Clapham 2003
Clapham 2003
Clapham 2003
527–544
Matthew E. Clapham,* Guy M. Narbonne, and James G. Gehling.† Department of Geological Sciences and
Geological Engineering, Queen’s University, Kingston, Ontario K7L 3N6, Canada.
E-mail: narbonne@geol.queensu.ca
*Present address: Department of Earth Sciences, University of Southern California, Los Angeles 90089-0740.
E-mail: clapham@usc.edu
†
Present address: South Australian Museum, Division of Natural Science, North Terrace, Adelaide, South
Australia 5000, Australia
TABLE 1. Summary description of stratigraphic position (below the top of the Trepassey Formation), area studied,
smallest feature preserved, number of fossils, and dominant taxa for the studied fossil surfaces. The smallest feature
preserved is a taphonomic variable quantifying the finest detail visible on each surface.
Smallest
Stratigraphic Area feature No. of
Surface position studied preserved fossils Dominant taxa
SH 210 m 47.0 m 2
3.0 mm 370 Pectinate
G 2350 m 7.05 m2 1.2 mm 162 Bradgatia, Charniodiscus
E 2365 m 104.75 m2 0.5 mm 4188 Spindle, Charniodiscus
D 2368 m 63.4 m2 1.8 mm 1488 Spindle
LMP 2600 m 14.0 m2 2.0 mm 304 Charnia A
BC 2950 m 0.71 m2 0.3 mm 106 Spindle, Charnia B
PC 22200 m 16.7 m2 2.0 mm 239 Triangle, Ivesia
tical testing and accurate characterization of mixing of separate populations and time-av-
community attributes. eraging of living with recently dead material,
The studied surfaces were subdivided with which are common in Phanerozoic shelly fos-
a meter-square grid to facilitate data collec- sil assemblages (Miller 1986; Fürsich and
tion. Every fossil was identified and its posi- Aberhan 1990; Kidwell 1993; Powell et al.
tion recorded within the grid system; each of 2002), would not have affected the communi-
these records also contains measurements of ties at Mistaken Point. Preservation of the Mis-
fossil dimensions and orientation(s). In addi- taken Point fossil assemblages as census pop-
tion, taphonomic information (e.g., bending, ulations of in situ, entirely soft-bodied, epifau-
folding, incompleteness, partial preservation nal organisms, with no evidence of spatial or
due to fracturing or ash cover) was recorded temporal taphonomic mixing, provides a
where applicable. Superpositional relation- nearly ideal situation to recreate benthic com-
ships were also noted, indicating which fossil munity paleoecology. Nektonic and plankton-
was preserved underneath in an overlapping ic animals would have had a low preservation
pair. potential so their abundance (or even pres-
The resulting database contains detailed re- ence) in the Mistaken Point biota is uncertain;
cords for hundreds to thousands of fossils on however, pelagic organisms are typically not
each surface. However, the Mistaken Point included in studies of modern slope benthos
area has been subjected to pervasive tectonic either (e.g., Grassle et al. 1975; Smith and
deformation and every bedding plane has un- Hamilton 1983; Mayer and Piepenburg 1996;
dergone significant shortening, altering ab- Gutt et al. 1999).
solute fossil positions as well as both dimen- Observed differences between communities
sion and orientation values (Seilacher 1999). are indicative of significant biotic patterns
Deformed positions, dimensions, and orien- only if the variation within a single commu-
tations were restored to their original values nity is less than that between fossil surfaces.
by mathematically removing the apparent Our qualitative observations of surfaces that
bed-parallel shortening in a process called re- crop out in several localities along the sea
trodeformation (see Wood et al. in press for a coast suggested that, although even strati-
mathematical description of the methodolo- graphically closely spaced surfaces can differ
gy). All ecological tests were conducted on the dramatically in fossil content, lateral variabil-
retrodeformed database. ity in the fossil composition of any surface is
minimal (Narbonne et al. 2001). The amount
How Good Are the Data? and effects of within-surface variation can be
All Mistaken Point taxa represent soft-bod- quantified by comparing the classic E ‘‘Yale
ied, sessile organisms. There is little evidence surface’’ at Mistaken Point (used for the de-
of postmortem transport and no evidence of tailed paleocommunity studies presented be-
any infauna (Narbonne et al. 2001), so com- low) with a small sample of the E ‘‘Queen’s
plicating taphonomic effects, such as spatial surface,’’ also at Mistaken Point but separated
EDIACARAN COMMUNITY ECOLOGY 531
from the Yale surface by a 10-m-wide expo- TABLE 2. Summary attributes of Mistaken Point fossil
sure gap, and with the same stratigraphic sur- communities.
