Journal of Experimental Marine Biology and Ecology
                     311 (2004) 117 – 145
                                       www.elsevier.com/locate/jembe




The control of the development of a marine benthic
    community by predation on recruits
         Richard W. Osman a,*, Robert B. Whitlatch b,1
      a
      Smithsonian Environmental Research Center, P.O. Box 28, 647 Contees Wharf Road,
                    Edgewater, MA 21037, USA
b
 Department of Marine Sciences, University of Connecticut, 1080 Shennecossett Road, Groton, CT 06340, USA

      Received 23 February 2004; received in revised form 1 April 2004; accepted 8 May 2004



Abstract

  Recruitment is an important process in regulating many marine benthic communities and many
studies have examined factors controlling the dispersal and distribution of larval immigrants.
However, benthic species also have early post-settlement life-stages that are dramatically different
from adult and larval stages. Predation on these stages potentially impacts measured recruitment and
the benthic populations and communities that ultimately develop.
  We examined the consequences of post-settlement predation on 1-day-old to 1-month-old recruits
of sessile invertebrates at two field sites in southern New England. One site (Breakwater) was in a
protected area with few predators and the other (Pine Island) was < 1 km away in an open coast area
with three different predator guilds: small and large invertebrates and fish. The Breakwater site had
been dominated for >10 years by colonial and solitary ascidians. These species were absent from the
Pine Island site which was dominated by bryozoans. Our goal was to examine whether post-
settlement predation influenced the development and subsequent structure of the epifaunal
community.
  Here we examine long-term changes in community development resulting from post-settlement
predation, and contrast these results to those of earlier experiments examining the reductions in
observed recruitment by post-settlement predation. Our first long-term experiment examined natural
community development at the two sites and whether transplanted communities changed when
exposed to the different levels of predation at these sites. The communities that developed at both
sites were consistently different from each other and similar to resident communities at their
respective sites. On panels transplanted from the Breakwater to Pine Island, solitary ascidians and the
colonial ascidian, Botryllus schlosseri, suffered high mortalities on both caged and uncaged
treatments, indicative of predation by small predators that could enter cages. Some solitary ascidians


  * Corresponding author. Tel.: +1-443-482-2213; fax: +1-443-482-2380.
  E-mail addresses: Osmanr@si.edu (R.W. Osman), Robert.Whitlatch@uconn.edu (R.B. Whitlatch).
  1
   Tel.: +1-860-405-9154; fax: +1-860-405-9153.


0022-0981/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jembe.2004.05.001
118       R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

did survive inside cages and the colonial ascidian, Botrylloides violaceus, became dominant on all
transplanted treatments. On panels transplanted from Pine Island to the Breakwater, ascidians
invaded and dominated all treatments except those that were originally caged at Pine Island.
  In the second long-term experiment, natural communities were allowed to develop on panels
exposed at the Breakwater for 1, 2, 3, and 4 weeks. Each set was transplanted to three treatments at
Pine Island: open uncaged pilings, caged pilings to exclude fish and large invertebrates, and racks
suspended above the bottom to exclude all predators. When 1-week-old communities were
transplanted, after 2 – 3 weeks only bryozoans were found on the open and caged pilings, while
colonial ascidians dominated the suspended rack treatment. When older 2-week-old communities
were transplanted, colonial ascidians also became dominant in the caged piling treatment and when
3- and 4-week-old communities were transplanted colonial ascidians dominated all three treatments.
Solitary ascidians were never abundant on open pilings exposed to fish and large benthic invertebrate
predators.
  Post-settlement predator – prey interactions involved newly settled and juvenile life-stages of a
variety of prey species and many invertebrate and vertebrate predator species. The effects of these
interactions on recruitment did result in differences in the development and eventual species
composition of the communities, even though predators had little if any effect on the adults of the
prey species.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Anachis; Ascidians; Botrylloides; Botryllus; Bryozoans; Dispersal; Epifauna; Long Island Sound;
Marine benthic invertebrates; Mitrella; Predation; Recruitment




1. Introduction

  Many studies examining the contribution of recruitment to the regulation of marine
benthic communities have focused on processes controlling the dispersal and distribution
of larvae and how these newly arrived individuals interact as adults with the existing
community (e.g. Roughgarden et al., 1985; Menge and Sutherland, 1987; Underwood and
Keough, 2001 for a recent review). However, benthic species often have post-settlement
life-stages that are dramatically different from adult and/or larval stages. As such, these
stages have a unique ecology which may have a strong impact on recruitment and its
influence on community development (e.g. Thorson, 1966; Keough and Downes, 1982;
Watzin, 1983; Luckenbach, 1984; Young and Chia, 1984; Connell, 1985; McGuinness and
Davis, 1989; Stoner, 1990; Osman et al., 1990, 1992; Roegner, 1991; Olafsson et al.,
1994; Osman and Abbe, 1994; Osman and Whitlatch, 1995, 1996, 1998; Caley et al.,
1996). Their small size, if nothing else, makes them vulnerable to predators that may not
interact with adults (e.g. Thorson, 1966; Sutherland, 1974; Watzin, 1983; Young and Chia,
1984; Stoner, 1990). Given that benthic systems are often limited by recruitment, post-
settlement control may have a continuing influence on the composition of adult
communities.
  Our research has focused on a temperate subtidal marine benthic community composed
principally of sessile invertebrates which, as adults, permanently attach to hard substrates
such as cobbles, rocks, boulders, reefs, etc. These subtidal communities are often viewed
as extensions of their intertidal counterparts (Witman and Dayton, 2001). Important
        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145  119

processes within intertidal systems such as (1) periodic disturbances of the community
(e.g. Dayton, 1971; Sousa, 1979a,b, 1980), (2) strong competitive and trophic relation-
ships controlled by one or a few key species (e.g. Connell, 1961a,b; Paine, 1966; Dayton,
1971; Underwood, 1980), and (3) recruitment into open space which is variable and often
a function of large-scale hydrographic phenomena (e.g. Shanks, 1983; Wethey, 1986;
Roughgarden et al., 1988; Farrell et al., 1991; Shkedy and Roughgarden, 1997) are also
seen as important in these subtidal communities (e.g. Sutherland, 1974, Osman, 1977,
1978; Moran, 1980; Ayling, 1981; Witman, 1985). With larvae that are released into the
water and presumably carried away from a site, the recruitment of many species within
these communities is usually viewed as uncoupled from local changes in adult abundance
and more dependent on the potentially unpredictable spatial and temporal variability of
larvae produced throughout the whole region. In essence, these communities are viewed as
open systems with competition among adults controlling local species dominance,
physical disturbances and predation on various scales opening space for recruits, but the
subsequent supply of new recruits (and, perhaps new species) uncoupled from local
processes.
  Because many of the species in these systems are relatively short-lived, recruitment is a
critical component. Competition and predation may control dominance and disturbance
may establish the extent of free space, but recruitment controls which species are present
and their initial abundances. Key issues that have been foci of much research are the
processes controlling the supply of available larvae, processes controlling their distribu-
tion, and factors affecting the selection of the substrate onto which they permanently
attach. However, it is also critical to examine the degree to which post-settlement
processes influence recruitment. Such processes have been examined infrequently in
marine communities and their relative influence is largely unknown. Our previous research
(Osman and Whitlatch, 1995, 1996, 1998; Osman et al., 1990, 1992) has indicated that on
small, isolated substrates, predation on newly settled individuals has the potential to
control recruitment, regardless of larval supply. The present research is directed at
determining whether this predation can influence the development and subsequent
structure of the prey community. To do this we examined the development of epifaunal
communities under experimental conditions designed to contrast the effects of the predator
guilds. The results of these experiments are also contrasted with those of 3 –6-day
experiments that examined predation on recruits of individual prey species.


2. Methods

2.1. Study sites

  The studies were conducted at two field sites in eastern Long Island Sound (Groton,
CT, USA) at the mouth of the Poquonnock River estuary (41j19V15WN, 72j2V40WW). One
site was at the Avery Point Breakwater and the other was near Pine Island < 1 km from the
Breakwater site. Although both sites are similar in their temperature and salinity regimes
and tidal range, they differ in several important environmental parameters. Behind the
protected Breakwater the bottom quickly grades into mud, water depth is 2.5 m below
120      R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

MLW and tidal currents are generally < 15 cm sÀ 1. During the study, hard substrates at the
site were dominated by colonial and solitary ascidians and, except for green crabs,
Carcinus maenas (Linnaeus), predators were rare or absent (Rogers, 1998, pers. obs).
  The Pine Island site is dominated by large boulders (1– 2 m diameter) with much of the
remaining bottom covered by smaller rocks and cobbles with a few pockets of sand. The
site is more exposed to wind and waves and subject to strong tidal currents of f 70 cm
sÀ 1 during peak flood or ebb flows. Our experiments were conducted at a depth of 5 to 6
m below MLW. The sessile invertebrate community at this site was dominated by
bryozoans. Ascidians were extremely rare. On the other hand, both small and large
invertebrate predators as well as benthic-feeding fish were extremely common (Rogers,
1998, pers. obs.).

