Genetic Divergence Among Massachusetts Populations of
the Vernal Pool Fairy Shrimp Eubranchipus vernalis (Crustacea: Anostraca)
S. Shawn McCafferty, Nicolas Warren, Christopher Wilbur,
and Scott Shumway
Northeastern Naturalist, Volume 17, Issue 2 (2010): 285–304
Full-text pdf (Accessible only to subscribers.To subscribe click here.)
Access Journal Content
Open access browsing of table of contents and abstract pages. Full text pdfs available for download for subscribers.
Current Issue: Vol. 30 (3)
Check out NENA's latest Monograph:
Monograph 22
2010 NORTHEASTERN NATURALIST 17(2):285–304
Genetic Divergence Among Massachusetts Populations of
the Vernal Pool Fairy Shrimp Eubranchipus vernalis
(Crustacea: Anostraca)
S. Shawn McCafferty1,*, Nicolas Warren1,2, Christopher Wilbur1,3,
and Scott Shumway1
Abstract - The vernal pool fairy shrimp Eubranchipus vernalis (Eastern Fairy
Shrimp) is an obligate freshwater invertebrate that is an important member of
northeastern United States vernal pool ecosystems. The extent of gene flow among
populations of vernal pool fairy shrimp in the Northeast is currently unknown, yet
this information is important for understanding the ecology and evolution of this species.
In order to infer the level of gene flow among E. vernalis populations, we used
mtDNA sequence data to estimate the level of genetic divergence and the pattern of
geographic variation among 22 vernal pool populations sampled across Massachusetts
during spring 2003. We found significant differences among vernal pool populations
in the frequency of unique mtDNA sequences, but no coherent geographic pattern in
the distribution of these sequences, and there was no significant correlation between
levels of genetic divergence and geographic distance. We argue that these results
reflect priority effects rather than limited dispersal among these populations.
Introduction
Dispersal of aquatic organisms inhabiting isolated freshwaters has long
been of interest to ecologists. Darwin pondered this question in On The Origin
of Species and later in his last published work (Darwin 1882), inspired
by a small freshwater bivalve attached to the leg of a water beetle sent to him
by W.D. Crick, Francis Crick’s grandfather (Ridley 2004). Darwin reasoned,
and many others have since demonstrated, that dispersal among isolated
freshwater habitats readily occurs even in organisms that lack dispersing
adult stages (reviewed in Bilton et al. 2001). Though understanding the
pattern and rate of dispersal is important to understanding the evolutionary
and ecological mechanisms driving populations of freshwater organisms, it
is gene flow that provides the critical link between dispersal and evolution
(Bohonak 1999). Clearly dispersal and gene flow are related to some degree,
yet dispersal does not necessarily translate into gene flow (DeMeester et al.
2002). Therefore, in order to better understand the ecology and evolution
of freshwater habitats and the species that live there, a key first step is to
understand the extent of gene flow among populations. This understanding is
particularly important in organisms that depend on small isolated freshwater
habitats like vernal pools.
1Department of Biology, Wheaton College, 26 East Main Street, Norton, MA 02766.
2Current address - MRC Room 313, Marine Biological Laboratory, 7 MBL Street,
Woods Hole, MA 02543. 3Current address - Chicago Medical School, Rosalind
Franklin University of Medicine and Science, 3333 Green Bay Road, North Chicago,
IL 60064. *Corresponding author - smccaffe@wheatonma.edu.
286 Northeastern Naturalist Vol. 17, No. 2
Vernal pools are shallow, seasonal water bodies that provide critical
habitat for many woodland amphibian and invertebrate species in New
England (Colburn 2004). They vary in size, depth, hydrology, and degree of
isolation, and characteristically experience a hydrologic cycle of seasonal
flooding and drying that prevents fish populations from becoming established.
Consequently, a variety of organisms lives and breeds only in these
specialized environments. For this reason, vernal pools play an important
role in regional ecosystem biodiversity (e.g., Colburn 2004, DeMeester et
al. 2005, Zedler 2003), yet they are often threatened by man-made changes
to the landscape (Belk 1998, Burne and Griffin 2005, Colburn 2004).
Generally, a defining feature of vernal pools is the lack of any surface
connection with other water bodies, making them, for all practical purposes,
small isolated islands surrounded by inhospitable habitat (Holland and
Jain 1981, Zedler 2003). This isolation is an important constraint on any
obligate vernal pool organism; active dispersal of these organisms among
vernal pools can only occur in adult stages adapted for such dispersal. For
crustaceans lacking such an adult stage (such as fairy shrimp and many
zooplankton species), dispersal can only occur via the passive transport of
adults or diapausing eggs (Bilton et al. 2001, Havel and Shurin 2004, Panov
et al. 2004). Though many authors consider passive dispersal mechanisms
to be quite effective, recently this assumption has come under increasing
scrutiny (e.g., Bohonak and Jenkins 2003, Bohonak et al. 2006, Cáceres and
Soluk 2002).
High levels of dispersal can potentially lead to high levels of gene flow,
which can result in low levels of genetic divergence among populations.
Therefore, one expectation is that there should be little if any genetic divergence
among vernal pool populations of obligate freshwater invertebrates
with diapausing cysts. However, recent studies on various freshwater invertebrates
tend to show just the opposite, i.e., a general pattern of high levels
of genetic divergence among populations, suggesting limited gene flow
(reviewed in DeMeester et al. 2002). These findings suggest an unexpected
disconnect between dispersal and gene flow in some species. How common
this disconnect is in obligate vernal pool species is currently unknown.
Key members of North American vernal pool ecosystems are the diminutive
crustaceans referred to as fairy shrimp (Class Branchiopoda, Order
Anostraca). In fact, fairy shrimp are considered to be a key indicator group
for identifying and certifying vernal pools (Colburn 2004, Kenney and
Burne 2000). These small crustaceans swim ventral side up, are comparatively
weak swimmers that are unable to escape fish and other fast visual
predators, and therefore are restricted to temporary waters. Most fairy
shrimp reproduce sexually, producing diapausing eggs that can withstand
freezing and desiccation. In New England vernal pools, fairy shrimp hatch
from diapausing cysts in the early spring, grow rapidly, mate, and deposit
diapausing cysts in the sediment (where they contribute to the ongoing egg
bank) before dying and completing their life cycle, typically by early May.
