Dragons in the mist An interview with Steve Wagstaff, evolutionary biologist at Landcare Research

David Glenny1

Introduction Steve: The aims Dr Steve Wagstaff came to New of the project Zealand in 1994 from the USA as the were to defi ne the fi rst botanist in to have underlying reasons both classical taxonomic training and for the differences expertise in the then relatively new in species richness technique of DNA sequencing. Steve between New has used DNA sequencing to solve Zealand, Australia, many of the long-standing problems and New Caledonia in the of the New Zealand (see Box 1, p. 18). fl owering and conifer fl ora. He Q: Which countries has now published work on Hebe and have the most several groups of daisies including species? Brachyglottis, Celmisia, and Olearia. His results often tell stories about the Steve: New Zealand biogeography of New Zealand : has by far the where they came from, how long most species of they have been in New Zealand, how Dracophyllumm (35 often and when they dispersed from spp.), followed by New Zealand to neighbouring islands Fig. 1 The world distribution of and its near New Caledonia (9 or continents. Steve’s work, along relatives, and Sphenotoma. spp.), and Tasmania with that of others in the same fi eld, resemblance of the leaves to those of to a lesser degree (2 spp.). We has changed our understanding of the dragon-tree ( draco), a compared islands that are species the origins of the New Zealand fl ora. monocot tree of central America that rich with the Australian continent (4 As the interview below illustrates, it is like our cabbage trees (Cordyline spp.). We wanted to determine the seems the majority of our fl owering species). phylogeny, to fi nd the direction of plants evolved from ancestors that evolution. arrived in New Zealand about the late Steve’s fi eld work was supported by Q: What do you mean by direction of Pliocene (less than 5 million years the National Geographic Committee evolution? ago) to early Pleistocene (about 1.6 for Research and Exploration to study mya), often from Australia. The phrase Dracophyllumm throughout its entire Steve: The direction of dispersal ‘Goodbye Gondwana’ (McGlone, range. This involved fi eld work in New – where the Dracophyllum 2005) sums up this rejection of the Zealand, Australia including Tasmania, originated and where it dispersed to. view of New Zealand as being an ‘ark’ Lord Howe Island, and New Caledonia Also, we wanted to fi nd out why there carrying the remnants of an ancient to collect specimens of most of the are so many species in New Zealand fl ora. This new view is mainly due Dracophyllumm species and many relative to Australia. Dracophyllum to the results of DNA sequencing species of related genera (Fig. 1). is paraphyletic3 and could include studies. the genus Richea, and if you do Interview this, then the number of species of A recent project, now complete and Dracophyllumm is not so imbalanced soon to be published, was a study Q: You have just completed this in Australia and New Zealand. So of the genus Dracophyllum, a tree project on Dracophyllum. What was the imbalance might be an artefact and shrub genus best known to New the topic of this and why did National of the taxonomy. We wanted to fi nd Zealanders as turpentine scrub, Geographic fund you to do the fi eld out how old the lineages are in each pineapple scrub, or Dr Seuss trees. work? country. New Zealand species might The name Dracophyllumm means have had longer to speciate, or the dragon-leaves and comes from the Australian lineage might be older

1 Landcare Research, PO Box 40, Lincoln 7640; [email protected]. 2 See the next article by Dawson and Winkworth (pp. 18–23). 3 A paraphyletic group is one that is incomplete, i.e., it contains only some of the species that are descendants of the root ancestor species. See the previous article by Glenny and Dawson (pp. 10–12).

