A competitive hierarchy model integrating roles of physiological competence and competitive ability does not provide a mechanistic explanation for the zonation of three intertidal Fucus species in Europe

TL;DR: Keddy's competitive hierarchy model, which has not been tested in marine systems, provides an explanation for the zonation of plant species on environmental gradients but requires empirical testing in each species assemblage, to ensure that the negative outcome reported here does not invalidate the concept in general.

Abstract: It has been clear for the last 20 yr that both competitive ability and physiological competence on the intertidal gradient of rocky shores determine the zoned distribution of fucoid seaweeds in the North Atlantic. However to this point, there has been no concept integrating these functions for seaweed vegetation in a single mechanistic explanation. Keddy's competitive hierarchy model, which has not been tested in marine systems (i.e. not on seaweed vegetation) provides an explanation for the zonation of plant species on environmental gradients. The model proposes that competitive abilities of species on a stress or resource gradient are inversely related to fundamental niche breadths. We tested 2 assumptions of the model to determine whether it provided a comprehensive explanation of zonation of 3 Fucus species on the island of Helgoland in the North Sea. The 2 assumptions translate into specific predictions for the Fucus assemblage (where F. serratus occurs on the low shore, F. spiralis on the high shore, and F. vesiculosus in between): I. competitive ranking is F. serratus > F. vesiculosus > F. spiralis; 2. fundamental niche breadth rankings are the reverse of competitive ability rankings. Pairwise competition experiments were done in the field. A highly modified substitutive design was used, in order to take into account the shortcomings of this approach. The empirically derived competitive hierarchy did not fit prediction I. Transplant experiments with adults and juveniles provided results contrary to prediction 2. Since none of the assumptions of the model fit, it cannot be used to provide a mechanistic explanation for the zonation of Fucus species on Helgoland. Unlike other important models of competition, Keddy's approach does not claim universal validity in all communities. This means that it requires empirical testing in each species assemblage, and more importantly, that the negative outcome reported here does not invalidate the concept in general.

Summary

\ competitive hierarchy

• The surveyed gradient was parallel to the shoreline of Canadian lakes.
• Only partial supporting evidence could be found (Reader et al. 1994) .

Description of the study site

• Thus, the mean angle of gradient is less than 0.5".
• In the uppermost zone of the main NE-Intertidal.
• The area surrounding these boulders is populated from ca March to September by dense stands of green ephemeral algae.
• More comprehensive descriptions of the distribution of Helgoland's intertidal species are given by Markham and Munda (1980) , Janke (1986) and some of the studies summarized by Harms (1993) .

Competition

• Crnsztses of tlutural Fucus gernzling dmsit), in the Helgolarzd intrrtidc~l zone Censuses in the Helgoland intertidal zone of subjectively "dense" stands of Fucus spp. were done in order to estimate appropriate experimental densities.
• Numbers were assigned to 20 patches of Fzrcus germlings distributed across the fucoid belt.
• Seven of them were chosen randomly and algae from 5 small areas (5 x 5 cm. coordinates with random number tables) of each were scraped from the rock and brought to the laboratory.
• Densities of samples and size class frequency distributions from subsamples were estimated.

Replacet7zent series esprrinzent

• A suspension hierarchy hypothesis, it was not necessary to test comof zygotes was poured into each tray and agitated by petitive interactions between F. spiralis and F. serrutus filling the tray with additional sterile seawater immedidirectly.
• Later in spring, extensive deterioration of receptacles occurred.

Data collection

• Before installation, the lengths of 25 unbranched randomly chosen individal Fucus juveniles (from each species in mixtures) were measured from each plot.
• The lengths and dry weights were measured for 50 juveniles of each species not used in the replacement series.
• With polynomial regression equations the lengths (L. in cm) of the juveniles on the experimental plots give an estimate of the average mass (in mg) of individual shoots and, with knowledge of the density, of the total starting biomasses of each Fucus species in each plot.
• At the end of the experiment, the remaining shoots were separated to species, censused and their dry weights measured individually.
• Dry weight was obtained by drying (at 60°C for 3 d) thalli that had been rinsed in tap water.

Statistical analyses

• Replacement series graphs were drawn by plotting final yield of each component species of a replacement series experiment against its proportion in mixtures, as explained by Khan et al. (1975) .
• Individual and relative crowding coefficients were calculated from initial dry weight of germlings and final biomass.
• Initial weight was calculated with polynomial regression equations as indicated above.
• The individual crowding coefficient K, (Khan et al. 1975) is calculated for species i as: where o, is the per plant yield in mixed culture, o, is the initial per plant mass in mixed culture, mi,, the per plant yield in the corresponding monoculture and o,, the inital per plant mass in monoculture.
• Input-output diagrams were plotted for mixtures of F. serrutlis with F. cesiculosus and for mixtures of F. cesicillosus with F. spircllis for each total density separately.