In North America, there are seven families of Anostracans. The most common
Anostracans in northeastern vernal pools are in the genus Eubranchipus
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 287
(Family Chirocephalidae; Smith 2001). In Massachusetts, there are two
species of Eubranchipus, the common and widespread E. vernalis (Verrill)
(Eastern Fairy Shrimp) and the more rare E. intricatus (Hartland-Rowe)
(Smooth-Lipped Fairy Shrimp). Eubranchipus vernalis has a fairly wide range
east of the Appalachians, from South Carolina north into Massachusetts and
perhaps New Hampshire and Maine (Belk et al. 1998). Unfortunately very little
is published on E. vernalis biology, though the species’ life history appears
to be consistent with most other Anostracans (summarized in Colburn 2004).
The purpose of this study was to determine the level and pattern of
genetic divergence among Massachusetts populations of the fairy shrimp
E. vernalis. We describe for the first time the levels and distribution of genetic
variation within and among northeastern US vernal pool populations of
an obligate freshwater invertebrate based on mitochondrial DNA (mtDNA)
sequence data. The levels of genetic divergence among populations, the distribution
of unique mtDNA sequences, and the relationship between levels of
genetic divergence and geographic distance are used to infer the geographic
structure of vernal populations in this recently deglaciated landscape. We
interpret our results with respect to levels of dispersal and gene flow among
vernal pools.
Methods
Study design and field collections
To determine the level of genetic divergence among vernal populations,
the authors and volunteers with the Vernal Pool Association (http://www.vernalpool.
org, Reading, MA) used dipnets to collect samples of fairy shrimp
from 22 vernal pools across Massachusetts during Spring 2003 (Fig. 1).
Locales and sample sizes are presented in Table 1 (latitudes and longitudes
are available from S.S. McCafferty upon request). Distances between the
pools range from 0.1 km to more than 200 km. Individuals were identified to
species according to Kenney and Burne (2000). All samples were preserved
in 95% ethanol in the field.
Laboratory methods
The use of mtDNA in studies of geographic variation among populations
has been an accepted approach for over 25 years and the methods for
studying genetic variation in mtDNA are now fairly routine (Avise 2004).
Genomic DNA is extracted from individual samples, and a region of the
mitochondrial genome is sequenced using the polymerase chain reaction
(PCR) coupled with direct sequencing. The resulting sequences can be used
to estimate both the evolutionary relationship (phylogeny) among unique
mtDNA sequences (haplotypes) and the frequency and occurrence of haplotypes
within and among populations using routine methods.
Genomic DNA was extracted from individual fairy shrimp using the
Wizard Genomic DNA Extraction Kit (Promega Corp., Madison, WI) or
by a standard chelex method (Walsh et al. 1991). Roughly 700 nucleotide
288 Northeastern Naturalist Vol. 17, No. 2
base pairs (bp) of the mitochonrial cytochrome oxidase I (COI) gene region
were amplified in a standard PCR reaction using the primer pairs HCO2198
(Folmer et al. 1994) and FS-COIL (5’ CTGCTGGGTCACAGAATGAA 3’)
under the following conditions: 94 °C for 2 minutes, followed by 35 cycles
of 94 °C for 30 seconds, 52 °C for 30 seconds, and 72 °C for one minute,
followed by a finishing step of 72 °C for five minutes. For each sample,
the resulting PCR reactions produced a single fragment that was cleaned
(Wizard Gel/PCR Cleanup Kit, Promega Corp., Madison, WI, or AmPure,
Argencourt Bioscience Corp, Beverley, MA) and sequenced (BigDye 3.1,
Applied Biosystems, Foster City, CA) in both directions. The resulting DNA
sequences were analyzed on an ABI 310 Genetic Analyzer (Applied Biosystems,
Foster City, CA) following the manufacturer’s recommendations.
Statistical and phylogenetic analysis of mtDNA
The analytical approach taken here is fairly routine when dealing with
population samples of DNA sequence data (Avise 2004). First, the data
are manually edited and aligned to detect ambiguities or problems in the
sequences. Once aligned, haplotypes are identified, and their frequency
estimated for each population. The phylogenetic relationship among haplotypes
is inferred and subsequently used to determine if any pattern exists
between the evolutionary relationship of the different haplotypes and their
geographic origin, referred to as their phylogeographic relationship.
Figure 1. Map of 22 sampling locations in Massachusetts. Locale ID numbers are by
town alphabetically: (1) Belchertown, (2) Boxford, (3) Concord, (4) East Brookfield,
(5) Groton, (6) Ipswich, (7) Leverett, (8) Nantucket, (9) Newburyport, (10) Norfolk,
(11) North Hampton, (12) North Reading, (13–14) Norton, (15) Rowley, (16) Sturbridge,
(17) Sudbury, (18) Wellfleet, and (19–22) Westborough.
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 289
Table 1. Frequency of mtDNA haplotypes from 22 vernal pool locations throughout Massachusetts. The order of haplotypes follows the evolutionary relationship
among haplotypes found in Figure 2. The order of the first 14 locales (n > 4) follows the level of genetic similarity based on the frequency of haplotypes determined
as in Figure 3. The remaining 8 locales (n < 4) are ordered alphabetically. Locale ID numbers were determined alphabetically. π is the level of nucleotide
diversity within a population, estimated only for those populations with n > 4.
