Download Recruitment of marine invertebrates - University of California, Santa

Oer
9 Springer-Verlag 1982
Oecologia (Berl) (1982) 54:348-352
Recruitment of Marine Invertebrates:
the Role of Active Larval Choices and Early Mortality
Michael J. Keough and Barbara J. Downes
Department of Biological Sciences, University of California, Santa Barbara, Ca 93106, USA
Summary. Spatial variation in the recruitment of sessile
marine invertebrates with planktonic larvae may be derived
from a number of sources: events within the plankton,
choices made by larvae at the time of settlement, and mortality of juvenile organisms after settlement, but before a
census by an observer. These sources usually are not distinguished.
A study of the recruitment of four species of sessile
invertebrates living on rock walls beneath a kelp canopy
showed that both selection of microhabitats by settling
larvae and predation by fish may be important. Two microhabitats were of interest; open, flat rock surfaces, and small
pits and crevices that act as refuges from fish predators.
The polychaete Spirorbis eximus and the cyclostome
bryozoan Tubulipora spp. showed no preference for refuges ,
but settled apparently at random on the available substrata.
Tubulipora was preyed upon heavily by fish, while Spiro~rbis
was relatively unaffected. The bryozoans Celleporaria
brunnea and Scrupocellaria bertholetti both recruited preferentially into refuges. Scrupocellaria were preyed upon, while
Celleporaria juveniles seemed unaffected. Predation by fish
modified the spatial distribution (Tubulipora), abundance
(Tubulipora), or size distribution (Scrupocellaria) of the juvenile population, or had relatively little effect (Cellepor-
aria, Spirorbis).
All of the above events occur within three weeks of
settlement. Since inferences about the effect of larval events
on the population dynamics of adult organisms are often
based on observations of the patterns of recruitment after
one or two months, they are therefore likely to be misleading.
Introduction.
The colonisation of habitats by marine organisms with
planktonic larvae involves three phases: development (including dispersal as a planktonic form), testing of a habitat
for suitability, and settlement (summarised by Underwood
1979). For sessile invertebrates, the latter phase also includes attachment to the substratum and metamorphosis.
The organism is unlikely to be detected immediately, because of small size, cryptic habitat, etc., and there is a fourth
"phase", survival until the organisms is counted by an
observer. This phase may last from hours to months (Scheltema 1974), but it is not a true life-history stage, merely
Offprint requests to." M.J. Keough
0029-8549/82/0054/0348/$01.00
a reflection of the limitations of the observer. The number
of organisms passing through the fourth phase is termed
recruitment, while the number passing to the third phase
is termed settlement. Recruitment is a composite of larval
and juvenile stages, while settlement involves only larval
stages.
It is important to distinguish between settlement and
recruitment. Non-random patterns of recruitment, such as
differences in the density of recruitment with height on the
shore (Underwood 1979) or differences in the density of
recruitment with patch size (Jackson 1977; Keough 1982a),
or with microhabitat, may have two causes: (1) differential
settlement, and (2), different probabilities of early mortality
in different parts of an organism's habitat. The first may
involve an active response by larvae at the time of settlement that may be an evolved response to patterns of mortality, while the second involves no active choice by larvae.
Failure to distinguish between these two phenomena
may lead to misleading inferences in a number of areas.
First, explanations for the spatial distributions of adult organisms have frequently neglected the importance of recruitment (Underwood and Denley 1982), and second,
many of the studies that have included recruitment in the
interdidal zone of rocky shores, have not distinguished between recruitment and settlement. This may have led to
an overestimation of the importance of interactions between
adult organisms and physical factors in limiting these distributions. The same is true of work on subtidal hard substrata; the terms recruitment and settlement are used interchangeably, when recruitment is actually measured. Patterns of recruitment have then been used to make inferences
about larval settling behaviour (e.g. Day and Osman 1982;
Dean and Hurd 1980; Jackson 1977; Osman 1977;
Schoener and Schoener 1981).
At the community level, it has been suggested for some
subtidal systems, species arriving first may be able to resist
further invasion by other species, so that the abundance
of sessile species in such communities can be explained by
measuring the colonising ability of component species (e.g.
Dean and Hurd 1980; Sutherland and Karlson 1977). It
is often considered that the ability of a species to colonise
a habitat can be measured accurately by its recruitment
rate (usually over a time period of 1-2 mo.).
In a similar way, many other studies have considered
the effect of predation in natural communities (e.g. Day
1977; Day and Osman 1982; Keough and Butler 1979;
Osman 1977; Paine 1966; Russ 1980; Sammarco 1980).
