Handed Behavior in Hagfish--an Ancient Vertebrate Lineage--and a Survey of Lateralized Behaviors in Other Invertebrate Chordates and Elongate Vertebrates.

Abstract

Hagfish represent an ancient lineage of boneless and jawless vertebrates. Among several curious behaviors they exhibit, solitary individuals in one dominant genus of hagfish (Eptatretus spp.) regularly rest in a tightly coiled posture. We present the first systematic treatment of this distinctive behavior. Individual northeastern Pacific hagfish (E. stoutii) exhibited significant handedness (preferred orientation of coiling). However, right-coiling and left-coiling individuals were equally common in the population. Individual hagfish likely develop a preference for one direction by repeating the preceding coiling direction. We also revisit classical accounts of chordate natural history and compare the coiling behavior of Eptatretus with other handed or lateralized behaviors in non-vertebrate chordates, lampreys, and derived vertebrates with elongate bodies. Handed behaviors occur in many of these groups, but they likely evolved independently. In contrast to vertebrates, morphological asymmetries may bias lateralized larval behaviors toward one side in cephalochordates and tunicates. As a consequence, no known handed behavior can be inferred to have existed in the common ancestor of vertebrates.

Full text

MBL Published by the Marine Biological Laboratory
Biological
Discovery
in Woods Hole
THE
BIOLOGICAL
BULLETIN
April 2014 Volume 226 • Number 2

Handed Behavior in Hagfish—an Ancient Vertebrate
Lineage—and a Survey of Lateralized Behaviors in
Other Invertebrate Chordates and Elongate
Vertebrates
TETSUTO MIYASHITA* AND A. RICHARD PALMER
Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2E9
Abstract. Hagfish represent an ancient lineage of bone-
less and jawless vertebrates. Among several curious behav-
iors they exhibit, solitary individuals in one dominant genus
of hagfish (Eptatretus spp.) regularly rest in a tightly coiled
posture. We present the first systematic treatment of this
distinctive behavior. Individual northeastern Pacific hagfish
(E. stoutii) exhibited significant handedness (preferred ori-
entation of coiling). However, right-coiling and left-coiling
individuals were equally common in the population. Indi-
vidual hagfish likely develop a preference for one direction
by repeating the preceding coiling direction. We also revisit
classical accounts of chordate natural history and compare
the coiling behavior of Eptatretus with other handed or
lateralized behaviors in non-vertebrate chordates, lampreys,
and derived vertebrates with elongate bodies. Handed be-
haviors occur in many of these groups, but they likely
evolved independently. In contrast to vertebrates, morpho-
logical asymmetries may bias lateralized larval behaviors
toward one side in cephalochordates and tunicates. As a
consequence, no known handed behavior can be inferred to
have existed in the common ancestor of vertebrates.
Introduction
Vertebrates exhibit a surprising variety of handed behav-
iors from left-footed food grasping by parrots (Harris, 1989;
Rogers and Workman, 1993) to right-pawed face wiping by
toads (Naitoh and Wassersug, 1996). However, uncertain-
ties remain about how widespread individual- or popula-
tion-level handed behaviors are across the more basal lin-
eages of vertebrates. Rooted between jawed vertebrates and
non-vertebrate chordates along with lampreys (Fig. 1;
Heimberg et al., 2010), hagfish (Myxinoidea) are peculiar
eel-like, boneless, jawless, and sightless fish that exhibit
several unique behaviors including a conspicuously asym-
metrical one. They produce copious amounts of slime to
deter predators (Strahan, 1963; Martini, 1998; Zintzen et al.,
2011). They form a “traveling knot” to escape a grip, to get
them rid of their own slime, or to anchor the body when
ripping flesh while scavanging (Adam, 1960). And, at least
in one dominant genus (Eptatretus), the body forms a tight
coil when at rest (Strahan, 1963; Martini, 1998) (Fig. 2A).
A deep origin of the lineage, its primitive appearance, and
these other unique behaviors make hagfish an interesting
model for studying lateralized behavior in the earliest ver-
tebrates.
Both clockwise and counterclockwise coiling occurs fre-
quently in Eptatretus, a genus of hagfish that contains more
than 70% of the approximately 70 known hagfish species
(Fernholm, 1998; Kuo et al., 2003). Previous accounts of
hagfish behaviors casually mention that healthy individuals
of Eptatretus coil at rest, both in the field and in captivity,
unless burrowing in soft sediment or confined in a narrow
space (Strahan, 1963; Martini, 1998). In the first systematic
treatment of this behavior, we use the northeastern Pacific
inshore hagfish Eptatretus stoutii (Lockington, 1878) to test
whether (1) individual E.stoutii exhibit a preferred direction
of coiling, and (2) the population exhibits an overall left or
right bias in this behavior. We also present a synthesis of
coiling and other lateralized behaviors from a broad sample
of other elongate vertebrates and non-vertebrate chordates.
Do lateralized behaviors occur in other elongate vertebrates
Received 16 August 2013; accepted 18 February 2014.
* To whom correspondence should be addressed. E-mail: tetsuto@
ualberta.ca
Reference: Biol. Bull. 226: 111–120. (April 2014)
© 2014 Marine Biological Laboratory
111

and non-vertebrate chordates? Are coiling or other lateral-
ized behaviors functionally related or phylogenetically
linked? Answers to these questions test whether or not
lateralized behaviors in non-vertebrate chordates and verte-
brates share a common evolutionary origin.
Materials and Methods
We trapped 40 hagfish (Eptatretus stoutii) from a depth
of about 80 m in Barkley Sound, British Columbia, Canada
(48°84⬘96.37⬙N; 125°13⬘18.01⬙W). Each individual was
held separately in a 100-liter tank with running seawater
(12–15° C) and a dark cover for 2 to 3 months and moni-
tored as frequently as every 10 to 30 min for 10 to 12 h per
day at the Bamfield Marine Sciences Centre. Coiling direc-
tion was recorded until 50 observations were obtained per
individual. Coiling direction was determined moving from
the tail to the head as viewed from the dorsal side (Fig. 1A).
