Content uploaded by Jan Jeffrey Hoover
Author content
All content in this area was uploaded by Jan Jeffrey Hoover on Jul 14, 2016
Content may be subject to copyright.
1
American Fisheries Society Symposium 66:000–000, 2009
© 2009 by the American Fisheries Society
Morphological Variation in Juvenile Paddlesh
Ja n Je f f r e y Ho o v e r *, Kr i s t a a. Bo y s e n ,
Ca t H e r i n e e. Mu r p H y , a n d st e v e n G. Ge o r G e
U.S. Army Engineer Research and Development Center
Waterways Experiment Station
3909 Halls Ferry Road, Vicksburg, Mississippi 39180, USA
Abstract.—Juvenile paddlefish Polyodon spathula exhibit conspicuous
variation in the shape of their rostra and caudal fins. We quantified morpho-
logical variation for a composite collection of young-of-year paddlefish (N
= 55; 61.9–403.7 mm total length) using nine measurements of the rostrum,
body, and caudal fin. Sheared principal component analysis of morphologi-
cal data resulted in three distinct groups of fish corresponding to three dif-
ferent localities: hatchery-reared fish from the Mermentau River, Louisiana;
hatchery-reared fish from the Tombigbee River, Alabama; and field-collected
fish from the Mississippi River, Mississippi. Series were segregated from each
other based on size of caudal lobes and width of rostrum. With increased
body size of fish, relative length of rostrum increased, and mesal expansion
of rostrum increased for all three series. For Tombigbee and Mississippi Riv-
er series, with increased size of fish, caudal asymmetry decreased. Overall,
smaller fish had shorter, narrower rostra and highly asymmetrical caudal
lobes; larger fish had longer, broader rostra and more symmetrical caudal
lobes. Morphological differences among series were most conspicuous for
fish greater than 85 mm eye-to-fork length. Fish from Mermentau River had
shorter, narrower (leaf-shaped) rostra and asymmetrical (conspicuously het-
erocercal) caudal lobes. Those from the Tombigbee River had longer, broader
(spoon-shaped) rostra and more symmetrical caudal lobes. Those from the
Mississippi River had the longest, broadest (paddle-shaped) rostra and sym-
metrical (superficially “homocercal”) caudal lobes. Locality-based gradient
in paddlefish morphology corresponds to a gradient in river hydrology: lon-
ger, broader rostra and symmetrical caudal lobes were associated with larger
basins, higher gradients, and greater discharge.
* Corresponding author: jan.j.hoover@usace.
army.mil
Introduction
Nineteenth-century naturalists observed
sufficient morphological variation in pad-
dlefish to recognize two different “spe-
cies”: a small, toothed form with a longer
rostrum and asymmetrical tail, Polyodon
folium, and a large, toothless form with
shorter rostrum and symmetrical tail, Plat-
irostra edentula (LeSueur 1818; Rafinesque
1820; Kirtland 1842, 1845). Some of this
variation was subsequently attributed to
allometric growth of the rostrum of the
fish (Brehm and Hacke 1892; Wagner 1904;
Hildebrand and Towers 1927), which was
later confirmed and quantified (Thompson
1934; Grande and Bemis 1991; Hoover et al.
2000). As young-of-year paddlefish grow,
2h o o v e r e t a l .
rostrum growth accelerates and it becomes
longer relative to the body of the fish (posi-
tive allometry). Relative rostrum length is
greatest in fish 240–310 mm total length
(TL). At larger sizes, growth of the rostrum
decelerates and matches growth of the
body (isometry). After age 1, as paddlefish
continue to grow, growth of the rostrum
slows even further and becomes shorter
relative to the body of the fish (negative al-
lometry). Breadth of the rostrum exhibits a
similar pattern (Grande and Bemis 1991).
However, not all variation in morphology
is related to growth.
An early field study suggested the pos-
sibility of distinct paddlefish morphotypes
in the Mississippi Delta (Stockard 1907). It
noted that adult fish with longer, narrower
rostra inhabited oxbow lakes and those
with shorter, broader rostra inhabited riv-
ers. Recent study of that population sup-
ported the existence of these two rostral
morphotypes, but the form with the long,
narrow rostrum was substantially more
abundant and phenotypically variable,
thus making unequivocal field identifica-
tion of morphotypes difficult (Hoover et
al. 2000). Juvenile paddlefish, however, are
morphologically distinctive from adults
and appear to show greater variation in
appearance (e.g., pigmentation, rostrum
size and shape). Identification of distinc-
tive morphotypes, then, may be easier with
smaller fish. To date, the extent of morpho-
logical variation and presence of distinct
morphotypes in juvenile paddlefish has
not been investigated.
Methods
Paddlefish were obtained from three sourc-
es: hatchery-reared fish from broodstock
obtained in the Mermentau River, Louisi-
ana, near Jennings, Louisiana in 2000 (N =
20); hatchery-reared fish from broodstock
obtained in the Tombigbee River, Alabama
near Demopolis in 2004 (N = 20); and field-
collected fish from the lower Mississippi
River, Mississippi, 2001–2006 (N = 15).
Two-thirds (10 of 15) of the Mississippi
River fish were collected at Mhoon Bend
(Rivermile 694), near Tunica, Mississippi,
midway between Helena, Arkansas and
Memphis, Tennessee. Drainage, eleva-
tion, and stream discharge are lowest in
the Mermentau River, intermediate in the
Tombigbee River, and greatest in the lower
Mississippi River (Table 1). Fish from all
three locations were obtained live, fixed
in 10% formalin, rinsed, and transferred to
70% ethanol.
Nine measurements were made on
each fish to the nearest 0.01 mm using
Mitutoyo Absolute IP66 digital calipers.
Measurements were (1) eye-to-fork length
(EFL), from the anterior orbit of the eye to
Table 1. Area, elevation, and discharge of three rivers: drainage area is for entire Mer-
mentau River (Demcheck and Skrobialowski 2003) and Tombigbee River (Mettee et al.