FIGURE 3. Species-area relations for Mistaken Point communities. Species-area curves for each surface were gen-
erated by measuring the species richness of randomly placed subsamples.
er (39.7 ind/m2 on the E surface) or lower (7.9 Shannon evenness values from Mistaken Point
and 14.3 ind/m2 at SH and PC, respectively). communities span nearly the entire possible
BC is the only surface displaying significantly range of equitability, from 0.26 (SH) to 0.86
anomalous density values (149.3 ind/m2), re- (Gautam et al. 2000:). The seven communities
sulting from the abundance of small fossils. are distributed over the entire equitability
range, rather than clustering in discrete
‘‘Biomass’’ groups. Other Shannon evenness values are
Biomass, as measured in modern ecosys- 0.33 (D), 0.48 (BC), 0.55 (LMP), 0.61 (E), and
tems, cannot be directly calculated from Mis- 0.80 (PC).
taken Point communities because of uncer-
tainties as to the three-dimensional shapes Single-Species Spatial Pattern
and material properties of the Ediacaran or- Methodology. Many nearest-neighbor tech-
ganisms. Instead, two-dimensional area oc- niques for the analysis of spatial pattern exist,
cupied by the fossils was adopted as a proxy although most are derived for sample areas
for biomass, allowing approximate compari- with simple polygonal boundaries (Clark and
son of total community areal coverage be- Evans 1954; Campbell 1992, 1996). Because all
tween Mistaken Point surfaces. Although this six communities (soft-sediment deformation
is not ideal, it is a far more realistic estimate at SH did not permit the original positions of
of biomass than census counts of individuals the organisms to be measured accurately) ex-
(irrespective of size) could provide. Fossil ar- amined at Mistaken Point had irregular
eal coverage, expressed as a percentage of to- boundaries imposed by the edge of preserved
tal surface area, ranges from 3.4% (SH) to outcrop, more-sophisticated statistical tests of
12.4% (E). Three communities (G, BC, LMP) spatial pattern are required. Monte Carlo sim-
have relatively low areal coverage, at 6.2%, ulation methods are able to compensate for ir-
6.3%, and 7.2% respectively, whereas the re- regular outcrop areas by comparing the ob-
maining three communities (PC, D, E) have served community spatial pattern to the sim-
higher areal coverage (10.7%, 11.2%, and ulated random spatial patterns of populations
12.4% respectively). with identical boundaries (Coomes et al.
1999). Complete spatial randomness is simu-
Diversity and Evenness lated by a two-dimensional Poisson process,
Lowest diversity is observed on the SH sur- and the cumulative distribution of nearest-
face, which has a Shannon diversity value of neighbor distances from that random popu-
0.46. The most diverse communities are E and lation is compared with the observed distri-
G, with Shannon coefficients of 1.52 and 1.54, bution of distances from the actual population
respectively. Intermediate diversity values are (Coomes et al. 1999).
0.67 (BC), 0.70 (D), 0.87 (PC), and 1.31 (LMP). We modified this procedure slightly for the
EDIACARAN COMMUNITY ECOLOGY 533
spatial pattern. The ‘‘Frondose’’ label refers to the grouping of Charniodiscus and dusters. The minimum population size for the testing was approximately 20 fossils;
TABLE 3. Single-species spatial patterns. Significance levels of aggregation (AGG), regularity (REG), and randomness (Rand) derived from Monte Carlo simulation of
1992; Dixon 1994; Zou and Wu 1995). Monte
Rand (p 5 0.094)
Carlo methods, similar to those used by Coo-
REG (p 5 0.04)
REG (p 5 0.05)
mes et al. (1999), allow simulation of multi-
species interactions in irregular sample areas.
G
NP
NP
NP
NP
NP
NS
NS
same procedure as single-species simulation,
but extended to simulate a community with
the same number of species and the same-
sized populations as the comparison fossil
AGG (p , 0.001)
AGG (p , 0.001)
AGG (p , 0.001)
AGG (p 5 0.037)
Rand (p 5 0.10)
Rand (p 5 0.17)
Rand (p 5 0.28)
community. For each species, we recorded the
identity of each nearest neighbor and tabulat-
E
NP
NP
NP
than expected.
NP
NP
NP
NP
NP
NS
NS
NS
Rand (p 5 0.14)
NP
NP
NP
NS
NS
NS
NS
Rand (p 5 0.14)
Between-Community Variation
NP
NP
NP
NP
NP
NP
NP
NS
NS
NP
NP
NP
NP
NP
NP
NP
NS
Pectinate
Bradgatia
Triangle
FIGURE 5. Q-mode and R-mode cluster analyses of Mistaken Point communities (complete linkage, Bray-Curtis
similarity coefficient).
alteration on community composition. Small and SH). Cluster analysis clearly shows
within-community variation also suggests groupings of similar communities but does
that observed between-surface differences are not display environmentally or ecologically
not simply reflections of variability within a mediated gradients in community composi-
single community type. In addition, the uni- tion. Such trends are better displayed by or-
formity within and between all E surface sam- dination techniques that map the relationship
ples implies that spatial heterogeneity was between communities in two- or three-dimen-
minimal in Ediacaran communities at Mistak- sional space (Clarke 1993). We chose nonmet-
en Point, even at kilometer-scale, in contrast to ric multidimensional scaling (MDS) to com-
well-developed patchiness in many modern pare community similarity because it is a mul-
slope communities (Grassle et al. 1975; Smith tivariate ordination technique that does not re-
and Hamilton 1983; Vetter and Dayton 1999). quire the data to be normally distributed,
Minimal within-surface variation suggests making it especially suited to analysis of com-
that observed compositional differences be- munity abundance data (Field et al. 1982;
tween Mistaken Point communities are signif- Clarke 1993). MDS ordination was performed
icant and reflect the ecological or environmen- with the PC-ORD software package (McCune
tal processes that structured those communi- and Mefford 1999) and the results are dis-
ties. Q- and R-mode cluster analysis (Fig. 5) played in Figure 6. Although MDS does not
reveals a fundamental division between two rigidly structure variability along the major
major groupings of communities: frond-dom- ordination axes (as in PCA, for example), com-
inated communities (G, E, LMP, and to some munity trends in ordination space may still be
extent PC) and frond-poor (typically spindle- linked to environmental or ecological vari-
or pectinate-dominated) communities (BC, D, ables through regression analysis (Clarke
536 MATTHEW E. CLAPHAM ET AL.
FIGURE 6. Nonmetric multidimensional scaling (MDS) ordination plot for Mistaken Point communities. Stress
,0.01 for three-dimensional solution. Regression lines for preservation quality (A), stratigraphic position (B), and
proposed ecological succession model (C) are shown. Length of regression line is proportional to the strength of
correlation. Shaded ellipses correspond to frond-poor (light gray) and frond-rich (dark gray) groupings from cluster
analysis (Fig. 5).
