Small Size, Brooding, and Protandry in the Apodid Sea Cucumber Leptosynapta clarki

12 Pages·1994·5.2 MB·English

by  M A Sewell

Reference: Bioi Bull. 187: 112-123. (August, 1994)
Small Size, Brooding, and Protandry in the Apodid
Sea Cucumber Leptosynapta clarki
MARY SEWELL
A.
Bamfield MarineStation, Bamfield. British Columbia, Canada \'OR 1BO, and
Department ofZoology. University ofAlberta, Edmonton, Alberta, Canada T6G 2E9
Abstract. The apodid seacucumberLeptosynapta clarki ( 1974) suggested that the trend for lecithotrophy, brood-
exhibits the three commonly associated traits of small ing, orviviparity in small animals resulted from thelower
adult size (max. length 1 13 mm), brooding (intraovarian energetic reserves forgamete production. Strathmann and
= viviparity), and hermaphroditism (protandric). Juvenile Strathmann (1982) described three categories of hy-
L. clarki are released from the ovary at a length of 1- potheseson theassociation ofsmall adult sizeand brood-
mm
2 in the early spring (April-May) and are reproduc- ing related to (1) dispersal, (2) adult longevity and re-
tivelyactiveas males in the reproductive season (Novem- cruitment, and (3) the allometry ofstructures associated
ber) following their birth. In their second year, some in- with gamete production and brood care. These hypotheses
dividuals continue to reproduce as males, but others un- are not mutually exclusive and no single one is entirely
dergo protandric sex change to reproduce as females. satisfactory (Strathmann and Strathmann, 1982). How-
Analysisofthe relationship between sizeand sex revealed ever, ofthese four hypotheses, the allometry hypothesis
a "critical" size forsex changeata weight of200-400 mg has received the most attention, primarily because it can
with a 1:1 sex ratio above 500 mg total weight. Transi- be tested within a single species.
tional gonads with previtellogenic oocytes and mature The allometry hypothesis predicts that brooding will
spermatozoawere observed, suggestingthat sex change is not occur in large species, because ofspatial limitations
initiated priorto reproducing in the current reproductive on brood size with increasing animal size (Heath, 1977;
season. A test of the allometric hypothesis on the asso- Strathmann and Strathmann, 1982). In numerical terms,
ciation between small size and brooding found no evi- fecundity is likely to increase with the cube oflength, but
dence for scaling constraints on brood size in L. clarki. the space for brooding may increase with a lower scaling
These allometric constraints may be avoided because of constant, resulting in insufficient space for brooding the
potentially low fertilization success and brooding within embryos. Although originally proposed for an external
a distensible structure. The sequential hermaphroditism brooder (Strathmann and Strathmann, 1982), the allo-
in L. clarki may additionally be a method to reduce in- metry hypothesis has been tested by comparing the rela-
breeding in a species with limited dispersal. tionship between animal size, egg number, and number
ofbrooded embryos in several marine invertebrates(Daly,
Introduction 1972; Rutherford, 1973; Menge, 1974; Ockelmann and
Muus, 1978;Rumrill, 1982; Strathmann eta/.. 1984; Ka-
Two characters ofthe adult have been associated with bat, 1985; Gremare and Olive, 1986; McGrath and
brooding in marine invertebrates: small size and the ten- OFoighil. 1986; Russell and Huelsenbeck, 1989; Byrne,
dency towards hermaphroditism. Several hypotheses have 1991; Brey and Hain, 1992; Hines, 1992; Hess, 1993).
been proposed to explain the association of small adult Hermaphroditism, in which both male and female ga-
size with brooding. The energetic hypothesis of Chia metes are produced by a single individual in its lifetime,
iscommon in marine invertebratesand isoften associated
Received 4January 1994;accepted 25 May 1994. with brooding(Ghiselin, 1969, 1974; but see Heller, 1993,
Presentaddress: Dept. ofBiologicalSciences,Simon FraserUniversity, for arguments against this association in gastropods).
Burnabv, B.C., Canada V5A 1S6. Ghiselin (1969, 1974) proposed three models fortheevo-
112
PROTANDRY AND BROODING IN L CLARK] 113
Table I
Patternsofsexchangeanilreproductivecharacters in thephylum Echinodermata
Species Sex change Brood1 Evidence forsexchange2 Reference
ClassAsteroidea
14 M. A. SEWELL
sequencesofthatsexchange,andexaminesthehypotheses 1994). Juvenileseacucumbersorthosewithsmall gonads
fortheassociation between small adult size and brooding were dissected to verify sex identification.
