A new workerless social parasite in the ant genus Pseudomyrmex (Hymenoptera: Formicidae), with a discussion of the origin of social parasitism in ants.

Systematic Entomology (1996) 21, 253-263

A new workerless social parasite in the ant genus
Pseudomyrmex (HymenopXera: Formicidae), with a
discussion of the origin of social parasitism in ants

PHILIP S. WARD Department of Entomology, University of California at Davis

Abstract. The New World ant genus Pseudomyrmex (subfamily Pseudomyrmecinae)
contains about 180 species, of which only one workerless social parasite, P.leptosus,
from Florida, has been previously recorded. A new species discovered recently in
northern Argentina, P.inquilinus sp. nov., is more derived morphologically and
behaviourally than P.leptosus and has convergently developed features characteristic
of the workerless inquiUnes known in other ant subfamiUes. These features include
diminutive size, reduced mouthparts, a broadened petiole and postpetiole, well-
developed subpetiolar and subpostpetiolar processes, and the habit of straddling the
gaster of the host queen. A cladistic analysis confirms that the new species is not
closely related to its host nor to P.leptosus; it belongs in fact to a different species
complex within the genus Pseudomyrmex. Two widespread Neotropical species,
P.filiformis and P.subater, to which P.inquilinus appears to be related, show indications
of being temporary social parasites on other, unrelated species of Pseudomyrmex.
Thus, there is evidence from comparative natural history that the extreme social
parasitism seen in the new species from Argentina was preceded by a period of
temporary social parasitism, but 'Emery's Rule' - the claim that social parasites are
close relatives of, and evolve from, their hosts - is not supported. Rather, parasitism
in Pseudomyrmex seems to have arisen through interspecific colonization and
exploitation of congeners. A review of the evidence suggests that Emery's Rule has
received too uncritical an acceptance in the literature on social parasitism. Uncertainties
about the generality of Emery's Rule and about the plausibility of achieving prezygotic
isolation under sympatric conditions undermine the theory that social parasites evolve
from within populations of their host species.

Introduction

One of the more intriguing phenomena in ants is the occurrence
of species that depend upon the workers of other species for
the provision of labour. Such social parasites include dulotic
(slave-making) species that raid other ant nests for worker
brood, and temporary social parasites whose colony-founding
queens invade the nests of other species, kill the host queen,
and use the workers as labourers until such time as their
own worker brood has developed. In more extreme cases of
permanent inquilinism, the invading queen produces sexual
offspring only, typically allowing the host queen to survive
and furnish a continuing supply of workers (Wilson, 1971;
Buschinger, 1986; HoUdobler & Wilson, 1990). Instances of
social parasitism have a patchy taxonomic and geographic
occurrence, being well known among north temperate ants in
the subfamilies Formicinae and Myrmicinae, but rather poorly

documented from other taxa and regions of the world (Wilson,
1984; HoUdobler & Wilson, 1990).

An important and largely unchallenged generalization about
social parasitism has emerged: that there is a close phylogenetic
relationship between the parasite and host species (Emery,
1909; Wasmann, 1909; Kutter, 1968; Buschinger, 1970, 1986,
1990; Heinze, 1991; Bourke & Franks, 1991). In a broad sense
this is hardly disputable: most ant social parasites are either
considered congeneric with their hosts or they have been
artifically segregrated into derivative 'satellite genera' that
render the genus of the host species paraphyletic. But the close
morphological resemblance between parasite and host has led
•to the stronger claim, known as Emery's Rule (Le Masne,
1956), that social parasites are more closely related to their
hosts than to any other free-living taxa. To allow for the
possibiUty of subsequent speciation in both the host and parasite
Uneages this rule can be stated as follows: for any monophyletic
group of closely related social parasites the nearest non-

1996 Blackwell Science Ltd

253

254 Philip S. Ward

parasitic outgroup is a clade that includes the host species
(Buschinger, 1990). This assumes, of course, that social
parasitism is an irreversible trait, an assumption which appears
reasonable at least for permanent social parasites.

The apparent veracity of Emery's Rule has in turn inspired
the hypothesis that social parasites originate from their hosts
by sympatric speciation (Buschinger, 1970, 1990; Bourke &
Franks, 1991). This has been argued most cogently for
workerless inquilines. Bourke & Franks (1991) present a
model that posits a facultatively polygynous ancestral species
exhibiting variation in queen size. Small queens (microgynes)
come to specialize in the production of sexual offspring (i.e.
they become intraspecific social parasites), and size-based
assortative mating leads eventually to reproductive isolation
between parasite and host. Apart from raising questions about
the efficacy of disruptive selection in producing prezygotic
isolation (cf. Rice & Hostert, 1994), this model and others like
it (e.g. Buschinger, 1990) rest on the assumption that Emery's
Rule is generally valid. Yet, there have been surprisingly few
explicit phylogenetic tests of this assumption (Ward, 1989;
Agosti, 1994).

This paper describes a new workerless social parasite in the
Neotropical ant genus Pseudomyrmex. It is the second known
example in its subfamily (Pseudomyrmecinae), and it is
considerably more modified in appearance than the only other
workerless inquiline heretofore recorded in the group. At the
same time it provides an opportunity to determine whether
Emery's Rule applies in this case, and to reexamine the
applicability of the 'rule' to other taxa. The results caution
against the uncritical acceptance of Emery's Rule and suggest
the need to reevaluate the theory that social parasites evolve
sympatrically from their hosts.

Materials and Methods

The new species described herein was discovered during
fieldwork in northern Argentina, in late January and early
February, 1995. Other species examined in this study were
collected at various times during a long-term investigation of
the genus Pseudomyrmex in the Nootropics. Material in museum
collections has also been studied. The abbreviations for museum
collections cited here are listed in Amett etal. (1993) and
Ward (1993).

