Induction of seaweed chemical defenses by amphipod grazing

Induction of seaweed chemical defenses by amphipod grazing

Greg Cronin

INTRODUCTION

Ecologists are becoming increasingly aware that plants are active participants in the dynamics of plant-animal interactions (Rhoades 1979, 1985, Belsky 1986, Karban and Myers 1989, Tallamy and Raupp 1991, Baldwin 1994). Besides producing nectar, flowers, and fruits to attract animal pollinators and dispersers, plants also produce structures that mimic butterfly eggs (Williams and Gilbert 1981) or grazer scars (Niemela and Tuomi 1987) to deter herbivores or attract enemies of herbivores, respectively. One of the most commonly documented responses of plants to herbivory has been the induction of increased concentrations of chemical defenses (Rhoades 1985, Karban and Myers 1989, Tallamy and Raupp 1991, Baldwin 1994).

The production of chemical defenses is believed to be costly because defenses use resources that could have been allocated to growth or reproduction (Coley 1986, Herms and Mattson 1992). Constitutive defenses require expenditure of resources even when consumers are absent and the benefits of protection are not realized. In contrast, inducible defenses allow costs of defenses to be deferred until enemies have been detected, at which time the costs can be offset by the benefits of protection. Induced resistance may therefore be an adaptation that minimizes costs by keeping defenses low until they are needed (Harvell 1990, Baldwin 1994).

The pattern of variation in the chemical defenses of some seaweed species suggests herbivore-induced increases of chemical defenses may be responsible for intraspecific variation in chemical defenses. For example, seaweeds from areas of coral reefs where herbivory is intense often produce more potent and higher concentrations of chemical defenses than plants from habitats where herbivory is less intense (Paul and Fenical 1986, Paul and Van Alstyne 1988). However, clipping experiments have failed to induce increased terpenoid chemical defenses in the green seaweeds Halimeda, Udotea, and Caulerpa (Paul and Van Alstyne 1992) and clipping or urchin grazing failed to induce higher levels of phlorotannins in the kelps Ecklonia and Alaria or in the rockweed Sargassum (Pfister 1992, Steinberg 1994, 1995). Thus, the higher levels of constitutive chemical defenses from sites with many herbivores could have been generated by preferential grazing, founder effects, selection, or among-habitat variation in other environmental variables. Despite the common occurrence of secondary metabolites in seaweeds (Hay and Fenical 1988, Faulkner 1993 and references therein), there is only one example of an induced increase in the chemical defenses of a seaweed (Van Alstyne 1988). In contrast, there are many examples of inducible chemical defenses in the terrestrial literature as evidenced by the number of reviews on the topic (Baldwin 1994 and references therein). This discrepancy in the prevalence of inducible defenses in terrestrial vs. marine plants could be due to the different biologies of vascular terrestrial plants vs. nonvascular seaweeds (e.g., the induction stimulus from localized damage may not be efficiently translocated in seaweeds), but the lesser amount of research on seaweeds relative to terrestrial plants may also explain much of the disparity.

In our study, we document within-thallus, within-site, among-site, and among-year variation in chemical defenses, toughness, protein content, and susceptibility to herbivory of the chemically defended brown alga Dictyota menstrualis and how this variation corresponds to the density of herbivorous amphipods. We also manipulate amphipod density in a controlled field experiment to determine if D. menstrualis induces higher levels of chemical defenses in response to amphipod grazing.

METHODS

Study sites and organisms

The collection and experimental sites used in this study are located within 8 km of each other near More-head City and Beaufort, North Carolina, USA. Dictyota menstrualis is found intertidally and subtidally at the mudflat at Lennoxville Point and at the seagrass bed at Mitchell Village. The alga is mostly subtidal at Radio Island Jetty, although shallow plants sometimes experience aerial exposure during extreme low tides. Herbivores used in our assays included the sea urchin Arbacia punctulata collected from the rock jetty at Radio Island, the pinfish Lagodon rhomboides collected near the Morehead City, North Carolina waterfront, and the amphipod Ampithoe longimana collected from Dictyota menstrualis in grass beds in Bogue Sound, North Carolina.

Dictyota menstrualis in North Carolina produces three diterpenoid secondary metabolites: dictyol E, pachydictyol A, and dictyodial (Cronin et al. 1995). Previous experiments show that pachydictyol A or dictyol E significantly deter feeding by most local fishes, urchins, and mesograzers, but that the amphipod Ampithoe longimana is unusually tolerant of these chemical defenses and selectively consumes local species of Dictyota (Hay et al. 1987, Duffy and Hay 1991, 1994). The feeding deterrent effects of dictyodial are relatively unstudied due to its chemical instability (Cronin et al. 1995), but indirect assays suggest that it may deter the urchin, but not the fish or amphipod that we studied in this investigation (Cronin and Hay 1996a). Thus, natural concentrations of dictyols in D. menstrualis are effective defenses against urchins and fishes but are much less effective against the amphipod A. longimana. In fact, because fishes rarely visit or consume Dictyota, the alga represents a partial refuge from predation for small mesograzers like the amphipod A. longimana and the polychaete Platynereis dumerilii, which preferentially live on and consume Dictyota spp. (Hay et al. 1987, 1988, Duffy and Hay 1991, 1994, Cronin and Hay 1996a) despite their defensive chemistry.

Effects of secondary metabolites on herbivore feeding behavior

Dictyol E from Dictyota was purified using methods described by Cronin et al. (1995) and its effects on feeding by Lagodon, Arbacia, and Ampithoe were tested using an agar- and seaweed-based artificial food (see Hay et al. 1994 for methods) that had been used previously to test the effects of secondary metabolites from Dictyota ciliolata on these same herbivores (Cronin and Hay 1996a). Because D. ciliolata contains pachydictyol A, its effects, but not those of dictyol E, had already been determined using this methodology for these herbivores, allowing us to concentrate on dictyol E alone.

General feeding assay procedures

This section explains general methods used to assess the palatability of Dictyota to Arbacia and Ampithoe. Specific details (e.g., number of replicates, duration, and statistical analysis) for each separate feeding assay are described later. Some specifics of the feeding assays were altered slightly over the 3 yr of this project as we developed improved methodologies. For all feeding assays, herbivores had access to nonassay food until just prior to the assay (i.e., animals were not starved).

Arbacia’s willingness to feed on Dictyota from different sites was evaluated by offering plants to individual urchins held in 1.8-L plastic tubs with flow-through seawater. Urchins often shred algae making it difficult to determine the plant from which the pieces originated, therefore no-choice feeding assays were performed to avoid placing more than one plant with each urchin. The wet mass of each alga was determined by spinning the seaweed in a salad spinner to remove excess water, weighing the alga to the nearest milligram, and quickly returning the plant to seawater. Each Arbacia was offered 200-250 mg wet mass (WM) of Dictyota and urchins were allowed to feed for 3-4 d, or until most of the urchins had consumed roughly half their available alga. For each assay plant, a similar portion of the same plant was set up in a similar manner without urchins to control for autogenic changes in mass (Renaud et al. 1990).