ID Town Code O H10 H12 H13 H11 H4 H2 H3 H9 H8 H7 H1 H6 H5 n π
21 Westborough3 ANF 4 5 1 10 0.01011
22 Westborough4 ANN 1 8 4 13 0.00970
17 Sudbury SUD 3 2 1 6 0.00467
6 Ipswich IPS 2 2 1 5 0.00300
10 Norfolk NOR 4 1 5 0.00133
8 Nantucket NAN 6 1 1 8 0.00470
15 Rowley ROW 1 1 5 1 8 0.00315
7 Leverett LEV 1 4 5 0.00200
19 Westborough1 W5 7 3 10 0.00233
4 East Brookfield EBRK 5 5 0.00000
5 Groton GRO 5 5 0.00000
13 Norton NORT 7 7 0.00000
14 Norton Woods NORTW 1 5 6 0.00167
20 Westborough2 W9 10 10 0.00000
1 Belchertown BEL 1 1 2 -
2 Boxford BOX 4 4 -
3 Concord CON 1 1 -
9 Newburyport NEW 1 1 2 -
11 North Hampton NHAM 2 1 3 -
12 North Reading NRED 2 2 -
16 Sturbridge STU 1 1 2 -
18 Wellfleet CCNS 2 1 3 -
Total 1 14 17 2 2 11 3 4 6 2 1 14 5 40 122
290 Northeastern Naturalist Vol. 17, No. 2
The resulting sequences were edited by eye and aligned using Sequencer
4.1 (Gene Codes Corp., Ann Arbor, MI). The aligned sequences were imported
into MacClade 4.1 (Maddison and Maddison 2001), manually rechecked, and
trimmed to 600 bp. An unweighted pair group method with arithmetic mean
(UPGMA) tree was constructed based on the simple percent difference (with
pair-wise deletion) among all mtDNA sequences using MEGA 3.1 (Kumar et
al. 2004), and the resulting tree was used as a guide to determine the number of
haplotypes. Representative haplotype sequences were constructed using majority
rule consensus in MacClade when ambiguous base calls were present.
The phylogenetic relationship among the resulting unique mtDNA haplotypes
was inferred by neighbor joining (NJ) based on Kimura’s 2-parameter model,
using the program Mega 3.1. In addition, a nonhierarchical phylogenetic view
of the relationship among haplotypes was inferred by statistical parsimony using
the program TCS (Clement et al. 2000).
Nucleotide diversity (π) within vernal pools was estimated using the program
Arlequin 3.11 (Excoffier et al. 2005). To test for spatial patterns in the
data, an analysis of molecular variation (AMOVA) and estimates of pairwise
Φst (analogous to Fst) were used to test for significant differences in mtDNA
haplotype frequencies among populations (vernal pools) using Arlequin 3.11.
Φst is a commonly used measure of genetic divergence among populations
based on the variance in allele (or haplotype) frequencies. Values of Φst range
from a low of 0.0 (no divergence) to a high of 1.0. AMOVA tests for significant
differences in the frequency of genetic markers within and among hierarchical
groups of populations and is analogous to an analysis of variance. The key difference
is that AMOVA also factors in levels of genetic divergence. Because
there was no a priori hierarchical structure to our sampling scheme, the AMOVA
analysis simply tested for significance among vernal pools.
The relationship among vernal pool populations based on the population
pairwise estimates of Φst (negative values were converted to 0.000) was
visualized using hierarchical clustering (UPGMA) and the nonhierarchical
ordination method nonmetric multidimensional scaling (NMDS) coupled
with a minimum spanning tree using the program NTSYSpc (Rohlf 2002).
Used together, these methods provide a powerful multivariate approach for
detecting groups of populations based on overall levels of genetic similarity
(Allendorf and Luikart 2007).
Isolation by distance (IBD), the relationship between geographic distance
and genetic divergence (Φst), was tested using the program IBDWS
(Bohonak 2002, Jensen et al. 2005) with 1000 permutations of the Mantel
(1967) test. Geographic distance among vernal pools was calculated from
latitude and longitude using the geographic distance matrix generator (Estes
2007). If the latitude and longitude of a particular vernal pool was unknown,
we used the latitude and longitude of the town center where the vernal pool
was located as a proxy estimate.
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 291
Results
Thirteen unique mtDNA sequences, arbitrarily labeled haplotypes H1–
H13, were found in 121 E. vernalis from 22 vernal pools distributed across
Massachusetts (GENBANK accession numbers EU169880-EU169893). A
single individual E. intricatus was also sequenced and used as an outgroup
in the phylogenetic analyses. There were 18 variable nucleotide positions
found in the aligned sequences, 16 of which allowed us to determine the
relationship among haplotypes (parsimony informative). No insertions or
deletions were found, and within E. vernalis, all substitutions were transitions
(C/T or G/C) located at the third codon position, meaning that none of
these substitutions leads to an amino acid substitution. The levels of nucleotide
diversity within populations were relatively low, ranging from 0.00000
to 0.01010 with an average of 0.00775 ± 0.00213 (Table 1).
The relationship among the 13 haplotypes can be found in the NJ tree
and as an unrooted parsimony network in Figure 2. The two trees are very
similar in terms of the evolutionary relationship or branching order among
haplotypes. The 13 haplotypes are closely related and generally fall into two
primary groups (clades), one consisting of haplotypes H10–H13, the other of
haplotypes H1–H9. There is a tendency for subgroups to form within these
two main clades (H11–H13; H1–H6, H7–H9), though these are not well
supported by bootstrap analysis, a measure of confidence in the resulting
branching pattern (results not presented).
The frequency of the 13 haplotypes can be found in Table 1. The most
common haplotype is H5, which occurs in a little over a third of the pools and
Figure 2. The phylogenetic relationship among haplotypes as determined by (a) the
neighbor joining (NJ) method and (b) statistical parsimony (see text for details).
292 Northeastern Naturalist Vol. 17, No. 2
represents 10–100% (when n > 4) of the populations where it occurs. Three
other haplotypes (H1, H4, and H12) are less common, occurring in 5–6 of the
pools at frequencies of 12–80%. The remaining eight haplotypes (H2, H3,
H6, H7, H8, H9, H10, H11, and H13) are relatively rare, occurring in only
1–3 of the sampled pools, although at within-pool frequencies of up to 62%.
Several haplotypes are confined to one (H6, H7, and H13) or two (H7, H9,
and H11) populations. No single haplotype is found in all populations.
There is a highly significant difference among populations based on the
frequency of mtDNA haplotypes, with a little over half the variation accounted
for among populations (Table 2). We include only those populations
with a sample size of n > 4, which results in a final data set containing 104 of
the 121 original sequences from 14 populations and all 13 haplotypes. The
pairwise estimates of Φst are highly variable (ranging from 0.000 to 0.961)
but on the average quite high (mean Φst of 0.490 ± 0.038; Table 3).
