Lontroi Ocennog,-, 43(3), 1998, 5 16-521 0 1998, by the Amcr~. in Sac~ety of L~mnology and Oceanography, Inc A new device for studying benthic invertebrate recruitment Robert B. Whitlatch Department of Marine Sciences, University of Connecticut, 1084 Shennecossett Rd., Groton, Connecticut 06340-6097 Richard W. Osman Estuarine Research Center, Academy of Natural Sciences, 10545 Mackall Rd., St. Leonard, Maryland 20685 Abstract An automated larval settlement sampler has been designed and tested in the field to examine the effects of environmental conditions on the recruitment of epifaunal and infaunal benthic invertebrates. The sampler (1) exposes substrates to settling larvae at four discrete levels of naturally varying environmental conditions (e.g. different tidal states, current speeds, light regimes), (2) repeatedly exposes substrates over a period (2-8 weeks) sufficient to accumulate measurable numbers of recruits, and (3) maintains unexposed substrates in an environment that does not result in unnatural levels of mortality. Preliminary field trials indicate that the device was effective in assessing recruitment responses of epifaunal invertebrates to diurnal variations in light level and infaunal invertebrates to variations in tidal state. The life-cycles of many marine benthic invertebrates in- habiting coastal environments include a planktonic larval phase lasting for minutes to months followed by metamor- phosis and settlement to the seafloor. Cues affecting larval settlement have been extensively studied in many types of epifaunal and infaunal invertebrate species (e.g. reviews of Crisp 1974; Woodin 1986; Butman 1987; Pawlik 1992), and it is now recognized that a variety of physical, chemical, and biological factors have the potential to influence the settle- ment of invertebrate larvae. These factors include (1) the near-bottom hydrodynamic regime (Eckman 1983; Pawlik et al. 1991; Mullineaux and Butman 1991; Jonsson et al. 1991; Snelgrove et al. 1993), (2) substrate texture and type (Crisp and Barnes 1954; Wethey 1986; Snelgrove et al. 1996), (3) physical factors (Korrigna 1949; Thorson 1964; Crisp 1974; Pechenik 1990), (4) chemical attractants and toxins (Dubilier 1988; Davis et al. 1991), and (5) biotic interactions (Gross- berg 1991; Woodin 1985; 0lfasson et al. 1994). Early work on larval settlement generally focused on physical cues (e.g. light, temperature, salinity, substrate type) in a laboratory setting (e.g. reviews of Butman 1987; Young 1990). While these studies advanced our understanding of the proximate factors affecting settlement, they were often conducted under somewhat artificial conditions that neglect- ed the potential contribution of other co-occurring factors on Acknowledgments We thank G. Grenier for invaluable assistance in helping to de- sign and construct ELVIS. B. Dziomba and D. Arbige aided with construction and electronics development, respectively. M. Berger, V. J. Cummings, A. Frese, E. Miles, and R. Stankelis assisted with field and laboratory work. D. Cohen and K. Howard-Strobe1 sup- plied current meter data. C. A. Butman, L. A. Levin, L. S. Mullin- eaux, I? V. R. Snelgrove, and two anonymous reviewers provided thoughtful comments of an earlier version of the manuscript. To all: thank you, thank you very much. This work was supported by NSF grants OCE 91-01 815 and OCE 91-23890 and the University of Connecticut Marine Sciences and Technology Center (MSTC). settlement behavior. For example, we now recognize the in- fluence of hydrodynamic processes on benthic invertebrate settlement behavior and that larval substrate selection can vary with the hydrodynamic regime (Pawlik and Butman 1993; Snelgrove et al. 1993). In addition, field studies of recruitmentfrequently demonstrate high spatial and temporal variability (Hawkins and Hartnoll 1982; Gaines and Rough- garden 1987). This variation often precludes the ability to rank or develop a hierarchy of factors contributing to settle- ment patterns (but see Wethey 1986). Lastly, larval settle- ment cues often co-vary (e.g. gas solubility with tempera- ture; organic content and bacterial abundance with sediment grain size; water velocity with tidal state), further adding complexity to our interpretation and partitioning of the rel- ative importance of each factor in affecting recruitment dy- namics of benthic invertebrates. We have developed a new device for field experimental studies in order to examine recruitment patterns for a broad array of infaunal and epifaunal invertebrates with the goals of establishing general trends and exceptions and determin- ing the influence of specific environmental patterns on nat- ural in situ recruitment patterns. This instrument pennits us to control exposure of larval settlement substrates to param- eters that naturally vary, expose the substrates under specific conditions for periods up to 8 weeks so that measurable re- cruitment differences can be determined for an array of spe- cies, and maintain unexposed substrates in an environment that does not affect organisms already recruited to the ex- perimental substrates. Here we describe the design and operational tests of the device and present results of several preliminary experi- ments on known recruitment patterns of shallow-water ben- thic invertebrates using either hard substrates (for sessile epi- faunal species) or sediment-filled containers (for infaunal species). Specifically, we examined diurnal patterns of epi- faunal recruitment since it is well documented that larval release in many species is often affected by variations in light intensity (Ryland 1977; Svane and Young 1989). Our Benthic recruitment device intent was to verify whether epifaunal recruitment also ex- hibited diurnal periodicity and to use this experiment as an initial test of the efficacy of the recruitment apparatus. Re- cruitment patterns of infaunal invertebrates relative to tidal state were also examined. While generally less is known about settlement cues of infaunal species (Woodin 1986), tidal action can influence larval availability for many species (Cronin and Forward 1982; Stancyk and Feller 1986). Sev- eral studies have documented tidally induced variations in the larval abundance of infaunal species (Srikrishnadhas and Ramarnoorthi 1982; Stancyk and Feller 1986; Levin 1986). There are, however, no published field studies that have ex- plicitly examined the relationship of tidal state to patterns of infaunal recruitment. Materials and methods Recruitment device design-The design and operation of the electronic larval in situ settlement (ELVIS) device was driven by several basic requirements and constraints. First, we needed to expose substrates to benthic invertebrate re- cruitment in a variety of shallow-water habitats under known environmental conditions. Second, to collect sufficient num- bers of organisms on the recruitment substrates, the device had to repeatedly expose a given substrate or set of sub- strates under similar conditions for periods up to 60 d. Third, repeated exposure of substrates under similar conditions re- quired sequestering the substrates in a manner that elimi- nated larval settlement (e.g. larval "contamination") without affecting the mortality rate of recruits already present on (or in) the substrates. Fourth, exposed substrates needed to be deployed in a manner that minimized hydrodynamic artifacts that could potentially bias or influence larval transport and(or) recruitment to the substrates. It is well recognized that recruitment responses of benthic invertebrates to alter- ations in boundary-layer flow conditions can occur in a wide variety of benthic species (Crisp 1955; Eckman 1983; Keen 1987; Mullineaux and Butman 1990; Walters 1992). Lastly, the device needed to be designed for relative ease of de- ployment and recovery of experimental substrates under a variety of field conditions. ELVIS consists of three main components: a rotating cir- cular faceplate, four removable settlement substrate units, and a pressure housing that contains components to control the movement of the faceplate (Fig. 1). The rotating face- plate has four holes that allow one of the four substrates within each of the four substrate units to be exposed to set- tling larvae. Each settlement substrate unit protects the three unexposed substrates from settling larvae while housing them in an environment as similar as possible to natural field conditions. Finally, electronic components in the pressure housing control the sequence of rotation of the faceplate such that each set of four substrates (one from ,each of four substrate units) is exposed during a specific set of environ- mental conditions. The faceplate and cam design, however, can be modified to expose two sets of substrates at eight different times or one substrate at 16 separate intervals de- pending upon specific needs. Fig. 1. The new device for studying benthic invertebrate re- cruitment. I, Cutaway view showing the (A) faceplate and position of four substrate exposure positions, (B) aluminum ring with cam attached to the underside of the faceplate, and (C) pressure housing. Diameters: faceplate, 91.5 cm; circular substrates, 7.8 cm; pressure housing, 32 cm. 11, Internal components of the pressure housing showing the (D) microswitch used to align the faceplate over the substrate, (E) gear-drive and faceplate attachment plate, (F) low- speed motor, and (G) microprocessor. 111, Outside view of substrate settlement unit. Diameters: PVC pipes, 9.0 cm; holes, 1.1 cm. See text for further description and operation. Faceplate-The faceplate, made of 0.7-em-thick PVC plate, is 91.5 cm in diameter and has four 8.3-em-diameter holes that are positioned at equidistant (27 em) from adjacent holes and 16.5 em from the edge of the plate. An aluminum ring, attached to the underside of the faceplate, has four cams that are aligned with the faceplate holes (Fig. 1). The face- plate assembly is secured to a circular aluminum frame that can be attached to a stand or suspended in the water column. The faceplate assembly is driven by a I/,,-hp, 12-V DC low- speed (6 rpm) gear-motor mounted inside a 32-em-diameter, 28-em-long PVC pipe secured to the aluminum frame. The outer end is sealed with a PVC plate bolted to the pressure housing. The gear-motor is attached to a step-down gear- drive that turns the faceplate at -2 rpm. In this configura- tion, 2 s are required to index the faceplate from one sub- strate position to the next. Pressure housing-The pressure housing contains two switches-one for positioning the faceplate holes so they are 518 Whitlatch and Osman precisely aligned over each exposed substrate, the other be- ing used to index the faceplate without interrupting the pro- grammed timing of a low-power microprocessor. The