See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/275583629

A New Case of Poecilogony From South America

and the Implications of Nurse Eggs, Capsule
Structure, and Maternal Brooding Behavior on
the Development of Different Larval Types

ARTICLE in BIOLOGICAL BULLETIN · APRIL 2015

Impact Factor: 1.64

CITATION READS

1 75

2 AUTHORS, INCLUDING:

Fernanda Oyarzun
Catholic University of the Most Holy Conce…
8 PUBLICATIONS 92 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Fernanda Oyarzun
letting you access and read them immediately. Retrieved on: 01 February 2016
Reference: Biol. Bull. 228: 85–97. (April 2015)
© 2015 Marine Biological Laboratory
A New Case of Poecilogony From South America and
the Implications of Nurse Eggs, Capsule Structure,
and Maternal Brooding Behavior on the Development
of Different Larval Types
FERNANDA X. OYARZUN1,2* AND ANTONIO BRANTE2
1
Departamento de Zoologı́a, Facultad de Ciencias Naturales y Oceanográficas, Universidad de
Concepción, Chile; and 2Facultad de Ciencias, Universidad Católica de la Santı́sima Concepción, Chile
Abstract. Poecilogony is the production of different lar-
val types within the same species. Although rare, poecil-
ogonous species are ideal systems for testing the evolution-
ary and ecological implication of different developmental
modes in marine invertebrates. Here, we described a new
case of poecilogony, the Southern Hemisphere spionid Boc-
cardia wellingtonensis. We used a combination of common-
garden experiments, video recordings, and in vitro manip-
ulations of individuals from three sites to (1) document the
type of poecilogony, the brooding behavior of the mother,
and the hatching process; (2) experimentally measure the
effect of nurse eggs on the growth and type of larvae
produced; and (3) document variation in the length of the
brooding period, number of capsules, larvae, and nurse eggs
of mothers from three sites to explore the potential for
plasticity in reproductive traits. These results were com-
pared to the previously reported poecilogonous species B.
proboscidea, which resembles B. wellingtonensis in size,
morphology, ecology, and reproductive strategy but differs
in capsule structure. We found that in contrast to B. pro-
boscidea, B. wellingtonensis produced larvae that, in isola-
tion and in the presence of nurse eggs, developed into a
wide range of offspring sizes. Mothers brood and hatch the
larvae with frequent partial hatching of the brood during the
brooding period. Although larvae could not liberate them-
selves, larvae crossed to other capsules as interconnections
between capsules broke during the developmental period,
potentially affecting food availability, sibling competition
Received 30 May 2013; accepted 30 January 2015.
* To whom correspondence should be addressed. E-mail:
fernanda.oyarzun@gmail.com
85
for nurse eggs, and cannibalism. Variation in brooding time
and number of capsules deposited among sites suggest local
adaptations.
Introduction
Developmental modes in marine invertebrates have im-
portant implications for ecological and evolutionary pro-
cesses, such as population dynamics, connectivity among
populations, geographic range, speciation, and extinction
probabilities (Palumbi, 1994; Siegel et al., 2003; Kinlan and
Gaines, 2003; Shanks, 2009, Haye et al., 2014). At the
extremes of a continuum, there are (1) species with indirect
embryonic development in which offspring develop entirely
or partially in the water column, and (2) species with direct
embryonic development characterized by the absence of a
free-living larval phase (Thorson, 1946, 1950; Strathmann,
1985; McEdward, 1995). Indirect developers usually have
small eggs, higher fecundity, smaller hatching sizes, and
higher dispersal potential than direct developers; while di-
rect developers usually have some degree of parental care in
the form of brooding, encapsulation, or the presence of a
protective jelly mass (Thorson, 1946, 1950; Christiansen
and Fenchel, 1979; Strathmann, 1985; Rumrill, 1990; Todd,
1998). In addition, in species that have brooding or some
sort of encapsulation and offspring that are isolated from the
external environment, different mechanisms of embryonic
nutrition have evolved. For instance, embryonic nutrition
can be in the form of an additional amount of yolk incor-
porated by the females into the eggs (lecithotrophy), or by
active feeding behavior of the embryos that eat other viable
86 F. X. OYARZUN AND A. BRANTE

Figure 1. Adults, larvae, and capsules of Type I (i.e., capsules that contain only planktotrophic larvae) and
Type III reproductive mode (i.e., capsules that contain nurse eggs, adelphophages, and larvae with partially
arrested development) of Boccardia proboscidea and B. wellingtonensis.

