A Fossil Fuchsia (Onagraceae) Flower and an Anther Mass with in Situ Pollen from the Early Miocene of New Zealand 1

Home , Alepis, Circaea, Fuchsia, Fuchsia (moth), Fuchsia excorticata, Fuchsia magellanica

American Journal of Botany 100(10): 2052–2065. 2013.

A FOSSIL FUCHSIA (ONAGRACEAE) FLOWER AND AN ANTHER MASS WITH IN SITU POLLEN FROM THE EARLY MIOCENE OF NEW ZEALAND 1

D APHNE E. LEE 2 , J OHN G. CONRAN 3,6 , J ENNIFER M. BANNISTER 4 , U WE K AULFUSS 2 , AND D ALLAS C. MILDENHALL 5

2 Department of Geology, University of Otago, P.O. Box 56, Dunedin, New Zealand; 3 Australian Centre for Evolutionary Biology and Biodiversity & Sprigg Geobiology Centre, School of Earth and Environmental Sciences, Benham Bldg DX 650 312, The University of Adelaide, SA 5005 Australia; 4 Department of Botany, University of Otago, P.O. Box 56, Dunedin, New Zealand; and 5 GNS Science, P.O. Box 30-368, Lower Hutt, New Zealand

• Premise of the study: Fuchsia (Onagraceae) anthers, pollen, and an ornithophilous Fuchsia -like fl ower from an earliest Mio- cene lacustrine diatomite deposit at Foulden Maar, southern New Zealand confi rm a long record for Fuchsia in New Zealand and probably an equally long history for its distinctive honeyeater pollination syndrome. The anthers contain in situ pollen of the fossil palynomorph previously assigned to Diporites aspis Pocknall et Mildenh. (Onagraceae: Fuchsia L.). • Methods: We undertook comparative studies of the fl ower and anther morphology of the newly discovered macrofossils and compared the in situ pollen grains from the anthers with dispersed pollen grains from extant species. • Key results: The anther mass is referred to a new, extinct species, Fuchsia antiqua D.E.Lee, Conran, Bannister, U.Kaulfuss & Mildenh. (Onagraceae), and is associated with a fossilized Fuchsia -like fl ower from the same small mining pit. Because Dipo- rites van der Hammen is typifi ed by a fungal sporomorph, the replacement name for D. aspis is Koninidites aspis (Pocknall & Mildenh.) Mildenh. gen. & comb. nov. Phylogenetic placement of the fossils agrees with a proximal position to either sect. Skinnera or sect. Procumbentes. These are the oldest macrofossils of Fuchsia globally. • Conclusions: The fl oral structures are remarkably similar to those of modern New Zealand Fuchsia. They suggest that the distinctive honeyeater bird-pollination syndrome/association seen in modern New Zealand was already established by the late Oligocene–earliest Miocene. The implications for the biogeography and paleoecology of Fuchsia in Australasia are discussed.

Key words: anthers; biogeography; ﬂ ower; Fuchsia antiqua; in situ pollen; Koninidites aspis; Miocene; Oligocene; Onagraceae.

In contrast to the fl ora of neighboring landmasses of Austra- plants include Alepis Tiegh. and Peraxilla Tiegh. (Loranthaceae), lia and New Caledonia, the present-day fl ora of New Zealand is Alseuosmia A.Cunn. (Alseuosmiaceae), Fuchsia L. (Onagraceae), characterized by a depauperate array of fl owering plants, many Knightia R.Br. (Proteaceae), Metrosideros Banks ex Gaertn. monotypic woody genera, a predominance of white-fl owered (Myrtaceae), Phormium J.R.Forst. & G.Forst. (Xanthorrhoeaceae), taxa and unspecialized fl oral structures for generalist pollina- Rhabdothamnus A.Cunn. (Gesneriaceae), and Sophora L. tion by small birds and insects (Lloyd, 1985; Lee et al., 2001). (Fabaceae) ( Kelly et al., 2010 ). Of the 29 species of fl owering trees and shrubs in New Zealand Relatively little is known about the antiquity of the fl oral that possess ornithophilous fl owers, most are mainly pollinated structures and syndromes that relate to bird pollination in New by honeyeaters (Meliphagidae), including tūī ( Prosthemadera Zealand (and elsewhere). Lloyd (1985) postulated that the de- novaeseelandiae) and bellbirds (Anthornis melaneura ). These velopment of simplifi ed fl oral structures in New Zealand plants was related to a depauperate insect and bird fauna, although he did not speculate on the cause or antiquity of this general biotic 1 Manuscript received 20 December 2012; revision accepted 5 July 2013. The authors thank the Gibson family, Dr. Alan Walker, and Featherston impoverishment. Similarly, Kelly et al. (2010) showed that de- Resources Ltd for kindly allowing us access to the site. Liz Girvan (Otago pendence on bird pollination is unexpectedly high, with 85 na- Centre for Electron Microscopy) provided valuable assistance with the tive species (including 30% of the tree fl ora) having reports of SEM photographs. Jack Lamb of the World living collection of Fuchsia visits to fl owers, despite the very small number of nectarivorous species for BGCI, Fuchsia Species Conservation and Research, U.K. is bird species. They concluded that bird pollination in New Zealand thanked for the photo of F. cyrtandroides . Dr. Jamie Wood helped with is widespread and important for numerous taxa, including many pollination and dispersal information on modern Fuchsia in New Zealand. lacking apparently ornithophilous fl owers. The Departments of Geology and Botany, University of Otago, Dunedin Recent research in richly fossiliferous lagerstätten deposits and the School of Earth and Environmental Sciences, The University of of late Oligocene–early Miocene age in southern New Zealand Adelaide are thanked for the provision of resources to undertake this has revealed a wide variety of particularly well-preserved plant research. Funding for this study was provided by a Marsden Grant from the Royal Society of New Zealand. Prof. Bill Lee (Landcare Research, NZ) is fossils, including a fertile fern (Conran et al., 2010), leaves with thanked for comments on the manuscript. cuticle (e.g., Conran et al., 2009 ; Carpenter et al., 2010a , b ; 6 Author for correspondence (e-mail: [email protected]) Ferguson et al., 2010; Jordan et al., 2010; Lee et al., 2010; Bannister et al., 2012 ), fruits and seeds, and importantly, fl ow- doi:10.3732/ajb.1200643 ers with in situ pollen (Bannister et al., 2005; Lee et al., 2010).

