The Cleveland Museum of Natural History
November 2010 Number 57:36–45
PALEOECOLOGY OF DUNKLEOSTEUS TERRELLI
ROBERT K. CARR
Department of Biological Sciences
Ohio University, Athens, Ohio 45701-2979
The Cleveland Member (Late Famennian) fish fauna represents one of the most speciose and
well-collected faunas from the Late Devonian; however, our understanding of the fauna’s
paleoecology is limited. Published interpretations of placoderm paleoecology typically suggest that
most species are obligate bottom-living forms or are tied closely to life on the bottom. Dunkleosteus
terrelli (Newberry, 1873) (Placodermi, Arthrodira), however, was a pelagic organism. This
interpretation is based on an analysis of the nature of the distal Appalachian Basin depositional
habitat and the distribution of this species’ remains within the basin. The species presence as fossils
represents the remains of organisms that lived within the water column in the basin rather than an
allochthonous accumulation of floating carcasses. Thus, the disarticulation seen in specimens of
Dunkleosteus terrelli is a result of local post-mortem flotation and possible scavenging within the
water column. This hypothesis is supported by a Chi-squared statistical analysis for the geographic
distribution of Dunkleosteus terrelli fossil remains. Finally, estimates of dry and wet weights for
Dunkleosteus terrelli suggest (1) an inability to rest on the bottom given the physical propertiesof
the sediments and (2) the presence of a mechanism for static buoyancy, which would account for
post-mortem flotation in a basin with published estimated depths of 30 to 100 meters.
Our understanding of the paleoecology of Late Devonian fish
faunas is limited by the lack of high diversity faunas available for
analysis. The Cleveland Member (Late Famennian) fauna is one
of the best known, possessing 66 species (Hlavin, 1976; Denison,
1979; Williams, 1990; Carr, 1996; Carr and Jackson, 2010)
consisting of sharks (33), placoderms (28), and osteichthyans (5).
Paleoecological interpretations of the fauna historically were
based on analogy to putatively similar taxa from other faunas.
Coccosteus (Miles and Westoll, 1968), for example, is often used
as an analogue to interpret other arthrodires (e.g., Heintz, 1932,
p. 202, used the form of Coccosteus to interpret ‘‘Dinichthys’’ or
Dunkleosteus, a comparison across subordinal-level distinct taxa).
These analogies often involve the comparison of taxa from
dissimilar facies and faunal compositions.
Hlavin (1976) reviewed sedimentological-depositional models
to interpret the Catskill Delta and the distal black shales (the
Cleveland Member of the Ohio Shale along with other regional
shales and their biostratigraphic relationships) and provided an
up-to-date faunal analysis for the Cleveland Member. However,
he did not consider the impact of sedimentological interpretations
on the potential lifestyles of the fishes associated with the
depositional habitat of the Cleveland Member.
Williams (1990) provided the first comprehensive predator-prey
analysis of the Cleveland Member fish fauna. His analysis
centered on chondrichthyans due to their unusually complete
preservation including stomach contents (53 specimens reported
with identifiable gastric contents, Williams, 1990, p. 280).
Evaluation of non-chondrichthyans was limited to a few cases
of associated remains possibly suggesting predator-prey relation-
ships (the presence of a shark spine imbedded in the oral region of
an arthrodire; associated arthrodire and shark remains suggesting
a possible predator-prey interaction; and a palaeoniscid with
shark and arthrodire stomach contents; Williams, 1990, p. 286–
Despite the vast amount of effort dedicated to the sedimento-
logical history of the Appalachian Basin (e.g., Woodrow and
Sevon, 1985) and the long history of vertebrate paleontology in
the region, little work has addressed the paleoecology of the
vertebrate fauna associated with these sediments. The placo-
derms, which represent the most speciose Devonian clade (Carr,
1995) and numerically the largest part of regional collections,
represent the least known clade in terms of their paleoecology.
As seen in the examples above, estimates concerning the
paleoecology of Paleozoic fishes are based on evidence such as
stomach contents, body form, or indirectly on facies analyses.
Little is known concerning placoderm stomach contents, although
Dennis and Miles (1981) and Miles and Westoll (1968, p. 462)
have evaluated stomach contents in Incisoscutum and Coccosteus,
respectively. The interpretation of gnathal-plate morphology is
used to suggest specializations for various feeding strategies such
as durophagy or piscivory (Denison, 1978, p. 17). However, the
feeding strategies of generalized forms or forms without recent
analogues are difficult, if not impossible, to interpret without
additional supporting evidence.
Lindsey (1978) and Webb (1982) used body form to interpret
locomotor styles among extant fishes. Among placoderms this
approach has found limited application being applied primarily to
laterally and dorsoventrally compressed forms. Miles (1969,
p. 129; see also Moy-Thomas and Miles, 1971, p. 175) suggested
a nektonic lifestyle for the laterally compressed brachydeirid
arthrodires, while several authors (Stensio¨, 1963, p. 13; Miles,
1969, p. 129; Moy-Thomas and Miles, 1971, p. 185, 197–198)
interpreted dorsoventrally compressed forms, such as the raylike
rhenanids, as bottom living. Between these extremes, a vast
number of taxa are interpreted generally as bottom living
(Stensio¨, 1963, p. 13) or specialized for ‘‘life just above the
bottom’’ (Moy-Thomas and Miles, 1971, p. 197).
Finally, analysis of the facies in which placoderm fossils are
found has provided an additional source of information, but relies
on the assumption that fishes lived in a habitat on or above the
accumulating sediments. The utility of this approach is difficult to
evaluate when post-mortem transport is involved and the fishes
lived outside the area of deposition.
Of the methods that have been used to determine aspects of
placoderm paleoecology, none individually can provide a
complete picture, and thus far, few attempts have been made to
combine analyses. Ideal preservation, having complete organisms
preserved with soft tissue and stomach contents, is an exceedingly
rare occurrence. Within the Devonian, this ideal is achieved
among the chondrichthyan remains found in carbonate concre-
tions from the Cleveland Member (Williams, 1990). For
placoderms, researchers have retrieved some of the best-preserved
fossil material from the Hunsru¨ckschiefer of Germany (Lower
Emsian) and the acid-prepared concretions from the Gogo
Formation of Western Australia (Frasnian). Even in these
conservation Lagersta¨ tten (sensu Seilacher, 1990) information is
lost. However, a total-evidence approach including, e.g., sedi-
mentology, geochemistry, and taphonomic mechanisms, may
recover sufficient information to reconstruct details of life history
and paleoecology (Tasch, 1965; Elder and Smith, 1988).
