Seaweed Adaptations to Herbivory Author(S): J

Seaweed Adaptations to Herbivory Author(s): J. Emmett Duffy and Mark E. Hay Source: BioScience, Vol. 40, No. 5 (May, 1990), pp. 368-375 Published by: American Institute of Biological Sciences Stable URL: http://www.jstor.org/stable/1311214 Accessed: 06/01/2010 15:17

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=aibs.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Institute of Biological Sciences is collaborating with JSTOR to digitize, preserve and extend access to BioScience.

http://www.jstor.org Seaweed Adaptations to Herbivory Chemical,structural, and morphologicaldefenses are often adjustedto spatial or temporalpatterns of attack

J. EmmettDuffy and Mark E. Hay

erbivoryon seaweedscan be kelp beds along several kilometersof intense,with nearly100% of and coastline (Tegnerand Dayton 1987), production being consumed Fish, urchins, these events are rare. Mesograzer in some habitats (Carpenter1986). gastropods can each populationsare usuallymaintained at Seaweedsminimize damage from her- low densities by predation, and they bivores by any of three strategies strongly affect seaweed commonly have little obvious effect (Lubchencoand Gaines 1981). They communities on most seaweeds (Duffy in press, can escapein space or time so they do Hay et al. 1988c). However, meso- not co-occur with important herbi- . grazerscan have subtle but important vores or are not detected when they effects on some seaweeds, and the do co-occur. They can deter feeding ture; in temperate communities, ur- types of effectscan differdramatically by herbivoresthat encounterand rec- chins and gastropods tend to be most among mesograzer species (Brawley ognize the plant. And they can mini- important (Carpenter 1986, Gaines and Adey 1981, Duffy in press). mize the decrease in fitness that re- and Lubchenco 1982, Hawkins and Fish may have been especially im- sults from herbivoreattack. Hartnoll 1983, Hay 1984a, in press a, portant in selecting for antiherbivore Herbivory'sprofound effects on the Lubchenco and Gaines 1981). Al- traits in seaweeds (Hay in press a, organizationof seaweedcommunities though mesograzers such as amphi- Steneck1983) becauseof their mobil- has been extensively reviewed (Car- pods can dramatically affect seaweed ity, abundance,high metabolic rate, penter 1986, Dayton 1975, Hay communities by destroying entire and, for some abundanttropical spe- 1985, Lewis 1986, Lubchenco and Gaines 1981). In this article, we de- scribeseaweed characteristicsthat di- minish losses to herbivores. Seaweeds are consumed by a diverse assemblage of herbivores that includes fishes, urchins, gastropods, crabs, and numerous smaller herbivores such as amphipods, isopods, and polychaetes (collectivelytermed mesograzers). In tropical habitats, fish and urchins have the greatest effect on seaweed community struc- J. EmmettDuffy is a postdoctoralfellow in the Departmentof InvertebrateZool- ogy, NationalMuseum of NaturalHis- tory, SmithsonianInstitution, Washing- ton, DC 20560.Mark Hay is an associate professorof marinesciences, biology, and ecology at the Universityof North Caro- lina at Chapel Hill, Institute of Marine Sciences, Morehead City, NC 28557. ? 1990 AmericanInstitute of Biological Sciences. A dense school of parrotfishgraze algae on a coral reef. Photo: M. Hay.

368 BioScienceVol. 40 No. 5 cies, fused teeth (Figure 1), which Hypnea musciformis was quickly allow them to feed on tough or calci- eaten by fish when growing alone, but fied seaweeds. Urchins and gastro- it grew more rapidlythan it was con- pods can also be importantin habi- sumedwhen placedin contactwith its tats where they reach high densities unpalatable competitor, the brown (Carpenter1986, Hawkins and Hart- alga Sargassum filipendula (Hay noll 1983). 1986). During the yearlypeak in her- Fish, urchins, and gastropods can bivore abundance,Hypnea and other each strongly affect present-daysea- palatableseaweeds were found exclu- weed communities (Carpenter1986, sively in associationwith unpalatable Hawkins and Hartnoll 1983, Hay plants; at times of year when herbi- 1985, Lewis 1986, Lubchenco and vores were less active, palatable Gaines 1981), and the fossil record plants were not strongly associated suggests they have played a major with unpalatablespecies. role throughout evolutionary time In communities where associa- (Steneck 1983). In this article, we tional refuges are important, species discuss how seaweeds escape, deter, richnesscan increase rather than de- or tolerate herbivory. The depth in crease as the community becomes which we cover each of these topics dominatedby a few unpalatablesea- reflects the availability of rigorous weeds (Hay 1986). In these cases, experimentalstudies; we know little palatablealgae that would have been of the relative importance of these eliminatedby grazerscan persistonly differenttraits. in association with their unpalatable competitors.Similarly, on coral reefs, Figure1. The beak-liketeeth (top) and twice as taxa occur from herbivores mill and in nearly many Escaping pharyngeal (middle bottom) within 10 cm of the chemically de- Seaweeds can avoid theback of thethroat from two speciesof fended brown zo- Spatial refuges. Caribbeanparrotfishes. The fusedteeth alga Stypopodium herbivoresby growing in habitats, in enablethe fishto bite into hardcalcified nale as in similar-sized areas away from et al. microhabitats,or at times