The head was always to the outside of the spiral regardless
of coiling direction and regardless of whether individuals
were resting on their ventral or dorsal side. We logged a
new observation only after confirming, either by direct
observation of movement or by the animal’s change of
posture or location in the tank, that a hagfish had uncoiled
from its previous resting position. Individuals sometimes
responded to the disturbance of opening the tank cover by
uncoiling (scored for direct observation of movement), or
sometimes they did not respond at all. Hagfish are generally
quite sedentary and rest in the same position and location in
the aquarium over a week in some cases (TM, pers. obs.).
We tested whether coiling direction in each individual
occurred in one direction more often than expected due to
chance by comparing observed frequencies to critical values
for a test of equal proportions. For each individual hagfish,
we calculated a handedness score h:
h⫽nC⫺nCC (1)
Where n
c
is the number of clockwise coiling and n
cc
is the
number of counterclockwise coiling events. The frequency
distribution of hreveals modes of asymmetry (Palmer,
2004). If individual hagfish do not have a preferred coiling
direction, the distribution of hwill be unimodal with a mean
near zero. If individuals have a preferred direction, and if
preferred directions are not heritable, the distribution will be
symmetrically bimodal with a mean near zero. If individuals
have a preferred direction, and if the degree of side domi-
nance is heritable, the distribution could take many shapes,
but the mean should depart from zero.
We also tested whether individual hagfish repeated an
immediately preceding coiling direction more than expected
from the observed frequencies of clockwise or counter-
clockwise coiling for that individual, regardless of preferred
orientation. We calculated zscores for repeating the preced-
ing coiling direction for each individual, using the standard
formula for ztest statistics for test of proportion (calculated
separately for clockwise and counterclockwise coiling):
zi⫽p
ˆ⫺p0
冑
p0共1⫺p0兲
n
⫽
冉
Ri
fi⫺1
冊
⫺pi
2
冑
pi
2共1⫺pi
2兲
fi⫺1
(2)
Figure 2. Hagfish coiling is handed at the individual level but not
biased at the population level. (A) Examples of clockwise, C, and coun-
terclockwise, CC, coiling of individual hagfish at rest. (B) Bimodal fre-
quency distribution of individual handedness h; a mean near zero (n⫽40;
␮
⫽⫺2.46 ⫾7.07; 95% confidence interval) indicates random orientation
of preferred coiling direction.
Figure 1. Phylogenetic relationships among major chordate lineages
and outgroups discussed in this paper. Hagfish and lampreys represent two
living lineages of jawless vertebrates. Tree topology follows Heimberg et
al. (2010).
112 T. MIYASHITA AND A. R. PALMER

Where R
i
is the frequency of events in which an individual
irepeated the same coiling direction as the immediately
preceding observation, where f
i
is the overall frequency of
that coiling orientation, and where p
i
is a proportion of that
direction in 50 observations. A square of p
i
estimates the
probability of the same coiling direction occurring twice
consecutively due to chance based on a proportion of that
direction in a full sample of 50 observations for that indi-
vidual. The first observation of each coiling orientation was
excluded because the preceding coiling direction was un-
known. To evaluate whether hagfish repeat the same coiling
orientation at the population level more often than expected
due to chance, we calculated z
t
for the entire data set for
each coiling direction by substituting each term in Equation
2 with the sum of all individuals.
To rule out the possibility that individual hagfish devel-
oped a preference during the experiment, we performed
paired Wilcoxon signed rank tests on both the strength of
handedness 兩h兩and the frequency of repeated coiling orien-
tation R
t
in the data set of all individuals between the first 10
observations and the last 10 observations. We also calcu-
lated Spearman’s rank correlation coefficient (r
s
) with body
length for these two variables to test whether body size
affected individual handedness or repeatability of coiling
direction.
Results
Out of 40 individual hagfish, 29 exhibited significantly
handed coiling at a⫽0.05. However, all but one individual
coiled in one direction (clockwise [C] or counterclockwise
[CC]) 30 or more times in the same direction out of 50
observations (Table 1; h⬎10 or h⬍⫺10). Only three
individuals departed from the critical values by a count of
two or more. The ratio between statistically significant C
coilers and CC coilers (16:13) did not differ significantly
from equal odds (50:50) (ztest for equal proportions; P⫽
0.71), and the frequency distribution of individual handed-
ness scores was clearly bimodal with a mean near zero (Fig.
2B). No individual fell in the interval around equal odds (h
⫽0). Within the clearly identified C and CC coilers, about
half of the individuals in each group coiled in that direction
15 to 25 times more than in the other direction. Both C and
CC coilers have long tails of frequency distribution away
from even odds, and four individuals coiled almost exclu-
sively in one direction (40 or more times out of 50). Among
sequential observations, 30 of 40 individuals repeated the
immediately preceding coiling direction more often than
expected from the observed frequencies of C and CC coil-
ing, regardless of their overall preferred directions, and four
more individuals did so in one of the coiling directions
(Table 1). In the entire data set, z
t
scores of repeated coiling
calculated for each coiling direction (C: n⫽40; df ⫽39;
z
sC
⫽30.47; P⬍0.01; CC: n⫽40; df ⫽39; z
sCC
⫽29.93;
P⬍0.01) were highly significant. Therefore, most individ-
uals repeated the same coiling directions more often than
expected due to chance at the population level.
Table 1
Coiling statistics for individual hagfish used in this study, ordered by
increasing body length (BL) in mm
Specimen BL nCn
CC
n
rC
n
rCC
1 260 44 6 39 1
2 294 32 18 (§) 26 12
3 302 14 36 4 26
4 317 35 15 27 7
5 319 21 29 (§§) 8 17
6 322 31 19 (§§) 21 9
7 324 14 36 5 27
8 328 20 30 (§§) 9 19
9 345 17 33 10 27
10 348 12 38 4 30
11 350 19 31 (§§) 13 24
12 366 17 33 9 25
13 367 31 19 (§§) 22 11
14 368 7 43 5 40
15 371 19 31 7 21
16 380 35 15 26 6
17 386 37 13 31 8
18 393 15 35 7 27
19 419 31 19 (§§) 18 7
20 419 37 13 26 (§) 3
21 421 12 38 4 29
22 433 35 15 26 7
23* 446 10 40 1 (§§) 30 (§)
24 448 34 16 21 (§) 4 (§)
25 452 36 14 29 8
26 465 5 45 0 (§§) 40 (§)
27* 470 32 18 (§) 16 (§§) 3 (§§)
28 472 37 13 27 4
29 477 15 35 8 28
30 480 36 14 29 6
31 482 15 35 6 26
32 493 13 37 3 26 (§)
33 498 36 14 27 5
34* 502 17 33 5 20 (§)
35 508 15 35 4 24
36* 514 14 36 3 (§) 24 (§)
37* 523 35 15 23 3 (§§)
38 576 20 30 (§§) 10 19
39 586 48 2 45 (§§) 0 (§§)
40 698 18 32 (§) 10 24
* Gravid females.