Metric measurements were carried out as described in Ward
(1989a, 1993). The following measurements and indices are
used here: HW (head width), HL (head length), EL (eye
length), EW (eye width), MFC (minimum distance between
the frontal carinae), MDl (basal width of the mandible), MD2
(width of mandible at the juncture of the basal and masticatory
margins), MD3 (mandible length), EW (eye width), SL (scape
length), FL (profemur length), FW (profemur width), PL
(petiole length), PH (petiole height), DPW (dorsal petiolar
width), MPW (minimum petiolar width), PPL (postpetiole
length), PPW (postpetiole width), LHT (length of metatibia),
CI (cephalic index: HW/HL), OI (ocular index: EW/EL), REL
(relative eye length: EL/HL), REL2 (relative eye length, using
HW: EL/HW), FCI (frontal carinal index: MFC/HW), SI (scape

Table 1 . Data set used for cladistic analysis of Pseudomyrmex
inquilinus and other taxa. '?' signifies unknown, polymorphic or
inapplicable. See text for description of character states.

00000

00001

mil

nil

12345

67890

12345

^89?

Tetraponera

0000?

0000?

00000

00?0

Pseudomyrmex inquilinus

1100?

?20U

00001

0100

Pseudomyrmex subater

01100

12011

00001

1201

Pseudomyrmex duckei

01100

10010

00001

1201

Pseudomyrmex filiformis

01100

01010

00001

0000

Pseudomyrmex oki

OHIO

00010

11100

0010

Pseudomyrmex gebelli

oun

01010

11100

0010

Pseudomyrmex pallidus group

01111

01010

12110

0010

Pseudomyrmex viduus group

01100

?011?

00000

0000

index: SL/HW), FI (profemur index: FW/FL), PLI (petiole
length index: PH/PL), PWI (petiole width index: DPW/PL),
and PPWI (postpetiolar width index: PPW/PPL). For further
details see Ward (1989a).

The procedure used for cladistic analysis was as follows. A
data set of nineteen characters was developed for the social
parasite, its host species, and allied taxa. Further discussion of
the choice of terminal taxa, including the outgroup, appears later
in this paper (under 'Phylogenetic affinities of Pseudomyrmex
inquilinus'). Emphasis was placed on features previously found
to be useful at the species and species group level in
Pseudomyrmex (Ward, 1989a, 1991, 1993). The characters and
character states used for cladistic analysis are listed below, and
the data matrix is given in Table 1 . The data set was analysed
using HENNIG86, version 1.5 (Farris, 1988).

1 . Queen and male: mandible, masticatory margin (0) straight,
dentate, (1) concave, edentate (Figs 3 and 4).

2. Queen and worker: mandible, ventral ridge (0) absent,
(1) present.

3. Queen and worker: basal margin of mandible, proximal
tooth (0) absent, (1) present.

4. Queen and worker: median clypeal lobe, laterally (0)
rounded, (1) angulate.

5. Worker: standing pilosity on mesosoma dorsum (0)
common (> 11 setae visible in profile), (1) sparse (< 9 setae
visible in profile).

6. Worker: standing pilosity on mesosoma (0) sparse to
common, not underlain by dense subdecumbent pilosity, (1)
abundant, and underlain by dense subdecumbent pilosity, which
gives worker a 'scruffy' appearance.

7. Queen: standing pilosity on dorsum of body (0) common,
conspicuous in profile on head, mesosoma, petiole and
postpetiole, (1) sparse: 2-3 pairs of setae on pronotum, none
visible in profile on propodeum, 1-3 pairs on petiole and
postpetiole, (2) very sparse, absent from pronotum, propodeum,
petiole and postpetiole.

8. Queen and worker: standing pilosity on extensor faces of
mid- and hind-tibiae (0) sparse (fewer than 8 standing hairs),
(1) common (> 10 hairs).

9. Queen and worker: hind basitarsal sulcus (0) present,
(1) absent.

© 1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

New

social parasite in the ant genus Pseudomyrmex 255

Figs 1-4. Pseudomyrmex inquilinus, sp. nov. Fig. 1, lateral view of holotype queen (excluding appendages and most of gaster); Fig. 2, dorsal
view of holotype queen, excluding head; Fig. 3, full-face, dorsal view of head, excluding antennae, holotype queen; Fig. 4, same, paratype male.

10. Queen: profemur (0) relatively broad (FT &0.42), (1)
slender (FI < 0.42).

11. Male: posterolateral comers of sterna VI, VII and VIII (0)
rounded, not produced, (1) angulate and produced ventrally.

12. Male: pygidium, posterior margin (0) directed
posteroventrally, not notably recurved, (1) recurved, directed
ventrally, (2) strongly recurved, forming an anteroventrally
directed pocket.

13. Male: paramere, mesal dorsoventral lobe (0) attached
continuously to inner wall of paramere, (1) in the form of an
isolated digitiform structure attached to inner wall of paramere
at base only (Fig. 14).

14. Male: paramere, large thin posterodorsal lobe (0) absent,
(1) present (Figs 13 and 14).

15. Male: paramere, distal end, inner face (0) without a
dorsomesally or posterodorsomesally directed, irregular
concavity, bordered anteriorly by a mesal dorsoventral lobe
which is connected by a low saddle to the highest point of the
posterodorsal ridge, (1) of such a form (Figs 15, 17, 19, 21).

16. Male: paramere, distal end (0) highest point of
posterodorsal ridge not developed as truncate, subcorneal
protrusion (as seen in lateral profile), (1) so developed
(Figs 7, 9).

17. Male: paramere, distal end (0) without broadly rounded
(semicirculEir) distal extremity, (1) of such a form (Figs 7, 9).

18. Male: aedeagus, outer face (0) without an anterodorsal to
posteroventral impression, bordered anteriorly by an oblique
ridge and posteriorly by a carina or lamella that terminates
near a small incision or tooth on the posterior margin of the
aedeagus, (1) of such a form, the posterior border a denticulate
carina which does not cover most of the posterior margin, in
lateral view, (2) of such a form, the posterior border a lamella
which forms most of the posterior margin.

