Content uploaded by Will Selman
Author content
All content in this area was uploaded by Will Selman on Jan 16, 2025
Content may be subject to copyright.
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
448
SOUTHEASTERN NATURALIST
2024 23(X):448–473
Distribution and Abundance of Graptemys Species
(G. nigrinoda and G. pulchra) in the Tombigbee River System
of Northeastern Mississippi
Will Selman*
Abstract - The river turtles Graptemys nigrinoda (Black-knobbed Sawback) and G. pulchra
(Alabama Map Turtle) are endemic to the Mobile River drainage of Alabama, northwest-
ern Georgia, and northeastern Mississippi. Little is known of their current distribution
and abundance in the Tombigbee River (TR) system in northeastern Mississippi even
though both are being considered for protection under the US Endangered Species Act.
During May–July 2019, 2022, and 2023, I used a combination of methods (e.g., point-
count surveys, basking-density surveys, trapping, mark–resight population estimates) to
determine the presence and densities of both species in the TR drainage of northeastern
Mississippi. Both Graptemys species were detected throughout most of the TR and Ten-
nessee–Tombigbee Waterway (TTW). Black-knobbed Sawbacks and Alabama Map Turtles
were both found in 5 previously undocumented creeks, and both species were found near or
upstream of almost all historic localities. Basking densities at 33 sites were 15× greater for
Black-knobbed Sawbacks (15.0 individuals/rkm) compared to Alabama Map Turtles (1.0
individuals/ rkm), while different habitat types had different densities of Black-knobbed
Sawbacks and Alabama Map Turtles; both species occurred at their lowest densities in the
TTW lakes. Trapping at 5 sites yielded 107 Black-knobbed Sawbacks and 14 Alabama Map
Turtles, and mean mark–resight population estimates were 173 individuals/rkm (min–max:
82–397 individuals/rkm) at 5 sites for Black-knobbed Sawbacks and 13 individuals/rkm
(4–20 individuals/rkm) at 3 sites for Alabama Map Turtles. Alabama Map Turtles are of
greater conservation concern in Mississippi compared to Black-knobbed Sawbacks due to
lower densities and apparent extirpation from parts of the drainage.
Introduction
Two endemic Graptemys species occur sympatrically in the Mobile River sys-
tem of Alabama and northeastern Mississippi: G. nigrinoda Cagle (Black-knobbed
Sawback) and G. pulchra Baur (Alabama Map Turtle). The Black-knobbed Saw-
back is considered imperiled (S2) in Mississippi (Mississippi Museum of Natural
Science 2015) and considered vulnerable (G3) throughout its range (NatureServe
Explorer 2024a); it is currently listed with a status of least concern by the Interna-
tional Union for the Conservation of Nature (IUCN; van Dijk 2011a). The Alabama
Map Turtle is similarly considered imperiled (S2) in Mississippi (Mississippi Mu-
seum of Natural Science 2015) and considered apparently secure (G4) throughout
its range (NatureServe Explorer 2024b); it is currently considered near threatened
by the IUCN (van Dijk 2011b). Further, both Graptemys species were petitioned
for listing as threatened under the Endangered Species Act in 2011 (US Fish and
*Biology Department, Millsaps College, 1701 N. State Street, Jackson, MS 39201;
will.selman@millsaps.edu.
Manuscript Editor: Christopher Lechowicz
Southeastern Naturalist
449
W.Selman
2024 Vol. 23, No. 4
Wildlife Service 2011).
Even though both species are considered species of greatest conservation need
(SGCN) in Mississippi, most of the ecological information and population data
available for both species is from Alabama. In the early 2000s, Godwin (2003) sur-
veyed both species throughout the Mobile River system in Alabama. He found that
Black-knobbed Sawbacks were the most common basking turtle observed below
the Fall Line in the Mobile River system, and densities of Black-knobbed Sawbacks
in the Tombigbee River (TR) were markedly lower than sites in the Alabama or Ca-
haba rivers. Similar to Black-knobbed Sawbacks, Godwin (2003) also found lower
densities of Alabama Map Turtles at TR sites compared to other rivers surveyed.
Further, Black-knobbed Sawbacks were found at ~6.5× higher densities than Ala-
bama Map Turtles throughout the Mobile River system, but Alabama Map Turtles
outnumbered Black-knobbed Sawbacks at sites above the Fall Line (Godwin 2003).
It is unknown why the Graptemys densities observed by Godwin (2003) were
lower in the TR compared to other rivers, but an explanation could be the major
hydrologic alterations that occurred to the TR system when the Tennessee–Tombig-
bee Waterway (TTW) was constructed in northeastern Mississippi. The TTW was
constructed from 1972 to 1985 to connect the Tennessee River system and the TR
system to facilitate a shipping connection between the Gulf of Mexico and the inte-
rior region of the US (Stine 1991). During and after construction of the TTW, direct
impacts occurred to the native sh (Taylor et al. 2008) and mussel communities,
with some mussel species becoming extirpated or extinct because of the TTW con-
struction (Williams et al. 2008). Along with direct impacts to aquatic organisms,
the TTW also made a direct path to connect 2 diverse and unique aquatic faunas.
Indeed, researchers have documented numerous sh (e.g., Etnier and Starnes 1993)
and mussels (Harteld and Jones 1989) that have exchanged between the Tennessee
and TR systems via the TTW. Berry et al. (2020) also documented turtle exchanges
among the drainages via the TTW; Alabama Map Turtles were found in the Ten-
nessee River system and Graptemys ouachitensis Cagle (Ouachita Map Turtle)—a
species native to the Mississippi and Tennessee rivers—was found in the upper TR.
To date, neither native Graptemys species of the TR has been surveyed exten-
sively in Mississippi. Early surveys recorded both species in the TR (Cagle 1954,
Cliburn 1971, McCoy and Vogt 1979, Shoop 1967), but these studies lacked some
modern survey methods (e.g., large-zoom digital cameras) and lacked extensive
sampling, particularly in the smaller tributaries. As a further measure of the knowl-
edge on both species, there are only 37 Black-knobbed Sawbacks and 8 Alabama
Map Turtle specimens from Mississippi, respectively, in the Mississippi Museum
of Natural Science (MMNS) herpetology collections. Records of both species in
Mississippi are also rare in other museum collections (Black-knobbed Sawbacks:
n = 5, Alabama Map Turtles: n = 1; www.vertnet.org, accessed 26 June 2022). As a
consequence, both of these species are the most poorly understood Graptemys spe-
cies in Mississippi, and little is known about their conservation status in the state.
Because there is a lack of records and observations from the Mississippi por-
tion of their ranges, the rankings discussed above may not be representative of
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
450
the species’ true conservation status. Therefore, the objective of this study is to
complete distribution and abundance sampling in the Mississippi-portion of the TR
drainage for Graptemys species and also further dene the conservation status of
both imperiled species in Mississippi.
Field-Site Description
I completed this research within the historical TR system of northeastern
Mississippi. In Mississippi, the drainage basin of the TR is ~1.58 million ha
and drains a portion 19 eastern and northeastern Mississippi counties. Major
tributaries of the TR in Mississippi include (organized from north to south): Big
Brown Creek, Twentymile Creek, Bull Mountain Creek, Town Creek, Mattubby
Creek, Buttahatchie River, Tibbee Creek, Luxapallila Creek, Noxubee River,
Sucarnoochee Creek, and Ponta Creek. These streams are primarily low-gradient
streams, and many contain a sandy and/or silty substrate. However, portions of
some streams on the east side of the drainage contain a gravel substrate (e.g., Bull
Mountain Creek, Buttahatchie River), while others contain a hard-bottom sub-
strate along some stream reaches (e.g., Mattubby Creek, Noxubee River). Small
point bars and sandbars exists along most stretches, but numerous kilometers of
some streams have been channelized and lack sandbars and meander bends. Many
of the latter drain heavy agricultural areas on the west side of the drainage. Along
with these streams, the TTW has 9 impoundments created by a series of locks
and dams. From north to south, these impoundments include Bay Springs Lake,
Montgomery Lake (aka, Pool E), Rankin Lake (Pool D), Fulton Lake (Pool C),
Wilkins Lake (Pool B), Amory Lake (Pool A), Aberdeen Lake, Columbus Lake,
and Aliceville Lake. The latter is primarily in Alabama, but a portion extends
northwesterly into southeastern Lowndes County, MS.
Methods
Rapid assessment surveys
I used 2 methods to determine the population distribution of Graptemys in
the TR system of northeastern Mississippi: point-count surveys of rivers/streams
crossed by roads or at river/stream access points and basking-density surveys by
boat or canoe.
