An experimental analysis of ecological dominance in a rocky subtidal macroalgal community

J. Exp. Mar. Biol. Ecol., 1990, Vol. 136, pp. 170-195

179

Elsevier

JEMBE 01388

An experimental analysis of ecological dominance in a rocky subtidal macroalgal community Verena Rapp de E&on’ and Wilton 0. Bussab’ ‘Institute Oceanogrbifco. Universidade de Sdo Paulo, SCo Paulo, SBo Paulo. Brazil; 21nstitutode Matenuitica e Estatistica, Vniversidade de Sa”oPaulo. Scio Pa&o, &?a Paulo. Brazil

(Received 2 August 1989; revision received 14 December 1989; accepted 9 January 1990) Abstract: Closed canopies of Sa~assum occur both at the low intertidal and shallow subtidal levels of southern Brazilian rocky coasts in the absence of sea urchin grazing. Sargussurn stenophyllum (Mertens) Martius possess a flexible frond, ~40 cm long, which is stirred by the waves. To investigate the effects of S. stenophyllurnon settlement and persistence of macroalgae, experiments were conducted for a 3-yr period. The treatments comprised untreated control, removal of all macroorganisms except Sargassum, Sargassum removal and scraped clean, representing different levels of perturbation, and were initiated in three different seasons. Three distinct periods of colonization of the cleared surfaces could be observed, relative to competitive interactions among species: (1) an initial period of succession, with increased abundance of ephemerals, (2) an intermediate period, with maintenance of Dictyoptezi spp. cover and (3) the reestablishment of closed canopies of S. stenophyllum. S. stenophyllum was both the competitive dominant and an opportunist colonist. Although it promptly colonizes newly cleared surfaces, it grows slowly, therefore, formation of a closed canopy is delayed. The brown algae Dictyopterir delicatula and D.plagiogramma occupied a high proportion of the rocky surface with continued removal of Sargassum colonists; nevertheless, frequent clearings opened in their cover due to poor ability of Dictyopteris spp. to adhere to the rock allowed eventual establishment of Sargassum. Key words: Competition; Community structure; Macroalgal community; Recovery pattern; Sargassum

Rocky seashore communities are often approximately homogeneous, with a single species or aggregates of comparable life forms occupying at least 80% of the surface (Paine, 1984). Several species of canopy-forming brown algae of the orders Laminariales and Fucales have been reported to monopolize rocky substrata in temperate and high-latitude regions of both hemispheres (Schiel & Foster, 1986), whenever sea urchin densities are low enough that the algae are not removed by their grazing (Dayton, 1985; Pringle, 1986; Johnson & Mann, 1988). Disturbances, either physical or biological, which disrupt the community organization, release resources in space and time that enable recruitment of new individuals (Sousa, 1984). Various survival patterns which reflect physioio~cal and morpholo~c~ adaptations allow different algae to coexist in Correspondence address: V. Rapp de Eston, Av. Pde. Pereira de Andrade 545 apto 103/E, S&o Paula, SP 05469, Brazil. 0022-0981/90/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)

i’ K I)E ESTON ,AND IV 0. BUSSAB

1x11

similar environments Littler, 1980). Closed canopies levels of southern reported

through spatial and temporal

partitioning

of resources

of Surgussum occur both at the low intertidal Brazilian

for other localities

rocky coasts (Paula & Eston, (Umezaki,

1974; Wanders,

(Littler J;r

and shallow subtidal

1987), and have also been

1976; Schiel & Choat,

1980;

McCourt, 1984; Ang, 1986). Despite its abundance and importance as a primary producer (Mann, 1982) plus its invasive abilities (Norton, 1981; Deysher 8r. Norton. 1982; Critchley, 1983a-c; Paula & Eston, 1987) interactions among Sargassum spp. and other algae are little understood. The 13 taxa of Sargussum which inhabit the rocky coast of the state of Sao Paula may be grouped into five morphological types, from dwarf to flexible or stiff elongatedfrond plants, each type observed under characteristic conditions of wave exposure and stress or depth (Paula, 1988). Sargassum stenophyllum (Mertens) Martius possesses a flexible frond, z 40 cm long, which is stirred by the waves, and which can form a closed canopy over understory and primary space occupiers. To investigate the effects of S. stenophylfum on settlement and persistence of macroalgae, experiments were conducted in a shallow subtidal macroalgal community for a 3-yr study period. desiccation

METHODS STUDY

SITE

Experiments were done at Lazaro (23”3O’S, 45”08’W) in Fortaleza Cove on the northeastern coast of the state of S%o Paulo, Brazil. The site is moderately exposed to waves and S. .stenophylZum is subtidally abundant on large gneiss boulders. These occur to a depth of 8 m and are distributed over an area of 1200 rn’. Because of their large size, the boulders are not dislodged by the waves. Dense cover of S. stenophyllum (95-100%) reduces the incident irradiance to l-100;, of that at surface (Eston, 1987). A sparse understory composed primarily of the brown algae Dictyopteris delicatula Lamouroux and D. plagiogramma Montagne (Vickers) occurs beneath the Sargassum canopy. In addition,

variable amounts

of tubes of young polychaetes

of Chaetopterus sp.