The Catskill Delta and its associated foreland basin (Appala-
chian Basin, Figure 1C) provide a unique opportunity to analyze
the paleoecology of fishes in the region. The black-shale facies
found within the distal basin (Figure 1D) represent an anoxic
depositional environment, potentially inhospitable to benthic
organisms or capable only of supporting a low-diversity fauna
(see Discussion below). Fossil fishes found in these shales
represent either fishes living and dying within the region, post-
mortem allochthonous accumulations of floating carcasses, or a
mixture (Brett and Baird, 1993, p. 254).
Historically, most placoderms have been seen as obligate
benthic organisms, which accessed the water column only to feed
(Stensio¨, 1963, p. 13; Moy-Thomas and Miles, 1971, p. 197;
Denison, 1978, p. 17), suggesting that the presence of carcasses
within the low-diversity deep basin represents an accumulation of
remains originating in the established benthic communities of the
aerated shallow regions of the basin. To evaluate this hypothesis
in one species, the current study analyses the distribution of
gnathal elements from Dunkleosteus terrelli (Newberry, 1873)
within the Appalachian Basin. Dunkleosteus terrelli (Figure 2A–
C) is a large arthrodire (4.5–6 m in length) found within the
Appalachian Basin. This species is oval in cross-section and
possesses a well-developed dermal skeleton (Figure 2B, C) with
individual bones up to 5–7.5 cm thick along the lateral and
occipital thickenings of the head shield. The gnathal elements (IG,
ASG, PSG, Figure 2B, C) in this species represent bones that
detach relatively early in the disarticulation process (see Scha¨fer,
1972, p. 49–91, and Elder, 1985, for a discussion of disarticulation
patterns) and would be expected to accumulate close to the site of
death (secondary transport after these remains settle is discussed
below). The geographic distribution of the gnathal elements for
Dunkleosteus terrelli does not support a hypothesis where remains
of this arthrodire represent long-distance floating of carcasses.
Methods and Materials
Dunkleosteus terrelli specimens from the Cleveland Member
(Upper Famennian; postera to late expansa–early praesulcata
conodont zones, Zagger, 1995) of the Ohio Shale analyzed in this
study are housed within the collections of the Cleveland Museum of
Natural History. Data were retrieved from the Museum vertebrate
paleontology catalog using those specimens with entries providing
detailed locality information. The presence or absence of individual
plates indicated in the catalog was confirmed from the actual
specimens. The specimen localities included seven north-south river
or creek basins and the excavation site for Interstate-71 in northern
Ohio near Cleveland (Figure 3). These localities were grouped into
five drainage systems representing five north-south transects in
northern Ohio paralleling the ancient Catskill shoreline (from west-
to-east: 1, Huron and Vermillion Rivers and Chance Creek; 2,
Beaver Creek; 3, Black River; 4, Rocky River; 5, Big Creek and
Interstate-71). These drainage basins expose the black shales
(Cleveland Member) of the distal foreland basin. Recorded
specimens included those with and without gnathal elements and
isolated gnathal elements. The relative abundance of specimens
with and without gnathal elements was evaluated for the localities
A Chi-squared nonparametric analysis for k independent
samples (Equation 1; Siegel and Castellan, 1988) was conducted
to test the null hypothesis (H
) that the proportion of specimens
with and without gnathal elements was statistically the same in
each geographic sample (i.e., there was no geographic trend and
the distribution of fossils was random). The research hypothesis,
based on published interpretations of placoderm paleoecology
(e.g., Stensio¨ , 1963; Denison, 1978), was that the proportions
differ across samples (an increased concentration of gnathal
elements shoreward, suggesting an allochthonous accumulation
of fossils within the distal basin from an eastward shallower
source). A significance level of alpha 5 0.05 was chosen with a
sample size of n 5 199.
5 number of expected cases in the ith row of the jth column
when the null hypothesis is true
i 5 variables within the r 3 k contingency table
j 5 groups within the r 3 k contingency table
5 number of observed cases categorized in the ith row of the
jth column (Siegel and Castellan, 1988, p. 191)
Dry weight represents the weight of a living specimen out of
water, while wet weight is its weight in water. Body mass (wet
weight) is an important factor influencing buoyancy (static or
2010 PALEOECOLOGY OF DUNKLEOSTEUS TERRELLI 37
dynamic) in active swimmers, equally important to benthic forms
when considering an unstable substrate, and an important
consideration in the post-mortem transport of carcasses. To
evaluate locomotor patterns in Dunkleosteus terrelli, estimates of
its wet and dry weight were calculated based on a conservative
comparison to extant Western Atlantic sharks and ten Pacific
examples for three species with limited or no Atlantic weight data.
A total of 59 length-weight examples from 18 species were used
(Bigelow and Schroeder, 1948, all of their reported length-weights
were included in this analysis). Bigelow and Schroeder (1948)
provided a conversion factor for dry to wet weight in sharks (5.5%
for noncontinuous and 2.6% for continuous swimmers) that was
applied to Dunkleosteus terrelli. The lack of bone in sharks gives a
conservative estimate for dry and wet weight in D. terrelli;
however, the purpose of this estimate was to evaluate its impact
on potential settling in bottom sediments and buoyancy (static or
After Allison et al. (1995, p. 98), the root aerobic is used
here to refer to ‘‘modes of life or biofaces,’’ while oxic refers
to oxygen levels within the environment. The specimen-
number prefix CMNH denotes specimens from the Cleveland
Museum of Natural History Department of Vertebrate
The raw data of specimens with or without gnathal plates are
summarized in an r 3 k contingency table (Table 1). A Chi-
squared analysis for the five independent samples gives a Chi-
square of 6.1 with four degrees of freedom. The critical Chi-
square value for alpha 5 0.05 is 9.49, thus the null hypothesis (H
Figure 1. A, paleogeographic reconstruction with paleolatitudes based on Middle Devonian map of Streel et al., 2000. B, detailed
reconstruction of North American and northern European blocks after Ettensohn (1985, fig. 2).The position of Ohio (OH) is indicated
and probable landmasses are shaded (LL5 mountains). C, close-up of the western Appalachian Basin in Ohio and other regional
basins (from Elliot et al., 2000, fig. 2). States (dash, double-dotted lines), Great Lakes’ boundaries (dotted lines), basin depositional
contours (solid thin lines), hypothesized basin boundaries (dashed lines), and regional geographic highs (solid thick lines) are indicated.
D, schematic cross-section of the Ohio Shale (Cleveland and Huron Members) and the laterally equivalent shoreward
38 CARR No. 57
is not rejected (0.10 , P , 0.20). An analysis of three samples
(transects 1, 3, and 5, thus eliminating any potential geographic
overlap of adjacent transects) does not lead to a rejection of the
null hypothesis (X
5 0.64, d.f. 5 2; 0.70 , P , 0.80).
