Swimming Behavior of the Nudibranch Melibe leonina

8 Pages·2002·3.7 MB·

by  K A Lawrence

Reference: BiW. Bull. 203: 144-151. (October 2002)
Swimming Behavior of the Nudibranch Melibe leonina
K. A. LAWRENCE* AND W. H. WATSON III
Zoology Department & Center for Marine Biology, University of New Hampshire. Durham,
New Hampshire 03824; and Friday Harbor Laboratories, University of Washington,
Friday Harbor, Washington 98250
Abstract. Swimming in the nudibranch Melibe leonina Introduction
consists of five types of movements that occur in the fol-
lowing sequence: (1 ) withdrawal, (2) lateral flattening, (3) a Swimming is a common form of locomotion in marine
series oflateral flexions. (4) unrolling and swinging, and (5) gastropods. It has been described for at least 47 species
termination. Melibe swims spontaneously, as well as in (reviewed by Farmer. 1970). and the neural mechanisms
response to different types ofaversive stimuli. In this study, underlying swimming have been investigated in detail in
swimming was elicited by contact with the tube feet of the four of these: Pleurobranchaea californica (Jing and Gil-
predatory sea star Pycnopodia helianthoides, pinching with lette. 1995. 1999), Tritoniadiomedea (Willowsetal., 1973:
forceps, or application of a 1 M KC1 solution. During an Hume ei al.. 1982; Getting, 1983), Clioiw linuicina (Ar-
episode of swimming, the duration of swim cycles (2.7 shavsky et al., 1985: Satterlie. 1985, 1991: Satterlie and
=
0.2 s [mean SEM], n 29) and the amplitude oflateral Spencer, 1985; Satterlie etal.. 1985; Satterlie and Norekian,
flexions remained relatively constant. However, the latency 1996), and Aplysia brasiliana (von der Porten et al., 1982;
between the application of a stimulus and initiation of McPherson and Blankenship, 1991a.b). In general, both
swimming was more variable, as was the duration of an rhythmic swimming and fictive swimming in molluscs are
episode of swimming. For example, when touched with a highly stereotyped and reliably expressed in intact and
single tube foot from a sea star (;; = 32). the latency to semi-intact preparations, as well as in isolated ganglia.
swim was 7.0 2.4 s, and swimming continued for 53.7 These characteristics of molluscan swimming, combined
9.4 s, whereas application of KC1 resulted in a longer with the suitability oftheir nervous systems for neurophys-
latency to swim (22.3 4.5 s) and more prolonged swim- iological experimentation (Willows, 1965). have allowed
ming episodes (174.9 32.1 s). Swimming individuals scientists toelucidate someofthe fundamental neural mech-
tended to move in a direction perpendicular to the long axis anisms responsible for producing swimming and other
of the foot, which propelled them laterally when they were rhythmic behaviors (reviewed in Audesirk and Audesirk,
oriented with the oral hood toward the surface ofthe water. 1985; Getting, 1989).
The results of this study indicate that swimming in Melibe. In opisthobranchs, five general types of swimming have
like that in several other molluscs, is a stereotyped fixed been described, but only three types are common (Farmer,
Tachteisoen cphaatrtaecrtnertihsatticcsa,nableonrgelwiiabtlhytehleicliatregdeinidtehnetifliaabbolreatnoeruy-. t1e9r7o0n),:H(eIx)abpraraanpcohdnisa.l aonrdmAapnltylseiaf)l:ap(p2i)ndgor(saosveinntGraalstuenrdoup--
rons typical of many molluscs, make swimming in this lation (as in Tritonia and Pleurobranchaea): and (3) lateral
nudibranch amenable to neuroethological analyses. bliesntdedinbgy(aFsarinmeDrcm(l1m9n7o0)t.us2)1. sOwfitmheb4y7fslwapipmimnginegitsheprecitehes
mantle or some part ofthe foot, 5 swim using dorsoventral
undulation, and 18 swim using lateral flexions. The latter is
Received 7 November 2001; accepted 28 June 2002. the most common type used by aeolidaceans and dendrono-
*Current address: Highline Community College. 2400 S. 240th Street. taceans. Lateral-bending swimming in these animals does
P.O. Box 98000, Des Moines, WA 4X198-9800. not seem to propel them in a particular direction; rather, it
tTowhomcorrespondence should he addressed. E-mail: win<3'unh.edu appears as ifswimming moves these animals into the water
144
MELIBE SWIMMING BEHAVIOR 145
column where the current may carry them away from po- We found that swimming in Melibe is a stereotyped rhyth-
tential predators. mic behavior that is most readily elicited in the laboratory
Thompson (1976) hypothesized that swimming in opis- by touching animals with the tube feet of the predatory sea
thobranchs evolved as a means of escape. For most of the star Pvcnopoeliu helianthoides. Furthermore, what appears
species studied, this seems to be a feasible explanation, to be random motion during swimming has a fairly predict-
since swimming can be elicited by noxious stimuli such as able directional component, with an animal typically mov-
strong tactile stimulation orcontact with apotential predator ing in apath perpendiculartothe long axis ofits foot. These
(Mauzey et ul.. 1968: Edmunds. 1968; Farmer, 1970; Wil- studies clarify some controversial issues concerning the
lows ct ul., 1973: Page, 1993). Some animals, such as T. swimming behavior of M. leonina and lay the framework
clioineiiea. apparently swim solely as a means of escape for the neurophysiological studies presented in the subse-
(Willows er ai. 1973). In others, such as A. brasiliana. quent paper (Watson et ai, 2002).
