Pacific Spiny Lumpsucker armor—Development, damage, and
defense in the intertidal
Eleanor C. Woodruff
| Jonathan M. Huie
| Adam P. Summers
Karly E. Cohen
Department of Biology, Carleton College,
Northfield, Minnesota, USA
Department of Biology, George Washington
University, Washington, District of
Friday Harbor Laboratories, University of
Washington, Friday Harbor, Washington, USA
Department of Biology, University of
Washington, Seattle, Washington, USA
Karly E. Cohen, 620 University Rd, Friday
Harbor Labs, Friday Harbor, WA 98250, USA.
National Science Foundation, Grant/Award
Numbers: DBI-1759637, DBI-1852096; NSF
graduate research fellowship, Grant/Award
Number: DGE-1746914; Eugster Endowed
Student Research and Internship Fund
Predation, combat, and the slings and arrows of an abrasive and high impact environ-
ment, represent just some of the biotic and abiotic stressors that fishes are armored
against. The Pacific Spiny Lumpsucker (Eumicrotremus orbis) found in the subtidal of the
Northern Pacific Ocean is a rotund fish covered with epidermal, cone-shaped, enamel
odontodes. The Lumpsucker is a poor swimmer in the wave swept rocky intertidal, and
this armor may be a lightweight solution to the problem of collisions with abiotic obsta-
cles. We use micro-CT and scanning electron microscopy to reveal the morphology and
ontogeny of the armor, and to quantify the amount of mineralization relative to the
endoskeleton. The non-overlapping odontodes are organized into eight rows—six rows
on the body, one row surrounding the eye, and one row underneath the chin. Odontodes
start as a single, hooked cone; and they grow by the addition of cusps that accrete into a
spiral. The mineral investment in armor compared to skeleton increases over ontogeny.
Damage to the armor occurs both through passive abrasion and breakage from impact;
and there is no evidence of replacement, or repair of damaged odontodes.
allometry, growth, odontodes, ontogeny, protection
including defense, offense, display, restriction of movement, and camou-
flage (Buser et al., 2019; Kawai, 2019; Kolmann, Peixoto, et al., 2020;
Kolmann, Urban, & Summers, 2020; Kruppert et al., 2020; Lowe
et al., 2021; Porter et al., 2013; Reichert & Steffen, 2010; Sherman
et al., 2017; Song et al., 2011; Yang et al., 2015). The morphology of armor,
whether it is thick or thin, the presence or absence of sculpturing, and the
material it is made of, can reveal function. A close examination of wear and
breakage can tell some of the story of how the armor is used (Kruppert
et al., 2020). Armor that has been abraded away is defending against a dif-
ferent assault than armor that is broken or deeply scratched. Armor repre-
sents an investment in mineral that reflects an animal's ecology and
natural history; a relationship that is only revealed by a paired investigation
into gross morphology and a fine scale examination of the surface plates.
Gross morphology of armor layout can hold information about its
utility as a defensive structure, as well as the penalties imposed by the
armor on mobility or maneuverability. For example, the lightweight
armor plates, with minimal overlap, seen in sticklebacks offer less pro-
tection than the fully imbricate armor of seahorses and pipefishes
(Browning, 2012; Porter et al., 2013; Song et al., 2010; Vamosi &
Schluter, 2004; Webb et al., 1992). However, stickleback armor
imposes a far smaller penalty on both swimming speed and turning
ability. When the armor fully encases the fish there are still informa-
tive nuances—the fully fused armor of a boxfish restricts body undula-
tion, while the rail and channel system that connects the overlapping
plates of a Northern Spearnose Poacher allows these fish to retain the
ability to c-start (Kolmann, Peixoto, et al., 2020; Yang et al., 2015).
Armor, whether made up of dermal bone or epidermally derived
odontodes such as ganoid scales, placoid scales, or denticles, bear wit-
ness to the types of damage they are subjected to. The surface will
Received: 17 September 2021 Revised: 1 December 2021 Accepted: 5 December 2021
Journal of Morphology. 2022;1–10. wileyonlinelibrary.com/journal/jmor © 2022 Wiley Periodicals LLC. 1
wear with fine striations when colliding with an abrasive environment,
but encounters with teeth or sharp substrates lead to wider scratches,
and powerful impacts leave broken edges (Kruppert et al., 2020).
Spalling, or the removal of a section of surface, is also evidence of a
shearing impact. These damage modes can be quantified across or
among individuals and species by a close examination of the surface,
either by scanning electron microscopy (SEM) or CT scanning
(Kolmann, Peixoto, et al., 2020; Kolmann, Urban, & Summers, 2020;
Kruppert et al., 2020). Intact armor subjected to known damage can
then be visualized for ground truthing the results in field caught speci-
mens. Looking at damage patterns and the frequency of wear is some-
times the best way to see what the plates are made of. Broken edges
may show the characteristic columnar morphology of enamel and
fibrous or tubular areas of dentine. Also, armor that lacks damage sug-
gests that the structures may be for display or may act as a substrate
for epibionts that will conceal the fish from predators.