of year in surfacesto extractalgal tissues. The pha- Stypopodium (Littler which herbivoresare not active. For ryngealmill functionslike a mortarand 1986). As a final example, brown example, on tropical coral reefs pestle,grinding the seaweedsalong with algae in the genus Desmarestia con- whereherbivores are abundant,many sedimentsand coralrock. Because these centrate sulfuric acid up to 18% of seaweeds that are physiologicallyca- fish have no cellulases,they use their plant dry mass (Andersonand Velim- pable of growing on topographically mouthpartsto physicallyrupture the algal irov 1982). In Chilean kelp beds complex reef slopes occur only where cells.Photo: S. M. Lewis. heavily grazed by sea urchins, the herbivoredensities are low-on topo- palatable kelp Macrocystis cannot graphically simple sand plains and successfullycolonize unless it invades reef flats (Hay 1981a, 1984b, 1985, spatial refuges for seaweeds. Reef an area encircled by Desmarestia Lewis 1986). Similarly, Caribbean damselfishare small, aggressiveher- plants, which appear to act as "acid patch reefs are often surroundedby bivoresthat establishgardens of pal- brooms"that sweep the substrateand halos of bare sand that are kept free atablealgae by removingunpalatable prohibit urchins from entering the of seagrassesby grazing fish and ur- species and by vigorously defending area (Dayton 1985). chins (Ogden et al. 1973, Randall their gardens against other herbi- Associational refuges can thus be 1965). Because many reef-associated vores. Though the damselfishfeed in effectivein a variety of marine habi- grazersare reluctantto venturemore the gardens, these plots experience tats and against different kinds of than approximately 10 m from the reduced herbivory relative to areas herbivores. Interestingly, although shelter of the reef, seagrassesthrive not defended by damselfish.A few the patternsand processesof associa- beyond this radius. seaweedsdepend on these herbivore- tional resistancein marineand terres- On a smaller spatial scale, sea- created refuges, and algal diversity trial systems are similar, the mecha- weeds may live in cracks and holes can be higherin damselfishterritories nisms producing them can be inaccessibleto herbivores. On inter- than in eithercaged or uncagedareas strikingly different. In the best- tidal rock surfaces along Pacific Pa- outside territories(Hixon and Brost- studied marine example (Pfisterand nama, where herbivoryis particularly off 1983). Hay 1988), the palatable seaweed intense, herbivore-resistant algal Gracilariawas grazed more by sea crusts dominate exposed surfaces, Associational refuges. Herbivore- urchinswhen it occurredas a mono- whereas less-resistantleafy algae oc- influenced communities are often culture than when it occurred in a cur primarilyor exclusively in holes dominated by unpalatableseaweeds polyculturewith the unpalatablealga or cracks(Menge et al. 1985). Spatial that may createrefuges for more pal- Sargassum. The greatest grazing in escapes have been discussedin detail atablespecies (Hay 1986, Littleret al. monoculturesoccurred because indi- by Lubchencoand Gaines (1981) and 1986, Pfister and Hay 1988). For vidual urchinsate faster there. Hay (1985). example, field experiments demon- This observationcontrasts with ter- In some cases, herbivores create strated that the palatable red alga restrial systems, where associational

May 1990 369 8- small sporeling stage that is vulnerable to grazers, size-related susceptibil- LLJ ity differences may be common. Even uJ_U1 species that herbivores avoid as adults have that are eaten t) may sporelings by 6- unselective grazers because they are too small to be distinguished from PACHYDICTYOL-A other plants, they are nutritionally m superior to older plants of the same or are easier for herbi- cz 4- species, they vores to manipulate. The challenge then is for the plant OH18 ) a at COCz to survive until it reaches size which it is less vulnerable to herbivores. of the mechanisms oZ 2- Any refuge mentioned above may be employed. .-_ For example, kelp sporelings are y////1111 more vulnerable to herbivores than (18) (18) (18) (17) are larger kelps, and the kelp sporel- 0- ings escape herbivores more often 0 0.2 0.5 1.0 when hidden among filamentous CONCENTRATIONOF PACHYDICTYOL-A(%) brown algae than when growing on bare rock (Harris et al. 1984). Simi- larly, small (less than 3 cm) Fucus Figure 2. The number of amphipods (Pseudamphithoides incurvaria) building domiciles vesiculosus are more were treated with plants suscepti- in 48 hours from discs of the green seaweed Ulva that differing ble to snails than are concentrations of the alcohol = P = This grazing larger diterpene pachydictyol-A (r2 0.98, 0.011). and survival of small is deters fish and is the metabolite by the plants, plants compound feeding by major secondary produced enhanced settlement in small host alga Dictyota bartayresii. When in domiciles built of Dictyota, the by amphipod's her- amphipods are immune to fish predation. (From Hay et al. in press.) cracks or among barnacles where bivores cannot graze effectively (Lub- chenco 1983). refuges have been reported to occur but are suspectible to herbivores. because herbivorous insects become Therefore, they occur primarily dur- Decreasing attractiveness more numerous in monocultures. ing times of the year when herbivore to herbivores These increased densities are due to activity is low. The less