The importance of form: Differences in competitive ability, resistance to consumers and environmental stress in an assemblage of coralline algae

J. Exp. Mar. Biol. Ecol., 1984, Vol. 19, pp, 105-127 Elsevier

105

JEM 285

THE IMPORTANCE

OF FORM: DIFFERENCES

ABILITY, RESISTANCE

TO CONSUMERS

IN COMPETITIVE

AND ENVIRONMENTAL

STRESS IN AN ASSEMBLAGE OF CORALLINE ALGAE

DIANNA

K. PADILLA

Department of Zoology, Oregon State University, Corvallir, OR 97331. U.S.A.’

Abstract: Competitors, consumers, and environmental stresses are everpresent factors that most organisms face during their lifetime. On the Pacific northwest coast of North America there are three forms of coralline algae (Rhodophyta, fam. Corallinaceae) differentially distributed within and around large tide pools in the mid and low intertidal region. Observations and experiments were conducted to test whether species ofthree different forms (finely branched, coarsely branched, and encrusting forms) were differentially susceptible to consumers, competitors and environmental stress. The results indicate that the distributions of these three forms are distinctly different. Although no single factor appears to be controlling this zonation during the time scale of this study, the three forms are differentially susceptible to these potential selective agents. The finely branched coralline (Corallina vancouveriensis Yendo), found immediately outside of the pools on emergent substrata, is the most resistant to desiccation. This is apparently because its bushy structure holds more water than the other forms preventing drying during times of emersion. The coarsely branched coralline, Calliarthron tuberculosum (Post. & Rupr.), is the most resistant to consumers; it was not eaten by any of four molluscan herbivores tested. Again it appears to be the structure of the plant that provides this protection. When this species is ground and placed in agar, it is readily eaten by all four herbivores. The encrusting corallines (approximately six species) appear to be the best spatial competitors, occupying the majority of the space within the pools, and were the winners of all overgrowth encounters with articulated coralline species.

INTRODUCTION

Competitors, consumers, and environmental stress are three potential selective agents that organisms are faced with during their lifetime. Organisms can find escapes in time and space from the potential negative effects of these factors by changes in their life history or behavior, thereby reducing the probability of encounter with these agents. Alternatively, organisms can possess means of dealing and coexisting with them. Sessile organisms, particularly plants, have a limited array of potential behaviors, particularly once a site for growth has been established. Therefore, one would expect them to deal with these factors primarily through physiological (i.e., growth) and/or structural (i.e., morphological or chemical) responses. Grime (1977) has suggested that terrestrial plants have evolved along a continuum of three basic strategies: (1) C-Strategy, being a superior competitor, (2) S-Strategy, being able to tolerate environmental stress, and (3) R-Strategy, able to deal with the ’ Present addresses: Department of Zoology, The University of Alberta, Edmonton, Alberta, Canada T6G 2E9; and Bamtield Marine Station, Barnfield, B.C., Canada VOR 1BO. 0022-0981/84/$03.00 0 1984 Elsevier Science Publishers B.V.

106

DIANNA K. PADiLLA

effects of disturbance, including consumers. Steneck & Watling (1982) suggest that marine algae can be categorized into functional groups on the basis of their form regarding their relative abilities to withstand consumers, and, Norton it al. (1982) have considered many of the conflicting factors that are influenced by the form of an alga including the influences of drag forces caused by moving water and the efficiency of nutrient uptake and photosynthesis. Coralline algae exhibit a wide range of morphological types. These algae incorporate calcium carbonate (in the form of calcite) into their walls, making them very hard. They are a prominent feature of rocky intertidal regions world wide (Stephenson & Stephenson, 1972; Johansen, 1976, 1981) and they are common in and around tidepools along the Pacific northwest coast of North America (Ricketts et al. 1968). Here the saxicolous species (those that live on rock) are found in two basic morphological types: crustose (strictly prostrate) and upright species with a crustose basal holdfast which resembles crustose species and from which arise branches with noncalcified articulations (genicula). The articulated corallines can be further subdivided on the basis of their size and branch width into finely branched species which appear thick and bushy, and coarsely branched species which are more stick-like. Most large, deep tide pools (2 0.6 m across and 2 0.3 m deep) in the mid to low intertidal region of the Pacilic northwest coast are characterized by an abundance of coralline algae. Although various corallines can and do exist elsewhere, this study was restricted to the species within and around these deep tide pools. Experimental and observational data were collected to: (1) quantify the differential distributions of the various coralline morphotypes, (2) determine the factors that could be responsible for their distributions, including physiological stress, consumers, and recruitment, (3) examine the relative ability of the three forms to resist environmental stress, (4) determine the susceptibility of each of the three morphotypes to specific herbivores, and (5) study aspects of the relative abilities of the coralline forms to compete for primary space.

STUDY SITES AND ORGANISMS

Two sites on the central Oregon coast (Boiler Bay and Yaquina Head), two sites on the outer coast of Washington (Shi Shi Beach and Waadah Island), and three sites in the San Juan Archipelago, Washington (Cattle Point and Pile Point, San Juan Island, and Deadman Island) were selected for study (Fig. 1). At the Oregon sites the rocky substratum varied from a mudstone/siltstone to sandstone conglomerate. This substratum is quite porous and holds water. There is a mixed semi-diurnal tide with the lowest low tides occurring early in the morning in the spring and summer and late in the afternoon and evening in the fall and winter. The sites on the outer coast of Washington are similar in both substratum type and tidal regime. The San Juan sites, however, are characterized by hard, less porous rock which cannot hold water (Dayton, 1975), and

IMPORTANCE OF FORM IN CORALLINES

107

although there is a mixed semi-diurnal tide, the timing of the tides is different from the outer coast sites. In the San Juan Islands the lowest low tides of spring and summer are mid-day when temperatures are warmest, and, in the fall and winter, they are late at night or before dawn when temperatures are the coldest (U.S. National Ocean Survey, 1980).