Pacific Spiny Lumpsuckers (Eumicrotremus orbis) are small, charis-
matic, and densely armored fish of the North Pacific Ocean, and their
armor may serve all, or most, of the previously mentioned functions
(Arita, 1969; Berge & Nahrgang, 2013). They are found in the heavily
fouled and rugose near-shore intertidal and subtidal environment.
Because of their rotund bodies, Pacific Spiny Lumpsuckers are poor
swimmers, but they have a pelvic suctorial disc that helps them stick
to surfaces (Arita, 1969; Budney & Hall, 2010; Hart, 1973;
Tietbohl, 2014). Male lumpsuckers tend nests in evacuated barnacles
covered in red coralline algae, and their whole body fluoresces in the
same red as their substrate. This fluorescence, in concert with the out-
line breaking armor, camouflages them against the algae covered
background (Cohen & Summers, in press). The spinous armor of
lumpsuckers may also play a defensive role as the fish face a myriad
of assaults from biotic and abiotic stressors (Figure 1).
Here, we quantify the development and morphology of armor in
the Pacific Spiny Lumpsucker. Our goals were fourfold: 1) describe
the development of lumpsucker armor, 2) quantify the amount of min-
eral investment in armor relative to skeleton across ontogeny, 3) qual-
itatively and quantitatively assess damage to lumpsucker armor, and
4) determine the material that constitutes the armor.
2|MATERIALS AND METHODS
We used a number of methods to investigate Pacific Spiny
Lumpsucker armor. Historically the plates of the armor are referred to
as tubercles (Arita, 1969). But this terminology does not reflect the
FIGURE 1 Eumicrotremus orbis, life
images. (a) Dorsal view, (b) lateral view,
displaying the cone-shaped, non-
overlapping, odontodes covering the body
2WOODRUFF ET AL.
developmental origin, the morphology of the plates, nor their material
properties. We will refer to these structures as odontodes because
this term is used for any epidermal structure that contains enamel
and/or dentine (Fraser et al., 2010). For example, this separates this
armor from the sturgeon's dermally derived bony plates. It also makes
a distinction between these structures and the keratinous breeding
tubercles of cyprinid fishes, which, though sharp and cone shaped, are
in no way related to Pacific Spiny Lumpsucker armor.
2.1 |Collection and CT scanning
We used micro-CT scanning to compare the morphology and mineral
investment of the Pacific Spiny Lumpsucker (E. orbis Günther, 1861)
armor across ontogeny. Specimens (n=39) were obtained through
the University of Washington Fish Collection, and arrived formalin-
fixed and preserved in 70% EtOH. Because the specimens arrived
fixed, we had no information on the animals' original color and could
not sex them. This was not a problem because the armor of male and
female lumpsuckers are subjected to similar abiotic factors and will
likely show similar patterns of wear. All scanning was done at the
Karel F. Liem Bio-Imaging Center at Friday Harbor Laboratories,
Friday Harbor, WA with a Bruker Skyscan 1173. Specimens ranged
from 9.2 to 97.0 mm standard length (SL) and were scanned with a
voxel size between 6.1 and 35.5 μm, a voltage of 55 or 65 kV, an
amperage of 123 or 133 μA, and an exposure of 1150 or 1115 ms
(Table S1). We used a 1 mm Al filter for all scans to reduce attenua-
tion artifacts. We scanned specimens together in batches of one to
four, and for each batch we scanned a separate can with two stan-
dards (i.e., phantoms) with known densities (25% and 75% hydroxyap-
atite, respectively) at the exact same settings and resolution. Phantom
scans were reconstructed in NRecon (Bruker, 2005–2011) with the
same settings as their corresponding fish scans so that relative bright-
ness of the fish could be converted into mineral density estimates. All
scans used in this study were uploaded to MorphoSource.org
(Table S1) and are freely available for download.
2.2 |Armor development and mineral investment
Reconstructed CT-scans were processed in the open-source image
analysis software 3D Slicer (version r29738), with the SlicerMorph
extension (Kikinis et al., 2014; Rolfe et al., 2021). We primarily used
the thresholding and scissor tools to include all skeletal material while
removing unnecessary voxels and background noise. The whole fish
was separated into individual segmentation nodes consisting of only
the armor and only the skeleton. We define the skeleton as a combi-
nation of both the appendicular and axial components. Volume and
mean brightness of the armor and skeleton were calculated using the
Segment Statistics tool in Slicer. We removed two individuals (84.5
and 86.5 mm SL) from our data set because either 100% or the major-
ity of their armor was damaged or lost. To investigate the scaling rela-
tionships of armor and skeleton volume with SL, we performed
standardized major axis regressions on log-transformed data using the
“lmodel2”R package (Legendre, 2018). Volume is proportional to
length cubed so the predicted slope for isometric growth was 3. Scal-
ing relationships were considered allometric if the predicted slope fell
outside of the bounds of the 95% confidence intervals.
