Connective appendages in Huberia bradeana (Melastomataceae) affect pollen release during buzz pollination

Abstract

Floral structures, such as stamen appendages, play crucial roles in pollinator attraction, pollen release dynamics and, ultimately, the reproductive success of plants. The pollen-rewarding, bee buzz-pollinated flowers of Melastomataceae often bear conspicuous staminal appendages. Surprisingly, their functional role in the pollination process remains largely unclear. We use Huberia bradeana Bochorny & R. Goldenb. (Melastomataceae) with conspicuously elongated, twisted stamen appendages to investigate their functional role in the pollination process. We studied the effect of stamen appendages on pollinator behaviour and reproductive success by comparing manipulated flowers (appendages removed) with unmanipulated flowers. To assess bee pollinator behaviour, we measured three properties of buzzes (vibrations) produced by bees on Huberia flowers: frequency, duration and number of buzzes per flower visit. We measured male and female reproductive success by monitoring pollen release and deposition after single bee visits. Finally, we used artificial vibrations and laser vibrometry to assess how flower vibrational properties change with the removal of stamen appendages. Our results show that the absence of staminal appendages does not modify bee buzzing behaviour. Pollen release was higher in unmanipulated flowers, but stigmatic pollen loads differ only marginally between the two treatments. We also detected lower vibration amplitudes in intact flowers as compared to manipulated flowers in artificial vibration experiments. The presence of connective appendages are crucial in transmitting vibrations and assuring optimal pollen release. Therefore, we propose that the high diversity of colours, shapes and sizes of connective appendages in buzz-pollinated flowers may have evolved by selection through male fitness.

Keywords: buzz pollination; carpenter bees; pollen release; sonication.

Fig. 1

Flower of Huberia bradeana Bochorny & R. Goldenb. (Melastomataceae). (A) Anthers (stamens all grouped in a zygomorphic bundle that the bee grasps during buzz pollination), style and connective with dorsal appendages; (B) Natural stamen with connective appendage; (C) Manipulated stamen without connective appendage. (Photos: Renato Goldenberg and Thuane Bochorny).

Fig. 2

Flowers of Huberia bradeana. Two treatment groups: (A) manipulated flowers, in which connective appendages were artificially removed, and (B) natural flowers without manipulation (Photo: Thuane Bochorny).

Fig. 3

Xylocopa cf. brasilionorum. (Apidae) buzz pollinating Huberia bradeana. (Photo: Vinícius L. G. Brito).

Fig. 4

Comparison of bee sonication behaviour on Huberia bradeana flowers with or without connective dorsal appendages. (A) Frequency of vibrations; (B) Duration of vibrations; (C) Number of buzzes. Box plots show minimum and maximum values on the external whiskers and the first and third quartiles on the internal whiskers. Internal line represents the median.

Fig. 5

(A) spectrogram of flower vibration during artficial sonication in flowers of Huberia bradeana with (purple) and without (green) connective dorsal appendages (dashed line indicates vibration of the paper‐tag directly on the vibrator metal rod); (B) Comparison between relative amplitude of flower vibration in each harmonic peak during artificial vibration of flowers with and without appendages.

Fig. 6

Effects of natural bee buzzes on (A) pollen released by anthers and (B) pollen receipt by stigmas of Huberia bradeana flowers with (purple) and without (green) appendages. (A) Number of pollen grains remaining in anthers before and after the first bee visit (pollen released) NA (not applicable); (B) Counts of stigma load indices after first bee visit (pollen receipt). The stigma load index was estimated by dividing the stigma in three sections. We used the following stigma load indices: 0 = 0%, 1 = 10%, 2 = 25%, 3 = 50% and 4 = 100% for section covered with pollen grains under a microscope (100× objective lens). *P < 0.05, ***P < 0.01.