are found, mostly empty, as well as a few algal crusts, sponges and colonial tunicates. Hypnea musciformis (Wulfen) Lamouroux is observed especially during summer months, as a Sargassum epiphyte, which increases the overstory cover. EXPERIMENTAL

DESIGN

A series of manipulations was performed to investigate the effects of S. stenophyllum on settlement and persistence of algae (settlement here is based on visual observations of tiny macroscopic plants). Four treatments were established: scraped clean of all macroorganisms, Sargassum removal, removal of all macroorganisms except Sargassum (for convenience, we will call this understory removal) and untreated control (Fig. 1). Each treatment was replicated three times. Treatments and replicates were assigned to

SARGASSKW-DOMINATED

181

MACROALGAL COMMUNITY MACRO-ORGANISMS

;: E A

E

E N f s

untreated scraped

control

I

+

+

I

c/eon

Sofgossum

removol

unders tory

removal

+

i--:

+

Fig. 1. Design of experiment to investigate effects of S. stenop~y~Zum on settlement and persistence of macroalgae. Each treatment was replicated three times. Replicated sets of four treatments were initiated in December 1983 (summer) and September 1984 (early spring). In April 1985 (autumn), three treatments were set up: scraped clean, Sargassum removal and untreated control. Four treatments were maintained clear of sea urchins.

different boulders, which were tagged with PVC labels glued to the rocky surface with epoxy putty (Z-Spar A-788 Splashzone Compound, Koppers). Since the initiation of clearances at different times of the year may result in colonization by different species, replicated sets of the four treatments were initiated in December 1983 (summer experiment) and September 1984 (early spring experiment). In April 1985 (autumn experiment), only three treatments were set up: scraped clean, Sargassum removal and untreated control. Therefore, 33 boulders were used. Five additional Sargassum removal boulders were maintained clear of Surgassum plants for a prolonged period (until January 1986) to enable identification of species able to establish and maintain a closed cover without Sargussum interference: two boulders for removals initiated in January 1984 and three for those set up in September 1984. Treatments were assigned to boulders using a r~domized block design. Because there were not enough boulders of approximately the same size, surface area of the tops of most of the boulders used ranged from 1.5 to 5.5 m*. Depth of these top surfaces varied from 0.50 to 2.50 m below ELWS and they were situated at least 0.50 m above the sandy bottom. To reduce variation in colonization patterns due to the presence of cracks and crevices, boulders with large-scale heterogeneity in surface texture were avoided. Treatments were carried out over the entire top surface of the boulders in order to hinder inward vegetative growth from the perimeter of the cleared patches. The organisms were removed with a scraper and the rocky substrata of scraped clean treatment were scrubbed with a wire brush after scraping. Entire holdfasts of Surgassum plants were removed to avoid reg~eration and growth (Paula & Eston, 1987). Boulders maintained clear of Sargassum plants for a prolonged period were cleared as soon as S. stenophyllum began to grow noticeably.

k R. DF: ESrON ANI) IV.0

1K.I

BlISSAH

Percent cover of the macroorganisms established on each boulder was estimatccr visually every month using haphazardly placed quadrats (50 x SO cm). Since size of the boulders

varied. a different number

of quadrats

one to give a complete picture. These subsamples replicates

(two to four) was sampled

within each

were averaged before comparison

wnh

and other treatments.

Percent

cover was obtained

layering of plants. Percent cover because a smaller quadrat (12.5 aid cover estimation within the Organisms were identified as for further identification in the

for each species

separately

to take into account

the

of each species could be estimated to the nearest 1.Z “(, x 12.5 cm, divided into four equal parts), was used to 50 x 50-cm quadrat. far as possible in the field and samples were collected laboratory if necessary.

HERBIVORk

Astrrreu phoehicl Roding and Tegulu viridula (Gmelin), herbivorous gastropods that may feed on microscopic stages of macroalgae were observed within the algal bed. Nevertheless, a Kruskal-Wallis test showed no differences in their densities among boulders submitted to different treatments (Eston, 1987). Therefore, it was assumed that any effect they may have had in consuming newly settled algae did not alter the results as grazing should have been equivalent on all experimental rocky surfaces. Small numbers of sea urchins [Arbaciu lixula (L.) and Echinometra experimental area were removed.

lucunter (L.)] close to the

Among high-mobility herbivores, there were the roundspot porgy Diplodus urgenteus (Valenciennes), the Bermuda chub Kyphosus sectatrix (L.), the molly miller Scartelfa cristatu (L.), the crab Mithrax hispidus (Herbst) and the sea hare Aplysiu brusiliana Rang. Sporadically, one young green turtle Chelonia mydus (L.) was observed. Nevertheless, cages were not employed since there were few individuals of each herbivorous species and no evidence of high algal consumption. STATISTICAL