Taken in conjunction with the extensive, continuous palyno- 1999 ; Martin, 2003 ), is considered to represent the modern morph record for the New Zealand Cenozoic ( Mildenhall, 1980 ; semiaquatic genus Ludwigia L. (including Jussiaea L.), which Pocknall and Mildenhall, 1984 ; Mildenhall and Pocknall, 1989 ), is now extinct in New Zealand, possibly indicating climatic these macrofossils make it possible to investigate and clarify change to more open and drier ecosystems ( Macphail, 1999 ). aspects of the history of angiosperms in New Zealand in some The affi nity of C. oculusnoctis was listed in Bannister et al. detail. Of particular interest is the antiquity of ecological inter- (2005 , p. 517) as “Onagraceae, cf. Epilobium .” actions relating to pollination and dispersal in relation to the Onagraceae appears to have diverged from its sister family evolution of the distinctive modern New Zealand fl ora. Lythraceae ~93 Ma in the early Cenomanian, while Fuchsia is One notable genus is Fuchsia, which is widespread through- thought to have differentiated from Circaea L. ~40 Ma in the out New Zealand today and has distinctively colored fl owers Eocene (Conti et al., 1993; Berry et al., 2004; Sytsma et al., and pollen and specialized fl oral structures for bird pollination 2004 ; Wagner et al., 2007 ; Xie et al., 2009 ). The presence of a ( Robertson et al., 2008 ). Similarly, the fruits display a bird- well-dated 23 Ma Fuchsia fossil fl ower and pollen-bearing an- dispersal syndrome (Armesto and Rozzi, 1989 ; Robertson et al., thers at Foulden Maar provides new information on the fl oral 2008 ). structure of early Fuchsia, in particular the shape of the fl oral Fuchsia also displays an unusual biogeographic pattern. Of tube and arrangement of the anthers. Furthermore, as macrofos- the 107 species in 12 sections ( Heywood et al., 2007 ; Wagner sil records for Onagraceae are extremely rare, these fl oral mac- et al., 2007 ; Mabberley, 2008 ), three species (and a natural hy- rofossils are very important for an improved understanding of brid) in two sections are endemic to New Zealand: sect. Skinnera the fossil record of Fuchsia and unraveling the molecular phy- (J.R.Forst. & G.Forst.) DC.: F. excorticata (J.R.Forst. & G.Forst.) logenetics and biogeographic history of Onagraceae. L.f. ( Fig. 1A ), F. perscandens Cockayne & Allan (Fig. 1B) and F. × colensoi Hook.f. (Fig. 1C); and sect. Procumbentes Godley & P.E.Berry: F. procumbens R.Cunn. ex A.Cunn. (Fig. 1E). The MATERIALS AND METHODS remaining species in sect. Skinnera , F. cyrtandroides J.W.Moore (Fig. 1D), is confi ned to Tahiti (Godley and Berry, 1995). The Fossil preparation — The fossils were recovered from small mining pits remainder of the genus is found throughout South and Central near Middlemarch in Otago, southern New Zealand. The plant fossils occur in America as far north as Mexico, but Fuchsia is notably absent the Foulden Hills Diatomite (Lee et al., 2007; Lindqvist and Lee, 2009), a fi nely from present-day Australia, New Caledonia, and other Pacifi c laminated biogenic lake deposit comprised mainly of pennate diatom frustules, islands ( Berry et al., 2004 ). siliceous chrysophycean algal stomatocysts, sponge spicules, insects, articu- The global fossil record of Onagraceae has until now been lated fi sh skeletons, and plant material ( Pole, 1996 ; Lee et al., 2007 ). The fossil based almost entirely on distinctive pollen, with macrofossils macrofl ora comprises numerous leaf compressions, typically with well-preserved cuticle, fruits, occasional bark and wood fragments, and rare fl owers. (seeds and fruits) rare for the family (Grímsson et al., 2011 and The anthers and fl ower described here, as with other fl ower parts from the site references therein) and none reported for Fuchsia ( Martin, (Bannister et al., 2005), were preserved on light-colored layers in the varved 2003). Within the family, Fuchsia pollen is highly recognizable diatomite ( Lindqvist and Lee, 2009 ). (Erdtman, 1952; Balme, 1995). Distinctive diporate pollen as- The Foulden Hills Diatomite was deposited in a small maar crater lake Diporites aspis (Lindqvist and Lee, 2009). Dating of volcanics associated with the maar forma- signed to the palynomorph Pocknall & Mildenh. ± from the late Oligocene–early Miocene Southland Gore Lignite tion indicate an earliest Miocene age of 23 0.2 Ma ( Coombs et al., 1986 ; Lee et al., 2007 ; Lindqvist and Lee, 2009 ; Kaulfuss et al., 2011 ). Measures was ascribed to an extinct species of Fuchsia (Pock nall The anther clump presented as a tight cluster lying on a small twig/branch and Mildenhall, 1984 ). It was also reported subsequently from (Fig. 2A, B). The fl ower was collected as an isolated compression on a light- the Oligocene to Miocene of Australia (Macphail, 1999 ; Martin, colored bedding plane ( Fig. 2C, D ). Initially, it was assumed that the anthers 2003) and the lower Miocene of South America (Zetter et al., were all that remained of a caulifl orous bud still in life position on a branch of 1999 ), the last given originally as Middle Eocene, but since re- the parent tree. However, this assumption was refuted by examination of the dated ( Barreda et al., 2009 ). wood structure of the branch, which proved instead to be more probably from a member of the Myrsinaceae, based on differences in vessel form and scalari- In this paper we describe and illustrate fossil Fuchsia anthers form perforation in particular (Carlquist, 1977; Meylan and Butterfi eld, 1978 ). in life position and with in situ pollen, apparently from a fossil- The bud had thus fallen into the lake and settled fortuitously on one of the rare ized bud, and a separate Fuchsia fl ower without pollen. Both wood fragments found in the diatomite before decomposing and leaving only were collected from Foulden Maar, a well-dated diatomite de- the anther mass. posit of earliest Miocene (23 ± 0.2 Ma) age in Otago, New Zealand Pollen grains were squeezed out of one of the anthers, removed using a that has yielded numbers of other fossil fl owers and infl ores- very fi ne paintbrush, and rinsed in water. The grains were mounted in glycerin jelly on a slide for light microscopy (LM) and photography, while others were cences, several of which have in situ pollen (Bannister et al., placed on a stub, sputter-coated with gold and palladium, and examined and 2005 ; Lee et al., 2010 ). photographed with a Cambridge S360 Stereoscan electron microscope. Other Because the form genus Diporites van der Hammen is typi- similar, dispersed pollen grains were retrieved from various-sized samples fi ed by a fungal sporomorph, Koninidites Mildenh. is proposed ranging between 0.70 and 10.76 g from a 183 m long core through the diato- herein as a replacement generic name for the fossil angiosperm mite deposit and prepared at GNS Science, Lower Hutt. These were treated palynomorphs placed previously in D. aspis and which are now with 10% hydrochloric acid and 50% hydrofl uoric acid digestion to eliminate carbonates and silicates respectively, nitric acid oxidation (varying between referred to Koninidites aspis (Pocknall & Mildenh.) Mildenh. 3–12 min) to eliminate unwanted amorphous organic debris, and potassium gen. & comb. nov. There is also an extensive fossil record of at hydroxide treatment to remove the soluble oxidation products and to release least two other genera of Onagraceae in New Zealand, both pollen. As Martin (2003) noted, all chemical treatments of pollen have an in- with triporate pollen (see Raine et al., 2011). The Neogene to fl uence on size. However, the pollen grains from the anthers were not pro- Corsinipollenites epilobioides cessed, but mounted directly from the fossil anther mass and viewed under an Holocene pollen type Krutzsch × represents the extant, largely alpine genus Epilobium L., which Axioplan 2 light microscope using a 40 objective with and without oil immersion. has 47 native species in New Zealand ( de Lange and Rolfe, 2010 ). The fl ower, anthers, and a pollen sample are held in the collections of the In contrast, the Paleogene to Neogene pollen taxon C. oculusnoctis Department of Geology, University of Otago. All other pollen samples are (Thiergart) Nakoman, also known from Australia (Macphail, stored at GNS Science, Lower Hutt. 2054 AMERICAN JOURNAL OF BOTANY [Vol. 100 October 2013] LEE ET AL.—NEW ZEALAND FOSSIL FUCHSIA 2055

Phylogenetic placement—To place the fossil phylogenetically, the extant Age— Earliest Miocene (23 Ma). members of the Old World Fuchsia sect. Skinnera, sect. Procumbentes , and the fossil were coded for 19 morphological and phytochemical characteristics Associated material (Table 1), derived largely from data in Averett et al. (1986), Williams and — An incomplete solitary ﬂ ower 19 mm Garnock-Jones (1986), Crisci and Berry (1990), and Godley and Berry (1995), high lacking petals, with spreading sepals, anther remnants and with missing data coded as “?”. The outgroup taxon represents a combination a well-developed, narrowly urceolate ﬂ oral tube ( Fig. 2C, D ). of characteristics seen in the New World sections of Fuchsia , as per Crisci and Berry (1990). Material examined— OU33220. These data were analyzed using the parsimony ratchet option (10 000 repli- cates; random addition; mult*TBR; hold 20 trees; sample 6 characters; charac- Pollen — Anteturma— Pollenites R.Potonié, 1931. ters nonadditive) in the program ASADO version 1.89 ( Nixon, 2004 ). The robustness of the trees was assessed using both bootstrapping (10 000 reps; 33% resampling) and Bremer (1994) decay analysis (20 steps limit) with the program Turma— Poroses Naumova, 1939. TNT 1.1 ( Goloboff et al., 2008 ), following Jordan and Hill (1999) and Conran et al. (2009). Character state mapping was performed in ASADO using the AC- Subturma— Diporines Naumova, 1939. CTRAN option. Genus— Koninidites Mildenh. gen. nov. ( Fig. 3 ).

SYSTEMATIC RESULTS Diagnosis—Grains oblate, barrel-shaped; diporate, apertural protrusions large, with atrium; exine psilate (LM) to granulate, Flower — Family— Onagraceae Juss., 1789 nom. cons. two-layered, nexine thickening at the base of apertural protrusions, thickening around apertures; viscin threads on proximal Tribe— Circaeeae Dumort., 1827. side.

Genus— Fuchsia L., 1753. Derivation of name—From the Māori word k ōnini given to the fruit of the New Zealand tree fuchsia, F. excorticata Section— Skinnera (J.R.Forst. & G.Forst.) DC., 1828. ( Moorﬁ eld, 2011 ).

Species—Fuchsia antiqua D.E.Lee, Conran, Bannister, Comparisons—This new genus differs from Diporites van U.Kaulfuss & Mildenh. sp. nov. ( Fig. 2A–B ). der Hammen (1954), Diporisporites van der Hammen (1954) and Granodiporites Varma & Rawat (1963) in having a 2-layered Diagnosis— Anther mass consisting of long and short anthers, exine (LM), viscin threads, and structured apertures. arranged in two levels with the short anthers uppermost; anthers containing pollen of Koninidites aspis. Type species— Koninidites aspis (Pocknall & Mildenh.) Mildenh. gen. & comb. nov. Description— An anther mass showing long (1.6–1.8 × 0.8– 1.0 mm) and short anthers (0.9–1.1 × 0.6–0.7 mm), arranged in Type locality and age— Drillhole 1075 at 65.5 m, 0.5 km east two levels with the short anthers uppermost ( Fig. 2A, B ). The of Underwood–Linds Bridge Road and Provincial Highway 96 anther sacs separated by wide, contrastingly colored anther conjunction, 9 km north of Invercargill, Southland; early Miocene nectives, and on one of the lower anthers, there is an oval scar ( Pocknall and Mildenhall, 1984 ). representing the attachment point of the subterminal fi lament ( Fig. 2A, B ). No pollen was visible initially, but compression of Botanical affi nity— Fuchsia (Onagraceae). the anthers forced out a clump of mature pollen grains tied together by viscin threads ( Fig. 3A–C ). Description— Pollen grains monads, diporate, shape oblate (barrel-shaped), outline elliptic in polar and equatorial view Holotype—OU33298 (Fig. 2A, B), includes material mounted (Fig. 3A–E), apertural protrusions sharply delimited, equatorial on slides and SEM stubs. All material is stored in the Depart- diameter 54–60 µm wide, polar axis 29–40 µm long; apertures ment of Geology, University of Otago, Dunedin, New Zealand. compound, 18–22 µm wide, 12–14 µm high, consisting of cir- cular pores often with ragged margins, voluminous atria, ellip- Etymology—The specifi c epithet comes from the Latin anti- tical endopores; exine 2–4 µm thick, two-layered (LM), nexine qua meaning old or ancient, in reference to the taxon being the thinner than sexine, at base of apertures the nexine shows a oldest Fuchsia macrofossil taxon yet discovered. markedly lamellate granular thickening, representing the boundary between apertural protrusion and central body of the Type locality— The anther mass was collected from a mining grain ( Fig. 3F, G ); sculpturing psilate in LM, consisting of seg- pit, Foulden Hills Diatomite, Foulden Maar, Otago, New Zea- mented rod-like and globular elements in SEM, sculptural ele- land. The site is registered as I43/f8503 in the New Zealand Fos- ments elongated in apertural regions; viscin threads on proximal sil Record File administered by the Geoscience Society of New side (Fig. 3H, I), simple, up to at least 60 µm long, 0.14–0.32 Zealand and GNS Science (New Zealand Map Grid Infomap I43 µm thick, smooth, segmented, threads multiple at attachment reference 929166: 45.5271° S, 170.2218 ° E). point, branching and fusing near the grain.