which has no known predators, swimming is fairly direc-
tional and may serve a migratory role (Hamilton and Am- Materials and Methods
brose, 1975).
Studies of the opisthobranch Melthe Iconina offer con- Animals
flicting hypotheses as to the function and general character- Specimens of the nudibranch Melibe leonina (Gould.
istics of swimming behavior in this species. In one of the 1852) were collected during May and June of 1994 and
earliest papers on the subject. Agersborg (1921) states that January through June of 1995 in sheltered bays throughout
the position of animals during a swimming episode may the San Juan Archipelago, Washington. Collections were
vary from dorsal aspect up to ventral aspect up. He further made by scuba divers and the animals were maintained in
notes that swimming seems tobe correlated with copulating flow-through seawater tables, typically between 10 and 12"
masses of animals, suggesting that it may be a voluntary C. at the University of Washington's Friday Harbor Labo-
method forfinding mates. In the same paper. Agersborg also ratories. Animals were provided with blades of eelgrass
refers to a methodof"falling" through the watercolumn, by (Zostera marina) to crawl upon, and they fed on planktonic
completely relaxing the body musculature, which looks like crustaceans from the unnltered water supply, supplemented
"a feigned death." Hurst (1968) briefly describes the swim- with Anemia nauplii twice weekly.
ming behavior of this species as occurring only dorsal
aspect up, and does not mention the ecological significance Anal\sis ofnormal swimming
of the behavior. Farmer (1970) concluded that Melibe uses
swimming to move from one kelp blade to the next. Most We analyzed the progression ofacomplete boutofswim-
recently, Bickell-Page (1991; Page, 1993) suggested that ming (from initiation to termination) in 29 animals. Each
swimming is an escape response. However, it is unclear animal was placed in a 50-1 aquarium with a small amount
whichorganisms in the natural habitatofMelibe mightelicit ofeelgrass. Swimming was initiatedby using a 3-ml syringe
M
escape swimming. Mauzey et ul. (1968) and Bickell-Page without a needle to apply 1 ml of 1 KC1 to the skin of
( 1991) have observed the sea starCrossasterpaposuseating either the oral hood or body wall. Not every animal re-
Melibe, and Ajeska and Nybakken (1976). Mauzey et at. sponded to the salt stimulus. Our analyses are based on the
(1968), and Bickell-Page ( 1991 ) have repotted that several 29 animals that swam. The following parameters were then
crab species, including Pugettia producta, will capture and measured: (1) latency between application of the stimulus
eat Melibe. In contrast, it has also been reported that several and initiation ofthe swim; (2) swim duration; (3) numberof
species ofsea stars avoid Melibe, presumably because they complete swim cycles in each swimming episode; (4) av-
find the secretions of its repugnatorial gland repulsive erage swim cycle duration (duration of an episode divided
(Ajeska and Nybakken. 1976; Bickell-Page, 1991). One by the number of swim cycles); and (5) direction (right or
goal ofthis study was todetermine ifMelibe would swim in left) ofthe first and last flexions. Finally, forone animal, we
response to these potential predators, which would indicate measured the duration ofeach individual swim cycle from a
that one function of swimming in this species is escape. videotape of a complete swim episode. All averages are
In this paper we present results from three types of presented as the mean SEM.
experiments concerned with swimming behavior in M. leo- Toassess the magnitude oflateral flexions throughoutthe
nina. First, we analyzed swimming in 29 animals toenhance course ofa swim, three animals were videotaped while they
our understanding of the behavior and to quantify the var- swam in place and the tapes weredigitized formeasurement
ious components of the swim. Second, we sought to deter- of flexion angles. Two loops of 4-0 surgical silk were
mine the types of stimuli and potential predators that elicit attached to the middle ofthe body wall ofeach animal, one
swimming. Third, we assessed the movement of animals on either side, at the point where their body pivots during
through the water column to determine whether swimming swimming. After one day of recovery, animals were indi-
propels animals in random directions or predictable ones. vidually suspended by these loops, ventral aspect up. in an
146 K. A. LAWRENCE AND W. H. WATSON
acrylic plastic chamber. The chamber was supplied with a bottom, then induced to swim using a brief touch on the
continuous flow of natural seawater so that it remained at back of the oral hood with a single sea star tube foot. The
10-12 C. A video camera was mounted above the cham- tube foothadbeenexcisedfromalive Pycnopodia with tine
ber, and the output ofthe camera was recorded onto video- scissors and was held with a pair of self-closing forceps. A
tape. Swimming was induced by dislodging the foot of the line representing the orientation of the foot was marked on
animal from its attachment to the surface tension of the the A-V grids at 5-s intervals. These lines were then plotted
water in the chamber. The video recordings were digitized, in two dimensions and used to calculate the predicted po-
one frame/second, and version 1.55 of the public domain sition of the midpoint of the foot at successive time inter-
NIH Image software program (developed at the U.S. Na- vals. The "variance angle" (theanglebetweenthe actual and
tional Institutes of Health and available on the Internet at predicted position of the foot at the next time point) was
(http://rsb.info.nih.gov/nih-image/)) was used to measure then calculated for each 5-s time interval, averaged, plotted
changes in the angle of the portion of the body anterior to usingpolarcoordinates, andcomparedtothe predicted path.
