BENTHIC WALKING, CRYPTIC LOCOMOTION, AND EVOLUTIONARY 
MORPHOLOGY OF BODY FORM IN FLATFISHES (PLEURONECTIFORMES) 
 
A Dissertation 
Presented to the Faculty of the Graduate School 
of Cornell University 
In Partial Fulfillment of the Requirements for the Degree of 
Doctor of Philosophy 
 
 
 
 
 
 
by 
Claire Hope Fox 
December 2020 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
© 2020 Claire Hope Fox
 
BENTHIC WALKING, CRYPTIC LOCOMOTION, AND EVOLUTIONARY 
MORPHOLOGY OF BODY FORM IN FLATFISHES (PLEURONECTIFORMES) 
 
Claire Hope Fox, Ph. D.  
Cornell University 2020 
 
Flatfishes (Pleuronectiformes) are an ecologically and economically important group of 
benthic predators with unique adaptations to life on the bottom. Flatfishes’ key 
innovations—eye-migration and lateralized behavior —result in evolutionary constraints 
as well as novel opportunities for this group (Chapter 1). How does an organism find 
food and evade predators when it is lying on one side? Using kinematic and comparative 
analyses (Chapter 2) I show that flatfishes use a mode of locomotion that is structurally 
and functionally unique amongst vertebrates, in which successive portions of median fins 
are co-opted to form functional “feet” that push against the substrate. This unusual form 
of locomotion allows the flatfish to maneuver, forage, and traverse across the substrate at 
slow to moderate speeds while maintaining a low profile. Using Digital Particle Image 
Velocimetry (Chapter 3) I measured the velocity magnitude and vorticity of water flow 
around walking and swimming flatfishes. I found that water disturbance was much lower 
during walking, without the trailing vortices produced during swimming, suggesting that 
benthic walking in flatfishes is hydrodynamically cryptic. To better understand the 
evolutionary morphology of this large and diverse group I analyzed body and fin shapes 
of 67 species from 12 flatfish families using geometric morphometric and phylogenetic 
methods (Chapter 1). Lastly, in fulfillment of my NSF Graduate Research Fellowship in 
K-12 STEM Education, I evaluated the impact of a new mobile science program using a 
combination of survey and interview methods (Chapter 4). 
 ii 
 
BIOGRAPHICAL SKETCH 
 
Claire Hope Fox was born in Burlington, Vermont, and spent her childhood running 
around in the forest, catching animals, and splashing in the streams of the Blue Ridge 
Mountains, Virginia. She graduated from Cornell University with a B.A. in Visual 
Studies in 2001. Fascinated by the intersection of art and technology, she created an 
interdisciplinary course of study that integrated cultural theory, studio art, art history, 
technology studies, architecture, and the science of perception. Claire settled in Ithaca 
NY to raise two sons, Felix and Casper, with her husband Mark. A life-long learner at 
heart she has served as co-founder, director, and teacher at a series of educational 
nonprofit organizations including a cooperative learning center, community makerspace, 
mobile science program, and community science workshops. She is dedicated to building 
resources that enable anyone, especially from communities that have been historically 
oppressed and marginalized, to get excited about and enjoy science. It was her work with 
under-served youth that inspired her to pursue a PhD. In encouraging young people to 
dream big and overcome barriers she realized that she, too, could pursue her childhood 
dream of studying Zoology.   
 iii 
 
Dedicated to all the young people out there who are curious and fascinated by the natural 
world around them, who remind me daily to find the joy  
and never stop asking questions 
 iv 
 
ACKNOWLEDGMENTS 
 
I owe deepest gratitude to my advisor, Willy Bemis, who introduced me to the 
remarkable evolutionary and morphological diversity of fishes, encouraged me to pursue 
a PhD, and believed in me every step of the way. I could not have done this without you! 
Thank you to Betty McGuire, Kate Bemis, Joshua Moyer and Stacy Farina, for creating a 
supportive lab environment where I felt welcome and appreciated.  
 
Many thanks to Alice Gibb and Adam Summers for the inspiration and insight that 
sparked this thesis, during the fantastic Functional Morphology of Fishes Course at 
Friday Harbor Lab. I am grateful for the thoughtful feedback of my committee members: 
Amy McCune, John Hermanson, Casey Dillman, Adam Summers, and Willy Bemis. 
Thank you to the Department of Ecology and Evolutionary Biology for the opportunity to 
pursue this research - it has been a privilege and an honor.  
 
To my mother, who knows the name of every plant and still swims in the ocean in the 
winter, whose love of nature placed me on this path; and my father, whose 
entrepreneurial spirit and work ethic helped me to persevere. To my sister for always 
understanding and being there for me. To my sons, Felix and Casper, for believing in me. 
And most of all thank you to my husband, Mark, for unconditional support in all 
dimensions of our life together. 
  
 v 
 
TABLE OF CONTENTS 
 
Chapter 1 .......................................................................................................................................... 1 
Chapter 2 ........................................................................................................................................ 57 
Chapter 3 ........................................................................................................................................ 94 
Chapter 4 ...................................................................................................................................... 126 
Appendix ..................................................................................................................................... 174 
 vi 
 
LIST OF FIGURES 
 
Chapter 1 
Figure 1.1. Diversity of body form in flatfishes .............................................................................5 
Figure 1.2. Landmarks used for analyses......................................................................................12 
Figure 1.3. Example of a Turbot-type feeder, Psettodes erumei ..................................................17 
Figure 1.4. Example of a Plaice-type feeder, Bothus ocellatus ....................................................18 
Figure 1.5. Example of a Sole-type feeder, Paraplagusia japonica ............................................. 19 
Figure 1.6. Number of valid species per family of Pleuronectiformes .......................................... 22 
Figure 1.7. Landmark-based morphospace of Pleuronectiformes ................................................. 24 
Figure 1.8. Loadings plot for landmark-based PCA ...................................................................... 25 
Figure 1.9. Meristic-based morphospace of Pleuronectiformes .................................................... 26 
Figure 1.10. Boxplot of relative eye diameter and distance between the eyes .............................. 28 
Figure 1.11. Boxplot of jaw length and abdominal size ................................................................ 29 
Figure 1.12. Ancestral state reconstruction of feeding type .......................................................... 30 
Figure 1.13. Ancestral state reconstruction of substrate type ........................................................ 31 
Figure 1.14. Ancestral state reconstruction of water type ............................................................. 32 
Figure 1.15. Body shape does not correlate with feeding type ...................................................... 38 
Chapter 2 
Figure 2.1. Percentage time spent walking, bounding, rotating, and swimming ........................... 66 
Figure 2.2. Frames from a high-speed video (120 fps) ................................................................. 68 
Figure 2.3. Tracing of the tip of a single fin ray from the middle of the anal fin .......................... 70 
Figure 2.4. Fin rays in the dorsal and anal fin move symmetrically ............................................. 71 
 vii 
 
Figure 2.5. Undulations of the body during walking and swimming ............................................ 72 
Figure 2.6. Frames from high-speed videos of Parophrys vetulus bounding ................................ 74 
Figure 2.7. Movements of two species of flatfish ......................................................................... 75 
Figure 2.8. Frames from high-speed videos of Platichthys stellatus ............................................. 77 
Figure 2.9. Comparison of fin ray angles for three locomotor behaviors ..................................... 78 
Figure 2.10. Fin ray angles during a rotation of Parophrys vetulus .............................................. 80 
Figure 2.11. Speed in BL/s as a function of wave frequency ........................................................ 81 
Figure 2.12. Average speed of body and median fins ................................................................... 82 
Figure 2.13. Comparisons of speeds of bounding, walking, and swimming ................................. 83 
Chapter 3 
Figure 3.1. DPIV experimental set-up ......................................................................................... 100 
Figure 3.2 Orientation of the fish in the laser plane during DPIV capture .................................. 101 
Figure 3.3. Morphology features of Trinectes maculatus ........................................................... 104 
Figure 3.4. High velocity and strong turbulent flow in the wake ................................................ 106 
Figure 3.5. Vorticity plots of the flow in the wake of swimming and walking ........................... 109 
Figure 3.6. Fluid vorticity of flow around swimming and walking ............................................ 110 
Figure 3.7. Sequence of vector fields surrounding swimming Trinectes maculatus ................... 111 
Figure 3.8. Schematic illustration of flow around Trinectes maculatus ...................................... 112 
Chapter 4 
Figure 4.1. Venues hosting the Physics Bus (Ithaca-based program) ......................................... 134 
Figure 4.2. Populations served by the Physics Bus program in 2014-2019 ................................ 135 
Figure 4.3. The Physics bus mobile facility overview ................................................................ 136 
Figure 4.4. Documentation of favorite exhibits in order of popularity ....................................... 145 
 viii 
 
Figure 4.5. Twelve most popular exhibits ................................................................................... 146 
Figure 4.6. Surprising sensory experiences of physics phenomena ............................................ 150 
Figure 4.7. Familiar, everyday materials repurposed to produce physics exhibits ..................... 152 
 ix 
 
LIST OF TABLES 
 
Table 1.1. Categorization of feeding types .................................................................................... 17 
Table 2.1. Analyses of Walking Speeds ........................................................................................ 63 
Table 2.2. Analyses of Bounding for Parophrys vetulus .............................................................. 63 
Table 3.1. Average velocity magnitude and vorticity in three flow regions ............................... 107 
 
 
  
 x 
 
APPENDIX 
 
Supplemental Figure 1. Landmark based PCA morphospace with additional outgroups ........... 174 
Supplemental Figure 2. Ancestral State Reconstruction with alternate phylogeny .................... 175 
Supplemental Figure 3. Reconstruction of feeding type without prior at root node ................... 176 
Supplemental Table 1.1. Specimens studied and vertebral number ............................................ 177 
Supplemental Table 1.2. Feeding type sources and categorization ............................................. 180 
 xi 
 
CHAPTER 1 
DIVERSITY OF BODY FORM IN FLATFISHES (CARANGARIA: 
PLEURONECTIFORMES) WITH COMMENTS ON FUNCTIONAL BIOLOGY AND 
FEEDING BEHAVIOR 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1 Claire Fox and William E Bemis 
1  
 
 
Abstract 
Flatfishes (Pleuronectiformes) are a diverse and ecologically important group of 
benthic predators with unique morphological innovations for life on the bottom. How the 
unique Bauplan of flatfishes arose and diversified to fill ecological niches world-wide has 
been the subject of historic and recent interest. Here, to explore the morphological 
diversification of flatfishes, we use geometric morphometric analysis of radiographs and 
photographs in a phylogenetic context to investigate patterns of morphological diversity 
for 67 species (139 specimens) representing 13 of the 14 extant families of 
Pleuronectiformes. Separate PCA analyses of landmark-based data and meristic data are 
compared. Our analyses include transitional Eocene fossils, both species in the basal-
most extant family of Pleuronectiformes, and outgroups within Carangaria. We resurrect 
and redefine three feeding types among flatfishes and use ancestral state reconstruction to 
evaluate the evolution of three ecologically important variables: feeding type, substrate 
type, and water type. We confirm family-level trends in eye size, interocular distance, 
jaw size, and abdomen size that relate to feeding type. The largest component of body 
shape evolution in flatfishes is elongation, which we show is correlated with increased 
vertebral counts. We discuss some functional implications of morphological differences 
in body, fin, and jaw shapes, and note limitations of our categories of feeding types 
because some flatfishes are broadly opportunistic predators.
2  
 
 
“The Pleuronectidae, or Flat-fish, are remarkable for their asymmetrical bodies…the eyes 
offer the most peculiarity, for they are both placed on the upper side of the head. That the 
Pleuronectidae are remarkably adapted by their flattened and asymmetrical structure for 
their habits of life, is manifest from several species, such as soles, flounders, etc., being 
extremely common. The chief advantages thus gained seem to be protection from their 
enemies, and facility for feeding on the ground. The different members of the family 
present, as Schiödte remarks, “a long series of forms exhibiting a gradual transition 
from Hippoglossus pinguis [Reinhardtius hippoglossoides] which does not in any 
considerable degree alter the shape in which it leaves the ovum, to the soles, which are 
entirely thrown to one side.” - Charles Darwin  
1872 p. 387 
3  
 
 
1. Introduction 
Flatfishes (Pleuronectiformes) are an ecologically and economically important 
order of primarily marine fishes with about 800 species in 14 families occupying a wide 
range of ecosystems worldwide, structuring benthic communities and filling niches as 
keystone predators and prey, contributing largely to benthic biomass, and including 
commercially valuable food fish species (Munroe, 2015). Post-metamorphic flatfishes 
exhibit a key innovation, one that is unique among vertebrates, for life on the benthos: 
asymmetry and eye migration during ontogeny that allow flatfishes to lie flat on the 
substrate on one side of the body with both eyes pointing upwards. This innovation is 
accompanied by a suite of secondary adaptations for predation, locomotion, and 
camouflage, including: protrusible eyes (Chapleau, 1993); ability to alter body 
pigmentation to match the background (Ramachandran et al., 1996; Akkaynak, 2017); 
dorsal and anal fins with long bases (the dorsal fin originates on cranium) that are used 
for benthic locomotion (Wirtz and Davenport, 2017; Fox et al., 2018); asymmetric 
development of the lateral line (Voronina et al., 2019); asymmetry of trunk musculature; 
asymmetry of skin coloration; asymmetry of jaw morphology and jaw movement (Gibb, 
2003); and often, asymmetry of pectoral fins. Although flatfishes share this broad suite of 
secondary adaptations, functional aspects of their morphology including body elongation, 
eye size, jaw size and shape, length of fin rays, and shapes of fins, vary greatly (Fig. 1.1).  
 
4  
 
 
 
Figure 1.1. Diversity of body form in flatfishes. Pleuronectiformes from six families display a 
range of body and fin shapes and features related to prey sensing and capture (eye size, eye 
location, jaw size and jaw symmetry). Images were created by layering a partial-transparency 
photograph of each fluid-preserved specimen over its inverted radiograph. 
5  
 
 
1.1. Historical and Evolutionary context  
The flatfish Bauplan has long puzzled and fascinated evolutionary biologists, and 
periodically over the past 150 years flatfishes have been at the center of heated debates 
over evolutionary theory and methods. Transitional flatfish fossils were unknown in 
the19th century, and discussion about the origin and history of Pleuronectiformes was 
mostly speculative. Mivart, an early advocate of Darwin’s evolutionary theory who is 
notorious for later becoming one of his most vociferous critics, used flatfishes as a prime 
example of the problems he perceived with natural selection. Mivart (1871: 38) asserted 
that the gradual selection-driven change proposed by Darwin was not feasible because a 
partly-migrated eye “must rather have been injurious” to a pleuronectid in its 
intermediate form; he allowed that a sudden change in eye position could be beneficial 
but “an accidental occurrence of such a spontaneous transformation is hardly 
conceivable.” Sixty two years later, Goldschmidt (1933: 545) posited the spontaneous 
transformation of body form in flatfishes that Mivart had thought impossible. 
Pleuronectids were, according to Goldschmidt, likely examples of “Hopeful Monsters” – 
organisms in which a single mutation with a large affect results, randomly and rarely, in a 
trait with adaptive value that founds a new group. Goldschmidt believed that eye 
migration in flounders could be accomplished with a single mutation; all other cranial 
and neural transformations would be “necessary consequences of the first step” and 
smaller mutations would be gradually accumulated over time through natural selection.  
Recent paleontological evidence shows that evolution of the flatfish Bauplan was 
gradual and incremental, with ocular asymmetry being one of the first derived features to 
arise. In his description of †Heteronectes chaneti and re-examination of †Amphistium 
6  
 
 
paradoxum, both from localities near Bolca, Italy, Friedman (2008, 2012) recognized 
intermediate eye migration along with a mixture of ancestral and derived character states 
that place them unequivocally as stem Pleuronectiformes. Partial eye migration found in 
large-sized, fully ossified adult †Heteronectes and †Amphistium and their presence in the 
fossil record over two geological stages proves that the intermediate condition was not 
maladaptive in this group. Piscivory, most likely through ambush, is likely to be the 
feeding mode of stem flatfishes, and †Amphistium and †Heteronectes were piscivorous 
based on morphological features and stomach contents preserved in the fossils (Friedman 
2008, 2012). Experimental studies of diet and behavioral observations for the basal-most 
extant genus of Pleuronectiformes, Psettodes, exclusively foraged by ambush predation, 
accepting only live demersal fish (Kizhakudan et al., 2013). Juvenile and adult Psettodes 
lie buried in the sand waiting for prey fish to approach and “snap the fish with wide open 
flexible jaws bearing large number[s] of teeth and swallow it”, typically consuming one 
or two before retreating to the sand (Kizhakudan et al., 2013: 28).  
Which came first, the behavior or morphology? Other examples of extant fishes 
outside the flatfish clade that employ similar predation strategies include the cichlid 
Haplochromis livingstoni, which lies on one side on the substrate, apparently feigning 
death, until small fishes come close enough to ambush (McKaye, 1981); similar behavior 
has been noted in another cichlid (Tobler, 2005) and a serranid, the comb grouper 
Mycteroperca acutirostris (Gibran and Armbruster, 2004). For fish already exhibiting 
this mode of ambush behavior, improvements in visual perception of prey could be 
advantageous. Schreiber (2005) showed that lateralized behavior initiates before eye 
7  
 
 
migration in larval flatfishes and is decoupled from it. It may be that behavior preceded 
morphology in this case.  
Despite recent progress in understanding the evolutionary history and 
interrelationships of Pleuronectiformes, many questions remain (Munroe, 2015). Even the 
fundamental question of whether flatfishes are monophyletic has been controversial 
(Campbell et al., 2013; Betancur-R. and Ortí 2014) with early molecular support 
inconclusive, most likely due to long-branch attraction (Wei et al., 2018); the subsequent 
debates spurred both criticisms of and refinements to molecular phylogenetic methods. 
Here, we regard Pleuronectiformes as monophyletic and sister to Pleuronectoidei based 
on anatomical synapomorphies (Chapleau, 1993), and molecular analyses with sound 
hypothesis-testing procedures and adequate taxonomic coverage (Betancur-R. et al., 
2013b; Betancur-R. and Ortí 2014), and a large number of ultraconserved DNA element 
loci and broader clade sampling (Harrington et al., 2016). Molecular phylogenetic studies 
place Pleuronectiformes within Carangaria (Carangimorpha, sensu Li et al. 2009, 
Harrington et al. 2016; Carangimorpharia, Betancur-R. et al., 2013a,b), which includes 
archerfishes, Nile perches, cobias, remoras, swordfishes, moonfishes, barracudas, and 
jacks. Pleuronectiformes originated in the Paleocene and rapidly diversified in an 
adaptive radiation (Harrington et al., 2016; Ribeiro et al., 2018), a pattern of 
diversification common within Carangaria and acanthomorph clades more generally 
(Friedman, 2010). The earliest anatomically modern flatfish, †Eobothus minimus 
(minimum age 63 Ma, Rabosky et al., 2018) has morphological apomorphies 
characteristic of the bothid lineage of Pleuronectoidei; †Eobothus minimus is abundant at 
Monte Bolca along with †Amphistium paradoxum (Chanet, 1999). One of the earliest 
8  
 
 
soles (Soleidae), †Eobuglossus eocenicus from a locality near Cairo, also dates from this 
period (Chanet, 1994).  
Early taxonomic efforts grouped flatfishes by the morphological trait of 
“handedness”, i.e., to which side of the body eyes migrated, a trait now understood to be 
an unreliable indicator of relatedness because of familial-, generic-, and individual-level 
reversals (Palmer, 2009). Monophyly of six families of Pleuronectiformes is well 
established by both morphological and molecular phylogenetic analyses: Psettodidae 
(Chapleau, 1988), Pleuronectidae (Cooper and Chapleau, 1998), Cynoglossidae 
(Chapleau, 1988), Samaridae (Sakamoto, 1984), Achiridae (Ramos, 1998), and Soleidae 
(Chapleau and Keast, 1988). Four groups traditionally regarded as subfamilies of 
Pleuronectidae have been elevated to families: Paralichthodidae, Rhombosoleidae, and 
Poecilopsetttidae (Chapleau and Keast, 1988; Chapleau, 1993). Scophthalmids, 
traditionally regarded as a sub-family of Bothidae (Norman, 1934), are now confirmed to 
be monophyletic and are broadly recognized as Scophthalmidae (Hensley and Ahlstrom, 
1984; Chapleau 1993; Chanet 1998). Achiropsettidae, Paralichthodidae and 
Rhombosoleidae are small families whose affiliations and familial status remain 
uncertain. Paralichthyidae is problematic: in addition to a lack of morphological 
synapomorphies (Chapleau, 1993; Hoshino, 1999), molecular phylogenies consistently 
divide the family into two or more groups. Monophyly of Citharidae and its recognition 
as a family was reinforced by identification of new morphological synapomorphies 
(Hoshino, 2001) following decades of contention over shared derived characters and the 
inclusion of genera with opposite “handedness”. 
9  
 
 
Recent molecular phylogenetic studies provide enough taxon sampling within 
Pleuronectiformes to begin exploring phenotypic evolution in this important and curious 
clade of fishes (Byrne et al., 2018, Rabosky et al., 2018). In this study we use geometric 
morphometrics to examine body form variation of flatfishes in a phylogenetic context, 
investigate possible ecological correlates of body form, and assess conservation and 
convergence of aspects of feeding and habitat within Pleuronectiformes using ancestral 
character state reconstruction. 
2. Methods 
2.1. Specimen Selection 
Specimens were selected from the fluid ichthyology collections of the 
Smithsonian Institution, American Museum of Natural History, and Cornell Museum of 
the Vertebrates; selection was informed by specimen condition, morphological diversity, 
and availability of phylogenetic and ecological data. We studied 139 specimens 
representing 67 species of Pleuronectiformes from 13 families: Psettodidae: 2 species 
(N=9), Citharidae: 5 species (N=12), Samaridae: 3 species (N=12), Cynoglossidae: 5 
species (N=11), Achiridae: 6 species (N=12), Soleidae: 5 species (N=9), Pleuronectidae: 
7 species (N=15), Paralichthyidae: 9 species (N=17), Bothidae: 9 species (N=16), 
Scophthalmidae: 5 species (N=12), Poecilopsettidae: 4 species (N=4), Rhombosoleidae: 5 
species (N=5), Achiropsettidae: 2 species (N=2). We chose the pilotfish, Naucrates 
ductor (Carangaria: Carangidae), as an outgroup. We included two genera of fossil 
flatfishes (N=2) with exceptional preservation: †Amphistium paradoxum 
(MNHNFBOL86) and †Heteronectes chaneti (NMNH_1974-1639-24); we used 
10  
 
 
photographs of fossils for landmark-based morphometrics and interpretive drawings and 
descriptions in Friedman (2008, 2012) for meristic counts. We estimated eye sizes and 
position based on the ocular opening. We grouped species into families for PCA, 
ANOVA, and ASR analyses according to the 14 families recognized by Munroe (2015) 
and Nelson et al. (2016). We obtained current counts for the number of species in each 
family from Eschmeyer’s Catalog of Fishes (Eschmeyer et al., 2020). We did not include 
three small families (Leptobramidae, 2 species, Oncopteridae, 1 species, 
Paralichthodidae, 1 species) recognized by Eschmeyer et al. (2020). 
2.2. Radiography and Photography  
NMNH specimens were radiographed at the Smithsonian Museum Support Center 
using a Varian PaxScan 4030 E panel with Kodak Lanex Fine Screen scintillator and 40 x 
28 cm dimension of capture; Xray source was Thermo Scientific Kevex PXS5-927EA, 
Focal spot 4 microns @ 2 watts, Target-Tungsten-Window Be, Beam angle 45 
deg.; Image capture by VIVA K.03 Image Acquisition/Control Software. Additional 
specimens were radiographed at the AMNH (Carestream DRX, Rochester, NY) and 
CUMV specimens at the Cornell Veterinary Clinic (imaging system: Computed 
Radiography System, AGFA NX, Canton MA; radiography tube: Linear MC 150, North 
American Imaging, Camarillo, CA). A Cannon EOS Rebel t3 was used to photograph the 
eyed and blind side of each specimen. Radiographs and photographs of the eyed side of 
each specimen were overlaid in Photoshop, and the opacity of the top layer (the 
photograph) was reduced to allow visibility of both soft tissue and skeletal landmarks, 
resulting in two-layer images saved in .tiff format. 
11  
 
 
2.3. Selection and Digitization of Landmarks  
We used 2-D landmarks because the highly compressed body forms of flatfishes 
allow 2-D morphometrics to capture shape variation for co-planar landmarks (Buser et 
al., 2018). We defined 25 homologous fixed landmarks and 16 sliding semi-landmarks, 
which are used to create point to point correspondence along homologous curves 
(Bookstein 1997) (Fig. 1.2).Skeletal and soft-tissue landmarks deemed of functional and 
ecological relevance were selected: landmarks 1-12 capture eye size and eye position, 
jaw size and jaw orientation (relevant to prey-capture); landmarks 13 and 17 define the 
abdominal region (relevant to feeding mode); landmarks 16-20 and 31, 32 capture fin 
length and placement (relevant to locomotion); semi-landmarks 23-30 and 33-40 capture 
the width and curvature of the body (relevant to locomotion). Landmarks related to 
features that were not present in all species, such as pectoral fins, were excluded from the 
morphometric data set. Measurements of caudal fin rays were excluded because damage 
to tails is common in fluid fish collections. We used sliding semi-landmarks to define the 
exterior margin of the body form where homology was impossible to determine. All 
landmarks were placed on specimen images by one individual (senior author) to 
minimize errors of interpretation.  
12  
 
 
 
Figure 1.2. 2-D Landmarks: (1-4) exterior margin of left eye (dorsal-most eye); (5-8) exterior 
margin of right eye (ventral-most eye); (9) anterior tip of premaxilla; (10) anterior tip of dentary; 
(11) posterior end of maxilla/corner of oral opening; (12) articular-quadrate jaw joint; (13) 
posterior of cranium; (14) posterior-most point of operculum; (15) interopercular joint; (16) 
origin of pelvic fin; (17) anterior-most caudal vertebrae; (18) dorsal-most intersection of caudal 
fin rays with the body; (19) hypural joint; (20) ventral-most intersection of caudal fin rays with 
the body; (21) mid-point of posterior edge of urostyle; (22) origin of dorsal fin; (23-30) sliding 
semi-landmarks along base of dorsal fin; (31) last anal fin ray (32) last anal fin ray; (33-40) 
sliding semi-landmarks along base of anal fin; (41) origin of anal fin. Note that right-facing 
(“right-handed” or dextral) specimens were transformed horizontally for geometric morphometric 
analyses, to match left-facing (“left-handed” or sinistral) specimens. 
 
We used tpsUtil ver. 1.76 x64 to build tps files from images and append tps 
curves to landmarks. Landmarks and semi-landmarks were placed using tpsDig2 ver. 
2.31 (Rohlf, 2004). 
2.4. Landmark-based PCA Analyses 
We used landmark-based morphometric methods (R ver. 4.0.2 and package 
geomorph ver. 3.3.1, Adams et al., 2020) to quantify shape variation in the 13 families of 
flatfishes together with representatives of two genera of fossil flatfishes (†Amphistium 
paradoxum and †Heteronectes chaneti) and one outgroup to Pleuronectiformes, the 
13  
 
 
Pilotfish Naucrates ductor. We used rotate.coords flipX to flip “right-eyed” specimens to 
face left; without this transformation the overwhelming driver of the morphospace would 
be the character state of sinistral vs. dextral, which is fixed at the family level for nearly 
all flatfishes, with a few exceptions such as Psettodidae and Platichthys stellatus 
(Munroe, 2015). Generalized Procrustes Analysis (Gower, 1975) was performed by 
gpagen (Adams et al., 2013) to align specimen coordinates and correct for differences in 
specimen sizes, rotation, and location in the frame. gpagen was used to slide semi-
landmarks along their tangent directions to a position optimized by algorithms that 
minimize bending energy (Bookstein 1997). Principal component analyses (PCA) were 
performed on the aligned Procrustes coordinates using gm.prcomp. We plotted PCA 
functions with family and species labels and then drew and colored convex hulls in 
Adobe Illustrator 16.0.0 for easy visualization of the 13 families within the morphospace.  
2.5. Meristic Counts and Measures 
Values for the meristic PCA were counted and measured on radiographs imported 
into Adobe Photoshop. Total length was measured from the tip of the rostrum to the 
posterior edge of the caudal fin, and body width was measured perpendicular to the spine 
at the deepest spot on the body. To represent variation in dorsal and anal fin morphology, 
we counted the number of dorsal and anal fin rays and measured the longest and shortest 
fin rays on both fins (some species of flatfishes have very long fin rays that they use to 
prop up their bodies off the substrate, while others are uniformly short). Fin-ray lengths 
were measured with a straight line from the base of the fin ray to its distal tip, excluding 
fin rays that were broken or unnaturally bent. Vertebrae and fin rays were counted in the 
image using a small tic mark to keep track of progress. Because intraspecific variability 
14  
 
 
in vertebral counts is well documented due to developmental plasticity in response to 
temperature (Hubbs, 1922) and water turbulence (Corral and Aguirre, 2019) and 
opposing direct and indirect selection (Swain, 1992; Tibbin et al., 2016), our meristic 
analyses incorporated multiple individuals for most species that had been collected at 
different locations and times. JMP Pro 14 was used to perform a PCA on the meristic 
data; lengths were converted to ratios (ratio of length to width, ratio of shortest to longest 
fin rays) to control for specimen size. The resulting plot, with points colored by family, 
was imported into Adobe Illustrator and hulls were drawn and filled with colors to match 
those of the landmark PCA. 
We used interlmkdist in geomorph (Adams et al., 2020) to obtain specific inter-
landmark measures relevant to feeding type (post-Procrustes transformation). Inter-
landmark measure included in boxplots were jaw length (landmarks 10 to 12, anterior tip 
of dentary to jaw joint); eye size (landmarks 1 and 3, diameter measured in the vertical 
orientation of the migrated eye); interocular spacing (landmarks 3 and 5, distance 
between eyes); and abdominal length (landmarks 13 and 17 from posterior of cranium to 
first caudal vertebrae).  
2.6. Feeding Type Categorization 
We obtained diet data from peer-reviewed journal articles and federally-
sponsored literature for 33 of our study species (for sources and categorization, see 
Supplementary Table 1). Diet and life history information are unknown for many species 
of flatfishes, particularly smaller-bodied species with tropical distributions and those that 
are uncommon or have low commercial value (Munroe, 2015). For each species with 
15  
 
 
data, we determined the major prey categories e.g., “Fishes and Small Crustaceans” (see 
Supplementary Table 1). 
We defined three feeding types, inspired by categories proposed by Yazdani 
(1969) and de Groot (1969): Turbot-type, Plaice-type, and Sole-type (Table 1, Figs. 1.2-
1.4). We dropped the fourth category proposed by Yazdani (1969), “Ammotretis-type” 
because it is easily represented by features of Sole-type feeders. We also expand 
Yazdani’s (1969) concept for Plaice-type feeders to include fast-moving benthic prey. To 
assign each of the 33 species included in our ancestral state reconstructions to one of the 
three functional feeding types, we combined diet data, our observations of morphology, 
and published accounts of feeding behavior (where available). Our feeding-type 
categorization allows us to distinguish among flatfish species that consume the same prey 
category using different behavioral and functional methods. For example, “Fish” prey can 
be large or small, and pelagic, demersal, or benthic; flatfishes can locate prey using 
vision, mechanoreception, or olfaction; flatfishes can ambush attack from a hiding 
position on the substrate, locate and pursue prey by swimming in the water column, or 
forage and pursue prey on the substrate by creeping, bounding, stalking, and lunging 
(Olla et al., 1972, Fox et al., 2018). Some species of flatfishes exhibit behavioral 
plasticity and complexity in feeding, while others have stereotyped behavior and a 
limited predatory repertoire (Homes and Gibson, 1983). Our three feeding categories are 
intended to reduce this complexity into groups that retain essential form-function 
characteristics.  
Table 1.1 summarizes the three feeding types, together with prey types, predatory 
behaviors, and morphological features.  
16  
 
 
Feeding Prey Type Predatory Behavior Morphology 
Type 
Turbot-type Active, motile prey, e.g. Ambush predation from Jaws: large and nearly symmetrical with long 
fishes, cephalopods. substrate and active pursuit in sharp teeth 
water column.  Caudal fin: large 
Visual detection of prey. Abdominal cavity: large Eyes: large, 
moderately protuberant; migrate just past the 
dorsal midline of body  
 
Plaice-type Motile and sessile Active foraging on and above Jaws: medium sized somewhat asymmetrical  
benthic prey e.g. the benthos. Eyes: large, extremely protuberant 
crustaceans, benthic Generally diurnal. Abdominal cavity: variable within the mid-
fishes, mollusks. Primarily visual detection of range  
prey but can also use olfaction.  
Sole-type Low-mobility infaunal Crawls on substrate searching Jaws: small and highly asymmetrical, minute 
and epifaunal prey e.g. for food.  teeth 
polychaetes, worms, Generally nocturnal. Eyes: small and closely spaced  
siphon tips of bivalves. Mechanosensory and Abdominal cavity: small 
chemosensory detection of prey 
but can also use vision. 
Table 1.1 Categorization of feeding types, prey types, predatory behavior, and morphology. 
 