The palatability of Dictyota to Ampithoe was determined using both choice and no-choice feeding assays. Although we felt offering the amphipods a choice between different types of Dictyota (e.g., different sites, plant parts, or level of grazing damage) was the better way to assess feeding preferences, no-choice assays were sometimes performed in conjunction with choice assays or when more than two plant types were compared so that we could maintain independence among treatments (Peterson and Renaud 1989). For no-choice feeding assays, [approximately equal to] 100 mg WM of each alga was offered to a separate amphipod in a dish with [approximately equal to]50 mL of non-flow-through seawater. For assays where Ampithoe had a choice of plant types, replicate groups of amphipods were all offered two pieces of algae ([approximately equal to]100 mg WM of each) in a dish with [approximately equal to]100 mL of seawater; the two pieces were distinguished by sewing threads of slightly different lengths through them. For each assay plant, a portion of the same plant was placed in a separate dish without amphipods to control for changes in mass not due to amphipods.

The amount of Dictyota consumed in each replicate was calculated as: [([H.sub.o] x [C.sub.f]/[C.sub.o]) – [H.sub.f]] where [H.sub.o] and [H.sub.f] were pre-assay and post-assay wet masses of tissue exposed to herbivores and [C.sub.o] and [C.sub.f] were pre-assay and post-assay wet masses of controls for autogenic changes in mass. Replicates in which the alga or herbivore died were excluded from the experiment. The highest number of excluded replicates was 6 of 25. Sample size, whether it was a choice or no-choice assay, and the statistical analysis used for each assay are given below, or in the pertinent figure.

Field sampling of amphipods and grazing damage

The relative abundance of grazing scars on Dictyota collected from different sites was determined in November 1991, July 1992, and October 1993. Ampithoe appears to be the most important consumer of Dictyota at our field sites (Hay et al. 1987, Duffy and Hay 1991, 1994) and it makes small semicircular or circular feeding scars that are distinguishable from other types of damage. Haphazardly selected plants from each site (N = 12-24) were placed in coded bags of seawater and an unbiased observer (i.e., one that did not know the code) scored each plant as being ungrazed, slightly grazed, moderately grazed, or heavily grazed. These scores were given a value of 0, 1, 2, or 3, respectively, and among-site differences in these scores were analyzed with a Kruskal-Wallis test for each date. Because observers assessed grazing damage after looking over all plants collected during a single date, numerical scores were subjectively “scaled” to the range of damage seen for that collection. Scaling of these subjective values was not attempted among dates. Thus, among-site differences for a particular date are valid, but comparisons of numerical scores among-dates should not be done because our scale was relative, not absolute.

To assess amphipod densities on Dictyota, 10-20 plants and their associated faunas from each site were carefully sealed in plastic bags while still underwater and returned to the laboratory. After sieving (163-[[micro]meter] mesh) the water from the bag and three freshwater rinses of each plant to retrieve the amphipods, Ampithoe longimana were separated from other amphipods. Ampithoe and “other amphipods” were counted, and each alga was spun in a salad spinner and weighed to the nearest milligram. Densities per mass of each alga were square-root transformed prior to statistical analysis to decrease the heterogeneity among variances.

Measurement of tissue traits

Plant traits that are potentially important proximal cues for the feeding decisions of herbivores were quantified for plants used in feeding assays. Because herbivores are often nitrogen limited and rarely carbon limited (Mattson 1980), protein is generally considered a good measure of the nutritional quality of plant material. Soluble protein was extracted with 1 mol/L NaOH from freeze-dried tissue and analyzed using the Bradford (1976) method with bovine serum albumin as a standard. The protein data were supplemented by analyzing total nitrogen and carbon with a C/N analyzer for plants collected in November 1991. Concentrations of the three secondary metabolites produced by D. menstrualis (dictyol E, dictyodial, and pachydictyol A) were determined using high-pressure liquid chromatography (HPLC) methodologies as described by Cronin et al. (1995).

Temporal and among-site variation in plant traits and in Dictyota-herbivore interactions

To evaluate among-site differences in algal palatability, plant characteristics, and how these patterns changed with time and location, haphazardly chosen plants of Dictyota were collected from two or three different sites on five dates spanning a 3-yr period. Most comparisons were made between Dictyota collected from a rock jetty at Radio Island and a mudflat at Lennoxville Point; however, Dictyota from a seagrass bed at Mitchell Village was included in the initial assessment of among-site variation. These sites were chosen because plants at these locations appeared to differ in grazing damage and associated faunas. Densities of amphipods at the jetty were low, apparently due to predation by fishes (Duffy and Hay 1991), while amphipod densities on plants at the mudflat (Duffy and Hay 1994) and seagrass bed sites seemed high.

On 28-29 October 1991, plants were collected from Radio Island Jetty, Lennoxville Point, and Mitchell Village and kept in flow-through seawater until used in assays within I d. The palatability of Dictyota from each site was determined by individually offering pieces of 25 separate plants from each site to Ampithoe (120 mg WM tissue per amphipod, N = 25) or Arbacia (200 mg WM per urchin, N = 25) in no-choice assays. Amphipods and sea urchins fed for 3.8 and 4.1 d, respectively. The soluble protein (N = 18-24), total nitrogen (N = 8), total carbon (N = 8), and the concentrations of secondary metabolites (N = 13-14) in Dictyota from each site also were analyzed. The density of amphipods and the relative amounts of grazing scars were determined for 15 individuals of Dictyota from each site collected on 8 November 1991. These data were analyzed with a one-way ANOVA and means were compared with a Tukey hsd multiple comparisons test.

On 13 July 1992, Dictyota plants were collected in individual plastic bags from the jetty at Radio Island and the mudflat at Lennoxville Point. After assessing amphipod grazing damage (N = 24) and amphipod densities (N = 26), portions of each plant were used for palatability assays with Ampithoe (choice and no-choice; N = 24), protein quantification (N = 8), and secondary metabolite analysis (N = 12). In the no-choice feeding assay, 100 mg WM of each plant was offered to a single amphipod. In the choice feeding assay, 100 mg WM of each plant was offered to a pair of amphipods. Amphipods fed for 2.7 d. The data for each measured variable and no-choice feeding assay were analyzed with two-sample t tests. Data from the choice feeding assays were analyzed with a paired-sample t test.

On 23 September 1992, Dictyota was collected again from the jetty and mudflat and its palatability to amphipods determined as described above. Concentrations of soluble protein (N = 5-6), dictyol E (N = 7-8), dictyodial (N = 15), and pachydictyol A (N = 15) were evaluated with two-sample t tests.

Collections were made again on 13 July 1993. Between-site differences in palatability of Dictyota to Ampithoe were determined by offering separate groups of 5-7 amphipods (N = 24) a 120 mg WM plant portion from each site. Palatability to Arbacia was compared by offering [approximately equal to]200 mg WM from each plant (N = 24 for each site) to individual urchins for 3 d. Concentrations of soluble protein (N = 15), dictyol E (N = 22), dictyodial (N = 22), and pachydictyol A (N = 17) were measured for plants from the two sites.

Final collections of Dictyota from the jetty and mudflat were made on 25 October (feeding assay and tissue measurement plants) and 26-27 October 1993 (for grazing damage and amphipod density assessment). Feeding assays and plant characteristics were determined as in the previous assays.

Within-site and within-plant variation of D. menstrualis

Differences in the apparency of amphipod grazing scars not only occurred among sites, but also, at times, among plants within a site. By contrasting the palatability of grazed vs. ungrazed plants from a single site, we hoped to determine if previous amphipod damage, rather than between-site differences in physical factors, could be altering the palatability of Dictyota to the amphipod Ampithoe longimana. From Lennoxville Point, 30 plants with no apparent grazing scars and 30 plants with obvious amphipod grazing damage were collected and assayed for palatability using Ampithoe. Groups of 10 amphipods (N = 23) were offered a choice of [approximately equal to]100 mg WM each of ungrazed and previously grazed Dictyota for 2.1 d. The remaining seven plants from each treatment were held in similar containers but without amphipods. These served as controls for autogenic changes in mass.