or not viable eggs of the clutch (ovophagy), or by canni-
balizing other embryos or larvae of the clutch (adel-
phophagy). What factors determine that some species re-
produce via direct development while others produce
intermediate free larval stages and an assortment of protec-
tive and nutritional strategies is still not well understood
(e.g., Collin, 2003a, b; Gibson and Gibson, 2004; Ellingson
and Krug, 2006; Vendetti et al., 2012; Thomsen et al., 2014;
Rawlinson, 2014).
Poecilogonous species produce more than one type of
larvae (Giard, 1905), which usually involves the production
of both planktonic and benthic larvae or juveniles that differ
morphologically or in feeding modes. Poecilogony is ex-
tremely rare and has been well documented only in six
species of sacoglossan gastropods (Elysia pusilla, E.
zuleica, E. chlorotica, Alderia willowi, Costasiella ocel-
lifera, Calyptraea lichen; reviewed in Vendetti et al., 2012;
McDonald et al., 2014) and in six species of spionid
polychaetes (Boccardia proboscidea, B. polybranchia,
Polydora cornuta, P. hoplura, Streblospio benedicti, Pygos-
pio elegans, reviewed in Collin, 2012; David et al., 2014).
In these cases, poecilogony usually involves large non-
dispersing and small dispersing offspring, a period of en-
capsulation (capsules or egg masses), nurse egg production,
or sibling cannibalism (Chia et al., 1996; Collin, 2012). This
peculiar strategy, in which two contrasting developmental
modes are present in one species, offers the opportunity to
comparatively study advantages and constraints that deter-
mine the evolution of different developmental modes, while
avoiding incidental confounding factors of interspecific
comparisons and potentially providing insight into the
mechanisms that generate developmental transitions (Levin
and Huggett, 1990; Chia et al., 1996; Krug, 1998; Gibson
and Gibson, 2004; Ellingson and Krug, 2006; Krug et al.,
2007; Clemens-Seely and Phillips, 2011). Although it is not
clear which factors determine the evolution of poecilogony,
some evidence suggests that food availability for adults or
larvae, feeding behavior of larvae, and maternal control of
hatching are key determinants driving its evolution (Gibson,
1997; Gibson and Gibson, 2004; Kamel et al., 2010; Oyar-
zun and Strathmann, 2011). Previous work on the Northern
Hemisphere poecilogonous polychaete Boccardia pro-
boscidea (Fig. 1), revealed great reproductive variation
(Gibson et al., 1999; Gibson and Gibson, 2004; Oyarzun
and Strathmann, 2011; Oyarzun et al., 2011). In this species,
some females (Type I reproductive mode) produce only
planktotrophic larvae within capsules, while other females
(Type III reproductive mode) produce within the same
capsule planktotrophic larvae, nurse eggs, and larvae that
are morphologically similar to planktotrophs but that ingest
nurse eggs and cannibalize siblings while encapsulated
(adelphophagic larvae). These adelphophagic larvae com-
plete most of their development inside the capsule and hatch
as advanced larvae or juveniles, while their planktotrophic
siblings require at least 15 days of planktonic life before
becoming competent to settle—that is, they show arrested
 POECILOGONY IN B. WELLINGTONENSIS
development (Gibson and Gibson, 2004). Experimental in
vitro manipulations supported the hypothesis that the type
of larvae (i.e., planktotroph or adelphophage) is determined
early in development, before setigers are formed (Oyarzun
and Brante, 2014). Along the west coast of North America
only Type I females have been reported in populations
that inhabit lower latitudes, below Point Conception
(34°26⬘53.211⁄2N, 120°28⬘17.18⬙W), while Type III females
are found throughout the species’ distribution along the
west coast of North America (Oyarzun et al., 2011). Under
common-garden conditions, Type III females from higher
latitudes produce embryos with a greater number of nurse
eggs and a longer brooding time. Finally, Oyarzun and
Strathmann (2011) showed that mothers can regulate larval
stage at hatching, cannibalism, and overall sibling compe-
tition by opening capsules earlier or later in response to
environmental temperature. At higher temperatures, females
not only reduced brooding periods but also released em-
bryos at earlier stages of development, which reduces the
time during which sibling cannibalism can occur.
In this study, we report a new case of poecilogony, the
Southern Hemisphere species Boccardia wellingtonensis,
which resembles B. proboscidea in size, morphology, ecol-
ogy, and reproductive strategy, but which lays capsules that
are interconnected, potentially intensifying sibling compe-
tition for food and by cannibalism (Fig. 1). This species was
originally described in New Zealand, but it has also been
reported in Chile and South Africa. Previous studies re-
ported capsules with nurse eggs and larvae of different sizes
inside and suggested that separation among capsules could
break down over time and that larvae might hatched unaided
(Kamel et al., 2010; Oyarzun and Strathmann, 2011). In the
present study, we document the poecilogony of this species
and hypothesize that, similarly to B. proboscidea, B. wel-
lingtonensis (1) has at least two larval types, which are
determined very early in development, before setigers are
formed; (2) shows active brooding behavior; and (3) has
intracapsular food availability that should affect larval
growth rate and larval type. In contrast to B. proboscidea
and based on previous reports, females should not control
the hatching process in B. wellingtonensis and larvae should
be able to liberate themselves (Kamel et al., 2010, Oyarzun
and Strathmann, 2011). Using a combination of common-
garden experiments, video recordings, and in vitro manip-
ulations of individuals from three sites, we (1) documented
the type of poecilogony, the type of larvae produced, the
brooding behavior of the mother, and the hatching process;
(2) experimentally measured the effect of nurse eggs on the
growth and type of larvae produced; and (3) documented
variation in the length of the brooding period, and number
of capsules, larvae, and nurse eggs by using mothers from
three sites to explore the potential for plasticity in repro-
ductive traits.
87
Materials and Methods
Collection and maintenance of animals
Boccardia wellingtonensis Read 1975 was collected
at three localities in Chile: Coliumo (36°33⬘11⬙S;
72°57⬘24⬙W), Los Molinos (39°50⬘46⬙S; 73°23⬘54⬙W), and
Quempillén on Chiloé Island (41°52⬘8⬙S; 73°46⬘16⬙W).
Worms were found at high densities in the high intertidal
zone, where they construct U-shaped sedimentary tubes
directly in the sediment, in crevices, or by perforating
sedimentary rock. Individuals were easily removed and
taken to the laboratory, where species identity was con-
firmed under dissecting and compound microscopes. The
Chilean spionid species used in our study was originally
assigned by Blake (1983) to Boccardia polybranchia Has-
well 1885, but Sato-Okoshi and Takatsuka (2001) reas-
signed it to the New Zealand species Boccardia welling-
tonensis. Recently, the correspondence of the Chilean
species to the New Zealand counterpart has been questioned
(pers. comm., V. Radashevsky, Institute of Marine Biology
of the Russian Academy of Sciences), but we have main-
tained the species name of B. wellingtonensis in this study
until further information is published. Simon et al. (2010)
review in detail the morphological characteristics of B.
wellingtonensis and B. polybranchia across the sites where
they have been reported. We used descriptions in Sato-
Okoshi and Takatsuka (2001) and Simon et al. (2010) to
identify the species in this study.
Methods of rearing worms in the laboratory were slightly
modified from Oyarzun and Strathmann (2011). Worms
were kept in pairs, initially identified as male and female, in
individual 50-ml beakers with 5 cm3 of sediment. Each
beaker was placed individually inside a 500-ml transparent
plastic container, with 300 ml of seawater. Containers
were maintained with constant aeration, at a temperature of
20 °C, and with a 16:8-h light/dark cycle. Water was
changed three times per week, and worms were fed after
water changes with a 50:50 mix of Nestum mixed cereal
(Nestle) and Zoplan (Advanced Zooplanton Diet). Animals
were acclimated for one month before using them in experi-
ments. A total of 70 pair of worms was used during the
experiment, and the number of replicates in each experiment
is indicated in the following sections.
Experiments on the effect of food on larval growth rate
and larval type
Previous work on B. proboscidea had shown that when
embryos were raised in isolation and with nurse eggs as the
only food source, they grew into two different sizes, dem-
onstrating that the adelphophagic or planktotrophic feeding
behavior of the offspring is established very early in devel-
opment (Oyarzun and Brante, 2014). To test whether em-
bryos of B. wellingtonensis had intrinsic differences in
88 F. X. OYARZUN AND A. BRANTE
growth rate and feeding preferences, or if different larval
sizes are simply the result of intracapsular competition for
food, we separated embryos at early developmental stages
(metatrochophore stage) and followed their development in
vitro. In this study we use the term embryo to refer to all
developmental stages in which segments with chaeta (seti-
gers) have not been formed, and larva when setigers have
been formed, independent of the location of the individual
(inside or outside the capsule). Two small clutches (20
capsules and 26 capsules) respectively produced by two
different females were divided among three treatments: (1)
one-third of the clutch remained untouched as a control, (2)
one-third of the embryos were fed with nurse eggs (10 nurse
eggs per embryo), and (3) one-third of the embryos were fed
with Dunaliella salina (10,000 cells/ml). The untouched
control capsules were kept together as a whole, while each
individual embryo of other treatments was separated and
placed in a 1.5-ml Eppendorf tube, with 0.2 ml of filtered
seawater and antibiotics (a solution of 50 ␮g/ml streptomy-
cin and 50 ␮g/ml penicillin; Strathmann, 1987). Although
capsules are physically attached to each other (intercon-
nected), the capsules can be separated by pulling them apart