.Y

P

ss-

.:_ *

i

t \

r-----l\

CFW

Y :;

NP

Fig. 1. Map of study sites in Washington and Oregon: indicated on the map are the locations of the two sites in Oregon, (1) Boiler Bay (BB), near Depoe Bay (DB), north of Cape Foul Weather (CFW); (2) Yaquina Head (YH)near Newport (NP); the outer coast ofwashington sites, (1) Shi Shi (SS), and (2) Waadah Island (WI); and the sites in the San Juan Archipelago, Cattle Point (CP) and Pile Point (PP) on San Juan Island (SJI), and Deadman Island (DI) next to Lopez Island (LI).

At all sites, only the larger and deeper pools were examined. These pools ranged in depth from 0.3 to 1.5 m and in width from 0.6 to 10.0 m. The coralline species present were: (1) the finely branched upright Corullina vancouveriensis Yendo, (2) the coarsely branched Calliarthron tuberculosum (Post. & Rupr.) Dawson, and Corallina ofl?cinalis var. cilensis (Dec.) Kutz (this species was very rare and when present, was mixed with Calliarthron, therefore, I lumped it with Calliarthron),

108

DIANNAK. PADILLA

and (3) approximately six species of crustose corallines primarily in the genera Pseudolithophyllum, Lithothamnium, and Lithophyllum (the taxonomy of this group is currently under revision, R. Steneck, pers. comm.). The major herbivores in this system include two coralline specialists Acmaea mitra Rathke (a limpet, 2 to 5 cm) and Tonicella lineata (Wood) (a chiton, 2 to 5 cm), and two generalist herbivores Notoacmea scutum (Rathke) (a limpet, 2 to 6.5 cm) and Katharina tunicata (Wood) (a chiton, 7 to 12 cm) (Kozloff, 1973; Ricketts et al. 1968). The purple urchin, Strongylocentrotus putpuratus (Stimpson), was also found at these locations, but other work (Irvine, 1973) indicated that they generally do not eat coralline algae. Although there are other herbivores higher in the intertidal or found on emergent substratum at these same tidal heights, these were the only herbivores that were both regularly found and abundant at all sites.

METHODS DISTRIBUTIONALDATA Distributional data for coralline algae within and around large, deep tide pools were taken at all sites. Areas were sampled with a 20 x 40 cm quadrat subdivided into 5 x 5 cm squares. These were scored according to the major species occupying that area (0.5 for each species if it was half occupied by each). All additional species present were recorded. This sampling method was compared to one using 100 randomly placed dots over the same area. The two methods were consistent with 5% error. Densities of all animals were also determined for the same quadrats. Four quadrats (or as many as would fit) roughly centered on the four compass directions were sampled in each of three zones for each pool. The zones were defined as follows: (1) the range of the distribution of Corallina ( = Cv-zone), (2) the range of the distribution of Calliarthron ( = Call-zone), and (3) the range of the distribution of encrusting corallines (= L&ho-zone). Initial observations indicated that the morphotypes of corallines were differentially distributed between these zones. Nine pools in Oregon were examined four times per year (during each season), for two consecutive years (1979, 1980). Fifty-four pools in Washington (42 in the San Juans and 12 at the outer coast) were sampled once during the summer of 1980. TRANSPLANTEXPERIMENTSAND SUSCEPTIBILITYTO DESICCATION Transplant experiments were conducted to determine the relative susceptibility of the three morphotypes of corallines to physiological stress and to determine the role of physical factors in influencing the distributions of coralline algae. Each of the three morphological types of corallines was transplanted to each of the three zones. Chiseled pieces of rock (minimum 4 cm diameter) with intact algae were secured with an epoxy putty to the rock. Experimental treatments were transplanted into the zones in which the alga did not normally occur. Controls were transplanted within the zone in which

IMPORTANCE OF FORM IN CORALLINES

109

the alga was usually found. Transplants were conducted for all three forms in the summer of 1979 and spring of 1980 at both Oregon sites. In the summer of 1980 similar experiments were conducted with Corullina higher in the intertidal ( + 3.0 m) at Boiler Bay, and in the mid to low intertidal ( + 0.8 m) at Cattle Point where it was found in pools and not on the emergent substratum. Initial distributional and transplant data suggested morphology was important in influencing desiccation resistance in these algae. The finely branched coralline, Corullina, appeared the most resistant to desiccation stress. Therefore to determine the mechanism by which Corallina avoids desiccation outside of tide pools in the mid intertidal region of Boiler Bay, a series of field and laboratory experiments was conducted. Six replicate experimental plots 15 x 15 cm were marked with an epoxy putty as were adjacent controls. The experimental plots (tidal height + 0.6 to 0.9 m) were thinned evenly so that half of the algal branches in the plot were removed. In the six removal controls, half of the corallines were removed from one side leaving the other half intact. Six nonmanipulated controls were also followed. To determine the water-binding ability of each of the morphological types, 10 individual fronds from each of the three morphotypes were dipped in water, shaken twice and then weighed. This procedure was repeated three times for each individual specimen to ensure equal treatment was given to each alga. The algae were placed in an 80 ‘C drying oven for 48 h, and then in a desiccation chamber for 12 h. Dry weights were determined and the amount of weight lost due to water was calculated. Differences between the three forms should be an indication of their relative abilities to bind extracellular and intracellular water, both of which would affect desiccation resistance (Dromgoole, 1980). HERBIVORE DIETS AND FEEDING EXPERIMENTS

Individuals of the four herbivore species were collected from large coralline pools in the field and either isolated for the collection of fecal material or dissected to determine gut contents. This information was used to determine the natural diets of these species for comparison with the laboratory feeding experiments. Animals were collected at the Washington sites during the summer of 1980 and the Oregon sites through all seasons from the summer of 1979 through the winter of 198 1. Laboratory feeding experiments were conducted at the Friday Harbor Laboratories on San Juan Island, Washington, to determine the relative susceptibility of algal species to the different herbivores. The four herbivore species and the three algal types were collected from tide pools on San Juan Island and used the same day in feeding experiments. Each animal was placed with a single alga in a l-liter, plastic container with four sides replaced by plastic screening to increase water flow. All containers were exposed to natural lighting conditions in an outdoor running sea-water tank. Four to 10 replicates of each alga/herbivore combination were run. Each trial lasted from 25 to 40 days. Single herbivores were used in each trial and no animal was used twice. The crustose algae used in the experiment were single pieces, (3 to 4 cm across)