For point-count surveys, I completed a total of 176 point counts at 157 bridge
crossings or river access points on the Tombigbee, Buttahatchie, Tibbee, and Nox-
ubee rivers as well as minor tributaries (Fig. 1, Table 1) during May–July in 2019,
2022, and 2023. At each survey locality, I identied basking turtles with a 20–60×
spotting scope (Vortex Intrepid HD; Barneveld, Wisconsin) with a tripod or 12×36
Canon Image-Stabilized Binoculars (Tokyo, Japan). I also photographed individu-
als for locality vouchers with a Nikon Coolpix p900 digital camera (Tokyo, Japan)
and deposited the images in the Florida Museum of Natural History (FMNH)
Herpetology Digital Archives. At each survey point, I identied turtles using a
combination of body size, carapace shape, head size, and head/carapace marking
patterns and recorded the number of each species observed.
Southeastern Naturalist
451
W.Selman
2024 Vol. 23, No. 4
Figure 1. Point-count (dots) and
basking-density (squares) survey
locations for Graptemys species
throughout the Tombigbee River
system in northeastern Mississippi,
2019–2022. The numbers inside the
squares correspond to the numbers
in Table 1.
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
452
For the second survey method, I conducted 108 basking-density surveys by
boat at 33 locations during May/June 2022 and 2023 (Fig. 1). Along with bridge
surveys, I used these survey sites to determine species presence, as well as their
densities. Sites were located on the TR, the TTW (i.e., channels where commer-
cial barge traffic occurs and channel modifications have occurred), tributaries,
lakes (i.e., still water formed by lock and dams along the TTW), and old channels.
The latter sites were old meander bends of the TR that were isolated following the
channelization of the river; some of these old channels had both outlets (upstream
and downstream) connected to the TTW, while others only had a single outlet
connected (downstream). I completed replicate surveys at most of these locations
to provide higher confidence in observed basking densities. During these surveys
and similar to methods described by Selman (2020), I located basking turtles
with 12×36 Canon Image-Stabilized Binoculars from the motorboat while idling
downstream; if extensive sandbars were present (sites 13, 27), I walked sandbars
and spotted turtles using a 20–60× spotting scope with tripod. Identification of
basking turtles was similar to point counts, and basking densities are reported in
number of turtles observed per river kilometer (rkm) or per shoreline kilometer
(skm) for lake sections.
Because count data were non-normally distributed, I used separate nonpara-
metric Kruskal–Wallis tests to determine if basking densities of Black-knobbed
Sawbacks and Alabama Map Turtles were equal across the 5 different aquatic habi-
tat types: TR, TTW, old channels, tributaries, and lakes (see Table 1 for sites). If
differences were observed, I used Wilcoxon nonparametric multiple comparisons
post hoc analyses to determine differences among habitat types.
Table 1. Summary results of point-count surveys for Graptemys and other turtle species in the Tom-
bigbee River system in northeastern Mississippi. This table is sorted alphabetically by primary drain-
age and also subdrainage. Abbreviations: R. = river, Cr. = creek, #S = number of surveys, # Loc =
number of unique locations surveyed, No turtles = no turtles observed, Gn = Graptemys nigrinoda,
Gpu = Graptemys pulchra, Am = Apalone mutica, As = Apalone spinifera, Cd = Chrysemys dorsalis,
Pc = Pseudemys concinna, Sp = Sternotherus peltifer, Ts = Trachemys scripta. [Table continued on
following page.]
Primary drainage/ No
subdrainage # S # Loc turtles Gn Gpu Am As Cd Pc Sp Ts
Bull Mountain Cr. 5 5 2 1 6 5
Gum Cr. 4 4 1 4
Sipsey Cr. 1 1 1
Buttahatchie R. 9 7 13 7 1 27 2
East Fork Tombigbee R. 5 5 19 1 7 2
Big Brown Cr. 2 2 5
Boguefala Cr. 6 6 3
Little Brown Cr. 3 3 1 2 2 2
Mackey’s Cr. 1 1 X
Luxapallila Cr. 5 4 5 1 2
Magby Cr. 2 2 X
Yellow 4 2 1 1 2 4
Southeastern Naturalist
453
W.Selman
2024 Vol. 23, No. 4
Trapping and mark–resight population estimates
In May/June 2022 and 2023, I sampled turtles at 1 site on the TR (site 13), 1
site on an old channel of the Tombigbee (site 22), and 3 sites on the TTW (sites 14,
19, 27) (Fig. 1); specic site names are not used to avoid potential collection for
the pet trade. I trapped turtles by shallowly submerging basking traps (made of ¾”
Table 1, continued.
Primary drainage/ No
subdrainage # S # Loc turtles Gn Gpu Am As Cd Pc Sp Ts
Mattubby 2 2 5 1 1 4 2
Noxubee R. 14 12 1 8 3 1 1 8
Cypress Cr. 1 1 4
Little Noxubee R. 1 1 X
Tibby Cr. 1 1 X
Wahalak Cr. 4 2 3 2
Wet Water Cr. 1 1 X
Sucarnoochee R. 22 13 5 9 1
Alamuchee Cr. 1 1 1
Blackwater Cr. 1 1 X
Pawtickfaw Cr. 1 1 1
Ponta Cr. 9 8 2 1 4
Running Tiger Cr. 1 1 X
Sucatolba Cr. 1 1 X
Tibbee Cr. 2 2 5 5 3
Catalpa Cr. 4 4 1 9
Chuquatonchee Cr. 3 3 X
Houlka Cr. 1 1 X
Line Cr. 2 2 2 2
Trim Cane Cr. 3 3 2
Town Cr. (Lowndes County) 3 3 1 1 5
TTW 7 7 13 2 5 11
Bay Springs Lake 12 12 1 42 27
Twentymile Cr. 4 4 9 9 11
Osborne Cr. 1 1 3
Wolf Cr. 2 2 1 3 1
West Fork Tombigbee R. 1 1 2 2
Bridge Cr. 1 1 7
Brown Cr. 1 1 X
Chiwapa Cr. 2 2 X
Coonewaha Cr. 3 3 1 2
Little Dry Cr. 1 1 X 3
Mubby Cr. 1 1 X
Mud Cr. 5 5 2 5 4 7
Tabinella Cr. 2 2 X
Town Cr. (Lee County) 5 5 1 1 1 14
Yonaba Cr. 3 3 1
Total 176 157 60 28 1 34 3 139 4 156
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
454
PVC coated crawsh wire and varying in size from 61×40×40 cm to 137×30×61
cm) from observed turtle-basking structures using similar methods described by
Selman et al. (2012). I used nails and cotton twine to attach traps to logs or branches
known to be Graptemys basking sites. I also opportunistically captured turtles by
dip net while turtles were basking on emergent deadwood or swimming near the
water surface.
Following capture, I permanently marked individuals on their marginal scutes
with drill holes using the method described by Cagle (1939) and injected passive
integrated transponders (PIT tags) into the inguinal region of the body cavity,
positioned anteriorly and parallel to the bridge of the shell (Buhlmann and Tuber-
ville 1998). I also paint-marked Graptemys on the 2nd and 3rd vertebral scutes with
non-toxic, tree-marking spray paint to aid in mark–resight population surveys. I
released all marked turtles at their point of capture after the paint-mark dried. For
catch-per-unit-effort (CPUE), I used trap days (i.e., number of days on the river
trapping) for the calculation denominator. Similar to Selman (2024), I took this
approach because the number of basking traps used to capture Graptemys varied
from 10 to 17 traps per day, different numbers of trap runs were made by boat for
each trap on different days and different locations, and traps occasionally had to be
moved due to basking turtles actively avoiding the trap.
I conducted visual mark–resight surveys at each site during optimal basking
times (0900 hr–1700 hr) within 2 weeks of the rst paint-marked individual be-
ing processed in order to meet the assumption of a closed population. Prior to
any mark–resight surveys, I attempted to have a minimum of 10–15 turtles paint-
marked in the sampled river stretch to improve the condence of each population
estimate. I was able to complete this prerequisite for Black-knobbed Sawbacks
at all sites and surveys, but this prerequisite could not be met for Alabama Map
Turtles at any sites or surveys due to the rarity of the species and inability to capture
more than 6 individuals at any site (4 individuals at site 13, three at site 14, one
at site 19, six at site 22, and none at site 27). Similar to basking-density surveys,
when sandbars were present (sites 13 and 17), I walked the banks/sandbars and
used a 20–60× spotting scope with tripod to locate unmarked, as well as marked
individuals along the outer cutbank and inner sandbars. At 3 sites where sandbars
were not present (sites 14, 19, and 22), I completed surveys by idling downstream
in a motorboat and spotted turtles with 12×36 Canon Image Stabilized Binoculars.
I also used a Nikon Coolpix p900 camera to photograph and later identify turtles,
especially those basking in large aggregations that are more prone to abandon their
basking locations from a longer distance. I identied and distinguished basking
turtles similar to point-count and basking-density surveys.
I generated all population estimates with the program NOREMARK (White
1996) using the joint hypergeometric maximum likelihood estimator (Bartmann
et al. 1987). I used this estimator because paint-marked turtles were not uniquely
marked and were resighted and not recaptured. NOREMARK also accounts for ad-
ditional marked individuals added to the population between survey intervals. The
assumptions for this test were as follows: each individual in the population has an
Southeastern Naturalist
455
W.Selman
2024 Vol. 23, No. 4
equal and random chance of being resighted, marking of one individual does not af-
fect another individual from being resighted, no paint marks are lost in the process
(via emigration and mortality, i.e, closed population), and marking does not affect
an individual’s chance of emigration or mortality.