Psettodes erumei (Fig. 1.33) has the characteristically large, nearly symmetrical 
jaws and long sharp teeth of a Turbot-type feeder that eats fishes and squid in the water 
column. Its streamlined body and robust caudal fin allow active swimming pursuit of 
prey. The abdominal region can accommodate large prey. The relatively low number of 
vertebrae (9 abdominal and 14 caudal) and fin rays are associated with a stiffer, less 
flexible body. Large eyes used for visual detection of prey are moderately protuberant, 
with eye migration barely past the mid-line. 
 
17  
 
 
 
Figure 1.3. Example of a Turbot-type feeder, Psettodes erumei. Inset of eyed side (B) and blind 
side (C) of head showing large eyes and large symmetrical mouth. Note the fish prey item in the 
stomach. 
18  
 
 
The Plaice-type feeder Bothus ocellatus (Fig. 1.4) diurnally forages on small to 
medium benthic prey using vision. In comparison to a Turbot-type feeder, it has smaller, 
moderately asymmetrical jaws used to attack prey on the substrate. Large and extremely 
protuberant eyes allow visual detection and tracking of prey from a position on the 
substrate. Consumes generally small benthic prey. A higher number of caudal vertebrae 
(26) is associated with a more flexible body and an increased number of fin rays that can 
be used to walk on the substrate.  
 
Figure 1.4. Example of a Plaice-type feeder, Bothus ocellatus. Inset of eyed side (B) and blind 
side (C) of head showing large eyes and medium-sized, slightly asymmetrical mouth. 
19  
 
 
 
 Paraplagusia japonica (Fig. 1.5) is a Sole-type flatfish that eats small, low-
mobility prey such as siphon tips of bivalves. The dorsal fin originates anterior to the 
head, and the caudal fin skeleton is absent, with the dorsal and anal fins confluent around 
the posterior tip of the body. In comparison with Turbot-type and Plaice-type feeders, 
Sole-type feeders have small, closely spaced eyes correlated with nocturnal feeding and 
reliance on other sensory systems for prey detection. The small, highly asymmetrical 
mouth opens on the blind side of the body; it enables feeding from the substrate on small 
infaunal and epibenthic prey. The abdomen is small, but the number of caudal vertebrae 
(43) is greater than in either the Turbot-type or Plaice-type feeders. 
 
Figure 1.5. Example of a Sole-type feeder, Paraplagusia japonica. Inset of eyed side (B) and 
blind side (C) of head showing tiny, closely spaced eyes and small, highly asymmetrical mouth. 
 
In all three feeding types, the abdomen has 9 vertebrae. Body elongation is 
indicated by the number of caudal vertebrae. 
20  
 
 
2.7. Ecological Character Mapping and Ancestral State Reconstruction 
Only species that had both detailed diet data and were included in Rabosky et al. 
(2018) were used in ancestral state reconstruction of feeding type (N=27). We obtained 
substrate and depth data from FishBase ver. 12/2019 (Froese and Pauly, 2000) for N=47 
species. If a single substrate-type (or no substrate type) was included in FishBase, we 
consulted additional sources. We combined mud substrate type with sand for our “Soft 
Bottom” substrate type because all species in our study listed in FishBase as residing on 
mud were, with further research, found to utilize sand bottoms in certain locations or to 
change preference in different life stages.  
We used phylogenetic data from the Fish Tree of Life project (Rabosky et al., 
2018) which combines sequences from the Euteleost TOL (Betancur-R. et al., 2013a), 
Rabosky set (Rabosky et al., 2018), and Near set (Near et al., 2012). Timetree files were 
downloaded from fishtreeoflife.org, imported into R, and pruned to our study species. An 
alternate tree based on data from Byrne et al. (2018) was used for comparison (see 
Supplemental Materials based on data downloaded from the Github repository 
https://github.com/sarisbro; Byrne_etal_SeqAlignments.tar.bz2).  
We performed stochastic character mapping using the function make.simmap in 
phytools v. 0.7.47 (Revell, 2012) which modifies ace (Paradis et al., 2004) to compute 
likelihood using Felsenstein’s pruning algorithm. The routine make.simmap uses a 
continuous-time reversible Markov model for the evolution of discrete character traits, 
which is used along with tip states on the tree to simulate stochastic character histories. 
Stochastic character mapping using a probabilistic approach avoids the drawbacks of 
parsimony (Bollback, 2006). We used the default “empirical” method in make.simmap, 
21  
 
 
which maximizes the likelihood of the transition matrix and samples character histories 
for the discrete trait, from their posterior distribution. We created two ancestral state 
reconstructions for feeding mode: 1) with a prior imposed, setting the probability = 1 for 
“Turbot-type” at the root node, reflecting fossil evidence that this is the ancestral state for 
Pleuronectiformes; 2) using an “equal” prior distribution which produces a root node 
sampled from the conditional scaled likelihood distribution at the root (see Supplemental 
Materials for this Ancestor State Reconstruction). 
3. Results 
3.1. Feeding Types 
The general classification of feeding types for 11 families studied is shown in 
Figure 1.6 along with estimates of total species diversity for each family. These are broad 
generalizations because all members of a family may not conform to one of the three 
feeding types. Families generally classified as Turbot-type are the least diverse (127 
species); Plaice-type families have twice as many species (284), largely due to the 
diversity of Bothidae. Families generally classified as Sole-type are the most diverse, 
with 380 species. For details on classification see Supplemental Table 1.2 in Appendix. 
22  
 
 
 
Figure 1.6. Number of valid species per family of Pleuronectiformes assigned to each of our 
feeding types, color coded. Species number is based on valid species listed in Eschmeyer’s 
Catalog of Fishes. Paralichthyidae shows the highest species richness for turbot-type feeders. 
Bothidae are the most species-rich family of Plaice-type visual foragers on motile benthic prey 
species. Cynoglossidae and Soleidae show high diversity of Sole-type nocturnal foraging feeders. 
Species counts per family are as follows. Among turbot-type feeders: Psettodidae (3), Citharidae 
(6), Scophthalmidae (9), Paralichthyidae* (109). For Plaice-type feeders: Rhombosoleidae (19), 
Samaridae (30), Pleuronectidae (63), Bothidae (172). For Sole-type feeders: Achiridae (35), 
Cynoglossidae (161), Soleidae (184). Two families, Poecilopsettidae (21) and Achiropsettidae 
(4), were data deficient for diet and predation mode and were excluded. *Paralichthyidae is 
broadly considered paraphyletic. 
 
3.2. PCA Morphospace 
Principle component analysis of landmark-based shape variation in 67 species of 
flatfishes from 13 families found PC1 (46%) and PC2 (24%) explained 70% of the total 
variance (Figs. 1.7 and 1.8). PC2 and PC3 contributed 5.7% and 5.3%. PC1 primarily 
reflects variation in the primary and secondary body axes, i.e. increase in body length vs. 
width (because flatfishes undergo axial re-orientation during metamorphosis, body depth 
is functionally body width). Forms with wide bodies are on the high end of PC1 with 
23  
 
 
Scophthalmus maximus (Scophthalmidae) in the bottom right representing the positive 
extreme. The elongate body form of Cynoglossus senegalensis (Cynoglossidae) 
represents the negative extreme.  
 
Figure 1.7. Landmark-based morphospace of Pleuronectiformes produced by PCA (N=67 
species). Polygons represent the distribution of species examined within each family. Combined 
radiographs and photographs of species representing morphological extremes are placed adjacent 
to their point in the plot. The greatest variance within PC1 relates to body elongation in the upper 
left) and deep-bodied forms in the lower right) with additional influence from abdominal size 
(larger in the upper right), jaw size (smaller in lower left). PC2 reflects variance in origin of the 
dorsal, anal, and pelvic fins from a more ancestral position (top and upper right) vs. shifted far 
forward (bottom and lower left), and degree of eye migration from minimal (top) to extreme 
(bottom). 
 
24  
 
 
 
 
Figure 1.8. Loadings plot for landmark-based PCA. Variables that are far from the origin have 
the highest loadings on the principal components. PC1 is strongly influenced by variables on the 
far left (negative), e.g. elongate body form, and far right (positive), e.g. disk-shaped body. 
Variables that have the largest influence on PC2 are at the bottom of the plot (negative), e.g. 
extensive eye migration, at the top (positive), e.g. more posterior origin of dorsal, anal, and pelvic 
fins. 
 
 
Jaw and abdominal sizes also contribute to PC1. Families in the lower left 
quadrant of Figure 1.7 have small jaws and small abdominal regions, typified by Solea 
solea (Soleidae), while those in the upper right have large jaws and large abdominal 
regions, e.g., fossil flatfishes and Psettodes.  
PC2 reflects variation in fin and jaw position, and degree of eye migration. At the 
top of Figure 7, occupying the positive PC2 extreme, is the outgroup Naucrates ductor 
(Carangidae) along with Psettodes spp. and Lyopsetta exilis (Pleuronectidae); these taxa 
have dorsal, pelvic, and anal fins originating farther back on the body. Families with 
negative PC2 at the bottom of the plot, exemplified by Gymnachirus sp. (Achiridae), 
have dorsal and anal fins shifted far forward that wrap nearly all the way around the head 
and upper abdomen. Families with positive PC2 scores have superior or terminal mouths 
25  
 
 
while those at the negative extreme have small subterminal jaws. Lastly, variation in eye 
position contributes to PC2, with flatfishes at the top of the morphospace in Figure 7 
having an eye that has barely migrated past the body mid-line, in contrast with those with 
negative PC2, which exhibit eye migration ~180° onto the opposite side of the body. 
3.3. Meristic Morphospace 
A PCA morphospace based on meristic data (Fig. 1.9) includes variation in 
vertebral number and aspects of dorsal and anal fin morphology assessed by fin-ray 
counts and measures of fin-ray lengths.  
 
Figure 1.9. Morphospace of Pleuronectiformes produced by PCA based on vertebral and 
fin-ray counts and lengths of body axes and fin rays. Polygons represent the distribution of 
species within each family. Representatives of each quadrant are Platichthys stellatus (upper left) 
with a wide body and dorsal/anal fin rays that are lengthened into peaks in the mid-region of the 
body; Cynoglossus senegalensis (upper right) with high vertebral count and short dorsal and anal 
26  
 
 
fin rays; Samariscus longimanus, with an elongate body and a consistent length to fin rays; and 
Psettodes erumei in the lower left with low vertebral and fin-ray counts. (N=139 individuals, 67 
species) 
 
We found a large overlap in the positions of families within the meristic 
morphospace. PC1 (50%) and PC2 (34%) explained 84% of the total variance. Families 
with positive PC1 and PC2 on the upper right of the plot have high vertebral counts, 
which are in turn correlated with high median fin-ray counts, exemplified by tonguefishes 
such as Cynoglossus senegalensis, which has 61 vertebrae (9 abdominal; 52 caudal), 133 
dorsal fin rays, and 106 anal fin rays. Taxa that are negative for PC1 have fewer than half 
as many vertebrae and fin rays, e.g., members of Achiridae have notably negative PC1 
scores including Trinectes maculatus with 26 vertebrae (9 abdominal; 17 caudal), 51 
dorsal fin rays and 37 anal fin rays. Psettodes erumei has 23 vertebrae (9 abdominal; 14 
caudal), 50 dorsal fin rays and 38 anal fin rays. Vertebral and fin ray counts are even 
lower for the fossil flatfishes, †Heteronectes and †Amphistium. Families with positive 
PC1 and negative PC2 (lower right quadrant) have relatively elongate bodies and mid-
range vertebral counts, e.g. Solea solea has 46 vertebrae (9 abdominal; 37 caudal), 77 
dorsal fins-rays, and 63 anal fin rays; their fin rays are more consistent in length, i.e., the 
longest and shortest fin rays have a 1:1 or 1:2 ratio. Families in the upper left quadrant 
have wider bodies and dorsal and anal fins with a high ratio of shortest to longest fin rays, 
e.g. the longest fin ray is eight times as long as the shortest in Platichthy stellatus 
(Pleuronectidae).  
3.4. Comparison of morphological measures relating to feeding type 
To assess variation in morphological features within and between families related 
to foraging and predation, we compared means of four inter-landmark measures: eye size 
27  
 
 
and inter-ocular distance (Fig. 1.10); and jaw-length and abdominal size (Fig. 1.11). 
Measures were taken after Procrustes transformation in scale to account for size 
differences in specimens. Eye sizes and interocular distances were larger for families we 
identified as primarily visual predators (Turbot-type and Plaice-type, Fig. 1.6). Jaw 
length and abdominal size were largest for the two species of Psettodes and the fossil 
flatfishes †Amphistium and †Heteronectes. Other piscivorous flatfishes had large to 
intermediate jaws and abdominal size. Plaice-type feeders were consistently intermediate 
on jaw size, however they showed high variability in abdominal size, with Pleuronectidae 
on the large side and Samaridae on the small. Again Achiridae, Soleidae, and 
Cynoglossidae form a distinct cluster with small abdominal and jaw size. 
 
28  
 
 
Figure 1.10. Boxplot of relative eye diameter (top) and distance between the eyes (bottom). 
Includes thirteen families of extant flatfishes, two transitional fossil species, and the outgroup, 
Naucrates ductor.  
 
Figure 1.11. Boxplot of jaw length (top) and abdominal size (bottom) for thirteen families of 
extant flatfishes, two transitional fossil species, and an outgroup. Jaw length was measured from 
the anterior tip of dentary to the jaw joint; abdominal size was measured from the posterior of 
cranium to first caudal vertebrae. 
 
3.5 Ancestral State Reconstruction 
Ancestral state reconstructions for feeding (Fig. 1.12), substrate (Fig. 1.13), and 
water type (Fig. 1.14) infer a common ancestor for Pleuronectiformes that was a marine-
dwelling ambush predator with a preference for soft-bottom substrates. 
29  
 
 
  
Figure 1.12. Ancestral state reconstruction of feeding type in flatfishes using stochastic character 
mapping methods infers multiple transitions from Turbot to Plaice feeding type, and a single 
transition from Turbot to Sole feeding type on the branch leading to Achiridae, Samaridae, 
Cynoglossidae, and Soleidae*. Feeding type categories are “Turbot-type” ambush and active 
pursuit predators on primarily fish prey; “Plaice-type” visual predators of benthic organisms; and 
“Sole-type” nocturnal predators on primarily epibenthic and infaunal prey. Root node prior was 
set to “Turbot type” based on fossil evidence of transitional forms and extant character state in the 
basal-most family. *Symphurus plagiusa has been placed within Cynoglossidae by other 
molecular and morphological analyses. 
30  
 
 
 
 
Figure 1.13. Ancestral state reconstruction of substrate type in Pleuronectiformes inferred a soft-
bottom (sand/mud) at the root node with preferences for sand/gravel, sand/gravel/rock, and 
gravel/rock arising multiple times in distantly related families. 
31  
 
 
 
Figure 1.14. Ancestral state reconstruction of water type in Pleuronectiformes inferred a marine 
origin for the clade. Multiple independent originations of deep-water, reef-associated, and 
brackish/freshwater types are represented, with frequent transitions to brackish/freshwater 
inferred within Cynoglossidae and Soleidae. 
 
Character mapping of feeding type (Fig. 1.12) showed Sole-type feeding arising 
once (excluding Symphurus plagiusa, which is generally placed in Cynoglossidae 
(Eschmeyer 2020) but was not placed near Cynoglossidae based on the phylogenetic data 
32  
 
 
set that we used to generate Fig. 1.12). Turbot-type feeding independently transitioned to 
Plaice-type several times; e.g., in Bothidae and Pleuronectidae (Fig. 1.12). Character 
mapping for substrate type (Fig. 1.13) showed an overwhelming preference for soft 
bottoms (sand and mud) with a few species living on other substrates such as sand, sand 
and gravel, mixtures of sand, gravel and rock, or gravel and rock. Ancestral water  type 
(Fig. 1.14) was inferred to be shallow marine, with independent transitions to 
bathymarine, reef-associated, and brackish/fresh water. The turbot-type families 
Psettodidae, Scophthalmidae, and one branch of Paralichthyidae, are nearly exclusively 
marine, while the primarily Sole-type families Achiridae, Soleidae, and Cynoglossidae 
showed many transitions to brackish and freshwater.  
4. Discussion 
Although Pleuronectiformes share a suite of adaptations for hunting and hiding in 
benthic environments, this large and ecologically important group also displays 
incredible morphological variation that we explored using geometric morphometrics in a 
phylogenetic context. In summary: we found that the largest component of shape 
variation in flatfishes is body elongation, which is correlated with an increase in the 
number of caudal vertebrae, which is in turn correlated with an increasing number of 
dorsal and anal fin-rays. Some families have a highly conserved body shape while others 
display great variation. Specific measures such as jaw size and position, abdomen size, 
and eye size and position correlate to feeding type however body shape does not. 
Ancestral state reconstructions for feeding, substrate, and water type infer a common 
ancestor for Pleuronectiformes that was a marine-dwelling ambush predator with a 
preference for soft-bottom substrates. Multiple independent transitions to Plaice-type 
33  
 
 
feeding suggest convergent evolution on feeding and foraging on the bottom on a range 
of visually detected motile and sessile benthic prey.  
Body elongation, the largest component of shape variation in flatfishes, is an axis 
of shape change that dominates diversification of fishes more broadly (Claverie and 
Wainwright, 2014; Price et al., 2019). We found that abdominal size and the origin and 
length of the dorsal and anal fins and their shape (curvature of the body and length of fin-
rays) also contributed largely to shape variation. Jaw length, eye position, eye size, and 
body curvature emerged as additional important components of morphological variation.  
Psettodidae, †Amphistium, †Heteronectes, and Naucrates ductor occupy opposite 
sides of the morphospace from Cynoglossidae and Soleidae (Fig. 1.7), with each of these 
groups distinctly clustered and non-overlapping. 
Bothidae had the largest variance and overlapped with the greatest number of 
other families in both landmark-based and meristic morphospaces (Figs. 1.7 and 1.9). 
This overlap reflects both conservation and convergence of traits. Within Bothidae 
extremes of body form were associated with transitions to bathymarine habitats, such as 
the gulper-type jaws, large gular pouch, and elongated body of Chascanopsetta 
crumenalis (reminiscent of the pelican eel, Eurypharynx pelecanoides), and to reef 
habitats, such as the wide-eyed and wide-bodied Bothus ocellatus (Fig. 1.1). Character 
mapping inferred a marine demersal, soft-bottom origin for Pleuronectiformes (Fig. 
1.14); occupation of different aquatic environments and substrates may be driving 
morphological and species diversification. Members of the most speciose families 
(Bothidae, Cynoglossidae, and Soleidae) in our sample show more transitions to different 
34  
 
 
water types compared to the least speciose families (Psettodidae, Scophthalmidae, 
Citharidae). 
According to both Rabosky et al. (2018) and Byrne et al. (2018), Samaridae is 
more closely related to Cynoglossidae and Soleidae yet in our landmark-based 
morphospace its body form overlaps entirely with Bothidae (Fig. 1.7). The large eyes and 
relatively symmetrical jaws of samarids as well as descriptions of active crepuscular 
foraging on the benthos, waving of the elongated pectoral fin (Kuiter and Tonozuka, 
2001) perhaps as a lure or as a sensory appendage, and documentation of small benthic 
fishes and invertebrates as prey, place samarids closer to a Plaice-type rather than Sole-
type feeder. If the phylogenetic placement is correct, then this is an example of 
convergence in feeding type. 
Elongation in vertebrates can be achieved by increasing the number or length of 
vertebrae (Ward and Brainerd, 2007; Wake, 2009); here we show that elongation in 
flatfishes is associated with an increase in the number of caudal vertebrae (Figs. 1.3-1.5, 
1.9; Supplementary Table 1). Transitional fossil flatfishes and basal-most families such 
as Psettodidae have low vertebral counts (22-25), while members of the highly elongated 
and later-diverging family Cynoglossidae have more than twice as many vertebrae (45-
61). There are several possible hypotheses for the functional value of increased vertebral 
number and elongation in flatfishes. Vertebral counts may increase flexibility, which in 
turn may lead to more effective C-start and escape maneuvering (Brainerd and Patek, 
1998). Conformation to benthic surfaces might be aided by increasing vertebral count. 
Intriguing footage of highly elongated tonguefish crawling (Okeanos ROV mission, 
2016, EX1605L1 CAPSTONE: ROV Exploration of the Marianas Trench MNM, 
35  
 
 
location Esmerelda Bank) document this flexibility, with the flatfish conforming its long 
body to dips and rises of large rocks on the seafloor, allowing its many fin rays to contact 
and push against the substrate (also see Fox et al. 2018). Axial flexibility and high 
vertebral counts may be a factor in burying, which flatfishes achieve by undulating the 
body rapidly in waves while in contact with the substrate. Flatfishes with a range of body 
and fin morphologies have been shown to bury in generally similar ways, however there 
may be functional specializations for different substrate types (McKee et al., 2016); 
flatfishes can exhibit strong preference for different substrates, which changes 
ontogenetically and has been shown to correlate to burying ability as well as food 
availability (Tanda, 1990; Phelan et al., 2001). Elongation may allow physical access to 
restricted spaces. Lastly, increased vertebral number in some groups may reflect a 
developmental response, rather than adaptation, to environmental conditions. 
Achiropsettidae was centrally located in the landmark morphospace (Fig. 1.7) with an 
intermediate length to width ratio yet emerged as an outlier on the meristic morphospace 
(Fig. 1.9) due to its high number of vertebrae (51 vertebrae, 105 dorsal fin rays, 93 anal 
fin rays). “Jordan’s Rule,” in which there is an inverse relationship between water 
temperature and vertebral number, has been experimentally demonstrated to link to 
developmental conditions in some groups (Corral and Aguirre, 2019) and may play a role 
in the morphology of this exclusively circumpolar family. 
 Adaptive benefits of having a dorsal fin that is advanced onto the head – a 
synapomorphy for Pleuronectiformes (Chapleau 1993) – are unclear. Flatfishes show 
large variation in the position of the origin of the dorsal fin, from the ancestral condition 
far back on the head (Fig. 1.3) to its extreme forward location in Achiridae and 
36  
 
 
Cynoglossidae (Fig. 1.5). Fox et al. (2018) showed that the anterior portion of the dorsal 
fin is not used to push against the substrate during benthic walking in two species of 
Pleuronectidae; observations of the achirid Trinectes maculatus and the tonguefish 
Symphurus plagiusa confirm that fin rays on the head are not used for propulsion during 
walking Fox et al. (submitted). Fin rays on the head may have been co-opted for sensory 
purposes, particularly for species that hunt nocturnally or in deep-sea or turbid estuary 
environments. Pterygiophores supporting fin rays extend and enlarge the anterior margin 
of the head in members of Achiridae and Cynoglossidae (some of which have a rostral 
hook extending the anterior portion of their head). Species in these families show 
frequent evolutionary transitions to brackish and freshwater habitats (Fig. 13). Having a 
nearly continuous margin of fin rays may help these groups adhere to the substrate (like 
the thin edge of a suction cup) to avoid displacement by flowing tidal or river water. In 
handling Trinectes maculatus in locomotor studies Fox et al. (submitted) demonstrated 
firm attachment to the substrate and resistance to removal; spreading their dorsal and 
caudal fin rays broadly and arching their backs slightly created a tight seal with the 
potential for a negative pressure zone beneath. Anecdotal observations of other round-
bodied species specializing in adhering to rock surfaces, e.g. the Topknot, Zeugopterus 
(Scophthalmidae) (Gosse, 1865) suggests that the round-body type with a fringe of fins 
surrounding is may be an adaptation for adherence. 
The close phylogenetic relationship between Achiridae and Cynoglossidae + 
Soleidae raises interesting questions: our results show that the three families share a 
common ancestor and overlap in diet, feeding mode, and occupy similar environments, 
37  
 
 
yet have body shapes and vertebral and fin-ray counts on opposite extremes (Fig. 1.15 
and Figs. 1.7 and 1.9).  
 
Figure 1.15. Body shape does not always correlate with feeding type. Species with similar body 
forms were found to have different feeding types, for example within Achiridae Trinectes 
maculatus is Sole-type and Achirus declivis is Turbot-type. Conversely, the same feeding type 
(Sole-type) was found in very different body forms, including the  highly elongate members of 
Soleidae and Cynoglossidae, as well as in Achiridae, whose members are extremely wide bodied. 
 
38  
 
 
 
A wide body could help protect against gape limited predators (recall the 
colloquial name for Trinectes maculatus, “Hogchoker”); it could also increase 
maneuverability in the water column or contribute to ground effect efficiencies when 
swimming above the substrate (Fox et al., submitted). Our character state reconstruction 
(Fig. 1.12) shows a single origin for Sole-type feeding, suggesting that these families 
diverged onto different morphological paths to being successful nocturnal benthic 
predator of small infaunal and epifaunal prey.  
In the diversity of natural systems there is not necessarily a one-to-one mapping 
of form to function (Collar and Wainwright, 2006), and flatfishes are no exception. 
Although eye and jaw size and position (Figs. 1.10 and 1.11) distinguish certain groups 
of flatfishes and correlate with feeding type, we also observed similarity in body forms 
between species with different diets and predation modes. This is evident from the 
overlapping of families with generally different feeding modes in the morphospace (Figs. 
1.6 and 1.7) and a close examination of diet and behavior data. For example, Achiris 
declivis and Trinectes maculatus are closely aligned on the far positive end of PC1 in the 
family Achiridae, with nearly identical body shape, fin shape, jaw and eye size and 
position, as well as similar maximum body size (Fig. 1.15). Trinectes maculatus is a 
Sole-type feeder that forages nocturnally and primarily consumes polychaetes, clam 
siphons, and amphipods (Derrick and Kennedy, 1997; Curti, 2005; VIMS 2020) while 
Achiris declivis was identified by Duarte and Andreata (2003) as a Turbot-type feeder 
following the Yazdani 1969 classification, because in the locality studied it consumed 
mostly fishes (Gobiidae) and was thought to perform ambush predation. Complicating 
simple interpretations is the fact that many flatfish species are opportunistic predators that 
39  
 
 
vary their diet depending on their life stage, location, season, and prey availability. In a 
different study, Achiris declivis was found to consume primarily polychaetes and 
crustaceans in the Summer season with polychaetes reaching 100% of stomach contents 
in mid-size individuals and crustaceans reaching 80% of stomach contents in large-size 
individuals (Couto and Farias, 2001). Synapturichthys kleinii (Soleidae) has a typical sole 
body and fin shape and clusters in the morphospace alongside other soleids such as Solea 
solea, which is a well-studied nocturnal predator (Kruuk 1963) that uses chemoreception 
and mechanoreception (Applebaum 1983) to consume mainly polychaetes (Allen 2005). 
In contrast, S. kleinii been classified as a Turbot-type feeder by Dallaville and Chanet 
(2009); because the behavior of this species is known only from the single record of 
ambush-type predation on small fishes it is difficult to draw firm conclusions.  
Our findings agree with the conclusion of Black and Berendzen (2020) that the 
largest component of body shape variation within flatfishes is body elongation, and we 
found similar patterns in family spread and clustering. Our landmark-based morphospace 
(Fig. 1.7) shows more differentiation among families than found by Black and Berendzen 
(2020; Fig. 1.2), likely because we included soft-tissue landmarks and made broader use 
of sliding semi-landmarks. Our study is the first to include fossils with intermediate body 
forms, as well as outgroups. Black and Berendzen (2020) concluded that more robust 
analyses are needed to determine the influence of ecological traits on body form. We 
agree and believe that current analyses are limited by data deficiencies for diet and 
unresolved phylogenetic relationships.  
40  
 
 
As the ecological roles, behavior, and interrelationships of the remarkable 
diversity of flatfishes become better known we look forward to a more nuanced and 
conclusive understanding of their evolution, form, and function. 
Acknowledgements 
Radiology services were provided by Sandra Raredon at the Smithsonian 
Institution, Radford Arrindell at the American Museum of Natural History, and Barb 
Catlin at the CU College for Veterinary Medicine. Erika Mudrake of the CU Statistical 
Consulting Unit and Mark Zifchock provided extremely helpful support for R. We thank 
Casey Dillman for guidance on geometric morphometric methods, Charles Dardia at the 
CUMV for collections support, and Stacy Farina for helpful advice. 
Support 
This work was supported by a Raney Fund Award from the American Society of 
Ichthyologists and Herpetologists, a Lerner-Gray Award for Marine Research from the 
American Museum of Natural History, and NSF GRFP.  
 