2.2. Background data

2.2.1. Sessile epifauna
  Our studies focused on the marine sessile invertebrate community found on hard
substrates. Previous studies have shown that on individual substrates as small as 100 cm2,
>25 species can coexist for years (e.g. Osman, 1977, 1978). The dominant taxa were often
colonial species, mostly bryozoans and ascidians. Although some colonies and individuals
survive years, most live months to 1 – 2 years. The majority of species produce
lecithotrophic larvae that live for only a few minutes to a day.
  To test whether new recruits of species missing at the Pine Island site could survive
there in the absence of predators, we exposed twelve 100-cm2 PVC panels at the
Breakwater site for 2 days after which all recruits were counted using a dissecting
microscope. These panels were then exposed at the Pine Island site on racks suspended 2
m below a surface float ( f 4 m above the bottom) to eliminate exposure to invertebrate
predators (Young, 1985). Four panels were also randomly assigned to cages (1-cm2
plastic-coated wire mesh) in order to test for effects of any fish predators that could reach
the suspended panels. Panels were collected after 5 days and recounted to determine
survival.

2.2.2. Predator guilds
  Three guilds of predators can potentially impact the epifaunal community at the two
sites: large invertebrates, small invertebrates, and benthic-feeding fish.
  The most common large invertebrate predators were the seastar, Asterias forbesi (Desor
1848), and the crabs, Libinia emarginata (Leach, 1815), Carcinus maenas, and Cancer
borealis (Stimpson, 1859). Adults of these species prey on both sessile species (partic-
ularly mussels and barnacles) and smaller motile invertebrates (e.g. Elner, 1978; Hughes
and Elner, 1979; Jubb et al., 1983; Schneider and Mann, 1991; Cohen et al., 1995; Berger,
1998). In addition, as newly recruited juveniles they also prey on a wider array of sessile
species (Cohen et al., 1995, pers. obs.).
  Epifaunal communities also harbor a great variety of small motile invertebrates
(amphipods, isopods, pycnogonids, gastropods, polychaetes), some of which are active
predators on one or more of the sessile species. Previously, we found that the very small
predaceous gastropods Mitrella lunata (Say) ( < 5 mm shell length), Anachis lafresnayi
        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145  121

(P. Fischer and Bernardi, 1856), and A. avara (Say) ( < 15 mm) are capable of consuming
>200 newly settled ascidians snailÀ 1 dayÀ 1 (Osman et al., 1990, 1992; Osman and
Whitlatch, 1995). These gastropods also prey on new recruits of a variety of other
epifaunal species including barnacles, mussels, and bryozoans, but the predation rates are
either low or the prey species escape through a rapid increase in size or the production of
external calcification.
  Finally, benthic fish, particularly the cunner Tautogolabrus adspersus (Walbaum,
1792), can be extremely common and prey on a variety of sessile species (Olla et al.,
1975; Green et al., 1984). In earlier studies (Osman et al., 1990) we found that juvenile
solitary ascidians may be particularly vulnerable to predation by cunner.
  Numbers and sizes of predators sampled on the experimental pilings (described below)
at the end of every experiment at the Pine Island site (Osman and Whitlatch, 1996, 1998)
were used to estimate temporal changes in the populations of these species at that site. To
estimate predation rates of fish (cunner) we deployed 100-cm2 panels with juveniles or
adults of the solitary ascidian Molgula manhattensis (DeKay 1843) at the Pine Island site
with a video camera mounted to record any activity of predators over a 40 – 60-min period.
Molgula were spawned in the laboratory and panels were exposed to the larvae that
developed. After larval settlement the panels were placed at the Breakwater site and
periodically ‘gardened’ to remove other species. Juvenile panels had individuals approx-
imately 2 weeks old and individuals on adult panels were 4 weeks old. Video recordings
were analyzed by counting the number of Molgula removed and the number and sizes of
fish present each minute.

2.3. Experiments

2.3.1. Predation on recruits and juveniles of individual species
  We exposed 1 –3-day-old newly settled individuals and 1 –3-week-old juveniles of
dominant species to natural levels of predation at the field sites. In 1993 and 1994 over 20
recruitment experiments were conducted. Most experiments had four piling treatments
with five pilings treatmentÀ 1 and one to four substrates pilingÀ 1. Treatments were: (1)
open, uncaged pilings, exposed to all three guilds of predators; (2) caged (1-cm2 plastic-
coated wire mesh) pilings designed to exclude larger invertebrate and fish predators but
allowing access to Mitrella and Anachis and other small predators; (3) screened (1-mm2
fiberglass mesh) pilings to exclude all predator guilds; and (4) partially screened pilings to
control for potential artifacts (e.g. changes in water flow or sedimentation) associated with
screening the pilings.
  The pilings were 75-cm-tall, 28-cm-diameter sections of PVC pipe secured upright to
weighted steel frames. All cages (including those used in screened and partially screened
treatments) were cylinders 50 cm in diameter and 85 cm tall with sealed tops and bottoms.
Cages completely enclosed individual pilings. Cages used in the partially screened
treatments had a 10-cm-wide band at their bottom that was devoid of screening (but still
covered by the 1-cm2 mesh). Thus, this treatment was designed to exclude the same
predators as the caged treatment and, in the absence of any artifacts, to produce similar
effects on prey species as the caged treatment. Because no individual experiment lasted for
more than a week, long-term artifacts resulting from caging or screening (e.g. reduced
122      R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

larval settlement) were assumed to be minimal. All cages were periodically cleaned by
divers and replaced monthly or if they needed repair.
  For each experiment, clean substrates were exposed to larvae of the chosen species in
the laboratory or the field. After exposure substrates were hung beneath a raft at the
Breakwater site until the selected species had reached its proper life-stage. Panels were
gardened periodically to remove all other species. Prior to beginning an experiment, all
non-desired species were removed from each substrate and the substrates were photo-
graphed and randomly assigned to treatments. In addition, individuals or colonies of the
chosen species were usually counted on each substrate before deployment. Experiments
ran for 3– 6 days after which substrates were retrieved, photographed, and surviving
individuals counted. Estimates of mortality and/or growth were made by comparing the
initial and final counts or by comparing photographs using computer-assisted image
analysis (NIH Image).

2.3.2. Predation effects on community development
  We conducted two 3-month experiments, one in 1993 and one in 1994, to test whether
predation on post-settlement life-stages alters long-term development and species domi-
nance within the sessile community. Both experiments used the experimental pilings. At
the Pine Island site these pilings were placed among boulders and separated from each
other by at least 3 m. At the Breakwater site the pilings were placed on the mud bottom
within 2 m of the Breakwater and separated from each other by at least 2 m. In each
experiment 100-cm2 panels attached to the pilings were used as sampling units.
  The 1993 study was conducted to measure (1) differences in community development
at the two field sites, 2) whether excluding fish and large invertebrate predators influenced
community development, and 3) whether established communities would change when
transplanted between the two sites. To accomplish this, 20 panels were deployed in late
July 1993 on caged and uncaged pilings at both sites. Tests for cage artifacts conducted at
both sites were negative for all 14 species tested (unpublished data). After 2 weeks, 10
panels of each treatment were reciprocally transplanted between sites with five panels from
each treatment transplanted to caged pilings and five transplanted to open pilings. At the
time of transplant the remaining 10 panels at each site were also removed, photographed,
and then returned to pilings at that site. All panels were periodically photographed in situ
over a 3-month period and these photographs were analyzed to determine differences in
species composition and abundance among treatments. NIH Image software was used to
measure the area covered by each recognizable taxon in each photographic image.
  The 1994 study examined how the age of the community when first exposed to
predators influenced the development of the community and species. A total of sixty 100-
cm2 panels were initially exposed in early July at the Breakwater site. After 1, 2, 3, and 4
weeks, 15 of the panels were transplanted to the Pine Island site. A panel was assigned to
one of: five uncaged, open pilings exposed to all predator guilds; five caged pilings
exposed only to the small predators that could enter through the 1-cm2 cage mesh; or five
panel racks suspended 2 –3 m above the bottom and located in sandy areas >5 m from the
main boulder area to reduce (and usually eliminate) exposure to all predators (e.g. Young,
1985). To minimize handling, the selection and assignment of panels to treatments were
both done haphazardly (‘randomized’ by being blindly chosen and assigned by different
        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145  123

groups of people). Unlike short-term experiments (Osman and Whitlatch, 1996, 1998),
pilings could not be screened without potentially influencing other environmental
parameters, particularly larval settlement. All panels were sampled weekly, transported
to the laboratory, kept in filtered running seawater, photographed, and returned to the field
in < 1 day. During the experiment, only treatment conditions were maintained, i.e.
assignment to a particular piling or rack, location on a piling (or rack), or site on the
bottom was haphazard with panels blindly selected by one person and assigned to each
treatment piling or rack by a second individual, and a separate group of divers placing the
pilings or racks at bottom sites. This experiment was concluded after approximately 3
months of exposure. Weekly photographs were analyzed using NIH image software.