Conzparison of intruspecific and iriterspecjfic interactions in additice designs.

• In substitutive designs of competition experiments, such as the replacement series, individuals of one species are "replaced" by individuals of a second species in the different plots.
• The density of one species is held constant while that of the other is varied (Firbank and Watkinson 1990) .
• The competition experiment performed here had a replacement series configuration, but some plots can also be analysed according to the additive design.
• Prior to ANOVA, homoscedasticity of data was tested with Cochran's test.
• To obtain the ranking predicted by the competitive hierarchy hypothesis, competition should be asymmetric.

Fundamental niche breadths

• Transplants oj' laboratory gernzlirigs In.stullation and rnairztenunce of the esperiment.
• 3 tiles bearing germlings of one of the Fucus species were removed and distributed among the 3 zones on the seawall (one per zone = total of 3: repeated for each of 3 F~rcusspecies).
• Remaining algae were scraped from the tiles, censused and dry weights were determined.
• Tr-ansplants uf' ad~tlt ,field plunts I11,srallution atld nzait~tenanceof the e.~perimerzt.
• 'A' compares the interspecific effect of adding 500 F. srrrutus, 'B' the intraspecific effect of adding 500 F. crsiculo.rus. 'C' and 'D' depict substitutive comparisons between intra-and interspecific effects at medium density = 500 and high density = 1000 shoots, 10 x 10 cm, respectively.

Replacement series graplzs

• At none of the 3 densities was there a linear relationship between yield and mixture proportion.
• Yields of F. ~~esiculosus were elevated in all mixed cultures, while yields of F. serratus were depressed.
• Hence, for F. cesiculosus, the curves are totally (medium density) or partially (low and high density) convex, while those of F. serratus have a concave shape.
• F. spirulis performed poorly in most plots.

Cro~vding coejj7cienr.s

• This means that growth of F. serratu.7 was depressed in mixtures (compared with its monocultures).
• Conversely, all individual crowding coefficients except one for F. cesiculosus.
• K,. from the same mixtures were > 1, indicating that growth of F. cesiculo.s~iswas elevated in these mixtures.
• All values of K, were bigger than the corresponding K,. hence all relative crowding coefficients K,, were > 1. all K,, < 1. in concordance with analysis by replacement series graphs.

Input-output ratlo diagrunzs

• All output ratios are less than input ratios.
• Three tiles, each bearing a different Fucus species (upper right), were fixed to each Fucus zone on the seawall .
• All data points for F. cesiculosus-F. spiralis mixtures lie above the line of unit slope; hence, in all mixtures, more biomass of F. resiculo.sus was gained per initial mass at the expense of F. ,spiralis and.
• Thus, the analysis through input-output diagrams is in accord with the graphical analysis and with crowding coefficients.
• Although unrelated to the main experimental objectives.

Cornparison of intraspecific and interspecific interactions in additire designs

• Information on intra-and interspecific effect was obtained from a few plots of the replacement series experiments that fit to the design suggested by Underwood (1986) (EAD: extended additive design).
• Field-grown fucoid thalli originating from concrete blocks in the Helgoland intertidal and transplanted to the three Fuclrs zones.
• When grown together with another 250 vs another 750 F. serrcrtus.
• At the medium density, the yield of 250 F. serratus was more depressed by inter-than by intraspecific competition.

Ej'ects

• Comparisons of intra-with interspecific effects revealed a stronger reduction (assumed. since there is no control) in yield through Intraspecific competition.
• According to Shipley and Keddy's (1994) defin~tion.
• The competitive relationship between F. t.eaiculosu~and F. serratus is asymmetric, the former being the dominant species.
• The findings of data analysis in an additive design are in agreement with those of substitutive replacement series.

F.serratus:F. vesiculosus F.vesiculosus: F. spiralis

• Neither of the higher density treatments differed significantly from control mean yield (EAD, Tukey-Kramer .* > 0.05, Fig. 8c ).
• Intra-and interspecific effects were significantly different at medium density, in that yield of 250 F. cesiculosus was less when grown with 250 additional F. cesiculostrs (8.95 g) as opposed to 250 additional F. mixture proportion spiralis (21.47 g).
• Since differences in I-way ANOVA were nearly significant ( p = 0.066) and differences in TK test were close to critical differences for interspecific effects (diff. = 4.4, crit. diff. = 4.9).
• The only significant pairwise difference was from an irrelevant comparison (i.e. between treatments with different species and different total density).
• Yield of F. spiralis was reduced both by intra-and interspecific competition, the latter having the stronger effect.

