Oecologia (Berl) (1982) 54:23-31 Oecologia ~�9 Springer-Verlag 1982

Disturbance and Predation in an Assemblage of Herbivorous Diptera and Algae on Rocky Shores Carlos Robles Department of Biology, California State University, Los Angeles, CA 90032, U.S.A.

Summary. Speculation about the effects of disturbance in marine benthic communities is often based on competition theory. Disturbances are thought to "provision" numeri- cally depleted or competitively inferior species with re- sources associated with open substrate. However, distur- bances that remove entire assemblages of sessile species also alter trophic structure, and thereby, influence the outcome of predator/prey relations. Aspects of community structure may be determined by patterns of disturbance and preda- tion.

The influence of disturbance and predation on the distri- bution and seasonality of blooms of ephemeral algae and associated Diptera was investigated with field experiments at several rocky beaches in central California. Blooms of ephemeral algae developed on high intertidal rock faces that were subject to severe seasonal disturbances caused by shifting sediment. These were subsequently colonized by the herbivorous larvae of several Diptera species for predictable periods each year. Other areas, without blooms, were not so disturbed.

Experiments were done to determine if seasonal blooms were caused by seasonal disturbances that remove predators which otherwise might prevent the establishment of the Diptera/algae assemblage. The predators were crabs and limpets which eat both algae and larvae while foraging. Blooms of algae and larvae did not develop when limpets were transplanted to disturbed areas in periods between disturbances. Adjacent control areas did support blooms. Transplanted limpets did not survive periods of burial. When both limpets and crabs were excluded from treatment plots in undisturbed areas, blooms developed where they would not otherwise have occurred; controls remained un- changed. Crabs and limpets differed in their effects on this assemblage. Crabs recruited quickly to the site of a bloom, but did not crop algal cover as closely, nor decrease larval density as much as the slowly recruiting limpets.

The results suggest that disturbances favor blooms of some species by reducing predation. Severe localized distur- bances increased the variability of the upper shore commu- nity by creating a patchwork of differing predator/prey abundances.

Introduction

Natural disturbances take many forms in marine communi- ties. Coral reefs are subject to innundations of fresh water

(Connell 1973, 1978) and sediment (Glynn 1976), destruc- tive swells caused by storms and earthquakes (Grigg and Maragos 1974; Stoddard 1972), and exposure to very low tides (Glynn 1976; Loya 1976). Soft-bottom communities are disrupted by surge during storms and ploughing by foraging predators (Woodin 1978). Rock surfaces in shal- low subtidal and intertidal are scoured by moving ice (Day- ton et al. 1970) and battered by storm swells, surf-driven cobble, and drift logs (Dayton 1970; Kitching, 1937; Rfitzler 1970; and Sousa, 1979).

Such disturbances are often thought to influence com- munity structure by interrupting "the process of competi- tive exclusion" (Menge and Sutherland 1976; Connell 1978). How this occurs depends in part on the area, fre- quency, and severity of disturbance (Levin and Paine 1974; Connell 1978; Paine and Levin 1981). Disturbances of inter- mediate severity, which kill only some species in an area, increase local diversity if these species are the most numer- ous or competitively dominant (Connell 1978). More severe disturbances, which kill all species, may open new patches of substrate to colonization. These disturbances are thought to increase regional diversity when the initial colonists in newly opened patches are competitively subordinant species (e.g. Dayton 1971). Such "fugitive species" (Hutchinson 1951) persist, despite mortality from subsequent distur- bances and superior competitors, by virtue of short genera- tion times, high vagility, and high fecundity. These charac- teristics permit rapid colonization and growth to maturity, in the relatively brief period between the opening of a patch and the colonization of competitive dominants. Thus, as- pects of community development are thought to result from differences in the patterns of recruitment and competition among colonists.

The effects of disturbance on predator/prey relations have received less attention from community ecologists, but may be no less important. There is evidence that frequent disturbances of intermediate severity increase prey densities by continually removing predators (Menge and Sutherland 1976) or by hampering their foraging activities (Menge 1978). Other intermediate disturbances, those which period- ically reduce prey abundances relative to predators, may limit the evolution of predator efficiency (Connell 1971). In this case periodic disturbances occasionally deprive pred- ators of resources, thereby preventing predator specializa- tion and hence efficiency. Prey populations presumably ob- tain greater abundances over the long run despite the short- term impact of disturbance directly on their populations.

0029-8549)82/0054/0023/$01.80

24

More severe periodic disturbances cause the removal of entire species assemblages (e.g. Daly and Mathieson 1977). In this instance, vulnerable but dispersive prey spe- cies may become abundant in the lag between the distur- bance and the recruitment of the predators (Connell 1971). When localized, these disturbances presumably result in a patchwork of predator/prey abundances, and in this sense, increase the variability of communities in a region.

Therefore, disturbances potentially affect various as- pects of community structure by mediating the intensity of predation. But few studies specifically consider patterns of disturbance and predation. The form and generality of this relationship remain to be explored in different marine environments. The present study is an investigation of how severe periodic disturbances interact with predation to con- strain the distribution and seasonal occurrance of an assem- blage of herbivorous dipteran larvae and ephemeral algae in an upper rocky shore community (above +1.0 m MLLW).

Natural History of the Diptera/Algae Assemblage

The species studied include a little known group of marine Diptera and their algal forage. The Diptera are the midges Paraclunio alaskensis Coquillet, P. trilobatus Kieffer, and Saunderia marinus Saunders; the tipulid Limonia marma- rata Osten-sacken; and the cycloraphan fly Noeticanace ar- naudi Wirth. The larvae live in tubes spun on the rock surface, are subject to periodic inundations of the tides, and graze on algae during low tides. With the exception of Nocticanace sp., which apparently graze on films of dia- toms and other microscopic algae (per. obsr.), the adults of these species do not feed, and probably do not live beyond the span of a single low tide (Saunders 1928).

