Site by Bennett Web & Design Co.
2012 SOUTHEASTERN NATURALIST 11(2):219–238
Population Characteristics of the Mussel Villosa iris (Lea)
(Rainbow Shell) in the Spring River Watershed, Arkansas
Allison M. Asher1,2 and Alan D. Christian1,3,*
Abstract - the goal of this study was to better understand population characteristics of
Villosa iris (Lea) (Rainbow Shell) in the Spring River drainage in north-central Arkansas
through documenting seasonal spatial patterns, movement behavior, population size,
size-frequency distributions, sex ratios, and fecundity. We conducted monthly mark and
recapture sampling between May and September 2007 (i.e., before, during, and after
spawning) and documented the sex, size, fecundity, and spatial location of individual
Rainbow Shell at 2 sites (SFSR1 and SR1). Population estimates were relatively high at
both sites with 166 ± 32 (SE) and 381 ± 37 (SE) individuals at SFSR1 and SR1, respectively.
Sex ratio was highly skewed toward males at SFSR1 with a ratio of 1.0:2.6, but only
slightly skewed at SR1 with a ratio of 1.0:1.3. Mean fecundity was 27,849 ± 11,653 (SE)
and 15,089 ± 11,710 (SE) glochidia at SFSR1 and SR1, respectively. Spatially, statistically
more males were found upstream of non-gravid females during the spawning period.
Mean movement for all sampling events was 1.6 ± 0.53 cm/day and 1.9 ± 0.58 cm/day
for SFSR1 and SR1, respectively. Home range was 29.3 ± 27.7 cm2 and 43.0 ± 42.5 cm2 for
SFSR1 and SR1, respectively. From our study, we conclude that Rainbow Shell exhibits
traits, such as male-skewed sex ratio and non-uniform distribution of males and females,
that may influence fertilization rates of females.
Freshwater mussels (Bivalvia: Unioniformes) reach their greatest richness in
North America with ≈300 taxa (Bogan and Roe 2008); however, starting in the
early 1990s, it was documented that ≈70% of these taxa were imperiled (Williams
et al. 1993, Lydeard et al. 2004). These apparent declines resulted in development
of a national strategy for the conservation of native freshwater mussels (National
Native Mussel Conservation Committee 1998) that spurred basic and applied
biological and environmental research on freshwater mussels. However, for
many species, there is still a paucity of information on basic biology, especially
concerning reproductive behavior and biology.
Freshwater mussels move both vertically and horizontally; however, the
reasons for this behavior are not well understood but may be associated with
reproduction. Vertical movement, burrowing into substrate and rising to the
substrate surface, has been associated with day length (Perles et al. 2003,
Schwalb and Pusch 2007), water temperature (Amyot and Downing 1997,
1Environmental Sciences Graduate Program, Arkansas State University, PO 847, State
University, AR 7246. 2Current address - Department of Zoology, Southern Illinois University,
1125 Lincoln Drive, Carbondale, IL 62901. 3Current address- Department of
Biology, University of Massachusetts Boston, 100 Morrissey Boulevard, Boston, MA
02125. * Corresponding author - email@example.com.
220 Southeastern Naturalist Vol. 11, No. 2
Schwalb and Pusch 2007, Watters et al. 2001), discharge (Schwalb and Pusch
2007) and spawning (Watters et al. 2001). Vertical movement is believed to be
an avoidance response to unfavorable conditions such as high water velocities
(Di Maio and Corkum 1995) or predation (Amyot and Downing 1997). It is
plausible that horizontal movement across the substrate surface could occur for
similar reasons. For example, horizontal movement in freshwater mussels has
been shown to increase spatial aggregation (Amyot and Downing 1998, Balfour
and Smock 1995, Downing and Downing 1992, Downing et al. 1993) and
is thought to occur in response to reproductive efforts to increase the chance
for successful reproduction (Amyot and Downing 1998; Downing and Downing
1992; Downing et al. 1989, 1993).
Information relating to population characteristics, especially reproductive
characteristics, of mussels is more abundant, but still lacking for most species.
Two characteristics that are of particular interest in studying reproduction are
sex ratios and fecundity. Sex ratios in mussels vary by species, with femaleskewed
(Garner et al. 1999), near equal, (Hanlon and Levine 2004, Rogers
et al. 2001, Yeager and Neves 1986), and male-skewed sex ratios varying by
drainages (Hagg and Staton 2003). Mean fecundity values range from lows of
9647 for Quadrula asperata (Lea) (Alabama Orb), 23,890 for Fusconaia cerina
(Conrad) (Southern Pigtoe), and 25,767 for Obliquaria reflexa Rafinesque
(Three-horn Wartyback), to higher values of 281,776 for Lampsilis ornata
(Conrad) (Southern Pocketbook), 325,709 for Amblema plicata (Say) (Threeridge),
and 566,000 for L. siliquoidea (Barnes) (Fatmucket) (Haag and Staton
2003, Perles et al. 2003). Furthermore, fecundity typically increases with
length (Haag and Staton 2003); however, this relationship is not always observed
(Perles et al. 2003).
Villosa iris (Lea) (Rainbow Shell) has a global conservation rank of G5Q
(common, widespread, abundant, but taxonomic classification is a matter
of conjecture among scientists) and a state conservation ranking in Arkansas of
S2S3 (imperiled, very few populations; vulnerable, relatively few populations)
(Harris et al. 2009). As its state status suggests, little is known about Rainbow
Shell populations in Arkansas. Villosa iris is dioecious, has a maximum
reported shell length of 75 mm, and is considered sexually dimorphic (Parmalee
and Bogan 1998, Williams et al. 2008). Reproductively, Rainbow Shell is
a long-term brooder (bradytictic) (Watters et al. 2001, Williams et al. 2008)
and broods glochidia from May to July (Parmalee and Bogan 1998). It should
be noted that the current taxonomic status of Rainbow Shell is uncertain as a
recent phylogenetic analysis across the range revealed that the genus Villosa is
polyphyletic with at least 9 clades that include species from 5 genera (Kuehnl
2009, Williams et al. 2008).