ANALYSIS

As an exploratory analysis, one-way ANOVA was employed on pretreatment evaluation of Sargassum cover. Angular transformation was applied to percent cover data to obtain 32 values, SD values and to look for between-differences of Sargassum cover. To reduce the dimensionality of the data, we created an indicator based on the competitive interactions among the most common species. We used the difference between S. stenophyllum percent cover and that of the species able to occupy a high proportion of the substrata whenever Sargassum cover was low. Thus, the indicator was defined as: Y = Sargassum

- (Dictyopteris + ephemeral)

SARGASSUM-DOMINATED MACROALGAL COMMUNITY

183

where Sargassum = percent cover of S. stenophyllum, Dictyopteris = percent cover of D. delicatula + D. plagiogramma and ephemeral = percent cover of ephemeral green algae and tube-forming benthic diatoms. Due to the multilayered nature of this community, Y may take over values from - 100% [with no Sargassum and 100% cover of (Dictyopteris + ephemeral)] to + 100% [with no (Dictyopferis + ephemeral) and 100% cover of Sargassum]. Therefore, higher values denote prevalence of S. ~tenoF~yl~urnover the two other groups of algae. Missing values were imputed by averaging values from samples either side of the missed sample, if these data were available, otherwise those were excluded from the analysis, The computed data matrix is indicated in Table I as well as those values imputed by averaging. T0 indicates pretreatment values and the month of treatment as well. Exception for the early spring experiment in which pretreatment values were obtained in July 1984 but rough water conditions postponed for a month initiation of treatments. Therefore, T, = September 1984 in this case. ’ Dictyopteris spp. cover data were grouped according to similar colonization patterns such as ability to increase their cover by vegetative growth as well as colonization through propagules from the water. Nevertheless, D. deli~a~la increased its cover especially on boulders manipulated in early spring and autumn and D. plagiogramma especially on boulders treated in the summer. Cluster analysis of data of the competitive interaction matrix, defined by Y, was utilized to search into the effects caused by the treatments performed as well as time since they were initiated. Euclidean distance was chosen as the similarity measure and linkage method was the centroid (Ever&t, 1980). If we denote as Y,, each observation of the b-th row (b = boulder) and t-th column (t = time of sampling) of the matrix of values assumed by Y (Table I), similarity between time t and t’ is given by: d2 (t, t’) = f

(Y,, - Yb,J2

b=l

where k is the total number of boulders of each set of treatments. For sets of treatments started in the summer and in early spring, k = 12, while for the set of treatments initiated during the autumn, k = 9. Therefore, it is possible to observe how the clusters are formed in response to time since the treatments were started, related to the prevalence of Sargazsum among different boulders. Afterwards, we used the same procedure to cluster boulders through time. In consequence, similarity between two boulders & and 6’ is given by: d’(b,b’)=

f f=

(Y,,-

Y6.1)2

1

where T is the total number of months in which samples were obtained within each season (T = 20 in the summer, T = 14 in early spring and T = 8 in autumn). We preferred this descriptive technique for the examination of the interactions among

cri‘a;impling

__~.,,

rreatnrcnts

50 bY 61

72 84 -35

I 2 3

f 2 3

Se

s

.- ._-.

_2..

84 12 51

13 71 51 14 -6 2x

15 39 34

..__.--__.._...

X6 23 22

90 67 18

__._I. _

7;

.-I -3 -'83

2Y -7

-50 -23 --6-i -30 ~-58 67'

79 69 3

Th

82 68 39

--I2 - I5 -24 -5 -8 -23 -7 20 -53

84 IO 16

86 88 58

T,

JB4 J84

..^^^

__~. ..--_.--.,-.

i-4

4% Ki,

_L---_

7 39 -4 7 -41 -5 1; -1 -.55 .-40 -46 -75

-50 -19 -1

48 -13 25

X4 83 42

T2 T.3 ._“._“~”

.._.2..

__..“.__ ...-.

60 -4s -71

-50 '-SO 50

56 -.22 38

93 25 51

z 3

c

1

92 82 39

87 64

c

82

‘I, _. ..-.

22

I

7;,

.-

a,

2 3

Summer .“_._. ,____..---. ._-...-_

The

-.

% 23 ." 77

-47 -3 -86

8i 32 14

86 89 2

.

64 78 -87

-83 --38 83

95 57 55

91 97 39

99 --66

69

29 73 1

98 $6 88

Yl 87 33

94 -7

75

55 86 70

76 94 93

96 78 14

__.._ _ ._..-I .“i-.

88 -91

5i

-50 26 -6?

98 43 67

97 80 39

G -Tic, T,I T,I TL4 ..-.__.___. .-. .____-__-

I‘M

96 -14

81

.

” . ..,._

96 -90

70

81 x2 5Y

YB 80 70 83 76 68 79 93 -27

87 73 85

89 84 86

33

80 77

rz, -.

86 91 79

89 89 67

7^17 . .___.

