You don't have to be a specialist to know that flowering plants (angiosperms) divide into monocots and dicots - a distinction immediately apparent when a seed germinates. Unfortunately, on closer inspection this simple distinction turns out not to be so simple after all.
Current gene-based research suggests the way in which plants are understood and presented to the public needs to be substantially revised. This was the motivation for a recent conference at the National Botanic Garden of Wales, organised by PlantNetwork, an umbrella group for bota-nic gardens and plant collections in Britain and Ireland.
Royal Botanic Gardens, Kew head of genetics Mike Fay says: "The distinction between monocots and dicots is seen as basic, yet we now know that plants like magnolias, which are dicots, are more closely related to monocots than to other dicots."
Along with Magnolia, a very mixed bag of plant families including the peppers, laurels, water lilies and nutmegs now belong to this group. This is just one consequence of a major reordering of the family tree of plants conducted by the Angiosperm Phylogeny Group (APG), of which Fay's team at Kew's Jodrell Laboratory is a member.
The revised classifications are based largely on the comparison of plant DNA. "In the 1990s there was a revolution in easy DNA splicing," says Fay. "It meant that you no longer had to worry about things like flowers. Although often different data give you the same answers."
For example, an analysis of pollen supports its findings, Fay adds. The so-called "eudicots" (literally "true dicotyledons") show a common "triaperturate" or "triculpate" form, ie having three openings, whereas the monocots and their newly discovered relatives have "monocolpate" or "monosulcate" pollen, ie with a single groove.
Fay explains why it is worth making the effort to get plant classification right. "By understanding the relationships, you can better predict likely plant behaviour," he says. "You can standardise the naming."
Ireland's National Botanic Gardens herbarium curator Matthew Jebb adds that the behaviour of pathogens can also indicate a hitherto misunderstood relationship, so understanding such relationships can help anticipate problems.
"The fungal pathogen Phloeospora heracles attacks both giant hogweed (Heracleum mantegazzianum) and the common parsnip (Pastinaca sativa)," he says. "Without understanding the plant relationship, you might have thought the fungus would be a candidate biological control agent."
All this has consequences for how plants are presented to the public. Traditionally, botanic gardens and herbaria have displayed plants in "order beds" - a term dating from Bentham & Hooker's 19th-century description of around 200 "natural orders", which roughly correspond to plant families, rather than orders in the modern sense.
"Several herbaria are reorganising their collections - a consensus is growing," says Fay.
Indeed, when the University of Bristol Botanic Gardens was moved to its current site closer to the main university campus, it was seen as an opportunity to put the new insights into practice, says its director, Professor Simon Hiscock. "There was much debate about how formal a structure the Angiosperm Phylogeny Display should have," he says. An informal curved "tree of paths" layout was settled on, with planting plans laid out in accordance with APG II.
"We have had a delegation from Nanjing Botanic Garden in China who plan to create an APG garden 15 to 20 times the size of ours," he adds.
The Royal Botanic Garden Edinburgh also took the opportunity of a renovation of its three-million-specimen herbarium to make the change to the APG system. "It's not as difficult as actually doing the research," says head of horticulture Dr David Rae.
"After much debate, the motion to change was won only narrowly," he says. "But we use the same database, BG-BASE, as the living collection, so if we were to retain full integration between the two it would have to have happened anyway."
The re-engraving of nearly 4,000 plant labels not only in the Edinburgh garden, but also its outstations at Benmore, Dawick and Logan, cost only £5,000 — less than half of what was budgeted — but took nearly 600 hours of staff time, he adds.
Rae says the new system will be easier for staff, students and visitors, and that other gardens should consider following suit. "My advice is, just do it," he says.
"Don't worry about any adverse public reaction — there won't be any."
The University of Oxford Botanic Garden has also taken the plunge and reordered its beds. Britain's oldest botanic garden, it has incorporated both the Linnaean and Bentham and Hooker taxonomies in its time.
Hardy-collection curator Tom Price says: "We thought it would take three years, but it took 18 months. Some of the plants are only now recovering from the move. But some annual beds, which only looked good four or five months in the year, now have more woody plants. We want it to be both botanically accurate and visually appealing."
Matthew Jebb, who chairs PlantNetwork, also sees the new system as an opportunity for the Glasnevin garden. "We can tell interesting stories to the visitors to our gardens," he points out.
"Everyone knows nettles and roses, yet until recently not even experts knew how closely they were related. With good interpretation we can make people excited about this."
CLASSIFICATION THROUGH THE AGES
1540s Order beds are a feature of first botanic gardens in northern Italy, but there is as yet no agreement on the basis for plant relationships.
1754 Linnaeus publishes a system dividing the plant kingdom into 25 classes.
1859 Darwin's On the Origin of Species links taxonomy with evolutionary history for the first time, proposing that taxonomy "must be strictly genealogical in order to be natural".
1862-1883 Bentham and Hooker, emulating an earlier French classification by de Candolle, publish their three-volume taxonomy, dividing plants into gymnosperms, monocotyledons and three classes of dicotyledons. "All pre-APG systems make gaffes, but Bentham and Hooker is still 90 per cent congruent with APG," says Matthew Jebb.
1998 The first Angiosperm Phylogeny Group - APG I - proposes a radical reclassification of plant taxonomy, with 462 families of plants in 40 orders. The announcement makes front-page news in The Independent. "There was a lot of interest in what might seem a fairly dry topic," recounts Mike Fay.
2003 APG II creates five new orders and over 70 families that had been unplaced in APG I, while conflating others.
2009 A further minor update, APG III, will be published shortly. "After a vote, it did away with several small, difficult orders," says Fay.
PHYLOGENY INTO PRACTICE
First opened only in 2000, the National Botanic Garden of Wales was fortunate to have a blank canvas to work on. And the restoration of its walled garden also benefited from coinciding with the appearance of the APG scheme.
Its walled garden reopened to the public in 2003, a tropical glasshouse was incorporated into the display in 2007, and it continues to be developed.
Senior horticulturist at the garden Laura Davies says: "We wanted to marry science, art and horticulture, and it was a fantastic opportunity to develop the garden alongside the new classification scheme."
Three of the walled garden's quadrants are now given over to the Evolution of Flowering Plants display, which is laid out in line with APG III - the remaining quarter housing a kitchen garden, which Davies describes as "a sop to historic gardeners".
Public engagement with the display is enhanced with a five-minute video shown at the entrance and daily talks from April to October.
"An order bed should be attractive," Davies says, with contrasts in colours and foliage, and seasonal interest from bulbs and annuals. "Even if you don't understand the botany, it should still stand on its own horticultural merit."