1996) and for the lower Mississippi River, below its conuence with the Ohio River (Baker
et al. 1991). U.S. Geological Survey (USGS) elevation data are for Mermentau River (HUC
08080202), middle Tombigbee River-Chicasaw (HUC 0316021), and lower Mississippi Riv-
er-Memphis (HUC 08010100). USGS discharge data are for the same localities for water
year 2005 (Mermentau River, Tombigbee River) and for the period of record 1934–1988
(Mississippi River).
Characteristic Mermentau River Tombigbee River Mississippi River
Drainage area (km2) 9,842 35,628 130,000
Mean (SD) elevation (m) 1.6 (2.5) 57.4 (28.5) 80.5 (15.9)
Discharge (ft3/s) <10,000 (<3 m3/s) 6,019–52,380 >130,000
(1,834.6–15,965.4 m3/s) (>39,624 m3/s)
3m o r p h o l o g i c a l va r i a t i o n i n j u v e n i l e p a d d l e f i s h
the innermost curve of the caudal fin; (2)
rostrum length (RL), from the end of the
rostrum to the anterior orbit of the eye; (3)
length of upper caudal lobe (UCL), from
the innermost curve of the caudal fin to the
tip of the upper caudal (epicaudal) lobe; (4)
length of lower caudal lobe (LCL), from the
innermost curve of the caudal fin to the tip
of the lower caudal (hypocaudal) lobe; (5)
eye-to-dorsal distance, from the anterior
orbit of the eye to the origin of the dorsal
fin; (6) proximal rostrum breadth (PRB),
width of the rostrum approximately 20%
of the distance from the eye to the end of
the rostrum (rostral base); (7) mesal rostrum
breadth (MRB), width of the rostrum ap-
proximately 50% of the distance from eye
to the end of the rostrum (rostral shaft);
(8) distal rostrum breadth, width of the
rostrum at 80% of the distance from the
eye to the end of the rostrum (rostral tip);
(9) eye to barbel distance (EBD), from the
anterior orbit of the eye to the origin of
the left barbel. Presence or absence of an
upper caudal (epicaudal) notch was not-
ed. Total length was calculated as the sum
of rostrum length, eye-to-fork length, and
length of the upper caudal lobe. It was
used as a descriptor of overall size of fish
used in study and for comparing preva-
lence of qualitative characters (barbels,
epicaudal notches). Eye-to-fork length,
because it is not influenced by size of
highly variable structures (rostrum, cau-
dal fin), was used to standardize measure-
ments for fish used in comparisons among
groups (e.g., locality).
Data were compiled and analyzed us-
ing SAS 9.1 (SAS Institute, Cary, North Car-
olina). For each series, mean values for all
nine characters were calculated. For each
character, significant differences among
series was determined using analysis of
variance (ANOVA) and Student-Newman-
Keuls (SNK) test, with alpha = 0.05. To
identify morphologically important vari-
ables and to quantify degree of morpholog-
ical similarity within and between groups,
we performed ordination analyses on raw
(nontransformed, nonstandardized) data.
Fish were divided into two size-classes
(<85 mm EFL, $85 mm EFL) such that
sample size of Mississippi River specimens
(field-collected fish) was approximately
equal in each class. An analysis of mor-
photypes within the collective series was
done using sheared principal component
analysis (sPCA). This technique is similar
to traditional principal component analy-
sis (PCA). It differs from PCA by “shear-
ing” the size effects from morphometric
relationships so that only shape compo-
nents remain. It is especially useful when
evaluating morphology of fish exhibit-
ing allometric growth and has been used
previously in studies of river sturgeons
Scaphirhynchus spp. (Mayden and Kuhaj-
da 1996; Kuhajda et al. 2007; Murphy et al.
2007). Following protocol used for previ-
ous analysis of hatchery-reared sturgeon,
separate ordinations were performed for
the small size-class (<85 mm EFL) and the
large size-class ($85 mm EFL) (Kuhajda et
al. 2007). All nine variables were used in
sPCA. First axis is interpreted as size com-
ponent, second and third axes (sheared PC2
and sheared PC3) as shape components (D.
L. Swofford, SAS Program for computing
sheared PCA, unpublished data).
To describe overall trends in allometric
growth of the rostrum and caudal fin, we
used three morphometric indices: (1) rela-
tive rostrum length (RL/EFL), (2) mesal
rostrum expansion (MRB/PRB), and (3)
caudal asymmetry (UCL/LCL). Relative
rostrum length describes the rostrum size
relative to body size with values for adult
fish typically ranging from 0.10 to 0.50 or
100–500‰ (thousandths) EFL (Thompson
1934; Hoover et al. 2000). Mesal rostrum
expansion describes width of the rostrum
shaft relative to its base. Values can range
from less than 1.00 when tapered or wedge-
shaped, approximately 1.00 when paddle-
4h o o v e r e t a l .
or blade-shaped, or greather than 1.50 if
leaf- or spoon-shaped (Hoover et al. 2000).
Caudal asymmetry describes the dispar-
ity in size between the upper and lower
caudal lobes of the caudal fin. Values less
than 1.00 reflect an asymmetrical tail with a
large lower caudal lobe (rare in extant spe-
cies of fish), 1.00 a symmetrical tail with
equal-sized lobes, and values greater than
1.00 an asymmetrical tail with a larger up-
per lobe. Relationships between morpho-
metric indices and EFL of paddlefish were
quantified using bivariate plots and linear
and polynomial regressions. Regression
model selected as representative was that
which accounted for the higher amount of
variance (i.e., greatest R2) and for which all
estimates of model parameters (intercept,
EFL, EFL2) were statistically significant (p
< 0.05). Correlations among certain other
morphological characters were also evalu-
ated using regression analyses. For each
character within a size-class showing mini-
mal allometric growth, significant differ-
ences among series was determined using
ANOVA and SNK test, with alpha = 0.05.
Results
Fish examined in this study ranged in size
from 61.9 to 403.7 mm TL. Length range
was low for Mermentau River series and
specimens were large: 153.4–255.6 mm TL.
