How Nectar-Feeding Bats Localize their Food: Echolocation Behavior of Leptonycteris yerbabuenae Approaching Cactus Flowers

Abstract

Nectar-feeding bats show morphological, physiological, and behavioral adaptations for feeding on nectar. How they find and localize flowers is still poorly understood. While scent cues alone allow no precise localization of a floral target, the spatial properties of flower echoes are very precise and could play a major role, particularly at close range. The aim of this study is to understand the role of echolocation for classification and localization of flowers. We compared the approach behavior of Leptonycteris yerbabuenae to flowers of a columnar cactus, Pachycereus pringlei, to that to an acrylic hollow hemisphere that is acoustically conspicuous to bats, but has different acoustic properties and, contrary to the cactus flower, present no scent. For recording the flight and echolocation behaviour we used two infrared video cameras under stroboscopic illumination synchronized with ultrasound recordings. During search flights all individuals identified both targets as a possible food source and initiated an approach flight; however, they visited only the cactus flower. In experiments with the acrylic hemisphere bats aborted the approach at ca. 40-50 cm. In the last instant before the flower visit the bats emitted a long terminal group of 10-20 calls. This is the first report of this behaviour for a nectar-feeding bat. Our findings suggest that L. yerbabuenae use echolocation for classification and localization of cactus flowers and that the echo-acoustic characteristics of the flower guide the bats directly to the flower opening.

Fig 1. Experimental set up.

Setup used during the flight cage experiments with both targets. We worked with one target at a time, respectively, each fixed to a cactus branch. The flight and echolocation behavior was recorded with an ultrasound microphone and two synchronized video cameras supported by stroboscopic light. Flight cage dimensions (4m x 4m x 3m).

Fig 2. 3D Flight paths reconstructions of a bat approaching the cactus flower and the hemisphere.

Side (1), top (2) and 3D (3) view of three exemplary flight path reconstructions from one bat approaching cactus flower (A) and hemisphere (B). Small dots in the flight path show the position of the bat while emitting echolocation calls. Larger black and green dots represent the position of the microphone and the center of the cactus, respectively, and the pink star shows the position of the target.

Fig 3

Examples of flight and echolocation behavior (spectrograms) of a single bat approaching the cactus flower (A) and the hemisphere (B). Small dots in the flight path show the position of the bat while emitting the echolocation calls. Larger black and green dots represent the position of the microphone and the center of the cactus, respectively, and the pink star shows the position of the target. The red asterisk indicates the first call that appears in the 3D flight path reconstruction. We could identify different phases: search (S), approach (Ap) and before inserting the snout into the cactus flower the bats broadcast an exceptionally long terminal group with increased call rate (TG). In contrast, when approaching the hemisphere, the last group (LG) emitted by the bats before giving up and flying away did not differ from other groups emitted during the entire approach. Spectrogram settings: FFT length: 512, window: Hamming and overlap: 50%.

Fig 4

Pulse duration (A and B), pulse interval (D and C) and number of calls per group (E and F) of all calls recorded from bats approaching cactus flower and hemisphere. The line represents the beginning of the overlap zone and all calls located to the right of this line overlap with their echo. Pulse interval data are divided in inter-group interval (GI, higher values) and inter-pulse interval (IGI, smaller values). E and F represent the number of calls emitted per group while the bats approach to the targets; the plotted distance is the first call of each group.

Fig 5. Echolocation call parameters during the approach to the cactus flower and the hemisphere.

Pulse duration (A), initial frequency (B), pulse interval (C), end frequency (D), peak amplitude (E) and bandwidth (F) per distance interval (mean values ± SD) during the approach to the flower and hemisphere targets. The peak amplitude was corrected for the expected decrease of 6 dB per halving of the distance and by angle flight (θ, φ) with respect to the target.