349
Most have used exclusions (cages, fences, etc.) that modify
the physical environment, such as light, water flow, sedimentation, in some way. Differences in the abundance of
taxa between controls and exclusions may be a result of
two (not exclusive) factors; the presence or absence of predators, and larval responses to two different physical regimes. These alternatives have rarely been separated (Choat
1982; Keough 1982b).
The above examples have in common the question of
how much the observed pattern of recruitment reflects
active choices by larvae, and how much it reflects mortality
subsequent to settlement. As a consequence, the implicit
assumption of many studies is that settlement can be measured with sufficient accuracy by recruitment, i.e. either
there is little mortality during the first few weeks after settlement, or the mortality processes affect all species equally,
so that their relative abundances are unchanged.
Here, we describe the patterns of recruitment for four
sessile invertebrate taxa and estimate mortality rates for
those taxa during the first three to four weeks after settlement.
The four taxa are Spirorbis eximus (Polychaeta: Serpulidae), and the bryozoans Tubulipora spp. (T. concinna and
T. tuba; Cyclostomata: Tubuliporidae) Scrupocellaria
bertholetti (Cheilostomata: Scrupocellariidae), and Celleporaria brunnea (Cheilostomata: Celliporidae). Predation
by fishes is an important source of mortality of sessile organisms in the study area (Downes& Keough unpubl, obs.),
of which the most important fish species are the garibaldi,
Hypsypops rubicundus (Pomacentridae), the rock wrasse,
HaIichoeres semicinetus (Labridae), and juvenile black surfperch, Embioticajacksoni (Embiotocidae).
The sessile organisms live attached to hard substrata,
inter alia vertical rock faces. These rock surfaces are uneven,
and often bear small pits and cracks, which may offer protection for settling larvae since they are inaccessible to fish.
Our aim was to measure the extent to which such refuges
are used by recruits, and to separate the presence of active
choice by settling larvae from subsequent mortality to yield
observed patterns of recruitment.
There are a number of ways in which larvae may
respond to the presence of such refuges, and in the absence
of in situ observations of the behaviour of the larvae, inferences about such responses can be made only from the
spatial distribution of juveniles, more specifically the proportion of juveniles that occur in refuges, relative to the
number on more exposed surfaces. A number of models
of larval behaviour can be erected that predict different
spatial patterns:
1. No searching (Dropped-egg model.) Larvae encounter
a substratum, but only test it for suitability, and do not
search extensively over the substratum. The predicted
pattern: Recruits are distributed in refuges and on exposed
surfaces in proportion to the surface areas of substratum
in the plane of the substratum surface. In this study, the
ratio of cross-sectional surface area of refuges to exposed
surfaces was 2 : 98.
2. Searching, but no active choice (Ping-pong ball model). Larvae encounter a substratum, and search over it, but
settle at random on any surface that is suitable, i.e. no
choice of microhabitats. The predicted pattern : Recruits are
distributed in proportion to the total surface area of refuges, compared to exposed surfaces. In this case, the ratio
is 8:92.
3. Searching and active choice of microhabitat. Larvae
encounter the substratum, search over it, and then select
microhabitats (refuges) for settlement. The predicted
pattern." Recruits are found disproportionately in refuges.
In this case, the ratio exceeds 8:92.
Methods
The study site was characterised by north-east facing vertical rock faces beneath a canopy of Macro cystis pyrifera
at Isthmus Reef, approximately 500 m from the Catalina
Marine Science Center on Santa Catalina Island in southern
California (33~
118~
Experimental substrata were 150 m m • 150 m m unglazed clay tiles, chosen for their similarity in colour and
texture to natural rock surfaces in the area. They were
mounted flush against vertical faces at a depth of about
10 m. Refuges from fish were provided on plate surfaces
by drilling small pits, 5 mm in diameter and 5 m m deep,
on the exposed face of each tile. Twenty randomly positioned pits were drilled on each tile. Surfaces of tlhe panels
are thus termed either "pits ", or "flats".
Fish were excluded from half of the panels by placing
small cages over each experimental panel. The cages were
made from 1.5-2 m m diameter plastic-coated wire, with
mesh sizes 65 r a m •
ram. Previous observations suggested that fish avoided these meshes, and we have never
observed any fish feeding through the meshes.
Panels were placed at three experimental "sites", about
3-5 m from each other. The experiment was done twice,
in November and December 1981. Both caged and uncaged
panels were placed at each site. In November, each combination of treatment and site was replicated once, whilst
two replicates were used in December. The experiment thus
had three factors, caging, site, and time, and was analysed
by analysis of variance.