New Zealand Garden Journal, 2008, Vol. 11(1) 14 but extinction there might have been make fi eld work in New Caledonia National Park. The second trip was greater. Rates of speciation and possible. Murray Dawson assisted a conference fi eld trip led by Greg extinction were issues to address. To with fi eld work in New Zealand, and Jordan, a paleobotanist in Tasmania. get at these issues we had to study in the laboratory did all the DNA He’s studied fossils of Richeaa and the larger group that Dracophyllum extraction and sequencing work. took us to three fossil sites and also belongs to outside New Zealand. showed us sites with living Richea It was critical to include all species (Fig. 5) and Dracophyllum. from the wider regional perspective. That’s partly why we were funded by National Geographic – they saw the importance of doing fi eld work outside of New Zealand. Q: How much did you know about the group when you started and what put the idea in your head to do this project? Steve: We already knew quite a bit about the genus. One major reason we got the National Geographic fund is that we had a strong team in all the countries we worked in. International collaboration produces the best Fig. 3 , a New science, and besides, it’s more fun Zealand species in fl ower (Photo: ©Paul Fig. 5 on Mt Field, Tasmania. that way. Ashford, www.NZplantpics.com). Q: You did some interesting fi eld work. And fi nally I organised two trips to Where was this? Who did you do the eastern Australia. The fi r st started fi eld work with? at Sydney where we collected a Dracophyllumm species on coastal Steve: I organised two trips to cliffs, then went up to the New South Western Australia with Kristina Wales – Queensland border, and Lemson of Perth to collect then south to Melbourne. A lot of road Sphenotomaa throughout its range. travelling. It only occurs in the forests of South West Australia south of Perth, like Q: Who did you do that trip with? in the Stirling Range National Park Steve: I hired a research fellow at north-east of Albany. Sphenotoma Sydney Botanic Gardens, Andrew is a genus of 7 or 8 species, and Perkins, who had just fi nished his Kristina showed me what she believes PhD and was doing various pieces of to be an undescribed species on the contract work—he’d just published a Fig. 2 Dracophyllum elegantissimum in the summit of the Stirling Range (Fig. 4). book on orchids. Brunner Range, New Zealand. Fanie Venter discovered this new tree species in Nelson Then Fanie Venter and I travelled to province while working on his PhD (Venter, northern Queensland, and to Lord 2004a). Howe Island. Dracophyllum sayeri Fanie Venter was working on a is found only on a few peaks in the revision of the genus for his PhD at Bellenden Ker Range south of Cairns Victoria University, Wellington, and where it’s a forest gap species on the had samples from his work of most peaks of the range. Dracophyllum of the New Zealand species (Fig. 2, fi tzgeraldiii is only on the summit of 3). He had studied specimens from Lord Howe Island (Fig. 6). overseas but had done no fi eld work there. Darren Crayn in Canberra had completed a DNA sequencing survey of the Epacridoideae and contributed sequences from this Fig. 4 A beautiful undescribed Sphenotoma species on the summit ridge of the Stirling earlier study. Kristina Lemson of Perth Range, Western Australia. was preparing a fl ora treatment of a closely related genus, Sphenotoma. There were two trips to Tasmania, Dorothy Steane I had known before I the fi rst trip with Dorothy Steane in came to New Zealand, and she is an autumn when the Dracophyllumm and expert on the Tasmanian fl ora. Jérôme Richea species were in fl ower or fruit, to a number of sites like Mt Field Fig. 6 Steve Wagstaff on Lord Howe Island. Munzinger is writing treatments for The summit of the island, the sole habitat of the New Caledonian fl ora; he helped National Park and Cradle Mountain Dracophyllum fi tzgeraldii, is in the distance.