Dictyota often has adventitious branches growing from areas of the thallus that have been damaged by grazing or possibly other factors (G. Cronin and M. E. Hay, personal observations). To see if induction is localized to plant parts near damage (i.e., if adventitious branches have more defenses than other tissues), tissue traits of adventitious branches, apices, and the middle of individual plants were compared by collecting 20 individuals from Lennoxville Point that had numerous adventitious branches. Each plant was dissected into its respective parts: adventitious branches growing from the damaged margins of plants were excised with scissors, apices were cut from the top 1 cm of branches, and middles were defined as tissue located 2-6 cm below the tips of branches. Half of the middles were cut into pieces similar in size to the adventitious branches and apices to control, to some extent, for the effects of cutting and of size; the remaining half of the middles was not cut into smaller pieces. The palatability of the four tissue types (i.e., adventitious branches, apices, cut middles, and uncut middles) to Ampithoe was determined in a no-choice feeding assay with appropriate controls. About 110 mg WM of each tissue type was offered to 20 separate pairs of amphipods for 1.5 d. Because there was no difference in the amount of cut and uncut middles eaten by Ampithoe (mean [+ or -] 1 SE, 27 [+ or -] 3 vs. 25 [+ or -] 2 mg WM, respectively; P = 0.6 paired-sample t test; paired by plant; N = 20), additional measurements were not performed on the cut middles.

Tissue toughness, nutritive value, and concentrations of secondary metabolites were measured for each plant part. The toughness of the different plant parts (N = 20) was measured with a penetrometer as described by Duffy and Hay (1991). The concentration of secondary metabolites and soluble protein was determined for the different parts of eight individuals. Data from the feeding assay and tissue measurements were analyzed with a mixed-model ANOVA (the fixed factor was “plant portion” and the random factor was “individual plant”) followed by a Tukey hsd multiple comparisons test for the “plant portion” term.

Manipulation of amphipod grazing on D. menstrualis

To test directly the effect of previous Ampithoe grazing on Dictyota, we performed field manipulations to alter Ampithoe grazing damage on different halves of individual Dictyota plants. Thirty relatively undamaged plants were removed from the rocks at Radio Island Jetty, divided in half underwater, entwined in a 10 cm length of three-strand polypropylene rope, and anchored along the west side of Radio Island Jetty where Dictyota grew naturally. Dividing plants in half required localized damage near the base of the plant. We assumed that this minimal damage would cause little, if any, induction of chemical defenses; however, tissue near this site was not used for subsequent measurements and any response to this damage should have occurred equally in both halves. One haphazardly chosen half was designated to be attacked by Ampithoe while the remaining half was left ungrazed as a control. Grazing damage was manipulated by periodically enclosing the plants in cages with or without added amphipods. The cages were made by stretching nylon stockings over a cylindrical frame ([approximately equal to]10 cm diameter x 20 cm tall) made of plastic-coated copper wire. Plant halves designated to be grazed were placed in the cages with 8-13 amphipods for 1-3 d, producing levels of damage that were visually similar to levels observed at Lennoxville Point. Control plants remained in similar cages without amphipods for the same time period to control for cage effects. For 2-5 d between grazing periods, all plants were placed back on the jetty to allow them to recover under natural conditions. Plants within a pair were separated by 20-30 cm. The experiment was run once in 1993 and once in 1994. In the 1993 experiment, plants were caged on days 0-1, 3-5, and 7-8 of the 16-d experiment. In the 1994 experiment, plants were caged on days 0-2 and 6-8 of the 10-d experiment.

Despite efforts to minimize plant stress, many plants were lost from our marked ropes and overall recovery of plants was modest. The first time the experiment was performed, from 25 October-10 November 1993, we recovered 8 of 30 paired halves. After using some tissues from these plants to determine their palatability to amphipods, only four of the pairs had enough healthy tissue to analyze for secondary metabolites, and only two pairs had enough tissue remaining for protein analysis. The experiment was repeated from 8 to 18 September 1994, and 18 of 30 paired halves were recovered. Immediately after the first experiment in 1993, we used a digitizer to measure amphipod grazing scars and determine the percentage of thallus area removed by amphipods in each of our eight treatment and control plant portions.

The palatability to Ampithoe of amphipod-damaged vs. undamaged plant halves was determined by offering groups of four amphipods a choice of 90 mg WM of each plant half. The amphipods from the 1993 experiment fed for 2.0 d (N = 8) and the amphipods from the 1994 experiment fed for 1.0 d (N = 18). The concentrations of secondary metabolites (N = 14) and soluble protein (N = 14) were determined as described above. The data from the 2 yr were pooled and analyzed with paired-sample t tests (i.e., paired by plant). Directed P values with [Gamma]/[Alpha] = 0.8, as suggested by Rice and Gaines (1994), were calculated when testing for differences in palatability and secondary metabolites based on the prediction that previously damaged Dictyota would be less palatable and have more chemical defenses than undamaged plants.

RESULTS

Effect of secondary metabolites on herbivore feeding behavior

Dictyol E reduced feeding by the three herbivores that we assayed [ILLUSTRATION FOR FIGURE 1 OMITTED]. Natural concentrations of dictyol E in Dictyota menstrualis range from 0.02 to 0.045% WM ([ILLUSTRATION FOR FIGURE 2 OMITTED], Cronin et al. 1995). This compound thus deterred the pinfish Lagodon rhomboides at only 27-60% of natural concentrations, but deterred the sea urchin Arbacia and the amphipod Ampithoe only at the highest range of its natural concentration [ILLUSTRATION FOR FIGURE 1 OMITTED]. Another secondary metabolite in Dictyota, pachydictyol A, deters the sea urchin at, or below, natural concentrations, but does not deter the pinfish or the amphipod until concentrations reach 3-7 times natural levels (see Cronin and Hay 1996a). Thus, feeding by pinfish is very sensitive to natural levels of dictyol E, feeding by urchins is deterred by natural levels of pachydictyol A, and feeding by the amphipod is relatively insensitive to average concentrations of either compound; however, dictyol E can depress feeding at the highest of natural concentrations. Our field observations are consistent with these findings in that we rarely see fish or urchin grazing scars on Dictyota, but amphipod scars are common in some habitats.

The instability of dictyodial (a third dictyol class diterpene produced by Dictyota) once it is purified prevented us from obtaining enough pure compound to construct a standard curve in order to determine rigorously its absolute concentrations in our HPLC analysis of Dictyota. Relative concentrations (i.e., peak area per tissue mass) of dictyodial, however, varied by a factor of 5-6 among individual plants. Because we currently lack the techniques to prevent dictyodial from degrading during feeding assays, we were unable to test rigorously the effects of this compound on herbivore feeding. Some data suggest that dictyodial might deter the sea urchin but not the pinfish or amphipod (Cronin and Hay 1996a).

Temporal and among-site variation in Dictyota-herbivore interactions

In 1991, 1992, and 1993, subjective assessments of amphipod grazing scars (i.e., assigned a number of 03 that ranged from 0 = not grazed to 3 = heavily grazed) indicated that scars were more common on plants collected from the seagrass bed or mudflat than on plants collected from the jetty (P [less than] 0.003 for each year, Table 1). Plants from the mudflat and grass bed habitats were usually ranked as moderately (=2) to heavily (=3) grazed, while jetty plants were usually ranked as undamaged (=0) to little damaged (=1).