Statistical abbreviations: n
C
, number of clockwise events; n
CC
, number
of counterclockwise events (H
0
:n
C
⫽n
CC
; critical values from Table Q;
Rohlf and Sokal 1995); n
rC
, number of clockwise events that were pre-
ceded by a clockwise coiling orientation (tested for zscore with H
0
:p
rC
⫽
p
C
2
; see Methods); n
rCC
, number of counterclockwise events that were
preceded by a counterclockwise coiling orientation (tested for zscore with
H
0
:p
rCC
⫽p
CC
2
; see Methods).
Symbols in parentheses indicate level of significance (§: failure to reject
H
0
at
␣
⫽0.01; §§: failure to reject H
0
at
␣
⫽0.05).
Appendix presents the complete table of statistical tests.
113HANDED HAGFISH

Individual hagfish did not reinforce their coiling orienta-
tion during the study. The paired Wilcoxon signed rank test
between the first 10 and last 10 observations revealed no
significant difference either in the strength of handedness
(兩h兩:n⫽40; df ⫽39; WS [sum of Wilcoxon’s signed ranks]
⫽222.5;
␣
⫽0.01) or in the frequency of repeating the
same coiling orientation twice consecutively (R
t
:n⫽40;
df ⫽39; WS ⫽457.5;
␣
⫽0.01). No significant correlation
was observed between body size and the strength of handed
behavior (兩h兩:n⫽40; df ⫽39; r
s
⫽0.096; T⫽0.59; p⫽
0.55) or the frequency of repeated coiling in the same
direction (R
t
:n⫽40; df ⫽39; r
s
⫽-0.304; T⫽1.97; p⫽
0.056) even though body lengths ranged from 260 to 698
mm.
Discussion
Coiling behavior of individual hagfish is handed
Individual hagfish (Eptatretus stoutii) clearly exhibited
handed coiling behavior in a laboratory setting. The major-
ity (29 of 40 individuals) showed a statistically significant
preference toward one coiling direction. The bimodal dis-
tribution of individual handedness scores (Fig. 2B) indicates
that preferred coiling orientation varied at random among
individuals, which implies no genetic determination to pre-
ferred direction (Palmer, 2004, 2005). Frequency-dependent
selection has been argued to maintain roughly equal fre-
quencies of right- and left-bending mouth-morphs in one
animal example: scale-eating cichlid fish (Hori, 1993).
However, doubts have been raised about whether direction
of mouth bending is actually inherited in this species, and
increasing evidence suggests that it may be induced via
developmental plasticity in response to strongly lateralized
behavior (Palmer, 2010; Kusche et al., 2012; Lee et al.,
2012).
For hagfish, coiling is a stable resting posture that may
help them avoid detection by predators or reduce drag. But
we see no obvious advantage to coiling in a particular
direction. We have also not found any anatomical correlates
of coiling orientation. Even the right-biased gonads appear
not to bias coiling orientation, likely because they are sus-
pended near the midline in the visceral cavity (Marinelli and
Strenger, 1956). Among five gravid females we observed
(noted with an asterisk [*] in Table 1), the ratio of signifi-
cant C to CC coilers is 2:3 at
␣
⫽0.05 and 1:3 at
␣
⫽0.01.
So both C and CC coilers occurred among gravid females
bearing large gonads. The number of gill pouches occasion-
ally differs between the right and left sides (Martini and
Beulig, 2013), but such asymmetry is rare. None of more
than 20 Eptatretus specimens we dissected exhibited any
gill asymmetry, and no consistent bias toward one side
appears to exist in cases reported in the literature.
The handed coiling behavior of individual Eptatretus is
most easily explained as a reinforced behavior following a
random initial choice, as observed in paw use by mice
(Ribeiro et al., 2011). Indeed, current coiling direction
appears to strongly bias subsequent choice of coiling direc-
tion in hagfish. Even though no significant relationship
existed between body size and either the strength of hand-
edness or the tendency to repeat the same coiling orienta-
tion, young E. stoutii clearly developed preferred coiling
directions at or before the smallest sizes we examined (body
length of 260 mm).
Coiling and knotting behavior in hagfish
Although many species of the dominant hagfish genus
Eptatretus coil, coiling does not appear to occur in one
well-studied species of the other common genus, Myxine
glutinosa (Strahan, 1963). So, either coiling behavior
evolved once in Eptatretus, or the lack of coiling behavior
in Myxine is a secondary loss. Although neither lineage is
paraphyletic relative to the other (Fernholm, 1998; Kuo et
al., 2003, 2010; Chen et al., 2005; Fernholm et al., 2013),
Eptatretus is generally believed to retain more plesiomor-
phic morphological features relative to the specialized bur-
rower Myxine (Strahan, 1963; Martini, 1998; Miyashita,
2012), so the lack of coiling in Myxine may be derived. A
recent molecular phylogenetic analysis resolved two previ-
ously known species of Eptatretus into a newly designated
genus Rubicundus as a sister group to eptatretines and
myxinines (Fernholm et al., 2013). Tests for coiling behav-
ior in Rubicundus and other species of Eptatretus would
therefore resolve whether coiling behavior was an ancestral
state in hagfish.