with forceps at an early stage. We hypothesize that capsules
are laid very close together and the touching sides stick to
each other and remain “soft” while the areas that are ex-
posed to seawater harden over time. This would explain
why the divisions among capsules are broken by larvae, but
not the outer wall. Each larva was measured every 3rd day
for a total of 15 days, recording maximum length (in ␮m)
and setiger number (number of segments with chaeta, a
common measurement used to describe polychaete size),
without magnesium chloride and using an ocular microme-
ter. Water was changed after measurements, and algal den-
sity and nurse egg number were restored. To rule out the
possibility that differences in larval size were due to pater-
nal effects (given that females can store spermatophores,
and thus the same clutch could potentially have siblings
with different fathers), we used females that had hatched as
larvae in the laboratory and had been exposed to only a
single male for a few days following metamorphosis. For
this experiment we used capsules produced by females
obtained from Coliumo.
We evaluated differences in the frequency of the final
size of larvae from the nurse egg and the D. salina treat-
ments. We calculated the coefficient of variation (CV) and
the range of larval size obtained from the two treatments.
Larvae from the control group were grown together and
experienced cannibalism, therefore their data are not inde-
pendent; however, the control group was used to help assess
whether the maximum size attained by larvae inside the
untouched capsule differed from the maximum size of the
larvae in the other two treatments. Normality and homoge-
neity of variance were checked with the Shapiro-Wilk W
and Levene tests. When appropriate, the post hoc Tukey test
was used to evaluate differences between groups. Analyses
were performed using the statistical package JMP 10.0.0
(SAS Institute, 1989 –2008). In one case of non-compliance
with assumptions (larval size comparison between two
types of food), which could not be resolved by transforma-
tions, we used a randomization-based approximation anal-
ogous to the one-way ANOVA (PERMANOVA) with
50,000 iterations, in ECOSIM (Gotelli and Entsminger,
2004).
Description of maternal brooding and hatching process
Previous work conducted with B. proboscidea had shown
that mothers actively brood their capsules and hatch larvae
by pulling capsules until they tear open. In B. proboscidea,
hatching occurs as a single event, where larvae are not able
to hatch unaided, and shortening the time to hatching re-
duces levels of cannibalism (Oyarzun and Strathmann,
2011). In the present study we made descriptive observa-
tions to document whether B. wellingtonensis females also
brooded their capsules and how they did so (e.g., ventilating
them, cleaning them, or using other caring behavior). We
also observed whether larvae were able to hatch unaided.
Female brooding behavior was filmed through the bottom of
the beaker with a Canon Vixia HF camera attached to a
Nikon dissecting scope, which was inverted for the obser-
vations. Filming was possible because, like B. proboscidea,
B. wellingtonensis females tend to deposit capsules against
a hard structure. Given that the beakers had only a small
amount of sediment, females deposited capsules against the
bottom of the beaker, making the capsules and the brooding
behavior of the female visible through the glass. Female
behavior was followed and periodically documented
throughout the brooding period and hatching process for
five females, coming from each of the three localities. Each
female was checked daily, and videos and notes were taken
to document the types of behavior observed. When larvae
appeared to be advanced in their development, we filmed
females constantly in order to capture the hatching process
of the larvae on digital videorecordings.
Common-garden comparison of brooding time and
reproductive traits of individuals from three localities
Work on B. proboscidea had shown that under common-
garden conditions females from a high-latitude population
brooded their offspring for longer periods than did females
coming from a low-latitude population. Also, females from
the lower latitude population provided their offspring with
fewer nurse eggs per embryo per capsule, although there
were no differences in the number of capsules among sites
(Oyarzun et al., 2011). To test for variation of reproductive
traits in B. wellingtonensis, we compared the number of
capsules, embryos, and nurse eggs, and the brooding time in
clutches from females originating from the three localities,
 POECILOGONY IN B. WELLINGTONENSIS
Coliumo, Los Molinos, and Quempillén, under common-
garden conditions. A total of 70 females, 35 from Coliumo,
15 from Los Molinos, and 20 from Quempillen, were used
in these experiments and kept under the common-garden
conditions previously described in the “Collection and
maintenance of animals” section, throughout the brooding
process. For each female we recorded the number and type
(i.e., Type I or Type III) of capsules deposited in each brood
as well as the size of the female (measured in number of
setigers).
Additionally, we followed the brooding period for
a subset of 40 females—from the 70 females described
above—20 from Coliumo, 10 from Los Molinos, and 10
from Quempillén, by checking daily for liberation of cap-
sules. When capsules disappeared from the female’s tube,
we checked for larvae swimming in the water or juveniles
settled in the sediment to confirm hatching (versus possible
ingestion by the mother). Data were analyzed using a one-
way ANOVA and Student’s t-test to compare brooding
times from different localities and reproductive types (Type
I vs. Type III), respectively. Correlation analyses were used
to evaluate whether brooding time and female size were
related. For a second subset of 18 females, 9 from Coliumo,
6 from Los Molinos, and 4 from Quempillén, we also
followed two consecutive clutches from each female to
confirm that females are able to produce more than one
clutch in their lifespan, and to test for within-female vari-
ability in the number of capsules produced. We calculated
the coefficient of variation (CV) for each female, compared
it with the CV at the locality level, and then tested for
differences among localities using a one-way ANOVA. We
kept 10 of these 18 females in order to quantify the maxi-
mum number of clutches they could deposit in their lifes-
pan.
Finally, to test for geographic variation in the number of
nurse eggs and embryos deposited by Type III females, we
removed clutches from 19 females coming from the three
localities, 5 from Coliumo, 7 from Los Molinos, and 7 from
Quempillén, and counted the number of capsules, number of
embryos, and number of nurse eggs in 10 capsules of the
clutch, as well as measuring female size. To ensure that
cannibalism had not yet taken place and that the number of
nurse eggs had not been altered by consumption, we made
these measurements only a few days after deposition, when
embryos were distinguishable but had not yet developed
setigers, and their feeding activity had not been enough to
modify the number of nurse eggs present. To count the
content of each capsule, we gently broke one side of a single
capsule, releasing its content. Once the content of that
capsule had been quantified, we cleaned the area around the
string of capsules and broke the next capsule. We analyzed
data using a one-way ANOVA, after checking that worm
size was not correlated with either response variable (i.e.,
number of capsules, number of embryos, and number of
89
nurse eggs). We compared the average and CV of the three
variables with those obtained for B. proboscidea in Oyarzun
et al. (2011). We checked for normality of data using the
Shapiro-Wilk W test and homogeneity of the variances
using the Levene test. The post hoc Tukey test was used to
determine which groups differed. Analyses were performed
in the statistical package JMP 10.0.0 (SAS Institute, 1989).
Finally, we observed and documented the brooding behav-
ior of all females with video recordings to complement
observations of brooding made in the previous section.
Results
Two reproductive modes were found in females of Boc-
cardia wellingtonensis (similar to B. proboscidea) from all
of the studied localities: Type III females with nurse eggs,
adelphophagic larvae, and larvae with partially arrested
development (i.e., the expected mode; Fig. 1) which was the
expected type of poecilogony, and Type I females with only
planktrotrophic larvae, which had not previously been re-
ported for this species (Fig. 1). Although Type I females
were observed in the field at the localities of Los Molinos
and Quempillén, they were not as abundant as at Coliumo.
Actually, of all the females raised in the laboratory, 55%
from Coliumo were Type I, while only 26% from Quem-
pillen and ⬍1% from Molinos were of this type. Females of
both reproductive types were morphologically indistin-
guishable (on the basis of the characters studied to date) and
coexisted in the same environment. We expanded our ex-
periments and observations on maternal brooding and
hatching to include both reproductive types when possible,
but most of our discussion is focused on the Type III
developmental mode, which has the highest potential for
intrafamilial conflict.
Experiments on the effect of food on larval growth rate
and larval type
Pre-setiger embryos of Type III capsules of females from
Coliumo that were grown in isolation and with the constant
presence of 10 nurse eggs grew into a wide range of sizes,
from 5 to 18 setigers with an average of 13.3 setigers and a
mode of 17 setigers (Fig. 2A, B). In contrast, larvae grown
with Dunaliella salina as their only food source grew into a
narrower range of sizes and a smaller maximum size than
larvae grown with nurse eggs, oscillating between 9 and 14
setigers with an average of 12.5 and a mode of 14 setigers
(Fig. 2A, C). The coefficient of variation of the offspring
size when embryos were grown with nurse eggs was double
(27.8) the amount of when they were grown with D. salina
(14.4). The PERMANOVA did not detect differences in
the average final size of larvae raised on each diet
(Fpseudo(1,11) ⫽ 0.43, pperm ⫽ 0.54).
90 F. X. OYARZUN AND A. BRANTE