competitive, insects finding the monocultures more but more herbivore-resistant, crustose For long-lived seaweeds, and those easily, rarely leaving them, and repro- forms dominate when herbivores are growing in areas such as tropical reefs ducing while there (Bach 1980). In active. Though upright forms nor- where herbivores are abundant and the marine study, urchins did not find mally occur in the field only during active year-round, it may be difficult monocultures more easily, and they winter, these forms persist during the for seaweeds to escape their enemies emigrated from them more rapidly summer if grazers are experimentally through spatial or temporal refuges. than from polycultures. excluded (Lubchenco and Cubit In these situations, selection should 1980). favor seaweeds that are unattractive Temporal refuges. Often, herbivore Seaweeds may also minimize herbi- to herbivores. activity is predictable in time as well vore damage by producing their new, as in space. For example, in temperate most palatable, growth during peri- Chemical defenses. Seaweeds produce marine systems, herbivores are often ods of relatively low herbivory. An a large number of secondary metabo- active primarily during the summer. extreme example is the tropical alga lites, including terpenes, aromatic Several genera of algae have life his- Halimeda; it synchronously produces compounds, acetogenins, amino ac- tories that appear to capitalize on its youngest and most nutritious por- id-derived substances, and phloro- these seasonal changes in grazing tions at night while herbivorous reef tannin polyphenolics (Faulkner 1984, pressure (Lubchenco and Cubit fishes are not feeding (Hay et al. 1986). Although many of these com- 1980). 1988b, Paul and Van Alstyne 1988a). pounds are known to function as de- Complex life histories, including al- fenses against herbivores (reviewed ternation of morphologically distinct Size-related refuges. Seaweeds can by Hay and Fenical 1988), they might generations that differ in growth po- also avoid being eaten if they become also serve to deter pathogens and tential and resistance to herbivores, too large for herbivores to handle fouling organisms (Wahl 1989). are common among intertidal algae effectively. Such size-related immu- VARIANCE IN SUSCEPTIBILITY (e.g., Ulothrix, Urospora, Petalonia, nity presumably results from a larger AMONG HERBIVORES. Recent field Scytosiphon, Bangia, and Porphyra). plant being tougher, or more easily tests have demonstrated the effective- Upright forms of these species are recognized, if it is unpalatable. Be- ness of a variety of terpenoid com- good competitors against other algae cause all algae must pass through a pounds in decreasing the overall

370 BioScience Vol. 40 No. 5 herbivore damage that seaweeds Hay et al. 1989, in press, Paul and or in patches of a palatable green alga experience in their natural environ- Van Alstyne 1988b). were eaten rapidly. ment (Hay et al. 1987b, Hay and For example, the amphipod Pseud- Last, many shell-less gastropods in Fenical 1988, Paul and Van Alstyne amphithoides incurvaria lives in a the order Ascoglossa (Figure 4) are 1988a). Laboratoryexperiments have mobile bivalved domicile it builds specialized feeders on chemically rich clarifiedsome of the interactionsoc- from the chemically defended sea- green algae (Jensen 1980). These her- curringin the field and have shown weed Dictyota bartayresii (Hay et al. bivores sequester both chemical de- that a given compoundis more effec- in press). When in domiciles, the am- fenses (Paul and Van Alstyne 1988b) tive against some types of herbivores phipods are rejected by predatory and functional chloroplasts from than against others. Often com- fish; when removed from domiciles, their algal hosts. Sequestered chloro- pounds that deter grazing fish either they are rapidly eaten. The alga they plasts continue to photosynthesize stimulateor do not affect feeding by use for domiciles produces the diter- within the animal's cells for as long as small invertebrateherbivores (meso- pene alcohol pachydictyol-A. This three months, helping to meet the grazers)that live on seaweeds; these compound deters feeding by fishes energetic needs of the herbivore mesograzers may gain protection but stimulates domicile building by (Trench 1975). Sequestered seaweed from their predators by associating the amphipods. Amphipods will not metabolites make the ascoglossan dis- with chemicallydefended algae (Hay normally build domiciles from the tasteful to predators (Paul and Van in pressb, Hay et al. 1987a, 1988a,c, green seaweed Ulva; however, they Alstyne 1988b). Ascoglossans are the 1989, in press). will do so when Ulva is treated with only group of marine herbivores that As examples, the amphipod Am- pachydictyol-A (Figure 2). show a high proportion of specialized pithoe longimanaand the polychaete The Great Barrier Reef crab Ca- feeders. Platynereisdumerilii both build and phyra rotundifrons is found only in DISTRIBUTION OF CHEMICAL DE- live in mucilaginoustubes attachedto patches of the chemically defended FENSES WITHIN AND AMONG PLANTS. the algae on which they feed. Feeding green alga Chlorodesmis fastigiata, Studies of variation in defensive and living on seaweeds that are pre- which appears to be its only food chemistry among conspecific plants, ferredby fish might expose these me- (Hay et al. 1989). Chlorodesmis pro- and among parts of individual