New Zealand Garden Journal, 2008, Vol. 11(1) 15 Q: How did you get onto those training in recognition of plant families, so there was no point continuing to peaks? but the New Caledonian fl ora was sequence very similar New Zealand so exotic I was right out of my depth species. Steve: Well, as it happened, only trying to identify what families the a week before we went, Cyclone We included all Australian species, all plants were in. After the New Zealand Larry went through Queensland and New Caledonian species, and 60% of fl ora with mostly small white fl owers the national parks and many roads the New Zealand species. it was amazing to see these huge, were closed. When we got there, we red, bird-pollinated fl owers. Fanie Q: Did the results give you the were told there was a cable car that and I climbed Mt Humboldt, the information you needed to solve the went to the TV transmitter tower on highest point on the main island. We questions you’d asked? Mt Bellenden Ker. We contacted the traversed through Araucaria forests, studio executives, and said we were Steve: The results were not very eerie in fog that was coming and visiting scientists. They rolled out completely resolved, not as much going (Fig. 7). At one point below the the red carpet for us, much to the as we’d have liked. We were able to summit we walked through a 10 m annoyance of the cable car operators answer the questions we had, but not high Dracophyllum forest with a tree who had just got it working again with as much certainty as we’d have fern understorey. That particular and wanted to get supplies up to the liked. And we found out some other Dracophyllumm has been regarded as a transmitter. We had half a day on the things along the way (Box 2, p.18). variety of Dracophyllum ramosum, but summit. The forest around the summit Fanie considers it a distinct species Q: What did the results say about the was completely devastated, torn tree called Dracophyllum thiebautii (Fig. 8). questions you were interested in? trunks, jagged branch ends, the forest At the summit, the fog cleared and we was shredded. Steve: The results told us how old the got phenomenal views of the island in Dracophyllumm lineage is. It separated Q: You also went to New Caledonia? every direction. from other genera of Epacridoideae The maquis, sparse shrublands on 20 million years ago (mya), about ultramafi c rock (igneous rocks that are early Miocene (Fig. 10). very poor in silica but rich in iron and They told us that Dracophyllumm in magnesium), was some of the hardest New Zealand is younger than in walking I’ve ever done because of the Australia, and that there was one, slippery clay soil – it’s like standing possibly two, dispersals to New on ice. Fanie had already described Zealand; the results are ambiguous Dracophyllum mackeeanumm as a new on that, but my hunch is that there species from herbarium specimens Fig. 7 Fanie Venter in Araucaria humboldtii was a single dispersal event. This forest near the summit of Mont Humboldt, but had never seen it in the wild, and occurred about 2.2 mya, at the New Caledonia. This is as tall as the trees we found it on the maquis, in a distinct beginning of the Pleistocene era. get. habitat that confi rmed it as being This corresponds to the time when a distinct species (Fig. 9). It turned the mountains in New Zealand out to be more widespread than the began to uplift, and the Pleistocene herbarium specimens indicated. glaciations started. We didn’t know Q: Did you manage to sample all before the study if Dracophyllumm in species in the genus? New Caledonia had its origin from New Zealand or Australia. The results Steve: We sampled most species, but showed that there was a single the experimental design only required dispersal to New Caledonia from single representatives of each Australia, about 7 mya. Fig. 8 Dracophyllum thiebautiii in cloud, species. Our aim was to do a broad near the summit of Mont Humboldt, New Subgenus Oreothamnuss contains the Caledonia. survey, to allow calibration of the tree with distantly related genera for small-leaved Dracophyllumm species which there are fossils with dates. We and is very morphologically diverse in used gene regions called rbcL and New Zealand with 28 species ranging mattK – these are fairly conservative from tall shrubs to cushion plants. regions that allowed sampling across However, within this subgenus there the entire subfamily that Dracophyllum is very little genetic diversity and belongs to. species with quite different forms are able to hybridise. Its greatest species The trade off in using conservative diversity is in the eastern Southern gene regions was that many of the Alps (Fig. 11). It came as a surprise sequences were identical, especially Fig. 9 Dracophyllum mackeeanum, on to discover that DNA sequence in the New Zealand species belonging the ultramafi c maquis of New Caledonia, a evidence suggests that the diminutive species described recently by Fanie Venter to subgenus Oreothamnus. Once we Tasmanian species, D. minimum, is (Venter, 2004b). saw identical sequences we could see misplaced in subgenus Oreothamnus that we had sampled enough of the Steve: Yes, with Fanie Venter. For me, meaning that the subgenus may now New Zealand subgenus. Sequencing New Caledonia was the highlight of be confi ned to New Zealand and its is expensive and time-consuming all the fi eld work. I have a traditional offshore islands. However, this result

New Zealand Garden Journal, 2008, Vol. 11(1) 16 appears to Enkianthus fl owers were used. Dated contradict Fanie fossils of Rhododendronn and Richea Vener’s PhD leaves were also available and the work based on time of emergence of Lord Howe morphology. Island was known. Fossil seeds of Rhododendron are known from 60 Subgenus mya and fossil pollen from 50 mya. Oreothamnus The fi rst appearance of Richeaa leaves diverged from in the fossil record was from 35–30 the rest of mya. Lord Howe Island is the remnant Dracophyllum of a large volcano that erupted (subgenus 6.9 mya. Each gives a different Dracophyllum) in perspective: fossils set a lower age New Zealand, 1.2 limit, while the emergence of an island mya. sets an upper age limit. We used the Dispersal of single calibration point of Enkianthus Dracophyllumm to and checked that the other fossils and Lord Howe Island, the emergence of Lord Howe Island from eastern were consistent with that, and got that Australia occurred consistency pretty well. just after the Q: What do your results say about emergence of the the imbalance of species between the island 6.9 mya. islands and continent of Australia? Q: How do you Steve: Most of the species diversity is arrive at the dates in New Zealand and New Caledonia. you mentioned? But interestingly, most of the genetic Steve: We diversity is in Australia. There are detected a smaller number of species in differences in Australia, but they are the oldest the DNA isolated lineages and so we suggest that from the species there have been more extinctions in in our study. DNA Australia. mutations are There was dispersal to New Zealand passed on from in much more recent times and this one generation to was followed by a large species the next. When we Fig. 10 Chronogram showing the inferred evolutionary history of radiation that happened mainly in the study a species’ Dracophyllum and its relatives. A = the calibration point using the New Zealand mountains. dated Enkianthuss fossil 90 mya; B = the point at which Southern DNA profi le, a and Northern Hemisphere diverged, 40.5 mya; C = mutation shared Q: These extinctions in Australia – the point at which the tribe containing Dracophyllum diverged, why would that have happened? 37.8 mya; D = New Zealand and New Caledonian species of by two or more species suggests Dracophyllum diverged from eastern Australian species 7.5 mya; Steve: The three related genera, E = the start of the species radiation of Dracophyllum subg. they are related Dracophyllum, Richea, and Oreothamnuss in New Zealand, 1.2 mya. Geographic regions are to one another by highlighted in colour and an evolutionary time-scale is shown at the Sphenotoma, may have been descent. bottom. widespread in Australia in the We assume most Miocene at which time temperate of these changes are ‘silent’, not forests were prevalent in Australia. expressed in the morphology of the Then as a result of climate change plants, and that they are also evolving at the end of the Tertiary, Australia in a clock-like manner (i.e., the rate became a lot drier and the great at which those point mutations occur inland desert formed. The temperate is uniform over time). To calibrate the forests shrank to a few localities on DNA molecular clock, we use the the eastern and western edges of the fossil record (or the known date of continent. The genus Sphenotoma emergence of a volcanic island), and became confi ned to a small area Fig. 11 Dracophyllum rosmarinifolium on determine the rate at which mutations of mountains at the southern tip of Porters Pass, Canterbury. The dry eastern are occurring in the group under Western Australia. Dracophyllumm for ranges are habitat for most of the shrubby Dracophyllum species that make study. the most part became confi ned to the up the New Zealand Dracophyllum subgenus mountains of eastern Australia and Q: How did you calibrate the Oreothamnus. Dracophyllum rosmarinifolium Tasmania. Dracophyllumm likes damp Dracophyllum phylogeny? has been known as D. unifl orum, but Fanie cloud forest. Most Australian species Venter found these names refer to the same species and D. rosmarinifolium is the older of Steve: Using the fossil record. Ninety are now confi ned to small areas on the two names (and thus takes priority). million year-old fossils of Australian the summits of mountain ranges like