DIANNA

110

K. PADILLA

attached to a small piece of rock. Articulated species were presented as entire branches with little or no rock attached at their bases. Fecal pellets were used to determine the diets of experimental animals. Fecal materials from the first 24 h were removed from the containers and used to determine what these animals had fed on in the field. After 3 days (to allow herbivores to clear their guts), fecal materials were collected every 3 days for a period of at least 3 wk. These were examined with a microscope to determine contents, and tested with dilute hydrochloric acid for the presence of calcium carbonate. If corallines had been eaten, the undigested calcium carbonate would be present in fecal pellets. Field feeding experiments were conducted with Katharina and ~ora~~~nadue to inconsistent laboratory results. These experiments were conducted at Boiler Bay in the Fall of 1980. Stainless steel cages 15 x 15 cm were attached with stainless steel screws to rock covered 100% by Coral&a. The area was censused visually and photographs were taken to check estimates of percent cover. The edges of the cages were sealed with an underwater epoxy putty. Three replicates were placed at a tidal height of + 0.85 m. Each replicate had two experimental cages each enclosing a small Katharina, (3.5 to 4.5 cm), and one control cage excluding Katharina for a total of six enclosures and three exclosure controls. Two cages were lost due to waves, leaving four exclosures and three controls. After 4 wk the cages were removed and plots were monitored as at the beginning of the experiment. All animals were collected and gut contents were examined. Caliiarthron was the only coralline not eaten by any of the herbivores. It was hypothesized that the structure of this alga was responsible for its protection from consumers. This was tested by grinding Cal~iarthronin a blender (750 g algae with 300 ml filtered sea water) to make a finely ground algal “soup”. This algal homogenate was then mixed with an equal amount of 4.5 “/, marine agar solution (resulting in a 2.25 y0 marine agar solution) and poured into glass Petri dishes. These Calliarthron-enriched agar plates were s~~t~~y prostrate rather than erect or cylindrical like the entire plant. Controls consisted of 4.5% marine agar mixed with an equal amount of sea water. An herbivore of each species was placed with each of the two treatments to test whether the herbivore would feed on ground up Calliarthron. Each treatment had four replicates for each herbivore species and four agar controls Animals were isolated for 48 h to clear their guts before the experiments began. Individual herbivores were scored during the course of the experiment (4 days) as to whether they: (1) did not feed or sample the substratum ( = no fecal pellet production during the course of the experiment), (2) sampled substratum but did not feed ( = produced one to two fecal pellets), or (3) fed on the experimental material ( = produced > 15 fecal pellets). HERBIVORE

EXCLOSURES

flerbivore exclusion fences were placed in all zones of poois at Boiler Bay and Yaquina Head in the summer of 1979 to determine the role of consumers in controlling

IMPORTANCE

OF FORM IN CORALLINES

111

the distribution of coralline algae. Exclosures made of 6 mm plastic mesh and lined with 3 mm mesh, 18 cm across and 8 cm high were attached to the rock surface with either masonry nails or stainless steel screws. Lower edges of the fences were sealed with an epoxy putty. Fence controls consisted of open half-fences aligned with the open side perpendicular to the edge of the pool, and attached in a similar fashion. Unmanipulated controls of an equal area were also monitored. Four exclosure replicates and two of each control were placed in the Cv-zone, and six exclosures and three of each of the controls were placed in both the Call-zone and the Litho-zone. The fences were effective against all herbivores 2 3 mm. The small herbivores entering or settling within the exclosures were periodically removed by hand every 2 to 4 wk. These were monitored approximately every 4 wk. URCHiN

REMOVALS

To test the effects of urchins on the distribution of corallines, the purple urchin, Strongylocentrotuspurpumtus, was removed from two tide pools (0.7 m across and 0.6 to 0.8 m deep) containing corallines. Removals were made during the summer (August and June). Three adjacent pools of similar size with urchins present served as controls. The distributions of the corallines were monitored over time (l-2 yr, 1979-1981). Urchins could have direct effects on the corallines, or secondary effects (e.g., influencing other algae in the pools which might affect coralline distributions). OVERGROWTH

COMPETITION

One aspect of competitive ability, dominance in overgrowth interactions, of the three morphotypes in tide pools was assessed. Overgrowth interactions were examined in pools on San Juan and Deadman Islands and Boiler Bay and Yaquina Head. The winner of an overgrowth was designated as that individual overgrowing the other. Ties were scored for two individuals abutting but neither overgrowing the other. RECRUITMENT

PATTERNS

To examine the role of colonization in affecting the distribution patterns of corallines in tide pools, three artificial tide pools were placed in the intertidal region in areas surrounded by coralline algae. These artificial pools were slip-on pipe caps made of high grade polyvinyl chloride plastic, 20 cm in diameter and 20 cm deep. The interiors of the pools were roughened with sand paper and they were attached to the substratum using metal braces, concrete nails, and an epoxy putty. Colonization within these pools was followed for 12 months.

56.4 (23.8)

Litho

80.1 (14.7)

28.8 (16.5)

Percent of total depth

TABLE I

0 (0)

0.5 (1.2)

70.3 (11.5) 15.8 (5.8)

(8.4)

(9.2)

(0)

5.0

19.2

55.8 (10.7)

0

40.0 (23.0)

0

Bare rock

(0)

cover

(7.4)

Crustose corallines

primary

Oregon

29.2

Calliarthron

i(s) percent

of tide pool data for the central

13.3 (8.2)

20.0 (7.7)

28.0 (19.6)

Otherd

coas@‘.

54.8 (25.3)

15.3 (13.9)

(6.0)

7.8

I(s) No. m herbivores

’

3s)

(1.5)

4.5

(2.5)

2.8

(3.6)

herbivore

d 4(s) pool depth = 76.6 cm (44.5); X(s) pool width = 1.5 m (1.7); n = 9 pools. Data entries are the average (7) above and the SD (s) below in parentheses. h There was no significant correlation between pool depth, pool width and any other factors. c Zone depth is measured from the surface of the water in the pool. -’ Others includes (in order of decreasing abundance): (1) Anthopleuro wnfho,yammica, (2) sponges. (3) red. brown and green non-calcified algal crusts, and (4) red and brown bladed algae.

