Eggleston et al Artificial shelters and survival of juvenile Panulirus argus 



695 



acrylate cement ensured that a piece of carapace re- 

 mained on the line as evidence of predator-induced mor- 

 tality. An empty cable tie without a piece of carapace 

 attached to it was scored as an escape. Each cable-tie 

 was connected with 301b test monofilament line either 

 to another cable-tie and attached to a shelter, or at- 

 tached to a square wire-metal frame that was posi- 

 tioned outside of the triangular casita station (Fig. 3) 

 on the seagrass bed with lead weights. The wire-metal 

 frame had the same length-width dimensions as the 

 large casita but did not provide shelter. The metal 

 frame was chosen over stainless-steel stakes because 

 stakes could not penetrate the underlying carbonate 

 platform at the inner-bay site. The metal frames were 

 visually inconspicuous because they were covered by 

 a thin layer of sediment. Tether lengths of 0.7m pro- 

 vided a foraging area of about l.Sm^ and prevented 

 tangling between adjacent lobsters. Although tether- 

 ing does not necessarily measure absolute rates of 

 predation, it does measure relative rates of predation 

 (Heck and Thoman 1981), which can serve to compare 

 mortality rates as a function of different experimen- 

 tal treatments. 



We used a stationary visual census technique (Bohn- 

 sack and Bannerot 1986) to quantify the community 

 structure of potential predators associated with casita 

 and no-casita stations during the experimental period 

 (January and August 1989). Visual censuses were per- 

 formed between 10:00 and 14:00 hours with three 

 replicate samples taken during the experimental peri- 

 od. By performing the visual censuses during midday, 

 we maximized the visibility available for species iden- 

 tification. Nighttime observations were not performed 

 because our previous study (Eggleston et al. 1990) in- 

 dicated that the predator guild normally associated 

 with the casitas dispersed widely over the seagrass bed 

 at night. However, predator movements were observed 

 during one dawn and dusk crepuscular period. 



We examined the daytime foraging ranges of casita- 

 associated predators by swimming along a transect 

 perpendicular to each casita. When potential predators 

 were observed, a float was set to mark the location, 

 whereupon a scuba diver then followed the predators 

 to assure that they were associated with the casita. Our 

 initial observations indicated that piscine predators 

 associated with casitas seldom moved more than 

 30-40 m away from a casita. 



Experimental design 



Before initiating the tethering experiments in 1989, we 

 deployed casitas at the inner-bay and outer-bay sites 

 during 1988. During July 1988 at the inner-bay site, 

 we positioned a row of six large casitas 25 m apart from 

 one another (Fig. 3). Each large casita had one medium 



and one small casita placed 10m away, yielding six 

 stations with one small, medium, and large casita 

 arranged in a triangle (Fig. 3). At the outer-bay site 

 during August 1988, we positioned six small, medium 

 and large casitas equidistant between the shore and 

 reef line and arranged these in two rows, each contain- 

 ing three triangular stations (Fig. 3b). See Eggleston 

 et al. (1990) and Eggleston and Lipcius (1992) for a 

 complete description of the small and medium casitas 

 and their use in other field experiments. Two separate 

 tethering experiments were then performed during 

 January and August 1989. 



The first experiment was performed during January 

 1989. In this study we examined the survival of two 

 sizes of juvenile lobsters with and without access to 

 shelter at both the inner-bay and outer-bay sites. Six 

 metal-frame, no-casita stations were placed 60-70 m 

 away and perpendicular to the casitas in sparse-to- 

 moderate-density Thalassia at both sites (Fig. 3). 

 Juvenile lobsters were divided into two size-classes: 

 small, 46-55 mm carapace length (CL) as measured dor- 

 sally from the base of the supraorbital spines to the 

 posterior border of the cephalothorax; and large, 

 56-65 mm CL. Lobsters were tethered for 7 days. Each 

 casita and no-casita station at both sites had six 

 tethered lobsters from either of two size-classes for a 

 total of 144 tethered lobsters (2 sites x 6 lobsters x 

 2 sizes X 2 treatments (casita vs. no-casita) x 3 

 replicate stations). 



Based on our initial observations of predator forag- 

 ing ranges (see above), we assumed that our choice of 

 60-70 m for the no-casita station was well beyond the 

 foraging range of diurnally active predators, thereby 

 providing unbiased estimates of lobster survival in the 

 absence of artificial shelters (i.e., mortality estimates 

 were not biased towards finding significantly higher 

 predation rates on lobsters tethered within the forag- 

 ing range of casita-associated predators). However, our 

 observations during the January 1989 experiment in- 

 dicated that some predators moved nearly 60 m from 

 the casitas (see Results). Thus, although the 70 m 

 distance from the large casitas was probably beyond 

 the foraging range of casita-associated predators, the 

 60 m distance from the small and medium casitas was 

 probably not. 



Before initiating the second tethering experiment, 

 we positioned a row of three large casitas equidistant 

 between the shore and reef line in July 1989 at the 

 outer-bay site (Fig. 4). In August 1989 we examined 

 how lobster survival varied with distance from the 

 casitas. Three metal-frame no-casita stations were 

 placed 15, 30, and 70 m away and perpendicular to the 

 large casitas (Fig. 4). Based on the foraging ranges of 

 predators during the January experiment (see above), 

 we assumed that 70 m was an adequate distance to 