New Zealand Garden Journal, 2008, Vol. 11(1) 17 the Blue Mountains in New South Many groups are morphologically References Wales, also in protected valleys in the diverse: Dracophyllumm for example McGlone, M.S. (2005). Goodbye area where the famous Wollemi pine varies from trees down to small Gondwana. Journal of was discovered. Dracophyllum sayeri cushion plants with nearly identical Biogeography 32: 739–740. is only on a few peaks in tropical sequences for the gene regions Venter, S. (2004a). Dracophyllum Queensland. The Lord Howe Island we’ve looked at. elegantissimumm (Ericaceae), a species, Dracophyllum fi tzgeraldii, is new species from north-west Q: What about the disparity in species only on the summit of the island. Nelson, New Zealand. New richness between Australia and New Zealand Journal of Botany 42: Getting back to the question of the Zealand? Do you see that in other 37–43. imbalance in the species numbers, groups? Venter, S. (2004b). Dracophyllum the DNA results don’t convincingly Steve: I’m not sure about that, results mackeeanumm (Ericaceae: separate Richeaa from Dracophyllum. in other groups vary quite a bit. The Richeeae), a new species from If you include Richeaa in Dracophyllum, most diverse elements in Australia are New Caledonia. New Zealand then the imbalance in species Journal of Botany 42: 747–752. absent or rare in New Zealand and numbers is less marked. vice versa. For instance, Eucalyptuss is Further reading Q: So there are two explanations of very diverse in Australia but is absent Steve Wagstaff maintains a website the imbalance: one that it’s an artefact from the current New Zealand native called ‘Phylogeny of New Zealand of the taxonomy, and the other is fl ora (but fossil evidence shows that it plants’ which has evolutionary that it results from climate change was here once). In the other direction, trees and links to phylogeny at the beginning of the Pleistocene hebes and related genera are very literature (see http://plantphylogeny. with extinctions due to increasing diverse in New Zealand but Australia landcareresearch.co.nz/). Major aridity in Australia, at the same time has very few species. papers that Steve has coauthored are as new cold-climate and high altitude listed below: The Australian fl ora is highly adapted environments appeared in New to the arid climate and to fi re, while Wagstaff, S.J. and Garnock- Zealand. the New Zealand fl ora has almost Jones, P.J. (1998). Evolution What about New Caledonia? Isn’t the none of those adaptations – features and biogeography of the Hebe fl ora of New Caledonia said to be an like microphyllous leaves and the complex (Scrophulariaceae) ancient piece of Gondwana? How did ability for burnt plants to resprout. inferred from ITS sequences. a species radiation happen there? New Zealand Journal of Botany Q: Does this explain why the onset 36: 425–437. Steve: People are now challenging of the Pleistocene glaciations seems Wagstaff, S.J. and Wardle, P. (1999). that perception of New Caledonia, to have been the time when much of Whipcord hebes: systematics, and take the view that while some of our current fl ora had its beginnings in distribution, ecology and the fl ora is very old other elements New Zealand? evolution. New Zealand Journal of have arrived by long distance Botany 377: 17–39. Steve: The aridifi cation of Australia dispersal in geologically recent times. Wagstaff, S.J. and Garnock- could be an explanation for groups Jones, P.J. (2000). Patterns of Q: How does this work fi t with other with higher morphological species diversifi cation in Chionohebee and work done on the evolution of the diversity in New Zealand than Parahebee (Scrophulariaceae) New Zealand fl ora? Is this the sort Australia. But also there was rapid inferred from ITS sequences. of distribution pattern seen in other speciation in New Zealand throughout New Zealand Journal of Botany groups? Is the age of this group the Pleistocene, and that can be 38: 389–407. similar to others in New Zealand? seen in a number of groups of Wagstaff, S.J. and Breitwieser, I. plants, particularly in the alpine zone. Steve: They are good questions. (2002). Phylogenetic relationships That shift in climate has been New I think we’ve generated enough of New Zealand Asteraceae Zealand’s gain and Australia’s loss. independent studies on the New inferred from ITS sequences. Zealand fl owering plant fl ora to show Q: It seems that to understand the Plant Systematics and Evolution a common theme on the origin and New Zealand fl ora we need to relate it 231: 203–224. time of its arrival. There are two clear to the climate history of the region? Wagstaff, S.J. and Breitwieser, patterns: I. (2004). Phylogeny and Steve: Yes, something that was classifi cation of Brachyglottis 1. A large proportion of the New apparent in doing this work is that it’s (Senecioneae: Asteraceae): Zealand fl ora arrived by long vital that we study the New Zealand an example of a rapid species distance dispersal at the beginning fl ora in a wider regional context. We radiation in New Zealand. of the Pleistocene and radiated can make incorrect interpretations of Systematic Botany 29: 1003– after those dispersal events into the New Zealand fl ora if we study it in 1010. the fl ora we see now. isolation. The wider context allows us Winkworth, R.C.; Wagstaff, S.J.; to appreciate how unique our fl ora is, Glenny, D.; Lockhart, P.J. (2002). 2. There is a disparity between the but also how its history is strongly tied Plant dispersal N.E.W.S from New diversity in morphological forms to the fl oras of nearby countries. Zealand. Trends in Ecology and seen in New Zealand fl owering Evolution 177: 514–520. plants and the low level of genetic difference between the species.

