ORIGINAL ARTICLE
doi:10.1111/j.1558-5646.2007.00119.x
MUSEUM SPECIMENS AND PHYLOGENIES
ELUCIDATE ECOLOGY’S ROLE IN
COEVOLUTIONARY ASSOCIATIONS BETWEEN
MITES AND THEIR BEE HOSTS
Pavel B. Klimov,1 Barry M. OConnor,2 and L. Lacey Knowles3
University of Michigan, Museum of Zoology, 1109 Geddes Ave., Ann Arbor, Michigan 48109-1079
1 E-mail:
pklimov@umich.edu
2 E-mail:
bmoc@umich.edu
3 E-mail:
knowlesl@umich.edu
Received December 26, 2006
Accepted February 19, 2007
Coevolutionary associations between hosts and symbionts (or parasites) are often reflected in correlated patterns of divergence
as a consequence of limitations on dispersal and establishment on new hosts. Here we show that a phylogenetic correlation is
observed between chaetodactylid mites and their hosts, the long-tongued bees; however, this association manifests itself in an
atypical fashion. Recently derived mites tend to be associated with basal bee lineages, and vice versa, ruling out a process of
cospeciation, and the existence of mites on multiple hosts also suggests ample opportunity for host shifts. An extensive survey
of museum collections reveals a pattern of infrequent host shifts at a higher taxonomic level, and yet, frequent shifts at a lower
level, which suggests that ecological constraints structure the coevolutionary history of the mites and bees. Certain bee traits,
particularly aspects of their nesting behavior, provide a highly predictive framework for the observed pattern of host use, with
82.1% of taxa correctly classified. Thus, the museum survey and phylogenetic analyses provide a unique window into the central
role ecology plays in this coevolutionary association. This role is apparent from two different perspectives—as (a) a constraining
force evident in the historical processes underlying the significant correlation between the mite and bee phylogenies, as well as
(b) by the highly nonrandom composition of bee taxa that serve as hosts to chaetodactylid mites.
KEY WORDS:
Apidae, chaetodactylidae, coevolution, long-tongued bees, Megachilidae, mites, nest architecture.
The intimate interactions of symbionts (or parasites) with their
hosts predicts not only the coevolution of symbiont and host biology, but also nonrandom associations among species. A strict
history of cospeciation will generate a concordance between symbiont and host phylogenies (Fahrenholz’s rule) (Eichler 1948;
Klassen 1992; Peek et al. 1998; Clark et al. 2000; Lo et al.
2003; Degnan et al. 2004). However, a suite of factors influences the degree of phylogenetic concordance (Reed and Hafner
1997; Johnson et al. 2002; Rannala and Michalakis 2003; Ronquist 2003; Taylor and Purvis 2003; Quek et al. 2004; Ricklefs
C 2007 The Society for the Study of Evolution.
2007 The Author(s). Journal compilation
Evolution 61-6: 1368–1379
C
1368
et al. 2004; Smith et al. 2004; Banks et al. 2005), including shifts
to new hosts or speciation of the symbiont on the same host (Clayton and Johnson 2003; Clayton et al. 2003; Clayton et al. 2004;
Weckstein 2004).
Chaetodactylid mites represent a compelling group to investigate factors influencing divergence across their hosts. Chaetodactylids are obligate associates of solitary and facultatively social
long-tongued bees (Apidae and Megachilidae), representing the
most species-rich group among 30 other bee–mite lineages (reviewed Eickwort 1994). More than 200 species of chaetodactylid
ECOLOGY’S ROLE IN BEE–MITE COEVOLUTION
mites live in bee nests feeding throughout bee development as
either mutualists (feeding on nest waste), parasitoids (killing the
bee egg or larvae), or as commensals or cleptoparasites (feeding on provisioned pollen) (Krombein 1962; Roubik 1987; Abrahamovich and Alzuet de 1990; Qu et al. 2002). During each bee
generation, the nonfeeding immature mites disperse to new bee
nests on the newly emerged adult bees. Successful dispersal to
new nests is critical for the mites because the bees do not commonly reuse their old nests. The dispersing life-history stage of
the mite (the deutonymph) is tightly synchronized with the last
stages of its bee host’s development. Moreover, some mites are
carried in specialized pouches (acarinaria) of their bee hosts (Fain
and Pauly 2001; Okabe and Makino 2002).
These striking adaptations suggest that the mites and bees
have been involved in long-term coevolutionary interactions, and
that shifts among distantly related bee hosts would be rare events.
For example, the 170 mite species of the genus Sennertia (Klimov
and OConnor) have remained strictly associated with the more
than 500 species of carpenter bees (Michener 2000), even tracking
the ancient dispersal (34–34.6 Mya) of Old World bee lineages to
the New World (Leys et al. 2002). Nevertheless, highly nonrandom
host shifts on unrelated hosts probably have occurred in the early
stages of mite evolution.
Using a comparison of the chaetodactylid mite and longtongued bee phylogenies, and an extensive survey of museum
material, we investigate what factors have structured the historical associations of these symbionts/parasites and their hosts. Our
goals were to (1) determine whether the evolutionary history of
major clades of mites and bee supports a model of coevolutionary
divergence, and (2) understand the underlying determinants of the
observed phylogenetic associations, and specifically (a) what historical events (e.g., cospeciation, host switching, and speciation or
extinction on the host (Johnson et al. 2003; Ronquist 2003) structure this symbiont/parasite–host assemblage, and (b) whether bee
ecology predicts which potential hosts are likely to be part of the
bee–mite coevolutionary association.
Materials and Methods
SPECIMENS AND DETERMINATION OF HOST
ASSOCIATIONS
A thorough survey of museum specimens determined that most
chaetodactylid mites are not associated with a single host species,
but instead, exist on several closely related sympatric hosts. The
only well-supported exception is Sennertia americana, which is
exclusive to a single host; several poorly known species are also
described from a single host, but may reflect the relative rarity of
these species (e.g., species associated with Chalepogenus, Ancyloscelis, Ptilothrix, and Diadasia). The results of the survey on
host ranges of the different chaetodactylid mites are the focus of
this study and are summarized at the generic level of the bees
(Table 1).
Mites from roughly 1500 museum specimens of longtongued bees from 18 museum collections in the United States
and abroad were examined for mites, as were specimens freshly
collected by the authors in North America and Africa. All major
groups of short- and long-tongued bees were sampled, but only
the long-tongued bees had chaetodactylid mites (Table 1). This
survey yielded about 230 mite species that includes all currently
known species and species groups, as well as a large number of
undescribed taxa, with the exception of a few species of Sennertia for which their bee hosts were not available for inspection.
The majority of these mite specimens (about 5000 slide-mounted
specimens) are vouchered in the University of Michigan, Museum
of Zoology (Klimov and OConnor 2003).
