BULLETIN OF MARINE SCIENCE, 40(1): 11-21, 1987

ROLE OF CUPREA BOSC (POLYCHAETA: ) TUBES IN STRUCTURING A SUBTROPICAL INFAUNAL COMMUNITY

Suzanne M. Ban and Walter G. Nelson

ABSTRACT An a priori hypothesis predicted that in the vicinity of aggregated Diopatra cuprea tubes an enhanced infaunal density and richness would be found, resulting from a biological refuge effect of the tubes. To test this hypothesis, cores were taken over a 5-month period in both vegetated, Halodule wrighti! Aschers. beds, and unvegetated areas of a site in the Indian River lagoon, Florida. An inner, 0.01 m2, frame was placed to enclose densities of 0, I, or 4 D. cuprea tubes, while an outer concentric, 0.02 m2, frame was placed so that it enclosed the smaller frame, plus a surrounding area lacking in D. cuprea tubes. The presence of D. cuprea tubes was found to have no consistent significant effect on the abundance and number of infaunal species found in either the vegetated or unvegetated areas. Laboratory experiments employing a benthic predator, Callinectes, were carried out in order to determine whether D. cuprea tubes andlor H. wrightii rhizome mats actually constitute a barrier to predation. Significantly higher survivorship of the bivalve Mulinia lateralis Say, used as prey, was found in laboratory treatments containing 10 tubes per 0.01 m2 versus treatments containing 4 or a tubes per 0.01 m2. Highest survivorship of bivalves was found in treatments containing a H. wrightii rhizome mat; tubes placed within the mat did not enhance clam survivorship. The discrepancy between the findings of this study, and previous studies on the refuge effect of D. cuprea tubes which found evidence to support the refuge hypothesis, indicates that there may be a critical lower limit of tube density that is needed to establish an effective refuge. This density was not found at the Indian River study site and may account for the lack of an observed refuge effect in the field data.

Natural disturbance can be important as a factor influencing the community structure of the marine benthos (Woodin, 1978; Thistle, 1980; 1981; Miller, 1982). Biological perturbation by motile predators (e.g., decapod crustaceans, Virnstein, 1977; Woodin, 1978; 1981; Holland et a1., 1980; Nelson, 1981; bottom- feeding fishes, Arntz, 1977; Virnstein, 1977; shorebirds, Goss-Custard, 1977; Luckenbach, 1984) can be a significant source of mortality to the infaunal benthos. Physical structures that reduce the susceptibility ofthe infauna to mortality caused by either physical or biological disturbance have been termed refuges (Woodin, 1978). Several biologically generated refuge structures have received recent study. The density and diversity of infaunal organisms have been found to be increased in association with biogenic refuges such as beds (Young et a1., 1976; Orth, 1977; Reise, 1977; Nelson, 1981; Peterson, 1982; Virnstein et a1., 1983), oyster shells (Dauer et a1., 1982), and worm tubes (Woodin, 1978; 1981; Wilson, 1979). Brenchley (1982) found that two kinds of biogenic structures, roots of seagrass and tubes of invertebrat~s, are able to deter the mobility of a number of burrowing predators. The combined effect of roots and tubes in the substrate has the greatest impact on predator mobility. The tube structures of the onuphid Diopatra cuprea have been found to be one such biogenic refuge. Woodin (1978; 1981) found elevated infaunal density and diversity around aggregated D. cuprea tubes in a Virginia mud flat. In addition, Luckenbach (1984) found that D. cuprea tubes appear to provide an effective refuge for macrofauna from shorebird predation. However, in a recent

II 12 BULLETIN OF MARINE SCIENCE, VOL. 40, NO.1, 1987 study, Belland Woodin (1984) found the spatial patterns ofmeiofaunaand selected macrofauna in relation to tubes to be highly variable. The discrepancy between the findings of Bell and Woodin (1984) and earlier reports of increased macrofaunal density in D. cuprea rich areas (Woodin, 1978; 1981) suggests that a number of unanswered questions remain. Most previous work on the refuge effect of D. cuprea tubes has been conducted in temperate areas. First, it is important to investigate whether results found in these temperate zone studies can be extrapolated to a functionally different subtropical system where small decapod predators are more numerous (Virnstein, 1977; 1978) and D. cuprea densities are lower. In addition, an understanding of the relative im- portance of the refuge effect of D. cuprea tubes within H. wrightii beds is needed since vegetated areas constitute a large part of subtropical subtidal systems. Fi- nally, although Woodin (1978; 1981) suggests that disturbance to the infauna by Callinectes sapidus Rathbun is reduced by the presence of D. cuprea tubes, she does not document their specific behavior around the tubes. The behavior of Callinectes in the vicinity of aggregated D. cuprea tubes must be examined in order to determine whether in fact tubes constitute an effective refuge from Cal- linectes predation. The present study seeks to examine these three questions.