plants, sograzersto indirectpredation as the duces a cytotoxic terpene that deters are just beginning. Two examples fish eat the algae. In North Carolina, fish feeding but significantly stimu- suggest that plants allocate chemical both these mesograzers selectively lates feeding by the crab (Figure 3), defenses preferentially to parts that feed on the brown alga Dictyota di- which experiences reduced predation are especially valuable or vulnerable chotoma, which produces diterpene by associating with the chemically to attack. First, Steinberg (1984) alcohols that strongly deter fish and defended seaweed (Hay et al. 1989). found higher concentrations of phe- urchin feeding. These compoundsei- Crabs tethered in patches of Chlo- nolics in reproductive fronds of the ther significantlystimulate or do not rodesmis were rarely consumed by kelp Alaria marginata than in vegeta- affect feeding by the mesograzers predators; those tethered in the open tive blades, and both laboratory and (Hay et al. 1987a, 1988c). A similar field experiments demonstrated re- pattern has been documented for cr duced grazing on the reproductive other +15- II,f and chemi- . . 7 . fronds. in the fish, ,- Hal- mesograzers, I ., -, ! Second, green alga -7 cally defendedseaweeds on a tropical z '9 imeda, young plants and the nitrogen- reef et al. w+10- (Hay 1988a). X/I rich (and thus highly nutritious) new In each of these cases, of older contain con- mesograzers 'w5- growth plants selectively feed on chemically de- centrations of defensive metabolites CI) fended seaweedsthat are repellentto 0- an order of magnitude higher than fish. However, the mesograzersare ? older, less valuable, plants or plant not to these restrictivelyspecialized E-5- 0 0 0 0 parts (Hay et al. 1988b, Paul and Van 0 0o chemically rich seaweeds; they are (D Alstyne 1988a). These patterns sug- found on 0 LO 0 and feed from a wide range Cj r;O gest that allocation of defenses repre- of unrelated algae. It appears that 0 sents a trade-off between costs and these mesograzers selectively graze benefits. toxic host becausethese chlorodesmin plants plants AcO AcO In several tropical green algae, pop- provide micrositesof reducedpreda- ulations on herbivore-rich reefs have tion or by herbivorous,omnivorous, OAc OAc higher metabolite concentrations, predatory fish (Hay et al. 1987a, produce a larger number of metabo- 1988b,c). lites, and are more resistant to herbi- Although feeding specializationis Figure 3. Amount of nonhost seaweed vores than consumed the conspecific populations terrestrial it is by specialistcrab Caphyra from beds where her- typical among insects, when the seaweed was nearly grass rare among marine herbivores rotundifrons is much lower and Feni- (Hay coated with increasingconcentrations of bivory (Paul in press b). The few marine herbi- the cal 1986, Paul and Van Alstyne vores that do are cytotoxic compound chlorodesmin, specialize small, which is producedby the crab's normal 1988a). This pattern is consistent relatively sedentary, and live on host plant Chlorodesmisfastigiata. Verti- with the hypothesis that multiple de- plants that provide them protection cal lines are standarderrors. (Figure mod- fenses become more necessary as her- against predators (Hay in press b, ified from Hay et al. 1989.) bivorediversity increases (Hay 1984b,

May 1990 371 Pacific (Steinberg 1989, in press). Estes and Steinberg (1988) suggest that this patternresults from different histories of herbivorepressure in the two regions. These authors argue that, in the North Pacific, predation by sea otters kept herbivorousurchin populations low, resulting in relatively weak selection for phenolic defenses in brown seaweeds from this region.Conversely, the absenceof sea otters in the South Pacific allowed herbivoresto maintainhigh densities there, and large brown seaweeds evolved high phenolic contents in response to their grazing.Interestingly, South Pacific herbivores appear to have adapted to these chemicallydefended plants, and they are much more tolerant of phenolics than are their North American counterparts (Steinbergin press). Morphological/structural defenses. Marine algae come in a great variety of sizes, shapes, and textures, from delicate threadlikefilaments to giant kelps to rock-hardcrusts. In several cases, seaweedmorphology correlates with susceptibilityto grazers (Lewis et al. 1987, Littler and Littler 1980, Steneckand Watling 1982). Although the possible confounding effects of chemical defenses need to be more rigorously investigated (Hay 1984b, Lewis 1985, Paul and Hay 1986), it is clear that some structuraland morphological features of algae reduce their vulnerabilityto some grazers. In an extensive literature review, Steneck and Watling (1982) showed that the morphology of plants eaten by gastropods correlates reasonably well with the structure of their mouthparts. Morphological catego- for Figure4. An ascoglossangastropod (commonly called a sea slug) grazingon the ries have been less useful explain- chemicallyrich green alga, Caulerpa racemosa. Most ascoglossan gastropods (like this ing food preferencesby other groups Elysiasp.) sequester both chemical defenses and functional chloroplasts from their algal (Duffy in press, Hay 1984b, Lewis hosts.The chemicals defend ascoglossans against their own predators; the chloroplasts 1985), although soft filaments are continuephotosynthesis for as long as threemonths and meetmost of