The log of dry weights versus log of body length for 59 extant
sharks (taken from Bigelow and Schroeder, 1948) provided an
estimated dry weight for Dunkleosteus terrelli (Figure 4). Dry
weight for a 4.6 m (15 ft) Dunkleosteus terrelli is estimated at
665.0 kg (1466.3 lbs). A wet weight estimate is 36.6 kg (80.6 lbs;
or using a continuous swimmer conversion factor, 17.3 kg
Given the precautions concerning interpreting paleoecology
based on a fragmentary fossil record (e.g., Elder, 1985; Elder and
Smith, 1988; Allison et al., 1991), it is important to evaluate all
available evidence. Elder and Smith (1988) presented a method to
interpret fish ecology from taphonomic evidence. Using principles
of information theory (Tasch, 1965) in their studies of fish
taphonomy within lakes, they noted that taphonomic processes
provide an informational trade-off. As information about an
organism is lost during the taphonomic process, these processes
themselves provide new information about the organism’s
physical and chemical environment and the post-mortem history
of the organism. Taphonomy provides not only information on
the environment of burial, but if explored fully, the post-mortem
history may reveal details about the source of these organisms and
their life habitat. Unlike a laboratory experiment where individual
variables can be controlled, historical events require a thorough
consideration of all taphonomic variables.
The fossil record for placoderms is meager, limiting evaluation
of habitat and life history. Within the Appalachian Basin,
Dunkleosteus terrelli is numerically the predominant Late
Devonian vertebrate fossil. It is one of the best-described
members of the fauna (Heintz, 1932) and is recognized easily
due to its large size and distinctive osteology relative to the other
large members of the fauna (e.g., the thin plates and unique
morphology of Titanichthys clearly distinguish this taxon from
Dunkleosteus within the Cleveland Member fauna). Dunkleosteus
terrelli specimens range in estimated size from 25 cm to 6 m with
some even smaller specimens questionably attributed to the
species. This size range represents a wide range of age classes.
The available Dunkleosteus terrelli material housed within the
Cleveland Museum of Natural History was collected over many
decades. Peter A. Bungart (Hlavin, 1976) collected, from 1923 to
1946, the bulk of the material in the region of Cuyahoga and
Lorain Counties in northern Ohio (Figure 3), diligently and
meticulously recording all material associated with a single
individual even when collecting continued over several years.
This collection possibly represents the best Paleozoic material
amenable to statistical analysis with the least amount of sampling
Although questions remain concerning the origin of anoxic
basins (Hoover, 1960, p. 31; Degens et al., 1986; Wignall, 1994;
Schieber, 1998), a number of factors are common to the various
hypotheses relating to the development of black shales such as the
Cleveland Member, including: (1) minimal or lack of available
oxygen (Hoover, 1960; Heckel, 1972; Degens and Ross, 1974;
Brumsack and Thurow, 1986; Wignall, 1994; Allison et al., 1995);
(2) presence of hydrogen sulfide representing a potential toxin
(Hoover, 1960, p. 34; Heckel, 1972; Jannasch et al., 1974); (3) an
unstable substrate consisting of fine sediments with high porosity
(Keller, 1974); (4) water depths with low light or below the photic-
zone (additionally, below wave base), and (5) limited or lack of
trace fossils and bioturbation (bioturbation in the Cleveland
Member is reported by Lewis, 1988, p. 25; Schieber, 2003, notes
Figure 2. A, life reconstruction of Dunkleosteus terrelli in pursuit of Cladoselache. Redrawn from Carr (1995, fig. 17). B and C, anterior
and right lateral views of D. terrelli ossified skeleton (composite after Heintz, 1932 and 1968). ASG 5 anterior superognathal, IG 5
inferognathal, PSG 5 posterior superognathal, dashed lines 5 sensory canal grooves.
2010 PALEOECOLOGY OF DUNKLEOSTEUS TERRELLI 39
the unrecognized presence of bioturbation in black shales; see
Brett and Allison, 1998, for a review of paleontological
approaches to interpreting the environment of deposition). The
combined effects of the first four factors above clearly impact the
potential for the establishment of a benthic community. Common
to all models are anoxic pore waters (factor 1). What differs is
whether anoxia extends from the sediment-water interface up into
the water column. Published accounts of the benthic community
associated with black shales (factor 5) range from a complete
absence of benthic organisms (Hoover, 1960, p. 32, 42; Conant
and Swanson, 1961, p. 56–62; Heckel, 1972; Hlavin, 1976; Allison
et al., 1995, p. 100) to low diversity communities capable of
tolerating the physical properties of the depositional environment
(anoxic-dysoxic, stable or intermittent; hydrogen sulfide; depths
potentially below the photic-zone; and unstable substrate;
Hannibal et al., 2005). Detailed geochemical (e.g., degree of
pyritization) and published sedimentological data for the Cleve-
land Member are limited, although recent studies of laterally
equivalent black shales, other non-contemporaneous black shales,
or modern examples provide useful analogues for the Cleveland
Member (hopefully similar studies will be expanded to include the
Modern oxygen-minimum zones that intersect the sediment-
water interface are known from numerous regions of high
primary productivity, for example, the northwest Indian Ocean
(Degens et al., 1986) and the Gulf of California (Brumsack and
Thurow, 1986). However, there are only a few examples of
potential analogues for stagnant foreland basins, including the
Black Sea (Degens and Ross, 1974) and Norwegian fjords
(Brumsack and Thurow, 1986). The Black Sea, a potential
modern analogue to the ancient distal Appalachian Basin, is
characterized by a reduced sedimentation rate (variable within the
basin) and high sediment porosity with the sediments containing
greater than or equal to 71% water by volume (Keller, 1974).
Sedimentation occurs within an anoxic water column with toxic
levels of hydrogen sulfide. The upper few millimeters of sediment
Figure 3. Drainage map of northern Ohio showing the seven river or creek drainages that provided specimens in the current study. Inset
map of Ohio indicates the study area (box). Labels: counties, intermediate sized font; and creeks/rivers, small sized font. Compiled from
Ohio Department of Natural Resources, Principal streams and their drainage areas (map), 1985, and U.S. Geological Survey Kipton,
Ohio, 7.5-minute quadrangle topographic map, 1969.