in marine invertebrates. The relationship between sea cucumber size and sex
was determined by calculating the percentage offemales
Materials and Methods ineachweightorlengthclass. Forweightslessthan 800 mg
the data were divided into 50-mg size classes. Few sea
Collections andsize measurements cucumbers were found above 800 mg, so the size class
SpecimensofLeptosynaptaclarkiwerecollectedattwo was increased to 100 mg. The final category was sea cu-
sites in Grappler Inlet, Bamfield. British Columbia (Site cumbers with weights > 1300 mg. Sea cucumbers in this
1:4849'57"N, 12506'45"W;Site2: 4850'N, 12506'36" category ranged from 1412 to 1684 mg (TV = 4), so are
W). Site 1 is located on the northern side of Barge Bay plottedasasinglepointat 1500 mg. Seacucumberlength
and has a short intertidal area (ca. 20 m) and a high pro- was divided into 5-mm size classes until 100mm; the
portion ofgravel in the substrate. Site 2 isagentlersloping three sea cucumbers with greater length (104, 108,
mudflat (intertidal area ca. 60 m) at the opposite end of 1 13 mm) are plotted as a single point at 1 10 mm. The
the channel separating "No-name" Island from Vancou- total number of sea cucumbers sampled was 693 (454
verIsland proper(see Sewell and Chia, 1994, Fig. 1). This males, 239 females).
site has a finer sediment composition with little gravel The breeding sex ratio was determined from haphazard
and contains L. clarki ofa larger size than Site 1 (mean cores used in following the recruitment of L. clarki on
total weight SD: Site 1: 307.8 153.99 mg, N = 720: the intertidal mudflatsat Site 1 (Sewell, 1993a). Ten hap-
Site 2: 490.4 268.80 mg, N = 660; Sewell and Chia, hazardly thrown cores were taken on the mid-intertidal
1994). Both siteshaveextensive Zosterabedsinthelower in Barge Bay (Sewell, 1993a), the animals removed by
intertidal that extend into the shallow subtidal. sieving, and the size and sex ofeach sea cucumber deter-
The seacucumbers werecollected on the mid-intertidal mined in four samples during the winter of 1990-1991
mudflats, either by using cores or by sweeping away the and the summer of 1991.
upper 5 cm of sediment to reveal the animals in their
burrows; put in buckets with the surrounding mud; and
transported tothe laboratory. Adultswere removed from Gonadal changes
the sediment with a 2- or0.85-mm sieve. The remaining Sea cucumbers sampled during regular gonad index
fraction was then sieved through a 0.25-mm sieve to re- dissectionsinJanuaryand February 1990and 1991 (Sew-
movejuveniles. Seacucumberswere kept in seawateruntil ell and Chia, 1994) were used to determine the size of
length and weight were determined. animals with transitional gonads (that is, animals in the
Animals to be measured were placed in a 90-mm petri processofchangingfrom maleto female). Seacucumbers
dwiasshmceoanstuariendi,ngin2.m5i%lliMmgeCtelr:s,infrsoemawtahteeran(twe/rvi)o.rLeenndgtohf wweerigehceodllaenctdeddifsrsoecmteSditaess d1eascnrdib2ed(TiontaSlewNell= a2n4d0)Cahnida
the calcamremous ring to the posterior. Sea cucumbers less (1994). The gonad in L. clarki consists oftwo separate
than 20 were measured on a dissecting microscope
tubules thatjoin through the dorsal mesentery to a single
equipped with an eyepiece graticule calibrated against a gonoduct (Sewell and Chia, 1994). Both gonad tubules
stage micrometer. Largerseacucumbers weregently held weredissected, placed on aglassslide (38 X 75 mm) with
straight with tweezers and length was measured to the a small amount of seawater, and temporarily mounted
nearest millimeter with reference to a small plastic ruler with a coverslip (35 X 50 X 0.02 mm). The weight ofthe
wsueanatdeecrruactnuhdmebtpeherterniwawdseiisphgl.haecAdefdtoenornamaefaitnsiesu-srsuceeamlteeonbrtaelmoafoncvleeentegotxhtd,eertnteahrle- ceonvceersolfipoorceyvteeasl,edwidtethaoiults otfhethneeegodnafodr,hsiusctholoagsy.theOfprtehse-
240 sea cucumbers dissected, 87 were females brooding
mine its total weight (TW) in milligrams. pentactulae or undergoing resorption. Ofthe remaining
153 seacucumbers, 75 had transitional gonads; these were
Size-sex relationship
divided into 50-mg size classes for analysis.
The relationship between size and sex in L. clarki was Histological descriptionsofprotandricsexchangewere
determined in July 1991 when the gonads were well de- obtained during a detailed study of reproduction in L.
veloped (Sewell and Chia. 1994). After a specimen was clarki(Sewell and Chia, 1994). Gonadswere preserved in
weighed, its sex was determined by examining the color Bourn's fixative, stored in 70% ethanol, and embedded in
ofthe gonads through the semitransparent body wall. In Paraplast. Sections were cut at 7 ^m, stained in hema-
July, females had yellow ovaries with large eggs (ca. toxylin and eosin, and photographed on a Zeiss photo-
200 nm) and males had white testes (Sewell and Chia, microscope.