The ephemeral (short-lived) algae include species of mi- croscopic diatoms and blue-green algae; filamentous algae, Urospora sp. and Bangia sp. ; and foliose algae, Enteromor- pha spp., Ulva spp., and Porphyra spp. In the Pacific North- west of the United States these algae often occur as seasonal blooms briefly covering high intertidal rock surfaces in a dense mat (Castenholz 1961; Lubchenco and Cubit 1980). Algal spores initiating new blooms apparently immigrate from older blooms or cryptic populations. In addition to the upright thali, Urospora, Bangia, and Porphyra have cryptic phases that bore into rock, wood, bits of shell, or crustose algae (Abbott and Hollenberg 1976; Conway et al. 1976; Sommerfeld and Nichols 1970). Lubchenco and Cu- bit (1980) propose that such heteromorphology is an adap- tive response to variable environments. Cryptic stages en- dure periods of intense herbivory and harsh physical condi- tions; apparently, upright phases allow higher reproduction and growth during favorable periods.

The greatest densities of grazing dipteran larvae occur during seasonal blooms of the ephemerals (Saunders 1928; Leonard 1972; Morely and Ring 1972). In central Califor- nia I recorded densities during blooms as high as 44, 58, 48, and 52 per 50 cm 2 for Paraclunio spp., Saunderia, Li- monia, and Nocticanaee, respectively. A combination of flight and quick development afford the Diptera good colo- nizing ability. Oviposition begins at the start of a bloom on rocks supporting only a thin film of microscopic algae and sporlings. Within a month thousands of larvae may colonize a square meter (Robles and Cubit 1981, for a de- tailed sequence of larvae and algal succession). With the

exception of Limonia, which was not abundant enough to form recognizable cohorts in field experiments, all species were found to mature in less than three months.

Like the algae, the larvae apparently have cryptic popu- lations that, in addition to those in older blooms, serve as a source of immigrants to new blooms. Long and careful searching reveals few larvae living in areas perennially bare of macroscopic algae. The tubes of these larvae are secreted in grooves along the sides of barnacles, in tiny rock crevices, and within tufts of the alga Endocladia muricata.

The most conspicuous invertebrates sharing the upper intertidal environment with the assemblage are potential predators. Acmaeid limpets, Collisella digitalis and C. sca- bra, littorines, Littorina planaxis and L. scutulata; and the grapsid crab, Pachygrapsus crassipes are among the most common grazers in the upper intertidal zone of California (Ricketts and Calvin 1972). I observed that foraging crabs and limpets scrape, rasp-up, and injest the eggs and larvae along with the algae. The larger grazers therefore act as generalized predators on the assemblage. Littorines may have a similar effect, but this was not directly observed.

Patterns of Disturbance and Predation

I made regular observations among four central California sites: Slide Ranch (37~176 Pillar Point (37~176 Duran State Beach (38~ - 123~ and Montara (37~176 Observa- tions at Slide Ranch and Duran were made at intervals not longer than three months over a period of 18 months. Observations were made at Pillar Point at monthly intervals for a period of 18 months followed by weekly observations at both Pillar Point and Montara for a period of 13 months. Observations were made at all hours of the day and under tidal conditions ranging from awash to fully exposed.

The distribution and timing of blooms were observed to correspond to the occurrence of severe physical distur- bances. Summer blooms occurred at Slide Ranch after high intertidal rock faces were pummeled by surf-driven cobble in winter and spring storms. At Duran State Beach and Pillar Point, high intertidal boulders were abraided by the summer build-up of a sand berm. Algal blooms began in winter, when turbulance swept away the sand. Winter storms thus played different roles among these beaches de- pending on the character of the sediments and the local topography of the beaches. Upper intertidal rocks that were free of such disturbances did not support blooms. Montara and certain areas of the other beaches remained free of sediments and developed no blooms, regardless of whether they were exposed or protected. The distribution of blooms of the assemblage seldom overlapped that of dense popula- tions of the larger grazers, In contrast to the Diptera/algae assemblage, limpets, littorines, and crabs were abundant only in areas free of abrasion.

I documented relative abundances of the molluscs and the assemblage with a simple field survey. Sets of 14 to 30 quadrats, each 25 c m 2, w e r e sampled in disturbed and undisturbed areas of Slide Ranch, Pillar Point, or Montara. The quadrats were placed at random within a 1 m z grid or continuously along a vertical transect. Per cent cover of algae was estimated in 5% intervals, the numbers of molluscs counted, and fly larvae removed for later identifi- cation. Sampling in disturbed areas was done in periods between disturbances, in undisturbed areas, at arbitrary

Table 1. Locations, type of disturbance, sample size, and dates of sampling in the field survey

Sample Location Disturbance No. of quadrats Date

A Pillar Point shifting sand 14 February 6, ]976 B Pillar Point shifting sand 20 April 15, 1976 C Pillar Point shifting sand 20 September 10, 1975 D Slide Ranch surf-driven cobble 30 September 19, 1975 E Slide Ranch surf-driven cobble 24 October 31, 1975 F Pillar Point none 20 September 12, 1977 G Pillar Point none 20 November 25, 1977 H Montara none 20 October 5, 1978 I Montara none 20 October 5, 1978 J Montara none 20 October 5, 1978 K Montara none 20 October 5, 1978

25

Table 2. Per cent cover of algae and barnacles in field survey. The first number in each column is mean per cent cover; the second, standard error

Sample Disturbance Micro Urospora Bangia Enteromorpha Porphyra Balanus Chthamalus