(Fig. 6). Some community parameters, espe- others are more diverse (H9 5 1.31 at LMP,
cially richness and diversity, may also have 1.54 at G).
been influenced by stratigraphic position in
the oldest communities, where the regional Ecological Succession
species pool was smaller (Caley and Schluter Important ecological factors in modern
1997; Lukaszewski et al. 1999). It should also communities include predation (Menge et al.
be noted that, although there is some corre- 1985; Seitz and Lipicus 2001), competition
lation between stratigraphic position and (Drobner et al. 1998; Menge 2000), and ecolog-
community similarity, biostratigraphic chang- ical succession (Connell 1978; Visser 1995).
es may not be the proximate cause of variation Predation can be eliminated as an important
if stratigraphic position is instead linked to a structuring mechanism for Mistaken Point
different environmental or ecological factor. communities because there is no evidence for
However, the results of cluster analysis and macropredators in any Ediacaran ecosystems
MDS ordination suggest that other environ- (Narbonne 1998). Evidence for intra- or inter-
mental or ecological variables were more im- specific competition is also limited: regular
portant in determining overall community spatial patterns are rare, as are nonrandom
structure and composition. pairwise patterns. In addition, interspecific
competition, as an isolated factor, may not be
Environmental Controls able to influence such disparate community
Taxa that inhabit modern slope communi- aspects as composition, species richness, di-
ties often display marked substrate preference versity, evenness, tiering structure, and spatial
(Mayer and Piepenburg 1996; Gutt et al. 1999). pattern.
However, all seven studied communities were Ecological succession produces orderly
living on silty bottom sediments that are in- changes in community composition, diversity,
distinguishable in thin sections, suggesting spatial pattern, and tiering. Ecological succes-
that differences in community composition sion is abundantly evident in modern ecosys-
did not result from variability in substrate tex- tems (Helm and Allen 1995; Visser 1995; Lich-
ture and/or composition. Similarly, major dif- ter 1998) and has been inferred from progres-
ferences between the seven diverse Mistaken sive upward changes in Phanerozoic shelly
Point communities do not appear related to fossil assemblages (Nicol 1962; Walker and Al-
resource levels. Four communities (BC, LMP, berstadt 1975; Copper 1988). However, time-
G, SH) have low areal coverage (‘‘biomass’’) averaging in Phanerozoic shelly assemblages
values (Table 2), implying that resource levels may have reduced temporal resolution so that
were low. In addition, the G surface is domi- ecological succession may not be resolvable in
nated by regular spatial patterns, suggesting these records; cases of ‘‘succession’’ may in-
that competition, likely for food, was intense. stead have been longer-term community re-
However, these low-‘‘biomass’’ communities placement mediated by environmental change
do not show strong similarity on the cluster (Miller 1986). Preservation of Mistaken Point
dendrogram (Fig. 5), nor does areal coverage assemblages as census populations of the ben-
appear as a significant regression trend on the thic communities has the disadvantage of pre-
MDS ordination plot. Correlations between senting them as ‘‘snapshots’’ of the living
areal coverage and species richness or diver- community rather than as a continuous record
sity, both of which should vary predictably of the accumulation of hardparts, but the ab-
with resource levels (Wright 1983; Menge et al. sence of time averaging in these entirely soft-
1985; Cosson-Sarradin et al. 1998), are also bodied communities provides suitable tem-
weak. Some low-‘‘biomass’’ communities have poral resolution to recognize ecological suc-
low to moderate species richness (BC, G, SH; cession, if present. The Ediacaran organisms
four to six species), but the LMP community at Mistaken Point inhabited a tectonically ac-
has high species richness (11 species). Like- tive basin, suggesting that the communities
wise, some of these communities have low di- may have been affected by occasional distur-
versity (H9 5 0.46 at SH, 0.67 at BC) whereas bances. These disturbances, such as turbidity
538 MATTHEW E. CLAPHAM ET AL.
currents, volcanic ashfalls, and anoxia and/or and D recording slightly later stages in suc-
reduced food supply resulting from slowing cession. The grouping of BC, D, and SH com-
or cessation of the contour current (Wood et munities as potential pioneer stages is sup-
al. in press), would have resulted in mass mor- ported by the cluster dendrogram, which, de-
tality of the local biota at random intervals spite their occurrence at different stratigraph-
during community development. ic levels, groups those communities together
If Mistaken Point communities preserve dif- as one of two fundamental clusters (Fig. 5).