Species of both Eptatreutus and Myxine exhibit another
conspicuously asymmetric behavior. They tie themselves
in—and slip through—a knot to escape from attack, to rid
themselves of their mucous secretions and other debris on
the body, and in macrophagous feeding (Adam, 1960; Stra-
han, 1963; Martini, 1998; Zintzen et al., 2011). Such knots
also come in right-handed (the body crosses over the head
before the tail passes through the loop) and left-handed
forms (the body crosses under the head before the tail passes
through the loop) from either head or tail. We did observe
this behavior occasionally, but not frequently enough for us
to test for concordance with coiling orientation. Nonethe-
less, it would be interesting to know whether either Myxine
or Eptatretus species exhibit consistent handed behavior in
knotting and whether, in Eptatretus, chirality of knotting
correlates with chirality of coiling.
Hagfish coiling behavior is distinct from those of other
vertebrates
Coiling behaviors occur in several elongate vertebrates
(Table 2). Some notable examples are the following:
114 T. MIYASHITA AND A. R. PALMER

Table 2
Examples of coiling, curling, or spiral behavior discussed in this paper
Asymmetry mode* Life history† Morphology§ Ecological context¶
Taxa and behavior Ind. Pop. Pref. Life stage Sex Head pos. Coil form Setting Function Seasonality Ref#
Hemichordata
Enteropneust (various
abyssal taxa); spiral
fecal trail
— — — Adult — Out Tight; multiple Aquatic;
substrate
Feeding Perennial 1
Enteropneust (Glandiceps
hacksi); spiral
swimming
Yes Yes C Adult — N/A Loose; multiple Aquatic;
suspension
Dispersal Seasonal? 2
Cephalochordata
Lancelets
(Branchiostoma); spiral
swimming
Yes Yes C Late larva — N/A N/A Aquatic;
suspension
Dispersal Perennial 3
Lancelets
(Branchiostoma);
resting on side
No No N/A Adult — N/A N/A Aquatic;
substrate
Resting;
defense?
Perennial 3
Tunicata
Ascidian tadpoles; spiral
swimming
Yes Yes C Larva — N/A N/A Aquatic;
suspension
Dispersal Perennial 3
Myxinoidea
Hagfish (Eptatretus);
coiling at rest
Yes No C, CC Adult Both Out Tight; multiple Aquatic;
substrate
Resting;
defense?
Perennial 4
Hagfish (Eptatretus,
Myxine): traveling knot
— — — Adult Both N/A Tight; single Aquatic Defense;
feeding
Perennial 5
Petromyzontiformes
Lamprey (Petromyzon);
resting on side
No No N/A Larva — N/A N/A Aquatic;
substrate
Resting;
defense?
Perennial 4
Lamprey (Petromyzon);
males wrapping around
females
— — — Adult M Out Loose; multiple Aquatic;
suspension
Mating Seasonal 6
Gnathostomata
Sturgeon (Acipenser);
rotational swimming
Yes No C, CC Juvenile — N/A N/A Aquatic;
suspension
Locomotion Perennial 7
Eel (chlopsids, congrids,
and muraenids);
leptocephalus curling
— — — Larva — In Loose; multiple Aquatic;
suspension
Defense Perennial 8
Pricklebacks & gunnels
(stichaeids & pholids);
curling around egg
mass
— — — Adult M Out Loose; single Aquatic;
substrate
Parental
care
Seasonal 9
Catfish (silurids) and
loaches (cobitids);
males enfolding
females
— — — Adult M Out Loose; single Aquatic;
suspension
Maring Seasonal 10
Wolffish (Anarichas);
male rolls over to side
and bends in U-shape
— — — Adult M Out Loose; single Aquatic;
substrate
Courtship Seasonal 11
Lungfish (Protopterus);
curling within
aestivation burrow
— — — Adult Both Out Loose; single Terrestrial;
burrow
Resting Seasonal 12
Elongate amphibians
(salamanders,
caecilians, and
lysorophoids); curling
for various purposes
— — — Adult Both In, out Loose; single,
multiple
Terrestrial;
substrate;
burrow
Resting;
defense;
parental
care
Perennial,
seasonal
13
continued
115HANDED HAGFISH

●Male lampreys wrap around females during spawning
(Beamish and Neville, 1992; Lorion et al., 2000).
●Leptocephalus larvae of three families of eels (Chlop-
sidae, Congridae, and Muraenidae) coil when sus-
pended in water, and at least the latter two families
exhibit both clockwise and counterclockwise coiling in
published figures (Miller, 2009; Miller et al., 2013).
●Stichaeid and pholid perciforms (pricklebacks and gun-
nels) guard their egg masses by curling around them
(Qasim, 1957; Hughes, 1986; Coleman, 1992).
●Silurid and cobitid males (catfishes and loaches) coil
around conspecific females during spawning (Maehata,
2002; Bohlen, 2008).
●Males of wolffish (Anarhichas) roll over to one side
and bend the body in inverted U shape prior to mating
(Johannessen et al., 1993).
●Lungfish curl from the tail first with the snout pointing
upward in estivation burrows (Johnels and Svennson,
1954; Greenwood, 1986).
●Lysorophoids (an extinct family of elongate lepospon-
dyl tetrapods) are often found coiled within burrows or
nodules (Wellstead, 1991). At least one specimen
(UCLA VP 2801; Olson, 1971) is coiled clockwise; a
reconstruction based on several specimens (Olson and
Bolles, 1975) also exhibits clockwise coiling.
●Elongate amphibians such as salamanders and caecil-
ians coil under many conditions including resting,
brooding eggs, defense, and estivation (Cochran, 1911;
Heatwole, 1960; Brodie, 1977; Brodie et al., 1984;
Trauth et al., 2006; Fontenot and Lutterschmidt, 2011).
●Elongate squamates, especially snakes, coil for many
functional purposes including resting, feeding, defense,
estivation, and hibernation (Roth, 2003; Heatwole et
al., 2007).
Although these behaviors are all described as coiling or
curling, a comparative analysis reveals many differences
among them and from the coiling behavior of hagfish (Table
2). These differences concern modes of asymmetries, life-
history traits, morphology of coiling, and ecological con-
texts. Unfortunately, for none of these behaviors do we
know whether (a) individuals have a preferred orientation;
or (b) a population bias exists. Either they have not been
studied systematically, or the presence of handedness is
inconclusive. For example, 3 out of 30 individuals of cot-
tonmouth snakes coiled clockwise more frequently than
counterclockwise (P⬍0.05; Roth, 2003). However, a fol-
low-up experiment revealed no such preference in the same
taxon or in the closely related copperhead snakes (Heatwole
et al., 2007). In some cases, sample size appears sufficiently
large for a statistical test but quantitative data were not
reported (e.g., for the lysorophoid Brachydectes, a single
locality has yielded more than 40 nodules, each likely
containing a coiled skeleton; Hembree et al., 2005).