A + 10.8/3.0 0.0/0.0 2.4/1.5 10.0 B + 10.1/1.6 37.0/7.5 4.4/1.7 6.5 C + 53.3/7.2 20.0/6.0 2.8/1.2 17.3 D + 11.3/3.0 0.0/0.0 35.3/8.3 29.7 E + 54.6/6.8 0.0/0.0 8.3/4.2 12.0 F - 0.0/0.0 0.0/0.0 0.0/0.0 0.0 G - 0.0/0.0 0.0/0.0 0.0/0.0 0.0 H - 6.0/5.0 0.0/0.0 0.0/0.0 0.0 I - 2.5/4.1 0.0/0.0 0.0/0.0 0.0 J - 99.5/0.5 0.0/0.0 0.0/0.0 0.0 ~ : - 0.0/0.0 0.0/0.0 0.0/0.0 0.0

%8 0.0 '2.3 I0.0 '4.5 1.5 '7.2 24.3 '5.2 13.8 '0.0 0.0 '0.0 0.0 '0.0 0.0 '0.0 0.0 '0.0 0.0 '0.0 0.0

/0.0 3.4/2.2 0.0/0.0 /3.6 0.0/0.0 0.0/0.0 /1.0 1.8/1.1 0.0/0.0 /6.5 0.0/0.0 0.0/0.0 /4.2 0.0/0.0 0.0/0.0 /0.0 4.3/1.3 0.0/0.0 /0.0 1.3/0.9 1.3/0.6 /0.0 0.0/0.0 0.0/0.0 /0.0 0.0/0.0 0.0/0.0 /0.0 0.0/0.0 0.0/0.0 /0.0 0.3/0.3 3.0/1.3

Table 3. Densities of larvae, limpets and littorines in field survey. The first number in each column is mean number of individuals per 25 cm 2; the second, standard error

Sample Disturbance Total larvae C. digitalis C. scabra Total littorines

A + 29.1/3.4 9.0 B + 5.6/1.3 0.0 C + 2.1/0.5 0.0 D + 15.3/2.0 0.0 E + 2.9/0.7 0.0 F - 0.0/0.0 2.5 G - 0.0/0.0 1.1 H - 0.0/0.0 1.6 i - 0.0/0.0 0.2 J - 0.3/0.1 0.1 K - 0.0/0.0 2.5

'0.0 0.0/0.0 0.0/0.0 '0.0 0.0/0.0 0.0/0.0 '0.0 0.0/0.0 0.6/0.5 '0.0 0.0/0.0 0.0/0.0 'o.o o.o/o.o o.o/o.o '0.8 0.0/0.0 0.0/0.0 '0.8 0.5/0.2 0.0/0.0 q .7 0.0/0.0 3.0/2.2 '0.1 0.1/0.1 0.8/0.4 ~0.1 0.0/0.0 0.7/0.6 '0.4 0.1/0.1 2.3/0.6

times afforded by the schedule of experiments. At Pillar Point and Montara, the exact locations of samples were chosen to be adjacent to surfaces used in experiments. The conditions of sampling are summarized in Table 1. The rela- tive abundances of the algae and sessile invertebrates are listed in Table 2. Relative abundances of larvae, limpets and littorines are listed in Table 3.

The other large grazer on these sites, Pachygrapsus cras- sipes, foraged mostly at night. The crabs' motility and secre- tive habits made accurate field censuses difficult. However, in evenings of field work, I observed only newly settled Pachygrapsus in disturbed areas at Pillar Point or Slide Ranch. Adult crabs were nearly always observed in experi- mental and survey areas at Montara, and occasionally in undisturbed areas at Pillar Point.

In the discussion of the study sites I used "d i s tu rbance" to mean the abrasion caused by shifting sediments. This was severe enough to remove all sessile life and the surface layer of rock. Desiccation, wave shock, or other physical perturbations probably affected these areas as well, but only shifting sediments were observed to cause the wholesale removal of populations on a regular basis. Detailed descrip- tions of the effects of sand scouring on populations of lim- pets, barnacles, ephemeral algae and other common upper intertidal species appear in Daly and Mathieson (1977).

The seasonal build-up and loss of an intertidal sand berm is a regular feature of sandy coasts (Bascom 1964). When the berm overlaps onto rock formations the area of disturbance varies from a few square meters to hectares. Along-shore sorting processes and the proximity of source

26

areas (rapidly eroding bluffs and coastal rivers) favor the formation of cobble and boulder fields on some beaches (Komar 1976). Storms can turn cobbles to millstones whose abrasive force may be seen in concavities and potholes in rock faces. Although localized to specific areas within the study sites, such disturbances are common along the central California coast.

A Hypothesis and Predictions

Although caution is warranted in making generalizations about a species' demography in different locations, results of previous studies (e.g. Hiatt 1948; Frank 1965; Nicotri 1974; Choat 1977) suggest that the species of crabs and limpets settle and grow more slowly than species in the assemblage. Unlike the algae and larvae, these larger long- er-lived grazers may not be able to attain high densities when disturbances occur yearly. I f the larger grazers are effective predators, then the observed pattern of blooms could be simply determined by the pattern of disturbance and predation. Stated as a specific hypothesis: Crabs, lim- pets, and littorines prevent the establishment of blooms in undisturbed areas. Elsewhere disturbances are sufficiently severe to remove the larger grazers and allow the coloniza- tion of algae and larvae. I f this hypothesis is true we would predict that:

1) If the larger grazers were moved to disturbed areas in periods between disturbances, then these transplants will thrive and prevent blooms of algae and larvae, and

2) If the larger grazers were removed from undisturbed areas, then dense populations of algae and larvae will devel- op where they would not otherwise occur.

I f the predictions are true, then it would be difficult to explain the distribution of seasonal blooms by any other mechanism than the pattern of disturbance and predation. If blooms were regulated solely by fluctuations in the re- quirements of growth or other factors unrelated to preda- tion, then transplants and removals should have no effect. I f predation, competition, or physical factors exclusive of disturbance were preventing the establishment of the larger grazers in disturbed areas, then the transplants should not survive.

