Optimal foraging theory assumes that individuals maximizing a given en
ergy currency maximize their fitness. Though models with different cur
rencies (and assumptions) have been successful at describing forager b
ehavior, discriminating among currencies has proven difficult because
models optimizing different currencies often make similar predictions.
A field experiment was designed to use flight speeds of Black Terns (
Chlidonias nigra) feeding their young to test predictions from optimal
foraging models that maximize (1) the net energy gained per unit of e
nergy expended (efficiency; EFF), (2) the net rate of energy intake (N
REI), and (3) the daily delivery rate (DR), respectively. Manipulating
the distance an adult had to fly between a feeding enclosure and its
nest enabled us to discriminate among the three currencies. Predicted
speeds of the EFF model best matched the observed flight speeds. Maxim
izing efficiency resulted in the lowest delivery rate to the nest, but
it was also the only currency that predicted adult daily energy expen
ditures within the theoretical limit (DEE(max)) estimated for Black Te
rns (206 kJ/d; Kirkwood 1983). The feeding enclosure represented an un
usually profitable food source for the terns, and this may have result
ed in artificially high predictions for flight speed, particularly for
the NREI and DR currencies. We therefore made another set of predicti
ons from the three models using estimates of foraging success from par
ents who had never foraged from the feeding enclosure, and compared th
ese predictions with the measured airspeeds. The EFF model was apain t
he best predictor of flight speeds even though none of the three model
s predicted energy expenditures in excess of DEE(max). Though the DR m
odel resulted in the highest total daily energy delivered to the nest,
this represented only a small gain relative to the NREI and EFF model
s. Black Tem parents may maximize EFF because this currency is least l
ikely to result in daily energy expenditures that exceed their metabol
ic limit.