The goal of this study was to document population characteristics of Rainbow
Shell at 2 sites in the Spring River drainage, AR (Spring River and South
Fork Spring River) and to compare these characteristics through time and
between sites. The first objective was to determine population characteristics
2012 A.M. Asher and A.D. Christian 221
(i.e., population size, size frequency, fecundity, sex ratio) at each location and
to compare these characteristics between sites and among sampling events.
The second objective was to determine movement and spatial patterns (i.e.,
distance traveled between captures, total displacement, male and female spatial
patterns) of Rainbow Shell for each sampling event at each site. To do so,
we conducted a monthly mark and recapture study of Rainbow Shell from May
to September 2007 and recorded the location, sex, and reproductive status
(gravid or not gravid) of females.
Two sampling sites were established in the 2 main streams of the Spring
River drainage that is located in the Ozark Mountains of north-central Arkansas
and southwestern Missouri. One site was located in the South Fork Spring River
(SFSR1), and the other site was located in the Spring River (SR1) (Fig. 1). The
SFSR begins in southeastern Missouri and flows southeast for 120 river km to
its confluence with the Spring River just upstream of Hardy, AR. The underlying
geology of the SFSR watershed is primarily limestone and land use is primarily
pasture land (Martin 2008, Martin et al. 2009). Martin et al. (2009) reported
Rainbow Shell from 11 locations within the SFSR, including the population
we investigated for this study (SFSR1). Based on surface-exposed mussels, the
mussel assemblage at SFSR1 was composed of 8 species, occupied an area approximately
70 m in length and 10 m in width, was located in a run habitat on
the left descending bank upstream of the low-water bridge on Red Bud Road
(County Road 26), and had substrate composed of primarily cobble and fines
Figure 1. Villosa iris sampling sites on the South Fork Spring River (SFSR1; GPS coordinates:
15S 603340 4037709 [UTM]) and Spring River (SR1; GPS coordinates: 15S
633870 4022487 [UTM]) in Fulton County, AR. Map layers downloaded from GeoStor
222 Southeastern Naturalist Vol. 11, No. 2
(Martin 2008). The SR begins at Mammoth Spring, AR, which discharges ≈34
million liters of water/hour (Trauth et al. 2007), and flows south 92 river km
until its confluence with the Black River. The geology and land use of the SR is
similar to the SFSR, with underlying limestone and land use of primarily pasture
land. Trauth et al. (2007) reported Rainbow Shell from 4 locations within the
SR, including the population we studied (SR1; Fig. 1). Based on surface-exposed
mussels, the mussel assemblage at SR1 was composed of 3 species, occupied an
area approximately 73 m long and 13 m wide, was located in a run habitat on the
left descending side channel, and had substrate composed of boulders with fines
and cobble (Trauth et al. 2007).
Mark and recapture sampling
Mark and recapture sampling was conducted monthly at each site from May
to September, 2007. However, we collected preliminary data in July 2006 at
SR1 and included these data to estimate population size, calculate sex ratio,
and compare shell length at SR1. During each sampling event, we surveyed
mussels using a single snorkeling pass. Only mussels that could be visually
observed at the substrate surface were located and marked with surveying flags
after being completely removed from the substrate and being visually identified
as Rainbow Shell; in other words, buried mussels were not excavated from
below the substrate.
Spatial locations of flagged Rainbow Shell were determined by establishing a
long-term X-axis on the first sampling date by inserting metal rods into the soil
on the right and left banks. These rods remained in place throughout the sampling
period and were documented by recording the GPS coordinates of both rods.
During each sampling event, a temporary X-axis was established by attaching
a meter tape to the right and left bank rods. A Y-axis was created by using another
meter tape run parallel to the banks. The flagged mussel X and Y location
was recorded to the nearest 0.01 meter. Variation in measurements of individual
mussels, i.e. measurement error, was accounted for at each site by establishing
a “dummy” marker within the study reach at each site. Dummy markers were
measured exactly as flagged mussels. This allowed us to estimate the precision
of our X and Y measurements at each site.
Each previously unmarked Rainbow Shell collected was marked by etching
a unique identification code on its right valve using a Dremel® tool. For each
individual collected, we recorded the unique identification code, measured and
recorded anterior-to-posterior shell length to the nearest 0.1 mm, and recorded
the sex and reproductive status (gravid or not gravid) of females. The reproductive
status of females was determined by gently prying open the valves and
observing the gills for glochidia. Individuals were returned to their original location
after being processed.
Fecundity was estimated by counting glochidia collected from 5 gravid females
at SR1 in July 2006 and 6 gravid females at SFSR1 in May 2007. Glochidia
were harvested in two ways. At SR1, gravid females were collected, and pre2012
A.M. Asher and A.D. Christian 223
served in 100% ethanol, and glochidia were removed from the gills in the laboratory.
At SFSR1, glochidia were collected from females in the field by prying
open the valves, flushing de-ionized water over the gills, collecting the flushed
glochidia, and preserving the glochidia in 100% ethanol. Females from SFSR1
were returned after glochidia harvest to the exact location they were collected. In
the laboratory, the number of glochidia per female was determined by counting 5
glochidia subsamples from each female as follows. Each female’s glochidia were
placed in a known volume of ethanol. The glochidia and ethanol were thoroughly
mixed, a 1-mL subsample of the glochidia and ethanol mixture was removed and
discharged into a gridded petri dish, and glochidia were counted under a dissection
microscope. This was repeated 5 times per female. Total fecundity per
female was estimated by multiplying the mean number of glochidia per/mL by
the total volume of the glochidia and ethanol mixture.