the pivot point, relative to the region of the body posterior
to that point. Results
Stimuli that elicit swimming Analysis ofnormal swimming
Different stimuli were applied to animals to determine Although Melibe is occasionally observed swimming
their effectiveness in eliciting swimming behavior. The near the ocean surface (Mills, 1994; pers. obs.), typically it
stimuli included ( 1 ) pinching ofthe cerata with self-closing is found attached to eelgrass or kelp blades. Melibe swims
forceps; (2) proddMing of the foot with a glass rod: (3) spontaneously, as well as when forced off the substrate or
application of a 1 KC1 solution to the oral hood or body exposed to a noxious stimulus. Its swimming behavior is
wall; and (4) presence of, or contact with, potential natural characterized by slow, rhythmic, lateral flexions that last for
predators (sea stars Pycnopodiu helianthoides, Henricia 1-2 s (2-4 s fora full swim cycle). During each flexion, the
leviuscula. Pisaster spp: crabs Cancer magister, Scyra body bends laterally into a shape resembling the letter C,
acittifrons. Oregonia gracilis. Cancer prodiictus; and an with the oral hood approaching the tip of the tail (Fig. 1).
anemone found on eelgrass Epiactus prolifera). In each A digitized video of swimming can be viewed at the fol-
experiment, the stimulus was given at time zero, and the lowingwebsite: http://zoology.unh.edu/faculty/win/Melibe/
latency to swim and the duration of any ensuing swim
melibeswimming.htm.
episodes were recorded. In the sea starcontact experiments, The overall organizationofaswimepisode can be broken
a single tube foot was excised from a live sea star, held in down into five basic components, some of which are illus-
self-closing forceps, and brought into contact with the back trated in Figure 1. When not swimming, Melibe is often
of the oral hood of a specimen of Melibe. found attached, by means of its flattened foot, to a vertical
To determine if certain animals commonly found in the substrate, such as a blade of eelgrass, with oral hood open
natural habitat (eelgrass beds) of Melibe were potential and cerata extended (Fig. 1, top). A spontaneous swimming
predators, we performed a series of predation experiments. episode begins with closure ofthe oral hood and release of
5iI0nm-de1invaioqdfuuaMalerliPiuvbmcenowwpiaotihslipfall.aocwEe-pdtihairncottuuhgseh.tsaoenrakcw,arwtahebrsi.cwhTehraleesnoplcoaoncneetdasipinenecd-a csthuoeblsuftmoronatt,,etuh(seucaoblmolptyotnsotemanrottfint1ghewiwftiohtohttdhriesawaranoltle)lre.idoOrmnepdcoieratlilionyn,,thtfehreowmbaottdehyre
6a-s8malhltahmeounundtiborfaeneclhgrawsass, aenxdamlienftedthefroer feovrid2e4nche. Eovferayn i(s2comlaptrereaslsefldatltaetneirnagl)l.y,Sawnidmtmhiencgeramtoavaermeeenxttsencdoensdidsotrosaflalny
attack. Three trials were carried out with each potential alternating lateral-bending, or flexion, of the body that
predator. brings the closed oral hood in close proximity to the tail
(3 flexions). These rhythmic flexions may continue for a
Direction ofswimming period of a few seconds to over an hour. The conclusion ot
The movement ofMelihe through the water column dur- a swim episode is preceded by the anterior tip of the foot
ing a swimming episode was determined for seven animals. unrolling and "probing" for an appropriate substrate on
The goal of this study was to test the hypothesis that this which to settle (4 unrolling and swinging). A swim epi-
species moves ventrally, in a direction roughly perpendic- sode is terminated (5 termination) when the anterior foot
ulartothe long axis ofthefoot. Individual nudibranchs were comes in contact with a suitable substrate and the animal
induced to swim in a 50-1 aquarium that hadx-y coordinates uses its flattened foot to attach to it. Termination does not
drawn on a clear plastic cover placed over the top and on necessarily occur during the first encounterofthe foot with
one side. Each animal was placed on a blade ofeelgrass that a substrate; often the animal will make multiple contacts
was located in the center of the tank and secured to the before ceasing to swim. Moreover, animals will occasion-
\1/1IRE SWIMMING BEHAVIOR 147
Figure 1. Illustrations ofMelihe leonina at rest (top) and swimming (all ventral views). The sequence of
drawingson thebottomshow 50% ofaswimcycle,asthe nudibranchgoesfrom being fully flexed totheright,
to fully flexed tothe left. Note the slight twisting ofihe posteriorportion ofthe foot, which creates a sculling
motion that propels animals in the ventral direction. Scale bar = 1 cm.