micro- processor (8-bit, 28K RAM Tattle-tale Mark V data logger) is programmed, via an RS-232 serial connection, and all con- trol and data-logging functions are established in software (PBASIC). This configuration permits a high degree of flex- ibility in the faceplate indexing schedule. When the faceplate is indexed, the microprocessor logs the time along with any possible error flags that can be downloaded at the end of each deployment. The gear motor and microprocessor are powered by a 12 V, 10 A hs' sealed rechargeable battery. Under these power constraints, the faceplate can be indexed up to 16 times per day for at least 60 d. Settlement units-Each substrate settlement unit consists of four (9.0-cm-diam., 13.5 cm long) PVC pipes that are attached at one end to a curved section of 0.7-cm PVC sheet (Fig. 1). The other end of each pipe, with a series of 1.1- cm-diameter holes drilled around it, has a 1.0-cm-long PVC ring that moves up and down on four stainless steel pins. The ring is attached to the pipe by gluing an 80-pm Nytex mesh netting around the upper portion of the pipe and ring. Each pipe contains a spring-loaded inner PVC pipe assembly (7.5 cm in diam., 10 cm long) that is kept in position by perpendicularly mounted roller-pins. The settlement units are secured in position on the frame with stainless steel clips. When in position, the four cams allow one substrate from each of the four units to be flush with the outer surface of the faceplate (Fig. 1). The other substrates are depressed -2 cm below the tops of the pipes, creating a head space of -93 cm3 over each unexposed substrate. The upper pipe surfaces containing the unexposed substrates are flush with the underside of the faceplate. Therefore, the only water ex- change available to the unexposed substrates is through the plankton netting surrounding the upper portion of each pipe assembly. The settlement units can be fitted with PVC panels or cups for use in studies of epifaunal or infaunal recruitment, re- spectively. Circular panels (7.8 cm in diam.), made of rough- ened PVC sheet with glued threaded stems, are screwed into the inner sleeves of the settlement unit pipes (Fig. 1). Cups (5.5 cm in diam., 2.5 cm deep), machined from solid PVC pipe, are designed to fit snugly into the inner sleeves of the substrate holders. PVC covers, which are clipped on the cups, ensure that any material contained in cups will not be lost or disturbed when the cups are placed on or removed from ELVIS. Unlike the panels, the cups are loaded from the top of ELVIS after the settlement units are secured into position. The switch that overrides the microprocessor (see above) moves the faceplate and minimizes disturbance of material in the cups during their deployment and recovery. Operational constraints and tests-When in position, the four exposed substrates are flush with the surface of the face- plate. The purpose of this design was to minimize potential flow-related artifacts in substrate exposure and position that may influence larval transport to, and their retention on, the substrates. At the currrent stage of development, the sub- strate exposure method is designed to minimize potential flow artifacts associated with the leading edge of the face- plate. In laminar flow, vertical advection, shear stress, and turbulence will vary with distance downstream from the leading edge of the faceplate (e.g. Schlichting 1979); and the settlement substrates are located 16.5 cm from the leading edge (as~ject ratio >20: 1) in order to maintain similar flow conditions over each of the exposed substrates. Mullineaux and Butman (1991) demonstrated that for 1.0-cm-thick poly- carbonate substrates, advection, shear stress, and turbulence all stabilized at distances 4-10 cm from the leading edge in free-stream velocities of 5 and 10 cm ssl in a laboratory flume. At higher flow speeds, advection, shear stress and turbulence will stabilize more quickly in the downstream direction. Importantly, note that in naturally turbulent con- ditions, flow-related artifacts in larval settlement are likely and should be examined on a case-by-case basis. Another potential source of flow disturbance on the device is the -0.5-cm ring of space between each substrate and the faceplate. Field deployments of ELVIS in relatively low-wa- ter-flow conditions (e.g. 0-10 cm ssl water velocities) have found nonrandom epifaunal recruitment patterns on the sur- faces of the substrates; preliminary evidence suggests that the lack of small-scale variations in the boundary-layer flow on the substrates could influence larval transport and reten- tion (e.g. Mullineaux and Butman 1990; Mullineaux and Garland 1993). Deployments of the device in a wider range of conditions are required in order to more fully address this issue. The faceplate design and diameter can be altered to accommodate different types of flow regimes or habitat con- ditions. Three different sets of experiments were conducted to test operating conditions and sampling assumptions of the re- cruitment device. In all of these trials except one (see below), ELVIS was deployed faceplate-down from a floating raft moored in 2-3 m of water behind a jetty at Avery Point, Conneticut, eastern Long Island Sound (see Osman and Whi- tlatch 1995a,b for site description). The faceplate was po- sitioned -1 m below the water surface to collect recruiting sessile