ferent stages of ecological succession they Middle Succession. Mid-succession com-
should also display predictable changes in munities typically have the highest diversity
composition, diversity, spatial pattern, and ti- and evenness, as well as displaying increas-
ering structure as the conditions become op- ingly complex tiering and spatial pattern. The
timized for the growth of different organisms G surface community has high diversity and
(Walker and Alberstadt 1975; Whittaker 1993). evenness values, with complex tiering and
Species composition should change from the random multispecies spatial patterns, and is
pioneer community, characterized by low di- most consistent with mid-successional posi-
versity and evenness, limited tiering, and ran- tion. The E surface community is similar to the
dom spatial patterns (Walker and Alberstadt G community in terms of diversity and tier-
1975), through a mixed mid-successional ing, but it has lower evenness and contains
stage, characterized by highest diversity and nonrandom multispecies spatial interactions,
evenness (Walker and Alberstadt 1975), in- possibly indicating greater similarity to a later
creasing tiering complexity (Helm and Allen successional position. The PC community is
1995; Lichter 1998), and more nonrandom spa- difficult to place in the succession, owing to its
tial patterns from competitive exclusion and low stratigraphic position and corresponding
preferential colonization of empty spaces (Bel- depauperate fauna. The diversity value of 0.87
lingham 1998), to the climax community, char- is more similar to an early successional stage
acterized by high or slightly decreased diver- but the high evenness value (0.80) is more con-
sity (Walker and Alberstadt 1975; Clebsch and sistent with a mid-successional position, sim-
Busing 1989), lower evenness values from en- ilar to the G surface community.
hanced competitive exclusion (Death 1996; Late Succession. The LMP community is the
Drobner et al. 1998), and the greatest tiering most consistent with a late succession stage,
and spatial pattern complexity (Helm and Al- having both lower diversity and evenness
len 1995; Bellingham 1998; Lichter 1998). The than the potential mid-succession communi-
following section will investigate whether the ties (E, G), and nonrandom multispecies pat-
‘‘snapshots’’ represented by the Mistaken terns. The community has a unique species
Point fossil surfaces might reflect different composition, with a much greater proportion
stages in this idealized succession model. of frondose taxa, distinguishing it from the
Early Succession. Early successional (‘‘pio- less frond-rich E and G communities. The
neer’’) stages typically display low diversity LMP community also contains unique meter-
and highly uneven species abundances, with tall organisms not found on any other surfaces
minimal tiering and spatial pattern complex- in the study area (Clapham and Narbonne
ity. Community parameters measured from 2002).
BC, D, and SH communities are most similar Proposed Ecological Succession Model. The
to the values expected from early stages in fundamental division between the probable
ecological succession. Diversity is low (H9 5 mid- to late-succession PC, LMP, E, and G
0.46–0.70), as is evenness (E 5 0.26–0.48). Both communities and the pioneer-like BC, D, and
BC and D have random multispecies spatial SH communities is well supported by cluster
patterns (spatial patterning at SH could not be analysis (Fig. 5). The proposed succession pat-
studied) and tiering is present only in the D tern (earliest SH to early BC/D to middle PC/
surface community. The extremely low diver- E/G to late LMP) was examined by coding
sity and evenness at SH suggest it could rep- each stage with an integer value (1 through 4)
resent an earlier, pioneer community with BC and performing regression analysis in MDS
EDIACARAN COMMUNITY ECOLOGY 539
FIGURE 7. Approximate trends in species composition between different Mistaken Point communities. Position of
communities along horizontal scale is based on community parameters, cluster analysis, and MDS ordination and
parallels the proposed ecological succession model.
ordination space. The correlation between a dles and Bradgatia in the lower tiers, and the
priori successional stage (1 through 4) and uppermost tier was occupied by rare meter-
community variability is strong (regression tall forms such as the whip stem and Xmas
line C, Fig. 6). tree (Clapham and Narbonne 2002).
The trends in community species composi- If the communities reflected different suc-
tion are consistent with the proposed ecolog- cessional stages, trends in species composi-
ical succession model (Figs. 7, 8). The general tion (Fig. 7) should result in predictable
trend through succession appears to have changes in multispecies interactions (where
been the replacement of flat-lying organisms present) as early species were replaced by new
with upright, frondose organisms, possibly as colonizers (Bellingham 1998). If succession
sediment became increasingly stabilized by
was based on a tolerance model, in which the
microbial mat. Early successional stages (SH,
community changed through progressive dis-
BC, D) may have been dominated by pecti-
placement by species adapted to lower re-
nates and/or spindles, whereas intermediate
source levels (Miller 1986), later colonizers
stages (E, G) were characterized by abundant
should have preferentially settled in locations
Charniodiscus, dusters, and Bradgatia. The pro-
posed mid-succession PC community did not where preexisting species were less densely
contain these organisms because it predated packed. Observed nonrandom multispecies
their first appearances; however, it is domi- spatial patterns are consistent with a tolerance
nated by the triangle form, which was typical model of succession. Ostrich feathers (inter-
of other mid-succession assemblages (E sur- preted late-stage colonizers of the lower tier)
face). The potential late-stage community in the LMP community tend to be segregated
(LMP) was completely dominated by frondose from Charnia Type A, which occurs at every
organisms: small fronds such as Charnia Type stage of succession. On the E surface, mid-suc-
A and ostrich feathers largely replaced spin- cession frondose forms (Charniodiscus and
540 MATTHEW E. CLAPHAM ET AL.
FIGURE 8. Diorama illustrating idealized progression of communities during ecological succession. A, Charnia
Type B. B, Pectinate. C, Charnia Type A. D, Spindle. E, Bradgatia. F, Duster. G, Charniodiscus. H, Triangle. I, Ostrich
feather. J, Xmas tree.