Spatial patterns of recruits were examined by testing
the observed patterns against the predictions made by the
three models of larval behaviour by log-likelihood ratio
goodness-of-fit test (Bishop et al. 1975).
Results
"Settlement" - Distribution of Recruits
in the Absence ofFish
The distribution of juvenile Tubulipora on caged panels was
consistent with the non-searching (dropped-egg) model,
while that for Spirorbis differed from the pattern predicted
by this model. The distributions of both Spirorbis and Tubulipora were in accord with models that do not invoke selection of microhabitats (Table 1).
Scrupocellaria and Celleporaria both showed patterns
of recruitment on caged panels that differed from both of
the first two models (Table 1), with disproportion ately more
recruits in the pits. This was especially so for ScrupoceIlaria,
where 88% of recruits were found in the pits.
Recruitment - Distribution of Recruits
in the Presence ofFish
Again, TubuIipora and Spirorbis showed patterns of recruitment that were consistent with the null hypothesis of no
selection of microhabitats (Table 1), but in this case, the
350
Table 1. Distribution of recruits in pits and on flat surfaces of caged (C) and uncaged (U) panels. Data were pooled across all times
and sites. Three G statistics are shown. All are log-likelihood ratio tests with df= 1. (1)Goodness-of-fit of observed distribution to
an expected ratio of 2: 98 (H 1: Ratio > 2: 98). (2) Goodness-of-fit of observed distribution to an expected ratio of 8 : 92 (H 1: Ratio > 8 : 92).
(3) Test of independence of the spatial distribution of recruits on caged and uncaged panels (2 • 2 contingency table). The models
that generate the expected spatial distributions of recruits are described on P-2. n s = P > 0 . 0 5 ; * P<0.05; ** P<0.01 ; *** P<0.00t
Spirorbis
Tubulipora spp.
Scrupocellaria
Celleporaria
C
U
C
U
C
U
C
U
Pits
Flat surfaces
47
967
47
820
20
811
22
311
66
9
26
2
20
109
10
61
(1) Goodness-of-fit
G
(2) Goodness-of-fit
G
(3) Independence
G
26.3***
35.5***
22.6***
460***
189"**
49.6***
23.0***
18.2 ns
8.8 ns
0.9 ns
280***
117"**
7.9**
3.0*
0.66 ns
47.7***
0.61 ns
Table 2, Analyses of variance for the
effects of time, site, and caging
on the numbers of recruits on the exposed
surfaces of panels,
n s = P > 0 . 0 5 ; * P<0.05; ** P<0.01
10.85"
Source
of
variation
df
Caging
Site
Times
CxS
CxT
SxT
CxSxT
Residual
1
2
1
2
1
2
2
6
Table 3. Means and standard deviations (in parentheses) of the
number of recruits per panel for caged and uncaged panels at
three sites. Data were pooled across time periods, and only recruits
on flat surfaces are included, n = 3 in all cases
Taxon
Site
Caged
Uncaged
Spirorbis
1
2
3
178 (22)
87 (88)
8 (6)
131 (58)
80 (60)
29 (33)
Tubulipora
1
2
3
34 (16)
223 (26)
116 (54)
15 (9)
92 (77)
62 (55)
Scrupocellaria
1
2
3
2.3 (1.5)
0.3 (0.6)
0.3 (0.6)
0
0.7 (1.2)
0
Celleporaria
1
2
3
12 (8)
12 (7)
5 (4)
4 (3)
9 (9)
5 (4)
distribution of Tubulipora recruits differed significantly
from the predictions of the no-searching (dropped-egg)
model, but was in accordance with the second (ping-pong
ball) model. The distribution of Spirorbis recruits was unchanged.
Celleporaria and Scrupoeellaria were still found disproportionately more often in the pits (Table 1).
0.13 ns
Spirorbis
0.07 ns
Tubulipora
Celleporaria
MS
F
MS
F
MS
F
567
28,108
2,434
1,734
1,320
728
7,405
2,138
0.27 ns
13.14"*
1.14 ns
0.81 ns
0.62 ns
0.34 ns
3.46 ns
20,808
26,274
7,168
4,925
256
3,618
522
1,698
12.25"
15.47'*
4.22 ns
2.90 ns
0.15 ns
2.13 ns
0.3i ns
46.7
43.5
103.4
20.2
34.0
8.4
7.5
55.3
0.85 ns
0.79 ns
1.87 ns
0.37 ns
0.62 ns
0.15 ns
0.14 ns
Effects offish on the Abundance of Recruits
The survival rate of juveniles was assessed by comparing
the n u m b e r of recruits on flat surfaces of caged and uncaged
panels. It is unlikely that the cages used have major effects
on recruitment, but this was checked by comparing the
n u m b e r of recruits in pits, since they are protected from
predation by fish whether in cages or not.