epifaunal species. Colonial (Botryllus schlosseri, Bo- trylloides diagensis, and Diplosoma macdonaldi) and soli- tary (Molgula manhattensis, Ciona intestinalis, Styela clava, and Ascidiella aspersa) ascidians are the most dominant (>80% cover) epifaunal species encrusting rocks, jetties, and pilings at the site. Other epifaunal species include spirorbid (Spirorbis spp.) and serpulid (Hydroides dianthus) poly- chaetes, barnacles (Balanus eburneus, B. amphitrite, Semi- balanus balanoides), bryozoans (Cryptosula pallasiana, Schizoporella errata, Bowerbankia gracilis, Bugula turrita, and Electra sp.), and hydroids (Obelia sp.). All these species have been regularly collected on PVC substrates exposed from 1 d to several months (Osman and Whitlatch 1995b,c). Electromagnetic current meter (Interocean S4) deployments indicate that water flow at the site is characterized by a tid- ally driven current (mean peak flows of -6 cm s I ) running predominantly in northeast-southwest and east-west direc- tions. Episodic velocities of 20-30 cm s - ' have been re- corded, and their occurrence correlates well with local wind events (unpubl. data, Dept. Mar. Sci., Univ. Conn., collected -300 m from the study site). The first series of ELVIS operation tests determined Benthic recruitment device 519 whether substrates positioned in an unexposed mode re- mained isolated from settling larvae. This information was needed to test whether substrates were being contaminated by settling larvae when kept in an unexposed mode. ELVIS was deployed for 7 d on two occasions, 17-24 July and 1- 7 August. ELVIS was programmed to expose one set of panels from 0600 to 2100 h (daytime treatment) and another set of panels from 2100 to 0600 h (nighttime treatment). The remaining two sets of panels were exposed for 30 s peri- ods-one at 2100 h (30-s 1 treatment) and one at 0600 h (30-s 2 treatment). At the end of each deployment, substrates were removed from the settlement units and all individuals or colonies of all epifaunal species were counted and iden- tified in the laboratory under a dissecting microscope. A second set of ELVIS deployments was conducted to assess whether there were differences in the mortality rates of epifaunal recruits attached to exposed and unexposed sub- strates. These data were necessary to determine whether any observed difference in the number of recruits on a given set of panels was due to differences in the number of settling larvae or to differences in postsettlement mortality of re- cruits attached to panels with different exposure schedules. ELVIS panels were exposed to competent larvae of the as- cidian Ascidella aspersa in the laboratory. After 24 h the panels were examined under a dissecting microscope and number of recruits were "gardened" to a known number of individuals (20-40 panel-') by haphazardly removing ex- cess individuals from panel surfaces. Two experiments were conducted-one in the laboratory in order to reduce potential confounding effects of recruitment of conspecifics and(or) other species, the other at the field site when Ascidiella was not naturally recruiting. In the laboratory, ELVIS was posi- tioned, faceplate upward, in a large (0.91 X 1.36 X 4.38 m) tank supplied by continuously flowing, filtered (10-pm nom- inal size) seawater. Water-flow rates in the tank averaged -10-12 volume exchanges per day. ELVIS was pro- grammed to repeatedly expose individual sets of four panels at 16-h, 8-h, 1-min, and 30-s intervals over a 7-d period (3- 10 November). Two additional sets of four identical sub- strates were placed in the tank-one with known numbers of Ascidiella recruits to serve as a reference, the other con- sisting of blank panels to record the number and type of recruits that may have not been removed by the water filtra- tion system. The field deployment (17-24 November) had a similar panel exposure schedule as the laboratory experi- ment. Two sets of four additional panels were also deployed at the site in a manner similar to the position of ELVIS- one set with known numbers of Ascidiella to assess mortality rates of recruits on substrates not attached to the device, another set with blank panels to record the number and type of epifaunal organisms recruiting during the deployment pe- riod. At the end of the deployments, substrates were exam- ined as previously described. The last set of trials assessed recruitment patterns of epi- faunal recruits relative to the distance from the leading edge of the faceplate. Evidence of nonrandom recruitment across the ELVIS faceplate may indicate flow-related alterations in larval settlement behavior (Mullineaux and Butman 1990; Mullineaux and Garland 1993). ELVIS was deployed in the field for 7 d, and the numbers of recruits of the colonial ascidian Botrylloides diagensis were counted along four equidistant transects in five consecutive 47.8-cm2 sampling areas running from the edge to the center of the faceplate. Although other epifaunal species were observed settling on the faceplate, only Botrylloides recruits could be recorded by eye because of their relatively large postset size (-1 mm in diam.) and distinctive bright orange color. Epifaunal recruitment in relation to photoperiod-Two ELVIS deployments examined patterns of epifaunal recruit- ment in relation to photoperiod by repeatedly exposing sets of substrates at intervals of 0600-0900, 0900-1200, 1200- 