40 CARR No. 57
are unable to support even the remains of microorganisms. The
next 30–60 cm are reported to have the consistency of a ‘‘slurry’’
(Keller, 1974, p. 333). Shear strength does not improve within the
sediments until a depth of approximately 140 cm (Keller, 1974,
fig. 2). Associated with the anoxic water column and presence of
hydrogen sulfide is the formation and accumulation of iron
sulfides (Berner, 1974; Rozanov et al., 1974). Neither the Black
Sea or fjords (episutural basin of extreme depth or drowned
glacial valley with restricted surface area, respectively) are truly
representative of the large epicontinental basins of the Devonian;
however, the muds deposited in these anoxic environments share
the physical properties of sediments that potentially form black
The Cleveland Member (Nelson, 1955; Hoover, 1960; Mausser,
1982) is typical of black shales with a high organic content
accumulated in an anoxic environment. The shale produces a
petroliferous odor when freshly broken due to this organic
content (Hoover, 1960, p. 23). Its composition of fine sediments,
with little or no carbonates in the bulk of the shale (Mausser,
1982, p. 86; although discontinuous cone-in-cone and carbonate
concretions are present), suggests an unstable high-porosity
substrate at the time of deposition (estimated water content at
the time of deposition for the Cleveland Member is 75–80%, J.
Schieber, personal communication, 2004). The presence of
disseminated pyrite (Mausser, 1982) throughout the formation
indicates the presence of hydrogen sulfide at the time of
In northern Ohio, Upper Devonian Appalachian Basin rocks
are exposed in a number of north-south river basins that drain
into Lake Erie (Figure 3). These river basins represent a series of
north-south transects that roughly parallel the ancient Catskill
Delta (Figure 1B, C) thereby providing a series of proximal-to-
distal samples within the black shales of the Appalachian Basin.
The Cleveland Member represents the distal element of the basin
facies that are laterally equivalent to the eastward slope deposits
of the Chagrin Member (Figure 1D). This shoreward facies
transition is a temporally repeated pattern within the Appalachian
Basin as noted by Baird and Brett (1986) for the Genesee sequence
of New York.
Evaluation of the paleogeographic distribution of Dunkleosteus
terrelli provides two potential hypotheses: (H1) D. terrelli lived,
died, and was preserved within the basin; (H2) D. terrelli lived and
died elsewhere, but was transported into the basin where it was
preserved. The paleogeography of D. terrelli is reflected in the
nature and distribution of its remains within the Cleveland
Member. Hypothesis 1 would suggest the possibility that remains
of this species are either intact (no disarticulation due to post-
Table 1. Contingency tables (r 3 k) for the Dunkleosteus terrelli specimens in the current study. Table for five river/creek drainage
basins represents north-south transects paralleling the ancient Catskill shoreline. The critical X
value for alpha 5 0.05 is 9.49;
therefore, the null hypothesis is not rejected (0.10 , P , 0.20). Table for three samples (drainage basins) eliminates any potential
geographic overlap of adjacent transects. The critical X
value for alpha 5 0.05 is 5.99; therefore, the null hypothesis is not rejected
(0.70 , P , 0.80). Calculations based on Equation 1. Abbreviations: df, degrees of freedom [(r 2 1)(k 2 1)]; E, calculated expected
occurrences; O, observed occurrences; 1, Huron and Vermilion Rivers and Chance Creek; 2, Beaver Creek; 3, Black River; 4, Rocky
River; and 5, Big Creek and Interstate-71.
Contingency table for five drainage basins
Gnathals 1-O 1-E 2-O 2-E 3-O 3-E 4-O 4-E 5-O 5-E Total
With 18 17.48 4 3.35 7 7.44 13 8.18 32 37.56 74
Without 29 29.52 5 5.65 13 12.56 9 13.82 69 63.44 125
Total 47 47.00 9 9.00 20 20.00 22 22.00 101 101.00 N 5 199
/E 18.54 4.78 6.59 20.66 27.26
28.49 4.42 13.45 5.86 75.04
5 S 2 N 6.10
P value 0.1 , P , 0.2
Contingency table for three drainage basins
With 18 15.95 7 6.8 32 34.27 57
Without 29 31.05 13 20.66 69 66.73 111
Total 47 47 20 20 101 101.00 N 5 168
/E 20.32 7.22 29.88
27.08 12.78 71.34
5 S 2 N 0.64
P value 0.7 , P , 0.8
Figure 4. Plot of log dry weight versus log body length for extant
North Atlantic sharks (circles). Estimated dry weight for
Dunkleosteus terrelli of 665.0 kg (1466.3 lbs) for a body length
of 4.6 m (15 ft) is indicated by a star.
2010 PALEOECOLOGY OF DUNKLEOSTEUS TERRELLI 41
mortem transport or through other processes) or disarticulated (if
disarticulated in transport then there should be no pattern
suggesting an eastward source of remains from outside the deep
basin). Hypothesis 2 requires transport of D. terrelli remains into
the deep basin. Associated with the latter hypothesis would be a
pattern reflecting the distance traveled from the source (e.g., more
complete specimens proportionally found closer to the source or
elements that disarticulate early in the taphonomic process found
closer to the source).
The remains of Dunkleosteus terrelli within the Cleveland
Member are typically disarticulated with a variable number of
plates missing (only 5.1% of the evaluated specimens possess over
25% of the dermal and perichondral bones). Elder (1985) and
Elder and Smith (1988) noted three potential sources for the
disarticulation and removal of organic remains in an aquatic
environment: flotation, scavenging, or transport by currents (see
also Allison et al., 1991). Within the distal Appalachian Basin, the
latter two mechanisms have little or no influence on organic
remains once they have settled to the bottom. Proposed
mechanisms to account for anoxia and the deposition of black
shales require an environment below wave base that may be
further restricted from mixing by the presence of a density
stratification (formation of a pycnocline; see Murphy et al., 2000,
and Sageman et al., 2003, for alternative models for the formation
of an anaerobic benthos). An anoxic environment (and presence
of hydrogen sulfide in some models) would make the bottom
unsuitable for macroscopic scavengers; however, that does not
preclude scavenging upon floating carcasses within the water
column. The presence of an unsuitable bottom habitat is
supported in part by the limited or lack of bioturbation and
trace fossils. Additionally there is little evidence for the presence
of a diverse macroscopic infauna or epifauna, and no evidence for
the presence of epibionts encrusting bones. Fossil sharks within
the Cleveland Member often are found fully articulated with soft
tissues preserved (Williams, 1990). These fossils have been
reported from carbonate concretion zones in the shales of Big
Creek and Interstate-71. This represents a sampling bias in that
concretions are easily seen in profile, in contrast to flattened
isolated specimens. Shark fossils are found from outside the
concretion zones in westward deposits either associated with
cone-in-cone deposits or isolated within the shale. The skeletons
of these fishes consist of perichondrally ossified prismatic
cartilage that is highly susceptible to mechanical damage and is
destroyed easily by scavengers or currents.
