2.1 Study System and Experimental Set-up
The facultative biennial herb D. purpurea (Plantaginaceae) produces long showy inflorescences (Fig. 1) of nectar-rich flowers that open in sequence from the bottom up. The flowers are protandrous, with anthers dehiscing following anthesis and the stigma becoming receptive three to five days after anthesis. Animal pollination is required to achieve full seed set (Nazir et al., 2008; Mackin et al ., 2021), and the primary pollinators in the UK are the long-tongued bumblebees Bombus hortorum and B. pascuorum (Broadbent & Bourke, 2012). In introduced populations in Central and South America, hummingbirds are also important pollinators, in addition to bumblebees (Mackin et al ., 2021). In those populations we also observe high levels of nectar robbing by bumblebees (Riveros et al. 2006),Diglossa flower piercers, and hummingbirds. We have not observed bumblebees making holes on the corollas; instead, bees and hummingbirds are secondary robbers using the holes pierced by Diglossa birds. In surveys in two localities in Colombia (Floresta N=50 and Choachí N=60) we found that 36.2% and 47.8% (respectively) of all post-anthesis flowers were robbed, with robbing making up to 14.1% to 19.4% of all visits to flowers in those two populations.
Nectar production and secretion begins the day before the first pair of anthers dehisce, and peaks during anthesis and stigma maturation (Percival and Morgan 1965). Nectar is produced in floral nectaries located at the base of the ovary and escapes through modified stomata that are permanently open (Gaffal et al ., 1998). Nectar sugar concentration ranges from 16-27% and is predominately made up of sucrose (78.5%) with some glucose and fructose sugars (Gaffal et al ., 1998). In our study populations, a single flower secretes between 3.1 to 10.5µl of nectar over 24 hours without visitation (pers. obs).
We grew D. purpurea potted plants from seed collected from a wild population near Portsmouth (UK) in 2017, that were grown until flowering in summer 2019 and then transported to the University of Sussex campus in Falmer. Between 16-18 plants were selected at random to be in one of two treatment groups – “robbed” and “control” (non-robbed flowers). To standardise the rate of natural pollination that all plants experienced, plants were exposed for 3 hours each day to receive visits by local bumblebees, and for the other period of 21 hours the inflorescences were covered with a mesh bag to block visitation.
Plants in the robbed treatment group had all post-anthesis (nectar-producing) flowers manually robbed by piercing a hole in the proximal corolla tube with a microcapillary tube twice during the 3 hour pollination period (once at the start, and then again after 90 minutes). This rate of robbing is based on field observations in the non-native range where D. purpurea flowers are robbed less than once per hour, and on greenhouse trials on the experimental plants that showed that nectar is replenished at a rate of 2.3µl per hour during three hours following nectar depletion (N = 30 flowers over 10 plants). Compared to other species tested, this is a fast rate of replenishment (Castellanos et al ., 2002) as a single foxglove flower secretes up to 10.5µl of nectar per day.
Bumblebees will often avoid plants that have recently been visited by other floral visitors using olfactory and other cues (Stout et al ., 2001), so we manually handled the inflorescences (both controls and robbed) to mimic contact during simulated robbing. This entire procedure was repeated daily for the period that plants flowered between June 17 and July 3, 2019.