dusters) are segregated from early-succession in a slope community, they can be compared
spindles. directly with modern bathyal megafaunal
Although the preservation of the Mistaken communities. Comparisons with Phanerozoic
Point communities as census populations shelly fossil assemblages are less secure be-
makes it impossible to prove conclusively that cause of time-averaging and the taphonomic
between-community differences result from bias against soft-bodied organisms inherent
ecological succession, the succession model is in those assemblages. If complex community
internally consistent and agrees with ob- structure is an inherent property of assem-
served variation in community structure. The blages of complex animals, the structure (rich-
proposed succession model is well supported ness, diversity, spatial patterning) of Mistaken
by cluster analysis and MDS ordination, and Point communities should be similar to that of
corresponds well to changes in diversity, equivalent modern communities. If, however,
evenness, spatial pattern complexity, and ti- complex communities are a feature that
ering structure. Species composition trends evolved gradually during early animal evo-
and interspecific interactions also vary pre- lution, then Mistaken Point may show lower
dictably in correspondence with the proposed species richness, lower diversity, or less de-
model. veloped spatial patterns than modern coun-
terparts.
Comparison with Modern Communities Species richness, Shannon diversity indices,
Mistaken Point fossil assemblages provide and spatial patterning from Mistaken Point
the opportunity to study the relationships be- communities were compared with values
tween the early evolution of animals and the from modern bathyal megafaunal communi-
evolution of animal ecosystems. Because com- ties. In the comparison studies of modern
munities at Mistaken Point preserve census slopes, megafauna refers to organisms visible
populations of the benthic megafauna living on photographic transects and generally in-
EDIACARAN COMMUNITY ECOLOGY 541
FIGURE 9. Comparison of species richness, average fossil density, and diversity values from Mistaken Point com-
munities (stars) with typical (filled rectangle) and extreme (solid line) range from modern slope communities of
epibenthic megafauna.
cludes all epifaunal organisms larger than Mistaken Point communities are as low as 8
0.5–3 cm, depending on photographic reso- ind/m2 and as high as 148 ind/m2 (Fig. 9), al-
lution, a size limit that is comparable with the though evidence from the E surface suggests
taphonomically controlled minimum visible that typical density values for Mistaken Point
size at Mistaken Point. Only photographic communities may have been 50–150 ind/m2,
studies were used for comparison purposes much greater than the animal density found
because trawl samples typically include infau- on modern slopes.
na, which are far more abundant than epifau- Spatial patterns, especially multispecies
na in modern settings (Grassle et al. 1975), distributions, are thought to indicate higher-
and which were absent from Ediacaran com- level community structuring including intra-
munities at Mistaken Point. and interspecies interactions. Random single-
Species richness varies greatly on the mod- species patterns are most common in some
ern slopes, with typical values between 10 and slope communities (Grassle et al. 1975),
30 species, for areas of ca. 10–100 m2, and a whereas aggregation or regularity predomi-
maximum range of 2 to 40 species (Grassle et nate in others (Smith and Hamilton 1983). Of
al. 1975; Smith and Hamilton 1983; Mayer and 211 single-species distributions examined by
Piepenburg 1996; Gutt et al. 1999). Mistaken Mayer and Piepenburg (1996), 81 showed sig-
Point communities fall within the typical nificant aggregation and the other 130 were
range of modern communities (Fig. 9), al- random. Mistaken Point communities also
though in general they are moderately spe- displayed significant levels of spatial pattern-
cies-poor with a maximum of 12 species per ing: more than one-half of single-species dis-
community. Mistaken Point communities also tributions deviate from randomness, with ag-
fit well within the observed range of Shannon gregation more common than regularity. Mul-
diversity coefficients from modern commu- tispecies spatial patterns are rare at Mistaken
nities (Fig. 9). Shannon diversity can be as low Point, with 60 of 64 pairwise interactions con-
as 0.05 (Grassle et al. 1975; Smith and Ham- forming to a random distribution, but the fre-
ilton 1983) or as high as 3.7 (Mayer and Pie- quency of multispecies patterning in the mod-
penburg 1996) but typically ranges between ern deep sea is poorly understood, making it
0.8 and 2.0 (Gutt et al. 1999), comparing well difficult to determine if Mistaken Point com-
with Mistaken Point communities (H9 5 0.5– munities are in fact less complexly structured.
1.5). Although fossil density measures at Mis- Mistaken Point communities fall within the
taken Point are strongly confounded by taph- typical range of species richness and diversity
onomic bias, density seems to be equal to or for modern marine epibenthic communities
greater than density values observed on mod- and displayed similar levels of single-species
ern slopes. Values as high as 70 ind/m2 have spatial patterning. Fossil density is consis-
been reported from modern settings (Gutt et tently higher than in modern slope commu-
al. 1999), but typical values range from 15 to nities but interspecies interactions may have
40 ind/m2 (Smith and Hamilton 1983; Mayer been less common. These results imply that
and Piepenburg 1996). Observed values in the structural organization of the oldest ani-
542 MATTHEW E. CLAPHAM ET AL.
mal communities at Mistaken Point was sim- onized rapidly at an early stage of animal evo-
ilar to community structure of modern slope lution. Mistaken Point communities are sig-
communities. nificantly richer than the deep-water biota
from northwestern Canada (Narbonne and
Conclusions Aitken 1990) and England (Ford 1999), and
Mistaken Point communities display signif- similar studies of these localities are necessary
icant between-community variation, likely re- to fully elucidate the paleoecology of Ediacar-
sulting from ecological processes with super- an slope environments. Similarly, shallow-wa-
imposed evolutionary and environmental ef- ter Ediacaran assemblages such as those in
fects. Modern communities are influenced by Australia, the White Sea, and Namibia are
a myriad of interrelated factors, including nu- markedly different in composition from those
trient levels, disturbance frequency, environ- at Mistaken Point, and quantitative studies are
mental heterogeneity, competition, predation, needed to determine how these differences af-
and ecological succession, to name a few. fected their ecological structure.