,__ ..__..---

84 -8

72

56 90 79

82 97 94

89 86 82

T,, .._

55 -

59

100

.__

Percem cover data of Y = Sargasfuwl - (Dict.vopterfs i ephemeral) which represents matrix of competitive interactions. Dictyopteris represents both D. tlrliitrruiu xxi D pitigiogramma. Ephemeral comprises especially C. wgubundu and U.&zwiatu as well as E, flesuasu and tube-forming diatoms C,floreurus, G. rrmrina and AC~fflc~te[~~~,

z

> _ _

3

39 82 68 - 76 - 71 - 25 -21 -8 - 19

78 40 16 35 67 59 60 65

1

s

1

2 3

-79 - 53 - 15

- 59 - 38 -63

- 16 -77 -23

- 100 -86 -100 -100 -96 -92 - 82 -26 - 13

3 49 63

-21 70 48

14 66 80

6

1

2 3

3

T5 82 89 51

T3 92 91 75

94 98 82

88 87 70

T2

71 71 57

T,

1 2

SC 2 3

u

c

T,,

-22

-5 -1

1 3

S

-7

-11

-2 0 -8

- 35 -5 - 36

- 93

1 2 3 2

45 88

40 88

SC

75

68

I 2 3

C

-7 18 43

90 -Ii 93

78 79 13

84 88 74

TW

Treatments

91 -3 93

50 48 43

79 76 66

T*

52 -7 46

81 -7 91

47 26 45

88 89 83

i-8

2 -59 71

35 -6 79

72 69 60

90 89 12

T7

-20 - 16 19

- 74 -21 15

- 12 -23 - 21

34 76 74

90 89 62

T6

-71

-33 -57

-30 -2 -20

10 100

77

-20 15 49

94 35 75

54 10 13

80 93 71

-1o*

19 0

-44 0 -401

29 100

65

7-12 TM

51

90 92

7 45 100

60 85

65

59 66 85

91 97 97

94 91 55

90 96 92

53

82 75

82 51 100

60 90

73

62 72

73 68 25

91 80 39*

63 80 100

58 95 52 53 100*

76

68 - 25 94

61 12

66

94 88 100 97 12 100 29

89 59 28

87 83 75

T,9

89 53 37

91 83 56

7.17 l-18

C, untreated control; U, understory removal; SC, scraped clean; S, Sargassum removal; I-3, replicates; 7’9,pretreatment values; T,-T,,, months since treatments initiated; summer, spring and autumn, season of start of treatment. Missing sampling occasions derived from rough water conditions which hindered scuba diving. * Missing values were imputed through average between either side of missed sample.

Treatments

Spring

k? 7 4

5

0”

z? G 9 r

5 rl

El

H 7

B k 2 5

2

Srrr;q~.r.su~nand other algae to the usual ANOVA with repeated measures because ( I ) the unsuitability of assuming similar structures for the variances and covarianccs the errors among data derived

from treatments

as diverse as untreated

controls

of of and

scraped clean surfaces: m the latter the variances are almost null, therefore, we expect high correlations between consecutive measures, at least for the first months ofrecolonization;

and because

covariance

INITIAL

of (2) the low number

matrix when using unrestricted

PERIOD

OF SlJCCESSION

WITHIN

of replicates

from which to estimate

models of repeated

CL.EARED

intra-

measures.

PATCHES

Wherever bare space was available, several species of ephemeral green algae [especially, Cfudophoru vagabunda (Linnaeus) Hoek and Ulva fasciata Delile] and tube-forming diatoms (mostly, Nitzschia pulchelfu Peragallo et Peragallo) colonized the

60

20

100

60

2u

DFAJAODFAJAODF +-

1984 F

_.-.

-

,96,

_.___i

Fig. 2. Monthly response of algae to treatments performed in summer: C, untreated control; U, understory removal; SC, scraped clean; S, Sargassum removal. Mean cover values are indicated. SD > 8% are distinguished. For clearness, only one side of SD bars are drawn where necessary. Angular transformation was applied to percent cover data to obtain X and SD values. For graphical presentation, these data were transformed back into percentages. 1 points month treatment was accomplished. Variability among replicates of treatment S denotes presence of an outlier (Replicate 3).

Sz4RGASSUzWDOMINATED MACROALGAL COMMUNITY

187

EARLY SPRlNG

&

JSNJMMJSNJM

AJAODF

-

fDBC

-

Fig. 3. Monthly response of algae to treatments performed in early spring and autumn: notation as in Fig. 2.

rocky surfaces (Figs. 2-3). Their percent cover was variable among different treatments and among the seasons during which bare space became available. This colonization also occurred under Sargassum canopy when understory organisms were removed although in lesser amounts. Ephemeral green algae also occurred on untreated control boulders, predo~n~tly associated with initial colonization of natural open patches (300-750 cm2 in size) caused by detachment of S. stenophylum plants. Settlement of S. stenophyllum was common during the first 3 months where both scraped clean and Sargassum removal treatments were performed, regardless of season (Figs. 2-3). Nevertheless, it settled much later when the rocky surfaces were almost completely covered by Uiva fasciata (boulders scraped clean in early spring - Fig. 3). Settlement of Sargassum also occurred under its own canopy but sporadically and in low amounts. The number of S. stenophy~~um colonists was very variable within cleared patches and it reached a maximum settlement density of > 200 colonists * m - ’ which far exceeded the density of adult plants within untreated control boulders. In this last case, X and SD values were 64 & 14.8 plants * me2 in June 1985 and 70.7 + 13.3 plants-m-* in March 1986 (obtained from nine replicates in both cases).