A) Experimental design and results B) Fed with nurse eggs

5
adelphophages
4

Frequency
partially arrested larvae
3

2
pre-setiger
embryos 1

2 4 6 8 10 12 14 16 18
Experimental
Treatments

C) Fed with Dunaliella salina

5

Frequency
D. salina Nurse eggs 4
12.5 13.3 Average
(# setigers)

11 16 N 3
Results

1.8 3.7 Stdev.

2
14.4 27.8 CV
5 13 Range 1

2 4 6 8 10 12 14 16 18
size of larvae (# of setigers)

Figure 2. Results of the experiment on the effect of food on larval fate of partially arrested larvae of Boccardia
wellingtonensis. (A) Experimental design and descriptive statistics of the results for both experimental treat-
ments. (B) Frequency of the size of larvae grown individually in Eppendorf tubes after 15 days with nurse eggs
as their only food source. Dotted line indicates that the limit separating the definition of larval types is diffuse.
(C) Frequency of the size of larvae grown individually in Eppendorf tubes after 15 days with Dunaliella salina
as their only food source. The temperature in this experiment was maintained at 20 °C.

Description of maternal brooding and the hatching
process
Of the five B. wellingtonensis females that were filmed
throughout the brooding period, two proved to be Type III
and three were Type I. Females of both reproductive modes
brooded their capsules and stayed with them until hatching.
Females ventilated capsules with undulatory movements,
cleaning the tubes by removing any foreign object; they
used their palps to touch capsules and mouthed the capsules,
presumably to clean the surface. Females continued feeding
while brooding, but frequently maintained nearly all of their
bodies in contact with the capsules, only ingesting particles
that entered their tubes or that could be caught with the part
of the palps that they were able to extend outside of the
tubes, without moving away from the capsules.
Digital video recordings and observations suggest that
cannibalism does occur in the Type III reproductive mode
in B. wellingtonensis. Digital video recordings showed
smaller larvae with arrested development progressively and
abruptly disappearing in advanced capsules (Fig. 3A, B),
although the actual ingestion of a sibling was not recorded
on video, and given that the visibility of the whole capsule
space simultaneously was limited, we could not quantify the
rate of cannibalism. Connections among capsules broke
over time and larger larvae (with up to 20 setigers and
already metamorphosed into juveniles) crossed between
capsules. Frequently, capsules ran out of nurse eggs before
larvae hatched. When the connections between capsules
were broken, adelphophages broke into adjacent cap-
sules and continued ingesting nurse eggs from the other
capsules. In addition, we observed not two, but a range of
larval sizes during development in Type III reproductive
mode, both while filming broods and after larvae had been
released from the capsules.
Females of both reproductive modes (Type I and III)
from the three sites hatched their embryos in a way similar
to what has been described for B. proboscidea (Fig. 3C, D);
however, the liberation process sometimes occurred in more
than one event and across several days. At liberation events,
females executed violent movements to open or break cap-
sules by pulling and also pushing them, as opposed to the
previous smoother movements of the brooding period. Their
movements were similar to those previously observed in B.
proboscidea (Oyarzun and Strathmann, 2011); however, the
differences in capsule structure (interconnected capsules)
seem to be important in determining the type of movements
that females were able to accomplish when cleaning and
breaking the capsules, and resulted in less precise manipu-
lation. B. wellingtonensis females cannot pull individual
capsules as B. proboscidea do; instead we observed females
pulling a bunch of capsules from one tip, and also pushing
them forward with the mouth wide open. Although some
capsules had been previously lost in what could be uninten-
tional liberation due to the friction of the body, one final
 POECILOGONY IN B. WELLINGTONENSIS 91

A) B)

partially arrested
larva

nurse egg

adelphophage

A)
C) D)

female pulling capsule

(ventral view) larvae
hatching

Figure 3. Observations of the capsules and female hatching behavior in Boccardia wellingtonensis. (A) Single
frames of digital video recordings and (B) schematic representations of capsules of the Type III reproductive mode,
inside the female tube. One capsule with two arrested larvae, and two capsules with one adelphophagic larva are
visible. The female’s body is visible in the bottom right corner of the video frame. (C) Single frame of video recording
and (D) schematic representation of Type I female pulling capsules to liberate larvae. The female pulled and pushed
capsules until they broke open, then larvae swam and were expelled from the tube by the undulating motions of the
female; this process is similar to that in B. proboscidea (Oyarzun and Strathmann, 2011).

major hatching event took place in 4 of the 5 females
recorded. The three Type I females deposited 21, 32, and 35
capsules, respectively, and took 10, 7, and 18 min to open
capsules during the final hatching event that liberated most
of the larvae. The two Type III females deposited 30 cap-
sules each. One of them took 19 minutes to hatch the larvae
in the final hatching event that liberated most of the cap-
sules; however, the other female liberated a few capsules
each day for several days. In the last hatching event it took
her 26 minutes to open only 5 of the remaining 30 capsules.
Additional observations on brooding behavior of the fe-
males that were used for the common-garden experiment
confirmed that the hatching process frequently occurred
over several days. Among these females, 45% of the brood-
ing events resulted in at least some capsules been opened
before the end of the brooding period of the female. Given
that capsules are interconnected, when females made undu-
latory movements they usually stressed one or two sections
of the clutch more than other areas. After a few days,
capsules were absent from those sections. This “partial
liberation of the clutch” seemed to be accidental. However,
two females were filmed pulling capsules unsuccessfully in
the middle of the brooding period, and others were recorded
liberating some of the capsules and continuing to brood for
a couple of additional days. Although large larvae could
break the connections between capsules, we did not find
evidence that larvae could break the capsules to release
themselves while they were inside the mother’s tube. When
capsules of both types were cultivated without the presence
of the female, either because the mother had abandoned the
clutch (in three cases) or because capsules were left outside
the tube in circulating seawater when they were removed for
observations after collection from the field, capsules got
infected after a few days if no antibiotics were used. This
observation supports the hypothesis that the mother is es-
sential to the successful development and hatching of her
offspring. Finally, two females were observed aborting a
complete clutch by pushing it outside the tube, one female
was observed ingesting a complete capsule, two females
ingesting loose nurse eggs, and one female defecating par-
tially digested capsules.
Common-garden comparison of brooding time and
reproductive traits of individuals from three localities
(see Fig. 4)
Of the 40 females that we followed throughout the brood-
ing period, 23 were Type III (7 from Coliumo, 10 from Los
Molinos, and 6 from Quempillén), 10 were Type I from
Coliumo, and the rest did not complete their brooding either
due to abortion of the clutch or death of the female. It was
not possible to obtain brooding time data for Type I females
from Los Molinos and Quempillén in this study, due to their
low abundance in the field. On average, the Tukey post hoc
92 F. X. OYARZUN AND A. BRANTE

A) Boccardia proboscidea (redrawn from Oyarzun et. al 2011)

50 Female size restricted

Brooding time (days)

20
B to ~55-60 setigers
40

# of capsules
15 AB A
30
B
10 A B
20
5 10

Type I Type III

Type I Type III

B) Boccardia wellingtonensis (this study)

Both reproductive types
Brooding time (days)

B 50

# of capsules
20
AB 40
15 A
A 30
10 20
5 10

40 60 80
Type I Type III Female size (# of setigers)

Figure 4. Comparative results of common-garden experiments for Boccardia wellingtonensis and B. pro-
boscidea. (A) Brooding time and number of capsules deposited by Type I and Type III females of B. proboscidea
coming from two sites. Data were redrawn from Oyarzun et al. (2011). (B) Brooding time and number of
capsules deposited by Type I and Type III females of B. wellingtonensis coming from three sites (this study).
Common-garden conditions for both species were 20 °C, 16:8 L/D cycle, and constant aeration.