← Fig. 1. Flowers of modern New Zealand Fuchsia taxa for comparison. (A) F. excorticata (Orokonui Nature Reserve, Otago) showing color change to red when no longer receptive. (B) F. perscandens (cult. Dunedin Botanic Gardens). (C) F. × colensoi (cult. Dunedin Botanic Gardens). (D) F. cyrtandroides (cult. FUCHS garden (BCGI), Rotheram, UK ex Berry 4628 [PAP, MO]), labeled for comparison with the fossil ﬂ ower in Fig. 2 (T = ﬂ oral tube, P = petal, S = sepal, A = anthers). (E) F. procumbens (cult. Landcare New Zealand, Lincoln). Scale bars A–D = 10 mm, E = 5 mm. 2056 AMERICAN JOURNAL OF BOTANY [Vol. 100 October 2013] LEE ET AL.—NEW ZEALAND FOSSIL FUCHSIA 2057

Discussion— The genus Diporites was fi rst described by van Skinnera and Procumbentes , although the pollen of the fossil der Hammen (1954) , encompassing pollen grains with two appears to be somewhat more elongate than the modern taxa. pores—“diporate (con 2 poros)”. The only species described Daghlian et al. (1985) noted that species differentiation between was Diporites grandiporus van der Hammen, and it is quite both fossil and modern taxa in this group was diffi cult. Thus, clear from the description (p. 83) that van der Hammen regarded given the 23 million-year time gap separating the macrofossils the palynomorph as a pollen grain. We have not seen the origi- from the modern species and the inability to place the fossil nal specimen, and unfortunately, the species was illustrated defi nitively into any extant taxon, it seems more appropriate to only by one drawing, and its morphology is unclear. Subse- put it into a new species. quently, van der Hammen (1956) described Diporites amplis- The associated fossil fl ower also conforms closely to the gen- simus van der Hammen as a lectotype of the genus Diporites, eral shape of modern South Pacifi c Fuchsia fl owers (Fig. 1) in but selected a modern pollen grain of Fuchsia hartwegii Benth. as possessing spreading sepals and a well-developed, narrowly ur- the type species. If this taxon were to be regarded as the type for ceolate fl oral tube (Godley and Berry, 1995). In the most wide- Diporites, the genus would be a junior synonym of Fuchsia . spread modern New Zealand tree species, F. excorticata , each Potonié (1960) provided a more complete diagnosis of the fl ower consists of a basal ovary joined to a funnel-shaped fl oral genus Diporites in which he correctly and formally designated tube. This terminates in four sepals that are pointed and spread- D. grandiporus as the type species for the genus, but gave no ing to slightly recurved, four small petals, and eight stamens (or indication as to whether he regarded the genus as representing staminodes in female fl owers). The ovary is cylindrical with a spores or pollen. In contrast, Rouse (1962) accepted D. amplis- long slender style, usually longer than the stamens and topped simus as the type species of Diporites and the use of a modern by a large, globular stigma, and bisexual fl owers have stamens pollen as the holotype, describing a new species. Potonié’s de- with long fi laments and anthers. The fossil fl ower resembles scription was accepted by Jansonius and Hills (1976) in their members of sect. Skinnera in size, relative proportions of the translation of his German and van der Hammen’s Portuguese fl oral tube and anthers, and the spreading arrangement of the works into English, and they commented on the problems caused sepals ( Fig. 2C, D ). However, it lacks the more swollen base of by van der Hammen’s systematic work. the extant New Zealand species in the section and is rather like Diporites grandiporus was described originally from late the Tahitian F. cyrtandroides . The lack of a swollen base to the Cretaceous sediments of Columbia and is now regarded as a fl oral tube is also shared with F. procumbens (sect. Procum- fungal spore, being listed in a number of fungal catalogues (e.g., bentes ), but the fossil anther mass differs in having the anthers Index Fungorum : http://www.indexfungorum.org ). Mildenhall in two ranks ( Fig. 2A, B ). and Pocknall (1989) also discussed the problems created by van Note that we have made the assumption that the isolated der Hammen’s systematic descriptions, but decided to use fl ower that lacks pollen is from the same taxon that produced Diporites as a form generic name for a new species of pollen the anther mass with in situ pollen. Both macrofossils were col- grain clearly related to the modern genus Fuchsia . However, the lected from a similar stratigraphic level in the same small min- morphology of D. grandiporus with its simple pore morphol- ing pit. Of the over 5000 fossils collected from this site, we ogy and single layered exine is quite unlike the morphology of have about 30 individual fl owers and infl orescences from a fossil Fuchsia -like pollen, and a new morphogenus is therefore range of families. At least 10 of these have in situ pollen, in- required for the latter. cluding Fouldenia staminosa Bannister, D.E.Lee & I.R.Raine Pollen of K. aspis was only found at Foulden Maar in two (aff. Rutaceae; Bannister et al., 2005 ), Mallotus Lour. –Macar- samples from a 183 m core (42 m and 105 m), as well as in anga Thours (Euphorbiaceae; Lee et al., 2010), Hedycarya surface outcrops ( Bannister et al., 2005 ). The grains fall into the J.R.Forst. & G.Forst. (Monimiaceae), and representatives of higher size range of the dispersed specimens described by Cunoniaceae, Elaeocarpaceae, Loranthaceae, Luzuriagaceae, Pocknall and Mildenhall (1984) as D. aspis , but are otherwise Euphorbiaceae or Picrodendraceae, and Rutaceae that are yet to identical in morphology. This taxon has been recorded from be described formally (Lee et al., 2012). Although Loranthaceae both Australia (Berry et al., 1990) and South America (Barreda also have 4-partite fl owers often with an apparent fl oral tube et al., 2009) and is discussed in numerous other papers (e.g., and refl exed tepals and might be considered possible candidates Grímsson et al., 2011 ). In contrast, Diporites granulatus Rouse for the Fuchsia -like fl ower, the anther shape and attachment is (1962) with very small pores and D. iszkaszentgyorgyi Kedves quite different, and the undescribed mistletoe fl ower at Foulden (1986) with large, nonaspidate pores are both excluded from Maar (with in situ pollen) is very different from the fossil fl ower Koninidites . Other species of Diporites are clearly fungal. described here. The only other tubular fl ower found at the site does not look like Fuchsia , is 5-partite, and instead closely re- Justifi cation of assignment to a new species of Fuchsia — sembles the modern New Zealand endemic genus Alseuosmia The fossil anther mass with its long and short anthers arranged A.Cunn. (Alseuosmiaceae). in two levels (Fig. 2A, B) closely resembles the structure of There is remarkably close correspondence between the list of extant Fuchsia in size, shape and disposition of the anthers, and fl ower taxa and dispersed pollen types in the lake sediment subterminal fi lament attachment (Fig. 2E–G). The K. aspis (D. C. Mildenhall et al., unpublished manuscript). The 1500-m- pollen also matches closely to modern members of sections diameter maar lake in which the fossils occur was effectively a

← Fig. 2. (A) Fuchsia antiqua isolated anther mass showing long and short anthers arranged in two levels; holotype OU33298. (B) Outline drawing of same. (C) Associated fl ower with spreading sepals (S), anther remnants (A), and well-developed, narrowly urceolate fl oral tube (T); OU33220. (D) Outline drawing of same. (E) F. excorticata anther mass from a dissected, unopened bud (cultivated plant, Dunedin Botanic Gardens); (F) detail of wide anther connective and subterminal fi lament attachment. (G) F. magellanica Lam. anther detail (cultivated plant, Dunedin Botanic Gardens). White arrows indicate the anther sacs separated by wide, contrastingly colored anther connectives; black arrows mark point of attachment of the subterminal fi lament, or in the fossil, its scar. Scale bars: A, B = 5 mm, C–G = 1 mm. 2058 AMERICAN JOURNAL OF BOTANY [Vol. 100 .

5 closed system (Lindqvist and Lee, 2009), and the presence of a Fuchsia- like fl ower and an anther mass indicates that the source plants lived immediately adjacent to the lake margin. In these circumstances and given the very unlikely chances of fi nding fossil fl owers at all, even under ideal preservation conditions, it

. Leaf underside: seems most parsimonious to assume that a ﬂ ower with Fuchsia - 16

. Flower disposition: . Flower type morphology and an anther mass containing undoubted 8 Fuchsia pollen represent the same taxon. . Sepals: 0, spreading; 1, 12 Phylogenetic analysis— Analysis of the fossil together with

. Leaf phyllotaxis: 0, opposite; 1, extant taxa from Fuchsia sections Procumbentes and Skinnera

19 for the data in Table 1 produced three equally parsimonious trees of length 25 steps, CI = 88, RI = 76 ( Fig. 4 ), differing only in the position of the fossil as either sister to F. procumbens . Flavone sulphates: 0, absent; 1, present. . Flavone 4 (sect. Procumbentes), proximal to the New Zealand species of sect. Skinnera , or as sister to all Old World Fuchsia species. 200 . ≥ However, when these trees are combined as a strict consensus tree ( Fig. 4A ), the fossil sits in a polytomy consisting of a well- supported crown clade representing the New Zealand members (data for extant taxa derived mainly from Crisci and Berry mainly from taxa derived (data for extant of sect. Skinnera (89% bootstrap support), F. procumbens, and F. cyrtandroides .

. Leaf texture: 0, membranous; 1, subcrassate. . Leaf texture: Character mapping onto one of the trees showed that the ma- Skinnera 15 jority of the character states represent unique apomorphies

sect. ( Fig. 4B ) with the Old World Fuchsia species distinguished from the outgroup New World species by the presence of ﬂ a- . Stigma: 0, exceeding anthers; 1, equaling . Stigma: 0, exceeding

. Flower color change: 0, absent; 1, present. . Flower vones, blue pollen, ﬂ avone sulphates, and alternate leaves, and 11 Fuchsia

18 sect. Skinnera by gynodioecious ﬂ owers, an obconical ﬂ oral tube with strongly constricted base, and a vine/scrambler habit

. Ovule number per ovary: 0, <200; 1, . Ovule number per ovary: (but with a reversal in F. excorticata to a tree habit). Placement 3

. Sexual system: 0, hermaphrodite; 1, gynodioecious; 2, subdioecious. . Sexual of the fossil in the analysis was hampered by the amount of miss- 7 ing data, as well as character states shared with the outgroup, and the characters placing it with the New Zealand members of sect. Skinnera or F. procumbens (in the other two trees) were in both cases features for which the fossil was coded as missing.