I’ \l”I-I:KNS

II.

Ot

drlic~urulu

Increased

IINIII~KST-OIIY

C’O\‘I+

and II. pku~k~~runznz~

vegetative growth or colonization canopy

although

Whenever

in

low

the algae able to mamtain

were

cover once S. .vteno&&lm

has been removed

from propagules. amounts

S. srmoph~~llumcolonists

and

not

(Figs.

t’o~ :I w hilt ~1

Z-3).

both through

Both species grow under SU~~~LS.~UUI forming

have been constantly

a

continuous

removed,

cover.

Dicfjwpreris spl).

occupied a high proportion of the rocky surface (Fig. 4). Nevertheless, great fluctuation in their cover, which reflects a poor ability to adhere to the rock, provided constant formation of patches. These could be recolonized either by lateral encroachment of the two species of Dicf~vpteris or by settlement of other species including S. vfmophy/hrm. No obligate understory species, which die after canopy removal. have been observed.

.

5’.

stenophyllum

.

Dicfyopteris

b

ephemeral

SPP

DMJSDMJSDMJSD -!984w

-

Km--

w

,986--”

Fig. 4. Monthly response of algae to constant removal of Sargussum. Mean cover values are indicated. SD > 8% are distinguished. For clearness, only one side of SD bars are drawn where necessary. Angular transformation was applied to percent cover data to obtain X and SD values. For graphical presentation, these data were transformed back into percentages. 1 denote initial and final Sargussum removal.

CLUSTERING

WITH

RESPECT

TO TIME

AND

TREATMENTS

Three clusters of time emerged from the analysis of data from the summer experiment: times T,-T,, T,-T,, and T,,-T2, (Fig. 5a). Spring and autumn data gave similar results (Fig. 5b,c), notwithstanding between-season differences observed for pretreatment percent cover of S. stenophyllum (one-way ANOVA: F = 4.41, tail probability = 0.0201; df: between factors = 2 and within factors = 33). The three clusters indicate three distinct groups: T,-T6 (or T,) which corresponds to the initial period of succession, with increased abundance of ephemerals, T,-T,, to an intermediate period with maintenance of Dictyopteriscover, and T,,- Tz7 to the late successional period with

SARGASSUM-DOMINATED

MACROALGAL COMMUNITY

EUCLIDEAN 0 I

.

5 I

I

189

DISTANCE IO ,

IS

20

,

T2 TI T4 ;: T7 TS

T

0

Ts

A

Tll

TIOI

M

I

TIJ T& T/B T27 T24 T25 T17

E

0 ,

1

5 1

3

IO .

I.5

I

I

I

20 ,

Tb r: TJ T2

0B

z TB T/o Tt2 T/6 Tl9 T/7 TO 0

3

ik

T2 T4

IO

a

TIO I Tll TS TB Tl

20

JO

I

AUTUMN

40

0 C

Fig. 5. Dendrograms of Y = Surgassum-(Dictyopteriv + ephemeral) between time t and ?‘, produced by using average Euclidean distance coetlicient and centroid clustering method. Graph shows similarities between number of months (T,-T,,) since treatments were initiated, irrespective of treatment. To = initial composition of boulders. Season when treatments were performed is pointed.

a similar configuration to TO, when closed canopies of S. stenophyllum were reestablished. Note that T,, (pretreatment values of the boulders) is included in the last group, denoting complete recovery of the macroalgal community. Average percent cover data of homogeneous groups produced by using Euclidean distance coefficient, point to an equivalence among replicates of the different treatments for the reestablished period (T,, on), regardless the season in which the treatments were initiated (Table II). Replicate number three of the summer Sargassum removal treatment was an outlier, increasing the variability of the cover data observed (Table I). This was a boulder densely covered by D.plagiogramma at the time Sargassum plants were removed. Although it was not quantified, abrasion of the top surface of this boulder

I Y(!

V K. DE IXON

AND W 0 I’ARII

RIJSSAB

II

Average average

percent cover data of homogeneous groups (I-111, for treatments and trme) produced by usmg Euclidean distance coefficient and centroid clustering method on percent cover dat3 I Surgcrssunz-/Di~r~~~~~~~i.~ t ephemeral). C, untreated control; L, understory removal: s. Sarg~~s.ctc~rt removal; SC, scraped clean. T,-7, or T,, initial period of succession, with increased abundance of ephemerals; ‘r, or 7x-T,,. intermediate recovery period, with maintenance of Dicryoptrric cover; ;md 7‘, 1-Tz7, late successional period with a similar configuration to 71,. -Group

