Uneven distribution of grazing negatively impacts rangelands through o
ver- and under utilization of resources. The goal of this study was to
quantify the role of experience on search pattern and foraging effici
ency of cattle. Steers (Bos taurus x B. indicus) were exposed once dai
ly during 15-20 min. sessions to 3 food-distribution treatments: VR (v
ariable-random, food locations were changed randomly and daily), CR (c
onstant-random, food locations were randomly set at the beginning and
remained the same throughout the experiment), and CC (constant-clumped
, food locations were constant and clumped in groups of 5). Pelleted f
eed was available in 20 out of 64 feeders arranged in 8 rows and 8 col
umns, with neighboring locations 5 m apart. Encounter rate of food loc
ations was partitioned into search speed, total number of visits per u
nit distance walked, ratio of different (not previously visited within
the session) locations to total visits (including revisits), and rati
o of food locations to different locations visited. Intake rate increa
sed (P < 0.01) as animals gained experience, but more slowly in variab
le-random than constant-clumped and constant-random. Residence time at
food locations declined (P < 0.01) with increasing experience. Intake
rate was negatively affected (P < 0.01) by search time per food locat
ion, which in turn was determined by the steers' ability to remember f
ood locations. Steers in constant random and constant clumped used lon
g-term spatial memory to return to food locations, and ignored areas w
here no food was found (P < 0.01). Conversely, steers in variable rand
om used a strategy based on avoidance of locations already visited wit
hin sessions. Thus, in constant random and constant clumped food searc
h was more efficient (P < 0.01) and concentrated in certain areas, whe
reas in variable random it was less efficient and more evenly distribu
ted over the whole area The results of this study suggest that impedin
g spatial memory could improve grazing patterns.