Some of these controls, such as predation, As a final conclusion of this study, it is in-
were absent from Ediacaran communities at teresting to note that the earliest complex
Mistaken Point. In addition, the homogeneous communities in the fossil record have struc-
nature of the E surface community, even at lo- tural attributes strikingly similar to those of
calities over 1 km apart, suggests that envi- modern counterparts. Species richness, organ-
ronmental heterogeneity was negligible. Mod- ism abundance, and diversity values, as well
ern communities display extreme variability as levels of spatial patterning, all fall within
over small and intermediate scales (Grassle et the norms of modern epibenthic slope com-
al. 1975; Mayer and Piepenburg 1996), in con- munities. Only interspecies interactions ap-
trast to the spatially uniform community pear less common. These community similar-
structure observed at Mistaken Point. Major ities suggest that, although the taxonomic af-
community variability is consistent with con- finities of Ediacaran organisms are unknown,
trol by ecological succession, and there is also they had many of the same ecological respons-
evidence for intraspecific competition, and es as present-day animals.
possibly limited interspecific segregation, in
several communities. The presence of these Acknowledgments
complex controls implies that Mistaken Point We thank R. Sala and M. Laflamme for help-
communities were largely structured by the ful assistance during fieldwork. Fieldwork in
same set of parameters that are active in mod- the Mistaken Point Ecological Reserve was
ern communities. carried out under Scientific Research Permits
Although this study does not constrain the granted by the Parks and Natural Areas Di-
affinities of component organisms, it does al- vision, Department of Tourism, Culture, and
low speculation on their environmental toler- Recreation, Government of Newfoundland
ances and reproductive strategies. For exam- and Labrador. Comments by L. Aarssen and
ple, the abundance of random and regular N. James and reviews by A. Miller and B. Wag-
spatial distributions in Mistaken Point com- goner greatly improved the manuscript. This
munities implies that the constituent organ- research was supported by a Natural Sciences
isms had a dispersal phase in their life cycle. and Engineering Research Council of Canada
The global distribution of Charniodiscus, Char- (NSERC) grant (to Narbonne) and by an
nia, and Hiemalora further suggests that at NSERC postgraduate scholarship (to Cla-
least some Edicaran taxa may have possessed pham).
a teleplanic larva.
The census populations at Mistaken Point Literature Cited
preserve diverse communities that inhabited Anderson, M. 1992. Spatial analysis of two-species interactions.
the deep slope during the Neoproterozoic. Oecologia 91:134–140.
Anderson, M. M., and S. Conway Morris. 1982. A review, with
The presence of a diverse slope biota at Mis- descriptions of four unusual forms, of the soft-bodied fauna
taken Point suggests that the deep sea was col- of the Conception and St. John’s groups (late-Precambrian),
EDIACARAN COMMUNITY ECOLOGY 543
Avalon Peninsula, Newfoundland. Pp. 1–8 in B. Mamet and M. Ford, T. D. 1999. The Precambrian fossils of Charnwood Forest.
J. Copeland, eds. Proceedings of the Third North American Geology Today 14:230–234.
Paleontological Convention. Université de Montreal, Montre- Fürsich, F. T., and M. Aberhan. 1990. Significance of time-aver-
al, and Geological Survey of Canada, Ottawa. aging for paleocommunity analysis. Lethaia 23:143–152.
Anderson, M. M., and S. B. Misra. 1968. Fossils found in the Pre- Gehling, J. G., and J. K. Rigby. 1996. Long expected sponges
Cambrian Conception Group of south-eastern Newfound- from the Neoproterozoic Ediacara fauna of South Australia.
land. Nature 220:680–681. Journal of Paleontology 70:185–195.
Bellingham, P. J. 1998. Shrub succession and invasibility in a Gehling, J. G., G. M. Narbonne, and M. M. Anderson. 2000. The
New Zealand montane grassland. Australian Journal of Ecol- first named Ediacaran body fossil, Aspidella terranovica. Pa-
ogy 23:562–573. laeontology 43:427–256.
Benus, A. P. 1988. Sedimentological context of a deep-water Edi- Glaessner, M. F. 1984. The dawn of animal life: a biohistorical
acaran fauna (Mistaken Point, Avalon Zone, eastern New- study. Cambridge University Press, Cambridge.
foundland). Pp. 8–9 in E. Landing, G. M. Narbonne, and P. M. Grassle, J. F., and N. J. Maciolek. 1992. Deep-sea species richness:
Myrow, eds. Trace fossils, small shelly fossils and the Precam- regional and local diversity estimates from quantitative bot-
brian-Cambrian Boundary. New York State Museum and Geo- tom samples. American Naturalist 139:313–341.
logical Survey Bulletin 463. Grassle, J. F., H. L. Sanders, R. R. Hessler, G. T. Rowe, and T.