DISCUSSION . Floral tube obconical: 0, absent; 1, present. 14 . Stamens: 0, whorled; 1, subequal. History of Onagraceae— 10 There is considerable biogeo-

. Pollen color: 0, cream; 1 blue. graphic interest in relationships within the 22 genera in the fam- 2 ily (e.g., Levin et al., 2004; Wagner et al., 2007; Grímsson et al., 2011). The family contains two subfamilies: Ludwigioideae ( Ludwigia) and Onagroideae with six tribes (Wagner et al., 2007), but the majority of species belong to 15 genera in the tribes Onagreae and Epilobieae, most of which are found in western North America. In contrast, the tribe Circaeeae consists . Eriodictyol 7-glucuronide: 0, absent; 1, present.

6 of the temperate boreal genus Circaea (eight species) and Fuch- sia with 107 species in 12 sections and largely restricted to Cen- . Habit: 0, shrub or tree; 1, climbing vine; 3, procumbent creeper. . Habit: 0, shrub or tree; 1, climbing vine; 3, procumbent creeper. tral and South America (Conti et al., 1993; Berry et al., 2004; 17 Wagner et al., 2007 ), but with three species from two sections

. Flavones: 0, absent; 1, present. . Flavones: in New Zealand and one in Tahiti ( Godley and Berry, 1995 ; 1 Wagner et al., 2007 ). Most fossil evidence for Onagraceae comes from the distinctive pollen that can be distinguished from other angiosperm . Petals: 0, conspicuous; 1, reduced; 2, absent. 9 families using combined LM and SEM studies (e.g., Zetter and 11110010 0 011111110 0 0 0 0 1 0 1 0 0 1 1 1 1 0 1 2 1 0 1 1 11100000 0 011010021 0 1 0 1 0 1 0 1 1 0 2 0 0 0 1 0 0 2 0 1 ????,00 ? 1 0 0 ? ? ? ? 11111110 0 0????????1,20 0 0 1 1 0 2 1 1 1 Keri, 1987; Martin, 2003; Punt et al., 2003; Grímsson et al., 2011 ). Onagraceae-like pollen (mostly of the form genus Cors- . Floral tube base: 0, unconstricted; 1, strongly constricted.

inipollinites) is widespread in Paleogene fossil deposits from 13

North and South America and Europe (see summary in Ramírez-

Arriaga et al., 2006), as well from as the Eocene of Australia exed. exed.

1. to other members of Characters and character states used to place the fossil phylogenetically relative ( Foster, 1982 ). Neogene seeds and/or fruits have also been re- [1990] and references therein). [1990] 0, green; 1, whitish; 2, silvery-white. 0, green; 1, whitish; 2, silvery-white. alternate. Missing data coded as “?”. Flavonol diglycosides: 0, absent; 1, present. Flavonol 0, pendant; 1, erect. reﬂ colensoi ported for Circaea , Epilobium , and Ludwigia ( Tobe et al., 1988 ; × ABLE F. antiqua F. F. excorticata F. F. F. procumbens F. perscandens F. T 12345678 9 10111213141516171819 Taxa Outgroup cyrtandroides F. 0 0 0 0 0 0 0 0 0,1 0 0 0 0 0 0 0 0 0 0 † Martin, 2003 : table 5; Grímsson et al., 2011 ). October 2013] LEE ET AL.—NEW ZEALAND FOSSIL FUCHSIA 2059