METHODS AND MATERIALS

Field Study. -Fieldwork was conducted at a site 1.5 km south of the Sebastian Inlet in the Indian River lagoon, Florida (80029'W, 27°50'N). The study site was located in the shallow subtidal zone along a mangrove dominated shoreline. The bottom closest inshore consisted of interspersed vegetated and unvegetated patches, grading into larger grassbeds in the offshore direction. An offshore sandbar separated H. wrightii grassbeds from those dominated by Syringodium filiforme. Maximum depth inshore of the sandbar was approximately I m. A detailed description of the study site can be found in Ban (1985). To examine the association between the presence (and/or abundance) of D. cuprea tubes and the abundance and diversity of the remainder of the infauna, sediment cores were taken from both vegetated H. wrightii beds and unvegetated sandy areas following the methods of Woodin (1978). Sampling was done over a 5-month period, from May to September 1983. Observations at the site indicated highest D. cuprea population densities were found during this period. Sampling was dis- continued in October 1983 due to the low density of D. cuprea. Sampling employed the use of two aluminum frames, 0.01 m2 x 14 cm deep and 0.02 m2 x 14 cm deep. The smaller frame.was oriented so that it enclosed an area with 0, 1, or 4 D. cuprea tubes. The maximum of 4 tubes was chosen because this is the largest number of worm tubes that could be consistently found together in 0.0 I m2• The outer 0.02-m2 frame was placed so that it enclosed the smaller frame plus the surrounding area containing 0 D. cuprea. Thus, two concentric 0.01 m2 x 14-cm-deep samples were obtained, the inner containing a known density of worm tubes, the outer containing no D. cuprea. Tube caps and seagrass blades were cut off at the sediment surface prior to the removal of the core in order to minimize contamination by epifaunal organisms. Sampling was done at low tide, in water depths of between 10 and 30 cm, to facilitate the search for the tubes and the removal of the cores. The tidal range in this area is small, less than 50 cm, and subject to weather conditions (G. Swain, pers. comm.). Three replicates of each D. cuprea density were taken each month. Upon removal from the substrate with a shovel, cores were separated and washed on 0.5-mm Nitex mesh squares to remove sediment. Samples were immediately fixed in 7% Formalin. At least 48 h later, the preserved samples were again washed through a 0.5-mm screen and stored in a 70% alcohol- rose bengal solution. These samples were sorted, identified, and only whole and heads were enumerated. Samples were tested by three-way analysis of variance (ANOVA) with replication to analyze the response of macrofauna I abundance and species richness to the main effects of time, D. cuprea density, and vegetation presence or absence. Prior to computation of the ANOV A statistics, the homogeneity of variances was checked by the Bartlett test (Sokal and Rohlf, 1981) and a 10glO(x+ I) transformation was employed where needed. Laboratory Experiments. - These were conducted in glass aquaria each divided into three sections, designed to allow direct observation and determination of the effect of D. cuprea tubes on the foraging techniques of Callinectes. Juvenile Callinectes (approximately 8 cm in carapace width) were collected in the early spring from the Sebastian sampling site. The crabs were only identified to the level of BAN AND NELSON: ROLE OF DIOPATRA CUPREA TUBES IN AN INFAUNAL COMMUNITY 13 genus. The dominant species in this area is Callinectes sapidus, but C. ornatus and C. similis are also present (Gore et aI., 1981). Mulinia latera lis of the size range 10-20 mm were collected from the Indian River and used as a food source. Sediment from the field study site was allowed to become azoic by air drying for at least 30 days prior to the commencement of the experiment, and was used as a sediment base in the aquaria. Experiment I tested the null hypothesis that higher tube densities are more effective in deterring crab foraging. Three treatments, each replicated three times were used: I) a tube density of 4/0.0 I m2, the maximum consistently found at the Indian River study site; 2) a tube density of 10/0.01 m2, the maximum found by Woodin (1978) at Tom's Cove, Virginia; and 3) zero tubes, to act as a control. Large D. cuprea tubes collected from the field study site were pushed into the sediment and reinforced from the inside with toothpicks. Twenty M. latera lis were placed under the azoic sediment proximate to the tubes in each section, or in the center of the control treatments, and served as prey items. One crab that had been starved for 48 h was placed in each replicate and was allowed to forage for I h. In a series of preliminary tests, this feeding exposure time was determined to be sufficient for feeding to take place without complete removal of the prey. After I h, the crabs were removed, and the sand examined for uneaten clams. Differences in number of clams remaining among treatments were tested for significance by one-way ANOVA and a posteriori Welsch Step-Up Tests (Sokal and Rohlf, 1981). Assumption of heterogeneity of variances was examined by the Bartlett test (Sokal and Rohlf, 1981) and transformation was found to be unnecessary. Experiment 2 tested the null hypothesis that the combination of vegetation and tubes has the greatest effect on crab foraging. H. wrightii rhizome mats were removed from the field sampling site, isolated, and washed in seawater to remove infauna and epifaunal organisms. The mat was placed on a base of azoic sediment and more sediment was added, such that the original three dimensionality of the mat was retained (Brenchley, 1982). The four treatments, again replicated three times, were: 1) a vegetation mat and D. cuprea tubes (4/0.01 m2), 2) a vegetation mat alone, 3) D. cuprea tubes alone (4/0.01 m2), and 4) sediment alone (control). As in Experiment 1,20 M. latera lis were provided per replicate as a food source. Duration of the feeding exposure was the same as the previous experiment. The data were analyzed by one-way ANOVA and a posteriori Welsh Step-Up Tests. The final set of experiments was designed to allow the observation and documentation of crab behavior in the presence of D. cuprea tubes. The treatments, replicated three times and conducted in a randomized block design because it was impossible to observe more than four crabs simultaneously, were: 1) D. cuprea tubes (4/0.01 m2) and M. lateralis (20), to observe crab behavior in the presence of the tubes with a food stimulus; 2) D. cuprea tubes (4/0.01 m2) but no M. latera lis. to observe crab behavior around tubes without a food stimulus; 3) M. latera lis (20) but no tubes, to observe crab behavior without the presence of tubes; and 4) azoic sediment only, to act as a control. The following four types of crab behavior were noted: 1) foraging, or searching through the sediment with the chelipeds and walking legs; 2) feeding, or actually retrieving a prey item, breaking it open, and ingesting it; 3) roving, or walking about the aquarium but not foraging; and 4) resting, or hiding beneath the sediment. The duration of each of these actions was recorded for a total of I h of exposure time. A two-way ANOV A without replication was done to determine if block effects could be ruled out and if differences in crab foraging time occurred among the treatments.