Calculations and statistical analyses
Population estimates at each site were derived using mark and recapture
methods outlined in Schumacher and Eschmeyer (1943) and only when a few individuals
could be collected in each sampling event. Briefly, the population sizes
(N) of Rainbow Shell were estimated using the formula:
N = (Σk[n2(m + u)]) / Σk(nm),
where N is the population estimate, k represents the total number of sampling
events, n is the total number of marked individuals from prior sampling events,
m is the number of marked individuals captured for the current sampling event,
and u is the number of unmarked individuals captured for the current sampling
event. The relative standard error of the population with a probability of 0.95
for each estimate was also estimated using the following formula where s2 is the
2[Ns2 / Σk(nm)]
To determine if there were differences in male and female length among
sampling events within a site and between sites on common sampling events, a
Mann-Whitney U test was used (JMP IN® , 2001, SAS Institute Inc., Cary, NC).
Locations of every male and female captured were mapped, distance traveled
between captures calculated, and minimum convex polygons calculated
using ArcMap 9.1 (ESRI, Inc. 2006). To convert X-Y meter coordinates to
latitude and longitude coordinates of individual mussels, GPS coordinates
of the X-axis were loaded into ArcMap and the X-Y coordinates (in meters) of
all recaptured mussels were added to the data file. These data were edited to
align the X-Y coordinates of the GPS coordinates recorded for the permanent
To calculate horizontal movement (i.e., distance traveled between captures),
we used Arc GIS and the Hawth’s Tools (Beyer 2004) function of “convert locations
to paths”. To calculate individual mussel home ranges for individuals
captured 3 or more times, we created minimum convex polygons by calculating
224 Southeastern Naturalist Vol. 11, No. 2
the area of the polygon using the “create minimum convex polygons function” in
To determine if males and females for each sampling event had uniform or
non-uniform distribution patterns, we created matrices in which the location
of each male was compared to the location of each female. This allowed us
to determine the upstream and downstream distribution of males in relation
to females. In addition, females were coded as either gravid or non-gravid. A
likelihood ratio test for goodness of fit was conducted to determine the significance
of expected versus observed male and female spatial patterns (Sokal
and Rohlf 1995).
Population estimates, size, fecundity, and reproductive status
At SFSR1, 111 captures representing 83 individuals and 28 recaptures from
7 sampling events (Table 1) resulted in a population estimate of 166 ± 32(SE)
individuals. Both first time captures and recaptures were highest during the last
sampling event (September) and lowest during the second sampling event (May).
The female (n = 23) to male (n = 60) ratio at SFSR1 was 1.0:2.6.
At SR1, 194 captures represented by 163 individuals and 31 recaptures
resulted in a population estimate of 451 ± 43(SE) individuals (Table 1). Both
first time captures and recaptures was highest during the last sampling event
Table 1. Mark and recapture data for Villosa iris in the South Fork Spring River (SFSR1) and Spring
River (SR1) sites for each sampling event. n = number of individuals previously marked for entire
study period, m = number of individuals captured during a sampling event that were previously
captured and marked, u = number of individuals captured during sampling event that were not previously
captured and marked, N = population estimate, s2 = variance, and CI = confidence interval.
Date n m u m + u n2 (m + u) nm m2/( m + u) N s2 95% CI
May-1 0 0 9 9 0 0 0.00000
May-2 9 1 7 8 648 9 0.12500
May-3 16 7 18 25 6400 112 1.96000
June 34 3 7 10 11,560 102 0.90000
July 41 2 17 19 31,939 82 0.21053
Aug. 58 5 5 10 33,640 290 2.50000
Sept. 63 10 20 30 119,070 630 3.33333
Summary 83 28 83 111 203,257 1225 9.02886 166 0.274328 32
July 07 0 0 26 26 0 0 0.00000
May-1 26 0 2 2 1352 0 0.00000
May-2 28 1 24 25 19,600 28 0.04000
June 52 5 21 26 70,304 260 0.96154
July 73 2 24 26 138,554 146 0.15385
Aug. 96 8 32 40 368,640 768 1.60000
Sept. 129 15 34 49 815,409 1935 4.59184
Summary 163 31 163 194 1,413,859 3137 7.34722 451 0.064500 43
2012 A.M. Asher and A.D. Christian 225
(September) and lowest during the second sampling event (May). The female (n =
70) to male (n = 9 3) ratio at SR1 was 1.0:1.3.
Male Rainbow Shell lengths at SFSR1 ranged from 35.9 to 59.1 mm
throughout the sampling period with sampling event means ranging from
42.6 ± 6.7 (SE) mm to 48.7 ± 1.7 (SE) mm (Fig. 2, Table 2). Female lengths
Figure 2. Length - frequency (number of individuals) distribution of male (black) and
female (white) Villosa iris individuals per size class at the South Fork Spring River site
(SFSR1) from May to September 2007.
226 Southeastern Naturalist Vol. 11, No. 2
at SFSR1 ranged from 29.4 to 54.6 mm throughout the sampling period, with
individual sampling event means ranging from 39.4 ± 9.14 (SE) mm to 42.9 ±
3.6 (SE) mm (Fig. 2, Table 2). For SR1, mean male length ranged from 23.2
mm to 62.3 mm throughout the sampling period, and individual sampling
event means ranged from 44.8 ± 3.8 (SE) mm to 50.3 ± 4.3 (SE) mm (Fig. 3,
Table 2). Mean female length ranged from 23.5 to 54.9 mm throughout the
sampling period, and individual sampling event means ranged from 39.8 ± 2.7
(SE) to 43.4 ± 2.3 (SE) mm (Fig. 3, Table 2).
SFSR1 males were slightly significantly larger than SR1 males only during
the May 2007 sampling period (S1: U = -1.97, P < 0.049). There were no signifi-
cant differences in male lengths at all other sampling events nor were there any
significant differences in female lengths between sites at any sampling event (S1:
P > 0.05).
Across all sampling events, 10 of 23 females at SFSR1 were gravid. The
percentage of gravid females at SFSR1 declined from 77.8% in May to 0% in
June, then gradually increased to 80% throughout the rest of the sampling period
(Table 3). During the 2007 sampling period, 30 of 54 females at SR1 were gravid,
and patterns of gravidity followed that of SFSR1, with the percentage of gravid females
declining after May and slowly increasing through September (Table 3).
