Factors Affecting the Recruitment of Juvenile Caribbean Spiny Lobsters Dwelling in Macroalgae

In south Florida, Caribbean spiny lobsters (Panulirus argus) settle and spend their first few months in macroalgae or seagrass. After a few months, these “algal-phase” juveniles emerge from vegetation and, as “postalgal-phase” juveniles, seek refuge in crevices, often dwelling in groups. The importance of crevice shelters in determining the abundance of postalgal-phase juvenile spiny lobsters has been studied, but we know little about the processes affecting lobster distribution and survival during their cryptic algal-dwelling phase. We found that postlarval supply varied independently of changes in the structure of macroalgal settlement habitat. For this reason, postlarval supply alone can not reliably predict local settlement density. Changes in the size of macroalgal patches in particular tend to increase the variability in settlement density among locations and times. Field and mesocosm experiments indicate that social interactions and individual movements are unlikely to alter the general distribution of algal-phase lobsters established at settlement. But if algal-phase lobsters are aggregated at scales <1 m (e.g., due to patchy settlement), they experience higher mortality than non-aggregated lobsters, as revealed in field experiments where lobsters were tethered alone or in pairs and at varying inter-individual distances. Field manipulations of settlement density indicate that recapture (survival) of microwire tagged algalphase juveniles is positively associated with features of the habitat that affect lobster density (e.g., site area, macroalgal patch size), but survival and growth of lobsters are unrelated to artificially manipulated settlement density. Collectively, these results imply that the population dynamics of juvenile P. argus dwelling in macroalgae are not typically regulated by density-dependent processes, although density-dependent predation may be locally important in patches when settlement is episodically high. Extreme variance in recruitment, defined as the number of new individuals within a population that survive to a specified size, age, or ontogenetic stage, is characteristic of marine animals with “open” populations whose meroplanktonic larval stages drift for weeks or months in the water column. Many of these species also have complex life histories with multiple developmental stages requiring distinctly different habitats (Roughgarden et al., 1988). Therefore, understanding the processes affecting recruitment of such species requires information on larval availability, larval settlement, ontogenetic shifts in stage-specific habitat requirements, changes in nursery habitat availability and structure, and other factors that might alter post-settlement survival and growth. A recurrent theme in the marine ecological and fishery literature centers on the extent that recruitment is limited either by larval settlement or post-settlement processes (Underwood and Fairweather, 1988; Grossberg and Levitan, 1992). A significant relationship between larval settlement and recruitment to some later life stage is typically considered evidence for supply-side (“recruitment”) population regulation and also the absence of density-dependent post-settlement mortality (Connell, 1985; Doherty and Williams, 1988). If postsettlement events, such as interor intraspecific competition or predation, are to regulate population density, then it is generally believed that they must act in a density-dependent manner (Hughes, 1990). If so, the relationship between larval settlement and recruitment is destroyed so that the age structure of recruits no longer reflects prior settlement. Yet, the 4 BULLETIN OF MARINE SCIENCE, 61(1): 3–19, 1997 results of at least one mathematical model indicate that a settlement “signal” can persist in a population’s age structure despite density-dependent mortality, especially where settlement density is low or density-dependence is weak (Holm, 1990). The literature describing the relative importance of preand post-settlement processes in regulating populations of marine organisms is rich with examples from crustacea. Barnacle populations on rocky shores in California (USA) (Gaines and Roughgarden, 1985; Possingham and Roughgarden, 1990 ) and spiny lobster populations in Western Australia (Chittleborough and Phillips, 