REFERENCES 
 
Adams, E. D., F. J. Rohlf, and D. E. Slice. 2013. A field comes of age: Geometric 
morphometrics in the 21st century. Hystrix. 24:7-14. 
Adams, D., Collyer, M., and A. Kaliontzopoulou. 2020. Geomorph: Software for 
geometric morphometric analyses. R package version 3.2.1.  
41  
 
 
Akkaynak, D., L. A. Siemann, A. Barbosa, and L. M. Mathger. 2017. Changeable 
camouflage: How well can flounder resemble the color and spatial scale of 
substrates in their natural habitats? Royal Society Open Science. 4, 160824. 
Argolo, L.A., R. T. C. Ramos, S. B. Barreto, J. A. Bitencourt, I. Sampaio, H. Schneider, 
and P. R. A. M. Affonso. 2018. The flounder next door: Closer evolutionary 
relationship between allopatric than sympatric Bothus (Rafinesque, 1810) species 
(Pleuronectiformes, Bothidae). Zoologischer Anzeiger 277; 131-142. 
Betancur-R, R., R. E. Broughton, E. O. Wiley, K. Carpenter, J. A. López, C. Li, N. I. 
Holcroft, D. Arcila, M. Sanciangco, J. C. Cureton II, F. Xhang, T. Buser, M. A. 
Campbell, J. A. Ballesteros, A. Roa-Varon, S. Willis, W. C. Borden, T. Rowley, 
P. C. Reneau, D. J. Hough, G. Lu, T. Grande, G. Arratia, and G. Ortí. 2013a. The 
Tree of Life and a new classification of bony Fishes. PLOS. doi: 
10.1371/currents.tol.53ba26640df0ccaee75bb165c8c26288 
Betancur-R., R., C. Li, T. A. Munroe, J. A. Ballesteros, and G. Ortí, G. 2013b. 
Addressing gene tree discordance and non-stationarity to resolve a multi-locus 
phylogeny of the flatfishes (Teleostei: Pleuronectiformes). Systematic Biology. 
62:763-785.  
Betancur-R, R. and G. Ortí. 2014. Molecular evidence for the monophyly of flatfishes 
(Carangimorpharia: Pleuronectiformes). Molecular Phylogenetics and Evolution 
73; 18-22. 
Black, C. R. and P. B. Berendzen. 2020. Shared ecological traits influence shape of the 
skeleton in flatfishes (Pleuronectiformes). PeerJ. 8:e8919 DOI 10.7717/peerj.8919 
42  
 
 
Bollback J. P. 2006. SIMMAP: Stochastic character mapping of discrete traits on 
phylogenies. BMC Bioinformatics. 7:88.  
Bookstein, F. L. 1997. Landmark methods for forms without landmarks: morphometrics 
of group differences in outline shape. 1:225-243. 
Brainerd, E.L. and S. N. Patek. 1998. Vertebral column morphology, C-start curvature, 
and the evolution of mechanical defenses in Tetraodontiform fishes. Copeia. 
4:971-984. 
Buser, T .J., B. L. Sidlauskas, and A. P. Summers. 2018. 2D or Not 2D? Testing the 
utility of 2D vs. 3D landmark data in geometric morphometrics of the Sculpin 
subfamily Oligocottinae (Pisces; Cottoidea). The Anatomical Record. 301: 806-
818 
Byrne, L., F. Chapleau, and S. Aris-Brosou. 2018. How the Central American seaway 
and an ancient northern passage affected flatfish diversification. Molecular 
Biology and Evolution. 35:1982-1989.  
Campbell, M. A., W. J. Chen, and J. A. Lopez. 2013. Are flatfishes (Pleuronectiformes) 
monophyletic? Molecular Phylogenetics and Evolution. 69; 664-673 
Chanet, B. 1994. Eubuglossus eocenicus (Woodward 1910) from the Upper Lutetian of 
Egypt, one of the oldest soleids (Teleostei, Pleuronectiformes). Neues Jahrbuch 
für Paläntologie Monatehefte. 7:391-398.  
Chanet, B. 1999. Supposed and true flatfishes [Teleostei: Pleuronectiformes] from the 
Eocene of Monte Bolca, Italy. 
43  
 
 
Chapleau, F. 1988. Comparative osteology and intergeneric relationships of the tongue 
soles (Pisces: Pleuronectiformes: Cynoglossidae). Canadian Journal of Zoology. 
66:1214-1232. 
Chapleau, F. and A. Keast. 1988. A phylogenetic reassessment of the monophyletic status 
of the family Soleidae, with comments on the suborder Soleoidei (Pisces; 
Pleuronectiformes). Canadian Journal of Zoology. 2792-2810.Chapleau, F. 1993. 
Pleuronectiform relationships: a cladistic reassessment. Bulletin of Marine 
Science. 52:516-540. 
Chapleau, F. 1993. Pleuronectiform relationships: a cladistic reassessment. Bulletin of 
Marine Science. 52:516-540. 
Claverie, T., and P. C. Wainwright. 2014. A morphospace for reef fishes: Elongation is 
the dominant axis of body shape evolution. PLoS ONE. 9:e112732. 
Collar, D. C., and P. C. Wainwright. 2006. Discordance between morphological and 
mechanical diversity in the feeding mechanism of centrarchid fishes. Evolution. 
60:2575-2584. 
Cooper, A. J. and F. Chapleau. 1998. Monophyly and intrarelationships of the family 
Pleuronectidae (Pleuronectiformes), with a revised classification. Fishery 
Bulletin. 96:686-726. 
Corral, W.D.R, and W. E. Aguirre. 2019. Effects of temperature and water turbulence on 
vertebral number and body shape in Astyanax mexicanus (Teleostei: Characidae). 
Phenotypic plasticity in response to temperature PLoS One. 2019;14:e0219677. 
Darwin, C. 1872. On the origin of species: By means of Natural Selection,6th Ed. The 
Floating Press. 
44  
 
 
DeGroot, S. J. 1969. Digestive system and sensorial factors in relation to the feeding 
behavior of flatfish (Pleuronectiformes). Conseil Permanent International pour 
L’Exploration de la Mer. 32:385-395. 
Eschmeyer, W. N., R. Fricke, R., and R. van der Laan. (Eds). 2020. Eschmeyer’s catalog 
of fishes: genera, species, references. California Academy of Sciences. Accessed 
online August 2020.  
Fox, C. H., A. C. Gibb, A. P. Summers, and W. E. Bemis. 2018. Benthic walking, 
bounding, and maneuvering in flatfishes (Pleuronectiformes: Pleuronectidae): 
new vertebrate gaits. Journal of Zoology. 130: 19-29. 
Friedman, M. 2008. The evolutionary origin of flatfish asymmetry. Nature Letters. 
454:209-212. doi:10.1038/nature07108 
Friedman, M. 2012. Osteology of †Heteronectes chaneti (Acanthomorpha, 
Pleuronectiformes), an Eocene stem flatfish, with a discussion of flatfish sister-
group relationships. Journal of Vertebrate Paleontology. 32:735-756. 
Friedman, M. and G. Carnevale. 2018. The Bolca Lagerstätten: Shallow marine life in the 
Eocene. Journal of the Geological Society. 175: 569-579. 
Froese, R. and D. Pauly (Eds.) 2000. FishBase 2000: concepts, design and data sources. 
ICLARM, Los Baños, Laguna, Philippines.  
Gibb, A. C. 2003. Modeling the jaw mechanism of Pleuronichthys verticalis: the 
morphological basis of asymmetrical jaw movements in a flatfish. Journal of 
Morphology. 256:1-12 
45  
 
 
Gibran, F.Z. and J. W. Armbruster. 2014. Dying or illness feigning: an unreported 
feeding tactic of the comb grouper Mycteroperca acutirostris (Serranidae) from 
the Southwest Atlantic. Copeia. 2004:403-405. 
Goldschmidt, R. 1933. Some aspects of evolution. Science. 78:539-547. 
Gosse, P. H. 1865. A Year at the Shore. Gutenberg E-book # 49022.  
Gower, J. C. 1975. Generalized Procrustes analysis. Psychometrika. 40:33-51. 
Harrington, R. C., B. C. Faircloth, R. I. Eytan, W. L. Smith, T. J. Near, M. E. Alfaro, and 
M. Friedman. 2016. Phylogenomic analysis of carangimorph fishes reveals 
flatfish asymmetry arose in a blink of the evolutionary eye. BMC Evolutionary 
Biology. 16:224 
Hensley, D. A. and E. H. Ahlstrom. 1984. Pleuronectiformes: relationships, in Ontogeny 
and Systematics of Fishes (eds. H.G. Moser, W.J. Richards, D. M. Cohen et al.) 
American Society of Ichthyologists and Herpetologists, Lawrence, KA. pp. 670-
687. 
Honshino, K. 2001. Monophyly of the Citharidae (Pleuronectoidei: Pleuronectiformes: 
Teleostei) with considerations on pleuronectid phylogeny. Ichthyological 
Research. 48: 391-404. 
Hubbs, C.L. 1922. Variations in the number of vertebrae and other meristic characters of 
fishes correlated with the temperature of water during development. American 
Naturalist. 56: 360-372.  
Kizhakudan, J. K., S. J. Kizhakudan, J. V. Xavier, S. Krishnamoorthi, K. S. S. M. 
Yousuf, R. Sundar, C. Manibal, R. Ponniah, D. Pakkiri, A. Janakiraman, S. 
Chandrasekharan, G. Chakrapani, and M. P. Chandrashekaran. 2013. Preliminary 
46  
 
 
observations on broodstock development and spawning of Indian Halibut 
Psettodes erumei (Block and Schneider, 1801) in captivity. Mar. Fish. Infor. Serv, 
T and E Ser. 218, 27-29. 
Kuiter, R. H. and T. Tonozuka. 2001. Pictorial guide to Indonesian reef fishes. Part 3. 
Jawfishes – Sunfishes, Opistognathidae - Molidae. Zoonetics, Australia. p. 623-
893. 
JMP, Version Pro 14. SAS Institute Inc, Cary, NC, 1989-2019. 
McKaye, K. R. 1981. Field observations on death feigning: a unique hunting behavior by 
the predatory cichlid, Haplochromis livingstoni, of Lake Malawi. Environmental 
Biology of Fishes. 6:361-365. 
McKee, A., I. MacDonald, S. C. Farina, and Summers, A.P. 2016. Undulation frequency  
affects burial performance in living and model flatfishes. Zoology. 119:75-80. 
Mivart, St. G. J. 1871. On the Genesis of Species. MacMillan, London. Gutenberg e-book 
#20818  
Munroe, T. 2015. Systematic diversity of the Pleuronectiformes. In: Gibson, R., Nash, R., 
Geffen, A., Veer, van der, John, H. (Eds.), Flatfishes: Biology and Exploitation, 
2nd ed. Wiley and Sons, New York, pp. 13–51.  
Near, T. J., R. I. Eytan, A. Dornburg, K. L. Kuhn, J. A. Moore, M. P. Davis, P. C. 
Wainwright, M. Friedman, and W. L. Smith. 2012. Resolution of ray-finned fish 
phylogeny and timing of diversification. PNAS. 109:13698-13703. 
Nelson, J. S., T. C. Grande, and M. V. H. Wilson. 2016. Fishes of the World, 5th edition. 
Wiley and Sons, Inc, Hoboken.  
Palmer, A. R. 2009. Animal asymmetry. Current Biology 19:474-477. 
47  
 
 
Paradis, E., J. Claude, and K. Strimmer. 2004. APE: Analyses of phylogenetics and 
evolution in R language. Bioinformatics. 20:289-290. 
Phelan, B. A., J. P. Manderson, A. W. Stoner, and A. J. Bejda. 2001. Size-related shifts in 
the habitat associations of young-of-the-year winter flounder (Pseudopleuronectes 
americanus): Field observations and laboratory experiments with sediments and 
prey. Journal of Experimental Marine Biology and Ecology. 257:297-315. 
Price, S.A., S. T. Friedman, K. A. Corn, C. M. Martinez, O. Larouche, and P. C. 
Wainwright. 2019. Building a body shape morphospace of teleostean fishes. 
Integrative and Comparative Biology. 59:716-730.  
Olla, B. L., R. Wicklund, and S. Wilk. 1969. Behavior of winter flounder in a natural 
habitat. American Fisheries Society. 98:717-718. 
Rabosky, D. L., J. Chang, P. O. Title, P. F. Cowman, L. Sallan, M. Friedman, K. 
Kaschner, C. Garilao, T. J. Near, M. Coll, and M. E. Alfaro. 2018. An inverse 
latitudinal gradient in speciation rate for marine fishes. Nature. 559:392-395. 
Ramachandran, V. S., C. W. Tyler, R. L. Gregory, and D. Rogers-Ramachandran. 1996. 
Rapid adaptive camouflage in tropical flounders. Nature. 379, 815-818. 
Ramos, R. T. C. 1998. Estudio filogenético da familía Achiridae (Teleostei: 
Pleuronectiformes: Pleuronectoidei), com a revisão das formas de água doce da 
América do Sul Cis-Andina e a reavaliação do monofiletismo de Soleomorpha 
(‘Soleoidei’). PhD thesis, Universidade de São Paulo.  
Revell, L. J. 2012. Phytools: An R package for phylogenetic comparative biology (and 
other things). Methods Ecol. Evol. 3:217-223 
48  
 
 
Ribeiro, E., A. M. Davis, R. A. Rivero-Vega, G. Ortí, and R. Betancur-R. 2018. Post-
Cretaceous bursts of evolution along the benthic-pelagic axis in marine fishes. 
Proceedings Royal Society B. 285:20182010. 
Rohlf, F. J. 2004. tpsDig, digitize landmarks and outlines, version 2.0. Department of 
Ecology and Evolution, State University of New York at Stony Brook. 
Sakamoto, K. 1984. Interrelationships of the family Pleuronectidae (Pisces: 
Pleuronectiformes). Memoirs of the Faculty of Fisheries Hokkaido University. 
31;95-215. 
Schreiber, A. M. 2005. Asymmetric craniofacial remodeling and lateralized behavior in 
larval flatfish. Journal of Experimental Biology. 209;610-621. 
Shi, W., S. Chen, X. Kong, L. Si, L. Gong, Y. Zhang, and H. Yu. 2018. Flatfish 
monophyly refereed by the relationship of Psettodes in Carangimorphariae. BMC 
Genomics. 19:400 
Shiödte, J. C. 1868. Flatfish eye development. Annals and Magazine of Natural History, 
including Zoology, Botany, and Geology. Vol. 1 Series Four. pgs. 378-383. 
Taylor and Francis, Ltd., London. 
Swain, D. P. 1992. Selective predation for vertebral phenotype in Gasterosteus aculeatus: 
reversal in the direction of selection at different larval sizes. Evolution. 46:998-
1013. 
Tanda, M. 1990. Studies on burying ability in sand and selection of the grain size for 
Hatchery-reared Marbled Sole and Japanese Flounder. Nippon Suisan Gakkaishi. 
56:1543-1548. 
49  
 
 
Tibbin, P., H. Berggren, O. Nordahl, P. Larsson and A. Forsman. 2016. Causes and 
consequences of intra-specific variation in vertebral number. Scientific Reports. 
6:26372 
Tobler, M. 2005. Feigning death in the Central American cichlid Parachromis 
friedrichsthalii. Journal of Fish Biology. 66:877-881. 
Voronina, E. P., V. G. Sideleva, and D. R. Hughes. 2019. Lateral line system of flatfishes 
(Pleuronectiformes): Diversity and taxonomic distribution of its characters. Acta 
Zoologica. 00:1-25. 
Wake, D. B. 2009. What salamanders have taught us about evolution. Annual Review of 
Ecology, Evolution, and systematics. 40:333-52. 
Ward, A. B. and E. L. Brainerd. 2007. Evolution of axial patterning in elongate fishes. 
Biological Journal of the Linnean Society. 90:97-116. 
Wirtz, P. and J. Davenport. 2017. Fin crawling in flatfishes (Teleostei: 
Pleuronectiformes). Bulletin of Fish Biology. 17:29-38. 
Yazdani, G. M. 1969. Adaptations in the jaws of flatfishes (Pleuronectiformes). Journal 
of Zoology, London. 159:181-222. 
 
Diet References 
Allen, B.M., P.A. King, and J.M. Fives. 2005. The Irish Naturalist's Journal. 28:162-165. Food 
consumed by Common Sole Solea solea (L.) in Galway Bay, on the West Coast of 
Ireland 
50  
 
 
Baeck, G.W., J.M. Park, and H. Hashimoto. 2011. Feeding ecology of three tonguefishes, 
genus Cynoglossus (Cynoglossidae) in the Seto Inland Sea, Japan. 15:325-336. 
Carpenter, K.E. and V.H. Niem (eds) FAO species identification guide for fishery purposes. 
The living marine resources of the Western Central Pacific. Volume 4. Bony fishes part 
2 (Mugilidae to Garangidae). Rome, FAO. 1999. pp. 2069-2790. 
Carpentieri, P., T. Cantarelli, F. Colloca, A. Criscoli, and G. Ardizzone. 2010. Feeding 
behavior and daily ration of the spotted flounder Citharus linguatula (Osteichthyes: 
Pleuronectiformes) in the central Tyrrhenian Sea. Scientia Marina. 74:659-667.  
Chang, S., P.L. Berrien, D.L. Johnson, and W.W. Morse. 1999. Windowpane, Scophthalmus 
aquosus, life history and habitat characteristics. NOAA Technical Memorandum 
NMFS-NE-137 
Couto, E.C.G., and M.C.V. Farias. 2001. Feeding habits of Sole fish (Achirus declivis 
(Chabanaud, 1940) (Teleostei: Soleidae)) in the River Sergipe estuary (Northeastern, 
Brazil). XII Coloquio Argentino di Oceanografia, Mar Del Plata, IB, 101p.Curti, K. L. 
(2005). Patterns in the distribution, diet and trophic demand of the hogchoker, 
Trinectes maculatus, in the Chesapeake Bay, USA. University of Maryland, College 
Park, ProQuest Dissertations Publishing. 
Dallaville, G. and B. Chanet. 2009. New data on the biology of Klein's sole Synapturichthys 
kleinii (Pleuronectiformes: Soleidae). Marine Biodiversity Records. 
doi:10.1017/S1755267209000876; Vol. 2; e70 
Derrick, P.A., and V.S. Kennedy. 1997. Prey selection by the hogchoker, Trinectes maculatus 
(Pisces: Soleidae), along summer salinity gradients in Chesapeake Bay, USA. Marine 
Biology. 129:699-711. 
51  
 
 
Duarte, G.A.S., and J.V. Andreata. 2003. Feeding habits of Achiridae and Cynoglossidae 
species from Ribeira Bay, Angra dos Reis, Rio de Janeiro, Brazil. Bioikos, PUC-
Campinas, 17:39-48.  
Earl, J. and S. Dittman. 2011. Variation in the diet of greenback flounder (Rhomobosolea 
tapirina) affected by freshwater inflows in Australia's largest temperate estuary. 
Conference paper based on PhD project: Population biology and ecology of the 
greenback flounder (Rhombosolea tapirina) in the Coorong estuary, south Australia. 
This reference is incomplete. 
Evseenko, S.A. 2004. Family Pleuronectidae (Cuvier 1816) righteye flounders. California 
Academy of Sciences Annotated Checklists of Fishes Number 37. 
Fischer, W., I. Sousa, C. Silva, A. de Freitas, J.M. Poutiers, W. Schneider, T.C. Borges, J.P. 
Feral, and A. Massinga. 1990. Fichas FAO de identificaçao de espécies para 
actividades de pesca. Guia de campo das espécies comerciais marinhas e de águas 
salobras de Moçambique. Publicaçao preparada em collaboraçao com o Instituto de 
Investigaçao Pesquiera de Moçambique, com financiamento do Projecto PNUD/FAO 
MOZ/86/030 e de NORAD. Roma, FAO. 1990. 424 p. 
Gloeckner, D.R. and J.J. Luczkovich. 2008. Experimental assessment of trophic impacts from 
a network model of a seagrass ecosystem: Direct and indirect effects of gulf flounder, 
spot and pinfish on benthic polychaetes. Journal of Experimental Marine Biology and 
Ecology. 357:109-120. 
Holmes, R.A. and R.N. Gibson. 1983. A Comparison of predatory behavior in flatfish. Animal 
Behavior. 31:1244-1255. 
52  
 
 
Kizhakudan, J.K., S.J. Kizhakudan, J.V. Xavier, S. Krishnamoorthi, K.S.S.M.Yousuf, R. 
Sundar, C. Manibal, R. Pohhiah, D. Pakkiri, A. Janakiraman, S. Chandrasekharan, G. 
Chakrrapani, and M.P. Chandrachekaran. (2013). Preliminary observations on 
broodstock development and spawning of Indian Halibut Psettodes erumei (Block and 
Schneider, 1801) in captivity. Marine Fisheries Information Service, Technical and 
Extension Series. 218, 27-29. 
Kruuk, H. 1963. Diurnal periodicity in the activity of the common sole, Solea vulgaris 
quensel. Netherlands Journal of Sea Research. 2:1-28. 
Kuiter, R.H. and T. Tonozuka. 2001. Pictorial guide to Indonesian reef fishes. Part 3. 
Jawfishes - Sunfishes, Opistognathidae - Molidae. Zoonetics, Australia. p. 623-893. 
Livingston, M.E. 1987. Food resources use among five flatfish species (Pleuronectiformes) in 
Wellington Harbour, New Zealand. New Zealand Journal of Marine and Freshwater 
Research. 21:281-293. 
Maigret, J. and B. Ly. 1986. Les poissons de mer de Mauritanie. Science Nat., Compiègne. 
213 p. 
Lowe-McConnell, R. H. 1987. Ecological Studies in Tropical Fish Communities. Cambridge 
University Press. 382 pgs. 
Mikawa, M. 1953. Studies on the digestive system and feeding habits of bottom fishes in the 
north-easter sea area along the Pacific coast of Japan. II. Clidoderma asperriumum 
(Temminck et Schlegel) and Microstomus stelleri Schmidt. Bulletin of Tohoku 
Regional Fisheries Research Laboratory. 2:26-48. 
53  
 
 
Nakane, Y., Y. Suda, and M. Sano. 2011. Food habits of fishes on an exposed sandy beach at 
Fukiagehama, South-West Kyushu Island, Japan. Helgoland Marine Resources. 
65:123-131 
NOAA Fisheries. Groundfish diet composition for Southern rock sole, Lepidopsetta bilineata 
from Aleutian Islands, Bering Sea, and Gulf of Alaska. NOAA Fisheries: Alaska 
Fisheries Science Center. Accessed 08/2020. 
Olla, B.L., C.E. Samet, and A.L. Strudholme. 1972. Activity and feeding behavior of the 
summer flounder (Paralichthys dentatus) under controlled laboratory conditions. 
Fishery Bulletin. 70:1127 
Oñate, F.C. 1990. Food and daily ration of the rock sole Lepidopsetta bilineata 
(Pleuronectidae) in the eastern Bering Sea. Marine Biology. 108:185-191.  
Pearcy, W.G. and D. Hancock. 1978. Feeding habits of Dover Sole, Microstomus pacificus; 
Rex Sole, Glyptocephalus zachirus; Slender Sole, Lyopsetta exilis; and Pacific 
Sanddab, Citharichthys sordidus, in a region of diverse sediments and bathymetry off 
Oregon. Fishery Bulletin. 76:1978. 
Randall, J.E. 1996. Caribbean Reef Fishes. Third Edition – revised and enlarged. T.F.H. 
Publications, Inc., Ltd., Hong Kong. 368 p. 
Redon M.J., M.S. Morte, and A. Sanz-Brau. 1994. Feeding habits of the spotted flounder 
Citharus linguatula off the eastern coast of Spain. Marine Biology. 120:197-201. 
Robertson, D.R. and J. Van Tassell. 2015. Fishes: Greater Caribbean a guide to the shorefishes 
of the Caribbean & adjacent areas. Smithsonian Tropical Research Institute. 
54  
 
 
Rodrigues, F.L., M.V. Condini, and C. Brito. 2018. The feeding habits of the endemic Remo 
flounder, Oncopterus darwinii (Actinopterygii: Pleuronectidae), in an exposed sandy 
beach's surf zone in southern Brazil. Zoologia. 25:e15038.  
Santic, M., M. Podvinski, A. Pallaoro, I. Jardas, and M. Kirincic. 2009. Feeding habits of 
megrim, Lepidorhombus whiffiagonis Walbaum, 1792), from the central Adriatic Sea. 
Journal of Applied Ichthyology. 25:417-422.  
Sedberry, G.R. 1983. Food habits and trophic relationships of a community of fishes on the 
outer continental shelf. NOAA Technical Report NMFS SSRF-773. Pg 8. 
Sommer, C., W. Schneider, and J.M. Poutiers. 1996. FAO species identification field guide for 
fishery purposes: the living marine resources of Somalia. FAO, Rome. 383 pp. 
Staudinger, M.D. and F. Juanes. 2009. Feeding tactics of a behaviorally plastic predator, 
summer flounder (Paralichthys dentatus). Journal of Sea Research. 64:68-75.  
Stehlik, L.L. and C.J. Meise. 2000. Diet of Winter flounder in a New Jersey estuary: 
Ontogenetic change and spatial variation. Estuaries. 23:381-391.  
Stickney, R.R. 1976. Food habits of Georgia Estuarine Fishes II. Symphurus plagiusa 
(Pleuronectiformes: Cynoglossidae). Transactions of the American Fisheries Society. 
105:202-207.  
Tokranov, A.M. and A.M. Orlov. 2011.Some biological features of roughscale sole 
Clidoderma asperrimum (Temminck et Schelegel, 1846) in the Pacific waters off the 
Northern Kuril Islands and Southeastern Kamchatka. DOI: 
10.1109/OCEANS.2001.968229.  
VIMS Multispecies Research Group. 2020, March 3. Fish food habits online data base. 
Retrieved from http://www.vims.edu/fisheries/fishfood. 
55  
 
 
Vinagre, C., A. Silva, M. Lara, and H.N. Cabral. 2011. Diet and niche overlap of southern 
populations of brill Scophthalmus rhombus and turbot Scophthalmus maximus. Journal 
of Fish Biology. 79:1383-1391. 
56  
 
 
CHAPTER 2 
 
BENTHIC WALKING, BOUNDING, AND MANEUVERING IN FLATFISHES 
(PLEURONECTIFORMES: PLEURONECTIDAE): NEW VERTIBRATE GAITS  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1 Claire H. Fox, Alice C. Gibb, Adam P. Summers, and William E. Bemis. (2018). Zoology 
130:19-29. 
 