Results
Rapid assessment surveys
A total of 484 turtles of 8 species were observed during 176 point-count surveys
at 157 locations in the Tombigbee River system of northeastern Mississippi (Fig. 1,
Table 1). The 2 dominant basking species observed were Trachemys scripta (Thun-
berg in Schoepff) (Slider Turtle; 156 individuals observed at 61 sites [32.2% of the
total turtles observed]) and Pseudemys concinna (Le Conte) (River Cooter; 139 in-
dividuals at 54 sites [28.7% of total observations]). Other species observed in order
of abundance include Black-knobbed Sawbacks (60 at 15 sites [12.4%]), Apalone
spinifera (Le Sueur) (Spiny Softshell; 34 at 22 sites [7.0%]), Alabama Map Turtles
(28 at 18 sites [5.7%]), Sternotherus peltifer (Smith and Glass) (Stripe-necked
Musk Turtle; 4 individuals at 4 sites [<1%]), Chrysemys dorsalis Agassiz (South-
ern Painted Turtle; 3 at 2 sites [<1%]), and Apalone mutica (Le Sueur) (Smooth
Softshell; 1 at 1 site [<1%]). The remainder of turtle observations were unknown
Graptemys (4 [<1%]), unknown Emydid (39 [8%]), unknown Apalone (7 [1%]),
and unknown turtle (9 [<2%]).
I also conducted basking-density surveys along 249.6 rkm/skm of the TR (n =
4 sites), TTW (n = 8 sites), old channels (n = 5), tributary (n = 8), and lake sites
(n = 8) (Fig. 1, Table 2); these are inclusive of the 5 mark–resight survey sites. At
all locations and all counts, I made 4558 turtle observations (Table 2). The species
observed in order of abundance during these surveys was Black-knobbed Sawbacks
(3042 turtle observations [67% of the total turtles observed]), River Cooter (711
[16%]), Slider Turtle (273 [6%]), Alabama Map Turtle (226 [5%]), Spiny Softshell
Table 2. Basking densities of Graptemys species by site in northeastern Mississippi. Sites are
numbered similar to Figure 1 and are sorted upstream to downstream within each habitat section.
Different superscript letters indicate signicant differences among groups as determined by non-
parametric multiple comparisons post hoc tests and as described in text. Species abbreviations: Gn =
Graptemys nigrinoda, Gpu = Graptemys pulchra, Am = Apalone mutica, As = Apalone spinifera, Cd
= Chrysemys dorsalis, Go = Graptemys ouachitensis, Pc = Pseudemys concinna, Sp = Sternotherus
peltifer, and Ts = Trachemys scripta. Other abbreviations: # = number of surveys, TR = Tombigbee
River, TTW = Tennessee-Tombigbee Waterway, rkm = river kilometer, SD = standard deviation, and
“-”= none observed. [Table continued on following page.]
Total Gn/rkm Gpu/rkm
Site rkm # Gn (± SD) Gpu (± SD) Am As Cd Go Pc Sp Ts
TR
3 2.8 4 5 1.8 ± 1.8 5 1.8 ± 0.7 - 2 - - - - 2
6 11.6 5 74 6.3 ± 2.9 5 0.4 ± 0.6 - - - - 35 - -
9 4.3 3 55 12.7+2.0 7 1.6 ± 1.7 - 2 - - 7 - 2
13 6.6 3 470 71.2 ± 9.9 21 3.2 ± 2.1 - 5 - - 47 - 6
Subtotal 25.3 15 604 19.4 ± 27.4B 38 1.6 ± 1.5A 0 9 0 0 89 0 10
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
456
(43 [1%]), Ouachita Map Turtle (5 [<1%]), Southern Painted Turtle (2 [<1%]),
Stripe-necked Musk Turtle (1 [<1%]), and Smooth Softshell (1 [<1%]). The re-
mainder of observations were unknown Emydid (143 [3%]), unknown Apalone (46
[1%]), unknown Graptemys (35 [<1%]), unknown turtle (29 [<1%]), and unknown
Sternotherus (1 [<1%]).
Table 2, continued.
Total Gn/rkm Gpu/rkm
Site rkm # Gn (± SD) Gpu (± SD) Am As Cd Go Pc Sp Ts
TTW
7 8.6 4 12 1.4 ± 0.7 0 - - - - - 3 - -
11 3.2 1 2 0.6 0 - - 1 - - 1 - -
14 7.5 5 221 29.5 ± 5.5 6 0.8 ± 0.6 - 5 - - 18 - 2
15 1.8 1 40 22.2 3 1.7 - 3 - - 23 - 21
19 8.4 4 328 39.0 ± 2.8 5 0.6 ± 0.2 - - - 2 10 - -
21 2.8 2 334 121.0 ± 60 12 4.3 ± 1.0 - 1 - - 11 - 2
24 8.7 3 216 24.8 ± 15.1 18 2.1 ± 1.0 - 1 - - 14 - 2
27 4.6 4 390 84.1 ± 27.5 14 3.0 ± 1.1 1 - - - 15 - 2
Subtotal 45.6 24 1543 41.0 ± 39.8A 58 1.5 ± 1.5A 1 11 0 2 95 0 29
Old Channels
18 6.0 2 53 8.8 ± 1.6 6 1.0 ± 0.9 - 2 - - 10 - 3
20 4.3 2 21 4.9 ± 0.3 9 2.1 ± 1.0 - - - - 2 - 3
22 7.9 4 322 40.9 ± 10.4 33 4.2 ± 1.5 - - - - 32 - 3
25 15.6 3 193 12.4 ± 2.6 3 0.19 - 1 - - 24 1 14
26 6.0 2 61 10.1 ± 0.2 4 0.7 - - - - 16 - -
Subtotal 39.8 13 650 19.1 ± 16.2AB 55 1.9 ± 1.9A 0 3 0 0 84 1 23
Tributaries
10 3.2 2 - - - - - 3 2 - 52 - 25
17 1.9 2 19 10.2 ± 0.8 2 1.0 ± 1.5 - - - - 3 - 1
28 3.3 3 - - 3 0.9 ± 0.5 - 2 - - 1 - 2
29 7.8 3 4 0.7 ± 1.0 14 1.5 ± 1.9 - - - - 33 - 3
30 7.3 2 40 5.8 ± 3.3 15 2.0 ± 0.2 - - - - 11 - 6
31 8.2 2 132 16.1 ± 7.2 19 2.3 ± 0.5 - 1 - - 34 - 4
32 10.0 2 1 0.1 ± 0.1 2 0.2 - - - - 19 - 9
33 6.6 3 13 2.0 ± 1.1 12 1.8 ± 0.5 - - - - 4 - -
Subtotal 48.3 19 209 3.8 ± 5.8B 66 1.3 ± 1.1A 0 6 2 0 157 0 50
Lakes (density values are per shoreline km)
1 34.1 9 3 0.1 ± 0.3 5 0.1 ± 0.2 - - - 3 26 - 17
2 7.9 4 1 0.2 ± 0.4 2 0.2 ± 0.4 - - - - 107 - 14
4 16.1 5 - - - - - 4 - - 26 - 19
5 4.4 3 - - - - - 1 - - 40 - 22
8 9.8 6 14 1.0 ± 1.6 - - - 2 - - 33 - 16
12 4.3 2 - - - - - 2 - - 15 - 16
16 4.5 5 11 2.3 ± 2.5 1 0.2 ± 0.5 - 3 - - 18 - 36
23 9.6 3 7 0.9 ± 1.0 1 0.1 ± 0.2 - 2 - - 21 - 21
Subtotal 90.7 37 36 0.6 ± 1.3C 9 0.1 ± 0.2B - 14 - 3 286 - 161
Total 249.7 108 3042 15.0 ± 26.9 226 1.0 ± 1.4 1 43 2 5 711 1 273
Southeastern Naturalist
457
W.Selman
2024 Vol. 23, No. 4
Black-knobbed Sawback. Combining information from both point-count and
basking-density surveys, Black-knobbed Sawbacks were documented primarily in
the TR, East Fork of the TR, TTW, and old TR channels. I also conrmed the species
upstream of or near all previously documented localities (Fig. 2, Appendix 1), with
the exception of a historic locality from the southern end of Bull Mountain Creek
near Hwy 50 (MMNS 7733). For the latter, this exact locality was surveyed, and no
Graptemys species were observed. Black-knobbed Sawbacks were observed farther
north in the TR system than previously reported. Several individuals were observed
in northern Itawamba County and into Tishomingo County (Appendix 1), and these
locations are ~75 rkm and ~109 rkm, respectively, upstream of the northernmost
historical record near Smithville (Monroe County; MMNS 3390). Further, Black-
knobbed Sawbacks were also observed in 3 previously undocumented counties
(Tishomingo, Prentiss, Lee) and 5 previously undocumented TR tributaries: Tibbee
Creek, the Luxapallila/Yellow Creek system, Buttahatchie River, Mud Creek, and
Town Creek (Lee County) (Fig. 2, Appendix 1). Black-knobbed Sawbacks were
observed in all 5 of the TR old-channel sites, and they were also present in 5 of the
8 TTW lake sites surveyed (Table 2, Appendix 1); the species had not been previ-
ously reported from any of the TTW lake locations. Black-knobbed Sawbacks were
not observed in several of the smaller tributaries including the Sucarnoochee River
and Bull Mountain Creek (Fig. 2).
