The widespread evolution of tube-like anthers releasing pollen from apical pores is associated with buzz pollination, in which bees vibrate flowers to remove pollen. The mechanical connection among anthers in buzz-pollinated species varies from loosely held conformations, to anthers tightly held together with trichomes or bio-adhesives forming a functionally joined conical structure (anther cone). Joined anther cones in buzz-pollinated species have evolved independently across plant families and via different genetic mechanisms, yet their functional significance remains mostly untested. We used experimental manipulations to compare vibrational and functional (pollen release) consequences of joined anther cones in three buzz-pollinated species of Solanum (Solanaceae). We applied bee-like vibrations to focal anthers in flowers with (“joined”) and without (“free”) experimentally created joined anther cones, and characterised vibrations transmitted to other anthers and the amount of pollen released. We found that joined anther architectures cause non-focal anthers to vibrate at higher amplitudes than free architectures. Moreover, in the two species with naturally loosely held anthers, anther fusion increases pollen release, while in the species with a free but naturally compact architecture it does not. We discuss hypotheses for the adaptive significance of the convergent evolution of joined anther cones.

To measure the velocity of the vibrations applied and measured, we used a Doppler laser vibrometer (PDV-100, Polytec Ltd, Coventry, UK) set to 500mm/s maximum velocity and a Low Pass Filter at 22kHz. . The force applied by the shaker was simultaneously measured using the miniature force sensor. The signals of both the laser vibrometer and the force sensor were simultaneously acquired using a two-channel NI9250 Sound and Vibration module (NI Corporation (UK) Ltd, Newbury, UK) and a USB-powered data acquisition module (cDAQ-9171, NI). For each recorded vibration, we estimated the dominant frequency (Hz) and the RMS amplitude for both the velocity (mm/s) measured in the distal anther, and the force (mN) measured in the proximate anther. To estimate the number of pollen grains released after experimental buzzes, we used a particle counter (Multisizer 4e Coulter Counter, Indianapolis, USA).