the energetic grazed by most herbivores,whereas needsof the ascoglossans.Photo: M. Hay. crusts and heavily calcified upright forms are exploited by only a few. The clearest examples of morpho- Lubchencoand Gaines 1981). ever, this trend is not universal.The logically defended algae occur in the There are intriguing geographic concentrationsof phenolics in tem- red algal family Corallinaceae.These patterns in chemical defense. Most perate brown algae are substantially algae grow as thin, flat, heavilycalci- seaweed secondary metabolites have higher than in related species from fied crusts.Eighty to ninetypercent of been isolated from tropical species tropicalregions (Steinbergin press). the dry mass of these crusts may be (Faulkner1984, 1986). This chemical Curiously, many brown seaweeds ash (primarily calcium carbonate), diversitypresumably reflects the more in Australiaand New Zealand have and they are often the most abundant intense herbivory in the tropics two to three times the phenolic con- plants in intensely grazed habitats (Gainesand Lubchenco1982). How- tent of similarspecies from the North such as reefs and urchin barrens(Ste-

372 BioScienceVol. 40 No. 5 neck 1983, 1986).1 Because of their were available as supplements, this defenses. The high concentrationof excellent preservation in the fossil strategy may be effective only for spe- defensive compounds in the most record, Steneck (1983) was able to cies that form large monocultures valuable or vulnerable plant parts trace their morphological changes (e.g., some seagrass or brown algal (Hay et al. 1988b, Steinberg 1984), through geologic time; he described beds). and correspondingly low concentra- an adaptive radiation of corallines in tions in less vulnerable parts, suggests response to the evolution of herbi- Defense costs. Seaweedsemploy a re- that allocation of defenses represents vores with increasingly powerful markable array of structural and a trade-off between costs and benefits. mouthparts and abilities to excavate chemical armament against herbi- Anotherway to allocate defensesis to calcareous substrates. Grazing scars vores. The benefits of these features increasemetabolite production when on fossil corallines suggested that the prompt the question, Why don't all herbivores are more active, as in the fused teeth of parrotfish (Figure 1) seaweeds possess such defenses?De- brown alga Fucus vesiculosus,which represented a dramatic advance in the fenses against herbivores may be increases phenolic concentration by ability of herbivores to damage sea- costly and selected against in habitats 20% after simulated herbivoredam- weeds and may have favored coral- with predictably low rates of her- age (Van Alstyne 1988). When trans- lines over other seaweeds (Steneck bivory (Hay and Fenical 1988). planted into the field, these clipped 1983). This explanation is supported by plants lost only half as much tissue as Examples of morphological de- the differential allocation of chemical undamagedcontrols. fenses exist among the noncalcareous defenses to more suspectible or valu- fleshy algae as well. Like other sessile able plant parts (Hay et al. 1988b, Integratedapproaches organisms, many seaweeds can adjust Paul and Fenical1986, Paul and Van to defense their morphology to prevailing phys- Alstyne 1988a, Steinberg 1984); the ical and biological conditions, includ- induction of chemical (Van Alstyne In the field, structural,chemical, and ing grazing pressure. For instance, 1988) or morphological (Lewis et al. nutritionalcharacteristics may act in clonal seaweeds often form short, 1987) defenses after attack; the con- concert to decrease a seaweed's at- tightly packed and highly branched centration of chemically and morpho- tractivenessto herbivores, and these turfs in grazed areas; this turf growth logically defended seaweeds in habi- characteristicsmay be coordinated form can significantly reduce losses to tats most affected by herbivores and with patternsof temporaland micro- grazing (Hay 1981b). Additionally, their low abundance in habitats with habitatescape (Hay 1984b, Hay et al. some reef seaweeds have two mor- minimal herbivory(Hay 1984b, Ste- 1988b, Lewis et al. 1987, Paul and phologically distinct forms that differ neck 1986); and the competitivein- Hay 1986, Paul and Van Alstyne in growth rate and susceptibility to teractions between defended and un- 1988a). Such integrated defense herbivory; transition between the defended plants in the absence of should be particularlyimportant in forms is mediated by changes in graz- herbivores (Estes and Steinberg 1988, environments such as coral reefs, ing pressure (Lewis et al. 1987). As an Hay 1985, Lewis 1986). However, where the diversity of potential ene- example, in high-herbivory habitats, there are no direct experimentalas- mies is high. For example, Littlerand the brown alga Padina occurs as a sessments of these costs, and it is Littler(1980) suggestedthat late suc- prostrate turf that is relatively resist- unclearhow such costs could be mea- cessional seaweeds tend to be less ant to grazing. Within 96 hours of sured or what currency(e.g., energy nutritious, tougher, and more toxic excluding herbivorous fishes, the turf or nitrogen) is appropriate (Hay and because of a suite of physical and changes into a rapidly growing blade- Fenical 1988). biological selectivepressures. like form, which is susceptible to her- Spatial escapes can also entail