Length ranges were greater and similar for
series from the Tombigbee River, 75.8–315.9
mm TL, and Mississippi River, 61.9–403.7
mm TL. The largest fish collected from the
Mississippi River (403.7 mm TL) was sub-
stantially larger than the next largest speci-
men from that series (305.7 mm TL) but
was retained in subsequent analyses for
evaluation of relationships between size
and morphology.
The majority of paddlefish examined
had barbels and an epicaudal notch. Every
fish examined had two intact barbels ex-
cept one 122.4-mm-TL specimen from the
Tombigbee series in which only the barbel
bases were present. Eighty-five percent of
fish (47/55) had a notch in the upper cau-
dal lobe. Fish having an epicaudal notch
comprised 70% of the Mermentau River se-
ries, 90% of the Tombigbee River series, and
100% of the Mississippi River series. Fish
lacking an upper caudal notch were large,
167.5–242.5 mm TL, suggesting that the fea-
ture is lost as fish grow; however, notches
were observed on fish of comparable size
and larger, 165.0–403.7 mm TL, suggesting
that notches are typical of most specimens
less than 400 mm TL. Because these traits
were ubiquitous or nearly-so, they were ex-
cluded from subsequent analyses.
Significant differences among series
were observed for the majority of morpho-
logical characters (Table 2). Highly signifi-
cant differences were observed for length
of upper caudal lobe, proximal rostrum
breadth, mesal rostrum breadth, and dis-
tance from eye to barbel (F > 6.1, p < 0.0005,
df = 2/52). Length of the upper caudal lobe
was significantly greater in Mississippi
River and Tombigbee River fish than in
Mermentau River fish. Proximal and mesal
rostrum breadth were significantly greater
in Mississippi River fish than in Mermentau
River fish, which in turn were significantly
greater than Tombigbee River fish. Eye to
barbel distance was significantly greater in
Mississippi River and Tombigbee River fish
than in Mermentau River fish. Significant
differences among series were observed for
eye-to-fork length EFL and length of lower
caudal lobe (F > 3.35, p < 0.05, df = 2/52).
Eye-to-fork length was greatest for the
Mermentau River specimens, significantly
lower for the Tombigbee River specimens,
and intermediate (and not significantly dif-
ferent) for fish from the Mississippi River.
Length of the lower caudal lobe was sig-
nificantly greater in Mississippi River fish
than in Mermentau River fish and was in-
termediate in Tombigbee River fish. There
were no significant differences among the
5m o r p h o l o g i c a l va r i a t i o n i n j u v e n i l e p a d d l e f i s h
Table 2. Morphometric data for three series of juvenile paddlesh. Values for eye-to-fork
length are in milimeters; all other values are expressed as 1,000s (‰) of eye-to-fork
length. Superscripts indicate values for a character that are signicantly different among
series based on analysis of variance and Student-Newman-Keuls test.
Mermentau River Tombigbee River Mississippi River
Character N = 20 N = 20 N = 15
Eye-to-fork length 116.8a 87.9b 95.37ab
Rostrum length 600.5a 543.9a 625.7a
Upper caudal lobe 198.6b 214.9a 224.9a
Lower caudal lobe 79.8b 97.7ab 113.3a
Eye-to-dorsal distance 553.3a 536.7a 547.5a
Proximal rostrum breadth (base) 126.7b 117.2c 142.4a
Mesal rostrum breadth (shaft) 155.3b 121.6c 164.2a
Distal rostrum breadth (tip) 129.9a 125.4a 137.0a
Eye-to-barbel distance 108.9b 123.1a 126.0a
series in relative rostrum length, distance
from eye to dorsal fin, and distal rostrum
breadth (F < 3.2, p > 0.05, df = 2/52).
Mean values for proximal and mesal
rostrum breadth and for length of the up-
per and lower caudal lobes suggested dif-
ferent morphotypes from each locality: one
form with a leaf-shaped rostrum (moder-
ately wide base, wider shaft, narrow tip)
and small, asymmetrical tail from the Mer-
mentau River; a second form with a spoon-
shaped rostrum (narrow base, wider shaft,
widest at the tip) and large, more symmet-
rical tail from the Tombigbee River; a third
form with a paddle- or blade-shaped ros-
trum (wide base, wider shaft) and larger
tail of greater symmetry from the Missis-
sippi River (Table 2). Measures of central
tendency for individual characters, howev-
er, do not address individual variation and
distinctive combinations of characters that
may take place as the animal grows.
Sheared PCA for each size-class identi-
fied rostral and caudal characters that ac-
counted for individual variation among se-
ries (Table 3). For small juveniles, sheared
PC2 was positively associated with mesal
and proximal rostrum breadth. Sheared
PC3 was positively associated with distal
rostrum breadth and length of upper cau-
Table 3. Character loadings from sheared principal components analysis of nine morpho-
metric characters for juvenile paddlesh. High-loading variables (|>0.40|) on sheared
principal components (PC) indicated in boldface. Eye-to-fork length = EFL.
<85 mm EFL #85 mm EFL
(N = 20) (N = 35)
Sheared Sheared Sheared Sheared
Size PC2 PC3 Size PC2 PC3
Eye-to-fork length –0.2335 –0.2210 0.1131 –0.3041 0.3226 –0.2401
Rostrum length –0.4121 –0.3194 0.0137 –0.3384 0.1609 0.2909
Upper caudal lobe length –0.2507 –0.0489 0.4251 –0.4040 0.1443 –0.5522
Lower caudal lobe length –0.4676 –0.2580 –0.4534 –0.4497 –0.7699 0.0023
Eye to dorsal n distance –0.2581 –0.2291 0.1424 –0.3200 0.3348 –0.1834
Proximal rostrum breadth –0.2181 0.4589 –0.3107 –0.2492 0.1476 0.2115
Mesal rostrum breadth –0.4027 0.5296 –0.2874 –0.2778 0.2227 0.6746
Distal rostrum breadth –0.3704 0.3900 0.6316 –0.2928 –0.1698 0.1201
Eye to barbel distance –0.2838 –0.2907 –0.0489 –0.3164 –0.0652 –0.0633
6h o o v e r e t a l .
dal lobe, negatively with length of the low-
er caudal lobe. For large juveniles, sheared
PC2 was negatively associated with length
of lower caudal lobe. Sheared PC3 was
positively associated with mesal rostrum
breadth and negatively associated with
length of the upper caudal lobe.