Basking-density counts for Black-knobbed Sawbacks at all 33 sites aver-
aged 15.0 individuals/rkm (min–max: 0–121 individuals/rkm; Table 1) and were
different among the habitat types (χ2 = 65.9, df = 4, P < 0.0001; Table 2, Fig. 3).
Nonparametric multiple comparisons indicated that Black-knobbed Sawback den-
sities were highest in the TTW and old channels, and the TTW was greater than
TR, tributary, or lake sections. Densities in old-channels were equal to densities in
TR localities, densities in TR localities were greater than densities in tributaries
or lakes, and densities in tributaries were greater than densities in lakes (Table 2,
Fig. 3). In general, densities were >22 individuals/rkm in the southern portions of
the TR and TTW, downstream of the conuence of East and West Forks of the TR
(sites 13–15, 19, 21, 24, 27), while densities decreased to <12/rkm upstream in the
East Fork of the TR. Black-knobbed Sawback densities in tributaries were low (<6
individuals/rkm) at almost all sites except the Lower Tibbee Creek section (site 31;
16.1 individuals/rkm) and Mattubby Creek (site 10; 10.2 individuals/rkm), while
densities were very low in lakes (<2.3 individuals/skm).
Alabama Map Turtle. Combining information from both point-count and bask-
ing-density surveys, Alabama Map Turtles were documented primarily in the TR,
TTW, larger tributaries (e.g., Buttahatchie, Noxubee, Tibbee), and old channels,
and I conrmed them from upstream of or near all previously documented localities
(Fig. 4). Further, Alabama Map Turtles were observed in 3 previously undocument-
ed counties (Prentiss, Winston, Lauderdale) and 5 previously undocumented TR
tributaries including Tibbee Creek, Luxapallila/Yellow Creek system, Buttahatchie
River, Ponta Creek, and Bull Mountain Creek. A major range extension for Ala-
bama Map Turtles was also observed in the Noxubee River upstream into Winston
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
458
Figure 2. Distribution of
Graptemys nigrinoda (Black-
knobbed Sawback) in north-
eastern Mississippi. Black
dots represent a location
where the species was ob-
served in this study, whereas
small white dots represent
locations where surveys oc-
curred but no individuals
were detected. Gray triangles
represent historical specimen
records.
Southeastern Naturalist
459
W.Selman
2024 Vol. 23, No. 4
County (UF 194496), ~49 rkm upstream of the nearest record (MMNS 5362). The
species was also observed farther north in the TR system than previously reported,
including several individuals observed in Bay Springs Lake (Tishomingo County),
~11 rkm upstream of the nearest record (UF 188673). Alabama Map Turtles were
observed in all 5 of the TR old channel sites, and they were also observed in 4 of
the 8 lake sites surveyed (Table 1, Appendix 1); the species has not been previ-
ously reported from any lake locations. Alabama Map Turtles were not observed
in some tributaries, including the Sucarnoochee River, the entire West Fork of the
Tombigbee (i.e., Mud/Town Creek), upper Tibbee Creek tributaries, and upper Bull
Mountain Creek (Fig. 4).
Basking-density counts of Alabama Map Turtles at the same 33 sites averaged
1.0 individuals/rkm (min–max: 0–4.3 individuals/rkm; Table 1), and densities of
Alabama Map Turtles were low at all sites and habitat types (i.e., all <1.9 individu-
als/rkm). There was a difference in Alabama Map Turtle densities by habitat type
(χ2 = 27.6, df = 3, P < 0.0001). Nonparametric multiple comparisons indicated that
Figure 3. Box plots showing basking densities of Graptemys nigrinoda (Black-knobbed
Sawback; top panel) and Graptemys pulchra (Alabama Map Turtle; bottom panel) in differ-
ent habitats of the Tombigbee River system in northeastern Mississippi. Note the different y-
axes scales on the 2 different species panels. Top and bottom of boxes indicates 75th and 25th
percentiles, respectively, liness in the middle of the boxes are the means, whiskers represent
SE, and dots are outliers. Different letters above the boxes represent signicant differences
by habitat as determined by post-hoc analyses and as described in text. Abbreviations: TR
= Tombigbee River, TTW = Tennessee-Tombigbee Waterway, Gn/rkm = G. nigrinoda per
river km, Gpu/rkm = G. pulchra per river km.
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
460
Figure 4. Distribution of Grapte-
mys pulchra (Alabama Map Tur-
tle) in northeastern Mississippi.
Black dots represent a location
where the species was observed
in this study, whereas small white
dots represent locations where
surveys occurred but no individu-
als were detected. Gray triangles
represent historical specimen re-
cords.
Southeastern Naturalist
461
W.Selman
2024 Vol. 23, No. 4
densities of Alabama Map Turtles were higher in TR, TTW, old channel, and tribu-
tary sections compared to lake sections; there were no other pairwise comparison
differences (Table 2). For sites with densities ≥0.2 for both species, Alabama Map
Turtle densities did not meet or exceed Black-knobbed Sawback densities except
for at 1 site on the Upper East Fork of the TR (site 3), 2 sites on the Buttahatchie
River (sites 28, 29), and 1 site on the lower Luxapallila Creek (site 32).
Trapping and mark–resight population estimates
Black-knobbed Sawbacks. I trapped a total of 107 Black-knobbed Sawbacks at
5 sites in 10.75 total trap days, and the trapping results as measured by catch per
unit effort (CPUE) were high across all sites (mean CPUE: 9.95 captures/trap day)
with the exception of site 14 (5.33 caputures/trap day; Table 3). Males, females, and
juveniles were captured at almost all sites with the exception of site 22 (Table 3),
where no adult females were captured.
Similar to basking-density surveys, mark–resight population estimates varied
across the 5 study sites (mean estimate: 173 individuals/rkm, min–max: 82–397
individuals/rkm; Table 3). Site 27 (397 individuals/rkm) had the largest population
estimate compared to the other sites, while site 13 (167 individuals/rkm) and site
19 (133 individuals/rkm) were moderate in size. Site 14 (87 individuals/rkm) and
21 (82 individuals/rkm) had the lowest population estimates, with the former being
at the upper end of Aberdeen Lake, while the latter was an old channel site.
Table 3. Graptemys captures by site, species, and sex and population estimates in the Tombigbee
River system of northeastern Mississippi, May–June 2022 and 2023. Sites are numbered similar to in
Figure 1 and Table 2 and are sorted upstream to downstream. Abbreviations: M = male, F = female,
Juv = juvenile, CPUE = catch per unit effort, rkm = river kilometer, CI = 95% condence interval, Min
alive = minimum alive estimate from mark-resight population estimate, and * = indicates a population
estimate that was not estimable due to no individuals captured or none resighted.
Pop. estimate:
Trap individuals Min
Site Species # M # F # Juv Total days CPUE /rkm (CI) alive
13 G. nigrinoda 16 2 2 20 1.75 11.40 167 (132–226) 200
G. pulchra 2 1 1 4 1.75 2.30 20 (9–108) 15
14 G. nigrinoda 8 3 5 16 3.00 5.33 87 (68–120) 65
G. pulchra 2 1 - 3 3.00 1.00 4 (5–13) 5
19 G. nigrinoda 21 2 1 24 2.00 12.00 133 (102–185) 106
G. pulchra - 1 - 1 2.00 0.50 * *
22 G. nigrinoda 17 - 4 21 2.00 10.50 82 (70–101) 109
G. pulchra 3 - 3 6 2.00 3.00 16 (10–36) 16
27 G. nigrinoda 19 5 2 26 2.00 13.00 397 (289–588) 159
G. pulchra - - - 0 2.00 0.00 * *
Trapping totals (all sites)
G. nigrinoda 81 12 14 107 10.75 9.95
G. pulchra 7 3 4 14 10.75 1.30
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
462
Alabama Map Turtles. I captured only 14 Alabama Map Turtles in 10.75 trap
days (mean CPUE: 1.3 captures/trap day, min–max: 0–3.0 captures/trap day;
Table 3), and the trapping results were very low across all sites. The highest CPUE
was at site 3 (3.0 captures/trap day), while the lowest was at site 27 where no Ala-
bama Map Turtle individuals were captured; they were observed, however, in very
low densities at site 27 via basking-density surveys. The full complement of males,
females, and juveniles were captured at only 2 of the 5 sites (sites 13). Collectively
at all sites, Alabama Map Turtle CPUE was 7.7× lower than the Black-knobbed
Sawback CPUE.