The fine grain size of Cleveland Member sediments further
suggests a relatively low energy depositional environment
(Nelson, 1955; Mausser, 1982) and the absence of currents
sufficient to move the relatively large bones of Dunkleosteus
terrelli. Additionally, there are no sedimentary structures (bed-
ding features) suggesting the presence of higher energy bottom
currents within the Cleveland Member. Sediment winnowing does
occur, demonstrated by the presence of several isolated lag
deposits (invertebrate thanatocoenoses, Hlavin, 1976; silty lags, J.
Schieber, personal communication, 2004); however, winnowing
tends to accumulate bone and pelagic invertebrates rather than
remove them. The large and relatively heavy Dunkleosteus terrelli
bones would preferentially remain within these deposits.
Flotation remains as the most plausible mechanism to account
for the disarticulated remains of Dunkleosteus terrelli. Flotation
could occur in fishes living within the basin or it may represent a
mechanism for the transport of organisms into the basin.
Common to the patterns of disarticulation in both aquatic
(Elder, 1985; Scha¨fer, 1972) and terrestrial (Hill, 1979; Weigelt,
1989) organisms is the loss of elements near sites of access by
scavengers (microscopic and macroscopic). The oropharyngeal
cavity represents such a site with scavenging activity in this region
resulting in the disarticulation and potential loss of gnathal
elements. In teleosts and arthrodires (Figure 2B, C) both upper
and lower gnathal elements, which lack ossified connections to the
axial skeleton, are lost. In contrast, within mammals where the
maxillae and premaxillae are connected intimately with the skull,
the dentaries alone are detached from the skull. Work by Elder
(1985) and others (Scha¨fer, 1972; Hill, 1979; Smith and Elder,
1985; Elder and Smith, 1988) suggests that gnathal elements
(dentaries only in mammals) in particular are lost early in the
disarticulation process, thus presence of specimens with gnathal
elements suggests a relatively short time of exposure to pre-burial
taphonomic processes (including both transport and scavenging).
In the case of Dunkleosteus terrelli within the Appalachian Basin,
the presence of isolated gnathal plates suggests that floating
carcasses may have dropped elements relatively close to the
original life habitat.
Elder (1985), Smith and Elder (1985), and Elder and Smith
(1988) documented the importance of water depth and temper-
ature to the process of flotation. The physical principles outlined
in the gas laws of Charles and Gay-Lusac (k 5 V/T) and Boyle (k
5 PV) clearly delineate limitations for the potential of an
organism to float upon bacterial decomposition within finite
ranges of water depth and temperature. Thus, the volume of
accumulating decomposition gases within the tissues or body
cavities to induce flotation will increase with warming (a
latitudinal or climate factor) and decrease with depth (increased
pressure). Elder (1985) suggested a limiting depth for flotation in
a number of teleosts although her experimental work did not
consider larger species or variable densities. She suggested that
settling below a depth of 10 m even within tropical waters would
severely limit or prevent flotation. In contrast, Allison et al. (1991)
noted the calculated potential of flotation in whales up to a depth
of 1200 m. Unlike the teleosts studied by Elder (1985), whale
specimens may be nearly neutrally buoyant at the time of their
settling to the bottom (some are positively buoyant and float at
the time of death, representing an inherent static buoyancy,
Allison et al., 1991).
The Appalachian Basin in the Late Devonian was located in the
subtropics (30–35u S), suggesting moderate to warm water
temperatures (17–18u C) based on oxygen isotopic data from
brachiopods (Streel et al., 2000, fig. 30, p. 154). Several authors
dispute the water depth within the open Appalachian Basin;
however, all estimates are well within the computed depths of
Allison et al. (1991) that are necessary for flotation (61.0–91.4 m,
Nelson (1955); 30.5–45.7 m, Hlavin (1976); 50–100 m, Brett et al.
(2003); up to 100 m using the depth limits, noted by Ferguson
(1963), for Lingula found within the Cleveland Member; contrast
these estimates with Schieber’s (1998) estimate of 10–20 m for the
Chattanooga Shale of Tennessee, although, the Chattanooga
Shale is characterized by a number of lag deposits and erosional
The presence of Dunkleosteus terrelli remains in the black shales
of the Appalachian Basin can be explained either as a post-
mortem accumulation of organisms that lived within the basin or
a rain of parts falling from carcasses transported via flotation
from aerated regions. The results of a Chi-squared statistical
analysis of these fossils did not support the hypothesis that D.
42 CARR No. 57
terrelli was restricted to the shallow aerated eastern part of the
basin and that its presence in the distal sediments was the result of
carcasses floating into the basin (Table 1; X
5 6.1, d.f. 5 4, 0.10
, P , 0.20). In an additional analysis of non-adjacent transects
(nos. 1, 3, and 5), the null hypothesis was not rejected (X
d.f. 5 2, 0.70 , P , 0.80). Non-rejection of the null hypothesis is
consistent with a random distribution of D. terrelli within the
basin, thus it is more likely that this species was a pelagic form
(living and dying within the basin) and not an obligate bottom
dweller (restricted to the eastern aerated shallower benthos, e.g.,
the basin slope represented by the Chagrin Member or more
shoreward regions, Figure 1D).
Given our knowledge of the bottom habitat associated with
black-shale formation, Dunkleosteus terrelli would have had to be
a continuous swimmer. The paleontological, sedimentological,
and geochemical evidence all suggest that the bottom environ-
ment was inhospitable; however, even if fishes could reach the
bottom within the basin, the stability of the substrate to support
the weight of these organisms would have proven to be a problem.
At present there are few estimates for the size or weight of
Dunkleosteus terrelli (Heintz, 1932). The large amount of bone
and concomitantly high body mass has led several researchers to
conclude that D. terrelli must be a benthic form (e.g., Denison,
1978; although rejected by later authors, Heintz, 1932, considered
Dunkleosteus to be an active swimmer and predator). Comparison
of Dunkleosteus terrelli (using an adult length of 4.6 m (15 ft) to
modern sharks (Figure 4) suggests a dry weight of approximately
665.0 kg (1466.3 lbs). Bigelow and Schroeder (1948) suggested
that wet weight is 5.5% of dry weight in extant sharks (2.6% for
free or continuous swimming forms). Using this estimate, the wet
weight for D. terrelli can be calculated to be 36.6 kg (80.6 lbs; or
using a free swimming estimate, 17.3 kg or 38.1 lbs). This
represents a conservative estimate due to the lack of bone in
sharks and the conversion to wet weight being based on
organisms possessing lipids for static buoyancy. A wet weight of
36.6 kg (or 17.3 kg for continuous swimmers) would generate
sufficient shear forces to permit the settling of D. terrelli into the
substrate exposing it to the toxic properties of the sediments and
potentially clogging the gills with fine silt (refer to Keller, 1974,
p. 333–334, for estimates of shear strength in black muds). If D.
terrelli lived within the basin it can be assumed that it did not rest
on the bottom, but swam continuously.
