In search of unusual interactions. A commentary on: ‘Pollen adaptation to ant pollination: a case study from the Proteaceae’

Abstract

This article comments on:

Nicola Delnevo, Eddie J van Etten, Nicola Clemente, Luna Fogu, Evelina Pavarani, Margaret Byrne and William D Stock, Pollen adaptation to ant pollination: a case study from the Proteaceae, Annals of Botany, Volume 126, Issue 3, 1 September 2020, Pages 377–389, https://doi.org/10.1093/aob/mcaa058.

Whenever we think of pollination, bees, hummingbirds, butterflies, perhaps beetles and flies or even bats spring to mind – but never ants. However, ants are omnipresent in almost all terrestrial ecosystems and ant–plant associations are widely diverse and distributed throughout the world, contributing to complex ecological networks (Del-Claro et al., 2016). For this reason, conclusive evidence for ants acting as pollinators now emerging from field work (Del-Claro et al., 2019) should not come as a surprise. Indeed, as de Vega et al. (2014) point out, ‘our understanding of ant–flower systems is still in its infancy’.

In the overall context of ant–plant mutualism, ant pollination is regarded as a rare interaction, with few studies showing the role of ants as the effective pollinators (de Vega et al., 2014; Domingos-Melo et al., 2017; Del-Claro et al., 2019). On the basis of published data, the main reasons why ants are considered poor agents of cross-pollination are: (1) their small size, smaller than most of the floral reproductive structures; (2) ant self-cleaning behaviour, removing pollen before transport; (3) limited displacement, as foragers, ants cannot fly and thus only visit resources near the nest; and (4) the interference in pollen viability by antibiotics secreted by the metapleural glands and spread over the surface of the insect’s body (Hickman, 1974; Beattie, 1985; de Vega et al., 2009; Del-Claro et al., 2019; Fig. 1).

Fig. 1.

Summary of the main factors triggering or that limit ant pollination. Data from Hickman (1974), Beattie (1985), de Vega et al. (2009) and Del-Claro et al. (2019).

In this issue of Annals of BotanyDelnevo et al. (2020) further explore the roles of insect secretions and the environment in ant pollination of Western Australian plants. To evaluate the antibiotic function of ant secretions (the so-called antibiotic hypothesis; Beattie, 1985), Delnevo et al. (2020) tested the hypothesis that a ‘lack of a negative response to ants could result from either the low production of secretions by local ants or because a plant species has adapted to potentially use ants as pollen vectors by producing pollen resistant to secretions’. Data from three complementary experiments strongly indicate that ants act as secondary pollinators of Conospermum shrubs (Proteaceae) in Western Australia. In the first, floral fidelity was tested, revealing that the native bee Leioproctus conospermi (Colletidae) is the primary pollen vector and ants the secondary pollen vectors of C. undulatum. Secondly, they performed classical field exclusion experiments (control; no ants; and no flying insects; e.g. Del-Claro et al., 2019) that confirmed that ants are important complementary pollinators of C. undulatum. Thirdly, a well-designed pollen germination assay was employed to compare the germination of pollen collected from C. undulatum with that of five other plant species after contact with three species of Australian ants, as well as honeybees and, of course, a control (no contact with insects). This last set of experiments is important because it indicates coevolution between C. undulatum and ants while the other five species tested had significantly lower rates (~50 %) of pollen germination. The pollen germination in Conospermum species (C. undulatum, C. stoechadis and C. canaliculatum) was reduced by only 5–9 % after exposure to ants, similar to the effect of exposure to bees. Delnevo et al. (2020) thus conclude that the pollination systems of these plants have evolved the capacity to overcome the negative effects of ant secretions.

Lipoidal secretions of ants have been shown to penetrate pollen via a hydrophobic pathway and to compromise the function of both plasma membrane and organellar membranes (Beattie, 1985). In their study, Delnevo et al. (2020) propose that Conospermum possesses a mechanism to mitigate the osmotic shock resulting from lysis of the bilayer membrane of pollen. It is our view that, while interesting, this remains a speculation in need of verification through further careful experimentation, including chemical and genetic assays. This chemical ‘conversation’ between plant and pollinator is unlikely to be one-sided, and it is important that the roles of plant secretions, including volatiles in floral nectar, in modulating ant behaviour are also explored (Del-Claro et al. 2016 and references therein) so that a ‘holistic’ picture of this aspect of ant/plant mutualism can finally emerge.

Delnevo et al. (2020) also focus on the role of environmental background – the habitat – in ant pollination. This pollination syndrome has been reported mainly in dry environments (with dramatic differences between rainy and dry, and hot and cold seasons), with poor-quality soil and where shrubs are the predominant vegetation (e.g. de Vega et al., 2009, 2014; Del-Claro et al., 2019 and references therein). In these environments, flying insects are not abundant, and some features of ants may be well-fitted to pollination (Fig. 1). Furthermore, plants that are short, with near ground-level inflorescences, are common in these environments, featuring small, sessile flowers that produce little or no odour, where nectar is the main reward, and generating low amounts of pollen – the latter suggested to avoid ant self-grooming behaviour (‘ant-pollination syndrome’; Hickman 1974; Fig. 1). Thus the high diversity and abundance of ants and their active behaviour (rendering them capable of visiting different plants and flowers many times in the same day) may present these plants with an opportunity to recruit an unlikely pollinator at a low energetic cost.

This study of Delnevo et al. (2020) provides valuable new insights on ant pollination and highlights the need for chemical aspects of plant/ant mutualism to be investigated through a combination of ‘natural history’ and laboratory-based studies. We strongly encourage further search for ant pollination systems in dry ecosystems, such as tropical savannahs and tropical high-altitude fields. We predict that a directed focus on shrubs and herbs with very small white flowers arranged in capitula and usually producing nectar will frequently reveal ants as their pollinators!