FirstCite?
e-publishing
Use of diploid male
frequency data as an
indicator of pollinator
decline
Amro Zayed1*, David W. Roubik2
and Laurence Packer1
1Department of Biology, York University, 4700 Keele Street, Toronto,
Ontario M3J 1P3, Canada
2Smithsonian Tropical Research Institute, Box 2072, Balboa, Ancon,
Republic of Panama
*Author for correspondence (amro@torontodigital.com).
Recd 10.08.03; Accptd 17.09.03; Online 31.10.03
Pollination deficits in agricultural and natural sys-
tems are suggestive of large reductions in pollinator
populations. However, actual declines are difficult
to demonstrate using census data. Here, we show
census data to be misleading because many abun-
dant pollinators exhibit high levels of production
of sterile diploid males usually found only in
small inbred hymenopteran populations; Euglossa
imperialis exhibits high levels of diploid male pro-
duction induced by low effective population sizes
(Ne 
 15), despite being the most abundant orchid
bee in lowland tropical forests in Panama. We cau-
tion that although some pollinators appear abundant
on the basis of census data, their long-term persist-
ence may be highly tenuous based on genetic evi-
dence. We propose the use of diploid male frequency
data as a metric for assessing the sustainability of
bee populations.
Keywords: pollinator decline; diploid males; Euglossini;
Hymenoptera
1. INTRODUCTION
Owing to their role in pollination, bees are crucial for the
maintenance of terrestrial biodiversity and their conser-
vation is thereby of prime importance (Kearns et al. 1998;
Kevan 1999). Pollination deficits suggest that bees are
undergoing large population reduction (Kevan & Phillips
2001); however, actual pollinator declines have been diffi-
cult to demonstrate using census data (Roubik 2001;
Williams et al. 2001, but see Frankie et al. 1997; Kevan
et al. 1997). Because our ability to conserve pollinators
depends on our ability to detect when their populations
are at risk, developing accurate parameters to assess pol-
linator declines is essential. Genetic parameters have the
potential to indicate small populations at risk of extir-
pation (Avise & Hamrick 1996; Petit et al. 1998). How-
ever, their role in bee conservation remains largely
unexplored as standard conservation genetic models are
not directly applicable to haplodiploids (Packer & Owen
2001).
One criterion that we show is ideal for detecting loss of
genetic diversity in haplodiploids is the production of dip-
loid males (Packer & Owen 2001; Zayed & Packer 2001).
Proc. R. Soc. Lond. B (Suppl.) ? 2003 The Royal Society
DOI 10.1098/rsbl.2003.0109
Diploid males arise from fertilized eggs homozygous at the
sex determination locus, whereas heterozygotes develop
into females, and hemizygotes into haploid males (Cook &
Crozier 1995). Through their sterility or inviability
(Stouthamer et al. 1992; Agoze et al. 1994), the pro-
duction of diploid males is disadvantageous because it
increases the genetic load and decreases reproductive fit-
ness (Page 1980; Ross et al. 1993) and, in social species,
increases colony mortality and decreases colony growth
rates (Plowright & Pallett 1979; Ross & Fletcher 1986).
However, natural populations usually maintain many
alleles at the sex locus and thus diploid males occur at
very low frequencies (Cook & Crozier 1995). The number
of alleles at the sex locus is maintained at mutation-drift
equilibrium, with larger populations maintaining more
alleles than smaller ones (Yokoyama & Nei 1979).
A study of Euglossine-orchid bees in Panama revealed
that many species have unusually large numbers of diploid
males, with frequencies ranging from 12% to 100%
(Roubik et al. 1996). This is paradoxical considering that
censuses indicate that orchid bees exist in large popu-
lations (Roubik 2001). High frequencies of diploid males
were not found in Brazilian orchid bees (Takahashi et al.
2001), prompting our research into the causes of such an
unusual phenomenon in apparently healthy Panamanian
bee populations. Here, we provide an explanation for this
paradox. Through an extensive study of one species,
Euglossa imperialis Cockerell, the most abundant orchid
bee in lowland forest in Panama (Roubik 2001), we con-
firm that there are indeed high levels of diploid males and
that this results from small effective population sizes (Ne).