Sunimary of results for competition experirnerzt

• Replacement series graphs, crowding coefficients, ratio diagrams and EAD and MASD comparisons unequivocally gave this result.
• F. .spiralis performed poorly in replacement series experiments at all three initial densities.
• The findings are not concordant with the prediction from the competitive hierarchy model (F. serratus > F. cesiculo.sus > F. spirrilis).

Transplants of laboratory gerrnlirzgs

• And total dry weight per experimental tile for each Fucus species-zone combination are shown in Table 5 .
• For all parameters considered here, F. spiralis and F. r:esiculo.szr.s clearly performed worst in the lowerlnost zone, contrary to the predictions of the competitive hierarchy model.

Tran.splants of adult jell-gron.n thalli

• Initial and final numbers of transplants of all three Fucus species are shown in Table 2 .
• Differences in percentage survival were tested with ANOVA after arcsine transformation.
• All three Fucus species showed the greatest percentage survival in the zone of F. srrra-tus and successively fewer surviving plants in the zones above this tidal level.
• The most complete data for F. serratu.r parallel those from the summer experiment.
• This may also have led to the high survival, compared with summer transplants. of F. rc~siculosus and F. .spiralis in the highest zone.

Discussion

• The authors tested whether the Fucus species on Helgoland conform to the predictions of the "competitive hierarchy hypothesis" (Keddy 1989a ).
• The authors determined the competitive ranks of Helgoland Fucus species, and their fundamental niche breadths in the section of the intertidal zone naturally populated by members of the genus.

Competitive rank

• In replacement series experiments set up in the Helgoland intertidal zone.
• The competitive exclusion of F. spiralis from the mid-intertidal zone occupied by F. cesiculosus was also shown by Schonbeck and Norton (1980) .
• The competitive dominance of F. cesiculosus over F. serratus was less pronounced than that over F. spimlis.
• Hence, the hypothesis that competitive ranks of Helgoland Fucus species can be explained with the predictions from the competitive hierarchy model must be rejected.

Assessment of replacement series as experimental design

• Initial and final numbers of shoots from concrete fragments that withstood the storm.
• Few went as far as Connolly (1986) who stated "that it [the replacement series] is usually a misleading tool for research on mixtures", and most concluded that it may be a valuable method.

Zone

• Replacement series experiments have their difficulties, as have other designs, in detecting competition.
• For this to be true, both species must achieve constant final yield, when grown alone at the density of their mixture proportion (Taylor and Aarssen 1989.
• In the present study, no tests of this kind were done explicitly.
• For each Funrs species, final densities of monospecific treatments were not significantly different, even though starting densities ranged between 100 and 1000 plants per tile.
• The assumption of similar sizes was probably met with congeneric Fucus species in this study.

Transplant experiments and fundamental niche breadths in 3 Fucus species

• A fundamental tenet of the competitive hierarchy model is that competitive abilities are negatively correlated with fundamental niche breadth, perhaps because of an inherent trade-off between ability for interference competition and ability to tolerate low resource levels (Keddy 1989a ).
• The model predictions for fundamental niche breadths are: Patterns of transplants of laboratory germlings were similar in that F. ~r~iciilo.s~is was able clearly to survive in the uppermost zone and F. srrrarus performed poorly when transplanted above its natural zone.
• Effects of high temperature and desiccation should be ameliorated on this NE-facing side of the seawall and.
• It is not appropriate in this context to compare yields among species among zones simultaneously.

Application of a "relaxed" version of the competitive hiearchy model

• The survival of F. z~esiculosusabove the zone of its natural occurrence has been observed repeatedly before (Schonbeck and Norton 1978 .
• This variant seems more realistic compared with the originally strict assumptions of sharp borders of occurrence inevitably linked with invariable competitive ability and permits dominant subordinate pairs to change rank under different environmental conditions and, hence.
• Gradient" would have been a term more appropriate, not suggesting a commonly used definition of "resource" (emphasizing the possibility of consun~ption).
• Only controlled experiments, as performed in this study, will reveal the validity of Keddy's model.
• It can be argued that the mid-intertidal zone occupied by F. cesiculosus is the central benign habitat.

Figures

Table 1 . Configurations of treatments used in replacement series. Total (N) frequencies per 10 x 10 cm. Total densities: Low = I0 000: Medium = 50 000: High = 100 000 shoots,'m2. V = Fucus re.siculo~us; S = F. srrrutus; P = F. sp~mlis.Each configuration was replicated 5-fold. The same monocultures of F. ae.sicu1o.su.s were used for S:V- and V:P-replacement series.