PHYLOGENETIC ANALYSES
Only the relationships among mite genera are considered (Table 1)
for comparison with the phylogeny of the bee hosts because of
constraints imposed by incomplete host phylogenies (Michener
2000) and to emphasize the unique cophylogenetic pattern observed at this level. Maximum parsimony analyses of 51 morphological characters (the appendix) from chaetodactylid heteromorphic deutonymphs was used to estimate the mite phylogeny
were conducted in PAUP∗ 4.0b10 (Swofford 2002) using either
equal character weights, or characters weighted according to the
degree of homoplasy using Goloboff’s concave weighting function (Goloboff 1993) with the constant of concavity (k) set to
2 (implied weights parsimony). A bootstrap majority rule consensus tree was calculated using the branch-and-bound algorithm
and 10,000 bootstrap replicates. Bremer branch support or decay
indices were also calculated using PAUP∗ with a command file
generated in TreeRot.v2 (Sorenson 1999). A Bayesian analysis
was also conducted; four chains (three hot, one cold) of 5 × 106
generations each with a burn-in of 6300 and a sampling frequency
of 100 were used (MrBayes ver. 3.1.1 [Ronquist and Huelsenbeck
2003]). Five independent analyses were conducted to confirm convergence; all resulted in similar topologies. All analyses produced
the same topology, except for unresolved relationships in Achaetodactylus in the two parsimony analyses. Megacanestrinia (family
Canestriniidae) was selected as the outgroup for the mite phylogeny as Chaetodactylidae is most likely a basal group of the
superfamily Hemisarcoptoidea (OConnor 1993). The influence
of outgroup choice on the phylogeny and position of the root were
thoroughly investigated and the topology of the tree was robust
to various potential outgroups (54 astigmatid families, including
Canestriniidae, Aeroglyphidae, Glycyphagidae, Winterschmidtiidae, Hyadesiidae, and Algophagidae). The trees are deposited
EVOLUTION JUNE 2007
1369
PAVEL B. KLIMOV ET AL.
Table 1. Distribution of chaetodactylid mites across their bee hosts showing their geographic affiliations. Aust, Australian region; Orient, Oriental region; Madag, Madagascar; Afr, Afrotropical region; Palear, Palearctic region; Near, Nearctic Region; Antill, the Greater
and Lesser Antilles, excluding Trinidad; Arauc, Araucanian region (Michener 2000). Some associations that are marginally occurred in a
region are omitted. Unusual finding of chaetodactylids on Andrena, Halictus, Anthophora, Apis, Vespula, Passalidae (Zachvatkin 1941;
Chmielewski 1993; Haitlinger 1999), and Bombus (our data) are omitted. Cleptoparasites of the principal hosts (parenthesis) that may
transfer chaetodactylids are also not included: Stelis (Osmia), sapygids Polochrum (Xylocopa), Sapyga (Chelostoma) (Zachvatkin 1941;
Samšiňák 1973), and Coelioxoides (Tetrapedia) (our data).
Mite taxon
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Sennertia
Sennertia
Achaetodactylus
Chaetodactylus
Roubikia
Chaetodactylus
Chaetodactylus
Chaetodactylus
Chaetodactylus
Centriacarus
Bee taxon
Megachilidae
Lithurgini
Lithurgus
Trichothurgus
Microthurge
Osmiini
Osmia
Hoplitis
Chelostoma
Anthidiini
Rhodanthidium
Anthidium
Megachilini
Megachile
Apidae (Xylocopinae)
Xylocopini
Xylocopa
Ceratinini
Ceratina
Ceratina
Aust
Orient
Madag
Afr
Palear
Near
Neotr
Antill
Arauc
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Apidae (Apinae)
Tapinotaspidini
Chalepogenus
Tetrapediini
Tetrapedia
Emphorini
Melitoma
Diadasia
Ptilothrix
Ancyloscelis
Centridini
Centris
EVOLUTION JUNE 2007
+
+
in TreeBase (SN3139). Taxon selection for the analyses above
(Table 2) was done on the basis of a larger analysis representing all species groups and genera of chaetodactylids (46 species)
(online Supplementary Figure S1).
The results from the morphology-based phylogenetic analysis (Fig. 1) were confirmed by a molecular phylogenetic analysis of
nuclear protein-coding and ribosomal gene sequence data (1.1 kb
of EF1-␣, 1.8 kb of 18S, and 2.15 kb of domains 1–5 and 9–10 of
28S rDNA) for a subset of taxa (one species of Achaetodactylus,
five species of Chaetodactylus, and seven species of Sennertia).
1370
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
This analysis is part of a larger ongoing molecular phylogenetic
project (P. Klimov and B. M. OConnor, unpubl. data); as with the
morphological based analyses, robustness of the phylogenetic tree
to different outgroups (specifically, representatives from 54 different astigmatid mite families) was confirmed with the molecular
phylogenetic analyses as well.
As a phylgoenetic estimate of the bee hosts, the generic topology of Roig-Alsina and Michener, analysis C (Roig-Alsina and
Michener 1993) and the tribal phylogeny of Engel (2001) were
used. These findings are widely accepted by bee systematists and
ECOLOGY’S ROLE IN BEE–MITE COEVOLUTION
Taxa used in the generic-level phylogenetic analysis of Chaetodactylidae. Early derivative mite lineages representing all known
host associations at the generic level were selected for the phylogenetic analyses (except for little known derived species of Chaetodactylus
associated with Emphorini and Tapinotaspidini, see Table 1. The taxon selection was based on the results of the 46-taxon phylogenetic
Table 2.
analysis (online Supplemental Figure S3).
Taxon
Host
Collection locality
Megacanestrinia sp.
Centriacarus turbator
Centriacarus guahibo
Roubikia panamensis
Tefflus zanzibaricus
Centris (Heterocentsris) vittata, C. sp.
Centris sp.
Tetrapedia diversipes, T. peckholtii, T. sp., Coelioxoides
waltheriae (cleptoparasite of Tetrapedia)
Tetrapedia sp.
Ceratina opaca
Ceratina diloensis
Ceratina sp., C. spilota, C. aereola,
C. excavata (Fain 1981)
Melitoma marginella, M. segmentaria, Melitoma sp.
Lithurgus dentipes, L. scabrosus,
L. atratus
Osmia rufa, O. tricornis,
O. fulviventris, O. cornuta
Xylocopa olivieri
Ceratina chloris, C. laeta
Tanzania
Brazil, Peru, Colombia, Panama, Mexico
Venezuela
Panama, Mexico, French Guiana,
Brazil, Bolivia
Peru
South Africa
Democratic Republic of Congo
Tanzania, Cameroon, Democratic Republic
of the Congo (Fain 1981)
Mexico, Honduras
Micronesia, Indonesia, New Caledonia,
French Polynesia, South India
France, Belgium, England, Germany,
Hungary, Croatia, Spain
Greece
Suriname, French Guiana, Panama
Roubikia latebrosa
Achaetodactylus ceratinae
Achaetodactylus leleupi
Ochaetodactylus decellei
Chaetodactylus melitomae
Chaetodactylus ludwigi
Chaetodactylus osmiae
Sennertia zhelochovtsevi
Sennertia surinamensis
largely match molecular data (unpublished phylogeny, Danforth,
pers. comm., 2005); family-level relationships (Danforth et al.