PROTANDRY AND BROODING IN L. CLARKI 115
AHornetry ofreproduction Results
The relationshipbetween thesizeofthefemaleandthe Si:e-se.\ relationship
number of unfertilized eggs was determined in late Oc- Leptosynapta clarki is a protandric hermaphrodite
tober 1989 and 1990 priorto the spawning period in No- in which sex change occurs over a broad range ofsize
vember and December (Everingham, 1961; McEuen, Sea cucumbers in the smallest weight class (0-
(Fig. 1).
1986; Sewell and Chia, 1994). Sea cucumbers were col- 50 mg) arejuveniles ofthe year(Sewell, 1993a) and are
alsecdteesdcrfirboemd aSbitoevse.1Aafntedr2totaanldwesiigehvted(TfrWo)mwtahsemseeadsiumreendt, a7l7l,m1a2l0e,(aFnigd. 115A0).mTgh(e21s,ma3l6l,es3t0fmemmaleisn fleonugntdh,wereisgpheecd-
a longitudinal incision was made from the base of the
tively; Fig. 1).
calcareous ring to the posterior, the coelomic water was Sexchangedid not occurin adistinctweightclass(Fig.
drained from the body cavity, and the animal was re- 1A). Few females were found until 200-250 mg total
weighed to obtain itsdrainedweight (DW) in milligrams. weight (Fig. 1A). Between 200 and 400 mg there was an
Thegonadwasdissectedandtemporarily mounted under increasing percentage offemales in each size class. This
atmwactoouvsreiretselseigipgn.seCpaoecruhngytoesnaarwde(rT(oeNtaml=aNd2et=uob8fu0ltehfsee)matilonteas2l)0.nfuemmableers oatf p5e0r0Acemnsgitmai(gFleiagr.sti1anAbci)rl.eiazseed iant paebrocuetnt5f0e%m,alaets wweaisghstesenabwoivteh
The relationship betweenthesizeofthe femaleandthe increased seacucumberlength(Fig. 1B). Nofemaleswere
number of fertilized eggs/embryos was examined in sea found until the 20-25 mm size class. The percentage of
cucumbers in late November and early December 1989 females increased between 25 and 45 mm, plateauing at
and mid-November 1990. This wasduring the spawning 50%, above a length ofabout 50 mm (Fig. IB).
period of L. clarki at Bamfield (Sewell and Chia, 1994)
but before degeneration ofunfertilized eggs orsignificant 100 A
pentactulae mortality (Sewell, 1993a). Countsweremade -,
ofthe number of unfertilized eggs, early stage embryos,
or pentactulae in each gonad for 50 females. The allo-
metric relationship between female size and number of
pentactulae was determined for those females that had
=
early embryos or pentactulae present (N 28).
The allometric relationship between size and number
ofeggs or pentactulae was tested using a Model II re-
duced major axis (RMA) regression. Ordinary least
squares(OLS) regressions were calculated using Micro-
soft Excel Version 3.0. Because oftheconsiderable error
rinegmreeasssiuorninwigllsiuznedienrseesaticmuactuemtbheersva(lSueeweolfl,the19s9l0o)p,eO(LseSe I *~~2100 4010 6010 80r0 1000 1200 1400 1600
McArdle, 1988; LaBarbera, 1989), and may suggest that Total Weight (mg)
scaling constraints exist when there are none (Hess, 100-1 B
1993). RMA regressions were performed on raw and
natural log (loge) transformed data. Standard error (s2) 80-
and asymmetric confidence intervals were attached to
the slope ofthe RMA regression using the formulas in
McArdle (1988). E
Ifthere are no constraints on brood size, the number LoL>
ofembryos brooded isassumed to scale with body size in 3? 40
the same way as the fecundity ofbroadcast spawning an-
imals; i.e.. linearly with bodyweight orthecube oflength 20
(Strathmann and Strathmann, 1982). To test whether
brood sizewas limited by scalingconstraints, the slope of -r
the RMA line was thus compared to the slope predicted 20 40 60 80 100 120
by isometry (^ = 1 forthe relationship between log.body Length (mm)
wtetiesgtht(Calnardkelo,g,,19n8u0m;bMecrArodflpee,nta1c9t88u)l.aeA) usslionpgealeSstsutdheannt's1 GraFpipgluererI1n.let,LeBpatmafsivelnda.pliandsaizrekic.laPsseersceonft(Af)emtaoltaelsweeaigchutcu(mmbge)rasndfr(oBm)
in this relationship would indicate scaling constraints on length(mm). Point isplottedat upperlimitofsizeclassexcept forfinal
brood size in larger sea cucumbers. category shown with (x);seedetailsin text.
16 M. A. SEWELL
The presence of large numbers of small, but sexually
active, males in the population strongly biases the sex
ratio in L. clarki (Table II). In all samples the sex ratio
was above 70% male, with a maximum of88.5% male in
August (Table II).