ally stop swimming in the water column, without contact of 29 different animals, the average duration of a swim cycle
the foot with a substrate. was 2.7 0.2 s. The magnitude ofthe lateral flexions was
The flexions involved in the swim are not equivalent also quite consistent throughout a swim episode (Fig. 2).
along the entire length ofthe body. In addition to the lateral Other than the first and last few flexions in the swim
flexions, there is also a concurrent twisting of the posterior episode, the contractions ofthebody in bothdirections were
part of the body, so that the foot becomes the leading edge similar in amplitude for most of the episode.
during each lateral flexion (Fig. 1). This "sculling" motion In contrast to the stereotyped swimming escape response
provides a propulsive force that pushes water dorsally and ofTritonia, where the durationofthe swimandthedirection
movesthe animal in aventraldirection, much as the sculling of the first and last swimming flexion can be reasonably-
movements ofthe wings of the pteropod Clione cause it to predicted (Willows et al., 1973: Hume et al, 1982). the
move in the anterior direction (Satterlie el al., 1985). This swims of Melibe show little consistency in those parame-
sculling movement was originally described by Hurst ters. Fifty-eight percent of the animals tested began with a
(1968), but no further mention of it has appeared in the left flexion (n = 29), and 53% of the animals finished a
literature on Melibe. The combination of lateral bending of swimmingepisode with aflexionto left. The variation in the
the entire body and dorsal twisting of the foot typically duration of swim episodes was also quite large. The mean
propels the animal in the ventral direction. If the animal is swim duration in response to a salt stimulus was 174.9
=
oriented with its oral hood toward the surface, swimming 32.1 s (n 11: data were taken from the I I animals that
will propel it in a lateral direction. swam, out of the 49 tested with a salt stimulus); some
Although some aspects ofswimming are quite variable in animals swam foronly 33 s. but otherscontinued foras long
Melibe, the duration of each swim cycle, the magnitude of as 1546 s (25.7 mini.
rhythmic lateral flexions, and the instantaneous swimming One ofthe unique features ofswimming in Melibe is the
velocity are all very consistent during a swim episode. For motionless floating behavior which Agersborg (1921) re-
example, in a single swim episode lasting 58 s. the average ferredtoas "feigneddeath." During these floatingeventsthe
duration of each swim cycle was 2.03 0.03 s, with no animals lie in one place, dorsal aspect up. with the cerata
appreciable variation throughout the course ofthe swim. In inflated and spread parallel to the surface ofthe water; they
148 K. A. LAWRENCE AND W. H. WATSON
150- occurred significantly faster than the response to a salt
stimulus, but not significantly faster than the response to a
3 ioo- pinch (Fig. 3B; P < 0.05 Kruskal-Wallis nonparametric
oof'J ANOVA, Dunn's multiple comparison post-test). The three
Ji a! 50- stimuli also elicited swims with variable durations (Fig.
3B). Swims in response to contact with a sea star tube foot
<c c lasted an average of 53.7 9.4 s (H == 20) and were
-50- significantly shorter than swims elicited by salt (174.9
32.1 s, H = 1 1 , P < 0.01, Kruskal-Wallis nonparametric
-100- ANOVA. Dunn's multiple comparison post-test), but not
significantly different in duration from swims triggered by a
-150 2i5 5i0 75 100 125 150 pinch (91.0 76.9 s, n = 6, P > 0.05). There was no
Time (s) significantdifference betweentheduration ofswimselicited
>
by salt and swims triggered by a pinch with forceps (P
fleFxiigonusreof2.theSbwodiym.miThnegseinmoMtciloinhseairsecrheapreaactteedriwzietdhobuytvsiiggnoirfoiucsantlavtaerrial- 0.05). Finally, there was nocorrelation between the latency
Tathiiosngirnaptihmiisngapolrotamopfltihteudaengtlheroofugthhoeubtotdhyeodfuraatsiinognleoftetthheersewdiimndeipviisdoudael., tsoubrseesqpuoenndt tsowaimpaertpiicsuoldaer s(rt2imu=lu0s.0a5n)d.tHheowdeuvrearti,onanoifmatlhse
as viewed from the ventral side, during slightly more than 2 min of
swimming. Animals were induced to swimby dislodging them from their
attachmenttothesurfacetensionofthewater. Intheexampleshowninthis 80
figure, the animal was slightly flexed to the left when the stimulus was A.
applied, and also remained slightly flexed to the left when it stopped
swimming. 60 -
will occasionally remain in this position forseveral minutes. 40 -
During a separate set of long-term swimming experiments oo
(60-90 min), during which the animals were not allowed to
attach to any substrate, these pauses occurred every 10-20 20 -
min.