1500, and 1500-0600 h over a 7-d period. The two experi- ments were conducted from 20 to 27 September and 4 to 11 October. Sunriselsunset at the start of each experiment were 060511941 h and 0701118 10 h, respectively. Current speed and direction, temperature, conductivity, and pressure were monitored at 1 m above the seafloor and 5 m west of the raft with an electromagnetic current meter (Interocean S4) for the duration of both experiments. All variables were av- eraged over 1-min intervals recorded every 15 min. At the end of the deployments, panels were examined using a dis- secting microscope and all individuals were counted. Infaunal recruitment in relation to tidal state-Two 7-d experiments were conducted to examine patterns of infaunal recruitment in relation to tidal state. ELVIS was deployed faceplate upward in - 1 m (mlw) of water at a site located -200 m north of the floating raft. Sediments at the site are composed of a silty sand with a 0.6% organic content (ash- free dry wt). The infaunal community at the time of deploy- ment was dominated by spionid (Streblospio benedicti, Polydora cornuta) and syllid (Parapionosyllis longicirrata, Brania welveetensis, Streptosyllis arenae) polychaetes. Oth- er less abundant taxa included bivalves (Gemma gemma, Mya arenaria, and Mercenaria mercenaria) and an uniden- tified oligochaete. The site is more protected than the break- water site owing to the existence of a concrete dock posi- tioned to the south and -100 m from the site. Before ELVIS was deployed, a hole (-50 cm in diam., 1 m deep) was dug at the site and a plastic ring (26 cm in diam., 80 cm long) was positioned in the hole so that one end of the ring was nearly flush with the sediment surface. ELVIS was placed over the ring so that the faceplate rested parallel to and -10 cm above the sediment-water interface. Sediment CUDS were filled with a medium sand (0.3% or- ganic content) previously screened through a 1-mm mesh sieve and defaunated by freezing and thawing. Although it is recognized that placing the faceplate above the sediment surface mav limit colonization of the exverimental substrates by postsettlement juvenile and adult infauna that may be transported across the seabed (Smith and Brumsickle 1989; Snelgrove et al. 1995; Commito et al. 1995), the primary intent of these initial deployments was to reduce potential confounding interactions of organisms that are transported along the bed from those settling from the water column. Also, comparisons were limited to collections between cups exposed at different times within an ELVIS depolyment and not between cups and surrounding ambient sediments. The sediment exposure schedule for both experiments co- 520 Whitlatch and Osman Table 1. Average abundance ( 2 1 SD) of the total recruits and numbers of the most abundant taxa found on substrates exposed during the day (0600-2100 h), night (21004600 h), and two 30-s (30-s 1 and 30-s 2) periods during two 7-d deployment periods. Empty cells indicate that no organisms were found. Exposure period Day 30-s 1 Night 30-s 2 Taxon 17-24 Jul 1-7 Aug 17-24 Jul 1-7 Aug 17-24 Jul 1-7 Aug 17-24 Jul 1-7 Aug Total recruits Diplosoma macdonaldi Botryllus schlosseri Ascidiella aspersa Botrylloides diagensis Mogulu manhattensis Cryptosula pullasiana Spirorbis spp. Barnacles incided with the following tidal states: 1 h before and after high tide (high tide period); 1 h after high tide to 1 h before low tide (ebb tide period); 1 h before and after low tide (low tide period); and 1 h after low tide to 1 h before high tide (flood tide period). Times for the exposure schedule were derived from a continuously recording tide gauge located -100 m from the deployment site (unpubl. data, Conn. Dept. Environ. Protection). At the end of each deployment, the cups were capped and removed; cup contents were preserved in 10% buffered Formalin, screened with a 100-pm mesh sieve, and transferred to 70% ethanol containing Rose Ben- gal. Sample residues were sorted under a dissecting micro- scope and organisms were counted and identified to the low- est possible taxon. availability in the filtered water. These operation tests demon- strated that this epifaunal species can survive on substrates for relatively long periods sequestered in an unexposed mode, and that observed differences in the abundance of recruits on panels with different exposure schedules was not due to exposure- dependent mortality. In our examination of recruitment of Botrylloides diagen- sis relative to the distance from the edge of the ELVIS, we found significantly lower numbers of recruits in the outer- most 7.8 cm of the disc (Fig. 3). No other differences in recruit abundance were found. These data suggest that the positioning of the experimental substrates 16.5 cm from the faceplate edge minimized any potential variations in flow- related recruitment for this species of ascidian. Results ELVIS operation tests-The experiments assessing the ade- quacy of the design for isolating unexposed substrates from settling larvae indicated that sets of substrates repeatedly ex- posed for 30-s daily intervals over the 7-d deployment periods averaged 0.5-2.5 epifauna recruits (Table 1). The few numbers of recruits found on the panels either settled when the sub- strates were positioned in the unexposed mode or were resulted from larvae trapped in the headspace of water captured above