Gonadal changes
MostseacucumberswithtransitionalgonadsinJanuary
or February were in the 200-400 mg size range (Fig. 2).
Sex change was, however, observed in some very large
specimens (Fig. 2). The largest sea cucumber found with
a transitional gonad weighed 855 mg (TW).
Afterspawning, thetestesbecame thin and yellow, and
shrank in length towardsthegonad basis(Sewelland Chia,
1994). In protandric males, new oocytes were observed
in fresh gonad tubules at the basal end adjacent to the
gonadbasis(Fig. 3A). Unspawned sperm wereoften pres-
ent in thetubuleswith the newoocytes(Fig. 3B)andwere
active when exposed to seawater.
Histological examination ofmalegonads revealed that
oocyteswere present alongthetubule wall in some mature
testes(Fig. 3C). The testis lumen waspacked with mature
spermatozoa, whereas previtellogenic oocytes 10to 20 /jm
in diameter were found along the tubule wall (Fig. 3C).
Afterthe spermatozoa were spawned, the oocytescontin-
uedtogrowwhilethe remnant spermatozoawere resorbed
(Fig. 3D). Sex change must, therefore, be initiated prior
to spawning in the current reproductive season.
Allometry ofreproduction
Usingrawand logc-transformed variables, reduced ma-
joraxis(RMA)regressionswerecalculated between female
size and number ofeggs orpentactulae (Table III). In the
period before spawning, female sea cucumbers show a
positive significant relationship between weight (TW or
DW)and egg number(Fig. 4, Table III). Regressions using
natural logtransformations ofthese variables are alsosig-
nificant, although there is a slight reduction ofthe r in
both cases (Table III). The allometric exponent of the
loge-transformed egg number/drained weight relationship
Table II
.W.Yratio<>/ Leptosynapta clarki tilSite I. (irupi'/erIntel
Date
PROTANDRY AND BROODING IN L Cl.lRkl 117
Figure 3. Leplosynapta c/arki. Photomicrographs oftransitional gonads. (A) Small resorbing testis in
processofsexchange.Almosttheentiregonadisshown. Notetheswollengonadbasis(gb)with madreporite
(m), andsmall oocytes(o)inthetubules. Bar = 500^m. (B)Transitional testis. Tubulecontainsnumerous
oocytes and remnant packets ofsperm (arrow). Some sperm have broken from tubule onto base ofslide.
Bar= 200^m.(C)Histologicalsectionofmaturetestiswithspermatozoa(sz)inthelumenandearlyoocytes
forming along the tubule wall. Section stained with hematoxylin/eosin. Bar = 25 ^m. (D) Histological
sectionofpostspawnedtestisin transitionalstage. Someremnantspermatozoaareseen inthelumen.Oocytes
arepresentalongthe tubule wall. Bar = 50nm.
Discussion ever, sex change in H. atra is difficult to demonstrate be-
cause the species also reproduces asexually by binary fis-
Protandry in Leptosynapta clarki sion and because the gonads regress aftereach spawning,
The histological and population information pre- making determination of the previous sex impossible
sented here confirms that Leptosynapta clarki is a pro- (Harriott, 1982).
tandric hermaphrodite and provides the first direct ev- Sex allocation theory predicts that sex change should
idence of sex change within the class Holothuroidea. bean "all-or-nothing" response, with the animal orplant
Two other holothurians in the order Apodida have been producing gametes ofone sex until it is more profitable
suggested to be protandric, though the evidence in both to produce gametes ofthe other sex, and then switching
cases is somewhat circumstantial. Leptosynapta in- entirely tothe latter(Charnov, 1982). Research on marine
haerens was believed to be protandric because only 4 invertebrates has, however, shown a number of species
of80 spawning specimens shed eggs (Runnstrom, 1927; that do not display an all-or-nothing response but have a
cited in Hyman, 1955), and the lack ofmales in a sam- 1:1 sex ratio above a certain size (Patella argenvillei
pled population of Labidopla\ media led Gotto and Branch, 1981; Luttia giganteaWrighl, 1989; Patella
Gotto (1972) to suggest that only very small and young kermadecensis Creese et al.. 1990; present study), or a
specimens produce sperm. wide size range or time over which sex change occurs
Thereare fewreportsofsexchange inthefiveremaining (Fromia ghardaqana Achituv and Delavault, 1972;
holothurian orders. Despite some anecdotal reports of Crepidula fornicata Hoagland, 1978; Ophionereis oli-
protandry among dendrochirote holothurians (Smiley et vacea Byrne, 1991; Coralliophila violacea Soong and
a!.. 1991), protandry has not been clearly demonstrated Chen, 1991; Puiiriellci c.\ixua Byrne, 1992). Thedebate
in any species. In aspidochirote holothurians, protandry continues as to whether this is the result ofa genetically
is suggested in Holothuria atra on the basis ofa change programmed sex change, an environmental sex deter-
in thesex ratiowith increased size (Harriott, 1982). How- mination, or both.