We highlight the factors responsible for creating the large
discrepancies between the census population size and Ne
in this species, and discus the utility of data on the pro-
duction of diploid males as indicators of the long-term
viability of pollinator populations.
2. MATERIAL AND METHODS
Euglossa imperialis males were collected using cineole and methyl
salicylate baits (Ackerman 1983) during February 2002. Site sym-
bols, names and sample sizes are: (A) Pipeline Road-40; (B) El Llano
Carti Road-18; (C) Howard Airforce Base-32; (D) Fort Clayton-21;
(E) Barro Colorado Island-93; (F) Cerro Jefe-41; (G) Fort Sherman-
60; (H) Madden Dam-11; (I) Cerro Campana-30; (J) Gigante Pen-
insula-51; (K) Cerro Hoya-94; (L) Cuesta de Piedra-25; (M) Santa
Rita Ridge-99; (N) Maria Chiquita-25; and (O) Rio Caimito-55.
Genetic variation was assayed at the allozyme loci esterase (EST),
hexokinase (HK) and phosphoglucomutase (PGM) following
Packer & Owen (1990). The frequency of diploid males, 
, was esti-
mated for each locus using maximum likelihood (Owen & Packer
1994). The effective number of alleles K was estimated given the
expected frequency of haploids in a population h (with a 1 : 1 sex ratio;
Peruquetti & Campos 1997) using the equation (derivation to be
presented elsewhere) K = (1  h)(1  
 )/h
. Single-locus estimates of

 and K were averaged over loci to obtain population values.
Ne was estimated in two ways, as follows.
(i) Iteratively by fixing a mutation-drift model (Yokoyama & Nei
1979, eqn (30)) for the frequency of sex locus homozygotes
given the average K estimated and assuming a mutation rate
of u = 105.
(ii) By regressing the log of pairwise geneflow (Nm), estimated from
FST (Wright 1951) against the log of pairwise distance for all
population pairs.
Ne is equivalent to 10b, where b is the y-intercept of the isolation by
distance plot (Slatkin 1993). We simulated diploid populations of
equal sizes to those sampled based on male allele frequencies to esti-
mate FST using FSTAT 2.9.3 (Goudet 1995). Average pairwise FST
from 100 simulations was used to estimate Nm.
Rates of heterozygosity loss by drift in ideal isolated populations
were estimated following Crow & Kimura (1970). Populations were
03BL0243.S2 A. Zayed and others Assessing pollinator decline
L
(mono)
Republic of Panama
K (27%, 3)
O
(mono)
I
(mono)
(b) B (52%, 1) G
(25%, 4)
J
(52%, 1)
N (35%, 2)
F
(mono)M (28%, 3)
H (39%, 2)
A (13%, 7)
D (56%, 3)
C
(mono)
E
(16%, 7)
(b)(a)
Figure 1. (a) Map of Panama and (b) magnification of boxed section, showing sampling locations, average frequencies of
males that are diploid 
, and the effective number of sex alleles K for Euglossa imperialis. Monomorphic populations are
labelled ?(mono)?.
classified as monomorphic if the frequency of the most common allele
at all loci was greater than 95%.
To estimate the census population size, we used data recorded dur-
ing May 1980 to May 2000 (Roubik 2001). Because only males are
attracted to the baits, the total population size is double the number
of males observed, given a 1 : 1 sex ratio (Peruquetti & Campos
1997).
3. RESULTS
Ten out of the 15 populations sampled were polymor-
phic, with an average expected heterozygosity of
Hexp = 0.257. The average proportion of males that are
diploid 
 is 34.3% (ranging from 13% to over 50%), with
a mean of only 3.1 sex alleles (figure 1). Using the
mutation-drift model, we obtained an estimate of
Ne = 23.1 individuals. The isolation by distance model
provided an independent estimate of Ne = 7.3 individuals
(95% CI = 3.5 ? 17), giving an average for the two esti-
mates of only 15.2 individuals, representing ca. 5% of the
average censused population size of 295.