Materials and Methods

Description of experimental sites: All experiments and most field survey and observation

were done at Pillar Point and Montara. Pillar Point is a broad, semi-protected shelf of mudstone projecting along a north-south axis for approximately 1.5 km. The Montara site is a row of rugged granite bluffs aligned on a north- south axis. Its profile rises abruptly from the subtidal to above the upper intertidal zone and receives the full impact of ground swell. Detailed descriptions of the sites appear in Robles and Cubit (1981).

Experimental methods: Most experiments were begun by scraping and scorching natural or artificial colonization surfaces with a putty knife and propane torch to remove all growth. Similar techniques have been used by other au- thors studying algal growth and succession (e.g. Castenholz 1961; Lubchenco 1980).

In some experiments, circular concrete plates (garden "stepping stones") 30 cm dia. and 5 cm thick served as colonization surfaces. To minimize possible effects of con-

Fig. 1. Arrangement of crab exclosure and settling plates. Iron frame (right background) and assembled crab exclosure (center). Plates attached directly to stone (foreground) do not exclude crabs. Plate in right foreground has been circled with copper paint, which excludes limpets

crete on algal growth, plates were cured in the surf for at least two weeks prior to use in an experiment. Patterns of colonization on most plates did not appear to differ greatly from those on natural stone (Robles and Cubit 1981). On a few plates minerals leached through and were deposited in a crust on the lower half of the exposed surface. When such leaching occurred I sampled only the upper half of the plate. For some replicates done at Pillar Point, algal growth seemed to be faster on the plates than on nearby rock surfaces. This may have been the result of the greater porosity of the concrete, which retained water longer than the fine-grained surface of the mudstone.

Previous studies of the effects of biotic interactions in the rocky intertidal usually employed cages to exclude pred- ators or competitors from experimental plots. However, in this study cages might have interfered with insect oviposi- tion, algal spore settlement, and local patterns of drainage and desiccation. Consequently, I employed a variety of ex- clusion techniques that did not include caging. Conclusive tests of hypotheses required the separate or simultaneous exclusion of two groups of herbivores: crabs and limpets. These required different exclusion techniques.

In experiments requiring the exclusion of crabs, the con- crete plates served as colonization surfaces. Each plate was held 8 to 10 cm above the rock surface by a small iron platform fitted with an acrylic plastic baffle (Fig. 1). Crabs were not able to cross the slippery surface of the plastic and reach the colonization surface. Some experimental con- trols involved concrete plates to which the crabs did have access. These plates were attached directly to the stone sur- face. Crabs readily foraged on these plates.

Limpets were excluded from colonization surfaces using a plastic paint enriched with a powder of elemental copper (Cubit 1975, for details). A band of copper paint was ap- plied around concrete figures attached directly to the rock, to the stone around the base of the iron legs of crab exclo- sures, or around plots laid out directly on stone. Control plots on stone were surrounded by a discontinuous ring of copper. I observed no obvious differences in the growth of algae in areas surrounded by continuous copper bands compared to those surrounded by discontinuous bands.

I challenged the predictions by manipulating the densi-

Table 4. Densities of limpets, heights, and dates of replicates in the transplant experiment

27

Replicate Coil&ella digitalis C. scabra Total Limpets Height above Date of 4/25 Cm 2 4/25 Cm 2 ~t/25 Cm 2 MLLW Clearance

A 0.14 0.25 0.39 + 1.0 m November 13, 1977 B 0.60 0.04 0.64 + 1.4 m November 13, 1977 C 0.57 0.00 0.57 + 1.6 m April 19, 1979 D 0.67 0.04 0.71 + 1.4 m April 19, 1979

Fig. 2. Two replicates of the limpet transplant experiment begun April 19, 1979. Limpets have prevented growth in treatment enclo- sures (center circles). A bloom of ephemeral algae, predominantly Enteromorpha sp., covers control plots (outer circles) and the re- mainder of rock surface

ties of limpets and crabs. Littorine snails, though probably important in some areas (Behrens 1971), were not manipu- lated and their densities remained low on the surfaces used in experiments.

Changes in larval and algal abundances were estimated with sets of "mic roquadra t s" randomly spaced on plastic sheets (see Robles and Cubit 1981, for details). With the exception of foliose green algae, per cent cover has been recorded for each genus of macroscopic algae. Determina- tion of foliose green algae in the field was impractical. Since Enteromorpha was by far the most common genus, all fo- liose green algae were scored as "En te romorpha" . When sufficiently abundant, solitary and colonial diatoms, blue- green algae, and the spores of macroscopic algae formed a film. This film was scored as "microalgae ." A little inac- curacy in the per cent cover of microalgae was inevitable since, when thin, this film was indistinguishable from the bare substrate.

A Test of the First Prediction

I tested the first prediction - that the larger grazers can survive and control blooms in periods between disturbances with a transplant experiment. The abundance of algae and larvae in plots to which limpets were transplanted was com- pared with abundances in control plots to which no limpets were transplanted. Treatment and control plots were pre- pared on the surface of a large boulder in the disturbed southeastern corner of Pillar Point. Blooms of algae and larvae occurred naturally on this surface during fall and winter. These areas of stone were free of limpets and crabs, and were subject to the longest periods of burial by shifting sands.

Several square meters of stone were cleared and pairs of copper enamel rings, 30 cm dia., painted directly on the stone. The stone was scraped and brushed, not scorched. By not sterilizing the substrate I hoped to speed algal colo- nization, thus minimizing the chance of shifting sands pre- maturally ending the experiment. After about 10 days a thin film of algae developed. The plots were then dowsed with sea water and groups of Collisella digitalis and C. scabra were transplanted to the center of the treatment plots; controls received no limpets. Grooves had been chis- eled in the center of each plot to provide shelter for newly transplanted limpets. The dates, heights, and limpet densi- ties of four replicates are listed in Table 4. Only a few of the transplanted limpets failed to " t a ke" . Final transplant densities were below the densities recorded on undisturbed surfaces at the same tidal level nearby.