118 M. A. SEWELL
One explanation applied in a number of marine in- 5000i
vertebrates isthat there are two sexual typeswithin a spe-
cies. Bacci (1951), who proposed this terminology in As- 4000-
terina gibbosa, described one "race," or type, whose in-
dividuals (known asbalanced hermaphrodites) change sex
ataspecifictimeintheirexistence. Theothertypecontains 3000-
twocategories: unbalanced hermaphrodites, which change
sex earlier or later, depending upon the individual, and "
true males and females, which remain one sex all their 2000-
lives(Bacci, 1951). Inthelattertype,thepresenceof"true" I
males or females suggests a genetic predetermination for 1000-
thatsex. Evidence forpredetermination isseen in Bonellia
viridis: 10% ofthe larvae ofthisechiuran develop as males
even ifthe "male determining factors" ofa female pro-
100 200 300 400 500 600 700 800 900
boscis are absent (Leutert, 1975). Similarly, in Patella
vulgata. some individuals neverchangesex(Orton, 1928: Drained Weight(mg)
Ballantine, 1961), and in Lottia gigantea, about 15% of Figure-4. Leplosynaptaclarki. Relationshipbetweendrainedweight
the population change sex annually, regardless of envi- (RmMg)Aand numberofeggs(filled squares)orpentactulae(open squares).
regressions:
ronmental setting or age (Lindberg and Wright. 1985). Eggs: v = 6.348.V - 237.726, r = 0.829, N= 80.
InL. clarki, all individualsstarttheirreproductive lives Pentactulae: r = 1.795.Y + 44.960. r2 = 0.375, N = 28.
as males (Sewell. 1991; Sewell and Chia, 1994; Sewell,
1993a); thusany "true sex" individuals(thosethatdo not
change sex during theirlifetimes) in the population would The size-sex relationship presented here isa static data
be male. The rest ofthe population must be divided into set that combines sea cucumbers ofmany ages within a
thosethatchangesexatsomespecifictime,andthosethat size class (Fig. 1). Juveniles released from the mother in
change sex under the influence ofsome genetic or envi- May 1990 were maintained in the laboratory for 1 year
ronmental factor. Male L. clarki may either delay sex and used for measurements of growth (Sewell, 1993a).
change, as shown by the large size of some transitional These individuals reproduced as malesin November 1990
males (Fig. 2), or have a labile sexuality that alternates (Sewell, 1993a). When the growth experiment finished in
between male and female. There is some circumstantial May 1991, the surviving sea cucumbers were kept in the
evidence that sex can be reversed from female to male laboratory until thegonads had regrown afterpostspawn-
(Sewell and Chia. 1994), and it is conceivable that sex ing resorption (Sewell and Chia, 1994). In July 1991, eight
may change more than once in the lifetime ofa holothu- animalsofthesurviving78 hadchangedsex(unpub. data).
rian (Everingham, 1961; Harriott, 1982). These protandric seacucumberswereabove 185 mgTW,
Table III
Leptosynaptaclarki: Scalingconstantsforrelationshipsbetweenseacucumbersizeandnumberofeggsorpentactulae
Regression' F2 ln(a) s2? filower upper
TW
ivv#eggs
PROTANDRV AND BROODING IN L CI.ARK1 119
high number ofpentactulae do not show a higher degree
of mortality (Sewell, I993a). These results suggest that
5- brood size is not constrained in larger L. clarki by either
brood space or differential embryonic mortality.
4- Although the allometry hypothesis is intuitively ap-
pealing(Hess, 1993), theevidenceforscalingconstraints
.cOu on brood size in marine invertebrates is remarkably
poor. Studies that suggest that brood size is uncon-
strained by brood space include species ofpolychaetes
2- (Daly, 1972; Gremare and Olive, 1986; Hess, 1993),
bivalves (Ockelmann and Muus, 1978; Kabat, 1985;
McGrath andOFoighil. 1986; Russell and Huelsenbeck,
1989; Brey and Hain, 1992), asteroids (Menge, 1974),
ophiuroids(Rumrill, 1982; Byrne, 1991), holothuroids
10 20 30 40 50 60 70 80 90 100 (Rutherford, 1973), and pinnotherid crabs (Hines.
% 1992). In fact, when the study on Asterina phylactica
Fertilization
that provided the first empirical evidence for the hy-
Figure5. Leptosvnaptaclarki. Numberoffemaleseacucumbersin pothesis (Strathmann el al.. 1984) was reanalyzed with
eaancdh1p9e9r0c.enA't=fer2t8ilfizeamtailoens.class during November and December 1989 RMA regression, no scaling constraints on brooding
were indicated (Hess, 1993).