Stimuli that elicit swimming Pinch Salt Sea Star
To determine what external factor probably causes Me-
libe to swim in its natural habitat, and how to reliably
stimulate swimming in the laboratory, we screened a num-
ber of possible noxious stimuli, including pinches with
forceps, salt (KC1). and contact with several different puta-
tive predators. There was a significant effect of these treat-
ments on the tendency of Melibe to swim, with some
<
treatments being more effective than others (P 0.001,
G-test for independence). Of the three stimuli, the touch of
the predatory sea star Pycnopodia yielded the most reliable
response (Fig. 3A; 62% ofthe 32 animals that were touched
swam, P < 0.0001 Fisher's exact test, comparing sea star
responses to pooled KC1 and pinch responses). In fact, a Pinch Salt Sea star
very brief(<1 s) touch with an individual Pycnopodia tube
foot was usually sufficient to elicit a swim. This finding Stimulus
contrasts with an earlier report that "A/. Iconimi rarely swim Figure 3. The influence of various stimuli on swimming behavior in
following sea star contact" (Page, 1993). Single pinches to Mt'lihe. Animals were exposed to three different stimuli (pinch, n = 20;
a cerata, as well as trains of pinches, caused rapid escape salt, n = 49: tube foot of the sea star Pycnopodia. n = 32) and their
crawling but rarely swimming (5% swam, n = 20). A salt tendency toswim(panel A. %thatswam), latency toswim (Panel B. time
esloilcuittieodns(w1immlmionfg1 iMn 2K2C%1)oafppthleietdritaolsth(en s=kin49o)f.the head fe(fPrfaoenmcetlisvtBeilmtuhwlauensretthoemeioantsihtueiraretidso.tnimTouhflie.swsAienmailm.staalarsndsttiodmuuurclahutesidonwwaiostfhstsihegenaisfsiwtcaiarmnttlueybpeimsofoerdeeet
Ofthe three stimuli that were found to be most effective, responded very rapidly and consistently in comparison to those given the
contact with sea stars elicited a rapid escape response that other stimuli.
MELIBE SWIMMING BEHAVIOR 149
often swam multiple times in response to a single sea star
contact. If these multiple swims are viewed as one long 330 30
swim episode, then in general, "stronger" stimuli (sea
star>salt>pinch) caused animals to respond more quickly 300 60
andAlslwtihemclroanbgsearncdoamnpeamroendetsot"eswteeadk,ears"wesltlimaulsit.he sea stars -1SD(14)
other than Pvcnopodia. elicited no responses at all in Me-
libe. Neitheranimal seemedtotake any notice ofthe other's 270
When contact between crabs and nudibranchs
presence.
occurred, the nudibranch would often simply crawl over the + 1 SD (14)
carapace of the crab without incident. No contact between
the anemone and the nudibranch was ever observed. Some 240 120
nudibranchs were left withcrab and seastarpredators for up
to 48 h, with no signs of predation. Finally, in a number of 210 150
cases, nudibranchs were placed on the oral surface of po- 180
tential sea star predators, and no ingestion occurred. How-
epvreerd,atowres,diadndnootncoontthreorlofcocrasthieonsstawteeohfahvuengoebrseorfvetdhebtoetsht MweelrFieibgoeubrsweeirt4vh.edaAfporrceodtihmceptaderudirasptoaintohnof(opfteharenpesinwnddiiucmcumleiadnrsgwtoidmitr,heecatnfidooontt)ho.efaSanegvslepeneocfianmtieramnvaelolsf,
P\cnopodin and anemones eating small specimens of Me- relativetothelong-axisofthefoot,wasmeasuredat5-sintervals.Anangle
libe in the laboratory. In addition, Ajeska and Nybakken of90degreesrepresentsmovementinadirectionperpendiculartothelong
n(i1a97k6e)lphabveedsr,episoratepdretdhaattoPrugoefttMieal,ibae.crab found in Califor- aaxniismaolf.tThheefoaovte.raTgheeaanvgeleratgreavaenlgeldebsyw=feirveeocfalsceuvleanteadniamnadlsplfoetltlewditfhorineoanceh
standard deviation ofthe prediction ( 14 degrees).