the panels during the daily 30-s periods when the substrates were being indexed from an exposed to an unexposed position. The two trials indicate that the method devised for holding substrates in a unexposed mode was very effective in isolating them from larval contamination. Both deployments also re- vealed pronounced daylnight differences in the abundance of recruiting epifauna (Table 1). Both the laboratory and field deployments determining dif- ferences in mortality rates of recruits attached to exposed and unexposed substrates on ELVIS revealed no natural settlement of Ascidiella aspersa and no significant differences in the mor- tality of Ascidiella recruits on substrates exposed on schedules ranging from 30 s to 16 h d-'. Also, mortality rates of recruits on ELVIS panels did not differ from recruits attached to iden- tical substrates suspended in the water column adjacent to EL- VIS (Fig. 2). Ascidiella mortality was slightly higher in the laboratory than in the field, possibly the result of reduced food Relationship of photoperiod to epifaunal recruitment pat- terns-ELVIS field deployments examining the presence of 1 min 8 h Control Treatment Fig. 2. Effects of ELVIS substrate exposure (30 s, 1 min, and 8 and 16 h) on survival (avg. -C 1 SD) on recruits of Ascidiella aspersa in the laboratory (open histograms) and field (hatched his- tograms) during a 7-d deployment period. "Control" refers to pan- els containing Ascidiella recruits placed adjacent to ELVIS at the initiation of the experiment. Benthic recruitnzent device 0.0-7.8 7.8-15.6 15.6-23.4 23.4-31.2 31.2-39.0 Distance from Faceplate Edge (cm) Fig. 3. Recruitment of Botrylloides diagensis on ELVIS as a function of distance from the faceplate edge. Average abundance (?I SD) on 7.8-cm-diameter adjacent samples running from the edge to 39.0 cm from the faceplate edge. Lines connecting sample distances are not significantly different (one-way ANOVA, with data blocked by transect, and Bonferroni-Dunn contrast test; P > 0.05) diurnal variations in the recruitment patterns of epifauna are summarized in Fig. 4. Total recruit abundance varied signif- icantly with exposure period; highest numbers were recorded on panels exposed from 0600 to 1200 h, while substrates exposed from 1500 to 0600 h had the fewest recruits. As- cidian (Diplosonza nzacdonaldi, Ascidiella aspersa, and Bo- trylloides diagensis) recruitment patterns, the most abundant species recorded in both deployments, were similar to the overall diurnal recruitment patterns. Owing to low individual numbers of species, all bryozoan recruits (Bugula turrita, Crytosula pallasiana, and Schizoporella errata) were com- bined. Most bryozoan recruits were found on panels exposed in the early morning (0600-0900 h). Spirorbid polychaete (Spirorbis spp.) recruits, in contrast, were only found on sub- strates exposed during the middle of the day (0900-1500 h). observed diurnal variations in epifauna recruitment gener- ally are consistent with data collected from the daylnight exposure deployments (Table 1). Infaunal recruitnzent relative to variations in tidal state- Unfortunately, relatively few individuals recruited to the sandy substrates in both ELVIS deployments. Of the three species (Mya arenaria, Polydora cornuta, and an unidenti- fied oligochaete) found in the sediment cups, only Polydora recruits averaged more than one individual per exposure pe- riod. Polydora recruitment varied significantly with tidal state-greater numbers of recruits were found in substrates exposed during the low tide period than during any other tidal states (Fig. 5). The same tidally related recruitment pat- tern of Polydora was found in both deployments. Discussion Experimental tests of the operating assumptions of the benthic invertebrate recruitment device revealed that the de- Diplosoma _h, ,200-1- ,-em 5'" I 1 Ascidiella Diplosoma _1TL ,m-,YO ,ra--Om Bryozoans Time of Day (h) Fig. 4. Total recruits and numbers of the most abundant taxa (Diplosoma macdonaldi, Ascidiella aspersa, Botrylloides diagensis, and bryozoans) found on substrates repeatedly exposed at 0600- 0900, 0900-1200, 1200-1500, and 1500-0600 h for a 7-d period. Values are averages (+ 1 SD) of recruit abundance. Lines connecting treatments are not significantly different from each other (one-way ANOVAs and Bonferroni contrasts tests; P > 0.05). A. 20-27 Sep- tember deployment; B. 4-1 1 October deployment. sign could repeatedly expose sets of substrates to a broad array of naturally recruiting infaunal and epifaunal species at repeated, predefined environmental conditions for periods up to 2 weeks. The relatively simple design of substrate ex- posure and maintenance of organisms on unexposed sub- strates also indicated that the device was reasonably effective in isolating substrates from larval contamination while not adversely affecting recruits (at least for the test species As- cidiella aspersa) previously attached to the sequestered sub- strates. Users of the device, however, should carefully con- sider the potential effects of environmental conditions on particular species in enclosed vs. exposed substrates. For in- stance, using sediment as substratum may lead to possible artifacts from varying exposure times due to differential re- cruit mortality in the enclosures and sediment scour and re- suspension in the exposed sediment cups. conditions in which there is a