4. DISCUSSION
Populations of E. imperialis in Panama have unusually
high frequencies of diploid males induced by chronically
low Ne. Our finding of five monomorphic populations
(figure 1), representing a third of the populations sampled,
is consistent with low Ne and extreme loss of allelic diver-
sity. Complete loss of heterozygosity from the average of
Hexp = 0.257 found in polymorphic populations, occurs in
50 generations in an isolated population with Ne = 15.2,
but requires 967 generations given the census population
size. How can we reconcile our low estimates of Ne with
the high census estimates for E. imperialis?
First, population sizes of orchid bees in Panama vary
dramatically over time (Roubik 2001) and Ne is bound by
the lowest census size in a series of fluctuations
(Nunney & Elam 1994). Second, male reproductive suc-
cess is influenced strongly by the complexity of the chemi-
cal bouquets that they obtain from flowers, particularly
rare orchids (Roubik 1998; Eltz et al. 1999). High variance
in reproductive success, combined with overlapping gen-
erations (Roubik 1989), considerably reduce Ne with
respect to N (Felsenstein 1971; Crow & Denniston 1988).
Also, census estimates are based on sampling both the
Proc. R. Soc. Lond. B (Suppl.)
Figure 2. Male orchid bees (species from the genuses
Euglossa, Eulaema, Eufriesea and Exaerete are depicted in the
above photograph) are easily censused using chemical baits.
However, despite their apparent abundance based on
census data, several species exhibit a genetic phenomenon
prevalent in small haplodiploid populations: the production
of sterile diploid males. Census data obtained from baiting
greatly overestimate the effective population size of these
important pollinators, and often lack sufficient power to
detect pollinator declines. Photograph, D.W.R. and M.
Guerra.
haploid and the diploid males attracted to chemical baits
(figure 2; Ackerman 1983), but sterile diploid males do
not contribute to reproduction (Cook & Crozier 1995).
Furthermore, since diploid males arise from fertilized
eggs, they are failed attempts at female production, and
their presence in large frequencies drains the population
of females, which also reduces Ne (Crozier 1976). Taken
in combination, these factors suggest why census data
overestimate Ne by more than an order of magnitude.
Finally, the dynamics of metapopulation extinction and
recolonization, which can reduce the effective population
size by several orders of magnitude (Gilpin 1991; McCau-
ley 1993), might be responsible for the low effective popu-
lation size of E. imperialis. Although it is generally believed
that Euglossines are capable of long-range dispersal, based
on their flight abilities (Roubik 1989), our study shows
Assessing pollinator decline A. Zayed and others 03BL0243.S3
that gene flow is geographically restricted. Geographical
variance in the number of sex alleles (figure 1) and
expected heterozygosity across Panama, are consistent
with a metapopulation structure with a few large source
populations and many small sink populations.
Orchid bees play a major role in neotropical ecosystems.
The males are the sole pollinators of several hundred spec-
ies of orchids, and females pollinate many trees, shrubs
and vines (Ramirez et al. 2002). Small Ne, as shown for
E. imperialis, is probably occurring in other Panamanian
orchid bees, several of which have even higher estimates
of production of diploid males (Roubik et al. 1996). Sig-
nificantly, many of our sample sites lie in protected forests,
some of which are over 10 000 ha in area. The long-term
persistence of these bees is unlikely simply on genetic
grounds (Saccheri et al. 1998; Westemeier et al. 1998;
Higgins & Lynch 2001), and indeed, E. imperialis popu-
lations, although apparently large, are declining even in
extensively protected forests (Roubik 2001).
Our findings have far-reaching implications for the con-
servation of the Hymenoptera. First, the frequency of dip-
loid males is a sensitive indicator of genetic diversity and
its loss. Baseline data on 
 and K for native bee species
should be compiled to facilitate future population moni-
toring. Second, the conservation status of hymenopteran
populations should be addressed with genetic data, as
assessments based solely on census data can be mislead-
ing. Additionally, high temporal variability in censused
abundance of tropical and temperate pollinators impedes
the detection of meaningful trends (Roubik 2001;
Williams et al. 2001). Genetic parameters such as Ne and

 are insensitive to such short-term fluctuations. Further
diploid male frequency data can be easily gathered and
translated to Ne, making the Hymenoptera a good indi-
cator taxon to assess the current health of ecosystems.
Acknowledgements
The authors thank Mr R. Hartmann for providing accommodation
in Santa Clara, Mr B. Boyd for providing access to Rio Caimito, J.