Direction ofswim
stimuli, such as contact with the tube feet of the predatory
Preliminary observations indicatedthat, when swimming, sea star Pycnopodia. It is also very stereotyped in terms of
Melibe moved in a ventral direction, perpendicular to the the consistent rhythmic flexions used to propel the animals
long axis of the foot. To test this hypothesis, we analyzed through the water.
the instantaneous swimming direction of seven animals, in Melibe is rarely observed swimming in its natural eel-
5-s intervals, as described in the Materials and Methods. grass and kelp habitats, even though potential predators
Five ofthese animals moved, on average, in a direction that such as Pycnopodia. anemones, and crabs are present
varied less than one standard deviation (14) from the (Ajeska and Nybakken. 1976; unpubl. obs.). One explana-
predicted direction (90 from the long axis ofthe foot). The tion for this apparent low frequency ofswimming might be
variance angles ofthe other two animals were only slightly that Melibe rarely encounters predators. This nudibranch
different than predicted (Fig. 4). These data support the tends to situate itselfnearthe distal portions ofeelgrass and
hypothesis thatthe generaldirectionofmovement, from one kelpblades, where large seastars andcrabs arerarelyfound.
swimming flexion ofMelibe to the next, can be predicted if Melibe may also have less of a tendency to swim where
the orientation ofthe foot is known. This prediction is most currents are strong, because swimming animals have a high
accurate afterthe firsttwoswimming flexions, which tendto probability ofbeing carried away from their preferred hab-
propel the animal upward. Subsequently, most movement itat, and potential mates, by these currents. The high density
generated by an individual flexion is in a plane that is of swimming and floating individuals observed intermit-
perpendicular to the long axis of the foot. Therefore, if an tently at considerable distances from local eelgrass beds
animal is positioned vertically in the water column, with its suggests that other factors, besides predators, may also
oorfaleehlgoroadssto(pwearrsd. otbhse.)s.ursfwaciem,maisngit wisooufltdenmofsotunldikoenlybmlaodvees tsruiggggeesrtesdwithmamtitnhgeseinanMiemlailbsem(iMgihltlsu,se19v9o4l)u.ntFaarrymsewri(m1m9i7n0g)
it in a lateral direction. episodes to move from one kelp blade to another, and we
have observed spontaneous bouts of swimming on many
Discussion occasions in the laboratory. Voluntary swimming may also
In this study we examined the swimming behavior of be a means of dispersal, which would allow mixing of the
Melibe leonina, from initiation, elicited by a variety of gene pools from spatially isolated populations inhabiting
stimuli, totermination, markedby reattachment to a suitable eelgrass beds located several kilometers from each other
substrate. As in other lateral-bending swimmers (Agers- (Mills, 1994).
borg, 1922), and many other swimming molluscs (Thomp- A unique feature of swimming in Melibe. in comparison
son, 1976), the behavior is elicited most reliably by noxious with other swimming molluscs, is motionless floating be-
150 K. A. LAWRENCE AND W. H. WATSON
havior. These pauses may represent an energy-saving strat- ria must be fulfilled for a behavior to be useful in neuro-
egy that allows the animals to"rest" and remain in the water ethological studies: reliability, robustness, and stereotypy.
column for a time, at a relatively low energy cost. An All of these criteria are characteristic of the swimming
alternative hypothesis is that floating may enable this nudi- behaviorofMelibe leonina. It canbe reliably initiated in the
branch toperiodically open its oral hood to sample the water laboratory with natural stimuli or a salt solution. The ro-
column for prey. It cannot feed and swim at the same time, bustness and stereotypy are illustrated in Figure 2, which
so this sampling activity would allow it to "forage" while shows that over the time course of a swim, the flexion
floating. When it encountered a high density of food, it amplitude and frequency do not change significantly. Fur-
could stop, seek a suitable substrate, and feed (Trimarchi thermore. Melibe is amenable to electrophysiological inves-
and Watson. 1992; Watson and Chester. 1993). Preliminary tigations, as are many other opisthobranch species, because
studies in our laboratory indicate that prey (Anemia) reduce it has large, identifiable neurons, and impulses from these
both the rate ofcrawling and the frequency ofspontaneous neurons can be recorded both in swimming, semi-intact
swimming episodes in Melibe. animals and in isolated brains (see companion paper,
The most effective stimulus for eliciting swimming in Watson et til.. 2002). Finally, in Melibe, relatively few
Melibe is contact with the tube feet of Pycnopodia. This higherorderinterneuronsconstitute the swimcentral pattern
stimulus is probably effective due to the surfactants found generator, so a very thorough neuroethological understand-
on the tube feet ofcertain predatory seastars (Mauzeyetal., ing of the behavior is possible (Watson et al.. 2001).