We measured bone density by comparing mean voxel brightness
in each lumpsucker to the mean voxel brightness of the phantoms
with known densities. We used Segment Statistics to find the mean
voxel brightness for the 25% and 75% hydroxyapatite phantoms and
derived a standard curve for each corresponding lumpsucker scan.
From there, we calculated the mean concentration of hydroxyapatite
in the armor and skeleton of each specimen by transforming their
respective mean voxel brightnesses with the standard curves. We cal-
culated the ratio between the concentration of hydroxyapatite in the
armor and the concentration in the skeleton. We also calculated the
ratio of total mineral investment in the armor versus the skeleton for
each specimen. Total mineral investment was calculated by multiply-
ing the total voxel volume of the armor and skeleton by their respec-
tive mean hydroxyapatite concentrations.
An issue that can arise when segmenting thin structures with a
threshold is a partial volume effect, where intermediate gray scale
voxels just miss (or just make) the threshold cutoff and skew the mea-
sure of volume in a way that is not reproducible from specimen to
specimen. The odontodes certainly are thin enough to warrant con-
cern about partial volume, but three factors mitigate against this being
a significant effect. First, we scanned each fish at the best resolution
for its size, so an odontode, whether on an 80 mm or 15 mm fish, had
similar numbers of voxels—between 25,000 and 100,000 voxels. Sec-
ond, the skeleton, like the odontodes, is without large, dense, space
filling elements—instead every bone has many thin areas and sculptur-
ing. Third, when we looked at a threshold five grayscale values darker,
and five grayscale values lighter, than our selected threshold, the vol-
ume changed less than 8%. This establishes an upper bound for error
in the ratio of armor to bone because the bone volume would also
change with thresholding, and in the same direction.
2.3 |Assessment of damage
We divided the Pacific Spiny Lumpsucker armor into eight distinct
rows to help track the growth and orientation of lumpsucker armor
throughout ontogeny. These rows were initially selected by 3D-
printing an oversized lumpsucker and examining the armor for pat-
terns. We generated a hypothesis, based on our 3D-models, about
the eight rows present and mapped this pattern over lumpsucker
ontogeny using the micro-CT scans. In 3D Slicer, we segmented each
row of armor from the left side of the fish and recorded the total
number of odontodes as well as the number of damaged odontodes
to track the amount and location of damage over ontogeny.
We used SEM to assess types of damage to the armor. Lumpsucker
odontodes were carefully dissected away from the body and placed in a
1% trypsin solution for 24 h to let epithelial covering dissolve.
WOODRUFF ET AL.3
dehydration series in ethanol before being placed in 100% EtOH for 2 h.
We removed odontodes from 100% EtOH and allowed them to dry
uncovered for at least 24 h. We chose the first odontode in the fourth
row, a large odontode located in the middle of the body, to compare
damage and growth of an individual scale across ontogeny. These
odontodes were present in our smallest individual (9.2 mm) and were
easy to recognize from both the mico-CT scans and dissection. We sur-
veyed additional odontodes that had significant damage to evaluate the
extent of denticular damage. Once specimens were completely dried we
sputter coated them using a Cressington 108 Sputter Coater (Ted Pella,
Inc). To visualize each odontode we used a SEM (Neoscope JCM-5000).
Images were taken of the whole odontode, as well as the regions of
interest, including damage and cone patterns. The stage was tilted to a
maximum of 45to capture the sides of the odontode.
2.4 |Armor material
We used cross polarized microscope and SEM to determine the material
basis for lumpsucker odontodes. Individual scales were carefully dis-
sected away from the body and cleaned using a 1% Trypsin solution.
Cleaned odontodes were placed between the cross polarizers and
imaged using a ZEISS SteREO v20 Discovery microscope (Zeiss
Oberkochen, Germany). Each piece of armor was rotated between the
polarizers from when light was fully able to pass through the specimen
until light was perpendicular to the direction of the mineral resulting in a
black field of view. For SEM analysis, previously sampled and prepped
odontodes were frozen to 80C and broken. Cracked pieces were then
remounted on SEM stubs and imaged with a Neoscope JCM-5000.
3.1 |Investment in armor
The concentration of hydroxyapatite in Pacific Spiny Lumpsucker
armor versus their skeleton showed a weak positive relationship with
=.2; p=.005; Figure 2c). The armor: skeletal density ratio
ranged from 0.72 to 1.44, though 85% of fish had nearly equal
hydroxyapatite density in their armor as their skeleton (1 ± 0.2 s.e.).