Kirchner (1898) described a fl oral impression from the Ma)”, but the locality was listed erroneously as S136/824, in- Eocene North American Florissant Formation, regarding it as stead of S146/824. The second record is given as “?Awamoan possibly representing Onagraceae. Manchester (2001) reexam- (N91/547)”, which is in the Awakino Valley. However, the fossil ined the specimen, concluding that this assessment was reason- pollen grain illustrated came from “N133/518, Taranaki?”, able, despite the fl ower being apparently 5-partite. In addition, which would be of Taranakian age (10.9–5.3 Ma). Onagroideae pollen from two morphospecies of Corsinipollin- The oldest confi rmed New Zealand record is from the late ites was reported at Florissant ( Wingate and Nichols, 2001 ), as Oligocene Pomahaka Formation and the species ranges through was pollen similar to Xylonagra J. Donnell Smith & Rose (On- to the Pleistocene (Raine et al., 2011). Further records of New agreae) and/or Lopezia Cavanilles sect. Riesenbachia Plitmann, Zealand fossil Fuchsia pollen include Daghlian et al. (1985), P.H. Raven & Breedlove (Lopezieae) (Leopold et al., 2008), Mildenhall and Pocknall (1989) , Mildenhall (1993) and Raine suggesting that the family was relatively diverse at the site. et al. (2011). Nevertheless, none of the pre-Pliocene/Pleisto- Three genera of Onagraceae: Epilobium , Fuchsia , and Lud- cene records is considered to represent any of the three extant wigia have been reported from the Cenozoic pollen record of New Zealand species ( Mildenhall and Pocknall, 1989 ; Lee et al., New Zealand ( Raine et al., 2011 ), although Ludwigia is no lon- 2012 ). ger represented in the native fl ora. Monads and regular, tetrahe- Diporites aspis (now Koninidites aspis ) has been used as a dral tetrads of triporate pollen attributed to Corsinipollenites general name for any fossil Fuchsia pollen, without any attempt oculusnoctis were reported from late Eocene and Oligocene being made to assign it to a particular section. In a detailed sediments ( Pocknall, 1985 ; Raine et al., 2011 ), as well as from study of fossil Fuchsia pollen from New Zealand, Daghlian et al. the early Miocene Foulden Hills Diatomite ( Bannister et al., (1985) discussed the differences between modern New Zealand 2005). This pollen has an affi nity with Ludwigia s.l. (Pocknall, and Tahitian Fuchsia species and those from the Central 1989 as Jussiaea), a pantropical herbaceous to woody and fl oat- and South American sections of the genus. They concluded that ing to emergent aquatic macrophyte or helophyte that is mainly D. aspis was aligned most closely with section Skinnera , which pollinated diurnally by bees, but also by fl ies and butterfl ies at the time incorporated all the South Pacifi c species. Godley ( Heywood et al., 2007 ; Mabberley, 2008 ). Corsinipollenites and Berry (1995) transferred the New Zealand F. procumbens to pollen has been reported from middle–late Eocene sediments of a monospecifi c section, Procumbentes , with the remaining New the Yamba Basin, Queensland (Foster, 1982). Pollen, fruits, and Zealand species and that from Tahiti retained in section Skinnera . seeds of Ludwigia are also reported from the Paleocene to Early There appear to be minor differences in viscin threads between Eocene of South America, Eocene of North America and the the species: in F. excorticata and F. procumbens the threads are Oligocene of Europe ( Martin, 2003 ). segmented; those of F. cyrtandroides are only lightly segmented, Pollen assigned to Epilobium is fi rst recorded in North America while those of F. perscandens may appear smooth although in the Oligocene, Europe in the Pliocene, and from Australia in they are even more lightly segmented. However, Daghlian et al. the Quaternary (Berry et al., 1990; Martin, 2003). The oldest (1985, p. 1046) concluded that it was “in most cases not possi- records from New Zealand are of late Oligocene age ( Pocknall, ble to determine species based on pollen morphology.” 1982: as Corsinipollenites epilobioides), but the modern form Diporites aspis was recognized subsequently from the early of Epilobium pollen appears fi rst in the Early Pliocene ( Nelson Oligocene to mid-Miocene of Australia (Macphail, 1999; Martin, et al., 1988 ). 2003). More recently, Macphail and Cantrill (2006) compiled a record of selected miospores from the southwest Atlantic region History of Fuchsia— The fl ower and pollen-bearing anthers and listed Diporites aspis from the ~late? Oligocene of Cen- from the earliest Miocene of southern New Zealand provide the tral Argentina ( Barreda, 1997), the late Oligocene and early fi rst record of Fuchsia macrofossils globally. No leaves, fruit, Miocene of southern Argentina and Tierra del Fuego (Barreda or seeds attributable to Fuchsia have been reported previously and Palamarczuk, 2000a, b ;) and the early late Oligocene of (Martin, 2003), apart from a very recent record of Fuchsia seeds southernmost South America ( Barreda et al., 2009 ). from Holocene moa coprolites from New Zealand ( Wood et al., 2012: table 1). The cuticle of F. excorticata is “either absent or Biogeographic origins of Fuchsia — Godley and Berry (1995 , poorly developed, and the stomata are not sunken or otherwise p. 498) noted that “The South Pacifi c species of Fuchsia are key protected” (Godley and Berry, 1995, p. 491), making Fuchsia elements to understanding evolution in the genus because of leaves, the soft fl eshy fruits, and tiny seeds unlikely candidates their unusual combination of characters and their widely dis- for fossilization. junct distribution.” These South Pacifi c species are character- Fuchsia pollen is highly distinctive within Onagraceae and ized by a wide variety of habit types from 12 m tall trees or a generally defi nable by the characteristics of 2-colporate, po- woody liana to a low-growing scrambler, possession of fl a- rorate grains with abundant viscin threads ( Nowicke et al., 1984 ; vones, greenish-colored fl owers that change color during anthe- Grímsson et al., 2011 ). There are also some South American sis, strongly reduced or missing petals, scentless fl owers that species with three apertures, but this is thought to be a derived produce copious nectar at the base of the fl oral tubes and dis- condition in the genus and possibly related to polyploidy ( Daghlian tinctive bright blue pollen (Godley and Berry, 1995). These et al., 1985). The genus has been reported from the early Oligo- differences (summarized in Table 1 ) help to confi rm that sec- cene of Australia (see review in Martin, 2003 ) and lower Miocene tions Skinnera and Procumbentes diverged possibly as early of South America ( Zetter et al., 1999 ). as the Eocene from their South American relatives. This di- The fi rst mention of fossil Fuchsia pollen from New Zealand vergence time is confi rmed by the well-supported crown clade was by Couper (1953 : 44), who gave a stratigraphic range of representing the New Zealand members of sect. Skinnera Pleistocene to Holocene and noted that this pollen was very (89% bootstrap support), F. procumbens and F. cyrtandroides rare. Couper (1960) later added two Miocene fossil localities ( Fig. 4A ) and the fact that most of the characters used in the with Fuchsia pollen, but these records are problematic. The fi rst analysis represent unique synapomorphies ( Fig. 4B ). How- is listed in the text ( Couper, 1960 , p. 48) as “Awamoan (20–17.5 ever, the lack of clear fl oral synapomorphies in the fossil make 2060 AMERICAN JOURNAL OF BOTANY [Vol. 100 its placement other than as a member of the Old World Fuch- large petals ( Wagner et al., 2007 ). A few sections of Fuchsia are sia lineage diffi cult. also pollinated by bumblebees or in one case, by tachinid fl ies Berry (1982, p. 6 ) postulated that New Zealand species of ( Wagner et al., 2007 ). Fuchsia are almost certainly derived from American ancestors, In contrast, the red, orange, or greenish nectar-rich fl owers of so the genus probably evolved considerably earlier. Bird polli- the New Zealand species are smaller, have strongly reduced to nation is a dominant theme in Fuchsia and is present in both absent petals, and are all fl orally heteromorphic, bearing her- New World and Old World groups; the fruits are also adapted maphrodite and female fl owers (Godley, 1955; Delph and Lively, to bird dispersal. Specialized fl ower-visiting birds probably did 1985 ; Delph, 1989 ). Fuchsia excorticata (tree fuchsia) is polli- not evolve until the Eocene (Sussman and Raven, 1978), but it nated mainly by perching New Zealand honeyeaters such as bell- is uncertain whether the common ancestor of Fuchsia would birds and tūī (Robertson et al., 2008), with visitation related to have been bird-pollinated or not. At any event, plants similar to fl oral tube color change; red fl owers indicating that the fl ower is the existing species of Fuchsia probably could not have evolved postanthesis, no longer bears nectar and is thus avoided ( Delph prior to the Eocene. The most likely hypothesis, therefore, is and Lively, 1985 ; Delph, 1989 ). Other native birds observed tak- that the genus originated in the Eocene or Oligocene in South ing nectar from F. excorticata include the stitchbird ( Notiomystis America.” cincta); Perrott and Armstrong, 2000); saddleback ( Philesturnus The occurrence of Fuchsia pollen in South America, Australia carunculatus; Higgins et al., 2006). The volant parrots the kea and New Zealand by the early to mid-Oligocene suggests that ( Nestor notabilis) and kākā (N. meridionalis ) have also been ob- the genus originated in the Eocene or earlier. Berry et al. (2004) served eating the fl owers ( O’Donnell and Dilks, 1994 ). raised the possibility of overland dispersal between New Zealand, The blue-green fl owers of the lianas F. perscandens and Australia, and South America via Antarctica during the Eocene F. × colensoi are similarly bird-pollinated ( Montgomery et al., to Oligocene, which is possible, but as yet Fuchsia pollen has 2001 ). In contrast, Whitaker (1987) speculated that the erect, not been reported from Antarctica. bright-orange fl owers of the prostrate groundcover species The major biogeographic disjunction between the South Pa- F. procumbens (the only member of sect. Procumbentes ) might cifi c and Central and South American sections within Fuchsia be lizard (gecko) pollinated and dispersed, although this is un- has been the subject of considerable debate. Earlier work sug- confi rmed. The fl owers in modern New Zealand species are gested that Fuchsia might have reached New Zealand via Tahiti also visited by a few other native and introduced birds, but from South America (Raven and Axelrod, 1972; Berry et al., some of these such as silvereyes, Zosterops lateralis Latham, 2004 ). However, this is now considered unlikely, following the 1802, are nectar-robbers and do not assist pollination ( Delph discovery of K. aspis pollen of Oligocene age in New Zealand and Lively, 1985). Thus, honeyeaters and possibly lizards have and Australia. Dispersal to Tahiti from New Zealand is likely to been the likely pollinators of the New Zealand Fuchsia since its have occurred much more recently ( Sytsma et al., 1991 ; Godley probable arrival into the region in the Eocene–Oligocene, as no and Berry, 1995 ), particularly given the recent geological age representatives of the hummingbird clade are yet known from of 0.6–1.2 Ma for Tahiti ( Neall and Trewick, 2008 ). Although the southwest Pacifi c, and recent fossil fi nds from Central Otago the present island of Tahiti is ~4000 km from New Zealand and suggest bird and reptile diversity in the Miocene of New Zealand ~6000 km from South America, there were probably many was comparable to today ( Tennyson et al., 2011 ). ephemeral volcanic islands between New Zealand and Tahiti Dispersal of the red to purple fl eshy berries that encase tiny ( Lee et al., 2001 ; Steadman, 2006 ), making biotic fruit dispersal seeds is carried out in part by the same honeyeater pollinators, possible. including bellbirds and tūī; this double-dependency on birds for The disappearance of Fuchsia from Australia in the late both pollination and dispersal is rare globally (W. G. Lee, Land- Miocene appears to have resulted from increasing aridity or care Research, personal communication, 2012). Possible con- climate cooling (Godley and Berry, 1995). However, several ex- tenders as dispersers include the strong-fl ying fruit-eating New tant New Zealand Fuchsia species are either fully or semidecidu- Zealand pigeon Hemiphaga novaeseelandiae, which has been ous, especially in colder areas (McGlone et al., 2004), and the observed to eat fruit/seeds, buds, fl owers and leaves of F. ex- most cold-tolerant species (F. excorticata ) is limited to lowland corticata ( McEwan, 1978 ; O’Donnell and Dilks, 1994 ; Emeny conditions in cool-temperate forests. It is likely that fossil et al., 2009). Other Fuchsia -eating birds include the fl ightless species of Fuchsia were fully evergreen during warm late kākāpō ( Strigops habroptilus; Butler, 2006) and weka ( Gallirallus Oligocene and early Miocene conditions in both New Zealand australis ), silvereye ( Williams and Karl, 1996 ), gray fantail (Rhipi- and Australia. dura fuliginosa ; O’Donnell and Dilks, 1994 ), yellow-crowned parakeet ( Cyanoramphus auriceps; Buller, 1888), and yellow- Reproductive syndromes — Floral evolution and specializa- head ( Mohoua albicilla; Higgins and Peter, 2002). There is also tion in extant Fuchsia shows the development of several strate- some evidence for dispersal in New Zealand by tree wetas gies relating to bird pollinators (Berry, 1982, 1985 ; Breedlove ( Hemideina ; King et al., 2011 ), although the effectiveness of et al., 1982 ; Eyde, 1982 ; Clout and Hay, 1989 ; Godley and this is disputed ( Wyman et al., 2011 ). South American Fuchsia Berry, 1995). Most South and Central American Fuchsia are species are similarly dispersed primarily by a range of birds pollinated by hovering hummingbirds attracted by the nectar (Armesto and Rozzi, 1989), but also by small mammal species associated with brightly colored, pendant terminal fl owers with ( Traveset et al., 1998 ).

← Fig. 3. Koninidites aspis pollen from Foulden Maar extracted from the anthers of Fuchsia antiqua (OU33298a). (A) Pollen clump; OU33298b. (B) Close-up of previous photograph showing viscin threads (arrows). (C) SEM of two pollen grains entangled by beaded viscin threads. (D, E) SEM, lateral view of two pollen grains. (F) Subterminal view showing terminal pore opening and exine surface. (G) Close-up of terminal pore. (H, I) Detail of exine and viscin thread showing simple thread base and semibeaded structure. Scale bars: A = 100 µm, B = 25 µm, C–F = 20 µm, G = 2 µm, H, I = 1 µm. October 2013] LEE ET AL.—NEW ZEALAND FOSSIL FUCHSIA 2061 2062 AMERICAN JOURNAL OF BOTANY [Vol. 100