One species that might show allometric constraints
which isaboutthecritical sizeof200-250 mgdetermined on brooding is the hermaphroditic, bursal-brooding
from field collections. The other 70 individuals, although Axiognathusscniamata (Rumrill, 1982). In thisspecies,
ofthesameage,were maleorofundetermined sex. Seven the relationship between brood number and parental
of these 70 sea cucumbers were above 185 mg TW but body size is significantly different from zero, but with
=
had notchanged sex. Theseresultssuggest, therefore,that a low slope (untransformed data b 0.57; Rumrill.
size in combination with some environmental or genetic 1982). This indicates that smalleradultsbrood numbers
factor may be more important than age to the timing of ofembryos nearly equivalent to those brooded by larger
adults (Rumrill. 1982). However, inAxiognathussqua-
sex change in L. clarki.
mata. staggered oocyte maturation and brood release
ensures that the production of oocytes never exceeds
Brooding andsmall body size the capacity ofthe adult to brood the developing em-
Brooding is size-dependent in many marine inverte- bryos (Rumrill, 1982). Therefore, the observed allo-
brates, and species with smaller adults protect their em- metric constraint on brood size in this species may de-
bryosto moreadvanced stagesofdevelopment (see review pend more upon the rate ofoocyte maturation than on
by Strathmann, 1990). This trend has been noted in two body size (Rumrill, 1982).
familiesofholothurians, theCucumaridae (Menge, 1975) In L. clarki. sperm released externally entersthe ovary
and the Synaptidae (Strathmann, 1990). L. clarki con- through an as yet unknown means (Everingham, 1961;
formstothese trends in beingsmall (adult size maximum McEuen, 1987; Hess elal, 1988; Sewell andChia, 1994),
at Bamfield is 1 13 mm) and brooding pentactulae in the but probably via the gonopore. Large amounts ofsperm
ovary (Everingham, 1961; McEuen, 1986; Sewell and are available during the spawning season as a result of
Chia, 1994). early reproduction, which skews the sex ratiodramatically
A test ofthe allometry hypothesis to explain the asso- towards males (Table II), and the production of large
ciation ofsmall adult size with brooding(Strathmann and quantitiesofsperm byeach male(Sewell andChia, 1994).
Strathmann. 1982) in L. clarkidid not show proportion- Although all males invest a similar proportion ofenergy
ately smaller broods in larger sea cucumbers. The slope in reproduction (about 5% ofdrained weight, Sewell and
ofthe relationship between loge drained weight and log,, Chia, 1994), larger males will have larger volumes of
number ofpentactulae was not less than 1, indicating no sperm, which might result in higher fertilization success.
spatial constraints on brood size in larger L. clarki. However, even though sperm does not appear to be lim-
Constraints on brood size can also occur ifthe larger iting because of high-density populations (at Bamfield,
broods from larger adults develop more slowly or suffer mean density = 169 per m:; Sewell, 1993a), a skewed sex
higher mortality (Strathmann and Strathmann, 1982). In ratio, and large amounts ofsperm per male, in some fe-
L. clarki. despiteatrend forlargerfemalestohaveahigher males fertilization isloworzero. There isalso noevidence
number ofpentactulae, larger females or females with a to suggest that larger females have a higher fertilization
120 M. A. SEWELL
success. Iffertilization isaconstraint forL. clarkiorother Brooding and hennaphroditism
brooders, then limitations on brood size due to space or The association between brooding and hermaphrodit-
embryonic requirements for dissolved materials (Strath- ism has been noted in a variety ofmarine invertebrates
mann and Strathmann. 1982) may never be reached. (Ghiselin, 1969, 1974; Charnov, 1982; Strathmann and
In brooding species where fertilization success is high Strathmann, 1982), including asteroid, ophiuroid, and
(e.g., Cucumaria pseudocurata Rutherford, 1973; holothuroid echinoderms (Ghiselin, 1969, 1974). The
Ophioplocus esmarki Rumrill, 1982; Ophionereis oli- fairly high incidence ofprotandry in ophiuroids and hol-
vacea Byrne, 1991), allometricconstraintson space may othuroids was believed by Ghiselin (1969) to favor the
be avoided by brooding ofembryos to a limited stage of gene-dispersal model for hermaphroditism.
development (Daly, 1972; Ockelmann and Muus, 1978), In the gene-dispersal model, sequential hermaphrodit-
three-dimensional packingofembryos in thebrood space ism can act to prevent inbreeding due to fertilization be-
(Kabat, 1985), sequential broodingthroughouttherepro- tween siblings, especially in forms with restricted gene
ductive season (Kabat, 1985; Hess, 1993), or brooding flow (Ghiselin, 1969, 1974). For example, ifall members
within adistensible structure (Hines, 1992; present study). ofa clutch are males at the onset ofsexual maturity and
In the latter case, the ability for the brood structure to change into females simultaneously, then they are never
expand to accommodate the number of fertilized eggs able to mate with each other (Ghiselin, 1974). The ob-
might ensurethat, even when fertilization successis high, servation that there is a higher incidence of sequential
all the embryos can be brooded. In such cases, a test of hermaphroditism in forms with restricted gene flow (spe-
the allometry hypothesis may be inappropriate because, cies with no larvae or brooders) provides some support
in contrast to external brooders or species with a well- for the hypothesis (Ghiselin, 1974), though there is little
defined, solid brood space, there may not be a limit to direct evidence for reduced levels of inbreeding in se-
brood space. quential hermaphrodites (but see Hunt. 1993), and none
Alternative hypotheses to explain the association be- for brooding species. Studies on the scale ofgenetic dif-
tween small size and brooding such as lower energetic ferentiation in L. clarki (Hess et ai. 1988) and the con-
reserves (Chia, 1974), dispersal, or recruitment (Strath- firmation ofprotandry in the present study may provide
mann and Strathmann, 1982) cannot be considered in direct evidence for this hypothesis in a brooding species.