Most studies have focused on how bivores but is a superior competitor costs. Algae often escape herbivores one particulartrait affects herbivory; against other benthic algae and sessile by growing in physically stressful few have assessedthe potentialimpor- invertebrates (e.g., corals); it may habitats, thereby entailing large costs tanceof covariationof deterrenttraits. even overgrow and kill corals (Lewis in growth rate and fecundityrelative The tropicalgreen alga Halimeda,for 1986, Lewis et al. 1987). to plants protected from grazers in example, minimizes losses to herbi- morephysically benign but herbivore- vores by usinga combinationof heavy Low nutritional quality. A potential rich habitats(Hay 1981a). Physiolog- calcification and chemical defenses, defense that has been mentioned ical adaptationsto the rigorsof these both of which vary in time as tissue (Lubchenco and Gaines 1981), but habitats may entail additional costs. value and susceptibilityto herbivory not addressed experimentally, is low In a similar way, palatable plants that change. This alga also produces its nutritional quality. Plants that are escape herbivores by associating with valuable new growth at night, when critically low in required nutrients, unpalatable competitors may experi- herbivorousfish are not feeding (Hay such as nitrogen, might be little used ence an 80% reduction in growth rate et al. 1988b). Such complex defensive by herbivores. Because herbivores due to competition (Hay 1986). Al- strategiescould be common. would probablybe minimallyaffected though the costs of using these ref- by nutritional deficiencyif many al- uges can be high, the alternative is ternate foods of adequate quality often local extinction. Tolerating herbivory Patterns in the allocation of chem- It is surprising that the dominant 1M. M. Littler,1990, personalcommunication. ical defenses within a plant also pro- plants in areas of most intense her- SmithsonianInstitution, Washington, DC. vide evidence for the costliness of bivory representthe extremes of sus-

May 1990 373 ceptibility to herbivores: small, highly Ugarte 1987). against where herbivory is predict- susceptible filamentous forms (Car- In contrast, spores and vegetative ably low. Some seaweedsbenefit from penter 1986), heavily calcified, very portions of late successional plants moderategrazing because herbivores resistant, crustose corallines (Steneck are rarely resistant to digestion. How- remove competitors, supply limiting 1983, 1986), and upright, calcified ever, grazing may still provide some nutrients, or disperse the seaweed's seaweeds like Halimeda that are also benefits. Amphipods graze on cysto- propagules. defended chemically (Hay et al. carps (reproductive structures) of the 1988b, Paul and Van Alstyne 1988a). red alga Iridaea laminarioides, releas- Acknowledgments These seaweeds may depend on grazing spores into the water column ing to prevent overgrowth by superior when the cystocarps are opened Preparationof this paper and our competitors (Carpenter 1986, Hay (Buschmann and Santelices 1987). recent studies of seaweed-herbivore 1981a, Lewis 1986, Steneck 1986). Spores released by amphipods are as interactions were supported by the Sometimes this service entails loss of viable as naturally released spores, National Science Foundation (OCE the plant's own tissue. and, surprisingly, ungrazed fronds of- 86-08663 and 89-00131), the Na- One alternative to costly structural ten have cystocarps that remain tional GeographicSociety (3400-86), and chemical defenses is to channel closed and do not naturally release the Lizard Island Research Founda- all energy into rapid growth and re- their spores. The percentage of such tion of the AustralianMuseum, and production in an effort to outpace unopened cystocarps is significantly the CharlesA. LindberghFund, Inc. herbivores. By producing tissue rap- greater at sites without amphipods We thankJ. A. Miller, V. J. Paul, and idly, and by having basal portions than at sites with high amphipod den- two anonymousreviewers for helpful that escape herbivory due to the to- sities. Though a lower proportion of commentson the manuscript. pographic complexity of the sub- spores ingested by amphipods germi- strate, small filamentous algae can nated compared with uningested Referencescited often persist despite losing much of spores, those that did germinate had their tissue to due Anderson,R.J., and B. Velimirov. 1982. An grazers (Carpenter higher growth rates, presumably of the Lewis This al- to nutrients absorbed cells experimentalinvestigation palatability 1986, 1986). strategy by plant of kelp bed algae to the sea urchinParechi- lows inconspicuous filamentous algae during passage through the herbivore nus angulosusLeske. Mar. Ecol. 3: 357-373. to make up most of the plant biomass gut (Buschmann and Santelices Bach,C. E. 1980. Effectsof plantdiversity and on moderately grazed areas of some 1987). time of colonization on an herbivore-plant reefs. These in fact Some freshwater interaction.Oecologia 44: 391-326. tropical plants phytoplankton Brawley, S. A., and W. H. Adey. 1981. The produce more mass when grazed, be- show even stronger positive reactions effect of micrograzerson algal community cause continuous cropping prevents to grazing. When the green alga structurein a coral reef microcosm. Mar. self-shading, and herbivore excretion Sphaerocystis schroeteri is consumed Biol. 61: 167-177. increase