20.2 (14.4)

Call

zone depth’

Z(s)

Summary

6

g F

F

>

IMPORTANCE

OF FORM IN CORALLINES

113

RESULTS

DISTRIBUTIONAL

DATA

Sampling at the Oregon sites confirmed my initial impressions that finely branched corallines were found out of pools, coarsely branched corallines were limited to the upper portions of the pools and crusts were found over the remainder of the pool (Table I). The exact depth or width of the Call-zone and the depth to which the crusts extended in pools was not constant between pools. However they did remain constant over a 29 period within the same pool. A high variance of herbivore abundance was also found, not only between pools (Table I) but also within a single pool over time (Table II). Consistent zonation patterns were found at the other sites. However, in the San Juan Archipelago the Cv-zone was not on the emergent substrata, but rather at the uppermost edge of the pool at the air-water interface (Table III). Another major difference between the Oregon sites and the San Juan Islands sites was the presence of two species of sea grass, Phyllospudix scouleri and P. torreyi. In the San Juan Islands Phyllospudix was found in every large pool that contained corallines. The rhizomes were usually found only in the Call-zone but occasionally extended lower, into the Litho-zone and were found covering the bottom of some pools. Culliurthron was less abundant in the San Juans than in Oregon (Table III) and when present was found immediately below the Phyllospudix. The average herbivore densities were higher in the San Juans, but these differences were not significant due to the high variance in herbivore abundance (analysis of variance). The outer coast of Washington was similar to the San Juans with the presence of PhyZlospudixin most, but not all coralline pools (Table IV). There was an additional articulated coralline alga, Bossiellaplumosu, in pools, particularly at Shi Shi Beach. This species, however, does not occur in the pools studied in Oregon and was rare in the pools studied in the San Juan Islands. Therefore, such pools were not included in the analysis. Of the pools that did not contain Bossiellu, the patterns of distributions were similar to the San Juans. TRANSPLANT

EXPERIMENTS

AND SUSCEPTIBILITY

TO DESICCATION

The transplant experiments showed that all morphotypes of coralline algae could live successfully in both the Call-zone and the Litho-zone (Table V). However, only Corullinu was able to survive outside of pools (in the Cv-zone) in the mid- to low intertidal region of Oregon. When Culliurthron or coralline crusts were transplanted to the emergent substratum they turned white, died and flaked off of the rock within a few days. In the high intertidal in Oregon and in the San Juans where Corullinu occurs in pools, transplants out of pools died within a matter of days while those in pools survived (Table V). Plots of Corullinu which were thinned to one half of the natural densities bleached and died within several weeks (Fig. 2). In all cases where the corallines had been thinned

1 2 3 Cl RI c2 c3 R2 Al A2 A3

Pool

Date:

50.2 25.3 40.2 12.5 100.0 137.5

(3) (2) (2) (1) (2) (3)

Aug. 79 41.7 (2) 55.0 (3) 77.0 (4)

May 80

0 0 25.0 62.5 125.0 289.3 150.9 (2) (2) (2, (I) (1)

m

188.7 (1) 106.9 (1) 75.5 (1)

(3) (2) (3) (1) (2, (1)

446.6 (2) 377.4 (1) 6.3 (1)

12.7 10.0 20.0 12.5 37.5 12.5

Oct. 80

’ (number of herbivore

10 Sept. 80

of herbivores.

9 Sept. 80

Number

421.4 (2) 106.9 (1) 44.0 (3)

Nov. 80

species)

673.0 (1) 132.1 (1) 6.3 (1)

Mar. 81

XI

50.0 (1) 0

Ma!

Herbivore densities within pools were monitored over time at Boiler Bay, Oregon: the densities of herbivores varied from pool to pool and from sampling period to sampling period within the same pool; the herbivores in the pools included Acmaea mitra, Katharina tunicata, Notoacmea scutum. Strongylocentrotus purpuratus, Tonicella lineata, and small limpets which were not identified to species.

TABLE 11

w j;

30.5 (15.2)

Litho

96.5 (6.9)

17.5 (17.9)

(i)

(Z)

Coral&a

0.8 (2.3)

17.4 (25.2)

Calliarthron

74.1 (19.9)

15.0 (21.2)

Crustose corallines

(1:::)

(1:::)

89.0 (13.4)

Bare rock

17.3 (15.5)

63.6 (31.7)

OtheP

106.3 (100.2)

21.0 (14.6)

(6.5)

1.4

X(s) No:m-’ herbivores

(:::)

0.2 (0.8)

(0.2)

0.1

X(s) herbivore species richness

a X(s) pool depth = 31.2 cm (14.7); X(s) pool width = 1.4 m (0.7); n = 42 pools. Data entries are as in Table I. b There was no significant correlation between pool depth, pool width and any other factors. ’ Zone depth was measured from the top of the water in the pool. d Other includes (in decreasing abundance): (1) Phyllospadir, (2) sponges, (3) non-calcified red and brown crustose algae, (4) red bladed algae, and (5) bryozoans.

3.9 (3.9)

2.8

(4.9)

1.0

Percent of total depth

(1.7)

Call

cv

Zone

Z(s) zone depth”

X(s) percent primary cover

Summary of tide pool data for the San Juan Archipelagoa*b.

TABLE III

51.4 (26.4) .--..____-

(7.4)

8.8

(3.7)

3.0

(3.4)

(i, -_

(i,

17.0 (14.0)

98.7

(iI

21.4 (18.0)

6.0 (7.6)

(0)

0

57.9 (69.6)

Cailiarthron

Corallinu

Percent of total depth

X(s) percent

71.7 (24.3)

15.0 (11.2)

(iI

Crustose corallines

primary

cover

(7.9)

4.3

2.9 (7.7)

50.0 (16.1)

Bare rock

24.0 (24.7)

38.6 (26.7)

28.6 (16.8)

OtheF

of tide pool data for the outer coast of Washingtot@.