Fig. 4. Phylogenetic placement of the fossil. (A) Strict consensus tree resulting from three equally most-parsimonious cladograms (length 25, CI = 88, RI = 76) of Old World Fuchsia taxa plus the fossil; numbers at nodes represent bootstrap support/Bremer decay values. (B) One of the three trees showing character evolution; character numbers and states are those from Table 1 ; fi lled squares are unique apomorphies; hollow squares, homoplasies; dashed branches are those that collapse under strict consensus. Godley and Berry (1995) reported that no native birds had and Christidis, 2004 ; Driskell et al., 2007 ). An ancient associa- been observed visiting the fl owers of the greenish-pink-fl owered tion between Fuchsia and nectar-feeding birds (hummingbirds hermaphroditic Tahitian endemic F. cyrtandroides, and no extinct in the New World and honeyeaters in the southwest Pacifi c) is meliphagid birds have been reported from Tahiti (Steadman, also supported indirectly by the presence of fossilized honeyeater 2006; T. H. Worthy, Flinders University, personal communica- bones in Miocene deposits from Australia (Boles, 2005) and tion, 2012). However, prior to extinctions following human New Zealand ( Tennyson et al., 2011 ), as well as Paleogene-aged settlement, fl ighted fruit-eating pigeons related to the modern hummingbird fossils from Europe (Bochenski and Bochenski, New Zealand species, together with parrots and parakeets may 2008 and references therein) and South America ( Poinar et al., have acted as dispersers of Fuchsia fruit to Tahiti and possibly 2007 ). other Pacifi c islands, and large fruit pigeons are known from the Miocene of New Zealand ( Worthy et al., 2009 ). Conclusions — This paper reports a direct association between The Fuchsia fossils reported in this study suggest that the an anther mass and the fossil Fuchsia pollen type Koninidites current reproductive syndromes dependent on bird pollination aspis and an indirect association with a solitary fl ower with dis- in the New Zealand tree fl ora are of considerable antiquity. The tinctive morphology that closely resembles modern New Zealand presence of honeyeaters in early Miocene fossil deposits sug- Fuchsia. Both the fl ower and the bud originally containing the gest that this guild has coevolved with Fuchsia and other orni- anthers were blown or fell into a small maar lake from sur- thophilous plant species over at least the last 23 million years as rounding trees, shrubs and lianas that were part of a diverse rain- this group of nectarivorous birds radiated in Australasia (Driskell forest adjacent to the lake. Because leaves are not present, it is October 2013] LEE ET AL.—NEW ZEALAND FOSSIL FUCHSIA 2063 not possible to determine whether the fossil species was decidu- B REEDLOVE , D. E. , P. E. BERRY , AND P . H . R AVEN . 1982 . The Mexican and ous, but the majority of the species in the Lauraceae-dominated Central American species of Fuchsia (Onagraceae) except for sect. forest at the Foulden Maar site described by Bannister et al. Encliandra. Annals of the Missouri Botanical Garden 69 : 209 – 234 . (2012) appears to have been multiannual evergreen in the sense B REMER , K. 1994 . Branch support and tree stability. Cladistics 10 : of McGlone et al. (2004) . 295 – 304 . B ULLER, W. L. 1888 . A history of the birds of New Zealand, 2nd ed.. Sir The fl oral morphology incorporating a long, pendulous fl oral Walter Lawry Buller, London, UK. tube and refl exive sepals indicates a likely close association with B UTLER , D. J. 2006 . The habitat, food and feeding ecology of kakapo in bird pollination (ornithophily), suggesting that this evolution- Fiordland: A synopsis from the unpublished MSc thesis of Richard ary shift to bird-mediated pollination in the genus Fuchsia took Gray. Notornis 53: 55–79. place before the late Oligocene–earliest Miocene, if not earlier. C ARLQUIST , S. 1977 . Wood anatomy of Onagraceae: Additional species Species of Fuchsia carrying the pollen type Koninidites aspis and concepts. Annals of the Missouri Botanical Garden 64 : 627 – 637 . appear to have been an important component of the New C ARPENTER , R. J. , J. M. BANNISTER , G . J . J ORDAN , AND D. E. LEE . 2010a . Leaf Zealand vegetation for ~25 million years since the Oligocene, fossils of Proteaceae tribe Persoonieae from the Late Oligocene–Early becoming extinct in the region by the early Pliocene, when they Miocene of New Zealand. Australian Systematic Botany 23 : 1 – 15 . C ARPENTER , R. J. , G. J. JORDAN , D. E. LEE , AND R. S. HILL . 2010b . Leaf were replaced by species with pollen indistinguishable from fossils of Banksia (Proteaceae) from New Zealand: An Australian modern taxa. abroad. American Journal of Botany 97 : 288 – 297 . C LOUT , M. N. , AND J. R. H AY . 1989 . The importance of birds as browsers, LITERATURE CITED pollinators and seed dispersers in New Zealand forests. New Zealand Journal of Ecology 12 ( supplement ): 27 – 33 . A RMESTO , J. J. , AND R . R OZZI. 1989 . Seed dispersal syndromes in the rain C ONRAN , J. G. , J. M. BANNISTER , AND D. E. LEE . 2009 . Earliest orchid forest of Chiloé: Evidence for the importance of biotic dispersal in a macrofossils: Early Miocene Dendrobium and Earina (Orchidaceae: temperate rain forest. Journal of Biogeography 16 : 219 – 226 . Epidendroideae) from New Zealand. American Journal of Botany 96 : A VERETT , J. E. , W. J. HAHN , P. E. BERRY , AND P. H. RAVEN . 1986 . Flavonoids 466 – 474 . and fl avonoid evolution in Fuchsia (Onagraceae). American Journal C ONRAN , J. G. , U. K AULFUSS , J. M. B ANNISTER , D. C. MILDENHALL , AND D. E. of Botany 73 : 1525 – 1534 . L EE. 2010 . Davallia (Polypodiales: Davalliaceae) macrofossils from B ALME , B. E. 1995 . Fossil in situ spores and pollen grains: An annotated Early Miocene Otago (New Zealand) with in situ spores. Review of catalogue. Review of Palaeobotany and Palynology 87 : 81 – 323 . Palaeobotany and Palynology 162 : 84 – 94 . B ANNISTER , J. M. , D. E. LEE , AND J . G . C ONRAN . 2012 . Lauraceae from C ONTI , E . , A . F ISCHBACH , AND K. J. SYTSMA . 1993 . Tribal relationships rainforest surrounding an early Miocene maar lake, Otago, southern in Onagraceae: Implications from rbc L sequence data. Annals of the New Zealand. Review of Palaeobotany and Palynology 178 : 13 – 34 . Missouri Botanical Garden 80 : 672 – 685 . B ANNISTER , J. M. , D. E. LEE , AND J. I. RAINE . 2005 . Morphology and palae- C OOMBS , D. S. , R. A. CAS , Y . K AWACHI , C . A . L ANDIS , W. F. MCDONOUGH , oenvironmental context of Fouldenia staminosa, a fossil fl ower with AND A . R EAY . 1986 . Cenozoic volcanism in north, east and Central associated pollen from the Early Miocene of Otago, New Zealand. Otago. In I. E. M. Smith [ed.], Late Cenozoic volcanism in New New Zealand Journal of Botany 43 : 515 – 525 . Zealand. Royal Society of New Zealand Bulletin 23 : 278 – 312 . B ARREDA , V. D. 1997 . Palinostratigrafía de la Formación San Julián en C OUPER , R. A. 1953 . Upper Mesozoic and Cainozoic spores and pol- el área de Playa La Mina (Provincia de Santa Cruz), Oligoceno de la len grains from New Zealand. Palaeontological Bulletin of the New Cuenca Austral. Ameghiniana 34 : 283 – 294 . Zealand Geological Survey 22 : 1 – 77 . B ARREDA , V. D. , AND S . P ALAMARCZUK . 2000a . Estudio palinoestratigrá- C OUPER , R. A. 1960 . New Zealand Mesozoic and Cainozoic plant micro- fi co integrado del entorno Oligoceno Tardío–Mioceno en secciones fossils. Palaeontological Bulletin of the New Zealand Geological de la costa patagónica y plataforma continental Argentina. In F. G. Survey 32 : 1 – 87 . Aceñolaza and R. Herbst [eds.], El Néogeno de Argentina, 103–138. C RISCI , J. V. , AND P. E. BERRY. 1990 . A phylogenetic reevaluation of the INSUEGO, Serie Correlación Geológica vol. 14, San Miguel de Old World species of Fuchsia (Onagraceae). Annals of the Missouri Tucumán, Argentina. Botanical Garden 77 : 517 – 522 . B ARREDA , V. D. , AND S . P ALAMARCZUK . 2000b . Palinoestratigrafía de depósi- D AGHLIAN , C. P. , J. J. SKVARLA , D . POCKNALL , AND P . H . R AVEN . 1985 . tos del Oligoceno tardío–Mioceno, en el área sur del Golfo San Jorge, Fuchsia pollen from the Early Miocene of New Zealand. American provincia de Santa Cruz, Argentina. Ameghiniana 37 : 103 – 117 . Journal of Botany 72 : 1039 – 1047 . B ARREDA , V. D. , L. PALAZZESI , AND S . A . M ARENSSI . 2009 . Palynological DE LANGE , P. J. , AND J. R. ROLFE . 2010 . New Zealand indigenous vas- record of the Paleogene Río Leona Formation (southernmost South cular plant checklist. New Zealand Plant Conservation Network, America): Stratigraphical and paleoenvironmental implications. Wellington, New Zealand. Review of Palaeobotany and Palynology 154 : 22 – 33 . D ELPH , L. F. 1989 . The evolution of fl oral color change: Pollinator attrac- B ERRY , P. E. 1982 . The systematics and evolution of Fuchsia sect. tion versus physiological constraints in Fuchsia excorticata. Evolution Fuchsia (Onagraceae). Annals of the Missouri Botanical Garden 6 9 : 43 : 1252 – 1262 . 1 – 198 . D ELPH, L. F., AND C . M . L IVELY. 1985 . Pollinator visits to fl oral colour phases B ERRY , P. E. 1985 . The systematics of the apetalous fuchsias of South of Fuchsia excorticata. New Zealand Journal of Zoology 12 : 599 – 603 . America, Fuchsia sect. Hemsleyella (Onagraceae). Annals of the D RISKELL , A. C. , AND L . C HRISTIDIS . 2004 . Phylogeny and evolution of Missouri Botanical Garden 72 : 213 – 251 . the Australo-Papuan honeyeaters (Passeriformes, Meliphagidae). B ERRY , P. E. , W. J. HAHN , K. J. SYTSMA , J. C. HALL , AND A . M AST . 2004 . Molecular Phylogenetics and Evolution 31 : 943 – 960 . Phylogenetic relationships and biogeography of Fuchsia (Onagraceae) D RISKELL , A. , L. CHRISTIDIS , B . J . G ILL , W. E. BOLES , F. K. B ARKER , AND N. based on noncoding nuclear and chloroplast DNA data. American W. LONGMORE. 2007 . A new endemic family of New Zealand pas- Journal of Botany 91 : 601 – 614 . serine birds: Adding heat to a biodiversity hotspot. Australian Journal B ERRY , P. E. , J. J. SKVARLA , D. PARTRIDGE , AND M. K. MACPHAIL . 1990 . of Zoology 55 : 73 – 78 . Fuchsia pollen from the Tertiary of Australia. Australian Systematic E MENY , M. T. , R. G. POWLESLAND , I. M. HENDERSON , AND R. A. F ORDHAM . Botany 3 : 739 – 744 . 2009 . Feeding ecology of kererū ( Hemiphaga novaeseelandiae ) in B OCHENSKI , Z . , AND Z. M. BOCHENSKI . 2008 . An Old World hummingbird podocarp-hardwood forest, Whirinaki Forest Park, New Zealand. New from the Oligocene: A new fossil from Polish Carpathians. Journal Zealand Journal of Ecology 33 : 114 – 124 . fur Ornithologie 149 : 211 – 216 . E RDTMAN , G. 1952 . Pollen morphology and plant taxonomy–angiosperms B OLES , W. E. 2005 . Fossil honeyeaters (Meliphagidae) from the Late (An introduction to palynology, vol. I). Chronica Botanica, Waltham, Tertiary of Riversleigh, north-western Queensland. Emu 105 : 21 – 26 . Massachusetts, USA. 2064 AMERICAN JOURNAL OF BOTANY [Vol. 100