2006); however, some relationships require further investigation
given weak support and character conflict in the dataset (RoigAlsina and Michener 1993).
ANALYSES OF HISTORICAL ASSOCIATIONS BETWEEN
MITES AND THEIR HOSTS
A test for a significant correlation between the phylogenies of
the mites and bee hosts was conducted with PARAFIT (Legendre
et al. 2002); this approach accommodates the association of mite
taxa on more than one host, where taxon distances are used in the
test, rather than topology. Correlation was evaluated by determining whether the degree of association between species distances
(computed from a principal coordinate analysis of the patristic
distance matrices from the host [Roig-Alsina and Michener 1993;
Engel 2001] and parasite phylogenies) (Fig. 1, excluding the outgroup) and an incidence matrix that describes the associations
between mites and their hosts (Table 2) differs significantly from
expectations of random association between mites and hosts (for
details see Legendre et al. 2002). The program DISTPCOA (Legendre and Anderson 1999) was used to transform patristic matrices
to principal coordinates. Permutation tests (9999 randomizations)
were used to assess the probability that the detected coevolutionary correlation differs significantly from that expected by chance.
The program TREEFITTER (Ronquist 1995, 2003) was used
to examine what historical processes might have generated the
observed correlation between the mite and bee histories. Possible historical scenarios were investigated by evaluating the cost
space of the four events: cospeciation, speciation within a host
lineage, host shifts, and extinction (which are referred to as codivergence, duplication, switching, and sorting, respectively in the
program; Ronquist [2003]). General cost optimization was performed in which the cost of each event was inversely related to
the likelihood of the event and the cost was varied incrementally
within a specified range (i.e., between 0 to 10 for all events, except
for extinction, for which a range of 0.5–4.0 was considered) (see
Ronquist [2003]). The significance of each historical scenario was
evaluated against the null hypothesis that such a combination of
events was statistically indistinguishable from a pattern arising
by chance using randomization tests. Phylogenetically significant
scenarios were identified as those for which the probability of
observing a particular number of events (e.g., Table 3) was P <
0.05 as assessed by 10,000 permutations of both mite- and hosttree terminals. For this test, the topology of the host tree used in
the analysis was the same as that of Roig-Alsina and Michener
(1993) with a few additions of bee tribes from recent smaller
scale phylogenetic analyses (none of which is chaetodactylid
hosts).
Four associations of Chaetodactylus with Emphorini and
Tapinotaspidini (Table 1) were excluded. They constitute only
1.8% of all known associations at the species level, comprise only
derived mite taxa (see online Supplemental Fig. S1) collected from
a single bee species each and may be accidental.
EVOLUTION JUNE 2007
1371
PAVEL B. KLIMOV ET AL.
Phylogenetic relationships within the family Chaetodactylidae based on Bayesian analysis of morphological data (on the right)
of 12 representative taxa (Table 2) selected on the basis of a larger analyses (Fig. S1); posterior probabilities, bootstrap values, and Bremer
Figure 1.
indices are shown. A phylogeny of the long-tongued bees (Engel 2001) is shown on the left. Links between mites and their bee hosts are
shown (a few other links formed by rare and relatively derived mite species are given in Table 1; their exclusion is justified by phylogenetic
analyses presented in supplemental material available online). Note that the chaetodactylid phylogeny is drawn “upside-down” for ease
of showing the inverse phylogenetic correlation.
ANALYSES OF THE ECOLOGICAL DETERMINANTS
OF THE MITE AND BEE ASSEMBLAGES
Four aspects of bee biology that may affect the suitability of a
particular taxon as a chaetodactylid host were considered: nest
construction site, the arrangement of cells within a nest, the provisioning of cells, and the degree of sociality characterizing the
bee taxa. These data were collected primarily from Radchenko and
Pesenko (1994), Radchenko (1996), and Michener (2000), references cited therein, and more recent publications (online Supplementary Table S1). In a few cases, data were extrapolated from
other species when the trait appeared to be similar across the
genus.
The six significant models (as shown in Fig. 2B) that are consistent with the observed correlation between the mite and bee
phylogenies, and the number of invoked historical events specified under a particular model (cospeciation, speciation within a host
lineage, host shifts, and extinction); see Figure 2B for the distribution of costs for host switching and speciation within host associated
with each model.
Table 3.
Historical processes
Models
1
Cospeciation
Speciation within a host lineage
Host shifts
Extinction
Total number of events
Total cost
1372
0
5
6
0
11
3.3–16.5
EVOLUTION JUNE 2007
2
1
5
5
3
14
19.5–24.75
3
2
5
4
7
18
24.6–31.8
4
3
5
3
12
23
28.5–50.9
5
3
7
1
26
37
33.15–49.9
6
3
8
0
35
46
35–39
ECOLOGY’S ROLE IN BEE–MITE COEVOLUTION
2003), might account for the correlation between the mite and beehost phylogenies. These historical scenarios (i.e., models) were
explored using a general cost-optimization procedure (Fig. 2A).
Costs were assigned and varied incrementally to each of the
four historical processes (i.e., cospeciation, host shifts, speciation
within a host lineage, and extinction), where the cost assigned to
each event was inversely related to the likelihood of the process,
to produce a cost space surface for the number of specific historical events that would have to be invoked in order to produce the
observed association between the mite and bee phylogenies under
different models (Ronquist 2003).
Six significant models (Table 3) were identified from the
general cost-optimization procedure as having a probability of
less than 5% of being generated by chance (Fig. 2B), as assessed
by permutations of both mite and host taxa across the respective trees. However, these models are not necessarily biologically
equivalent. Some models can be rejected as unlikely because of
the excessive number of historical events required to generate the
observed correlation between mite and bee phylogenies and high
total cost. For example, the absence of host shifts postulated by
model 6 (Table 3) requires invoking 35 extinctions, which results
in a relative high total cost for the model (lower, right corner
of Fig. 2B). Consideration of both the total costs and number
of individual events required to produce the observed mite–bee
assemblage identifies a model of speciation within hosts and host
switching as more parsimonious than the other scenarios; a difference of 11 total events (model 1, Table 3, vs. 14, 18, 25,
37, and 46 events for models 2, 3, 4, 5, and 6, respectively).
Logistic regression analysis was used to investigate how well
these traits predict whether chaetodactylid mites will (or will not)
be associated with a particular bee taxon. The fit of the data to
the model was evaluated using a likelihood-ratio test. The predictive power of the model (i.e., the contribution of host biological
traits to the observed pattern of bee–mite associations) was evaluated with the program SPSS version 11.0.4 (2005) by calculating
the posterior probabilities for each bee taxon and estimating the
percentage of correctly predicted associations.
