Use of a lure in visual census significantly improves probability of detecting wait-ambushing and fast cruising predatory fish

A lure assisted, visual census, belt transect method has been developed and compared with the traditional visual census. Presentation of a lure was predicted to improve visibility of fish that are difficult to detect or identify. These include species hidden within three dimensional structures, buried in or camouflaged with the substratum, and fast swimming. A snorkeler presenting an artificial lure near the seabed along 3-m transects recorded all fish visiting the lure or present near the lure's path. Totals of 1527 lure-assisted transects and 1292 lure-absent transects intersecting a diversity of benthic habitats were carried out at thirty sites spanning the Croatian Adriatic coast in spring 2010. More than half the observed fish visited the lure, and lure presentations increased the probability of census for 76% of the 58 observed taxa. The overall significantly higher fish density at lure-assisted transects than at lure-absent transects was also significant within all common habitat types and for the entire investigated depth range (0.25–9 m). The positive effect of the lure on estimates of density was significant for 33% of taxa, including: ambush predators (Trachinus draco), fast moving cruise-chase predators (Sparidae), predators chasing prey from a hidden waiting position (Coris julis and Serranidae) or from a position of camouflaging against bare sand (Gobiidae). Adding lure presentation to regular visual census is logistically simple and flexible in the detail of its design and application.