Fig. 4. Additional additive and substitutive comparisons ('mixed additivesubstitutive design', MASD) for selected treatments from replacement series experiments, analysable with ANOVA. A hypothetical example is presented here where the intra- and interspecific effects on the yield of an initial 250 F, t.rsiculosus germlings (symbolized as black area indicated with an arrow) is given. 'A' compares the interspecific effect of adding 500 F. srrrutus, 'B' the intraspecific effect of adding 500 F. crsiculo.rus. 'C' and 'D' depict substitutive comparisons between intra- and interspecific effects at medium density = 500 and high density = 1000 shoots, 10 x 10 cm, respectively. E, F: included as 'unplanned comparisons' in one-way ANOVA but irrelevant, since total density and species combination are changed simultaneously. For meaning of abbreviations (e.g. "SV 0:4 High Density") see Table 1. Note that germlings of 2 species are interdispersed in mixtures and not clumped in separate corners as symbolized here.

Table 4. Individual and relative crowding coefficients from replacement series field experiments with Fuczts cesiculosus and F. spiralis at 3 total densities, and in 3 mixtures. Mean values of 5 replicates. K , , and K, ,,, are individual crowding $, coefficients. K, = relative crowdng coefficient F. nesiculosus on F. spiralis. <,,= relative crowding coefficient F. ~piralison F. resiculosus. See text for explanation of coefficients.

Table 3. Individual and relative crowding coefficients from replacement series field experiments with Fucus serratus and F. " '/ Density oesiculosus at 3 total densities and 3 mixtures. Mean values of 5 replicates. K,,,,, ,, ,! and K are individual crowding + Medium + Highcoefficients. K,, = relative crowdmg coefficient F. serratus on F. nesiculosus. K,,, = relative crowding coefficient F. ~.esiculosus on F. serratus. See text for explanation of coefficients.

Fig. 12. Curves of competitive ability along an enk~ronmental gradient as predicted from the "relaxed" version of the competitive hierarchy model (Keddy 1989a). Shown at bottom are realized niches (observed in nature under the effects of competition). Two kinds of competitive dominance are distinguished: I . in the sense of the narrow competitive hierarchy model (CHM-dominance). and 2. actual dominance (ACT-dominance) in nature, in terms of the more relaxed version of the model (see text for further explanation of 2 model versions). The borders of realized distributions derive from the intersection points of competitive ability curves of neighbouring species. Thus. a species:species boundary is set by neither the final exhaustion of tolerance of low resource levels (that would coincide with zero competitive ability), nor by competitive exclusion of other species by a CHM-dominant species through its whole fundamental niche. Black sections of gradient indicate where occupying species are CHM-dominants. Hatched areas indicate sections uhere a CHM-dominant species becomes an ACT-subordinate, and a CHM-subordinate becomes an ACT-dominant (thus excluding the CHM-dominant from this section of gradient).

Fig. 6 . Replacement series graphs for competition experiments with Fucus serralus and F. t.rsiculosus (left) or with F. cesiculosus and F. spiralis (right) at three total densities (low = 100, medium = 500, high = 1000 initial shoots per 10 x 10 cm). Mean values of 5 replicates.

Fig. 8b). Yield reduction due to interspecific competition was very slight (17.55 g vs 17.90 g in control). Yield of F. ce.siculosus was reduced to 10.45 g when grown together with another 500 individuals of F. ~esicu1osu.s. but this intraspecific effect was not significant.

Fig. 1. Two models explaining a) species distribution along field observation = realized niches resource gradients. a) Upper graph: physiological response curves of 4 species realized in the field. Removal of neighbours may lead to the pattern explained by niche differentiation (lower left) or by the competitive hierarchy hypothesis (lower right). b) Fundamental and realized niches of the models on the same resource gradient. \competitive hierarchy Modified after Keddy (1989a). niche

Fig. 10. Yields of transplanted Fucus germlings as percentages of yield in the zone of natural occurrence. Weight loss during the course of the experiment led to negative yield of F. serratus in the F. spiralis zone. Mean values of 7 replicates.

Fig. 9. Results of "mixcd additive.substitutive design" comparisons (MASD) from certain field experiment treatments of replacement series performed with Fucus serrntzrs and F. ccsic.ulosus (a and b) or with F. t.csicitlosus and I;: spiralis (c and d). Data are yields of an initial 250 F. serrntus (a). 250 F. resic.ulo.szi,s(b and c) or 250 F. spiralis (d) under different intra- or interspecific configurations (mean values of 5 replicates). Arrows indicate significant differences in Tukey-Kramer post hoc tests at r = 0.05 done subsequently to ANOVA. Dashed lines indicate no significant difference. For further explanations see Fig. 4 and text. Note that germlings of 2 species !\ere interdispersed in mixtllres and not clumped in separate corners as symbolized here.