The n u m b e r of Spirorbis recruits per panel did not differ
between uncaged and caged panels (Table 2), although
there was a strong difference between the sites (Tables 2,
3). The n u m b e r of recruits in the pits in the two treatments
was almost identical (Table 1). Tubulipora was markedly
less a b u n d a n t on uncaged panels (Tables 2, 3), although
there was no difference in the n u m b e r in the pits (Table 1).
Celleporaria showed no marked difference with caging (Table 2), despite a difference in the n u m b e r of recruits in the
pits (Table 1). This was due to a single caged panel in December, which received thirty recruits. There was no difference between the sites for Celleporaria (Table 2).
Only two Scrupocellaria recruits were observed on exposed surfaces of uncaged panels, and analysis of variance
was not possible. Nine recruits were observed on the caged
panels, but this difference is not sufficient to reject the null
hypothesis that recruits occur equally frequently on both
types of panels (Binomial test, P = 0.065). A more sensitive
test of the effect of predation on Scrupocellaria comes from
examination of recruits in the pits. Scrupocellaria has an
351
arborescent growth form, and colonies in the pits quickly
grow out of the pits, i.e. taller than 5 mm. For the December series of panels, all colonies were categorized as
being less than or greater than 5 m m high, i.e. accessible
or not accessible to fish. On caged panels, there were 14
"tall" colonies out of 36, while on the uncaged panels,
none of the 17 colonies were higher than 5 mm. These distributions are different from each other (log-likelihood ratio
test on 2 x 2 contingency table, G=9.06, d f = l , P<0.01).
These data suggest that as colonies grow tall, the pits no
longer act as refuges, and the colonies may be picked off.
Observations in the laboratory showed that when the protruding parts of a colony were pulled, the whole colony
was removed, and so the small colonies in the pits of
uncaged panels were juvenile colonies rather than remnants
of older ones. Thus, the number of juvenile colonies did
not differ between the two treatments (22 in caged vs. 17
in uncaged). That is, there was no evidence of differential
settlement into caged areas. Although sample sizes were
small, neither of the colonies on uncaged exposed surfaces
were large, while three of nine on caged panels were large.
Discussion
The observations on uncaged panels correspond to recruitment as reported in the literature; in fact, the time before
first census by us was shorter than in many studies. The
relationship between settlement and recruitment is not
strong, nor is it constant between species. A few field studies
exist that have documented the mortality of juveniles of
a single species (Connell 1961; Denley 1981; Goodbody
1965), but these studies have attributed much of the mortality of recruits to physical factors or to competition. These
results show that there may also be substantial mortality
due to predation. Further, the small, poorly-calcified stages
of the four taxa differ in their susceptibility to predation.
The observed patterns of early mortality are also likely to
vary in time, since reproduction of many invertebrate
species is seasonal. The intensity of predation may also
vary, both seasonally (e.g. Haldorsen and Moser 1979), or
with age of fish (S. Holbrook and R. Schmitt, personal
communication). Similar changes in foraging behaviour are
known for invertebrate predators (Keough and Butler 1979;
Menge 1972; Paine 1969).
The processes shown here operate on time scales shorter
than the interval between censuses in many studies of predation (e.g. Day 1977; Day and Osman 1982; Keough and
Butler 1979; Russ 1980; Sammarco 1980). The actual response of larvae to cages has not been investigated, and
it is therefore possible that apparent effects, or lack of
effects could be due not only to the exclusion of predators,
but also to a complex interaction between larval behaviour
and predation that occurs before the first census by an
observer.
Inferring Larval Behaviour
The observed patterns of recruitment can be classified as
suggesting no selection of microhabitats (Spirorbis, Tubulipora), or suggesting varying degrees of active selection of
microhabitats. Juveniles may suffer extensive mortality during the first weeks after settlement. Predation may alter
the abundance (Tubulipora), spatial distribution (Tubulipora), or size distribution of the population of juvenile or-
ganisms (Scrupocellaria). Alternatively, species may be relatively unaffected by predation (Celleporaria, Spirorbis). The
susceptibility to predation by fish is thus not clearly related
to the inferred behaviour of larvae, since of the species
that are affected by predation, one shows strong selection
of microhabitats, while the other settles apparently at
random. A similar pattern is seen for species not affected
by predation.
Direct observations of larvae often can not be made,
and the decisions made by larvae must be inferred from
the distribution of juveniles. Such inferences are usually
made from exposed substrata, and thus may be misleading.