of treatments

Group

II

I T,-T, Summer

I

c + II

II III

s SC Total

T.,- T, ,

51.23 .- 13.28 32.41 16.61

~___. T,-T, Spring

I

c+u

II

s t SC Total

66.31 - 55.33 5.52 Ti-T,

Autumn

I II

C s + SC Total

of time

65.17 - 21.37 7.47

68.22 60.33 47.33 35.21

Total III T,,-Ty,,

Tv

80.35 83.50 70.63 78.27

68.33 46.15 16.X5 43.40

_~ ~~ ~~~- ____~~_~ 7,~T,,, T, T,- T,1 64.30 34.07 49.18

6Y.62 71.25 70.43

66.59 12.76 39.68

TR-T,, 66.58 68.83 68.08

65.X7 23.73 37.78

seemed to be greater than on others during turbulent periods, due to the orientation of this rocky surface towards the open sea. This boulder was not included in the cluster analysis, otherwise answers of the community to time of recovery would be misleading. Since after the 13th month of colonization all the boulders presented similar cover of macroalgae (Fig. 5a,b), only data from T, to T,, were used to assess the effect of experimental treatments (the use of information obtained for a period of > 1 yr to identify the effect of treatments would disguise the results). Boulders submitted either to untreated control or removal of understory organisms clustered together regardless of season treatments were initiated (Fig. 6). The exceptions were: (1) Replicate 2 of the summer understory removal, due to a decrease in the initial cover of Surgassum as a result of weakening of some plants’ holdfasts during removal of the understory stratum; this replicate behaved as if S. scenophyllum plants had been removed; and (2) Replicate 2 of autumn untreated contrd, which had a significant decrease in the amount of S. stenophyllum due to occurrence of large natural openings; this replicate clustered with the Sargussum removal and scraped clean group. Therefore, this community, when subjected to understory removal treatment, recovers extremely fast and soon becomes

SARGASSUWDOMINATED

MACROALGAL COMMUNITY

191

similar to the untreated control group as denoted by average percent cover data of homogeneous groups produced by using Euclidean distance coefficient {Table II). Scraped clean and ~~rg~~~ removal boulders clustered together both for treatments initiated in early spring and autumn (Fig. 6b,c). In the first case, this was a consequence of the confounding effect of the cover of both D. delicatula and U. fasciata on Y; the latter species was dominant in the initial colonization period

EUCLIDEAN 0 ,

I

10 I

DISTANCE 1

20 I

30 1

1

c2 SUMMER

Cl

0

UI us

T

c3

0

A

$2 SI

R

UP SC3

E

SC1 SC2

iiiii!l

A T M

0B

u2

E N

UI u.3 $3 SC3 SC1

T

s

St SC2 S2

~

0 I

10 I

20 I

b

30 7

C3 Cl c2 SC2 SP S! SC3 S3 SC/

Fig. 6. Dendrograms of Y = Surgussam - (&@opreti + ephemeral) between boulders b and 6 ’ , produced by using average Euclidean distance coefficient and centroid cfustering method. Graph shows similarities between pairs of boulders as indicated by their labels, e.g., C,, Ilrst replicate untreated control; S,, third replicate Surgurm~~~removal; U,, second replicate understory removal; SC,, first replicate scraped clean. Season when treatments were performed is pointed. Only data from T, to T,, (number of months since treatments were initiated) were used to assess effect of experimental treatments.

141

V K.I)EESTONANllWO.BUSSAR

on scraped boulders while D. delicatula increased its cover on Sargassum removal boulders at once (Fig. 3). In the second case (autumn), this similarity was the result of a low initial colonizati~)n by ephemerals (Fig. 3). foliowed by later establishment of i>. delicI_rtul~~.On the other hand, boulders submitted to Sargassum removal and scraped clean treatments in the summer clustered apart (Fig. 6a). reflecting the invasion of D. pfagiogrammo on boulders scraped clean, with formation of a closed cover for a short period: surfaces scraped clean in the summer were tbc last to reestablish (Table II). Superior competitive abilities of S. sten~phyllum within this macroalgal community were evidenced through the recovery of its closed canopy regardless of season or treatment performed.

DISCUSSION S. stenophyllum is the dominant species in this macroalgal community by maintaining a closed canopy year-round. It is able to recruit into cleared surfaces as soon as they arise, together with ephemeral algae. As the canopy grows, the abundance of understory species is reduced. Shading of the substratum is considered an important means of competition of established sporophytes of macroalgae such as the kelps Laminaria, Macrocystis and Pterygophora (Kitching, 1941; Dayton et al., 1984; Reed 8z Foster, 1984). Nevertheless, understory species were not prevented from growing beneath S. ste~o~hyllum canopy, surviving quite well at reduced abundances. Kitching (1941) and Gerard (1984) argue that negative effects caused by low light levels to plants trapped below the canopy may be reduced if water motion allows the incidence of light flashes through the stands. This may be the case in this macroalgal community since S. stenophylfumcanopy is constantly stirred by the waves. S. .~tenophy~~um moving canopy may impose a barrier against effective settling of spores or germlings by creating local turbulence as mentioned by Ang (1985) for Sargassum spp. from Philippines. Low canopies may impede the arrival of propagules to the substratum (Schiel & Foster, 1986), such as canopies of Cystoseira, another member of the Fucales, that also appear to preclude the recruitment and survival of other algae through physical interference (Dayton et al., 1984). S. sten~phy~~umpromptly colonizes bare surfaces year-round, even when a high proportion of the substratum is covered by Dictyopteris spp. This ability to colonize bare patches year-round is not shared by all Sargassum spp. since settlement of some species of the genus is seasonal (De Wreede, 1983; Ang, 1985). S. stenophyllum reproductive and settlement patterns suggest an opportunistic life history which includes year-round fertility (Paula, 1988), probably dispersal by abscission of fertile branches that float away, a mechanism common to many species of Sargassum (Paula & Eston, 1987), and the ability to regenerate and grow from cells of the