However, there was a relatively strong positive relationship between
the armor: skeleton mineral investment ratio and SL (R
p< .001; Figure 2d). The total mineral (volume mean hydroxyapa-
tite concentration) devoted to armor increased over ontogeny from
FIGURE 2 Eumicrotremus
orbis, armor and skeletal density,
investment, and volume over
ontogeny. (a) Strong correlation
between the total number of
odontodes and standard length,
and (b) weak correlation between
the proportion of damaged
odontodes and standard length,
but the smallest individuals had
no damage. (c) A slight correlation
between armor density relative to
skeletal density and standard
=.2), (d) a strong
correlation between the mineral
investment in armor relative to
skeletal investment and standard
=.59), (e) positive
allometric growth in armor
volume over ontogeny (R
and (f) slight negative allometry in
skeleton volume over ontogeny
=.96). Axes in graphs (c–f) are
log transformed. Scale bar set
to 1 cm
4WOODRUFF ET AL.
12% of the skeleton in a 9.2 mm SL fish, to 121% in a 90.5 mm SL ani-
mal (Table S2). Only large lumpsuckers (89.8 mm SL-97 mm SL)
invested more mineral in their armor than their skeleton. The most
visibly armored fish (62 mm SL), whose armor left little unprotected,
and the only fish with ventrolateral odontodes fully abutting one
another, only invested half as much in its armor as in its inner skele-
ton. The changes in investment were primarily due to an increase in
armor volume relative to the skeleton over ontogeny (Figure 2e,f).
The relationship between armor volume and SL showed positive
allometry over ontogeny for all individuals excluding the most heavily
armored fish (slope =3.32; 95% C.I. =3.14–3.52, R
=.97, p< .001),
while the skeleton volume scaled with negative allometry (slope =
2.74; 95% C.I. =2.57–2.93, R
=.96, p< .001).
3.2 |Armor development
Armor develops in eight rows starting from the rostrum and extending
to the caudal fin (Figures 3 and 4). Six of the rows run horizontally
along the length of the body, one runs under the orbit, and the last
one is present only on the “chin”of the fish under the lower jaw. In
adults, armor covered the entire body except for the most ventral part
of the fish where the suctorial disc is located. These eight rows were
not fully established in smaller lumpsuckers; however, by 9.2 mm SL
there was at least one scale in each of the eight partitions, and a
sequence was evident in each of the rows by 11 mm SL. Armor begins
developing from anterior to posterior with more odontodes on the
skull and operculum than on posterior aspects of the body. As
lumpsuckers grow, the rows of armor are maintained and odontodes
in each of those rows maintains a clear identity relative to early devel-
opment (Figures 3 and 4). That is to say, odontodes do not migrate
along their row or across rows as new odontodes are added, and they
There was a positive correlation between the number of
odontodes and SL (R
=.71, p< .001), with the largest fish (97 mm
SL) having more than eight times the number of odontodes in each
row than the smallest fish (9.2 mm SL, Figure 5e). There were consis-
tently more odontodes in the first three rows of armor covering the
FIGURE 3 Eumicrotremus
orbis, armor over ontogeny. Left,
center, and right views showcase
dorsal, ventral, and lateral aspects
of lumpsucker armor. Odontodes
are arranged in eight rows: Rows
one through six (starting dorsally)
cover the body, one runs under
the eye, and one is under the
chin. Scale bar is 1 cm. Different
colors correspond to row
numbers where row 1 =red, row
2=orange, row 3 =yellow, row
4=green, row 5 =blue, row
6 (eye) =pink, row 7 =indigo,
row 8 (chin) =purple
WOODRUFF ET AL.5
neurocranium, operculum, and surrounding the eye than the rest of
the body. On average, adult lumpsuckers had 19.7 odontodes in each
of the first three rows, but just 17.5 in the others (Table S3). In smaller
fish (11 mm SL) the first three rows had 6–8 odontodes while the rest
only had 1–5 odontodes (Figure 3). In larger lumpsuckers, the first
three rows were composed of numerous smaller odontodes while
rows on the sides of their body had fewer odontodes that were larger.
Bigger lumpsuckers have more odontodes overall on their body and in
each of their rows than smaller lumpsuckers. Lumpsucker armor first
develops with large odontodes, and as those structures reach their
maximum size, smaller odontodes begin to fill in empty space.
3.3 |Odontode development
The primordial Pacific Spiny Lumpsucker odontode is a single, simple
cone of mineralized tissue. This cone is joined by others of varying
size and at arbitrary angles; and with each new addition, a spiral of
cones becomes more apparent (Figure 4). The odontodes of smaller
fish looked drastically different from the odontodes of adults. In juve-
nile lumpsuckers (9.2–20 mm SL), odontodes were highly topographic
structures formed of aggregated cones of similar size. The aggregate
made a hollow, volcano-shaped plate, but not all cones pointed
straight up or in the same direction. Rather, depending on their posi-
tion along the body, the cones pointed in different directions
(Figure 4). Odontodes oriented in a particular direction when they are
initiated stay in that direction. For instance, the largest odontode in
the sixth row remained curved compared to those along the midline
(Figure 4). Odontodes on the anterior end of the skeleton and the
head pointed rostrally, odontodes along the midline pointed outwards,
and odontodes on the posterior of the fish pointed caudally.