Grixti and G. Arbabi for assistance in running gels, J. Young for
sharing expertise on iterative functions, and two anonymous referees
for helpful comments on the manuscript. This work was supported
by an American Orchid Society grant to the authors, a NSERC grant
to L.P., and NSERC and OGS scholarships to A.Z.
Ackerman, J. D. 1983 Specificity and mutual dependency of the
orchid?euglossine bee interaction. Biol. J. Linn. Soc. 20, 301?314.
Agoze, M. E., Drezen, J. M., Renalt, S. & Preiquet, G. 1994 Analysis
of the reproductive potential of diploid males in the wasp
Diadromus pulchellus (Hymenoptera: Ichneumonidae). Bull. Ento-
mol. Res. 84, 213?218.
Avise, J. C. & Hamrick, J. L. (eds) 1996 Conservation genetics, case
histories from nature. New York: Chapman & Hall.
Cook, J. M. & Crozier, R. H. 1995 Sex determination and population
biology of the Hymenoptera. Trends Ecol. Evol. 10, 281?286.
Crow, J. F. & Denniston, C. 1988 Inbreeding and variance effective
population numbers. Evolution 42, 482?495.
Crow, J. F. & Kimura, M. 1970 An introduction to population genetics
theory. New York: Harper & Row.
Crozier, R. H. 1976 Counter-intuitive property of effective popu-
lation size. Nature 262, 384.
Eltz, T., Whitten, W. M., Roubik, D. W. & Linsenmair, K. E. 1999
Fragrance collection, storage, and accumulation by individual male
orchid bees. J. Chem. Ecol. 25, 157?176.
Felsenstein, J. 1971 Inbreeding and variance effective numbers in
populations with overlapping generations. Genetics 68, 581?597.
Proc. R. Soc. Lond. B (Suppl.)
Frankie, G. W., Vinson, S. B., Rizzardi, M. A., Grisold, T. L.,
O?Keefe, S. & Snelling, R. R. 1997 Diversity and abundance of
bees visiting a mass flowering tree species in disturbed sesonal dry
forest, Costa Rica. J. Kansas Entomol. Soc. 70, 281?296.
Gilpin, M. E. 1991 The genetic effective size of a metapopulation.
Biol. J. Linn. Soc. 42, 165?175.
Goudet, J. 1995 FSTAT (vers. 1.2): a computer program to calculate
F-statistics. J. Heredity 86, 485?486.
Higgins, K. & Lynch, M. 2001 Metapopulation extinction caused by
mutation accumulation. Proc. Natl Acad. Sci. USA 98, 2928?2933.
Kearns, C. A., Inouye, D. W. & Waser, N. M. 1998 Endangered
mutualisms: the conservation of plant?pollinator interactions. A.
Rev. Ecol. Syst. 29, 83?112.
Kevan, P. G. 1999 Pollinators as bioindicators of the state of the
environment: species, activity and diversity. Agric. Ecosyst. Environ.
74, 373?393.
Kevan, P. G. & Phillips, T. P. 2001 The economic impacts of pollin-
ator declines: an approach to assessing the consequences. Conserv.
Ecol. 5, 8. See http://www.consecol.org/vol5/iss1/art8.
Kevan, P. G., Greco, C. F. & Belaoussoff, S. 1997 Log-normality of
biodiversity and abundance in diagnosis and measuring eco-
systemic health: pesticide stress on pollinators on blueberry heaths.
J. Appl. Ecol. 34, 1122?1136.
McCauley, D. E. 1993 Genetic consequences of extinction and reco-
lonization in fragmented habitats. In Biotic interactions and global
change (ed. P. M. Karevia, J. G. Kingsolver & R. B. Huey), pp.
217?233. Sunderland, MA: Sinauer.
Nunney, L. & Elam, D. R. 1994 Estimating the effective population
size of conserved populations. Conserv. Biol. 8, 175?184.
Owen, R. E. & Packer, L. 1994 Estimation of the proportion of
diploid males in populations of Hymenoptera. Heredity 72, 219?
227.
Packer, L. & Owen, R. 1990 Allozyme variation, linkage disequilib-
rium and diploid male production in a primitively social bee
Augochlorella striata (Hymenoptera: Halictidae). Heredity 65,
241?248.
