Locating prey in a vast ocean. (A) The bird travels to a productive area using navigation mechanisms that have not yet been defined. Upon arriving, the bird may recognize the productive area by a change in the odor landscape, depicted here as a change in color. This change in the way the ocean smells triggers the bird to begin area-restricted search (ARS). (B) ARS may involve tracking odor plumes upwind to a prey patch or item (P), in combination with visually monitoring the foraging activity of other birds (Nevitt, 2000).

An olfactory feature. A bathymetric feature (in this case, a seamount) where phytoplankton accumulate leads to a change in the odor landscape over the seamount that a bird might recognize upon arrival [adapted from Nevitt (Nevitt, 2000)].

Dimethyl sulfide (DMS) emissions increase when phytoplankton are grazed by zooplankton. Its precursor, dimethylsulfoniopropionate (DMSP) is a metabolite of phytoplankton, and an excretion product of zooplankton and other predators (Dacey and Wakeham, 1986; Hill and Dacey, 2006).

Burrow nesting is correlated to behavioral sensitivity to DMS. (A) Southern giant petrel (Macronectes giganteus) brooding a chick on a surface nest. Photograph courtesy of R. Van Buskirk. (B) White-chinned petrel (Procellaria aequinoctialis) incubating an egg in its dark, underground burrow. Photograph provided by G.A.N. (C) Marginal probability reconstructions of ancestral states for nesting habit and DMS responsiveness using the topography of Nunn and Stanley (Nunn and Stanley, 1998). Each tree uses a global analysis and a Markov k-state one-parameter model (Pagel, 1999). Areas of pie charts indicate relative support for each ancestral state. Positive DMS attraction indicates those species that showed a statistically significant behavioral response to DMS in experimental trials. Log likelihood scores for the trees are –6.869894700 for nesting habit and 7.607159452 for DMS behavioral sensitivity. All reconstructions are significant for the ancestral state occupying the majority of each node, except for two nodes labeled as not significant (NS) [reproduced with permission from Van Buskirk and Nevitt (Van Buskirk and Nevitt, 2008)].

Sketch illustrating the general phylogenetic relationships among the procellariiforms (see Kennedy and Page, 2002). DP indicates diving petrels. We have proposed that the procellariiforms may have arisen from a burrow-nesting lineage, with surface nesting arising independently in two groups (shown in red) as a derived condition. In the context of foraging, this change could have lessened the degree to which surface nesters rely on olfactory tracking to locate ephemeral prey patches (DMS–), while at the same time promoting a cascade of changes that ultimately led to a multi-modal foraging strategy, and the exploitation of distant and more consistently productive areas (Van Buskirk and Nevitt, 2008).

Testing the response of blue petrel (Halobaena caerulea) chicks and fledglings to odors in the field. (A) The Porter method. A light bulb warms the chick to induce a `sleeping' state (1). Puffs of odor are presented to the chick (2) and the reaction is scored following the convention of Porter et al. (Porter et al., 1999) [adapted from Cunningham et al. (Cunningham et al., 2003)]. (B) Responses to DMS using the Porter method. Average response scores to DMS (10 pmol l^–1, filled bar) and control (open bar, distilled water) solutions. Values are means + s.e.m. Differences are significant (Wilcoxon signed-rank test, P<0.03, N=22 chicks, 15–20 days old). (C) A blue petrel near fledging age. (D) Blue petrel fledglings orient towards DMS (10^–12 mol l^–1) in a Y-maze. The histogram shows the percentage of birds that chose DMS or control. NC indicates no-choice. P<0.01, binomial test, N=24. Photograph provided by G.A.N. [data from Bonadonna et al. (Bonadonna et al., 2006)].