Fecundity at SFSR1 (n = 6) ranged from 11,687 to 52,265 glochidia per female
with a mean of 27,849 ± 11,653(SE). Fecundity at SR1 ranged from 3080
to 34,434 glochidia with a mean of 15,089 ± 11,710 (SE).
Movement and spatial patterns
Sample size was based on recaptures, and for some sampling periods was
not large enough to perform statistical analysis. However, we did observe 3
Table 2. Mean lengths (mm) ± standard errors of South Fork Spring River site (SFSR1) and Spring
River site (SR1) Villosa iris males and females for each sampling event from May to September
2007 (including the May 2006 sample event at SR).
May 2006 May 2007 June July August September
SFSR1 - 48.7 ± 4.4 42.6 ± 6.9 48.2 ± 3.2 46.4 ± 2.1 48.6 ± 1.9
SR1 50.3 ± 4.3 44.8 ± 3.8 44.8 ± 4.5 45.3 ± 4.0 47.1 ± 2.1 47.0 ± 2.2
SFSR1 - 42.9 ± 3.6 39.4 ± 9.1 40.7 ± 3.6 39.8 ± 5.5 42.3 ± 2.2
SR1 43.4 ± 2.3 40.3 ± 2.8 41.1 ± 3 42.0 ± 6.8 42.2 ± 3.7 39.8 ± 2.7
Table 3. The percentage and sample size (n) of Villosa iris females gravid for each sampling period
(May–September) at the South Fork Spring River (SFSR1) and Spring River (SR1) sites from May
to September 2007.
May June July August September
SFSR1 77.8 (9) 0.0 (5) 25.0 (4) 50.0 (2) 80.0 (5)
SR1 47.1 (17) 21.4 (14) 50.0 (6) 90.0 (10) 88.2 (17)
2012 A.M. Asher and A.D. Christian 227
interesting patterns (Table 4). First, overall movement for males and females
combined for the entire sampling period was greater at SR1 (1.89 ± 0.58 cm/
day) than at SFSR1 (1.64 ± 0.53 cm/day). Second, overall female movement was
greater than male movement at both sites. Third, male movement over the entire
study period was greater at SR1, while female movement over the entire study
period was slightly greater at SFSR1.
Figure 3. Length-frequency (number of individuals) distribution of male (black) and female
(white) Villosa iris individuals per size class at the Spring River site (SR1) in May
2006 and from May to September 2007.
228 Southeastern Naturalist Vol. 11, No. 2
For those individuals with at least 3 observations (a capture and at least 2
recaptures), movement during the sampling period resulted in an overall mean
home range of 35.1 ± 22.6 cm for both sites combined. The home range of Rainbow
Shell was larger at SR1 (43.0 ± 42.5 cm; n = 3) than at SFSR1 (29.3 ± 27.7
cm; n = 4).
Spatial distributions of males and females at SFSR1 were significantly
different from a uniform distribution for the May, July, August, and September
sampling periods (Table 5, Fig. 4). In May, more males were observed
upstream of all females (G1 = 5.8, P < 0.016) and gravid females (G3 = 18.58,
P < 0.001) than expected from a uniform distribution. During July (G3 =
22.57, P < 0.001) and August (G3 = 13.18, P < 0.004), more males were
observed upstream of non-gravid females than expected from a uniform distribution.
During September, more males were observed downstream of all
females (G1 = 23.19, P < 0.001) and gravid females (G3 = 81.79, P < 0.001)
than expected from a uniform distribution.
Spatial distributions of males and females at SR1 were significantly different
than a uniform distribution for the June, August, and September sampling dates
(Table 6, Fig. 5). In June, significantly more males were observed upstream
of non-gravid females (G3 = 60.08, P < 0.001) than expected from a uniform
Table 4. Sample size of recaptures (n) and mean movement in cm/day (x̅) and standard error (SE)
for recaptured male (M) and female (F) Villosa iris from the South Fork Spring River (SFSR1) and
Spring River (SR1) sites. Recaptured individuals were not encountered during all sampling events
and values are for each sampling period between May to September 2007. O = overall.
May– May– May– May– June– June– June– July– July– Aug.–
June July Aug. Sept. July Aug. Sept. August Sept. Sept. O
M n 3 2 4 5 - - 2 3 3 1 23
x̅ 0.26 3.35 2.17 1.44 - - 3.24 0.61 1.08 0.72 1.55
SE 0.25 1.27 1.06 0.53 - - 3.04 1.19 0.41 - 0.52
F n - - 1 - - - 1 - - - 2
x̅ - - 1.11 - - - 4.22 - - - 2.67
SE - - - - - - - - - - 3.05
M+F n 3 2 5 5 - - 3 3 3 1 25
x̅ 0.26 3.35 1.96 1.44 - - 3.57 0.61 1.08 0.72 1.64
SE 0.25 1.27 0.92 0.53 - - 1.87 1.19 0.41 - 0.53
M n 1 - 2 2 1 1 3 2 2 6 20
x̅ 5.66 - 1.46 1.12 2.39 0 1.00 0.07 1.19 2.33 1.63
SE - - 2.28 1.68 - - 0.99 0.13 1.75 1.47 0.74
F n 3 1 - 1 - 3 1 - 1 - 10
x̅ 2.98 0.56 - 2.46 - 3.08 1.26 - 1.48 - 2.39
SE 2.00 - - - - 1.89 - - - - 0.92
M+F n 4 1 2 3 1 4 4 2 3 6 30
x̅ 3.65 0.56 1.46 1.57 2.39 2.31 1.06 0.07 1.29 2.33 1.89
SE 1.93 - 2.28 1.31 - 2.02 0.71 0.13 1.03 1.47 0.58
2012 A.M. Asher and A.D. Christian 229
distribution. Interestingly, significantly more males were observed downstream
of all females (G3 = 155.43, P < 0.001) and gravid females (G3 = 377.16, P less than
0.001) in August. However, the pattern changed in September, with significantly
Table 5. The number of observed and expected (based on uniform distribution) observations of
each individual Villosa iris male location compared to each individual female location with the
calculated G and P-values (bold indicates significance at P ≤ 0.05) for each comparison for each
month of the study period at the South Fork Spring River site (SFSR1). For the strictly male versus
female comparison: MUF = males upstream of females; MDF = males downstream of females.