Ordination of paddlefish in multi-
variate space demonstrated morphologi-
cal separation among the three series. For
small juveniles, clusters of Tombigbee and
Mississippi River fish were nearly dis-
crete along sheared PC2 (Figure 1A). Fish
from the Tombigbee River typically had
-0.15
-0.1
-0.05
0
0.05
0.1
-0.18 -0.13 -0.08 -0.03 0.02 0.07 0.12 0.17
Sheared PC3
Sheared PC2
-0.1
-0.05
0
0.05
0.1
-0.2 -0.15-0.1 -0.0500.05 0.1 0.15 0.2
Sheared PC3
Sheared PC2
A
B
Figure 1. Sheared principal components analyses of paddlesh. Black symbols and solid
polygon represent Mermentau River sh. Gray symbols and dashed polygon represents Tom-
bigbee River sh; open symbols and stippled line represents Mississippi River sh. A. Fish
less than 85 mm eye-to-fork length (EFL). B. Fish is greater than or equal to 85 mm EFL.
7m o r p h o l o g i c a l va r i a t i o n i n j u v e n i l e p a d d l e f i s h
rostra that were narrow proximally (base)
and mesally (shaft). Fish from the Missis-
sippi River typically had rostra that were
broad proximally and mesally. Ordination
of large juveniles indicated three discrete
clusters of the series along both axes (Fig-
ure 1B). Fish from the Mississippi River
and Tombigbee River had relatively large
lower caudal lobes when compared with
fish from the Mermentau River and were
well separated along sheared PC2. Rostra
of Mississippi River fish were broader me-
sally than those of the Tombigbee River and
were well separated along sheared PC3.
Relative rostrum length ranged from
342.9‰ to 835.0‰ EFL overall and was
positively correlated with size of fish for
each series (Figure 2A). Relationships were
highly significant for Mississippi River
and Tombigbee River series (R2 > 0.90, p
< 0.0001) and significant for Mermentau
River series (R2 = 0.27, p = 0.0231). For
Mississippi River and Tombigbee River
series, relative size of rostrum increased
substantially in fish up to 85 mm EFL, was
approximately isometric for fish 85–165
mm EFL, and then decreased. Differences
in values among series were apparent in
fish greater than 65 mm EFL but were most
pronounced for fishes greater than 95 mm
EFL. At those sizes, relative rostrum length
was more than 689‰ EFL for all Mississip-
pi River specimens, 609–675‰ EFL for all
Tombigbee River specimens, and less than
609‰ EFL for 68% of the Mermentau River
specimens.
Growth of both anterior and poste-
rior portions of the rostrum contributed
to overall relative length of rostrum. Rela-
tive distance from tip of rostrum to barbel,
(RL – EBD)/EFL, was positively related
with EFL for Mississippi River (R2 = 0.95,
p < 0.0001), Tombigbee River (R2 = 0.92, p <
0.0001), and Mermentau River (R2 = 0.40, p
= 0.0037) series. Relative distance from bar-
bel to eye, EBD/EFL, was positively relat-
ed with EFL for Mississippi River series (R2
= 0.59, p = 0.0375), Tombigbee River series
(R2 = 0.60, p = 0.0001), but not for Mermen-
tau River series (R2 = 0.02, p = 0.5143) for
which there were no small specimens.
Mesal rostrum expansion ranged from
0.65 (shaft 65% as wide as base) to 1.50
(shaft 50% wider than base) and was posi-
tively correlated with size of paddlefish for
each series (Figure 2B). Relationships were
highly significant for all three series (R2 >
0.62, p < 0.0001). The regression curve for
Mermentau River series intersected that
of the Tombigbee River series at approxi-
mately 110 mm EFL. This indicates more
pronounced allometric growth of the ros-
trum in Mermentau River paddlefish, with
smaller specimens having rostra narrower
or comparable in expansion to Tombigbee
River fish. For Mississippi River and Tom-
bigbee River series, mesal rostrum expan-
sion increased in fish up to 85 mm EFL, was
approximately isometric for fish 85–165
mm EFL, and then decreased. Differences
in values were less apparent among series
than those for relative rostrum length. For
fish greater than 95 mm EFL, mesal ros-
trum expansion was more than 1.30 for
65% of Mississippi River specimens, 1.08–
1.27 for all Tombigbee River specimens,
and 1.08–1.27 for 79% of Mermentau River
specimens.
Caudal asymmetry ranged from 5.27
(upper caudal lobe 5.3 times longer than
lower caudal lobe) to 1.24 (upper caudal
lobe only 24% longer than lower caudal
lobe) overall and was negatively related
with size of fish for the Mississippi and
Tombigbee rivers (R2 > 0.80, p < 0.0001), but
not for the Mermentau River fish (R2= 0.02,
p = 0.5212) (Figure 2C). Regression curves
for Mississippi River and Tombigbee River
series intersected at approximately 70 mm
EFL, suggesting that fish at that approxi-
mate size would be difficult to distinguish
by this particular character. For those two
series, caudal asymmetry decreased in fish
up to 85 mm EFL and was approximately
8h o o v e r e t a l .
