Behavioural measurements enable to establish visual abilities of the system as a whole. The measurements should be concentrated on a specific ability, like acuity, colour discrimination, threshold value, polarization sensitivity, etc. One can also measure natural behaviour that is visually generated, like food detection and gender recognition but one must be sure that no other sense organs are involved.
In case of the haplochromine cichlids from Lake Victoria the used behavioural measurements were operant conditioning (acuity), optomotor nystagmus (thresholds) and particle detection at low light levels (acuity and sensitivity).
A review on behavioural studies of fish vision is found in Douglas & Hawryshyn (1990) showing the relation between results and used method.

Operant conditioning
Operant conditioning as a tool to measure visual resolution in fish has been applied on several occasions, usually to measure resolution as a function of background illumination (Brunner, 1934; Yamanouchi, 1957; Nakamura, 1968; Hodos & Yolen, 1976; Penzlin & Stubbe, 1977). Measurements of visual resolution as a function of ontogenetic development has been carried out by Baerends et al. (1960) and Clark (1981). The lowest behaviourally measured values of the minimum resolvable angle in cichlids came to 5,8 min. of arc in Aequidens portalegrensis (Baerends et al., 1960) to 7,0 min. of arc in Hemichromis bimaculatus (Baburina et al., 1968).
Operant conditioning as a tool has also been used in fish to measure colour processing (e.g. Risner et al., 2006) and polarization sensitivity (e.g. Mussi et al., 2005). In haplochromine cichlids it was also used to measure spectral sensitivity in H. burtoni (Allen & Fernald, 1985).

Optomotor response
Throughout the animal kingdom, organisms with eyes tend to fix a moving pattern on their retina. This visual reflex is called optomotor response (movement of head or entire body) or optokinetic nystagmus (eye movement) and can be used as a tool to measure visual abilities. The optokinetic fixation may involve the entire surrounding which occurs in animals without a fovea (e.g. rabbits; Collewijn, 1977) or the fixation of a small object on the fovea centralis (e.g. man; Cheng & Outerbridge, 1975). When confronted with a moving line-pattern, fishes tend to swim along or, at least, show rather abrupt eye-movements.
Optomotor response has been used in fish to measure scotopic thresholds (Kawamura et al., 1984) and critical flicker frequencies (Crozier & Wolf, 1940). Optomotor response was more recently used in cichlids to measure spectral sensitivity (e.g. Maan et al., 2006; Kröger et al., 2003). Visual acuity in cichlids have also been measured using optomotor and optokinetic responses with resulting lowest values of the minimum resolvable angle of more than 50 min. of arc in Xenotilapia leptura (Dobberfuhl et al., 2005) which indicates a rather poor acuity in this species. It is a less accurate tool to measure resolution as the distance between the eye and the pattern is not constant. For comparative purpose, however, the study of Dobberfuhl et al. (2005) is quite interesting. 

Visual feeding behaviour
There is a positive correlation between intraspecific eye-size and visual prey detection (Hairston et al., 1982; Li et al., 1985; Meyer, 1986). In other words, as fish grow they become more successful in capturing a prey. Yet, it is not self-evident that the improved visual behaviour is a direct result of increased spatial resolution, as acuity is also a function of contrast sensitivity and temporal resolution. Beside, increasing eye-size may also effect the sensitivity thresholds. Some authors consider the developmental changes of retinal structure a necessity to preserve constant visual capabilities rather than an actual improvement of visual abilities (Fernald, 1988). Still, possible changes of the visual behaviour and/or the photic environment during the natural development of a species may conceal its improved visual potentials.