We find no geographically coherent pattern in the relationship among
vernal pools based on the Φst values. The results of the UPGMA and NMDS
(Fig. 3) are very similar; those populations forming clusters in the UPGMA
are also closely associated in the ordination. There is a tendency for the
vernal pool populations to fall into three groups, one consisting of vernal
pools from East Brookfield (EBRK), Groton (GRO), Westborough (W9),
and Norton (NORT and NORTW), a second group consisting of pools from
Nantucket (NAN), Sudbury (SUD), and Westborough (ANF and ANN),
and a third group consisting of pools from Ipswich (IPS), Leverette (LEV),
Norwood (NOR), Rowley (ROW), and Westborough (W5). However, these
groups do not appear to have any coherent geographic basis. There is extensive
geographic overlap in the distribution of the three groups (Fig. 4),
clearly demonstrating a lack of any coherent geographic pattern in the data.
In addition, there is no significant relationship between geographic proximity
and cluster membership (tested by a Mantel test, P = 0.85), and there is
no obvious geographic feature separating the three groups.
The clustering of vernal pools into these three groups appears to be due to
the frequency of haplotype H5 and of clades 1 and 2 (Fig 2). Group 1 (EBRK,
GRO, NORT, NORTW, and W9) is characterized by either fixation for or a
high frequency of haplotype H5. Group 2 is distinguished from Group 3 by
having a high frequency of haplotypes from clade 2, while Group 3 is dominated
by haplotypes from clade 1. These results are clearly seen in Figure 3
where the frequency of the two clades is plotted on the results from the NMDS.
Table 2. Results of the AMOVA testing for significant differences among vernal pool populations
of E. vernalis (***P < 0.001).
Variance Percentage
Source of variation d.f. Sum of squares components of variation
Among populations 13 136.976 1.30157 53.59***
Within populations 88 99.200 1.12727 46.41
Total 101 236.176 2.42884
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 293
Table 3. Pairwise estimates of Φst and geographic distance among pools. The Φst values are below the diagonal, the estimates of geographic distance (km) above.
ANF ANN EBRK GRO IPS LEV NORT NW NAN NOR ROW SUD W5 W9
ANF 0.000 0.16 34.23 39.46 80.34 74.23 52.20 49.40 167.67 30.12 74.53 22.82 2.97 2.98
ANN 0.047 0.0000 0.16 34.23 39.46 80.34 74.23 52.20 49.40 167.67 30.12 74.53 2.82 2.97
EBRK 0.316 0.526 0.000 57.97 111.36 44.80 79.45 78.18 193.05 61.01 104.35 55.05 36.34 36.36
GRO 0.316 0.526 0.000 0.000 60.52 78.08 81.13 76.66 191.72 58.45 51.75 28.49 41.16 41.15
IPS 0.249 0.432 0.654 0.654 0.000 138.45 85.61 80.36 167.06 73.92 9.56 57.53 79.70 79.67
LEV 0.288 0.494 0.813 0.813 0.440 0.000 123.66 121.96 237.83 103.71 129.81 89.52 76.97 76.98
NORT 0.369 0.566 0.000 0.000 0.709 0.846 0.000 5.34 115.50 23.20 85.53 52.68 49.23 49.22
NW 0.280 0.507 0.000 0.000 0.479 0.605 0.028 0.000 118.41 19.59 80.20 48.16 46.44 46.24
NAN 0.305 0.325 0.773 0.773 0.586 0.658 0.803 0.728 0.000 137.98 172.91 164.07 164.7 164.69
NOR 0.232 0.449 0.867 0.867 0.145 0.359 0.891 0.672 0.600 0.000 71.53 30.32 27.19 27.18
ROW 0.348 0.494 0.571 0.571 0.228 0.537 0.620 0.458 0.651 0.514 0.000 51.80 74.18 74.16
SUD 0.278 0.313 0.791 0.791 0.596 0.651 0.824 0.733 0.000 0.618 0.652 0.000 22.38 22.36
W5 0.331 0.545 0.577 0.577 0.433 0.049 0.620 0.353 0.705 0.422 0.483 0.696 0.000 0.02
W9 0.428 0.612 0.000 0.000 0.764 0.878 0.000 0.091 0.835 0.915 0.673 0.858 0.667 0.000
294 Northeastern Naturalist Vol. 17, No. 2
Furthermore, there is no obvious coherent spatial pattern based on the frequency
of the two clades when mapped onto sampling locales (Fig. 4), reinforcing
our conclusion that the three groups are not defined by geographic proximity.
This overall lack of coherent geographic structure is further seen when
testing for isolation by distance. No significant relationship is found between
geographic distance and genetic divergence Φst (P = 0.557; P = 0.727 for the
log-log analysis; Fig. 5).
Discussion
Our results clearly demonstrate significant differences among fairy shrimp
populations in the frequency and occurrence of mtDNA haplotypes. However,
even though we find a significantly high level of spatial variation among
populations (Φst values and the AMOVA results), no coherent geographic
pattern can be detected. There are no distinctive geographic boundaries
that would suggest population breaks, and there is no relationship between
Figure 3. The relationship among sites (n > 4) as determined by (a) unweighted pair
group method with arithmetic mean (UPGMA) and (b) nonmetric multidimensional
scaling (NMDS) of Φst values (see text for details). A minimum spanning tree connecting
similar populations is overlain on the plot. Abbreviation for locations can be
found in Table 1. Also shown are the frequencies of clade 1 (black) and 2 (white)
found at each site.
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 295
geographic distance and levels of genetic divergence. Populations that are
geographically close to each other are no more genetically similar to each
other than they are to populations that are more geographically distant.
There are two important aspects of our results. The first is finding significant
levels of genetic divergence among vernal pool populations of
Figure 4. (a) Plot of the three groups of populations identified by UPGMA and NMDS
(Fig. 3) onto their geographic location. Group 1 is black, Group2 is grey, and Group 3
is white. (b) Plot of the frequency of mtDNA clades 1 (black) and 2 (white) onto their
geographic locations. All locations sampled are plotted in (b), whereas only locations
with n > 4 are plotted in (a).
296 Northeastern Naturalist Vol. 17, No. 2
E. vernalis. This result is fairly consistent with previous studies of other
fairy shrimp species based on protein electrophoresis (Bohonak 1998, Boileau
et al. 1992, Davies et al. 1997, Hulsmans et al. 2007, Ketamaier et al.