RESULTS Field Study. - Table 1 compares mean macrofaunal densities associated with D. cuprea tubes each month in the inner and outer samples. Vegetated samples possessed consistently higher abundances of organisms than did the unvegetated samples. In the unvegetated areas, densities in May were generally higher than during the summer months of June through August (Table 1). In the vegetated areas, densities were generally lowest in June, while peak periods of abundance varied among treatments (Table 1). Table 2 summarizes the effects of three pa- rameters (substrate type, tube density, and time of year) on total abundance and species richness of the infauna (see Ban, 1985, for individual ANOVA tables). The presence of D. cuprea tubes had no effect on the abundance or number of species found in the Indian River cores (Table 2). For both abundance and species richness, there was a significant substrate effect due to higher infaunal numbers and diversity in the vegetated areas (Table 1). Seasonal changes significantly affected the infaunal community; however, this study was not primarily concerned with seasonal factors, and the time effect was not investigated further. There were no significant interaction terms. 14 BULLETIN OF MARINE SCIENCE, VOL. 40, NO. I, 1987

Table 1. Comparison of mean densities of macrofauna (±standard deviation) in inner and outer cores (N = 3). Surface area of cores = 0.01 m2

Unvegetated areas Vegetated areas Month o Tubes 1 Tube 4 Tubes o Tubes 1 Tube 4 Tubes Inner cores May 16.0 8.0 34.6 55.3 72.0 44.3 (7.8) (4.3) (12.1) (20.0) (44.0) (32.6) June 2.0 5.3 5.7 6.7 9.3 20.0 (3.5) (2.1) (5.5) (4.2) (6.5) (17.0) July 3.0 6.0 11.3 80.3 38.0 55.3 (1.0) (1.7) (9.5) (54.4) (20.7) (29.7) August 8.0 4.3 10.7 35.7 40.0 28.3 (5.0) (0.6) (6.5) (11.2) (9.2) (6.4) September 15.8 15.7 7.7 44.7 37.7 14.3 (13.8) (12.5) (3.8) (23.6) (5.0) (8.9) Outer cores May 45.0 22.7 24.0 138.3 144.0 127.7 (8.8) (2.8) (10.5) (108.1) (5.6) (66.1) June 0.3 4.7 8.3 10.3 10.7 42.0 (0.6) (2.1) (4.0) (5.5) (7.6) (21.2) July 1.7 4.0 9.7 64.3 29.7 41.3 (1.2) (1.7) (2.1) (18.9) (17.6) (15.0) August 6.0 3.3 7.0 39.3 35.3 27.0 (2.0) (3.5) (7.0) (11.0) (12.7) (14.7) September 5.7 8.3 8.3 20.3 20.0 19.3 (4.0) (4.2) (5.5) (6.2) (7.5) (14.0)

The effect of vegetation overshadowed any tube effects in the vegetated cores and thus might mask significant tube effects in bare substrate cores. Therefore, one-way ANOVA tests were done on the inner and outer cores from unvegetated areas [log10 (x + 1) transformation] to compare total infaunal abundance in cores containing 0, 1, or 4 D. cuprea tubes for each month separately. While these ANOVAs are not independent of the three-way ANOVA and should be interpreted cautiously, they suggest that the relationship of infaunal abundance with tubes in unvegetated areas is highly variable (Table 3). In the inner cores taken in May 1983, infaunal densities of the treatments containing 0 and 4 tubes are not sig- nificantly different while both are significantly different from the 1 tube samples (Table 4). However, in the May outer cores, the samples from the vicinity of 0 tube treatments contain the highest number of infaunal organisms (Table 4). Further ambiguity is added by the data from the outer cores taken in June and July. In June, the densities of infauna in the outer areas surrounding the 1 and 4 tube treatments