A second point bearing upon the interpretation of placoderm
lifestyles is their mode of locomotion. Placoderms apparently
possessed a low profile and poorly supported heterocercal tail
and have been interpreted to have swum using a sine-wave
undulation of the body (Thomson, 1971; anguilliform locomo-
tion of Lindsey, 1978). They have been compared to extant
macrurid or chimaerid fishes (Stensio¨, 1963, p. 13), which
demonstrate a bottom-dwelling lifestyle. Although anguilliform
locomotion may be considered a relatively inefficient form of
locomotion (e.g., relative velocity between anguilliform and
other forms of locomotion) it does not mean that anguilliform
swimmers are not effective prolonged or continuous swimmers.
One needs only to look at the migratory patterns of extant eels
(Anguilla anguilla), which migrate between North America and
Europe (McDowall, 1988) to recognize this point. Although
placoderms apparently never achieved some of the advanced
forms of locomotion seen in modern teleosts (e.g., carangiform
and thunniform locomotion, Lindsey, 1978), they did develop a
number of locomotor adaptations associated with increased lift
and reduced drag (Carr, 1995).
The taphonomic evidence did not support an interpretation of
Dunkleosteus terrelli as an obligate benthic organism. A Chi-
squared statistical analysis of the distribution for the remains of
this species failed to support a restricted bottom-living hypothesis.
Dunkleosteus terrelli was a free-swimming species living within the
Appalachian Basin, which may help to explain its wide North
American distribution (California to eastern United States; if
synonymised with D. marsaisi (Lehman, 1954) from Morocco
(Ru¨cklin, 2002) the range would extend via the Rheic Ocean to
east of the Old Red Sandstone Continent, Figure 1A). This
interpretation is consistent with the analysis of Carr (1995), which
noted the development of locomotor specializations within
pachyosteomorph arthrodires and was further supported by the
preservation of relatively complete specimens or the distribution
of elements lost early in the disarticulation process throughout the
Potential objections to viewing placoderms as free-swimming
organisms have included their possession of heavy body armor
and an anguilliform form of locomotion. The distribution of
Dunkleosteus terrelli remains in the Appalachian Basin is
consistent with disarticulation due to localized post-mortem
flotation. The presence of post-mortem flotation in a basin with
published depth estimates ranging from 30 to 100 meters suggests
an organism with some level of additional static buoyancy beyond
inherent tissue buoyancy (refer to Allison et al., 1991). An
interpretation of anguilliform locomotion does not necessarily
imply an ineffective form of locomotion. It appears that an
interpretation of bottom life is a consequence of the researcher’s
choice of extant analogue rather than any necessary correlation
between lifestyle and form of locomotion.
The implications of this study are three-fold and form the
basis of continuing work on the Cleveland Member fauna. Life
within the basin raises questions concerning the: (1) reproductive
strategy; (2) life history; and (3) static buoyancy in Dunkleosteus
terrelli. An inhospitable bottom limits potential nesting sites. It is
not possible at this time to determine whether this species was
viviparous or oviparous; although, the presence of putative egg
cases within the Cleveland Museum of Natural History
collections (CMNH 8133–8136, 9461) raises some interesting
questions. Further work is needed to provide information on the
distribution of age classes for D. terrelli within the Appalachian
Basin. Finally, the presence of disarticulation associated with
flotation raises questions concerning static buoyancy in D.
terrelli. Elder (1985) has pointed out the physical limits (depth
and temperature) associated with flotation; however, Allison et
al. (1991) have demonstrated that given sufficient static
buoyancy, flotation may occur up to depths of 1200 m. The
presence of air sacs in placoderms has been questioned (Denison,
1941, recognized the presence of air sacs in Bothriolepis, but did
not consider them to be a feature of placoderms in general,
although this interpretation has never been confirmed in other
specimens of Bothriolepis or in any other placoderm taxa (D.
Goujet, personal communication, 2004); Gardiner, 1984, con-
sidered air sacs to be a derived feature within Osteichthyes).
However, the predominance of disarticulation associated with
floating in D. terrelli strongly suggests the presence of a static
buoyancy mechanism within this species (if not an air sac then
potentially lipids as seen in chondrichthyans). Further work on
taphonomic patterns and flotation within D. terrelli may help to
shed light on these questions.
2010 PALEOECOLOGY OF DUNKLEOSTEUS TERRELLI 43
I would like to thank D. Goujet and G. R. Smith for their early
reviews and the latter for discussions on taphonomy, R. Cox and
D. Dunn for our many discussions, and C. Brett, J. Schieber, and
D. Goujet for review of the final manuscript. Posthumous thanks
and regards are expressed to M. E. Williams for our many
discussions concerning the Cleveland Member fauna over a
twenty-year period. Finally, I want to additionally thank D.
Fisher, C. Gans, and P. D. Gingerich who provided encourage-
ment and reviews on an early version of this work submitted in
partial fulfillment of the requirements for a Doctor of Philosophy
in Geological Sciences at The University of Michigan.
Allison, P. A., C. R. Smith, H. Kukert, J. W. Deming, and B. A.
Bennett. 1991. Deep-water taphonomy of vertebrate carcasses:
a whale skeleton in the bathyal Santa Catalina Basin.
Allison, P. A., P. B. Wignall, and C. E. Brett. 1995. Palaeo-oxygen-
ation: effects and recognition, p. 97–112. In D. W. J. Bosence and
P. A. Allison (eds.), Marine Palaeoenvironmental Analysis from
Fossils. Geological Society Special Publication No. 83.
Baird, G. C., and C. E. Brett. 1986. Erosion on an anaerobic
seafloor: significance of reworked pyrite deposits from the
Devonian of New York State. Palaeogeography, Palaeoclima-
tology, Palaeoecology, 57:157–193.
Berner, R. A. 1974. Iron sulfides in Pleistocene deep Black Sea
sediments and their paleo-oceanographic significance, p. 524–
531. In E. T. Degens and D. A. Ross (eds.), The Black Sea—
Geology, Chemistry, and Biology. American Association of
Petroleum Geologists Memoir 20.