2003, Meglecz and Theiry 2005). Those few examples of low levels of genetic
divergence tended to be among populations that were geographically
very close (e.g,. Bohonak 1998, Riddoch et al. 1994). In a general review
of the literature on Anostracans, Hulsmans et al. (2007) suggest an overall
trend for high levels of genetic differentiation on geographic scales of
greater than 0.1 km, with low levels of differentiation at geographic scales
of less than 0.1 km due to the effects of dispersal and gene flow. Similar
patterns of high levels of divergence among populations are also reported
for a variety of other diapausing freshwater invertebrates (e.g., DeMeester
et al. 2002, Marten et al. 2006, Zeller et al. 2006, and references therein),
suggesting this may in fact be a general phenomenon in organisms with this
particular reproductive strategy.
The second aspect of our results that is worth discussing is that even
though we find significant levels of divergence among populations, there
is no coherent geographic pattern to this variation. How do we explain this
seemingly paradoxical result? Though phylogeographic boundaries (defined
here as large breaks in the distribution of haplotypes that correspond to
some geographic feature) have been reported in other studies of diapausing
freshwater invertebrates, these were invariably found among populations
separated by very large geographic distances (hundreds to thousands of
kilometers, e.g., De Gelas and DeMeester 2005, Dooh et al. 2006, and references
therein). The general tendency is for studies at medium or smaller
geographic scales to show limited if any evidence of phylogeographic structure
(see references above). This can be attributed to the effects of ongoing
dispersal and gene flow coupled with a general absence of any major geologic
features that would interfere with historical or contemporary dispersal.
Our data would certainly be consistent with these findings. Considering the
Figure 5. Plot
of untransformed
pairwise
genetic
distance (Φst)
versus geographic
distance
(km).
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 297
geographic range of our sampling sites, the topography of eastern and central
Massachusetts, and the potential for high rates of dispersal among fairy
shrimp populations, there may be no compelling reason to expect substantial
phylogeographic structure among vernal pool populations in this region.
Though dispersal and gene flow tend to erase phylogeographic signals,
they can, however, lead to a correlation between levels of genetic divergence
and geographic distance among populations. Termed isolation by distance
(IBD), this relationship is expected to occur among populations if levels
of gene flow are dependent on geographic distance (e.g., Hutchison and
Templeton 1999, Slatkin 1993). As geographic distance among populations
increases, levels of gene flow tend to decrease. This trend leads to increasing
levels of divergence among populations that are geographically distant,
which translates into a positive relationship between levels of genetic divergence
and geographic distance.
Bohonak (1998) interpreted the results from previous studies on various
species of fairy shrimp as demonstrating clear evidence for IBD. In fact, a
general consensus seems to be emerging from the literature that IBD may
be common in diapausing freshwater species. However, this relationship
may also depend on geographic scale. IBD has been demonstrated among
populations that are geographically close, but the relationship appears to
break down at larger geographic distances (Haag et al. 2006, Hulsman et al.
2007, Zeller et al. 2006). At what point IBD breaks down tends to be species
specific, ranging from ca. 0.1 km in Branchipodopsis wolfiDaday and in
two species of Daphnia (Haag et al. 2006, Hulsman et al. 2007), to between
0.28 km and 50 km in Branchinecta sandiegonensis Fugate (San Diego Fairy
Shrimp) (Davies et al. 1997), to less than 100 km in a species of diapausing
copepod (Zeller et al. 2006). In E. vernalis, we found no evidence for IBD.
However, this may simply be a reflection of the geographic scale of our sampling,
since only 3 of the 91 possible pairwise comparisons were less than 1
km distant (mean distance of 71.5 km ± 5.27 km). However, it is interesting
to note that the only comparison less than 0.1 km in our data has a Φst value
of 0.667, which suggests a high level of divergence even in geographically
close populations and an overall lack of IBD in E. vernalis.
Dispersal, gene flow, and priority effects
Our results strongly suggest that there is limited gene flow among Massachusetts
vernal pool populations of E. vernalis, resulting in a pattern of
high levels of divergence among vernal pool populations with little coherent
geographic pattern. How do we account for this limited gene flow? Either
(a) there is limited dispersal among populations, which translates into limited
gene flow, or (b) dispersal is occurring but it does not lead to gene flow
among populations; in other words, there is a decoupling between dispersal
and gene flow.
Limited dispersal means limited gene flow, which results in increasing
levels of genetic divergence among populations over time. It is generally
assumed that diapausing cysts are an effective mechanism of dispersal in
298 Northeastern Naturalist Vol. 17, No. 2
obligate freshwater invertebrates. However, Bohonak and Jenkins (2003)
convincingly argue that passive dispersal via diapausing cysts may not
be as frequent as commonly assumed. Without clear direct estimates, the
assumption of passive dispersal should be carefully evaluated on a species
by species basis. However, though Bohonak and Jenkins’ (2003) argument is
compelling and would certainly explain our results, there are aspects of the
life history of E. vernalis that suggest an alternative explanation.
Many early studies on the population genetics of freshwater organisms
with passive dispersal revealed unanticipated high levels of genetic divergence
among populations (Boileau et al. 1992, DeMeester et al. 2002). This
result was considered a paradox; the life-history strategy clearly suggests
substantial levels of dispersal, yet the genetic data point otherwise. Boileau
et al. (1992) first proposed a solution to this apparent paradox, which
they termed persistent founder effects. In this model, colonization of new
habitats by a limited number of individuals followed by rapid population
expansions lead to the preservation of founder effects (stochastic differences
among populations) over substantial periods of time (Boileau et al. 1992).
This long-term lack of equilibrium between gene flow and drift can result
in a random spatial pattern of genetic divergence that can persist even when
levels of dispersal are high. Because the populations are at or near carrying
capacity, there is little opportunity for immigrant genotypes to contribute to
the population, dampening the homogenizing effects of gene flow. In effect,
dispersal occurs, but it is effectively decoupled from gene flow. The presence
of a large cyst bank would further reinforce the long-term persistence
of founder effects by swamping out any immigrant genotypes entering the
population (Hairston 1996). In a further expansion of this model, DeMeester
et al. (2002) proposed that rapid local adaptation tended to even further accentuate
persistent founder effects by locking out immigrant genotypes with
lower fitness, a model they referred to as the monopolization hypothesis.