were not significantly different, but in July, the most dense infaunal assemblage was found in the outer area of the 4 tube treatments (Table 4). To determine whether D. cuprea tube densities affect the community compo- sition of the surrounding infauna, species in the inner and outer samples were ranked by abundance for each month. Replicates of each tube density were pooled. Each species received a score that corresponded to its rank; the most abundant received 10 points, the next 9, etc. (Virnstein, 1977; Woodin, 1978.) Cumulative scores were used to determine the species rank for the entire sampling period (May-September 1983). Rank analyses from cores taken in vegetated and un- vegetated areas were determined and compared separately. Clymenella mucosa BAN AND NELSON: ROLE OF DIOPATRA CUPREA TUBES IN AN INFAUNAL COMMUNITY 15

Table 2. Summary of F statistics from the three-way ANOV A's for the field data. Analyses for individuals were loglO(x+ 1) transformed to correct heterogeneity of variances. *** = P < 0.005, ** = P < 0.01, n.s. = P > 0.05. All interaction terms were non-significant (P > 0.05) and are not shown here

Parameter Substrate effect Diopatra effect Time effect Inner cores Number of species 18.9365*** 0.0464 n.s. 3.935** Number of individuals 13.0652*** 0.0206 n.s. 2.224 n.s. Outer cores Number of species 23.6769*** 0.6235 n.s. 10.0891 *** Number of individuals 34.2789*** 0.5123 n.S. 6.6392*** was ranked 1 or 2 in all vegetated cores (Table 5). All of the dominant taxa from the vegetated treatments were or nemerteans. Inner and outer cores were extremely similar with four of the top five dominant taxa being the same for each of the three tube densities. Arenicola cristata was ranked either 1 or 2 in all unvegetated cores. Again the most dominant species were polychaetes or ne- merteans with the exception of one bivalve, Macoma constricta. Inner and outer cores were only slightly less uniform in their dominant species composition for the vegetated samples (Table 5). Thus there did not appear to be any major differences in community composition between inner and outer cores in either substrate type. Laboratory Experiments. -Experiment 1 tested the hypothesis that high tube densities are more effective than low tube densities in deterring crab foraging. A one-way ANOVA (Ban, 1985) showed there to be no significant difference in the number of clams remaining among the three treatments (0, 4, or 10 tubes per 2 0.01 m ). However, higher M. latera/is survivorship in the 10 tube treatments was suggested (Ban, 1985) and the experiment was repeated. The combined ex- periments were analyzed by a two-way ANOVA (Table 6) that included an ex- amination of the linear and non-linear components of the main effect of the treatments, and these components of the interaction between treatments and trials (Ray, 1960). This test revealed there to be no significant effectsin the test due to the combined trials; however, there was a significant difference (P < 0.005) among treatments. Significantly more clams remained in the treatments containing 10 tubes compared to the 0 and 4 tube treatments. In Experiment 2, testing the hypothesis that the combination of vegetation and tubes has the greatest effect on crab foraging, a one-way ANOVA and an a pos- teriori Welsch Step-Up Test indicated a significant difference (P < 0.005) in