Bigelow, H. B., and W. C. Schroeder. 1948. Sharks, p. 59–546. In
J. Tee-Van, C. M. Breder, S. F. Hildebrand, A. E. Parr, and
W. C. Schroeder (eds.), Fishes of the Western North Atlantic,
Lancelets, Cyclostomes, and Sharks. Sears Foundation for
Marine Research, Memoir 1, Part 1. Yale University Press,
Brett, C. E., and P. A. Allison. 1998. Paleontological approaches
to the environmental interpretation of marine mudrocks,
p. 302–349. In J. Schieber, W. Zimmerle, and P. S. Sethi
(eds.), Shales and Mudstones, Vol. 1. E. Schweizebart’sche
Brett, C. E., and G. C. Baird. 1993. Taphonomic approaches to
temporal resolution in stratigraphy: examples from Paleozoic
marine mudrocks, p. 250–274. In S. M. Kidwell and A. K.
Behrensmeyer (eds.), Taphonomic Approaches to Time
Resolution in Fossil Assemblages. Paleontological Society
Short Courses in Paleontology, No. 6.
Brett, C. E., A. H. Turner, P. I. McLaughlin, D. J. Over, G. W.
Storrs, and G. C. Baird. 2003. Middle-Upper Devonian
(Givetian-Famennian) bone/conodont beds from central Ken-
tucky, USA: reworking and event condensation in the distal
Acadian foreland basin. Courier Forschunginstitut Sencken-
Brumsack, H.-J., and J. Thurow. 1986. The geochemical facies of
black shales from the Cenomanian/Turonian boundary event
(CTBE), p. 247–265. In E. T. Degens, P. A. Meyers, and S. C.
Brassell (eds.), Biogeochemistry of Black Shales. Im Selbst-
verlag des Geologisch-Pala¨ontologischen Institutes der Uni-
versita¨t Hamburg, No. 60, Hamburg.
Carr, R. K. 1995. Placoderm diversity and evolution. VIIth
International Symposium: Studies on Early Vertebrates.
Bulletin du Muse´um d’Histoire Naturelle, 17:85–125.
Carr, R. K. 1996. Stenosteus angustopectus sp. nov. from the
Cleveland Shale (Famennian) of northern Ohio with a review
of selenosteid (Placodermi) systematics. Kirtlandia, 49:
Carr, R. K., and G. L. Jackson. 2010. The vertebrate fauna of the
Cleveland Member (Famennian) of the Ohio Shale, Chapter 5.
In J. T. Hannibal (ed.), Guide to the Geology and Paleontol-
ogy of the Cleveland Member of the Ohio Shale. Ohio
Geological Survey Guidebook 22.
Conant, L. C., and V. E. Swanson. 1961. Chattanooga Shale and
related rocks of central Tennessee and nearby areas. U.S.
Geological Survey Professional Paper 357. 91 p.
Degens, E. T., P. A. Meyers, and S. C. Brassell (eds.). 1986.
Biogeochemistry of Black Shales. Im Selbstverlag des
Geologisch-Pala¨ontologischen Institutes der Universita¨ t Ham-
burg, Hamburg. 421 p.
Degens, E. T., and D. A. Ross (eds.). 1974. The Black Sea—
Geology, Chemistry, and Biology. American Association of
Petroleum Geologists Memoir 20. 663 p.
Denison, R. H. 1941. The soft anatomy of Bothriolepis. Journal of
Denison, R. H. 1978. Handbook of Paleoichthyology, Vol. 2,
Placodermi. Gustav Fischer, Stuttgart. 128 p.
Denison, R. H. 1979. Handbook of Paleoichthyology, Vol. 5,
Acanthodii. Gustav Fischer, Stuttgart. 62 p.
Dennis, K. D., and R. S. Miles. 1981. A pachyosteomorph
arthrodire from Gogo, Western Australia. Zoological Journal
of the Linnean Society, 73:213–258.
Elder, R. 1985. Principles of aquatic taphonomy with examples
from the fossil record. Unpublished Ph.D. dissertation,
University of Michigan, Ann Arbor. 351 p.
Elder, R. L., and G. R. Smith. 1988. Fish taphonomy and
environmental inference in paleolimnology. Palaeogeography,
Palaeoclimatology, Palaeoecology, 62:577–592.
Elliot, D. K., H. G. Johnson, R. Cloutier, R. K. Carr, and E. B.
Daeschler. 2000. Middle and Late Devonian vertebrates of the
western Old Red Sandstone Continent. Courier Forschungin-
stitut Senckenberg, 223:291–308.
Ettensohn, R. 1985. Controls on development of Catskill Delta
complex basin-facies, p. 65–77. In W. L. Woodrow and W. D.
Sevon (eds.), The Catskill Delta. Geological Society of
America Special Paper 201.
Ferguson, L. 1963. The paleoecology of Lingula squamiformes
Phillips during a Scottish Mississippian marine transgression.
Journal of Paleontology, 37:669–681.
Gardiner, B. G. 1984. The relationships of the palaeoniscid fishes,
a review based on new specimens of Mimia and Moythomasia
from the Upper Devonian of Western Australia. Bulletin
British Museum (Natural History), 37:173–428.
Hannibal, J. T., K. L. Gallup, L. E. Hulslander, E. F. Kennedy,
C. A. Moglovkin, B. J. Olsen, M. V. Prarat, and J. A. Von
Glahn. 2005. The black shale paradox: presumed pelagic,
pseudoplanktonic, and burrowing and other benthic organ-
isms preserved in the Cleveland Shale (Famennian), a ‘‘deep
water’’ black shale deposited in anoxic conditions. Geological
Society of America Abstracts with Programs, 37(2):4.
Heckel, P. H. 1972. Recognition of ancient shallow marine
environments, p. 226–286. In J. K. Rigby and W. K. Hamblin
(eds.), Recognition of Ancient Sedimentary Environments.
Society of Economic Paleontologists and Mineralogists Special
44 CARR No. 57
Heintz, A. 1932. The structure of Dinichthys: a contribution to
our knowledge of the Arthrodira. Bashford Dean Memorial
Volume Archaic Fishes, 4:115–224.
Hill, A. 1979. Disarticulation and scattering of mammal skeletons.
Hlavin, W. J. 1976. Biostratigraphy of the Late Devonian black
shales on the cratonal margin of the Appalachian geosyncline.
Unpublished Ph.D. dissertation, Boston University. 211 p.
Hoover, K. V. 1960. Devonian–Mississippian shale sequence in
Ohio. Ohio Division of Geological Survey Information Circular
27. 154 p.
Jannasch, H. W., H. G. Truper, and J. H. Tuttle. 1974. Microbial
sulfur cycle in the Black Sea, p. 419–425. In E. T. Degens and
D. A. Ross (eds.), The Black Sea—Geology, Chemistry, and
Biology. American Association of Petroleum Geologists
Keller, G. H. 1974. Mass physical properties of some western
Black Sea sediments, p. 332–337. In E. T. Degens and D. A.