Results
INVERSE CORRELATION BETWEEN MITE
AND BEE PHYLOGENIES
A significant (P = 0.029) inverse correlation between the phylogenies of the mites and bees was detected (Fig. 1) (based on
analyses from the program PARAFIT [Legendre et al. 2002]), but it
is not the pattern expected from a history of cospeciation (Eichler
1948; Klassen 1992; Peek et al. 1998; Clark et al. 2000; Lo et al.
2003; Degnan et al. 2004). Recently derived mites are not associated with recently derived hosts. Instead the converse is observed.
This intriguing pattern of bee–mite associations at the level of
bee tribes creates an unprecedented case of a reverse “codivergence” that violates Fahrenholz’s rule and is obviously not caused
by cospeciation.
FACTORS UNDERLYING THE BEE–MITE ASSOCIATIONS
Different historical scenarios, involving cospeciation, host shifts,
speciation within a host lineage, and/or extinction (Ronquist 1995,
(A)
(B)
0
1.00
2
0.75
4
3
0.50
6
0.25
0.0
cost of host switch
2
ch
4
3.3
8
sw
it
P that historical sce
anario is random
1
6.6
8
6
within
of
iation
a hos
4
t linea
6
st
f spec
ge
2
9.9
0
co
cost o
5
ho
10
st
0.00
10
3.0
2.0
1.0
0.0
cost of speciation within a host lineage
Exploration of the cost space to evaluate the different processes underlying the observed coevolutionary association of chaetodactylid and bees: (A) P-values were estimated from 10,000 random permutations of both host and symbiont terminals with a codivergence
and extinction cost of 0 and1, respectively; significant values (P < 0.05) are shaded and shown in detail on the right (B), where the dashed
lines demark the cost space that corresponds with the specific suite of historical events represented in the six significant models (as given
Figure 2.
in Table 3).
EVOLUTION JUNE 2007
1373
PAVEL B. KLIMOV ET AL.
Classification accuracy of the four-variable logistic regression analysis, indicating the analyses high classification accuracy of predicting the presence or absence of chaetodactylid mites
in the nest of a long-tongued bee; the four-variables used in the
model were nest construction site, cell arrangement, cell provision,
and the degree of sociality.
Table 4.
Mites
Inferences from model
Absent
Present
Overall
Absent
Present
Accuracy (%)
20
3
4
12
83.3
80.0
82.1
Because model 1 (speciation within hosts and host switching) is
the most parsimonious and biologically meaningful (many species
of chaetodactylids are known to exist on different hosts and several species of the mites can be associated with a single host
species), we consider it as the most preferred historical scenario for
this association.
A significant relationship between the presence of chaetodactylid mites and specific ecological aspects of the bees was
detected by the logistic regression analysis (P = 0.024). The classification accuracy of bees as potential hosts based on four bee
traits (nesting site, cell arrangement, cell provisioning, and sociality; see online Supplemental Table S1) is 82.1%, indicating the
importance of the traits (Table 4). Any combination of bee traits,
where one or more were removed, resulted in a decreased predictive power, suggesting all four aspects of bee ecology structure
the coevolutionary assemblage of these symbionts/parasites and
hosts. For relative contribution of each variable to the model see
online Supplementary Table S1; logistic regression coefficients
(online Supplementary Table S2) can be used to calculate the
probability of mite association with any given long-tongued bee.
Discussion
We propose that ecology has had a predominant influence on
the history of coevolutionary associations between chaetodactylid
mites and long-tongued bees, affecting both the opportunity and
probability of successful host shifts, and thereby dictating which
bees are potential hosts (Michener 2000). One expectation resulting from the dependency of the mites on their bee hosts for food,
habitat, and a means of dispersal (Krombein 1962; Abrahamovich
and Alzuet de 1990) is the significant correlation between the mite
and bee phylogenies, as with other symbiont/parasite–host assemblages (Reed and Hafner 1997; Peek et al. 1998; Clark et al. 2000;
Lo et al. 2003; Degnan et al. 2004). Indeed, such a correlation
is observed but it is inverse (Fig. 1). The signature of ecological constraint in this case manifests itself in an atypical fashion.
1374
EVOLUTION JUNE 2007
First, the temporal disjunction between the diversification of the
mite and bee lineages as evidenced by the bee and mite topologies
rules out a process of cospeciation (Ronquist 2003). Second, the
existence of mite species on multiple related hosts indicates there
are ample opportunities for host shifts. The museum survey and
the phylogenetic analyses identify two ways in which ecology has
constrained this coevolutionary history—interestingly, the nonrandom association between the mites and bees (Fig. 1) provides
evidence for ecological constraint in terms of the infrequency of
shifts among distantly related hosts and their frequency among
closely related hosts (Table 1).
Most of the mite genera and species groups are associated
with hosts from a single bee tribe, with the notable exception of
Chaetodactylus (see below). Conservatism in the pattern of host
use evident at this higher taxonomic level (Fig. 1) does not apparently reflect the lack of opportunities for host shifting at lower
taxonomic levels. For example, species of Sennertia have experienced multiple host shifts within and between the host genera Ceratina and Xylocopa and these were followed by speciation events.
All this without any obvious mechanism of transfer between host
taxa. These bees differ in nesting sites and lack a common hymenopteran cleptoparasite (a potential means of dispersal among
host species). Ecological constraints imposed by similarities in
life-cycles of these bees may structure patterns of host shifting in
these mite genera. The absence of Sennertia mites on numerous
alternative and sympatric hosts supports the hypothesized constraint of mite dispersal imposed by the life-cycle of its host, as
opposed to geography explaining patterns of host shifts.
Mites in the genus Chaetodactylus represent an interesting
contrast to the pattern of host conservatism, and have shifted onto
many unrelated and different bee tribes in the families Megachilidae and Apidae, with particular mite species still associated with
closely related hosts (Fig. 1; Table 1). Many unrelated hosts of
Chaetodactylus exhibit similar ecologies, as evidenced by the relatively high accuracy of the predictive classification based on aspects of bee ecology (Table 4). The broad host range of Chaetodactylus may be explained by two particular characteristics of
this mite genus. These mites most likely experience increased opportunities for host shifting as a result of an inert deutonymph,
a cyst-like life-history stage that can survive off the host. By remaining in the nest cavity, the mite is able to infest the next bees
to reuse the cavity (Krombein 1962). For example, the nests of the
hosts of the C. osmiae species group (i.e., several taxa of bees in
the genus Osmia) are reused by other bee taxa, providing means
for dispersal between different bee species. Antagonistic interactions of Chaetodactylus with its host may also incur evolutionary
pressures driving the utilization of new unrelated hosts. Species
in this mite genus often kill the developing bee larvae (van Lith
1957; Krombein 1962; Qu et al. 2002). Phylogenetic reconstruction of the genus (Klimov and OConnor; online Supplementary
ECOLOGY’S ROLE IN BEE–MITE COEVOLUTION
Fig. S1) shows that associations with Apidae (except for Melitoma) (Table 1) are clearly secondary and resulted from host shifts
from unrelated megachilid hosts.