There are various aspects of the life history of E. vernalis and the nature
of vernal pools that makes persistent founder effects and/or the monopolization
hypothesis (here we combine the two models under the general category
of priority effects) compelling explanations for our results. First, populations
of E. vernalis in New England tend to be rather young, reflecting colonization
events that can be no more than 11,000–14,000 years old, based on the glacial
history of eastern North America (Ruddiman 1987). Some populations may
be even younger, considering the agricultural history of this region may have
led to extensive habitat destruction prior to recent reforestation (Gerhardt and
Foster 2002, Hall et al. 2002). Because of their relatively young ages, these
populations may not have had sufficient time to reach an equilibrium with respect
to gene flow and drift (e.g., Nurnberger and Harrison 1995), a situation
that can lead to persistent founder effects. Second, population sizes of E. vernalis
vary greatly by location and year, though they can be quite high in some
instances (M. Burne, Vernal Pool Association, MA, pers. observ.). Rapidly
fluctuating populations can also lead to a lack of equilibrium between gene
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 299
flow and drift. Third, though no formal estimates on cyst-bank size has been
reported in the literature for E. vernalis, it is reasonable to assume they tend to
be rather high. For example, estimates of cyst-bank density from other Anostraca
tend to be around 103 to 105 per square meter (Brendonck and DeMeester
2003, Hulsman et al. 2006). In addition, hatching success from cysts may also
be quite variable. Hatching rates appear to be highly dependent on the environmental
conditions, ranging from around 3% to 85%, depending on the species
and environment (Belk 1998, Hulsman et al. 2006). A large cyst bank coupled
with variable hatching rates would act to limit the contribution of immigrant
genotypes to a population.
The physical and chemical nature of vernal pools also suggest that
priority effects may play an important role in determining the pattern of
genetic divergence we find in E. vernalis. Vernal pools tend to be relatively
limited but variable in size. Persistent founder effects are thought to be
more important in species occupying small habitats than in those occupying
larger, more stable environments (Boileau and Taylor 1994, DeMeester
et al. 2002, Vanoverbeke et al. 2007). In addition, vernal pools can vary
extensively in their hydrology, water and sediment chemistry, and other
variables such as dissolved oxygen and temperature (Colburn 2004). Recent
studies have shown that variation in physical and chemical variables
play an important role in determining the community composition of vernal
pools (E.A. Colburn, Harvard Forest, Petersham, MA, unpubl. data).
This variation could lead to adaptive differences among vernal fairy shrimp
populations as has been shown for other freshwater crustaceans (De-
Meester et al. 2002). Though we have no evidence for the role of natural
selection in driving levels of genetic divergence among vernal pool fairy
shrimp populations, and we certainly are not implying selection acting on
the mitochondrial gene region studied here, adaptive differences among
populations could further reduce levels of gene flow among populations,
as proposed in the monopolization hypothesis (DeMeester et al 2002). Observed
geographic patterns of genetic divergence among populations would
then result from an interaction of persistent founder effects and natural
selection with the end result being a complex spatial pattern among vernal
pools, with no clear phylogeographic signal.
Priority effects versus lack of dispersal
Which explanation bests accounts for our results, a lack of dispersal
among vernal pools or priority effects? Resolving this issue clearly depends
on direct estimates of dispersal among vernal pools. However, though passive
dispersal may not be as prevalent as previously thought (Bohonak and
Jenkins 2003), it is clear that some level of dispersal occurs in diapausing
invertebrates, as demonstrated in recent studies using experimental pools
(Cáceres and Soluk 2002, Frish and Green 2007, Louette et al. 2007). Extrapolating
from these results should proceed cautiously, however, since we
should expect dispersal rates to vary greatly among species regardless of
the fact that they share the common character of having a diapausing cyst.
300 Northeastern Naturalist Vol. 17, No. 2
Finding strong evidence for dispersal in one species having a diapausing egg
stage need not imply extensive dispersal occurs in all such species.
One point does seem to be generally agreed upon: dispersal does occur
at least on a microgeographic scale. Geographically proximate populations
may experience a fairly high level of dispersal among populations due to
wind, waterfowl, and other animal vectors (Bilton et al. 2001, Bohonak
and Jenkins 2003, DeMeester et al. 2002, Figuerola et al. 2005, Green and
Figuerola 2005). Therefore, the primary mechanism driving patterns of divergence
in species with passive dispersal may vary according to geographic
scale. Priority effects may explain high levels of population differentiation
in geographically close populations, while a lack of dispersal leading to
genetic drift (stochastic lineage sorting) predominates at greater geographic
distances. That transition zone where we see a shift from the lack of gene
flow being driven by priority effects versus lack of dispersal should vary by
species according to species specific patterns of dispersal and should be reflected in the observed patterns of genetic divergence among populations.
In our particular example, the extent of contemporary dispersal among
vernal pool populations of E. vernalis is unknown at any scale; we are not
aware of any direct estimates of dispersal in this species. It is difficult to
imagine a lack of contemporary dispersal of fairy shrimp among Massachusetts
vernal pools when clearly it had to have occurred in the recent past,
but this is of course no evidence that dispersal continues today. Regardless,
the high levels of genetic divergence among E. vernalis populations, a lack
of any coherent phylogeographic association among haplotypes, and no
isolation by distance all suggest limited gene flow among Massachusetts’
populations of the vernal pool fairy shrimp. Whether this is due to priority
effects or simply limited dispersal remains to be determined. Be that as it
may, we have clearly demonstrated that vernal pool fairy shrimp populations
tend to be genetically distinct entities with limited gene flow among them.
In light of the growing awareness of the critical link between community
ecology and evolutionary processes (e.g., DeMeester et al. 2007, Emmerson
and Gillespie 2008, Stockwell et al. 2003, Vellend and Geber 2005), these
results lend valuable insights to our overall understanding of the ecology and
evolution of vernal pool organisms.
Acknowledgments
The majority of the laboratory research reported here was performed by C.
Wilbur and N. Warren under the guidance of S.S. McCafferty as part of their undergraduate
Independent Research Projects (BIO499) at Wheaton College. The authors
wish to thank Leo Kenny, Matt Burne, and the numerous Vernal Pool Association
volunteers for their invaluable assistance in collecting samples, and E. Colburn and
two anonymous reviewers for their many valuable and constructive comments. This
study was funded in part by the Wheaton College Grants for Undergraduate Research
to C. Wilbur and N. Warren, the Merck Foundation, and Richard White and Sons Science
Fund.
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 301
Literature Cited
Allendorf, F.W., and G. Luikart. 2007. Conservation and the Genetics of Populations.
Blackwell Publishing Ltd, Malden, MA. 642 pp.