Table 3. Summary of F statistics from the one-way ANOV A's comparing the abundance of infauna from inner and outer unvegetated cores associated with 0, 1, or 4 D. cup rea tubes per 0.01 m2• All data were loglO(x + 1) transformed. *** = P < 0.005, ** = P < 0.01, * = P < 0.05, n.s. = P > 0.05

Month Inner core Outer core May 7.1217* 7.5684* June 1.0274 n.s. 17.2067*** July 4.2472 n.s. 12.9838** August 1.1342 n.s. 0.4097 n.s. September 0.4577 n.s. 0.4034 n.s. 16 BULLETIN OF MARINE SCIENCE, VOL. 40, NO.1, 1987

Table 4. Results of Welsch Step-Up Tests comparing mean infaunal densities of either the inner or the outer unvegetated cores. Underlined values not significantly different at a = 0.05. Values are back- transformed means

May inner cores 1 Tube o Tubes 4 Tubes 8.5 18.1 31.4 May outer cores 1 Tube 4 Tubes o Tubes 21.9 22.1 47.5 June outer cores o Tubes 1 Tube 4 Tubes 0.3 4.4 7.7 July outer cores o Tubes 1 Tube 4 Tubes 0.4 3.7 9.5

number of M. latera/is remaining among the 4 treatments (Table 7). A comparison of Experiments 1 and 2 (Tables 6 and 7) indicates that the survivorship of M. latera/is is lowest in tanks with either a or 4 tubes (equivalent to control and tubes treatments of Experiment 2). A significantly greater number of clams survive in treatments with 10 tubes, and survivorship of M. latera/is is highest in treat- ments containing either vegetation alone, or tubes and vegetation. Experiment 3 employed a randomized blocks design to allow the observation and documentation of crab behavior in the presence of D. cuprea tubes. A two- way ANOVA (Table 8) showed that no significant block effectswere present during the test (P < 0.05); however, there was a significant difference (P < 0.005) among the four treatments in the amount of time Callinectes spent foraging. The results of an a posteriori Welsch Step-Up Test showed that foraging time was equal in the treatments containing tubes and clams and clams alone, and that these treat- ments were significantly greater than treatments containing tubes alone and the control. Thus, foraging time was enhanced when M. latera/is were present, but the presence of D. cuprea tubes did not change total foraging time. In the treat- ments where a food stimulus was not provided, the crabs spent most of the test time resting or roving at the edges of the aquaria. As an additional analysis of Experiment 3, time spent actually feeding in the 2 treatments where clams were provided was compared by a t-test. A significant difference (P < 0.005) in time spent feeding was found between the treatment with tubes and clams (x = 8.3 min) versus the treatment with clams alone (x = 11.7). Thus juvenile Callinectes spent the same amount of time foraging in these two treatments, but less time actually feeding when tubes were present.

DISCUSSION In two separate studies, Woodin (1978; 1981) found the abundance and species richness of other species of the infauna to be positively associated with the presence of tube structures, regardless of whether the structures were actual Diopatra cuprea tubes or drinking straws used to simulate tubes. In the absence of all predators, presumably excluded by 0.625-cm mesh cages, infaunal abundance no longer increased as a function of tube abundance. Thus the tubes of D. cuprea appeared to act as a refuge from predation for the infauna of the Virginia mud flat that BAN AND NELSON: ROLE OF DWPATRA CUPREA TUBES IN AN INFAUNAL COMMUNITY 17

N '-0 N -

00 N o 00 r-- '

u o~ '-' u~ 15 o

-'

-r-- -M -o 18 BULLETIN OF MARINE SCIENCE, VOL. 40, NO.1, 1987

Table 6. Results of two-way ANOVA and Welsch Step-Up Test for Experiment I comparing the number of clams remaining in two trials of treatments containing 10, 4, or 0 clams per 0.01 m2• *** = p < 0.001, n.s. = P > 0.05