Billings, E. 1872. Fossils in Huronian rocks. Canadian Naturalist McLellan. 1975. Pattern and zonation: a study of the bathyal
and Quarterly Journal of Science 6:478. megafauna using the research submersible Alvin. Deep-Sea
Caley, M. J., and D. Schluter. 1997. The relationship between lo- Research 22:457–481.
cal and regional diversity. Ecology 78:70–80. Gutt, J., E. Helsen, W. Arntz, and A. Buschmann. 1999. Biodi-
Campbell, D. J. 1992. Nearest-neighbour graphical analysis of versity and community structure of the mega-epibenthos in
spatial pattern and a test for competition in populations of the Magellan region (South America). Scientia Marina
singing crickets (Teleogryllus commodus). Oecologia 92:548– 63(Suppl. 1):155–170.
551. Haase, P. 2001. Can isotropy vs. anisotropy in the spatial asso-
———. 1996. Aggregation and regularity: an inclusive one- ciation of plant species reveal physical vs. biotic facilitation?
tailed nearest-neighbour analysis of small spatially patchy Journal of Vegetation Science 12:127–136.
populations. Oecologia 106:206–211. Helm, D. J., and E. B. Allen. 1995. Vegetation chronosequence
Clapham, M. E., and G. M. Narbonne. 2002. Ediacaran epifaunal near Exit Glacier, Kenai Fjords National Park, Alaska, U.S.A.
tiering. Geology 30:627–630. Arctic and Alpine Research 27:246–257.
Clark, P. J., and F. C. Evans. 1954. Distance to nearest neighbor Jenkins, R. J. F. 1992. Functional and ecological aspects of Edi-
as a measure of spatial relationships in a population. Ecology acaran assemblages. Pp. 131–176 in J. H. Jipps and P. W. Si-
35:445–453. gnor, eds. Origin and early evolution of the Metazoa. Plenum,
Clarke, K. R. 1993. Non-parametric multivariate analyses of New York.
changes in community structure. Australian Journal of Ecol- Jenkins, R. J. F., and J. G. Gehling. 1978. A review of the frond-
ogy 18:117–143. like fossils of the Ediacara assemblage. South Australian Mu-
Clebsch, E. E. C., and R. T. Busing. 1989. Secondary succession, seum Records 17:347–359.
gap dynamics, and community structure in a southern Ap- Kidwell, S. M. 1993. Patterns of time-averaging in the shallow
palachian cove forest. Ecology 70:728–735. marine fossil record. In S. M. Kidwell and A. K. Behrensmey-
Connell, J. H. 1978. Diversity in tropical rain forests and coral er, eds. Taphonomic approaches to time resolution in fossil
reefs. Science 199:1302–1310. assemblages. Short Courses in Paleontology 9:275–300.
Coomes, D. A., M. Rees, and L. Turnbull. 1999. Identifying ag- Lichter, J. 1998. Primary succession and forest development on
gregation and association in fully mapped spatial data. Ecol- coastal Lake Michigan sand dunes. Ecological Monographs
ogy 800:554–565. 68:487–510.
Copper, P. 1988. Ecological succession in Phanerozoic reef eco- Lukaszewski, Y., S. E. Arnott, and T. M. Frost. 1999. Regional
systems: is it real? Palaios 3:136–152. versus local processes in determining zooplankton commu-
Cosson-Sarradin, N., M. Sibuet, G. L. J. Paterson, and A. Van- nity composition of Little Rock Lake, Wisconsin, USA. Journal
griesheim. 1998. Polychaete diversity at tropical Atlantic of Plankton Research 21:991–1003.
deep-sea sites: environmental effects. Marine Ecology Pro- MacNaughton, R. B., G. M. Narbonne, and R. W. Dalrymple.
gress Series 165:173–185. 2000. Neoproterozoic slope deposits, Mackenzie Mountains,
Dalrymple, R. W., and G. M. Narbonne. 1996. Continental slope northwestern Canada: implications for passive-margin de-
sedimentation in the Sheepbed Formation (Neoproterozoic, velopment and Ediacaran faunal ecology. Canadian Journal of
Windermere Supergroup), Mackenzie Mountains, N.W.T. Ca- Earth Sciences 37:997–1020.
nadian Journal of Earth Sciences 33:848–862. Martin, M. W., D. V. Grazhdankin, S. A. Bowring, D. A. D. Evans,
Death, R. G. 1996. The effect of habitat stability on benthic in- M. A. Fedonkin, and J. L. Kirschvink. 2000. Age of Neopro-
vertebrate communities: the utility of species abundance dis- terozoic bilaterian body and trace fossils, White Sea, Russia:
tributions. Hydrobiologia 317:97–107. implications for metazoan evolution. Science 288:841–845.
Dixon, P. 1994. Testing spatial segregation using a nearest- Mayer, M., and D. Piepenburg. 1996. Epibenthic community
neighbor contingency table. Ecology 75:1940–1948. patterns on the continental slope off East Greenland at 758 N.
Drobner, U., J. Bibby, B. Smith, and J. B. Wilson. 1998. The re- Marine Ecology Progress Series 143:151–164.
lation between community biomass and evenness: what does McCune, B., and M. J. Mefford. 1999. PC-ORD. Multivariate
community theory predict, and can these predictions be test- Analysis of Ecological Data, Version 4.0. MjM Software De-
ed? Oikos 82:295–302. sign, Gleneden Beach, Ore.
