Summary
It has long been recognized that many hoverfly species (Diptera: Syrphidae) mimic the morphological appearance of defended Hymenoptera, such as wasps and bees. However, it has also been repeatedly suggested that some mimetic hoverflies respond with sounds on attack that resemble the warning or startle sounds of their hymenopteran models. In this study, we set out to quantitatively compare the spectral characteristics of the sounds produced by a range of nonmimetic flies, wasps, bumblebees, honeybees, and their hoverfly mimics when they were artificially attacked. The sounds made by wasps and honeybees after simulated attacks were statistically distinguishable from their hoverfly mimics. Bumblebee models of their hoverfly mimics share some similarities in the sound they produce on attack, but they were no closer acoustically to their model than a range of other hoverfly species that morphologically resemble other models. All the mimetic hoverflies tested in this study tended to sound similar to one another, regardless of the model they resemble morphologically. Overall, we found little evidence that mimetic hoverflies sound like their hymenopteran models on attack, and we question whether acoustic mimicry has evolved in this complex.
Methodology
Species investigated
The sounds produced by a total of 162 individuals from 14 different insect species were measured, analyzed, and compared (Table 1). The majority of the specimens were collected between 28 June and 15 September 2005 at the Queens University Biological Station, Ontario, Canada (44°34′N, 79°15′W). The exceptions were bumblebees, Bombus impatiens Cresson, which were reared in the laboratory, and the wasp species, Vespula germanica Fabricius, and honeybees, Apis mellifera Linnaeus, which were collected from a residential area in Ottawa, Canada. Honeybee mimics, Eristalis transversa Wiedemann and Eristalis arbustorum Linnaeus, were collected from Ottawa's Central Experimental Flower Garden. Specimens were collected using hand nets and on capture were placed in wooden and mesh cages measuring 30 × 30 × 50 cm. Insects were transported back to the laboratory where acoustic recordings were made. Insects were typically tested within 2 h of capture but never more than 4 h after capture.
Behaviors associated with sound production
We observed the mechanism of sound production on simulated attack by filming the response of a typical bumblebee, Bombus sp., and a wasp mimic hoverfly, S. longicornis, using a high-speed digital camera (2000 frames per second, Fastcam PCI HSI High-speed Imaging Incorporated, Ontario, Canada) and simultaneously recording sounds using an Earthworks QTC40 microphone. Videos and accompanying sounds were digitized and analyzed using Midas software (Xcitex, Cambridge, MA).