Norris (Smithsonian Institution). Each species was ... er specimens of seaweeds were identified by J. Norris and. S. Fredericq ...... James Norris. Valerie Paul ...
oecologia 0Springer-Verlag
Occologia (Bcrlin) (1984) 64: 396-407
1984
Predictable spatial escapes from herbivory : how do these affect the evolution of herbivore resistance in tropical marine communities? Mark E. Hay Llnivcrsity of Norlh Carolind a t Chdpel Hill, I n ~ l i l u l eof Mdrine Sciences, Morehedd City, N C 28557, USA
Summary. Between-habitat differences in macrophyte consumption by herbivorous fishes were examined on three Caribbean and two Indian Ocean coral reefs. Transplanted sections of seagrasses were used as a bioassay to compare removal rates in reef-slope, reef-flat. sand-plain, and lagoon habitats. Herbivore susceptibility of fifty-two species of seaweeds from these habitats was also measured in the field. Seagrass consumption on shallow reef slopes was always significantly greater than on shallow reef flats, deep sand plains, or sandy lagoons. Reef-slope seaweeds were consistently resistant to herbivory while reef-flat seaweeds were consistently very susceptible to herbivory. This pattern supports the hypothesis that defenses against herbivores are costly in terms of fitness and are selected against in habitats with predictably low rates of herbivory. Sand-plain and lagoon seaweeds showed a mixed response when placed in habitats with high herbivore pressure; most fleshy red seaweeds were eaten rapidly, most fleshy green seaweeds were eaten at intermediate rates, and most calcified green seaweeds were avoided or eaten at very low rates. Differences in susceptibility between red and green seaweeds from sand-plain or lagoon habitats may result from differential competitive pressures experienced by these seaweed groups or from the differential probability of being encountered by herbivores. The susceptibility of a species to removal by herbivorous fishes was relatively consistent between reefs. Preferences of the sea urchin Diackma anfillarum were also similar to those of the fish guilds. Unique skcondary metabolites were characteristic of almost all of the most herbivore resistant seaweeds. However, some of the herbivore susceptible species also contain chemicals that have been proposed as defensive compounds. Genera such as Sargassum, Turbinaria, Thalassia. Halodule, and Thalassodendron, which produce polyphenolics or phenolic acids, were consumed at high to intermediate rates, suggesting that these compounds are not effective deterrents for some herbivorous fishes. Additionally, potential for the production of the compounds caulerpin, caulerpicin and caulerpenyne in various species of Caulerpa did not assure low susceptibility to herbivory. Heavily calcified seaweeds were very resistant to herbivory, but all of these species also produce toxic secondary metabolites which makes it difficult to distinguish between the effects of morphological and chemical defenses. Predictions of susceptibility to herbivory based on aigal toughness
and external morphology were of limited value in explaining differing resistances to herbivory.
Herbivory plays a primary role in determining the distribution of seaweeds on coral reefs (Stephenson and Searles 1960, Randall 1961, 1965, Ogden et al. 1973, Sammarco et al. 1974, Wanders 1977, Sammarco 1980, Hay 1981a, b, Lubchenco and Gaines 1981, Hay et al. 1983, Hay and Goertemiller 1983). Sharp, between-habitat changes in herbivory have been shown to occur where reef flats join reef slopes (Steneck and Adey 1976, Hay 1981 b, c, Hay et al. 1983), where reef slopesjoin unstructured sand plains (Earle 1972, Hay 1981 a, Hay et al. 1983), and where patch reefs join surrounding grass beds (Randall 1965, Ogden et al. 1973, Hay 1984). Neither the between-reef variations in these patterns, nor the relative susceptibility to herbivory of seaweeds from each of these habitats have been systematically assessed. Recent investigations have focused on between- and within-habitat variations in herbivory on individual reefs (Hay 1981 a, b, Hatcher 1982, Hatcher and Larkum 1983, Hay et al. 1983, Lewis 1984) and on changes in herbivory that occur over depth gradients on several reef slopes scattered throughout the Caribbean (Hay and Goertemiller 1983, Hay 1984). In this study, I assess between habitat differences in herbivory on 3 Caribbean and 2 Indian Ocean reefs. I used transplanted sections of the seagrasses Thalassia testudinum (Caribbean Sea) and Thalassia hemprichii (Indian Ocean) as a bioassay for herbivore activity in different habitats (reef flats, reef sloDes, sand plains, and lagoons). These assessments are followed by algal and seagrass transplants that measure the relative susceptibility to herbivores (primarily reef fishes) of seaweeds from the different habitats. A growing literature on herbivory and the evolution of plant defensive characteristics suggests that decreased susceptibility to herbivores can be achieved only by divcrting energy and nutrients from other plant needs (Feeny 1976, Rhoades and Cates 1976, Rhoades 1979, Lubchenco and Gaines 1981). Thus, defenses are costly and, in the absence of herbivores, less defended individuals or species will have higher fitness than more heavily defended individ-
A
397
uals or species. Rhoades (1979) cites several terrestrial examples supporting this hypothesis. If this reasoning is also valid for marine plants, then habitats that serve as predictable escapes from herbivory should be populated primarily by species that are highly susceptible to herbivore damage. Low susceptibility to herbivory should be characteristic of those species that occur in habitats where herbivory is predictably high. In this investigation I ask the following questions: 1) Are between-habitat differences in herbivory consistent among reefs within a geographic region and among reefs in different oceans? 2) When placed in areas of high herbivore activity, are seaweeds from high herbivory habitats consistently resistant to herbivory and seaweeds from low herbivory habitats consistently susceptible to herbivory? 3) Do the herbivore resistant species have common chemical or morphological characteristics that could account for their low susceptibility to herbivory? 4) Are similar preferences found in guilds of herbivorous fishes on widely separated reefs? If so, are seaweeds that are preferred or avoided by fishes treated in a similar way by the sea urchin Diadema antillarum? Methods