19. Male: aedeagus, posterior margin, (0) entire or with at

most a single distinct tooth, (1) with several (six) to many
teeth (Fig. 24).

Where characters varied within one of the supraspecific taxa
(P.pallidus group, P.viduus group and genus Tetraponera) they
were coded as unknown, except for character 4 (the shape of
the median clypeal lobe). For this character the groundplan of
the P.viduus group was inferred to be state 0, and the angulate
condition seen in one derivative and nonbasal species,
P.tachigaliae, is presumed to be derived within the group.

Descriptive taxonomy

Pseudomyrmex inquilinus, sp. n. (Figs 1^, 11, 12, 21, 22)

Holotype queen. ARGENTINA, Catamarca: Santa Maria,
2000 m, 26°42'S, 66°03'W, 2 February 1995; P. S. Ward ace.
no. 12845; in nest of host species, Pseudomyrmex sp. PSW-64
(Ward ace. no. 12844), in dead twig of Baccharis sp. HW
0.63, HL 0.74, EL 0.36, LHT 0.48. (MCZC)

Paratypes. Series of (i) eight queens and one male from
same host nest as the holotype (ii) two queens from a second
nest, same host species, locality and date. (BMNH, IMLA,
LACM, MZSP, PSWC)

Queen measurements (n = 7). HW 0.63-0.65, HL 0.74-
0.78, EL 0.36-0.38, DPW 0.29-0.33, PPW 0.41-0.46, LHT
0.47-0.50, CI 0.83-0.87, OI 0.60-0.62, REL 0.49-0.50, REL2
0.57-0.59, FCI 0.091-0.115, SI 0.40-0.45, FI 0.37-0.40, PLI
1.12-1.27, PWI 1.08-1.20, PPWI 1.65-1.83.

Description, queen. Of diminutive size for the genus
Pseudomyrmex (HW < 0.68, LHT < 0.52). Mandible relatively
broad (MD2/MD3 =0.55), with basal and external margins
subparallel (MD1/MD2 =0.96); basal and masticatory margins
distinctly differentiated, the former lacking teeth or denticles.

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

256 Philip S. Ward

1

11

12

h

Figs 5-14. Pseudomyrmex males, left parameres, lateral (left) and mesal (right) views. Figs 5-6, P.filiformis (Peru); Figs 7-8, P.duckei (Panama);
Figs 9-10, P.subater (Panama); Figs 11-12, P.inquilinus (Argentina); Figs 13-14, P. sp. PSW-64 (Argentina). Scale line equals 0.25 mm for Figs 5-
8, 13-14, and 0.20 mm for Figs 9-12.

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

New social parasite in the ant genus Pseudomyrmex 257

16

21

22

\-

Figs 15-24. Pseudomyrmex males, dorsal views of left paramere (left) paired with left lateral views of the aedaegus (right). Figs 15-16, P.filiformis;
Figs 17-18, P.duckei; Figs 19-20, P.subater; Figs 21-22, P.inquilinus; Figs 23-24, P. sp. PSW-64. These are the same males depicted in Figs 5-
14. Scale line equals 0.25 mm for Figs 15-18, 23-24 and 0.20 mm for Figs 19-22.

except for an acute apicobasal tooth; masticatory margin
strongly concave and essentially edentate (Fig. 3), at best very
weakly crenulate, with a small denticle, discemable only at
high magnification, at about two thirds of the distance between
the apicobasal tooth and the apical tooth. Closed mandibles

with a gap (0.03-0.04 mm wide) between the midpoints of their
masticatory margins as a consequence of the aforementioned
concavity. Palp formula 3,2. Median clypeal lobe broadly
rounded, and lacking a distinctive transverse truncation on the
upper surface (typical of nearly all other Pseudomyrmex - see

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

258 Philip S. Ward

Ward, 1990). Frontal carinae well separated, the distance
between them subequal to the maximum scape width
(FCI « 0.10). Head moderately elongate (CI 0.83-0.87), with
weakly convex sides and a straight to shallowly concave
posterior margin, in frontal view (Fig. 3). Evidently fully
winged: all known specimens dealate, with a full complement
of thoracic sclerites (and queen pupa seen to have wings).
Metapleural gland opening and bulla reduced, inconspicuous.
Legs relatively short, the profemur in particular small and
slender (FL 0.44-0.46, FI 0.37-0.40). Meso- and metatibiae
each with a pair of apical spurs, the posterior one of each pair
conspicuously pectinate. In dorsal view propodeum
subtrapezoidal, the sides converging posteriorly; in lateral view,
the basal (dorsal) face of propodeum with a long weakly
declining surface, followed by a shorter steeper portion, which
is nevertheless well differentiated from the short, vertical
declivitous face of the propodeum (Fig. 1). Petiole short,
broad, apedunculate, submarginate laterally, and with a large
conspicuous ventral keel that protrudes posteroventrally
(Fig. 1). Dorsal face of petiole flat, with a slight median
depression; posterior face absent, owing to an expanded
helcium. Postpetiole short and broad, with a bulging ventral
protrusion that is thick and wide, not keel-like. Remainder of
gaster about the same length as the mesosoma.

Mandible sublucid, with very weak, almost obsolete,
transverse striolae. Head moderately shiny, with numerous
punctulae, about 0.01 mm in diameter or less, separated by
one to several diameters, on a finely coriarious background
scupture. Mesoscutum and mesoscutellum similarly sculptured,
punctures more scattered; in contrast, pronotum, propodeum
and petiole densely coriarious-reticulate, and subopaque,
especially the propodeum; epistemum and sides of pronotum
less strongly sculptured. Postpetiole and gaster sublucid, with
scattered, very fine punctures on a weakly coriarious
background.