2.4. Data analysis

  All data were analyzed using analysis of variance (ANOVA), either as one-way
analyses of treatment effects or as two-way analyses of treatment  site effects. In
each analysis a posteriori multiple comparisons of all pairs of means were done
using the Tukey – Kramer HSD test. When variances were not homogeneous, data
were log-transformed.
  Both colonization experiments involved the repeated analysis of the same experimental
panels over time. Our main goal was to characterize the progression over time in
differences among treatment in the abundances of the dominant taxa. We used separate
ANOVAs for each sampling period to compare these treatment differences. To compare
overall treatment effects in the 1994 experiment in which experimental panels were
analyzed 9– 11 times, we also used a repeated measures ANOVA. We excluded data for
the first sampling period of each series (pre-transplant data) in the repeated measure
analyses. Each of the four series of transplants was analyzed separately.


3. Results

3.1. Background data

3.1.1. Recruit survival in the absence of predators
  The survival of 2-day-old recruits on panels suspended f 4 m above the bottom at the
Pine Island site was generally high (Table 1). For most species the mean number of recruits
observed increased after the 5-day exposure. The increases most likely resulted from larval
settlement. Regardless, we analyzed both the change in the percentage of recruits of each
species (which included increases from settlement) and the percentage surviving in which
we assumed any additional individuals on a replicate were newly settled and that the
survival of transplanted recruits on that panel was 100%.
  Survival for all five species was >95% on caged panels (Table 1), indicating that in the
absence of predators these species can survive at the Pine Island site. In addition, we found
no difference in survival between open and caged panels for the colonial ascidian
Botrylloides violaceus (Oka, 1927), the erect bryozoan Bugula turrita (Desor, 1848),
and the serpulid polychaetes Spirorbis spp. However, we found significantly lower
                                                                                    124
                                                                                    R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145
Table 1
Differences in survival of newly settled recruits on four caged and eight uncaged, open panels on racks suspended above the bottom at the Pine Island site
Species      Treatment    Initial        F       Final         F       % Change     F       % Survival    F
Botrylloides   CAGE       10.00 F 2.04       0.11    12.00 F 1.80      12.8    124 F 15       6.4    100 F 0        1.9
         OPEN        5.63 F 1.44     p > 0.10    4.13 F 1.27     p < 0.01    78 F 11     p < 0.05    80 F 10     p > 0.10
Botryllus     CAGE       101.25 F 22.20      0.2    111.75 F 18.22      7.2    116 F 15      13.7     97 F 3       10.7
         OPEN       112.75 F 15.69    p > 0.60    51.88 F 12.88    p < 0.05    47 F 11     p < 0.01    47 F 10     p < 0.01
Bugula      CAGE       18.50 F 22.30      2.3    25.25 F 24.49      0.6    157 F 29       3.5    100 F 0        3.1
         OPEN       60.00 F 15.77    p > 0.15    48.88 F 17.32    p > 0.40    91 F 20     p > 0.05    78 F 9      p > 0.10
Cryptosula    CAGE       10.75 F 2.39      11.7     13.00 F 3.88       0.8    130 F 24       8.1    100 F 0        5.1
         OPEN       20.75 F 1.69     p < 0.01    8.75 F 2.74     p > 0.35    44 F 17     p < 0.05    51 F 15     p < 0.05
Spirorbis     CAGE       30.50 F 9.73       0.02    47.67 F 10.62      0.7    117 F 16       1.8    100 F 0        2.4
         OPEN       42.25 F 6.88     p > 0.85    37.13 F 6.51     p > 0.40    92 F 10     p > 0.20    81 F 7      p > 0.10
Means F SE are shown. Survival was calculated using 100% for all replicates in which the final number of recruits was equal to or greater than the initial number.
Spirorbis was not present on one caged panel, reducing the number of caged replicates for this species to 3.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145         125

survival on open panels for the encrusting bryozoan Cryptosula pallasiana (Moll, 1803)
and the colonial ascidian Botryllus schlosseri (Pallas), suggesting some predation by fish.
Although this experiment was compromised by settlement of new individuals, the results
do show that in the absence of predators, recruits of common epifaunal species not found
at Pine Island (e.g. Botrylloides, Botryllus) will survive in that environment. In addition,
recruits of the bryozoans Schizoporella errata (Waters, 1878) and Bowerbankia gracilis
(Leidy, 1855), the ascidian Molgula, and the barnacles Balanus spp. that were in low
abundance or found on only a few panels also survived on these panels.

3.1.2. Predator abundances
  The most abundant predator species were the gastropods Mitrella lunata, Anachis
lafresnayi, and Anachis avara. Abundances of Mitrella collected on experimental pilings
at Pine Island increased from < 50/piling in June to >400/piling in late August and then
declined in early September (Fig. 1). The combined abundances of both species of Anachis
(species-level field identifications were not always made) were much lower, more variable,
and did not change systematically over the summer. The mean sizes of both gastropods
increased in early July but then showed sharp declines in September reflecting the




Fig. 1. Mean abundances and shell lengths of the gastropods, Mitrella lunata and Anachis spp. (A. avara and A.
lafresnayi) on uncaged experimental pilings at the Pine Island field site. Data were collected each time an
experiment was recovered and reflect the numbers on the pilings that had accumulated by the end of each
deployment period of 3 – 6 days. Error bars represent standard errors.
126        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145




Fig. 2. Comparison of juvenile and adult Molgula manhattensis exposed to predation by the cunner,
Tautogolabrus adspersus, at the Pine Island field site. The number of Molgula eaten and fish present were
tabulated for each minute after the experiments were deployed at the field site. The data were collected from video
recordings of the events. The juvenile experiment lasted 54 min and the adult experiment 42 min. Interference
resulted in the loss of data for minutes 16 and 17 in the adult experiment.


recruitment of new cohorts. Thus these predators were very abundant during the late
summer and early autumn when recruitment of many epifaunal species occurred.

3.1.3. Fish predation rates
  When 144 juvenile Molgula were deployed at Pine Island only 1 remained after 54 min
(Fig. 2). Conversely, when larger adult Molgula were placed at the same location 179 of
200 individuals remained at the end of the 42-min deployment. Large numbers of cunners
were attracted to the Molgula, often five or six at a time. Initially, small 2 –4-cm fish were
observed attacking the ascidians, but larger individuals (8 –16 cm) soon followed. The
main difference between the two deployments was that fish numbers increased more
rapidly and remained high longer in the juvenile Molgula experiment. After 25 min few


Fig. 3. The mean survival of new 1 – 2-day-old recruits of seven species at the Pine Island field site. Separate
experiments were conducted for each species in 1993 and 1994. The means are plotted as a function of the total
abundance of Mitrella lunata, Anachis lafresnayi, and Anachis avara on each experimental piling. The results of
Bonferonni a posteriori tests based on one-way ANOVAs of treatment effects with experimental panels nested by
piling are shown. Treatments were: O = open, uncaged pilings, C = caged pilings (fish and large invertebrate
predators excluded), P = partially screened pilings (fish and large invertebrate predators excluded, control for
screening), and S = fully screened (all predators excluded) pilings. Error bars represent standard errors.
R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145  127
128        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

fish visited the adult Molgula and no Molgula were removed. This contrasted with the
continued predation of the juvenile Molgula, regardless of their density.

3.2. Experiments

3.2.1. Predator effects on recruits of individual species
  The survival of recruits of individual prey species was analyzed using a one-way
ANOVA of the four piling treatments (see also Osman and Whitlatch, 1996, 1998) and as a
function of the measured abundances of the predators, Mitrella and Anachis (Fig. 3). The
number of surviving recruits of the colonial ascidians Botryllus schlosseri and Diplosoma
listerianum (Milne Edwards, 1841) (previously identified by us as Diplosoma macdonaldi
(Herdman, 1886)) and the erect bryozoan Bugula turrita (weakly) varied inversely with




Fig. 4. Mean abundances in the 1993 Colonization Experiment after 1 month of exposure. Means are for four taxa
on 100-cm2 panels attached to experimental pilings at the Avery Point Breakwater and Pine Island sites. These
data represent abundances on six piling treatments at each site prior to the reciprocal transplanting of some panels
between the two field sites. The treatments are: Op = control or untransplanted panels on open, uncaged pilings,
Cg = control panels on caged pilings, OO = transplanted panels from open pilings at the source site to open pilings
at the receiving site, CC = transplanted panels from caged pilings at the source site to caged pilings at the
receiving site, OC = transplanted panels from open pilings at the source site to caged pilings at the receiving site,
and CO = transplanted panels from caged pilings at the source site to open pilings at the receiving site. Open bars
represent treatments at Pine Island and black bars represent treatments at the Breakwater. Letters indicate means
for treatments that are not significantly different. Bars without letters were not significantly different from one
another but were significantly different from those with letters. Error bars represent standard errors.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145          129

the abundances of predators on the pilings, showing high survival only on screened pilings
which excluded these small predators (Fig. 3). The survival of the solitary ascidian,
Molgula manhattensis, also varied inversely (weakly) with predator abundance, but two
other solitary ascidians, Ascidiella aspersa (Mu¨ller, 1776) and Ciona intestinalis (Lin-
naeus, 1767), showed no relationship to snail abundances. However, none of these species
survived on open pilings exposed to large invertebrates and fish. Finally, survival of non-
prey species such as the encrusting bryozoan, Cryptosula pallasiana, showed no
significant variation with predator abundance and no differences among treatments.

