J. Avian Biol. 39: 479483, 2008 doi: 10.1111/j.2008.0908-8857.04394.x # 2008 The Authors. J. Compilation # 2008 J. Avian Biol. Received 4 October 2007, accepted 9 June 2008 Sugar preferences of nectar feeding birds  a comparison of experimental techniques Mark Brown, Colleen T. Downs and Steven D. Johnson M. Brown (correspondence), C. T. Downs and S. D. Johnson, School of Biological & Conservation Sciences, University of KwaZulu-Natal, Private Bag X01, Scottsville, 3201, South Africa. E-mail: brownma@ukzn.ac.za Experiments to determine sugar preferences of nectarivorous animals have been conducted using a wide variety of experimental procedures, all of which aim at ensuring that the solutions offered in choices are ‘‘equivalent’’. Each method used historically has controlled for a particular variable, such as number of molecules in solution, weight of sugar in solution, or amount of energy in solution, depending on what question the researchers have tried to answer. Biologists interpreting these results in terms of bird sugar preference have seldom taken these differences into account. The consequences of using different experimental procedures for sugar preferences exhibited by a nectarivorous bird, the malachite sunbird Nectarinia famosa, were examined using paired sucrose and hexose sugar solutions made up to be either equimolar, equiweight or equicaloric. We found the effect of methodology on bird sugar preference to be quite distinct, especially at low concentrations, where malachite sunbirds showed either sucrose preference, no preference, or hexose preference, depending on the method used. This study highlights the need for researchers to consider methodology when interpreting, or comparing among, results from previous studies. It has often been claimed that nectarivorous animals, such as birds, bats and insects, exhibit preferences for particular sugars in choice tests (Martinez del Rio 1990, Downs 1997a, 2000, Jackson et al. 1998a, b, Johnson et al. 1999). However, the results of choice tests may be highly dependent on experimental procedure. Previous tests have involved solutions that, in terms of sugars, are equimolar (Downs 1997a, b, 2000), equivalent by weight (hereafter termed equiweight; Lotz and Nicolson 1996, Jackson et al. 1998a,b, Johnson et al. 1999, Blem et al. 2000) or equicaloric (Fleming et al. 2004). Choice tests using equimolar solutions (Downs 1997a,b) have standardised solutions by offering equal numbers of molecules per solution, but have the obvious problem that a given sucrose solution has approximately twice the energy of an equimolar hexose solution (Downs 1997a, Stiles 1976). Choice tests using solutions that are equivalent in terms of sugar weight (Stiles 1976, Martinez del Rio and Karasov 1990) suffer a similar problem because of the c. 5% higher energy value for sucrose over the equivalent weight of hexose sugar (Bumstead 1980, Fleming et al. 2004). For these reasons, recent authors have advocated that choice tests be conducted between solutions that are potentially equivalent in energetic terms for the animal (assuming complete or similar digestion efficiency of the two choices), since the results of previous studies suggest that energy and concentration appear to be the most important features determining preference in specialist nectarivores (Fleming et al. 2004). None of these experimental procedures are ideal from a botanical perspective (indeed, it could be argued that experiments designed to explain why nectar sugar composi- tion varies among plants should actually involve solutions that are equivalent in terms of the energy invested into nectar by plants, as this would be the most biologically realistic null model), however from an ornithocentric view it becomes necessary to standardise various variables. Standardising energy makes sense in terms of optimal foraging theory (Hixon and Carpenter 1988) and optimal diet theory (Schaefer et al. 2003), which has shown that nectarivorous birds, for the most part, discriminate among food items according to their energetic value. Using equicaloric solutions, Fleming et al. (2004) found that broadtailed hummingbirds Selasphorus platycercus and whitebellied sunbirds Nectarinia talatala showed very little discrimination between sucrose and hexose solutions at most concentrations, except at 0.1 mol/l sucrose equivalent (SE) concentrations where the sunbirds preferred hexose solutions. These results seem intuitive, and were predicted by Martinez del Rio and Karasov’s (1990) hummingbird model for situations where hexose hydrolysis is limiting. Unless there is an underlying physiological reason (Marti- nez del Rio and Karasov 1990, Lotz and Schondube 2006), or a taste issue (Mata and Bosque 2004), birds should show no preference when presented with equicaloric solutions at most concentrations (Martinez del Rio and Karasov 1990). Based on their results, Fleming et al. (2004c) called for renewed research on specialist nectarivores using equicaloric 479