1975; Pearse and Phillips, 1988; Caputi et al., 1995) are strongly affected by variation in larval (or postlarval) supply, a large part of which is driven by alterations in coastal currents. The recruitment of clawed lobsters in New England (USA) (Wahle and Steneck, 1991; Incze and Wahle, 1991), stomatopods in Panama (Steger, 1987), and Dungeness crabs in Washington (USA) (Eggleston and Armstrong, 1995) are thought to be limited by post-settlement mortality, notably the availability of suitable shelter for protection from predators. Similarly, shelter availability is critical to the recruitment of spiny lobsters in the Caribbean (Herrnkind and Butler, 1986; Butler and Herrnkind, 1992; Eggleston et al., 1990; Mintz et al., 1994; Field and Butler, 1994) and in Hawaii (Parrish and Polovina, 1994; Polovina et al., 1995), where it appears to regulate recruitment by setting thresholds (i.e., a local carrying capacity) for late-stage juveniles and subadults. One hypothesis is that recruitment is limited above these shelter-imposed thresholds, but below them the population fluctuates in response to larval supply (Forcucci et al., 1994; Polovina et al., 1995; Herrnkind and Butler, 1994; Butler and Herrnkind, 1997). In this paper we examine the relationship between postlarval supply, settlement habitat dynamics, and early post-settlement processes that are likely to affect the distribution and survival of algal-phase juvenile Caribbean spiny lobster (Panulirus argus Latreille) in the Florida Keys (USA). We do this by drawing inference from: 1) field observations of postlarval supply and changes in macroalgal habitat structure that may alter settlement density, 2) mesocosm experiments that test whether algal-phase lobsters are gregarious, which could alter their post-settlement distribution, 3) field experiments testing whether residency in macroalgae is affected by short-term changes in prey availability, 4) tethering studies that test if cryptic algal-phase lobsters experience higher mortality when aggregated, and 5) field experiments that directly test whether microwire-tagged lobsters released into the wild experience density-related differences in growth or survival. METHODS AND MATERIALS THE COMPLEX LIFE CYCLE OF THE CARIBBEAN SPINY LOBSTER. The Caribbean spiny lobster supports the most economically important fishery in Florida (Hunt, 1994) and is heavily fished throughout its range from Bermuda to southern Brazil. It has a complex life cycle requiring three distinct habitats: coral reef (adults), open ocean (larvae), and shallow, vegetated coastal zone (juveniles). The early life history and ecology of P. argus is reviewed elsewhere (Herrnkind et al., 1994), so we only summarize the ecology of life history stages relevant to this study, namely the postlarval and early benthic juvenile stages. Following a prolonged oceanic larval period of 9 months or more, the phyllosome larvae metamorphose near the continental shelf break into non-feeding, strongly swimming puerulus postlarvae. Postlarvae enter Florida Bay, the major spiny-lobster nursery in Florida, over a period of a several days every lunar cycle, usually at night during new-moon flood tides. They settle and metamorphose on macroalgae-covered hard-bottom habitat, particularly among clumps of the ubiquitous red macroalga Laurencia spp. and, less frequently, in seagrass. In the Florida Keys, juvenile spiny lobsters exhibit 5 BUTLER ET AL.: SPINY LOBSTERS IN MACROALGAE three behaviorally and ecologically distinct phases following settlement: algal -phase, postalgal-phase, and subadult. The algal-phase juveniles (5 15 mm carapace length; CL) remain for a few months in vegetation, where they are sheltered from predators and have abundant prey. Upon reaching 15-20 mm CL, they emerge from settlement habitat and, as postalgal juveniles, take up daytime refuge under crevices provided by rocks, sponges, octocorals, or other structures. Approximately 1 yr after settling the juveniles become nomadic within the nursery and after 2 yr, they mature and migrate seaward to the reef tract. ESTIMATING POSTLARVAL SUPPLY AND MACROALGAL STRUCTURE. We estimated monthly changes in postlarval supply from the number of postlarval spiny lobsters collected from five Witham-type surface collectors deployed approximately 200 m offshore of the southwest end of Long Key, FL, adjacent to Long Key Channel. The collectors were