For the males upstream or downstream of gravid females and non-gravid females: MUGF = males
upstream of gravid females, MDGF = males downstream of gravid females, MUNGF = males upstream
of non-gravid females, and MDNGF = males downstream of non-gravid females. NA = no
gravid females observed for June.
Month Treatment Observed Expected G P-value
MUF 82 68.00
MDF 54 68.00 5.80614 0.0160
MUGF 44 25.50
MDGF 24 25.50
MUNGF 14 25.50
MDNGF 20 25.50 18.58776 0.0003
MUF 16 12.50
MDF 9 12.50 1.98645 0.1587
MUGF NA NA
MDGF NA NA
MUNGF NA NA
MDNGF NA NA
MUF 29 30.00
MDF 31 30.00 0.06668 0.7962
MUGF 3 15.00
MDGF 12 15.00
MUNGF 26 15.00
MDNGF 19 15.00 22.57311 0.0000
MUF 10 8.00
MDF 6 8.00 1.01069 0.3147
MUGF 2 4.00
MDGF 6 4.00
MUNGF 8 4.00
MDNGF 0 4.00 13.18334 0.0043
MUF 36 62.50
MDF 89 62.50 23.19900 0.0000
MUGF 22 31.25
MDGF 78 31.25
MUNGF 14 31.25
MDNGF 11 31.25 81.79490 0.0000
230 Southeastern Naturalist Vol. 11, No. 2
more males observed upstream of all females (G1 = 11.22, P < 0.001) and gravid
females (G3 = 388.76, P < 0.001).
Population estimates and characteristics
Our population estimates for both sites were higher than previous reports at
both SFSR1 and SR1. In a study just 1 year prior to this study, Martin (2008) used
a stratified random sampling protocol and estimated the population size at SFSR1
to be 16 ± 14 Rainbow Shell individuals. This estimate is considerably lower than
our estimate of 166 ± 32 individuals. At the same time, our population estimate
at SR1 (451 ± 43) was over 100 individuals higher than Trauth et al. (2007),
which estimated the Rainbow Shell population there to be 273 ± 109 individuals
just 2 years prior to our sampling. Both studies (Martin 2008, Trauth et al. 2007)
used the same stratified random sampling design of a defined assemblage that
estimates populations using equations described by Christian and Harris (2005).
This protocol is ideal for large-scale surveys that sample all species encountered;
however, the stratified random sampling method tends to both over and under
Figure 4. Observed and expected occurrences of Villosa iris male locations compared
to female locations for monthly sampling period at the South Fork Spring River site
(SFSR1). MDNGF = males downstream of non-gravid females, MUNGF = males upstream
of non-gravid females, MDGF = males downstream of gravid females, MUGF
= males upstream of gravid females, MDF = males downstream of females, and MUF =
males upstream of females.
2012 A.M. Asher and A.D. Christian 231
estimate population size under certain conditions. For example, because the
Christian and Harris (2005) method only samples the assemblage once during
a year, it does not account for seasonal vertical and horizontal movements or
Table 6. The number of observed and expected (based on uniform distribution) observations of each
individual Villosa iris male location compared to each individual female location with the calculated G
and P-values (bold indicates significance at P ≤ 0.05) for each comparison for each month of the study
period at the Spring River site (SR1). For the strictly male versus female comparison: MUF = males
upstream of females; MDF = males downstream of females. For the males upstream or downstream of
gravid females and non-gravid females: MUGF = males upstream of gravid females, MDGF = males
downstream of gravid females, MUNGF = males upstream of non-gravid females, and MDNGF =
males downstream of non-gravid females.
Month Treatment Observed Expected G P-value
MUF 73 68
MDF 63 68 0.7359 0.3910
MUGF 41 34
MDGF 23 34
MUNGF 32 34
MDNGF 40 34 6.4930 0.0108
MUF 86 84
MDF 82 84 0.0952 0.7576
MUGF 15 42
MDGF 21 42
MUNGF 71 42
MDNGF 61 42 60.0816 0.0000
MUF 60 60
MDF 60 60 0.0000 0.0000
MUGF 22 30
MDGF 38 30
MUNGF 38 30
MDNGF 22 30 8.7375 0.0033
MUF 47 150
MDF 253 150 155.4278 0.0000
MUGF 44 75
MDGF 226 75
MUNGF 3 75
MDNGF 27 75 377.1645 0.0000
MUF 311 272
MDF 233 272 11.2225 0.0008
MUGF 290 136
MDGF 190 136
MUNGF 21 136
MDNGF 43 136 388.7647 0.0000
232 Southeastern Naturalist Vol. 11, No. 2
underrepresented species like mark and recapture techniques have the ability to
do. Nevertheless, we believe our population estimates are also an underestimate
of the true population because we were not able to capture all of the individuals
in the population. For example, Rogers et al. (2001) concluded that the methods
of Schumacher and Eschmeyer (1943) underestimate the true population size
when only surfaced individuals are sampled. Additionally, Berg et al. (2008) recognized
true abundances of smaller individuals are often underestimated due to
sampling techniques. Freshwater mussels remain burrowed through the juvenile
stage, often not surfacing until reproductively mature (Balfour and Smock 1995,
Yeager et al. 1994). Since only surfaced individuals were sampled, juveniles
were underrepresented in our study. This bias can be seen in the size-frequency
graphs of each site, in which none of the 246 individuals were less than 20 mm
The sex ratio in our study was skewed toward males in both populations.