How do host shifts occur in mites that typically inhabit nests
of a single bee are dependent on their host for dispersal to new
nests, and therefore a single host species? Various mechanisms
could provide opportunities for host shifting, which include nest
supersedure (i.e., the take over of a nest partly provisioned by
a different individual of the same or different species), utilization of a shared entrance to intraspecific or interspecific nest tunnels, and hibernating aggregations (Rust 1974; Linsley et al. 1980;
Gerling et al. 1989; McCorquodale and Owen 1994; Hogendoorn
1996). Host shifts may also occur from an incidental transfer on
flowers (e.g., Sennertia) or with loose dirt collected as nest material (e.g., Roubikia) (Roubik 1987; Vicidomini 1996). Lastly, hymenopteran cleptoparasites attacking multiple host species may
facilitate movement of mites among populations and different host
species (Zachvatkin 1941; Samšiňák 1973; Munster-Swendsen
and Calabuig 2000). For example, deutonymphs of C. krombeini
and C. reaumuri were found on the cleptoparasites Stelis montana and S. murina, respectively (Klimov and OConnor, in press;
Türk and Türk 1957). Transfer among their hymenopteran hosts
via cleptoparasites is known to influence the host ranges of the
unrelated mite genera Vidia (OConnor and Eickwort 1988) and
Parasitellus (Richards and Richards 1976), and dispersal on parasitic hippoboscid flies is similarly considered a major cause of
incongruence between phylogenies of the louse genus Brueelia
and their avian hosts (Johnson et al. 2002). Although it is not possible to determine how prevalent one mode of transfer might be
over another, irrespective of the specific mechanism involved in
a host shift, the success of such events in chaetodactylid mites is
highly predictable based on aspects of bee ecology (Table 4).
Host characteristics play a critical role in structuring the host–
mite assemblage (Table 4), no doubt by influencing the opportunity and probability of success of a host shift. Properties of the
bee nests can significantly constrain mite dispersal. Cell partitions
constructed by the majority of bees are impenetrable for mites and
do not allow them to move across the brood; the mites die if the
bee in an infested cell dies early (Krombein 1962, 1967; Michener
2000). Consequently, chaetodactylids are associated with bees
where the dispersal ecology of the mites is not limited by nest
architecture. For example, in nests with cells arranged in a linear sequence, bees in the inner cells usually complete development sooner and break through partitions of the outermost cells
to emerge (Skaife 1952; Krombein 1962; Linsley et al. 1980) and
cross-contaminate other members of the nest. Not surprisingly,
chaetodactylids are associated with bees exhibiting these qualities as opposed to bees with independent emergence of broods
(e.g., branching nests or nests composed of clusters of cells, see
online Supplementary Table S1). In addition to the bee traits that
would foster a host–mite association, certain developmental and
biological characteristics of hosts that would negatively impact the
mite could also contribute to the observed coevolutionary patterns.
For example, the bee tribes Allodapini and Ceratinini are closely
related, with similar nest architectures (excluding the absence of
cell partitions in the former), but only the Ceratinini provision
their cells prior to laying eggs. Chaetodactylids are not associated
with bee lineages of the Allodapini whereas they do exist on the
Ceratinini (Table 1, online Supplementary Table S1), suggesting
that progressive feeding of larvae, rather than mass provisioning,
may make them unsuitable hosts for the mites. Bees in both the
Apini and Bombini lay their eggs with a little or no associated
food, and neither hosts any chaetodactylids, although it possible that the production of different castes in highly eusocial bees
might also reduce the chances of mite dispersal and contribute to
the complete absence of chaetodactylids.
Conclusions
The extensive museum survey and phylogenetic analyses provide
a unique window into how ecological constraint has shaped the coevolutionary associations of chaetodactylid mites and their hosts,
the long-tongued bees. The temporal disjunction between the diversification of the mite and bee lineages rules out a process of
cospeciation. Moreover, although the existence of mites on multiple hosts species suggests ample opportunity for dispersal among
hosts, the infrequency of host shifts at one level of taxonomic
resolution and their frequency at another suggest how ecological
characteristics of the bees affect both the opportunity for dispersal
and the probability of successful infestation of the mites. When
certain characteristics of the bees and, in particular, aspects of
their nesting behavior are considered, a highly predictive framework for this coevolutionary association emerges, reflecting the
critical role ecology plays in governing the distribution of mites
across the bee hosts.
ACKNOWLEDGMENTS
We would like to acknowledge the curators of 18 museums who provided
access to host bee specimens and mite collections used in this study, including personal thanks for the help and hospitality during our visits to
J. G. Rozen and J. S. Ascher (American Museum of Natural History,
New York), T. Griswold (USDA—Bee Biology and Systematics Laboratory, Logan, Utah), B. V. Brown (Natural History Museum of Los Angeles
County), C. Michener and the late B. Alexander (University of Kansas,
Lawrence), and W. Pulawski (California Academy of Sciences, San Francisco). We also thank P. Legendre (Universite de Montreal, Canada) for his
very useful assistance with ParaFit, C. Michener, J. Rozen, and J. Ascher
for useful comments on the manuscript and, especially, its bee component, and identification of some critical host taxa, M. Terzo (Université de
Mons-Hainaut, Belgium) for sharing his unpublished data on phylogeny
of Ceratina, J. Bosch (Universitat Autònoma de Barcelona, Spain) for providing information on the nest architecture of Rhodanthidium sticticum,
EVOLUTION JUNE 2007
1375
PAVEL B. KLIMOV ET AL.
and G. Hammond (University of Michigan) for his valuable comments
on earlier drafts of the manuscript. We also extend our appreciation to J.
Dykema, J. Diesel, and R. Tao (undergraduate assistants at the University
of Michigan) for their help in mounting, labeling, and databasing mite
specimens. This work was supported by grants from the National Science
Foundation, DEB-0118766 (PEET) and the United States Department of
Agriculture (CSREES #2002-35302-12654).
LITERATURE CITED
Abrahamovich, A. H., and A. B. Alzuet de. 1990. Tipos de asociación entre
acaros e Hymenópteros. I: Sennertia (A.) splendidulae Alzuet y Abrahamovich, 1988 (Acarina: Chaetodactylidae) comensal y forético de
Xylocopa (S.) splendidula splendidula Lepeletier, 1841 (Hymenoptera:
Anthophoridae). Turrialba 40:319–322.
Banks, J. C., R. L. Palma, and A. M. Paterson. 2005. Cophylogenetic relationships between penguins and their chewing lice. J. Evol. Biol. 19:156–166.
Chmielewski, W. 1993. Biologia Chaetodactylus osmiae (Duf., 1866)
(Acarida, Chaetodactylidae - pylkozernego rozkruszka foretycznie
zwiazanego z pszczolami samotnymi (Apoidea). Pszczelnicze Zeszyty
Naukowe 37:133–143.
