RobertsonNew Zealand et al.—Pollination Journal of Botany, and 2008, dispersal Vol. in46 Fuchsia: 299–314 299 0028–825X/08/4603–0299 © The Royal Society of New Zealand 2008 Assessing pollination and fruit dispersal in Fuchsia excorticata (Onagraceae) ALastaIR W. ROBErtsON cases very frequently by silvereyes, which also oc- Ecology, Institute of Natural Resources casionally rob nectar from flowers. We confirmed Massey University that hermaphrodites account for more than half the Private Bag 11222 plants in all populations, are fully self-compatible, Palmerston North 4474, New Zealand and can autonomously self in the absence of pollina- [email protected] tors (especially in plants with smaller herkogamy). Fruit production in hermaphrodites and (particularly) JENNY J. LADLEY females was frequently pollen-limited (mean Pollen DAVE KELLY Limitation Indices of 0.17 and 0.40, respectively), School of Biological Sciences and was correlated with visual assessments of pol- University of Canterbury len loads on the stigma, a useful index of pollinator Private Bag 4800 service. A comparison of the proportion of ripe or Christchurch 8140, New Zealand overripe fruit on branches exposed to birds versus KatE L. MCNUTT branches enclosed in wire cages showed that un- Ecology, Institute of Natural Resources caged fruit on Kapiti Island is removed almost as Massey University soon as it is ripe but on the mainland it persists for Private Bag 11222 much longer. The proportion of ripe or overripe Palmerston North 4474, New Zealand compared to green fruit is therefore an approximate index of dispersal service. Both indices may be use- PAUL G. PETERSON ful to managers concerned with measuring the level Landcare Research of mutualism service provided by native birds. Private Bag 11052 Palmerston North 4442, New Zealand Keywords reproductive assurance; gynodioecy; MERILYN F. MERRETT fruit-dispersal rates; frugivory; nectar robbing; or- Landcare Research nithophily; rapid assessment indices; bellbird; tui; Private Bag 3127 kereru; silvereyes; pollen limitation Hamilton 3240, New Zealand BRIAN J. KARL INTRODUCTION Landcare Research Private Bag 6 The widespread decline in abundance and ranges of Nelson 7042, New Zealand bird species worldwide has raised concerns about the failure of ecological interactions involving birds, especially fruit-dispersal and pollination mutual- Abstract We assessed the degree of pollen and isms (Sekercioglu et al. 2004). This is particularly fruit-dispersal limitation in the endemic gynodioe- relevant to New Zealand (Craig et al. 2000) as many cious New Zealand tree fuchsia Fuchsia excorticata terrestrial bird species are extinct (Holdaway 1989) over several seasons and at 17 sites throughout New or declining (Hitchmough et al. 2007), putting at Zealand. The flowers were visited mainly by two risk pollination and dispersal services (Clout & Hay endemic honeyeaters (bellbirds and tui), and in some 1989; Craig et al. 2000; Newstrom & Robertson 2005). Significant disruptions to bird-plant mutu- alisms have been recently demonstrated in New B07038; Online publication date 29 July 2008 Zealand forest species, and it appears that many bird- Received 27 August 2007; accepted 1 July 2008 visited plants now suffer from incomplete pollination 300 New Zealand Journal of Botany, 2008, Vol. 46 and/or fruit dispersal (McNutt 1998; Robertson et et al. 2007). This raises questions about whether fruit al. 1999; Montgomery et al. 2001; Kelly et al. 2004; dispersal by native birds is still effective (Clout & Merrett et al. 2007). Hay 1989) and whether other introduced birds may The risks from these kinds of disruptions de- substitute (Kelly et al. 2004, 2006). The exotic black- pend on whether there are reproductive-assurance bird (Turdus merula) and European starling (Sturnus mechanisms that can replace the role played by the vulgaris) also take the fruit (Burrows 1994) and may mutualists (Bond 1994). For example, autonomous therefore be candidates. In South Westland, where self-pollination and passive dispersal of fruits may tui, bellbirds and kereru remain abundant, O’Donnell be enough to at least replace the adult plants as they & Dilks (1994; summarised in Kelly et al. 2006) die. However, reproductive assurance through self- found that kereru accounted for 64% of visits to F. ing may incur its own costs if inbred offspring are excorticata fruits, followed by blackbirds 17%, sil- less fit than outcrossed seedlings due to inbreeding vereyes 12%, bellbirds 5%, and fantails (Rhipidura depression (Charlesworth & Charlesworth 1987; fuliginosa) 2%. In this paper, by measuring rates Herlihy & Eckert 2002; Scofield & Schultz 2006), of fruit removal from a range of sites throughout and passage through a bird gut may be required for the country, we test the extent of fruit dispersal adequate seed germination (Traveset 1998; Robert- failure. son et al. 2006). The aims of this study on F. excorticata were: (1) Fuchsia excorticata (tree fuchsia or kotukutuku) to