While SR1 had a slight male bias of 1.0:1.3, the SFSR1 population had 2.6 males
per female. Sex ratios can be a response to overcome limitations of reproductive
mechanisms (van Ekrom Schurink and Griffiths 1991) or an indicator of a
Figure 5. Observed and expected occurrences of Villosa iris male locations compared to female
locations for monthly sampling period at the Spring River site (SR1). MDNGF = males
downstream of non-gravid females, MUNGF= males upstream of non-gravid females,
MDGF = males downstream of gravid females, MUGF = males upstream of gravid
females, MDF = males downstream of females, and MUF = males upstream of females.
2012 A.M. Asher and A.D. Christian 233
changing population (Heard 1975). Male-skewed sex ratios have been reported
for Quadrula asperata (1.0:3.8) and Lampsilis ornata (1.0:5.5) in the Sipsey
River, AL (Haag and Staton 2003). However, male-biased sex ratios are not observed
in all freshwater mussel populations (Berg et al. 2008, Haag and Staton
2003, Rogers et al. 2001, Yeager and Neves 1986), and some populations exhibit
female-skewed sex ratios (Garner et al. 1999). Downing et al. (1989) suggested
that a male-skewed sex ratio emphasizes the importance of female fertilization.
However, current evolutionary theory for dioecious species is that sex ratios at
conception commonly should be 1:1 (Charnov 1982). One class of factors that
have been shown to influence sex ratios post conception is environmental factors.
Martin (2008) found that SFSR1 had a low index of biotic integrity (IBI) score
(i.e., impaired), but higher habitat assessment score (marginal) and microinvertebrate
index score (i.e., very good). The lower IBI score, indicative of altered fish
composition compared to a less altered site, could suggest reproductive barriers
for freshwater mussels either due to poor water quality, as indicated by low IBI
scores, or lack of host fish or host-fish migration.
Fecundity is a measure of reproductive potential and is linked with female
size, as larger females generally have higher fecundity (Downing et al. 1993,
Haag and Staton 2003), although exceptions have been reported (Perles et al.
2003). Fecundity was much lower at SR1 than SFSR1, even though we did not
detect any differences in female size between the two sites. Villosa iris is considered
bradytictic (Watters et al. 2001), releasing glochidia in late spring to early
summer (Parmalee and Bogan 1998). Because glochidia were collected from
SR1 females in July, compared to May collection at SFSR1, seasonal differences
in fecundity may be the reason for the differences we observed. For example,
it is likely the females at SR1 had already discharged glochidia and were only
partially recharged from the current spawning period. Thus, we believe that fecundity
estimates from SFSR1 are likely to be more representative measurement
of pre-glochidia release fecundity.
Both populations had size distributions similar to those often reported in
the literature showing a conspicuous absence of individuals under 20 mm
(Christian et al. 2005, Payne and Miler 1989). When lengths of females and
males were compared between rivers, no differences were observed except
during May 2007, in which SFSR1 males were longer than SR1. At first
glance, both populations would appear to lack recruitment. However, as previously
mentioned, juvenile mussels burrow into the substrate (Balfour and
Smock 1995, Yeager et al. 1994). Sediments were not extensively sampled;
therefore, we believe that juveniles were not effectively sampled. Thus, population
size structure and recruitment are likely underestimated and reflective
of individuals over 20 mm in length.
Our observed horizontal movement rates were higher than movement rates
of an Elliptio complanata (Lightfoot) (Eastern Elliptio) assemblage at Buzzards
234 Southeastern Naturalist Vol. 11, No. 2
Branch, VA, a headwater stream consisting of primarily sandy substrate with silt
and gravel (Balfour and Smock 1995). However, our movement rates are much
lower than those of Peck et al. (2007), in which native and relocated Potamilus
capax (Green) (Fat Pocketbook) had combined displacement values ranging up
to ≈225 cm/day (27 m over a 3-month period). Displacement values, measures
of distance between first and last known location, are not reflective of the overall
movement and would be equal to or less than actual movement distances. Horizontal
movement rates in this study were more comparable to those observed by
Schwalb and Pusch (2007) in the River Spree, Germany, a stream with sandy
substrate. Schwalb and Pusch (2007) observed mean movement of 1.9 cm/day
for Unio pictorum (L.) (Painter’s Mussel), 1.4 cm/day for U. tumidus (Philipsson)
(Swollen River Mussel), and 2.1 cm/day for Anodonta anatina (L.) (Duck
Mussel). Neither study determined the exact position of an individual, but instead
used the relative movement between quadrats. Our Rainbow Shell movement
rates also were close to those of E. complanata in Lac de l’Achigan (Amyot and
Downing 1997). Although movement rates were similar among the unionids, the
habitats were contrasting, as Lac de l’Achigan is a lentic system. Numerous factors,
such as substrate type, stream order, discharge (Schwalb and Pusch 2007),
temperature (Perles et al. 2003, Schwalb and Pusch 2007), day length (Amyot and
Downing 1997, Perles et al. 2003, Schwalb and Pusch 2007), spawning period
(Amyot and Downing 1997, 1998), and density and position of other individuals
(Downing and Downing 1992, Huang et al. 2007) have been shown to influence
movements of unionids and could have influenced the horizontal movement of
the Rainbow Shell in our study.
Since females brood glochidia, we hypothesized females would have lower
movement rates than males, allocating energy toward brooding rather than movement
(Amyot and Downing 1998). Observations in our study were the same as
Amyot and Downing (1998), with females having higher mean movement rates
than males. However, due to low sample size per sampling event, statistical tests
were not feasible. Comparisons of male and female unionid movement rates have
been made for lentic systems; however, they have not been investigated for lotic
systems. Amyot and Downing (1998) reported the mean distance travelled by
females in the lentic study area was slightly greater than that of males, but differences
in mean distances were not significant.