Between-habitat variations in herbivory were assessed using the seagrasses Thalassia testudinum in the Caribbean Sea and Thalassia hemprichii in the Indian Ocean. Weighted clothespins holding 5 cm lengths of clean Thalassia were transplanted into different habitats and herbivory measured as the loss in length of each blade over a period of 2.54.5 h (see Hay et al. 1983 and Hay 1984). In this paper, I consider only the shallower portions of reef slopes (2-10 m deep) and compare herbivory in this habitat to nearby reef-flat, sand-plain or lagoon habitats. Extensive sand areas at the base of fore reefs are called sand plains (see Earle 1972, Dahl 1973, Hay 1981 a). Shallower, sand areas on the lee side of the fore reef are considered lagoons and are usually dominated by seagrasses and associated seaweeds. When distributing Thalassia sections within either sand-plain or lagoon habitats, I avoided patch reefs. Shallow back-reef and reef-crest areas that are exposed or only a few centimeters deep at low tide are called reef flats. Susceptibility of different seaweeds to removal by herbivorous fishes was assessed by weaving small ( 3 4 cm long) pieces of each species into a 3-strand rope that was then fastened to the reef slope at depths of 5-10m (N=1452 ropes/study site). Plants within a length of rope were separatcd from each other by about 5 cm. Thus, when an herbivore encountered a rope, all species of seaweed should have been equally apparent and available. Grazing on the ropes was allowed to continue until there was a clear distinction between the most and least susceptible seaweeds. This resulted in exposure times of between 1.5 and 36 h for seaweeds on different reefs. At the end of an experiment, each species on each rope was recorded as either still present or totally consumed. Ropes were only placed in the fieid under completely calm conditions and were shaken before attachment to the reef to insure that all species were securely attached. Small portions of consumed individuals were inaccessible to fish since they were between the strands of rope: these portions almost always showed the crescent
shaped bite marks of herbivorous fishes. In addition, on reefs where feeding trials were of short duration (Becerro 1.75 h, Media Luna 3 h and Lighthouse 1.5 h), I was able to observe the ropes directly for most of the test period; I saw n o loss of individuals to any source other than herbivory. For the longer duration experiments (9-36 h) where the ropes were not continually watched, I cannot rule out the possibility that some individuals were lost to breakage. However, the magnitude of such losses would have to be very small given the calm conditions and my inability to observe breakage during any of the observation periods. These methods measure only the relative susceptibility of a species to being removed from the reef by herbivorous fishes. They d o not differentiate among fish species and they d o not yield preference data for any given species or type of fish. Based on qualitative observations at each study site, seaweeds used in the experiments on susceptibility to herbivory were listed as typical of reef-flat, reef-slope, sand-plain or lagoon habitats. Assignment of species to habitats does not encompass seasonal variation or spatial variation between different study areas. Thus, on Becerro, Laurencia papillosa was considered a reef flat species because it was not found in other habitats. On Lighthouse, L. papillosa was abundant on reef flats and occasionally found in seagrass beds within the lagoon; on this reef, it was considered both a reef-flat and a lagoon species. Feeding preference data for the sea urchin Diadema antillarum were gathered only a t Galeta Reef, Panama. Forty urchins were collected from the reef and placed in individual glass containers (3.8 1 in volume) that were supplied with running seawater. Small pieces (-2 cm long) of 9 different species ,of seaweeds were pushed into slits in a heavy rubber disc that was placed on the bottom of each container. The top of each container was covered by a small mesh cloth so that dislodged portions of algae could not float out. After variable periods of time (3-24 h) each disc was recovered, the container searched for dislodged pieces of algae, and each species recorded as either present or completely eaten. The experiment was repeated until there were 116 replicates in which some, but not all species, had been eaten and in which no species had been dislodged from the mat. A replicate was not included in the final data set i f (1) any alga floated loose from the rubber disc. (2) no algal species was completely eaten, or (3) all the species were completely eaten. The effect of morphology on susceptibility to herbivory was assessed by assigning species used in the susceptibility trials to morphological groups similar to those proposed by Littler and Littler (1980), Littler et al. (1983a, b), and Steneck and Watling (1982). Assignments followed Littler et al. (1983a, b) as closely as possible since these studies had been conducted on similar species. Information on species that produce secondary metabolites that can be interpreted as feeding deterrents was gathered by searching the available chemical literature and by consulting with W. Fenical, V. Paul (Scripps Institution of Oceanography) and 3. Norris (Smithsonian Institution). Each species was assigned to one of five categories depending upon its chemical characteristics. Categories were: (1) species known to produce unusual secondary metabolites or whose extracts show significant biological activity. (2) species in which chemists have been unable to kind unusual compounds, (3) genera in which some species produce
398
secondary metabolites but this species remains uninvestigated, (4) genera in which species have been analyzed without detecting any unusual chemicals but this species has not yet been analyzed, and (5) species and genera for which no information on secondary chemistry is available. Voucher specimens of seaweeds were identified by J. Norris and S. Fredericq (Smithsonian Institution) and deposited in the algal collection of the U.S. National Herbarium, Smithsonian Institution. Results
1401
,
LIGHTHWSE. BELIZE
BECERRO, HOHWRAS
MEDIA LUNI. HONDURAS.
CARIBBEAN SEA
Thalassia sections transplanted to shallow (2-10 m), reefslope habitats were consumed at rates that were significantly higher than in any of the other habitats tested (Fig. 1, P