Dictyota collected from different sites often differed in nutritive value (e.g., nitrogen and soluble protein), concentrations of secondary metabolites, and/or the number of amphipods associated with the plants [ILLUSTRATION FOR FIGURE 2 OMITTED], but differences in these traits often varied among sampling dates. In October-November 1991, Dictyota collected from the Mitchell Village seagrass bed, Lennoxville Point mudflat, and Radio Island Jetty had differing estimates of amphipod damage that were positively related to the density of Ampithoe and other amphipods on the seaweeds (i.e., Mitchell Village [greater than] Lennoxville Point [greater than] Radio Island for both measures; Table 1, [ILLUSTRATION FOR FIGURE 2 OMITTED]). In no-choice feeding assays, Dictyota plants from the jetty were more susceptible to attack by Ampithoe than plants collected from the seagrass bed or mudflat [ILLUSTRATION FOR FIGURE 2-1b OMITTED], but sea urchins did not distinguish among plants from Mitchell Village, Lennoxville Point, and Radio Island (54 [+ or -] 13, 78 [+ or -] 17, and 66 [+ or -] 17 mg consumed per urchin per 4.1 d, respectively; mean [+ or -] 1 SE; P = 0.55). Because amphipods were significantly deterred by dictyol E at 0.058% WM but not at 0.035% WM [ILLUSTRATION FOR FIGURE 1C OMITTED], the significant 100% higher concentration of dictyol E in mudflat and seagrass bed plants ([approximately equal to]0.045% WM) relative to jetty plants ([approximately equal to]0.022% WM) [ILLUSTRATION FOR FIGURE 2-1d OMITTED] could explain the decreased feeding by Ampithoe on the mudflat and seagrass bed plants relative to jetty plants. Pachydictyol A and dictyodial were also significantly higher in the Ampithoe-resistant plants from the soft substrate sites [ILLUSTRATION FOR FIGURE 2-1d OMITTED], so avoidance of these plants could have been due to the combined effects of these dictyol class metabolites.

Dictyota from the Mitchell Village seagrass and the Lennoxville Point mudflat bed did not differ significantly in concentration of secondary metabolites [ILLUSTRATION FOR FIGURE 2-1d OMITTED], but plants from Mitchell Village had significantly lower concentrations of soluble protein [ILLUSTRATION FOR FIGURE 2-1c OMITTED], total nitrogen (1.2 [+ or -] 0.1 vs. 2.1 [+ or -] 0.1% dry mass [DM]), and total carbon (24 [+ or -] 1 vs. 28 [+ or -] 1% DM). Among-site differences in soluble protein, nitrogen, or carbon content are unlikely to have strongly affected the preference of Ampithoe because the more susceptible jetty plants had intermediate levels of soluble [TABULAR DATA FOR TABLE 1 OMITTED] protein [ILLUSTRATION FOR FIGURE 2-1C OMITTED], levels of total nitrogen (1.3 [+ or -] 0.1% DM) similar to plants from the seagrass bed, and levels of total carbon (21 [+ or -] 1% DM) lower than the other two sites.

The pattern of differences in Dictyota between Radio Island Jetty and the Lennoxville Point mudflat observed in October-November 1991 was not observed in July 1992 (Fig. 2-1 vs. 2-2). Densities of Ampithoe did not differ between the sites; none were found on the 20 plants collected from Radio Island Jetty, and the number of Ampithoe per gram wet mass of alga did not differ significantly from zero for plants from Lennoxville Point. However, grazing scars appeared more abundant on mudflat than jetty plants (Table 1, P [less than] 0.0001) and the density of non-Ampithoe amphipods was greater at the mudflat than the jetty (“other pods” in [ILLUSTRATION foR FIGURE 2-2A OMITTED]). Ampithoe exhibited no feeding preference between mudflat or jetty plants whether the seaweeds were offered individually (i.e., no-choice) or simultaneously (i.e., choice) [ILLUSTRATION FOR FIGURE 2-2B OMITTED]. Plants from these two sites also did not differ significantly in concentrations of soluble protein [ILLUSTRATION FOR FIGURE 2-2C OMITTED] or secondary metabolites [ILLUSTRATION FOR FIGURE 2-2D OMITTED]. Two months later, jetty and mudflat Dictyota again did not differ significantly in their susceptibility to Ampithoe grazing in either choice or no-choice a[ILLUSTRATION FOR FIGURE 2-3B OMITTED], although plants from the jetty had 47% less soluble protein ([ILLUSTRATION FOR FIGURE 2-3C OMITTED], P = 0.001) and 63% more dictyodial than mudflat plants ([ILLUSTRATION FOR FIGURE 2-3D OMITTED], P = 0.025). As with the July 1992 collection, concentrations of dictyol E and pachydictyol A did not differ significantly between sites [ILLUSTRATION FOR FIGURE 2-3D OMITTED].

In 1993, the susceptibility of mudflat and jetty Dictyota to Ampithoe did not differ significantly in July [ILLUSTRATION FOR FIGURE 2-4B OMITTED] or October [ILLUSTRATION FOR FIGURE 2-5B OMITTED], although, in October, mudflat plants had more Ampithoe, and more non-Ampithoe amphipods than jetty plants [ILLUSTRATION FOR FIGURE 2-5A OMITTED] and were subjectively classified as having more amphipod grazing scars (Table 1). However, Ampithoe densities at Radio Island Jetty in 1993 were considerably higher than in the previous 2 yr [ILLUSTRATION FOR FIGURE 2-A OMITTED]. The susceptibility of Dictyota to the urchin Arbacia also did not differ significantly in July 1993 (Lennoxville Point vs. Radio Island; 86 [+ or -] 19 vs. 102 [+ or -] 19 mg consumed per urchin per 3 d; P = 0.58). As in the previous year, the amount of soluble protein in Dictyota did not differ between sites in July [ILLUSTRATION FOR FIGURE 2-4C OMITTED], but the mudflat plants did have significantly more soluble protein than jetty plants a few months later [ILLUSTRATION FOR FIGURE 2-5C OMITTED]. One significant difference in secondary metabolite concentration was detected on both dates in 1993; jetty plants had 31% less pachydictyol A than mudflat plants in July [ILLUSTRATION FOR FIGURE 2-4D OMITTED] and 12% less dictyol E in October [ILLUSTRATION FOR FIGURE 2-5D OMITTED].

Within-site and within-plant variation of D. menstrualis characteristics

When undamaged and amphipod-damaged Dictyota were collected from the Lennoxville Point mudflat and offered to Ampithoe as a choice, the amphipods consumed [approximately equal to]37% more of the ungrazed than the previously grazed algae [ILLUSTRATION FOR FIGURE 3 OMITTED]. When these data were analyzed according to Peterson and Renaud (1989), the differences in mass change between damaged and undamaged Dictyota in replicates with amphipods (N = 23, difference = – 11.7 [+ or -] 6.6, mean [+ or -] 1 SE) vs. replicates without amphipods (N = 7, difference = +7.0 [+ or -] 2.3) was statistically significant (P = 0.012, two-sample t test with separate variances), indicating that Ampithoe prefers ungrazed over previously grazed plants [ILLUSTRATION FOR FIGURE 3 OMITTED].