analysis indicated that under common-garden conditions
Type III females from Coliumo brooded their offspring for
a shorter period than Type III females from Quempillén, and
that the average brooding time for females from Los Moli-
nos was intermediate between the brooding times for fe-
males from Coliumo and Quempillén, and not significantly
different from either (Fig. 4B; one-way ANOVA: F(2,20) ⫽
4.88; P ⫽ 0.02).Type I and Type III females from Coliumo
brooded their capsules for a similar amount of time (t-test:
t(15) ⫽ ⫺1.23; P ⫽ 0.24), although larvae from the Type I
reproductive mode were ready to hatch (i.e., had 3– 4 seti-
gers) much earlier than liberation actually took place. Fe-
male size was not correlated with brooding time (correla-
tion: r(23) ⫽ 0.03; P ⫽ 0.89). To fulfill the assumption of
normality for the t-test for Coliumo, data were transformed
using the equation 1/sqrt(brooding time ⫹1). For the cor-
relation analysis, body size data were transformed using
“log10”. In all other cases data were normal and variance
was homogeneous without transformation.
Females of both types deposited capsules against the
beaker as expected; however, in some reproductive events
part of the brood of a female was hidden from view and it
was not possible to count the number of capsules. For 70
females showing complete bodies (i.e., the pygidium was
present) and coming from the three sites, we had clutches
that were completely counted (it is easy to distinguish when
the clutch is complete because the capsules at the tip of a
clutch are much smaller. Female size (i.e., number of seti-
gers) was measured as soon as liberation took place. Log
transformation was used to meet the normality assumption,
and the slopes of the regressions between female size and
number of capsules were not statistically different from
each other. Female size, considering both female types, was
positively related to the number of capsules deposited
(ANCOVA: F(1,66) ⫽ 45.93; P ⬍ 0.001), and the site of
origin of the females significantly affected the number of
capsules that were deposited (ANCOVA: F(1,66) ⫽ 10.01;
P ⬍ 0.001; Fig. 4B). The Tukey post hoc analysis indicated
that females from Coliumo deposited more capsules than
females from Los Molinos and Quempillén (P ⬍ 0.05). A
comparison of the number of capsules deposited by Type I
and Type III females did not reveal significant differences in
Coliumo (ANCOVA: F(1,29) ⫽ 0.70; P ⫽ 0.41), which was
the only site for which we had enough data on different
reproductive types to allow this comparison.
To test whether variation in the number of capsules
deposited by females at different times is larger than the
variation found among females within a site or between
sites, we counted the number of capsules deposited in two
consecutive clutches by females of both types coming from
the three sites. There were no locality differences in the CV
of the number of capsules deposited on two consecutive
 POECILOGONY IN B. WELLINGTONENSIS
Summary of average number of nurse eggs and embryos per capsule for two species of Boccardia
Average number of nurse Average number of embryos
Species eggs per capsule
B. proboscidea ␮ ⫽ 74.2 ⫾ 4.0
CV ⫽ 36.9
B. wellingtonensis ␮ ⫽ 64.7 ⫾ 5.3
CV ⫽ 25.9
Average (␮), standard error, and coefficient of variation (CV). In each clutch we measured 10 capsules. There were no significant differences between
species for any of the response variables.
reproductive events by each female (one-way ANOVA:
F(2,15) ⫽ 2.00; P ⫽ 0.17). Overall average CV, considering
one clutch per female at all localities, was higher (27.8) than
any single CV for a female (range ⫽ 2.3 to 25.7), and higher
than the average CV for each locality (Coliumo ⫽ 9.7; Los
Molinos ⫽ 16.8; Quempillén ⫽ 12.5). In 72% of the fe-
males, the second clutch was larger than the first clutch, as
expected due to female growth.
For females with Type III reproductive mode, we did not
find a correlation between female size and either average
number of embryos per capsule (r(15) ⫽ 0.13; P ⫽ 0.14) or
average number of nurse eggs per embryo (r(15) ⫽ 0.06;
P ⫽ 0.33). We did find a significant correlation between
female size and the average number of nurse eggs per
capsule (r(15) ⫽ 0.36; P ⫽ 0.01), but it was driven by only
one datapoint from Coliumo and the correlation was lost
when that datapoint was excluded (r(15) ⫽ 0.13; P ⫽ 0.17).
To evaluate differences between sites for the three variables,
we performed one-way ANOVA tests. The three variables
were normal and homogeneous after transforming the av-
erage number of embryos per capsule and the average
number of nurse eggs per embryo using a Log10 transfor-
mation. There were no significant differences among sites
for all three response variables: (1) average number of nurse
eggs per capsule (one-way ANOVA: F(2,16) ⫽ 1.48, P ⫽
0.26), (2) average number of embryos per capsule (one-way
ANOVA: F(2,16) ⫽ 0.29, P ⫽ 0.75), and (3) average number
of nurse eggs per embryo (one-way ANOVA: F(2,16) ⫽
0.80, P ⫽ 0.47).
The average number of nurse eggs allocated to each
embryo per capsule was 22.2 ⫾ 4.8 (average ⫾ standard
error) in B. wellingtonensis, and it was not significantly
different than the amount allocated by B. proboscidea—
20.1 ⫾ 1.6 (t-test: t(37) ⫽ 0.42; P ⫽ 0.66)— considering all
sites for which information was available. The CV for this
response variable was also not significantly different be-
tween the two species (t-test: t(37) ⫽ 0.42; P ⫽ 0.66). Log10
transformation was used to meet the assumption of normal-
ity. We summarized the average number of nurse eggs and
embryos for both species in Table 1.
93
Table 1
Average number of nurse
eggs per embryos per Number of clutches
per capsule capsule measured
␮ ⫽ 5.3 ⫾ 0.4 ␮ ⫽ 20.1 ⫾ 1.6 20
CV ⫽ 53.3 CV ⫽ 64.6
␮ ⫽ 5.9 ⫾ 1.2 ␮ ⫽ 22.2 ⫾ 4.8 19
CV ⫽ 53.0 CV ⫽ 52.8 19
Discussion
Our results support the hypothesis that in the poecil-
ogonous Type III reproductive strategy of Boccardia wel-
lingtonensis, larval types are established early in develop-
ment, before setigers are formed, and that larval type and
growth rate is influenced by intracapsular food availability,
food competition, and sibling cannibalism. However, in
contrast to the congeneric species B. proboscidea, we found
a wider range of larval sizes established earlier in development,
which can be further modified by the crossing of larvae to
adjacent capsules, thus modifying food availability and sibling
interaction. We reject the hypothesis that larvae can hatch by
themselves, and we found strong evidence that females control
hatching of capsules, potentially limiting the amount of com-
petition and cannibalism that takes place. B. wellingtonensis
females coming from the three study sites investigated in this
study showed geographic variation and potential local adapta-
tion to different environmental conditions.
The role of capsule structure on larval feeding and
maternal behavior
The Polydora-complex represents a group of about 150
morphologically similar spionids that all have a modified
5th setiger; it includes the genera Boccardia, Polydora,
Dipolydora, and Pseudopolydora, among others (Blake,
1996; Walker, 2011). Blake (1969) described three types of
egg capsules within the Polydora-complex (Fig. 5): (A) A
simple cylindrical sac that is attached to the sedimentary
tube via filaments, which can be found in Dipolydora
quadrilobata, D. giardi, and Boccardia semibranchiata
(Blake, 1969; Day and Blake, 1979; Guerin, 1991). (B) A
bead-like string, which is the most frequent form, occurring
in the capsules of B. wellingtonensis and is also found in
Polydora websteri, P. uncinata, P. rickettsi, P. hoplura, and
Dipolydora aciculata (Blake and Arnofsky, 1999; Sato-
Okoshi and Takatsuka, 2001; Radashevsky and Olivares,
2005; Sato-Okoshi et al., 2008, David et al., 2014). (C) A
series of separated individual capsules attached only to the
wall of the tube but not to other capsules, such as the
94 F. X. OYARZUN AND A. BRANTE

Type I Type III

(only larvae) (larvae + nurse eggs)
+
Capsule structure complexity

C1 C2

B1 B2

A1 A2

sibling cannibalism 0% ++
maternal control 100%
Figure 5. Schematic representation of trade-offs involved in the evolution of capsule structure and poecilogony in
the Polydora-complex. Letters represent the three types of capsules found in the Polydora-complex: A, a simple
cylindrical sac that is attached to the sedimentary tube via filaments; B, a bead-like string, which is the most frequent
form; and C, a series of individual capsules attached to the wall of the tube but not to other capsules.