In our study, the male and female distribution patterns were associated with
the spawning period of Rainbow Shell. In May, when locations of surfaced males
were compared to surfaced gravid and non-gravid females, more males were
located upstream of gravid females than compared to a uniform distribution. In
July and August, more males were located upstream of non-gravid females than
expected from a uniform distribution (but not at SR1 for August). This finding
is consistent with the observation that female Rainbow Shell released glochidia
in spring and early summer, resulting in a spawning period during summer to
2012 A.M. Asher and A.D. Christian 235
early fall (Parmalee and Bogan 1998). It was also consistent with our findings
in which the percentage of gravid females peaked in May, was lowest in June,
and increased in each subsequent month until 80% of the females were observed
gravid in September. Male and female spatial distributions at SR1 were similar
to that of SFSR1, with 78.6% of the females non-gravid in June. Our observation
of more males upstream (see comment about SR1) of non-gravid females
(July and August) at both sites corresponds to the spawning period and might be
interpreted as a behavior to increase fertilization success (Downing et al. 1989).
Meanwhile, our observation of more males downstream of females in September
corresponded with a cessation of spawning and spatial position being less
relevant. Thus, the observation that more males are located upstream of females
supports the hypothesis that during spawning events, more males would be located
upstream of females than other times of the year.
Information on the population characteristics and spatial patterns of freshwater
mussels has been generally lacking. Our study resulted in 2 findings.
First, population estimates at both study sites were larger than previously reported,
and we conclude that mark and recapture methods tend to provide more
accurate and precise estimates than quadrat-based sampling. Quadrat-based
sampling surveys a portion of the mussel assemblage during 1 sampling event,
whereas mark and recapture surveys the entire assemblage during multiple
sampling events. Second, based on a male-skewed sex ratio, movement, and
spatial patterns observed in our study, we conclude that Rainbow Shell exhibits
behaviors that may lead to increased fertilization rates of females. For example,
although mussels are considered sedentary animals with small home ranges,
movement potentially increases the chance for fertilization and may aid mussels
in avoiding harmful conditions or locating more suitable microhabitat.
Furthermore, males and females exhibited a non-uniform distribution, at least
in association with spawning events, with males being spatially located upstream
of females. Thus, due to external fertilization of freshwater mussels, this
spatial positioning of males should increase fertilization rates and ultimately
Funding for this project was provided by the Arkansas State University Environmental
Sciences Graduate program and the Arkansas Biosciences Institute at ASU. We thank
M.N. Asher, E.A. Daniells, K. Inoue, D.M. Hayes, R.L. Lawson, H.C. Martin, and A.J.
Peck for field and lab assistance. We thank N. Young and A.J. Peck for assistance with
ArcGIS. We thank J. Harris for early comments on this manuscript and anonymous reviewers
for comments during the review process.
Amyot, J.P., and J.A. Downing. 1997. Seasonal variation in vertical and horizontal movement
of the freshwater bivalve Elliptio complanata (Mollusca: Unionidae). Freshwater
236 Southeastern Naturalist Vol. 11, No. 2
Amyot, J.P., and J.A. Downing. 1998. Locomotion in Elliptio complanata (Mollusca:
Unionidae): A reproductive function? Freshwater Biology 39:351–358.
Balfour, D.L., and L.A. Smock. 1995. Distribution, age structure, and movement of the
freshwater mussel Elliptio complanata (Mollusca: Unionidae) in a headwater stream.
Journal of Freshwater Ecology 10:255–268.
Berg, D.J., T.D. Levine, J.A. Stoeckel, and B.K. Lang. 2008. A conceptual model linking
demography and population genetics of freshwater mussels. Journal of the North
American Benthological Society of America 27:395–408.
Beyer, H.L. 2004. Hawth’s Analysis Tools for ArcGIS. Available online at http://www.
spatialecology.com/htools. Accessed July 2007.
Bogan, A.E., and K.J. Roe. 2008. Freshwater bivalve (Unioniformes) diversity, systematics,
and evolution: Status and future directions. Journal of the North American
Benthological Society 27:395–408.
Charnov, E.L. 1982. The Theory of Sex Allocation. Princeton Univerity Press, Princeton,
NJ. 355 pages.
Christian, A.D., and J.L. Harris. 2005. Development and assessment of a sampling design
for mussel assemblages in large stream. American Midland Naturalist 153:284–292.
Christian, A.D., J.L. Harris, W.R. Posey, J.F. Hockmuth, and G.L. Harp. 2005. Freshwater
mussel (Bivalvia : Unionidae) assemblages of the Lower Cache River, Arkansas.
Southeastern Naturalist 4:487–512.
Di Maio, J., and L.D. Corkum. 1995. Relationship between the spatial distribution of
freshwater mussels (Bivalvia: Unionidae) and the hydrological variability of rivers.
Canadian Journal of Zoology 15:663–671.
Downing, J.A., and W.L. Downing. 1992. Spatial aggregation, precision, and power in
surveys of freshwater mussel populations. Canadian Journal of Fisheries and Aquatic
Downing, J.A., J.P. Amyot, M. Perusse, and Y. Rochon. 1989. Visceral sex, hermaphroditism,
and protandry in a population of the freshwater bivalve Elliptio complanata.
Journal of the North American Benthological Society 8:92–99.
Downing, J.A., Y. Rochon Perusse, and H. Harvey. 1993. Spatial aggregation, body size,
and reproductive success in the freshwater mussel Elliptio complanata. Journal of the
North American Benthological Society 12:148–156.
ESRI, Inc. 2006. ArcGIS Version 9.2. Redlands, CA.
Garner, J.T., T.M. Haggerty, and R.F. Modlin. 1999. Reproductive cycle of Quadrula metanevra
(Bivalvia: Unionidae) in the Pickwick Dam tailwater of the Tennessee River.
America Midland Naturalist 141:277–283.
Haag, W.R., and J.L. Staton. 2003. Variation in fecundity and other reproductive traits in
freshwater mussels. Freshwater Biology 48:2118–2130.