67.0 (349.7)

A

(iI

E(s) No. m herbivores

(1.8) .--...

1.9

(i,

(i,

X(s) herbivore species richness

J X(s) pool depth = 52.1 cm (26.6); E(s) pool width = 1.64 m (0.56); n = 12 pools. Data entries are as in Table I. b There was no significant correlation between pool depth, pool width and any other factors. c Zone depth was measured from the surface of the water in the pool. ” Other includes (in decreasing abundance\: I I i f’hvilospadi.~. (?)Anthopkur~r ranthogrurcimmicc~. (3 I non-calcified red and brown algal crusty. 1J) fleshy red algae. (5) sponges. and (6) bryozoans

Litho

Call

CV

Zone __~

X(s) zone depth’ ..~~

Summary

TABLE IV

F

E F

7:

5

$

Ei

IMPORTANCE OF FORM IN CORALLINES

117

TABLE V

Results of coralline transplant experiments: algae were recorded as surviving if they remained alive and attached to the substratum for a minimum of 2 months; they were recorded as dead if they were totally white and flaking off the rock and never regained color before flaking entirely off the rocks (n = sample size).

n

Percent survived

Statistical* significance

100 100 100

NS

0

*

100 100

NS

A. Boiler Bay ( + 0.4 m) Corallina vancouveriensis

Control (Cv-zone) Call-zone Litho-zone Calliarthron

Cv-zone Control (Call-zone) Litho-zone Crustose corallines &zone Call-zone Control (Litho-zone)

NS

0

*

100

NS

100

B. Boiler Bay ( + 3 m) high intertidal C. vancouvenfrnsir out of pool 4 In pool (control) 4

100

C. San Juan Island (+ 0.2 m) C. vancouverimsis out of pool In pool (control) In bottom of pool

0 100 100

4 4 4

0

*

* NS

a Statistically different from controls using a Fisher Exact test.

the remaining branches and basal crust were white after 2 wk. These never regained color, and most had flaked off of the rock by 6 wk. After 8 wk, all manipulated plots were bare rock. In all cases the adjacent controls where no branches were removed showed no effects. Corallina bound more water than did Calliarthron or crustose species (Table VI). Furthermore, there was a lower percent of inorganic matter per .unit weight in Corallina than the other two forms. This could be an indication of the amount of cellular water these species could hold.

118

DIANNA

K. PADILLA

Fig. 2. Results of the Corallina wncouveriensis thinning experments: the white putty dots mark the corners of one of the control (left) and thinned (right) quadrats after 4 wk; in the thinned plot the coralline algae Fare white and flaking off of the rock, while in the control the algae appear intact and healthy.

The weight lost during drying (48 h at 80 C) was calculated for each algal type: entries are an average 10 samples; SD in parentheses. -Percent weight lost .u(s)

Alga Corollina

** c‘. vancouveriensis was significantly Kruskal-Wallis test.

DIETS

AND

xi.3

tuberculosunt

Crustose corallines (two species sampled, no difference between

HERBIVORE

53.4 (h.4)*”

vancouveriensic

Caliiarthron

different

FEEDING

of

(1.9)

27.8 (2.3) them) (P < 0.001) from Calliarthron

and crustose

species using a

EXPERIMENTS

Results indicate that all four herbivores regularly ate some type of coralline. The herbivores can be divided into two general groups, those that ate crustose corallines, and those that ate finely branched corallines. Acmaea, Tonida and Notoacmea ate only crustose corallines. Katharina on the other hand did not appear to have eaten crustose

119

IMPORTANCE OF FORM IN CORALLINES

corallines nor the coarsely branched Calliarthron, but did consume finely branched corallines (Tables VII and VIII). The results of the laboratory feeding experiments were consistent with the natural diets of these herbivores (compare Table VII with Table VIII). Acmaea, Tonicella, and Notoacmea all consistently ate crustose, but not articulated corallines (Table VIII). Several different species of crustose corallines were used, all yielding the same result. Katharina, on the other hand, did not eat crustose corallines or the coarsely branched articulated coralline, but did eat Corallina in four of six trials. The ability ofKatharina to eat Corallina was confirmed in the field feeding experiments. All four experimental animals had Corallina branches in their guts, and the coralline

TABLE

VII

Gut and fecal samples were collected from animals inside and adjacent to coralline dominated pools in Oregon (during all seasons) and from the San Juan Archipelago, Washington (during the summer).

Herbivore species

Percent with

Percent with

Calliarthron

C. vancouveriensis

Percent with CaCO, in feces and/or gut”

intergenicula

100 100

0 0

0 0

0 0

0 85

Acmaea mitra Tonicella lineata Notoacmea swum Katharina tunicata

88 85b

intergenicula

Number of animals 17 32 25 20

a The presence of CaCO, would indicate that coralline algae had been recently consumed. b CaCO, was only in the form of intergenicula of C. vancouveriensis. Intergenicula of articulated forms can be identified to species in guts and feces of herbivores.

TABLE VIII

Laboratory feeding experiments were conducted at the Friday Harbor Laboratories, San Juan Island, Washington: data are given as the number of animals that ate the alga presented (determined by production of fecal material) over the number of animals tested. Algae tested

Herbivore Acmaea mitra ToniceIla lineata Notoacmea scutum Katharina tunicata

Coralline crusts 515 717

lo/lo O/4

Sig.” ** **

*+ *+

Calliarthron

Sig.

l/gb

** ** *+ **

O/9 015 O/6

Corallina O/7 O/7 O/5 416

Sig. ** ** ** NS

a Significant differences between numbers of individuals feeding and not feeding, using a Fisher Exact test, P < 0.05. b One A. mitra made bite marks on the flat surface of the last intergenicula of a branch of CaNiarthron during

the final 3 days of the experiment.