3.2.2. Community development differences and transplant effects
  After 1 month and prior to any transplanting, the panels at the two field sites generally
had less than 50% of their available surface covered by the sessile community. Individuals
and colonies that had recruited onto them were still fairly small. Nevertheless, panels on
pilings at the Breakwater were dominated by ascidians and those at Pine Island were
dominated by encrusting bryozoans (Fig. 4). No ascidians occurred on any of the panels on
pilings at Pine Island. Although covering less than 10% of any panel, encrusting




Fig. 5. Mean abundances in the 1993 Colonization Experiment 1 month after the reciprocal transplanting of
panels between sites (2 months after initial exposure). The symbols are the same as those in Fig. 3. Data are
shown by their destination site. Therefore, the means for four treatments transplanted from the Breakwater to Pine
Island (OO, CC, OC, and CO, cross-hatched bars) are shown at Pine Island, and those treatments transplanted
from Pine Island to the Breakwater (diagonally lined) are shown at the Breakwater. Error bars represent standard
errors.
130        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

bryozoans were significantly more abundant on the three caged piling treatments at Pine
Island than on the three open piling treatments.
  After 2 months of exposure (1 month after transplanting), control panels on caged
and open pilings at the Breakwater were still dominated by ascidians (Fig. 5). Botryllus
was more abundant on these panels than on all other treatments while solitary ascidians
were significantly more abundant on Breakwater caged control panels than on all other
treatments. Both taxa were absent or in low abundance on panels transplanted to pilings
at Pine Island with solitary ascidians only occurring on caged pilings. On the other
hand, both taxa did colonize panels transplanted from Pine Island to open and caged
pilings at the Breakwater. Botrylloides reached significantly higher abundances on two
groups of panels transplanted from the Breakwater to Pine Island. Botrylloides
abundance was not significantly different among any of the other treatments, but it
did remain absent from control panels at Pine Island. Finally, control panels on caged
pilings at Pine Island were dominated by encrusting bryozoans. Bryozoan abundances
on these panels were significantly greater than in all other treatments. Panels trans-
planted from Pine Island caged pilings to the Breakwater were also dominated by
bryozoans, and bryozoan abundances were significantly greater than in all the remaining
treatments.




Fig. 6. Mean abundances in the 1993 Colonization Experiment 2 months after the reciprocal transplanting of
panels between sites (3 months after initial exposure). The symbols and bar shading are the same as those in Fig. 5.
Error bars represent standard errors.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145         131

  After 3 months (Fig. 6), Botryllus was found only on pilings at the Breakwater and was
significantly more abundant on caged pilings. Solitary ascidians had a distribution similar
to Botryllus and were dominant on control panels on caged pilings at the Breakwater.
However, a few solitary ascidians also remained on panels transplanted to caged pilings at
Pine Island. Botrylloides was in greatest abundance on all treatments transplanted from the
Breakwater to Pine Island. It was also abundant on all other treatments except the Pine
Island control panels. Finally, the pattern of bryozoan abundance did not change
substantially from the previous month; abundances on panels originally on caged pilings
at Pine Island were significantly greater than on other treatments. The large difference in
bryozoan abundance on control panels on caged and open pilings at Pine Island did not
result from the presence of other species on the open pilings. The control panels on these
pilings had 50 – 75% of the remaining space unoccupied.




Fig. 7. The change in the mean abundances on the 1-week series panels in the 1994 Colonization Experiment.
Data are for panels transplanted to Pine Island after 1 week of exposure at the Avery Point Breakwater site. The
mean abundance F S.E. on each of three treatments is compared for each taxon. Treatments were open (O) = open,
uncaged experimental pilings, cage (C) = caged pilings, and rack (R) = racks suspended above the sea floor. Week
1 represents the initial mean abundance prior to transplanting. Significant results of repeated measures ANOVA
for each taxon are shown and time periods in which a significant difference (one-way ANOVA) was found among
the treatments are indicated by an asterisk (*). Error bars represent standard errors.
132        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

3.2.3. Variation in response to predation with initial community age
  In the 1994 experiment, all colonies and individuals on panels transplanted to the Pine
Island site after a 1-week exposure at the Breakwater site were still extremely small and
could only be identified and counted using a dissecting microscope. The total area covered
by all species was never greater than 2% of the available surface area of any panel.
However, after transplanting, striking differences among the treatments developed. The
colonial ascidians, Botryllus and Botrylloides, dominated the panels on the racks but were
absent from the caged and open pilings (Fig. 7). Although solitary ascidians never became
abundant and did not differ significantly among the treatments, they survived only on
panels suspended on racks. Botryllus abundance declined near the end of the study, but this
probably resulted from the post-reproductive loss of adult colonies. In the absence of
ascidians, encrusting bryozoans dominated both open and caged pilings. Bryozoan
abundance differed significantly between the caged and open pilings with panels on
caged pilings becoming almost completely covered by bryozoans (Fig. 7). Mean bryozoan
cover on the open pilings reached only 40% of the available surface area and did not differ




Fig. 8. The change in the mean abundances on the 2-week series panels in the 1994 Colonization Experiment.
Data are for panels transplanted to Pine Island after 2 weeks of exposure at the Avery Point Breakwater site. Week
2 represents the initial mean abundance prior to transplanting. Descriptions and symbols are the same as in Fig. 7.
Error bars represent standard errors.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145           133

significantly from that on racks. Similar to the 1993 experiment, most of the remaining
surface on open-piling replicates still remained open and devoid of epifaunal invertebrates
by the end of the experiment.
  The relationships among the three treatments in the communities that developed on
panels that were transplanted to Pine Island after a 2-week exposure at the Breakwater
were different than those seen in the 1-week transplants. Botryllus and Botrylloides
continued to be the dominant species on the suspended racks (Fig. 8). Botrylloides also
dominated the two piling treatments and was not significantly different among the
treatments. Some Botryllus were present in the caged piling treatment, but its abundance
remained significantly higher on the racks than on the other two treatments. Bryozoan
abundance also differed among the treatments, with their abundance on open pilings
(reaching almost 80% cover) being significantly different from the racks with bryozoan
abundance on caged pilings being intermediate. The abundance of bryozoans on the open
pilings also contrasted with their much lower abundance on open pilings in the 1-week
series.




Fig. 9. The change in the mean abundances on the 3-week series panels in the 1994 Colonization Experiment.
Data are for panels transplanted to Pine Island after 3 weeks of exposure at the Avery Point Breakwater site. Week
3 represents the initial mean abundance prior to transplanting. Descriptions and symbols are the same as in Fig. 7.
Error bars represent standard errors.
134        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

  Three major changes occurred on the panels transplanted after 3 weeks. First, Botryllus
as well as Botrylloides became dominant on piling treatments as well as on the suspended
racks with no significant differences among the three treatments in the abundance of either
species (Fig. 9). Secondly, solitary ascidians increased to 20% cover on the panels in the
rack treatment and were significantly more abundant than in the open piling treatments
with caged pilings being intermediate. Thirdly, colonial hydroids became abundant on the
panels on suspended racks and significantly different from the two piling treatments (Fig.
10). Finally, as in the 2-week series, encrusting bryozoans were significantly more
abundant on open pilings, probably as a consequence of the reduced abundance of
Botryllus and solitary ascidians.
  The results of the 4-week series (Fig. 11) were fairly similar to those of the 3-week
series. The major difference was the significantly greater abundance of solitary ascidians
on caged pilings than on the open pilings. In addition, Botryllus decreased to significantly
lower abundances on open pilings, contributing to significantly higher abundances of




Fig. 10. The change in the mean abundances of colonial hydroids on each of the four transplant series in the 1994
Colonization Experiment. Data are for panels transplanted to Pine Island after 1, 2, 3, and 4 weeks of exposure at
the Avery Point Breakwater site. Descriptions and symbols are the same as in Fig. 7. Error bars represent standard
errors.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145           135




Fig. 11. The change in the mean abundances on the 4-week series panels in the 1994 Colonization Experiment.
Data are for panels transplanted to Pine Island after 4 weeks of exposure at the Avery Point Breakwater site. Week
4 represents the initial mean abundance prior to transplanting. Descriptions and symbols are the same as in Fig. 7.
Error bars represent standard errors.


encrusting bryozoans in this treatment. As in the 3-week series, hydroids colonized and
dominated on the suspended racks (Fig. 10).


