Griffiths et al.: Fish recolonization in temperate Australian rockpools 



635 



Tasman Sea 



•^Shellharbniir 

 Bass Point 



5 km 



Figure 1 



Map illustrating the four sampling sites at Bass Point and the location of the study loca- 

 tion in the Illawarra region. New South Wales, Australia. 



tidal rockpool fish assemblages has received considerable 

 attention in many countries of the world (Bussing, 1972; 

 Matson et al., 1986; Yoshiyama et al., 1986; Prochazka 

 and Griffiths, 1992; Lardner et al., 1993; Prochazka, 

 1996; Faria and Almada, 1999; Silberschneider and 

 Booth, 2001). Such studies have identified patterns in 

 the rates of recovery, variation in species and size com- 

 position of recolonizing fish assemblages (Polivka and 

 Chotkowski, 1998; Beckley, 2000), and homing abilities 

 of many intertidal fishes (Green 1971; Yoshiyama et al., 

 1992; Griffiths. 2003b). 



Rockpools can be regarded as "island' habitats (Under- 

 wood and Skilleter, 1996) among an inhospitable rocky 

 landscape. Therefore, there is probably a balance be- 

 tween immigration (recruitment and relocation) and 

 emigration (mortality) of fishes after a disturbance, 

 sensu the equilibrium theory of island biogeography 

 (MacArthur and Wilson, 1967). After a period of time, 

 the number of species and individuals in a defaunated 

 rockpool can be expected to reach an asymptote when a 

 carrying capacity is reached. It is difficult to generalize 

 about recolonization rates of rockpools by fishes from 

 the current literature mainly owing to the diversity of 

 methods used, their differing effectiveness in sampling 

 fish, and the varying intensity of the sampling regime. 

 For example, most studies have used only small sample 

 sizes (<10 pools) and have sampled at a range of time 

 intervals from days (Mistry et al., 1989; Matson et al., 

 1986; Polivka and Chotkowski, 1998) to years (Thomson 

 and Lehner, 1976; Lardner et al., 1993; Mahon and 

 Mahon, 1994). A second problem in measuring and 

 comparing fish recolonization patterns between stud- 

 ies is that many researchers have sampled fish using 

 an anesthetic (Mahon and Mahon. 1994; Pfister, 1995, 

 1997) or ichthyocide (Beckley, 1985a, 1985b, 2000; Wil- 



lis and Roberts, 1996; Silberschneider and Booth, 2001), 

 which may affect subsequent catches (Yoshiyama et 

 al., 1986) and possibly result in fish assemblages never 

 reaching preperturbation conditions (see Mok and Wen, 

 1985; Lockett, 1998). 



Nonetheless, recolonization of rockpools by fishes is 

 generally a rapid process, beginning within days, or 

 even hours, after defaunation (Collette, 1986). and com- 

 plete recolonization to preperturbation levels can take a 

 few weeks (Collette, 1986; Faria and Almada. 1999) to 

 several months (Mok and Wen. 1985; Willis and Rob- 

 erts, 1996; Polivka and Chotkowski, 1998). 



The aims of this study were to quantitatively deter- 

 mine 1) the period required for intertidal rockpools to 

 recover to preperturbation levels, 2) the fish species 

 (permanent residents, opportunist, or transients) respon- 

 sible for recolonizing rockpools, 3) whether recoloniza- 

 tion patterns differ between the four sites at Bass Point 

 and between the times of year when defaunation took 

 place, and 4) whether fish comprise different life-history 

 stages before and after a disturbance (sampling) — by 

 examination of length-frequency distributions. 



Methods 



Study site and experimental design 



Spatial and temporal variation in fish recolonization pat- 

 terns were investigated in three separate studies under- 

 taken along the north- and south-facing rocky platforms 

 at Bass Point (34°58'S, 150°93'E), New South Wales, 

 Australia (Fig. 1). Bass Point is a large rocky headland 

 that extends approximately 3 km into the Tasman Sea. 

 Two short-term recolonization studies (each around 5 