Clearly, such coiling behaviors warrant systematic study
in these other animals. Nonetheless, the available evidence
suggests that Eptatretus coiling is unique for having the
head outside the spiral, showing strong individual prefer-
ence for either direction, and using it as a primary resting
posture in contact with the substrate, regardless of sex,
season, or life stage.
Table 2 (Continued)
Asymmetry mode* Life history† Morphology§ Ecological context¶
Taxa and behavior Ind. Pop. Pref. Life stage Sex Head pos. Coil form Setting Function Seasonality Ref#
Snakes (Agkistrodon);
coiling during rest
Yes? Yes? C? Adult Both In, out Tight; multiple Terrestrial;
substrate
Resting;
defense
Perennial 14
Two different categories of behaviors are listed under the higher taxonomic headings: coiling or curling behaviors in chordates; potentially lateralized
behaviors in outgroups of vertebrates. Behaviors are compared in different categories: mode of asymmetry, life history, morphology, and ecological
contexts. Tests of individual preference and population bias indicate whether handedness develops at the individual or population level. Dash (—) indicates
no information. N/A ⫽not applicable.
ⴱAsymmetry mode. Ind: individuals exhibit a bias towards one side?; Pop: population bias toward one side?; Pref: preferred orientation? (C— clockwise,
CC— counterclockwise).
†Life history. Life stage: life stage that exhibits the behavior (larva, late larva, adult); Sex: (M—male only, both— evident in both sexes).
§Morphology. Head pos: location of head in coiled position? (out— outside of coil, in—inside of coil); Coil form: tightness of coil (tight— body wall
in contact between revolutions, loose— body wall not in contact between revolutions), number of revolutions (single, multiple).
¶Ecological context. Setting: where coiling behavior is observed (aquatic vs. terrestrial; suspension—in the water column, substrate— on the surface of
substratum, burrow—in below-ground burrows); Function: the possible adaptive significance of the behavior; Seasonality: does behavior occur year-round
(perennial) or only during certain seasons (seasonal)?
#Ref. Source of observations: 1) Smith et al. (2005), Anderson et al. (2011), 2) Urata et al. (2012), 3) Gisle´n (1930), 4) this study, 5) Adam (1960),
Zintzen et al. (2011), 6) Lorion et al. (2000), 7) Izvekov et al. (2014), 8) Miller (2009), Miller et al. (2013), 9) Qasim (1957), Hughes (1986), 10) Maehata
(2002), Bohlen (2008), 11) Johannessen et al. (1993), 12) Greenwood (1986), 13) Trauth et al. (2006), Olson (1971), 14) Roth (2003), Heatwole et al.
(2007).
116 T. MIYASHITA AND A. R. PALMER

Asymmetric behaviors in basal vertebrates and their
relatives likely have independent origins
Even if hagfish coiling is unique in detail, the mere
presence of lateralized behavior may be informative. Is any
lateralized behavior likely to have existed in the last com-
mon ancestor of living vertebrates? This question was pre-
viously considered in the context of various hypotheses that
incorporated bilaterally asymmetrical fossil echinoderms in
chordate evolution (Gisle´n, 1930; Jefferies, 1986; Gee,
1996). However, none of the authors addressed whether or
not any lateralized behaviors in non-vertebrate chordates
could be compared with those in the living vertebrates.
Besides hagfish, some basal vertebrates and vertebrate
relatives exhibit asymmetrical resting postures. Both am-
mocoete larvae of lampreys and the non-vertebrate chordate
lancelets (Cephalochordata) rest on one side of the body. No
marked difference exists between the frequencies of right
and left sides at the individual level (ammocoetes: 41 to 45
observations each for six specimens, with 134 total right-
side down, 125 total left; lancelet adults: 18 to 31 observa-
tions each for four individuals, with 42 right and 59 left;
Gisle´n, 1930). Deep-sea acorn worms (Hemichordata) leave
spiral fecal trails on the sea floor while feeding, and these
are clockwise or counterclockwise with the head oriented
out as in hagfish (Smith et al., 2005; Anderson et al., 2011).
However, no consecutive observations were made to deter-
mine preferred orientation at the individual or population
level.
Swimming does appear to be lateralized in some verte-
brate relatives. Lancelet larvae switch from cilia-driven
counterclockwise spiral swimming to muscle-driven clock-
wise spiraling after their peculiarly asymmetric metamor-
phosis, whereas ascidian tadpole larvae (Tunicata) consis-
tently swim in a clockwise spiral (Gisle´n, 1930). Even some
benthic acorn worms rotate clockwise while swimming
(Urata et al., 2012). Juvenile sterlet sturgeons (Acipenser
ruthenus) have either clockwise or counterclockwise bias in
rotational swimming at the individual level, but no statisti-
cally significant bias was observed at the population level
(Izvekov et al., 2014).
However, these swimming behaviors have a different
morphological basis. In lancelet adults, the preoral hood
extends anteriorly from the right metapleural fold, and the
right series of somites develop half a segment posterior to
the left series (Schubert et al., 2001). Removal of the hood
leads to the loss of spiral swimming (Gisle´n, 1930). Among
tunicates, Grave (1926) attributed the clockwise spiral
swimming of ascidian larvae to oblique contractile fibrillae
within the tail muscles. The tail undulates predominantly
toward the left in Distaplia occidentalis (Berrill, 1950;
McHenry, 2001). Such torsion and yawing may be aug-
mented by a bilaterally uneven mass distribution and the
lack of a bilateral sensory circuit that would facilitate cor-
rective torque against yawing (McHenry and Strother, 2003;
McHenry, 2005). There is no information on what drives
clockwise swimming in acorn worms. As such, these curi-
ous behaviors cannot predict ancestral condition for verte-
brates, and they have no bearing on the origin of coiling
behavior in hagfish.