capsules of B. proboscidea, which are also found in B.
columbiana and Polydora cornuta (Blake and Arnofsky,
1999; Gibson et al., 2012; Surugiu, 2012). In our study we
found that the interconnected capsules of B. wellingtonensis
restricted females from cleaning capsules and generated
incidental content release when females ventilated capsules
with body movements. In addition, it allowed adel-
phophagic larvae to intrude into other capsules, increasing
the potential for competition and cannibalism among sib-
lings. In contrast, B. proboscidea produced individual cap-
sules that restricted both cannibalism and competition, and
allowed females to clean and manipulate each capsule in a
more precise manner than in B. wellingtonensis. Females of
B. proboscidea are able to adjust the stage of development
at hatching in response to changes in temperature, and their
hatching process occurs as one single event (Oyarzun and
Strathmann, 2011). Overall, the individual capsules of B.
proboscidea give mothers more control over the distribution
of resources among siblings than in B. wellingtonensis. This
difference between species could be increasingly important
if there are differences in the number of patrilines within
clutches and throughout the reproductive season, and if
multiple paternity plays a role in sibling cannibalism as it
does in the gastropods Solenosteira macrospira (Kamel
et al., 2010) and Crepidula coquimbensis (Brante et al.,
2013).
If bead-like strings of capsules are more common than
individual capsules within the Polydora-complex, and if
there are greater conflicts of interest over the distribution of
resources when offspring share the same space, then simple
cylindrical sacs should be an ancestral character in this
group, and increasing division among embryos could have
evolved in response to intrafamilial conflicts over resources.
However, this would imply that nurse eggs and adel-
phophagy evolved first and capsule modifications in differ-
ent species occurred later. Alternatively, capsule structure
could have become more complex over time (Fig. 5 A1, B1,
C1) in response to other selective forces or constraints, and
the evolution of nurse eggs as separate evolutionary events
could have imposed different conflicts of interest over the
distribution of parental resources based on the already ex-
istent differences in capsule structure between species (Fig.
5 A2, B2, C2). The fact that B. proboscidea and B. welling-
tonensis both have Type I (Fig. 5 B1, C1) and Type III (Fig.
5 B2, C2) reproductive modes supports the latter hypothe-
sis. Also, we are not aware of any case of a species that
produces a simple cylindrical sac and also produces nurse
eggs and adelphophages (Fig. 5 A2), except in the sense that
 POECILOGONY IN B. WELLINGTONENSIS
B. wellingtonensis moves toward this stage when intercon-
nections of capsules start forming as walls break. Further
investigation of capsule characteristics and dynamics in
other Polydora-complex species and mapping of capsule
structure and larval traits over molecularly constructed phy-
logenetic trees could shed light on this problem.
Poecilogony and nurse egg allocation
Species can cope with environmental fluctuations by pro-
ducing diverse types of offspring via genetic or non-genetic
mechanisms (Meyers and Bull, 2002). In some cases, pro-
duction of specific larval types is determined by females
before the environment is experienced or it can involve the
production of multiple types to maximize the geometric
mean of fitness and reduce its variance, as in a bet-hedging
strategy. Previous studies suggest that variability in off-
spring sizes is the result of differences in nurse egg con-
sumption during embryonic development (Rivest, 1983,
1986; Chaparro et al., 1999; Moran and Emlet, 2001; Strath-
mann and Strathmann, 2006; Marshall et al., 2008). For
example, Smith and Thatje (2013) found large size differ-
ences among siblings developing inside capsules in the
whelk Buccinum undatum, including some embryos that
had not consumed nurse eggs. In their study they did not
find evidence of cannibalism, implying that size differences
were due only to differences in nurse egg consumption. Our
in vitro manipulation experiments with B. wellingtonensis
showed that differences in offspring size are reached
through nurse egg consumption, but that the differences are
partly generated by a previously determined difference in an
embryo’s ability to consume nurse eggs. Our results showed
that the coefficient of variation in offspring size doubled in
isolated embryos that were grown in the presence of nurse
eggs as their only food source, compared to that of embryos
that were raised in the presence of Dunaliella salina (Fig.
2). Similar differences in feeding preferences are found in
B. proboscidea (Gibson and Carver, 2013; Oyarzun and
Brante, 2014) and are not explained by morphology alone
(Gibson and Carver, 2013). Overall, arrested larvae of both
species seem to be morphologically well equipped to con-
sume nurse eggs (i.e., they have cilia, a mouth, the ability to
ingest particles, etc.), so their lack of feeding remains puz-
zling (Gibson and Carver, 2013; Oyarzun and Brante,
2014). In B. proboscidea, feeding preference acts as an
all-or-nothing response by producing two sizes of offspring
(Oyarzun and Brante, 2014), while in B. wellingtonensis
there seems to be a gradient of feeding capacity that gen-
erates a gradient of offspring sizes (Fig. 2B). Thus, although
both species show high variation in larval feeding types and
offspring size, the gradient observed in B. wellingtonensis
should result in a higher variability of the time that larvae
expend in the plankton, which could have important impli-
cations for dispersal potential, population dynamics, and
connectivity.
95
The mechanisms by which nurse eggs (or trophic eggs)
are produced vary enormously across species (Perry and
Roitberg, 2006). There is evidence in at least two species of
the Polydora-complex, B. proboscidea and Polydora cor-
nuta, that the formation of nurse eggs is determined by the
mother and occurs through a process similar to apoptosis
(Smith and Gibson, 1999; Gibson et al., 2012; but see Rice
and Rice 2009). However, the molecular mechanisms that
determine differences in embryonic fate in poecilogonous
spionids are less clear. Phylogenetic studies using mito-
chondrial genes, random amplified polymorphic DNA, allo-
zymes, and single nucleotide polymorphism markers have
uncovered geographic differences among populations, but
not genetic differences among individuals with different
reproductive strategies in B. proboscidea (Gibson et al.,
1999; Oyarzun et al., 2011) or in the poecilogonous spio-
nids Streblospio benedicti (Zakas and Wares, 2012) and
Pygospio elegans (Morgan et al., 1999; Kesäniemi et al.,
2012). The results on growth rates of isolated embryos in
the in vitro manipulations of B. proboscidea and B. wel-
lingtonensis suggest that there could be genetic differences
between embryos.
In contrast with the poecilogonous species B. probocidea,
B. wellingtonensis, and Polydora cornuta, the reproductive
plasticity of the well-documented poecilogonous spionid
Streblospio benedicti does not involve nurse eggs, but rather
the production of larger eggs that develop as lecithotrophic
larvae and smaller eggs that develop as planktotrophic lar-
vae (Levin et al., 1987; Schulze et al., 2000). Therefore,
although nurse egg production could be a mechanism by
which poecilogony is generated in spionids, it is certainly
not the only one.
Geographic differences