120

DIANNA

K. PADILLA

abundance was reduced to an average of 50% cover as compared to lOOo/, cover for the controls. This difference was significant (P -C0.05; Mann-Whitney U-test). The agar experiments demonstrated that all of the herbivores would feed on Cuiliurthron when it was ground up, although they did not eat this plant when entire. One hundred percent of the animals tested fed on the Cdkzrthron-agar medium (4/4 for each of the four species). None of the agar controls were fed upon or sampled by any of the animals tested (O/4 for each of the four species). This difference was significant (P < 0.05; X2-test). HERBIVORE

EXCLOSURES

Herbivore exclusion cages remained in the field an average of 9 months before being removed by storms. During this time, there was no significant change in the percent cover or distribution of corallines in the exclosures or any of the controls, and the variance between samples was < 5%. URCHIN

REMOVALS

The urchin removal pools did not show any change in the distributions of corallines and remained similar to controls (Table IX). In the removal pools there was a small

Urchin removals: S. purpuratus were removed from two pools at Boiler Bay, Oregon in consecutive years; three matched pools served as controls; pools were monitored four times each year in each season; the table contains the average percent cover (I) of each of the coralline forms. Deptha (cm) of alga in pools Beginning of expt.

After 1 yrb

Percent primary space in pool After 2 yr’

Beginning of expt. ~___~~.~

After I yr”

After 2 yr”

C. vancouveriensis X Removal pools i: Control pools

0 0

0

0

0

0 0

0

0

Ii

0

Calliarthron Z Removal pools X Control pools

17 23

20 21

19 20

17 12

IY IO

20 II

Crustose corallines X Removal pools 2 Control pools

90 87

88 87

XY X6

64 60

60 62

62 56

0

* Depth in pool is measured from the top of the water in the pool. ’ After 1 yr data are averaged for two experimental and three control pools. ’ After 2 yr data are averaged for one experimental pool and two control pools. There is no significant difference between any of the experimental or control pools

IMPORTANCE

OF FORM IN CORALLINES

121

settlement of fleshy algae, however, they occupied < 5 y0 of the primary space after 2 yr and had not affected the coralline distributions. OVERGROWTH

COMPETITION

Crustose corallines were the consistent dominants in overgrowth interactions, winning 100% of the interactions with articulated forms at all study sites in over 120 observations. There were no ties. Crustose forms usually covered the basal crusts of articulated algae to the point where erect branches were found. Subsequently, they have been found to overgrow erect branches as well. RECRUITMENT

PATTERNS

Although the artificial pools were placed in the tield in June, corallines did not colonize until November. Diatoms were the first to colonize these pools and within 2 wk completely covered the interiors of these pools. Shortly thereafter a large number of small limpets (mostly Collissellu strigatellu) were found in the pools and eventually diatom cover decreased. In November there was a settlement of corallines in all three pools. Although the initial settlement was quite dense, the number of individuals decreased greatly over time. This was particularly true in one pool which contained a single, large Tonicella and where all corallines within the pool disappeared; only those on the rim above the water survived. Coralline individuals were too small to be identified until March of 198 1. At that time the remaining plants could ail be identified as Corullina by the size and pattern of small branches that had sprouted. It was impossible to tell if all of the individuals that had previously settled were also of this species. Only this single settlement of corallines was seen in the artificial pools. Nonquantitied observations were also made of the colonization of coarsely branched corallines and crustose corallines on plastic caging material and epoxy putty used in previously described experiments. Calliarthron and C. officinalisvar. chilensis were found only on caging material and only when these were placed within dense growths of these species. Neither was found to colonize caging materials in the Litho-zones or Cv-zones of these same pools, the epoxy putty, or the artificial pools. The crustose corallines did colonize epoxy putty and cage materials in all parts of the pools, but were not found in the artificial pools or outside of pools.

DISCUSSION

A distinct zonation of corallines within and around pools was found. The finely branched corallines were generally found out of the pool while coarsely branched species were found in the pools, primarily in the uppermost portions of these pools. The many crustose species were found predominantly in the lower portions of the pools. Although similar patterns were found at different study sites, there were some

“ Most of the crustos~ ;orailines

look

Bossiella

the same between

sites.

below it

~6 spp. crusts”

Phyllospadix with either Calliarthron or C. oficinatis var. chdlensic

( ‘. vuncouveriensis or

Outer Coast. Washington

26 spp. crusts”

Phyllospadkr with Calliarthron below it

none

San Juan, Washington

Major species in tower portions of pools ~6 spp. crusts”

Major species in upper parts of pools Caliiarthron

C. vuncouveriensis

Major algal species just out of pools

Oregon

._--.--.

TABLE X

Summary of comparisons between pools in Oregon and Washington.

9x

100

0

Percent of pools examined w jPhyllospadix

50

5

0

w:Bossiellu

Percent of pools

0

6

F

g

?

F

5

IMPORTANCE

OF FORM IN CORALLINES

123

important differences (Table X). The presence of Phyllospadix in tide pools in Washington, but not in Oregon, was probably the most striking difference. In small tide pools Dethier (1984) has found that Phyllospadix is the dominant competitor, and has a large influence on other species (probably through shading and/or whiplash effects). However, the corallines in the large pools that I observed were very similar to the large pools in Oregon which do not contain Phyllospadix. The absence of Phyllospadix in the Oregon pools studied may be due to the fact that Phyllospadix requires particular algal species (including Corallina) for colonization. These species live outside of tide pools at the Oregon sites. Phyllospadix germlings are very sensitive to desiccation, which at these tidal heights probably prohibits their successful colonization (Turner, 1983). Further studies need to be done at the Washington sites to determine how biotic and abiotic factors affect these species. Observational and experimental data suggest no single factor was predominantly responsible for the distribution patterns of corallines observed. It does appear that physiological limitations restricted coarsely branched and crustose corallines to tide pools in the mid to low intertidal region. When these two morphotypes were transplanted out of pools they desiccated and died. Only the finely branched coralline survived the desiccation stress of emergent substrata. However, all three of these algal types were able to survive within the tide pools, even at levels where they were not normally found. Therefore physiological stress appears to be restricting the crustose and coarsely branched corallines to tide pools, but cannot explain why Corallina was not found in tidepools in the mid and low intertidal regions, nor why one does not generally find coarsely branched corallines in the lower portions of pools. Morphology appears to be important in influencing the relative abilities of these algae to resist desiccation. Of the three morphotypes, only the finely branched coralline survived the desiccation stress of emergent substrata in the mid to low intertidal region. This is probably due to the bushy structure of this plant and the dense mat it forms on the rock. This morphology appears particularly adapted to trapping water; it is potentially this trapped water, not available to Calliarthron or crusts because of their morphology, that buffers desiccation stress (Dromgoole, 1980). When dense assemblages of Corallina were thinned, thereby reducing their ability to trap water, the plants died (Table V). This adaptation appears to be effective only in the less harsh environment of the mid to low intertidal region in Oregon. In the high intertidal of the same area where desiccation stress is much greater, Corallina was found only in pools, and transplants out of pools did not survive. The same was true for the San Juan Islands. Therefore it appears that the morphology of the finely branched coralline allows it to survive some desiccation and live outside pools in lower, more moderate zones, but is ineffective in more stressful environs. Laboratory experiments support this (Table VI). Corallina had the highest water content of all three morphotypes. The coarsely branched coralline and crustose coralline were not significantly different from each other. The three different morphological forms of corallines also differed regarding suscepti-