5 7 
 
 
Abstract 
Video-based observations of voluntary movements reveal that six species of pleuronectid 
flatfishes use sequential portions of long-based dorsal and anal fins as “feet” (hereafter, 
fin-feet) to move on the substrate. All six species used a gait that we term “walking,” 
which produced constant forward movement, and several of these species also used a 
second gait that we call “bounding” for intermittent movements over the substrate. We 
selected Pacific Sand Sole, Psettichthys melanostictus, and English Sole, Parophrys 
vetulus, for kinematic analyses of these two gaits. Psettichthys melanostictus consistently 
used walking for benthic locomotion; Parophrys vetulus primarily used a bounding gait. 
During forward walking, a fin ray swings up off the substrate, protracts and converges 
with neighboring fin rays to contribute to a fin-foot. The fin-foot pushes down on the 
substrate and rotates posteriorly by sequential recruitment of fin rays, a pattern known as 
a metachronal wave. As one fin-foot passes off the posterior end of the fin, a new fin-foot 
forms anteriorly. During bounding, undulations of the body and tail assist one or two 
waves of fin-feet, producing rapid but intermittent forward acceleration of the body. 
Flatfishes also use fin-feet to maneuver on the substrate. The Starry Flounder, Platichthys 
stellatus, performs near zero displacement rotation by running waves of fin-feet in 
opposing directions along the dorsal and anal fins. Although other teleosts use specialized 
pectoral fin rays for bottom walking (e.g., Sea Robins: Triglidae), the duplication of 
structures and patterns of movement in the median fins of flatfishes more closely 
resembles metachronal motions of millipede feet or the parapodia of polychaete worms. 
Sequential use of median fin rays in flatfishes resembles that of other teleosts that swim 
58  
 
with elongate median fins, including Amiiformes, Gymnotiformes, and some 
Tetraodontiformes, but flatfishes offer a novel form of substrate locomotion based on 
dorsal and anal fins. 
1. Introduction 
Fishes as diverse as epaulette sharks (Hernandez et al., 2016), batoids (Koester 
and Spirito, 2003; Macesic and Kaijura, 2010; Macesic et al., 2013), lungfishes (King et 
al., 2011), sea robins (Jamon et al., 2007) and batfishes (Ward, 2002) walk on the 
substrate using paired appendages. Benthic walking may minimize pressure waves that 
could alert predators or potential prey to a fish’s movements (Macesic and Kaijura, 2010; 
King et al., 2011). For fishes in fast-flowing or turbulent currents, such as the waterfall-
climbing cave fish Cryptotora thamicola, walking enables the fish to adhere to the 
substrate (Flammang et al., 2016). Predatory benthic fishes may benefit from pushing 
against the substrate to produce a forward lunge or a rapid rotation when closing on a 
potential prey item. 
Metamorphosed flatfishes (Pleuronectiformes) — including flounders, halibuts, 
sole, and plaice — are unusual within vertebrates because their median fins are in direct 
contact with the substrate. Bauplan remodeling and reorientation of the eyes during early 
development in pleuronectiforms results in an exceptionally asymmetrical body and 
cranium. As adults, flatfishes lie on one side on the bottom and survey the surrounding 
environment with eyes that protrude from their upward-facing side. When adult flatfishes 
rest on the substrate, the dorsal and anal fins are positioned such that they can interact 
with the substrate — a functional task usually assumed by vertebrate paired appendages. 
Flatfishes are a highly successful and abundant group of bottom-dwelling, ambush 
5 9 
 
predators, with 14 extant families, 123 genera, and nearly 800 species (Nelson et al., 
2016); additionally, many species are economically important as seafood and make up 
large portions of the benthic-living biomass in marine ecosystems (Munroe, 2015). 
There has been extensive research on the development (Schreiber, 2013) and 
evolutionary origin of cranial novelties in Pleuronectiformes (Friedman, 2008; 
Harrington et al., 2016). Specializations for benthic life include: modifications for 
feeding on or near the substrate (Gibb, 1995; Gibb, 1996; Gibb, 1997); mechanisms that 
enable flatfishes to adhere to and rapidly separate from substrates (Brainerd et al., 1997); 
adaptations to facilitate color changes in the skin to enhance crypsis (Ramachandran et 
al., 1996); and burying behaviors (McKee et al., 2017). Webb (2002) evaluated 
swimming performance of Plaice (Pleuronectes platessa) above the substrate, finding 
that they swim only at moderate to high speeds, use symmetrical fin-beats on the 
upstroke and downstroke, maintain a positive but variable tilt angle (inclination of the 
body plane), and benefit from ground effects when swimming close to the substrate. 
There are some qualitative observations of walking behavior of flatfishes (Orcutt, 1950; 
Kruuk, 1963; Olla et al., 1969; Stickney et al., 1973) but no kinematic analyses. 
We used high-speed video analyses of benthic locomotion in six species of 
flatfishes from the Pacific Northwest (all from the family Pleuronectidae) to address the 
following questions: Do flatfishes use their median fins to move along the substrate? If 
so, what patterns of fin movement are used to generate these behaviors? Do benthic gaits 
(patterns of fin movements) vary within and among species of flatfishes? How similar is 
the benthic walking of flatfishes to locomotor modes exhibited by other benthic fishes, 
the swimming gaits of midwater fishes with elongate median fins, and invertebrate 
60  
 
locomotor behaviors? Ultimately, we ask does benthic walking of flatfishes constitute an 
undescribed vertebrate gait? 
2.0. Methods 
2.1. Animal Collection and Care 
We collected flatfishes from the Puget Sound near Friday Harbor Laboratories 
(Friday Harbor, WA) using bottom trawls and beach seines in July and August 2016, held 
them in flow-through seawater tanks (11 – 13 °C), and fed them every 1 - 2 days on 
mysid shrimp (University of Washington IACUC protocol 4208-03). We video-recorded 
fishes moving along the substrate within four weeks of collection, and, after the study 
was complete, either released or euthanized them. 
2.2. Videography 
We made high-speed digital-video recordings of 31 individual flatfishes 
representing six species in the family Pleuronectidae: Pacific Sand Sole, Psettichthys 
melanostictus (N=7); English Sole, Parophrys vetulus (N=12); Starry Flounder, 
Platichthys stellatus (N=5); Slender Sole, Lyopsetta exilis (N=2); Butter Sole, Isopsetta 
isolepis (N=3); and Rock Sole, Lepidopsetta bilineata (N=2). Prior to videography, we 
photographed each fish from above with a scale-bar in the frame. Specimens ranged from 
6.3 cm – 32.2 cm TL. Five of these six species are very closely related (Roje, 2010) and 
two hybridize (P. vetulus and P. stellatus; Eschmeyer and Herald, 1983). 
During the trials, individuals were transferred to two tanks (61 cm x 33 cm x 43 
cm and 122 cm x 20.5 cm x 30.5 cm) for filming sessions lasting 10 - 30 min. Water in 
the tanks was completely replaced once per hour. The substrate was a transparent, 
smooth, acrylic plate elevated 4 cm off the tank bottom to allow videography from the 
6 1 
 
side to capture animal-substrate interactions. Locomotion was elicited by gently prodding 
fish with a soft plastic probe (46 cm X 0.9 cm). Two Panasonic Lumix DMC - FZ200 
video cameras (120 frames per second, fps, at a resolution of 1280 X 720 pixels) were 
used to record five to ten locomotor sequences per individual from above and side. 
Two species (Psettichthys melanostictus N=3, Parophrys vetulus N=7) also were 
filmed moving in a flow-through seawater tank with a fiberglass bottom (60 cm by 120 
cm). The video camera was positioned above the tank to record self-initiated locomotion 
in a less constrained space. Filming took place during feeding sessions to capture 
intrinsically motivated movements.  
We selected 67 sequences of benthic locomotion of Psettichthys melanostictus 
(N=6: 27 walking sequences (see Table 1) and 3 walking-swimming transition 
sequences) and Parophrys vetulus (N=9, 22 walking sequences (see Table 1) and 15 
bounding sequences (see Table 2)) for detailed analyses because of the number of 
individuals available and their use of distinct walking and bounding gaits. We also 
quantified rotation in one sequence for Parophrys vetulus (N =1) and three sequences for 
Platichthys stellatus (N=3).  
6 2 
 
Table 2.1. Analyses of Walking Speeds for Psettichthys melanostictus and Parophrys vetulus. 
Species Individual TL (cm) Sequences Average Average Wave 
analyzed Walking Speed Frequency (Hz) 
(cm/s) 
5 19.6 9 5.37 1.69 
16 15.1 4 4.01 2.02 
Psettichthys 18 8.8 4 3.73 1.62 
melanostictus 24 14.4 4 5.40 2.27 
30 12.2 3 4.60 1.52 
34 12.7 3 6.02 2.47 
2 16.3 3 4.68 2.24 
19 6.8 2 3.17 3.83 
Parophrys 25 12.8 2 3.95 2.20 
vetulus 28 9.2 8 3.65 2.86 
31 15.6 3 3.49 1.76 
32 6.3 4 3.53 2.69 
 
Table 2.2. Analyses of Bounding for Parophrys vetulus. 
Average 
Bounding 
Sequence Frequency Speed 
Individual TL Length (s) # bounds (bounds/s) (cm/s) 
19 6.84 29.2 5 0.23 1.35 
50.4 5 0.10 1.43 
2.0 1 0.51 5.55 
20 18.30 20.6 3 0.15 0.67 
7.4 3 0.40 3.30 
51.3 12 0.23 1.89 
25 12.80 10.5 6 0.57 2.87 
8.9 4 0.45 2.28 
28 9.20 2.0 2 1.01 0.81 
3.1 2 0.64 1.59 
46.4 16 0.34 1.89 
35 5.32 31.2 11 0.35 2.13 
30.0 11 0.37 1.88 
36.4 11 0.30 2.67 
37 10.21 
60.3 15 0.25 1.79 
63  
 
To interpret differences in substrate locomotion of flatfishes, we first had to 
characterize and clearly define locomotor modes; the first three of these modes are newly 
reported here. (1) Forward walking is based on convergence of fin-rays to form a “fin-
foot” in which fin rays angle toward one another; the convergence of fin rays is most 
visible when viewed from above as a darker region of the fin that propagates in a wave, 
from anterior to posterior; three or more sequential fin waves produce continuous forward 
movement. (2) Forward bounding is generated by one or two fin waves with fin-ray 
convergence; this produces intermittent bouts of movement with the body coming to a 
complete stop between successive bounds. We defined a single bound as a push by 
median fin rays against the bottom, which launches the body into a glide phase during 
which there is no contact with the bottom. At the end of the glide, the fish comes to rest 
with the tips of its median fin rays in contact with the bottom. (3) Rotation against the 
substrate refers to a change in direction generated by out-of-phase waves of fin-ray 
convergence along the dorsal and anal fins. (4) Swimming is rapid and continuous 
forward movement in the water column generated by undulations of the body and median 
fin surfaces; no fin-ray convergence occurs. A final locomotor behavior not studied here 
is burying, which can be distinguished from other locomotor modes because the body is 
not displaced along the substrate; instead, the body rapidly undulates to displace substrate 
particles to cover the fish.  
We studied locomotor preferences in a 7 min video of freely-moving Parophrys 
vetulus (N=7) and Psettichthys melanostictus (N=3) as they foraged in the flow-through 
seawater tank. We coded all movements during the 7 min period as walking, bounding, 
64  
 
rotating, or swimming (using the definitions given above), and recorded the frequency 
and duration of each of these four locomotor modes. 
For kinematic analyses of walking, bounding, and swimming, body velocity was 
calculated by auto tracking the eye in top-view video sequences using DLTdv5 digitizing 
tools for MATLAB (Hendrick, 2008); fin-foot velocity was measured using the same 
software in manual tracking mode. Each walking sequence used for kinematic analysis 
contained three or more successive fin-foot strides. 
To determine if movements of the dorsal and anal fins were symmetrical, we 
measured the angle formed by the midline and the leading edge of the fin-foot at the 
beginning, middle, and end of one walk cycle per individual (N = 10) using the angle tool 
in Fiji (Schneider et al., 2012). To calculate wavelength (= distance between successive 
fin-feet), top-view video frames showing the beginning and end of a wave were measured 
in Adobe Illustrator by finding the length of a spline that follows the curvature of the fin 
base between fin-foot peaks.  
To analyze rotation, we first observed metachronal waves of fin movement in top-
view video footage and characterized them as uni-directional or bi-directional. Waves 
traveling in the same direction in the dorsal and anal fins produce either forward or 
backward walking; waves traveling in opposite directions produce rotation. For a subset 
of top-view rotation sequences (N=3) the turning radius was calculated by selecting three 
frames from the beginning, middle and end of each rotation sequence, marking the 
centroid of the fish in each frame, drawing a circle that intercepts all three points, and 
calculating the radius.  
6 5 
 
For fin ray angle analyses, we imported every 8th video frame from representative 
locomotor sequences into Adobe Illustrator; the bases of dorsal and anal fins were then 
traced with a spline. Two custom scripts were used. The first script divided the splines for 
dorsal fin and anal fin as described above, and placed a straight vector line at each point. 
The distal end of the line was moved manually to align with the end of the fin-ray whose 
base was closest to the anchor point. The second script calculated the angles for each fin-
ray line. 
2.3. Statistics 
We used Microsoft Excel to calculate means, standard deviations, t-tests, and graph data. 
All values are reported with standard deviation. We used JMP Version 10 for a Wilcoxon 
two-sample test to compare gait preferences, a regression analysis, and an ANOVA to 
examine the bivariate fit of speed in BL/s as a function of wave frequency. 
3.0. Results 
3.1. Locomotor Preferences 
A 7-minute video of foraging in two species of flatfishes freely moving in the sea table 
revealed species preferences in locomotor modes (Fig. 2.1).  
6 6 
 
 
Figure 2.1. Percentage time spent walking, bounding, rotating, and swimming for two species of 
flatfishes, Psettichthys melanostictus (N=3) and Parophrys vetulus (N=3) during voluntary 
movements in seven-minute observation period. 
 
Psettichthys melanostictus (N=3) moved forward on the substrate exclusively using a 
walking gait, while Parophrys vetulus (N=7) moved forward by bounding and, less 
frequently, walking (Fig. 2.1). The percentage of time spent walking, swimming, and 
bounding differs between P. melanostictus and P. vetulus (p < 0.05 for each comparison, 
Wilcoxon two-sample tests), but species differences for rotation are not significant. 
3.2. Forward Walking 
Large anterior and posterior inclinations of fin rays occur during walking (Figs. 
2.2 and 2.3).  
67  
 
 
Figure 2.2. Frames from a high-speed video (120 fps) showing positions of dorsal fin rays during 
a benthic walking sequence of the Pacific Sand Sole, Psettichthys melanostictus. A single stride 
in a walk cycle is shown. 
 
An exemplar forward walking cycle begins with an individual fin ray held approximately 
perpendicular (90°) to the body (Fig. 2.2). The fin ray lifts away from the substrate, 
6 8 
 
toward the eyed side and into the water column. It moves anteriorly to converge with 
additional fin rays to form a fin-foot. When one fin ray starts to retract from its maximum 
anterior position, the fin ray immediately behind sweeps forward to reach its maximum 
anterior position. As it retracts posteriorly, a fin ray pushes down onto the substrate, 
causing forward movement of the body. During forward walking, more posterior rays are 
progressively recruited to the fin-foot, creating an anterior-to-posterior metachronal wave 
of fin-rays in contact with the substrate. In the sequence shown, a new fin-foot originates 
anteriorly just as the prior fin-foot reaches the posterior end of the fin (Fig 2.2. 0.00 s and 
0.53 s; see also Supplemental Video 1). The wavelength, measured as the distance 
between two successive fin-feet, is slightly shorter than the length of the anal fin. This 
results in continuous support and propulsion: as the new fin-foot forms, the prior fin-foot 
is still present at the posterior end of the fin. In this way, steady locomotor progress is 
maintained. The fin ray concludes its posterior stroke, then is again lifted from the 
substrate and rotated forward to begin the next cycle, describing during its complete cycle 
a movement like a spoke in a wheel (Fig. 2.3). We observed this basic pattern of fin ray 
movements in the dorsal and anal fins during forward walking in all six species of 
flatfishes studied (Supplemental Video 2). 
69  
 
 
Figure 2.3. Tracing of the tip of a single fin ray from the middle of the anal fin of Platichthys 
stellatus in side view during one stride cycle (120 fps). Dotted lines connect the position of the fin 
ray tip (red dots) to its base (black dot). Fin rays angle down from the fin base to contact the 
substrate. Beginning at time zero, the fin ray begins to lift off of the substrate. At 0.30 s, the fin 
ray again contacts the substrate. The distance moved in the horizontal (anterior-posterior) 
direction is greater than the vertical (lateral) distance moved between the substrate and the 
maximum elevation of the fin ray. 
 
Waves of fin-foot propulsion along the median fins are in phase and operate in 
parallel, producing a synchronous, symmetrical gait (Fig. 2.4). As measured here, perfect 
symmetry would result in a 90° angle between the mid-line and the span line connecting 
the anterior-most portion of each fin-foot. For 10 strides analyzed, the average angle near 
the start of a wave was 89.6 ± 2.3°, at mid-wave 89.6 ± 4.2°, and end-wave 88.9 ± 4.1°. 
In analyses of 24 walk cycles from four Psettichthys melanostictus (N=12) and four 
Parophrys vetulus (N=12) wavelengths (= distance between fin-feet) were as follows: 
dorsal fin P. melanostictus 0.47 ± 0.01 TL; dorsal fin of P. vetulus 0.49 ± 0.01 TL; anal 
fin P. melanostictus 0.46 ± 0.01 TL; and anal fin P. vetulus 0.47 ± 0.01 TL.  
70  
 
 
Figure 2.4. Fin rays in the dorsal and anal fin move symmetrically to produce steady forward 
locomotion. A. The anterior-most 35% of the dorsal fin of Psettichthys melanostictus, shaded in 
red, is not used in the propulsive wave. Fin regions active in the propulsive wave, shaded in 
purple, are equal in the dorsal and anal fins. B. Beginning of the wave mid-way down the body. 
C. Middle of the wave approximately ¾ of the way along the body. D. End of the wave at the 
posterior of the body. 
 
The dorsal and anal fins of all species studied are unequal in length and have 
unequal numbers of fin rays. Both fins end at the caudal peduncle, but the anterior end of 
71  
 
the dorsal fin is near the anterior margin of the eye, whereas the anterior end of the anal 
fin is posterior to the pelvic fin. In four Psettichthys melanostictus the anal fin averaged 
65% of the length of the dorsal fin, and in four Parophrys vetulus the anal fin averaged 
69% of the length of the dorsal fin. Presumably because of the mismatch in overall fin 
lengths, the anterior-most portion of the dorsal fin does not form fin-feet during walking 
or bounding. Rather, the wave of movement that produces the fin-foot starts posterior to 
the head, in the same position in the transverse plane as the anterior-most end of the anal 
fin. 
 
Figure 2.5. Undulations of the body during walking and swimming. Areas above and below the 
mid-sagittal line are shaded red. A. Video frame from a walking sequence of Psettichthys 
melanostictus showing no body undulation; 0.34 BL/s. B. Video frame from a walking sequence 
of Parophrys vetulus with little body undulation; 0.28 BL/s. C. Video frame from a swimming 
sequence of P. melanostictus with some body undulation; 0.66 BL/s. D. Video frame from a 
swimming sequence of P. melanostictus with strong body undulation; 1.13 BL/s.  
 
Walking requires little to no undulation of the body (Fig. 2.5A - B; also see 
Supplemental Video 1) whereas larger amplitude undulations are used in swimming (Fig. 
2.5C - D). The amount of body undulation during walking and swimming of Psettichthys 
melanostictus and Parophrys vetulus (4 swimming sequences for two species, N=3 
individuals; 7 walking sequences for two species, N=7 individuals) was significantly 
different (t-test p < 0.005).  
7 2 
 
 
3.3. Forward Bounding 
During voluntary movements, Parophrys vetulus preferentially used a second 
substrate-based gait, which we term bounding (Fig. 2.6; also see Supplemental Video 3). 
During bounding, one or two rapid waves of fin propulsion are followed by a glide, 
during which the fish is no longer in contact with the substrate and body and tail 
undulation contribute to forward thrust (Fig. 2.6). During the landing phase, the fin rays 
are protracted and deflected towards the substrate to brake the body. A distinct pause 
with no forward movement follows the landing phase, before another bound is initiated.  
73  
 
 
Figure 2.6. Frames from high-speed videos of Parophrys vetulus showing a bound. White arrows 
indicate areas of fin-ray contact with the substrate. Fin rays successively push against the 
substrate, lift the body, and propel it forward. An initial wave of propulsion against the substrate 
is followed by a single undulation of the tail and body before the fish lands on the substrate. 
There is a glide phase at 0.60 s where the body is not in contact with the substrate. 
 
Plots of locomotion in X-Y space highlight key differences between walking and 
bounding (Fig. 2.7). Walking is a continuous forward movement produced by successive 
waves of fin-feet (there are six waves in the sequence in Fig. 2.7A), each following 
immediately after the other. Bounding (Fig. 2.7B) involves intermittent forward 
movement with one to two waves of propulsion followed by distinctive short pauses 
74  
 
when the body is at a full stop; these pauses lasted 0.48 s – 3.61 s in the sequence shown 
in Figure 7B (also see Supplemental Video 3). Frequent changes in direction highlight the 
maneuverability typical of bounding behavior (Fig. 7B).  
 
 
Figure 2.7. Movements of two species of flatfish, Psettichthys melanostictus and Parophrys 
vetulus, as seen from above in X-Y space. Movement paths indicated with red dashed lines and 
arrows. A. Walking sequence of Psettichthys melanostictus (4.2 s total) with six successive waves 
and times noted. B. Bounding sequence of Parophrys vetulus (46.4 s total) with pauses (black 
dots) between bounds. Durations of pauses indicated with black numbers; direction of bounds 
indicated with red dashed lines, and durations of bounds indicated with red numbers. 
 
 
 
 
 
3.4. Maneuvering and Rotation 
75  
 
All six species used variations in the relative wavelength, duration, and direction 
of the waves of fin-feet along their dorsal and anal fins to produce rotating and pivoting 
maneuvers (N = 31 rotation sequences total). Figure 8 shows a locomotor sequence of 
Platichthys stellatus with fin-rays moving in opposite directions against the substrate on 
opposite sides of the body to achieve a near-zero turn rotation (Fig. 2.8).  
 
7 6 
 
Figure 2.8. Frames from high-speed videos of Platichthys stellatus showing rotation in place. 
White lines highlight the fin-ray angles. The fin rays push against the substrate in opposite 
directions to produce rotation. 
 
Analysis of three such rotations of Platichthys stellatus had an average turning radius of 
1.1 ± 0.4 cm. The blind-side pectoral fin of P. stellatus is near the center of mass, and 
appears to serve as a rotation point (Supplemental Video 4). Rotations with larger turning 
radii occurred when a wave of forward propulsion passed along one median fin, with the 
other median fin held stationary as a pivot point. Rotation combined with forward 
lunging occurred during predatory strikes. 
3.5 Comparisons of Locomotor Behaviors 
 
77  
 
Figure 2.9. Comparison of fin ray angles for three locomotor behaviors at 0.4 s frame sampling. 
Heavy solid black lines indicate the five dorsal fin rays and four anal fin rays; angles are reported 
for two focal fin rays in the dorsal fin and two focal fin rays in the anal fin. A. Fin ray angles 
during walking of Psettichthys melanostictus. Fin rays converge to form fin-feet (0.00-0.13 s) that 
rapidly progress backwards (0.26 s) and begin their return to the start condition (0.40 s). B. Fin 
ray angles during bounding of Parophrys vetulus. Fin rays converge to push away from the 
substrate (0.13 s) followed by a glide (0.26 s) and fin protraction (0.40 s) before initiating 
landing. C. Swimming in P. melanostictus. Fin rays show minimal convergence. 
 
The angles subtended by individual fin rays differ during three locomotor 
behaviors: bounding, walking, and swimming (Fig. 2.9). We focus here on analyses of 
the two dorsal fin rays and two anal fin rays nearest to the mid-point of the body because 
we could clearly measure them throughout all sequences and highlight the differences 
between the locomotor modes. To assess the ranges of fin-ray motion, we calculated the 
differences between the smallest angle (anterior inclination) and the largest angle 
(posterior declination) reached by the focal fin rays. The anterior to posterior range of 
fin-ray motion during walking was as great as 93° (i.e., the fin-ray moved from 41° 
anterior inclination to 134° posterior declination). The mean range of motion during 
walking = 74 ± 14° N=44; for bounding: mean = 37 ± 9° N=36; for swimming: mean = 
20 ± 7° N=44. The range of motion for walking was significantly greater than for 
bounding (t-test; p < 0.005), and significantly greater than for swimming (t-test; p < 
0.001). The range of motion for bounding was significantly greater than for swimming (t-
test; p < 0.001). 
To assess the convergence of fin rays, we calculated differences in angles 
between adjacent focal fin rays. For example, during the walking sequence shown in 2ure 
9B 0.13 s, the convergence of dorsal fin rays is 102° (i.e., 145° - 43°); this was the 
greatest fin ray convergence observed across all locomotor behaviors. Mean convergence 
angles were: walking 30 ± 39°; bounding -20 ± 19°; and swimming -15 ± 9°. 
78  
 
Convergence during walking differed significantly from convergence during swimming 
(t-test; p < 0.001).  
As in Figure 2.4, fin rays in the anterior portions of the dorsal fin do not 
contribute to the formation of fin-feet and show a limited range of motion in all four 
locomotor behaviors (Figs. 2.9 - 2.10). In particular, the anterior-most dorsal fin rays of 
Psettichthys melanostictus extend in a fixed position nearly perpendicularly from their 
insertion above the eye during walking and swimming (Fig. 2.9 B - C).  
Fin ray angles during rotation of Parophrys vetulus show a large range of motion 
and are asymmetrical in the dorsal and anal fins (Fig. 2.10).  
79  
 
 
Figure 2.10. Fin ray angles during a rotation of Parophrys vetulus at 0.4 s frame sampling. Heavy 
solid black lines indicate the five dorsal fin rays and four anal fin rays; angles are reported for two 
focal fin rays in the dorsal fin and two focal fin rays in the anal fin.  
 
The mean range of motion of the four focal fin rays during rotation was 65 ± 18° 
(N=36). The mid-body fin rays on dorsal and anal fins reach asymmetrical obtuse and 
acute angles at the beginning and end of the rotation maneuver (0.00 s 64° and 95°; 0.20 s 
112° and 32°; 0.27 s 117° and 45°; 0.47 s 65° and 98°) and are symmetrical in the middle 
of the rotation maneuver (0.07 s 74° and 75°; 0.40 s 101° and 89°). Posterior fin rays 
80  
 
parallel the position of mid-body fin rays as they reach maximum inclination and 
declination (0.27 s anal fin 45° and 43°; 0.33 s dorsal fin 115° and 115°) and show 
convergence in the middle of the rotation maneuver (0.20 s dorsal 48° and 112°; anal 98° 
and 42°).  
3.6 Speed Analyses 
Walking speed in BL/s increases significantly for both Psettichthys melanostictus 
and Parophrys vetulus as wave frequency increases (Fig. 2.11; F < 0.0001).  
0.7
Psettichthys melanostictus
Parophrys vetulus
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Wave Frequency (Hz)
 
Fig. 11. Speed BL per s as function of wave frequency
Fig9u0re m 2m.11 x.  S7p0e medm in BL/s as a function of wave frequency for Psettichthys melanostictus (N = 
27 sequences) and Parophrys vetulus (N = 22 sequences). 
 
We compared the speed of 106 fin-feet (54 for Psettichthys melanostictus, 52 for 
Parophrys vetulus) moving posteriorly during walking to the average forward speed of 
the body (Fig. 2.12). Average fin speed was greater than body speed with a slip ratio 
(ratio between forward speed and wave speed) of 0.53 ± 0.1 for P. melanostictus and 0.36 
± 0.1 for P. vetulus. Comparisons of bounding, walking, and swimming speeds show 
8 1 
Speed (BL/s)
 
significant differences between the two substrate-based benthic gaits and swimming (Fig. 
2.13).  
 
 
Figure 2.12. Average speed of body and median fins for Psettichthys melanostictus (N=3) and 
Parophrys vetulus (N=3). For each walking sequence analyzed 6 or 7 waves were included. Ratio 
between forward speed and fin speed is reported as Slip Ratio. 
 
Because of the pauses between each set of one or two waves, bounding is a slow 
locomotor mode, with Parophrys vetulus achieving an average bounding speed of 0.22 ± 
0.1 BL/sec. Walking was slightly faster than bounding, with an average walking speed in 
Psettichthys melanostictus of 0.34 ± 0.1 BL/s (27 sequences N=6) and average walking 
speed for P. vetulus of 0.39 ± 0.1 BL/s. In both species, swimming speeds were more 
than twice as fast as walking: 0.91 ± 0.3 BL/s for Psettichthys melanostictus and 0.96 ± 
0.3 BL/s for Parophrys vetulus.  
82  
 
 
Figure 2.13. Comparisons of speeds of bounding, walking, and swimming in Psettichthys 
melanostictus and Parophrys vetulus. Note that none of 67 locomotor sequences captured for P. 
melanostictus included bounding. For P. vetulus bounding, 15 sequences from 6 individuals were 
analyzed; for P. melanostictus walking, 27 sequences from 6 individuals were analyzed; for P. 
vetulus walking, 22 sequences from six individuals were analyzed; for P. melanostictus 
swimming, 6 sequences from 5 individuals were analyzed; for P. vetulus swimming, 3 sequences 
from 3 individuals were analyzed. 
 
Three transition sequences lasting 4.37 to 27.6 s in length for three individual 
Psettichthys melanostictus showed a smooth transition between walking and swimming 
and a significantly faster swimming speed than walking speed (p < 0.04). Average 
walking speed during transition sequences was 0.29 ± 0.1 BL/s, and average swimming 
speed was 0.96 ± 0.3 BL/s. 
4.0. Discussion 
Six species of pleuronectids studied use median fins to walk and maneuver in 
direct contact with the substrate using successive recruitment of fin rays to form a fin-
foot that travels down the body in a metachronal wave. This metachronal locomotion 
resembles the movements of feet of millipedes and parapodia of polychaetes (Sleigh and 
83  
 
Barlow, 1980; Hesselberg, 2007) and enables a previously undescribed mode of 
substrate-walking in vertebrates. Because transformation of body orientation during 
metamorphosis of flatfishes results in dorsal and anal fins positioned on the functional 
lateral edges of the fish, flatfishes use median appendages for substrate-based 
locomotion. Flatfishes walk on the substrate using fins that are not homologous to the 
paired appendages of other vertebrates, yet, like vertebrates that use paired appendages, 
flatfishes coordinate movements of their median fins to produce symmetrical gaits. 
We also found that similar body and fin morphologies can produce functional 
diversity, highlighting the importance of examining both morphology and behavior to 
assess locomotor capabilities. Flatfishes can use their unpaired fins to produce both 
walking and bounding gaits, and they also use the dorsal and anal fins to reorient their 
bodies during benthic locomotion. However, species of flatfishes with only subtle 
morphological differences prefer different locomotor modes. For example, Psettichthys 
melanostictus routinely used the tips of its median fin rays to walk with a continuous 
forward motion and minimal body undulation, reserving rapid and intermittent 
acceleration for direct attacks on prey. Parophrys vetulus, in contrast, foraged using a 
distinct bounding gait characterized by rapid and intermittent acceleration with pauses 
between bounds to reorient and arch the body upwards, effectively placing the eyes in a 
higher visual plane above the substrate. Although all six species studied exhibited 
lunging and gliding behaviors, none employed a bounding gait as consistently as P. 
vetulus. Bounding is a conspicuous and therefore potentially risky behavior depending on 
environmental conditions. For example, Lemke and Ryer (2006: 267) showed that 
juvenile P. vetulus are subject to greater predation in clear water than are two other 
84  
 
flatfishes studied (Pacific Halibut, Hippoglossus stenolepis, and Northern Rock Sole, 
Lepidopsetta polyxystra), a difference they attributed to the more conspicuous foraging 
behaviors of P. vetulus. When tested in turbid water that simulates the estuarine habitats 
to which juvenile P. vetulus recruit, the predation rate decreased and was comparable to 
those found for H. stenolepis and L. polyxystra in clear water. Subsequently, Ryer et al. 
(2012) reported a higher intrinsic growth rate for P. vetulus than for H. stenolepis or L. 
polyxystra, which may be linked to the more active foraging behavior of P. vetulus. Thus, 
the different gaits we observed in morphologically similar species may reflect trade-offs 
related to water clarity, foraging strategy, and intrinsic growth rates. 
How does benthic locomotion of flatfishes compare to that of other fishes that 
walk on the bottom? Batoids such as skates and rays punt along the substrate using their 
pelvic fins in a symmetrical gait, with or without assistance from their pectoral fins 
(Macesic et al., 2013); the ceratotrichia of chondrichthyans are incapable of the 
convergence observed in the fin rays that form the fin-feet of flatfishes. Lungfishes and 
epaulette sharks use diagonal sequence gaits, while sea robins move via sequential 
movements of three pairs of free (that is, not connected by integument) pectoral fin rays 
(Jamon et al., 2007). Like the median fins of flatfishes, the pectoral fin rays of sea robins 
move symmetrically, using successive rays for propulsion. However, unlike flatfishes, 
sea robins are functionally hexapods because they use only six propulsive rays and lack a 
membrane between the rays. It will be interesting to examine movements of fin-rays in 
other benthic fishes that have enlarged pectoral and pelvic fins, such as Hillstream 
Loaches (Balitoridae), to learn if they also use metachronal waves of fin rays to move on 
the substrate. 
85  
 
Metachronal waves that flatfishes use for substrate locomotion are both similar to, 
and different from, those seen in other teleosts that use elongate median fins for 
swimming. For example, the Bowfin (Amia calva) has 48 - 51 dorsal fin rays (Grande 
and Bemis, 1998) undulated laterally in waves to swim forward or backward; faster 
swimming speeds are produced by adding body and caudal fin undulations (Jagnandan 
and Sanford, 2013). The Black Ghost Knifefish, Apteronotus albifrons, has 
approximately 140 - 160 anal fin rays (Albert, 2001) and is capable of exceptionally rapid 
movements and maneuvers—forward, backward, up, and down—achieved by holding the 
body immobile and modulating wave direction and amplitude, as well as curving the fin 
rays (Ruiz-Torres, 2014). Balistiform locomotion of triggerfishes similarly involves a 
rigid body and undulation of median fins; specializations of the fin skeleton and 
associated inclinator and declinator muscles allow strong lateral movements of the fin 
rays (Sorenson, 2007). Metachronal waves of pleuronectid flatfishes differ in at least two 
ways: 1) the tips of the fin rays contact the substrate, differentially slowing and distorting 
the waveform; 2) fin rays converge to form a fin-foot, which requires both mobility and 
control in which individual fin rays can be protracted, retracted, inclined, or declined with 
respect to adjacent rays in a joystick-like rotation. 
The diverse body shapes and dorsal and anal fin morphology within 
Pleuronectiformes are undoubtedly suited for different locomotor functions, but 
relationships between form and function are yet unclear. It is likely that the range of body 
and fin morphologies exhibited by flatfishes reflects morphological trade-offs to facilitate 
walking, burying, and swimming behaviors. Members of the most basal lineage of 
flatfishes (Psettodidae) have fewer vertebrae and fin rays than more derived flatfishes and 
86  
 
swim vertically in the water column. There are anecdotal reports of enhanced body-
substrate adhering abilities in the disk-shaped American soles (Achiridae). Preliminary 
observations of benthic walking in tonguefishes (Cynoglossidae), which have elongate 
bodies, a diphycercal caudal fin, and uniformly short dorsal and anal fin-rays, suggest 
that they are crawling specialists. Investigating the morphology and locomotor 
capabilities of diverse flatfish groups will increase our understanding of functional 
diversity and evolutionary trade-offs in this group, and shed light on the potential 
locomotor capabilities of transitional flatfishes in the fossil record. 
5.0. Conclusion 
The convergence of flatfish fin rays to form a fin-foot that travels down the length 
of the body and pushes directly against the substrate is a previously undescribed 
locomotor mode in fishes. The anatomical and functional specializations that underlie 
benthic locomotion in flatfishes and the diversity of locomotor capabilities across flatfish 
phylogeny are poorly understood and warrant future investigation. To our knowledge, no 
other vertebrate uses non-paired appendages to produce a walking gait. Thus, benthic 
locomotion performed by flatfishes appears to be structurally and functionally unique 
among vertebrates. 
Acknowledgements 
This project started as part of the Functional Morphology of Fishes course at 
Friday Harbor Laboratories in summer 2016; many course participants provided early 
and helpful input. Amy McCune, John Hermanson, and Peter Wainwright offered 
comments on locomotion in other teleosts and gaits. Frances Lux, Isaiah Gutman, Betty 
McGuire, and Mark Zifchock assisted with data collection and analyses. Stephen Parry 
8 7 
 
offered statistical support. CHF was supported by NSF GRFP 2016231353 and a 
Wainwright Fellowship for study at FHL. APS and the fish class were supported by NSF 
grant IOS -1256602. Additional support from the Tontogany Creek Fund. 
 