Avise, J.C. 2004. Molecular Markers, Natural History, and Evolution. Sinauer Associates,
Inc. Sinauer, MA. 684 pp.
Belk, D. 1998. Global status and trends in ephemeral pool invertebrate conservation:
Implications for Californian Fairy Shrimp. Pp. 147–150, In C.W. Witham, E.T.
Bauder, D. Belk, W.R. Ferren, Jr., and R. Ornduff (Eds.). Ecology, Conservation,
and Management of Vernal Pool Ecosystems—Proceedings from a 1996 Conference.
California Native Plant Society, Sacramento, CA.
Belk, D., G. Mura, and S.C. Weeks. 1998. Untangling confusion between Eubranchipus
vernalis and Eubranchipus neglectus (Branchiopoda: Anostraca). Journal of
Crustacean Biology 18:147–152.
Bilton, D.T., J.R. Freeland, and B. Okamura. 2001. Dispersal in freshwater invertebrates.
Annual Review of Ecology and Systematics 32:159–181.
Bohonak, A.J. 1998. Genetic population structure of the fairy shrimp Branchinecta
coloradensis (Anostraca) in the Rocky Mountains of Colorado. Canadian Journal
of Zoology 76:2049–2057.
Bohonak, A.J. 1999. Dispersal, gene flow, and population structure. Quarterly Review
of Biology 74:21–45.
Bohonak, A.J. 2002. IBD (Isolation By Distance): A program for analyses of isolation
by distance. Journal of Heredity 93:153–154.
Bohonak, A.J., and D.G. Jenkins. 2003. Ecological and evolutionary significance of
dispersal by freshwater invertebrates. Ecological Letters 6:783–796.
Bohonak, A.J., M.D. Holland, B. Santer, M. Zeller, C.M. Kearns, and N.G. Hairston.
2006. The population genetic consequences of diapause in Eudiaptomus copepods.
Archiv für Hydrobiologie 167:183–202.
Boileau, M.G., and B.E. Taylor. 1994. Chance events, habitat age, and the genetic
structure of pond populations. Archiv für Hydrobiologie 132:191–202.
Boileau, M.G., P.D.N. Hebert, and S.S. Schwartz. 1992. Nonequilibrium gene frequency
divergence: Persistent founder effects in natural populations. Journal of
Evolutionary Biology 5:25–39.
Brendonck, L., and L. DeMeester. 2003. Egg banks in freshwater zooplankton: Evolutionary
and ecological archives in the sediment. Hydrobiologia 491:65–84.
Burne, M.R., and C.R. Griffin. 2005. Protecting vernal pools: A model from Massachusetts,
USA. Wetlands Ecology and Management 13:367–375.
Cáceres, C.E., and D.A. Soluk. 2002. Blowing in the wind: A field test of overland
dispersal and colonization by aquatic invertebrates. Oecologia 131:402–408.
Clement, M., D. Posada, K.A. Crandall. 2000. TCS: a computer program to estimate
gene genealogies. Molecular Ecology 9:1657–1659.
Colburn, E.A. 2004. Vernal Pools: Natural History and Conservation. McDonald and
Woodward Publishing Company, Blacksburg, VA.
Darwin, C. 1882. On the dispersal of freshwater bivalves. Nature 25:1882.
Davies, C.P., M.A. Simovich, and S.A. Hathaway. 1997. Population genetic structure
of a California endemic branchiopod, Branchinecta sandiegonensis. Hydrobiologia
359:149–158.
De Gelas, K., and L. DeMeester. 2005. Phylogeography of Daphnia magna in Europe.
Molecular Ecology 14:753–764.
302 Northeastern Naturalist Vol. 17, No. 2
DeMeester, L., A. Gomez, B. Okamura, and K. Schwenk. 2002. The monopolization
hypothesis and the dispersal-gene flow paradox in aquatic organisms. Acta
Oecologica 23:121–135.
DeMeester, L., S. Declerck, R. Stoks, G. Louette, F. van de Meutter, T. de Bie, E.
Michels, and L. Brendonck. 2005. Ponds and pools as model systems in conservation
biology, ecology, and evolutionary biology. Aquatic Conservation: Marine
and Freshwater Ecosystems 15:715–725.
DeMeester, L., G. Louette, C. Duvivier, C. Van Damme, and E. Michels. 2007.
Genetic composition of resident populations influences establishment success of
immigrant species. Oecologia 153:431–440.
Dooh, R.T., J. Adamowicz, and P.D.N. Hebert. 2006. Comparative phylogeography
of two North American “glacial relict” crustaceans. Molecular Ecology
15:4459–4475.
Emerson, B.C., and R.G. Gillespie. 2008. Phylogenetic analysis of community assembly
and structure over space and time. Trends in Ecology and Evolution,
23:619–630.
Estes, P.J. 2007. Geographic Distance Matrix Generator (version 1.2.0). American
Museum of Natural History, Center for Biodiversity and Conservation. Available
online from http://geospatial.amnh.org/open_source/gdmg. Accessed
April 2007.
Excoffier, L., G. Laval, and S. Schneider. 2005. Arlequin ver. 3.0: An integrated software
package for population genetics data analysis. Evolutionary Bioinformatics
Online 1:47–50.
Figuerola, J., A.J. Green, and T.C. Michot. 2005. Invertebrate eggs can fly: Evidence
of waterfowl-mediated gene flow in aquatic invertebrates. American Naturalist
165:274–280.
Folmer, O., M. Black, W.R. Hoeh, R. Lutz, and R.C. Vrijenhoek. 1994. DNA primers
for amplification of mitochondrial cytochrome c oxidase subunit I from
diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology
3:294–299.
Frish, D., and A.J. Green. 2007. Copepods come first: Rapid colonization of new
temporary ponds. Fundamental and Applied Limnology 168:289–297.
Gerhardt, F., and D.R. Foster. 2002. Physiographical and historical effects on
forest vegetation in central New England, USA. Journal of Biogeography
29:1421–1437.
Green, A.J., and J. Figuerola. 2005. Recent advances in the study of long distance
dispersal of aquatic invertebrates via birds. Diversity and Distributions
11:149–156.
Haag, C.R., M. Riek, J.W. Hottinger, V.I. Pajunen, and D. Ebert. 2006. Founder
events as determinants of within-island and among-island genetic structure of
Daphnia metapopulations. Heredity 96:150–158.