Fig. 5. Experimental construction for transplanting laboratory-raisedFucus germlings to the Helgoland northern seawall ("N-Mole"). Ceramic tiles (upper left) had 8 elevated areas. Three tiles, each bearing a different Fucus species (upper right), were fixed to each Fucus zone on the seawall (bottom). Natural vegetation around experimental units was removed so that there were no plants foreign to the experiment within 2 m of each tile (i.e. there were no interspecific competitibe effects). For further explanations see text.

Table 6. Winter transplants of Fuclrs spp. from an experiment mostly destroyed by a storm. Initial and final numbers of shoots from concrete fragments that withstood the storm. In parentheses percentage survival.

Fig. 1. Two models explaining a) species distribution along field observation = realized niches resource gradients. a) Upper graph: physiological response curves of 4 species realized in the field. Removal of neighbours may lead to the pattern explained by niche differentiation (lower left) or by the competitive hierarchy hypothesis (lower right). b) Fundamental and realized niches of the models on the same resource gradient. \competitive hierarchy Modified after Keddy (1989a). niche

Fig. 8b). Yield reduction due to interspecific competition was very slight (17.55 g vs 17.90 g in control). Yield of F. ce.siculosus was reduced to 10.45 g when grown together with another 500 individuals of F. ~esicu1osu.s. but this intraspecific effect was not significant.

Full text

A Competitive Hierarchy Model Integrating Roles of Physiological Competence
and Competitive Ability Does Not Provide a Mechanistic Explanation for the
Zonation of Three Intertidal Fucus Species in Europe
Rolf Karez; Anthony R. O. Chapman
Oikos, Vol. 81, No. 3. (Apr., 1998), pp. 471-494.
Stable URL:
http://links.jstor.org/sici?sici=0030-1299%28199804%2981%3A3%3C471%3AACHMIR%3E2.0.CO%3B2-0
Oikos is currently published by Nordic Society Oikos.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/oikos.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic
journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,
and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take
advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org.
http://www.jstor.org
Tue Dec 11 12:51:02 2007

OlKOS
81: 471-494.
Copenhagen
1998
A
competitive hierarchy model integrating roles of physiological
competence and competitive ability does not provide a mechanistic
explanation for the zonation of three intertidal
Fucus
species in Europe
Rolf Karez and Anthony
R.
0.
Chapman
Karcz,
R.
and Chapman, A.
R.
0. 1998. A competitive hierarchy model integrating
roles of physiological competence and competitive ability does not provide a mecha-
nistic explanation for the zonation of three intertidal
Fucus
species in Europe.
-
Oikos 81: 371 394.
It has been clear for the last 20 yr that both competitive ability and physiological
competence on the intertidal gradient of rocky shores determine the zoned distribu-
tion of fucoid seaweeds in the North Atlantic. However to this point. there has been
no concept integrating these functions for seaweed vegetation in a single mechanistic
explanation. Keddy's competitive hierarchy model, which has not been tested in
marine systems
(i.e. not on seaweed vegetation) provides an explanation for the
zonation of plant species on environmental gradients. The model proposes that
competitive abilities of spccies on a stress or resource gradient are inversely related to
fundamental niche breadths. We tested
2
assumptions of the model to detem~ine
whether it provided
a
comprehensive explanation of zonation of 3
Fttcus
species on
the island of Helgoland in the North Sea. The
2
assumptions translate into specific
predictions for the
Fticzrs
assemblage (where
F.
srrrurus
occurs on the low shore.
F.
spiraiis
on the high shore. and
F.
t.c~siculu.sus
in between): I. competitive ranking is
F.
srrratu5
>
F.
cesiculosus
>
F. spirali.~:
2.
fundamental niche breadth rankings are the
reverse of cornpctitive ability rankings. Pairwise competition experiments were done
in the field. A highly modified substitutive design was used, in order to take into
account the shortcomings of this approach. The empirically
deri~ed competitive
hierarchy did not fit prediction 1. Transplant experiments with adults and juveniles
provided results contrary to prediction
2.
Since none of the assumptions of the model
fit. it cannot be used to provide a mechanistic explanation for the zonation
of
FU(.US
species
011
Helgoland. Unlike other inlportant models of competition, Keddy's
approach does not claim universal validity in all communities. This means that it
requires empirical testing in each species
assemblage.
and more importantly. that the
negative outcome reported here does not invalidate the concept in general.
R.
Kure:. Ahreilung Aleeresbofmrik.
Illst.
jZr ~%leereskcrrzde,Diisternhrooker Weg
20.
0-24105
Kiel, Gertr~rmnj.
(rknre~@ifi?~.~mi-h-if1.h).
A.
0.
Clinpr~ian. Depr
qf
-
R.
Biologj,, Dtrlliotrsie (Itiit.., Hulifcis.
.VS.
Canatla
5-31?
451.
One of the steepest environmental gradients for plants
marshes has mixed terrestrial and marine origins, and
occurs at the interface between land and sea. In the
for the two components. the landward and seaward
North Atlantic Ocean, in areas removed from human
ends of the gradient represent different poles of
physio-
engineering influence, salt marshes tend to dominate the
logical stress. However. on rocky shores of the North
intertidal gradient on soft sediments. The flora of
Atlantic. this level of complexity is reduced because the
Accepted 11 September 1997
Copyrtght
d
OIKOS 1998
lSSN 0030-1299
Prtnted In l~eland
-
all r~ghts reserved