Clark, M. A., N. A. Moran, P. Baumann, and J. J. Wernegreen. 2000. Cospeciation between bacterial endosymbionts (Buchnera) and a recent radiation
of aphids (Uroleucon) and pitfalls of testing for phylogenetic congruence. Evolution 54:517–525.
Clayton, D. H., S. Al-Tamimi, and K. Johnson. 2003. The ecological basis
of coevolutionary history. Pp. 310–341 in R. D. M. Page, ed. Tangled
trees: phylogeny, cospeciation, and coevolution. Univ. Chicago Press,
Chicago.
Clayton, D. H., S. E. Bush, and K. P. Johnson. 2004. Ecology of congruence:
past meets present. Syst. Biol. 53:165–173.
Clayton, D. H., and K. P. Johnson. 2003. Linking coevolutionary history to
ecological process: doves and lice. Evolution 57:2335–2341.
Danforth, B. N., S. Sipes, J. Fang, and S. G. Brady. 2006. The history of
early bee diversification based on five genes plus morphology. Proc.
Natl Acad. Sci. USA 103:15118–15123.
Degnan, P. H., A. B. Lazarus, C. D. Brock, and J. J. Wernegreen. 2004. Hostsymbiont stability and fast evolutionary rates in an ant-bacterium association: cospeciation of Camponotus species and their endosymbionts,
Candidatus Blochmannia. Syst. Biol. 53:95–110.
Eichler, W. 1948. Some rules on ectoparasitism. Ann. Mag. Nat. Hist. 12:588–
598.
Eickwort, G. C. 1994. Evolution and life-history patterns of mites associated
with bees. Pp. 218–251 in M. A. Houck, ed. Mites: ecological and evolutionary analyses of life-history patterns. Chapman and Hall, London.
Engel, M. S. 2001. A monograph of the Baltic amber bees and evolution of
the Apoidea (Hymenoptera). Bull. Am. Mus. Nat. Hist. 259:1–192.
Fain, A. 1981. Notes on the hypopi of the genus Chaetodactylus Rondani,
1866 (Acari, Chaetodactylidae). Bull. Inst. R. Sci. Nat. Belg. Entomol.
53:1–9.
Fain, A., and A. Pauly. 2001. Notes on phoretic deutonymphs of mites (Acari)
associated with Old World Megachilidae and Anthophoridae (Insecta
Hymenoptera), mainly from Madagascar 1. Families Chaetodactylidae,
Acaridae, Histiostomatidae and Winterschmidtiidae (Astigmata). Belg.
J. Entomol. 3:125–142.
Gerling, D., H. H. W. Velthuis, and A. Hefetz. 1989. Bionomics of the
large carpenter bees of the genus Xylocopa. Annu. Rev. Entomol. 34:
163–190.
Goloboff, P. A. 1993. Estimating character weights during tree search. Cladistics 9:83–91.
1376
EVOLUTION JUNE 2007
Haitlinger, R. 1999. Sennertia herminae sp. n. (Acari, Astigmata: Chaetodactylidae) a phoretic mite associated with Xylocopa sp. (Hymenoptera:
Anthophoridae) from Madagascar. Folia Entomol. Hung. 60:57–59.
Hogendoorn, K. 1996. Socio-economics of brood destruction during supersedure in the carpenter bee Xylocopa pubescens. J. Evol. Biol. 9:
931–952.
Johnson, K. P., R. J. Adams, and D. H. Clayton. 2002. The phylogeny of the
louse genus Brueelia does not reflect host phylogeny. Biol. J. Linn. Soc.
Lond. 77:233–247.
Johnson, K. P., R. J. Adams, R. D. M. Page, and D. H. Clayton. 2003. When
do parasites fail to speciate in response to host speciation? Syst. Biol.
52:37–47.
Klassen, G. J. 1992. Coevolution: a history of the macroevolutionary approach
to studying host–parasite associations. J. Parasitol. 78:573–587.
Klimov, P. B., and B. M. OConnor. 2003. Mites associated with bees,
University of Michigan. Available at http://141.211.243.61/bee mites/
index.html.
Klimov, P. B., and B. M. OConnor. Morphology, evolution and host associations of beeassociated mites of the family Chaetodactylidae (Acari:
Astigmata), with monographic revision of North American taxa. Misc.
Publ. Mus. Zool. Univ. Mich. 198. In Press.
Krombein, K. V. 1962. Natural history of Plummers Island, Maryland. XVI.
Biological Notes on Chaetodactylus krombeini Baker, a parasitic mite
of the megachilid bee, Osmia (Osmia) lignaria Say (Acarina, Chaetodactylidae). Proc. Biol. Soc. Wash. 75:237–250.
———. 1967. Trap-nesting wasps and bees: life histories, nests, and associates. Smithsonian Institution Press, Washington, DC.
Legendre, P., and M. J. Anderson. 1999. Distance-based redundancy analysis:
testing multispecies responses in multifactorial ecological experiments.
Ecol. Monogr. 69:1–24.
Legendre, P., Y. Desdevises and E. Bazin 2002. A statistical test for host–
parasite coevolution. Syst. Biol. 51:217–234.
Leys, R., S. J. B. Cooper, and M. P. Schwarz. 2002. Molecular phylogeny
and historical biogeography of the large carpenter bees, genus Xylocopa
(Hymenoptera: Apidae). Biol. J. Linn. Soc. Lond. 77:249–266.
Linsley, E. G., J. W. MacSwain, and C. D. Michener. 1980. Nesting Biology
and Associates of Melitoma (Hymenoptera, Anthophoridae). Univ. Calif.
Publ. Entomol. 90:1–45.
Lo, N., C. Bandi, H. Watanabe, C. Nalepa, and T. Beninati. 2003. Evidence
for cocladogenesis between diverse dictyopteran lineages and their intracellular endosymbionts. Mol. Biol. Evol. 20:907–913.
McCorquodale, D. B., and R. E. Owen. 1994. Laying sequence, diploid males,
and nest usurpation in the leafcutter bee, Megachile rotundata (Hymenoptera: Megachilidae). J. Insect Behav. 7:731–738.
Michener, C. D. 2000. The bees of the world. John Hopkins Univ. Press,
Baltimore and London.
Munster-Swendsen, M., and I. Calabuig. 2000. Interaction between the solitary
bee Chelostoma florisomne and its nest parasite Sapyga clavicornis—
empty cells reduce the impact of parasites. Ecol. Entomol. 25:63–70.
OConnor, B. M. 1993. Generic relationships in the Chaetodactylidae (Acari:
Astigmata) with description of a new genus. Acarologia 34:345–362.
OConnor, B. M., and G. C. Eickwort. 1988. Morphology, ontogeny and systematics of the genus Vidia (Acari: Winterschmidtiidae). Acarologia 29:147–
174.
Okabe, K., and S. Makino. 2002. Phoretic mite fauna on the large carpenter
bee Xylocopa appendiculata circumvolans (Hymenoptera: Apidae) with
descriptions of its acarinaria on both sexes. J. Acarol. Soc. Japan 11:73–
84.