Odontodes begin to flatten out around 23.2 mm SL as new cones
(approximately the same size as earlier ones) continue to accrete at
the outer edge, and the whole structure takes on a broad-conical
FIGURE 4 Eumicrotremus
orbis, individual scale morphology
over ontogeny. (a) Shows the
morphogenesis of lumpsucker
armor from a single cone to a
large conical plate from row two;
(b) shows the growth of the
largest scale from row one;
(c) shows the growth of the first
scale in row four; (d) shows the
growth of the first scale in row
six. As the fish grows, cusps are
added, eventually forming a spiral.
All odontodes are shown in lateral
view, scale bar is 1 mm. Arrows
point to example of cones
6WOODRUFF ET AL.
shape. Odontodes maintain their hollow architecture and continue to
grow. Those in the eighth row were the most curved, while those in
row 4 were closest to a right circular cone (Figure 4c). Adult
lumpsucker odontodes (62–97 mm SL) appeared simpler than those
of smaller fish because the added cones remained the same size as
that first cone. Over ontogeny, the individual odontode shape is pre-
served, though the shape varies between rows. In a particular row,
the odontodes maintain the size relationship established when they
first appeared. The biggest odontode in row 8, for example, was
always the largest in that row. However, rows have different growth
rates: row 1 develops first, but the size of its odontodes is outstripped
by those in row 4 which develop last.
3.4 |Damage and material
We observed three different types of damage to the odontodes: com-
plete abrasion (cones are worn down to the base), partial abrasion
(tops of cones are worn down), and breakage (cones absent). These
types of damage are visible in micro-CT and SEM, though partial and
complete abrasion was more obvious in SEM and typically found on
the top of odontodes (Figure 5a–d). We found broken cones all over
the odontode, and smashed or chipped whole odontodes (Figure 5e,f).
Larger fish had more damaged odontodes than smaller fish, with the
five smallest fishes having no damage at all, but there was no signifi-
cant correlation between damage and size (R
Figure 5f). We excluded two fish (84.5 and 86.5 mm SL) where every
odontode or large patches were damaged or removed, with the next
most damaged fish being of average size between 25 and 40 mm
SL. The two largest fish (97 mm and 90 mm SL) had fewer than 5% of
their odontodes damaged (Figure 5f, Table S4). The fourth row,
located in the middle of the body, had the highest proportion of dam-
aged odontodes for any row (6%). The fourth row contained some of
the largest individual odontodes and the fewest number. Odontodes
under the chin had the smallest proportion of damage (1%) and were
the smallest and most numerous on the body.
FIGURE 5 Eumicrotremus
orbis, damage of the armor.
(a) Doral scanning electron
microscopy (SEM) image of
largest scale in row four, (b) inset
of the odontode from image
(a) showcasing abrasion damage.
(c) Dorsal SEM of odontode
shower complete breakage and
abrasion. (d) Inset from panel
(c) showering spalling (arrow) that
resulted in severely damaged
odontodes. (e) Lateral SEM image
of largest scale in row four
showcasing the spiraling effect of
adding more cones through
ontogeny, (f) inset of odontode
from image (e) looking at cone
WOODRUFF ET AL.7
SEMs of broken cones on the odontodes revealed a fibrous orga-
nization of mineralized tissue, while polarized light showed that the
mineral specularly refracts light like enamel (Figure 6a). There was no
evidence of repair or regrowth of broken or abraded odontodes or
cones as we would expect from enamel tissue (Figure 5f).
Pacific Spiny Lumpsucker armor is lightweight relative to that of other
fishes, like seahorses, bichirs, sea robins, boxfish, catfish, and gar
(Kawai, 2019; Lowe et al., 2021; Porter et al., 2013; Reichert &
Steffen, 2010; Yang et al., 2013, 2015). The largest adults invest
around half of the total mineral in their body in armor, compared to
more than 90% in some poachers (Agonidae) (Kruppert et al., 2020).
In terms of mineral investment this is more like the highly focused
defensive systems of lionfish or stingrays, where mineralized spines
represent a localized defensive approach (Galloway & Porter, 2019;
Shea-Vantine & Kajiura, 2021). However, the lightweight armor in
lumpsuckers covers nearly the entire body, so it superficially appears
to play as important a role as the heavy, whole-body armor of other
lineages. Pacific Spiny Lumpsuckers live in the rocky intertidal and are
poor, even comical, swimmers, powering locomotion with paired and
median fin undulation (Allen & Smith, 1988; Arita, 1969; Hart, 1973).
They must be at risk of collision with the structure in their habitat and
they cannot evade attacks of predators. We propose that lumpsucker
armor represents an innovation—lightweight armor that nevertheless
serves to protect against abrasion and impact.