4. Discussion

  Taken together the results from two colonization experiments demonstrate that the
predation on recruits observed in short-term transplant experiments (Fig. 3; Osman and
Whitlatch, 1995, 1996, 1998) has predictable consequences on the communities that
develop at the two sites. In the 1993 experiment (Figs. 4 – 6), bryozoans dominated panels
that remained at the Pine Island site for the entire length of the experiment and ascidians
dominated controls at the Breakwater. When panels were transplanted from Pine Island to
the Breakwater, absent ascidians rapidly colonized the communities on them. However,
panels from caged pilings at Pine Island with higher abundances of bryozoans when
transplanted became dominated by bryozoans rather than ascidians. This suggests that
136      R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

some established encrusting bryozoans can successfully compete with the ascidians and
maintain dominance. Panels transplanted from the Breakwater to Pine Island (Figs. 4 –6)
suffered losses of Botryllus and solitary ascidians from both caged and open pilings,
indicating the effects of the small and abundant snail predators that could enter cages.
Some solitary ascidians survived on caged pilings, suggesting that larger individuals do
escape predation by smaller invertebrates but not large invertebrates and fish. The large
difference in the abundance of solitary ascidians between caged and open pilings at the
Breakwater indicates that some losses to large predators (probably green crabs) occurred at
that site. Because Botryllus is usually semelparous (Grosberg, 1988) its overall decline on
all treatments was a consequence of the post-reproductive mortality of adult colonies and
the lack of subsequent recruitment. Finally, the dominance of Botrylloides on all panels
transplanted to Pine Island is consistent with the results of recruitment experiments
(Osman and Whitlatch, 1996) in which colonies greater than 1 week old were not affected
by predators.
  The results from the 1994 experiment emphasize how predation by different predator
guilds on different life-stages of different species influences recruitment and overall
community development (Figs. 7 – 11). All of the ascidian species were vulnerable to
predation when newly settled or less than 1 week old (Fig. 7). When predation was
reduced or removed by using suspended racks, even the 1-week transplants (Fig. 7) with
the smallest and youngest prey developed an ascidian-dominated community. However, no
ascidians survived on either piling treatment and bryozoans were the dominant taxon.
When the transplanted community was older (Figs. 8,9 and 11), Botrylloides, as in the
1993 experiment, became a dominant species in all treatments. Botryllus was absent from
open pilings in the 2-week series (Fig. 8) and was rare on caged pilings. Only when the
transplanted community was 3 weeks old did this species become abundant on pilings
(Figs. 9 and 11). Bryozoan abundance reflects the differences among treatments in
ascidian abundance. On racks where both Botrylloides and Botryllus dominated, bryozoan
abundance remained low (Figs. 7– 9 and 11). It was somewhat higher in the caged
treatment where only Botrylloides was dominant and highest on open pilings where
Botrylloides abundance was lower. The bryozoans also increased in abundance as ascidian
abundance declined near the end of the study. The decline in both species of colonial
ascidians reflects their mortality after reproduction coupled with an absence of subsequent
recruitment at the Pine Island site. Finally, solitary ascidians never survived on open
pilings and only became abundant on caged pilings when the transplant community was 4
weeks old (Fig. 11). This progression of additions of species to the community is what we
would expect based on the earlier predation studies (Osman and Whitlatch, 1995, 1996). In
addition, the ascidians appeared to become part of the community on caged pilings at an
earlier age than on open pilings. This is consistent with earlier experiments that suggested
that as prey grew older they first escaped predation by the small predators and then by the
macro-predators.
  It appears that for the most part the short-term effects of predators on the survival of
early life-stages of the dominant species do have a longer term effect on the development
of the community. We did not follow the experiment into the winter and do not know
whether the treatments would have remained different. Because the predators are likely to
prevent recruitment by any of the ascidians, we would expect that all treatments would
        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145  137

become dominated by bryozoans as the transplanted ascidians died. In fact, this can be
seen as in the decline of Botryllus and Botrylloides in all three treatments. Nevertheless,
the experiments show that ascidians can thrive as part of the community at the Pine Island
site if recruits and young juveniles can escape predation, which seems unlikely unless
there are habitat refuges from the predators. It should also be noted that the hydroids that
became dominant in the 3- and 4-week series on suspended racks (Fig. 10) may have also
escaped the effects of predators, since they never became abundant in either of the piling
treatments.
  Earlier studies of subtidal marine epifaunal communities stressed the importance of
temporal and spatial variation in larval recruitment coupled with competition for space and
predation on established adults in controlling the development and structure of these
communities (e.g. Osman, 1977, 1978; Sutherland and Karlson, 1977; Karlson, 1978). Our
present study demonstrates that when examined more closely, predation on the earliest
post-settlement life-stages rather than larval supply can control recruitment completely and
the community that develops. Because post-settlement predation occurs so early in the
development of communities it was largely unnoticed in those earlier studies in which
recruitment was measured over longer time periods (e.g. monthly). Hixon et al. (2002)
have pointed to similar effects of unnoticed post-settlement predation on estimates of fish
recruitment and an increasing number of studies (e.g. Carr and Hixon, 1995; Hixon and
Carr, 1997; Cowen et al., 2000; Steele, 1997; Steele et al., 1998; Searcy and Sponaugle,
2001; Carr et al., 2002; Taylor and Hellberg, 2003) document the importance of local
control of early life-stages to recruitment and populations. In their review, Hunt and
Scheibling (1997) showed that post-settlement mortality from various causes can be
important in numerous other marine benthic communities. Thus, the post-settlement
control of recruitment and subsequent community development by predators is not unique
to our system and may actually be fairly common.
  The strong control of these New England subtidal epifaunal communities by predation
on recruits is the consequence of the cumulative influence of a number of generalist
invertebrate and vertebrate predators feeding on early post-settlement life-stages of species
that vary sharply in their susceptibility to the different predators. The results of the
experiments reported here (also see Osman and Whitlatch, 1996, 1998; Whitlatch and
Osman, 1998) as well as earlier work examining individual predator species (Osman et al.,
1990, 1992; Osman and Whitlatch, 1995) all suggest that Anachis and Mitrella have broad
diets, but they exert their greatest influence on newly settled ascidians. Likewise, the
cunner has a broad diet (Olla et al., 1975; Green et al., 1984) and in the epifaunal
communities at our sites exerts its greatest influence on recruitment by solitary ascidians.
Vulnerability to these predators not only varies among prey species but also among life-
stages of the same species.
  For most species vulnerability to predation is high immediately after metamorphosis
and attachment to the substrate. Small invertebrates, similar in size to the newly settled
prey, are the likely predators. Some prey taxa, particularly calcified ones (e.g. barnacles,
mussels, some bryozoans), remain vulnerable for only a short period of time (hours to
days). For example, in an experiment in which different densities of recruits of the
bryozoan Schizoporella errata were exposed to Mitrella, mean survival was reduced to
approximately 50% of that observed in controls (Table 2). The observed mortality
138        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

Table 2
Comparison of the mortality of recruits of the encrusting bryozoan Schizoporella errata on control surfaces and
surfaces exposed to Mitrella lunata
Treatment       % Mortality         % Ancestrula mortality       % Survivors with ancestrula
Control        15.07 F 7.76        18.13 F 8.59            95.91 F 1.70
Mitrella        52.83 F 5.85        78.45 F 6.40            37.81 F 8.51
Also shown is the percent mortality of the first zooid or ancestrula of the colonies and the percentage of surviving
colonies in which the ancestrula was alive.


resulted from predation exclusively on the first zooid or ancestrula of each Schizoporella
colony. The ancestrulae are both smaller than normal zooids and uncalcified (Waters,
1924), making them more vulnerable to predators (Harvell, 1984). Less than 50% of the
surviving colonies exposed to Mitrella predation had a living ancestrula. All observed
dead recruits were ancestrulae without other zooids. Thus for Schizoporella, there is an
escape from Mitrella predation with the formation of the second zooid of the colony,
which should make recruits vulnerable to these small predators for only a 1– 2-day
period.
  On the other hand, recruits of most colonial and solitary ascidians remain vulnerable to
small predators for much longer (usually weeks). Even for these species vulnerability to
predators can vary as a function of size. For example, we exposed 0.5-, 1.0-, and 4.7-mm-
diameter recruits of the solitary ascidian Molgula to predation by individual Anachis and
Mitrella. For the two smallest size classes, < 10% of the Molgula survived predation by
Mitrella compared to >85% survival in controls (Table 3). However, for the 4.7-mm
recruits we found >90% survival in both the control and Mitrella treatments. Molgula
survival in the Anachis treatment decreased from 63% to 36% as the size of the recruits
increased. Thus, as Molgula recruits increase in size they escape predation by Mitrella
only to suffer increased predation by the larger Anachis. The only exception to this several
week-long period of vulnerability seems to be the colonial ascidian Botrylloides which
remains vulnerable to predators for only a few days to a week. We have observed partially
eaten recruits of this bright orange species, suggesting that it may be defended chemically,
as in some other ascidians (e.g. Young and Bingham, 1987; Pisut and Pawlik, 2002;
Tarjuelo et al., 2002).
  Ultimately, most prey species reach a size at which the small predators have no
measurable effect on them. However, some species then become vulnerable to predation