sampled 7 d after each new moon from June 1991 August 1992. Details of their construction and sampling is described elsewhere (Heatwole et al., 1992; Phillips and Booth, 1994). Tidal currents from Long Key Channel surge into Florida Bay and pass over an area just south of the Arsnicker Keys (Herrnkind and Butler, 1994) where we monitored macroalgal habitat structure (described below). This channel therefore serves as an important “source” of postlarvae for our macroalgal monitoring sites and the Witham collectors we deployed near the channels are likely to yield good estimates of postlarval supply to the region (Herrnkind and Butler, 1994). Monthly changes in macroalgal habitat structure were monitored at 27 separate hard-bottom sites, each surrounded by seagrass and situated approximately 0.5 km south of the Arsnicker Keys in Florida Bay. The sites were spread roughly east-west over an area > 2 km. The size of the sites ranged from about 200 – 1000 m and all were 2 – 3 m deep. In June 1991, we established four haphazardly selected, non-overlapping permanent transect locations within each site by driving stakes into the substrate at the ends of each transect. Transects varied in length depending on the configuration of the site. Divers visited each site once a month from June–November 1991, February–March 1992, and June–August 1992, stretched underwater tape measures between each set of stakes, and then measured the length (cm) of all substrate types lying under the tape. When a patch of red macroalgae Laurencia spp. was encountered, the divers also measured its height (cm) from the substrate every 0.25 cm along the tape. Data from the four replicate transects were then used to calculate the mean (and 1 SD) percent cover, patch size, and height of macroalgae at each site during each sample month. During some months, our observations were precluded by dense blooms of cyanobacteria (Butler et al., in press), so the data set is temporally discontinuous. The monthly correspondence between postlarval supply and the three separate measures of macroalgae structure (i.e., percent cover, patch size, patch height), averaged across sites, were examined using the Pearson correlation statistic. In addition to the standard estimate of postlarval supply derived from the Witham collectors, we created two additional indices of potential settlement density by expressing postlarval supply (numerator) in terms of macroalgal percent cover and patch size (separate denominators) for each site. Postlarval supply/percent cover provides a relative estimate of postlarval density per site, whereas postlarval supply/patch size provides an index of crowding. Macroalgal height was not included as an index because it is positively correlated with the percent cover of macroalgae (r = 0.82, P < 0.001) and, therefore, not independent. ESTIMATING GREGARIOUSNESS. This experiment was conducted in September, 1988, at the Keys Marine Laboratory on Long Key, FL in nine 2.5 m diameter x 0.3 m tall plastic tanks with sand covered bottoms. We recirculated sea-water through each tank and covered them with shade-cloth. Two 2-liter clumps of Laurencia spp., collected fresh from the field, were placed 1 m apart in the center of each tank. Between 16:00 and 18:00 h, two algal-phase lobsters (5.5 – 11.4 mm CL) were released on the open sand in the center of each tank. Forty-eight hours later, the macroalgal clumps were recovered with a fine-mesh handnet and the frequency of lobster cohabitation within macroalgal clumps was recorded. A chi-square goodness-of-fit test was used to evaluate whether the frequency of cohabitation or solitary habitation by lobsters in macroalgal clumps differed from a random distribution. Greater than expected cohabitation would suggest that algal-phase lobsters are gregarious. Less than expected 6 BULLETIN OF MARINE SCIENCE, 61(1): 3–19, 1997 cohabitation would suggest that lobsters are anti-social. A random distribution of lobsters would indicate that lobsters in this ontogenetic stage are asocial, neither preferring the company of conspecifics nor avoiding it. RESIDENCY OF ALGAL-PHASE LOBSTERS IN RESPONSE TO FOOD AVAILABILITY. The presence of prey in macroalgal clumps (Marx and Herrnkind, 1985b) reportedly can affect the small-scale dispersal and residency patterns of algal-phase juvenile P. argus. We tested this hypothesis in a field experiment conducted in