Standing pilosity very scarce, essentially absent from the
scapes (except apex), head capsule (above the clypeus),
mesosoma, legs, petiole, postpetiole and fourth abdominal (first
gastric) tergite. A single short seta present consistently on each
side of the mesoscutum, near the lateral margin. Gastric pilosity
beginning on the fifth abdominal tergite (or on the posterior
margin of the fourth), and becoming long and dense near the
apex (Fig. 2). Very fine, short, submicroscopic pubescence
present on much of the body but scarcely visible at normal
magnification (50-lOOX). Body orange-brown, mesoscutellum
and metanotum with variable darker infuscation; mandibles,
clypeus and gula pale yellow-brown.

Description, male. Mandibles similar to those of queen;
masticatory margin edentate and with same peculiar concave
edge, although not as pronounced (Fig. 4). Palp formula 3,2
as in queen. Head broad (HW 0.65, CI 0.96), and posterior
margin very rounded (Fig. 4). Eyes and ocelli not strongly
protruding from head capsule. Scape short, subequal in length
to second funicular segment. Total length of antenna about
1.4 mm. Propodeum, petiole and postpetiole similar to those
of queen, but less extreme in shape: propodeal faces less
distinct, petiole and postpetiole less broad, and petiole less
strongly margined. Subgenital plate much wider than long, its

posterior margin shallowly and broadly concave. Pygidium
with a truncate posterior margin, not recurved anteroventrally.
Paramere as in Figs 11, 12, 21; mesal dorsoventral lobe
developed as an obUque, rounded, dorsomesal protrusion,
joined through a low saddle to a posterodorsal ridge, the surface
between the two concave and directed posterodorsomesally
(Fig. 21). Aedeagal plate simple, somewhat circular in outline,
outer margin entire except for a weak medial incision and one
or two adjacent, very small denticulae. Outer surface of
aedeagus with a broad oblique impression (anterodorsal to
posteroventral), bordered on either side by corresponding weak
ridges, the anterior ridge situated about midway between the
posterior margin and the lateral apodeme, the posterior ridge
merging with the posterior margin above the level of the afore-
mentioned incision (Fig. 22). Body sculpture paralleling that of
queen, but integument a little less shiny; subopaque, shagreened
areas on the head between the antennal insertions and in the
ocellar triangle; mesoscutum and mesoscutellum with coarser
sculpture than in queen. Standing pilosity very scarce,
distributed as in the queen. Pubescence slightly more noticeable
than in queen, but still rather inconspicuous. Dark brown, with
paler, transverse yellow-brown bands on the posterior margins
of the gastric segments; mandibles, fronto-clypeal complex,
gula and forecoxa constrastingly pale, almost white; antennae
and legs light brown.

Worker. The worker caste is unknown and probably non-
existent.

Comments. The queen of this species is immediately
distinguishable from all known congeners by any one of the
following features: the concave and edentate masticatory margin
of the mandibles, the reduced palp formula (3,2), the shape of
the petiole (especially the lack of a posterior face), shape of
the postpetiole (short, broad and deep), and the lack of standing
pilosity on nearly all parts of the body except the mouthparts
and apex. Also distinctive are the edentate basal margin of the
mandible, the lack of a transverse truncation on the median
clypeal lobe, the well separated frontal carinae, the slender
forefemur, the short legs, and the long basal face of the
propodeum. The male is diagnosable by the mandible shape,
palp formula, lack of a posterior face on the petiole, highly
reduced standing pilosity, and the male genitalia (see
description above).

Host species. Based on worker and queen morphology, as
well as male genitalia, the host species is easily recognizable
as a member of the P.pallidus group (diagnosis in Ward,
1989a). Within this group it belongs to a taxonomically vexing
assemblage of species that may be termed the P.flavidulus
complex. Members of this complex are characterized by the
following combination of character states: male paramere with
large posterodorsal lobe preceded anteriorly by a much smaller
finger-like lobe (Figs 13 and 14); abdominal tergite IV of
worker and queen densely pubescent; head relatively elongate
(worker CI ^ 0.90); worker profemur relatively slender (worker
FI s 0.47); and worker and queen predominantly orange or
orange-brown in colour. There are several species within the
P.flavidulus complex, but the limits of intra- and interspecific
variation are not yet clear. For the moment the host species is
referred to using a code number, Pseudomyrmex sp. PSW-64.

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

New social parasite in the ant genus Pseudomyrmex 259

Table 2. Contents of two nests of Pseudomyrmex sp. PSW-64 occupied
by the parasitic species, P.inquilinus.

Nest no. 12841

Nest no. 12844

Host species

Dealate queens

1

1

Workers

19

22

Alate queens

5

Alate males

3

6

Worker pupae

2

present

Queen pupae

8

Male pupae

present

24

Parasite

Dealate queens

2

9

Alate males

1

Queen pupae

1

Queen prepupae

4

Indeterminate

Eggs, larvae

present

present

As currently interpreted this species is known only from
Catamarca and Tucuman provinces in Argentina, although a
related (and possibly conspecific) form occurs in Bolivia. The
flavidulus complex as a whole is widely distributed and common
in the Neotropics, from Costa Rica to Argentina.

Biology, behaviour. The host species was found nesting in
dead stems of Baccharis salicifolia and a second Baccharis
species, probably B.angulata, in pastured riparian grassland
next to the Rio Santa Maria. The Baccharis species are dominant
in this community; other plants include Salix humboldtiana,
Prosopis, Juncus, and various grasses. The town of Santa
Maria has an average annual rainfall of 182 mm and is
located in the Argentinian desert biome known as monte
(Morello, 1958).

Pseudomyrmex inquilinus queens were recovered from two
out of six nests of the host species. The contents of the two
parasitized nests are summarized in Table 2. Both nests
contained a dealate host queen and sexual pupae of the
host species. Thus P.inquilinus was not inhibiting sexual
reproduction in its host.