Source df Ms F Among treatments (A) 2 Liner component (LJ 1 161.33 33.38*** Quadratic component (QJ 1 18.78 3.89 n.s. Betweeen trials (B) 1 4.50 1.03 n.s. Interaction (A x B) 2 La X B 1 3.00 0.62 n.s. Qa X B I 0.00 0.00 n.s. Error 12 4.83 Total 17 245.61

Means: 0 tubes (1.3) = 4 tubes (2.8) < 10 tubes (8.7)

Woodin investigated, and results were consistent over a 3-year period. Luckenbach (1984) also supports the hypothesis of a refuge effect for D. cuprea tubes with studies in the North Inlet marsh, South Carolina. In contrast, Bell and Woodin (1984) conducted a study at the same field site as that used for Woodin's (1978; 1981) earlier work, and found that selected macrofauna increased in response to predator exclusion, but infaunal spatial pat- terns and relationships with tubes were highly variable. In this case the analyses of spatial patterns of macrofauna did not fit the a priori predictions necessary to support the hypothesis that D. cuprea tubes can act as a refuge for all macrofauna. Bell and Woodin (1984) suggest that the discrepancy between the findings of the latter study and earlier reports of a refuge effect for D. cuprea rich areas may be related to 1) changes in community composition, 2) predator activity, or 3) the time scale of the latter experiments (2-4 weeks versus 2-6 months). The present study does not support the general hypothesis that aggregations of D. cuprea tubes constitute a biogenic refuge for the other infauna. Specifically,an analysis of cores taken around tubes in unvegetated areas showed there to be no clear pattern of association between infaunal abundance or diversity and D. cuprea tubes. The differences in infaunal densities found among treatments in the May inner cores did not follow the predictions of a refuge hypothesis. In addition, the highest infaunal densities were found in the 4 tube treatments of both the June and July outer cores where a refuge effect of D. cuprea tubes was not expected. These results might be due to the seasonal increase of predators that occurs in May and begins to taper off in June and July. It is possible that these predators are attracted to the areas of tube aggregates and thus feed comparatively less in

Table 7. Results of one-way ANOV A and Welsch Step-Up Test for Experiment 2 comparing the number of clams remaining in treatments containing an H. wrightii rhizome mat and D. cuprea tubes, tubes alone, a rhizome mat alone, and neither a rhizome mat nor tubes. *** = P < 0.001

Source df Ms F Among treatments 3 160.97 39.69*** Experimental error 8 4.06 Total 12 165.03 Means: control (2.0) = tubes alone (4.0) < rhizome mat (14.3) = rhizome mat and tubes (16.7) BAN AND NELSON: ROLE OF DlOPATRA CUPREA TUBES IN AN INFAUNAL COMMUNITY 19

Table 8. Results of two-way ANOYA (without replication) for Experiment 3 comparing Ca/linectes spp. foraging time in trltatments containing D. cuprea tubes and M. fateralis. tubes alone, clams alone, and neither tubes nor clams. *** = p < 0.001, n.s. = P > 0.05

Source df Ms F Among treatments 3 309.86 52.87*** Among blocks 2 4.75 0.81 n.s. Interaction 6 5.86 Total 11 320.47