The plots were sampled eight to ten weeks after trans- planting. At this time control areas (without limpets) sup- ported a conspicuous cover of microscopic algae; most treatment plots were completely free of algae (Fig. 2 and Table 5). Larval abundances were similarly affected, being higher in control than treatment plots (Table 5). Microscop- ic and macroscopic algal covers, and larval densities dif- fered significantly between treatments and controls (Mann- Whitney U, P < 0.05).

These results show that the limpets are capable of pre- venting the establishment of the Diptera/algae assemblage

Table 5. Results of limpet transplant experiment. Mean per cent covers of algae and mean densities of larvae are tabulated; standard errors are listed below the means. Limpets were transplanted to treatment enclosures; control enclosures received no limpets

Per cent cover of algae Number of larvae per 50 cm 2

Microscopie Urospora B a n g i a Enteromorpha Porphyra Saunderia Paraclunio

Mean of Treatments 4.5 Standard Error 4.5 Mean of Controls 23.0 Standard Error 8.7

0.0 0.0 5.5 1.0 0.5 0.25 0.0 0,0 5.5 1.0 0.5 0.25

23.0 7.0 39.5 4.0 11.25 15.0 11.7 4.4 4.6 2.8 4.9 6.9

28

Fig. 3. A single replicate in the limpet exclosure experiment at Pillar Point. The right, treatment exclosure is covered by a nearly contin- uous film of Urospora. The control exclosure was prepared just to the left of the treatment exclosure. Four dabs of copper paint mark its location

on these surfaces in periods between disturbances. Dayton (1971) demonstrated that the limpets also reduce the densi- ties of young barnacles.

These results also suggest that disturbance is a primary factor limiting the abundance of adult limpets in this area. Transplanted limpets survived in enclosures for months un- til shifting sands buried them for a period of several weeks. Newly settled limpets were seen in periods between distur- bances, but did not survive to maturity.

Transplants of Pachygrapsus were impractical because their movements over a continuous rock surface were diffi- cult to control without cages.

Tests of the Second Prediction

The second prediction holds that the exclusion of crabs and limpets from undisturbed areas will result in the estab- lishment of abundant algal and larval populations. At Pillar Point the effects of limpets were studied by comparing algal and larval densities in undisturbed areas from which limpets had been excluded with adjacent areas to which limpets had access. Limpets were excluded from 5 treatment plots (continuous copper ring) located between +1.8 m and + 2.3 m M L L W on vertical surfaces along the western edge of the point. Each replicate consisted o f two circular plots, 30 cm in diameter, that had been cleared and scorched. Controls were surrounded by a discontinuous ring of cop- per paint, permitting limpets to enter and forage within the area. The replicates were prepared on November 23, 1977.

Percent cover of algae and larval densities were recorded eight weeks later on January 20, 1978. A thick cover of macroscopic algae developed within most exclosure areas. Control areas remained completely bare of visible algae, as did all of the rock surface outside the exclosures (Fig. 3 and Table 6). The mean per cent cover of microscopic and macroscopic algae, and the mean density of Paraclunio lar- vae differed significantly between treatment and control ar- eas (Mann-Whitney U, P < 0.05). Over the period of obser- vation, grazing and predation by limpets accounted for the rarity of macroscopic algae and fly larvae on these un- disturbed rock faces.

On a few evenings during the course of the limpet exclo- sure experiment, I observed Pachygrapsus foraging within the exclosures. The differences between treatment and con- trol plots were maintained, therefore, despite the levels of crab foraging that occurred at that time and location. Dif- ferences in limpet densities were sufficient to explain the pattern of blooms at this site. This does not rule out the possibility that crabs affected the relative abundance o f spe- cies in the assemblage. When numerous, crabs might severe- ly reduce abundances.

I tested the effect of both crabs and limpets on the assemblage with an exclosure experiment at Montara, a site where dozens of foraging crabs could be seen on any evening. This was a comparison of three groups of concrete colonization surfaces. In the first treatment group the con- crete plates were raised on the crab-exclosure frames de- scribed previously. The stone around the base of these frames was painted with copper so that neither crabs or limpets had access to these colonization surfaces. In the second treatment group, plates were attached directly to the rock and circled with copper paint. Crabs, but not lim- pets, had access to these colonization surfaces. In the third treatment group plates were attached directly to the stone, and not surrounded by a ring of copper paint. Both crabs and limpets had access to these plates. Hereafter I refer to these treatment groups as Group I, II, and III, respective- ly. A replicate consisted of one plate from each treatment group clustered on the same rock face (Fig. 1). With this arrangement I was able to compare the effects of crab pre- dation alone with the effects of simultaneous crab and lim- pet predation.

A total of five replicates were installed between + 1.9 m and + 3.3 m M L L W at Montara. All plates raised on iron frames and all but two pairs of plates attached directly to the stone were cleared on April 11, 1978. Due to equip- ment difficulties, the two remaining pairs were not cleared until two weeks later, April 29, 1978. This asynchrony does not appear to have affected the results. Differences between treatment groups within the two asynchronous comparisons

Table 6. Results of limpet exclosure experiment. Data are mean per cent cover of the algal species and mean densities of species of larvae in the limpet exclosure experiment at Pillar Point; standard errors are listed below the means. Limpets were excluded from treatment areas, but allowed to forage in control areas