Brief observations on behaviour were made in the field
during the collection of nest no. 1 2844. Most of the P. inquilinus
queens were close to the host queen. One was riding on the
back of the host queen gaster, holding on with her mandibles
around the anterior peduncle of the petiole. (The concave
masticatory margins of the P.inquilinus mandibles make them
well-suited for this task.) As the nest was broken open
Pseudomyrmex host workers were seen encountering
P.inquilinus queens and no antagonistic behaviour ensued. The
single parasite male was in the upper part of the twig nest,
where most of the host males were concentrated, while the
nine parasite queens and the host queen occurred in the lower
portions of the nest. The entire host nest occupied a hollow
section of the dead twig about 0.5 m in length (out of a total
length of about 1 m).

At the type locality (Santa Maria) and at neighbouring sites

near Amaicha del Valle (Tucuman Province) two species of
Pseudomyrmex were found to be common inhabitants of dead
Baccharis stems, both members of the P.pallidus group: P. sp.
PSW-64 (six nests collected, two parasitized by P.inquilinus)
and P.rufiventris (Forel) (five nests sampled, none parasitized).
The sample sizes are too small to draw any firm conclusions
about host specificity. During a 9-day period of collecting
Pseudomyrmex in northern Argentina (in Cordoba, Catamarca
and Tucuman provinces) a total of 24 nests of P.rufiventris
was sampled and none was parasitized. Both the parasite and
its only known host (P. sp. PSW-64) were found only in the
vicinity of Santa Maria.

Phylogenetic affinities of Pseudomyrmex inquilinus

Pseudomyrmex inquilinus bears a superficial resemblance to
its host species: the queen is the same bright orange-brown
colour, standing pilosity is scarce on the head and mesosoma,
the integument has a similar lustrous sheen, head shape is
similar, and the male is dark brown with light appendages, as
in the host species. Hence, the first impression in the field was
that this was a member of the P.pallidus group, modified for
life as a social parasite on another member of the same group,
as in the case of Pseudomyrmex leptosus, a workerless inquiline
from Florida (Ward, 1985; Klein, 1987) and the only other
example known in the genus and subfamily.

More careful scrutiny of P. inquilinus soon forced the
abandonment of this idea. It proved to possess none of the
apomorphic characteristics of the P.pallidus group. In fact it is
not placeable in any of the nine major species groups which
have been recognized in the genus (Ward, 1989a). The nine
species groups are reasonably well defined and most or all of
them are likely to be monophyletic. Collectively they
encompass about 85% of the 180-odd species in the genus.
There is a remaining heterogeneous assemblage of about
twenty-five unplaced (incertae sedis) species, and P. inquilinus
is phenetically most similar to a small subset of these, centred
around P.subater (Wheeler & Mann) and Pduckei (Forel).

Thus it would seem to be a foregone conclusion that Emery's
Rule is contradicted here. None the less it could be salvaged -
albeit with special pleading -if P.inquilinus could be shown
to be a sister group of the P.pallidus group. While an
investigation of the phylogenetic relationships of all 180 or
more Pseudomyrmex species was beyond the scope of this
study, a subset of taxa was chosen, appropriate for probing the
cladistic relationships of parasite and host. These taxa included
the P.pallidus group, two species (P.gebelli, P.oki) allied to that
group, P.inquilinus, and three taxonomically isolated species
(P.duckei, P.filiformis, P.subater), with suggestive similarities
to P.inquilinus. Also included was the Pviduus group, which
appeared in a basal position in an earlier cladistic analysis of
the genus (Ward, 1991), and the genus Tetraponera which
served as an outgroup. Tetraponera is the probable sister genus
of Pseudomyrmex (Ward, 1990). Note that because some species
and species groups of Pseudomyrmex are being excluded this
test is biased in favour of support of Emery's Rule.

A data set of nineteen morphological characters was

© 1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

260 Philip S. Ward

7.1 15

18.1

=4=

1 3 7.2 10

+4=++=

7.0

6 16 17 18.2

7.2 10

I I I I " n

4 11 12.1 13 19

I M [ I

5 7.1

12.2 14

=+H=

Tetraponera

viduus grp

filiformis

inquilinus (parasite)

duckei

subater

old

gebelli

pallidus grp (40 spp., incl. host)

Fig. 25. Phylogenetic relationships among Pseudomyrmex inquilinus, its host's species group {P.pallidus group), and other Pseudomyrmex taxa.
This is one of three most parsimonious trees, and also the strict consensus of those three trees. Character state changes are marked, with solid bars
indicating unique forward changes, cross-hatched bars convergences, and open bars reversals. Numbers refer to the characters (see text under
'Materials and methods'). This tree has a length of 27, consistency index 0.81.

developed, of which ten were worker- or queen-based, and nine
were male-based (primarily male genital characters) (Table 1).
Cladistic analysis of the data set using the ie (implicit
enumeration) option of hennig86 yielded three most
parsimonious trees of length 27 and consistency index (CI)
0.81. The consistency index drops to 0.77 if one excludes
autapomorphies of the terminal taxa (three characters) and
non-reversing synapomorphies of the ingroup (two characters).
Figure 25 illustrates the strict consensus tree; it is also one of
the three most parsimonious trees, the other two involving
resolutions of the basal trichotomy.

The results confirm that P.inquilinus does not have close
phylogenetic affinities to its host species or its host species
group. It appears to be more closely related to P.subater
and P.duckei, two of the taxonomically isolated species not
belonging to any of the larger Pseudomyrmex species groups.
The most parsimonious trees place P.inquilinus in a clade with
P.subater and P.duckei. P.filiformis is a sister species to these
three, and together they can be said to form a rather weakly
defined 'P.filiformis complex'. The two species, P.gebelli and
P.oki, placed in the incertae sedis group of Pseudomyrmex
(Ward, 1989a), appear as successive outgroups of the pallidus
group. Finally, relationships among the viduus group, the
filiformis complex, and [oki + {gebelli + pallidus group)] are
not well resolved, leaving a basal trichotomy.