Ross (eds.), The Black Sea—Geology, Chemistry, and Biology.
American Association of Petroleum Geologists Memoir 20.
Lehman, J. P. 1954. Les Arthrodires du Maroc me´ridional
International Geological Congress, Algiers
Lewis, T. L. 1988. Late Devonian and Early Mississippian distal
basin-margin sedimentation of northern Ohio. Ohio Journal of
Lindsey, C. C. 1978. Form, function, and locomotory habits in
fish, p. 1–100. In W. S. Hoar and D. J. Randall (eds.), Fish
Physiology. Academic Press, New York.
Mausser, H. F. 1982. Stratigraphy and sedimentology of the
Cleveland Shale (Devonian) in northeast Ohio. Unpublished
master’s thesis, Case Western Reserve University, Cleveland.
McDowall, R. M. 1988. Diadromy in Fishes: Migrations between
Freshwater and Marine Environments. Timber Press, Port-
land. 308 p.
Miles, R. S. 1969. Features of placoderm diversification and the
evolution of the arthrodire feeding mechanism. Transactions
of the Royal Society of Edinburgh, 68:123–170.
Miles, R. S., and T. S. Westoll. 1968. The placoderm fish
Coccosteus cuspidatus Miller ex Agassiz from the Middle Old
Red Sandstone of Scotland. Part 1, descriptive morphology.
Transactions of the Royal Society of Edinburgh, 67:373–476.
Moy-Thomas, J. A., and R. S. Miles. 1971. Paleozoic Fishes. W.
B. Saunders Company, Philadelphia. 259 p.
Murphy, A. E., B. B. Sageman, D. J. Hollander, T. W. Lyons,
and C. E. Brett. 2000. Black shale deposition and faunal
overturn in the Devonian Appalachian Basin: clastic starva-
tion, seasonal water-column mixing, and efficient biolimiting
nutrient recycling. Paleoceanography, 15(3):280–291.
Nelson, B. 1955. Mineralogy and stratigraphy of the pre-Berea
sedimentary rocks exposed in northern Ohio. Unpublished
Ph.D. dissertation, University of Illinois, Urbana-Champaign.
Newberry, J. S. 1873. Descriptions of fossil fishes. Report of the
Geological Survey of Ohio. Vol. 1, Pt. 2, Palaeontology,
Rozanov, A. G., I. I. Volkov, and T. A. Yagodinskaya. 1974.
Forms of iron in surface layer of Black Sea sediments, p. 532–
541. In E. T. Degens and D. A. Ross (eds.), The Black Sea—
Geology, Chemistry, and Biology. American Association of
Petroleum Geologists Memoir 20.
Ru¨cklin, M. 2002. New finds of placoderms from the Late
Devonian of Morocco. 7
European Workshop on Vertebrate
Palaeontology, Sibiu, Romania, Abstracts. 31 p.
Sageman, B. B., A. E. Murphy, J. P. Werne, C. A. Ver Straeten,
D. J. Hollander, and T. W. Lyons. 2003. A tale of shales: the
relative roles of production, decomposition, and dilution in the
accumulation of organic-rich strata, Middle-Upper Devonian,
Appalachian basin. Chemical Geology, 195:229–273.
Scha¨fer, W. 1972. Ecology and Palaeoecology of Marine Environ-
ments. Translated by I. Oertel. Oliver and Boyd, Edinburgh.
Schieber, J. 1998. Sedimentary features indicating erosion,
condensation, and hiatuses in the Chattanooga Shale of
central Tennessee: relevance for sedimentary and stratigraphic
evolution, p. 187–215. In J. Schieber, W. Zimmerle, and P. S.
Sethi (eds.), Shales and Mudstones (Vol. 1). E. Schweize-
bart’sche Verlagsbuchhandlung, Stuttgart.
Schieber, J. 2003. Simple gifts and hidden treasures—implications
of finding bioturbation and erosion surfaces in black shales.
The Sedimentary Record, 1:4–8.
Seilacher, A. 1990. Taphonomy of Fossil-Lagersta¨tten, p. 266–270.
In D. E. G. Briggs and P. R. Crowther (eds.), Palaeobiology: A
Synthesis. Blackwell Scientific Publications, Oxford.
Siegel, S., and N. J. Castellan, Jr. 1988. Nonparametric Statistics
for the Behavioral Sciences. McGraw-Hill Book Company,
New York. 399 p.
Smith, G. R., and R. Elder. 1985. Environmental interpretation of
burial and preservation of Clarkia fishes, p. 85–93. In C. J.
Smiley (ed.), Late Cenozoic History of the Pacific Northwest.
Pacific Division of the American Association for the Advance-
ment of Science, California Academy of Science, San Francisco.
Stensio¨ , E. A. 1963. Anatomical studies on the arthrodiran head.
Part I. Preface, geological and geographical distribution, the
organization of the arthrodires, the anatomy of the head in the
Dolichothoraci, Coccosteomorphi, and Pachyosteomorphi.
Taxonomic appendix. Kungliga Svenska Vetenskapsakade-
miens, Handlingar, 9(2):1–419.
Streel, M., M. V. Caputo, S. Loboziak, and J. H. G. Melo. 2000.
Late Frasnian-Famennian climates based on palynomorph
analyses and the question of the Late Devonian glaciations.
Earth-Science Reviews, 52:121–173.
Tasch, P. 1965. Communications theory and the fossil record of
invertebrates. Transactions Kansas Academy of Science, 68:
Thomson, K. S. 1971. The adaptation and evolution of early
fishes. Quarterly Review of Biology, 46:139–166.
Webb, P. W. 1982. Locomotor patterns in the evolution of
actinopterygian fishes. American Zoologist, 22:329–342.
Weigelt, J. 1989. Recent Vertebrate Carcasses and their Paleobi-
ological Implications. Translated by J. Schaefer. University of
Chicago Press, Chicago. 188 p.
Wignall, P. B. 1994. Black Shales. Clarendon Press, Oxford.
Williams, M. E. 1990. Feeding behavior in Cleveland Shale fishes,
p. 273–287. In A. J. Boucot, Evolutionary Paleobiology of
Behavior and Coevolution. Elsevier, Amsterdam.
Woodrow, D. L., and W. D. Sevon (eds.). 1985. The Catskill
Delta. Geological Society of America Special Paper 201. 246 p.
Zagger, G. W. 1995. Conodont biostratigraphy and sedimentology
of the latest Devonian of northern Ohio. Unpublished master’s
thesis, Case Western Reserve University, Cleveland. 112 p.
2010 PALEOECOLOGY OF DUNKLEOSTEUS TERRELLI 45