The three different portions (apices, middles, and adventitious branches) of amphipod-damaged Dictyota from the mudflat demonstrated within-plant variation in toughness, soluble protein, and concentrations of all three secondary metabolites, but this variation did not result in significant differences in susceptibility to grazing by Ampithoe when offered in no-choice assays [ILLUSTRATION FOR FIGURE 4 OMITTED]. Middle portions of plants were significantly tougher than adventitious branches, which were significantly tougher than apices [ILLUSTRATION FOR FIGURE 4B OMITTED]. The concentration of soluble protein in adventitious branches did not differ significantly from that of middles or apices; however, apices did have significantly higher concentrations of soluble protein than middles. The within-plant pattern of dictyol E and dictyodial concentrations were similar; adventitious branches had significantly higher concentrations of both compounds than middles or apices, both of which had similar concentrations [ILLUSTRATION FOR FIGURE 4D OMITTED]. The concentrations of pachydictyol A in middles and adventitious branches did not differ significantly, but this compound was significantly less concentrated in apices than the other plant parts [ILLUSTRATION FOR FIGURE 4D OMITTED]. No significant within-plant difference in the concentration of sterols was detected.

Manipulation of amphipod grazing on D. menstrualis

When grazing by Ampithoe was manipulated while other (e.g., physical environment and genotype) factors were controlled, minimally grazed control halves were eaten 62% faster than the heavily grazed plant halves by Ampithoe that were offered a choice between the two types of tissue in the laboratory ([P.sub.dir] = 0.036, [ILLUSTRATION FOR FIGURE 5A OMITTED]). Thallus toughness and soluble protein were unaffected by the grazing treatment [ILLUSTRATION FOR FIGURE 5B AND C OMITTED]. However, heavily grazed plants had significantly higher concentrations of dictyol E (+19%), dictyodial (+22%), pachydictyol A (+34%), and sterols (+9%) [ILLUSTRATION FOR FIGURE 5D OMITTED]. During our first effort at this experiment in 1993, plant portions subjected to supplemental amphipod grazing had grazing scars that accounted for 3.3 [+ or -] 1.1% of their total thallus area, while control plant portions lost only 0.2 [+ or -] 0.1% of their thallus area to amphipods ([P.sub.dir] = 0.013, paired-sample t test, N = 8). We did not quantify grazing scars in 1994 but damage levels appeared similar to those measured in 1993.

DISCUSSION

Mensurative experiments suggest that Dictyota menstrualis induces increased levels of chemical defenses in response to grazing by Ampithoe longimana [ILLUSTRATION FOR FIGURES 2 AND 3 OMITTED], the herbivore that appears to be the most important consumer of Dictyota at our study sites (Hay et al. 1987, Duffy and Hay 1991, 1994, Cronin and Hay 1996b). Dictyota collected from sites with 1-2 orders of magnitude higher densities of Ampithoe was less susceptible to further attack and had higher concentrations of secondary metabolites than Dictyota collected from a site with few Ampithoe [ILLUSTRATION FOR FIGURE 2-1 OMITTED]. When the densities of Ampithoe were more similar between sites, the susceptibility of Dictyota to Ampithoe and levels of chemical defenses were more similar [ILLUSTRATION FOR FIGURE 2-2 OMITTED]. Between-site differences in Ampithoe density and grazing scars, however, did not always result in differences in palatability and chemical defenses [ILLUSTRATION FOR FIGURE 2-5 OMITTED]. This may occur because grazing thresholds need to be exceeded before induction occurs, or because these between-site and between-year comparisons are complex and affected by factors other than grazing damage by Ampithoe. Site-specific variables were eliminated by comparing amphipod-grazed and ungrazed Dictyota from within the mudflat site. That some of these plants were ungrazed while neighboring plants had grazing scars was not due to high herbivore resistance in undamaged plants. Ampithoe significantly preferred undamaged over amphipod-damaged plants when these were made equally available in laboratory assays [ILLUSTRATION FOR FIGURE 3 OMITTED]. These observations of lower susceptibility to herbivory of grazer-damaged plants are consistent with the hypothesis that Dictyota increases levels of chemical defenses when attacked by amphipods; however, they do not exclude other possible explanations.

A controlled field experiment where individual plants from the jetty were split and subjected to increased densities of Ampithoe or to the ambient, low levels of Ampithoe found on the jetty demonstrated: (1) that Dictyota portions attacked by Ampithoe ultimately contained higher concentrations of secondary metabolites than paired portions not attacked by amphipods, (2) that plant portions with these increased chemical defenses were less susceptible to further attack, and (3) that this change in plant susceptibility to amphipods did not occur due to changes in plant toughness or concentrations of soluble protein [ILLUSTRATION FOR FIGURE 5 OMITTED]. The increase in levels of secondary metabolites in amphipod-damaged plants potentially could have been caused by (1) plants increasing their production of secondary compounds or (2) amphipods selectively removing low-dictyol portions of the plant and leaving portions that were on average higher in secondary metabolites. However, explanation 2 is unlikely because feeding scars on grazed plants recovered in 1993 represented only 3.3% of the surface area in the damaged plants, and only 0.2% of the surface area of control plants. Plants in the 1994 experiment appeared to have similar levels of damage. If we assume that amphipods consumed 3.3% of the experimental plants, and that this 3.3% totally lacked secondary metabolites (extremely unlikely given the pattern in [ILLUSTRATION FOR FIGURE 4D OMITTED]), then the potential effect of preferential grazing could account for only a 3.4% increase in whole plant concentrations of secondary metabolites. We documented increases of 19-34% in Dictyota secondary metabolites [ILLUSTRATION FOR FIGURE 5D OMITTED], indicating that the plants were actually inducing increased levels of defensive metabolites. Induction is also indicated by the finding that adventitious branches growing from damaged areas of the plant produced higher levels of secondary metabolites than did apices [ILLUSTRATION FOR FIGURE 4 OMITTED].

Because small structural differences among the secondary metabolites produced by Dictyota can have large effects on plant susceptibility to specific herbivores (see review in Hay and Steinberg 1992), it seems possible that the alga could respond to herbivores by preferentially increasing the most effective compound against the currently harmful herbivore. However, when grazed by Ampithoe, Dictyota responded by increasing the concentrations of all the secondary metabolites by 19-34%. In contrast, the primary metabolites we measured (sterols) were increased by only 9%, although this was statistically significant. The alga might have produced this general increase in lipophilic metabolites because it is a general response to thallus damage, because the different secondary compounds may be equally effective against amphipods (e.g., the minimum deterrent concentration assayed with Ampithoe was 0.058% WM for both pachydictyol A and dictyol E; [ILLUSTRATION FOR FIGURE 1C OMITTED], Cronin and Hay 1996a), or as a third possibility, because there could be important synergistic or additive effects of the different compounds affecting feeding by Ampithoe (see Hay et al. 1994).

Unlike Ampithoe, which has high and similar thresholds for tolerance of both pachydictyol A and dictyol E, other herbivores were more sensitive to small structural differences among Dictyota secondary metabolites. For example, dictyol E differs from pachydictyol A only by the substitution of a hydroxyl group on the side branch (compare structures in [ILLUSTRATION FOR FIGURE 5D OMITTED]), yet dictyol E deterred feeding by the pinfish Lagodon rhomboides at 0.023% WM [ILLUSTRATION FOR FIGURE 1A OMITTED] while pachydictyol A did not significantly affect the pinfish at 2.5 times this concentration (Cronin and Hay 1996a). The different effects of similar compounds also depend on the herbivore species; 0.023% WM dictyol E did not significantly affect the feeding behavior of the sea urchin Arbacia [ILLUSTRATION FOR FIGURE 1B OMITTED], but pachydictyol A significantly deterred urchins at only 56% of this concentration (Cronin and Hay 1996a). That Dictyota coexists with a diverse herbivore guild (Hay et al. 1987, 1988), of which only a subset is deterred by natural concentrations of any one compound, may explain why the alga produces multiple secondary metabolites.