relation tobrooding in L. clarki until there iscomparative Inan electrophoreticstudy ofthreegeographically sep-
information on other sea cucumber species. Hess (1993) arate L. clarki populations on SanJuan Island, Washing-
hasrecentlysuggested that the slowerdevelopmental rate ton, Hess et ai (1988) found no evidence for inbreeding:
ofembryos in large brooders may be a factor in the evo- thetwopolymorphicloci studieddid notshowthereduced
lution ofbrooding. This time-constraint hypothesis sug- numbers ofheterozygotes expected in inbred populations.
gests that longer developmental times may be disadvan- The authorsconcluded thattherewasenough movement
tageous because the stage-specific mortality rate is expe- ofeither individuals or gametes within each population
rienced over a longer period oftime, and the number of to allow random mating and maintain heterozygosity.
broods per unit time is reduced. In the development of Similar results were obtained by Hunt (1993) in the as-
this hypothesis, Hess assumes that embryonic mortality teroidPatirit'llaexigua, which isprotandricandoviposits
may be low and relatively unimportant in brooders, but direct-developing eggs on the undersides of intertidal
this may not be true for all brooding species (Sewell, boulders (Byrne, 1992). This species has no known
1993a, b). Ifembryonic mortality is high, as in some fe- method fordispersal, butshowed noevidence ofinbreed-
male L. clarki(Sewell, 1993a, b), then thecostsofalonger ing (Hunt, 1993). In both cases, low dispersing and pro-
development in larger brooders will be even greater (H. tandric species, one with direct development, the other a
Hess, pers. comm.). brooder, showed evidence forrandom matingwithin iso-
Although the time-constraint hypothesis may be ap- lated local populations (Hess et ai, 1988; Hunt, 1993).
plicabletospeciesthat spawn more than once perseason, L. clarki juveniles released from the mother rapidly
alternative hypotheses will be needed to explain the as- form shallow burrows, and dispersal though rare-
sociation between small sizeand broodingin semelparous might occur by rafting, floating, swimming, or incidental
brooders or in individuals that produce only one brood dispersal as a result ofwater movement (Sewell, 1993a).
per season (Hess, 1993). As the latter category includes The combination ofoccasional dispersal and sequential
manyechinodermsinwhichtheassociation between small hermaphroditism might reduce levelsofinbreeding in an
size and brooding is particularly pronounced (see Strath- L. clarki population. However, because growth rates in
mann, 1990), this emphasizes that no single hypothesis juvenile L. clarki are extremely variable (Sewell, 1993a)
willbeapplicableinalltaxa(Strathmann andStrathmann, and protandry appears to be dependent upon the com-
1982). bined effects ofsize and environmental/genetic effects, it
PROTANDRY AND BROODING IN L. CLARKI 121
is unlikely that all siblings undergo a simultaneous sex ably L. inhaerens are protandric, whereas L. minuta is
change. Although L. clarki does show some evidence to reported to be hermaphroditic (Smiley el al., 1991). Fur-
support the gene-dispersal model for sequential her- ther comparative work on this genus would therefore be
maphroditism, further clarification is required to deter- ofconsiderable interest because it removes phylogenetic
mine how farsiblingsdisperse from oneanother; whether constraints from theconsideration oflife historyquestions
all, or only some, individuals in a clutch change sex at relating small size, brooding, and hermaphroditism.
the same time; and ifthere is a reduction in inbreeding
in the Bamneld population. Acknowledgments
Ifwe assume that sequential hermaphroditism is an
advantageous strategy for L. clarki, the question to be I thank my supervisor, F-S. Chia; thesis committee, J.
answered is. Why is L. clarki protandric and not pro- Holmes, H. Clifford, J. Spence; and M. Byrne and H.
togynic? Hess forcomments on the manuscript. I am also grateful
Juvenile sea cucumbers are released from the mother to D. Levitan for many thought-provoking discussions
into the adult habitat in April or May (Sewell and Chia, and E. Charnov for discussions on sex-change theory.