large vertical component to water flow may Whitlatch and Osman Low T~de Flood T~de Hlgh T~de Ebb Tide Low T~de Flood T~de Hlgh T~de Ebb T~de Treatment Treatment Fig. 5. Average number ( 2 1 SD) of total recruits (hatched histograms) and Polydora cornutu recruits (open histograms) found in substrates repeatedly exposed at four different tidal states for two different 7-d deployment intervals (A. 7-14 July; B. 3-10 September). result in biased larvae transport to, and retention on, the experimental substrates. ELVIS deployments to date have been conducted in relatively low-water-flow conditions. The device is likely most effective in relatively calm conditions, or when deployed several meters below the surface of the water. Habitat- and species-specific operation tests of the de- vice similar to those conducted are recommended in order to fully understand potential artifacts and limitations of the recruitment device. Preliminary short-term field trials examining recruitment patterns of a variety of epifaunal and infaunal species re- vealed the device provided temporally repeatable results that were generally consistent with previously published studies. Pronounced diurnal patterns in recruitment were found in five different taxa of sessile epifaunal invertebrates (bryo- zoans and spirorid polychaetes and three species of ascidi- ans). Previous studies on ascidian and bryozoan reproductive behavior have indicated that larval release can be controlled by light intensity (Ryland 1977; Svane and Young 1989). The nonfeeding ascidian larvae typically reside in the water column for relatively short periods (minutes to hours for colonial forms; days to a week for soliary forms), and sev- eral species have been observed to settle during the midday period (Olson 1983; Lindquist et al. 1992). Although the device found that recruitment of the infaunal polychaete Polydora cornuta was related to tidal state, there are conflicting reports regarding the generality of this find- ing. Levin (1986) reported 10-fold variations in the abun- dance of polychaete larvae over a tidal cycle and that Polydora cornuta (=ligni) tended to be most abundant dur- ing mid-ebb and late-flood tide periods during the daytime and high and ebb tides during the nighttime. The abundance of competent Polydora larvae, however, exhibited no dis- cernable tidally mediated abundance patterns. In contrast, Stancyk and Feller (1986) reported higher numbers of poly- chaete larvae collected at high slack water than in ebb or flood conditions. The modeling study of Gross et al. (1992) predicted that larval availability near the seabed should be higher during slack water and lower during peak flow and ebb tides because of enhanced mixing with higher flow that mixes larvae away from the bed. Further studies are clearly required to assess whether there are general patterns of in- faunal recruitment relative to tidal state and current velocity. We have attempted to keep the design and operation of the recruitment device as simple as possible in order to meet our specific needs and basic assumptions. Once we have more fully explored the behavior of the device, we will con- trol its operation using environmental sensors (e.g. light me- ters, pressure transducers, current meters) interfaced to the microprocessor. This step will allow us to more easily study specific types of environmental conditions and deploy the device in a fully automated mode. Because the recruitment device can repeatedly expose a substrate over an extended period (e.g. up to 60 d), final collections should contain an- imals of different sizes depending on whether they recruited early or late in the collection period. It may therefore be possible to obtain added information on when the recruits were collected and determine if behavioral patterns change over a given long-term deployment period. References BUTMAN, C. A. 1987. Larval settlement of soft-sediment inverte- brates: The spatial scales of pattern explained by active habitat selection and the emerging role of hydrodynamical processes. Oceanogr. Mar. Biol. Ann. Rev. 25: 113-165. COMMITO, J. A., S. THUSH, R. D. PRIDMORE, J. E. HEWITT, AND V. J. CUMMINGS. 1995. Dispersal dynamics in a wind driven benthic system. Limnol. Oceanogr. 40: 1513-151 8. CRISP, D. J. 1955. The behaviour of b'arnacle cyprids in relation to water movements over a surface. J. Exp. Biol. 32: 569-590. . 1974. Factors influencing the settlement of marine inver- tebrate larvae, p. 177-265. In P. T. Grant and A. M. Mackie [eds.], Chemoreception in marine organisms. Academic. , AND H. BARNES. 1954. The orientation and distribution of barnacles at settlement with particular reference to surface con- tour. J. Anim. Ecol. 23: 142-162. CRONIN, T. W., AND R. B. FORWARD. 1982. Tidally timed behavior: Effects on larval distributions in estuaries, p. 505-520. In V. S. Kennedy led.], Estuarine comparisons. Academic. DAVIS, A. R., A. J. BUTLER, AND I. VAN ALTENA. 1991. Settlement behaviour of ascidian larvae: Preliminary evidence for inhibi- tion by sponge alleolochemicals. Mar. Ecol. Prog. Ser. 72: 117- 123. Benthic recruitment device DIJRILIER, N. 1988. H,S-a settlement cue or toxic substance for Capitella sp. I larvae? Biol. Bull. 174: 30-38. ECKIMAN, J. E. 1983. Hydrodynamic processes affecting benthic recruitment. Limnol. Oceanogr. 28: 241-257. GAINES, S., AND J. ROUGHGARDEN. 