Because I captured/paint-marked no Alabama Map Turtles at site 27 and only 1
at site 19, mark–resight estimates could not be generated at those sites. Thus, only
3 estimates could be made, with a mean site estimate of 13.3 individuals/rkm (min–
max: 4–20 individuals/rkm; Table 3), and all 3 estimates were low. In comparison,
population estimates for Alabama Map Turtles were 13× lower than Black-knobbed
Sawbacks at the same sites.
Discussion
Black-knobbed Sawback distribution and abundance
Black-knobbed Sawbacks were found upstream of or near almost all histori-
cal localities, but most of the historical records were collected in the 1960s and
1970s prior to the construction of the TTW. Their continued presence at these lo-
calities indicate that they can persist even though the TR is considered a “strongly
affected” basin due to river fragmentation and flow regulation (Dynesius and
Nilsson 1994). In the short term, this finding appears promising, but given the
long lifespan of Graptemys (>40 years for males, >75 years for females; Jones
2017), it seems possible that it has only been 1–2 generations since the TTW
alterations have been made. Thus, many turtles present today could have been
present during construction. Therefore, this relatively short time-frame is not
likely enough time to see population-level changes, and continued monitoring is
needed at these sites in the future.
Along with their persistence in historic localities, I found the species present in
several new drainages. I believe that these new drainage records are not reective
of a range expansion by Black-knobbed Sawbacks, but instead I suspect many of
these drainages have not been previously surveyed for the species. While Cliburn
(1971:16) mentions that “extensive eld work during the spring and summer of
1968 did not reveal Black-knobbed Sawbacks in smaller streams”, there is no docu-
mentation of the locations or streams explored for that study. Because Graptemys
species in smaller streams are warier of potential disturbances (Selman et al. 2013)
and optics/cameras in the 1960s were not as powerful as today, I suspect many
smaller streams where Cliburn (1971) surveyed may have had Black-knobbed
Sawbacks present that were just not detected. An unexpected nding was the ob-
servation of Black-knobbed Sawbacks in Mud and Town creeks near Tupelo (Lee
County), both tributaries of the West Fork of the TR. It is surprising because most
of the surrounding area is row-crop agricultural elds, and consequently, many of
Southeastern Naturalist
463
W.Selman
2024 Vol. 23, No. 4
the streams that make up the West Fork of the TR (e.g., Mud, Town, Chiwapa, and
Yonaba creeks)—along with many other Blackbelt Prairie streams to the north and
south (e.g., Mantachie, Twentymile, Big Brown, Chuquatonchee, Houlka, Trim
Cane, and Catalpa creeks)—have been heavily channelized to convey water quickly
to prevent ooding of agricultural lands. Along with channelization, considerable
amounts of sediment enter the streams from agricultural elds. Therefore, these
channelized streams are shallow, slow-ow, and muddy, and they also lack large
riparian buffers and regular inputs of deadwood. As a consequence, these streams
hosted turtle communities that were more similar to lentic turtle communities
(i.e., more Slider Turtles, Spiny Softshells, and River Cooters). Because this area
has been strongly altered by people over the last 200 years, it seems likely that
historical Black-knobbed Sawback populations were larger and more widespread
in the West Fork of the Tombigbee and other western tributaries prior to the large,
landscape-level anthropogenic alterations.
I was also able to make observations of Black-knobbed Sawbacks farther north
in the drainage compared to historical records. Previously, the most northern record
for Black-knobbed Sawbacks (MMNS 1055) was at the conuence of Bull Moun-
tain Creek and the East Fork of the TR (Southern Itawamba County). However, I
observed the species in the TR and TTW in Itawamba County, and the farthest north
record was into southern Tishomingo County where I captured a male in north-
eastern Bay Springs Lake (UF 194412); the latter was ~56 straight-line km farther
north than MMNS 1055 record. Given the presence of Black-knobbed Sawbacks
in these locations, it seems likely that the species historically occurred upstream
into Mackey’s Creek in southwestern Prentiss and southern Tishomingo counties
prior to the construction of the TTW. The lack of observations from lower Bull
Mountain Creek, a 1968 locality record (MMNS 7733), appears to be a product of
a habitat conversion associated with the TTW construction. Historically, this area
was a moderate-sized stream surrounded by bottomland hardwood forest, whereas
the region now contains areas of swamp and marsh habitats. In this location, I ob-
served many Slider Turtles, River Cooters, and a single Southern Painted Turtle—
all lentic turtle species—but I made no observations of Black-knobbed Sawbacks
or Alabama Map Turtles likely due to the unsuitable habitat currently at the site.
Mean Black-knobbed Sawback densities at all sites in Mississippi (15.0 individ-
uals/rkm) were nearly identical to the mean densities observed by Godwin (2003)
in Alabama (15.3 individuals/rkm). Using similar methods to this study, Godwin
(2003) reported the maximum Black-knobbed Sawback basking density in Alabama
to be 66 individuals/rkm on a site in the Alabama River system, and most sites in
Alabama varied between 10 and 20 individuals/rkm. However, 3 sites on the lower
TR and the TTW (sites 13, 21, 37) had higher basking densities for the species than
previously reported (71–121 individuals/rkm). It is unknown why Black-knobbed
Sawback densities are greater upstream in Mississippi than downstream in Ala-
bama, as this is contrary to previous ndings of narrow-headed Graptemys species
where the highest densities were typically observed in downstream locales (e.g.,
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
464
Selman and Qualls 2009). It is possible that the contrasting ndings could be due
to minor differences in observation techniques used, or it could represent a true dif-
ference in densities.
Further, this study is the rst to conduct mark–resight population estimates
for the species, and the estimates (82–397 individuals/rkm) for Black-knobbed
Sawbacks also far exceed the basking densities reported to date. Relatively few
mark–resight population estimates have been published for Graptemys species for
comparison. Mark–resight estimates in this study for Black-knobbed Sawbacks
(mean: 173 individuals/rkm) was higher than previous estimates for G. oculifera
(Baur) (Ringed Sawback) from the Pearl River system in Louisiana (99 individuals/
rkm; Selman 2024), but they were similar to 2 species in Mississippi including the
Graptemys avimaculata Cagle (Yellow-blotched Sawback; 188 individuals/rkm,
Selman and Qualls 2009) and Ringed Sawback (175 individuals/rkm; Jones and
Harteld 1995). Also, the site with the highest Black-knobbed Sawback density,
site 27 (397 individuals/rkm), was greater than all Ringed Sawback site densities
in Louisiana and Mississippi, but smaller than 1 Yellow-blotched Sawback estimate
(602 individuals/rkm; Selman and Qualls 2009) prior to a decline that followed
Hurricane Katrina (Selman and Qualls 2008). Nonetheless, it would be fruitful
for contemporary surveys to be completed throughout the Alabama portion of the
Black-knobbed Sawback’s range while also including mark–resight population es-
timates at these same sites.
One observed concern is that Black-knobbed Sawback densities were sig-
nicantly lower in lake settings compared to river or TTW settings. This nding
suggests that while Black-knobbed Sawbacks can occur in lakes, lake habitat is
the least suitable of all 5 habitat types surveyed for the species. A similar species,
Ringed Sawback, was also observed in lower densities in lake settings (i.e., ox-
bows) compared to the main river system (Selman 2020).
Alabama Map Turtle distribution and abundance
Similar to Black-knobbed Sawbacks, Alabama Map Turtles were found up-
stream of or near almost all historical localities. Along with their presence near
all historic localities, I observed the species in several new drainages. Like Black-
knobbed Sawbacks, I suspect these drainages were not surveyed previously, and
Cliburn (1971) makes no mention of nding any Alabama Map Turtles in the TR
system. One surprising new drainage for Alabama Map Turtles was the observation
of the species throughout the entire Mississippi portion of the Buttahatchie River
(Lowndes and Monroe counties) to the Alabama state line. This waterway is a small
to moderate-sized river and is one of the last unchannelized and undammed rivers in
the southeastern US, a major juxtaposition to the major hydrologic alterations to the
TR system. It holds a signicant native mollusk fauna including several federally
endangered/threatened freshwater mussels (Williams et al. 2008). Further, ~3240
ha of land surrounding the river is considered conservation lands protected by the
non-prot organization Wildlife Mississippi. Along with the Buttahatchie River,
the Noxubee River basin is also highly forested and lacks high levels of agriculture.
Southeastern Naturalist
465
W.Selman
2024 Vol. 23, No. 4
Federal lands in the Noxubee River basin include the Ackerman Unit of Tombigbee
National Forest (16,109 ha) and the Noxubee Wildlife Refuge (19,424 ha). Both the
Buttahatchie and Noxubee rivers should be considered priority conservation areas
for Alabama Map Turtles in Mississippi; they both have considerable amounts of
riparian habitat that is protected, lack high-density row-crop agriculture, contain
a signicant mussel fauna for this molluscivorous species (Lindeman 2016), and
have had observations of Alabama Map Turtles consistently made throughout the
river systems.