Peek, A. S., R. A. Feldman, R. A. Lutz, and R. C. Vrijenhoek. 1998. Cospeciation of chemoautotrophic bacteria and deep sea clams. Proc. Natl Acad.
Sci. USA 95:9962–9966.
ECOLOGY’S ROLE IN BEE–MITE COEVOLUTION
Qu, D., Y. Maeta, M. Goubara, K. J. Nakatsuka, J. Kozo, and K. Kenji.
2002. Reproductive strategy in the two species of cleptoparasitic astigmatid mites, Chaetodactylus nipponicus and Tortonia sp. (Acari: Chaetodactylidae and Suidasiidae), infesting Osmia cornifrons (Hymenoptera:
Megachilidae). I. Invasion/infestation patterns and partial use of the host
food. Jpn. J. Entomol. 5:121–141.
Quek, S. -P., S. J. Davies, T. Itino, and N. E. Pierce. 2004. Codiversification
in an ant-plant mutualism: stem texture and the evolution of host use
in Crematogaster (Formicidae: Myrmicinae) inhabitants of Macaranga
(Euphorbiaceae). Evolution 58:554–570.
Radchenko, V. G. 1996. Evolution of nest building in bees (Hymenoptera,
Apoidea). Entomol. Rev. 75:20–32.
Radchenko, V. G., and Y. A. Pesenko. 1994. Biologiya pchel (Hymenoptera:
Apoidea) [ = Biology of bees (Hymenoptera: Apoidea)]. Russian
Academy of Sciences, St. Petersburg.
Rannala, B., and Y. Michalakis. 2003. Population genetics and cospeciation:
from process to pattern. Pp. 120–143 in R. D. M. Page, ed. Tangled
trees: phylogeny, cospeciation, and coevolution. Univ. Chicago Press,
Chicago.
Reed, D. L., and M. S. Hafner. 1997. Host specificity of chewing lice on
pocket gophers: a potential mechanism for cospeciation. J. Mammal.
78:655–660.
Richards, L. A., and K. W. Richards. 1976. Parasitid mites associated with
bumblebees in Alberta, Canada (Acarina: Parasitidae; Hymenoptera: Apidae). II. Biology. Univ. Kans. Sci. Bull. 51:1–18.
Ricklefs, R. E., S. M. Fallon, and E. Bermingham. 2004. Evolutionary relationships, zcospeciation, and host switching in avian malaria parasites.
Syst. Biol. 53:111–119.
Roig-Alsina, A., and C. D. Michener. 1993. Studies of the phylogeny and classification of long-tongued bees (Hymenoptera: Apoidea). Univ. Kans.
Sci. Bull. 55
Ronquist, F. 1995. Reconstructing the history of host–parasite associations
using generalised parsimony. Cladistics 11:73–89.
———. 2003. Parsimony analysis of coevolving species associations. Pp. 22–
64 in R. D. M. Page, ed. Tangled trees: phylogeny, cospeciation, and
coevolution. Univ. Chicago Press, Chicago.
Ronquist, F., and J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574.
Roubik, D. W. 1987. Notes on the biology of anthophorid bee Tetrapedia
and the mite Chaetodactylus panamensis Baker, Roubik and DelfinadoBaker (Acari: Chaetodactylidae). Int. J. Acarol. 13:75–76.
Rust, R. W. 1974. The systematics and biology of the genus Osmia, subgenera
Osmia, Chalcosmia, and Cephalosmia (Hymenoptera: Megachilidae).
Wasmann J. Biol. 32:1–93.
Samšiňák, K. 1973. Zwei neue Arten der Gattung Chaetodactylus Rondani,
1866 (Acari, Acaroidea). Zool. Anz. 190:400–404.
Skaife, S. H. 1952. The yellow-banded carpenter bee Mesotrichia caffra Linn.,
and its symbiotic mite, Dinogamasus braunsi Vitzthum. J. Entomol. Soc.
South. Afr. 15:63–76.
Smith, V. S., P. R. D. M., and K. P. Johnson. 2004. Data incongruence and the
problem of avian louse phylogeny. Zool. Scr. 33:239–259.
Sorenson, M. D. 1999. TreeRot. Boston University, Boston, MA.
SPSS. 2005. SPSS for Macintosh. SPSS Inc., Chicago, IL.
Swofford, D. L. 2002. PAUP∗ . Phylogenetic analysis using parsimony (∗ and
other methods). Sinauer Associates, Sunderland, MA.
Taylor, J., and A. Purvis. 2003. Have mammals and their chewing lice diversified in parallel. Pp. 240–261 in R. D. M. Page, ed. Tangled trees:
phylogeny, cospeciation, and coevolution. Univ. Chicago Press, Chicago.
Türk, E., and F. Türk. 1957. Systematik und Ökologie der Tyroglyphiden
Mitteleuropas. Pp. 4–384 in H. -J. Stammer, ed. Beiträge zur Systematik
und Ökologie mitteleuropäischer Acarina. Band 1. Tyroglyphidae und
Tarsonemini. Akademische Verlagsgesellschaft Geest & Portig K.-G.,
Leipzig.
van Lith, J. P. 1957. On the behaviour of Chaetodactylus mites (Acar., Tyr.)
in the nests of Osmia rufa L. and Chelostoma florisomne (L.) (Apidae,
Megachilidae). Entomol. Berichten 17:197–198.
Vicidomini, S. 1996. Biologia di Xylocopa (Xylocopa) violacea (L., 1758)
(Hymenoptera: Apidae): interazione con Sennertia (Sennertia) cerambycina (Acari Chaetodactylidae). Boll. Zool. Agrar. Bachic. 28:71–76.
Weckstein, J. D. 2004. Biogeography explains cophylogenetic patterns in toucan chewing lice. Syst. Biol. 53:154–164.
Zachvatkin, A. A. 1941. Tiroglifoidnue kleshchi Tyroglyphoidea [ = Tyroglyphoid mites Tyroglyphoidea]. Pp. 1–475 in S. A. Zernov, ed. Fauna
SSSR: Paukoobraznuye. Akademiya Nauk SSSR, Moscow-Leningrad.
Associate Editor: W. O. McMillan
Appendix. Data matrix of morphological characters (based on the character list below) used in phylogenetic reconstruction of Chaetodactylidae relationships: ?, unknown and –, nonapplicable characters; characters 4, 10, and 32 are uninformative. This is an exemplar
subset of a larger matrix containing 46 chaetodactylid taxa (Fig. S1 available online).