1968: Mackie. 1970). Oddly enough, we rarely observe sea
stars attacking an adult Melibe. Page (1993) suggests that Acknowledgments
sea stars avoid Melibe because its repugnatorial glands,
located throughout the epidermis, release a chemical that We thank A. O. Dennis Willows and Glen Brown for
renders it repulsive to predators. These glandsdo not mature their advice during the experimental phases of this study
until the animals are 4-7 weeks old, and sea stars do attack and for providing critical comments on this paper, Eric
and consume younger individuals. It is interesting that even Abrahamson and Jen Wishinski for their help with several
though their repugnatorial glands help deter potential pred- behavioral experiments, Jim Newcomb and Stuart Thomp-
ators, mature specimens retain their tendency to escape son for constructive comments on the manuscript, and the
when they sense the presence of certain sea stars. entire staff of FHL for their continuous support and assis-
The direction that Melibe travels during a swim appears tance with all phases ofthe work. This study was supported
to be random, upon casual observation. However, certain by Centerfor Marine Biology grants and SummerTeaching
features of the path taken during a swim episode are fairly fellowships to K.A.L. through the University of New
predictable. When this nudibranch starts to swim, it first Hampshire and an NIH grant to W.H.W. It is contribution
releases the anteriorpartofits footfrom the substrate. Then, number 386 of the Center for Marine Biology/Jackson Es-
no matter what the initial orientation of the animal is. its tuarine Laboratory series.
head moves toward the surface and the first few lateral
flexions tend to move its body in the anterior direction. Literature Cited
Once an individual has "pushed off" and is in the water
Agersborg, H. P. v. W. K. 1921. Contributionstotheknowledgeofthe
column, the combination of lateral flexions and twisting of nudibranchiate mullusk. Melibe leonina (Gould). Am. Nat. 55: 222-
the posterior portions of the foot and tail region creates a 253.
"sculling" motion that reliably propels it along a plane Agersborg, H. P. v. W. K. 1922. Notes on the locomotion of the
perpendicular to the long axis of the foot (Fig. 4). Thus, nudibranchiate mollusk. Dendronotus giganteux O'Donoghue. Biol.
although its swimming behavior has less of a directional Bull. 42: 257-266.
component than seen in some molluscs that use parapodial AjesMkeal,ibRe. lA.e,onainnad .(1.GoNuylhda,kk1eX5n2.)1(9M7o6l.lusCcoan:trOipbiusttihoonbsratnocthhiea)b.iolVoeglyigoefr
flapping, such as Aplysia brusiliunii and Clione, Melibe 19: 19-26.
appears to have more control than the animals that use Arshavsky,Yu. I., I. N. Beloozerova,G. N.Orlovsky,Vu. V. Panchin,
dorsal-ventral flexions, like Tritonia and Pleurobranchaea. and G. A. Pavlova. 1985. Control oflocomotion in marine mollusc
This raises the question ofwhether Melibe has the ability to Clione linuicinu. 11. Rhythmic neurons of pedal ganglia. Exp. Brain
Res. 58: 263-272.
seekout its preferred habitats orpotential mates, or whether Audesirk,T.,and G. Audesirk. 1985. Behaviorofgastropod molluscs.
it attempts to move laterally from one eelgrass orkelp blade Pp. 1-94 in Neurohiiilogy amiBehavior. Vol. 8. A.O.D. Willows, ed.
to another. Certainly, in the laboratory, it swims spontane- Academic Press, New York.
ously, especially during the night (Watson and Newcomb, Bickell-Page, L. R. 1991. Repugnatorial glands with associated striated
unpubl. obs.). and our working hypothesis is that it uses muscle and sensorycells in Melibe leonina (Mollusca: Nudibranchia).
swimming both as a response to predators and as ameans of EdmZuinHdismo,rpMh.olo1g9y681.10:O2n81-t2he91.swimming and defensive response of
intermittent locomotion. Hexabranclnts inuri>iimtux (Mollusca: Nudibranchia). J. Linn. Soc.
According to Audesirk and Audesirk ( 1985), three crite- Luml. 47: 425-429.
MELIBE SWIMMING BEHAVIOR 151
Farmer, \V. M. 1470. Swimming gastropods (Opisthobranchia aiul Page, I.. R. 1993. IX'\elopment of behavior in juveniles of Melibe
Prosobranchia). IV/i'vc'1 13: 73-89. lamina (Gastropoda: Nudibranchia). Mar. Beluiv. Physiol. 22: 141-
Getting, I'. A. 1983. Neural control ofswimming in Tritonia. Pp Sll 161.
128 in Neural Origin ofRhythmic Movements, A. Roberts and B. L. Satterlie, R. A. 1985. Reciprocal inhibition and postinhibitory rebound
Roberts, eds. Cambridge University Press. New York. produce reverberation in a locomotor pattern generator. Science 229:
Gelling. P. A. IMS'*. Emerging principles governing the operation of 402-404.
neural networks. Annu. Rev. Neurosci. 12: 185-204. Satterlie, R. A. 1991. Neural control of speed changes in an opistho-
Hamilton.P.V.,andH.W.Ambrose.1975. Swimmingandorientation branch locomotory system. Bid. Bull. 180: 228-233.
mAplysiabrasiliana (Mollusca: Gastropoda). Mar. Behav. Physiol. 3: Satterlie, R. A., and T. P. Norekian. 1996. Modulation of swimming
131-144 speedinthepteropodmollusc. Clionelimacina: roleofacompartmen-
Hume. R. I.. P. A. Getting, and M. A. Del Beccaro. 1982. Motororga- tal serotonergic system, tmcrtebr. Neurosci. 2: 157-165.
nization of Tritimia swimming. I. Quantitatne analysis of swim be- Satterlie, R. A., and A. N. Spencer. 1985. Swimming in the pteropod
haviorand flexion neuron firing patterns. J. Neurophvsiol. 47: 60-74. mollusc. Clione limacina. II. Physiology. J. E.\p. Bid. 116: 205-2!