Along with several new drainage records, I was also able to make observa-
tions of Alabama Map Turtles farther north in the drainage compared to historical
records, and these records include a small number of individuals within several
of the TTW lakes. Similar to Black-knobbed Sawbacks, the most northern record
prior to this study for Alabama Map Turtles (UF 188673) was in the TTW, and I
also observed the species in the TTW (Pool E, Prentiss County; Bay Springs Lake,
southern Tishomingo County) and in the East Fork of the TR in central and northern
Itawamba County. Because Alabama Map Turtles are present in these locations, it
seems likely that, as with Black-knobbed Sawbacks, the species historically oc-
curred upstream into Mackey’s Creek.
The lack of observations from all of the streams that form the West Fork of
the TR and from several other streams draining the historic Blackbelt Prairie (see
above) is suggestive that Alabama Map Turtles have been extirpated from portions
of their historical range. There are several lines of supportive evidence for this state-
ment. First, the presence of Black-knobbed Sawbacks in Mud/Town creeks (i.e.,
tributaries of the West Fork of the Tombigbee) is suggestive that these streams are
large enough for Graptemys. Second, I observed Alabama Map Turtles in smaller
streams than Black-knobbed Sawbacks (e.g., Bull Mountain, Ponta Creek, farther
north in the Buttahatchie River). In the Alabama portion of their range, Alabama
Map Turtles are found farther upstream into smaller tributaries than Black-knobbed
Sawbacks, and Black-knobbed Sawbacks and Alabama Map Turtles occur sym-
patrically in downstream portions of the drainage (Godwin 2003). This pattern
is similar to other drainages like the Pascagoula (Selman and Qualls 2009) and
Pearl (Lindeman et al. 2020) rivers, whereby the broad-headed Graptemys species
occurs farther upstream than the narrow-headed species. Thus, one would expect
this same pattern to occur in the West Fork of the TR, but that was not observed.
Third, Alabama Map Turtles are a dietary specialist, with females consuming na-
tive mollusks (Lindeman 2016). Native Unionid mussels are nearly absent from
the West Fork of the TR (M. Wagner, US Fish and Wildlife Service, Mississippi
Ecological Field Ofce, Jackson, MS, pers. comm.), and Unionid mussel declines
are associated with heavy sedimentation runoff from agricultural landscapes and
the sensitivity of mussels to declines/extirpations in areas with heavy sedimentation
(Goldsmith et al. 2021, Österling et al. 2010, Williams et al. 2008). Therefore, it is
likely that Alabama Map Turtles would be more affected in these areas compared
to Black-knobbed Sawbacks, a species that specializes in consuming freshwater
sponges (Lindeman 2016). Even though historical records are lacking from this
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
466
portion of the drainage, the evidence is strongly supportive of local extirpation of
Alabama Map Turtles in the West Fork of the Tombigbee River and possibly other
tributaries within the TR (e.g., upper Tibbee Creek tributaries). Similarly, Ilgen et
al. (2014) found that G. pseudogeographica (Gray) (False Map Turtle), another
mollusk-consuming Graptemys species, was nearly extirpated in the Mermentau
River drainage of southern Louisiana. The authors cited extensive channel dredging
and sedimentation throughout the Mermentau River basin that decimated mussel
populations (Vidrine 1993) and subsequent near extirpation of False Map Turtles
in that system. A further similarity is that the Mermentau River also drains another
intensive agricultural landscape, the historic Cajun Prairie (Vidrine 2010).
In contrast, the lack of Alabama Map Turtle observations from the Sucarnoochee
River (Kemper County) is not as easily explained. Throughout the study, I com-
pleted 23 point-count surveys at 14 locations in the subdrainage without making a
single observation of the species; this effort included a series of 3 repeated surveys
across 3 different years at 4 localities, and these localities all visually appeared
to be suitable habitat for the species. In Ponta Creek, a southern tributary of the
Sucarnoochee River, I completed 9 point-count surveys at 8 locations, document-
ing Alabama Map Turtles at 2 localities, so one might also expect to nd them in
the Sucarnoochee River, the larger of the 2 stream systems. It is possible that they
went undetected at these locations due to very low densities. Additionally, in the
Sucarnoochee River, there were no point-count locations to survey in the most
downstream reaches in southeastern Kemper County (i.e., no bridge crossings and
all land adjacent to the stream is privately owned). Even though the species has
not been documented in Alabama portions of the Sucarnoochee River, I suspect
that Alabama Map Turtles should be present in Alabama portions of the river and
upstream at least to its conuence with Ponta Creek in Sumter County, AL. There-
fore, it may be possible that the species occurs downstream of the survey locations
I completed in Mississippi and may be detected in the Mississippi portions of the
Sucarnoochee River in future surveys.
Mean densities of Alabama Map Turtles were low to very low in Mississippi,
averaging 1.0 individuals/rkm in all habitats surveyed. This estimate is 55% lower
than the mean densities observed by Godwin (2003) in Alabama (2.3 individu-
als/rkm). Similarly, the maximum site density I observed (4.3 individuals/rkm)
is much lower than the maximum density reported in Alabama from a site on the
Coosa River (9.5 individuals/rkm). Furthermore, most sites in Alabama varied be-
tween 1 and 6 individuals/rkm, whereas most sites in Mississippi varied between
0.1 and 2.0 individuals/rkm.
This study is also the rst to conduct mark–resight population estimates for
Alabama Map Turtles. The 2 highest estimates (16 and 20 individuals/rkm) for Ala-
bama Map Turtles exceed the highest basking densities reported, but I suspect that
sites with higher basking densities reported by Godwin (2003) should have higher
population estimates than those reported herein. Many of the highest densities in
Alabama come from the Coosa River, a primarily rocky river system that has a more
stable substrate better suited to benthic mollusks species (Goldsmith et al. 2021);
Southeastern Naturalist
467
W.Selman
2024 Vol. 23, No. 4
this is in stark comparison to the soft/mucky bottom substrate of many of the TR
sites I surveyed in Mississippi. Indeed, Godwin (2003) found Alabama Map Turtles
to be more abundant and outnumber Black-knobbed Sawbacks at sites above the
Fall Line (i.e., rockier substrate), whereas, similar to what I observed, Alabama
Map Turtles were less abundant and were greatly outnumbered by Black-knobbed
Sawbacks at sites below the Fall Line. Additionally, mean mark–resight densities
in this study for Alabama Map Turtles (13.3 individuals/rkm) were lower than
previous estimates for similar megacephalic Graptemys species from other drain-
ages: G. pearlensis Ennen, Lovich, Kreiser, Selman, and Qualls (Pearl Map Turtle)
from the Pearl River system in Louisiana (29.5 individuals/rkm; Selman 2024) and
G. gibbonsi Lovich & McCoy (Pascagoula Map Turtle) from the Pascagoula River
system (39 individuals/rkm; Selman and Qualls 2009).
Even though I documented the species in more locations than previously report-
ed in Mississippi, there are some concerning aspects about Alabama Map Turtles in
Mississippi. First is their aforementioned likely extirpation from the West Fork of
the Tombigbee River. Second, like Black-knobbed Sawbacks, Alabama Map Turtles
appear to be maladapted to lake settings in the TTW. The species was almost nonex-
istent in lake habitat in Mississippi, but that nding does not seem to be consistent
with upstream portions of the drainage where they can occur in higher densities
(e.g., upper Coosa River impoundments; G. Brown, Jacksonville State University,
Jacksonville, AL, pers. comm.). Again, rocky bottom lakes above the Fall Line may
be more conducive to Alabama Map Turtles due to mollusk persistence associated
with more stable substrates. Third, Alabama Map Turtles also occur in much lower
densities than Black-knobbed Sawbacks, and smaller populations are more sus-
ceptible to localized extirpation due to environmental and demographic stochastic
events (Lacy 2000).
Conservation and management considerations
There are several existing threats to Graptemys in the TR system of northeastern
Mississippi including excess sedimentation, lack of river scour on TTW (i.e., to
produce sandbars and sandbanks), additive predation pressure, long-term dyscon-
nectivity of populations, and conversion of lotic systems to lentic systems. Each of
these concerns are discussed hereafter.
First, sedimentation in the TR system appears to be high, particularly from
streams draining intensive agricultural regions on the northwestern side of the
drainage. At numerous stream conuences, I observed higher sediment loads com-
ing from streams draining agricultural regions (e.g., West Fork of the TR, Big
Brown Creek) compared to streams that were draining primarily forested regions
(e.g., East Fork of the TR, Mackey’s Creek). Further, I also observed large amounts
of sediment that accumulated at boat launches following intense rains. Higher
sediment loads in rivers are associated with declines in freshwater mussels, sh, and
other species (Williams et al. 2008, Goldsmith et al. 2021), and this factor could be a
contributor to the extirpation of Alabama Map Turtles from the West Fork of the TR
and possibly other western tributaries. I recommend that efforts be made to provide
landowner incentives to farmers in the western portion of the drainage to improve
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
468
riparian forests margins and implement bafes/barriers in agricultural elds to slow
and limit runoff during heavy rainfall events. This could have measurable benets to
Graptemys species and other aquatic species of concern in the TR system.