July, 1986, at a hard-bottom site 1.5 m deep and approximately 50 m offshore (north) of the Keys Marine Laboratory on Long Key, FL. The 14 m x 16 m site was devoid of any prominent physical structures other than sparse sprigs of calcareous green macroalgae, which we removed by hand. We then anchored 42 2-liter clumps of Laurencia spp. to the sea floor 2 m apart in a 6 x 7 array using monofilament and lead weights. Prior to placement on the bottom, every other clump in the array was rinsed in sea-water to reduce the number of lobster prey (e.g., amphipods, isopods, copepods, gastropods, echinoderms, etc.) dwelling in the macroalgae. Thus, half the clumps contained natural prey densities (high food treatment) and half contained a reduced number of prey (low food treatment). We did not control for artifacts caused by the rinsing procedure (e.g., rinse the high food treatment macroalgal clumps and then to replace the prey) because it was shown to have no significant impact by Marx and Herrnkind (1985b). Forty-two algal-phase lobsters, obtained from Witham surface collectors, then received unique paint marks on their legs before each was implanted individually by divers into each of the macroalgal clumps in the array. After 24 h, we retrieved each clump in a fine-mesh hand net; the clumps were dismantled in the laboratory and we recorded the number and identification of those lobsters found. Two trials of this experiment were conducted. EFFECT OF AGGREGATION ON SURVIVAL OF ALGAL-PHASE LOBSTERS. Three separate tethering experiments were conducted to address different questions (listed below as 1 3) concerning the effects of aggregation on the probability of lobster mortality by predators. 1. Does the mortality of solitary algal-phase lobsters differ from that measured for lobsters aggregated in pairs? In July, 1986, we obtained algal-phase lobsters from Witham collectors and tethered them alone (n = 30) or in pairs (n = 15 pairs; 30 lobsters total) in seagrass and macroalgal habitat located approximately 50 – 100 m offshore (north) of the Keys Marine Laboratory on Long Key, FL. Our tethering protocol is described elsewhere (Herrnkind and Butler, 1986; Smith and Herrnkind, 1992). After 24 h, the number of lobsters found alive or missing in each treatment was recorded, and a log-linear categorical analysis was used to determine if lobster condition (i.e., alive vs. missing) was independent of tethering condition (i.e., solitary vs. paired). 2. Does the distance between pairs of lobsters affect the probability that one or both individuals are killed? Pairs of algal-phase lobsters were tethered in July 1986, 1990, 1991, and 1995 in seagrass and macroalgal habitat at the same location near the Keys Marine Laboratory. Pairs of tethered lobsters were separated by either 25, 75, or 200 cm, forming three treatment groups. Again, lobster presence or absence was recorded after 24 h and the data were analyzed using log-linear categorical analyses that tested whether lobster survival was independent of distance between individuals and experimental trial (year). 3. Does the spatial pattern of predation on tethered lobsters differ with respect to the spatial scale over which lobsters are distributed? Algal-phase lobsters were tethered in seagrass in three separate 5 x 5 arrays. Inter-individual distances were 25, 75, and 200 cm. Thus, arrays of 25 lobsters covered 1, 9 and 100 m, respectively. After 24 h, lobster presence or absence was recorded as was their position within the array. We used a Monte-Carlo based simulation model and statistic (C) designed for analysis of spatial patterns on grids (Stapanian et al., 1982) to determine if the occurrence of predation was random, clumped, or uniform when lobsters are aggregated at the three different densities. The density of lobsters in the three arrays corresponded to densities of 25 lobsters m in the small array, 2.78 lobsters m in the 9 m array, and 0.25 lobsters m in the 100 m array. The two lower densities are comparable to previously published natural densities of about 0.03 lobsters m (Marx and Herrnkind, 1985b; Herrnkind and Butler, 1994) and recently recorded settlement densities in the range of 2 – 4 lobsters m (M. Butler, W. Herrnkind, J. Hunt, R. Bertelsen, unpubl. data), whereas the highest density (= smallest array) is considerably greater than any recorded. 