available nutrients the Buschmann,A., and B. Santelices.1987. Mi- may (Car- by zooplankter Daphnia magna, and releasein Iridaealami- more than 90% of the cells are un- crograzers spore penter 1986). narioidesBory (Rhodophyta:Gigartinales). Some intertidal seaweeds that are damaged, and the nutrient enrich- J. Exp. Mar. Biol. Ecol. 108: 171-179. resistant to herbivores as adults en- ment that occurs during gut passage Carpenter,R. C. 1986. Partitioningherbivory sure survival through the vulnerable may increase growth rates by as much and its effectson coral reef algal communi- a massive recruit- as this enrichment more than ties. Ecol. Monogr.56: 345-363. juvenile stage by 63%; P. K. 1975. evaluationof ment that the of mol- for the the Dayton, Experimental swamps ability compensates slight damage ecologicaldominance in a rockyintertidal al- luscan herbivores to eat the young grazers do to the algal population gal community.Ecol. Monogr.45: 137-159. plants. Under these conditions, the (Porter 1976). In situ grazing experi- . 1985. The structureand regulationof plants may quickly reach a size at ments have indicated that this alga some South American kelp communities. which no suffer increased in number as the number of Ecol. Monogr.55: 447-468. they longer mortality E. In on seaweeds: at normal herbivore densities increased. It to Duffy,J. press. Amphipods (Dayton grazers appeared rely partnersor pests? Oecologia. 1975). In extreme cases, algal growth on grazers as a rich localized source Estes,J. A., and P. D. Steinberg.1988. Preda- may be heavy enough to alter the of nutrients. tion, herbivory,and kelp evolution.Paleobi- habitat in such a way that it becomes ology 14: 19-36. unsuitable for some herbivores Faulkner,D. J. 1984. Marinenatural products: (Un- Conclusions metabolitesof marine and herbivorous derwood and algae Jernakoff 1981). marine molluscs. Nat. Prod. Rep. 1: 251- Losses to herbivores are sometimes Seaweeds employ a wealth of chemi- 280. minimized by ingested propagules or cal, structural, and morphological de- 1986. Marine naturalproducts. Nat. vegetative portions remaining viable fenses against herbivory and often Prod. Rep. 3: 2-33. and the these defenses Gaines,S. D., and J. Lubchenco.1982. A uni- being dispersed by herbivore, manipflate according fied to marine in- a situation reminiscent of seed dis- to or of approach plant-herbivore spatial temporal patterns teractions. II. Biogeography. Annu. Rev. persal by birds and mammals. For herbivore attack. Defenses may differ Ecol. Syst. 13: 111-138. several opportunistic algae exposed greatly in effectiveness against dif- Harris,L. G., A. W. Ebeling,D. R. Laur, and to grazing by gastropods, gut passage ferent herbivores, and multiple de- R.J. Rowley. 1984. Community recovery increases the fenses are common in areas where after storm damage:a case of facilitationin significantly production primary succession. Science 224: 1336- of motile spores and also the growth herbivores are diverse and abundant. 1338. rate of sporelingsrelative to the un- Several lines of evidence suggest that Hawkins,S. J., and R. G. Hartnoll.1983. ingested controls (Santelices and defenses are costly and are selected Grazingof intertidalalgae by marineinver-

374 BioScience Vol. 40 No. 5 tebrates. Oceanogr. Mar. Biol. Annu. Rev. 75: 233-245. (Halimedaceae; Chlorophyta). Coral Reefs 21: 195-282. Hay, M. E., J. R. Pawlik, J. E. Duffy, and W. 6: 263-269. Hay, M. E. 1981a. Herbivory, algal distribu- Fenical. 1989. Host-plant specialization re- . 1988b. The use of ingested algal diter- tion, and the maintenance of between- duces predation on small herbivores: sea- penoids by the ascoglossan opisthobranch habitat diversity on a tropical fringing reef. weed-herbivore-predator interactions. Oeco- Elysia halimedae Macnae as antipredator Am. Nat. 118: 520-540. logia 81: 418-427. defenses.J. Exp. Mar. Biol. Ecol. 119: 15- .1981b. The functional morphology of Hay, M. E., P. E. Renaud, and W. Fenical. 29. turf-forming seaweeds: persistence in stress- 1988c. Large mobile versus small sedentary Pfister, C. A., and M. E. Hay. 1988. Associa- ful marine habitats. Ecology 62: 739-750. herbivores and their resistance to seaweed tional plant refuges: convergent patterns in . 1984a. Patterns of fish and urchin chemical defenses. Oecologia 75: 246-252. marine and terrestrial communities result grazing on Caribbean coral reefs: are previ- Hixon, M. A., and W. N. Brostoff. 1983. Dam- from differing mechanisms. Oecologia 77: ous results typical? Ecology 65: 446-454. selfish as keystone species in reverse: inter- 118-129. . 1984b. Predictable spatial escapes mediate disturbance and diversity of reef Porter, K. G. 1976. Enhancement of algal from herbivory: how do these affect the algae. Science 220: 511-513. growth and productivity by grazing zoo- evolution of herbivore resistance in tropical Jensen, K. R. 1980. A review of sacoglossan plankton. Science 192: 1332-1334. marine communities? Oecologia 64: 396- diets, with comparative notes on radular and Randall, J. E. 1965. Grazing effects on 407. buccal anatomy. Malacol. Rev. 13: 55-77. seagrasses by herbivorous reef fishes in the . 1985. Spatial patterns of herbivore Lewis, S. M. 1985. Herbivory on coral reefs: West Indies. Ecology 46: 255-260. impact and their importance in maintaining algal susceptibility to herbivorous fishes. Santelices, B., and R. Ugarte. 