REFERENCES 
Albert, J. S., 2001. Species diversity and phylogenetic systematics of American 
knifefishes (Gymnotiformes, Teleostei). Misc. Publ. Mus. Zool. Univ. Michigan 
190, 1–127. 
Brainerd, E., Page, B., Fish, F., 1997. Opercular jetting during fast-starts by flatfishes. J. 
Exp. Biol. 200, 1179-1188. 
Eschmeyer, W. N., Herald, E. S. 1983. A Field Guide to Pacific Coast Fishes of North 
America. (Peterson Field Guides). Houghton Mifflin, New York.  
Flammang, B. E., Suvarnaraksha, A., Markiewicz, J., Soares, D., 2016. Tetrapod-like 
pelvic girdle in a walking cavefish. Sci. Rep. 6, 23711. 
Friedman, M., 2008. The evolutionary origin of flatfish asymmetry. Nature 454, 209-212. 
Gibb, A., 1995. Kinematics of prey capture in a flatfish, Pleuronichthys verticalis. J. Exp. 
Biol.198, 1173-1183. 
Gibb, A., 1996. The kinematics of prey capture in Xystreurys liolepis: Do all flatfish feed 
asymmetrically? J. Exp. Biol. 199, 2269-2283. 
Gibb, A., 1997. Do flatfish feed like other fishes? A comparative study of percomorph 
prey-capture kinematics. J. Exp. Biol. 200, 2841–2859 
Grande, L. and Bemis, W. E., 1998. A Comprehensive Phylogenetic Study of Amiid 
Fishes (Amiidae) Based on Comparative Skeletal Anatomy. An Empirical Search 
88  
 
for Interconnected Patterns of Natural History. J. Vert. Paleontol. Special Memoir 
#4 (supplement to Vol. 18).  
Harrington, R. C., Faircloth, B. C., Eytan, R. I., Smith, R. I., Near, T. J., Alfaro, M. E., 
Friedman, M., 2016. Phylogenomic analysis of carangimorph fishes reveals 
flatfish asymmetry arose in a blink of the evolutionary eye. BMC Evol. Biol. 16, 
224.  
Hendrick, T. L., 2008. Software techniques for two and three dimensional kinematic 
measurements of biological and biomimetic systems. Bioinspir. and Biomim. 3, 
034001. 
Hernandez, A. V., Gervais, C. R., Rummer, J. L., Porter, M. E., 2016. Comparing 
submerged walking and swimming kinematics in epaulette sharks. Integr. Comp. 
Biol. 56, E301 P2. 185. 
Hesselberg, T., 2007. Biomimetics and the case of the remarkable ragworm. Sci. Nat. 94, 
613-621. 
Jagnandan, K., Sanford, C. P., 2013. Kinematics of ribbon-fin locomotion in the bowfin, 
Amia calva. J. Exp. Zool. 319A, 569-583. 
Jamon, M., Renous, S., Gasc, J. P., Bels, V., Davenport, J., 2007. Evidence of force 
exchanges during the six-legged walking of the bottom-dwelling fish, 
Cheidonichthys lucerna. J. Exp. Biol. 307A, 542-547. 
King, H., Shubin, N., Coats, M., Hale, M., 2011. Behavioral evidence of the evolution of 
walking and bounding before terrestriality in sarcopterygian fishes. Proc. Natl. 
Acad. Sci. U.S.A. 2011 108, 21146-21151. 
89  
 
Koester, D.M., Spirito, C.P., 2003. Punting: An unusual mode of locomotion in the little 
skate, Leucoraja erinacea (Chondrichthyes: Rajidae). Copeia 2003, 553-561. 
Kruuk, H., 1963. Diurnal periodicity in the activity of the common sole, Solea vulgaris 
Quensel. Neth. J. Sea Res. 2, 1-28. 
Lemke, J. L., Ryer, C. H., 2006. Relative predation vulnerability of three juvenile (Age-
0) North Pacific flatfish species: Possible influence of nursery-specific predation 
pressures. Mar. Ecol. Prog. Ser. 328, 267-273. 
Macesic, L. J., Kaijura, S. M., 2010. Comparative punting kinematics and pelvic fin 
musculature of benthic batoids. J. Morphol. 271, 1219-1228. 
Macesic, L.J., Mulvaney, D., Blevins, E.L., 2013. Synchronized swimming: Coordination 
of pelvic and pectoral fins during augmented punting by the freshwater stingray 
Potamotrygon orbigny. Zoology 116, 144-150. 
McKee, A., MacDonald, I., Farina, S., Summers, A., 2016. Undulation frequency affects 
burial performance in living and model flatfishes. Zoology 119, 75-80. 
Munroe, T. 2015. Systematic diversity of the Pleuronectiformes. Pages 13-51 in Ed. 
Gibson, R. Nash, R., Geffen, A., van der Veer, H. John. Flatfishes: Biology and 
Exploitation, 2nd Ed. Wiley and Sons, Ltd., New York. 
Nelson, J. S., Grande, T. C., Wilson, M. V. H., 2016. Fishes of the World, 5th Edition. 
Wiley and Sons, Inc., Hoboken. 
Olla, B. L., Wicklund, R., Wilk, S., 1969. Behavior of winter flounder in a natural 
habitat. Trans. Am. Fish. Soc. 98, 717-720. 
Orcutt, H. G., 1950. The life history of the Starry Flounder, Platichthys stellatus. Cal. 
Dept. Nat. Res. Div. Fish Game Bur. Mar. Fish. Fish Bull. No. 78, 1-68. 
9 0 
 
Ramachandran, V. S., Tyler, C. W., Gregory, R. L., Rogers-Ramachandran, D., 
Duensing, S., Pillsbury, C., Ramachandran, C., 1996. Rapid adaptive camouflage 
in tropical flounders. Nature 379, 815-818. 
Roje, D. M., 2010. Incorporating molecular phylogenetics with larval morphology while 
mitigating the effects of substitution saturation on phylogeny estimation: A new 
hypothesis of relationships for the flatfish family Pleuronectidae (Percomorpha: 
Pleuronectiformes). Mol. Phylogenet. Evol. 56, 586–600. 
Ruiz-Torres, R., Curet, O. M., Lauder, G. V., MacIver, M. A., 2014. Kinematics of the 
ribbon fin in hovering and swimming of the electric ghost knifefish. J. Exp. Biol. 
216, 823-824. 
Ryer, C. H., Boersma, K. S., Hurst, T. P., 2012. Growth and distributional correlates of 
behavior in three co-occurring juvenile flatfishes. Mar. Ecol. Prog. Ser. 460, 183-
193.  
Schneider, C. A.; Rasband, W. S., Eliceiri, K. W., 2012. NIH Image to ImageJ: 25 years 
of image analysis. Nat. Methods 9, 671-675. 
Schreiber, A. M., 2013. Flatfish: An asymmetric perspective on metamorphosis. Curr. 
Top. Dev. Biol. 103, 167-194. 
Sleigh, M. A., Barlow, D. I., 1980. Metachronism and control of locomotion in animals 
with many propulsive structures. Society for Experimental Biology Seminar 
Series: Volume 5, Aspects of Animal Movement. Ed. Elder, H.Y., Trueman, E.R. 
Cambridge University Press, Cambridge. 
91  
 
Sorenson, L. 2007. Examining morphological characteristics of triggerfish: A three-
dimensional reconstruction of skeletal and muscular features of Rhinecanthus 
rectangulus. J. Young Investig. 16:3.  
Stickney, R. R., White, D. B., Miller, D., 1973. Observations of fin use in relation to 
feeding and resting behavior in flatfishes (Pleuronectiformes). Copeia 1973, 154-
156. 
Ward A. B., 2002. Kinematics of the pectoral fins in batfishes (Ogcocephalidae) during 
aquatic walking. Integr. Comp. Biol. 42,1331–1331. 
Webb, P., 2002. Kinematics of Pleuronectes platessa, and Cod, Gadus morhua, 
swimming near the bottom. J. Exp. Biol. 205, 2125. 
  
92  
 
Supplemental Videos 
Supplemental Video 1. Side view of benthic walking in Psettichthys melanostictus. 
Supplemental Video 2. Top views of benthic walking sequences in six species of 
pleuronectid flatfishes. 
Supplemental Video 3. Bounding of Parophrys vetulus. 
Supplemental Video 4. Rotation of Platichthys stellatus. 
9 3 
 
CHAPTER 3 
 
STEALTH LOCOMOTION: HYDRODYNAMIC CRYPSIS IN A FLATFISH, THE 
HOGCHOKER (TRINECTES MACULATUS) 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1Claire H. Fox, Sunghwan Jung, Seungho Kim, and William E. Bemis 
 
94  
 
Abstract 
We found benthic walking in the Hogchoker, Trinectes maculatus (Pleuronectiformes: 
Achiridae) to be hydrodynamically cryptic – akin to “pressure camouflage” that enables 
the flatfish to move stealthily through its aquatic environment. Using Digital Particle 
Image Velocimetry methods we compared hydrodynamic flow in a 2-D longitudinal 
plane during swimming and walking sequences. Analyses of three flow regions around 
the fish (bow-wave, along the body, and in the wake) demonstrated that walking 
Hogchokers produce significantly lower mean velocity magnitude and vorticity in the 
surrounding water when walking than during swimming. Swimming Hogchokers 
produced thrust by undulating their body, generating counter-rotating proto-vortices on 
the upward and downward facing sides of their wide flat bodies that are shed into the 
wake and form a distinct and persistent hydrodynamic trail, visualized here in vorticity 
plots. Walking Hogchokers lie on their side on the bottom and use sequential portions of 
their dorsal and anal fins as “fin feet” to maneuver and move slowly along the bottom, 
with little to no body undulation. We found that fluid disturbance around the body and in 
the wake of a walking Hogchoker was comparatively small and left no persistent vortex 
signature. Vortex trails from swimming fishes are quickly identified and followed by 
many piscivorous predators. We interpret that hydrodynamic crypsis of the benthic-
walking Hogchoker is advantageous because it does not leave a wake that could be 
detected by a predator. Hydrodynamic crypsis might also facilitate location and capture 
of epibenthic and infaunal prey.  
 
 
9 5 
 
Introduction 
Visual crypsis is a familiar and well-explored concept, whereas hydrodynamic 
crypsis, of great importance in aquatic environments, is comparatively understudied 
(Ruxton, 2011). Swimming fishes produce turbulent hydrodynamic signals that can 
inform and alert potential predators and prey of their location, size, and direction of 
travel. The turbulent trail left behind a goldfish is detectable in the water up to five 
minutes after its passage (Hanke et al., 2000) and distance, speed, and size of the fish can 
be derived from the decay of vorticity, velocity gradient, and width of track, respectively. 
Different fishes (e.g., Lepomis, Colomesus, and Thysochromis) have distinct wake 
signatures that last from 30 s to >5 min producing water velocities detectable by 
piscivorous predators and at distances where vision and hearing often fail (Hanke and 
Bleckmann, 2004). Marine mammals and piscivorous fishes use these hydrodynamic 
signals to locate and evaluate potential prey. Harbor seals can detect vortex trails of 
swimming fishes and respiratory jets of flatfishes camouflaged on the substrate, rapidly 
locating the sources of these hydrodynamic stimuli without additional visual or chemical 
cues (Wieskotten et al., 2010; Niesterok et al., 2017). The specialized vibrissae of harbor 
seals can distinguish between hydrodynamic trails generated by objects of different sizes 
and shapes (Wieskotten et al., 2011). Fishes can sense even weak hydrodynamic changes 
using superficial neuromasts on their body surface and neuromasts in lateral line canals 
(Bleckmann and Zelick, 2009; Schwalbe et al., 2012). Goldfish discriminate between 
hydrodynamic signals that vary in speed, direction, size, and shape using their lateral line 
system (Vogel and Bleckmann, 2001). Silurus glanis, a nocturnal piscivorous catfish, 
detects and follows a series of past locations of prey fishes by tracking their wakes, 
96  
 
succeeding in prey capture using only their hydrodynamic sense (Pohlmann et al., 2004). 
Even fishes that are primarily visual predators integrate sensory information from both 
canal and superficial neuromast cells in localizing prey, resulting in more direct search 
paths and extending the area within which prey can be detected (Montgomery et al., 
2002). 
In addition to generating a turbulent wake that predators can track, swimming 
fishes also produce a bow-wave that can alert prey to their approach. Small fishes can 
respond to flow stimulus of a predator in less than 4 ms (Liu and Fetcho, 1999) compared 
to the 200 ms or longer that it takes them to respond to a visual stimulus (Burgess and 
Granato, 2007), and, using only hydrodynamic stimuli, respond with a fast start within 2 
cm of the predator’s approach (Stewart et al., 2014).  
Flatfishes, the subject of our study, have been characterized as belonging to one 
of three major feeding groups (Yazdani, 1969; DeGroot, 1969, 1971): 1) Visual feeders 
that generally take free-swimming food, e.g., other fishes; 2) Visual feeders that mainly 
take bottom-living and slow-moving food, e.g., mollusks, polychaetes; 3) Nocturnal 
feeders that strictly take bottom-living food, e.g. polychaetes. There is evidence that stem 
flatfishes were piscivorous (Friedman, 2008), and species of Psettodes, the basal-most 
extant genus of Pleuronectiformes, belongs to feeding group 1(they are ambush predators 
that eat other fishes, Kizhakudan et al., 2013). In contrast, the model flatfish chosen for 
our study, the Hogchoker (Pleuronectiformes; Achiridae: Trinectes maculatus shown in 
walking posture in Figure 2.1) is in feeding group 3, with derived morphological and 
behavioral specializations for nocturnal predation on infaunal and epifaunal organisms. 
Hogchokers exhibit both opportunistic and selective feeding patterns with primary prey 
97  
 
items being polychaetes, followed by siphons of tellinid clams, and amphipods (Derrick 
and Kennedy, 1997; Curti, 2005; VIMS Multispecies Research Group, 2020). 
Flatfishes (Pleuronectiformes) have exemplary visual crypsis, achieving adaptive 
surface marking changes as rapidly as 2-8 s (Ramachandran et al., 1996) to match the 
luminance, color contrast, and spatial scales of a range of natural substrates (Akkaynak, 
2017). Because flatfishes undergo a unique axial transformation during metamorphosis, 
the side-to-side body undulation of typical teleosts becomes oriented up-and-down. When 
inactive, flatfishes bury to hide in soft-bottom substrates (Fig. 2.2B) (Ellis et al., 1997). A 
unique walking gait along the substrate (Fox et al., 2018) entails little to no body 
undulation, which begs the question: is flatfish walking a form of hydrodynamic crypsis? 
We investigated this question using Digital Particle Image Velocimetry (DPIV) analyses 
of the hydrodynamics of the Hogchoker, Trinectes maculatus (Pleuronectiformes: 
Achiridae), a small, estuarine sole abundant along the Atlantic Coast of North America.  
2.0. Methods 
2.1. Animal Collection and Care 
We chose Trinectes maculatus as a model flatfish for this study because of its 
relatively small adult size, availability, and tolerance of a wide range of salinity (a factor 
in water density). For DPIV analyses conducted at Cornell University we ordered six 
specimens of Trinectes maculatus from Gulf Marine Specimen Marine Lab, held them in 
recirculating tanks (15 – 18 °C, salinity 20 ppt) equipped with biological filtration units, a 
sand substrate, 12:12 day/night cycle, and fed them daily on bloodworms (Cornell 
University IACUC protocol 2017-0010). We video recorded individuals moving along 
the substrate and after the study was complete, adopted out the animals. Individuals spent 
98  
 
most of their time inactive and cryptic, buried under the sand, and instinctively perform 
burying movements even when substrate was not present. 
For a locomotor studies of Trinectes maculatus conducted at the Virginia Institute 
of Marine Science (VIMS) in June 2017 specimens were collected via bottom-trawl 
during the monthly Chesapeake Bay survey conducted by VIMS. Fish were transported 
in aerated coolers using water collected from the capture site, held for five days in the 
VIMS Seawater Research Laboratory in recirculating tanks supplied with low-salinity 
water from the estuary and fed daily on bloodworms (VIMS protocol IACUC-2017-04-
14-12050-ejhilton). After study, fish were released near their collection location. 
2.2. Videography 
Prior to videography, we photographed each fish from above with a scale-bar in 
the frame. We captured high-speed digital-video recordings of Trinectes maculatus 
walking and swimming. We video recorded sequences for locomotor analyses from the 
top and side using a GoPro Hero4 at 60 fps at a resolution 1280 X 720 pixels. DPIV 
videography was performed with a Photron Fastcam SA-Z camera (Photron USA, Inc., 
San Diego, CA, USA), Nikon 105mm 1:2.8 D lens with manual focus and f-stop 22, at a 
shutter speed of 1000 s-1, record rate 500 fps, resolution 1024 X 1024 pixels. For 
illumination of particles during DPIV we used an array of three laser modules generating 
150 mW green line (532 nm Diode-Pumped Solid State laser with integrated cylindrical 
lens optics and aperture angle of 90° manufactured by Apinex Inc., Durham, NC, USA). 
For particles we used hollow glass spheres with particle diameter = 10 m, s.g. = 1.05 
(catalog # 1108952 Lavision Inc., Ypsilanti, MI, USA). 
9 9 
 
 
Figure 3.1 DPIV experimental set-up. Three green lasers with built-in light sheet optics were 
aligned with custom holders to point down into a still-water tank seeded with 10 m hollow glass 
spheres, producing an illuminated plane that bisects the fish longitudinally when it moves left or 
right down the center of the long axis of the tank. The Photron Fastcam SA-Z camera is located in 
front of the tank with a 8.9cm X 8.9cm capture window positioned in the middle region, away 
from the tank side walls and just above the substrate. The substrate consists of fine sand 
embedded in a thin layer of aquarium-grade silicone adhered to an acrylic sheet. 
 
During filming sessions lasting 10-30 min, individuals were transferred to a 
filming tank (76.2 cm x 30.5 cm x 45.7 cm) with non-circulating water of the same 
temperature and salinity as their holding tank. Still-water was used to permit visualization 
of weak flow structures (Hank and Bleckmann, 2004). A challenge in working with still 
water videography environments is dealing with residual water motion in the tank 
generated, for example, by previous swimming behaviors. As much as possible we 
selected video sequences with a minimum of residual water movement from prior 
locomotion and respiratory jets from gill ventilation. Wall interactions were minimized 
by using a tank over five times the width and eight times the length of the animal, 
situating the camera to capture movement in the center region of the tank, and centering 
10 0 
 
the laser line down the center of the long tank axis (Figs 3.1 and 3.2). The substrate was 
an acrylic plate with a smooth layer of aquarium silicon with fine sand embedded to 
provide traction, elevated 4 cm off the tank bottom to allow videography from the side to 
capture animal-substrate interactions. Locomotion was elicited by gently prodding fish 
with a flexible plastic or metal probe. Photographs for Figure 1 were taken using a 
Google Pixel 3 phone camera with 12.2 Mp, 1/2.55” dual-pixel sensor and f/1.8 aperture 
lens. We adjusted levels, contrast, and tone in Adobe Photoshop and digitally removed 
the background from Fig. 3.3A.  
 
Figure 3.2. Orientation of Trinectes maculatus during DPIV experiments and 
example capture frames. 
10 1 
 
2.3. Locomotor and DPIV Analyses 
We calculated body velocity during walking and swimming by auto tracking the 
eye using DLTdv5 digitizing tools for MATLAB (Hedrick, 2008). Each walking 
sequence used for kinematic analysis contained three or more successive fin-foot strides 
(Fox et al. 2018). 
Digital Particle Image Velocimetry analyses were performed using PIVlab 2.31 
and MATLAB R2018a. We pre-processed video frames in Tiff format with the “sharpen” 
filter in ImageJ.  
In PIVlab, we used interrogation window sizes of 128 pixels in pass 1 and 64 
pixels in pass 2 for full-frame analyses, and 64 pixels in pass 1 and 32 pixels in pass 2 for 
analysis of smaller flow region (i.e., bow-waves). We used an inter-frame temporal 
spacing of 2 ms – 4 ms to measure the velocity field, with swimming sequences requiring 
a higher frame sampling than walking sequences. Erroneous vectors (outliers) were 
eliminated, when necessary, by selecting velocity limits that excluded them and 
interpolating the missing vectors. We applied a low level of data smoothing. Vector fields 
and derivatives (magnitude, vorticity) were plotted in PIVlab to produce Figs. 3.4-3.7 and 
were exported in .csv format and imported into Excel to produce Table 1. 
2.4. Statistics 
We used Mann-Whitney U-Tests to compare velocity magnitude and absolute 
vorticity during swimming and walking for three body regions to estimate bow wave, 
flow regimes along the body, and flow regimes of the wake. 
3.0. Results 
3.1. Observations on Functional Morphology 
10 2 
 
Fin-feet form during walking when successive fin-rays converge together at the 
tip (Fig. 3.3A). In periods of inactivity the Hogchoker lies buried completely under the 
sand or mud with only a cluster of small features protruding above the substrate: two 
small eyes, an upward-facing nostril tube, and fringed lip edge (Fig. 3.3B). The family 
Achiridae is named for “achirine lines”, accessory branches of the lateral line that extend 
at right angles across the body surface toward the dorsal and anal fins, increasing the 
mechanosensory capacity of the eyed side (Fig. 3.3C). Fimbriae, fleshy projections of the 
epidermis, cover the blind side of the head from the operculum to the anterior margin, 
and the anterior dorsal and anal fin rays (Fig. 3.3D). 
10 3 
 
 
Figure 3.3. Morphology features of Trinectes maculatus for walking on the bottom, avoiding 
detection by predators, and sensing prey. (A). Eyed side during benthic walking. Dorsal and 
anal fin rays converge into symmetrical fin-feet that push against the substrate. Note camouflaged 
body. (B). Cryptic appearance when buried in soft bottom sediments with only the eyes and 
mouth protruding. C. Accessory branches of the lateral line extend the mechanosensory system. 
(D). Blind side showing sensory papillae near naris and mouth. 
 
10 4 
 
3.2. Kinematics of Walking and Swimming 
During swimming, the entire body surface of the flatfish, from the tip of the 
rostrum to the tip of the tail, flexes in an undulating wave. Hogchokers lack pectoral fins 
and the small pelvic fins are fused with the anal fin. Thrust generation from paired fins 
can thus be discounted. The dorsal and anal fins have soft, flexible fin-rays and no bony 
spines. During steady swimming, the caudal, dorsal, and anal fins undulated along with 
the body. Together, the body and fins form essentially a flat, flexible, thrust-producing 
surface that increases in flexibility along its margins and toward the tail. Moving the fins 
out of line with the body during swimming creates drag used during steering and braking 
maneuvers. When swimming close to the substrate, the flow on the eyed side (the right 
side of the body in Trinectes maculatus) is impacted differently than the bottom facing, 
or blind (left) side. 
During walking there is little to no body undulation. Forward propulsion results 
from sequential bunches of dorsal and anal fin-rays on the margins of the body (Fig. 
3.3A) cycling forward and back in contact with the substrate to push the body forward 
(Supplemental Video 1). The fin rays are small appendages that lift only slightly off the 
substrate during the recovery phase of each stride. 
3.3. Swimming and walking speeds  
Swimming and walking differed significantly in mean speed (Mann-Whitney 
U=0.5, p < 0.05 two-tailed). For 24 sequences from 7 specimens of Trinectes maculatus, 
mean swimming speed was 12.46 ± 2.12 cm s1 (1.20 ± 0.26 BL s-1) and mean walking 
speed was 2.79 ± 2.49 cm s-1 (0.26 ± 0.21 BL s-1). 
3.4. Velocity magnitude of flow during swimming vs. walking 
10 5 
 
Figure 3.4 shows the magnitude of 2D velocity fields around either swimming or 
walking flatfish. As a Hogchoker swims or walks, it stirs fluid around. Walking causes 
small disturbances to the fluid surrounding the body. However, a swimming Hogchoker 
creates a strong fluid flow trailing the body. As the strong flow is formed, the 
surrounding fluids swirl around. To compare the amount of water disturbance and 
turbulence produced by walking and swimming we found the mean velocity magnitude 
and vorticity for three regions: 1) bow wave surrounding the head; 2) flow above and 
under the body; and 3) wake behind the body (30 video sequences, 17 trials, N=4). 
Results are summarized in Table 3.1. 
 
Figure 3.4. High velocity and strong turbulent flow in the wake of a swimming Hogchoker 
contrasts with lower velocity and minimal water disturbance during walking. Velocity 
magnitude of fluid in the wake of swimming (A-C) and walking (D-F) Trinectes maculatus. High 
velocity flow regions are white and yellow, mid-velocity flow regions are orange and red, low-
velocity regions are dark red to black. Flatfish are shaded in grey, and direction of travel is left to 
right. 
 
10 6 
 
 
Table 3.1. Average velocity magnitude and vorticity in three flow regions (bow wave, body, and 
572 wake) for swimming and walking Trinectes. (30 sequences from 17 videos; N = 4) 
 
The mean velocity magnitude of flow around and behind the flatfish differed 
significantly between swimming and walking, with an average 15 times higher velocity 
magnitude in the wake during swimming compared to walking. Mean velocity magnitude 
of flow around the body in walking and swimming was 3.6 cm s−1 and 0.7 cm s−1; the 
distributions in the two groups differed significantly (Mann-Whitney U=0.5, p < 0.05 
two-tailed). Mean velocity magnitude of the wake in walking and swimming was 3.0 cm 
s−1 and 0.2 cm s−1; the distributions in the two groups differed significantly (Mann-
Whitney U=0, p < 0.05 two-tailed).  
Mean vorticity of the fluid around and behind the flatfish was significantly higher 
during swimming and averaged 12 times higher in the wake. Mean vorticity in the fluid 
around the body in walking and swimming was 2.8 s−1 and 0.6 s−1; the distributions in the 
two groups differed significantly (Mann-Whitney U=1.5, p < 0.05 two-tailed). Mean 
vorticity in the wake during walking and swimming was 3.9 s−1 and 0.3 s−1; the 
10 7 
 
distributions in the two groups differed significantly (Mann-Whitney U=0, p < 0.05 two-
tailed).  
The mean velocity magnitude of the bow-wave was three times higher for 
swimming (2.9 cm s−1 for swimming and 1.1 cm s−1 for walking), and vorticity was eight 
times higher (4.0 s−1 for swimming and 0.53 s−1 for walking). Due to the small number of 
video sequences that included the flow region in the bow wave front of the head, we did 
not apply a statistical measure of significance. 
3.5. Vorticity and wake structure during swimming vs. walking 
When the Hogchoker swims forward, a series of reversed von Karman vortices 
form in the wake. In our experiments, we did not observe clear alternating von Karman 
vortices, but still observed strong vortex underneath the trajectory of the tail fin (Fig. 
3.5A-C). This vortex structure benefits the fish swimming forward by creating pressure 
variation around the body and fins. In contrast, when the flatfish walks in contact with the 
surface, we did not observe any clear vortex or swirling flow structure behind the body.  
10 8 
 
 
Figure 3.5. Vorticity plots of the flow in the wake of swimming (A-C) and walking (D-F) 
Trinectes maculatus show a persistent hydrodynamic trail of vortices left behind after 
swimming that is absent after walking. Dark blue colors represent positive vorticity (clockwise 
rotation) and yellow-green colors represents negative vorticity (counter-clockwise). Flatfish are 
shaded in grey, and direction of travel is left to right. 
 
Interpreting complex flow accurately is difficult with only 2-D data, thus we can offer 
limited observations based on velocity fields in what is functionally sagittal plane 
(Supplemental Video 2). An undulatory swimming sequence begins with the Hogchoker 
flexing the body upwards. As the fish raises the head and moves forward into the 
10 9 
 
surrounding fluid, the snout pushes fluid upwards and upstream (Fig. 3.6A; Fig. 3.7B). 
 
Figure 3.6. Fluid vorticity of flow around swimming (A-C) and walking (D-F) Trinectes 
maculatus shows high vorticity around the head, along the body, and in the wake during 
swimming, and low vorticity during walking. Counter-rotating flow is indicated in blue 
(clockwise) and green (counter-clockwise). Flatfish are shaded in grey, and direction of travel is 
left to right. 
 