S~~G~SSU~-DOMINATED MACROALGAL COMMUNITY

193

holdfast (Eston, 1987). On the other hand, S. stenophyllum exhibits slow growth and a perennial thallus, characteristics of late successional forms. Damage incurred from the algal removal reagents is typical of that which occurs following episodes of natural disturbance at the study site. Natural openings in the middle of S. stenophyllum canopy provide opportunity for ephemeral algae and Dictyopteris spp. to colonize until the Sargassum canopy recloses. The creation of these small holes in the canopy by detachment of S. stenophyIium may be related to the age of individual plants. With increasing age, the holdfast grows in diameter but remains alive only at the perimeter (E. J. de Paula, pers. comm.), reducing its adherence capacity. Understory species probably coexist with S. sten~~hy~~um through the periodic liberation of space by Sargussum mortality and through their ability to grow in low densities under its canopy. S. stenophyllum apparently does not withstand intense herbivory by sea urchins since closed canopies of the species are not observed whenever high densities of E. lucunter occur (Paula 8~ Oliveira Filho, 1980; Giordano, 1986). Within this S. stenophyilum -dominated community, without sea urchins, selective grazing of other algae by herbivores, such as fishes and gastropods, cannot be considered to promote S. stenophyllum dominance since there is no evidence of high algal consumption. Also, Bermuda chubs, that are among the most conspicuous herbivores in the region, feed especially on Sargassum and Dictyopteris (Sazima & Sazima, 1983), the most common species. The response of this macroalgal community to increasing levels of disturbance, such as understory removal, Surgassum removal and scraped clean, indicate its resistance to displacement by other species. The community was able to return to its original composition even following a disturbance that markedly changed for a considerable time the abundance of the species, demonstrated through the boulders dominated by ~ic~~pten~ for an extended period due to constant removal of Sargassum.

This paper is based on a dissertation submitted by the first author in partial fulfillment of the requirements for the Ph.D. in the Instituto Oce~o~~co, University of S%o Paulo. Careful revision by C. R. Johnson, IL H. Mann, T. Moulton and anonymous reviewers greatly improved the manuscript. The work was supported by a FAPESP fellowship (Proc. 83/2413-3) and a CNPqgrant (Proc. 401560/83) to V. R. Eston, and a FAPESP grant (Proc. 8312412-7) to M.B.B. Kutner.

REFERENCES Ang, Jr., P.O., 1985. Studies on the recruitment of Surgumum spp (Fucales : Phaeophyta) in Balibago, Calatagan, Philippines. J. Exp. Mar. Biol. Ecoi., Vol. 91, pp. 293-301. Ang, Jr., P.O., 1986.Analysis of vegetation structure of a Sa~gassumcommunity in the Philippines. Mar. Ecol. Prog. Ser., Vol. 28, pp. 9-19.

1Y3

V. R. DE ESTON

AND W.0.