Per cent cover of algae Number of larvae per 50 cm 2

Microscopic Urospora Bangia Enteromorpha Porphyra Saunderia Paraclunio

Treatment Mean 45.2 44.0 4.4 0.0 1.2 0.0 19.2 S.E. 14.3 17.8 2.6 0.0 0.8 0.0 2.5

Control Mean 0.0 0.0 0.0 0.0 0.0 0.0 0.4 S.E. 0.0 0.0 0.0 0.0 0.0 0.0 0.4

29

Fig. 4. Differences in algal cover resulting from three different treatments 18 weeks after the initial clearance of these surfaces (see Fig. i for the initial condition of this replicate). The raised plate, excluding both crabs and limpets, supports a diverse cover of microscopic and macroscopic algae; the lower right plate, ex- cluding onIy limpets, supports a continuous cover of "turf"; and the lower left plate, the corttrol, is bare of algae. The small dark object on the lower plate in the left foreground are limpets which recruited to plate in Group III relatively late in the observation period

Algal Cover in Treatment Groups I , Z , I I [

I00

80 T 1

6 C

2 4 6 8 lO 12 14 16 I8

o o 60

2o

o) I I I I : ~ 2 4 6 8 I0 12 14 16 18

I00

8 0

6 0

4 0

2 0

0 0 2 4 6 8 I0 12 ~4 16 18

Week after cleQrance Fig. 5. Changes in the mean per cent cover of microscopic, macro- scopic, and "turf", algae among the tbree treatment groups at Montara. Treatment Group I excluded both crabs and limpets, Group II, only limpets, and Group III served as control. Symbols are placed at the mean of five replicates; bars on either side of the symbols represent one standard error. Triangles represent microscopic algae; squares represent macroscopic algae; and dots represent algal " tur f " . Microscopic algae included species of soli- tary and colonial diatoms, blue-green algae, and the spores of macroscopic species. Macroscopic algae included species of Uro- spora, Bangia, Enteromorpha, Ulva, and Porphyra

span the range of variation found in synchronous compari- sons, and nonparametric statistical tests of differences be- tween treatment groups are significant regardless of whether or not the two asynchronous comparisons are in- cluded. For this reason data analysis and graphic represen- tations are reckoned by week of succession rather than date of sterilization. After clearing, algal per cent cover was sam- pled weekly and larval densities sampled every other week.

Within 18 weeks clear differences in algal cover devel- oped among treatment groups (Fig. 4). In Group I (crabs and limpets excluded), plates developed a thick cover of microscopic and macroscopic algae (Fig. 5). As in Group I, plates in Group II (only limpets excluded), developed high total per cent covers. However, the species composition and general aspect of this cover differed markedly from Group I. Weekly nocturnal observations revealed that after an initial bloom of microscopic and macroscopic algae crabs grazed algae down to a thin film less than 3 mm thick. Once established, this closely cropped turf persisted throughout the period of observation (Fig. 5). The turf ap- peared to be comprised of microscopic algae growing in a matrix of holdfasts of Enteromorpha. In the initial weeks of the experiment, some plates in Group III (no grazers excluded) developed a partial cover of microscopic and macroscopic algae. This was soon discovered by Pachygrap- sus and grazed down to turf. As the weeks passed, however, limpets recruited to the upper surfaces and sides of these plates; algal turf declined until all plates in Group III were completely free of visible algae (Fig. 5).

At the end of 18 weeks, differences among the treat- ments in the proportion of microscopic, macroscopic, and " t u r f " algae were statistically significant (X 2, P<0.005). The total per cent cover also differed significantly among the three treatment groups (X 2, P < 0.005).

The foraging pattern of Pachygrapsus in Groups II and III changed during the course of the experiment. During evening low tides in the initial weeks of observation, crabs could be seen foraging on plates in both groups. However, as limpets recruited to plates in Group III and the cover of turf declined, Pachygrapsus confined its foraging to plates in Group II, from which limpets were excluded. A spot-check done the 18th week revealed a significant differ- ence in the number of foraging crabs between the two groups (Mann-Whitney U, P < 0.05).

Similar turfs were observed to occur naturally at Mon- tara (Robles 1979). These turfs were limited to rock breaks and recently overturned stones, which were free of limpets, but frequently visited by foraging crabs. After foraging, the crabs retreated to deep crevices in nearby rock faces. Apparently, these small-scale limpet removals in an other- wise undisturbed area allowed growth that was soon cropped back by the mobile crabs, just as occurred in the second treatment group.

Grazing crabs and limpets also affected the larvae. The greatest larval densities developed in the absence of crabs and limpets (Group I); larvae were less abundant when only crabs were present (Group II); and were scarce or complete- ly absent when both crabs and limpets were present (Group III). Fig. 6 describes the changes in mean densities of larvae over the 18 weeks of observation. Two genera, Paraclunio and Nocticanace, accounted for all of the larvae collected. For unknown reasons Limonia and Saunderia did not occur in the samples from this experiment. Of a total of 207 Paraclunio larvae collected, 93%, 7%, and 0% were collect- ed from Groups I, II and III, respectively. However, Nocti- canace arnaudi colonized all three treatment groups. Of a

30

14

12

tO

8

6

4

~ 2 o

O o L O

E2

> IO

~b 6 ~4 E 2

E 0 - - 2

P, N ,4 ]]Z

t2

I0

8

6

4

2

Larvas Density in Treatment Groups I,rr,]I[

- - i i ! j _ i , 2 4 6 8 I0 12 14 16 18

4 6 8 I0 12 [4 16 [8

_ _ ~ ~ ~ T . ~ , - d ~ - - - m - ~ - - ~ _ z 4 e 8 i o 12 64 16 le

Week after clearance Fig. 6. Changes in mean density of species of larvae among the three treatment groups at Montara. Treatment Group I excluded both crabs and limpets; Group II, only limpets; and Group III served as control. Symbols are placed at the mean of five replicates; bars on either side of the symbols represent one standard error. Squares represent Nocticanace; dots represent Paraclunio; the other fly genera did not occur in this experiment

total of 159 Nocticanace, 47%, 47%, and 6% were collected from Groups I, [I, and III, respectively. Apparently, Para- chmio were affected by limpets or crabs, but foraging crabs alone had little effect on Nocticanace populations. Both the differences in the proportion of the two genera and total densities differed significantly among treatment groups (X 2, P<0.005).