Negative pressure zones form mid-way down the concave back and under the head that 
pull fluid backwards and inwards to form proto-vortices, or “regions of elevated vorticity 
adjacent to the body” (from Müller et al., 2001). The flow rotates in opposite directions 
on either side of the body as the proto-vortices travel from high to low pressure zones 
formed by the crests and troughs of the body wave, towards the posterior (Fig. 7B-D). As 
the undulation progresses down the body the tail lifts upwards, with the top side of the 
tail pushing water up and forward into the negative pressure zone formed by the concave 
top side of the body; on the underside, the undulating body pushes a bolus of reverse-
rotating fluid underneath the fish towards the posterior (Fig. 7E-F). The flatfish swims 
11 0 
 
very close to the substrate and likely exploits benefits of ground effect. The underside of 
the body and the substrate form a narrow channel or a lubricating layer for fluid flow, 
which allows the flatfish to slide over it. As the proto-vortex underneath the body reaches 
the posterior it joins the upstream-directed water formed by the upward movement of the 
tail. When the tail changes direction and moves downwards a reverse vortex is shed into 
the wake of the fish (Fig. 3.6B; Fig. 3.7I). Counter-rotating vortices are shed off the top 
surface of the tail (Fig. 3.7H-I) or form secondarily in the surrounding fluid, in response 
to the movement of the primary vortices shed from the bottom surface of the flatfish (Fig. 
3.5A-C). Such strong vortices with lower pressure shed off from the tail will suck and lift 
sediment particles upward, increasing the turbidity of the trailing fluid. The vortices form 
a distinctive hydrodynamic trail behind the fish that persists long after the fish has moved 
on (Fig. 3.6C; Fig. 3.7L). 
 
11 1 
 
Figure 3.7. Sequence of vector fields surrounding swimming Trinectes maculatus show 
proto-vortices forming in regions of high and low pressure along the undulating body and 
vortex shedding in the wake. White arrows visualize magnitude and direction of fluid 
movement. Semi-circles indicating proto-vortices and circles indicating vortices are drawn in blue 
for clockwise flow and red for counter-clockwise flow.  
 
Figure 3.8. Schematic illustration of flow around Trinectes maculatus swimming close to the 
substrate. Opposite-rotating proto-vortices form above and below the undulating body and are 
shed as vortices in the wake. The turbulent flow lifts sediment up from the substrate and increases 
turbidity. Semi-circles indicating proto-vortices and circles indicating vortices are drawn in blue 
for clockwise flow and red for counter-clockwise flow. 
 
Walking Hogchokers, in contrast, disturb the fluid very little (Fig. 4D-F; Fig. 5D-
F; Fig 6D-F). To initiate a bout of walking the flatfish raises the head slightly off the 
substrate—likely to reduce adhesion—and steps forward, forming a small bow-wave. 
Because the flatfish is moving at slow speed the momentum imparted to the water is 
minimal. The head does not make large up and down movements during walking and 
there is low velocity magnitude and low vorticity in the bow-wave region (Fig. 3.6D), as 
well as along the body and in the wake (Table 1; Fig. 3.5D-F; Fig. 3.6E-F). A thin region 
of shear flow forms in the boundary layer between the body and surrounding fluid (Fig. 
63.E-F). Walking flatfish have little to no body undulation and proto-vortices are not 
apparent as they were during swimming. The tail lifts slightly off of the substrate during 
some stride cycles (Fig. 3.4F; Fig 6F) and stays in contact with the substrate during 
11 2 
 
others (Fig. 3.3.5F); in both cases the fluid behind the fish is minimally disturbed and 
there are no clear vortices or persistent trail left in the wake.  
4.0. Discussion 
4.1. Benthic walking in flatfishes is hydrodynamically cryptic 
We found benthic walking in Hogchokers, Trinectes maculatus, produces 
minimal hydrodynamic signals, a phenomenon termed hydrodynamic crypsis. 
Hydrodynamic crypsis enables the fish to move stealthily through its aquatic 
environment. Swimming produced a distinct hydrodynamic foot-print of vortices behind 
the fish that was entirely absent during walking (Figs. 3.5 and 3.6). Measures of water 
disturbance (mean velocity magnitude) in three flow regions – the bow-wave in front of 
the fish, along the body, and in the wake – differed significantly (Table 1). The greatest 
difference was found in the wake, where a strong backwards flow (Fig. 3.4) and trail of 
vortices (Figs. 3.5 and 3.6) was produced by flatfish during swimming, and not during 
walking. 
4.2. Characteristics of the hydrodynamic flow during walking and swimming  
Flatfishes can initiate swimming using opercular jetting to reduce adhesion to the 
substrate, particularly in escape responses (Brainerd et al., 1997). Once they are 
swimming (schematic Fig. 3.8) they undulate their body producing regions of high and 
low pressure. Resulting flow forms proto-vortices, or regions of increased vorticity, that 
are posteriorly-directed in opposite-rotating flow along the top and the bottom surfaces of 
the body. Negative or clockwise-rotating vortices were shed from the trailing edge once 
per tail-beat and were joined by counter-rotating vortices shed from the top surface of the 
body or induced in the surrounding flow. Hogchokers voluntarily swam close enough to 
11 3 
 
the substrate to impact the flow on the bottom side of the fish, likely benefitting from 
ground-effect. Comparing the kinematics and hydrodynamics of flatfishes to carangiform 
and anguilliform swimmers is an interesting challenge because flatfishes have 
characteristics of both. The mode of swimming used by flatfishes, with specific 
amplitudes that increase continuously from nose to trailing edge, is a pattern 
characteristic of a more anguilliform species (Webb, 2002). However, unlike eels, 
flatfishes have a flat, wide, plate-like body form that allows them to “surf” above the 
substrate. Like many other flatfishes, Hogchokers have a distinct caudal peduncle 
characteristic of carangiform swimmers (e.g., Bluegill Sunfish, Lepomis macrochirus; 
Drucker and Lauder, 1999). However, swimming flatfishes do not produce thrust by 
oscillating the caudal fin separately from the body and lack pectoral fins; in that way, 
they are more similar to anguilliform swimmers, which produce thrust continuously with 
their body (Müller et al., 2001). Anguilliform swimmers produce a more complex wake, 
with each tail beat shedding at least two pairs of vortices and no significant downstream 
flow, compared to carangiform swimmers, which have two staggered vortices per tail 
beat and a strong jet (Drucker and Lauder, 1999, 2002; Tytell et al., 2004; 2010). 
Hogchokers appear to leave a wake more similar to the staggered vortices model of 
carangiform swimmers because we observed single vortices being shed with each half-
tailbeat and a strong backwards flow behind the fish. Additional study with 3-D DPIV 
methods would help to resolve this comparison. 
4.3. Relevance to predator detection 
Piscivorous fishes including Northern Stargazer, Southern Stingray, Clearnose 
Skate, Spiny Dogfish (VIMS Multispecies Research Group, 2020) and piscivorous birds 
11 4 
 
such as the Little Blue Heron (Rodgers, 1982) prey on Hogchokers. Flatfishes provide up 
to 30-33% of seal diets (Hammond et al., 1994; Lance and Jeffries, 2009) and 79% of 
cormorant diets (Leopold et al., 1998). Burying behavior and skin camouflage are 
adaptations that help flatfishes evade visual predators. Predatory fishes and seals use their 
hydrodynamic sense to detect and localize swimming prey fishes by tracking 
hydrodynamic trails left behind when they swim; benthic walking, which is 
hydrodynamically cryptic, may be another class of anti-predator adaptation. Future work 
on responses of different potential predators to benthic walking and swimming flatfishes 
will be needed to confirm hydrodynamic crypsis.  
4.4. Relevance to prey capture 
Benthic walking in Hogchokers is a relatively slow locomotor mode as it is in 
other flatfishes (Wirtz and Davenport, 2017, Fox et al., 2018) and walking produces a 
bow-wave with reduced velocity magnitude compared to swimming. A slow-ram strike 
speed reduces bow-wave induced flows and may contribute to the ability of predatory 
fishes to surprise hydrodynamically sensitive prey (Holzman and Wainwright, 2009). The 
polychaete Nereis (Hediste) diversicolor responds rapidly to mechanical and visual 
stimuli (Evans, 1969) as well as chemical cues from fishes dispersed in the water 
(Schaum, 2013) by withdrawing into its burrow and reducing emergence. Reflexive 
withdrawal of clam siphons, extended during filter feeding, is controlled by fast-acting 
nerve pathways (Mellon 1965). Amphipods decrease their movement when exposed to 
chemical cues of macerated conspecifics (Sehr and Gall, 2015). Minimizing water 
disturbance by benthic walking rather than swimming may slow the dispersal of prey-
alarming scent cues generated by the Hogchokers and its feeding activity. Compared to 
11 5 
 
swimming, benthic walking allows a Hogchoker to approach benthic prey with a reduced 
visual profile (flattened against the substrate) and a reduced magnitude bow-wave, which, 
given the reaction speed and sensitivity of its prey to pressure and chemical cues, is 
advantageous. 
Benthic walking facilitates pause-travel predation in flatfishes. Pause-travel 
predation (Andersson, 1981) is characterized by low speed, frequent turning, and short 
move and attack lengths, movement patterns documented in Parophrys vetulus (Fox et 
al., 2018) and juvenile Pleuronectes platessa as they locate and nip the siphons from 
buried clams (Ansel, 1995; Hill et al., 2000) or capture worms and amphipods (Holmes 
and Gibson, 1983). Pleuronectes platessa search for prey using a “shuffle” up to 10 cm 
s−1 and approach prey using a slow “creep” 1 cm s−1 or less (Holmes and Gibson, 1983). 
Feeding behavior of the Common Sole, Solea solea, a nocturnal predator of sessile 
benthic prey with morphological and behavioral similarities to the Hogchoker, has been 
well studied. Sensory papillae on the snout and underside of the head of the Common 
Sole, that resemble those of the Hogchoker (Fig. 3.1C) – once thought erroneously to be 
chemosensory – appear to protect a protrusion of superficial neuromasts clustered around 
the naris (Appelbaum and Schemmel, 1983). Foraging behavior of the Common Sole is 
induced by the onset of darkness (Kruuk, 1963), and they locate prey using a combination 
of mechanoreception and olfaction (De Groot, 1969, Appelbaum and Schemmel, 1983). 
In the presence of live prey, Common Sole cease swimming, reduce their activity level, 
and forage by moving their head side to side, pivoting on the substrate (Batty and Hoyt, 
1995), and tapping their head against the surface (Maia et al., 2009), bringing their head 
in close sensory contact with their prey (Holmes and Gibson, 1983). Benthic walking 
11 6 
 
allows flatfishes that specialize on epibenthic and infaunal prey to approach slowly, with 
less water disturbance to alert prey, and use their mechanosensory and olfactory senses in 
close proximity to their prey. 
4.5. Additional potential advantages of benthic walking 
Diverse fishes independently evolved the ability to walk along the bottom, from 
lungfishes (King et al., 2011) to frogfishes (Dickson and Pierce, 2019). In addition to 
providing hydrodynamic crypsis, benthic walking may be energetically advantageous. 
Theoretically the metabolic costs associated with fish swimming should be lowest at 
intermediate cruising speeds and follow a U-shaped curve (Di Santo et al., 2017). The 
aerobic costs of swimming by Pleuronectes platessa are highest at 0-5 BL s-1 (Duthie, 
1982), the locomotor speed range within which benthic walking has been observed in 
other species of pleuronectid flatfishes (Fox et al., 2018). In kinematic tests, P. platessa 
held station on the bottom and only began to swim at current speeds > 25 cms–1 (Webb, 
2002); Arnold and Weihs (1978) described P. platessa assuming a “clamped-down 
posture” to adhere to the substrate at low current speeds, and as the current increased, 
gradually switching to “posterior-fin beating response” with successive movements of fin 
rays on the dorsal and anal fins (Arnold and Weihs, 1978: p. 148), a locomotor mode that 
we equate with benthic walking (Fox et al. 2018). Analyses of energetic costs of 
swimming – including anaerobic metabolism – in the Clearnose Skate Raja eglanteria, a 
negatively buoyant flat-bodied benthic fish, substantiated the U shaped curve for 
metabolic cost, with higher costs at lower speeds attributed to more energy needed for 
postural equilibrium and to counteract induced drag (Di Santo et al., 2017). Although it 
appears that flatfishes benefit from ground effect when swimming at slow speeds close to 
11 7 
 
the substrate (Webb, 2002) energetic studies to date suggest that swimming at low speeds 
is still relatively costly for flatfishes. Neutrally buoyant fishes must expend energy 
simply to maintain position in the water column (Lauder and Madden, 2007). The ability 
of flatfishes to rest immobile on the substrate and use an alternate locomotor mode – 
walking – at lower speeds is likely to be energetically advantageous. Future work on the 
energetic costs of benthic locomotion in fishes is needed to evaluate this potential benefit. 
Acknowledgements 
CHF was supported by NSF GRFP 2016231353 and a Lerner-Grey Fellowship 
for study at VIMS. Thank you to: James Brister and Eric Hilton at VIMS for hosting a 
research visit for CHF, Gregory Bewley and Reece Kearney of Cornell’s Sibley School 
of Mechanical and Aerospace Engineering for advice on DPIV methods, and Adam 
Summers, Amy McCune, John Hermanson, and Casey Dillman for thoughtful guidance 
on CHF’s dissertation. George Lauder provided greatly appreciated encouragement and 
suggestions for analyses and materials. 
REFERENCES 
Akkaynak, D., Siemann, L.A., Barbosa, A., and Mathger, L.M. (2017). Changeable 
camouflage: How well can flounder resemble the color and spatial scale of 
substrates in their natural habitats? R. Soc. Open Sci. 4, 160824. 
Appelbaum, S., and Schemmel, C. (1983). Dermal sense organs and their significance in 
the feeding behavior of the common sole Solea vulgaris. Mar. Ecol. Prog. Ser. 
13, 29-36.  
Andersson, M. (1981). On optimal predator search. Theor. Popul. Biol. 19, 58-86. 
11 8 
 
Ansell, AD. (1995). Surface activity of some benthic invertebrate prey in relation to the 
foraging activity of juvenile flatfishes. Biol. Ecol. Shall. Coast. Waters. 1, 245-
252. 
Arnold, G.P. and Weihs, D. 1978. The hydrodynamics of rheotaxis in the plaice 
(Pleuronectes platessa L.). J. Exp. Biol. 75, 147-169.  
Batty, R.S., Hoyt, R.D. (1995). The role of sense organs in the feeding behavior of 
juvenile sole and plaice. J. Fish Biol. 47, 931-939. 
Bleckmann, H., and Zelick, R. (2009) Lateral line system of fish. Integr. Zool. 4, 13-25. 
Brainerd, E.L., Page, B.N., and Fish, F.E. (1997). Opercular jetting during fast-starts by 
flatfishes. J. Exp. Biol. 200, 1179-1188. 
Curti, K. L. (2005). Patterns in the distribution, diet and trophic demand of the 
hogchoker, Trinectes maculatus, in the Chesapeake Bay, USA. University of 
Maryland, College Park, ProQuest Dissertations Publishing. 
De Groot, S.J. (1969). Digestive system and sensorial factors in relation to the feeding 
behavior of flatfish (Pleuronectiformes). ICES J Mar. Sci. 32, 385-394. 
De Groot, S.J. (1971). On the interrelationships between morphology of the alimentary 
tract, food and feeding behavior in flatfishes (Pisces: Pleuronectiformes). Neth. J. 
Sea Res. 5, 121-196. 
Derrick, P.A. and Kennedy V.S. (1997) Prey selection by the American Sole, Trinectes 
maculatus (Pisces: Soleidae), along summer salinity gradients in Chesapeake Bay, 
USA. Mar. Biol. 129, 699-711. 
Dickson, B.V., and Pierce, S.E. (2019). How (and why) fins turn into limbs: insights from 
anglerfish. Earth Environ. Sci. Trans. R. Soc. Edinb. 109, 87-103. 
11 9 
 
Di Santo, V., Kenaley, C.P., and Lauder G.V. (2017). High postural costs and anaerobic 
metabolism during swimming support the hypothesis of a U-shaped metabolism-
speed curve in fishes. Proc. Natl. Acad. Sci. U.S.A. 114, 13048-13053. 
Drucker, E. G. and Lauder, G. V. (1999). Locomotor forces on a swimming fish: Three-
dimensional vortex wake dynamics quantified using digital particle image 
velocimetry. J. Exp. Biol. 202, 2393. 
Drucker, E. G. and Lauder, G. V. (2002). Experimental hydrodynamics of fish 
locomotion: functional insights from wake visualization. Integ. Comp. Biol. 
42:243–257. 
Ellis, T., Howell, B.R., and Hughes R.N. (1996). The cryptic response of hatchery-reared 
sole to a natural sand substratum. J. Fish Biol. 51, 389-401. 
Fox, C.H., Gibb, A.C., Summers, A.P., and Bemis, W.E. (2018). Benthic walking, 
bounding, and maneuvering in flatfishes (Pleuronectiformes: Pleuronectidae): 
new vertebrate gaits. J. Zool. 130, 19-29. 
Friedman, M. (2008). The evolutionary origin of flatfish asymmetry. Nature. 454, 209-
211. 
Hammond, P.S., Hall, A.J., and Prime, J.H. (1994). The diet of grey seals in the Inner and 
Outer Hebrides. J. Appl. Ecol. 31, 737-746. 
Hanke, W., Brücker, C., and Bleckmann, H. (2000). The ageing of the low-frequency 
water disturbances caused by swimming goldfish and its possible relevance to 
prey detection. J. Exp. Biol. 203, 1193-1200.  
Hanke, W., and Bleckmann, H. (2004). The hydrodynamic trails of Lepomis gibbosus 
(Centrarchidae), Colomesus psittacus (Tetraodontidae) and Thysochromis 
12 0 
 
ansorgii (Cichlidae) investigated with scanning particle image velocimetry. J. 
Exp. Biol. 207, 1585-1596.  
Hedrick, T.L. (2008). Software techniques for two and three dimensional kinematic 
measurements of biological and biomimetic systems. Bioinspir. Biomim. 3, 
034001. 
Hill, S., Burrows, M.T., and Hughes, R.N. (2000). Increased turning per unit distance as 
an area-restricted search mechanism in a pause-travel predator, juvenile plaice, 
foraging for buried bivalves. J. Fish Biol. 56, 1497-1508. 
Holmes, R.A., and Gibson R.N. (1983). A comparison of predatory behavior in flatfish. 
Anim. Behav. 31, 1244-1255. 
Holzman, R., and Wainwright, P. C. (2009). How to surprise a copepod: Strike 
kinematics reduce hydrodynamic disturbance and increase stealth of suction-
feeding fish. Limnol. Oceanogr. 54:2201-2212. 
Kalmijn, A. J. (1988). Hydrodynamic and acoustic field detection. In Sensory Biology of 
Aquatic Animals (ed. J. Atema, R. R. Fay, A. N. Popper and W. N. Tavolga), pp. 
83-130. New York: Springer-Verlag. 
King, H.M., Shubin, N.H., Coates, M.I., and Hale, M.E. (2011). Behavioral evidence for 
the evolution of walking and bounding before terrestriality in sarcopterygian 
fishes. Proc. Natl. Acad. Sci. USA. 108, 21146-21151. 
Kizhakudan, J.K., Kizhakudan, S.J., Xavier, J.V., Krishnamoorthi, S., Yousuf, K.S.S.M., 
Sundar, R., Manibal, C., Pohhiah, R., Pakkiri, D., Janakiraman, A., 
Chandrasekharan, S., Chakrrapani, G., Chandrachekaran, M.P. (2013). 
Preliminary observations on broodstock development and spawning of Indian 
12 1 
 
Halibut Psettodes erumei (Block & Schneider, 1801) in captivity. Mar. Fish. 
Infor. Serv, T & E Ser. 218, 27-29. 
Kruuk, H. (1963). Diurnal periodicity in the activity of the Common Sole, Solea vulgaris 
quensel. Neth. J. Sea Res. 2, 1-16. 
Lance, M.M., and Jeffries, S.J. (2009). Harbor seal diet in Hood Canal, South Puget 
Sound and the San Juan Island archipelago. Contract Report to Pacific States 
Marine Fisheries Commission for Job Code 497; NOAA Award No. 
NA05NMF4391151. Washington Department of Fish and Wildlife, Olympia WA. 
30 pp.  
Lauder, G.V., Madden, P.G.A., Tangorra, J.L., Anderson, E., and Baker, T.V. (2011). 
Bioinspiration from fish for smart material design and function. Smart Mater. 
Struct. 20, 1-13. 
Leopold, M.F., van Damme, J.G., and van der Veer, H.W. (1998). Diet of cormorants and 
the impact of cormorant predation on juvenile flatfish in the Dutch Wadden Sea. 
J. Sea Res.  
Maia, A., Vinagre, C., and Cabral, H.N. (2009). Impact of a predator in the foraging 
behavior of Solea senegalensis. J. Mar. Biol. Assoc. U.K. 89, 645-649. 
Mellon, D.J. (1965). Nerve pathway and reflex siphon withdrawal in the Surf Clam. J. 
Exp. Biol. 43, 455-472. 
Montgomery, J.C., Macdonald, F., Baker, C.F., and Carton, A.G. (2002). 
Hydrodynamic contributions to multimodal guidance of prey capture behavior 
in fish. Brain Behav. Evol. 59, 190–198 
12 2 
 
Müller, U.K., Smit, J., Stamhuis, E.J., and Videler, J.J. 2001. How the body contributes 
to the wake in undulatory fish swimming: flow fields in a swimming eel (Anguilla 
anguilla). J. Exp. Biol. 204:2751-2762. 
Muthuramalingam, M., and C. Brücker. (2018). Seal and sea lion whiskers detect slips of 
vortices similar as rats sense textures. Sci. Rep. 9, 12808. 
Niesterok, B., Dehnhardt, G., and Hanke, W. (2017). Hydrodynamic sensory threshold in 
harbour seals (Phoca vitulina) for artificial flatfish breathing currents. J. Exp. 
Biol. 220, 2364-2371. doi: 10.1242/jeb.158055 
Pohlmann, K., Atema, J., and Breithaupt, T. (2004). The importance of the lateral line in 
nocturnal predation of piscivorous catfish. J. Exp. Biol. 207, 2971-2978. doi: 
10.1242/jeb.01129 
Quinn, D.B., Lauder, G.V., and Smits, A.J. (2014). Flexible propulsors in ground effect. 
Bioinspir. Biomim. 9, 036008. 
Ramachandran, V.S., Tyler, C.W., Gregory, R.L., and Rogers-Ramachandran, D. (1996). 
Rapid adaptive camouflage in tropical flounders. Nature. 379, 815-818. 
Rogers, J.A., Jr. (1982) Food of nestling Little Blue Herons on the west coast of Florida. 
Fla. Field Nat. 10, 25-44. 
Ruxton, G.D. (2011). Evidence for camouflage involving senses other than vision. In 
Animal Camouflage: Mechanisms and Function (ed. M. Stevens and S. Merilaita), 
pp. 330-350. Cambridge: Cambridge University Press. 
Schaum, C.E., Batty, R., and Last, K.S. (2013). Smelling danger – alarm cue responses in 
the polychaete Nereis (Hediste) diversicolor (Müller, 1776) to potential fish 
predation. PLoS One.  
12 3 
 
Schwalbe, M.A.B., Bassett, D.K., and Webb, J.F. (2012). Feeding in the dark: Lateral-
line-mediated prey detection in the peacock cichlid Aulonocara stuartgranti. J. 
Exp. Biol. 215, 2060-2071.  
Sehr E.K., and Gall B.G. (2015). Responses of an aquatic isopod and amphipod to 
chemical alarm cues from damaged conspecifics. J. Freshw. Ecol. 31, 231-237. 
Sfakiotakis, M., Lane, D.M., Bruce, J., and Davies, C. (1999). Review of fish swimming 
modes for aquatic locomotion. IEEE J. Oceanic Eng. 24, 237-252. doi: 
10.1109/48.757275 
Stewart, W.J., Nair, A., Jiang, H.S., and McHenry, M.J. 2014. Prey fish escape by 
sensing the bow wave of a predator. J. Exp. Biol. 24, 4328-4336.  
Thieklicke, W., and Stamhuis, E.J. (2014). PIVlab – towards user-friendly, affordable 
and accurate Digital Particle Image Velocimetry in MATLAB. J. Open Res. 
Softw., v. 2, Ubiquity Press, Ltd., Oct. 2014, doi:10.5334/jors.bl. 
Tytell, E.D., and Lauder, G.V. (2004). The hydrodynamics of eel swimming I. Wake 
structure. J. Exp. Biol. 207, 1825-1841. 
Tytell, E.D., Borazjani, I., Sotiropoulos, F., Baker, T.V., Anderson, E.J., and Lauder, 
G.V. (2010). Disentangling the functional roles of morphology and motion in the 
swimming of fish. Integr. Comp. Biol. 50, 1140-1154. 
VIMS Multispecies Research Group. 2020, March 3. Fish food habits online data base. 
Retrieved from http://www.vims.edu/fisheries/fishfood. 
Vogel, D., and Bleckmann, H., (2001). Behavioral discrimination of water motions 
caused by moving objects. J. Comp. Physiol. A. 186: 1107-1117. 
12 4 
 
Webb, P.W. (2002). Kinematics of plaice, Pleuronectes latessa, and cod, Gadus 
morhua, swimming near the bottom. J. Exp. Biol. 205, 2125-2134.  
Wieskotten, G.D., Mauck, B., Miersch, L., and Hanke, W. (2010). Hydrodynamic 
determination of the moving direction of an artificial fin by a Harbour Seal 
(Phoca vitulina). J. Exp. Biol. 213, 2194-2200. doi: 10.1242/jeb.041699 
Wieskotten, S., Mauck, B., Miersch, L., Dehnhardt, G., and Hanke, W. (2011) 
Hydrodynamic discrimination of wakes caused by objects of different size or 
shape in a Harbour Seal (Phoca vitulina). J. Exp. Biol. 214, 1922-1930. 
Wirtz, P., and Davenport, J. (2017). Fin crawling in flatfishes (Teleostei: 
Pleuronectiformes). Bull. Fish Biol. 17, 29-38 
Yazdani, G.M. (1969) Adaptations in the jaws of flatfish (Pleuronectiformes). J. Zool. 
159, 181-222. 
12 5 
 
CHAPTER 4 
 
IMPACT OF A MOBILE SCIENCE PROGRAM ON PHYSICS ATTITUDES AND 
ENGAGEMENT 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1 Claire H. Fox and Bruce V. Lewenstein 
 
  
12 6 
 
Abstract 
A mobile physics program serving primarily small cities, towns, and villages in New 
York State (USA) was evaluated for its impact on students aged 8-13. In countries where 
student interest in STEM careers is declining and equity and diversity in science has 
made slow gains, informal science programs have the potential to engage underserved 
audiences and change attitudes; however, claims of positive outcomes need to be tested. 
We investigated student attitudes and responses to the “Physics Bus” program using 
surveys and focus group interviews. The 120 students in our study indicated low levels of 
awareness of Physics as a subject before the program. After the program we found 
increased positivity and enthusiasm for Physics, increased interest in discovering how 
things work, as well as confidence in doing this type of science themselves. Although 
there was little evidence of specific knowledge gain, we did find evidence that the 
program activated wonder, curiosity, and interest in learning more. An emergent theme 
was the appeal of the “Do It Yourself” style of the interactive science exhibits, in which 
inexpensive and everyday materials assembled with duct-tape and hot-glue are used to 
explore physics phenomena in surprising ways. Informal science programs like this one, 
where Physics is perceived as relevant and fun, can help shift the perception of science as 
difficult, boring, and “not for me”.  
 