vegetation consists of seaweeds with a marine origin for
which terrestrial conditions become increasingly more
stressful landwards. On this tidal gradient, there is a
very obvious zonation of algal species (Stephenson and
Stephenson 1972), and, in the boreal waters of the
North Atlantic. fucoids (division Heterokontophyta:
order Fucales) are often dominant in
terms of total
biomass. There are 6 species of fucoids in these waters,
and they have been the subject of intensive ecological
investigation (reviewed by Chapman 1995). This paper
is concerned with 3 common species of
Fzicus
L.
occur-
ring in Europe. Where there is a closed vegetation,
FZICIISserratus
L..
F. cesiculosi/s
L.
and
F.
spira1i.s
L.
frequently occur in consecutive, contiguous zones be-
tween the tide marks from sea to land. Seaweeds other
than fucoids and species of fucoids other than the
3
above may form zones, but these are not of concern
here. It is quite clear that the formation of zones by the
3 species under consideration is under the dual control
of varying physiological competences and competitive
interactions along the intertidal gradient (reviewed by
Chapman 1986. 1995). The gradient encompasses
changes in resources
(e.g. dissolved minerals [Schon-
beck and Norton 1979a. Hurd and Dring 19901) and
also in stresses which cannot be considered resources
(e.g. temperature, Davison et al. 1989). The
3
species of
fucoids which occur on different parts of the gradient
have different physiological competences in the face of
changing stresses landwards. Drought tolerance of the
photosynthetic machinery of
F.
serrutus
is reduced in
relation to
F. cesiculosus,
which is, in turn, reduced
below that of
F.
spiralis
(Dring and Brown 1982).
F.
spiralis
is much more tolerant of freezing than low-
shore species (Pearson and Davison 1993). This same
species has enhanced phosphorus uptake capability in
comparison with
F.
resiculosus,
which. in turn, sur-
passes
F. serratus
(Hurd and Dring 1990). Fucoids
obtain all of their phosphorus during immersion, and,
landwards, the time of immersion is diminished up to
5-fold through the portion of the gradient occupied by
the three species. In all of these cases, it is the landward
borders of the species
that are related to the limits of
physiological competences. Furthermore, at least in
special circumstances, seaward borders may be physio-
logically determined. Dring (1987) showed that, in the
turbid
waters of the Bristol Channel in England, the
lower boundary of
F. serratus
is likely determined by
inadequate light availability for photosynthesis. In the
subtidal populations of
Fucus
occurring in the Baltic
Sea, light penetration may well regulate the depth of
the downslope boundary (Kautsky et al. 1986).
There is also evidence that competition among the
Fucla
species plays a major role in the formation of
zones. Both upper and lower bounds of fucoids may be
controlled by competition. In a rather simple experi-
mental design, Hawkins and Hartnoll (1985) and
Hawkins and Harkin (1985) demonstrated that the
seaward boundaries of the 3 European
F~~cus
species
under consideration could be disrupted by removing the
canopy occupants. Hence. downslope boundaries were
influenced by competitive interaction. Upslope migra-
tions indicated that competition has a similar role in
setting upper boundaries for some species pairs. Fur-
ther cases are reviewed by Chapman (1986, 1995).
Simply identifying physiological competence and
competitive ability as determinants of zonation does
little to provide an explanatory mechanism which inte-
grates the two. However, Chapman (1995) recognized,
retroactively, that there was an appropriate integrating
model (Keddy
1989a). He attempted to apply the re-
sults of earlier experiments, not done specifically with
model assumptions in mind, to explain the zonation of
Fucus
species of eastern Canada. The experimental
results did not fit model assumptions, but the designs of
the tests used by Chapman (1995) are not considered as
appropriate here. For
example, a replacement series
design (a substitutive model) was used to test competi-
tive ability without taking into account all of the short-
comings of such a design. For this reason. we began a
study to test whether Keddy's
(1989a) model (Fig. 1)
could provide a mechanistic explanation (which inte-
grates the roles of competition and physiological com-
petence) for the phenomenon of fucoid zonation in the
intertidal zone of the island of Helgoland in the North
Sea.
According to Keddy's
(1989a) model (Fig. 1). plant
species which are zoned on an environmental gradient
may have overlapping fundamental niche breadths (i.e.
portions of the gradient in which species are able to
exist in the absence of interspecific competition). so that
all species are able to exist (physiologically) at the
benign end of the gradient. The species which occupies
space at this benign end of the gradient is the top
competitor which displaces other species to positions
higher on the gradient. The top competitor cannot
survive conditions outside its realized distribution. The
species which occurs in the zone contiguous with that
of the top competitor is the second most effective
competitor. This second-ranking species displaces all
other species (except the top competitor) to positions
higher up the gradient where it is unable to exist. These
relationships continue on down through the hierarchy
to the worst competitor which is displaced to the
highest position on the gradient where physiological
conditions are least benign. There may be a strategic
resource allocation trade-off between competitive abil-
ity and tolerance of stress at the less benign end of the
gradient, but this is not known with certainty for any
single seaweed species.
The essence of Keddy's
(1989a) model can be found
in its
3
assumptions: 1. "[Ilt is assumed that the species
in the community have inclusive niches: i.e. the gradient
is a gradient of resource quantity. with all species
having best performance (size; growth rate and
repro-