For Tubulipora, such observations of recruitment would
support the second (ping-pong ball) model, and would lead
to the rejection of the no-searching (dropped-egg) modeI.
If the effect of fish is removed, the reverse is true, and
the data suggest that Tubulipora larvae settle as they encounter suitable substrata.
In sum, the period immediately following settlement and
metamorphosis of sessile marine invertebrates may involve
heavy mortality, so that variation in recruitment naay be
due to a combination of planktonic events, active choices
by larvae, and subsequent mortality. Inferences about
causes of distributions of adult organisms or of patterns
of community structure that are based on observations of
recruitment may be erroneous, since they can not distinguish between these sources of variation. From the viewpoint of management of commercial populations of marine
invertebrates, the distinction is important, since a major
component due to mortality of juveniles is amenable to
experimental modification, while variations that are derived
from events in the plankton are much less manipulable.
Acknowledgements. We are grateful to R. Schmitt, A. Butler, S.
Holbrook, and S. Swarbrick for their helpful discussions and for
their comments on the manuscript. We also thank Dr. R. Given
for his help, and permission to use the facilities at Catalina Manne
Science Center.
This is publication number 68 from the Catalina Marine
Science Center.
Financial support to M.J.K. was provided by an NSF grant
to J.H. Connell.
References
Choat JH (1982) The influence of fish predation on benthic communities. Ms
Connell JH (1961) Effect of competition, predation by Thais lapillus, and other factors on natural populations of the barnacle,
Balanus balanoides. Ecol Monogr 31:61 104
Day RW (1977) Two contrasting effects of predation on species
richness in coral reef habitats. Mar Biol 44:1-5
Day RW, Osman RW (1982) Predation by Patiria miniata (Asteroidea) on bryozoans: prey diversity may depend on the mechanism of succession. Oecologia (Berl.) 51 : 300-309
Dean TA, Hurd LE (1980) Development in an estuarine fouling
community: the influence of early colonists on later arrivals.
Oecologia (Berl) 45:295-301
Denley EJ (1981) The ecology of the intertidal barnacle, Tesseropera rosea. PhD thesis, University of Sydney, Australia
Goodbody I (1965) The biology of Ascidia nigra (Savigny). III The
annual pattern of colonization. Biol Bull 129:128-133
Haldorsen L, Moser M (1979) Geographic patterns of prey utilization in two species of surfperch (Embiotocidae) Copeia
1979 : 567-572
Jackson JBC (1977) Habitat area, colonization and development
of epibenthic community structure. In" BF Keegan, PO Ceidigh
352
and PJS Boaden (eds) Biology of benthic organisms. Pergamon
Press, Oxford, p 349-358
Keough MJ (1982a) Dynamics of the epifauna of Pinna bicolor:
interactions between recruitment, predation and competition.
Submitted Ms
Keough MJ (1982b) Effects of patch size on the abundance of
sessile epibenthic invertebrates. Ms
Keough M J, Butler AJ (i 979) The importance of asteroid predators
in the organisation of a sessile community on pier pilings. Mar
Biol 51 : 167-177
Menge BA (1972) Foraging strategy of a starfish in relation to
actual prey availability and environmental predictability. Ecol
Monogr 42 :25-50
Osman RW (1977) The establishment and development of a marine
epifaunal community. Ecol Monogr 47:37-63
Paine RT (1969) The Pisaster- Tegula interaction: prey patches,
predator preference, and intertidal community strucutre. Ecology 50: 950-961
Russ GR (1980) Effects of predation by fishes, competition, and
structural complexity of the substratum on the establishment
of a marine epifaunal community. J Exp mar Biol Ecol
42: 55-69
Sammarco PW (1980) Diadema and its relationship to coral spat
mortality: grazing, competition and biological disturbance. J
Exp mar Biol Ecol 45:245-272
Scheltema RS (1974) Biological interactions determining larval settlement of marine invertebrates. Thalassia Jugoslaviea
10: 263-296
Schoener A, Schoener TW (1981) The dynamics of the species-area
relation in marine fouling systems: 1 Biological correlates of
changes in the species-area slope. A~zaNat 118:339-360
Sutherland JP, Karlson RH (1977) Development and stability of
the fouling community at Beaufort, North Carolina. Ecol
Monogr 47 :425-446
Underwood AJ (1979) Ecology of intertidal gastropods. Adv mar
Res 16:111-210
Underwood AJ, Denley EJ (1982) Paradigms, explanations, and
generalizations in models for the structure of intertidal communities on rocky shores. In: Ecological communities : Conceptual
Issues and the Evidence. Princeton University Monograph,
Princeton University Press
Received March 22, 1982