determine the degree of self-compatibility and is a long-lived New Zealand endemic tree up to potential for autonomous selfing in hermaphrodites; 12 m tall, with trunks up to 70 cm diameter, and is (2) to test for pollen limitation in the field at a range found from North Cape to Auckland Island. Fuchsia of mainland sites; (3) to measure rates of fruit dis- excorticata has been reduced throughout much of persal; and (4) to develop field indices of pollination its range by brushtailed possum (Trichosurus vul- and dispersal service as quick assessment tools to pecula) herbivory (Pekelharing et al. 1998; Sweet- measure the status of bird-plant mutualisms. apple & Nugent 1999), and thus establishment of new seedlings is important to replace dying adults. Also, it is a partially seral species that is eventually overtopped by other trees. The production and dis- MATERIALS AND METHODS persal of viable seed is therefore important for the persistence of this species. Study species Fuchsia excorticata is gynodioecious (plants are Originally the main pollinators of F. excorticata either hermaphrodite or female) and both sexes are would have been bellbirds, tui, and the now ex- capable of high fruit set. Godley (1955) showed that tremely rare stitchbird (Notiomystis cincta; Thomson the hermaphrodites are self-compatible but have var- 1881; Kirk 1892; Delph & Lively 1985; Kelly et al. iable levels of herkogamy (the physical separation 2006). The silvereye (self-introduced in the 1850s) of the anthers from the stigma), which he suggested is now frequently observed on the flowers, but the could affect their chances of self-pollinating in the hypanthium tube is long in hermaphrodite flowers absence of pollinator visits. The primary flower (c. 20 mm) and silvereyes are thought to be unable visitors to F. excorticata are bellbirds (Anthornis to reach the nectar from the flower opening as their melanura), tui (Prosthemadera novaeseelandiae), beaks are too short (14.3 mm compared to 20 mm and silvereyes (Zosterops lateralis) (Delph & Lively for bellbirds), although silvereyes can access female 1985; Kelly et al. 2006). Since bellbird and tui popu- flowers which are about half as long (Delph & Lively lations have declined throughout much of mainland 1985). Instead, silvereyes frequently make holes in New Zealand in the last 150 years (Robertson et al. the side of the corolla of hermaphrodites to reach 1999), and silvereyes are currently the predominant the nectar and probably do not achieve pollination flower visitor at many sites (Kelly et al. 2006) but are this way. Bumblebees rob flowers by probing the considered to be primarily nectar robbers (Delph & cuts made by silvereyes for nectar and thus they are Lively 1985), we wanted to test whether this change probably not generally effective as pollinators and has compromised F. excorticata pollination. furthermore may deter subsequent bird visitation The fruits are also dispersed by tui and bellbirds by reducing nectar availability (Delph & Lively (Craig et al. 1981), and by kereru (New Zealand pi- 1985). geon, Hemiphaga novae-zelandiae; McEwan 1978), The 12 × 5 mm fruits are a berry containing which is listed as in “gradual decline” (Hitchmough c. 300–600 small seeds in four locules (Godley & Robertson et al.—Pollination and dispersal in Fuchsia 301 Berry 1995). Fruits ripen from green through maroon bagged to prevent cross-pollination. Tagged flow- to purple-black. Overripe fruits begin to desiccate ers were left for c. 8 weeks to allow fruit to set and and wrinkle, which distinguishes them from the then both types of fruits were harvested at both shiny ripe fruits. sites to count seeds. We counted the seeds in one of the fruit locules on up to 10 fruits of each type at Study sites and sex ratios each site. Two types of seeds were identified: thin We used 17 study sites from Lake Okataina in the and translucent, or fat and opaque. A study using North Island to Hinewai Reserve on the South Is- eight fruits from Ahuriri Reserve showed that the land’s Banks Peninsula during the course of the germination rate for fat seeds was 84.6% (n = 800) study (Table 1). The sites varied from scattered trees compared to only 0.92% of the thin seeds (n = 327), along roadsides and pasture margins, to plants on the so we present here totals for only fat seeds. We edges of forest patches, or along streams in forested used Generalised Linear Models (GLMs) in S-Plus catchments. Fuchsia excorticata was common at version 4.5 (Mathsoft, Inc., Seattle, WA) to analyse all study sites. At 13 of the sites we recorded the the data—binomial GLMs for fruit set, and Poisson ratio of hermaphrodites to female plants by noting GLMs for seeds per locule. the gender of all accessible flowering plants within the vicinity of our focal plants used for aspects of Bird visitation and nectar robbing the study. In the 1999/2000 season we recorded birds visiting flowers at nine sites and at a further two sites in the Self-compatibility 2002/03 season.