Table 3
Comparison of the survival of different sized Molgula manhattensis when preyed on by Mitrella lunata and
Anachis spp.
Molgula size    % Molgula survival                  A posteriori test   F2,27   p
          Control      Mitrella     Anachis
0.53 F 0.03    99.3 F 0.5     2.7 F 1.2    61.4 F 8.6    C>A>M         90.5    < 0.0001
1.00 F 0.06    85.4 F 5.6     7.2 F 3.0    53.6 F 11.7   C>A>M         26.4    < 0.0001
4.65 F 0.20    95.0 F 1.6     94.3 F 1.4    36.2 F 7.2    C = M>A        50.0    < 0.0001
In each experiment thirty 2.5 Â 7.5-cm panels with at least 40 Molgula were exposed to no predator (control), a
single Mitrella, or a single Anachis (10 panels per treatment). After 2 days survival was estimated as the
percentage of Molgula remaining. Data were transformed (arcsine square root) before analysis.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145           139

by larger invertebrates (e.g. sea stars, crabs) and fish. Juveniles of all species of solitary
ascidians, as well as mussel recruits (Osman and Whitlatch, 1998), were found to be
vulnerable to fish predation. In addition, mussels and barnacles are known prey of sea
stars, gastropods, and crabs (e.g. Connell, 1961a; Paine, 1966; Menge, 1976, 1978; Elner,
1978; Jubb et al., 1983). Eventually, even the solitary ascidians can reach a size at which
they escape predation from fish (Osman et al., 1990). Barnacles and mussels, which can
find refuge from some predators in the intertidal zone, may remain vulnerable to some
predators in these subtidal communities.
  Most of our inferences regarding the particularly strong effects of predators on post-
settlement life-stages and thus on recruitment have resulted from observations of three
small species of gastropods. These species can be extremely abundant and their
recruitment occurs when prey recruitment is highest (Rogers, 1998; Fig. 3). There is
clearly a wide array of other similarly sized species that potentially function as predators of
the epifaunal communities. Some of these (e.g. nudibranchs) prey on particular species
(e.g. Thompson, 1964; Clark, 1975; Bloom, 1981; Lambert, 1991) while others (e.g.
flatworms, nemerteans, small crustaceans, some polychaetes) may be more general in their
prey selection (e.g. Bell, 1980; Watzin, 1983; Palmer, 1988; Ejdung and Elmgren, 1998).
In addition, new recruits of larger predators (e.g. crabs, sea stars, whelks) are often similar
in size to epifaunal recruits and may feed on them or juveniles (e.g. Hughes et al., 1992;
Cohen et al., 1995). For example, in two 1-week experiments contrasting epifaunal
recruitment onto 20 sets of paired panel surfaces, one exposed to a newly settled spider
crab, Libinia emarginata, and one acting as a control (see Osman et al., 1992; Osman and
Whitlatch, 1995 for methods) we found significantly reduced recruitment of Botryllus,
Diplosoma, Bugula, Cryptosula, and Botrylloides, but not Spirorbis or Molgula (Table 4).
These recruiting predators often eat different prey species than adults (Birkeland et al.,
1971, Barker, 1979) and even those species that are herbivorous as adults can be predators
as juveniles (Barker, 1979). Given the wide array of likely predators of epifaunal recruits
and their potential to influence community development, it is likely that our results provide
a conservative estimate of the influence of post-settlement mortality resulting from
predation.


Table 4
Comparison of the recruitment of seven epifaunal species in the presence and absence of newly recruited Libinia
emarginata
Species          Control           Libinia present        t          p
Botrylloides        1.45 F 0.21         0.78 F 0.15          2.37          0.0229
Botryllus         2.15 F 0.27         0.30 F 0.11          5.91         < 0.0001
Diplosoma         0.92 F 0.21         0.15 F 0.08          3.44          0.0014
Molgula          0.30 F 0.11         0.18 F 0.07          0.96          0.3423
Bugula           1.08 F 0.21         0.05 F 0.03          4.95         < 0.0001
Cryptosula         4.62 F 0.61         2.65 F 0.36          2.86          0.0067
Spirorbis         1.28 F 0.18         1.55 F 0.26          0.96          0.3434
Means are for two experiments conducted in September – October 1993 each with twenty 2.5 Â 7.5-cm panels
with each panel having one surface exposed to Libinia and a paired surface acting as a control. A paired t-test was
used to contrast the two treatments.
140      R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

  Ultimately, the strong effect of these predators on the community as a whole can be
seen in our colonization experiments (Figs. 4– 11). When all predators were excluded the
most vulnerable ascidians were common or dominant members of the communities. This
occurred regardless of the age of the transplanted community and implies that their
absence from sites with predators is not a consequence of their inability to survive in these
more exposed habitats (also Table 1). The presence of solitary ascidian larvae in larval
traps at Pine Island (Rogers, 1998) and small numbers of ascidian larvae in plankton
samples from this site (unpublished data) demonstrate that larvae of these species do reach
this site. The dominance of these species in the 1-week series suspended rack treatment
(Fig. 7) indicates that if these larvae settle at this site they can survive if predators are
absent.
  It remains that dominance within these communities is controlled by differences in
recruitment. However, the control of recruitment is at the post-settlement stage. This
control is so strong that the same communities develop at the same sites year after year
despite the ephemeral nature of several dominants and the almost certainty that
propagules of all species reach all the habitats. We observed clear differences in
dominant species at the Pine Island and Breakwater sites over more than 10 years,
yet many of the dominants are short-lived with individuals surviving much less than a
year. For these sessile species, larval recruitment is critical for maintaining a population
at a particular site. Given the potential of larvae to be well-dispersed and be distributed
among the sites, this long-term maintenance of distinctly different communities at the
sites which are < 1 km apart is both remarkable and a testament to the strength of the
predators in exerting strong local control on recruitment into these communities, altering
any patterns that would result from changes in the supply of larval settlers. This strong
control would seem to differentiate this subtidal system from intertidal ones that are
more closely linked to larval supply.
  Although we can document the role of predators in determining the development of
different communities at different sites, we have not determined what controls the
distribution of the predators and why they are rare at the Breakwater site. In earlier work
(Osman et al., 1990, 1992; Osman and Whitlatch, 1995) we used both Anachis and
Mitrella in experiments conducted at the Breakwater site. The snails were often held at
this site for several months, with only incidental mortality. Thus adult snails can survive
at this site, and given the large ascidian populations found there an ample food supply
exists. Although all three species produce benthic egg capsules, larvae that hatch from
these capsules still spend several weeks in the plankton (Scheltema and Scheltema,
1963; Scheltema, 1969; Thiriot-Quievreux, 1983; Rogers, 1998). We would expect
larvae to be transported to both sites and that the long-term maintenance of differences
in population densities between sites must result from post-settlement differences in
mortality.
  Another difference between the sites is the higher abundance of green crabs (Carcinus
maenas) at the Breakwater (Berger, 1998). This corresponds to previous studies that have
shown green crabs are often more abundant in protected areas (Menge, 1983, 1991;
Moksnes et al., 1998). In addition, Berger (1998) found an inverse correlation between the
abundance of green crabs and the densities of Anachis and Mitrella based on habitat type.
In laboratory experiments Carcinus readily preyed on Anachis and Mitrella (Berger,
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145        141

1998). It is possible that green crabs or other predators influence the distribution of the
small predators which in turn influence the epifaunal community.


5. Conclusions

  Our experimental studies demonstrate that post-settlement predator– prey interactions
can involve multiple early-ontogenetic life-stages of a variety of prey species and many
predator species, both invertebrate and vertebrate. There is no single dominant prey nor a
single key predator species. These dominant processes are the most likely cause of the
striking and persistent differences in species composition between sites that are inversely
correlated with the presence of predators which consume newly settled and juvenile (post-
settlement) life-stages. At sites without abundant predators, ascidians dominate adult
populations and recruitment. At sites where predators are present ascidians are rare or
absent and bryozoans dominate.
  We feel that post-settlement predation needs to be recognized as a generally important
process affecting community composition. For those systems in which recruitment is a
dominant or controlling process, post-settlement predation may often be the most
important process. In this light, the intertidal barnacle– mussel communities in which
supply often seems to control recruitment (e.g. Gaines and Roughgarden, 1985; Raimondi,
1990; Sutherland, 1990; Minchinton and Scheibling, 1991) may be viewed as an end
member of a continuum. In such harsh environments the diversity of possible prey and
predators is greatly reduced, limiting both the potential number of post-settlement
predators as well as their ability to control recruitment. However, even within intertidal
systems there have been indications that predation on recruiting life-stages can be
important (e.g. Palmer, 1990; Gosselin and Qian, 1996).