124

DIANNA

K. PADILLA

bility to the four herbivore species examined. Again, form appeared to be an important factor determining algal susceptability to specific herbivores. Although all of the herbivore species were found to consume some form of coralline algae in their natural diets, the role of consumers in determining the observed patterns is unclear. Exclusion of herbivores by fences had no effect on the distribution of corallines, even though spores of corallines were available to colonize these areas (they colonized fences and putty during this time period). This could have been due to insufficient time, or the action of tiny herbivores which could not be totally excluded by the fences. Other experiments and observations indicated that most of the coralline types were eaten by herbivores. Acmaea, Notoacmaea, and Tonicella can do severe damage to coralline crusts, completely grazing them down to bare rock (laboratory and field observations). Katharina was capable of severely damaging Corallina in the laboratory and field as well. Although the densities of herbivores varied greatly between zones. this pattern reelects the relative abundance of attachment surface available for these animals. These herbivores cannoteasily hold on to branched corallines, particularly in wave swept areas. Therefore, they are restricted to flat surfaces (bare rock, encrusting species, and basal crust areas). If one recalculates herbivore densities on the basis of available attachment space, there is much less difference between zones. Calliarthron was the only alga not eaten by any of the herbivores tested. The agar experiments support the hypothesis that the morphology of Calliarthron was the key to its protection against herbivores. Ground, it was readily eaten by all of the herbivores, while the agar controls were not. This suggests that it is not low palatability that is protecting Calliarthron from grazers. Gieselman (1980), using a similar technique, found that certain marine algae appear to be protected from herbivory by the use of secondary plant compounds. When these algae were ground they were still not eaten by herbivores. These data suggest that Calliarthron is not protected by such chemicals; although this technique cannot rule out the possibility that the grinding in some way altered or caused the release of some chemical that protects this alga from herbivores. The two limpets and the chiton Tonicella are small relative to the fronds of the articulated corallines. Also, the shield-shaped shells of the limpets restrict their ability to raise up from the substratum. Therefore, it appears as though these herbivores would be most effective at grazing flat surfaces such as prostrate algae, or fronds or blades which could be easily pushed over to lie relatively prostrate or were large enough for the animals to crawl on. The coralline branches, due to their calcification and cylindrical nature, would not be found prostrate nor provide a sufficiently flat surface for these herbivores. Katharina is large enough relative to the size of the Cordlina branches to easily push them over or swallow them whole (pers. obs.). However, Calliarthron appears to be too large and too stiff to be pushed over or swallowed by Katharina and consequently is not eaten. The susceptibility of newly-settled individuals of the articulated forms to the various herbivore species may be very different as the young plants would be small relative to all of the herbivores. The effect of herbivores on newly settled corallines may provide insight into the observed coralline patterns.

IMPORTANCE

OF FORM IN CORALLINES

125

The crustose corallines were susceptible to more types of herbivores than individuals of either of the erect forms. The finely branched Corullina was susceptible to a single large herbivore Kutharinu, and Culliurthron was immune to all common herbivores present. Urchins did not appear to influence the coralline algae. Corallines generally are not eaten by these urchins (Paine & Vadas, 1969; Irvine, 1973). No direct or indirect urchin effects (via controlling other algal competitors) were observed in my experiments. These results differ from those of Paine & Vadas (1969). They removed urchins from a tide pool with quick lime and found the pool was quickly dominated by brown algae. It is unclear whether the pool contained coralline algae and whether their removal technique killed or damaged encrusting algae on the primary substratum. In my experiments, there was no change in the primary cover of corallines; this coralline system does not appear influenced by urchins. Alternatively, consumers may be primarily important to the newly settled algae. Colonization experiments and observations indicated that spore settlement and growth may be limited or patchy. Although only Corullinu settled in the artificial pools, settlement occurred through all levels of the artificial pools. In June 1981 these plants were still growing. There was no preferential settlement or success in any particular area within these pools. Observations made regarding the settlement of corallines on other artificial substrata such as cage materials and epoxy putty give further support to this notion. During the same time period, I found Culliurthron settling on exclosure fences, but only those among Culliurthron plants within the Call-zone. These observations suggest that the distance of spore dispersal may be very limited for some corallines, but not all. Although the artificial pools were in close proximity to all three morphotypes of corallines, only the species not normally found in pools had successfully invaded. These artificial pools were not as large or deep as the observed natural pools. This difference may be important to the corallines, as small pools are much more susceptible than large pools to fluctuations in temperature and other physical and chemical parameters (Stephenson & Stephenson, 1972). Further colonization experiments need to be done within large, natural pools and in all areas of the pools. Regarding overgrowth competitive ability, crustose corallines were superior. In overgrowth competition the crustose corallines won all interactions with articulated corallines. Edge morphology may play an important role in this dominance. Crustose coralline species studied generally have a thicker or raised growing edge as compared to the thin and closely adherent edge of the basal crusts of articulated corallines. Also, crustose corallines with thickened, raised edges tend to overgrow crustose species with thin, closely adhering edges. Since lateral expansion and vegetative growth would be very limited if the basal crust of an articulated plant was overgrown, and because crusts can overgrow erect branches (pers. obs.), this interaction may be very important in determining overall relative competitive ability for the articulated corallines. Although examination of overgrowth interactions at a single point in time represents a static view of the interactions between species, in combination with the distributional data it

