74
P.P. Molloy et al.
Fish may also spawn in aggregations if certain sites are particularly safe for
adults, rather than for their eggs (Johannes 1978 ; Shapiro et al. 1988 ) . This could
happen if there are particularly few predators in an area, or if the physical structure
of a site protects spawners as they release their gametes, often at the apex of an
upward spawning rush during which they are most vulnerable to predation (Sancho
et al. 2000 ) . Several studies have noted that aggregations occur at sites that have
more relief compared to nearby non-spawning areas. For example, red hind aggregate to spawn around highly complex and rare habitats (Beets and Friedlander
1998 ) , benthic-spawning brown puller damselfi sh select spawning sites that are
signifi cantly more rugose than neighbouring non-spawning sites, and Sancho et al.
( 2000 ) found that several group-spawning species aggregated at particularly complex
sites to spawn. In all these examples, the authors hypothesise that reef complexity
serves to reduce predation risk (Beets and Friedlander 1998 ; Sancho et al. 2000 ;
Gladstone 2007a ) .
The ultimate experiment to test whether specifi c physical features are sought by
spawning fi sh would be to remove all fi sh from an area where they aggregate, replace
them with site-naïve fi sh, and see whether these new fi sh aggregate to spawn at the
same location as the original population. This extraordinary experiment was conducted by Robert Warner using the wrasse as a model system (Warner 1988 ) .
Bluehead wrasse spawn either in pairs or groups, and the groups form and spawn
daily at predictable times and locations (Figs. 12.38, 12.39). Some mating sites
have remained in use for over 12 years without changing location (Warner 1988 ) .
Experimental replacement of entire populations from small patch reefs led to the
use of sites for spawning that differed from those used by the original populations.
Thus mating site locations in bluehead wrasse are not solely the result of individual
assessment of current resource quality, but rather represent culturally transmitted
traditions. Further experiments in which either the male or the female portion of
each population was replaced demonstrated that transmission of spatial traditions
occurred via females (Warner 1990 ) . Males simply go where females aggregate
to spawn.
At this point, it is unknown whether these results can be extrapolated to other
species that spawn in aggregations. It would be extremely diffi cult to carry out such
an experiment with larger species, such as groupers or snappers. Unfortunately, the
fi rst part of this experiment is effectively taking place through overfi shing of spawning
aggregations in many regions of the world (Sala et al. 2001 ; Sadovy and Domeier
2005a ) . Should protection and management result in recovery of populations to
former levels, it will be of interest to observe whether spawning aggregations
re-form in locations from which they were extirpated, which would support the idea
that relatively static, benefi cial environmental features exist at these sites. By contrast,
the formation of aggregations at novel locations would indicate that spawning sites
are not chosen on the basis of benefi cial features, that many places have benefi cial
features but not all are used, or that the physical features that prompted site choice
by previous generations are no longer present.
P.P. Molloy et al.
Fish may also spawn in aggregations if certain sites are particularly safe for
adults, rather than for their eggs (Johannes 1978 ; Shapiro et al. 1988 ) . This could
happen if there are particularly few predators in an area, or if the physical structure
of a site protects spawners as they release their gametes, often at the apex of an
upward spawning rush during which they are most vulnerable to predation (Sancho
et al. 2000 ) . Several studies have noted that aggregations occur at sites that have
more relief compared to nearby non-spawning areas. For example, red hind aggregate to spawn around highly complex and rare habitats (Beets and Friedlander
1998 ) , benthic-spawning brown puller damselfi sh select spawning sites that are
signifi cantly more rugose than neighbouring non-spawning sites, and Sancho et al.
( 2000 ) found that several group-spawning species aggregated at particularly complex
sites to spawn. In all these examples, the authors hypothesise that reef complexity
serves to reduce predation risk (Beets and Friedlander 1998 ; Sancho et al. 2000 ;
Gladstone 2007a ) .
The ultimate experiment to test whether specifi c physical features are sought by
spawning fi sh would be to remove all fi sh from an area where they aggregate, replace
them with site-naïve fi sh, and see whether these new fi sh aggregate to spawn at the
same location as the original population. This extraordinary experiment was conducted by Robert Warner using the wrasse as a model system (Warner 1988 ) .
Bluehead wrasse spawn either in pairs or groups, and the groups form and spawn
daily at predictable times and locations (Figs. 12.38, 12.39). Some mating sites
have remained in use for over 12 years without changing location (Warner 1988 ) .
Experimental replacement of entire populations from small patch reefs led to the
use of sites for spawning that differed from those used by the original populations.
Thus mating site locations in bluehead wrasse are not solely the result of individual
assessment of current resource quality, but rather represent culturally transmitted
traditions. Further experiments in which either the male or the female portion of
each population was replaced demonstrated that transmission of spatial traditions
occurred via females (Warner 1990 ) . Males simply go where females aggregate
to spawn.
At this point, it is unknown whether these results can be extrapolated to other
species that spawn in aggregations. It would be extremely diffi cult to carry out such
an experiment with larger species, such as groupers or snappers. Unfortunately, the
fi rst part of this experiment is effectively taking place through overfi shing of spawning
aggregations in many regions of the world (Sala et al. 2001 ; Sadovy and Domeier
2005a ) . Should protection and management result in recovery of populations to
former levels, it will be of interest to observe whether spawning aggregations
re-form in locations from which they were extirpated, which would support the idea
that relatively static, benefi cial environmental features exist at these sites. By contrast,
the formation of aggregations at novel locations would indicate that spawning sites
are not chosen on the basis of benefi cial features, that many places have benefi cial
features but not all are used, or that the physical features that prompted site choice
by previous generations are no longer present.