Hurst. A. 1968. The feeding mechanism and behaviour ofthe opistho- Satterlie, R. A., M. LaBarbera, and A. N. Spencer. 1985. Swimming
braneh Melibe lamina. Symp. Zoo/. Soc. Land. 22: 151-166. inthepteropod mollusc. Clionelimacina. I. Behaviorandmorphnlo.js
Jin. J.. and R. Gillette. 1995. Neuronal elements that mediate escape J. Exp. Bid. 116: IS1)-204.
swminima and suppress feeding behavior in the predator\ sea slug Thompson,T.E. 1976. BiologyofOpisthohranchMolluscs. Vol. 1.The
Pleinvbramhaea. J. Neuroplt\.siol. 74: 1900-1910. Ray Society. London. 206 pp.
Jing, J.. and R. Gillette. 1999. Central pattern generator for escape Trimarchi, J., and W. H. Watson. 1992. The role of Melibe buccal
swimming in the notaspid sea slug Pleurobranchaea californica. ganglia in feeding behavior. Mar. Beluiv. Physiol. 19: 195-209.
J. Neurophvsiol. 81: 654-667. von der Porten, K., D. W. Parsons, B. S. Rothman, and H. Pinsker.
Mackie, A. M. 1970. Avoidance reactions of marine invertebrates to 1982. Swimming in Aplv.iia hrasiliana: analysis of behavior and
eithersteroidgl>cosidesofstarfishorsyntheticsurface-activeagents.J. neuronal pathways. Behav. Neural. Biol. 36: 1-23.
MauEizoxerpy.o.fMKaa.srt.eP.r.BoiiCdd.s.BaEincrdoklee.lsac5n:adp6.e3ar-en6s9dp.oPn.seKs.oDfayttheoinr.pr1e9y68i.n thFeeePduignegtbSeohuanvd- Watsa1on8dn52,t)aWc(.tOilpHei.sstIthliolm,burlaainndcohnCi.at:hMe.DefCnehdeedrisontngeortb.aech1e9aa9v)3i..orVeolTfihgeMeeriln3if6bl:eue3nl1ce1eo-noi3fn1ao6l.f(aGcotuolrdy
McPsRhewegirimsoomnn.i,nEgcDo.ilnoRAg.p,yly4as9ni:da6bJ0r.3a-sEi6.l1i9Ba.lnaa.nkIe.nIsnhniepr.vat1i9o9n1ao.fpaNreauproadliaclonmtursoclleosf WatsNeounr,oeWt.holHo.gyIIoIf.MKe.libDe.lLeaonwirneancsew,imamnidngJ.behMa.vioNre.wAcmo.mbZ.ool2.00411.:
1026-1035.
McPbshyweirpsmeomdnail,nggDa.inngRlA.ip,olyna.snmidoatoJb.rrunsEei.ulriBoalnnasa..nkJIe.In.sNOhernigrpao.npih1zy9as9ti1ioobln..o6f6N:epue1rdaa3ll38mc-oo1nt3to5rr1on.leuo-f Watcsoornr,elaWt.esHo.fIsllw,iJm.mMi.ngNebewhcaovimobr.iannMdelSi.bTehloemopnisnoan..Bi2o0l0.2.Bull.Neu2r0a3l:
Millrso,nsC.anE.d p1a9r9a4p.odiSaelasmoontaolrsfwieilmdsm.iJn.gNeoufrsoepxhuyasliloyl.ma6t6u:re13b5e2n-th1i3c65o.pis- Will1o5w2s-,16A0..O. D. 1965. Giant nerve cells in the ganglia ofnudibranch
thobranchmolluscs(MelibeleoninaandGasteropteronpacificum)may molluscs. Comp. Bioehem. Physiol, 14: 707-710.
aveulgompemnetntpoopfulMaatriionnedIinsvpeerrstaelb.raPtpe.s.3W1.3-H3.19WiilnsoRne.prSo.duAc.tiStornieakenrd,Daen-d Willboaswiss,oAf.beOh.avDi.o,rDionrsTeritito,niDa.. AII.I,. NaenudroGn.alHomyelceh.an1i97s3m.ofTahfeixendeuarcotniaoln
G L. Shinn, eds. The Johns Hopkins University Press. Baltimore. pattern. J. Neurobiol. 4: 255-285.