Hairston, N.G. 1996. Zooplankton egg banks as biotic reservoirs in changing environments.
Limnology and Oceanography 41:1087–1092.
Hall, B., G. Motzkin, D.R. Foster, M. Syfert, and J. Burk. 2002. Three hundred years
of forest and land-use change in Massachusetts, USA. Journal of Biogeography
29:1319–1335.
Havel, J.E., and J.B. Shurin. 2004. Mechanisms, effects, and scales of dispersal in
freshwater zooplankton. Limnology and Oceanography 49:1229–1238.
2010 S.S. McCafferty, N. Warren, C. Wilbur, and S. Shumway 303
Holland, R.F., and S.K. Jain. 1981. Vernal pools. Pp. 515–36, In M.G. Barbour and
J. Major (Eds.). Terrestrial Vegetation of California. California Native Plant Society,
Sacramento, CA.
Hulsman, A., S. Bracke, K. Moreau, B.J. Riddoch, L. DeMeester, and L. Brendonck.
2006. Dormant egg-bank characteristics and hatching pattern of the Phallocryptus
spinosa (Anostraca) population in the Makgadikgadi Pans (Botswana). Hydrobiologia
571:123–132.
Hulsmans, A., K. Moreau, L. DeMeester, B.J. Riddock, and L. Brendonck. 2007.
Direct and indirect measures of dispersal in the fairy shrimp Branchiopodopsis
wolfiindicate a small-scale isolation-by-distance pattern. Limnology and Oceanography
52:676–684.
Hutchison, D.W., and A.R. Templeton. 1999. Correlation of pairwise genetic and
geographic distance measures: Inferring the relative influences of gene flow and
drift on the distribution of genetic variability. Evolution 53:1898–1914.
Jensen, J.L., A.J. Bohonak, and S.T. Kelley. 2005. Isolation by distance, web service.
BMC Genetics 6:13. Available online at http://ibdws.sdsu.edu/.
Kenney, L.P., and M.R. Burne. 2000. A Field Guide To The Animals of Vernal Pools.
MA Division of Fisheries and Wildlife, Westborough, MA. 73 pp.
Ketamaier, V., P. Zarattini, E. DeMatthaeis, M. Cobolli, and G. Mura. 2003. Intraand
inter-specific relationships in the six Italian species of the fairy shrimp genus
Chirocephalus: combining allozyme and mtDNA data. Journal of Zoology, Systematics,
and Evolutionary Research 41:276–285.
Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: Integrated Software for Molecular
Evolutionary Genetics Analysis and Sequence Alignment. Briefings in
Bioinformatics 5:150–163.
Louette, G., J. Vanoverbeke, R. Ortells, and L. DeMeester. 2007. The founding
mothers: The genetic structure of newly established Daphnia populations. Oikos
116:728–741.
Maddison, D.R., and W.P. Maddison. 2001. MacClade, Version 4.08. Sinauer Associates,
Sunderland, MA.
Mantel, N. 1967. The detection of disease clustering and a generalized regression
approach. Cancer Research 27:209–220.
Marten, A., M. Brändle, and R. Brandl. 2006. Habitat type predicts genetic population
differentiation in freshwater invertebrates. Molecular Ecology 15:2643–2651.
Meglecz, E., and A. Theiry. 2005. Genetic population structure of a rare, temporary
pond-dwelling fairy shrimp Linderiella massaliensis (Crustacea: Anostraca).
Archiv fur Hydrobiologia 164:277–286.
Nurnberger, B., and R.G. Harrison. 1995. Spatial population structure in the whirligig
beetle Dineutus assimilis: Evolutionary inferences based on mitochondrial
DNA and field data. Evolution 49:266–275.
Panov, V.E., P.I. Krylov, and N. Riccardi. 2004. Role of diapause in dispersal and
invasion success by aquatic invertebrates. Journal of Limnology 63(Suppl.
1):56–69.
Riddoch, B.J., S.W. Mpoloka, and M. Cantrell. 1994. Genetic variation and localized
gene flow in the fairy shrimp, Branchipodopsis wolfi, in temporary rainwater
pools in southeastern Botswana. Pp. 96–102, In A.R. Beaumont (Ed.). Genetics
and Evolution of Aquatic Organisms. Chapman and Hall, New York, NY.
Ridley, M. 2004. Crick and Darwin’s shared publication in Nature. Nature 431:244.
304 Northeastern Naturalist Vol. 17, No. 2
Rohlf, J.F. 2002. NTSYSpc: Numerical Taxonomy and Mutivariate Analysis System,
ver 2.1. Exeter Software, Setauket, NY.
Ruddiman, W.F. 1987. Synthesis: The ocean/ice sheets record. Pp. 463–478, In W.F.
Ruddiman and H.E. Wright (Eds.). North America and Adjacent Oceans During
the Last Deglaciation. The Geology of North America, Vol. K3. Geological Society
of America, Boulder, CO.
Slatkin, M. 1993. Isolation by distance in equilibrium and nonequilibrium populations.
Evolution 47:264–279.
Smith, D.G. 2001. Pennak’s Freshwater Invertebrates of the United States, Fourth
Edition. John Wiley and Sons, New York, NY.
Stockwell, C.A., A.P. Hendry, and M.T. Kinnison. 2003. Contemporary evolution
meets conservation biology. Trends in Ecology and Evolution 18:94–101.
Vanoverbeke, J., K. De Gelas, and L. DeMeester. 2007. Habitat size and the genetic
structure of a cyclical parthenogen, Daphnia magna. Heredity 98:419–426.
Vellend, M., and M.A. Geber. 2005. Connections between species diversity and genetic
diversity. Ecology Letters 8:767–781.
Walsh, S.P., D.A. Metzger, and R. Higuchi. 1991. Chelex 100 as a medium for simple
extraction of DNA for PCR-based typing from forensic material. Biotechniques
10:506–513.
Zedler, P. 2003. Vernal pools and the concept of ‘‘isolated wetlands.’’ Wetlands
23:597–607.
Zeller, M., T.B.H. Reusch, and W. Lampert. 2006. A comparative population genetic
study on calanoid freshwater copepods: Investigation of isolation-by-distance in
two Eudiaptomus species with a different potential for dispersal. Limnology and
Oceanography 51:117–124.