E YDE , R. H. 1982 . Evolution and systematics of the Onagraceae: Floral Cenozoic angiosperm fl ora in relation to palaeogeography and cli- anatomy. Annals of the Missouri Botanical Garden 69 : 735 – 747 . mate. Australian Journal of Botany 49 : 341 – 356 . F ERGUSON , D. K. , D. E. LEE , J. M. BANNISTER , R . Z ETTER , G. J. JORDAN , N . L EE , D. E. , R. M. M CDOWALL , AND J. K. LINDQVIST . 2007 . Galaxias fossils V AVRA , AND D. C. MILDENHALL . 2010 . The taphonomy of a remark- from Miocene lake deposits, Otago, New Zealand: the earliest records able leaf bed assemblage from the Late Oligocene–Early Miocene of the Southern Hemisphere family Galaxiidae (Teleostei). Journal of Gore Lignite Measures, southern New Zealand. International Journal the Royal Society of New Zealand 37 : 109 – 130 . of Coal Geology 83 : 173 – 181 . L EOPOLD , E. B. , S. R. MANCHESTER , AND H. W. M EYER . 2008 . Phytogeography F OSTER , C. B. 1982 . Illustrations of early Tertiary (Eocene) plant micro- of the Late Eocene Florissant fl ora reconsidered. In H. W. Meyer and fossils from the Yaamba Basin, Queensland. Geological Survey of D. M. Smith [eds.], Paleontology of the Upper Eocene Florissant Queensland publication no. 381, Queensland Department of Mines, Formation, Colorado. Geological Society of America Special Papers Brisbane, Australia. 435, 53–70. Geological Society of America, Boulder, Colorado, USA. G ODLEY , E. J. 1955 . Breeding systems in New Zealand plants. I. Fuchsia. L EVIN , R . A . , W . L . W AGNER , P . C . H OCH , W . J . H AHN , A . R ODRIGUEZ , Annals of Botany 19 : 549 – 559 . D. A. BAUM , L . K ATINAS , E. A. Z IMMER , AND K. J. SYTSMA . 2004 . G ODLEY , E. J. , AND P. E. BERRY . 1995 . The biology and systematics of Paraphyly in Tribe Onagreae: Insights into phylogenetic relation- Fuchsia in the South Pacifi c. Annals of the Missouri Botanical Garden ships of Onagraceae based on nuclear and chloroplast sequence data. 82 : 473 – 516 . Systematic Botany Monographs 29 : 147 – 164 . G OLOBOFF , P. A. , J. S. FARRIS , AND K . C . N IXON . 2008 . TNT, a free pro- L INDQVIST, J. K., AND D . E . L EE. 2009 . High-frequency paleoclimate signals gram for phylogenetic analysis. Cladistics 24 : 774 – 786 . from Foulden Maar, Waipiata Volcanic Field, southern New Zealand: G RÍMSSON , F . , R . Z ETTER , AND Q . L ENG . 2011 . Diverse fossil Onagraceae An Early Miocene varved lacustrine diatomite deposit. Sedimentary pollen from a Miocene palynofl ora of north-east China: Early steps in Geology 222 : 98 – 110 . resolving the phytogeographic history of the family. Plant Systematics L LOYD , D. G. 1985 . Progress in understanding the natural history of and Evolution 298 : 671 – 687 . New Zealand plants. New Zealand Journal of Botany 23 : 707 – 722 . H EYWOOD , V. H. , R. K. BRUMMITT , A . C ULHAM , AND O . S EBERG . 2007 . M ABBERLEY , D. J. 2008 . Mabberley’s plant-book. A portable dictionary Flowering plant families of the world. Royal Botanic Gardens, Kew, of plants, their classifi cations, and uses, 3rd ed. Cambridge University UK. Press, Cambridge, UK. H IGGINS , P. J. , AND J. M. PETER. 2002 . Handbook of Australian, New M ACPHAIL , M . , AND D. J. CANTRILL . 2006 . Age and implications of the Zealand and Antarctic birds, vol. 6, Pardalotes to Shrike-thrushes. Forest Bed, Falkland Islands, southwest Atlantic Ocean: Evidence Oxford University Press, Melbourne, Australia. from fossil pollen and spores. Palaeogeography, Palaeoclimatology, H IGGINS , P. J. , J. M. PETER , AND S. J. COWLING . 2006 . Handbook of Palaeoecology 240 : 602 – 629 . Australian, New Zealand & Antarctic birds, vol. 7, Boatbill to star- M ACPHAIL , M. K. 1999 . Palynostratigraphy of the Murray Basin, inland lings. Oxford University Press, Melbourne, Australia. southeastern Australia. Palynology 23 : 197 – 240 . J ANSONIUS , J . , AND L. V. HILLS . 1976 . Genera fi le of fossil spores. Special M ANCHESTER , S. R. 2001 . Update on the megafossil fl ora of Florissant, Publication, Department of Geology, University of Calgary, Calgary, Colorado. In E. Evanoff, K. M. Gregory-Wodzicki, and K. R. Johnson Alberta, Canada. [eds.], Fossil fl ora and stratigraphy of the Florissant Formation, J ORDAN , G. J. , J. M. BANNISTER , D. C. MILDENHALL , R . Z ETTER , AND D. E. Colorado. Proceedings of the Denver Museum of Nature and Science, L EE . 2010 . Fossil Ericaceae from New Zealand: Deconstructing the series 4, no. 1, 137–161. Denver Museum of Nature & Science, use of fossil evidence in historical biogeography. American Journal Denver, Colorado, USA. of Botany 97 : 59 – 70 . M ARTIN , H. A. 2003 . The history of the family Onagraceae in Australia J ORDAN , G. J. , AND R. S. HILL . 1999 . The phylogenetic affi nities of and its relevance to biogeography. Australian Journal of Botany 51 : Nothofagus (Nothofagaceae) leaf fossils based on combined molecu- 585 – 598 . lar and morphological data. International Journal of Plant Sciences M CEWAN , W. M. 1978 . The food of the New Zealand pigeon ( Hemiphaga 160 : 1177 – 1188 . novaeseelandiae novaeseelandiae ). New Zealand Journal of Ecology K AULFUSS , U . , T . W APPLER , E . H EISS , AND M.-C. LARIVIÈRE . 2011 . Aneurus 1 : 99 – 108 . sp. from the early Miocene Foulden Maar, New Zealand: The fi rst M CGLONE , M. S. , R. J. DUNGAN , G. M. J. HALL , AND R . B . A LLEN . 2004 . Southern Hemisphere record of fossil Aradidae (Insecta: Hemiptera: Winter leaf loss in the New Zealand woody fl ora. New Zealand Heteroptera). Journal of the Royal Society of New Zealand 41 : Journal of Botany 42 : 1 – 19 . 279 – 285 . M EYLAN, B. A., AND B. G. B UTTERFIELD . 1978 . The structure of New Zealand K EDVES , M. 1986 . Paleogene fossil sporomorphs of the Bakony Mountains woods. New Zealand Department of Scientifi c and Industrial Research IV. Studia Biologica Hungarica 21 : 1 – 120 . Bulletin 222 : 1 – 250 . K ELLY , D. , J. J. LADLEY , A. W. ROBERTSON , S . H. ANDERSON , D. M. WOTTON , M ILDENHALL , D. C. 1980 . New Zealand Late Cretaceous and Cenozoic plant AND S. K. WISER . 2010 . Mutualisms with the wreckage of an avifauna: biogeography: A contribution. Palaeogeography, Palaeoclimatology, The status of bird pollination and fruit dispersal in New Zealand. Palaeoecology 31 : 197 – 233 . New Zealand Journal of Ecology 34 : 66 – 85 . M ILDENHALL , D. C. 1993 . Last glaciation/postglacial pollen record for K ING , P . , L . M ILICICH , AND K. C. BURNS . 2011 . Body size determines Porirua, near Wellington, New Zealand. Tuatara 32 : 22 – 27 . rates of seed dispersal by giant king crickets. Population Ecology 53 : M ILDENHALL , D. C. , AND D. T. POCKNALL . 1989 . Miocene–Pleistocene 73 – 80 . spores and pollen from Central Otago, South Island, New Zealand. K IRCHNER, W. C. G. 1898 . Contributions to the fossil fl ora of Florissant, New Zealand Geological Survey Palaeontological Bulletin 59 : 1 – 128 . Colorado . Transactions of the Academy of Science of St. Louis 8 : M ONTGOMERY , B. R. , D. K ELLY , AND J. J. LADLEY . 2001 . Pollinator limi- 161 – 188 , plates 111–115. tation of seed set in Fuchsia perscandens (Onagraceae) on Banks L EE , D. E. , J. M. BANNISTER , J. I. RAINE , AND J. G. CONRAN . 2010 . Peninsula, South Island, New Zealand. New Zealand Journal of Euphorbiaceae: Acalyphoideae fossils from Early Miocene New Botany 39 : 559 – 565 . Zealand: Mallotus–Macaranga leaves, fruits, and infl orescence with M OORFIELD , J. C. 2011 . Te Aka Māori–English, English–Māori dictionary in situ Nyssapollenites endobalteus pollen. Review of Palaeobotany and index, 3rd ed. Pearson Education, New Zealand, Auckland, New and Palynology 163 : 127 – 138 . Zealand. L EE , D. E. , J. G. CONRAN , J. K. LINDQVIST , J. M. BANNISTER , AND D. C. N EALL , V. E. , AND S. A. TREWICK. 2008 . The age and origin of the Pacifi c M ILDENHALL. 2012 . New Zealand Eocene, Oligocene and Miocene Islands: A geological overview. Philosophical Transactions of the macrofossil and pollen records and modern plant distributions in the Royal Society, B, Biological Sciences 363 : 3293 – 3308 . Southern Hemisphere. Botanical Review 78 : 235 – 260 . N ELSON , C. S. , D. C. MILDENHALL , A. J. TODD , AND D. T. POCKNALL . 1988 . L EE , D. E. , W. G. LEE , AND N . M ORTIMER . 2001 . Where and why have Subsurface stratigraphy, paleoenvironments, palynology, and depo- all the fl owers gone? Depletion and turnover in the New Zealand sitional history of the late Neogene Tauranga Group at Ohinewai, October 2013] LEE ET AL.—NEW ZEALAND FOSSIL FUCHSIA 2065