1987. Fish in offshore kelp forests affect recruitment to intertidal barnacle populations. Science 235: 479-481. GROSS, T. E, E E. WERNER, AND J. E. ECKMAN. 1992 Numerical modeling of larval settlement in turbulent bottom boundary layers. J. Mar. Res. 50: 61 1-642. GROSSBERG, R. K. 198 1. Competitive ability influences habitat choice in marine invertebrates. Nature 290: 700-702. HAWKINS, . J., AND R. G. HARTNOLL. 1982. Settlement patterns of Semibalanus balanoides (L.) in the Isle of Man (1977- 1981). J. Exp. Mar. Biol. Ecol. 62: 271-283. JONSSON, I? R., C. ANDRB, AND M. LINDEGARTH. 1991. Hydrody- namic near-bottom confinement of settling marine bivalve lar- vae in a flume boundary-layer flow. Mar. Ecol. Progr. ser. 79: 67-76. KEEN, S. L. 1987. Recruitment of Aurelia aurita (Cnidaria: Scy- phozoa) larvae is position-dependent, and independent of con- specific density, within a settling surface. Mar. Ecol. Progr. Ser. 83: 151-160. KORRIGNA, I? 1949. Experiments and observations on swarming, pelagic life and setting in the European flat oyster, Ostrea ed- ulis L. Arch. Neerl. Zool. 5: 1-249. LEVIN, L. A. 1986. The influence of tides on larval availability in shallow waters overlying a mudflat. Bull. Mar. Sci. 39: 224- 233. LINDQUIST, N., M. E. HAY, AND W. FENICAL. 1992. Defense of ascidians and their conspicuous larvae: Adult vs. larval chem- ical defenses. Ecol. Monogr. 62: 547-568. MULLINEAUX, L. S., AND C. A. BIJTMAN. 1990. Recruitment of encrusting benthic invertebrates in boundary-layer flows: A deep-water experiment on Cross Seamount. Limnol. Oceanogr. 35: 409-423. , AND . 1991. Initial contact, exploration and attach- ment of barnacle (Balanus amphitrite) cyprids settling in flow. Mar. Biol. 110: 93-103. , AND E. D. GARLAND. 1993. Larval recruitment in response to manipulated field flows. Mar. Biol. 116: 667-683. OLFASSON, E. B., C. H. PETERSON, AND W. G. AMRROSE, JR. 1994. Does recruitment limitation structure populations and com- munities of macro-invertebrates in marine soft sediments: The relative significance of pre- and post-settlement processes. Oceanogr. Mar. Biol. Ann. Rev. 32: 65-109. OLSON, R. R. 1983. Ascidian-Prochloron symbiosis: The role of larval photoadaptation in midday larval release and settlement. Biol. Bull. 165: 221-240. OSMAN, R. B., AND R. B. WHITLATCH. 1995a. Predation on early ontogenetic life stages and its effect on recruitment into a ma- rine epifaunal community. Mar. Ecol. Progr. ser. 117: 11 1-126. , AND - . 19956. The influence of resident adults on recruitment: A comparison to settlement. J. Exp. Mar. Biol. Eco~. 190: 169-198. , AND - . 1995c. The influence of resident adults on larval settlement: experiments with four species of ascidians. J. Exp. Mar. Biol. Ecol. 190: 199-220. PAWLIK, J. 1992. Chemical ecology of the settlement of benthic marine invertebrates. Oceanogr. Mar. Biol. Ann. Rev. 30: 273- 335. , C. A. BUTMAN, AND V. R. STARCZAK. 1991. Hydrodynam- ic facilitation of gregarious settlement of a reef-building tube worm. Science 251: 421-424. PECHENIK, J. 1990. Delayed metamorphosis by larvae of benthic marine invertebrates: Does it occur? Is there a price to pay? Ophelia 32: 63-94. RYLAND, J. S. 1977. Taxes and tropisms in bryozoans, p. 41 1-435. In R. M. Woollacott and R. L. Zinn reds.], Biology of bryo- zoans. Academic. SCHLICHTING, H. 1979. Boundary-layer theory. 7th ed. McGraw- Hill. SMITH, C. R., AND S. J. BRUMSICK~~E. 1989. The effect of patch size and substrate isolation on colonization modes and rate in an intertidal sediment. Limnol. Oceanogr. 34: 1263-1277. SNELGROVE, P. V. R., C. A. BIJTMAN, AND J. E GRASSLE. 1993. Hydrodynamic enhancement of larval settlement in the bivalve Mulinia lateralis (Say) and the polcyhaete Capitella sp. I in microdepositional environments. J. Exp. Mar. Biol. Ecol. 168: 71-109. - , AND - . 1995. Potential flow artifacts asso- ciated with benthic experimental gear: Deep-sea mudbox ex- amples. J. Mar. Res. 53: 821-845. , J. E GRASSLE, AND R. E PETRECCA. 1996. Experimental evidence for aging food patches as a factor contributing to high macrofauna density. Limnol. Oceanogr. 41: 605-614. SRIKRISHNADHAS, B., AND K. RAMAMOORTHI. 1982. Ecology of polychaete larvae of Porto Novo waters. Indian J. Mar. Sci. 11: 296-302. STANCYK, S. E., AND R. J. FELLER. 1986. Transport of non-decapod invertebrate larvae in estuaries: An overview. Bull. Mar. Sci. 39: 257-268. SVANE, I., AND C. M. YOUNG. 1989. The ecology and behaviour of ascidian larvae. Oceanogr. Mar. Biol. Ann. Rev. 27: 45-90. THORSON, G. 1964. Light as an ecological factor in the dispersal and settlement of larvae of marine bottom invertebrates. Ophe- lia 1: 167-208. WALTERS, L. J. 1992. Field settlement locations on subtidal marine hard substrata: Is active larvae exploration involved? Limnol. Oceanogr. 37: 1101-1 107. WETHEY, D. 1986. Ranking of settlement cues by b,arnacles larvae: Influence of surface contour. Bull. Mar. Sci. 39: 393-400. WOODIN, S. A. 1985. Effects of defecation by arenicolid polychaete adults on spionid polychaete juveniles in field experiments: Se- lective settlement or differential mortality. J. Exp. Mar. Biol. Ecol. 87: 119-132. . 1986. Settlement of infauna: Larval choice? Bull. Mar. Sci. 39: 401-407. YOUNG, C. M. 1990. Larval ecology of marine invertebrates: A sesquicentennial history. Ophelia 32: 1-48. Received: 31 October 1996 Accepted: 13 August 1997 Amended: 22 September 1997