Figure 2. Relationships between size of juvenile paddlesh and morphometric indices. Black
symbols and solid line represent Mermentau River sh. Gray symbols and dashed line rep-
resent Tombigbee River sh. Open symbols and stippled line represent Mississippi River
sh. A. Relative rostrum length. B. Mesal rostrum expansion. C. Caudal asymmetry.
y = 1.4198x + 430.69
R2= 0.2683, p = 0.0231
y = -0 .024x2+ 6.9675x + 148.39
R2= 0.9196, p < 0.0001
y = -0.0403x2+ 11.664x -23.049
R2= 0.9422, p < 0.0001
250
350
450
550
650
750
850
25 45 65 85 105125 145165 185 205225
Relative Rostrum Length
Eye to Fork Length (mm)
y = 0.0045x + 1.9865
R2= 0.0246, p = 0.5212
y = 0.0002x2-0.0522x + 5.666
R2= 0.8137, p < 0.0001
y = 0.0003x2-0.0814x + 7.2315
R2= 0.8594, p < 0.0001
0
1
2
3
4
5
6
25 45 65 85 105125 145 165185 205225
Caudal Assymetry
Eye to Fork Length (mm)
C
y = 0.0066x + 0.4603
R2= 0.6307, p < 0.0001
y = -5E-05x2+ 0.0132x + 0.3101
R2= 0.7329, p < 0.0001
y = -7E-05x2+ 0.018x + 0.1953
R2= 0.8187, p < 0.0001
0.6
0.8
1
1.2
1.4
1.6
25 45 65 85 105125 145165 185205 225
Mesal Rostrum Expansion
Eye to Fork Length (mm)
A
B
9m o r p h o l o g i c a l va r i a t i o n i n j u v e n i l e p a d d l e f i s h
isometric in larger fish. The apparent in-
crease in caudal asymmetry for Mississippi
River fish greater than 165 mm EFL is driv-
en by a single large specimen. Differences
in values for fish larger than 95 mm EFL
were apparent for Mermentau River speci-
mens, all of which had caudal asymmetry
greater than 2.0. Values for the other two
series all ranged from 1.24 to 1.81.
Three-way comparisons of regression
models were precluded by disparities in
length distributions. Length range of Mer-
mentau River series (98.36–136.77 mm
EFL) was substantially smaller than that
of Tombigbee River series (46.77–168.00
mm EFL) and the Mississippi River series
(39.5–209.00 mm EFL). Differences in range
of Mermentau River fish was 30% of Tom-
bigbee River series and only 22% of the
Mississippi River series.
Effects of allometric growth on mor-
phometric indices were most pronounced
for fish less than 85 mm EFL (Figure 2A–C).
When fish less than 85 mm EFL were elimi-
nated from analyses, correlations between
the indices and EFL were weak and non-
significant (R2 < 0.08, p > 0.110). Although,
length effects on morphometric indices
were nonsignificant for fish greater than
85 mm EFL, locality effects were appar-
ent. At any given size greater than 85 mm
EFL, fish from different localities separated
vertically from each other, indicating site-
specific variation in morphology.
For large juveniles, variation within
series was low for morphological indices,
but differences among series were signifi-
cant (Table 4). Coefficients of variation for
rostrum length and rostral expansion were
only 5–8%, and for caudal asymmetry only
10–13%. Eye-to-fork length did not dif-
fer among series (F = 1.22, p = 0.3074, df =
2/32), but differences in morphological in-
dices were highly significant (F > 6.35, p <
0.005, df = 2/32). Rostrum length was sig-
nificantly greater in Mississippi River fish,
lower in Tombigbee River fish, and lowest
in Mermentau River fish. Rostral expan-
sion was significantly greater in Mississip-
pi River fish than in fish from the other two
rivers (Table 4). Caudal asymmetry was
greater in fish from the Mermentau River
than in fish from the other two rivers.
Discussion
Three series of fish originating from three
different basins represented three distinc-
tive morphotypes. Mermentau River fish
had short, narrow rostra, widest in the
middle, with minimally expanded tips;
upper caudal lobes were on average 2.5
times longer than lower caudal lobes, and
70% were notched. Tombigbee River fish
had longer, wider rostra, expanded at the
tip; upper caudal lobes were on average
1.6 times longer than lower caudal lobes,
and 90% were notched. Differences in the
leaf-shaped rostra and strongly hetero-
cercal tails of Mermentau River fish and
spatulate or spoon-shaped rostra and more
symmetrical tails of Tombigbee River fish
Table 4. Morphometric indices describing relative rostrum length, mesal rostrum expan-
sion, and caudal asymmetry for larger juvenile paddlesh ($85 mm eye-to-fork length).
Superscripts indicate values for a character that are signicantly different among series
based on analysis of variance and Student-Newman-Keuls test.
Mermentau River Tombigbee River Mississippi River
N = 19 N = 8 N = 8
Eye-to-fork length 118.47 (11.71)a 127.18 (26.06)a 132.97 (37.47)a
Relative rostrum length 598.89 (32.10)c 634.14 (32.26)b 764.79 (45.21)a
Mesal rostrum expansion 1.24 (0.10)b 1.17 (0.08)b 1.33 (0.08)a
Caudal asymmetry 2.52 (0.34)a 1.64 (0.17)b 1.53 (0.17)b
10 h o o v e r e t a l .
are apparent when examining individual
specimens (Figure 3). Mississippi River
fish had the longest and widest rostra, only
moderately expanded at the tip; upper cau-
dal lobes were on average 1.6 times longer
than lower caudal lobes, but 100% were
notched. Lobes were narrower than those
of the other series, giving the caudal fin a
scissor-tail-like appearance in some speci-
mens (Figure 4). Linear models generated
for Mermentau River series may have re-
sulted from length distribution of fish since
there were no small specimens to represent
allometric growth trends in smaller speci-
mens. Data for the three series, however,
clearly represent gradients in surface area
of the rostrum (lowest in Mermentau River
fish, highest in Mississippi River fish) and
asymmetry of the caudal fin (highest in
Mermentau River fish, lowest in Mississippi
River fish). These morphological gradients
should correspond to hydrodynamic gradi-
ents for the three series of fish: reduced drag,
increased lift, and greater thrust (Webb 1975,
1984). Paddlefish from the relatively low ve-
locity Mermentau River would be expected
to have lowest swimming performance, and
fish from the relative high velocity Missis-
sippi River would have the greatest highest
swimming performance.
Morphotypes may reflect responses
to different rearing conditions in the two
hatcheries and the river, genetic variation,
or a combination of both. Sturgeon may
vary morphologically along longitudinal
gradients (Keenlyne et al. 1994) and from
Figure 3. Hatchery-reared paddlesh: upper specimen is from Mermentau River brood-
stock; lower specimen from Tombigbee River broodstock. Fish were nearly identical in size.