Differences in the proportion of Paraclunio and Nocti- canace between Group I and Group II may be related to the foraging ability of Pachygrapsus. All larvae collected on plates from Group II were found deep within the relief of the surface of the concrete. These larvae included the smaller instars of Paraclunio and all instars and pupae of Nocticanace, all less then 5 mm in length. Older, larger in- stars and pupae of Paraclunio were found only on plates in Group I. Differences in algal cover between Groups II and III show that crabs do not crop back algal cover as close to the surface as do limpets. The surface of the con- crete may have provided a spatial refuge from crab preda- tion for larvae small enough to fit into its contours.

The extreme differences that obtained between plates of Group I (crabs and limpets excluded) and Group III (no grazers excluded) demonstrate that crabs and limpets to- gether were capable of completely preventing blooms of ephemeral algae and subsequent larva recruitment over the period of study. Since Diptera and algae readily recruited to cleared experimental surfaces, the absence of seasonal blooms in the upper intertidal of Montara can be explained

by the absence of seasonal disturbance that would otherwise remove crabs and limpets in these areas.

Discussion

Early work on seasonality in marine ecosystems attributed seasonal algal blooms directly to fluctuations of the physical requirements of growth water, light, and nutrients (e.g. Johnstone 1908; Johnson and Scutch 1928; Aleem 1950; Lawson 1957). More recent work indicates that seasonal changes in algal standing crops result from an interplay between fluctuating physical and biotic factors (e.g. Casten- holz 1961; Cubit 1975; Cushing 1975). This interaction pro- ceeds by several mechanisms. For example, fluctuating physical resources may change algal productivity. When seasonal fluctuations in primary productivity are suffi- ciently large, herbivores are alternately subjected to periods of starvation and surfeit. Algal blooms occur in these latter periods when algal production is high and herbivore popu- lations have not recovered their maximum densities (e.g. Cubit 1975).

Changing physical conditions - particularly mechanical disturbance also affect herbivore populations directly. At the sites I studied, disturbances in the form of shifting sedi- ments appeared to be the overriding factor causing lags between rates of algal production and consumption. Beaches subjected to disturbances at different times sup- ported blooms at correspondingly different times, regard- less of season. Severe abrasion apparently kept local areas free of dominant predators. Vagile, short-lived species, flourished in periods between disturbances. In this sense, disturbance mitigated the effects of predation and deter- mined the timing of seasonal blooms of some species. Cas- tenholz (1961) and Southward (1956, 1964) propose similar mechanisms to account for changes in ephemeral algal abundance.

Although natural disturbances were not manipulated, alternatives to this "mitigation hypothesis" are implausi- ble. Verification of the predictions contradicts alternatives based on factors other than the pattern of disturbance and predation.

Investigations restricted to a few locations cannot char- acterize all the community dynamics of even a single spe- cies. A variety of factors doubtless constrain the distribu- tion and seasonal occurrence of the algae and Diptera else- where in their ranges. However, severe disturbances are common, and the mechanisms described in this study may operate in many populations.

The vulnerability of individuals in the Diptera/algae as- semblage is apparently compensated by cryptic habit, high vagility, and quick development. The latter attributes are associated with "opportunistic species". However, the op- portunities exploited by members of the assemblage are not times of abundant resources, but rather instances of reduced predation.

Disturbance is most commonly viewed as affecting com- munity structure by interrupting the process of competitive exclusion. In this and several other studies (e.g. Connell 1971; Menge and Sutherland 1976) disturbance influences community structure by interrupting processes of predatory removal. Local disturbances removed dominant predators creating a patchwork of prey densities and successional states (see Robles and Cubit 1981, for a successional se- quence of Diptera and algae). In this sense disturbances increased the variability of the upper shore community.

31

Acknowledgement. This paper is based on a Doctoral dissertation in the Department of Zoology, University of California, Berkeley. I thank my major advisor, Dr. R.K. Colwell, for his support and advice in this work. Dr. John Cubit provided many stimulating discussions and helpful insights into the structure of high intertidal communities.

The quality of this manuscript was much improved by reviews of preliminary drafts by Drs. R.K. Colwell, J.H. Connell, J. Ka- stendiek, and W.M. Murdock. I was helped with the sometimes arduous field work by R. Bowman, E. Lord, M. Lord, W. Sousa, J. Torres, and my father.

References

Abbott IA, Hollenberg GJ (1976) Marine algae of California. Stan- ford University Press, Stanford

Aleem AA (1950) Distribution and ecology of British marine litto- ral diatoms. J of Ecol 38 : 75-106

Bascom W (1964) Waves and beaches: The dynamics of the ocean surface. Anchor, Doubleday, New York

Behrens S (1971) The distribution and abundance of the intertidal prosobranch, Littorina scutulata (Gould 1899) and L. sitkana (Philippi 1845). M.S. thesis, Univ of British Columbia, Van- couver

Castenholz RW (1961) The effect of grazing on marine littoral diatom populations. Ecol 42:783-794

Choat JH (1977) The influence of sessile organisms on the popula- tion biology of Acmaeid limpets. J Exp Mar Biol Ecol 26:1-26

Connell JH (1971) On the role of natural enemies in preventing competitive exclusion in some marine animals and rainforest trees. In PJ den Boer and GR Gradwell editors, Dynamics of Populations, Center for Agricultural Publications and Docu- mentations, Wageningen, Netherlands

Connell JH (1973) Population ecology of reef building corals, pp 205-245. In OA Jones and R Endean, editors, Biology and geology of coral reefs, Vol 2, Biology I. Academic Press, New York