The lack of overlap of the odontodes could indicate the armor is
not proof against attack because defensive armor often has over-
lapping regions of high stiffness (Browning, 2012; Ehrlich, 2015; Lowe
et al., 2021; Sherman et al., 2017). However, lumpsucker armor bears
the scars of not one, but two types of damage. Abrasion and impact
leave different signatures on an armored surface. Abrasion wears
away high points, leaving characteristic grooves and scratches that
can reveal the nature of the abrading surface (Amini & Miserez, 2013;
Reif, 1978). Impact cleaves, or spalls, armor, leaving sharply defined
edges and shatter patterns explained by the direction of the impactor
(Ehrlich, 2015). Lumpsucker odontodes have both types of damage
suggesting both gentle and persistent abrasion wear, and that the
acute stresses of impact are part of the lumpsucker's life. Odontode
row four, which girdles the widest part of the body, had the most
damage, suggesting inadvertent contact with the substrate is an
important factor. Odontodes are only abraded on the top, never along
the sides, so the abrasive surface must be broad. Though lightweight,
the armor is also a barrier to gape limited predators, and the cones on
the odontodes would make deglutition a challenge.
As the fish grows from pea to grapefruit size, we expect collisions
to be more frequent and at greater velocity. Drag and lift imposed by
the complex flow of the intertidal should grow with the area of the
fish, or length squared (Vogel, 2008). Forces on larger fishes will be
far larger than on small ones. Inertia and collision energy are also
expected to be far greater for larger individuals since both scale with
mass, which is growing with the cube of length (Vogel, 2008). As a
result, large Pacific Spiny Lumpsuckers are likely to be less able to
resist currents with their poor swimming performance, leading to
more collisions. The development of denticulation, and greater invest-
ment in armor, over ontogeny supports the notion that frequent and
traumatic collisions become increasingly important. Smaller fish also
have almost no damage, compared to some larger individuals where
up to 100% of the odontodes are damaged, and are more poorly
armored than larger ones, indicative of their collision-free life. Fur-
thermore, lumpsuckers are unique among armored fishes in that they
possess an adhesive disc that may help them reduce the chances of
collision by staying put (Budney & Hall, 2010; Tietbohl, 2014).
Fish scales, whether the unadorned disks of cycloid form or the
spikey plates of ctenoid form, develop first on the tail and proceed to
the head (Hughes, 1981). Ctenoid scales, which have simple, conical,
enamel spikes in the anterior region, start as plates of bone before
sprouting spikes (Roberts, 1993; Shono et al., 2019; Williamson &
FIGURE 6 Eumicrotremus
orbis, material of armor. (a) Cross-
polarized light microscopy of
adult lumpsucker scale. Refraction
patterns indicated that the
material is not amorphous bone
but rather enamel. (b and c)
scanning electron microscopy
(SEM) images showing the
layered enamel sheets that make
up lumpsucker armor
8WOODRUFF ET AL.
Carpenter, 1851). In contrast, the denticulated enamel cones of the
Pacific Spiny Lumpsucker first develop on the head, and they manifest
as a single cone. More cones are added, initially closely adherent to
the first, then more widely separated by a flat area; and these form a
spiral of hooked teeth on a conical surface. Like teeth, these
odontodes show damage and there is no evidence they ever repair.
As organisms grow and interact with their environment, we expect to
see evidence of collision, predation, and defense. Armor and special-
ized teeth bear the most obvious scars, but enamel and dentine, the
material basis for odontodes, do not repair and we see no signs of
replacement of lumpsucker armor. The question then becomes how
does armor stay protective as the animal grows. The answer should lie
in the maintenance of a developmental pathway that ensures continu-
ous production of odontodes but not necessarily replacements.
Odontic regions in teleosts are not unheard of, for example, the den-
ticular apparatus of ceratioid anglerfishes and the odontodes of
armored catfishes (Pietsch, 2005; Schaefer & Buitrago-Suárez, 2002).
Odontodes and teeth are unified by their developmental toolbox, and
despite differences in regeneration or shape, all require a source of
odontogenic tissue. We propose that the odontodes of lumpsuckers
are of odontogenic origin and the species might be of interest in
understanding the evolution and development of specialized teeth.
Overall, we find that only large lumpsuckers invest more in their
armor and bigger lumpsuckers have more and larger scales.
Odontodes begin as simple cones that aggregate into a flatten plate
studded with smaller cones that cover the fish in eight distinct
regions. The non-overlapping odontodes of Pacific Spiny Lumpsuckers
may aid in camouflage against a background of rocks and barnacles
while protecting this fish as it bounces through rough, subtidal waters.
Special thanks to the Eugster Endowed Student Research and Intern-
ship Fund for funding to E.C.W., Stephen and Ruth Wainwright
Endowment, Edwards award, Wingfield-Ramenofsy Award, and Ori-
ans Award to K.E.C., Karel F. Liem Bioimaging center, DBI-1852096,
and DBI-1759637 to A.P.S., NSF Graduate Research Fellowship
(DGE-1746914) to J.M.H., Katherine Maslenikov at the University of
Washington Burke Museum for access to specimens.