Discussion

The preceding analysis, although based on an incomplete set
of taxa and a relatively small number of characters, nevertheless
indicates rather clearly that, in the case of P.inquilinus, there
is not a recent sister-group relationship between the parasite
and host lineages. As such, one can exclude the possibility
that P.inquilinus arose by sympatric speciation from its current
host or an immediate ancestor of that host. The results leave

open the question of the origin of parasitic habits in P.inquilinus.
It may be possible to gain some insight into this problem by
examining the natural histories of three species that are
apparently related to P.inquilinus, namely P.filiformis, P.subater
and P.duckei.

P.duckei is a rarely encountered rainforest species ranging
from Guatemala to Bolivia. Only one queenright colony has
been collected by the author: it was monogynous and
unremarkable in composition. The queens of P.duckei are
'normal' in morphology, i.e. similar to workers but larger in
size, and they give no obvious indication of being parasitic at
any stage of their life cycle. P.subater is widely but patchily
distributed from Mexico and the West Indies south to Bolivia
and Brazil. The queens of this species are unusually small,
being about the same size as the workers but differing in the
scarcity of their pilosity. Little is known about the biology of
P.subater, but field observations in the dry forests of western
Mexico (Jalisco) and the Dominican Republic suggest that it
is a rather aggressive species that is both polygynous and
polydomous (P. S. Ward, unpublished data). Collections of this
species from the Bahamas, sent to the author by Blaine Cole,
included a heterospecific nest sample, containing workers of
both P.subater and P.cubaensis (Forel), the latter being an
unrelated species in the P.oculatus group (Ward, 1985, 1989a).
This suggests the possibility of parasitic behaviour on the
part of P.subater. Moreover, old references to mixed-nest
associations of Bahamian P.'flavidula' and P'elongata'
(Wheeler, 1905; Mann, 1920) can be reinterpreted in the light
of the preceding observations. Examination of museum material
shows that Wheeler's and Mann's records of P. 'elongata' refer
to a mixture of P.subater and P.cubaensis (Ward, 1985; present
study). Pfiavidulus does not occur in the West Indies and the
records of 'flavidula' are mostly misidentifications of P.pallidus
(Ward, 1985). It is also true, however, that queens of Bahamian
P.subater are a bright-orange colour (in contrast to the dark
brown of the workers) that makes them look superficially like

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

New social parasite in the ant genus Pseudomyrmex 261

Table 3. Social parasites in the ant genus Pseudomyrmex.

Parasite

Host

Type of parasitism

Reference

P. Seminole
P.fiUformis

P.subater
P.leptosus

P. inquilinus

P.pallidus

P.elongatus, P.holmgreni
P. sp. PSW-14
P.cubaensis, ?P.pallidus
P.ejectus

P sp. PSW-64

Temporary social parasitism
Temporary social parasitism

Temporary social parasitism (?)
Workerless inquilinism,
intolerant of host queen
Workerless inquilinism,
tolerant of host queen

Ward, 1985
Present study

Ward, 1985; present study
Ward, 1985; Klein, 1987

Present study

those of P.pallidus. So there is circumstantial evidence to
indicate that, at least in the Bahamas, P.subater parasitizes the
nests of P.cubaensis (and perhaps also those of P.pallidus).
Because monospecific nests of P.subater are also encountered,
in the West Indies and elsewhere, the behaviour apparently
represents temporary social parasitism or facultative dulosis.
The situation begs for a detailed study.

For P.fiUformis there is more definitive evidence of temporary
social parasitism. This species ranges from Mexico to Peru,
Bolivia and Brazil. The queens are peculiar in appearance,
with strikingly elongate heads and sturdy mandibles (Wheeler,
1919; Fig. 2). Over the last 10 years the author has collected
about twenty-four nest series of P.fiUformis. Most of these
were large and mature colonies, monospecific in composition,
and containing no more than a single functional queen (one
colony had two physogastric queens). The author has
encountered only three incipient colonies of P.fiUformis, and
all involved dealate queens cohabiting with several (4-8)
workers of a different species, in small dead twigs. The host
species were P. sp. PSW-14 (an undescribed species near
P.ejectus), P.holmgreni (a member of the P.pallidus group, like
the first species), and P.elongatus (Mayr) (in the P.oculatus
group), and they were accompanied by one, two and one
dealate queens, respectively, of P.fiUformis. Most of the brood
could not be identified to species, but in all three instances it
included some worker pupae of the host species. No host
queens were present. The host species are common and
widespread in Central and/or South America.

Thus, two of the three species to which P.inquilinus appears
to be related exhibit evidence of being temporary social
parasites on common but distantly related congeners. This
suggests a comparable intermediate stage preceding the
evolution of the workerless condition in P.inquilinus. It would
be useful to have more detailed phylogenetic and behavioural
data on the species allied to P.fiUformis, of which there are
several undescribed species in addition to P.duckei and P.subater
(P. S. Ward, unpublished data). If P.duckei is both non-parasitic
and a sister species of P.subater (as in Fig. 25) then, in view
of the likely irreversibility of the parasitic habit, one would
have to postulate multiple origins of temporary social parasitism
in this group of ants.