Differential sensitivity to plant traits among herbivore species may explain why Dictyota from three sites were differentially susceptible to Ampithoe but not to Arbacia in 1991. Although the survivorship of Ampithoe is positively correlated with the protein content of food plants (Duffy and Hay 1991), protein content does not appear to be an important proximal cue for their feeding decisions (Duffy and Hay 1991, Cronin and Hay 1996a, b). Feeding by the amphipod appears to be more influenced by variation in the concentrations of secondary metabolites than protein ([ILLUSTRATION FOR FIGURES 2 AND 5 OMITTED], Cronin and Hay 1996a, b). Arbacia is more sensitive than Ampithoe to Dictyota secondary metabolites (Hay et al. 1987, Cronin and Hay 1996a), but the urchin’s feeding decisions may also depend on the protein content of food plants (Renaud et al. 1990, Cronin and Hay 1996c). It is possible that Arbacia did not feed preferentially on Dictyota from the jetty compared to the mudflat because the higher protein content of mud-flat plants [ILLUSTRATION FOR FIGURE 2-1C OMITTED] counter balanced the deterrent effects of their chemical defenses. Other studies have demonstrated that the effectiveness of prey defenses can vary according to the nutritional quality of the prey (Duffy and Paul 1992, Hay et al. 1994).

When we used plants collected from the mudflat or plant portions from our induction experiment at Radio Island Jetty, Ampithoe avoided previously damaged plants or plant portions compared to undamaged plants or plant portions ([ILLUSTRATION FOR FIGURE 2-1B OMITTED], [ILLUSTRATION FOR FIGURES 3 AND 5 OMITTED]). These observations, together with our demonstration of significantly greater concentrations of secondary metabolites in previously grazed plants, suggest that amphipods avoid previously damaged plants because these plants are induced to produce greater levels of chemical defenses. This interpretation could be in error if Ampithoe avoid plants with amphipod grazing scars as a way of minimizing intraspecific competition or detection by predators that cue on amphipod feeding scars. However, the amphipod did not distinguish between plants that differed in amounts of visible damage (Table 1) but had more similar levels of secondary metabolites [ILLUSTRATION FOR FIGURE 2-2 AND 2-5 OMITTED]. Additionally, artificial foods that were identical except for their concentrations of dictyol E [ILLUSTRATION FOR FIGURE 1 OMITTED] or pachydictyol A (Cronin and Hay 1996a) were treated differently by Ampithoe, with more chemically defended ones being avoided once dictyol E or pachydictyol A concentrations reached 0.058% WM.

Previous studies conflict somewhat with the above findings in that they show that Ampithoe preferentially consumes Dictyota over other seaweeds and that its feeding is unaffected, or even stimulated, when dictyol E or pachydictyol A are coated onto fresh seaweeds (Hay et al. 1987, Duffy and Hay 1994); however, evenly mixing the compound throughout our agar-based test food demonstrated that feeding by Ampithoe can be deterred by relatively high concentrations of these compounds ([ILLUSTRATION FOR FIGURE 1 OMITTED], Cronin and Hay 1996a). Although less affected by Dictyota metabolites than urchins and fishes, Ampithoe does choose Dictyota tissues with lower levels of these secondary metabolites ([ILLUSTRATION FOR FIGURE 2-5 OMITTED]; Cronin and Hay 1996a, b). Differences in secondary metabolites between the amphipod-damaged and undamaged halves of our experimental plants is the best supported explanation for their differential susceptibility to Ampithoe [ILLUSTRATION FOR FIGURE 5 OMITTED]. In contrast to these generally consistent patterns of amphipods selecting plants with lower levels of chemical defenses, Ampithoe did not avoid chemically rich adventitious branches in contrast to less chemically rich apices and middles [ILLUSTRATION FOR FIGURE 4 OMITTED]. Thus, secondary chemistry alone is not always sufficient to explain Ampithoe feeding choices.

The limited number of investigations evaluating induced chemical defenses in seaweeds suggest that induction may be uncommon. Although both Van Alstyne (1988) and Yates and Peckol (1993) found increased levels of phlorotannins in different species of Fucus that they subjected to artificial clipping (Van Alstyne also showed that herbivorous snails fed less on the induced plants), numerous other investigations of both brown and green seaweeds have failed to find evidence of induction (Paul 1992, Paul and Van Alstyne 1992 – for terpenoids in Halimeda, Udotea, and Caulerpa; Pfister 1992 – for phlorotannins in Alaria; and Steinberg 1994, 1995 – for phlorotannins in Sargassum and Ecklonia). With the exception of Steinberg (1995), all of the above studies used artificial clipping instead of herbivores to damage plants. Because simulated herbivory may not mimic the effects of herbivores (Baldwin 1990, Renaud et al. 1990), studies should be conducted using the herbivores rather than artificial clipping whenever possible. Additionally, most previous studies have searched for induction by measuring the levels of specific secondary metabolites or classes of compounds in damaged vs. undamaged plants rather than by measuring the susceptibility of these plants to relevant herbivores. Because plants could be inducing increased defenses by altering levels of chemical defenses that are not known (and, thus, are not measured) or via plant traits other than chemistry (see Steneck and Adey 1976, Lubchenco and Cubit 1980, Hay 1981, Lewis et al. 1987 for examples), feeding by herbivores may provide a much more sensitive measure of induction than analysis via HPLC or other common chemical techniques. As one possible example, Renaud et al. (1990) found that the brown alga Padina became significantly less palatable to sea urchins within 16 h of previous attack; however, they were unable to identify a specific chemical responsible for this effect.

The apparent rarity of induced chemical defenses in seaweeds has been suggested to be due to (1) a greater intensity and spatial predictability of herbivory in marine vs. terrestrial systems, which could select for constitutive rather than inducible defenses (Paul and Van Alstyne 1992, Steinberg 1994), or (2) the difficulty for structurally simple seaweeds (i.e., lacking a vascular system, etc.) of efficiently propagating a signal from the area of grazing damage to the rest of the plant thallus or translocating defenses to the area currently under attack (Cronin and Hay 1996b). Alternately, (3) the limited number of seaweed studies that are available may not be representative of the general patterns exhibited by seaweeds, or (4) the common use of artificial clipping instead of herbivores to damage seaweeds may result in an underestimation of induction.

In vascular plants, a systemic induction of defense in response to localized grazing damage could occur via the vascular transport system, although airborne cues may also be important (Baldwin and Schultz 1983). Although most seaweeds lack a vascular system, waterborne cues could elicit a systemic response in seaweeds, as they do in some marine bryozoans (Harvell 1990). This mechanism was probably unimportant for the induced response observed by Van Alstyne (1988) in Fucus disticus because the noninduced, control plants and the induced, damaged plants were separated by only 2-3 cm. The grazed and ungrazed portions of Dictyota used in our field experiments were separated by 20-30 cm, so if any waterborne cues were used, they were ineffective at stimulating equal levels of induction in plants that were separated by these distances.