Concluding remarks
Given the information outlined so far, neither coiling
behaviors nor lateralized behaviors are readily comparable
between chordate lineages. Neither are they phylogeneti-
cally congruent in the currently accepted tree (summarized
in Fig. 1). That is, it takes a smaller number of changes to
assume independent origins of the behaviors than to assume
that a behavior arose in a common ancestor and was vari-
ably modified in some lineages and lost in others. Taken
together, these observations lead to the following conclu-
sions. First, the coiling behavior of Eptatretus has no ap-
parent parallel among non-vertebrate chordates, and various
coiling behaviors in vertebrates almost certainly evolved
independently in each lineage. Second, individual prefer-
ences in coiling directions of Eptatretus likely develop via
repetition of previous coiling directions, but are not biased
at the population level, analogous to the development of
paw preference in mice (Ribeiro et al., 2011) and the
individual bias in direction of rotational swimming in stur-
geons (Izvekov et al., 2014). So, whereas coiling itself may
be advantageous in some way, direction of coiling appears
not to matter. Finally, no known lateralized behavior in
living chordates can be inferred to have been present in the
ancestral vertebrate.
Acknowledgments
This research was supported by scholarships from the
Alberta Ingenuity Fund, Bamfield Marine Sciences Centre,
and Vanier CGS to T.M. and NSERC Discovery Grant
(A7245) to A.R.P., and approved under the animal use
policy at BMSC and the University of Alberta. We thank B.
Anholt, P. Currie, K. Gale, G. Goss, E. Koppelhus, K.
Miyashita, E. Montgomery, J. Pierce, and D. Riddell for
logistical support. T.M. thanks the faculty, students, and
assistants for Embryology 2013— especially A. Accorsi, R.
Behringer, A. Edgar, J. Henry, L. Maya Ramos, L. Linden,
J. Park, and G. Stooke-Vaughan—for six weeks of scientific
adventures at the Marine Biological Laboratory. This paper
is a tribute to one of the themes during the course: linking
natural historical observations to yield an evolutionary in-
sight.
Literature Cited
Adam, H. 1960. Different types of body movement in the hagfish,
Myxine glutinosa. Nature 188: 595–596.
Anderson, T. J., R. Przeslawski, and M. Tran. 2011. Distribution,
117HANDED HAGFISH

abundance and trail characteristics of acorn worms at Australian con-
tinental margin. Deep Sea Res. II 58: 970 –978.
Beamish, R. J., and C.-E. M. Neville. 1992. The importance of size as
an isolating mechanism of lampreys. Copeia 1992: 191–196.
Berrill, N. J. 1950. The Tunicata. The Ray Society, London. 354 pp.
Bohlen, J. 2008. First report on the spawning behaviour of a golden
spined loach, Sabanejewia vallachica (Teleostei: Cobitidae). Folia
Zool. 57: 139 –146.
Brodie, E. D. 1977. Salamander antipredator postures. Copeia 1977:
523–535.
Brodie, E. D., R. A. Nussbaum, and M. DiGiovanni. 1984. Antipreda-
tor adaptations of Asian salamanders (Salamandridae). Herpetologica
40: 56 – 68.
Chen, Y.-W., H.-W. Chang, and H.-K. Mok. 2005. Phylogenetic
position of Eptatretus chinensis (Myxinidae: Myxiniformes) inferred
by 16S rRNA gene sequence and morphology. Zool. Stud. 44: 111–
118.
Cochran, M. E. 1911. The biology of the red-backed salamander
(Plethodon cinereus erythronotus Green). Biol. Bull. 20: 332–349.
Coleman, R. M. 1992. Reproductive biology and female parental care in
the cockscomb prickleback, Anoplarchus purprescens (Pisces:
Stichaeidae). Environ. Biol. Fishes 35: 177–186.
Fernholm, B. 1998. Hagfish systematics. Pp. 33– 44 in The Biology of
Hagfishes, J. M. Jørgensen, J. P. Lonholt, R. E. Weber, and H. Malte,
eds. Chapman, London.
Fernholm, B., M. Nore´n, S. O. Kullander, A. M. Quattrini, V. Zintzen,
C. D. Roberts, H.-K. Mok, and C. H. Kuo. 2013. Hagfish phylog-
eny and taxonomy, with description of the new genus Rubicundus
(Craniata, Myxinidae). J. Zoolog. Syst. Evol. Res. 51: 296 –307.
Fontenot, C. L., and W. I. Lutterschmidt. 2011. Thermal selection and
temperature preference of the aquatic salamander, Amphiuma tridacty-
lum. Herpetol. Conserv. Biol. 6: 395–399.
Gee, H. 1996. Before the Backbone. Chapman and Hall, London. 346
pp.
Gisle´n, T. 1930. Affinities between the Echinodermata, Enteropneusta,
and Chordonia. Zool. Bidr. Upps. 12: 199 –304.
Grave, C. 1926. Molgula citrine (Ald. and Hanc.). Activities and
structure of the free-swimming larva. J. Morphol. 42: 453– 468.
Greenwood, P. H. 1986. The natural history of African lungfishes. J.
Morphol. 190 (S1): 163–179.
Harris, L. J. 1989. Footedness in parrots: three centuries of research,
theories and mere surmise. Can. J. Psychol. 43: 369 –396.
Heatwole, H. 1960. Burrowing ability and behavioral responses to
desiccation of the salamander, Plethodon cinereus. Ecology 41: 661–
668.
Heatwole, H., P. King, and S. G. Levine. 2007. Laterality in coiling
behaviour of snakes: another interpretation. Laterality 12: 536 –542.
Heimberg, A. M., R. Cowper-Sal䡠lari, M. Se´mon, P. J. C. Donoghue,
and K. J. Peterson. 2010. microRNAs reveal the interrelationships
of hagfish, lampreys, and gnathostomes and the nature of the ancestral
vertebrate. Proc. Natl. Acad. Sci. USA 107: 19379 –19383.
Hembree, D. I., S. T. Hasiotis, and L. D. Martin. 2005. Torridorefu-
gium eskridgensis (new ichnogenus and ichnospecies): amphibian aes-
tivation burrows from the Lower Permian Speiser Shale of Kansas. J.