Lower Waikato Lowland, South Auckland, New Zealand. New systems, and fl avonoids in Fuchsia sect. Skinnera (Onagraceae). Zealand Journal of Geology and Geophysics 31 : 21 – 40 . Systematic Botany 16 : 257 – 269 . N IXON , K. C. 2004 . ASADO version 1.89. Program and documentation T ENNYSON , A. J. D. , R. P. SCOFIELD , AND T. H. WORTHY . 2011 . Vertebrate distributed by the author, Ithaca, New York. survivors of Zealandia’s Oligocene “drowning”. In N. J. Litchfi eld N OWICKE , J. W. , J. J. SKVARLA , P. H. RAVEN , AND P. E. BERRY . 1984 . A and K. Clark [eds.], Abstract volume, Geosciences 2011 Conference, palynological study of the genus Fuchsia (Onagraceae). Annals of the Nelson, New Zealand. Geoscience Society of New Zealand Missouri Botanical Garden 71 : 35 – 91 . Miscellaneous Publication 130A, 107. Geoscience Society of New O ’ D ONNELL , C. F. J. , AND P. J. DILKS . 1994 . Foods and foraging of for- Zealand, Nelson, New Zealand. est birds in temperate rainforest, South Westland, New Zealand. New T OBE , H . , P. H. RAVEN , AND C.-I. PENG . 1988 . Seed coat anatomy and Zealand Journal of Ecology 18 : 87 – 107 . relationships of Ludwigia sects. Microcarpium , Dantia , and Miquelia P ERROTT , J. K. , AND D. P. ARMSTRONG . 2000 . Vegetation composition and (Onagraceae), and notes on fossil seeds of Ludwigia from Europe. phenology of Mokoia Island, and implications for the reintroduced Botanical Gazette 149 : 450 – 457 . hihi population. New Zealand Journal of Ecology 24 : 19 – 30 . T RAVESET , A . , M . F . W ILLSON , AND C . S ABAG . 1998 . Effect of nectar-rob- P OCKNALL , D. T. 1982 . Palynology of Late Oligocene Pomahaka Bed sed- bing birds on fruit set of Fuchsia magellanica in Tierra Del Fuego: A iments, Waikoikoi, Southland, New Zealand. New Zealand Journal of disrupted mutualism. Functional Ecology 12 : 459 – 464 . Botany 20 : 263 – 287 . VAN DER HAMMEN , T. 1954 . El desarrollo de la fl ora colombiana en los P OCKNALL , D. T. 1985 . Palynology of Waikato Coal Measures (Late periodos geológicos. I, Maestrichtiano hasta terciario más inferior Eocene–Late Oligocene) from the Raglan area, North Island, New (una investigación palinológica de la formación Guaduas y equiva- Zealand. New Zealand Journal of Geology and Geophysics 28 : lentes). Boletíno Geológíco (Bogotá) 2 : 49 – 106 . 329 – 349 . VAN DER HAMMEN , T. 1956 . A palynological systematic nomenclature. P OCKNALL , D. T. 1989 . Late Eocene to Early Miocene vegetation and cli- Boletíno Geológíco (Bogotá) 4 : 63 – 101 . mate history of New Zealand. Journal of the Royal Society of New V ARMA , C. P. , AND M . S . R AWAT . 1963 . A note on some diporate grains Zealand 19 : 1 – 18 . recovered from Tertiary horizons of India and their potential marker P OCKNALL , D. T. , AND D. C. MILDENHALL . 1984 . Late Oligocene–Early value. Grana Palynologica 4 : 130 – 139 . Miocene spores and pollen from Southland, New Zealand. New W AGNER, W. L., P. C. H OCH , AND P . H . R AVEN. 2007 . Revised clas- Zealand Geological Survey Palaeontological Bulletin 51 : 1 – 66 . sifi cation of the Onagraceae. Systematic Botany Monographs 8 3 : P OINAR , G . , C . V OISIN , AND J.-F. VOISIN . 2007 . Bird eggshell in Dominican 1 – 243 . amber. Palaeontology 50 : 1381 – 1383 . W HITAKER , R. D. 1987 . The roles of lizards in New Zealand plant repro- P OLE , M. S. 1996 . Plant macrofossils from the Foulden Hills Diatomite ductive strategies. New Zealand Journal of Botany 25 : 315 – 328 . (Miocene), Central Otago, New Zealand. Journal of the Royal Society W ILLIAMS , C. A. , AND P. J. GARNOCK-JONES . 1986 . Leaf fl avonoids and of New Zealand 26 : 1 – 39 . other phenolic glycosides and the taxonomy and phylogeny of Fuchsia P OTONIÉ , R. 1960 . Synopsis der Gattungen der Sporae dispersae. III. Teil: sect. Skinnera (Onagraceae). Phytochemistry 25 : 2547 – 2549 . Nachträge Sporites, Fortsetzung Pollenites. Mit Generalregister zu W ILLIAMS , P. A. , AND B . J . K ARL. 1996 . Fleshy fruits of indigenous and Teil I–III. Beihefte zum Geologischen Jahrbuch 39 : 1 – 189 . adventive plants in the diet of birds in forest remnants, Nelson, New P UNT , W . , J . R OVERS , AND P. P. HOEN . 2003 . The northwest European pol- Zealand. New Zealand Journal of Ecology 20 : 127 – 145 . len fl ora, 67. Onagraceae. Review of Palaeobotany and Palynology W INGATE , F. H. , AND D. J. NICHOLS . 2001 . Palynology of the upper- 123 : 107 – 161 . most Eocene lacustrine deposits at Florissant National Monument, R AINE , J. I. , D. C. MILDENHALL , AND E. M. K ENNEDY . 2011 . New Zealand Colorado. In E. Evanoff, K. M. Gregory-Wodzicki, and K. R. Johnson fossil spores and pollen: An illustrated catalogue, 4th ed. GNS [eds.], Fossil fl ora and stratigraphy of the Florissant Formation, Science Miscellaneous Series No. 4. Website http://www.gns.cri.nz/ Colorado. Proceedings of the Denver Museum of Nature and Science, what/earthhist/fossils/spore_pollen/catalog/index.htm [accessed 20 Series 4, no. 1, 71–135. Denver Museum of Nature & Science, Denver, April 2012]. Colorado, USA. R AMÍREZ-ARRIAGA , E. , M. B. PRÁMPARO , E . M ARTÍNEZ-HERNÁNDEZ , AND W OOD , J. R. , J. M. WILMSHURST , S. J. WAGSTAFF , T . H . W ORTHY , N. J. A . V ALIENTE-BAUNET. 2006 . Palynology of the Paleogene Cuayuca R AWLENCE , AND A . C OOPER . 2012 . High-resolution coproecol- Formation (stratotype sections), southern Mexico: Chronostratigraphical ogy: Using coprolites to reconstruct the habits and habitats of New and palaeoecological implications. Review of Palaeobotany and Zealand’s extinct Upland Moa (Megalapteryx didinus ). PLoS ONE Palynology 141 : 259 – 275 . 7 : e40025 . R AVEN , P. H. , AND D. I. AXELROD . 1972 . Plate tectonics and Australasian W ORTHY , T. H. , S. J. HAND , J. P. WORTHY , A. J. D. TENNYSON , AND R . P AUL paleobiogeography. Science 176 : 1379 – 1386 . SCOFIELD . 2009 . A large fruit pigeon (Columbidae) from the Early R OBERTSON , A. W. , J. J. LADLEY , D. KELLY , K. L. MCNUTT , P. G. PETERSON , Miocene of New Zealand. Auk 126 : 649 – 656 . M. F. MERRETT , AND B. J. KARL . 2008 . Assessing pollination and fruit W YMAN , T. E. , S. A. TREWICK , M . M ORGAN-RICHARDS , AND A. D. L. NOBLE . dispersal in Fuchsia excorticata (Onagraceae). New Zealand Journal 2011 . Mutualism or opportunism? Tree fuchsia (Fuchsia excorti- of Botany 46 : 299 – 314 . cata ) and tree weta (Hemideina ) interactions. Austral Ecology 36 : R OUSE , G. E. 1962 . Plant microfossils from the Burrard Formation of 261 – 268 . western British Columbia. Micropaleontology 8 : 187 – 218 . X IE , L . , W. L. WAGNER , R. H. R EE , P. E. BERRY , AND J . W EN . 2009 . S TEADMAN , D. W. 2006 . Extinction and biogeography of tropical Pacifi c Molecular phylogeny, divergence time estimates, and historical bio- birds. University of Chicago Press, Chicago, Illinois, USA. geography of Circaea (Onagraceae) in the Northern Hemisphere. S USSMAN , R. W. , AND P. H. RAVEN. 1978 . Pollination by lemurs and mar- Molecular Phylogenetics and Evolution 53 : 995 – 1009 . supials: An archaic coevolutionary system. Science 200 : 731 – 736 . Z ETTER , R. , C.-C. HOFMANN , I . D RAXLER , J . D URANGO DE CABRERA , M. S YTSMA , K. J. , A. LITT , M. L. ZJHRA , J . C . P IRES , M . N EPOKROEFF , E . CONTI , D EL , M . V ERGAL , AND F . V ERVOOSRT . 1999 . A rich Middle Eocene J . W ALKER , AND P. G. WILSON . 2004 . Clades, clocks, and continents: microfl ora at Arroyo de los Mineros, near Cañadón Beta, NE Tierra Historical and biogeographical analysis of Myrtaceae, Vochyiaceae, del Fuego Province, Argentina. Abhandlungen der Geologischen and relatives in the Southern Hemisphere. International Journal of Bundesanstalt 56 : 439 – 460 . Plant Sciences 165 : S85 – S105 . Z ETTER , R . , AND C . K ERI. 1987 . New details on the morphology of fos- S YTSMA , K. J. , J. F. SMITH , AND P . E . B ERRY . 1991 . The use of chloroplast sil onagraceous pollen grains. Plant Systematics and Evolution 157 : DNA to assess biogeography and evolution of morphology, breeding 1 – 7 .