Means: control (5.7) = tubes (8.3) < tubes and clams (24.3) = clams alone (24.7) the outer areas. However, the discrepancies in infaunal abundance might also simply be due to the general patchiness of the infaunal community of the bare sand areas. In areas vegetated with H. wrightii. the infaunal community was not enhanced in abundance or diversity in the vicinity of aggregates of D. cuprea tubes. As expected, infaunal density and diversity in the vegetated cores was much greater than in those from unvegetated areas. The presence of the rhizome mats appears to outweigh any refuge effect generated by the tubes of D. cuprea. The similar community composition of cores taken from vegetated and unvegetated areas is consistent with the findings of Virnstein et al. (1983). The lack of an observed refuge effect in the present study may be partly due to a possibly greater abundance of small decapod predators in the Indian River, as compared to more temperate soft sediment communities (Heck and Orth, 1980 vs. Gore et al., 1981). Also, the occurrence of many infaunal carnivores such as nemerteans may serve to erase the refuge effect of D. cuprea tubes in the Indian River system. The abundant H. wrightii beds of the Indian River site have the ability to harbor a great number of predators (Virnstein et al., 1983) that can venture into unvegetated areas to feed, an element lacking in other studies (Wood- in, 1978; 1981; Bell and Woodin, 1984; Luckenbach, 1984). Secondly, previous studies (Woodin 1978; 1981; Bell and Woodin, 1984; Luckenbach, 1984) were conducted on intertidal mud flats as compared to the shallow subtidal area studied in the Indian River lagoon. Therefore, feeding by crabs and fishes at the Indian River site would not be interrupted by the periods of total sediment exposure that occur in an intertidal system, and predation pressure might be greater in the Indian River system. Differences in the mean size of crabs (e.g.,juveniles versus adults) might alter the relative impact of predation in the vicinity of tubes. Un- fortunately, this information is not available for comparing the various areas where tube effects have been studied. Discrepancies between the results of the present study and those of Woodin (1978; 1981) and Luckenbach (1984) may be attributed in part to the low tube densities found at the Indian River site. Luckenbach sampled in areas of 10 tubes per 0.01 m2• Woodin (1978) found increased infaunal density and diversity in 2 areas with a density of 6 D. cuprea tubes per 0.01 m ; the refuge effect was not 2 observed in areas with 1 tube per 0.01 m • At the Indian River site, a maximum density of only 4 D. cuprea tubes per 0.01 m2 was consistently found. The laboratory studies indicate an agreement between observed Callinectes activity in the presence of tubes and the infaunal abundance data from the field- work. Crabs could easily forage among 4 tubes per 0.01 m2 in the laboratory, and cores taken in the vicinity of 4 tubes per 0.0 1 m2 in the field did not exhibit enhanced infaunal density and diversity. In addition, tubes within vegetation mats 20 BULLETIN OF MARINE SCIENCE, VOL. 40, NO. I, 1987 in the laboratory experiments did not add to M. lateralis survivorship, and the abundance and species richness in samples taken from the Halodule wrightii beds was not significantly different among tube treatments. Blue crabs are known to feed mostly on mollusks, though stomachs of smaller crabs are also known to contain polychaetes, amphipods, sand, mud, and detritus (Paul, 1981). At a car- apace width of less than 80 mm they tend to ingest and triturate large amounts of the substrate (Tagatz, 1968),and thus foragingaction alone can be an important source of disturbance to the infauna. The laboratory observation that tubes lessen actual feeding time but do not affect foraging time of Callinectes may not be important when considering the overall impact of juvenile crabs on the benthos. It is not presently known whether aggregates of 10 tubes per 0.0 I m2 if trans- ferred to the field study site would convey the refuge effect as suggested by the laboratory studies. Presently, the maximum density of four D. cuprea tubes per 0.0 I m2 appears to be below a critical lower limit of tube density required to establish an effective refuge.

ACKNOWLEDGMENTS

The authors would like to thank Drs. K. Clark, G. Swain and W. H. Wilson for their critical analyses of this work. The following persons provided much needed assistance in the field: K. Baltz, J. Bomber, K. Durkin, K. Lorking, M. Main, W. J. Miller, and P. Pendoley. Contribution No. 69 from the Department of Oceanography and Ocean Engineering, Florida Institute of Technology.