Acknowledgements

  We wish to thank A. Frese, M. Holt, B. Lussier, R. Malatesta, P. Mitchell, R. Stankelis,
and J. Siemon, E. Rogers, M. Berger, and J. Street for field and laboratory assistance and
G. Grenier for the design and construction of various field equipment. The research has
also benefited by discussions with R. Zajac, J. Stachowicz, D. Breitburg, S. Bullard, and J.
Carlton and the comments of P. Petraitis and two anonymous reviewers. This work was
supported by grants from the National Science Foundation (OCE 9101815, OCE 9123890,
OCE 9901139, and OCE 9819489), the Jessie B. Cox Charitable Trust, and National Sea
Grant. [SS]


References

Ayling, A.M., 1981. The role of biological disturbance in temperate subtidal encrusting communities. Ecology
  62, 830 – 847.
Barker, M.F., 1979. Breeding and recruitment in a population of the New Zealand starfish Stichaster australis
  (Verrill). J. Exp. Mar. Biol. Ecol. 41, 195 – 211.
142        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

Bell, S.S., 1980. Meiofauna – macrofauna interactions in a high salt marsh habitat. Ecol. Monogr. 50,
  487 – 505.
Berger, M.S., 1998. Predator – prey relations between the green crab, Carcinus maenas and three small gastro-
  pods, Mitrella lunata, Anachis lafresnayi, and Anachis avara. Senior Honors Thesis, University of Connecti-
  cut, Storrs, CT. 65 pp.
Birkeland, C., Chia, F.-S., Strathmann, R.R., 1971. Development, substratum selection, delay of metamorphosis
  and growth in the seastar, Mediaster aequalis Simpson. Biol. Bull. 141, 99 – 108.
Bloom, S.A., 1981. Specialization and noncompetitive resource partitioning among sponge-eating dorid nudi-
  branchs. Oecologia 49, 305 – 315.
Caley, M.J., Carr, M.H., Hixon, M.A., Hughes, T.P., Jones, G.P., Menge, B.A., 1996. Recruitment and the local
  dynamics of open marine populations. Annu. Rev. Ecol. Syst. 27, 477 – 500.
Clark, K.B., 1975. Nudibranch life cycles in the Northwest Atlantic and their relationship to the ecology of
  fouling communities. Helgol. Wiss. Meereunters. 27, 28 – 69.
Carr, M.H., Hixon, M.A., 1995. Predation effects on early post-settlement survivorship of coral-reef fishes. Mar.
  Ecol. Prog. Ser. 124, 31 – 42.
Carr, M.H., Anderson, T.W., Hixon, M.A., 2002. Bio-diversity, population regulation, and the stability of coral-
  reef fish communities. Proc. Natl. Acad. Sci. U. S. A. 99, 11241 – 11245.
Cohen, A.N., Carlton, J.T., Fountain, M.C., 1995. Introduction, dispersal and potential impacts of the green crab
  Carcinus maenas in San Francisco Bay, California. Mar. Biol. 122, 225 – 237.
Connell, J.H., 1961a. Effects of competition, predation by Thais lapillus, and other factors on natural populations
  of the barnacle Balanus balanoides. Ecol. Monogr. 31, 61 – 104.
Connell, J.H., 1961b. The influence of interspecific competition and other factors on the distribution of the
  barnacle Chthamalus stellatus. Ecology 42, 710 – 723.
Connell, J.H., 1985. The consequences of variation in initial settlement vs. post-settlement mortality in rocky
  intertidal communities. J. Exp. Mar. Biol. Ecol. 93, 11 – 45.
Cowen, R.K., Lwiza, K.M.M., Sponaurgle, S., Paris, C.B., Olson, D.B., 2000. Connectivity of marine popula-
  tions: open or closed? Science 287, 857 – 859.
Dayton, P.K., 1971. Competition, disturbance, and community organization: the provision and subsequent
  utilization of space in a rocky intertidal community. Ecol. Monogr. 41, 351 – 389.
Ejdung, G., Elmgren, R., 1998. Predation on newly settled bivalves by deposit-feeding amphipods: a Baltic Sea
  case study. Mar. Ecol. Prog. Ser. 168, 87 – 94.
Elner, R.W., 1978. The mechanics of predation by the shore crab, Carcinus maenas (L.), on the edible mussel,
  Mytilus edulis (L.). Oecologia 36, 333 – 334.
Farrell, T.M., Bracher, D., Roughgarden, J., 1991. Cross-shelf transport causes recruitment to intertidal popula-
  tions in central California. Limnol. Oceanogr. 36, 279 – 288.
Gaines, S.D., Roughgarden, J., 1985. Larval settlement rate: a leading determinant of structure in an ecological
  community of the marine intertidal zone. Proc. Natl. Acad. Sci. U.S.A. 82, 3707 – 3711.
Gosselin, L.A., Qian, P.-Y., 1996. Early post-settlement mortality of an intertidal barnacle: a critical period for
  survival. Mar. Ecol. Prog. Ser. 135, 69 – 75.
Green, J.M., Martel, G., Martin, D.W., 1984. Comparisons of the feeding activity and diets of male and female
  cunners, Tautogolabrus adspersus (Pisces: Labridae). Mar. Biol. 84, 7 – 11.
Grosberg, R.K., 1988. Life-history variation within a population of the colonial ascidian Botryllus schlosseri: I.
  The genetic and environmental control of seasonal variation. Evolution 42, 900 – 920.
Harvell, D., 1984. Why nudibranchs are partial predators: intracolonial variation in bryozoan palatability. Ecology
  65, 716 – 724.
Hixon, M.A., Carr, M.H., 1997. Synergistic predation, density dependence, and population regulation in marine
  fish. Science 277, 946 – 949.
Hixon, M.A., Pacala, S.W., Sandin, S.A., 2002. Population regulation: historical context and contemporary
  challenges of open vs. closed systems. Ecology 83, 1490 – 1508.
Hughes, R.N., Elner, R.W., 1979. Tactics of a predator, Carcinus maenas, and morphological responses of the
  prey, Nucella lapillus. J. Anim. Ecol. 48, 65 – 78.
Hughes, R.N., Burrows, M.T., Rogers, H.M., 1992. Ontogenetic changes in the foraging behaviour of the
  dogwhelk Nucella lapillus (L.). J. Exp. Mar. Biol. Ecol. 155, 199 – 212.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145           143

Hunt, H.L., Scheibling, R.E., 1997. Role of early post-settlement mortality in recruitment of benthic marine
  invertebrates. Mar. Ecol. Prog. Ser. 155, 269 – 301.
Jubb, C.A., Hughes, R.N., Rheinallt, T.A., 1983. Behavioural mechanisms of size-selection by crabs, Carcinus
  maenas (L.) feeding on mussels, Mytilus edulis L. J. Exp. Mar. Biol. Ecol. 66, 81 – 87.
Karlson, R., 1978. Predation and space utilization patterns in a marine epifaunal community. J. Exp. Mar. Biol.
  Ecol. 31, 225 – 239.
Keough, M.J., Downes, B.J., 1982. Recruitment of marine invertebrates: the role of active larval choices and
  early mortality. Oecologia 54, 348 – 352.
Lambert, W.J., 1991. Coexistence of hydroid-eating nudibranchs: recruitment and non-equilibrial patterns of
  occurrence. J. Molluscan. Stud. 57, 35 – 47.
Luckenbach, M.W., 1984. Settlement and early postsettlement survival in the recruitment of Mulinia lateralis
  (Bivalvia). Mar. Ecol. Prog. Ser. 17, 245 – 250.
McGuinness, K.A., Davis, A.R., 1989. Analysis and interpretation of the recruit – settler relationship. J. Exp. Mar.
  Biol. Ecol. 134, 197 – 202.
Menge, B.A., 1976. Organization of the New England rocky intertidal community: role of predation, competition,
  and environmental heterogeneity. Ecol. Monogr. 46, 355 – 393.
Menge, B.A., 1978. Predation intensity in a rocky intertidal community. Relation between predator foraging
  activity and environmental harshness. Oecologia 34, 1 – 16.
Menge, B.A., 1983. Components of predation intensity in the low zone of the New England rocky intertidal
  region. Oecologia 58, 141 – 155.
Menge, B.A., 1991. Generalizing from experiments: is predation strong or weak in the New England rocky
  intertidal? Oecologia 88, 1 – 8.
Menge, B.A., Sutherland, J.P., 1987. Community regulation: variation in disturbance, competition, and predation
  in relation to environmental stress and recruitment. Am. Natur. Mag. 130, 730 – 757.
Minchinton, T.E., Scheibling, R.E., 1991. The influence of larval supply and settlement on the population
  structure of barnacles. Ecology 72, 1867 – 1879.
Moksnes, P.-O., Pihl, L., van Montfrans, J., 1998. Predation on postlarvae and juveniles of the shore crab
  Carcinus maenas: importance of shelter, size and cannibalism. Mar. Ecol. Prog. Ser. 166, 211 – 225.
Moran, J., 1980. Natural physical disturbance and predation: their importance in structuring a marine sessile
  community. Aust. J. Ecol. 5, 193 – 200.
Olafsson, E.B., Peterson, C.H., Ambrose Jr., W.G. 1994. Does recruitment limitation structure populations and
  communities of macro-invertebrates in marine soft sediments: the relative significance of pre- and post-
  settlement processes. Oceanogr. Mar. Biol. Ann. Rev. 32, 65 – 109.
Olla, B.L., Bejda, A.J., Martin, A.D., 1975. Activity, movements, and feeding behavior of the cunner, Tautogo-
  labrus adspersus, and comparison of food habits with young tautog, Tautoga onitis, off Long Island, New
  York. Fish. Bull. 73, 895 – 900.
Osman, R.W., 1977. The establishment and development of a marine epifaunal community. Ecol. Monogr. 47,
  37 – 63.
Osman, R.W., 1978. The influence of seasonality and stability on the species equilibrium. Ecology 59, 383 – 399.
Osman, R.W., Abbe, G.R., 1994. Post-settlement factors affecting oyster recruitment in the Chesapeake Bay,
  USA. In: Dyer, K. (Ed.), Changes in Fluxes in Estuaries: Implications from Science to Management. Olsen
  and Olsen Press, Denmark, pp. 335 – 340.
Osman, R.W., Whitlatch, R.B., 1995. Predation on early ontogenetic life-stages and its effect on recruitment into
  a marine community. Mar. Ecol. Prog. Ser. 117, 111 – 126.
Osman, R.W., Whitlatch, R.B., 1996. Processes affecting newly-settled juveniles and the consequences to
  subsequent community development. Invertebr. Reprod. Dev. 30, 217 – 225.
Osman, R.W., Whitlatch, R.B., 1998. Local control of recruitment in an epifaunal community and the conse-
  quences to colonization processes. Hydrobiologia 375/376, 113 – 123.
Osman, R.W., Whitlatch, R.B., Malatesta, R.J., Zajac, R.N., 1990. Ontogenetic changes in trophic relationships
  and their effects on recruitment. In: Barnes, M., Gibson, R.N. (Eds.), Trophic Relationships in the Marine
  Environment, Proc. 24th European Mar. Biol. Symp. Aberdeen Univ. Press, Scotland, pp. 117 – 129.
Osman, R.W., Whitlatch, R.B., Malatesta, R.J., 1992. Potential role of micro-predators in determining recruitment
  into a marine community. Mar. Ecol. Prog. Ser. 83, 35 – 43.
144        R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145