- -
-
Fig.
1.
Two models explaining
a)
species distribution along
field observation
=
realized niches
resource gradients. a) Upper
graph: physiological response
curves of
4
species realized in
the field. Removal of
neighbours may lead to the
pattern explained by niche
differentiation (lower left) or
by the competitive hierarchy
hypothesis (lower right).
b)
Fundamental and realized
niches of the models on the
same resource gradient.
\
competitive hierarchy
Modified after Keddy (1989a).
niche
b)
niche differentiation
competitive hierarchy
fundamental niches fundamental niches
realized niches
realized niches
ductive output) at the same end of the gradient."
2.
"[Tlhe species vary in competitive ability in a pre-
dictable manner and [...I competitive ability is an inher-
ent characteristic of a species. perhaps having
something to do with rates of resource acquisition and
capacity to interfere with neighbours."
3.
"[Clompeti-
tive abilities are negatively correlated with fundamental
niche width, perhaps because of an inherent trade-off
between ability for interference competition and ability
to tolerate low resource levels.".
The model assumptions indicate explicitly that the
gradient is a resource gradient, but Keddy (1989a: p.
73) made it clear that it is also applicable to an environ-
mental gradient. Keddy's model may be tested in a
straightforward manner: zoned species along a gradient
should show transitive ranks in their competitive abili-
ties and. in reverse order, in the widths of their funda-
mental niches. According to the second of Keddy's
assumptions these ranks should be consistent among
different environmental conditions (environments).
Keddy and his co-workers provided much evidence for
the competitive hierarchy hypothesis from their own
work. In most cases. the surveyed gradient was parallel
to the shoreline of Canadian lakes. Gradients mediated
by wave exposure exist between sheltered bays with rich
organic content, high soil nutrient levels (=low stress
sensu Grime
1974), high plant biomass and low distur-
bance, and exposed sites with infertile soils of coarser
grain structure, low plant biomass, with high levels of
stress and disturbance (Keddy 1984, Wilson and Keddy
1985, 1986a, b). It should be noted that this gradient is
a very complex one. However, even on gradients of, for
example, only one mineral nutrient, interactions with
other factors (ion balances) may produce complex gra-
dients (Austin and Austin 1980).
In his studies with lakeshore plant communities,
Keddy (1984) found no evidence that niche differentia-
tion (Fig. l) led to species' coexistence. He found that
most species had highest biomasses at the same (high
fertility) end of the gradient (Wilson and Keddy 1985),
that the competitive abilities of species from the high
fertility end were highest (Wilson and Keddy 1986'0)
and that diffuse competition was greatest there (Wilson
and Keddy 1986a). In the following years, Keddy found
that plant biomass (and height) is a trait that often
explains competitive dominance (Gaudet and Keddy
1988, Keddy
1989b). He came to the conviction that in
plant communities consistent hierarchies of competitive