In our study we found that, similar to B. proboscidea, B.
wellingtonensis females from higher latitudes brood their
offspring for longer periods under common-garden condi-
tions, showing local adaptation(Fig. 4A). Furthermore, fe-
males from Coliumo deposited more capsules than females
coming from the higher latitude locations. However, we
found no evidence that B. wellingtonensis females from
higher latitudes provide offspring with a greater number of
nurse eggs per capsule, as occurs with B. proboscidea,
indicating that this might not be a mechanism by which
females control the rate of cannibalism among the offspring.
The overall number of embryos per capsule, number of
nurse eggs per capsule, and average number of nurse eggs
per embryo were not significantly different between the two
species (B. proboscidea and B. wellingtonensis) (Table 1),
but it is important to note that these numbers apply to early
capsules. Although the initial allocation of nurse eggs and
embryos per capsule is not different between species, in B.
wellingtonensis the variation in the number of nurse eggs
per embryo changes over time as connections form between
96 F. X. OYARZUN AND A. BRANTE
capsules. Therefore, capsule structure is directly affecting
the mechanisms by which females could control resource
allocation among their offspring.
In B. wellingtonensis, female body size was not corre-
lated with brooding time, the number of nurse eggs per
capsule, or the number of embryos per capsule, but was
significantly correlated with the number of capsules depos-
ited per clutch. We were unable to test whether body size is
correlated with reproductive traits in B. proboscidea given
that in Oyarzun et al. (2011) females were restricted to a
small range of sizes. Body size is well known to have effects
on reproductive traits across phyla (Calder, 1996); further
exploration of its effect on the number of capsules, food
supply, and brooding time in B. proboscidea could increase
our understanding of the consequences and trade-offs be-
tween capsule structure of these two species.
Finally, molecular work and interbreeding experiments
are necessary to assess whether Type III and the previously
unreported Type I reproductive mode in B. wellingtonensis
do in fact correspond to the same species. If this is the case,
it will add another level of reproductive variation to the
already poecilogonous Type III reproductive mode, increas-
ing the similarities with B. proboscidea and adding to the
enormous potential of this species to serve as a model to
study the implications of different developmental modes.
Acknowledgments
We thank Cristian Canales, Oscar Chaparro, Paula Neill,
and Richard Strathmann for suggestions, assistance, or dis-
cussion of ideas. Daniela Mardones, Loreto Mardones,
Hans Maucar, Angela Montes, Adriana Quiñones, and Kat-
terina Villagrán helped in the field or lab. We thank the staff
of the Universidad Católica de la Santı́sima Concepción and
of the Universidad Austral de Chile. This research was
funded by Postdoctoral FONDECYT grant #3120202 to
Fernanda Oyarzun.
Literature Cited
Blake, J. A. 1969. Reproduction and larval development of Polydora
from northern New England (Polychaeta: Spionidae). Ophelia 7: 1– 63.
Blake, J. A. 1983. Polychaetes of the family Spionidae from South
America, Antarctica, and adjacent seas and islands. Pp. 205–288 in
Biology of the Antarctic Seas, L. Kornicker, ed. Antarctic Research
Series, Vol. 39, American Geophysical Union, Washington, DC.
Blake, J. A. 1996. Family Spionidae. Pp. 81–223 in Taxonomic Atlas of
Benthic Fauna of the Santa Maria Basin and Western Santa Barbara
Channel, Vol. 6, Annelida Part 3: Polychaeta: Orbiniidae to Cossuri-
dae, J. A. Blake, B. Hilbig, and P. H. Scott, eds. Santa Barbara Museum
of Natural History, Santa Barbara, CA.
Blake, J. A., and P. L. Arnofsky. 1999. Reproduction and larval
development of the spioniform Polychaeta with application to system-
atics and phylogeny. Hydrobiologia 402: 57–106.
Brante, A., M. Fernández, and F. Viard. 2013. Non-random sibling
cannibalism in the marine gastropod Crepidula coquimbensis. PLoS
One 8: e67050.
Calder, W. A. 1996. Size, Function, and Life History. Dover Publica-
tions, Mineola, NY.
Chaparro, O. R., R. F. Oyarzun, A. M. Vergara, and R. J. Thompson.
1999. Energy investment in nurse eggs and egg capsules in
Crepidula dilatata Lamarck (Gastropoda, Calyptraeidae) and its influ-
ence on the hatching size of the juvenile. Exp. Mar. Biol. Ecol. 232:
261–274.
Chia, F. S., G. Gibson, and P. Y. Qian. 1996. Poecilogony as a repro-
ductive strategy of marine invertebrates. Oceanol. Acta 19: 203–208.
Christiansen, F. B., and T. M. Fenchel. 1979. Evolution of marine
invertebrate reproductive patterns. Theor. Popul. Biol. 16: 267–282.
Clemens-Seely, K., and N. E. Phillips. 2011. Effects of temperature on
hatching time and hatchling proportions in a poecilogonous population
of Haminoea zelandiae. Biol. Bull. 221: 189 –196.
Collin, R. 2003a. Worldwide patterns in mode of development in
calyptraeid gastropods. Mar. Ecol. Progr. Ser. 247: 103–122.
Collin, R. 2003b. Phylogenetic relationships among calyptraeid gastro-
pods and their implications for the biogeography of marine speciation.
Syst. Biol. 52: 618 – 640.
Collin, R. 2012. Nontraditional life-history choices: what can “interme-
diates” tell us about evolutionary transitions between modes of inver-
tebrate development? Integr. Comp. Biol. 52: 128 –137.
David, A. A., C. A. Matthee, and C. A. Simon. 2014. Poecilogony in
Polydora hoplura (Polychaeta: Spionidae) from commercially impor-
tant molluscs in South Africa. Mar. Biol. 161: 887– 898.
Day, R. L., and J. A. Blake. 1979. Reproduction and larval develop-
ment of Polydora giardi Mesnil (Polychaeta, Spionidae). Biol. Bull.
156: 20 –30.
Ellingson, R. A., and P. J. Krug. 2006. Evolution of poecilogony from
planktotrophy: cryptic speciation, phylogeography, and larval develop-
ment in the gastropod genus Alderia. Evolution 60: 2293–2310.
Giard, A. 1905. La poecilogonie. Pp. 617– 646 in Compte-Rendu des
Seances du Sixième Congrès International de Zoologie, International
Congress of Zoology, 14 –16 August 1904 Bern. W. Kündig and Son,
Geneva.
Gibson, G. D. 1997. Variable development in the spionid Boccardia
proboscidea (Polychaeta) is linked to nurse egg production and larval
trophic mode. Invertebr. Biol. 116: 213–226.
Gibson, G. D., and D. Carver. 2013. Effects of extra-embryonic pro-
visioning on larval morphology and histogenesis in Boccardia pro-
boscidea (Annelida, Spionidae). J. Morphol. 274: 11–23.
Gibson, G. D., and A. J. F. Gibson. 2004. Heterochrony and the
evolution of poecilogony: generating larval diversity. Evolution 58:
2704 –2717.
Gibson, G., I. G. Paterson, H. Taylor, and B. Woolridge. 1999. Mo-
lecular and morphological evidence of a single species, Boccardia
proboscidea (Polychaeta:Spionidae), with multiple development
modes. Mar. Biol. 134: 743–751.
Gibson, G. D., H. C. Hart, C. Coulter, and H. X. Xu. 2012. Nurse
eggs form through an active process of apoptosis in the spionid Poly-
dora cornuta. Integr. Comp. Biol. 52: 151–160.
Gotelli, N. J., and G. L. Entsminger. 2004. EcoSim: Null models
software for ecology. Version 7. Acquired Intelligence Inc. and Kesey-
Bear. Jerico, VT 05465. [Online]. Available: http://garyentsminger.
com/ecosim/index.htm. [2015, 17 March].
Guerin, J. P. 1991. Cultivation of spionids (Annelida, Polychaeta). 3.
Description of the larval development of Boccardia semibranchiata.
Ann. Inst. Oceanogr. 67: 145–154.
Haye, P. A., N. I. Segovia, N. Muñoz-Herrera, F. E. Gálvez, A.
Martı́nez, A. Meynard, and S. Faugeron. 2014. Phylogeographic