BUSSAB

Critchley, A.‘1 ., IY83a. Sargassum muticum: a taxonomic history mcluding world-wide and western Pacttic distributions. J. Mar. Biol. .Issoc~. U.K., Vol. 63, pp. 617-625. Critchley. A.T., lY83b. The establishment and increase ofSorgussum muticum (Yendo) Fensholt populattons within the Solent area of southern Britain. I. An investigation of the increase in number of populations individuals. Bat. Mar., Vol. 26, pp. 539-545. Critchley. A. T., 1983~. The establishment and increase of Sorgussum muticum (Yendo) Fensholt populations within the Solent area of southern Britain. Il. An investigation of the increase in canopy cover of the alga at low water. Bot. Mar., Vol. 26, pp. 547-552. Dayton, P. K., 1985. The structure and regulation ofsome South American kelp communities. 6’co/. Monugr., Vol. 55. pp. 447-468. Dayton, P. K., V. Currie. T. Gerrodette, B. D. Keller, R. Rosenthal & D. Ven Tresca, 1984. Patch dynamics and stability of some California kelp communities. Ecol. Monogr., Vol. 54, pp. 253-289. De Wreede, R.E.. 1983. Sargussum muricum (Fucales, Phaeophyta): regrowth and interaction with Rhodomela kuri.u(Ceramiales, Rhodophyta). Phycologiu, Vol. 22, pp. 153-160. Deysher, 1.. & T.A. Norton. 1982. Dispersal and colonization in Sargussum muticum (Yendo) Fensholt. ,I. E\/J. Mar Bioi. hcoi.. Vol. 56. pp. 179-195 Eston, V.R., 1987. Avaliacao experimental da dominancia ecologica em uma comunidade de macroalgas do infralitoral rochoso (Ubatuba, SP, Brasil). Ph.D. dissertation, University of SBo Paulo, S%o Paulo, SP, Brazil, 129 pp. Everitt, B.S., lY80. Clusrer analysis. Halsted Press, New York, 136 pp. Gerard. V. A., 1984. The light environment in a giant kelp forest: influence of Macrocystis pyrijera on spatial and temporal variability. Mar. Biol., Vol. 54, pp. 189-195. Giordano, F., 1986. Ouricos do sub-litoral rochoso da regilo de SHo Sebasti&o. SP: uma abordagem ecologica. Master thesis, Unicamp, Campinas, SP, Brazil, 128 pp. Johnson, C. R. L K. H. Mann, 1988. Diversity, patterns of adaptation, and stability of Nova Scotian kelp beds. Ecol. Monogr., Vol. 58, pp. 129-154. Kitching, J.A., 1941. Studies in sublittoral ecology. III. Laminariu forest on the west coast of Scotland, a study of zonation in relation to wave action and illumination. Biol. Bull. Mar. Biol. Lab. (Woods Hole, Mass.), Vol. 80, pp. 324-337. Littler, M. M. & D. S. Littler, 1980. The evolution of thallus form and survival strategies in benthic marine macroalgae: field and laboratory test of a functional form model. Am. Nat., Vol. 116, pp. 25-44. Mann, K. H.. 1982. Ecology qfcoastal waters: a systems approach. Blackwell Scientific Publications, Oxford, 322 pp. McCourt, R. M., 1984. Seasonal patterns of abundance, distribution, and phenology in relation to growth strategies of three Sargassum species. J. E.xp. Mar. Biol. Ecol., Vol. 74, pp. 141-156. Norton, T. A.. 1981. The varied dispersal mechanisms of an invasive seaweed, Sargassum muticum. Phyxcologiu,Vol. 20, pp. 110-l I I. Paine, R.T., 1984. Ecological determinism in the competition for space. Ecology, Vol. 65, pp. 1339-1348. Paula, E.J., 1988. 0 g&nero Sargassum C. Ag. (Phaeophyta, Fucales) no litoral do Estado de SBo Paulo, Brasil. Bol. BOI. Univ. Sdo Puulo, Vol. 10, pp. 65-l 18. Paula, E.J. & V. R. Eston, 1987. Are there other Sargussum species potentially as invasive as S. muticum? Bot. Mar., Vol. 30, pp. 405-410. Paula, E. J. & E. C. Oliveira Filho, 1980. Aspectos fenolbgicos de duas popula@es de Surgussum cymosum (Phaeophyta, Fucales) do litoral de Sgo Paulo. Bol. Bot. LG. Sdo Puulo, Vol. 8, pp. 21-39. Pringle, J. D., 1986. A review of urchin/macroalgal associations with a new synthesis for nearshore, eastern Canadian waters. Monogr. Biol., Vol. 4, pp. 191-218. Reed, D.C. & M. S. Foster, 1984. The effects of canopy shading on algal recruitment and growth in a giant kelp forest. Ecology, Vol. 65, pp. 937-948. Rohlf, F. J. & R.R. Sokal, 1969. Statistical tables. W.H. Freeman & Co., San Francisco, 253 pp. Sazima. 1. & M. Sazima, 1983. Aspectos do comportamento alimentar e dieta da tartaruga marinha, Cheloniu mydas, no litoral norte paulista. Bol. Inst. Oceunogr. Sao Paula, Vol. 32, pp. 199-203. Schiel, D.R. & J.H. Choat, 1980. Effects of density on monospecific stands of marine algae. Nature (London), Vol. 285, pp. 324-326. Schiel, D. R. & M. S. Foster, 1986. The structure of subtidal algal stands in temperate waters. Oceanogr. Mur. Biol. Annu. Rev.. Vol. 24, pp. 265-308.

SARGASXSf-DOMINATED

MACROALGAL COMMUNITY

195

Sousa, W. P., 1984. Intertidal mosaics: patch size, propagule availability, and spatially variable patterns of succession. Ecology, Vol. 65, pp. 1918-1935. Umezaki, I., 1974. Ecological studies of Sff~g~su~ r~u~~e~~ (Mertens) 0. Kuntze in Maizuru Bay, Japan sea. Bob. Mug. Tokyo, Vol. 87, pp. 285-292. Wanders, J. B. W., 1976. The role of benthic algae in the shallow reef of Curacao (Netherlands Antilles). II. Primary productivity of the Sargassum beds on the Northeast coastal submarine plateau. Aquac. Bat., Vol. 2, pp. 327-335.