Note leaf-shaped rostrum and asymmetrical caudal lobes of upper specimen, spatulate or
spoon-shaped rostrum, and nearly symmetrical caudal lobes of lower specimen. Epicaudal
notch is visible on both sh.
11m o r p h o l o g i c a l v a r i at i o n i n j u v e n i l e p a d d l e f i s h
Figure 4. Field-collected juvenile paddlesh from the Mississippi River. Upper sh is less
than 85 mm eye-to-fork length (EFL); lower sh is greater than 85 mm EFL (penny is
provided for scale). Note asymmetrical caudal lobes in smaller specimen and nearly sym-
metrical caudal lobes in larger specimen.
waters of different temperature (Ruban
and Sokolov 1986). Paddlefish vary in ap-
pearance among hatcheries using different
sources of artesian water, tanks of various
size and flow, and artificial versus natural
lighting (R. Campbell, U.S. Fish and Wild-
life Service, personal communication), and
those conditions would differ substantially
from the warmer, turbid water, and hy-
draulically challenging environments of a
river. Genetic variation and adaptation to
local environments may also be a factor.
Although genetic diversity in paddlefish
is low and geographic clustering is not ap-
parent, population divergences in paddle-
fish have been observed (Epifanio et al.
1996). Mobile Bay paddlefish are segregat-
ed from Mississippi and Pearl River fish,
and population divergences are observed
in some tributaries of the Mississippi River
(e.g., unusual haplotypes and rare nuclear
alleles in fish from the White and Arkan-
sas River system). Genetic differences then
among the Mermentau River (an isolated
drainage near the western edge of the pad-
dlefish range), the main-stem Mississippi
River, and the Tombigbee River (in the Mo-
bile Bay basin) would not be unexpected.
The basins from which paddlefish were
obtained are geographically isolated from
each other (all drain into the Gulf of Mex-
ico) and represent a physiographic gradi-
ent (Table 1). The Mermentau River basin
is confined to lowlands of the Gulf Coastal
Plain and is small and flat; discharge in the
river is low but varies with tide and wind
(Demcheck and Skrobialowski 2003). The
Tombigbee River basin flows through the
Black Belt, Southern Red Hills, and Lime
Hills districts of the Gulf Coastal Plain
12 h o o v e r e t a l .
(Mettee et al. 1996) and is intermediate in
size, elevation, and discharge. The lower
Mississippi River basin lies within the lower
Mississippi embayment, part of the central
Gulf Coastal Plain and is large, elevated in
the upper reaches, with moderate gradient
and high discharge (Baker et al. 1991).
Hydrodynamic characteristics of the
paddlefish rostrum have not been investi-
gated but can be inferred from other spe-
cies (Webb 1975). Longer, broader rostra,
like those of Tombigbee River and espe-
cially Mississippi River paddlefish, can
provide advantages for swimming in faster
flow by improving, respectively, keel and
hydrofoil properties. Long rostra of some
paddlefish could function similarly to the
long rostra of sailfish Istiophorus platypterus
(also known as Histiophorus americanus)
and swordfish Xiphias gladius, increasing
overall streamlining and minimizing sepa-
ration of the boundary layer from poste-
rior surfaces of the body, thereby reducing
wake and drag. Broad rostra of paddlefish
could function in a way similar to that of
pectoral fins of various species, deflecting
flow at an angle to the line of fish move-
ment, thereby generating lift and changes
in vertical position. The tapered base of
the rostrum in all three populations should
result in highest lift forces produced clos-
est to the center of mass and reducing the
tendency to “roll,” which will provide ef-
ficient changes in vertical position.
Overall shape of the paddlefish is most
similar to species classified as specialized
“cruisers,” which rely on sustained swim-
ming and have stiff bodies, slim caudal
peduncles, deep narrow caudal fins, and
a fusiform shape that is deepest anteriorly
narrowing gradually toward the tail (Webb
1984). Cruisers include ichthyosaurs,
sharks, tuna, and shad. The combination
of a relatively small peduncle with a deep
narrow caudal fin, called “narrow neck-
ing,” minimizes side forces in swimming
that would otherwise make the head of a
fish oscillate laterally, while generating and
maintaining substantial thrust. This enables
paddlefish to swim efficiently in a straight
line, facilitating ram-ventilation and filter-
feeding (Burggren and Bemis 1992; Sand-
erson et al. 1994). Significant differences
among the series in the size of the caudal
lobes result in apparent differences in the
caudal aspect ratio, the ratio of span (dis-
tance measured between upper and lower
tips of the caudal fin) to chord (distance
measured from peduncle to edge of caudal
fin). Aspect ratios are conspicuously lower
in the pictured Mermentau River specimen
(approximately 2.0) than in the Tombig-
bee and Mississippi River specimens (ap-
proximately 2.8). Since higher aspect ratios
typically result in greater thrust, reduced
resistance, and greater swim speeds (Webb
and Buffrénil 1990), fish with smaller, wid-
er caudal fins from the Mermentau River
series would be expected to have lower
sustained swimming capabilities than fish
with more falcate fins from the Tombig-
bee River and particularly the Mississippi
River. Although size and shape of rostrum
and caudal fin differed significantly among
the three series of fish, qualitative features
were less variable. Barbels occurred on all
specimens, and epicaudal notches occurred
on the majority of fish.
Presence of barbels on the paddlefish
rostrum was unreported by anatomists
and naturalists until 1903 when the struc-
tures were first described in the scientific
literature (Allis 1903). Regional variation in
presence/absence of barbels was suggested
after examination of a large series of adults
from Lake Pepin, Minnesota indicated that
all specimens possessed them (Wagner
1904; 1908). Failure to report barbels then
and later (e.g., Norman 1948) is attributable
to their inconspicuous appearance and their
small size in adults, which is minute, even
when compared to small barbels of some
cyprinids (Fox 1999). Paddlefish from the
Big Sunflower River, Mississippi, have bar-
13m o r p h o l o g i c a l va r i a t i o n i n j u v e n i l e p a d d l e f i s h
bels less than 5 mm long (J. J. Hoover, un-
published data). They are more conspicu-
ous on smaller fish (<<200 mm TL) when
they are presumably functional (Larimore
1949) and are an important morphological
marker. Allometric growth of the rostrum
by distal elongation, observed previously
(Thompson 1934), was supported by data
for all three series in this study (i.e., positive
relations between eye-to-fork length and
relative distance between the barbels and
the tip of the rostrum). For two of the series,
however, significant allometric growth was
observed for the posterior portion of the
rostrum as well (i.e., Tombigbee and Mis-
sissippi rivers). Proximal elongation of the
rostrum with growth of fish, although not
studied explicitly, is suggested by dispro-
portionate increase in size of cranial case by
anterior growth (Larimore 1949).