  
 
 
 
12 7 
 
Introduction 
Mobile Science Labs are uniquely positioned to provide active, hands-on science to 
under-resourced rural and urban youth and communities (Franzblau et al., 2011; Jones & 
Stapleton, 2017), providing equitable access to the types of positive early childhood 
experiences that spark interest in science careers (Maltese et al., 2010, 2014). Attracting 
more students to STEM fields is critical to maintaining a globally competitive economy - 
in the past decade growth in STEM jobs in the United States (24.4%) was much faster 
than non-STEM occupations (4%), a trend that is predicted to continue (Noonan, 2017). 
Increasing equity and effectiveness of STEM education can help address gaps in 
workforce demands, create a more scientifically literate populace, and increase the 
representation of women, Blacks, and Hispanics in STEM jobs (Morrell & Parker, 2015). 
Although much emphasis has been placed on improving STEM Education in schools 
(National Research Council, 2011) there is growing acknowledgement that out-of-school 
experiences play a significant role in developing interest in science and are well suited to 
broadening participation in STEM (Bell et al., 2009; National Research Council, 2015). 
Evaluation of informal STEM experiences is an inherently challenging (Allen & 
Peterman, 2019) but necessary endeavor, to test claims of efficacy and impact (Banerjee, 
2017) and understand the mechanisms behind any positive outcomes.  
Informal science experiences outside of the school classroom have been shown to 
have a strong association with interest in, and pursuit of, STEM careers (Dabney, 2012; 
Simpkins et al., 2006). How does interest start, and how is it sustained? Motivational 
research posits that novel, surprising, and intense situations can trigger interest in 
individuals who have no previous exposure or prior personal interest; according to 
12 8 
 
expectancy-value theory if a subject is perceived as personally relevant, useful, and 
meaningful, over time short-term situational interest is more likely to advance to long-
term individual interest (Hidi & Renninger, 2006); and that interest development process 
can be supported (Renninger & Hidi, 2002). Interest and identity formation in science is a 
complex variable that is shaped by an interplay of factors (Regan & DeWitt, 2014) 
including parental beliefs (Perera, 2014), culture (Coburn & Aikenhead, 1997), media 
(Steinke, 2017), informal science experiences (Dabney et al, 2012), classroom curricula 
(Häussler & Hoffman, 2002), science teachers and the ways they teach and treat students 
(Xu et al., 2012). Attitudes towards science are complex: interest in specific science 
subjects varies, and attitudes towards science in general are often different from attitudes 
towards ‘school science’ (Osborne et al., 2003). A drop-off in personal interest in science 
and science careers is broadly observed starting between ages 10 and 14 in developed 
countries (Jenkins & Nelson, 2005; DeWitt & Archer, 2015). Student interest may be 
“switched off” by school science (Barmby et al., 2008), in contrast to developing 
countries where interest in science careers remain high (Sjøberg 2005; Schreiner & 
Sjøberg 2007).  
Of particular concern for personal interest is Physics. Physics has a reputation for 
being obscure, difficult, and boring. Fascination and curiosity about physical phenomena 
in early years may be undermined by student’s later classroom experiences, as 
exemplified by this excerpt from large longitudinal study, from a girl who deeply loved 
physics in primary school but found that in high school ‘physics got more 
complicated…It’s too dusty … too one-tracked, maybe … Sometimes I think that physics 
can be seen as very cold’ (Krogh & Thomsen, 2005, p. 296). Krogh and Thomsen’s 
12 9 
 
metaphor of (cultural) border crossing investigates how student attitudes towards physics 
reflect the degree of congruency between their home and peer values and ways of being, 
and the culture of their classroom science – similar to Costa’s conceptualization of 
science as ‘another world’ (Costa, 1995). The majority of students reported a negative 
reputation for physics as strange, difficult, or boring; along with physics self-concept and 
teacher’s treatment of students, the reputation of physics explained 51% of the variation 
in student’s attitudes towards physics (Krogh & Thomsen, 2005). Many students perceive 
physics as isolated facts that are disconnected from the real world (Milner-Bolotin et al., 
2011). The rigid discourse around physics needs reconfiguration for a wider range of 
student gender and ethnic identities or it will continue to be rejected as “horrible, 
difficult, and nasty” by individuals who are not compatible with the currently available 
dominant scientist identities (Hughes, 2001). 
Costa (1993) calls for science courses to be more compelling and relevant, 
organized around the question ‘How does chemistry (or biology or physics) impact my 
personal life and society?” Perceiving the utility of science and the relevance of science 
to everyday life is strongly and positively associated with the desire to study and work in 
science (Sheldrake, 2017). Researchers have shown experimentally that when under-
performing students with low expectations of success are encouraged to connect science 
topics with their daily life, their interest and performance increases substantially 
(Hulleman & Harackiewicz, 2009). These connections may be especially important for 
traditionally under-represented communities (Dawson, 2018). There may be a limit, 
however, to what curricular reform can accomplish in terms of encouraging authentic, 
intrinsically motivated learning experiences and positive science identities within the 
13 0 
 
context of a school environment that is focused on tests and achievement (Carlone, 
2004). The U.S. Next Generation Science Standards (NGSS), released in 2013, are a 
reform effort designed to increase equity, student engagement and conceptual learning in 
STEM. A 2019 study of Hispanic/Latino and Black middle school students’ self-efficacy 
in STEM found lower confidence, self-efficacy, and career interest in physical sciences 
compared to life sciences; no improvement was found in students from NGSS aligned 
classes (McCormick, 2019).  
Informal science experiences can be learner-centered and interest-driven in ways 
that can be challenging to provide in a classroom environment. Motivation and identity 
are key elements of the strands of science learning described in the National Research 
Council’s report on Learning Science in Informal Environments (Bell et al., 2009). The 
first strand states that learners will “experience excitement, interest, and motivation to 
learn about phenomena in the natural and physical world.” The sixth and final strand says 
that learners will “think about themselves as science learners and develop an identity as 
someone who knows about, uses, and sometimes contributes to science.” This paper 
explores whether a particular instance of informal science education – a mobile physics 
lab – can make the connections between physics and daily life that research suggests are 
13 1 
 
useful for generating the motivation and identity that lead to longer-term engagement 
with science. 
Program Description  
Background and Philosophy: 
The Physics Bus is an affordable, accessible, innovative public science program with the 
mission to re-establish physics as something anyone can do with the stuff around them; 
whether for fun, utilitarian purposes, or for the beauty in it.  
The Ithaca, NY (USA), based Physics Bus program was founded in 2013 by Erik 
Herman and a small team of coworkers and community volunteers. The impetus for the 
Physics Bus program grew out of Herman’s personal struggles and dissatisfaction with 
classroom education both as a student and as a teacher – and the joy and accomplishment 
he felt when tinkering, inventing in the garage, and sharing the wonder of physics 
phenomena. Herman was inspired to start a science outreach program by his experience 
traveling with the Arizona Mathematics Roadshow and Physics Factory demo shows. He 
wanted to combine the reach and charisma of the mobile bus (e.g., Barbagallo, 1997; 
Fox, 2015) with a program model that allowed visitors a self-directed, hands-on 
experience. The program was inexpensive to start - the Ithaca-based program began with 
the gift of a free vehicle, a small vintage schoolbus that had been used as a party bus by a 
local fraternity, and was subsequently filled with physics exhibits built by volunteers and 
high school students out of junk. Additional funding from a local Family Foundation 
supported an upgrade to a larger schoolbus in better condition (2001 International 3800 
body, 10 windows on each side, approximately 10.7 meters in length). The idea of using 
cheap everyday materials was influenced by Brian Jones’ Little Shop of Physics 
13 2 
 
(Colorado State University, 2020), Arvind Gupta’s Toys from Trash (Gupta, 2020), 
Science Toymaker (Harrison, 2020), and the Global Alliance of Community Science 
Workshops (GACSW, 2020). Similar efforts have been tested elsewhere (García-
Guerrero & Lewenstein, 2020). The Physics Bus program received some support from 
Cornell University during the years 2014-2019 while Herman was employed as a science 
engagement specialist at the Cornell Laboratory for Accelerator based Sciences and 
Education. However, the Physics Bus operated during the study period as a separate non-
governmental organization (in the United States, usually called a nonprofit organization 
or 501(c)3 organization after the relevant tax code) and considered itself to be a 
grassroots, community-based program. The program model has experienced national-
level growth. As of May 2020 there are five Physics Buses operating out of Ithaca NY, 
Tucson AZ, St. Louis MO, Gainesville FL, and Watsonville, CA; the programs operate 
under a common nonprofit organization and while each one has its own style, they share 
13 3 
 
similar methods and motivations. This study only evaluated aspects of the Ithaca-based 
Physics Bus program; hereafter referred to for brevity as “the Physics Bus”. 
 
Figure 4.1. Venues hosting the Physics Bus (Ithaca-based program) from 2014-2019. 
 
Population served: 
Between 2014 and 2019 the Physics Bus reached approximately 45,000 visitors through a 
variety of venues (Fig. 4.1). The largest demand for the Physics Bus program came from 
K-12 Schools (36%) predominantly at the Elementary level, followed by Community 
Events (25%) encompassing a wide range of seasonal/regional festivals, music events, 
and block parties. The Physics Bus is a popular request for Public Libraries (17%). A 
small but significant percentage of programs took place at STEM themed fairs (5%) and 
the remaining miscellaneous venues (5%) include visits to churches, country clubs, 
academic conferences, and private events. Most programs were in the Upstate NY region 
with occasional daytrips and overnight trips, and one or two 7-14 day tours each year in 
13 4 
 
other parts of the country. The Physics Bus disproportionately serves towns, villages, and 
small cities with populations under 50,000. 
 
Figure 4.2. Populations served by the Physics Bus program in 2014-2019 compared to the U.S. 
population. The program disproportionately visited rural towns and small cities. Within school-
age populations the program reached more low to mid poverty students than average, and fewer 
students on the extremes of high or low poverty. 
13 5 
 
Typical visit experience: 
The bus parks in a pre-arranged location and one or two folding tables of exhibits and 
interactive features are put outside, under a crank-out awning if the weather is inclement 
(Fig. 4.3).  
 
Figure 4.3. The Physics bus is a decommissioned school-bus covered in aluminum foil, with 
garbage-can “rocket thrusters” on the back and functioning solar panels on the roof. (1) Tables of 
exhibits and large interactive physics experiences are placed outside the bus. (2) Inside most of 
the passenger seats have been removed and the remaining seats form a bench and café table with 
exhibits. (3) Two built-in display tables run down either side of the bus. They have storage 
underneath and a metal countertop which allows the exhibits, with strong magnets embedded in 
the base, to stay in place during transit. 
 
The front and back door of the bus are open with a set of metal steps out the back. The 
program consists of around two dozen colorful, homemade physics exhibits on two long 
counters down either side of the bus interior and at a couple of outdoor stations. The 
13 6 
 
counters are surfaced with metal and the exhibits are attached with magnets allowing 
transport with the exhibits in place. Solar panels on the roof of the bus and a large array 
of batteries provide power to exhibits and interior LED light strips. If it is an event with 
scheduled groups such as a school visit, a program representative (usually doubling as the 
driver) greets the group, introduces themselves, and may choose to give a short 1 minute 
introduction. They oversee the visit for safety, repair and refill exhibits as needed, and 
field any questions. In the early years of the program 2-5 trained volunteers (generally 
teens from a charter school or college students) accompanied an adult staff member, but 
recently the program has used a single facilitator/driver and requests that the venue 
provide helpers. The visits in this study had one staff member and one or two trained 
volunteers. The interior space, behind the entry stairs and driver’s seat, is approximately 
8 meters in length. Twenty small children or 15 teens/adults fit comfortably onboard at 
once and at large public events the bus fills shoulder-to-shoulder. Visitors move at their 
own pace around the interior encountering exhibits, interacting with them by poking 
buttons, singing, swinging, dropping, drawing, cranking, with simple signs (that are often 
ignored) providing basic instructions like “push down here.” It is a noisy, active 
environment - there is a lot of laughter, shrieks of excitement, people often call their 
friends or family over to what they just experienced to share, laugh, and wonder. Outside 
13 7 
 
the bus visitors try out more exhibits, make giant bubbles, launch water rockets, get 
bottles stuck on the roof, and run around and play. 
Expected outcomes 
Expected outcomes based on extended conversations with the program’s founder and 
director over a period of years (and confirmed for this study) are that visitors will:  
1. Experience joy, wonder, and excitement in association with physics  
2. Feel increased curiosity and an intrinsic desire to learn more  
3. Gain an intuitive understanding of a variety of physics phenomena  
4. Become aware of the relevancy of physics in their daily lives  
5. Be inspired to tinker with electronics and mechanics and explore science at 
home  
6. Believe that science is something they can engage in – that it can happen 
outside of academia and be done by everyday people with free materials.  
Although this list of expected outcomes was developed inductively, we note that many of 
them map onto the “strands of learning” from Bell et al., (2009). 
Methods 
For this study we obtained Institutional Review Board approval for human participant 
research from Cornell University’s Office of Research Integrity and Assurance, protocol 
#1903008709. 
Recruitment 
In Fall 2018 and Winter 2019 we contacted Elementary and Middle Schools within a 2 
hour radius from Ithaca, NY that receive school-wide Title 1 funding (a federal aid 
program awarded to public schools in the United States with high percentages of low-
13 8 
 
income students). Schools were offered a free program in exchange for participation in 
the study. Previous visit logs were consulted to ensure that the schools and communities 
considered for inclusion had never received a Physics Bus visit prior to the study. One 
Elementary School and one Middle School agreed to participate. The Elementary School 
is in a city with a population under 50,000; it is classified as a high-poverty school and 
performance on state standardized tests are far below average. The Middle School is in a 
village with a population under 2,500; it is classified as a mid-high poverty school and 
performance on state standardized tests are close to average. Parent consent forms were 
sent home with students prior to the visit. Students who turned in parent consent forms 
also provided verbal assent prior to participating in the paper survey or focus group. A 
total of 120 students (102 3rd and 5th graders and 18 6th and 7th graders) took the survey, 
and 12 students participated in the focus group interviews, in two groups of 6. 
Approximately 70% of visiting students in 3rd and 5th grade participated in the survey. 
Schools were asked to randomly select 6 student participants, from the sub-set with 
parental permission, to take part in the focus group interviews. We do not know what 
specific methods the teachers used to select students, thus the sample was neither 
representative nor random; rather, the focus-groups allowed the researchers to explore 
questions raised by the study in more depth. Students who participated in the focus 
groups reported a range of school science achievement.  
Demographics 
Student eligibility for free or reduced-price lunch (FRPL) was used as a proxy measure 
for the concentration of low-income students in a school, following the categorization of 
the U.S. National Center for Educational Statistics: High-poverty schools are defined as 
13 9 
 
public schools where more than 75.0 percent of the students are eligible for FRPL; mid-
high poverty schools are those where 50.1 to 75.0 percent of the students are eligible for 
FRPL; mid-low poverty schools are those where 25.1 to 50.0 percent of the students are 
eligible for FRPL; and low-poverty schools are those where 25.0 percent or less of the 
students are eligible for FRPL. 
We applied U.S. Census Bureau data to categorize program locations: ‘Rural’ 
encompasses all population, housing, and territory not included within an urban area. 
‘Urban’ is subdivided into Urban Clusters (UCs) of at least 2,5000 and less than 50,000 
people, and Urbanized Areas (UAs) of 50,000 or more people. 
Instruments and Analyses 
We used a mixed-methods approach utilizing surveys, focus-group interviews, and 
observation.  
Survey: All students with permission (both parental consent and child assent) were given 
a one-page paper pre-survey immediately prior to the program, and a one-page paper 
post-survey immediately after the program. The survey was developed specifically for 
this study. The pre-program portion consisted of two short answer questions and six 5-
point Likert scale questions. The post-survey portion consisted of three short answer 
questions, one yes/no/maybe question, and four 5-point Likert scale questions. The Likert 
scale questions had verbal prompts and a smiley face visual analogue scale. The survey 
was kept short in order to minimize disruption to classroom teaching time, and to reduce 
survey fatigue especially for the younger Elementary students. Short-answer questions 
were included to allow students unlimited expression in their understandings and feelings 
about Physics, what they liked most and why, and the thoughts and questions raised by 
14 0 
 
their experience. Short-answer responses were transcribed as written (preserving spelling 
and grammatical errors, which were at times extensive enough to require interpretation) 
and went through two rounds of coding. Emergent categories were developed during the 
first round, which were then split, merged, reconsidered and renamed as needed during 
the second round. Some short answer questions included here were edited for spelling. 
Interview: The focus group interviews were 45min in length and were conducted at the 
end of each visit day. The interviews were semi-structured, with questions exploring the 
student’s attitudes towards school science, future science careers, and science in general, 
their thoughts and reactions to the Physics Bus program, what they liked and disliked 
about the experience, and whether they gained any new understanding of physics 
phenomena. The focus group interviews were audio-recorded at the time and later 
transcribed without names or identifying data, and the audio files were then deleted.  
Observations: One author of the study, CHF, observed K-12 student and public 
interaction with the program in her role as employee and volunteer for the Physics Bus 
from 2014-2018. These observations were combined with the survey and focus group 
results to identify the primary factors responsible for increased science interest and 
enthusiasm.  
Results 
Background levels of Interest  
To provide a background level of interest before the program, students reported their 
interest in a range of activities on a five-point Likert scale: Make arts or crafts, Explore 
how things work, Play a sport, or Read a story. Third and fifth graders loved Making arts 
and crafts more than any other activity (66%), followed by Sports (60%) and Exploring 
14 1 
 
how things worked (51%). Reading was by a large margin the least-loved activity (34%) 
as well as the most hated, with 16% of students reporting that they either disliked (7%) or 
hated (9%) reading. Sixth and Seventh grade students loved Sports (75%) most of all, 
followed distantly by Making art or crafts (31%). Only 17% loved Exploring how things 
work. Again, Reading came in last with only 7% of students loving the activity, and 
nearly 47% hating it. 
Focus group participants were interviewed about what kind of science they liked, 
whether they liked science in school, or were interested in becoming scientists. Reflecting 
national and international data and as expected based on the literature review (Jenkins 
and Nelson, 2005; DeWitt and Archer, 2015) the Elementary students were generally 
positive about science. Every student was able to quickly name a science topic that they 
enjoyed, e.g. learning about animals (‘I like doing research on animals, that’s what we’re 
doing in science, and I really like it’), engineering and making things, and learning about 
the universe (because it is ‘so mysterious’). Some of their school science experiences 
were positive, others were negative – they all vigorously agreed that learning about the 
weather ‘was so boring… because we would read the same thing over and over again. It 
was just memorizing.’ Two students who felt that they would not be able to be scientists 
when they grew up explained why: one said they were ‘bad at memorizing’ and the other 
lamented ‘I have a lot of ideas… but I don’t think I would be very good at making them a 
reality… because when someone tries teaching me something, like my brother, when 
they’re not showing me, and they’re just talking to me, I really don’t get it.’ They 
reflected a belief that science is a transmissive subject where achievement is dependent 
14 2 
 
on factual knowledge and ability to learn from lecture. In the terms of the Bell et al. 
(2009) strands of learning, these students did not identify as science learners.  
Middle school students in the focus group expressed mixed interest in science. 
None participated in science outside of school, and none had met a scientist before. When 
asked if they liked science in school, they replied with a range of responses from ‘not 
really’, ‘for the most part’, and ‘depends’. One student said it ‘can be very boring’ and 
another said that they ‘look forward to the hands-on parts.’ One student felt that she could 
become a scientist because she had very good grades in science ‘it takes a lot of stuff 
though, like needing to study in school, a lot of effort.’ The others were not interested in 
science careers, because they were ‘not good at it’ or because it took ‘too much work.’ 
Awareness of Physics as a Science subject 
The majority of 3rd and 5th grade students (51%) and 6th and 7th grade students (54%) 
surveyed had no idea when asked ‘What is Physics’, with answers such as ‘IDK’ and ‘I 
don’t know what Physics are?’ Of the Elementary students who had an idea of what 
Physics was, 26% indicated that it had something to do with science e.g. ‘it is some kind 
of science’, and 16% identified one or more correct Physics topic e.g. gravity, motion, 
energy, or how things work. Survey results from middle school students indicated that 
40% knew that it was some kind of science; 13% thought it had something to do with 
movement e.g. ‘physical stuff that you do every day’ suggesting a possible confusion 
with Physical Education, exercise, or using your physical body.  
Enthusiasm and positivity towards Physics  
After students participated in the Physics Bus program, we found increased enthusiasm 
and positivity towards Physics as a subject, with the greatest gains observed in the 
14 3 
 
Elementary age group. Positivity was defined as the voluntary use of positive words (e.g., 
awesome, amazing, fun, cool, like it, or excited) when writing in their own words what 
they thought of Physics. For 3rd and 5th grade 30.5% used positive words to describe what 
they thought of physics before the program; Post-visit this increased to 78%. Post-visit 29 
elementary students specifically used the word ‘Fun’ to describe Physics, an increase of 
21 students. Only one student used the word ‘Awesome’ to describe Physics pre-
program; 11% of students described Physics as ‘Awesome’ after the program. Middle 
schoolers in the focus group, when asked to describe their Physics Bus experience to a 
friend, all expressed a positive impression, e.g. ‘fun and interesting’, ‘I’d say it was cool’, 
and ‘a fun experience and I did a lot of learning…about what Physics was and how we 
could use it in our daily lives and how we can use it with different scraps’.  
Favorite exhibits and why they were engaging 
Students favored physics exhibits that multiple people could explore at once, and that had 
surprising results or complex interaction (Figs. 4.4 & 4.5).  
14 4 
 
 
Figure 4.4. Favorite exhibits in order of popularity: 1) Water bottle rockets 2) Giant Bubbles 3) 
String Loop 4) Mini-plasma Cutter 5) Van de Graaff 6) Magnet TV 7) Fresnel Lens Helmet 8) 
Laser writer 
 
When asked which exhibit was their favorite, two were picked 3-4 times more frequently 
than any others: Water Bottle Rockets and Giant Bubbles. Other popular exhibits were a 
mini-plasma cutter used to cut aluminum foil, a vacuum cleaner that blew a string loop, 
and a Van de Graaf machine that built up a charge and shocked people. Also mentioned 
were a Fresnel lens mounted on a recycling bin that is worn like a helmet, a magnet 
dangling in front of a Cathode Ray Tube color TV, a ‘spy screen’ made from an LCD 
monitor with the outer polarizing filter removed, a tornado produced by water vapor 
14 5 
 
pulled upward by an old computer fan, a laser writer made from a blue laser pointer and 
glow-in-the-dark paper, an electromagnet ring jump machine, and a mini-vortex ring 
shooter made out of a cool mist humidifier covered by a Frisbee with a hole.  
 
Figure 4.5. Twelve most popular exhibits and the number of students who chose them as their 
favorite 
Students gave a range of answers for what they liked about their favorite exhibit. Aside 
from saying that they liked ‘everything’ about it, they most often noted the impressive 
height or size (21%), e.g. ‘I liked them because the rocket amazed me on how high it 
went.’ ‘I had to hold the bottle and it went soooo high.’ They also liked that it was fun or 
funny (17%), e.g. ‘it was awesome and really fun to do’, ‘I liked that it was very fun and 
we asked really funny questions.’ Students liked that it was made of junk (9%; see 
section in results on “Everyday materials and DIY aesthetic). Some liked the memorable 
or intense physical sensation ‘getting electrocuted and hurt that’s why’, ‘it went through 
my body.’ Other students liked it because it was an interactive (8%) or social (7%) 
14 6 
 
experience ‘when people came by you shot them with string,’ ‘I like it because if you 
touch someone they get electric,’ or ‘playing around with friends with the bubbles.’ Some 
students appreciated the beauty of some of the physics phenomena ‘I liked all of the cool 
color.’ Other students liked the unexpected ‘I like that everything surprised you.’ Several 
liked the variety ‘that there are a lot of things to do.’ Lastly, the students appreciated 
having freedom ‘I liked that you could write anything on it’ and ‘we got to go off by are 
self’s.’ Students liked the idea of science that is self-initiated and experimental. As one 
focus group participant explained, ‘I am not up to the science that people are like…all the 
things they know I know, like this is how you do slime…I like science where you just 
come up with it in your mind and you’re like, ‘this might be science, why not do it?’ 
Student interest and confidence 
After participation in the program, students made large gains in interest and confidence. 
Asked on a five point scale how interested they were in discovering how things work, 
after the visit 80% of elementary students said that they were ‘definitely’ interested in 
discovering how things worked, a 30% increase from before the program. After the 
program nearly half of the elementary school students (48%) indicated that they 
‘definitely’ could make a science gadget, a 17% increase. The percentage of middle 
school students who felt that they ‘definitely’ could make a science gadget jumped from 
8% before the program to 42% after, and those saying they ‘probably’ could increased 
from 21% to 50%. As a student in the focus group explained, ‘I never made one before, 
but I would, if I had like all the supplies and stuff and had all the old and new stuff to like 
make it.’ Students in both focus groups easily identified what exhibit they would make, 
and spontaneously shared different ways they’d like to use the exhibits in non-scientific 
14 7 
 
ways, such as using the lie-detector circuit as a musical instrument to DJ, using the 
Fresnel lens helmet in a comedy routine, or using the swirling colors of the magnet TV to 
increase focus and calm feelings.  
Student curiosity, wonder, and desire to learn more 
Student curiosity, wonder, and interest in learning was activated by the Physics Bus 
program. When asked whether they ‘would like to learn more about’ their favorite exhibit 
87% of elementary students said yes. When asked ‘what did you wonder about when 
using the exhibit’ there were only two student who didn’t wonder at all. One offered the 
explanation ‘Nothing…I was so distracted by how cool it was.’ What students wondered 
about largely fell into two categories: how it was made (39%), and how it worked (34%). 
Typical responses were ‘how they made it, what junk they used’, ‘that I could make it at 
home’, and ‘I wonder what materials did they need.’ Students were curious about the 
specific physics phenomena they observed, such as ‘how the foil burned’, ‘how did the 
magnet go so smooth thru,’ ‘how when you twist the handle the lit come on,’ ‘how do 
tornados form,’ and larger questions such as ‘how we get electricity.’ Other students 
wanted to push the limits of the exhibits and try new things, and wondered what would 
happen (9%), e.g. ‘how much air could go into the bottle’, ‘how high it could go’ and ‘if 
we shocked the bubble what would happen.’ Some students wondered about how the 
program was run: ‘how long it takes to set up,’ ‘how they replace the magnets’ and where 
the ideas for the exhibits came from,’ ‘how people found out how to make these things’, 
and ‘how people thought about this stuff.’ 
Content knowledge  
14 8 
 
We found that students were applying their previous experiences and reasoning abilities 
to make sense of the physics phenomena, however we found little evidence that new 
knowledge or intuition was gained from the program experience. Although the purpose of 
this study was not to assess student content learning, we did ask a couple of questions of 
focus group participants to see if they had gained intuition about underlying physics 
phenomena (one of the expected outcomes of the program experience). Here we report on 
Elementary student’s post-visit understandings of two physics topics that are explored in 
multiple exhibits on the bus (Fig. 4.6). 
1) Three exhibits on the bus visualize the wave nature of sound: an oscilloscope made out 
of an old cathode ray TV with a microphone attached, a guitar with a spinning strobe disk 
under the strings, and a transparent tube attached to a speaker with Styrofoam beads that 
stand up in waves when the sound is adjusted to certain frequencies. After the program, 
when asked ‘If you could see sound, what do you think it would it look like?’ none of the 
Elementary students referred to exhibits in their answers. One of the students said 
immediately ‘I think it would just be air waves’ which likely reflects pre-existing 
knowledge, as the waves in the program exhibits were visualized in physical materials. 
Another said that it would ‘look like how it sounds’, i.e. ‘if you hear a clap you would see 
someone clap.’ A third student recalled what sound waves looked like in a Magic School 
Bus video (a popular television and online media edutainment show produced by 
Scholastic Inc.), i.e. little green circles for small noises and large red circles for loud 
noises. When the researcher prompted the students by asking them what they saw when 
they plucked the guitar exhibit, one student said, ‘I just plucked it, I didn’t see anything’, 
another said, ‘I saw things moving over and over,’ and another reported that when they 
14 9 
 
‘plucked it more hard it [the strobe drum] would spin faster.’ None of the students in the 
focus group mentioned wiggly or curvy lines or waves. 
 
Figure 4.6. Physics Bus exhibits provide surprising sensory experience with a range of physics 
phenomena. Exhibits 1-5 visualize the wave nature of sound. (1-2) Modified acoustic guitar with 
strobe belt behind the strings. (3-4) Oscilloscope made from an old cathode ray tube TV hooked 
up to a microphone. (5) Clear tube with small styrofoam beads showing standing waves, 
connected to a speaker and adjustable tone generator. Exhibits 6-7 feature electromagnetism. (6) 
Large wire coil connected to A/C power and push button, surrounding a transparent bowl with 
permanent magnet inside. (7) Classic “Ring Jump” demonstration. 
2) At least three exhibits on the bus are based on electromagnets, including a popular 
exhibit consisting of large transparent plastic bowl with a button in the middle, which 
controls AC electrical current in a large coil of wire around the outside of the bowl. The 
15 0 
 
coil of wire is partly obscured by patterned duct-tape holding the coil onto the bowl. 
When the button is pushed a magnet wrapped in tape vibrates and bounces around in the 
bowl. When asked how they thought it worked, most of the focus group students knew 
that the object bouncing was a magnet and several used words like ‘repel’ and ‘interact.’ 
Answers that went into more detail were incorrect: ‘there was something in it, like hitting 
on the top’, ‘the bowl that it was in would make the magnet spin, because it would like 
tap on one spot and would keep going around.’ One of the students thought ‘maybe it was 
kind of alive, maybe it wanted to get away from whatever that was.’ None of the 
Elementary or Middle school students mentioned the coil of wire, or electricity – 
observations that would be necessary to make an intuitive leap to understanding an 
electromagnet.  
When asked the broad question ‘did you discover anything or learn anything?’ on 
the bus, a student in the focus group shared that he ‘learned that you can make plasma’ 
because the ‘guy said that the rod and tinfoil make plasma.’ He had heard of plasma 
before the program and thought that it was ‘a kind of energy’, but he didn’t realize that 
the arc on the aluminum cutter exhibit was actually plasma, or that you could ‘just make 
it,’ until the teenage volunteer on the Physics Bus engaged with him. This suggests that 
program staff or trained volunteers have the potential to aid the process of knowledge 
acquisition from the Physics Bus exhibits. 
Everyday materials and DIY aesthetic  
The use of inexpensive, common materials and homemade nature of the exhibits (Fig. 
4.7) made a strong positive impression on students. When asked why they liked their 
favorite exhibit, some students specifically noted that it was because it was made of 
15 1 
 
everyday and reused materials: ‘I liked how they used junk and made it so cool,’ ‘It was 
cool that you could burn stuff with simple objects’, ‘That you can make new stuff out of 
trash,’ ‘I liked that most of it was made out of trash then it was turned into science.’ 
When asked how the science on the Physics bus was different than science in their school 
a student explained ‘I think it’s different because you don’t need to buy the things you 
need to make a project. You already have the things you need you can just re-use them. 
You already bought it in the past and you can just save it and use it for something else.’ 
 
Figure 4.7. Familiar, everyday materials are repurposed to produce physics exhibits. (1) Home-
made “spy screen” LCD monitor with polarizing filter removed and polarized sunglasses (2) 
Vortex ring exhibit made from a Frisbee with a hole in the center over a basin with cool mist 
humidifier components (3) Microwave oven with small neon bulbs lying on the glass tray 
15 2 
 
Students wanted more science like this, ‘I want science to be more fun, especially 
science, which is like something that can be really fun, and the teachers usually want to 
do science where you go out and buy stuff, where it seems like they really don’t have to, 
they could just… the stuff they’re buying isn’t really the best, you can get other things 
just for free.’ 
Physics Bus compared favorably to school science and other informal science 
opportunities 
The Physics Bus program was exceptionally well rated by students relative to other 
science experiences. On a five point scale 97% of 3rd and 5th graders, and 92% of 6th and 
7th graders indicated that the Physics Bus was either ‘better’ or ‘the best’ compared to 
other science that they have done.  
When focus group participants were asked to share what they thought of the Physics Bus 
experience, and how it was different from science in school, the Elementary students felt 
that they learned more because it was fun, and they were able to directly experience 
science. They were also impressed by the re-use and the simplicity of the materials: 
S1: It’s very different because on the Physics Bus you learn more. 
S2: Because it’s like science in Ms.X’s class... it’s like ‘eh’... it feels like work, but 
when you’re actually learning stuff, it’s fun. 
S3: It’s different because, like S2 said, we just do working and reading and 
writing, and on the Physics Bus you actually get to do it, instead of like writing 
about it. 
S4: They are COMPLETELY different because in school you have to be 
[indistinguishable] the safest thing you possibly can do. And you really don’t do 
anything because you are usually working in a group. But when you are on the bus 
you do something and you don’t have to wait. 
15 3 
 
S5: So its like also more fun, and it’s not expensive….you can just find stuff in a 
dumpster, you can just dumpster dive. So you can, just like he said with the 
trashcan thing, I was like it was simple to make, you don’t need a ton of money. 
S6: Something so simple can just be so amazing. 
The middle school students appreciated the hands-on nature and freedom to self-
guide: 
S1: I loved it better…because it had more hands-on experience. 
S2: And we had more freedom - instead of telling us we needed to focus on one 
topic, and one thing, and we had to do all the experiments by ourselves, we could 
explore the different science topics we learned about for our whole school career.  
S3: Yeah, and you get to figure it out for yourself without someone like telling 
you. 
Middle schoolers also recognized the relevance of physics to their daily lives: 
R: Did anyone else discover or learn anything? 
S1: I learned how important physics is in your everyday life and how much we use 
them even if we don’t notice 
R: Did you not know that before? 
S1: Because I didn’t really think about it. Because everyday we use different 
things that we are used to using every single day, that we don’t think about how 
there is so much science behind it and how physics is used.  
R: So you never realized that. 
S2: [another student] I’m usually too busy using the thing, to think about it 
R: Does this make you curious about anything that’s in your daily life, like how it 
works? 
At this point, students explained how the delivery method (outside the classroom, 
onboard and around a modified schoolbus) was a critical part of the successful 
engagement: 
S2: I feel like it was like better, because it was outside of like our normal 
environment, like outside of the classroom. 
R: So you liked it because it was outside of your normal classroom? Like if we’d 
taken all of the exhibits and we’d put them in your classroom, and it wasn’t in a 
bus… 
S2: I’d still be attracted to them, but like I’d just be like ‘um ok, whatever’ 
15 4 
 
R: So you think the bus makes a difference? What do you all think? 
[chorus of agreement]: OUTSIDE. 
R: The fact that it’s outside and you got to run around? OK 
S3:…I feel like if we were in a classroom I’d probably get 20 minutes and then I’d 
be bored instead it was like an hour 
S4: Yes and it’s because we’ve been in the same areas for more than 100 days. 
Like in the classroom over and over and over again. Even if we were just on the 
bus, I still feel like since we’re in a different area than we are used it would be 
more fun. 
 