7 BUTLER ET AL.: SPINY LOBSTERS IN MACROALGAE FIELD MANIPULATIONS OF ALGAL-PHASE LOBSTER DENSITY. We directly tested for the possible effects of increased density on algal-phase juvenile lobster population dynamics in two separate field experiments that differed in protocol. The first experiment ran for 20 d and we manipulated lobster density by altering site size rather than the number of lobsters released. The second experiment continued for over a year and lobster density was manipulated by releasing two different numbers of lobsters into sites with natural boundaries and which varied in size. Thus, the experiments differed both in the way settler density was manipulated and duration of the experiment. The first experiment was designed to reveal short-term density effects occurring during the first three weeks that lobsters resided in macroalgae, whereas the second assessed the effect of density over the entire algal-dwelling period and the early postalgal phase when lobsters were finally collected. The first experiment was conducted in July, 1990, 1 km north of the northeast tip of Grassy Key, FL and just west of Tom’s Harbor Channel and Channel Key. Four separate circular macroalgal-covered sites (1.5 m deep) were established by removing all vegetation and structure (e.g., rocks, sponges) from a 2 m wide belt surrounding the sites; two sites were 100 m and two were 50 m. Algal-phase lobsters obtained from Witham collectors were then implanted into the sites: 25 lobsters added per site. It was impossible to search all the macroalgae to search for algal-phase lobsters, so after 20 d we removed all of the macroalgae from 30% of the bottom at each of the four sites. That is, we sampled the same proportion of area in each site. Divers sampled by gathering algae in mesh bags at 2 m areas haphazardly selected on each site. The bags of macroalgae were then meticulously searched at boatside. The second experiment was conducted from June 1991–August 1992 at 18 of the 27 natural hardbottom sites (193 722 m area) near the Arsnicker Keys where we were also monitoring changes in macroalgal habitat structure (described earlier in the section, Estimating Postlarval Supply and Macroalgal Structure). Once a month from June–December 1991, we collected algal-phase juvenile lobsters from Witham collectors, marked them with internal microwire tags, and then implanted each tagged lobster directly into clumps of macroalgae on each of the sites. Nine sites received high levels of enhancement (“High Seed” treatment; 182 lobsters added site total) which was four times as many tagged lobsters as implanted in the nine “Low Seed” treatment sites (46 lobsters added site total). Microwire tagging has only minor effects on the survival and growth of juvenile spiny lobsters and has been used in estimating growth of juvenile lobsters in the field (Lellis and Pardee, 1991; Phillips et al., 1992). The number of lobsters implanted on high and low density sites varied among months in accordance with natural fluctuations in postlarval supply to our Witham collectors. Divers searched each site and collected all lobsters found each month from June 1991–August 1992. We searched for lobsters around structures that provide shelter for postalgal-phase juveniles (e.g., sponges, corals, solution holes, and rock crevices); macroalgae was not searched, so we primarily recovered postalgalstage juveniles. All the lobsters collected were measured, and evaluated for the presence of a microwire tag. Untagged lobsters and tagged lobsters < 25 mm CL were returned to the site. As described earlier (see section on Estimating Postlarval Supply and Macroalgal Structure), we also surveyed the macroalgal habitat on each site each month to follow changes in habitat structure that might be associated with the recapture of tagged lobsters. The results of this experiment were evaluated by comparing the number and sizes of microwire tagged lobsters recovered from high seed and low seed treatment sites (MannWhitney Rank sum test), and by examining (best-fit multiple regression) whether the proportion of tagged lobsters that were recaptured could be predicted from variables describing the experimental conditions at each site (e.g., number of tagged lobsters added to site, number of untagged lobsters captured, area of the site, macroalgal percent cover, macroalgal patch size, and height of macroalgae). 8 BULLETIN OF MARINE SCIENCE, 61(1): 3–19, 1997