Petrik et al.: Recruitment of Mkropogonias undulatus 



955 



1982); 2) Atlantic cod grow faster in seagi-ass habitats 

 than in sand, rocky reef, or cobble habitats (Tupper 

 and Boutilier, 1995); and 3) pinfish exhibit higher 

 growth rates in seagrass than in sand habitats (Levin 

 etal., 1997). 



Atlantic croaker, Micropogonias undulatus (here- 

 after referred to as croaker), range from Cape Cod to 

 Campeche Bank, Mexico (Johnson, 1978), and occur 

 both offshore and in estuaries in a variety of habi- 

 tats including mud, sand, and seagrass (White and 

 Chittenden, 1977; Johnson, 1978; Rooker et al., 

 1998). Croaker are an important component of com- 

 mercial fisheries in the Gulf of Mexico and south- 

 eastern United States, often dominating bottom fish 

 landings, and are an important sport fisheiy in this 

 region (Lassuy, 1983). In the Gulf of Mexico, croaker 

 spawn over the continental shelf or near inlets from 

 September to May with peak levels occurring before 

 January (Johnson, 1978; Cowan, 1988; Cowan and 

 Shaw, 1988). Larval croaker then move toward shore 

 and may be transported hundreds of kilometers be- 

 fore entering estuarine nursery grounds (Cowan and 

 Shaw, 1988; Norcross, 1991). In Texas, recmitment of 

 croaker peaks in November (Rooker et al., 1998). It is 

 not clear whether variability in abundance of juvenile 

 croaker is the result of variability in lai-val supply or 

 differential postsettlement growth and mortality. 



The delivery of larval croaker recruits to estua- 

 rine nursery habitats is dependent on large-scale 

 oceanographic processes (Cowan and Shaw, 1988). 

 Once fish arrive at estuaries, delivery into suitable 

 habitats is dependent on currents and tidal processes 

 (Norcross, 1991 ). As fish that are competent to settle 

 approach nursery grounds they have the opportu- 

 nity to choose specific microhabitats. In this paper 

 we examined patterns of microhabitat preference and 

 use by newly settled croaker, as well as the conse- 

 quences of microhabitat associations. Specifically we 

 asked 1) Do croaker have specific microhabitat pref- 

 erences and are these preferences reflected in pat- 

 terns of abundance in the field? 2) Does food supply 

 limit the number or gi'owth rates of croaker recruits 

 in different habitats? 3) Does predation determine 

 the number of recruits in different habitats? 



Methods 



Habitat use by newly recruited croaker 



To determine what habitats newly recruited Atlan- 

 tic croaker use, we conducted a field sui-vey during 

 November 1996 at Christmas Bay (29"03'N, 95'10'W), 

 near Galveston, TX. Christmas Bay is a shallow estu- 

 ary and contains the most easterly well-developed 



seagrass bed in Texas. A detailed description of this 

 site can be found in Thomas et al. (1990). The 

 seagi-ass bed is dominated hy Halodule wrightii with 

 an average density of 10,469 shoots/m^ (SE=461). An 

 epibenthic sled was used to quantify fish abundance 

 in three habitats: bare sand, seagrass meadow, and 

 marsh edge. We defined marsh edge habitat as the 

 subtidal substrata directly adjacent to a Spartina 

 alterniflora marsh. The sled consisted of a 0.66 m x 

 0.5 m opening fitted with a3-mlongnet(l-mmmesh) 

 with a removable codend. Habitats were sampled by 

 placing the sled on the substratum, extending a 15-m 

 rope in a semicircular fashion (to avoid disturbing 

 sampling area) and pulling the sled through a lO-m^ 

 area. Each habitat was sampled four times at two 

 different sites, resulting in eight samples per habi- 

 tat type. Differences in croaker density were exam- 

 ined with a two-way analysis of variance with both 

 site and habitat type as fixed effects. In this and sub- 

 sequent analyses, if we failed to reject the null hy- 

 pothesis of no difference in croaker abundance be- 

 tween habitats, then power analysis was performed. 

 If statistical power was low, we calculated the num- 

 ber of replicates required to achieve sufficient power 

 to accept the null hypothesis. 



To examine habitat preference we performed choice 

 experiments in laboratory mesocosms. Six 117-L 

 mesocosms were constructed from round circular 

 plastic tanks (41.3 cm diameter x 60 cm). The 

 mesocosms were filled with 5 cm of sand, a plastic 

 mesh screen was placed on top of the sand, and an 

 additional 5 cm of sand was placed over the mesh. 

 Each tank was filled with filtered seawater and main- 

 tained at ambient light and temperatures. We divided 

 mesocosms in half, with each half randomly receiv- 

 ing a sand or grass habitat. Sand habitat was the 

 sand bottom described above. To construct seagrass 

 habitats, cores of seagrass were randomly collected 

 from the field and brought to the laboratory where 

 they were washed and dipped in fresh water. After 

 leaves were wiped to remove any epiphytic growth, 

 the cores were planted in each mesocosm. 



One croaker ( 15-20 mm SL) was introduced to the 

 center of each mesocosm and monitored for any ab- 

 normal behavior for 24 h. After the initial acclima- 

 tion period, the location of each croaker was visually 

 determined hourly for ten consecutive hours. Visual 

 observations were performed by a single observer 

 peering into the mesocosm, without disturbing the 

 fish. This was repeated for six mesocosms over two 

 days for a total of 12 mesocosm observations. New 

 fish were used for each trial. Percent occurrence in 

 each habitat was determined for all twelve trials. A 

 one-way Ntest determined if percent occurrence in 

 seagi'ass was different from 509f . 