Paleontol. 79: 583–593.
Hori, M. 1993. Frequency-dependent natural-selection in the handed-
ness of scale-eating cichlid fish. Science 260: 216 –219.
Hughes, G. W. 1986. Observations on the reproductive ecology of the
crescent gunnel, Pholis laeta, from marine inshore waters of southern
British Columbia. Can. Field-Nat. 100: 367–370.
Izvekov, E. I., E. A. Kuternitskaya, N. A. Pankova, Y. B. Malashichev,
and V. A. Nepomnyashchikh. 2014. Lateralisation of rotational
swimming but not fast escape response in the juvenile sterlet sturgeon,
Acipenser ruthenus (Chondrostei: Acipenseridae). Laterality 19: 302–
324.
Jefferies, R. P. S. 1986. The Ancestry of the Vertebrates. British
Museum (Natural History), London. 376 pp.
Johannessen, T., J. Gjøsæter, and E. Moksness. 1993. Reproduction,
spawning behaviour and captive breeding of the common wolffish
Anarhichas lupus L. Aquaculture 115: 41–51.
Kuo, C.-H., S. Huang, and S.-C. Lee. 2003. Phylogeny of hagfish
based on the mitochondrial 16S rRNA gene. Mol. Phylogenet. Evol. 28:
448 – 457.
Kuo, C.-H., S.-C. Lee, and H.-K. Mok. 2010. A new species of hagfish
Eptatretus rubicundus (Myxinidae: Myxiniformes) from Taiwan, with
reference to its phylogenetic position based on its mitochondrial DNA
sequence. Zool. Stud. 49: 855– 864.
Kusche, H., H. J. Lee, and A. Meyer. 2012. Mouth asymmetry in the
textbook example of scale-eating cichlid fish is not a discrete dimor-
phism at all. Proc. R. Soc. B 279: 4715– 4723.
Lee, H. J., H. Kusche, and A. Meyer. 2012. Handed foraging behavior
in scale-eating cichlid fish: its potential role in shaping morphological
asymmetry. PLoS One 7:e44670. doi: 10.1371/journal.pone.0044670.
Lorion, C. M., D. F. Markie, S. B. Reid, and M. F. Docker. 2000.
Redescription of the presumed-extinct Miller Lake lamprey, Lampetra
minima. Copeia 2000: 1019 –1028.
Maehata, M. 2002. Stereotyped sequence of mating behavior in the far
Eastern catfish, Silurus asotus, from Late Biwa. Ichthyol. Res. 49:
202–205.
Marinelli, W., and A. Strenger. 1956. Vergleichende Anatomie und
Morphologie der Wirbeltiere. II Lieferung. Myxine glutinosa (L). Franz
Deutlicke, Vienna. Pp. 81–172.
Martini, F. H. 1998. The ecology of hagfishes. Pp. 57–77 in The
Biology of Hagfishes, J. M. Jørgensen, J. P. Lonholt, R. E. Weber, and
H. Malte, eds. Chapman, London.
Martini, F. H., and A. Beulig. 2013. Morphometics and gonadal
development of the hagfish Eptatretus cirrhatus in New Zealand. PLoS
One 8: e78740.
McHenry, M. J. 2001. Mechanisms of helical swimming: asymmetries
in the morphology, movement and mechanics of larvae of the ascidian
Distaplia occidentalis. J. Exp. Biol. 204: 2959 – 2973.
McHenry, M. J. 2005. The morphology, behavior, and biomechanics of
swimming in ascidian larvae. Can. J. Zool. 83: 62–74.
McHenry, M. J., and J. A. Strother. 2003. The kinematics of photo-
taxis in larvae of the ascidian Aplidium constellatum. Mar. Biol. 142:
173–184.
Miller, M. J. 2009. Ecology of anguilliform leptocephali: remarkable
transparent fish larvae of the ocean surface layer. Aqua-Biosci. Monogr.
2: 1–94.
Miller, M. J., M. D. Norman, K. Tsukamoto, and J. K. Finn. 2013.
Evidence of mimicry of gelatinous zooplankton by anguilliform lepto-
cephali for predator avoidance. Mar. Freshw. Behav. Physiol. 45:
375–384.
Miyashita, T. 2012. Comparative Analysis of the Anatomy of the
Myxinoidea and the Ancestry of Early Vertebrate Lineages. Unpub-
lished M.Sc. thesis, the University of Alberta, Edmonton. 407 pp.
Naitoh, T., and R. Wassersug. 1996. Why are toads right-handed?
Nature 380: 30 –31.
Olson, E. C. 1971. A skeleton of Lysorophus tricacarinatus (Amphibia:
Lepospondyli) from the Hennessay Formation (Permian) of Oklahoma.
J. Paleontol. 45: 443– 449.
Olson, E. C., and K. Bolles. 1975. Permo-Carboniferous freshwater
burrows. Fieldiana Geol. 33: 271–290.
Palmer, A. R. 2004. Symmetry breaking and the evolution of develop-
ment. Science 306: 828 – 833.
Palmer, A. R. 2005. Antisymmetry. Pp. 359 –397 in Variation, B.
Hallgrı´msson and B. K. Hall, eds. Elsevier, Cambridge.
118 T. MIYASHITA AND A. R. PALMER

Palmer, A. R. 2010. Scale-eating cichlids: from hand(ed) to mouth.
J. Biol. 9: 11.
Qasim, S. Z. 1957. The biology of Centronotus gunnellus (L.) (Te-
leostei). J. Anim. Ecol. 26: 389 – 401.
Ribeiro, A. S., B. A. Eales, and F. G. Biddle. 2011. Learning of paw
preference in mice is strain dependent, gradual and based on short-term
memory of previous reaches. Anim. Behav. 81: 249 –257.
Rogers, L. J., and L. Workman. 1993. Footedness in birds. Anim.
Behav. 45: 409 – 411.
Rohlf, F. J., and R. R. Sokal. 1995. Statistical Tables, 3
rd
ed. W. H.
Freeman, New York. 199 pp.
Roth, E. D. 2003. ‘Handedness’ in snakes? Lateralization of coiling
behaviour in a cottonmouth, Agkistrodon piscivorus leucostoma, pop-
ulation. Anim. Behav. 66: 337–341.