The only other workerless social parasite known in the genus
Pseudomyrmex, P.leptosus Ward from Florida, is a member of
the P.pallidus group (Ward, 1985) and hence represents an
origination of the workerless condition separate from that of

P.inquilinus. The host species of P.leptosus is P.ejectus (F.
Smith), a member of the same species group. A cladistic
analysis indicated that P.leptosus is not a sister species of its
host; rather it is more closely related to two non-host species,
P.pallidus and Pseminole (Ward, 1989b). Interestingly,
P.seminole is a facultative temporary social parasite of
P.pallidus (Ward, 1985). One important difference between
P.leptosus and P.inquilinus is that the former species apparently
does not tolerate the presence of the host queen (Klein, 1987).
To summarize, there are now two known independent
originations of workerless inquilinism in the ant genus
Pseudomyrmex, and in both cases phylogenetic analysis
indicates that (i) the parasite is more closely related to other
free-living taxa than to its host, and (ii) among the parasite's
close relatives are some species exhibiting habits of temporary
social parasitism. Although the comparative data are still
sparse, the most plausible route to permanent social parasitism
in Pseudomyrmex involves interspecific exploitation of
congeners. Looking at the known social parasites in this genus
we observe a suggestive ethocline from temporary social
parasitism (P.seminole, P.fiUformis, and probably P.subater) to
workerless, queen-intolerant parasitism (P.leptosus) to queen-
tolerant inquilinism (P.inquilinus) (Table 3). The known hosts
of these parasites are common, facultatively polygynous species
drawn from two different species groups (P.oculatus group,
P.pallidus group). Most unparasitized colonies of the host
species contain a single functional queen, but multiple-queen
colonies are encountered at least occasionally. Insofar as this
polygyny is due to readoption of recently mated daughter
queens into the nest - the pattern observed in most polygynous
ants (Buschinger, 1990; Holldobler & Wilson, 1977, 1990)-
these species may be more vulnerable than strictly monogynous
species to interspecific (congeneric) parasitism.

Concluding remarks

In all three instances where cladistic methods have been used
to test the validity of Emery's Rule in ants, it has been found
wanting (Table 4). A fourth study, involving socially parasitic
polistine wasps, also concluded that the rule did not hold
(Carpenter etal, 1993; see also Choudhary etal., 1994). To
the extent that other studies fail to support Emery's Rule, the
theory that social parasites evolve sympatrically from their
hosts is seriously undermined. Because social parasitism has

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

262 Philip S. Ward

Table 4. Cladistic tests of Emery's Rule in ants.

Parasite

Host

Emery's Rule supported?

Reference

Pseudomyrmex leptosus
Pseudomyrmex inquilinus
Cataglyphis hannae

P.pallidus
P sp. PSW-64
C.bicolor

No
No
No

Ward, 1989b
Present study
Agosti, 1994

arisen numerous times in ants (and in other social Hymenoptera)
it is certainly too early to issue a blanket rejection of Emery's
Rule. There is no reason to believe that the ant taxa that have
been tested {Pseudomyrmex, subfamily Pseudomyrmecinae;
Cataglyphis, subfamily Formicinae) are representative of the
diverse array of social parasites seen among the ants. At the
very least, however, the rule should be treated with greater
circumspection. More detailed taxonomic, phylogenetic and
genetic investigations are needed of other groups of social
parasites, particularly in the species-rich genera Myrmica and
Leptothorax (sensu lato). A recent promising study by Baur
etal. (1996), using sequences from the ITS-1 region near the
5.8S rRNA gene, demonstrated a relationship at the level of
tribes and species groups between several myrmicine social
parasites and their respective hosts, but the low level of
sequence variation among closely related species and the
incomplete sampling of relevant taxa (admittedly a complete
survey would be a daunting task) precluded effective
phylogenetic inference at the species level.

A final point of discussion concerns the plausibility of the
sympatric speciation model. Even if Emery's Rule is found to
be valid for some ant parasites, simple alternative explanations
involving allopatric differentiation are available (Wilson, 1971;
HoUdobler & Wilson, 1990; see also Pearson, 1981). The
sympatric models, favoured by Buschinger (1970, 1990) and
Bourke & Franks (1991) (see also West-Eberhard, 1986), face
the obstacle of producing prezygotic isolation by disruptive
selection in the presence of continuing gene flow. The
antagonism between selection and recombination that
jeopardizes this process (Felsenstein, 1981) would be
manifested in the case of an incipient microgyne parasite by
the difficulty in building up linkage disequilibrium among
genes affecting three different traits: selfish (parasitic)
behaviour, body size, and mating habits. This could be
ameliorated if positive assortative mating were linked by
pleiotropy to the disruptively selected trait(s) (Rice & Hostert,
1994), but it is difficult to imagine that choice of mating venue
(e.g. near the nest rather than in a mating swarm) and therefore
mate would be ineluctably tied to parasitic behaviour and small
body size (cf. Bourke & Franks, 1991). In a species that
remains facultatively polygynous one would expect that some
non-parasitic queens would continue to mate near, and return
to, the nest. Moreover, dispersal and colonization of new nests
would have to persist (or develop?) as a behaviour in the
proto-parasite. Finally, the parasitic reproductive behaviour is
unexpressed as a male phenotype, further diluting the strength
of disruptive selection. In the light of these considerations, the
intraspecific route to interspecific social parasitism remains
problematic.

Acknowledgements

I am grateful to Abraham Willink and Fabianna Cuezzo for
facilitating my fieldwork in northern Argentina, and to Barry
Bolton, Sean Brady and an anonymous referee for helpful
comments. This study was supported by NSF DEB 9419242.

References

Agosti, D. (1994) A new inquiline ant (Hymenoptera: Formicidae) in

Cataglyphis and its phylogenetic relationship. Journal of Natural

History, 28, 913-919.
Amett, R.H., Jr., Samuelson, G.A., Nishida, G.M. (1993) The Insect

and Spider Collections of the World, 2nd edn. Sandhill Crane Press,

Gainesville, vi -H 310 pp.
Baur, A., Sanetra, M., Chalwatzis, N., Buschinger, A., Zimmermann,

F.K. (1996) Sequence comparisons of the internal transcribed spacer

region of ribosomal genes support close relationships between

parasitic ants and their respective host species (Hymenoptera:

Formicidae). Insectes Sociaux, 43, 53-67.
Bourke, A.F.G., Franks, N.R. (1991) Alternative adaptations, sympatric

speciation and the evolution of parasitic, inquiline ants. Biological

Journal of the Linnean Society, 43, 157-178.
Buschinger, A. (1970) Neue Vorstellungen zur Evolution des

Sozialparasitismus und der Dulosis bei Ameisen (Hym., Formicidae).