New Zealand Garden Journal, 2008, Vol. 11(1) 18 Box 1 Species diversity in Dracophyllumm and related genera Family Ericaceae Subfamily Epacridoideae (the Southern Hemisphere part of the family) Tribe Richeeae Dracophyllum subgenus Dracophyllum New Caledonia: 8 species New Zealand: 7 species eastern Australia: 4 species Lord Howe Island: 1 species Tasmania: 1 species subgenus Oreothamnus New Zealand: 28 species Tasmania: 1 species (but probably misplaced in this subgenus) subgenus Cordophyllum New Caledonia: 1 species Richea Tasmania: 9 species Eastern Australia: 2 species Sphenotoma Western Australia: 8 species

Box 2 Summary of the fi ndings • Tribe Richeeae (Dracophyllum, Richea, Sphenotoma) is monophyletic. • Sections & subgenera • New Zealand Dracophyllumm subg. Oreothamnuss is monophyletic. • Tasmanian D. minimumm may be misplaced. • Dracophyllum strictumm is sister to D. subg. Oreothamnus. • Dracophyllum subg. Dracophyllumm is polyphyletic. • Dracophyllum subg. Cordophyllumm is nested among New Caledonian spp. of D. subg. Dracophyllum. • The two Richea clades correspond with R. sect. Cystanthee and R. sect. Dracophylloides. • Genera • Dracophyllumm is paraphyletic. • Richeaa is polyphyletic. • Sphenotomaa is monophyletic. • Biogeography • While most of the morphological species diversity is found in New Zealand and New Caledonia, the phylogenetic diversity is greatest in Australia. • Australian species of Dracophyllum are remnants of older lineages; their present distributions are fragmented and disjunct. • In contrast, New Zealand (esp. subg. Oreothamnus) and New Caledonian species have recently radiated following long-distance dispersal from Australia. • Evolution in the group was spurred by geological and climatic changes during the late Tertiary.

New Zealand Garden Journal, 2008, Vol. 11(1) 19