126

DIANNAKPADILLA

appears as though crustose corallines were better spatial competitors within pools. They were the spatial dominants and they were the winners of all overgrowth interactions. The dynamics of the interactions, however, must be determined before precise conclusions can be made. The apparent winner of an overgrowth may not always be the spatial competitive dominant. The coralline form best suited to desiccation resistance is the finely branched form, while the form most resistant to the consumers tested was the coarsely branched form. The form that appears to be the best spatial competitor is the crustose form. Therefore, it appears that the morphological type that is best adapted to one of these potential selective agents is not also best adapted to the others. Evidence from the thinning experiments with Coralha and the agar experiments with Calliarthron suggest that structure may be the key to mechanisms by which these organisms are adapted. These results are in agreement with the predictions of the Principle of Allocation proposed by Levins (1968). An underling premise of this theory is that organisms face constraints such that allocation of energy to one structure or function detracts from energies available for others; the Jack-of-all-trades is master of none (Ricklefs, 1973; Pianka, 1978). Alternatively, organisms may be under developmental constraints such that committal to one form or structure would eliminate the potential for others. Such reasoning would predict inverse correlations between successful adaptations to opposing selective factors. This is expected only if structures beneficial for one of these factors are not also good for others. Although this theory and its assumptions are an underlying premise of much current ecological theory, neither has been rigorously tested. However, the rest&s of these experiments support this hypothesis. In conclusion, the three morphotypes of corallines studied have different distributions and demonstrate a distinct zonation pattern within and immediately surrounding large, deep tide pools. No single factor appears to be controlling the zonation of the corallines, at least within the time scale of this study. Desiccation stress appears to restrict coarsely branched and crustose corallines to areas within pools, but does not explain the distribution of corallines within pools. The finely branched coralline is the most desiccation resistant form, evidently as a result of its dense, tufted morphology. The coarsely branched articulated coralline appears to be the most resistant to the consumers tested. The finely branched coralline is eaten by a single herbivore, and the crustose corallines are eaten by three of the four herbivores tested. Again, it appears that morphology is playing an important role in influencing this pattern. Crustose corallines appear to be the best spatial competitors in overgrowth competition, and occupy a majority of the space within pools. The three morphoiogical types of corallines appear to be differentially adapted to various selective agents. ACKNOWLEDGEMENTS

This research was supported by a Grant-in-Aid of Research from Sigma Xi, the ZoRF fund from the Zoology Department at OSU, and M. A. and G. I. Padilla. I would

IMPORTANCE OF FORM IN CORALLINES

127

like to thank Dr. E. N. Kozloff for kindly allowing me the use of the Friday Harbor Laboratories. Useful comments have been given by J. Lubchenco, B. Menge, P. Dawson, B. Mate, H. Cleator, C. D’Antonio, M. Dethier, M. Geber, D. Harvell, C. Marsh, T. Turner, and P. Wilzbach. J. Cureton typed many of the tables. REFERENCES DAYTON,P.K., 1975. Experimental evaluation of ecological dominance in a rocky intertidal algal community. Ecol. Monogr., Vol. 45, pp. 137-159. DETHIER,M.N., 1984. The role of disturbance in tide pools. Ecol. Monogr., Vol. 54, pp. 99-118. DROMGOOLE,F.I., 1980. Desiccation resistance of intertidal and subtidal algae. Bar. Mar., Vol. 23, pp. 149-159. GIESELMAN,J., 1980. Ecology of chemical defenses of algae against the herbivorous snail, Linorina liftorea, in the New England rocky intertidal community. Ph.D. dissertation, Woods Hole Oceanographic Inst./MIT, Woods Hole, Mass., 209 pp. GRIME,J. P., 1977. Evidence for the existence ofthree strategies in plants. Am. Nut., Vol. 111, pp. 1169-l 194. IRVINE,G. V., 1973. The effect of selective feeding by two species of sea urchins on the structure of algal communities. M. S. thesis, University of Washington, Seattle, 93 pp. JOHANSEN,H. W., 1976. Current status of generic concepts in coralline algae (Rhodophyta). Phycologia, Vol. 15, pp. 221-244. JOHANSEN,H.W., 1981. Coralline algae, afirst synthesis. CRC Press, Inc., Boca Raton, FL, 100 pp. KOZLOFF, E., 1973. Seashore lif of Puget Sound, the Strait of Georgia, and the San Juan Archipelago. University of Washington Press, Seattle, 282 pp. LEVINS,R., 1968. Evolurion in changing environments. Princeton University Press, Princeton, N. J., 120 pp. NORTON,T. A., A. C. MATHIESON& M. NEUSHUL, 1982. A review of some aspects of form and function in seaweeds. Bar. Mar., Vol. 25, pp. 501-510. PAINE, R. T. & R. L. VADAS, 1969. The effects of grazing by sea urchins, Strongylocenirotus spp., on benthic algal populations. Limnol. Oceanogr., Vol. 14, pp. 710-719. PIANKA, E.R., 1978. Evolutionary ecology. Harper and Row, New York, 365 pp. RICKETTS, E., J. CALVIN & J.W. HEDGPETH, 1968. Between Pacific tides. Stanford University Press, Stanford, Calif., 614 pp. RICKLEFS,R.E., 1973. Ecology. Chiron Press, New York, 861 pp. STENECK,R.S. & L. WATLING, 1982. Feeding capabilities and limitation of herbivorous molluscs: A functional group approach. Mar. Biol., Vol. 68, pp. 299-319. STEPHENSON,T.A.& A. STEPHENSON,1972. Lye between tidemarks on rocky shores. W.H. Freeman, San Francisco, 425 pp. TURNER,T., 1983. Facilitation as a successional mechanism in a rocky intertidal community. Am. Nat., Vol. 121, pp. 729-738. U. S. NATIONALOCEANSURVEY,1980. Tide tables, high and low water predictions, West coast of North and South America including the Hawaiian Islands. U.S. Department of Commerce, NOAA, 234 pp.