Paine, R.T., 1966. Food web complexity and species diversity. Am. Natur. Mag. 100, 65 – 75.
Palmer, M.A., 1988. Epibenthic predators and marine meiofauna: separating predation, disturbance, and hydro-
  dynamic effects. Ecology 69, 1251 – 1259.
Palmer, A.R., 1990. Predator size, prey size, and the scaling of vulnerability: hatchling gastropods vs. barnacles.
  Ecology 71, 759 – 775.
Pisut, D.P., Pawlik, J.R., 2002. Anti-predatory chemical defenses of ascidians: secondary metabolites or inorganic
  acids? J. Exp. Mar. Biol. Ecol. 270, 203 – 214.
Raimondi, P.T., 1990. Patterns, mechanisms, consequences of variability in settlement and recruitment of an
  intertidal barnacle. Ecol. Monogr. 60, 283 – 309.
Roegner, G.C., 1991. Temporal analysis of the relationship between settlers and early recruits of the oyster
  Crassostrea virginica (Gmelin). J. Exp. Mar. Biol. Ecol. 151, 57 – 69.
Rogers, E.L., 1998. The autecology of three predatory gastropods—Mitrella lunata, Anachis avara, and Anachis
  lafrenayi. M.Sc. Thesis, University of Connecticut, 144 pp.
Roughgarden, J., Iwasa, Y., Baxter, C., 1985. Demographic theory for an open marine population with space-
  limited recruitment. Ecology 66, 54 – 67.
Roughgarden, J., Gaines, S., Possingham, H., 1988. Recruitment dynamics in complex life cycles. Science 241,
  1460 – 1466.
Scheltema, A.H., 1969. Pelagic larvae of New England gastropods: IV. Anachis translirata and Anachis avara
  (Columbellidae, Prosobranchia). Vie Milieu 20, 95 – 103.
Scheltema, R.S., Scheltema, A.H., 1963. Pelagic larvae of New England intertidal gastropods: II. Anachis avara.
  Hydrobiologia 22, 85 – 91.
Schneider, F.I., Mann, K.H., 1991. Species specific relationships of invertebrates to vegetation in a seagrass bed:
  II. Experiments on the importance of macrophyte shape, epiphyte cover and predation. J. Exp. Mar. Biol.
  Ecol. 145, 119 – 139.
Searcy, S.P., Sponaugle, S., 2001. Selective mortality during larval – juvenile transition in two coral reef fishes.
  Ecology 82, 2452 – 2470.
Shanks, A.L., 1983. Surface slicks associated with tidally forced internal waves may transport pelagic larvae of
  benthic invertebrates and fishes shoreward. Mar. Ecol. Prog. Ser. 13, 311 – 315.
Shkedy, Y., Roughgarden, J., 1997. Barnacle recruitment and population dynamics predicted from coastal
  upwelling. Oikos 80, 487 – 498.
Sousa, W.P., 1979a. Disturbance in marine intertidal boulder fields: the nonequilibrium maintenance of species
  diversity. Ecology 60, 1225 – 1239.
Sousa, W.P., 1979b. Experimental investigations of disturbance and ecological succession in a rocky intertidal
  algal community. Ecol. Monogr. 49, 227 – 254.
Sousa, W.P., 1980. The responses of a community to disturbance: the importance of successional age and species’
  life histories. Oecologia 45, 72 – 81.
Steele, M.A., 1997. Population regulation by post-settlement mortality in two temperate reef fishes. Oecologia,
  11264 – 11274.
Steele, M.A., Forrester, G.E., Almany, G.R., 1998. Influences of predators and conspecifics on recruitment of a
  tropical and a temperate reef fish. Mar. Ecol. Prog. Ser., 172115 – 172125.
Stoner, D.S., 1990. Recruitment of a tropical colonial ascidian: relative importance of pre-settlement vs. post-
  settlement processes. Ecology 71, 1682 – 1690.
Sutherland, J.P., 1974. Multiple stable points in natural communities. Am. Natur. Mag. 108, 859 – 873.
Sutherland, J.P., 1990. Recruitment regulates demographic variation in a tropical intertidal barnacle. Ecology 71,
  955 – 972.
Sutherland, J.P., Karlson, R.H., 1977. Development and stability of the fouling community at Beaufort, North
  Carolina. Ecol. Monogr. 47, 425 – 446.
Tarjuelo, I., Lopez-Legentil, S., Codina, M., Turon, X., 2002. Defense mechanisms of adult and larvae of colonial
  ascidians: patterns of palatability and toxicity. Mar. Ecol. Progr. Ser. 235, 103 – 115.
Taylor, M.S., Hellberg, M.E., 2003. Genetic evidence for local retention of pelagic larvae in a Caribbean fish.
  Science 299, 107 – 109.
Thiriot-Quievreux, C., 1983. Summer meroplanktonic larvae occurring off Beaufort, North Carolina. Estuaries 6,
  387 – 398.
         R.W. Osman, R.B. Whitlatch / J. Exp. Mar. Biol. Ecol. 311 (2004) 117–145          145

Thompson, T.E., 1964. Grazing and the life cycles of British nudibranchs. Symp. Br. Ecol. Soc. 4, 275 – 297.
Thorson, G., 1966. Some factors influencing the recruitment and establishment of marine benthic communities.
  Neth. J. Sea Res. 3, 267 – 293.
Underwood, A.J., 1980. The effects of grazing by gastropods and physical factors on the upper limits of
  distribution of intertidal macroalgae. Oecologia 46, 201 – 213.
Underwood, A.J., Keough, M.J., 2001. Supply-side ecology. The nature and consequences of variations in
  recruitment of intertidal organisms. In: Bertness, M.D., Gaines, S.D., Hay, M.E. (Eds.), Marine Community
  Ecology. Sinauer Associates, Sunderland, MA, pp. 183 – 200.
Waters, A.W., 1924. The ancestrula of Membranipora pilosa L., and other cheilostomatous Bryozoa. Ann. Mag.
  Nat. Hist. Ser. 14, 594 – 612.
Watzin, M.C., 1983. The effects of meiofauna on settling macrofauna: meiofauna may structure macrofaunal
  communities. Oecologia 59, 163 – 166.
Wethey, D.S., 1986. Local and regional variation in settlement and survival in the littoral barnacle Semibalanus
  balanoides (L.): patterns and consequences. In: Moore, P.G., Seed, R. (Eds.), The Ecology of Rocky Coasts.
  Columbia Univ. Press, New York, pp. 194 – 202.
Whitlatch, R.B., Osman, R.W., 1998. A new device for studying benthic invertebrate recruitment. Limnol.
  Oceanogr. 43, 516 – 523.
Witman, J.D., 1985. Refuges, biological disturbance, and rocky subtidal community structure in New England.
  Ecol. Monogr. 55, 421 – 445.
Witman, J.D., Dayton, P.K., 2001. Rocky subtidal communities. In: Bertness, M.D., Gaines, S.D., Hay, M.E.
  (Eds.), Marine Community Ecology. Sinauer Associates, Sunderland, MA, pp. 339 – 366.
Young, C.M., 1985. Abundance patterns of subtidal ascidians in the San Juan Islands, Washington, as influenced
  by food preferences of the predatory snail Fusitriton oregonensis. Mar. Biol. 84, 309 – 321.
Young, C.M., Bingham, B.L., 1987. Chemical defense and aposematic coloration in larvae of the ascidian
  Ecteinascidia turbinata. Mar. Biol., 96539 – 96544.
Young, C.M., Chia, F.-S., 1984. Microhabitat-associated variability in survival and growth of subtidal solitary
  ascidians during the first 21 days after settlement. Mar. Biol. 81, 61 – 68.