structure in benthic marine invertebrates of the southeast Pacific coast
of Chile with differing dispersal potential. PloS One 9: e88613.
Kamel, S. J., F. X. Oyarzun, and R. K. Grosberg. 2010. Reproductive
biology, family conflict, and size of offspring in marine invertebrates.
Integr. Comp. Biol. 50: 619 – 629.
Kesäniemi J. E., P. D. Rawson, S. M. Lindsay, and K. E. Knott. 2012.
 POECILOGONY IN B. WELLINGTONENSIS
Phylogenetic analysis of cryptic speciation in the polychaete Pygospio
elegans. Ecol. Evol. 2: 994 –1007.
Kinlan, B. P., and S. D. Gaines. 2003. Propagule dispersal in marine
and terrestrial environments: a community perspective. Ecology 84:
2007–2020.
Krug, P. J. 1998. Poecilogony in an estuarine opisthobranch: plankto-
trophy, lecithotrophy, and mixed clutches in a population of the asco-
glossan Alderia modesta. Mar. Biol. 132: 483– 494.
Krug, P. J., R. A. Ellingson, R. Burton, and Á. Valdés. 2007. A new
poecilogonous species of sea slug (Opisthobranchia: Sacoglossa) from
California: comparison with the planktotrophic congener Alderia mod-
esta (Lovén, 1844). J. Molluscan Stud. 73: 29 –38.
Levin, L. A., and D. V. Huggett. 1990. Implications of alternative
reproductive modes for seasonality and demography in an estuarine
polychaete. Ecology 71: 2191–2208.
Levin, L. A., H. Caswell, K. D. DePatra, and E. L. Creed. 1987.
Demographic consequences of larval development mode: planktotro-
phy vs. lecithotrophy in Streblospio benedicti. Ecology 68: 1877–1886.
Marshall, D. J., R. Bonduriansky, and L. F. Bussiére. 2008. Offspring
size variation within broods as a bet-hedging strategy in unpredictable
environments. Ecology 89: 2506 –2517.
McDonald, K. A., R. Collin, and M. P. Lesoway. 2014. Poecilogony
in the caenogastropod Calyptraea lichen (Mollusca: Gastropoda). In-
vertebr. Biol. 133: 213–220.
McEdward, L. R. 1995. Ecology of Marine Invertebrate Larvae. CRC
Press, Boca Raton, FL.
Meyers, L. A., and J. J. Bull. 2002. Fighting change with change:
adaptive variation in an uncertain world. Trends Ecol. Evol. 17: 551–
557.
Moran, A. L., and R. B. Emlet. 2001. Offspring size and performance
in variable environments: field studies on a marine snail. Ecology 82:
1597–1612.
Morgan, T. S., A. D. Rogers, G. L. J. Paterson, L. E. Hawkins, and M.
Sheader. 1999. Evidence for poecilogony in Pygospio elegans
(Polychaeta: Spionidae). Mar. Ecol. Prog. Ser. 178: 121–132.
Oyarzun, F. X., and A. Brante. 2014. The effects of nurse eggs and
sibling interactions on the larval development of the poecilogonous
annelid Boccardia proboscidea (Spionidae). Invertebr. Biol. 133: 340 –
353.
Oyarzun, F. X., and R. R. Strathmann. 2011. Plasticity of hatching
and the duration of planktonic development in marine invertebrates.
Integr. Comp. Biol. 51: 81–90.
Oyarzun, F. X., A. R. Mahon, B. J. Swalla, and K. M. Halanych. 2011.
Phylogeography and reproductive variation of the poecilogonous
polychaete Boccardia proboscidea (Annelida: Spionidae) along the
west coast of North America. Evol. Dev. 13: 489 –503.
Palumbi, S. R. 1994. Genetic divergence, reproductive isolation, and
marine speciation. Annu. Rev. Ecol. Syst. 25: 547–572.
Perry, J. C., and B. D. Roitberg. 2006. Trophic egg laying: hypothesis
and tests. Oikos 112: 706 –714.
Radashevsky, V. I., and C. Olivares. 2005. Polydora uncinata
(Polychaeta: Spionidae) in Chile: an accidental transportation across
the Pacific. Biol. Invasions 7: 489 – 496.
Rawlinson, K. A. 2014. The diversity, development and evolution of
polyclad flatworm larvae. Evodevo 5: doi: 10.1186/2041-9139-5-9.
Rice, S. A., and K. A. Rice. 2009. Variable modes of larval develop-
ment in the Polydora cornuta complex (Polychaeta: Spionidae) are
directly related to stored sperm availability. Zoosymposia 2: 397– 414.
Rivest, B. R. 1983. Development and the influence of nurse egg allot-
ment on hatching size in Searlesia dira (Reeve, 1846) (Prosobranchia,
Buccinidae). J. Exp. Mar. Biol. Ecol. 69: 117–241.
Rivest, B. R. 1986. Extra-embryonic nutrition in the prosobranch gas-
tropod Urosalpinx cinerea (Say, 1822). Bull. Mar. Sci. 39: 498 –505.
97
Rumrill, S. S. 1990. Natural mortality of marine invertebrate larvae.
Ophelia 323: 163–198.
SAS Institute. 1989 –2008. JMP, ver. 8.0.1. SAS Institute, Cary, NC.
Sato-Okoshi, W., and M. Takatsuka. 2001. Polydora and related
genera (Polychaeta, Spionidae) around Puerto Montt and Chiloe Island
(Chile), with description of a new species of Dipolydora. Bull. Mar.
Sci. 68: 485–503.
Sato-Okoshi, W., K. Okoshi, and J. Shaw. 2008. Polydorid species
(Polychaeta : Spionidae) in south-western Australian waters with spe-
cial reference to Polydora uncinata and Boccardia knoxi. J. Mar. Biol.
Assoc. UK 88: 491–501.
Schulze, S. R., S. A. Rice, J. L. Simon, and S. A. Karl. 2000. Evo-
lution of poecilogony and the biogeography of North American popu-
lations of the polychaete Streblospio. Evolution 54: 1247–1259.
Shanks, A. L. 2009. Pelagic larval duration and dispersal distance
revisited. Biol. Bull. 216: 373–385.
Siegel, D. A., B. P. Kinlan, B. Gaylord, and S. D. Gaines. 2003.
Langrangian descriptions of marine larval dispersion. Mar. Ecol. Prog.
Ser. 260: 83–96.
Simon, C. A., T. M. Worsfold, L. Lange, and J. Sterley. 2010. The
genus Boccardia (Polychaeta: Spionidae) associated with mollusk
shells on the south coast of South Africa. J. Mar. Biol. Assoc. UK 90:
585–598.
Smith, H. L., and G. D. Gibson. 1999. Nurse egg origin in the
polychaete Boccardia proboscidea (Spionidae). Invertebr. Reprod.
Dev. 35: 177–185.
Smith, K. E., and S. Thatje. 2013. Nurse egg consumption and intra-
capsular development in the common whelk Buccinum undatum. Hel-
gol. Mar. Res. 67: 109 –120.
Strathmann, M. F. 1987. Reproduction and Development of Marine
Invertebrates of the Northern Pacific Coast: Data and Methods for the
Study of Eggs, Embryos, and Larvae. University of Washington Press,
Seattle.
Strathmann, M. F., and R. R. Strathmann. 2006. A vermetid gastro-
pod with complex intracapsular cannibalism of nurse eggs and sibling
larvae and a high potential for invasion. Pac. Sci. 60: 97–108.
Strathmann, R. R. 1985. Feeding and nonfeeding larval development
and life-history evolution in marine invertebrates. Annu. Rev. Ecol.
Syst. 16: 339 –361.
Surugiu, V. 2012. Systematics and ecology of species of the Polydora-
complex (Polychaeta: Spionidae) of the Black Sea. Zootaxa 3518:
45– 65.
Thomsen, O., R. Collin, and A. Carrillo-Baltodano. 2014. The effects
of experimentally induced adelphophagy in gastropod embryos. PLoS
One 9: 0103366.
Thorson, G. 1946. Reproductive and larval development of Danish
marine bottom invertebrates. Meddelelser fra Kommissionen for Ha-
vundersoegelser Serie Plankton 4: 1–523.
Thorson, G. 1950. Reproduction and larval ecology of marine bottom
invertebrates. Biol. Rev. 25: 1– 45.
Todd, C. D. 1998. Larval supply and recruitment of benthic inverte-
brates: Do larvae always disperse as much as we believe? Hydrobio-
logia 375/376: 1–21.
Vendetti, J. E., C. D. Trowbridge, and P. J. Krug. 2012. Poecilogony
and population structure in Elysia pusilla (Heterobranchia: Saco-
glossa), and reproductive data for five sacoglossans that express dimor-
phisms in larval development. Integr. Comp. Biol. 52: 138 –150.
Walker, L. M. 2011. A review of the current status of the Polydora-
complex (Polychaeta: Spionidae) in Australia and a checklist of re-
corded species. Zootaxa 2751: 40 – 62.
Zakas, C., and J. P. Wares. 2012. Consequences of a poecilogonous
life history for genetic structure in coastal populations of the polychaete
Streblospio benedicti. Mol. Ecol. 21: 5447–5460.