Biologisches Zentralblatt, 89, 273-299.
Buschinger, A. (1986) Evolution of social parasitism in ants. Trends

in Ecology and Evolution, 1, 155-160.
Buschinger, A. (1990) Sympatric speciation and radiative evolution

of socially parasitic ants - Heretic hypotheses and their factual

background. Zeitschrift filr Zoologische Systematilc und

Evolutionsforschung, 28, 241—260.
Carpenter, J.M., Strassmann, J.E., Turillazzi, S., Hughes, C.R., Soli's,

C.R., Cervo, R. (1993) Phylogenetic relationships among paper

wasp social parasites and their hosts (Hymenoptera: Vespidae;

Polistinae). Cladistics, 9, 129-146.
Choudhary, M., Strassmann, J.E., Queller, D.C., Turillazzi, S., Cervo,

R. (1994) Social parasites in polistine wasps are monophyletic:

implications for sympatric speciation. Proceedings of the Royal

Society of London, Series B, 257, 31-35.
Emery, C. (1909) Uber den Ursprung der dulotischen, parasitischen und

myrmekophilen Ameisen. Biologisches Centralblatt, 29, 352-362.
Farris, J.S. (1988) Hennig86 reference. Version 1.5. Published by the

author. Port Jefferson, New York.
Felsenstein, J. (1981) Skepticism towards Santa Rosalia, or why are

there so few kinds of animals? Evolution, 35, 124-138.
Heinze, J. (1991) Biochemical studies on the relationship between

socially parasitic ants and their hosts. Biochemical Systematics and

Ecology, 19, 195-206.
HoUdobler, B., Wilson, E.O. (1977) The number of queens: an

important trait in ant evolution. Naturwissenschaften, 64, 8-15.

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263

New social parasite in the ant genus Pseudomyrmex 263

Holldobler, B., Wilson, E.O. (1990) The Ants. Harvard University

Press, Cambridge, MA. xii + 732 pp.
Klein, R. (1987) A workerless inquiline in Pseudomyrmex

(Hymenoptera: Formicidae: Pseudomyrmecinae). In: Chemistry and

Biology of Social Insects (ed. by J. Eder and H. Rembold), pp. 623-

624. Verlag J. Pepemy, Munich.
Kutter, H. (1968) Die sozialparasitischen Ameisen der Schweiz.

Neujahrsblatt. Naturforschende Gesellschaft in Zurich, 171, 1-62.
Le Masne, G. (1956) Recherches sur les fourmis parasites. Plagiolepis

grassei et revolution des Plagiolepis parasites. Comptes Rendus

des Seances de I'Academie des Sciences, 243, 673-675.
Mann, W.M. (1920) Additions to the ant fauna of the West Indies and

Central America. Bulletin of the American Museum of Natural

History, 42, 403^39.
Morello, J. (1958) La provincia fitogeografica del monte. Opera

Lilloana, 2, 1-155.
Rice, W.R., Hostert, E.E. (1994) Laboratory experiments on speciation:

what have we learned in forty years? Evolution, 47, 1637-1653.
Ward, P.S. (1985) The Nearctic species of the genus Pseudomyrmex

(Hymenoptera: Formicidae). Quaestiones Entomologicae, 21, 209-

246.
Ward, P.S. (1989a) Systematic studies on pseudomyrmecine ants:

revision of the Pseudomyrmex oculatus and P. subtilissimus species

groups, with taxonomic comments on other species. Quaestiones

Entomologicae, 25, 393^68.
Ward, P.S. (1989b) Genetic and social changes associated with ant

speciation. In: The Genetics of Social Evolution (ed. by M. D. Breed

and R. E. Page), pp. 123-148. Westview Press, Boulder, Colorado.

Ward, P.S. (1990) The ant subfamily Pseudomyrmecinae (Hymenoptera:

Formicidae): generic revision and relationship to other formicids.

Systematic Entomology, 15, 449^89.
Ward, P.S. (1991) Phylogenetic analysis of pseudomyrmecine ants

associated with domatia-bearing plants. In: Ant-Plant Interactions

(ed. by C. R. Huxley and D. F Cutler), pp. 335-352. Oxford

University Press, Oxford.
Ward, P.S. (1993) Systematic studies on Pseudomyrmex acacia-ants

(Hymenoptera: Formicidae; Pseudomyrmecinae). Journal of

Hymenoptera Research, 2, 117-168.
Wasmann, E. (1909) Uber den Ursprung des sozialen Parasitismus,

der Sklaverei und der Myrmekophilie bei den Ameisen. Biologisches

Centralblatt, 29, 587-604, 619-637, 651-663, 683-703.
West-Eberhard, M.J. (1986) Alternate adaptations, speciation and

phylogeny. Proceedings of the National Academy of Science of the

United States of America, 83, 1388-1392.
Wheeler, W.M. (1905) The ants of the Bahamas, with a list of the

known West Indian species. Bulletin of the American Museum of

Natural History, 21, 79-135.
Wheeler, W.M. (1919) A singular neotropical ant {Pseudomyrma

filiformis Fabricius). Psyche (Cambridge), 26, 124—131.
Wilson, E.O. (1971) The Insect Societies. Harvard University Press,

Cambridge, MA. x + 548 pp.
Wilson, E.O. (1984) Tropical social parasites in the ant genus Pheidole,

with an analysis of the anatomical parasitic syndrome (Hymenoptera:

Formicidae). Insectes Sociaux, 31, 316-334.

Accepted 20 June 1996

1996 Blackwell Science Ltd, Systematic Entomology, 21, 253-263