Second, because of river regulation via the TTW, many of the TTW sites lack
high-volume river ood pulses needed to create and maintain turtle-nesting habitats
via sediment transport and the scouring of vegetation on sandbars. Without seasonal
ood pulses, only a few minor sand banks currently occur along TTW stretches
of the river. Further, without regular scouring of sandbars, vegetation begins to
encroach upon sandy nesting habitats, and with more vegetation, nests will have
lower temperatures (Horne 1999). With lower temperatures, these nest sites—many
of which are selected through natal homing by freshwater turtles (Freedberg et al.
2005)—will produce a higher number of males, because both Graptemys species
exhibit temperature-dependent sex-determination type 1a (i.e., females produced at
high temperatures, males at low temperatures; Hertwig 2001). Therefore, I recom-
mend that the US Army Corps of Engineers identify existing sand banks along the
TTW and regularly limit/remove the vegetation through mechanical or herbicidal
applications. Similar treatments were completed with habitat for Yellow-blotched
Sawbacks in the Pascagoula River system, and nesting females preferred to nest
in treated, unvegetated sections of sandbar, and these nests produced higher peak
temperatures during incubation (Horne 1999). Further, creating new sand banks ad-
jacent to the TTW during future channel-dredging operations via dredge spoil could
be benecial to riverine turtles. Articial sand banks are known to be attractive to
nesting turtle species (Buhlmann and Osborn 2011, Patterson et al. 2013) and could
provide open habitat, necessary to produce high temperatures for nests, that may be
lacking from the system.
Third, because sandy nesting areas are small pockets on the TTW, nests are
likely more concentrated compared to larger sandbars that existed prior to the
TTW construction. When more concentrated, predators can focus on and decimate
nests in a relatively small area. At many of these small sand banks on the TTW, I
qualitatively observed high rates of predation of turtle nests. Along with sandbank
vegetation management along the TTW, predator control along these stretches
could benet nesting turtles (Engeman et al. 2006) and be a cost-effective manage-
ment tool to improve nesting success and recruitment (Engeman et al. 2002).
Fourth, the numerous lock and dam systems on the TTW contribute to the isola-
tion of tributaries from the main river system (i.e., Bull Mountain Creek) and are
articial barriers to movement. It seems likely that Graptemys populations are less
connected to or entirely disconnected from other populations compared to his-
torical populations. As a consequence, limited gene ow will be a likely result for
these isolated populations (Turcotte et al. 2022). For example, the small number of
Black-knobbed Sawbacks and Alabama Map Turtles observed in Bay Springs Lake
are physically separated from downstream populations by the Bay Springs Lock
and Dam. It seems unlikely that many individuals will move upstream through the
lock and dam, and therefore, the individuals in the lake may represent a small, in-
breeding population that may not be viable in the long-term. Future studies should
Southeastern Naturalist
469
W.Selman
2024 Vol. 23, No. 4
continue to survey Graptemys in these lake habitats, which will help elucidate if
individuals in lakes are “zombie” populations (i.e., lingering around until they die
off) or reproducing and viable populations.
Lastly, there has been little research on impacts of hydrologic alterations to Gulf
Coast Graptemys species, and this study is the rst to show that both Tombigbee
River species are ill-suited to lake settings in the Gulf Coastal portion of the drain-
age. Given the proclivity of state/municipal districts to dam rivers for either ood
control and/or economic development (e.g., One Lake Project on the Pearl River
in Jackson, MS), the data herein should be a bellwether for the dramatic shifts that
occur in turtle communities when they are in lotic systems (e.g., primarily endemic
Graptemys) versus lentic systems (e.g., primarily cosmopolitan Trachemys and
Pseudemys; Selman 2020). It would be benecial for future studies to investigate
the specic mechanisms associated with community shifts and whether they are
due to competition from cosmopolitan species, lack of suitable nesting habitat, lack
of adequate prey sources, migration away from lentic settings, changes in aquatic
predator communities, or a combination thereof.
Acknowledgments
This project was funded by the Mississippi Department of Wildlife, Fisheries, and Parks
(MDWFP), the US Fish and Wildlife Service Division of Federal Aid, and the Cincin-
nati Zoo and Botanical Garden CREW Conservation Fund. Field research activities were
approved by the Mississippi Museum of Natural Science through scientic research and
collecting permit numbers 0510221 and 0510231. Millsaps College Animal Care and Use
Committee approved of the project (WS041717). I appreciate the eld assistance of Ryan
Dumas and Marks McWhorter in 2019 and also the quick and effective boat repairs provid-
ed by RJ’s Outboard (Flowood, MS) and Performance Marine (Columbus, MS) in 2022 and
2023. Justin Hughes (MDWFP) and his team at Canal Section Wildlife Management Area
were helpful in removing excess mud from boat ramps that greatly assisted with launching
my boat at several sites on the East Fork of the Tombigbee River. I am also grateful to Cole-
man Sheehy (Florida Natural History Museum) for archiving photographic vouchers and
Ruth Elsey for reviewing an earlier version of the manuscript. Special thanks to Christine
for “holding down the fort” while I was away; you are a great wife, mom, and teammate.
Literature Cited
Bartmann, R.M., G.C. White, L.H. Carpenter, and R.A. Garrott. 1987. Aerial mark–recap-
ture estimates of conned Mule Deer in pinyon–juniper woodland. Journal of Wildlife
Management 51:41–46.
Berry, G., G. Brown, R.L. Jones, L. Haden, L. Pearson, and W. Selman. 2020. Chutes and
ladders: Drainage exchange of map turtles (Genus Graptemys) across the Tennessee-
Tombigbee Waterway in northeastern Mississippi. Chelonian Conservation and Biology
19:262–267.
Buhlmann, K.A., and C.P. Osborn. 2011. Use of an articial nesting mound by Wood
Turtles (Glyptemys insculpta): A tool for turtle conservation. Northeastern Naturalist
18:315–334.
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
470
Buhlmann, K.A., and T.D. Tuberville. 1998. Use of passive integrated transponder
(PIT) tags for marking small freshwater turtles. Chelonian Conservation and Biology
3:102–104.
Cagle, F.R. 1939. A system of marking turtles for future identification. Copeia
1939:170–173.
Cagle, F.R. 1954. Two new species of the genus Graptemys. Tulane Studies in Zoology
1:165–186.
Cliburn, J.W. 1971. The ranges of four species of Graptemys in Mississippi. Journal of the
Mississippi Academy of Sciences 16:16–19.
Dynesius, M., and C. Nilsson. 1994. Fragmentation and ow regulation of river systems in
the northern third of the world. Science 266:753–762.
Engeman, R.M., S.A. Shwiff, B. Constantin, M. Stahl, and H.T. Smith. 2002. An economic
analysis of predator removal approaches for protecting marine turtle nests at Hobe
Sound National Wildlife Refuge. Ecological Economics 42:469–478.
Engeman, R.A., R.E. Martin, H.T. Smith, J. Woolard, C.K. Crady, B. Constantin, M. Stahl,
and N.P. Groninger. 2006. Impact on predation of sea turtle nests when predator control
was removed midway through the nesting season. Wildlife Research 33:187–192.
Etnier, D.A., and W.C. Starnes. 1993. The Fishes of Tennessee. The University of Tennessee
Press, Knoxville, TN. 681 pp.
Freedberg, S., M.A. Ewert, B.J. Ridenhour, M. Neiman, and C.E. Nelson, 2005. Nesting
delity and molecular evidence for natal homing in the freshwater turtle Graptemys
kohnii. Proceedings of the Royal Society B: Biological Sciences 272:1345–1350.
Godwin, J.C. 2003. Alabama Map Turtle (Graptemys pulchra) status survey. Unpublished
report submitted to the Alabama Department of Conservation and Natural Resources,
Division of Wildlife and Freshwater Fisheries, Montgomery, AL. 20 pp.
Goldsmith, A.M., F.H. Jaber, H. Ahmari, and C.R. Randklev. 2021. Clearing up cloudy
waters: A review of sediment impacts to unionid freshwater mussels. Environmental
Review 29:100–108.
Harteld, P.J., and R.L. Jones. 1989. The status of Epioblasma penita, Pleurobema curtum,
and P. taitianum in the East Fork Tombigbee River. Mississippi Museum of Natural Sci-
ence Technical Report 4, Jackson, MS. 33 pp.
Hertwig, S. 2001. Ökologie, halting, und fortpanzung im terrarium von Graptemys caglei,
G. avimaculata, G. nigrinoda, und G. oculifera. Salamandra 37:21–48.
Horne, B. 1999. Female reproductive ecology and experimental manipulation of nest-
site vegetation: Implications for the conservation of the Yellow-blotched Map Turtle
(Graptemys avimaculata). M.Sc. Thesis. Southeastern Louisiana University, Ham-
mond, LA. 90 pp.