To see if induction of increased defenses in Dictyota was localized within the immediate vicinity of the grazing damage or occurred over the entire thallus, we evaluated the concentrations of secondary metabolites in portions of individual thalli that differed in recent history of grazing, as evidenced by adventitious branches growing from what we assumed were amphipod grazing scars. Our assumption that adventitious branches usually resulted from previous amphipod grazing seems reasonable in that (1) adventitious branching from damaged areas of seaweed thalli is common among a variety of different seaweeds (Isaac 1956, Dixon 1958, 1960, Moss 1961, Fulcher and McCully 1969, Van Alstyne 1989), (2) plants with adventitious branches were common at sites where amphipods were abundant and rare where amphipods were rare, and (3) we commonly observed adventitious branches growing from what we recognize as amphipod grazing scars. Adventitious branches of Dictyota had elevated levels of dictyol E and dictyodial and intermediate levels of thallus toughness and soluble protein compared to the apices and middles of individual plants [ILLUSTRATION FOR FIGURE 4 OMITTED]. Although the elevated concentrations of secondary metabolites in adventitious branches are consistent with localized induction of chemical defenses near damaged portions of the thallus, these within-plant differences of Dictyota did not result in differential feeding by Ampithoe in the laboratory [ILLUSTRATION FOR FIGURE 4A OMITTED]. However, levels of secondary metabolites in all portions of the plants used in this experiment were relatively high compared to levels in the jetty plants that were selectively consumed by amphipods (compare [ILLUSTRATION FOR FIGURE 2-1D AND FIGURE 4 OMITTED]). These patterns suggest that defensive compounds are elevated in new tissues that grow from damaged cells near grazing scars, but that a lesser degree of induction may also occur throughout the entire plant thallus. In our field experiment [ILLUSTRATION FOR FIGURE 5 OMITTED] and in the plant portions we analyzed [ILLUSTRATION FOR FIGURE 4 OMITTED], the grazing activity of the numerous amphipods may have been diffuse enough that a mosaic of isolated responses produced increased levels of defenses throughout the damaged thalli. Clearly, more research is needed to determine the mechanisms and stimuli responsible for induction in seaweeds.

For a congener of Dictyota menstrualis (D. ciliolata), Ampithoe did respond to within-plant variation in chemical defenses, preferring to consume less chemically rich apices over more chemically rich middles (Cronin and Hay 1996b). Australian amphipods feeding on the brown alga Zonaria angustata also appear to minimize exposure to its chemical defenses by feeding preferentially on those plant portions that are lowest in the physodes that contain phlorotannins (Poore 1994). Additionally, in a situation similar to ours, Van Alstyne (1989) demonstrated that the brown seaweed Fucus distichus produced adventitious branches at hero bivore grazing scars and that these adventitious branches were less susceptible to grazing by littorine snails and had higher concentrations of phlorotannins relative to apical meristems. Just as these marine herbivores avoid the chemical defenses of seaweeds they commonly eat, many terrestrial insects avoid the chemical defenses of favored host plants, and perform behaviors like vein cutting or leaf trenching to reduce exposure to the defenses (Dussourd and Denno 1991).

Just as herbivore activity creates intraspecific and interspecific variation in the distribution and abundance of plants, it is becoming increasingly clear that plants actively alter their tissue quality in ways that can affect the distribution of herbivores and enemies of herbivores (Rowell-Rahier and Pasteels 1990, 1992, Schultz 1992). Indirect effects associated with induced changes in Dictyota could be important. For example, experiments have shown that Ampithoe avoids detection and consumption by omnivorous fishes by being relatively immobile and living on chemically-defended plants that are avoided by fishes (Hay et al. 1987, Duffy and Hay 1991, 1994). If these amphipods, or other herbivores, must increase movement as algal resources change, they could become more susceptible to these visual predators. Alternatively, if grazing induces plant secondary metabolites that the amphipods can tolerate better than the fishes, then plants may become better refuges when grazed. Additionally, the production of adventitious branches increases the morphological complexity of the plants, which could alter the value of plants as refuges from predators (Hacker and Steneck 1990). The active responses of seaweeds to herbivores affect the dynamics of seaweed – herbivore interactions and could influence higher order interactions and have repercussions throughout the community.

ACKNOWLEDGMENTS

Funding was provided by NSF grants OCE 89-11872 and OCE 92-02847. Emmett Duffy allowed Us to use his unpublished data on the palatability of grazer-damaged vs. undamaged plants from Lennoxville Point. We thank Julie Cronin, Margaret Miller, and Buffy Turner for help with experiments. Emmett Duffy, James Estes, Niels Lindquist, David Lodge, Hans Paerl, Joe Pawlik, Charles Peterson, and two anonymous reviewers made helpful comments that improved the manuscript.

LITERATURE CITED

Baldwin, I. T. 1990. Herbivory simulation in ecological research. Trends in Ecology and Evolution 5:91-93.

—– 1994. Chemical changes rapidly induced by folivory. Pages 1-23 in E. A. Bernays, editor. Insect-plant interactions. Volume 5. CRC Press, Boca Raton, Florida, USA.

Baldwin, I. T., and J. C. Schultz. 1983. Rapid changes in tree leaf chemistry induced by damage: evidence for communication between plants. Science 221:227-229.

Belsky, A. J. 1986. Does herbivory benefit plants? A review of the evidence. American Naturalist 127:870-892.

Bradford, M. M. 1976. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72:248-254.

Coley, P. D. 1986. Costs and benefits of defense by tannins in a neotropical tree. Oecologia 70:238-241.

Cronin, G., and M. E. Hay. 1996a. Susceptibility to herbivores depends on recent history of both the plant and animal. Ecology 77:1531-1543.

Cronin, G., and M. E. Hay. 1996b. Within-plant variation in seaweed palatibility and chemical defenses: optimal defense theory versus the growth-differentiation balance hypothesis. Oecologia 105:361-368.

Cronin, G., and M. E. Hay. 1996c. Effects of light and nutrient availability on the growth, secondary chemistry, and resisitance to herbivory of two brown seaweeds. Oikos, in press.

Cronin, G., N. Lindquist, M. E. Hay, and W. Fenical. 1995. Effects of storage and extraction procedures on yields of lipophilic metabolites from the brown seaweeds Dictyota ciliolata and Dictyota menstrualis. Marine Ecology Progress Series 119:265-273.

Dixon, P.S. 1958. The structure and development of the thallus in the British species of Gelidium and Pterocladia. Annals of Botany Volume XXII 87:353-368.

—–. 1960. Studies on marine algae of the British Isles: Ceramium shuttleworthianum (Kutz) Silva. Journal of the Marine Biological Association of the United Kingdom 39: 375-390.

Duffy, J. E., and M. E. Hay. 1991. Food and shelter as determinants of food choice by an herbivorous marine amphipod. Ecology 72:1286-1298.

Duffy, J. E., and M. E. Hay. 1994. Herbivore resistance to seaweed chemical defense: the roles of mobility and predation risk. Ecology 75:1304-1319.

Duffy, J. E., and V. J. Paul. 1992. Prey nutritional quality and the effectiveness of chemical defenses against tropical reef fishes. Oecologia 90:333-339.

Dussourd, D. E., and R. F. Denno. 1991. Deactivation of plant defense: correspondence between insect behavior and secretory canal architecture. Ecology 72:1383-1396.

Faulkner, D. J. 1993. Marine natural products. Natural Products Reports 10:497-539.