1987. Algal life- algal species richness. Pages 29-34 in Pro- Oecologia65: 370-375. history strategies and resistance to digestion. ceedings of the Fifth InternationalCoral .1986. The role of herbivorous fishes in Mar. Ecol. Prog. Ser. 35: 267-275. Reef Congress. Antenne Museum-Ephe, the organization of a Caribbean reef commu- Steinberg, P. D. 1984. Algal chemical defense Moorea, French Polynesia. nity. Ecol. Monogr. 56: 183-200. against herbivores: allocation of phenolic . 1986. Associational plant defenses Lewis, S. M., J. N. Norris, and R. B. Searles. compounds in the kelp Alaria marginata. and the maintenance of species diversity: 1987. The regulation of morphological plas- Science 223: 405-407. turning competitors into accomplices. Am ticity in tropical reef algae by herbivory. . 1989. Biogeographical variation in Nat. 128: 617-641. Ecology 68: 636-641. brown algal polyphenolics and other second- . In press a. Fish-seaweed interactions Littler, M. M., and D. S. Littler. 1980. The ary metabolites: comparison between tem- on coral reefs: effects of herbivorous fishes evolution of thallus form and survival strat- perate Australasia and North America. and adaptations of their prey. In P. F. Sale, egies in benthic marine macroalgae: field and Oecologia 78: 373-382. ed. The Ecology of Coral Reef Fishes.Aca- laboratory tests of a functional form model. . In press. Geographical variation in the demic Press, San Diego, CA. Am. Nat. 116: 25-44. interaction between marine herbivores and . In press b. Seaweed chemical defenses: Littler, M. M., P. R. Taylor, and D. S. Littler. brown algal secondary metabolites. In V. J. their role in the evolution of feeding special- 1986. Plant defense associations in the ma- Paul, ed. Ecological Roles for Marine Sec- ization and in mediating complex interac- rine environment. Coral Reefs 5: 63-71. ondary Metabolites. Comstock Publishing, tions. In V. J. Paul, ed. EcologicalRoles for Lubchenco, J. 1983. Littorina and Fucus: ef- Ithaca, NY. Marine Secondary Metabolites. Comstock fects of herbivores, substratum heterogene- Steneck, R. S. 1983. Escalating herbivory and Publishing, Ithaca, NY. ity, and plant escapes during succession. resulting adaptive trends in calcareous algal crusts. 9: 44-61. Hay, M. E., J. E. Duffy, and W. Fenical. In Ecology64: 1116-1123. Paleobiology . 1986. The of coralline press. Host-plant specialization decreases Lubchenco,J., andJ. Cubit. 1980. Heteromor- ecology algal crusts: and predation on a marine amphipod: an herbi- phic life histories of certain marine algae as convergent patterns adaptive Annu. Rev. Ecol. 17: 273- vore in plant's clothing. Ecology. adaptations to variations in herbivory. Ecol- strategies. Syst. 303. M. E. W. and K. ogy 61: 676-687. Hay, E., J. Duffy, Fenical, R. S., and L. 1982. Gustafson. 1988a. Chemical defense in the Lubchenco, J., and S. D. Gaines. 1981. A uni- Steneck, Watling. Feeding capabilities and limitations of herbivorous seaweed Dictyopteris delicatula: differential fied approach to marine plant-herbivore in- molluscs: a functional group approach. Mar. effects against reef fishes and amphipods. teractions. I. Populations and communities. Mar. Ecol. Ser. 48: 185-192. Annu. Rev. Ecol. 12: 405-437. Biol. 68: 299-319. Prog. Syst. M. and P. K. 1987. El Nino M. E. C. A. and B. and L. R. Ashke- Tegner, J., Dayton. Hay, E., J. Duffy, Pfister, W. Menge, A., J. Lubchenco, effects on southern California forest Fenical. 1987a. Chemical nas. 1985. and con- kelp defenses against Diversity, heterogeneity, communities. Adv. Ecol. Res. 17: 243-279. different marine herbivores: are sumer pressure in a intertidal amphipods tropical rocky Trench, R. K. 1975. Of leaves that crawl: func- insect 68: 1567-1580. community. Oecologia 65: 394-405. equivalents? Ecology tional chloroplasts in animal cells. Pages and W. Fenical. 1988. Marine C., R. and N. 1973. Hay, M.E., Ogden, J. Brown, Salesky. 229-265 in D. H. Jennings and D. L. Lee, interactions: the of the echinoid Diadema antillarum plant-herbivore ecology Grazing by eds. Symbiosis. Cambridge University Press, chemical defense. Annu. Rev. Ecol. Syst. 19: Phillipi: formation of halos around West New York. Indian reefs. Science 182: 715-717. 111-145. patch Underwood,A. J., and P. Jernakoff.1981. In- and W. Fenical. 1986. Chemical Hay, M. E., K. Gustafson, and W. Fenical. Paul, V.J., teractions between algae and grazing gastro- defense in order Caul- 1987b. Chemical defense against diverse tropical green algae, pods in the structure of a low-shore intertidal Mar. Ecol. Ser. 33: 255-264. coral reef herbivores. Ecology 68: 1581- erpales. Prog. algal community. Oecologia 48: 221-233. 1590. Paul, V. J., and M. E. Hay. 1986. Seaweed Van Alstyne, K. L. 1988. Herbivore grazing Hay, M. E., V. J. Paul, S. M. Lewis, K. Gustaf- susceptibility to herbivory: chemical and increases polyphenolic defenses in the inter- son, J. Tucker, and R. N. Trindell. 1988b. morphological correlates. J. Exp. Mar. Biol. tidal brown alga Fucus distichus. Ecology Can tropical seaweeds reduce herbivory by Ecol. 33: 255-264. 69: 655-663. growing at night? diel patterns of growth, Paul, V.J., and K. L. Van Alstyne. 1988a. Wahl, M. 1989. Marine epibiosis. I. Fouling nitrogen content, herbivory, and chemical Chemical defense and chemical variation in and antifouling: some basic aspects. Mar. versus morphological defenses. Oecologia some tropical Pacific species of Halimeda Ecol. Prog. Ser. 58: 175-189.

May 1990 375