Eleanor C. Woodruff: Data curation (equal); investigation (equal); visu-
alization (equal); writing –review and editing (equal). Jonathan M.
Huie: Conceptualization (equal); formal analysis (equal); investigation
(equal); methodology (equal); validation (equal); visualization (equal);
writing –review and editing (equal). Adam P. Summers: Funding
acquisition (equal); methodology (equal); project administration (equal);
resources (equal); validation (equal); writing –original draft (equal);
writing –review and editing (equal). Karly E. Cohen: Conceptualization
(lead); data curation (equal); funding acquisition (equal); investigation
(equal); methodology (equal); project administration (lead); resources
(equal); software (equal); supervision (equal); validation (equal); writing
–original draft (equal); writing –review and editing (equal).
The peer review history for this article is available at https://publons.
DATA AVAILABILITY STATEMENT
All CT scans and data are uploaded and freely available for download
Eleanor C. Woodruff https://orcid.org/0000-0002-0996-7678
Jonathan M. Huie https://orcid.org/0000-0002-7925-7372
Adam P. Summers https://orcid.org/0000-0003-1930-9748
Karly E. Cohen https://orcid.org/0000-0001-6556-5414
Allen, M. J., & Smith, G. B. (1988). Atlas and zoogeography of common fishes
in the Bering Sea and northeastern Pacific (NOAA Technical Report
NMFS 66), p. 151.
Amini, S., & Miserez, A. (2013). Wear and abrasion resistance selection
maps of biological materials. Acta Biomaterialia,9, 7895–7907.
Arita, G. S. (1969). Sexual dimorphism in the cyclopterid fish Eumicrotremus
orbis.Journal of the Fisheries Board of Canada,26, 3262–3265.
Berge, J., & Nahrgang, J. (2013). The Atlantic spiny lumpsucker
Eumicrotremus spinosus: Life history traits and the seemingly unlikely
interaction with the pelagic amphipod Themisto libellula.Polish Polar
Browning, A. (2012). Mechanics and design of flexible composite fish armor
(Thesis). Massachusetts Institute of Technology.
Buser, T. J., Summers, A. P., & Sidlauskas, B. L. (2019). Stags of the sea? Com-
parisons of territoriality and cranial weapon morphology in the fish sub-
family Oligocottinae (Pisces: Cottoidea). Journal of Morphology,280, S92.
Budney, L., & Hall, B. (2010). Comparative morphology and osteology of
pelvic fin-derived midline suckers in lumpfishes, snailfishes and gobies.
Journal of Applied Ichthyology,26, 167–175.
Cohen, Karly E & Summers, Adam P. Dimorphic Florescence in the Pacific
Spiny Lumpsucker (In Press). Ichthyology and Herpetology.
Ehrlich, H. (2015). Materials design principles of fish scales and armor. In
H. Ehrlich (Ed.), Biological materials of marine origin: Vertebrates
(pp. 237–262). Springer Netherlands.
Fraser, G. J., Cerny, R., Soukup, V., Bronner-Fraser, M., & Streelman, J. T.
(2010). The odontode explosion: The origin of tooth-like structures in
vertebrates. BioEssays,32, 808–817.
Hart, J. L. (1973). Pacific fishes of Canada. Bulletin - Fisheries Research
Board of Canada,180, 740.
Galloway K. A., Porter M. E. (2019). Mechanical properties of the venom-
ous spines of Pterois volitans and morphology among lionfish species.
Journal of Experimental Biology, 22(6), jeb197905. https://doi.org/10.
Günther, A. (1861). Catalogue of the acanthopterygian fishes in the collec-
tion of the British Museum. 3. Gobiidae, Discoboli, Pediculati,
Blenniidae, Labyrinthici, Mugilidae, Notacanthi. London. I-xxv +1-586
Hughes, D. R. (1981). Development and organization of the posterior field
of ctenoid scales in the Platycephalidae. Copeia,1981, 596–606.
Kawai, T. (2019). Revision of an armored searobin genus Scalicus Jordan
1923 (Actinopterygii: Teleostei: Peristediidae) with a single new spe-
cies. Ichthyological Research,66, 437–459.
Kikinis, R., Pieper, S. D., & Vosburgh, K. G. (2014). 3D slicer: A platform for
subject-specific image analysis, visualization, and clinical support. In
Intraoperative imaging and image-guided therapy (pp. 277–289). Springer.
WOODRUFF ET AL.9
Kolmann, M. A., Peixoto, T., Pfeiffenberger, J. A., Summers, A. P., &
Donatelli, C. M. (2020). Swimming and defence: Competing needs
across ontogeny in armoured fishes (Agonidae). Journal of the Royal
Society Interface,17, 20200301.
Kolmann, M. A., Urban, P., & Summers, A. P. (2020). Structure and function
of the armored keel in piranhas, pacus, and their allies. The Anatomical
Kruppert, S., Chu, F., Stewart, M. C., Schmitz, L., & Summers, A. P. (2020).