ability prevail (Keddy and Shipley 1989, Shipley 1993.
Shipley and Keddy 1994) rather than intransitive net-
works as known from, for example, bryozoan assem-
blages (e.g. Buss and Jackson 1979. Karlson and
Jackson 1981. Russ 1982). For hierarchies to develop.
competitive interactions have to be asymmetric (Keddy
and
Shipley 1989. Shipley and Keddy 1994). Keddy et
al. (1994) partly confirmed that hierarchies of competi-
tive effects (sensu
Goldberg 1990) were mostly consis-
tent between different environments. although in single
species pairs there were some reversals. Recently,
Keddy recommended looking for general patterns of
plant traits rather than compiling more and
more single
species pairs observations (Keddy 1992). and he co-
operated in an intercontinental experiment designed to
test the assumption that competitive intensity increases
with habitat productivity. Only partial supporting evi-
dence could be found (Reader et al. 1994).
The assumptions of Keddy's
(1989a) model may be
used to formulate testable hypotheses: 1. The competi-
tive ranking of the 3
Fucus
species on Helgoland is:
F.
.serratu.r
>
F.
resiculosus
>
F. .spirali.s:
2. The fundamen-
tal niche breadth rankings are:
F. .serratus
<
F,
ce.riculo-
.sus
<
F. .spiralis.
Competitive ranks were tested in a highly modified
replacement series design deployed in field experiments.
Fundamental niche breadths were tested in transplant
experiments in the absence of "alien" competitors.
Materials and methods
Description of the study site
The study was done on the sandstone north-east shore
of the island of Helgoland in the North Sea (54'1
1'N.
7"53'E). This shore is known locally as the "Nordost-
Felswatt". and will be referred to hereafter as the
"NE-Intertidal". From a high cliff close to shore, a flat
rocky terrace extends
seawards horizontally
>
300 m
from mean high water to mean low water, with a
difference in vertical level of 2.4 m (Janke 1990). Thus,
the mean angle of gradient is less than 0.5". Along the
whole stretch.
the NE-Intertidal is sheltered from west-
erly storms by a seawall ("N-Mole").
A pattern of grooves in the bedrock distorts the
normal zonation, leading to two tidal gradients: a large
scale gradient along the land-sea axis and a small scale
gradient. right-angled to the former. Only zonation
patterns of the most conspicuous macroalgae along the
large scale gradient will be considered here.
On the lower shore, there is a dense canopy of
Fuci~s
serratus
(leaf area index
>>
I)
which becomes more scat-
tered with a lower cover landwards. Dense patches of
Myti1~t.s edulis L.
populate the tops of the sandstone
slabs.
F.
r~esic~~1osu.s
alone or together with
F. serratus
builds patches of sparse canopies. In the uppermost
zone of the main NE-Intertidal. plateaux of rocks more
recently fallen from the cliff are often populated by a
dense mixed canopy. mostly of
F. aesiculosus
and
F.
spiralis
and, in a few cases,
F. serratus.
The area
surrounding these boulders is populated from ca March
to September by dense stands of green ephemeral algae.
mainly
Enteromorpha
Link in Nees species (Janke
1986).
F.
spiralis
forms a distinct zone on large concrete
blocks of the eastward extents of the NE-Intertidal. The
typical zonation pattern of
Fucus
spp. (landwards
F.
.serrutus-F. ~~esiculosus-F. .spiralis),
conspicuous on
most of Helgoland's seawalls. and the main subject of
this study, is not clearly discernible in the main
NE-
Intertidal.
The grazer guild consists mainly of
Littorina
L.
spp.
and small crustaceans, while limpets, a major structur-
ing force on e.g. many British coasts (Southward and
Southward 1978. Hawkins 1981). are absent from the
Helgoland intertidal.
More comprehensive descriptions of the distribution
of Helgoland's intertidal species are given by
Markham
and Munda (1980), Janke (1986) and some of the
studies summarized by Harms (1993).
Competition
Crnsztses of tlutural
Fucus
gernzling dmsit), in the
Helgolarzd intrrtidc~l zone
Censuses in the Helgoland intertidal zone of subjec-
tively "dense" stands of
Fucus
spp. were done in or-
der to estimate appropriate experimental densities. In
July 1993. numbers were assigned to 20 patches of
Fzrcus
germlings distributed across the fucoid belt.
Seven of them were chosen
randomly and algae from
5 small areas (5
x
5 cm. coordinates with random
number tables) of each were scraped from the rock
and brought to the laboratory. Densities of samples
and size class frequency distributions from subsamples
were estimated. Patches of
Fucus spiralis
juveniles (1-
2 cm length) occurred at densities up to 70 000
shoots m'. Since the experimental size of fucoids was
to be 1.0-1.5 cm, this density was considered to be
the approximate natural (maximum) density, although
only
F. spira1i.s
seems to establish such dense stands
of juveniles (pers. obs. and see Schonbeck and Nor-
ton
1979b: but see Creed et al.
1996 for similar data
on
F, .serrntus).
Replacet7zent series esprrinzent
For experimental examination of competition. the re-
placement series design (de Wit 1960) was used. Here,
the original design was extended by using 3 replacement
series with 3 total densities. Densities of low
=
10 000.
medium
=
50 000 and high
=
100 000 shootslm"ere
chosen for the experiment so that the highest field
density for fucoid germlings of the experimental size