LITERATURE CITED

Arntz, W. E. 1977. Results and problems of an "unsuccessful" benthos cage predation experiment (western Baltic). Pages 31-44 in B. F. Keegan, P. O. Ceidigh and P. J. S. Boaden, eds. Biology of marine benthos. Pergamon Press, New York. Ban, S. M. 1985. The role of Diopatra cuprea Bosc 1802 (Polychaeta: Onuphidae) tubes in the structure of a subtropical infaunal community. M.S. Thesis, Florida Institute of Technology, Melbourne, Florida. 79 pp. Bell, S. S. and S. A. Woodin. 1984. Community unity in soft-bottom benthos: experimental evidence from a disturbance study in Virginia. J. Mar. Res. 42: 605-632. Brenchley, G. A. 1982. Mechanisms of spatial competition in marine soft-bottom communities. J. Exp. Mar. BioI. Ecol. 60: 17-23. Dauer, D. M., G. H. Tourtellotte and R. M. Ewing. 1982. Oyster shells and artificial worm tubes; the role of refuges in structuring benthic communities of the lower Chesapeake Bay. Int. Rev. Gesamten. Hydrobiol. 67: 661-677. Gore, R. H., E. E. Gallaher, L. E. Scotto and K. A. Wilson. 1981. Studies on decapod crustacea from thc Indian River region of Florida. XI. Community composition, structure, biomass and species-areal relationships of seagrass and drift-algae associated macrocrustaceans. Estuar. Coastal Mar. Sci. 12: 485-508. Goss-Custard, J. D, 1977. Predator responses and prey mortality in redshank, Tringa totanus and a preferred prey Corophium vo/utator, J. Anim. Ecol. 46: 21-35. Heck, K. L., Jr. and R. J. Orth. 1980. Structural components of eelgrass (Zostera marina) meadows in the lower Chesapeake Bay-decapod crustacea. Estuaries 3: 289-295, Holland, A. F., N. K. Mountford, M. H. Hiegel, K. R. Klaumeyer and J. A. Mihursky. 1980. Influence of predation on infaunal abundance in upper Chesapeake Bay, U.S.A. Mar. BioI. 57: 221-235. Luckenbach, M. W. 1984. Biogenic structure and foraging by five species of shorebirds (Charadrii). Estuar. Coast. Shelf Sci. 19: 691-696. Miller, T. E. 1982. Community diversity and interactions between the size and frequency of dis- turbance. Am. Nat. 120: 533-535. Nelson, W. G. 1981. Experimental studies of decapod and fish predation on seagrass macrobenthos. Mar. Ecol. Prog, Ser. 5: 141-149. Orth, R. J. 1977. The importance of sediment stability in seagrass communities. Pages 281-300 in B. C. Coull, ed. Ecology of marine benthos, University of South Carolina Press, Columbia, Paul, R. K. G. 1981. Natural diet, feeding and predatory activity of the crabs Callinectes arcuatus and C. toxotes (Decapoda, Brachyura, Portunidae). Mar. Ecol. Prog. Ser, 6: 91-99, Peterson, C. H. 1982. Clam predation by welks (Busycon spp.). Experimental tests of the importance of prey size, prey density, and seagrass cover. Mar. BioI. 66: 159-170. BANANDNELSON:ROLEOFDIOPATRA CUPREA TUBESINANINFAUNALCOMMUNITY 21

Ray, W. S. 1960. An introduction to experimental design. The MacMillan Company, New York. 254 pp. Reise, K. 1977. Predation exclusion experiments in an intertidal mud flat. Helgol. Wiss. Meeresunters. 30: 263-271. Sokal, R. R. and F. J. Rohlf. 1981. Biometry. W. H. Freeman and Company, San Francisco, California. 859 pp. Tagatz, M. E. 1968. Biology of the blue crab Callinectes sapidus in the St. Johns River, Florida. Fish. Bull. U.S. 67: 17-33. Thistle, D. 1980. The response of a harpacticoid community to a small scale natural disturbance. J. Mar. Res. 38: 381-395. ---. 1981. Natural physical disturbances and communities of marine soft-bottoms. Mar. Ecol. Prog. Ser. 6: 223-228. Virnstein, R. W. 1977. The importance of predation by crabs and fishes on benthic infauna in Chesapeake Bay. Ecology 58: 1199-1217. ---. 1978. Predator caging experiments in soft sediments: caution advised. Pages 261-272 in M. L. Wiley, ed. Estuarine interactions. Academic Press, New York. ---, P. S. Mikkelsen, K. D. Cairns and M. A. Capone. 1983. Seagrass bed versus sand bottoms; the trophic importance of their associated benthic invertebrates. Fla. Sci. 46: 363-381. Wilson, W. H., Jr. 1979. Community structure and species diversity of the sedimentary reefs con- structed by Petaloproctus socialis (Polychaeta: Maldanidae). J. Mar. Res. 37: 623-641. Woodin, S. A. 1978. Refuges, disturbance and community structure: a marine soft-bottom example. Ecology 59: 274-284. ---. 1981. Disturbance and community structure in a shallow water sand flat. Ecology 62: 1052- 1066. Young, D. K., M. A. Buzas and M. W. Young. 1976. Species densities of macrobenthos associated with seagrass: a field experimental study of predation. J. Mar. Res. 34: 577-592.

DATEACCEPTED:July 17, 1985.

ADDRESSES:Department of Oceanography and Ocean Engineering, Florida Institute of Technology, Melbourne. Florida 32901. PRESENTADDRESS:(S.M.B.) Woodward-Clyde Consultants. 701 Sesame St., Anchorage, Alaska 99503.