Data on occurrence of epicaudal notch-
es are not available in previous accounts of
paddlefish morphology. Epicaudal notch-
es in paddlefish may reduce functional
length of the upper caudal lobe, reducing
disparity in thrust produced by the two
different sized lobes. This notch, common
in sharks, allows the terminal flap to wag,
reducing the propulsive arc of the tail and
eliminating vortices near the tip of the tail,
thereby equalizing thrust and reducing
drag (Martin, undated). Equalizing thrust
from the upper and lower lobes minimiz-
es downward torque typically associated
with heterocercal tails and would require
equal and opposite upward torque for ef-
fective horizontal movement. In addition
to epicaudal notches, sharks have broad
pectoral fins that generate lift to counter-
balance such torque (Gray 1953; Lighthill
1969; Alexander 1992; Ferry and Lauder
1996). Pectoral fins of sturgeon do not gen-
erate lift per se, but by changes in their
vertical position, enable rising and sinking
in the water column by inducing a “pitch-
ing moment” that reorients the body of the
fish (Wilga and Lauder 1999). Paddlefish
may do something similar, especially since
they have small pectoral fins that develop
and function differently from that of other
fishes (Mabee and Noordsy 2004). Slight
changes in pitch of a paddlefish would
generate substantial lift due to the large
size of the rostrum.
Drainage-specific adaptations to hy-
drology, although not documented for
paddlefish, are established for other spe-
cies of far-ranging fishes. The blacktail
shiner Cyprinella venusta, which occurs in
a wide range of habitats throughout much
of the southcentral United States (Baker
et al. 1991; Mettee et al. 1996), also exhib-
its allometric growth during development
(Hood and Heins 2000) and morphological
variation among basins (Gibbs 1957). Pop-
ulations from basins with low mean annual
runoff have smaller females that lay smaller
eggs than populations from streams with
greater mean annual runoff having larger
females laying larger eggs (Heins and Baker
1987). Populations from streams with low
discharge spawn in a wider range of water
velocities than those that came from high
discharge rivers, which spawned preferen-
tially in swifter current (Baker et al. 1994).
Unlike blacktail shiner, covariation in mor-
phology and habitat over wide geographic
ranges remains largely unexplored in the
early life history of paddlefish. Significant
morphological variation in the rostrum
and caudal fin demonstrated in this study
could have effects on juvenile paddlefish
locomotion and behavior, which would in-
fluence paddlefish dispersal, habitat selec-
tion, feeding, and possibly survival. Conse-
quently, fishery managers should consider
morphological compatibility of hatchery
stock with river hydrology when making
decisions regarding stocking.
Acknowledgments
Hatchery-reared fish were provided by
Richard Shelton and DeWayne French,
14 h o o v e r e t a l .
Mammoth Spring National Fish Hatchery
and Ricky Campbell, Private John Allen
National Fish Hatchery. Assistance in the
field was provided by Neil Douglas, Wil-
liam Lancaster, Jack Killgore, Bradley Lew-
is, and Jay Collins. Funding for field stud-
ies of Mississippi River fishes was provided
by the U.S. Army Corps of Engineers: Mis-
sissippi Valley Division, Memphis and St.
Louis Districts. Permission to publish was
provided the Chief of Engineers.
References
Alexander, R. M. 1992. Exploring biomechanics:
animals in motion. Scientific American Li-
brary, New York.
Allis, E. A., Jr. 1903. On certain features of the
lateral canals and cranial bones of Polyodon
folium. Zoologische Jahrbucher 17:659–678.
Baker, J. A., K. J. Killgore, and R. L. Kasul. 1991.
Aquatic habitats and fish communities in
the lower Mississippi River. Reviews in
Aquatic Sciences 3:313–356.
Baker, J. A., K. J. Killgore, and S. A. Foster. 1994.
Population variation in spawning current
speed selection in the blacktail shiner, Cy-
prinella venusta (Pisces: Cyprinidae). Envi-
ronmental Biology of Fishes 39:357–364.
Brehm, A. E., and W. Hacke. 1892. Die fische.
Brehm’s Tierleben, Bibliographisches In-
stitu, Lipzig, Germany.
Burggren, W. W., and W. E. Bemis. 1992. Me-
tabolism and ram ventilation in juvenile
paddlefish, Polyodon spathula (Chondrostei;
Polyodontidae). Physiological Zoology
65:515–539.
Demcheck, D. K., and S. C. Skrobialowski. 2003.
Fipronil and degradation products in the
rice-producing areas of the Mermentau
River basin, Louisiana, February–Septem-
ber 2000. U.S. Geological Survey Fact Sheet
FS-010–03, Baton Rouge, Louisianna.
Epifanio, J. M., J. B. Koppelman, M.A. Nedbal,
and D. P. Phillip. 1996. Geographic variation
of paddlefish allozymes and mitochondrial
DNA. Transactions of the American Fisher-
ies Society 125:546–561.
Ferry, L. A., and G. V. Lauder. 1996. Heterocer-
cal tail function in leopard sharks: a three-
dimensional kinematic analysis of two
models. Journal of Experimental Biology
199:2253–2268.
Fox, H. 1999. Barbels and barbel-like structures
in sub-mammalian vertebrates: a review.
Hydrobiologia 403:153–193.
Gibbs, R. H., Jr. 1957. Cyprinid fishes of the sub-
genus Cyprinella of Notropis III variation and
subspecies of Notropis venustus (Girard).
Tulane Studies in Zoology 5:175–203.