Schubert, M., L. Z. Holland, M. D. Stokes, and N. D. Holland. 2001.
Three amphioxus Wnt genes (AmphiWnt3, AmphiWnt5, and Amphi-
Wnt6) associated with the tail bud: the evolution of somitogenesis in
chordates. Dev. Biol. 240: 262–273.
Smith, K. L., N. D. Holland, and H. A. Ruhl. 2005. Enteropneust
production of spiral fecal trails on the deep-sea floor observed with
time-lapse photography. Deep Sea Res. I 52: 1228 –1240.
Strahan, R. 1963. The behaviour of myxinoids. Acta Zool. 44: 73–102.
Trauth, S. E., M. L. McCallum, R. R. Jordan, and D. A. Saugey. 2006.
Brooding postures and nest site fidelity in the western slimy salaman-
der, Plethodon albagula (Caudata: Plethodontidae), from an abandoned
mine shaft in Arkansas. Herpetol. Nat. Hist. 9: 141–149.
Urata, M., S. Iwasaki, and S. Ohtsuka. 2012. Biology of the swim-
ming acorn worm Glandiceps hacksi from the Seto Inland Sea of Japan.
Zool. Sci. 29: 305–310.
Wellstead, C. F. 1991. Taxonomic revision of the Lysorophia, Permo-
Carboniferous lepospondyl amphibians. Bull. Am. Mus. Nat. Hist. 209:
1–90.
Zintzen, V., C. D. Roberts, M. J. Anderson, A. L. Stewart, C. D.
Struthers, and E. S. Harvey. 2011. Hagfish predatory behaviour
and slime defence mechanism. Sci. Rep. 1: 131. 6 pp.
Appendix
List of specimens (BL, body length in mm) with statistics used in this paper
Specimen BL n
C
n
CC
hn
rC
n
rCC
P
C
P
CC
z
rC
z
rCC
1 260 44 6 38 39 1 0.88 0.12 2.08 3.48
2 294 32 18 14 26 12 0.64 0.36 4.86 7.07
3 302 14 36 ⫺22 4 26 0.28 0.72 3.08 2.66
4 317 35 15 20 27 7 0.70 0.30 3.55 5.36
5 319 21 29 -8 8 17 0.42 0.58 2.62 3.03
6 322 31 19 12 21 9 0.62 0.38 3.55 4.29
7 324 14 36 ⫺22 5 27 0.28 0.72 4.11 3.00
8 328 20 30 ⫺10 9 19 0.40 0.60 3.73 3.31
9 345 17 33 ⫺16 10 27 0.34 0.66 6.37 4.66
10 348 12 38 ⫺26 4 30 0.24 0.76 4.36 2.87
11 350 19 31 ⫺12 13 24 0.38 0.62 6.97 4.68
12 366 17 33 ⫺16 9 25 0.34 0.66 5.59 3.94
13 367 31 19 12 22 11 0.62 0.38 3.93 5.63
14 368 7 43 ⫺36 5 40 0.14 0.86 14.38 3.14
15 371 19 31 ⫺12 7 21 0.38 0.62 2.95 3.55
16 380 35 15 20 26 6 0.70 0.30 3.20 4.43
17 386 37 13 24 31 8 0.74 0.26 3.78 8.27
18 393 15 35 ⫺20 7 27 0.30 0.70 5.36 3.55
19 419 31 19 12 18 7 0.62 0.38 2.43 2.95
20 419 37 13 24 26 3 0.74 0.26 2.11 2.52
21 421 12 38 ⫺26 4 29 0.24 0.76 4.36 2.54
22 433 35 15 20 26 7 0.70 0.30 3.20 5.36
23* 446 10 40 ⫺30 1 30 0.20 0.80 1.09 1.68
24 448 34 16 18 21 4 0.68 0.32 2.00 2.10
25 452 36 14 22 29 8 0.72 0.28 3.67 7.20
26 465 5 45 ⫺40 0 40 0.10 0.90 -0.20 1.68
27* 470 32 18 14 16 3 0.64 0.36 1.21 0.58
28 472 37 13 24 27 4 0.74 0.26 2.44 3.67
29 477 15 35 ⫺20 8 28 0.30 0.70 6.29 3.89
30 480 36 14 22 29 6 0.72 0.28 3.67 5.14
31 482 15 35 ⫺20 6 26 0.30 0.70 4.43 3.20
32 493 13 37 ⫺24 3 26 0.26 0.74 2.52 2.11
33 498 36 14 22 27 5 0.72 0.28 3.00 4.11
continued
119HANDED HAGFISH

Appendix (Continued)
Specimen BL n
C
n
CC
hn
rC
n
rCC
P
C
P
CC
z
rC
z
rCC
34* 502 17 33 ⫺16 5 20 0.34 0.66 2.46 2.16
35 508 15 35 ⫺20 4 24 0.30 0.70 2.56 2.52
36* 514 14 36 ⫺22 3 24 0.28 0.72 2.04 1.98
37* 523 35 15 20 23 3 0.70 0.30 2.18 1.62
38 576 20 30 ⫺10 10 19 0.40 0.60 4.36 3.31
39 586 48 2 46 45 0 0.96 0.04 0.91 -0.04
40 698 18 32 ⫺14 10 24 0.36 0.64 5.63 4.13
Abbreviations as in text. n
C
, number of clockwise events; n
CC
, number of counterclockwise events; h, handedness score (n
C
⫺n
CC
); n
rC
, number of
clockwise events that were preceded by a clockwise coiling orientation; n
rCC
, number of counterclockwise events that were preceded by a counterclockwise
coiling orientation; P
C
, proportion of clockwise events in 50 observations; P
CC
, proportion of counterclockwise events in 50 observations; z
rC
,zscore for
number of clockwise events that were preceded by a clockwise coiling orientation for test of proportion (P
rC
⫽P
C
2
); z
rCC
,zscore for number of
counterclockwise events that were preceded by a counterclockwise coiling orientation for test of proportion (P
rCC
⫽P
CC
2
). Asterisks (*) indicate gravid
females.
120 T. MIYASHITA AND A. R. PALMER