Discussion 
The Physics Bus program generated strong interest and enthusiasm in Elementary 
and Middle School students, with 97% of third and fifth graders rating it “better” or “the 
best” compared to other science they have experienced. Participants left the program 
feeling animated, excited, and happy, and described ‘what they thought of Physics’ 
afterwards as fun, awesome and cool. These responses highlight the importance of 
including “excitement, interest, and motivation” in the definition of science learning (Bell 
et al. 2009). Positive informal science experiences like those offered on the Physics Bus 
counteract the reputation that physics is boring, detached from daily life, and difficult. 
The experiences develop an alternative narrative that young children pick up on quickly, 
in only one visit: physics is exciting and fascinating, it is part of our everyday life, and 
exploring it is intuitive and fun. The quotes above also highlight the degree of self-
reflection allowed by Physics Bus experiences (“I’m usually too busy using the thing, to 
think about it”), as suggested by the fourth strand of science learning: being able to 
“reflect on...their own process of learning about phenomena.” 
Students in our study had low awareness of Physics as a subject prior to the 
program, suggesting the potential for informal science programs to fill an important gap 
15 5 
 
in STEM ecosystems. Early childhood is a key time period for sparking interest--65% of 
physical scientists report that their interest in science began before middle school 
(Maltese et al., 2014). We found that most students (51-53%) did not have any idea what 
‘Physics’ was before going onboard the Physics Bus, and only 16% showed awareness of 
what topics physics covers aside from being generally “science” related. As noted above, 
those numbers increased substantially after visiting the Physics Bus. Lack of awareness 
of physics in primary school students may reflect their parents and teacher’s limited 
familiarity or interest. Parent interest and attitudes towards science have a significant 
impact on their children’s interest in science (Perera, 2014, Burt & Johnson, 2018), and 
many parents never studied physics – some from lack of interest, others for lack of 
opportunity. Nationwide only 63% of High Schools offer physics (U.S. Department of 
Education, 2014). The percentages of U.S. high school graduates who completed 
coursework in physics increased from 21% in 1990 to 36% in 2009. The increase is good 
news, as Physics is required for many STEM careers, but the percentage of high school 
students taking Physics lags far behind Chemistry (70%) and Biology, which increased 
from 91% to 96% in the same time period (U.S. Department of Education, 2014). 
Primary school teachers are an important source for encouragement and inspiration in 
early childhood, but they broadly lack confidence teaching science, particularly Physics 
and Engineering. Although 77% feel very well prepared to teach Reading/Language arts 
and 73% Mathematics, only 31% feel very well prepared to teach Science; this breaks 
down further with only 13% feel very well prepared to teach grade-level Physical Science 
and 3% feel prepared to teach Engineering. (Banilower et al., 2018). This is a widespread 
problem – in the majority of 21 European countries 4th grade teachers have pronounced 
15 6 
 
lower confidence teaching Physical Science compared to Life Science and Earth Science 
(Klepaker & Almendingen, 2017). Many of us can recall an elementary teacher happily 
bringing in tadpoles, identifying birds, sprouting seeds, and similar life-science 
experiences; how many joyfully took apart electronics to see how they worked, hooked 
up a hobby motor, or generated a high-voltage arc? When typical home and school 
environments are not providing encouragement and positive early childhood impressions 
of Physics, mobile programs offer a way to access those experiences. Though specific 
physics knowledge was not produced by the experience on the bus, the experience did 
contribute to greater understanding of the general field and way of approaching the 
world. These areas fall under Strands 2 and 3 of the National Research Council’s (Bell et 
al., 2009) definition of learning in informal environments (Strand 2: “Come to generate, 
understand, remember, and use concepts, explanations, arguments, models, and facts 
related to science”; Strand 3: “Manipulate, test, explore, predict, question, observe, and 
make sense of the natural and physical world”). 
Mobile science programs reach audiences that might otherwise be underserved. 
The Physics Bus program disproportionately served rural audiences – 28% of programs 
took place in towns and villages of less than 2,500 people, and 81% of programs took 
place in communities of under 50,000 people, where easy access to science museums, 
zoos, aquariums, and other informal science institutions is often limited. Recent work in 
the United Kingdom has highlighted the importance of moving beyond institutions, to 
address the sense of disconnectedness that underserved audiences feel (Dawson, 2018). 
By appearing at public schools and free community events the Physics Bus reaches a 
wide swath of the populace, not just those that are already interested in science. People 
15 7 
 
attending a Reggae festival or chili cook-off probably do not leave home expecting to 
have a science experience, but when they spot the Physics Bus they are drawn in by 
curiosity, peer pressure, or both, and end up having a great time. Organizations 
sponsoring science events at music festivals have made similar observations (Bisbee et 
al., 2020). We believe that offering engagement opportunities broadly and for free is an 
important step towards increasing equity in science. Research has shown that parents 
offer science opportunities to girls only if they express interest, whereas boys receive 
science opportunities regardless of their interest (Alexander et al., 2012). How is anyone 
to know that they like physics if they aren’t given opportunities to have fun with it? 
Ultimately, issues of equity serve the sixth strand of science learning: the ability of 
learners to “think about themselves as science learners and develop an identity as 
someone who knows about, uses, and sometimes contributes to science” (Bell et al. 
2009). 
From our data we identified four primary factors in the Physics Bus’s approach to 
STEM outreach/enrichment that we believe are responsible for the high levels of science 
interest and enthusiasm reported by students, and observed by the researchers: 
1) The experience is self-directed. Self-determination theory posits that honoring 
natural curiosity and giving people the freedom to explore, experiment, and take initiative 
results in positive gains in intrinsic motivation and well-being (Ryan and Deci, 2000). 
Similarly, the concept of free-choice learning highlights the importance of individual 
agency (Falk et al., 2001). Visitors to the Physics Bus have a rich and stimulating 
environment with many opportunities. They can spend time with whatever they like the 
most and exploration is self-paced. There is no didactic signage explaining what you 
15 8 
 
should learn, no required content, no instructors, and no rules. Visitors generate their own 
questions and apply their own ideas and prior experiences to make sense of what they 
experience. They are filled with excitement and wonder. After the program many visitors 
find that they want to learn more and have a keener interest in discovering how things 
work.  
2) Novelty and intensity of the fun. Climbing onboard a transformed schoolbus is, 
in and of itself, a memorable experience. Once inside, people are amazed and surprised 
by what they encounter. As one 3rd grader put it, “you are basically in an unknown 
territory - you don’t know what’s going to happen!” The bright lights, rapidly moving 
parts, and cacophony of sound produced simultaneously by many physics devices in a 
small space—most of which are only held together with hot glue and duct-tape—could be 
chaotic and overstimulating, but nearly all visitors find the experience to be very fun and 
exciting. Reaching out to get shocked by a friend’s fingertip, feeling a cold mist vortex 
ring on your face, marveling at the swirling of rainbow colors, feeling and hearing the 
explosion of pressure from a water rocket. These are the types of novel, surprising, 
intense experiences that have high potential to spark interest (Hidi & Renninger, 2006). 
3) Relevancy. Expectancy-value theory holds that interest is reinforced when a 
topic is perceived as personally relevant, useful, and meaningful (Hidi & Renninger, 
2006; Hulleman & Harackiewicz, 2009). Students were generally very interested in 
making arts & crafts and playing sports and were less interested (or actively resistant to) 
figuring out how things worked or reading. The creative, physical experience of science 
onboard the Physics Bus appealed to them and related to their pre-existing interests. They 
also were impressed with how it used everyday household materials such as a microwave 
15 9 
 
oven, pinata, toy keyboard, vacuum, speaker, humidifier, treadmill motor, slinky, foil-
covered salad bowl, and hand-held mixer. Scientific equipment comes from “another 
world” but these items come from children’s home culture, and their relevance is 
immediately apparent.  
4) Authenticity. The unpolished, homemade nature of the exhibits further 
connects visitors to the world they already know. This authenticity encourages people to 
believe that they too are capable of scientific tinkering. After the Physics Bus program 
many more students felt that they would be able to create a science gadget than before the 
program. When people believe that they can be successful in a given field (in this case, 
applied physics and engineering) they are more likely to maintain interest and persist 
(Pajares, 2005; Wang et al., 2013).  
These factors suggest that the Physics Bus provides a “third space” or hybrid 
space where students and community members can apply funds of knowledge (González 
et al., 2005) and discourse from their home lives. Third-space or hybrid space learning 
environments support individuals who are traditionally marginalized in formal learning 
settings (Moje et al., 2004, Barton & Tan, 2008), facilitating the types of cultural border 
crossing (Krogh & Thomsen, 2005) needed to develop positive science attitudes and self-
concept. Providing a space where individuals’ rich funds of everyday knowledge 
experiences are drawn upon and valued helps to “legitimize multiple ways of 
participating within the science learning community” (Barton & Tan, 2008).  
The importance of these third spaces is highlighted by evidence that anti-science 
beliefs in the U.S. stem not from a lack of science literacy, but rather from attitudinal and 
cultural differences (Morgan et al., 2018; Kahan, 2012). Although confidence in science 
16 0 
 
leaders is relatively high in the U.S. and science as generally perceived as important and 
beneficial, concerns remain about cooperation from the public during health crises, 
continued funding for basic research and climate-related R&D, and maintenance for 
environmental protections (Funk et al., 2019; Besley & Hill, 2020; Scheufele et al., 
2020). Instead of focusing on raising test scores, programs like the Physics Bus focus on 
changing perceptions and attitudes towards science.  
Limitations of this Study 
This study was based on student responses from eight 3rd and 5th grade classes 
from one Elementary School and seven 6th and 7th grade classrooms in one Middle school 
in a rural region of upstate New York, U.S.A. Student recruitment for the study in the 
Middle School was low resulting in a small sample size for this age group; however, 
informal program assessment with a larger group of middle school students (data that 
could not be used for this study) yielded very similar results. Elementary student 
responses used in the study were consistent with a pilot study conducted the year before 
with 65 3rd and 5th grade students from a different rural upstate NY school. 
Future Directions 
Most directly, we would like to see this study extended longitudinally, to find out what 
students remember about the experience a few years later, how it fits into their 
developing attitudes towards science and physics. We would also like to hear from 
children and adults who encounter the program at community events. In some 
16 1 
 
communities, simple postcards sent home with school students may allow data to be 
gathered several years after an original experience (Trautmann & Lewenstein, 2009). 
In addition to research implications, the value demonstrated by this study also has 
implications for practice: With grant funding and private support, programs like the 
Physics Bus could supplement their fee-for-service business model and positively impact 
more underserved regions and high-poverty schools (Fig. 4.2). One-time interventions 
can initiate interest but because beliefs are only moderately stable across time, continuous 
promotion is needed for maintenance (Simpkins et al., 2006). The Physics Bus program 
could increase long-term impact by pointing interested visitors towards ways to extend 
their experience such as activities for the home and classroom, afterschool programs, or 
media. 
Ultimately, we would like to see more consideration of recreational and 
community-based science engagement. Research and program assessment have been 
heavily focused on the end-goals of school achievement and the professional STEM 
pipeline. When the learning of lay people are considered, it is often in the contexts of 
museums (Bell et al., 2009) or in Citizen Science, in which ordinary people contribute 
data to institutionally based science research projects or participate in community-based 
research (Pandya & Dibner 2018). We are interested in the ways that non-scientists can 
(and already do) engage in science and technology related pursuits voluntarily and on 
their own terms by gardening, hunting, fixing things, tinkering in the garage, and sharing 
their knowledge and enjoyment with other community members (Avery, 2013; 
Lewenstein, 2013; Liu & Falk 2013). The founders of the Physics Bus envision a world 
16 2 
 
in which people of all ages dabble and play with science for the fun and joy of it. In the 
process, they will pick up skills and understandings that are useful in daily life. 
Acknowledgements 
CHF was supported by NSF GRFP 2016231353. Many thanks to Isaiah Gutman and Ava 
Cohen for their contributions in processing survey data and program statistics. Thank you 
to Charles Izzo from the Bronfenbrenner Center for Translational Research for early 
advice on evaluation methods.  
Competing Interests  
The authors of this publication have no financial or competing interests. CHF was a 
volunteer co-founder and employee of the Physics Bus program from 2014-2016 and 
volunteered occasionally from 2017-2018. CHF and the Physics Bus directors Erik 
Herman and Chris Bell served together on the Board of the Free Science Inc. from 2017-
2019, a program which received fiscal sponsorship from the parent nonprofit of the 
Physics Bus, the Physics Factory Inc, from 2016-2018. Data for the study was collected 
in 2019. 
Author contributions 
CHF conceived and led design of the study, conducted data collection, analyzed the data, 
and led writing of the manuscript. BVL contributed to study design and theoretical 
framing and helped in drafting the manuscript. 
 
16 3 
 
REFERENCES 
Alexander, J.M., Johnson K.E., & Kelley K. (2012). Longitudinal Analysis of the 
Relations between Opportunities to Learn about Science and the Development of 
Interests Related to Science. Science Education, 39(9), 870-888. 
Allen, S., & Peterman, K. (2019). Evaluating informal STEM education: issues and 
challenges in context. New Directions for Evaluation, 2019(161), 17-33. DOI: 
10.1002/ev.20354 
Avery, Leanne M. (2013). Rural Science Education: Valuing Local Knowledge. Theory 
Into Practice, 52, 28-35. DOI: 10.1080/07351690.2013.743769 
Banerjee, Pallavi A. (2017). Is Informal Education the Answer to Increasing and 
Widening Participation in STEM Education? Review of Education, 5(2), 202-224. 
DOI: 10.1002/rev3.3093 
Banilower, E.R., Smith, P.S., Malzahn, K.A., Plumley, C.L., Gordon, E.M., & Hayes, 
M.L. (2018). Report of the 2018 NSSME+. Chapel Hill, NC: Horizon Research 
Inc.  
Barbagallo, F. (1997). On the road with the Shell Questacon Science Circus: An 
investigation of its contribution to public understanding of science. [Master of 
Science thesis, Australian National University, Canberra] Australian National 
University Open Research Library. openresearch-
repository.anu.edu/au/handle/1885/145927  
Barmby, P., Kind, P.M., & Jones, K. (2008). Examining changing attitudes in secondary 
school science. International Journal of Science Education, 30(8), 1075-1093. 
DOI: 10.1080/09500690701344966 
16 4 
 
Barton, A.C., & Tan, E. (2008). Funds of knowledge and discourses and hybrid space. 
Journal of Research in Science Teaching, 46(1), 50-73. DOI 10.1002/tea.20269 
Basu, S.J., & Barton, A.C. (2007). Developing a sustained interest in science among 
urban minority youth. Journal of Research in Science Teaching, 44(3), 466-489. 
DOI 10.1002/tea.20143 
Bell, P., Lewenstein, B., Shouse, A.W., & Feder, M.A. (2009). Learning science in 
informal environments; people, places, and pursuits. National Academies Press, 
Washington DC. 
Besley, J. C. & Hill, D. (2020). Science and Technology: Public Attitudes, Knowledge, 
and Interest. Washington, DC, National Science Board. 
Bisbee O’Connell, K., Keys, B., Storksdieck, M., & Rosin, M. (2020). Context matters: 
using art-based science experiences to broaden participation beyond the choir. 
International Journal of Science Education, Part B, 10, 166-185. 
DOI:10.1080/21548455.2020.1727587 
Burt, B. & Johnson, J. (2018). Origins of early STEM interest for black male graduate 
students in engineering: A community cultural wealth perspective. School Science 
and Mathematics, 118, 257-270. DOI: 10.111/ssm.12294  
Carlone, H.B. (2004). The cultural production of science in reform-based Physics: girl’s 
access, participation, and resistance. Journal of Research in Science Teaching, 41, 
392-414. DOI:10.1002/tea.20006 
Coburn, W.W., & Aikenhead, G. (1997). Cultural Aspects of Learning Science. Scientific 
Literacy and Cultural Studies Project. 13. 
16 5 
 
Colorado State University. (2020). ‘Little Shop of Physics’ 
https://www.lsop.colostate.edu  
Costa, V.B. (1993). The use of a course question to facilitate learning: How does 
chemistry impact my personal life and society? Journal of College Science 
Teaching, 23(1), 49-53. 
Costa, V.B. (1995). When science is ‘another world’: relationships between worlds of 
family, friends, school, and science. Science Education. 79(3), 313-333. 
Dabney, K.P., Tai, R.H., Almarode, J.T., Miller-Friedmann, J.L., Sonnert, G., Sadler, 
P.M., & Hazari, Z. (2012). Out-of-school time science activities and their 
association with career interest in STEM. International Journal of Science 
Education, Part B, 2(1), 63-79. DOI: 10.1080/21548455.2011.629455 
Dawson, E. (2018). Reimagining publics and (non)participation: Exploring exclusion 
from science communication through the experiences of low-income, minority 
ethnic groups. Public Understanding of Science, 27(7), 772-786. DOI: 
10.1177/0963662517750072  
DeWitt, J. and Archer, L. (2015). Who aspires to a science career? A comparison of 
survey responses from primary and secondary school students. International 
Journal of Science Education, 37(13), 2170-2192, DOI: 
10.1080/09500693.2015.1071899 
Falk, John H., Donovan, Elizabeth, & Woods, Rosalie (Eds.). (2001). Free-choice 
science education : how we learn science outside of school. New York: Teachers 
College Press. 
16 6 
 
Falk, J. H., Pattison, S., Meier, D., Bibas, D., & Livingston, K.. (2018). The contribution 
of science-rich resources to public science interest. Journal of Research in 
Science Teaching, 55, 422-445. DOI: 10.1002/tea.21425 
Funk, C., M. Hefferon, B. Kennedy & C. Johnson. (2019). Trust and Mistrust in 
Americans' View of Scientific Experts. Washington, DC, Pew Research Center. 
Fox, J. (2015). Changes in Urban Youth's Attitude Towards Science and Perception of a 
Mobile Science Lab Experience. [Doctoral dissertation, Columbia University 
Teacher’s College], Columbia Academic Commons. 
https://academiccommons.columbia.edu/doi/10.7916/D85Q4V5X  
Franzblau, C., Romney, C.A., Faux, R., & DeRosa, D. (2011). Mobile laboratory 
programs as vehicles to promote STEM education in K-12 and beyond. 
Proceedings – Frontiers in Education Conference. DOI: 
10.1109/FIE.2011.6142774  
García-Guerrero, M, & Lewenstein, B.V. (2020). Science recreation workshops groups in 
Mexico: a study on an emergent community. International Journal of Science 
Education, Part B, 10(2), 133-148. 
https://doi.org/10.1080/21548455.2020.1719293  
Global Alliance of Community Science Workshops. (2020). 
http://communityscienceworkshops.org  
González, N., Moll, L.C., & Amanti, C. (2005). (Eds.) Funds of knowledge: theorizing 
practices in households, communities, and classrooms. Lawrence Erlbaum 
Associates Publishers.  
Gupta, Arvind. (2020). ‘Toys from Trash’ www.arvindguptatoys.com/toys.html  
16 7 
 
Harrison, Slater. (2020). ‘Science Toymaker’ sciencetoymaker.org  
Häussler, P. & Hoffman, L. (2002). An intervention study to enhance girl’s interest, self-
concept, and achievement in physics classes. Journal of Research in Science 
Teaching, 39(9), 870-888. DOI: 10.1002/tea.10048 
Hidi, S. & Renninger, KA. (2006). The four-phase model of interest development. 
Educational Psychologist, 41(2), 111-127.  
Hughes, G. (2001). Exploring the availability of student scientist identities within 
curriculum discourse: an anti-essentialist approach to gender-inclusive science. 
Gender and Education, 13(3), 275-290. DOI: 10.1080/0954025012006356 2 
Hulleman, C.S. & Harackiewicz, J.M. (2009). Promoting interest and performance in 
high school science classes. Science Reports, 326, 1410-1412. DOI: 
10.1126/science.1177067 
Jenkins, E., & Nelson, N.W. (2005). Important but not for me: student attitudes towards 
secondary school science in England. Research in Science & Technology 
Education, 23(1), 41-57. DOI: 10.1080/02635140500068435 
Kahan, D. M. (2012). Cultural Cognition as a Conception of the Cultural Theory of Risk. 
In S. Roeser, R. Hillerbrand, P. Sandin, & M. Peterson (Eds.), Handbook of Risk 
Theory (pp. 725-759). Springer. 
Klepaker, T.O, & Almendingen, S.F. (2017). How confident are primary school teachers 
to teach science? A comparative European study. Conexão Ciência, 12, 176-184. 
Krogh, L.B, & Thomsen, P.V. (2005). Studying students’ attitudes towards science from 
a cultural perspective but with a quantitative methodology: border crossing into 
the physics classroom. International Journal of Science Education, 27, 281-302. 
16 8 
 
Lewenstein, B. (2013). Recreation in the public communication of science and 
technology. In Elaine Reynoso Haynes, Bertha Michel Sandoval, Miguel Garcia 
Guerrero, and Carmina de la Luz Ramirez (Eds.), Memorias de la XIII Reunión de 
la Red de Popularización de la Ciencia y la Técnica en América Latina y el 
Caribe – XIX Congreso Nacional de Divulgación de la Ciencia y la Técnica (pp. 
89-101). Mexico: SOMEDCyT 
Lyons, T. (2007). Different countries, same science classes: students experiences of 
school science in their own words. International Journal of Science Education, 
28, 591-613. DOI: 10.1080/09500690500339621 
Jones, A.L., & Stapleton, M.K. (2017). 1.2 Million kids and counting – Mobile science 
laboratories drive student interest in STEM. PLOS Biology, 15(7), e1002609 DOI: 
10.1371/journal.pbio.2001692 
Liu, C.C., & Falk, J. H. (2013). Serious Fun: Viewing Hobbyist Activities through a 
Learning Lens. International Journal of Science Education, Part B, 4(4), 343-
355. DOI: 10.1080/21548455.2013.824130 
Maltese, A.V., Melki, C. S., & Wiebke H. L. (2014). The Nature of Experiences 
Responsible for the Generation and Maintenance of Interest in STEM. Science 
Education, 98(6), 937-962. 
Maltese, A. V., & Tai, R. H. (2010). Eyeballs in the Fridge: Sources of early interest in 
science. International Journal of Science Education, 32(5), 669-685. 
Maltese, A.V., Tai, & R. H. (2011). Pipeline Persistence: examining the association of 
educational experiences with earned degrees in STEM among US students. 
Science Education Policy, 95, 877-907. DOI: 10.1002/sce.20441 
16 9 
 
McCormick, W. (2019). Self-efficacy and STEM career interest in Black and Latino 
middle school students: a study on the Next Generation Science Standards. 
[Doctoral dissertation, Loyola Marymount University and Loyola Law School.] 
LMU/LLS Theses and Dissertations. 805. https://digitalcommons.lmu.edu/etd/805 
Milner-Bolotin, M., Antimirova, T., Noack, A., & Petrov, A. (2011). Attitudes about 
science and conceptual physics learning in university introductory physics 
courses. American Physical Society, 7(2), 020107. DOI: 
10.1103/PhysRevSTPER.7.020107 
Moje E.B., Ciechanowski, K., Kramer, K., Ellis, L., Carrillo, R., & Collazo, T. (2004). 
Working toward third space in content area literacy: An examination of everyday 
funds of knowledge and discourse. Reading Research Quarterly, 39(1), 38-70. 
Morgan, M., Collins, W.B., Sparks, G.G., & Welch, J.R. (2018). Identifying relevant 
anti-science perceptions to improve science-based communication: the negative 
perceptions of science scale. Social Sciences, 7(4), 64 DOI: 
10.3390/socsci7040064 
Morrell, C. & Parker, C. (2015). Solving the Education Equation: a new model for 
improving STEM workforce outcomes through academic equity. National Alliance 
for Partnerships in Equity Report. 
National Research Council. (2015). Identifying and Supporting Productive STEM 
Programs in Out-of-School Settings. Board of Science Education, Division of 
Behavioral and Social Sciences and Education. Washington, DC: The National 
Academies Press. DOI: 10.17226/21740 
17 0 
 
National Research Council. (2011). Successful K-12 STEM Education: Identifying 
Effective Approaches in Science, Technology, Engineering, and Mathematics. 
Washington, DC: The National Academies Press. DOI: 10.17226/13158  
Noonan, R. (2017). STEM Jobs: 2017 Update (ESA Issue Brief # 02-17). Office of the 
Chief Economist, Economics and Statistics Administration, U.S. Department of 
Commerce.  
Osborne, J., & Collins, S. (2001). Pupils’ views of the role and value of the science 
curriculum: a focus-group study. International Journal of Science Education, 
23(5), 441-467. DOI: 10.1080/09500690010006518 
Osborne, J., Simon, S., & Collins, S. (2003). Attitudes towards science: A review of the 
literature and its implications. International Journal of Science Education, 25(9), 
1049-1079 DOI: 10.1080/0950069032000032199 
Pajares, F. 2005. Gender Differences in Mathematics Self-Efficacy Beliefs. In A. M. 
Gallagher & J. C. Kaufman (Eds.), Gender differences in mathematics: An 
integrative psychological approach (pp. 294–315). Cambridge University Press. 
Pandya, R. & K. A. Dibner. (Eds). (2018). Learning through Citizen Science: Enhancing 
Opportunities by Design. Washington, DC, National Academies Press. 
Perera, L.D.H. (2014). Parent’s attitudes towards science and their children’s science 
achievement. International Journal of Science Education, 36, 3021-3041.  
Regan, E. & DeWitt, J. (2014). Attitudes, interest, and factors influencing STEM 
enrolment behavior: an overview of relevant literature. In E.K. Henriksen., J. 
Dillon, & J. Ryder, (Eds.), Understanding Student Participation and Choice in 
Science and Technology Education (pp 63-88). Springer. 
17 1 
 
Renninger, K.A., & Hidi, S. (2002). Student interest and achievement: Developmental 
issues raised by a case study. In A. Wigfield & J.S. Eccles (Eds.), A Vol. in the 
educational psychology series. Development of achievement motivation. (pp. 173-
195). Academic Press. DOI:10.1016/B978-012750053-9/50009-7  
Ryan, R.M., & Deci, E.L. (2000). Self-determination theory and the facilitation of 
intrinsic motivation, social development, and well-being. American Psychologist, 
55, 68-78. 
Scheufele, D.A., Brossard, D. Freiling, I., & Krause, N.M. (2020). How Not to Lose the 
COVID-19 Communication War. Issues in Science & Technology, 2020 (17 
April). 
Schreiner, C. & Sjøberg, S. (2007). Science education and youth’s identity construction – 
two incompatible projects? In D. Corrigan, J. Dillon, & R. Gunstone. (Eds.), The 
Re-emergence of Values in the Science Curriculum (pp. 231-247). Sense 
Publishers. 
Simpkins, S.D., Davis-Kean, P.E., & Eccles, J.S. (2006). Math and science motivation: a 
longitudinal examination of the links between choices and beliefs. Developmental 
Psychology, 42, 70-83. 
Sjøberg, S. (2005). How do learners in different cultures relate to science and 
technology? Asia-Pacific Forum on Science Learning and Teaching, 6. 
Steinke, J. (2017). Adolescent girl’s STEM identity formation and media images of 
STEM professionals: considering the influence of contextual cues. Frontiers in 
Psychology, 8, 716. 
17 2 
 
St. John, M., Carroll, B., Hirabayashi, J., Huntworth, D., Ramage, K., & Shattuck, J. 
(2000). The Community Science Workshops: a report on their progress. Inverness 
Research Associates. Inverness-research.org/reports/2000-01-Rpt-CSW-
ProgressRpt.PDF 
Trautmann, C. & B. V. Lewenstein (2009). Long-term Impact of Museums. Informal 
Learning Review, 95, 1-5. 
Tytler, R., & Osborne, J. (2012). Student attitudes and aspirations towards science. In B. 
Fraser, K. Tobin, & C.J., McRobbie (Eds.), Second International Handbook of 
Science Education (pp 597-626). Springer Publishing. 
U.S. Dept of Commerce Economics and Statistics Administration. (2017). Women in 
STEM: 2017 Update. (ESA Issue Brief #06-17). 
U.S. Department of Education. (2014). Data Snapshot (College and Career Readiness). 
(Issue Brief #03-14). Office for Civil Rights.  
Wang, M.T., & Degol, J. (2013). Motivational pathways to STEM career choices: Using 
expectancy-value perspective to understand individual and gender differences in 
STEM fields. Developmental Review, 33:304-340. 
Xu, J., Coats, L.T., & Davidson, M.L. (2012). Promoting student interest in science: the 
perspectives of exemplary African American teachers. American Educational 
Research Journal, 49(1), 124–154. DOI: 10.3102/0002831211426200 
  
17 3 
 
APPENDIX 
 
Chapter 1 Supplemental Figures 
 
 
Supplemental Figure 1. Landmark-based PCA morphospace with a range of carangimorph 
representatives. Outgroups used: Coryphaena hippurus, Mene maculata, Naucrates ductor, 
Leptobrama mulleri, and Sphyraena lucasana. Only the transitional fossil flatfishes (†Amphistium 
and †Heteronectes) and Psettodidae overlap with the outgroups in the morphospace. N=71 
species. 
 
 
  
17 4 
 
 
 
Supplemental Figure 2. Ancestral state reconstruction of feeding type using an alternate 
phylogeny (Byrne et al. 2018) that includes members of Rhombosoleidae. As discussed in the 
text, we used outgroup comparison to fossil flatfishes to establish Turbot-type feeding as the 
ancestral state for Pleuronectiformes. Turbot-type feeding transitioned to Plaice-type feeding 
independently multiple times, including in the lineage leading to Rhombosoleidae. 
17 5 
 
 
 
Supplemental Figure 3. Reconstruction of feeding type (without imposing a prior likelihood at the 
root-node) infers a Plaice-type ancestral state. There is convergent evolution of Sole-type feeding. 
Plaice-type transitioned to Turbot-type multiple times in the lineages leading to Psettodidae, 
Scophthalmidae, and Paralichthyidae. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17 6 
 
Chapter 1 Supplemental Tables 
 
Supplemental Table 1.1. Specimens Studied and Vertebral Number. 
 
17 7 
 
 
17 8 
 
 
 
17 9 
 
Supplemental Table 1.2.  
 
 
 
18 0