1
Centriacarus turbator
Centriacarus guahibo
Roubikia panamensis
Roubikia latebrosa
Achaetodactylus ceratinae
Achaetodactylus leleupi
Ochaetodactylus decellei
Chaetodactylus melitomae
Chaetodactylus ludwigi
Chaetodactylus osmiae
Sennertia zhelochovtsevi
10
20
30
40
50
000000101100001000001000000000000000001010000100001
000100101100001000001000000000000000001010000100001
001-00001100001000000000000001101100002010100100101
001-00001000001000000000000001101100002010100100101
111-01000110111110111000010112101211112101111111111
111-01000110111110111000010112101211112110111111111
111-?1000110111110111000010112111211112110111111111
010010001111111111011111110012101211112110111111110
010010001111111011011111111012101211112110111111110
010010001111111011011111111112101211112110111111110
011-1--11111111110111110111112101211112101111011111
EVOLUTION JUNE 2007
1377
PAVEL B. KLIMOV ET AL.
CHARACTER LIST
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Gnathosomal solenidion: 0, present; 1, absent
Setae on free palpomeres: 0, present; 1, absent
Free palpomeres: 0, present; 1, absent
Free palpomeres: 0, longer than width at base; 1, shorter
than width at base
Alveoli ve: 0, dorsal, distinctly anterior to se; 1, dorsal, approximately at level of se; 2, ventral
Prodorsal shield striation: 0, longitudinal anteriorly and
transverse posteriorly; 1, longitudinal; 2, absent
Posterior edge of prodorsal shield: 0, longer than lateral
edges; 1, shorter than lateral edges
Prodorsal shield: 0, present; 1, absent
Setae se situated: 0, on prodorsal shield; 1, on soft cuticle
Setae si: 0, about twice or more longer than se; 1, less than
twice longer than se
Setae c 2 situated: 0, on same transverse level as c 1 ; 1, distinctly anterior to level of c 1
Setae e 2 situated: 0, on hysterosomal shield; 1, outside hysterosomal shield or touch it
Setae 1a and 3a: 0, touching posterior borders of respective
coxal fields and filiform (conoids in outgroup,; 1, Setae 1a
and 3a not touching posterior borders of respective coxal
fields, if touching then inflated and elongated
Cupules ia situated: 0, on hysterosomal shield; 1, outside
hysterosomal shield
Cupules im situated: 0, at level of acetabules III, approximately at middle of line between setae d 2 and e 2 ; 1, distinctly
posterior to acetabules III, situated off line between d 2 and
e2
Cupules im: 0, ventral, ventro-lateral; 1, dorsal
Cupules ip are: 0, anterior to setae f 2 ; 1, posterior to setae f 2
Cupules ih situated: 0, on sides of attachment organ; 1, incorporated into lateral sclerotized borders of attachment organ
Posterior part of posterior apodemes of coxal fields II : 0, not
displaced posteriorly to anterior apodemes III; 1, displaced
posteriorly to anterior apodemes III
Coxal fields III: 0, enclosed; 1, open
Coxal fields IV: 0, enclosed; 1, open
Transverse medial extension of posterior apodemes IV: 0,
well developed; 1, absent
Anterior extension of posterior apodemes IV: 0, present,
connecting with anterior apodeme III; 1, absent or not connecting
Ventral longitudinal sclerites of progenital chamber at posterior part: 0, conspicuous; 1, inconspicuous
Ventral longitudinal sclerites of progenital chamber at anterior part: 0, conspicuous; 1, inconspicuous
Posterior and lateral cuticular suckers: 0, present; 1, absent
Suckers ad 3 (excluding transparent margin,: 0, larger than
1378
EVOLUTION JUNE 2007
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
inner unsclerotized area of suckers ad 1 +ad 2 ; 1, smaller
or nearly equal to inner unsclerotized area of suckers
ad 1 +ad 2
Anterior cuticular suckers: 0, present; 1, absent or vestigial
Bases of anterior cuticular suckers: 0, inserted on separate
apodeme (may touch or overlap posterio-lateral sclerotized
border of the attachment organ,; 1, incorporated to the border
Apodemes of ps 1 : 0, separated; 1, partially fused anteriorly;
2, completely fused
Setae wa and f I–II: 0, wa I–II submedial, f I–II apical, near
tarsal apices; 1, wa I–II apical or subapical, f I–II at level or
proximal to wa I–II and far from tarsal apices
Solenidion w2: 0, present; 1, absent
Empodial claws I–III: 0, not twisted; 1, twisted
Dorsal cuticular folds of ambulacra I–III: 0, absent; 1,
weakly developed, with distal part smaller than proximal;
2, well developed, with distal part distinctly larger than any
of proximal folds
Condylophores of tarsi I–III: 0, weakly developed, almost
symmetrical; 1, well developed, distinctly assymetrical—
anterior longer, posterior shorter, incorporated into posteriolateral lobe
Supporting sclerites of condylophores (latero-apical sclerites of tarsus,: 0, not distinct from the tarsus, not connected
by dorsal bridge; 1, distinct from the tarsus, connected by
dorsal bridge
Disto-dorsal lobe of distal part of the caruncle: 0, absent; 1,
present, well developed
Dorsal condylar plate of femur-tibia joint: 0, broad; 1, absent
or indistinct
Tarsi I–II with: 0, eight setae (e present,; 1) 7 setae (e absent,
p and q present); 2, 5 setae (e, p, and q absent)
Tarsal setae ra and la I–II: 0, foliate; 1, simple or spiniform
Genual seta cG I : 0, distinctly shorter than genu I and unmodified; 1, longer or slightly shorter than genu I and modified
Genual setae: 0, cG I longer than cG II; 1, cG I–II subequal
Tarsal setae q III: 0, present; 1, absent
Tarsal setae w, r, and p III: 0, present; 1, absent
Tarsal seta s III: 0, foliate; 1, simple
Sigma III: 0, present; 1, absent, represented by alveola
Tarsus IV with: 0, 8 setae (s, p, q present); 1, maximum five
setae (s, p, q always absent)
Tarsal setae e, f IV: 0, foliate or slightly lanceolate; 1, simple
or absent
Tarsal setae w IV: 0, longer than leg IV; 1, distinctly shorter
than leg IV or absent
Tibial setae kT IV: 0, present; 1, absent
Solenidion phi IV: 0, present; 1, absent, represented by alveola
ECOLOGY’S ROLE IN BEE–MITE COEVOLUTION
Supplementary Material
The following supplementary material is available for this article:
Table S1. Aspects of host biology, and in particular nest architecture, considered to investigate the factors influencing associations
between the chaetodactylid mites and their bee-hosts. The variable “cell construction material” was not included in the analysis
because of difficulties with uniform coding and a possibility of model overfitting. Cleptoparasitic bees were also not included
because they do not have chaetodactylids by definition.
Table S2. Logistic regression model for prediction of the presence of chaetodactylid mites in the nests of long-tongued bees
(raw data from Table S1).
Figure S1. Phylogenetic relationships of Chaetodactylidae reconstructed by maximum parsimony analysis of morphological
data; Bremer indices are shown.
This material is available as part of the online article from:
http://www.blackwell-synergy.com/doi/abs/10.1111/j.1558-5646.2007.00119.x
(This link will take you to the article abstract).
Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by
the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
EVOLUTION JUNE 2007
1379