inpress;Sewell, 1993a). Thetransitionalgonadswith new FundstosupportthisresearchwereprovidedbytheZool-
eggsare,however,seeninsex-changinganimalsinJanuary ogy Department, University of Alberta, a Sigurd Tveit
while these pentactulae were still being brooded. There- Memorial Scholarship from Bamfield Marine Station to
fore, iftheseacucumbersweretobereproductively active the author, and an NSERC Grant to F-S. Chia.
as females in their first year, they would need to start egg
development while being brooded within the female. In Literature Cited
addition, based on the regression line for drained weight
versus number ofeggs (Table III, Fig. 4), to produceDanWy Achimtaupv,hrYo.d,iasnmdeRd.eFDreloamviaaulglh.ar1d9a7q2.anaNMorutvsenll(eEschriencohdeerrcmhees,Assutrer1i'dhee)r.-
eggs at all a sea cucumber must weigh over 37 mg Cah. Biol. Mar. 13: 433-442.
during late October. Although growth to sizes at which Aldrich, F. A., and M. M. Aldrich. 1955. Studiesonthe reproductive
eggs could be produced is possible on the basis of labo- functional morphologyofAsteroideans.I.Sexreversalinthecommon
ratory measurements (Sewell, 1993a). the time required sea star, Asteriasforbesi (Desor). Not. Nail. Acad. Nat. Sci. Phi/a.
toproduce matureeggs(10months,Jan-Nov.;Sewell and Bacc2i7,6.G.21p9p5.1. On twosexual racesofAsterinagibbosa(Penn.). Ex-
Chia, 1994) might be a constraint to early reproduction perienlia7: 31-33.
as a female. Bacci,G. 1965. Se.\Determination. Pergamon Press.Oxford. 306pp.
In terms of the size-advantage hypothesis, although Ballantine. \V. J. 1961. The population dynamics ofPatella rulgaia
there isclearly an increase in the numberofeggsproduced andotherlimpets. Ph.D.Thesis,Queen MaryCollege, LondonUni-
per female with size (Fig. 4), the potential for low fertil- Branvcerhs,itGy..2M3.61p9p.81. The biology oflimpets: physical factors, energy
ization (as number ofeggs fertilized) in all females may flow, and ecological interactions. Oceanogr. Mar. Biol. Annu. Rev.
reduce any advantage in large females. In addition, there 19: 235-380.
is a strong relationship between testis weight and male Brey,T.,andS. Hain. 1992. Growth, reproductionandproductionof
size (unpub. data), so larger males are producing larger Lissarca notorcadensis(Bivalvia: Philobryidae) in theWeddell Sea,
volumes of sperm. Until the mechanism of fertilization ByrnAen,taMrc.ti1c9a9.1.Mar.ReEpcroold.ucPrtoigo.n,Sedre.v8e2l:o2p1m9e-n2t26a.ndpopulation biology
in L. clarki is determined, it is not known whether male ofthe Caribbean ophiuroid Ophionereisolivacea. aprotandric her-
size is important to an individual's fertilization success. maphrodite thatbroodsits young. Mar. Biol. Ill: 387-399.
Byrne, M. 1992. Reproduction ofsympatricpopulationsofPatiriella
gunnii. P calcarandP. exigiiainNewSouthWales,asterinidseastars
Future research withdirectdevelopment. Mar. Biol. 114: 297-316.
Charnov, E. L. 1982. Thetheory ofsex allocation. Monographs in
Leptosynapla is an ideal genus in which to conduct Population Biology 18, PrincetonUniversity Press.Princeton. 355
further research on the association of small size, her- pp.
maphroditism, and brooding in marine invertebrates. Charnov,E.L. 1986. Sizeadvantagemaynotalwaysfavorsexchange.
These sea cucumbers are ofsmall to moderate size (Hy- / Theor. Biol. 119:283-285.
man, 1955; 12-30 cm maximum adult length [except for Chiao,pmFe.nSt.al19p7a4t.terCnlsasisnimfiacraitnieonianvnedrtaedbarpattiesv.eTsihganliafsisciaancJeugoofxdl.eve1l0-:
L. minuta, 1 cm]; McEuen, 1986); are characterized by 121-130.
burrowinghabits(Hyman, 1955);andexhibitawide range Clarke, M. R. B. 1980. Thereduced majoraxisofabivariatesample.
ofreproductive strategies from lecithotrophic broadcast Biometrika67: 441-446.
spawners(L. inhaerens, L.galliennei, L.girardii, L. tennis) Cognetti, G. 1954. La proteroginia in una popolazione di Aslerina
pancernGascodel Golfodi Napoli. Boll. Zoo/. 21: 77-80.
tocoelomic (L. minuta) and ovarian brooders (i.e., vivip- Creese, R.G.,D. R.Schiel,and M.J. Kingsford. 1990. Sexchangein
arous species: L. clarki, L. transgressor; forreferencessee agiantendemic limpet. Patellakermadecensis. from the Kermadec
Smiley el al., 1991). Ofthese species, L. clarkiand prob- Islands. Mar. Biol 104: 419-426.