Ontogeny and potential function of poacher armor (Actinopterygii:
Agonidae). Journal of Morphology,281, 1018–1028.
Lowe, A., Summers, A. P., Walter, R. P., Walker, S., & Misty Paig-Tran, E. W.
(2021). Scale performance and composition in a small Amazonian
armored catfish, Corydoras trilineatus.Acta Biomaterialia,121, 359–370.
Pietsch, T. W. (2005). Dimorphism, parasitism, and sex revisited: Modes of
reproduction among deep-sea ceratioid anglerfishes (Teleostei:
Lophiiformes). Ichthyological Research,52, 207–236.
Porter, M. M., Novitskaya, E., Castro-Ceseña, A. B., Meyers, M. A., &
McKittrick, J. (2013). Highly deformable bones: Unusual deformation
mechanisms of seahorse armor. Acta Biomaterialia,9, 6763–6770.
Reichert, S. H., & Steffen, H. (2010). Reverse engineering nature: Design prin-
ciples for flexible protection inspired by ancient fish armor of Polypteridae
(Thesis). Massachusetts Institute of Technology.
Reif, W. E. (1978). Protective and hydrodynamic function of the dermal
skeleton of elasmobranchs. Neues Jahrbuch für Geologie und Pal-
äontologie Abhandlungen,157, 133–141.
Roberts, C. D. (1993). Comparative morphology of spined scales and their
phylogenetic significance in the Teleostei. Bulletin of Marine Science,
Rolfe, S., Pieper, S., Porto, A., Diamond, K., Winchester, J., Shan, S.,
An open and extensible platform to retrieve, visualize and analyze 3D mor-
phology. Methods in Ecology and Evolution,12, 1816–1825.
Schaefer, S. A., & Buitrago-Suárez, U. A. (2002). Odontode morphology
and skin surface features of Andean astroblepid catfishes
(Siluriformes, Astroblepidae). Journal of Morphology,254, 139–148.
Sherman, V. R., Quan, H., Yang, W., Ritchie, R. O., & Meyers, M. A. (2017).
A comparative study of piscine defense: The scales of Arapaima gigas,
Latimeria chalumnae and Atractosteus spatula.Journal of the Mechanical
Behavior of Biomedical Materials,73,1–16.
Shea-Vantine C. S., Galloway K. A., Ingle D. N., Porter M. E., Kajiura S. M.
(2021). Caudal Spine Morphology and Puncture Performance of Two
Coastal Stingrays. Integrative and Comparative Biology,61, 749–758.
Shono, T., Thiery, A. P., Cooper, R. L., Kurokawa, D., Britz, R., Okabe, M., &
Fraser, G. J. (2019). Evolution and developmental diversity of skin
spines in pufferfishes. IScience,19, 1248–1259.
Song, J., Reichert, S., Kallai, I., Gazit, D., Wund, M., Boyce, M. C., &
Ortiz, C. (2010). Quantitative microstructural studies of the armor of
the marine threecone stickleback (Gasterosteus aculeatus). Journal of
Structural Biology,171, 318–331.
Song, J., Ortiz, C., & Boyce, M. C. (2011). Threat-protection mechanics of
an armored fish. Journal of the Mechanical Behavior of Biomedical Mate-
Tietbohl, M. (2014). Fishes that suck: Comparison of the adhesive discs of
three fishes of the Pacific Northwest (p. 17). Friday Harbor Laboratories.
Vamosi, S. M., & Schluter, D. (2004). Character shifts in the defensive
armor of sympatric sticklebacks. Evolution,58, 376–385.
Vogel, S. (2008). Modes and scaling in aquatic locomotion. Integrative and
Comparative Biology,48, 702–712.
Webb, P. W., Hardy, D. H., & Mehl, V. L. (1992). The effect of armored skin
on the swimming of longnose gar, Lepisosteus osseus.Canadian Journal
of Zoology,70, 1173–1179.
Williamson, W. C., & Carpenter, W. B. (1851). Investigations into the struc-
ture and development of the scales and bones of fishes. Abstracts of
the Papers Communicated to the Royal Society of London,5, 969–971.
Yang, W., Chen, I. H., Gludovatz, B., Zimmermann, E. A., Ritchie, R. O., &
Meyers, M. A. (2013). Natural flexible dermal armor. Advanced Mate-
Yang, W., Naleway, S. E., Porter, M. M., Meyers, M. A., & McKittrick, J.
(2015). The armored carapace of the boxfish. Acta Biomaterialia,23,
Additional supporting information may be found in the online version
of the article at the publisher's website.
How to cite this article: Woodruff, E. C., Huie, J. M., Summers,
A. P., & Cohen, K. E. (2022). Pacific Spiny Lumpsucker
armor—Development, damage, and defense in the intertidal.
Journal of Morphology,1–10. https://doi.org/10.1002/jmor.
10 WOODRUFF ET AL.