Fulcher, R. G., and M. E. McCully. 1969. Histological studies on the genus Fucus. IV. Regeneration and adventive embryony. Canadian Journal of Botany 47:1643-1649.

Hacker, S. D., and R. S. Steneck. 1990. Habitat architecture and the abundance and body-size-dependent habitat selection of a phytal amphipod. Ecology 71:2269-2285.

Harvell, C. D. 1990. The ecology and evolution of inducible defenses. Quarterly Review of Biology 65:323-340.

Hay, M. E. 1981. The functional morphology of turf forming seaweeds: persistence in stressful marine habitats. Ecology 62:739-750.

—–. 1986. Associational plant defenses and the maintenance of species diversity: turning competitors into accomplices. American Naturalist 128:617-641.

Hay, M. E., J. E. Duffy, C. A. Pfister, and W. Fenical. 1987. Chemical defense against different marine herbivores: are amphipods insect equivalents? Ecology 68:1567-1580.

Hay, M. E., and W. Fenical. 1988. Marine plant-herbivore interactions: the ecology of chemical defense. Annual Review of Ecology and Systematics 19:111-145.

Hay, M. E., Q. E. Kappel, and W. Fenical. 1994. Synergisms in plant defenses against herbivores: interactions of chemistry, calcification, and plant quality. Ecology 75:1714-1726.

Hay, M. E., P. E. Renaud, and W. Fenical. 1988. Large mobile versus small sedentary herbivores and their resistance to seaweed chemical defenses. Oecologia 75:246-252.

Hay, M. E., and P. D. Steinberg. 1992. The chemical ecology of plant-herbivore interactions in marine versus terrestrial communities. Pages 371-413 in G. A. Rosenthal and M. R. Berenbaum, editors. Herbivores: their interactions with secondary plant metabolites. Volume II. Evolutionary and ecological processes. Academic Press, New York, New York, USA.

Herms, D. A., and W. J. Mattson. 1992. The dilemma of plants: to grow or defend. Quarterly Review of Biology 67:283-335.

Isaac, W. E. 1956. The ecology of Gracilaria confervoides (L.) Grey. in South Africa with special reference to its ecology in the Saldanha-Langebaan Lagoon. Pages 173-185 in T. Braarud and N. A. Sorenson, editors. Second International Seaweed Symposium. Pergamon, New York, New York, USA.

Karban, R., and J. H. Myers. 1989. Induced plant responses to herbivory. Annual Review of Ecology and Systematics 20:331-348.

Lewis, S. M., J. N. Norris, and R. B. Searles. 1987. The regulation of morphological plasticity in tropical reef algae by herbivory. Ecology 68:636-641.

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

Mattson, W.J. 1980. Herbivory in relation to plant nitrogen. Annual Review of Ecology and Systematics 11:119-161.

Moss, B. L. 1961. Wound healing and regeneration in Fucus vesiculosus L. Proceedings of the International Seaweed Symposium 4:117-122.

Niemela, P., and J. Tuomi. 1987. Does the leaf morphology of some plants mimic caterpillar damage? Oikos 50:256-257.

Paul, V. J., editor. 1992. Ecological roles of marine natural products. Cornell University Press, Ithaca, New York, USA.

Paul, V. J., and W. Fenical. 1986. Chemical defense in tropical green algae, order Caulerpales. Marine Ecology Progress Series 34:157-169.

Paul, V. J., and K. L. Van Alstyne. 1988. Chemical defense and chemical variation in some tropical Pacific species of Halimeda (Halimedaceae; Chlorophyta). Coral Reefs 6: 263-269.

Paul, V. J., and K. L. Van Alstyne. 1992. Activation of chemical defenses in the tropical green algae Halimeda spp. Journal of Experimental Marine Biology and Ecology 160: 191-203.

Peterson, C. H., and P. E. Renaud. 1989. Analysis of feeding preference experiments. Oecologia 80:82-86.

Pfister, C. A. 1992. Costs of reproduction in an intertidal kelp: patterns of allocation and life history consequences. Ecology 73:1586-1596.

Poore, A. G. B. 1994. Selective herbivory by amphipods inhabiting the brown alga Zonaria angusrata. Marine Ecology Progress Series 107:113-123.

Renaud, P. E., M. E. Hay, and T. M. Schmitt. 1990. Interactions of plant stress and herbivory: intraspecific variation in the susceptibility of a palatable versus an unpalatable seaweed to sea urchin grazing. Oecologia 82:217-226.

Rhoades, D. F. 1979. Evolution of plant chemical defense against herbivores. Pages 3-54 in G. A. Rosenthal and D. H. Janzen, editors. Herbivores: their interaction with secondary plant metabolites. Academic Press, New York, New York, USA.

—–. 1985. Offensive-defensive interactions between insects and plants: their relevance in herbivore population dynamics and ecological theory. American Naturalist 125: 205-238.

Rice, W. R., and S. D. Gaines. 1994. ‘Heads I win, tails you lose’: testing directional alternative hypotheses in ecological and evolutionary research. Trends in Ecology and Evolution 9:235-237.

Rowell-Rahier, M., and J. M. Pasteels. 1990. Phenolglucosides and interactions at three trophic levels: Salicaceae-herbivores-predators. Pages 75-94 in E. A. Bernays, editor. Insect-plant interactions. Volume II. CRC Press, Boca Raton, Florida, USA.

Rowell-Rahier, M., and J. M. Pasteels. 1992. Third trophic level influences of plant allelochemicals. Pages 243-277 in G. A. Rosenthal and M. R. Berenbaum, editors. Her-bivores: their interactions with secondary plant metabolites. Volume II. Evolutionary and ecological processes. Academic Press, New York, New York, USA.

Schultz, J. C. 1992. Factoring natural enemies into plant tissue availability to herbivores. Pages 175-197 in M. D. Hunter, T. Ohgushi, and P. W. Price, editors. Effects of resource distribution on animal-plant interactions. Academic Press, New York, New York, USA.

Steinberg, P. D. 1994. Lack of short-term induction of phlorotannins in the brown algae Ecklonia radiata and Sargassum vestitum. Marine Ecology Progress Series 112:129-133.

—–. 1995. Interaction between the canopy dwelling echinoid Holopneustes purpurescens and its host plant Ecklonia radiata. Marine Ecology Progress Series 127:169-181.

Steneck, R. S., and W. H. Adey. 1976. The role of environment in control of morphology in Lithophyllum congestum, a Caribbean algal ridge builder. Botanica Marina 19:197-215.

Tallamy, D. W., and M. J. Raupp. 1991. Phytochemical induction by herbivores. John Wiley & Sons, New York, New York, USA.

Van Alstyne, K. L. 1988. Herbivore grazing increases polyphenolic defenses in the intertidal brown alga Fucus distichus. Ecology 69:655-663.

—–. 1989. Adventitious branching as a herbivore-induced defense in the intertidal brown alga Fucus disticus. Marine Ecology Progress Series 56:169-176.

Williams, K. S., and L. E. Gilbert. 1981. Insects as selective agents on plant vegetative morphology: egg mimicry reduces egg laying by butterflies. Science 212:467-469.

Yates, J. L., and P. Peckol. 1993. Effects of nutrient availability and herbivory on polyphenolics in the seaweed Fucus vesiculosus. Ecology 74:1757-1766.

COPYRIGHT 1996 Ecological Society of America

COPYRIGHT 2004 Gale Group