (Formicidae, Cera-
pachyinae). – Insectes Sociaux 49: 114-119.
RAVARY, F. & JAISSON, P. 2004: Absence of individual sterility in
thelytokous colonies of the ant Cerapachys biroi FOREL (For-
micidae, Cerapachyinae). – Insectes Sociaux: 51: 67-73.
SANADA-MORIMURA, S., MINAI, M., YOKOYAMA, M., HIROTA, T.,
SATO, T. & OBARA, Y. 2003: Encounter-induced hostility to
neighbors in the ant Pristomyrmex pungens. – Behavioral Ecol-
ogy 14: 713-718.
SASAKI, T. & TSUJI, K. 2003: Behavioral property of the unusu-
al large workers in the ant, Pristomyrmex pungens. – Journal
of Ethology 21: 145-151.
SCHILDER, K., HEINZE, J. & HÖLLDOBLER, B. 1999: Colony struc-
ture and reproduction in the thelytokous parthenogenetic ant
Platythyrea punctata (F. SMITH) (Hymenoptera, Formicidae).
– Insectes Sociaux 46: 150-158.
TERANISHI, C. 1929: The habits and distributions of Japanese ants
(II). – Zoological Magazine Tokyo 41: 312-332. (in Japanese)
TIMMERMANIS, I., HEFETZ, A., FOURNIER, D. & ARON, S. 2008:
Population genetic structure, worker reproduction and thelytok-
ous parthenogenesis in the desert ant Cataglyphis sabulosa. –
Heredity 101: 490-498.
TSUJI, K. 1988a: Obligate parthenogenesis and reproductive di-
vision of labor in the Japanese queenless ant Pristomyrmex
pungens: comparison of intranidal and extranidal workers. –
Behavioral Ecology and Sociobiology 23: 247-255.
TSUJI, K. 1988b: Inter-colonial incompatibility and aggressive in-
teractions in Pristomyrmex pungens (Hymenoptera: Formici-
dae). – Journal of Ethology 6: 77-81.
TSUJI, K. 1988c: Nest relocation in the Japanese queenless ant
Pristomyrmex pungens MAYR (Hymenoptera: Formicidae). –
Insectes Sociaux 35: 321-340.
TSUJI, K. 1990a: Reproductive division of labour related to age
in the Japanese queenless ant Pristomyrmex pungens. – Ani-
mal Behaviour 39: 843-849.
98

99
TSUJI, K. 1990b: Kin recognition in Pristomyrmex pungens (Hy-
menoptera: Formicidae): asymmetrical change in acceptance
and rejection due to odour transfer. – Animal Behaviour 40:
306-312.
TSUJI, K. 1992: Sterility for life: applying the concept of eusoci-
ality. – Animal Behaviour 44: 572-573.
TSUJI, K. 1994: Inter-colonial selection for the maintenance of
cooperative breeding in the ant Pristomyrmex pungens: a la-
boratory experiment. – Behavioral Ecology and Sociobiology
35: 109-113.
TSUJI, K. 1995: Reproductive conflicts and levels of selection in
the ant Pristomyrmex pungens: contextual analysis and parti-
tioning of covariance. – The American Naturalist 146: 586-607.
TSUJI, K. & ITÔ, Y. 1986: Territoriality in a queenless ant, Pristo-
myrmex pungens (Hymenoptera, Myrmicinae). – Applied En-
tomology and Zoology 21: 377-381.
TSUJI, K. & YAMAUCHI, K. 1995: Production of females by par-
thenogenesis in the ant Cerapachys biroi. – Insectes Sociaux
42: 333-336.
WANG, M. 2003: A monographic revision of the ant genus Pristo-
myrmex (Hymenoptera: Formicidae). – Bulletin of the Museum
of Comparative Zoology 157: 383-542.
WIERNASZ, D.C., HINES, J., PARKER, D.G. & COLE, B.J. 2008: Mat-
ing for variety increases foraging activity in the harvester ant,
Pogonomyrmex occidentalis. – Molecular Ecology 17: 1137-1144.