Connell JH (1978) Diversity in tropical rain forests and coral reefs. Science 199:1302-1310

Conway E, Mumford TF Jr, Scagel RF (1976) The genus Porphyra in British Columbia and Washington. Syesis:185 244

Cubit J (1975) Interactions of seasonally changing physical factors and grazing affecting high intertidal communities on a rocky shore. Ph.D. thesis, Univ of Oregon, Eugene

Cushing DH 0975) Marine Ecology and Fisheries. Cambridge University Press, London

Daly MA, Mathieson AC (1977) The effect of sand movement on intertidal seaweeds and selected invertebrates at Bound Rock, New Hampshire Mar Biol 43 : 45-55

Dayton PK (1970) Competition, predation and community struc- ture: the allocation and subsequent utilization of space in a rocky intertidal community. Ph.D. thesis. Univ of Washington, Seattle

Dayton PK (1971) Competition, disturbance and community orga- nization: the provision and subsequent use of the rocky intertid- al. Ecol Monogr 41 : 351-389

Dayton PK, RobilIiard GA, Paine RT, Dayton ZB (1970) Benthic faunal zonation as a result of anchor ice at McMurdo Sound, Antarctica. In: Holdgate MW (ed) Antarctic Ecology, Aca- demic Press, London

Frank PW (1965) The biodemography of an intertidal snail popula- tion. Ecol 46:831 844

Glynn PW (1976) Some physical and biological determinants of coral community structure in the Eastern Pacific. Ecol Monogr 46 : 431-456

Grigg RW, Maragos GE (1974) Diversity and colonization of Ha- waiian coral reefs. Ecol 55 : 387

Hiatt RM (1948) The biology of the lined shore crab, Pachygrapsus crassipes. Pacific Science 2:135-213

Hutchinson GE (1951) Copepodology for the ornithologist. Ecol 32:571 577

Johnson DS, Skutch AF (1928) Littoral vegetation on a headland

of Mt. Desert Island, Maine. I. Submersible or strictly littoral vegetation. Ecol 9:188-215

Johnstone J (1908) Conditions of Life in the Sea. Cambridge Univ Press, London

Kitching JA (1937) Studies in sublittoral ecology. II. Recoloniza- tion at the upper margin of the sublittoral region; with a note on the denudation of Laminaria forests by storms. J of Ecol 25 : 482495

Komar PD (1976) Beach Processes and Sedimentation. Prentice- Hall Inc. New Jersey

Lawson GW (1957) Seasonal variation of the intertidal zonation on the coast of Ghana in relation to tidal factors. J EcoI 45 : 831-860

Leonard SD (1972) The natural history of Paraclunio alaskensis and Paraclunio trilobatus (Diptera, Chironomidae), two inter- tidal flies. M.S. thesis. Humbolt State College, Humbolt

Levin S, Paine RT (1974) Disturbance, patch formation, and com- munity structure. Proc Nat Acad Sci Vol 71, pp 2744-2747

Loya Y (1976) Recolonization of Red Sea corals affected by natu- ral catastrophies and man-made perturbations. Ecol 57 : 278-289

Lubchenco J (1980) Algal zonation in the New England rocky intertidal community: an experimental analysis. Ecol 61 : 333-344

Lubchenco J, Cubit J (1980) Heteromorphic life histories of certain marine algae as adaptations to variations in herbivory. Ecol 61 : 676-687

Menge BA (1978) Predation intensity in a rocky intertidal com- munity: relation between predator foraging activity and envi- ronmental harshness. Oecol 34:1-16

Menge BA, Sutherland JP (1976) Species diversity gradients: syn- thesis of the roles of predation, competition and temporal heter- ogeneity. Amer Nat 110: 351-369

Morely RH, Ring RA (1972) Intertitdal Chironomidae (diptera) of British Columbia : II. Life history and population dynamics. Can Ent 104:1099-1121

Nicotri ME (1974) Resource partitioning, grazing activities, and influence on the microflora by intertidal limpets. Ph.D. thesis. Univ of Washington, Seattle

Paine RT, Levin SA (1981) Intertidal iandscapes: disturbance and the dynamics of pattern. Ecol Monogr 51:145-173

Ricketts RF, Calvin J (1972) Between Pacific Tides. 4th ed. J Hedg- peth (ed). Stanford University Press

Rtitzler K (1970) Spatial competition among Porifera: solution by epizoism. Oecol 5 : 85-95

Robles CD (1979) Disturbance and the interactions of herbivorous diptera with their resources, competitors, and predators on rocky shores. Ph.D. thesis. Univ of California, Berkeley

Robles CD, Cubit JD (1981) Influence of biotic factors in an upper intertidal community: effects of grazing diptera larvae on algae. Ecol 62 : 1536-1547

Saunders LG (1928) Some marine insects of the Pacific Coast of Canada. Ann Ent Soc Amer 21 : 521 531

Sommerfeld MR, Nichols HW (1970) Developmental and cytologi- cal studies of Bangia fuscopurpurea in culture. Amer J Bot 57 : 640-648

Sousa WP (1979) Disturbance in marine intertidal boulder fields: the non-equilibrium maintenance of species diversity. Ecol 60:1225-1239

Southward AJ (1956) The population balance between limpets and seaweeds on wave-beaten rocky shores. Rept Mar Bio Sta Port Evin 68 : 20-29

Southward AJ (1964) Limpet grazing and the control of vegetation on rocky shores. In: Crisp DJ (ed) Grazing in terrestrial and marine environments. Blackwell, Oxford

Stoddart DR (1972) Catastrophic damage to coral reef communi- ties by earthquake. Nature 239 : 51 52

Woodin SA (1978) Refuges, disturbance, and community struc- ture: a marine soft-bottom example. Ecol 59: 274-284

Received December 3, 1981