Ilgen, E.L., C.A. Hartson, O.S. Zaleski, and P.V. Lindeman. 2014. Map turtles of the Mer-
mentau: Status surveys of forgotten populations. Chelonian Conservation and Biology
13:1–8.
Jones, R.L. 2017. Long-term trends in Ringed Sawback (Graptemys oculifera) growth,
survivorship, sex ratios, and population sizes in the Pearl River, Mississippi. Chelonian
Conservation and Biology 16:215–228.
Jones, R.L., and P.D. Harteld. 1995. Population size and growth in the turtle Graptemys
oculifera. Journal of Herpetology 29:426–436.
Lacy, R.C. 2000. Considering threats to the viability of small populations using individual-
based models. Ecological Bulletin 48:39–51.
Southeastern Naturalist
471
W.Selman
2024 Vol. 23, No. 4
Lindeman, P.V. 2016. Diets of syntopic Black-knobbed Sawbacks (Graptemys nigrinoda)
and Alabama Map Turtles (Graptemys pulchra) in the Alabama River. American Mid-
land Naturalist 175:194–205.
Lindeman, P.V., A.G. Gibson, W. Selman, R.L. Jones, G.J. Brown, C.C. Huntzinger, and
C.P. Qualls. 2020. Population status of megacephalic map turtles Graptemys pearlensis
and Graptemys gibbonsi and recommendations regarding their listing under the US En-
dangered Species Act. Chelonian Conservation and Biology 19:165–185.
McCoy, C.J., and R.C. Vogt. 1979. Distribution and population status of the Ringed Saw-
back (Graptemys oculifera), Blotched Sawback (Graptemys avimaculata), and Black-
knobbed Sawback (Graptemys nigrinoda) in Alabama and Mississippi. Final Report, US
Fish and Wildlife Contract No. 14-16-0004-79-038. USFWS, Washington, DC. 30 pp.
Mississippi Museum of Natural Science. 2015. Mississippi state wildlife action plan. Mis-
sissippi Department of Wildlife, Fisheries, and Parks. Mississippi Museum of Natural
Science, Jackson, MS. 692 pp.
NatureServe Explorer. 2024a. Graptemys nigrinoda: Black-knobbed Map Turtle. Avail-
able online at https://explorer.natureserve.org/Taxon/ELEMENT_GLOBAL.2.106155/
Graptemys_nigrinoda. Accessed 7 November 2024.
NatureServe Explorer. 2024ab. Graptemys pulchra: Alabama Map Turtle. Available online
at https://explorer.natureserve.org/Taxon/ELEMENT_GLOBAL.2.103963/Graptemys_
pulchra. Accessed 7 November 2024.
Österling, M.E., B.L. Arvidsson, and L.A. Greenberg. 2010. Habitat degradation and the
decline of the threatened mussel Margaritifera margaritifera: Inuence of turbidity and
sedimentation on the mussel and its host. Journal of Applied Ecology 47:759–768.
Patterson, J.E., B.D. Steinberg, and J.D. Litzgus. 2013. Not just any old pile of dirt: Evalu-
ating the use of articial nesting mounds as a conservation tool for freshwater turtles.
Oryx 47:607–615.
Selman, W. 2020. River turtles and one dam lake: Two imperiled Graptemys species in the
Pearl River and potential impacts of the proposed of One Lake Project. Chelonian Con-
servation and Biology 19:186–196.
Selman, W. 2024. Distribution and abundance of Graptemys oculifera (Ringed Sawback)
and Graptemys pearlensis (Pearl Map Turtle) in the Pearl River system of Louisiana.
Herpetological Conservation and Biology 19:129–143.
Selman, W., and C. Qualls. 2008. The impacts of Hurricane Katrina on a population of
Yellow-blotched Sawbacks (Graptemys avimaculata) in the lower Pascagoula River.
Herpetological Conservation and Biology 3:224–230.
Selman, W., and C. Qualls. 2009. Distribution and abundance of two imperiled Grapte-
mys species of the Pascagoula River system. Herpetological Conservation and Biology
4:171–184.
Selman, W., J. Jawor, and C. Qualls. 2012. Seasonal variation of corticosterone levels in the
Yellow-blotched Sawback (Graptemys avimaculata), an imperiled freshwater turtle.
Copeia 2012:698–705.
Selman, W., C. Qualls, and J.C. Owen. 2013. Effects of human disturbance on the behav-
ior and physiology of an imperiled freshwater turtle. Journal of Wildlife Management
77:877–885.
Shoop, C.R. 1967. Graptemys nigrinoda in Mississippi. Herpetologica 23:56.
Stine, J. K. 1991. Environmental politics in the American south: The ght over the Tennes-
see- Tombigbee Waterway. Environmental History Review 15:1–24.
Southeastern Naturalist
W.Selman
2024 Vol. 23, No. 4
472
Taylor, C.M., D.S. Millican, M.E. Roberts, and W.T. Slack. 2008. Long-term change to
sh assemblages and the ow regime in a southeastern US river system after extensive
aquatic ecosystem fragmentation. Ecography 31:787–797.
Turcotte, A., G. Blouin-Demers, and D. Garant. 2022. Exploring the effect of 195 years-old
locks on species movement: Landscape genetics of painted turtles in the Rideau Canal,
Canada. Conservation Genetics 23:467–479.
US Fish and Wildlife Service. 2011. Endangered and threatened wildlife and plants; partial
90-day nding on a petition to list 404 species in the southeastern United States as en-
dangered or threatened with critical habitat. Federal Register 76:59836–59862.
van Dijk, P.P. 2011a. Graptemys nigrinoda. The IUCN Red List of Threatened Spe-
cies 2011: e.T9502A97420750. Available online at https://www.iucnredlist.org/spe-
cies/9502/97420750. Accessed on 13 July 2023.
van Dijk, P.P. 2011b. Graptemys pulchra. The IUCN Red List of Threatened Species
2011: e.T170494A97424569. Available online at https://www.iucnredlist.org/spe-
cies/170494/97424569. Accessed on 13 July 2023.
Vidrine, M.F. 1993. The historical Distributions of Freshwater Mussels in Louisiana. Gail
Q. Vidrine Collectables, Eunice, LA. 225 pp.
Vidrine, M.F. 2010. The Cajun Prairie: A Natural History. Gail Q. Vidrine Collectables,
Eunice, LA. 314 pp.
White, G.C. 1996. NOREMARK: Population estimation from mark–resighting surveys.
Wildlife Society Bulletin 24:50–52.
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.
Southeastern Naturalist
473
W.Selman
2024 Vol. 23, No. 4
Appendix 1. Graptemys nigrinoda (Black-knobbed Sawback) and Graptemys pulchra (Alabama
Map Turtle) vouchered photographs by county and drainage in northeastern Mississippi. The table is
organized by counties north to south. Abbreviations: TTW = Tennessee-Tombigbee Waterway, UF =
University of Florida Herpetological Collections, * = new county record, + = new drainage record, #
= new lake record.
Species County Water body UF voucher number(s)
G. nigrinoda Tishomingo* Bay Springs Lake# 194412
Prentiss* Montgomery Lake# 194421
Lee* Mud Creek+ 190995
Town Creek+ 190956
Itawamba Tombigbee River 190957–190958, 194447–
194448, 194505
TTW 194501
Wilkins Lake# 194498-194500
Monroe Aberdeen Lake# 194458, 194459, 194461
Buttahatchie River+ 194479, 194416
Mattubby Creek+ 194491, 194493
Old Tombigbee River Channel 194482
Tombigbee River 194468
TTW 194462, 194464, 194474
Clay Tibbee Creek+ 194437, 194440
Old Tombigbee River Channel 194451
Lowndes Columbus Lake# 194456
Old Tombigbee River Channel 194442, 194443
TTW 194454, 194503
Yellow Creek+ 194509
Noxubee Noxubee River 194432, 194434
G. pulchra Tishomingo Bay Springs Lake# 194402–194405, 194408,
194411
TTW 194418, 194514
Prentiss* Montgomery Lake# 194419
Itawamba Bull Mountain Creek+ 194512
Tombigbee River 194449, 194504, 194506
Monroe Aberdeen Lake# 194460
Buttahatchie River+ 194413, 194414, 194476,
194478
Mattubby Creek+ 194490, 194494
Old Tombigbee River Channel 194446, 194481
Tombigbee River 194466
TTW 194471, 194463
Clay Old Tombigbee River Channel 194452
Tibbee Creek+ 194438, 194441
Lowndes Buttahatchie River+ 194477
Columbus Lake# 194457
Luxapallila Creek+ 194417
Old Tombigbee River Channel 194444
TTW 194455, 194502
Yellow Creek+ 194508, 194511
Winston* Noxubee River 194496
Noxubee Noxubee River 194425, 194426, 194431,
194433, 194435
Lauderdale* Ponta Creek+ 190959