Grande, L., and W. E. Bemis. 1991. Osteology
and phylogenetic relationships of fossil
and recent paddlefishes (Polyodontidae)
with comments on the interrelationships of
Acipenseriformes. Journal of Vertebrate Pa-
laeontology Special Memoir 1(Supplement
to Volume 11):1–121.
Gray, J. 1953. How animals move. Cambridge
University Press, London.
Heins, D. C., and J. A. Baker. 1987. Analysis
of factors associated with intraspecific
variation in propagule size size of a steam-
dwelling fish. Pages 223–231 in W. J. Mat-
thews and D. C. Heins editors. Community
and evolutionary ecology of North Ameri-
can stream fishes. University of Oklahoma
Press, Norman.
Hildebrand, S. F.. and I. L. Towers. 1927. Anno-
tated list of fishes collected in the vicinity
of Greenwood, Mississippi with descrip-
tions of three new species. Bulletin of the
Bureau of the United States Bureau of Fish-
eries 43:105–136.
Hood, C. S., and D. C. Heins. 2000. Ontog-
eny and allometry of body shape in the
blacktail shiner, Cyprinella venusta. Copeia
2000:270–275.
Hoover, J. J., S. G. George, and K. J. Killgore.
2000. Rostrum size of paddlefish (Polyodon
spathula) (Acipenseriformes: Polyodon-
tidae) from the Mississippi delta. Copeia
2000:288–290.
Keenlyne, K. D., C. J. Henry, A. Tews, and P.
Clancey. 1994. Morphometric comparisons
of upper Missouri River sturgeons. Trans-
actions of the American Fisheries Society
123:779–785.
Kirtland, J. P. 1842. Descriptions of the fishes of
the Ohio River and its tributaries. Boston
Journal of Natural History 4:16–26, 231–
308.
Kirtland, J. P. 1845. Descriptions of the fishes
of the Ohio River and its tributaries. Bos-
15m o r p h o l o g i c a l va r i a t i o n i n j u v e n i l e p a d d l e f i s h
ton Journal of Natural History 5: 21–32,
265–344.
Kuhajda, B. R., R. L. Mayden, and R. M. Wood.
2007. Morphologic comparisons of hatch-
ery-reared specimens of Scaphirhynchus
albus, Scaphirhynchus platorynchus, and S.
albus X S. platorynchus hybrids (Acipenseri-
formes: Acipenseridae). Journal of Applied
Ichthyology 23:324–347.
Larimore, R. W. 1949. Changes in the cranial
nerves of the paddlefish, Polyodon spathula,
accompanying development of the ros-
trum. Copeia 1949:204–212.
LeSueur, C. A. 1818. Descriptions of several new
species of North American fishes. Journal
of the Academy of Natural Sciences of Phil-
adelphia 1:222–235, 359–368.
Lighthill, M. J. 1969. Hydromechanics of aquat-
ic animal propulsion. Annual Review of
Fluid Mechanics 1:413–446.
Mabee, P. M., and M. Noordsy. 2004. Develop-
ment of the paired fins in the paddlefish,
Polyodon spathula. Journal of Morphology
261:334–344.
Martin, R.A. Undated. The swimming sea-saw.
Reefquest Centre for Shark Research, World
Wide Web publication. Available: www.
elasmo-research.org/education/topics/p_
caudal_lobe.htm (March 2009)
Mayden, R. L., and B. R. Kuhajda. 1996. System-
atics, taxonomy, and conservation status of
the Alabama sturgeon, Scaphirhynchus sutt-
kusi Williams and Clemmer (Actinoptery-
gii, Acipenseridae). Copeia 1996:241–273.
Mettee, M. F., P. E. O’Neil, and J. M. Pierson. 1996.
Fishes of Alabama and the Mobile basin.
Oxmoor House, Birmingham, Alabama.
Murphy, C. E., J. J. Hoover, S. G. George, and
K. J. Killgore. 2007. Morphometric varia-
tion among river sturgeons (Scaphirhynchus
spp.) of the middle and lower Mississippi
River. Journal of Applied Ichthyology
23:313–323.
Norman, J. R. 1948. A history of fishes. A.A.
Wyn, New York.
Rafinesque, C. S. 1820. Ichthyologia ohiensis, or
natural history of the fishes inhabiting the
Ohio River and its tributary streams, Pre-
ceded by a physical description of the Ohio
and its branches. W.G. Hunt, Lexington,
Kentucky, 1970 reprint by Arno Press, New
York.
Ruban, G. L., and L. J. Sokolov. 1986. Morpho-
logical variability of Siberian sturgeon,
Acipenser baeri, in the Lena River in relation
with its culture in warm waters. Journal of
Ichthyology 26:88–93.
Sanderson, S. L., J. J. Cech, Jr., and A. Y. Cheer.
1994. Paddlefish buccal flow velocity dur-
ing ram suspension feeding and ram ven-
tilation. Journal of Experimental Biology
186:145–156.
Stockard, C. R. 1907. Observations on the natu-
ral history of Polyodon spathula. American
Naturalist 41:753–766.
Thompson, D. C. 1934. Relative growth in Poly-
odon. Illinois Natural History Survey, Bio-
logical Notes No. 2, Urbana.
Wagner, G. 1904. Notes on Polyodon, I. Science
[new series] 19(483):554–555.
Wagner, G. 1908. Notes on the fish fauna of Lake
Pepin. Transactions of the Wisconsin Acad-
emy of Sciences Arts and Letters 16:23–37.
Webb, P. W. 1975. Hydrodynamics and energet-
ics of fish propulsion. Bulletin of the Fish-
eries Research Board of Canada 190.
Webb, P. W. 1984. Form and function in fish
swimming. Scientific American 251:72–82.
Webb, P. W., and V. de Buffrénil. 1990. Locomo-
tion in the biology of large aquatic verte-
brates. Transactions of the American Fish-
eries Society 119:629–641.
Wilga, C. D., and G. V. Lauder. 1999. Locomo-
tion in sturgeon: function of the pecto-
ral fins. Journal of Experimental Biology
202:2413–2432.