Fedonkin, M. A. 1992. Vendian faunas and the early evolution McMenamin, M. A. S. 1986. The Garden of Ediacara. Palaios 1:
of Metazoa. Pp. 87–129 in J. H. Lipps and P. W. Signor, eds. 178–182.
Origin and early evolution of the Metazoa. Plenum, New York. Menge, B. A. 2000. Recruitment vs. postrecruitment processes
Field, J. G., K. R. Clarke, and R. M. Warwick. 1982. A practical as determinants of barnacle population abundance. Ecologi-
strategy for analysing multispecies distribution patterns. Ma- cal Monographs 70:265–288.
rine Ecology Progress Series 8:37–52. Menge, B. A., J. Lubchenco, and L. R. Ashkenas. 1985. Diversity,
544 MATTHEW E. CLAPHAM ET AL.
heterogeneity and consumer pressure in a tropical rocky in- structions of Precambrian evolution. Journal of the Geological
tertidal community. Oecologia 65:394–405. Society, London 149:607–613.
Miller, W. 1986. Paleoecology of benthic community replace- ———. 1999. Biomat-related lifestyles in the Precambrian. Pa-
ment. Lethaia 19:225–231. laios 14:86–93.
Misra, S. B. 1969. Late Precambrian (?) fossils from southeastern Seitz, R. D., and R. N. Lipicus. 2001. Variation in top-down and
Newfoundland. Geological Society of America Bulletin 80: bottom-up control of marine bivalves at differing spatial
2133–2140. scales. ICES Journal of Marine Science 58:689–699.
———. 1971. Stratigraphy and depositional history of late Pre- Smith, C. R., and S. C. Hamilton. 1983. Epibenthic megafauna
cambrian coelenterate-bearing rocks, southeastern New- of a bathyal basin off southern California: patterns of abun-
foundland. Geological Society of America Bulletin 82:979– dance, biomass, and dispersion. Deep-Sea Research 30:907–
988. 928.
———. 1981. Depositional environment of the late Precambrian Thrush, S. F. 1991. Spatial patterns in soft-bottom communities.
fossil-bearing rocks of southeastern Newfoundland, Canada. Trends in Ecology and Evolution 6:75–79.
Journal of the Geological Society of India 22:375–382. Vetter, E. W., and P. K. Dayton. 1999. Organic enrichment by
Myrow, P. M. 1995. Neoproterozoic rocks of the Newfoundland macrophyte detritus, and abundance patterns of megafaunal
Avalon Zone. Precambrian Research 73:123–136. populations in submarine canyons. Marine Ecology Progress
Narbonne, G. M. 1998. The Ediacara biota: a terminal Neopro- Series 186:137–148.
Visser, S. 1995. Ectomycorrhizal fungal succession in jack pine
terozoic experiment in the evolution of life. GSA Today 8(2):
stands following wildfire. New Phytologist 129:389–401.
1–6.
Waggoner, B. M. 1998. Interpreting the earliest Metazoan fos-
Narbonne, G. M., and J. D. Aitken. 1990. Ediacaran fossils from
sils: what can we learn? American Zoologist 38:975–982.
the Sekwi Brook area, Mackenzie Mountains, northwestern
———. 1999. Biogeographic analyses of the Ediacara biota: a
Canada. Palaeontology 33:945–980.
conflict with paleotectonic reconstructions. Paleobiology 24:
Narbonne, G. M., and J. G. Gehling. 2003. Life after Snowball:
440–458.
the oldest complex Ediacaran fossils. Geology 31:27–30.
Walker, K. R., and L. P. Alberstadt. 1975. Ecological succession
Narbonne, G. M., R. W. Dalrymple, and J. G. Gehling. 2001. Neo-
as an aspect of structure in fossil communities. Paleobiology
proterozoic fossils and environments of the Avalon Peninsula,
1:238–257.
Newfoundland. Guidebook, Trip B5, Geological Association
Whittaker, R. J. 1993. Plant population patterns in a glacier fore-
of Canada/Mineralogical Association of Canada Annual
land succession: pioneer herbs and later-colonizing shrubs.
Meeting, St. John’s, Newfoundland. Ecography 16:117–136.
Nicol, D. 1962. The biotic development of some Niagaran reefs: Wood, D. A., R. W. Dalrymple, G. M. Narbonne, J. G. Gehling,
an example of an ecological succession or sere. Journal of Pa- and M. E. Clapham. In press. Paleoenvironmental analysis of
leontology 36:172–176. the Late Neoproterozoic Mistaken Point and Trepassey For-
Powell, E. N., K. M. Parsons-Hubbard, W. R. Callender, G. M. mations, southeastern Newfoundland. Canadian Journal of
Staff, G. T. Rowe, C. E. Brett, S. E. Walker, A. Raymond, D. D. Earth Sciences 40.
Carlson, S. White, and E. A. Heise. 2002. Taphonomy on the Wright, D. H. 1983. Species-energy theory: an extention of spe-
continental shelf and slope: two-year trends, Gulf of Mexico cies-area theory. Oikos 41:496–506.
and Bahamas. Palaeogeography, Palaeoclimatology, Palaeoe- Zou, G., and H.-I. Wu. 1995. Nearest-neighbor distribution of
cology 184:1–35. interacting biological entities. Journal of Theoretical Biology
Seilacher, A. 1992. Vendobionta and Psammocorallia: lost con- 172:347–353.