Packer, L. & Owen, R. 2001 Population genetic aspects of pollinator
decline. Conserv. Ecol. 5, article 4. See http://www.consecol.org/
vol5/iss1/art4.
Page, R. E. 1980 The evolution of multiple mating behavior by honey
bee queens (Apis mellifera). Genetics 96, 263?273.
Peruquetti, R. C. & Campos, L. A. O. 1997 Aspectos da biologia de
Eufriesea violacea (Blanchard) (Hymenoptera, Apidae, Euglossini).
Rev. Bras. Zool. 14, 91?97.
Petit, R. J., Mousadik, A. E. & Pons, O. 1998 Identifying populations
for conservation on the basis of genetic markers. Conserv. Biol. 121,
844?855.
Plowright, R. C. & Pallett, M. J. 1979 Worker-male conflict and
inbreeding in bumble bees (Hymenoptera: Apidae). Can. Entomol.
111, 289?294.
Ramirez, S., Dressler, R. L. & Ospina, M. 2002 Abejas euglosinas
(Hymenoptera: Apidae) de la Regio?n Neotropical: Listado de
especies con notas sobre su biolog??a. Biota Colombiana 3, 7?118.
Ross, K. G. & Fletcher, D. J. C. 1986 Diploid male production?a
significant colony mortality factor in the fire ant Solenopsis invicta
(Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 19, 283?291.
Ross, K. G., Vargo, E. L., Keller, L. & Trager, J. C. 1993 Effect of
a founder event on variation in the genetic sex-determining system
of the fire ant Salenopsis invicta. Genetics 135, 843?854.
Roubik, D. W. 1989 Ecology and natural history of tropical bees. Cam-
bridge University Press.
Roubik, D. W. 1998 Grave-robbing by male Eulaema (Hymenoptera,
Apidae): implications for euglossine biology. J. Kansas Entomol.
Soc. 71, 188?191.
Roubik, D. W. 2001 Ups and downs in pollinator populations: when
is there a decline? Conserv. Ecol. 5, article 2. See http://www.
consecol.org/vol5/iss1/art2.
Roubik, D. W., Weigt, L. A. & Bonilla, M. A. 1996 Population gen-
etics, diploid males, and limits to social evolution of euglossine
bees. Evolution 50, 931?935.
Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius,
W. & Hanski, I. 1998 Inbreeding and extinction in a butterfly
metapopulation. Nature 392, 491?494.
Slatkin, M. 1993 Isolation by distance in equilibrium and non-
equilibrium populations. Evolution 47, 264?279.
Stouthamer, R., Luck, R. F. & Werren, J. H. 1992 Genetics of sex
determination and the improvement of biological control using
parasitoids. Environ. Entomol. 21, 427?435.
03BL0243.S4 A. Zayed and others Assessing pollinator decline
Takahashi, N. C., Peruquetti, R. C., Del Lama, M. A. & de Oliveira
Campos, L. A. 2001 A reanalysis of diploid male frequencies in
euglossine bees (Hymenoptera: Apidae). Evolution 55, 1897?
1899.
Westemeier, R. L., Brawn, J. D., Simpson, S. A., Esker, T. L.,
Jansen, R. W., Walk, J. W., Kershner, E. L., Bouzat, J. L. & Paige,
K. N. 1998 Tracking the long-term decline and recovery of an iso-
lated population. Science 282, 1695?1698.
Williams, N., Minckley, R. & Silviera, F. 2001 Variation in native
bee faunas and its implications for detecting community changes.
Proc. R. Soc. Lond. B (Suppl.)
Conserv. Ecol. 5, article 7. See http://www.consecol.org/vol5/
iss1/art7.
Wright, S. 1951 The genetical structure of populations. Ann. Eugenics
15, 323?354.
Yokoyama, S. & Nei, M. 1979 Population dynamics of sex-determin-
ing alleles in honey bees and self-incompatibility alleles in plants.
Genetics 91, 609?626.
Zayed, A. & Packer, L. 2001 High levels of diploid male production
in a primitively eusocial bee (Hymenoptera: Halictidae). Heredity
87, 631?636.