Hanlon, S.D., and J.F. Levine. 2004. Notes on the life history and demographics of the
Savannah Lilliput (Toxolasma pullus) (Bivalvia: Unionidae) in University Lake, NC.
Southeastern Naturalist 3:289–296.
Harris, J.L., W.R. Posey II, C.L. Davidson, J.L. Farris, S.R. Oetker, J.N. Stoeckel, B.G.
Crump, M.S. Barnett, H.C. Martin, M.W. Matthews, J.H. Seagraves, N.J. Wentz, R.
Winterringer, C. Osborne, and A.D. Christian. 2009. Unionoida (Mollusca: Margaritiferidae,
Unionidae) in Arkansas, third status review. Journal of the Arkansas Academy
of Science 63:50–86.
Heard, W.H. 1975. Sexuality and other aspects of reproduction of Anodonta (Pelecypoda:
Unionidae). Malacologia 15:81–103.
2012 A.M. Asher and A.D. Christian 237
Huang, D., P.A. Todd, and J.R. Guest. 2007. Movement and aggregation in the Fluted
Giant Clam (Tridacna squamosa L.). Journal of Experimental Marine Biology and
Kuehnl, K.F. 2009. Exploring levels of genetic variation in the freshwater mussel genus
Villosa (Bivalvia: Unionidae) at different spatial and systematic scales: Implications
for biogeography, taxonomy, and conservation. Ph.D. Dissertation. Ohio State University,
Columbus, OH. 261 pp.
Lydeard, C., R.H. Cowie, W.F. Ponder, A.E. Bogan, P. Bouchet, S.A. Clark, K.S. Cummings,
T.J. Frest, O. Gargominy, D.G. Herbert, R. Hershler, K.E. Perez, B. Roth, M.
Seddon, E. E. Strong, and F.G. Thompson. 2004. The global decline of nonmarine
mollusks. Bioscience 54:321–330.
Martin, H.C. 2008. Physical and biological assessment of the South Fork of the Spring
River, Arkansas. M.Sc. Thesis. Arkansas State University, Jonesboro, AR.
Martin, H.C., J.L. Harris, and A.D. Christian. 2009. A qualitative freshwater mussel survey
of the South Fork Spring River, Missouri and Arkansas. Journal of the Arkansas
Academy of Science 63:106–112.
National Native Mussel Conservation Committee. 1998. National strategy for the conservation
of native freshwater mussels. Journal of Shellfish Research. 17:1419–1428.
Parmalee, P., and A. Bogan. 1998. The Freshwater Mussels of Tennessee. The University
of Tennessee Press, Knoxville, TN.
Payne, B.S., and A.C. Miller. 1989. Growth and survival of recent recruits to a population
of Fusconaia ebena (Bivalvia: Unionidae) in the lower Ohio River. American
Midland Naturalist 121:99–104.
Peck, A.J., J.L. Harris, J.L. Farris, and A.D. Christian. 2007. Assessment of freshwater
mussel relocation as a conservation strategy. Pp. 115–124, In C. Leroy Irwin, Debra
Nelson, and K.P. McDermott (Eds.). Proceedings of the 2007 International Conference
on Ecology and Transportation. Center for Transportation and the Environment,
North Carolina State University, Raleigh, NC.
Perles, S.J., A.D. Christian, and D.J. Berg. 2003. Vertical migration, orientation, aggregation,
and fecundity of the freshwater mussel Lampsilis siliquoidea. Ohio Journal of
Rogers, S.O., B.T. Watson, R.J. Neves. 2001. Life history and population biology of the
endangered Tan Riffleshell (Epioblasma florentina walkeri) (Bivalvia: Unionidae).
Journal of the North American Benthological Society 20:582–594.
Schumacher, F.X., and R.W. Eschmeyer. 1943. The recapture and distribution of tagged
bass in Norris Reservoir, Tennessee. Journal of the Tennessee Academy of Science
Schwalb, A.N., and M.Y. Pusch. 2007. Horizontal and vertical movement of unionid
mussels in a lowland river. Journal of North American Benthological Society
Sokal, R.R., and F.J. Rohlf. 1995. Biometry: The Principals and Practice of Statistics in
Biological Research. W.H. Freeman and Company, New York, NY.
Trauth, S.E., B.A. Wheeler, W.R. Hiler, R.L. Lawson, H.C. Martin, and A.D. Christian.
2007. Current distribution and relative abundance of the crayfish, mussels, and aquatic
salamanders of the Spring River, AR. Final Report to the Arkansas Game and Fish
Commission, Little Rock, AR. Department of Biological Sciences, Arkansas State
University, Jonesboro, AR.
238 Southeastern Naturalist Vol. 11, No. 2
van Erkom Schurink, C., and C.L. Griffiths. 1991. A comparison of reproductive cycles
and reproductive output in four southern African mussel species. Marine Ecology
Progress Series 76:123–134.
Watters, G.T., S.H. O’Dee, and S.W. Chordas III. 2001. Patterns of vertical migration
in freshwater mussels (Bivalvia: Unionidae). Journal of Freshwater Ecology
Williams, J.D., M.L. Warren, K.S. Cummings, J.L. Harris, and R.J. Neves. 1993. Conservation
status of freshwater mussels of the United States and Canada. Fisheries
Williams, J.D., A.E. Bogan, and J.T. Garner. 2008. Freshwater Mussels of Alabama and
the Mobile Basin in Georgia, Mississippi, and Tennessee. University of Alabama
Press, Tuscaloosa, AL. 908 pp.
Yeager, B.L. and R.J. Neves. 1986. Reproductive cycle and fish hosts of the Rabbit’s Foot
Mussel, Quadrula cylindrica strigillata (Mollusca: Unionidae) in the upper Tennessee
River drainage. American Midland Naturalist 116:329–340.
Yeager, M.M., D S. Cherry and R.J. Neves. 1994. Feeding and burrowing behavior of
juvenile Rainbow Mussels, Villosa iris (Bivalvia: Unionidae). Journal of the North
American Benthological Society 13:217–222.